Quantitative Analysis of Cellular Ferroptosis Using Dual-Channel Colorimetry Based on Gold Nanoparticle Catalytic Activity

Materials and Instruments

We purchased HAuCl₄·3H₂O (≥ 99.99%) and sodium citrate dihydrate (≥ 99%) from Aladdin Industrial Corporation (China). FeCl₂·4H₂O (analytical grade, ≥ 99%) and FeCl₃·6H₂O (analytical grade, ≥ 99%) served as the sources of Fe²⁺ and Fe³⁺ respectively. Commercial suppliers provided the chromogenic substrates 3-amino-9-ethylcarbazole (AEC, C₁₄H₁₄N₂) and o-phenylenediamine (OPD, C₆H₈N₂). The Fe²⁺, Fe³⁺, AEC, and OPD were purchased from Sinopharm Chemical Reagent Co., Ltd. (China). Single-stranded DNAs (ssDNA, 5′-GGT TGG TGT GGT TGG-3′) were synthesized by Sangon Biotechnology Co. Ltd. (China). ssDNA purification using high-performance liquid chromatography (HPLC). All other chemicals were analytical grade and used without purification. Ultrapure water (18.2 MΩ·cm) from a Direct-Q3 system (Merck Millipore) was used for preparing all aqueous solutions.

The ultraviolet–visible (UV–Vis) spectra were recorded using a UV–Vis spectrometer (Mettler Toledo). An incubator shaker (Hangzhou Miu Instruments Co., Ltd.) was used to mix solutions and maintain them at 37 °C.

Synthesis of Citrate-Capped AuNPs

Gold nanoparticles (AuNPs) with an approximate diameter of 12 nm were synthesized via the citrate reduction method (Turkevich method), ¹⁴ yielding citrate-capped AuNPs. Briefly, an aqueous solution of HAuCl₄ (100 mL, 0.25 mM) was brought to a vigorous boil under stirring. Then, a sodium citrate solution (1 mL, 34 mM) was rapidly injected into the boiling solution. The mixture was kept boiling for another 15–20 minutes until the color changed from pale yellow to deep red, indicating the formation of AuNPs. The colloid was allowed to cool to room temperature with continuous stirring. The concentration of the AuNP colloid was estimated according to the Beer–Lambert law using its characteristic surface plasmon resonance absorbance.

Functionalization of Gold Nanoparticles (AuNPs) with ssDNA

The placement of ssDNA onto citrate-capped gold nanoparticles (AuNPs, ∼12 nm) was performed based on the physical adsorption of DNA bases (particularly adenine) onto the gold surface. The procedure is outlined as follows: (i) The ssDNA oligonucleotide solution was heated to 95 ^oC for 5–10 min and then rapidly cooled on ice. This step ensures the DNA remains in a single-stranded, unfolded conformation, maximizing the exposure of its nucleobases for subsequent adsorption. (ii) The denatured ssDNA (100 nM, 100 μL) was directly added to the AuNP colloid (11 nM, 100 μL). The mixture was incubated under gentle agitation at room temperature for 12 h to allow for sufficient physical adsorption of the DNA strands onto the nanoparticle surface.

To remove excess, unbound ssDNA, the functionalized AuNPs (AuNPs-ssDNA) were purified by centrifugation. The supernatant was carefully discarded, and the resultant pellet was resuspended in an appropriate buffer (100 μL, 10 mM Tris–HCl, pH 7.4). This wash-resuspension cycle was typically repeated three times to obtain purified AuNPs-ssDNA.

Preparation of Chromogen Solutions for Horseradish Peroxidase (HRP)

3-Amino-9-ethylcarbazole stock solution (4 mg/mL), 20 mg AEC was dissolved in 5 mL N,N-dimethylformamide (DMF) by vortexing. This solution was aliquoted, protected from light, and stored at –20 ^oC for up to several months. The working solution (prepared fresh for immediate use) and 0.5 mL the AEC stock solution was added to 9.5 mL of 0.05 M sodium acetate buffer (pH 5.0–5.2) and mixed thoroughly. Immediately prior to application, 5 μL of 30% (v/v) hydrogen peroxide (H₂O₂) was added and mixed gently. The working solution was kept in the dark and used promptly. It is recommended to filter the solution through a 0.45 μm filter before use to remove any potential micro-crystals and minimize background.

o-Phenylenediamine (OPD) Solution

The OPD substrate, which produces a soluble orange-yellow product, was prepared with caution due to its potential carcinogenicity. All handling was performed wearing gloves.

The substrate buffer (0.1 M citrate–phosphate buffer, pH 5.0) was prepared by mixing 24.3 mL of 0.1 M citric acid solution with 25.7 mL of 0.2 M disodium hydrogen phosphate (Na₂HPO₄) solution. The final volume was adjusted to 100 mL with deionized water, and the pH was verified to be 5.0.

The working solution (prepared fresh for use) of 4 mg of OPD tablets or powder was dissolved in 10 mL of the pre-warmed substrate buffer (pH 5.0). Immediately before adding to the assay, 5 μL of 30% H₂O₂ was added to the solution, which was then mixed well and used immediately under light-protected conditions.

Preparation of the Dual-Channel Detection

The sensor array was constructed based on the modulation of the peroxidase-like activity of AuNP-ssDNA by Fe²⁺ and Fe³⁺ ions. In a typical procedure, 10 μL AuNP-ssDNA colloid was mixed with 10 μL ultrapure water (as a control unit) or solutions containing different concentrations of Fe²⁺ or Fe³⁺. The mixture was incubated at 37 °C for 10 min to allow interaction between the metal ions and the AuNP-ssDNA surfaces.

Subsequently, the chromogenic reaction was initiated by adding 500 μL these substrate solutions (AEC or OPD). For the dual-channel detection, separate aliquots of the AuNP–metal ion mixture were transferred to wells containing either AEC or OPD substrate solution. The final reaction mixture was incubated at 37 °C for a fixed time (e.g., 10, 20, 30, 40, 50, and 60 minutes) to allow color development catalyzed by the peroxidase-like activity of the AuNPs-ssDNA-M^n + .

Data Acquisition and Multivariate Analysis

The response of the dual-channel detection was captured in multiple forms, (i) digital photographs of the colored solutions, and (ii) full UV–Vis spectra (e.g., 400–700 nm). For simultaneous quantitative analysis of Fe²⁺ and Fe³⁺ concentrations, PLS regression models were developed. The dataset was randomly split into a calibration set (∼80%) and an external validation set (∼20%). Model robustness was evaluated using bootstrapping Latin partitions, and performance was assessed based on the square correlation coefficient (R²), root mean square error of cross-validation (RMSECV), and standard deviation (SD)–root mean square error of cross-validation for the external set.^17–19 All chemometric analyses were performed using appropriate toolboxes in Matlab (The MathWorks Inc.).

Detection of Cellular Ferroptosis

A549 cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37 °C in a humidified atmosphere containing 5% CO₂. For experiments, cells were seeded into appropriate culture plates and allowed to adhere for 24 hours. To establish a cellular ferroptosis model, the culture medium was replaced with fresh medium containing varying concentrations of the ferroptosis inducer GPX4-IN-7 (MedChemExpress, USA). ³ Cells were then incubated for an additional 24 h to simulate different degrees of ferroptotic stress. Following treatment, cells were washed three times with phosphate-buffered saline (PBS). Intracellular labile Fe²⁺ was stained using the FerroOrange ferroptosis detection kit according to the manufacturer's instructions. Fluorescence imaging was performed using a confocal microscope. The fluorescence intensity, corresponding to Fe²⁺ levels, was quantified using ImageJ software. In parallel, treated cells were collected, washed with PBS, and resuspended in 0.2 mL of PBS. The cell suspension was thoroughly homogenized via grinding. The homogenate was centrifuged to remove cell debris, and the clear supernatant was collected for analysis. A 10 µL aliquot of the processed cell supernatant was introduced into the self-assembled sensor array, comprising the OPD and AEC sensing units. The array was incubated for 60 minutes (OPD unit) or 20 minutes (AEC unit), as previously optimized. Following incubation, the UV–Vis absorption spectra of the reaction systems were recorded. The spectral data were then input into the pre-established PLS model to predict the concentrations of Fe²⁺ concentration in the samples. A spike-recovery experiment conducted in cell lysis buffer to robustly demonstrate accuracy and precision under conditions mimicking real application. In this study, cell lysates is PBS. Known concentrations of Fe²⁺ and Fe³⁺ standards (10 µL, 50 µM) were spiked into the PBS buffer and quantified using our established dual-channel method and PLS calibration models. The agreement between the iron concentrations predicted by our PLS mold and the values indicated by the FerroOrange fluorescence intensity was assessed using Spearman correlation analysis. A P-value of ≤ 0.005 was considered indicative of statistically significant concordance.

Mechanisms

The core of the method relies on the horseradish peroxidase (HRP)-like catalytic activity of AuNPs–ssDNA–M^n + . In the presence of hydrogen peroxide (H₂O₂), these AuNPs–ssDNA–Mⁿ⁺ can catalyze the oxidation of specific chromogenic substrates, such as AEC and OPD, leading to the formation of colored products with distinct absorption spectra. The catalytic efficiency of AuNPs is highly sensitive to their surface properties, including charge, aggregation state, and the adsorption of molecules or ions. ²⁰

Our initial work ¹² utilized the classic HRP-mimicking substrate, 3,3’,5,5'-tetramethylbenzidine (TMB), with AuNPs-ssDNA. While successful for single-analyte detection, we found that a single substrate like TMB provided insufficient “cross-reactive” signal diversity to resolve mixtures. To overcome this limitation, in our subsequent research, ⁹ we conducted a comprehensive screening. We evaluated a panel of seven common HRP chromogens, including 4-chloro-1-naphthol (4-CN), AEC, OPD, 2,2′-azinobis (3-ethylbenzthiazpline-6-sulfonic acid)-ammonium salt (ABTS), 3,3′-diaminobenzidine (DAB), benzidine, and 5-aminosalicylic acid (5-ASA), for their ability to generate differential response patterns with AuNPs-ssDNA-M^n + . This work successfully enabled the discrimination of 14 metal ions and the quantification of ternary mixtures (Tl⁺/Pb²⁺/Hg²⁺), demonstrating the power of using multiple substrates to create unique “fingerprint” responses. Building on that foundation, the present study focuses on the detection of iron species (Fe²⁺/Fe³⁺), which are crucial in biomedical research (e.g., ferroptosis). For this specific application, we re-evaluated the substrate panel and found that the combination of AEC and OPD provided the most distinct and complementary response patterns for Fe²⁺ and Fe³⁺, leading to superior analytical selectivity for these ions compared to other substrate pairs.

The principle for ion detection exploits the fact that different metal ions interact with the AuNP surface in unique ways, thereby modulating their catalytic activity to different extents. The sensor operates through a cascade of synergistic interactions among the three key components, which is summarized below.

The principle for ion detection exploits the fact that different metal ions interact with the AuNPs surface, thereby modulating their catalytic activity to different extents. The sensor operates through a cascade of synergistic interactions among the three key components:

Platform preparation (AuNPs–ssDNA): Citrate-capped AuNPs are first functionalized with unmodified ssDNA. The ssDNA adsorbs onto the AuNP surface primarily via physical interactions, which concomitantly displaces or covers the surface citrate molecules. This ssDNA layer serves a triple purpose: (a) it stabilizes the AuNPs against salt-induced aggregation, (b) its exposed nucleobases provide specific binding sites for metal ions, and (c) it removes the potentially interfering reductant (citrate) from the nanoparticle surface, thereby creating a cleaner and more defined catalytic interface.

The innovation lies in using two different chromogenic substrates (AEC and OPD) to form a dual-channel detection. For quantitative and simultaneous analysis, the method integrates chemometrics. The data from the dual-channel, full UV–Vis spectra are processed using multivariate calibration models. PLS regression models are built to correlate the multidimensional response pattern from the two-substrate channel with the known concentrations of Fe²⁺ and Fe³⁺ in calibration samples. This model can then deconvolute the combined signal from a mixture, allowing the simultaneous quantification of both ions in an unknown sample.

In summary, the principle combines (i) the ion-specific modulation of AuNP-ssDNA peroxidase-like activity, (ii) the generation of differential colorimetric fingerprints using a dual-channel, and (iii) multivariate calibration models to extract quantitative information on both analytes from the complex response pattern.

Characterization of Gold Nanoparticles

Figure 1 shows transmission electron spectroscopy (TEM) images of the synthesized AuNPs that we prepared following the Experimental section. TEM imaging confirms AuNPs consist of individual spherical-shaped gold nanoparticles (Figure 1a) with average size of about 12 nm. Figure 1b shown the spherical AuNPs have their absorption peak at about 520 nm. Meanwhile, AuNPs were conjugated with ssDNA, the absorption peak position is not change, which indicated the image of well dispersed and morphology AuNPs is not affected by the ssDNA. The AuNPs exhibited an average diameter of 12 nm (polydispersity index, or PDI = 0.135) and a zeta potential of –14.17 ± 1.29 mV. Following ssDNA conjugation, the AuNPs-ssDNA complex showed an increased average diameter of 13 nm (PDI = 0.069) and a zeta potential of –37.23 ± 5.23 mV (Figure 1c). The dynamic light scattering experiments indicated the AuNPs were functionalized with ssDNA. The synthesized AuNPs and their size distribution are appropriate for our subsequent experiments. Figure 1d are the photograph and UV–Vis spectra of AuNPs, AuNPs-ssDNA, AuNPs-ssDNA + Fe²⁺ and AuNPs-ssDNA + Fe³⁺. Fe²⁺ and Fe³⁺ cannot change the colors and absorption of AuNPs-ssDNA, which conclude the AuNPs-ssDNA AuNP aggregation colorimetric method is not suitable to detect Fe²⁺ and Fe³⁺.

OPEN IN VIEWER

Figure 1.

Characterization of gold nano particles (AuNPs). (a) TEM characterization of AuNPs. (b) UV–Vis spectra of AuNPs and AuNPs-ssDNA. (c) Dynamic light scattering data of AuNPs and AuNPs-ssDNA. (d) Photos and spectra of AuNPs-ssDNA response Fe²⁺ and Fe³⁺.

Identification of Fe²⁺ And Fe³⁺ Based on the Color Map and UV–Vis Spectra

Based on the color change of the chromogenic substrate, the catalytic activity of AuNPs–ssDNA can classify the metal ions into different groups. To rigorously evaluate the selectivity of dual-channel detection, fourteen metal ions (Cu²⁺, Fe²⁺, Fe³⁺, Mn²⁺, Ni²⁺, Zn²⁺, Cd²⁺, Cr³⁺, Co²⁺, Ba²⁺, K⁺, Tl⁺, Pb²⁺, and Hg²⁺) was selected. This is a critical experiment to demonstrate that the observed colorimetric response is specific to or significantly distinct for the target ions (Fe²⁺ and Fe³⁺), rather than a general reaction to metal ions. Figures 2a and 2b presents the photographic response of the array, which comprises the two sensing units OPD and AEC, toward each of these ions at a concentration of 10 μM. OPD sensor units exhibited different colors (Figure 2a) when AuNPs–ssDNA was mixed with Fe²⁺, Fe³⁺. The fundamental reason different metal ions produce varied color changes lies in their distinct electronic structures and physicochemical properties. Key factors include: (i) The unique electron configuration (e.g., d-orbital occupancy in transition metals) and stable oxidation states of each ion (e.g., Fe²⁺ versus Fe³⁺) dictate their binding affinity and mode of interaction with the nucleobases of the AuNPs-ssDNA surface. (ii) Upon binding, each metal ion uniquely modulates the peroxidase-mimicking activity of the resulting AuNPs-ssDNA-Mⁿ⁺ complex. This is because the ion's electronic properties influence charge transfer dynamics and the catalytic pathway for oxidizing the chromogenic substrates (AEC/OPD) in the presence of H₂O₂. (iii) Consequently, ions with different electronic structures (e.g., alkali metal Na⁺ versus transition metal Fe³⁺) induce different catalytic efficiencies and kinetics. This leads to variations in the rate and extent of the oxidation reaction, ultimately producing distinct color intensities and hues for each ion-substrate combination in the array.

OPEN IN VIEWER

Figure 2.

Identification of Fe²⁺ and Fe³⁺ based on the color map and UV–Vis spectra. (a) Photograph of the colorimetric OPD sensor unit used for the detection of 14 metal ions (10 μM each). (b) Photograph of the colorimetric AEC sensor unit used for the detection of 14 metal ions (10 μM each). (c) Absorbance at the maximum absorption wavelength in the dual-channel (AEC and OPD) in the presence in the presence of Fe²⁺ at different time. (d) UV–Vis spectra of the colorimetric OPD sensor units used for the detection of 14 metal ions (10 μM each). (e) UV–Vis spectra of the colorimetric AEC sensor units used for the detection of 14 metal ions (10 μM each).

However, the OPD sensor unit cannot distinguish Fe²⁺ and Fe³⁺. Given the differing biological roles of Fe²⁺ and Fe³⁺, it is crucial to distinguish between them. To achieve this, AEC was incorporated as a sensing unit to improve the selectivity of our approach (Figure 2b). The OPD sensing unit effectively identifies Fe²⁺ and Fe³⁺ from other metal ions. Although the AEC unit responds to multiple ions (including Cu²⁺, Fe²⁺, Fe³⁺, and Cr³⁺), it provides clear discrimination between Fe²⁺ and Fe³⁺. By integrating the advantages of both OPD and AEC, the sensor array achieves not only the detection of iron ions but also their specific differentiation into Fe²⁺ and Fe³⁺. The identification performance of the sensing units was highly dependent on the color development time. Notably, the OPD unit required an incubation period of 60 minutes to achieve optimal performance, at which point its colorimetric response was most distinct. UV–Vis spectral analysis further verified that this 60-minute time point enabled the OPD unit to provide clear discrimination between iron ions (Fe²⁺/Fe³⁺) and the other tested metal ions. In contrast, the AEC sensing unit achieved rapid color development, enabling clear discrimination between different metal ions within 20 minutes. While a longer incubation time (60 minutes) was optimal for the OPD unit, it proved detrimental for AEC, as bubble formation occurred. These bubbles severely compromised the assay by distorting visual color assessment. Consequently, incubation times of 60 min (OPD) and 20 min (AEC) were used in further experiments.

The control experiments performed in this study provide definitive insights into the catalytic mechanism and decisively address concerns regarding solution-phase redox interference. First, the absolute necessity of the AuNPs–ssDNA platform for signal generation was unequivocally established. As shown in Figure S1a (Supplemental Material), in the absence of AuNPs (with only ssDNA and metal ions present), the target ions Fe²⁺ and Fe³⁺, along with most other tested metal ions, failed to induce any significant color change in the chromogenic substrates (AEC–OPD) within the 60-minute observation window. Critically, this key experiment, i.e., containing all reaction components except AuNPs, also serves as the most direct evidence that potential side reactions (including any homogeneous Fenton-like processes ²¹ ) do not contribute to the detectable signal under our assay conditions (Figure S2). The system remains optically silent, confirming that the complex redox network potentially involving free metal ions, H₂O₂, and the substrates is kinetically negligible and does not interfere with the measurement.

Therefore, the significant chromogenic activity observed in the complete system is exclusively “gated” by the formation of the ternary AuNPs–ssDNA–Mⁿ⁺ complex. This conclusion is further reinforced by the result that solutions containing only free Fe²⁺ or Fe³⁺ (without AuNPs or ssDNA, Figure S1b) were equally inactive. The stark contrast between the silent “no-AuNP” control and the active complete system underscores that the peroxidase-mimicking activity is not a solution-phase phenomenon but a surface-catalyzed process. The measured signal thus correlates directly with the concentration of metal ions specifically adsorbed and activated at the nanostructured interface, ensuring the specificity of the detection.

Quantitation Analysis of Fe²⁺ and Fe³⁺ in Binary Mixtures

Given its capability to identify Fe²⁺ and Fe³⁺, our sensor array can further be applied to the quantitative determination of each specific ion. In practical samples, Fe²⁺ and Fe³⁺ ions often coexist in varying and unknown ratios. To simulate this realistic scenario, we prepared random mixtures of Fe²⁺ and Fe³⁺ at different proportions. These mixtures were then introduced to the sensor array comprising the OPD and AEC sensing units, and the corresponding UV–Vis absorption spectra (Figure 3a) were collected for analysis. A PLS regression model was employed to enable the simultaneous quantitative analysis of Fe²⁺ and Fe³⁺ in binary mixtures based on their UV–Vis absorption spectra. This method overcomes the limitation of traditional univariate calibration, which depends on a single wavelength and often fails in mixtures due to spectral overlap and poor selectivity. The predictive PLS models for Fe²⁺ and Fe³⁺ were established using the optimal number of PLS components (13 for Fe²⁺, 10 for Fe³⁺), which were selected to minimize the average RMSECV. Model precision was assessed via 10 bootstrap iterations, with the 95% confidence intervals presented in Figures 3b (Fe²⁺) and 3c (Fe³⁺). The models showed strong agreement between predicted and reference concentrations, as evidenced by low RMSECV values (1.0019 ± 0.04 μM for Fe²⁺; 4.3796 ± 0.07 μM for Fe³⁺) and high coefficients of determination (R² = 0.9994 for Fe²⁺; for R² = 0.9879 Fe³⁺). According to IUPAC guidelines, the calculated limits of detection were 76.34 nM for Fe²⁺ and 102.33 nM for Fe³⁺, with both ions exhibiting a linear response across the 0.1–100 μM concentration range.

OPEN IN VIEWER

Figure 3.

Quantitation analysis of Fe²⁺/Fe³⁺ in binary mixtures. (a) The UV–Vis spectra OPD and AEC sensor units for quantitation of binary mixture (Fe²⁺/Fe³⁺). (b) The PLS results for Fe²⁺. (c) The PLS results for Fe³⁺.

The strategy does not rely on a single, perfectly specific probe. Instead, we employ multiple sensing units (AuNPs-ssDNA complexes tested with different chromogenic substrates like AEC and OPD). While ions like Fe³⁺ and Cr³⁺ may produce similar signals in one channel (e.g., the AEC unit), they generate distinct and differential response patterns across the dual-channel. This collective pattern forms a unique “fingerprint” for each ion. To interpret these complex fingerprint patterns and quantitatively determine concentration in potential mixtures, we employed PLS regression modeling. Unlike univariate analysis which looks at a single signal, PLS is a powerful multivariate calibration method that utilizes the entire response profile from all sensor units simultaneously. It is specifically designed to handle correlated data and extract the latent variables that best correlate with the concentration of the target analyte, even in the presence of interfering species that cause similar responses in some channels. The PLS model effectively learns and utilizes the entire unique pattern caused by Fe²⁺ or Fe³⁺, differentiating it from the pattern caused by Cr³⁺ or other ions. Therefore, the perceived similarity in individual channels is not a flaw but a characteristic of the system that is effectively managed and turned into an advantage by the array-based design combined with chemometrics. The combination of a cross-reactive dual-channel and multivariate statistical analysis (PLS) is a well-established paradigm to overcome the limited specificity of individual receptors, enabling accurate quantification in complex environments. Unlike univariate models, PLS is built using UV–Vis spectral data from the dual-channel (AEC and OPD) response. During calibration, the model learns the complete, multidimensional fingerprint of the system, which intrinsically incorporates any minor but consistent systematic effects. Consequently, as long as these effects are reproducible, the model mathematically accounts for and corrects their influence when deconvoluting the signals from unknown mixtures. This ensures that the final predicted concentrations of Fe²⁺ and Fe³⁺ reflect the amount of ions participating in the surface-catalyzed reaction with high accuracy and reliability.

Detection of Cellular Ferroptosis

This study aimed to detect the concentrations of Fe²⁺ and Fe³⁺ in biological processes, intending to provide a new method for biological research and medical diagnosis. To validate the practical application capability of this method, A549 cells were treated with different concentrations of the ferroptosis inducer GPX4-IN-7 (an indirubin derivative) to simulate varying degrees of cellular ferroptosis. Subsequently, the self-assembled sensor array was used to detect the Fe²⁺ concentration in the treated cells (Figure 4a), and the established PLS model was employed to predict the total iron ion concentration (Figure 4b). To verify the accuracy of our method, the FerroOrange ferroptosis detection kit was used simultaneously to detect the iron ion concentration in the same cell samples (Figure 5). Spearman correlation analysis (Figure 6) showed a strong concordance (P ≤ 0.001) between the results obtained by our method and those from the FerroOrange kit, indicating the potential applicability of our method in real complex systems. In this study, cell lysates is PBS. Known concentrations of Fe²⁺ and Fe³⁺ standards were spiked into the PBS buffer and quantified using our established dual-channel method and PLS calibration models. Recoveries were 101.92 and 99.07% for Fe²⁺ and Fe³⁺, respectively. The relative standard deviation (RSD) was 2.34 and 3.01% for Fe²⁺ and Fe³⁺, respectively. This data will directly confirm the method's reliability and resistance to matrix effects in complex biological environments.

OPEN IN VIEWER

Figure 4.

Detection of cellular ferroptosis based on sensor array. (a) The UV–Vis spectra OPD and AEC sensor units for detection of Fe²⁺. (b) The PLS predicted results for Fe²⁺.

OPEN IN VIEWER

Figure 5.

Detection of cellular ferroptosis based on FerroOrange ferroptosis detection kit. (a) Fluorescence of Hoechst 33342 (to stain nuclei, blue), FerroOrange (Fe^2 + in the cell, yellow), Scale: 10 μm. (b) Relative FerroOrange level.

OPEN IN VIEWER

Figure 6.

Analysis of the concordance between the sensor array and FerroOrange ferroptosis detection kit.