Intermittent low-oxygen treatment may reduce Alzheimer's-related protein buildup by boosting cell cleanup in mice

Specific pathogen-free (SPF), male APP/PS1 (B6C3-Tg [APPswe, PSEN1de9] 85Dbo/J [005864]) and wild-type littermate control C57BL/6 J mice on a C57BL6/J background aged 6–8 months were purchased form Nanjing Junke Bioengineering Corporation, Ltd. (Nanjing, China; certification number SCXK 2020-0009). Animals were maintained at 23 ± 2 °C under 45–60% humidity and a standard 12/12-h light/dark cycle. For TFEB activation, 6-month-old mice were orally administered TFEB activator 1 (TA1, 10 mg/kg/day; MedChemExpress, HY-135825, CAS: 39777-61-2) or vehicle (corn oil) for 3 months [31]. During IHT, mice were placed in a 60 cm × 30 cm × 25 cm chamber with rapidly adjustable oxygen levels. In this therapy, the oxygen concentration in the chamber was initially reduced to 8% within 30 s and maintained for 8 min by introducing compressed nitrogen, followed by compressed oxygen to restore the oxygen concentration to 21% within 30 s and sustained for 8 min. A total of 10 cycles of this hypoxia–normoxia exposure was performed daily for 28 days between 09:00 and 11:00. The IHT protocol was modified from a procedure that has proven to be neuroprotective in 5 × FAD mice [33]. After IHT, the mice were euthanized and perfused with 0.9% saline via the left ventricle to remove the blood.

MWM tests for spatial learning-memory behavior were conducted as previously described by Zha et al. [34]. The maze was divided into four equal quadrants: the northeast, southeast, southwest, and northwest. Visual cues were posted on the laboratory walls around the maze to facilitate the spatial learning of the platform's location. After IHT, mice were placed in a 150-cm circular pool with water temperature maintained at 21 ± 1 °C. A circular escape platform (10 cm in diameter) was placed in the middle of the southwest quadrant at 1.5 cm below the water surface. The water was made opaque with a non-toxic, white pigment. The mice were allotted 60 s to locate the platform and were allowed to stay on it for 20 s. The mice were trained four times per day for 5 days. On the 6th day, the hidden platform was removed from the destination quadrant. The mice were then released in the northeast quadrant and allowed to swim freely for 60 s. Video recordings were used to analyze and record the swimming trajectory of each mouse, along with measurement of the time required by the mice to initially find the escape platform, time spent in the target quadrant, and the average swimming speed.

The OFT was used to assess the anxious behavior of the mice and was performed according to a prior study by Kraeuter et al. [35]. After IHT treatment, the mice were individually placed in a white polyvinyl chloride box (40 cm × 40 cm × 40 cm) and allowed to move freely for 5 min. An overhead camera was used to capture the movement of the test animal in the peripheral (15 cm from the wall) and central zones (25 cm × 25 cm). The number of entries and time spent in the central area by the mice were also recorded.

The EPM test was employed to measure anxiety-related behavior in the mice and was conducted as previously described [36]. Briefly, the mice were positioned at the junction of the four arms (two open and two closed arms) of a maze (35 cm × 5 cm, 50 cm from the ground) facing an open arm and allowed to freely explore the maze for 5 min. Their behavior was recorded using a video camera above the maze. In addition, the entries and time spent by the mice in the open and closed arms were recorded.

Brain sections (30 μm) of the euthanized mice were fixed using 4% neutral paraformaldehyde and stained with 1% tar violet. The samples were subsequently dehydrated using an ethanol gradient and transparent xylene. The samples were imaged using a Leica DM4000B microscope.

Primary microglia culture Based on previously described protocols [37–39], primary microglial cells were obtained from the cerebral cortices of 3-day-old C57BL/6 mice. After removing the meninges, cortical tissue was digested using 0.05% trypsin. The separated cells were then cultured in Dulbecco's modified Eagle's medium-F12 supplemented with 10% fetal bovine serum, 5 ng/ml of Granulocyte–macrophage colony-stimulating factor (STEMCELL Technologies, 78017), and penicillin/streptomycin (100 U/ml and 100 mg/ml, respectively) at 37 °C in a 5% CO₂ humidified incubator. After 10 days, the mixed cell population was dominated by astrocytes and formed a fused trophoblast. The microglia gradually proliferated and floated in the supernatant. Finally, the cells from the supernatant were harvested and seeded in 35-mm confocal dishes.

IHT for the primary microglial cells The cell cultures were placed in the same intermittent hypoxia chamber used for the experimental mice. The cells were exposed to 21% oxygen (8 min) and 8% oxygen (8 min) for 10 cycles.

Lyophilized Aβ1–42 (Anaspec, AS-20276) were dissolved in PBS and incubated overnight at 4 °C to form oligomers (oAβ) [40]. The microglial cells were treated with 1 μM of oAβ for 0, 3, 6, 9, 12, 18, and 24 h. The residual oAβ in the supernatant was detected via ELISA to evaluate the ability of the microglia to metabolize oAβ. Phagocytosis ability and autophagy flux were determined as described below.

TA1 (CAS 39777-61-2) purchased from MedChemExpress (HY-135825, 99.69% purity) was dissolved in DMSO to yield a final concentration of 25 mg/ml. It was further diluted tenfold to 2.5 mg/ml with corn oil for animal treatment. In the case of the mice treatment, 6-month-old APP/PS1 mice were administered TA1 orally at 10 mg/kg for 3 months. For cell treatment, TA1 was further diluted to 1 mM in DMSO and directly added to the culture medium to produce a working concentration of 1 μM.

Bafilomycin A1 (CAS 88899-55-2) was procured from MedChemExpress (HY-100558, 98.77% purity). The colocalization of Aβ and LAMP1 in PAM in vitro was examined by adding 100 nM of bafilomycin A1 to the culture medium at 1 h before cell fixation.

TFEB pathway activation were determined by transcriptome screening and qRT-PCR validation. After IHT treatment, total RNA was isolated with TRIzol reagent. The purified total RNA was then submitted to Gene Denovo Biotechnology Co. for RNA-seq. Furthermore, purified total RNA was reverse-transcribed using the HiScript III 1st Strand cDNA Synthesis Kit (Vazyme Biotech, R312-02), according to the manufacturer's instructions and qRT-PCR was performed via the AceQ qPCR SYBR Green Master Mix (Vazyme Biotech, Q141-02).

To further confirm the relationship between IHT and TFEB, primary microglia were transfected by lentivirus expressing mouse shTfeb (#1, #2 and #3) to silence TFEB expression. Aβ-555 endocytosis and degradation, and autophagy level were determined.

After 70% fusion of the trophoblast cells (approximately 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 70% in the harvested microglia. Lentivirus expressing LC3-GFP was used to observe autophagosomes. In addition, lentivirus expressing mouse shTfeb (#1 target sequence: CGGCAGTACTATGACTATGAT, 2# target sequence: GCGGCAGAAGAAAGACAATCA, and #3 target sequence: GGAGATGACTAACAAGCAGCT) was used to silence TFEB expression.

Confocal imaging Primary microglial cells were incubated with 1 μg/ml of oligomer Aβ1–42, HiLyte™ Fluor 555 (Aβ-555) (Anaspec, AS-60480-01) for 30 min, followed by fixation with 4% paraformaldehyde for endocytosis analysis. In the degradation assay, the microglia were further chased with fresh medium and incubated for 30 min, followed by immediate fixation with 4% paraformaldehyde. Finally, the microglial cells were counterstained with DAPI, and confocal imaging was performed (552 nm; Leica SP8). Positive signals within a single cell were estimated using FIJI Image J (National Institutes of Health).

Flow cytometry The harvested microglial cells were incubated directly with Aβ-555 for 30 min. The ratio of positive cells containing the HiLyte 555 signal was determined using flow cytometry.

Tissue sections (40 μm) and cultured cells were fixed with 4% paraformaldehyde. The samples were subsequently permeabilized with 0.5% Triton X-100. The samples were then blocked in 10% donkey serum, followed by incubation with the following antibodies: anti-MAP2 (Sigma-Aldrich, M4403), anti-Aβ1–16 (BioLegend, SIG-39300), anti-Iba1 (Abcam, ab5076), anti-TFEB (Proteintech, 13372-1-AP), anti-LC3A/B (Cell Signaling Technology, 12741S), anti-LAMP1 (Abcam, ab24170), and anti-synaptophysin (Cell Signaling Technology, 9020). The binding of the primary antibodies was visualized using Alexa Fluor 555-conjugated donkey anti-rabbit IgG (Thermo Fisher Scientific, A31572), Alexa Fluor 488-conjugated donkey anti-mouse IgG (Thermo Fisher Scientific, A21202), or Alexa Fluor 647-conjugated donkey anti-goat IgG (Abcam, ab150131). Finally, the samples were counterstained with DAPI, and confocal microscopy (Leica Thunder or Leica SP8 confocal microscope) was used to capture the fluorescence images.

Total RNA from the samples was isolated with a TRIzol reagent. Purified total RNA was reverse-transcribed using the HiScript III 1st Strand cDNA Synthesis Kit (Vazyme Biotech, R312-02), according to the manufacturer's instructions. The purified total RNA was then submitted to Gene Denovo Biotechnology Co. for RNA-seq. Furthermore, qRT-PCR was performed via the AceQ qPCR SYBR Green Master Mix (Vazyme Biotech, Q141-02).

All primers used for qRT-PCR were designed as follows.

Lamp1, forward: 5′-CAGCACTCTTTGAGGTGAAAAAC-3′, reverse: 5′-ACGATCTGAGAACCATTCGCA-3′.

Il1b, forward: 5′-TGCCACCTTTTGACAGTGATG-3′, reverse: 5′-TGATGTGCTGCTGCGAGATT-3′.

Tnf, forward: 5′-AAGCCTGTAGCCCACGTCGTA-3′, reverse: 5′-GGCACCACTAGTTGGTTGTCTTTG-3′.

Tgfb, forward: 5′-TGATACGCCTGAGTGGCTGTCT-3′, reverse: 5′-CACAAGAGCAGTGAGCGCTGAA-3′.

Wipi2, forward: 5′-TGCTGGTAGGAGCATCAGATGG-3′, reverse: 5′-TCACTGGTCGTCTCCATACTGC-3′.

Vps11, forward: 5′-ATCGGCAGTCTCTGGCTAATGC-3′, reverse: 5′-GGACCTTGATGGCTGTCTCTAC-3′.

Vps18, forward: 5′-AAGTGAGCCCAACCGTGTGGAA-3′, reverse: 5′-AAAGGACCTCGGTGCTACTCAG-3′.

Actb, forward: 5′-CATCCGTAAAGACCTCTATGCCAAC-3′, reverse: 5′-ATGGAGCCACCGATCCACA-3′.

The relative amount of gene expression was calculated using ΔCt values, where Ct represents the threshold cycle of PCR.

Soluble and insoluble proteins from the mice brains, supernatant, or cultured microglia were diluted to 21% using PBS to determine Aβ1–42 and Aβ1–40 levels via specific ELISA methods. The test procedures were performed according to the manufacturer's instructions for the Human Aβ1–40 ELISA Kit (Immunoway Biotechnology, KE1389) and Human Aβ1–42 ELISA Kit (Immunoway Biotechnology, KE1390). The optical density was measured at 450 nm. In addition, the standard curve was created using a four-parameter logistic curve fit with Origin 8.0 software (https://www.originlab.com↗).

The samples were lysed in RIPA buffer to obtain soluble proteins. The precipitates were further cleaved with RIPA buffer containing SDS and urea to obtain insoluble proteins. The protein concentration was determined via bicinchoninic acid assay. The proteins were then isolated and transferred to polyvinylidene fluoride membranes. Next, the membranes were incubated with primary antibodies, including anti-Aβ1–16 (BioLegend, SIG-39300), anti-TFEB (Proteintech, 13372-1-AP), anti-phospho-MAPK (Cell Signaling Technology, 4370), MAPK (Cell Signaling Technology, 4695), anti-p-Akt (Cell Signaling Technology, 4060 T), anti-Akt (Cell Signaling Technology, 4685 s), anti-ribosomal protein S6 (RPS6; Santa Cruz Biotechnology, sc-74459), anti-p-RPS6 (Cell Signaling Technology, 2211), anti-mTOR (Santa Cruz Biotechnology, sc-517464), anti-p-mTOR (Santa Cruz Biotechnology, sc-293133), anti-p62 (Proteintech, 18420-1-AP), anti-LC3A/B (Cell Signaling Technology, 12741S), anti-LAMP1 ([Santa Cruz Biotechnology, sc-20011] or [Abcam, ab24170]), and anti-β-actin (Sigma-Aldrich, A5316). The binding of the primary antibodies was visualized using HRP-conjugated secondary antibodies. Furthermore, grayscale analysis was conducted using FIJI ImageJ (National Institutes of Health).

GraphPad Prism version 8 (GraphPad) was used to analyze the data. A Student's t test and one- or two-way ANOVA followed by Dunnett's multiple comparisons were applied for statistical assessment. Data are presented as mean ± Standard Error of Measurement (SEM). The significance levels for all the graphs were as follows: *P < 0.05, **P < 0.01, and ***P < 0.001. P values with no statistical difference have been marked in the graph.

We administered IHT (8% O₂ for 8 min and 21% O₂ for 8 min, 10 cycles/day) in 8-month-old APP/PS1 and wild-type (WT) littermate control mice for 28 days (Fig. 1A). The Morris water maze (MWM) results revealed that the swimming trajectories to reach the platform in the training trials were longer in the APP/PS1 mice than in the WT mice, while IHT-treated APP/PS1 mice demonstrated a shorter swimming route to the target quadrant (Fig. 1B). In the probe trials, the WT mice remained in the target quadrant, whereas the APP/PS1 mice exhibited indiscriminate movement to all quadrants. In contrast, the IHT-treated APP/PS1 mice displayed extended dwelling in the target quadrant (Fig. 1D). Furthermore, the latency to reach the escape platform for the first time was significantly longer in the APP/PS1 mice than in the WT mice, whereas IHT significantly shortened the latency time in the APP/PS1 mice during the 5-day training duration (Fig. 1C). Consistent with these findings, the results of the probe trials on the 6^th day showed that the APP/PS1 mice had a shorter dwelling time in the target quadrant and a longer latency time than the WT mice, and these values were largely reversed by IHT (Fig. 1D, E). However, the swimming speed was similar across all mice groups, suggesting that IHT has little effect on the muscle (Fig. 1F). Moreover, the open field and elevated plus maze tests indicated that IHT did not improve the anxiety of the mice (Additional file 1: Fig. S1). All these results suggest that the learning and memory of 8-month-old APP/PS1 mice were significantly impaired and that IHT effectively improved the cognitive function of these mice.

We further investigated whether the behavioral improvement of the IHT-treated APP/PS1 mice was associated with changes in their neural structures. Our examination showed a significant loss of axons and dendrites in the brain of the APP/PS1 mice. Furthermore, IHT significantly blocked this neuronal loss in the hippocampal CA1 and CA3 and cortical regions of the APP/PS1 mice (Fig. 1G, H, Additional file 1: Fig. S2). In addition, compared with the WT mice, the APP/PS1 mice exhibited a significantly reduced number of brain neurons, which was reversed after IHT (Fig. 1I, J). Considering these findings, IHT may improve cognitive function by mitigating neuronal impairment in APP/PS1 mice.

To further confirm whether IHT-induced cognitive function improvement in the APP/PS1 mice was related to changes in Aβ pathology, we co-labeled Aβ plaques with the microglial marker Iba1 in mouse brain sections (Fig. 2A). Our findings revealed that IHT significantly reduced the number and area of Aβ plaques in the cortex and hippocampus (Fig. 2B–D), which was accompanied by fewer activated microglia (Fig. 2E). Furthermore, IHT significantly reduced the levels of Aβ1–40 and Aβ1–42 (Fig. 2F, G), the most common Aβ subtypes, suggesting that IHT effectively suppresses Aβ plaque deposition.

We additionally explored the effect of IHT on Aβ clearance by PAM. As demonstrated in Fig. 2H, I, IHT significantly upregulated lysosome formation in PAM, along with significantly increasing the colocalization of the lysosomes with Aβ (Fig. 2J), suggesting that IHT promotes the lysosomal clearance of Aβ in PAM. Furthermore, IHT alleviated neuroinflammation in the APP/PS1 mice, as evidenced by the reduced levels of pro-inflammatory factors IL-1β and TNF-α and upregulated anti-inflammatory factor TGF-β (Fig. 2K). Therefore, IHT enhances Aβ clearance by PAM and reduces Aβ plaque accumulation and neuroinflammation.

We hypothesized that the effect of IHT on AD pathology is mediated by the upregulation of the ALP. After IHT, lysosomal-associated membrane protein 1 (LAMP1) and LC3I/II levels in the hippocampus of APP/PS1 mice were significantly upregulated and that of the insoluble content of the autophagy substrate SQSTM1/p62 was decreased, suggesting that IHT upregulates the ALP (Fig. 3A, B). Moreover, considering that Aβ clearance in the brain depends primarily on microglial autophagy [6], we further examined the effects of IHT on autophagosome formation in PAM. As shown in Fig. 3C, andD, the punctate intensities of LC3 (a well-known autophagosome marker) increased significantly in PAM after IHT. The colocalization of LC3 with LAMP1 was also increased in PAM, suggesting enhanced autolysosome formation (Fig. 3C, E). Consistent with the autophagy-dependent degradation of Aβ in PAM, elevated Aβ levels and increased colocalization of Aβ with LC3 were observed in PAM after IHT (Fig. 3F–H). These findings suggest that IHT enhances Aβ clearance by stimulating the ALP in PAM.

To further validate that IHT modulates its effects via microglia, an in vitro PAM model was established to enable the direct exposure of microglia to prolonged oAβ activation. As shown in Fig. 3I, microglia quickly degraded oAβ in the medium within 3 h of oAβ exposure, but the subsequent clearance activity was significantly diminished, suggesting that prolonged oAβ stimulation might impair microglial autophagic degradation of oAβ. Furthermore, a single day of IHT treatment (10 cycles of 8% O₂ for 8 min and 21% O₂ for 8 min) dramatically enhanced the clearance of the remaining oAβ by the microglia (Fig. 3J). In addition, we observed that the internalization of Aβ-555 was significantly improved in the IHT-treated PAM (Fig. 3K, L), wherein IHT was demonstrated to accelerate Aβ clearance by PAM after 30 min of intracellular degradation, i.e., chase period (Fig. 3M, N). This finding indicates that IHT enhances cellular Aβ clearance by promoting endocytosis and intracellular Aβ degradation by PAM. Furthermore, IHT significantly increased the colocalization of Aβ-555 with LC3 (Fig. 3O, P) and LAMP1 (Fig. 3Q, R), suggesting that IHT can stimulate autophagic degradation of oAβ by directly acting on microglia.

To explore the mechanism by which IHT upregulates oAβ degradation in PAM, we screened the PAM cells via RNA sequencing (RNA-seq). As depicted in the heatmap analysis, IHT treatment of PAM in vitro significantly increased the expression of ALP-related proteins (Fig. 4A). In addition, the levels of pro-inflammatory factors were decreased and those of anti-inflammatory factors were increased in IHT-treated PAM in vitro (Fig. 4A). Similarly, the KEGG pathway analysis revealed that the autophagy and inflammation-related pathways were among the top upregulated pathways (Fig. 4B). In particular, the mRNA levels of autophagy/lysosome-related genes, such as Lamp1, Vps11, Wipi2, and Vps18, were upregulated in the IHT-treated PAM in vitro (Fig. 4C–F).

Thus, we hypothesized that the IHT-enhanced autophagy was related to TFEB activation. In support of this hypothesis, our in vitro and in vivo results showed that the nuclear expression of TFEB in PAM was significantly increased after IHT (Fig. 4G–J). Moreover, IHT upregulated the transcription activity of TFEB (Additional file 1: Figs. S3 and S6). In consistent, IHT significantly upregulated the expression of ALP-related genes in mice brains, including Lamp1, Vps11, Wipi2, and Vps18, which were also target genes of TFEB (Fig. 4K–N). Next, to address the mechanism of TFEB activation by IHT in PAM, we investigated the activities of AKT, MAPK, and mTOR signaling, which serve as the negative regulators of TFEB activation. Our results demonstrated that IHT inhibited the phosphorylation levels of AKT, MAPK, mTOR, and RPS6 in vivo (Additional file 1: Fig. S4A, B) and in vitro (Additional file 1: Fig. S4C, D), indicating that IHT enhanced TFEB activation in PAM by inhibiting AKT–MAPK–mTOR signaling. These results suggest that IHT-enhanced autophagic degradation of Aβ by PAM is associated with the upregulation of ALP-related proteins, which in turn is mediated by TFEB activation.

TA1, also named as Curcumin analog C1, was reported as a TFEB-specific activator by binding to the N-terminus of TFEB, promoting TFEB nuclear translocation and activating TFEB-mediated autophagy [41], which has no effect on the pathways regulating autophagy, including mTOR, MAPK1/ERK2 and MAPK3/ERK1. TA1 has been reported to reduce Aβ pathology and improve cognitive function by effectively activating TFEB in the brain of 3xTg AD model mice [31]. In our study, TA1induced TFEB nuclear translocation in primary microglial cells (Additional file 1: Fig. S5). Thus, we further examined whether TA1 can alleviate AD pathology by activating the TFEB pathway in the PAM of APP/PS1 mice. Our findings demonstrated that TA1 reduced Aβ plaques in the mouse brain (Fig. 5A, B) and Aβ load in the cortex (Fig. 5C). In addition, TA1 increased the transcription activity of TFEB (Additional file 1: Fig. S6) and nuclear translocation of TFEB in PAM (Fig. 5D), which was in line with TA1's established role in TFEB activation [41]. We also revealed that TA1 upregulated the colocalization of Aβ with LC3 as well as the colocalization of LC3 with LAMP1 in PAM (Fig. 5E–H), consistent with the observation of TFEB-enhanced autophagic degradation of Aβ.

Furthermore, we examined whether TA1 directly affects microglia. For this purpose, we co-treated microglia with TA1 and oAβ. Our results showed that TA1 upregulated the nuclear levels of TFEB (Additional file 1: Fig. S7A, B). In consistent, TA1 upregulated the transcription activity of TFEB (Additional file 1: Fig. S3) and its target genes Lamp1, Vps11, Wipi2, and Vps18 in PAM in vitro (Additional file 1: Fig. S7C–F). TA1 was also demonstrated to upregulate the colocalization of LC3 with LAMP1 (Fig. 5I, J) and enhance Aβ clearance by microglia (Fig. 5K), indicating that TFEB activation contributed to oAβ degradation. Furthermore, TA1 increased the colocalization ratio of Aβ-555 with both LC3 and LAMP1 in PAM in vitro (Fig. 5L–O). These findings suggested that TFEB activation by TA1 enhanced Aβ degradation via the upregulation of the ALP in PAM in vitro, supporting the in vivo results.

To determine the role of TFEB in the autophagic degradation of Aβ by microglia, we inhibited TFEB expression by infecting primary microglial cells with lentivirus expressing shTfeb (#1, #2, and #3) (Fig. 6A). Our results showed that the colocalization of LC3 with LAMP1 was significantly decreased after TFEB silencing (Fig. 6B, C). Furthermore, oAβ degradation by microglia was impaired after TFEB knockdown (Fig. 6D), indicating that TFEB is important for Aβ degradation by microglia.

To further investigate the involvement of TFEB in IHT-enhanced Aβ degradation by PAM, we silenced TFEB in IHT-treated PAM in vitro. We found that TFEB silencing (Fig. 6F) inhibited the IHT-associated improvement of oAβ clearance by microglia (Fig. 6E). Similar to the findings in the primary microglial cells, the absence of TFEB in IHT-treated PAM in vitro prevented the enhanced colocalization of LC3 with LAMP1 (Fig. 6G, H), suggesting that IHT upregulated the ALP via TFEB. Furthermore, TFEB interference restricted IHT-stimulated augmentation of Aβ-555 colocalization with both LC3 and LAMP1 in PAM in vitro (Fig. 6I–L). Thus, the effect of IHT on the ALP-dependent degradation of Aβ in PAM is mediated by TFEB signaling.

All the study protocols of 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).

The authors declare that they have no competing financial interests.

The authors confirm that the data supporting the findings of this study are available within the article and its Supplementary material. Raw data that support the findings of this study are available from the corresponding author, upon reasonable request.