Pinch-current injection defines two discharge profiles in mouse superficial dorsal horn neurones, in vitro

Abstract

Neurones in the superficial dorsal horn (SDH) are a major target for nociceptive afferents and play an important role in pain processing. One approach to understanding the role of SDH neurones has been to study their action potential (AP) discharge in spinal cord slices during injection of depolarizing step-currents. Four or five neurone subpopulations are typically identified based on AP discharge, with various roles proposed for each in pain processing. During noxious peripheral stimulation in vivo, however, SDH neurones are activated via synaptic inputs. This produces a conductance change with different somato-dendritic distributions and temporal characteristics to that provided by a somatic step-current injection. Here we introduce an alternative approach to studying SDH neurone discharge under in vitro conditions. We recorded voltage-clamp responses in SDH neurones, in vivo, during noxious mechanical stimulation of the hindpaw (1 s pinch, ∼100 g mm⁻²). From these recordings a representative ‘pinch-current’ was selected and subsequently injected into SDH neurones in spinal cord slices (recording temperature 32°C). Pinch-current-evoked discharge was compared to that evoked by rectangular step-current injections. Pinch- and step-current-evoked AP discharge frequency was highly correlated (r² = 0.61). This was also true for rheobase current comparisons (r² = 0.61). Conversely, latency to discharge and discharge duration were not correlated when step- and pinch-current responses were compared. When neurones were grouped according to step-current-evoked discharge, five distinct patterns were apparent (tonic firing, initial bursting, delayed firing, single spiking, and reluctant firing). In contrast, pinch-current responses separated into two clear patterns of activity (robust and resistant firing). During pinch-current injection, tonic-firing and initial-bursting neurones exhibited robust AP discharge with similar characteristics. In contrast, single-spiking and reluctant-firing neurones were resistant to AP discharge. Delayed-firing neurones exhibited pinch-current responses that were transitional between those of tonic-firing/initial-bursting and single-spiking/reluctant-firing neurones. Injection of digitally filtered pinch-currents indicated that transient current fluctuations are necessary for robust repetitive discharge in initial-bursting neurones. These data suggest the functional significance of the diverse step-current-evoked firing patterns, previously reported in SDH neurones remains to be fully understood. When a ‘facsimile’ current profile or pinch-current is used in place of step-currents, AP discharge diversity is much reduced.

In vitro studies of superficial dorsal horn (SDH) neurones typically use somatic injection of depolarizing step-currents to identify subpopulations based on action potential (AP) discharge patterns (Thomson et al. 1989; Yoshimura & Jessell, 1989; Lopez-Garcia & King, 1994; Grudt & Perl, 2002; Prescott & De Koninck, 2002; Ruscheweyh & Sandkuhler, 2002). Under in vivo conditions, however, SDH neurones discharge APs during synaptically driven depolarization. Recently, we used an in vivo mouse spinal cord preparation to analyse AP discharge evoked by depolarizing step-currents and noxious cutaneous mechanical stimulation (Graham et al. 2004b). We found similar subpopulations (evoked by step-currents) to those previously observed in vitro, and that noxious mechanical stimulation (pinch) evoked AP discharge in approximately 50% of these neurones. To address how neurones with different discharge patterns respond to pinch, we compared tonic-firing and initial-bursting neurone responses during stimulation of the hindpaw. Despite clear qualitative and quantitative differences in the AP discharge evoked by depolarizing step-currents in the two groups (tonic firing and initial bursting), the pinch-evoked discharge was remarkably similar. This result is perhaps understandable given that depolarizing step-currents present a neurone's soma with a conductance change that has vastly different somato-dendritic distribution and temporal characteristics from those received by neurones during noxious peripheral stimulation in vivo.

In this study, we analyse the AP discharge of SDH neurones in a slice preparation (in vitro) as we replay the current from an in vivo voltage-clamp recording of the collective synaptic bombardment received by an individual SDH neurone during hindpaw pinch. This ‘pinch-current’ recording was transformed into a depolarizing stimulus and injected into SDH neurones by current clamp. The results of this alternative form of in vitro stimulation are compared to the more common approach of injecting step-currents to evoke AP discharge in SDH neurones.

Methods

Acquisition of in vivo pinch responses

The University of Newcastle Animal Care and Ethics Committee approved all procedures. Details of the in vivo mouse spinal cord preparation have been previously described (Graham et al. 2004a,b). Briefly, animals (C57Bl/6 mice, both sexes, aged 29–35 days) were anaesthetized with urethane (2.2 g kg⁻¹i.p.), and recordings were obtained from SDH neurones (L3–L5 segments), using an AxoClamp 2B amplifier (Axon Instruments, Union City, CA, USA). Patch pipettes (8–12 MΩ) were filled with a K⁺-based internal solution containing (mm): 135 KMeSO₄, 6 NaCl, 2 MgCl₂, 10 Hepes, 0.1 EGTA, 2 MgATP, 0.3 NaGTP, pH 7.3 (with KOH). The exposed surface of the spinal cord was continually irrigated with warmed (37°C) artificial cerebrospinal fluid (ACSF) containing (mm): 118 NaCl, 25 NaHCO₃, 11 glucose, 2.5 KCl, 1 NaH₂PO₄, 1 MgCl₂ and 2.5 CaCl₂. Data were digitized online (sampled at 10–20 kHz, filtered at 10 kHz) via an ITC-16 computer interface (Instrutech, Long Island, NY, USA) and stored on a Macintosh G4 computer using Axograph v4.6 software (Axon Instruments). After obtaining the whole-cell recording configuration (series resistance <30 MΩ), the neurone's receptive field was identified by applying a soft-bristled brush over the glabrous skin of the hindpaw. Recordings were then made in voltage- (holding potential −60 mV) and current-clamp modes as a custom-built computer-controlled pincher (Graham et al. 2004a) delivered a noxious mechanical stimulus (1 s pinch ∼100 g mm⁻²) to the centre of the neurone's receptive field.

In vitro slice preparation

Spinal cord slices were prepared using standard techniques (Yoshimura & Jessell, 1989; Graham et al. 2003). Briefly, animals (C57Bl/6 mice, both sexes, aged 22–59 days) were anaesthetized with ketamine (100 mg kg⁻¹i.p.) and decapitated. The lumbosacral enlargement of the spinal cord was rapidly removed and placed in ice-cold sucrose-substituted ACSF (250 mm sucrose for Na⁺). After cutting transverse slices (L3–L5 segments; 300 μm-thick) using a vibrating microtome (Leica VT-1000S, Heidelberg, Germany), they were transferred to a storage chamber containing oxygenated ACSF, and allowed to equilibrate for 1 h at room temperature (22–24°C) before recording.

In vitro electrophysiology

Slices were transferred to a recording chamber and continually superfused with oxygenated ACSF (see above). All recordings were made at elevated temperature (32°C) from visualized SDH neurones, using infrared differential interference contrast (IR-DIC) optics. Recordings were limited to the SDH by targeting neurones located within or dorsal to the substantia gelatinosa, which is easily identified by its translucent appearance in spinal cord slices. Patch pipettes (2–4 MΩ) contained the same internal solution as used for in vivo experiments. The whole-cell recording configuration was established in voltage clamp (holding potential −60 mV) before switching to current clamp. The membrane potential immediately after this switch (within 15 s) was designated as resting membrane potential (RMP), and all recordings were subsequently made from this potential. Data were acquired using an Axopatch 200B amplifier (Axon Instruments) and digitized as above. SDH neurones received two types of stimulation via the recording electrode: (1) a series of depolarizing rectangular ‘step-current’ injections (800 ms duration, 20 pA increments, delivered every 8 s), and (2) a series of depolarizing ‘pinch-current’ injections.

Data analysis

Criteria for inclusion of a neurone in the in vitro data set was a RMP more negative than −50 mV and a series resistance < 20 MΩ. All data were analysed offline using Axograph software. Individual APs elicited by step- and pinch-currents were captured using a derivative threshold method (dV/dt = 20 V s⁻¹) with the inflection point during spike initiation being defined as AP threshold. Rheobase current was determined as the smallest step- and pinch-current to elicit at least one AP. The difference between AP threshold and its positive peak was used to define AP amplitude. AP base-width was measured at AP threshold. AP afterhyperpolarization (AHP) amplitude was measured as the difference between AP threshold and the maximum negative peak following the AP.

Several parameters were measured to describe each neurone's AP discharge during step- and pinch-currents. Discharge latency was the time from the onset of step- or pinch-current injection to the first evoked AP in a train. Discharge duration was the time between the onset of the first and last AP during a step- or pinch-current. Instantaneous AP frequency, for all step- or pinch-currents evoking two or more APs, was the calculated reciprocal of the time interval between successive APs. For step- and pinch-currents that contained multiple APs, mean discharge frequency was the average of all instantaneous AP frequencies.

All statistical analysis was carried out using SPSS v10 (SPSS Inc., Chicago, IL, USA). Analysis of variance (ANOVA) was used to compare variables between/across SDH neurone groupings based on step- and pinch-current responses. The Student–Newman–Keuls post hoc test was used to determine which groups differed. Data that failed Levene's test for homogeneity of variance were compared using the non-parametric Kruskal–Wallis test. Student's t test was used to compare filtered and unfiltered pinch-current responses. Statistical significance was set at P < 0.05. All values are presented as means ± s.e.m.

Results

Whole-cell patch-clamp recordings were first obtained from five SDH neurones, from four animals under in vivo conditions. These recordings allowed us to assess and then determine a representative response from SDH neurones following noxious mechanical stimulation via pinch (see Graham et al. 2004b). The resultant response (i.e. due to synaptic excitation following peripheral stimulation in vivo) was acquired in both current- and voltage-clamp modes. Subsequently, a voltage-clamped pinch response was selected and used to study SDH neurone discharge properties under in vitro conditions in 52 neurones from 12 animals. These pinch-current responses were compared to the responses to step-current injection in the same neurone.

In vivo pinch responses

The first goal of this study was to obtain responses from SDH neurones, in vivo, during the application of a 1 s pinch to the hindpaw. Figure 1 shows current and voltage traces recorded from three SDH neurones during hindpaw pinch. In current clamp, AP discharge was elicited in all neurones; however, the temporal features of this discharge varied (upper traces in Fig. 1A). In voltage clamp, the shape of the underlying pinch-evoked current was similar in all SDH neurones tested (n = 5). The characteristics of these pinch-currents are provided in Table 1. The pinch-evoked current reached a maximum during the first 500 ms and then gradually returned to baseline (lower traces in Fig. 1A).

Figure 1.

In vivo current- and voltage-clamp recordings from SDH neurones during noxious mechanical stimulation of the hindpaw A, recordings from three neurones during the application of a 1 s pinch to the hindpaw. Current-clamp recordings (upper black traces) show that responses range from robust AP discharge throughout the pinch stimulus (left), to a single AP at pinch onset (right). Arrowheads denote −60 mV. Voltage-clamp recordings (lower grey traces) show that peak amplitude, area under the curve (total charge), and decay of the inward current evoked by pinch is qualitatively and quantitatively similar in all neurones (also see Table 1). B, Expanded view of selected pinch-current (asterisk in A), after conversion to a depolarizing current-stimulus for injection into SDH neurones under in vitro conditions. This current was scaled to provide a series of five pinch-currents (20%, 40%, 60%, 80% and 100% of the original amplitude). C, the mean current delivered during each pinch-current appears in the left bar plot. Dashed lines extending from the depolarizing step-currents indicate the mean current delivered by each step. The five pinch-currents are best matched by the first five step-currents (P1, 20 pA; P2, 40 pA; P3, 60 pA; P4, 80 pA; and P5, 100 pA). D, the total charge delivered during each pinch-current appears in the left bar plot. Dashed lines extending from the depolarizing step-currents indicate the total charge delivered by each step. The five pinch-currents are best matched by the first three, fifth, and sixth step-currents (P1, 20 pA; P2, 40 pA; P3, 60 pA; P4, 100 pA; and P5, 120 pA).

Table 1.

Comparison of group and selected pinch-current attributes

	Group mean ± s.e.m	Group range	Selected pinch-current
Peak (pA)	225 ± 20	151–264	246
Area (nA s)	102 ± 15	73–140	101
Rise time (ms)	258 ± 35	56–263	143
Decay time constant (ms)	1.2 ± 0.3	0.6–2.2	1.6
Current variance (nA2)	2.38 ± 0.72	1.04–4.35	1.70

Group data are based on characteristics of five in vivo-recorded pinch-currents.

One voltage-clamped pinch response (Fig. 1A, right lower trace) that had representative features (see Table 1) of the current responses during pinching was selected for use in subsequent in vitro experiments. A 2 s epoch from this trace, containing the pinch response, was extracted and converted to a depolarizing stimulus (Fig. 1B). This current resulted from the summation of multiple synaptic inputs, and had two prominent features: (1) an overall, slow, triangular depolarization that rose rapidly and decayed slowly; and (2) numerous fast transients that produced a highly fluctuating baseline. The selected pinch-current was scaled to produce five traces of increasing amplitude (20%, 40%, 60%, 80%, and 100% of the original amplitude). This provided a set of pinch-stimuli of increasing intensity. These scaled ‘representative’ pinch-currents were subsequently injected into SDH neurones under in vitro conditions to examine AP discharge features. This approach is similar to step-current injection, a technique that has been used extensively to study AP discharge in SDH neurones in vitro (Thomson et al. 1989; Yoshimura & Jessell, 1989; Lopez-Garcia & King, 1994; Grudt & Perl, 2002; Prescott & De Koninck, 2002; Ruscheweyh & Sandkuhler, 2002).

In order to make appropriate comparisons between the AP discharge evoked by pinch- and step-currents, the mean amplitude and total charge for these two qualitatively different stimuli were measured (Fig. 1C and D). The average current delivered by the five pinch-currents (pinch 1 to pinch 5) closely matched the average current delivered by the five step-currents (20–100 pA) (Fig. 1C). The total charge delivered by the five pinch-currents most closely matched that of the first three, fifth and sixth step-currents (Fig. 1D). For comparative purposes we chose to emphasize the charge equivalence of step- and pinch-currents rather than amplitude alone, since charge combines both amplitude and duration of responses. The first three pinch-currents were paired with the first three step-currents (i.e. pinch 1 with 20 pA, pinch 2 with 40 pA, pinch 3 with 60 pA) and the last two pinch-currents were paired with the 100 pA and 120 pA step-currents (i.e. pinch 4 with 100 pA, pinch 5 with 120 pA).

Intrinsic, AP- and discharge properties of SDH neurones in vitro

The intrinsic membrane, AP, and discharge properties for 43 SDH neurones are reported in Table 2. Mean input resistance and resting membrane potential were 367 ± 21 MΩ and −66.5 ± 1.1 mV, respectively. AP characteristics were analysed using responses to depolarizing step-currents. The values for AP threshold, AP amplitude, AP base-width and AP afterhyperpolarization were −37.9 ± 0.6 mV, 48.3 ± 1.6 mV, 1.76 ± 0.06 ms and −30.5 ± 1.2 mV, respectively. Some of these values differ from those previously reported. In particular, our input resistances are lower, AP amplitude is smaller and AP base width is shorter than values reported in other in vitro studies (Prescott & De Koninck, 2002; Ruscheweyh & Sandkuhler, 2002). These differences are consistent with the known influence of elevated temperature (32°C versus 23°C), as employed in this study, on active and passive membrane properties of neurones (Hille, 2001).

Table 2.

Intrinsic and AP properties for SDH neurones grouped according to discharge pattern

	Sample	Tonic firing (TF)	Initial bursting (IB)	Delayed firing (DF)	Single spiking (SS)	Reluctant firing (RF)
% of sample (n)	100% (43)	19% (8)	25% (11)	21% (9)	23% (10)	12% (5)
Input Resistance (MΩ)	367 ± 21	385 ± 61	342 ± 51	380 ± 26	356 ± 34	388 ± 83
RMP (mV)	−66.1 ± 1.1	−62 ± 1.4	−62.2 ± 1.7	−69.2 ± 2.6	−70.5 ± 2.6	−70.5 ± 3.1
Rheobase Current (pA)	116 ± 18	25 ± 3	40 ± 5	128 ± 26	262 ± 27	—
		*DF and SS	*DF and SS	*TF, IB and SS	*TF, IB and DF
AP Threshold (mV)	−37.9 ± 0.6	−39.6 ± 1.4	−40.4 ± 0.7	−34.6 ± 1.0	−36.5 ± 1.3	—
		*DF	*DF	*TF & IB
AP amplitude (mV)	48.3 ± 1.6	55.9 ± 3.2	48.2 ± 2.4	43.0 ± 3.6	47.1 ± 3.6	—
AP base width (ms)	1.76 ± 0.06	1.58 ± 0.07	1.91 ± 0.12	1.92 ± 0.18	1.61 ± 0.09	—
AHP amplitude (mV)	−30.5 ± 1.2	−35.1 ± 1.6	−29.6 ± 1.8	−28.9 ± 0.8	−29.2 ± 1.1	—

Values are means ± s.e.m. Membrane potential values are corrected for a 10 mV liquid junction potential (Barry & Lynch, 1991). AP properties could not be analysed in reluctant-firing neurones.

^*Significant difference (P < 0.05) and discharge categories compared.

SDH neurones in our sample exhibited the discharge patterns that have previously been described under in vitro and in vivo recording conditions. This included tonic-firing, initial-bursting, delayed-firing, single-spiking and reluctant-firing neurones (comprising 19%, 25%, 21%, 23%, and 12% of the population, respectively). These subpopulations also retained the characteristic membrane and AP properties reported in other studies (Grudt & Perl, 2002; Prescott & De Koninck, 2002; Ruscheweyh & Sandkuhler, 2002). For example, tonic-firing neurones exhibited the largest and briefest APs, whereas single-spiking and delayed-firing neurones required significantly larger step-currents to evoke AP discharge. From these data we conclude that the present sample captures the typical heterogeneity of SDH neurones, and thus provides a representative sample to examine responses to pinch-current injections under in vitro conditions.

Responses of SDH neurones to pinch-current injection in vitro

In 35/43 SDH neurones, two major types of response were readily identified during pinch-current injections in vitro (Fig. 2). The first was characterized by increased AP discharge frequency and duration as pinch-current amplitude increased (Fig. 2A). In the second, little or no AP discharge was observed in response to pinch-currents of increasing amplitude (Fig. 2B). Close examination of the AP discharge characteristics displayed by the two response types support our separation of the two groups. Specifically, most cells exhibiting responses like those shown in Fig. 2A were tonic-firing or initial-bursting neurones (14/15). Neurones with responses in Fig. 2B were mostly (19/20) single-spiking, delayed-firing or reluctant-firing neurones. Together, these observations suggest a predictive relationship exists between step-current- and pinch-current-evoked AP discharge in SDH neurones.

Figure 2.

SDH neurones show variable responses to pinch- and step-currents in vitro A and B, recordings from two SDH neurones showing responses during pinch- (upper traces) and step-currents (lower traces) of increasing magnitude (arrowheads denote −60 mV). The neurone in A responded during pinch- and step-currents with robust AP discharge that increased in frequency and duration as pinch- and step-current magnitude increased (lower traces). In contrast, the neurone in B did not discharge APs during pinch-currents, and only discharged a single AP at the highest step-current intensity. C, plots comparing AP discharge characteristics elicited by step- versus pinch-currents. The first plot compares the minimum step- and pinch-currents required to elicit AP discharge (rheobase-steps versus rheobase-pinches). This relationship was highly correlated. The remaining plots compare data from pinch 5 (P5) with the 120 pA step-current. Mean AP discharge frequency, like rheobase, was significantly correlated. The latency between stimulus onset and AP discharge and the discharge duration were not significantly correlated. Note there is extensive overlap of data points in the plot of AP latency.

To further explore the relationship between pinch- and step-current-evoked AP discharge we used correlation analysis on a number of variables that distinguish SDH neurone discharge patterns during step-current injections (Fig. 2C). Inclusion in this analysis required that SDH neurones discharged APs during both pinch- and step-currents. Furthermore, neurones that only discharged a single AP were not included in comparisons of discharge duration and mean discharge frequency. These criteria were fulfilled by all tonic-firing and initial-bursting neurones (n = 8 and n = 11, respectively), and half the delayed-firing neurones (4/9) but excluded single-spiking (n = 10) and reluctant-firing (n = 5) neurones. These analyses compared rheobase current, AP discharge onset, duration and mean frequency during pinch- and step-current injection (Fig. 2C). A significant correlation (r² = 0.61) existed between the rheobase current required to evoke AP discharge following both types of stimuli. Mean discharge frequency during maximum pinch- and step-current injections were also significantly correlated (r² = 0.61, pinch 5 versus 120 pA step). No significant correlation, however, existed between latency from stimulus onset to AP discharge, and the duration of AP discharge (r² = 0.14 and 0.12, respectively, pinch 5 versus 120 pA step).

Pinch-evoked discharge in SDH neurones with different discharge patterns

We next separated the data from SDH neurones (n = 43), grouped it by discharge pattern during step-current injection (i.e. tonic-firing, initial-bursting, delayed-firing, single-spiking and reluctant-firing), and constructed averaged peristimulus histograms of AP discharge evoked during pinch-currents (Fig. 3). This allowed quantitative comparisons between the onset and temporal characteristics of pinch-evoked discharge in neurones with different step-evoked AP discharge patterns. Once again, tonic-firing and initial-bursting neurones responded with similar characteristics to most pinch-currents, with duration and discharge frequency increasing as pinch-current amplitude increased (Fig. 3A and B, respectively). In delayed-firing neurones, pinch-current responses were shifted to the right (Fig. 3C). These neurones only responded to high-intensity pinch-currents and resembled the tonic-firing and initial-bursting responses observed during low-intensity pinch-currents. Finally, single-spiking and reluctant-firing neurones only responded to the highest-intensity pinch-currents, and these responses were brief, comprising a few APs (Fig. 3D and E, respectively). Thus, when a ‘facsimile’ current profile or pinch-current is used to assess AP discharge, only two distinguishable categories are revealed.

Figure 3.

Pinch-current-evoked discharge in SDH neurones expressing different step-current-evoked discharge patterns A–E, pinch-current data for recorded SDH neurones were grouped according to the discharge pattern exhibited during step-current injections (left, overlayed voltage traces, arrowheads denote −60 mV). Averaged peristimulus histograms (100 ms bins) of AP discharge were constructed for each pinch-current (pinch 1 to pinch 5, left to right). A, tonic-firing neurones typically exhibited responses to most pinch-currents with increasing discharge frequency and duration as pinch-current intensity increased. B, initial-bursting neurones also exhibited responses to most pinch-currents with increasing discharge frequency and duration as pinch-current intensity increased. Moreover, these responses had similar characteristics to those of tonic-firing neurones. C, delayed-firing neurones did not respond to low-intensity pinch-currents, but exhibited responses to high-intensity pinch-currents with discharge characteristics similar to tonic-firing and initial-bursting responses during low-intensity pinch-currents. D, single-spiking neurones only responded to the highest-intensity pinch-currents, and these responses were restricted to a brief discharge of one or two APs. E, reluctant-firing neurone responses during pinch-currents were similar to those of single-spiking neurones, showing a brief discharge of one or two APs during the high-intensity pinch-currents.

The role of rapid transients in sustaining pinch-evoked discharge

Previous work has shown that both tonic-firing and initial-bursting neurones can fire repetitively under some stimulus conditions (Prescott & De Koninck, 2002). To establish which features of our pinch-current allowed tonic-firing and initial-bursting neurones to exhibit surprisingly similar discharge (Fig. 3A and B), we generated a series of smoothed pinch-currents that successively reduced complexity (i.e. reduced rapid transients) on the original pinch-current (Fig. 4). Two digitally smoothed pinch-currents were obtained from the original trace using a box-car filter. This procedure replaces individual points with an average of a number of surrounding points – smoothed-pinch 1 averaged 100 surrounding points; and smoothed-pinch 2 averaged 1000 surrounding points. A ramp current was also generated that replicated the underlying ‘triangular’ depolarization found in the original pinch-current. Importantly, these three new traces provided approximately the same charge as the original, unfiltered, pinch-current (∼100 pA s, Fig. 1D). We quantified the reduced complexity of these alternative pinch-currents by subtracting the derived pinch-current from the original and measuring the remaining variance. The current variance of smoothed-pinch 1 was reduced by 45%, smoothed-pinch 2 was reduced by 87%, and the variance on ramp-pinch was reduced by 100%.

Figure 4.

Reducing pinch-current complexity Three alternative pinch-current stimuli were generated to determine how the overall level of depolarization versus the transient fluctuating nature of the pinch-current baseline influenced AP discharge. The original pinch-current (top left) is essentially a triangular depolarization with a fluctuating baseline (inset shows segment of each pinch-current on an expanded time scale; indicated by black bar). The original pinch-current was filtered at two levels to smooth its baseline, producing two pinch-currents with progressively reduced baseline fluctuations (smoothed-pinch 1, bottom left; and smoothed-pinch 2, top right). A ramp-pinch current was also generated that followed the underlying ‘triangular’ depolarization of the original pinch-current (bottom right).

We next injected the alternative pinch-currents into a selected sample of tonic-firing (n = 4) and initial-bursting (n = 5) neurones, with similar input resistances and RMPs (317 ± 54 MΩversus 338 ± 21 MΩ, and −58 ± 6 mV versus−55 ± 3 mV, respectively), and compared the resulting discharge (Fig. 5). As may be expected, both neurone classes responded similarly to the original pinch-current with robust AP discharge. Likewise, AP discharge during application of smoothed-pinch 1 was similar in both groups; however, smoothed-pinch 2 and ramp-pinch produced markedly different results in tonic-firing and initial-bursting neurones. Tonic-firing neurones sustained AP discharge during application of smoothed-pinch 2, with a similar number of APs and over a similar duration to the original pinch-current. In contrast, initial-bursting neurones discharged significantly fewer APs over a shorter duration during application of smoothed-pinch 2. Application of the ramp-pinch also separated the neurone classes by evoking sustained discharge in tonic-firing neurones but not in initial-bursting neurones, which responded with significantly fewer APs over a shorter duration. These data suggest that the rapid transients featured on the original pinch-current play an important role in allowing initial-bursting neurones to maintain robust repetitive discharge. In contrast, tonic-firing neurones are relatively insensitive to these rapid transients and continue to show repetitive discharge even as transients are progressively removed.

Figure 5.

Initial-bursting, but not tonic-firing, neurones require pinch-current complexity (transient fluctuations) for sustained AP discharge A, representative voltage recordings from tonic-firing (upper traces) and initial-bursting (middle traces) neurones during injection of pinch-currents of decreasing complexity (lower traces, left to right). Arrowheads denote −60 mV. Tonic-firing neurone responses are unchanged as transient fluctuations are progressively removed from pinch-currents. In contrast, initial-bursting neurone responses are decreased as transient fluctuations are removed. B, group data comparing the number of APs (left) and the discharge duration (right) in tonic-firing and initial-bursting neurones during pinch-current injections of decreasing complexity. For tonic-firing neurones, AP number and discharge duration were similar for all pinch-currents. Alternatively, AP number and discharge duration were significantly reduced during pinch-current injections of decreasing variance (SP2 and RP) in initial-bursting neurones. *Significant difference from original (OP) pinch-current response.

Discussion

Pinch-current features and caveats

The rationale for injecting pinch-currents into SDH neurones, under in vitro conditions, was to provide a depolarizing stimulus that more closely matched the characteristics of the excitation these neurones receive in vivo. The voltage-clamped pinch-current waveform selected for our experiments was representative of the responses we recorded in vivo, having a similar peak amplitude, total charge, and decay to the mean of five neurones studied (Table 1, Fig. 1). Support for this approach can be found in other in vivo voltage-clamp studies that used pinch (noxious mechanical stimulation) to investigate synaptic excitation of SDH neurones (Furue et al. 1999; Light & Willcockson, 1999; Narikawa et al. 2000). The pinch-current response reported here closely resembled the first second of the pinch responses reported in these studies, despite longer pinch durations (∼5 s versus 1 s) in these studies. Thus, we are confident our facsimile pinch-current has properties that are similar to the excitation SDH neurones receive during noxious cutaneous mechanical stimulation.

It is important to recognize the limitations of this approach, and that certain features of in vivo excitation are lost when a pinch-current is injected directly into the soma. First, in vivo, many of the synapses that contribute to membrane excitation during pinch are located on dendrites, some distance from the soma (Willis & Coggeshall, 2004). Consequently, as voltage-gated ion channels are distributed non-uniformly throughout a SDH neurone's soma and dendritic tree (Wolff et al. 1998; Safronov, 1999), the interactions of synaptic (dendritic) versus electrode-injected (somatic) pinch-currents with spike-generating mechanisms are likely to differ. Second, ligand-gated ion channels normally activated during synaptic bombardment by pinch together with downstream intracellular processes are circumvented by electrode-injected pinch-currents. Finally, not all neurones in the SDH receive nociceptive inputs, and thus some of our in vitro sample may not normally be subjected to direct pinch-evoked excitation (Light & Willcockson, 1999; Graham et al. 2004a). Despite these caveats, the pinch-current responses we have recorded under in vitro conditions are strikingly similar to in vivo pinch responses reported in this and previous studies (Graham et al. 2004a,b). Furthermore, the bypassing of dendritic input may not be so serious a problem since the pinch-current was originally recorded in vivo from the soma of a SDH neurone and has therefore already undergone dendritic filtering. Taken together, these considerations support injection of a pinch-current at the soma, as a realistic means of assessing responses to depolarizations using a synaptic profile previously generated by noxious peripheral stimuli in vivo.

Responses to pinch-currents versus step-currents

These experiments compare the input/output properties of in vitro SDH neurones using two stimuli with very different profiles. Previously, we delivered step-currents to SDH neurones in an in vivo preparation and compared these responses to AP discharge evoked during pinching of the hindpaw (Graham et al. 2004a). We concluded that the mechanisms underlying some of the distinct discharge patterns observed during step-current responses, specifically tonic firing and initial bursting, were less important during pinch-evoked responses. Delayed-firing, single-spiking and reluctant-firing neurones were not included in that analysis because of our limited sample. In agreement with this in vivo study, we also find that under in vitro conditions, SDH neurones with distinctly different discharge properties during step-currents, such as tonic-firing and initial-bursting neurones, exhibit indistinguishable responses during pinch-currents. Likewise, single-spiking and reluctant-firing neurones maybe grouped together since they exhibit similar pinch-current responses. The response of delayed-firing neurones during pinch-currents was partway between tonic-firing/initial-bursting and single-spiking/reluctant-firing responses. Thus, in terms of their input/output function during a fluctuating depolarizing current, SDH neurone responses can be classified into two canonical groups; robust with vigorous AP discharge or resistant with limited AP discharge.

The above findings indicate that although five different forms of AP discharge can be readily discerned in SDH neurones when step-currents are used, the same five populations cannot be identified by responses during a hindpaw pinch. Interestingly, one in vitro study in lamina I neurones, has observed a range of responses in spontaneously active neurones that are in some respects equivalent to each of the discharge patterns observed during step-currents (Prescott & De Koninck, 2002). This study also provided evidence that SDH neurones other than tonic-firing, particularly initial-bursting neurones, can sustain repetitive discharge given the appropriate stimulus. Our findings provide further support for the notion that tonic-firing and initial-bursting neurones can respond similarly during pinch-current injection. More recently the above authors have identified two active conductances, selectively expressed in tonic-firing neurones, that allow them to act as integrators: i.e. encode stimulus intensity (Prescott & De Koninck, 2005). From our results it appears that whatever role is proposed for tonic firers, initial-bursting neurones serve a similar role during our pinch-current injection.

In this study we have demonstrated that rapid transients within the pinch-current waveform are requisite for initial-bursting neurones to support robust pinch-evoked discharge. These results showed that when filtering is used to remove rapid transients (i.e. smoothed-pinch 2, Fig. 5) the same neurones no longer exhibit robust pinch-evoked discharge. This is perhaps not surprising as the profile of smoothed and ramp currents progressively approach that of a step-current, which after all, is the basis of separating initial-bursting and tonic-firing neurones according to their capacity to sustain AP discharge.

Functional implications and future directions

Overwhelmingly, the in vitro literature describing SDH neurone excitability has implied that various discharge patterns observed during step-current injections shape in vivo AP discharge during nociceptive processing under both normal and pathological conditions (Thomson et al. 1989; Lopez-Garcia & King, 1994; Prescott & De Koninck, 2002; Ruscheweyh & Sandkuhler, 2002). However, in vivo studies that record SDH neurone responses during noxious cutaneous mechanical stimulation do not describe the same patterns of AP discharge (Cervero et al. 1976; Kumazawa & Perl, 1978; Light et al. 1979; Laird & Cervero, 1989; Furue et al. 1999; Light & Willcockson, 1999). This discrepancy might be explained by the intact nature of the in vivo preparation, where different intrinsic membrane properties between neurones may be just one of many explanations that could account for different responses. Alternatively, different pinch-evoked responses could easily be explained by stimulus application to a different location within a neurone's receptive field rather than different intrinsic membrane properties. Using this reduced preparation, with a standardized pinch-current for all SDH neurones, we have demonstrated that some intrinsic membrane properties greatly influence the response of SDH neurones during pinches (tonic firing/initial bursting versus single spiking/reluctant firing). Conversely, other intrinsic membrane properties that might be expected to produce distinct pinch-evoked AP discharge in various neurones do not result in markedly different pinch-evoked responses (tonic firing versus initial bursting or single spiking versus reluctant firing).

These findings have functional implications for models of nociceptive processing under normal and pathological conditions. Our data suggests that tonic-firing, initial-bursting and, to a lesser degree, delayed-firing neurones can respond with AP discharge during a brief noxious stimulus (1 s pinch), and thus participate in the processing of noxious stimuli. During such stimulations single-spiking and reluctant-firing neurones remain largely silent. Importantly, these conclusions can now be drawn from data collected with both in vitro spinal slices and an in vivo spinal cord preparation. Thus, if the mechanisms underlying delayed firing, single spiking and reluctant firing were altered to allow greater AP discharge, a predicted consequence would be profoundly changed nociceptive processing, with many more SDH neurones becoming hyperexcitable. Future studies using various models of altered pain states are now necessary to test this hypothesis.

Conclusions

In this study we suggest that the five discharge patterns routinely described during step-current injections collapse into two groups based on responses recorded during pinch-current injections. We acknowledge, however, that our facsimile of a pinch-current only models a single brief noxious stimulus. We cannot therefore rule out the possibility that there may be conditions, such as longer noxious stimulation intensities and durations or chronic pain states, where neurones exhibiting each of the five described discharge patterns selectively mediate different responses during natural stimulation, possibly correlating with each discharge pattern. Future in vivo studies, using noxious and innocuous stimuli of varying durations and intensities, are needed to acquire voltage-clamp recordings of other excitation profiles that SDH neurones receive in the intact animal. These profiles can then be used to generate a more complete repertoire of naturally acquired stimuli for future use under in vitro conditions.