|Home | About | Journals | Submit | Contact Us | Français|
Migraineurs experience debilitating headaches that result from neurogenic inflammation of the dura and subsequent sensitization of dural afferents. Given the importance of inflammatory mediator (IM)-induced dural afferent sensitization to this pain syndrome, the present study was designed to identify ionic mechanisms underlying this process. Trigeminal ganglion neurons from adult female Sprague-Dawley rats were acutely dissociated 10–14 days after application of retrograde tracer DiI onto the dura. Modulation of ion channels and changes in excitability were measured in the absence and presence of IM (prostaglandin 1, bradykinin 10, and histamine 1 (µM)) using whole cell and perforated patch recordings. Fura-2 was used to assess changes in intracellular Ca2+. IM modulated of a number of currents in dural afferents including those both expected and/or previously described, (i.e,. an increase in tetrodotoxin resistant voltage-gated Na+ current (TTX-R INa) and a decrease in voltage-gated Ca2+ current) as well currents never before described in sensory neurons (i.e., a decrease in a Ca2+-dependent K+ current and an increase in a Cl− current), and produced a sustained elevation in intracellular Ca2+. While several of these currents, in particular TTX-R INa appear to contribute to the sensitization of dural afferents, the Cl− current is the primary mechanism underlying this process. Activation of this current plays a dominant role in the sensitization of dural afferents due to the combination of the density and biophysical properties of TTX-R INa, and the high level of intracellular Cl− in these neurons. These results suggest novel targets for the development of anti-migraine agents.
Migraine is a neurological disorder characterized by incapacitating head pain. Compelling evidence indicates that neurogenic inflammation of the dura and subsequent dural afferent sensitization are fundamentally important for initiating migraine pain (Ray and Wolff, 1940; Strassman et al., 1996; Sarchielli et al., 2000; Burstein, 2001). However, the ionic mechanisms involved in dural afferent sensitization have yet to be discovered.
A feature that distinguishes nociceptive afferents from all other primary sensory neurons is that they can be sensitized by mediators released at sites of inflammation. Identification of the underlying mechanisms of sensitization remains an active area of investigation because of the critical role this increase in afferent excitability plays in ongoing pain and hypersensitivity (hyperalgesia) observed in the presence of tissue injury. Given the direct link between ion channel activity and neuronal excitability, ion channels have remained a primary focus of this line of investigation. Recent evidence suggests that the specific ion channels underlying the sensitization of nociceptive afferents varies as a function of target of innervation. For example, there are differences between the mechanisms underlying the acute sensitization of afferents innervating the colon and those innervating glabrous skin (Gold and Traub, 2004). Similar differences have been described for afferents innervating the muscle (Harriott et al., 2006), ileum (Stewart et al., 2003), bladder (Yoshimura and de Groat, 1999) and colon (Beyak et al., 2004) in response to persistent inflammation. While the differences described to date suggest the involvement of distinct neurobiological processes, increases in voltage-gated Na+ and/or decreases in voltage gated or Ca2+-dependent K+ currents appear to be mechanisms common to sensitization of nociceptive afferents in these previous studies. In marked contrast to the results from these previous studies, we recently described inflammatory-mediator (IM)-induced changes in dural afferents that appeared to reflect processes in addition to those previous described (Harriott and Gold, 2009). Given that IM-induced sensitization of this population of afferents appears to play an essential role in the headache associated with migraine (Ray and Wolff, 1940; Strassman et al., 1996; Sarchielli et al., 2000; Burstein, 2001), the purpose of the present study was to identify the mechanisms underlying the IM-induced sensitization of these afferents.
Acutely dissociated retrogradely labeled dural afferents from adult female rats were studied with whole cell patch clamp and Ca2+ imaging techniques. Results from this analysis suggest that in addition to changes in Na+ currents common to the sensitization of other afferent populations, and the inhibition of a K+ current that appears to be unique to dural afferents, the primary mechanism of IM-induced sensitization of dural afferents appears to be the activation of a Cl− current.
Adult female Sprague Dawley rats (Harlan, Indianapolis, IN) weighing between 180–290 g were used for all experiments. Rats were housed two per cage at the University of Pittsburgh animal facility on a 12:12 light: dark schedule with food and water freely available. Prior to all procedures, animals were deeply anesthetized with an i.p. injection (1 ml/kg) of cocktail containing ketamine (55mg/kg), xylazine (5.5 mg/kg) and acepromazine (1.1 mg/kg). Experiments were approved by the University of Pittsburgh Institutional Animal Care and Use Committee and performed in accordance with National Institutes of Health guidelines for the use of laboratory animals in research. All efforts were employed to minimize the total number of animals used.
Afferents innervating the dura were identified as previously described following application of DiI to the dura (Harriott and Gold, 2008). Ten to fourteen days following DiI application, trigeminal ganglia (TG) were removed, enzymatically treated and mechanically dissociated as previously described (Harriott et al., 2006). Changes in currents, intracellular Ca2+, and excitability were measured 2–8 hours after cells were plated.
All whole cell and perforated patch-clamp recordings were performed with a HEKA EPC10 amplifier (HEKA Electronik, Lambrecht/Rhineland-Pfalz, Germany). Data were low-pass filtered at 5–10 kHz with a four-pole Bessel filter and digitally sampled at 25–100 kHz. Ionic solutions were chosen to study specific currents in isolation. For all solutions, pH was adjusted to between 7.2 and 7.4 with Tris-base (unless otherwise stated) and the osmolality adjusted to between 310 and 325 mOsm with sucrose. Thick walled borosilitate glass (1.5 mm internal diameter, WPI) electrodes were pulled (Sutter P2000) such that when filled with electrode solution the resistance was < 5 MΩ.
A standard protocol was used to facilitate comparisons between neurons. After establishing whole cell access, membrane resistance and capacitance were determined with hyperpolarizing voltage steps from −60 mV. Baseline data were collected over 2 to 10 minutes to ensure the particular current under study was stable as well as facilitate detection of IM-induced changes. IM were then applied and changes in currents were monitored. Voltage-clamp protocols were employed to assess the impact of IM on current activation, inactivation and deactivation as well as Ca2+ dependence with the specific details of each protocol described in conjunction with the description of the specific currents. To facilitate assessment of the reversal potential for IM-induced currents, a voltage ramp from +50 to −100 mV over 100 ms was used. Finally, gramicidin perforated patch recording was used to assess the resting concentration of intracellular Cl− in dural afferents. Bath and electrode solutions were constructed to either reflect physiological solutions or to facilitate the study of specific currents in isolation. The details of each solution used are as follows:
To isolate Na+ currents, the electrode solution was composed of (mM) Cs-methanesulfonate 100, tetraethylammonium (TEA) -Cl 40, NaCl 5, CaCl2 1, MgCl2 2, HEPES 10, EGTA 11, Mg-ATP 2, and Li-GTP 1. Bath solution contained (mM) Na-methanesulfonate 35, Choline-Cl 65, TEA-Cl 30, CaCl2 2.5, MgCl2 5, CdCl2 0.05, HEPES 10, glucose 10.
To isolate K+ currents, the electrode solution was composed of (mM) K-methanesulfonate 110, KCl 30, NaCl 5, CaCl2 1, MgCl2 2, HEPES 10, EGTA 11, Mg-ATP 2, and Li-GTP 1. Bath solution contained (mM) KCl 3, Choline-Cl 130, CaCl2 2.5, MgCl2 0.6, NFA 0.1, HEPES 10, glucose 10.
To isolated Ca2+ currents, the electrode solution was composed (mM) Cs-methanesulfonate 100, Na-methanesulfonate 5, TEA-Cl 40, CaCl2 1, MgCl2 2, EGTA 11, HEPES 10. Bath solution contained Choline-Cl 100, TEA-Cl 30, CaCl2 2.5, MgCl2 0.6, NFA 0.1, HEPES 10, glucose 10.
Initial characterization of IM evoked currents were performed with an electrode solution composed of (mM) K-methanesulfonate 110, KCl 30, NaCl 5, CaCl2 1, MgCl2 2, HEPES 10, EGTA 11, Mg-ATP 2, Li-GTP 1 and bath solution containing (mM) KCl 3, NaCl 130, CaCl2 2.5, MgCl2 0.6, HEPES 10, glucose 10.
The contribution of monovalent and divalent cations to the IM-induced current was assessed by manipulating concentrations of K+ and Na+ and Ca2+. The contribution of monovalent cations to the IM-induced current was minimized with an electrode solution containing (mM) Cs-methanesulfonate 100, CsCl 30, CaCl2 1, MgCl2 2, HEPES 10, EGTA 11, Mg-ATP 2, Li-GTP 1 and bath solution containing (mM) Choline-Cl 130, CaCl2 2.5, MgCl2 0.6, HEPES 10, glucose 10.
To further analyze the source of Ca2+ responsible for the IM-induced activation of Cl− currents, 4 different manipulations were formed: 1) the addition of Cd2+ (50 µM) to the bath solution to block influx via voltage-dependent Ca2+ channels, 2) the addition of ruthenium red to the bath solution to block influx via ligand gated ion channels, 3) the substitution of BAPTA (10 mM) for EGTA (11 mM) in the electrode solution, and 4) the use of the combination of an electrode solution with Ca2+ artificially buffered to ~620 nM with an electrode solution containing EGTA (1.2 mM), Ca2+ (1 mM) and Mg2+ (2 mM), in which influx via voltage-gated Ca2+ channels was also blocked by the addition of Cd2+ (50 µM) to the bath solution. MaxChelator© was used to generate estimates of resting free intracellular Ca2+.
The reversal potential of IM-induced currents were determined using gramicidin perforated patch recordings and a bath solution containing (mM) Choline-Cl 100, TEA-Cl 30, CaCl2 2.5, MgCl2 0.6, HEPES 10, glucose 10.
Excitability was assessed in the presence and absence of test compounds as previously described (Harriott and Gold, 2009). A neuron was considered sensitized if application of a test solution resulted in a hyperpolarization of action potential threshold, decrease in rheobase, and/or an increase in the response to suprathreshold stimulation greater than 2 SD’s from the baseline mean. Current injection was used in the present study as a means to bypass natural processes underlying the transduction of what would be largely mechanical (changes in vessel diameter) and chemical stimuli in the dura. Given evidence that IM can sensitize transduction processes (i.e., TRPV1) as well as the ion channels underlying action potential initiation and propagation, changes in excitability described in the present study would serve to amplify IM-induced modulation of transduction processes which should also occur at afferent terminals in vivo.
Neurons were incubated with 2.5 µM Ca2+ indicator fura-2 AM ester with 0.025% pluronic as described previously (Lu et al., 2006) and IM-induced Ca2+ transients were acquired on a PC running Metafluor software (Molecular Devices, Sunnyvale, CA) via a CCD camera (Roper Scientific, Trenton, NJ, Model RTE/CCD 1300). The ratio (R) of fluorescence emission (510 nm) in response to 340/380 nm excitation [controlled by a lambda 10–2 filter changer (Sutter Inst.; CA)] was acquired at 1Hz during IM applications.
All salts and test compounds were obtained from Sigma-Aldrich (St. Louis, MO). IM consisted of bradykinin (10 µM), histamine (1 µM), and prostaglandin E2 (1 µM), where bradykinin was dissolved in 1% acetic acid (23.58mM stock concentration), PGE2 was dissolved in 100% ETOH (10mM stock concentration), and histamine was dissolved in water (100mM stock concentration). All stock solutions were stored at −20°C until the day of use. IM-vehicle bath containing the final concentration of ETOH (0.01%) and acetic acid (0.001%) was used as a control. Niflumic acid (NFA) was dissolved in 100% ETOH. Ruthenium red was dissolved in water.
There were 4 primary reasons for the use of an “inflammatory soup” in the present study rather than single mediators. First, we (Gold and Traub, 2004; Harriott and Gold, 2009) and more importantly others (Strassman et al., 1996; Oshinsky, 2006; Jakubowski et al., 2007; Levy et al., 2008; Edelmayer et al., 2009), have used a cocktail to sensitize dural afferents, with the studies performed by others on the afferent terminals. Second, while an important and interesting question, the focus of the study was not on which mediators and/or mediator receptors underlie the sensitization of dural afferents, but which ion channels are down stream from these mediators. Third, data from micro-dialysis studies of injured tissue (Hargreaves et al., 1994; Roszkowski et al., 1997; Lepinski et al., 2000) as well as the cerebral spinal fluid from migraine patients during a migraine (Sarchielli et al., 2000) indicates that multiple mediators are released at the same time. Furthermore, evidence suggests that a combination of mediators such as mast cell (Levy et al., 2007) and cyclooxygenase (Jakubowski et al., 2005) products contribute to the sensitization of dural afferents and/or migraine. Fourth, there is already evidence that the IM-induced modulation of at least one ion channel (i.e., NaV1.9) requires a combination of mediators (Maingret et al., 2008). Therefore, we chose to pool mediators rather than attempt to identify the mediator or specific combination of mediators that underlies the sensitization of dural afferents.
Data were analyzed with PulseFit (HEKA), Sigma Plot and Sigma Stat software (Systat Software Inc., Richmond, CA). Conductance-voltage (G–V) curves were constructed from I–V curves by dividing the evoked current by the driving force on the current, such that G = I/(Vm - Vrev) where Vm is the potential at which current was evoked and Vrev is the reversal potential for the current was measured directly (for K+, Na+, and Ca2+ currents). Instantaneous I–V data, obtained from the tail currents measured following activation of voltage-gated Ca2+ currents, was used to construct G–V curves for voltage-gated Ca2+ currents. Activation and steady state availability data were fitted with a Boltzmann equation of the form: G = Gmax/1 + exp[(V0.5 − Vm)/k], where G = observed conductance, Gmax = the calculated maximal conductance, V0.5 = the potential for half activation or inactivation, Vm = command potential and k = the slope factor. K+ currents were corrected for voltage error. For Ca2+ imaging, the IM-induced change in the fluorescence ratio was determined by subtracting the baseline ratio from the peak value. The decay of the IM-induced Ca2+ transient was analyzed as time to 50% decay of the peak (T50).
For comparisons of parametric data collected before and after IM application, either a paired t-test or Repeated Measures ANOVA were used. Otherwise, a Wilcoxin or Friedman test was used for nonparametric analysis. For unpaired comparisons of the percent reduction in rheobase, a t-test was used for parametric data and a Mann Whitney U for nonparametric analysis. Data were considered statistically significant when p<0.05. All data are represented as mean ± standard error.
Data was collected from 186 dural afferents acutely dissociated from 36 female Sprague Dawley rats. Of these, 131 were studied in voltage-clamp and 45 were studied in current clamp and 10 were studied with fura-2 based microfluorimetry. The size distribution of these neurons was similar to that of our previous study (Harriott and Gold, 2009) with a median cell body capacitance of 29.54 pF (with 21.3 and 37.0 as 25th and 75th percentiles).
We recently described significant increases in the action potential (AP) overshoot and rate of rise in dural afferents following application of inflammatory mediators (IM, bradykinin 10 µM, histamine 1 µM, and prostaglandin E2 1 µM) (Harriott and Gold, 2009). Since voltage gated Na+ channels (VGSC) are largely responsible for the upstroke of the action potential, we predicted that these IM-induced changes in the AP waveform reflected an increase in Na+ currents (INa). For steady state availability, INa was elicited with a 15ms test pulse to −10 mV following a series of 500ms pre-pulses from −120 to −5 mV. To isolate tetrodotoxin resistant currents (TTX-R INa) from TTX sensitive currents (TTX-S INa), a 500 ms pre-puse to a potential between −40 and −30 mV was used to inactivate TTX-S INa. The pre-pulse potential used for each neuron was based on results from the steady-state availability data for total Na+ current evoked in each neuron. To validate this approach for separation of TTX-S from TTX-R INa, TTX (100 nM) was used in 5 dural afferents and the currents isolated with both methods were identical (not shown). TTX-S currents were isolated by digitally subtracting TTX-R currents from the total INa. To examine changes in the voltage dependence of activation, INa was elicited with test pulses from −60 to +65 mV following a 100 ms pre-pulse to −110 mV. TTX-R INa was again isolated with a pre-pulse to a potential between −40 and −30 mV and TTX-S currents were isolated by digital subtraction (Fig 1A, B).
As described in other afferent populations, two general types of INa were detected in dural afferents. These included a relatively low threshold, rapidly activating, rapidly inactivating TTX-S INa (Fig 1A) and a relatively high threshold, more slowly activating and inactivating TTX-R INa (Fig 1A). No persistent current was detected, although the voltage-protocols employed were not optimized to detect the presence of a low threshold persistent current. In dural afferents, peak TTX-S INa density was 28.6 ± 10.6 pA/pF while that of TTX-R INa was 65.3 ± 19.3 pA/pF at a test potential of −10 mV (n = 13). TTX-S and TTX-R INa were detectable in every dural afferent studied. Parameters describing the voltage-dependence of activation and inactivation of TTX-S and R INa are summarized in Table 1.
Consistent with our prediction, TTX-R INa was significantly (i.e., >2 SDs from baseline) increased in 12 of 13 dural afferents following IM application (Fig 1C). There was no significant increase in TTX-S INa following IM application (Fig 1D). There were small but significant IM-induced changes in the biophysical properties of TTX-R and TTX-S INa which included a hyperpolarizating shift in the voltage-dependence of activation (Table 1).
There is a growing body of evidence from a number of investigators that persistent inflammation results in a significant decrease in voltage dependent K+ currents (Yoshimura and de Groat, 1999; Stewart et al., 2003; Harriott et al., 2006) and more relevantly, that inflammatory mediators including PGE2 (Nicol et al., 1997) and NGF (Zhang et al., 2002) produce a rapid decrease in K+ current. Therefore, we examined the possibility that decreases in K+ currents (IK) also contribute to IM-induced dural afferent sensitization. K+ currents were elicited with10 mV, 500ms voltage-steps between −60 and +60 mV following a 500ms pre-pulse to −120 mV. The reversal potential for K+ was determined by eliciting tail currents with voltage steps from −110 to −50 mV in 10 mV increments, following a test pulse to +40 mV. Furthermore, since both voltage- and Ca2+-dependent K+ currents contribute to primary afferent excitability, currents were recorded in the absence and presence of the voltage-dependent Ca2+ channel blocker Cd2+ (50 µM). IM had no detectable influence on voltage-dependent K+ currents evoked in the presence of Cd2+ (n = 7), as assessed by the voltage-dependence of channel activation (Fig 2B) and the maximal conductance (Fig 2B). In contrast, when Cd2+ was omitted from the bath solution (n = 7), an IM-induced suppression of K+ current was detected. The result was a significant decrease in maximal conductance for total outward current (Fig 2C). To begin to identify the channel sensitive to the actions of IM, the experiment was repeated in the presence of iberiotoxin (IbTx, 100 nM, n = 3), to selectively block large conductance Ca2+ dependent K+ currents (BK). Pre-application of IbTx failed to occlude the actions of IM, suggesting BK channels do not underlie the IM sensitive current. Taken together, results from this series of experiments suggest IM suppress a Ca2+-dependent K+ current in dural afferents different from that previously described in other sensory neurons.
While an increase in INa and decrease in IK could account for the IM-induced sensitization of dural afferents, given minimal shifts in the voltage dependence of activation of either current and the fact that these channels are not active close to the resting membrane potential; they are unlikely to account for the IM-induced changes in passive electrophysiological properties, in particular, membrane depolarization and decreased input resistance reported recently (Harriott and Gold, 2009). Therefore to identify the ion channel(s) involved in IM-induced changes in passive electrophysiological properties of dural afferents, IM activated currents were recorded at a holding potential of −60 mV. Consistent with the observed IM-induced membrane depolarization, a significant increase in inward current (mean 10.7± 2.9 pA/pF, Fig 3A, Fig 3C) was observed following application of IM to dural afferents (n = 7). To begin to identify the basis for this IM-induced current, current was evoked from a series of holding potentials ranging between −60 and +10 mV. Data from this series of experiments (n = 7) indicated that the IM-induced current had a reversal potential of approximately −30 mV (Fig 3B) which was close to the calculated equilibrium potential for Cl− under our recording conditions (−34 mV). Therefore, we tested the prediction that the IM-induced current was mediated by anion flux. IM-induced currents were measured again after reducing extracellular Cl− concentration from 140mM to 36mM which produced an equimolar concentration of Cl− inside and outside the cell. Consistent with our prediction, the decrease in extracellular Cl− concentration (n = 6) resulted in an IM-induced current with a reversal potential at 0 mV (Fig 3B) and consequently, an increase in current density evoked at −60 mV (Fig 3C). Moreover, bath application of non-selective Cl− channel blocker, niflumic acid (NFA, 10 µM) reversed the IM-induced increase in holding current (Fig 3C, n = 5). Given the nature of this IM-induced current, we subsequently refer to it as IIM-Cl
While a Ca2+ dependent Cl− current has yet to be described in trigeminal ganglion neurons, evidence from nodose ganglion neurons suggests that bradykinin is capable of activating a Cl− current that is Ca2+ dependent (Oh and Weinreich, 2004). Therefore to assess the contribution of Ca2+ to IIM-Cl, currents were recorded using intracellular BAPTA (10mM) to replace EGTA (11mM). Consistent with the suggestion that an increase in [Ca2+]in is important for activation of the IIM-Cl, the increase in Cl− current was significantly attenuated by 10mM BAPTA (Fig 4A, n = 7).
To determine if influx through voltage-gated Ca2+ channels (VGCC) provides the source of Ca2+ that activates the IIM-Cl, currents were measured under conditions in which Na+ and K+ currents were minimized by replacing intracellular K+ with Cs+ and extracellular Na+ with choline, leaving Cl − and Ca2+ unchanged. Currents were elicited with 100 ms test pulses from −70 mV to +50 mV following a 40ms pre-pulse to 0 mV to evoke Ca2+ influx (Fig 4B). Driving an increase in intracellular Ca2+ with this protocol alone was insufficient to activate a Cl− current in dural afferents. However, the current-voltage relationship for IIM-Cl (resolved by digital subtraction (Fig 4C, n=8) of currents evoked before and after application of IM), was clearly altered by this protocol, exhibiting pronounced inward rectification that appeared to reflect an increase in outward current between −30 and 0 mV. Furthermore, IIM-Cl decreased at voltage steps greater than 0 mV (Fig 4C). These data were consistent with the possibility that Ca2+ influx via voltage-gated Ca2+ currents facilitates IIM-Cl, particularly at potentials between −30 and 0 mV.
To confirm the shape of the I–V curve in Fig 4C reflected the influence of Ca2+ influx through VGCC, IIM-Cl was examined again under recording conditions in which intracellular Ca2+ was artificially elevated to ~620 nM (with the reduction of EGTA to 1.2 mM) and Ca2+ influx via VGCC was blocked by the addition 50µM Cd2+ to the bath solution. Despite the elevated intracellular Ca2+ concentration in dural afferents studied under these recording conditions (n = 6), the NFA sensitive current at −60 mV, was comparable to that observed under our standard whole cell recording conditions when intracellular Ca2+ was buffered to ~45 nM (i.e., < 1 pA/pF). Furthermore, the IM-induced inward current was comparable to that observed when Ca2+ influx was facilitated with a pre-pulse to 0 mV (Fig 4C) and in the absence of a pre-pulse to facilitate Ca2+ influx (Fig. 3C). However, under conditions in which Ca2+ influx was block by Cd2+, the I–V relationship for IIM-Cl was linear (i.e., no increase in outward current between −30 and 0 mV, and no decrease in outward current at potentials > 0 mV). These observations further suggest that Ca2+ influx via VGCC may facilitate IIM-Cl efflux but is neither necessary nor sufficient for the activation of IIM-Cl.
To further substantiate this suggestion and rule out the possibility that Ca2+ influx via VGCC contributes to the activation of Cl− current secondary to an IM-induced shift in the voltage dependence of activation of VGCC, we recorded Ca2+ currents directly (n = 5), before and after application of IM. High threshold voltage-gated Ca2+ currents were the only currents detected in dural afferents. While these currents did exhibit run-down, it was monitored until currents were stabilized (~10 minutes) prior to the application of IM. No evidence of a low threshold current was detected following application of IM (Fig 5B). In contrast, however, application of IM produced a significant (p < 0.05) decrease in peak Ca2+ current (Fig 5A, B). The voltage dependence of Ca2+ current activation was examined with the instantaneous current-voltage relationship derived from tail currents. IM application did not shift the voltage dependence of activation of Ca2+ currents (Fig 5C), suggesting a voltage-independent mode of current suppression as has been described by others (Dolphin and Scott, 1987).
There is evidence that bradykinin (Vellani et al., 2001) and PGE2 (Pitchford and Levine, 1991; Schnizler et al., 2008) can sensitize TRPV1, providing a source of Ca2+ for the activation of IIM-Cl. And while an IM-induced activation of such a current should have impacted reversal potential measurements of IIM-Cl, we sought to rule out Ca2+ influx via a TRP channel as a source of Ca2+ contributing to the activation of IIM-Cl at resting membrane potential. Towards that end, neurons were studied in the presence of the non-selective Ca2+ channel blocker ruthenium red (10 µM, n = 9). While we confirmed the ability of ruthenium red to block capsaicin evoked currents (n = 4, data not shown), ruthenium red failed to block the IM-induced membrane depolarization, increase in holding current and decrease in input resistance (e.g., IM-induced decrease in input resistance was 74 ±13 and 52 ± 22% of baseline in the absence and presence of ruthenium red, respectively; p > 0.05). These results suggested an IM-induced release of Ca2+ from intracellular stores as a source of Ca2+ for the activation of IIM-Cl.
To assess this possibility, we examined the impact of IM on [Ca2+]in in dural afferents measured directly with fura-2. Application of IM to dural afferents resulted in a dramatic increase in [Ca2+]in in 9 of 10 neurons tested (Fig 6) while an IM-induced increase in [Ca2+]in was seen in only 41 of 63 non-labeled neurons. Interestingly, the magnitude of the evoked increase was significantly larger (Fig 6D, E) and the decay of the transient (in response to a 30 second application of IM) significantly slower (Fig 6D, F) in dural afferents than in unlabeled afferents. To rule out the possibility that the augmented Ca2+ transients in dural afferents were due to DiI labeling, IM were applied to 2 additional populations of DiI labeled afferents: cutaneous afferents retrogradely labeled from the glabrous skin of the hindpaw and muscle afferents retrogradely labeled from the temporalis muscle. While both populations of afferents can be sensitized by inflammatory mediators (Gold and Traub, 2004; Harriott and Gold, 2009), IM-induced increases in intracellular Ca2+ were only observed in 1/10 muscle afferents and in 0/12 capsaicin sensitive cutaneous afferents.
To begin to determine whether a combination of inflammatory mediators was actually needed for the sensitization of dural afferents, data was collected from 15 neurons in when single mediators were applied sequentially with a 5 minute inter-application interval. The order of application was histamine, PGE2, then bradykinin. As with the combination of mediators, an increase of 20% above baseline was considered a response. Of these, 6 neurons responded to histamine, 9 responded to PGE2 and 5 responded to bradykinin. Of these only 3 neurons “responsive” to histamine responded to at least one other mediator while the same was true for 6 neurons responsive to PGE2 and all 5 neurons responsive to bradykinin. While the fraction of neurons responsive to single mediator was not significantly different than the fraction responsive to the combination, the magnitude of the response was smaller (p < 0.05) and the decay faster (p < 0.05) than was observed when mediators were applied in combination. The increase in fluorescence above baseline was 0.26 ± 0.03, 0.61 ± 0.18 and 0.94 ± 0.09 for histamine, PGE2 and bradykinin, respectively. While T50 for decay of the evoked transient was 25.6 ± 6.1 seconds, 65.9 ± 14.2 seconds and 45.2 ± 10.5 seconds for histamine, PGE2 and bradykinin. These results suggest that the IM-induced increase in [Ca2+]in observed in dural afferents reflects the actions of the combination of mediators rather than the actions of any individual mediator.
In the context of our previous current clamp data (Harriott and Gold, 2009), the voltage clamp data presented here suggest that activation of IIM-Cl is a depolarizing current that increases the excitability of dural afferents. However, these previous data were recorded using whole cell patch configuration with ECl (−34 mV) determined by the composition of the bath and electrode solutions. As a result, the “excitatory” effects observed may have been an artifact of our recording conditions. In light of this possibility, it was important to determine if ECl in dural afferents was sufficiently depolarized to account for an IM-induced increase in excitability. IIM-Cl was recorded in response to a ramp protocol from +50 mV to −100 mV using gramicidin perforated patch to prevent dialysis of intracellular Cl− (Akaike, 1996). To isolate the current, extracellular TEA (30 mM) was used to block K+ currents and choline was used to substitute the remaining (100 mM) extracellular Na+. Under these recording conditions, IIM-Cl was similar to those recorded under traditional whole cell patch configuration in Fig 4C (Fig 7, n=6). More importantly, results from these experiments indicated that the IIM-Cl reversal potential in dural afferents is −27.73 ± 2.6 mV (Fig 7B inset) which is close to the action potential threshold previously recorded in dural afferents (Harriott and Gold, 2009). These observations suggest that activation of IIM-Cl is excitatory in dural afferents.
However, ECl is generally depolarized relative to the resting membrane potential in sensory neurons, resulting in the phenomenon of primary afferent depolarization first described almost 50 years ago (Eccles et al., 1962). Despite this fact, activation of Cl− channels via GABAA receptors, is generally thought to be inhibitory as a result of membrane shunting and depolarization-induced inactivation of voltage gated Na+ channels (Price et al., 2009). Therefore, we assessed the impact of IM on the excitability of dural afferents with gramicidin patch recording (Fig 8A). Three sets of experiments were performed. In the first, the relative impact of IM-induced activation of IIM-Cl on the sensitization of dural afferents (n = 7) was assessed with the application of NFA (100µM). Blocking Cl− channels with NFA alone reduced baseline excitability as indicated by an increase in rheobase (Fig 8B) and a decrease in the slope of the stimulus response function (Fig 8D); although NFA did not impact the AP threshold (Fig 8C). These changes were associated with an NFA induced hyperpolarization of the resting membrane potential. The effects of NFA on rheobase and the stimulus response function (SRF, i.e., response to suprathreshold stimulation) returned to baseline following a 2 minute washout (Fig 8B, D &E). Consistent with previous data (Harriott and Gold, 2009), application of IM produced a significant reduction in rheobase (Fig 8B), hyperpolarization of AP threshold (Fig 8C) and increase in the slope of the SRF (Fig 8D). When NFA was applied in the presence of IM, NFA reversed the IM-induced decrease in rheobase and the IM-induced increase in the SRF slope (Fig 8B, D). NFA also reversed the IM-induced membrane depolarization (Fig 8E). Interestingly, however, NFA did not reverse the hyperpolarization of AP threshold (Fig 8C), suggesting that while activation of IIM-Cl is responsible for the most dramatic changes in excitability, it is not responsible for all IM-induced changes.
In the second set of experiments we tested the prediction that if membrane depolarization associated with a depolarized ECl underlies IM-induced dural afferent sensitization, hyperpolarization of ECl should block the sensitizing effects of IM on dural afferents. To test this prediction, ECl was artificially hyperpolarized by decreasing the concentration of Cl− in the electrode solution to 10 mM. Under these recording conditions, the predicted ECl is −68 mV (n = 15). In stark contrast to the 19 of 19 dural afferents sensitized by IM (Harriott and Gold, 2009) with an ECl of −34 mV and the 7 of 7 neurons sensitized by IM with gramicidin patch recording, an ECl of −68 mV produced an IM-induced decrease in excitability in 5 of 15 dural afferents as indicated by an increase in rheobase, and produced no detectable change in excitability in an additional 6 of the 15 dural afferents tested. Moreover, an increase in excitability as indicated by a decrease in rheobase was detected in only 4 of 15 dural afferents tested (Fig 9A). The net effect of IM in the presence of low intracellular Cl− was a 15.35 ± 26.9% increase in rheobase (Fig 9B) and no changes in the response to suprathreshold stimuli (Fig 9D). Nevertheless, as with NFA, the IM-induced hyperpolarization of AP threshold was still present in 14 of 15 neurons tested with a hyperpolarized ECl (Fig 9C).
In the third set of experiments, we sought to confirm the link between the IM-induced Ca2+ transient and the activation of an excitatory Cl− current. IM-induced sensitization of dural afferents was assessed with an electrode solution in which we again substituted 11mM EGTA with 10mM BAPTA (n=8). In the presence of 10mM BAPTA in the electrode solution, a significant decrease in rheobase was only detected 4 of 8 neurons tested. Furthermore, this decrease in rheobase was significantly less than the percent decrease in rheobase observed in the presence of 11mM EGTA (Fig 10A, B). The IM-induced increase in the slope of the SRF was also significantly attenuated in the presence of BAPTA (Fig 10D). In contrast to the results obtained with NFA and the hyperpolarizing shift in ECl, 10mM BAPTA also prevented the IM-induced hyperpolarization of AP threshold (Fig 10H). These results are consistent with the suggestion that a rapid increase in intracellular Ca2+ is necessary for the activation of IIM-Cl and subsequent sensitization of dural afferents and suggest that the Ca2+ transient also contributes to the modulation of voltage-gated Na+ currents critical for the establishment of AP threshold.
The purpose of this study was to identify ionic mechanisms underlying IM-induced sensitization of dural afferents. While at least one experimental outcome, modulation of TTX-R INa, was expected, our results contain several novel, and potentially important observations: 1) the IM-induced block of a Ca2+ dependent K+ current, 2) the IM-induced inhibition of voltage-gated Ca2+ currents, 3) the IM-induced activation of IIM-Cl, and 4) the critical role IIM-Cl plays in the sensitization of dural afferents.
The IM-induced increase TTX-R INa in dural afferents observed in the present study is consistent with a growing body of literature indicating that NaV1.8, the channel underlying the slowly inactivating TTX-R current in nociceptive afferents, is not only a common target for a wide variety of inflammatory mediators, but is also the dominant channel in nociceptive afferents innervating structures throughout the body (Gold and Caterina, 2008). Results from in vivo studies of dural afferents suggests this channel is also present in the peripheral terminals of nociceptive dural afferents (Strassman and Raymond, 1999) where it is positioned to be regulated by inflammatory mediators released in this structure. The absence of detectable IM-induced changes in TTX-S INa, suggests that the IM-induced increase in TTX-R INa is responsible for increases in AP overshoot and rate of rise (Harriott and Gold, 2009). This increase in Na+ current is also the most likely mechanism underlying the IM-induced decrease in AP threshold. A unique role for TTX-R INa in the regulation of AP threshold would account for the observation that the IM-induced hyperpolarization of AP threshold was still observed in the presence of NFA and a hyperpolarized ECl. While there is evidence that PGE2-induced modulation of TTX-R INa involves the activation of protein kinase A (PKA) (England et al., 1996; Gold et al., 1998), there is also evidence that PKA activation may be upsteam of PKC in dissociated neurons as we were able to block PKA mediated modulation of TTX-R INa with PKC inhibitors (Gold et al., 1998). The results of the present study are consistent with these previous results if the inhibitor effect of BAPTA reflects the inhibition of IM-induced PKC activation. That is, the inhibition of IM-induced activation of PKC by BAPTA would account for the observation that BAPTA attenuated the IM-induced decrease in AP threshold. However, we suggest that it is the biophysical properties of this channel in combination with the fact that it carries the vast majority of Na+ current underlying spike initiation in dural afferents that enables activation of IIM-Cl to have such a profound influence on afferent excitability.
IM-induced suppression of two K+ currents has been described in sensory neurons. One was a Ca2+ dependent K+ current underlying a slow after hyperpolarization in vagal afferents that is blocked by bradykinin and PGE2 (Cordoba-Rodriguez et al., 1999) that was subsequently described in a small subpopulation of DRG neurons (Gold et al., 1996a). These K+ currents do not appear to be present in dural afferents. The second, a K+ current suppressed by inflammatory mediators including PGE2, is a α-dendrotoxin-sensitive current, likely mediated by a K+ channel with Kv1.1 properties (Chi and Nicol, 2007). This current has a relatively low threshold for activation and contributes to a decrease in current threshold, action potential threshold and the increase in burst duration in capsaicin sensitive small diameter DRG neurons. Despite the inclusion of PGE2 in the IM used in the present study, no such current was suppressed in dural afferents. Instead, IM suppressed an iberiotoxin insensitive Ca2+-dependent K+ current. Given the biophysical properties of the IM-sensitive K+ current in dural afferents, it is unlikely to contribute to the changes in excitability measured here, but may contribute more readily after prolonged stimuli where it may begin to play a more dominant role as other K+ currents begin to inactivate.
The efficacy of BAPTA to block both the activation of IIM-Cl and the IM-induced sensitization of dural afferents suggests that activation of IIM-Cl is dependent on a transient increase in [Ca2+]in. Whether it is the rapid IM-induced increase in [Ca2+]in that is essential for the activation of IIM-Cl, or that activation of this current requires the rise in Ca2+ coincident to additional second messenger pathway(s) activated by IM, remains to be determined. Nevertheless, what is clear is that an increase in [Ca2+]in alone, is insufficient for the activation of IIM-Cl. BAPTA will only influence the kinetics of the Ca2+ transient without significantly attenuating the magnitude or duration of the IM-induced increase in [Ca2+]in. Furthermore, neither the increase in [Ca2+]in alone via the activation of voltage-gated Ca2+ currents or artificially buffering [Ca2+]in to over 600 nM was sufficient to activate a Cl− current in dural afferents. Our data also suggest that once activated, IIM-Cl is not influenced by further increases in [Ca2+]in. That is, given IM-induced suppression of voltage-gated Ca2+ currents, a decrease in an inward current (i.e., Ca2+) at potentials >−30 mV (see below) is likely to account for odd shape of the IIM-Cl I–V curve (i.e., Fig 4C), rather than an influence of Ca2+ that is selective for Cl− infflux.
Recent data suggest that following nerve injury, expression of the Ca2+-dependent Cl− channel, bestrophin-1, expression upregulated in dorsal root ganglion (DRG) neurons (Boudes et al., 2009), which may be responsible for the nerve injury-induced increase in Ca2+ dependent Cl− current in DRG (Andre et al., 2003; Boudes et al., 2009) and nodose ganglion (Lancaster et al., 2002) neurons. Such a channel may also account for the Cl− current activated by bradykinin in vagal afferents (Oh and Weinreich, 2004; Lee et al., 2005). However, this is the first description of a Cl− current contributing to the IM-induced sensitization of somatic afferents. Furthermore, the IM-activated Cl− current in dural afferents appears to be distinct from Cl− channels up-regulated following nerve injury, and bestrophin-1, given that these channels are readily activated by any increase in [Ca2+]in, such as that associated with action potential generation (Boudes et al., 2009), while IIM-Cl, once activated, appears to be largely Ca2+ independent.
Suppression of voltage-gated Ca2+ currents in sensory neurons is generally believed to be one of the primary mechanisms underlying the actions of several spinally administered analgesics (Gold and Caterina, 2008). Nevertheless, PGE2-induced suppression of Ca2+ currents in mouse trigeminal ganglion neurons has been described (Borgland et al., 2002). It is, therefore possible that the IM-induced suppression of Ca2+ current in dural afferents reflects a similar mechanism of action. However, in contrast to the results obtained with PGE2 in mouse trigeminal ganglion neurons, suppression of Ca2+ currents in dural afferents did not appear to reflect a membrane delimited binding of G-protein subunits that would be expected to shift the voltage-dependence of channels activation. Alternatively, the large sustained IM-induced increases in [Ca2+]in, raise the possibility that inhibition of Ca2+ currents in dural afferents is mediated by a Ca2+-induced inactivation of Ca2+ channels (Catterall, 2000).
IM-induced suppression of Ca2+ currents in dural afferents may serve as a form of a “brake” to the excitatory IM-induced processes, for example by attenuating the peripheral release of transmitters from nociceptive afferents, and consequently the neurogenic inflammation mediated by vasoactive neuropeptides in dural afferents (McIlvried et al., 2009). However, a decrease in voltage-gated Ca2+ currents may also contribute, indirectly to dural afferent sensitization, as the IM-induced decrease in Ca2+ current is likely to be responsible for the IM-induced decrease in high threshold Ca2+-dependent K+ current.
The most dramatic IM-induced changes in dural afferent excitability are the decrease in rheobase and the shift in the stimulus response function. That IM-induced activation of IIM-Cl plays a dominant role in both of these changes in suggested by the observation that both changes were reversed by NFA, attenuated by BAPTA, and switched from being excitatory to inhibitory when ECl was hyperpolarized. These observations suggest that the mix of ion channels underlying the excitability of dural afferents is unique from those regulating the excitability of other populations of afferents where the membrane depolarization associated with the activation of Cl− channels is inhibitory (Price et al., 2009). That IIM-Cl does not appear to contribute to the sensitization of other populations of afferents we have studied (Gold and Traub, 2004; Harriott and Gold, 2009) does not rule out the possibility that the channel contributes to the sensitization of other afferent populations, but it does raise the intriguing possibility that it may be unique to dural afferents.
While the results of the present study have added several important aspects to our understanding of dural afferents, several caveats must be kept in mind. First, while several lines of evidence suggest that the impact of dural afferent labeling on the properties of dural afferents should be minimal (mast-cell degranulation normalized by 10 days post labeling ((Harriott and Gold, 2009); the IM-induced changes in excitability are comparable to those observed in unlabeled sensory neurons (Gold et al., 1996b); and the IM-induced sensitization of other populations of trigeminal ganglion neurons (Harriott and Gold, 2009), even those such as pulpal afferents that require significant tissue destruction to enable retrograde labeling (unpublished observation), still appears to involve processes distinct from those observed in dural afferents), it is possible that the process of cell labeling has changed the properties of dural afferents. Second, despite evidence from other afferent populations suggesting that mechanisms underlying the actions of inflammatory mediators, as revealed through the study of the afferent cell body in vitro (Gold et al., 1996c) contribute to the sensitization of afferent terminals in vivo (Khasar et al., 1998), the afferent cell body in vitro, is only a model of the afferent terminal. With a number of potentially important limitations (i.e., injury, differences in anatomical constraints, etc) associated with this model, additional experiments will be necessary to confirm the contribution of processes identified in the present study to the sensitization of dural afferent terminals in vivo.
Although there is evidence that central mechanisms are involved in triggering migraine, dural afferents appear to be critical for the initiation of migraine pain (Moskowitz et al., 1993; Strassman et al., 1996; Burstein, 2001; Bolay and Moskowitz, 2002). Activating the peripheral terminals of dural afferents produces pain that is identical to a migraine (Ray and Wolff, 1940). Furthermore, there is evidence that neurogenic inflammation of the dura can activate and sensitize this population of afferents during a migraine attack (Strassman et al., 1996; Sarchielli et al., 2000). The data presented here suggest ionic mechanisms that may serve as targets for the development of novel anti-migraine therapies. The fact that NaV1.8 is present in all dural afferents where it appears to play the dominant role in spike initiation (Strassman and Raymond, 1999) supports the notion that a selective blocker would not only help other types of pain but may be effective for migraine pain as well. The critical role of IIM-Cl in mediating IM-induced sensitization of dural afferents suggest that the ion channel underlying this current and/or the mechanisms underlying the maintenance of the Cl− gradients in dural afferent may be a particularly useful target for the treatment of migraine.
We would like to thank Drs. Daniel Weinreich and Brian Davis for helpful comments during the preparation of the manuscript. Work was support by NIH grants: NS059153 (AHV) NS41384 (MSG) and DE018252 (MSG).