Enhancing ROS-Inducing Nanozyme through Intraparticle Electron Transport

Abstract

Iron oxide nanoparticles (IONPs) have garnered significant attention as a promising platform for reactive oxygen species (ROS)-dependent disease treatment, owing to their remarkable biocompatibility and Fenton catalytic activity. However, the low catalytic activity of IONPs has been a major hurdle in their clinical translation. To overcome this challenge, IONPs of different compositions are examined for their Fenton reaction under pharmacologically relevant conditions. The results show that wüstite (FeO) nanoparticles exhibit higher catalytic activity than magnetite (Fe₃O₄) or maghemite (γ-Fe₂O₃) of matched size and coating, despite having a similar surface oxidation state. Further analyses suggest that the high catalytic activity of wüstite nanoparticles can be attributed to the presence of internal low-valence iron (Fe⁰ and Fe²⁺), which accelerates the recycling of surface Fe³⁺ to Fe²⁺ through intraparticle electron transport. Additionally, ultrasmall wüstite nanoparticles are generated by tuning the thermodecomposition-based nanocrystal synthesis, resulting in a Fenton reaction rate 5.3 times higher than that of ferumoxytol, an FDA-approved IONP. Compared with ferumoxytol, wüstite nanoparticles substantially increase the level of intracellular ROS in mouse mammary carcinoma cells. This study presents a novel mechanism and pivotal improvement for the development of highly efficient ROS-inducing nanozymes, thereby expanding the horizons for their therapeutic applications.

Graphical Abstract

In the IONP-catalyzed heterogeneous Fenton reaction, surface Fe²⁺ is more reactive than Fe³⁺. In a composite nanoparticle of magnetite (Fe₃O₄) shell and a core of low-valence iron (Fe⁰ & Fe²⁺), the low-valence iron can accelerate the catalysis by reducing the octahedral Fe³⁺ in the magnetite shell to Fe²⁺ through intraparticle electron transport.

1. Introduction

Synthetic and biogenic iron oxide nanoparticles (IONPs) possess versatile physical properties, which make them an ideal choice for biomedical applications such as magnetic resonance imaging, thermal therapy, drug/gene delivery, and magnetogenetics.^[1–3] Recent studies showed that IONPs could catalyze the generation of reactive oxygen species (ROS), especially highly reactive hydroxyl free radicals (·OH), under biologically relevant conditions.^[4] IONP-mediated ROS generation may explain the enigmas in previous studies, such as unexpected cell behaviors and adverse effects in in vivo applications.^{[5, 6]} On the other hand, the ROS-inducing property is of significant interest for disease treatment since ROS is inherently involved in numerous physiological processes.^[7] For instance, ferumoxytol, an FDA-approved IONP for treating iron deficiency, can induce the polarization of macrophages to a tumor-suppressing phenotype by generating intracellular ROS.^[8] In addition, many types of IONPs, alone or in combination with other therapeutic agents, can induce ferroptosis of cancer cells, which holds great promise for treating refractory cancers resistant to current chemotherapy.^[9–11] Compared with other ROS-inducing metal or metal oxide nanomaterials, IONPs exhibit superior biocompatibility, and several formulations have been approved for clinical use.^{[12, 13]} However, the low catalytic activity of IONPs has been a hurdle in the clinical translation of ROS-dependent therapies.

Soluble ferrous (Fe²⁺) and ferric (Fe³⁺) ions are known to catalyze the decomposition of hydrogen peroxide (H₂O₂) through classic Fenton or Fenton-like reactions, respectively.^{[14, 15]} In contrast, iron oxide-mediated catalysis follows a heterogeneous Fenton process characterized by the complex interfacial interactions among the iron ions, H₂O₂, and other co-existing compounds.^[14–16] Although the mechanism underlying the heterogeneous Fenton reaction remains unclear, studies on solid catalysts have manifested the versatility of iron oxide-mediated catalysis. Surface Fe²⁺ in solid catalysts is more reactive than Fe³⁺.^[16] However, Fe²⁺ is converted to Fe³⁺ upon reacting with H₂O₂, and the recycling of Fe³⁺ to Fe²⁺ becomes the rate-limiting step in the catalytic reaction. Interestingly, zero valence iron (Fe⁰), while not an effective Fenton agent, can enhance the catalytic activity of neighboring magnetite (Fe₃O₄).^{[17, 18]} It was proposed that the electrons in the Fe⁰/magnetite composite catalyst could flow from regions of Fe⁰ to magnetite, reducing the surface Fe³⁺ to Fe²⁺ since magnetite has high electrical conductivity and the reduction of Fe³⁺ by Fe⁰ is thermodynamically favorable.^[19] The most likely target of the reduction is the octahedral Fe³⁺ in the magnetite, as the change from Fe³⁺ to Fe²⁺ will not increase the crystal structural strain.^[17] An X-ray-absorption spectroscopy study confirmed that the reduction could occur across a few atom layers at the interface of Fe⁰ and magnetite.^[20] Likewise, a recent study found that intraparticle electron transfer via the Fe²⁺-O-Fe³⁺ in magnetite nanocrystals played a crucial role in maintaining the surface oxidation state during the Fenton reaction.^[21]

In biomedical applications, small nanocrystals are preferred due to the need for optimal colloidal stability, tissue penetration, and cellular uptake.^[22] Moreover, smaller nanocrystals are more favorable for surface-mediated catalysis due to the large surface area-to-volume ratio. However, at this size range, low-valence iron (Fe⁰ and Fe²⁺) on the crystal surface is susceptible to oxidation when exposed to air or biological medium. While various monometallic iron or iron oxide nanocrystals can be synthesized using organic iron salts under an inert atmosphere, the surface layer of nanocrystals undergoes rapid oxidation to magnetite and eventually iron (III) oxide.^[23–26] As a result, few studies on IONP-catalyzed Fenton reaction have explored nanocrystals beyond magnetite and maghemite (γ-Fe₂O₃), and research on compounds of other transition metals with high surface reactivity has been pursued instead, despite their lower tolerability by the body.

Wüstite (FeO) is a non-stoichiometric iron oxide that is metastable at temperatures below 570°C.^[27] Previous studies showed that wüstite nanocrystals could be easily produced by thermodecomposition of organic salts of iron under a reductive atmosphere.^[23] Although wüstite nanocrystals form a surface layer of magnetite through oxidation, they also undergo a disproportion reaction to Fe⁰ and magnetite, resulting in a composite structure with a magnetite surface and a core containing scattered regions of Fe⁰ and FeO. Therefore, the catalytic activity of wüstite nanoparticles would be significantly higher than that of nanocrystals of pure magnetite or maghemite if there is intraparticle electron transport from internal low-valence iron (Fe⁰ and Fe²⁺). In this regard, we systemically investigate the Fenton reaction catalyzed by wüstite, magnetite, and maghemite nanoparticles under pharmacologically relevant conditions to gain a better understanding of the Fenton catalysts at the nanoscale and provide new leverage for ROS-dependent therapeutic applications.

2. Results and Discussion

2.1. Synthesis and Characterization of IONPs

We initially generated wüstite nanocrystals by thermodecomposition of iron acetylacetonate in a mixture of oleic acid and oleylamine (Figure 1A).^{[23, 28]} Transmitted electron microscopy (TEM) images reveal that many wüstite nanocrystals exhibit a distinctive contrast variation, indicating a composition variation inside the nanocrystals (Figure 1A, inset). To ensure an impartial analysis of the Fenton reaction, we synthesized two other types of iron oxide nanocrystals that were of the same size and capping molecules. Magnetite nanocrystals were synthesized by thermodecomposition of iron acetylacetonate using a published method.^{[25, 29]} Maghemite nanocrystals (γ-Fe₂O₃) were obtained by oxidizing magnetite nanocrystals (Figure 1C).^[30] The synthesized nanocrystals were coated by a layer of oleic acid and oleylamine and were only dispersible in nonpolar solvents. After purification, the nanocrystals were dispersed in toluene and stored under nitrogen to prevent further oxidation. The nanocrystals have a uniform size distribution of approximately 15 nm (Table S1). We included ferumoxytol as a control, which consisted of heterogeneous nanocrystals (5.6 ± 2.5 nm) (Figure 1D).

Figure 1.

Synthesis of IONPs. A), B), C), and D). Representative TEM images of wüstite, magnetite, and maghemite nanocrystals and ferumoxytol. Scale bars: 50 nm. Inset: higher magnification image of nanocrystals circled in the white box. Scale bar: 10 nm. E) XRD patterns of iron oxide nanocrystals. Red, blue, and orange lines label the standard peaks of wüstite, magnetite, and maghemite, respectively.

The X-ray powder diffraction (XRD) patterns of magnetite and maghemite nanocrystals reveal the characteristic cubic spinel structure of magnetite and maghemite (Figure 1E). Notably, the peaks of maghemite nanocrystals shift toward higher angles, and two superlattice diffraction peaks, (210) and (211), characteristic of maghemite, appear around 24 °C and 26 °C. The lattice parameters of the two nanocrystals are 8.368 Å and 8.331 Å, respectively, which are close to the values reported for bulk magnetite (8.396 Å, JCPDS 19-0629) and maghemite (8.346 Å, JPCDS 39-1346), respectively (Table S1). In contrast, wüstite nanocrystals exhibit diffraction peaks of both the face-centered cubic wüstite, including (111), (200), and (220) peaks, as well as (311) and (511) peaks of magnetite (Figure 1E). However, the (311) and (511) peaks are less prominent in larger wüstite nanocrystals (Figure S1), indicating that the magnetite phase in wüstite nanocrystals is mainly formed by surface oxidation. Smaller wüstite nanocrystals are more oxidized due to their higher surface area-to-volume ratio.

The compositions of nanocrystals were further analyzed using high-resolution TEM (HR-TEM) and X-ray photoelectron spectroscopy (XPS) (Figure 2). The atomic lattice fringes observed in the dark regions inside the wüstite nanocrystal have a distance of 0.210 nm, which is close to the interspacing of (200) planes (0.216 nm) in wüstite (Figure 2A). The interplanar distance of 0.291 nm is only observed near the surface of the wüstite nanocrystal and is assigned to the 220 planes (0.296 nm) of magnetite. In contrast, the lattice fringes in the magnetite and maghemite nanocrystals are distributed uniformly throughout the nanocrystals (Figure 2B and 2C). The oxidation state of Fe in the nanocrystals was examined by XPS spectra of Fe 2p (Figure 2D). For wüstite nanocrystals, the spectrum can be fitted to four main peaks in the 2p3/2 region.^{[31, 32]} Notably, the lowest binding energy peak at 707.0 eV is attributed to Fe⁰, which is undetectable in magnetite and maghemite nanocrystals. The XPS spectrum of maghemite nanocrystals shows only peaks associated with octahedral and tetrahedral Fe³⁺, which are consistent with the oxidation state.

Figure 2.

Compositions of IONPs. A), B), and C) HR-TEM images of single wüstite, magnetite and maghemite nanocrystals. Scale bar: 5nm. Insets: higher magnification images of regions circled in the corresponding color box. Scale bar: 1 nm. D) XPS spectra of wüstite, magnetite, and maghemite nanocrystals.

The selected area electron diffraction (SAED) pattern of wüstite nanocrystals confirms the presence of wüstite, magnetite, and likely a trace amount of Fe⁰ (Figure S2). Using a modified ferrozine assay, we determined that low-valence iron (Fe⁰ and Fe²⁺) in wüstite, magnetite, and maghemite nanocrystals accounted for 50.3%, 15.8%, and 1.5% of the total iron content, respectively (Figure S3). Note that the ferrozine method cannot distinguish between Fe⁰ and Fe²⁺. The results indicate partial oxidation of wüstite and magnetite nanocrystals compared with pure wüstite and magnetite. Despite the complex compositions of the nanocrystals, we refer to them as wüstite and magnetite respectively, based on their synthesis methods for simplicity. The XRD pattern (Figure 1E), lattice parameter (8.333 Å), and oxidation state (Figure S3) of ferumoxytol indicate a composition of maghemite.

The XPS spectra of C 1s of the nanocrystals show three peaks related to the carboxylic group, amine group, and the alkyl chains of oleic acid and oleylamine binding to the nanocrystals (Figure S4). Wüstite and magnetite nanocrystals have similar surface ligand compositions, and maghemite shows a larger peak at 286.6 eV, possibly due to the oxidation reaction in the presence of oleylamine.

2.2. IONP-Catalyzed Fenton Reaction

The as-synthesized nanocrystals are not dispersible in aqueous buffers, so we utilized two different coating methods to generate water-dispersible wüstite nanoparticles. Firstly, we coated the nanocrystals with a biocompatible amphiphilic DSPE-PEG copolymer using a dual solvent exchange method.^[33] Alternatively, we replaced the oleic acid and oleylamine on the nanocrystal surface with dextran or citric acid through surface ligand exchange. Nanocrystals coated with dextran and citric acid showed additional oxidation compared with uncoated nanocrystals, while the nanocrystals with DSPE-PEG remained unchanged (Figure S5A). We evaluated the catalytic activities of wüstite nanoparticles using a colorimetric TMB (3,3’,5,5’-tetramethylbenzidine) assay, which measures the hydroxyl free radicals in the solution.^[34] The DSPE-PEG-coated nanocrystals exhibited the highest catalytic activity, despite being encapsulated inside a lipid layer formed of oleic acid, oleylamine, and phospholipids (Figure S5B). For this reason, we used DSPE-PEG coating in the following experiments. Wüstite nanoparticles displayed pH-dependent catalytic activity, which slightly shifted toward neutral pH compared with soluble Fe²⁺ and reached maximum efficiency at pH 4 (Figure S6). We carefully examined the reaction kinetics and found that at pH 4, the IONP-mediated Fenton reaction was mainly driven by iron oxide nanocrystals, while at pH 2, the reaction was partly catalyzed by surface-leached iron ions (Figure S7).

We investigated the catalytic activities of all IONPs with H₂O₂ concentrations set within a pharmacologically relevant concentration range (0-200 μM).^{[35, 36]} We monitored the real-time reaction rate of the Fenton reaction by measuring the optical absorption of the oxidation product of TMB (oxTMB).^[34] We set the concentrations of reagents to ensure that the reaction rate remained constant during the initial state of the reaction and met the assumptions of the Michaelis-Menten equation.^[37] As shown in Figure 3A, both wüstite and magnetite nanoparticles exhibited higher catalytic activities than ferumoxytol. Moreover, wüstite nanoparticles had the highest catalytic activity among all nanoparticles, with a V_max 4.62 folds higher than that of maghemite nanoparticles (Table S4).

Figure 3.

Effect of composition on the catalytic activity of IONPs. A) Catalytic activities of IONPs. B) Catalytic activities of soluble Fe²⁺ and Fe³⁺. Soluble Fe²⁺ and Fe³⁺ were generated by dissolving FeCl₂ and FeCl₃ to a final concentration of 0.1 μM. The data points were fitted by solid lines using the Michaelis-Menten equation (Table S4). Data represents mean ± standard deviation (n = 3).

It is important to note that the wüstite nanoparticles are enclosed in a magnetite shell, i.e., the surface is similar to that of magnetite nanocrystals. Most of the low-valence iron (Fe⁰ and Fe²⁺) is likely stored in the dark regions 2-3 nm below the surface (Figure 1A). Furthermore, the V_max of soluble Fe²⁺ is only 1.61 times of soluble Fe³⁺ under similar conditions (Figure 3B). If the kinetics of the IONP-catalyzed heterogeneous Fenton reaction follow that of soluble Fe²⁺ and Fe³⁺, there should be only a 1.5-fold difference even if all maghemite is replaced by wüstite. Therefore, comparisons of the catalytic activity and the composition of nanoparticles strongly support the hypothesis that in both wüstite and magnetite nanocrystals, surface-catalyzed Fenton reactions can be accelerated by internal low-valence iron (Fe⁰ and Fe²⁺), presumably by reducing surface Fe³⁺ through intraparticle electron transport (Figure 4).^{[17, 18]} The relative contribution of Fe⁰ and Fe²⁺ cannot be determined accurately due to the lack of methods to analyze the distribution of the two species at this length scale.

Figure 4.

Schematic diagram of enhancing surface Fenton reaction via intraparticle electron transport. In the IONP-catalyzed heterogeneous Fenton reaction, surface Fe²⁺ is more reactive than Fe³⁺. However, Fe²⁺ is converted to Fe³⁺ upon reacting with H₂O₂ to generate hydroxyl free radicals and recycling of Fe³⁺ to Fe²⁺ is the rate-limiting step in the catalytic reaction. In a composite nanoparticle of magnetite (Fe₃O₄) shell and a core of low-valence iron (Fe⁰ & Fe²⁺), the low-valence iron can accelerate the catalysis by reducing the octahedral Fe³⁺ in the magnetite shell to Fe²⁺ through intraparticle electron transport.

2.3. Ultrasmall Wüstite Nanoparticle-Catalyzed Fenton Reaction

To explore if high catalytic activity can be achieved by smaller wüstite nanoparticles, which are more susceptible to oxidation due to their increased surface area to volume ratio, we undertook a further examination. Although wüstite nanocrystals larger than 15 nm could be readily generated through thermodecomposition of iron (III) acetylacetonate in a reductive solution,^{[23, 38–41]} few studies have investigated the synthesis of small wüstite nanocrystals due to their instability. Therefore, we systematically explored the conditions that favor the generation of small nanocrystals. Our study found that the size of wüstite nanocrystals could be reduced by decreasing the incubation temperature (Figure 5A–F). Additionally, the nanocrystal size could be tuned by changing the concentration of the iron salt (Figure S8A) or the ratio between the solvent to capping molecules (Figure S8B). Interestingly, all methods converged at a minimum size of approximately 8 nm. As the nanocrystals decreased in size, their XRD patterns shifted from that of wüstite toward magnetite due to increased surface oxidation (Figure 5G).

Figure 5.

Synthesis of ultrasmall wüstite nanocrystals. Representative TEM images are shown for the wüstite nanocrystals synthesized with the incubation temperature set at A) 250°C, B) 260°C, C) 270°C, D) 280°C, and E) 290°C. Scale bars: 50 nm. F) Size distribution of wüstite nanocrystals vs. incubation temperature. Data represents mean ± standard deviation (n = 500). G). XRD patterns of wüstite nanocrystals in A) (8 nm), C) (12 nm), and E) (15 nm). Red and blue lines label the standard peaks of wüstite and magnetite, respectively.

Multiple batches of magnetite nanocrystals were synthesized over the same size range for comparison (Figure S9). The results showed that both wüstite and magnetite nanocrystals exhibit increased oxidation as the particle size decreases (Figure 6A). As the size decreases below 10 nm, the percentage of the low-valence iron in the wüstite nanocrystals gradually approaches that in the magnetite nanocrystals. Moreover, in these wüstite nanocrystals, the composition variations are no longer obvious in the TEM images of these small wüstite nanocrystals, indicating that the oxidized surface layer became the dominant component (Figures 5A and 5B). However, the reaction rate of the wüstite nanoparticles is inversely proportional to the size and consistently higher than that of the magnetite nanoparticles (Figure 6B). These findings suggest that a trace amount of Fe⁰ and FeO may serve as the electron pool for reducing surface Fe³⁺ via intraparticle electron transport. Importantly, the reaction rate of 8 nm wüstite nanoparticles is 5.3 times higher than that of ferumoxytol (Figure 6B). Furthermore, 8 nm wüstite nanoparticles could maintain high reactivity even after extended incubation with H₂O₂, suggesting that nanoparticles with a magnetite shell are relatively resistant to further oxidation (Figure S10).

Figure 6.

Effect of size and composition on catalytic efficiency. A) Low-valence iron (Fe⁰ and Fe²⁺) vs. the size of IONPs measured by the ferrozine method. B) Catalytic activities of representative nanoparticles circled in green lines in A). Data represents mean ± standard deviation.

2.4. IONP-Induced Intracellular ROS

The pH-dependency of wüstite nanoparticles suggests their potential as effective ROS-inducing agents in acidic intracellular organelles like lysosomes. To evaluate this potential in a biological environment, we incubated highly metastatic mouse 4T1 mammary carcinoma cells with a culture medium containing IONPs (50 μg Fe mL⁻¹ of either 8 nm wüstite nanoparticles or ferumoxytol). After 4 hours of incubation, the IONPs were internalized by 4T1 cells through endocytosis, as evidenced by their accumulation in the lysosomes (Figures 7A and 7B). The cells were then exposed to H₂O₂ (250 μM) and the level of intracellular hydroxyl free radicals was detected with CM-H₂DCFDA. While the cells treated with ferumoxytol and H₂O₂ showed a similar level of ROS as the cells treated with H₂O₂ alone, those incubated with wüstite nanoparticles and H₂O₂ exhibited significantly higher levels of ROS compared with all other groups (Figures 7C and 7D). The results indicate that wüstite nanoparticles are highly efficient in catalyzing the decomposition of H₂O₂ to intracellular ROS. Furthermore, in the presence of H₂O₂ at pharmacologically relevant concentrations, wüstite nanoparticles exhibited significantly higher cytotoxicity in the 4T1 cells compared to ferumoxytol (Figure S11).

Figure 7.

IONP-mediated ROS generation in cancer cells. Mouse 4T1 mammary carcinoma cells were incubated with wüstite nanoparticles or ferumoxytol. A) and B) Fluorescence imaging and flow cytometry evaluation of intracellular distribution of wüstite nanoparticles. The wüstite nanoparticles were labeled with DiI. Lysosomes were stained by LysoTracker^™ Deep Red. Scale bars: 20 μm. C) and D) Fluorescence imaging and flow cytometry evaluation of intracellular hydroxyl free radicals. Scale bars: 50 μm. Data represent mean ± standard deviation (n = 3). **, P< 0.01, and ns, not significant.

3. Conclusion

In summary, our study highlights the remarkable potential of wüstite nanoparticles as efficient ROS-catalytic agents, surpassing magnetite and maghemite nanoparticles such as ferumoxytol. Our findings demonstrate that wüstite nanocrystals readily form composite nanocrystals with a magnetite/maghemite shell and a core containing low-valence iron (Fe⁰ and Fe²⁺), achieved through surface oxidation and disproportionation. A comparison of size- and surface-matched IONPs indicates that the enhanced catalytic activities of wüstite nanoparticles can be attributed to the pool of low-valence iron within the nanocrystals. Analyzing electron flow at the nanoscale poses challenges due to the limitations of available characterization techniques. Nonetheless, to the best of our knowledge, this is the first study to demonstrate that intraparticle electron transport in IONPs significantly enhances the Fenton reaction. This finding also sheds light on other metal/metal oxide nanozyme-catalyzed Redox reactions. It is worth noting that IONPs with interior low valence iron can be synthesized using alternative methods. For instance, Fe⁰ core and magnetite shell nanocrystals can be synthesized through controlled surface oxidation or seed-mediated growth to form a magnetite shell around Fe⁰ nanocrystals.^[26] However, the synthesis of such a structure is considerably more challenging due to the volatility and toxicity of the precursor compound (iron pentacarbonyl) and the laborious processing involved in the two-step methods.

ROS plays a crucial role in various physiological and pathological processes and has been targeted for therapeutic purposes in numerous diseases. In the field of cancer medicine, mounting evidence suggests that inducing a high level of intracellular ROS can overcome multidrug resistance, induce cancer cell death, and trigger anticancer immunity.^[42–44] Nanoparticles offer a range of advanced therapeutic strategies, including targeted delivery, image-guided therapy, and combination therapy. Several formulations of IONPs, such as ferumoxytol and Feridex, have already gained clinical approval.^[45] Consequently, the discovery of the Fenton catalytic activity of IONPs has generated immense excitement within the field.^{[47, 48]} However, early studies utilizing ferumoxytol had to compensate for its low catalytic activity by administering exceedingly high doses, which presents a potential hurdle for clinical applications.^[8] Our study represents a crucial advancement in IONP-based cancer therapy. We demonstrate the potential of DSPE-PEG-coated wüstite nanoparticles as highly effective ROS-inducing agents in a biological environment. The DSPE-PEG coating preserves the oxidative state of the nanocrystals and provides a biocompatible surface. Importantly, DSPE-PEG is a component in FDA-approved liposomal drugs and offers a multifunctional platform for drug delivery and conjugation with targeting ligands or bioactive molecules.^{[45, 49]} These findings pave the way for further exploration of wüstite nanoparticles in ROS-driven cancer therapies.

4. Experimental Section

Materials:

Iron acetylacetonate (99%), oleic acid (technical grade, 90%), oleylamine(technical grade, 70%), benzyl ether (99%) hydrochloric acid (>37%), hydroxylamine hydrochloride (99.9%), iron(Ⅱ) chloride tetrahydrate (≥99.5%), iron(Ⅲ) chloride hexahydrate (98 - 102%), sodium hydroxide solution (10M), citric acid (≥99.5%), dextran (MW 9,000-11,000), thiazolyl blue tetrazolium bromide (≥97.5%), hydrogen peroxide (30% w/w), 3,3’,5,5’-tetramethylbenzidine (TMB), horseradish peroxidase (HRP), RPMI 1640 medium, ammonium acetate, ferrozine, and iron standard solution were purchased from Sigma-Aldrich. CM-H₂DCFDA, Amplex^® Red, DiI, and LysoTracker^™ Deep Red were purchased from Thermo Fisher Scientific. Ammonium hydroxide was purchased from VWR. 1,2-Distearoyl-sn-glycero-3-phosphoethanolamine-N [methoxy (polyethylene glycol)-2000] (ammonium salt) (DSPE-PEG) was purchased from Avanti Polar Lipids. Ferumoxytol was a gift from AMAG Pharmaceuticals. All reagents were used without further modifications.

Synthesis of iron oxide nanocrystals:

Wüstite nanocrystals were synthesized by thermal decomposition of iron acetylacetonate.^{[23, 28]} To synthesize 15 nm wüstite nanocrystals, iron acetylacetonate (12 mmol), oleic acid (75 mmol), oleylamine (105 mmol), and benzyl ether (20 mL) were mixed in a three-neck flask. The solution was heated at 120 °C for 2 hours under vacuum to remove moisture and trapped air. Then the solution was heated to 220 °C at a ramping rate of 5 °C minute⁻¹ under a nitrogen flow and held at this temperature for 1 hour. After that, the solution was heated to 300 °C at a ramping rate of 2 °C minute⁻¹ and incubated at this temperature for 30 minutes. After incubation, the solution was let to cool down to room temperature. The nanoparticles were precipitated by the addition of ethanol and centrifugation at 4,000 g for 10 minutes. The pellets were redispersed in toluene and washed with ethanol two more times. After the last wash, the nanocrystals were precipitated with acetone and the pellets were dispersed in toluene and stored under nitrogen. Wüstite nanocrystals of other sizes were synthesized using the same protocol except that one condition was adjusted as designated (Figures 1A, 4A through 4E, and S8).

Magnetite nanocrystals were synthesized using a thermodecomposition method published previously (Figures 1B and S9).^{[25, 29]} Maghemite nanocrystals were obtained by heating magnetite nanocrystals (100 mg) with oleylamine (9 mL) and benzyl ether (17.9 mL) at 130°C in the air for 50 hours (Figure 1C).

Characterization of iron oxide nanocrystals:

The size and compositions of iron oxide nanocrystals were analyzed by TEM, XRD, SAED, and XPS. TEM and SAED images were acquired using a transmission electron microscope (Talos F200X TEM) connected to a CCD camera. The size of nanocrystals was measured in TEM images using Image J software. For XRD measurements, the nanocrystals were precipitated with acetone and dried under a vacuum. XRD measurements were performed with an X-ray diffractometer (Bruker D8 ADVANCE). The elemental composition of IONPs was measured using a Thermo K-Alpha X-ray photoelectron spectroscopy instrument.

Synthesis of water-dispersible IONPs:

Water-dispersible IONPs were synthesized using two different methods. First, DSPE-PEG-coated IONPs were generated via a dual solvent exchange method.^[33] In a typical procedure, 15 nm nanocrystals (2 mg Fe in 0.4 mL toluene) were mixed with DSPE-PEG (6 mg in 0.6 mL chloroform) in a glove box filled with nitrogen. DMSO (8 mL) was added to the mixture under vigorous stirring. After the toluene and chloroform were removed under vacuum, Millipore water (16 mL) was added to the solution. DMSO was removed by centrifugation in a centrifugal filter tube (MW = 100K). Finally, free DSPE-PEG was removed by ultracentrifugation at 80,000 g.

Dextran-coated IONPs were generated using a published protocol.^[50] In brief, to remove the oleic acid and oleylamine on the nanocrystal surface, the nanocrystals (20 mg Fe) were incubated sequentially in methanol (50 mL) at 40°C for 8 hours and acetone (50 mL) at 40°C for 48 hours with stirring. Next, the nanocrystals were incubated with dextran solution (50 mL, 200 mg in 0.1 M ammonium hydroxide, pH 11) at 40°C for 48 hours. After incubation, IONPs were washed with Millipore water using a centrifugal filter tube (MW = 100K). To generate citric acid-coated IONPs, the nanocrystals (10 mg Fe in 5 mL of toluene) were mixed with citric acid (40 mg in 5 mL of DMSO). The mixture was sonicated at 40°C for 3 hours. After sonication, IONPs were washed with ethanol to remove the free citric acid.

All coated IONPs were dispersed in Millipore water, sterile filtered by passing through a 0.22 μm syringe filter, and stored at 4°C until further use. The hydrodynamic diameter of the IONPs was measured by dynamic light scattering (DynaPro Nanostar, Wyatt Technology) (Tables S2 and S3). The mass-weighted diameter and polydispersity index were reported.

Quantification of iron content of IONPs:

The iron content of samples was determined using a ferrozine assay.^[29] Briefly, the sample (50 μL) was mixed with HCl (12 M, 50 μL) for 10 minutes. Then, NaOH (2 M, 240 μL), ammonium acetate (4 M, 50 μL), hydroxylamine HCl (5% in water, 110 μL), and Millipore water (500 μL) were added sequentially. After 5 minutes of incubation, the solution (40 μL) was mixed with ferrozine solution (0.1% in water, 60 μL) in a 384-well plate. Light absorption was read at 562 nm with 810 nm as the reference wavelength using a microplate reader (Tecan Spark Microplate Reader). The iron concentrations of the samples were determined by comparing them to a standard curve generated using iron standards. To quantify the low-valence iron (Fe⁰ and Fe²⁺), the samples were processed as above except that the hydroxylamine HCl solution was replaced by Millipore water.

Quantification of catalytic activity of IONPs:

The IONP-catalyzed hydroxyl free radical generation was measured using a 3,3’,5,5’-tetramethylbenzidine (TMB) assay.^[34] Upon oxidation by hydroxyl radicals, TMB will generate 3,3’,5,5’-tetramethylbenzidine diamine (oxTMB). In a typical experiment, IONPs were dispersed in sodium acetate buffer (0.2 M). TMB was dissolved in DMF. TMB solution was first mixed with H₂O₂ solution at designated concentrations in sodium acetate buffer in a 96-well plate. Then IONPs were added to the mixture. The final concentrations of IONPs, H₂O₂, and TMB were fixed at 20 μg Fe mL⁻¹, 0.2 mM, and 1 mM, respectively and the sodium acetate buffer was adjusted to pH 4.0 unless otherwise specified. The production of oxTMB was monitored by its light absorbance at 652 nm. The concentration of oxTMB was calculated using a molar extinction coefficient of 3.9 × 10⁴ M⁻¹cm⁻¹ for reactions at pH 4.0. The molar extinction coefficient of oxTMB at other pH values was measured based on a pH titration curve. The reaction rate was calculated as the slope of oxTMB concentration vs. time. The reaction rate was fitted with the Michaelis–Menten equation,

$$V=\frac{V_{max}\left[S\right]}{K_m+\left[S\right]}$$ (S1)

where V is the reaction rate, V_max is the maximal reaction rate, [S] is the concentration of the substrate (H₂O₂) and K_m is the Michaelis constant.

Cellular uptake of IONPs:

Mouse 4T1 breast adenocarcinoma cells were purchased from ATCC. The cells were cultured in RPMI-1640 supplemented with 10% fetal bovine serum according to the vendor’s instructions. 8 nm wüstite nanoparticles were labeled with DiI according to an established method.^[45] 5×10⁴ 4T1 cells were seeded in a chambered glass slide and incubated overnight. Then the cells were incubated with culture media containing IONPs (50 μg Fe mL⁻¹) for 4 hours at 37°C. Next, cells were treated with LysoTracker^™ Deep Red and Hoechst to stain the lysosomes and nuclei. The fluorescence images were acquired with a Nikon Eclipse Ti2 Inverted Fluorescence Microscope.

IONP-induced intracellular hydroxyl free radicals:

4T1 cells were seeded in a chambered glass slide. 24 hours later, the cells were incubated with the culture medium containing IONPs (50 μg Fe mL⁻¹) for 4 hours. After that, the cells were washed with PBS twice and incubated with the culture medium containing H₂O₂ (250 μM) for 4 hours. Then the cells were incubated with the fresh medium containing CM-H₂DCFDA (10 μM) for 1 hour. After washed with PBS, the cells were imaged with the fluorescence microscope. Flow cytometry analysis was performed with a BD FACSymphony cell analyzer.

Statistics:

All measurements were performed at least in triplicate. GraphPad Prism 9 (GraphPad Software) was used for all the calculations. Data were analyzed using one-tailed Student’s t-tests or one-way analysis of variance (ANOVA) with post-hoc Tukey tests. A difference of P < 0.05 was considered statistically significant (*P < 0.05; **P < 0.01; ns, not significant).

Data Availability Statement

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