Lidocaine Blocks the Hyperpolarization-activated Mixed Cation Current, I  ^h, in Rat Thalamocortical Neurons

• The mechanisms underlying the supraspinal central nervous system effects of lidocaine are poorly understood and not solely explained by Na⁺channel blockade, but may involve the hyperpolarization-activated cation current, I  ^h, which is blocked by lidocaine in peripheral neurons

• Lidocaine concentration-dependently inhibited I  ^hin rat thalamocortical neurons in vitro , with high efficacy and a potency similar to Na⁺channel blockade

• A resultant reduction of intrinsic burst firing and δ rhythms may contribute to the alterations in oscillatory cerebral activity produced by systemic lidocaine in vivo 

LIDOCAINE is a widely exploited local anesthetic exerting its main peripheral therapeutic effects by blocking voltage-gated Na⁺channels. It also is useful systemically in the management of acute postoperative and chronic neuropathic pain syndromes, in the maintenance of general anesthesia, and as a class IB antiarrhythmic.1,–,5In addition, systemic lidocaine exhibits concentration-dependent central nervous system (CNS) toxicity that begins with alterations in sensorium at low plasma concentrations that overlap with those associated with the therapeutic effects (in humans, typically less than 5 μg/ml or approximately 20 μM) and progresses to generalized seizures, coma, and death at higher levels (approximately >15–50 μg/ml or 60–200 μM).6,7 

Though poorly understood, the mechanisms that underlie lidocaine's complex concentration-dependent supraspinal CNS effects are not solely explained by its classic action on Na⁺channels.8,–,10Among the list of other possible targets is the hyperpolarization-activated mixed Na⁺/K⁺current, I  ^h, which is blocked by lidocaine in peripheral sensory neurons.11,I  ^h, predominantly its underlying channel isoform, HCN2,12is highly expressed in the thalamus,13,–,15a brain area that plays an important role in the generation of the different physiologic conscious states and associated cerebral rhythms; in drug-induced sedation, anesthesia, and analgesia; and in epileptogenesis.16,–,19In mammals, lidocaine at subconvulsive doses has long been known to produce slow-wave electroencephalographic rhythmic activity and “spindling” associated with sedation and reduced responsiveness to noxious stimuli,20,–,23implicating the thalamus as a site of action. More recently, in vitro  24,25and human in vivo  26reports have focused on lidocaine's actions in the ventrobasal thalamus, the main supraspinal relay station for somatosensory and nociceptive signals.27However, lidocaine's effects on I  ^hin ventrobasal thalamocortical neurons are unknown.

I  ^h, whose activation produces a depolarizing noninactivating inward current,12is crucial for controlling excitability in thalamocortical neurons in multiple ways. First, it contributes to the setting of the resting membrane potential (RMP), as a significant fraction of I  ^hchannels is active near rest.12,28,–,30Second, because of its leak and negative-feedback properties, I  ^hoperates as a “voltage-clamp,” passively shunting incoming impulses and actively opposing hyperpolarization and depolarization.12,31,32As a result, I  ^his critical for determining the distinct voltage-dependent firing mode of these neurons. At depolarized potentials positive to approximately −60 mV, they exhibit a “relay” or “tonic” mode that is associated with vigilance and wakefulness in vivo  and characterized by tonic repetitive firing of singleton Na⁺-dependent action potentials.16,33At hyperpolarized potentials, neurons switch to the “oscillatory” or “burst” mode that occurs in states of slow-wave electroencephalographic activity (e.g. , nonrapid eye movement sleep) and features action potential bursts mediated by the low-threshold Ca²⁺current, I  ^T.12,16,33Third, through interaction with I  ^Tto generate rhythmic burst firing, I  ^hserves as a pacemaker current and is central to the generation of slow intrinsic neuronal28and network oscillations in the thalamocortical system during nonrapid eye movement sleep and drowsiness.34,–,36 

Here, we tested the hypothesis that lidocaine blocks I  ^hin rat ventrobasal thalamocortical neurons in vitro  and explored the functional consequences of I  ^hblockade by lidocaine in these neurons.

Ethics approval for all animal experiments was obtained from the Committee on Animal Care of The University of British Columbia (Vancouver, British Columbia, Canada). Wistar rats of postnatal age P13–P16 were deeply anesthetized with isoflurane (Abbott Laboratories, Montreal, Canada) and decapitated. The cerebrum was rapidly removed and placed in oxygenated (5% CO²/95% O²), cold (1–4°C), artificial cerebrospinal fluid (ACSF) of the following composition (mM): NaCl, 124; KCl, 2.5; NaH²PO⁴, 1.25; CaCl², 2; MgCl², 2; NaHCO³, 26, dextrose, 10 (pH, 7.3–7.4; 290 mOsm). After trimming the chilled brain, a block containing the ventrobasal thalamus was glued onto a tissue-slicer stage with cyanoacrylate adhesive. Coronal slices of the thalamus were cut at 250–300 μm on a Leica VT1200S vibratome (Leica Biosystems, Nussloch, Germany) while the block was submerged in oxygenated, 1–4°C ACSF. Immediately after cutting, the slices were incubated at room temperature (22–24°C) in oxygenated ACSF.

For recording, slices were submerged in a Perspex chamber with a volume of 1.5 ml, fixed between two pieces of polypropylene mesh, and maintained at room temperature. The slices were continuously perfused by gravity with oxygenated ACSF at a flow rate of 2.5 ml/min controlled by a FR-50 flow valve (Harvard Apparatus, St. Laurent, QC, Canada). Individual neurons were visualized with the aid of differential interference contrast infrared videomicroscopy (Zeiss Axioskop FS, Carl Zeiss, Göttingen, Germany). The images were recorded with a Hamamatsu C2400 video camera system (Hamamatsu Photonics K.K., Hamamatsu, Japan). Patch pipettes were pulled from borosilicate glass (World Presicion Instruments, Inc., Sarasota, FL) using a PP-83 two-stage electrode puller (Narishige Scientific Instrument Laboratory, Tokyo, Japan) and filled with a solution containing (mM): K-gluconate, 139; EGTA, 10; KCl, 6; NaCl, 4; MgCl², 3; HEPES, 10; CaCl², 0.5; adenosine-5′-triphosphate (disodium salt), 3; guanosine-5′-triphosphate (sodium salt), 0.3, titrated to pH 7.3–7.4 with 10% gluconate. Typical electrode resistances were 5–6 MΩ and access resistance ranged from 10 to 20 MΩ. Whole cell patch-clamp recordings from ventrobasal thalamic neurons were performed in both current- and voltage-clamp modes with a HEKA EPC-7 amplifier (HEKA Elektronik Dr. Schulze GmbH, Lambrecht, Germany) via  a Digidata 1322A 16 bit data acquisition system (Axon Instruments, Inc., Foster City, CA) using pCLAMP software (Axon Instruments, Inc.). The membrane currents were low-pass filtered (three-pole Bessel filter) at a frequency of 3 kHz and digitized at 10 kHz. Data were collected more than 10 min after whole cell access to allow the internal pipette solution to equilibrate with the neuron. Membrane potentials were corrected off-line for a liquid junction potential of −8 mV.37No leak subtraction was performed.

Data were analyzed using ORIGIN 7 (OriginLab Corporation, Northampton, MA) and Prism 5 software (GraphPad, La Jolla, CA). To determine the IC⁵⁰and Hill coefficient, concentration-response curves were normalized and fitted using the Hill equation as follows:

where I  represents the current measured in the presence of a given drug concentration; I  ^maxis the control current measured in the absence of the drug; [C] is the drug concentration; IC⁵⁰is the half-maximal inhibitory concentration; and h  is the Hill coefficient.

The conductance-voltage relationship for I  ^hsteady-state activation was fitted by the Boltzmann equation:

where G  ^his the I  ^hconductance (calculated as G  ^h= I /[V −V  ^r]: I , amplitude of the I  ^htail current following a hyperpolarizing step [V ]; V  ^r, estimated I  ^hreversal potential); G  ^h(max)is the maximum conductance obtained after the most hyperpolarizing step; V  ^0.5is the half-maximal activation potential; and k  is the slope factor.

We estimated the reversal potential of I  ^hfrom the intersection of extrapolated linear regression fits of instantaneous voltage-current relationships at two different holding membrane potentials.38The slopes of the plots were assumed to vary depending on the degree of activation of I  ^hand to intersect at the reversal potential of I  ^hwhere there is no driving force.

Data are presented as mean ± SEM unless mentioned otherwise; baseline membrane properties of all included neurons are given as mean ± SD as indicated. We used one-way ANOVA to test for concentration-dependent drug effects and comparisons of more than two groups. Comparisons between two groups were conducted with the use of a paired Student t  test; a one-sample Student t  test was used to test for differences of normalized data from baseline (i.e. , a hypothetic mean of 1.0). Statistical tests were two-tailed and results were considered significant at α = 0.05.

Lidocaine HCl, tetrodotoxin, and CsCl were purchased from Sigma–Aldrich Canada Ltd. (Mississauga, ON, Canada). ZD7288 was obtained from Ascent Scientific (Princeton, NJ). BaCl²was obtained from ICN Biomedicals (Aurora, OH). Lidocaine, tetrodotoxin, and ZD7288 were dissolved in fresh ACSF to prepare concentrated stock solutions stored at 4°C. Before application, required aliquots of the stock solutions were dissolved in ACSF to obtain the respective concentrations. All drugs were applied to the bath by switching from the control perfusate to ACSF containing a desired drug concentration. Recordings were conducted after 6 min of perfusion (approximately 2 ml/min) of the slices with a test solution except ZD7288 (20 min). All results reported reflect steady state responses.

We investigated n = 62 thalamocortical neurons of the ventrobasal complex (ventral posterior lateral/medial nuclei). The neurons had an average (± SD) RMP of −67.2 ± 3.1 mV, consistent with the results of previous studies.24,39,40When voltage-clamped at −68 mV, the neurons had an average (± SD) input resistance (R  ⁱ) of 271 ± 84 MΩ, determined from the responses to a 5 mV hyperpolarizing voltage step. Their average (± SD) membrane capacitance (C  ⁱ= τ^m/R  ⁱ) was 197 ± 50 pF. All neurons voltage-dependently exhibited both the relay and oscillatory modes of operation characteristic for thalamocortical relay neurons (see Introduction, third paragraph). 33Accordingly, they responded with tonic repetitive firing to depolarizing current pulses from membrane potentials positive to approximately −60 mV, and, when depolarized from hyperpolarized membrane potentials less than approximately −70 mV, responded with burst firing, generated by a low-threshold spike (LTS; known to be mediated by I  ^T; see Introduction, third paragraph) crowned by a burst of action potentials.

Hyperpolarization of neurons voltage-clamped at −68 mV induced an inwardly rectifying, noninactivating current consisting of an instantaneous component and a slow-activating component (fig. 1A), known to be generated by the inwardly rectifying K⁺current, I  ^Kir, and the hyperpolarization-activated mixed cation current, I  ^h, respectively.41,42Extracellular application of 600 μM lidocaine inhibited I  ^h(calculated as the difference between the instantaneous current [I  ^inst] and the steady-state current [I  ^ss] at the beginning and end of the voltage step, respectively) without affecting I  ^Kir(n = 4; fig. 1, A and B). Lidocaine's effects were mirrored by the specific I  ^hantagonist, ZD7288 (50 μM),43,44which similarly blocked only the I  ^hcomponent as predicted (n = 4; fig. 1, C and D), whereas Cs⁺(CsCl, 2 mM), a nonspecific I  ^hblocker,28inhibited both I  ^hand I  ^Kir(n = 4; fig. 1, E and F). Conversely, application of the I  ^Kirblocker, BaCl²(0.1 mM)28,45almost completely abolished this current and effectively unmasked I  ^h(fig. 1, G and H).46In neurons recorded in the presence of extracellular Ba²⁺at −128 mV, the average magnitude of I  ^hwas 233 ± 97 pA (n = 16). The estimated I  ^hreversal potential (V  ^r) was −43.4 ± 2.4 mV (n = 5; fig. 2). Lidocaine reversibly blocked I  ^hin a concentration-dependent manner, with an IC⁵⁰of 72 ± 7 μM (n = 4; ANOVA, P < 0.001) and an estimated Hill coefficient of 1.19 ± 0.12 (fig. 3, A and B). The I  ^hblock was not voltage-dependent at 100 μM (n = 4; ANOVA, P = 0.57; fig. 3C).

Because intrinsic and network thalamocortical oscillations are critically dependent on the activation and deactivation properties of I  ^h,47we further investigated whether lidocaine would alter these. To examine the conductance-voltage relationship for I  ^hsteady-state activation, we measured I  ^htail currents on repolarization to −78 mV (shown in fig. 3A; see Materials and Methods, Data Analysis sections). As illustrated in figure 4A, the activation curve rose between −58 and −128 mV, with a half-maximal activation potential (V  ^0.5) of −94.6 ± 1.9 mV and a slope factor of 11.4 ± 1.0 (n = 4), yielding an estimated maximum G  ^hin the range of 3–12 nS. Lidocaine (100 μM) did not significantly shift the V  ^0.5(−90.5 ± 3.1 mV; n = 4; P = 0.11), but decreased the slope factor to 7.6 ± 1.1 (P = 0.02; fig. 4A). These minor effects possibly reflected an improved quality of the voltage-clamp due to an increase in neuronal R  ⁱ.

We examined the rate of I  ^hactivation by stepping neurons to potentials from −98 to −128 mV (fig. 3A). The resulting kinetics were examined by fitting the activation phase of the current with a double-exponential function. The fast time constant decreased from 1,553 ± 510 ms at −98 mV to 274 ± 33 ms at −128 mV (n = 4; ANOVA, P = 0.02). Lidocaine (100 μM) had no significant effect on the fast time constant of I  ^hactivation in the voltage range from −98 to −128 mV (n = 4; each voltage tested, P > 0.05; fig. 4B). Because the I  ^hwas still increasing slightly even at the end of a 3-s activation pulse, the slow time constant, especially at less negative potentials (fig. 3A), could not be estimated accurately; longer activation pulse durations compromised the whole cell patch.

We determined the rate of I  ^hdeactivation by examining its tail current relaxation kinetics upon repolarization to −78 mV after a 3-s hyperpolarizing pulse to −128 mV. Repolarization to more depolarized potentials resulted in contamination of I  ^htail currents with T type, low-threshold Ca²⁺conductances (see Results section, first paragraph). Fitting the kinetics of I  ^hdepolarization with a single-exponential function produced a deactivation time constant of 991 ± 17 ms (n = 4). Lidocaine (100 μM) substantially delayed I  ^hdeactivation (fig. 4C), such that the deactivation time constant could not be estimated correctly in three of four neurons.

To investigate the implications of lidocaine's actions on I  ^hfor thalamocortical neurons' membrane electrical properties, we performed a series of current-clamp experiments in neurons pretreated with the tonic Na⁺channel blocker, tetrodotoxin. At 600 nM, tetrodotoxin did not significantly alter passive membrane properties of neurons: the baseline R  ⁱ, RMP, and C  ⁱwere 303 ± 21 MΩ, −65.9 ± 0.8 mV, and 206 ± 15 pF, respectively, compared with 277 ± 26 MΩ (P = 0.22), −66.4 ± 1.1 mV (P = 0.45), and 201 ± 26 pF (P = 0.77; for all variables, n = 12) in the presence of tetrodotoxin. Application of tetrodotoxin also did not greatly change current-voltage relationships at potentials negative to −60 mV, but reduced the apparent R  ⁱat depolarized potentials (not shown).

Application of lidocaine at concentrations blocking I  ^hproduced a reversible increase in the R  ⁱof neurons and a hyperpolarization of their RMP (table 1). These effects were concentration-dependent with a peak at 600 μM, but diminished in magnitude at 1 mM (fig. 5, A and B). Of note, application of 1 mM lidocaine initially (within the first 2 min) resulted in hyperpolarization of the RMP followed by its eventual depolarization to a steady-state value. The C  ⁱwas not significantly different from the baseline values over the range of 0.1–1 mM (data not shown), indicating a primary effect of lidocaine on membrane conductance (1/R  ⁱ).24 

To define the effects of I  ^hblockade by lidocaine on the active membrane properties of neurons, we conducted current-clamp experiments at potentials negative to −45 mV, corresponding to the activation range of I  ^h. Neurons current-clamped at −62 to −64 mV exhibited in their voltage responses to hyperpolarizing current pulses a typical inward (“anomalous”)41rectification consisting of instantaneous and time-dependent components (fig. 5C). Bath application of lidocaine (600 μM) inhibited only the time-dependent, I  ^h-mediated inward rectification and produced an increase in the voltage responses to injected current pulses most pronounced at hyperpolarized potentials (n = 4). Current-voltage relationship analyses of the lidocaine-induced changes (fig. 5D) revealed inhibition of a conductance with an average reversal potential of −58.1 ± 0.9 mV (n = 4) and implicated other conductance(s) in addition to I  ^hand voltage-gated Na⁺currents. In addition to the previously mentioned effects, lidocaine also increased the latency of rebound LTSs (arrows, fig. 5C; calculated as the time required by the membrane potential to reach the LTS peak following the termination of the hyperpolarizing current pulse; see Results, first paragraph) from 128 ± 5 ms to 279 ± 47 ms (n = 4, P = 0.04). This increase occurred despite greater hyperpolarization responses, which would result in a larger population of deinactivated T-type Ca²⁺channels.

We compared the effects of lidocaine on the passive and active membrane properties of thalamocortical neurons with those of Cs⁺and ZD7288. Extracellular application of both CsCl (2 mM) and ZD7288 (50 μM) led to an increase in the R  ⁱof neurons, comparable with that produced by 600 μM lidocaine, as well as a significant hyperpolarizing shift in their RMP (table 1). In the current-clamp mode, extracellular Cs⁺reversibly inhibited both components of the inward rectification in the voltage responses of neurons current-clamped at −62 to −66 mV (n = 4; fig. 6) whereas ZD7288 irreversibly abolished only the time-dependent, I  ^h-mediated component (n = 4; fig. 7). Both Cs⁺and ZD7288 increased the voltage response magnitudes at potentials negative to approximately −50 mV. More depolarized holding potentials, from −60 to −63 mV and from −55 to −56 mV, were required to trigger rebound LTSs in three of four and in two of four neurons in the presence of Cs⁺and ZD7288, respectively. However, even in those neurons whose holding potentials were depolarized, Cs⁺and ZD7288 application both increased the LTS latencies (figs. 6and 7). In addition to delaying the activation of rebound LTSs, ZD7288 also decreased the number of action potentials in the LTS-evoked bursts from 3.8 ± 0.6 to 2.0 ± 0.4 (n = 4, P = 0.006). In contrast, Cs⁺had no effect on LTS burst firing.

We also examined the effects of lidocaine at concentrations blocking I  ^hon firing properties of neurons not pretreated with tetrodotoxin. For this purpose, we used a current-clamp protocol to generate both tonic and rebound burst firing,39applied in 2-min intervals to neurons constantly injected with a depolarizing current required to shift their membrane potential from RMP to −58 mV (associated with the tonic mode of firing that occurs in states of vigilance and wakefulness in vivo ; see Introduction, third paragraph). As expected from its action on voltage-gated Na⁺channels, lidocaine, at 100 μM, abolished tonic firing of Na⁺-dependent action potentials. However, the number of action potentials in the rebound LTS-evoked bursts decreased only slightly at this concentration, from 5.3 ± 0.3 to 4.3 ± 0.3 (n = 3; P < 0.001) (fig. 8A1). The burst discharges disappeared only at 600 μM, with the exception of the first spike (n = 4; fig. 8A2), which, consistent with our previous findings,24was resistant to lidocaine in all neurons tested (but blocked by 600 nM tetrodotoxin; fig. 5C), raising the possibility that lidocaine's potency for voltage-gated Na⁺channel blockade in thalamocortical neurons might be lower than for its blockade of I  ^h. At 600 μM, lidocaine produced a significant (7–10 mV) depolarization of the holding membrane potential and triggered repetitive firing of high-threshold spikes in response to the depolarizing current pulses. Application of 1 mM lidocaine led neither to a depolarizing shift in the holding potential nor to firing of high-threshold spikes (n = 5; fig. 8A3), although the latter could be evoked by increasing the amplitude of the depolarizing pulse (not shown). Similar to the findings in neurons pretreated with tetrodotoxin (fig. 5C), lidocaine at all three concentrations (100 μM, 600 μM, and 1 mM) concentration-dependently and reversibly increased LTS latencies (fig. 8A1–3). Consistent with previous findings of others,48we observed little effect of lidocaine on LTS magnitude.

Also similar to our results obtained in tetrodotoxin-pretreated neurons, application of lidocaine produced a reversible increase in slope resistance in the range from approximately −50 to −85 mV, a hyperpolarization of the RMP (table 1), and suppression of the time-dependent inward rectification. The lidocaine-induced changes (600 μM) in the current-voltage relationships reflected inhibition of a conductance with a reversal potential of −66.5 ± 2.4 mV (n = 3; fig. 8B shows the current-voltage curves of a representative neuron). Application of Ba²⁺(0.1 mM; fig. 8C) shifted this value above −55 mV (n = 4), toward the reversal potential of I  ^h. We found that the lidocaine-induced increase in R  ⁱwas smaller in the presence of Ba²⁺than that observed at baseline or in the presence of tetrodotoxin (table 1). Collectively, these data implicate a K⁺conductance (other than I  ^Kir) besides I  ^hin the actions of lidocaine at 600 μM.

Here, we have demonstrated that lidocaine reversibly and voltage-independently inhibited the hyperpolarization-activated mixed cation current, I  ^h, in rat ventrobasal thalamocortical relay neurons. Lidocaine blocked I  ^hwith high efficacy (producing near-complete blockade; fig. 3B) and a potency (IC⁵⁰, 72 μM) similar or higher in comparison with that associated with its best-known effect, voltage-gated Na⁺channel blockade.10,49Our findings in the thalamus are overall comparable with previous observations in the periphery, i.e. , rat dorsal root ganglion neurons (IC⁵⁰, 99 μM)11and also cardiac (sinoatrial) myocytes (IC⁵⁰, 38 μM).50 

The biophysical and pharmacologic properties of I  ^hin our experiments were similar to those previously reported in thalamocortical neurons42,51and correspond well to those characteristic for the underlying HCN2 channel isoform dominant in these neurons.12Lidocaine blocked the I  ^h-mediated time-dependent inward rectification without affecting the instantaneous inward rectification due to the K⁺current, I  ^Kir. In addition, lidocaine substantially delayed I  ^hdeactivation while exhibiting no effects on the rate and voltage dependence of I  ^hactivation. These observations suggest that lidocaine's action is unlikely to reflect a primary effect on channel gating.

Consistent with the identified role of I  ^hin determining membrane electrical properties,12,28,29,52its blockade by lidocaine was accompanied by a concentration-dependent hyperpolarization and led to large increases in the voltage responses to hyperpolarizing current pulses. The magnitude of lidocaine's effects declined at 1 mM, suggesting that other conductances counteracting the hyperpolarization and increase in R  ⁱwere activated. This observation is in agreement with the lidocaine-induced depolarization at more than 3 mM previously reported in cultured dorsal root ganglion neurons.10The precise mechanisms are unknown and have been speculated to involve blockade of ion channels and pumps playing a role in the maintenance of RMP.

In the current study, the effects of lidocaine on membrane potential critically depended on holding voltage. At potentials associated with the relay mode of operation (> approximately −60 mV; see Introduction), lidocaine, at 600 μM, depolarized neurons. Current-voltage analyses showed that the depolarization occurred due to the hyperpolarized reversal potential of the lidocaine-blocked conductance relative to the holding membrane potential. At the same time, 100 μM lidocaine did not depolarize neurons and produced a smaller increase in R  ⁱthan expected based on the concentration dependence of the I  ^hinhibition. Combined with the effects of Ba²⁺on reversal potential and R  ⁱchanges, these observations implicate the contribution of a K⁺conductance (other than the inward rectifier, I  ^Kir) blockade to lidocaine's actions that is substantially increasing at 600 μM. In good agreement is the report that lidocaine inhibits the hTREK1 current underlying leak K⁺conductance, with an IC⁵⁰of 180 μM.53The lidocaine-induced depolarization increased neuronal excitability mediated by (under the conditions of Na⁺-dependent action potential blockade) a high-threshold Ca²⁺conductance.33,39,46In this regard, our findings support the hypothesis of Mulle et al. ,54explaining the occurrence of dendritic high-threshold spikes in thalamocortical neurons in the presence of intracellular QX-314 (100 μM), a permanently charged lidocaine analog, as resulting from an increase in R  ⁱdue to inhibition of persistent Na⁺and/or K⁺conductances. With regard to lidocaine's concentration-dependent effects on passive membrane properties, it is of note that an older study with “blind” recordings in the ventral posterior lateral nucleus of Sprague-Dawley rats yielded some results at variance with the current findings on R  ⁱ, failing to find a statistically significant effect at 600 μM.24Whereas the precise reason is unclear, differences in recording technique and associated quality, species, animal age, neuronal homogeneity, and/or a type II error (an increase in R  ⁱoccurred in some neurons) may have contributed.

In the current investigations, we also found that by blocking I  ^h, lidocaine concentration-dependently altered firing properties of thalamocortical neurons in the burst mode, which in vivo  is associated with nonrapid eye movement sleep and drowsiness.16Specifically, lidocaine increased the latency of rebound LTSs. Most likely, inhibition of I  ^htail currents, known to evoke hyperpolarization-activated membrane potential overshoots,28,55accounts for these effects. Furthermore, at 100 μM, lidocaine reduced the number of Na⁺action potentials in LTS-evoked bursts. Our findings that the I  ^hblocker, ZD7288, produced the same effects are indicative of this action being due to inhibition of I  ^hrather than voltage-gated Na⁺channels. At the same time, the almost complete suppression of burst firing at 600 μM likely is mediated by both I  ^hinhibition and Na⁺channel blockade. Our results are consistent with previous observations that both pharmacologic (ZD7288) and “electronic” (dynamic clamp) I  ^hblockade increase LTS latency, partially suppress LTS-evoked bursts, and decrease the propensity of thalamocortical neurons to generate intrinsic δ oscillations.35,56We would therefore predict that lidocaine blockade of I  ^h, despite hyperpolarization, will reduce the ability of thalamocortical neurons to produce intrinsic burst firing and δ oscillations.57The fact that intracellular QX-314 (100 μM) inhibits intrinsic slow oscillatory activity in cat thalamocortical neurons in vivo  supports this prediction.54 

In a discussion of the potential clinical relevance of the current findings it is important to note that results from in vitro  animal investigations obviously cannot easily be translated to the in vivo  domain without consideration of experimental limitations. For example, whereas our current work has a specific focus on intrinsic properties of single ventrobasal thalamocortical neurons, I  ^halso is expressed in other neurons involved in thalamocortical networks, such as those in the thalamic reticular nucleus and cortex.13,42Future studies using such approaches as multiunit and field potential recordings, imaging, and neural network modeling will help define the I  ^h-mediated effects of lidocaine on the entire thalamocorticothalamic system and aid in filling the gap between findings from single cells in brain slices and higher levels of organization (and ultimately, the human patient).

These considerations notwithstanding, numerous lines of evidence render it a plausible possibility that the mechanisms of systemic lidocaine's concentration-dependent CNS effects involve varying degrees of thalamic I  ^hinhibition, thereby affecting neuronal excitability and oscillatory behavior. For example, with regard to the higher (and presumably epileptogenic/CNS-toxic) concentrations producing close to maximal I  ^hblockade in the current study, an absence of I  ^hin thalamocortical neurons of HCN2-deficient (−/−) knockout mice produces abnormal synchronized (3–5 Hz) electroencephalographic oscillations and facilitates the occurrence of spike-and-wave discharges.12In rat models, a decreased responsiveness of I  ^hto cyclic adenosine monophosphate in the ventrobasal thalamus promotes epileptogenesis.19,58,I  ^hblockade in other brain regions involved in the actions of systemic lidocaine59,60obviously may contribute to the complex array of this agent's CNS effects. Finally, given that the pathogenesis of lidocaine neurotoxicity involves an increase in intracellular Ca²⁺,9,10,61lidocaine-induced depolarization and Ca²⁺-mediated increase in excitability at high concentrations may well play a part. Clearly, a body of future research is required to further elucidate these mechanisms.

In addition to its toxic effects on the CNS, systemic lidocaine, at low, subconvulsive plasma concentrations (ranging from approximately 1–7 μg/ml or approximately 4–30 μM), is efficacious in alleviating acute postoperative as well as chronic neuropathic pain in humans2,4,5,62,63and animal models.64We recognize that extrapolation of ACSF concentrations from in vitro  rodent studies to in vivo  human plasma concentrations (where, among other factors, protein binding occurs and species differences play a role) requires caution. However, the therapeutic lidocaine concentration range in humans is near the lower end of that producing the I  ^hinhibition in this study (e.g. , approximately 23% suppression at 30 μM), which is noteworthy particularly because the current experiments were conducted at room temperature to facilitate stable recording conditions and slice viability. Given the pivotal role of the ventrobasal thalamus in pain and analgesia, our observations raise the possibility that moderate inhibition of thalamic I  ^hin the low micromolar range might represent a contributing mechanism for lidocaine's systemic analgesic actions. Again, future studies are required to test this hypothesis.

In addition to shedding new light on the mechanisms of lidocaine's supraspinal CNS effects, the current findings also emphasize on the role of I  ^has an emerging anesthetic drug target. For example, our findings with lidocaine, which has well-known general anesthetic properties (see Introduction),1,3share some noteworthy similarities with those recently obtained with propofol.57In thalamocortical neurons, propofol inhibited I  ^h-HCN2 at clinically relevant concentrations (e.g. , 36% at 5 μM; 23°C) and slowed I  ^hactivation. Consistent with our predictions on the in vivo  implications of lidocaine's I  ^hblockade, propofol's actions resulted in decreased regularity and frequency of δ oscillations in the neurons. Another example is ketamine, recently reported to block I  ^h-HCN1 in mouse cortical pyramidal neurons (approximately 29% at 20 μM; room temperature); in HCN1 knockout mice, ketamine showed a dramatically decreased hypnotic efficacy.65In addition, there is growing evidence that the anesthetic mechanisms of volatile agents involve I  ^hinhibition.66,67A comprehensive review on the role of I  ^hand HCN channels in anesthesia and other physiologic and pathologic conditions (including pain and epilepsy) has appeared recently.68 

In this work, we have shown that lidocaine concentration-dependently inhibited I  ^hin ventrobasal thalamocortical neurons at micromolar concentrations in vitro , with high efficacy and a potency similar or higher compared with that associated with its blockade of voltage-gated Na⁺channels. By inhibiting I  ^h, lidocaine profoundly altered membrane properties of neurons and reduced their ability to generate the rebound burst firing associated with slow oscillatory cerebral activity. Our findings provide new insight into the multiple overlapping mechanisms that underlie the complex array of concentration-dependent therapeutic and toxic effects that intravenous lidocaine exerts on the CNS and emphasize on the significance of I  ^has an emerging anesthetic drug target.

The authors express their gratitude to Christian Caritey (Engineering Technician) and Andy Jeffries (Operations Manager; both, Department of Anesthesiology, Pharmacology & Therapeutics, The University of British Columbia, Vancouver, British Columbia, Canada) for excellent technical assistance; James E. Cooke, Ph.D. (recent doctoral graduate, Department of Anesthesiology, Pharmacology & Therapeutics, The University of British Columbia), for valuable advice; Dietrich W. F. Schwarz, M.D., Ph.D. (Professor Emeritus, Department of Surgery, The University of British Columbia), for generously gifting equipment; and Ernest Puil, Ph.D. (Professor Emeritus, Department of Anesthesiology, Pharmacology & Therapeutics, The University of British Columbia), for his inspiration and constructive comments.