Toxicity Assessment of Iron Oxide Nanoparticles Based on Cellular

Subscriber access provided by UNIV OF DURHAM

Article

Toxicity Assessment of Iron Oxide Nanoparticles Based on Cellular Magnetic Loading Using Magnetophoretic Sorting in a Trapezoidal Microchannel Fengshan Shen, and Je-Kyun Park Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b03875 • Publication Date (Web): 06 Dec 2017 Downloaded from http://pubs.acs.org on December 14, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Toxicity Assessment of Iron Oxide Nanoparticles Based on Cellular Magnetic Loading Using Magnetophoretic Sorting in a Trapezoidal Microchannel Fengshan Shen and Je-Kyun Park*

Department of Bio and Brain Engineering, Korea Advanced Institute of Science and Technology (KAIST), 291 Daehak-ro,Yuseong-gu, Daejeon 34141, Republic of Korea

* To whom correspondence should be addressed. E-mail: [email protected]. Phone: +82-42350-4315. Fax: +82-42-350-4310.

1

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 31

ABSTRACT: To accurately assess potential nanotoxicity on the basis of cellular iron content, the precise separation of cells into subpopulations according to their magnetic nanoparticle loading is of crucial importance. In this study, we developed a microfluidic magnetophoresis device consisting of a trapezoidal channel containing five side outlet branches and a narrow rectangular channel with three outlet branches. This unique structure enabled the sequential separation of cells loaded with tiny amounts of iron oxide and cells heavily labeled with iron oxide, in a single device. As a proof of concept, we attempted the sequential separation of Raw 264.7 cells with a large heterogeneity in uptake capabilities (1–50 pg of iron per cell). Consequently, we were able to differentiate the bulk cell population into seven subpopulations according to their mean iron oxide loading. We also evaluated potential nanotoxicity effects using the production of excess reactive oxygen species (ROS) and the inhibition of proliferation on the separated subpopulations, and found that 46.6% of cells loaded with iron above the threshold value (16.4 pg) had higher ROS levels than the control group. Cells loaded with more than 3.7 pg of iron exhibited transiently inhibited cell cycle progression. In particular, cells loaded with more than 35.4 pg of iron exerted a significant effect on cell proliferation. The proposed system could be useful in the investigation of nanotoxicity effects of iron oxide nanoparticle-induced cells, based on their iron oxide nanoparticle loading.

2


Page 3 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Magnetic iron oxide nanoparticles, consisting of a maghemite or magnetite core coated with a functionalized biocompatible polymer surface shell, have a wide range of interesting properties which make them useful in biological applications.1 With sizes ranging from a few nanometers to several hundred nanometers, magnetic oxide nanoparticles have a range of magnetic features, including ferromagnetism, paramagnetism, and even superparamagnetism.2 The combination of these unique magnetic properties of iron oxide nanoparticles with their tunable size, chemical composition, and varying surface chemistries due to coating materials, offer considerable promise in biomedical applications such as magnetic resonance imaging (MRI) contrast agents,3 separation of magnetically labeled cells, and magnetic heat generators for hyperthermia.4 To exploit these applications, the cells must be labeled with sufficient amounts of magnetic nanoparticles via an internalization process. The main process of cellular internalization is endocytosis,5 which varies from cell to cell.6−8 For example, inflammatory cells such as macrophages show a high degree of heterogeneity in magnetic nanoparticle uptake,9−11 with some cells being heavily labeled and other cells containing only tiny amounts of iron oxide nanoparticles. Recently, several studies have shown that iron oxide nanoparticles, despite the initial belief that they were non-cytotoxic, cause undesirable interference in cellular function, such as the production of excess reactive oxygen species (ROS),12,13 cytoskeleton defects,14 and proliferation inhibition.15 These potential toxicity issues are dose-dependent,16 indicating that cells with higher iron loading cause greater toxicity. Therefore, toxicity effects induced by iron oxide nanoparticles based on cellular iron content15,17−19 should be analyzed in cells exhibiting significant heterogeneity in uptake capabilities. To implement this concept, cells must first be sorted into subpopulations according to their magnetic loading after the labeling. The assessment of iron oxide nanoparticle cellular uptake 3



Page 4 of 31

and potential toxicity at the subpopulation level is more reliable than that obtained from particles in bulk. However, most current techniques analyze the constituent iron atoms or ions dissociated from iron oxide nanoparticles; for example, inductively coupled plasma atomic emission spectroscopy (ICP-AES)20 and ultraviolet/visible (UV/VIS) spectrometry.21 In general, these techniques digest cells in transparent solution before analysis, and measure the concentration of iron in the solution, making further toxicity analysis of the sample impossible. These approaches provide information on the average value of the bulk cell sample and evaluate only non-viable cells. To overcome these limitations, magnetophoresis, a method that not only quantifies the amount of iron uptake by cells but also sorts cells by their iron loads, has been recommended.9 This method is based on the determination of the magnetically induced velocity that results from a magnetically labeled cell in a magnetic field gradient. Chalmers et al.22 quantified the velocity of magnetically labeled cells that remained active and also displayed differences in uptake capabilities between individual cells. Thereafter, Robert et al.10 demonstrated that cells could be separated in a microfluidic magnetophoresis device based on their endocytotic capacity. During magnetophoresis, cells are deflected by the flow from a permanent magnet positioned near the separation channel. The magnetophoretic force on a cell depends on the quantity of magnetic nanoparticles internalized by the cells. In this study, we proposed a strategy to apply a microfluidics-based magnetophoretic separation method that distinguishes magnetically labeled cells via the amounts of iron loading to directly analyze corresponding cell toxicities induced by the magnetic nanoparticles. The use of microfluidics-based magnetophoresis in toxicity assays makes it possible to link the amount of internalized nanoparticles to the toxicity data, thereby providing more reliable toxicity results.17 In particular, when cells exhibit significant variation in uptake capability, the effectiveness of the 4


Page 5 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


proposed method is more evident. Here, we investigated the nanoparticle toxicity using a macrophage cell line (Raw 264.7) that shows the wide variation in internalization capability from 1 to 50 pg of iron per cell, as a proof-of-concept demonstration. Specifically, we examined the effects of magnetic iron oxide nanoparticles toxicity on viability, ROS production, and proliferation. We also demonstrated improved separation resolution of poorly labeled cells using a microchannel with a trapezoidal structure.

Theory Magnetophoresis is a migration phenomenon exhibited by particles under the influence of magnetic fields. In the x-direction, the liquids flow straight through the channel. Without the magnetic field, the cells are focused to the one sidewall which is away from the magnet. When an inhomogeneous magnetic field is applied perpendicular (y-direction) to the direction of the flow, the magnetically labeled cells following the laminar flow undergo a global magnetic moment and experience an attracting force from the magnetic field deflecting from the straight flow path. As the surrounding medium for the cells is nonmagnetic, the magnetic force that causes magnetically labeled cells to move is the total magnetic forces of the magnetic nanoparticles inside the cells. The total magnetic force, Ftm, of the magnetic nanoparticles inside the cell is the sum of the magnetic forces acting on each magnetic nanoparticle internalized by the cell.27

۴୲୫ = ܰ۴୫ = ܰ

௏ౣ ∆ఞ∇۰ ૛ ଶఓబ

(1)

where N is the number of magnetic nanoparticle within a cell, Fm is the magnetic force on a magnetic nanoparticle in the cell, Vm is the volume of magnetic material per particle, B is the 5



Page 6 of 31

magnetic field, and µ0 is the vacuum permeability. ∆χ is the net volume magnetic susceptibility of magnetic nanoparticles inside the cell and is assumed to be approximated with that of magnetic nanoparticles. When the cells are moved by the total magnetic force, the cell experiences a viscous drag force, FD that is equal but opposite to the total magnetic force.

۴ୈ = −6ߨܴୡ ߟ‫ݒ‬୫ୟ୥

(2)

۴ୈ = −۴୲୫

(3)

where Rc is the cell radius, η is the medium viscosity, vmag is the velocity of the cell resulting from the magnetic force. Combining eqs 1 and 2 into eq 3, the magnetic velocity of the cell is represented by eq 4.

‫ݒ‬୫ୟ୥ =

ே௏ౣ ∆ఞ∇۰ మ ଵଶగோౙ ఎఓబ

(4)

The magnetic velocity of magnetically labeled cells is proportional to the number of nanoparticles internalized by cells and the strength of the magnetic field. In a rectangular structure, which is commonly used, the cells are initially exposed to a low magnetic field because they are focused on a sidewall, away from the magnet. As the cells are drawn toward the magnet, they experience a larger magnetic force due to the stronger magnetic fields near the magnet. In such a situation, cells with a very small iron oxide content, which initially experience a negligibly small magnetic force, show slight divergence from the focusing flow path; it is difficult to differentiate such subpopulations from other cells in a single device. To overcome this problem, we developed a microfluidic magnetophoresis device consisting of a trapezoidal channel with five side outlet branches toward the magnet and a narrow rectangular channel with three outlet branches. This trapezoidal structure facilitates the exposure of cells with low iron oxide content to stronger magnetic fields by narrowing the width of the channels. 6


Page 7 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Although the five side branches cause the increasing flow rate to be slightly reduced, the magnetic exposure time is still shortened due to the narrowing of the channel width. To compensate for this effect, a long narrow rectangular channel is added to the system. As shown in Figure 1, cells heavily labeled with iron oxide nanoparticles are rapidly deflected to the upper sidewall and therefore take a lower outlet number, whereas cells with tiny amounts of iron oxide nanoparticles receive only a negligibly small magnetic force. When cells with tiny amounts of iron oxide nanoparticles flow alongside the focusing stream, the tapered geometry forcibly exposes the cells to stronger magnetic fields. Under the influence of these enhanced magnetic fields, the cells are deflected to a greater extent and are later retrieved at the corresponding outlets. Therefore, cells loaded with higher iron oxide content will exit more quickly from the separation channel, whereas cells with lower iron oxide content are retrieved more slowly from the cascading side outlets.

EXPERIMENTAL SECTION Chip Design. A microfluidic device for magnetophoretic separation was fabricated by a poly(dimethylsiloxane) (PDMS) (Sylgard 184; Dow Corning, MI, USA) molding process. The device featured a trapezoidal structure, including five outlets of 100-µm wide side branches (outlets 1–5), a 300-µm-wide and 15-mm-long rectangular channel connected with three outlets of 100-µm wide branches (outlets 6–8), an 8-mm-long inlet channel for cell suspension, and a focusing inlet for the buffer solution, as shown in Figure 2. Additionally, the trapezoidal channel had two pairs of the parallel and non-parallel sides. The parallel pair consisted of a narrow 300µm-wide segment and a broad 3000-µm-wide segment, whereas the two nonparallel sides had a 30-mm-long straight sidewall and a 5° inclined side. A total of five lateral outlet channels 7



Page 8 of 31

(outlets 1–5) for the recovery of the separated cells were connected to the nonparallel straight sidewall, each 5 mm apart. The overall height of the channel was designed to be 100 µm. The two-dimensional numerical analysis of the magnetic flux density was solved by a finite-element method program (Aladdin Enterprises, Menlo Park, CA, USA) and the cell trajectories were calculated using MATLAB 7.0 (MathWorks, MA, USA). Magnetic Labeling by Endocytosis. Superparamagnetic nanoparticles (diameter: 120 nm, Carboxyl Adembeads 0211; Ademtech, Pessac, France) consisted of a Fe2O3 magnetic core encapsulated by a highly cross-linked hydrophilic polymer shell. Their surface displayed a carboxylic acid functionality. For simulation, the volume magnetic susceptibility of superparamagnetic nanoparticles was measured by a superconducting quantum interference device (SQUID) magnetometer (Quantum Design MPMS; Quantum Design, San Diego, CA, USA), as previously reported.28 The nanoparticle solution of 15 µL with 3% volume fraction of magnetic materials was analyzed. Before magnetic labeling, the magnetic nanoparticles were incubated in medium containing human AB serum (Sigma-Aldrich) that generated a protein corona to render nanoparticles more suitable for the intracellular uptake process.23 RAW 264.7 cells (murine macrophage-like cell line) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin at 37°C in a 5% CO2 atmosphere. For magnetic labeling, 3.6 × 104 cells/cm2 were incubated for 4 h at an iron concentration of 0.75 mM. The cells were then collected, centrifuged, and resuspended at a cell concentration of 2 × 106 cells/mL. Cell Sorting Experiments. A neodymium magnet (25-mm width × 25-mm height × 50-mm length) was placed 1.5 mm from the separation channel containing trapezoidal and straight regions. Cell suspension and focusing buffer were injected into the microfluidic device using a 8


Page 9 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


syringe pump (Harvard Apparatus, Inc., MA, USA) at a flow rate of 60 µL/ min. To prevent cell adhesion, the channel walls were coated with 10% Pluronic F-68 for 30 min. For each separation experiment, 100 µL of cell suspension with 2 × 106 cells/mL were introduced into the channel. Intracellular Iron Concentration Measurements by ICP-MS. The concentration of the iron elements in the cell was measured by atomic mass spectrometry, using an inductively coupled plasma mass spectrometer (ICP-MS; PerkinElmer, Shelton, CT, USA). Prior to this analysis, concentrated nitric acid (HNO3, Optima grade; Fisher Scientific, Hampton, NH, USA) was added to the cell samples at a ratio of 5:1. The digested samples were then diluted 10-fold using distilled water.24 Prussian Blue Staining of Intracellular Iron-Oxide. Prussian blue staining was used to demonstrate the presence of iron oxide in the cells. Prior to staining, cells treated with iron oxide nanoparticles were washed three times with phosphate buffered saline (PBS) and fixed in 2% glutaraldehyde. The cells were then stained with a working solution containing equal amounts of 4% potassium ferrocyanide and 1.2 mM hydrochloric acid (Sigma-Aldrich) for 10 min at room temperature. After washing with distilled water, the samples were observed using an inverted microscope (Carl Zeiss, Oberkochen, Germany) with a 40× objective lens.

Detection of Reactive Oxygen Species (ROS). ROS production was quantitated using nitroblue tetrazolium (NBT) salt (Thermo Fisher Scientific, Seoul, Korea). The separated cells were washed twice with PBS and incubated with fresh medium containing 1 mg/mL NBT (Sigma-Aldrich) for 4 h at 37°C and 5% CO2. The media were then removed, and the cells washed once with PBS, after which 120 µL KOH was added to solubilize cell membranes. We then added 140 µL dimethyl sulfoxide (DMSO) to dissolve blue formazan with gentle shaking 9



Page 10 of 31

for 10 min at room temperature. The dissolved NBT solution was transferred to a 96-well plate and the absorbance was read on a microplate reader at 620 nm. The results were expressed as a percentage of a non-labeled control cells.25

Proliferation Assay. Cell proliferation was assessed by manual counting using an automated cell counter (LUNA-II; Logos Biosystems, Anyang, Korea). In preparation for this assay, the cells were collected from each outlet, kept in culture, and re-seeded every 24 h at 2 × 104 cells/mL. The average cell division time for four independent tests was measured as described previously.15

Statistical Analysis. All data acquired from experiments were represented as mean ± standard deviation. At least four independent tests were performed to reach the final results.

10


Page 11 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


RESULTS AND DISCUSSION Simulation Analysis. To evaluate whether the proposed unique structure was more effective than the structure commonly used, we investigated the trajectories of cells with several iron loads through simulation studies. To study the deflection trajectories of the magnetically susceptible cells, the distribution of the magnetic fields generated by the external permanent magnet must first be estimated. We assume that the origin is at the point of 1.5 mm from the magnet. The numerical analysis of the magnetic field gradient along the cross-sectional distance is shown in Figure 3A. As expected, the strength of the magnetic field gradient decreased exponentially with increasing distance from the magnet. Strong magnetic field gradients were observed at a crosssectional distance of less than 0.5 mm, resulting in a large magnetic force acting on the cells in this region. Conversely, the magnetic field gradients were weak at a cross-sectional distance of greater than 3 mm, which would result in a negligible magnetic force acting on the cells. Based on this simulation, the rectangular channel width was determined to be 3 mm. In the trapezoidal channel, the long and short widths were designed to be 3 mm and 0.3 mm (shorter than 0.5 mm), respectively. Using numerical magnetic field analysis, the trajectories of cells with seven different iron loads (2,6,10, 15, 25, 35, and 50 pg per cell) were calculated for a single rectangular channel (3000-mm width × 45-mm length), with an applied flow rate of 60 µL/min, corresponding to a magnetic exposure time of 5 s. Here, the origin of the axis is set at the entry point of the separation channel. The initial focusing position of cells was 100 µm. The magnetic susceptibility of iron oxide nanoparticles was measured by a SQUID magnetometer and the value was approximately 1.528. The theoretical trajectory of a cell in a rectangular channel based on its magnetic loading is shown in Figure 3B; a cell loaded with 50 pg of iron experienced a 11



Page 12 of 31

large degree of deflection, nearly reaching the upper sidewall of the channel. However, the degree of deflection of cells containing less than 10 pg of iron was so small that it was difficult to distinguish. Given that particle distribution is not considered in the path calculation and the assumption that the particles enter the channel from exactly the same position, it is even more difficult to distinguish between cells containing 2, 6, and 10 pg of iron. In the rectangular channel, the magnetic velocity of the cells at the initial entry point differed depending on the degree of magnetic labeling. From eq 4, the initial magnetic velocities of cells loaded with 2, 6 and 10 pg of iron were calculated to be 0.5, 6, and 15 µm/s, respectively, whereas that of cells containing 50 pg of iron was 150 µm/s. Because the initial velocity of cells loaded with 50 pg of iron was overwhelmingly high, it was impossible to simultaneously discriminate such cells with large differences in uptake capability using a single device. The proposed device, which consisted of a trapezoidal channel with five side outlet branches toward the magnet and a narrow rectangular channel with three outlet branches, was expected to be more effective in separating cells that have significant heterogeneity in intracellular iron content. A theoretical trajectory of a cell in the proposed device based on its magnetic loading from 2 to 50 pg is shown in Figure 3C. The focusing stream was aligned along the lower sidewall. The tapered structure allowed the magnetically labeled cells to be deflected by the magnetic force as well as the 5 ° tilted lower sidewall. The movement of ∆d along the x-axis results in a ∆d × tanθ (θ = 5°) shift in the y-axis direction. This shift in the y-direction moved cells to the region where magnetic fields were higher, resulting in a higher magnetic velocity. Particularly in the cells with a magnetic load less than 10 pg of iron, the deflection increased as the cells approached the magnet, thus promoting efficient separation of cells with low iron content. Indeed, relying on geometric movement, the calculated instantaneous magnetic velocity 12


Page 13 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


of the cells loaded with 2 pg of iron increased from 0.5 to 10 µm/s, such that the gap distance between the focusing path and the cell lateral displacement also increased. Cells with higher magnetic loading (e.g., 50, 35, 25, or 15 pg of iron), were sequentially deflected to the upper channel wall and cells with lower magnetic loading were collected later at a higher outlet number. The cells loaded with 2, 6, and 10 pg of iron entered the narrow channel at the position of 245 µm, 182 µm, and 19 µm, respectively, from the upper wall. Since the pressure-drive laminar flow represents a parabolic profile, the particles that flow mainly through the channel at the middle position would be faster than the particle near the wall. Consequently, the cells loaded with 10 pg of iron already reached the upper wall that was close to the magnet and would slowly flow into the corresponding outlet. The cell with 2 pg of iron was close to the lower wall experiencing a slower velocity, while the cells loaded with 6 pg entering the channel in the position near the center moved faster. As a result, even the cells loaded with 6 pg experience a larger magnetic force, they have a slightly shorter residual time in the microchannel compared to cells containing 2 pg of iron.

Cell Separation Based on Magnetic Loading. The uptake of magnetic iron oxide nanoparticles varies considerably within a given cell population. Such cell-to-cell variation is more severe in macrophages, whose main function is to engulf and digest cellular debris and foreign substances. This high heterogeneity in nanoparticle uptake may lead to misleading conclusions about the nanotoxicity of magnetic nanoparticles in which toxic effects appear to be dose-dependent. To obtain more appropriate data, cell populations with more narrowly defined magnetic properties are required. We attempted to achieve this through magnetophoresis using trapezoidal and rectangular channels. 13



Page 14 of 31

In the device, when no magnetic field was present, all cells were focused onto the lower sidewall by the hydrodynamic effect and therefore exited through outlet 8. By applying the magnetic field perpendicular with respect to the flow direction, the cells were deflected by laminar flow and successively distributed over a wide range of outlets, depending on their magnetic loading. The ratios of cells recovered from each outlet are shown in Figure 4A. From outlet 1, however, a very small number of cells were collected (0.1 ± 0.2%), too small to obtain statistically significant data. Therefore, we analyzed only cells from outlets 2–8 in subsequent experiments. The frequency of cells exiting from outlets 2–7 exhibited a normal distribution. The sum of the ratios of the cells recovered from the outlets 2, 7, and 8 was 18.5%, whereas most cells were retrieved through the four remaining outlets (outlets 3–6). The distribution ratios of outlets 3–6 were approximately 19.8%, 23.3%, 20.2%, and 18.2%, respectively. The ratio of cells from outlet 8 was slightly greater than that from outlet 7. The most probable reason for such a small difference is that the subpopulation from outlet 8 may include both cells loaded with tiny amounts of iron and non-magnetically labeled cells. The iron load of cells exiting at each outlet was evaluated qualitatively and quantitatively. For qualitative assessment, we used Prussian blue staining on the cell subpopulations from each outlet; intra-cytoplasmic iron inclusions appeared as dense blue-stained agglomerates (Figure 4B). The number of blue-stained agglomerates increased and the color became darker as the nanoparticle dose increased. The number of blue-stained agglomerates in the separated cell subpopulations gradually decreased from outlet 2 to outlet 8, because the cellular magnetic loadings of the cells decreased accordingly. These findings were consistent with the results of the quantitative measurements performed using ICP-MS (Figure 4A). We recovered an average of 46.8 pg of iron per cell through outlet 2, 14


Page 15 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


and averages of 8.3, 3.7, and 1.5 pg of iron per cell in cells exiting through outlets 6, 7, and 8, respectively. The magnetic loading of outlets 3, 4, and 5 again displayed a declining trend, with average values of 35.4, 27.6, and 16.4 pg, respectively. The experimentally measured intracellular iron content at each outlet agreed well with the theoretically predicted simulation shown in Figure 3C. Considering the results of the cell separation based on the magnetic loading, we can conclude that the proposed device combining a trapezoidal channel with a narrow long rectangular channel can efficiently separate cells with tiny amounts of iron (less than 10 pg of iron per cell) as well as heavily labeled cells, in a single device. Before conducting post analysis, we evaluated the cell viability at each outlet using a live/dead cell assay (Figure 4C). At every outlet, at least 98% of collected cells were found alive, and no significant differences were found among high to low iron-loaded cells, indicating that the nanoparticles were non-toxic over this loading range.

ROS Detection in Separated Cell Subpopulations. The ROS levels produced by the bulk cell population and the separated cell subpopulations were compared using an NBT assay. ROS production by iron oxide nanoparticle is well known to be involved in undesirable interferences in cellular function. Fe3O4, a mixture of FeO and Fe2O3, is unstable and can easily undergo oxidation to yield γ-Fe2O3 and Fe2+. The free Fe2+ ions can react with hydrogen peroxide and oxygen to generate highly reactive free radicals that could damage DNAs, proteins, and lipids in vivo.26 The production of free radicals implies the possible degradation of nanoparticles and the release of ferric ions from the iron oxide nanoparticles into the acidic endolysosomal compartments. Figure 5 shows that the ROS level of the bulk cell population was approximately 1.5-fold higher than the control population, which was not labeled with magnetic nanoparticles. 15



Page 16 of 31

In contrast, the ROS levels of the subpopulations from outlets 2–4 showed a dose-dependent decrease. The ROS levels of outlets 2, 3, and 4 were approximately 2.3-, 1.5-, and 1.2 times higher than the control, respectively. The cells from outlet 2 produced the highest amount of ROS, much higher than in the bulk population, whereas the cells from outlet 4 produced slightly less ROS than the bulk population. Cells collected from outlets 5–8 produced similar levels of ROS to those of the control group, suggesting that the subpopulations with iron contents lower than 16.4 pg (outlet 5) showed little toxicity related to ROS, which indicated minimal intracellular degradation. Cellular iron loading of 16.4 pg may be the threshold for internalized nanoparticles to exert effects on cell physiology. Taken together, the average ROS level of the subpopulations was similar to that determined by bulk measurement, indicating that the subpopulation results provide a comprehensive interpretation of bulk data. Furthermore, approximately 46.6% of the cells (cells from outlets 2–4) in the bulk population showed higher ROS levels than the control, whereas 3.5% of the cells (outlet 2) caused a significant ROS production. Conversely, approximately 53.3% of the cells presented negligible effects on ROS production, similar to the control group (outlets 5–8). Prior to evaluating the clinical potential of magnetic nanoparticles, in such applications as cell transplantation studies, these effects should be carefully investigated.

Effects of Intracellular Iron Contents on Cell Proliferation. The actin cytoskeleton is involved in many cellular processes, such as apoptosis and proliferation, through integrinmediated signaling, as integrins are mechanically linked to the actin cytoskeleton in so-called focal adhesion complexes.14 To determine the semi-long-term toxic effects of iron oxide nanoparticles, we assessed and verified the cell proliferation rate by cell counting 1, 2 and 3 days 16


Page 17 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


after labeling (Figure 6A). We found that the bulk group of cells reduced cellular proliferation at 1 and 2 days after nanoparticle internalization (Figure 6B). These results were consistent with previous findings that high intracellular amounts of iron oxide nanoparticles could transiently reduce the rate of the cell cycle.14 One of the greatest causes of diminished proliferation could be the interference of internalized nanoparticles with cytoskeleton architecture. These effects have been shown to be only transient and to recover near control levels when intracellular nanoparticles are reduced by cell division. These results support the dose-dependent effects of nanoparticles on cell proliferation, because the cells appear to recover when particles are diluted by a factor of 2n (n = the number of cell doublings). The proliferation of separated subpopulation cells was affected in a similar manner, with a reduction in proliferation of subpopulations at 1 and 2 days after labeling. In particular, the reduction in cell proliferation was most pronounced 1 day after labeling in cells collected from outlets 2 and 3 (a sum of 23.2%). The significant increase in cell doubling times observed in cells from outlets 2 and 3 is likely due to their high intracellular nanoparticle loadings, because the subpopulations of cells with the highest uptake efficiency exerted significant effects on cell proliferation. Cells from outlets 5 and 6 showed a slightly lower reduction effect than bulk group cells. Additionally, the transient diminished proliferation rate was not detectable in cells from outlets 7 and 8. These data suggest that the amounts of nanoparticles obtained from cells from outlets 7 and 8 may provide a threshold for internalized nanoparticles to exert effects on the cell cycle. In agreement with the bulk group data, after 3 days post labeling, the proliferative capacity of the separated subpopulations recovered to almost normal levels, which corresponds to approximately six cell doublings. Consequently, our methods to assess cell proliferation at the subpopulation level are

17



Page 18 of 31

more reasonable and provide more detailed information about toxic effects than bulk cell-based assays.

CONCLUSIONS The assessment of the potential nanotoxicity of iron oxide nanoparticles on cells requires a more reliable approach, wherein toxic effects and cell iron loading should be analyzed together, due to heterogeneity in nanoparticle uptake during labeling via endocytosis. In this study, we proposed and evaluated a simple and efficient method for magnetic cell sorting based on iron oxide loading using a macrophage-like cell line as a proof of concept. Unlike current techniques used to digest cells to assess the constituent iron atoms or ions dissociated from iron oxide nanoparticles, such as ICP-ACE and UV/VIS spectrometry, the proposed system is capable of sequentially quantifying and sorting cells based on the uptake of nanoparticles in their solid form, and thus post-nanotoxicity assays are made possible as a function of the endocytotic capacities of the cells. Our microfluidic device for magnetophoresis was able to sort cells accurately. By combining trapezoidal and rectangular structures, we were able to separate cells highly loaded with magnetic nanoparticles as well as those poorly loaded with magnetic nanoparticles, in a single device. In this manner, we demonstrated that the assessment of the potential nanotoxicity of subpopulations according to their iron content is more reliable and provides more detailed information than the assessment of the bulk population. Magnetophoresis using a combination of trapezoidal and rectangular microchannels could be useful in future studies of nanotoxicity induced by magnetic nanoparticles.

ASSOCIATED CONTENT 18


Page 19 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


SUPPORTING INFORMATION The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.#######. Cells were stained with Cell Tracker Red™ CMTPX (Invitrogen). Fluorescently stained cells were visualized by a CCD camera. Since the field of view of the microscope that covered only parts of the channel, we separately showed the trajectories of the cells. As the distinguishable trajectories are focused on the trapezoidal region, the trajectories of magnetically labeled cells are clearly shown at the trapezoidal channel (20× speed). (Movie S1.MP4) (ZIP) AUTHOR INFORMATION Corresponding Author * E-mail: [email protected]. Phone: +82-42-350-4315. Fax: +82-42-350-4310. Notes The authors declare no competing financial interest. ACKNOWLEDGMENTS This research was supported by the National Research Foundation of Korea (NRF) (Grant No. NRF-2016R1A2B3015986, NRF-2015M3A9B3028685) funded by the Ministry of Science and ICT. The authors also acknowledge a KAIST Systems Healthcare Program.

19



Page 20 of 31

REFERENCES (1) Colombo, M.; Carregal-Romero, S.; Casula, M. F.; Gutierrez, L.; Morales, M. P.; Bohm, I. B.; Heverhagen, J.T.; Prosperi, D.; Parak, W. J., Chem. Soc. Rev. 2012, 41, 4306–4334. (2) Gupta, A. K.; Gupta, M., Biomaterials 2005, 26, 3995–4021. (3) Wang, Y. X., Quant. Imaging Med. Surg. 2011, 1, 35–40. (4) Deatsch, A. E.; Evans, B. A., J. Magn. Magn. Mater. 2014, 354, 163–172. (5) Canton, I.; Battaglia, G., Chem. Soc. Rev. 2012, 41, 2718–2739. (6) Cromer Berman, S. M.; Kshitiz; Wang, C. J.; Orukari, I.; Levchenko, A.; Bulte, J. W.; Walczak, P., Magn. Reson. Med. 2013, 69, 255–262. (7) Heyn, C.; Bowen, C. V.; Rutt, B. K.; Foster, P. J., Magn. Reson. Med. 2005, 53, 312-320. (8) Wang, D.; Bodovitz, S., Trends Biotechnol. 2010, 28, 281–290. (9) Jing, Y.; Mal, N.; Williams, P. S.; Mayorga, M.; Penn, M. S.; Chalmers, J. J.; Zborowski, M., FASEB J. 2008, 22, 4239–4247. (10) Robert, D.; Pamme, N.; Conjeaud, H.; Gazeau, F.; Iles, A.; Wilhelm, C., Lab Chip 2011, 11, 1902–1910. (11) Wang, H.; Wu, L.; Reinhard, B. M., ACS Nano 2012, 6, 7122–7132. (12) Liu, G.; Gao, J.; Ai, H.; Chen, X., Small 2013, 9, 1533–1545.

20


Page 21 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(13) Sharifi, S.; Behzadi, S.; Laurent, S.; Forrest, M. L.; Stroeve, P.; Mahmoudi, M., Chem. Soc. Rev. 2012, 41, 2323–2343. (14) Soenen, S. J.; Nuytten, N.; De Meyer, S. F.; De Smedt, S. C.; De Cuyper, M., Small 2010, 6, 832–842. (15) Soenen, S. J.; Himmelreich, U.; Nuytten, N.; De Cuyper, M., Biomaterials 2011, 32, 195– 205. (16) Lehmann, A. D.; Parak, W. J.; Zhang, F.; Ali, Z.; Rocker, C.; Nienhaus, G. U.; Gehr, P.; Rothen-Rutishauser, B., Small 2010, 6, 753–762. (17) Soenen, S. J.; De Cuyper, M., Nanomedicine 2010, 5, 1261–1275. (18) Soenen, S. J.; De Cuyper, M., Contrast Media Mol. Imaging 2011, 6, 153–164. (19) Soenen, S. J.; Rivera-Gil, P.; Montenegro, J.; Parak, W. J.; De Smedt, S. C.; Braeckmans, K., Nano Today 2011, 6, 446–465 (20) Marquis, B. J.; Love, S. A.; Braun, K. L.; Haynes, C. L., Analyst 2009, 134, 425–439. (21) Rad, A.; Janic, B.; Iskander, A. S. M.; Soltanian-Zadeh, H.; Arbab, A., BioTechniques 2007, 43, 627–636. (22) Chalmers, J. J.; Haam, S.; Zhao, Y.; McCloskey, K.; Moore, L.; Zborowski, M.; Williams, P. S., Biotechnol. Bioeng. 1999, 64, 519–526. (23) Lunov, O.; Syrovets, T.; Loos, C.; Beil, J.; Delacher, M.; Tron, K.; Nienhaus, G. U.; Musyanovych, A.; Mailander, V.; Landfester, K.; Simmet, T., ACS Nano 2011, 5, 1657–1669. 21



Page 22 of 31

(24) Jin, X.; Chalmers, J. J.; Zborowski, M., Anal. Chem. 2012, 84, 4520–4526. (25) Choi, H. S.; Kim, J. W.; Cha, Y.-N.; Kim, C., J. Immunoassay Immunochem. 2006, 27, 31–44. (26) Watanabe, M.; Yoneda, M.; Morohashi, A.; Hori, Y.; Okamoto, D.; Sato, A.; Kurioka, D.; Nittami, T.; Hirokawa, Y.; Shiraishi, T.; Kawai, K.; Kasai, H.; Totsuka, Y., Int. J. Mol. Sci. 2013, 14, 15546–15560. (27) Wilhelm, C.; Gazeau, F.; Bacri, J.-C., Eur. Biophys. J. 2002, 31, 118–125. (28) Hahn, Y. K..; Park, J-K., Lab Chip 2011, 11, 2045–2048.

22


Page 23 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


FIGURE CAPTIONS

Figure 1. Schematic illustration of the magnetophoresis device with trapezoidal and straight regions. In the trapezoidal region, cells heavily labeled with iron oxide nanoparticles are first deflected to the upper sidewall close to the magnet, whereas cells with tiny amounts of iron oxide nanoparticles receive only a negligibly small magnetic force. Due to the tapered structure of the device, cells with low iron oxide loading are subjected to stronger magnetic fields, resulting in an improvement in the separation resolution.

Figure 2. (A) Schematic diagram of the proposed microfluidic device. The microfluidic chip design consists of two separation regions. The trapezoidal region included a tapered channel with five upward side branches (outlets 1–5) and the straight region featured a narrow rectangular channel with three branches (outlets 6–8). (B) A photograph of the fabricated microfluidic device. The enlarged images show the structures of sample inlet and outlets.

Figure 3. Simulation results of magnetic field gradient distribution and lateral trajectories of magnetically susceptible cells in a microchannel with a flow rate of 60 µL/min. (A) Numerical analysis of the magnetic field gradients in a microchannel at a cross sectional distance 1.5 mm from the permanent magnet. Theoretical estimation of the trajectories of magnetically susceptible cells in (B) a single rectangular chamber and (C) the proposed design combining a trapezoidal channel and a narrow rectangular channel. The origin of the axis was set at the entry point. Magnetic movement paths for cells loaded with 2, 6, 10, 15, 25, 35, and 50 pg of iron oxide nanoparticles were calculated. 23



Page 24 of 31

Figure 4. Separation of magnetically labeled cells based on their iron oxide loading. (A) Ratios of cells recovered from each outlet measured and the mean iron load of the cell subpopulations from each outlet were quantitatively determined using inductively coupled-plasma mass spectrometry (ICP–MS). (B) Prussian blue staining of cells separated from each outlet for qualitative assessment. (C) Viability of sorted cells was evaluated by a live/dead assay. Live cells stained with green dye and dead cells showed red fluorescence. Due to insufficient amounts of sorted cells, we obtained cell images for outlet 2 in panels B and C after the cells were concentrated for optical inspection.

Figure 5. Relative levels of reactive oxygen species (ROS) production by sorted cell subpopulations and by the bulk group.

Figure 6. Cellular proliferation of the sorted cell subpopulation. Cell doubling times were compared to (A) the sorted cells at each outlet and (B) the bulk group and the control. The control group represents untreated cells, whereas the bulk group represents the magnetically labeled cells, which were not treated with the microfluidic device. The cells were counted 1, 2, and 3 days after the labeling and separating procedures.

24


Page 25 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1. Schematic illustration of the magnetophoresis device with trapezoidal and straight regions. In the trapezoidal region, cells heavily labeled with iron oxide nanoparticles are first deflected to the upper sidewall close to the magnet, whereas cells with tiny amounts of iron oxide nanoparticles receive only a negligibly small magnetic force. Due to the tapered structure of the device, cells with low iron oxide loading are subjected to stronger magnetic fields, resulting in an improvement in the separation resolution. 160x67mm (300 x 300 DPI)



Figure 2. (A) Schematic diagram of the proposed microfluidic device. The microfluidic chip design consists of two separation regions. The trapezoidal region included a tapered channel with five upward side branches (outlets 1–5) and the straight region featured a narrow rectangular channel with three branches (outlets 6– 8). (B) A photograph of the fabricated microfluidic device. The enlarged images show the structures of sample inlet and outlets. 160x166mm (300 x 300 DPI)


Page 26 of 31

Page 27 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. Simulation results of magnetic field gradient distribution and lateral trajectories of magnetically susceptible cells in a microchannel with a flow rate of 60 µL/min. (A) Numerical analysis of the magnetic field gradients in a microchannel at a cross sectional distance 1.5 mm from the permanent magnet. Theoretical estimation of the trajectories of magnetically susceptible cells in (B) a single rectangular chamber and (C) the proposed design combining a trapezoidal channel and a narrow rectangular channel. The origin of the axis was set at the entry point. Magnetic movement paths for cells loaded with 2, 6, 10, 15, 25, 35, and 50 pg of iron oxide nanoparticles were calculated. 160x194mm (300 x 300 DPI)



Figure 4. Separation of magnetically labeled cells based on their iron oxide loading. (A) Ratios of cells recovered from each outlet measured and the mean iron load of the cell subpopulations from each outlet were quantitatively determined using inductively coupled-plasma mass spectrometry (ICP–MS). (B) Prussian blue staining of cells separated from each outlet for qualitative assessment. (C) Viability of sorted cells was evaluated by a live/dead assay. Live cells stained with green dye and dead cells showed red fluorescence. Due to insufficient amounts of sorted cells, we obtained cell images for outlet 2 in panels B and C after the cells were concentrated for optical inspection. 160x132mm (300 x 300 DPI)


Page 28 of 31

Page 29 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. Relative levels of reactive oxygen species (ROS) production by sorted cell subpopulations and by the bulk group. 160x54mm (300 x 300 DPI)



Figure 6. Cellular proliferation of the sorted cell subpopulation. Cell doubling times were compared to (A) the sorted cells at each outlet and (B) the bulk group and the control. The control group represents untreated cells, whereas the bulk group represents the magnetically labeled cells, which were not treated with the microfluidic device. The cells were counted 1, 2, and 3 days after the labeling and separating procedures. 160x66mm (300 x 300 DPI)


Page 30 of 31

Page 31 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


For TOC only 84x27mm (300 x 300 DPI)


Toxicity Assessment of Iron Oxide Nanoparticles Based on Cellular

Recommend Documents