|Home | About | Journals | Submit | Contact Us | Français|
Chronic neuropathic pain resulting from neuronal damage remains difficult to treat, in part due to incomplete understanding of underlying cellular mechanisms. We have previously shown that inward Ca2+ flux (ICa) across the sensory neuron plasmalemma is decreased in a rodent model of chronic neuropathic pain, but the direct consequence of this loss of ICa on function of the sensory neuron has not been defined. We therefore examined the extent to which altered membrane properties after nerve injury, especially increased excitability that may contribute to chronic pain, are attributable to diminished Ca2+ entry.
Intracellular microelectrode measurements were obtained from A-type neurons of dorsal root ganglia excised from control rats and those with neuropathic pain behavior following spinal nerve ligation. Recording conditions were varied to suppress or promote ICa while biophysical parameters and excitability were determined.
Both lowered external bath Ca2+ concentration and blockade of ICa with bath cadmium diminished the duration and area of the afterhyperpolarization (AHP), accompanied by decreased current threshold for action potential (AP) initiation and increased repetitive firing during sustained depolarization. Reciprocally, elevated bath Ca2+ increased the AHP and suppressed repetitive firing. Voltage sag during neuronal hyperpolarization, indicative of the cation-nonselective H-current, diminished with lowered bath Ca2+, cadmium application, or chelation of intracellular Ca2+. Additional recordings with selective blockers of ICa subtypes showed that N-, P/Q, L-, and R-type currents each contribute to generation of the AHP, and that blockade of any of these as well as the T-type current slows the AP upstroke, prolongs the AP duration, and (except for L-type current) decreases the current threshold for AP initiation.
Taken together, our findings show that suppression of ICa decreases the AHP, reduces the hyperpolarization-induced voltage sag, and increases excitability in sensory neurons, replicating changes that follow peripheral nerve trauma. This suggests that the loss of ICa previously demonstrated in injured sensory neurons contributes to their dysfunction and hyperexcitability, and may lead to neuropathic pain.
Loss of inward Ca2+ current in A-type neurons, such as follows peripheral nerve injury, contributes to increased sensory neuron excitability. Measures that increase inward Ca2+ flux may potentially be therapeutic for painful peripheral neuropathy.
Activity of primary sensory neurons induced by natural stimulation or by processes such as trauma, inflammation and nerve injury, is the origin of all but a small fraction of painful sensations. Sensory afferents are additionally the afferent source of intraoperative nociceptive reflexes that trigger cardiovascular, ventilatory and neurohumoral efferent pathways. The central role of these neurons in painful conditions and anesthesia makes it critical to understand the regulation of their excitability. The sensory neuron plasmalemma is equipped with a variety of voltage-activated ion channels conducting Na+, K+ or Ca2+, which together determine the biophysical function of the membrane. Currents through voltage-activated Ca2+ channels play a critical double role. First, inward Ca2+ flux (ICa) depolarizes the cell, and thus contributes to action potential (AP) formation. Once reaching the intracellular compartment, however, Ca2+ is also a key second messenger, controlling a broad range of neuronal functions including kinase activity, neurotransmitter release, cell differentiation and growth, genetic expression, and cell death. Membrane events are regulated by intracellular Ca2+ through depression of ICa (1), stimulation of Ca2+-activated Cl− channels (ICl(Ca)) (2), and opening of Ca2+-activated K+ channels (IK(Ca)), which are widely distributed among dorsal root ganglion (DRG) neurons of all sizes (3).
Pain resulting from neuronal damage is a special case in which the primary pathology involves disrupted regulation of sensory neuron excitability at the site of injury and in the DRG proximal to the injury (4–6). We and others have reported loss of both high-voltage activated (HVA) (3,7–9) and low voltage-activated (LVA) subtypes (10) of ICa in DRG neurons following peripheral nerve injury. Although the effects of decreased ICa upon neuronal membrane properties have been detailed in other cell types, the only studies examining ICa regulation of DRG neuronal excitability have employed the patch-clamp electrophysiological technique on dissociated neurons (11,12). This is problematic, since dialysis of the cytoplasm by the solution in the patch pipette disrupts natural biophysical events, as demonstrated by substantially prolonged action potential (AP) durations, more frequent afterdepolarizations following the AP, and diminished AHP amplitude, compared to microelectrode recording (2,6,13,14). For instance, although AP durations in control neurons are consistently less than 4mS when recorded extracellularly or by high resistance intracellular microelectrodes (6,15), durations may be as long as 60mS using the patch-clamp method (11), which indicates clearly abnormal recording conditions. Furthermore, intracellular regulation of Ca2+ concentrations is disrupted by the patch clamp technique (16), and amplifiers typically used for patch clamp recording are not ideally suited for recording APs (17). Finally, dissociation itself produces an injury-like effect on neuronal excitability (18).
Since injury also produces substantial changes in voltage-gated Na+ and K+ channels of sensory neurons (19,20), the overall purpose of this study is to clarify the extent to which altered membrane functions after injury, especially increased excitability, are specifically attributable to diminished Ca2+ entry. Therefore, the present study examines the influence of changes in ICa and intracellular Ca2+ upon AP dimensions, threshold for initiation of APs, and repetitive firing. We employed a technique designed to minimize disruption of cellular Ca2+ signaling, by using intracellular microelectrode recording from intact DRGs.
All procedures used in the study were approved by the Animal Resource Center of the Medical College of Wisconsin. Experiments were performed on male Sprague-Dawley rats (200–300g) obtained from a single vendor (Taconic Inc., Germanville, NY). Our general strategy was to record from intact DRGs, which avoids disruptive effects of cell dissociation, and to employ multiple pharmacologic measures that together will delineate the role of ICa in sensory neuron activation.
Ganglia were removed 21±4 days after surgery, a time at which hyperalgesia has fully developed (21). Rats were anesthetized with halothane in oxygen, and a laminectomy was performed up to the second thoracic level while the surgical field was perfused with artificial cerebrospinal fluid (aCSF, in mM: NaCl 128, KCl 3.5, MgCl2 1.2, CaCl2 2.3, NaH2PO4 1.2, NaHCO3 24.0, glucose 11.0) aerated with 5% CO2 and 95% O2 to maintain a pH of 7.35. The fourth and fifth lumbar DRGs and attached dorsal roots were removed and the connective tissue capsule dissected away from the ganglia under 20× magnification. Ganglia were transferred to a glass-bottomed recording chamber and perfused with 35°C aCSF. The proximal cut end of dorsal roots was placed on a pair of platinum wire stimulating electrodes. DRG neurons were viewed using an upright microscope equipped with differential interference contrast optics and infrared illumination. Neuronal soma diameter was determined with the focal plane adjusted to reveal the maximum somatic area using a calibrated video image.
Intracellular recordings were performed with microelectrodes fashioned from borosilicate glass (1mmOD, 0.5mmID, with Omega fiber – FHC Bowdoinham, USA) using a programmable micropipette puller (P-97, Sutter Instrument Company, Novato, CA). Microelectrode resistances were 80–120 MΩ when filled with 2M potassium acetate. Neurons were predominantly selected from the two outermost cell layers of the dorsal medial aspect of the DRG, and were impaled under direct vision by minimally indenting the membrane and then applying an oscillating current to the microelectrode. Data were acquired after stable recordings were achieved, typically within 5min of puncture. Membrane potential was recorded using an active bridge amplifier (Axoclamp 2B, Axon Instruments, Foster City, CA), except in protocols requiring simultaneous current injection through the electrode, for which we used discontinuous current clamp recording mode. This allows accurate recording of voltage, although the noise level is increased compared to bridge mode. During discontinuous mode recording, the switching frequency from current injection to recording was 2kHz, and full settling of the electrode charge was confirmed. Currents were filtered at 1 kHz (discontinuous mode) and 10kHz (bridge mode), and then digitized at 10kHz (discontinuous mode) or 40 kHz (bridge mode; Digidata 1322A and Axograph 4.9, Axon Instruments) for data acquisition and analysis. Somatic APs were produced by axonal conduction of APs from a dorsal root site 18–22mm distant, where the root was stimulated with square-wave pulses of up to 90mA lasting 0.06mS via bipolar electrodes, except in specified protocols in which current was injected through the recording electrode.
Inclusion criteria were a resting membrane potential (RMP) negative to −50mV, and an AP amplitude greater than 40mV. Together, these excluded approximately 6% of recordings. APs measures (Fig. 1) were obtained from single traces after consistency of dimensions was confirmed by comparison to 10 sequential APs. The presence of a hump or inflection on the descending limb of the AP was determined by examination of the differentiated trace (Fig. 1B and 1C). Determinations of AP and AHP durations and the area under the curve for the AHP (AHParea) were performed digitally. Input resistance was calculated from the minimum potential achieved during 100ms hyperpolarizing current (0.5nA) injection through the recording electrode (15). Sensory neurons show time-dependent rectification (“sag”) in response to hyperpolarization, predominantly attributable to the hyperpolarization-induced H-current (22). If a neuron showed sag with 0.5nA injection, additional traces during injection of lower depolarizing currents were examined to assure that the input resistance calculation was not affected. Sag was quantified as the fractional return from the peak hyperpolarized membrane potential back towards RMP during injection of a 1.2nA hyperpolarization current for 100ms. Rheobase was determined as the minimum current able to elicit an AP during incremental depolarizing current injection of 0.5–10nA for 100mS. The pattern of impulse generation was determined during current injection steps beyond rheobase, at which neurons either produced single APs or fired repetitively. The influence of bath Ca2+ level or drug upon AP firing pattern was measured at a depolarizing voltage that first produced a bath Ca2+- or drug-induced difference in the number of APs generated.
Neurons were classified by the method of Villiere and McLachlan (15). Conduction velocity (CV) was measured by dividing the distance between stimulation and recording sites by the conduction latency. Neurons with dorsal root CV<1.5m/S were considered C-type, neurons with CV>15m/S were considered Aα/β-type, and neurons with CV>1.5m/S but CV<10m/S were considered Aδ-type. For neurons with CV between 10 and 15m/S, long AP duration was used to categorize the cells as Aδ-types (15). Too few C-type neurons were successfully recorded to provide statistical analysis, and are not reported here.
The influence of ICa on membrane biophysical events was revealed by various techniques that alter ICa or intracellular Ca2+ concentration. Baseline electrophysiologic parameters were measured in neurons in aCSF, after which they were exposed by bath change to an external solution identical except for a low Ca2+ concentration, in which CaCl2 was substituted with MgCl2 (n = 25 Aα/β, 6 Aδ). This resulted in a measured Ca2+ concentration of 0.35 ± 0.02 mM (n = 3 measurements, by Ca2+-sensitive electrode, ABL 515, Radiometer, Copenhagen). Magnesium was added (final concentration 3.5mM) to exclude the possible influence of changed surface charge, which is otherwise a potent source of neuronal stimulation (23), and to maintain a constant divalent cation effect on potassium channels (24). Stable recordings were achieved after a wash-in interval of 3 min. Using conventional patch-clamp recording of whole-cell ICa (7) in dissociated DRG neurons, we confirmed that only 6±4% of baseline ICa remains during these low bath Ca2+ conditions (n=4; Fig. 2). In supplementary experiments, ICa was decreased in other neurons (n = 13 Aα/β, 7 Aδ) by bath exchange to aCSF containing cadmium 200μM, a widely-used blocker of neuronal voltage-activated Ca2+ channels (13,25).
In other recordings (n = 6 Aα/β and 1 Aδ neurons), ethylenediamine-tetraacetic acid (EDTA) 500μM was added to the potassium acetate micropipette solution to chelate intracellular Ca2+ upon entry, using a previously established iontophoretic method in which injection is achieved using a 1nA hyperpolarizing current for 60s (26). Preliminary studies showed that current injection of this magnitude alone had no effect on measured parameters. Further studies (n = 10 Aα/β, 1 Aδ neurons) used the fast Ca2+ chelator 1,2-Bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid as the cell-permeant acetoxymethyl ester (BAPTA-AM) to lower intracellular concentration (27). This was dissolved in bath solution (200μM) and delivered by a microperfusion technique from a pipette with a 10μ diameter tip that was positioned 200μ from the impaled neuron, and ejected continuously by pressure applied to the back end of the pipette (Picospritzer II, General Valve Corp, Fairfield, NJ). Preliminary experiments in which cell depolarization was measured during microperfusion application of solutions of various K+ concentrations indicated an effective 5-fold dilution of pipette solution into the bath at the cell membrane, so the BAPTA-AM concentration was approximately 40μM at the cell surface.
In order to further test the role of intracellular Ca2+ on sensory neuron function, increased ICa was provoked in other neurons by bath change to a solution with Ca2+ concentration elevated to 5mM or 7mM. Since the results of these two concentrations were comparable, the results were pooled for analysis (n = 8 Aα/β and 9 Aδ neurons). Patch-clamp measurement of ICa showed a 35±16% increase of ICa under these conditions (n=4; Fig. 2).
Sensory neurons express a variety of voltage-gated Ca2+ currents, including subtypes with both LVA and HVA activation characteristics (28). Since the methods described above affect all voltage-activated Ca2+ currents indiscriminately, we examined the contributions of specific Ca2+ channel subtypes to membrane activation before and after microperfusion application of selective blockers. Specifically, the agents and concentrations in the application pipette were as follows: nickel (2mM) and mibefradil (5μM) to block the T-type LVA current, nitrendipine (50μM) to block the L-type HVA current, BayK 8644 (50μM) to amplify L-type current, SNX-111 (2.5μM) to block the N-type HVA current, ω-Aga-IVA (2.5μM) to block the P/Q-type HVA current, and SNX-482 (2.5μM) to block the R-type HVA current. Since the peptide toxins are irreversible, no more than 2 recordings were made from widely separated sites on each ganglion. In other experiments, all the selective blockers (nitrendipine, SNX-111, ω-Aga-IVA, SNX-482) were administered together to examine the effect of complete HVA ICa blockade. The vital dye fast-green, which is compatible with neuronal electrophysiological recording (29,30), was included at a final concentration of 0.5% in order to permit visual confirmation of drug delivery to the recorded cell during application, and to ensure that new recordings were made from parts of the ganglion that had received no previous drug exposure. Cytochrome C 0.1% was included with the peptide toxins to inhibit nonspecific binding to glass surfaces. The nitrendipine solution included dimethyl sulfoxide at a final concentration of 0.25% in the delivery pipette. Control experiments were performed to determine if pressure ejection of vehicle alone altered recorded parameters. In this portion of the study, dimensions were measured from APs averaged from 10 sequential traces before and 5min after initiation of drug administration, at which time stable changes had evolved.
SNX-111 was the kind gift of Dr. Scott Bowersox from Neurex Corporation (Menlo Park, CA), while ω-Aga-IVA was the gift of Dr. Nicholas Saccomano from Pfizer (Groton, CT), and SNX-482 was purchased from Peptides International (Louisville, KY). All other agents were purchased from Sigma-Aldrich Co. (St. Louis, MO).
The control group included cells from both L4 and L5 ganglia in animals having undergone skin incision only. It was not the purpose of the study to determine differences in biophysical parameters between cell types or injury categories, as has been done by us before (6), so these comparisons were not tested statistically. Data were analyzed using Excel (Microsoft Co. Seattle, USA) and Statistica 6.0 (StatSoft, Tulsa, USA). Averaged results are expressed as means ± SD. The effect of drug application was evaluated using paired Student’s t-test, which allows identification of significant drug effects upon measured parameters in the context of natural variability between neurons that results in large SD and differences between group means under baseline conditions. Nonparametric data were evaluated by cross-tabulation with significance tested by Pearson Chi-square for a main effect and planned post hoc paired comparisons by Fisher’s exact test. The threshold for statistical significance was accepted at P<0.05.
Electrophysiological data were obtained from a full range of DRG neuronal sizes (average diameter 39.7±0.6μ for Aα/β, 34.2±0.8μ for Aδ). As shown in Figure 3, the distributions of neuronal diameters for Aα/β and Aδ categories extensively overlap. Main results for altered bath Ca2+ for are summarized in Table 1. Effects of lowered and raised external bath Ca2+ and cadmium administration were reversible on washout (data not shown).
Low bath Ca2+ levels depolarized the RMP in Aδ neurons and decreased AP amplitude in Aα/β and Aδ neurons (Fig. 4). Bath Ca2+ withdrawal also led to the disappearance of the hump on the descending limb of the AP in 2 of 4 inflected Aα/β neurons and in 3 of 5 inflected Aδ neurons. Similarly, cadmium application abolished the inflection in both of 2 inflected neurons, so that decreasing ICa eliminated the inflection in 7 of 11 inflected neurons overall (P<0.01). However, high bath Ca2+ did not elicit the appearance of de novo inflections. Thus, while ICa is necessary to produce an inflection, not all cells can do so even if intracellular Ca2+ levels are increased. Bath Ca2+ withdrawal had no significant effect on AP duration in either Aα/β or Aδ neurons, and cadmium similarly had no effect on AP duration in Aα/β neurons (from 0.85±0.19 to 0.82±0.18ms; P=0.25, n = 13) or in Aδ neurons (from 0.96±0.13 to 0.90±0.12ms; P=0.08, n = 7). Contrasting effects on inflection and duration are attributable to our measurement of the duration at the base of the AP. Buffering of intracellular Ca2+ with BAPTA-AM did prolong the AP duration (from 0.88±0.08 to 0.96±0.11ms; P<0.05, n = 9). BAPTA-AM also decreased the AP upstroke velocity (from 425±54 to 366±58mV/ms; P<0.05, n = 9), as did cadmium blockade of ICa in Aδ neurons (from 256±51 to 246±52 mV/ms; P<0.05, n = 7).
Low bath Ca2+ levels caused a decrease in AHP duration and area for both Aα/β and Aδ neurons (Fig. 4, Table 1). This observation was supported by decreased AHP area when ICa was blocked by cadmium in Aα/β neurons (from 238±269 to 147±152mV·mS; P = 0.01, n = 13), although a change was not confirmed in Aδ neurons (from 227±234 to 147±161mV·mS; P=0.11, n = 7). Combined application of all HVA ICa blockers to neurons (n = 7 Aα/β and 5 Aδ neurons) similarly reduced AHP duration (from 21.1±22.0 to 13.0±15.2mS; P<0.01) and AHP area (from 397±570 to 204±344mV·mS; P<0.05). The importance of Ca2+ influx for generating the AHP was further substantiated by chelating intracellular Ca2+ with EDTA iontophoresis, which similarly decreased AHParea (from 103±76 to 79±67mV·mS; P=0.05, n = 7; Fig. 4), as did BAPTA-AM application (from 187±163 to 134±111mV·mS; P<0.05, n = 9). Reciprocally, bath change to high Ca2+ concentration increased the AHP duration and area in Aα/β neurons (Fig. 5), providing confirmation of the dependence of the AHP on ICa. Variability in baseline AHP dimensions between groups reflects natural heterogeneity of IK(Ca) ion sensory neurons (31).
Low bath Ca2+ solution decreased rheobase in Aα/β neurons (Fig. 6, Table 1), which also followed blocking ICa using cadmium (from 1.77±1.6nA to 1.45±1.27nA; P = 0.02, n=13), and by application of BAPTA-AM (from 3.10±1.7nA to 2.39±1.72nA; P < 0.01, n=7). Three of 38 Aα/β and 1 of 10 Aδ neurons fired repetitively during sustained depolarization under baseline conditions, but 9 additional Aα/β neurons fired repetitively in low Ca2+ solution (P <0.05). The average number of APs in Aα/β neurons during these sustained (100ms) depolarizations followed a similar pattern, increasing from 1.3±0.9 to 10.9±13.4 (P = 0.01) with lowered bath Ca2+. Cadmium similarly provoked novel repetitive firing in Aα/β neurons (from 3 to 6 of 13 cells), and Aδ neurons (from 2 to 4 of 6 cells, Fig. 7), as did BAPTA-AM (from 2 to 5 of 7 cells; average rate increased from 1.1±0.8 to 13.1±5.6, P<0.05, n=7). During combined application of HVA blockers, 4 of 4 nonaccommodating neurons fired more rapidly.
Changing bath to high Ca2+ solution had the opposite effect on the pattern of neuronal firing, causing 2 of 7 repetitively firing Aα/β neurons (Fig. 6) and 1 of 2 repetitively firing Aδ neurons to become adapting and fire only once during sustained depolarization.
The depolarizing sag in membrane potential during hyperpolarization was diminished by low Ca2+ bath solution in both Aα/β and Aδ neurons (Fig. 7A, Table 1), as did combined application of HVA ICa blockers (from 21.8±18.0% to 12.4±18.0%; P<0.05, n = 7 Aα/β and 5 Aδ neurons), cadmium application (from 37.6±17.6% to 23.2±11.4%; P<0.05, n = 5 Aδ neurons), and chelation of intracellular Ca2+ with either EDTA (from 19.23±17.2% to 14.9±16.8%; P<0.05, n = 5; Fig. 7B) or BAPTA-AM (from 28.0.6±19.5% to 13.9±17.6%; P<0.05, n = 8). Together, these observations indicate a Ca2+ dependency for the hyperpolarization-induced inward current.
The regulation of AP dimensions and neuronal excitability by individual ICa subtypes was further examined using selective blockers on 59 Aα/β neurons (Table 3, Fig. 8). Mibefradil, in contrast to low Ca2+ bath solution, hyperpolarized the RMP, which may be attributable to nonspecific effects of this agent (32,33). AP amplitude was increased by nickel and SNX-482 but decreased by nitrendipine. All agents prolonged AP duration, and all agents except nickel and nitrendipine slowed the CV. AHParea was decreased by SNX-111, nitrendipine, ω-Aga-IVA, and less so by SNX-482. All blockers except nitrendipine decreased rheobase, indicating an excitatory influence, although repetitive firing during sustained depolarization was not affected by individual application of the selective blockers. Voltage sag after membrane hyperpolarization was decreased only by mibefradil, ω-Aga-IVA, and SNX-482. None of the agents produced a consistent change in the presence or absence of inflection. The only effect of pressure ejection of vehicle alone (n = 9 Aα/β neurons and 1 Aδ neuron) was to decrease the slope of the ascending limb of the AP (dV/dt), which was decreased to a greater extent by all agents.
We extended our examination of the role of Ca2+ entry by amplifying ICa with the L-current agonist BayK 8644 (n = 10 Aα/β neurons and 1Aδ neuron; Table 3), which slowed the CV, increased the AP amplitude, and substantially prolonged the AP. The effect on the AHP showed divergent patterns. Maximum AHP amplitude was decreased in 10 neurons (Figs. 9A and 9B), but increased in another (Fig. 9C). Expansion of the AHP area was also seen in two other neurons that developed new AHPs following the afterdepolarization, which were more delayed than the AHP present at baseline (Fig. 11B). Although there were no afterdepolarizations under baseline conditions, these appeared in 7 of 11 neurons during application of BayK 8644, either with loss (Fig. 9A) or expansion (Fig. 9B) of the AHP. In contrast, only 2 small afterdepolarizations developed during application of the selective blockers to 59 neurons (one cell each for SNX-111 and ω-Aga-IVA, P<0.001). BayK 8644 decreased rheobase in all neurons, and 9 of 10 neurons either initiated repetitive firing during application of BayK 8644 or increased firing rate.
This study was designed to identify changes in membrane electrical properties of sensory neurons attributable to the loss of inward ICa, and thereby indicate the consequences of injury-related ICa loss that we have previously identified. To this end, we manipulated ionic conditions and applied a variety of pharmacologic agents to DRG neurons in their non-dissociated state, while recording membrane events in a manner least likely to disrupt intracellular Ca2+ handling. This approach lacks the ability to identify effects on isolated specific currents that patch-clamp membrane control provides, but it allows characterization of the natural interactions of the cell’s full complement of membrane voltage-gated and Ca2+-activated channels. Moreover, processes that sequester and extrude intracellular Ca2+ remain largely intact. Therefore, the role of ICa can be put into a physiologically relevant context in the sensory neuron soma. Our observations collectively indicate that loss of ICa and depression of the intracellular Ca2+ level are accompanied by elevated excitability of DRG neurons. Although similar events are likely in C-type neurons, this cannot be proved as those were not examined here.
At any moment, the neuronal transmembrane potential is the result of ionic conductance across the membrane through multiple channel types. The function of these channels is linked, since membrane voltage and intracellular Ca2+ regulate opening of these channels, but levels of both voltage and Ca2+ concentration are simultaneously controlled by currents through the various channels. Thus, it is impossible a priori to predict the final effect on AP dynamics of a change in a single channel. Central to this study is the specific case of divergent electrical effects that may arise from a loss of ICa. On the one hand, inward Ca2+ flux depolarizes the membrane such that decreasing ICa should inhibit generation of APs and diminish axonal conduction success. However, the entry of a diminished amount of Ca2+ results in less recruitment of IK(Ca), which in turn will lead to the excitatory effects of membrane depolarization and elevated membrane resistance (34). In cerebellar Purkinje neurons, for instance, block of ICa enhances firing since the diminished outward K+ flux outweighs the decreased inward Ca2+ flux (25). Further complexity is added by participation of ICl(Ca) (2), a depolarizing current that competes with inward IK(Ca). Even membrane hyperpolarization may either directly stabilize the neuron or create conditions for repetitive firing by phasically reactivating Na+ currents and LVA Ca2+ currents. As a result of the intricate interplay of these various currents, experimental observation is necessary to resolve the balance of effects in a natural system.
To determine the end result of decreased ICa, we modulated inward ICa using techniques each of which has imperfections. Bath Ca2+ withdrawal dependably decreases the driving force for ionic flux through the Ca2+ channels but may also change the surface potential and thus disrupt voltage gradients in the immediate microdomain of the channel (23). We compensated for this by reciprocal Mg2+ addition, although divalent cations may differ in extent of effects on various channels (24). A factor that further complicates interpretation of Ca2+ withdrawal experiments is the presence in neurons of a nonselective cation channel that is suppressed by external Ca2+ (35,36), particularly in nociceptive sensory neurons (37). Calcium withdrawal may thus initiate a separate excitatory inward current, which potentially accounts for differences between findings with bath Ca2+ withdrawal and ICa blockade. Also, a direct interaction of Ca2+ with the Na+ channel pore has been demonstrated in which Na+ channels that are not Ca2+-occupied fail to close (38). We elevated bath Ca2+ to predictably increase the concentration gradient driving Ca2+ entry. This, however, does not necessarily produce opposite electrophysiological effects compared to Ca2+ withdrawal, for instance on AP duration (39), as may be the case if Ca2+-dependent processes are saturated in the baseline condition. Cadmium inhibits ICa globally, but lacks perfect selectivity since it may also minimally modulate Na+ currents (25) and K+ currents (24). Although peptide toxins for HVA ICa subtypes are highly specific, there is no comparably specific toxin for T-type currents. We therefore used both nickel and mibefradil (40), which provided largely congruent results. Nitrendipine was chosen to block L-type current because of its greater specificity in DRG neurons compared to other dihydropyridine blockers (41). Despite these precautions, overlapping effects are noted even with the most selective toxins (42). Finally, modulation of excitability may conceivably also result through ICa regulation of release of regulatory peptides from glia and neurons in the DRG (43). Because of these various limitations, we employed multiple complementary techniques for decreasing and increasing Ca2+ entry to resolve the role of ICa.
Withdrawal of bath Ca2+ and application of cadmium both reduced the incidence of inflection of the descending limb of sensory neuron APs, presumably through the depression of ICa. Blockade of individual ICa subtypes had no effect on AP inflection, perhaps indicating the partial contribution of several subtypes to this feature. Our findings confirm observation on cranial nerve sensory neurons that identified a contribution of ICa to the inflection (39,44). This contrasts with the effect of nerve injury, which is accompanied by an increase in the frequency of AP inflection (6) despite the development of a concurrent loss of high-voltage and low-voltage activated Ca2+ currents (7,10). Since IK(Ca) participates in repolarization of the AP and competes with ICa in the formation of the inflection (31,45), a possible explanation is that injury induces a proportionately greater loss of IK(Ca) than ICa, through a decrease of Ca2+-activated K+ channel number or function. Indeed, we have recently identified such a post-injury deficit (46). Loss of voltage-gated K+ current (20) might also contribute to more frequent inflections after injury. Although slow TTX-resistant Na+ current also contributes to the AP inflection (39,47), injury is associated with a decrease of this current (48).
The RMP is depolarized and the AP amplitude is diminished by low bath Ca2+ levels but not by cadmium administration. This may be due to activation of an inward current during low extracellular Ca2+ conditions as noted above, although combined administration of HVA blockers also diminishes AP amplitude. Individual selective blockers have a divergent effects on AP amplitude, perhaps indicative of their variable contributions of ICa versus IK(Ca) during AP depolarization and repolarization. The decrease of AP amplitude following nerve injury (6) may thus be attributed either to loss of L-type current or simultaneous changes in Na+ channels. Selective blockers of ICa consistently prolong AP duration, presumably by a dominant indirect inhibition of inward IK(Ca) (25), and thus duplicate the effect of nerve injury (6). The lack of effect of low bath Ca2+ and cadmium on AP duration may result from a less complete loss of ICa. Alternatively, the difference may be due to secondary effects of low extracellular Ca2+ noted above. Increased AP duration may contribute to neuropathic pain by enhancing synaptic neurotransmitter release (49).
Although Ca2+-independent K+ currents may contribute to the formation of the AHP (50,51), previous studies have shown a critical dependence of the AHP upon ICa (52–54), which our observations now confirm in adult DRG neurons. AHP regulates a critical aspect of neuronal excitability by determining spike frequency adaptation and the ability of sensory neurons to maintain rapidly firing pulse trains (15,55,56). In this study, repetitive firing was amplified by lowering extracellular Ca2+, while elevated Ca2+ had the opposite effect, indicating that the elevated excitability found in neurons after axonal injury may be the result of diminished ICa.
In various neuronal types, the source of Ca2+ that activates IK(Ca) is not uniformly distributed among Ca2+ channel subtypes, due to colocalization of specific isoforms of Ca2+ channels and Ca2+-activated K+ channels (57,58). In contrast, we found in sensory neurons that N-type, P/Q-type and L-type currents all contribute to the generation of the AHP, with a lesser role for R-type current, and that LVA current (T-type) contributes minimally to activation of channels producing the AHP. This may explain why selective blockade of ICa subtypes minimally contributed to repetitive firing (2 cells of 43 became nonaccommodating), since burst firing is regulated by the combined contributions of more than a single channel subtype alone.
Enhancement of L-type current with the agonist BayK 8644 is associated with substantial prolongation of the AP and generation of afterdepolarizations, which may result directly from sustained inward Ca2+ flux or from activation of Cl− channels. This latter option is supported by previous observations of large afterdepolarizations through activation of ICl(Ca) (2), and BayK 8644 has previously been reported to activate Ca2+-sensitive Cl− current in rat DRG neurons (59). Repetitive firing was induced or enhanced by BayK 8644, and can be attributed to the influence of afterdepolarizations (60). Heterogeneity is evident, however, since increased Ca2+ influx from BayK 8644 initiates a new, sustained AHP in a subset of neurons, which highlights the complexity of competing Ca2+ effects.
The rheobase current, which is the depolarizing current just adequate to initiate an AP, is decreased in Aα/β neurons during low bath Ca2+ and application of cadmium and most selective ICa blockers. We cannot explain the decrease in rheobase subsequent to lowering of ICa, but we note that it is not due in all cases to an increase in input resistance, such as might occur with decreased IK(Ca) or Ih. Therefore, loss of ICa must lead to an elevated intrinsic excitability of the neuronal membrane, making it easier for generator currents at natural and pathologic sensory transduction sites to initiate neuronal activity. The ubiquitous actions of Ca2+/calmodulin (61) and Ca2+-sensitive kinases (62–65) on a variety of membrane channels may contribute to membrane hyperexcitability during disruption of intracellular Ca2+ levels. Increased Ca2+ loading during Bay K 8644 application also destabilizes the sensory neuron, however, which points to a system in which normal membrane excitability is maintained only within a certain range of intracellular Ca2+ concentrations.
Voltage sag as recorded in this study in sensory neurons is mostly due to the hyperpolarization-induced current Ih (15), which aids in setting the RMP and may generate excitability by contributing to the pacemaker potential of spontaneously active cells. However, Ih can also suppress excitation by decreasing input resistance in the resting neuron. In our study, depression of ICa decreased hyperpolarization-induced voltage sag. The loss of Ih in sensory neurons after bath Ca2+ withdrawal, application of ICa blockers, and intracellular Ca2+ buffering with EDTA and BAPTA-AM is similar in this respect to the effect of peripheral nerve injury (6). These findings indicate that cytoplasmic Ca2+ level regulates the hyperpolarization-induced inwardly rectifying current in sensory neurons, as has been reported for cortical and brainstem neurons (66,67), possibly by acting through the Ca2+-calmodulin signaling pathway and Ca2+/calmodulin-dependent protein kinase II activation (68). Nerve injury may additionally decrease expression of the channel underlying the Ih current (69).
The findings of this study reveal that suppression of ICa increases DRG neuron excitability. Therefore, the loss of ICa associated with peripheral nerve injury may in part account for the hyperexcitability of the sensory pathway in neuropathic pain states. Furthermore, processes that lower extracellular Ca2+ concentration may be expected to result in neuronal excitation. Both neural trauma (70) and intense neuronal activity, such as happens at the moment of peripheral nerve injury, depress extracellular Ca2+ concentrations (71,72), and therefore may initiate processes that amplify sensory processing even in the absence of changes in Ca2+ channels.
These findings may appear to contradict reports of analgesia after intrathecal administration of Ca2+ channel blockers in experimental nerve injury (73) or clinical neuropathic pain (74). However, neurophysiological processes, including Ca2+ signaling, are spatially distinct. For instance, relief of pain by intrathecal ICa blockade may result from interference with neurotransmitter release in the spinal cord, which is not applicable at peripheral sites. Block of ICa at the SNL injury site has no analgesic effect (73). Our findings point to the possibility that medications anatomically and pharmacologically targeted at elevating inward Ca2+ flux of sensory neurons in the DRG may provide novel treatments for neuropathic pain.
Supported by grant NS-42150 from the National Institutes of Health, Bethesda, Maryland, USA to QH. Presented in part in abstract form at the European Society of Anaesthesiology 2005 Annual Meeting, Vienna, Austria.