Chronic pain conditions may result from peripheral nerve injury, chronic peripheral inflammation, or sensory ganglia inflammation. However, inflammatory processes may also contribute to peripheral nerve injury responses. To isolate the contribution of local inflammation of sensory ganglia to chronic pain states, the authors previously developed a rat model in which long-lasting pain is induced by inflaming sensory ganglia without injuring the neurons. This results in prolonged mechanical pain, local increases in proinflammatory cytokines, increased neuronal hyperexcitability, and abnormal spontaneous activity.
The authors used whole cell patch clamp in acutely isolated small-diameter neurons to determine how localized inflammation (3-5 days) of L4 and L5 ganglia altered voltage-gated K and Na currents.
Tetrodotoxin-sensitive Na currents increased twofold to threefold in neurons from inflamed ganglia. Tetrodotoxin-resistant Na currents increased more than twofold, but only in cells that bound isolectin B4. These increases occurred without shifts in voltage dependence of activation and inactivation. Similar results are seen in models of peripheral inflammation, except for the large magnitudes. Unlike most pain models, localized inflammation increased rather than decreased voltage-gated K currents, due to increased amplitudes of the sustained (delayed rectifier) and fast-inactivating transient components. The overall effect in current clamp experiments was an increase in excitability as indicated by decreased rheobase and lower action potential threshold.
Neuronal inflammation per se, in the absence of nerve injury, causes large increases in Na channel density and enhanced excitability. The unusual finding of increased K current may reflect regulation of excitability in the face of such large increases in Na current.
ANIMAL models of chronic pain have generally been classified as either nerve injury or inflammation. Both nerve injury and inflammation can produce spontaneous pain, hyperalgesia, and allodynia.1Both are associated with increased excitability and ectopic spontaneous discharges of dorsal root ganglion (DRG) neurons.2–6These changes in sensory neurons, along with altered information processing in the spinal cord or higher centers, are thought to contribute to increased pain sensation.7However, the underlying ionic mechanisms mediating the increased neuronal excitability are thought to differ between nerve injury and chronic inflammation. For example, nerve injury generally results in overall down-regulation of Na+channels along with changes in the isoforms expressed, whereas chronic peripheral inflammation often results in up-regulation of tetrodotoxin-resistant Na+channels.8Voltage-dependent K+currents are down-regulated in almost all nerve injury and peripheral inflammation models, but the details differ between different models.2,9–15
Recent work has suggested that the distinction between inflammation and nerve injury may not be complete. Nerve injury models often include some inflammation. For example, in peripheral nerve injury, macrophage infiltration, release of localized proinflammatory cytokines, and their retrograde transport to the DRG have been shown to be important for hyperalgesia.16–18Inflammatory processes and glial activation within the DRG have also been proposed to play important roles in some nerve injury models, although these processes have been more extensively studied in the central nervous system and peripheral nerves (for review, see Ji and Strichartz,19Marchand et al. ,20and Moalem and Tracey21).
To gain a better understanding of the inflammatory contributions to pathologic pain, we developed a rodent pain model in which the somata of sensory neurons are subjected to direct inflammatory stimulus by depositing a small drop of the immune activator zymosan into the epidural space in the vicinity of the DRG.22Localized inflammation of the DRG (LID), in the absence of nerve trauma or damage, occurs in certain clinical pain states such as postherpetic neuralgias and some forms of low back pain, such as after lumbar disc rupture. Material released from the nucleus pulposus is known to possess immunogenic and chemogenic capacities.23,24However, we also view the LID model as one that allows us to study the effects of DRG inflammation per se (i.e. , in the absence of nerve damage or axotomy). This is of more general interest because inflammation is one component of most pain models, including those based on nerve injury. Local inflammation results in bilateral mechanical hyperalgesia, which begins by day 1, peaks at days 3–7, and lasts up to 28 days. LID also results in rapid macrophage infiltration and glial activation, increased levels of proinflammatory cytokines, and abnormal sprouting of sympathetic fibers in the DRG.22Like other pain models, LID leads to increased excitability and spontaneous activity of the DRG cells, including C cells. In the current study, we used patch clamp techniques on inflamed small-diameter DRG cells to investigate changes in Na+and K+channels that might contribute to the excitability changes observed in LID.
Materials and Methods
Young female Sprague-Dawley rats (body weight, 100–150 g) were housed one or two per cage under a controlled diurnal cycle of 12 h light and 12 h dark with free access to water and food. The ambient environment was maintained at constant temperature (22°± 0.5°C) and relative humidity (60–70%). All of the surgical procedures and the experimental protocol were approved by the institutional animal care and use committees of the University of Cincinnati (Cincinnati, Ohio).
In Vivo Localized Inflammatory Irritation of the DRG
Rats were anesthetized with isoflurane. The paraspinal muscles on the right side were separated surgically from the L4 to L6 vertebrae as described previously.22,25The superior articular and the transverse processes between L4 and L6 vertebrae were cleaned, the L5 and L4 spinal nerves that course through the intervertebral foramen were identified, and a small hole (0.49 mm in diameter) was drilled through the junction of transverse process and lamina over the L5 and L4 DRG. The drill bit had a limited length (1.5–2.0 mm) that could be finely adjusted during the surgery to ensure that the drill just barely penetrated the overlying bone without injuring the underlying ganglion. The location of the hole was determined according to the course of the L5 or L4 spinal nerve and was approximately 2 mm from the inferior edge of the transverse process. Before the surgery, a 25-gauge needle (0.5 mm in diameter) was cut to a length less than the thickness of the transverse process (2–3 mm in length) and was forced into the hole in close proximity to the DRG but without contracting the ganglion. Then, 20 μl freshly prepared zymosan (0.5 μg/μl; 10 μg total, in incomplete Freund adjuvant) was slowly injected into the hole through the inserted needle. The needle with attached syringe was left in place for at least 3 min after the injection to ensure complete delivery of the solution and avoid possible extravasations outside the DRG site.
Acute Culture of Sensory Neurons
Dorsal root ganglia were isolated from the animals on postoperative days (POD) 3–6, a time period during which the behavioral measurements of mechanical hyperalgesia (von Frey test) and mechanical allodynia (cotton wisp test) have reached their maximum values and are not changing rapidly.22The rats were deeply anesthetized by pentobarbital sodium (40 mg/kg, intraperitoneal). The ipsilateral DRGs (L4 and L5) were isolated and carefully removed in ice-cold normal Ringer's solution for cleaning. The solution contained 130 mm NaCl, 3 mm KCl, 2 mm CaCl2, 2 mm MgSO4, 1.25 mm KH2PO4, 26.2 mm NaHCO3, and 10 mm dextrose and was bubbled with 95% O2and 5% CO2. The pH of the solution was adjusted to 7.4, and the osmolarity was approximately 290–310 mOsm. The connective tissue of the DRGs was dissolved by exposure to Ca2+-free normal Ringer's solution containing 1.0% collagenase II (265 u/mg; Worthington Biochemical Corp., Lakewood, NJ) for 30 min at 37°C followed by rinsing in normal Ringer's solution for another 10 min. The DRG cells were then dissociated by trituration with a fire-polished Pasteur pipette and were plated onto poly-d-lysine–coated glass coverslips in Medium 199 (Sigma Chemicals Co., St. Louis, MO) containing 10% heat-inactivated fetal bovine serum and 1,000 U/ml each of penicillin and streptomycin. The DRG cells were incubated at 37°C (5% CO2balance air) before recording.
After short-term culture (2–30 h), coverslips were transferred to a recording chamber, and DRG cells were visualized under differential interference contrast using an inverted microscope (IX71; Olympus America Inc., Center Valley, PA). Whole cell voltage clamp recordings of small DRG neurons (diameter, 9–25 μm) were conducted at room temperature with an AxoPatch-200B amplifier (Axon Instruments, Foster City, CA). Patch pipettes (2.5–4.0 MΩ) were fabricated from borosilicate glass (Sutter Instruments, Novato, CA). The recording chamber was continuously perfused at room temperature with oxygenated bath solution at a flow rate of 2 ml/min. Data were acquired on a Pentium IV computer with the Clampex 8 program (Axon Instruments). The cell capacitance artifact was canceled by the nulling circuit of the recording amplifier. Ohmic leakage currents were subtracted off-line based on the input resistance measured during voltage steps in the hyperpolarizing range where no voltage-gated currents were evoked. Voltage errors were minimized by using 80% or greater series resistance compensation. The current was filtered at 5 kHz and sampled at 50 kHz. All chemicals were purchased from Sigma (Sigma Chemical Co., St. Louis, MO) unless otherwise indicated.
K+Current and Action Potential Parameters
Excitability measurements were made under current clamp conditions, followed by voltage clamp measurements of voltage activated K+currents using the same internal solution. After obtaining the whole cell configuration, only neurons that had a stable membrane potential more than −45 mV were used for further study. Excitability measurements included the threshold current (rheobase), action potential (AP) threshold, resting membrane potential, AP maximum rising rate (mV/ms), AP maximum falling rate (mV/ms), amplitude and duration of afterhyperpolarization, and input resistance. Resting membrane potential was measured 1 min after a stable recording was obtained. Current pulses from −0.2 to 0.37 nA (80-ms pulse duration) were delivered in increments of 0.03 nA until one or more APs were evoked. The threshold current (rheobase) was defined as the minimum current required to evoke an AP. The AP voltage threshold was defined as the first point on the upstroke of an AP.26The AP amplitude was measured between the peak and AP threshold level. The AP rising rate was defined as the maximum rising rate from baseline to the AP peak, and the AP falling rate was defined as the maximum falling rate from the AP peak to the afterhyperpolarization valley peak.27The afterhyperpolarization amplitude was measured between the maximum hyperpolarization and the final plateau voltage, and the afterhyperpolarization duration was measured at the voltage halfway between these two points.28,29The input resistance for each cell was obtained from the slope of a steady state current–voltage plot in response to a series of hyperpolarizing currents of 80 ms in duration, delivered in decreasing steps of 0.03 nA between 0 and −0.2 nA. Any cells without AP were excluded from the study. After measurement of the excitability parameters in current clamp mode, the amplifier was switched to whole cell voltage clamp mode for measurements of K+currents. Total potassium current was measured by depolarizing voltage steps after a 1-s prepulse to −120 mV. The difference between the peak current and the current at the end of the pulse was used as a measure of the transient K+current, and the current at the end of the pulse was used as a measure of the steady state current. Voltage dependence of inactivation was measured by test pulses to 20 mV after 1-s prepulses to voltages between −100 and 0 mV.
For the sodium current study, we labeled cells with isolectin B4 (IB4) conjugated to fluorescein isothiocyanate (1 μg/μl) for 30–60 min before the start of the recording session, to identify IB4-positive and IB4-negative neurons.30,31This procedure has been previously shown not to affect the currents measured.32Neurons showing a robust fluorescence signal were classified as IB4 positive, and those displaying no signal at all were classified as IB4 negative. Any cells showing an intermediate or weak signal were excluded from the study. The fluorescence was not examined until after completion of current recordings from each cell; i.e. , IB4 binding was not used as a criterion in selecting which cells to record. In some experiments, 100 nm tetrodotoxin was added to the Na+bath solution after currents were recorded in the absence of tetrodotoxin. In most experiments, the tetrodotoxin-resistant currents were isolated from the tetrodotoxin-sensitive currents by using a holding potential of −50 mV. Both tetrodotoxin-resistant and tetrodotoxin-sensitive current were activated by depolarizations from a holding potential of −80 mV. The amplitude of the tetrodotoxin-sensitive current in each cell was measured during depolarizations to −10 mV from −80 mV in these activation protocols, based on the distinct kinetic properties of the two currents. At this voltage, tetrodotoxin-sensitive current was measured as the difference between the early peak current and the current 2 ms after the beginning of the depolarization, a time window in which the tetrodotoxin-sensitive current essentially completely inactivates but tetrodotoxin-resistant current changes little. Similar separation of the two components could sometimes be obtained by digitally subtracting the currents evoked from −50 mV (tetrodotoxin resistant) from the currents evoked from −80 mV (tetrodotoxin resistant plus tetrodotoxin sensitive). However, many cells especially from control animals had a very small tetrodotoxin-sensitive component, and it was observed that in these cases the subtraction method gave difference currents that were dominated by residual tetrodotoxin-resistant currents due to slight differences in inactivation of that current at −50 versus −80, or due to small time-dependent changes. Hence, the activation data for the tetrodotoxin-sensitive current over the full voltage range was obtained only from cells in which that was the predominant current and readily separated from the tetrodotoxin-resistant current by the subtraction method.
Solutions for Recording Potassium Current and Action Potential
The normal bath solution for recording excitability parameters contained 130 mm NaCl, 5 mm KCl, 2 mm CaCl2, 1 mm MgCl2, 10 mm HEPES, and 10 mm glucose. The pH was adjusted to 7.4 with NaOH, and the osmolarity was adjusted to approximately 300–310 mOsm with sucrose. The bath solution for recording K+currents contained 130 mm choline Cl, 5 mm KCl, 1 mm MgCl2, 2 mm CoCl2, 10 mm HEPES, and 10 mm glucose. The pH was adjusted to 7.4 with Tris base, and the osmolarity was adjusted to approximately 300–310 mOsm with sucrose. The pipette solution contained 140 mm KCl, 1 mm CaCl2, 2 mm MgCl2, 11 mm EGTA, 10 mm HEPES, 2 mm Mg adenosine triphosphate, and 1 mm Li guanosine triphosphate. The pH was adjusted to 7.2 with Tris base, and osmolarity was adjusted to approximately 290–300 mOsm with sucrose. Voltages were not corrected for liquid junction potentials, which were estimated to be less than 10 mV in all cases.
Solutions for Recording Sodium Currents
Sodium currents were measured under a reduced sodium gradient to ensure better voltage control. The Na+bath solution contained 30 mm NaOH, 110 mm tetramethyl ammonium chloride, 5 mm tetraethyl ammonium chloride, 2 mm MgCl2, 1 mm CaCl2, 0.1 mm CdCl2, 10 mm HEPES, and 10 mm glucose. The pH was adjusted to 7.4 with HCl, and the osmolarity was adjusted to approximately 300–310 mOsm with sucrose. The pipette solution contained 140 mm CsCl, 5 mm NaOH, 2 mm MgCl2, 1 mm CaCl2, 11 mm EGTA, 10 mm HEPES, 2 mm Mg adenosine triphosphate, and 1 mm Li guanosine triphosphate. The pH was adjusted to 7.2 with CsOH, and the osmolarity was adjusted to approximately 290–300 mOsm with sucrose.
Data Analysis and Statistics
Data were analyzed using pClampfit 8.2 (Axon Instruments) and Origin 7 (Origin Lab Corp., Northampton, MA). Currents were normalized by cell capacitance. Data are expressed as means ± SEM. Differences in proportion of cells binding IB4 were examined with the Fisher exact test. Statistical significance of differences between average values in LID and control neurons were analyzed by the Student t test or, for data that were not normally distributed, the Mann–Whitney rank sum test, as indicated, using SigmaStat Software (Systat Software, Inc., San Jose, CA). In the case of multiple comparisons over a voltage range for activation or inactivation data, the data were analyzed by two-way repeated-measures analysis of variance using SigmaStat software, with pairwise multiple comparison (Holm–Sidak method) to determine at which voltages the differences between LID and control cells were significant if an overall effect of LID was observed. Significance was ascribed for P < 0.05. Any results found to be significant with this method of analysis were reanalyzed at individual voltages with two-way analysis of variance using both condition (LID vs. control) and time in culture (short, 2.5–8 h vs. long, 18–30 h) as factors, to ensure that differences between LID and control values were still significantly different after time in culture was accounted for. Average culture times did not differ between cells from control and LID animals in any of the experiments. Results presented were obtained from three to seven different cultures for each condition (control and LID), as indicated. To further ensure that significant findings were not the results of spurious differences between cultures, we reevaluated the primary significant results by performing t tests on the average values for each culture, instead of on the values for individual cells. Despite the much smaller n values in these comparisons, the following effects of localized inflammation still approached or attained significance (P values are in parentheses after each value): the increase in amplitude of the steady state component (0.055) and fast component (0.09) of the K current at −100 mV; the increase in tetrodotoxin-sensitive current at −20 mV in IB4-negative cells (0.006) and IB4-positive cells (0.19); and the reduced action potential threshold (0.005), and rheobase (0.01), but not the increase in action potential rising rate (0.89) or falling rate (0.35).
LID Increases Excitability
Membrane properties were recorded in 60 acutely isolated small-diameter DRG neurons using the current clamp configuration of the whole cell patch clamp. As summarized in table 1, several measures of excitability were increased in cells isolated from animals 3–6 days after LID, compared with cells isolated from control animals. The mean rheobase current was significantly smaller, and the voltage threshold was significantly more negative, for LID than for control neurons. In addition, both the maximum AP rising rate and falling rate were significantly faster for LID than for control neurons, although this effect was less robust when differences between different batches of cells was examined (see Materials and Methods). No significant difference was found between control and LID neurons in cell capacitance, resting membrane potential, input resistance, AP amplitude, AP duration, afterhyperpolarization amplitude, or afterhyperpolarization duration (table 1).
Voltage-gated K+Currents Increase after LID
Voltage-activated K+currents were recorded as shown in figure 1. To measure total potassium current, the membrane was held at −60 mV, and voltage steps were applied in 10-mV increments to a value of +60 mV after a 1-s prepulse to −120 mV. The total K+current was comprised of a sustained current as measured at the end of the 500-ms pulse, and a transient component that decayed during a 500-ms depolarizing pulse. In comparison with control neurons, LID neurons had significantly increased density of both the transient (overall P value = 0.009) and the sustained components (overall P value < 0.001). Two-way repeated-measures analysis of variance indicated that the effect of LID was significant at individual voltages above 10–20 mV (figs. 1C and D). These effects were due to increased amplitude of the current, without any change in the voltage dependence of activation.
The voltage dependence of steady state inactivation was also determined for the voltage-activated K+currents. The protocol consisted of a 1-s conditioning prepulse to potentials ranging between −100 and 0 mV followed by a depolarizing voltage step to a +20 mV test pulse for 1 s. The outward current at +20 mV could be well described in most cells as the sum of a sustained component and two exponentially decaying transient components whose time constants differed by an order of magnitude (fig. 2). In 5 of 24 control cells and 3 of 26 LID cells, only a slow component was observed; in 3 of 26 LID cells, only a fast component was observed. As shown in figures 2B and C, the average values for the slow and fast time constants were unchanged by LID. The steady state inactivation of each of these three components was studied individually by fitting the data as shown in figure 2. As shown in figure 3, the fast inactivating (figs. 3B and E) component showed a simple increase in amplitude after LID (overall P value from two-way repeated-measures analysis of variance = 0.0004), with no shift in the voltage dependence of inactivation (P = 0.124). Similar results were obtained for the steady state component (figs. 3C and F): The increase in amplitude was significant (overall P value = 0.01), but the voltage dependence was not affected (P = 0.17). This analysis further confirms the increase in sustained current shown in figure 1D, because fitting the final steady state value of the exponential decay should be less contaminated by the incompletely inactivated currents present at some voltages at the end of the 500-ms depolarizations shown in figure 1. Figure 3also demonstrates that the “steady state” component as determined by fits to the current decay during 500-ms pulses can actually show some partial inactivation at more positive potentials, when measured with the 1-s-long prepulses used in this protocol.
The effects on the slow inactivating component were more complex: Although after LID there was a tendency for an increase in the maximum amplitude of the current evoked from most negative potential (fig. 3D; unadjusted P value = 0.04), the overall effect of LID on amplitude was not significant (overall P = 0.98). In addition, there was also a leftward shift of the inactivation curve (fig. 3D; overall P value = 0.001), which would tend to offset any increases in maximum amplitude. Because of this shift, the slow inactivating current density evoked from the more physiologically relevant value of −50 mV was 49.7 ± 7.5 pA/pF in control neurons versus 38.2 ± 5.2 pA/pF in LID neurons, a difference that was not statistically significant.
Because the increased K+currents could not account for the increased excitability observed after LID, we next examined the effects of LID on voltage-gated Na+currents. Because previous studies showed differences between Na currents of the IB4-positive and IB4-negative subclasses of small-diameter DRG neurons,32–34for this study we determined whether the cell bound IB4. The fraction of IB4-positive cells (0.5) did not change significantly with LID (P = 0.8, Fisher exact test).
Examples of Na+currents are shown in figure 4. As described in many previous studies, the Na+current (in both control and LID cells) included tetrodotoxin-sensitive and tetrodotoxin-resistant components. The tetrodotoxin-resistant currents were distinguished by their incomplete steady state inactivation at a holding potential of −50 mV, and by the slower time course of their inactivation during depolarizing pulses. Both before and after LID, and in both IB4-positive and IB4-negative cells, it was possible to observe cells with predominantly tetrodotoxin-sensitive or predominantly tetrodotoxin-resistant current. However, both before and after LID, the ratio of tetrodotoxin-sensitive to total Na current (measured at −10 mV) was significantly higher in IB4-negative cells than in IB4-positive cells. In control cells, this fraction was 0.70 ± 0.08 in IB4-negative cells versus 0.36 ± 0.09 in IB4-positive cells (P = 0.008, Mann–Whitney rank sum test, T = 309, T = 308). In LID cells, this fraction was 0.30 ± 0.08 in IB4-negative cells versus 0.08 ± 0.03 in IB4-positive cells (P = 0.007, Mann–Whitney rank sum test, T = 309). The time constants for decay of the currents during a depolarizing pulse were not significantly altered by LID (figs. 4D and E).
LID Increases the Amplitude of Tetrodotoxin-resistant Na+Currents
The amplitude of the tetrodotoxin-resistant Na+current, as measured from a holding potential of −50 mV, which largely inactivates tetrodotoxin-sensitive current, was significantly increased after LID. This increase was observed only in the IB4-positive cells (figs. 5A and B; overall P value = 0.017), where it was more than twofold at 0 mV. In IB4-negative cells, the current densities were not significantly different (P = 0.50). The increased amplitude occurred without any shift in the voltage dependence of activation (fig. 5C; P = 0.54) or inactivation (fig. 5D; P = 0.32). In figures 5C and D, the data from IB4-positive and IB4-negative cells have been combined because differences between the two cell types were not observed for these parameters.
LID Increases the Amplitude of Tetrodotoxin-sensitive Na+Currents
The amplitude of the tetrodotoxin-sensitive current, measured as described in the Materials and Methods, was also markedly increased after LID (fig. 6A). In contrast to the tetrodotoxin-resistant current, the tetrodotoxin-sensitive current markedly increased in both IB4-negative and IB4-positive cells. The voltage dependence of activation (fig. 6B) and inactivation (fig. 6C) were not significantly altered by LID. In figures 6B and C, the data from IB4-positive and IB4-negative cells have been combined because differences between the two cell types were not observed for these parameters.
Excitability Changes after LID
Our current clamp experiments demonstrated that several measures of excitability increased in small-diameter DRG neurons after localized inflammation. Hence, rheobase and action potential threshold were both reduced after LID. In addition, the rising and falling rates of the action potential increased. Increased excitability is a feature of virtually all animal models of chronic pain, whether they model nerve injury, peripheral inflammation, or compression of the DRG.35It can be difficult to directly predict effects on excitability from voltage clamp measurements.36However, the decreased rheobase and increased rising rate of the action potential seem at least qualitatively consistent with the observed increases in Na+currents (particularly tetrodotoxin-sensitive currents, which activate more quickly and at more negative potentials). Although this study focuses on voltage-gated K+and Na+channels, it should be noted that other ion channels in sensory neurons may also play roles in chronic pain models, including the capsaicin receptor,37hyperpolarization-activated cation channels,38and voltage-gated calcium channels.39
The general characteristics of the tetrodotoxin-sensitive and tetrodotoxin-resistant Na+currents described in this study are similar to those reported in many other patch clamp studies of small DRG neurons.40–43As in most such studies, the tetrodotoxin-resistant currents we observed had characteristics attributed to the Nav1.8 isoform. We did not routinely observe persistent Na+currents that activate at very negative potentials and that are thought to be mediated by the Nav1.9 channel, probably because of slow inactivation at the holding potentials used, and/or because the current may “wash out” during whole cell patch clamp.44,45In addition, the work of Maruyama et al. 46suggests that under our recording conditions (adenosine triphosphate in the pipette, Cl−rather than F−as the anion), the persistent current would have a very small magnitude.
Changes in Na+channel expression or activity have been observed in other models of chronic pain. Nerve injury models generally lead to down-regulation of tetrodotoxin-resistant channels, with more complex effects on tetrodotoxin-sensitive channels. For example, after sciatic nerve transection, there is an up-regulation of tetrodotoxin-sensitive Nav1.3 channels and a down-regulation of tetrodotoxin-resistant Na+channels and of some other isoforms of tetrodotoxin-sensitive channels. In this case, increased excitability is thought to result from redistribution of Na channels along the axon and into the neuroma, as well as from the lower threshold and higher firing frequencies permitted by the switch to the Nav1.3 channel. Other nerve injury models such as spinal nerve ligation or chronic constriction injury also lead to reduced expression of tetrodotoxin-resistant Na+channels, although the findings about tetrodotoxin-sensitive currents are more variable (for review, see Waxman et al. 8and Rogers et al. 47). In contrast, models of peripheral inflammation almost always lead to an up-regulation of tetrodotoxin-resistant current, although effects on the tetrodotoxin-sensitive current are less consistent. In this regard, our results are perhaps most similar to those seen in the subcutaneous carrageenan injection model, in which both tetrodotoxin-sensitive and tetrodotoxin-resistant currents (the latter due to increased Nav1.8 but not 1.9) are up-regulated without shifts in voltage dependence of activation and inactivation.48,49These results are similar to ours except that the magnitude of the changes was much larger in our experiments—e.g. , twofold to threefold increase in tetrodotoxin-sensitive current after LID versus 40% increase after peripheral carrageenan, and 90% overall increase in tetrodotoxin-resistant current after LID versus 55% after peripheral carrageenan. This may be simply due to the fact that only a subset of L4 and L5 neurons should be affected by peripheral carrageenan, whereas all of the neurons in the LID model experience inflammation. Effects of peripheral inflammation on the tetrodotoxin-sensitive current are less consistent; particularly in models of visceral pain, there are examples of no increase or even decreases of tetrodotoxin-sensitive current after inflammation,15,50,51although most of these studies do report increases of varying magnitude of the tetrodotoxin-resistant current. The reasons for these differences between these various models are not yet clear. One possible explanation is that the visceral pain studies looked at a longer time point, 7–10 days instead of 3–5 days in our study and the carrageenan studies; perhaps the effect on tetrodotoxin-sensitive current is transient. Another possibility is that our study examined both cutaneous and visceral afferent neurons, which are mixed in the L4 and L5 DRG; if the increase in tetrodotoxin-sensitive current occurs only in cutaneous neurons, the differences between the above-cited studies could be accounted for.
The changes in Na+channels in nerve injury and inflammation models discussed above have in general been observed at the level of messenger RNA (mRNA) and protein as well as in functional studies, and are generally thought to reflect changes in channel expression. However, posttranslational modifications and alternative splicing have also been proposed as additional possible mechanisms.47These mechanisms are distinct from the rapid, reversible increase in Na+current caused by inflammatory mediators such as prostaglandins and bradykinin. Such effects, if present in the LID model in vivo , would not have been preserved in the acute culture methods used by this study and others cited above.
Small sensory neurons that are primarily nociceptors can be divided neurochemically into two populations: IB4-positive, primarily nonpeptidergic neurons, and IB4-negative peptidergic neurons. It has been shown that IB4-positive neurons depend on glial-derived neurotrophic factor, whereas IB4-negative neurons depend on nerve growth factor for survival during postnatal development.52Furthermore, these two populations of nociceptors terminate in distinct regions of the superficial spinal cord. Our finding that the IB4-negative cells had a higher expression of tetrodotoxin-sensitive current than did IB4-positive cells is in general agreement with previous studies. In our study, tetrodotoxin-resistant currents with the characteristics of Nav1.8 were commonly observed in both IB4-positive and IB4-negative cells, also in general agreement with previous electrophysiologic and protein or mRNA expression studies.32–34In our study, the increase in tetrodotoxin-resistant current after DRG inflammation was confined to IB4-positive cells, suggesting that this may be an important variable to consider when conducting such studies.
The voltage-gated K+currents described in this study are similar to those reported in numerous other patch clamp studies of acutely isolated small neurons, which have described sustained (or delayed rectifier) currents and transient currents with fast inactivation and slow inactivation during a depolarizing test pulse.53–57Calcium-activated K+currents also present in these cells58were not examined in the current study because of the presence of EGTA in the recording pipette and Co2+replacing Ca2+in the extracellular solution.
A number of studies of voltage-gated K+current changes in acutely isolated, small DRG neurons after peripheral inflammation or nerve injury have been reported, in different laboratories and animal models. These studies differ markedly from ours in that K+currents were increased in our study, but are reduced in most studies involving either peripheral inflammation or nerve injury. A common finding is the functional reduction of fast or slow inactivation due to a reduced total conductance and/or to a leftward shift in steady state inactivation that reduces channel availability at rest. This has been observed in several models of visceral inflammation as well as in the axotomy model of nerve injury. In some but not all of these models, the sustained current is also reduced.2,9–15In contrast, we observed a 36% increase in the maximum magnitude (i.e. , that evoked from a holding potential of −100 mV) of the sustained current, with comparable and significant increases at −60 mV and −50, closer to the physiologic resting potential. The fast inactivating component increased in maximum amplitude after LID, also in contrast to most other models. As in some other studies, we observed a leftward shift in steady state inactivation of the slow inactivating component; however, in effect this was apparently offset by small increases in magnitude such that there was no significant effect on this current at physiologic resting potentials.
Our excitability data indicate that the overall effect of the increases in both K+and Na+currents after LID is to increase excitability, which is in agreement with other models. A possible explanation for the differences in K+channel regulation between our study and other chronic pain models may lie in the magnitude of the Na+current changes observed. Although our findings of increased Na+current density are qualitatively similar to the results obtained in many models of peripheral inflammation, the magnitude of the changes is generally much larger in our LID model. It may be that basic intrinsic mechanisms regulating overall excitability of the neuron59dictate an increase in K+currents to partially compensate for the large increases in Na+currents. In addition, the LID model may expose cells to different types or higher concentrations of the regulatory molecules such as cytokines and growth factors that may regulate the ion channels.22In this vein, it is interesting that 24-h exposure to the proinflammatory cytokine interleukin 1β can increase tetrodotoxin-sensitive Na+currents by 67% in trigeminal nociceptors.60In preliminary experiments, we have also observed marked up-regulation of both tetrodotoxin-sensitive and tetrodotoxin-resistant Na+currents after overnight exposure to proinflammatory cytokines at concentrations similar to those observed in vivo after LID.
This study demonstrates that localized inflammatory irritation of lumbar DRG increases the magnitude of both voltage-gated K+and Na+currents. The effect on Na+currents is similar to that seen in some models of peripheral inflammation, except in being particularly large, but the finding of increased K+current is rather unique to this model. Measurements of action potential parameters in current clamp mode indicated that the overall effect was increased excitability. Therefore, these ion channel changes may help to account for the pain and hyperalgesia associated with LID.