Impaired mitochondrial calcium efflux contributes to disease progression in models of Alzheimer's disease

Abstract

Impairments in neuronal intracellular calcium (_iCa²⁺) handling may contribute to Alzheimer's disease (AD) development. Metabolic dysfunction and progressive neuronal loss are associated with AD progression, and mitochondrial calcium (_mCa²⁺) signaling is a key regulator of both of these processes. Here, we report remodeling of the _mCa²⁺ exchange machinery in the prefrontal cortex of individuals with AD. In the 3xTg-AD mouse model impaired _mCa²⁺ efflux capacity precedes neuropathology. Neuronal deletion of the mitochondrial Na⁺/Ca²⁺ exchanger (NCLX, Slc8b1 gene) accelerated memory decline and increased amyloidosis and tau pathology. Further, genetic rescue of neuronal NCLX in 3xTg-AD mice is sufficient to impede AD-associated pathology and memory loss. We show that _mCa²⁺ overload contributes to AD progression by promoting superoxide generation, metabolic dysfunction and neuronal cell death. These results provide a link between the calcium dysregulation and metabolic dysfunction hypotheses of AD and suggest _mCa²⁺ exchange as potential therapeutic target in AD.

Fig. 1

NCLX expression and mtCU components are significantly altered in AD. a Western blots for proteins associated with _mCa²⁺ exchange in postmortem brains of patients diagnosed with non-familial, sporadic AD, and age-matched controls, n = 7 for both groups. MCU, Mitochondrial Calcium Uniporter; MCUB, Mitochondrial Calcium Uniporter β-subunit; MICU1, Mitochondrial Calcium Uptake 1; MICU2, Mitochondrial Calcium Uptake 2; EMRE, Essential MCU Regulator; NCLX, Mitochondrial Na⁺/Ca²⁺ Exchanger. Voltage-dependent anion channel (VDAC) and oxidative phosphorylation component, Complex V α-subunit (CV-Sα) were used as mitochondrial loading controls. b NCLX mRNA expression in tissue isolated from the frontal cortex of 3xTg-AD mutant mice and age-matched outbred non-transgenic controls (NTg). c Western blots for expression of _mCa²⁺ exchanger in 12-month-old 3xTg-AD mutant mice and age-matched NTg controls, n = 3 for both groups. d NCLX mRNA expression in N2a control + Ad-NCLX, APPswe and APPswe + Ad-NCLX corrected to N2a control. e Western blots for NCLX expression in N2a control, Control + Ad-NCLX, APPswe, and APPswe + Ad-NCLX; representative of three independent experiments. f Mitochondrial Ca²⁺ transients, mean shown as solid line, thin lines display ± SEM; n = 9 for Con, n = 9 for Con + Ad-NCLX, n = 11 APPswe and n = 9 for APPswe + Ad-NCLX. g Quantification of _mCa²⁺ transient peak amplitude. h Percent _mCa²⁺ efflux rate vs. control. i Cytosolic Ca²⁺ transients, mean shown as solid line, thin lines display ± SEM. j Quantification of cytosolic Ca²⁺ peak amplitude. k Representative recordings of mitochondrial Ca²⁺ retention capacity. l Percent change in _mCa²⁺ retention capacity vs. N2a control cells. m Representative traces for basal mitochondrial Ca²⁺ content. n Quantification of _mCa²⁺ content. (n = individual dots shown for each group in all graphs. All data presented as mean ± SEM; ****p < 0.001, **p < 0.01, *p < 0.05; one-way ANOVA with Sidak's multiple comparisons test.) Source data are available as a Source Data file

Fig. 2

Neuronal deletion of NCLX accelerates AD pathology. a Schematic of NCLX knockout 3xTg-AD mutant mouse gene-targeting strategy. b NCLX mRNA expression, corrected to the housekeeping gene, Rps13; expressed as fold change vs. Camk2a-Cre control, n = 4 for all groups. c Western blots for NCLX expression in tissue isolated from the hippocampus of mice. VDAC and CV-Sα, served as mitochondrial loading controls. d, e Y-maze spontaneous alternation test. d Percentage of spontaneous alternation. e Total number of arm entries. f–h Fear-conditioning test. f Freezing responses in the training phase. g Contextual recall freezing responses, h Cued recall freezing responses. i, j Soluble and insoluble Aβ_1–40 and Aβ_1–42 levels in cortex of 12-month-old mice. k Representative immunohistochemical staining for 4G8-reactive β-amyloid; 4× scale bar = 100 μm, 40× scale bar = 50 μm. l Quantification of the integrated optical density area for Aβ immunoreactivity, n = 4 for all groups. m Western blots of full-length APP, ADAM-10, BACE1, PS1, Nicastrin, APH, and tubulin (loading control) for cortex homogenate of 12-month-old mice. n Representative western blots of soluble and insoluble total tau (HT7), phosphorylated tau at residues S202/T205 (AT8), T231/S235 (AT180), T181 (AT270), and S396 (PHF13) in cortex homogenate of 12-month-old mice, n = 3 for all groups. o Representative immunohistochemical staining for total tau (HT7) and phospho-tau S202/T205 (AT8) in hippocampus of mice; scale bar = 50 μm. p, q Quantification of the integrated optical density area of HT7 and AT8 immunoreactivity, n = 4 for all groups. (n = individual dots shown for each group in all graphs. All data presented as mean ± SEM; ****p < 0.001, **p < 0.01, *p < 0.05; one-way ANOVA with Sidak's multiple comparisons test.) Source data are available as a Source Data file

Fig. 3

Neuronal rescue of NCLX expression impedes cognitive decline and AD pathology. a Schematic of tetracycline-responsive (TRE) transgenic construct and neuronal-specific driver, Camk2a-tTA. Neuronal-specific gain-of-function models were crossed with 3xTg-AD mutant mice to generate 3xTg-AD × TRE-NCLX × Camk2a-tTA mice. b NCLX mRNA expression corrected to the housekeeping gene, Rps13, expressed as fold change vs. tTA controls, n = 3 for all groups. c Western blots for NCLX expression and mitochondrial calcium uniporter channel components, tissue isolated from the hippocampus of 2 months old mice. d, e Y-maze spontaneous alternation test. d Percentage spontaneous alternation, e Total number of arm entries. f–h Fear-conditioning test. f Freezing responses in the training phase, g contextual recall freezing responses, h Cued recall freezing responses. i, j Soluble and insoluble Aβ_1–40 and Aβ_1–42 levels in brain cortex of 12 months old mice. k Representative immunohistochemical staining for 4G8-reactive β-amyloid; 4 × scale bar = 100 μm, 40 × scale bar = 50 μm. l Quantification of the integrated optical density area for Aβ immunoreactivity, n = 4 for all groups. m Western blots of full-length APP, ADAM-10, BACE1, PS1, Nicastrin, APH, and tubulin (loading control) in cortex homogenate of 12 months old mice, n = 3 for all groups. n Representative western blots of soluble and insoluble total tau (HT7), phosphorylated tau at residues S202/T205 (AT8), T231/S235 (AT180), T181 (AT270), and S396 (PHF13) in soluble brain cortex homogenate of 12 months old mice, n = 3 for all groups. o Representative hippocampal staining for total tau (HT7) and phospho-tau S202/T205 (AT8) immunoreactivity in 12 months old mice; scale bar = 50 μm. p, q Quantification of HT7 and AT8 integrated optical density area correct to 3xTg-AD × Camk2a-tTA controls, n = 4 for all groups. (n = individual dots shown for each group in all graphs. All data presented as mean ± SEM; ****p < 0.001, **p < 0.01, *p < 0.05; one-way ANOVA with Sidak's multiple comparisons test.) Source data are available as a Source Data file

Fig. 4

Enhancing _mCa²⁺ efflux rescues mitochondrial dysfunction in APPswe cells. a Experimental protocol timeline for N2a maturation and adenovirus encoding NCLX (Ad-NCLX) transduction. b Oxygen consumption rate (OCR) at baseline and following: oligomycin (oligo; Complex V inhibitor; to uncover ATP-linked respiration), FCCP (protonophore to induce maximum respiration), and rotenone + antimycin A (Rot/AA; complex I and III inhibitor for complete ETC inhibition). c Quantification of basal respiration (base OCR – non-mito respiration (post-Rot/AA). d Quantification of ATP-linked respiration (post-oligo OCR−base OCR). e Maximum respiratory capacity (post-FCCP OCR−post-Rot/AA). f Spare respiratory capacity (post-FCCP OCR−basal OCR). g Proton leak (post-Oligo OCR−post-Rot/AA OCR). h Quantification of Cell Rox green fluorescent intensity (total cellular ROS production); fold change vs. N2a controls, n = 8 for all groups. i Quantification of DHE fluorescent intensity; fold change vs. N2a controls. j Quantification of MitoSOX fluorescent intensity; fold change vs. N2a controls, n = 52 for N2a control, n = 59 APPswe, and n = 59 for APPswe + Ad-NCLX. k Western blots of full-length APP, ADAM-10 (α-secretase) BACE1 (β-secretase), PS1, Nicastrin, APH (γ-secretase), and tubulin (loading con). l Quantification of β-secretase activity, n = 3 for all groups. m Quantification of extracellular Aβ_1–40 and Aβ_1–42 levels. n Representative images of intracellular protein aggregates in N2a control, APPswe and APPswe + Ad-NCLX cells stained with proteostat aggresome detection reagent (red) and Hoechst 33342 nuclear stain (blue), scale bars = 20 μm. o Total aggregates per cell, n = 41 for N2a control, n = 62 APPswe and n = 69 APPswe + Ad-NCLX. (P-R) Control, APPswe and APPswe transduced with Ad-NCLX for 48 h were assessed for plasma membrane rupture, Sytox Green, after treatment with: p Ionomycin (Ca²⁺ ionophore, 1–5 µM), q glutamate (-NMDAR agonist, 10–50 µM), r tert-Butyl hydroperoxide (TBH, oxidizing agent, 10–30 µM), n = 4 experiments for each reagent. (n = individual dots shown for each group in all graphs. All data presented as mean ± SEM; ****p < 0.001, **p < 0.01, *p < 0.05; one-way ANOVA with Sidak's multiple comparisons test.) Source data are available as a Source Data file

Fig. 5

Restoration of _mCa²⁺ efflux improves mitochondrial function in AD mutant mice. a Representative traces for _mCa²⁺ retention capacity (CRC). b Percent change in CRC of 3xTg-AD × Camk2a-Cre and 3xTg-AD × NCLX^fl/fl × Camk2a-Cre vs. Camk2a-Cre control. c Representative trace for CRC in Camk2a-tTA, 3xTg-AD × Camk2a-tTA and 3xTg-AD × TRE- NCLX × Camk2a-tTA mice. d Percent change in _mCa²⁺ retention capacity of 3xTg-AD × Camk2a-tTA and 3xTg-AD × TRE- NCLX × Camk2a-tTA vs. Camk2a-tTA control. e DHE staining for ex vivo detection of superoxide production in freshly prepared cortical and hippocampal sections from 12 months old mice. f, g DHE fluorescent intensity, percent change vs. 3xTg-AD × Camk2a-Cre controls. h DHE staining for ex vivo detection of superoxide production in freshly prepared cortical and hippocampal sections from 12 months old mice. i, j Quantification of DHE fluorescent intensity, percent change vs. 3xTg-AD × Camk2a-tTA controls. k Representative images of hippocampal 4-HNE immunohistochemistry to detect lipid peroxidation in 12 months old mice. l Percent change in 4-HNE-integrated optical density area corrected to 3xTg-AD × Camk2a-Cre controls. m Representative images of hippocampal 4-HNE immunohistochemistry to detect lipid peroxidation in 12 months old mice. n Percent change in 4-HNE-integrated optical density area corrected with 3xTg-AD × Camk2a-tTA controls. o Mitochondrial DNA (mtDNA)/nuclear DNA (nDNA) ratio in tissue isolated from the cortex of 2 and 12 months old mice, fold change vs. 2 months old Camk2a-Cre controls. p mtDNA/nDNA ratio in tissue isolated from the brain cortex of 2 and 12 months old mice expressed as fold change vs. 2 months old Camk2a-tTA controls. (n = individual dots shown for each group in all graphs. All data presented as mean ± SEM; ****p < 0.001, **p < 0.01, *p < 0.05; one-way ANOVA with Sidak's multiple comparisons test.) Source data are available as a Source Data file

Fig. 6

Working hypothesis of _mCa²⁺ exchange dysfunction in AD pathogenesis. Alzheimer’s disease (AD) initiators, such as aging and metabolic dysfunction, elicit a compensatory increase in intracellular calcium (_iCa²⁺) and remodeling of the _mCa²⁺ exchange machinery to elevate matrix Ca²⁺ levels and activate mitochondrial dehydrogenases to augment cellular energetics (blue boxes). However, these alterations in _iCa²⁺ handling quickly turn maladaptive leading to _mCa²⁺-overload, which results in mitochondrial dysfunction and AD pathophysiology. The proposed sequence of events initiates positive feedback at multiple levels of the disease pathway potentiating the progression of neurodegeneration