PINK1-dependent NFKB signaling contributes to amyloid pathology in Alzheimer disease

ABSTRACT

Mitochondrial dysfunction plays a preponderant role in the development of Alzheimer disease (AD). We have demonstrated that activation of PINK1 (PTEN induced kinase 1)-dependent mitophagy ameliorates amyloid pathology, attenuates mitochondrial and synaptic dysfunction, and improves cognitive function. However, the underlying mechanisms remain largely unknown. Using a newly generated PINK1-AD transgenic mouse model and AD neuronal cell lines, we provide substantial evidence supporting the contribution of PINK1-mediated mitochondrial ROS (reactive oxygen species) and NFKB/NF-κB (nuclear factor kappa B) signaling to altering APP (amyloid beta precursor protein) processing and Aβ metabolism. Enhancing neuronal PINK1 is sufficient to suppress Aβ-induced activation of NFKB signal transduction in PINK1-overexpressed Aβ-AD mice and Aβ-producing neurons. Blocking PINK1-mediated NFKB activation inhibits activities of BACE1 (beta-secretase 1) and γ-secretase, which are key enzymes for cleavage of APP processing to produce Aβ. Conversely, loss or knockdown of PINK1 produces excessive ROS, along with increased phosphorylated NFKB1/p50 and RELA/p65 subunits, APP-related BACE1 and γ-secretase, and Aβ accumulation. Importantly, these detrimental effects were robustly blocked by the addition of scavenging PINK1 Aβ-induced mitochondrial ROS, leading to the suppression of NFKB activation, restoration of normal APP processing, and limitation of Aβ accumulation. Thus, our findings highlight a novel mechanism underlying PINK1-mediated modulation of Aβ metabolism via a ROS-NFKB-APP processing nexus. Activation of PINK1 signaling could be a potential therapeutic avenue for the early stages of AD by combining improving mitochondrial quality control with limiting amyloid pathology in AD.

Introduction

Alzheimer disease (AD) is a progressive neurodegenerative disorder most prevalent among people between 60 and 80 years of age [1]. Clinically characterized by progressive impairment of behavioral and cognitive functions, AD is pathologically associated with cerebral amyloid-β peptide (Aβ) accumulation [2–6]. Other key features of AD pathology include mitochondrial dysfunction, local neuroinflammation, oxidative stress, synaptic dysfunction and electrophilic damage [7]. Amyloid pathology is one of the factors involved in the pathogenesis of AD. Aβ directly or indirectly perturbs neuronal and mitochondrial function [5,8–11]. The underlying mechanisms of Aβ-mediated progressive mitochondrial and synaptic degeneration remain elusive. An effective treatment to rescue such injuries remains unavailable [12].

The transcription factor NFKB/NF-κB (nuclear factor kappa B) has been implicated in immune function, the nervous system development, synaptic plasticity, memory, spatial navigation and cell survival in physiological and pathological settings [13,14]. NFKB is composed of five subunits, RELA/p65, RELB, REL/c-Rel, NFKB1/p105/p50, and NFKB2/p100/52, that bind to the promoter regions of target genes as homodimers or heterodimers, with the NFKB subunits (NFKB1/p50 and RELA/p65) heterodimer being the most common dimer [15]. Studies have shown that reactive oxygen species (ROS) and proinflammatory cytokines-induced NFKB activation play a key role in the progression and pathogenesis of AD [16]. Neuronal degeneration surrounding early plaques from AD patients is associated with activated NFKB [17]. Indeed, ROS activate various downstream signaling molecules such as PRKC/PKC (protein kinase C) and MAPK/p38 (mitogen-activated protein kinase) that further induce nuclear translocation and activation of NFKB and the expression of pro-inflammatory genes [18,19].

Mitochondrial dysfunction and synaptic failure are prominent and early pathological features of AD [3–5,20]. Defective mitophagy induces accumulation of dysfunctional mitochondria, thereby promoting AD pathology. PINK1 (PTEN induced kinase 1) is critical for the maintenance of mitochondrial integrity and function by promoting removal of damaged mitochondria via mitophagy [9,21–27]. We have previously demonstrated that adenovirus delivery of PINK1 protein in Aβ-rich brains resulted in cerebral Aβ reduction through enhanced autophagy signaling and mitochondrial quality control [9]. However, the underlying mechanisms of PINK1-involved Aβ metabolism have yet to be fully explored.

In the present study, we further established the significance of PINK1 function by regulating NFKB signal transduction in Aβ metabolism using the genetically manipulated neuronal PINK1 mouse model and an Aβ-producing AD neuronal cell culture system. The result of this study addresses the following unexplored fundamental questions: 1) Are PINK1 expression levels associated with Aβ pathology and stage of AD pathology in AD-affected brain regions? 2) Does loss of PINK1 function enhance mitochondrial ROS production? If yes, does PINK1-induced excessive mitochondrial ROS activate NFKB? 3) Is PINK1-mediated activation of NFKB linked to abnormal Aβ metabolism by altering APP processing and related cleavage enzyme activities? We believe that answering these questions will significantly enhance our understanding of the role of mitochondrial signaling in Aβ metabolism and amyloid pathology observed in AD brain. Our studies display that PINK1-mediated ROS-NFKB signaling modulates production of Aβ, importantly, mitochondrial Aβ through BACE1 (beta-secretase 1) and the tetrameric γ-secretase complex processing. Together, our findings provide new insight into the contribution of PINK1-mediated NFKB signal transduction to Aβ pathology.

Results

PINK1 deficiency aggravates mitochondrial and cytosolic Aβ accumulation

We and other independent groups have demonstrated downregulation of PINK1 in AD-affected brains and several well-established AD mouse models [9,26,28,29]. Given that mitochondrial dysfunction is associated with Aβ pathology in AD-affected brains, we evaluated the effect of PINK1 on mitochondrial Aβ accumulation. Compared to cortical mitochondria from mAPP mice at 6 and 9 months of age, levels of Aβ40 and Aβ42 were significantly elevated in the cortical mitochondria from pink1 knockout mAPP mice (pink1^−/− mAPP) mice (Figures. 1A-B). This was further confirmed by immunoblotting of mitochondrial fractions, which showed ~ 3 fold increased mitochondrial pool of Aβ in 9-month-old pink1^−/− mAPP mice in comparison with age matched mAPP mice (Figures. 1C-E). Similarly, Aβ levels were also elevated by ~ 4.9 fold in the cytosolic fractions derived from 9-month-old pink1^−/− mAPP mice compared to age-matched mAPP mice (Figures. 1H-J). Intriguingly, cortical mitochondria from pink1^−/− mice revealed significant increases in naturally produced endogenous mouse Aβ 40 (1.30 fold in 12 months old mice, Figure 1F) and Aβ42 levels (1.4–1.5 fold in 9–12 months old mice, Figure 1G), compared to age-matched nonTg cortical mitochondria. Thus, lack of PINK1 increases mitochondrial Aβ accumulation.

Figure 1.

PINK1 deficiency accelerates Aβ accumulation in cortical mitochondrial and cytosolic fractions of Aβ-ad mice. (A-B) ELISA measurement of human Aβ levels in the hippocampi of 6- and 9-months-old mAPP and pink1^−/− mAPP mice. N = 6 mice, 3 male and 3 female mice in each group. ##p < 0.01 vs. 6-months-old mAPP mice and **p < 0.01 vs. 6-months-old pink1^−/− mAPP mice. (C-E) Immunoblotting of mitochondrial fraction from the brains of indicated mice for PINK1 and Aβ using 6E10 antibody to detect the mitochondrial pool of Aβ. (C) the purity of mitochondria fraction from indicated 9-month-old mice brains was verified by immunoblotting for the presence of specific mitochondrial marker TOMM20 and SOD2 and the absence of lysosomal marker LAMP1 and endoplasmic reticulum marker CANX (calnexin). (D-E) Densitometry of immunoblotting of PINK1 (D) and Aβ (E) in mitochondrial fractions normalized by TOMM20. N = 3 mice, 1 ~ 2 male and 1 ~ 2 female mice in each group. (F-G) ELISA measurement of endogenous mouse Aβ levels in the hippocampi of 6, 9 and 12 months old nonTg and pink1^−/− mice. N = 6 mice, 2 ~ 4 male and 2 ~ 4 female mice in each group. #p < 0.05 vs. 6 months old nonTg mice and **p < 0.01 vs. 6 months old pink1^−/− mice. (H-J) Immunoblotting of cytosolic fraction from indicated mice brains for PINK1 and Aβ in H. Quantifications of immunoblotting of PINK1 (I) and Aβ (J) in the cytosolic fraction, normalized to ACTB/β-actin. n = 3 mice, 1 ~ 2 male and 1 ~ 2 female mice in each group.

Effect of neuronal PINK1 on amyloid pathology in PINK1 mAPP mice

We next assessed the impact of neuronal PINK1 on Aβ pathology in AD mice in vivo by creating transgenic mice (Tg PINK1) with overexpression of human PINK1 in neurons under the control of the Thy1 promoter (Figure S1A). Tg PINK1 mice were identified by PCR genotyping using primers specific for PINK1 (Figure S1B). Immunoblotting using antibody specific to PINK1 showed that PINK1 levels were elevated by ~ 2.5–3 fold in the cortical homogenates of Tg PINK1 mice compared to nonTg littermates (Figure S1C). PINK1 expression was increased in cortical neurons and mitochondria of Tg PINK1 mice, as shown by extensive colocalization of PINK1 with mitochondrial protein HSPD1/Hsp60 under confocal microscopy (Figure S1D).

To determine the relationship of PINK1 expression with amyloid plaque deposition, we visualized Aβ plaques and PINK1 expression patterns by double immunostaining of Aβ and PINK1 under confocal microscopy (Quantification data for amyloid plaque load in Figure 2A and PINK1 expression in Figure 2B. Representative images in Figure 2C). PINK1 mAPP mice displayed a ~ 85% reduction in Aβ deposition (Figures. 2A, C) compared to mAPP mice. The degree of Aβ deposition was negatively correlated with PINK1 expression in mAPP and PINK1 mAPP hippocampi (R² = 0.8464 in mAPP, Figure 2D; and R² = 0.6788 in PINK1 mAPP, Figure 2E), suggesting the significance of neuronal PINK1 in amyloid pathology.

Figure 2.

Negative correlation of PINK1 expression with Aβ aggregation Aβ-ad mice. (A-C) Double immunostaining of Aβ (red) and PINK1 (green) showed significant Aβ deposits in 12-month-old Tg mAPP hippocampi (up panel), while increased PINK1 expression and reduced size of Aβ deposits were observed in the hippocampi of Tg PINK1 mAPP mice (below panel). Scale bars: 100 μm. Quantifications of the size of Aβ plaque were shown in A and PINK1 staining in B and representative immunofluorescence images in C. N = 5 mice, 2 ~ 3 male and 2 ~ 3 female mice in each group. (D-E) Correlation analysis of the relationship between the size of Aβ plaque and areas occupied by PINK1 immunofluorescence in hippocampi of Tg mAPP in D and PINK1 mAPP mice in E.

Upregulated PINK1 has been demonstrated to induce cerebral Aβ degradation through enhanced autophagy signaling [9,30]. PINK1 has been implicated to initiate the mitophagy process via phosphorylating mitochondrial preexisting ubiquitin (Ub) at Ser65 (p-Ser65-Ub) [31]. In the mitochondrial fraction of AD mice with PINK1 deficiency, we detected a downregulation of lysosomal-autophagic signaling as compared to AD mice (Figure S2A-D). Restoring PINK1 level significantly upregulated p-Ser65-Ub expression in AD mice (Representative images in Figure S2E and quantification data for amyloid plaque load and p-Ser65-Ub expression in Figure S2F and S2G, respectively). There is a negative correlation between p-Ser65-Ub expression and Aβ deposition in mAPP (Figure S2H) and in PINK1 mAPP mice (Figure S2I), indicating that phosphorylation of ubiquitin (Ub) at Ser 65 is involved in the impact of PINK1 on amyloid pathology.

We then evaluated whether increased PINK1 expression would rescue mitochondrial dysfunction in mAPP mice. Consistent with our previous studies, mAPP mice displayed a ~ 50% decrease in mitochondrial complex IV activity (Figure 3A) and ATP levels (Figure 3B), whereas PINK1 mAPP hippocampi had almost completely restored complex IV activity and ATP levels compared to age-matched nonTg mice (Figures. 3A-B). Furthermore, PINK1 overexpression in mAPP mice improved cognitive impairment in mAPP mice as assessed by the Morris Water Maze (MWM, Figs. 3C-F). Compared to mAPP mice, PINK1 mAPP mice displayed decreased latencies for locating the hidden platform during the training session (Figure 3C) and an increase in both number of crossings of the target and time spent in the target quadrant in the probe test (Figures. 3D-F). These data suggest that increased PINK1 expression/activity attenuates mitochondrial dysfunction and improves learning and memory in mAPP mice.

Figure 3.

PINK1 attenuates mitochondrial defects and improves learning memory in Aβ-ad mice. (A-B) Cytochrome c oxidase (CcO, complex IV, A) activity and ATP levels (B) in hippocampi of the indicated 12-month-old Tg mice. (C-F). Effect of PINK1 on learning and memory in mAPP mice. C. Mean escape latency to the hidden platform during each day of the training session. p < 0.01, mAPP mice compared to nonTg, PINK1 and PINK1 mAPP mice. D. Mean number of crossings of the target during probe trials. E. Time spent in the quadrant with the hidden platform in probe trials. F. The pattern of representative searching traces for the indicated Tg mice in search of the target. N = 7 ~ 9 mice, 1 ~ 6 male and 2 ~ 7 female mice in each group.

PINK1-dependent activation of NFKB signaling in Aβ-producing neurons

We next investigated the possible mechanism underlying PINK1-mediated amyloid pathology via alteration of Aβ production. Oxidative stress induces activation of transcription factor NFKB. In view of the protective effects of PINK1 on mitochondrial stress and ROS production and the involvement of NFKB signaling in APP processing, we assessed whether PINK1 activates NFKB transcription factor and whether PINK1-mediated NFKB activation contributes to Aβ production and accumulation. To do so, we first investigated the effect of PINK1 on Aβ-induced phosphorylation of NFKB1/p50 and RELA/p65 subunits. Levels of phosphorylated NFKB1/p50 (Figures. 4A-B) and RELA/p65 (Figures. 4A, C) were robustly increased by 4.2 ~ 7.0 fold in Aβ-overexpressed mAPP mice compared to nonTg mice. However, PINK1 overexpression remarkably suppressed the elevations of phosphorylated NFKB1/p50 and RELA/p65 in mAPP mice by 80–90% (Figures. 4A-C). Conversely, pink1^−/− mAPP mice further increased phosphorylated NFKB1/p50 and RELA/p65 by 1.8 ~ 2.1 fold relative to mAPP mice (Figures. 4D-F).

Figure 4.

Effect of PINK1 on the phosphorylation of NFKB1/p50 and RELA/p65 subunits in Aβ-ad mouse brain. (A-C) Effect of PINK1 overexpression on the phosphorylation of NFKB1/p50 and RELA/p65 subunits in Aβ-ad mouse brain. Immunoblotting of the hippocampal homogenates for phospho-NFKB1/p50 and RELA/p65 expressions in the indicated Tg mice at 12 months of age. Quantifications of immunoblotting of phospho-NFKB1/p50 in B and RELA/p65 in C. (D-F) Effect of PINK deficiency on the phosphorylation of NFKB1/p50 and RELA/p65 subunits in Aβ-ad mouse brain. Immunoblotting of phospho-NFKB1/p50 and RELA/p65 expression in the hippocampi of the indicated Tg mice at 6 months of age. Quantifications of immunoblotting of phospho-NFKB1/p50 in E and phospho-RELA/p65 in F. N = 3 mice, 1 ~ 2 male and 1 ~ 2 female mice in each group.

To further verify the effect of PINK1 on Aβ-induced NFKB activation, we assessed nuclear translocation of phosphorylated NFKB1/p50 and RELA/p65 subunits in Aβ-producing neurons by high-resolution immunostaining images. Compared to nonTg cortical neurons, the intensity of phospho-NFKB1/p50 and RELA/p65 was significantly increased in cortical neurons of mAPP mice by ~ 3 fold at age of 9 months (Figures. 5A-B and Figure S3A) and by 1.7 ~ 2 fold at 6 months old (Figures. 5D-E and Figure S3B). Clearly, increased phosphorylated NFKB1/p50 (Figures 5A, 5C and Figure S3A) and RELA/p65 (Figures 5D, 5F and Figure S3B) were mainly localized in nuclei of mAPP cortical neurons, demonstrating increased translocation of NFKB to nucleus which is a sign of NFKB activation. Notably, PINK1 mAPP brains almost completely blocked upregulation of phosphorylated NFKB1/p50 and RELA/p65 levels (Figures. 5A-B and Figure S3A) and nuclear translocation/staining signals (Figures 5A, 5C and Figure 3A). Conversely, pink1^−/− mAPP mice significantly increased phosphorylated NFKB1/p50 and RELA/p65 expression (Figures. 5D-E and Figure S3B) and their nuclear localization compared to mAPP mice (Figures 5D, 5F and Figure S3B), suggesting that loss of PINK1 enhances Aβ-induced activation of NFKB. There were no significant changes in NFKB1/p50 and RELA/p65 expression (Figures. 5D-E and Figure S3B) and translocalizations into the nucleus in pink1 null mice vs. nonTg mice at 6 months old mice (Figures 5D, 5F and Figure S3B). These data indicate that PINK1 restoration inhibits, while loss of PINK1 activates, NFKB signal transduction in Aβ-producing neurons. Thus, there is a link between PINK1-mediated NFKB pathway activation and Aβ metabolism.

Figure 5.

Effect of PINK1 on NFKB nuclear activation and translocation in the cortical neurons of Aβ-ad mouse brain. (A-C) Effect of PINK1 overexpression on NFKB nuclear activation and translocation in the cortical neurons of Aβ-ad mouse brain. (A) Representative immunofluorescent staining images of phospho-NFKB1/p50 and RELA/p65 and MAP-2 in cortical neurons of the indicated Tg mice at 12 months of age. (B-C) Quantification of total (B) and nuclear (C) phospho-NFKB1/p50 and RELA/p65 fluorescent intensity in the cortex of the indicated 12-month-old Tg mice. (D-F) Effect of loss of PINK1 on NFKB nuclear activation and translocation in the cortical neurons of Aβ/AD mouse brain. (D) Representative immunofluorescent staining images of phospho-NFKB1/p50 and RELA/p65 and MAP-2 in the cortical neurons of the indicated 6-month-old Tg mice. (E-F) Density of total (E) and nuclear (F) phospho-NFKB1/p50 and RELA/p65 fluorescent intensity in the cortex of the indicated 6-month-old Tg mice. The panels of A and D merge images denote co-localization of immunofluorescence of phospho-NFKB1/p50, RELA/p65 and nuclear staining by DRAQ-5 dye. MAP-2 reveals neuronal localization. Scale bars: 50 μm. N = 5 mice, 2 ~ 3 male and 2 ~ 3 female mice in each group.

PINK1-dependent NFKB signaling contributes to Aβ production

To determine the contribution of PINK1-mediated NFKB signal to Aβ production, we assessed whether blocking PINK1-induced NFKB activation reduced Aβ levels. N2a-APPsw cells were treated with two different siRNAs, each targeting a different region of Pink mRNA to determine the effect of Pink1 knockdown on NFKB activation and Aβ accumulation. PINK1 expression levels were reduced by ~ 80% in cells treated with siRNA-Pink1 compared to vehicle-treated control cells (Figure S4A). In parallel, cells knocked down for PINK1 by siRNA-Pink1 transfection significantly upregulated phosphorylated NFKB1/p50 and RELA/p65 levels (Figures. 6A-C), demonstrating that defective PINK1 activates NFKB activation. Total NFKB1/p50 or RELA/p65 expression level was not altered in PINK1 knockdown cells (Figures. 6D, E). Accordingly, cellular Aβ40 and Aβ42 levels were higher in siRNA-Pink1-treated cells than control siRNA-treated cells (Figures. 6F-G). Secondly, to validate PINK1-mediated Aβ accumulation dependents on NFKB activation, N2a-APPsw cells with reduced levels of PINK1 were co-transfected with siRNA-Nfkb to silence the Nfkb gene. As expected, APP cells transfected with siRNA-Nfkb reduced the expression of NFKB1/p50 and RELA/p65 subunits by 80% (Figures. 6A-E). Importantly, knockdown NFKB not only suppressed phosphorylation of NFKB1/p50 and RELA/p65 (Figures. 6A-C), but also abolished elevation of cellular Aβ40 and Aβ42 resulting from knockdown of PINK1 in siRNA-Pink1-treated cells relative to the control siRNA-treated cells (Figure 6F-G). Together, these results indicate that activation of NFKB regulated by PINK1 is responsible for Aβ production and metabolism.

Figure 6.

NFKB signaling pathway contributes to Aβ accumulation in N2a-APPsw cells in vitro. (A-E) the efficiency of Nfkb siRNA transfection in N2a-APPsw cells was evaluated by immunoblotting analysis. The expression levels of phospho-NFKB1/p50 (A-B), phospho-RELA/p65 (A, C), NFKB1/p50 (A, D), and RELA/p65 (A, E) were reduced by Nfkb siRNA transfection in N2a-APPsw cells with or without Pink1 siRNA co-transfection. PINK1 knockdown by Pink1 siRNA transfection induced increases in phospho-NFKB1/p50 (C-D) and phospho-RELA/p65 (A, C). The levels of total NFKB1/p50 (A, D) and RELA/p65 (A, E) were also significantly inhibited by Nfkb siRNA transfection in the indicated N2a-APPsw cells. N = 3 in each group. (F-G) ELISA analysis for measurement of Aβ levels (F for 40 and G for 42) in the N2a-APPsw cells with Pink1 siRNA and/or Nfkb siRNA transfection(s). N = 8 in each group.

PINK1-dependent NFKB signaling regulates BACE1 and γ-secretases activities in Aβ-producing neurons

Since Aβ peptides are generated through sequential proteolytic cleavage of APP by two secretase entities, BACE1 and the tetrameric γ-secretase complex [32], we analyzed the effect of PINK1 on BACE1 and γ-secretase. We performed an in vitro cleavage assay and followed by immunoblotting to detect the fragments of CTFβ (~11 KD) and CTFγ (6 KD) arisen from BACE1 or γ-secretase cleavage of full-length APP that are the indexes of BACE1 and γ-secretase activity, respectively. The intensity of immunoreactive bands migrating at ~ 11 KD and 6 KD corresponding to APP-CTFβ and CTFγ were robustly increased in the cortex of mAPP mice at 6 and 9 months of age by 11.5–15 fold, respectively, compared to nonTg mice (Figures. 7A-B), indicating increased BACE1 and γ-secretase activities in Aβ-producing neurons. However, compared to mAPP mice, levels of APP-CTFβ and CTFγ were significantly reduced by 80.7% and 74.6% respectively in PINK1 mAPP mice (Figure 7A). In contrast, PINK1-deficient mAPP mice had significantly elevated CTFβ and CTFγ by 2.31- and 2.39-fold as early as 6 months of age, respectively (Figure 7B).

Figure 7.

PINK1 regulates BACE1 and gamma secretase expression through NFKB signaling pathway. (A-B) Densitometry of APP C-terminal fragment cleavage by BACE1 and gamma secretase (APP-CTFβ/γ, open bars for CTFβ and solid bars for CTFγ) in the hippocampi of PINK1 mAPP mice at 12 months of age in A and PINK1-deficient mAPP mice (pink1^−/− mAPP) at 6 months of age in B. Decreased CTFβ and CTFγ levels in hippocampal tissues from Tg PINK1 mAPP mice as compared to Tg mAPP mice in A. Upregulated CTFβ and CTFγ in hippocampal tissues assessed by APP-CTFβ and APP-CTFγ from membrane-associated APP in Tg mAPP mice with pink1 knockout (pink1^−/− mAPP) in comparison to the Tg mAPP mice in B. N = 3 mice, 1 ~ 2 male and 1 ~ 2 female mice in each group. (C-D) APP-CTFβ and APP-CTFγ (open bars for CTFβ and solid bars for CTFγ) in the N2a-APPsw cells with lentivirus delivery of Pink1 and triple-mutant Pink1 (mPINK1) in C, or Pink1 siRNA and/or Nfkb siRNA transfection(s) in D. Decreased CTFβ and CTFγ levels in the N2a-APPsw cells with lentivirus delivery of PINK1, but not mPinkk1 compared to N2a-APPsw cells with vehicle treatment in C. Upregulated CTFβ and CTFγ levels in the N2a-APPsw cells with Pink1 siRNA, and downregulated CTFβ and CTFγ levels in the N2a-APPsw cells with Nfkb siRNA and Pink1 siRNA co-transfections in comparison with the cells with Control siRNA transfection in D. N = 3 in each group.

To confirm the involvement of PINK1 in the modulation of BACE1 and γ-secretase activities in mAPP mice in vivo, we examined the effect of PINK1 overexpression on BACE1 and γ-secretase levels and activities in the N2a-APPsw cells overproducing Aβ. High levels of CTFβ and CTFγ expressions were observed in N2a-APPsw cells in comparison with N2a-Ctrl cells (Figure 7C). Introduction of PINK1 with lentivirus infection significantly reduced both CTFβ and CTFγ expressions (Figure 7C), while mPINK1 has no effect on CTFβ and CTFγ (Figure 7C) and their enzyme activities (Figure S6H-I). The efficacy of lentivirus-PINK1 or -mPINK1 infection was evaluated by immunoblotting, showing increased expressions of PINK1 in lentivirus-PINK1 or -mPINK1 infected cells by ~ 2.3 fold. The expression levels of PINK1 were comparable between PINK1- and mutant PINK1-infected cells (Figure S4B). These results indicate that PINK1 mediates Aβ production and accumulation through modulation of APP BACE1 and γ-secretase activity.

To determine whether PINK1 modulates BACE1 and γ-secretase activities via NFKB signaling pathway, we assessed the effect of NFKB on PINK1-induced BACE1 and γ-secretase activity. As shown in Figure 7D and Figure S4C-D, siRNA-Pink1-treated N2a-APPsw cells significantly increased CTFβ and CTFγ immunoreactive bands and their enzyme activities compared to control siRNA-treated N2a-APPsw cells. However, co-transfection of siRNA-Nfkb with siRNA-Pink1 reduced both CTFβ and CTFγ immunoreactive bands (Figure 7D), and their enzyme activities (Figure S4C-D) relative to N2a-APPsw cells with Pink1-knockdown alone. Thus, blockade of NFKB suppresses defective PINK1-mediated upregulation of BACE1 and γ-secretases activities insulted by Aβ. Loss of PINK1 accelerates amyloidogenic APP processing through activation of NFKB signaling, which results in Aβ overproduction and amyloid pathology.

Similarly, in wild-type APP (N2a-APPwt) cells, introduction of PINK1, but not mPINK1, with lentivirus infection significantly inhibited both CTFβ and CTFγ levels (Figure S4E). NFKB-knockdown with transfection of Nfkb siRNA, also decreased both CTFβ and CTFγ levels in N2a-APPwt cells (Figure S4F), or the elevated CTFβ and CTFγ expressions in N2a-APPwt cells triggered by PINK1-knockdown (Fig. S4G). The data obtained from wild-type N2a-APP cells further confirm that PINK1 regulates BACE1 and γ-secretase activities through NFKB signaling pathway.

Conversely, application of BACE1 and γ-secretase inhibitors significantly decreases the upregulated phosphorylated-NFKB levels mediated by PINK1-knockdown in N2a-APPsw cells (phospho-NFKB1/p50: Fig. S4H, I, and phospho-RELA/p65: Fig. S4H, J), indicating that secretases also actively regulate NFKB phosphorylation mediated by PINK1.

Scavenging ROS reduces PINK1-mediated Aβ accumulation

We determined whether loss of PINK1 enhanced cellular and mitochondrial Aβ accumulation and if so, whether PINK1-deficiency-induced elevation of Aβ is linked to ROS production. To mimic downregulation of PINK1 expression in AD, PINK1 expression was knocked down in Aβ-producing N2a-APPsw neuronal cells by Pink1 siRNA transfection (Fig. S4A and Fig. S5A-B). Cellular levels of Aβ significantly increased in siRNA-Pink1-treated cells compared to control siRNA treatment (Fig. S5A-C). Intriguingly, mito-TEMPO treatment blocked loss-of-PINK1-mediated Aβ accumulation. Accordingly, cells with a reduced level of PINK1 exhibited greatly enhanced ROS production, which was reversed by mito-TEMPO treatment (Fig. S5D-E). These results suggest a link between ROS and PINK1-mediated Aβ accumulation.

Overproduction of ROS enhance Aβ accumulation and NFKB activation

Next, we evaluated the direct effect of ROS on Aβ accumulation and production in N2a-APPsw cells. Treatment with mito-TEMPO to scavenge mitochondrial ROS largely reduces cellular Aβ40 and 42 measured by quantitative ELISA (Fig. S5F-G) and by immunoblotting (Fig. S5H) in human Aβ-producing N2A-APPsw cells in a dose-dependent manner. Immunocytochemistry with specific antibody to Aβ40 or Aβ42 demonstrated that accumulation of Aβ occurs in the mitochondria of N2a-APPsw cells (Yellow staining shows colocalization of MitoGreen with Aβ staining, indicated by arrows in Fig. S5I). Treatment with mito-TEMPO significantly decreased mitochondrial Aβ accumulation (Images in Fig. S5I and quantification in Fig. S5J), and robustly inhibited Aβ-induced activation of NFKB by suppressing phosphorylation of NFKB subunits NFKB1/p50 and RELA/p65 along with reduced Aβ levels (Fig. S5K-L). Furthermore, mito-TEMPO inhibited the upregulated phosphorylated NFKB1/p50 and RELA/p65 levels in N2A-APPsw cells with PINK1 knockdown by siRNA-Pink1 transfection (Fig. S5M-O). These results suggest a link between ROS-PINK1-NFKB signaling and Aβ metabolism.

Involvement of PRKN on PINK1-mediated NFKB phosphorylation

Given that the PINK1-PRKN-mediated pathway has been identified as an essential molecular mechanism of mitophagy [31], we also investigated the effect of PRKN in PINK1-mediated NFKB-APP processing nexus. To determine PRKN-dependent on PINK1-involved activation of NFKB, we knocked down Prkn gene expression with Prkn siRNA infection in PINK1-overexpressed N2a-WT and APPsw cells with downregulation of phospho-NFKB1/p50 (Fig. S6A-B) and phospho-RELA/p65 levels (Fig. S6A, C) by upregulation of PINK1. Figures S6F-G show that Prkn siRNA transfection reduced PRKN expression by 90% by immunoblotting. Importantly, PRKN knockdown partially blocked the inhibitory effect of PINK1-induced BACE1 and γ-secretases activities in N2a-APPsw cells (Fig. S6H-I), indicating that PRKN mediates suppression of NFKB, BACE1, and γ-secretase activities by PINK1 overexpression.

PINK1-mediated phosphorylated-nfkb nexus correlates with the progression of AD pathology

Finally, we investigate the correlations between PINK1-related phosphorylation of NFKB and the progression of AD pathology from human AD patients (Braak score). AD hippocampus displayed a ~ 34% reduction in PINK1 expression (Fig. S7A, B), but 5.3- and 2.6-fold increases in phospho-NFKB1/p50 and phospho-RELA/p65 (Fig. S7A, C and D) expressions in comparison with those in ND hippocampi. The degree of Braak score negatively correlated with PINK1 (R² = 0.6776, Fig. S7G), while positively with phospho-NFKB levels (R² = 0.6232 for phospho-NFKB1/p50 in Fig. S7H and R² = 0.5707 for phospho-RELA/p65 in Fig. S7I). Total NFKB1/p50 or RELA/p65 was not changed in AD hippocampus compared to ND patients (Fig. S7E-F and S7J-K). These results demonstrate a negative association of PINK1 expression with NFKB activation and stage of AD pathology related to the progression of AD.

Discussion

Mitochondrial abnormalities and synaptic degeneration are early pathological features in AD progression [2–5,20]. Excessive accumulation of Aβ sets in motion the pathological cascade of AD. The underlying mechanisms of Aβ-mediated progressive mitochondrial and synaptic degeneration remain elusive and the strategies to rescue such damages remain unavailable. We have previously demonstrated that gene delivery of PINK1 to the hippocampus of a transgenic amyloid AD mouse model alleviates amyloid pathology and mitochondrial and synaptic dysfunction through modulating autophagy signaling [9]. Here, we further validate the involvement of PINK1 on Aβ pathology in newly generated transgenic PINK1-AD mice expressing human PINK1 and Aβ in neurons. Consistent with our previous studies in gene delivery of PINK1 in mAPP mice, gain of neuronal PINK1 function protects against Aβ accumulation, mitochondrial dysfunction, and ROS production. Conversely, lack of PINK1 enhances mitochondrial Aβ accumulation. Most importantly, functionally inactive PINK1 is unable to reverse Aβ accumulation, indicating that PINK1 kinase activity is required for modulation of Aβ metabolism. These results inform us of the possible mechanisms underlying PINK1-mediated Aβ pathology.

In AD, reactive oxygen species (ROS) overproduction induces excessive fragmentation of mitochondria and promotes defective mitophagy [33]. We have previously demonstrated that restoring PINK1 levels in AD activated mitophagy signaling and mitochondrial quality control resulting in Aβ reduction [9,30,34]. Conversely, loss of PINK1 function causes defective lysosomal-autophagic signaling in the cortical mitochondrial fraction of PINK1-deficient AD mice (Figure S2A-D), indicating contribution of PINK1-involved mitophagy to Aβ accumulation and Aβ pathology. In addition to modifying the E3 ubiquitin ligase PRKN as its partner to activate the process of mitophagy, PINK1 has been recently implicated in the mitophagy process via phosphorylating mitochondrial preexisting ubiquitin (Ub) at Ser 65 (p-Ser65-Ub) [31]. P-Ser65-Ub chains serve as PRKN receptors to recruit PRKN from the cytosol to mitochondria, which leads to the phosphorylation of PRKN by PINK1 [35], or recruit primary mitophagy receptors, such as OPTN (optineurin) and CALCOCO2/NDP52 to promote mitophagy [36]. Thus, upregulation of p-Ser65-Ub along with PINK1 expression occurs near Aβ plaques of AD mice and could explain the possible mechanism of the negative correlation between cerebral Aβ accumulation and plaque deposition and PINK1-involved mitophagy signaling (Figure S2E-I).

Aβ is derived from APP (amyloid beta precursor protein) through processing by three proteases, α-secretase, BACE1, and γ-secretase [32]. Cleavage by BACE1 and γ-secretase produces Aβ peptides, while processing by α-secretase is protective as it involves cleaving within the Aβ sequence, thereby preventing the production of Aβ peptides [37,38]. Here, we demonstrate that PINK1 overexpression inhibits BACE1 and γ-secretase activities, while PINK1 deficiency in AD mice results in significant elevation of BACE1 and γ-secretases at an early stage of AD progression (6 months old). Similarly, overexpression of PINK1, but not mutant PINK1, by the lentivirus infection in N2a-APPsw cells also decreases BACE1 and γ-secretase expression and activity, while knockdown of PINK1 elevated levels of both secretases. Thus, PINK1-mediated activation of BACE1 and γ-secretase is responsible for abnormal Aβ production.

NFKB is one of the transcription factors involved in changes of gene transcription associated with receptor activation by extracellular agents. NFKB has diverse functions in the central nervous system, including roles in neuronal plasticity and synaptogenesis [39,40]. Emerging evidence demonstrates the role of NFKB in the processes of aging and neurodegenerative diseases [17,41,42]. NFKB is sequestered in the cytoplasm by inhibitory I kappa B (IKB) proteins in unstimulated cells. Stimulation of cells by diverse inducers causes phosphorylation of IKB, stimulating NFKB translocation to the nucleus. This leads to NFKB induced transcription of target genes. Oxidative stress-mediated activation of transcription factors, including the redox-sensitive activation protein (AP-1) and NFKB, enhances BACE1 expression [43–45]. In fact, the promoter and 5’ untranslated region of the BACE gene contain binding sites for multiple transcription factors, including NFKB; therefore, activation of NFKB by oxidative stress may in turn promote BACE1 expression [46]. We demonstrated that levels of NFKB subunit NFKB1/p50 and RELA/p65 phosphorylation and nuclear translocation were increased in Aβ-insulted N2a-APP cells and mAPP mouse brains as early as 6 months of age and further elevated at 9 months. These results show that Aβ-induced ROS accumulation contributes to NFKB activation.

Conversely, suppressing BACE1 and γ-secretase activations by through enzyme inhibitors block the effect of PINK1-deficiency-induced activation of NFKB. Indeed, BACE1 and γ-secretase inhibitors alone inhibit NFKB activations in N2a-APPsw cells (Figure S4H-L). These results indicate that suppression of BACE1 and γ-secretases activate a negative feedback loop in upregulated NFKB phosphorylation mediated by PINK1 deficiency.

Several in vivo and in vitro studies have implied NFKB-dependent regulation of Aβ production [47–49]. Activated NFKB is found surrounding early plaques in the entorhinal cortex and hippocampus of AD patients, and inhibition of NFKB activation reduces Aβ secretion in vitro [49–53]. Accordingly, we demonstrated significant upregulations of NFKB1/p50 and RELA/p65 associated with amyloid pathology, as observed in AD-affected hippocampus and entorhinal cortex of mAPP mice. In AD-affected brains of human AD patients, downregulation of PINK1 associates with significantly activated NFKB. Importantly, NFKB activation and PINK1 reduction were significantly correlated with the Braak score, stage of AD pathology (Figure S7), indicating the impact of PINK1-NFKB pathway on the progression and pathology of AD.

Overexpression of PINK1, but not inactive PINK1, inhibited Aβ-induced NFKB1/p50 and RELA/p65 activation as shown by reduced phosphorylated NGKB1/p50 and RELA/p65 and their nuclear localization. Conversely, loss of PINK1 in AD mice sustained activation of NFKB1/p50 and RELA/p65, suggesting the impact of NFKB signaling on PINK1-involved Aβ accumulation. Indeed, knockdown of NFKB completely abolished defective PINK1-mediated upregulation of BACE1 and γ-secretases along with cellular Aβ accumulation in N2a-APP cells. Clearly, PINK1 mediates abnormal Aβ production via NFKB-dependent modulating amyloidogenic APP processing.

Progressive accumulation of mitochondrial Aβ promotes and exacerbates mitochondrial perturbation and ROS production [2,6,10,54,55]. Oxidative stress, induced by excessive production of ROS and/or deficiency in the antioxidant system, has been recognized to play a crucial role in the processes of aging and the origin and development of AD [56]. Oxidative stress decreases the activity of α-secretase while promoting the expression and activation of BACE1 and γ-secretase, enzymes critical to the generation of Aβ from APP [57–63]. Sustained Aβ-related toxic stimuli or excessive ROS suppress PINK1 expression [9]. Therefore, we propose that compromised PINK1 function contributes to aberrant mitochondrial function and decreased ability for clearance and degradation of Aβ. Indeed, scavenging mitochondrial ROS blocks Aβ- and H₂O₂-induced suppression of PINK1 expression and ROS overproduction. PINK1 knockdown not only increased mitochondrial ROS production but also significantly elevated Aβ accumulation. Scavenging defective PINK1-mediated mitochondrial ROS by the treatment of mito-TEMPO, a scavenger of mito-ROS, robustly reduced Aβ accumulation (Figure S5A-E), implying a link between oxidative stress and PINK1-mediated-Aβ accumulation. In addition to reducing Aβ levels, scavenging mitochondrial ROS robustly inhibited Aβ-induced activation of NFKB by suppressing phosphorylation of NFKB subunits NFKB1/p50 and RELA/p65 (Figure S5F-O). Together, these results show that the NFKB pathway, controlled by oxidative stress, is crucially involved in the regulatory effect of PINK1 on Aβ accumulations.

PRKN serves as a downstream substrate for PINK1. PRKN amplifies the mitophagy signal triggered by PINK1 and removes damaged mitochondria [64,65]. However, PINK1 has also been reported to be able to activate mitophagy by recruiting the autophagy receptors directly without the involvement of PRKN [21]. Our previous work has demonstrated that PINK1 promotes the clearance of damaged mitochondria by augmenting autophagy signaling in AD mice [9]. In this study, defective PRKN blocks the protective effects of PINK1 overexpression on NFKB activation and BACE1 and γ-secretase activities related to abnormal APP processing. Thus, PRKN and PINK1 coordination plays at least a part in mitochondrial signaling on amyloid pathology.

Conclusions

In summary, using newly generated PINK1-Aβ mice in vivo and Aβ-producing AD neuronal cell lines in vitro, we report a new mechanism for how PINK1-mediated signaling alters Aβ metabolism and processing, leading to cerebral amyloid pathology. We propose that AD risk factors, such as mitochondrial dysfunction and oxidative stress, compromise PINK1 function, which in turn exacerbates mitochondrial damage and/or ROS accumulation/overproduction. Activation of the transcription factor NFKB by oxidative stress interrupts physiologically normal APP processing by abnormally upregulating BACE1 and γ-secretase levels/activities, in addition to deactivating PINK1-mediated autophagy signaling, resulting in Aβ accumulation and hindered Aβ clearance. Thus, enhancing/activating PINK1 function to suppress ROS-NFKB-APP-processing signaling could be a potential therapeutic strategy as it combines enhancement of mitochondrial quality control with elimination of amyloid pathology during the early stages of AD progression.

Materials and Methods

Animal studies

Animal studies were carried out with the approval of the Institutional Animal Care and Use Committees of the University of Kansas-Lawrence and Columbia University in accordance with the National Institutes of Health guidelines for animal care.

Human subjects

We obtained hippocampal tissues from individuals with Alzheimer disease and age-matched, non – Alzheimer disease controls from the New York Brain Bank at Columbia University and the University of Arizona. Detailed information for each of the cases studied is shown in Table S1 (1A: Information for human hippocampal tissues from New York Brain Bank at Columbia University, and 1B: Information for human hippocampal tissues from the University of Arizona) online. We obtained informed consent from all subjects.

Generation of PINK1 mice in an Aβ-rich environment (mAPP mice)

To generate transgenic mice overexpressing human PINK1 in neurons, we created a transgenic expression cassette bearing the full-length human PINK1 gene driven by the Thy1 promoter. A schematic depiction of transgenic cassettes is shown in Figure S1A. To make a DNA construct PTSC21k-HsPINK1 for generating transgenic (Tg) mice, a full-length fragment (1746 bp) containing complete codes for human PINK1 matched to human PINK1 Genebank number AB053323.1 was released by XhoI and subcloned to the vector backbone (pTsKC21k which was pre-inserted with the Thy1 promoter (generously provided by Dr. Herman van der Putten [66]). The founders of Tg Pink1 mice were identified as bearing the transgene by PCR analysis of tail genomic DNA using genotyping primers (5”-GAGACCTGAAATCCGACAACA-3‘ [forward] and Pink2-R − 5’-CAGAAGCAGCCCTGGAATAA-3” [backward]) (Figure S1B). PINK1 levels were increased in cerebral cortex of Tg Pink1 mice compared with nonTg littermates. Increased expression of PINK1 was present in cortical neurons of Tg Pink1 mice (Figure S1C-D).

PINK1 mice were cross-bred with mAPP mice overexpressing a mutant human form of APP (amyloid beta precursor protein) bearing both the Swedish (K670N M671L) and the Indiana (V717F) mutations (APPSwInd, J-20 line, obtained from the Jackson Lab, MMRRC strain #034836-Jax) to generate double transgenic (Tg) mice expressing neuronal PINK1 and mAPP/human Aβ (PINK1 mAPP), single Tg mice (PINK1, or mAPP), and nonTg littermate offspring.

The pink1 knockout (B6.129S4-Pink1^tm1Shn/J, Jackson Laboratory 017,946 [termed KO]) mice in the C57BL6/J background were crossed with mAPP mice (J-20 line) to generate PINK1-deficient mAPP mice (pink1^-/- MAPP) and single Tg mice (mAPP, pink1^-/- and nonTg). Both male and female mice were used for the described experiments. The investigators were blinded to the mouse genotype during experiments.

Cell culture

Neuroblastoma N2a cells transfected with Swedish-mutated human APP (N2a-APPsw) and wild-type APP (N2a-APPwt) cells (Obtained from Dr. Huaxi Xu, Center for Neuroscience and Aging, Burnham Institute for Medical Research, La Jolla, CA, USA) were maintained in medium containing 50% DMEM/F-12 (Fisher 11,330,032) and 50% OptiMEM medium (Fisher 11,058,021) supplemented with 5% fetal bovine serum, 200 μg/ml geneticin (Fisher, G418, 10131035) and penicillin-streptomycin [6,9,54]. N2a-APP cells were transduced with the indicated lentivirus and/or siRNA for 48 h, and then cultured in the same medium for one additional day.

Lentivirus construction of Lenti-PINK1, Lenti-mPINK1, and Lenti-gfp

PINK1 constructs and viral packaging-p FUGW (Addgene 14,883; deposited by David Baltimore) were used to generate lentiviral plasmids expressing wild-type Pink1 (PINK1) and triple kinase dead mutant of PINK1 (K219A D362A D384A, triple-mPINK1) with EGFP at their C terminus as described previously [9].

Blockage of PINK1, NFKB and PRKN with siRNAs transfection

siRNA was introduced by reverse transfection using Lipofectamine RNAiMAX transfection reagent (Thermo Fisher 13,778–030) according to the manufacturer’s protocol. For Pink1, Prkn and Nfkb siRNAs and scrambled siRNA control transfections, cells were plated at 0.5 × 10⁶ cells/well in 6-well plates and were transfected with siRNAs. The siRNAs (Thermo Fisher) were Nfkb1 mouse siRNA1 (s70542), Nfkb1 mouse siRNA2 (s70544), Pink1 mouse siRNA1 (s206144), Pink1 mouse siRNA2 (s87249), Prkn/Park2 mouse siRNA (s78529) and the non-targeting scrambled control RNA (4390843, Silencer® Select Negative Control No. 1 siRNA).

Isolation of mitochondria.

We isolated mitochondria from the cortices of mouse brain as previously described [9,11]. Briefly, mitochondria were isolated by centrifuging brain homogenates at 1,500g for 5 min at 4°C. We adjusted the supernatant to 10% Percoll (Sigma, P1644) and recentrifuged at 12,000g for 10 min. We resuspended the mitochondrial pellet in the isolation buffer (225 mm d-mannitol, 75 mm sucrose [Sigma 84,097], 2 mm K₂HPO₄, 5 mm HEPES, pH 7.2) containing 0.01% digitonin (Sigma, D141) and recentrifuged at 6,000g for 10 min. The protein concentration was determined by the Bio-Rad DC protein assay (Bio-Rad 5,000,006).

Immunoblotting Analysis

Samples were lysed in extraction buffer (10 mm Tris-HCl, pH 7.4, 100 mm sodium chloride, 1 mm EDTA, 1 mm EGTA, 1 mm sodium fluoride, 20 mm sodium pyrophosphate, 2 mm sodium orthovanadate, 1% Triton X-100 (Fisher Scientific, BP151–100), 10% glycerol, 0.1% SDS, 0.5% deoxycholate (ThermoFisher scientific 89,905), 1 mm PMSF (ThermoFisher scientific 36,978 containing protease inhibitor mixture (Calbiochem 539,137; set V, EDTA-free), separated by SDS/PAGE (Invitrogen, 12% Bis-Tris gel), and then transferred to a nitrocellulose membrane (Bio-Rad 1,620,094, USA). After blocking in TBST buffer (20 mm Tris-HCl, 150 mm sodium chloride, 0.1% Tween-20 (Fisher Scientific, BP337) containing 5% (wt:vol) nonfat dry milk (Santa Cruz Biotechnology, sc-2324) for 1 h at room temperature, the membrane was then incubated and gently shaken overnight at 4°C with primary antibodies. This was followed by incubation with the corresponding secondary antibody for 1 h at room temperature. The following antibodies were used in this experiment: rabbit anti-PINK1 (Santa Cruz Biotechnology, sc -33,796; 1:1000), mouse anti-human Aβ 1–17 clone 6E10 (Signet, 9320–02; 1:2000), rabbit anti-LAMP1 (Cell Signaling Technology, 3243; C54H11, 1:1000), rabbit anti-CANX/calnexin (Cell Signaling Technology, 2433; 1:1000), rabbit anti-SOD2 (Enzo Life Sciences, ADI-SOD111; 1:2500), rabbit anti-phospho‐RELA/p65 (Ser536; Cell Signaling Technology, 3033; 93H1, 1:1000), mouse anti-RELA/p65 (Cell Signaling Technology, 6956; L8F6, 1:3000), mouse anti-phospho‐NFKB1/p50 (Santa Cruz Biotechnology, sc -271,908; 1:1000), mouse anti-NFKB1/p50 (E-10; Santa Cruz, sc-8414, 1:3000), rabbit anti-OPTN/optineurin polyclonal antibody (ProteinTech Group 10,837–1-AP; 1:1000), rabbit anti-CALCOCO2/NDP52 (Cell Signaling Technology, 9036; 1:1000) and rabbit anti-LC3A/B (Cell Signaling Technology 12,741; 1:1000).After three washes with TBST, membranes were incubated for 2 h with horseradish peroxidase-conjugated secondary antibody: goat anti-rabbit IgG or goat anti-mouse (Sigma Aldrich, A6154 or A4416, 1:10,000) and developed using enhanced chemiluminescence (ECLThermo Fisher Scientific 32,016). To ensure equal protein loading of the samples, the same membrane was probed with mouse anti-ACTB/β-actin monoclonal (Sigma Aldrich, A5441) or mouse anti-TOMM20 (Santa Cruz Biotechnology, sc -17,764) antibody at 1:10,000 or 1:5000 dilutions.

Immunofluorescent Staining

Brain slices from the indicated Tg mice were subjected to double immuno-staining with rabbit anti-PINK1 (Santa Cruz Biotechnology, sc -33,796, 1:1000) and mouse anti-HSPD1/HSP60 (Enzo Life Sciences, ADI-SPA-806-D; 1:2500) or mouse anti-Aβ (Eli Lilly, 3D6, 1:2000) and rabbit anti-phospho-ubiquitin (Millipore, ABS1513-I; Ser65, 1:1000) and mouse anti-Aβ (Eli Lilly, 3D6, 1:2000), rabbit anti-phospho‐RELA/p65 (Ser536) (Cell Signaling Technology, 3033; 93H1, 1:1000) and mouse anti-MAP2 (Santa Cruz Biotechnology, sc -33,796; 1:5000), mouse anti-phospho‐NFKB1/p50 (Santa Cruz, sc -271,908, 1:1000) and rabbit anti-MAP2 (Thermo Fisher, PA517646; 1:5000) at 4°C overnight, followed by the conjugation of goat anti-mouse Alexa Fluor 594 and goat anti-rabbit Alex Fluor 488 or the conjugation of goat anti-mouse Alex Fluor 488 and goat anti-rabbit Alex Fluor 594 at 1:1000 dilutions. The staining images were taken under a Leica Confocal Microscope.

For the quantification of immunofluorescent staining, in Figure 2C, a total of 28 slides from 5 mice for each group, 5~6 slides (35 µm) from each mouse showing hippocampal area (bregma -1.94 mm to -2.46 mm) according to a standard mouse brain atlas, were used for immunofluorescent staining and quantification. In S2, a total of 10 slides from 3 mice for each group, 3~4 slides (35 µm) from each mouse showing hippocampal area (bregma -1.94 mm to -2.18 mm) according to a standard mouse brain atlas, were used for immunofluorescent staining and quantification. The percentage of area occupied by immunofluorescence was quantified. In Figure 5 and S3, a total of 10 slides from 5 mice for each group, 2 slides (30 µm) from each mouse showing entorhinal cortex (bregma -2.2 mm to -2.5 mm) according to a standard mouse brain atlas, were used for immunofluorescent staining and quantification. The immunofluorescence intensities in Figures 5B,E and the occupied areas co-localized with nuclear staining in Figures 5C,F were quantified. The images were taken at equal exposure for all the different groups under confocal microscopy (Leica) and analyzed by the ImageJ software for fluorescence intensity quantification and Universal MetaMorph Image Program for occupation of quantification (Molecular Devices, Sunnyvale, CA, USA). The investigator was blinded to the mouse genotype.

Aβ measurement

Brain homogenates, mitochondrial pellets or in vitro cultured cells were incubated in 5 M guanidine HCl and 50 mm Tris HCl overnight and then subjected to Aβ concentration detection using human and mouse Aβ1–40 and Aβ1–42 ELISA kits (Invitrogen, KHB3441, KHB3481, KMB3441, and KMB3481) following the manufacturer’s instructions. For quantification of Aβ deposits, brain sections were prepared from 4% paraformaldehyde-fixed brain and stained with antibody 3D6 (Eli Lilly, 3D6,1:2000;) to identify deposits [9–11]. The area of plaque in the hippocampi from multiple sections at the same level was determined by image analysis using Universal MetaMorph Image Program (Molecular Devices, Sunnyvale, CA, USA).

Cytochrome c oxidase (CCO) activity assay

The CCO activity of mitochondrial fraction was measured as previously described [10]. Briefly, 60 μg of brain or 100 μg of neuronal homogenates and a suitable volume of enzyme dilution solution were added to 950 μL of assay buffer. The reaction was initiated by the addition of 50 μL of ferrocytochrome c substrate solution. The change in fluorescence at 550 nm was recorded immediately using a kinetic program with a 5-s delay, 10-s interval, and total of six readings on an Ultrospect 3100 Pro spectrophotometer.

ATP levels

ATP levels were determined using an ATP Bioluminescence Assay Kit (Roche 11,699,695,001) following the manufacturer’s instruction as we previously described [10]. Briefly, the brain was quickly removed from anesthetized mice, hippocampi were dissected and immediately placed into liquid nitrogen, and then transferred to −80°C freezer until use. The hippocampus or cultured neurons were homogenized in the lysis buffer provided, incubated on ice for 30 min, and centrifuged at 12,000 g for 10 min. ATP levels were then measured in the subsequent supernatants using a Luminescence plate reader (Molecular Devices). A 1.3 second delay time after substrate injection and 10 s integration time were used.

Levels of BACE1 and γ-secretase

The levels of these enzymes were measured as previously described [67]. Freshly dissected mouse hippocampus was homogenized and the cultured N2a-APPsw cells were collected in homogenization buffer (10 mm MOPS [4-morpholinepropanesulfonic acid], pH 7.0, 10 mm KCL, 1 × complete protease inhibitor cocktail) and centrifuged at 587 g for 15 min at 4°C. The supernatant was collected and centrifuged at 18,407 g for 20 min. The pellet was rinsed once in homogenization buffer and then resuspended in assay buffer (150 mm sodium citrate, pH 6.4, 1 × complete protease inhibitor cocktail, protein concentration 4 μg/μl). Samples were incubated for 2 h at 37°C, and then subjected to immunoblotting analysis. The APP-CTF antibody (polyclonal anti-amyloid-β protein precursor C-terminal 751–770; Millipore 171,610; 1:5000) was used for APP-CTFβ (10 kDa) and CTFγ (6 kDa) identification.

Behavioral test

The Morris Water Maze (MWM) test was performed according to the method described in previous publications [9,10]. During the spatial acquisition session, mice were trained for 5 consecutive days with 4 trials per mouse per day. Maximum time for each trial was capped at 60 s. Escape latency was analyzed by the HVS Image software (2020). On day 6, a probe trial was performed to assess the spatial memory of the mice. The platform was removed from the pool and the mice were allowed to swim freely for 60 s. Traces of the mice swim paths were recorded and analyzed by HVS Image. Investigators were blind to mouse genotypes.

Statistical analysis

All data were expressed as the mean ± SEM. Statistical significance was determined by unpaired t test, one ANOVA according to the number of independent variables using commercially available software, StatView (version 5.0.1, Berkeley, Calif) and GraphPad Prism 9. Fisher’s multiple comparisons test was performed for post hoc comparisons in StatView and GraphPad Prism 9, respectively. p < 0.05 was considered significant.

Funding Statement

The work was supported by the National Institute on Aging.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Consent for publication:

All authors have given their consent for publication.

Availability of supporting data:

All data generated or analyzed during this study are included in this published article [and its supplementary information files].

Authors Contributions

Fang Du, Qing Yu, and Shirley ShiDu Yan: Department of Surgery, Columbia University, New York, NY 10,032; Chyuan-Sheng Lin, Herber Irving Comprehensive Cancer Center, Columbia University, New York, NY 1003; Gang Hu, Higuchi Bioscience Center, University of Kansas, KS 66,407.

Abbreviations

AD

Alzheimer disease

Aβ

amyloid-β peptide

APP

amyloid beta precursor protein

ATP

adenosine triphosphate

BACE1

beta-secretase 1

CcO

cytochrome c oxidase

NFKB/NF-κB

nuclear factor kappa B

PINK1

PTEN induced kinase 1

Ub

ubiquitin

Supplementary Materials

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2025.2463322.