Graphene oxide electrodes enable electrical stimulation of distinct calcium signalling in brain astrocytes.
Free full text in Europe PMC
Abstract
Free full text
Graphene oxide electrodes enable electrical stimulation of distinct calcium signalling in brain astrocytes
Abstract
Astrocytes are responsible for maintaining homoeostasis and cognitive functions through calcium signalling, a process that is altered in brain diseases. Current bioelectronic tools are designed to study neurons and are not suitable for controlling calcium signals in astrocytes. Here, we show that electrical stimulation of astrocytes using electrodes coated with graphene oxide and reduced graphene oxide induces respectively a slow response to calcium, mediated by external calcium influx, and a sharp one, exclusively due to calcium release from intracellular stores. Our results suggest that the different conductivities of the substrate influence the electric field at the cell–electrolyte or cell–material interfaces, favouring different signalling events in vitro and ex vivo. Patch-clamp, voltage-sensitive dye and calcium imaging data support the proposed model. In summary, we provide evidence of a simple tool to selectively control distinct calcium signals in brain astrocytes for straightforward investigations in neuroscience and bioelectronic medicine.
Main
Astrocytes are brain glial cells responsible for maintaining brain homoeostasis; they are capable of sensing and responding to different extracellular chemophysical stimuli (such as neurotransmitters, temperature, osmotic and ion gradients, mechanical stimulus) by changes in their concentration of intracellular calcium ([Ca2+]i)1–3. Astrocytic [Ca2+]i mediates the release of so-called gliotransmitters (such as glutamate, d-serine and ATP)4–8 and modulates the activity of neighbouring astrocytes, neurons and vascular cells. Through these molecules, astrocytes regulate cerebral blood perfusion9,10 or modulate the excitation/inhibition balance of brain excitability at different spatiotemporal scales4,11–13. Dysfunctions in [Ca2+]i dynamics contribute to the pathogenesis of all neurological disorders14–16 characterized either by cognitive impairment or by alteration in the vascular flow.
The increase in astroglial [Ca2+]i can be caused by (1) the influx of extracellular Ca2+ (EXT-Ca2+) or (2) the release of intracellular Ca2+ (INT-Ca2+) from the cytoplasmic stores2. The two distinct mechanisms display different spatiotemporal dynamics of [Ca2+]i (ref. 3). EXT-Ca2+ influx across the cellular membrane is a slow process, typically lasting hundreds of seconds17–19, although some reports identify a faster EXT-Ca2+ signal at astrocytic endfeet facing blood vessels20–22. EXT-Ca2+ mainly enters the cell via ion channels1–3, such as members of the transient receptor potential superfamily, which includes transient receptor potential vanilloid 4 (TRPV4)11,23,24 and transient receptor potential ankyrin 1 (TRPA1)25, and voltage-gated calcium channels (VGCCs), which are expressed in primary and brain astrocytes26–28. INT-Ca2+ relies on triggered release of Ca2+ stored in endoplasmic reticulum or in mitochondria2,6,21, on a scale of seconds. The release of Ca2+ from the intracellular stores is largely mediated by inositol 1,4,5-trisphosphate (IP3), as a consequence of the increase of [Ca2+]i (refs. 21,29) or of activation of plasma membrane metabotropic G-protein coupled receptors (GPCRs)3,29–31, or possibly by ryanodine receptor (RyR), located on the astrocytic endoplasmic reticulum32. The Ca2+ signalling of astrocytes is also spatially distinct, as it occurs in the astrocytic soma, and in localized Ca2+ domains within or at the endfeet of fine astrocytic elongations (microdomains)4,6,24,25, or spread between astroglial cells through gap junctions (Ca2+ waves)6,7.
The diversity of [Ca2+]i signals in astrocytes has distinct functional roles3,4,24,25, which, however, are far from being fully understood, and controversial reports exist in the literature4,20,21,33,34.
In this context, a major issue is the limited availability of technologies and tools to selectively activate and control distinct Ca2+ pathways in astrocytes35, avoiding the use of complex, and potentially artefactual, genetic modifications36,37.
A bottleneck of current bioelectronic devices and protocols is that they are designed to interface or trigger neuronal cells and they are rarely suitable and useful to study astrocytes, whose functional properties differ from those of neurons37,38. In this respect, bioelectronic approaches to selectively drive specific and distinct Ca2+ signalling in astrocytes are lacking.
Graphene, a single two-dimensional layer of hexagonal structure consisting of sp2-hybridized carbon atoms, due to its combination of high electrical conductivity, high flexibility, chemical inertness and biocompatibility properties39,40, has been used in devices to perform high-resolution brain mapping41 or as electrodes to alter the excitability of neurons in vitro and in vivo42. The broad family of graphene materials includes graphene oxide (GO) and reduced GO (rGO), which, while maintaining similar morphologies, display different physicochemical and conductive properties43,44. In this respect, the potential advantages of using GO-based devices to interrogate signalling in glial cells have been recently highlighted, but never exploited to control [Ca2+]i dynamics in astrocytes38,45.
Here we prove the possibility of selectively evoking diverse astroglial [Ca2+]i responses with electrical stimulation through different GO/rGO-coated electrodes. We propose a biophysical model in which the insulating/conductive properties of the GO or rGO interfaces are determinant to drive distinct [Ca2+]i responses. Experimental data achieved with patch-clamp, calcium and voltage-sensitive dye imaging experiments support the proposed model. In brain tissues, GO/rGO devices stimulate distinct astrocytic Ca2+ dynamics, but astrocytes respond more promptly and more efficiently to electrical stimulation than neurons. Notably, electrical stimulation by GO/rGO electrodes in vitro and ex vivo activates different GPRC intracellular pathways, known to be involved in astrocyte–neuron bidirectional communication, in astrocytes and neurons2,30,31.
GO and rGO electrical stimulation activates distinct Ca2+ dynamics in astrocytes
The procedures to obtain GO- and rGO-coated electrodes as well as the characterization of their chemophysical properties are reported in Methods and in Supplementary Information (Supplementary Fig. 1 and Supplementary Results and Discussion Section 1). The electric properties of the substrates were studied at the nanoscale using conductive atomic force microscopy (AFM) (Fig. 1a,b). The current map of GO substrates shows that most of the surface is uniform44 and insulating (Fig. (Fig.1a),1a), with some highly conductive areas of uncovered indium tin oxide (ITO) delimited by the linear edges of GO flakes, while only conductive areas are present in the rGO film (Fig. (Fig.1b).1b). Insulating and conductive properties of GO, rGO and ITO are confirmed by current-voltage (I–V) curves (Supplementary Fig. 1h).
a,b, AFM characterization of GO and rGO coatings. Topography (left panels) and corresponding PeakForce TUNA current (right panels) images of GO- (a) and rGO-coated ITO electrodes (b). The current image is taken with a voltage bias Vb=
1
V. The scale bars are all 1
µm. c, Fluorescent images of astrocytes stained with fluorescein diacetate (FDA) and Hoechst 33342 for GO (left panel) and rGO (right panel). Scale bars, 25
μm. d, Bar–dot graph reporting the number of cells per area counted on the different samples. Data are presented as mean
±
s.e.m. For ITO, n
=
5, N
=
3, no. of cells/area
=
20.2
±
2.6. For rGO, n
=
6, N
=
3, no. of cells/area
=
15.8
±
1.9. For GO, n
=
6, N
=
3, no. of cells/area
=
17
±
3.7. For GO×10, n
=
6; N
=
3, no. of cells/area
=
41.8
±
5.7. n, number of analysed images, N, number of experiments. Statistical significance was calculated via one-way analysis of variance (ANOVA) with Bonferroni post-test. P values are reported in the graph when P
≤
0.05, which was considered significant. No statistically significant differences were observed between ITO and rGO (P
=
0.2), GO and rGO (P
=
0.8) or ITO and GO (P
=
0.5). e, Scheme of the experimental set-up for the electrical stimulation, showing the direction of the applied electric field (E). Electrical stimulus was delivered by ramping up the substrate voltage using as a reference an Ag/AgCl grounded electrode immersed in the same saline solution as the sample. The applied voltage protocol was low enough to provide an electrical field suitable for cell stimulation. The voltage protocol consisted in a continuous voltage ramp increasing from 0.1 to 0.8
V in 85
s at a rate of 8.24
mV
s−1. The total length of the experiment was 300
s, and the voltage stimulus was applied 25
s after the start of the recording. f, Bar–dot graph reporting transcript levels of the inflammatory marker gfap in astrocytes plated on ITO, rGO and GO before (NO STIM) and after (STIM) electrical stimulation. The y value corresponds to the levels of expression of gfap mRNA normalized with respect to the relative values for β-actin. Data are presented as mean
±
s.e.m. N
=
4 per condition, in triplicate (pooled). Normalized gfap mRNA levels are for ITO NO STIM 1
±
0, for ITO STIM 0.5
±
0.01, for rGO NO STIM 1
±
0, for rGO STIM 0.6
±
0.1, for GO NO STIM 1
±
0 and for GO STIM 0.9
±
0.2. Statistical significance was calculated via one-way ANOVA with Bonferroni post-test. P values are reported in the graph when P
≤
0.05, which was considered significant. No statistically significant difference was observed between GO NO STIM and GO STIM (P
=
0.6). g,h, Typical [Ca2+]i variations observed in the majority of astrocytes grown respectively on GO- (g) and rGO-coated electrodes (h), in an external standard solution containing Ca2+, when stimulated according to the protocol described above (inset). Sustained S-type signal observed in astrocytes on GO (g) and P-type signal observed on rGO (h).
We then tested the viability of primary neocortical astrocytes on ITO and on ITO coated with rGO, GO and GO×10 (Fig. 1c,d and Supplementary Results and Discussion Section 1). The analyses revealed that viable astrocytes with a typical flat polygonal shape45,46 adhered and grew on all the substrates analysed (Fig. (Fig.1c)1c) and that the numbers of astrocytes were comparable on GO and rGO and significantly higher on GO×10 (Fig. (Fig.1d).1d). These results confirm that GO and rGO, in both film45 and nanoflakes47–49, promote direct growth of astrocytes, without the need for treatment with adhesion molecules, which potentially decrease the electrical coupling with the astrocytic membrane45.
In a previous work, we demonstrated that surface and mechanical properties of GO substrates do not cause an adverse inflammatory reaction, typically observed in response to biomaterial implants50,51, characterized by astrogliotic reactivity and increase in glial fibrillary acidic protein (GFAP) expression35,38,50,51. Thus, we analysed the messenger RNA expression level of gfap, in cells grown on GO/rGO-coated and on bare ITO electrodes (Fig. (Fig.1f),1f), by quantitative real-time PCR performed before and after application of the extracellular stimulation protocol schematized in Fig. Fig.11 (Fig. 1e,f). The mRNA levels of gfap were significantly reduced after stimulation with rGO and ITO but not with GO-coated electrodes (Fig. (Fig.1f).1f). The data reported here suggest that GO/rGO coatings of brain implants may improve the biocompatibility and stability of the bioelectronic neural interface over the longer term38,51.
We next performed calcium imaging in astrocytes before, during and after electrical stimulation (Fig. 1g,h). Ca2+ dynamics were almost absent when stimulation was not applied (NO STIM) to GO- and rGO-coated electrodes (Supplementary Fig. 2a,b), while we observed spontaneous [Ca2+]i in cells on ITO (Supplementary Fig. 2c). The data show that, compared with ITO and GO nanoflakes47,48, GO and rGO films have the advantage of not harming basal Ca2+ signalling, that is, the basal excitation level, in primary astrocytes.
Surprisingly, in response to electrical stimulation, astrocytes display diverse temporal Ca2+ dynamics, depending on the GO/rGO interface used to deliver the stimulation. In the majority of cells on GO, electrical stimulation caused a slow, ‘sustained’ rise in [Ca2+]i, which we will call hereafter an ‘S’-type signal (Fig. (Fig.1g),1g), acting on a timescale of hundreds of seconds and lasting up to 5min from the end of the stimulation (Supplementary Fig. 3a,b). The response in the remaining cells was negligible, while spontaneous activity was observed in a few cells.
Conversely, electrical stimulation caused in most of the cells on rGO a rapid response, with well defined, sharp oscillating ‘peaks’ of ~30s duration, which we will describe hereafter as a ‘P’-type [Ca2+]i signal (Fig. (Fig.1h),1h), that were still persistent 10
min after the stimulation (Supplementary Fig. 3c,d). The maximal amplitude (ΔF/F) of the response to electrical stimulation was stronger and had a longer time to peak on GO than on rGO substrates, while the P-type signal on rGO also showed a significantly higher number of peaks (Supplementary Fig. 4b–d).
We then performed comparative pharmacological analyses using selective inhibitors of different Ca2+ paths in astrocytes1–4,17,23–30, while stimulating cells on GO- and rGO-coated electrodes (Fig. (Fig.22 and Supplementary Fig. 4a–d; additional information is reported in Supplementary Results and Discussion Section 3) and ITO electrodes (Supplementary Fig. 5).
a,b, Representative traces of Ca2+ imaging observed after positive voltage bias stimulation of astrocytes, starting at time t≈
25
s from the beginning of the experiment (insets to panels 1) plated on GO–ITO- (a) and on rGO–ITO-coated electrodes (b). Different panels refer to the different conditions of the cells exposed to standard bath solution (CTRL, 1) and solution without extracellular Ca2+ (NO EXT-Ca2+, 2) and in the presence of VGCC inhibitor verapamil (VERAP, 25
μM, 3), TRPV4 inhibitor RN-1734 (RN, 10
μM, 4), TRPA1 inhibitor HC-030031 (HC, 40
μM, 5), IP3 receptor pathway inhibitor 2-aminoethoxy diphenyl borate (2-APB, 100
μM, 6), SERCA inhibitor cyclopiazonic acid (CPA, 10
μM, 7), RyR activator caffeine (CAFF, 20
mM, 8), RyR inhibitor ryanodine (RYAN, 50
μM, 9), Gq–PLC inhibitor U73122 (0.5
μM, 10) and Gi/o inhibitor pertussis toxin (PTX, 500
ng
ml−1, 11).
We found that EXT-Ca2+ influx through TRPV417 and TRPA125 was essential for the S-type Ca2+ signalling amplitude and onset, observed in response to the electrical stimulation provided by GO-coated electrodes. Blocking of VGCCs26 delayed the onset of the response on GO, but not its magnitude (Fig. (Fig.2a2a panels 3,4,5 and Supplementary Fig. 4a,e). P-type signals became visible and more frequent in GO samples in the absence of EXT-Ca2+ and when VGCCs were inhibited (Fig. (Fig.22 panels 2,3 and Supplementary Fig. 4b). The results also suggest that the IP3 path and SERCA are not essential for the onset of the response on GO (Fig. (Fig.2a2a panels 6,7 and Supplementary Fig. 4e). However, the significant decrease in the ΔF/F observed on blocking these INT-Ca2+ pathways52 (Supplementary Fig. 4a) suggests that they could be implicated in sustaining the [Ca2+]i response over time, through a Ca2+-induced Ca2+ increase mechanism, as previously described in astrocytes11. The data indicate that blocking RyR52 had no significant effects on the S-type amplitude and dynamic, observed in response to GO-coated electrodes, while the effect of the application of the RyR agonist could be artefactual53 (Fig. (Fig.2a2a panels 8,9 and Supplementary Fig. 4a,e). Finally, experiments stimulating astrocytes on GO, while blocking Gq signalling or Gi/o signalling (Fig. (Fig.2a2a panels 10,11)30,31, revealed that, while the Gq–PLC–IP3 pathway (PLC, phospholipase C) is activated by stimulation with GO-coated electrodes and is implicated in the dynamics (onset and time to peak) and in the magnitude of the response, Gi/o is not relevant for this effect (Supplementary Fig. 4a,d,e).
Conversely, the pharmacological analyses indicated that stimulation by rGO-coated electrodes exclusively triggers INT-Ca2+ release, through IP3, SERCA and Gi/o paths, blocking of which alters either the magnitude, the percentage of responding cells or the dynamics of the P-type calcium response (Fig. (Fig.2b2b panels 6,7, 11 and Supplementary Fig. 4a–e). The Gq–PLC–IP3 pathway is not activated by stimulation through rGO electrodes, while RyR is important for the onset of the response (Fig. (Fig.2b2b panels 8–10 and Supplementary Fig. 4a–e).
Notably, P-type response in cells on rGO-coated electrodes is not abolished by blocking EXT-Ca2+ influx, through omission of EXT-Ca2+ or inhibition of VGCCs, TRPV4 or TRPA1 (Fig. (Fig.2b2b panels 2–5 and Supplementary S4a–e). However, the results indicate that EXT-Ca2+ influx through VGCCs and TRPA1 may set the basal Ca2+ levels18,26 needed for INT-Ca2+ to occur25,29 (Fig. (Fig.2b2b panels 3,5 and Supplementary Fig. 4b,d,e).
Bioelectrical modelling of GO/rGO–astrocyte interface and experimental validation
Although the detailed description of our system would require the microscopic modelling of the charged ion distribution induced by the applied potential, following previous models developed for electrolytic solutions54,55 we propose a qualitative bioelectric scheme at the GO/rGO–astrocyte interface, where the main differences are due to the insulating/conducting properties of the substrates. Despite its simplicity, our picture can account for all the observed behaviour (Supplementary Results and Discussion Section 4).
The electrical circuit (Fig. (Fig.3a3a lower panel) is approximated to its main components: Rsub, Csub are the resistance and capacitance of the substrate; Rsystem, Csystem are the total resistance and capacitance of all the other parts (cells, bulk solution, Ag/AgCl electrode), which are constant for all samples.
a,b, Schematic representation of the proposed mechanism taking place during GO (a) and rGO (b) stimulation and of the consequent cellular response. Upper panels: a, In the case of GO, charge accumulation at the GO–cell interface (1) causes depolarization of the membrane, which promotes opening of VGCCs or TRPA1 and EXT-Ca2+ influx (2). 3, Ca2+ entry leads to calcium-induced calcium release via IP3 or SERCA but not RyR. 4, The IP3 path potentiates the Ca2+ influx mediated by TRPV4 via the calcium-induced calcium increase mechanism13,57. The entrance of further external Ca2+ into the cell causes a steady increase of cytoplasmic Ca2+ (S-type signal). TRPA1 might be involved in this process as a cooperative channel promoting either maintenance of basal Ca2+ levels or potentiation of the Ca2+ influx over time25,56. IP3Rs, IP3 receptors. b, In the case of rGO, charge accumulation occurs at the cell–solution interface, inducing depolarization of the cell membrane at the electrolyte–cell interface (1) which might stimulate directly electrically/mechanically the endoplasmic reticulum58 (2) causing the release of INT-Ca2+ from the stores. 3, The above-mentioned electric field might repulse cations at the cell–electrolyte interface, thus hampering the EXT channel in mediating Ca2+ influx. Lower panels: the potential drop across the substrate (GO, a, and rGO, b), and the direction of electric fields created by the potential applied to the substrate (Esub). The electric fields created on the cell walls by the membrane potential (Emem), pointing inside the cell, are also shown. Bottom panel: the scheme of the equivalent electric circuit, as described in the text.
GO substrate
In GO coatings, RsubRsystem and most of the potential drop takes place inside the GO film, with a minor voltage drop at the interface with the cells. In an electrolytic system, the interface voltage drop generates an electric field EF oriented from the substrate towards the inside of the cell. This field is synergic with the potential of the membrane in contact with the substrate, and almost negligible at the more distant upper membrane (Fig. (Fig.3a),3a), allowing cell membrane depolarization.
Considering the pharmacological evidence as well, we propose that GO stimulation induces the S-type signal through the following mechanism.
EF at the interface induces opening of VGCCs and TRPA1, which are activated by depolarizing voltages26,56 (Fig. (Fig.3a),3a), allowing the influx of EXT-Ca2+.
EXT-Ca2+ stimulates INT-Ca2+ release from intracellular stores via the IP3 path, which potentiates the Ca2+ influx mediated by TRPV4 via the calcium-induced calcium increase mechanism11,57.
The further entrance of EXT-Ca2+ causes a sustained, steady increase of cytoplasmic Ca2+. TRPA1 might be involved in this process, promoting maintenance of basal Ca2+ levels and potentiation of the Ca2+ influx25. Other molecular players, such as pannexin 1, could be involved (Supplementary Results and Discussion Section 5), but additional studies employing genetic deletion or short interfering RNA will be needed to clarify its role.
rGO substrate
In rGO coatings, Rsub is low and the potential decreases only marginally across the rGO layer, while there is a substantial potential drop, and thus an electric field (oriented from the substrate to the solution), within the cell and at the cell–solution interface (Fig. (Fig.3b).3b). EF is opposite to the potential of the upper membrane cell, hindering the intake of external Ca2+. Assuming a cell thickness d=
10
μm, the magnitude (E
=
V/d) can be as high as several kilovolts per metre, generating a negative electrochemical gradient that
depolarizes the cell membrane at the electrolyte–cell interface,
stimulates electrically the endoplasmic reticulum58 or generates an osmotic gradient resulting in an electrical or mechanical stimulation that
causes the release of INT-Ca2+ from the stores.
EF repulses cations at the cell–electrolyte interface, hampering the EXT channel in mediating Ca2+ influx.
Our interpretation is corroborated by the observation that, in the case of ITO substrate, cells behave similarly to the rGO case (Supplementary Fig. 5). Also, a previous work studying astrocytes on insulating substrates46 displayed the S signal, in agreement with the GO case.
Finally, the results reported in Supplementary Fig. 6 indicate that the thickness of GO does not alter the functional response of the cells and that a quasimonolayer coating of GO is sufficient to trigger the onset of EXT-Ca2+ entry in astrocytes (Supplementary Results and Discussion Section 6).
To obtain insights into the mechanism through which astrocytes sense the voltage stimulation and to validate the bioelectrical modelling, we performed single-cell patch-clamp experiments and FluoVolt voltage-sensitive dye (VSD) imaging in a plurality of cells (Extended Data Fig. Fig.11 and Supplementary Figs. 7 and 8).
a, Typical patch-clamp current clamp traces, showing Voltage membrane variations over time, recorded before, during and after the application of the same voltage protocol of Ca2+ experiments (insets). b, Bar–dot graph reporting the averaged variation of voltage membrane (Vmem (mV)), before the stimulation (PRE) and at the maximal value recorded after the stimulation (POST) of astrocytes plated on GO (red bar) and rGO (black bar) devices. Data are presented as mean±
Standard Error of the mean. n=number of analysed cells, N=number of experiments. For GO: n
=
16, N
=
4, Vmem PRE (mV)=-13.8
±
3.3, Vmem POST (mV) =-1.75
±
4.7. For rGO, n
=
11, N
=
4, Vmem PRE (mV)=-15.3
±
3.7, Vmem POST (mV)=-3
±
7. Statistical significance was calculated via one-way ANOVA with Bonferroni post-test. p values are reported in the graph when p
≤
0.05, which was considered significant. No significant differences were observed between Vmem PRE GO and Vmem PRE rGO (p
=
0.8) and between Vmem POST GO and Vmem POST rGO (p
=
0.6). c, Bar–dot graph reporting the time point of the onset (Tonset(s)) of the response for astrocytes plated on GO (red bar) and rGO devices (black bar) devices. Data are presented as mean
±
Standard Error of the mean. n=number of analysed cells, N=number of experiments. For GO: n
=
14, N
=
4, Tonset (s)= 14.4
±
3.1. For rGO: n
=
5, N
=
4, Tonset (s)=31.8
±
7.8. Statistical significance was calculated via one-way ANOVA with Bonferroni post-test. p values are reported in the graph when p
≤
0.05, which was considered significant. d, e, Typical ΔF/F over time of Fluovolt VSD imaging, observed in response to the same voltage stimulations protocols applied for Ca2+ imaging experiments, on astrocytes plated on GO (d, e, left panels, red traces) and on rGO (d, e, right panels, black traces) before, during and after the stimulation with positive (d) and negative biases (e). Positive voltage protocol (d) is the same described in Fig. Fig.1,1, negative voltage protocol had the same duration but inverted polarity (e, V from -0.1
V to -0.8
V). f, g, Bar–dot graphs of fluorescence variation measurements performed on GO and rGO: f) maximal averaged fluorescence variation after the stimulation (ΔF/F), g) onset time of the response (Tonset (s)), measured using positive (+) and negative (-) voltage protocols for astrocytes stimulation. Data are presented as mean
±
Standard Error of the mean. nROI= number of Regions Of Interest, N=number of experiments. For GO (+): n ROI
=
270, N
=
5, ΔF/F
=
0.010
±
0.0008, Tonset (s)=51.2
±
7. For rGO (+): n ROI
=
335, N
=
5, ΔF/F
=
0.009
±
0.002, Tonset (s)=101.6
±
14.7. For GO (-): nROI=156, N
=
3, ΔF/F
=
0.008
±
0.002, Tonset (s)=62
±
7. For rGO (-): nROI=367, N
=
6, ΔF/F
=
0.022
±
0.003, Tonset (s)=56.5
±
2.8. Statistical significance was calculated via one-way ANOVA with Bonferroni post-test. p values are reported in the graph when p
≤
0.05, which was considered significant. No significant differences were observed between ΔF/F GO (+) and ΔF/F GO (-) (p
=
0.3), between ΔF/F GO (+) and ΔF/F rGO (+) (p
=
0.7), between Tonset (s) GO (+) and Tonset (s) GO (-) (p
=
0.3).
Current-clamp recordings revealed that in resting condition the membrane voltages (Vmem) of cells on GO and rGO did not differ significantly (Supplementary Fig. 7b), and that in response to electrical stimulation (Extended Data Fig. Fig.1a)1a) cells depolarize, reaching values close to 0mV, for both GO and rGO (Extended Data Fig. Fig.1b).1b). However, the onset of the depolarization was much faster on GO than on rGO (Extended Data Fig. Fig.1c),1c), thus indicating the different kinetics of the two effects.
The FluoVolt voltage-sensitive dye imaging analyses59 indicated that in response to electrical stimulation the depolarization occurs in a plurality of cells, with variation in Vmem values (ΔF/F) comparable between GO and rGO (Extended Data Fig. Fig.1f),1f), and that the onset of the response was much faster in cells on GO when compared with rGO (Extended Data Fig. Fig.1g).1g). Thus, voltage-sensitive dye results are in line with the patch-clamp analyses (compare Supplementary Fig. 7c with Extended Data Fig. Fig.1f1f and Extended Data Fig. Fig.1c1c with Extended Data Fig. Fig.1g1g).
In experiments inverting the polarity of the same voltage ramp stimulation protocol, the variation in Vmem in cells on rGO was significantly higher than that of cells on GO, but comparable to that of cells on ITO (Supplementary Fig. 7d). The onsets of the response to negative biases were comparable in the cases of rGO and GO (Extended Data Fig. 1f,g).
Calcium imaging experiments indicated that negative voltage ramp stimulation did not elicit any significant Ca2+ response on GO (Supplementary Fig. 8a lower panel), while it triggered a slow Ca2+ variation on rGO, which resembled the S type, with a significantly lower number of detected peaks (Supplementary Fig. 8b lower panel,e).
Overall, the data are in agreement with the model and suggest the following.
Positive bias on GO-coated electrodes induces rapid astrocyte depolarization due to voltage membrane variation at the GO–cell interface.
Positive bias on rGO-coated electrodes alters whole-cell potential, hindering external Ca2+ intake and activating a different Ca2+ signalling pathway when compared with GO.
Negative bias on GO-coated electrodes does not significantly depolarize astrocytes but creates an electric field against the Ca2+ intake (Supplementary Results and Discussion Section 7).
Negative bias on rGO alters whole-cell potential, favouring intake of external Ca2+.
GO and rGO stimulate S-type and P-type Ca2+ dynamics in astrocytes and neurons ex vivo
We next performed experiments ex vivo, in brain slice samples from GFAP–enhanced green fluorescent protein (eGFP) transgenic mice (Supplementary Fig. 10d–f), using flexible ITO electrodes coated with GO or rGO (Supplementary Fig. 10a–c and Supplementary Results and Discussion Section 8).
Notably, S-type and P-type calcium dynamics (Fig. (Fig.4a)4a) were observed predominantly in slices respectively stimulated with GO- or rGO-coated electrodes. Statistical analysis (Fig. 4b,c) confirms that the different mechanisms observed on GO and rGO in vitro take place also in astrocytes ex vivo. In addition, high-magnification two-photon imaging evidences that the Ca2+ signal evoked by electrical stimulation either by GO or rGO is first triggered in the soma and then propagates along the astrocytic elongation5–7 (Fig. 4d,e and Supplementary Fig. 10h–n). The results were also confirmed in differentiated astrocytes in vitro (Supplementary Fig. 9 and Supplementary Results and Discussion Section 9).
a, Representative traces of Ca2+ imaging experiments performed on brain slices lying on GO (left) and rGO devices (right), using the same voltage protocol as described before (inset). b,c, Bar–dot graphs of maximal averaged fluorescence variation (ΔF/F, b) and number of peaks (c), measured on GO and rGO devices. Data are presented as mean±
s.e.m. For GO, N
=
6, s
=
13, n
=
108, ΔF/F
=
0.20
±
0.02, no. of peaks
=
1.02
±
0.01. For rGO, N
=
4, s
=
9, n
=
142, ΔF/F
=
0.16
±
0.01, no. of peaks
=
1.37
±
0.07. For GO NO STIM, N
=
2, s
=
2, n
=
13, ΔF/F
=
0.10
±
0.01. For rGO NO STIM, N
=
2, s
=
2, n
=
16, ΔF/F
=
0.09
±
0.01. n, number of analysed cells; s, number of slices. Statistical significance was calculated via one-way ANOVA with Bonferroni post-test. P values are reported in the graph when P
≤
0.05, which was considered significant. d,e, Representative traces of [Ca2+]i over time (right) performed with high magnification on X-Rhod-1/GFAP–eGFP-labelled astrocytes (merged images, left) for slices on GO (d) and on rGO (e), analysed in astrocytic soma and in astrocytic processes.
Given that the surface of the slice or of the cell is smaller than the surface of the electrode interface, it is plausible to suppose that the whole cell is responding to the changes in the electrolyte bath due to electrical stimulation provided by the GO and rGO interfaces.
We next studied the effects of electrical stimulation in the neurons close to the analysed astrocytes of the same brain slice (Fig. 5a,b and Supplementary Results and Discussion Section 10). We found that S-type and P-type Ca2+ dynamics could be recorded in astrocytes and in the nearby neurons of brain slices respectively on GO- (Fig. (Fig.5c5c left) and on rGO-coated electrodes (Fig. (Fig.5c5c right). Remarkably, in both cases, the response was faster and the magnitude of the response was significantly higher in astrocytes than in neurons of slices lying on rGO (Supplementary Fig. 11a,d).
a, Confocal fluorescence microscopy image of GFAP–eGFP/X-Rhod-1-AM-labelled cells revealing the co-presence of astrocytes (yellow cells) and neurons (red cells). b, Immunohistochemical image of GFAP–eGFP-labelled astrocytes (green cells) and neuronal cell protein marker (NeuN)-positive neurons (red cells) in brain slice. c–e, Representative traces of Ca2+ imaging experiments performed on brain slices lying on GO and rGO, analysed in neurons and astrocytes, recorded in control saline (c) and after exposure to U73122 (4μM) (d) and Gi/o–GPCR inhibitor PTX (7.5
μg
ml−1) (e).
Given the results described above and the importance of GPCRs in astroglial Ca2+ signalling participating in astrocyte–neuron cross-talk2,30,31 (Supplementary Results and Discussion Section 10), we investigated the role of Gq and Gi/o in the response to electrical stimulation by GO- or rGO-coated electrodes, in astrocytes and in the nearby neurons (Fig. 5d,e).
Collectively, the results indicated that GO and rGO activate different GPCR signalling pathways in cortical brain astrocytes and in neurons ex vivo.
- i.
The Gq–PLC–IP3 pathway is critically involved in the Ca2+ signal evoked in astrocytes in response to GO electrodes, but not in the response of astrocytes to stimulation by rGO electrodes (Fig. (Fig.5d5d and Supplementary Fig. 11).
- ii.
Gi/o–Ca2+ signalling is activated in astrocytes when stimulated by rGO and not by GO (Fig. (Fig.5e5e and Supplementary Fig. 11).
These data are in agreement with results achieved in vitro (compare Fig. Fig.22 and Supplementary Fig. 4 with Fig. Fig.55 and Supplementary Fig. 11; Supplementary Results and Discussion Section 10).
- iii.
Neuronal response to GO electrical stimulation involves activation of Gq signalling.
- iv.
Gi/o activation is implicated in the onset of the Ca2+ signal induced by GO and rGO stimulation in neurons.
The results (Fig. (Fig.55 and Supplementary Fig. 11) show that, in the case of GO, when the astrocyte Ca2+ signal is decreased by Gq inhibition, the neuronal calcium signal is also significantly smaller (Fig. (Fig.5d5d left and Supplementary Fig. 11a). Also, the delay in the onset observed in neurons, in the control condition, is no longer significant in the presence of Gq inhibition (Fig. (Fig.5c5c and Supplementary Fig. 11d). Thus, it is plausible to suppose that GO stimulation evokes Ca2+ signalling in astrocytes and in neurons that might be correlated temporally, spatially and in amplitude through a Gq–PLC–IP3-mediated event, such as gliotransmission2,4,30,31. However, we cannot rule out that GO electrical stimulation directly evokes a Gq pathway in neurons that could lead to neuronal Ca2+ rise60. Recent evidence suggests that activation of Gi/o–Ca2+ signalling could mediate the release of glutamate from astrocytes30,31,61, which in turn might excite neurons. Although we found that rGO stimulation activates Gi/o-mediated Ca2+ signalling in astrocytes, the dynamics of neuronal and astroglial calcium signalling seem to be independent.
We cannot exclude the possibility that the release of gliotransmitters through pathways other than GPCRs61 could occur in response to electrical stimulation by GO and rGO, or that the impact of the astroglial stimulation on neurons might result in a different outcome than the Ca2+ rise. Future studies will aim to elucidate these aspects.
Conclusions
We have demonstrated that the electrical properties of GO and rGO (insulating/conductive) can be used to selectively stimulate electrically the EXT-Ca2+ influx or INT-Ca2+ release from primary and brain astrocytes. Our observations are rationalized within an electrostatic/bioelectric qualitative model. Specifically, we found the following.
Electrical stimulation provided by GO-coated electrodes triggers EXT-Ca2+ influx. VGCCs, TRPA1 and TRPV4 are critical for sensing, transduction and onset of the response on GO, while IP3/SERCA-mediated release of INT-Ca2+ is implicated in sustaining the Ca2+ response over time.
Electrical stimulation provided by rGO electrodes exclusively triggers IP3-, SERCA- and RyR-mediated release of INT-Ca2+.
Stimulation by GO and rGO depolarizes astrocytes, with different onsets, evidencing their ability to sense the different electric fields caused by the presence of the insulating GO or conductive rGO interface.
The GO or rGO interface elicits distinct GPCR Ca2+ signalling in astrocytes in vitro and ex vivo: the Gq–PLC–IP3 pathway is activated only by stimulation with GO electrodes; Gi/o–Ca2+ signalling is activated only when astrocytes are stimulated by rGO.
Cortical neurons respond more slowly than astrocytes ex vivo. In neurons, Gq signalling is activated by GO electrical stimulation and Gi/o activation is implicated in the onset of the Ca2+ signal induced by both GO and rGO stimulation.
The spatiotemporal complexity of astrocyte Ca2+ patterns caused uncertainties about how different signals participate in the physiology and pathophysiology of astrocytes. Indeed, controversies exist in the literature on the implication of INT-Ca2+ release in the modulation of synaptic function33, and the beneficial/detrimental role of its alteration in brain pathologies14–17. Similarly, the contribution of EXT-Ca2+ influx in neurovascular coupling and arterial contraction needs to be clarified12,34. Decoding the physiological meaning of these dynamic changes in astrocytic Ca2+ activity to explain the underlying mechanisms has remained a major challenge.
The possibility to trigger different cellular events by electrical stimuli is particularly attractive on our platform since GO can be easily transformed into rGO and vice versa, allowing us to produce GO/rGO patterns at the nanometric scale by simple methods. We anticipate that micrometric array devices can allow simultaneous stimulation/modulation in a spatially selective way from local circuits up to large brain areas.
In a longer-timescale perspective, the use of our devices might target selectively novel neuromodulatory effects in physiological and pathological conditions such as ischaemia, epilepsy and spreading depression in which the diverse nature of astrocyte [Ca2+]i signalling is implicated14–17.
In vitro, in vivo and clinical studies indicated that different electrical stimulations46,62,63 are able to excite neuronal cells, but questions arise of whether the effect is on neurons directly64 or, as proposed, on surrounding glial cells63. In this respect, when we interfere with GPCR signalling, which is critical for astrocyte–neuron cross-talk30,31, we find that stimulation by a GO-coated electrode induces a Gq-mediated astrocyte Ca2+ signalling that might impact on neuronal Ca2+ signalling. However, given the lack of cell specificity of the GPCR inhibitors and the existing constraints of experimental models studying gliotransmission ex vivo and in vivo4, the data herein reported cannot be conclusive on the question of whether the stimulation by GO/rGO interferes with astrocyte vesicular release and neuro-glial communication or if there is any possible gliomodulation effect. These questions will require a devoted future study.
We have not considered possible different osmotic–mechanical gradients induced by the applied electric field46. Future studies, targeting the expression of ions and water channels mediating mechanic/osmotic sensation in astrocytes24,65, will be useful to clarify this aspect.
Overall, the simple electronic interface approach proposed here could be useful to explore the function of astrocytes in fundamental neuroscience investigation and in neuropathologies.
Methods
Device fabrication
The GO solution used for the preparation of the samples was obtained through a modified Hummer method. The substrates of ITO on glass (Kintec) and of ITO on polyethylene terephthalate (Techinstro) with dimensions 2.5cm
×
2.5
cm were cleaned using sonication at 60
°C in acetone and subsequently in isopropyl alcohol, followed by a cleaning treatment with air plasma. Aqueous solutions of GO with a concentration of 1 or 2
mg
ml−1 were deposited on these samples by spin coating at 2,000
r.p.m. The rGO samples were prepared by annealing some of the deposited GO samples (from aqueous solutions with concentration of 2
mg
ml−1) at 200
°C for 2
h under vacuum. The GO×10 samples were prepared by repeating ten times the spin coating of a 2
mg
ml−1 GO solution on the ITO-covered glass to obtain a thick (approximately 20
nm) layer.
AFM investigation
Surface topography and current images were taken with a commercial Multimode 8 microscope (Bruker) operated in air, using the PeakForce-TUNA module and employing a Bruker cantilever. PeakForce-TUNA allows the simultaneous acquisition of the sample topography and spatial-resolved tunnelling current (TUNA) with picoampere resolution.
Cell culture preparations
Primary rat cortical astroglial cultures were prepared as described previously46, according to the Italian law on protection of laboratory animals, with the approval of bioethical committees of the University of Bologna and of the Ministry of Health (ID 1138, code number 2DBFE.N.3CN, ex-protocol number 360/2017-PR) under the supervision of the veterinary commission for animal care and comfort at the University of Bologna. Every effort was made to minimize the number of animals used and their suffering. Briefly, after removing the meninges, the cerebral cortices of 1–2-day-old Sprague Dawley pups (P0–P2) were mechanically dissociated and placed in cell culture flasks containing DMEM–GlutaMAX medium supplemented with 15% fetal bovine serum, 100U
ml−1 penicillin and 100
mg
ml−1 streptomycin (all products were purchased from Gibco-Invitrogen). Culture flasks were maintained in a humidified atmosphere incubator at 37
°C and 5% CO2 for three to four weeks. The culture medium was replaced every 3
d. Before medium change, flasks were gently shaken to detach microglial cells seeded on top of the astrocytic monolayer. At confluence, astroglial cells were enzymatically dispersed using trypsin–EDTA. Cells were then seeded at high concentration on GO-based devices and maintained in culture medium containing 10% fetal bovine serum.
For Ca2+ imaging experiments conducted in differentiated astrocytes, subconfluent astrocytes plated on GO devices were treated with 500µm adenosine 3ʹ,5ʹ-cyclic monophosphate, N6,O2ʹ-dibutyryl-, sodium salt and maintained in vitro before Ca2+ imaging measurements46,66.
Acute brain slice preparation
Ex vivo experiments were performed on acute brain slices of GFAP–eGFP transgenic mice, at the ages of 15–25d and of 4
months (ref. 17).
All procedures were performed at the Department of Cellular Neurophysiology, Institute of Experimental Medicine, Czech Academy of Science, in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC) and animal care guidelines approved by the Institute of Experimental Medicine ASCR Animal Care Committee on 17 April 2009, approval number 02/2017.
Mice were anaesthetized with an intraperitoneal injection of 1% pentobarbital diluted in physiological saline and decapitated. Brains were dissected out and placed into a cold (4–8°C) N-methyl-d-glucamine-based isolation solution containing (mM) 110 NMDG-Cl, 3 KCl, 23 NaHCO3, 1.25 Na2HPO4, 0.5 CaCl2, 7 MgCl2, 20 glucose, osmolality ~300
mOsm
kg−1. Coronal 300-µm-thick slices were cut using a vibrating microtome (HM 650
V, Thermo Scientific Microm) and incubated for 30
min at 34
°C in the isolation solution, oxygenated with carbogen. The brain slices were then transferred to artificial cerebrospinal fluid containing (mM) 122 NaCl, 3 KCl, 28 NaHCO3, 1.25 Na2HPO4, 1.5 CaCl2, 1.3 MgCl2, 10 glucose, osmolality ~305
mOsm
kg−1, at room temperature.
Cell viability assays
Cell viability was investigated by FDA/Hoescht assay. The FDA (Sigma) stock solution (5mg
ml−1) was diluted in PBS. Hoechst 33342 (1:2000) was added to the solution46. Astrocytes plated on GO devices were incubated for 5
min at room temperature (22–24
°C), washed with PBS and characterized using a Nikon Eclipse 80i fluorescence microscope, equipped with a ×40 objective. A series of five to ten images was taken from each replicate, after 5
d in vitro, from the date of replating.
Electrical stimulation and calcium microfluorometry in vitro and ex vivo
For experiments in primary culture, variations in [Ca2+]i were monitored with calcium microfluorometry using the single-wavelength fluorescent Ca2+ indicator Fluo-4 AM (Life Technologies). Before measurements, high-density astrocytes seeded on GO devices were loaded with 2µM Fluo-4 AM dissolved in standard bath solution for 45
min at room temperature.
Electrical stimulation was performed by immersing the samples and a standard Ag/AgCl reference electrode in saline bath solution and applying voltage using a custom-made 2612A Dual-channel System SourceMeter instrument (Keithley). Electrical stimulus was delivered by ramping up substrate voltage using as a reference an Ag/AgCl grounded electrode immersed in the same saline solution as the sample. The applied voltage protocol was low enough to provide an electrical field suitable for cell stimulation, while avoiding generation of detrimental Faradaic currents66–68. The voltage protocol consisted in a continuous voltage ramp increasing from 0.1 to 0.8V in 85
s at a rate of 8.24
mV
s−1. The total length of the experiment was 300
s, and the voltage stimulus was applied 25
s after the start of the recording.
Samples were rinsed with standard bath solution after incubation. Measurements of [Ca2+]i were performed using a fluorescence microscope (Nikon Eclipse Ti-S) equipped with a long-distance dry objective (×40) and appropriate filters. The excitation wavelength was 450nm with a light pulse duration of 200
ms and a sampling rate of 2
Hz. The whole data acquisition was controlled using MetaFluor software (Molecular Devices).
Blockers were diluted in standard bath saline to their respective final concentrations and added after rinsing. For in vitro calcium imaging experiments, cells were considered responding to the stimulus when the maximal variation in fluorescence after the stimulus was higher than 0.02 ΔF/F.
To evaluate the temporal features of [Ca2+]i dynamics, we extracted the average number of peaks by detecting the number of fluorescence oscillations recorded over time, from the beginning of electrical stimulation until the end of the experiment (Supplementary Fig. 4b). When a slow variation occurred, we quantified one peak on average. The average peak number was significantly higher on rGO than on GO, indicating that the response on rGO samples was characterized by a more oscillatory behaviour (Supplementary Fig. 4b). To characterize the diverse [Ca2+]i temporal dynamics observed after GO/rGO-coated electrode operation, we also estimated, for each cell, the average time to reach the maximal fluorescence increase after the voltage stimulus (time to peak, Supplementary Fig. 4c).
The onset was calculated at the time point where we could measure the minimal variation (0.02 ΔF/F) in ΔF/F after the electrical stimulation.
For experiments in brain slices, calcium imaging measurements were performed using an FV1200MPE multiphoton laser scanning microscope (Olympus) equipped with a ×20 water objective. Fluorescence was excited in a two-photon absorption mode at 850nm using a MaiTai DeepSee tunable Ti–sapphire laser system (Spectra Physics). The laser system operated in a pulse mode with 80
MHz repetition rate, <100
fs pulse width, using an IR average power of ~91
mW. A fluorescence signal selected with a 495–540
nm band-pass emission filter was detected using a GaAsP detector. A fluorescence signal emitted from cells loaded with X-Rhod-1-AM dye was selected with a 575–630
nm band-pass emission filter and detected using a GaAsP detector. Data acquisition was controlled using FluoView FV1000 software.
For experiments in brain slices, calcium microfluorometry was performed using the single-wavelength fluorescent Ca2+ indicator X-Rhod-1-AM (Life Technologies) to distinguish the signal from eGFP. Brain slices were loaded with 2µM X-Rhod-1-AM dissolved in artificial cerebrospinal fluid solution for 30
min at 34
°C. After the incubation period, the slices were kept at room temperature (23–25
°C) in artificial cerebrospinal fluid before calcium imaging measurements.
Astrocytes were identified by green fluorescence emission (Supplementary Fig. 10d). The brain slice was also stained with a red-emitting Ca2+ probe, X-Rhod-1-AM, to detect calcium variations (Supplementary Fig. 10e).
RNA extraction and quantitative polymerase chain reaction
RNA was extracted from astrocytes plated on ITO, GO and rGO using 600μl of PureZOL (Bio-Rad 7326880) following the manufacturer instructions. The RNA concentration was measured via a Varioskan LUX microplate reader (Thermo Fisher Scientific) using the specific support μDrop plate (Thermo Fisher Scientific). Samples with an absorbance (260/180) ratio between 1.8 and 2.2 were considered acceptable. Five hundred micrograms of mRNA were used to produce complementary DNA by reverse transcription using iScript reverse transcriptase (Bio-Rad 1708890). Quantitative PCR was performed using a CFX96 Touch real-time PCR detection system (Bio-Rad). Each reaction tube contained 2
μl cDNA, 10
μl iTaq Universal SYBR Green Supermix (Bio-Rad 1725120), 1
μl PrimePCR SYBR Green Assay: GFAP, rat (Bio-Rad 10025636) and RNAse-free water to a final volume of 20
μl. Data were analysed and normalized to the expression of β-actin.
Patch clamp
GO/rGO devices were included in a standard patch-clamp set-up and primary astrocytes plated on GO and rGO were recorded using continuous current-clamp recording mode (Supplementary Fig. 7a). Current-clamp traces were recorded before, during and after extracellular electrical stimulation by GO- or rGO-coated electrodes. Electrophysiological experiments were performed on a set-up based on a Nikon Eclipse Ti-S microscope. Patch-clamp recordings were performed 48–72h after replating in continuous current-clamp mode at room temperature (20–24
°C). Patch pipettes were prepared from thin-walled borosilicate glass capillaries to have a tip resistance of 2–4
MΩ when filled with the standard internal solution. Responses were amplified (Multiclamp 700B, Axon Instruments) and stored on a computer for off-line analysis (pClamp 10, Axon Instruments). The access resistance (below 10
MΩ) was corrected up to 70–90% of the original value by real-time automatic software correction.
Voltage-sensitive dye imaging
For membrane potential imaging, cultured astrocytes plated on GO devices were loaded for 30min at room temperature with a voltage-sensitive dye using the FluoVolt Membrane Potential Kit (Thermo Fisher Scientific, F10488). Samples were rinsed with standard bath solution after incubation and membrane potential variations were measured, using the same electrical stimulation protocols as applied for Ca2+ imaging experiments.
Immunohistochemistry
For immunohistochemical experiments, mice were deeply anaesthetized with sodium pentobarbital (100mg
kg−1 intraperitoneally), and transcardially perfused with 20
ml of saline with heparin (2,500
IU per 100
ml; Zentiva) followed by 20
ml of 4% paraformaldehyde. Brains were dissected, post-fixed in 4% paraformaldehyde overnight, and placed stepwise in solutions with gradually increasing sucrose concentrations (10, 20 and 30%) for cryoprotection. Coronal slices (30
μm) were prepared using a Hyrax C50 cryostat (Zeiss). The slices were incubated in a blocking solution containing 5% ChemiBLOCKER (Merck) and 0.5% Triton X-100 (Merck) in PBS for 1
h. They were then incubated overnight at 4
°C with primary antibodies diluted in a blocking solution, followed by a 2
h incubation with species-specific secondary antibodies diluted in a blocking solution at room temperature. A primary antibody against neuron-specific nuclear protein NeuN (diluted 1:200, Merck) was used as a marker for neurons. Corresponding secondary antibody (goat anti-mouse IgG conjugated with Alexa-Fluor 594) was diluted at 1:200. After immunostaining, the slices were mounted onto microscope slides using Aqua-Poly/Mount (Polysciences)17.
Chemical composition
The chemical state and composition of GO and rGO devices were studied using X-ray photoelectron spectroscopy. Freshly prepared GO on ITO and rGO on ITO were measured immediately after the preparation to avoid contamination. The X-ray photoelectron spectroscopy set-up was composed by a hemispherical analyser (Phoibos 100, SPECS) and a Mg Kα X-ray source (XR50, SPECS).
Solutions and chemicals
Salts and other chemicals of the highest purity grade were purchased from Sigma. For calcium microfluorometry experiments the standard bath solution was composed of (mM) 140 NaCl, 4 KCl, 2 MgCl2, 2 CaCl2, 10 HEPES, 5 glucose, pH 7.4 with NaOH and osmolarity adjusted to ~318mOsm with mannitol.
Calcium-free extracellular saline (NO EXT-Ca2+) contained (mM) 140 NaCl, 4 KCl, 4 MgCl2, 10 HEPES, 0.5 EGTA, pH 7.4 with NaOH and osmolarity adjusted to ~318mOsm with mannitol.
Stock solutions of 2-APB (100mM) and CPA (40
mM) were prepared by dissolving in methanol and stored at −20
°C. Stock solutions of RN-1734 (14.7
mM), HC-030031 (40
mM), verapamil (10
mM), ryanodine (1.25
mM) and U71322 (1.93
mM) were prepared by dissolving in dimethylsulfoxide and stored at −20
°C. Stock solutions of caffeine were prepared by dissolving in water (20
mM) and stored at −20
°C. Stock solutions of PTX (7.5
µg
ml−1) were prepared by dissolving in dimethylsulfoxide and stored at 2–8
°C.
In experiments stimulating astrocytes while blocking Gq–PLC–IP3 signalling, we added U73122 to the standard bath solution. To block Gi/o signalling, cells were incubated in standard solution containing the Gi/o inhibitor PTX for 2h before experiments31.
Statistical analysis
For in vitro experiments, somatic or process cellular fluorescence time series were manually extracted in both MetaFluor (Molecular Devices) and a dynamic-data-exchange Excel file (Microsoft Office 365). Representative traces and statistical analyses of extracted data from in vitro calcium imaging, voltage-sensitive dye and patch-clamp experiments were then performed using Microcal Origin 8.5. Bar–dot plots were generated using Prism GraphPad 8.0.2.
Data were compared using one-way ANOVA with Bonferroni post-test. A statistically significant difference was reported if P≤
0.05. All data are presented as mean
±
s.e.m. Sample size (n) for each statistical analysis is reported in the figure caption referring to the specific result. The data were analysed from at least four independent experiments.
In calcium imaging experiments, the ratio of the fluorescence intensity at each time point and the initial fluorescence was continuously recorded during the experiment (ΔF/F).
For ex vivo experiments, in each slice, the images of eGFP fluorescence were recorded to visualize properly astrocytic cell soma, to set the region of interest in each image during analysis or to perform correction of cell movement. Time series of images of a fluorescence signal emitted by cells loaded with X-Rhod-1-AM dye and eGFP fluorescence were analysed using Fiji (ImageJ software, general public license). The obtained time series of fluorescent intensities (numbers) for individual cells were then exported and further analysed using an Excel template that was generated in the Department of Cellular Neurophysiology, IEM, Prague. Using the Excel template, fluorescence intensities were corrected for photobleaching and fluorescence variation was calculated. In the final analyses changes in fluorescent intensity above 20% of the baseline were considered as responses. The template is available upon request.
For voltage-sensitive dye imaging59, the average of fluorescence intensity over time of the total number of regions of interest for each experiment was calculated and normalized with respect to the initial fluorescence (ΔF/F).
All the calculated means, s.e.m., P values, numbers of experiments (N) and numbers of cells or replicates (n) are reported in the graph and legend of each figure.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Online content
Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at 10.1038/s41565-024-01711-4.
Supplementary information
Supplementary Figs. 1–11, Results and Discussion, and Table 1.
Source data for bar–dot plot and statistics in Supplementary Fig. 4.
Source data for bar–dot plot and statistics in Supplementary Fig. 5.
Source data for bar–dot plot and statistics in Supplementary Fig. 6.
Source data for bar–dot plot and statistics in Supplementary Fig. 7.
Source data for bar–dot plot and statistics in Supplementary Fig. 8.
Source data for bar–dot plot and statistics in Supplementary Fig. 9.
Source data for bar–dot plot and statistics in Supplementary Fig. 10.
Source data for bar–dot plot and statistics in Supplementary Fig. 11.
Source data for statistics in Supplementary Table 1.
Source data
Source Data for statistics in Fig. 1.
Source Data for statistics in Fig. 4.
Source Data for statistics in Extended Data Fig. 1.
Acknowledgements
This work was supported by the Air Force Office of Scientific Research (AFOSR) projects ASTRONIR (FA9550-17-1-0052), ASTROLIGHT (FA9550-20-1-0386), ASTROTALK (FA9550-23-1-0736) and ICARUS (FA9550-21-1047). The research leading to these results has received funding from the European Union’s Horizon 2020 research and innovation programme through the following projects: MSCA-ITN-2020-ASTROTECH (GA956325) and GrapheneCore3 881603—Graphene Flagship. This study was also supported by the Czech Science Foundation (grant number 20-05770S) and by PRIN-PNRR-2022 (P2022Z27NS—Nanotechnological interfaces and devices enabling selective control of ion and water channels dynamics—NANODYN). Financial support from PNRR MUR project ECS_00000033_ECOSISTER is also acknowledged. In situ experiments were carried out at the Microscopy Service Centre of the Institute of Experimental Medicine CAS supported by MEYS CR (LM2023050 Czech-Bioimaging). S. Bin-Salamon, AFOSR Biophysics Programme Manager, and L. Ambrosio from CNR-IPCB are acknowledged for their trust in the vision and the discussion of the results. We are grateful to V. Ragona and F. Bona from CNR-ISMN for their technical support.
Author contributions
Extended data
Notes
All the authors contributed to the manuscript preparation and revision and approved the present version. V.B. conceived the concept, designed the experiments, interpreted the results, discussed the mechanism and coordinated the work. E.T. and V.P. designed the experiments, interpreted the results, discussed the mechanism and coordinated the work. E.T. and A.S. performed the GO/rGO sample preparation and discussed the mechanism. A.C. performed the AFM investigation and discussed the mechanism. A. Kovtun performed the X-ray photoelectron spectroscopy measurements and discussed the mechanism. R.F. performed the calcium imaging and voltage-sensitive dye experiments on primary astrocytes and brain slices, analysed and discussed the results and wrote the manuscript. E.S. performed patch-clamp experiments. R.Z., M.M. and M.U. discussed the results concerning the model. M.C., R.F., E.S., D.S., C.L. and A. Konstantoulaki prepared and maintained primary rat cortical astroglial cultures and analysed biocompatibility data. C.L. contributed to the analysis of voltage-sensitive dye experiments and of calcium imaging in brain slices. A. Konstantoulaki revised the manuscript and provided English proofing of the manuscript. P.D. performed NMR analyses. G.C. performed the real-time quantitative PCR experiments and analysis and revised all the figure formats to comply with the Nature Nanotechnology policy standard. D.K. prepared brain slices and co-supervised experiments on brain slices. V.M. performed immunostaining experiments on brain slices. M.A. supervised the ex vivo experiments, discussed the results and revised the manuscript.
Peer review
Peer review information
Nature Nanotechnology thanks Kuan Zhang and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.
Data availability
Source data are provided with this paper.
Code availability
The template of the code used for analyses of calcium imaging ex vivo is available upon request.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Change history
9/9/2024
A Correction to this paper has been published: 10.1038/s41565-024-01797-w
Contributor Information
Emanuele Treossi, Email: ti.rnc.fosi@issoert.eleuname.
Vincenzo Palermo, Email: ti.rnc.fosi@omrelap.oznecniv.
Valentina Benfenati, Email: ti.rnc@itanefneb.anitnelav.
Extended data
is available for this paper at 10.1038/s41565-024-01711-4.
Supplementary information
The online version contains supplementary material available at 10.1038/s41565-024-01711-4.
References
Citations & impact
This article has not been cited yet.
Impact metrics
Alternative metrics
Discover the attention surrounding your research
https://www.altmetric.com/details/165288767
Similar Articles
To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.
Electrical Stimulation by an Organic Transistor Architecture Induces Calcium Signaling in Nonexcitable Brain Cells.
Adv Healthc Mater, 8(3):e1801139, 19 Dec 2018
Cited by: 5 articles | PMID: 30565894
Selective stimulation of calcium signalling pathways in astrocytes with graphene electrodes.
Nat Nanotechnol, 19(9):1253-1254, 01 Sep 2024
Cited by: 0 articles | PMID: 39014183
An Increase in Membrane Cholesterol by Graphene Oxide Disrupts Calcium Homeostasis in Primary Astrocytes.
Small, 15(15):e1900147, 20 Mar 2019
Cited by: 10 articles | PMID: 30891923
The active role of astrocytes in synaptic transmission.
Cell Mol Life Sci, 56(11-12):991-1000, 01 Dec 1999
Cited by: 54 articles | PMID: 11212330 | PMCID: PMC11146931
Review Free full text in Europe PMC
Funding
Funders who supported this work.
Akademie Věd České Republiky (1)
Grant ID: 20-05770S
Akademie Věd České Republiky (Academy of Sciences of the Czech Republic) (1)
Grant ID: 20-05770S
European Commission (4)
Grant ID: Graphene Flagship
Grant ID: Graphene Flagship- GA 20-05770S
Grant ID: GrapheneCore3 881603 – Graphene Flagship
Grant ID: GA956325
European Commission (EC) (4)
Grant ID: Graphene Flagship
Grant ID: GrapheneCore3 881603 - Graphene Flagship
Grant ID: GA956325
Grant ID: Graphene Flagship- GA 20-05770S
Ministero dell'Istruzione, dell'Università e della Ricerca (2)
Grant ID: PNRR MUR project ECS_00000033_ECOSISTER
Grant ID: PRIN-PNRR-NANODYNE
Ministero dell'Istruzione, dell'Università e della Ricerca (Ministry of Education, University and Research) (2)
Grant ID: PRIN-PNRR-NANODYNE
Grant ID: PNRR MUR project ECS_00000033_ECOSISTER
United States Department of Defense | United States Air Force | AFMC | Air Force Office of Scientific Research (3)
Grant ID: FA9550-20-1-0386
Grant ID: FA9550-23-1-0736
Grant ID: FA9550-21-1-0424
United States Department of Defense | United States Air Force | AFMC | Air Force Office of Scientific Research (AF Office of Scientific Research) (3)
Grant ID: FA9550-21-1-0424
Grant ID: FA9550-23-1-0736
Grant ID: FA9550-20-1-0386