Introduction

Hyperpolarization-activated currents (Ihs) have aroused continuing interest of physiologists since their initial discovery in the early 1980s [1]. The current-designated Ih flowing through hyperpolarization-activated cyclic-nucleotide-gated cation nonselective (HCN) channels plays a key role in the control of neuronal rhythmicity [2, 3]. These channels are also proposed to play a fundamental role in determining the neuronal resting membrane potential [4], dendritic integration [5], synaptic plasticity [6], neurotransmitter release [7], and the periodicity of network oscillations [8]. With the development of molecular biology, a family of four mammalian genes that encode the HCN channels (HCN 1–4) are identified, and two (HCN1 and 2) are substantially expressed in rodent hippocampus [9]. In contrast, HCN3 and HCN4 are expressed at very low levels in hippocampus [3]. It has been reported that HCN1 channels constrain learning and memory by regulating dendritic integration of distal synaptic inputs to pyramidal cells [5]; the deletion of HCN1 channels causes profound motor learning and memory deficits in swimming and rotarod tasks [10]. In our previous study, we have found that HCN1 messenger RNA (mRNA) is shown to be decreased significantly in the hippocampus and neocortex in the chronic ischemic rats [11]. Besides, HCN2 channel dysfunction has been implicated in neurological disorders such as ataxia [12] and absence epilepsy [13] as well as reduced inflammatory and neuropathic pain [14]. Double immunofluorescence revealed colocalization of immunoreactivity for HCN1 and HCN2 in distal dendrites of pyramidal cells in the hippocampus and neocortex [15]. However, whether HCN2 may be altered has not yet been adequately established in chronic cerebral hypoperfusion rats with possible features of vascular dementia.

HCN channels and γ-aminobutyric acid (GABA) receptors can mutually co-regulate neuronal excitability in many brain areas [16–19]. In the central nervous system, three main classes of GABA receptors exist and are termed GABAA, GABAB, and GABAC receptors [20]. The GABAA and GABAC are ionotropic receptors, belonging to the Cys-loop family of ligand-gated ion channels. GABAB receptors are members of the metabotropic receptor family; these receptors couple via G proteins (Gi/o) to interact with neuronal inwardly rectifying potassium and voltage-gated calcium channels, mediating slow synaptic inhibition by increasing potassium (GIRK) and decreasing calcium conductances. Atherton et al. have reported that HCN channels in subthalamic nucleus neurons are conformed to selectively counteract GABAA receptor-mediated inhibition arising from the globus pallidus [18]. It has been found that the genetic deletion of HCN increased the expression of α5GABAA receptors in cortical pyramidal neurons [16], but the expression of HCN1 protein is significantly reduced in the hippocampus of Gabrα5−/− mice (lack the α5 GABAA subunit gene) [17]. Studies have shown that activation of GABAB receptors can inhibit Ih in rat ventral tegmental neurons and substantia nigra zona compacta principal neurons [19, 21]. However, little is known about the co-regulation of GABAB receptors and HCN channels in the hippocampal CA1 area.

Cerebral ischemia can cause the imbalance of excitation—inhibition. Glutamatergic and GABAergic transmissions work by each counterbalancing the function of the other; enhancing GABAergic activity should balance excessive glutamatergic excitation, which is the pivotal event leading to cell death. Baclofen, a selective GABAB receptors agonist, has been shown to be neuroprotective in animal models of transient brain ischemia [22–27], middle cerebral artery occlusion (MCAO) [28], and oxygen-glucose deprivation (OGD) [24, 27, 29, 30]. However, the potential roles of GABAB receptors and the co-regulation of GABAB receptors and HCN channel in chronic cerebral hypoperfusion are poorly understood.

A brain-specific protein termed TPR containing Rab8b interacting protein (TRIP8b) interacts through a conserved tripeptide sequence in the COOH terminus of HCN channels [31], which is thought to play an important role in the HCN channels trafficking [32]. TRIP8b colocalizes with HCN channels in cortical and hippocampal pyramidal cells, with more than ten isoforms expressed in the brain [33]. Based on quantitative real-time PCR analysis of brain tissue, TRIP8b (1a-4) and TRIP8b (1a) represent the two most prominently expressed isoforms (account for ∼30–40 and ∼25–30 % of total TRIP8b mRNA, respectively), with TRIP8b (1b-2) expressed at somewhat lower levels (∼10–15 % of total TRIP8b mRNA) [34]. Functional coexpression of TRIP8b with HCN protein, either in native cells or heterologous systems, results in a strong up or downregulation of HCN channels in the plasma membrane. TRIP8b (1a) prevents mislocalization of HCN1 in the axons of CA1 pyramidal neurons [33], and its physiological functions in HCN2 remain unclear. Indeed, whereas TRIP8b (1a-4) strongly increases HCN1 surface expression [34, 35], this isoform decreases the surface expression of HCN2 [36]. A study has reported that TRIP8b (1b-2) produces a potent downregulation in the surface expression of HCN1 and HCN2 [31]. In the present study, we investigated the functional role of TRIP8b (1a-4) and TRIP8b (1b-2) in chronic cerebral hypoperfusion. Studies have reported that subunits of adaptor protein 2 (AP2) complex appeared to be the major TRIP8b-interacting proteins [37, 38]. AP2 is a member of the family of adaptor proteins (today, comprising AP1, AP2, AP3, and AP4) and is composed of four subunits, which are termed α, β2, μ2, and σ2, respectively [39]. A recent study has been reported that AP2 may function as two partially independent hemicomplexes (α and μ2 subunits) [40]. Adaptor-associated kinase 1 (AAK1), a member of the Prk/Ark family of serine/threonine kinases, can decrease AP2-stimulated transferrin internalization via phosphorylating the AP2 μ2 subunit [41]. Besides, activation of protein kinase A (PKA) promotes the stabilization of GABAB receptors, presumably by protecting the receptors from proteolysis at or near the cell surface [42]. Until now, the role of these regulatory factors in chronic cerebral hypoperfusion is unclear.

In the present study, a widely accepted model of chronic cerebral hypoperfusion induced by permanent occlusion of bilateral common carotid arteries (two-vessel occlusion, 2VO) in rats was used to evaluate whether and how HCN channels were pathologically altered in the hippocampal CA1 area. We also investigated the neuroprotective effects of baclofen and cross-talk networks between GABAB receptors and HCN channels in the hippocampus of 2VO rats.

Materials and Methods

Animals

Adult male Sprague–Dawley rats of clean grade (approval number SCXK(E)2010-0007, No. 00008512), aged 2–3 months, weighing 220–250 g, were provided from the Experimental Animal Center, Tongji Medical College, Huazhong University of Science and Technology. Animals were group-housed with free access to water and food with a 12-h light/dark cycle and a thermoregulated environment and adapted to these conditions for at least 7 days before experiments. All experiments were approved by the Review Committee for the Care and Use of Laboratory Animals of Tongji Medical College, Huazhong University of Science and Technology. All efforts were made to minimize both the suffering and number of animals used.

Experimental Design

The experimental design is detailed in Fig. 1a. The rats were anesthetized with chloral hydrate (350 mg/kg, intraperitoneal injection, i.p.). Bilateral common carotid arteries were exposed and freed from surrounding connective tissue through a midline cervical incision. Both common carotid arteries were ligated with 4-0 type surgical silk and cut between the double ligatures to induce cerebral ischemia. The same operation was carried out on the sham-operated rats except for ligation and cutting of the carotids. During the surgery, their body temperature was monitored and maintained at 37.5 ± 0.5 °C by means of a heating lamp until the rats recovered to thermal homeostasis. Two weeks after induction of hypoperfusion, the repeated drug treatment of 2VO animals was performed once daily at 20:00–21:00 during the last 21 days. Baclofen (Meryer Chemical Technology Co., Ltd, Shanghai, China) was dissolved in saline at concentration of 1.25 and 2.5 mg/ml [24–26]. Baclofen was administered by i.p. injection in a volume of 10 ml/kg. Sham-operated rats were given normal saline (NS), also in the same volume. Four weeks after induction of hypoperfusion, the Morris water maze (MWM) was utilized to evaluate the spatial learning and memory performances of rats. One day after the Morris water maze test, novel object recognition (NOR) test evaluated nonspatial working memory. Five weeks after 2VO, we recorded the field excitatory postsynaptic potential (fEPSP) and carried out biochemical studies as described below.

Fig. 1
figure 1

Activation of GABAB receptors rescued the learning and memory deficits induced by 2VO. a Experimental design. b The average speed of rats from the 1st day to the 5th day. Swimming speed of each group did not differ significantly (n = 10 in sham-operated and 2VO groups, n = 9 in baclofen-treated groups). c The escape latencies to find the hidden platform from the 1st day to the 5th day. Latency to find the platform of 2VO rats was longer than sham-operated rats; after treatment of sham-operated rats with baclofen at 12.5 and 25 mg/kg, no changes in escape latencies compared to normal saline treated rats were observed. However, 2VO rats given baclofen at 12.5 and 25 mg/kg showed learning latency similar to that of sham-operated rats (n = 10 in sham-operated and 2VO groups, n = 9 in baclofen-treated groups). d The percentage of quadrant dwell time in Q2 on the sixth training day. The sham-operated rats sojourned longer in Q2 than 2VO rats; treatment with baclofen at 12.5 and 25 mg/kg improved the 2VO-induced deficit of acquisition in the Morris water maze (n = 10 in sham-operated and 2VO groups, n = 9 in baclofen-treated groups). e Effects of baclofen on novel object recognition test. Rats with 2VO significantly reduced their discrimination ability; baclofen at 12.5 and 25 mg/kg significantly ameliorated 2VO-induced cognitive impairment; and treatment with baclofen at 12. and 25 mg/kg in sham-operated rats resulted in a significant decrease in discrimination index (n = 10 in sham-operated and 2VO groups, n = 9 in baclofen-treated groups). *P < 0.05 and **P < 0.01 vs sham-operated rats; •• P < 0.01 vs 2VO rats; # P <0.05 and ## P < 0.01 vs 2VO rats

Full size image

Water Maze Task

The swimming pool named Morris water maze was a circular water tank of 150 cm in diameter and 50-cm deep. It was filled to a depth of 21 cm with water at 23 ± 2 °C. A platform of 15 cm in diameter and 20 cm in height was placed inside the tank to hold the rats, the horizontal plane being 1.5 cm above the platform top. Many cues external to the maze were around the pool. Each rat received four trials everyday with an inter-trial interval of approximately 10 min. Animals were trained for five consecutive days. Latency to escape onto the hidden platform was recorded. On the 6th day, each rat was subjected to a 60-s probe trial, in which the platform was removed and quadrant dwell time in Q2 was recorded. Parameters of escape latency, swim speed provided information about the ability of learning and about the motor performance. The amount of time spent in the quadrant of the former platform position was one of the measures for the performance in spatial memory [43, 44]. We used a video camera linked to a computer-based image analyzer Morris water-maze tracking system MT-200 (Chengdu Technology and Market Co., Ltd, Chengdu, Sichuan province, China) to monitor swimming activities. Tests were carried out between 10:00 and 12:00 A.M.

Novel Object Recognition Test [45–47]

The apparatus was formed by a gray-colored Plywood box (50 × 50 × 90 cm) with a grid floor that could be easily cleaned. The apparatus was illuminated by a 40-W lamp suspended 120 cm above the box. The objects to be discriminated were in three different shapes and colors: polyvinyl chloride cylinders (blue, 6-cm height), glass cubes (pink, 6-cm side), and tennis balls (yellow, 6.5–6.8-cm diameter). The day before testing, rats were allowed to explore the box for 5 min without objects. On the day of the test, a session of two trials was given. The inter-trial interval was 60 min. In the first trial, two objects, cylinders and glass cubes, were presented in two opposite corners of the box, and the amount of time taken by each rat to complete 20 s of object exploration was recorded. Exploration was considered directing the nose at a distance ≤2 cm from the object and/or touching it with the nose. During the second trial, one of the familiar objects, cylinder, was replaced by a novel object, tennis ball, and the rats were left in the box for 5 min. The time spent for the exploration of the familiar (F) and the novel (N) object was recorded separately. In normal rats, the time spent to explore a novel object is significantly higher than that spent to explore a familiar one. A discrimination index that normalized the data excluding the individual variability was calculated (N − F / N + F) to compare different groups. Care was taken to avoid place preference and olfactory stimuli by randomly changing the position of the two objects (familiar and novel object) during the second trial and cleaning them carefully.

Electrophysiological Recordings In Vivo

Extracellular recordings were described in detail previously [48]. Rats in each group were anesthetized with urethane (1.5 g/kg, i.p.), and the body temperature was maintained at 37 °C via a constant temperature water cycling system. The head was mounted in a stereotaxic frame (SN-3, Narishige, Japan), and both the skin and fascia were retracted to expose the skull under sterile conditions. The tissue was kept moist with gauze moistened with sterile saline throughout the surgical procedures. The ground electrode was inserted into subcutaneous tissue of the hind limb, and the reference electrode was placed on the scalp. A concentric bipolar stimulating electrode was placed in the Schaffer collaterals, which arose from hippocampal CA3 neurons and terminated on more proximal CA1 dendrites in stratum radiatum (SR) (3.5–4.0 mm posterior to bregma, 3.1–3.5 mm lateral to midline, and 2.2–2.5 mm ventral to dura). Recording electrode was positioned under stereotactic guidance in the CA1 stratum radiatum (3.3–3.6 mm posterior to bregma, 2.2–2.5 mm lateral to midline, 1.8–2.2 mm ventral to dura). There was a 30-min rest period after electrodes insertion. Stimulation was achieved with pulses of 0.15-ms duration delivered at 0.3 Hz (30 s intervals); evoked field responses were acquired, amplified, monitored, and analyzed with RM6240BD biology signal processing system (Chengdu Instrument Factory, China). For recordings of the slope of the fEPSP, the stimulation intensity for the tests pulses was set to elicit a response for which the fEPSP slope was 50–60 % of the maximum response so as to allow increases or decreases of the fEPSP slope to be detected. Long-term potentiation (LTP) was induced by high-frequency stimulation (HFS, four trains of 100 Hz, 1-s stimulation separated by 5 min at a stimulus intensity that evoked a fEPSP of approximately 80 % of maximum response). After HFS, the slope of fEPSP was recorded for at least 60 min. The percentage of the ratio of absolute fEPSP slope to baseline value was used to represent the fEPSP slope level. It was defined as a successful induction of LTP if the amplitude of fEPSP change exceeded 20 % [49]. After the experiment, correct placement of the electrodes was confirmed by histology.

Immunohistochemistry

Rats in each group were deeply anesthetized. Following the perfusion of rats by intracardiac injection of 0.9 % saline solution, followed by 4 % paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB). After perfusion, brains were harvested and postfixed in PFA overnight. After fixation and dehydration with gradient ethanol, the tissues were embedded in paraffin. Five-micrometer sections were cut, and the slides were deparaffinized in xylene and then dehydrated by incubation in graded ethanol solutions. Slides were boiled in sodium citrate buffer for 20 min for antigen retrieval and then incubated sequentially with 3 % H2O2 in PBS for 30 min and 5 % bovine serum albumin (BSA) for 1 h. The primary antibodies used were rabbit anti-mouse NeuN (1:400, MAB377, Millipore). The slides were incubated with the indicated antibody overnight at 4 °C, followed by sequential incubation with the secondary antibody, DyLight 488 Affinipure Goat Anti-Mouse IgG (H + L) (E032210, EarthOx, LLC) for 2 h. The sections were analyzed with Olympus FluoView 1200 confocal microscope system (Olympus Corporation, Japan), and photomicrographs of representative fields were taken. Image analysis was then performed using analySIS software (analySIS 3.0; Soft Imaging System) [50].

Golgi Silver Staining

Rats in each group were deeply anesthetized with urethane and perfused intranscardially with 0.9 % saline solution. The brains were removed and stained by the modified Golgi-Cox method [51, 52], and then stored for 14 days in Golgi-Cox solution, followed by 3 days in 30 % sucrose solution. Coronal sections of 50-μm thickness of regions to be studied were obtained using a vibratome (Camden Instrument, MA752, Leicester, UK). Sections were then treated with ammonium hydroxide for 30 min, followed by 30 min in Kodak Film Fixer (Eastman Kodak Company, Rochester, NY, USA) and finally rinsed with distilled water, dehydrated, and mounted with 90 % glycerol. Golgi-impregnated pyramidal neurons in the CA1 subfield of hippocampus were studied. Spine density calculation was performed as follows: a length of dendrite (at least ≥10-μm long) was traced (at ×1,000), the exact length of the dendritic segment was calculated, and the number of spines along the length counted (to yield spines/10 μm).

Western Blotting Analysis

Rats in each group were killed by decapitation under anesthesia. Brains rapidly removed and plunged into ice-cold artificial cerebrospinal fluid (ACSF). The cutting ACSF solution contained (mM) 125 NaCl, 3 KCl, 6 MgCl2, 1 CaCl2, 1.25 NaH2PO4, 25 NaHCO3, and 10.6 glucose, and was saturated with 95 % O2/5 % CO2. Coronal slices (400 μm) of brain tissue through the hippocampus were cut and subdissected into CA1, CA3, and dentate gyrus (DG) regions under a dissecting microscope (as shown in Fig. 2b) [53]. Samples were stored at −80 °C until required. Membrane protein extracts were prepared using a ProteoExtract Native Membrane Protein Extraction kit (71772–3, Calbiochem/Merck Biosciences). Protein concentration was determined using a BCA Protein Assay kit (Pierce). Equal amounts of protein samples (80 μg) were separated by 10 % SDS/PAGE gel electrophoresis and then transferred to Polyvinylidene difluoride (PVDF) membranes (03010040001, Roche). After blocking with 5 % nonfat milk powder in tris-buffered saline containing 0.1 % Tween-20 (TBST) for 1 h at room temperature, transferred membranes were incubated overnight at 4 °C with primary antibodies to anti-neuronal nuclear antigen (NeuN) (1:2,000, MAB377, Millipore), anti-GABAB R1 (1:1,000, AF7000, R&D systems), anti-GABAB R2 (1:200, MABN488, Millipore), anti-HCN1 (1:800, NBP1-20250, Novus), anti-HCN2 (1:200, APC-030, Alomone labs), anti-TRIP8b (1a-4) (1:500, NeuroMab clone N212/3), anti-TRIP8b (1b-2) (1:200, NeuroMab clone N212A/34), PKA α/β/γ (H-56) (1:100, sc-98951, Santa Cruz), anti-phospho-PKA (Thr197) (1:500, 4781, Cell Signaling), anti-AAK1 (H-140) (1:200, sc-134662, Santa Cruz), anti-adaptin 2 (M-16) (1:400, sc-6422, Santa Cruz), and p-APμ2 (pT156) (1:4,000, 3312–1, Epitomics) or GAPDH (1:5,000, cw0100, Cwbiotech). The antigen-antibody complexes were visualized with goat anti-rabbit, goat anti-mouse, mouse anti-goat or rabbit anti-sheep horseradish peroxidase (HRP)-conjugated secondary antibodies (1:5,000; Proteintech Group Inc, china) by using immobilon Western chemiluminescent HRP substrate (WBKLS0500, Millipore). The optical density of the bands was determined using NIH ImageJ software. The results were normalized to the quantity of GAPDH in each sample lane. All assays were performed at least three times.

Fig. 2
figure 2

Baclofen improved neuronal survival and dendritic spine density in rat hippocampal CA1 area after 5 weeks of 2VO. a Representative photomicrographs of immunohistochemical staining with an anti-NeuN antibody in the hippocampal CA1 (×400, scale bar, 100 μm), CA3 (×200, scale bar, 100 μm), and DG (×200, scale bar, 100 μm) areas and quantitative analysis of neuron density, which was performed by NeuN staining. Significant neuronal loss was observed in hippocampal CA1 area; however, the CA3 and DG exhibited no detectable cell loss (n = 4 in each group). b Schematic illustration of sectioning of hippocampus technique for regional analysis, which is based upon Fig. 3 of K. L. Powell et al. [53]. c Western blotting showed that baclofen treatment reversed the 2VO-induced reduction of NeuN protein in hippocampal CA1 area, but not in CA3 and DG area (n = 5 in each group). d Representative photomicrographs of immunohistochemical staining with an anti-NeuN antibody in the hippocampal CA1 of each group (×400, scale bar, 100 μm) and quantitative analysis of neuron density, which was performed by NeuN staining (n = 4 in each group). e Treatment with baclofen at 12.5 and 25 mg/kg markedly diminished the reduction of NeuN protein of hippocampal CA1 area (n = 5 in each group). f Representative photomicrographs of Golgi staining from each group. g Summary data showed that the decrease in the spine density induced by 2VO was partly reversed by baclofen (n = 4 in each group). **P < 0.01 vs sham-operated rats; • P < 0.05 and •• P < 0.01 vs 2VO rats

Full size image

Statistical Analysis

All analyses were performed using SPSS 16.0 software (SPSS Inc., USA) and data are presented as mean ± SEM. Differences between mean values were evaluated using one- or two-way analysis of variance (ANOVA), as appropriate. The t test is used for testing differences between two groups. P < 0.05 was considered statistically significant.

Results

Activation of GABAB receptors reversed the learning and memory deficits induced by 2VO.

To investigate the potential protective effects against chronic cerebral hypoperfusion injury, we first examined the effect of baclofen on the water maze task for five consecutive days. As shown in Fig. 1b, swimming speed of each group did not differ significantly. Latency to find the platform of 2VO rats was longer than sham-operated rats (Fig. 1c). It demonstrated that 2VO rats usually spent much more time to locate the platform than sham-operated rats, which found the platform in an appropriate path. After treatment of sham-operated rats with baclofen at 12.5 and 25 mg/kg, no changes in swim speed and escape latencies compared with normal saline-treated rats were observed. However, 2VO rats given baclofen at 12.5 and 25 mg/kg showed learning latency similar to that of sham-operated rats (Fig. 1c). On the sixth training day, the retention experiment with the platform removed was carried out. The percentage of quadrant dwell time in Q2 was adopted to evaluate the retention performance. The sham-operated rats sojourned longer in Q2 than 2VO rats (sham-operated rats 25.53 ± 1.08 %, 2VO rats 12.49 ± 1.03 %, Fig. 1d). Treatment with baclofen at 12.5 and 25 mg/kg improved the 2VO-induced deficit of acquisition in the Morris water maze (20.80 ± 2.65 and 23.38 ± 2.31 %, respectively, Fig. 1d).

Next, in order to evaluate nonspatial learning and memory, novel object recognition test was performed. Rats with 2VO significantly reduced their discrimination ability (discrimination index, −0.23 ± 0.06 vs. 0.44 ± 0.09 compared with sham-operated rats,); baclofen at 12.5 and 25 mg/kg significantly ameliorated 2VO-induced cognitive impairment (0.28 ± 0.05, 0.27 ± 0.13, respectively); and treatment with baclofen at 12.5 and 25 mg/kg in sham-operated rats resulted in a significant decrease in discrimination index (−0.19 ± 0.14, −0.29 ± 0.15, respectively) (Fig. 1e). Our results showed that even though the GABAB receptor agonist baclofen itself reduced discrimination ability of sham-operated rats, it improved the 2VO-induced reduction of exploratory preference to the novel object in 2VO rats.

Baclofen partially reversed the 2VO-induced inhibitory action on LTP at the Schaffer collateral-CA1 synapses in rats in vivo.

To determine the electrophysiological basis for the changes of cognitive functions, we recorded the LTP in Schaffer collateral synapses on CA1 pyramidal cells. Long-lasting significant enhancement appears in the slope of fEPSP induced by HFS (0–30 min after HFS, 204.78 ± 9.25 % of baseline values; 30–60 min, 179.43 ± 9.30 % of baseline values), and treatment with baclofen at 12.5 and 25 mg/kg in sham-operated rats had no significant effect on LTP (Fig. 3a–c). As shown in Fig. 3d, LTP was significantly inhibited in 2VO rats (0–30 min after HFS, 164.29 ± 14.32 % of baseline values; 30–60 min, 118.92 ± 7.31 % of baseline values). Administration of 12.5 and 25 mg/kg baclofen partially reversed the 2VO-induced inhibitory action on LTP (30–50 min after HFS, 210.38 ± 22.84 and 201.92 ± 10.00 % of baseline values, respectively; 50–80 min, 159.76 ± 18.09 and 170.17 ± 8.23 % of baseline values, respectively) (Fig. 3d–f).

Fig. 3
figure 3

Baclofen partially reversed the 2VO-induced inhibitory action on LTP at the Schaffer collateral-CA1 synapses in rats in vivo. a, d Representative fEPSP recorded before (solid line) and after HFS (dashed line). The downward-filled arrow indicates HFS. b. Treatment with baclofen at 12.5 and 25 mg/kg in sham-operated rats had no significant effect on LTP (n = 5 in each group). d LTP was significantly inhibited in 2VO rats at the Schaffer collateral-CA1 synapses; administration of 12.5 and 25 mg/kg baclofen partially reversed the 2VO-induced inhibitory action on LTP (n = 5 in each group). c, d. The change of average fEPSP slopes 30–50 and 50–80 min after HFS in different groups. *P < 0.05 and **P < 0.01 vs sham-operated rats; • P < 0.05 and •• P < 0.01 vs 2VO rats• P < 0.05 and •• P < 0.01 vs 2VO rats

Full size image

Baclofen improved neuronal survival and dendritic spine density in rat hippocampal CA1 area after 5 weeks of 2VO.

The most obvious signs of neurodegeneration are the loss of neuronal cell bodies and synaptic contacts. Immunohistochemical staining with an anti-NeuN antibody was performed to assess loss of neurons in hippocampus after 5 weeks of 2VO. Significant neuronal loss was observed in hippocampal CA1 area; however, the CA3 and DG exhibited no detectable cell loss (CA1, ×400; CA3, ×200; DG, ×200; Fig. 2a). Correspondingly, as shown in Fig. 2c, the protein levels of NeuN were significantly decreased in hippocampal CA1 area, but no significant change in hippocampal CA3 and DG areas. Treatment with baclofen at 12.5 and 25 mg/kg markedly diminished the neuronal loss of hippocampal CA1 area (Fig. 2d–e). Besides, our results revealed that the dendritic spine density was markedly decreased in 2VO rats compared with sham-operated rats, and baclofen reversed the reduction of spine density (Fig. 2f–g).

Chronic GABAB receptors agonist exposure results in a specific loss of cell surface GABAB receptors in the hippocampal CA1 area of sham-operated rats, but not in 2VO rats.

In this study, we identified the surface expression of GABAB R1 and GABAB R2 in the hippocampal CA1 area with Western blot analyses. No contamination with cytosolic protein was observed as GAPDH was not seen by Western blot in these samples (data not shown). Our results showed that, 5 weeks after induction of hypoperfusion, the surface expression of GABAB R1 and GABAB R2 was significantly decreased and intracellular expression of GABAB R1 was significantly increased; by contrast, intracellular expression of GABAB R2 was not significantly changed, and treatment of sham operated rats with baclofen accelerated the decrease in the population of surface GABAB R1 and GABAB R2 and the increase in the population of intracellular GABAB R1 and GABAB R2 (Fig. 4a–b). But, our data revealed that baclofen could attenuate 2VO-induced reduction of GABAB R1 and GABAB R2 surface expression (Fig. 4a–b). In order to further investigate possible mechanism for modulate GABAB receptor cell surface stability under chronic cerebral hypoperfusion, we examined the expression of PKA α/β/γ and p-PKA α/β/γ (the active form of PKA α/β/γ) in the hippocampal CA1 area. PKA α/β/γ levels remained unchanged among all treatment groups. Five weeks after induction of hypoperfusion, the expression of p-PKA α/β/γ was not significantly changed, but chronic treatment with baclofen significantly increased p-PKA α/β/γ (Fig. 4c).

Fig. 4
figure 4

Chronic GABAB receptors agonist exposure results in a specific loss of cell surface GABAB receptors in the hippocampal CA1 area of sham-operated rats, but not in 2VO rats. a, b. Five weeks after induction of hypoperfusion, the surface expression of GABAB R1 was significantly decreased, and intracellular expression of GABAB R1 was significantly increased; the surface expression of GABAB R1 and GABAB R2 was also significantly decreased, but intracellular expression of GABAB R2 was not significantly changed; baclofen could attenuate 2VO-induced reduction of membrane GABAB R1 and GABAB R2 expression; treatment of sham operated rats with baclofen dose-dependently accelerated the decrease in the population of surface GABAB receptors and the increase in the population of intracellular GABAB receptors (n = 5 in each group); c PKA α/β/γ levels remained unchanged among all treatment groups; and 5 weeks after induction of hypoperfusion, the expression of p-PKA α/β/γ was not significantly changed, but chronic treatment with baclofen significantly increased p-PKA α/β/γ (n = 5 in each group). *P < 0.05 and **P < 0.01 vs sham-operated rats; # P <0.05 and ## P < 0.01 vs 2VO rats; • P < 0.05 and •• P < 0.01 vs sham-operated rats

Full size image

Activation of GABAB receptors restored the balance of HCN1/HCN2 surface expression in the hippocampal CA1 area in 2VO rats.

Next, we evaluated the interaction between GABAB receptors and HCN channels under chronic cerebral hypoperfusion. We observed only a slight reduction of the surface expression of HCN1 and HCN2 in sham-operated rats which were treated with baclofen (25 mg/kg) (Fig. 5a); 5 weeks after induction of hypoperfusion, the surface expression of HCN1 was significantly reduced and intracellular expression of HCN1 was significantly increased in the hippocampal CA1 region (Fig. 5a); and however, hypoperfusion resulted in a significantly increase in HCN2 surface expression of hippocampal CA1 cells (Fig. 5b), but the intracellular HCN2 had no obvious change. Chronic treatment with baclofen significantly increased HCN1 surface expression, but reduced HCN2 surface expression; correspondingly, intracellular expression of HCN1 was reduced and that of HCN2 was increased (Fig. 5a–b). Immunoblotting of membrane protein extracts demonstrated no reactivity with anti-GAPDH antibodies, excluding the possibility of cross-contamination by cytoplasmic fractions (data not shown).

Fig. 5
figure 5

Activation of GABAB receptors restored the balance of HCN1/HCN2 surface expression in the hippocampal CA1 area in 2VO rats. a, b The surface expression of HCN1 was significantly reduced, and intracellular expression of HCN1 was significantly increased in the hippocampal CA1 region 5 weeks after induction of hypoperfusion; hypoperfusion significantly increased HCN2 in the surface of hippocampal CA1 cells but not in the intracellular of it; chronic treatment with baclofen significantly increased surface expression of HCN1, but reduced surface expression of HCN2; correspondingly, intracellular expression of HCN1 was reduced and that of HCN2 was increased; and treatment of sham-operated rats with baclofen had no effect on the surface and intracellular expression of HCN1 and HCN2 (n = 5 in each group). *P < 0.05 and **P < 0.01 vs sham-operated rats; # P <0.05 and ## P < 0.01 vs 2VO rats

Full size image

Activation of GABAB receptors could not only increase the expression of TRIP8b (1a-4) and TRIP8b (1b-2) but also regulate the function of TRIP8b via AAK1 and AP2 μ2.

We further explored the possible mechanisms of association between GABAB receptors and HCN channel in chronic cerebral hypoperfusion. The results showed that TRIP8b (1a-4) and TRIP8b (1b-2) protein levels were significantly reduced in the hippocampal CA1 region in 2VO rats. The expression of TRIP8b (1a-4) and TRIP8b (1b-2) were increased in baclofen-treated rats, especially in 2VO rats (Fig. 6a). AP2 α levels remained unchanged among all treatment groups (Fig. 6b); the expression of p-AP μ2 was not significantly changed in 2VO rats, but chronic treatment with baclofen significantly increased p-AP μ2 (Fig. 6b). Under chronic cerebral hypoperfusion, AAK1 levels were significantly reduced in the hippocampal CA1 region; baclofen restored the hippocampal AAK1 expression to the control level (Fig. 6c).

Fig. 6
figure 6

Activation of GABAB receptors could regulate not only the expression of TRIP8b (1a-4) and TRIP8b (1b-2) but also the function of TRIP8b via AAK1 and p-AP2 μ2. a TRIP8b (1a-4) and TRIP8b (1b-2) protein levels were significantly reduced in the hippocampal CA1 region in 2VO rats; the expression of TRIP8b (1a-4) and TRIP8b (1b-2) were increased in baclofen-treated rats, especially in 2VO rats (n = 5 in each group). b The expression of p-AP μ2 was not significantly changed in 2VO rats, but chronic treatment with baclofen significantly increased p-AP μ2; and AP2 α levels remained unchanged among all treatment groups (n = 6 in each group). c AAK1 levels was significantly reduced in the hippocampal CA1 region, and baclofen restored the hippocampal AAK1 expression to the control level (n = 5 in each group). *P < 0.05 and **P < 0.01 vs sham-operated rats; # P <0.05 and ## P < 0.01 vs 2VO rats; • P < 0.05 vs sham-operated rats

Full size image

Discussion

In the present study, we have demonstrated for the first time that activation of GABAB receptors has neuroprotective effects in chronic cerebral hypoperfusion by attenuating the decrease of surface expression of GABAB R1 and GABAB R2 and restoring the balance of HCN1/HCN2 surface expression.

Our results indicated that chronic administration of GABAB receptors agonist baclofen markedly alleviated the 2VO-induced deficits in the Morris water maze. Although some reports have demonstrated that baclofen injected 15–90 min prior to behavioral testing impairs spatial learning of normal animals [54–57], we did not observe any influence of chronic treatment with baclofen on the acquisition in the Morris water maze test in sham operated rats. We may suggest that the influence of baclofen on the acquisition process in the Morris maze depends also on the time of administration. In a nonspatial learning paradigm, our results showed that discrimination ability of SD rats was significantly reduced after 5 weeks of 2VO. It has been reported that Wistar rats performed as well as the controls after 30 days of 2VO in the object recognition test [58]. The reason for this inconsistent result may be that the process of memory impairments induced by 2VO varies according to the strain of rat involved [59]. In the present study, we found that baclofen could reduce the discrimination ability of sham-operated rats in the object recognition, which might be associated with the nitrergic system [60] and the decrease in the population of surface GABAB receptors as described below. Nevertheless, baclofen improved the 2VO-induced deficit of acquisition the object recognition test.

2VO-induced, permanent neuronal damage was correlated with the memory failure [61]. Besides, dendritic spines play a crucial role in information transmission [62]. The morphology of dendritic spines is an important component of synaptic plasticity which is vital in learning and memory. Changes in dendritic spines have been shown to occur with synaptic plasticity and cognitive function [63]. In this study, chronic administration of baclofen attenuated chronic hypoperfusion-induced neuronal damage and the reduction of spine density. Furthermore, baclofen also reversed 2VO-induced suppression of LTP at the Schaffer collateral-CA1 synapse.

Changes in the GABAB receptor expression and function during hypoxia and ischemia play an important role in central nervous system damage [26, 29, 64]. It has been reported that baclofen stimulation may provide some level of neuroprotection against OGD-induced GABAB R2 degradation [29]. Our data revealed that baclofen could attenuate 2VO-induced reduction of GABAB R1 and GABAB R2 surface expression. However, treatment of sham-operated rats with baclofen accelerated the decrease in the population of surface GABAB receptors, which was in agreement with previous results [42]. It has been reported that the number of cell surface GABAB receptors is modulated by chronic agonist exposure via a mechanism that is independent of agonist-mediated phosphorylation and arrestin recruitment, and activation of PKA promotes the stabilization of GABAB receptors, presumably by protecting the receptors from proteolysis at or near the cell surface [42]. Our results showed that treatment with baclofen could increase the levels of p-PKA (Thr-197) under chronic cerebral hypoperfusion, which was not observed in sham operated rats. Together, our observations indicated that baclofen might indirectly enhance PKA activity, and then attenuated 2VO-induced reduction of GABAB R1 and GABAB R2 surface expression.

In hippocampus and cortical pyramidal neurons of mice, GABAA receptors and HCN1 channels are mutually regulated [16, 17]. But we did not observe a similar phenomenon between GABAB receptors and HCN1/HCN2 channels in sham-operated rats: after treatment of sham operated rats with baclofen, the surface expression of GABAB R1 and GABAB R2 was significantly decreased; however, no significant changes in the surface expression of HCN1 and HCN2 were observed. Our results showed that, 5 weeks after induction of hypoperfusion, the surface expression of GABAB R1 and GABAB R2 was significantly decreased, accompanying with the reduction of HCN1 surface expression and increase in HCN2 surface expression. In hippocampal CA1 neurons, Ih is likely to be comprised of both HCN1 and HCN2 [65]. As described for other members of the voltage-gated cation channel family, four HCN subunits assemble to form a channel [66]. Homomeric HCN1 channels conduct fast kinetic currents with modest cAMP gating, consistent with currents recorded in hippocampal pyramidal cells where HCN1 expression is high [67]. By contrast, homomeric HCN2 channels conduct Ih currents with slower kinetics and robust cAMP-evoked shifts in voltage dependence [68]. In naive rat hippocampus, amounts of HCN1 protein are over eightfold higher than those of HCN2, so that most HCN1 channel molecules are likely to interact with other HCN1 subunits and form homomeric channels [69]. Chronic cerebral hypoperfusion-induced, an increased HCN2/HCN1 ratio should increase the stochastic probability of the HCN2 isoform’s interaction with HCN1, which might augment HCN1/HCN2 heteromerization [69]. Because the properties of heteromeric channels are distinct from those of homomeric ones, their contribution should significantly modify the properties of the neuronal Ih. The resulting changes in the Ih properties of individual neurons, a form of intrinsic neuronal plasticity, may contribute to long-term pathological alteration of the hippocampal network. Our results suggested that activation of GABAB receptors restored the balance of HCN1/HCN2 surface expression in the hippocampal CA1 area in 2VO rats.

Alternative mechanisms for hypoperfusion-driven enhanced HCN1/HCN2 heteromerization may involve activity-evoked changes in the intracellular trafficking. In this report, we found that TRIP8b (1a-4) and TRIP8b (1b-2) protein levels were significantly reduced in the hippocampal CA1 region in 2VO rats. Reduction of TRIP8b (1a-4) inhibited HCN1 channel trafficking to the plasma membrane, and reduction of TRIP8b (1b-2) inhibited HCN1 channel trafficking from the plasma membrane. Given that expression of TRIP8b (1a-4) is much more abundant than TRIP8b (1b-2), TRIP8b (1a-4) may play a more significant role in HCN1 channel trafficking than TRIP8b (1b-2) under chronic cerebral hypoperfusion. Thus, the reduction of HCN1 surface expression might be caused by the combined effect of a decrease in expression of both TRIP8b (1a-4) and TRIP8b (1b-2). Similarly, reduction of TRIP8b (1a-4) and TRIP8b (1b-2) inhibited HCN2 channel trafficking from the plasma membrane, which induced retention of HCN2 channels. During chronic cerebral hypoperfusion, a significant decrease in protein expression of AAK1 was not accompanied by a remarkable reduction of p-AP2 μ2, which might be attributed to enhanced activity AAK1. Our results indicated that the expression of TRIP8b (1a-4) and TRIP8b (1b-2) was increased in baclofen-treated rats, especially 2VO rats. Although treatment of sham-operated rats with baclofen caused slight but significant increase in TRIP8b (1a-4) and TRIP8b (1b-2), no significant changes in the surface expression of HCN1 and HCN2 were observed (as mentioned above). Possible causes include the following: the degree of increase in TRIP8b is not enough to make significant changes in the surface expression of HCN channels, and other hippocampal TRIP8b isoforms are also involved in the trafficking of HCN channels. However, during chronic cerebral hypoperfusion, activation of GABAB receptors could regulate the surface expression of HCN1 and HCN2 proteins via possible mechanisms as follows: increase in protein expression of TRIP8b (1a-4) increased HCN1 surface expression, but decreased HCN2 surface expression; TRIP8b (1b-2), which was also increased in baclofen treatment groups, decreased HCN1 and HCN2 surface expression simultaneously. In addition, baclofen could regulate the function of TRIP8b via AAK1 and p-AP2 μ2 under chronic cerebral hypoperfusion: baclofen restored the hippocampal AAK1 expression to the control level, and baclofen-induced enhancement of PKA activity (as mentioned previously) could further promote the action of AAK1 on phosphorylation of AP2 μ2 subunit [41, 70], which resulted in a decrease in TRIP8b (1b-2)-mediated HCN1 internalization. The combined effect of these on HCN1 channel was that activation of GABAB receptors attenuated the decrease of surface expression of HCN1 induced by chronic cerebral hypoperfusion. Coincidentally, increased phosphorylation of AP2 μ2 subunit might prevent HCN2 surface expression from downregulating exceedingly. Together, activation of GABAB receptors could not only increase the expression of TRIP8b (1a-4) and TRIP8b (1b-2) but also regulate the function of TRIP8b via AAK1 and p-AP2 μ2, which restored the balance of HCN1/HCN2 surface expression. Similarly, we also did not exclude the effects of other TRIP8b isoforms on the trafficking of HCN channels under chronic cerebral hypoperfusion.

Although the present study has yielded some preliminary findings, several limitations to this pilot study need to be acknowledged. Firstly, in order to eliminate the interactions with fluctuating female hormones that influence memory, we only used the male rat model, but it also limited the findings to a single sex. More studies clarifying gender differences in the drug treatment of chronic cerebral hypoperfusion-induced memory impairment are needed. Secondly, it remains to be seen precisely whether and how other HCN channels and TRIP8b isoforms were pathologically altered in chronic cerebral hypoperfusion. Thirdly, neuroprotective effects of GABAB receptors activation on chronic cerebral hypoperfusion should be further evaluated in nonhuman primates, which appear to represent the best model for the study of vascular dementia, because they have well-developed white matter and vascular architectures which closely resemble those in human brains [71].

In conclusion, our present results demonstrated that, under chronic cerebral hypoperfusion, GABAB receptor agonist baclofen attenuated the decrease of surface expression of GABAB R1 and GABAB R2 possibly by activation of PKA, and activation of GABAB receptors restored the balance of HCN1/HCN2 surface expression possibly via not only increasing the expression of TRIP8b (1a-4) and TRIP8b (1b-2) but also regulating the function of TRTP8b via AAK1 and p-AP2 μ2 (Fig. 7), which reversed the learning and memory deficits induced by 2VO.

Fig. 7
figure 7

A mechanistic explanation for the co-regulation of GABAB receptors and HCN channels under chronic cerebral hypoperfusion. Five weeks after 2VO, hypoperfusion significantly reduced surface expression of GABAB R1, GABAB R2, and HCN1, but increased HCN2 surface expression (red arrow). Baclofen attenuated 2VO-induced reduction of surface expression of GABAB R1 and GABAB R2 via activation of PKA (orange arrow). Besides, baclofen increased the expression of TRIP8b (1a-4) (blue arrow) and TRIP8b (1b-2) (green arrow) in 2VO rats. TRIP8b (1a-4) strongly increases HCN1 surface expression, and TRIP8b (1b-2) produces a potent downregulation in the surface expression of HCN1. Baclofen can enhance PKA activity and restored the hippocampal AAK1 expression to the control level under chronic cerebral hypoperfusion. Enhancement of PKA activity can promote the action of AAK1 on phosphorylation of AP2 μ2 subunit (black arrow), which results in a decrease in TRIP8b (1b-2)-mediated HCN1 internalization. The combined effect of these on HCN1 channel is that activation of GABAB receptors attenuated the decrease of surface expression of HCN1 induced by chronic cerebral hypoperfusion. TRIP8b (1b-2) produces a potent downregulation in the surface expression of HCN2; however, TRIP8b (1a-4) also decreases the surface expression of HCN2. Increased phosphorylation of AP2 μ2 subunit (black arrow) may prevent HCN2 surface expression from downregulating exceedingly

Full size image