Animals

APP/PS1 mice (B6C3-Tg[APPswe, PSEN1de9]85Dbo/J [005864]) and CX3CR1Cre − ER mice (021160, B6.129P2[Cg]-Cx3cr1tm2.1[cre/ERT2]Litt/WganJ) were purchased from Nanjing Junke Bioengineering Corporation, Ltd. (Nanjing, China; certification number: SCXK 2020-0009). VPS35fl/fl mice (C57BL/6 N-Vps35 tm1c[EUCOMM]Hmgu/Cmsu) constructed by CAM-SU GRC were provided by Professor Tong Liu from Nantong University. The VPS35fl/fl mice were crossed with the CX3CR1Cre − ER mice as previously reported [42], followed by crossing with the APP/PS1 mice to generate VPS35fl/fl: CX3CR1Cre − ER: APP/PS1 mice. At postnatal days 45, 95, and 150, the VPS35fl/fl: CX3CR1Cre − ER: APP/PS1 mice and their littermate control VPS35fl/fl: APP/PS1 mice were intraperitoneally injected with 100 mg/kg of tamoxifen (Sigma, #T5648) for 5 days. Mice showing a successful knockout of microglial VPS35 were designated as MG VPS35 KO: APP/PS1 mice and their littermate controls were denoted as VPS35 fl/fl:TG mice. Experimental animals were maintained at 23 ± 2 °C and 45–60% humidity with a standard 12/12-h light/dark cycle. All study protocols involving animal experiments were reviewed and approved by the Animal Care and Use Committee of Nantong University and the Jiangsu Province Animal Care Ethics Committee (approval ID: SYXK[SU]2007-0021).

Animal experimental design

IHT treatment. IHT treatment was performed as described previously [38]. Briefly, 8-month-old APP/PS1 (TG) mice or 6-month-old MG VPS35 KO: APP/PS1 were placed in a 60 cm × 30 cm × 25 cm chamber. The oxygen concentration in the chamber was initially reduced to 8% within 30 s and maintained for 8 min by introducing compressed nitrogen. Compressed oxygen was then introduced to restore the oxygen concentration to 20% within 30 s and sustained for 8 min. A total of 10 cycles of this hypoxia-normoxia exposure was conducted daily for 28 days between 09:00 a.m. and 11:00 a.m.

TFEB activator 1 (TA1) administration. TA1 (MedChemExpress, HY-135,825, CAS: 39777-61-2, 99.69% purity) was dissolved in DMSO to a concentration of 25 mg/ml and further diluted 10-fold with corn oil to 2.5 mg/ml for animal treatment. TA1 at 10 mg/kg/day or vehicle (corn oil) was orally administered to the 6-month-old APP/PS1 mice for 3 months [43].

Eltrombopag (EO) administration. EO (MedChemExpress, HY-15,306, CAS: 3 496775-61-2, 99.73% purity) was dissolved in DMSO to a concentration of 30 mg/ml, followed by 10-fold dilution with corn oil to 3 mg/ml for animal treatment. EO (30 mg/kg/day) or vehicle (corn oil) was injected intraperitoneally in the 8-month-old mice for 14 days [44].

After these treatments, the mice were euthanized and perfused with 0.9% saline via the left ventricle to remove the blood.

Morris water maze (MWM) test

MWM test for assessing spatial learning-memory behavior was conducted as previously described by Zha et al. [45]. In this test, a 150-cm circular pool was divided into four equal quadrants, namely the northeast, southeast, southwest, and northwest quadrants. Visual cues were placed on the walls around the maze to facilitate the spatial learning of the platform’s location. The water used in the maze was made opaque with a non-toxic, white pigment and maintained at a temperature of 21 ± 1 °C. During the training period (4 times/day for 5 days), a 10-cm diameter circular platform was placed in the middle of the southwest quadrant at 1.5 cm below the water surface. Individual mice were then allowed to swim freely for 60 s to find the platform and stay on it for 20 s. On the probe test day (the 6th day), the platform was removed. The mice were released in the northeast quadrant and allowed to swim freely for 60 s. Video recordings were utilized to analyze and record the swimming trajectory, along with measurement of escape latency (i.e., time to find the escape platform) of the mice and crossing frequency to the target quadrant.

Primary microglia culture and construction of Aβ-exposed microglia

Primary microglia culture. Based on earlier established protocols [46,47,48], primary microglia were obtained from the cerebral cortices of 2-day-old C57BL/6 mice. In this procedure, single-cell suspensions of brain tissue were prepared by digestion in 0.05% trypsin, followed by culturing in DMEM-F12 medium (Thermo Fisher, 11,320,033) containing 10% fetal bovine serum (Celligent, CG0430B), GlutaMAX supplement (Thermo Fisher, 35,050,079), 5 ng/ml of granulocyte-macrophage colony-stimulating factor (STEMCELL Technologies, 78,017), and penicillin/streptomycin (100 U/ml and 100 mg/ml, respectively) at 37 °C in a 5% CO2 humidified incubator. After 10 days of growth, the mixed cell population was dominated by astrocytes, forming a fused trophoblast. Microglial cells gradually proliferated and floated in the supernatant, and they were harvested on the 14th day.

Construction of Aβ-exposed microglia. Lyophilized Aβ1–42 (AnaSpec, AS-20,276) was dissolved in PBS and incubated overnight at 4 °C to form oligomers (oAβ) [49]. Finally, primary microglia or BV2 cells were treated with 1 µM of oAβ for 12 h.

Cell transfection and treatment

Lentivirus transfection in primary microglia. Microglial VPS35 was knocked out by transfecting lentivirus expressing Cre-GFP into trophoblast cells obtained from Vps35fl/fl mice. Additionally, TFEB was silenced by transfecting lentivirus expressing shTfeb (target sequence: GCGGCAGAAGAAAGACAATCA) into trophoblast cells isolated from C57BL/6 mice. After 7 days of primary culture, the cells were incubated with 8 × 107 TU of lentiviruses per 25 cm2 culture flask and allowed to grow until microglia production. The lentivirus transfection efficiency was approximately 50% in the harvested microglia [38].

Construction of shTfeb BV2 cells. BV2 cells were transfected with shTfeb-expressing lentivirus at multiplicity of infection = 10. The GFP+ cells were then isolated via flow cytometry and cultured to form single-cell clones.

IHT treatment. Cultured cells were placed in the same intermittent hypoxia chamber used for the experimental mice. The cells were exposed to 21% oxygen and 8% oxygen (30 s for the oxygen to jump between the two concentrations) for 8 min cycles. The viability of cells after IHT treatment was measured by a Cell Counting Kit-8 (CCK-8, HY-K0301, MedChemExpress).

TA1 treatment. TA1 was diluted to 1 mM in DMSO and directly added to the culture medium to yield a working concentration of 1 µM.

Oligomeric Aβ-555 endocytosis and TREM2 or TFR1 recycling assays

Aβ-555 endocytosis assay. To test the effect of IHT or TA1 treatment on Aβ-exposed microglia internalized oAβ, cells were incubated with 1 µg/ml oligomeric Aβ1–42, HiLyte™ Fluor 555 (Aβ-555) (AnaSpec, AS-60480-01) for 30 min and fixed with 4% paraformaldehyde, followed by counterstaining with DAPI. Since non-fluorescently labeled oAβ was used in the construction of the Aβ-exposed microglia, the red fluorescence observed under the microscope were all internalized oAβ in the cells within 30 min.

TREM2 or TFR1 internalization and recycling assays. After IHT treatment, primary microglia were incubated in DMEM-F12 medium containing sheep anti-TREM2 antibodies (R&D Systems, AF1729) or mouse anti-TFR1 antibodies (Thermo Fisher, 13-6800) for 1 h at 4 °C. The microglia were then washed with a pre-cooled DMEM-F12 medium to remove unbound antibodies. Subsequently, the cells were incubated with DMEM-F12 medium containing 10% fetal bovine serum for 30 min at 21% O2, 5% CO2, 37 °C to stimulate TREM2 or TFR1 internalization. Cells were then washed with wash solution (hydrochloric acid added to DMEM/F12 medium to reduce the pH to 2.0) for 5 s to remove surface antibodies, and then the wash solution was immediately removed with fresh DMEM/F12 medium for 2 min to restore the extracellular pH environment. In the internalization assay, the cells were directly fixed and labeled with anti-sheep Cy3 antibodies (Jackson ImmunoResearch, 713-165-174) or anti-mouse Alexa 488 antibodies (Jackson ImmunoResearch, 715-545-150) (Fig. 2A). In the case of the recycling assay, the cells were further incubated with DMEM-F12 medium containing 10% fetal bovine serum for 60 min at 21% O2, 5% CO2, 37 °C. The live cells were then probed by incubating in DMEM-F12 medium containing anti-sheep Cy3 antibodies for 1 h at 4 °C, followed by fixation (Fig. 2D).

Immunofluorescence staining

Brain tissue sections or cultured cells were fixed and subsequently permeabilized with 0.5% Triton X-100. Next, the samples were blocked in 10% donkey serum, followed by incubation with the primary antibodies at 4 °C overnight. Secondary antibodies were then employed to visualize the binding of the primary antibodies. Lastly, the samples were counterstained with DAPI, and confocal microscopy (Leica SP8 confocal microscope) was performed to capture the fluorescence images. The antibodies utilized for immunofluorescence staining are listed in Table 1. Individual Iba1-positive cells were selected and the average fluorescence intensity of VPS35, TREM2 and TFEB in the selected cells was counted using FIJI software. The data were subsequently normalized with Nor-TG.

Table 1 Antibody informationRNA isolation and qRT-PCR

Purified total RNA from the brain tissue or cells underwent reverse transcription using the HiScript III 1st Strand cDNA Synthesis Kit (Vazyme Biotech, R312-02). qRT-PCR was performed for a total of 40 cycles at 95 °C for 10 s, 60 °C for 30 s, and 72 °C for 20 s, according to the instructions in the AceQ qPCR SYBR Green Master Mix (Vazyme Biotech, Q141-02). The relative amount of gene expression was calculated using ΔCt values. The primers used for qRT-PCR are provided in Table 2.

Table 2 Primer informationProtein isolation and Western blotting

The samples were lysed in RIPA buffer, and their protein concentrations were determined via the bicinchoninic acid assay. The proteins in the remaining lysate were isolated and transferred to polyvinylidene fluoride membranes for primary antibody hybridization. Finally, HRP-conjugated secondary antibodies were utilized to visualize the binding of the primary antibodies.

Statistical analysis

The fluorescent signal intensity on microscopy imaging and the grayscale values of the Western blotting protein bands were calculated using FIJI software (National Institutes of Health). Data analyses were performed using GraphPad Prism version 8 software. Statistical assessment was conducted using Student’s t-test and two-way ANOVA, followed by Dunnett’s multiple comparisons test. Data were presented as mean ± SEM. The significance levels for all graphs were as follows: *p < 0.05, **p < 0.01, and ***p < 0.001. n.s. indicated no statistical difference.