Simvastatin impairs hippocampal synaptic plasticity and cognitive function in mice

All procedures have been approved by the Institutional Animal Use and Care Committee of School of Life Sciences, University of Science & Technology of China. Adult C57BL/6J male mice at 5 weeks of age were obtained from Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). After acclimating for a week, mice received administration of simvastatin (S.C., 30 mg/kg) or vehicle for 26 consecutive days. All behavioral tests were performed from Day 21 to Day 26. All mice were housed at 18–23 ℃ with 40–60% humidity under a 12-h dark/light cycle (lights off at 7 p.m) and free access to food and water.

After receiving 20 consecutive days (Day 1-Day 20) of vehicle/simvastatin treatments, the mice were arranged for the MWM tests (Day 21). Mice of each group were trained in a large tank (120 cm in diameter and 40 cm in depth) which was divided into four quadrants. A hidden 10-cm-diameter platform (1 cm below the surface of water) was placed in the center of a quadrant. The pool was surrounded by a black curtain with four visual cues on the wall of pool. Water was kept at 20° C and opacified with titanium dioxide. The trials were conducted 4 times daily at the same time point for 5 successive days followed by a probe test on Day 6. Mice were placed into four quadrants in order (20 min interval) and swam freely for a maximum of 60 s. If a mouse did not find the platform within a 60-s period, it was gently guided to the platform and allowed to stay on the platform for 15 s. The latency, distance and speed of mice to find platform were recorded. For probe test, the platform was removed from the pool and the mouse was put into the quadrant opposite to where the platform located and allowed to swim for 30 s. The time of the mice spent in each quadrant was recorded.

After receiving 20 consecutive days of vehicle/simvastatin treatments, another group of mice were arranged for the NOR tests (Day 21–22). The open-field apparatus consisted of an acrylic chamber (40 cm × 40 cm × 30 cm). Two different objects were prepared in duplicate: towers of rectangular Lego bricks (built from blue, green and yellow bricks) and circular Lego bricks (built from yellow and red bricks). The objects were placed 10 cm away from the walls and attached to the floor. Mice were tested in the dark (active phase between 7:00 p.m. and 7:00 a.m.). During the familiarization session, mice were allowed to freely explore two identical objects (rectangular Lego) placed into the arena at fixed locations for 3 min. The ANY-maze video-tracking system (Stoelting, Wood Dale, USA), which is based on nose-point detection, was used to record the time mice spent exploring objects. Active exploration was defined as mice sniffing or touching the object when the gap between the nose and the object was less than 2 cm. Climbing over the object or gnawing the object was not considered as exploratory activity. At the end of the test, each mouse was returned to its home cage, and the chamber and objects were cleaned using 75% ethanol, then air-dried for 3 min. The mice with no significant preference for the two identical objects were selected for further tests. In the NOR tests, 6 of 34 mice were excluded based on their abnormal preference to specific legos. After an intersession interval (ISI) of 24 h, one of the familiar objects was replaced by a novel object (circular Lego). The location of the novel object (left or right) was randomized among the mice and the groups tested. Object preference was calculated by using the following formula: preference % = (time to explore the individual object/total exploration time for both objects) × 100%. Data were excluded if the total of exploration time was less than 10 s. After the novel object recognition test, mice were allowed to recover for 2 days before further behavioral tests.

The open field test was performed 2 days after the NOR test (Day 24). An open field test system (XR-XZ301, Xinruan, Shanghai, China) was used. Mice were individually transferred from their home cages to an open field chamber (width, 45 cm; length, 45 cm; height, 45 cm) for locomotion tests for 15 min. Locomotor activity was recorded by a camera and the distance each mouse travelled was analyzed by the ANY-MAZE software (Global Biotech Inc.).

The rotarod test was performed on the next day after the OFT (Day 25). A rotarod training system (XR1514, Xinruan, Shanghai, China) was used. Before the first training session, mice were habituated to stay on a stationary rod for 2 min. A total of six trials for the rotarod test were carried out using an accelerating protocol from 4 to 60 rpm in 300 s with 20-min inter-trial intervals. After falling, the mice were immediately placed back to their home cages and the time to fall was automatically recorded by the rotarod software. Once the trial reached to 300 s, the mice were manually removed from the rod immediately. The apparatus and testing area were cleaned with 75% ethanol (w/v) after each trial.

The elevated plus maze was performed on the next day after RT (Day 26). The EPM apparatus consisted of a cross-shaped maze (with 25 cm × 5 cm arms) elevated by a 60-cm support. Two opposite arms were surrounded by a 20-cm wall, while the other two were open (only with a 1-cm contention step). Mice were individually placed in the central area of the apparatus, facing one of the closed arms, and their mobility within the maze was assessed over 5 min. The exploration profile within the different areas of the maze (open arms, closed arms and center) was analyzed. The anxiety behavior was assessed by examination of the open arm exploration. Animals that fell from the apparatus had to be censored from the analyses. Arm preference was automatically analyzed by the ANYmaze video tracking software.

The mice were sacrificed on the next day after all behavioral tests were finished (Day 27). Coronal hippocampal slices (350-μm thick) from adult male mice were prepared with Leica Vibratome in ice-cold cutting solution containing (in mM) 30 NaCl, 26 NaHCO₃, 10 Glucose, 194 sucrose, 4.5 KCl, 1.2 NaH₂PO₄, 1 MgCl₂ and continuously bubbled with carbogen (95% O₂ + 5% CO₂). The slices were then recovered at room temperature for 1 h. Slices were transferred into the recording chamber continuously perfused at 12 ml/min with artificial cerebrospinal fluid (ACSF) at 37 ℃. The constituent of ACSF are the followings: (in mM): 124 NaCl, 4.5 KCl, 1 MgCl₂, 2 CaCl₂, 1.2 NaH₂PO₄, and 26 NaHCO₃, continuously bubbled in carbogen. Long-term potentiation (LTP) was triggered by high frequency stimulations (HFS, 100 Hz, 1 s) in the hippocampal CA3 area. Field excitatory postsynaptic potentials (fEPSPs) were recorded using a glass electrode (filled with NaCl, 3–6 MΩ) placed into the stratum radiatum of the CA1 area. Signals were amplified (gain 100) and filtered (3 kHz), then digitized (10–100 kHz; National Instruments). After a 20-min baseline recording, recordings were continued for at least 50 min following LTP induction. The LTP was quantified by the fEPSP slope normalized to the baseline. Paired-pulse ratio (PPR) was obtained by delivering two stimulation pulses with interstimulus intervals of 50 ms. PPR values were quantified by calculating the ratio between the mean amplitude of the second and the first fEPSP. Synaptic responses were evoked at 0.1 Hz using a bipolar tungsten electrode. Data were collected and analyzed on or off-line by using pClamp 10.4 software (Molecular Devices, Sunnyvale, CA) software.

The mice used for paDESI-MS imaging also received vehicle/simvastatin treatments and behavioral tests except those in Fig. 3c and were then sacrificed on the next day after behavioral tests (Day 27). The brain was immediately removed from the skull and flash frozen in liquid nitrogen for 15 s. The frozen mouse brain was transferred to the cryostat chamber of a Vibratome (VT 1200S, Leica, Germany) at − 20 °C. Brains from vehicle group and simvastatin group were separately cut into 16-μm-thick coronal sections. In each group, three adjacent hippocampal slices were collected for parallel experiments. One slice from control group and one slice from simvastatin group were placed on the same microscope slide to avoid the matrix effects caused by different slides. The slide was then scanned by paDESI-MS. The cholesterol intensity was normalized to ¹³C3-cholesterol (0.1 mg/mL) which has been added into the spray. The major fragment of cholesterol is at m/z = 369.3532 [M − H₂O + H] + and the major fragment of ¹³C3-cholesterol is at m/z = 372.3628. Thus, we can examine the cholesterol intensity semi-quantitatively by normalizing the brain cholesterol to the signal intensity of [M − H₂O + H] + ions of ¹³C3-cholesterol. The changes in cholesterol were calculated as changes in cholesterol = ((C − Cmean)/Cmean)*100. C represents the normalized cholesterol level (normalized to the signal intensity of ¹³C3-cholesterol of the hippocampus and Cmean represents the mean of the normalized cholesterol levels in the hippocampus of vehicle-treated mice.

paDESI-MS imaging system consisted of a DESI sprayer, a 2D scanning stage, and a postphotoionization interface. A solvent was infused at a flow rate of 3 μL/min through a DESI sprayer (50 μm i.d. and 150 μm o.d. inner fused silica capillary and a 250 μm i.d. and 350 μm o.d. outer fused silica capillary) and directed onto the surface of a tissue slice with a 53° angle of incidence with the assistance of the nebulizing N₂ gas (120 psi). The flow of the solvent was driven by a syringe pump, and the metal needle tip was connected to a high-voltage power supply (3500 V for the positive ion mode and − 4000 V for the negative ion mode). The desorbed compounds were sucked in the heated transfer tube (i.d. 0.5 mm, o.d. 1/16in.) with a 10° angle of collection, and the un-ionized neutral molecules were ionized in an ionization tube (i.d. 4 mm, o.d. 10 mm) by a coaxially oriented krypton DC discharge vacuum ultraviolet (VUV) lamp, which was positioned to shine toward the exit of the transfer tube. Then the ionized species was transferred into a capillary of mass spectrometer. In order to improve the transfer efficiency, an air-flow assisted transport arrangement was added in this interface, and a pneumatic diaphragm pump (60 L/min, model GM-1.0A, Jinteng Experimental Equipment Co., Ltd., Tianjin, China) was connected to the side port of the ionization tube. In experiments, the transfer tube and ionization tube were kept at 300 °C. Note that the krypton lamp was turned off in the DESI mode and turned on in the DESI/PI mode. All imaging data were collected on an Agilent 6224 Accurate-Mass TOF mass spectrometer (Agilent, USA). The flow rate and temperature of drying gas of the mass spectrometer were set at 5 L/min and 325 °C, respectively. A programmable motorized X–Y scanning stage (GCD-203050 M, Daheng, Beijing, China) was used for tissue imaging, and the scanning process was allowed to be synchronized with the Agilent mass spectrometer data acquisition by the customized stage control software. The sample surface was line scanned in the X direction with a stepper motor at a velocity of 370 μm/s while acquiring mass spectra every 0.5 s. The distance between adjacent scan lines in the Y direction was 200 μm. The acquired multiple scan lines were combined in one data file for ion distribution images by using the freely available standalone version of the MSiReader software.

For simvastatin discontinuation experiments, the control group and simvastatin group received 26-day vehicle or simvastatin treatments and were then sacrificed on Day 27. The brain was then removed and frozen at − 80 °C for further MS imaging. The discontinuation group suffered 4-week simvastatin discontinuation after 26-day simvastatin treatments. After the discontinuation session, the mice were sacrificed and the brain was removed and frozen at − 80 °C. Brains from vehicle group, simvastatin group and simvastatin discontinuation group were separately cut into 16-μm-thick coronal sections. In each group, three adjacent hippocampal slices were collected for parallel experiments. One slice from control group, one slice from simvastatin group and one slice from simvastatin discontinuation group were placed on the same microscope slide to avoid the matrix effects caused by different slides. The slide was then scanned by paDESI-MS. The identifications for most of these peaks were facilitated by accurate m/z values, comparison of isotope distribution patterns, and tandem mass spectrometry.

All experiments and data analysis were conducted in a blinded way. All statistical analyses for in vitro recording and behavioral experiments were performed using Prism7 software (GraphPad). Data were statistically compared by unpaired t tests, as indicated in the specific figure legends. Average values are expressed as the mean ± SEM. P < 0.05 was considered significant.

First, we examined the LTP, a main form of synaptic plasticity that underlies synaptic information storage within the CNS [27], in the hippocampal slices of mice receiving chronic subcutaneous (S.C.) simvastatin administration (30 mg/kg/day, 26 days). Field excitatory postsynaptic potentials (fEPSPs) were recorded in CA1 area in response to the electrical stimulation of Schaffer commissural pathway (Fig. 1a). After setting of stimulating and recording electrodes into hippocampal CA3 and CA1, an input–output curve was constructed by stimulating at intensities ranging from 0 to 0.6 mA. Before LTP recording, we assessed the effects of simvastatin on presynaptic function of CA1 using a paired-pulse ratio (PPR) test. The results showed that simvastatin-treated mice showed a similar PPR compared with vehicle-treated mice, suggesting that the presynaptic release probability is unchanged (Fig. 1b, c). We then examined whether the basal synaptic field responses in the hippocampus were altered by simvastatin, by comparing input–output curves constructed from the stimulation intensity vs fEPSP slope. No significant differences between vehicle- and simvastatin-treated mice in the overall input–output curves were observed (Fig. 1d). These results suggest that long-term treatment of simvastatin does not affect the basal synaptic transmission. We next investigated whether simvastatin would affect synaptic plasticity induced by HFS. High frequency stimulation (HFS, 100 Hz, 1 s) was used to achieve LTP, before which a 20-min baseline recording was performed. The HFS-induced potentiation of fEPSP was significantly reduced in the simvastatin-treated mice when compared with the vehicle-treated mice (Fig. 1e, f). These results indicate that chronic simvastatin usage may impair the hippocampal synaptic plasticity.

We next conducted behavioral tests including novel object recognition (NOR) and Morris water maze (MWM) to examine the effects of simvastatin on the development of recognition and spatial memory, both greatly involving the hippocampal synaptic plasticity (Fig. 2a, g).

For the MWM test, mice were required to find a hidden platform to escape from swimming in a pool of water. The pool contained four quadrants and the mice were placed into four quadrants orderly (20-min interval) to swim freely for a maximum of 60 s. Four consecutive trials were conducted daily at the same time point for five successive days from Day 1 to Day 5. The simvastatin-treated mice showed an increased latency to find the platform compared with vehicle-treated mice on Day 5 (Fig. 2b, c). Additional probe trials demonstrated that simvastatin-treated mice spent less time in the target quadrant than the vehicle-treated mice (Fig. 2d). Similarly, simvastatin-treated mice also travelled a long distance compared with vehicle-treated mice on Day 5 (Fig. 2e). These results showed that long-term simvastatin treatments may cause deficiency in spatial memory. Such impairment seems to be independent of the swimming ability and sensitivity to water because the swimming speed was unchanged in simvastatin-treated mice (Fig. 2f).

For the NOR tests, the vehicle- and simvastatin-treated mice were adapted to the training room for 30 min. Then, the mice were allowed to freely explore two identical objects (rectangular lego) placed into the arena at fixed locations for 3 min. The mice with no significant preference for the two identical objects were selected for further tests. After an intersession interval (ISI) of 24 h, one of the original objects was replaced by a novel object (circular lego) and the object preference was calculated (Fig. 2h). The vehicle-treated mice spent more time exploring the novel object compared with the familiar object. Such preference to the novel object was significantly inhibited in the simvastatin-treated mice, indicating a deficiency in recognition memory (Fig. 2i). Such deficiency in memory is certainly not due to the preference of mice to the shape of lego itself (Fig. 2j).

We further examined the effects of simvastatin on other neurological behaviors. Simvastatin did not affect locomotor activity and motor coordination of mice, reflected by unchanged travel distance in the open field test and unaltered time to fall in the rotarod test (Fig. 2k, l). In the elevated plus maze test, time spent in the open and closed arms was not changed in the simvastatin treated mice compared with the vehicle-treated mice (Fig. 2m).

To examine whether long-term usage of the BBB-permeable simvastatin affects the hippocampal cholesterol level, we used our recently developed paDESI-MS imaging technique [26] to quantify the intensity of cholesterol in the hippocampus of mouse brain sections (Fig. 3a, b). The paDESI-MS technique combines conventional DESI with a postphotoionization. The advantage of this technology is that it enhances the ionization and imaging of desorbed neutral molecules such as cholesterol in biological tissue sections. Considering that it will take a long time for paDESI-MS to scan a whole brain slice and such a long time may cause degradation of metabolites, in this study we only screened and analyzed a small brain area containing the hippocampus (Fig. 3c). Long-term simvastatin administration significantly reduced brain cholesterol concentration in the hippocampus of mice. There was a strong correlation between hippocampal cholesterol intensities with the recognition memory (Fig. 3d) and the spatial memory of mice (Fig. 3e). Taken together, these results suggest that the simvastatin-induced synaptic plasticity impairment and cognition deficiency are correlated with the down-regulation of cholesterol level in hippocampus.

For investigating whether the neurological side effects of simvastatin are reversible, the medication was then weaned over a 4-week period in the simvastatin-treated mice. After that, the hippocampal cholesterol levels, LTP amplitude and the memory capacity were all re-examined in these mice. The hippocampal cholesterol concentration was restored to normal level testified by paDESI-MS imaging (Fig. 4a). Both the simvastatin-impaired recognition memory and spatial memory were significantly restored after simvastatin discontinuation (Fig. 4b–f). In addition, the LTP of fEPSP slopes in hippocampal CA1 slices were also recovered (Fig. 4g, h). These results suggest that the simvastatin-induced impairment of hippocampal cholesterol, synaptic plasticity and memory is transient and reversible.