Iron dyshomeostasis, lipid peroxidation and perturbed expression of cystine/glutamate antiporter in Alzheimer's disease: Evidence of ferroptosis

Abstract

Iron dyshomeostasis is implicated in Alzheimer's disease (AD) alongside β-amyloid and tau pathologies. Despite the recent discovery of ferroptosis, an iron-dependent form cell death, hitherto, in vivo evidence of ferroptosis in AD is lacking. The present study uniquely adopts an integrated multi-disciplinary approach, combining protein (Western blot) and elemental analysis (total reflection X-ray fluorescence) with metabolomics (¹H nuclear magnetic resonance spectroscopy) to identify iron dyshomeostasis and ferroptosis, and possible novel interactions with metabolic dysfunction in age-matched male cognitively normal (CN) and AD post-mortem brain tissue (n = 7/group). Statistical analysis was used to compute differences between CN and AD, and to examine associations between proteins, elements and/or metabolites. Iron dyshomeostasis with elevated levels of ferritin, in the absence of increased elemental iron, was observed in AD. Moreover, AD was characterised by enhanced expression of the light-chain subunit of the cystine/glutamate transporter (xCT) and lipid peroxidation, reminiscent of ferroptosis, alongside an augmented excitatory glutamate to inhibitory GABA ratio. Protein, element and metabolite associations also greatly differed between CN and AD suggesting widespread metabolic dysregulation in AD. We demonstrate iron dyshomeostasis, upregulated xCT (impaired glutathione metabolism) and lipid peroxidation in AD, suggesting anti-ferroptotic therapies may be efficacious in AD.

Keywords: Alzheimer’s disease; Excitotoxicity; Ferroptosis; Glutamate/cystine antiporter; Iron dyshomeostasis; Lipid peroxidation.

Fig. 1

Overview of multidisciplinary techniques used in study protocol involving western blotting, total X-ray reflection fluorescence (TXRF) and ¹H nuclear magnetic resonance (NMR) spectroscopy of human medial temporal cortical post-mortem tissue. Abbreviations: trimethylsilyl-[2,2,3,3,-²H₄]-propionate (TSP), deuterium oxide (D₂O), deuterochloroform (CDCl₃) and radioimmunoprecipitation assay buffer (RIPA).

Fig. 2

Typical partial ¹H nuclear magnetic resonance spectrum of the δ 0.80 – 8.70 ppm region (excluding δ 4.50 - 5.80 ppm) of the aqueous phase obtained following dual phase extraction with chloroform, methanol and water of medial temporal cortex of a cognitively normal subject. Resonances from formate and lactate have been truncated. Assignments are based on Chenomx™, literature values (see Supplemental) and from ¹H-JRES and ¹H-¹H COSY data (see Fig. 4, Fig. 5). Abbreviation: γ-aminobutyrate (GABA).

Fig. 3

Typical partial ¹H nuclear magnetic resonance spectrum of the δ 0.75 – 6.00 ppm region of the organic phase obtained following dual phase extraction with chloroform, methanol and water of tissue from the medial temporal lobe of a cognitively normal subject. The overlapping methylene resonances from the fatty acid chains have been truncated. Assignments are based on literature values. Abbreviations: fatty acid chains (FA), polyunsaturated fatty acids (PUFA) and triglycerides (TG).

Fig. 4

Typical partial (A) ¹H-¹H COSY and (B) ¹H-JRES nuclear magnetic resonance (NMR) spectra (with the corresponding 1D-¹H NMR spectrum) of an aqueous phase sample obtained from dual phase extraction of human medial temporal gyrus tissue. Abbreviations include: N-acetyl-aspartate (NAA) and γ-aminobutyric acid (GABA).

Fig. 5

Typical partial ¹H-¹H-COSY nuclear magnetic resonance (NMR) spectrum (with the corresponding 1D-¹H NMR spectrum) of an aqueous phase sample obtained from dual phase extraction of human medial temporal gyrus tissue.

Fig. 6

Expression levels of ferritin-light chain (FTL), ferritin-heavy chain (FTH), ceruloplasmin (Cp), ferroportin (Fpn), cystine/glutamate transporter (Xc-) and 4-hydroxynonenal (4-HNE) adducts in the medial temporal cortex of cognitively normal (CN) and Alzheimer’s disease (AD) subjects, along with representative western blots. Graph presents individual values and the mean ± standard deviation. A p-value ≤ 0.05 was considered significant; with * and ***, p < 0.05 and 0.005, respectively.

Fig. 7

Total Reflection X-ray Fluorescence (TXRF) measurements of elemental concentrations of iron, copper, zinc, calcium and phosphorus in the medial temporal cortex of cognitively normal (CN) and Alzheimer’s disease (AD) subjects. Graph shows individual values and the mean ± standard deviation mg/g protein. The significance threshold was set at *p ≤ 0.05. Abbreviation: not significant, ns.

Fig. 8

Levels of selected metabolites obtained from dual phase extraction of medial temporal cortical tissues from cognitively normal (CN) and Alzheimer’s disease (AD) subjects. Graphs shows individual values and the mean ± standard deviation μmol/g wet tissue. Significance was set at a threshold of p ≤ 0.05, with * and ** being p < 0.05 and 0.01, respectively. Abbreviation: not significant, ns.

Fig. 9

Heatmap showing the Pearson’s correlation coefficients between metabolites, metals/elements and proteins measured in the cognitively normal (CN) and Alzheimer’s disease (AD) groups. Abbreviations are: γ-aminobutyrate (GABA), N-acetylaspartate (NAA), polyunsaturated fatty acids (PUFA), ferritin-light chain (FTL), ferritin-heavy chain (FTH), transferrin-receptor (TfR), divalent metal transporter 1 (DMT1), Iron Responsive Element Binding Protein 2 (IREB2), ceruloplasmin (Cp), ferroportin (Fpn), heme-oxygenase-1 (HO-1), melanotransferrin (MTf), lactoferrin (LTf), Nuclear factor erythroid 2-related factor 2 (Nrf2), Acyl-CoA Synthetase Long Chain Family Member 4 (ACSL4), the light-subunit of the cystine/glutamate antiporter (xCT), glutathione peroxidase 4 (GPX4) and 4-hydroxynonenal (4-HNE). Note, unsaturation (CC) refers to quantification of –CH2-CHCH-CH2- of fatty acid chains.

Fig. 10

Significant and trend Pearson correlations of γ-aminobutyrate (GABA) with metals/elements and metabolites in cognitively normal and Alzheimer’s disease. Significance is set at p≤0.05, with *, ** and *** being p<0.05, <0.01 and < 0.005, respectively. Abbreviation: N-acetylaspartate (NAA).

Fig. 11

Significant and trend Pearson correlations of iron with proteins, elements/metals and metabolites in cognitively normal and Alzheimer’s disease. Significance is set at p ≤ 0.05, with * and ** being p<0.05, <0.01, respectively. Abbreviations: ferritin-heavy chain (FTH), transferrin-receptor (TfR), ceruloplasmin (Cp) and heme-oxygenase-1 (HO-1).

Fig. 12

Overview of ferroptotic-associated changes in Alzheimer’s disease. Iron dyshomeostasis is observed in the form of increased ferritin levels and decreased export of iron. The consequent increases in labile iron pool enables redox-active (ferrous) iron to precipitate oxidative stress via Fenton reaction. The upregulated expression of the light-subunit (xCT) of the cystine/glutamate transporter (X_c^-) may contribute to excitotoxicity via NR2B-containing N-methyl-d-aspartate receptors (NMDARs), evidenced by augmented excitatory glutamate to inhibitory γ-aminobutyrate (GABA) ratio, alongside zinc-deficiency and attenuated glutamine levels, culminating in lipid peroxidation and iron-dependent cell death, ferroptosis. Abbreviations are: ferritin-light chain (FTL), ferritin-heavy chain (FTH), transferrin-receptor (TfR), divalent metal transporter 1 (DMT1), Iron Responsive Element Binding Protein 2 (IREB2), ceruloplasmin (Cp), ferroportin (Fpn), heme-oxygenase-1 (HO-1), melanotransferrin (MTf), lactoferrin (LTf), Nuclear factor erythroid 2-related factor 2 (Nrf2), Acyl-CoA Synthetase Long Chain Family Member 4 (ACSL4), glutathione peroxidase 4 (GPX4), 4-hydroxynonenal (4-HNE), polyunsaturated fatty acids (PUFA), excitatory amino acid transporter (EAAT), and lipid peroxides and alcohols (Lipid-OOH and Lipid-OH, respectively).

Fig. 13

Overview of selected metabolic pathways involving glutamate, glutamine and γ-aminobutyrate (GABA). Abbreviations: N-acetylaspartate (NAA), S-adenosyl-methionine (SAM), S-adenosylhomocysteine (SAH), inorganic phosphate (Pi).