Cytokine-induced stable neuronal differentiation of human bone marrow mesenchymal stem cells in a serum/feeder cell-free condition

Authors


*Author to whom all correspondence should be addressed.
Email: h.tao@garvan.unsw.edu.au

Abstract

The characteristics and multilineage differentiation potential of bone marrow mesenchymal stem cells (BM MSC) remain controversial. This study aimed to characterize human BM MSC isolated by plastic adherent or antibody selection and their neuronal differentiation potential using growth factors or chemical inducing agents. MSC were found to express low levels of neuronal markers: neurofilament-M, β tubulin III, and neuron specific enolase. Under a serum- and feeder cell-free condition, basic fibroblast growth factor, epidermal growth factor, and platelet-derived growth factor induced neuronal morphology in MSC. In addition to the above markers, these cells expressed neurotransmitters or associated proteins: γ-aminobutyric acid, tyrosine hydroxylase and serotonin. These changes were maintained for up to 3 months in all bone marrow specimens (N = 6). In contrast, butylated hydroxyanisole and dimethylsulfoxide were unable to induce sustained neuronal differentiation. Our results show that MSC isolated by two different procedures produced identical lineage differentiation with defined growth factors in a serum- and feeder cell-free condition.

Introduction

The adult bone marrow (BM) compartment contains a population of stem cells, which have extensive capacities for self-renewal and proliferation. Under appropriate conditions, they are able to give rise to multiple cell lineages, such as osteoblasts, adipocytes, chondrocytes, myoblasts and hepatocytes, that can be further developed into a broad range of non-hematopoietic tissues (Tao & Ma 2003). These stem cells have been named as mesenchymal stem cells (MSC), multipotent adult progenitor cells (MAPC), or bone marrow stromal progenitor cells (Gronthos & Simmons 1996; Pittenger et al. 1999; Reyes et al. 2001). However, it is unclear if these cells defined by in vitro assays represent an identical cell population.

Over the last few years, BM MSC derived from both rodents and humans have been shown to differentiate into neuronal-like cells in response to chemical reagents or growth factors (Sanchez-Ramos et al. 2000; Woodbury et al. 2000; Deng et al. 2001; Kabos et al. 2002; Kim et al. 2002; Jiang et al. 2003; Jin et al. 2003; Joannides et al. 2003; Hermann et al. 2004). These in vitro generated cells have been shown to possess some morphological and phenotypic characteristics of neuronal cells normally resident in the central nervous system. However, variability currently exists between laboratories in the efficiency and stability of BM MSC-derived neuronal differentiation, especially when originating from human BM MSC. This variability is most likely due to differences in the starting cell populations, inducing reagents and differentiation media used.

The aim of this study was to characterize these cells and their neuronal differentiation potential. BM MSC were isolated from multiple human bone marrow samples using both plastic adherent and immunomagnetic bead selection methods. Under a serum- and feeder cell-free condition, BM MSC were induced to differentiate towards the neuronal lineage using a combination of basic fibroblast growth factor (bFGF), epidermal growth factor (EGF) and platelet-derived growth factor (PDGF). This was compared with that of chemical induction reagents: butylated hydroxyanisole (BHA) and dimethylsulfoxide (DMSO) as described previously (Woodbury et al. 2002; Woodbury et al. 2000). The differentiated cells were then characterized by immunophenotyping, Western blot and reverse transcription–polymerase chain reaction (RT–PCR) analysis.

Materials and Methods

Human bone marrow collection

Bone marrow collections were performed following written informed consent and approval by the Human Research Ethics Committee of St Vincent's Hospital Sydney. BM specimens were collected from the posterior iliac crest under local anesthesia. Approximately 0.5 mL of BM was collected from six hematologically normal donors (age range 29-63 years) and used in the study (Table 1).

Table 1.  Donor and cell culture characteristics in the determination of colony-forming efficiency of human bone marrow mesenchymal stem cells
A
Plastic selectedAgeSexNo. of colony/5 × 106 mononuclear cells (MNC)
Sample 163M5
Sample 229F1
Sample 355M2
B
Bead selectedAgeSex% CD45/G-ANo. of Colony/5 × 104
cells of MNCCD45/G-A cells
Sample 139F0.7510
Sample 249M0.84 2
Sample 331F1.10 1

Isolation and expansion of BM MSC

Bone marrow specimens were diluted with Ca++/Mg++-free phosphate buffered saline (PBS, Thermo Trace, Noble Park, Australia). Mononuclear cells (MNC) were separated by Ficoll-Paque (Pharmacia, Uppsala, Sweden) density (1.077 g/cm3) gradient centrifugation. Two isolation methods were employed to isolate MSC.

Plastic adherent selection

Bone marrow MNC were suspended in complete culture medium containing Dulbecco's modified Eagle's medium-low glucose (DMEM-LG; Gibco, Invitrogen, Carlsbad, CA, USA), 10% heat inactivated fetal bovine serum (FBS; Thermo Trace), 100 U/mL penicillin and 100 µg/mL streptomycin (Gibco). Five milliliters of a 1 × 106 cells/mL cell suspension were plated in 25 cm2 culture flasks (BD Scientifics Discovery Labware, Bedford, MA, USA) and incubated at 37°C with 5% humidified CO2. After 24 h, non-adherent cells were discarded, and adherent cells were thoroughly washed twice with PBS. Fresh complete culture medium was added and replaced every 3-4 days. Colonies consisting of 50 or more fibroblast-like cells were counted 14 days after plating. When the culture reached 80-90% confluency, the cells were harvested using 0.05% trypsin/0.53 mm ethylenediamine tetraacetic acid (EDTA) (Gibco). After washing, BM MSC were re-plated in complete culture medium at a concentration of 1 × 105 cells/mL and termed as passage 1. The cells were subcultured weekly at a 1:2 dilution under the same culture conditions. MSC beyond passage 6 were used for neuronal differentiation.

Immunomagnetic bead selection  Bone marrow MSC were isolated from MNC populations by negative selection using anti-CD45 and antiglycophorin-A (G-A) magnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany). After magnetic activated cell sorting (MACS; Miltenyi Biotec), CD45/G-A cells were suspended in an expansion medium consisting of 60% DMEM-LG, 40% MCDB-201 (Sigma-Aldrich, St. Louis, MO, USA), 1 × insulin transferrin selenium, 1 × linoleic acid bovine serum albumin, 10−9 M dexamethasone, 10−4 M ascorbic acid 2-phosphate (Sigma-Aldrich), 100 U/mL penicillin, 100 µg/mL streptomycin and 10% FBS. Five milliliters of cell suspension at a density of 1 × 104 cells/mL were plated per 25 cm2 of a culture flask precoated with 10 ng/mL fibronectin (Gibco), and grown in the expansion medium. Fresh complete culture medium was added and replaced every 2-3 days. Colonies consisting of 50 or more fibroblast-like cells were counted at day 14 of culture. At 80-90% confluency, cells were recovered by trypsinization as described above.

Characterization of BM MSC by flow cytometry

Ex vivo expanded BM MSC obtained by both plastic adherent and immunomagnetic bead selection methods were characterized by flow cytometry. In brief, cells were trypsinized and washed with fluorescence activated cell sorting (FACS) buffer, consisting of Ca++/Mg++-free PBS, 13.6 mm tri-sodium citrate and 1% bovine serum albumin (BSA, Sigma-Aldrich). After centrifugation at 100 g for 5 min at 4°C, cells were suspended in FACS buffer at a concentration of 1 × 106/mL and incubated with normal human AB plasma at a final dilution of 1:10 at 4°C for 30 min to block non-specific binding. After washing, less than 1 × 106 cells were resuspended in 50 µL of FACS buffer for each antibody tested. Cells were then labeled with 10 µL of fluorescein conjugated monoclonal antibody at 4°C for 30 min. The antibodies used in this study included CD14, CD29, CD34, CD44, CD45, CD71, CD106 (BD Pharmingen, San Diego, CA, USA) and G-A (Dako, Carpinteria, CA, USA). The labeled cells were analyzed by flow cytometry (Beckman Coulter, Fullerton, CA, USA) within 2 h. At least 20 000 cells for each sample were acquired and analyzed.

Neuronal differentiation

Growth factors

Human BM MSC obtained by both isolation methods were maintained in a neural differentiation medium consisting of DMEM/nutrient mixture F12 (DMEM/F12; Gibco), 25 µg/mL insulin, 100 µg/mL transferrin, 100 µm putrescine, 0.02 µm progesterone, and 0.03 µm sodium selenite (Sigma-Aldrich). The medium was supplemented with 10 ng/mL basic fibroblast growth factor (bFGF), 10 ng/mL epidermal growth factor (EGF) and 1 ng/mL PDGF (R&D Systems, Minneapolis, MN, USA). Five milliliters of cell suspension at a density of 1 × 104/mL was plated in a 25 cm2 culture flask precoated with 10 ng/mL fibronectin. The cultures were incubated at 37°C in 5% CO2 atmosphere with daily replenishment of growth factors and medium change every 2 days. The cells were subcultured every 1-2 weeks. Cultures were harvested for characterization at between 2 weeks to 3 months post-differentiation.

Chemical reagents

Prior to neuronal induction, human BM MSC obtained using both selection methods were maintained in DMEM-LG with 10% FBS, 1 mmβ-mercaptoethanol (2-ME; Sigma-Aldrich) and 10 ng/mL bFGF for 24 h. Following pre-induction, the cells were washed with PBS and maintained in DMEM-LG with 2% dimethylsulfoxide (DMSO; Sigma-Aldrich) and 200 µm butylated hydroxyanisole (BHA; Sigma-Aldrich). Cells were harvested for characterization at various time points ranging from 6 h to 7 days post-induction. As a positive control, human neuroblastoma SK-N-SH cells (ATCC, Rockville, MD, USA) were treated with 3 µm all-trans-retinoic acid (RA; Sigma-Aldrich) for 7 days (Lombet et al. 2001).

Characterization of neuronal differentiation

Immunofluorescence staining

Untreated and treated human BM MSC were fixed with 4% paraformaldehyde (Sigma-Aldrich) at room temperature for 20 min. Cells were permeabilized using 0.1% Triton X-100 (Sigma-Aldrich) in PBS for 20 min. After blocking with 10% normal goat serum and 1% BSA in PBS at room temperature for 20 min, slides were incubated with primary antibody at 4°C overnight. Following a wash, the slides were further incubated with Alexa Fluor dye coupled antimouse or antirabbit IgG antibodies (Molecular Probes, Eugene, OR, USA) at room temperature for 45 min. Slides were mounted in Vectashield mounting medium with 4,6-diamino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA) counterstain for visualization. The cells were examined under an Axioskop fluorescent microscope (Carl Zeiss, Jena, Germany) and analyzed using Cytovision 2.7 software (Applied Imaging International, Newcastle Upon Tyne, UK).

Antibodies against β tubulin III (1:250), microtubule-associated protein 2 (MAP2, 1:100), γ-aminobutyric acid (GABA, 1:50), tyrosine hydroxylase (TH, 1:50), serotonin (1:100) and 2′,3′-cyclic nucleotide 3′-phosphohydrolase (CNPase; 1:100) were from Sigma-Aldrich. Antibodies against neuron specific enolase (NSE; 1:100), neurofilament-M (NF-M; 1: 2000), neuron-specific nuclear protein (NeuN; 1:100) and tau (1:100) were from Chemicon International, Temecula, CA, USA. The antibody against glial fibrillary acidic protein (GFAP; 1:50) was from Dako.

Western Blot

Cell lysates were prepared using standard procedures. Protein content was quantified using the BCA protein assay kit (Pierce, Rockford, IL, USA). Ten micrograms of total protein from each lysate was separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and transferred electrophoretically to a polyvinylidene fluoride (PVDF) transfer membrane (NEN Life Science Products, Boston, MA, USA). Immunodetection was performed with rabbit anti-NSE (1:1000, Chemicon) and detected with a horseradish peroxidase conjugated goat antirabbit IgG (Chemicon). The membranes were visualized using enhanced chemiluminescence (ECL Western blotting detection reagents; Amersham Pharmacia Biotech, Piscataway, NJ, USA). Approximately 0.5 µg human brain extract (BD Biosciences Clontech Laboratories, Palo Alto, CA, USA) was used as an internal control. The same blot was stripped and probed for α-tubulin (Sigma) expression.

Reverse transcriptase–polymerase chain reaction

Total RNA was extracted using Trizol reagent (Gibco) and RNeasy mini kit (Qiagen, Hilden, Germany). Advantage RT-for-PCR kit (BD Biosciences Clontech) was used for reverse transcription (RT; 1 µg RNA per condition). Polymerase chain reaction (PCR) conditions and cycle numbers for a linear amplification range were determined and optimized. Primers used are shown in Table 2. Actin was used as the housekeeping gene. The gene expression level was quantitated using NIH ImageJ program (National Institutes of Health, USA). The expression level of house-keeping gene was defined as 1.0. The expression ratio of BM MSC-derived neural specific genes to house-keeping gene was determined.

Table 2.  Sequences of oligonucleotides used for reverse transcription–polymerase chain reaction (RT–PCR) analysis
  Primer Sequence 5′-3′  
GeneAccession no.*ForwardReverseNo. of cycleAnnealing temp °C
  • *

    Accession numbers are derived from GenBank/EMBL/DDBJ database for the sequence information of targeted genes.

ActinNM001614cgcaccactggcattgtcatgtggccatctcctgctcgaa2261
NF-MY00067gagcgcaaagactacctgaagacagcgatttctatatccagagcc3253
MAP2NM002374tggctctctgaagaacatcgacagtggatctgcctggggactgtgt3265
MBPBC008749tgataggccttccaaagagtgtctggagttttgttctttg3360
GFAPNM002055tcatcgctcaggaggtccttctgttgccagagatggaggtt3556
GADNM000817agtaaagatggtgatgggatgccagcagttgcattgacataa3560

Results

Isolation of BM MSC

In this study, the isolation of BM MSC from hematologically normal donors was achieved by plastic adherence or immunomagnetic bead selection (Table 1A,B). Our data showed that using the plastic adherence method, the frequency of BM MSC obtained at day 14 ranged from 0.2-1 per 1 × 106 mononuclear cells (Table 1A), which is consistent with previous reports (Gronthos & Simmons 1996; Gronthos et al. 2003). Using immunomagnetic bead selection, hematopoietic CD45+/G-A+ cells were depleted via immunomagnetic beads and magnetic sorting. The CD45/G-A cells obtained constituted 0.7-1.1% of BM MNC (Table 1B). The numbers of MSC colonies derived from CD45/G-A cells ranged from 1-10 per 5 × 104 cells (Table 1B). Thus, the data showed that the two separation procedures isolated similar numbers of MSC (Table 1A,B).

Characterization of BM MSC

To investigate if progenitor cells obtained from these two methods represent the same cell type, the isolated BM MSC were characterized at passage 1. No morphological differences between the two populations were observed (see Fig. 1A,B).

Figure 1.

Representative morphological features of human bone marrow mesenchymal stem cells (BM MSC) pre- and post-neuronal differentiation. (A) Human BM MSC selected by the plastic adherence method were grown in culture at passage 1 without neuronal induction. (B) Human BM MSC selected by depleting CD45 and G-A positive cells using magnetic beads and grown in culture at passage 1 without neuronal induction. (C) BM MSC exposed to 10 ng/mL basic fibroblast growth factor (bFGF), 10 ng/mL epidermal growth factor (EGF) and 1 ng/mL platelet-derived growth factor (PDGF) for 3 months. (D) BM MSC exposed to 10 ng/mL bFGF, 10 ng/mL EGF and 1 ng/mL PDGF for 3 months, labeled with β tubulin III and immunofluorescence-conjugated antibody. (E) BM MSC treated with 2% dimethylsulfoxide (DMSO) and 200 µm butylated hydroxyanisole (BHA) for 30 h. (F) BM MSC treated with 2% DMSO and 200 µm BHA for 48 h (magnification ×100).

In this study, the immunophenotype of MSC obtained from the six BM samples was analyzed by flow cytometry. BM MSC isolated by both plastic adherence and immunomagnetic beads were negative for hemopoietic markers: CD34, CD45, CD14 and G-A, but strongly positive for CD29 and CD44, and weakly positive for CD71 and CD106 (Fig. 2).

Figure 2.

Representative immunophenotypic analysis of human BM MSC by flow cytometry. Human BM MSC at passage 1 were labeled with fluorescence-conjugated monoclonal antibodies. Twenty thousand cells from each sample were analyzed.

Bone marrow MSC were further characterized by immunofluorescence staining, RT-PCR and Western blot analysis. By immunofluorescence staining, undifferentiated BM MSC (passage 1) weakly expressed the following neuronal protein markers: β tubulin III and NF-M, and the oligodendrocyte marker, CNPase (Fig. 3A and Table 3). Western blot analysis also demonstrated low-level expression of NSE in these cells (Fig. 4). RT-PCR analysis confirmed that MSC expressed neuronal genes, with weak expression of NF-M and MAP2, as well as weak expression of the oligodendrocyte gene MBP (Fig. 5A,B).

Figure 3.

Representative images of BM MSC-derived neuronal cells characterized by immunofluorescence staining. (A) Cells were labeled with primary antibody against neuronal markers: β tubulin III, neurofilament-M (NF-M), neuron-specific nuclear protein (NeuN), microtubule-associated protein 2 (MAP2), tau or neuron specific enolase (NSE), followed by immunofluorescence-conjugated secondary antibodies. (1) Untreated passage 1 BM MSC. (2) BM MSC exposed to 10 ng/mL bFGF, 10 ng/mL EGF and 1 ng/mL PDGF for 5 weeks. (3) BM MSC treated with 2% DMSO/200 µm BHA for 30 h. 4,6-diamino-2-phenylindole (DAPI) counterstain was used for visualization (magnification ×200). (B) BM MSC exposed to 10 ng/mL bFGF, 10 ng/mL EGF and 1 ng/mL PDGF for 5 weeks and characterized by double immunofluorescence staining. TUB, β tubulin III; SER, serotonin (magnification ×200).

Table 3.  Phenotypic characterization of neuronal-like cells by immunofluorescence staining
Antigenic markerBone marrow undifferentiatedBone marrow cytokine 5 weeksBone marrow BHA/DMSO 30 hSK-N-SH RA 7 days
  1. Data shown summarizes results for all six BM samples. +, weak positive staining ranging between 20–40% of cells. ++, intermediate positive staining ranging between 40–60% of cells. +++, intensive positive staining of greater than 60% of cells. BHA, butylated hydroxyanisole; DMSO, dimethylsulfoxide; RA, all-trans-retinoic acid.

NSE++++++
β Tubulin III+/–+++++++++
NF-M+/–+++++++++
MAP2++++++
Tau+++++++++
GFAP
CNPase+/–+++
GABA++++++
TH++++ND
Serotonin++++++
CD44++++++
Figure 4.

Representative protein expression of BM MSC-derived neuronal cells by Western blot analysis. The expression of NSE and α-tubulin was monitored. (A) Lymphocytic cell line. (B) Untreated BM MSC. (C) BM MSC exposed to 10 ng/mL bFGF, 10 ng/mL EGF and 1 ng/mL PDGF for 5 weeks. (D) BM MSC treated with 2% DMSO and 200 µm BHA for 30 h.

Figure 5.

Representative gene expression of BM MSC-derived neuronal cells by RT-PCR analysis. Neural specific genes: NF-M, MAP2, GAD and MBP were monitored. Actin was used as a house-keeping gene. (A) Gene expression of BM MSC-derived neuronal cells by reverse transcription–polymerase chain reaction (RT–PCR) analysis. (B) The relative gene expression level was shown as the ratio of neural specific genes to house-keeping gene which is defined as 1.0. (1) Untreated BM MSC; (2) BM MSC treated with 2% DMSO and 200 µm BHA for 30 h; (3) BM MSC exposed to 10 ng/mL bFGF, 10 ng/mL EGF and 1 ng/mL PDGF for 5 weeks.

Cytokine induced neuronal differentiation

Subconfluent BM MSC were cultured in serum-free neural differentiation medium and maintained in fibronectin-coated cultureware without feeder cells. Morphological changes in BM MSC were noted after 2 weeks of culture with daily addition of bFGF, EGF and PDGF. The changes progressed over the subsequent 2-3 weeks and were maintained for up to 3 months. The majority (80-95%) of cells in both plastic and bead selected samples showed morphological changes that included retraction of the cytoplasm towards the nucleus, and formation of refractile cell bodies with long branching processes (Fig. 1C). There was no evidence of cell proliferation after commencement of differentiation. No observation of cell death or reversible differentiation was detected in any of the samples for up to 3 months.

To determine whether this differentiation procedure was reproducible, we repeated the same differentiation protocol on samples collected from a further five individuals. Similar morphological and phenotypic changes were observed in all samples.

Characterization of neuronal differentiation by immunofluorescent staining and Western blot analysis

To characterize the morphologically differentiated BM MSC, immunofluorescence staining was carried out on pretreated passage 1 MSC and on cells exposed to bFGF, EGF and PDGF for 5 weeks. The expression of BM MSC marker CD44 declined, whereas, the expression of neuronal markers (NSE, β tubulin III and NF-M) and mature neuronal markers (MAP2 and tau) increased in all BM MSC cultures treated with cytokines when compared with the untreated cells (Figs 1D and 3A and Table 3). The expression of β tubulin III was found in majority of differentiated cells (80-90%), NF-M was observed in 66-70% of the cells, while only 32-36% expressed the mature neuronal marker MAP2 (Table 4).

Table 4.  Quantitation of neuronal-like cells by immunofluorescence staining
Antigenic marker% of positive cells
  1. Bone marrow mesenchymal stem cells (BM MSC) exposed to 10 ng/mL basic fibroblast growth factor (bFGF), 10 ng/mL epidermal growth factor (EGF) and 1 ng/mL platelet-derived growth factor (PDGF) for 5 weeks. The frequency of positively stained cells was estimated by counting cells in random visual fields and a minimum of 200 cells for each sample was counted. The results presented are the mean ± SEM derived from triplicates of six samples.

NF-M70.60 ± 28.82
MAP236.33 ± 14.83
β Tubulin III83.48 ± 5.09
NF-M/GABA30.00 ± 12.25
NF-M/TH41.40 ± 16.90
β Tubulin III/Serotonin62.54 ± 7.95

Double immunofluorescence staining showed that 30% of the cells in samples treated by cytokines for 5 weeks expressed both NF-M and GABA, approximately 40% of cells expressed both NF-M and TH, and 60-70%β tubulin III positive cells were found to be positive for serotonin (Fig. 3B and Table 4). Our data suggest that GABAergic, dopaminergic and serotonergic neurons were generated in vitro.

The results of our immunofluorescence study were confirmed by Western blot analysis. Undifferentiated BM MSC expressed low levels of NSE and following cytokine exposure the expression levels of NSE were increased in all BM samples (Fig. 4).

In contrast to the expression of neuronal specific markers detected, immunofluorescence staining showed a weak expression of the oligodendrocyte specific marker CNPase in untreated BM MSC (passage 1) and a slight increase in expression of CNPase in all samples treated by both methods. However, none of the samples studied showed expression of the classical glial astrocyte marker, GFAP.

Gene expression analysis by RT-PCR

Bone marrow MSC were collected and subjected to RT-PCR analysis before neuronal induction (passage 1) as well as post-exposure to bFGF, EGF and PDGF for 5 weeks (Fig. 5A). The relative gene expression level was shown as the ratio of neural specific genes to house-keeping gene (Fig. 5B). Table 5 summarizes the expression levels of neuronal or neurotransmitter genes for all samples. The expression of neuronal genes NF-M and MAP2 was increased after treatment with growth factors. Increased gene expression was also observed in all the samples for glutamic acid decarboxylase (GAD), the rate-limiting enzyme in the synthesis of GABA. Additionally, a weak increase in the expression level of oligodendrocyte gene MBP was seen in all samples. This is consistent with the immunofluorescence staining results, which also showed a slight increase in expression of the oligodendrocyte marker CNPase in all samples. The transcript of astrocyte lineage-associated GFAP was undetectable by RT-PCR. Overall, gene expression analysis supported the results obtained by immunofluorescence staining, confirming that cells with a neuronal phenotype were the predominant cell type obtained.

Table 5.  Characterization of neuronal differentiation by gene expression analysis using RT-PCR
GeneBM undifferentiatedBM cytokine 5 weeksBM BHA/DMSO 30 hSK-N-SH RA 7 days
  1. Data shown summarizes results for all six BM samples. +, weak expression. ++, intermediate expression. +++, intensive expression.

NF-M+/–+++++++
MAP2+/–+++++++
GAD+++++++
MBP+/–++
GFAP+/–
Actin++++++++

Chemical reagent induced neuronal differentiation

Bone marrow MSC obtained by both isolation methods were subjected to undergo neural differentiation using chemical reagents. By exposing BM MSC to serum-free DMEM-LG medium containing 2% DMSO and 200 µm BHA, morphological changes were apparent within 30 min. These changes progressed over the first 6 h, and were maintained for up to 30 h post-treatment. The morphological changes included retraction of the cytoplasm towards the nucleus in the flat fibroblast-like cells. Subsequently, the cells exhibited a typical neuronal appearance by forming round and refractile cell bodies with peripheral extensions. By 30 h post-treatment, 80-95% of the cells in the plastic adherence selected samples and 70-90% in the bead selected samples responded with a change in morphology as described above (Fig. 1E). We observed that the number of cells had not increased after morphologic changes. Although some phenotypic and genetic markers of neuronal lineages were detected on the cells treated with DMSO/BHA for 30 h (Table 4 and Table 5), after approximately 48 h, there was loss of peripheral branches and detachment of some cell bodies from tissue culture flasks (Fig. 1F). The detached cells were confirmed to be non-viable by the Trypan blue dye exclusion assay, indicating that the phenomenon of neuronal differentiation driven by these chemical induction agents was not typical of physiological development.

Discussion

Isolation of BM MSC using the plastic adherent method was first described by Friedenstein et al. (1968). The precursor cells obtained grew in plastic cultureware and formed colonies capable of giving rise to chondrocytes and osteoblasts. The majority of research groups continue to use this traditional method for isolating BM MSC (Haynesworth et al. 1992b; Gronthos & Simmons 1996; Pittenger et al. 1999; Prockop et al. 2001; Gronthos et al. 2003). Recently, Reyes et al. (2001) have described a strategy using immunomagnetic beads for negative enrichment of these cells, which they named MAPC. In this study, BM MSC were isolated using both plastic adherence and immunomagnetic bead methods.

Bone marrow MSC have been described in several studies as positive for a number of antigenic markers, including Stro-1, SH2, SH3, CD29, CD44, CD71, CD90, CD106, CD120a and CD124 (Simmons & Torok-Storb 1991; Haynesworth et al. 1992a; Pittenger et al. 1999; Reyes et al. 2001). Importantly, MSC have not been found to express antigens of hematopoietic lineages (Pittenger et al. 1999; Reyes et al. 2001).

We found that the two separation procedures isolated similar numbers of MSC having the same morphology and immunophenotype. Furthermore, these cells are capable of differentiating into cells with neuronal phenotype, suggesting that both plastic adherence and immunomagnetic bead isolation procedures enrich the same cell population in BM progenitor cell population.

In this study, MSC derived from six BM samples were found negative for hemopoietic markers: CD34, CD45, CD14 and G-A, but positive for CD29, CD44 and CD106, indicating that the MSC obtained by both plastic adherence and immunomagnetic bead isolation methods possess the consistent phenotype as reported by other investigators (Haynesworth et al. 1992a; Pittenger et al. 1999; Reyes et al. 2001).

Our findings of neuronal gene and protein expression in undifferentiated BM MSC are consistent with the recent report that BM MSC express several neural associated proteins (Tondreau et al. 2004). This observation is supported by gene expression studies using microarray analysis, which demonstrate that BM MSC (Seshi et al. 2003) as well as hematopoietic stem cells (Georgantas et al. 2004; Steidl et al. 2004) express genes of non-mesodermal lineages, that is, the ectodermal and endodermal lineages. Furthermore, isolated single human BM MSC were found to express genes typically associated with osteoblasts, muscle cells, epithelial cells, endothelial cells, neural/glial cells and hematopoietic cells. The results of these studies support the concept of a pluripotent mesenchymal progenitor cell in the BM.

Although the idea of BM MSC as progenitor cells possessing the potential for multilineage differentiation remains controversial, we have provided evidence that these cells are capable of differentiating into cells of the neuronal lineage in vitro in a feeder cell and serum-free environment. In this study, we have successfully defined the conditions optimal for this in vitro induction that is reproducible in multiple adult BM specimens.

The choice of using serum-free medium and fibronectin as a coating agent was based on previous reports of neural differentiation of stem cells derived from the fetal and adult central nervous system (CNS) (Johe et al. 1996). Our data demonstrates that this serum-free medium and coating agent are appropriate culture conditions to induce differentiation of BM MSC towards the neuronal lineage.

Identifying the most suitable cytokines and formulating the best cytokine combinations are fundamental to regulating the in vitro differentiation of BM MSC to the neuronal lineage. Johe et al. (1996) reported that CNS precursor cells, after pretreatment with bFGF for 6 days, differentiated into the three neural cell types: neurons, astrocytes and oligodendrocytes. We applied this approach to human BM MSC but were unable to reproduce their results. We found that partially differentiated BM cells could not be maintained in culture beyond 3 days post-cytokine withdrawal (data not shown), suggesting that neuronal differentiation of human BM MSC under serum-free and feeder cell-free conditions is cytokine dependent. We found the combination of three growth factors to be crucial: bFGF to induce neuronal differentiation, EGF to maintain cell proliferation and differentiation, and PDGF to support neuronal differentiation. Our data suggests that growth factors, in this particular combination, can effectively induce the differentiation of BM MSC into neuronal cells.

Using these factors and culture conditions, changes in morphology were observed in the majority of BM MSC from both isolation methods. The neuronal-like cells obtained expressed increased levels of neuronal markers: NSE, β tubulin III, NF-M, MAP2 and tau. The neuronal developmental process was confirmed by the elevated expression of neuronal genes: NF-M and MAP2. Our data demonstrated that BM MSC, isolated by either method, differentiated predominantly into cells with a neuronal phenotype. These findings are supported by a recent report on neuronal differentiation of isolated murine MSC (Jiang et al. 2003).

Neurotransmitters are considered the terminal differentiation products of neurons. Three types of neurotransmitters or enzymes involved in neurotransmitter production were monitored: GABA (the neurotransmitter of GABAergic neurons), TH (an enzyme involved in synthesis of the neurotransmitter dopamine of dopaminergic neurons), and serotonin (the neurotransmitter of serotonergic neurons). Additionally, enzymes involved in neurotransmitter production were also examined, including GAD (GABA production), and TH (dopamine production). Expression of these enzymes provides further indication that the corresponding neurotransmitters were expressed. Our data demonstrates that neuronal-like cells derived from BM MSC exhibit elevated expression of GABA, GAD, TH and serotonin, indicating that GABAergic, dopaminergic and serotonergic neurons have been generated in vitro.

To determine whether our differentiation procedure is reproducible, we repeated the same differentiation protocol on samples collected from six individuals. Similar findings were seen with all samples. We believe that the event of cell fusion with non-MSC is unlikely since our in vitro neuronal differentiation approach employed purified BM MSC as a starting cell population in a culture system free of feeder cells. Further experiments are required to provide formal proof.

The mechanisms of the neuronal induction by DMSO and BHA are unclear. Our data suggests that the phenomenon of neuronal differentiation driven by these chemical induction agents is a transient event. This is consistent with the results reported by two recent studies (Lu et al. 2004; Neuhuber et al. 2004). The morphological changes and increase in immunopositivity for certain cellular markers of BM MSC in response to BHA and DMSO have been interpreted to be the result of cellular toxicity, cell shrinkage and changes in the cytoskeleton, and do not represent regulated steps in the complex process of cellular differentiation.

In conclusion, we have shown that both plastic adherence and immunomagnetic bead isolation procedures enrich the same cell population in human BM. Undifferentiated BM MSC have been found to express a low level of specific phenotypic or genetic neuronal markers. Under serum- and feeder cell-free culture conditions, human BM MSC developed into cells with neuronal morphology and phenotype when induced by a unique cytokine combination. Cytokine-induced differentiation was also found to be stable for up to 3 months, in contrast to the reported chemical induction method. Our data supports the concept that BM-derived stem cells are capable of differentiating into cells of the neural lineage. The reproducible protocol we have described here will be useful in further defining this process and for the potential applications of BM MSC-derived cells in therapeutic cell replacement.

Acknowledgements

This work was supported by research grants from the Estate of the Late R. T. Hall Trust and St. Vincent's Clinic Foundation. The authors wish to thank their generous support. We acknowledge Dr Yi mo Deng for technical advice on RT-PCR.

Ancillary