Simultaneous Imaging of Cu⁺ and Cu^{2 +} in Neural Cells Using DNAzyme Probes Reveals Mechanistic Link Between Copper Redox Imbalance and Amyloid Pathology

ABSTRACT

Copper dysregulation is implicated in neurodegenerative diseases such as Alzheimer's disease (AD), yet its precise role in neuronal death remains unclear. To address this issue, here, we introduce a pair of Cu⁺ and Cu^{2 +} specific DNAzymes‐based fluorescent probes, for the first time, enabling simultaneous visualization of both redox states of copper in single living neurons. Using this dual‐color system, we found that amyloid‐beta (Aβ) oligomerization promotes intracellular copper accumulation, distinct from that induced by artificial ionophore loading. Elevated Cu⁺ drives reactive oxygen species (ROS) generation, lipoylated protein aggregation, and FDX1‐dependent cuproptosis, while Cu⁺ chelation or FDX1 knockdown completely prevents cell death. In contrast, ROS scavengers only partially rescue viability, demonstrating that neuronal death is driven by copper overload, not oxidative stress itself. These findings redefine the mechanistic framework linking copper redox imbalance to Aβ pathology and neuronal vulnerability and demonstrate a selective, sensitive approach for monitoring copper homeostasis and its disruption in neurodegenerative disease.

We develop Cu⁺‐ and Cu^{2 +}‐specific DNAzyme fluorescent probes to visualize both copper redox states in living neurons. Aβ oligomers cause intracellular copper overload, triggering FDX1 dependent cuproptosis and generating ROS. This work links copper redox imbalance with Aβ pathology and offers a selective, sensitive way to monitor copper homeostasis in neurodegenerative disease.

1. Introduction

Copper (Cu) is an essential transition metal that serves as a cofactor for numerous metabolic enzymes and thus plays a critical role in regulating various physiological processes [1]. As a result, disruptions in Cu homeostasis have been implicated in the pathogenesis of several diseases, including cancer, Wilson's disease (WD), and neurodegenerative disorders such as Alzheimer's disease (AD), Parkinson's disease, and Huntington's disease [2]. Under physiological conditions, Cu readily cycles between the Cu⁺ and Cu^{2 +} oxidation states. While this redox flexibility is biologically indispensable, excessive Cu accumulation can cause toxicity through uncontrolled reactive oxygen species (ROS) generation, oxidative stress, and mitochondrial dysfunction [3]. A recently identified Cu‐induced cell death pathway, known as cuproptosis, arises when Cu ions bind to and aggregate lipoylated proteins, leading to proteotoxic stress, tricarboxylic acid (TCA) cycle failure, and elevated ROS [4, 5]. The subsequent oxidative damage and disruption of the mitochondrial membrane contribute to severe cellular dysfunction and ultimately result in cell death [6].

A primary example of Cu and its redox cycles playing important roles is its involvement in AD as Cu accumulation in the brain has been linked to oxidative damage and the aggregation of neurotoxic proteins, including amyloid‐beta (Aβ) peptides [7, 8]. Cu ions can bind to Aβ peptides, facilitating ROS generation and enhancing Aβ aggregation, which in turn increases neurotoxicity [9, 10]. Emerging evidence suggests that Cu dyshomeostasis plays a role in AD neurodegeneration by mediating oxidative stress, promoting Aβ plaque deposition, inducing neuronal death, and impairing synaptic function [6, 11, 12, 13, 14]. Recent studies have reported that cuproptosis in neurons is driven by oxidative stress caused by the redox cycling between Cu⁺ to Cu^{2 +}, leading to proteotoxic stress and mitochondrial dysfunction [15]. However, the mechanistic bridge between the Cu⁺ and Cu^{2 +} redox cycle, Aβ pathology, and cuproptosis in AD neurodegeneration remains unresolved, largely because it has been technically impossible to simultaneously monitor Cu⁺ and Cu^{2 +} in the same living neuron.

To detect and quantify Cu and other metal ions, several instrumental techniques, including inductively coupled plasma mass spectrometry (ICP‐MS) [16], x‐ray fluorescence microscopy (XFM) [17] and nanosecondary ion mass spectrometry (Nano‐SIMS) [18] have been employed. While these methods can provide accurate information on the metal ions, they are not suitable for real‐time tracking of Cu ions in living cells. Consequently, there is a growing interest in developing selective sensors that enable direct visualization of Cu ions in live‐cell environments.

Over the past two decades, numerous fluorescent probes have been designed for either Cu⁺ [19, 20, 21, 22, 23, 24, 25] or Cu^{2 +} [22, 26, 27, 28] in biological systems. Cu⁺ sensors such as Coppersensor‐1 and its successors enabled live‐cell imaging of labile copper pools [29]. Notable among these is a Förster resonance energy transfer (FRET)‐based probe FCP‐1 for labile Cu⁺ [23] and a separate fluorescent sensor for labile Cu^{2 +} [30], demonstrated by the Chang group, which together established live‐cell imaging of specific Cu oxidation states separately. Most recently, Chang and colleagues reported an alkyne‐directed, activity‐based Cu(I) probe that operates under O₂‐independent conditions and revealed Mn(II)‐mediated sensitization of cuproptosis, further expanding the copper imaging toolkit [31]. While many insights have been gained from separate experiments using either a Cu⁺ or Cu^{2 +} sensor, there is still a need to monitor both Cu⁺ and Cu^{2 +} simultaneously in the same living cells.

To overcome this limitation, surface‐enhanced Raman Scattering (SERS) probes have been developed for the simultaneous detection of Cu⁺ and Cu^{2 +} [15, 32, 33]. These nanoplasmonic platforms enable ratiometric, oxidation‐state‐resolved readouts in living systems. For example, the Tian group reported a single SERS probe to provide real‐time tracking and sensing of Cu⁺ and Cu^{2 +} in the live brain and uncovered ischemia‐linked increases in extracellular copper with mechanistic routes for Cu⁺/Cu^{2 +} elevation [32]. Beyond SERS, the Chang group has reported an activity‐based SRS strategy that duplex‐maps labile copper pools in live cells using copper‐directed acyl‐imidazole chemistry (CRP2181 for Cu(I/II) and CRP2153.2 for Cu(II)), expanding Raman‐based copper imaging to standard SRS microscopes [34]. While the above SERS/SRS studies demonstrate the power of simultaneous monitoring of Cu⁺ and Cu^{2 +} in gaining insight into the role of Cu ions in living cells, these methods require specialized instrumentation not routinely available in biological laboratories. It is desirable to develop fluorescent sensors as fluorescent microscopes are more readily available to biomedical researchers. In addition, since the distributions of Cu ions in different biological systems are diverse and complex, alternative approaches are needed to provide complementary information and a more comprehensive understanding.

Toward the goal of developing alternative methods for simultaneous imaging of Cu⁺ and Cu^{2 +} with high sensitivity and fast response in data collection, we are interested in leveraging DNAzymes (also called deoxyribozymes), a class of catalytic DNA molecules that recruit metal ions as cofactors for catalysis [35]. Unlike small‐molecule or protein‐based sensors, DNAzymes can be selected in vitro for optimal sensitivity and selectivity toward not only different metal ions but also different oxidation states of the same metal ions [36]. Since their discovery in 1994, DNAzymes have emerged as a versatile platform for metal ion sensing due to their high specificity, cost‐effective synthesis, and biocompatibility [37, 38, 39, 40]. More importantly, unlike small molecular fluorescence sensors, where the fluorophore is often integrated into the metal ion recognition and thus difficult to change for different metal ions, the DNAzyme sensors separate fluorescent signaling from metal recognition [41] (see Figure 1). As a result, the DNAzyme sensor can freely choose different fluorophores that do not overlap with each other for multiplex sensing [42].

FIGURE 1.

Fluorescence‐based activity assay for Cu⁺ and Cu^{2 +} DNAzymes. (a) The secondary structure of the Cu⁺ DNAzyme. The Cu⁺ DNAzyme substrate strand was labeled with Cy5 and quenched by Iowa Black RQ. (b) The secondary structure of the Cu^{2 +} DNAzyme. The Cu²⁺ DNAzyme substrate strand was labeled with FAM and quenched by Iowa Black FQ, enabling orthogonal fluorescence detection with minimal spectral crosstalk. (c) Quantification of Cu⁺, the concentration of Cu^{2 +} is 0 nM, 80 nM, 120 nM, 240 nM, 320 nM, 480 nM, 800 nM, 1.6 µM, 4.0 µM, 8.0 µM, and 20 µM, respectively. The concentration of ascorbate is 50 µM. Insert: The linear response at low Cu⁺ concentrations. (d) Quantification of Cu^{2 +}, the concentration of Cu^{2 +} is 0 nM, 30 nM, 50 nM, 100 nM, 200 nM, 300 nM, 400 nM, 500 nM, 750 nM, 1.0 µM, 1.5 µM, 2.0 µM, and 3.0 µM, respectively. Insert: The linear response at low Cu^{2 +} concentrations. Data are presented as mean ± SD from n = 3 independent experiments.

Taking advantage of this sensing platform, we herein report Cu⁺ specific and Cu^{2 +}‐specific DNAzyme fluorescent sensors to simultaneously detect and visualize both Cu⁺ and Cu^{2 +} in a single living neuronal cell. Using DNAzyme sensors, we discovered that Aβ oligomerization induces copper stress in an oligomerization status‐dependent manner, promoting Cu accumulation in neurons. Aβ oligomerization promotes intracellular accumulation, accompanied by a coordinated elevation of Cu⁺ and Cu^{2 +}, distinct from artificial ionophore loading. Mechanistically, Aβ oligomerization triggers intracellular ROS production, DLAT aggregation, and FDX1‐dependent cuproptosis during neuronal death. Cu⁺ chelation or FDX1 knockdown completely prevents neuronal death, whereas ROS scavengers only partially rescue viability, revealing two separable causal modules: copper overload and oxidative stress. These data resolve inconsistencies in prior antioxidant only studies and establish that neuronal fate is governed by absolute Cu⁺ load rather than a global Cu⁺ to Cu^{2 +} ratio shift. Together, our DNAzyme platform achieves the first fluorescence‐based, oxidation‐state‐resolved imaging of copper ions in living neurons, links Aβ‐driven Cu accumulation to FDX1‐mediated cuproptosis and redefines the mechanistic control variable underlying copper neurotoxicity. This work paves the way for further studies exploring the implications of Cu homeostasis in neurodegeneration and potential therapeutic interventions.

2. Results and Discussion

2.1. In Vitro Activity Assay for Cu⁺ and Cu^{2 +} DNAzymes

The optimized sequences and predicted secondary structures of the Cu⁺ and Cu^{2 +} DNAzymes used in this study are shown in Figure 1a,b, respectively. While the design of both DNAzymes are derived from previously reported Cu⁺ [43] and Cu^{2 +} DNAzymes [44], we have engineered the binding arms of both DNAzymes so that they are optimal for the sensor activation at 37 °C with low sequence crosstalk, enabling their application in live‐cell environments. A previous study showed that the Cu⁺ DNAzyme can selectively detect Cu⁺ ions using either Cu⁺‐acetonitrile or Cu^{2 +} in the presence of ascorbate as a water‐stable Cu⁺ source [43]. As it has been shown, Cu^{2 +} can be reduced to Cu⁺ by ascorbate [45], a common reductant in cells. In contrast, the Cu⁺ DNAzyme cannot detect Cu^{2 +} in the absence of ascorbate [43]. To determine the optimal ascorbate concentration, we used polyacrylamide gel electrophoresis (PAGE) to measure the cleavage activity of Cu⁺ and Cu^{2 +}DNAzyme in the presence of various concentrations of ascorbate. As shown in Figure S1a, no cleavage was observed when the Cu⁺ DNAzyme was incubated with 1 µM Cu^{2 +} alone, confirming its high selectivity for Cu⁺ over Cu^{2 +}. When the ascorbate concentration was increased to 50 µM, the cleavage activity was observed. When we tested the Cu^{2 +} DNAzyme activity with increasing levels of ascorbate, we found that, while it can cleave its substrate in the presence of Cu^{2 +} but in the absence of ascorbate, no detectable cleavage activity was observed beyond 5 µM ascorbate (Figure S1b), probably because the Cu^{2 +} is reduced to Cu⁺ and the Cu^{2 +} DNAzyme cannot cleave its substrate in the presence of Cu⁺. Therefore, we chose 50 µM ascorbate as the optimal condition to measure the activity of the Cu⁺ DNAzyme. Using this condition, we found that the cleavage activity increased with increasing concentrations of Cu^{2 +} and 50 µM ascorbate (Figure S2a), while the cleavage activity of the Cu^{2 +} DNAzyme increased with increasing Cu^{2 +} concentrations (Figure S2b). These results indicate that the Cu⁺ and Cu^{2 +} DNAzymes are selective toward their respective Cu⁺ and Cu^{2 +} and do not have cross‐reactivity. Notably, Cu^{2 +} alone does not activate the Cu⁺ DNAzyme in the absence of ascorbate, while Cu^{2 +} DNAzyme activity is attenuated under Cu⁺ generating conditions (with Cu^{2 +} and ascorbate), confirming oxidation‐state‐selective activation and minimal cross‐reactivity between the two DNAzymes (Figure S3).

After confirming the selectivity of Cu⁺ and Cu^{2 +} DNAzymes for their respective metal ions using PAGE, we further investigated their sensitivity by fluorescence assay using the constructs shown in Figure 1a,b. Specifically, we designed catalytic beacons by labeling the fluorophore (F) at one end of the substrate and an intermolecular quencher (Q) on opposite termini of the enzyme strands [46]. In addition, a second intramolecular quencher (Q) was added at the other end of the substrate strand to minimize background fluorescence [47]. The Cu DNAzyme complexes are formed because the melting temperature of the entire complex is higher than room and cellular temperatures. As shown in Figure S4, the fluorescence of DNAzyme was quenched at 37°C with the enzyme/substrate ratio increased, suggesting the substrate and enzyme strands can stably hybridize to each other at 37°C. Based on these results, a substrate/enzyme ratio of 1:1.8 for Cu⁺ DNAzyme, and 1:1.5 for Cu^{2 +} DNAzyme was chosen for further kinetics analysis and live cells imaging. The fluorescent signals at 660 nm (for Cy5 emission) and 520 nm (for FAM emission) in the presence of varying concentrations of Cu⁺ and Cu^{2 +} were monitored. Upon addition of Cu ions, Cu DNAzymes were cleaved and fluorescence increased, because the melting temperature of the fluorophore‐containing substrate decreased after cleavage [48]. Indeed, as shown in Figure 1c,d, the fluorescence enhancement rates were accelerated with additional Cu ions. For Cu⁺ DNAzyme (Figure 1c), the observed fluorescence rate saturates at around 480 nM Cu^{2 +} in the presence of 50 µM ascorbate. For Cu^{2 +} DNAzyme (Figure 1d), the monitored fluorescence rate saturates at around 1.0 µM Cu^{2 +}. The detection limit of Cu⁺ was determined to be 0.95 nM (3σ/slope), with a dynamic range up to 320 nM (insert Figure 1c). The detection limit of Cu^{2 +} was determined to be 52 nM (3σ/slope), with a dynamic range up to 400 nM (insert Figure 1d).

2.2. Simultaneous Imaging of Cu⁺ and Cu^{2 +} in Living Cells Using the DNAzyme Sensors

After we demonstrated Cu⁺ and Cu^{2 +} DNAzymes’ performance in test tubes, we explored the application of Cu⁺ and Cu^{2 +} DNAzymes for simultaneous imaging of Cu⁺ and Cu^{2 +} in living cells. We first co‐delivered Cu⁺ and Cu^{2 +} DNAzymes with defined molar ratio to HeLa cells using the Turbofect transfection reagent. As shown in Figure S5, the fluorescence signal of both Cu⁺ and Cu^{2 +} sensors increased upon the addition of 50 µM Cu^{2 +} and pyrrolidine dithiocarbamate (PDTC), which has been shown to facilitate Cu^{2 +} uptake into cells [49]. To rule out any imaging artifacts unrelated to the Cu ions, we developed an inactive version of the Cu⁺ and Cu^{2 +} DNAzymes, in which the loop in the enzyme strand Cu⁺ DNAzyme was inverted, while the loop in the enzyme strand of the Cu^{2 +} DNAzyme was replaced with polyT, as negative controls. These inactive copper sensors no longer show any Cu ions‐responsive cleavage activity (Figure S6). We next tested the ability of Cu⁺ and Cu^{2 +} DNAzymes to respond to changes in endogenous copper ions in live SH‐SY5Y neuroblastoma cells under either depleted or elevated levels of Cu ions. Specifically, we applied Cu⁺ and Cu^{2 +} DNAzymes to SH‐SY5Y cells, and cellular copper levels were perturbed by incubating with either bathocuproine disulphonate (BCS) to decrease intracellular copper levels [22] or PDTC to increase intracellular copper levels [49]. As shown in Figure 2, we observed a significant decrease in fluorescence intensity following overnight pretreatment with 200 µM BCS to induce copper depletion, compared with the control endogenous group. In contrast, SH‐SY5Y cells treated with 50 µM Cu^{2 +} and PDTC for 30 min displayed a significant increase in fluorescence intensity. As shown in Figure S7, when these inactive DNAzyme controls were delivered into SH‐SY5Y cells, their fluorescence intensity was significantly lower than the active DNAzymes. Together, the data established that the Cu⁺ and Cu^{2 +} DNAzymes enable the detection of relative, condition‐dependent changes in intracellular copper levels in living cells. In biological measurements of copper, it is important to distinguish total cellular copper from the labile, probe‐accessible copper pool. Bulk analytical methods such as ICP‐MS quantify total copper but do not differentiate between Cu⁺ and Cu²⁺ after sample digestion. Consistent with this, a published ICP‐MS measurement for PC12 cells reports ∼7 ng Cu per 10⁶ cells [50], corresponding to a total intracellular copper level on the order of ∼120–210 µM. By contrast, our DNAzyme imaging readouts reflect relative changes in labile copper availability under different conditions, rather than the total copper content. This interpretation is consistent with prior fluorescent and small molecule‐based probe studies in mammalian cells, including HeLa [51], SH‐SY5Y [52, 53], and brain‐derived cell types [22], which report labile copper pools in the nanomolar range and emphasize that probe responses track condition‐dependent changes in bioavailable copper rather than total cellular copper.

FIGURE 2.

Images of SH‐SY5Y cells treated with BCS and PDTC. (a) SH‐SY5Y cells transfected with Cu⁺ and Cu^{2 +} DNAzymes along with BCS and PDTC. The cells were incubated with 200 µM BCS in DMEM/10% FBS medium overnight, followed by DNAzyme transfection and images. The concentration of Cu^{2 +} and PDTC is 50 µM. After DNAzyme transfection, the cells were incubated with 50 µM Cu^{2 +} and PDTC in Opti‐MEM for 30 min, the medium was replaced by Opti‐MEM for images. (b) Distribution of FAM and Cy5 intensity of confocal images of SH‐SY5Y cells treated with BCS and PDTC. The green channel is from FAM fluorescence. The red channel is from Cy5 fluorescence. Scale bar 20 µm. Data were analyzed using an unpaired t‐test. For all graphs, data are presented as mean ± SD, with n = 35∼45 cells per condition collected from more than five randomly selected fields of view. Statistical analysis was performed using an unpaired two‐tailed t‐test. ****p < 0.0001.

2.3. Imaging Cu⁺ and Cu^{2 +} Homeostasis in Cuproptosis Process

Cuproptosis is a recently discovered form of cell death that is mediated by copper and is a non‐apoptotic form of cell death [5]. Current evidence suggests that mitochondrial ferredoxin (FDX1) is implicated in cellular copper metabolism and is linked to copper‐dependent cell death pathways. While some studies describe FDX1 as a reductase that potentially converts Cu^{2 +} to Cu⁺, direct biochemical proof of this activity in vitro or in vivo remains limited. To reveal the cellular function of FDX1 in reducing Cu^{2 +} to Cu⁺, we carried out redox imaging of Cu by simultaneously quantifying Cu⁺ and Cu^{2 +} homeostasis using DNAzyme sensors across different FDX1 genetic backgrounds. We constructed CRISPR‐Cas9‐based FDX1 knockout in DLD‐1 cells, which is a human colorectal adenocarcinoma cell line. As shown in Figure 3, with the genetic knockout of the FDX1 gene, the fluorescence intensity in the Cu⁺ channel decreased around 1.6‐fold, in the Cu^{2 +} channel increased around 2.0‐fold, while that Cu^{2 + /}Cu⁺ ratios significantly increased, compared with wild‐type groups. These data demonstrate that DNAzyme is well suited for the detection of Cu⁺ and Cu^{2 +} homeostasis in cuproptosis, and FDX1 is a key regulator of Cu^{2 +} to Cu⁺ reduction and mitochondria Cu toxicity.

FIGURE 3.

Images of Cu⁺ and Cu^{2 +} in wild‐type and FDX1‐knocked out DLD‐1cells. (a) Wild type and FDX1‐knocked out DLD‐1 cells transfected with Cu⁺ and Cu^{2 +} DNAzymes. (b) Distribution of FAM and Cy5 intensity of Cu⁺ and Cu^{2 +} in wild‐type and FDX1 knocked off DLD‐1 cells. Scale bar 20 µm. Data were analyzed using an unpaired t‐test. For all graphs, data are presented as mean ± SD, with n = 35∼45 cells per condition collected from more than five randomly selected fields of view. Statistical analysis was performed using an unpaired two‐tailed t‐test. ****p < 0.0001.

2.4. Imaging Cu⁺ and Cu^{2 +} Under Treatment of Amyloid‐β in Neuronal Cells

AD is the most common type of dementia, and the amyloid cascade hypothesis proposed that the misfolding of the extracellular Aβ protein into plaques leads to neurotoxicity and contributes to the AD pathogenesis. Aβ monomers aggregate into different forms of oligomers, which can then accumulate regular fibrils. The oligomeric form of Aβ has been confirmed as the most neurotoxic aggregate and proposed as a historic hallmark of AD [54, 55]. Aβ can bind Cu ions and thus modulate the aggregation process and reduce Cu^{2 +} to Cu⁺, which perhaps facilitating Aβ‐mediated oxidative damage in AD [56]. Aβ_1–40 and Aβ_1–42 are still regarded as the most abundant and their binding to metal ions is the most investigated [57]. After verifying that Cu DNAzyme can be used to detect the intracellular Cu⁺ and Cu^{2 +} homeostasis in live cells, we explored the intracellular copper levels influenced by Aβ peptides in neuronal SH‐SY5Y cells by pretreating cells with 10 µM Aβ40 monomers, oligomers, and fibrils for 24 h. As shown in Figure 4a, after treatment with Aβ40, the fluorescence signals of intracellular Cu⁺ and Cu^{2 +} DNAzymes treated with Aβ40 monomers, oligomers, and fibrils were significantly enhanced compared with the control cells without Aβ40 treatment. Notably, the quantitative data in Figure 4b show that the highest fluorescence intensity was observed when the cells were treated with 10 µM Aβ40 oligomers, with no statistically significant change in the Cu⁺/Cu^{2 +} ratio. Aβ40 oligomer is known to be the more toxic species to cells because they can diffuse into cell membranes, forming ion‐permeable pores and inducing membrane fragmentation [58], leading to the enrichment of intracellular copper ions. We also transfected the Aβ40 peptides preincubated cells with inactive Cu⁺ and Cu^{2 +} DNAzymes as a negative control, as shown in Figure S8, the Aβ40 peptides have no influence on the background fluorescence signal.

FIGURE 4.

Images of intracellular copper levels influenced by Aβ peptides in SH‐SY5Y cells. (a) SH‐SY5Y incubated with 0 µM (endogenous, control), 10 µM Aβ40 monomer, 10 µM Aβ40 oligomers, and 10 µM Aβ40 fibrils, for 24 h. Then the cells were transfected with Cu⁺ and Cu^{2 +} DNAzymes for imaging. (b) Quantification of FAM(Cu^{2 +}) and Cy5 (Cu⁺) fluorescence intensity from confocal images of SH‐SY5Y cells treated with Aβ peptides. The green channel is from FAM fluorescence. The red channel is from Cy5 fluorescence. Scale bar 20 µm. Data were analyzed using an unpaired t‐test. For all graphs, data are presented as mean ± SD, with n = 35∼45 cells per condition collected from more than five randomly selected fields of view. Statistical analysis was performed using an unpaired two‐tailed t‐test. ****p < 0.0001, ***p < 0.001, and ns is p > 0.05.

We next expanded to characterize copper homeostasis directly in human‐induced pluripotent stem cell (iPSC)‐derived neuron‐based model, a functional human neuron as a powerful tool for investigating the pathogenesis of neurodegenerative disease. In this cell model, Cu ions are released from synaptic vesicles into the synaptic cleft and can transiently reach micromolar levels upon neuronal excitation during neurotransmission, which suggests the potential interaction between Cu ions and Aβ40 at the synaptic cleft [59]. To examine this possibility, we incubated 10 µM Aβ40 oligomers with iPSC‐derived neuronal cells and observed a pronounced increase in intracellular Cu fluorescence signals not only within neuronal cell bodies but also along outgrowing axons (Figure S9). The enlarged views in Figure S9 were therefore included to highlight the subcellular distribution of copper accumulation and to visualize signal propagation from the soma into axons. In contrast, a low fluorescence background was observed when the inactive DNAzymes were used as a negative control (Figure S10). These results demonstrate that Cu⁺ and Cu^{2 +} DNAzymes sensors can be used to monitor changes in intracellular copper homeostasis associated with AD.

2.5. Copper Homeostasis and Cuproptosis in Neuronal Cells

With extensive research on regulatory cell death in AD, increasing evidence has revealed that copper dyshomeostasis and neurotoxicity could play a contributing role in AD neurodegeneration. Encouraged by the above demonstration of using DNAzyme sensors to visualize Cu⁺ and Cu^{2 +} in cuproptosis within cancer cells and in neuronal Cu homeostasis, we next investigated whether the same platform could resolve copper redox balance during neuronal cuproptosis. To achieve the goal, we constructed SH‐SY5Y cells with FDX1‐knockdown using RNAi technology. The depletion of FDX1 by RNAi in SH‐SY5Y cells caused a strong decrease in Cu⁺ fluorescence signal and a significant increase of Cu^{2 +} fluorescence signal, while the Cu^{2 +}/Cu⁺ ratios significantly increased compared with SH‐SY5Y cells without this FDX1‐knocked down (Figure 5). The FDX1 knockdown efficiency was determined through qPCR, the result in Figure 5b shows that the expression of FDX1 was inhibited by the siRNAs (Figure 5b). This data demonstrated that DNAzyme can be used for imaging Cu⁺ and Cu^{2 +} homeostasis during cuproptosis in live neuronal cells, establishing their utility for dissecting redox‐metal regulation in neurodegeneration.

FIGURE 5.

Images of Cu⁺ and Cu^{2 +} in wild type and FDX1‐deficient SH‐SY5Y cells using small interference RNA and determined the knockdown efficiency through qPCR. (a) Wild‐type and FDX1‐knockdown SH‐SY5Y cells transfected with Cu⁺ and Cu^{2 +} DNAzymes. Scale bars, 20 µm. (b) Distribution of FAM and Cy5 intensity of Cu⁺ and Cu^{2 +} in wild‐type and FDX1‐knockdown SH‐SY5Y cells. Determined the knockdown efficiency through qPCR. For all graphs, data are presented as mean ± SD, with n = 35∼45 cells per condition collected from more than five randomly selected fields of view. Statistical analysis was performed using an unpaired two‐tailed t‐test. ****p < 0.0001, ***p < 0.001, **p < 0.01, and *p < 0.05.

Cuproptosis mediated by both oxidative stresses and proteotoxic stresses was also observed in neurons during Cu overload by elesclomol (ES), which is a copper ionophore, contributing to neuronal death [15]. The high redox properties of Cu can promote the Aβ‐induced oxidative stress and subsequent neuronal death. Our imaging data in Figure 4 show that the Aβ40 treatment resulted in an increase in the intracellular Cu level. This result suggests that Aβ acts as an endogenous copper ionophore, thereby potentiating cuproptotic stress in AD. To determine whether Aβ‐Cu complexes directly trigger cuproptosis‐linked neuronal death, SH‐SY5Y cells were treated overnight with 10 µM Aβ40 and varying Cu^{2 +} concentrations. As shown in Figure 6a, Aβ‐Cu complexes exhibited significantly higher cytotoxicity than Aβ40 alone, whereas FDX1‐knockdown cells displayed strong resistance (Figure S11). The cytotoxic effect was enhanced at higher Aβ/Cu ratios, consistent with previous reports showing that Cu^{2 +}‐Aβ stoichiometry critically regulates amyloid fibril formation and associated neurotoxicity [60]. These results indicate that Cu complexed with Aβ induces neuronal death through an FDX1‐dependent cuproptosis pathway.

FIGURE 6.

Cell death in FDX1 regulated SH‐SY5Y cells with Aβ‐Cu treatment. (a) SH‐SY5Y cells treated with 10 µM Aβ40 and 100, 50, 10, 2, 1, and 0.5 µM CuCl₂. (b) Western blot analysis of oligomerization state of DLAT treated with elesclomol‐Cu or Aβ‐Cu complex from extracts of SH‐SY5Y cells with siRNA knockdowns. (c) Schematic diagram of the mechanism of cell death induced by copper overload by Aβ‐Cu complex. (d) FDX1‐knockdown SH‐SY5Y cells treated with Aβ‐Cu (20:1) and 25 µM Tempol. (e) Intracellular ROS levels measured using CellROX Orange in FDX1‐knocdown SH‐SY5Y cells. For all graphs, data are presented as mean ± SD from n = 3 independent experiments. Statistical analysis was performed using an unpaired two‐tailed t‐test. ****p < 0.0001, *p < 0.05, and ns is p > 0.05.

To confirm pathway activation, we analyzed the oligomerization state of DLAT, a hallmark of cuproptosis, by Western blot (Figure 6b). In the control group, DLAT predominantly appeared as monomeric species under basal conditions, indicating low cuproptosis activity. Densitometric analysis of DLAT oligomers normalized to β‐actin (Figure S12) supports delivering Cu into cells by elesclomol ionophore treatment (Cu/Ele group) increased the intensity of oligomeric DLAT bands, confirming that Cu/Ele promotes cuproptosis through DLAT oligomerization beyond its baseline level. Compared with the untreated Ctrl group delivered with non‐targeting siRNA, treatment with Aβ‐Cu under a 20:1 ratio (Cu/Aβ group) markedly increased oligomer formation, although to a lesser extent than in the Cu/Ele positive control group. This observation suggests that Cu/Aβ treatments promote DLAT oligomerization‐dependent cuproptosis. In contrast, in cells with FDX1 knocked down by siRNA, the same treatments resulted in a noticeable reduction in oligomeric DLAT species in both Cu/Ele and Cu/Aβ groups, showing less cuproptosis activity. These results indicate that Aβ‐Cu treatment induces FDX1‐dependent DLAT oligomerization and cell cuproptosis, which can be rescued by lowering FDX1 expression, and support a role for cuproptosis in Aβ‐Cu induced neural cell death.

We next examined whether oxidative stress also contributes to Aβ‐Cu cytotoxicity. SH‐SY5Y cells with FDX1 knockdown were treated with ROS scavengers Tempol, which is a cell‐permeable nitroxide compound with superoxide dismutase (SOD) mimetic activity. As shown in Figure 6d, Tempol significantly attenuated Aβ‐Cu induced cell death, indicating that reactive oxygen species contribute to cytotoxicity in this process. However, because Tempol broadly scavenges multiple ROS types, the incomplete rescue of cell viability suggests that ROS is not the sole driver of toxicity. Together with the complete protection afforded by FDX1 knockdown, these results support the presence of a separable copper‐dependent mechanism, consistent with Cu⁺‐driven proteotoxic stress and cuproptosis pathways, in addition to ROS‐mediated damage. The fluorescence intensity of ROS in the control group is higher than that of FDX1 gene knockout cells (Figure 6e), which means oxidative stress decreased with lower Cu⁺ generated in FDX1 knocked out of SH‐SY5Y cells. Together, the data showed that Aβ‐Cu is involved in both oxidative stress and cuproptosis‐mediated proteotoxic stress during neuronal death. These findings support the interpretation that ROS‐dependent mechanisms, beyond the lipid radical axis, contribute significantly to the observed cell death.

3. Conclusions

By leveraging Cu oxidation state‐specific DNAzymes, we established the first fluorescence‐based sensing platform capable of simultaneously detecting and imaging both Cu⁺ and Cu^{2 +} in single living neurons. This dual‐sensor approach enables real‐time visualization of Cu redox dynamics in physiologically relevant environments and provides an accessible tool for dissecting the molecular basis of copper‐induced neurotoxicity. We note that this approach requires intracellular delivery of DNAzyme probes, which may perturb cellular physiology beyond copper homeostasis, and that DNAzyme kinetic parameters determined in vitro do not directly reflect intracellular reaction kinetics. Accordingly, the primary utility of this method lies in comparing relative changes in Cu⁺ and Cu^{2 +} levels across biological conditions rather than providing absolute quantification.

Using this platform under Aβ‐induced, disease‐mimetic copper loading conditions, we uncovered that Aβ‐induced copper stress in neurons is aggregation‐dependent and leads to both FDX1‐mediated Cu⁺/Cu^{2 +} redox cycling and elevated intracellular ROS levels. Mechanistically, we delineate two separable modules, Cu⁺ driven proteotoxic stress and ROS mediated oxidative stress, showing that Cu⁺ chelation or FDX1 knockdown abolishes cell death, whereas ROS scavengers afford only partial rescue. These findings clarify why antioxidant treatments alone fail to fully protect neurons and reveal that neuronal fate is governed by absolute Cu⁺ load rather than by a shift in the global Cu⁺ to Cu^{2 +} ratio.

Conceptually, this work refines the field's understanding of copper neurotoxicity by linking Aβ‐Cu redox imbalance to regulated, FDX1‐dependent cuproptosis, extending beyond the traditional ROS‐centric model of Alzheimer's disease. The DNAzyme sensors not only fill a long‐standing methodological gap in redox state resolved copper ions imaging but also introduce a generalizable fluorescence strategy for probing metal ions dynamics in living systems. Together, these advances provide a framework for investigating Cu homeostasis as a therapeutic target and for uncovering the broader relevance of cuproptosis in neurodegeneration and other metal dysregulated pathologies.

Funding

This material is based on work supported by the US National Institute of Health (GM141931) and Cancer Prevention and Research Institute of Texas (CPRIT) under grant ID RP240061. The authors also acknowledge the Robert A. Welch Foundation for support of the Lu group research (F‐0020), and the Collaborative Accelerator for Transformative Research Endeavors grant, jointly awarded by The University of Texas at Austin and The University of Texas MD Anderson Cancer Center: the Collaborative Accelerator for Transformative Research Endeavors grant (to Y.L. and B.G.).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: anie71548‐sup‐0001‐SuppMat.pdf.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.