Identification of Ca_V channel types expressed in muscle afferent neurons

Abstract

Cardiovascular adjustments to exercise are partially mediated by group III/IV (small to medium) muscle afferents comprising the exercise pressor reflex (EPR). However, this reflex can be inappropriately activated in disease states (e.g., peripheral vascular disease), leading to increased risk of myocardial infarction. Here we investigate the voltage-dependent calcium (Ca_V) channels expressed in small to medium muscle afferent neurons as a first step toward determining their potential role in controlling the EPR. Using specific blockers and 5 mM Ba²⁺ as the charge carrier, we found the major calcium channel types to be Ca_V2.2 (N-type) > Ca_V2.1 (P/Q-type) > Ca_V1.2 (L-type). Surprisingly, the Ca_V2.3 channel (R-type) blocker SNX482 was without effect. However, R-type currents are more prominent when recorded in Ca²⁺ (Liang and Elmslie 2001). We reexamined the channel types using 10 mM Ca²⁺ as the charge carrier, but results were similar to those in Ba²⁺. SNX482 was without effect even though ∼27% of the current was blocker insensitive. Using multiple methods, we demonstrate that Ca_V2.3 channels are functionally expressed in muscle afferent neurons. Finally, ATP is an important modulator of the EPR, and we examined the effect on Ca_V currents. ATP reduced Ca_V current primarily via G protein βγ-mediated inhibition of Ca_V2.2 channels. We conclude that small to medium muscle afferent neurons primarily express Ca_V2.2 > Ca_V2.1 ≥ Ca_V2.3 > Ca_V1.2 channels. As with chronic pain, Ca_V2.2 channel blockers may be useful in controlling inappropriate activation of the EPR.

the group i (aα) and ii (Aβ) muscle afferents provide sensory information needed to guide motor activity (Houk 1974). The group III (Aδ) and IV (C) afferents transmit muscle pain signals and also mediate the exercise pressor reflex (EPR) (Kaufman and Hayes 2002), which is a critical neural mechanism that regulates the cardiovascular response to exercise (Kaufman and Hayes 2002). This reflex is clinically important because certain diseases, such as peripheral vascular disease and heart failure, can produce muscle ischemia that drives EPR activity, which can generate additional cardiac stress to increase the risk of myocardial infarction (Baccelli et al. 1999; Bakke et al. 2007; Smith et al. 2006). Here we investigate the voltage-dependent calcium (Ca_V) channels expressed in muscle afferent neurons as an initial step in understanding the role of these channels in controlling excitability of neurons that mediate the EPR.

Ca_V channels play a prominent role in neuronal excitability (Khosravani and Zamponi 2006). At synaptic terminals, they deliver the Ca²⁺ needed to induce the release of excitatory and inhibitory neurotransmitters (Lisman et al. 2007). There are 10 genes that encode Ca_V channels (Catterall et al. 2005). In dorsal root ganglia (DRG) neurons, the evidence supports the expression of L-type (Ca_V1.2), P/Q-type (Ca_V2.1), N-type (Ca_V2.2), R-type (Ca_V2.3, also called E-type or α_1E), and T-type (Ca_V3.2) (Vanegas and Schaible 2000; Zamponi et al. 2009). The importance of Ca_V2.2 channels in chronic pain has been highlighted by the success of the specific blocker ziconotide (SNX-111 or ω-conotoxin MVIIA) in the treatment of chronic pain in many patients (Elmslie 2004; Molinski et al. 2009). This intrathecally delivered drug blocks transmission from primary to secondary nociceptors by blocking presynaptic Ca_V2.2 channels (Motin and Adams 2008). While the effect of blocking presynaptic Ca_V channels is to reduce central nervous system excitability, blocking Ca_V channels can also increase neuronal excitability through reduced activation of closely associated Ca²⁺-activated potassium channels (Marrion and Tavalin 1998; Yu et al. 2010). Indeed, blocking Ca_V channels increases excitability of sensory neurons (Lirk et al. 2008). Thus it is important that we understand the Ca_V channels expressed in muscle afferent neurons, along with the modulation of these channels by molecular activators of the EPR. ATP is one of these activators, which works by acting on P_2X receptors (Cui et al. 2011; Hayes et al. 2008). However, ATP can inhibit Ca_V channels by activation of G protein-coupled P_2Y receptors (Filippov et al. 2003; Gerevich et al. 2004). Here we identify the Ca_V channels that are functionally expressed by muscle afferent neurons and show that ATP can inhibit the Ca_V current by reducing activity of Ca_V2.2 (N-type) channels. As with chronic pain, we show that Ca_V2.2 channels could be an important target for the treatment of symptoms resulting from excessive EPR activity.

MATERIALS AND METHODS

Isolation of muscle afferent neurons.

Adult male Sprague-Dawley rats (weight 150–400 g) were obtained from Hill Top Laboratories (Scottdale, PA). All procedures were reviewed and approved by the Institutional Animal Care and Use Committee and followed National Institutes of Health (NIH) guidelines. The labeling and neuronal isolation followed previously described procedures (Ramachandra et al. 2012). Briefly, muscle afferent neurons were labeled by retrograde transport of the lipophilic dye DiI (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate) [100 μl of 1.5% DiI in dimethyl sulfoxide (DMSO)] that was injected into both left and right gastrocnemius muscles of anesthetized rats (ketamine, xylazine, and acepromazine). Four to five days postinjection, the rats were killed with CO₂ followed by decapitation. Neurons from the lumbar DRG L₄ and L₅ were isolated using an enzyme mixture of trypsin, collagenase, and DNAse and plated onto polylysine-coated glass coverslips. The isolated neurons were maintained overnight at 37°C in a 5% CO₂ incubator in minimal essential medium with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin.

Expressed Ca_V channels.

Ca_V2.3 channels (α1, β2A, and α2δ cDNA constructs, generous gifts from Dr. Henry L. Puhl, National Institutes of Health/National Institute on Alcohol Abuse and Alcoholism) were heterologously expressed in LN229 cells by intranuclear microinjection at a ratio of 1:1:2. The cells were cultured in Dulbecco's modified Eagle's medium, supplemented with 10% FBS, 1% glutamine, and 1% penicillin-streptomycin. Following cDNA injection, the cells were incubated overnight, and electrophysiological recordings were performed the next day.

Solutions.

Most DiI-positive DRG neurons were initially recorded in an external solution containing (in mM) 45 NaCl, 100 N-methyl d-glucosamine (NMG)·Cl, 4 MnCl₂, 10 Na·HEPES, 10 glucose, and 0.0003 tetrodotoxin (TTX), with pH = 7.4 and osmolarity = 320 mosM, which was used to identify voltage-gated sodium (Na_V) 1.8 expressing muscle afferent neurons (Ramachandra et al. 2012). When recording Ca_V current, the external solution was switched to a barium external solution containing (in mM) 145 NMG·Cl, 5 BaCl₂, 10 NMG·HEPES, and 5 glucose, with pH = 7.4 and osmolarity = 320 mosM. The pipet solution contained (in mM) 104 NMG·Cl, 14 creatine·PO₄, 6 MgCl₂, 10 NMG·HEPES, 5 Tris·ATP, 10 NMG₂·EGTA, and 0.3 Tris₂·GTP, with pH 7.4 and osmolarity = 300 mosM.

For some experiments, the Ca_V currents in muscle afferent neurons were recorded in external Ca²⁺. Na_V1.8 expressing muscle afferents were identified by expression of enhanced green fluorescent protein (Puhl and Ikeda 2008), as previously described (Hassan and Ruiz-Velasco 2013). Briefly, DRG neurons were microinjected with a cDNA plasmid containing the enhanced green fluorescent protein gene under control of the putative Na_V1.8 promoter. The external solution for these experiments contained (in mM) 145 tetraethanolamine (TEA)·CH₃SO₃H, 10 TEA·HEPES, 10 CaCl₂, 15 glucose, and 0.0003 TTX, pH 7.4, 325 mosM. The internal solution contained (in mM) 90 NMG·CH₃SO₃H, 25 TEA·CH₃SO₃H, 14 creatine·PO₄, 11 EGTA, 10 HEPES, 1 CaCl₂, 20 CsOH, 4 MgATP, 0.3 Na₂GTP, pH 7.2, 300 mosM. The expressed Ca_V2.3 channels were recorded in a solution containing (in mM) 140 NaCl, 5.4 KCl, 10 CaCl₂, 1 MgCl₂, 10 glucose, and 10 HEPES, pH 7.4 with NaOH. The internal solution was the same as that used for muscle afferent neurons recorded in Ca²⁺.

Ion channel blockers were made up as stock solutions in either water or DMSO and diluted into the external solution to make the working concentration. All external solutions contained the same DMSO concentration (maximum 0.03%) so that the only change was presence or absence of the blocker. In experiments where Ca²⁺ was the charge carrier, 0.1 mg/ml cytochrome c was added to all external solutions to minimize potential binding of the peptide toxins to the capillary columns used for drug delivery. Solutions were applied using a gravity fed perfusion system with a solution exchange time of 2 s.

Measurement of ionic currents.

DiI-labeled DRG neurons were identified using a Nikon Diaphot microscope with epifluorescence and voltage-clamped using the whole cell configuration of the patch-clamp technique. Pipettes were pulled from glass capillaries (King Precision Glass, Claremont, CA) on a Sutter P-97 puller (Sutter Instruments, Novato, CA). Currents were recorded using either an Axopatch 200A or 200B amplifier (Molecular Devices, Sunnyvale, CA) and digitized with an ITC-18 data acquisition interface (Instrutech, Port Washington, NY). Experiments were controlled using S5 data acquisition software written by Dr. Stephen Ikeda (NIH/National Institute on Alcohol Abuse and Alcoholism, Rockville, MD). Leak current was subtracted online using a P/4 protocol. Recordings were carried out at room temperature, and the holding potential was −80 mV.

Data analysis.

Data were analyzed using IgorPro (WaveMetrics, Lake Oswego, OR) running on a Macintosh computer. Cell diameter was calculated using the cell capacitance as previously described (Ramachandra et al. 2012). Group data were calculated as means ± SD throughout the paper. Student's T-test (unpaired, two-tailed) was calculated to determine significant differences (P < 0.05).

Immunocytochemistry.

Neurons were fixed with 4% formaldehyde and permeabilized with 2% Tween 20, as previously described (Ramachandra et al. 2012). Neurons were labeled with primary antibodies for both Na_V1.8 (mouse, Abcam, Cambridge, MA) and Ca_V2.3 (rabbit, Alomone Laboratories, Jerusalem, Israel) (1:500) and visualized using secondary antibodies Alexa Fluor 488 IgG goat anti-rabbit and Alexa Fluor 635 IgG goat anti-mouse (Invitrogen, Carlsbad, CA). Images were captured using a Nikon ECLIPSE 80i epifluorescence microscope, and neurons were measured using ImageJ (rsbweb.nih.gov/ij/index.html). Cell size was calculated and positive fluorescent labeling was determined, as described previously (Ramachandra et al. 2012).

Chemicals.

DiI, minimal essential medium, Dulbecco's modified Eagle's medium, FBS, and penicillin/streptomycin were obtained from Invitrogen. TTX citrate, nifedipine (Nif), and SNX482 (SNX) were obtained from Ascent Scientific (Princeton, NJ). ω-Conotoxin GVIA (GVIA) and ω-agatoxin IVa (AgaIVa) were obtained from Bachem America (King of Prussia, PA). All other chemicals were from Sigma (St. Louis, MO).

RESULTS

We were interested in determining the Ca_V channels that are functionally expressed in muscle afferent neurons. Consistent with previous work, we used pharmacology to determine the percentage of total Ca_V current generated by each channel type. Ca_V2.2 (N-type) was determined from block by 10 μM GVIA, Ca_V2.1 (P/Q-type) from block by 0.2 μM AgaIVa, Ca_V1.2 from block by 3 μM Nif and Ca_V2.3 (R-type) from block by 0.3 μM SNX (Fuchs et al. 2007; Huang et al. 1997; Ikeda and Matsumoto 2003; Lu et al. 2010; Ohnami et al. 2011). Using our 5 mM Ba²⁺ external solution, we found that the largest block was produced by GVIA (47 ± 19%, n = 20) > AgaIVa (25 ± 14%, n = 16) > Nif (13 ± 14%, n = 16) (Fig. 1). The block produced by each of these three drugs was significant (P < 0.05). However, SNX produced no significant block (−7 ± 11%, n = 6). This was surprising since previous work had shown SNX-sensitive currents in DRG neurons (Fang et al. 2007; Fuchs et al. 2007). Perhaps muscle afferent neurons fail to express Ca_V2.3 channels. However, 22 ± 16% (P < 0.05, n = 11) of the total Ca_V current was resistant to all blockers (resistant current). As expected, only block by GVIA was irreversible (Boland et al. 1994; Liang and Elmslie 2002).

Fig. 1.

Voltage-dependent calcium (Ca_V) 2.2 channels generate the dominant Ca_V current in muscle afferent neurons. Ca_V current in 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI) labeled muscle afferent neurons was measured using 5 mM Ba²⁺ external solution. A: example currents showing the effect of 0.3 μM SNX482 (SNX), 0.2 μM ω-agatoxin IVa (AgaIVa), 3 μM nifedipine (Nif), and 10 μM ω-conotoxin GVIA (GVIA). The voltage protocol is shown below the current traces. Cntl, control. B: comparison of mean ± SD calcium current block in small (S; 20–30 μm) vs. medium (M; 30–40 μm) muscle afferent neurons by the indicated blockers. The blocker concentrations are the same as indicated for A. C–E: the distribution of percentage block vs. neuron diameter for 3 μM Nif (C), 0.2 μM AgaIVa (D), and 10 μM GVIA (E). Note the different y-axis scale for GVIA. The dashed line indicates average block.

Previous work has shown differences in the percentage Ca_V current blocked by various specific blockers in small vs. medium or large DRG neurons (Fuchs et al. 2007; Scroggs and Fox 1992). Thus it appears that expression of Ca_V channel types can vary across classes of DRG neurons. Unmyelinated C-type and thinly myelinated Aδ-type cutaneous afferents have cell body diameters < 35 μm, with C fibers having diameters < 23 μm (Djouhri et al. 2003). However, muscle afferents were found to have larger soma diameters relative to cutaneous afferents (Hu and McLachlan 2003; Ramachandra et al. 2012). Thus we include muscle afferent neurons up to 40 μm in the group III (Aδ) and define small group IV (C) afferent neurons as those with cell diameters < 30 μm (Ramachandra et al. 2012, 2013). While the Ca_V current was significantly blocked by Nif, AgaIVa, and GVIA (but not SNX) in both groups, we found no differences in percentage block by any Ca_V channel blocker between small and medium neurons (Fig. 1B), which is consistent with previous findings from cutaneous afferents (Lu et al. 2010). In addition, the resistant current was not different between these two groups at 21 ± 12% (n = 7) of total Ca_V current in small neurons and 25 ± 24% (n = 4) in medium neurons. A more detailed examination of the Ca_V channel block data is shown using scatter plots (Fig. 1, C–E), which shows that there are no systematic differences among muscle afferent neurons ranging from 20 to 40 μm diameter. Thus neurons in both the group III (Aδ) and group IV (C) size range most prominently express Ca_V2.2 > Ca_V2.1 > Ca_V1.2 (Fig. 1).

The functional expression of Ca_V3 (T-type) channels in DRG neurons has been demonstrated (Ikeda and Matsumoto 2003; Jagodic et al. 2008; Scroggs and Fox 1992), but we failed to observe evidence of Ca_V3 currents in our recordings in Ba²⁺. Such currents are characterized by low-voltage activation (less than −30 mV) and rapid inactivation (Scroggs and Fox 1992). To quantitatively determine whether Ca_V3 currents contributed to the total Ca_V current, we calculated the percentage inhibition over a range of voltages for each of the three blockers that produced significant inhibitions. Ca_V3 currents would be revealed by the relatively smaller block at hyperpolarized voltages where Ca_V3 channels dominate (Ikeda and Matsumoto 2003; Jagodic et al. 2008; Scroggs and Fox 1992). Using our standard holding potential (−80 mV), we found no significant difference in the inhibition at −30 vs. +30 mV produced by any of our blockers (Fig. 2). It was not possible to do this analysis using more hyperpolarized voltages, since −30 mV was the most hyperpolarized voltage to produce measureable current. Under these conditions, it appears that Ca_V3 channels do not significantly contribute to total Ca_V current in muscle afferent neurons.

Fig. 2.

Voltage-independent block by Ca_V channel blockers. A: current-voltage relationships from a muscle afferent neuron in the presence of either 3 μM Nif, 0.2 μM AgaIVa, or 10 μM GVIA. Ca_V currents were measured in 5 mM Ba²⁺ at the end of 25-ms voltage steps to the indicated voltage. B: the mean inhibition (±SD) for each Ca_V channel blocker is shown over voltages ranging from −30 to 30 mV. The currents were measured as described for A. There was no statistical difference between the block at −30 mV vs. that at 30 mV for each blocker (n = 5 for each blocker).

Ca_V2.1 (P/Q) currents are typically blocked using submicromolar concentrations of AgaIVa (Fuchs et al. 2007; Huang et al. 1997), since Ca_V2.2 channels have been shown to be blocked by 1 μM AgaIVa (Sidach and Mintz 2000). However, Q-currents are a Ca_V2.1 isoform that is less sensitive to AgaIVa block and is typically blocked using micromolar AgaIVa concentrations (Randall and Tsien 1995; Sidach and Mintz 2000). We wondered if a component of our resistant current (22% of total Ca_V current) was contributed by Q-channels. Therefore, we compared block of muscle afferent Ca_V current in 0.2 vs. 1 μM AgaIVa after block by GVIA to remove the Ca_V2.2 component (Fig. 3). In these neurons, GVIA blocked 50 ± 2% of current, and 0.2 μM AgaIVa blocked 26 ± 9% of current, but increasing AgaIVa to 1 μM blocked only an additional 4 ± 1% of current. Thus a weakly AgaIVa-sensitive Ca_V2.1 channel type (e.g., Q-current) does not substantially contribute to resistant current in muscle afferent neurons.

Fig. 3.

Little or no Q-current in muscle afferent neurons. Ca_V current from muscle afferent neurons was recorded in 5 mM Ba²⁺ external solution. A: example currents from a muscle afferent neuron in Cntl, 10 μM GVIA, 0.2 μM AgaIVa, and 1 μM AgaIVa. B: the mean (±SD) percentage block by GVIA and the two concentrations of AgaIVa tested on the same neurons. The number of cells tested is indicated in the GVIA bar. *Ca_V current was significantly blocked.

The absence of SNX-sensitive Ca_V current was puzzling. We wondered if our use of Ba²⁺ as the charge carrier was the problem, since it was previously demonstrated that R-like Ca_V current was enhanced when Ca²⁺ was the charge carrier (Boland et al. 1994; Liang and Elmslie 2001). Using the same blockers, we redid our study using a 10 mM Ca²⁺ external solution. In addition, we used a cumulative blocker application strategy to better gauge the resistant current (Fig. 4). It should be noted that, under these recording conditions (HP −80 mV), we observed T-type currents, but only in ∼10% of muscle afferent neurons. The percentage block by Nif, GVIA, and SNX was statistically similar (P > 0.05) to that in Ba²⁺, and the resistant current was also similar between Ca²⁺ and Ba²⁺ (P > 0.05). However, the block by AgaIVa was smaller in Ca²⁺ (13 ± 2%) vs. Ba²⁺ (25 ± 14%). We do not understand the reasons for the smaller AgaIVa block in Ca²⁺, but it is clear that Ca²⁺ did not enhance SNX-sensitive current, as we predicted.

Fig. 4.

Ca_V current block in external Ca²⁺ is similar to that in Ba²⁺. A: an example time course for Ca_V current block in 10 mM external Ca²⁺. This experiment utilized cumulative blocker application to better determine the size of the resistant current (Resistant). The blocker concentrations are indicated and were the same as those used when recording in external Ba²⁺. B: the average (±SD) percentage block by each Ca_V channel blocker. *Ca_V current was significantly blocked. The number of cells tested is indicated in each bar.

As a positive control for SNX, we tested 0.3 μM SNX on Ca_V2.3 channels heterologously expressed in LN229 cells. The recordings were done using a 10 mM Ca²⁺ external solution, and SNX blocked 93 ± 8% (n = 3) of the expressed current. Thus two possible explanations for the absence of a SNX effect on muscle afferent neurons are that Ca_V2.3 channels are not expressed by these neurons, or the expressed Ca_V2.3 channels are insensitive to SNX.

Ca_V2.3 in muscle afferent neurons.

SNX has been used in many experiments to define Ca_V2.3 current, but not all Ca_V2.3 channels are sensitive to SNX (Tottene et al. 2000). To determine whether Ca_V2.3 channels were expressed in muscle afferent neurons, we exposed isolated DRG neurons to a rabbit Ca_V2.3 antibody. For consistency with our electrophysiological recordings, we also tested for the presence of Na_V1.8 using a specific mouse antibody. We found that all muscle afferent neurons were positively labeled by the Ca_V2.3 antibody (Fig. 5, A and B), and all but one of these neurons were also labeled by the Na_V1.8 antibody (Fig. 5C). Most of the identified muscle afferent neurons had soma diameters between 20 and 40 μm, which overlaps with the electrophysiological recordings. Thus muscle afferent neurons appear to express Ca_V2.3 channels.

Fig. 5.

Immunocytochemistry shows Ca_V2.3 channels expressed in muscle afferent neurons. A: the upper four images show sensory neurons stained with both voltage-gated sodium (Na_V) 1.8 and Ca_V2.3 antibodies. The bright-field (left) image shows the three neurons, and the DiI panel (right) shows the one labeled muscle afferent neuron in this field. All three neurons were positive for both Na_V1.8 and Ca_V2.3. The white bar in the bright-field image indicates 50 μm. The lower four images show Na_V1.8 and Ca_V2.3 controls along with bright-field (left) image and the DiI image showing one labeled muscle afferent neuron. B: there is no correlation between Ca_V2.3 label intensity and muscle afferent neuron size. The threshold line was determined from measurement of Cntl muscle afferent neurons (e.g., lower images in A). The units for the Y-axis are arbitrary fluorescence units. C: comparison of Ca_V2.3 label intensity vs. that of Na_V1.8 in muscle afferent neurons. The threshold lines were determined from Cntl muscle afferent neurons. Only one Ca_V2.3-positive muscle afferent neuron was negative for Na_V1.8.

We also wanted to test if these channels were functional, but our blocker (SNX) did not work. One strategy was to use roscovitine, which is a Ca_V2-specific agonist (Buraei et al. 2005, 2007). While roscovitine affects many ion channels (Buraei et al. 2007; Ganapathi et al. 2009; Yarotskyy and Elmslie 2007, 2012), it uniquely affects Ca_V2 channels by slowing deactivation (Buraei et al. 2005, 2007), which produces slow tail currents (Fig. 6). For these experiments, we tested the effect of 100 μM roscovitine after blocking Ca_V2.1 and Ca_V2.2 with 0.2 μM AgaIVa and 10 μM GVIA, respectively. Figure 6A shows the effect of roscovitine to inhibit step current and slow deactivation of total Ca_V current (before block of Ca_V2.1 and 2.2). After toxin application, the tail current is still reversibly slowed by roscovitine (Fig. 6, B and C). The roscovitine-induced inhibition was similar before and after applications of the toxins, with average reductions of 39 ± 15% and 52 ± 18% for control and in toxins, respectively. The slowed deactivation shows that functional Ca_V2 channels produce current in the presence of AgaIVa and GVIA. The similar inhibition before and after toxin application suggests that roscovitine did not reverse toxin block to reveal Ca_V2.2 or 2.1 channels, which would have been caused either by an increase in step current or a decrease in inhibition. The most likely conclusion is that SNX-resistant Ca_V2.3 channels comprise a large fraction of the resistant Ca_V current in small to medium muscle afferent neurons.

Fig. 6.

Roscovitine (Rosc)-induced slowed deactivation reveals functional Ca_V2 channels in the presence of GVIA and AgaIVa. All currents were recorded in 5 mM Ba²⁺. A: a muscle afferent neuron shows the effect of Rosc on Ca_V2 currents (no toxins present). Rosc (100 μM) slowed deactivation (black trace) compared with Cntl and recovery (Recov; gray traces). B: current traces from the same muscle afferent neuron as shown in A recorded in 10 μM GVIA and 0.2 μM AgaIVa. Rosc (100 μM) (black trace) slowed deactivation compared with Cntl (Toxin) and Recov (gray traces). C: a single exponential equation was fit to the deactivating currents at −40 mV to determine the deactivation τ in presence of toxin (GVIA and AgaIVa) and toxin + 100 μM Rosc. The average deactivation τ (±SD) is shown. *Significant slowing of deactivation induced by Rosc. The number of muscle afferent neurons tested is indicated in the middle bar.

As a second test, we examined the Ni²⁺ sensitivity of the resistant current. It has been demonstrated in major pelvic ganglion neurons that SNX-resistant R-current is blocked by Ni²⁺ with an IC₅₀ (22 μM) similar to that of expressed Ca_V2.3 channels (21 μM) (Won et al. 2006). Using Ba²⁺ as the charge carrier, we tested the dose response for Ni²⁺ block of total Ca_V current and calculated an IC₅₀ = 246 μM (n = 3–4 neurons), which is expected for Ca_V currents dominated by Ca_V2.2 and 2.1 channels (Liang and Elmslie 2001; Zamponi et al. 1996). However, following application of 1 μM GVIA to block the dominant Ca_V2.2 current, the Ni²⁺ block IC₅₀ did not change with IC₅₀ = 319 μM (n = 7–8 neurons), which was against our expectation of a decrease in IC₅₀. It has been previously reported that the Ni²⁺ sensitivity of Ca_V2.3 current was ∼300 μM when recorded in 10 mM Ba²⁺, but < 30 μM when recorded in 10 mM Ca²⁺ (Zamponi et al. 1996). In addition, the Ni²⁺-sensitive R-current of major pelvic ganglion neurons was recorded in 10 mM Ca²⁺ (Won et al. 2006). Therefore, we reexamined the Ni²⁺ block of Ca_V currents recorded in 10 mM Ca²⁺. We found that the Ni²⁺ block of the resistant current (Nif, GVIA, and AgaIVa) recorded in Ca²⁺ was much more potent than when recorded in Ba²⁺ with IC₅₀ = 4.4 μM (n = 2 to 5) and the maximum block = 64%. The potent Ni²⁺ block suggests that the majority of the resistant current (64%) was generated by the activity of Ca_V2.3 channels.

ATP modulates Ca_V2.2 channels.

ATP helps to generate the EPR by activating P_2X ionotropic receptors (Cui et al. 2011; Hayes et al. 2008), but ATP modulation of the reflex is also possible by activation of G protein-coupled P_2Y receptors. ATP has been shown to inhibit Ca_V currents by activation of P_2Y receptors (Filippov et al. 2003; Gerevich et al. 2004), so we tested the effect on muscle afferent Ca_V currents (Fig. 7). G protein-mediated inhibition of Ca_V2 currents is often found to be voltage dependent in that the inhibition is temporarily reversed by brief, strong depolarization (Elmslie et al. 1990; Ikeda 1991). ATP (10 μM) produced an average 32 ± 15% (n = 13) (Fig. 7B) inhibition of Ca_V current that was reversed following strong depolarization (Fig. 7A). In addition, there was no clear difference in the percentage inhibition observed among neurons with diameters ranging from 22 to 48 μm (Fig. 7B). Previous reports have demonstrated Ca_V2.2 (N-type) to be a major target of G protein-mediated inhibition (Elmslie et al. 1990; Gerevich et al. 2004; Ikeda 1991), but other Ca_V2 channels can be modulated as well (Colecraft et al. 2000; Meza and Adams 1998). Thus we determined the effect of GVIA on ATP modulation (Fig. 7D). Prior to GVIA application, ATP inhibited current (prepulse) by an average of 19 ± 8%, while the inhibition was reduced to an average of 3 ± 4% following GVIA block of Ca_V2.2 channels in the same five neurons. Ca_V2.2 channels are the primary Ca_V channel target for ATP-induced inhibition in muscle afferent neurons.

Fig. 7.

ATP inhibits Ca_V2.2 channels in muscle afferent neurons. All currents were recorded in 10 mM Ca²⁺. A: the inhibition of Ca_V current induced by 10 μM ATP (black trace) compared with Cntl (gray trace) recorded from a muscle afferent neuron. The inhibition is transiently reversed by strong depolarization (+80 mV), which can be seen by comparing the prepulse (before the +80-mV step) and postpulse (following the +80-mV step) currents. B: there was no clear differences in ATP (10 μM) induced inhibition in small (<30 μm) vs. medium (30–40 μm) vs. large (>40 μm) muscle afferent neurons. The percent inhibition of prepulse current measured from 13 muscle afferent neurons is plotted vs. neuron diameter. C: the ATP (10 μM) inhibition is blocked by preapplication of 10 μM GVIA. This time course shows the inhibition induced by ATP prior to GVIA application and little or no ATP response in the presence of GVIA. The prepulse (solid circle) and postpulse (open circle) current amplitudes are plotted. D: the average (±SD) inhibition induced by 10 μM ATP is shown before (ATP) and during (GVIA + ATP) application of 10 μM GVIA. *ATP response in GVIA is significantly different from that in Cntl. The number of muscle afferent neurons tested is indicated.

DISCUSSION

We have identified the Ca_V channels that generate Ca_V current in muscle afferent neurons along with examining the modulation of that current by ATP. The results demonstrate that, regardless of using external Ba²⁺ or Ca²⁺, Ca_V2.2 channels dominate by generating 40–50% of the total Ca_V current. Ca_V2.1 channels generated between 13 and 25% of the current, while Ca_V1.2 channels generated ∼15% of the total current. The Ca_V2.3 channel blocker SNX had no effect on muscle afferent Ca_V current, while the portion of the current that was resistant to all blockers averaged ∼25% of the total current. Using roscovitine and Ni²⁺, we demonstrated that the majority of resistant current is generated by Ca_V2.3, which shows that the current generated by this channel type is roughly equal to that generated by Ca_V2.1. In the small to medium neurons recorded in our study, there was no difference in Ca_V channel expression between these two groups. ATP inhibited Ca_V current by activation of G protein-coupled receptors, and that inhibition primarily targeted Ca_V2.2 channels. Based on the neuronal size and the expression of Na_V1.8 by these neurons (Ramachandra et al. 2012), we conclude that group IV (C) and group III (Aδ) neurons primarily express Ca_V2.2 channels that are inhibited by extracellular ATP.

Ca_V channels in sensory neurons.

The percentage of Ca_V channel types comprising the total current in our work matches well with most studies on small- and medium-diameter sensory neurons with Ca_V2.2 > Ca_V2.1 = Ca_V2.3 > Ca_V1.2 (Huang et al. 1997; Lu et al. 2010). However, Scroggs and Fox (1992) found a much larger block by 2 μM nimodipine (53%) in small sensory neurons (20–27 μm) and a much smaller block (7%) in medium-sized neurons (33–38 μm) than we and others have observed. The source of this difference is unclear, but it is possible that there is a population of DRG neurons in which Ca_V1.2 channels dominate.

One important issue when using blockers to identify component channels is the specificity and concentration of the chosen compounds. The blockers used in this study were the same as used in many previous studies. GVIA is a highly specific blocker of Ca_V2.2 channels, but the concentration we used was higher than some other studies (Fuchs et al. 2007; Huang et al. 1997; Lu et al. 2010; Scroggs and Fox 1992). Early studies of Ca_V2.3 showed this channel was reversibly blocked by 5 μM GVIA when expressed in Xenopus oocytes (Zhang et al. 1993), but we did not observe a reversible component to the GVIA block in our studies. In addition, GVIA has been shown to be specific for Ca_V2.2 in neurons at concentrations up to 100 μM (Boland et al. 1994; Elmslie 1997). Finally, the percentage Ca_V current blocked by GVIA in our study was similar to that in studies using lower concentrations (Huang et al. 1997; Lu et al. 2010). The higher GVIA concentration in our study was used to ensure we could obtain data from multiple blockers applied to the same neuron. GVIA is a very high-affinity blocker, but the access to the blocking site is highly sensitive to divalent cation concentration (Boland et al. 1994; Liang and Elmslie 2002; Wagner et al. 1988). Thus the blocking rate is much lower in higher divalent cation concentrations (Boland et al. 1994; Liang and Elmslie 2002; Wagner et al. 1988) and also much slower in Ca²⁺ vs. Ba²⁺ (Liang and Elmslie 2001). Thus our data collection when using 10 mM Ca²⁺ was greatly enhanced by the fast block achieved using a higher GVIA concentration. Unlike GVIA, there is evidence that Nif and AgaIVa lose selectivity at higher concentrations. Nif has been shown to block non-L-type Ca_V currents at 10 μM (Jones and Jacobs 1990), which was one reason for our use of 3 μM. We were also able to achieve fast block of L-type channels using this Nif concentration. AgaIVa is highly specific for Ca_V2.1 channels at submicromolar concentrations, but Ca_V2.2 channels are blocked at concentrations ≥ 1 μM (Sidach and Mintz 2000). Thus, like other studies, we utilized 0.2 μM AgaIVa for most of our experiments. However, the Q-type variant of Ca_V2.1 is less sensitive to AgaIVa than the P-type variant (Bourinet et al. 1999; Randall and Tsien 1995). Thus, when determining if lower affinity Ca_V2.1 channel variants were expressed in muscle afferents, we preblocked Ca_V2.2 channels with GVIA, which allowed us to show that only 4% of the total Ca_V current could be attributed to Q-like Ca_V2.1 channels. Since resistant current comprises ∼25% of Ca_V current, our results suggest that Q-like Ca_V2.1 channels generate ∼16% (4%/25%) of resistant current.

Current that is resistant to the classic Ca_V channel blockers (e.g., GVIA, AgaIVa, and Nif) was originally termed R-current (Randall and Tsien 1995). The identification of Ca_V2.3 (α1_E) fit the characteristics of R-current, and this channel type was called R-type (Zhang et al. 1993). SNX was identified as a blocker of Ca_V2.3 channels, but only a subpopulation of those channels was sensitive to this blocker (Tottene et al. 2000). We found that 22–27% of our Ca_V current was resistant to the classic Ca_V channel blockers, but that current was also insensitive to SNX. Two obvious possibilities were that Ca_V2.3 channels were not expressed in muscle afferent neurons, or SNX-insensitive variants were expressed in these neurons. We conclude that Ca_V2.3 channels are expressed in muscle afferent neurons, but these channels are insensitive to SNX, regardless of whether Ca²⁺ or Ba²⁺ was used to record current. This conclusion is based on three lines of evidence. First, muscle afferent neurons were labeled by a Ca_V2.3 antibody. Second, the deactivation of Ca_V current in the presence of GVIA and AgaIVa was slowed by roscovitine. GVIA and AgaIVa block Ca_V2.2 and 2.1, respectively, and roscovitine has been shown to specifically slow deactivation of Ca_V2 channels, including Ca_V2.3 channels (Buraei et al. 2007). A small component (∼16%) of this effect can be attributed to the Q-like current, but the robust roscovitine response supports expression of Ca_V2.3 in muscle afferent neurons. Finally, the potent Ni²⁺ block of the resistant current recorded in 10 mM Ca²⁺ (IC₅₀ = 4.4 μM) is consistent with the majority (64%) of that current being generated by the activity of Ca_V2.3 channels (Won et al. 2006). The difference in Ni²⁺ blocking affinity in Ca²⁺ vs. Ba²⁺ external solutions is also consistent with the current being generated by Ca_V2.3 channels (Zamponi et al. 1996).

ATP modulation of Ca_V2.2 channels.

ATP is an important regulator of the EPR (Cui et al. 2011; Hanna and Kaufman 2003; Hayes et al. 2008) primarily via activation of ionotropic P_2X receptors (Cui et al. 2011; Hanna and Kaufman 2004; Hayes et al. 2008). While it is clear that activation of G protein-coupled P_2Y receptors cannot generate the EPR (Hayes et al. 2008), it is possible that activation of these receptors could modulate the reflex. One potential mechanism for that modulation is via Ca_V channels. ATP has been shown to inhibit Ca_V2.2 channels by activation of P_2Y receptors (Filippov et al. 2003; Gerevich et al. 2004). Since the majority of neurons recorded for this study had small to medium diameters and all expressed Na_V1.8, these neurons are likely group III and IV neurons that participate in the EPR (Kaufman and Hayes 2002). Thus we were interested in the effect of ATP on Ca_V current in our muscle afferent neurons. We found that ATP inhibited Ca_V current, and this inhibition primarily involved Ca_V2.2 channels. While we did not specifically test the involvement of P_2Y receptors, there is no doubt that the inhibition is mediated via activation of G protein-coupled receptors. G protein-mediated inhibition of Ca_V2 channels is characterized by rapid reversal following strong depolarization (Elmslie 1992; Ikeda 1991; Meza and Adams 1998), which was exhibited by the ATP-induced inhibition of Ca_V current in our study. In addition, this type of inhibition is specifically mediated by the G protein βγ-subunit (Ikeda 1996). Thus ATP activation of G protein-coupled receptors (likely P_2Y) inhibits Ca_V2.2 channels in muscle afferent neurons by G protein βγ-subunits binding directly to the channels (Elmslie and Jones 1994; Zamponi and Snutch 1998).

Ca_V2.2 channels are one of the primary Ca_V channel types that trigger excitatory neurotransmitter release from nociceptor synaptic terminals in the dorsal horn (Elmslie 2004; Vanegas and Schaible 2000; Zamponi et al. 2009), which is the reason for the clinical effectiveness of the Ca_V2.2 channel blocker ziconotide in controlling pain in chronic pain patients (Elmslie 2004; Snutch 2005; Zamponi et al. 2009). Activation of P_2Y receptors in the dorsal horn of the spinal cord was shown to inhibit nociceptor synaptic transmission by inhibition of Ca_V2.2 (N-type channels), which was associated with reduced pain responses (Gerevich et al. 2004). Thus activation of P_2Y receptors on group III and/or IV muscle afferent synaptic terminals could inhibit Ca_V2.2 channel activity to reduce cardiovascular effects of the EPR. These results also suggest that intrathecal application of ziconotide could be used to treat excessive EPR activity resulting from peripheral vascular disease or heart failure (Baccelli et al. 1999; Bakke et al. 2007; Smith et al. 2006).

Ca_V channel control of neuronal excitability.

Ca_V channel activity can produce either an increase or decrease in neuronal excitability, depending on the Ca²⁺-sensitive proteins activated by the Ca²⁺ influx. As discussed above, inhibition of Ca_V channels in excitatory presynaptic terminals (e.g., primary sensory neurons) inhibits excitability. On the other hand, inhibition of Ca_V channel activity can also lead to enhanced neuronal excitability through reduced activation of Ca²⁺-activated potassium channels (Marrion and Tavalin 1998; Yu et al. 2010). In a animal models of chronic pain, a reduction of Ca_V2.2 currents in small DRG neurons (Fuchs et al. 2007), and small to medium cutaneous afferent neurons (Lu et al. 2010) has been demonstrated. This reduction leads to enhanced excitability, which may help to produce chronic pain (Hogan et al. 2008; Lirk et al. 2008). The prominent expression of Ca_V2.2 channels by group III and IV neurons suggests that peripheral application of ziconotide or GVIA could enhance action potential activity in these neurons to increase EPR-induced effects on the cardiovascular system. However, natural activators could achieve the same effect. The ATP level in muscle has been demonstrated to depend on muscle contraction (Li et al. 2003; Mortensen et al. 2011), and the concentration (10 μM) is consistent with that used in this study (Li et al. 2003). If P_2Y receptors and Ca_V2.2 channels are expressed on afferent endings in skeletal muscle, muscle released ATP could enhance the EPR via inhibition of Ca_V2.2 channels.

GRANTS

This work was funded by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant AR-059397 (K. S. Elmslie and V. Ruiz-Velasco).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: R.R., S.G.M., V.R.-V., and K.S.E. conception and design of research; R.R., B.H., S.G.M., J.D., M.F., and V.R.-V. performed experiments; R.R., B.H., S.G.M., J.D., M.F., V.R.-V., and K.S.E. analyzed data; R.R., B.H., S.G.M., J.D., M.F., V.R.-V., and K.S.E. interpreted results of experiments; R.R., S.G.M., V.R.-V., and K.S.E. prepared figures; R.R., V.R.-V., and K.S.E. drafted manuscript; R.R., B.H., S.G.M., J.D., M.F., V.R.-V., and K.S.E. edited and revised manuscript; R.R., B.H., S.G.M., J.D., M.F., V.R.-V., and K.S.E. approved final version of manuscript.