AMPA Receptor-Mediated Ca2+ Transients in Mouse Olfactory Ensheathing Cells

Ca2+ signaling in glial cells is primarily triggered by metabotropic pathways and the subsequent Ca2+ release from internal Ca2+ stores. However, there is upcoming evidence that various ion channels might also initiate Ca2+ rises in glial cells by Ca2+ influx. We investigated AMPA receptor-mediated inward currents and Ca2+ transients in olfactory ensheathing cells (OECs), a specialized glial cell population in the olfactory bulb (OB), using whole-cell voltage-clamp recordings and confocal Ca2+ imaging. By immunohistochemistry we showed immunoreactivity to the AMPA receptor subunits GluA1, GluA2 and GluA4 in OECs, suggesting the presence of AMPA receptors in OECs. Kainate-induced inward currents were mediated exclusively by AMPA receptors, as they were sensitive to the specific AMPA receptor antagonist, GYKI53655. Moreover, kainate-induced inward currents were reduced by the selective Ca2+-permeable AMPA receptor inhibitor, NASPM, suggesting the presence of functional Ca2+-permeable AMPA receptors in OECs. Additionally, kainate application evoked Ca2+ transients in OECs which were abolished in the absence of extracellular Ca2+, indicating that Ca2+ influx via Ca2+-permeable AMPA receptors contribute to kainate-induced Ca2+ transients. However, kainate-induced Ca2+ transients were partly reduced upon Ca2+ store depletion, leading to the conclusion that Ca2+ influx via AMPA receptor channels is essential to trigger Ca2+ transients in OECs, whereas Ca2+ release from internal stores contributes in part to the kainate-evoked Ca2+ response. Endogenous glutamate release by OSN axons initiated Ca2+ transients in OECs, equally mediated by metabotropic receptors (glutamatergic and purinergic) and AMPA receptors, suggesting a prominent role for AMPA receptor mediated Ca2+ signaling in axon-OEC communication.


INTRODUCTION
Ca 2+ signaling in glial cells is involved in various intercellular processes such as the release of gliotransmitters, modulation of synaptic transmission, long-range Ca 2+ wave propagation, and neurovascular coupling (Haydon, 2001;Carmignoto and Gomez-Gonzalo, 2010). Moreover, intracellular processes such as apoptosis, transcription as well as posttranslational modification are regulated by Ca 2+ signaling (McConkey and Orrenius, 1997;Flavell and Greenberg, 2008). In glial cells, rises of intracellular Ca 2+ are predominantly triggered by Ca 2+ release from internal Ca 2+ stores induced by metabotropic pathways (Deitmer et al., 1998;Verkhratsky et al., 1998), while Ca 2+ influx was reported in rare cases in specialized astrocytelike glial cells such as cerebellar Bergmann glia and retinal Müller glia (Burnashev et al., 1992;Muller et al., 1992;Wakakura and Yamamoto, 1994). However, more recent studies suggest that astroglial cells of different regions of the brain also express ligandgated ion channels that may trigger Ca 2+ responses via Ca 2+ influx from the extracellular space (Schipke et al., 2001;Lalo et al., 2006;Mishra et al., 2016;Droste et al., 2017). In addition, storeoperated channels as well as transient receptor potential (TRP) channels contribute to Ca 2+ signaling in astrocytes (Singaravelu et al., 2006;Rungta et al., 2016;Belkacemi et al., 2017;Rakers et al., 2017;Toth et al., 2019). These studies indicate that Ca 2+ influx might play a previously underestimated role in glial cell physiology and function. Olfactory ensheathing cells (OECs) represent a specialized population of glial cells, exclusively located in the olfactory nerve layer (ONL) in the olfactory bulb (OB) and the peripheral olfactory mucosa (Su and He, 2010;Lohr et al., 2014). They support growth and guidance of axons of olfactory sensory neurons (OSN) from the olfactory epithelium (OE) into the main OB (Graziadei and Graziadei, 1979;Doucette, 1984;Yang et al., 2015). It is assumed, that OECs generate an environment allowing the lifelong re-growth and reintegration of axons into the cellular network after degeneration by the expression of different neurite growth factors and cell adhesions molecules (Crandall et al., 2000;Woodhall et al., 2001;Cao et al., 2007;Yang et al., 2015). In addition, studies showed that Ca 2+ signaling in OECs is a critical regulator for neurite outgrowth in OEC/retinal ganglion cell (RGC) cocultures, suggesting a prominent role for Ca 2+ signaling in OEC-axon interaction (Hayat et al., 2003). Extrasynaptic release of glutamate and ATP by OSN axons initiates Gq-mediated Ca 2+ release from internal stores in OECs via mGluR1 and P2Y1 receptors (Rieger et al., 2007;Thyssen et al., 2010). Montague and Greer (1999) investigated the distribution of the ionotropic AMPA receptor subunits, GluA1, GluA2/3 and GluA4 in the ONL and found expression by nerve-associated glial cells. However, whether OECs exhibit functional AMPA receptors, which might conduct membrane currents and mediate Ca 2+ influx is not known. We performed whole-cell voltage-clamp recordings, showing that kainate application induced AMPA receptor-mediated inward currents in OECs, which are partly mediated by Ca 2+ -permeable AMPA receptors. Kainate also induced Ca 2+ transients that depended on Ca 2+ influx via the receptor channel itself, but additionally comprised Ca 2+ release from internal stores. Electrical stimulation of OSN axons evoked Ca 2+ transients mediated by metabotropic receptors as well as Ca 2+ -permeable AMPA receptors, suggesting a role for AMPA receptor-mediated Ca 2+ signaling in axon-OEC communication.

Electrophysiology and Analysis
For electrophysiological experiments, OB slices (220 µm, horizontal) of GLAST-Cre ERT2 × tdTomato fl/fl mice were prepared using a vibratome (Leica VT1200S, Nußloch, Germany). Slices were transferred into a recording chamber and continuously superfused with ACSF via the perfusion system. The experiments were performed at room temperature. Whole-cell voltage-clamp recordings (Multiclamp 700B, Molecular Devices) were performed on OECs, identified by tdTomato expression (excited at 543 nm, emission 553-618 nm), and the distinct localization in the ONL (Au et al., 2002). Recordings were digitized (Digidata 1440A, Molecular Devices) at 10-20 kHz and filtered (Bessel filter 2 kHz). The holding potential was −80 mV, achieved by current injection of −35.14 ± 13.5 pA on average. Patch pipettes had a resistance of 4-7 M when filled with internal solution containing (mM): 105 K-gluconate, 30 KCl, 10 HEPES, 10 phosphocreatine, 4 Mg-ATP, 0.3 Na-GTP, 0.3 EGTA, and pH = 7.43. For visualization of the recorded OEC, ATTO-488 carboxy (20 µM, ATTO-Tech GmbH, Siegen, Germany) was added to the internal solution before the experiments. Agonists were applied via the perfusion system for 30 s. Antagonists were applied via the perfusion system 10 min prior to agonist application. The flow rate of the perfusion system amounted to approximately 2 mL/min. The experimental bath was oval in shape and had a diameter of 2.1 cm, the bath volume amounted to 1 mL. Therefore, agonist and antagonist concentrations continuously increased in the experimental bath, until the final concentration is reached. Vice versa, agonists were slowly washed out, accounting for long lasting effects. Kainateinduced inward currents in OECs were analyzed by measuring amplitude the currents.

Ca 2+ Imaging and Analysis
For Ca 2+ imaging experiments, GLAST-Cre ERT2 × tdTomato fl/fl and NMRI mice were used. Whole OBs were glued onto coverslips, transferred into a recording chamber and the coverslip secured with a U-shaped platinum wire. For multi-cell bolus loading (Stosiek et al., 2003), a glass pipette with a resistance of ∼3.5 M (when filled with ACSF) was filled with 200 µM Fluo-8 AM (Thermo Fisher Scientific, Darmstadt, Germany) in ACSF, made from a 4 mM stock solution (dissolved in DMSO and 20% pluronic acid). After inserting the injection pipette into the ONL, the Ca 2+ indicator was pressure-injected with 0.7 bar for 30 s into the tissue (PDES-01 AM, NPI electronic GmbH, Tamm, Germany), followed by an incubation of 20 min. Changes in the cytosolic Ca 2+ concentration in OECs were detected by the fluorescence of Fluo-8 (excitation: 488 nm; emission: 500-530 nm) using a confocal microscope (eC1, Nikon, Düsseldorf, Germany). Images were acquired at a time rate of one frame every 3 s. A similar perfusion system and flow rate was used as described above. However, the experimental bath for Ca 2+ imaging experiments was differently shaped (diameter 1 cm, volume 1.9 mL) resulting in differences in the kinetics of washin and wash-out of drugs between the electrophysiological and Ca 2+ imaging experiments. To analyze changes in cytosolic Ca 2+ in single cell somata, regions of interest (ROIs) were defined using Nikon EZ-C1 3.90 software. OECs were identified by GLAST promoter-driven tdTomato expression in GLAST-Cre ERT2 × tdTomato fl/fl mice in most experiments. In wild type mice, OECs were identified by their distinct localization in the ONL. The changes in Ca 2+ were recorded throughout the experiments as relative changes in Fluo-8 fluorescence ( F) with respect to the resting fluorescence, which was normalized to 100%. Quantification of the Ca 2+ transients was achieved by calculating the amplitude of F.

Statistics
All values are stated as mean values ± standard error of the mean. The number of experiments is given as n = x, where x is the number of analyzed cells. At least 3 animals were analyzed for each set of experiments. Statistical significance was estimated by comparing three means using Friedmann ANOVA and the Wilcoxon post hoc test for paired data, and for comparing two means using the Mann-Whitney U-Test, with the error probability p ( * p ≤ 0.05; * * p ≤ 0.01; * * * p ≤ 0.001).

Immunohistochemistry
Immunohistochemistry on OBs of PLP-Cre ERT2 × tdTomato fl/fl mice (≥P28) was performed as described before (Droste et al., 2017). After dissection, the OBs were kept for 1 h at room temperature (RT) in formaldehyde (4% in PBS, pH 7.4). Afterward, 100 µm thick sagittal slices were prepared with a vibratome (VT1000S, Leica, Nußloch, Germany) and incubated for 1 h in blocking solution (10% normal goat serum (NGS), 0.5% Triton X-100 in PBS) at RT. Subsequently, the slices were incubated for 48 h at 4 • C with the following primary antibodies: Guinea pig anti-GluA1 (Alomone labs; 1:200); rabbit anti-GluA2 (Millipore, 1:200); rabbit anti-GluA4 (Millipore; 1:200). To validate the specificity of the GluA antibodies we used cerebellar slices as control, since the distribution of GluA subunits is well documented in this brain area (Supplementary Figure S1). Moreover, the antibodies against GluA1 and GluA4 have been validated in glia-specific GluA1 and GluA4 double knockout mice before (Saab et al., 2012). In our control experiments, the GluA2 antibody only labeled cells known to express GluA2, but not cells that lack GluA2 such as Bergmann glial cells (Burnashev et al., 1992;Muller et al., 1992;Saab et al., 2012), as shown before for the used antibody (Droste et al., 2017). Hence, we consider the used antibodies as efficient and specific. The antibodies were diluted in 1% NGS, 0.05% TritonX-100 in PBS. Slices were incubated in PBS with the following secondary antibodies for 24 h at 4 • C: goat anti-rabbit Alexa Fluor 488 (Thermo Fisher Scientific; 1:1000) or goat antiguinea pig Alexa Fluor 488 (Thermo Fisher Scientific; 1:1000). Moreover, Hoechst 33342 (5 µM; Molecular Probes, Eugene, OR, United States) was added to stain nuclei. Slices were mounted on slides using self-hardening embedding medium (Immu-Mount, Thermo Fisher Scientific). Immunohistological stainings were analyzed using a confocal microscope (Nikon eC1). Confocal images were adjusted for brightness and contrast using ImageJ and Adobe Photoshop CS6.

Distribution of GluA Subunits in the ONL
AMPA receptors constitute of four subunits (GluA1-GluA4), which form complex heteromeric cation channels, activated by glutamate or selective receptor agonists such as AMPA and kainate (Steinhauser and Gallo, 1996). The composition of the subunits is especially critical to determine the Ca 2+ permeability of the channel. Thus, GluA2-lacking AMPA receptors show a high permeability for Ca 2+ whereas GluA2-containing AMPA receptors are impermeable for Ca 2+ (Hollmann et al., 1991). The OB is divided into several clearly distinguishable layers (Figures 1A,B) and immunoreactivity against GluA1, GluA2/3, and GluA4 demonstrated a wide distribution of these subunits in all layers, while mRNA expression analyses showed low amounts of GluA3 in the OB (Montague and Greer, 1999;Horning et al., 2004). We aimed to analyze the distribution of GluA1, GluA2, and GluA4 specifically in the ONL by immunohistochemistry using PLP-Cre ERT2 × tdTomato fl/fl mice. The plp gene encodes for the PLP, expressed by oligodendrocytes, and its slice variant DM20, expressed by oligodendrocytes and OECs, enabling the specific identification of oligodendrocytes and OECs by Cre-recombinase-driven tdTomato expression (Griffiths et al., 1995;Dickinson et al., 1997). OECs are present exclusively in the ONL, whereas oligodendrocytes are located in the glomerular layer (GL) and deeper layers but not in the ONL (Figures 1C,D; Doucette, 1990Doucette, , 1991. Therefore, PLP-Cre ERT2 -driven tdTomato expression in the ONL is assumed to be restricted to OECs (Figures 1C,D; Beiersdorfer et al., 2019;Piantanida et al., 2019). GluA1 immunoreactivity was widely distributed in the entire ONL, colocalizing intensely with tdTomato-expressing OECs (Figures 1E,F). Homogenous GluA2 immunoreactivity was detected in the ONL, with scattered colocalization with tdTomato-positive OECs (Figures 1G,H). In the GL, GluA2-positive juxtaglomerular cells were present (Figure 1G, arrow; Droste et al., 2017). Additionally, GluA4 immunoreactivity was found in the ONL, however, it sparsely colocalized with tdTomato-positive OEC somata (Figures 1I,J). At higher magnification, GluA4 immunoreactivity was found in close approximation of tdTomato-expressing OECs, likely indicating GluA4 located in cell processes of OECs ( Figure 1J, see arrowheads). The results suggest that OECs express GluA1 and GluA2, while GluA4 appears not to be present in OEC somata but presumably in cell processes.

Kainate Induces AMPA Receptor-Mediated Inward Currents in OECs
Functional AMPA/kainate receptors have been investigated in macroglial cells, namely astrocytes and oligodendrocyte precursor cells in gray and white matter as well as microglial cells (Burnashev et al., 1992;Muller et al., 1992;Jabs et al., 1994;Seifert and Steinhauser, 1995;Noda et al., 2000;Zonouzi et al., 2011;Hoft et al., 2014;Droste et al., 2017). Based on our immunohistochemical data, we suggested that OECs in the ONL might also express functional AMPA receptors. Therefore, we performed electrophysiological recordings in OECs in acute OB slices of GLAST-Cre ERT2 × tdTomato fl/fl mice. OECs were identified by GLAST promoter-driven tdTomato expression in the ONL (Figure 2A). Additionally, a voltage-step protocol was applied to characterize the I/V relationship of the recorded cell. All OECs analyzed showed a characteristic linear I/V relationship (Figures 2B,C; Rela et al., 2010;Beiersdorfer et al., 2019). We used kainate to agonize AMPA/kainate receptors, since kainate induced a non-desensitizing AMPA receptor-mediated inward current in acutely isolated hippocampal glial cells (Seifert and Steinhauser, 1995). Kainate application (100 µM, 30 s) induced a prominent slowly rising and long lasting inward current with a mean amplitude of 231.5 ± 58.9 pA (at −80 mV holding potential) in OECs (Figure 2D). To specifically isolate kainateinduced responses in OECs, all experiments were performed in the presence of the gap junction inhibitor carbenoxolone (CBX, 100 µM) to avoid intercellular panglial communication to juxtaglomerular astrocytes that respond to kainate and might interfere with recordings in OECs (Droste et al., 2017;Beiersdorfer et al., 2019). Additionally, neuronal activity was suppressed by inhibiting voltage-gated sodium channels by tetrodotoxin (TTX, 1 µM). Application of GYKI53655 (100 µM), a selective AMPA receptor antagonist, significantly reduced kainate-induced inward currents in OECs by 88.8 ± 3.2% of the control (n = 5; p = 0.022), suggesting that the kainate-induced response is mainly mediated by AMPA receptors in OECs (Figures 2D,E). To further characterize the AMPA receptor subtype involved in kainate-induced inward currents, GluA2lacking Ca 2+ -permeable AMPA receptors were inhibited by the selective antagonist, NASPM (100 µM) (Koike et al., 1997). NASPM reduced kainate-mediated inward currents in OECs significantly by 63.9 ± 2.7% of the control (n = 6; p = 0.019), suggesting that OECs exhibit functional Ca 2+ -permeable AMPA receptors (Figures 2F,G).
To estimate the impact of intracellular Ca 2+ stores on kainateinduced Ca 2+ transients in OECs, we applied cyclopiazonic acid (CPA, 20 µM) to inhibit Ca 2+ uptake into Ca 2+ stores via SERCA (sarcoplasmic/endoplasmic reticulum calcium ATPase). Kainate-induced Ca 2+ transients were significantly reduced by 57.6 ± 4.2% upon Ca 2+ store depletion (n = 70; p = 8.18 × 10 −8 ) (Figures 3G,H). Our results suggest that Ca 2+ influx via AMPA receptor channels is essential to trigger Ca 2+ transients in OECs, whereas Ca 2+ release from internal stores is a secondary effect as a result of the kainate-evoked Ca 2+ influx. To verify that the AMPA-evoked membrane current and associated Ca 2+ influx was not elicited by a preceding Ca 2+ increase, we evoked a Ca 2+ transient by application of ADP and recorded membrane currents. ADP induced a large Ca 2+ increase but failed to elicit membrane currents, demonstrating that Ca 2+dependent membrane conductances are negligible in OECs (Supplementary Figure S2).

Endogenous Glutamate Release Induces AMPA Receptor-Mediated Ca 2+ Transients in OECs
Electrical stimulation of OSN axons evokes ectopic release of glutamate and ATP, which initiates mGluR1 and P2Y1 receptor-mediated Ca 2+ rises in OECs (Berkowicz et al., 1994;Rieger et al., 2007;Thyssen et al., 2010). Here, we showed that OECs express functional Ca 2+ -permeable AMPA receptors which may also contribute to Ca 2+ transients in response to neuronal activity. We endogenously released glutamate upon electrical stimulation of OSN axons (20 Hz, 2 s) and recorded Ca 2+ transients in Fluo-8-loaded OECs in wild type animals. Electrical stimulation of OSN axons induced Ca 2+ transients in OECs with an amplitude of 51.8 ± 5.0% F (n = 46). By eliminating purinergic (MRS2179), dopaminergic (SCH23390) and mGluR1-mediated Ca 2+ transients (CPCCOEt) as well as possible indirect NMDA-mediated neuronal responses (D-APV), the mean amplitude of stimulation-induced Ca 2+ transients in OECs were significantly reduced to 61.0 ± 9.3% of the control (p = 3.68 × 10 −4 ). Additional inhibition of Ca 2+ -permeable AMPA receptors by the selective antagonist NASPM further reduced the mean amplitude to 30.6 ± 5.8 (p = 1.36 × 10 −5 ) of the control. NBQX almost completely abolished stimulationinduced Ca 2+ transients in OECs and the amplitude further decreased significantly to 13.01 ± 3.3% of the control (p = 0.002). Receptor inhibition was reversible and stimulationinduced Ca 2+ transients in OECs recovered after wash out (Figure 4). The results indicate that endogenous glutamate release triggers AMPA receptor-mediated Ca 2+ responses in OECs, suggesting a possible role for AMPA receptors in axon-OEC communication.

DISCUSSION
In the present study, we investigated the role of AMPA receptors in OEC physiology and axon-OEC communication. Kainate induced inward currents and Ca 2+ transients in OECs that were partially mediated by the GluA2-lacking Ca 2+ -permeable AMPA receptor subtype. Endogenous glutamate release initiated AMPA receptor-mediated Ca 2+ transients in OECs, indicating the relevance of AMPA receptors for axon-OEC communication.
We showed clear GluA1 and GluA2 immunoreactivity in the ONL colocalized with tdTomato-expressing OECs, while GluA4 immunoreactivity rarely colocalized with tdTomato in OECs. However, in the present study tdTomato was predominantly localized in OEC somata, limiting the detection of OEC processes, in which GluA4 might be enriched, as it has been demonstrated in a study by Montague and Greer (1999), showing strong GluA4 immunoreactivity in cell processes of presumed olfactory nerve-associated glial cells. Therefore, the results of our and other studies suggest the presence of at least three of four GluA subunits in OECs.
While the presence of GluA2 in OECs argues against Ca 2+ permeability of the AMPA receptors, our physiological data implies the involvement of Ca 2+ influx via Ca 2+ -permeable AMPA receptors. AMPA receptor-dependent Ca 2+ influx could either be conducted by Ca 2+ -permeable AMPA receptors or by voltage-gated Ca 2+ channels activated by AMPA receptormediated depolarization (Porter and McCarthy, 1995). However, OECs do not generate voltage-dependent Ca 2+ influx, e.g., via voltage-gated Ca 2+ channels, ruling out that a depolarization by AMPA receptors is causative for the kainate-evoked Ca 2+ signals (Thyssen et al., 2010). In addition, both kainate-evoked membrane currents and Ca 2+ transients were largely reduced by NASPM, a selective blocker of Ca 2+ -permeable AMPA receptors (Koike et al., 1997). Hence, taking into account that on the one hand side OECs show immunoreactivity to the GluA2 subunit and on the other hand side the current and Ca 2+ responses are NASPM-sensitive, the results suggest that OECs express both Ca 2+ -permeable as well as Ca 2+ -impermeable AMPA receptors. This is confirmed by the NASPM-insensitive but GYKI-sensitive fraction of kainate-evoked inward currents. In addition to Ca 2+ influx via Ca 2+ -permeable AMPA receptors, Ca 2+ release from internal stores contributes to AMPA receptormediated Ca 2+ transients. OECs generate prominent Ca 2+induced Ca 2+ release (CICR) and even a small Ca 2+ increase generates a fast and large Ca 2+ transient by CICR (Stavermann et al., 2012). Thus, Ca 2+ influx through AMPA receptors upon kainate application could induce CICR that boosts the Ca 2+ increase and accounts for the fast rising kinetics of the Ca 2+ signal as compared to the rising phase of the kainate-evoked membrane current. We conclude that Ca 2+ influx is the initial Ca 2+ signal that triggers subsequent internal Ca 2+ release such as CICR, as shown by the entire suppression of the Ca 2+ signal by external Ca 2+ withdrawal but only partial reduction upon Ca 2+ store depletion.
Frontiers in Cellular Neuroscience | www.frontiersin.org concentration has been reported to evoke release of so-called gliotransmitters that affect nearby neurons and synapses, but also release of vasoactive substances such as prostaglandins and arachidonic acid (Bazargani and Attwell, 2016;Guerra-Gomes et al., 2017). OECs enwrap blood vessels in the ONL (Herrera et al., 2005) and respond to neurotransmitters extrasynaptically released from axons of OSN with Ca 2+ signals (Rieger et al., 2007;Thyssen et al., 2010;this study). Increases in intracellular Ca 2+ in OECs results in constriction of adjacent capillaries (Thyssen et al., 2010). Hence, Ca 2+ signals evoked by glutamate release from axons in the ONL and subsequent activation of Ca 2+ -permeable AMPA receptors may lead to vasoresponses that adjust blood flow to the metabolic demand upon action potential firing, a mechanism termed neurovascular coupling.
In summary, our results show that release of glutamate from axons of OSN results in Ca 2+ transients in OECs that are partially mediated by AMPA receptors. AMPA receptor-mediated Ca 2+ transients were due to Ca 2+ influx, most likely through Ca 2+permeable AMPA receptors themselves, and subsequent Ca 2+ release from internal Ca 2+ stores.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding authors.

AUTHOR CONTRIBUTIONS
AB and CL designed the experiments and wrote the manuscript. AB performed the experiments and analyzed the data.

FUNDING
This work was supported by the Deutsche Forschungsgemeinschaft (LO779/10 and SFB 1328 TP-A07).

ACKNOWLEDGMENTS
We thank A. C. Rakete and M. Fink for technical assistance. We also thank A. Scheller and F. Kirchhoff for providing transgenic animals.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fncel. 2019.00451/full#supplementary-material FIGURE S1 | Control stainings of the cerebellum to validate the specificity of anti-GluA antibodies. GluA1 was detected in Purkinje neurons (arrowhead) and Bergmann glial cells (arrow), GluA2 in Purkinje neurons and GluA4 in Bergmann glial cells, in line with published data. Scale bars: 50 µm.