Anthropogenic Iron Oxide Nanoparticles Induce Damage to Brain Microvascular Endothelial Cells Forming the Blood-Brain Barrier

Abstract

Background:

Iron nanoparticles, mainly in magnetite phase (Fe₃O₄ NPs), are released to the environment in areas with high traffic density and braking frequency. Fe₃O₄ NPs were found in postmortem human brains and are assumed to get directly into the brain through the olfactory nerve. However, these pollution-derived NPs may also translocate from the lungs to the bloodstream and then, through the blood-brain barrier (BBB), into the brain inducing oxidative and inflammatory responses that contribute to neurodegeneration.

Objective:

To describe the interaction and toxicity of pollution-derived Fe₃O₄ NPs on primary rat brain microvascular endothelial cells (rBMECs), main constituents of in vitro BBB models.

Methods:

Synthetic bare Fe₃O₄ NPs that mimic the environmental ones (miFe₃O₄) were synthesized by co-precipitation and characterized using complementary techniques. The rBMECs were cultured in Transwell® plates. The NPs-cell interaction was evaluated through transmission electron microscopy and standard colorimetric in vitro assays.

Results:

The miFe₃O₄ NPs, with a mean diameter of 8.45±0.14 nm, presented both magnetite and maghemite phases, and showed super-paramagnetic properties. Results suggest that miFe₃O₄ NPs are internalized by rBMECs through endocytosis and that they are able to cross the cells monolayer. The lowest miFe₃O₄ NPs concentration tested induced mid cytotoxicity in terms of 1) membrane integrity (LDH release) and 2) metabolic activity (MTS transformation).

Conclusion:

Pollution-derived Fe₃O₄ NPs may interact and cross the microvascular endothelial cells forming the BBB and cause biological damage.

Keywords

Blood-brain barrier cytotoxicity magnetite nanoparticles

INTRODUCTION

Synthetic iron oxide (Fe₃O₄) is found in a wide variety of materials used essentially for the industrial sector. It is released to the environment as particulate matter (PM), which may impact human health [1]. Recently, Maher and co-workers showed the abundant presence of Fe₃O₄ nanoparticles in human brains from subjects who were born, lived and died in Mexico City, Mexico and Manchester, England [2]. The emission of this airborne PM is mainly attributed to the abrasion of vehicle’s combustion cylinders and disc brakes, subway networks and industrial activity [3]. Their findings highlighted a greater nanoparticles (NPs) concentration in those people from Mexico City and two different Fe₃O₄ morphologies were observed in the human brain tissue: angular (or faceted) particles, with sizes ranging from some tens to hundreds nanometers, attributed to endogenous formation, and rounded Fe₃O₄ (magnetite) NPs containing a certain Fe₂O₃ (maghemite) phase amount with diameters between 8 and 60 nm, which were attributed to external pollution-derived sources [4]. The impact of air pollution to human health was also discussed in terms of other PM-associated metals and their co-occurrence with that found in anthropogenic magnetite particles. Moreover, authors proposed the olfactory nerve axons as the direct entry route of up to 200 nm magnetite NPs to the brain, this by an analogy with other PM-associated materials [5, 6] that have been found in the olfactory bulb. However, smaller PMs have the ability to penetrate to the respiratory tract and translocate to the bloodstream and from there to extra-pulmonary organs, including the brain [7, 8].

Some studies have demonstrated the ability of bare Fe₃O₄ NPs to cross the blood-brain barrier (BBB) [9, 10] in a significant amount. In the brain, Fe^+3, +4 ions are used in different fundamental biological processes, which include oxygen transportation, neurotransmitter synthesis, DNA and myelin synthesis, mitochondrial respiration, and metabolism [11]. However, iron accumulation into the brain has been associated with neurodegenerative diseases, including Alzheimer’s and Parkinson’s diseases, mainly due to its ability to induce oxidative stress and activate inflammatory responses, which cause cellular damage [12 –16]. There are also a number of studies describing the toxic effects of Fe₃O₄ NPs, which have been used as drug delivery systems or diagnostic tools. However, there are just a few studies describing the interaction of Fe₃O₄ NPs with the microvascular endothelial cells that form the BBB and the possible harmful consequences that airborne PM-derived metal oxide nanoparticles could trigger on this critical natural barrier [17 –23]. Using a simple design, the present work describes the interaction between as pollution-derived Fe₃O₄ NPs and primary rat brain microvascular endothelial cells (rBMECs) grown in monolayer. Hence, we synthetized bare Fe₃O₄ NPs that mimic the environmental ones (miFe₃O₄ NPs) and we evaluated how they interact with the rBMECs monolayer that represents the main type of cells forming the BBB.

MATERIALS AND METHODS

Materials

The FeCl₂.4H₂O (99%), FeCl₃.6H₂O (99%), and NaOH chemicals, used to prepare the Fe₃O₄ NPs, were purchased from Sigma-Aldrich, St. Louis, MO. The rat brain microvascular endothelial cells (rBMECs, Catalog #R1000), endothelial cell medium (ECM, Catalog #1001), endothelial cell growth supplement (ECGS, Catalog #1052), fetal bovine serum (FBS, Catalog #0025), and antibiotics (penicillin and streptomycin, Catalog #0503) were purchased from ScienCell Research Laboratories, Carlsbad, CA. Dulbecco’s Modified Eagle Medium (DMEM) was obtained from Mediatech, Manassas, VA. Trypsin-EDTA (0.25%) phenol red (Gibco™), lactate dehydrogenase (LDH) Cytotoxicity Assay Kit and MTS Cell Proliferation Colorimetric Assay Kit were purchased from Thermo Fisher Scientific, Waltham, MA. Transwell^® filters were purchased from Corning Costar, Lowell, MA. 96-wells plates were obtained from BD Falcon, San Jose, CA.

Synthesis of iron oxide nanoparticles

The miFe₃O₄ NPs were synthetized through co-precipitation as described elsewhere [24]. They were prepared under alkaline conditions using a 1:2 molar ratio of Fe²⁺ and Fe³⁺ chloride. The particles precipitated immediately as a black powder. The overall reaction may be written as follows [25]: $F e^{2} + 2 F e^{3 +} + 8 O H^{-} \to F e_{3} O_{4} + 4 H_{2} O$ (1)

Prior characterization, the obtained powders were rinsed with distillated water and dried at 60°C.

Physicochemical characterization

The morphology and particle size were determined by scanning electron microscopy (SEM) using a dual beam scanning electron microscope (FIB/SEM) FEI-Helios Nanolab 600 at the maximum operating voltage of 5 kV. Specific features, such as crystalline arrangement and periodicity of the sample, were analyzed by high-resolution transmission electron microscopy (HR-TEM) using a JEOL ARM200-F microscope operated in scanning mode (STEM). The size distribution histogram was built by analyzing the obtained SEM micrographs using ImageJ (version 1.51j8, Wayne Rasband, USA). The diameter of more than 150 objects was measured and statistically analyzed.

The atomic arrangement, structural fingerprint and material identification were determined by X-ray diffraction (XRD) and Raman spectroscopy. The diffraction pattern was obtained using a SmartLab Rigaku diffractometer equipped with a Cu-K_α radiation source (λ= 1.5406 Å). The sample was measured under atmospheric conditions at the 2θ angle range from 15 to 100 degrees. Rietveld refinement was used to find the crystallographic parameters of the analyzed material. The Raman spectrum was recorded, from 200 and 1200 cm^–1, at room temperature in a Micro-Raman Renishaw spectrophotometer using a He-Ne excitation laser (632.8 nm).

The NPs oxidation state was evaluated by X-ray photoelectron spectroscopy (XPS) using a PHI 5000 VersaProbe-II spectrometer (Physical Electronics) equipped with a monochromatic aluminum source (Al-K_α, 1486.7 eV) operating at 15 kV. The sample was introduced in an ultra-high vacuum chamber (1.333×10^–7 Pa) for the measurement. The Fe-2p and O-1s core levels were analyzed.

The magnetization of the sample was evaluated as a function of the applied field (0±25000 Oe) at room temperature using a DynaCool magnetometer, PPMS (Physical Properties Measurement System) from Quantum Design.

The hydrodynamic diameter (H_d) was determined in phosphate buffered saline (PBS 1×, pH = 7.4) suspensions (5 μg/mL) at room temperature using a Nanotrac Wave II (Microtrac MRB) dynamic light scattering (DLS) analyzer equipped with a solid state near-infrared laser-diode (λ= 780 nm).

Cytotoxicity evaluation

Adherent rBMECs were grown in ECM medium enriched with 2% FBS, 1% ECGS and 1% antibiotics (10,000 units/mL penicillin + 10,000 μg/mL streptomycin) at 37°C under a humidified 95% air/5% CO₂ atmosphere.

The cytotoxicity of the miFe₃O₄ NPs was evaluated by performing two different colorimetric assays based on: 1) the measurement of the lactate dehydrogenase (LDH) released from the cytosol of damaged cells to the supernatant as a marker of cellular membrane damage, and 2) the metabolic activity of viable cells capable to reduce the MTS water-soluble tetrazolium salt.

Hence, the integrity of the rBMECs’ membrane, after exposure to miFe₃O₄ NPs, was determined by LDH assay (Cytotoxicity Detection Kit LDH 11644793001, Roche®). Briefly, 6×10³ cells per well were seeded in 96-well microplates and incubated overnight for adhesion. Prior to NPs exposure, the ECM culture medium was replaced by high glucose DMEM (Dulbecco’s Modified Eagle’s medium, HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid), phenol red indicator free) enriched with 2% FBS, 1% ECGS growth factor and antibiotics (10,000 IU/mL penicillin + 10,000 IU streptomycin, ScienCell). Cells were then incubated for 24 h with miFe₃O₄ NPs at different concentrations (0, 1, 5, 15, 25, 40, 75, and 100 μg/mL), conditions based on previous in vitro reports [26 –28]. A positive control, consisting of rBMECs incubated with a lysis solution (2% Triton X-100 in DMEM), was included in the analysis. At the end of the exposure period, cell culture supernatants were recovered and the monolayer was gently rinsed with PBS (Ca⁺⁺ and Mg⁺⁺ free, pH = 7.4) at room temperature for their posterior use in the MTS transformation assay. LDH assay was carried out with culture supernatants according to the manufacturer’s specifications. Absorbance was measured at 492 nm, using 630 nm readings as reference in order to eliminate chemical interference of compounds. The results are expressed as the mean±standard deviation (SD) of triplicate measurements applying one-way ANOVA analysis.

Cellular viability, as a function of the metabolic activity, was determined using the tetrazolium reagent MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, Promega®). The reduction of MTS by viable cells generates a colored formazan product, which is soluble in culture media and can be quantified by measuring the absorbance at 490–500 nm. The MTS transformation consists in transferring a H⁺, from the NAD(P)H-dependent cellular oxidoreductase enzymes, to the tetrazolium salt [29]. Briefly, The MTS/PMS reagent was added according to the manufacturer’s recommendations to the cells that were gently rinsed with PBS after removing the supernatant for the LDH assay and kept at 37°C under 5% CO₂ atmosphere for 2 h. The absorbance was measured as indicated before; the results are the mean (±SD) of triplicate measurements analyzed by one-way ANOVA.

miFe₃O₄ NPs uptake, localization, and transport across a BBB model

The cell-cell and NPs-cell interactions were studied by TEM, which allows monitoring the morphological quality of in vitro biological barrier models. The resolution of an electronic microscope allows scrutinizing structures at the cellular and nanoscale level. Thus, the quality of a cell monolayer can be easily characterized before and after performing transport studies. In this work, rBMECs monolayers grown in the apical chamber of a Trans-well® were used as in vitro model. Illustration in Fig. 1 shows a comparison between the structure of the human BBB and the rBMECs monolayer cultured in a Transwell® system.

Fig.1

a) Hypothetical scheme showing miFe₃O₄ NPs crossing the BBB from the brain capillaries to the central nervous system. b) In vitro Transwell®-system based experimental model to study miFe₃O₄ NPs interaction (internalization and transcytosis) using rBMECs.

The samples were prepared after adapting a protocol described elsewhere [30]. Briefly, 1 mL of an optimal seeding density of 5×10⁴ cells in ECM were added to the apical side of 6-well Transwell® plates containing a polyester (PET) membrane (0.45 μm pore size). The inserts were coated with a 2% fibronectin matrix while the basolateral chamber was filled with 3 mL of enriched ECM without cells. Cells were cultured at 37°C in a humidified 5% CO₂ atmosphere for 24 h for adhesion. The medium at the apical chamber was then removed and replaced with fresh medium to avoid any non-attached cells that may disturb the monolayer. The medium was changed every 3 days until reaching a 100% confluence (5–6 days). The monolayers were cultured for another 48 h to assure 100% cellular confluence and to avoid empty intercellular spaces. The ECM was then replaced by DMEM (2% FBS) containing miFe₃O₄ NPs (5 and 40 μg/mL) and monolayers were incubated for 24 h under standard conditions. The control samples consisted of cells incubated with DMEM high glucose, 2% FBS only. To evaluate NPs transcytosis in the rBMEC monolayer, after incubation, culture medium from the basolateral Transwell® chambers was recovered to determine total Fe content by inductively coupled plasma (ICP-OES) spectroscopy; the monolayers were then gently rinsed with PBS. Cells were fixed with a 4% glutaraldehyde-PBS solution at 4°C for 23 h. Afterwards, cells were rinsed again with PBS to remove all reagent traces, 0.5 mL and 1.5 mL of an osmium tetroxide solution (1% – OsO₄) were added to the apical and basolateral chamber, respectively, and left for 1 h in order to stain cellular monolayers. Excess of OsO₄ was discarded and washed out using a Sorensen phosphate buffer bath for 10 min. The monolayers were then dehydrated with increased concentrations of ethanol (from 30% to absolute ethanol), and embedded in Spurr’s resin (kit R1032, AGAR Scientific, UK). The resin on samples was polymerized at 65°C and finally the samples were cut into ultrathin (70–80 nm) square sections using an ultramicrotome. The obtained slices were deposited on 3.05 mm diameter copper grids 300 mesh (IACCSA, Mexico), stained with uranyl acetate and lead citrate, and observed using a JEOL 200 CX TEM electron microscope, settled to 80 kV.

RESULTS

Physicochemical characterization of the miFe₃O₄ NPs

The miFe₃O₄ NPs were synthesized by co-precipitation of FeCl₂ and FeCl₃ at alkaline conditions. Their physicochemical properties, such as morphology, size, atomic arrangement, oxidation state and degree of magnetization, were determined by complementary microscopic and spectroscopic techniques.

In Fig. 2, scanning electron microscopy (SEM; a,b), scanning transmission electron microscopy (STEM; c), and high-resolution transmission electron microscopy (HR-TEM; d) micrographs display the morphology, particle size distribution and atomic arrangement of the obtained material.

Fig.2

Electron micrographs of the miFe₃O₄ NPs synthesized by co-precipitation using a Fe²⁺ and Fe³⁺ 1:2 molar ratio. Scale bars: a) and b) SEM, 200 and 50 nm, c) STEM, 10 nm and d) HR-TEM, 2 nm allowing crystal arrangement observations. Particle size distribution from e) SEM and f) DLS analysis.

Figure 2a and 2b show particles with an average diameter of 8.45±0.14 nm (histogram, Fig. 2e). Fig. 2c reveals objects crystallinity and in Fig. 2d, the HR-TEM micrograph clearly evidences the atomic arrangement of nanoparticles. The interplanar spacing of 4.957 Å and 2.971 Å correspond respectively to the (111) and (220) planes of Fe₃O₄ (JCPDS 99-0073).

The sharp diffraction peaks in Fig. 3a are in good agreement with the d-spacing values of magnetite (Fe₃O₄) and maghemite (Fe₂O₃) (96-900-5813 and 96-900-6317 entries). The contribution of each phase was determined using the Rietveld refinement technique [31]. Magnetite was found in a slightly larger amount (54.1%) in comparison with maghemite (45.9%) in the sample, with an estimated crystallite size of 70.9 nm (±0.01 nm), being this result most likely due to an agglomeration state. This result comes from the analysis performed using the Debye-Scherrer equation (Match! 3 software, Crystal Impact, Bonn, Germany).

Fig.3

a) XRD diffractogram compared with the 96-900-5813 and 96-900-6317 entries for respectively Fe₃O₄ and Fe₂O₃ phases identification. b) Raman spectrum showing vibrational modes of both Fe₃O₄ and Fe₂O₃ phases.

In agreement with XRD measurements, Raman spectroscopy indicated that the obtained material is composed of both magnetite and maghemite phases. In Fig. 3b, the band around 650 cm^–1 is attributed to the symmetric vibrational stretching mode (A_1g) of oxygen atoms along the Fe-O bonds of the magnetite phase [33, 34]. The bands observed at 717, 520 and 357 cm^–1 are consistent with the A_1g, T_2g, and E_g1 vibrational modes of the maghemite phase, which correspond respectively to the symmetric Fe-O stretching mode and, symmetric and asymmetric vibrational Fe-O bending modes [34, 35].

In Fig. 4a, the high-resolution Fe 2p XPS spectrum displays the characteristic Fe 2p_3/2 and Fe 2p_1/2 spin-orbit components, at respectively 711.1 and 724.4 eV. The spin–orbit splitting is close to 13.3 eV, corresponding to the standard value that confirms the presence of magnetite [36]. The satellite peaks, located at higher binding energies, are attributed to the oxide γ-Fe₂O₃ species [37]. At the O 1s core level (Fig. 4b), the main peak at 530.3 eV corresponds to Fe-O bonds, while the contribution around 531.1 eV is attributed to organic C-O, most likely from adventitious carbon [38].

Fig.4

XPS spectra of the obtained miFe₃O₄ NPs at the (a) Fe 2p and (b) O 1s core level regions.

The obtained bare miFe₃O₄ NPs are able to demagnetize almost spontaneously from their saturated state as shown in Fig. 5, conferring them partially (due the sample is not 100% in magnetite phase) super-paramagnetic properties. The magnetization curve shows the typical reversible S-shaped with near zero remanence field. They have a magnetic susceptibility of 27.6 emu.g^–1 acquiring most of their magnetization at fields <1000 Oe. This is in good agreement with the magnetic behavior of the airborne PM found in human brain tissue [2], indicating the dominant presence of ferrimagnetic minerals (e.g., magnetite and/or maghemite) (1000 Oe = 100 mT).

Fig.5

Magnetization curve of the bare miFe₃O₄ NPs displaying their magnetic susceptibility at room temperature.

In PBS, miFe₃O₄ NPs have a hydrodynamic diameter close to 130±33 nm (% volume > 90, Fig. 2f) and a Polydispersity Index (PdI) of about 0.203±0.066. These parameters reveal the presence of NPs aggregates as those observed by SEM, which are mainly attributed to the uncoated NPs surface. As mentioned above, this result is most likely due to agglomeration state of the mixed magnetic state of both found phases, i.e., magnetite and maghemite. At the studied conditions, the miFe₃O₄ NPs have an average ζ-potential of –25.6±2.3 mV. The physicochemical characteristics of the as synthetized bare miFe₃O₄ NPs allows emulating the human exposure to pollution-derived iron oxide nanoparticles and thus, to investigate in vitro the toxicity and pathway to reach the brain by crossing the rBMECs forming the BBB.

miFe₃O₄ NPs cytotoxicity determination by LDH and MTS

The observed rBMECs’ morphological changes (optical micrographs in Fig. 6a) were attributed to the presence of miFe₃O₄ NPs. At higher concentrations, 25 and 40 μg.mL^–1,bigger NPs agglomerates were observed (arrows, Fig. 6a), as well as, significant cellular structural changes. Figure 6b shows the interaction of NPs with the cellular plasma membrane measured by LDH leakage after 24 h exposure (A492 nm) at concentrations ranging from 0 to 100 μg.mL^–1. A significant cytotoxicity, characterized by the increase in LDH leakage, was found only at 5 μg.mL^–1 (p < 0.05) miFe₃O₄ NPs. Indeed, at this concentration, the LDH leakage increased by a factor of 2.4 in contrast with untreated rBMECs. We suggest that at 5 μg.mL^–1, bare miFe₃O₄ NPs suspensions are more stable and form smaller aggregates, which may preferentially interact with the cellular plasma membrane through their higher surface area.

Fig.6

a) Optical micrographs showing rBMECs’ monolayers in the presence of miFe₃O₄ NPs at concentrations ranging from 0 to 40 μg.mL^–1. b) NPs cytotoxicity evaluated by measuring LDH leakage after 24 h of incubation. Results represent the mean±SD (n = 3). *p < 0.05 compared with the control group. c) Cells viability evaluated by measuring formazan production (MTS transformation) after 24 h of incubation. Results represent the mean±SD (n = 3). Significant mean difference (p < 0.05) between the concentrations: 40-C and 40-1 (*), the indicated groups (**) and 100-1 (***).

The effect of miFe₃O₄ NPs on the metabolic activity of rBMECs was evaluated through the MTS tetrazolium salt transformation to formazan, which was determined at 492 nm. In Fig. 6c, it is observed that MTS transformation significantly increased after 24 h exposure to 15 and 40 μg.mL^–1 miFe₃O₄ NPs exposure compared to control unexposed endothelial cells, suggesting that at such concentrations NPs induce an increase on the cellular metabolic activity. We propose that the observed discrepancy in the LDH and MTS results, under the probed experimental conditions, is due to different type of interaction between NPs with cells or cells organelles, which in turn is linked to the NPs used concentration; while dispersed 5 μg.mL^–1 miFe₃O₄ NPs may be able to interact closely with rBMECs membrane and cellular organelles, most of the NP at higher concentrations (15 and 40 μg.mL^–1) form aggregates due to presence of serum in culture medium, and interact differently with cells, i.e., just a small proportion of these last NPs are able to interact closely with cells as the 5 μg.mL^–1 NPs do, causing damage to the membrane.

Internalization of miFe₃O₄ NPs by rBMECs

The internalization of miFe₃O₄ NPs by rBMECs was evaluated by TEM. Figure 7 shows TEM micrographs of rBMECs monolayers grown on Trans well apical chambers and incubated only with 0, 5 and 40 μg.mL^–1 miFe₃O₄ NPs. The chambers were embedded in epoxy resin and sectioned into 70 nm thick slices.

Fig.7

Internalization assay was conducted in rBMECs’ monolayers using two different Fe₃O₄ NPs concentrations (5 and 40 μg/mL) by TEM. a, b) Control cells showing TJs and an intact cell membrane. c, d) Representative images of rBMECs exposed to 5.0 μg/mL bare Fe₃O₄ NPs; images show NPs inside the cells and structures that resemble unorganized mitochondrial membranes. e, f) Representative images of rBMECs exposed to 40.0 μg/mL bare Fe₃O₄ NPs; multiple vesicles can be observed at the cell membrane that appeared to lose integrity, NPs agglomerates can be observed inside lysosomes and TJs present a greater contrast than those in control cells. TJs, tight junctions; V, vesicles; L, lysosome; M, mitochondria; EP, endocytic pit. Scale bars: 0.5 μm.

The micrographs clearly display the different sections of the culture; i.e., the filter, the rBMECs monolayer and the apical side. Control cells showed endocytic pits free of NPs, unloaded intracellular vesicles, as well as some electrodense structures that very likely correspond to tight junctions (TJs) located between two adjacent endothelial cells. On the other hand, treated rBMECs samples (Fig. 7c–f), show intracellular electron dense particles, corresponding to the miFe₃O₄ NPs, in the cytoplasm, inside vesicles that could correspond to lysosomes, as well as inside the mitochondria (Fig. 7e). Importantly exposed cells also showed intracellular structures that resembled damaged mitochondria (Fig. 7c) and may be assumed as another miFe₃O₄ NPs toxic effect. Other interesting observation, different from what it was expected, is that which seems to correspond to TJs proteins, due to their clear intercellular localization, look more electrodense in cells that were exposed to the NPs compared with non-exposed cells suggesting an increased expression in miFe₃O₄ NPs exposed cells. This hypothesis should be probed further in another work.

Transcytosis of miFe₃O₄ NPs across the rBMECs monolayer

To evaluate the ability of miFe₃O₄ NPs to conduct transcytosis on rBMEC monolayers, after exposing the rBMECs cultured in the Transwell® upper chamber to miFe₃O₄ NPs (5 and 40 μg.mL^–1) for 24 h, the culture medium located in the lower chamber was removed and analyzed by ICP-OES to determine the total Fe concentration. Medium from the lower (basolateral) chamber of not exposed rBMECs cells was included as negative control. Results revealed higher Fe concentrations in the culture medium from the basolateral chamber of cell monolayers exposed to the NPs than controls. After adjusting to control conditions, results displayed a 13.7% and 6.4% of miFe₃O₄ NPs transcytosis after 24 h incubation with 40 μg.mL^–1 and 5 μg.mL^–1 respectively. These observations strongly suggest that miFe₃O₄ NPs carry out transcytosis through the rBMECs monolayers under our experimental conditions.

DISCUSSION

Fe₃O₄ NPs are naturally found as crystals in the worldwide environment, but nowadays they are also abundant as emitted PM in urban environment, particularly in cities with high traffic [39]. Maher and co-workers had found airborne Fe₃O₄ NPs in postmortem brains from people who lived in highly trafficked areas in Mexico and England. These Fe₃O₄ NPs showed characteristics as shape, crystallinity, and magnetic properties that resembled superparamagnetic iron oxide nanoparticles (SPIONs), which clearly differ from biogenic magnetite NPs [2]. Such report suggested that environmental magnetite NPs could get into the brain through the olfactory nerve and eventually contribute to neurodegenerative diseases development in exposed populations all over the world. The hypothesis of Maher and co-workers is based in the fact that most of the literature studies point out to the fact that bare Fe₃O₄ NPs (independently of their size and concentration) are strong inducers of oxidative stress, cell cycle arrest, and apoptosis in different cell lines [39 –42]. In one of the most recent studies, Malvindi and co-workers demonstrated that bare Fe₃O₄ NPs are highly toxic for HeLa and A549 cells in terms of oxidative stress and genotoxicity. They demonstrated that these effects are induced by magnetite intracellular in situ degradation, releasing a significant amount of iron ions in lysosomes [39]. On the other hand, due to the wide size range of pollution-derived magnetite, it is also likely that smaller particles could ingress to the body’s blood circulation from the respiratory tract and then cross the BBB from the brain microcirculation through the brain microvascular endothelial cells. Whether coming from the brain microcirculation or entering through the olfactory nerve, NPs can exert their toxic effects on the microvascular endothelial cells forming the BBB. To date, there are only scarce in vitro studies evaluating the interaction between bare Fe₃O₄ NPs and these cells, therefore more studies are needed in this direction, giving their potential repercussions on neurodegenerative deceases, both from social and scientific standpoints.

In this sense, only some reported studies focus on the toxicity of magnetite NPs coated with different compounds on brain parenchyma cells to evaluate their biocompatibility as therapeutic agents as drugs delivery systems to treat brain tumors. In the present work, we evaluated the NPs ingress and toxicity in terms of membrane integrity and cellular metabolism on endothelial cells that form the BBB. Here it was shown that miFe₃O₄ NPs, sharing the main physicochemical characteristics to those found by Maher and colleagues in postmortem brains, seems to be internalized into the rBMECs due to the high amount of vesicles observed in the cellular membrane and the intracellular vesicles loaded with NPs, we propose that the most likely internalization pathway could be the endocytosis. Transcytosis through the rBMEC monolayer is also highly suggested for these NPs, since 6.4% and 13.7% of total initial Fe concentration was measured in the basolateral Transwell chamber when cellular monolayer cultured in the apical Transwell chamber was exposed to 5 μg.mL^–1 and 40 μg.mL^–1 of miFe₃O₄ NPs, respectively. Similar results have been already reported using a human cerebral microvascular endothelial cells (hCMECs) monolayer to compare NPs-cells interaction between bare and surface functionalized Fe₃O₄ NPs [43]. In such work, authors demonstrated 30% internalization and 13% transcytosis across hCMECs after 24 h exposure to bare magnetite (50 μg.mL^–1), amount that increased when Fe₃O₄ NPs were covered with poly(2-(methacryloyloxy)ethyl phosphorylcholine) (PC), poly(ethylene glycol) (PEG), or the combination of both (PC-PEG). In the same work, these authors evaluated the particles cytotoxicity in terms of cell viability and membrane fluidity, finding that 50 μg.mL^–1 bare Fe₃O₄ NPs had not effect on cells viability but lead to a significant decrease in the cell membrane fluidity compared with unexposed cells [43]. At our work conditions, cellular toxicity evaluation showed a discrete but significant effect on membrane integrity, measuring LDH release, but this only when cells where exposed to 5 μg.mL^–1, which correlated with the high amount of membrane pits observed by TEM. Although LDH results were not significantly different from the control to the rBMECs exposed to 40 μg.mL^–1, both TEM and even optical micrographs, showed a significant cellular damage after 24 h exposure. On the other hand, results from MTS transformation were inconsistent with LDH results since they suggest an increased mitochondrial activity in rBMECs treated with 15 and 40 μg.mL^–1. In this regard, it has been reported that due to the redox active surface of Fe₃O₄ NPs, it is commonly observed that they lead to deleterious effects on the mitochondrial electron flow, changing the mitochondrial functionality, leading to artifacts on the MTS assays (or its analogues MTT and XTT) and to wrong conclusions [44]. Therefore, the MTS results in this work are not conclusive. TEM images also suggest that Fe₃O₄ NPs may be not only inside cellular vesicles but also co-localizing with mitochondria. Interestingly, on images from cells exposed to 5 μg.mL^–1, there were observed structures that resemble highly disorganized mitochondria, characterized by multiple-layered membranous rounded forms that have been observed in other works when mitochondria fail to conduct fusion or fission [45]. In this regard, it has been reported swollen mitochondria with lysing cristae in human hepatoma BEL-7402 cells after exposure to 0.5 μg.mL^–1 of bare Fe₃O₄ NPs that correlated with early cellular apoptosis induction [42].

As the present work is a descriptive study, more in vitro and in vivo studies should be conducted in the future to better understand the kind of interaction existing between airborne bare magnetite NPs and the BBB. However, the observations from this preliminary work suggest that miFe₃O₄ NPs interact and penetrate the vascular endothelial cell monolayer most likely through endocytosis and appear to interact with cellular substructures such as lysosomes and mitochondria. These results also suggest that miFe₃O₄ NPs pass through the cells rapidly, causing discrete damage to the cell membrane, but without disintegrating the cell monolayer.

Footnotes

ACKNOWLEDGMENTS

The authors thank Drs. Mariela Bravo Sánchez, Olga Araceli Patron Soberano, José Luis Sánchez Llamazares, and M.Sc. Beatriz Adriana Rivera, Dulce Partida, Guillermo Vidriales and Ana Iris Peña Maldonado, for SEM, XRD, XPS, TEM, uranil cell stain and ultramicrotome preparation for TEM, PPMS, and ICP-OES characterization. We thank also to the Laboratory of Immunotoxicology and Genomics-Metabolic (FCQ-UASLP), Laboratory of Instrumental Analysis and to the National Laboratory for Research in Nanoscience and Nanotechnology (LINAN) at IPICYT, for the use of its facilities. L.G.V. acknowledges the scholarship number 453890 from CONACYT.

Authors’ disclosures available online ().

References

Gminski

, Decker

, Heinz

, Seidel

, Könczöl

, Goldenberg

, Grobéty

, Ebner

, Gieré

, Mersch-Sundermann

(2011) Genotoxic effects of three selected black toner powders and their dimethyl sulfoxide extracts in cultured human epithelial A549 lung cells in vitro. Environ Mol Mutagen 52, 296–309.

Maher

, Ahmed

IAM

, Karloukovski

, MacLaren

, Foulds

, Allsop

, Mann

DMA

, Torres-Jardón

, Calderon-Garciduenas

(2016) Magnetite pollution nanoparticles in the human brain. Proc Natl Acad Sci U S A 113, 10797–10801.

Gieré

, Querol

(2010) Solid particulate matter in the atmosphere. Elements 6, 215–222.

Kirschvink

, Kobayashi-Kirschvink

, Diaz-Ricci

, Kirschvink

(1992) Magnetite in human tissues: A mechanism for the biological effects of weak ELF magnetic fields. Bioelectromagnetics Suppl 1, 101–113.

Oberdörster

, Sharp

, Atudorei

, Elder

, Gelein

, Kreyling

(2004) Translocation of inhaled ultrafine particles to the brain. Inhal Toxicol 16, 437–445.

Wang

, Liu

, Jiao

, Lao

, Li

, Gu

, Li

, Ge

, Zhou

, Li

, Zhao

, Chai

, Chen

(2008) Time-dependent translocation and potential impairment on central nervous system by intranasally instilled TiO(2) nanoparticles. Toxicology 254, 82–90.

Calderón-Garcidueñas

, Serrano-Sierra

, Torres-Jardón

, Zhu

, Yuan

, Smith

, Delgado-Chávez

, Cross J

, Medina-Cortina

, Kavanaugh

, Guilarte

(2013) The impact of environmental metals in young urbanites’ brains. Exp Toxicol Pathol 65, 503–511.

Calderón-Garcidueñas

, Calderón-Garcidueñas

, Torres-Jardón

, Avila-Ramírez

, Kulesza

, Angiulli

(2015) Air pollution and your brain: What do you need to know right now. Prim Health Care Res Dev 16, 329–345.

Pinel

, Thomas

, Boura

, Barberi-Heyob

(2018) Approaches to physical stimulation of metallic nanoparticles for glioblastoma treatment. Adv Drug Deliv Rev 138, 344–357.

10.

Thomsen

, Linemann

, Pondman

, Lichota

, Kim

, Pieters

, Visser

, Moos

(2013) Uptake and transport of superparamagnetic iron oxide nanoparticles through human brain capillary endothelial cells. ACS Chem Neurosci 4, 1352–1360.

11.

Ward

, Zucca

, Duyn

, Crichton

, Zecca

(2014) The role of iron in brain ageing and neurodegenerative disorders. Lancet Neurol 13, 1045–1060.

12.

Kilburn

(2000) Effects of diesel exhaust on neurobehavioral and pulmonary functions. Arch Environ Health 55, 11–17.

13.

Chen

, Kwong

, Copes

, Tu

, Villeneuve

, van Donkelaar

, Hystad

, Martin R

, Murray

, Jessiman

, Wilton

, Kopp

, Burnett

(2017) Living near major roads and the incidence of dementia, Parkinson’s disease, and multiple sclerosis: A population-based cohort study. Lancet 389, 718–726.

14.

Costa

, Cole

, Coburn

, Chang

Y-C

, Dao

, Roqué

(2017) Neurotoxicity of traffic-related air pollution. Neurotoxicology 59, 133–139.

15.

Coburn

, Cole

, Dao

, Costa

(2018) Acute exposure to diesel exhaust impairs adult neurogenesis in mice: Prominence in males and protective effect of pioglitazone. Arch Toxicol 92, 1815–1829.

16.

Roqué

, Dao

, Costa

(2016) Microglia mediate diesel exhaust particle-induced cerebellar neuronal toxicity through neuroinflammatory mechanisms. Neurotoxicology 56, 204–214.

17.

Huang

, Atwood

, Hartshorn

, Multhaup

, Goldstein

, Scarpa

, Cuajungco

, Gray

, Lim

, Moir

, Tanzi

, Bush

(1999) The Aβ peptide of Alzheimer’s disease directly produces hydrogen peroxide through metal ion reduction. Biochemistry 38, 7609–7616.

18.

Huang

, Cuajungco

, Atwood

, Hartshorn

, Tyndall

, Hanson

, Stokes

, Leopold

, Multhaup

, Goldstein

, Scarpa

, Saunders

, Lim

, Moir

, Glabe

, Bowden

, Masters

, Fairlie

, Tanzi

, Bush

(1999) Cu(II) potentiation of Alzheimer Abeta neurotoxicity. Correlation with cell-free hydrogen peroxide production and metal reduction. J Biol Chem 274, 37111–37116.

19.

Fahmy

, Cormier

(2009) Copper oxide nanoparticles induce oxidative stress and cytotoxicity in airway epithelial cells. Toxicol Vitr 23, 1365–1371.

20.

Tabner

, Mayes

, Allsop

(2010) Hypothesis: Soluble Aβ oligomers in association with redox-active metal ions are the optimal generators of reactive oxygen species in Alzheimer’s disease. Int J Alzheimers Dis 2011, 1–6.

21.

Miller

, Wang

, Telivala

, Smith

, Lanzirotti

, Miklossy

(2006) Synchrotron-based infrared and X-ray imaging shows focalized accumulation of Cu and Zn co-localized with β-amyloid deposits in Alzheimer’s disease. J Struct Biol 155, 30–37.

22.

Everett

, Collingwood

, Tjendana-Tjhin

, Brooks

, Lermyte

, Plascencia-Villa

, Hands-Portman

, Dobson

, Perry

, Telling

(2018) Nanoscale synchrotron X-ray speciation of iron and calcium compounds in amyloid plaque cores from Alzheimer’s disease subjects. Nanoscale 10, 11782–11796.

23.

Khan

, Dobson

, Exley

(2006) Redox cycling of iron by Aβ₄₂. Free Radic Biol Med 40, 557–569.

24.

Gupta

, Gupta

(2005) Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications. Biomaterials 26, 3995–4021.

25.

Schwertmann

, Cornell

(2000) Iron oxides in the laboratory: Preparation and characterization, Wiley-VCH, Weinheim.

26.

Mahmoudi

, Simchi

, Imani

, Milani

, Stroeve

(2009) An in vitro study of bare and poly(ethylene glycol)-co-fumarate-coated superparamagnetic iron oxide nanoparticles: A new toxicity identification procedure. Nanotechnology 20, 1–8.

27.

Karlsson

, Cronholm

, Gustafsson

, Möller

(2008) Copper oxide nanoparticles are highly toxic: A comparison between metal oxide nanoparticles and carbon nanotubes. Chem Res Toxicol 21, 1726–1732.

28.

Feng

, Liu

, Huang

, Chen

, Huang

, Xiao

(2018) Uptake, distribution, clearance, and toxicity of iron oxide nanoparticles with different sizes and coatings. Sci Rep 8, 1–13.

29.

Präbst

, Engelhardt

, Ringgeler

, Hübner

(2017) Basic colorimetric proliferation assays: MTT, WST, and Resazurin. Methods Mol Biol 1601, 1–17.

30.

, Dawson

, Lynch

(2015) A TEM protocol for quality assurance of in vitro cellular barrier models and its application to the assessment of nanoparticle transport mechanisms across barriers. Analyst 140, 83–97.

31.

Guivar

, Martínez

, Anaya

, Valladares

, Félix

, Dominguez

(2014) Structural and magnetic properties of monophasic maghemite (ding-Fe₂O₃) nanocrystalline powder. Adv Nanoparticles 3, 114–121.

32.

Zhou

, Rahaman

(1993) Hydrothermal synthesis and sintering of ultrafine CeO₂ powders. J Mater Res 8, 1680–1686.

33.

Majumder

, Dey

, Bagani

, Dey

, Banerjee

, Kumar

(2015) A comparative study on the structural, optical and magnetic properties of Fe₃O₄ and Fe₃O₄@SiO₂ core-shell microspheres along with an assessment of their potentiality as electrochemical double layer capacitors. Dalton Trans 44, 7190–7202.

34.

Soler

MAG

, Qu

(2012) Raman spectroscopy of iron oxide nanoparticles. In Raman Spectroscopy for Nanomaterials Characterization. Challa

, ed. Springer Materials, Berlin, pp. 379–416.

35.

Shebanova

, Lazor

(2003) Raman spectroscopic study of magnetite (FeFe₂O₄): A new assignment for the vibrational spectrum. J Solid State Chem 174, 424–430.

36.

Yamashita

, Hayes

(2008) Analysis of XPS spectra of Fe²⁺ and Fe³⁺ ions in oxide materials. Appl Surf Sci 254, 2441–2449.

37.

Kokate

, Garadkar

, Gole

(2013) One pot synthesis of magnetite-silica nanocomposites: Applications as tags, entrapment matrix and in water purification. J Mater Chem A 1, 2022–2029.

38.

Greczynski

, Hultman

(2017) C 1s peak of adventitious carbon aligns to the vacuum level: Dire consequences for material’s bonding assignment by photoelectron spectroscopy. Chemphyschem 18, 1507–1512.

39.

Malvindi

, De Matteis

, Galeone

, Brunetti

, Anyfantis

, Athanassiou

, Cingolani

, Pompa

(2014) Toxicity assessment of silica coated iron oxide nanoparticles and biocompatibility improvement by surface engineering. PLoS One 9, 1–11.

40.

Mahmoudi

, Hofmann

, Rothen-Rutishauser

, Petri-Fink

(2012) Assessing the in vitro and in vivo toxicity of superparamagnetic iron oxide nanoparticles. Chem Rev 112, 2323–2338.

41.

Mai

, Hilt

(2017) Magnetic nanoparticles: Reactive oxygen species generation and potential therapeutic applications. J Nanopart Res 19, 1–10.

42.

Kai

, Xiaojun

, Ximing

, Zhenqing

, Qiqing

(2011) Cytotoxic effects and the mechanism of three types of magnetic nanoparticles on human hepatoma BEL-7402 cells. Nanoscale Res Lett 6, 1–10.

43.

Ivask

, Pilkington

, Blin

, Käkinen

, Vija

, Visnapuu

, Quinn

, Whittaker

, Qiao

, Davis

, Ke

, Voelcker

(2018) Uptake and transcytosis of functionalized superparamagnetic iron oxide nanoparticles in an in vitro blood brain barrier model. Biomater Sci 6, 314–323.

44.

Patil

, Thorat

, Shete

, Bedge

, Gavde

, Joshi

, Tofail

SAM

, Bohara

(2018) Comprehensive cytotoxicity studies of superparamagnetic iron oxide nanoparticles. Biochem Biophys Rep 13, 63–72.

45.

John

, Shang

, Li

, Renken

, Mannella

, Selker

JML

, Rangell

, Bennett

, Zha

(2005) The mitochondrial inner membrane protein mitofilin controls cristae morphology. Mol Biol Cell 16, 1543–1554.

Anthropogenic Iron Oxide Nanoparticles Induce Damage to Brain Microvascular Endothelial Cells Forming the Blood-Brain Barrier

Abstract

Background:

Objective:

Methods:

Results:

Conclusion:

Keywords

INTRODUCTION

MATERIALS AND METHODS

Materials

Synthesis of iron oxide nanoparticles

Physicochemical characterization

Cytotoxicity evaluation

miFe3O4 NPs uptake, localization, and transport across a BBB model

RESULTS

Physicochemical characterization of the miFe3O4 NPs

miFe3O4 NPs cytotoxicity determination by LDH and MTS

Internalization of miFe3O4 NPs by rBMECs

Transcytosis of miFe3O4 NPs across the rBMECs monolayer

DISCUSSION

Footnotes

ACKNOWLEDGMENTS

References

miFe₃O₄ NPs uptake, localization, and transport across a BBB model

Physicochemical characterization of the miFe₃O₄ NPs

miFe₃O₄ NPs cytotoxicity determination by LDH and MTS

Internalization of miFe₃O₄ NPs by rBMECs

Transcytosis of miFe₃O₄ NPs across the rBMECs monolayer