A Brake on Toxicity: Iron Suppresses Organic-Induced Superoxide Generation in Macrophages

Abstract

Iron (Fe) is the predominant metal in brake wear particles (BWPs) and urban particulate matter, yet its role in modulating oxidative stress toxicity remains poorly understood. Here, we show that BWPs induce slight cellular superoxide (^•O₂ ^–) at low doses (<10 μg mL^–1) but suppress it dose-dependently at higher doses. In contrast, highway PM_2.5 neither significantly induces nor suppresses cellular ^•O₂ ^– generation. Using Fe alone or mixed with isoprene secondary organic aerosols or 9,10-phenanthrenequinone, typical organic compounds known to induce cellular ^•O₂ ^– via nicotinamide adenine dinucleotide phosphate (NADPH) oxidase activation, we demonstrate that both Fe²⁺ and Fe³⁺ actively suppress baseline and organic-induced ^•O₂ ^– in a dose-dependent manner. Through inhibition experiments and qRT-PCR analysis, we suggest the following mechanisms for Fe suppression on cellular ^•O₂ ^–: exogenous Fe replenishes the intracellular labile iron pool or complexes with organics. This alleviates organic-induced iron deficiency, reduces NADPH oxidase activity, thereby reducing ^•O₂ ^– release. Fe simultaneously consumes ^•O₂ ^– via direct chemical reactions. This work reveals iron as a suppressor rather than a promoter of organic-induced ^•O₂ ^– generation in macrophages, operating through restoration of iron homeostasis and direct chemical quenching. These findings underscore the importance of considering both metal and organic constituents in assessing the health impacts of complex environmental mixtures.

Introduction

Epidemiological studies have consistently linked exposure to ambient particulate matter (PM), including metal-rich nontailpipe emissions, with adverse respiratory and cardiovascular outcomes, with oxidative stress serving as a central mechanistic paradigm in particle toxicity. − The oxidative potential of PM, defined as its redox activity or capacity to catalyze the formation of reactive oxygen species (ROS), is often attributed to its chemical constituents, including transition metals and organic compounds, which can generate ROS directly or through cellular activation. − Among transition metals, iron (Fe) is particularly abundant and ubiquitous in airborne particles, originating from diverse sources including brake and tire wear, industrial emissions, resuspension of road dust, and combustion processes. − In urban environments, brake wear particles (BWPs), contributing up to 55% of traffic-related nonexhaust PM₁₀, − represent a major contributor to ambient Fe level, with Fe constituting 33–67% of total metal mass. , Despite its predominance across multiple PM sources, the role of Fe in modulating PM-induced oxidative stress remains ambiguous.

On one hand, Fe is well-recognized for its pro-oxidant activity. Through Fenton chemistry and Haber-Weiss reaction, Fe²⁺ catalyzes the generation of highly reactive ^•OH from hydrogen peroxide (Fe²⁺ + H₂O₂ → Fe³⁺ + ^•OH + OH^–), − while Fe³⁺ is reduced back to Fe²⁺ by superoxide (^•O₂ ^–), sustaining the redox cycle (^•O₂ + Fe³⁺→ Fe²⁺ + O₂). Consistent with this pro-oxidant role, excess Fe is cytotoxic and can trigger ferroptosis, a form of regulated cell death driven by iron-dependent lipid peroxidation and disrupted redox homeostasis. −

On the other hand, Fe is an essential nutrient, and its intracellular levels are tightly controlled by iron homeostasis networks, including transferrin receptor-mediated uptake, ferritin storage, labile iron pool dynamics, and iron-responsive signaling pathways, which collectively determine cellular susceptibility to oxidative stress. , The disruption of such iron homeostasis due to iron deficiency impairs critical cellular functions, such as mitochondrial energy metabolism, DNA synthesis, cell proliferation, as well as immune cell differentiation and activation. − Emerging evidence also suggests that Fe can exert protective effects under specific conditions. For instance, Bidi et al. found that iron-based nanoparticles reduce intracellular levels of H₂O₂, ^•O₂ ^–, and methylglyoxal (MG) by upregulating antioxidant enzymes and activating the glyoxalase detoxification pathway, thereby mitigating as-induced oxidative stress. Another study suggests that Fe may regulate cellular antioxidant defenses by activating BMP6 transcription through the redox-sensitive transcriptional axis Nrf2-ARE pathway. These observations underscore the dual role of iron as both a pro-oxidant and a potential modulator of oxidative stress.

A similar complexity exists in studies on BWPs. BWPs and other traffic-related PM contain abundant Fe and redox-active and reactive organic constituents, both of which can significantly contribute to oxidative processes. ,,, For instance, our previous work found that Fe constitutes ∼30% of BWPs by mass, while OC and EC account for ∼5%, and phenolic resin-derived organic hydroperoxides have been shown to promote ^•OH formation from BWPs. While some in vitro studies have reported that BWPs can disrupt airway epithelial tight junctions and trigger inflammatory responses through oxidative stress mechanisms, − others have found no significant increase in intracellular ROS levels following BWP exposure. , These inconsistent findings highlight a clear research gap: the role of Fe, either alone or as the predominant metal in Fe-rich BWPs and urban traffic-related PM, in influencing cellular ROS generation and the underlying mechanisms remains largely unknown.

To address these knowledge gaps, we investigated the role of Fe in modulating cellular ^•O₂ ^– production, as ^•O₂ ^– serves as the pivotal upstream precursor of ROS. , We first examined the ^•O₂ ^– response in RAW264.7 macrophages exposed to lab-generated BWPs and ambient PM_2.5 collected near a highway. To dissect the underlying mechanisms, we then used standard compounds for exposure. We selected Fe as the representative metallic component of BWPs and employed 9,10-phenanthrenequinone (PQN) and isoprene SOA as model anthropogenic and biogenic organic compounds, respectively. Cellular ^•O₂ ^– generation was quantified using the Diogenes chemiluminescence assay combined with electron paramagnetic resonance (EPR) spectroscopy, complemented by gene expression analysis of key redox- and iron homeostasis-related pathways. Our findings show that iron suppresses cellular ^•O₂ ^– generation induced by organic compounds, acting through both cellular signaling and chemical quenching mechanisms. This work challenges the conventional view of iron solely as a pro-oxidant in particulate matter and highlights the importance of metal–organic interactions in shaping oxidative stress responses, with implications for understanding the health effects of Fe-rich PM pollution.

Methods

Sample Collection and Preparation

Brake wear particles (BWPs) were generated using a custom-built brake dynamometer system developed at the University of California, Irvine (UCI), built on a heavy-duty lathe (Lodge & Shipley) and equipped with a brake rotor and caliper housed in an aluminum chamber (87 L), which has been previously described in detail by Smith et al. and Fang et al. Braking was controlled by a hydraulic actuator coupled to a computer program, allowing simulation of different braking regimes. Nonasbestos ceramic and semimetallic brake pads were used for particle generation. Particles were collected onto filters and aluminum foils using a Sioutas cascade impactor. The metal contents were quantified by inductively coupled plasma mass spectrometry (ICP-MS).

Ambient PM samples were collected near the I-710 highway in Southern California using a PM_2.5 High-Volume Sampler (Tisch) from May 18 to 21, 2023. Samples were collected for 23.5 h from 1:00 p.m. to 12:30 p.m. the following day. Field samples were collected on prebaked quartz filters. PM mass on filters was estimated using PM_2.5 volume concentration from local South Coast Air Quality Management District (AQMD) PM_2.5 monitors. To account for potential background contamination during handling and transport, field blanks were collected alongside the ambient highway samples. Prebaked quartz filters were handled and loaded into the Tisch High-Volume Sampler following the same procedure as the ambient samples but were removed without initiating the airflow.

Isoprene SOA were generated with ^•OH oxidation using a Potential Aerosol Mass (PAM) oxidation flow reactor (PAM-OFR, Aerodyne Research, Inc.). A schematic of the experimental setup is provided in Figure S1. Isoprene was introduced into a PAM chamber, where ^•OH was generated through photolysis of water vapor and ozone. The total inlet flow rate to the chamber was 6.5 L min^–1, comprising 4.3 L min^–1 dry zero air (Model 111 Zero Air Supply, Thermo Scientific), 2 L min^–1 humidified zero air (Perma Pure, MH-110), and 0.2 L min^–1 precursor flow. The chamber was operated at approximately 40% relative humidity with a mean residence time of ∼123 s. The particle size distribution was monitored using a Scanning Mobility Particle Sizer (SMPS, TSI 3938), revealing a geometric mean particle diameter of 251 nm and a total particle number concentration of 3.58 × 10³ cm^–3 (Figure S2). The SOA particles were collected onto 47 mm polytetrafluoroethylene (PTFE) membrane filters (0.22 μm pore size, Millipore FGLP04700) and stored in Petri dishes at −20 °C before analysis. 9,10-phenanthrenequinone (PQN), ammonium iron­(II) sulfate hexahydrate ((NH₄)₂Fe­(SO₄)₂·6H₂O) and ammonium iron­(III) sulfate dodecahydrate (NH₄Fe­(SO₄)₂·12H₂O) were obtained from Sigma-Aldrich. A 5 mM PQN stock was prepared using dimethyl sulfoxide (DMSO) and stored at −20 °C.

Exposure Sample Preparation

On the day of exposure, samples were freshly prepared. BWPs, highway PM_2.5, and SOA samples were extracted by vortexing for 7 min using an incomplete medium (Dulbecco’s modified Eagle medium, DMEM, GIBCO). The extracts were then diluted with the incomplete medium to achieve the target exposure doses. PQN was diluted from the stock in incomplete medium. Fe²⁺ and Fe³⁺ solutions were prepared by directly dissolving the respective salts in incomplete medium. The final exposure doses in wells ranged from 1.4 to 204 μg mL^–1 for BWPs and from 23 to 84 μg mL^–1 for highway PM_2.5. For SOA and PQN, fixed doses of 253 μg mL^–1 and 1.74 μg mL^–1, respectively, were selected based on our prior dose–response study, which identified these levels as inducing maximal ^•O₂ ^– production. For Fe, a dose series was applied: 0.02, 0.07, 0.55, 2.19, 8.54 μg mL^–1 for Fe²⁺ and 0.02, 0.07, 0.55, 2.20, 8.61 μg mL^–1 for Fe³⁺. To provide environmental context for these exposure levels, a Multiple-Path Particle Dosimetry (MPPD) model was used to estimate pulmonary deposited doses from literature-reported atmospheric concentrations (Table S1). The estimates show that our data covers the range of environmentally relevant exposure levels. However, the real exposure dose is also influenced by bioavailability, cell number and density, uptake efficiency, metabolic capacity, antioxidant defenses, and intercellular signaling which were not considered in the MPPD model. The background Fe from DMEM (0.0105 μg mL^–1) is insignificant compared to our treatment concentrations and was further controlled by medium-only corrections.

Cell Culture and Cytotoxicity

Cell culture and cytotoxicity measurements were conducted following our previous study. In brief, RAW 264.7 macrophage cells (TIB-71, ATCC, USA) were cultured in complete medium consisting of DMEM (Gibco, 11965092) supplemented with 10% fetal bovine serum (FBS; Gibco, A5669701) and 1% penicillin-streptomycin (Gibco, 15140122). Cells were seeded in 96-well plates at a density of 4 × 10⁴ cells mL^–1 with 200 μL per well and incubated at 37 °C and 5% CO₂ to ensure complete cell adhesion prior to exposure experiments. Cytotoxicity was measured using the CellTox Green Cytotoxicity Assay (Promega), which evaluates membrane integrity alterations via fluorescence increases upon cell death. After 4 h of exposure, fluorescence intensity (excitation/emission: 475/525 nm) was measured using a microplate reader (GloMax, Promega or Varioskan LUX, Thermo Scientific or VICTOR Nivo, PerkinElmer), with data expressed as relative fluorescence units (RFU). For BWPs samples, an end point assay was performed with cells plated at 6 × 10³ cells mL^–1, exposed for 4 h, followed by cytotoxicity measurement.

Cellular ^•O₂ ^– Generation

Measurements of cellular ^•O₂ ^– generation were conducted based on our previous study using a Diogenes Cellular Luminescence Enhancement System (National Diagnostics, USA) combined with an EPR-spin probe method. In brief, following cell incubation, the complete medium was aspirated and replaced with probe, incomplete medium and exposure samples. For single exposures, cells were treated with a mixture containing 50 μL Diogenes reagent, 100 μL incomplete medium, and 50 μL of exposure samples. In coexposure experiments, cells received 50 μL incomplete medium combined with 50 μL Fe solution followed by 50 μL of either isoprene SOA extract or PQN solution. Vehicle controls were included for each experiment: blank filter extracts for BWPs, isoprene SOA or highway PM_2.5, 0.1% DMSO in incomplete medium for PQN, and incomplete medium for Fe. Parallel experiments were conducted with 15 μL of the incomplete medium replaced by an equivalent volume of 1250 U mL^–1 superoxide dismutase (SOD; MP Biomedicals) to assess the specificity of ^•O₂ ^– production. To account for potential chemical ^•O₂ ^– production, cell-free control experiments were performed to quantify background superoxide signals, which were subsequently subtracted from sample measurements. For NADPH oxidase inhibitor experiments, 10 μL of 100 μM Apocynin (Apo; Aladdin), a specific NADPH oxidase inhibitor, was substituted for the same volume of incomplete medium. The doses of Fe²⁺ and Fe³⁺ were fixed around 8.5 μg mL^–1. These doses were selected to isolate the modulatory effect of Fe on organic-induced ^•O₂ ^– production without introducing additional suppression from Fe itself. More details can be found in the discussion section. The ^•O₂ ^– production rate was monitored at 3 min intervals throughout the 4-h exposure period using a microplate reader (GloMax, Promega). Total ^•O₂ ^– generation level was calculated by integrating the measured production over the entire exposure duration.

qRT-PCR Analysis

Real-Time Quantitative Reverse Transcription PCR (qRT-PCR) was used to quantify the expression levels of oxidative stress-related genes in macrophages exposed to Fe²⁺ and Fe³⁺ at low (∼8.5 μg mL^–1) and high (∼28 μg mL^–1) dose levels, either alone or in combination with PQN and isoprene SOA. RAW264.7 macrophages were seeded in 12-well plates at a density of 6 × 10⁵ cells mL^–1 and allowed to fully adhere for 4 h. Then, the complete culture medium was removed. For single exposures, 1.5 mL of incomplete medium and 500 μL of either PQN, isoprene SOA, Fe²⁺, or Fe³⁺ were added to the wells. For coexposure groups, 1 mL of incomplete medium was supplemented with 500 μL of Fe²⁺ or Fe³⁺ and 500 μL of PQN or isoprene SOA, respectively. All treatments were conducted for 1 h in a CO₂ incubator at 37 °C. After exposure, the treatment media were aspirated, and cells were gently washed three times with PBS to remove residual compounds. One μg of total RNA was extracted from each group using an RNAeasy RNA Isolation Kit with Spin Column (Beyotime, China). RNA concentration was determined using a NanoDrop Spectrophotometer (NanoDrop One, Thermo Scientific), and RNA quality was determined by the A260/A280 ratio of 1.8–2.1. RNA integrity was also checked by 1% agarose gel electrophoresis to ensure 2 distinct and intact bands of 28S and 18S rRNA (Figure S3). One μg RNA template was reverse-transcribed into single-strand cDNA using a HiScript II Q RT SuperMix for qPCR (Vazyme Biotech, China). The cDNA template was mixed with an AceQ qPCR SYBR Green Master Mix (Vazyme Biotech, China) and the corresponding primers (Table S2) for quantitative real-time PCR, which was performed on a LightCycler 480 Instrument II (Roche, USA). The expression levels of mRNA were normalized to the expression of the housekeeping gene GAPDH and were calculated using the 2^–ΔΔCT method.

Statistical Analysis

All data are presented as the mean ± standard deviation (SD). For the particulate matter analysis, a total of 17 BWPs samples and 8 highway PM_2.5 samples were evaluated, with each sample performed in triplicate to ensure analytical reproducibility. Statistical analysis was performed using GraphPad Prism 10.1, which used unpaired t test with Welch’s correction for Apo inhibition experiments and One-way ANOVA analysis followed by Tukey post hoc multicomparison test for qRT-PCR. Statistical significance was set at p < 0.05.

Results and Discussion

Cellular ^•O₂ ^– Generation by Brake Wear Particles and Ambient Highway PM_2.5

Figure shows the dose–response relationship of cellular ^•O₂ ^– production following exposure to BWPs and highway PM_2.5. Interestingly, the production of cellular ^•O₂ ^– does not follow a conventional sigmoidal or biphasic dose–response pattern. For ceramic BWPs, a slight increase in ^•O₂ ^– is observed at doses below 10 μg mL^–1, while the production decreases with increasing dose. For semimetallic BWPs, no significant increase is observed below 10 μg mL^–1; above this level, the trend parallels that of ceramic BWPs. For highway PM_2.5, neither significant generation nor a discernible dose-dependent trend is observed across the entire tested dose range (roughly 10–100 μg mL^–1). We further investigate the temporal dynamics of these responses. The typical time profiles of cellular ^•O₂ ^– generation are shown in Figure S4. Both BWPs and highway PM_2.5 exhibit a characteristic inverted U shape, though their magnitudes and timing differ significantly. For BWPs (Figure S4a), the ^•O₂ ^– production rate exhibits a rapid increase within the first ∼20–30 min, followed by a pronounced peak and a gradual decline toward near-baseline levels by ∼200–240 min. In contrast, highway PM_2.5 (Figure S4b) exhibits a moderate increase, with the production rate peaking around 50–80 min and reaching a maximum value (∼0.01 μM min^–1) that is more than an order of magnitude smaller than that of the BWPs. To ensure that the observed decrease in cellular ^•O₂ ^– production is not due to cell death, cell cytotoxicity was assessed using the CellTox Green assay after 4 h of BWPs exposure (Figure S5). All exposed samples show significantly lower relative fluorescence units (RFU) than the lysed cells (p-value <0.0001), confirming that the reduced cellular ^•O₂ ^– generation is not attributed to cell death.

1.

Dose–response relationships of cellular ^•O₂ ^– production upon exposure to (a) lab-generated brake wear particles and (b) ambient PM_2.5 collected near a highway. Legends represent samples collected on different dates. For each sample, experiments were performed in triplicates, and data are plotted as the mean ± standard deviations.

Our results align with several previous studies demonstrating that exposure to BWPs do not induce significant cellular ROS production. For example, Zhao et al. (2015), using both acellular and cellular (A549 cells) assays, detected ROS generation only in cell-free systems, with no increase observed in exposed cells. Similarly, Alves et al. reported that BWPs did not induce ROS generation in A549 cells. In contrast, other studies have reported that BWP exposure can trigger ROS-mediated toxicity in airway epithelial cells. Gasser et al. and Puisney et al. observed ROS-related toxicity in epithelial models. Furthermore, Figliuzzi et al. and Hyman et al. have demonstrated a dose-dependent increase in ROS production following BWPs exposure, although they highlighted copper as a key determinant of this enhanced toxicity. The discrepancy between these findings and ours may be critically influenced by several factors. The concentrations of organic components and iron are likely key determinants, as their opposing effects on ROS generation may introduce competitive interactions (discussed further below). Other important factors include the specific formulation of the brake pads and the experimental conditions, such as cell model (epithelial vs macrophage cells) and exposure duration.

Cellular ^•O₂ ^– Generation by Fe Alone and with Organics

In both ceramic and semimetallic BWPs, Fe is the predominant metal, comprising 57–70% of the total metallic content (Figure S6), which is primarily attributed to wear from cast iron brake discs. Similarly, highway PM_2.5 contains substantial metallic content, with Fe as the most dominant metal originating from brake wear, as identified in a previous study at the same sampling site. , Given the predominance of Fe in both BWPs and highway PM_2.5, we hypothesized that Fe plays a key role in reducing cellular ^•O₂ ^– generation. To test this, we conducted experiments using standard Fe²⁺ and Fe³⁺ salts to directly evaluate the contribution of Fe to cellular ^•O₂ ^– generation. BWPs and highway PM_2.5 also contain organic components. Those in BWPs likely arising from thermal decomposition of phenolic resins and aramid fibers in brake pads , or formed from atmospheric reactions of the gaseous brake wear emissions, and those in highway PM_2.5 attributable to vehicle emissions and secondary organic aerosol formation. ,, Therefore, we further examined the combined effects of Fe²⁺ or Fe³⁺ with two representative organic compounds (PQN and isoprene-derived SOA) which are known to generate cellular ^•O₂ ^–. We note that our selection of PQN and isoprene SOA as model organics does not imply that these specific compounds are direct constituents of brake wear particles; rather, quinones were chosen as representative species because urban traffic emissions are known to produce quinone compounds, such as p-phenylenediamines-quinone (6PPD-Q) formed from tire antioxidants. Similarly, isoprene SOA was selected as a ubiquitous atmospheric organic component that can coexist with brake-derived Fe in ambient air, and upon codeposition in the respiratory tract, such Fe-organic mixtures may collectively contribute to cellular oxidative stress responses.

Figure shows the dose–response relationships between cellular ^•O₂ ^– generation and Fe doses following exposure to Fe alone or in combination with PQN or isoprene SOA. Exposure to Fe²⁺ or Fe³⁺ alone elicits negligible cellular ^•O₂ ^– generation (<5 μM, Figure a) across the low-to-moderate dose range (0.01–10 μg mL^–1), with Fe³⁺ yielding slightly lower than Fe²⁺. Notably, at elevated doses (>10 μg mL^–1), a sharp, dose-dependent decline into negative values is observed, indicating a suppression of cellular ^•O₂ ^– driven by Fe. PQN and isoprene SOA alone induce positive cellular ^•O₂ ^– (Figure b,c), consistent with our previous study. Specifically, exposure to PQN triggers macrophages to produce 45 ± 10 μM of total ^•O₂ ^–, whereas a much higher dose of isoprene SOA results in only 8 ± 3 μM. However, when Fe²⁺ or Fe³⁺ was coexposed with either PQN or isoprene SOA, the levels of cellular ^•O₂ ^– induced by these organics decrease in a dose-dependent manner with Fe doses, with Fe³⁺ exhibiting slightly stronger suppressive effects than Fe²⁺ (Figure b and c) (Reasons for this difference are discussed below). Cytotoxicity results show that the cells remain viable throughout the 4 h exposure period (p-value <0.05, Figure S7), confirming that the diminished cellular ^•O₂ ^– signals are not an artifact of cell death, but rather a genuine chemical or biological interference.

2.

Dose–response relationships of cellular ^•O₂ ^– production upon exposure to (a) Fe²⁺ or Fe³⁺ ions alone, (b) PQN coexposure with Fe, (c) isoprene SOA (ISO SOA) coexposure with Fe. PQN and isoprene SOA alone exposure are shown in dotted lines with shaded areas indicating the mean and standard deviation, respectively, from multiple measurements (n ≥ 4). Data points with error bars represent the mean ± standard deviations.

The finding that Fe²⁺ and Fe³⁺ alone do not induce cellular ^•O₂ ^–, yet at higher doses actively suppress both baseline and organic-induced cellular ^•O₂ ^–, points to a regulatory role for Fe. This Fe-mediated inhibition provides an explanation for the unconventional dose–response patterns observed in our BWPs and ambient highway PM. Specifically, for BWPs, the initial increase of cellular ^•O₂ ^– at low doses are likely attributable to organic components, while the subsequent decline at higher doses reflects the point at which iron content becomes sufficient to exert suppression. In contrast, the suppression effect is not as prominent in highway PM_2.5. This absence of clear suppression is likely because the Fe concentrations in these samples remain below the threshold required for suppression. Based on the typical Fe content in ambient PM_2.5 at our sampling site (<5% of total mass) and that Fe solubility is typically low (from <1% in mineral dust to 50–75% in combustion sources), − the estimated maximum Fe concentration in the tested dose range would be below 10 μg mL^–1, which is the level at which Fe-driven suppression becomes apparent (Figure a).

Mechanisms of Iron-Mediated Suppression of Cellular ^•O₂ ^–

To elucidate the underlying mechanisms, we employed qRT-PCR analysis to investigate key genes involved in cellular ^•O₂ ^– production and iron homeostasis, including superoxide dismutase 1 (SOD1), transferrin receptor (TFRC), and NADPH oxidase-2 (NOX2). For Fe exposure, we selected two representative doses (low at ∼8.5 and high at ∼28 μg mL^–1) that span the suppression thresholds identified in Figure . For coexposure experiments, we used only the low Fe dose to isolate the modulatory effect of Fe on organic-induced ^•O₂ ^– without introducing additional suppression from Fe alone.

We first examined whether the observed decline in cellular ^•O₂ ^– could result from enhanced enzymatic dismutation by SOD1, a cytosolic copper/zinc enzyme that converts ^•O₂ ^– into H₂O₂ and molecular oxygen. As shown in Figure a, SOD1 gene expressions remain unchanged across all exposure groups, including PQN alone, isoprene SOA alone, Fe²⁺ or Fe³⁺ alone, and their combinations. This indicates that neither organic compounds nor Fe modulate SOD1 transcription, and that the diminished ^•O₂ ^– signal is not attributable to conversion of ^•O₂ ^– to H₂O₂. Thus, alternative mechanisms must account for the Fe-dependent suppression.

3.

qRT-PCR analysis on (a) antioxidant, (b) iron homeostasis and (c) ROS generation in macrophages upon exposure to Fe²⁺, Fe³⁺, PQN, and coexposure to Fe and PQN. Low and high in subscripts denote two dose levels for Fe. qRT-PCR results were normalized to GAPDH. One-way ANOVA analysis followed by Tukey post hoc multicomparison test, **p < 0.005 and *p < 0.05. Data are presented as mean ± standard deviation (n ≥ 3).

We next evaluated the expression of TFRC (Figure b), which mediates cellular Fe uptake through receptor-mediated endocytosis and reflects cellular Fe status, as well as the expression of NOX2 (Figure c), the catalytic subunit of NADPH oxidase which is the main source of ^•O₂ ^– production during the respiratory burst in macrophages. , Exposure to Fe²⁺ or Fe³⁺ alone, at either dose, does not alter TFRC expression significantly relative to untreated controls. This confirms that Fe supplementation, in the absence of chelators, does not perturb cellular iron status. For NOX2, Fe²⁺ and Fe³⁺ at low doses have no effect on its expression, while at high doses, Fe³⁺ significantly downregulates NOX2. This suggests that Fe³⁺ can suppress NADPH oxidase transcription or stability. The downregulation of NOX2 provides a possible mechanistic basis for the stronger suppression of cellular ^•O₂ ^– observed with Fe³⁺ relative to Fe²⁺ at equivalent doses (Figure ). In contrast, high-dose Fe²⁺, but not Fe³⁺, significantly upregulates FTH1 (see Figure S8), which encodes ferritin heavy chain and is involved in iron sequestration and antioxidant defense. This differential response likely reflects the higher bioavailability of Fe²⁺, prompting cells to sequester excess Fe into ferritin as a protective mechanism. Notably, at low doses, neither Fe²⁺ nor Fe³⁺ alters FTH1 expression, consistent with the observation that low-dose Fe does not perturb cellular iron homeostasis to induce ROS generation or suppression and supports its use in coexposure experiments to isolate modulatory effects with inducing compensatory stress responses. Fe³⁺, being less bioavailable under physiological conditions, , does not trigger this storage response and may instead interact more directly with NADPH oxidase or its signaling pathways to suppress activity. In addition to transcriptional regulation, the reduction in intracellular ^•O₂ ^– levels observed with high Fe may also be attributed to chemically scavenging responses

$$\begin{array}{l}{\mathrm{Fe}}^{2+}+{}^{•}{\mathrm{O}}_2^{-}\stackrel{{\mathrm{H}}^{+}}{\to }{\mathrm{Fe}}^{3+}+{{\mathrm{H}}_2\mathrm{O}}_2\\ k_1=1.0\times 10{}^7\mathrm{\hspace{0.17em}\hspace{0.17em}}{\mathrm{M}}^{-1}{\mathrm{s}}^{-1}\end{array}$$ 1

$$\begin{array}{l}{\mathrm{Fe}}^{3+}+{}^{•}{\mathrm{O}}_2^{-}\to {\mathrm{Fe}}^{2+}+{\mathrm{O}}_2\\ k_2=1.5\times {10}^8{\mathrm{M}}^{-1}{\mathrm{s}}^{-1}\end{array}$$ 2

The higher reaction rate constant between Fe³⁺ and ^•O₂ ^– compared to Fe²⁺ further explains why, across all coexposure experiments, Fe³⁺ consistently exhibits stronger attenuation of ^•O₂ ^– signals than Fe²⁺.

Exposure to PQN significantly upregulates both TFRC and NOX2. Isoprene SOA shows a similar trend, although the increase does not reach statistical significance, likely due to its lower ^•O₂ ^– generation potential (Figure S9). In our previous study, we showed that PQN and isoprene SOA activate NADPH oxidase in macrophages to release massive amounts of superoxide, using a phasor approach to fluorescence lifetime imaging to observe the NADPH oxidase activation. In this work, we confirm the NADPH activation by organics through NOX2 gene expression. We further demonstrate that PQN and isoprene SOA exposure causes iron deficiency. PQN, as a quinone, contains an ortho-quinone motif with two adjacent carbonyl oxygen atoms, a classical bidentate chelating structure known to coordinate with transition metals. − Isoprene SOA may contain a range of oxygenated functional groups such as alcohol, diol, aldehyde, carboxylate groups, which may complex with Fe. − Together with our findings, we propose the following mechanism: following endocytosis, these organic constituents can access the intracellular environment, where their functional groups may complex with bioavailable Fe, thereby depleting the intracellular labile iron pool and inducing cellular iron deficiency. This interpretation is consistent with a proposed model in which air pollutants disrupt iron homeostasis by chelating or displacing Fe from cellular sites, leading to functional iron deficiency; in response, cells upregulate Fe importers and generate ^•O₂ ^– to support ferrireduction, a critical step for restoring Fe availability. ,

When PQN was coexposed with Fe, either Fe²⁺ or Fe³⁺, we observe that the upregulation of TFRC induced by PQN is significantly attenuated. Specifically, the TFRC elevation was reduced by approximately 22% for Fe²⁺ and 14% for Fe³⁺. Co-exposure also lowers NOX2 expression to near-baseline levels. This demonstrates that exogenous Fe replenishes the labile iron pool depleted by organic-mediated complexation, thereby relieving the iron-deficiency signal and dampening NADPH oxidase activity. Additionally, it is possible that Fe directly complexes with the organic compounds prior to or following cellular uptake, limiting the ability of the organics to chelate intracellular iron and further reducing their capacity to induce iron deficiency. Together, these results support the model that exogenous Fe suppresses organic-induced ^•O₂ ^– primarily by restoring Fe availability and reducing NADPH oxidase activity, and consequently, cellular ^•O₂ ^– production declines in a dose-dependent manner with Fe doses, as observed in Figure b. To confirm that NADPH oxidase is the major source of ^•O₂ ^– in the presence of Fe, we employed apocynin (Apo), a specific NADPH oxidase inhibitor. The addition of apocynin (100 μM) results in a near-complete reduction of cellular ^•O₂ ^– production across all treatment groups, reducing signals to baseline or below the detection limit (Figure S10). This result establishes that the ^•O₂ ^– elicited by both PQN and isoprene SOA, as well as its modulation by iron, is entirely dependent on NADPH oxidase activity.

We summarize the pathways governing organic-induced macrophage ^•O₂ ^– generation and its suppression by iron in Figure . Organic components, such as PQN and isoprene-derived SOA, can enter macrophages and complex intracellular iron. This complexation depletes the labile iron pool, creating a state of functional iron deficiency. Iron deficiency in turn promotes NADPH oxidase activity, leading to enhanced cellular ^•O₂ ^– production. When exogenous iron is present, for example, in BWPs or ambient PM, it can alleviate this response through two mechanisms: it may directly replenish the depleted labile iron pool, relieve iron deficiency and reduce NADPH oxidase activity; alternatively, it may complex with organic compounds, limiting their ability to chelate intracellular Fe. Both pathways result in decrease in ^•O₂ ^– release by macrophages. At high Fe doses, excessive expansion of the labile iron pool may lead to intracellular iron overload, triggering additional mechanisms including direct transcriptional suppression of NADPH oxidase and chemical scavenging of ^•O₂ ^– by Fe, further contributing to the suppression of cellular ^•O₂ ^–.

4.

Pathways governing organic-induced macrophage ^•O₂ ^– generation and its suppression by Fe. Organics induce iron deficiency and activate NADPH oxidase (NOX) activity, leading to cellular ^•O₂ ^– production; Exogenous iron alleviates this response by either complexing with organics or replenishing the labile iron pool (LIP), thereby restoring iron availability and reducing NOX activity and cellular ^•O₂ ^– production; At elevated doses, Fe-induced expansion of the labile iron pool may lead to intracellular iron overload, triggering additional suppressive mechanisms through transcriptional downregulation of NOX2 gene and direct chemical scavenging of ^•O₂ ^–.

Implication and Future Perspectives

Iron is an essential micronutrient required for normal cell function, yet its role in particulate matter toxicity extends beyond the conventional paradigm of oxidative stress generation. In this study, we have uncovered a “brake on toxicity”: Fe can actively suppress organic-induced ^•O₂ ^– generation in macrophages, thereby alleviating oxidative stress imposed by organic components. This reveals that the net biological outcome of PM exposure is not simply the sum of individual component toxicities, but rather an integrated response shaped by metal–organic interactions that can either amplify or quench oxidative pathways. It is important to note, however, that our findings are based on cellular ^•O₂ ^– measurements in macrophages, which reflect biological responses involving NADPH oxidase activation and intracellular metal handling. These conclusions may not be directly applicable to acellular oxidative potential assays (e.g., DTT or ascorbate), where Fe could exhibit different or even opposing effects due to its intrinsic redox chemistry. Our findings also help reconcile the seemingly contradictory observations on ROS generation by Fe by demonstrating that the net cellular ROS response is determined by the balance between organic-induced ROS production and iron-mediated suppression.

While the present study employed well-defined model systems to isolate the Fe-organic interaction mechanism, we recognize that future studies incorporating qRT-PCR analysis on cells exposed to BWPs or ambient samples, combined with detailed chemical speciation, will be valuable to assess whether the transcriptional responses observed in our surrogate system are recapitulated in real-world mixtures and to further establish the physiological relevance of Fe-mediated modulation of organic-induced oxidative stress.

Recognition that particle-induced biological effects can arise from a functional deficiency of iron suggests novel interventions such as Fe supplementation. However, such approach must be carefully calibrated, as excessive iron poses its own hazards. While our results demonstrate that Fe suppresses quinone-mediated ^•O₂ ^– production within the tested concentration range (<100 μg mL^–1), it is important to acknowledge that Fe can also induce cellular toxicity through alternative mechanisms that are endpoint-, exposure duration- and cell type-dependent, including mitochondrial dysfunction, inflammatory responses, as well as intracellular Fenton reactions that generate ^•OH. ,,− Given that our tested Fe concentrations extend up to 100 μg mL^–1, the suppressive effect of Fe on organic-induced ^•O₂ ^– should be interpreted within the specific context of the ^•O₂ ^– pathway and the concentration range examined, rather than as an indication of overall protection under all exposure conditions. Indeed, elevated Fe levels are known to trigger ferroptosis, a regulated form of cell death driven by iron-dependent lipid peroxidation. − Future studies investigating Fe-driven alternative toxicity pathways at higher doses will be valuable to comprehensively assess the risk of Fe-containing particles.

Beyond cellular toxicity, our findings carry implications for host defense and immune function. NOX2-mediated respiratory burst is essential for pathogen clearance and phagosomal signaling. , Chronic or excessive suppression of NADPH oxidase by high-Fe particles could therefore impair bacterial clearance and increase susceptibility to infection. While the present study focused on Fe as the most abundant metal constituent in BWPs, we recognize that other transition metals, such as Cu and Mn, may also contribute to the oxidative potential of complex environmental particles, even at relatively low mass fractions, owing to their high redox reactivity. , Systematic investigations of other transition metals and their potential interactions with redox-active organics remain an important direction for future research. Overall, our work opens a new avenue in assessing particulate matter toxicity: one that integrates metal speciation, organic composition, and their interactive effects on iron homeostasis and immune function.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.6c06554.

• Additional experimental details and results, including atmospheric concentration comparisons, experimental setup, particle characterization, cytotoxicity assays, metal composition, and gene expression analyses (PDF)

Y.S.: Laboratory experiments, data analysis, writingoriginal draft, writingreview and editing. K.E. and S.K.: Sample collection (brake wear particles, highway PM_2.5), data analysis, writingreview and editing. W.H., T.P., and S.Z.: PAM reactor provision, SOA generation support, writingreview and editing. M.S.: Writingreview and editing, funding acquisition. T.F.: Supervision, funding acquisition, writingreview and editing.

The authors declare no competing financial interest.