Neuronal input triggers Ca2+ influx through AMPA receptors and voltage-gated Ca2+ channels in oligodendrocytes.

Communication between neurons and developing oligodendrocytes (OLs) leading to OL Ca2+ rise is critical for axon myelination and OL development. Here, we investigate signaling factors and sources of Ca2+ rise in OLs in the mouse brainstem. Glutamate puff or axon fiber stimulation induces a Ca2+ rise in pre-myelinating OLs, which is primarily mediated by Ca2+ -permeable AMPA receptors. During glutamate application, inward currents via AMPA receptors and elevated extracellular K+ caused by increased neuronal activity collectively lead to OL depolarization, triggering Ca2+ influx via P/Q- and L-type voltage-gated Ca2+ (Cav ) channels. Thus, glutamate is a key signaling factor in dynamic communication between neurons and OLs that triggers Ca2+ transients via AMPARs and Cav channels in developing OLs. The results provide a mechanism for OL Ca2+ dynamics in response to neuronal input, which has implications for OL development and myelination.

INTRODUCTION

Oligodendrocytes (OLs) ensheath axons with their myelinating processes, providing protection and maintenance of axonal integrity and enabling saltatory conduction of action potentials. Dynamic interactions between neurons and OLs modulate development of OL lineage cells and myelination to control conduction speed in the central nervous system (CNS; Pajevic, Basser, & Fields, 2014; Sinclair et al., 2017). Recent studies demonstrated that a dynamic Ca2+ rise in OLs is associated with myelin sheath refinement in vivo in the zebrafish spinal cord (Baraban, Koudelka, & Lyons, 2018; Krasnow, Ford, Valdivia, Wilson, & Attwell, 2018). Ca2+ is an important second messenger that induces cellular processes that can lead to local translation and process extension in OL lineage cells (Baraban et al., 2018; Krasnow et al., 2018; Wake, Lee, & Fields, 2011). However, little is known about signaling factors from active axons that facilitate activity‐dependent neuron‐OL communication leading to Ca2+ transients in OLs, as well as the sources of OL Ca2+ rise in response to neuronal inputs.

Previous studies identified neuronal glutamate or GABA as potential mediators of neuron‐OL interaction (Bergles, Roberts, Somogyi, & Jahr, 2000; Kukley, Capetillo‐Zarate, & Dietrich, 2007; Lin & Bergles, 2004). In particular, OL lineage cells express Ca2+‐permeable or Ca2+‐impermeable AMPA receptors (AMPARs) and NMDA receptors, and thus can directly sense the glutamate release from neurons (Berret et al., 2017; Ge et al., 2006; Micu et al., 2016; Patneau, Wright, Winters, Mayer, & Gallo, 1994). Recent studies showed that inhibiting/knocking out neurotransmitter receptors in OLs impairs their response to neuronal activity and leads to alterations in OL development and function (Koudelka et al., 2016; Kougioumtzidou et al., 2017; Mensch et al., 2015), suggesting that glutamate can be an important signal for developing OLs to be able to sense the electrical activity of active axons.

Voltage‐gated Ca2+ (Cav) channels are potential indirect effectors of glutamatergic input, because activation of AMPA receptors is able to depolarize and activate voltage‐gated ion channels in OL lineage cells (Berret et al., 2017; Káradóttir, Hamilton, Bakiri, & Attwell, 2008). Previous studies have reported the expression of L‐, N‐, and R‐type Cav channels and the presence of their currents in OL lineage cells (Berger, Schnitzer, Orkand, & Kettenmann, 1992; Butt, 2006). In NG2+ glial cells, integration of synaptic input resulted in Ca2+ signals by recruiting low voltage‐activated R‐type or T‐type Cav channels (Sun, Matthews, Nicolas, Schoch, & Dietrich, 2016). In addition, high voltage‐activated L‐type Cav channels are known to modulate OL development and migration (Cheli et al., 2016; Cheli, Santiago González, Spreuer, & Paez, 2015; Paez, Fulton, Colwell, & Campagnoni, 2009). Despite functional implications of Cav channels in OL lineage cells, there is a lack of evidence for the functional link between glutamatergic inputs and Ca2+ influx through Cav channels in developing OLs beyond the precursor stage.

Here, we investigate the mechanism by which glutamate induces a Ca2+ rise in OLs in the medial nucleus of the trapezoid body (MNTB) of the auditory brainstem. The heavily myelinated axon fibers in the MNTB, which are important for ensuring the reliability and temporal fidelity of conduction, undergo activity‐dependent myelination during development (Kim, Renden, & von Gersdorff, 2013; Kim, Turkington, Kushmerick, & Kim, 2013; Sinclair et al., 2017). In the MNTB, we find glutamatergic input‐induced Ca2+ transients and Ca2+ currents via Cav channels in developing OLs using electrophysiology and Ca2+ imaging in mice expressing OL‐specific GCaMP6f, a genetically encoded Ca2+ indicator with fast kinetics (Chen et al., 2013). We identify AMPARs and L‐type and P/Q‐type Cav channels as crucial sources of Ca2+ influx in pre‐myelinating OLs in response to glutamatergic neuronal input, coordinating active communication between neurons and OLs.

MATERIALS AND METHODS

Animals

Experiments using transgenic mice were approved and performed in accordance with the guidelines and protocols of the University of Texas Health Science Center at San Antonio (UTHSCSA) Institutional Animal Care and Use Committee (IACUC, approved protocol #14005×). Heterozygous CNPase‐Cre mice, generated by Dr. Klaus Nave (Max Planck Institute of Experimental Medicine, Germany) and obtained from Dr. Manzoor Bhat (UTHSCSA), were bred with homozygote Cre‐dependent GCAMP6f mice (Ai95D, JAX #028865) obtained from Dr. Martin Paukert (UTHSCSA). Heterozygote CNPase‐Cre;GCaMP6f offspring of either sex between the ages of postnatal Day 8 (P8) and P12 were used for experiments. Genotyping was accomplished using PCR.

Slice preparation

Transverse brainstem slices (200 μm thick) were prepared from P8–P12 mice. After rapid decapitation, the brainstem was removed from the skull and immersed in ice‐cold artificial cerebrospinal fluid (aCSF) containing (in mM): 125 NaCl, 2.5 KCl, 3 MgCl2, 0.1 CaCl2, 25 glucose, 25 NaHCO3, 1.25 NaH2PO4, 0.4 ascorbic acid, 3 myo‐inositol, and 2 Na‐pyruvate, pH 7.3–7.4, when bubbled with carbogen (95% O2, 5% CO2; osmolality of 310–320 mOsm/kg water). After cutting, slices were incubated in recovery aCSF bubbled with carbogen at 35°C for 30 min, and thereafter at room temperature.

Immunohistochemistry

Brainstem slices (100 μm thick) were cut and subsequently fixed with 4% (wt/vol) paraformaldehyde in phosphate buffer solution (PBS) for 20 min. Free‐floating sections were blocked in 4% goat serum and 0.3% Triton X‐100 in PBS for 1 hr. Slices were incubated with antibodies for chicken anti‐green fluorescent protein (GFP; 1:300; Abcam, Cambridge, UK, Cat# ab13970, RRID:AB_300798), rabbit anti‐oligodendrocyte transcription factor 2 (Olig2; 1:200; Abcam, Cambridge, UK, Cat# ab109186, RRID:AB_10861310), mouse anti‐neuronal nuclei (NeuN; 1:600; Millipore, Burlington, MA, USA, Cat# MAB377, RRID:AB_2298772) overnight at 4°C and subsequently incubated with different Alexa dye–conjugated secondary antibodies (1:500, Invitrogen, Carlsbad, CA, USA) for 2 hr at room temperature. Slices were mounted onto Superfrost slides in photobleaching‐protective medium (Fluoroshield; Sigma‐Aldrich, St. Louis, MO, USA). Stained slices were viewed with laser lines at 488, 555, and 633 nm using a 40× oil‐immersion objective on a confocal laser‐scanning microscope (LSM‐510, Zeiss, Oberkochen, Germany). Stack images were acquired at a digital size of 1,024 × 1,024 pixels with optical section separation (z interval) of 0.5 μm and were later cropped to the relevant part of the field without changing the resolution.

Ca2+ imaging

Ca2+ imaging was performed in normal aCSF at room temperature (22–24°C). Normal aCSF was the same as slicing aCSF, but with 1 mM MgCl2 and 2 mM CaCl2. Fluorescence images were acquired at 20 Hz sampling frequency using an ORCA Flash4.0 CMOS camera (Hamamatsu Photonics, Hamamatsu, Japan) and HCImage software (Hamamatsu Photonics, Hamamatsu, Japan), using wide‐field fluorescence microscope illumination (X‐Cite 120Q excitation light source with 120‐watt lamp). 1 mM glutamate (Sigma, St. Louis, MO, USA) or 40 mM KCl (Sigma, St. Louis, MO, USA) was puff‐applied through a glass pipette (5–7 psi for ~1 s), the tip of which was placed 50–100 μm away from the cell body. GCaMP6f‐GFP or Fluo‐4 (50 μM, Life Technologies, Carlsbad, CA, USA) fluorescence was measured when excited at 488 nm. Background subtractions were done for cells filled with Fluo‐4. Fluorescent regions of interest (ROIs) were defined manually. All values are expressed as a change in the fluorescence calculated as ΔF/F, where F is the basal fluorescence of the ROI and ΔF is the change in fluorescence of the ROI at the peak response relative to the F value. 6‐cyano‐7‐nitroquinoxaline‐2,3‐dione (CNQX; 50 μM; Tocris, Bristol, UK), 1‐naphthyl acetyl spermine (Naspm; 50 μM; Tocris, Bristol, UK), D‐AP5 (50 μM; Tocris, Bristol, UK), dantrolene (20 μM; Tocris, Bristol, UK), CdCl2 (200 μM; Sigma St. Louis, MO, USA), 4‐aminopyridine (4‐AP; 2 mM; Sigma, St. Louis, MO, USA), tertraethylammonium (TEA; 10 mM; Sigma, St. Louis, MO, USA), BaCl2 (1 mM; Sigma, St. Louis, MO, USA), tetrodotoxin (TTX; 1 μM; Sigma, St. Louis, MO, USA), nifedipine (100 μM; Sigma, St. Louis, MO, USA), and ω‐agatoxin‐IVA (agatoxin; 200 nM; Alomone, Jerusalem, Israel) were bath‐applied.

Electrophysiology

Whole‐cell patch‐clamp recordings were performed in normal aCSF at room temperature (22–24°C) using the voltage‐ or current‐clamp mode of the EPC‐10 (HEKA Electronik, Lambrecht/Pfalz, Germany). For Ca2+ current recordings, CsCl‐based pipette solution included (in mM): 150 CsCl, 10 TEA‐Cl, 1 MgCl2, 10 HEPES, 2 ATP, 0.3 GTP, 10 phosphocreatine, and 50 μM Fluo‐4, pH = 7.3. Ca2+ currents were analyzed after leak subtraction using a “traditional” p/4 stimulus train in EPC10‐Patchmaster. For current clamp and AMPAR current recordings, K‐gluconate‐based pipette solution contained (in mM): 130 K‐gluconate, 20 KCl, 5 Na2‐phosphocreatine, 10 HEPES, 4 Mg‐ATP, 0.3 GTP, and 50 μM Fluo‐4, pH = 7.3. Occasionally, pipette solution additionally contained Alexa 568 (50 μM) to visualize the cell through dye‐filling during whole‐cell recordings. Patch electrodes had resistances of 4–6 MΩ, and the series resistance (R s) was <25 MΩ before compensation.

Data analysis

Ca2+ imaging and electrophysiology data were analyzed and presented using Igor Pro (Wavemetrics, Lake Oswego, OR, USA; RRID: SCR_000325). Statistical significance was determined using paired Student's t test (GraphPad Prism, San Diego, CA, USA; RRID: SCR_002798). p Values < .05 were considered significant. Data values are reported as means ± standard error of the mean (SEM).

RESULTS

Glutamate‐induced Ca2+ rise mediated by AMPA receptors on pre‐myelinating OLs

In the MNTB of the mouse auditory brainstem, GCaMP6f was expressed specifically in OLs, which were Olig2+ and NeuN‐, (Figure 1a) and were distinguishable under DIC and fluorescent imaging at 488 nm (Figure 1b). GCaMP6f + cells did not express PDGFRα, a marker of oligodendrocyte precursor cells (OPCs), but were positive for CC1, indicating that they are beyond the OPC stage (data not shown; Jang, Gould, Xu, Kim, & Kim, 2019). At P8‐P12, there is little compact myelin in the brainstem, and thus the majority of GCaMP6f + cells are presumed to be in the pre‐myelinating stage.

Figure 1

Glutamate induces Ca2+ rise in pre‐myelinating OLs. (a) GCaMP6f‐expressing cells (green, arrows) express OL marker Olig2 (red). Neuronal marker NeuN (blue) is expressed in nearby neurons, but not in GCaMP6f + OLs. (b) Fluorescent (left) and merge of DIC and fluorescent images (right) of a GCaMP6f + OL. (c) The cell responds to consecutive glutamate puffs (red arrow) with consistent Ca2+ rises. Two‐minute time intervals allowed for the recovery of the Ca2+ response to baseline. (d) A GCaMP6f+ cell (green) was dye‐filled with Alexa 568 (red) during whole‐cell patch‐clamp recording to discern its morphology. (e) Voltage‐activated currents in response to depolarization steps (from −80 to +50 mV, holding at −70 mV) in a GCaMP6f+ cell. OLs, oligodendrocytes [Color figure can be viewed at wileyonlinelibrary.com]

GCaMP6f was primarily detectable in the cytosol of OL cell bodies. Transient and local glutamate puff (1 mM) application 50–100 μm from the OL cell body resulted in a Ca2+ rise, as recorded by a transient increase in fluorescence (Figure 1c). The average amplitude of the Ca2+ response was 0.19 ± 0.01 ΔF/F (n = 113) and the decay time from the peak of the Ca2+ response to the resting Ca2+ level was 13.92 ± 0.52 s in 113 individual cells from 19 animals. Glutamate‐induced Ca2+ rise in OLs was replicable and consistent in amplitude after a two‐minute recovery period. Morphologically, the processes of GCaMP6f + cells did not align with axons to generate myelin sheaths (Figure 1d). In whole‐cell voltage‐clamp recordings, these GCaMP6f + cells displayed outwardly rectifying K+ currents and, and 37.5% of cells showed transient inward Na+ currents (Figure 1e). This electrophysiological profile aligns with what has previously been observed in pre‐myelinating OLs in the MNTB (Berret et al., 2017).

OL lineage cells express ionotropic glutamate receptors (e.g., AMPA and NMDA receptors) and receive glutamatergic input from neurons (Berret et al., 2017; Micu et al., 2016). We tested whether AMPA and NMDA receptors are involved in glutamate‐induced Ca2+ transients in OLs using their antagonists. Before drug application, we confirmed the reliability of a Ca2+ response by verifying consistent Ca2+ rises in the OL in response to two consecutive glutamate puffs. The individual Ca2+ responses in the presence of antagonists were normalized to this control response. In the presence of Naspm, a Ca2+‐permeable AMPAR antagonist, the amplitude of glutamate‐induced Ca2+ transient was significantly decreased to 40.8 ± 5.46% of the control response from the same cell (n = 13 cells from four animals; p < .0001, paired t test; Figure 2a,b). Furthermore, when CNQX, an AMPAR antagonist, was added to the bath along with Naspm, glutamate‐induced Ca2+ rise was almost completely inhibited (11.8 ± 2.45% of the control response from the same cell; n = 24 cells from six animals; p < .0001, paired t test). We tested whether NMDA receptor activation contributes to glutamate‐induced Ca2+ rise in OLs. D‐AP5, an NMDA receptor antagonist, had no significant effect on glutamate‐induced Ca2+ rise in OLs (93.8 ± 3.99% of the control response from the same cell; n = 15 cells from three animals; p = .1398, paired t test). Other studies have shown the presence of Ca2+‐induced Ca2+ release (CICR) in OLs (Haberlandt et al., 2011; Ruiz, Matute, & Alberdi, 2010). However, dantrolene to block CICR had no significant effect on Ca2+ rise in OLs in response to glutamate puff (103.4 ± 6.55% of the control response from the same cell; n = 11 cells from three animals; p = .6144, paired t test).

Figure 2

Ca2+ influx through Ca2+‐permeable AMPARs contributes to glutamate‐induced OL Ca2+ rise. (a) Ca2+ response to glutamate puff in a GCamP6f + OL in the presence of 50 μM Naspm and 50 μM CNQX to inhibit AMPARs. (b) Summary of Ca2+ responses in the presence of Naspm, Naspm, and CNQX, 50 μM D‐AP5, or 20 μM dantrolene. (c) Voltage clamp (V‐clamp) recordings of AMPAR‐mediated currents (red trace) in a single OL (at −70 mV) and associated Ca2+ rise (black trace) measured with Fluo‐4 in response to glutamate puff in control and the presence of Naspm and CNQX. (d) Summary of AMPAR‐mediated current amplitude in the presence of Naspm or Naspm and CNQX normalized to the control response. Data are represented as ± SEM. ** represents p < .01, and *** represents p < .001. OL, oligodendrocyte [Color figure can be viewed at wileyonlinelibrary.com]

To examine whether glutamate puff directly activates OL AMPARs to produce AMPAR currents, we recorded AMPAR currents and simultaneously measured Ca2+ rise in OLs. Fluo‐4, a Ca2+ indicator, was administered in the patch pipette during whole‐cell recording. In voltage‐clamp recordings (holding at −70 mV), the amplitude of AMPAR currents was −36.3 ± 6.04 pA in response to glutamate puff, which simultaneously induced a Ca2+ rise (n = 8 cells from seven animals; Figure 2c). Addition of Naspm to block Ca2+‐permeable AMPARs significantly reduced the amplitude of the AMPAR current to 44.5 ± 4.05% of control currents in the same cell (n = 8 cells from seven animals; p < .0001, paired t test; Figure 2d). CNQX application in addition to Naspm further inhibited the AMPAR current (8.8 ± 3.345% of control responses; n = 8 cells from seven animals; p < .0001, paired t test), indicating that glutamate induced AMPAR currents through both Ca2+‐permeable and ‐impermeable AMPARs in OLs. Interestingly, the Ca2+ rise in response to glutamate puff was completely inhibited in the presence of Naspm (1.3 ± 2.30% of control; p < .0001, paired t test). Thus, Ca2+ influx through Ca2+‐permeable AMPARs is mainly responsible for the Ca2+ rise when the membrane voltage was clamped at −70 mV. Notably, during GCaMP6f imaging when the membrane potential was not clamped, there was a significant portion of the Ca2+ rise remaining in the presence of Naspm, which was completely inhibited by CNQX (Figure 2a,b). These results indicate that AMPAR currents through both Ca2+‐permeable and ‐impermeable AMPAR activation contribute to glutamate‐induced Ca2+ response in developing OLs in the MNTB.

Voltage‐gated Ca2+ channels are a source for Ca2+ influx mediated by OL depolarization

How do Ca2+‐impermeable AMPAR currents contribute to Ca2+ responses in OLs? Our previous study showed that glutamatergic input depolarized pre‐myelinating OLs and consequently activated voltage‐gated ion channels (Berret et al., 2017). We examined whether Ca2+ influx via Cav channels could be a source of glutamate‐induced Ca2+ rise in OLs. After two consecutive applications of glutamate to confirm a replicable glutamate‐induced Ca2+ response, Naspm was bath applied to rule out Ca2+‐permeable AMPAR‐induced Ca2+ rise, reducing the amplitude of the Ca2+ response to 55.4 ± 6.70% of the control response from the same cell (n = 16 cells from four animals; p < .0001, paired t test; Figure 3a). The additional application of CdCl2, which blocks Cav channels, reduced the amplitude of the remaining Ca2+ response to 5.7 ± 2.34% of the control response from the same cell (n = 26 cells from four animals; p < .0001, paired t test, Figure 3b). CdCl2 alone reduced the amplitude of the Ca2+ response to 33.4 ± 5.31% of the control response from the same cell (n = 10 cells from three animals; p < .0001, paired t test). This result indicates that glutamate triggers Ca2+ entry through Cav channels in pre‐myelinating OLs.

Figure 3

Voltage‐gated Ca2+ channels activated by depolarization contribute to Ca2+ rises in OLs. (a) Glutamate‐induced Ca2+ responses from a GCaMP6f + OL in the presence of 50 μM Naspm and additional 200 μM CdCl2. A horizontal line indicates baseline Ca2+ level. (b) Summary of Ca2+ response after the application of Naspm, CdCl2, or both. (c) Fluorescent image during whole‐cell recording (top left) of OL filled with Alexa 568, corresponding confocal image after fixation (top right), the fluorescent image of Ca2+ indicator dye Fluo‐4 (50 μM, bottom left), and merge of DIC and fluorescent images (bottom right) during whole‐cell patch‐clamp recording. Circles indicate the ROIs taken for the fluorescence intensity analysis. (d) Current clamp (I‐clamp) recordings of membrane voltage (Vm, red trace) changes in OL in response to glutamate puff in the absence and presence of CdCl2. Glutamate puff depolarized OL to −40 mV from −65 mV. Note, CdCl2 had no effect on glutamate‐induced depolarization. (e, f) Subcellular Ca2+ responses of the same OL at the soma (e) and process (f) to glutamate puff correlated with depolarization in the absence or presence of CdCl2. (g) Summary of Ca2+ rise at soma and process in control and in the presence of CdCl2, which were normalized to control response in the soma. Data are represented as ± SEM. ** represents p < .01, and *** represents p < .001. OL, oligodendrocyte [Color figure can be viewed at wileyonlinelibrary.com]

Next, we examined whether glutamate application is sufficient to depolarize OL and consequently activate Cav channels. To study the correlation between glutamate‐induced depolarization and Ca2+ influx via Cav channel in pre‐myelinating OLs, we recorded membrane potential changes and simultaneously monitored a Ca2+ transient from a single OL (Figure 3c). In whole‐cell recordings of OLs with Ca2+ indicator Fluo‐4, glutamate puff elicited a depolarization from −65 mV to ~ − 40 mV (20.8 ± 2.00 mV change, n = 6; Figure 3d). Glutamate‐induced depolarization resulted in a Ca2+ rise (0.12 ± 0.021 ΔF/F, n = 6 cells from four animals; Figure 3e). By adding Fluo‐4 and Alexa 568 to the internal pipette solution, OL processes within 20 μm from the OL soma, which were not myelin sheaths, were clearly visible and showed detectable glutamate‐induced Ca2+ responses. The Ca2+ rise in the processes was 187.6 ± 22.76% of the response in the soma of the same cell (n = 6 cells from four animals; p = .0085, paired t test; Figure 3f). In the presence of CdCl2, the Ca2+ rise was significantly reduced in both processes and soma (n = 6 cells from four animals; p = .0012 for soma, p = .0018 for process, paired t test; Figure 3g), while glutamate‐induced depolarization was unaffected, indicating that glutamate puff is able to depolarize pre‐myelinating OLs and consequently activate Cav channels.

Increased neuronal activity increases extracellular K+ and glutamate, resulting in OL Ca2+ rise

During increased neuronal activity, K+ accumulates in the extracellular space and can depolarize OLs (Battefeld, Klooster, & Kole, 2016; Larson et al., 2018). We next tested the capacity of OL depolarization to recruit Cav channels allowing Ca2+ influx independent of OL AMPAR activation. KCl puff applied to OLs and surrounding tissue, which mimics elevated extracellular K+ when neuronal activity increases, resulted in a Ca2+ rise in GCaMP6f + OLs (Figure 4a). In the presence of Naspm and CNQX to isolate the OL depolarization‐induced Ca2+ entry, KCl‐induced Ca2+ rise was 70.5 ± 0.06% of control responses (n = 10 cells from three animals; p = .0008; paired t test). This remaining Ca2+ response was significantly reduced by the application of CdCl2 reduced to 24.3 ± 6.30% of responses in the presence of CNQX and Naspm (n = 10 cells from three animals; p < .0001, paired t test; Figure 4b). Together, these data indicate that OL depolarization is sufficient for Ca2+ influx via Cav channels.

Figure 4

OLs depolarize in response to increased extracellular K+ and/or glutamate when neuronal activity increases. (a) Ca2+ responses to KCl puff (40 mM) in GCaMP6f + OL in the presence of Naspm and CNQX, and addition of CdCl2. (b) Summary of OL Ca2+ response to KCl puff in the presence of Naspm and CNQX, and the addition of CdCl2. (c) Ca2+ rise (black) as measured by Fluo‐4 and associated membrane voltage (V m) change (red) of OL in current‐clamp (I‐clamp) recordings in response to glutamate puff in the presence of 1 μM TTX, and addition of Naspm and CNQX. (d, e) Summary of glutamate‐induced Ca2+ responses (d) and depolarization (e) in OLs in the presence of TTX, Naspm, and CNQX. Data are represented as ± SEM. ** represents p < .01, and *** represents p < .001. OLs, oligodendrocytes; TTX, tetrodotoxin [Color figure can be viewed at wileyonlinelibrary.com]

We next determined whether increased neuronal activity and elevated extracellular K+ as a result of glutamate puff induced OL depolarization or whether AMPAR‐mediated currents directly depolarized OLs, or whether both mechanisms contributed. Using TTX to inhibit neuronal activity, we isolated AMPAR activation‐mediated OL depolarization. In whole‐cell current‐clamp recordings with Fluo‐4 in the intracellular pipette solution, glutamate puff induced a Ca2+ rise in the presence of TTX (1 μM), although TTX application resulted in a decrease in the Ca2+ rise in the presence of TTX to 74.7 ± 6.04% of control responses (n = 17 cells from five animals; p = .0007, paired t test; Figure 4c,d). Simultaneously, glutamate puff still depolarized OLs by 14.64 ± 1.13 mV under the inhibition of neuronal activity with TTX, which was a slight decrease from control (depolarization was 21.44 ± 1.88 mV in control; n = 6 from three animals; p = .0021; paired t test; Figure 4c,e). The reduction in the presence of TTX is assumed to be caused by the inhibited neuronal activity, although voltage‐activated Na+ channels in OLs could be also involved in the TTX‐sensitive response (Berret et al., 2017). The addition of Naspm and CNQX blocked the remaining Ca2+ response (4.5 ± 2.26% of control response; n = 17 cells from five animals; p < .0001, paired t test) and depolarization (2.8 ± 0.64 mV; n = 6 cells from three animals; p < .0001, paired t test). These results suggest that elevated extracellular K+ due to increased firing rate of neurons partially contributes to depolarization and Ca2+ rise in OLs, and that the larger remaining response is due to AMPAR‐mediated currents.

OL Ca2+ rise through P/Q‐type and L‐type Cav channels

Furthermore, we identified the types of Cav channels involved in Ca2+ influx in OLs in response to glutamate puff. We specifically examined L‐type and P/Q‐type channels, which are expressed in OL lineage cells (Cheli et al., 2015, 2016; Paez et al., 2009; Zhang et al., 2014). After Ca2+‐permeable AMPARs were blocked with Naspm, nifedipine (100 μM), an L‐type Cav channel blocker, reduced the glutamate‐induced Ca2+ response in GCaMP6f + OLs to 38.0 ± 5.09% (n = 18 cells from four animals; p < .0001, paired t tests), while agatoxin (200 nM), a P/Q‐type Cav channel blocker, reduced the amplitude to 20.6 ± 2.61% of the control response from the same cell (n = 25 cells from four animals; p < .0001, paired t tests; Figure 5a,b). Combined, agatoxin and nifedipine nearly abolished the glutamate‐induced Ca2+ rise, reducing the amplitude to 6.2 ± 1.33% of control responses (n = 10 cells from three animals; p < .0001, paired t test). The combination of agatoxin and CdCl2 reduced the Ca2+‐impermeable AMPAR‐mediated Ca2+ response to 6.4 ± 2.26% of control responses in the same cell (n = 10 cells from three animals; p < .0001, paired t test). Thus, P/Q‐type and L‐type Cav channels play a significant role in OL Ca2+ rise in response to glutamate puff.

Figure 5

P/Q‐ and L‐type Cav channels mediated Ca2+ influx in response to glutamate input. (a) Glutamate‐induced Ca2+ rise in a GCaMP6f + OL in the presence of Naspm, agatoxin (200 nM) and nifedipine (100 μM). (b) Summary of Ca2+ responses after application of nifedipine or agatoxin, and combinations of agatoxin + nifedipine and agatoxin + CdCl2. (c) Cav channel currents in the presence of 4‐AP (2 mM), TEA (10 mM), BaCl2 (1 mM), and TTX (1 μM) in voltage‐clamp recording (holding at −65 mV), which were inhibited by CdCl2. Inset, step‐like depolarization protocol. (d) Current–voltage (I‐V) relationship of OL Ca2+ current (I Ca). I Ca was partially inhibited by agatoxin and completely inhibited by CdCl2. (e) Ca2+ rise in a GCaMP6f + OL when directly depolarized using a depolarizing pulse (diagram) during voltage‐clamp recording in the presence of agatoxin and nifedipine. The OL was brought from resting membrane potential at −65 mV to −40, −20, or 0 mV over the course of 100 ms, held at 0 mV for 100 ms, and brought back to resting membrane potential over a period of 2 s. (f) Summary of depolarization‐induced Ca2+ rise in the presence of agatoxin and nifedipine. Data are represented as ± SEM. *** represents p < .001. OL, oligodendrocyte; TEA, tertraethylammonium; TTX, tetrodotoxin [Color figure can be viewed at wileyonlinelibrary.com]

To confirm the presence of functional P/Q‐type Cav channel currents in OLs, voltage‐activated Ca2+ currents (I Ca) were elicited from OLs in whole‐cell voltage‐clamp recordings. I Ca was shown in response to depolarization steps from −80 mV to 60 mV and was completely inhibited by CdCl2 (Figure 5c). Partial Cav channel inactivation was observed, as expected for P/Q‐type Cav channels. Moreover, a portion of I Ca was inhibited by agatoxin, suggesting that P/Q‐type Cav channels are involved in Ca2+ influx due to depolarization (Figure 5d). The voltage dependency of OL I Ca was determined by plotting the current–voltage (I‐V) relationship curve. The peak of I Ca at 0 mV was −79.86 ± 10.47 pA (n = 14 cells from eight animals). The presence of I Ca and effect of Cav channel antagonists on glutamate‐induced Ca2+ responses demonstrates that Ca2+ influx via Cav channels importantly contributes to Ca2+ rise in pre‐myelinating OLs during glutamate‐induced depolarization.

We next tested the ability of direct OL depolarization to induce OL Ca2+ response mediated by P/Q‐type and L‐type Cav channels using whole‐cell voltage‐clamp. A depolarization pulse, mimicking the time course of the depolarization recorded in response to glutamate puff, resulted in a Ca2+ rise measured with GCaMP6f fluorescence. The Ca2+ rise was detectable in response to ~25 mV depolarization from −65 mV to −40 mV and was partially blocked by agatoxin application (52.3 ± 5.60% of control responses; n = 7 cells from four animals; p < .0001; paired t test; Figure 5e,f). The remaining Ca2+ response was inhibited by the addition of nifedipine (0.37 ± 0.08% of control responses; n = 7 cells from four animals; p < .0001; paired t test; Figure 5e,f), indicating that Ca2+ influx through a combination of P/Q‐type and L‐type Cav channels occurs in response to depolarization of OLs.

Axon stimulation provides sufficient glutamate for OL Ca2+ rise

We next examined whether neuronal activity causes glutamate‐mediated Ca2+ rise in OLs through neuron‐OL interaction. Axon fibers that innervate the MNTB were stimulated at 50 Hz for 10 s, inducing Ca2+ rise from GCaMP6f + OLs (Figure 6a,b). Blocking Ca2+‐permeable AMPARs with Naspm reduced the Ca2+ response to axon fiber stimulation to 58.2 ± 3.36% (n = 20 cells from four animals; p < .0001, paired t tests), and the combination of Naspm and CNQX reduced the response to 28.4 ± 3.16% of control responses in the same cell (n = 20 cells from four animals; p < .0001, paired t tests; Figure 6c,d), indicating that neuronal activity provides sufficient glutamate to induce Ca2+ rise. The Ca2+ response, which was not blocked by Naspm and CNQX, was significantly inhibited by Cav channel blocker CdCl2 (5.1 ± 0.99% of control responses in the same cell, n = 10 cells from two animals; p < .0001, paired t test). During axon fiber stimulation, increased neuronal activity leads to the accumulation of extracellular K+ and consequently OL depolarization (Battefeld et al., 2016; Larson et al., 2018). Indeed, we demonstrated that increasing extracellular K+ concentration using KCl puff depolarized and induced Ca2+ influx through Cav channels in OLs without AMPAR activation (Figure 4a). These data reveal that, when neuronal activity‐induced Ca2+ rise in OLs is mediated by AMPARs and Cav channels as sources of Ca2+ influx.

Figure 6

Axons provide sufficient glutamate to induce OL Ca2+ response. (a) Schematic of the experimental setup. A bipolar electrode was placed ~200 μm away from the MNTB to stimulate axon fibers (grey), and Ca2+ responses were recorded in GCaMP6f + OLs (green). (b) Fluorescent image of GCaMP6f + OL (white circle) before (left) and after (right) axon fiber stimulation at 50 Hz for 10 s. Note, some GCaMP6f + OLs did not respond to axonal stimulation. (c) OL Ca2+ response to axon fiber stimulation (red arrows) in the presence of Naspm, CNQX, and CdCl2. (d) Summary of Ca2+ rise in OLs in response to axon stimulation in the presence of Naspm, CNQX, and CdCl2. Data are represented as ± SEM. *** represents p < .001. MNTB, medial nucleus of the trapezoid body; OL, oligodendrocyte [Color figure can be viewed at wileyonlinelibrary.com]

DISCUSSION

During development, neuronal activity drives physiological responses in OL lineage cells that influence OL maturation and myelination. It is proposed that neuronal activity triggers release of signaling factors to elicit OL Ca2+ transients and initiate downstream signaling processes. Here, we studied what factors mediate the dynamic communication between neurons and OLs leading to OL Ca2+ rise and the primary source of OL Ca2+ transients in response to neuronal input. The results reveal that increased neuronal activity leads to Ca2+ influx in OLs through glutamatergic signaling, mediated by AMPARs, depolarization, and L‐type and P/Q‐type Cav channels. These findings support that glutamatergic signaling is a key mechanism for neuron‐OL interaction that could lead to various downstream Ca2+ signaling pathways important for myelination and OL function in the mammalian nervous system.

Due to the overlap in expression of channels and receptors in both neurons and glia in the brain slice, it was difficult to isolate effects on single OLs and to quantify the contribution of each cell type to OL Ca2+ responses using pharmacological approaches. For example, bath application of CdCl2 can block axonal or presynaptic Ca2+ channels, inhibiting vesicular glutamate release, as well as oligodendroglial Ca2+ channels, and thus a mixed response could be observed. Glutamate puff could be activating AMPARs on neighboring cells, resulting in increased firing and release of other molecules or ions to trigger OL Ca2+ transients. During periods of increased action potential firing, the concentration of K+ in the extracellular area increases, which in turn depolarizes OLs (Battefeld et al., 2016; Larson et al., 2018). Our findings in Figure 4 showed that increased extracellular K+ release from neighboring cells partially contributes to OL Ca2+ transients. This depolarization could then activate Cav channels in OLs, resulting in a Ca2+ rise. In addition, ATP and adenosine are other possible molecules, which are released in response to axon stimulation and have been shown to induce a Ca2+ rise in OL lineage cells (Hamilton et al., 2008; Hamilton, Vayro, Wigley, & Butt, 2010; Stevens, Porta, Haak, & Fields, 2002). Activation of neighboring cells may impact OL Ca2+ transients and explain the slow and prolonged kinetics of the Ca2+ rise in OLs in response to electrical stimulation of axons seen in Figure 6. In physiological conditions, Ca2+ dynamics in OLs are a result of a summation of various inputs over time when neuronal activity increases.

In this study, we examined OL Ca2+ rise to delineate sources of Ca2+ influx in OLs in response to glutamatergic input. An alternative mediator of Ca2+ transient duration may be a secondary Ca2+ source, such as release from internal stores in OLs. Amplitude and duration of Ca2+ transients in OLs induced by AMPA application were reduced after blocking ryanodine receptors (Ruiz et al., 2010), implicating a role for CICR from internal stores in AMPAR‐mediated Ca2+ transients. We tested the effect of dantrolene, a blocker of CICR, on glutamate‐induced Ca2+ rise in OLs, but there was no significant effect (Figure 2). In hippocampal NG2+ cells, Haberlandt et al. (2011) demonstrated glutamate‐induced Ca2+ rise mediated by Cav channels and Ca2+‐permeable AMPARs, similarly observed in the present study. In contrast to our results, Ca2+ influx triggered CICR and additionally amplified Ca2+ rise in hippocampal NG2+ cells (Haberlandt et al., 2011). As another intracellular Ca2+ source, a recent study demonstrates that mitochondrial Ca2+ release mediates microdomain Ca2+ transients and large Ca2+ waves in myelin sheaths (Battefeld, Popovic, de Vries, & Kole, 2019). This suggests that, in myelin sheaths of mature OLs, mitochondrial Ca2+ regulation is important for myelin remodeling.

What is the functional relevance of Ca2+ rise in OLs in response to glutamatergic input? Glutamate has been repeatedly shown to induce pathological damage to cultured OLs (Alberdi, Sánchez‐Gómez, Marino, & Matute, 2002). Similar to what occurs in neurons, excitotoxicity due to over‐activation of AMPARs leads to Ca2+ rise and downstream cascades leading to mitochondrial damage and apoptosis (Leuchtmann, Ratner, Vijitruth, Quab, & McDonald, 2003; Ruiz et al., 2010; Sánchez‐Gómez, Alberdi, Ibarretxe, Torre, & Matute, 2003). An irreversible elevation of Ca2+ through either AMPAR or Cav channels could induce necrosis of OLs in pathological conditions. However, dynamic Ca2+ transients are beneficial for signaling cascades of physiological processes of OLs, such as increased OL lineage cell survival and migration, as well as myelination (Gautier et al., 2015; Gudz, Komuro, & Macklin, 2006; Kougioumtzidou et al., 2017). Neuronal activity elicited a Ca2+ response in OLs in zebrafish, which was associated with myelin elongation or retraction, depending on the dynamics of the Ca2+ response (Baraban et al., 2018; Krasnow et al., 2018). In co‐cultures with dorsal root ganglion neurons, vesicular release of glutamate from axons stimulates local translation of myelin basic protein in OPCs and stimulates myelin induction through activation of NMDARs and mGluRs, rather than AMPARs (Wake et al., 2011, 2015). In the mouse brainstem, we identified AMPARs and Cav channels as an important Ca2+ influx pathway for Ca2+ dynamics in developing OLs. Recent studies suggest the potential role of AMPARs in OL development and myelination (Chen et al., 2018; Kougioumtzidou et al., 2017). Alterations in AMPAR expression in OPCs impact OL lineage cell differentiation, survival, and myelination in the corpus callosum (Chen et al., 2018; Kougioumtzidou et al., 2017). In addition, the electrical changes of the OL membrane are also important for OL development, because the knockdown of Cav channels or Nav channels throughout the OL lineage impaired OL development and myelination (Berret et al., 2017; Cheli et al., 2015, 2016). These studies support that AMPARs and Cav channels, as well as other signaling mechanisms, can play a role in OL Ca2+ dynamics as important mediators of glutamate‐mediated signaling to impact OL development.

Dynamic interaction between active axons and OLs in the auditory brainstem is particularly relevant to auditory function, because this brain region is heavily myelinated and operates at the upper limits of action potential frequency and speed required for sound localization (Kim, Renden, & von Gersdorff, 2013; Kim, Turkington, et al., 2013; Sinclair et al., 2017). Still, neuronal glutamate‐induced Ca2+ rise in OLs has been observed in brain areas other than the MNTB. Neuronal activity induced myelination‐associated Ca2+ rises in OLs in the zebrafish spinal cord (Baraban et al., 2018; Krasnow et al., 2018). Furthermore, OPC Ca2+ rise in response to glutamate or depolarization has been observed in the rodent optic nerve and hippocampus (Haberlandt et al., 2011; Hamilton et al., 2010; Sun et al., 2016). The present study provides a mechanistic explanation for neuronal activity‐dependent Ca2+ rise in pre‐myelinating OLs in the MNTB, but could be relevant for Ca2+ transients in developing OLs in other brain areas. Understanding the physiological properties of OLs and mechanisms of neuron‐OL interaction throughout the OL lineage and throughout the brain is critical for further comprehension of their roles in brain function. The mechanism of Ca2+ rise in pre‐myelinating OLs resulting from depolarization and neuronal input demonstrated in this study provides important insight into the dynamic physiological properties of OLs.

CONFLICT OF INTEREST

The authors declare no competing interests.

AUTHOR CONTRIBUTIONS

T.B. performed all experiments, analyses, and drafted the manuscript. J.H.K. supervised the project and revised the manuscript. All authors reviewed and approved the final manuscript.