Ion channels present in primary afferent neurons may be divided into two main classes: those involved in signal transduction and those involved in the control of excitability. Ion channels involved in signal transduction may be further subclassified on the basis of the stimulus modality(s) to which they are responsive. That is, ion channels have been identified that are opened or closed in response to mechanical,14,15 thermal,16 and chemical stimuli.9,17–19 This final subclass of signal transducing ion channels includes the prototypical ligand-gated ion channels.18,20–22 The second main subclass of ion channels includes background or leak channels, voltage-gated chan- nels, and Ca2+-dependent channels.
9.3.1 TRANSDUCERS
9.3.1.1 Changes in Transducer Properties Influence Nociceptor Processing
In the presence of injury, inflammatory mediators are released at the site of injury.
Many of these mediators are direct acting, resulting in the activation (a depolarization of sufficient magnitude for action potential generation) and/or sensitization (an increase in afferent sensitivity to subsequent stimuli) of primary afferent neurons.
At least one of the mechanisms underlying the actions of inflammatory mediators is a change in an ion channel responsible for the transduction of thermal stimuli, specifically a heat transducer. That is, inflammatory mediators such as bradykinin,12 prostaglandin E2,13 and protons,16,23 induce changes in a heat transducer such that the channel is activated at lower temperatures. Such a change could explain those in thermal sensitivity observed in the presence of inflammation. Inflammation also results in changes in the mechanosensitivity of primary afferent neurons.24–26 Simi- larly, injury results in changes in chemosensitivity.27–30 All of these observations point to potential therapeutic interventions for the treatment of pain associated with injury as well as underscore the importance of the identification and characterization of ion channels involved in sensory transduction.
While a detailed description of the methods used in the isolation and character- ization of ion channels involved in sensory transduction is beyond the scope of the present chapter, the following is a brief discussion of several issues that investigators new to this field may wish to keep in mind.
9.3.1.2 Characterization of Transducers
First, and probably most importantly, the study of ion channels involved in trans- duction will likely require specialized equipment above and beyond the equipment ordinarily required for electrophysiological recording; that is, able to apply adequate
Membrane Properties: Ion Channels 173
and appropriate stimuli. Investigators have employed several distinct stimuli in an effort to identify and characterize mechanical transducers. The application of positive and negative pressure to a patch pipette in the “cell-attached patch” configuration has long been known to activate stretch-activated channels in virtually every cell studied. However, upon establishing the “whole-cell” configuration, it was not been possible to demonstrate whole cell currents evoked with mechanical deformation of the plasma membrane. McBride and Hamill31 theorized that the reason for this failure was because the large volume of the patch electrode served to attenuate changes in pressure at the cell. Consequently, these investigators devised a system to clamp the pressure within the patch pipette.31–33 Recently, however, McCarter and colleagues10 reported whole-cell currents evoked with various mechanical stimuli while employ- ing no special equipment to clamp the pipette pressure.
9.3.1.2.1 Mechanical Transducers
Several different mechanical stimuli have been employed in the study of mechan- otransduction. One involves the application of hypo- and hyperosmotic extracellular solutions causing cells to swell or shrink in response to changes in osmotic pressure.10 Another involves a micromanipulator to apply a probe directly to the neuron studied.34 A third involves a hydraulic stimulus in which an extracellular solution is puffed onto a cell.10 Finally, Maingret and colleagues14 have employed trinitrophenol to produce crenation of the plasma membrane. While all of these stimuli produce electrophysiological changes, whether any are capable of utilizing the ion channels involved in the mechanotransduction that occurs in the peripheral terminals of sensory neurons has yet to be determined.
9.3.1.2.2 Thermal Transducers
As is to be expected, characterization of the ionic mechanisms underlying the transduction of heat requires an apparatus to heat and cool extracellular solutions.
The simplest approach to this problem is to apply solutions warmed to a specific temperature directly to the recording chamber.16 Minimally, one should be able to position a calibrated thermister close to the cell being studied so that the temperature
“seen” by the neuron may be determined. This approach becomes cumbersome if one wishes to test a series of different temperatures and/or hold the temperature at a specific value. Consequently, investigators have turned to the use of Peltier devices to heat and cool solutions.13 Use of a water-cooled Peltier system enables relatively rapid changes in temperature with the generation of virtually no electrical noise.
Resistor-based heating systems can also raise temperatures relatively rapidly. Unlike Peltier systems, however, it is not possible to actively cool solutions heated in this way. Implementing a feedback control circuit to a Peltier- or resistor-based heating system enables one to hold the temperature at a specific value. Several of these devices are commercially available.
9.3.1.2.3 Chemical Transducers
The ionic mechanisms underlying chemical transduction may require the least sophisticated equipment of the modalities. That is, many compounds will produce large sustained responses in neurons and so they may be applied via the recording
174 Methods in Pain Research
chamber perfusion system.9 The advantage of this approach is that the concentration of a test agent may be determined with certainty. However, there are two major disadvantages of this approach. First, it necessary to use relatively large volumes of test agents, thereby prohibiting the study of relatively scarce or expensive compounds.
Second, there are many ligand-gated responses that are rapidly desensitizing.35,36 Therefore, if agonists are applied through the relatively slow chamber perfusion system, the channels may desensitize before they mediate a detectable response. For example, bradykinin-evoked responses are virtually undetectable when applied through the chamber perfusion system, while they are clearly detectable when applied rapidly and briefly (unpublished observation).
Investigators have employed several different approaches to the problem of rapid drug application. One relatively inexpensive approach is the use of a “puffer” pipette.
In this approach, a microelectrode (tip diameter 30 to 100 àm) is positioned near (< 100 àm) the cell under study. This microelectrode, which may be filled with the test agent(s) of choice, is attached to a device that can apply a variable amount of pressure for a variable duration. The magnitude of applied pressure should be under experimental control so that adequate pressure may be applied to pipettes of different resistances to ensure ejection of the test agent. Furthermore, the time of application should be controllable for obvious reasons. The limitation of the puffer approach is that it is not possible to determine the concentration of drug seen by the cell (because it is mixed with normal extracellular solution), and the rate of drug removal depends on the flow of the extracellular solution through the recording chamber. A more sophis- ticated version of the puffer pipette involves a multi-barrel pipette and a switching valve, such that drug application is terminated rapidly by the application of a control solution in addition to the termination of the test solution. An alternate approach is the use of a “sewer-pipe.” This approach utilizes several large-bore pipettes that are attached to a stepping motor. The motor positions the pipette of choice in front of the cell, enabling a rapid exchange of pipettes. Because both the flow through the pipettes and the exchange of pipettes is rapid, solution exchanges may be achieved within milliseconds. Many of these devices are commercially available.
9.3.2 IONIC CURRENTS
9.3.2.1 Voltage-Gated K+ Currents 9.3.2.1.1 Voltage-Gated K+ Currents and
Nociceptive Processing
Voltage-activated K+ currents (VGKCs) have been shown to control action potential thresholds, resting potentials, and firing patterns in other neuronal preparations.37 In primary afferent neurons, at least six VGKCs have been characterized,38–43 includ- ing three inactivating currents and three non-inactivating currents;38 five of these currents are likely to be present in nociceptors.38 Two of the non-inactivating currents have low thresholds for activation and are likely, therefore, to contribute to the control of the action potential threshold. The inactivating K+ currents may also contribute to neuronal excitability as it has been demonstrated that a selective
Membrane Properties: Ion Channels 175
block of a slowly inactivating K+ current decreases both action potential threshold and accommodation.43 Recent evidence suggests VGKCs contribute to injury- induced increases in afferent excitability. For example, K+ current density is decreased by inflammatory mediators44 and nerve injury.3
9.3.2.1.2 Characterization of Voltage-Gated K+ Currents
The separation of VGKCs from other voltage-gated ionic currents simply requires replacing Na+ and Ca2+ in the extracellular solution with non-permeant, non-blocking ionic species. I replace Na+ with choline+ and Ca2+ with Co2+. Because the voltage- sensitivity of some VGKCs is influenced by the concentration and species of divalent ions in the extracellular solution,38,40 it is also possible to eliminate voltage-gated Ca2+ currents (VGCCs) with blockers. However, given that a single sensory neuron may express as many as five VGCCs,45,46 the use of selective VGCC blockers can become expensive. Consequently, many investigators employ non-selective VGCC blockers such as cadmium (50 àM).
While isolating VGKCs from other voltage-gated channels is relatively easy, distinguishing VGKCs from one another can be quite difficult. There are at least three reasons for this. First, several different VGKCs are often expressed in the same neuron.38 Second, the biophysical properties of VGKCs often overlap.38 Third, the pharmacological tools available for the isolation of VGKCs are relatively non- specific.37 Consequently, it is often difficult to determine with certainty the relative contribution of various VGKCs to the total outward current observed in a neuron.
As a result, an experimental design utilizing each neuron as its own control may provide the most unequivocal results in terms of specific VGKCs influenced by specific experimental manipulations; for example, a VGKC is inhibited by prosta- glandin E2 (PGE2).44 It is possible to determine the biophysical properties of the PGE2-sensitive VGKC by digitally subtracting current traces obtained after the application of PGE2 from those obtained before its application.
Because it is often not possible to use each neuron as its own control, it is worth considering some issues that may facilitate a more detailed characterization of the VGKC(s) influenced by a specific experimental manipulation. First, both transient and sustained VGKCs may be subject to steady-state inactivation over similar voltage ranges.38,39 Consequently, while studying currents evoked from two different conditioning potentials may provide valuable information, such a manip- ulation may be insufficient to separate transient and sustained currents. Second, sensory neurons express at least three transient currents that inactivate with mark- edly different time constants.38 Consequently, evoked currents must be sampled rapidly enough to resolve a current that inactivates within 10 msec, yet for long enough to resolve a current that inactivates with a time constant of more than 100 msec. Third, tail currents may be very informative. Various VGKCs may deactivate (i.e., a term used to describe the voltage-dependent transition of ion channels from open to the closed states) with markedly different time constants. Thus, the rate and complexity of current decay following a test potential may suggest the presence of one or more VGKCs contributing to the total outward current at the end of a test pulse.
176 Methods in Pain Research
9.3.2.2 Voltage-Gated Na+ Currents
9.3.2.2.1 Voltage-Gated Na+ Currents and Nociceptive Processing
At least seven distinct voltage-gated sodium currents (VGSCs) have been described in sensory neurons on the basis of distinct biophysical and pharmacological prop- erties.47–50 These currents are generally separated into two groups on the basis of their sensitivity to tetrodotoxin (TTX). One group is blocked by nanomolar concen- trations of TTX, while the other is resistant to TTX at concentrations greater than 1 àM. Of the currents that have been described in sensory neurons, three are TTX sensitive and four (possibly five) are TTX-resistant. Molecular biological studies support the suggestion that sensory neurons express many VGSCs,51 although the association between cloned channels and macroscopic currents is a question that is still being actively investigated. Furthermore, while expression of these currents appears to be tightly regulated, researchers are only just beginning to identify the factors underlying this regulation. For example, at least one TTX-sensitive current is expressed only during development or following nerve injury,52 and the expression of this current appears to be regulated by access to nerve growth factor.53
A detailed characterization of the biophysical properties of VGSCs present in sensory neurons has important implications for our understanding of the underlying mechanisms of pain for several reasons. First, activation of VGSCs is critical for the generation and propagation of action potentials. Second, clinical and basic research indicates that compounds known to block Na+ channels may be effective for the treatment of hyperalgesia and pain.54–59 Third, the biophysical properties of VGSCs have a profound impact on neuronal excitability. For example, the ratio of TTX-resistant to TTX-sensitive VGSCs available for activation influences excitability5 and can change rapidly in response to changes in resting membrane potential. Rapid changes in the ratio of these two currents reflect differences in steady-state inactivation properties and rates of recovery from inactivation of these two currents. Fourth, changes in the expression and/or biophysical properties of VGSCs are associated with the development and maintenance of pain, hyperalgesia, and allodynia associated with several distinct forms of injury including inflamma- tion,60–64 nerve injury,1,4,64 and diabetes.65
9.3.2.2.2 Characterization of Voltage-Gated Na+ Currents
VGSCs may be studied in isolation by blocking K+ and Ca2+ currents as described above. Eliminating VGKCs is generally achieved by replacing K+ with Cs+ in the electrode solution if whole-cell patch configuration is utilized. However, because a relatively large subpopulation of rat sensory neurons expresses a VGKC permeable to Cs+,38 a complete block of outward currents with this approach may be difficult.
Furthermore, this current is also incompletely blocked with very high (~90 mM) concentrations of tetraethylammonium (TEA) in bath solutions.38 However, I have found that including 30 to 40 mM TEA in the electrode solution is quite effective at eliminating residual outward currents in sensory neurons (Fig. 9.2). Voltage-gated Ca2+ currents (VGCCs) may be eliminated by decreasing the concentration of extra- cellular Ca2+ (≤ 0.1 mM) and/or adding non-selective Ca2+ channel blockers such as Cd2+ (50 àM) to the extracellular solution.
Because VGSCs activate rapidly and generally carry inward current, they gen- erate depolarizing voltage errors that can lead to loss of clamp control. Loss of clamp control can be mitigated through the use of relatively low resistance patch electrodes (< 2 MΩ), use of amplifier series resistance compensation circuitry, and decreasing the driving force on Na+ by lowering the concentration of extracellular Na+. I generally use 35 mM Na+ in the bath solution (with choline making up the bulk of the cation in the bath solution) in order to obtain peak whole-cell currents that are generally less than 5 nA; 5-nA current with a 2-MΩ electrode and 80%
series-resistance compensation will generate a 2-mV voltage error.
There are several other issues to keep in mind when designing voltage-clamp protocols for the characterization of VGSCs present in sensory neurons. First, these channels may undergo a process referred to as slow-inactivation. This is a voltage- dependent process similar to the more typical inactivation described with a steady- state inactivation protocol, except that it occurs over many seconds, rather than tens to hundreds of milliseconds.66 The result is that the fraction of available channels may be dramatically influenced by the duration that the membrane is held at a given potential. Ogata and Tatebayashi66 described the slow inactivation of TTX-resistant FIGURE 9.2 Voltage-gated sodium currents studied in isolation following elimination of voltage-gated potassium and calcium currents. Two small diameter DRG neurons were studied in voltage-clamp with the whole-cell patch configuration. The extracellular solution contained 35 mM NaCl (to mitigate voltage errors associated with large inward current), 65 mM choline (a non-blocking, non-permeant cation used as a substitute for extracellular NaCl), 40 mM tetraethylammonium (TEA) (to block some voltage-gated potassium currents), 0.1 mM CaCl2 (to limit the size of voltage-gated calcium currents and calcium-dependent chloride and/or potassium currents), 5 mM MgCl2 (to help stabilize the plasma membrane in the presence of a low concentration of extracellular calcium), 10 mM glucose (to help the cells with their energy requirements) and 10 mM HEPES (to buffer the pH to 7.4). The intracellular solution contained 1 mM CaCl2, 2 mM MgCl2, and 11 mM EGTA (resulting in a free calcium concentration of ~40 nM), 2 mM Mg-ATP (to mitigate run-down of currents and provide additional substrate for phosphorylation reactions), 1 mM Li-GTP (to facilitate G-protein- mediated reactions), and 10 mM HEPES (to buffer the pH to 7.2). The only difference between the solutions used to record the traces on the left and those in the middle was that 40 mM TEA was included in the electrode solution (the CsCl concentration was reduced to 100 mM).
Including TEA in the electrode solution largely eliminated the sustained outward current indicated by the bracket and arrow. The voltage-protocol used to evoke the currents is shown on the right.
2 nA
5 ms -50 mV
20 mV 50 mV
-60 mV 1 nA
5 ms
VGSCs in sensory neurons from neonatal rats that results in a dramatic reduction in peak current over tens to hundreds of seconds. Second, to achieve full recovery from steady-state inactivation of TTX-sensitive VGSCs, it is often necessary to hyperpolar- ize the membrane less than –120 mV. In my experience, sensory neurons do not tolerate hyperpolarized potentials well, often becoming unstable and leaky. However, I have found that increasing the concentration of divalent ions in solution can attenuate this problem. Third, it is generally possible to isolate TTX-resistant VGSCs from TTX- sensitive VGSCs with a 500-msec depolarizing voltage-step to –50 mV (a potential at which the steady-state inactivation of TTX-sensitive currents is complete). However, there does appear to be a TTX-sensitive current that requires membrane depolarizations of –40 mV or greater to fully inactivate (unpublished observation).
9.3.2.3 Voltage-Gated Ca2+ Currents
9.3.2.3.1 Voltage-Gated Ca2+ Currents and Nociceptive Processing
Four voltage-activated Ca2+ currents have been described in mammalian sensory neurons including the T, L, N, and P currents specifically blocked by amiloride, dihydropyridines, ω-conotoxin (ω-CTx), and ω-agatoxin (ω-AgTx), respectively.45,46 A fifth Ca2+ current may also exist in DRG that is resistant to these blockers.45,46 High-threshold Ca2+ currents (N, L, and P) are generally thought to control the efferent function of sensory neurons (i.e., by controlling Ca2+ entry and the subse- quent Ca2+-dependent release of transmitter from neuronal terminals).67 However, these currents may contribute indirectly to nociceptor excitability through effects on Ca2+-dependent currents,68 and/or through control of the release of transmitters from the primary afferent that may modulate nociceptor excitability directly and/or indi- rectly.69 In contrast, low-threshold T-type currents have been shown to underlie action potential bursts in sensory neurons,70 although the magnitude of this current is relatively small in putative nociceptors.45 Recent evidence suggests that voltage- gated Ca2+ currents may also contribute to injury-induced increases in afferent excitability as nerve injury results in changes in both T- and N-type currents.71,72 9.3.2.3.2 Characterization of Voltage-Gated Ca2+ Currents
VGCCs are studied in isolation following elimination of VGKCs and voltage-gated Na+ currents (VGSCs). Eliminating VGSCs may be achieved by substituting a choline for Na+ in the extracellular solution. Eliminating VGKCs is generally achieved by replacing K+ with Cs+ in the electrode solution if whole-cell patch- configuration is utilized. Because outward Cs+ current (passing through voltage- gated K+ channels) may contaminate VGCCs as well as VGSCs, 30 to 40 mM TEA in the electrode solution may be used to eliminate residual outward currents in sensory neurons (see Fig. 9.2).
While there are differences between VGCCs with respect to their biophysical properties, with the exception of T-type currents, these differences do not enable the isolation of one current from another with the use of unique voltage-clamp protocols and digital subtraction. Rather, the pharmacological differences between VGCCs provide the basis by which one current is distinguished from another. Because of this, the role of a particular VGCC is generally implicated with an occlusion exper-