Ketamine is a commonly used anesthetic, but the mechanistic basis for its clinically relevant actions remains to be determined. The authors previously showed that HCN1 channels are inhibited by ketamine and demonstrated that global HCN1 knockout mice are twofold less sensitive to hypnotic actions of ketamine. Although that work identified HCN1 channels as a viable molecular target for ketamine, it did not determine the relevant neural substrate.
To localize the brain region responsible for HCN1-mediated hypnotic actions of ketamine, the authors used a conditional knockout strategy to delete HCN1 channels selectively in excitatory cells of the mouse forebrain. A combination of molecular, immunohistochemical, and cellular electrophysiologic approaches was used to verify conditional HCN1 deletion; a loss-of-righting reflex assay served to ascertain effects of forebrain HCN1 channel ablation on hypnotic actions of ketamine.
In conditional knockout mice, HCN1 channels were selectively deleted in cortex and hippocampus, with expression retained in cerebellum. In cortical pyramidal neurons from forebrain-selective HCN1 knockout mice, effects of ketamine on HCN1-dependent membrane properties were absent; notably, ketamine was unable to evoke membrane hyperpolarization or enhance synaptic inputs. Finally, the EC50 for ketamine-induced loss-of-righting reflex was shifted to significantly higher concentrations (by approximately 31%).
These data indicate that forebrain principal cells represent a relevant neural substrate for HCN1-mediated hypnotic actions of ketamine. The authors suggest that ketamine inhibition of HCN1 shifts cortical neuron electroresponsive properties to contribute to ketamine-induced hypnosis.
HCN1 channels are inhibited by ketamine, but the neural substrate for this blockade and a full understanding of the mechanistic basis for ketamine’s clinically relevant actions remain to be determined
Using a conditional knockout strategy to selectively delete HCN1, a relevant neural substrate for ketamine’s clinical action, HCN1-mediated hypnotic actions of ketamine was found in forebrain principal (excitatory) cells, consistent with the view that ketamine inhibition of HCN1 alters cortical neuron electroresponsive properties
ANESTHETIC compounds represent some of the most useful drugs in the clinical arsenal. Despite extensive clinical experience and intensive experimental scrutiny, the molecular and neural mechanisms by which most of these drugs mediate their important actions remain elusive.1,2 Prevailing ideas hold that contributions of individual molecular or neuronal targets depend on the particular anesthetic compound and the specific clinical endpoint examined.1–3 This implies that understanding actions of any anesthetic drug demands information regarding how modulation of a given target within a defined neural system can yield each clinically relevant outcome.2,3 In this work, we use conditional knockout mice to test whether inhibition of HCN1 channels specifically in forebrain principal cells contributes to ketamine-induced hypnosis.
In our earlier work, we made the unexpected observation that ketamine inhibited recombinant HCN1 channels at clinically relevant concentrations via a decrease in maximal current amplitude and a hyperpolarizing shift in the V½ of channel activation.4 These inhibitory effects of ketamine were stereoselective, with greater potency of channel inhibition obtained with the S-(+)-ketamine isomer that is also the more potent anesthetic, and subunit selective, modulating HCN1-containing homomeric or HCN1–HCN2 heteromeric channels but not HCN2 homomeric channels. We validated this action of ketamine on an HCN1-mediated native neuronal hyperpolarization-activated cationic current (Ih) by showing that ketamine-mediated inhibition of neuronal Ih was absent in HCN1−/− mice. We also demonstrated that mice with global deletion of HCN1 were approximately twofold less sensitive to ketamine-induced hypnosis, providing strong evidence that HCN1 channels represent a relevant target for this anesthetic endpoint.4
A role for HCN1 channels in constraining oscillatory activity in forebrain pyramidal neurons is predicted based on their depolarizing influence on membrane potential and by virtue of their ability to shunt excitatory synaptic inputs.5–7 Indeed, we found that inhibition of HCN1 channels by ketamine led to membrane hyperpolarization and improved dendritosomatic synaptic transfer in cortical pyramidal neurons.4 Thus, insofar as these two effects—membrane hyperpolarization and increased synaptic efficacy—are associated with an enhanced propensity for coherent cortical rhythms,8,9 such as those accompanying ketamine-induced hypnosis,10 we proposed that deletion of this ketamine target in cortical neurons could contribute to the diminished hypnotic actions of ketamine we observed in HCN1−/− mice.4
Because HCN1 subunits were deleted globally in the ketamine-resistant HCN1 knockout mice used for our previous work, we sought here to use conditional knockout strategies to better localize the neural substrate for HCN1-dependent actions of ketamine. To this end, we took advantage of a line of mice in which HCN1 channels were deleted only from forebrain principal cells, and we focused on the hypnotic endpoint because that was strongly affected in global knockouts.4 Insofar as we find that forebrain-selective HCN1 knockout mice largely recapitulated the decreased sensitivity to ketamine-induced hypnosis observed previously in global HCN1 knockout animals, these new results reinforce the idea that inhibition of forebrain HCN1 channels can contribute to hypnotic actions of anesthetics.
Materials and Methods
We used a Cre-loxP strategy to generate forebrain-selective deletion of HCN1 channels, essentially as described previously.6,11 In this system, genetic engineering is performed in mice to incorporate two so-called loxP sites flanking a region of the targeted gene (i.e., to “flox” the gene); when mice bearing the floxed gene are crossed to a separate group of mice that express the bacteriophage Cre-recombinase, the enzyme will excise the region between the loxP sites, inactivating the gene.12 For this work, we obtained mice with loxP sites flanking the fourth exon of the HCN1 gene, the region that encodes the channel pore domain and S6 transmembrane segment, from Dr. Eric Kandel (Columbia University, New York, NY).6,11 We also obtained the R1ag#5 line of transgenic mice from Dr. Scott Zeitlin (University of Virginia, Charlottesville, VA).13,14 In these mice, Cre-recombinase is expressed selectively in forebrain principal cells under the control of the CaMKIIα-promoter,13,14 providing the opportunity to restrict Cre-mediated gene inactivation only to those cells. The CaMK-Cre mice were crossed with HCN1f/+ mice and the progeny intercrossed to obtain breeding lines in which CaMKCre:HCN1f/f mice were mated with HCN1f/f mice, yielding littermates homozygous for floxed HCN1 alleles that were either Cre+ (forebrain-selective knockouts) or Cre− (control). All procedures involving animals were approved by the University of Virginia Animal Care and Use Committee (Charlottesville, VA).
To confirm HCN1 channel deletion from HCN1−/− mice and from the forebrain of CaMKCre:HCN1ff mice, HCN1 immunohistochemistry was performed concurrently on tissue obtained from HCN1+/+, HCN1−/−, HCN1f/f, and CaMKCre:HCN1f/f mice (n = 3). Mice were anesthetized with an intramuscular injection of ketamine (ketamine, 200 mg/kg; xylazine, 14 mg/kg) and perfused transcardially with phosphate-buffered saline followed by 4% paraformaldehyde/0.1 m phosphate buffer. Brains were removed and postfixed in the same solution overnight. Coronal brain sections (30 μm) including the cerebellum, hippocampus, and cortex were cut on a vibrating microtome (VT1000S; Leica Microsystems, Wetzlar, Germany) and stored in cryoprotectant (50% phosphate-buffered saline, 30% ethylene glycol, and 20% glycerol) until they were used for immunohistochemistry. Free-floating sections were washed in 0.1 m Tris-buffered saline before incubation for 30 min in 1% H2O2 and again washed in buffered saline. After incubating sections in blocking reagent (Basic M.O.M Immunodetection Kit; Vector Laboratories, Burlingame, CA) for 1 h, tissue was incubated overnight at 4°C in mouse anti-HCN1 (1:1000; clone N70/28; UC Davis/NIH NeuroMab Facility, Davis, CA). Thereafter, sections were incubated for 1 h in M.O.M. biotinylated anti-mouse IgG reagent followed by 1 h with avidin–biotin–peroxidase complex (Vectastain ABC Elite Kit; Vector Laboratories), before visualization with 3,3′-diaminobenzidine (DAB peroxidase substrate kit; Vector Laboratories). The 3,3′-diaminobenzidine reaction was quenched by washing in 0.1 m phosphate buffer. Sections were then mounted onto gelatinized slides, dried, dehydrated in ethanol, cleared in xylene, and cover-slipped using DPX mounting media (VWR International, Radnor, PA). Images were obtained on a Zeiss FS microscope (Carl Zeiss Microscopy, LLC, Thornwood, NY) equipped with a PixelFly digital camera (Cooke Corp., Romulus, MI) using IPLab software (Scanalytics; Spectra Services Inc., Ontario, NY).
Quantitative Real-Time Polymerase Chain Reaction
HCN1 expression levels were determined in Cre− floxed control mice (HCN1f/f) and CaMKCre-expressing HCN1f/f mice (CaMKCre:HCN1f/f) by using quantitative real-time polymerase chain reaction (qRT-PCR), as described previously.15 The cortex, hippocampus, and cerebellum were microdissected for RNA isolation from adult mice (with five of more animals per genotype). Total RNA was extracted from all tissues using TRIzol reagent (Invitrogen, Carlsbad, CA) following the manufacturer’s instructions. An equal quantity of RNA from each sample was reverse transcribed using Superscript II (Invitrogen) enzyme. The resulting complementary DNA was assayed to quantify the relative abundance of various mRNA species using the SYBR Green Real-Time polymerase chain reaction kit (Qiagen, Valencia, CA) according to the manufacturer’s protocol. The qRT-PCR was performed from each sample in quadruplicate, using an ICycler (Bio-Rad Laboratories, Hercules, CA), a previously validated HCN1 primer set (upper, AGG TTA ATC AGA TAC ATA CACC; lower, GAG TGC GTA GGA ATA TTG TTTT; 231-bp amplicon) and PCR conditions (3 min at 95°C; 40 cycles: 10 s, 95°C; 40 s, 60°C; 40 s, 72°C) that yielded greater than or equal to 97% efficiency.15 We used cyclophilin as an internal standard (upper, GGC TCT TGA AAT GGA CCC TTC; lower, CAG CCA ATG CTT GAT CAT ATT CTT; 91-bp amplicon), and control samples with no added template were included with every experiment. Each animal contributed a single data point, and qRT-PCR data were analyzed using a modification of the so-called ΔΔCt normalization procedure to obtain HCN1 mRNA levels for each genotype, relative to cyclophilin.16
Electrophysiologic Recordings from Mouse Cortical Pyramidal Neurons
Mice of both sexes (14–22 days old) were anesthetized (ketamine, 200 mg/kg; xylazine, 14 mg/kg, intramuscular injection) and transverse brain slices prepared as described previously.4,15,17 Slices were submerged in a recording chamber on a Zeiss Axioskop FS microscope and visualized with Nomarski optics; pyramidal cells were targeted for recording based on location in the slice and characteristic size and shape. For voltage clamp recording, pipettes (2–4 MΩ) were filled with KCH3O3S, 120 mm; NaCl, 4 mm; MgCl2, 1 mm; CaCl2, 0.5 mm; HEPES, 10 mm; EGTA, 10 mm; MgATP, 3 mm; and GTP-Tris, 0.3 mm (pH 7.2); for current clamp, pipette solution contained KCl, 17.5 mm; potassium gluconate, 122.5 mm; HEPES, 10 mm; EGTA, 0.2 mm; NaCl, 9 mm; MgCl2, 1 mm; MgATP, 3 mm; and GTP-Tris, 0.3 mm (pH 7.2). Recordings were obtained with an Axopatch 200B amplifier (Axon Instruments, Inc., Union City, CA) at room temperature (~24°C) in standard bath solution containing NaCl, 140 mm; KCl, 3 mm; HEPES, 10 mm; CaCl2, 2 mm; MgCl2, 2 mm; and glucose, 10 mm; we routinely added tetrodotoxin (0.5 µm; Alomone Labs, Jerusalem, Israel), BaCl2 (200 µm), and bicuculline/strychnine (both at 30 µm; Sigma, St. Louis, MO) to the bath, except where noted. In some experiments, ZD-7288 was used to block Ih (50 µm; Tocris Cookson, Bristol, United Kingdom). Ketamine (Ketaset; Fort Dodge Animal Health, Fort Dodge, IA) was diluted into the bath solution at the indicated concentrations.
The properties of HCN channel currents and neuronal Ih were determined from whole-cell voltage clamp experiments, essentially as described.18 Under current clamp, we measured input resistance and the magnitude of depolarizing “sag” from voltage response to hyperpolarizing current injection. We evoked excitatory postsynaptic potentials (EPSPs) in cortical pyramidal neurons using a pipette that was placed in the superficial layers of the cortex and connected to a stimulator (Grass S48; Grass Technologies, West Warwick, RI) via a stimulus isolation unit (Grass SIU5; Grass Technologies); the pipette was filled with standard bath solution and synaptic responses were evoked by applying 40-Hz, 5- to 10-V pulses (corresponding to 2–10 μA). The EPSP summation was quantified as the ratio of the amplitudes of the last and first evoked EPSPs in the train (EPSP5:EPSP1).4,15 All data are corrected (−8 mV) for a measured liquid junction potential.
Analysis of Anesthetic Action in Mice
Bolus injections of ketamine (5–20 mg/kg) were administered via the tail vein of each mouse (2–4 months old). Each animal was injected with only a single concentration of the drug on any given day; some animals received multiple doses (on different days, always separated by at least 1 week), but no individual animal contributed more than one data point for any given dose. Hypnosis was established by loss-of-righting reflex (LORR) that occurred within 10 s after completion of the injection and that persisted for at least 10 s thereafter19 ; we also measured the latency to regain the righting reflex.19
To assay ketamine effects on gross sensorimotor function, we used RotaRod (Med Associates Inc., Georgia, VT) and tail flick assays, essentially as described.20 For motor function, mice were placed on a rod rotating at constant acceleration (5–40 rpm), and the length of time the animal was able to stay on the rod was recorded (Med Associates Inc., Georgia, VT), under control conditions and then immediately after receiving a subanesthetic dose of ketamine (2.5 mg/kg, IV). To measure sensitivity to painful stimuli, mice were lightly restrained in a commercially available apparatus (Tail-Flick Unit 7360; UgoBasile, Comerio, Italy) while a radiant heat source (intensity and cutoff time were set at 90 and 12 s) was directed onto the tail; the time required for the mice to remove their tails from the heat source was recorded, before and after treatment with ketamine (20 mg/kg, IV, at 30-s intervals for 5 min)
Data Acquisition and Analysis
Data were analyzed statistically with SigmaPlot 11.0 (Systat Software, Inc., San Jose, CA). Results are presented as mean ± SEM. Comparisons of mean values were performed using two-way ANOVA, with post hoc pairwise comparisons used the Tukey correction of the t test. For evaluation of neuronal electroresponsive properties, genotype and drug treatment (ketamine, ZD7288) represented principal factors, with a repeated measures design allowing comparisons for each cell under different conditions. For behavioral effects of ketamine analyzed by two-way ANOVA with genotype and dose as principal factors (fig. 5), each observation was treated as an independent data point because not all individual animals were assessed at each dose. Quantal dose–response data of hypnotic effects of ketamine were fitted using logistic equations to obtain EC50 values for hypnosis in Cre+ and Cre− HCN1f/f mice, and those values were analyzed statistically using GraphPad Prism 5.0 (GraphPad Software, Inc., La Jolla, CA) by ANOVA. Differences in mean values were considered significant for values of P < 0.05.
To test the role of forebrain HCN1 expression in anesthetic actions of ketamine, we crossed HCN1f/f mice with a line of CaMKCre transgenic mice that express Cre-recombinase under the control of the CaMKIIα-promoter (see Materials and Methods for details of this conditional knockout strategy).13,14 Of relevance here, CaMKIIα is expressed only in excitatory (i.e., principal) cells of the forebrain, with expression progressively increasing from barely detectable levels at birth to approximately 10-fold higher levels by postnatal day 21 and a further approximately 2.5-fold increase by postnatal day 90.21 Thus, this construct drives Cre expression selectively to pyramidal neurons of the hippocampus and neocortex in the postnatal period,13,14 where it is expected to inactivate the HCN1 channel gene. It is worth noting that forebrain-selective HCN1 knockout mice were described previously to have enhanced spatial memory but otherwise normal sensorimotor function.6,11
We first verified forebrain-specific deletion of HCN1 and loss of HCN1-dependent membrane properties in cortical pyramidal cells from CaMKCre:HCN1f/f mice, including ketamine modulation. After confirming HCN1 deletion and absence of ketamine actions on HCN1 channels in those cells, we tested the sensitivity of CaMKCre:HCN1f/f mice to ketamine-induced hypnosis and effects on sensorimotor function.
HCN1 Expression Is Selectively Deleted from Forebrain of CaMKCre:HCN1f/f Mice
To demonstrate that the HCN1 channel gene was deleted selectively from forebrain neurons in CaMKCre:HCN1f/f mice, we performed immunohistochemistry and qRT-PCR to detect HCN1 expression. As shown in figure 1A, immunostaining for HCN1 was observed in the cortex, hippocampus, and cerebellum from mice that were homozygous for either wild-type (HCN1+/+) or floxed (HCN1f/f) alleles at the HCN1 locus; note the prominent immunostaining in apical dendrites from cortical pyramidal neurons and in the molecular layer of the hippocampal CA1 region (see arrows) (stratum lacunosum-moleculare). In tissue from global HCN1 knockout mice (HCN1−/−), HCN1 expression was undetectable in all brain areas. However, consistent with forebrain selective knockout expected for CaMKCre:HCN1f/f mice,6 HCN1 staining was largely absent from cortex and hippocampus but retained in cerebellum. Corroborating these immunohistochemical results, our qRT-PCR analysis revealed that HCN1 mRNA was reduced by more than 90% in cortex and hippocampus of CaMKCre:HCN1f/f mice (decreased by 93.6% and 96.8%, respectively), with no difference in HCN1 transcript levels in cerebellum (fig. 1B). Together, these data confirm forebrain-specific deletion of HCN1 in CaMKCre:HCN1f/f mice.
Ih Is Reduced and Ketamine Effects Are Diminished in Cortical Pyramidal Neurons from CaMKCre:HCN1f/f Mice
To verify deletion of HCN1 channel subunits functionally, we determined whether HCN1-dependent membrane properties and HCN1-dependent ketamine actions were diminished in cortical pyramidal neurons of CaMKCre:HCN1f/f mice.
In cortical pyramidal neurons, Ih is primarily mediated by HCN1 and HCN2 channels, with HCN1 subunits conferring relatively faster activation kinetics and more depolarized voltage-dependence in homomeric HCN1 channels or heteromeric HCN1–HCN2 channels17,22 ; homomeric HCN2 channels activate with slower kinetics and at more hyperpolarized potentials.17,22 Accordingly, as shown in the voltage clamp recordings of figure 2, Ih was smaller (Ih amplitude: 316.5 ± 42.6 pA for Cre−vs. 102.0 ± 13.4 pA for Cre+) and V½ of current activation was more negative (V½: −92.6 ± 1.6 mV for Cre−vs. −101.2 ± 1.6 mV for Cre+) in pyramidal neurons from CaMKCre:HCN1f/f mice than in those same cells from control HCN1f/f mice (n = 5 and 7 for Cre− and Cre+, P < 0.05 for each variable). In addition, the residual Ih in pyramidal neurons from CaMKCre:HCN1f/f mice showed slower activation kinetics (τfast at −118 mV: 106.4 ± 24.0 ms for Cre−vs. 358.3 ± 50.0 ms for Cre+; data not shown). These results are consistent with deletion of HCN1 from pyramidal neurons of CaMKCre:HCN1f/f mice, leaving a residual Ih provided primarily by homomeric HCN2 channels.
We previously found that ketamine inhibits recombinant HCN channels that contain an HCN1 subunit in either HCN1 homomeric or HCN1–HCN2 heteromeric conformations; by contrast, ketamine had no appreciable effect on HCN2 homomeric channels.4 This subunit-selective action of ketamine was also apparent in its effects on native Ih in cortical pyramidal neurons, which express both HCN1 and HCN2: ketamine inhibited Ih in pyramidal neurons from wild-type mice but not from global HCN1 knockout mice.4 Likewise, as shown in figure 2, we found that ketamine inhibited Ih in cortical pyramidal neurons from control HCN1f/f mice by evoking a hyperpolarizing shift in voltage dependence of activation (ΔV½: −8.4 ± 1.7 mV, from −92.6 ± 1.6 mV to −101.0 ± 1.3 mV) together with a decrease in maximal current amplitude (% inhibition: 27.6 ± 4.5%, from 316.5 ± 42.6 pA to 228.0 ± 32.8 pA; n = 5, P < 0.05 for both variables), whereas ketamine had essentially no effect on the smaller and slower residual Ih in cortical pyramidal neurons from CaMKCre:HCN1f/f mice (ΔV½: −1.8 ± 1.0 mV, from −101.2 ± 1.6 mV to −103.0 ± 1.3 mV; % inhibition: 2.9 ± 4.2%, from 102.0 ± 13.4 pA to 97.5 ± 11.8 pA, n = 7, P < 0.05) (fig. 2). After ketamine treatment, ZD7288 completely blocked all remaining Ih in cells from both genotypes. Together, these data indicate that the ketamine-sensitive HCN1 subunit was effectively deleted in cortical pyramidal neurons from CaMKCre:HCN1f/f mice, leaving a residual Ih that is unaffected by ketamine.
It is known that Ih contributes a resting inward current that provides a depolarizing membrane potential bias in cortical pyramidal neurons; in addition, during a step hyperpolarization, voltage-dependent activation of Ih produces a depolarizing “sag” in the membrane potential trajectory.5,22 Thus, as shown in figure 3, and as expected for deletion of HCN1 channels, cortical pyramidal neurons from CaMKCre:HCN1f/f mice presented with a more hyperpolarized resting membrane potential (resting membrane potential: −70.4 ± 1.1 mV for Cre−vs. −78.1 ± 0.7 mV for Cre+), greater input resistance (RN: 164.1 ± 8.9 MΩ for Cre−vs. 244.3 ± 29.1 MΩ for Cre+), and diminished sag (sag at −90 mV: −3.0 ± 0.3 mV for Cre−vs. −0.8 ± 0.2 mV for Cre+), by comparison with pyramidal neurons from control HCN1f/f mice (n = 5 for Cre− and Cre+, P < 0.05 for each variable). Moreover, whereas ketamine caused membrane hyperpolarization (resting membrane potential: from −70.4 ± 1.1 mV to −73.1 ± 0.9 mV, n = 5, P < 0.05), increased input resistance (RN: from 164.1 ± 8.9 MΩ to 180.5 ± 11.1 MΩ, n = 5, P < 0.05), and reduced depolarizing sag (from −3.0 ± 0.3 mV to 1.7 ± 0.2 mV at −90 mV; n = 5, P < 0.05) in pyramidal neurons from control HCN1f/f mice, these actions of ketamine were occluded in pyramidal neurons from CaMKCre:HCN1f/f mice (resting membrane potential: from −78.1 ± 0.7 mV to −78.3 ± 0.8 mV; RN: from 244.3 ± 29.1 MΩ to 248.9 ± 27.4 MΩ; sag: from −0.8 ± 0.2 mV to −0.6 ± 0.2 mV at −90 mV; n = 5, P > 0.05 for all variables). Again, blocking the remaining Ih with ZD7288 eliminated all genotype-dependent differences in resting membrane potential, RN, or sag. Overall, these data are consistent with removal of the ketamine-sensitive HCN1 subunit in cortical pyramidal neurons from CaMKCre:HCN1f/f mice.
Ketamine Modulation of Synaptic Properties Is Abolished in Cortical Pyramidal Neurons of CaMKCre:HCN1f/f Mice
In cortical pyramidal neurons, HCN1 is most prominently expressed in distal dendritic domains (e.g., see fig. 1A) and the corresponding dendritic Ih provides a major current shunt that serves to dampen synaptic inputs onto those cells.7 We therefore examined whether selective HCN1 deletion in cortical pyramidal neurons was associated with improved dendritosomatic synaptic transfer.
To evaluate effects of HCN1 deletion on basal and ketamine-modulated synaptic properties, we used a measure of synaptic summation that is known to be influenced by dendritic Ih.7 For this, EPSPs were evoked in pyramidal neurons by 40-Hz extracellular stimulation of the superficial cortex (fig. 4A), and EPSP temporal summation was calculated as the ratio of the final evoked EPSP (EPSP5) relative to the first EPSP (EPSP5/EPSP1) (fig. 4B); the degree of summation is most clearly evident in the aligned and normalized traces (see lower panels of fig. 4A). Note that temporal summation of evoked EPSPs was enhanced after HCN1 deletion (EPSP5/EPSP1 ratio: 1.9 ± 0.2 for Cre−vs. 3.1 ± 0.2 for Cre+, n = 5, P < 0.05). Importantly, whereas ketamine increased EPSP temporal summation by approximately 28% (from 1.9 ± 0.2 to 2.5 ± 0.2, n = 5, P < 0.05) in HCN1f/f mice, this effect of ketamine was totally eliminated in pyramidal neurons of CaMKCre:HCN1f/f mice (from 3.1 ± 0.2 to 3.2 ± 0.2, n = 5, P > 0.05).
CaMKCre:HCN1f/f Mice Are Significantly Less Sensitive to Hypnotic Actions of Ketamine
We previously suggested that HCN1-containing channels may be important molecular substrates for hypnotic anesthetic actions of ketamine4 because those channels are inhibited by ketamine and because the associated synaptic enhancement could promote cortical synchrony8,9,23 like that observed during anesthetic-induced hypnosis.10 Indeed, we showed that HCN1−/− mice showed a marked decrease in sensitivity to ketamine-induced LORR, an established behavioral surrogate in mice for hypnotic anesthetic actions.1,2 However, those earlier experiments using global knockout mice could not localize that effect to deletion of HCN1 channels within forebrain neurons.4 Here, we examined whether altered ketamine sensitivity was also observed in CaMKCre:HCN1f/f mice, in which HCN1 subunits were deleted only from forebrain neurons. As shown in figure 5A, the EC50 that produced LORR was significantly higher in CaMKCre:HCN1f/f mice (by approximately 31%; 10.4 ± 0.4 mg/kg for Cre−vs. 13.7 ± 0.3 mg/kg for Cre+, n = 31 and 52, P < 0.0001); there was no overlap of the 95% CIs for the estimated EC50 values between genotypes (9.6–11.3 mg/kg for Cre−vs. 13.0–14.3 mg/kg for Cre+). Also consistent with decreased sensitivity to ketamine, the duration of LORR evoked by ketamine was reduced in CaMKCre:HCN1f/f mice, by comparison with control HCN1f/f mice (fig. 5B) (F1,211 = 10.4, P < 0.002). In contrast to these differences in ketamine-induced LORR, we found no genotype-dependent differences in ketamine actions on gross sensorimotor function. In a tail-flick assay to assess responses to noxious sensory stimulation (fig. 5C), HCN1f/f and CaMKCre:HCN1f/f mice displayed essentially identical latencies to remove their tails from a painful heat source under control conditions, and both showed a similar, transient increase in latency following high-dose ketamine administration (20 mg/kg). In a RotaRod assay to assess motor function (fig. 5D), both control and conditional HCN1 knockout mice were able to remain on an accelerating rod for equal durations initially, and both genotypes showed comparable decreases in duration following a subanesthetic dose of ketamine (2.5 mg/kg). Note that at higher ketamine doses, even those where all animals retained their righting reflex (i.e., 5 mg/kg), the mice were too incapacitated to remain on the RotaRod for more than a few seconds. In sum, these data indicate that actions of ketamine on behaviors that involve spinal nociceptive reflexes or cerebellar function were unaffected in mice with forebrain-selective deletion of HCN1, as expected, whereas those mice were significantly less sensitive to ketamine-induced hypnosis.
In this article, we have shown that selective deletion of HCN1 subunits from forebrain principal cells in mice leads to decreased sensitivity to hypnotic actions of ketamine, addressing both the molecular and the neural bases for this anesthetic action. That is, this work provides additional support for our earlier proposition that HCN1-containing channels represent a behaviorally relevant molecular target of ketamine.4 It also further restricts the potential neural substrate for HCN1-mediated hypnotic actions of ketamine to forebrain principal neurons. In this latter respect, we found that ketamine-mediated inhibition of HCN1-containing channels in cortical pyramidal neurons enhances dendritosomatic synaptic integration, an effect that would support the synaptically mediated slow cortical rhythms that accompany the hypnotic state induced by ketamine.8–10,23 Thus, these data remain consistent with the hypothesis that cortical pyramidal neurons represent a plausible neural substrate for HCN1 contributions to ketamine-induced hypnosis.
As described by others, a number of criteria should be met for a molecular target to qualify as relevant for an anesthetic action.1,24 These include the following: (1) the target is modulated by the anesthetic at clinically relevant concentrations and with a stereoselectivity that matches the in vivo actions of the drug; (2) the target is expressed in a specific anatomical location where its modulation could conceivably mediate the specific behavioral effects of the anesthetic (i.e., plausibility); and (3) elimination of the target (or disrupting its modulation by the anesthetic) diminishes the ability of the drug to achieve a specific anesthetic endpoint in vivo. With respect to pharmacology, we previously found a stereoselective, subunit-specific inhibition of HCN1 channels by clinically appropriate concentrations of ketamine4 ; we showed that racemic ketamine shifts the V½ of HCN1 activation with an EC50 of approximately 8 µm, whereas the S-(+)-enantiomer does so with an EC50 of approximately 4 µm,4 consistent with an approximately twofold greater anesthetic potency of S-(+)-ketamine.25,26 Importantly, we also demonstrated by global elimination of HCN1 channels that HCN1 knockout mice were strikingly less sensitive to the hypnotic effects of ketamine4 ; this did not appear to represent a nonspecific effect of HCN1 deletion because sensitivity of the mice to the anesthetic effects of etomidate was unchanged.4 The current study confirms and extends that previous work in terms of plausibility by showing that this relative resistance to ketamine-induced hypnosis was largely recapitulated in mice with deletion of HCN1 channels restricted to forebrain principal cells, a brain region relevant for anesthetic-induced hypnosis. Thus, the accumulated evidence available to this point is consistent with the idea that forebrain HCN1 channels indeed contribute to ketamine-induced hypnosis.
Mechanistically, we favor the idea that the relevant forebrain cells are cortical pyramidal neurons, because inhibition of dendritic HCN1 channels in those cells causes membrane hyperpolarization with enhanced synaptic efficacy and thus yields conditions conducive to coherent cortical activity.4,8,9 It should be noted that Cre expression is not limited only to cortical pyramidal neurons in this line of mice. However, those cells represent the predominant site of HCN1 expression in the forebrain,22 and this convergence of both Cre and HCN1 expression thus favors the idea that pyramidal neurons are the likely site of action in this mouse model. Nevertheless, further refinement of the relevant neural substrate, especially between pyramidal neurons of the hippocampus and neocortex, will require mouse lines with even more restricted Cre expression.
As was seen with global HCN1−/− mice, ketamine remained capable of inducing hypnosis in CaMKCre-HCN1f/f mice, albeit at higher concentrations. Thus, other molecular targets must also be able to contribute to ketamine-induced hypnosis. In this respect, it is important to emphasize that ketamine is well known to be a potent modulator of N-methyl-d-aspartic acid (NMDA) receptor channels.1,26 However, at present there remains no evidence from NMDA receptor knockout mice to support a role for NMDA receptor modulation in anesthetic actions of ketamine. Moreover, MK801 is a more potent and selective NMDA receptor blocker than ketamine but is actually a poor anesthetic.27,28 Thus, despite widespread recognition of NMDA receptors as a ketamine target, the contributions of NMDA receptor antagonism to specific anesthetic actions of ketamine remain to be delineated. It is possible that inhibitory effects on NMDA receptors underlie other nonanesthetic central actions of ketamine (e.g., antidepressant or psychotomimetic)26,29–31 ; it is also possible that modulation by ketamine of other molecular targets (e.g., activation of tonic γ-aminobutyric acid type A receptors containing α6 or δ subunits) may contribute to its other anesthetic actions.32–34
In conclusion, these results indicate that forebrain principal cells represent a relevant neural substrate for HCN1-mediated hypnotic actions of ketamine. As knowledge of the molecular determinants for ketamine inhibition of HCN1 channels becomes available, it may be possible to develop knock-in mice that express ketamine-insensitive mutant HCN1 channels, preferably in more restricted neuronal populations, to even more rigorously define the molecular and neuronal targets for this specific clinical endpoint of ketamine.