Exposure to CuO Nanoparticles Mediates NFκB Activation and Enhances Amyloid Precursor Protein Expression

Amyloid precursor protein (APP) is directly related to Aβ amyloidosis—a hallmark of Alzheimer’s disease (AD). However, the impact of environmental factors upon APP biology and Aβ amyloid pathology have not been well studied. The increased use of nanoparticles (NPs) or engineered nanomaterials (ENMs) has led to a growing body of evidence suggesting that exposure to metal/metal oxide NPs, such as Fe2O3, CuO, and ZnO, may contribute to the pathophysiology of neurodegenerative diseases such as AD through neuroinflammation. Our previous studies indicated that exposure to CuO nanoparticles (CuONPs) induce potent in vitro neurotoxicity. Herein, we investigated the effects on APP expression in neuronal cells exposed to different metal oxide NPs. We found a low dose of CuONPs effectively activated the NFκB signaling pathway and increased APP expression. Moreover, the inhibition of p65 expression using siRNA abolished CuONP-mediated APP expression, suggesting that NFκB-regulated APP expression in response to CuONP exposure may be associated with AD pathology.


Introduction
Alzheimer's Disease (AD) is a progressive brain disorder, defined by neuronal and synaptic loss in the hippocampus, cerebral cortex, and adjacent regions. Symptoms of the disease consist of memory loss, paranoia, cognitive impairment, confusion, and language deficiencies [1][2][3][4]. AD primarily affects the elderly [5], involving 50% of people over the age of 85 [6], and is the number one cause of dementia [7]. Regrettably, the rising cost of dementia is estimated at $818 billion making it a colossal socioeconomic burden [8]. Early studies of the disease have suggested the onset is associated with the extracellular deposition of abnormal β-amyloid peptides (Aβ) and intracellular accumulation of neurofibrillary tangles (NFT), which are aggregates of hyperphosphorylated tau proteins [6]. In addition to damaged brain tissue and apoptosis [9], clinical investigations have indicated the brains of AD patients exhibit high levels of inflammatory chemokines and cytokines. In addition, there is a notable proliferation of activated microglia [10], a characteristic of excessive inflammatory and innate immune response [11][12][13][14][15][16]. As such, advancements in the field have deemed of the brain, inhibiting the action potentials of hippocampal CA1 neurons [56]. The particles have also been reported to exert a wide array of genotoxic effects on brain cells as well [57]. Nonetheless, it remains unclear whether environmental exposure to CuONPs is involved in the AD pathogenesis and if these ENMs can activate the NFκB pathway or affect APP expression in human neuronal cells.
While it may be unclear how copper nanoparticles contribute to AD, there is compelling evidence that nanoparticles do in fact contribute to AD pathology. Indeed, there certainly seems to be an association, as various metal nanoparticles, including copper, are abnormally abundant in individuals who have Alzheimer's disease, possibly due to a tendency to promote Aβ aggregation [58,59].Young residents of Mexico City, who are exposed to particulate matter in the excessive air pollution in the area, were found to exhibit AD pathological hallmarks in their brainstems in 99.5% of cases; this included infant specimens [60]. Nonetheless, the exposure to NPs can implicate protective roles from aging, thus understanding the mechanism is critical.
Previously we identified an essential need for a model of cell death due to apoptosis pathways associated with debilitating pathologies versus normal physiology. As such, we presented a fully automated cellular image analysis to quantify the viability of human H4 cells exposed to CuONPs [61]. Then, we tested the cytotoxicity of CuONPs, specifically due to their versatile use in medicine and industrial applications, and found that the viability of human H4 and SH-SY5Y cells was negatively affected. We concluded that additional studies were needed to understand the neurotoxic pathways [42,61].
In our previous work, we investigated the molecular mechanism behind the detrimental effects of CuONP exposure and how it may be linked to AD pathology [43]. Herein, to determine the role of NFκB in the regulation of APP gene expression in response to CuONP exposure in human neuronal cells, we utilized the RNAi technique to knockout the expression of the NFκB p65 subunit to characterize its effects on APP gene expression. Lastly, we tested the effects of CuONPs upon NFκB activation and APP expression. Our data indicated that NFκB regulates APP expression in response to CuONP exposure, and CuONP-mediated NFκB activation may be associated with AD pathogenesis.

Nuclear and Cytoplasmic Protein Extraction
Naïve SH-SY5Y cells or SH-SY5Y cells transfected with either control siRNA or p65 siRNA were plated at 2 × 10 6 cells/well (6 well plate) and exposed to CuONPs (<50 nm in diameter, 29 m 2 /g in surface area, Sigma-Aldrich, St. Louis, MO, USA) in the presence or absence of a potent inhibitor of NFκB activation-ammonium pyrrolidine dithiocarbamate (PDTC) at times between 0-60 min. Cells were lysed, and the nuclear protein were extracted using NE-PER Nuclear and Cytoplasmic Extraction Reagent (PIERCE, Rockford, IL, USA). The total protein levels were determined by a bicinchoninic acid (BCA) protein assay kit (Thermo-Fisher, Waltham, MA, USA) and a Spectra Max M5e plate reader (Molecular Devices, San Jose, CA, USA). Equal amounts of protein per sample were added to NuPAGE loading buffer (Invitrogen, Carlsbad, CA, USA), and boiled for 10 min. Samples were electrophoresed on NuPAGE ® 4-12% Bis Tris gels with NuPAGE MES SDS running buffer and transferred to nitrocellulose membranes (Invitrogen, Carlsbad, CA, USA).

Results
We analyzed the activation properties of three different metal oxide NPs (CuO, Fe 2 O 3 , and ZnO) on the NFκB luciferase reporter gene in transfected SH-SY5Y cells. Similar to our previous findings, CuONPs affected cell viability [42] and activated the NFκB reporter 7-fold at a dose of 10 µM. This is higher than the 2-to3-fold activation by Fe 2 O 3 NPs or ZnONPs ( Figure 1). Notably, the observed fold change decreased substantially as the CuONP concentration increased from 10µM to 100 µM; perhaps this is due to the increased disruption from the toxic effects of 100 µM CuONPs, as the particles' toxic and genotoxic effects tend to increase with concentration [57,62]. Our previous study found that CuONPs were highly toxic at concentrations of 100 µM [43]. Other studies have also confirmed that any imbalance of some metal homeostasis induced adverse effects to brain pathology [63][64][65].
We further explored the effects of CuONPs on IκBα protein levels on SH-SY5Y cells exposed to CuONPs (10 µM) in the presence or absence of pyrrolidine dithiocarbamate (PDTC), an NFκB activation inhibitor (50 nM), and found decreased levels of IκBα protein at 30 min with the lowest at 60 min ( Figure 2A). PDTC inhibited the degradation of IκBα in response to CuONP treatment ( Figure 2B,C) at 60 min.
We further explored the effects of CuONPs on IκBα protein levels on SH-SY5Y cells exposed to CuONPs (10 μM) in the presence or absence of pyrrolidine dithiocarbamate (PDTC), an NFκB activation inhibitor (50 nM), and found decreased levels of IκBα protein at 30 min with the lowest at 60 min ( Figure  2A  Effect of nanoparticles on NFκB luciferase reporter activation. Human SH-SY5Y cells (2 × 10 4 /well) expressing a NFκB reporter gene (transfected plasmid-pNFκB-Lu-neo) were exposed to concentrations of nanoparticles (NPs) of CuO, Fe 2 O 3 , and ZnO at a range of 0 to 100 µM/mL for 6 h. Luciferase reagent was added and luciferase activation signals were measured using a Veritas™ microplate luminometer. Results are expressed as fold induction. * p < 0.01, ** p < 0.001, *** p < 0.0001, for CuONPs (10 µM, 100 µM) compared to control, untreated, # p < 0.01 for Fe 2 O 3 NPs compared to control,ˆp < 0.01 for ZnONPs compared to control. Effect of CuONPs on IĸB-α degradation. SH-SY5Y cells were plated at 2 × 10 6 cells/well (6 well plate) and exposed to CuONPs (10 μM) in the presence or absence of the potent NFκB inhibitor-pyrrolidine dithiocarbamate (PDTC, 50 nM) at the indicated time points. Cells were lysed, and lysates were Western blotted for the presence of IκB-α-a protein inhibitor for NFκB activation. Blots were collected, digitized, and quantified using a Bio-Rad VersaDoc™ Digital Imaging System (MP4000). Experiments were performed at n = 3 independent trials and representative Western blots were presented. (A) Western blot from cells exposed to CuONPs but not PDTC; (B) western blot from cells exposed to CuONPs and PDTC; (C) summary graph of relative degradation (compared to controls) in cells exposed to CuONPs and CuONPs and PDTC.
The influence of CuONPs on APP expression was analyzed by administering treatment for 72 h on SH-SY5Y cells. The levels of APP expression were increased up to 5-fold during 6 to 24 h time periods, declining thereafter with a dose level of 10 μM ( Figure 3A). No change in the protein concentration of βactin indicated specificity. The effects of the CuONP dose (0.01-100 μM) on APP protein expression after 6 h of exposure in SH-SY5Y and PC12 cells were detected. TNFα (50 ng/mL) was used as a control for APP induction, and higher levels (10-100 μM) of CuONPs resulted in increased APP expression in SH-SY5Y and PC12 cells ( Figure 3B). CuONPs of 10 μM increased APP in both cell types (SH-SY5Y and PC12). However, a 4-fold induction of APP in SH-SY5Y cells compared to 2-fold induction in PC12 cells was found.  Effect of CuONPs on IkB-α degradation. SH-SY5Y cells were plated at 2 × 10 6 cells/well (6 well plate) and exposed to CuONPs (10 µM) in the presence or absence of the potent NFκB inhibitor-pyrrolidine dithiocarbamate (PDTC, 50 nM) at the indicated time points. Cells were lysed, and lysates were Western blotted for the presence of IκB-α-a protein inhibitor for NFκB activation. Blots were collected, digitized, and quantified using a Bio-Rad VersaDoc™ Digital Imaging System (MP4000). Experiments were performed at n = 3 independent trials and representative Western blots were presented. (A) Western blot from cells exposed to CuONPs but not PDTC; (B) western blot from cells exposed to CuONPs and PDTC; (C) summary graph of relative degradation (compared to controls) in cells exposed to CuONPs and CuONPs and PDTC.
The influence of CuONPs on APP expression was analyzed by administering treatment for 72 h on SH-SY5Y cells. The levels of APP expression were increased up to 5-fold during 6 to 24 h time periods, declining thereafter with a dose level of 10 µM ( Figure 3A). No change in the protein concentration of β-actin indicated specificity. The effects of the CuONP dose (0.01-100 µM) on APP protein expression after 6 h of exposure in SH-SY5Y and PC12 cells were detected. TNFα (50 ng/mL) was used as a control for APP induction, and higher levels (10-100 µM) of CuONPs resulted in increased APP expression in SH-SY5Y and PC12 cells ( Figure 3B). CuONPs of 10 µM increased APP in both cell types (SH-SY5Y and PC12). However, a 4-fold induction of APP in SH-SY5Y cells compared to 2-fold induction in PC12 cells was found. Notably, this induced change in APP expression was not as large as that of NFκB. A study by Lv et al. showed the binding of Cu ions induced structural change in the amyloid dimer, which can induce AD pathology [66].  Moreover, we explored whether the NFκB pathway was involved in increased APP expression in response to CuONP treatment. The effects of CuONPs on nuclear p65 accumulation and cytoplasmic APP expression were detected in response to the incubation of either TNFα-an NFκB activator and inducer of APP expression-as a positive control ( Figure 4A) or CuONPs ( Figure 4B) in the presence or absence of PDTC (50 nM) (Figure 4). Lysates were blotted for the presence of p65, APP, and β-actin (control). TNFα of 50 ng/mL ( Figure 4A) and CuONPs of 10 μM ( Figure 4B) induced the nuclear accumulation of p65 and increased APP expression at 6 h. Additionally, PDTC partially inhibited p65 nuclear accumulation and APP expression in response to TNFα ( Figure 4A) and CuONP exposure ( Figure 4B). We conducted a p65 protein knockout test to confirm the involvement in APP regulation in response to CuONP exposure via p65 SiRNA or a siRNA control. The control lysates contained inducible p65 nuclear accumulation in response to CuONP and TNFα treatments, unlike cells transfected with p65 siRNA ( Figure  5A). Furthermore, β-actin levels were unaffected by the treatment. APP induction in response to CuONP or TNFα exposure was unaffected by the control siRNA. However, the transfection of p65 siRNA resulted in a significant reduction in basal and inducible APP expression in response to CuONP and TNFα treatments, thus confirming the role of NFκB p65 induction in response to CuONPs ( Figure 5B).  We conducted a p65 protein knockout test to confirm the involvement in APP regulation in response to CuONP exposure via p65 SiRNA or a siRNA control. The control lysates contained inducible p65 nuclear accumulation in response to CuONP and TNFα treatments, unlike cells transfected with p65 siRNA ( Figure 5A). Furthermore, β-actin levels were unaffected by the treatment. APP induction in response to CuONP or TNFα exposure was unaffected by the control siRNA. However, the transfection of p65 siRNA resulted in a significant reduction in basal and inducible APP expression in response to CuONP and TNFα treatments, thus confirming the role of NFκB p65 induction in response to CuONPs ( Figure 5B). We conducted a p65 protein knockout test to confirm the involvement in APP regulation in response to CuONP exposure via p65 SiRNA or a siRNA control. The control lysates contained inducible p65 nuclear accumulation in response to CuONP and TNFα treatments, unlike cells transfected with p65 siRNA ( Figure  5A). Furthermore, β-actin levels were unaffected by the treatment. APP induction in response to CuONP or TNFα exposure was unaffected by the control siRNA. However, the transfection of p65 siRNA resulted in a significant reduction in basal and inducible APP expression in response to CuONP and TNFα treatments, thus confirming the role of NFκB p65 induction in response to CuONPs ( Figure 5B).

Discussion
The interaction of Aβ with transition metals such as copper, zinc, and iron results in the aggregation of Aβ and the accumulation of reactive oxygen species (ROS), both of which promote AD pathogenesis. Indeed, metals are often found associated with amyloid plaques [37]. This is based on in vitro studies accelerating the aggregation and precipitation into plaques of Aβ, ultimately leading to synaptic dysfunction and accelerated amyloidogenesis [67,68].
Our current findings, that CuONPs are the more toxic than ZnONPs and Fe 2 O 3 NPs, confirm our previous study [42,61]. As indicated in Figure 1, CuONPs induced the most NFκB activation when concentrations of NPs were above 1 µM. This is because dissolute Cu ions from CuONPs are more redox active than Fe ions from Fe 2 O 3 NPs, and CuONPs also have smaller surface areas than Fe 2 O 3 NPs while dissolute Zn ions from ZnONPs are redox inert. Further analysis of the CuONP effects on NFκB activation in brain cells resulting from attenuated cell viability can produce insight into whether attenuation of NFκB activation may reduce AD pathogenesis [69]. Aβ levels were attenuated by suppressing BACE1 through NFκB deactivation, confirming the role of NFκB in Aβ formation [70]. Thus, modulating the NFκB-mediated neuroinflammatory pathway may be a valid therapeutic approach for AD pathology [71][72][73][74].
Experimental data have shown detrimental effects of NP exposure on brain health, which may lead to etiopathogenesis of AD [75], as exposure to ENMs continues to rise in both industrial and consumer applications [43]. There is overwhelming evidence, in the form of preclinical trials, in vitro and in vivo, which support the roles of metals and metal oxides alike in the pathogenesis of AD [43,76].
Herein, our data supports the argument that environmental exposure to ENMs is injurious to our health. As shown in our controlled in vitro assays, CuONP treatment activated NFκB more than ZnONPs and Fe 2 O 3 NPs in pNFκB-Lu-neo SH-SY5Y cell lines. Other studies have concluded that genetic manipulation of the NFκB pathway led to exhaustive inflammatory action, had serious consequences on aging as found in intracellular negative regulators, and led to toxic accumulations of antimicrobial peptides in flies [77]. Thus, it is critical we evaluate the relationships and toxic effects influenced by various NPs, as no index based on biological response is currently available [41]. It is evident that exposure to ambient particulate matter is a significant risk factor for cognitive decline and dementia [78]. Studies such as ours, which elucidate the mechanisms behind particulate toxicity, are vital to understanding how and where the problem needs to be addressed.
We have demonstrated that CuONPs at 10 µM activated NFkB after 6 h of treatment and enhanced APP expression from 6 to 72 h. At 24 h of treatment, the expression of APP was highest, while APP expression levels decreased at 72 h. Furthermore, APP increased as NFkB increased, indicating NFkB-regulated APP expression.
NFκB is made up of multiple subunits including IκBα, p65, and p50. Activation occurs by proteasome degradation of the inhibitor IκB, followed by entry of p65 and p50 into the nucleus. Our study indicated that exposure to CuONPs caused the degradation of IkBα starting between 30 and 60 min, and later activated NFκB. When NFkB was over expressed, APP levels were also high. On the other hand, when NFκB was inhibited with PDTC, APP expression was inhibited as well in SH-SY5Y and PC12 cells. This is critical in unearthing the role of inflammation and potential pathways to therapies, given the relevance of APP, inflammation, and the interplay between the two with regard to AD [1,10,12]. Overall, we found APP expression was significantly reduced when NFκB was inhibited, suggesting that NFκB influences APP levels. Overall, our data suggests that CuONP exposure induced NFκB-mediated neuroinflammation and increased AD risk. Thus, as current research efforts have failed to yield effective AD treatment so far, further understanding of the molecular mechanisms related to environmental exposure to CuONPs and its potential contribution to AD etiopathogenesis may offer an alternative perspective.

Conclusions
The effects of environmental factors, such as exposure to ENMs, on neurodegeneration are not well understood. Thus, understanding neurodegenerative mechanisms relating to exposure to ENMs, such as CuONPs, in neurological diseases such as AD can provide rationale for the regulation of these materials. In this study, we have determined that low-dose CuONP exposure has a direct role in activating the NFκB signaling pathway and increasing APP expression. Moreover, the inhibition of p65 expression using siRNA abolished CuONP-mediated APP expression, suggesting that NFκB-regulated APP expression in response to CuONP exposure may be associated with AD pathology. Hence, regulating environmental exposure to CuONPs may be necessary to lower AD risk and promote overall healthy brain aging. Funding: This work is partially funded by a grant from the NIH grant number R01AG056614 (to X.H.).

Conflicts of Interest:
The authors declare no conflict of interest.