ATP Modulates Cell Proliferation and Elicits Two Different Electrophysiological Responses in Human Mesenchymal Stem Cells

Authors


Abstract

Bone marrow-derived human mesenchymal stem cells (hMSCs) have the potential to differentiate into several cell lines. Extracellular adenosine 5′-triphosphate (ATP) acts as a potent signaling molecule mediating cell-to-cell communication. Particular interest has been focused in recent years on the role of ATP in stem cell proliferation and differentiation. In the present work, we demonstrate that hMSCs at early stages of culture (P0–P5) spontaneously release ATP, which decreases cell proliferation. Increased hMSC proliferation is induced by the unselective P2 antagonist pyridoxalphosphate-6-azophenyl-2′,4′-disulfonate (PPADS) and by the selective P2Y1 antagonist 2′-deoxy-N6-methyladenosine3′,5′-bisphosphate (MRS 2179). A functional role of extracellular ATP in modulating ionic conductances with the whole-cell and/or perforated patch-clamp techniques was also investigated. Exogenous ATP increased both the voltage-sensitive outward and inward currents in 47% of cells, whereas, in 31% of cells, only an increase in inward currents was found. Cells responding in this dual manner to ATP presented different resting membrane potentials. Both ATP-induced effects had varying sensitivity to the P2 antagonists PPADS and MRS 2179. Outward ATP-sensitive currents are carried by potassium ions, since they are blocked by cesium replacement and are Ca2+-dependent because they are eliminated in the presence of 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid. On the basis of different electrophysiological and pharmacological characteristics, we conclude that outward ATP-sensitive currents are due to Ca2+-dependent K+-channel activation following stimulation of P2Y receptors, whereas inward ATP-sensitive currents are mediated by P2X receptor activation. In summary, ATP released in early life stages of hMSCs modulates their proliferation rate and likely acts as one of the early factors determining their cell fate.

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

Introduction

Adenosine 5′-triphosphate (ATP) has emerged as one of the most versatile molecules implicated in a variety of cell processes, from energy supply to cell-to-cell signaling. In addition to the classic role of intracellular ATP as energy source and phosphate donor in enzymatic processes, extracellular ATP may activate specific membrane receptors belonging to the P2 purinergic receptor family, which are expressed in several cell types.

P2 receptors are classified as P2X ionotropic and P2Y metabotropic receptors [1, [2], [3]–4]. P2X receptors are unselectively permeable to cations (Na+, K+, Ca2+) and are present as homo- and heteromeric trimers [5, 6] on the cell surface. P2Y receptors are mostly coupled to Gq/11 proteins activating phospholipase C (PLC) and phosphoinositide cascade, with subsequent inositol triphosphate (IP3) formation and Ca2+ release from intracellular stores [7]. Nine subtypes of metabotropic P2Y (P2Y1,2,4,6,11,12,13,14 [8] and the recently deorphanized GPR17 receptor [9]) and seven subunits of the P2X (P2X1–7) receptor have been cloned to date in mammals [6].

It is generally known that extracellular ATP is involved in various cell processes such as neurotransmission [10], regulation of cardiovascular functions [11], smooth muscle contractility [12], and platelet aggregation [13]. In the last few years, several studies have demonstrated a trophic role for P2 receptor activation in various lineages of differentiated [14, 15] and undifferentiated cells. Hematopoietic stem cells [16, [17]–18], neuronal progenitor stem cells [19, [20]–21], and oligodendrocyte progenitors [22] respond to extracellular ATP with increased proliferation, differentiation, and/or migration by intracellular signaling cascades that are similar or synergic to those activated by numerous growth factors. In different cell systems, the trophic role of extracellular ATP is correlated with the occurrence of spontaneous intracellular Ca2+ waves, a self-renewing mechanism promoting periodic oscillation of intracellular Ca2+ levels that plays a crucial role in cellular proliferation and differentiation [23].

There has been great interest lately in human mesenchymal stem cells (hMSCs), a population of undifferentiated stromal stem cells derived from human adult bone marrow [24, [25], [26]–27] that can be isolated, expanded in culture, and stimulated to differentiate into a variety of cell lineages (bone, cartilage, muscle, marrow stroma, tendon, fat), including neurones [28]. Because large numbers of hMSCs can be generated in culture, tissue-engineered constructs principally composed of these cells may be reintroduced into the in vivo setting. It has been recently demonstrated that in vivo transplantations of hMSCs into the postischemic brain [29] or infarcted myocardium [30, 31] of experimental animals significantly improved tissue functions, suggesting a possible role for these cells in the therapy of degenerating pathologies [32, 33]. In spite of the huge employment of hMSCs in experimental cell transplantation that has been carried out in the last few years, the precise mechanisms determining in vivo proliferation, migration, and differentiation of engrafted hMSCs into the host tissue are still unclear.

The electrophysiological characteristics of hMSCs in early passages in culture have been recently reported by two different groups, which demonstrated that almost all hMSCs present outward K+ currents, whereas inward Na+ and Ca2+ conductances are present only in a minority of cells [34, 35]. Data in the literature have demonstrated that ATP released from hMSCs is responsible for the generation and propagation of intracellular Ca2+ waves [36]. Through this mechanism, ATP also promotes the activation and nuclear translocation of the transcription factor NFAT, suggesting a pivotal role of extracellular ATP in cell differentiation.

No evidence exists to date about a functional role of ATP released by hMSCs in culture, either on cell proliferation or on ionic conductances. We studied whether these cells spontaneously release ATP in the early stages of cell culture (from P0 to P5) with the luciferin/luciferase technique and whether P2 receptor activation modulates cell proliferation. Finally, we investigated a possible role of extracellular ATP in modulating membrane currents by using the patch-clamp technique.

Materials and Methods

Isolation, Culture, and Characterization of hMSCs

Human bone marrow cells were obtained from the iliac crest of marrow aspirates from healthy donors. Informed consent was obtained from all donors, and the institutional ethical committee approved all procedures. Details on isolation of hMSCs have been described previously [37]. Briefly, whole bone marrow aspirate was collected, and small aliquots were centrifuged for 10 minutes at 700g; the ring of white blood cells (buffy-coat) was recovered and plated in 75-cm2 flasks (1.6 × 105 cells per cm2) in Iscove's modified Dulbecco's medium (with l-glutamine and Hepes 25 mM; EuroClone, Milan, Italy, http://www.euroclone.net) with 50 μg/ml gentamicin (Schering-Plough, Milan, Italy, http://www.schering-plough.com), 10% fetal bovine serum (FBS) (HyClone, Logan, UT, http://www.hyclone.com), and 2% Ultroser G (Pall Biosepra, Cergy-Saint-Christophe, France, http://www.pall.com) and incubated at 37°C in a humidified atmosphere containing 95% air and 5% CO2. On reaching confluence, the adherent cells were harvested with 0.05% trypsin-0.02% EDTA (Eurobio, Courtaboeuf Cedex B, France, http://www.eurobio.fr) for 5–10 minutes at 37°C, washed with Hanks' balanced saline solution (HBSS, without calcium and magnesium; EuroClone) and 10% FBS, and resuspended in complete medium (primary culture, P0). Cells were plated again at 104 cells per cm2 in 100-mm dishes (P1) for the in vitro differentiation experiments and for determination of their growth kinetics; expansion of the cells was obtained with successive cycles of trypsinization and reseeding.

Phenotype and Differentiating Ability of hMSCs

At the first passage in culture, the morphologically homogeneous population of hMSCs was analyzed for the expression of cell surface molecules using flow cytometry procedures; hMSCs, recovered from flasks by trypsin-EDTA treatment, were washed in HBSS and 10% FBS and were resuspended in flow cytometry buffer consisting of CellWASH (0.1% sodium azide in phosphate-buffered saline; Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com) with 2% FBS. Aliquots were incubated with the following conjugated monoclonal antibodies: CD34-phycoerythrin (PE), CD45-fluorescein isothiocyanate (FITC), CD14-PE, CD29-PE, CD44-FITC, CD166-PE, CD90-PE, CD73-PE, human leucocyte antigen (HLA)-DP DQ DR, HLA-ABC (BD Pharmingen, San Diego, http://www.bdbiosciences.com/index_us.shtml), and CD105-PE (Ancell, Bayport, MI, http://www.ancell.com). Nonspecific fluorescence and morphologic parameters of the cells were determined by incubation of the same cell aliquot with isotype-matched mouse monoclonal antibodies. After incubations, cells were washed, and 7-amino-actinomycin (7-AAD) was added in order to exclude dead cells from the analysis. Flow cytometric acquisition was performed on a FACSort (Becton, Dickinson) instrument, and data were analyzed on dot-plot biparametric diagrams using Cell Quest software (Becton, Dickinson) on a Macintosh PC. The ability of hMSCs to differentiate among osteogenic, adipogenic, and chondrogenic lineages was assayed as previously described in more detail [37].

Adenosine 5′-Triphosphate Assay

ATP in the cell supernatant was analyzed by the ATPLite-M kit (PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com) according to the previously described method [38]. The luminescence was measured on the PerkinElmer TopCount luminescence counter (TopCount Packard Instruments; PerkinElmer).

hMSCs were plated onto a 24-well microplate at a density of 50,000 cells per well and were incubated for 24 hours before ATP assay. After 24 hours, a sample of extracellular medium (50 μl) was collected from each well and placed in a white 96-well microplate. The same volume of the substrate solution (50 μl) was added to each sample. Standard solutions of ATP (from 10−7 to 10−11 M) were run in parallel in the same microplate following an identical procedure as that of samples. The ATP values were extrapolated from the linear regression curve calculated on the basis of standard solutions and expressed in absolute values (nM).

Cell Proliferation

hMSCs at a culture passage from P2 to P4 were used for proliferation experiments. Cells were plated at 105 cells per cm2 in 100-mm dishes cultured with complete medium (10 ml) containing 10% FBS, and Ultroser G was not added. The P2 receptor agonist ATP (10 μM), the P2Y1 receptor antagonist 2′-deoxy-N6-methyladenosine3′,5′-bisphosphate (MRS 2179) (10 μM), and the unselective P2 antagonist pyridoxalphosphate-6-azophenyl-2′,4′-disulfonate (PPADS) (30 μM) were added daily. The concentrations used for the agonist or antagonists were chosen on the basis of the results obtained in electrophysiological experiments. The same volume of carrier solution was added to control dishes. After 5 days in culture, hMSCs were harvested with 0.05% trypsin-0.02% EDTA for 5–10 minutes at 37°C, washed with HBSS and 10% FBS, and resuspended in complete medium. The cell number was counted using an automated coulter. All experiments were performed in quadruplicate.

Electrophysiological Experiments

Solutions.

The tyrode solution contained (mM): HEPES 5, glucose 10, NaCl 140, KCl 5.4, MgCl2 1.2, and CaCl2 1.8. The pH was adjusted to 7.3 with NaOH. The pipette solution contained (mM): K-aspartate 110, KCl 20, Na2-ATP 5, MgCl2 2, HEPES 10, Na2-GTP 0.1, and EGTA 0.05. The pH was adjusted to 7.2 with KOH. In some experiments, to eliminate K+ currents, we replaced K+ with equimolar concentrations of Cs+ in both the extracellular and internal pipette solution, and the pH was adjusted with CsOH; 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) containing pipette solution was (mM): K-aspartate 130, Na2-ATP 5, MgCl2 2, HEPES 10, Na2-GTP 0.1, BAPTA 11, and CaCl2 5.

Recordings.

After enzymatic dissociation, cells were suspended in tyrode solution, transferred to a small chamber mounted on the platform of the inverted microscope (Olympus CKX41; Olympus, Tokyo, http://www.olympus-global.com), and superfused at a flow rate of 2 ml/minute with tyrode solution. Borosilicate glass electrodes were pulled with a Sutter instrument (model P-87; Sutter Instrument, Novato, CA, http://www.sutter.com). Tip resistance was between 2 and 3.5 MΩ. Data were acquired with an Axopatch 200B amplifier (Axon Instruments/Molecular Devices Corp., Union City, CA, http://www.moleculardevices.com), low pass filtered at 10 kHz, stored, and analyzed with pClamp 9.2 software (Axon Instruments/Molecular Devices Corp.). For perforated-patch experiments, pipettes were back-filled with an intracellular solution containing amphotericin-B (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) [39], solved in dimethyl sulfoxide (30 mg/ml), and diluted to 0.12 ng/ml in standard intracellular pipette solution before recording.

Protocols Used.

Membrane resistance was measured with fast hyperpolarizing voltage pulses (from −70 to −75 mV, 40 millisecond [ms] duration). Series resistance (Rs) and membrane capacitance (Cm) were routinely checked and compensated during the experiment. Only cells showing a stable Cm and Rs before, during, and after drug application were included in the analysis.

In order to study the electrophysiological characteristics and the response to P2 agonist and antagonist application in hMSCs, we used three different protocols. The first one was a voltage ramp protocol consisting of a 1,500-ms depolarization (from −90 to +40 mV) from a holding potential of −75 mV to evoke a wide range of overall voltage-dependent membrane currents. Each trace shown in the figures is the average of five consecutive episodes. The second protocol was a voltage step protocol to specifically generate outward currents. It consisted of 16 consecutive steps (10 mV each, 1,800-ms duration, −80 mV holding potential) starting from a −70 mV membrane potential to a +80 mV value. In these conditions, the current amplitude was measured at peak. A variation of this protocol was used to selectively block transient outward currents and consisted of a series of 10 consecutive depolarizing steps (10 mV each, 1,800-ms duration, −40 mV prestep potential) starting from a −30 mV membrane potential to a +60 mV value. Current values were measured at steady state. For the third protocol, a membrane potential of −60 mV was constantly imposed to the cell while the holding current was continuously measured (at a sampling rate of 100 microseconds) before, during, and after ATP application. Current amplitude and cell capacitance values are expressed as picoAmpere (pA) and picoFarad (pF), respectively.

In most experiments, drugs were applied by superfusing with a three-way gravity system. A complete exchange of perfusing solutions was achieved within 60 seconds. To better investigate the nature and kinetic characteristics of ATP-mediated inward currents, in a series of experiments we used a high-time resolution system. In that case, the patch-clamped cell was superfused by means of a microsuperfusor, which allowed rapid changes in the solution bathing the cell (less than 1 second).

Drugs

ATP was purchased from Sigma. PPADS and MRS 2179 were obtained from Tocris (Bristol, U.K., http://www.tocris.com). Each drug was dissolved in tyrode solution and applied by bath superfusion.

Statistical Analysis

Data were analyzed using Prism 3.02 software. All numerical data are expressed as the mean ± SE. Data were tested for statistical significance with the paired Student's t test, the unpaired Student's t test, or by analysis of variance (one-way ANOVA), as appropriate. When significant differences were observed, the Newman-Keuls multiple comparison test (one-way ANOVA) was done. A value of p < .05 was considered significant.

Results

Experiments were conducted on successfully culture-expanded hMSCs. A morphologically homogeneous population of fibroblast-like cells with more than 90% confluence was seen after 14 days. Thereafter, primary cultured cells were trypsinized and replated and, after the first passage in culture, grew exponentially by weekly passages in culture [37].

Flow cytometric analysis was used to assess the purity of hMSC cultures, which appeared uniformly positive for CD29, CD44, CD166, CD90, CD73, HLA-ABC, and CD105. HLA-DP DQ DR was expressed in less than 2% of the population. There was no detectable contamination of hematopoietic cells; in fact, markers of hematopoietic lineage, CD14, CD34, and CD45, were not detectable (Fig. 1). Similar results were obtained from P0 to P6 [37], indicating a morphologically homogeneous population of hMSCs in this life span.

Figure Figure 1..

Flow cytometric analysis of surface adhesion molecules on human mesenchymal stem cells. Cells were labeled with monoclonal antibodies specific for molecules indicated in each flow cytometric histogram (negative for CD45, HLA-DP DQ DR, CD34, and CD14 and positive for CD166, HLA-ABC, CD105, CD90, CD44, CD29, and CD73). Abbreviations: FITC, fluorescein isothiocyanate; HLA, human leucocyte antigen; PE, phycoerythrin.

Once a homogeneous hMSC population was obtained (Fig. 2A), we investigated whether it spontaneously released ATP into the culture medium. In previous experiments, we found that ATP concentrations measured in the supernatant collected after plating human astrocytoma cells (ADF1) decrease steadily over time, attaining a baseline in the low nanomolar range between 3 and 5 hours thereafter and remain unchanged for up to 24 hours (Fig. 2B). High ATP release measured soon after cell plating is in agreement with previous works and is attributed to “medium change” that is a mechanical stimulation [40]. As shown in Figure 2C, the extracellular ATP concentration measured in the medium containing hMSCs, 24 hours after a medium change, was 7.6 ± 0.57 nM. This value was significantly higher than that measured in the medium without cells, which was 0.3 ± 0.11 nM (p < .0001, unpaired Student's t test; n = 11). No difference was found in ATP concentration at cell passages in culture from P0 to P5. Therefore, all ATP concentrations from P0 to P5 were plotted together. Unfortunately, in our study, we could not determine ATP concentration in the presence of the ecto-ATPase inhibitor, ARL 67156, which prevents degradation of ATP, because this drug interfered with the luciferin/luciferase assay [38].

Figure Figure 2..

Human mesenchymal stem cells (hMSCs) in culture spontaneously release ATP, which modulates cell proliferation. (A): The figure shows adherent hMSCs as they appear in a primary culture (P0) 14 days after plating. (B): Extracellular ATP levels in ADF1 cells. The medium was collected at indicated times and ATP assayed with the luciferin/luciferase bioluminescence method. Data are the mean ± SE of n = 8 experiments. ∗ p < .0001 (unpaired Student's t test vs. 0, 1, and 3 hours). (C): Extracellular concentrations of ATP were measured in the medium containing hMSCs and in control medium not containing cells. Data are expressed as mean ± SE, n = 11. ∗ p < .0001 (unpaired Student's t test vs. medium alone). (D): Effect on hMSC proliferation after daily application of ATP (10 μM), PPADS (30 μM), and MRS 2179 (10 μM). Data are expressed as percentage of proliferation. Proliferation of untreated cells was assumed as 100%. The cell number was determined after 5 days of culture. Each column bar represents the mean ± SE of n = 4 for each experimental condition. ∗ p < .05 (paired Student's t test vs. respective control), § p < .05 versus 10 μM ATP-treated cells (one-way analysis of variance, Newman-Keuls post-test). Abbreviations: MRS 2179, 2′-deoxy-N6-methyladenosine3′,5′-bisphosphate; PPADS, pyridoxalphosphate-6-azophenyl-2′,4′- disulfonate.

We studied if stimulation of P2 receptors by endogenous ATP released from hMSCs affects their proliferation rate in culture. For 5 days, we cultured hMSC (P2–P4) in complete medium containing 10% FBS but no Ultroser G in order to identify a modulation of cell expansion by P2 agonist or antagonists, which were added daily. We observed a significant inhibitory effect on cell number in the presence of 10 μM ATP (−9% ± 0.03%, p < .01, paired Student's t test vs. respective control; n = 4) (Fig. 2D). Flow cytometric analysis of 7-AAD incubated cells revealed the same percentage (97%) of viable cells in treated and untreated hMSCs (data not shown). On the contrary, cell number was increased in the presence of the unselective P2 antagonist PPADS (26% ± 12%, 30 μM, p < .05, paired Student's t test vs. respective control; n = 4) and in the presence of MRS 2179 (12% ± 0.03%, 10 μM, p < .05, paired Student's t test vs. respective control; n = 4), a selective P2Y1 receptor antagonist (Fig. 2D).

Electrophysiological Experiments

To investigate the presence and role of P2 receptors in hMSCs, we used the patch-clamp technique to record whole-cell currents before and after the application of purinergic agonist and antagonists. The present work was carried out on 63 cells with a mean membrane resistance of 382.9 ± 71 MΩ and a mean membrane capacitance of 38.8 ± 4.3 pF. All of the experiments were carried out at room temperature on cells from P0 to P5 passage in culture.

Exogenous ATP Application Modulates Ionic Currents in hMSCs in Culture

In the first group of cells, we studied the effects of ATP (10 μM, 5 minutes application, n = 20) on hMSCs on the overall currents elicited by a voltage ramp protocol (inset Fig. 3A). In these experimental conditions, ATP elicited two different responses, shown in Figure 3A and 3B and Figure 3C and 3D, respectively.

Figure Figure 3..

Exogenous ATP increases nondesensitizing outward currents and/or desensitizing inward currents in human mesenchymal stem cells (hMSCs) in the whole-cell patch-clamp configuration. (A): Typical experiment in which ATP evoked the increase of outward currents in an isolated hMSC. Left panel: Voltage ramps (inset: from −90 to +40 mV, 1.5 s duration) were applied before and during 10 μM ATP superfusion. Note the increase in outward currents elicited by 10 μM ATP after 5 minutes of application. Right panel: The graph shows the ATP-sensitive current recorded in the same cells and obtained by subtraction of the current recorded at the maximal peak of effect (after 5 minutes of ATP application) from the control trace. (B): Time course of current amplitude activated by the ramp protocol at +40 mV before, during, and after 10 μM ATP application. Each point in the graph represents the mean ± SE of 10 experiments in which ATP increased outward currents. The effect was reversible after drug washout. (C): Typical experiment in which ATP evoked the increase in inward currents in an isolated hMSC. Left panel: Voltage ramps (from −90 to +40 mV, 1.5 s duration) were applied before and during 10 μM ATP superfusion. Note that only inward currents are increased in the presence of ATP (3-minute application). Right panel: Graph shows the ATP-sensitive current recorded in the same cells and obtained by subtraction of the current recorded at the maximal peak of effect (after 3 minutes of ATP application) from the control trace. Note the marked inward rectification. (D): Time course of current amplitude activated by the ramp protocol at −90 mV before, during, and after 10 μM ATP application. Each point in the graph represents the mean ± SE of six experiments in which ATP enhanced inward currents. Note that the response to ATP desensitized before the end of the 5-minute application. (E): Whole-cell recording from a single hMSC held at −60 mV before and during the first 70 s application of 10 μM ATP using a high resolution system. (F): Whole-cell recording from a single hMSC held at −60 mV before and during the first 15 s application of 10 μM ATP using a high resolution system. Abbreviations: min, minutes; pA, picoAmpere; s, seconds.

In Figure 3A, we report that ATP enhanced the outward currents at membrane potentials higher than −20 mV with a modest increase in the inward component at more negative voltages. The ATP-sensitive current, obtained by subtraction of the control ramp from the trace recorded in the presence of ATP, showed a reversal potential of −30.8 mV (Fig. 3A, right panel). Similar results were obtained in 10 cells, which presented a mean resting membrane potential of −37.9 ± 7.1 mV in control conditions and of −38.0 ± 7.1 mV during ATP application (p > .05, paired Student's t test). The effect elicited by 10 μM ATP was maximal since, when ATP was applied at a higher concentration (30 μM, n = 3, not shown), no difference in the effect on outward current was found. A lower concentration of ATP (500 nM) produced a modest and statistically nonsignificant (p > .05, paired Student's t test, n = 4) increase in outward currents (data not shown). The maximal increase in the outward currents was reached after 5 minutes of 10 μM ATP application and was reversible after drug washout (Fig. 3B). In this group of cells, the average current amplitude at +40 mV was 597.9 ± 123.3 pA in control conditions and 937.6 ± 226.1 pA after 5 minutes of ATP application (p < .05, paired Student's t test).

Figure 3C shows the other response elicited by ATP in hMSCs on overall membrane currents evoked by the ramp protocol. In this case, the effect produced by 10 μM ATP was an increase in inward currents without significant changes in outward currents. The ATP-sensitive current presented a reversal potential of 0.5 mV (Fig. 3C, right panel). A similar response was obtained in six cells that presented a mean resting membrane potential of −15.3 ± 3.2 mV in control condition. The mean resting membrane potential measured at the peak of ATP effect was −10.3 ± 2.8 mV, indicating a significant depolarization induced by ATP (p < .01, paired Student's t test). The maximal effect was reached after 3 minutes of ATP application, and it desensitized before the end of drug superfusion (Fig. 3D). The average current amplitude at −90 mV was −162.2 ± 76.8 pA in control and −292.5 ± 149.2 pA after 3 minutes of ATP superfusion (p < .01, paired Student's t test).

In 4 out of 20 cells investigated in this experimental section, no effects of ATP on membrane currents were observed. From this first set of experiments, we conclude that 80% of hMSCs respond to the application of exogenous ATP. These ATP-responding cells can be divided into two subgroups. One group presented a negative membrane potential (−37.9 ± 7.1 mV) and responded to exogenous ATP superfusion with the activation of an outward rectifying ATP-sensitive current, with a reversal potential of −45.9 ± 2.5 mV. The other group of hMSCs presented a more depolarized membrane potential (−15.3 ± 3.2 mV) and responded to exogenous ATP with the activation of an inward rectifying ATP-sensitive current. This effect was accompanied by membrane depolarization. Cell capacitance, as a measure of cell size, was not different in both subgroups. Similarly, cell passage in culture (P0–P5) did not influence ATP-mediated responses, and cells from P0 to P5 were pooled together.

In order to better describe the ATP-induced inward currents, a series of experiments was performed by using a rapid-exchange perfusion system with a high time-resolution rate. In these experiments, cells were constantly held to a membrane potential of −60 mV. In four out of nine cells investigated, a 90-second application of 10 μM ATP elicited an inward current (Fig. 3E), which peaked at 54.2 ± 7.3 seconds. In three other cells, the ATP-induced inward current peaked at 6.1 ± 1.4 seconds, as shown in Figure 3F. In the remaining two hMSCs, no effects of ATP were detected.

In order to investigate the properties of the outward currents activated by ATP, we studied the effects of ATP on ionic currents elicited by a voltage step protocol. As shown in Figure 4A, a consistent increase in outward currents evoked by the step protocol (inset of Fig. 4A) was induced in a typical cell by 10 μM ATP starting from membrane potentials of +40 mV. The ATP-sensitive current showed a more evident increase in the transient early phase followed by a moderate but significant enhancement in the late sustained current. Figure 4B shows the current-voltage relationship (I-V plot) of ATP-sensitive current obtained in this group of cells (n = 3). To isolate the late sustained currents activated by ATP, we used another protocol with a prestep potential of −40 mV (inset of Fig. 4C) that inactivated transient conductances. As expected, in this case only slowly activating, sustained currents were evoked by this protocol, and their amplitude was increased in the presence of 10 μM ATP (Fig. 4C). Figure 4D shows the averaged I-V plot of ATP-sensitive current obtained in this group of cells (n = 5).

Figure Figure 4..

Exogenous ATP increases both the early and late phases of outward currents in human mesenchymal stem cells (hMSCs). (A): Representative current traces recorded in a hMSC by using the voltage step protocol (inset: from −70 to +80 mV, 1.8 s duration, −80 mV holding potential) in the absence (left panel) and in the presence of 10 μM ATP (middle panel). The right panel shows the ATP-sensitive current obtained in the same cell by subtraction of the trace recorded at the maximal peak of effect (after 5 minutes of ATP application) from the control trace. Note the marked increase in fast-activating transient outward currents. (B): Current-voltage relationship of ATP-sensitive current amplitude (measured at the peak) evoked with the protocol described in (A) and obtained by averaging the results recorded in three cells. (C): Representative current traces recorded in a hMSC by using the voltage step protocol (inset: from −30 to +60 mV, 1.8 s duration, −40 mV prestep potential) in the absence (left panel) and in the presence of 10 μM ATP (central panel). The right panel shows the ATP-sensitive current obtained in the same cell by subtraction of the trace recorded at the maximal peak of effect (after 5 minutes of ATP application) from the control trace. Note the increase in late sustained outward currents. (D): Current-voltage relationship of ATP-sensitive currents (measured at the steady state) evoked with the protocol described in (C) and obtained by averaging the results recorded in five cells. Abbreviations: ms, milliseconds; pA, picoAmpere; s, seconds.

The ATP-induced increase in K+ currents in hMSCs was abolished in the presence of intracellular BAPTA (n = 4) as shown in Figure 5A, demonstrating that the ATP effect was mediated by intracellular Ca2+ increase. The effects of ATP were further investigated in the presence of different P2 purinergic antagonists. In the presence of a selective P2Y1 receptor antagonist, MRS 2179 (10 μM), ATP (10 μM) did not induce any effect on membrane currents, evoked either by the ramp (Fig. 5B, left panel) or by the step protocol (Fig. 5B, right panel). Figure 5C shows that the mean effects of ATP on both outward and inward currents evoked by ramp protocol, measured at +40 mV and −90 mV, respectively, were blocked in the presence of MRS 2179 (n = 7). On the contrary, the other P2 antagonist tested, the unselective PPADS, applied at a 30 μM concentration (n = 4), was efficacious in blocking only the ATP-induced increase in inward conductances, also revealing membrane hyperpolarization during ATP application (from −16.3 ± 3.7 mV to −23.9 ± 2.6 mV).

Figure Figure 5..

Effects of BAPTA and of different P2 receptor antagonists on ATP-evoked currents. (A): The graph shows the ATP-sensitive currents obtained by subtraction of the currents recorded after 5 minutes of a 10 μM ATP application from the control traces using BAPTA-filled electrodes. Each point in the graph represents the mean ± SE of four cells. Right panel: Time course of current amplitude activated by the ramp protocol at +40 mV before, during, and after 10 μM ATP application. Each point in the graph represents the mean ± SE of four experiments. Note that, in the BAPTA filled electrode, no effect of ATP was recorded. (B): The figure shows a typical experiment in which the presence of MRS 2179 (10 μM) prevents the increase of both inward and outward currents elicited by 10 μM ATP. Right panel: Current-voltage relationship of current amplitude (measured at the peak) evoked with the voltage protocol (from −70 to +60 mV, holding potential −80 mV) before and 5 minutes after ATP application in combination of 10 μM MRS 2179. Data represent the mean ± SE of four experiments. (C): The bar graphs report the average current values at +40 mV (left panel: 5 minutes ATP application) and at −90 mV (right panel: 3 minutes ATP application) evoked by the ramp protocol in the presence of ATP alone or in combination with different P2 purinergic antagonists. Left panel: The ATP-induced increase in outward currents is fully blocked by 10 μM MRS 2179 (n = 7) but not by 30 μM PPADS application (n = 4). Right panel: Both the antagonists are effective in blocking the ATP-evoked increase in inward currents. ∗ p < .05 versus 10 μM ATP alone (one-way analysis of variance, Newman-Keuls post-test). Abbreviations: BAPTA, 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid; min, minutes; MRS 2179, 2′-deoxy-N6-methyladenosine3′,5′-bisphosphate; pA, picoAmpere; pF, picofarad; PPADS, pyridoxalphosphate-6-azophenyl-2′,4′-disulfonate.

Effects of ATP on Ionic Currents in Potassium-Free Conditions

To better clarify the nature of the ATP-sensitive currents in hMSCs, we investigated the effect of ATP under K+-free conditions by substituting all of the K+ ions of the intracellular and extracellular solution with equimolar Cs+. In this experimental condition, cells showed a mean resting membrane potential of −5.9 ± 2.8 mV (n = 7). As shown in Figure 6A, 10 μM ATP increased the overall currents evoked by the ramp protocol, mainly affecting inward currents, as indicated also by the ATP-sensitive current (right panel, Fig. 6A). The same effect was recorded in all seven cells tested, where the maximal increase in inward currents (from −368.6 ± 73.5 pA in control conditions to −818.1 ± 175 pA in the presence of ATP) was reached after 3 minutes of ATP superfusion, and it desensitized after 5 minutes of agonist application (Fig. 6B). The averaged ATP-sensitive current showed a reversal potential of −6.5 ± 0.6 mV, indicating the unspecific nature of the ATP-activated conductance (Fig. 6C). It is to be mentioned that this effect resembles that observed in the presence of ATP in a group of cells, as shown in Figure 3C. The fact that, in K+-free conditions, ATP affected mainly inward current suggests that the ATP-sensitive outward currents, also shown in Figures 3A and 4, are carried by K+ ions.

Figure Figure 6..

In K+-free conditions, only the increase in inward currents was elicited by ATP. (A): Left panel: Representative experiment recorded in K+-free conditions obtained by substitution with equimolar Cs+. The application of 10 μM ATP mainly increases inward currents elicited by the ramp protocol. Right panel: ATP-sensitive current recorded in the same cell. (B): Time course of current amplitude recorded at −90 mV before, during, and after 10 μM ATP application. Each point in this graph represents the mean ± SE, which is shown only at representative times (n = 7). Note that the effect of ATP on inward currents reached a maximum at 3-minute application and then returned to the predrug level, showing a desensitizing response. (C): Averaged ATP-sensitive current amplitude elicited by the ramp protocol and recorded in the seven cells investigated in K+-free conditions. Abbreviations: min, minutes; pA, picoAmpere.

Effects of ATP on hMSCs in the Perforated Patch-Clamp Configuration

The effects of ATP were also studied in the perforated patch configuration. Also in these experimental conditions, a 5-minute application of 10 μM ATP elicited two different responses. In three out of five cells investigated, ATP induced a consistent increase in outward currents and a modest enhancement in the inward currents evoked by the ramp protocol. In these cells, the mean ATP-sensitive current (Fig. 7A, filled circles) showed a reversal potential of −38.7 ± 1.2 mV. In the remaining two cells, 10 μM ATP increased inward currents, producing an ATP-sensitive current with a reversal potential of −5.6 mV (Fig. 7A, open circles). By comparing the ATP-sensitive currents recorded in the perforated patch configuration (Fig. 7A) with that observed in the whole-cell configuration, shown in a typical experiment in Figure 3A and 3C and averaged in Figure 7B, it can be noted that, in both experimental conditions, ATP elicited comparable effects. In both cases, we can distinguish two groups of responding cells: cells responding to ATP with an increase in outward currents (filled circles) and cells responding to ATP with an increase in inward currents (open circles).

Figure Figure 7..

The same dual pattern of responses to exogenous ATP is observed in the perforated patch and whole-cell configurations. (A): The graph shows the ATP-sensitive currents evoked by the ramp protocol in the perforated patch configuration. Filled circles: In three out of five cells, the ATP-sensitive current presented a marked outward component with a minor effect on inward currents and a reversal potential of approximately −40 mV. Each point in this graph represents the mean ± SE, which is shown only at representative times. Open circles: In the remaining two cells, the ATP-sensitive current shows a prevalent inward component, with a reversal potential of approximately 0 mV. (B): The graph shows the ATP-sensitive currents evoked by the ramp protocol in the whole-cell configuration (K+-containing solutions: experimental section described in Fig. 3). Filled circles: In 10 cells, the ATP-sensitive current presents a marked outward component with a minor inward current and a reversal potential around −40 mV. Open circles: In six cells, the ATP-sensitive current shows a marked inward rectification with a reversal potential around 0 mV. Each point in this graph represents the mean ± SE, which is shown only at representative times. Abbreviation: pA, picoAmpere.

Discussion

In this work, we demonstrate for the first time that cultured hMSCs, from P0 to P5, spontaneously release ATP and express functional P2 purinergic receptors, which are involved in cell proliferation and in the modulation of ionic conductances. Our results that cultured hMSCs spontaneously release ATP are in agreement with Kawano et al. [36]. The mean ATP values estimated from P0 to P5 (∼7.5 nM) found in the present work are higher than mean extracellular ATP values (∼3 nM) estimated by Kawano and coworkers [36] at later culture passages, from P6 to P19. The difference might be due to the different culture passages of cells used.

In the present study we demonstrated, at early cell passages from P2 to P4, an enhanced cell number of hMSCs in comparison with saline-treated cells when the P2Y1 receptor antagonist MRS 2179 or the unselective P2 blocker PPADS are added to the culture medium. Conversely, exogenous application of ATP in hMSC culture medium induced a modest but significant decrease in cell proliferation. This reduction in cell number cannot be ascribed to a toxic effect of ATP on cell viability, as demonstrated by 7-AAD flow cytometric analysis of the cells and confirmed by a study demonstrating that hMSCs (differently from bone marrow hematopoietic stem cells) are not susceptible to ATP-induced cell death even at millimolar concentrations [41]. Different lines of evidence suggest a predominant role of extracellular ATP during early steps of hMSC differentiation. Kawano et al. [36] reported that ATP-induced initiation and propagation of intracellular Ca2+ waves in hMSCs promote activation of transcription factors (i.e., NFAT) that are involved in cell differentiation [36]. In the same work, the authors also demonstrated that ATP-induced Ca2+ waves in hMSCs disappear in the fully differentiated adipogenic phenotype. Furthermore, a high extracellular ATP concentration at earlier cell culture passages suggests an important role of ATP in regulating cell differentiation. On these bases, we hypothesize that an increased hMSC differentiation could account for our results, in which ATP decreased cell proliferation at early culture passages. Data in the literature support that ATP is involved in trophic mechanisms in different types of both differentiated and undifferentiated cells [14, [15], [16], [17]–18, 21]. It can be envisaged that ATP, as an autocrine/paracrine homeostatic regulator, exerts different effects on cell trophism according to the rate of efflux and receptor expression during cell life cycle.

Our data demonstrate that exogenous ATP modulates ionic conductances in hMSCs by inducing two different responses. By imposing voltage ramp protocols to evoke overall currents in a wide range of membrane potentials, we observed that a population of cells responded to exogenous ATP with a marked increase in outward currents that reached a maximum at the end of drug application (5 minutes). This effect was accompanied by a small enhancement of the inward component. In the other group of cells, exogenous ATP enhanced only the inward conductances. The inward conductances reached a maximum at 3 minutes and decayed at the end of the 5-minute agonist superfusion, indicating a desensitizing response. The same dual pattern of ATP responses was observed when the cells were recorded in the perforated patch-clamp configuration. When K+ ions were replaced by equimolar Cs+, ATP elicited only one type of response, that is, an increase in the inward currents, demonstrating that the outward ATP-sensitive currents are carried by K+ ions.

The increase in outward K+ currents observed in the presence of ATP is abolished by MRS 2179, a selective P2Y1 receptor antagonist, suggesting that such currents are modulated by P2Y1 receptor activation. Conversely, in the presence of PPADS, an unselective P2X-P2Y antagonist, an increase in outward currents was still elicited in the presence of ATP. This fact could be due to the expression and activation in hMSCs of PPADS-insensitive P2Y receptors (such as the human P2Y2 and P2Y11 receptors [42, 43]). An involvement of a metabotropic receptor in the modulation of outward K+ currents is also supported by the observation that the maximal effect is reached at the end of ATP application (5 minutes). Since P2Y receptors (P2Y1,2,4,6,11) stimulate PLC, with subsequent IP3 formation and Ca2+ release from intracellular stores [44], it is likely that this signaling cascade activates the ATP-sensitive K+ conductances (IKCa) observed in the present work, as also confirmed by experiments conducted in the presence of BAPTA. In fact, the presence of BAPTA in the pipette solution completely prevented the ATP-induced increase in outward K+ currents, demonstrating the Ca2+ dependence of such currents (IKCa).

These data are in agreement with the electrophysiological profile of hMSCs in culture present in the literature, which report the presence of three distinct K+ currents with different activation and inactivation kinetics in this cell line [34, 35]. In these works, 94% of tested cells present a rapidly activating and noise-like current, identified as the Ca2+-activated K+ current (IKCa); 48% of cells show a typical delayed rectified K+ current (IKDR), and a small fraction of cells (8%) presents a rapidly inactivating transient outward current, similar to the cardiac and neuronal Ito. Almost all investigated cells express a combination of the three K+ currents that is not correlated to cell diameter or passage in culture. In our experiments, the ATP-sensitive outward K+ conductances were studied in isolation by applying a voltage step protocol. Our results indicate that at least two different types of K+ conductances are involved in ATP effects. When the holding potential is clamped at −80 mV, the effect observed in the presence of ATP is the enhancement of both the transient early currents and late sustained currents. The ATP-sensitive conductance obtained in these cells shows a highly noisy trace, suggesting the involvement of high conductance Ca2+-dependent K+ channels (BK channels). However, when we inactivate transient conductances by a prestep potential at −40 mV, ATP still enhances outward currents, demonstrating that late sustained K+ conductances, most likely represented by IKDR, are also positively modulated by the P2 purinergic agonist. Both currents are activated by P2Y1 receptors since they are completely blocked in the presence of the selective P2Y1 antagonist MRS 2179.

The inward ATP-sensitive current presents the typical features of a nonselective cationic conductance, as expected after P2X receptor activation. It shows a reversal potential of 0 mV and a weak inward rectification, as determined by the I-V relationship. By using a fast perfusion technique, we could detect that the maximal peak of inward currents elicited by ATP is reached in 6 seconds, and desensitization is reached in mean after 1.3 seconds. This is a time span consistent with a P2X-mediated effect. In agreement with the involvement of a P2X-like receptor is the observation that ATP-induced inward current is associated with a concomitant membrane depolarization, as already demonstrated [45]. The increase in inward currents and membrane depolarization induced by ATP were prevented by PPADS, a nonselective P2 purinergic antagonist, and by MRS 2179, which mainly affects P2Y1 receptors. PPADS is known to block P2Y1,2,4,6 receptors and recombinant homomeric P2X1,2,3,5, as well as heteromeric P2X2/3 and P2X1/5 receptor subtypes [4]. Even if MRS 2179 is generally reported as a P2Y1 antagonist, it must be considered that it also blocks some P2X receptors in transfected cells [46]. In agreement with this observation, we envisage that a native system, such as immature stem cells at early stages in culture, can express peculiar assemblies of heteromeric P2X receptors that are blocked by MRS 2179, in addition to the classic MRS 2179-sensitive P2Y1 subtype. Alternatively, in the attempt to explain the MRS 2179 sensitivity of the ATP-sensitive inward current, we can ascribe this effect to a P2Y1-mediated activation of “transient receptor potential” (TRP) channels, a family of nonselective cation channels specifically activated by Gq protein-coupled receptors [47]. There is no evidence, to date, for TRP channel expression in hMSCs. Anyway, we cannot rule out the hypothesis that, in hMSCs, exogenous ATP stimulates P2Y1 receptors that lead to TRP channels opening with a consequent influx of cations in the cell, as demonstrated in megakaryocytes [48].

When voltage ramps are recorded in Cs+ containing solutions, even if all K+ conductances are blocked by Cs+ replacement in these conditions, the inward ATP-sensitive current also shows a small outward component. We hypothesize that, in this case, the outward current is carried by Cs+ ions themselves, permeating through P2X channels, since it has been demonstrated that some P2X receptors expressed in native systems are also partially permeable to Cs+ ions [49, [50]–51], given the large size of the pore diameter [6].

Finally, it is to be noted that the effects of ATP on inward currents are always present in hMSCs, even in cells where ATP mainly induces an increase in outward currents. This is confirmed by the observation that the reversal potential of the ATP-sensitive outward current is −40 mV, which is different from the Nernst-calculated K+ equilibrium potential in our experimental conditions (EK = −81.7 mV). These results suggest that all hMSCs can potentially respond to ATP with an increase in inward currents, but, when K+ channels or P2Y1 receptors are present, the outward current activation overlaps the effect on inward currents.

Accordingly, in the latter case, no changes in membrane potential were observed during ATP superfusion. This observation suggests that the depolarization due to the increase in inward currents elicited by ATP is counterbalanced by the outward current increase. In fact, when ATP is applied in the presence of PPADS, which only blocks the ATP-sensitive inward currents, a hyperpolarization is revealed, reasonably due to the ATP-induced activation of K+ channels.

The absence of the outward ATP-sensitive current in some cells may be justified in several ways. For example, it is possible that not all the cells tested express the same subtype(s) of P2X and/or P2Y receptors. Consequently, one possibility is that hMSCs responding to exogenous ATP with an increase in outward currents are provided with P2Y1 receptors, whereas cells that do not show this effect are not. Alternatively, since not all hMSCs express the same subtypes of K+ channels [34, 35], the lack of outward current increase in some hMSCs could be due to the absence of the particular subtype(s) of K+ channels activated by ATP, even in cells expressing P2Y1 receptors. In agreement with the last assumption is the fact that cells in which ATP increases outward currents present a more negative resting membrane potential in comparison with cells in which ATP enhances inward currents, in both perforated and whole cell configurations. The question arises as to whether the two populations of hMSCs, showing distinct responses to exogenous ATP, also present some differences in morphological parameters. However, we reported that cell capacitance, as a measure of cell size, is not different in hMSCs responding to ATP with the increase of inward or outward currents, and that the passage in culture from P0 to P5 does not affect the characteristics of ATP-sensitive currents.

Conclusion

In summary, we demonstrated that ATP is spontaneously released from hMSCs in culture and decreases their proliferation rate. In addition, exogenous ATP produces two different electrophysiological responses, which are coherent with P2Y and P2X receptor stimulation. Since the precise mechanisms by which hMSCs spontaneously differentiate into osteogenic, adipogenic, or chondrogenic phenotypes are still unclear, it is possible that the two different electrophysiological responses to exogenous ATP correlate with different potentialities of hMSC differentiation. Our results suggest that ATP is one of the early factors determining the fate of hMSCs.

Disclosures of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Acknowledgements

We thank Dr. Alessandro Fazzini for technical support and professor Maria Pia Abbracchio for helpful encouragement in our research. This work was supported by Grants from MURST (ex 40%) and Ente Cassa Di Risparmio of Florence, Italy (2006/40600). E.C. and A.M.P. contributed equally to this work.

Ancillary