Protein kinase C mediates up-regulation of tetrodotoxin-resistant, persistent Na⁺ current in rat and mouse sensory neurones

Abstract

The tetrodotoxin-resistant (TTX-r) persistent Na⁺ current, attributed to Na_V1.9, was recorded in small (< 25 μm apparent diameter) dorsal root ganglion (DRG) neurones cultured from P21 rats and from adult wild-type and Na_V1.8 null mice. In conventional whole-cell recordings intracellular GTP-γ-S caused current up-regulation, an effect inhibited by the PKC pseudosubstrate inhibitor, PKC19–36. The current amplitude was also up-regulated by 25 μm intracellular 1-oleoyl-2-acetyl-sn-glycerol (OAG) consistent with PKC involvement. In perforated-patch recordings, phorbol 12-myristate 13-acetate (PMA) up-regulated the current, whereas membrane-permeant activators of protein kinase A (PKA) were without effect. PGE₂ did not acutely up-regulate the current. Conversely, both PGE₂ and PKA activation up-regulated the major TTX-r Na⁺ current, Na_V1.8. Extracellular ATP up-regulated the persistent current with an average apparent K_d near 13 μm, possibly consistent with P2Y receptor activation. Numerical simulation of the up-regulation qualitatively reproduced changes in sensory neurone firing properties. The activation of PKC appears to be a necessary step in the GTP-dependent up-regulation of persistent Na⁺ current.

Many small diameter dorsal root ganglion neurones (DRG) generate a tetrodotoxin-resistant (TTX-r) Na⁺ current with very slow kinetics operating over a membrane potential range some 20 mV or more negative than the transient Na⁺ currents generated in the same cell (Cummins et al. 1999; Baker & Wood, 2001; Dib-Hajj et al. 2002). It is usually referred to as the persistent current and was first recorded by Cummins et al. (1999) in Na_V1.8 null neurones, where the major portion of the TTX-r Na⁺ current is missing. The current is attributed to the Na_V1.9 (NaN) gene known to be expressed in peripheral nociceptive neurones (e.g. Cummins et al. 1999; Fjell et al. 2000; Fang et al. 2002), and in myenteric neurones (Rugiero et al. 2002; Rugiero et al. 2003). Consistent with a possible role in peripheral nociception, the channel has been immunocytochemically localized to unmyelinated nerve endings in the cornea (Black & Waxman, 2002). Recently, the DRG neurones of an Na_V1.9 null mutant have been characterized as apparently missing the persistent TTX-r Na⁺ current, and the animals show deficits in pain behaviour during the second phase of the intraplantar formalin test, suggesting that the current is necessary for prolonged activation of peripheral nociceptive endings (Priest et al. 2005).

Although the activation kinetics of the persistent current are too slow for it to contribute to the up-swing of an action potential, it can participate in the control of excitability because it is regenerative and operates at subthreshold membrane potentials. The current exhibits wide activation/inactivation gating overlap (e.g. Herzog et al. 2001), and when up-regulated can generate steady-state currents large enough to depolarize the membrane potential (Herzog et al. 2001; Baker et al. 2003). The non-hydrolysable analogue of GTP, GTP-γ-S, strongly up-regulates the current leading to spontaneous action potential generation at potentials near −60 mV (Baker et al. 2003). The magnitude of this up-regulation suggests that G-protein pathway modulation of the current may be functionally important. Modulation probably involves a phosphorylation step, because it is inhibited by the non-selective protein kinase inhibitor, H7 (Baker et al. 2003), although the kinase involved has remained uncharacterized. Substantial protein kinase A- and C-dependent modulation has been demonstrated for Na_V1.8, where kinase A activation causes an increase in macroscopic current amplitude and makes the voltage dependence of activation steeper, shifting the midpoint in the hyperpolarizing direction (England et al. 1996; Gold et al. 1998; Fitzgerald et al. 1999). Consensus site serines that have been defined in a heterologous system (Fitzgerald et al. 1999) are involved in these effects. Block of background kinase C activity has been found also to prevent kinase A-dependent up-regulation, whereas blockers of kinase A had no effect on the protein kinase C dependence of Na_V1.8 (Gold et al. 1998).

G-protein-coupled ATP receptors on sensory nerve may cause changes in excitability lasting longer than those attributable to the activation of P2X receptors (e.g. Molliver et al. 2002; Stucky et al. 2004). Molliver et al. (2002) have found that selective activation of P2Y receptors can reproduce a prolonged period of action potential firing in cultured sensory neurones observed when applying ATP, although the mechanism causing the hyperexcitability is mysterious. One possibility is that the hyeperexcitability may ensue through the up-regulation of persistent Na⁺ current. The present experiments help to test this hypothesis by delineating the kinase type involved in persistent Na⁺ current up-regulation and by determining whether the current can be controlled by ATP.

Methods

Tissue preparation and culturing procedures have been previously described (e.g. Baker et al. 2003; Baker & Bostock, 1997). The adult mice and P21 Sprague-Dawley rats used were killed in accordance with Home Office guidelines by cervical dislocation. Neurones were used within 1 or 2 days of initially plating out on poly l-lysine coated coverslips.

Solutions and electrophysiology

These experiments utilized the conventional whole-cell patch-clamp technique with membrane patch rupture and also used amphotericin-B perforated patches. The solutions for both conventional whole-cell voltage-clamp and current-clamp are detailed in Baker et al. (2003) with a modification in some of the present experiments, where 1-oleoyl-2-acetyl-sn-glycerol (OAG) was applied during voltage-clamp recording from wild-type (WT) neurones, EGTA was equimolarly replaced by BAPTA in the pipette solution. Normal whole-cell voltage-clamp experiments utilized the same reduced [Na⁺] external solution as detailed previously and also below for perforated-patch recordings. For voltage-clamp using the perforated-patch technique, the solutions used were the following (mm). External: NaCl 43.3, tetraethylammonium chloride 96.7, Hepes 10, CaCl₂ 2.1, MgCl₂ 2.12, 4-aminopyridine 0.5, CsCl 10, KCl 7.5, CdCl₂ 0.1, TTX 0.25 × 10⁻³. Pipette: CsMetSO₄ 110.0 mm, CsCl 30.0 mm, MgCl₂ 1.0 mm, CaCl₂ 0.1–5.0 mm, Hepes 10 mm. Both solutions were buffered to pH 7.2–7.3 with the addition of CsOH. The pipette solution contained Ca²⁺ (0.1–5 mm) for two reasons. Firstly, the presence of Ca²⁺ seemed to improve seal formation, and secondly, the recording was lost following patch breakdown, helping to prevent recording during accidental dialysis of the cytoplasm. When almost simultaneous current-clamp and voltage-clamp recordings were made using the perforated-patch technique (see Fig. 6), CsMetSO₄ and CsCl were equimolarly replaced with KMetSO₄ and KCl, respectively, and the extracellular current-clamp solution detailed in Baker et al. (2003), with a normal external [Na⁺] was used. For perforated-patch recordings, electrode tips were filled with pipette solution without amphotericin-B, and subsequently solution containing amphotericin-B (final concentration of 200 μg ml⁻¹) was used to fill the electrode. In current-clamp experiments, the aim was to find whether persistent Na⁺ current up-regulation could alter excitability, measured in terms of applied depolarizing current, and an estimate of voltage threshold. In order to maximize the effects of the current on excitability, the neurones were routinely held at −80 or −90 mV, so that the amount of resting inactivation of the persistent Na⁺ current was low. This was achieved by applying a polarizing current, adjusted from time to time to maintain a steady membrane potential value.

Figure 6. Negative shift in voltage threshold and the simultaneous up-regulation of a low-threshold inward current.

A, current-clamp (perforated-patch) recordings in control solution and at 10 min following exposure to 1 μm external PMA in a WT neurone (left and right hand panels, respectively). Neurone held near −80 mV, before step depolarizations. Smooth lines are exponentials drawn according to best-fit parameters, whose trajectory approximates the passive membrane response to depolarization. An increase in excitability results in a negative shift in voltage threshold, from −43 to −52 mV (indicated by arrows), and an increase in the latency of the action potential generated by a near threshold depolarization. The estimated voltage-threshold in the control recording is indicated on the action potential in the right hand panel as a short line. B, voltage-clamp recordings from the same neurone as in A. Under control conditions, no low-threshold inward current was observed (left-hand panel). After exposure to PMA, a persistent inward current contributes at negative potentials (change in current is plotted, right-hand panel). With a voltage-clamp step to −50 mV, inadequacies in the voltage-clamp prevented control of the Na⁺ current and the current became regenerative (arrow).

Electrodes were made from thin-walled glass (Harvard Apparatus, Edenbridge, Kent, UK), and had a resistance of between 1.5 and 3 MΩ when initially filled with the pipette solution. The voltage-clamp protocols were generated using an Axopatch 2B amplifier (Axon Instruments, Union City, CA, USA) controlled by a PC running pCLAMP 9 (Axon Instruments). Series-resistance compensation was set near 70%, with a nominal feed-back lag of 12 μs for conventional whole-cell recordings and 20 μs for perforated-patch recordings. Recordings were filtered at 5 kHz (4-pole Bessel filter) and usually digitized at 10 kHz. For whole-cell recordings the holding potential was −110 mV, and incrementing clamp steps were preceded by a 20 ms duration prepulse to −130 mV (unless otherwise stated). The responses to three consecutive families of clamp steps were averaged. Leak subtraction was achieved on-line, using a P/N protocol, where five reverse polarity clamp steps were averaged to generate the leakage record. Each interstep interval was 1.5 s and the protocol repeat frequency for a 16 step protocol was 0.04 Hz. When using both the Cs⁺ and K⁺ perforated-patch pipette solutions, a junction-potential existed at the end of the pipette, measured as +13 mV. The junction potential was backed off at the start of recording to zero the current, but the offset disappeared on seal formation, so that subsequent holding and test potentials had to be corrected for the initial overcompensation. Following stable seal formation, the series resistance was repeatedly estimated until it fell to 10 MΩ or below (before compensation). In order to minimize problems in voltage clamp associated with the high series resistance of the perforated-patch recording, the size of the Na⁺ currents were normally reduced by using the reduced external [Na⁺] solution described above. Recordings in which Na⁺ current activation was not gradual with incrementing depolarization when using perforated patches, suggesting an inadequate voltage clamp, were excluded from the analysis. Conventional whole-cell recordings exhibited a substantially lower series resistance usually close to 3 MΩ before compensation.

Materials

Reagents were obtained from Sigma-Aldrich (Poole, Dorset, UK), except phorbol 12-myristate 13-acetate (PMA), which was purchased from Tocris Cookson (Bristol, UK), tetrodotoxin (TTX), purchased from Alomone Laboratories (TCS Biologicals, Botolph Claydon, Bucks, UK), and PKC 19–36, purchased from Calbiochem (Merk Biosciences Ltd, Nottingham, UK). PMA, ATP and PGE₂ solutions were applied by gravity fed local superfusion. 1-Oleoyl-2-acetyl-sn-glycerol (OAG) was incorporated into the pipette solution.

Data analysis

Peak sodium current values were converted to conductance values using the equation:

where G is the conductance, I the maximum current amplitude, V_m the membrane potential, and E_Na the equilibrium potential value for Na⁺, given in the text. The Boltzmann equation was used to describe steady-state activation:

G_max is the peak conductance, V_½ is the potential at half maximal activation, and a_g is the slope factor in mV.

Where possible, data are presented as mean values ±s.e.m.

Simulations

Simulations were of a space-clamped neurone soma (run in Visual Basic on a PC, employing the Euler method with an integration time step of 1 μs), using equations and voltage-dependent kinetic parameter values modified from Herzog et al. (2001) and from Bostock et al. (1991). Three forms of Na⁺ current (TTX-s, Na_V1.8 and persistent Na⁺ current) and two types of K⁺ current were included in the simulation (G_KF, kinetically fast K⁺ channels and G_KS, kinetically slow K channels). The voltage-activated K⁺ currents were assumed to be similar to those in myelinated axons, with shifted voltage dependence, matching the relatively depolarized voltage dependence of transient Na⁺ currents in small diameter sensory neurones. In the absence of a full kinetic analysis of Na_V1.8, the voltage-dependent rate constants were chosen to give currents in voltage clamp exhibiting plausible kinetics and voltage dependence, and in order to generate an action potential exhibiting an inflection. The forms of the equations used to derive the rate constant values were taken from the analysis of TTX-sensitive current of Herzog et al. (2001). The voltage-dependent parameters provided by Herzog et al. (2001) for persistent Na⁺ current were modified to more closely match Na⁺ current recordings made using conventional whole-cell methods, without internal F⁻, where the current exhibits a less negative voltage dependence than that reported by Dib-Hajj et al. (2002). Simulated TTX-r Na⁺ currents, which provide most of the inward Na⁺ current, are compared with real currents in Fig. 8. Conductance values of TTX-s Na⁺ current and Na_V1.8 were chosen to make the fractional conductance attributable to Na_V1.8 substantially larger than that of TTX-s Na⁺ current, appropriate for a nociceptive neurone, and analogous to nociceptive endings in the cornea and dura that retain their excitability even in the presence of micromolar concentrations of TTX (Brock et al. 1998; Strassman & Raymond, 1999). A minor fraction of the total TTX-s Na⁺ conductance, equivalent to 2% of the transient conductance, was incorporated as a non-inactivating persistent conductance, with an activation voltage dependence 20 mV more negative than for the transient current (Kiernan et al. 2003; Bostock & Rothwell, 1997). The up-regulation of persistent Na⁺ current was achieved simply by adding the current to a neurone first simulated without it. When the persistent conductance was set to 100 nS, it was sufficient to generate close to 4 nA peak current (i.e. in response to a voltage-clamp step to −10 mV) in simulated voltage-clamp with a one-third Na⁺ gradient. This is similar to a reasonably large example of a real up-regulated current (e.g. Fig. 5A in Baker et al. 2003). Using either 15 pF or 30 pF for the neuronal capacitance did not qualitatively affect the result of incorporating a persistent Na⁺ conductance. The simulation was run using room temperature parameter values.

Figure 8. Characteristics of Na⁺ currents incorporated into numerical simulation.

A, left panel, typical example Na_V1.8 currents recorded in the presence of TTX and simulated currents (right panel). Simulated Na⁺ currents without and with residual fast K⁺ current (upper and lower family of voltage-clamp traces, respectively). B, the corresponding peak I−V plot for simulated currents (without K⁺currents) using a one-third physiological Na⁺ gradient. C, example persistent Na⁺ currents recorded in Na_V1.8 null in the presence of TTX, and simulated currents. Real currents show some residual delayed rectification that remains following exposure to K⁺ channel blockers. D, corresponding peak I−V plot for simulated currents. E, G/G_max (activation, •) and h_∞ (inactivation, ○) curve for simulated Na_V1.8. The values of h_∞ were calculated directly from rate constants, whereas the activation curve was obtained by measuring the peak currents generated in a voltage-clamp protocol with a negative holding potential, and converting them to conductances. F, activation and inactivation curves for persistent Na⁺ current, G/G_max (•) and h_∞ (▴). The effect of inactivation rate on peak conductance can be observed at potentials more positive than −10 mV. The effects of ultra-slow inactivation on persistent Na⁺ current were not included in the simulation.

Figure 5. PGE₂ exposure and activation of PKA acutely up-regulates Na_V1.8.

A, superfusion of 5 μm PGE₂ caused an up-regulation of Na_V1.8 amplitude, recorded at +10 mV. Similar results were obtained in 6 neurones. B, peak current-voltage relations for the same neurone in A, recorded before (•) and after (○) exposure to PGE₂ (i.e. immediately before and after making the recordings for A). C, plot of normalized peak conductance values against membrane potential, derived from the current values in B, before (•) and after (○) exposure to PGE₂. Smooth curves are Boltzmann relations with maximal amplitudes of 1 and 1.62, and half-maximal activation potentials of −13.34 and −17.82 mV before and after exposure to PGE₂, respectively. Conductances were derived assuming E_Na to be +60 mV. D, normalized peak conductance versus membrane potential plot, before and during exposure to 8Br-cAMP (1 mm); same cell as in Fig. 4E and F. Smooth curves are Boltzmann relations with maximal amplitudes of 1 and 1.3 and half-maximal activation potentials of −4.98 and −11.32 mV before and after exposure to 8Br-cAMP. E_Na was assumed to be +62 mV. Data points were not corrected for residual series-resistance errors. All recordings made using perforated patches.

The transmembrane ionic currents were described by the following equations:

(1)

(2)

where I_ion is the total transmembrane ionic current and I_cap is the capacitative current, E_m is the membrane potential, R_in is the input resistance, E_Na and E_K are the Na⁺ and K⁺ equilibrium potentials, respectively. Equation (2) holds true in a space-clamp, without longitudinal currents. m and h are the voltage-dependent activation and inactivation gating parameters for the TTX-s Na⁺ current (Hodgkin & Huxley, 1952) and likewise m₈ and h₈ are the voltage-dependent activation and inactivation gating parameters for Na_V1.8, and m₉ and h₉ are for the persistent Na⁺ current. m_P is the activation gating parameter for the persistent TTX-s Na⁺ current. n_F and n_S represent the activation gating parameters for the fast and slow K⁺ currents, respectively. G_Na, G_NaP, G_Na8, G_Na9, G_KF and G_KS represent the maximum conductance for transient TTX-s Na⁺ current, persistent TTX-s Na⁺ current, Na_V1.8, persistent Na⁺ current, fast K⁺ current and slow K⁺ current, respectively.

For simulating the effects of up-regulating persistent Na⁺ current on a neurone held at −90 mV in current-clamp (cf. Baker et al. 2003), the following fixed parameter values were used: cell capacity = 15 pF; R_in= 200 MΩ; E_K=−85 mV; E_Na= 79.6 mV; G_Na= 18 nS; G_NaP= 0.36 nS; G_Na8= 220 nS; G_Na9= 0 or 100 nS; G_KF= 30 nS; G_KS= 2 nS. In addition, a polarizing current of 454 pA was applied to maintain a holding potential of −90 mV. This was for two reasons. Firstly, this removes resting inactivation from the persistent Na⁺ current, maximizing its subsequent effects on excitability, and secondly, it replicates what was done in Baker et al. (2003). In an attempt to simulate the generation of spontaneous activity by up-regulating persistent Na⁺ current in a resting neurone (seen in current-clamped neurones at membrane potentials close to −60 mV; Baker et al. 2003), R_in was increased to 2 GΩ, giving a resting potential of −58.4 mV. The G_Na9 conductance was increased up to 38 nS, and a hyperpolarizing ‘pump-current’ (indistinguishable from a small polarizing current) was simultaneously added. All other parameter values remained unchanged. The values of m, h, m_P, m₈, h₈, m₉, h₉, n_F and n_S were calculated from voltage-dependent rate constants α_x and β_x, where x indicates any single gating parameter. α_m, β_m, β_h, α_mP, β_mP, α_m8, β_m8, β_h8, α_m9, β_m9, α_h9, β_h9 were calculated according to the formula: A/(1 + exp((E_m+B)/C). α_h and α_h8 were given by Aexp(−(E_m+B)/C). α_nF and α_nS were given by A((E_m+B)/(1 − exp((−B −E_m)/C), and β_nF and β_nS were given by A(−B−E_m)/(1 − exp((E_m+B)/C). The values of A, B and C used are indicated in Table 1.

Table 1.

Values of A, B and C used to derive the values of voltage-dependent rate-constants

	A (ms−1)	B (mV)	C (mV)
αm	17.235	7.58	−11.47
βm	17.235	66.2	19.8
βh	10.8	−11.8	−11.998
αmP	17.235	27.58	−11.47
βmP	17.235	86.2	19.8
αm8	3.83	2.58	−11.47
βm8	6.894	61.2	19.8
βh8	0.61714	−21.8	−11.998
αm9	1.548	−11.01	−14.871
βm9	8.685	−112.4	22.9
αh9	0.2574	63.264	3.7193
βh9	0.53984	0.27853	−9.0933
αh	0.23688	115	46.33
αh8	0.013536	105	46.33
αnF	0.00798	72.2	1.1
αnS	0.00122	−10.5	23.6
βnF	0.0142	55	10.5
βnS	0.000739	57.1	21.8

Results

PKC inhibition prevents current up-regulation

Incorporating the non-selective protein kinase inhibitor H7 into the intracellular solution also containing GTP-γ-S and ATP, slowed and reduced the up-regulation of the persistent Na⁺ current in Na_V1.8 null neurones in conventional whole-cell voltage-clamp (Baker et al. 2003). A high concentration of kinase inhibitor (200 μm) was necessary because of the simultaneous exposure of the neuronal contents to GTP-γ-S. Subsequent experiments are consistent with pre-exposure of WT neurones to 20 μm H7, for at least 30 min prior to recording, also attenuating current up-regulation (Fig. 1A and B). In order to define more rigorously whether protein kinase C (PKC) is involved in up-regulating persistent Na⁺ current, the selective pseudosubstrate inhibitor PKC19–36 was incorporated into the intracellular solution along with GTP-γ-S and ATP. (In conventional whole-cell voltage-clamp recordings from Na_V1.8 null neurones, data are presented only from neurones in which a persistent Na⁺ current could be perceived from near the start of recording.) The increase in current amplitude with GTP-γ-S and ATP only was close to 300% at 4–5 min in Na_V1.8 null neurones and this increase was abolished by adding the peptide inhibitor (Fig. 1C), providing evidence that at least one PKC isoform is involved (P= 0.035, 2-way repeated measures ANOVA, total n = 10, n = 6 without and with PKC19-36, respectively, data collected at 30 s intervals; P < 0.01 at 2 min and P < 0.05 at 4.5 min, one-tailed t test). The peptide inhibitor was present at a maximum concentration of 5 μm, although 500 nm also appeared effective. A 10 μm concentration of the inhibitor is reported to have a selective functional effect on PKC pathways modulating Na_V1.8 (Gold et al. 1998).

Figure 1. Kinase inhibition blocks the up-regulation of persistent Na⁺ current.

A, recording of TTX-r Na⁺ currents, made immediately after attaining the whole-cell configuration in a wild-type (WT) neurone, which includes a low-threshold persistent portion. Heavy trace indicates current activated by a step to −30 mV. An interpretation of the currents evoked by depolarization (right) indicates that the current at −30 mV and more negative is largely persistent Na⁺ current. B, Na⁺ currents recorded in WT neurones in response to voltage-clamp steps to −30 mV. Dramatic up-regulation of the current amplitude followed the introduction of 500 μm GTP-γ-S to the neurone interior (by over 6 times within 5 min, same neurone as in A). After pre-exposure to 20 μm H7 in another neurone with a very similar initial current, up-regulation was near 150%. C, in Na_V1.8 null neurones, an intracellular solution containing GTP-γ-S and ATP (filled bars) gave rise to a substantial up-regulation of peak persistent current (I, measured at −10 mV or −20 mV, whichever was the larger) at 2 and 4.5 min (n = 10 and 8, respectively), in comparison to the initial current amplitude (I₀). The addition of 500 nm to 5 μm PKC19–36 to the intracellular solution with GTP-γ-S and ATP (open bars), blocked the up-regulation at 2 and 4.5 min (n = 6 and 4, respectively), consistent with a role for PKC. (*P < 0.05, **P < 0.01, one-tailed, unpaired t test.)

Activators of PKC, but not PKA, cause persistent current up-regulation

Confirmation of the involvement of PKC was obtained in conventional whole-cell patch-clamp experiments in WT neurones, incorporating a selective activator of PKC in the internal solution. When the activator 1-oleoyl-2-acetyl-sn-glycerol (OAG, 25 μm) (Fig. 2) was present in addition to 3 mm ATP and 500 μm GDP, the persistent current amplitude and density measured during a voltage-clamp step to −40 mV (and more negative potentials) was up-regulated. This result was not found with GDP and ATP alone. In Fig. 2A can be seen TTX-r currents initially dominated by Na_V1.8, but after exposure to 25 μm internal OAG, the low threshold, persistent portion of the current was selectively up-regulated. The enhancement of the most negatively activating currents can be seen in the current-voltage (I−V) plots shown in C, appropriate for persistent Na⁺ current up-regulation. Difference currents are clearly similar to persistent Na⁺ currents, consistent with the up-regulation of persistent Na⁺ current, with minimal changes in Na_V1.8 (Fig. 2B and D). In other WT neurones, the amplitude of the TTX-r current evoked at −40 mV was measured repeatedly over 5 min (Fig. 2E). Measurements from neurones that subsequently exhibited a persistent current and also those that did not were included in the analysis, because at these negative potentials and with internal GDP, it was initially difficult to tell whether a persistent current was present. The group of neurones exposed to OAG fell into two populations, either exhibiting up-regulation, or exhibiting no up-regulation (n = 5 and n = 4, respectively). These two groups differed significantly in their current densities (P= 0.002, two-way repeated measures ANOVA), a statistically significant difference being apparent after 30 s recording time, consistent with the presence of persistent current in only a fraction of neurones studied. I do not interpret the low initial current density (derived from the maximum inward current measured on the trace), or that found throughout in the presence of internal GDP, as an indication of the presence of a low level of persistent current in all neurones; these small values almost certainly fall within baseline noise. This result was expected because expression of Na_V1.9-like immunoreactivity occurs in only a subpopulation of sensory neurones (e.g. Fang et al. 2002). The frequency with which up-regulated persistent current was recorded (in 5 of 9 neurones exposed to 25 μm OAG) was significantly different from that obtained with GDP alone (P < 0.03, Fisher's exact test; Fig. 2E).

Figure 2. OAG in the pipette solution with GDP + ATP causes persistent Na⁺ current up-regulation in WT neurones.

A, families of TTX-r Na⁺ currents evoked by the inset voltage-clamp protocol. At −40 mV, a potential more negative than the activation threshold for Na_V1.8, a persistent current was up-regulated over 5.5 min (dark traces). B, subtraction of initially recorded currents from those recorded later gives the component that was up-regulated, corresponding with persistent Na⁺ current. C, I−V plots for peak Na⁺ current at the start of recording (•), 2 min (○), 5.5 min (▴) and 7.5 min (▵) into the voltage-clamp. D, I−V plots for initial peak Na⁺ current (•), and for that component up-regulated at 5.5 min (▴) and 7.5 min (▵). E, no up-regulation of current at −40 mV with GDP + ATP was apparent in 9 neurones (left hand panel), whereas 5 of 9 neurones with 25 μm internal OAG exhibited an up-regulation at −40 mV (right hand panel; P < 0.03, Fisher's exact test). In E those neurones showing no clear up-regulation in the right-hand panel (○) are presumably without functional persistent Na⁺ current.

Subsequently, using the perforated-patch technique to make recordings from both Na_V1.8 null and WT neurones, it was found that the membrane-permeant PKC activator phorbol 12-myristate 13-acetate (PMA) up-regulated the persistent current (n = 5), whereas 8-bromo-cyclic AMP (8Br-cAMP) and dibutyrylcyclic AMP (dbcAMP)(membrane permeant analogs of cAMP) did not (n = 4, when applied alone), even when tested in neurones also found to be responsive to PMA (n = 2; Fig. 3). The apparent changes in reversal potential during an experiment (e.g. Fig. 3A) may be partly caused by the progressive block of residual K⁺ currents, associated with the diffusion of Cs⁺ ions into the neurones. These data indicate that the up-regulation of persistent Na⁺ current may specifically require activation of PKC. In response to exposure to 100 nm PMA, current up-regulation in Na_V1.8 null neurones occurred gradually, with a near 65% average increase in maximal amplitude over 5 min (Fig. 3A and D). There was little change in persistent current activation voltage dependence for the neurone in Fig. 3A, even when the current amplitude doubled between 5 and 25 min exposure (Fig. 3B and C). The less than 6 mV apparent hyperpolarizing shift in the half-maximal activation potential can possibly be explained by the effect of residual K⁺ currents becoming proportionally less as the Na⁺ current amplitude increased.

Figure 3. PMA up-regulates persistent Na⁺ current, whereas membrane permeant activators of PKA do not.

A, peak current–membrane potential relation for persistent Na⁺ current in Na_V1.8 null neurone before (•) and during exposure to 100 μm PMA for up to 5 min (○), 15 min (▴) and 25 min (▵). Any change in apparent reversal potential on exposure to PMA is caused by the up-regulation of persistent Na⁺ current recorded against residual K⁺ currents, and it does not represent a true change in reversal potential for Na⁺. B, current amplitude values at 5, 15 and 25 min in A were converted to conductances, assuming E_Na to be +60 mV, normalized to the maximal conductance at 5 min, and plotted against membrane potential. The maximal conductance doubled between 5 and 25 min. C, normalized conductance values for 5 and 25 min, superimposed on Boltzmann relations, drawn with best-fit parameters, and giving half-maximal activation potentials of −32.6 and −38.3 mV, respectively. D, 100 μm PMA caused up-regulation of persistent Na⁺ current in Na_V1.8 null neurones (n = 5) where data were obtained following an exposure lasting close to 5 min. Data normalized to maximal pre-exposure inward current for each neurone. E, exposure of a WT neurone to 1 mm db-cAMP does not up-regulate current amplitude at −40 mV, but 1 μm PMA causes an immediate and profound up-regulation (n = 2 where both agents were applied). F, I−V relation for persistent Na⁺ current in an Na_V1.8 null neurone before, during and after (•, ○, ▵, respectively) exposure to 1 mm 8Br-cAMP for 2.5 min. Similar recordings with 500 μm 8Br-cAMP for up to 3 min were also without effect (total n = 3). G, peak TTX-r inward current densities in WT neurones, measured at −40 mV, before, following several minutes exposure to 1 mm 8Br-cAMP, and during wash (n = 3). The PKA activator does not appear to up-regulate persistent Na⁺ current. All recordings made using perforated patches.

In contrast, in WT neurones intracellular OAG (cf. Fig. 2) or extracellular PMA (100 nm, n = 4; 1 μm, n = 3) had a small effect on the peak amplitude of the major TTX-r Na⁺ current, Na_V1.8 (Fig. 4). There was a statistically significant increase in Na_V1.8 current amplitude (largest individual increase in maximal currents was 24%) with 100 nm PMA (P= 0.004, repeated measures 2-way ANOVA), between −10 and +40 mV (P < 0.05, post hoc Holm-Sidak test) although following exposure to 1 μm there was no significant change in current amplitude. Neither was there any obvious change in the voltage dependence of current recruitment after exposure to PMA. Superfusion of 8Br-cAMP gave a variable response (n = 4) from no effect to a substantial up-regulation (Fig. 4E and F) that was not reversed by washing. The gradual activation of Na_V1.8 with depolarization suggests that series-resistance errors across the perforated patches for presented data are likely to be small. That said, for Fig. 4E and F, taking the residual, uncompensated series resistance to be near to 1 MΩ for the during exposure data, the maximal series-resistance error at the peak would be expected to be close to 8 mV. However, the change in voltage errors between the control and exposure data is likely to be substantially less than this (probably less than 5 mV), because the voltage-clamped currents have similar amplitudes. Inadequacies in the quality of the voltage clamp caused by perforated-patch recording therefore cannot account for the pharmacological effect of PGE₂. This up-regulation is very similar to the effects reported by England et al. (1996) and Gold et al. (1996, 1998) and corresponds well with the effects of exposure to PGE₂ (n = 6; Fig. 5). PGE2 was without any acute effects on persistent Na⁺ current (see below). One possibility is that the hyperalgesic effects of PGE₂ are mediated largely or wholly through the functional modulation of Na_V1.8 alone, explaining the dramatic effect of antisense against Na_V1.8 on PGE₂ induced hyperalgesia in vivo (Khasar et al. 1998).

Figure 4. PMA exposure causes a small up-regulation in Na_V1.8 amplitude.

A and C, representative Na⁺ currents attributable to Na_V1.8 in WT neurones, and B and D, mean I−V relation (for 3 and 4 neurones, respectively) before and during superfusion of PMA at 1 μm (A) or 100 nm (C) over minutes (largest current amplitude values in both conditions before (•) and during (○) exposure, respectively). The values are normalized with respect to the maximal peak pre-exposure values. Larger increase in amplitude obtained with 100 nm PMA is a small but statistically significant change (P= 0.004, repeated measures, 2-way ANOVA), with the largest individual peak increase being close to 24%. E and F, superfusion of 1 mm 8Br-cAMP up-regulates Na_V1.8 in an example neurone. The peak current at 0 mV increased by 46% in minutes. F, I−V plots, for the same neurone in E, before (•), during exposure (○) and during washing (▵). All recordings made using perforated patches and data points are not corrected for residual series-resistance errors.

Increased excitability accompanies an increase in persistent inward current

To address the question as to whether exposure to PMA can give rise to increases in excitability by a mechanism consistent with the up-regulation of persistent Na⁺ current, perforated-patch recordings were made in voltage-clamp and current-clamp mode in the same neurone. In all perforated-patch experiments that included current-clamp recording, the extracellular solution was quasi-physiological, without TTX (Baker et al. 2003), and the intracellular solution contained KMetSO₄ (see Methods). In order to maximize the effects of the current on excitability, the neurones were routinely held at −80 or −90 mV by applying a holding current that was adjusted to maintain a steady membrane potential. Figure 6 includes recordings from an example neurone that show PMA-induced up-regulation of a low-threshold inward current, occurring simultaneously with a fall in the current threshold and a more negative voltage threshold for action potential firing. In six similar recordings, no increases in excitability were seen, and there was no perceptible up-regulation of a low-threshold inward current. This suggests that, at least in terms of changes in threshold measured from a negative holding potential, any effects on other transient Na⁺ currents operating in the small diameter neurones by PKC activation are negligible. Furthermore, when using either perforated-patch or conventional whole-cell recording, and a pipette solution containing a near physiological [K⁺], it was common to observe a fall in excitability over time. This was attributed to an increase in low-threshold K⁺ current possibly consistent with a gradually increasing intracellular [K⁺] in the neurone, although this phenomenon was not studied in detail.

Extracellular ATP up-regulates persistent current

Because intracellular proteins are not dialysed when using the perforated-patch recording method, acute functional study of potential endogenous activators of G-protein pathways becomes possible, as the second messenger pathways should remain intact. Exposure to ATP (5–150 μm) caused acute up-regulation of the persistent current (measured in voltage-clamp in the presence of extracellular TTX; Fig. 7A–C) in both Na_V1.8 null neurones (4 responded of 6 exposed) and WT neurones (6 of 9 exposed appeared to have an up-regulated persistent current). Thus the overall response rate for neurones already generating the current was 67%. The up-regulation was apparent on a time scale of seconds to minutes. Currents measured to produce the plot in Fig. 7C, are shown in Fig. 7D and E, where it is apparent that although there is substantial amplitude up-regulation, the voltage-dependent kinetics of the current are unchanged, consistent with minor or no changes in the voltage dependence of the persistent current, found also for PMA and GTP. In a WT neurone, the maximal current amplitude was measured at −40 mV (or more negative), potentials at which little or no contribution of Na_V1.8 would be expected. Based on cumulative concentration-response experiments, the effect of ATP exhibited an average apparent K_d near 13 μm. The largest up-regulation on exposure to a particular concentration of ATP was expressed as a fraction of that predicted to be the maximum possible after up-regulation, by fitting each individual data set by a Langmuir isotherm, and thus only data sets incorporating at least two ATP test concentrations could be included in the analysis (Fig. 7B). This analysis may underestimate the true potency of ATP for two reasons. Because these data were obtained by generating a cumulative concentration response to ATP, the up-regulation at the lowest concentrations tested may not have been fully complete before a higher concentration was applied (potentially reducing the apparent potency of low concentrations), and any tendency to spontaneous up-regulation during recording might increase the amount of apparent up-regulation caused by the higher concentrations, potentially skewing the estimate of K_d to a higher value.

Figure 7. Persistent Na⁺ current is up-regulated by exposure to external ATP.

A, TTX-r inward currents recorded at negative potentials in a WT neurone were up-regulated within a minute following exposure to a bolus of ATP introduced into the recording chamber (final concentration 50 μm). Left hand panel, and right hand panel before and after ATP introduction, respectively. B, increment in current amplitude (I_inc) on exposure to ATP (continuous superfusion) plotted as a fraction of maximum increase possible (I_{max inc}) in each neurone. The maximum increase for each neurone was found by fitting an isotherm to the raw amplitude values. Each symbol type represents an individual neurone. ○, ▴, ▵ indicate peak current values from Na_V1.8 null neurones and, ▪ from WT (in the latter the current amplitude was measured at −40 mV). Smooth curve is a Langmuir isotherm drawn according to parameters providing the best-fit to all data points, with an apparent K_d= 13 μm. C, example increases in Na⁺ current amplitude on exposure to increasing concentrations of ATP in an Na_V1.8 null neurone (exposure time indicated by the bars beneath the data points). D, increase in Na⁺ current amplitude measured at −28 mV, before and after exposure to 10 and 50 μm ATP in an Na_V1.8 null neurone (same as that in panel C). E, scaled version of control current in D (thin line), superimposed on current recorded in the presence of 50 μm ATP (thick line). No obvious changes in current kinetics occurred as the current amplitude increased, consistent with unchanged gating voltage dependence.

PGE₂ is known to up-regulate Na_V1.8 (e.g. Gold et al. 1996, 1998), on a time scale of seconds to minutes and this mechanism has been proposed to participate in hyperalgesia following inflammation. The actions of PGE₂ include effects on the amplitude and voltage dependence of Na_V1.8 and are most similar to the effects of directly activating PKA. The action of PGE₂ is mimicked by activators of PKA and blocked by PKA inhibition (England et al. 1996; Gold et al. 1998). PGE₂ exposure is also reported to have a chronic up-regulatory effect on the persistent Na⁺ current amplitude, when neurones are incubated in the presence of 10 μm PGE₂ for 1 h (Rush & Waxman, 2004), through an unknown mechanism involving G_i/o. The effects of micromolar range PGE₂ were therefore tested on persistent Na⁺ current, and no clear effects of any kind were seen in response to acute exposure over 5 min (n = 12).

Numerical simulation qualitatively reproduces excitability changes seen with persistent current up-regulation

Example real and simulated TTX-r Na⁺ currents recorded in voltage-clamp are shown in Fig. 8A and C (Na_V1.8 and persistent Na⁺ current, respectively). The peak-current-voltage relations for the two types of TTX-r current are shown in Fig. 8B and D. Simulated Na_V1.8 current activates around −30 mV, whereas the persistent Na⁺ current activates close to −60 to −65 mV (cf. Baker & Wood, 2001). In Fig. 8E and F are shown the simulated activation and h_∞ curves for Na_V1.8 and persistent Na⁺ current. Ultra-slow inactivation gating of the persistent Na⁺ current (e.g. Cummins et al. 1999; Herzog et al. 2001) was not included in the simulation, and this can be justified because the aim of the simulation was to help test the hypothesis that up-regulation of persistent Na⁺ current might be able to replicate the change in firing properties seen in current-clamped neurones, and also whether its up-regulation could cause spontaneous activity. Cleary ultra-slow inactivation would be expected to play a role in curtailing firing or any increase in excitability occurring as a result of persistent Na⁺ current up-regulation.

Action potentials and associated action currents in a simulated neurone normally held at −90 mV are shown in Fig. 9A. The neurone was held at −90 mV by passing a continuous hyperpolarizing current, clearly visible at the start and end of the current traces. The results of the simulation replicated experimental findings quite well (Baker et al. 2003). Up-regulation of persistent Na⁺ current is associated with a fall in current threshold, the prolongation of subthreshold depolarizations beyond the offset of an applied current, and an increase in the latency of an action potential evoked by a just-suprathreshold depolarization. Up-regulation of the current is also associated with repetitive firing at membrane potentials substantially more negative than the voltage threshold for action potential induction before current up-regulation (Fig. 9B). This present model was also modified to produce sustained firing in a neurone not receiving external stimulation. Before adding the persistent Na⁺ current, increasing the neurone input resistance from the initial 200 MΩ to 2 GΩ gave a membrane potential of −58.4 mV, without the application of any polarizing current, and this was taken as the resting state. If the persistent Na⁺ current only was subsequently included, the membrane potential was found to stabilize at a new, low value within the range of persistent current activation-inactivation overlap, and sustained firing was not seen. However, with the simultaneous addition of a hyperpolarizing ‘pump-current’, sustained firing occurred (Fig. 9C), providing at least a qualitative confirmation that persistent Na⁺ current may be able to drive spontaneous activity in a living neurone, but perhaps only where its tendency to depolarize is countered by the Na⁺ pump.

Figure 9. Simulated neurone exhibits characteristic changes in firing properties on up-regulation of persistent Na⁺ current.

A, just sub- and just supra-threshold responses from a control simulated neurone, lacking persistent Na⁺ current, being held at −90 mV (left hand panel). The action potential was evoked within the 10 ms stimulus duration. Just sub- and just supra-threshold rectangular, applied depolarizing currents, lasting 10 ms are shown inset, where the larger current represents a change in holding current of 400 pA. Beneath are plotted the total transmembrane currents for both responses (bottom of left hand panel). The supra-threshold response includes a regenerative Na⁺ current. Addition of persistent Na⁺ current with a maximal conductance of 100 nS dramatically reduced the applied current threshold and the action potential was elicited with a more negative voltage threshold (right hand panel). The regenerative inward current (and the action potential upswing) occurs after the applied depolarizing current has ceased (bottom of right-hand panel). B, up-regulation of persistent Na⁺ current gives rise to repetitive firing in response to a prolonged, just-supratheshold depolarization, a response not seen in the control neurone lacking the conductance. Initially the membrane potential was held at −90 mV by polarizing current and depolarization was brought about by a 34 pA change in holding current. The most negative interspike potential was close to −70 mV. C, simulation of a neurone with an input resistance of 2 GΩ, giving a resting membrane potential (without persistent Na⁺ conductance) of −58.4 mV (control). Addition of persistent Na⁺ conductance with a maximal value of 38 nS and with the simultaneous addition of a steady 60.5 pA hyperpolarizing ‘pump current’ produces a train of action potentials.

Discussion

The TTX-r Na⁺ channels present in primary afferents, Na_V1.8 and Na_V1.9, are both found in nociceptors (e.g. Akopian et al. 1999; Fjell et al. 2000; Fang et al. 2002). Functional evidence places TTX-r Na⁺ channels at unmyelinated nerve endings, where they are involved in action potential generation and propagation, indicative of a role certainly for Na_V1.8 (Brock et al. 1998; Strassman & Raymond, 1999). Immunohistochemical evidence indicates the expression of Na_V1.9 in the corneal leash fibres (Black & Waxman, 2002). One shared functional characteristic of the TTX-r Na⁺ channels in primary afferents is that they can be profoundly modulated by second messenger pathways, to an extent dramatic enough to be consistent with a contribution to inflammatory pain (e.g. Khasar et al. 1998). The mechanism by which membrane conductance is increased remains unresolved in both cases, but the data presented here show that PKC probably plays a crucial role in the up-regulation of persistent Na⁺ current, attributed to Na_V1.9. Protein kinase activation, especially PKA, underlies the up-regulation of Na_V1.8 (England et al. 1996; Gold et al. 1998; Fitzgerald et al. 1999). Gold et al. (1998) report that phorbol esters (including PMA) cause up-regulation of Na_V1.8, and that PKC activation is likely to contribute to the PGE₂-induced modification of Na_V1.8 Na⁺ currents, although the up-regulation of Na_V1.8 by exposure to PGE₂ is most similar to that brought about by PKA activation. While PKC may indeed contribute to the PGE₂-induced up-regulation of Na_V1.8, the effects of PMA reported here are small. The present experiments suggest that PGE₂ is without acute effects on persistent Na⁺ current, and this seems consistent with the lack of effect of PKA activation on the current.

Baker et al. (2003) reported that between 1 in 3 and 1 in 6 neurones exhibited persistent Na⁺ current, in voltage-clamp and current-clamp recordings, respectively. The currents are probably easier to spot in voltage clamp with the K⁺ channels blocked. The OAG experiments reported here found something like 1 in 2, quite similar to that expected from the immunocytochemical evidence (Fang et al. 2002). The apparent small discrepancy between the present voltage-clamp recordings and previous ones might indicate that internal application of OAG is very effective at up-regulating the current. A number of possibilities might explain the functional up-regulation of persistent Na⁺ current. These include an obligatory phosphorylation of the channels by PKC before they can conduct, the possibility that PKC is involved in facilitating channel trafficking to the membrane, or that phosphorylation by PKC increases the single channel conductance of already operable channels (for example by favouring a larger conductance state over another smaller one), or some combination.

Evidence exists implicating at least one isoform of PKC in the induction of painful states. Bradykinin sensitizes a heat-activated current in cultured DRG neurones, through a PKCɛ-mediated mechanism. The PKCɛ isoform exhibits membrane translocation in response to activation of the second messenger pathway (Cesare & McNaughton, 1996; Cesare et al. 1999). PKCɛ inhibition is reported to reduce (up to 50%) an enhancement of TTX-r Na⁺ current amplitude measured at −15 mV (in this case Na_V1.8) on exposure of cultured neurones to adrenaline (Khasar et al. 1999a, b) although it seems that the major modifier of the current in response to adrenaline exposure is PKA (Khasar et al. 1999b). The form of the enhancement appears very similar to that induced by PGE₂ (e.g. England et al. 1996; Gold et al. 1998) with an increase in the steepness of activation voltage dependence, leading to enhanced repetitiousness in current-clamp experiments.

Up-regulation of persistent Na⁺ current and the consequences for nociceptor excitability

The function of the persistent Na⁺ current has been controversial, because it begins to activate over a negative range of membrane potentials in sensory neurones (initially thought to be near −70 mV, but now reported to be between −50 and −70 mV; see Dib-Hajj et al. 2002; Rugiero et al. 2003; Baker et al. 2003) and with a negative steady-state inactivation voltage dependence, leading to the suggestion that the current is largely inactivated at the resting potential and may therefore be without function. Rugiero et al. (2003) have provided evidence that this is not the case in myenteric neurones, when recordings are made without internal F⁻. These authors recorded persistent Na⁺ current beginning to activate near −50 mV without intracellular F⁻, but at −65 mV with intracellular F⁻, suggesting that the very negative activation voltage dependence originally reported for DRG neurones may be unphysiological, and related to the presence of a high internal [F⁻] (cf. Cummins et al. 1999; Dib-Hajj et al. 2002). Rugiero et al. (2003) also defined the voltage dependence of the normal and ultra-slow inactivation of persistent Na⁺ current in the absence of F⁻, and their analysis indicates that steady-state inward currents are expected in myenteric neurones at −50 mV. In other experiments without intracellular F⁻, the steady-state component of the up-regulated current has been found to be substantial enough to depolarize DRG neurones by more than 10 mV starting from a resting potential near −60 mV (Baker et al. 2003). While moderate depolarization can increase excitability and generate action potentials, prolonged and strong depolarization enhances the resting inactivation of transient Na⁺ channels, leading to another speculation that the role of persistent Na⁺ current is to reduce excitability at nerve endings by inactivating transient Na⁺ channels (Herzog et al. 2001). However, persistent Na⁺ channels seem to exist in operable and initially inoperable pools, or more than one functional state, allowing for receptor activation-dependent up-regulation of the current, and the possibility that in the quiescent state, the channels are not operating to depolarize the membrane.

The real resting potentials of C-fibres remain unknown, but strong indirect evidence from the velocity recovery cycles of C-fibres in human skin (Bostock et al. 2003) shows that at rest few voltage-dependent K⁺ channels can be open, because activity-dependent increases in supernormality (attributable to electrogenic hyperpolarization) are not accompanied by a change in the membrane time constant. The resting potential of human C-fibres thus cannot be low and may be remarkably dependent on active pump current, rather than a resting K⁺ conductance. On this basis, the normal membrane potential estimates (around −60 mV) made in the soma (e.g. Wang et al. 1994) may either be representative or perhaps less negative than the true membrane potential in C-fibres and persistent Na⁺ channels might be able to elicit increases in excitability and depolarize the membrane potential in a manner dependent on the activation of G-protein coupled receptors. However, a different and apparently contradictory insight into the resting potentials of unmyelinated nerve endings is provided by the experiments of Carr et al. (2002) on the behaviour of Na⁺ channels in cold-sensitive and polymodal endings in the guinea pig cornea. All endings appeared to be chronically depolarized, with substantial resting inactivation of transient Na⁺ current, that could be reversed by passing hyperpolarizing current from an electrode. In fact, the inward Na⁺ current (in polymodal receptors being largely but not exclusively TTX-r) could be approximately doubled in amplitude by the application of polarizing current, leading the authors to suggest that the nerve terminal membrane potentials were somewhere near −30 mV or even more positive. In cold-sensitive fibres this depolarized condition could be partly attributed to the chronic activation of cold-gated ion channels (Carr et al. 2003). If this estimate of nerve terminal resting potential is close to the physiological value, it is difficult to envisage a function for persistent Na⁺ channels in terminals already operating at such depolarized potentials. However, as Na_V1.9-like immunoreactivity has a rather widespread expression pattern within the final axonal arborization (see Black & Waxman, 2002), then it might function to regulate excitability within a region of more negative membrane potential several microns or more from the transduction site.

ATP receptors

ATP is an algogenic agent that is released from damaged tissue (e.g. Burnstock, 2001), and cell damage can excite nociceptors in vitro, in a manner reproduced by exposure to ATP alone (e.g. Cook & McCleskey, 2002). P2X receptors (ATP-gated ion channels) are expressed by a subset of sensory neurones, and P2X₃ receptors have been implicated in nociception (Chen et al. 1995; Souslova et al. 2000). Activation of P2X₃ receptors in cultured neurones generates large, transient inward currents, that desensitize (e.g. Lewis et al. 1995; Ueno et al. 1999; Souslova et al. 2000). These receptors are implicated in an excitation of cultured neurones following exposure to ATP, and, in current-clamp, give rise to a period of firing lasting about a second (Molliver et al. 2002). Activation of P2Y receptors, on the other hand, give rise to a more prolonged period of action potential firing in cultured neurones (Molliver et al. 2002) through an unknown mechanism, and these receptors activate G-protein pathways probably acting on more than one final target (e.g. Molliver et al. 2002; Huang et al. 2003). The up-regulation of persistent Na⁺ current on exposure to ATP, where the up-regulated amplitude is maintained over minutes in the presence of the agonist, provides a mechanism by which sensory neurones could be made hyperexcitable and this may explain the prolonged period of action potential firing on exposure to ATP.

The P2Y receptors expressed in sensory neurones include P2Y₁, P2Y₂ and P2Y₄, although the levels of P2Y₄ mRNA are low (Nakamura & Strittmatter, 1996; Tominaga et al. 2001; Mollivier et al. 2002; Sanada et al. 2002; Huang et al. 2003; Moriyama et al. 2003). There is evidence that both P2Y₁ and P2Y₂ receptor activation can temperature sensitize TRPV1-mediated currents, the former via a PKC-coupled pathway (Tominaga et al. 2001) although it seems that P2Y₂ is the more important receptor in vivo (Moriyama et al. 2003). Evidence suggests that P2Y₂ is widely expressed by sensory neurones (e.g. Mollivier et al. 2002; Huang et al. 2003), whereas P2Y₁ may be primarily expressed in large, low-threshold mechanosensory neurones (Nakamura & Strittmatter, 1996). Furthermore, the P2Y agonist UTP has been found to activate functionally identified cutaneous afferents in a skin nerve preparation, almost all of which also responded to capsaicin (Stucky et al. 2004), indicating that they are nociceptive. As UTP acts relatively selectively on P2Y₂ and P2Y₄ receptor subtypes (Ralevic & Burnstock, 1998) these results may indicate the presence of these receptors in nociceptive endings, activation of which could stimulate PKC via a G_q/11 protein pathway. G protein pathway-coupled ATP receptors might therefore alter neuronal excitability, at least in part, through functionally up-regulating persistent Na⁺ current.