SPION^Citrate were synthesized according to an adjusted protocol of Elbialy et al. (2015), Muhlberger et al. (2019). In summary, iron (III) chloride and iron (II) chloride in a molar ratio of 2:1 (Fe³⁺: Fe²⁺) were precipitated to iron oxide nanoparticles using a 25% ammonia solution under rapid stirring. After 10 min of reaction, a sodium citrate solution was added, the temperature of the liquid was raised to 90°C within 15 min and was held at 90°C for another 15 min to stabilize the resulting SPION^Citrate. After cooling the particles to room temperature, they were magnetically washed from excess citrate five times using acetone. SPION^Citrate were dried from acetone and redispersed in deionized water.

SPION^Gold for subsequent gold-coating were synthesized accordingly with a difference in the washing procedure. The acetone washing steps were replaced with a dialysis treatment. The particles were given into 10 kDa dialysis tubes and dialyzed five times against 4.5 L of deionized water. Finally, the particles were ultrasonicated to fine-tune the hydrodynamic size to ∼80 nm for the following gold coating.

The SPION^Gold synthesis was performed following the route of Stein et al. (2020) with slight adaptations. In short, SPION^Citrate with a hydrodynamic size of 80 nm, instead of the previously reported 110 nm SPION^Citrate from Stein et al., were brought to a boil. Then, a HAuCl₄ solution was quickly added to the SPION dispersion to precipitate gold onto the particles. After 15 min of reaction, a citrate solution was injected to stabilize the generated SPION^Gold and the particles were cooled to room temperature.

SPION^Citrate were sterilized by filtration through syringe filters with 0.2 µm pore size (Sartorius, Goettingen, Germany), while SPION^Gold were autoclaved for 30 min at 121°C and 2 bar of pressure. Both particle systems were stored under sterile conditions at 8°C until characterization and further use.

Subsequently, the SPIONs were analyzed for their size, magnetic susceptibility, and zeta potential as described previously (15). For the determination of the iron content, the suspension was diluted 1:25 in deionized H₂O (produced in-house via the Merck Milli-Q^® Direct water purification system, Darmstadt, Germany) and dissolved in 65% nitric acid (Carl Roth, Karlsruhe, Germany), for analysis via atomic emission spectroscopy (AES), using the Agilent 4200 MP-AES (Agilent Technologies, Santa Clara, CA, United States) with an external standard iron solution of 1,000 mg/L (Bernd Kraft, Duisburg, Germany) as a basis for a calibration curve. Measurements were performed in triplicates at a wavelength of 371.993 nm and averaged.

The Z-Average, further called hydrodynamic size, of the particle systems was measured via dynamic light scattering (DLS) using a Zetasizer Nano (Malvern instruments, Worcestershire, United Kingdom). SPION^Citrate were diluted to an iron concentration of 50 μg/mL with deionized water, while SPION^Gold were diluted to 25 μg/mL for the analysis at 25°C.

The investigation of the zeta potential of SPION^Citrate and SPION^Gold was conducted using the Zetasizer Nano (Malvern instruments, Worcestershire, United Kingdom). For the measurement, SPION^Citrate and SPION^Gold dispersions with iron concentrations of 50 μg/mL and 25 μg/mL, respectively, were adjusted to a pH of 7.3 using minimal amounts of 10 mM HCl or 10 mM NaOH solutions.

The magnetic volume susceptibility of both SPION systems was measured using a magnetic susceptibility meter (MS2G, Bartington Instruments, Witney, UK). The iron concentrations of all samples were diluted to 1 mg/mL using deionized water in order to gain comparable values.

All characterizations were performed in triplicates and results are presented as the calculated mean value.

THP-1 monocytic cells were purchased from DSMZ (Braunschweig, Germany) and cultured in Gibco™ RPMI medium 1640 supplemented with 2 mmol/L glutamine, 100 U/mL penicillin-streptomycin, both from Thermo Fisher Scientific (Waltham, MA, United States), and 10% fetal calf serum (FCS, Biochrom, Berlin, Germany) under cell culture conditions.

To investigate their colloidal stability, SPION^Citrate and SPION^Gold (0, 20, 40 and 80 μg/mL) were incubated in 200 µL THP-1 medium overnight under cell culture conditions. The next day, the nanoparticles were either resuspended or not. Then, 100 µL supernatant of the not-resuspended SPIONs or the resuspended SPIONs were transferred to a separate 96-well-plate. Subsequently, the absorbance was measured in the SpectraMax iD3 (Molecular Devices, San José, CA, United States) from 230 nm to 750 nm in 10 nm steps. Additionally, Eppendorf tubes with SPION^Citrate or SPION^Gold (0, 20, 40 and 80 μg/mL) in THP-1 medium were incubated overnight and the nanoparticle sedimentation was photographed on the next day.

To evaluate the particle uptake, 2 × 10⁵ THP-1 cells per mL medium were incubated in a 24-well plate with SPION^Citrate or SPION^Gold at an iron concentration of 20, 40 or 80 μg/mL, using the same volume of H₂O or 2% DMSO as controls. After 24 h, cells were stained with 1 μL/mL Annexin V-FITC (AxV-FITC, ImmunoTools, Friesoythe, Germany) and 66.6 ng/mL propidium iodide (PI, Sigma-Aldrich, Taufkirchen, Germany) in Ringer’s solution (Fresenius Kabi, Bad Homburg, Germany). The side scatter, which describes the granularity of the cell and correlates with particle uptake, was measured with a Gallios flow cytometer and analyzed with Kaluza (1.3., 2.1) Analysis Software (both from Beckman Coulter, Brea, CA, United States).

Furthermore, the cellular iron content was quantified by AES. Therefore, 2 × 10⁵ THP-1 cells per mL medium were seeded in triplicates in a 24-well plate. The cells were either incubated with SPION^Citrate or SPION^Gold at an iron concentration of 20, 40 or 80 μg/mL, using the same volume of H₂O or 2% DMSO as controls. After 24 h incubation, cells were washed with phosphate-buffered saline (PBS) and centrifuged at 300 g for 5 min to remove residual particles. Then, the cells were resuspended in PBS and counted by a MUSE Cell Analyzer. Cell pellets were formed by centrifugation at 10,000 g for 5 min. The pellets were dried at 300 rpm at 95°C for 30 min and lysed with 65%-nitric acid at 300 rpm at 95°C for 15 min. The samples were diluted with water and the iron content was measured with Agilent 4200 MP-AES.

To visualize particle uptake and adherence on the cell surface transmission electron microscopy (TEM) was used. Therefore, 7 × 10⁶ THP-1 cells in 10 mL medium were incubated with 20 or 40 μg/mL of SPION^Citrate or SPION^Gold using H₂O as control. After 24 h cells were washed and then fixed in 2.5% glutaraldehyde (Carl Roth, Karlsruhe, Germany) in 0.1 M PO₄ buffer (Carl Roth) for 4 h. Afterwards cells were washed 3 times with 0.1 M PO₄ buffer and kept at 4°C overnight. The next day, cells were stained with 1% osmium tetra oxide (OsO₄) in 3% potassium ferricyanide (K3(Fe(Cn)6), (Science Service, Munich, Germany) for 2 h at room temperature and then washed with 0.1 M PO₄ buffer overnight at 4°C. Subsequently, cells were transferred into a matrix of 2% agarose (low melting point, Science Service) in 0.1 M PO₄ buffer overnight using an Eppendorf reaction tube. After being dehydrated in agarose through an ethanol step, the cells were finally transferred to Epon using an acetone-epon mixture and left to polymerize for 48 h at 60°C. With the Leica Ultracut UCT (Wetzlar, Germany) ultra-thin sections with a thickness of around 50 nm of the samples were prepared and put on copper grids (Science Service). The received sections were contrasted with lead citrate for 10 min followed by uranyl acetate (Serva, Heidelberg, Germany) for additional 10 min. For the images the TEM 906 LEO (Zeiss, Oberkochen, Germany) with a CCD-camera residual light amplifier from TRS Tröndle (Moorenweis, Germany) was used and further the ImageSP SysPROG software. An acceleration voltage of 80 kV and a magnification of 12,930-fold were used for image taking.

To determine the generation of oxidative stress in THP-1 cells by nanoparticles, THP-1 cells were stained with 20 µM 2′,7′-Dichlorodihydrofluoresceindiacetat (DCFH, Sigma Aldrich) and incubated under cell culture conditions for 30 min. Cells were washed, seeded in a concentration of 1.8 × 10⁵ THP-1 cells per mL medium in a 96-well plate, and incubated with either SPION^Citrate or SPION^Gold at a concentration of 20, 40 or 80 μg/mL, with H₂O as negative control and 2 mM H₂O₂ or 2% DMSO as positive controls. After 3 h, 6 h and 24 h, 50 µL of cells were stained with 66.6 ng/mL PI, 1 µl/ml Annexin V-APC (AxV, ImmunoTools, Friesoythe, Germany) and 30 µM monobromobimane (MBB, Sigma Aldrich) in Ringer’s solution at 4°C under light protection and analyzed in flow cytometry.

THP-1 cells were incubated with 0, 20, 40 or 80 μg/mL SPION^Citrate or SPION^Gold for 24 h and subsequently stained with 20 nM Hoechst 33342 (Hoe, Thermo Fisher Scientific, Waltham, MA, United States). µ-Slides I Luer with a channel height of 0.4 mm (Ibidi, Gräflingen, Germany) were used for microscopy. The slides were coated with FCS to prevent THP-1 cells from adhering to the channel surface. A NdFeB disc-shaped magnet (S-15-08-N, Webcraft GmbH, Gottmadingen, Germany) was used for magnetic attraction with 15 mm diameter, 8 mm height, N42 magnetization, Ni-Cu-Ni coating, and axial direction of magnetization.

Slides without magnet were used as controls for each SPION-concentration. Immediately after the magnet was placed onto the slide, cells were pipetted in the slide channel at the opposite side. After 10 min, the slide was analyzed with Zeiss Axio Observer Z1 fluorescence microscope (Carl Zeiss AG, Oberkochen, Germany). Pictures were taken below the center of the magnet, on the edge of the magnet and then in 5 mm increments up to 3 cm away from the magnet and analyzed with ImageJ software (version 1.52a).

Further, we confirm cellular viability in the presence of an external magnetic field (EMF). To do so, THP-1 cells were seeded in a concentration of 2 × 10⁴ cells per well in 96 well plates. The cells were incubated with 0, 20, 40 or 80 μg/mL SPION^Citrate or SPION^Gold for 24 h, washed and subsequently enriched by placing a NdFeB disc-shaped magnet underneath each well (S-15-08-N, Webcraft GmbH, Gottmadingen, Germany; 15 mm diameter, 8 mm height, N42 magnetization, Ni-Cu-Ni coating, and axial direction of magnetization). Cells were exposed to the EMF for 24 h and stained with 1 μL/mL Annexin V-APC (AxV-APC, ImmunoTools, Friesoythe, Germany), 66.6 ng/mL PI in Ringer’s solution at 4°C under light protection and analyzed in flow cytometry. Viability was measured with a Gallios flow cytometer and analyzed with Kaluza (1.3., 2.1) Analysis Software (both from Beckman Coulter, Brea, CA, United States).

We validated our experiments with a computational model. For the disc-shaped magnet used in the experiments, analytical expressions allow efficient computation of the magnetic field and the magnetic force on the SPION-loaded cells. We computed the magnetic field using the analytical expression presented by Derby and Olbert (2010). For the magnetic force on the SPION-loaded cells, we used the analytical expression derived by Wirthl et al. (2024), including a linear magnetization model with saturation magnetizations as measured by Stein et al. (2020).

Additionally, we computed the magnetic accumulation of the SPION-loaded cells. The force acting on a cell is the sum of the magnetic force F m a g and the drag force F d r a g . Based on Stokes’ law, the drag force is the resistive force exerted by the fluid on the cell, opposing its motion. Hence, the movement of the cell based on Newton’s second law of motion is given by F m a g + F d r a g = m a = m d v d t with m being the mass of the cell, v its velocity and a its acceleration. We used the backward Euler method to solve this ordinary differential equation for the cell velocity (Burden and Faires, 2015), which subsequently allowed us to compute the trajectory of the SPION-loaded cells over time.

Data were processed in Microsoft Excel (Microsoft, Redmond, WA, United States). Statistical analysis and graph creation were performed with GraphPad PRISM 9.0.2 from Graph-Pad Software, Inc. (San Diego, CA, United States). Data are shown as the mean ± standard deviation (SD) with at least three replicates, unless otherwise specified. Statistical p-values ≤ 0.05 were considered as statistically significant.

Hydrodynamic size, zeta potential at pH 7.3 and magnetic volume susceptibility of three (SPION^Citrate) or two (SPION^Gold) independently synthesized batches of SPIONs were investigated (Table 1). SPION^Citrate showed a hydrodynamic diameter of 52 nm ± 2 nm with a polydispersity index (PDI) of 0.15 ± 0.01 while SPION^Gold were 120 nm ± 0 nm with a PDI of 0.17 ± 0.01, respectively. For both SPION types, the size distribution was quite narrow. With −48.8 mV ± 6.2 mV and −41.6 mV ± 8.9 mV, SPION^Citrate and SPION^Gold showed comparable zeta potentials and a strong electrostatic stability. The magnetic susceptibility of SPION^Citrate ((4.18 ± 0.04) ⋅ 10⁻³) was slightly lower than that of SPION^Gold ((4.95 ± 1.59) ⋅ 10⁻³). The increased susceptibility of SPION^Gold is caused by the higher susceptibility of SPION^Citrate batches used as a basis for the gold-coating ((5.34 ± 1.60) ⋅ 10⁻³). The gold coating therefore reduced the susceptibility of SPION^Gold to 92.7% ± 2.1% of the initial value of primary SPIONs.

The colloidal stability of nanoparticles influences their cellular uptake and toxicity. To determine the colloidal stability of SPION^Citrate and SPION^Gold in THP-1 cell culture medium, the nanoparticles were incubated overnight to eventually let them agglomerate and sediment. Especially the SPION^Gold particles showed a macroscopically dose-dependent tendency for sedimentation. In contrast, sedimentation of SPION^Citrate particles was less pronounced (Figures 1A, D).

The measured spectrum (230–750 nm) of the nanoparticle dilutions showed an optimal wavelength to detect the SPIONs at 350 nm (Figures 1B, E). To analyze the colloidal stability of SPIONs, the samples were either resuspended or left not-resuspended, supernatants were taken and the optical densities (ODs) were compared. When nanoparticles are colloidally stable, the absorption of supernatants from not-resuspended and resuspended samples should be the same. As expected, SPION^Citrate and SPION^Gold displayed a dose-dependent increase in the absorption. SPION^Citrate showed only a slight difference in the OD values of the supernatants of not-resuspended and resuspended samples. SPION^Gold, on the other hand, showed a significant difference between the absorption values already at 20 μg/mL SPIONs (Figures 1C, F), indicating nanoparticle sedimentation. Thus, in direct comparison, SPION^Citrate showed a better colloidal stability than SPION^Gold under the tested conditions.

In order to control SPION-loaded THP-1 cells within a magnetic field, the cells require a sufficiently high loading with nanoparticles. Thereby their magnetic controllability is independent of whether the cells take up the particles during loading or whether the particles attach to the cell surface. Three different methods were used to assess the accumulation of SPION^Citrate and SPION^Gold in and/or on the THP-1 cells.

First, the side scatter properties of the cells were assessed by flow cytometry. Cells interacting with nanoparticles show an increase in granularity. It should be noted that cell morphology is also influenced by cell death processes, thus gating on viable cells is necessary to exclude cells with morphological alterations caused by dying processes such as blebbing or membrane rupture. Therefore, we analyzed only cells with intact plasma membrane (AxV-FITC-/PI-) to analyze cellular side scatter alterations caused by the SPIONs. When cells were incubated with SPIONs for 24 h, THP-1 cells showed a significant increase in cell granularity already at a concentration of 20 μg/mL, which increased dose-dependently (Figures 2A, C).

To quantify the iron content in the cells, THP-1 cells incubated with SPIONs for 24 h were washed to remove free nanoparticles. Subsequently, the cell number was determined and the iron content of the cell pellets was measured in AES. Again, significant amounts of cellular iron were already detected when cells were incubated with 20 μg/mL SPIONs. When incubated with increasing amounts of nanoparticles, the cellular iron content increased in a dose dependent manner, which was in line with the SSC measurements. When comparing the iron amounts per cell, a large increase was detected from 40 μg/mL to 80 μg/mL for both particle types, which is due to the fact, that the particles agglomerated and strongly sedimented in concentrations of 80 μg/mL, so that at the bottom of the wells, where the cells are growing, more particles arrived than nominally administered (Figures 2B, D).

For visualization of the nanoparticle localization TEM was performed. Internalization into intracellular vesicles was visible for both SPION types. When cells were incubated with 40 μg/mL SPION^Citrate, single particles were detected in intracellular vesicles, while with 20 μg/mL SPION^Gold no individual particles could be identified which was probably due to increased agglomeration of this type of nanoparticle (Figure 2E, see red arrows). Attachment of particles to the plasma membranes was detected as well.

To assess the generation of reactive oxygen species (ROS) by Fenton reactions leading eventually to cell death, THP-1 cells were incubated with 20, 40 or 80 μg/mL SPIONs and analyzed after 3, 6 and 24 h after particle incubation. H₂O served as negative control, H₂O₂ and DMSO served as positive controls for ROS or general toxicity, respectively. All values were standardized to the H₂O incubated cells.

DCFH was used as an indicator for general ROS. Non-polar DCFH-DA is taken up by the cells and converted by esterases into polar DCFH. Subsequently, DCFH is oxidized by ROS to fluorescent DCF (Rastogi et al., 2010). MBB served as indicator of the redox status of glutathione (GSH). As an antioxidant, GSH serves as an electron donor and is oxidized to glutathione disulfide (GSSG) in the presence of ROS. MBB reacts with thiols in non-oxidized GSH and shows strong fluorescence. As soon as GSSG occurs, the fluorescence is decreased (Cossarizza et al., 2009). PI discriminates between cells with intact (PI-) and disturbed plasma membranes (PI+).

When incubating the cells with H₂O₂ as positive control, ROS is generated which is indicated by loss of MBB fluorescence and increase in DCF fluorescence. After 3 h, there was hardly oxidative stress detectable in the SPION treated samples, no matter if cells were stained with MBB or DCFH. After 6 h, in the SPION^Gold treated cells general ROS was increased as reflected by DCF fluorescence. After 24 h, this effect has been further increased in SPION^Gold treated cells, while no DCF fluorescence was measurable in SPION^Citrate treated ones. Interestingly, MBB fluorescence was dose-dependently elevated in SPION^Gold treated cells, which indicates an attempt by the cells to upregulate the antioxidant systems in order to cope with the ROS (Figures 3A, B). When analyzing the toxicity of SPION^Gold and SPION^Citrate, we did not find a decrease in viability at any of the three measurement time points, while H₂O₂ and DMSO induced cell death in a time dependent manner (Figure 3C). Additionally, we investigated the effect of an EMF on cell viability alone and in combination with SPION^Citrate and SPION^Gold. After exposing non-loaded and loaded THP-1 cells to an EMF for 24 h we did not observe any significant alteration in cell viability (Figure 3D). We conclude that the ROS induced by SPION^Gold was neutralized by the cells by upregulation of the antioxidant systems so that both types of nanoparticles did not decrease cell viability.

The magnetic controllability of THP-1 cells after SPION-loading was evaluated under the influence of a magnetic field in Ibidi µ-Slides. A magnet was placed onto the channel on one side, while SPION-loaded Hoe-stained cells were added on the opposite side of the slide. Then, the slides were kept at room temperature for 10 min and subsequently the localization of the cells analyzed via fluorescence microscopy. Unloaded cells and slides without magnet served as controls. The experimental setup is shown as photography from above and as model from the side in Figures 4A, B.

When cells without particle loading were used, they were not magnetically attracted. In contrast, cells loaded with 80 μg/mL SPION^Citrate or SPION^Gold showed a clear accumulation at the edge of the magnet (position 1). For conditions with less particles, only 40 μg/mL SPION^Gold enabled a weak accumulation towards the edge of the magnet. Within half a centimeter distance from the edge of the magnet, for both SPION types the cell number was visibly reduced in the immediate vicinity (position 2) because the cells centered at the point of strongest attraction (position 1). If the distance of the cells exceeded 1 cm from the magnet, the cell counts become increasingly similar to the measurements of the controls (Figures 4C–F).

Based on the magnetic properties of the disc magnet and the SPION-loaded cells, we simulated the magnetic field and force with the result depicted in Figure 5. The strongest magnetic field occurs directly below the edge of the magnet (Figure 5A top). Similarly, the strongest magnetic force occurs below the edge of the magnet, while the magnetic force rapidly decreases with distance, so that the force at the inflow is five orders of magnitude smaller (Figure 5A bottom). Figure 5B shows the simulated movement of SPION^Citrate -loaded cells over time. We compared the movement of cells loaded with 40 μg/mL and with 80 μg/mL SPION^Citrate, which were initially located in different positions as marked on the slide (see Figure 4A). For both concentrations, cells initially located within 1 cm of the magnet (Positions 2 and 3) accumulated around the magnet within minutes. At greater distances, the magnetic force diminishes, causing minimal cell movement. Overall, the computer simulation results agree with the experimental findings presented in Figure 4. Both the experiment and simulation consistently demonstrated that effective magnetic control of SPION-loaded cells is feasible only within a short range from the magnet.

The agreement between experimental results and simulations validates the computational model for future use in estimating the potential for magnetic enrichment of cells under specific conditions. More broadly, a comprehensive computational model will facilitate a more systematic and faster design space exploration than relying solely on experimental methods.