Animals
AQP4 knockout mice were generated as described previously [17] (acc. No. CDB0758 K: http://www.cdb.riken.jp/arg/mutant%20mice%20list.html) and maintained by crossing with C57BL/6JJcl (CLEA Japan Inc. Tokyo, Japan). B6.Cg-Tg (APPSwFlLon,PSEN1*M146L*L286V)6799Vas/Mmjax (5xFAD) [40] were obtained from Mutant Mouse Resource & Research Centers. The mice were housed in polycarbonate cages (3–4 animals per cage) at 22–24 °C under a 12 h light /12 h dark cycle with food and water ad libitum. Numbers of mice used in this study are shown in Supplemental Table S1. We detected age-dependent reduction in body weight of 5xFAD mice (Supplemental Fig. S1a), which has also reported previously [23], and we did not see an effect of AQP4 deficiency on the reduction in body weight of 5xFAD aged around 8 month (Supplemental Fig. S1a).
Reverse transcription quantitative polymerase chain reaction (RT-qPCR)
Total RNA was extracted from cerebral hemispheres with Isogen (Nippon Gene Co., Ltd., Toyama, Japan). First-strand cDNAs were synthesized using SuperScript VILO Master Mix (Thermo Fisher Scientific, Waltham, MA). The qPCR analysis was performed using KOD SYBR qPCR mix (Toyobo) and the Applied Biosystems StepOne Real Time PCR system (Thermo Fisher Scientific). Primers used for the qPCR are listed in Supplemental Table S2.
Immunohistochemistry
Mouse brain slice sections were prepared as described previously [12] with slight modifications. Paraffin sections with thickness of 5 μm was subjected to immunostaining using a Lica Bond-Max automatic immunostainer (Leica Biosystems, Mount Waverley, VIC, Australia). Used antibodies are rabbit anti-beta Amyloid 1–42 antibody (mOC64, 1:1000, abcam, Cambridge, UK) rabbit anti-AQP4 (1:30000, Atlas Antibodies AB, Bromma, Sweden), rabbit anti-Glial fibrillary acidic protein (GFAP) (1:2000, abcam), and rabbit anti-Iba1 (1:2000, FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) antibodies. The number and size of plaques were analyzed using ImageJ ver. 2.
Clearing brains
We used tissue clearing protocol based on a protocol for Clear, Unobstructed Brain/Body Imaging Cocktails and Computational analysis (CUBIC) [29, 51]. Brains were removed from mice anaesthetized by inhalation administration of isoflurane followed by perfusion fixation with 4% paraformaldehyde (PFA). The brains were immersed in 4% PFA at 4 °C for 2 h and were cut sagittally along the midline. The cerebral hemispheres were further immersed in 4% PFA at 4 °C overnight and transferred to 50% CUBIC-L solution [51] to incubate at 37 °C for 6 h with gently shaking. Then the hemispheres were transferred to CUBIC-L [51] to incubate at 37 °C with gently shaking for 7 days, changing the solution every 2 days. For immunofluorescent staining, the hemispheres were washed with TBS-T for 1 h followed by washed twice with TBS containing 0.01% sodium azide for 2 h. Then the hemispheres were stained with Alexa Fluor 488-labeled monoclonal anti-β amyloid, 1–16 (6E10, 1:200, BioLegend, San Diego, CA) and Amylo-Glo RTD Amyloid Plaque Stain Reagent (1:100, Biosensis, Thebarton, Australia) in TBS containing 0.01% sodium azide, 0.5% Triton X-100, and 2.5% BSA at room temperature for 3 days. The stained hemisphares were washed three times with TBS containing 0.01% sodium azide and transferred to ScaleCUBIC-2 [29] to incubate at 37 °C for 2 days with gentle shaking. The cleared brains in 80% ScaleCUBIC-2 were observed with Lightsheet Z.1 (Carl Zeiss, Oberkochen, Germany). Obtained images were stitched with arivis Vision4D ver. 2.12.3 (arivis AG, Munich, Germany), and analyzed with Imaris ver. 9.1.2 (Biplane AG, Zurich. Switzerland).
One-hundred-μm sections were also cleared using the CUBIC method and stained with Amylo-Glo RTD Amyloid Plaque Stain Reagent, Cy3-conjugated monoclonal anti-GFAP (G-A-5, 1:200, Sigma-Aldrich corp. St. Louis, MO), Red Fluorochrome(635)-conjugated rabbit anti-Iba1 (1:200, FUJIFILM Wako Pure Chemical Corporation), and rabbit anti-AQP4 (1:200, Sigma-Aldrich corp.) followed by Alexa Fluor 488-labeled goat anti-rabbit IgG (1:200, Thermo Fisher Scientific). The sections were mounted on a slide using ProLong™ Glass Antifade Mountant (Thermo Fisher Scientific) and observed with LSM 710 laser scanning confocal microscope (Carl Zeiss).
Digital Vivarium
Vium Digital Smart Houses consist of standard individually ventilated cage slotted in Vium’s proprietary rack system. The Digital Vivarium test was described previously [35]. Briefly, Vium Digital Smart Houses are outfitted with sensors and a high-definition camera that enable continuous monitoring of animals. The Vium Digital Platform obtains and maintains a digital record of the data analytics on motion. This study used the validated Vium Motion (m/sec). Daytime motion (collected from 0600 to 1800 PDT), nighttime motion (collected from 1800 to 0600 PDT).
Measurement of Aβ
Extraction of soluble and insoluble Aβ was performed as described in Hashimoto et al. [13]. TBS and 70% formic acid fractions were used as soluble and insoluble Aβ fractions, respectively. Concentration of Aβ40 and Aβ42 was determined using Human/Rat βAmyloid(40) ELISA Kit Wako II and Human/Rat βAmyloid(42) ELISA Kit Wako (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan), respectively.
For in vivo measuring Aβ concentration in interstitial fluid, mice were anesthetized with isoflurane. A guide cannula (an outer diameter of 0.72 mm and an inner diameter of 0.64 mm; Eicom, Kyoto, Japan) was stereotactically inserted in the right hippocampus (AP: − 3.1 mm and LT: 2.5 mm from bregma, and V: 1.3 mm) [43] and cemented. A dummy cannula (Eicom) was placed in the guide cannula. After 3–5 days of recovery period, mice were subjected to microdialysis under the freely moving condition (AtmosLM system, Eicom). A 2-mm microdialysis probe (PEP-4-02, 2 mm, 1000 kDa molecular weight cut-off, Eicom) was inserted in the guide cannula and perfusion buffer (147 mM NaCl, 4 mM KCl, 2.3 mM CaCl2, 0.15% bovine serum albumin, filtrated with 0.2 μm syringe filter) was circulated by peristaltic pump (ERP-10, Eicom) and microsyringe pump (ESP-64, Eicom) at a flow rate of 10 μl/min for 3 h followed by 1 μl/min for 1 h. After 4 h of perfusion, samples were collected at 2 h/fraction in a fraction collector (EFC-96, Eicom) refrigerated at 4 °C. Collected samples were stored at − 80 °C until ELISA assay.
Electroencephalography
The animals were deeply anesthetized with a mixture of ketamine and xylazine (100 mg/kg and 10 mg/kg, respectively, i.p.) in accordance with guideline from Japanese Association for Laboratory Animal Medicine (JALAM), and then fixed to a stereotaxic apparatus (SM-15, Narishige Scientific Instrument, Tokyo, Japan). Their body temperature was maintained at 37 ± 0.5 °C using a heating pad (FHC-MO, Muromachi Kikai, Tokyo, Japan) during the surgical procedure. A longitudinal incision was made and the skull surface was exposed. The periosteum and blood were removed thoroughly. Two craniotomies with a diameter of 1 mm were prepared using a drill. The coordinates of the craniotomies were 0.5 mm posterior to Bregma, 2.5 mm lateral from the midline to the left for the somatosensory cortex, and 7.5 mm posterior to Bregma, 0.3 mm lateral from the midline to the left for the cerebellum [43]. In the individual craniotomies, Teflon-coated silver wire electrodes (786,500, A-M Systems, WA, USA) were implanted above the brain surface as an EEG electrode and a reference. The exposed skull was then covered with a dental cement (Super-Bond C&B, Sun Medical, Shiga, Japan). After completing surgical procedure, the mouse was returned to its homecage for recovery.
After recovery for approximately 1 week, EEG recording was performed. EEG signals were transferred through flexible cables (AWG32, Mogami Cable, Ca, USA) and amplified by 1000-fold, bandpass filtered between 1 and 1000 Hz (Model 3000, A-M Systems, WA, USA), and digitized at 1000 Hz using an analog to digital converter (cDAQ-9178, National Instruments, TX, USA).
EEG data were analyzed using custom-written programs in MATLAB (2017a, MathWorks, MA, USA). Epileptiform spikes were defined as sharp (< 50 ms) negative deflections with amplitudes exceeding twice the baseline EEG [37]. Number of the epileptiform large EEG waveform in each mouse were counted in all recording sessions.
Grip strength test
The forelimb grip strength was measured using grip strength meter (O’Hara & Co., Tokyo, Japan). The mice were lifted by their tails so that their forepaws could grasp a wire grid of the apparatus. The mice were gently pulled backward by the tail until they released the grip. The peak force was recorded in Newton (N).
T-maze test
Spontaneous alternation task was conducted to assess working memory, as previously described [48]. The mice were subjected to a session consisting of 10 trials. Each trial consisted of a forced choice followed by a free choice. In the forced-choice trial, the mice were forced to enter either the left or right arm of the T-shaped platform of the maze. After the 10-s period, the mouse could return to the starting compartment, and a free-choice trial was started. In the free-choice trial, the mice were allowed to choose one of the arms. The percentage of correct responses in which the mice entered the arm opposite to their choice in the forced-choice trial during the free-choice trial was calculated. Total distance traveled and latency were also measured. Data acquisition and analysis were performed automatically using ImageTM software, based on the public domain ImageJ program (http://rsb.info.nih.gov/ij/), was developed and modified by Tsuyoshi Miyakawa.
Statistical analysis
For Fig. 2, statistical analysis was performed using SPSS ver. 25 (IBM, Armonk, NY, USA). Data were analyzed using one-way ANOVA, followed by the Bonferroni method. For Figs. 3, 4, 5a, b, and 6; and Supplemental Fig. S5 statistical analysis was performed using JMP ver. 14.0.0 (SAS Institute Inc., Cary, NC, USA). Data were analyzed using two-sided student t-test for Figs. 4, 5a, and b; and one-way ANOVA followed by the Tukey-Kramer method for Figs. 3 and 6 and Supplemental Fig. S5. For Fig. 5c, d, e, and f, statistical analysis was conducted using Prism5 (GraphPad Software Inc., San Diego, CA, USA). Data were analyzed using two-way ANOVA. For Supplemental Fig. S1, statistical analysis was conducted using StatView (SAS Institute Inc.). Data were analyzed using one-way ANOVA (a, b), two-way repeated measures ANOVA (c, d).