The Upregulation of α₂δ-1 Subunit Modulates Activity-Dependent Ca²⁺ Signals in Sensory Neurons

Abstract

As auxiliary subunits of voltage-gated Ca²⁺ channels, the α₂δ proteins modulate membrane trafficking of the channels and their localization to specific presynaptic sites. Following nerve injury, upregulation of the α₂δ-1 subunit in sensory dorsal root ganglion neurons contributes to the generation of chronic pain states; however, very little is known about the underlying molecular mechanisms. Here we show that the increased expression of α₂δ-1 in rat sensory neurons leads to prolonged Ca²⁺ responses evoked by membrane depolarization. This mechanism is coupled to Ca_V2.2 channel-mediated responses, as it is blocked by a ω-conotoxin GVIA application. Once initiated, the prolonged Ca²⁺ transients are not dependent on extracellular Ca²⁺ and do not require Ca²⁺ release from the endoplasmic reticulum. The selective inhibition of mitochondrial Ca²⁺ uptake demonstrates that α₂δ-1-mediated prolonged Ca²⁺ signals are buffered by mitochondria, preferentially activated by Ca²⁺ influx through Ca_V2.2 channels. Thus, by controlling channel abundance at the plasma membrane, the α₂δ-1 subunit has a major impact on the organization of depolarization-induced intracellular Ca²⁺ signaling in dorsal root ganglion neurons.

Introduction

Calcium-activated signaling pathways underlie multiple cellular processes operating through complex spatial structures and a wide time range (Berridge et al., 2003). A tight balance between extracellular and intracellular calcium sources contributes to calcium dynamics in neurons where voltage-gated calcium channels (VGCCs) constitute the main regulators of calcium entry in response to membrane depolarization (Berridge, 1998). VGCCs are characterized by a pore-forming α1 subunit associated with two accessory proteins, a cytosolic β subunit, and a membrane-anchored α₂δ subunit (Bauer et al., 2010). α₂δ subunits modulate calcium channel current kinetics and also increase trafficking of the channel to the plasma membrane (Dolphin, 2012; Cassidy et al., 2014). Recent findings indicate that α₂δ proteins are crucial determinants of VGCC abundance at presynaptic terminals (Hoppa et al., 2012); thus, the overexpression of this subunit in hippocampal neurons promoted calcium channel localization at active zones, leading to an increase in vesicular release.

In sensory neurons, α₂δ-1 function has been associated with mechanisms for generation and maintenance of chronic pain. Peripheral nerve injury models of neuropathic pain in rodents resulted in a significant upregulation of α₂δ-1 protein levels in cell bodies and axon terminals of dorsal root ganglion (DRG) neurons, with a consequent accumulation of presynaptic α₂δ-1 protein in the dorsal horn of the spinal cord (Bauer et al., 2009). Conversely, damaged DRGs displayed no change in Ca_V2.2 mRNA or protein, which is the main VGCC type in sensory neurons (Xiao et al., 2002; Li et al., 2006). Although still debated, it is likely that the increased expression of α₂δ-1 subunit induced by nerve injury may increase VGCC trafficking toward the cell surface and presynaptic terminals. In-line with this hypothesis, experiments performed in transgenic mice overexpressing α₂δ-1 showed enhanced calcium currents recorded in DRG neurons, as well as nociceptive behavior characterized by hyperalgesia (Li et al., 2006). By contrast α₂δ-1 knock-out mice had reduced DRG calcium currents and lower baseline mechanical sensitivity (Patel et al., 2013).

DRG neurons exhibit diverse patterns for the regulation of intracellular calcium (Lu et al., 2006), among which the endoplasmic reticulum (ER) and mitochondria are the main contributors to activity-induced calcium increase (Fernyhough and Calcutt, 2010). The ER amplifies Ca²⁺ influx triggered by mild depolarization and promotes the propagation of a signal to the nucleus (Usachev and Thayer, 1997; Berridge, 1998), whereas mitochondria buffer high Ca²⁺ loads (Colegrove et al., 2000) particularly at synaptic terminals (Medvedeva et al., 2008). In this study, we show that the α₂δ-1 subunit has a key role in regulating the handling of intracellular calcium in sensory neurons. The overexpression of α₂δ-1 induces an upregulation of surface VGCCs and prolongs intracellular Ca²⁺ signals evoked by depolarization. Using pharmacological and genetic tools, we demonstrate that these sustained responses are mediated by augmented mitochondrial Ca²⁺ buffering of cytoplasmic Ca²⁺ increase induced by N-type channels.

Materials and Methods

DNA constructs.

The cDNAs used in this study were as follows: α₂δ-1 HA (Kadurin et al., 2012) α₂δ-1 MIDAS^AAA HA (Hoppa et al., 2012), HA Ca_V2.2 (Cassidy et al., 2014), and ratiometric Pericam (Nagai et al., 2001) expressed in pcDNA3.0; pEYFP, pECFP and pdsRed2-Mito (Clontech), pcDNA3.1 MCU^D260N,E263Q-FLAG and MCU–FLAG (Raffaello et al., 2013), pRK5 β1b, and Kir2.1-AAA (Tinker et al., 1996).

Reagents.

Fura-2AM was purchased from Invitrogen, ω-conotoxin GVIA from Alomone. Nifedipine, cyclopiazonic acid (CPA), antimycin, and oligomycin were obtained from Sigma-Aldrich.

Neuronal culture and transfection.

DRGs were isolated from P10 Sprague-Dawley rats of either sex. DRGs were dissociated in Hank's basal salt solution containing 5 mg/ml dispase (Invitrogen), 2 mg/ml collagenase (Worthington Biochemical), and 0.1 mg/ml DNase (Invitrogen) at 37°C for 30 min in a shaking water bath. Neuronal suspension was transfected by nucleofection following the manufacturer's instructions (Program G-13, Lonza). To improve cell viability after transfection, neurons were incubated in RPMI medium (Invitrogen) supplemented with 10% FBS (fetal bovine serum) and NGF (nerve growth factor; 50 ng/ml, Invitrogen) for 8 min at 37°C. DRGs were then plated on poly-l-lysine-coated coverslips (0.25 mg/ml, Sigma-Aldrich) and cultured in DMEM-F12 (Invitrogen) containing 10% FBS and 50 ng/ml NGF. α₂δ-1 HA cDNA was cotransfected with eCFP or eYFP in a 4:1 ratio (2 μg of total DNA). For the coexpression of α₂δ-1 HA, MCU^D260N,E263Q-FLAG, and eYFP cDNAs, the ratio used was 4:3:1. In control conditions, α₂δ-1 HA cDNA was replaced with an equivalent volume of empty vector. In live labeling experiments, control neurons were transfected with a control cDNA (Kir2.1-AAA cDNA encoding a nonfunctional potassium channel).

Calcium imaging.

Calcium imaging was performed on somata of small (≤25 μm) and medium (26–35 μm) DRGs, 40 h after transfection. Neurons were loaded with Fura-2-AM or Fura-FF-AM in DMEM-F12 medium supplemented with 2% FBS for 20 min at 37°C and washed for 5 min with bathing solution containing the following (in mm: 145 NaCl, 5 KCl, 2 CaCl₂, 1 MgSO₄, 10 HEPES, 10 Glucose, pH 7.4) and placed in a recording chamber under continuous superfusion (flow rate of 2.5–3 ml/min). Ca²⁺-free experiments were performed using a solution corresponding to the normal extracellular solution modified by the omission of CaCl₂, the addition of 0.1 mm EGTA and 2 mm MgCl₂ (Lu et al., 2006)_. Fura-2 and Fura-FF-loaded neurons were visualized on 20× objective with a Zeiss Axiovert 200M inverted microscope. Data were acquired using two imaging systems: Improvision Volocity software connected to a CCD camera (ORCA-ER; Hamamatsu Photonics) or MetaFluor Fluorescence Ratio Imaging Software (Cairn Research) via an iXon Ultra 897 camera (Andor Technology). Fura-2 excitation wavelengths at 340 and 380 nm were controlled either by a filter wheel or via an Optoscan monochromator (Cairn Research). Dual excitation filter at 340 and 380 nm, 400 nm dichroic mirror, and emission at 510/80 nm were purchased from Chroma Technology. Ratio signals were sampled at 0.5–1 Hz. Fluorescence was quantified within a region-of-interest after background subtraction. After confirmation of a stable baseline, neurons were depolarized by high K⁺ (50 or 100 mm, 10 s) or field stimulation (10 or 100 Hz). A positive response was defined as a 50% fluorescence increase with respect to the baseline. Only one field of neurons was recorded from each coverslip. No difference in the percentage of responding neurons was detected between control (41.5 ± 6.1%, total n = 86) and α₂δ-1 neurons (48.7 ± 5.8%, total n = 96; p = 0.4, t test). Peak amplitude was measured as the maximal signal observed within 20 s after depolarization, whereas the duration of a response was determined as the width at 25% of the maximal signal. For measurement of mitochondrial Ca²⁺, DRG neurons were transfected with ratiometric mtPericam (Nagai et al., 2001) and visualized on a 40× oil-immersion objective. The mitochondrial Ca²⁺ probe was excited at the pH insensitive wavelength of 380 nm using the Fura-2 excitation filter as described previously (Akimzhanov and Boehning, 2011). [Ca²⁺]mt was measured as −(F − F₀)/F₀ (Shutov et al., 2013) where F is the fluorescence at 380 nm and F₀ is the baseline fluorescence acquired before stimulation. Images were acquired at 1 Hz. Positively transfected neurons were identified by eYFP or mcherry fluorescence in Fura-2 or mtPericam imaging experiments respectively.

Neuron replating.

To remove neurite outgrowths and improve control of membrane potential in voltage-clamp experiments, neurons were replated as previously described (Page et al., 2010). Briefly, cells were incubated in a collagenase solution (0.2 mg/ml in serum free DMEM-F12) at 37°C for 10 min. Neurons were then resuspended in DMEM-F12/FBS and plated on poly-l-lysine-coated coverslips. Voltage-clamp experiments were performed 2–6 h after replating.

Electrophysiology.

As for calcium imaging, whole-cell voltage-clamp experiments were performed in small (<19 pF) and medium (20–38 pF) DRG neurons. Recordings were performed with Axopatch 200A amplifier (Molecular Devices) and analyzed with pClamp 9.0 software (Molecular Devices). Whole-cell voltage-clamp recordings were sampled at 10 kHz frequency, filtered at 2 kHz and digitized at 1 kHz. 80–85% series resistance compensation was applied and all recorded currents were leak subtracted using P/4 protocol. The extracellular solution for recording Ba²⁺ currents contained the following (in mm): 10 BaCl₂, 150 TEABr, 3 KCl, 1 NaHCO₃, 1 MgCl₂, 10 HEPES, 4 glucose, 0.001 TTX, pH 7.4, 320 mOsm. The patch internal solution contained the following (in mm): 140 Cs aspartate, 5 EGTA, 2 MgCl₂, 0.1 CaCl₂, 2 K₂ATP, 10 HEPES, pH 7.2, 292 mOsm. Membrane potential was held at −90 mV. For current-clamp experiments, the following solutions were used: extracellular (in mm): 145 NaCl, 5 KCl, 2 CaCl₂, 1 MgSO₄, 10 HEPES, and 10 glucose, pH 7.4, 316 mOsm; internal (in mm): 130 KCl, 10 EGTA, 10 HEPES, 8 NaCl, 4 Mg-ATP, 1 MgCl₂, 1 CaCl₂, 0.4 Na₂-GTP, pH 7.25 adjusted with 1 m KOH, 318 mOsm. Recording pipettes had access resistance of 1–4 MΩ. For recordings performed in the presence of CTX, the drug was superfused for 100 s at 1 μm concentration. Currents were measured before and after drug application at 6 ms after the start of a 15 mV test pulse.

Immunocytochemistry.

For live labeling experiments DRG cultures were incubated with a monoclonal anti-HA antibody (rat, 1:250; Roche) for 1 h at 37°C in bathing solution. This procedure, followed by cell fixation and application of a secondary antibody in nonpermeabilizing conditions, allowed the labeling of plasma membrane α₂δ-1 HA subunits without contamination from intracellular proteins. Briefly, after primary antibody incubation neurons were fixed with 4% PFA in PBS for 5 min and then blocked for 30–60 min with PBS supplemented with 10% goat serum. AlexaFluor-conjugated secondary antibody (1:500, Invitrogen) was incubated for 1 h at RT. For detection of intracellular proteins following surface labeling, neurons were permeabilized with 0.1% Triton for 8 min and then incubated with a primary antibody (1 h at RT). Samples were mounted on slides using Vectashield (Vector Laboratories) to reduce photobleaching. Imaging was performed using a confocal laser-scanning microscope (Zeiss) and a 40× oil-immersion objective. Optical sections of 1 μm thickness were acquired for each channel.

The approximate transfection efficiency of the α₂δ-1 HA subunit in DRG neurons was quantified by staining with an antibody against the HA epitope to be 32%, (n = 124 cells examined); this corresponded to an α₂δ-1 expression increase of 160 ± 30% (n = 139) compared with control endogenous proteins labeled with α₂δ-1 antibody (n = 142).

For quantitative analysis of neurite outgrowth patterns, live-labeled neurons were scored with respect to neurite length (average length of two longest neurites) and branching (average number of neuritic branch points per neurite length). The analysis was performed using NeuronJ software (Meijering et al., 2004).

Mitochondrial time lapse imaging.

Neurons were transfected with eCFP, pdsRed2-Mito, α₂δ-1 HA, or α₂δ-1MIDAS^AAA HA cDNAs in a 1:1:4 ratio. Forty hours after transfection, cultures were imaged at 37°C in bathing solution. Images were acquired every 2 s for 10 min. The percentage of moving mitochondria was analyzed in the distal part of neurites (100–150 μm from the soma). Before electrical stimulation (100 Hz, 10 s) neurons were imaged for 3 min in resting conditions. Mitochondria were defined as moving if they moved >1 μm in 1 min. Kymograph analysis was performed using ImageJ software as previously described (Macaskill et al., 2009). Kymographs were generated from live-imaging movies of 3 min before and after field stimulation. Moving mitochondria were identified using Manual Tracker plugin from ImageJ.

Western blotting.

DRG cultures transfected with eCFP (2 μg) or α₂δ-1 HA (2 μg) cDNAs were harvested in buffer A (50 mm Tris, pH7.5, 50 mm NaCl, and protease inhibitors; Complete, Roche). Neuronal suspensions were centrifuged at 60,000 × g for 1 h at 4°C. Pellets were lysed for 40 min at 4°C in buffer A supplemented with 1% Igepal and then centrifuged at 14,000 × g for 30 min at 4°C. Lysates were resolved by SDS-Page (3–8% NuPage Tris/acetate gels, Invitrogen), transferred to PVDF membranes and probed with antibodies to α₂δ-1 (mouse, 1:2000; Sigma-Aldrich) and β-tubulin III (rabbit, 1:2000; Sigma-Aldrich). Optical density quantification was performed with ImageJ. In every sample, the α₂δ-1 signal was normalized with respect to β-tubulin III content.

Statistical analysis.

Data were analyzed with GraphPad Prism 4.0 software or Origin7 (OriginLab). All data are shown as mean ± SEM; “n” refers to number of cells, unless indicated otherwise. The statistical significance between two groups was assessed by t test or Mann–Whitney U test. One-way ANOVA was used for comparison of means between three or more groups and two-way ANOVA to analyze the effect of two variables on an experimental response.

Results

Characterization of α₂δ-1 HA overexpressing neurons

α₂δ-1 Protein was overexpressed by nucleofection (Karra and Dahm, 2010) in DRG cultures obtained from postnatal rats. This in vitro model preserved most neuronal properties displayed by DRGs in vivo (Wood et al., 1988). To identify the expression of exogenous α₂δ-1 protein we used a construct engineered with an extracellular HA tag throughout these studies (α₂δ-1 HA; Kadurin et al., 2012). Exogenous α₂δ-1 HA subunits were found to be well expressed and localized at the cell surface in DRG cell bodies and neurites visualized with free eCFP (Fig. 1A). Upon transfection, the total expression of α₂δ-1 protein in neuronal lysates was increased by 63 ± 27% with respect to the control, as quantified by Western blotting (Fig. 1B, right; p = 0.02, Mann–Whitney U test). Overexpressed α₂δ-1 protein was also identified by an HA antibody (Fig. 1B, left). To examine whether α₂δ-1 transfection might affect neuronal morphology, we analyzed neurite outgrowth in control and α₂δ-1 HA overexpressing neurons (see Materials and Methods). No difference in neurite length (control: 106.4 ± 11.2 μm, n = 8; α₂δ-1 HA: 108.2 ± 15.7 μm n = 6; p = 0.92, t test) or number of branch points (control: 0.013 ± 0.004 μm⁻¹, α₂δ-1 HA: 0.010 ± 0.002 μm⁻¹; p = 0.52, t test) were measured between control and α₂δ-1 overexpressing DRGs.

Figure 1.

α₂δ-1 HA overexpression modulates Ca²⁺ response duration in DRG neuron cultures. A, 3D projection from z-stacks of confocal images of a DRG neuron transfected with α₂δ-1 HA and eCFP (right) or eCFP and nonfunctional Kir2.1-AAA cDNAs (left). Exogenous α₂δ-1 HA proteins were detected at the cell surface by live labeling (n = 6 experiments). No staining for the HA epitope was observed in neurons expressing eCFP only. Scale bars, 10 μm. B, Western blotting of neuronal lysates of eCFP or α₂δ-1 HA transfected DRGs. Lysates were loaded in duplicate and the membrane was probed for HA (left) or α₂δ-1 subunit expression. The neuronal content was quantified by β tubulin III (β-tub) staining (n = 8 DRG cultures, 4 gels). C, Fura-2 imaging of high K⁺-evoked Ca²⁺ transients performed in control (left; representative of n = 19) and α₂δ-1 HA overexpressing neurons (right; representative of n = 24). Traces are shown after baseline subtraction. D, Control, open bars; α₂δ-1 HA, gray bars. The width and total area of transients were significantly higher in α₂δ-1 HA overexpressing neurons with respect to control DRGs (*p = 0.01, t test). E, Representative Fura-FF transients imaged in control (n = 11) and α₂δ-1 HA neurons (n = 7). F, Examples of Ca²⁺ transients induced by field stimulation (100 Hz, 10 s) of control (left) and α₂δ-1 HA DRGs (right).

Next we addressed the impact of α₂δ-1 HA overexpression on Ca²⁺ entry. DRGs are a heterogeneous population of neurons, where function correlates with cell size (Basbaum et al., 2009). For this reason we focused our functional studies on the two major classes of nociceptors corresponding to small (<25 μm) and medium (25–35 μm) diameter DRG neurons, which are known to give rise respectively to C and Aδ afferent sensory fibers (Julius and Basbaum, 2001). We performed Fura-2 imaging on DRG cultures depolarized by a 10 s application of high (50 mm) K⁺. Table 1 summarizes some electrophysiological properties measured in current-clamp recordings before and during high K⁺ application to transfected neurons. Figure 1C shows representative high K⁺-evoked Ca²⁺ transients imaged 40–48 h after transfection in control and α₂δ-1 overexpressing DRG neurons. The peak of the response was unaltered by α₂δ-1 HA overexpression (control_F340/380: 2.03 ± 0.12, n = 19; α₂δ-1 HA_F340/380: 2.37 ± 0.16, n = 24; p = 0.13), although in these neurons the Ca²⁺ signals displayed a slower recovery after the peak, leading to a prolonged Ca²⁺ rise (“Ca²⁺ hump”; Fig. 1C, right trace). There was a significant increase of both response width and total area (Fig. 1D) of high K⁺-evoked Ca²⁺ transients in α₂δ-1 HA overexpressing DRGs compared with control neurons. As shown in Figure 1E, a slower recovery of the response in α₂δ-1 HA overexpressing neurons was also observed in the presence of the lower affinity Ca²⁺ dye, Fura-FF.

Table 1.

Electrophysiological properties of control and α₂δ-1 DRGs

Parameters	Control	α2δ-1	p
	(n = 26)	(n = 31)	(t test)
Vrest (mv)	−58.9 ± 0.8	−56.7 ± 1.0	0.1
VK+ (mv)	−18.4 ± 0.7	−17.7 ± 0.6	0.4
AP	4.5 (2, 8)	6.0 (4, 9)	0.9

V_rest, Resting membrane potential; V_K+, membrane potential during 50 mm K⁺; AP, median number of action potentials (25 and 75% percentile).

A similar modulation of Ca²⁺ signals was observed when cultures were field-stimulated at 100 Hz for 10 s. As illustrated in Figure 1F, intense electrical stimulation induced an extended Ca²⁺ rise in neurons transfected with α₂δ-1 HA (55 ± 7 s, n = 10), but not in the control neurons (32 ± 8 s, n = 7; p = 0.04, t test). By contrast, no change in response width was detected when neurons were stimulated at a lower (10 Hz) frequency (control: 37.6 ± 5.8 s, n = 18; α₂δ-1 HA DRGs: 41.1 ± 4.3 s, n = 18; p = 0.63, t test), indicating that the α₂δ-1-mediated Ca²⁺ hump depended on a sustained activation of VGCCs.

Mutation of the extracellular MIDAS motif prevents α₂δ-1 HA effects on Ca²⁺ signals

α₂δ Subunits are characterized by a conserved von Willebrand A domain (VWA), which mediates the interaction with extracellular proteins through the metal-ion-dependent adhesion site motif (MIDAS). This consensus sequence is involved in the coordination of divalent cations and has been found to be essential for the ability of α₂δ-1 and α₂δ-2 to modulate VGCC function (Cantí et al., 2005; Hoppa et al., 2012) and trafficking (Cassidy et al., 2014). We transfected DRG neurons with an α₂δ-1 construct carrying three point mutations within the MIDAS motif (α₂δ-1 MIDAS^AAA HA; Hoppa et al., 2012) and examined the effect of the mutant subunit overexpression on Ca²⁺ signaling. First we monitored α₂δ-1 MIDAS^AAA HA protein expression in DRG cultures using a combined staining of the cell surface via live labeling, followed by detection of intracellular α₂δ-1 HA after cell permeabilization (Fig. 2A). Figure 2B quantifies the ratio between surface and cytosolic HA staining in α₂δ-1 HA and α₂δ-1 MIDAS^AAA HA transfected neurons. There was a marked reduction of surface expression of α₂δ-1 MIDAS^AAA compared with wild-type α₂δ-1 whereas the expression of intracellular α₂δ-1 was unchanged in the examined cells (intra α₂δ-1 HA: 85.4 ± 15.1 A.U., n = 9; intra α₂δ-1 MIDAS^AAA HA: 79.8 ± 19.7 A.U., n = 5; p = 0. 82, t test).

Figure 2.

Expression of α₂δ-1 MIDAS^AAA HA in DRG cultures. A, Confocal images of DRG neurons transfected with α₂δ-1 HA or α₂δ-1 MIDAS^AAA HA subunits. Top, Cell surface HA (Surf HA). Scale bar, 10 μm. Positively transfected neurons were identified following detection of intracellular HA-tagged proteins (intra HA, bottom). B, Quantification of surface HA fluorescence for wild-type α₂δ-1 HA overexpressing neurons (gray bar; n = 9) and α₂δ-1 MIDAS^AAA HA (checked bar; n = 5). Membrane HA signal was normalized against the intracellular HA content (*p = 0.01, Mann–Whitney test). C, Examples of families of I_Ba current traces for control, α₂δ-1 HA and α₂δ-1 MIDAS^AAA HA overexpressing DRG neurons. Currents were evoked from −90 mV holding potential in 5 mV steps from −40 to + 70 mV. The scale bars refer to all panels. D, Calcium channel current density-voltage relationship for control (□; n = 36), α₂δ-1 HA (●; n = 32) and α₂δ-1 MIDAS^AAA HA (▴; n = 16). In each experiment, current density (pA/pF) recorded in wild-type and mutant α₂δ-1 HA overexpressing DRGs was normalized with respect to the control condition. At +10 and +15 mV, the current density of α₂δ-1 HA overexpressing DRGs was significantly higher than the control condition (p = 0.02 one-way ANOVA and Dunnett's test, *p < 0.05). E, Example of high K⁺-evoked Ca²⁺ transients in control neurons (left), wild-type α₂δ-1 HA (middle), and α₂δ-1 MIDAS^AAA HA (right) overexpressing DRG neurons. F, α₂δ-1 HA neurons (gray bar; n = 15) showed a prolonged response duration with respect to α₂δ-1 MIDAS^AAA HA (checked bar; n = 15) and control (open bar; n = 15) DRGs (p = 0.006 one-way ANOVA and Bonferroni post hoc test; *p < 0.05, **p < 0.01).

In-line with previous findings (Cantí et al., 2005; Hoppa et al., 2012), patch-clamp experiments confirmed that α₂δ-1 MIDAS^AAA was unable to potentiate Ca²⁺ current density, whereas the overexpression of wild-type α₂δ-1 strongly increased current density (Fig. 2C,D). When we examined the impact of the MIDAS mutation on intracellular Ca²⁺ signals we found that overexpression of α₂δ-1 MIDAS^AAA HA did not alter the shape of Ca²⁺ transients (Fig. 2E), producing responses with a duration similar to the control group and with faster recovery compared with wild-type α₂δ-1 HA overexpressing neurons (Fig. 2F). These results suggest that the presence of a functional α₂δ-1 subunit at the cell surface is critical for its regulatory role on neuronal Ca²⁺ pathways.

Role of Ca_V2.2 channels in α₂δ-1 HA mediated prolongation of Ca²⁺ signals

DRG neurons express a number of different VGCC subtypes and Ca_V2.2 (N-type) channels represent the main contributors to membrane depolarization-induced calcium influx (Scroggs and Fox, 1992; Bell et al., 2004). To investigate the influence of the different VGCCs on Ca²⁺ responses shaped by α₂δ-1 HA overexpression, we performed Fura-2 imaging in the presence of ω-conotoxin GVIA (CTX, 1 μm) or nifedipine (Nif, 1 μm), which are respectively N- and L-type (Ca_V1) channel blockers (Fig. 3A–F). As shown in Figure 3A,D, the peak amplitude of high K⁺-evoked Ca²⁺ transients of both control and α₂δ-1 overexpressing neurons was reduced by 15 min pretreatment with CTX. Moreover in α₂δ-1 HA overexpressing DRGs, the block of Ca_V2.2 channels abolished the prolongation of the Ca²⁺ signals (Fig. 3F). By contrast, in control neurons, CTX application had no effect on response width (Fig. 3C), suggesting that Ca_V2.2 channels are involved in α₂δ-1 modulation of the Ca²⁺ increase. Also we observed that Ca_V1 family channels were not critical for the generation of responses evoked by 50 mm K⁺, because continuous superfusion of DRG cultures with nifedipine did not change the shape of the Ca²⁺ transient, either in control (Fig. 3B,C) or in α₂δ-1 HA overexpressing neurons (Fig. 3E,F). Thus the contribution of Ca_V2.2 channels in response to a strong depolarization is much greater than the Ca_V1 channel contribution to intracellular Ca²⁺ signals in DRG somata, as previously described for sympathetic neurons (Wheeler et al., 2012).

Figure 3.

Ca_V2.2-mediated signaling in α₂δ-1 neurons. A, Imaging of high K⁺-evoked Ca²⁺ transients in control DRG neurons pretreated for 15 min with 1 μm CTX. B, In control DRG neurons, in the absence of α₂δ-1 HA overexpression CTX application (n = 7) decreased the peak amplitude of responses with respect to untreated (n = 21) and Nif (1 μm; n = 13) samples (p < 0.0001 one-way ANOVA and Bonferroni post hoc test, #p < 0.001). C, Inhibition of N- and L-type channels did not alter the width of signals imaged in control neurons (p = 0.4, one-way ANOVA). D, Examples of Ca²⁺ responses evoked in α₂δ-1 HA overexpressing neurons in the absence (left) or presence (right) of CTX. E, CTX pretreatment (n = 12) diminished the intensity of signals in α₂δ-1 HA overexpressing DRG neurons compared with responses obtained in untreated (n = 19) or Nif-treated (n = 19) DRGs (p = 0.0001 one-way ANOVA and Bonferroni post hoc test, #p < 0.001). F, Prolonged high K⁺-evoked Ca²⁺ transients were abolished following CTX pretreatment (p = 0.02 one-way ANOVA and Bonferroni post hoc test, *p < 0.05). G, Overexpression in DRG neurons of exofacially HA-tagged Ca_V2.2, β1B subunits and either α₂δ-1 (top) or α₂δ-1 MIDAS^AAA (bottom). Surface and intracellular detection of HA Ca_V2.2 proteins was performed by staining the HA epitope in nonpermeabilizing conditions and intracellular labeling with an anti Ca_V2.2 antibody, respectively. Scale bars, 10 μm. H, Quantification of surface HA staining in HA Ca_V2.2/α₂δ-1/β1B (black bar; n = 32) or HA Ca_V2.2/α₂δ-1 MIDAS^AAA/β1B (white bar; n = 36) transfected neurons (p = 0.02, t test). I, Labeling of overexpressed channels by Ca_V2.2 polyclonal antibody. DRG neurons were transfected with Ca_V2.2YFP and β1B cDNAs. J, Ca_V2.2 antibody failed to identify native subunits in eYFP overexpressing neurons indicating that the amount of protein is a limiting factor in immunocytochemistry experiments.

In agreement with these results, in whole-cell patch-clamp recordings CTX application significantly reduced Ba²⁺ current density of control (−64 ± 8 pA/pF, n = 10) and α₂δ-1 overexpressing neurons (−103 ± 19 pA/pF, n = 11) to −31 ± 5 pA/pF and −37 ± 7 pA/pF respectively (p < 0.01, t test), indicating a greater percentage of current blocked by CTX in the presence of α₂δ-1 subunit (63 ± 2%) compared with control condition (53 ± 3%; p = 0.016, t test).

Next, to address the hypothesis that α₂δ-1 overexpression could indeed modulate surface N-type calcium channels, neurons were cotransfected with α₂δ-1 cDNA and a Ca_V2.2 construct containing an HA tag in an extracellular loop (Cassidy et al., 2014). This technique was used because the lack of commercially available antibodies directed against extracellular epitopes prevented the detection of native surface-expressed Ca_V2.2 subunits. We compared the cell surface localization of HA-Ca_V2.2 when coexpressed with wild-type α₂δ-1 or α₂δ-1MIDAS^AAA (Fig. 3G). As shown in Figure 3G,H, surface expression of N-type channels was promoted by wild-type α₂δ-1 relative to α₂δ-1 MIDAS^AAA (Fig. 3H). Total expression of N-type channels was assessed by staining permeabilized neurons with a Ca_V2.2 antibody directed against the intracellular II–III loop (Raghib et al., 2001). In immunocytochemistry experiments this antibody enabled detection of overexpressed Cav2.2 subunits only (Fig. 3I,J).

α₂δ-1 HA modulation of Ca²⁺ transients is insensitive to depletion of the ER Ca²⁺ stores

What is the mechanism underlying the depolarization-evoked extended Ca²⁺ rise in α₂δ-1 HA overexpressing neurons? Figure 4A shows that initiation of the high K⁺-induced response was dependent on extracellular Ca²⁺, whereas the shape of the “hump” component did not change when α₂δ-1 HA overexpressing DRGs were incubated in a Ca²⁺-free bathing solution immediately after depolarization (Fig. 4B), indicating that the prolonged signal is mediated by Ca²⁺ release from intracellular compartments. The mechanism of calcium-induced calcium release from the ER has been found to regulate the depolarization-induced Ca²⁺ increase in a subpopulation of DRG neurons (Shmigol et al., 1995; Lu et al., 2006). To investigate the role of the ER in the α₂δ-1-mediated response, we depleted intracellular stores with CPA, a selective blocker of the sarcoendoplasmic reticulum Ca²⁺-ATPase (SERCA). Figure 4C,D shows example traces in untreated neurons, or after preincubation with CPA (5 μm, 15 min), for both control (Fig. 4C) and α₂δ-1 overexpressing cells (Fig. 4D). In the absence of α₂δ-1 HA overexpression, CPA induced a widening of the response, implying a role for ER stores in Ca²⁺ clearance after depolarization. However, in α₂δ-1 HA overexpressing neurons CPA did not affect the shape of Ca²⁺ transients (Fig. 4E). As expected, CPA application increased baseline Ca²⁺ levels both in control (baseline before CPA: 0.98 ± 0.02, n = 15; baseline after CPA: 1.11 ± 0.03 n = 15; p = 0.006, paired t test) and in α₂δ-1 HA overexpressing neurons (baseline before CPA: 0.92 ± 0.03 F340/380, n = 14; baseline after CPA: 1.15 ± 0.05 F340/380, n = 14; p = 0.002, paired t test), indicating a role for SERCA under resting conditions.

Figure 4.

ER stores are not involved in shaping α₂δ-1 HA prolonged responses. A, Preincubation of DRGs in Ca²⁺ free bathing solution during the period shown by the gray shading abolished high K⁺-induced transients. The response was recovered after a subsequent application of extracellular 2 mm Ca²⁺. B, The Ca²⁺ hump observed in high K⁺-evoked signal of α₂δ-1 HA overexpressing neurons is not affected by perfusion with 0 mm Ca²⁺ after the initiation of the peak response (representative of n = 4). C, D, High K⁺-evoked Ca²⁺ transients imaged before and after store depletion by CPA pretreatment (right trace; 5 μm, 15 min) in control (C) and α₂δ-1 HA overexpressing neurons (D). E, Ca²⁺ transients evoked in α₂δ-1 HA overexpressing DRGs (striped bars) were not altered by CPA preapplication (gray bars; n = 14) when compared with neurons processed in parallel without the drug (white bars; n = 15). By contrast, in control neurons, CPA pretreatment (n = 14) resulted in responses with a slower recovery than those from untreated neurons (n = 14; interaction: p = 0.001, two-way ANOVA and Bonferroni post hoc test, *p < 0.05, **p < 0.01). F, G, Application of CPA in Ca²⁺-free solution (gray bar; 20 μm, 5 min). Caffeine (Caff; 5 mm) was then applied, and the lack of response indicated store depletion. Adding Ca²⁺ into the extracellular solution (white bar) triggered a similar store-operated calcium channel response in both control DRG neurons (F) and those overexpressing α₂δ-1 HA (G).

To investigate whether overexpression of α₂δ-1 HA might interfere with the correct functioning of the ER stores, we applied CPA in a Ca²⁺-free bathing solution and measured Ca²⁺ leak from the ER. Caffeine was then applied to assess the effectiveness of CPA-mediated store depletion. In addition we then monitored store-operated Ca²⁺ entry by reverting to 2 mm Ca²⁺ after store depletion (Fig. 4F,G). No difference in ER Ca²⁺ content was detected in α₂δ-1 HA overexpressing DRGs (total area: 56.58 ± 7.36, n = 14), compared with control neurons (total area: 51.24 ± 8.07, n = 13; p = 0.6, t test). Furthermore upon CPA-mediated store depletion, the addition of extracellular Ca²⁺ to the bathing solution resulted in a store-operated Ca²⁺ channel response of equal peak amplitude between α₂δ-1 overexpressing and control neurons (control: 0.83 ± 0.06 F340/380, n = 11; α₂δ-1: 0.70 ± 0.04, n = 14; p = 0.1, t test).

Mitochondria buffer signals enhanced by α₂δ-1 HA upregulation

Mitochondria are important regulators of Ca²⁺ dynamics in DRG neuron cell bodies (Werth and Thayer, 1994) and neurite terminals (Medvedeva et al., 2008). These organelles accumulate Ca²⁺ into their matrix through a calcium uniporter mechanism driven by the negative mitochondrial membrane potential. Ca²⁺ is then slowly released back into the cytoplasm via Na⁺/Ca²⁺ exchangers (Rizzuto et al., 2012), and pumped out of the cell. We first used a pharmacological approach to identify the contribution of mitochondria to the Ca²⁺ response in α₂δ-1 HA overexpressing DRGs. Figure 5 shows representative Ca²⁺ signals and quantification in control (Fig. 5A,B) and α₂δ-1 HA overexpressing neurons (Fig. 5C,D) with or without the application of antimycin (anti; 0.3 μm) and oligomycin (oligo; 1 μm). The combination of these drugs has been previously found to be effective in blocking Ca²⁺ uptake into mitochondria while preventing ATP depletion (Medvedeva et al., 2008). When Ca²⁺ transients were evoked in the presence of antimycin and oligomycin, there was an increase in the peak amplitude of responses from both control (Fig. 5B) and α₂δ-1 HA overexpressing DRGs (Fig. 5D). However, in the latter condition the hump component was abolished giving rise to a marked reduction of response duration (Fig. 5C,E). A significant interaction (Fig. 5E; p = 0.004, two-way ANOVA) between the drug treatment and α₂δ-1 overexpression underlined a synergistic action of the two factors.

Figure 5.

Role of mitochondria in the intracellular Ca²⁺ response in control and α₂δ-1 HA overexpressing DRGs. A, High K⁺-triggered Ca²⁺ signals in the absence (left) or presence (right) of antimycin and oligomycin (anti + oligo; 5 min preapplication) in control DRG neurons. B, Quantification of peak response after the application of anti + oligo in control DRG neurons (striped bar;, n = 22) compared with untreated neurons (open bar; n = 19) processed in parallel. The peak amplitude of responses is increased by anti + oligo application (**p = 0.002, t test). C, Example of Ca²⁺ transients imaged in α₂δ-1 HA overexpressing DRGs following anti + oligo pretreatment, as in A. D, Quantification of peak response after the application of anti + oligo in α₂δ-1 HA overexpressing DRG neurons (gray striped bar; n = 20) compared with untreated neurons processed in parallel (gray bar; n = 27, *p = 0.03, t test). E, The increased Ca²⁺ transient duration mediated by α₂δ-1 HA overexpression (gray bars) compared with control DRG neurons (open bars) is inhibited by anti + oligo treatment (striped bars; two-way ANOVA and Bonferroni post hoc test, *p < 0.05).

The recent identification of the channel responsible for mitochondrial Ca²⁺ uptake (MCU) has allowed the development of more sensitive tools to study Ca²⁺ entry into mitochondria (De Stefani et al., 2011). Because the uniporter machinery involves the formation of a tetramer, we were able to use a dominant-negative form of MCU protein (MCU^D260N,E263Q; Raffaello et al., 2013) to knockdown mitochondrial calcium buffering in DRGs. As described previously (Raffaello et al., 2013), the overexpression of MCU^D260N,E263Q carrying two point mutations in the pore-forming domain oligomerize with endogenous subunits giving rise to a substantial reduction of calcium uptake into mitochondria (∼47% decrease of [Ca²⁺]mt). Calcium imaging experiments shown in Figure 6A indicate that in the absence of overexpressed α₂δ-1 HA subunit the overexpression of MCU^D260N,E263Q (MCU^NQ) did not affect the amplitude of high K⁺-evoked Ca²⁺ transients (Fig. 6B). The discrepancy between this result and the data obtained with antimycin and oligomycin treatment is likely to be due to only a partial loss of function induced by MCU^D260N,E263Q, as many mitochondria retained functioning MCU machinery (Raffaello et al., 2013). Nevertheless when MCU^D260N,E263Q was cotransfected together with α₂δ-1, the peak response showed a consistent increase of 21 ± 9% (Fig. 6C,D; p = 0.04). In addition the overexpression of MCU^D260N,E263Q exerted an effect on Ca²⁺ transient duration (Fig. 6E), leading to an inhibition of the prolonged Ca²⁺ rise in α₂δ-1 HA overexpressing neurons, whereas no significant change was detected in the shape of signals in control DRGs. These findings showed that in the presence of α₂δ-1 HA, increases in intracellular Ca²⁺ were rapidly taken up by mitochondria, and subsequently released into the cytoplasm resulting in a prolonged Ca²⁺ response.

Figure 6.

MCU contributes to generate the Ca²⁺ hump in DRGs overexpressing α₂δ-1 HA. A, High K⁺-induced Ca²⁺ transients in control DRG neurons, in the absence (left; n = 31) or presence (right; n = 22) of MCU^D260N,E263Q. B, No change in the peak amplitude after MCU^D260N,E263Q overexpression (checked bar compared with open bar) in control DRG neurons (p = 0.4, t test). C, High K⁺-induced Ca²⁺ transients in α₂δ-1 HA overexpressing (left; n = 28) and α₂δ-1 HA + MCU^D260N,E263Q overexpressing (right; n = 23) DRGs. D, Cotransfection of MCU^D260N,E263Q and α₂δ-1 HA subunit increased the peak amplitude of high K⁺-induced Ca²⁺ transients (gray checked bar compared with solid gray bar, *p < 0.05, t test). E, Following MCU^D260N,E263Q overexpression, α₂δ-1 HA neurons did not display prolonged Ca²⁺ response in response to 50 mm K⁺ (gray checked bar compared with solid gray bar; p = 0.01 two-way ANOVA and Bonferroni post hoc test, *p < 0.05). There was no effect of MCU^D260N,E263Q on the response width in control DRG neurons (checked bar compared with open bar). F, MtPericam responses stimulated by the application of 20 and 100 APs at 10 Hz in control (n = 26; black trace) and α₂δ-1 HA neurons (n = 31; top gray trace). G, Increase of 100 AP-induced mitochondrial Ca²⁺ uptake in the presence of α₂δ-1 HA and normalized to [Ca²⁺]mt in control neurons (p = 0.034; Kruskal–Wallis one-way ANOVA and Dunn's multiple-comparison test, *p < 0.05).

To directly monitor changes in mitochondrial Ca²⁺ uptake following membrane depolarization, control and α₂δ-1 HA overexpressing neurons were cotransfected with mtPericam (Nagai et al., 2001), a Ca²⁺ probe selectively localized within mitochondria. Stimulation of DRGs with 20 and 100 action potentials (APs) at 10 Hz induced a large increase in [Ca²⁺]mt (Fig. 6F), confirming the role of mitochondria in buffering depolarization-evoked Ca²⁺ signals (Colegrove et al., 2000). Moreover, comparison between control and α₂δ-1 HA overexpressing neurons demonstrated an augmented mitochondrial Ca²⁺ uptake in α₂δ-1 HA overexpressing DRGs stimulated with 100 APs (Fig. 6G).

Because the rate of mitochondrial Ca²⁺ buffering correlates with changes in cytosolic Ca²⁺ levels (Colegrove et al., 2000), we increased Ca²⁺ entry by the application of 100 mm K⁺ (experimental protocol shown in Fig. 7A) and assessed the contribution of MCU^D260N,E263Q overexpression to Ca²⁺ transients evoked in control and α₂δ-1 HA DRGs. This stronger depolarization protocol induced larger signals with a longer duration in both α₂δ-1 HA overexpressing and control neurons (Fig. 7B), resulting in the generation of an equal duration of Ca²⁺ response in both conditions. As expected the inhibition of prolonged Ca²⁺ transients linked with MCU^D260N,E263Q overexpression was strongly enhanced by 100 mm K⁺ application (Fig. 7C). However only when MCU^D260N,E263Q was coexpressed with α₂δ-1 HA, was the shortening of the response width associated with a significant rise of the peak amplitude (∼20% increase of F340/380 in α₂δ-1 HA/MCU^D260N,E263Q vs ∼5% increase in control/MCU^D260N,E263Q; Fig. 7D). This finding indicates that that the activation of mitochondrial Ca²⁺ uptake depends, not only on the intensity of the response but also may be linked to a specific cellular pathway. Thus the upregulation of Ca_V2.2 channels in α₂δ-1 HA overexpressing neurons may recruit mitochondria into a predominant role of shaping Ca²⁺ transients (Fig. 7E).

Figure 7.

MCU^D260N,E263Q effect on 100 mm K⁺-evoked Ca²⁺ response in control and α₂δ-1 neurons. A, Increasing concentrations of KCl (50, 100 mm) correspond to a consequent elevation of peak and duration of Ca²⁺ transients imaged in control (representative of n = 22; p < 0.0001, paired t test) and α₂δ-1 neurons (n = 10; p < 0.0001, p < 0.0008 for area and duration respectively, paired t test). B, Ca²⁺ traces triggered by the application of 100 mm K⁺ (10 s) in control and α₂δ-1 HA DRGs with or without the expression of MCU dominant negative. C, MCU^D260N,E263Q overexpression significantly reduced duration of Ca²⁺ transients activated by 100 mm K⁺ (p = 0.00002, two-way ANOVA and Bonferroni post hoc test, *p < 0.05, **p < 0.01). D, Quantification of peak amplitude of response in control (n = 20), control+ MCU^D260N,E263Q (n = 13), α₂δ-1 HA (n = 10), and α₂δ-1 HA + MCU^D260N,E263Q (n = 14) DRGs. Overexpression of mutant MCU increased peak amplitude of signals imaged in α₂δ-1 HA neurons only (p = 0.02, two-way ANOVA and Bonferroni post hoc test). E, Proposed mechanism for α₂δ-1 regulation of activity dependent-Ca²⁺ signals. Increased levels of α₂δ-1 protein lead to surface Ca_V2.2 channel upregulation. N-type channel-mediated Ca²⁺ influx is primarily buffered by mitochondria which contribute together with plasma membrane Ca²⁺-ATPase (PMCA) to limit the amount of Ca²⁺ increase in the cytoplasm.

Activity-dependent reduction of mitochondrial trafficking in α₂δ-1 HA overexpressing neurites

Mitochondria are dynamic organelles synthesized in the cell body and trafficked along neuronal processes at velocities ranging from 0.3 to 1 μm/s (MacAskill and Kittler, 2010). They are characterized by a bidirectional saltatory movement, which is regulated by intracellular Ca²⁺ levels (Wang and Schwarz, 2009). An increase in Ca²⁺ concentration induces the arrest of mitochondrial trafficking, leading to a rapid distribution of mitochondria in regions demanding high Ca²⁺ buffering. We hypothesized that by promoting VGCC trafficking, α₂δ-1 HA overexpression might increase local Ca²⁺ influx within the neurites and exert an effect on mitochondrial axonal transport. We measured mitochondrial motility in wild-type and α₂δ-1 MIDAS^AAA overexpressing DRGs, both in resting conditions and during field stimulation at 100 Hz. Mtdsred2 (Macaskill et al., 2009), a fluorescent protein exclusively localized in mitochondria (Fig. 8A, bottom), was used to visualize these organelles in the neurites during time lapse imaging experiments. Kymographs were generated to identify mobile and stationary mitochondria (Fig. 8B–D). The overexpression of α₂δ-1 HA or α₂δ-1 MIDAS^AAA mutant did not change the percentage of moving mitochondria in resting conditions (Fig. 8E). By contrast, intense electrical stimulation significantly reduced the fraction of moving mitochondria in α₂δ-1 HA-expressing neurons compared with nonstimulated conditions (Fig. 8F). In-line with the finding that calcium channel abundance at presynaptic sites is controlled by the level of expression of wild-type α₂δ-1 subunit (Hoppa et al., 2012), in the presence of α₂δ-1 MIDAS^AAA HA we found no difference in the percentage of moving mitochondria before and after stimulation. Furthermore after α₂δ-1 HA overexpressing cultures were pretreated with CTX (1 μm, 15 min; Fig. 8D), field stimulation did not cause mitochondria to stall, suggesting that the increased Ca²⁺ influx in response to depolarization is mainly mediated by Ca_V2.2 channels trafficked to the neurites by the α₂δ-1 subunit.

Figure 8.

Reduction of axonal mitochondrial transport following electrical stimulation of α₂δ-1 HA overexpressing DRG neurons. A, For time-lapse experiments, neurons were transfected with α₂δ-1 HA and eCFP to identify neuron morphology (top image) together with pdsRed2-Mito plasmid to label mitochondria (bottom image). Scale bar, 100 μm. B–D, Kymographs showing mitochondrial motility before and during the application of field stimulation in α₂δ-1 HA (B), α₂δ-1 MIDAS^AAA HA (C), or α₂δ-1 HA neurons pretreated with CTX (1 μm for 15 min; D). The first frame of the time-lapse movie is shown above each kymograph. Vertical arrows indicate a 2 min imaging period. Scale bars, 10 μm. E, No difference in the percentage of moving mitochondria between α₂δ-1 HA-expressing DRGs (gray bar; n = 239 mitochondria), α₂δ-1 MIDAS^AAA HA-expressing DRGs (black bar; n = 219 mitochondria), or α₂δ-1-expressing DRGs treated with CTX (gray striped bar; n = 234 mitochondria) before field stimulation (p = 0.7, Kruskal–Wallis one-way ANOVA). F, The proportion of moving mitochondria after field stimulation was normalized with respect to the moving mitochondria in each resting condition, and was decreased in α₂δ-1 HA-expressing DRGs (gray bar; p = 0.01, Kruskal–Wallis one-way ANOVA and Dunn's multiple-comparison test, **p < 0.01).

Discussion

VGCC α₂δ subunits are key molecules in the regulation of sensory neuron plasticity as their upregulation per se exerts effects on nociceptive behavior (Li et al., 2006). Moreover increased α₂δ-1 protein levels in damaged DRG neurons contribute to the enhanced neurotransmission and hyperexcitability observed in neuropathic pain models (Campbell and Meyer, 2006; Patel et al., 2013), although no details about the molecular mechanisms have been reported.

In this work, we describe an in vitro model to study cellular changes triggered by increased α₂δ-1 protein levels. Our data show that α₂δ-1 HA upregulation enhances Ca²⁺ signal duration in response to brief membrane depolarization. Ca²⁺ transients were characterized by an initial peak followed by a prolonged Ca²⁺ rise, which did not depend on extracellular calcium. Similar effects on evoked Ca²⁺ responses have been observed in vivo in DRGs subjected to inflammatory insult (Fuchs et al., 2007; Lu and Gold, 2008). Surprisingly, in our assays the generation of this Ca²⁺ hump was mainly associated with N-type calcium channel activity as it was blocked by CTX, but not by application of the L-type channel blocker nifedipine. Also α₂δ-1 HA overexpression prolonged the duration of Ca²⁺ transients evoked by field stimulation at 100 Hz frequency but did not change the shape of 10 Hz-triggered responses. This result is in agreement with the finding that 100 Hz-evoked Ca²⁺ signals are preferentially mediated by Ca_V2 channels (Wheeler et al., 2012). The experiments performed with mtPericam, confirmed this phenomenon, indicating a prominent role of mitochondria in buffering intracellular Ca²⁺ in response to a strong depolarization in α₂δ-1 overexpressing neurons, compared with the control condition.

Functional data related to the effect of the α₂δ-1 HA subunit on the intracellular Ca²⁺ rise through a CTX-sensitive pathway were confirmed by the increased surface detection of transfected Ca_V2.2 channels in α₂δ-1 overexpressing cultures.

The ability of the α₂δ-1 subunit to promote calcium channel expression at the cell surface is dependent on the MIDAS motif located in the VWA domain of α₂δ proteins, as also shown in a neuronal cell line (Cassidy et al., 2014). Mutation of this motif markedly decreased α₂δ-1 HA membrane expression, leading to a consequent reduction of Ca_V2.2 surface localization and complete inhibition of the prolonged Ca²⁺ response. Furthermore, the experiments performed with α₂δ-1 MIDAS^AAA HA showed that the modulation of Ca²⁺ responses is dependent on the control of VGCC trafficking by α₂δ-1, yet it provokes the question of the mechanism responsible for prolonged N-type VGCC-mediated Ca²⁺ transients.

Recently it has been postulated that intracellular Ca²⁺ signals triggered by different means can follow individual pathways coupled to specific cellular responses, such as the activation of gene transcription by Ca_V1 channels in sympathetic neurons (Wheeler et al., 2012). In the same neuronal model, using a Ca_V2.1 antibody, these VGCCs were found to be distributed in high-density patches in close proximity to ER–mitochondria interaction sites and it was postulated that Ca_V2-mediated Ca²⁺ influx was preferentially sequestered by mitochondria. ER stores and mitochondria constitute the main Ca²⁺ buffering compartments in DRG neurons (Medvedeva et al., 2008). In our study, the inhibition of SERCA in control neurons slowed the recovery from high-K⁺ evoked responses. This confirmed the role of the ER in Ca²⁺ clearance after depolarization. However, upon α₂δ-1 overexpression, ER stores failed to further prolong depolarization-induced Ca²⁺signals, despite contributing to resting Ca²⁺ levels, suggesting that the ER stores are full in this condition.

To investigate the role of mitochondria in the modulation of Ca²⁺ signaling by α₂δ-1, we selectively knocked down the mitochondrial Ca²⁺ uniporter mechanism through the expression of a mutated MCU protein, MCU^D260N,E263Q. Cotransfection of MCU^D260N,E263Q together with the α₂δ-1 HA subunit in DRGs abolished the elevation of the Ca²⁺ transient width and increased response peak amplitude, when compared with neurons overexpressing α₂δ-1 HA protein alone. Similar results were obtained when mitochondrial Ca²⁺ uptake was indirectly blocked by acute disruption of the mitochondrial membrane potential with antimycin and oligomycin treatment. Also with mtPericam imaging we were able to directly measure the increase in [Ca²⁺]mt evoked by 100 AP stimulation of α₂δ-1 HA-expressing neurons.

Altogether our genetic and pharmacological studies confirm that the activation of mitochondrial Ca²⁺ uptake is an essential mechanism called into play to limit the Ca²⁺ rise in response to Ca²⁺ entry (Friel, 2000), and define the α₂δ-1 protein as a crucial regulator of VGCC-mediated signaling in peripheral neurons. Because mitochondria are recruited to buffer high Ca²⁺ loads (Werth and Thayer, 1994), it is likely that their involvement in α₂δ-1-modulated signaling may be associated with an increase in the magnitude of Ca²⁺ response consequent on VGCC upregulation. Indeed, we found that stimulation of cultures with 100 mm K⁺ increased the duration of Ca²⁺ responses, as well as the peak amplitude. In this condition, we found that almost all responses in both α₂δ-1 HA overexpressing and control neurons were characterized by a large plateau phase, which followed the initial rise of the peak. However the inhibition of mitochondrial Ca²⁺ uptake by MCU^D260N,E263Q overexpression revealed that mitochondria contributed to blunting the Ca²⁺ response amplitude only in α₂δ-1 overexpressing DRGs, suggesting that even in the presence of high Ca²⁺ entry, α₂δ-1 HA overexpression displays a role in promoting mitochondrial Ca²⁺ buffering of N-type calcium channel-mediated responses, suggesting that they may be in proximity.

On the basis of data related to surface staining of Ca_V2.2 channels and lack of effect of the SERCA pump blocker in α₂δ-1 HA overexpressing DRGs, we can speculate that mitochondria may preferentially buffer Ca²⁺ in close proximity to the channels at the plasma membrane. Conversely, in the absence of the α₂δ-1-modulation of VGCC, mitochondria are more likely to be influenced by Ca²⁺ release from the ER. Previously it has been shown that mitochondria play an important role in the control of neurotransmission at presynaptic terminals both in capsaicin-sensitive DRGs and at central synapses (Billups and Forsythe, 2002; Medvedeva et al., 2008; Perkins et al., 2010). To regulate metabolic demand and the local intracellular Ca²⁺ concentration, mitochondria are rapidly trafficked through neuronal processes (Sheng and Cai, 2012). In agreement with the finding that mitochondrial stalling is Ca²⁺-dependent in neuronal processes (Macaskill et al., 2009), we found that mitochondrial axonal transport was selectively reduced after electrical stimulation of α₂δ-1 HA overexpressing cultures. In contrast, no change was observed in mitochondrial trafficking either in α₂δ-1 MIDAS^AAA HA-expressing DRGs or in CTX-treated α₂δ-1 HA overexpressing neurons. Our findings suggest that α₂δ-1 HA overexpression may contribute to the formation of domains predominantly buffered by mitochondria in the cell body, as well as the axons of DRGs.

In conclusion, this work describes the mechanism through which α₂δ-1 upregulation modulates the response of DRG neurons to depolarization, suggesting that N-type VGCC-mediated activation of mitochondrial Ca²⁺ buffering may contribute to intracellular signaling related to the aberrant neurotransmission in pathological conditions.