Rotating Magnetic Nanoparticle Clusters as Microdevices for Drug Delivery

Abstract

Background

Magnetic nanoparticles (MNPs) hold promise for enhancing delivery of therapeutic agents, either through direct binding or by functioning as miniature propellers. Fluid-filled conduits and reservoirs within the body offer avenues for MNP-enhanced drug delivery. MNP clusters can be rotated and moved across surfaces at clinically relevant distances in response to a rotating magnet. Limited data are available regarding issues affecting MNP delivery by this mechanism, such as adhesion to a cellular wall. Research reported here was initiated to better understand the fundamental principles important for successful implementation of rotational magnetic drug targeting (rMDT).

Methods

Translational movements of four different iron oxide MNPs were tested, in response to rotation (3 Hz) of a neodymium–boron–iron permanent magnet. MNP clusters moved along biomimetic channels of a custom-made acrylic tray, by surface walking. The effects of different distances and cellular coatings on MNP velocity were analyzed using videography. Dyes (as drug surrogates) and the drug etoposide were transported by rotating MNPs along channels over a 10 cm distance.

Results

MNP translational velocities could be predicted from magnetic separation times. Changes in distance or orientation from the magnet produced alterations in MNP velocities. Mean velocities of the fastest MNPs over HeLa, U251, U87, and E297 cells were 0.24 ± 0.02, 0.26 ± 0.02, 0.28 ± 0.01, and 0.18 ± 0.03 cm/sec, respectively. U138 cells showed marked MNP adherence and an 87.1% velocity reduction at 5.5 cm along the channel. Dye delivery helped visualize the effects of MNPs as microdevices for drug delivery. Dye delivery by MNP clusters was 21.7 times faster than by diffusion. MNPs successfully accelerated etoposide delivery, with retention of chemotherapeutic effect.

Conclusion

The in vitro system described here facilitates side-by-side comparisons of drug delivery by rotating MNP clusters, on a human scale. Such microdevices have the potential for augmenting drug delivery in a variety of clinical settings, as proposed.

Introduction

Nanoparticle technology offers ingenious strategies for improving the delivery and targeting of therapeutic molecules for cancer and many other diseases.^1–6 Magnetic nanoparticles (MNPs) may be used to enhance delivery of chemotherapeutic agents either through direct binding with the agent (which may or may not require a release mechanism), or by functioning as miniature devices (“micro-motors”) for facilitating drug transport via convection or an extended corona effect.^7–12 Despite the promise of MNPs, successful clinical use of magnetic drug targeting to date has been rare, underscoring the need for improved in vitro models to bridge the gap “from the bench to the bedside.”^10–15 In biological applications, individual MNPs tend to aggregate, forming chains and/or clusters.^16,17 The motion of such structures (under ideal conditions) has been mathematically described.^18–27 MNP aggregation produces changes in magnetic and physical properties, which may include an enhanced capacity for rotation and translational movement in response to a dynamic magnetic field.^{10,17,26–28}

Most MNPs, and certainly MNP aggregates, would be expected to have limited capacity for moving through solid tissue.^29–31 However, fluid-filled conduits and reservoirs within the body offer potential avenues for MNP-enhanced drug delivery.^11,32–34 Within these conduits, factors such as the nature of the transport medium, adhesion to the wall of the conduit, interaction with immune system cells, and distance and angle between the MNPs and the magneto-motive system, will have critical importance in determining the success or failure of the intended therapies. Surprisingly, little data is publically available regarding these specific issues, for systems that employ a rotating magnet for MNP propulsion, especially over human-sized distances.^10,11,35

Our group has been investigating a model system that allows MNPs to be “remotely” propelled on a human scale using a rotating permanent magnet.^10,11,32,34 MNPs synchronously counter-rotate in response to the magnet, act as microscopic stir rods and move by means of surface traction down lanes of specially designed cell culture trays. This surface walking phenomenon has recently been referred to as “magnetically enhanced diffusion (MED)” and “magnetically induced rotation and translation (MIRT).”^10,11 Use of the parallel “biomimetic” lanes of the tray can be used to facilitate comparisons between different experimental conditions. In two previous publications, we initially described the basic magnet and acrylic tray test configuration,¹⁰ then used it to demonstrate how tissue plasminogen activator (tPA) could be moved to the site of a blood clot in vitro using MNP clusters. In the primitive vascular model, the MNPs could be induced to move even through whole blood.¹¹ As the logical next step, we now address the principles of the system as they pertain to cancer cells, and the delivery of a chemotherapeutic agent.

Here we report our new observations regarding: 1) the relative velocities of different types of rotating MNPs, 2) the effect of magnet position and tray orientation on MNP velocity, 3) visualization of dye (as a surrogate for drug) movement in response to the MNP clusters, 4) movement of MNPs over monolayers of different cell types, and the resulting adherence to cells, and 5) MNP-facilitated delivery of the chemotherapeutic drug etoposide, as determined by cellular viability testing. Etoposide was chosen because it has been commonly used, has a well-understood mechanism of action, and has been used in the cerebrospinal fluid (CSF).^32,36 Our hypothesis was that MNP clusters could be used as microdevices to deliver etoposide by means of convection or bulk flow, in the absence of direct etoposide conjugation to the MNPs.

Methods

Magnetic Nanoparticles (MNPs): Synthesis, Size Determination, and Magnetic Separation Time

Four different iron oxide MNPs were tested in initial studies. These are referred to as Fe₃O₄ -MNPs, Fe₃O₄@Au -MNPs, Fe₃O₄@Au-GG -MNPs, and MBs. Synthesis of Fe₃O₄ -MNPs, Fe₃O₄@Au -MNPs, and Fe₃O₄@Au-GG -MNPs was according to Venugopal et al.³⁷ For this synthesis, all chemicals were purchased from MilliporeSigma (Burlington, MA) with the exception of hydrogen tetrachloroaurate (III) hydrate, which was purchased from Alfa Aesar (Haverhill, MA). Magnetic microbeads (MBs) were provided by Pulse Therapeutics, Inc. (St. Louis, MO), and have recently been described.^10,11,32 MBs are ferrimagnetic, with single-crystalline magnetite cores. The volume of MNP suspensions used in experiments described here was 20 µL (or 100 µL), from stock solutions of 25 mg/mL, unless otherwise noted.

For MNP synthesis, 3.46 g of iron (III) chloride hexahydrate (FeCl₃ • 6H₂O) and 1.27 g of iron (II) chloride (FeCl₂ • 4H₂O) were dissolved in 100 mL of 0.40 M hydrochloric acid aqueous solution (HCl_aq) to make the iron stock solution. The molar concentrations of the two iron salts in the solution were 0.128 M iron (III) and 0.064 M iron (II), respectively. Three mL of 1 M TX-100 solution was added to 250 mL of 1 M aqueous sodium hydroxide (NaOH), then heated to 75°C in a water bath. Twenty-five mL of the iron stock solution was added dropwise into this solution under vigorous non-magnetic stirring. After complete addition of the stock solution, the stirring was continued for 20 mins at 75°C. The Fe₃O₄ magnetic cores obtained were allowed to settle before magnetic separation, to remove the supernatant. These particles were then washed multiple times with distilled water to remove excess surfactant, then resuspended in distilled water, and designated “Fe₃O₄”.

For coating these MNPs with gold, 1 g of gold (III) chloride trihydrate was dissolved in 1 mL of dH₂O, while being protected from light, to create a gold-coating solution. Ten mL of the 25 mg/mL Fe₃O₄-MNP solution was transferred to a 150 mL beaker, and 90 mL of dH₂O was added for a total volume of 100 mL. 0.5 g of D (+) glucose was added to the iron core solution and sonicated for 15 mins before being placed in a 60°C water bath. Once the iron core solution reached 60°C, the gold-coating solution was added at a molar ratio of 1:2, then slowly stirred for 1 hr with non-magnetic stirring, protected from light. Gold-coated MNPs were transferred to a 50 mL Falcon tube (Corning, Corning, NY) and washed three times with dH₂O, with vortexing, sonication for 30 sec, and magnetic separation each time. The resulting “Fe₃O₄@Au” particles were stored in dH₂O at 25 mg/mL at room temperature (RT).

For the production of “Fe₃O₄@Au-GG” particles, an aqueous solution of gellan gum (Phytagel^TM) was prepared by slow addition of 5 mg of Phytagel^TM to 90 mL of water while heating using a magnetic stirring hotplate. Water was added until the solution volume reached 100 mL, and the pH was adjusted to 10 by dropwise addition of 1 M NaOH. This was then heated to 70°C while being stirred. After removing from heat, 3 mL of Fe₃O₄@Au -MNPs was added dropwise to the gellan gum solution while it was undergoing non-magnetic stirring. The solution was cooled to RT and stirred for an additional 1 hr, then washed and stored as described above.

MNP size determinations were performed using the Nanosight LM10, using the Nanosight NTA 3.0 software (Malvern, Malvern, UK). MNPs were prepared for analysis at a concentration of 1 mg/mL in PBS, then further diluted 1:100 to 10 ug/mL of MNPs. Directly prior to loading, the solution was sonicated using the Sonics Vibra-Cell (Sonics, Newton, CT). Nanoparticle motion was tracked for 30 sec in triplicate. For magnetic separation times, the Lifesep Series Biomagnetic Separator (Dexter Magnetic Technologies, Elk Grove Village, IL) was used, which was also used for MNP washing. Twenty-five mg of each MNP was loaded into separate 50 mL Falcon tubes suspended in 10 mL of phosphate-buffered saline (PBS), which were then placed into the magnetic separator. The clearance of the MNPs was video recorded using an Olympus SZ-12 camera (Olympus Corp., Center Valley, PA). The videos were analyzed using Quicktime (Apple Inc., Cupertino, CA). Separation time was measured in seconds. For CT and MRI imaging, evenly dispersed dilutions of MNPs were made in 5 mL of warm 2.5% agarose, and scanned in 12-well plates (Falcon) after refrigeration for 10 mins to harden the gel.

Rotating Permanent Magnet System (Mini-MED) and Acrylic Trays

The dynamic magnetic field used in these studies was generated by a rotating permanent magnet system (“mini-MED”) provided by Pulse Therapeutics, Inc. (St. Louis, MO). This contains a neodymium-boron-iron permanent magnet (US patents 8,313,422, 8,308,628, and 8,529,428), which spins in one direction at 3 Hz, causing agglomerated MNPs to synchronously rotate and surface walk even when separated at a distance from the magnet. The mini-Med has recently been described in detail.^10,11,32,34

Poly (methyl methacrylate) acrylic trays for testing the movement of MNPs by the mini-MED, were designed and produced as previously described.^10,11,32,34 These trays (US patent applied for), were specifically designed to study the magnetically induced rotation and translation (MIRT) of MNPs. MIRT trays are compatible with standard 96-well plate readers, as well as inverted microscopes, and therefore can also be utilized for fluorescence or dye studies. Each smooth, rounded lane in the straight lane tray is 1/8th inch (3.175 mm) wide. Here, a tray with a branching pattern was also used for dye studies. When used for tissue culture, trays were sterilized overnight using ethylene oxide, or with 100% ethanol and UV light.

Figure 1A shows a CAD image used in the design of the straight lane MIRT trays. Figure 1B is a photograph of a MIRT tray mounted on a plexiglass plate, over the mini-Med, in the offset above position (at a height of 20 cm). Figure 1C is a photograph of a portion of the tray, showing three of the eight lanes loaded with MBs (the most easily seen MNP) at their starting points, before the initiation of the rotation and translational movement resulting from starting the rotation of the magnet.

Figure 1.

(A) CAD image used in the design of the “magnetically induced rotation and translation (MIRT)” trays. The lanes have rounded bottoms to simulate tubular conduits within the body. (B) Photograph of the MIRT tray mounted on a plexiglass plate, over the mini-Med magnet, in the offset above position (at a height of 20 cm). (C) Photograph of a portion of the tray, showing three lanes (or channels) loaded with MNPs at their starting points, before the initiation of the rotation and translational movement induced by activating the mini-MED. Use of the tray lanes can be used to facilitate side-by-side comparisons of different types of MNPs, MNP modifications, the addition of drugs, or alterations of the media and cells within the channels. The tray is compatible with a standard 96-well plate reader.

Standard Tray Positions and Video Determination of MNP Velocity; Colored Dye Analysis

A diagram illustrating the different tray positions used for this paper is seen in Figure 2. The origins or starting points for the MNPs (from the vantage point of this figure), were always to the left. From this view, MNPs moved to the right, whether above or below the magnet, due to the influence of the mini-MED rotation. Use of these standardized tray positions (designated “pull”, “push”, “offset” and “offset below”) allowed comparisons to be made between MNP “runs,” and avoided non-movement areas that might be created by the formation of magnetic cancellation nodes.³⁸ For some experiments, additional vertical stacking heights were added to the basic tray position.

Figure 2.

Schematic illustrating the standardized MIRT tray positions used in this paper. This perspective is rotated 90 degrees to the left, from that shown in Figure 1B. The figure is not drawn to scale: in actuality, the tray in the pull position would be relatively closer to the rotating magnet, if the starting positions for the MNPs were all 20 cm from the center of the magnet.

Digital videos of MNPs being moved (in horizontal translation) by the mini-MED in the MIRT trays were made using an Olympus SZ-12 camera (Center Valley, PA). An “average” velocity determination for MNPs running down a lane was simply calculated from the MNP run time for the entire lane (distance: 10 cm). Velocity values in cm by cm analyses were calculated using MATLAB’s Video Viewer application (MathWorks, Natick, MA). As the leading edge of the MNP cluster passed each marked centimeter, the frame number was recorded. The difference in frames from cm to cm was divided by the frame rate (29.97 frames/sec) to determine the time for that particular span. The distance (i.e. 1 cm) was divided by the time to determine the velocity in cm increments. MNP acceleration was calculated by dividing the change in velocity readings over a given span, by the elapsed time.

In the cell-free system, MNP velocity was usually measured using 1 mL per lane of PBS at RT. 20 µL of 25 mg/mL MNP solution was added to the starting points (i.e. origins) of the lanes of the tray. MNPs were pre-magnetized by placing the tray at the side of the magnet, which consolidated the MNPs at each lane origin. The tray was then mounted at the position of interest, so that the median of imaged lanes would be along the central plane of the magnet. MBs were chosen over other MNPs for more extensive studies, since they had the greatest velocity in response to the mini-MED (as detailed in the Results section below). This improved velocity profile (other factors being equal) would be expected to convey an advantage in any future clinical setting. MB translation in the branched lane MIRT tray was studied using 8 mL of PBS, with 50 µL of 25 mg/mL MB solution being added to the starting point, again with pre-magnetization prior to the run.

Since most drugs are transparent, commercially available dyes (food coloring, trypan blue) were used to directly visualize the effects of MNP clusters on the movement of small molecules. Lanes of the straight lane tray were loaded with pre-mixed MNPs+yellow dye, versus 5 µL of yellow dye alone (which would indicate diffusion). Lanes were also loaded with 5 µL of blue dye placed 6 cm from the tray origin. Dye studies were performed with the magnet in the “pull” position. For branched lane studies, MNPs pre-mixed with yellow or blue dye were added to their respective lanes. Trypan blue was obtained from Lonza (Basel, Switzerland). To study the effect of trypan blue (molecular weight 872.88) on MB aggregation, 5 μL of trypan blue was added to 20 uL of 25 mg/mL of MBs, and the suspension vortexed for 30 sec, before being placed on a slide for photography. The effect of trypan blue on MB velocity was studied in the straight lane MIRT tray, using various tray positions.

Cell Culture, Staining, and Microscopy

Cell lines used in these experiments were obtained from the American Type Tissue Collection (Manassas, VA), along with E297 (DOI 10.1097/00006123-200103000-00035) and LKB1-KO (Jackson Laboratory, Bar Harbor, ME) cells. Cells were maintained using standard tissue culture technique at 37°C in 5% CO₂. HeLa cells were grown in MEM (Thermo Fisher, Waltham, MA.), and H2122 cells were grown in RPMI 1640 (Thermo Fisher, Waltham, MA). All other cells were grown in DMEM (Thermo Fisher, Waltham, MA). Cells were supplemented with 10% FBS (Thermo Fisher, Waltham, MA.), 1% Pen/Strep solution (Corning Inc., Corning, NY.), and 1% Ciprofloxacin (1 mg/mL) (Corning Inc., Corning, NY).

Cells were imaged either in 12-well culture plates, or MIRT trays, depending on the experiment. Phase-contrast images were obtained with an inverted microscope at 10× (Nikon Corp., Tokyo, Japan). For fluorescence or H&E staining, cells were washed with PBS before being fixed in 4% paraformaldehyde for 10 mins at RT, and washed again. Nuclei were stained in the MIRT tray by addition of 1 mL of Hoescht 33342 in PBS (1:1000 v/v; Sigma-Aldrich, St. Louis, MO) for 10 mins at 37°C. Cells in the MIRT tray were imaged using an EVOS FL Auto 2 microscope (Fisher Scientific, Waltham, MA). For scanning electron microscopy (SEM), samples were dehydrated using ethanol then hexamethyldisilazane. Glass cover slips were adhered to aluminum mounts with double-sided sticky tape and sputter coated with 6.0 nm of Pt/Pd in a low-pressure argon atmosphere for conductivity. Surfaces were analyzed at the UIC Electron Microscopy Service facility (Research Resource Center). Morphology was examined using a Hitachi S-3000N Variable Pressure Scanning Electron Microscope using secondary (SE) and backscatter (BSE) detectors.

For cell adhesion studies, cell lines were grown in 6 well plates to confluence, then treated with 1 mL of 10 µg/mL MBs for 30 mins at 37°C, then washed three times with PBS to eliminate non-adherent MBs. Images obtained by the inverted microscope were converted into binary files, and analyzed using ImageJ (National Institutes of Health) to quantify the relative number of MBs adhering to cells. For MIRT tray studies of MB velocities over cells, 1 x 10⁵ cells were seeded into each lane of the sterilized MIRT tray, which were pre-coated with 1 mL of 0.1% gelatin and incubated at 37°C for 15 min. Cells were grown for 24–48 hrs (until confluent) before further experimentation. Immediately prior to velocity determinations, the medium was replaced with 1 mL of PBS.

Evaluation of MB Delivery of Etoposide, by MTT Assay and Trypan Blue Exclusion

Since MBs were the fastest MNP, experiments were designed to determine whether or not they could be used to deliver a chemotherapeutic drug (etoposide), in the absence of direct conjugation to the drug. Etoposide was purchased from Cayman Chemical Company (Ann Arbor, MI), and stored at −20°C. Etoposide dilutions were made from a stock solution of 1 mg/mL in DMSO. The direct cytotoxic effects of different concentrations of etoposide (ranging from 1 to 100 µM) were studied prior to MB testing. Baseline (static) studies were performed in 12-well culture plates, with direct treatment of cells (U138 and H2122) in monolayer culture, using 1) etoposide alone, 2) MBs alone, or 3) a combination of MBs with etoposide, in comparison to untreated control cells. In these static studies, no magnet was used.

Standard MTT assays were performed in 12-well plates, as previously described.³⁹ Cells were seeded at a density of 10⁵ cells per well and grown for 24 hrs then treated for 24 hrs. Thiazolyl blue tetrazolium bromide powder (MilliporeSigma, Burlington, MA) was dissolved in 1X PBS at a concentration of 5 mg/mL to make the MTT working solution, which was diluted in media at 10% v/v. After cells had grown for 24 hrs, media was aspirated, cells were gently washed 3 times with 200 μL PBS, fresh media was added, and a 100 μL treatment volume was applied. After incubation at 37°C in 5% CO₂ for 24 hrs, media was aspirated, MTT-media was added, and cells were incubated at 37°C (5% CO₂) for 4 hrs. The MTT-media was aspirated, DMSO was added, and cells were incubated at 37°C for 30 mins. The resulting solution was transferred to 96-well plates in triplicate and read with a SPECTRAmax 340PC (Molecular Devices, San Jose, CA) at 590 nm in triplicate.

Translational (i.e. dynamic) experiments were then performed to determine whether or not etoposide could be transported 10 cm down the length of the MIRT tray by MB clusters in response to the mini-MED. The straight lane MIRT tray in the 20 cm pull position was used. Sixty μL of MBs alone and 60 µL of MBs plus 100 μM etoposide were loaded at the starting points of the tray. The mini-MED was turned on (“activated”) for 5 mins to ensure that all MBs had traveled the complete distance of the tray. Four hundred μL of MB solution was removed from the end of each lane and transferred to an Eppendorf tube. One hundred μL of this solution was applied to wells of a pre-seeded 12-well culture tray (U138 and H2122; 10⁵ cells/well, 24 hrs), in triplicate. After a treatment time of 2 hrs at 37°C in 5% CO₂, the supernatant (including any dead cells) was transferred into a 15 mL Falcon tube (Corning, Inc., Corning, NY) and 1 mL of media was added per well in order to harvest the remaining cells. Cells were detached with gentle use of a cell scraper and added to the 15 mL tube, which was centrifuged at 1800 rpm for 5 mins. Cell pellets were resuspended in 1 mL of media. One hundred μL of cell suspension was added to 400 μL of trypan blue. Ten μL of trypan blue cell suspension was then transferred to a hemocytometer, and live and dead cells were counted, with counts being performed 5 times per group.

Data Display and Tests of Statistical Significance

Data (e.g. the velocity curves) are expressed as mean and standard deviation (mean ± SEM) with n values as given in the text and/or figure legends. Statistical significance for the velocity and absorbance experiments was determined using two-factor ANOVA with replication. A value of p < 0.05 was considered to be statistically significant.

Results

Comparison of MNP Types: Magnetic Separation Times, Translational Velocities, and Particle Sizes

In initial studies, the magnetic separation times for the four different MNPs in PBS were determined using the biomagnetic separator. Results are shown in Figure 3A. MBs had the fastest clearing time (approximately 5 sec) due to their greater magnetic dipole moment. The movement of the four types of MNPs in the MIRT tray in response to the rotating permanent magnet (mini-MED) was then studied. MNPs form clusters in response to a magnetic field. MNP clusters counter-rotate and act as microscopic stir rods, in response to a rotating magnetic field, resulting in the surface-walking phenomenon at a distance from the magnet. Here, PBS was used as the transport media, and the pull position for the tray was used (as illustrated in Figure 2), with a starting distance of 20 cm from the magnet center. Not unexpectedly, the MBs moved the fastest in this situation as well, traversing a centimeter in approximately 2 seconds. Average velocities were plotted against the inverse of the magnetic separation times for the four MNPs, as shown in Figure 3B. Videography and digital analysis allowed for accurate quantification of MNP velocities, cm by cm, as they moved down the lanes of the MIRT tray.

Figure 3.

(A) Comparison of the magnetic separation times of the four different MNPs (n=3). (B) Average MIRT tray velocity (pull position) plotted as a function of the inverse of magnetic separation (clearance) time. The MBs separate most quickly and move the fastest in the MIRT tray. (C) MNP velocities plotted centimeter by centimeter as they move down the tray, in the pull position, showing the differences according to particle type (n≥3). (D) MNP acceleration, in the pull position, showing the greater acceleration as particles approach the magnet as demonstrated by the MBs.

Use of the parallel lanes was used to facilitate side-by-side comparisons of MNP translational velocities and cluster dispersion, according to particle type/coating. A plot of MIRT tray velocity versus distance from the tray origin for the different formulations, can be seen (for the pull position) in Figure 3C. With this tray position, MNPs accelerate due to the addition of the force of attraction to the permanent magnet, to the velocity produced by cluster rotation and surface traction, as the MNPs approach the mini-MED. This is shown in Figure 3D, which is a plot of particle acceleration as a function of distance down the tray.

MNP sizes were determined using the Nanosight LM10 instrument and NTA 3.0 software. Mean diameters for Fe₃O₄, Fe₃O₄@Au, and Fe₃O₄@Au-GG –MNPs were 95.4 ± 20.0, 105.9 ± 15.1 and 109.2 ± 9.7 nm, respectively, which indicated that the sequential coating process with gold and then gellan gum was successful and increased the particle diameter. Solitary MBs have a mean diameter of 60 nm and are prepared differently.^10,11,32 Translational velocity was therefore not just a function of particle core size. Based on these results, the MBs were used for further studies, due to their greater velocities and capacity for movement at greater distances from the mini-MED. Pre-magnetizing the MBs (by exposing them to the magnet prior to initiating magnet rotation) increased MB velocities, and was used as part of the standard procedure for subsequent experiments. MBs can be visualized radiographically, by x-ray, CT scan, and MRI. CT and MR images of MB dilutions (as indicated) are seen in Figure 4A and B. MRI, as expected, was found to be far more sensitive for MB detection than CT, especially when using T2-weighted or ultrafast gradient-echo imaging.

Figure 4.

(A) CT images of 0–500 µL of agarose-embedded MBs in tissue culture wells. (B) Ultrafast gradient-echo MR images of agarose-embedded MBs, with concentrations as indicated. Here, use of adjacent wells was avoided, in order to avoid imaging artifacts. (C) Average MB velocity according to tray position (n>3). MBs moved fastest in the pull position. (D) MB velocity according to tray position, centimeter by centimeter, as determined by video analysis. In the pull position, they accelerate as they approach the magnet; in the push position, velocity falls off. (E) MB acceleration according to tray position, which similarly shows the greatest acceleration close to the magnet, and acceleration decreasing in the push position farther away from the magnet. (F) Velocities of MBs in the offset above position, with the tray at three different heights above the mini-MED (n=9). As expected, increasing the distance from the rotating magnet resulted in decreased velocities of the particles.

Effect of MIRT Tray Position on MB Velocity

Changing the position (and/or distance) of the MIRT tray relative to the mini-MED changed the velocities of the MNPs - an important consideration (of course) when considering moving from in vitro to animal or especially human use. In data presented here, all tray positions were centered in the plane of the rotating magnetic field (as shown in Figure 2). Figure 4C shows average MB velocities in the four key positions: pull, push, offset and offset below. The standard starting point (origin) from the magnet center was 20 cm in all positions shown. The pull position produced the greatest average velocity. MNPs placed in a tray directly above the magnet (not offset) demonstrated complex patterns of motion, due to the combination of “pull” and “push” effects.

Figure 4D shows MB velocities centimeter by centimeter, as they move down the lanes of the MIRT tray. MB clusters seem to oscillate, with some expansion and contraction as they moved down the lane. MB velocity increases in the pull position at the end of the lane, since particles are approaching the magnet. This was confirmed in the acceleration plot, as seen in Figure 4E. Determinations were made to see how the amount of MBs used (i.e. aliquot size) affected their velocity, and whether lane position (e.g. lane 1 versus lane 3) was a factor in MB velocity, such as due to lack of homogeneity in the magnetic field. These results are shown in Table 1. While the effect of lane position on MB velocity was not found to be statistically significant, larger MB aliquots (100 µL in comparison to 20 µL) did indeed have greater velocities, especially in the pull position. Increasing the distance of the tray from the min-MED resulted in decreased velocities as expected, as seen in Figure 4F, for the offset (above) position.

Table 1.

Effect of Lane Position (Pull vs Offset Above) and Aliquot Size (20 vs 100 µL) on Average Velocity of MB Clusters

20 µL Pull	Average Velocity	SEM	20 µL Offset	Average Velocity	SEM
Inner	0.421	0.0513	Inner	0.228	0.0095
Middle	0.412	0.0516	Middle	0.187	0.0071
Outer	0.419	0.0463	Outer	0.189	0.0067
100 µL Pull	Average Velocity	SEM	100 µL Offset	Average Velocity	SEM
Inner	0.653	0.0762	Inner	0.241	0.0174
Middle	0.621	0.0896	Middle	0.232	0.0125
Outer	0.624	0.0826	Outer	0.215	0.0131

Dye Diffusion versus Magnetically Enhanced Dye Delivery

In order to model the effect of MBs on delivering a drug to a target area (through a conduit), studies with yellow and blue dyes were performed. Most drugs are transparent, but fluorescent drugs, or tagged therapeutic molecules, can also be studied in this fashion since the MIRT tray is compatible with standard 96-well plate readers. MBs (acting in response to the rotating magnet) were found to greatly accelerate dye delivery, when compared to diffusion alone. An example of this effect is illustrated in Figure 5A and B, for a cell-free system. Six lanes of the straight lane MIRT tray are shown, each of which contains yellow dye (representing drug) and blue dye (representing the target). Figure 5A is a photograph of the tray immediately after loading and initiating MB rotation with the mini-MED. Figure 5B shows the dispersion of the dye 30 sec later, when the MBs contact the target region. Successful dye delivery was confirmed once green coloring appeared.

Figure 5.

(A and B) A comparison of dye diffusion versus magnetically enhanced dye delivery, as demonstrated in the MIRT tray. (A) Positions of MBs with dye (top three lanes) and dye samples alone (bottom three lanes) at the time of mini-MED activation, with blue dye targets in the central portion of the tray. (B) Positions 30 sec later, when the MBs bring the yellow dye into contact with the blue target region (top three lanes), versus dye diffusion proceeding without MBs (bottom three lanes). (C) Bar graph showing a comparison of the velocities of MB-induced dye delivery versus movement by diffusion alone (n=3). (D) Photograph showing dye movement in response to MBs in a branched tray system, after 15 mins of mini-MED activation. MBs drag blue and yellow dyes (as drug surrogates) along with them, which then co-mingle to form green. (E and F) Effect of trypan blue dye on MB dispersion. (E) Photomicrograph of MBs in PBS. (F) Photomicrograph of MBs in PBS with addition of trypan blue, showing the resulting increase in particle dispersion (10× magnification).

The average velocity of the yellow dye as propelled by the MBs (0.18 + 0.01 cm/sec), versus by diffusion alone (0.005 + 0.0006 cm/sec), is graphed in Figure 5C. ANOVA two-factor analysis with replication showed that these values are statistically different, with p≪0.05. In both situations, dye does become diluted with movement down the lanes of the tray. MBs were able to transport dye (as a drug surrogate) in a branched lane tray as well, as illustrated in Figure 5D. Dye delivery could therefore be used to visualize the effect of MNPs as microdevices for drug delivery (i.e. by convection or bulk flow), in the absence of direct conjugation to the particles. MB-facilitated movement of trypan blue dye also studied. Trypan blue, which is negatively charged, inhibited the formation of MB aggregates when vortexed with MBs, as seen in Figure 5E and F. Combining MBs with trypan blue, did slow translational velocity somewhat, depending on tray position. Individual dyes were therefore found to respond differently in the model system, according to factors such as size, concentration, and surface charge, and the same would be anticipated for different drugs.

Movement of MBs Over Cellular Monolayers and Adherence of MBs to Cells

Interaction of MBs with cells cultured in monolayers was then studied, for several different cell lines, including HeLa (cervical cancer), human glioblastoma lines, and LKB1-KO. Different MB volumes, and cellular confluence percentages, were tested. MBs placed at the edge of a monolayer of cells in a standard 6-well plate, then exposed to the action of the mini-MED, tended to fan out to both sides, as seen in Figure 6A. Effects resulting from proximity to the side of the well, and at the fluid meniscus, were also observed. MBs were seen to adhere differently to different types of cells forming a confluent monolayer. Phase-contrast microphotographs, and images of cells stained with hematoxylin and eosin, are shown in Figure 6B and C, respectively. A scanning electron microscope (SEM) image of an MNP cluster on a glioblastoma cell is seen in Figure 6D. Fe₃O₄@Au-MNPs were less toxic than Fe₃O₄ -MNPs, at a treatment time of 24 hrs. No evidence was observed of any significant cellular toxicity resulting from exposure to MBs. Image J was used to quantify MB adhesion to the cancer cells after cell washes, with results shown in Figure 6E. MBs were seen to be more adherent to live cell monolayers than to paraformaldehyde-fixed cells, and more adherent (in general) to cancer cells than non-neoplastic cells. Figure 6E shows MB adherence to live and fixed glioma cells, versus cultured non-neoplastic astrocytes.

Figure 6.

(A) Photograph showing the pattern of MBs as they fan out across a cellular monolayer. (B) Photomicrograph showing the adherence of MBs to live cancer cells, as seen by phase-contrast microscopy. The MB clusters are seen here as highly refractile (white) particles. (C) Photomicrograph showing the adherence of MBs to fixed cancer cells (E297, human glioblastoma), after H&E staining. (D) A scanning electron microscope (SEM) image of an MNP cluster on a glioblastoma cell (arrow). (E) Bar graph indicating the relative adherence of MBs to different types of cells, unfixed (blue) versus fixed with paraformaldehyde (gray), as quantified using ImageJ (LKB1, E297 and U87 are glioma cells; Astrocytes are not neoplastic).

Cells were then grown in the lanes of the MIRT tray, in order to channel the translational movement produced by the mini-MED, as could be utilized in order to transport MNPs within physiologic conduits inside the body (such as blood vessels, the spinal subarachnoid space, ureters, lymphatics, etc.) Our data on vascular cells have been recently reported.¹¹ Cells were found to grow better in trays milled from acrylic, than in trays 3D printed in VeroClear. Bits of the poly (methyl methacrylate) acrylic were found to have less of a growth-inhibitory effect on U87 cells at 48 hrs, than similar pieces of VeroClear. Cells could be successfully grown in MIRT trays for up to 120 hrs, depending on the type of cell, and number of cells initially seeded.

Dynamic experiments were performed, with MBs transported down the lanes of the MIRT tray by the mini-MED, over cellular monolayers of a variety of cell types. This was to determine whether or not MB velocity would be affected differently, by different types of cells coating the lanes, and to determine whether or not there would be particle attrition due to adhesion of particles to the walls of the lanes. A clinically relevant scenario, for instance, might be that of clusters of cancer cells seeded within the spinal subarachnoid space, which might be more adherent to drug-laden MNP clusters. Of the cell lines tested, MIRT lane velocity was the least for U138 and E297 cells, which corresponded to their high MB adhesion rates. A photomicrograph of Hoechst 33342-stained U138 cells in a lane of the MIRT tray after MBs have passed, is shown in Figure 7A. Cells are blurry at the top of the photo, due to the curvature of the tray lane. Residual MBs can be seen adhering to the cells, as indicated in white using this technique.

Figure 7.

(A) Photomicrograph of residual (adherent) MBs in a MIRT lane containing U138 (human glioblastoma) cells. Cell nuclei were stained with Hoechst 33342. MBs clusters appear white on this image. Cells are blurry at the top of the photo, due to the curvature of the tray lane. (B) Graph showing the effect of cell type on MB velocity in the MIRT tray, in the offset above position. Different cell types were found to have different effects on the velocity and dispersion of the MB clusters.

The MB velocities over cellular monolayers (centimeter by centimeter) of different cell types are shown in Figure 7B. For these studies, the above offset position was used. The average velocities of MBs over HeLa, U251, U87, and E297 cells were 0.24 + 0.02 cm/sec, 0.26 + 0.02 cm/sec, 0.28 + 0.01 cm/sec, and 0.18 + 0.03 cm/sec, respectively. The attrition of MNP clusters with movement, due to MNP adhesion to the lanes, was also greater with some cells, such as U138 cells. MB movement over certain cell lines was found to be significantly slower (and with higher cluster dispersion), likely indicating their relative “stickiness”. U138 (human glioblastoma cells) for example, showed marked MNP adherence, and an 87.1% reduction in velocity (in comparison with U87 cells) at a distance of 5.5 cm along the channel. This difference was found to be statistically significant by ANOVA, with p≪0.05. It would be advantageous to be able to quantify (and exploit) such adherence to cancer cells in human applications of magnetic drug targeting.

Etoposide Cytotoxicity Testing and MB Drug Transport Over Human-Sized Distances

U138 cells in wells were treated with various concentrations of the classic chemotherapeutic agent, etoposide, in order to establish parameters for assessing cytotoxic activity. Direct treatment with 100 µM etoposide for 24 hrs (without employing the magnet) was found to decrease cellular viability to 23.0 + 0.4%, as studied by MTT assay. These results are shown in Figure 8A.

Figure 8.

Etoposide cytotoxicity in static vs dynamic (MIRT tray) studies. (A) Viability analysis of etoposide treatment of U138 cells in a 12-well plate, as determined by MTT assay. 1 µM, 10 µM, and 100 µM etoposide were used and compared to 10 µL of DMSO (10 µM volume equivalent) alone per well (n=3). (B and C) Trypan blue cell viability analyses after MBs were used to move etoposide down the lanes of the MIRT tray. (B) U138 (human glioblastoma) cells. (C) H2122 (human lung adenocarcinoma) cells. MBs successfully transported etoposide by convection/bulk flow (5 mins) to allow cancer cell killing as shown by this assay (n = 3; *p < 0.05).

Etoposide was added to lanes at the starting points of the MIRT tray, with 60 µL aliquots of MBs. Control lanes were also employed, which used MBs alone. MBs were remotely moved down the length of the tray (10 cm) in PBS using the mini-MED, with the tray positioned at 20 cm in the pull position. After 5 mins, MBs (with associated media) were removed from the ends of the lanes and used to treat U138 and H2122 cells for 2 hrs. Cells were analyzed for cytotoxic effect using trypan blue exclusion. Results are shown in Figure 8B and C, for U138 (human glioblastoma) and H2122 (human lung adenocarcinoma) cells, respectively. The cytotoxic effect of etoposide was retained after being delivered by MBs, and this result was found to be statistically significant (p < 0.05). Such an effect could not have been achieved by means of drug diffusion (with associated dilution) alone, therefore indicating the successful delivery of drug by the rotating MNP clusters.

Discussion

Data reported here show that: 1) The translational velocities of MNP clusters moving down the lanes of the MIRT tray (in response to the rotating magnet) can be roughly predicted from their clearance times in the magnetic separator; 2) Changes in the position of (or distance from) the rotating magnet predictably change the velocity of the MNPs, as does the size (i.e. mass) of the MNP aliquot; 3) Dye delivery can be used to visualize the effect of MNPs as microdevices for drug delivery; 4) MB clusters directed down MIRT tray channels lined with cells exhibit varying velocities and degrees of wall adherence, depending on cell type; and 5) MB clusters can be used to enhance the delivery of the chemotherapeutic agent etoposide, across the “mesoscale” distance of 10 cm. Non-functionalized iron oxide MNPs were studied as convective microdevices, and coating with gold and gellan gum did change the observed velocities. MBs were chosen over the other MNPs for more extensive studies, since they demonstrated the greatest velocity in response to the mini-MED. An improved velocity profile (other factors being equal) would be expected to convey an advantage in a clinical setting in which drug delivery is being augmented.

Admittedly, the in vitro model and test conditions used here are rudimentary and do poor justice to actual physiological conditions. Yet, the principles such experiments establish set the stage for achieving a more sophisticated understanding of MNP movement on a human scale, and ultimately for possible therapeutic use. Modeling of more complex conduit branching patterns in three dimensions would certainly be useful for predicting MNP movement and drug delivery within a variety of human structures and would add complexity in exponential fashion. With larger, more powerful magnets, the distance separating the magnet and MNPs could be greater, and with higher rotational speeds, the surface walking of the particles could potentially be faster. In the results reported here, MNP clusters functioned as medical devices (or “nanobots”) enhancing fluid transport for drug delivery^11,40,41 and did not require direct drug conjugation or a mechanism for triggering drug release.

When MNPs enter an environment containing cells, they adsorb biomolecules and form a corona, assuming a biological identity which is distinct from their synthetic identity.^42–47 The composition of the bio-corona will depend on the specific environment in which the particles are used, and can include solutes, lipids, protein, and other molecules.^1,42,48–50 Surface charge and corona formation can be altered by nanoparticle coatings, which in turn affect nanoparticle adherence to (and uptake by) cells.^13,51,52 Different types of cells have been studied for this effect including macrophages, fibroblasts, osteoblasts, and HeLa, neuroblastoma and ovarian cancer cells.^{44,51,53–59} Since corona formation may affect MNP targeting and drug delivery, this phenomenon provides another rationale for laboratory modeling (such as our own), prior to resource-intense in vivo studies. The MBs studied here have been reported to rotate and display translational movement through serum and even blood,^10,11 fluid environments shown to generate a corona. Use of the MIRT tray and mini-Med for primitive modeling of blood vessels, for analysis of tPA delivery to a blood clot, has recently been reported.¹¹

Considering MNP modeling relevant to cancer, studies have shown that a variety of chemotherapeutic drugs can be bound to, and transported by MNPs.^60–63 MNP-chemotherapy constructs have demonstrated growth-inhibitory effects in vitro and in animal studies for many types of cancer cells, including glioblastoma.^2,64-66 Ingenious larger nanostructures (“micro-robots”) which can be actuated by external magnetic fields have also been described, with the goal of targeting drug delivery.^67–70 These may or may not contain iron, or utilize a tumbling motion.^67,71 Few reports have described the use of unbound MNP clusters (also referred to as “aggregates” or, more accurately “agglomerates”) as microdevices for delivering therapeutic agents at human-sized distances.^11,33,72 As with solitary nanoparticle-drug conjugates, multiple strategies are envisioned that would allow MNP clusters to disperse and release drug at their target site.⁷³ Tumors demonstrate an “enhanced permeation and retention effect”, and individual tumor cells, as well, are likely to demonstrate heightened nonspecific binding and uptake of drugs and nanoparticles due to their greater metabolic needs.^74,75 In the CNS, reports have demonstrated that magnetic particles can be moved within the CSF-containing conduit of the spinal subarachnoid space.^76–78 These studies have employed either fixed (non-rotating) permanent magnets or static electromagnetic fields.

While the surface walking phenomenon of magnetic particles in response to a rotating or oscillating magnetic field has long been recognized, its potential clinical application for drug delivery has recently gained new appreciation.^{3,10,11,40,41} The main advantages of using a rotating permanent magnet as a magneto-motive unit (over that of a static or oscillating electromagnetic field) are that: 1) it may be able to induce MNP cluster movement at a greater distance; and 2) it can act unilaterally, i.e. it does not have to surround the target region.⁴⁰ Few animal studies have yet been reported which utilize this rotational mechanism.^41,79 Additional data are needed regarding many important and fundamental considerations of the variables influencing rotating MNP clusters as devices for drug delivery. While the human condition entails even more complexity, a list of factors that could affect rotational magnetic drug targeting (rMDT) – just in vitro – is given in Table 2.

Table 2.

Variables Influencing Rotational Magnetic Drug Targeting (rMDT) in vitro (Factors May Be Redundant, Overlap, Interact)

A. MNP / Cluster	B. Tray / Lane
MNP type / size / composition / heterogeneity	Composition
MNP magnetization / zeta potential	Lane width / radius
MNP charge / coating	Smoothness / irregularity / pitting
Prior magnetic exposure	Electrostatic effects / charge
Composition of corona	Elasticity / flexibility
Particle mixture / ratios	Erosion / friability
Aggregate / cluster size	Coating
Cluster stability / disassembly	Branching / curvature
Cluster cohesion	Wall shear / margination
Torque / traction	Wall adherence
Cluster adherence / interactions	Barrier phenomena (e.g. encountering target cell cluster or clot) / reflection
Coefficients of friction: rotational	Coefficients of friction: static, dynamic
Wave phenomena / energy dissipation
C. Media	D. Cellular lining
pH / temperature	Type / size
Viscosity / cavitation	Density / % confluence
Composition (protein, lipid, etc.)	Surface characteristics / waviness
Heterogeneity	Metabolic / endocytotic activity
Volume / depth / meniscus	Lane adherence
Cells suspended in media (e.g. RBCs, immune cells)	Cluster adherence
Bulk fluid flow / axial pulsations	Cellular mixture (e.g. endothelial plus cancer)
Webbing / matrix / scaffolds	Ciliation
E. Magnetic source	F. Drug / Miscellaneous
Distance	Drug type / size
Position / rotation (3D)	Drug charge / solubility / lipophilicity
Speed of rotation / oscillation	Nonspecific loading / adherence to MNPs
Field strength / heterogeneity	Ratio of cluster size to lane width
Permanent vs electromagnetic	External pulsations

Clinical Applications of Rotational Magnetic Drug Targeting

MNP clusters induce drug movement in the absence of direct drug conjugation to the MNP itself. They are envisioned as acting as implantable micro-motors, which may have a capacity for being (at least partially) removed, due to their mobile, magnetic nature. Accordingly, the FDA approval process may be different from that of a drug–nanoparticle conjugate.^73,75 The advantage of magnetic drug targeting (MDT) lies in its potential ability to control the concentration and distribution of drug within a desired target region. This is usually accomplished with external driving magnets, and is expected to minimize toxicity in untargeted tissue.^29,80 Maximal pull or push forces can be used within the body to direct MNPs to precise locations, even against currents.^38,80 To date, human trials of MDT have been restricted to a focusing depth of 5 cm.³⁸ Halbach arrays have been successfully used to optimally project magnetic forces in clinical applications.^81,82

While MNP clusters might not penetrate solid tissue, they may be able to directed through conduits (such as blood vessels, lymphatics, or the spinal subarachnoid space) or across surfaces (such as the uroepithelium, or peritoneal cavity).^65,83 Potential avenues and targets for MNP delivery, including implants such as catheters, are listed in Table 3, according to organ system and medical specialty. Adhesion of particles to the wall or surface, along with margination and wall shear, are major factors in determining whether or not MNPs will be successfully directed to their targets.^84–86 Orientation of the rotating magnet will also play a critical role, and this may need to change during targeting along a complex route. Strengths arguing for the clinical use of MNP clusters include their ability to be moved through stagnant fluid (such as to an area of obstruction), or even against flow and pulsations.^11,17,72 Uncertainties (as with MNPs in general) include their possible long-term toxicity, and/or generation of an immune or inflammatory response.¹⁴ Within the arterial system, formation of emboli by MNP clusters could be a major concern and would have to be avoided, especially in organs such as the brain and heart, until safety was ensured. Another difficulty arises from the general incompatibility of magnets with electronic devices, especially in complex medical environments. These issues need to be further addressed experimentally.

Table 3.

Applications for Remote MNP-Enhanced Drug Delivery for Representative Conduits and Surfaces in the Human Body

Specialty / System	Naturally Occurring Structure	Implanted Device	Potential Disease Application
CNS – brain	Ventricular system	VP shunt, EVD, Ommaya reservoir	Cancer, hydrocephalus, IVH
	Cerebral arterial system	Stent	Stroke
	Cerebral venous system		Venous thrombosis
	Virchow-Robin spaces		Multiple (general CNS drug delivery)
CNS – spine	Spinal subarachnoid space	Infusion pump, CSF port	Cancer, spinal cord injury, pain, etc.
GU	Bladder, kidney	Catheter	Cancer, obstruction, hemorrhage
	Ureter		Kidney stones
	Fallopian tube, seminal tubule		Conception
GI	Peritoneal cavity		Cancer, infection, endometriosis
Cardiac	Coronary arteries	Stent	MI
	Pericardium		Pericardial effusion
Vascular	Veins, arteries	Stent	Thrombosis, insufficiency
Pulmonary; thoracic	Bronchioles		Asthma
	Pleural cavity	Chest tube	Pleural effusion
Orthopedic	Joints: knee, shoulder, hip	Artificial joints	Arthritis
Ophthalmology	Retina, cornea, aqueous humor	Implants	Decreased vision, glaucoma

Abbreviations: VP, ventriculoperitoneal; EVD, external ventricular drain; IVH, intraventricular hemorrhage; MI, myocardial infarction.

MNP clusters can contain more than one type of particle. Sophisticated MNP-based drug delivery vehicles have recently been reported, which combine thermotherapy and chemotherapy with encouraging results in vitro and in lung-xenograft – bearing mice.⁸⁷ Hybrid MNPs have been developed, which allow for chemotherapy release in response to magnetically induced heating or a low-power radiofrequency field.^88,89 Delivering such particles, or particle clusters having multiple components, via conduits by the rotational mechanism we have described, may open up new avenues for successful MDT. Magnetically enhanced delivery of chemotherapy would be expected to have greater efficacy if administered earlier in the course of cancer dissemination. Early seedings of cancer cells could possibly be treated through access afforded by the lymphatic system, genitourinary system, and peritoneal cavity, as examples. An expected early clinical application of MIRT technology in the CNS would be in the treatment of patients with leptomeningeal metastases.^90–93 The greatest impact would be in treating (at an early stage) cancer cells disseminating via CSF from a primary brain tumor such as medulloblastoma, since systemic metastases would not have already occurred.⁹⁴

Conclusions

Rotating MNPs clusters can be used as simple microdevices for drug delivery, including chemotherapeutic agents as we demonstrate here. Precisely directing such drug carriers remains the “Holy Grail” of MDT.^8,95 Basic factors such as magnet position and the cellular lining of the access channels must be critically analyzed by in vitro modeling in order to predict MNP behavior in vivo. Velocity and cell surface adhesion of MNP clusters can be quantified, using video photography and digital analysis. Further studies need to be performed to determine whether or not MNP clusters – or their payload – may preferentially attach to (or be taken up by) cancer cells more than non-neoplastic cells, especially in vivo. Otherwise, additional targeting and release mechanisms may be required – functions for which hybrid MNPs are well suited.⁷³ Achieving an improved understanding of the factors that influence MNP agglomeration, stability, translational movement, drug binding and release, and ultimately uptake by the target cells, will be crucial for making the jump to successful human applications of rMDT.

Funding Statement

Funding for this project was through The Brain Tumor Research Fund at The University of Illinois at Chicago Cancer Center, Chicago, IL, USA.

Disclosure

Michael E Sabo was a full-time employee of Pulse Therapeutics Inc. from June 2009 to April 2019. Francis M Creighton was employed by Pulse Therapeutics Inc. Mr. Sabo and Dr. Creighton are currently employees of UNandUP, Inc., St. Louis, MO, USA. A patent “Tissue Culture Tray and System for the In Vitro Testing and Characterization of Magnetically Induced Rotation and Translational Motion of Magnetic Particles”; July 2017 US20180024109A1 was described and used in this paper. The authors report no other conflicts of interest in this work.