Ferrocene-derived magnetic fiber-particles from diesel exhaust: enhanced pulmonary toxicity via Bach1-SAT1-polyamine depletion

Abstract

Background and aim

Magnetic nanoparticles are key components of air pollution. The combustion of diesel engine fuels, especially with ferrocene doping to reduce emissions, may increase exposure to these particles and related health risks. This study aimed to reveal the generation and characterization of ferrocene-derived magnetic particles (FMP) in ferrocene-doped diesel exhaust, and to investigate its toxicities and associated mechanisms in an avian model.

Methods

FMP was observed in ferrocene-doped diesel exhaust particles, and extracted with neodymium magnets. Extracted FMP was characterized, and exposed to hatchling chickens via aerosol inhalation. Pulmonary toxicities were assessed with pathological and molecular methods. Associated mechanisms were investigated with RNA-seq, in vitro cell culture, and in vivo gene silencing.

Results

FMP was characterized to be fibrous, magnetic iron-containing carbon particles. Extracted FMP could directly induce pulmonary toxicity. Mechanistic investigations revealed molecular mechanism associated with ferroptosis via Bach1, SAT1 and polyamines depletion, and further confirmed with ferroptosis inhibitor treatment, Bach1 inhibitor treatment, supplementation of polyamines or SAT1 silencing.

Conclusions

Ferrocene doping could result in formation of magnetic particles in diesel exhaust. For the first time, magnetic fiber-like particles were extracted from ferrocene-doped DE particles, which is a potential source of magnetic particles in air pollution. To better balance emission control and health effects, further investigations are necessary.

Graphical Abstract

Introduction

Magnetic nanoparticles(MNPs), important contributors to air pollution, have raised concern due to their nanoscale size, magnetic properties, and potential to interact with biological systems [1, 2]. The magnetic properties of MNPs make them highly promising for applications in biomedical and environmental monitoring. These same properties also enable them to play a significant role in air pollution. With their extremely small size, these nanoparticles can easily penetrate biological barriers, entering the lungs, blood, and other tissues, posing a significant potential risk to human health. Research indicates that among various sources, fuel combustion is a major contributor to airborne MNPs [3, 4]. These combustion-derived nanoparticles have been widely detected, not only in the environment but also in human tissues, further highlighting their potential hazards to public health [5, 6]. Thus, the identification and characterization of combustion-generated MNPs, as well as the elucidation of their toxicities are necessary.

Among the common fuels, diesel is an important one. Diesel exhaust (DE) is a primary source of atmospheric particulate matter (PM), the major component of air pollution [7]. To minimize the various adverse health effects [8, 9], emission control measures, such as incorporation of various chemical compounds in diesel fuel (“doping”) had been taken [10,11,12]. Among the doping agents, Ferrocene (Fe(C2H2)2) has garnered attention for its stability in liquid fuels, solubility, and cost-effectiveness [10]. Introducing ferrocene into diesel has shown promise in reducing nitrogen oxides and particulate emissions [13]. However, this practice alters the chemical composition and physical structure of exhaust gases [14]. Studies have identified iron-rich, fiber-shaped nano-scale particles in diesel exhaust particles (DEPs) from ferrocene-doped diesel, instead of normal, spherical-shaped DEPs [15, 16, 17]. Such fiber-shaped nanoparticles raised toxicological concerns, since fibers are often considered to be more potent than spherical particles [18]. However, existing studies had predominantly focused on the chemical properties of these pollutants, lacking comprehensive toxicological studies to evaluate potential health risks [19].

Lung is the primary target organ of airborne inhalable pollutants, such as DEPs or MNPs. Animal studies had associated inhalation of DEPs with progressive pathophysiological changes, such as inflammatory cell infiltration and extracellular collagen matrix deposition, ultimately leading to pulmonary fibrosis [20]. Obviously, the translatability of laboratory studies remains uncertain due to species differences and high doses of diesel exhaust used, and no concrete evidence had been reported for the casual relationship between diesel exhaust and human pulmonary fibrosis yet. Nonetheless, MNPs, including traffic-generated ones, had been considered as an emerging health risk [21]. Specifically, iron-loading had been shown to convert a non-reactive particle into fibrogenic particles [22]. As a result, iron-doping may further increase the potential risks associated with diesel combustion-generated particles. Thorough evaluation of the potential adverse health impacts posed by ferrocene doping and resultant MNs in diesel engine exhaust is imperative.

In the current study, magnetic extraction was performed on collected ferrocene-doped diesel exhaust particles, from which nanoscale magnetic fiber-particles were successfully enriched. The toxicities and molecular mechanisms of these magnetic particles were comprehensively explored through transcriptomics, in vitro cell experiments, and in vivo experiments with lentivirus-mediated gene silencing. The current study aimed to characterize the ferrocene-derived magnetic fiber-particles (FMP) from ferrocene-doped diesel exhaust, unveiling, for the first time, the toxicological effects of FMP and the underlying molecular mechanisms, focusing on Bach1-SAT1-mediated disruption in polyamine metabolism and ferroptosis.

In summary, the current study provided a comprehensive assessment of potential health risks associated with FMP, providing new mechanistic insights and contributing to the establishment of safety standards for ferrous diesel additives.

Methods

Collection and characterization of ferrocene-doped diesel exhaust particles

Different concentrations of ferrocene (CAS No. 102-54-5, Yatai United Chemical, Qingdao, China) (0, 205, 410, 820 mg/L, equivalent to iron mass concentrations of 0, 75, 150, 300 ppm) were added to China VI grade 0 diesel fuel following standard practices [23]. The doses were close to those in existing studies with ferrocene as a diesel doping agent. A combustion model was established with a Yuchai YC9800XE diesel engine (Yuchai Power, Guangxi, China) generator operating at approximately 50% power output (5000 W). Whole diesel exhaust particulates were collected with a KB-120 F intelligent particulate samplers (Jinsida Co., Ltd, Qingdao, China) at a flow rate of 100 L/min for 10 min per Teflon membrane(MS-PTEE, Safelab, Beijing, China). Filter weights were acquired before and after collection, and filter integrities were also verified before and after collection. Collected membranes were processed with ultrasonication. Specifically, the membranes were cut into 1 cm² squares, and put in ultrapure water containing 0.01% Tween80 (10 ml per membrane), and ultrasonicated for 30 min twice with a sonicator (KQ-800DE, Kunshan Sonicator, Jiangsu, China). Sonication parameters were: output 800w, frequency 40 khz, temperature 25℃。Following ultrasonication, samples were filtered and vacuum freeze-dried for the collection of DEP. Freeze-dried DEPs were protected from light and stored in -20 ℃ freezers until further use. 1 mg/ml DEP suspensions were made in ultrapure water with sonication (800 w, 40khz, 20 min). DEP morphology and elemental composition were analyzed with scanning electron microscopy (SEM) (Hitachi, Japan) and energy-dispersive X-ray spectroscopy (EDX) (Carl Zeiss, Germany), respectively.

Extraction and characterization of ferrocene-derived magnetic fiber-particles (FMP) in ferrocene-doped diesel exhaust particles

Magnetic components from ferrocene-doped diesel exhaust particles were extracted using a strong magnetic extraction device based on Zhang et al.‘s method [2] (Fig. 2A). Three neodymium magnets (20 × 10 × 5 mm, strength 270–300 MT each) were aligned and attached to a sample vial. Diesel exhaust particle suspensions were added and shaken overnight at a speed of 60 rpm. The vial was then washed, the magnets removed, and the magnetic particles left on the vial’s wall were rinsed into a new tube. Mass fraction of FMP was calculated by dividing the weight of FMP by the total weight of diesel exhaust particle. The morphology and composition of the extracted magnetic components were examined using a scanning electron microscope (JEOL Ltd., JSM-6390LV). Nanoparticle size and Zeta potential in PBS and MEM cultures were measured with a nanoparticle size and Zeta potential analyzer (Brookhaven, USA). Measurement parameters were: temperature 25℃, measurement time: 3 min, equilibration time: 10 min, liquid: PBS/MEM, viscosity: 0.890, refractive index: 0, PH: PBS (7.2), MEM (7.4), refractive index of particles: real: 1.590, imaginary: 0. Shape: uniform spheres. Analytical model: NNLS. Please note that the particle size measurement may not be accurate for the fiber-like particles in the samples, thus six transmission electron microscopy images were taken per ferrocene dose, and the dimensions of the fiber-like particles were manually measured with ImageJ (NIH, US).

Animal tests

Chicken embryo incubation

Fertilized Plymouth Rock chicken eggs (Gallus gallus) were obtained from Hedi Farm (Jining, Shandong, China). This breed features distinctive black-and-white feathers for easy visual confirmation, reducing variability. Eggs were evenly grouped by weight and incubated in a Keyu incubator (Dezhou, Shandong, China) following standard protocols. The incubator automatically adjusted conditions: temperature decreased from 37.9 °C to 37.2 °C, and humidity increased from 50 to 70%. Eggs were rotated every 3 h until embryonic day 19 (ED19). Prior to hatching, eggs were transferred to separate hatching boxes. Newly hatched chicks were kept in a warm box with clean water until use. All procedures received approval from the Institutional Animal Care and Use Committee (IACUC) of Qingdao University (Approval number: QDU-AEC-2022051), with documentation available upon request.

In Ovo inhalation of ferrocene-doped diesel exhaust to hatching chicken embryo

On ED18-19, developing embryos were exposed to diluted ferrocene-derived diesel exhaust as described by Jiang et al. [24], with diesel exhaust diluted 1:1 with clean air, administering 10 ml per egg across four infusions. Briefly, two holes were opened with a metal probe at two ends of the air cell on the egg shell, 10 ml of diluted diesel exhaust were infused from one hole, expelling air from the other hole. After infusion, holes were sealed with duct tape. This method exposed hatching chicken embryos to diesel exhaust the moment they start pulmonary respiration. The chicks were sacrificed one month after hatching, and lung tissue samples were collected. Tissue samples were divided: one half fixed in 4% paraformaldehyde phosphate-buffered saline for histological assessment, and the other half stored at -80 °C for molecular analysis.

Intratracheal administration of FMP

FMP from diesel with 820 mg/L ferrocene (300 ppm iron) was selected for animal testing, a dose typical for commercial diesel doping agents [23]. Magnetic extraction results indicated that FMP constituted 7–11% of total DEP. The U.S. Mine Safety and Health Administration sets the occupational exposure limit for diesel exhaust at 106 µg/m³ [25], leading to an estimated average exposure contribution of about 9.5 µg/m³ for FMP. Taking into account the tidal volume, ventilation rate, body weight into account, to mimic a 90 day, 8-hour/day exposure, the total exposure dose would be approximately 0.6 mg/kg. In this study, FMP was administered to chicks at 0.3, 0.6 or 1.2 mg/kg (corresponding to approximately 12, 24–48 µg per chick, equivalent to 45, 90, or 180 days of 8-hour exposure). An 40, 80–160 µg/ml FMP suspension was prepared in phosphate-buffered saline containing 0.1% Tween and delivered via a high-pressure aerosol device HRH-MAG4 (Huironghe Technology, Beijing, China) in three 100 µl doses within 24-hour post-hatch. Generated aerosols were sprayed down the trachea of the animals. Half of the hatched chickens were sacrificed within 24 h, while the other half were reared until one month old before being sacrificed.

Air cell injection and in Ovo lentivirus transfection

Air cell injection and in ovo lentivirus-mediated transfection were performed as described in Zhong et al. [26]. Bach1 inhibitor Substituted Benzimidazole HPPE (4 mg/kg, egg weight) [27] was administered via air cell injection at embryonic day 17 (ED17), and SAT1 silencing virus was administered via microinjection at embryonic day 2 (ED2). For details, please refer to supplementary material.

Lung burden analysis

Preparation of FMP dispersion standards and lung tissue homogenate standards

Following the method of Lee et al. [28], a 3% FBS solution (C04400-500,Vivacell, Germany) was prepared with ultrapure water to disperse the FMP particles, and then gradient concentrations of standard FMP suspensions at 0, 3.1, 6.2, 12.5, 18.7, 25, 37.5, and 75 µg/ml were prepared. Resulting suspensions were subjected to 10 min of ultrasonication in an ice-water bath. To further eliminate interference from lung tissue components during subsequent UV-visible spectrophotometry measurements and to determine the appropriate measurement wavelength, untreated lung tissues were dried, homogenized in 60℃ and centrifuged at 12,000 g. The pellets was then re-dispersed in a 3% FMP solution to prepare lung tissue homogenate standards (LTM). Furthermore, assuming that all FMPs enter the lungs after intratracheal instillation in chick embryos (with 24 µg of FMP accumulated in each animal’s lung tissue after exposure to a 0.6 mg/kg dose), a 24 µg/ml FMP positive control standard in a 3% FBS solution was prepared and pre-treated with ultrasonication in an ice-water bath.

Digestion and recovery of FMP from lung tissue with proteinase K

Following intratracheal administration of FMP, lung tissue was collected from each group of animals within one day. Lung samples (one from each animal) from randomly selected animals were placed in a 60 °C incubator for 24 h to dry. After drying, an equal volume of 20 mg/ml proteinase K (Baisha, Shanghai, China) (prepared with ultrapure water) was added to the tissue, and the tissue was grounded using a tissue grinder (Jingxin, Shanghai, China). The samples were then treated with ultrasound for 3 min × 5 times (using an ultrasonic cleaner (KQ-800DE, Kunshan, China) set at 800 W power and a frequency of 400 kHz), followed by digestion in a 55 °C incubator for 24 h. After digestion, the samples were centrifuged at 12,000 g for 15 min, and the supernatants were discarded. The pellets were resuspended in proteinase K solution, vortexed thoroughly, and subjected to ultrasonication again, followed by re-digestion at 55 °C for 24 h. After 24 h, the sample was centrifuged again, and the steps were repeated. Finally, after tissue digestion, the pellets were dispersed in a 3% FBS solution and treated with water-bath ultrasonication for 10 min. Subsequent measurements were performed within 15 min.

UV-Visible spectrophotometry for lung burden measurement

To determine the optimal wavelength for measuring FMP with a UV spectrophotometer and to exclude interference from tissue homogenates, the lung homogenate standard (LTM) and FMP positive standards were measured using a UV-Vis spectrophotometer (UV-3600 Plus, A12616000298, Shimadzu, Japan) in quartz cuvettes in the range of 250–800 nm. Based on the wavelength distribution curve obtained from these measurements, 750 nm was selected as the experimental measurement wavelength. The prepared standards were then measured three times at 750 nm, and the absorbance values were used to plot a dose-absorbance standard curve for evaluating and calculating the FMP content in the lung tissue digestion fluid samples. The absorbance of the lung tissue digestion samples from each group was measured at 750 nm using the UV-Vis spectrophotometer. Based on the standard curve, the FMP exposure lung burden was calculated, and the results were normalized according to the lung tissue weight after drying at 60 °C. All measurements were completed within 15 min after ultrasonication.

Histology assessments

Fixed lung tissue samples were processed histologically, embedded in paraffin, and sectioned to 6 μm thickness using a Leica BIOCUT microtome (Leica, Germany). The slides were stained with hematoxylin and eosin (HE) (C0105, Beyotime, Beijing, China), modified Masson’s trichrome (G1346, Solarbio, Beijing, China), or Sirius Red staining kits (G1472, Solarbio, Beijing, China). Images of the stained slides were captured using an Olympus BX53F2 microscope (Olympus, Japan) and analyzed with ImageJ software (NIH, US). HE staining results were evaluated using the Ashcroft scoring system [29, 30, 31].

Immunohistochemistry

Immunohistochemistry was performed in lung tissue sections for Bach1 or SAT1. Please refer to supplementary material for detailed methods.

RNA-seq analysis

Chickens were exposed to PBS or 0.6 mg/kg FMP via tracheal aerosol administration as described in 2.3.2. Four animals per group were randomly chosen, and lung tissues were collected for RNA-seq analysis (OE Biotech, Shanghai, China). For details, please refer to supplementary material.

Quantitative Real-time PCR (qRT-PCR)

mRNA samples extracted from lung tissues were subjected to qRT-PCR for the mRNA expression levels of Bach1, SAT1, GPX4 and HOMX1. Please refer to supplementary material for detailed methods and primer information.

Cell culture tests

Cell culture

This study selected human bronchial epithelial cells (16HBE) and human fetal lung fibroblasts (HLF-1) for experiments, since bronchial epithelial cells are the primary targets of diesel exhaust particles once inhaled, and fibroblasts are the major contributors for pulmonary fibrosis. The 16HBE cell line was a kind gift from Dr. Yanjie Zhao [32]. HLF-1 cell line was purchased from Wuhan Procell Life Science & Technology Co., Ltd. 16HBE cells were cultured in MEM (GS0201, Source Bioscience, China) with 10% fetal bovine serum (FBS) (C04400-500, Vivacell, Germany), and HLF-1 cells were cultured in Ham’s F-12 K (PM150910, Procell, China) and 10% FBS. Both cells were cultured in a humidified 37 °C incubator with 5% CO2. Cells were seeded in 25 cm² culture flasks with 4 mL of medium and passaged to new dishes with 0.25% trypsin (T1300, Solabio, Beijing, China) after 2–3 days of proliferation.

Cell exposure to FMP and interventions

Ferrostatin-1 (Fer1), an effective antioxidant against ferroptosis, and HPPE, a potent Bach1-targeting inhibitor, were obtained from MedChemExpress (HY-100579 and HY-153040, Shanghai, China). 16HBE cells were divided into four groups: normal, Fer1/HPPE, FMP exposure, and FMP exposure + Fer1/HPPE. Upon reaching 30% confluence, the cells were pretreated with Fer1/HPPE for 4 h, followed by exposure to 20 µg/ml FMP for 48 h. While the in vitro surface area exposure level (6.25 ug/cm²) is higher than the in vivo exposure level (roughly 0.097–0.387 ug/cm² alveolar surface area), the in vivo exposure did not occur in 100% dispersion. Aggregation of particles was observed with TEM, which would result in higher localized exposure in vivo, which may explain the comparable toxicological results in vivo and in vitro.

After incubation, the culture medium was centrifuged at 14,000 rpm for 15 min at 4 °C to pellet FMP and debris. The supernatant was diluted 1:1 with Ham’s F-12 K medium and used as conditioned medium for HLF-1 cell culture for 48 h.

CCK-8 cell viability assay

16HBE cells were seeded in a 96-well plate at a density of 5,000 cells per well. Upon reaching 70% confluence, cells were pretreated with inhibitors and/or exposed to FMP as described in Sect. 2.4.2. In this experiment, 1 mM H2O2 was used to treat the cells for 4 h as a positive control. Subsequently, medium was replaced with fresh, serum-free medium, and then10 µL of CCK-8 working solution (BS350B, Biosharp, Beijing, China) was added to each well, and the cells were incubated at 37 °C for 2–4 h. Absorbance at 450 nm was measured using a BioTek Gen5 Microplate Reader (BioTek, USA).

MDA assay

16HBE cells were treated as described in Sect. 2.4.2 and exposed to 100 µM H2O2 for 4 h as a positive control. All cells were digested with 0.25% trypsin (T1300, Solarbio, Beijing, China), centrifuged at 3,000 g for 5 min, and then suspended in 0.9% saline. After sonication on ice (frequency 400 kHz, 30 s × 15), MDA levels were measured according to the instructions provided with the MDA assay kit (S0131S, Beyotime, Shanghai, China).

Intracellular ferrous Iron measurement

16HBE cells were treated as described in 2.4.2, and then lysed with ultrasonication on ice. Intracellular ferrous ion concentrations were subsequently measured using a ferrous ion detection kit (BL1147B, Biosharp, Beijing, China) as per the manufacturer’s instructions.

Intracellular ROS

Intracellular ROS levels were measured using the H2DCFDA probe (S0033S, Beyotime, Sahnghai, China). Cells were treated with 50 µg/ml Rosup reagent for 2 h as a positive control according to the manufacturer’s instructions, followed by incubation with the probe, and fluorescence intensity was measured with CytoFLEX (Beckman Coulter, USA) for each experimental group and analyzed statistically using FlowJo software (BD, USA). The H2DCFH-DA probe fluorescent analysis was conducted with a maximum excitation wavelength of 480 nm and a maximum emission wavelength of 525 nm.

Cell death assay

The Hoechst33342/PI double staining method was employed to assess cell damage and death. After treatment, cells were stained with Hoechst 33,342 and PI dyes according to the Biosharp kit instructions (BL116A-3, Biosharp, Beijing, China). In this experiment, 1 mM hydrogen peroxide treatment for 4 h was used as a positive control. Fluorescence levels were measured using CytoFLEX (Beckman Coulter, USA) and analyzed with FlowJo software (BD, USA). The ratio of PI-positive (red fluorescence) to Hoechst 33,342-negative or low-positive (blue fluorescence) cells indicated cell membrane damage and death.

Measurement of mitochondrial membrane potential

JC1 fluorescent probe (C2006-3, Beyotime, Shanghai, China) was used to detect mitochondrial membrane potential in FMP-exposed 16HBEcells following manufacturer’s instructions. The fluorescence intensities were analyzed with CytoFLEX (Beckman Coulter, USA). In this experiment, 10 µM CCCP treatment for 25 min was used as a positive control to adjust compensation. Using FlowJo software (BD, USA), the ratio of JC-1 monomers to JC-1 aggregates was analyzed to measure changes in mitochondrial membrane potential.

ELISA

16HBE cells were treated as described in 2.4.2, then the medium was collected and centrifuged at 14,000 rpm for 10 min. The resulting supernatant were subjected to an ELISA kit (F1767-B, Fankewi, Beijing, China) for TGF-β1 levels according to the manufacturer’s instructions. Additionally, treated cells (in 6-well plate) were collected and lysed with ultrasonication. Cell lysate was then subjected to ELISA kits (CB13658-Hu and CB12608-Hu, COIBO BIO, Shanghai, China) for intracellular spermidine (SPD) and spermine (SP) levels.

Transmission electron microscopy (TEM)

Lung tissue blocks (1 cm³) were fixed in 2.5% glutaraldehyde (P1126, Solabio, Beijing) at 4 °C. After dehydration through an ethanol gradient and 15 min in acetone, they were embedded in Epon 812 resin and sectioned with an ultramicrotome (Leica, Germany). Sections were stained with 2% uranyl acetate and lead citrate for 10–20 min each, dried overnight, and examined using a Zeiss AURIGA Compact transmission electron microscope.

For 16HBE cells exposed to FMP-containing medium, cells were collected with 0.25% trypsin, centrifuged at 5000 rpm for 5 min, and fixed with 2.5% glutaraldehyde at 4 °C. The fixed cell pellet underwent the same embedding and sectioning procedures and was observed with the same transmission electron microscope.

Western blotting

Proteins were extracted from animal lung tissues, 16HBE, and HLF-1 cells, and Western blotting was used to detect the expression levels of Bach1, SAT1, ALOX15, ALOX12, GPX4, SLC7A11, α-SMA, COL1A1, and TGF-β. Band analysis was performed using ImageJ (NIH, USA), and protein bands were normalized to GAPDH. Animal experiments were conducted with at least three samples from independent animals, and cell experiments were repeated at least three times independently. Animal lung tissue, 16HBE cells or HLF-1 cells were homogenized in RIPA buffer supplemented with 1:100 protease inhibitor cocktail (Epizyme Biomedical, Shanghai, China), and centrifuged at 14,000 g, 4 ℃ for 15 min. Protein levels in supernatants were measured with the BCA assay and adjusted to 1.5 µg/µl with PBS. Electrophoresis was then performed on 10–15% SDS-PAGE gels. Protein samples were then transferred to PVDF membranes, and blocked with 5% skim milk in PBST for 2 h. Membranes were then incubated overnight at 4 °C with specific primary antibodies: anti-GAPDH(1:2000, No.TA-08, ZSGBBIO, Beijing, China), anti-Bach1(1:1000, No.R389155, Zenbio, Chengdu, China), anti-SAT1(1:1000, No.bs-7244R, BIOSS, Beijing, China), anti-GPX4(1:1000, No.R381958, Zenbio, Chengdu, China), anti-SLC7A11(1:1000, No.R26116, Zenbio, Chengdu, China), anti-ALOX15(1:1000, No.YP-Ab-12198, Research Cloud Biology, Jinan, China), anti-ALOX12(1:1000, No.YP-Ab-17240, Research Cloud Biology, Jinan, China), anti-TGF-β1(1:1000, No.HA721143, HUABIO, Hangzhou, China), anti-α-SMA(1:1000, No.bs-0189R, BIOSS, Beijing, China), anti-Col1a1, (1:1000, No.R26615, Zenbio, Chengdu, China), HO1(1:1000, No.ER-1802-73, HUABIO, Hangzhou, China).After washing, membranes were exposed to secondary antibodies (goat anti-rabbit or anti-mouse IgG (Epizyme Biomedical, Shanghai, China) for 1:8000) for 2 h at room temperature. Bands were detected using ECL substrate (Boster, Beijing, China) and analyzed with ImageJ (NIH, US).

Statistical analysis

At least three independent samples were used per group. Data were presented as mean ± SD and graphs were created using Prism8 software (Graphpad, USA). IBM SPSS statistical software (SPSS Inc., Chicago, IL, USA) was used to analyze the data. In experiments with only DE or FMP exposure, one-way analysis of variance (ANOVA) was used to detect differences between groups. When ANOVA returned significant results, post hoc Least Significant Difference (LSD) tests were employed to compare differences between groups. For experiments involving FMP, inhibitors, and/or lentivirus combinations, factorial design analysis of variance was used to assess the effects of each factor. Statistical significance was set at P < 0.05.

Results

Ferrocene-doping in diesel resulted in formation of fiber-like particles in diesel exhaust

Fiber-like particles were observed in ferrocene-doped diesel exhaust particles (Fig. 1A). Morphological analysis revealed similar length and length-to-width ratio of these fiber-like particles across different ferrocene doses, with length of 6.25 ± 3.97 μm, diameter of 347.81 ± 207.87 nm, and length-to-width ratio of 16.70 ± 5.84 (Fig. 1B). EDX analysis revealed increased metallic elements (Fe, Ni, Co) on particle surfaces in ferrocene-doped exhaust comparing to ordinary diesel exhaust particulates (Fig. 1C).

Fig. 1

Characterization of ferrocene-doped diesel exhaust particle Ferrocene-doped diesel exhaust was collected and characterized with scanning electron microscopy. DEP-0: Diesel exhaust particles from diesel fuel with 0 mg/L ferrocene. DEP-205: Diesel exhaust particles from diesel fuel with 205 mg/L ferrocene. DEP-410: Diesel exhaust particles from diesel fuel with 410 mg/L ferrocene. DEP-820: Diesel exhaust particles from diesel fuel with 820 mg/L ferrocene. (A) Representative scanning electron microscopy images of ferrocene-doped diesel exhaust particles. (B) Quantification of length-diameter ratio and fiber-particles length (N = 6 per group). (C) Semi-quantification of element mass fraction of diesel exhaust particles

Extraction and characterization of ferrocene-derived magnetic fiber-particle (FMP) and assessment of lung burden/ pulmonary toxicities

Magnetic extraction successfully extracted magnetic components from ferrocene-doped diesel exhaust particles, with the yield is up to 10.11 ± 1.18% by weight at 820 mg/L ferrocene doping dose (300 ppm iron by mass) (Supplementary Fig. 1A and Fig. 2A). SEM revealed considerable amount of fiber-shaped particles with mean length of 5.64 ± 3.02 μm (Fig. 2B and Supplementary Fig. 1B). Further characterization indicated that FMPs from 820 mg/L ferrocene-doped diesel exhaust had hydrodynamic diameters of 257.64 ± 6.31 nm (in PBS) and 227.48 ± 8.08 nm (in MEM), with corresponding Zeta potentials of -38 mV and − 51 mV, respectively, indicating good dispersion in PBS or MEM (Supplementary Fig. 1C). Consistent with our previous study [34], in ovo exposure of ferrocene-doped diesel exhaust resulted in enhanced pulmonary toxicities in one-month-old animals (Supplementary Fig. 2). Meanwhile, exposing FMP to hatching chicken embryos revealed attachment of fiber-like particles on the bronchiole walls (Fig. 2C), and histological changes consistent with collagen deposition (Fig. 2DE).

The UV-Vis spectroscopy results show that FMP dispersed in ultrapure water containing 3% fetal bovine serum (FBS) exhibited a peak at 300 nm and maintained stable absorbance between 400 and 800 nm. In contrast, the absorbance of lung tissue homogenates(LTM) is near zero between 660 nm and 800 nm (Fig. 3ABC). Therefore, 750 nm was selected as the optimal wavelength for detecting FMPs in lung tissue. The standard calibration curve demonstrates a linear relationship between FMP absorbance at 750 nm and concentration(Fig. 3D). Protease K digestion of lung tissue to detect FMPs revealed that lung burden increased with the exposure dose. At exposure doses of 0.3, 0.6, and 1.2 mg/kg, the cumulative amounts of FMPs were 0.48 ± 0.12, 1.16 ± 0.41, and 1.24 ± 0.30 µg/mg of lung dry weight, respectively (Fig. 3E).

Fig. 2

Characterization and toxicological evaluation of ferrocene-derived magnetic nanoparticles (FMP). Extracted FMPs were characterized and exposed to hatchling chickens via aerosol inhalation, resulting pulmonary toxicities were evaluated in 0- and 1-month-old animals. Data were expressed as mean ± SD; * indicates significant difference from the blank control group (p < 0.05); ** indicates significant difference from the blank control group (p < 0.01). (A) Mass fraction of FMP in ferrocene-doped diesel exhaust particle samples (N = 3 per group). (B) Representative SEM images showing FMP enriched from 820 mg/L ferrocene-doped diesel exhaust particles. (C) Representative TEM images of lung tissues post-exposure, with red arrows indicating FMP attachment or entry into epithelium cells. (D) Representative pictures of lung tissue sections from animals exposed to FMP stained with H&E, Masson’s trichrome, and Sirius Red post-exposure (doses: 0.3 mg/kg, 0.6 mg/kg, 1.2 mg/kg; scale bars represent 100 μm). H&E staining indicated thickening of alveolar wall and loss of lung architecture in exposed samples, blue staining in Masson’s trichrome and red staining in Sirius Red staining represent collagen deposition. (E) Quantifications of lung tissue Ashcroft scores and fibrotic lesion ratios (N = 5 per group)

Fig. 3

Pulmonary load analysis in hatchling chickens exposed to FMP. Hatchling chickens were intratracheally administered with 0.3, 0.6 or 1.2 mg/kg FMP within 24-hour post-hatch. Animals were kept for 24 h after the exposure, and then lung tissues from each group were collected and digested with proteinase K. Pulmonary load was assessed using UV-visible spectrophotometry. Data are presented as mean ± SD. * indicates a statistically significant difference compared to the control group (p < 0.05); ** indicates a highly significant difference compared to the control group(p<0.01). (A) Experimental procedure and representative images showing the digestion of lung tissue from FMP-exposed chickens with proteinase K at 55 °C, followed by the preparation of the sample solution. FMP (only) refers to an FMP solution suspended in 3% FBS at a concentration of 24 µg/ml. (B) UV-visible absorbance measurements between 250–800 nm for lung tissue homogenates (LTM) from unexposed chickens and a 24 µg/ml FMP suspension (N = 3). (C) Quantification of absorbance at 750 nm for LTM and FMP samples (N = 3). (D) Standard curve fitting for FMP using a UV-Vis spectrophotometer, with a fixed wavelength of 750 nm (N = 3). (E) Evaluation of pulmonary load in chickens exposed to different doses of FMP using UV-Vis spectrophotometry, with results normalized to lung dry weight (N = 3 per group from independent animals)

RNA-seq and qPCR validation of FMP-exposed chicken lung tissues

RNA-seq analysis identified 2025 differentially expressed genes (DEGs) between control (PBS) group and FMP group animal lung samples, with 1402 upregulated and 1237 downregulated (Supplementary Fig. 3AB). SAT1 gene was particularly upregulated by FMP exposure (Supplementary Fig. 3C). These DEGs primarily participate in processes like phagocytosis and epithelial-to-mesenchymal transition (EMT) (Supplementary Fig. 3D). KEGG pathway analysis indicated pathways involving cell death and lipid metabolism in FMP-induced fibrosis (Supplementary Fig. 3E). Intersection with FerrDb genes revealed overlapping with Bach1, SAT1, and GPX4 genes, suggesting key gene regulation in the ferroptosis pathway may be activated. qPCR confirmed upregulation of Bach1 and SAT1, and downregulation of GPX4 following exposure to FMP, consistent with RNAseq findings (Supplementary Fig. 3F).

In vitro cytotoxic and fibrotic effects of FMP exposure

Following FMP exposure, remarkable cellular entry of FMP into 16HBE cells were observed both with optical microscopy and TEM (Fig. 2AB). Additionally, the cell viabilities of 16HBE cells exposed to 0, 30, 60, 90 µg/ml of FMP for 6, 12, 24, or 48 h were assessed with CCK-8 assay, revealing dose- and time-dependent cytotoxicity (Fig. 4C). Lower doses (0, 10, 20–40 µg/ml) FMP exposure to 16HBE cells for 48 h revealed significant direct cell damage in 20–40 µg/ml dosing groups (Fig. 4D) as well as significantly increased TGF-β concentrations in culture medium of 16-HBE cells (Supplementary Fig. 4A). Subsequently, HLF-1 cells cultured with this medium exhibited significantly enhanced expression of α-SMA and COL1 (Fig. 4E).

Fig. 4

FMP-induced toxicological effects in cultured 16HBE and HLF-1 Cells. 16HBE bronchial epithelium cells were exposed to FMP directly, cell morphology, and viability were assessed. HLF-1 cells were cultured with collected culture supernatants, α-SMA and COL1 expressions were assessed with western blotting. Data were expressed as mean ± SD; * indicates significant difference from the blank control group (p < 0.05); ** indicates significant difference from the blank control group (p < 0.01). (A) Representative optical microscope images of FMP-exposed 16HBE cells. (B) Representative TEM images of FMP-exposed 16HBE cells, showing entry of FMP into cells. (C) Representative optical microscope images of 16HBE cell morphology after exposure to different doses (0, 10, 20, 40 µg/ml) of FMP. Arrows indicate cell damage, including changes of shape and cell rupture. (D) 16 HBE cell viability assessed following exposure to varying concentrations of FMP (0, 30, 60, 90 µg/ml) after 6, 12, 24, and 48 h. (E) Representative western blotting images and semi-quantification of α-SMA and COL1 expression changes in HLF-1 cells cultured for 48 h with conditioned media from FMP-exposed 16HBE cells diluted in F’12 K culture medium (N = 3)

FMP-induced ferroptosis in 16HBE cells participated in fibrotic changes

Following exposure to 20 µg/ml FMP for 48 h, enhanced cytotoxicity (Fig. 5A), increased cellular ROS levels (Fig. 5B), damaged mitochondrial cristae and swelling rupture of mitochondrial membranes (Fig. 5E), increased MDA levels (Fig. 5D) and increased intracellular Fe2 + levels (Fig. 5C) were observed, while pretreatment with 1.0 µM Fer-1 for 4 h effectively alleviated such changes. FMP exposure remarkably decreased the expression levels of GPX4 and SLC7A11, but increased the expression levels of Bach1, SAT1 and ALOX15, which could be abolished by Fer-1 pretreatment (Fig. 5F and Supplementary Fig. 4BC). Additionally, the increased culture supernatant TGF-β levels following exposure to FMP were also significantly alleviated by Fer-1 pretreatment (Fig. 5G). In HLF-1 cells incubated with these culture supernatants, enhanced expression of α-SMA and COL1 were also abolished in those cultures from 16HBE cells pretreated with Fer-1 (Fig. 5H).

Fig. 5

FMP-induced Ferroptosis in 16HBE Cells Promotes HLF-1 Differentiation in Pulmonary Fibrosis. 16HBE cells were pretreated with ferrostatin-1 (Fer-1), then exposed to FMP. Cell viability, intracellular ROS levels, ferrous levels, MDA levels, along with key proteins participating in ferroptosis were assessed. Additionally, exposed cells were observed with TEM for characteristic mitochondria changes in ferroptosis. Data were expressed as mean ± SD, N = 3-5per group; * indicates significant difference from the blank control group (p < 0.05); ** indicates significant difference from the blank control group (p < 0.01); # indicates significant difference from 20 µg/ml FMP group (p < 0.05); ## indicates significant difference from 20 µg/ml FMP group (p < 0.01). (A) Viability of 16HBE cells following 48 h exposure to FMP (20 µg/ml), with or without pretreatment with 1µM Fer1 for 4 h. A 1 mM hydrogen peroxide treatment for 4 h was used as a positive control (N = 5). (B) Intracellular ROS levels following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 1µM Fer1 for 4 h. Cells were treated with 50 µg/ml ROSUP for 2 h as a positive control. (N = 5). (C) Intracellular Fe2 + content following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 1µM Fer1 for 4 h (N = 3). (D) Quantification of MDA content in 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 1mM Fer1 for 4 h. Cells were treated with 100 µM hydrogen peroxide for 4 h as a positive control (N = 5). (E) Representative TEM images of 16HBE cells from control group or FMP-exposed group, showing mitochondria (Mito). Arrow indicated representative morphological changes in mitochondria, including decreased mitochondria size, increased membrane density and decreased cristae. (F) Representative western blotting images for expression of key proteins (SLC7A11 and GPX4) in 16HBE cells. And semi-quantification of protein expression levels normalized to GAPDH (N = 4). (G) ELISA measurement of TGF-β in supernatants of 16HBE cell cultures following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 1µM Fer1 for 4 h (N = 5). (H) Representative western blotting images for α-SMA and COL1 expressions in HLF-1 cells cultured for 48 h with conditioned media from 16HBE cells exposed to FMP, diluted in F’12 K medium. And semi-quantification of COL1A1 and α-SMA protein expression levels relative to the internal control GAPDH in HLF-1 cells(N = 4)

FMP-induced ferroptosis is associated with Bach1-SAT1 and polyamine depletion

Pretreatment with specific Bach1 inhibitor HPPE (0.6 µM for 4 h) exerted similar protective effects as Fer-1 on cell viability, intracellular Fe²⁺ levels, ROS levels and MDA levels on 20 µg/ml FMP-exposed 16HBE cells (Fig. 6B-E). The results of JC-1 probing showed a significant decrease in the JC-1 aggregations (red dots) and increase in the JC-1 monomers (green dots), suggesting significantly decreased mitochondrial membrane potential following FMP exposure, while pretreatment with the HPPE inhibitor could alleviate the decline in membrane potential (Fig. 6F). Additionally, significantly decreased intracellular Spermidine (SPD) and Spermine (SP) levels were observed in FMP-exposed 16HBE cells, which was alleviated by pretreatment with HPPE (Fig. 6I). Western blotting results also revealed similar trends on the protein expression levels of GPX4, SLC7A11, Bach1, SAT1 and ALOX15 (Fig. 6GH).

Fig. 6

FMP promotes ferroptosis in 16HBE cells via Bach1-SAT1. 16HBE cells were pretreated with HPPE, a specific Bach1 inhibitor, and then exposed to FMP. Cell viability, intracellular ferrous levels, ROS levels, MDA levels, mitochondrial membrane potential levels, along with key proteins participating in ferroptosis were assessed. Data were expressed as mean ± SD, N = 3–5 per group; * indicates significant difference from the blank control group (p < 0.05); ** indicates significant difference from the blank control group (p < 0.01); # indicates significant difference from 20 µg/ml FMP group (p < 0.05); ## indicates significant difference from 20 µg/ml FMP group (p < 0.01). (A) Viability of 16HBE cells treated with varying HPPE concentrations under FMP exposure (IC50: 70 µg/ml). The optimal HPPE concentration was determined at 0.6 µM (N = 5). (B) Viability of 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 0.6 µM HPPE for 2 h. A 1 mM hydrogen peroxide treatment for 4 h was used as a positive control. (N = 5). (C) Quantification of intracellular Fe2 + levels in 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 0.6 µM HPPE for 2 h (N = 3). (D) Quantification of ROS levels in 16HBE cells. Cells were treated with 50 µg/ml ROSUP for 2 h as a positive control. (N = 5). (E) Quantification of MDA content in 16HBE cells. Cells were treated with 100 µM hydrogen peroxide for 4 h as a positive control(N = 5). (F) Assessment of mitochondrial membrane potential in 16HBE cells with JC-1. Cells were treated with 10 µM CCCP for 20 min as a positive control and used for compensation adjustment. (N = 3). (G) Representative Western blot images showing key molecular proteins (Bach1, ALOX15, SLC7A11, SAT1, GPX4) in 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 0.6 µM HPPE for 2 h. (H) Semi-quantification of target protein expression levels normalized to GAPDH (N = 4). (I) ELISA quantifications of spermine (SP) and spermidine (SPD) levels in 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 0.6 µM HPPE for 2 h (N = 5)

Polyamine depletion is critical for ferroptosis in FMP-exposed 16HBE cells

Both supplementation of SP or SPD partially alleviated FMP-induced cytotoxicity, as indicated by PI/Hoechst 33,342 flow cytometry results (Fig. 7A). Moreover, supplementation of low doses of SP (50 µM) or SPD (100 µM) could effectively increase their intracellular levels (Fig. 7B), and attenuate ferroptosis-related endpoints, including intracellular Fe2 + levels (Fig. 7C), MDA levels (Fig. 7D), ROS levels (Fig. 7E) and mitochondrial membrane potential (Supplementary Fig. 4E). Western blotting revealed that SP and SPD supplementation enhanced expression of HO1 and GPX4, while abolished FMP-induced upregulation of ALOX12 and ALOX15 (Fig. 7FG).

Fig. 7

Polyamine Supplementation Alleviates FMP-Induced Ferroptosis in 16HBE Cells. 16HBE cells were pretreated with spermine (SP, 50 µM) or spermidine (SPD, 100 µM), followed by 48-hour exposure to FMP (20 µg/ml). Cell death was assessed with Hoechst 33,342/PI staining. Intracellular SP/SPD levels and ROS levels, along with the protein expression levels of ALOX15, ALOX12, HO1 and GPX4 were assessed. Data were expressed as mean ± SD, N = 3 per group; * indicates significant difference from the blank control group (p < 0.05); ** indicates significant difference from the blank control group (p < 0.01); # indicates significant difference from 20 µg/ml FMP group (p < 0.05); ## indicates significant difference from 20 µg/ml FMP group (p < 0.01). (A) Cell death ratio in 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 50 µM SP or 100 µM SPD for 4 h (N = 3). The blue and red dots indicated cells marked with negative (low) PI staining or positive (high) PI staining. (B) ELISA quantification of intracellular SP levels in 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 50 µM SP or 100 µM SPD for 4 h (N = 5). And ELISA quantification of intracellular SPD levels in 16HBE cells (N = 5). (C)Quantification of intracellular Fe2 + levels in 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 50 µM SP or 100 µM SPD for 2 h (N = 3). (D)Quantification of intracellular MDA levels in 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 50 µM SP or 100 µM SPD for 2 h. Cells were treated with 100 µM hydrogen peroxide for 4 h as a positive control(N = 4). (E) Quantification of intracellular ROS levels in 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 50 µM SP or 100 µM SPD for 2 h. Cells were treated with 50 µg/ml ROSUP for 2 h as a positive control (N = 5). (F) Representative western blotting images for ALOX15, ALOX12, HO1 and GPX4 in 16HBE cells following 48 h exposure to FMP (20 µg/ml) with or without pretreatment with 50 µM SP or 100 µM SPD for 2 h. (G) Semi-quantification of ALOX15, ALOX12, HO1 and GPX4 protein expression levels (N = 4)

Inhibition of Bach1 mitigated FMP-induced pulmonary fibrosis in vivo

Embryos on day 17 pretreated with HPPE (Bach1 inhibitor, 1/2/4 mg/kg body weight) showed a significant decrease in the expression of Bach1 downstream gene HOMX1 at the 4 mg/kg dose. This dose was selected as the treatment dose for the experiment. (Fig. 8A and Supplementary Fig. 5AB). Immunohistochemistry results demonstrated that exposure to FMP resulted in increased expression of Bach1 and SAT1 in the bronchial epithelium, while pretreatment with HPPE effectively alleviated the enhanced SAT1 expression (Fig. 8B), suggesting FMP may initially activate the Bach1-SAT1 signaling pathway in bronchial epithelium, contributing to subsequent pulmonary fibrosis progression in the in vivo model. Decreased fibrotic changes were observed in animals pretreated with HPPE, as indicated by results from H&E, Masson’s trichrome, and Sirius Red staining (Fig. 8CD). Western blotting results revealed HPPE-mediated abolishment of FMP-induced changes in ferroptosis-related molecules, including SAT1, GPX4, ALOX15 and TGF-β (Fig. 8E).

Fig. 8

Inhibiting Bach1-SAT1 Pathway Alleviates FMP-Induced Pulmonary Fibrosis in vivo. Embryonic day 17 (ED17) chicken embryos were exposed to HPPE via air cell injection, and then 24 h-old chickens were exposed to 0.6 mg/kg FMP via aerosol inhalation. Bach1 and SAT1 expressions in chicken lung tissues were assessed with qRT-PCR and immunohistochemistry. Extent of pulmonary fibrosis were determined with H&E, Masson and Sirius red staining. Protein expression levels of ALOX15, SAT1, GPX4 and TGF-β were assessed with western blotting. Data were expressed as mean ± SD, N = 5 per group; * indicates significant difference from the blank control group (p < 0.05); ** indicates significant difference from the blank control group (p < 0.01); # indicates significant difference from 0.6 mg/kg FMP exposure group (p < 0.05); ## indicates significant difference from 0.6 mg/kg FMP exposure group (p < 0.01). (A) Schematic of animal treatments and qRT-PCR analysis of Bach1 and downstream HOMX1 expression in chicken lung tissues following air cell-injection of HPPE (4 mg/kg) at ED17 (N = 3). (B) Representative IHC images showing Bach1 and SAT1 protein expression in chicken lung tissues exposed to 0.6 mg/kg FMP, with or without pretreatment of 4 mg/kg HPPE, along with semi-quantifications (N = 4). Each image was captured under a 400× optical microscope, with the scale bar in the lower-right corner representing 100 μm. The lower-left corner of the image indicates a 200× magnification of the field of view, with the scale bar representing 100 μm. (C) Representative pathological pictures of chicken lung tissue sections from animals exposed to 0.6 mg/kg FMP, with or without pretreatment of 4 mg/kg HPPE, Yellow arrows represent positive areas. (D) Quantifications of Ashcroft score and fibrotic lesion ratio from C(N = 5). (E) Representative western blotting images for ALOX15, SAT1, GPX4 and TGF-β in the lung tissues samples from animals exposed to 0.6 mg/kg FMP, with or without pretreatment of 4 mg/kg HPPE, along with semi-quantifications (N = 4)

In vivo Silencing of SAT1 alleviates FMP-induced pulmonary fibrosis

In vivo silencing of SAT1 was achieved with lentivirus microinjection, as confirmed by qRT-PCR, the silencing efficacy was 52.2% (Fig. 9A). Histopathological evaluation (H&E, Masson and Sirus Red staining) revealed that SAT1 silencing effectively alleviated FMP-induced pulmonary inflammation and fibrotic changes (Fig. 9BC). Moreover, FMP exposure enhanced expression of ALOX12, ALOX15 and TGF-β, while suppressed expression of GPX4 in 1-month old lung tissue samples. SAT1 silencing abolished such changes, along with significantly elevated expression levels of HO1 (Fig. 9D).

Fig. 9

Silencing of SAT1 alleviates FMP-Induced Pulmonary Fibrosis in vivo. Embryonic day 2 (ED2) chicken embryos were micro-injected with SAT1-silencing lentivirus, and then 24 h-old-chickens were exposed to 0.6 mg/kg FMP via aerosol inhalation. SAT1 expression in chicken lung tissues were assessed with qRT-PCR. Extent of pulmonary fibrosis were determined with H&E, Masson and Sirius red staining. Protein expression levels of ALOX15, SAT1, GPX4, HO1 and TGF-β were assessed with western blotting. Data were expressed as mean ± SD, N = 5 per group; * indicates significant difference from the blank control group (p < 0.05); ** indicates significant difference from the blank control group (p < 0.01); # indicates significant difference from 0.6 mg/kg FMP exposure group (p < 0.05); ## indicates significant difference from 0.6 mg/kg FMP exposure group (p < 0.01). (A) Schematic of animal treatments and qRT-PCR confirmation of SAT1 silencing efficacy (N = 4). (B) Representative pictures of chicken lung tissue sections from animals exposed to 0.6 mg/kg FMP, with or without SAT1 silencing, Yellow arrows represent positive areas. (C) Quantification of Ashcroft score and fibrotic lesion ratio for the chicken lung tissue sections from animals exposed to 0.6 mg/kg FMP, with or without SAT1 silencing (N = 5). (D) Representative western blotting images for ALOX15, SAT1, GPX4, HO1 and TGF-β in the lung tissues samples from animals exposed to 0.6 mg/kg FMP, with or without SAT1 silencing, along with semi-quantifications, The quantification of GPX4 bands was performed using N = 5 animals, while the remaining proteins were quantified with N = 4 animals

Discussion

Diesel doping is a commonly used strategy for emission control. Iron-containing doping agents such as ferrocene had been demonstrated to improve diesel burning efficacy, decreasing emission. However, these iron-containing chemicals seem to impose intricate effects on diesel exhaust, potentially enhancing toxicities. The presence of magnetic fiber-particles in ferrocene-doped diesel exhaust particles is a reasonable potential explanation of enhanced toxicities.

Impacts of ferrocene-doping to the diesel exhaust particles

Since the identification of diesel exhaust-induced toxicities, various methods have been employed to reduce emissions. Among the strategies, doping is a cost-effective option. By adding chemicals to standard fuel, objectives such as extending engine life and reducing emissions can be achieved [33]. Ferrocene, an iron-containing doping agent, is commonly used and has been shown to enhance combustion efficiency and reduce emissions [24]. However, ferrocene doping could lead to the formation of nano-scale fibrous or tube-like particles in diesel exhaust [9, 15, 17]. In this study, consistent with these previous findings, fiber-like particles were identified in ferrocene-doped diesel exhaust, Our recent work characterized these ferrocene-doped diesel particles, revealing increased iron levels, larger particle sizes, and slightly reduced polycyclic aromatic hydrocarbons [34]. In the current study, enhanced pulmonary toxicity of ferrocene-doped diesel exhaust was confirmed with independent experiments and more comprehensive methods, clearly indicating the enhanced toxicity of ferrocene-doped diesel exhausts, thus further investigation for the source of this enhanced toxicity is necessary.

Toxicities of ferrocene-derived magnetic particle (FMP) in diesel exhaust

Magnetic particles had been reported to be an component of vehicle emission [35], the latter is an important contributor to air pollution [36]. Since significant magnet content were identified in the fiber particles identified in ferrocene-doped diesel exhaust particles, neodymium magnets were used to extract magnetic particles from ferrocene-doped diesel exhausts in the current study, which was a success, acquiring magnet particles equivalent to 7–11% of total diesel exhaust particles. This yield is consistent with Suzuki et al., who reported 9% yield of tube-like particles [16]. Fiber/tube-like, nano-scale particles with much higher prevalence were observed in the resulting FMP. Elongated, fiber-like particles tend to be exhibit higher inhalability than spherical particles [37], increasing the potential of toxicity. When this extracted FMP was exposed to animals very similar pulmonary toxicities were observed as reported in Wang et al. [34] suggesting that the potentiation of pulmonary toxicity in ferrocene-doped diesel exhaust could be at least partially explained by the existence of FMP. Very interestingly, both light microscopy and electron microscopy results also indicated that the fiber-like FMP directly entered 16HBE cells, and attached to the bronchial epithelium in exposed animal lung tissues similar to observed effects following carbon nanotube exposure [38], further suggesting bronchial epithelial cells are important potential targets. Moreover, the high iron content in the particles may also contribute to the toxicities, since disrupted iron homeostasis is an important factor in pulmonary fibrosis [39].

Ferroptosis in FMP-induced toxicities

Magnet particles often contain relatively high levels of ferroferric contents [35]. It had been reported that irons in particles are important contributors of reactive oxygen species [40], and magnetic particles-induced toxicities had also been associated with oxidative stress [3]. On the other hand, oxidative damage-associated lipid peroxidation is the key feature of ferroptosis [41], an important contributor to pulmonary fibrosis [42]. Additionally, fibrous or fiber-like particles (with a larger aspect ratio) often cause more severe toxicity than spherical particles, as their fibrous structure is more likely to disrupt intracellular ion homeostasis and inhibit cellular repair mechanisms, thus more readily inducing ferroptosis [43]. In the current study, following exposure to FMP, ferroptosis had been clearly identified in 16HBE cells, indicating that promoted ferroptosis in bronchial epithelial cells is an important mechanism of toxicity in FMP-induced pulmonary toxicity. Pretreatment with ferrostatin-1, a ferroptosis inhibitor, effectively alleviated observed fibrotic effects and ferroptosis, further confirming the crucial role of ferroptosis. Subsequent increase of TGF-beta secretion of 16HBE cells suggested that observed pulmonary fibrosis is likely contributed by ferroptosis in bronchial epithelial cells and the resulting TGF-beta secretion.

Bach1-SAT1-polyamine depletion in FMP-induced toxicities

Myriad factors could affect ferroptosis, such as iron balance, xenobiotic-induced oxidative stress and polyamine metabolism. Among them, polyamine metabolism is an important one. Spermine and spermidine serves as reactive oxygen species scavengers, regulating autophagy and intracellular metal ion homeostasis in pulmonary tissues, whose depletion had been associated with pulmonary fibrosis and lung cancer [44, 45]. The polyamine metabolism process includes polyamine synthesis and degradation pathways, with the degradation of spermidine and spermine being critical steps [46]. Previous studies have confirmed that intracellular ion homeostasis is closely related to polyamine metabolism [47]. There were also reports that SAT1 is directly involved in ferroptosis [48, 49]. In the current study, SAT1 had been identified to be up-regulated in pulmonary tissue samples from animals exposed to FMP according to RNA-seq results. Depletion of spermine and spermidine were also revealed in FMP-exposed cells. Supplementation of spermine or spermidine effectively alleviated FMP-induced fibrotic changes and ferroptosis, suggesting that SAT1-associated polyamine depletion is an important factor contributing to observed pulmonary toxicities following exposure to FMP. Furthermore, in ovo gene silencing of SAT1 also effectively abolished FMP-induced pulmonary fibrosis in animals exposed to FMP, further confirming SAT’s critical role. On the other hand, Bach1, a transcription factor, has a critical role in various biological processes, such as oxidative stress, cell cycle, heme homeostasis, and immune regulation [50]. The entry of FMP into cells may enhance oxidative stress, thereby promoting the activation of Bach1. Furthermore, Bach1 regulates intracellular iron homeostasis by modulating the expression and degradation of the downstream heme oxygenase HO1, and was also known to be associated with ferroptosis [51]. In the current study, Bach1 expression had been identified to be up-regulated in pulmonary tissue samples from FMP-exposed animals, just as SAT1. Interestingly, pretreatment with HPPE (Bach1 inhibitor) abolished SAT1 up-regulation and subsequent changes, suggesting that Bach1 may regulate SAT1 in FMP-induced ferroptosis and pulmonary fibrosis. Of course, more evidences are necessary to further prove the regulatory role of Bach1 towards SAT1 and polyamine metabolism in future. In summary, Bach1-SAT1-polyamine depletion is involved in FMP-induced ferroptosis and pulmonary fibrosis, further investigation of the exact roles of this pathway is guaranteed.

Limitations of the current study

The current study reported presence of FMP in ferrocene-doped diesel exhaust, its characterizations, toxicological impacts, and potential mechanisms. There are, however, some limitations of the current study, thus cautions should be paid in the data interpretation. First of all, this study is performed in a chicken model. Chicken model is an alternative animal model with several advantages such as easy developmental and/or early-in-life exposure, relatively large organ/tissues, cost-effectiveness, etc., but the physiological properties of the lung is not completely identical with mammals. Our studies had demonstrated that the responses in our chicken model are mostly comparable with mammal models, but cautions are still needed to extrapolate results of the current study directly to other species. Secondly, the dose used in the current study is calculated based upon the occupational exposure upper limit and a 90-day, 8-hour per day duration, which is higher than most real-world situation. Investigations with lower doses are planned. Thirdly, the particle size analyzation assumed spherical particle shape, which may not apply to the fiber-like portion of the FMP. Both the automatically analyzed particle size and manually acquired fiber dimensions were provided for reference. Last but not least, the mechanistic investigation of the current study did not fully distinguish the toxicities from the shape of particles or from the chemical components. Similarly, whether prevention of cellular entry contributes to the observed protective effects of Fer-1, HPPE, SP or SPD were not assessed. It is likely that both the shape (cellular entry) and chemical components contributes to the toxicities of FMP, but more work is planned to further elucidate these effects.

Conclusions

Magnetic particles are important contributors to air pollution. In the current study, under commonly used ferrocene doses, fiber-like magnet particles were detected, characterized, and extracted from ferrocene-doped diesel exhaust, which were shown to exert pulmonary toxicity. Underlying mechanism had been associated with ferroptosis and Bach1-SAT1-polyamine depletion. Our findings indicated that the commonly used doses of ferrocene doping in diesel fuel are potentially associated with formation of a new contaminant: toxic magnetic, fiber-like, nano-scale particles, thus further risk assessments and mechanistic works should be performed on ferrocene-based diesel doping agents and the resulting diesel exhaust particles. Moreover, the balance between emission control and health effects should be carefully considered in future development of doping agents.