Abstract
Background Tumor necrosis factor (TNF) receptor 1 (TNFR1) signaling mediates neuronal necroptosis in Alzheimer’s disease (AD). Interaction of TNFR1 signaling axis with autolysosomal pathway and the accumulation of necrosome molecules in impaired lysosomes have been shown to lead to necroptotic neuronal death. This has been attributed to the terminal failure of the autophagic process, primarily due to lysosomal degradation dysfunction. Being the final and determining step of the autolysosomal pathway, lysosomes with sufficient acidification as maintained by functional vacuolar (H+)-ATPase (V-ATPase) are required to achieve complete autophagic degradation of toxic cellular components. Here, we aim to investigate the role of defective lysosomal acidification in mediating TNFR1 induced neuronal necroptosis in AD.
Methods Neuropathological analysis of human post-mortem AD brains was performed to examine the correlation between TNFR1 induced neuronal necroptosis and autolysosomal dysfunction. Specifically, we probed for the level of V-ATPase subunits in AD brains to determine the extent of lysosomal acidification and function. Cell-based assays were conducted to understand the effect of TNFR1 activation in driving lysosomal acidification defect, autophagic impairment, mitochondrial dysfunction, and neuronal death in SH-SY5Y neuroblastoma cells. Furthermore, we applied lysosome-acidifying nanoparticles (AcNPs) to determine whether restoration of lysosomal acidification can rescue neuronal necroptosis in both TNF-treated SH-SY5Y cells and APPNL-G-F knock-in mouse model of AD.
Results We revealed that TNFR1 activated neuronal necroptosis correlates with autolysosomal dysfunction as characterized by downregulation of V-ATPase subunits and accumulation of autophagy receptor p62 in human AD brains. In cell culture, we showed for the first time that lysosomal acidification is only impaired in cells treated with TNF and not with other cytokines, contributing to inhibition of autophagic degradation in SH-SY5Y cells. We also illustrated that there is defective mitochondrial turnover, together with reduced mitochondrial functions and elevated reactive oxygen species, leading to neuronal death in SH-SY5Y cells. Importantly, we demonstrated that AcNPs restore lysosomal acidification, autophagic activity, and mitochondrial function, as well as rescue neuronal necroptosis in both TNF-treated SH-SY5Y cells and APPNL- G-F mice.
Conclusions Defective lysosomal acidification plays a key role in TNFR1 mediated neuronal necroptosis. This opens avenues for new therapeutic strategies to target lysosomal acidification dysfunction in AD.
Background
Alzheimer’s disease (AD) is neurological disorder characterized by sustained neuroinflammation leading to progressive neurodegeneration with clinical symptoms of cognitive decline and memory loss [1]. Alongside the classical pathological hallmarks of intracellular tau neurofibrillary tangles (NFTs) and extracellular β-amyloid (Aβ) plaques, increasing evidence suggests that neuroinflammation plays a prominent role in AD pathogenesis [2, 3]. This is supported by AD risk genes that are associated with immune response [4, 5] and the elevated levels of inflammatory cytokines in AD patients [6]. It has also been shown that neuroinflammation actively drives AD pathogenesis, rather than a mere bystander or consequence of insults [2, 3]. One of the key inflammatory mediators under neurodegenerative condition is the soluble tumor necrosis factor (TNF) that primarily acts through binding TNF receptor 1 (TNFR1) and activating downstream signaling, including cell death pathways [7]. There is also growing evidence that TNFR1 signaling is implicated in multiple neurodegenerative conditions [8], including AD [9], Parkinson’s disease [10], and multiple sclerosis [11].
The TNFR1 signaling pathway has been identified as a key regulator of both apoptosis and necroptosis. Binding of TNF to TNFR1 induces a conformational change in the receptor-ligand complex that enables the recruitment of TNFR1-associated death domain (TRADD) and receptor-interacting protein kinase 1 (RIPK1) and the formation of downstream complexes I and II [12, 13]. While complex I triggers the nuclear factor-κB (NF-κB) pathway responsible for inducing cytokine production and anti-apoptotic gene expression, complex II mediates cell death pathways including apoptosis and necroptosis, which is dependent on caspase-8 activity [14]. Caspase-8 function is regulated by its phosphorylation state, expression level, and interaction with the autolysosomal network or the ubiquitin-proteosome system [15]. In situations where caspase-8 is absent or inactive, RIPK1 is not cleaved, leading to the phosphorylation of RIPK1, RIPK3, and mixed-lineage kinase domain-like protein (MLKL). This results in the formation of phosphorylated MLKL (p-MLKL) oligomers, initiating the process of necroptosis. Rupturing of the cell membrane by p-MLKL oligomers and the subsequent release of intracellular components associated with damage-associated molecular patterns (DAMPs) are characteristic hallmarks of necroptotic cell death [14] (Fig. 1A).
(A) Under TNF stimulation, necroptosis pathway is activated and autolysosomal function is impaired due to defective lysosomal acidification. This results in a reduction in autophagic degradation of necrosome molecules such as toxic p-MLKL oligomers and a decrease in turnover of damaged mitochondria under inflammatory condition. Lysosomal membrane permeabilization (LMP) and the generation of ROS by impaired mitochondria further propagate p-MLKL oligomerization. The accumulated p-MLKL oligomers rupture the plasma membrane, resulting in necroptotic neuronal death and leading to the release and exposure of damage-associated molecular patterns (DAMPs) which are regarded as endogenous danger signals. (B) Lysosome-acidifying nanoparticles (AcNPs) restore lysosomal acidification by lowering luminal pH and promote autophagic degradation under TNF stimulated condition. This leads to enhanced autophagic degradation of p-MLKL oligomers, increased turnover of damaged mitochondria, and improved mitochondrial functions, contributing to decreased necroptosis. Overall, restoration of lysosomal acidification under TNF stimulated inflammatory condition attenuates necroptosis and improves neuronal functions and viability. The figure was created with BioRender.com.
Mounting evidence has shown that both TNF and TNFR1 expression are increased in post-mortem human AD brains [16, 17] and the levels of TNF and TNFR1 in the cerebrospinal fluid (CSF) and plasma are elevated [18, 19]. Studies in post-mortem human AD brains have revealed the role of TNFR1 activation in inducing neuronal necroptosis as characterized by specific expression of p-MLKL in degenerating neurons without activation of apoptosis [17, 20, 21]. Specifically, our recent study has demonstrated that these degenerating neurons are expressing phosphorylated tau and located near Aβ plaques, activated microglia, and cytotoxic T-cells, indicating that the neuroinflammatory microenvironment renders neurons susceptible to death [17]. In transgenic mouse models of familial AD [20, 22-24] and sporadic mouse model with Aβ oligomers injection [25], the elevated levels of necrosome molecules reinforce the mechanistic role of necroptosis in AD pathogenesis. Furthermore, stereotaxic injection of TNF in wild-type (WT) mice has been reported to activate necroptosis in mouse hippocampus [26]. In neuronal culture, studies have also shown that TNF activates necroptosis, with the highest percentage of cell death observed under the condition when caspases and inhibitor of apoptosis proteins (IAPs) are inactivated [17, 24, 26, 27]. Lowering the level of TNFR1 and treatment of necroptosis inhibitors or anti-TNF antibody are able to prevent necroptotic neuronal death, reduce the accumulation of toxic protein aggregates, and rescue cognitive impairments in neurons and AD mouse models [20, 22-24, 26, 27].
In addition to the presence of necroptosis in AD, recent studies have also highlighted the role of lysosomal dysfunction and autophagic impairment in TNFR1 mediated neuronal death [28, 29]. The autolysosome pathway initiates with phagophore and autophagosome formation followed by fusion of autophagosome and lysosome to form autolysosome which requires lysosomes with sufficient luminal acidification to have a complete autophagic degradation process [30]. It is important to note that while TNF has been suggested to promote autophagy initiation, studies have shown an increased accumulation of autophagic vesicles, indicating a terminal failure in the autophagic process [27, 31]. This is supported by strong evidence of lysosomal acidification dysfunction [24, 27, 32] and lysosomal membrane permeabilization (LMP) [28] under TNF stimulated condition. While V-ATPase dysfunction is a key culprit contributing to defective lysosomal acidification [33], reactive oxygen species (ROS) associated with mitochondrial dysfunction also drives LMP and leads to cathepsin release and cleavage of caspases, exacerbating necroptosis and cellular disintegration [28, 34]. Furthermore, interaction between TNFR1 signaling axis and autolysosomal pathway and the accumulation of necrosome molecules (e.g., p-MLKL) in poorly acidified lysosomes play key roles in mediating neuronal necroptosis [24, 35]. Altogether, these studies suggest that restoration of lysosomal acidification and function is a viable strategy to inhibit TNFR1 mediated neuronal necroptosis (Fig. 1B).
In this study, we revealed using neuropathological analysis of human post-mortem AD brains that TNFR1 mediated necroptosis correlates with autolysosomal dysfunction as indicated by downregulation of V-ATPase subunits and accumulation of autophagy receptor p62. Specifically, we showed that TNF induces lysosomal acidification defect, autophagy impairment, mitochondrial dysfunction, and neuronal death in SH-SY5Y neuroblastoma cells. Importantly, we further demonstrated that lysosome-acidifying nanoparticles (AcNPs) restore lysosomal acidification and metabolic functions and rescue neuronal necroptosis in both TNF-treated SH-SY5Y cells and APPNL-G-F mouse model of AD. We proposed that targeting lysosomal acidification dysfunction represents a novel therapeutic strategy to attenuate neuronal necroptosis in AD.
Materials and Methods
Data mining of human AD transcriptomic dataset
The RNA-sequencing (RNA-seq) transcriptomic dataset GSE173955 [36] was obtained from the Gene Expression Omnibus (GEO) database [37]. The 10 healthy control (HC) and 8 AD samples are from the hippocampus region of post-mortem brains with the age of the HC samples at 77.00 ± 8.50 years and AD samples at 91.75 ± 6.08 years. The dataset was analyzed using the GEO2R tool that makes use of the limma package in the Rstudio [38]. The mRNA expression from AD samples were compared against HC samples, and the differentially expressed genes (DEGs) were identified using the parameters of false discovery rate (FDR) adjusted P<0.01 and log2(fold change (FC))≥|0.5|. The pathway enrichment analysis was conducted using the Database for Annotation, Visualization and Integrated Discovery (DAVID) [39] which provides functional annotations of the DEGs based on Kyoto Encyclopedia of Genes and Genomes (KEGG), Gene Ontology Biological Process (GOBP), and Gene Ontology Cellular Component (GOCC) databases. In addition, gene set enrichment analysis (GSEA) [40] was conducted for both AD and HC samples of the GSE173955 dataset by comparing to a list of gene sets following default settings including 1000 permutations, Signal2Noise metric for ranking genes, and weighted enrichment statistics. The chip platform “Human_Gene_Symbol_with_Remapping_MSigDB.v2023.1.Hs.chip” was used for gene identification. The GOBP gene set from the Molecular Signatures Database (MSigDB) [41] was used as a reference.
Human post-mortem brain specimens
The human post-mortem AD brains used in this study were sourced from the Multiple Sclerosis and Parkinson’s Tissue Bank at Imperial College London and the South West Dementia Brain Bank at the University of Bristol. Snap-frozen tissue blocks from the hippocampus were acquired from 10 AD cases (82.90 ± 7.88 years) and 10 HC cases (84.20 ± 7.39 years). The post-mortem donations were obtained with full informed consent and under ethical approval by the National Research Ethics Committee (08/MRE09/31 and NHS REC No 18/SW/0029). The use of human samples in this study is also approved by the Institutional Review Board (IRB-2023-143) from the Nanyang Technological University (NTU), Singapore. Demographic data and neuropathological characteristics of the brain samples (Table S1) were provided by both brain banks, following standardized criteria of the UK MRC Brain Bank Network.
Animal maintenance and stereotaxic intrahippocampal injection
We used WT and APPNL-G-F homozygous knock-in mice (both male and female) with C57BL/6 background at 3-month-old for treatment procedures and the mice were sacrificed at 4-month-old for experiments. WT and APPNL-G-F mice were bred at the animal facility. The mice were housed in a temperature-controlled room (25 °C) in virus-free facilities on a 12 h light/dark cycle (7:30 a.m. on/7:30 p.m. off) and water ad libitum. All mice were maintained under specific pathogen-free conditions and treated with humane care. In each animal experiment, mice were randomly allocated to receive either PBS or AcNPs (75 mg/mL) at a total volume of 2 µL (total amount of 150 µg) via stereotaxic intrahippocampal injection at the respective coordinates from bregma (A/P: -2.1 mm; M/L: +/ -1.7 mm; D/V: 2.5 mm) [42]. For all stereotaxic injections, a 26G Hamilton syringe was used and the solution was applied at a speed of 0.5 µL/min. After injection, the needle was kept in place for an additional 5 min before gentle withdrawal. The mice were sacrificed 1 month post injection. Mice were sacrificed by perfusion and mouse brains were snap frozen in liquid nitrogen upon extraction and stored in -80°C for further processing. All procedures were conducted with the approval of the Institutional Animal Care and Use Committee (IACUC) animal use protocol A21043 and A22058 from NTU Singapore.
Tissue fixation and sectioning
For animal experiments, mouse brains were first fixed in 4% paraformaldehyde (PFA) in PBS solution (VWR, Cat# ALFAJ61899.AK) for 24 h, followed by overnight fixation in 30% sucrose solution (Sigma-Aldrich, Cat# S0389). The samples were then embedded in Tissue-Tek O.C.T compound (Sakura Finetek USA, Cat# 4583) using molds measuring 15x15 mm (Epredia, Cat# 58950). The mouse brains were then cut into 20 µm thick frozen tissue sections using a cryostat (Leica, Cat# CM1950) and placed on Superfrost Plus Adhesion Microscopic glass slides (Epredia, Cat# J1800AMNZ). For use of the human post-mortem brain tissues, 20 µm thick cryosections were cut from the snap frozen hippocampal blocks and placed on Superfrost Plus Adhesion Microscopic glass slides.
Tissue immunostaining and image acquisition
For immunostaining of both mouse and human brain tissues, the frozen sections were blocked with 5% normal goat serum (NGS) (Thermo Fisher Scientific, Cat# 31872) for 1 h at room temperature. Subsequently, the sections were incubated with primary antibodies followed by fluorescent secondary antibodies (Table S2). The tissue sections were subsequently incubated with DAPI (Thermo Fisher Scientific, Cat# 9542) for 10 min before washing with PBS. Lastly, antifade mounting medium (VectaShield, Cat# H-1000) was applied to the sections, before mounting of coverslip (24x50mm, Paul Marienfeld, Cat# 01-022-22). Slide scanner (Carl Zeiss Axio Scan Z1) was used to acquire fluorescent images at 40X magnification. Zen Blue software v2.6 was used to analyze the acquired images.
AcNPs synthesis and characterizations
The AcNPs were synthesized from poly(ethylene tetrafluorosuccinate-co-succinate) PEFSU polyesters as described previously [43]. Briefly, the polyesters (25% PEFSU) were dissolved in acetonitrile (Sigma Aldrich, Cat # 271004) and then passed through a 0.2 µm syringe filter (Millipore, Cat# Z741696) to eliminate any particles. Sodium dodecyl sulfate (SDS) (Sigma Aldrich, Cat# L3771) was dissolved in MilliQ water and stirred vigorously at 1700 rpm. The acetonitrile polyester solution was slowly added drop by drop into the vigorously stirred aqueous SDS solution. Once all the polyester solution had been added, the resulting emulsion was transferred into SnakeSkin dialysis tubing (MWCO 10 kDa) (Thermofisher Scientific, Cat# 68100) and dialyzed against MilliQ water for 24 h. For dynamic light scattering (DLS) measurements, 200 µL of the solution was diluted in 2.8 mL of MilliQ water, and the diameter was measured using the Brookhaven 90 plus NP sizer DLS instrument. For scanning electron microscopy (SEM) characterization, the AcNPs were diluted 50-fold in MilliQ water, and 10 µL droplets were applied onto silicon wafers, allowing them to air dry overnight. Later, these wafers were affixed to aluminium stubs using copper tape and coated with a 5 nm-thick Au/Pd layer through sputter coating. The samples were analyzed using a Supra 55VP field emission scanning electron microscope (ZEISS) with an accelerating voltage of 2 kV and a working distance of 5.5 cm.
Cell culture and cytotoxicity assay
SH-SY5Y cells (ATCC, Cat# CRL-2266) were cultured in DMEM/F12 media (Thermofisher scientific, Cat# 11320033) supplemented with 10% fetal bovine serum (Gibco FBS HI, Cat# A5256701) and 1% penicillin-streptomycin (Thermofisher Scientific, Cat# 15140122). The cytotoxicity assay was evaluated using an MTS Assay Kit (Abcam, Cat# ab197010). SH-SY5Y cells were cultured in a 96-well plate at 15000 cells/well for 24 h and subsequently treated with and without TNF and the respective doses of the AcNPs for 24 h. The cell viability was quantified relative to the no treatment control.
Lysosomal pH measurements
For live-cell imaging of lysosomal pH, SH-SY5Y cells subjected to respective treatments were stained with LysoSensor™ Yellow/Blue DND-160 (Thermofisher Scientific, Cat# L7545) following manufacturer’s protocol for 15 min. The stained cells were then imaged using a confocal microscope (ZEISS) with 360 nm excitation and images were acquired at 440 nm (blue channel) and 540 nm (yellow channel). The ratio between yellow and blue fluorescence intensities was calculated, and pH quantification was achieved by comparing it to a standard curve with known pH values. The standard curve was generated by correlating LysoSensor fluorescence ratio to known pH values using 2-(N-morpholino) ethanesulfonic acid buffer with varying pH.
Cathepsins activity assay
For cathepsin B activity assay, SH-SY5Y cells subjected to respective treatments were incubated with Magic red cathepsin B (MR-cathepsin B, Immunochemistry Technologies, Cat# 937) following manufacturer’s protocol for 1 h. For cathepsin D activity assay, the SH-SY5Y cells were treated according to manufacturer’s instructions (Abcam, Cat# ab65302). After treatment, the cells were washed three times with PBS, and the fluorescence was measured using a Synergy H1 model of hybrid multi-mode plate-reader with excitation/emission wavelengths of 531/629 nm for cathepsin B activity and 328/460 nm for cathepsin D activity.
Lysosomal membrane permeabilization (LMP) imaging
For LMP imaging, SH-SY5Y cells were first subjected to respective treatment conditions for 24 h. Subsequently, the cells were fixed using a 4% paraformaldehyde (PFA) in PBS solution for 20 min and rinsed twice with PBS. The fixed cells were washed three times with 0.1% Triton X-100 (Sigma Aldrich, Cat# X100) in PBS. Next, the cells were blocked with 5% normal goat serum (NGS) (Thermofisher Scientific, Cat# 31872) for 1 h at room temperature. The cells were then incubated with primary antibodies targeting galectin-3 (Proteintech, Cat# 14979-1-AP) overnight at 4 °C. The next day, cells were washed three times with 0.1% Triton X-100 in PBS, the cells were treated with Goat anti-Rabbit 488 nm (Thermofisher Scientific Cat# A-11008) fluorescent secondary antibody for 1 h at room temperature. Subsequently, the cells were washed with PBS. DAPI staining (Thermofisher Scientific, Cat# 9542) was performed for 10 min, followed by PBS washes. Fluorescent images were captured using a confocal microscope (ZEISS) at blue (excitation/emission 405/460 nm) and green (excitation/emission 490/520) channels.
Autophagy and mitochondrial turnover reporter assays
SH-SY5Y cells were transfected with GFP-LC3 reporter plasmid (Addgene, Cat# 11546) for 24 h, before treatment with different conditions. The cells expressing GFP-LC3 were then imaged using confocal microscopy (excitation/emission 490/510 nm) to visualize GFP-LC3 punctate dots within cells. To quantify mitochondria turnover under different conditions, SH-SY5Y cells were first transfected with mCherry-GFP-FIS1 (University of Dundee, Cat# DU55501) for 24 h to express the mitophagy reporter. Subsequently, the cells were subjected to different treatment conditions for another 24 h. Before imaging the cells, the LysoSensor™ Blue DND-167 dye (Thermofisher Scientific, Cat# L7533) was added for 15 min. Confocal microscopy was used to capture fluorescent images at blue (excitation/emission 405/460 nm), green (excitation/emission 490/520 nm), and red (excitation/emission 590/610 nm) channels.
Characterization of mitochondrial morphology and functions
SH-SY5Y cells were subjected to respective treatment conditions and then stained with MitoTracker Deep Red (Thermofisher Scientific, Cat# M22426) following manufacturer’s protocol for 30 min. The stained cells were washed and subsequently imaged using confocal microscopy with excitation/emission wavelengths of 640/665 nm. For mitochondrial morphology analysis, the acquired confocal images of mitochondria were processed using the Mitochondrial Network Analysis (MiNA) ImageJ macro [44]. Quantifiable measurements of mitochondrial morphology, such as mitochondrial footprint (µm2) and network branch length (µm), were plotted.
The processing parameters were kept consistent across images acquired from different treatment conditions. To measure mitochondrial membrane potential (MMP), SH-SY5Y cells were cultured in a 96-well plate and subjected to respective treatments, followed by analysis using the TMRE-Mitochondrial Membrane Potential Assay Kit (Abcam, Cat# ab113852). To measure mitochondrial superoxide release, the Mitochondrial Superoxide Assay Kit (Abcam, Cat# ab219943) was used. Both assays were conducted following the protocols provided by the manufacturer. The measurements were obtained using a Synergy H1 plate-reader with excitation/emission wavelengths of 549/575 nm for TMRE and 510/610 nm for the superoxide signals.
Western blotting
Sample preparations for cell lysates as well as mouse and human brain lysates were performed with the following procedures. SH-SY5Y cells treated with respective conditions were first washed with PBS, followed by the addition of lysis solution consisting of RIPA buffer (Thermofisher Scientific, Cat# 89900) supplemented with protease and phosphatase inhibitors (Thermofisher Scientific, Cat# 78440). Mouse whole brain tissues were homogenized using similar lysis solution as described above. The human hippocampus brain tissues were carefully extracted from frozen human HC and AD brain samples by delicately scoring the tissue block and then isolating the tissue through cryosectioning in a Leica cryostat. The resulting tissue samples were then homogenized using the lysis solution. All lysates were kept on ice for 15-30 min followed by centrifugation for 10 min at 13,500 g at 4 °C.
Once the cell and tissue lysates were prepared, Pierce™ BCA Protein Assay Kit (Thermofisher Scientific, Cat# 23225) was used to determine total protein concentrations for each sample. Samples were then mixed with 4X Laemmli Sample Buffer (Bio-Rad, Cat# 1610747) and boiled at 95 °C for 5 min before loading and running in 4-15% Mini-PROTEAN TGX Precast Protein Gels (Bio-Rad, Cat# 4561083) with Precision Plus Protein Dual Color Standards (Bio-Rad, Cat# 1610374). The proteins were transferred to polyvinylidene fluoride (PVDF) membrane (Millipore, Cat# ISEQ00010), and blocked with blocking buffer (Bio-Rad, Cat# 1706404). The membranes were then incubated with respective primary and secondary antibodies (Table S2). Densitometry was performed using ImageJ and protein expression levels were normalized to GAPDH loading control.
Statistical analysis
Statistical analyses were performed using the GraphPad Prism 9 software. Statistical analysis was conducted by using unpaired Student’s t test for difference between two conditions and one-way ANOVA with post hoc Tukey’s test for multiple comparisons. Statistical significance was determined by P<0.05 and indicated by *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 or ns for non-significance.
Results
Correlation of TNFR1 mediated necroptosis and autolysosomal dysfunction in hippocampal region of human AD brains
To investigate whether defective lysosomal acidification is present in human AD brains, we conducted data mining to analyze an existing RNA-seq transcriptomic dataset (GSE173955) [36] containing mRNA expression of hippocampus tissues from 8 AD and 10 HC samples. With a cutoff of FDR adjusted P<0.01 and log2(fold change (FC))>|0.5|, we identified 1769 DEGs with 1070 downregulated genes and 699 upregulated genes (Fig. 2A). We then performed pathway enrichment analysis using KEGG, GOBP, and GOCC databases to obtain the functional annotations of the DEGs. First, the identified DEGs were associated with pathways of neurodegeneration, AD, as well as glutamatergic and GABAergic synapses, validating the usefulness of this dataset for our study (Fig. 2B). We also showed that pathways related to synaptic functions (e.g., synaptic vesicle cycle and synaptic transmission) and metabolic activities (e.g., ROS and oxidative phosphorylation) are particularly enriched (Fig. 2B-D). Furthermore, our analysis highlighted the association of DEGs to lysosomal membrane, V-ATPase, proton transmembrane transport, vacuolar acidification, and regulation of TRP channels (Fig. 2B-D).
(A) Volcano plot illustrating the downregulated (blue) and upregulated (red) DEGs from the GSE173955 RNA-seq transcriptomic dataset. The analysis was conducted based on the mRNA expression of the 8 AD and 10 HC brain tissues using a cutoff of FDR adjusted P<0.01 and log2FC>|0.5|. (B-D) Pathway enrichment analyses performed using the (B) KEGG, (C) GOBP, and (D) GOCC databases, illustrating the functional annotations of the DEGs. (E) GSEA analysis of the GSE173955 dataset showing downregulation of proton transmembrane transport and mitochondrial membrane organization in AD compared to HC brains. (F) Specific V-ATPase subunits that are identified to be downregulated in AD brains. (G-H) Western blotting analysis of TNFR1 mediated necroptosis (TNFR1, p-MLKL, MLKL), top three downregulated V-ATPase subunits (ATP6V1A, ATP6V1H, and ATP6V1C1), and autophagy receptor p62 in the human HC and AD brains with GAPDH as a loading control. Data presented are relative to HC. Data are means ± SD of N=10 HC and N=10 AD brains. *P < 0.05, **P < 0.01, and ns indicates non-significance by unpaired Student’s t test for comparison between two samples.
To understand how the genes are involved in these pathways, we performed the pathway enrichment analysis separately for the upregulated and downregulated DEGs. From the analysis, upregulated genes were mainly related to regulation of RNA transcription. The analysis for downregulated genes suggests impairments in synaptic and metabolic activities as well as mitochondrial, autophagic, and lysosomal dysfunctions in AD (Fig. S1A-C). We then conducted GSEA analysis and illustrated downregulation of proton transmembrane transport, mitochondrial activities, and synaptic functions (Fig. 2E and S1D). With a focus in understanding lysosomal acidification alterations, we identified nine lysosomal V-ATPase subunits (ATP6V1A, ATP6V1H, ATP6V1C1, ATP6V1B2, ATP6V0D1, ATP6V1E1, ATP6V1D, ATP6AP1, and ATP6AP2) to be significantly downregulated in the dataset (Fig. 2F), indicating that lysosomal acidification was impaired in AD brains.
We further confirmed these observations in the protein level using 10 HC and 10 AD post-mortem human brains (Table S1). As an initial control, we illustrated the presence of activated microglia (Fig. S2A) as well as increased TNF and phospho-tau levels (Fig. S2B-C) in AD brains as compared to controls, indicative of neurodegenerative pathology. We then showed that there is an increase in both TNFR1 and p-MLKL levels with no significant changes in total MLKL level (Fig. 2G-H) and caspase activities (Fig. S2B-C), confirming the presence of TNFR1 induced necroptosis in AD brains. We further demonstrated reduction in the protein levels of the top three downregulated (log2FC≤-1) V-ATPase subunits (ATP6V1A, ATP6V1H, and ATP6V1C1) that we have identified from data mining, together with an increased accumulation of p62, illustrating both lysosomal dysfunction and autophagic impairment in AD brains (Fig. 2G-H). We further observed the expression of TNFR1, p-MLKL, MLKL, V-ATPase subunits and p62 in hippocampal neurons of human HC and AD brains (Fig. S3), thus establishing the correlation between TNFR1 induced neuronal necroptosis and autolysosomal dysfunction.
TNF induces lysosomal pH elevation in SH-SY5Y cells and AcNPs restore lysosomal acidification and cathepsin activity
To investigate the role of TNF-TNFR1 signaling axis in autolysosomal function, we first determined the effect of TNF in altering lysosomal acidification in SH-SY5Y cells. We showed that V-ATPase subunits (ATP6V1A, ATP6V1H, and ATP6V1C1) are decreased in TNF treated cells (Fig. 3A-B), recapitulating their downregulation in human AD brains. As expected, TNF induces lysosomal pH elevation from 4.5 to 5.6 in SH-SY5Y cells, making this cell model suitable to study lysosomal acidification dysfunction (Fig. 3C-D). Importantly, this lysosomal pH elevation is specific to TNF induction as stimulation by other cytokines did not change lysosomal pH (Fig. 3C-D). We then utilized a new type of lysosome-acidifying nanoparticles (AcNPs) that are capable of efficiently localizing into impaired lysosomes (Fig. S4A) [43, 45], to directly restore organelle luminal acidification in TNF stimulated SH-SY5Y cells. As a control, we showed that AcNPs do not affect the pH of control cells with sufficiently acidified lysosomes (Fig. S4B) and do not impart toxicity to SH-SY5Y cells (Fig. S4C). Importantly, we demonstrated that AcNPs acidify impaired lysosomes in a dose dependent manner (NP50 = 50 µg/mL and NP100 = 100 µg/mL) by lowering their pH from 5.6 to 4.6 (Fig. 3E-F). We also demonstrated that TNF induces LMP (Fig. 3G-H) and impairs lysosomal cathepsin B and D activities (Fig. 3I-J) in SH-SY5Y cells. AcNPs treatment prevented LMP and restored cathepsin enzyme activities. Overall, TNF induces lysosomal dysfunction and cathepsins inactivation in SH-SY5Y cells which can be rescued by AcNPs treatment.
(A-B) Western blotting analysis and quantification of V-ATPase subunits (ATP6V1A, ATP6V1H, and ATP6V1C1) in TNF-treated SH-SY5Y cells with GAPDH as a loading control. (C-F) Lysosomal pH measurement and quantification by confocal imaging with LysoSensor Yellow/Blue dye in (C-D) SH-SY5Y cells under different cytokine treatments and in (E-F) SH-SY5Y cells with and without addition of lysosome-acidifying nanoparticles (AcNPs). (G-H) Lysosomal membrane permeabilization as characterized by immunostaining with galectin-3 and image quantification in SH-SY5Y cells under respective treatment conditions. (I-J) Lysosomal cathepsin B and D enzyme activities in SH-SY5Y cells under respective treatment conditions. Data presented are relative to control without TNF and AcNPs treatment. NP50 and NP100 indicate AcNPs treatments at 50 µg/mL and 100 µg/mL respectively. Data are means ± SD of N=4 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and ns indicates non-significance by unpaired Student’s t test for comparison between two samples and one-way ANOVA with post hoc Tukey’s test for multiple comparisons.
AcNPs promote autophagic activity and increase mitochondrial turnover in TNF stimulated SH-SY5Y cells
As lysosome degradation is the final and determining step of complete autophagic degradation, we examined whether re-acidification of impaired lysosomes by AcNPs in TNF stimulated SH-SY5Y cells restores autophagic function. TNF caused autophagy inhibition as illustrated by the accumulation of autophagy-associated proteins p62 and LC3II in SH-SY5Y cells, and AcNPs treatment promoted their clearance (Fig. 4A-C). We further confirmed this observation by using a GFP-LC3 reporter, where TNF induced an aggregation of LC3 puncta and treatment of AcNPs increased the degradation and clearance of these puncta (Fig. 4D-E). Functional autophagy is required for proper turnover of mitochondria and maintenance of normal mitochondrial functions. To determine whether improved autophagic function with AcNPs treatment promotes mitochondrial turnover in TNF-treated SH-SY5Y cells, we transfected the cells with a mCherry-GFP-FIS1 reporter plasmid that reflects the accumulation of mitochondria in lysosomes with different extent of acidification by quantifying the fluorescence signals [46]. The green fluorescent protein (GFP) signal is quenched when mitochondria colocalize with sufficiently acidified lysosomes with low pH but the green fluorescence remained high in impaired lysosomes with elevated pH. Under all conditions, the mCherry fluorescence signal is unchanged and acts a control for the presence of mitochondria. In untreated SH-SY5Y control cells, normal mitochondrial turnover was illustrated by a basal red and green fluorescence of the mCherry-GFP-FIS1 reporter. TNF treatment increased the green fluorescence intensity (Fig. 4F-G), indicating the accumulation of mitochondria in poorly acidified lysosomes. Their colocalization was indicated by white puncta formed by overlapping red (mitochondrial protein), green (elevated lysosomal pH), and blue (lysosomes) fluorescence. Treatment of AcNPs restores lysosomal acidification and quenches the GFP fluorescence (Fig. 4F-G), leading to the appearance of purple puncta when mitochondria (red) colocalize with lysosomes (blue). This result indicates that AcNPs treatment improves autophagic clearance and mitochondrial turnover in TNF-treated SH-SY5Y cells.
(A-C) Western blotting analysis and quantification of autophagy-associated proteins p62 and LC3II in SH-SY5Y cells under respective treatment conditions with GAPDH as a loading control. (D-E) Confocal microscopy images and quantification of GFP-LC3 reporter assay in SH-SY5Y cells under respective treatment conditions. (F-G) Confocal microscopy images and quantification of SH-SY5Y cells transfected with mCherry-GFP-FIS1 mitophagy reporter plasmid followed by respective treatments. Data presented are relative to control without TNF and AcNPs treatment. NP50 and NP100 indicate AcNPs treatments at 50 µg/mL and 100 µg/mL respectively. Data are means ± SD of N=4 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, and ns indicates non-significance by one-way ANOVA with post hoc Tukey’s test for multiple comparisons.
AcNPs improve mitochondrial function and rescue TNF-induced necroptosis in SH-SY5Y cells With an increase in mitochondrial turnover, we then examined whether AcNPs restore mitochondrial functions by determining changes their morphology, MMP, and ROS production. By staining with MitoTracker Deep Red, TNF induced more mitochondrial fragmentation (Fig. 5A) in SH-SY5Y cells as quantified by the decrease in mitochondrial footprint (Fig. 5B) and network branches (Fig. 5C), which are associated with mitochondrial dysfunctions [47, 48]. With AcNPs treatment, there were more elongated mitochondria with tubular network (Fig. 5A-C), indicating an improvement in mitochondrial functions. Furthermore, AcNPs restored MMP as illustrated by the increase in TMRE intensity (Fig. 5D) and reduced ROS generation (Fig. 5E) in TNF-treated SH-SY5Y cells.
(A) Confocal microscopy images of SH-SY5Y cells treated with respective conditions and stained with MitoTracker Deep Red. (B-C) Quantification of mitochondrial images using MiNA analysis to obtain measurements for (B) mitochondrial footprint and (C) network branches. (D-E) Characterizations of mitochondrial functions including the measurements of (D) mitochondrial membrane potential (MMP) and (E) reactive oxygen species (ROS) generation in SH-SY5Y cells under respective treatment conditions. (F) Measurement of cell viability of SH-SY5Y cells with and without TNF (100 ng/mL) treatment. (G) Measurement of cell viability of TNF-treated SH-SY5Y cells with addition of increasing doses of AcNPs (1 to 200 µg/mL). AcNPs rescue TNF induced cell death in a dose-dependent manner with an EC50 of 46 µg/mL. (H-I) Western blotting analysis and quantification of necroptosis markers (p-MLKL and MLKL) and caspase-3 activity (aCasp-3) in SH-SY5Y cells under respective treatment conditions with GAPDH as a loading control. Data presented are relative to control without TNF and AcNPs treatment. NP50 and NP100 indicate AcNPs treatments at 50 µg/mL and 100 µg/mL respectively. Data are means ± SD of N=4 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and ns indicates non-significance by unpaired Student’s t test for comparison between two samples and one-way ANOVA with post hoc Tukey’s test for multiple comparisons.
Finally, we investigated whether AcNPs rescue TNF induced cell death in SH-SY5Y cells. TNF caused a significant cell death leading to 60% of cell viability (Fig. 5F) and treatment of AcNPs rescues cell death in a dose-dependent manner with a half maximal effective concentration (EC50) of 46 µg/mL, resulting in an increase in cell viability to 90% (Fig. 5G). As a control for specificity, we showed that AcNPs do not rescue lipopolysaccharides (LPS) induced cell death in SH-SY5Y cells (Fig. S5A-B) as the cellular impairment does not arise from lysosomal acidification dysfunction (Fig. 2C-D). In TNF-treated SH-SY5Y cells, both necroptosis and apoptosis pathways are activated (Fig. 5H-I), making it a good model to test whether AcNPs attenuate a specific pathway or both. Importantly, we observed that AcNPs reduced p-MLKL level without augmenting caspase-3 activities (Fig. 5H-I), suggesting that restoration of lysosomal acidification specifically attenuates necroptosis and not apoptosis. This demonstrates that re-acidification of impaired lysosomes under TNF stimulation restores mitochondrial functions, decreases toxic ROS production, and improves cellular functions and viability by attenuating TNFR1 mediated necroptosis.
Stereotaxic injection of AcNPs restore autolysosomal functions and attenuate necroptosis in the hippocampal region of the APPNL-G-F mouse model of AD
Having observed the effects of TNF in autolysosomal dysfunction and necroptosis in SH-SY5Y cells, we sought to investigate whether similar pathogenic mechanisms are present in the APPNL- G-F mouse model of AD and whether AcNPs can provide therapeutic effect in vivo. Glial activation and neuroinflammation are known to be present in the APPNL-G-F mice starting at 2-month-old [49], which has been shown to peak at 4-month-old with a sustained but lower level of inflammation at a later age [50]. We selected an early age of WT and APPNL-G-F mice at 3-month-old with glial activation (Fig. S6A) and elevated TNF level (Fig. S6B-C) for stereotaxic injection of PBS (control) and AcNPs into the mouse hippocampus and sacrificed the mice at 4-month-old.
We first showed that there is a significant increase in TNFR1 and p-MLKL levels with no change in the total MLKL level in APPNL-G-F mouse brains as compared to the WT controls (Fig. 6A-B), indicating the presence of necroptotic process. We also illustrated that there is no significant change in caspase activities in APPNL-G-F mouse brains (Fig. 6A-B). We then observed similar reduction in lysosomal V-ATPase subunits (ATP6V1A, ATP6V1H, and ATP6V1C1) and accumulation of p62, indicating autolysosomal impairment in APPNL-G-F mouse brains as compared to the WT controls (Fig. 6A-B). Importantly, we showed that necroptosis activation and autolysosomal impairments are present in the hippocampal CA1 region of APPNL-G-F mice (Fig. 6C-D and S7A-B). This makes APPNL-G-F mice a good model for our investigation of the effect of AcNPs in restoring lysosomal acidification and attenuating neuronal necroptosis in vivo.
(A-B) Western blotting analysis and quantification for TNFR1 mediated necroptosis (TNFR1, p-MLKL, and MLKL), caspase activities (aCasp-3 and aCasp-8), V-ATPase subunits (ATP6V1A, ATP6V1H, and ATP6V1C1), autophagy receptor p62, and GAPDH as a loading control. (C-D) Immunofluorescence images of p-MLKL, ATP6V1A, and p62 staining in the hippocampal CA1 region of mouse brains of WT, APPNL-G-F and APPNL-G-F injected with AcNPs. The amount of AcNPs (NP) injected is 150 µg per mouse brain hemisphere. Data presented are relative to WT control. Data are means ± SD of N=6-12 mice per group. *P < 0.05, **P < 0.01, ***P < 0.001, and ns indicates non-significance by one-way ANOVA with post hoc Tukey’s test for multiple comparisons.
Injection of AcNPs significantly reduced p-MLKL level without augmenting the TNFR1 level, indicating that the attenuation of necroptotic process was not due to reduced TNFR1 expression (Fig. 6A-B). We also observed restoration of lysosomal and autophagic functions as shown by an increase in V-ATPase subunits and a decrease in the accumulated p62 upon treatment of AcNPs, respectively (Fig. 6A-B). The rescue of autolysosomal dysfunction and attenuation of the necroptotic process were recapitulated in the hippocampal CA1 region (Fig. 6C-D and S7A-B). Altogether, our results demonstrated that autolysosomal impairment, in particular lysosomal acidification dysfunction, plays a key role in TNFR1 mediated neuronal necroptosis. Restoration of lysosomal acidification and function may provide an effective approach to attenuate neuronal necroptotic process in AD.
Discussion
Recent evidence has shown that the activation of neuroinflammatory processes is an early event of AD pathology that occurs due to multiple contributing factors including the accumulation of toxic protein aggregates and glial activation. While toxic protein aggregates may directly induce neuronal death [51, 52], recent studies have proposed that the inflammatory responses induced by these toxic protein aggregates may play a leading role in the initiation and subsequent progression of AD [53]. Hyperphosphorylated tau has been shown to induce necroptosis by promoting the formation of necrosome and stimulate cell-autonomous cytokine overexpression to propagate inflammation [23]. Aβ accumulation has been shown to lead to necroptosis in AD via microglia activation and cytokine production [25]. Furthermore, Aβ can act as DAMPs to increase the expression of proinflammatory cytokines [54] and elevated cytokine production can further lead to increased toxic protein expression and aggregation which results in an autocatalytic process that propagates the inflammatory processes [55]. Hence, targeting TNFR1 induced necroptosis by promoting autolysosomal degradation represents an important strategy in ameliorating the accumulation of toxic necrosome molecules and protein aggregates [55].
APPNL-G-F mice have been developed to recapitulate expression profiles of risk factor genes, key aspects of neuroinflammation, and Aβ plaque deposition, similar to that in AD patients [49]. With glial activation starting at 2 months of age, it has been suggested that intense neuroinflammation happens between 3-6 months of age in APPNL-G-F with a peak at around 4 months of age [50]. There is corresponding synaptic dysfunction without neuronal death, which is consistent with our observation in the initiation of a low level of necroptotic process. LPS treatment has been shown to increase TNF cytokine level in the hippocampus, accompanied with neuronal death indicated by increased TUNEL staining in APPNL-G-F compared to WT mice at 5-6 months of age [56]. APPNL-G-F mice have also demonstrated early metabolic defects starting from 2 months of age, such as mitochondrial dysfunction, autophagic impairment, and lysosomal defects [57-59]. Downregulation of lysosomal V-ATPase subunits has also been shown in APPNL-G-F [58] and human AD brains [60-62], consistent with our results and the notion of lysosomal impairment as the terminal failure in the autophagic processes.
Although sustained inflammation is present [63], a study has shown that there is no significant increase in the neuronal necroptotic process in APPNL-G-F mice at 12 months of age [64]. This could be due to a lower level of inflammation beyond the early ages. Another possibility is that the presence of necroptotic process observed in some studies could be due to externally induced inflammation such as exposure to inflammatory insults [56] and stereotaxic injection procedures [65] that may trigger further neuroinflammatory signaling and exacerbate neuronal impairment. Although not shown in brain cells, it is worth noting that a study has demonstrated that necroptotic cells show reduced release of conventionally secreted cytokines, although the levels of non-conventionally released cytokines were unaffected or elevated due to membrane permeabilization [66]. On the other hand, the presence of necroptosis is widely demonstrated in other rodent models, including APPNL-G-F rats [64] as well as APP/PS1 [24] and 5xFAD [20] mice. This warrants further investigations on the detailed molecular mechanisms in age-dependent cytokine regulation and neuronal dysfunction in different AD mouse models.
Recently, we have described a tripartite pathogenic mechanism of autophagy-lysosome-mitochondria axis in TNFR1 mediated neuronal necroptosis, with lysosome being the central mediator of this pathogenic axis [67]. While previous studies have focused on promoting autolysosome fusion by increasing autophagosome formation [24], we adopted a novel approach to target and remedy lysosomal acidification dysfunction and LMP to improve the autophagic processes. It has been suggested that limited release of proteases to the cytoplasm induces apoptosis, while massive LMP results in acute necroptosis [68], indicating that susceptibility of lysosomes to rupture is a key determinant for the propensity of occurrence of necroptotic process [69]. To counteract LMP, the endosomal sorting complex required for transport (ESCRT) machinery has been reported to repair lysosomal membrane and preserve lysosomal acidification [70], failure of which has been associated with autophagic inhibition and neurodegeneration. Studies have also shown that LMP related cell death is coordinated in a mitochondria-dependent manner. It has been suggested that p-MLKL can translocate to mitochondria to induce impairments [71] including decreased MMP and increased ROS production [17, 32, 72], consistent with our observations. This may form a positive feedback loop to further promote necrosome formation and necroptosis activation.
Both neuroinflammation [2, 3] and lysosomal acidification defect [45, 73] have been described as early pathologies of AD. Our study has demonstrated the presentation of these two pathogenic mechanisms in AD and the interaction between these two pathways. As TNF signaling can be neurodegenerative and neuroprotective, it is important to understand the contribution of individual type of brain cells to neuroinflammation to determine which inflammatory processes are good or bad under disease condition. Hence, it is important to elucidate the role of TNFR1 mediated necroptosis and lysosomal dysfunction in other cell types such as microglia [74, 75] and astrocytes [76]. In conclusion, targeting autolysosomal dysfunction and neuronal necroptosis are important therapeutic strategies for AD. With the notion of lysosome being a cellular signaling hub [77], we propose that lysosome-targeting therapies [45] may provide new therapeutic directions for treatment of AD and other neuroinflammatory diseases in general.
Competing interests
The authors declare that they have no competing interests.
Funding
C.H.L. is supported by a Lee Kong Chian School of Medicine Dean’s Postdoctoral Fellowship (021207-00001) from Nanyang Technological University (NTU) Singapore and a Mistletoe Research Fellowship (022522-00001) from the Momental Foundation USA. J.Z. is supported by a Presidential Postdoctoral Fellowship (021229-00001) from NTU Singapore.
Authors’ contributions
C.H.L. conceived and directed the research project. C.H.L. and J.Z. conducted all experiments.
G.W.Z.L. assisted in animal studies, immunostaining, and image acquisition. E.N.S. provided support in cell culture and biochemical assays. L.M.O. and J.I. assisted in bioinformatics and image analysis. R.R., and A.M.B. provided critical comments and edited the manuscript. C.H.L. and J.Z. wrote the manuscript. All authors have approved the final version of the manuscript.
Supplementary Figures
(A-C) Pathway enrichment analysis of downregulated DEGs in the GSE173955 dataset using (A) KEGG, (B) GOBP, and (C) GOCC databases. (D) Additional GSEA analysis of the GSE173955 dataset illustrating downregulation of synaptic and mitochondrial functions.
(A) Immunofluorescence images of HC and AD brains stained with Iba1 to determine the extent of microglia activation. (B-C) Western blotting analysis and quantification of the expression level of TNF, Tau (AT8), and caspase activities (aCasp-3 and aCasp-8) in human HC and AD brains with GAPDH as a loading control. Data presented are relative to HC. Data are means ± SD of N=10 HC and N=10 AD brains. **P < 0.01 and ***P < 0.001, and ns indicates non-significance by unpaired Student’s t test for comparison between two samples.
Immunofluorescence images showing the presence of TNFR1, p-MLKL, MLKL, V-ATPase subunits (ATP6V1A, ATP6V1H, and ATP6V1C1), and autophagy receptor p62 in hippocampal neurons of HC and AD brain tissues.
(A) Characterizations of AcNPs using SEM and DLS to determine their morphology and size. (B) Lysosomal pH measurement and quantification with LysoSensor Yellow/Blue dye in SH-SY5Y control cells with and without addition of AcNPs. (C) Measurement of cell viability in SH-SY5Y control cells with increasing doses of AcNPs treatment (1 to 400 µg/mL). Data presented are relative to control cells without AcNPs treatment. Data are means ± SD of N=4 independent experiments. NP100 indicates AcNPs treatment at 100 µg/mL. *P < 0.05 and ns indicates non-significance by unpaired Student’s t test for comparison between two samples and one-way ANOVA with post hoc Tukey’s test for multiple comparisons.
(A) Measurement of cell viability of SH-SY5Y cells with and without LPS (1 µg/mL) treatment. (B) Measurement of cell viability of LPS-treated SH-SY5Y cells with addition of increasing doses of AcNPs (1 to 200 µg/mL). AcNPs do not rescue LPS induced cell death in SH-SY5Y cells. Data presented are relative to control without LPS and AcNPs treatment. Data are means ± SD of N=4 independent experiments. *P < 0.05 and ns indicates non-significance by unpaired Student’s t test for comparison between two samples and one-way ANOVA with post hoc Tukey’s test for multiple comparisons.
(A) Immunofluorescence images of WT and APPNL-G-F mouse brains stained with Iba1 to determine the extent of microglia activation. (B-C) Western blotting analysis and quantification of the expression level of TNF in WT and APPNL-G-F mouse brains with GAPDH as a loading control. Data presented are relative to WT control. Data are means ± SD of N=12 mice per group. ***P < 0.001 by unpaired Student’s t test for comparison between two samples.
Immunofluorescence images showing the presence of TNFR1, MLKL, and V-ATPase subunits (ATP6V1H, and ATP6V1C1) as well as caspase activities (aCasp-3 and aCasp-8) in the hippocampal CA1 region of mouse brains under respective conditions. The amount of AcNPs (NP) injected is 150 µg per mouse brain hemisphere. Data presented are relative to WT control. Data are means ± SD of N=6 mice per group. *P < 0.05, **P < 0.01, and ns indicates non-significance by one-way ANOVA with post hoc Tukey’s test for multiple comparisons.
Acknowledgements
The authors thank Jia Hui Wong, Enoch Kwok, and Evridiki Asimakidou for their technical support and discussion in this study. The authors also thank the funding sources for supporting this work.
Footnotes
↵# Equal contribution
References
