The medial entorhinal cortex (MEC) plays a crucial role in spatial learning and memory. Whereas the MEC receives a dense histaminergic innervation from the tuberomamillary nucleus of the hypothalamus, the functions of histamine in this brain region remain unclear. Here, we show that histamine acts via H1Rs to directly depolarize the principal neurons in the superficial, but not deep, layers of the MEC when recording at somata. Moreover, histamine decreases the spontaneous GABA, but not glutamate, release onto principal neurons in the superficial layers by acting at presynaptic H3Rs without effect on synaptic release in the deep layers. Histamine-induced depolarization is mediated via inhibition of Kir channels and requires the activation of protein kinase C, whereas the inhibition of spontaneous GABA release by histamine depends on voltage-gated Ca2+ channels and extracellular Ca2+. Furthermore, microinjection of the H1R or H3R, but not H2R, antagonist respectively into the superficial, but not deep, layers of MEC impairs rat spatial learning as assessed by water maze tasks but does not affect the motor function and exploratory activity in an open field. Together, our study indicates that histamine plays an essential role in spatial learning by selectively regulating neuronal excitability and synaptic transmission in the superficial layers of the MEC.
The entorhinal cortex (EC) is widely regarded as the hub of cortico-hippocampal circuits (Witter et al. 1989; Witter et al. 2000). It has been divided into 2 subregions, the medial EC (MEC) and the lateral EC (LEC) (Canto et al. 2008; Van Cauter et al. 2013). The MEC contains spatially tuned neurons and is closely related to the path integration and spatial learning (Fyhn et al. 2004; Sargolini et al. 2006; Solstad et al. 2008; Suh et al. 2011; Van Cauter et al. 2013; Zhang et al. 2013; Zhang et al. 2014). Multimodal sensory information from various cortical areas converges onto neurons in the superficial layers (Layers II/III) of the EC, where it is processed and relayed into all subregions of the hippocampus (Burwell 2000; van Strien et al. 2009). In contrast, the deep layers of EC (Layers V/VI) receive the outputs from the CA1 and subiculum and then convey the highly processed sensory information to superficial layers of the EC (Dolorfo and Amaral 1998; van Haeften et al. 2003) and to other cortical areas (Witter et al. 1989; van Strien et al. 2009).
Consistent with the different anatomical connection patterns, the superficial and deep layers of the EC are functionally separated in the memory-encoding and replaying processes, respectively, and display different activity patterns across the distinct brain states (Chrobak and Buzsaki 1994; Buzsaki 1996; Battaglia et al. 2011). During the exploratory behavior, neurons in the superficial, but not deep, layers of the EC are highly active and their discharges are phase-locked to the negative peak of theta-coupled gamma oscillations (40–100 Hz) (Chrobak and Buzsaki 1994, 1998). The active state of the neurons in the superficial layers is required for the spatial information encoding as suppression of these neurons caused a dramatic impairment in rat learning phase (Chrobak et al. 2000; Deng et al. 2009; Igarashi et al. 2014). Interestingly, during awake immobility and slow-wave sleep, neurons in the superficial layers exhibit slow oscillation (<1 Hz) and become relatively silent compared with active wakefulness (Chrobak and Buzsaki 1994; Isomura et al. 2006; Hahn et al. 2012). At these off-line brain states, neuronal discharges in the deep, but not superficial, layers of the EC in conjunction with sharp-wave ripples are thought to be critical for the memory replaying and consolidation (Chrobak and Buzsaki 1994; Li et al. 2008; Ramadan et al. 2009; Inostroza and Born 2013).
Although it is evident that the neuronal activity in the superficial layers shows alternative potentiation and depression across distinct brain states, it is unknown which physiological modulators contribute to these effects. It is particularly interesting to consider that the arousal-promoting systems might to a large extent control the activity patterns in superficial layers of the EC since their activity exhibits a strong circadian variation. The histaminergic system, originating exclusively from the tuberomamillary nucleus (TMN) of the hypothalamus but projecting extensively to almost all the brain regions, represents a prominent arousal-promoting system in the central nervous system (CNS) (Lin 2000; Anaclet et al. 2009; Lin, Anaclet, et al. 2011; Lin, Sergeeva, et al. 2011) and is implicated in the regulation of several arousal-related physiological processes including learning and memory (Haas and Panula 2003; Haas et al. 2008; Alvarez 2009). Anatomically, the EC receives an extensive innervation from the histaminergic neurons (Airaksinen and Panula 1988). In normal human brain, it has been reported that histamine-containing nerve fibers were densely distributed in the superficial, but not the deep, layers of the EC (Panula et al. 1998). Moreover, histaminergic neurons specifically discharge at high rates during wakefulness, especially during attentive waking (Vanni-Mercier et al. 2003; Takahashi et al. 2006) and are closely related to theta oscillation in the hippocampal formation (Hajos et al. 2008). This activity pattern of the histaminergic neurons is highly correlated with that of neurons in the superficial, but not deep, layers of the EC.
Currently, the functions of histamine in the EC are still elusive. In the present study, we tested the hypothesis that histaminergic system might selectively increase the activity of the principal projection neurons in the superficial layers of the EC and thus facilitate superficial layer-dependent memory-encoding processes. Nevertheless, histamine might not affect the neuronal activity in the deep layers as they function during the off-lines states, during which the levels of histamine are relative low. Using a combination of electrophysiology, behavioral pharmacology, and immunohistochemical approaches, our results reveal the superficial layer-specific effects of histamine in regulation of neuronal excitability, neurotransmission, and spatial learning. Thus, these findings provide a novel pathway that at least partially explains the role of histaminergic system in spatial cognition and the different activity levels of the superficial layers in the MEC across distinct brain states.
Materials and Methods
Brain Slice Preparation
All experimental procedures involving animals were in compliance with the guidelines for the care and use of laboratory animals in the Third Military Medical University. Acute brain slices were obtained from male Sprague–Dawley rats (P14–20) by methods that have been described in detail previously (Li et al. 2011). Semi-horizontal slices (400 μm) containing the EC were prepared with an oscillating tissue slicer (Leica, VT1000) in an ice-cold section solution equilibrated with 95% O2 and 5% CO2 containing (in mm): sucrose, 220; KCl, 2.5; NaH2PO4, 1.25; NaHCO3, 26; MgCl2, 6; CaCl2, 1; and glucose, 10. The slices were initially incubated for at least 1 h at room temperature (20–24°C) in oxygenated (95%O2–5%CO2) artificial cerebrospinal fluid (ACSF, composition in mm: NaCl 124; KCl 3; NaHCO3, 26; MgCl2, 2; CaCl2, 2; and glucose 10). During recording sessions, the slices were transferred to a submerged chamber and continuously superfused with oxygenated (95%O2–5%CO2) ACSF at room temperature.
Whole-Cell Clamp Recordings
Targeted neurons were verified with an upright microscope equipped with Leica differential interference contrast optics and an infrared video imaging camera. Whole-cell recordings were performed on neurons with glass pipettes (3–5 MΩ) filled with an internal solution (composition in mm: potassium gluconate 125; KCl, 20; Hepes, 10; EGTA, 1; MgCl2, 2; ATP, 4; adjusted to pH 7.2–7.4 with 1 m KOH). For recordings of spontaneous miniature inhibitory postsynaptic currents (mIPSCs), at a holding potential of −60 mV, a Cs+ pipette solution (composition in mm: CsCl, 145; Hepes, 10; MgCl2, 2; EGTA, 1; ATP, 2; adjusted to pH 7.2–7.4 with 1 M CsOH) was used. Recording pipettes approached toward targeted neurons in the slice under positive pressure. After tight seal formation on the order of 1–2 GΩ made by negative pressure, the membrane patch was then ruptured by suction. Data were collected after at least 5 min of stabilization from the formation of whole-cell configuration unless stated otherwise. During recording sessions, series resistance was compensated 50–70% and cells were excluded from the study if the series resistance increased by >15% during recording or exceeded 20 MΩ. Unless stated otherwise, all the recordings were performed at room temperature.
Data were acquired with an EPC10 amplifier (HEKA Elektronik, Lambrecht/Pfalz) and stored for off-line analysis with Pulse/Pulseﬁt v.8.74 (HEKA Elektronik) and Igor Pro v.4.03 (WaveMatrics). The output signal was low-pass-filtered at 4 kHz and digitized at 10 kHz. To investigate the direct effect of histamine, histamine was puffed on the slice for 1 min. This drug application method has been described in detail in our previous study (Li et al. 2010). The mean membrane potential or holding current obtained during the last 6 s of the histamine application session was used for analysis and compared with baseline. For the miniature excitatory postsynaptic currents (mEPSCs) and mIPSCs, the methods for recording and analysis of these events were similar to those described elsewhere (Li et al. 2011). In brief, after stably recording of the baseline at least 2 min, histamine was bathed to the slices for 6 min. mEPSCs and mIPSCs recorded at last 2 min of the application session were analyzed, and their frequency and amplitude were compared with that of the baseline. Action potential, spontaneous mEPSCs, and mIPSCs were determined automatically by using Mini-analysis software (version 6.0, Synaptosoft). The detected synaptic events have fast onset and exponential decay kinetics, and obviously erroneous events were excluded manually by visual examination. The parameters for detecting synaptic events were identical in each cell in the absence or presence of drugs.
Cannula Implantation and Microinjection
Male Sprague–Dawley rats weighing 240–250 g were used in the behavioral tests. All animals were housed individually in thermoregulated (22–24°C) Plexiglas cages with a 12 h of light/dark cycle (lights on at 7:00 A.M.). The rats had ad libitum access to food and water.
All surgical procedures were performed under aseptic conditions. The rats were anesthetized with intraperitoneal injection of 2.5% sodium pentobarbital (2 mL/kg) and placed on a stereotaxic frame. The cranial surface was exposed by removal of the scalp, and the bregma was identified and used as the stereotactic reference point. Two stainless-steel guide cannulae (length 11 mm, o.d. 0.64 mm, and i.d. 0.45 mm) for the microinjection cannulae were bilaterally implanted into the superficial or deep layers of the MEC of each animal. The guide cannulae were secured to the skull with 2 stainless-steel skull screws and dental acrylic. The coordinates for the implantation were based on the rat brain atlas of Paxinos and Watson (2007) (superficial layers: AP −8.4 mm, ML ±4.4 mm, and DV −5.6 mm; deep layers: AP −8.3 mm, ML ±4.8 mm, and DV −4.6 mm). Each guide cannula was provided with a stainless-steel mandril to prevent obstruction. The drug microinjection approach is similar to that described in our previous study (Cun et al. 2014). In brief, at the time of infusion, an injection cannula (length 11.5 mm, o.d. 0.41 mm, and i.d. 0.25 mm) was inserted to protrude 0.5 mm beyond the tip of the guide cannula for microinjection of ACSF, H1R antagonist triprolidine (2 μm), H2R antagonist ranitidine (30 μm), and H3R antagonist clobenpropit (8 μm) using Hamilton syringes (0.5 μL/side, lasting 2 min). After microinjection, the injection cannula was left for an additional 2 min before a withdrawal to reduce efflux of ACSF or drugs.
Morris Water Maze Test
Behavior tests were conducted after 5 days of postsurgical recovery. All rats were trained in a white Morris water maze (diameter, 120 cm; height, 50 cm; water depth, 40 cm; and water temperature, 23 ± 2°C). The water was made opaque by the addition of semi-skimmed milk, which prevents the animals from seeing the platform. The pool was located on an elevated platform 20 cm above the floor in the center of a brightly lit room containing salient visual cues such as geometric shapes, wall posters, and electrical fittings on the wall. The swimming paths were tracked by a computer-assisted video-tracking device system that allows measurement of a number of parameters, including latency to find the platform, swimming distance, swimming speed, and quadrant analyses.
All rats were trained to find a hidden platform (10 cm in diameter, submerged 1 cm) in a water maze using protocols mentioned previously (Deng et al. 2009). They received 2 days of training, each day comprising 6 consecutive trials. The platform remained constant across all the trials. During the trials, rats were gently placed in the water maze facing the wall with 4 start positions varied in a predetermined and pseudorandom order. Each trial lasted for 60 s or until the rat successfully located the platform, with a 30-s intertrial interval. The rat that failed to find the platform within 60 s was guided to the platform by the experimenter and allowed to stay on the platform for 30 s before next trial. Six hours after completion of acquisition trials, a probe trial with the platform unavailable for 90 s was carried out. The rats were placed at the pool side opposite to the target quadrant. Probe trial performance of each group was expressed by the proportion of total time spent in each quadrant of the Morris water maze. The random level spent in each quadrant of the Morris water maze is 25%. Rats received ACSF or drug infusion 15 min prior to the training on Days 1 and 2 but no drug treatment before probe trial.
A square open field (50 cm × 50 cm × 40 cm) lit with 2 60-W floodlights above the field was used to assess general activity of rats. The movement and behavior of the rats were recorded by a computer-assisted video-tracking system, which allows measurement of a number of parameters, including movement distance and number of rearings. Rats received ACSF or drug infusion 15 min prior to the open-field test. After placing a rat at the center of the arena, the movement distance and number of rearings, which represent locomotor activity and exploratory activity, respectively, were recorded for 5 min.
The method used to identify the injection sites in the present study was similar to that described in previous study (Deng et al. 2009). In brief, after completion of the behavioral tests, rats were anesthetized with an overdose of sodium pentobarbital and perfused transcardially with saline followed by 4% paraformaldehyde. The brains were removed and then stored in 30% sucrose and 4% paraformaldehyde solution for dehydration. Frozen sections were prepared (30 μm) from the dorsal to the ventral and stained with cresyl violet. By visualizing the sections with an inverted microscope, the track of cannulae from dorsal to ventral can be identified. The last section from dorsal to ventral containing the injection hole was selected, and this site was regarded as the drug injection site. The injection sites were confirmed in the present study. Data from the rats with incorrect injection sites were excluded from analysis.
Sprague–Dawley rats (weighing 200–250 g) were anesthetized deeply with 2.5% sodium pentobarbital (2.4–2.8 mL/kg, intraperitoneal) and then perfused transcardially with 300 mL normal saline, followed by 500 mL of 4% paraformaldehyde in 0.1 m phosphate buffer. Brains were removed, postfixed for 12 h at 4°C in 4% paraformaldehyde, and then cryoprotected with a 4% paraformaldehyde and 30% sucrose solution in phosphate buffer for 48 h. Horizontal sections (30 μm) containing the MEC were prepared by using a freezing microtome (CM 3050S, Leica). The sections were first incubated in 10% normal bovine serum for 30 min and then incubated overnight at 4°C with primary antibodies to H1Rs or H2Rs: a rabbit anti-H1R polyclonal antibody (1:100; Santa Cruz Biotechnology) or a rabbit anti-H2R polyclonal antibody (1:50; Santa Cruz Biotechnology). The specificity of anti-H1R and anti-H2R primary antibodies has been tested in other studies (Peng et al. 2013; Zhuang et al. 2013). After washing in PBS, these sections were incubated in the ﬂuorescein-labeled secondary antibody solutions (goat anti-rabbit 1:2000; Invitrogen) for 2 h at room temperature in the dark and then mounted onto slides with coverslips. For double staining of immunofluorescence, the sections were washed with PBS 3 times, blocked with 1% BSA (bovine serum albumin) and 0.4% Triton X-100 for 30 min in 37°C. The sections were incubated with the first primary antibodies rabbit anti-H3Rs (1:300, Sigma) overnight at 4°C. The specificity of anti-H3R primary antibodies has been checked in the brain neurons according to the product information. After washing with PBS 3 times, the sections were incubated with Alexa 568-conjugated donkey anti-rabbit secondary antibody (1:500, Invitrogen) at 37°C for 2 h. Then, the sections were blocked again with 1% BSA and 0.4% Triton X-100 for 30 min in 37°C, incubated with the second primary antibodies goat anti-GAD-67 (1:200, Santa Cruz Biotechnology) overnight at 4°C. The sections were washed with PBS 3 times, incubated with Alexa 488-conjugated donkey anti-goat secondary antibody (1:500, Invitrogen) at 37°C for 2 h. The sections were washed with PBS 3 times, then nuclear-stained with 0.01% 4′, 6′-diamidino-2-phenylindole (DAB, Sigma), and mounted onto coverslips with Glycergel mounting medium (Dako). Negative controls had PBS or polyclonal rabbit immunoglobulin G (Santa Cruz Biotechnology) instead of the primary antibody. All images were taken with an inverted laser scanning confocal microscope (FV1000; Olympus).
Histamine, triprolidine, ranitidine, clobenpropit, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), D-2-amino-5-phosphonovaleric acid (AP5), picrotoxin (PIC), tetrodotoxin (TTX), BaCl2, CsCl, bisindolylmaleimide II (BIS-II), and tetraethylammonium (TEA) were purchased from Sigma. Voltage-dependent potassium channel blocker 4-aminopyridine (4AP) was obtained from Tocris Cookson. Histamine, triprolidine, ranitidine, clobenpropit, AP5, TTX, BaCl2, CsCl, TEA, and 4AP were dissolved in ACSF, whereas CNQX, PIC, and BIS-II were dissolved in dimethyl sulfoxide. All drugs were prepared as concentrated stock solutions and frozen at −20°C until use. The drugs from stock were freshly diluted to the desired concentrations. The final concentration of dimethyl sulfoxide did not exceed 0.1%.
Values were presented as the means ± SEM. Student's paired or unpaired t-test, Kolmogorov–Smirnov (K–S) test, one-way, and repeated-measures’ analysis of variance (ANOVA), and Fisher's protected least significant difference (LSD) post hoc testing were used for statistical analysis. Significant differences were accepted at P < 0.05.
Histamine Specifically Increases the Excitability of Principal Neurons in the Superficial, but not Deep, Layers of the MEC
Stellate neurons are the most prominent principal projection neurons in the superficial layers of the MEC. These neurons are primarily located in Layer II and the border between Layer II and Layer III and have a polygonal soma with multiple thick sparsely primary dendrites radiating out from the cell body (Supplementary Fig. 1A,B). A notable feature of stellate neurons is that they showed profound depolarizing voltage sags in response to hyperpolarizing current pulses (Fig. 1A and Supplementary Table 1) and had larger hyperpolarization-activated currents in voltage-clamp recordings (Supplementary Fig. 1C and Supplementary Table 1). This unique electrophysiological feature is attributed to stellate neurons expressing a high level of hyperpolarization-activated, cation nonselective channels at their somata (Nolan et al. 2007).
Considering the finding of moderately dense histaminergic innervation in the EC, we carried out current- or voltage-clamp recordings to examine the effect of histamine on the activity of stellate neurons in the superficial layers of the MEC. Under current-clamp mode, local application of histamine (1–1000 μm) significantly depolarized the membrane potential on a majority of the recorded stellate neurons (76.3%, 61/80). The depolarization was reversed following washout and sufficient to stimulate an increase in firing frequency in part of the recorded cells. The histamine-induced excitation on stellate neurons was concentration dependent. Local application of 1, 10, 100, and 1000 μm histamine generated a membrane depolarization of 1.1 ± 0.1 mV (n = 4), 3.3 ± 0.3 mV (n = 12), 6.3 ± 0.5 mV (n = 3), and 6.5 ± 0.6 mV (n = 3), respectively (F3, 21 = 28.30; P < 0.001; one-way ANOVA; Fig. 1B,C). We also recorded the holding currents of stellate neurons in voltage-clamp at −60 mV, a potential close to the resting membrane potential. Under these conditions, application of histamine (10 μm) significantly induced an inward current (control: −25.1 ± 11.7 pA; histamine: −52.0 ± 13.9 pA; n = 5; P < 0.01; paired t-test; Fig. 1D).
Accordingly, we also investigated the effect of histamine on the superficial pyramidal neuron activity, another type of principal projection neurons that were identified by their morphology, location, and electrophysiological properties (Supplementary Fig. 1D–I, Supplementary Fig. 2A,D, and Supplementary Table 1) (Beed et al. 2010; Canto and Witter 2012). Similarly, local application of histamine excited the pyramidal neurons (Supplementary Fig. 2B,E). Histamine (10 μm)-induced depolarization was not significantly different between stellate and pyramidal neurons in the Layer II of the MEC (stellate neurons: 3.3 ± 0.3 mV, n = 12; pyramidal neurons: 3.4 ± 0.5 mV, n = 7; P = 0.79; unpaired t-test; Supplementary Fig. 2C). Moreover, when recording at the somata, no significant difference was found in histamine-induced depolarization between pyramidal neurons in the Layers II and III (pyramidal neurons in Layer II: 3.4 ± 0.5 mV, n = 7; pyramidal neurons in Layer III: 3.8 ± 0.7 mV, n = 5; P = 0.67; unpaired t-test; Supplementary Fig. 2F).
EC-cortical projections primarily arise from Layer V pyramidal neurons (Canto et al. 2008). These neurons are characterized by a pyramidal or elongated soma (Supplementary Fig. 1K). Moreover, unlike the stellate neurons, they had no depolarizing voltage sags (Fig. 1E and Supplementary Table 1) and small hyperpolarization-activated currents (Supplementary Fig. 1L and Supplementary Table 1). We accordingly tested whether histamine affected the excitability of these neurons in the deep layers of the MEC. When recording at the somata, application of histamine at 10 μm failed to change the membrane potentials (control: −62.0 ± 1.6 mV; histamine: −61.8 ± 1.8 mV; n = 8; P = 0.61; paired t-test; Fig. 1F,H) and holding currents (control: −46.2 ± 11.2 pA; histamine: −46.0 ± 10.6 pA; n = 19; P = 0.94; paired t-test; Fig. 1G,I) in all the recorded pyramidal neurons from Layer V of the MEC. In addition, the membrane potentials (−60.7 ± 0.7 and −61.1 ± 0.6 mV before and during applying 100 μm histamine, n = 10, P = 0.22; −60.5 ± 0.6 and −60.9 ± 0.7 mV before and during applying 1 mm histamine, n = 10, P = 0.17; paired t-test; Fig. 1F,H) and holding currents (−32.3 ± 14.3 and −33.5 ± 14.6 pA before and during applying 100 μm histamine, n = 12, P = 0.07; −30.8 ± 12.4 and −29.9 ± 12.2 pA before and during applying 1 mm histamine, n = 13, P = 0.19; paired t-test; Fig. 1G,I) of Layer V pyramidal neurons were not affected by 100 μm and 1 mm histamine. We also investigated the effects of histamine on the interneuron excitability in the MEC. The superficial layer and deep layer interneurons have large input resistance (superficial layers: 317.2 ± 29.4 MΩ, n = 6; deep layers: 364.9 ± 51.8 MΩ, n = 7) and fast after hyperpolarization (fAHP) (superficial layers: 18.2 ± 1.5 mV, n = 6; deep layers: 17.4 ± 0.7 mV, n = 7) compared with the principal projection neurons. When recording at the somata, the membrane potentials of the interneurons in the MEC were not affected by histamine at concentrations of 10 μm (superficial layers: −63.4 ± 1.4 and −63.3 ± 1.0 mV before and during applying, n = 4, P = 0.78; deep layers: −61.3 ± 1.8 and −61.1 ± 1.4 mV before and during applying, n = 5, P = 0.66; paired t-test), 100 μm (superficial layers: −63.4 ± 0.9 and −63.6 ± 0.9 mV before and during applying, n = 5, P = 0.64; deep layers: −61.4 ± 1.3 and −61.0 ± 1.1 mV before and during applying, n = 6, P = 0.13; paired t-test), and 1 mm (superficial layers: −63.5 ± 0.9 and −63.8 ± 0.9 mV before and during applying, n = 4, P = 0.60; deep layers: −61.6 ± 1.8 and −61.0 ± 1.9 mV before and during applying, n = 6, P = 0.26; paired t-test). Together, these results suggest that histamine selectively increases excitability of principal projection neurons in the superficial, but not deep, layers of the MEC when recording at the somata. Considering the effect of histamine in the superficial layers is not cell-type specific and the axons of stellate neurons forming the perforant pathway provide the main output from the EC to the hippocampus (Alonso and Klink 1993; Canto et al. 2008), we primarily focused on the stellate neurons in the superficial layers to investigate the cellular mechanisms underlying histamine-induced excitation.
Histamine Selectively Decreases Spontaneous GABA, but Not Glutamate, Release in the Superficial, but Not Deep, Layers of the MEC
Since synaptic release is crucial for regulating individual neuronal as well as neural network excitability, it is important to know whether histamine affects synaptic transmission in the MEC. To test this, we recorded spontaneous miniature postsynaptic currents in the presence TTX (1 μm) to block voltage-gated Na+ channels. An alteration in the amplitude of spontaneous miniature postsynaptic currents reflects a postsynaptic mechanism, whereas a change in the frequency of these currents signifies a change in the presynaptic release (Li et al. 2011).
We initially recorded the mEPSCs in the presence of GABAA receptor antagonist PIC (100 μm) at a fixed potential of −60 mV. Similar to other cortical neurons, the stellate neurons in the MEC exhibited a continuous level of fast excitatory activities that were completely blocked by CNQX (10 μm) and AP5 (50 μm), confirming that they were mediated by ionotropic glutamate receptors (data not shown). After a stable recording, application of histamine (10 μm) did not influence the cumulative probability distributions of interevent interval (P = 0.31, K–S test, Fig. 2B, left) and amplitude (P = 0.14, K–S test, Fig. 2B, right). On average, histamine did not affect the frequency (control: 2.84 ± 0.47 Hz, histamine: 2.99 ± 0.36 Hz, washout: 2.97 ± 0.51 Hz, n = 7, P = 0.92; one-way repeated-measures ANOVA; Fig. 2C, left) and amplitude (control: 9.00 ± 0.61 pA, histamine: 8.95 ± 0.68 pA, washout: 8.74 ± 0.64 pA, n = 7, P = 0.64; one-way repeated-measures ANOVA; Fig. 2C, right) of mEPSCs in 7 stellate neurons examined. Together, these data indicate that histamine does not influence the glutamatergic transmission in the superficial layers.
To study modulation of GABAA receptor-mediated mIPSCs, ionotropic glutamate receptors were blocked with CNQX (10 μm) and AP5 (50 μm). After establishing the whole-cell configuration, a period of at least 15 min was allowed for the equilibration of intracellular and recording solutions. Since the pipette recording solution contained a high concentration of CsCl (145 mm), mIPSCs were recorded from the stellate neurons as inward currents that were abolished by PIC (100 μm), indicating that these currents were mediated by GABAA receptors (data not shown). After a stable baseline was achieved, application of histamine (10 μm) reversibly suppressed the mIPSC activity (Fig. 2D) and caused a rightward shift of the interevent interval distribution of mIPSCs (P < 0.001, K–S test, Fig. 2E, left), reflecting a decrease in mIPSC frequency. However, histamine did not affect the cumulative distribution of mIPSC amplitude (P = 0.34, K–S test, Fig. 2E, right). There was a group difference in the frequency of mIPSCs (control: 2.13 ± 0.34 Hz, histamine: 1.43 ± 0.24 Hz, washout: 1.93 ± 0.36 Hz, n = 7, P < 0.01; one-way repeated-measures ANOVA; Fig. 2F, left). LSD post hoc comparisons revealed that the frequency of mIPSCs was decreased to 67 ± 11% of control after application of histamine (P < 0.001, Fig. 2F, left). Application of histamine did not affect the mIPSC amplitude (control: 21.30 ± 1.62 pA, histamine: 21.80 ± 2.34 pA, washout: 22.49 ± 2.78 pA, n = 7, P = 0.62; one-way repeated-measures ANOVA; Fig. 2F, right). Together, these findings suggest that histamine selectively inhibits GABAergic synaptic transmission in the superficial layers of the MEC.
We also investigated the effect of histamine on the synaptic release on to the pyramidal neurons in the Layer V of the MEC. The protocols used for isolating the mEPSCs and mIPSCs are the same as those mentioned above. Application of histamine had no impact on the shape of mEPSC or mIPSC cumulative probability curves of interevent intervals (mEPSCs: P = 0.50; mIPSCs: P = 0.59; K–S test; Fig. 3B,E) and event amplitudes (mEPSCs: P = 0.42; mIPSCs: P = 0.43; K–S test; Fig. 3B,E), respectively. Histamine did not affect the frequency (control: 1.36 ± 0.23 Hz, histamine: 1.24 ± 0.25 Hz, washout: 1.23 ± 0.22 Hz, n = 9, P = 0.39; one-way repeated-measures ANOVA; Fig. 3C, left) and amplitude (control: 13.90 ± 1.37 pA, histamine: 14.06 ± 1.31 pA, washout: 13.90 ± 1.49 pA, n = 9, P = 0.98; one-way repeated-measures ANOVA; Fig. 3C, right) of mEPSCs in 9 pyramidal neurons examined. Similarly, the frequency (control: 2.24 ± 0.36 Hz, histamine: 2.27 ± 0.24 Hz, washout: 2.15 ± 0.26 Hz, n = 7, P = 0.84; one-way repeated-measures ANOVA; Fig. 3F, left) and amplitude (control: 31.50 ± 3.05 pA, histamine: 31.91 ± 2.85 pA, washout: 32.60 ± 3.74 pA, n = 7, P = 0.86; one-way repeated-measures ANOVA; Fig. 3F, right) of mIPSCs were unaffected by histamine. Together, these results imply that histamine selectively modulates the GABAergic, but not glutamatergic, synaptic transmission in a superficial layer-specific manner.
Histamine-Induced Depolarization in the Superficial Layers Is Mediated by H1R-Dependent Inhibition of Kir Channels
The actions of histamine are mediated via 3 distinct types of G-protein-coupled receptors (H1Rs, H2Rs, and H3Rs) in the CNS. H1Rs and H2Rs are postsynaptically coupled to Gq/11 and Gs proteins, respectively, whereas H3Rs, coupled to Gi/o protein, are preferentially located at presynaptic terminals acting as auto- or heteroreceptors to modulate the release of a number of neurotransmitters (Arrang et al. 2007; Haas et al. 2008; Passani and Blandina 2011). Here, the receptor mechanism underlying the direct excitatory action of histamine was determined by using histamine receptor-selective antagonists. The effects of histamine blocked by these antagonists can be always mimicked by the corresponding histamine receptor agonists (Zhou et al. 2006; Dai et al. 2007; Lundius et al. 2010). In the presence of triprolidine (2 μm), a selective H1R antagonist, the histamine (10 μm)-elicited depolarization on stellate was abolished (3.0 ± 0.7 and 0.1 ± 0.1 mV before and after applying triprolidine, n = 5, P < 0.05; paired t-test; Fig. 4A). However, application of H2R blocker ranitidine (30 μm) failed to alter histamine (10 μm)-induced membrane depolarization in stellate neurons (3.4 ± 0.2 and 3.1 ± 0.2 mV before and after applying ranitidine, n = 8, P = 0.32; paired t-test; Fig. 4B). Neither did application of clobenpropit (400 nm), a selective H3R antagonist, significantly change histamine (10 μm)-elicited membrane depolarization (3.5 ± 0.2 and 3.4 ± 0.2 mV before and after applying clobenpropit, n = 4, P = 0.59; paired t-test; Fig. 4C). Consistent with this, the results of immunofluorescence reveal that H1Rs were expressed at the cell bodies and dendritic processes of principal neurons in the superficial layers, but not in deep layers, of the MEC (Fig. 4D–F) whereas the distribution of H2Rs in the superficial layers and deep layers of the MEC was relative scarce compared with H1Rs (Supplementary Fig. 3A–C). Together, these results provide direct electrophysiological and morphological evidence that histamine depolarizes the stellate neurons in the superficial layers of the MEC by activation of H1Rs.
To clarify the ionic mechanisms underlying the histamine-induced depolarization in stellate neurons, we first tested whether the histamine-induced depolarization was a direct postsynaptic effect. Histamine (10 μm)-induced depolarization persisted in the presence of either TTX (1 μm) (3.6 ± 0.6 and 3.3 ± 0.5 mV before and after applying TTX, n = 5, P = 0.81; paired t-test; Supplementary Fig. 4A) or CNQX (10 μm), AP5 (50 μm), and PIC (100 μm) (3.5 ± 0.3 and 3.2 ± 0.4 mV before and after applying, n = 5, P = 0.38; paired t-test; Supplementary Fig. 4B) known to block synaptic transmission, thus indicating that the action of histamine is postsynaptic. Moreover, input conductance was determined before and after the application of histamine using a series of hyperpolarized voltage pulses. Accompanying the histamine-induced depolarization, 0.6 nS, n = 4, P < 0.05; paired t-test; Supplementary Fig. 4C), suggesting that histamine causes a closure of ion channels on the postsynaptic membrane.
To test whether the effect of histamine would result from the closure of a potassium conductance, we conducted slow ramp command tests (dV/dt = –10 mV/s) to evaluate the I–V curves for a large voltage range in the absence and presence of histamine. In an ACSF with Ko at 3 mm (estimated the theoretical K+ reversal potential by the Nernst equation: Ek = –88.59 mV), comparing voltage-clamp ramps in control and in the presence of histamine indicated a reversal of the histamine effect of approximately −90 mV (Fig. 5A). In 7 neurons, the mean (± SEM) reversal potential was −87 ± 1 mV, thus very close to the estimated Ek (Fig. 5D). Performing the same experiment in an ACSF with Ko at 10 mm (estimated Ek = −58.17 mV) indicated a reversal of the histamine effect of approximately −60 mV (Fig. 5C). In 4 cells, the mean reversal was now −59 ± 1 mV, thus again very close to the estimated Ek (Fig. 5D). Thus, these results indicate that the action of histamine on the stellate neurons in the MEC is mediated through the closure of a potassium conductance. Subtracting the control from the current recorded during histamine application yielded a difference current representing the histamine-induced current. The difference current obtained from the 7 stellate neurons in normal ACSF exhibited a strong inward rectification (Fig. 5B), indicating that the potassium channels inhibited by histamine are inwardly rectifying K+ (Kir) channels.
Next, we applied the classic K+ channel blockers to identify the properties of the involved K+ channels. Histamine (10 μm)-induced depolarization in stellate neurons was not significantly altered in the extracellular solution containing 2 mm 4AP (control: 3.1 ± 0.2 mV; 4AP: 3.0 ± 0.3 mV; n = 6; P = 0.77; paired t-test; Fig. 5G,I) or 10 mm TEA (control: 3.2 ± 0.8 mV; TEA: 3.3 ± 0.8 mV; n = 4; P = 0.89; paired t-test; Fig. 5H,I), suggesting that histamine-induced excitation is insensitive to the voltage-dependent potassium channel blockers. Since Kir channels are sensitive to Ba2+ (Hibino et al. 2010), we therefore tested the role of Ba2+ in histamine-induced membrane depolarization. Inclusion of Ba2+ (2 mm) in the extracellular solution induced membrane depolarization. After the baseline stabilized, application of histamine (10 μm) failed to elicit membrane depolarization (control: 3.2 ± 0.5 mV; Ba2+: 0.3 ± 0.2 mV; n = 6; P < 0.01; paired t-test; Fig. 5E,I). Considering some of two-pore-domain potassium channels are sensitive to Ba2+ (Deng et al. 2009; Xiao et al. 2009), we therefore tested the role of another Kir channel blocker Cs+, which has no effect on the two-pore-domain potassium channels. In the presence of Cs+ (3 mm), histamine (10 μm)-elicited membrane depolarization was also almost completely blocked (control: 3.1 ± 0.5 mV; Cs+: 0.1 ± 0.3 mV; n = 5; P < 0.01; paired t-test; Fig. 5F,I). Together, these results indicate that histamine-induced membrane depolarization is mediated by Kir channels, but not by two-pore-domain potassium channels.
H1Rs are coupled to Gq/11 protein to activate protein kinase C (PKC) signaling pathway (Passani and Blandina 2011). We next determined the role of PKC in histamine-induced depolarization. After bath incubation with PKC inhibitor BIS-II (500 nm) at least for 90 min, the effect of histamine on stellate neurons in the MEC was blocked (Supplementary Fig. 5A). The histamine-elicited depolarization was significantly reduced to 7.0 ± 4.2% of control (control: 3.3 ± 0.3 mV, n = 12; BIS-II: 0.23 ± 0.14 mV, n = 10; P < 0.001; unpaired t-test; Supplementary Fig. 5B) in the presence of BIS-II. Furthermore, histamine failed to induce an inward current in the slow ramp command tests (Supplementary Fig. 5C). These results together demonstrate that the activation of PKC is necessary for histamine-induced excitation.
Inhibition of Spontaneous GABA Release by Histamine in the Superficial Layers Requires H3R-Dependent Modulation of the VGCCs
Next, we investigated the histamine receptor subtype involved in the regulation of inhibitory synaptic transmission. In this set of experiments, we first added triprolidine (2 μm), ranitidine (30 μm), or clobenpropit (400 nm) into the bathing solution after a stable baseline recording and then the histamine (10 μm) was superfused. Neither triprolidine nor ranitidine changed the histamine-induced shift of the interevent interval distribution (triprolidine: P < 0.001; ranitidine: P < 0.001; K–S test; Fig. 6B,D). There was a significant group difference in the frequency of mIPSCs for the triprolidine (control: 2.05 ± 0.24 Hz, triprolidine: 1.91 ± 0.25 Hz, histamine + triprolidine: 1.31 ± 0.19 Hz, washout: 1.63 ± 0.36 Hz, n = 6, P < 0.05, one-way repeated-measures ANOVA on ranks, Fig. 6A,B)- or ranitidine (control: 1.97 ± 0.41 Hz, ranitidine: 2.14 ± 0.52 Hz, histamine + ranitidine: 1.26 ± 0.24 Hz, washout: 1.81 ± 0.37 Hz, n = 7, P < 0.05, one-way repeated-measures ANOVA, Fig. 6C,D)-treated experiments. Post hoc comparisons revealed that application of histamine still reduced the mIPSC frequency in the presence of triprolidine (P < 0.05, Dunn's test, Fig. 6A,B) or ranitidine (P < 0.05, LSD test, Fig. 6C,D). However, histamine-mediated decrease in the frequency of mIPSCs was blocked by H3R antagonist clobenpropit (control: 2.61 ± 0.55 Hz, clobenpropit: 2.56 ± 0.45 Hz, histamine + clobenpropit: 2.85 ± 0.65 Hz, washout: 2.76 ± 0.30 Hz, n = 6, P = 0.68; one-way repeated-measures ANOVA; Fig. 6E,F). These electrophysiological results suggest that the histamine-induced decrease in spontaneous GABA release is mediated by presynaptic H3Rs.
To better understand the role of H3Rs in regulating GABAergic transmission, we analyzed the receptor's colocalization with the well-known marker of GABAergic terminals, GABA-synthesizing enzyme glutamic acid decarboxylase (67 kDa form, GAD-67). In the superficial layers of the MEC, double labeling for H3Rs and GAD-67 revealed widespread colocalization (Supplementary Fig. 6A1–A3). At higher magnification, the colocalization was always observed on the inhibitory synaptic terminals that contact with the putative principal projection neuron bodies and the GABAergic neuron soma (Fig. 6G1–G3, H1–H3). Instead, the expression level of H3Rs in the deep layers of the MEC was relatively low compared with that in the superficial layers (Supplementary Fig. 6A1–A3). Double staining with H3Rs/GAD-67 showed scarce colocalization in the deep layers (Supplementary Fig. 6B1–B3 and C1–C3). GAD-67 mainly labeled inhibitory terminals that rarely expressed H3Rs. Overall, these morphological results combined with the electrophysiological data imply that histamine presynaptically suppresses GABA release in the superficial layers of the MEC by stimulating the H3Rs.
Recent studies suggest that spontaneous synaptic release depends on Ca2+ entry via voltage-gated Ca2+ channels (VGCCs) (Daw et al. 2009; Groffen et al. 2010). We therefore tested whether the histamine-elicited decrease in mIPSC frequency was VGCC dependent. Bath application of Cd2+ (100 μm), a general VGCC blocker, alone induced a rightward shift of the interevent interval distribution of mIPSCs (P < 0.001, K–S test, Fig. 7A) and significantly decreased the frequency to 57 ± 6% of control (control: 2.37 ± 0.22 Hz, Cd2+: 1.37 ± 0.15 Hz, histamine + Cd2+: 1.30 ± 0.25 Hz, n = 7, control vs. Cd2+: P < 0.01; LSD post hoc comparisons following one-way repeated-measures ANOVA; Fig. 7A). These results imply that activation of VGCCs contributes to the generation of mIPSCs. After at least 6-min pretreatment of Cd2+, application of histamine failed to alter the frequency of mIPSCs (n = 7, P = 0.80; LSD post hoc comparisons following one-way repeated-measures ANOVA; Fig. 7A), indicating that the inhibitory effect of histamine on GABA release is mediated by VGCCs. Moreover, application of histamine did not induce a significant change in the frequency of mIPSCs in Ca2+-free external solution (control: 2.47 ± 0.55 Hz, Ca2+-free: 1.46 ± 0.37 Hz, histamine + Ca2+-free: 1.30 ± 0.27 Hz, n = 5, Ca2+-free vs. histamine + Ca2+-free: P = 0.65; LSD post hoc comparisons following one-way repeated-measures ANOVA; Fig. 7B). Together, these results suggest that H3R-induced inhibition of spontaneous GABA release depends on extracellular Ca2+ passing through the VGCCs.
Histamine Increases the Sensitivities of Stellate Neurons to a Depolarizing Ramp-Like Current Stimulation
Given that histamine inhibits Kir channels and GABA release in the superficial layers, we asked whether these effects affect the firing properties and sensitivity of stellate neurons to a depolarizing ramp-like current stimulation (Fig. 8A), consisting of a 600-ms ramp (−150 to 150 pA, slope of 0.5 nA/s) followed by a long plateau of current (150 pA, 4400 ms). The initial membrane potentials of neurons were set at the similar level before and during the application of histamine. Application of histamine (10 μm) remarkably shortened the first-spike latency from 582.6 ± 18.8 to 524.1 ± 19.3 ms (n = 9; P < 0.01; paired t-test; Fig. 8A,B and D). Moreover, histamine increased the instantaneous firing rate (Fig. 8A,B) of the stellate neurons from 13.3 ± 2.5 spikes/s to 16.2 ± 3.0 spikes/s (n = 5; P < 0.05; paired t-test; Fig. 8C) during 150 pA current injection, implying an increase in the neuronal sensitivity to the current stimulation. However, histamine did not affect (control: 7.3 ± 2.1 spikes/s; histamine: 7.7 ± 3.5 spikes/s, n = 5; P = 0.87; paired t-test; Fig. 8E) the difference between the instantaneous firing rate at the end of the ramp and the stable discharge rate at the end of the plateau (firing frequency adaptation in spikes/s; Fig. 8E). Thus, histamine did not change the firing frequency adaptation. The increase in both the firing rate and sensitivity of stellate neurons implies that the histaminergic afferent system in the superficial layers of the MEC may have a physiological function.
Blockade of H1Rs or H3Rs in the Superficial, but Not Deep, Layers of the MEC Impairs Rat Spatial Learning in Morris Water Maze
Because of the key role of the MEC in spatial navigation, we asked whether the excitatory effect of histamine on the principal neurons would influence spatial learning. In this set of experiments, we first implanted cannula into the superficial layers of the MEC and tested effects of histamine receptor antagonists in this region on rat spatial learning in Morris water maze. Rats receiving sham-operation (n = 14) or intrasuperficial layer infusion of ACSF (n = 14), triprolidine (n = 9), ranitidine (n = 14), and clobenpropit (n = 11) showed a reduction in escape latencies to find the hidden platform throughout the trials in water maze tasks as two-way repeated-measures ANOVA revealed a significant main effect of trial (F11, 627 = 19.67; P < 0.001; Fig. 9A) and no groups-by-trials interaction (F44, 627 = 1.33; P = 0.08; Fig. 9A). Analysis of the latency traveled to the escape platform during training revealed a significant main effect of group (F4, 57 = 7.61; P < 0.001; two-way repeated-measures ANOVA; Fig. 9A). LSD post hoc comparisons revealed that the rats receiving intrasuperficial layer infusion of triprolidine or clobenpropit had longer escape latencies than the ACSF-treated group (triprolidine: P < 0.001; clobenpropit: P < 0.01), indicating that blockade of either H1Rs or H3Rs in the superficial layers of the MEC impairs rat spatial learning. Conversely, microinjection of the H2R antagonist ranitidine failed to alter the escape latencies compared with the ACSF-treated group (P = 0.86). In the probe trials, rats treated with ACSF and ranitidine showed a significant bias for the target quadrant where the platform had been originally located (P < 0.05 for each group compared with the chance level 25%, one-sample t-test, Fig 9C). However, rats receiving triprolidine or clobenpropit failed to show preference for the target quadrant (triprolidine: P = 0.08; clobenpropit: P = 0.18; one-sample t-test; Fig. 9C). Together, these results suggest that the activation of H1Rs and H3Rs by endogenous histamine in the superficial layers of the MEC contributes to spatial learning.
Nonspatial learning factors, such as the motivational, emotional, and motor functions of the tested subjects, have been reported to influence the water maze performance. Thus, we took the following measures to ensure that the effects of drugs were not generated by nonspatial learning factors. First, we analyzed the recorded swimming speed of the rats. Intrasuperficial layer application of the drugs did not alter the swimming speed of the rats compared with sham-operated group (F4, 57 = 0.99; P = 0.42; two-way repeated-measures ANOVA; Supplementary Fig. 7A). Second, we conducted the open-field test. Both spontaneous locomotor (F4, 33 = 0.59; P = 0.68; one-way ANOVA; Fig. 9E) and exploratory activity (F4, 33 = 0.23; P = 0.92; one-way ANOVA; Fig. 9E) of rats in an open field were not affected by microinjecting ACSF (n = 7), triprolidine (n = 6), ranitidine (n = 6), and clobenpropit (n = 6) into the superficial layers of the MEC compared with sham-operated group (n = 9). The microinjection site was located within the superficial layers of the MEC (Fig. 9G). Thus, these results demonstrate that either triprolidine or clobenpropit specifically affects the spatial learning rather than as a secondary effect on a motivational, emotional, or motor functions.
We also microinjected histamine receptor antagonists into the deep layers of the MEC and investigated their effects on rat spatial learning. Sham-operated (n = 10) or intradeep layer infusion of ACSF (n = 9), triprolidine (n = 11), ranitidine (n = 12), and clobenpropit (n = 7) groups exhibited a significant trial effect (F11, 484 = 13.20; P < 0.001; two-way repeated-measures ANOVA; Fig. 9B), showing some improvement by both groups over the trials. However, there was no overall group difference in escape latencies to find the hidden platform (F4, 44 = 1.19; P = 0.33; two-way repeated-measures ANOVA; Fig. 9B) and no groups-by-trials interaction (F44, 484 = 0.60; P = 0.98; two-way repeated-measures ANOVA; Fig. 9B), reflecting the equivalent improvement of both groups across the acquisition stage. In the probe trials, sham-operated, ACSF-, and histamine receptor antagonist-treated groups exhibited a remarkable bias for the target quadrant where the platform had been originally located (P < 0.05 for each group, one-sample t-test, Fig 9D). The placement of cannulation was confirmed at the end of the behavior study (Fig. 9H). Together, these results indicate that central histaminergic system in the deep layers of the MEC does not impact the learning phase, which is consistent with the electrophysiological results. Similar to the findings in the superficial layers, intradeep layer infusion of the drugs did not affect the swimming speed of the rats (F4, 44 = 1.38; P = 0.26; two-way repeated-measures ANOVA; Supplementary Fig. 7B). Also, the spontaneous locomotor (F4, 34 = 0.27; P = 0.90; one-way ANOVA; Fig. 9F) and exploratory activity (F4, 34 = 0.10; P = 0.98; one-way ANOVA; Fig. 9F) of rats in an open field were not affected by microinjecting ACSF (n = 6), triprolidine (n = 8), ranitidine (n = 6), and clobenpropit (n = 8) into the deep layers of the MEC compared with sham-operated group (n = 7).
Histamine exerts neuromodulatory roles and is implicated in the regulation of several pathological/physiological processes including spatial learning and memory (Masuoka et al. 2007; Haas et al. 2008; Alvarez 2009; Zlomuzica et al. 2009). In this study, our results demonstrate that the activation of H1Rs and H3Rs by histamine increases the excitability of principal neurons in the superficial layers, but not deep layers, of the MEC through direct and indirect mechanisms when recording at the somata. Moreover, pharmacological blockade of the H1Rs or H3Rs in the superficial, but not deep, layers of the MEC specifically impaired rat spatial performance in water maze. Therefore, these data provide novel mechanisms whereby histamine selectively modulates activity of superficial layer principal neurons that mediate the functional interaction between hippocampus and other cortical regions and thereby influences the neocortex–EC–hippocampus memory circuit.
Histaminergic neurons in the TMN are thought to play a crucial role in the maintenance of arousal state and cortical activation (Lin, Anaclet, et al. 2011; Lin, Sergeeva, et al. 2011). Pharmacologic inhibition or genetic knockout of histidine decarboxylase, the enzyme responsible for biosynthesis of histamine, inhibited the cortical activation and caused somnolence and exploratory behavioral deficits. In contrast, increased histamine level by suppression of its degradation can promote wakefulness (Monti 1993; Lin 2000; Parmentier et al. 2002). Since histaminergic varicosities have been shown to seldom make synapses (Takagi et al. 1986; Ellender et al. 2011) and currently no specific transporter for histamine has been identified (Haas and Panula 2003), histamine is thought to have widespread influence on neuronal activity possible via volume transmission. The EC, particularly its superficial layers, is innervated by densely histaminergic fibers (Steinbusch 1991; Panula et al. 1998). The present finding of excitatory effect for histamine in the MEC combined with previous anatomical results suggests that histamine signaling directly contributes to entorhinal arousal.
Mechanisms Underlying a Direct Effect of Histamine
Kir channels that govern the resting membrane potential are considered to be critically involved in the regulation of the neuronal excitability (Hibino et al. 2010). Our results reveal that histamine directly modulates the excitability of entorhinal neurons, particularly the stellate neurons, in the superficial layers by inhibiting Kir channels based on the following bodies of evidence. First, the net current induced by histamine had a reversal potential close to the K+ reversal potential. Second, the difference current obtained from the stellate neurons exhibited strong inward rectification, which is one of the most important features of Kir channels. Third, we have shown that the histamine-elicited depolarization is blocked by either Ba2+ or Cs+, both of which are most commonly used blockers for Kir channels (Hagiwara et al. 1976; Hagiwara et al. 1978; French and Shoukimas 1985) and insensitive to voltage-dependent potassium channels blockers TEA or 4AP. Whereas Ba2+ blocks some of two-pore-domain potassium channels that are also involved in controlling resting membrane potential (Deng et al. 2009; Xiao et al. 2009), the two-pore-domain potassium channels are unlikely to be the targets for histamine in the MEC based on the following 2 lines of evidence. First, the two-pore-domain potassium channels are insensitive to extracellular Cs+ whereas application of this classic K+ channel blocker abolished the histamine-induced depolarization. Second, the two-pore-domain potassium channel-mediated current does not exhibit strong inward rectification.
To the best of our knowledge, our results for the first time reveal that the observed effects of histamine on neurons are mediated primarily through the inhibition of Kir channels. Kir channels are shut down during membrane depolarization and open to accelerate repolarization (Nishida and MacKinnon 2002; Day et al. 2005; Hibino et al. 2010). Therefore, the inhibition of Kir channels by histamine can not only cause a depolarization of membrane potential and increase firing rate but also delay repolarization that consequently enhances sensitivity of stellate neurons to ramp-like current stimulation.
In the MEC, histamine-elicited depolarization was blocked by the potent and selective H1R, but not H2R or H3R, antagonist. Furthermore, immunochemical results in this study confirmed that there is a high density of H1Rs in the superficial, but not the deep, layers. Combined with the findings of low expression of H2Rs in the MEC, these results suggest that a greater number of H1R-mediated postsynaptic responses would be observed. The activation of H1Rs can excite neurons in cortex and hippocampus by the activation of PKC (Haas et al. 2008). Consistent with the previous findings, our results indicate that PKC is necessary for direct action of histamine in the MEC. Phosphorylation of the Ser residues in Kir2.1 and Kir2.3 by PKC has been showed to suppress the channel activity (Henry et al. 1996; Tang et al. 1999; Karle et al. 2002). Considering the fact that Kir2 mRNA are highly expressed in EC (Karschin et al. 1996), the most plausible mechanism for blockade of PKC to attenuate histamine-induced depolarization is the direct interaction between PKC and Kir channels. However, it still cannot rule out the possibility that PKC modulates Kir channels via an intermediary. Overall, the postsynaptic effect of 10 μm histamine in the superficial layers is primarily mediated by the H1Rs–PKC signaling pathway. Previous study found that histamine affected hippocampal excitatory postsynaptic potentials via an unknown mechanism when a slight shift of the pH in the acidic direction (Yanovsky et al. 1995). At a few occasions, we puffed 100 μm and 1 mm histamine onto slices to investigate its postsynaptic effects. Although the histamine was puffed and can be further diluted in the bath groove, the present study cannot exclude the possibility that the proton concentration might be slightly altered at these high concentrations, thus causing unspecific effect.
Mechanisms of Histamine-Induced Suppression of Presynaptic GABA Release
Our results showed that histamine suppresses GABAergic, but not glutamatergic, transmission in the superficial layers of the MEC by decreasing presynaptic GABA release without altering postsynaptic GABAA receptors because histamine only decreased the frequency without altering the amplitude of mIPSCs. Interestingly, the indirect action of histamine was also superficial layer-specific since application of histamine did not affect the synaptic release on to the pyramidal neurons in the deep layers. In the MEC, histamine-induced suppression of presynaptic GABA release was specifically blocked by the H3R antagonist. Furthermore, staining of immunofluorescence revealed that H3Rs were mainly expressed in the superficial layers of the MEC and preferentially colocalized with the GAD-67-labeled inhibitory terminals. Overall, these data indicate that histamine inhibited the GABAergic input to stellate neurons via H3Rs, which is in agreement with earlier reports indicating that H3Rs preferentially act as heteroreceptors at presynaptic terminals and constrain the release of various transmitters including the GABA (Garcia et al. 1997; Drutel et al. 2001; Bergquist et al. 2006; Ellender et al. 2011; Passani and Blandina 2011). To further strengthen our conclusions, future studies are needed to detect whether the H3Rs are specifically expressed at the axonal terminals of the superficial GABAergic neurons by using immunoelectron microscopy.
The presynaptic VGCCs are important targets for various neuromodulators to influence the synaptic release (Huang et al. 2011; Li et al. 2011). In these experiments, blockade of VGCCs or removing extracellular Ca2+ profoundly reduced the inhibiting effect of histamine on mIPSCs, implying that modulation of VGCCs by H3Rs is likely to contribute to the inhibition of spontaneous GABA release. Several studies in the MEC found that altered the release machinery at the presynaptic terminals is accompanied by a change in the activity-dependent and evoked synaptic release (Lei et al. 2007; Wang et al. 2012). Moreover, presynaptic H3Rs have been shown to mediate the histamine-induced reduction in activity-dependent and evoked GABA release in the striatum and ventrolateral preoptic nucleus (Ellender et al. 2011; Williams et al. 2014). Thus, it is reasonable to speculate that H3Rs expressed at the inhibitory presynaptic terminals in the MEC might also lead to reduce the activity-dependent or evoked GABA release.
Although H3Rs can suppress the inhibitory neurotransmission, it is noteworthy that H3Rs also act as autoreceptors to inhibit histamine release and synthesis and restrict the release of other arousal-promoting neurotransmitters (Lin, Sergeeva, et al. 2011; Passani and Blandina 2011). The final influence of H3Rs on brain state is governed by a balance of distinct actions across multiple brain regions. In previous behavior studies, oral administration of H3R antagonist primarily promoted wakefulness, whereas the H3R agonist increased cortical slow activity and slow-wave sleep (Lin et al. 1990; McLeod et al. 1998; Lin 2000). Thus, the H3Rs exhibit constitutive activity and exert a tight control over sleep-wake cycle, which have attracted a great scientific interest for the treatment of sleep-wake disorders (Lin, Sergeeva, et al. 2011; Passani and Blandina 2011).
A Role for Histamine Innervating Superficial Layers in Modulation of Neuronal Excitability and Spatial Learning
The MEC, instead of LEC, contains spatial information-related functional cells, including grid cells, head direction cells, and border cells (Fyhn et al. 2004; Sargolini et al. 2006; Solstad et al. 2008). These results suggest that MEC circuit not only passively serves as the interface to control the flow of information into and out of the hippocampus but also is a possible neural element aiding in the integration of information about location, direction, and distance and plays an active role in spatial cognition (Fyhn et al. 2004; Steffenach et al. 2005; Witter and Moser 2006; Moser et al. 2008; Van Cauter et al. 2013). According to a dual-stage model of memory formation, encoding and consolidation of spatial memory are achieved in wakefulness and sleep, respectively, and are independently operated in the superficial and deep layers of the MEC (Chrobak and Buzsaki 1994; Buzsaki 1996; Chrobak and Buzsaki 1998; Chrobak et al. 2000; Battaglia et al. 2011; Inostroza and Born 2013). Compared with the sleep state, the principal neurons in the superficial layers of the MEC exhibit higher activity during the attentive arousal and their discharges are synchronized to the theta-coupled gamma oscillation (Chrobak and Buzsaki 1994, 1998; Isomura et al. 2006; Hahn et al. 2012; Igarashi et al. 2014). This activity pattern of the superficial layers is correlated with that of the histaminergic system. The activity of histaminergic neurons and the central release of histamine vary in the different brain states, being lowest during sleep and awake immobility, moderate during active wakefulness, and highest during exploratory behavior (Vanni-Mercier et al. 2003; Chu et al. 2004; Takahashi et al. 2006). Moreover, histaminergic system exerts a strong control over the hippocampal theta oscillation (Masuoka et al. 2007; Hajos et al. 2008), which is known to be associated with a superficial layer-dependent memory encoding. Here, we found histamine modulates neuronal excitability and neurotransmission in a superficial layer-specific manner. Moreover, the blockade of the histamine receptors in the superficial, but not deep, layers caused impairment in rat learning phase, indicating that the endogenous histamine signaling is essential for the spatial information encoding. Despite histamine has no effects in the deep layers, dopamine, another arousal-promoting neurotransmitter (Young 2009; Qu et al. 2010), has been showed to reduce the excitability of Layer V pyramidal neurons in the EC (Rosenkranz and Johnston 2006). Thus, these results suggest that arousal-promoting systems differently regulate the neuronal activities in superficial and deep layers and thus might functionally separate the superficial and deep layers to operate memory encoding and consolidation across sleep-wakefulness cycle. Although we found negative effects of histamine in the deep layers and low expression of histamine receptors, it should be noted that deep layers of the MEC have H1R mRNA (Lintunen et al. 1998). It is possible that these receptors might express in deep layer neuron dendrites traveling toward superficial layers. As the electrophysiological recordings were conducted at the somata, we cannot rule out the possibility that histamine affects dendritic functions and future studies are needed.
The major function of the superficial layers of the MEC may be processing incoming visuospatial information and transmitting it onto the hippocampus (Chrobak et al. 2000). The excitability of principal projection neurons in the superficial layers of the MEC exerts a tight control over the spatial learning and memory. Inhibition of neuronal excitability by the activation of GABAA or GABAB receptors dramatically impaired the rat spatial learning, whereas the increased neuronal excitability by knockdown TREK-2 channels, the final targets of the GABAB receptors, improved the learning ability (Jerusalinsky et al. 1994; Deng et al. 2009). The activity of principal neurons in the superficial layers is tightly controlled by the local GABAergic neurons (Woodhall et al. 2005; Lei et al. 2007; Buetfering et al. 2014). During a spatial memory task, high levels of endogenous histamine reduce the presynaptic GABA release and thus might disinhibit the principal neurons in the superficial layers of the MEC. This effect combined with postsynaptic excitation largely enhances the sensitivity of principal neurons to excitatory inputs and alerts the superficial layers of the MEC, thus leading to potentiation of the connection between neocortex and hippocampus and facilitation of spatial information encoding within MEC-hippocampal network (Fig. 10). Conversely in sleep, low levels of histamine and other neuromodulators disconnect the hippocampus from neocortex, leaving only internal loops operative, and enforce a bias toward memory consolidation (Tononi and Cirelli 2014). Thus, the circadian variation of the histamine and other arousal-promoting neuromodulators might induce the systematic alternation between “connected” potentiation and depression and favor the spatial memory formation.
The stellate and pyramidal neurons in the superficial layers of the MEC target distinct subregions of the hippocampus and serve in several features of episodic-memory processing (Kesner et al. 2000; Nakazawa et al. 2004; Nakashiba et al. 2008; Suh et al. 2011; Kitamura et al. 2014). The present study revealed no significant difference in postsynaptic effects of histamine among these principal projection neurons. As these neurons form distinct memory circuits, histaminergic indistinguishable modulation thus might contribute different aspects of the spatial learning. The present study investigated the behavior effects of histamine signaling by using a pharmacological approach, which lacks specificity. To dissect circuit-specific function of histamine, in the future study knockout of the histamine receptors in neuron type- or layer-specific manner would be much helpful.
Being enlightened by electrophysiological findings, we tested the effects of histamine signaling on spatial learning in the MEC of adult rats. Due to the influence of development, electrophysiological effects observed in rats at P14-20 might not be exactly the same as they are during adulthood. In future studies, electrophysiological recording from older rats would be much more helpful. Anyhow, the present finding is that histaminergic modulation of the superficial layers of the MEC is required for the rat spatial learning. Currently, the effect of histamine system on spatial learning has been attributed to its influence on the neuronal excitability and synaptic plasticity in prefrontal cortex and hippocampus or interaction with other neuromodulatory systems (Bacciottini et al. 2002; Luo and Leung 2010; Tsujii et al. 2010). Thus, our results provide a novel mechanism to explain at least the roles of histamine in some physiological functions such as learning and memory. Patients with schizophrenia and Alzheimer's disease, in addition to displaying neuronal pathology and atrophy of the EC (Nasrallah et al. 1997; Joyal et al. 2002; Prasad et al. 2004; Jessen et al. 2006), often exhibit abnormalities of the histaminergic system (Kim et al. 2002; Iwabuchi et al. 2005; Arrang 2007; Ligneau et al. 2007; Yanai and Tashiro 2007). Thus, abnormal modulation of MEC physiology by histamine may underlie the cognitive dysfunction in these neurological disorders.
This work was funded by grants from the National Natural Science Foundation of China (No. 81071074).
We appreciate Prof. Xiaowei Chen in the Third Military Medical University, Chongqing, China, for useful advice and comments on this manuscript. Conflict of Interest: None declared.