Spermidine protects mitochondria against amyloid-β oligomers

Abstract

Background

Alzheimer's disease (AD) is a neurodegenerative condition characterized by amyloid-β (Aβ) plaques in the brain. In early pathological conditions, Aβ oligomers localize in mitochondria to promote disease progression by causing mitochondrial dysfunction.

Objective

Establish a robust AD pathological in vitro model and assess the therapeutic effect of spermidine (SPD) against extracellular Aβ oligomers.

Methods

In this study, high-resolution respirometry, flow cytometry, and immunoblotting techniques were used to characterize the negative effects of Aβ oligomer localization in mitochondria and assess the potential neuroprotective effects of SPD.

Results

Oligomeric Aβ accumulates in and causes mitochondrial dysfunction, as evidenced by increased mitochondrial superoxide production, decreased mitochondrial mass, and decreased maximal mitochondrial respiration and spare reserve capacity. SPD, an endogenously produced polyamine that can also be provided through exogenous supplementation, improved these markers and protected mitochondria of oligomer-treated cells.

Conclusions

SPD ameliorated mitochondrial dysfunction caused by Aβ oligomers in vitro.

Introduction

An estimated 55-million people world-wide are living with dementia, and this number will grow to nearly 78 million by 2030. ¹ Comprising 60–80% of dementia cases, Alzheimer's disease (AD) is projected to increase by 153 million world-wide by 2050. ² Characterized as a progressive neurodegenerative disease, AD differs from other pathologies by the presence of extracellular amyloid-β (Aβ) plaques and intracellular hyperphosphorylated tau neurofibrillary tangles. ³ Although these histological hallmarks are defined, the underlying mechanism(s) responsible for disease pathophysiology are still unknown. Recent evidence points to changes in mitochondria function (i.e., energy metabolism, and oxidative stress) in neuronal and neuroglial cells as potential early markers of AD development.^4,5 The brain is 2% of total human body weight but is responsible for 20% of daily energy consumption; ⁶ therefore, alterations to energy production caused by subcellular localization of Aβ could promote neurodegeneration and disease progression.

Endogenous production of spermidine (SPD), a polyamine that participates in several cell and metabolic functions, declines with age and plasma concentrations are significantly decreased in patients with AD compared to healthy controls. ⁷ Systemic decreases can be recovered through exogenous supplementation, and SPD permeates the blood-brain barrier, making it therapeutically relevant.^8,9 SPD induces autophagy, preserves mitochondrial function, elicits anti-inflammatory properties, and exerts protective effects in neurodegenerative models. ¹⁰ In this study, the subcellular accumulation of Aβ and its impact on mitochondrial function in neuronal and astrocytic cells were assessed. This was followed by an investigation into the potential in vitro protective effects of exogenous SPD.

Methods

Cell lines

Mouse neuroblastoma N2A and human U87 mg glioblastoma cell lines were cultured at 37°C with 5% CO₂ and 95% relative humidity in Dulbecco's Modified Eagle Medium (DMEM; Thermo-Fisher) supplemented with 10% fetal bovine serum (FBS), 4.5 g/L D-glucose, L-glutamine, and antibiotics. Spermidine, Aβ_1–42 peptides, chloroquine, and other chemicals were purchased from Sigma-Aldrich.

Immunoblotting

Cells were collected and lysed using RIPA buffer supplemented with 1% protease inhibitor solution. Lysates were then diluted into aliquots using RIPA buffer and 5 µL of sample buffer to reach 20–30 µg of protein. Samples were heated at 95°C for 5 min and loaded into an SDS-PAGE gel where proteins were separated at 150 V for 60 min. Proteins were then transferred to polyvinylidene fluoride (PVDF) membrane using the Trans-Blot Turbo Transfer System (Bio-Rad). Membranes were blocked in a skimmed milk Tris-buffered saline Tween (TBS-T) solution for 2 h. Primary antibody incubation was conducted overnight at 4°C on a shaker with LC3B, AMPK, p-AMPK, and PINK1 (1:1000–10,000, Abcam) antibodies in TBS-T. PVDF membranes were then washed repeatedly with TBS-T over a 30-min period and were eventually incubated with the appropriate secondary antibody at room temperature for 1 h (1:8000, Abcam). Protein visualization was conducted using enhanced chemiluminescence solution (Bio-Rad) and imaged using a Syngene Chemi Genius 2 Bio Imaging System. Densitometry was conducted using ImageJ analytical software (National Institutes of Health) and protein changes were normalized relative to the loading control protein, GAPDH.

Mitochondrial mass and ROS

Mitochondrial mass was assessed using MitoTracker Green (Life Technologies; M7514) and flow cytometry. Cells were plated and treated with SPD, Aβ oligomer, or a combination. Cells were then washed with PBS, unadhered, and collected in 5 mL centrifuge tubes (Falcon). Cells were then stained in 1 mL of MitoTracker Green working solution, as outlined by the manufacturer for 25-min in an incubator at 37°C. Cell viability was then assessed by 7AAD staining (1 µg/mL; Cayman Chemical) after an additional wash with FACS buffer solution. Cells were strained into 1 ml Eppendorf tubes using 20 µm cell strainers (Falcon) and plated in triplicate into a 96-well plate. Green-B and Red-B fluorescence was measured via flow cytometry (Guava^® easyCyte™ 8HT). ROS was measured by flow cytometry using MitoSOX staining, as per manufacturer protocol (Life Technologies; M36008). Staining occured over a 15-min period at 37°C. Red-B fluorescence was measured during flow cytometry to determine mitochondrial ROS content (Guava^® easyCyte™ 8HT).

Cellular respiration

Oxygen consumption rate (OCR) of U87 mg cells was measured using the MitoStress Test methodology on the Seahorse XFe24 Respirometer (Agilent) at a cell density of 2.5 × 10⁴ cells/well. One day prior, a XFe24 FluxPak (Agilent) cartridge was hydrated using XF Calibrant (Agilent) overnight at 37°C. Following 24 h, cells were washed twice with XF DMEM (Agilent) supplemented with 1 mM pyruvate (Gibco), 2 mM glutamine (Gibco), and 10 mM glucose and incubated at 37°C for 1 h. Port A of the cartridge was loaded with 56 µL of 20 µM Oligo A (Sigma-Aldrich) to reach a final concentration of 2 µM upon injection; Port B was loaded with 62 µL of 20 µm FCCP (Sigma-Aldrich) to reach a final concentration of 2 µM upon injection, and Port C was loaded with 69 µL of 20 µM Rot and 30 µM AA to reach a final concentration of 2 µM and 3 µM in each well following injection. Following calibration of the XFe24 FluxPak, the utility plate was removed, and the cell plate was loaded to start the assay. OCR results were analyzed using Wave Software (Agilent) and GraphPad 10.0 Prism software.

Mitochondrial isolation

Following a 1-h incubation with Aβ or vehicle control, U87 mg cells were collected, and mitochondria were isolated using a commercial kit (Abcam; 110170). Mitochondria-enriched pellets were resuspended in reagent C and subjected to protein quantification (i.e., using BCA) and immunoblotting to determine the protein levels of Aβ. All steps were performed on ice or at 4°C.

Aβ oligomer preparation

Aβ oligomers were prepared using the protocol outlined in Stine et al., 2011. Amyloid-β peptide 1–42 (Sigma-Aldrich) was prepared into a 1 M stock solution by injecting 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) directly into the vial. ¹¹ Once dissolved, the rubber septum is pierced and incubated at room temperature for 30 min before aliquoting 100 µL into 1.5 mL Eppendorf tubes (Thermo-Fisher). ¹¹ Samples evaporated overnight in a biological safety cabinet and were transferred to a SpeedVac and centrifuge for 1 h. Peptides were resuspended in DMSO to make a 5 mM oligomeric Aβ solution and once the sample was vortexed for 15 s, it was transfer to 4°C to incubate for 24 h. ¹¹

Statistical analysis

All statistical analyses were conducted using GraphPad 10.0 Prism software. Statistical significance was determined using Students t-tests, or one-way ANOVA tests with Tukey's multiple comparisons tests (p < 0.05).

Results

Amyloid-β oligomer exposure alters mitochondrial metabolism and abundance

Use of Aβ oligomers is a robust in vitro model to study AD ¹¹ (Figure 1). Aβ oligomers form the building blocks for AD plaques and, in the model, their cellular accumulation damages mitochondria and other organelles (Figure 1). To assess subcellular localization, mitochondria were isolated from U87 mg cells following Aβ oligomer exposure (Figure 2A); fraction purity was determined by the presence of cytochrome c (Cyt c). Aβ was found in mitochondrial (Mt) and supernatant (Sn) fractions (Figure 2A) showing that the oligomers accumulate in mitochondria.

Figure 1.

Uptake and cellular localization of Aβ oligomers. Schematic of the production and movement of Aβ oligomers into the mitochondria to how they induce mitochondrial damage and PINK1/Parkin mediated mitophagy. The location of AβPP processing has been simplified for conceptual comparison and do not reflect the cellular compartment in which processing is carried out; for a detailed description of compartment-specific AβPP processing, please refer to Chen et al., 2017. ¹² .

Figure 2.

Subcellular localization of Aβ causes dysfunction of mitochondrial respiration, increases superoxide production, and decreases mitochondrial mass. (A) Immunoblot of mitochondrial (Mt) and supernatant (Sn) enriched fractions of U87 mg cells treated with 0 or 10 µM of Aβ oligomers for 1 h. Representative blots are shown. Expression levels were determined using densitometry and were normalized to cytochrome c in mitochondrial fractions and GAPDH in supernatant fractions. (B) Seahorse XFe-24 instruments analysis of mitochondrial respiration represented as a function of OCR over time in U87 mg cells treated with or without 10 µM of Aβ oligomers under basal conditions or following the sequential addition of oligo A (2 µM), FCCP (2 µM), and AA/Rot (3 µM, 2 µM). (C) Basal respiration, (D) maximal respiration, or (E) percent spare reserve capacity (SRC) of U87 mg cells treated with or without 10 µM of Aβ oligomers. All measurements were made following 24 h of treatment. (F) Superoxide levels of U87 mg astrocyte cells following the addition of 10 µM of Aβ oligomers for 1 and 4 h; measurements were made using flow cytometry and MitoSOX staining. (G) Superoxide levels of N2A neuronal cells following the addition of 10 µM of Aβ oligomers for 1 and 4 h. (H) Mitochondrial mass of U87 mg astrocyte cells following treatment with 10 µM of Aβ oligomers for 24–72 h, respectively. (I) Mitochondrial mass of N2A neuronal cells following treatment with 10 µM of Aβ oligomers for 24–72 h, respectively. Figures show mean value from three biological replicates consisting of three technical replicates. Measurements were made using flow cytometry and MitoTracker Green staining (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Next, the impact of Aβ oligomer accumulation on mitochondrial respiration was assessed using respirometry. Aβ oligomers (10 µM for 24 h) did not change basal respiration rates (Figure 2B and C); however, there were significant decreases in maximal respiration (t₆ = 2.670; p < 0.05, Figure 2D). As a result, astrocytes exposed to Aβ oligomers had a significant decrease in SRC relative to controls (t₅ = 2.852; p < 0.01, Figure 2E).

To further assess the impact of Aβ oligomers on mitochondria function, superoxide production was measured using flow cytometry and MitoSOX staining. In the presence of oligomers, astrocytes produced significant amounts of superoxide (t₆ = 4.912; p < 0.01; t₄ = 4.580; p < 0.05; Figure 2F). A similar trend was observed in neuronal cells (t₄ = 5.978; p < 0.01; t₆ = 4.224; p < 0.01; Figure 2G). Astrocytes exposed to Aβ oligomers also showed a rapid decrease in mitochondrial mass, which remained significantly decreased over 72 h (t₄ = 6.375; p < 0.01; t₄ = 6.289; p < 0.01; t₄ = 7.198; p < 0.01, Figure 2H). Similarly, neuronal N2A cells exhibited a significant decrease in mitochondrial mass following 10.0 µM Aβ exposure at 24, 48, and 72 h (t₄ = 9.690; p < 0.001; t₆ = 6.291; p < 0.001; t₄ = 8.825; p < 0.001; Figure 2I). This data suggests that exposure to oligomeric Aβ negatively affected both mitochondrial function and abundance.

Spermidine decreases amyloid-β subcellular localization and protects against amyloid-β oligomer-induced mitochondrial dysfunction

To determine effects on mitochondrial localization, cells were treated with SPD in the presence of 10.0 µM Aβ oligomers for 1 h. Densitometry analysis of immunoblots indicate a decrease in Aβ in the mitochondrial and supernatant fractions following SPD treatment (Figure 3A).

Figure 3.

Spermidine decreased Aβ subcellular localization and protected against Aβ oligomer induced mitochondrial dysfunction. (A) Immunoblot of mitochondrial (Mt) and supernatant (Sn) enriched fractions of U87 mg cells treated with 10 µM SPD or 10 µM of Aβ oligomers for 1 h. Representative blots are shown. Expression levels were determined using densitometry and were normalized to cytochrome c in mitochondrial fractions and GAPDH in supernatant fractions. (B) Seahorse XFe-24 instruments analysis of mitochondrial respiration represented as a function of OCR over time in U87 mg cells administered with or without 10 µM of Aβ oligomers and 5 µM SPD. Measurements were made continuously at under basal conditions, following the sequential addition of oligo A (2 µM), FCCP (2 µM), and AA/Rot (3 µM, 2 µM). (C) Basal respiration of U87 mg cells treated with or without 10 µM of Aβ oligomers and 5 µM SPD. (D) Maximal respiration of U87 mg cells treated with or without 10 µM of Aβ oligomers and 5 µM SPD. (E) Percent spare reserve capacity (SRC) of U87 mg cells treated with or without 10 µM of Aβ oligomers and 5 µM SPD. All measurements were made following 24 h of treatment. (F) Superoxide production in U87 mg astrocyte cells treated with or without 5 µM of SPD in the presence of 10 µM of Aβ oligomers. (G) Superoxide production in N2A neuronal cells treated with or without 5 µM of SPD in the presence of 10 µM of Aβ oligomers. Figures show the mean of three technical replicates from three biological replicates. All measurements are normalized to the mean fluorescence value of the control from each respective biological replicate. All measurements were made using flow cytometry and MitoSOX staining (*p < 0.05, **p < 0.01, ***p < 0.001).

Functionally, SPD improved mitochondrial respiration in the presence of Aβ oligomers for 24 h. SPD alone had no significant effects on basal or maximal respiration (t₆ = 0.4108; t₆ = 0.7923; Figure 3C and D). In contrast, SPD protected cells from Aβ oligomer-induced SRC reductions and returned SRC to baseline (t₅ = 0.0446; p < 0.05, Figure 3E). Mitochondrial superoxide production showed a similar pattern. As previously shown, Aβ oligomer exposure significantly increased superoxide production after 1 and 4 h of exposure (F_(2,8)= 16.68; p < 0.01; F_(2,5)= 15.01; p < 0.01_, Figure 3F). Importantly, 1-h treatment with SPD significantly decreased superoxide production in Aβ exposed U87 mg cells (F_(2,8)= 16.68 ; p < 0.05; Figure 3F). SPD did not significantly decrease Aβ-induced superoxide production in N2A cells at the times and doses tested (Figure 3G).

Spermidine protects against mitochondrial loss caused by amyloid-β oligomer exposure

Aβ caused significant decreases in mitochondrial mass in U87 mg astrocytes (F_(2,6) = 26.1; p < 0.01; F_(2,5) = 19.2; p < 0.005; F_(2,5) = 19.2; p < 0.005; Figure 4A). However, Aβ oligomer-induced mitochondrial loss was attenuated by SPD, and a significant increase in mitochondrial mass was observed (F_(2,6) = 26.1; p < 0.001; F_(2,5)= 9.2; p < 0.005; Figure 4A). SPD imparted a similar protective effect against mitochondrial loss in neuronal cells. Following a 24-, and 48-, or 72-h exposure to Aβ oligomers, mitochondrial mass was significantly reduced in N2A cells that was attenuated with SPD (5 µM) treatment (F_(2,6) = 9.58; p < 0.05, F_(2,6) = 27.5; p < 0.001; F_(2,8) = 26.8; p < 0.001) (Figure 4B).

Figure 4.

Spermidine rescues losses in mitochondrial mass caused by Aβ oligomer treatment in astrocytes. (A) Mitochondrial mass of U87 mg astrocyte cells treated with or without 5 µM of SPD in the presence of 10 µM of Aβ oligomers. (B) Mitochondrial mass in N2A neuronal cells treated with or without 5 µM of SPD in the presence of 10 µM of Aβ oligomers. Figures show the mean of three technical replicates from three biological replicates. All measurements are normalized to the mean fluorescence value of the control from each respective biological replicate. All measurements were made using flow cytometry and MitoTracker Green staining. (C) LC3B I/II expression was measured in U87 mg cells were treated with 5 and 10 µM dosages of SPD in the presence or absent of 20 µM of the lysosomal inhibitor, CQ. Representative blot is shown. (D) LC3B I to LC3B II protein level ratios were quantified using densitometry analysis and expressed in arbitrary units (AU) relative to the loading control, GAPDH. (E) N2A cells were treated with 5 µM of SPD, 10 µM of Aβ oligomers, or in combination where full length PINK1 accumulation was assessed following a 4-h treatment duration. Representative blot is shown. (F) PINK1 protein levels were quantified using densitometry analysis and expressed in AU relative to the loading control, GAPDH. (G) p-AMPK and AMPK expression was assessed in U87 mg cells were treated with 5 and 10 µM doses over 4 h. Representative blots are shown. (H) p-AMPK and AMPK protein levels were quantified using densitometry analysis and expressed in AU relative to the loading control, GAPDH. Data shown consists of measurements from three biological replicates (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Mitophagy is the lysosomal degradation of mitochondria characterized by autophagosome formation and is initiated, at least in part, by AMPK activation or the presence of PINK1, a mitochondrial serine/threonine-protein kinase that senses damage. PINK1 increased following Aβ oligomer exposure relative to control and SPD treated groups (Figure 4E and F: F_(3,8) = 8.885; p < 0.01). In contrast, Aβ oligomer exposed cells treated with SPD (5 µM) did not change full length PINK1 accumulation relative to the control group (Figure 4E and F), suggesting a protective effect against Aβ oligomer-induced mitochondrial damage. SPD is an autophagy inducer so its effect on AMPK was tested. Here, SPD resulted in a 3.0–3.5-fold increase in the phosphorylation of AMPK (F_(2,6) = 9.350; p < 0.05; Figure 3G and H) as well as a change in the LC3BI/II ratio. Cytosolic LC3B-I is conjugated to phosphatidylethanolamine to form LC3B-II at the autophagosome and the LC3BII/LC3BI ratio is a typical marker of autophagic flux. SPD in combination with CQ, a lysosomal degradation inhibitor, significantly increased the LC3BII/I ratio (F_(5,12) = 10.39; p < 0.05; Figure 4C and D) relative to the control. SPD-induced autophagy did not change mitochondrial mass (not shown). These results show that SPD-induced autophagy may be a protect against Aβ oligomer-induced mitochondria damage.

Discussion

The hallmark requirements for AD diagnosis includes identification of extracellular Aβ plaque deposits and neurofibrillary tau tangles; however, their role in pathophysiology is well defined. ³ This study aimed to understand the negative effects of Aβ oligomers on mitochondria function and the potential therapeutic benefit of SPD. Aβ oligomers accumulated in and led to mitochondrial dysfunction by negatively impacting respiration, increasing ROS and decreasing mass. SPD ameliorated mitochondrial dysfunction in both neuronal and astrocyte cell lines by reducing Aβ mitochondrial accumulation and improving overall function. Mechanistically, SPD activated mitophagy-related pathways.

Aβ oligomers are the building blocks of extracellular senile plaques associated with AD pathology. ¹² Cleavage of the AβPP by amyloidogenic secretase enzymes generate Aβ₄₂ monomers that become more insoluble as the peptide’s secondary structure increases in complexity.^12,13 This results in peptide aggregation and the formation of oligomers that eventually generate insoluble Aβ fibrils that then aggregate to form senile plaques. ¹³ Prior to plaque formation, Aβ oligomers enter the cell from the extracellular space and can localize in mitochondria causing metabolic dysfunction in multiple cell types in the brain.^5,14,15 In this study, we verified that high-molecular weight Aβ peptides accumulate in mitochondria.

Mitochondrial alterations, including insufficient energy metabolism, ROS production and altered calcium homeostasis, are observed prior to the complete formation of the neuropathological AD hallmarks. ¹⁶ Our data showed that Aβ exposure increased mitochondrial ROS production two-fold, which was followed by a decrease in mitochondrial mass. Sustained elevation of ROS can damage mitochondrial structure, function, and DNA, and can initiate inflammatory responses in neuroglial cells.^14,16,17 Further, mitochondrial respiration was compromised following Aβ exposure. While basal oxygen consumption remained stable, maximal respiration and spare reserve capacity were significantly decreased in the presence of the oligomers. These data supports the hypothesis that Aβ oligomers, or less complex Aβ species, could potentially play a role in the initiation of the disease by elevating ROS and hindering mitochondrial function.

Mitophagy is the lysosomal degradation of damaged or non-functional mitochondria and its activation can lead to disease pathology or protect cells from disease pathology.^18–20 Commonly observed throughout pathological progression of AD, mitochondrial dynamics become unbalanced and significantly favor mitochondria fission and ultimately mitophagy.^5,15,21 In this study, both U87 mg and N2A cells decrease mitochondrial mass following Aβ exposure. Aβ oligomers increased accumulation of full length PINK1, which is a signal of mitochondria dysfunction that in turn activates mitophagy through Parkin-mediated ubiquitination that marks mitochondria for autophagic degradation.^22,23 Interestingly, the absence of PINK1 accumulation in SPD-treatment indicates that SPD-induced autophagy is independent of mitochondrial dysfunction. This observation is further supported by data showing SPD activates AMPK. Whether SPD-induced mitophagy protects against Aβ oligomer-induced mitochondrial dysfunction remains to be determined; however, other compounds such as urolithin A and actinonin stimulate mitophagy to ameliorate AD pathology.^24,25 Through the autophagic clearance of damaged mitochondria, Aβ and tau pathologies were decreased and reversed memory impairments in AD models. Moreover, targeting mitochondrial function with curcumin, resveratrol, melatonin, and coenzyme Q10 showed beneficials effects on Aβ pathology.^26–29 In these studies, improving mitochondrial respiration, reducing mitochondrial ROS, and restoring mitochondrial membrane potential imparted a neuroprotective effect against Aβ toxicity in both in vitro and in vivo AD models.^26–29 Together, these data show that Aβ oligomers induce mitochondrial damage and impart dysfunction that is ameliorated by SPD.

SPD protected cells from mitochondrial dysfunction caused by Aβ oligomers. SPD decreased mitochondrial accumulation of the oligomers which decreased ROS production and restored respiration and mass. This neuroprotective effect is imperative to the preservation of cellular function due to the high energy demands of the brain. Decreased mitochondrial abundance and impaired mitochondrial respiration occur in AD pathology and are often associated with late-stage disease. ³⁰ In this study, we demonstrate that these metabolic changes occur early at stages that would precede the formation of senile plaques. Similar protective alterations to cellular metabolism by SPD are documented in a variety of models. Through mitophagy, autophagy, and improvements to mitochondrial respiration, SPD extends mouse life-span through cardioprotective effects,^31,32 improved CD8+ T cell responses, ³³ reduced atherosclerosis and inflammation, ³⁴ and improved intestinal epithelial barrier function in obesity models. ³⁵ While the findings of this study are promising, pre-clinical studies are needed to validate these in vitro results. Transgenic AD models would also allow for the further characterization of Aβ oligomers on early pathological mitochondrial dynamics. Furthermore, evaluation of the endogenous synthesis and catabolism of SPD by spermidine synthase and spermidine/spermine-N(1)-acetyltransferase as well as exogenous supplementation will further justify viable prophylactic options.

Conclusions

In summary, oligomeric Aβ demonstrated pathological effects on mitochondrial function of astrocytes and neurons in vitro. Pro-pathological alterations to mitochondrial mass, superoxide production, and respiration caused by the subcellular localization of Aβ oligomers were ameliorated by SPD treatment, illustrating its potential therapeutic value against early AD progression.