Brain function is recognized to rely on neuronal activity and signaling processes between neurons, whereas astrocytes are generally considered to play supportive roles for proper neuronal function. However, accumulating evidence indicates that astrocytes sense and control neuronal and synaptic activity, indicating that neuron and astrocytes reciprocally communicate. While this evidence has been obtained in experimental animal models, whether this bidirectional signaling between astrocytes and neurons occurs in human brain remains unknown. We have investigated the existence of astrocyte–neuron communication in human brain tissue, using electrophysiological and Ca2+ imaging techniques in slices of the cortex and hippocampus obtained from biopsies from epileptic patients. Cortical and hippocampal human astrocytes displayed spontaneous Ca2+ elevations that were independent of neuronal activity. Local application of transmitter receptor agonists or nerve electrical stimulation transiently elevated Ca2+ in astrocytes, indicating that human astrocytes detect synaptic activity and respond to synaptically released neurotransmitters, suggesting the existence of neuron-to-astrocyte communication in human brain tissue. Electrophysiological recordings in neurons revealed the presence of slow inward currents (SICs) mediated by NMDA receptor activation. The frequency of SICs increased after local application of ATP that elevated astrocyte Ca2+. Therefore, human astrocytes are able to release the gliotransmitter glutamate, which affect neuronal excitability through activation of NMDA receptors in neurons. These results reveal the existence of reciprocal signaling between neurons and astrocytes in human brain tissue, indicating that astrocytes are relevant in human neurophysiology and are involved in human brain function.
Astrocytes, classically considered supportive cells for neurons without being directly involved in brain information processing, are emerging as important actors in brain physiology. They exhibit cellular excitability based on intracellular Ca2+ variations that occur spontaneously or evoked by different stimuli, such as mechanical stimulation, exogenous ligands, or neurotransmitters released from synaptic terminals (Porter and McCarthy 1996; Perea and Araque 2005a,b; Navarrete and Araque 2008; Shigetomi et al. 2008). In addition, astrocytes may release neuroactive substances, called gliotransmitters, such as glutamate, ATP, D-serine, adenosine, etc. that can influence neuronal excitability and synaptic transmission and plasticity (Serrano et al. 2006; Martín et al. 2007; Perea and Araque 2007; Navarrete and Araque 2008, 2010; Shigetomi et al. 2008; Henneberger et al. 2010; Santello et al. 2011; Navarrete et al. 2012). These findings have led to the establishment of the tripartite synapse concept, in which astrocytes and neurons reciprocally communicate, suggesting that brain physiology results from the coordinated signaling between neurons and astrocytes (Volterra and Meldolesi 2005; Haydon and Carmignoto 2006; Perea et al. 2009). While these ideas are primarily founded on observations made in rodent brain preparations, the existence of astrocyte–neuron bidirectional communication in human brain remains largely unknown.
We have therefore investigated the existence and properties of this astrocyte–neuron communication in human brain tissue, using electrophysiological and Ca2+ imaging techniques in slices of the cortex and the hippocampus obtained from biopsies from pharmacologically intractable epileptic patients undergoing surgical treatment. We found that cortical and hippocampal astrocytes exhibit Ca2+-based intrinsic excitability, and that they respond with transient Ca2+ elevations to neurotransmitters released during synaptic activity. These Ca2+ elevations affect neuronal excitability because they stimulate the release of glutamate from astrocytes that activating NMDARs evoke slow inward currents (SICs) in neurons. Therefore, human astrocytes detect synaptic activity and are capable of release gliotransmitters that act on neuronal receptors.
These findings indicate the existence of bidirectional communication between astrocytes and neurons in human tissue, and suggest that astrocytes may be more relevant cellular elements than previously thought in human brain function.
Materials and Methods
Human Brain Slice Preparation
Human brain tissue was obtained from biopsies from patients (11 males and 15 females; age: 28–59 and 21–51 years old, respectively) diagnosed of drug-resistant temporal lobe epilepsy for more than 7 years (cf. Arellano et al. 2004; Pastor et al. 2010) that underwent surgery to control their seizures. Patient's consent was obtained according to the Declaration of Helsinki, and protocols were approved by the institutional ethical committee (Hospital de la Princesa, Madrid, Spain). Slices (350–400 μm thickness) were prepared in cold ACSF containing (in mM): KCl 3, MgCl2 10, NaHCO3 25, CaCl2 1, glucose 10, and sucrose 250; and then incubated during >1 h at room temperature (21–24°C) in ACSF containing (in mM): NaCl 124, KCl 2.69, KH2PO4 1.25, MgSO4 2, NaHCO3 26, CaCl2 2, and glucose 10, continuously gassed with 95% O2/5% CO2 (pH 7.3). The time between the tissue removal from the patient and the slice preparation was typically 45 min and tissue was kept in cold ACSF-high sucrose during transportation. They were then transferred to an immersion recording chamber and continuously perfused with gassed ACSF at 2 mL/min. Cells were visualized under an Olympus BX50WI microscope or an Olympus FV300 laser-scanning confocal microscope (Olympus Optical). Astrocyte calcium signal and electrophysiological recordings were performed in CA1 and CA3 regions from hippocampus and layers 2/3–5 from the cortex. Because similar results were obtained, data collected from each area were pulled together.
Electrophysiological recordings were made using the whole-cell patch-clamp technique. For neuronal recordings, patch electrodes had resistances of 3–10 MΩ when filled with the internal solution (in mM): KGluconate 135, KCl 10, HEPES 10, MgCl2 1, ATP-Na2 2 (pH 7.3); and for astrocyte recordings, electrodes had resistances of 4–9 MΩ and filled with an intracellular solution (in mM): MgCl2 1, NaCl 8, ATP-Na2 2, GTP-Na2 0.4, HEPES 10, titrated with KOH to pH 7.2–7.25, and adjusted to 275–285 mOsm. Recordings were obtained with PC-ONE amplifiers (Dagan Instruments). Fast and slow whole-cell capacitances were neutralized and series resistance was compensated (≈70%), and the membrane potential was held at −70 mV. Signals were fed to a Pentium-based PC through a DigiData 1440A interface board. Signals were filtered at 1 kHz and acquired at 10 KHz sampling rate. The pCLAMP 10.2 (Axon instruments) software was used for stimulus generation, data display, acquisition, and storage. Theta capillaries (2–5 µm tip) filled with ACSF were used for bipolar synaptic stimulation. The electrodes were connected to a stimulator S-900 through an isolation unit and placed 75–150 µm away from the recording cells.
SICs were distinguished from miniature synaptic currents (mEPSCs) by their relatively slower time courses (mEPSCs: τon = 2.9 ± 0.3 ms, τoff fast = 3.6 ± 0.5 ms, τoff slow = 24.9 ± 3.1 ms, n = 18; SICs: τon = 8.9 ± 1.4 ms, τoff = 84.6 ± 13.8 ms, n = 35). Neuronal responses were considered SICs when they showed a τoff ≥ 2 times slower than the τoff, slow of mEPSCs (cf. Perea and Araque 2005a; Navarrete and Araque 2008; Shigetomi et al. 2008).
Human tissue was prepared as described previously for rodent tissue (Araque et al. 2002; Nimmerjahn et al. 2004; Perea and Araque 2005a). Briefly, Ca2+ levels in human astrocytes were monitored by fluorescence microscopy using the Ca2+ indicator fluo-4. Slices were incubated with fluo-4-AM (2–5 µL of 2 mM dye were dropped over the tissue, attaining a final concentration of 2–10 µM and 0.01% of pluronic) and Sulforhodamine 101 (100 µM) for 30–60 min at room temperature. In these conditions, most of the Fluo-4-loaded cells were astrocytes as indicated by their SR101 staining (Nimmerjahn et al. 2004; Dombeck et al. 2007; Kafitz et al. 2008; Takata and Hirase 2008), and confirmed in some cases by their electrophysiological properties (see Fig. 2B). Astrocytes were imaged with an Olympus FV300 laser-scanning confocal microscope or a CCD camera (Retiga EX) attached to the Olympus BX50WI microscope. Cells were illuminated during 200–500 ms with a xenon lamp at 490 nm using a monochromator Polychrome V (TILL Photonics), and images were acquired every 1–2 s. Polychrome V and CCD camera were controlled and synchronized by the IP Lab software (BD Biosciences) that was also used for quantitative epifluorescence measurements. Ca2+ variations recorded at the soma of the cells were estimated as changes of the fluorescence signal over baseline (ΔF/F0), and regions of interest were considered to respond to the stimulation when ΔF/F0 increased 3 times the standard deviation of the baseline for at least 2 consecutive images and with a delay ≤15 s after the stimulation. Local application of WIN (300 µM), ATP (20 mM), and glutamate (0.8 mM) was delivered by pressure pulses or by iontophoresis through a micropipette (2–5 s). Mean Ca2+ wave velocity was estimated from the ratio of the distance and the time delay of the calcium elevations between the nearest and the farthest responding cells relative to the stimulating pipette.
Experiments were performed at room temperature (21–24°C). Data are expressed as mean ± SEM. Results were compared using a 2-tailed Student's t-test (α = 0.05). Statistical differences were established with P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***).
Drugs and chemicals
D-(-)-2-Amino-5-phosphonopentanoic acid (D-AP5) and thapsigargin were purchased from Tocris. Fluo-4-AM was from Molecular Probes. All other drugs were purchased from Sigma-Aldrich Co.
We monitored intracellular Ca2+ levels in astrocytes in hippocampal and cortical slices (Fig. 1). Human astrocytes in both hippocampal and cortical slices displayed spontaneous Ca2+ elevations, but with differential Ca2+ activity manifested as different number of active cells and oscillation frequencies (Fig. 1A and B). These spontaneous Ca2+ elevations were abolished by perfusion with 1 µM thapsigargin for 30–45 min (3 slices), which depletes the internal stores by inhibiting the Ca2+ ATPase (Araque et al. 1998), but were insensitive to 1 µM TTX, which prevents action potential generation (Fig. 1B), indicating that they were mediated by Ca2+ release from internal stores and were independent of action potential-mediated neuronal activity. Therefore, human astrocytes display Ca2+ excitability that is due to intrinsic properties independent of neuronal network activity.
Murine astrocytes express a wide variety of neurotransmitter receptors coupled with intracellular Ca2+ signaling (Verkhratsky et al. 1998; Perea and Araque 2005b; Volterra and Meldolesi 2005; Haydon and Carmignoto 2006; Perea et al. 2009). We then asked whether the Ca2+ signal in human astrocytes could be triggered by activation of neurotransmitter receptors that are known to mediate neuron–astrocyte communication in rodents. We locally applied either ATP (20 mM), glutamate (0.8 mM), or the CB1 receptor agonist WIN (300 µM), from a micropipette (tip diameter: ∼2 µm) by iontophoresis or pressure pulses (2–5 s duration). Local application of these agonists evoked transient Ca2+ elevations in astrocytes in both hippocampal and cortical slices (Fig. 1C and D). While the Ca2+ signal evoked by WIN was restricted to a small area under the delivery micropipette, those evoked by glutamate or ATP propagated concentrically as waves to larger distances at similar rates (∼30 µm/s) in hippocampal and cortical slices (Fig. 1D and E) (cf. Oberheim et al. 2009; Kuga et al. 2011). Both the Ca2+ signal extension and the propagation speed were unaffected by TTX (Fig. 1E) but abolished by thapsigargin (5.7 ± 0.8% from control; 3 slices). Although an indirect effect mediated by neuronal activation cannot be totally ruled out due to the lack of receptor specificity, the insensitivity of the astrocytic responses evoked by these transmitters to TTX suggests a direct activation of astrocytic receptors. Taken together, these results indicate that human astrocytes responded to transmitter receptor activation with Ca2+ elevations mediated by Ca2+ mobilization from internal stores, and suggest that they could respond to synaptically released neurotransmitters.
To test this hypothesis, we recorded astrocyte Ca2+ levels in hippocampal and cortical slices while electrically stimulating axons with electrodes placed 75–150 µm away from the recorded cells. Electrical simulation effectively elicited neurotransmitter release, as confirmed by the excitatory postsynaptic potentials (EPSPs) and currents (EPSCs) recorded in neurons (Fig. 2A, C, F). Trains of electrically stimuli (30 Hz, 5 s) reliably evoked Ca2+ elevations in 76% of recorded astrocytes (n = 29 astrocytes; 11 slices) (Fig. 2D and E). In some cases, the Ca2+ signal was also monitored in patch-clamp recorded cells identified as astrocytes by their electrophysiological properties, that is, high membrane conductance and the absence of action potentials (Fig. 2B and D), confirming that responding cells were astrocytes (cf. Bordey and Sontheimer 1998; Hinterkeuser et al. 2000; Schröder et al. 2000). These astrocyte responses were abolished by TTX (11.4 ± 4.0% from control values, n = 4 slices; Fig. 2G and H), indicating that they were unlikely due to direct astrocyte stimulation but rather depended on synaptic activity. Furthermore, although trains of high frequency nerve stimulation were used to clearly reveal the astrocyte responsiveness to synaptic activity (cf. Araque et al. 2002, Perea and Araque 2005a,b; Haas et al. 2006), the sensitivity of the astrocytic responses to TTX (Fig. 2G and H), the fact that synaptic currents could be reliably evoked after train stimulation (Fig. 2F) and that successive trains of electrical stimuli could reliably induce astrocytic responses (Fig. 2E) indicate that they are mediated by a physiological process, that is, action potential generation, rather than resulting from artifact damage. Therefore, human astrocytes respond to synaptic activity with Ca2+ elevations, indicating the existence of neuron-to-astrocyte communication in human brain tissue.
We next investigated the consequences of astrocyte Ca2+ signal on human neurons. In hippocampal slices, local application of ATP evoked astrocyte Ca2+ elevations that propagated as a wave throughout the Stratum radiatum reaching the Stratum pyramidale, and then evoking Ca2+elevations in pyramidal neurons after a conspicuous delay from the initial astrocyte Ca2+elevations (Fig. 3A), suggesting that astrocyte Ca2+ stimulates the release of gliotransmitters that acting on transmitter receptors affect the intracellular Ca2+ levels in human neurons.
Murine astrocytes can release the gliotransmitter glutamate, which activating NMDARs evokes SICs in neurons (Parri et al. 2001; Fellin et al. 2004; Perea and Araque 2005a; Haydon and Carmignoto 2006; Navarrete and Araque 2008; Shigetomi et al. 2008; Perea et al. 2009; Bardoni et al. 2010; Sasaki et al. 2011), we asked whether this signaling also occurred in human brain slices. We recorded hippocampal and cortical whole-cell currents from neurons in the absence of extracellular Mg2+ to maximize NMDAR activation. In these conditions, the presence of spontaneous SICs, which were distinguished from mEPSCs by their relatively slower time courses (see Materials and Methods section) (Fig. 3B–D; cf. Perea and Araque 2005a; Shigetomi et al. 2008; Sasaki et al. 2011), was observed (mean amplitude: 23.3 ± 4.3pA; n = 46 from 6 neurons; Fig. 3B–D; cf. Perea and Araque 2005a; Shigetomi et al. 2008; Sasaki et al. 2011). Local application of ATP, which elevated Ca2+ levels in astrocytes (Fig. 3A), also increased the frequency of SICs in both hippocampal and cortical neurons (Fig. 3E and F). While SIC frequency was insensitive to TTX (n = 3 neurons), SICs were abolished by 50 µM AP5, indicating that they were independent of action potential-evoked neurotransmitter release and that they were mediated by NMDARs (Fig. 3E and F). Therefore, in agreement with compelling evidence obtained in rodents (Parri et al. 2001; Fellin et al. 2004; Perea and Araque 2005a; Navarrete and Araque 2008; Shigetomi et al. 2008; Bardoni et al. 2010; Sasaki et al. 2011), Ca2+ elevations in human astrocytes stimulate the release of glutamate that activates NMDARs in neurons, indicating the existence of gliotransmission and astrocyte-to-neuron communication in human brain tissue.
Present results indicate that human astrocytes in situ display spontaneous Ca2+ elevations, respond with Ca2+ elevations to applied neurotransmitter receptor agonists, and, more importantly, to synaptically released neurotransmitters. Human astrocytes are also able to release the gliotransmitter glutamate, which activates postsynaptic NMDARs and evoke SICs in neurons. Taken together, these results demonstrate that human astrocytes not only sense synaptic activity, but also regulate neuronal excitability through glutamate release, showing the existence of bidirectional communication between neurons and astrocytes in human brain tissue and suggesting that astrocytes are playing active roles in human brain function. Furthermore, the fact that astrocytes and neurons establish functional units in human brain tissue indicate that Tripartite Synapses, as functional entities originally postulated in rodents, are also present in more phylogenetically evolved brains, perhaps as an intrinsic property common to all nervous systems.
Cell physiology studies on human brain tissue are necessary limited and present several constraints that restrict the feasible experimental procedures. Furthermore, the accessible samples are far from ideal control samples. For example, present work is based on tissue obtained from patients that usually suffered long-lasting epileptic conditions and underwent pharmacological treatments. Nevertheless, present study shows that the basic principles of the neuron–astrocyte communication previously described in animal models, that is, astrocyte responsiveness to synaptic activity and astrocyte ability to release gliotransmitters that act on neurons, exist in human brain. Further studies are required to fully characterize the properties of the astrocyte–neuron signaling in the human epileptic tissue and their role in the pathophysiological process of the temporal lobe epilepsy.
One potential concern about the physiological relevance of data obtained in human brain tissue derives from the fact that the biological sample may correspond to unhealthy or abnormal tissue, which could render results more related to pathological rather than physiological phenomena. Furthermore, normal physiological conditions may be altered by the previous pharmacological treatment of patients as well as by the procedures involved, from the clinical surgery until the tissue samples reach the experimental bench. However, besides the well-documented morphological alterations of the epileptic tissue (e.g. Honavar and Meldrum 1997; Mathern et al. 2000; Arellano et al. 2004; Alonso-Nanclares et al. 2011), the cellular properties of the recorded cells in the present study indicate that the cellular viability is largely preserved in these cells. Indeed, recordings of calcium-based cellular activity as well as electrophysiological parameters of neurons and astrocytes (membrane resting potential, I-V curves, action potential amplitudes, synaptic currents, etc.; Fig. 2A and B) suggest that the observed phenomena do not result from a damaged tissue. Furthermore, present results show similar properties of the spontaneous astrocyte calcium signal in human and rodent tissues (Aguado et al. 2002; Nett et al. 2002; Takata and Hirase 2008; Kuchibhotla et al. 2009; Sasaki et al. 2011), suggesting that they reflect normal physiological characteristics. Nevertheless, we cannot discard that the normal properties were actually altered due to the pathological conditions.
The presence of NMDAR-mediated SICs induced by glutamate released from astrocytes have been shown in different rodent brain areas, specifically, CA1 hippocampal area, ventrobasal thalamus, nucleus accumbens, and neocortex (Parri et al. 2001; Angulo et al. 2004; Fellin et al. 2004; Fellin et al. 2006; Perea and Araque 2005a; Ding et al. 2007; Nestor et al. 2007; Navarrete and Araque 2008; Shigetomi et al. 2008; Pirttimaki et al. 2011; Sasaki et al. 2011). Whether these events are general or specific of certain brain regions remains unknown. Perhaps, the existence of SICs requires specific structural and functional interactions, such as spatial location of glutamate release sources, spatial distribution of glutamate transporters, or reduced glutamate accessibility to postsynaptic NMDA receptors. The physiological meaning of these SICs in human tissue is unknown, as in rodents, although these currents have been shown to depolarize postsynaptic neurons and to serve as a synchronizing mechanism for neuronal activity (Angulo et al. 2004; Fellin et al. 2004). In contrast to the available data obtained in rodent brain slices, the presence of SICs and their role in vivo still need experimental confirmation. Nevertheless, the presence of SICs in human brain tissue indicates the existence of structural and functional relationships between astrocytes and neurons in human brain, and demonstrates the ability of human astrocytes to release gliotransmitters that can activate neuronal receptors.
Studies performed in animal models reveal a high richness in the signaling processes and physiological consequences of the astrocyte–neuron communication (Volterra and Meldolesi 2005; Haydon and Carmignoto 2006; Perea et al. 2009). It is noteworthy that due to technical limitations, the analysis of the astrocyte calcium signal in the present study was restricted to global calcium events occurring at the astrocytic cell body. However, the actual calcium signaling events in astrocytes are underestimated in this analysis because recent studies have revealed the existence of important calcium signals limited to astrocytic processes that may have important implications in synaptic physiology (Di Castro et al. 2011; Panatier et al. 2011). The fact that the basic cellular machinery is present in human brain tissue, that is, astrocyte Ca2+ signal triggered by neurotransmitter receptors activated by neurotransmitters released during synaptic activity, astrocytic gliotransmitter release, and neuronal receptor activation by gliotransmitters, suggests that the more complex properties described in rodent brains, such as astrocyte processing of synaptic information and complex modulation of synaptic activity and plasticity by multiple astrocytic signals (Perea and Araque 2005a, 2007; Serrano et al. 2006; Henneberger et al. 2010; Navarrete and Araque 2010; Santello et al. 2011; Navarrete et al. 2012), may also take place in human brain. Perhaps even more intricate relationships between astrocytes and neurons may exist considering the higher structural complexity of human astrocytes (Ramón y Cajal 1913; Oberheim et al. 2006,2009; Matyash and Kettenmann 2010).
Our current knowledge of the morphological and physiological properties of human astrocytes is largely incomplete (see Matyash and Kettenmann 2010). However, Ramón y Cajal (1913) reported that human cerebral cortex differed from other animals in the higher number of astrocytes and their richer arborizations, and more recent studies have elegantly provided further insights in the morphological properties of human astrocytes, revealing their higher complexity and heterogeneity as well as unique characteristics (Oberheim et al. 2006, 2009). Little is also known about the physiological properties of human astrocytes. Some studies have explored their electrophysiological properties (e.g. Bordey and Sontheimer 1998; Hinterkeuser et al. 2000; Schröder et al. 2000; Seifert et al. 2004, 2006; Black et al. 2010; Matyash and Kettenmann 2010), and their Ca2+-based excitability has been recently reported (Oberheim et al. 2009). However, nothing is known about the possible functional interactions with neurons, that is, their ability to respond to synaptic activity and to release gliotransmitters that act on neuronal receptors. To our knowledge, this study represents the first demonstration of the existence of these functional properties in astrocytes interactions, suggesting the presence of bidirectional communication between astrocytes and neurons in the human brain.
While recent reports have questioned the relevance of astrocyte Ca2+ signaling on neurophysiology (see Agulhon et al. 2008), accumulating evidence continue to confirm the relevance of this signaling in the detection and control of neuronal and synaptic activity (Haydon and Carmignoto 2006; Serrano et al. 2006; Perea and Araque 2007; Perea et al. 2009; Henneberger et al. 2010; Navarrete and Araque 2010; Panatier et al. 2011; Santello et al. 2011; Navarrete et al. 2012). Present results add further evidence that supports such relevance, indicating that Tripartite Synapses also exist in human brain.
In conclusion, present data show that human astrocytes detect synaptic activity and release gliotransmitter that affect neuronal function, indicating the existence of functional bidirectional communication between neurons and astrocytes in human brain tissue. These results indicate that astrocytes are relevant in human neurophysiology and that astrocyte–neuron signaling is involved in human brain function.
Conflict of Interest: None declared.
This work was supported by grants from Ministerio de Ciencia e Innovación, Spain (BFU2010-15832, CSD2010- 00045), European Union (HEALTH-F2-2007-202167), and Cajal Blue Brain to AA, and ISCIII (PS09/02116), Spain, to J.P. and R.G.S. G.P. is supported by a Marie Curie International Outgoing Fellowship (FP7-253635).
We thank W. Buño, E.D. Martin, and A. Perez-Alvarez for helpful comments, and Iván Rodríguez for technical assistance with tissue samples.