SEARCH

SEARCH BY CITATION

Keywords:

  • bone marrow;
  • brain derived neurotrophic factor;
  • calcium signaling;
  • dopamine;
  • mesenchymal stem cells;
  • RE-1 silencing factor

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The generation of dopamine (DA) neurons from stem cells holds great promise in the treatment of Parkinson’s disease and other neural disease associated with dysfunction of DA neurons. Mesenchymal stem cells (MSCs) derived from the adult bone marrow show plasticity with regards to generating cells of other germ layers. In addition to reduced ethical concerns, MSCs could be transplanted across allogeneic barriers, making them desirable stem cells for clinical applications. We have reported on the generation of DA cells from human MSCs using sonic hedgehog (SHH), fibroblast growth factor 8 and basic fibroblast growth factor. Despite the secretion of DA, the cells did not show evidence of functional neurons, and were therefore designated DA progenitors. Here, we report on the role of brain-derived neurotrophic factor (BDNF) in the maturation of the MSC-derived DA progenitors. 9-day induced MSCs show significant tropomyosin-receptor-kinase B expression, which correlate with its ligand, BDNF, being able to induce functional maturation. The latter was based on Ca2+ imaging analyses and electrophysiology. BDNF-treated cells showed the following: increases in intracellular Ca2+ upon depolarization and after stimulation with the neurotransmitters acetylcholine and GABA and, post-synaptic currents by electrophysiological analyses. In addition, BDNF induced increased DA release upon depolarization. Taken together, these results demonstrate the crucial role for BDNF in the functional maturation of MSC-derived DA progenitors.

Abbreviations used
ACh

acetylcholine

BDNF

brain-derived neurotrophic factor

bFGF

basic fibroblast growth factor

BM

bone marrow

c-ret

receptor tyrosine kinase proto-oncogene

D1R

dopamine receptor D1

DA

dopamine

FGF8

fibroblast growth factor 8

GDNF

glial-derived neurotrophic factor

GFRα1

glial-derived neurotrophic factor family receptor alpha 1

Glut-R

glutamate receptor

l-glut

l-glutamate

MSC

mesenchymal stem cell

NGF

nerve growth factor

NT

neurotrophin

PBS

phosphate-buffered saline

SES

external balanced salt solution

SHH

sonic hedgehog

sPSCs

spontaneous post-synaptic currents

Trk

tropomyosin-receptor-kinase

Human stem cells show therapeutic potential for a number of neural disorders (Miller and Bai 2007). Moreover, stem cell therapy has been suggested to be an effective treatment for neurodegenerative disorders such as Parkinson’s disease, which is caused by the degeneration of dopamine (DA) neurons in the substantia nigra (Trzaska and Rameshwar 2007a). DA neurons derived stem cells are significant to cell replacement therapy for degenerated DA neurons of the nigro-striatal system, and also as a model system to study drug reaction and DA developmental processes.

Mesenchymal stem cells (MSCs) show significant potential for clinical applications (Zipori 2004; Giordano et al. 2007). In addition to pluripotency, MSCs secrete a wide variety of trophic factors that could be important for tissue regeneration and are plastic with respect to forming cells of other germ layers, such as neural cells (Cho et al. 2005; Caplan 2007; Giordano et al. 2007; Greco and Rameshwar 2007; Greco et al. 2007). MSCs have been shown to generate DA cells and show promise for Parkinson’s disease in animal models (Dezawa et al. 2004; Fu et al. 2006; Trzaska et al. 2007b). In addition, MSCs are easily isolated and expanded from small samples of bone marrow (BM) aspirate. The unique immune properties of MSCs could allow for transplantation across allogeneic barriers (Bianco et al. 2001; Potian et al. 2003).

We have reported on the induction of a DA phenotype from adult human MSCs with an induction cocktail that comprises sonic hedgehog (SHH), fibroblast growth factor 8 (FGF8) and basic FGF (bFGF) (Trzaska et al. 2007b). The MSC-derived DA cells express DA-specific markers, tyrosine hydroxylase, DA transporter, Nurr1, and Pitx3. However, despite the expression of these DA-specific genes and the synthesis and secretion of DA, they did show characteristics of mature neurons, in particular electrophysiological properties. Thus, the induced cells were considered to be at an immature state and were designated DA neuronal progenitors. The DA progenitors expressed the repressor of neural genes, REST, although at levels lower than the uninduced MSCs (Trzaska et al. 2008). Knockdown of REST led to the generation of electrophysiologically functional neurons (Trzaska et al. 2008).

As MSCs can form functional DA neurons, we sought for exogenous factors that can transition the DA progenitors to functional neurons. In this study, we examine the role of neurotrophins (NTs) in further differentiation of the MSC-derived DA progenitors to functional neurons. We hypothesize that the DA progenitors express receptors for neurotrophic factors and these factors can induce maturation of the cells. The goal was to identify what neurotrophic factor(s) can promote maturation to functional DA neurons. Temporal expression profile of NT receptors showed an increase in the expression of tropomyosin-receptor-kinase (Trk) B at day 9 induction. As brain-derived neurotrophic factor (BDNF) can bind to TrkB, we studied its role in the function of the DA progenitors, by Ca2+ imaging analyses electrophysiology.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Culture of human MSCs

The use of human BM aspirates followed a protocol approved by the Institutional Review Board of UMDNJ-Newark Campus. Informed consent was obtained from all subjects. Healthy volunteers ranged in age from 18 to 35. Briefly, BM aspirates (3 mL) were cultured with 3 mL Dulbecco’s modified Eagle’s medium (Invitrogen, Carlsbad, CA, USA) containing 10% Fetal Calf Sera (Hyclone, Logan, UT, USA) on Falcon 3003 culture dishes. At day 3, mononuclear cells were isolated using Ficoll Hypaque (Sigma, St Louis, MO, USA) density gradient and replaced back into the culture dishes. Fifty percent of fresh media was replaced at weekly intervals until MSCs were ∼80% confluency.

Antibodies and growth factors

Brain-derived neurotrophic factor and NT-3 were purchased from Cell Sciences (Canton, MA, USA); rabbit anti-GABA (α1 subunit), SHH, bFGF, FGF8, and glial-derived neurotrophic factor (GDNF) from R&D Systems (Minneapolis, MN, USA); TrkA, TrkB, TrkC, GDNF family receptor alpha 1 (GFRα1), receptor tyrosine kinase proto-oncogene (c-ret), β-III-tubulin and anti-DA receptor type 1 (D1R) from Santa Cruz Biotechnology (Santa Cruz, CA, USA); nerve growth factor (NGF), pan NaV channel and β-actin from Sigma; secondary antibodies (FITC and phycoerythrin conjugated); nicotinic acetylcholine receptor alpha 1 (nAchRα1) from Abcam (Cambridge, MA, USA); horseradish peroxidase-conjugated secondary antibody from Pierce (Rockford, IL, USA); rabbit anti-glutamate receptor (Glut-R) from Chemicon-Millipore (Burlington, MA, USA).

Neuronal induction of human MSCs

Mesenchymal stem cells were trypsinized and sub-cultured in multiwell plates, or on round glass coverslips, which pre-coated with BD Biocoat poly-d-lysine (BD Biosciences, San Jose, CA, USA). After adherence for 24 h incubation, the MSC medium, Dulbecco’s modified Eagle’s medium (Invitrogen) containing 10% Fetal Calf Sera (Hyclone), was replaced with neurobasal medium and 0.5% B27 supplement (Invitrogen) and with the following induction cocktail: 250 ng/mL SHH, 100 ng/mL FGF8, and 50 ng/mL bFGF. The media were not replaced during the induction period. The induced cells show neuronal morphology and are TH+, βIII-Tub+, NeuN+, and GFAP (Trzaska et al. 2007b).

Immunocytochemistry

Cells were fixed with 2% paraformaldehyde and next permeabilized and blocked with 0.1% Triton X-100, 1%bovine serum albumin in phosphate-buffered saline (PBS). Fixed cells were incubated with primary antibodies overnight at 4°C. Dilutions were as follows: TrkA, TrkB, TrkC, GFRα1, c-ret, 1/200; β-III-tubulin, 1/200; pan NaV channel, 1/500. Nuclear stain 4′,6-diamidino-2-phenylindole (300 nM) and cytoskeletal stain Texas Red Phalloidin (F-actin) (6.6 μM) (Molecular Probes, Carlsbad, CA, USA), were used for detection and visualization.

Western analysis

Whole cell protein lysates were obtained by resuspension of cell pellets in Lysis Buffer (Promega, Madison, WI, USA) and followed by freeze/thaw technique in dry ice/ethanol bath. The human neuroblastoma cell line, SH-Sy5y (ATCC, Manassas, VA, USA), was used as the positive control. Protein extracts (100 μg) were electrophoresed on 4–20% sodium dodecyl sulfate–polyacrylamide gel electrophoresis pre-cast gels (Bio-Rad, Hercules, CA, USA) and transferred onto polyvinylidene difluoride membranes (PerkinElmer, Boston, MA, USA). Membranes were incubated with primary antibodies in 5% milk-PBS overnight at 4°C. Primary antibody dilutions were: D1R, TrkA, TrkB, TrkC, GFRα1 and c-ret at 1/200; Glut-R, nAchα1, and GABA-R at 1/1000; β-actin at 1/1000. After 24 h incubation, membranes were washed and incubated with the appropriate horseradish peroxidase-conjugated secondary antibody (1/2000) for 2 h at 4°C. Membranes were subsequently exposed with chemiluminescence western blot detection reagents (Pierce) and stripped with Restore Stripping Buffer (Pierce) for reprobing.

Neuronal induction of human MSCs – addition of neurotrophins

After 9 days of induction with the initial induction cocktail containing SHH, FGF8, and bFGF, 50 ng/mL of BDNF, NT-3, GDNF, or NGF were added to the cultures for 3 days, for a total of a 12-day induction period. Control cells were incubated for 12 days with only the initial inductive cocktail, without addition of any NTs.

Calcium imaging

Intracellular calcium ([Ca2+]i) transients were studied in MSCs induced with the initial inductive cocktail and upon addition of NTs. Cells were cultured on coverslips and loaded with 5 μM fura-2AM (Molecular Probes) for 30 min at 37°C. After washing the cells with PBS to remove fura-2AM, cells were transferred to a recording chamber containing a external balanced salt solution (SES) for imaging. The SES contained, in mM: 145 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES pH 7.4, and 10 glucose. To induce changes in intracellular Ca2+ levels, bath solution was replaced with SES containing elevated K+ (KCl 50 mM equimolar substitution for NaCl), 100 μM ATP, 1 mM acetylcholine (ACh), 1 mM GABA, or 1 mM l-glutamate (l-glut). For the Ca2+-free SES, Ca2+ was substituted with 4 mM magnesium and 1 mM EGTA. Fura-2 fluorescence images were acquired by alternating the excitation wavelength between 340 and 380 nm using a TILL-Photonics monochramotor and capturing the resultant emissions through a 520 nm long-pass filter, using an Imago CCD camera. Fluorescence images were acquired at 4 s intervals, and the ratio of fluorescence elicited at 340 and 380 nm plotted. Images were developed and analyzed using SigmaPlot software (Chicago, IL, USA).

Electrophysiology

The whole-cell configuration was used to record the electrical activities with an Axopatch 200B amplifier (Molecular Devices Inc., Foster City, CA, USA), a Digidata 1320A analog-to-digital converter (Molecular Devices Inc.), and pClamp 9.2 software (Molecular Devices Inc.). Data were filtered at 1 kHz and sampled at 5 kHz. The external solution contains 140 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM HEPES, and 10 mM glucose (320 mOsm, pH set to 7.3 with Tris base). The patch electrodes had a resistance of 3–5 MΩ, when filled with pipette solution containing: 140 mM CsCl, 2 mM MgCl2, 4 mM EGTA, 0.4 mM CaCl2, 10 mM HEPES, 2 mM Mg-ATP, and 0.1 mM GTP. The pH was adjusted to 7.2 with Tris–base, and the osmolarity was adjusted to 280–300 mOsm with sucrose. Electrophysiological recordings were performed at room temperature (22–24°C). Spontaneous postsynaptic currents were counted and analyzed using Clampfit 9.2 (Molecular Devices Inc.).

Dopamine ELISA

Dopamine levels were acquired and quantitated using an ELISA kit obtained from Rocky Mountain Diagnostics (Colorado Springs, CO, USA) according to the manufacturer’s instructions. After induction, culture media was replaced with the SES. For the Ca2+-free SES, Ca2+ was substituted with 4 mM Mg2+ and 1 mM EGTA. Cells were stimulated with elevated K+ solution (50 mM) to induce depolarization, 100 μM ATP, 100 μM suramin, 1 mM ACh, or 100 μM scopolamine, all purchased from Sigma. Cells were treated for a 5 min period at 37°C, except for ATP, which was treated for 1 min. After this, 100 μL samples were collected and immediately analyzed for DA levels by ELISA.

Statistical analysis

Data is articulated as the mean ± SE. Statistical differences were calculated using anova and pair wise comparisons were evaluated using the post hoc Holms test. < 0.05 were considered significant. n refers to separate experiments.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

TrkB expression in induced MSCs

Mesenchymal stem cells were induced to generate DA progenitors, as described (Trzaska et al. 2007b). We have shown > 90% of the induced MSCs expressed NeuN and beta-III-tubulin and ∼70% expressed tyrosine hydroxylase, the enzyme involved in the biosynthesis of DA. We selected BDNF during the final induction process after we asked whether our previously reported inductive stimuli mediated increases in the expression of neurotrophic receptors. This study focused on neurotrophic factors because this family was implicated in DA development (Zhou et al. 1996; Erickson et al. 2001).

Mesenchymal stem cells were induced as previously described (Trzaska et al. 2007b). At different times after induction we performed Western analyses with extracts to determine if the following NT receptors were expressed: TrkA, TrkB, TrkC, GFRα1, and c-ret (Zhou et al. 1996; Erickson et al. 2001). TrkA, c-ret and TrkC were constitutively expressed (Fig. 1a), which is consistent with the neuroprotective effect of MSCs via low expression of Trk receptors (Pisati et al. 2007). TrkB and GFR1 bands were increased at days 9, 12, and 15 of induction (Fig. 1a, left panel). Despite these increases, normalized band densities show higher expression of TrkB as compared to GFR1 (Fig. 1a, right panel).

image

Figure 1.  Temporal expression profile of neurotrophin receptors in induced MSCs. (a) Representative western blots showing protein expression of neurotrophin receptors TrkA, TrkB, TrkC, GFRα1, and c-ret in MSCs induced for 0, 3, 6, 9, 12, and 15 days (left panel). The SH-SY5Y cell line was used as the positive control. β-actin was the normalizing control. Normalized densities showing differences in protein levels between TrkB (filled circle) and GFRα1 (empty circle) during neuronal induction (right panel). (b) Representative images of MSCs induced for 0, 3, 6, 9, 12, and 15 days. TrkB and GFRα1 protein expression is indicated by the green or yellow fluorescence (FITC). Cells were labeled with the cytoskeletal stain, Texas Red Phalloidin (F-actin) indicated by the red fluorescence, and the nuclear stain DAPI, indicated by the blue fluorescence. (c) Representative immunocytochemistry for β-III-Tubulin (PE) and NaV channel (FITC) in induced neurons with and without BDNF. Cells were counter-stained with DAPI.

Download figure to PowerPoint

We next confirmed the up-regulation of TrkB and GFRα1 (Fig. 1b) by immunocytochemistry. To obtain better visualization, the cells were also labeled for nuclear and cytoskeletal structures with 4′,6-diamidino-2-phenylindole and F-actin, respectively. TrkB was significantly up-regulated at day 9, 12, and 15 inductions as indicated by intense green fluorescence (Fig. 1b). GFRα1 was also up-regulated, but with less fluorescence intensity as compared to TrkB. These observations are consistent with the protein expression observed in the western analyses (Fig. 1a). These results show robust up-regulation of TrkB beginning at day 9 induction. Immunohistochemical studies showed approximately 90% positive cells for TrkB and 60–70% for GFRα1. The fluorescence intensity was weaker for GFR1 as compared to TrkB (Fig. 1b). As a final point, to confirm the expression of neuronal markers, we added BDNF to the induced cells at the time when its receptor is expressed. This resulted in the expressions of βIII tubulin and NaV channel (Fig. 1c), suggesting maturation.

The next set of studies focused on TrkB with its ligand, BDNF. As BDNF-TrkB signaling has been shown to up-regulate the expression of D1R (Do et al. 2007), we first performed western analyses to determine whether BDNF (50 ng/mL) stimulation at day 9 induction caused an increase in D1 receptor. Indeed, there were significantly denser bands for D1R after 6 and 24 h stimulations, as compared to uninduced MSCs (time 0) (Fig. 2a and b). These observations suggest that the cells respond to BDNF by up-regulating D1 receptor protein. These observations suggest functional TrkB receptor, further supporting BDNF-mediated maturation.

image

Figure 2.  D1 receptor expression in induced MSCs. (a) Representative western blots showing protein expression of D1 receptor in uninduced MSCs (day 0) and MSC induced for 9 days (Day 9). ‘−’ indicates no addition of BDNF and ‘+’ indicates addition of BDNF. Cells were stimulated with 50 ng/mL BDNF for 6 or 24 h. β-actin was the normalizing control. (b) Cumulative data showing the differences in normalized protein levels of D1 receptor between day 0, day 9, day 9 + 6 h BDNF, and day 9 + 24 h BDNF-treated MSCs. Normalized band density was presented as mean ± SE, n = 3. *, statistically significant < 0.05 vs. day 9 with no BDNF.

Download figure to PowerPoint

Maturation of induced MSCs by BDNF

In our previous studies, we showed no change in intracellular Ca2+ by 12-day induced MSCs, despite depolarization with elevated K+ extracellular solution (11). As the 9-day induced MSCs robustly up-regulated TrkB and responded to BDNF by inducing the expression of D1 receptor (Figs 1 and 2), we surmised that BDNF might be important for further maturation of MSCs to excitable DA neurons. To this end, we asked whether BDNF could induce the cells to express functional Ca2+ voltage-gated channels, which are necessary for communication between excitable cells.

We treated 9-day induced MSCs with 50 ng/mL BDNF for 3 days. This time point was selected as the goal is to induce maturation of cells that we designated as DA progenitors (Trzaska et al. 2007b). We performed Ca2+ imaging experiments to monitor changes in intracellular Ca2+ upon depolarization (Representative image, Fig. 3a). ATP showed robust increase in intracellular Ca2+ levels from intracellular stores, consistent with our previous results (Fig. 3d). Hereafter, ATP served as positive controls.

image

Figure 3.  Ca2+ imaging analyses of MSC-derived DA cells induced with neurotrophins. (a) Representative image of induced MSCs treated with BDNF, loaded with fura-2 and utilized for experimentation. The red fluorescence indicates the highest concentration of intracellular Ca2+ and the blue fluorescence indicates the lowest concentration of intracellular Ca2+. ATP was used as a positive control to monitor increases in intracellular Ca2+ levels. Arrows indicate time point of adding either 50 mM KCl or 100 μM ATP. (b) Treatment with BDNF induces elevated Ca2+ transient upon depolarization with KCl in 52 ± 8% (mean ± SEM, n = 5) of cells. (c) Absence of Ca2+ in the extracellular solution abolishes the KCl induced Ca2+ transient in BDNF-treated cells. (d) 12-day induced MSCs without any addition of neurotrophins were used as the negative control. (e) Treatment of the MSC-derived DA cells with NGF, NT-3, or GDNF did not evoke any changes in intracellular Ca2+ transient upon addition of KCl.

Download figure to PowerPoint

The BDNF-treated cells showed a robust increase in intracellular Ca2+ levels upon depolarization with 50 mM K+ solution (Fig. 3b). Quantitative analyses indicated that 52 ± 8% (mean ± SEM, n = 5) of the BDNF treated cells elicited Ca2+ transients in response to K+-induced depolarization. This observation was abolished by the removal of external calcium from the bathing solution (Fig. 3c), indicating that the increase in intracellular Ca2+ was dependent on external source. Thus, we deduce that the BDNF-treated cells express Ca2+ voltage-gated channels similar to native neurons.

Next, we tested whether the effects of BDNF were unique or similar for other NTs. We therefore tested NGF, NT-3, and GDNF by each to 9-day induced MSCs at 50 ng/mL. Interestingly the levels of intracellular Ca2+ did not change upon depolarization with 50 mM K+ external solution in the cells treated with NGF, NT-3, or GDNF (Fig. 3e). These results are similar to those of the untreated induced cells, in which only ATP increased intracellular Ca2+ levels through the release of Ca2+ from intracellular stores (Fig. 3b).

The calcium imaging studies were done by monitoring individual cells. We opted not to use gridded coverslip in the imaging studies as this method is intended for post hoc analyses of markers on the responding cells. Our previous studies have shown that > 90% of the neurons express NeuN and beta-III-tubulin and ∼70% of these express tyrosine hydroxylase. As ∼52% of the cells treated with BDNF yielded calcium transients, the responding cells are likely to be neurons.

To further access functionality in the BDNF-treated MSCs, we determined whether the cells could cause respond to neurotransmitters, ACh, l-glut, and GABA. First, we studied if the BDNF-treated cells express Glut-R, nAchα1. Western blots show increased expressions of GABA-R and nAchα1 in the BDNF-treated cells (Fig. 4a). The bands for Glut-R were similar, regardless of treatment (Fig. 4a, top row).

image

Figure 4.  Responsiveness of BDNF-treated cells to ACh and l-glut. (a) Western blots for neurotransmitter receptors with whole cell extracts from Sy5y (control); uninduced MSCs (day 0); and BDNF-treated induced cells (day 12). (b and c) Representative fura-2 Ca2+ imaging data of MSC-derived DA cells induced with BDNF and stimulated with the neurotransmitters, ACh or l-glut. ATP was used as a positive control to monitor increases in intracellular Ca2+ levels. Arrows indicate the time point of adding 1 mM ACh, 1 mM l-glut or 100 μM ATP. Untreated, induced MSCs (12 days) showed no change in intracellular Ca2+ transient upon addition of ACh or l-glut. (d and e) MSC-derived DA cells treated with BDNF show elevated intracellular Ca2+ levels upon stimulation with ACh (55 ± 13%, mean ± SEM, n = 3) but not with l-glut.

Download figure to PowerPoint

We next determined if the neurotransmitter receptors (Fig. 4a) are functional. This was addressed by stimulating the BDNF-treated and untreated cells with 1 mM ACh or l-glut. The 12 day induced control MSCs did not increase intracellular Ca2+ levels (Fig. 4b and c). The BDNF-treated cells showed robust increases in intracellular Ca2+ levels upon treatment with ACh (Fig. 4d). Quantitative analysis revealed that 55 ± 13% (mean ± SEM, n = 3) of the cells responded to ACh. This indicates that ACh receptors are functional. The BDNF-treated cells did not show an increase in intracellular Ca2+ upon stimulation with l-glut, suggesting that glutaminergic neurotransmission is not yet present in these cells, despite the light bands show for the respective receptors (Fig. 4a and d).

GABA, which is an inhibitory neurotransmitter that leads to hyperpolarization in mature native neurons showed an increase in intracellular Ca2+ levels in both the control 12-day induced cells and those with BDNF treatment (Fig. 5a and c). However, the intensity of the Ca2+ transients was significantly higher in the BDNF-treated cells (Fig. 5e). Quantitative analysis revealed that 52% ± 15% (mean ± SEM, n = 3) of the control 12-day cells responded to GABA, whereas 71 ± 10% (mean ± SEM, n = 3) of the BDNF-treated cells responded to GABA. The GABA-evoked increase in intracellular Ca2+ levels was prevented in Ca2+-free external solution, indicating presence of GABAA receptors (Fig. 5b and d). It has been widely reported that GABA responses via GABAA receptors are depolarizing in developing neurons because of high internal Cl levels and are able to increase intracellular Ca2+ (Ben-Ari 2002; Carrasco et al. 2007).

image

Figure 5.  Responsiveness of BDNF treated cells to GABA. Representative fura-2 Ca2+ imaging data of MSC-derived DA cells induced with BDNF and stimulated with GABA. ATP was used as a positive control to monitor increases in intracellular Ca2+ levels. Arrows indicate time point at which 1 mM GABA or 100 μM ATP was added. (a and b) Untreated 12-day induced MSCs (control) showed increases in intracellular Ca2+ levels upon addition of GABA (52% ± 15%, mean ± SEM, n = 3), which was abolished in Ca2+ free extracellular solution. (c and d) BDNF-treated induced MSCs show increases in intracellular Ca2+ levels upon GABA treatment (71 ± 10%, mean ± SEM, n = 3) were abolished in Ca2+-free extracellular solution. (e) Intracellular Ca2+ intensities are compared between GABA stimulation for untreated 12-day induced MSCs (control) and BDNF-treated induced MSCs. Data are normalized and presented as mean ± SEM, n = 3. *, statistically significant < 0.05 vs. control.

Download figure to PowerPoint

The hallmark of functional neuronal circuits is the formation of synapses between individual neurons. As this process is driven by an action potential in the pre-synaptic neuron and we have shown the presence of voltage-gated Ca2+ channels (Fig. 3c and d), we asked whether the cells would show synaptic transmission necessary for neuronal communication. In comparison to 12-day induced MSCs (Fig. 6a), the cells treated with BDNF exhibited inward spontaneous post-synaptic currents (sPSCs) in whole cell voltage clamp mode (Fig. 6b). The decay of the sPSC has a fast and a slow component (Fig. 6b2) with an amplitude of 221 ± 66 pA (mean ± SEM, n = 6). Future pharmacological studies are necessary to determine the properties of the sPSCs. Taken together, the data shows that BDNF promotes functional maturation of the MSC-derived DA cells, however the cells are not fully developed, due to the lack of glutamate transmission and the excitatory response produced by GABA, as determined by the Ca2+ imaging experiments.

image

Figure 6.  Electrophysiological properties of cells treated with BDNF. (a) Representative trace of control 12-day MSCs. (b1) Representative trace of spontaneous post-synaptic currents (sPSC) recorded with the whole-cell mode at a holding potential of −60 mV. Each sPSC is indicated with asterisks. (b2) Illustrates a sPSC in extended time scale. (b3) The decay of the sPSC has a fast and a slow component. The amplitude of the sPSC is 221 ± 66 pA (mean ± SEM, n = 6).

Download figure to PowerPoint

Analysis of dopamine secretion

Our previous studies showed that the 12-day induced MSCs constitutively secreted DA; however release was not dependent on neuronal activity (Trzaska et al. 2007b). Based on the robust increase in intracellular Ca2+ levels upon K+ depolarization in the BDNF-treated cells (Fig. 3c), we asked whether constitutive DA release would be increased or mediated by neuronal activity. Control experiments show that the constitutive release of DA is blocked by suramin, but not scopolamine (Fig. 7a). This suggests that spontaneous ATP mediates the constitutive release of DA. Scopolamine, a muscarinic ACh receptor antagonist, did not have any effect (Fig. 7a). As described previously, DA release in the control cells, 12-day induced MSCs, is not affected upon depolarizing conditions (Fig. 7b). Conversely, cells treated for 9 days with the induction cocktail and subsequently with BDNF, exhibited increased DA release upon depolarizing conditions (Hi K+ external solution), which was completely abolished in a Ca2+-free solution (Fig. 7b). This suggests that, in the BDNF-treated cells, DA release is mediated by neuronal activity.

image

Figure 7.  Mesenchymal stem cell (MSC)-derived DA cells treated with BDNF secrete dopamine. Levels of DA were measured using an ELISA kit. Cumulative data for DA secreted by the 12-day untreated MSCs and in parallel induced cells in which BDNF was added at day 9 induction. (a) Cells were treated for 5 min with control solution, external solution with 100 μM suramin or 100 μM scopolamine. (b) Cells were treated for 5 min with control SES solution, external solution with 50 mM K+, or Ca2+-free external solution with 50 mM K+. Results are presented as the mean DA concentration ± SEM, n = 3; (a) *, statistically significant < 0.05 vs. control. *, statistically significant < 0.05 vs. control 12-day induced MSCs. #, statistically significant < 0.05 vs. high K+ ES.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This report provides a novel finding showing BDNF promotes functional maturity in MSC-derived DA cells. We have previously reported on an induction cocktail comprising SHH, FGF8, and bFGF with ∼70% efficiency in the transdifferentiation of MSCs to DA-producing cells, based on tyrosine hydroxylase expression (Trzaska et al. 2007b). Among the induction was > 90% that expressed NeuN and beta-III-tubulin. Despite the markers of the latter neuronal markers, the cells were functionally immature, based on electrophysiology. In this study, we show that BDNF addition to 9-day induced MSCs results in the development of cells which elicit sPSCs and secrete DA in response to Ca2+-dependent depolarization. This is a significant finding that mimics the physiology of native neurons and has yet to be reported in MSC-derived DA cells. Furthermore, BDNF treatment results in KCl-induced depolarization and enhanced DA release (Figs 3 and 7). The ability of the induced cells to respond to KCl is a hallmark of excitable cells and the enhanced release of DA by the BDNF-treated cells indicate dopaminergic type.

As compared to other protocols in the literature, we did not change the media during the induction for two main reasons. Firstly, the evidence indicates that several unknown (or possibly neural inductive) factors are secreted by the cells during differentiation. The initial studies were performed with intermittent change in media and observed halting of differentiation with cell death (unpublished). This led us to surmise that change of media will remove key endogenously produced factors that are important for maturation. As neuronal induction is very complex and requires a multitude of factors, it is possible that the initial induction leads to the production of other secretory factors that might facilitate differentiation although autrocrine mechanism. It is also possible that addition of exogenous BDNF synergizes with factors secreted by the cells to induce differentiation. Secondly, SHH is tumorigenic (Yue et al. 2008). During induction, a key repressor of neuronal gene expression, REST, is down-regulated (Trzaska et al. 2008). As REST is a tumor suppressor gene (Coulson 2005; Reddy et al. 2009), its decrease, with the presence of the oncogeneic SHH could cause tumor formation. Thus, avoiding re-addition of SHH would be safer as we progress to apply the reach. These questions are follow-up to the present report. Also, we were trying to induce terminal differentiation and limit proliferation.

Neurotrophins, such as GDNF, BDNF, and NT-3, have an important role in the maintenance, development, and survival of DA neurons (Zhou et al. 1996; Erickson et al. 2001; Baquet et al. 2005; von Bohlen und Halbach et al. 2005). GDNF binds two receptors, GFRα1 and c-ret, and has been described as a potent trophic factor for DA neurons (Love et al. 2005). Although we detected GFRα1 on the induced cells, although with lower percentages (60–70% vs. 90% for TrkB), we focused on TrkB. We performed Ca2+ imaging experiments in which the cells were treated with combinations of BDNF and GDNF. The results were similar to studies with BDNF alone (not shown), suggesting that BDNF is the only contributor. This was a surprise as we observed GFRα1. The explanations for these observations could be explained by non-functional GFRα1 receptor or perhaps, another role for GDNF in the DA neurons. We plan to continue these studies in future experiments using knockdown of the various neurotrophic receptors.

Nerve growth factor, BDNF and NT-3 act through the Trk receptor tyrosine kinase family and regulate nervous system development (Bibel and Barde 2000; Chao 2003). NGF is the ligand for TrkA, whereas BDNF binds to TrkB and NT-3 binds to TrkC. BDNF and NT-3 have recognized roles in the differentiation and survival of DA neurons (Baquet et al. 2005). NGF has been implicated in the differentiation of a wide variety of neuronal types and particularly causes neurite outgrowth and increases Na+ voltage-gated channel expression (Kalman et al. 1990; Chao 2003). Thus, we were interested in the role that these NTs could have in further maturation of MSC-derived DA cells. The premise was that initial inductive stimuli would up-regulate receptors for additional factors required for maturation of the MSC-derived DA cells. These factors could be either too low or missing in our in vitro system and if the newly up-regulated receptors are not stimulated at the appropriate time points, the cells might fail to progress to the next stage of development. Analysis of receptor expression revealed up-regulation of TrkB and GFRα1 by day 9 of induction (Fig. 1). Interestingly, the expression of TrkA and TrkC did not change significantly during the course of differentiation (Fig. 1a). This was not surprising as MSCs have been reported to be neuroprotective and have low expression of Trk receptors (Pisati et al. 2007). Due to the high expression of TrkB (Fig. 1), we added BDNF to cells induced for 9 days and observed up-regulation of D1 receptor, indicating responsiveness to BDNF through TrkB signaling (Fig. 2) (Do et al. 2007). Thus, we hypothesized that addition of BDNF at initial up-regulation might be necessary for further maturation and functionality.

The hallmark of functional neurons is the presence of excitable properties and formation of synapses. We found that addition of BDNF to MSCs induced for 9 days facilitated further maturation evident by the intracellular Ca2+ increase upon depolarization, which was prevented in Ca2+-free external solution (Fig. 3b and c). This indicates that the cells express voltage-gated Ca2+ channels similar to native neurons. In contrast, cells stimulated with other neurotrophic factors, NGF, NT-3, or GDNF did not respond to K+-induced depolarization (Fig. 3e). The data suggest that depolarization causes direct activation of voltage-gated Ca2+ channels or activating voltage-gated Na+ channels, which then indirectly open Ca2+ channels in the 9 day MSCs treated with BDNF.

Dopamine neurons have been found to respond to neurotransmitters, ACh, l-glut, and GABA (Raye et al. 2007). In comparison to the control 12-day induced MSCs, 55 ± 13% of the cells treated with BDNF responded to ACh (Fig. 4a and c). ACh receptors are readily found on DA neurons and ACh transmission is important in striatal DA release and nerve development (Zhou et al. 2004; Calabresi et al. 2006; Raye et al. 2007). As DA neurons express both ACh nicotinic and muscarinic receptors (Calabresi et al. 2006), Ca2+ release could be mediated through excitatory nicotinic receptors or by the release of Ca2+ from intracellular stores via muscarinic receptors. Future studies are warranted to test this hypothesis. Surprisingly, the BDNF-treated cells did not respond to the other excitatory neurotransmitter, l-glut (Fig. 4d). This suggests that glutamatergic neurotransmission is not yet present in these cells. However, we found that GABA, which is an inhibitory neurotransmitter, caused an increase in intracellular Ca2+ levels upon stimulation in both the control 12-day induced cells (52% ± 15%) and the BDNF treated cells (71 ± 10%), albeit at different intensities (Fig. 5e). This may be due to the presence of higher numbers of GABAA receptors or Ca2+ voltage-gated channels in the BDNF treated cells. These observations show that the cells respond in an opposite, excitatory manner to GABA. This activity has been reported previously in developing neurons, in which GABA induces excitatory neurotransmission before the appearance of glutamatergic neurotransmission (Ben-Ari et al. 2007; Zhao et al. 2008). It was expected that the control 12-day induced MSCs would respond to GABA in an excitatory manner, as they are in an immature state, as determined by our previous studies (Trzaska et al. 2007b). However, the BDNF treated cells responded similarly, indicating that the cells are still developing and GABAergic neurotransmission has not yet shifted to inhibition. This is consistent with the lack of responses to l-glut (Fig. 4b and d), as GABA is excitatory before the presence of glutamatergic neurotransmission (Ben-Ari et al. 2007). This pattern allows for proper neuronal growth and synapse formation in developing neurons that have few synapses (Ben-Ari 2002). Moreover, this excitatory action of GABA occurs through the activation of ionotropic GABAA receptors (Maric et al. 2001). Our observations are consistent with previous reports, wherein the depolarizing response of GABA is abolished without extracellular Ca2+, indicating presence of GABAA ionotropic receptors (Fig. 5b and d). Taken together, the data indicate that the MSC-derived DA cells treated with BDNF are still developing mature synaptic connections. At this time we do not have any explanation why the BDNF-treated cells do not show fully functional neurons with regards to inhibitory and excitatory functions. Similar results were observed when the 3-day BDNF treatment was extended up to 1 week (data not shown). This suggests that another factor is required for glutamergic and GABAergic DA neurons. Further studies are planned to identify factor(s) that could show fully functional neurons.

Functional neuronal networks communicate through synaptic activity. Our data show that the BDNF-treated cells elicit inward sPSCs (Fig. 6). In comparison to the sPSCs recorded from MSCs treated with retinoic acid in our laboratories previous studies (Cho et al. 2005; Greco et al. 2007), the BDNF-treated cells exhibited kinetic properties of sPSCs that are closer to the sPSCs of those recorded from mammalian CNS neurons in brain slices (Xiao et al. 2007). Although future studies are necessary to characterize these responses, it is possible that the sPSCs could be GABAergic, judging by the time scale, rise and decay times, and recording conditions (Fig. 6b). Additionally, the sPSCs could be mediated by P2 receptor channels, based on our studies with ATP (Fig. 7a) (Safiulina et al. 2005; Saitow et al. 2005; Trzaska et al. 2007b).

Our observations on the maturation properties of BDNF are consistent with other reports. BDNF is an important regulator of synaptic transmission and addition of exogenous BDNF was found to enhance synaptic efficiency and accelerate synapse maturation (Bibel and Barde 2000; Bolton et al. 2000; Yamada et al. 2002; Gubellini et al. 2005; Carrasco et al. 2007). In addition, BDNF can enhance GABAergic post-synaptic currents in immature neurons, but had the reverse effect in mature neurons (Tanaka et al. 1997; Mizoguchi et al. 2003). Our observations are comparable with these previously reported studies, showing that BDNF promotes a maturation response in the MSC-derived DA cells.

In mature neuronal cells, the release of neurotransmitters is dependent on Ca2+ influx from the extracellular microenvironment. Our previous studies showed that the MSC-derived DA cells constitutively secrete DA that was not affected by K+-induced depolarization and not prevented by removal of extracellular Ca2+ (Trzaska et al. 2007b). We found that the basal release of DA was mediated by spontaneous ATP, as the ATP antagonist suramin, significantly decreased DA release (Fig. 7a). Conversely, this was not observed by scopolamine, an antagonist for ACh muscarinic receptors. The cells treated with BDNF showed a significant increase in DA release upon depolarization, which was abolished in a Ca2+-free extracellular solution (Fig. 7b). These results mimic the physiology of neurotransmitter release in native neurons which is not observed in the control 12-day MSCs. This indicates that day 9 of induction of MSCs is a critical point in the regulation of DA release, which is modulated by exogenous BDNF.

In summary, we have demonstrated that MSCs induced with our induction cocktail show robust up-regulation of TrkB, and addition of BDNF promotes functional maturation. The cells elicit sPSCs – an important finding that has yet to be reported in MSC-derived DA cells. The cells also secrete DA in manner similar to native neurons. Although the BDNF-treated cells exhibit many characteristics of neurons, they are not fully developed with regard to mature glutamateric and GABAergic transmissions (Figs 4 and 5). Altogether, these observations demonstrate the crucial function of TrkB and its ligand, BDNF, in the development of functional DA neurons from human MSCs.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This work was supported by the F. M. Kirby Foundation. K.A.T. was supported by the Benigno Fellowship for research in neural and regeneration and repair and the work is in partial fulfillment of a PhD thesis.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References