Based on our previous data indicating the inflammation-induced changes in the regulation of intracellular Ca²⁺ 5 and in voltage-gated Ca²⁺ currents 6 is restricted to small and medium diameter cutaneous neurons as well as in vivo data suggesting that nociceptive afferents innervating cutaneous tissue tend to have a small cell body diameter 13 , we focused on neurons with a cell body diameter <30 μm in the present study. Consistent with results of our previous study, cutaneous neurons from inflamed rats were significantly more excitable than those from naïve rats, where the increase in excitability was manifest with a small but significant decrease in action potential threshold (from −32 ± 0.8 to −34.9 ± 1.0 mV, p = 0.03: n = 53 and 38 for naïve and inflamed groups, respectively), decrease in rheobase (3.7 ± to 2.3 pA/pF, p < 0.05) and increase in the response to suprathreshold current injection (i.e., the number of action potentials evoked in response to current injection 3x rheobase increased from 2.5 ± 0.3 to 6.9 ± 0.9, p < 0.01).

Closer inspection of this dataset, suggested that there were at least two populations of small diameter neurons that could be defined by their response to inflammation: in one there was a clear increase in excitability while in the other, the increase was less apparent. Data suggest that sensory neurons defined by their binding to the lectin IB4 may play a differential role in the hypersensitivity that develops in the presence of inflammation 14 15 , We therefore repeated the excitability experiments on neurons incubated in FITC-labeled IB4 10 minutes prior to current clamp recording. The passive electrophysiological properties of the neurons in this dataset were analyzed with a two way ANOVA where IB4 binding and inflammation were the main factors. This analysis revealed a significant (p < 0.01, Table 1) influence of inflammation on membrane capacitance which was increased in cutaneous neurons from inflamed rats. There was also a significant interaction between inflammation and IB4 binding with respect to resting membrane potential where post-hoc analysis indicated that the difference between IB4+ and IB4- neurons in the inflamed group was significant (p = 0.03, Table 1), and the difference between IB4- neurons from the naïve and inflamed groups was also significant (p < 0.05, Table 1).

† Interaction between Manipulation and IB4 binding is significant with p = 0.03.

‡ The impact of Manipulation (inflammation) is significant with p < 0.01.

V_rest is resting membrane potential. R_in is input resistance. CFA is complete Freund’s adjuvant.

Results of the analysis of the excitability data indicated that there were not only differences between IB4+ and IB4- neurons with respect to baseline excitability, but the changes in excitability 3 days after the induction of inflammation. That is, while the IB4- neurons from naïve rats were more excitable than IB4+ neurons, there were no detectable inflammation-induced changes in the excitability of IB4- neurons: action potential threshold was −35.8 ± 2.3 mV (n = 13) in neurons from naïve rats and −29.0 ± 2.8 mV (p > 0.05, n = 9) in neurons from inflamed rats; rheobase was 4.2 ± 1.4 pA/pF and 6.9 ± 1.6 pA/pF (p > 0.05) in neurons from naïve and inflamed rats, respectively; and the slope of the stimulus response function to suprathreshold current injection was 0.52 ± 0.4 and 0.51 ± 0.5 (p > 0.05) in neurons from naïve and inflamed rats, respectively. In contrast, there was a significant increase in the excitability of IB4+ neurons with a significant decrease in action potential threshold (Figure 1A), decrease in rheobase (Figure 1B), and increase in the response to suprathreshold current injection (Figure 1C and D). Because persistent changes in excitability appeared to be restricted to IB4+ neurons, we focused on IB4+ neurons for the remainder of the study. As there were no significant differences between IB4+ neurons from naïve and inflamed rats with respect to resting membrane potential or input resistance (Table 1) and the increase in membrane capacitance, would work to attenuate excitability, we next assessed active electrophysiological properties to begin to identify mechanisms that may have contributed to the inflammation-induced increase in the excitability of IB4+ neurons. Of the active properties assessed, the action potential duration was the only property that was significant different between groups, and it was significantly longer in neurons from inflamed rats (Figure 1E, p < 0.01). There was also a trend toward an increase in the decay rate of the afterhyperpolarization (AHP) (Figure 1F, p = 0.06) in cutaneous neurons from inflamed rats. Given the role of K⁺ currents in both the action potential duration and the afterhyperpolarization (AHP), these changes in active electrophysiological properties are consistent with a decrease in a K⁺ current.

To determine whether a decrease in voltage-gated K⁺ (Kv) current contributed to the increase in excitability of cutaneous IB4+ neurons, we assessed the biophysical properties of Kv currents in cutaneous neurons from inflamed and naïve rats. Kv currents were recorded in Ca²⁺ free bath solution containing 2.5 mM Co²⁺ to eliminate contribution from BK_Ca channels. Based on previous data indicating that there are Kv currents both subject, and resistant, to steady-state inactivation 16 , we first assessed steady-state inactivation of Kv currents in each neuron studied (Figure 2A, B). With this protocol, it was possible to determine the total current available for activation, as well as the fractions of total current subject to (inactivatable), or resistant to (non-inactivatable), steady-state inactivation. There was no detectable influence of inflammation on either the voltage-dependence of steady of inactivation or the fraction of current resistant to steady-state inactivation (Figure 2B, Table 2). Nor was there a detectable influence of inflammation on the voltage-dependence of current activation (Figure 2B). Furthermore, analysis of the peak current density of the total (not shown), inactivatable (Figure 2C) and non-inactivatable (Figure 2D) current in cutaneous neurons from naïve rats as well as both ipsilateral and contralateral to the site of inflammation revealed no significant differences between groups (p > 0.05). Finally, there were no significant differences between cutaneous neurons from naïve and inflamed rats with respect to the rate of Kv current activation, inactivation, or deactivation (data not shown). Despite evidence for inflammation-induced changes in Kv current in other populations of neurons 9 , results of this series of experiments suggest that this is not the case for cutaneous neurons.

N is the number of neurons studied with the number of animals indicated in parenthesis. Cap is capacitance. Imax is maximal current estimated from steady-state inactivation data fitted with a Boltzmann function. Slope, V0.5 and a are parameters determined from the same Boltzmann function where Slope is the slope of the I-V curve, V0.5 is the potential at which half of the inactivatable current is inactivated, and “a” is the percentage of total current resistant to steady state inactivation.

Based on evidence that large conductance Ca²⁺-modulated K⁺ (BK_Ca) currents contribute to action potential duration as well as the decay of the AHP 7 8 17 , as well as evidence that inflammation results in a decrease in high threshold voltage-gated Ca²⁺ channels in cutaneous DRG neurons, we next assessed the impact of inflammation on BK_Ca currents in cutaneous neurons. Iberiotoxin (IbTx, 100 nM) or paxilline (10 μM) were used to isolate BK_Ca currents from the total current evoked in IB4+ cutaneous neurons from naïve and inflamed rats (Figure 3). The IbTx sensitive current was significantly smaller in cutaneous neurons from inflamed rats than from naïve rats (Figure 3), at voltage steps to potentials ≥ −10 mV. Comparable results were obtained with paxilline, where peak paxilline sensitive current in IB4+ neurons from naïve rats (153.3 ± 28.07 pA/pF, n = 17) was significantly (p < 0.01) larger than that in cutaneous neurons from inflamed rats (49.5 ± 18.04 pA/pF, n = 10).

As previously described 8 , BK_Ca currents in cutaneous neurons from naïve rats were in general, rapidly activating, with a variable degree of inactivation during sustained depolarization (Figure 4A). BK_Ca currents evoked in neurons from inflamed rats appeared to activate more slowly and demonstrated little, if any detectable inactivation during a 500 ms depolarizing voltage step (Figure 4B). Consistent with this impression, the average time constant for IbTx sensitive current activation was significantly larger in neurons from inflamed rats than in neurons from naïve rats (Figure 4C, D). Comparable results were obtained with paxilline. Deactivation of toxin sensitive currents was assessed with a tail current protocol evoked before and after toxin application (Figure 4E inset). While there was no influence of inflammation on the deactivation of IbTx sensitive currents, paxilline sensitive currents in cutaneous neurons from inflamed rats deactivated significantly more slowly than those from naïve rats (Figure 4E, F).

While the decrease in high threshold voltage-gated Ca²⁺ current previously described 6 could be sufficient to account for the inflammation-induced decrease in BK_Ca current density, two experiments were performed to begin to assess the possibility that a decrease in channel protein also contributed to the decrease in current. The first was a real time PCR analysis of BK_Ca subunit mRNA levels in L4/5 ganglia. GAPDH was used as an internal comparator as Ct values for amplification of GAPDH were comparable for ganglia from naïve (17.0 ± 0.2, n = 4) and inflamed (17.5 ± 0.3, n = 4) rats. No significant changes in mRNA levels were detected for either total α subunit or a splice variant of the α subunit containing the STREX insert, nor were there significant changes detected in mRNA levels for β2, 3 or 4: values for 2^ΔΔCt were all close to 1.

Given evidence for inflammation-induced changes in protein in the absence of detectable changes in mRNA 18 , the second experiment was to assess changes in BK_Ca subunit protein levels. Total protein from L4 and L5 ganglia from naïve (n = 4) and inflamed (n = 5) rats was recovered, processed for Western blot analysis, and probed with antibodies specific to the α and β2-4 BK_Ca channel subunits. The multiple bands present in the blots of the α subunit are consistent with the results of our previous PCR analysis of the BK_Ca channel in DRG which indicated that a number of splice variants of the α subunit are expressed 8 . Nevertheless, there was no detectable influence of inflammation on relative BK_Ca subunit protein levels (Figure 5).

In our previous analysis of BK_Ca currents in cutaneous DRG neurons, we noted the presence of considerable heterogeneity in the biophysical properties of currents between neurons, with variability in the rate of current activation as well as in the rate and extent of current inactivation 8 . In neurons from inflamed rats, the currents were considerably more homogeneous with properties similar to those illustrated in Figure 3. Given the influence of both splice variants in the α-subunit 19 , and β-subunits 20 on the biophysical properties of BK_Ca currents, this change raised the possibility that the changes in biophysical properties were due to changes in the subunit expression pattern. Subunit expression in cutaneous neurons was assessed with single cell RT-PCR (Figure 6). Results of this analysis indicated that there was a significant increase in the proportion of neurons in which all 4 splice variants (Figure 6, zero insert, as well as all 3 higher molecular weight species) were detected at the X4 splice site. This site corresponds to the stress-axis regulated exon (STREX) site previously described by others 20 . There was also a significant decrease in the proportion of neurons in which β2 and β3 subunits were detected.

Consistent with previous studies, persistent inflammation results in an increase in the excitability of sensory neurons innervating the site of inflammation that is detectable in acutely isolated sensory neurons. This increase in excitability was restricted to the IB4+ subpopulation of neurons and was associated with a significant increase in action potential duration with no additional changes in any other passive or active electrophysiological property. These changes in excitability and action potential waveform were not associated with any detectable change in Kv current. However, they were associated with a significant decrease in IbTx- and paxilline-sensitive current. This decrease in current was not associated with a change in total BK_Ca subunit mRNA or protein as assessed at the whole ganglion level. Furthermore, the decrease in IbTx-and paxilline-sensitive current was associated with changes in the biophysical properties of the current as well as changes in the pattern of splice variant expression of the BK_Ca channel α subunit and a decrease in the proportion of neurons in which the β2 and 3 subunits were detected. These results are consistent with the suggestion that a decrease in BK_Ca current contributes to the inflammation-induced sensitization of cutaneous afferents.

The pattern of changes in excitability observed in cutaneous neurons were comparable to patterns observed in afferents innervating other target tissue including muscle 9 , joint 21 and bladder 11 . However, there appear to be subtle, yet potentially important differences in the associated changes in passive and active electrophysiological properties. For example, the only significant change in the passive or active electrophysiological properties associated with the inflammation-induced sensitization of temporomandibular joint afferents is a decrease in the duration of the AHP 21 , while in bladder afferents, there is an increase in membrane capacitance with no change in action potential duration 11 . Because passive and active properties reflect the actions of ion channels, these differences suggest differences in underlying ionic mechanisms. These differences also highlight the fact that comparable increases in excitability can be achieved through a number of different mechanisms.

Changes in passive and active electrophysiological properties assessed in current clamp can be used to predict the ion channels that contribute to the inflammation-induced increase in excitability. For example, the absence of a detectable change in V_rest or R_in argues against a channel active at the resting membrane potential such as the Cl^- conductance activated by inflammatory mediators we recently described in dural afferents 22 . Similarly, the increase in action potential duration and trend toward a decrease in duration of the AHP in association with a decrease in rheobase and action potential threshold are consistent with a decrease in a K⁺ current activated with membrane depolarization that contributes to membrane repolarization on the falling phase of the action potential as well as the duration of the AHP. However, in contrast to masseter muscle 9 and bladder 9 afferents, where an inflammation-induced decrease in Kv current subject to steady-state inactivation was readily detectable, no changes in Kv currents were detected in cutaneous neurons. These results suggest that a decrease in a Ca²⁺ dependent K⁺ current is likely to contribute to the inflammation-induced increase in the excitability of cutaneous neurons.

Consistent with the prediction based on the negative results obtained with Kv currents, we observed significant decrease in IbTx and paxilline sensitive current. The observation that the paxilline sensitive currents were larger than the IbTx sensitive currents is consistent with our single cell PCR data suggesting that the β4 subunit is present in the majority of IB4+ cutaneous neurons and previous data suggesting that the β4 subunit confers resistance to IbTx 23 but not paxilline 24 . It is important to point out that a decrease in BK_Ca current was not necessarily a foregone conclusion. That is, in addition to the inflammation-induced decrease in voltage-gated Ca²⁺ currents recently described in cutaneous neurons 6 , there is also an inflammation-induced increase in the magnitude and decrease in the decay of the depolarization-evoked increase in the concentration of intracellular Ca²⁺ ([Ca²⁺ _i) in cutaneous neurons 5 . BK_Ca channels coupled to the latter process would have resulted in an increase in channel activity and a decrease in excitability 17 . However, the observed decrease in BK_Ca current suggests that the activation of these channels in cutaneous afferents is tightly coupled to an increase in intracellular Ca²⁺ mediated via Ca²⁺ influx through voltage-gated Ca²⁺ channels. Consistent with this suggestion is the observation that a decrease in high threshold Ca²⁺ current alone is sufficient to increase excitability, at least in a subpopulation of cutaneous afferents 8 . The observation that BK_Ca channels are largely restricted to IB4+ small diameter cutaneous neurons 8 is also consistent with this suggestion given the observation that the inflammation-induced increase in excitability was restricted to this subpopulation of neurons despite our previous observation that the inflammation-induced decrease in voltage-gated Ca²⁺ currents is detectable in both IB4+ and IB4- neurons 6 .

The absence of a detectable inflammation-induced change in BK_Ca channel subunit mRNA or protein at the whole ganglia level is also consistent with the suggestion that the decrease in BK_Ca current was largely secondary to the decrease in voltage-gated Ca²⁺ current rather than a decrease in BK_Ca channels. This suggestion is made with caution, however, given that afferents innervating the site of inflammation constitute far less than half of the neurons in L4 and L5 ganglia and as a result, a decrease in mRNA or protein in a subpopulation of these neurons may not have been detectable.

Changes in the biophysical properties of the BK_Ca currents in neurons from inflamed rats suggest that a decrease in high threshold voltage-gated Ca²⁺ current alone is insufficient to account for all of the changes in BK_Ca currents. At least some of the changes in biophysical properties appear to reflect changes in the pattern of BK_Ca channel subunit expression. That is, β3 subunits confer rapid channel activation, and β2 confers both rapid channel activation and channel inactivation 20 . Thus, a decrease in the number of cutaneous neurons in which these subunits were detected is consistent with the more slowly activating persistent currents observed in neurons from inflamed animals. However, at least two of the inserts at the X4 site of the BK_Ca α subunit result in a dramatic leftward shift in the voltage-dependence of channel activation 19 , an increase in the rate of channel activation as well as a decrease in the rate of channel deactivation. The decrease in rate of paxilline-sensitive, but not IbTx-sensitive current deactivation suggests there may be preferential assembly of channels with an X4 insert in the α-subunit and a β4 subunit (resulting in a channel with a slower deactivation rate that is resistant to IbTx), but additional biochemical data would be needed to confirm this prediction. Several possibilities could account for the failure to detect a larger influence of the splice variants on the whole cell current. These include: 1) that the larger α subunit transcripts are not translated as efficiently, 2) that the larger α subunits are not assembled in inflamed neurons, or 3) that the trafficking of these larger channels is also altered such that they are targeted to other parts of the afferent than the cell soma. These latter two possibilities are predicated on the assumption that at least some of the variability in biophysical properties of BK_Ca currents observed in neurons from naïve animals 8 is due to the presence of splice variants of the α subunit.

Cutaneous afferents were retrogradely labeled with DiI as previously described 5 8 . Briefly, rats were anesthetized with isofluorane and 10 μl of 1,1’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine perchlorate (DiI, Invitrogen, Carlsbad CA, 17 mg/ml in saline diluted from a stock of 170 mg/ml in DMSO) was injected at 3–5 sites (with 1.5 - 2 μl per site) with a 30 g needle directed into the epidermis. Rats were studied 14 to 17 days post DiI injection. Inflammation was induced in a subgroup of rats with a 100 μl subcutaneous injection of complete Freund’s adjuvant (CFA, mixed 1:1 in saline) into the same site previously labeled with DiI. This injection was also made under isofluorane-induced anesthesia. This group of rats was studied 3 days after CFA injection.

Acutely dissociated cutaneous neurons were obtained as previously described 5 8 . Briefly, rats were deeply anesthetized with 1 ml/kg rat cocktail (55 mg/ml ketamine, 5.5 mg/ml xylazine and 1.1 mg/ml acepromazine) and L4 and L5 dorsal root ganglion (DRG) were harvested, enzymatically treated, mechanically dissociated and plated onto laminin-ornithine coated cover slips. After 2 hrs of incubation at 37°C/3% CO₂, cover-slips were flooded with HEPES buffered L-15 media and stored at room temperature during the period of recording (< 8 hrs after removal from the animal).

Isolated cutaneous neurons were studied with conventional whole-cell and perforated patch configurations with an Axopatch 200B (Medical Devices Sunnyvale CA) controlled with pClamp (v 8.2, Molecular Devices) or a HEKA EPC9 amplifier (HEKA Electronik, Lambrecht/Pfalz Germany) controlled with Pulse software (V8.8, HEKA). The conventional whole cell configuration was used for current clamp recording and for voltage-clamp analysis of Kv currents while the perforated patch configuration was used for the analysis of BK_Ca currents. For current clamp recording, the bath solution contained (in mM): NaCl, 130; KCl, 5; CaCl₂, 2.5; MgCl₂, 0.6; HEPES, 5; and glucose, 10; pH adjusted to 7.4 with Tris-Base, and osmolality adjusted to 320 mOsm with sucrose. For voltage-clamp recording of Ca²⁺ dependent K⁺ currents, NaCl was replaced with choline-Cl to eliminate voltage-gated Na⁺ currents. This same solution was used for voltage clamp recording of voltage-gated K⁺ currents, except that CaCl₂ was replaced with CoCl₂ to eliminate voltage-gated Ca²⁺ currents. The electrode solution used for current clamp and to record voltage-gated K⁺ currents contained (in mM): KCl, 30; K-methanesulfonate (MES), 110; MgCl₂,1; CaCl₂, 0.1; EGTA, 1; HEPES, 10; ATP-Mg, 2; GTP, 1; pH adjusted to 7.2 with Tris-Base, and osmolality adjusted to 310 mOsm with sucrose. The electrode solution used for perforated patch clamp recordings contained (in mM): KCl, 30; K-MES, 110; MgCl₂, 1; HEPES, 10; EGTA, 0.1; pH adjusted to 7.2 with Trisbase, and osmolality adjusted to 310 mOsm with sucrose. Amphotericin B, used to obtain whole cell access for perforated patch recording, stock solution was prepared in DMSO (90 mg/ml) then diluted to a final concentration (600 μg/ml) in electrode solution immediately prior to use. All salts used for electrophysiology were obtained from Sigma-Aldrich (St Louis MO).

Dorsal root ganglia (DRG) from anesthetized male rats were harvested in a manner identical to that used for neuron isolation and plating. mRNA was extracted and cDNA synthesis performed as previously described 6 , except that that random hexomers were used to prime the reverse transcription reaction. SYBR Green was used to monitor amplification of template with primers on a real-time thermal cycler (Life Technologies, Grand Island NY) controlled by a PC running Prism 7000 SDS software. A melting curve was generated at the end of each experiment to assess for the presence of contamination. Amplification efficiency was determined for each target gene. The ΔΔCT method was used to assess differences in relative expression levels. Primers for amplification of the core BK_Ca α subunit were: F - TGTCATGATGACGTCACAGATCC, R -TTTTTTTGGTGACAGTGTTGGC; those for amplification of the BK_Ca α subunit with the STREX insert were: F - AGCCGAGCATGTTGTTTTGAT, R- ACGCACACGGCCTGACA; while those for GAPDH were: F - GGCCTACATGGCCTCCAA, R -TGGAATTGTGAGGGAGATGCT. Commercially available primer sets were used for the amplification of β2, 3 and 4 (Qiagen).

Single cell PCR was performed as previously described 26 . Because perforated patch recording is relatively slow and many neurons are needed from a single animal to obtain a reasonable estimate of the proportion of neurons from a single animal, a different set of neurons was used for single cell PCR analysis. Following identification of cutaneous neurons under epifluorescence illuminations, neurons were collected with large bore (~30 μm) glass pipettes and expelled into microcentrifuge tubes containing reverse transcriptase (RT) mix. RT-PCR was performed as described previously 26 utilizing an anchored primer (5’-ttttttttttttttttttvn-3’; v = a, c, or g; n = a, c, g, or t, from Life Technologies) for the RT reaction and a nested PCR amplification strategy for the PCR reaction. rslo primer sequences were identical to those described previously 8 . For each cell preparation, at least two tubes were run in which no cell was collected (although the electrode was manipulated in a manner identical to that used for cell collection) and at least two additional tubes were run in which no reverse transcriptase was added to the RT mix prior to the RT reaction. Cyclophillin (0.5 μl of cDNA) was used to monitor the success of the cell collection/RT reaction: only neurons in which cyclophillin was detected were used for further analysis. 5 μl of PCR products were loaded onto 2% agorase/TAE gel.

DRG (L4/L5) were rapidly removed from deeply anesthetized rats and homogenized in solubilization buffer (50 mM Tris.HCl, pH8.0; 150 mM NaCl, 1 mM EDTA, 1% NP40, 0.5% deoxycholic acid, 0.1% SDS, 1 mM Na₃VO₄, 1 U/ml aprotinin, 20 μg/ml leupetin, 20 μg/ml pepstatin A). The homogenate was centrifuged at 20,000 X g for 10 min at 4°C. The supernatant was removed. Protein (50–120 μg) was separated on a 7.5-10% SDS-PAGE gel and blotted to nitrocellulose membrane (Amersham) with a Trans-Blot Transfer Cell system (Bio-Rad). Blots were blocked with 5% milk in TBS buffer (20 mM Tris, 150 mM NaCl pH 7.4) at room temp for 1 hour. After decanting the blocking buffer, the blots were incubated with primary antibodies. These included: the α subunit of the BK_Ca channel (AKA slo1, NeuroMab clone L6/60, NeuroMab, Davis CA: 1:200), BK_Ca β2 (NeuroMab clone N53/32: 1:200), BK_Ca β3 (NeuroMab clone N40B/18: 1:200), BK_Ca β4 (NeuroMab clone L18A/3: 1:200), and GAPDH (sc-25778, Santa Cruz Biotechnology: 1:1000). The specificity of all BK_Ca subunit antibodies has been confirmed in heterologous expression systems where there was no evidence of cross reactivity with other BK_Ca subunits or KV2.1. Membranes were incubated with BK_Ca subunit antibodies overnight at 4°C, and GAPDH for 1 hour at room temperature. Membranes were washed with TBS buffer and incubated for 1 hour with anti-goat IgG horseradish peroxidase (1:3000, Santa Cruz) in 5% milk/TBS. Membranes were then washed with TBS buffer. The immunoreactivity was detected using Enhanced Chemiluminescence (ECL, Amersham). Chemiluminescence was captured with a CCD camera (Las-3000, Fujifilm) and analyzed with Fuji software Multi Gauge. The relative protein levels were obtained by comparing target protein to loading control (GAPDH) in the same membrane.

Data are expressed as mean ± S.E.M unless otherwise stated. Student’s t test, one- and two-way ANOVA with the Holm-Sidak post hoc test were used for comparisons of parametric data between groups. For single cell PCR analysis, between 30 and 40 cutaneous neurons were collected from each animal, although only 29 neurons were collected for one of the 4 naïve animals used to assess changes in α-subunit splice variants. The proportion of the total number of neurons from each animal in which a splice variant or β-subunit was detected was used as the “statistic” for that animal, where the mean proportion of expression in naïve animals was compared to that in inflamed animals. Statistical significance was assessed at p < 0.05.