Degradable and Biocompatible Magnesium Zinc Structures for Nanomedicine: Magnetically Actuated Liposome Microcarriers with Tunable Release

Inorganic therapeutic carriers and implants should not only be biocompatible, but should also degrade under physiological conditions. Ideally, the time of the degradation can be controlled, and ideally the degradation products are fully biocompatible and metabolized by the body. This proves a challenge for carriers used in nanomedicine, including microswimmers and nanorobotic systems destined for targeted delivery, as these generally require inorganic materials to enable coupling to external fields. Taking inspiration from macroscopic orthopedic implants that are made from magnesium (Mg) and zinc (Zn) and that are fully biocompatible and degradable, the growth of complex microstructures is demonstrated, including micropropellers, containing Mg and Zn. By varying the content of Mg, the corrosion time of the microstructures can be tuned from hours to over a month. Incorporation of biocompatible hard‐magnetic iron (Fe)‐platinum (Pt) permits the controlled motion of the micropropellers. The surface of the MgZn structures can be functionalized with liposomes, rendering the structures microcarriers that allow for a time‐dependent release of their cargo as a results of their degradation in aqueous environments. This suggests a powerful platform for targeted drug or gene delivery, that can be integrated with established systems for magnetic actuation and transfection.


Introduction
Actively propelled nano-and microswimmers are promising candidates for targeted delivery. [1,2]As opposed to the conventional systemic administration of therapeutics, which relies on passive distribution, local active targeting should have higher efficacy and in addition may even allow for the use of more potent therapeutics.To realize active propulsion, however, additional demands are placed on the structure and material composition of the carrier.[5] Similarly, the propulsion with external magnetic or ultrasound fields requires carriers that are fabricated with complex geometries and shapes, such as helices, [6][7][8] wires, [9][10][11] rods, [12] rockets, [13,14] and involve the use of inorganic materials to provide the desired functionalities.
Determining the longer-term effects of inorganic nanomaterials in vivo is, however, complex.For instance, a recent study by Ramachandran et al. [15] have found that while magnetic micromotors did not exhibit any signs of toxicity in vitro and in vivo over the course of one week, they nevertheless affected the tissue and activated an immune response.
It is therefore interesting to examine inorganic materials that provide the required performance when it comes to the coupling with external fields, but that at the same time degrade after their intended use. [16,17]In this case, questions of longer-term mechanical or chemical effects do not arise.In this context, it is interesting to ask if magnesium (Mg) and zinc (Zn), and their alloys, which degrade naturally under physiological conditions, and where some by-products can even be metabolized by the body, can be utilized in micro-and nanomedicine for this purpose.[20][21][22][23] Such degradable implants underly the biocompatibility testing standards for medical devices (ISO 10 993), which are currently being adapted to absorbable (magnesium) implants. [24]nvestigations into the in vitro [20,21,[25][26][27][28][29][30] and in vivo [22,[31][32][33] degradation behavior on the macro-and centimeter-scale have been numerous.It has been reported that the addition of Zn can stabilize the otherwise rapid degradation of Mg. [34][35][36][37] The corrosion rates are highly dependent on various parameters, such as the material purity, shape, and composition, as well as the surrounding medium, pH, fluid flows, etc., but overall Mg is expected to degrade two to twenty times faster than Zn. [23,27,29]On the micro-and nanoscale, a larger proportion of atoms is located on the surface, as compared to the bulk.Since decomposition reactions occur via the surface, it is expected to be much faster at the micro-scale.The rapid corrosion is potentially a hindrance to building stable microcarriers.[40] Another bottleneck for the use of MgZn in micro-and nanodevices might be the size limitation, since traditional fabrication processes for biodegradable metals and alloys, involve casting, powder metallurgy, transient directional solidification, and additive manufacturing [18] which do not have the precision to form structures below the millimeter scale.
While the utility of MgZn structures has been established at the macro-scale, it is, for the aforementioned reasons not clear, whether the degradation processes translate similarly to the micro-scale.In addition, the materials pose challenges for microand nanofabrication and it is unclear if suitable carriers can be realized that interact with external fields and that can actively carry therapeutics.
Here, we designed and fabricated complex biocompatible microcarriers in the form of helical micropropellers consisting of Mg and Zn, whose degradation time can be set by the Mg:Zn ratio to lie between hours and a month.As we show, the fabrication of MgZn structures with varying composition by physical means is possible despite high adatom mobilities.We demonstrate the successful growth of micropropellers containing varying ratios of Mg and Zn and the incorporation of biocompatible hard-magnetic iron-platinum (FePt) for magnetic manipulation and propulsion.We also show the functionalization of the pro-peller surface with liposomes, which thereby enables their guided delivery with subsequent, time-dependent release for potential use in nanomedicine.
We previously showed that a specialized physical vapor deposition method, termed glancing angle deposition (GLAD), [41] allows the fabrication of custom structures, including helical propellers in a wide range of sizes from tens of nanometers [42] to micrometers (Figure 1a). [43,44]Here, we use the method to grow structures with well-controlled MgZn ratios (Figure 1b), as this determines the degradation rate.The incorporation of a magnetic section based on Ni, Co, or Fe has been used for the actuation of these structures via weak rotating magnetic fields (below 10 mT). [43]Actuation of magnetic micropropellers through biomedically relevant fluids, such as the vitreous of the eye [45] or mucus of the stomach has been demonstrated. [46]Qiu et al. showed targeted gene delivery to human embryonic kidney (HEK 293) cells in vitro using a similar approach. [7]Pal et al. proved propulsion inside live cells. [47]Some of the magnetic materials used are either toxic [48] or possess weak magnetization, making them unattractive for biomedical applications.Recently, facecentered tetragonal (fct) or L1 0 FePt (henceforth shortened to FePt) emerged as a promising hard-magnetic material [49][50][51] that combines high magnetization with biocompatibility. [52]FePt provides a much stronger magnetic moment compared to Fe alone.This leads to a number of advantages and means that one can i) use less magnetic material, ii) apply stronger fields without re-magnetizing the structure, and/or iii) obtain larger torques for the same applied magnetic field strength.FePt thin films and structures can be grown via GLAD co-deposition of Fe and Pt, followed by an annealing step. [8]We further showed magnetically guided FePt propellers delivering a plasmid coding for an enhanced green fluorescent protein to A549 lung carcinoma cells. [8]This makes these propellers promising candidates as biocompatible carriers (Figure 1c,d) for therapeutic payloads.We have adapted the fabrication protocol (as described below) to enable the incorporation of Mg and Zn.Control over the material composition and shape of Mg and Zn based biodegradable microstructures allows for the rational design of guided carrier systems with time-dependent degradation and specific cargo binding and release capabilities.
To function as carriers, the propellers need to be loaded with cargo (Figure 1d).55][56] Liposomes can encapsulate a diversity of biological drugs, such as nucleic acids, pDNA, gene editing tools via CRIPPR/Cas9, proteins, antibodies, [53] as well as hydrophilic and hydrophobic chemical drugs, [57][58][59] while being biodegradable, and customizable in their size, composition, [60] and functionality.[63] Tuning the surface of liposomes with targeting moieties can enhance their ability to cross biological barriers. [64]However, this targeting relies on a stochastic encounter between the targeting moiety and a its receptor on the tissue, without any navigational capability of the particle, [65,66] although some recent work has indicated that chemical reactive asymmetries provide a potential route to (unguided) active motion. [67]Utilizing liposomes coupled to a magnetic microcarrier provides additional directional control not available with any other method.Qiu et al. show such a design, [68] but their 16 μm long propellers are not biodegradable and include nickel, which exhibits toxicity for human cells. [69]The liposome functionalization was achieved by unspecific adsorption, while the release is triggered through a heat stimulus.Unspecific adsorption or charge-based attachment strategies influence the selection process of potential cargoes.The surface of inorganic MgZn microswimmers discussed herein allows for covalent bonding of cargo through silane chemistry, giving rise to a wide range of cargo choices, e.g., neutral lipid nanoparticles.The degradation of the structure automatically causes the release of the cargo (Figure 1e).

Fabrication
The MgZn structures of this paper were fabricated using electron-beam assisted physical vapor deposition.This method affords excellent control over the composition and shape of the micro-and nanostructures.As a vacuum technique it does not require stabilizers, and thus ensures that pure materials can be used to grow highly uniform structures.These permit a detailed analysis of the degradation properties.From an array of possible shapes, including spherical and rod-shaped, the method we use also permits the growth of helical micropropellers that can also be actively propelled, when they are magnetic. [43]The more complex micropropeller shapes are also helpful in determining the state of degradation with electron microscope imaging.We fabricated ≈1.2 μm long helical propellers from MgZn as schematically depicted in Figure 2a.In the first step, polystyrene (PS) beads were patterned on a substrate to form a hexagonal closepacked monolayer and etched in O 2 plasma to reduce their size and to increase the inter-particle spacing.In a second step, the substrate was placed in a vacuum chamber that is equipped with a motorized stage for shadow growth with substrate rotation, also known as glancing angle deposition (GLAD). [41,42,70]The beads serve as the seed layer for the shadow deposition.The main aim is to grow Mg and Zn containing microstructures whose degradation characteristics can be determined at small length scales.In addition, we included a thin magnetic layer so that the structures can also be manipulated and propelled with magnetic fields.The latter extends the range of potential applications and provides, via its mobility that in itself requires a connected helical shape, another measure for the integrity of the structure during the course of its decomposition in water.In order to retain the full biocompatibility of the structure, FePt was chosen as the magnetic material.The fabrication of nano-and microstructures containing FePt has recently been realized. [8]It requires an annealing step to obtain the desired magnetic phase.The annealing step needs to precede the deposition of the MgZn, as is discussed below.First, a thin layer of MgO was deposited on the plasma-treated polystyrene beads, followed by the co-deposition of 40 nm of FePt at an atomic ratio of 1:1, followed by a thin Ti layer to facilitate adhesion of MgZn that is deposited subsequently.The co-deposited FePt has to be annealed at ≈680 °C to form the desired hard magnetic L1 0 phase. [8]Since the melting points of Mg (T m ≈ 650 °C) and Zn (T m ≈ 420 °C) are below the annealing temperature, the sample was first removed from the chamber and annealed in a furnace (step 3).During the annealing step the polystyrene seeds were also "burned" off and only the hard-magnetic FePt L1 0 halfspheres remained.
After the removal of polystyrene (see Experimental Section for further details) the FePt half-spheres remained on the substrate in their original arrangement, well separated for subsequent shadow growth (Figure S1, Supporting Information), which allowed for the GLAD co-deposition of MgZn (step 4) to form a left-handed helix with 1.3 turns. [44]The Mg content was adjusted and in different depositions set to 90, 80, 70, or 50 atomic %, as determined during fabrication by integrated quartz crystal microbalances (QCM).The corresponding Zn concentrations were set to 10, 20, 30, and 50 atomic%.All samples were deposited on the same seed layer at an angle of 87°, while cooling the substrate to ≈90 K.The relatively high adatom diffusion [71] of these light metals impedes the formation of complex shapes.The fabrication of helical Mg microstructures thus required special control over substrate spacing, as well as deposition angle, rate, and temper-ature.Hence, we have increased the spacing between the seed microparticles (as described above) to minimize the accidental fusion of particles during deposition.In addition, the steep angle of incidence (87°) and the cooled substrate (≈90 K) helped reduce adatom mobilities and improved the quality of the fabricated shapes.Post fabrication, the micropropellers were characterized with electron microscopy imaging (Figure 2b,c) and the magnetic properties were analyzed using a superconducting quantum interference device (SQUID), see Experimental Section and Figure S2 (Supporting Information).
The propellers have an outer diameter of 560 nm and a length of ≈1.2 μm with a pitch of ≈900 nm.The in-plane magnetic moment for the Mg:Zn 70:30 sample was determined to be 7.4 × 10 −12 emu per propeller with a remanent magnetization of 426 emu cm −3 , which is in agreement with previous measurements. [8]Since all structures are grown on highly uniform FePt half-spheres, the magnetic properties are also expected to be highly uniform.

Degradation
Micropropellers of different MgZn ratios were fabricated to study the corrosion behavior.Degradation tests for the whole range of fabricated MgZn ratios were conducted by dispersing the micropropellers in water at 37 °C after release from the substrate with a short sonication step.The integrity of the structures could be observed visually by monitoring over time after drying via SEM imaging at different intervals (Figure S3, Supporting Information).We found, that propellers with a Mg content of 90% and Zn content of 10% completely decomposed in under 6 h, making them difficult to handle.However, 80% Mg to 20% Zn propellers retained their shape for roughly 24 h in solution, while the 70% Mg to 30% Zn propellers maintained their shape for at least 140 h.A decrease of Mg by 10% and correspondingly an increase of Zn resulted in 5-6 times longer degradation times.Accordingly, propellers with Mg:Zn content of 50% took several weeks to exhibit any noticeable corrosion.This not only indicates that MgZn microstructures with varying composition can be grown that possess the intended complex shape at the microscale, but also that the obtained tunable degradation times are in the range that is of interest for applications.
Based on the typical workflow for utilizing micropropellers, for instance in a transfection setup, will include cargo loading, administration, propulsion, and delivery, and so the micropropellers need to remain structurally and functionally intact for at least a day.Propellers with a Mg content greater than 80% were therefore excluded in the subsequent experiments.Taking into account the degradation observed for different micropropellers, the ideal ratio of Mg:Zn is between 70% to 80% Mg, based on the desired lifetime.The exact degradation time will depend on the pH, as well as the temperature and ionic strength of the target environment.However, the structural integrity as observed by SEM (Figure S3, Supporting Information) does not alone guarantee the necessary functionality of a propelled transfection agent, therefore we also performed quantitative analyses as well as functional tests that are based on the propulsion of the MgZn structures.
First, we performed quantitative dissolution tests with inductively coupled plasma-optical emission spectroscopy (ICP-OES), to examine the degradation behavior of Mg and Zn.MgZn micropropellers grown on substrates (see Experimental Section) were placed in 5 mL of DI water at 20 °C and the supernatant was analyzed over 11 days.The increase in concentration of dissolved Mg 2+ and Zn 2+ over time was observed and after every measurement, the solution was completely replaced, to avoid ion saturation in the solution due to confinement and thereby ensure a constant degradation of the structures.The total amount of Mg and Zn (dissolved and undissolved) was determined by dissolving the remaining structures in a final sample in 0.1 m HCl and adding this amount to that determined over all the collected solutions.The Mg and Zn detected by ICP-OES is seen in Figure 3a,b which charts the corrosion progress over time relative to the total amount present in the fabricated sample.While the degradation time of wafer-bound structures will differ from the corresponding time for freely suspended structures, due to geometric effects and the proximity of the substrate, the substrate-based study was nevertheless chosen, as it allowed for more controlled observation conditions, and since it does not involve sonication steps.The ratio of Mg to Zn in the respective samples was determined to be 81%, 66%, and 48% (all determined with an accuracy of ≈±2%), which agrees very well with QCM measurements during the fabrication.This shows that the ratio of Mg and Zn can be precisely set during fabrication.In Figure 3a, the cumulative amounts of dissolved Mg 2+ and Zn 2+ show the overall degradation progress for each sample.As expected, the corrosion rate correlates with the amount of Mg in the structure: the higher the Mg content, the faster the structure degrades.Unlike the observations made in solution, however, the decomposition rates are somewhat lower, which might be due to the film geometry and slightly lower solution temperature, where the structures are less exposed to the solution compared to the freely suspended structures.The latter will in addition also experience mechanical sample damage, for instance during sonication, which would increase the surface area of the structures and therefore enhances the corrosion rates.
Analyzing the individual ion concentrations of Mg and Zn, shown for the MgZn 80:20 sample in Figure 3b, reveals that Mg is transferred to the solution at a higher rate compared to Zn in all samples.We therefore hypothesize that the transfer of Mg is mainly responsible for the degradation process ultimately leading to the collapse of the propeller structure.Zn does dissolve in aqueous media over longer periods, but not completely in the time of the experiment for all samples.After 11 days only ≈5-20% of Zn content initially present in the structures could be found in solution.This is consistent with the previously observed slow and incomplete decomposition of larger Zn structures in the literature, where a tube of 2 mm outer diameter and 0.15 mm wall thickness immersed in Hank's buffer solution showed a corrosion rate of ≈0.028-0.037mm per year compared to 0.61-0.73mm per year for Mg. [23]lectron micrographs of the submerged samples after 11 days were taken to further characterize the structural integrity.Surprisingly, we found structures still strongly resembling the initial helices on the substrate, as shown in Figure 3c,d.As seen in Figure 3a,b for MgZn ratios 70:30 and 80:20, aproximately bettween 20% and 40% of the total content is not dissolved after 11 days.This results in a porous, "hollowed out" helical structure, where most of the Mg is dissolved out, but most of the Zn remains.The degradation behavior depends on a multitude of factors, such as fluid flow, salt concentrations, temperature, which we cannot all test in this work, but we made a first attempt by repeating the same experiment and replacing water with Dulbecco's Modified Eagle medium (DMEM) for a more physiologically-relevant environment.The results are shown in Figure S4 (Supporting Information).The salt content in the medium influences the corrosion behavior, which would have to be accounted for when designing micropropellers for such environments.The overall time scale of degradation is still in the range required for applications.

Biocompatibility
To investigate the biocompatibility of our MgZn propellers during decomposition, toxicity assays were performed.Cell viability was assesed with an MTS-assay, and the cell density was determined with crystal violet staining and apoptosis induction with a caspase 3/7-assay.For the first two assays, negative ef-fects would cause a decrease in the analysis, whereas the induction of apoptosis would cause an increase of caspase 3/7-activity.Clinically relevant ocular cells (the human cell lines: ARPE-19 and MIO-M1, the first representing retinal pigment epothelium cells, the second Müller cells, as well as primary porcine retinal pigment epithelium cells) were treated, as they represent possible target sites, with high cell-to-propeller ratios of up to 1:20 for 48 h.The results are shown in Figure 4, where the cell viability and density remained unchanged for all tested cells.For ARPE-19 and MIO-M1 cells the caspase 3/7 activity was slightly reduced after 48 h for higher propeller concentrations.In short, all assays confirmed a high biocompability in all tested cell lines.

Functionalization
Fluorescent liposomes with outward facing amine functionalities were bound to the MgZn propellers via silane and Nhydroxysuccinimide (NHS) coupling chemistry.The surface of the propellers is oxidized when exposed to air and humidity, forming zinc and magnesium oxides and hydroxides, all of which are practically insoluble in water, [72] forming a stable foundation capable of condensation reactions with ethoxysilane groups.Using silane-PEG-NHS generates a reactive NHS-ester species, which forms amide bonds when exposed to amine groups located on the liposomes.This process creates a covalent link between the liposomes and the micropropellers.The fluorescent labels of the liposomes are used for co-localization with the microcarriers.
To determine if the carrier loading was successful, a sample of micropropellers, chemically functionalized with fluorescent liposomes, was placed in a closed glass chamber.A permanent magnet was used to manipulate the sample and to sediment the magnetic propellers to the bottom of the chamber, as confirmed through microscopy under brightfield illumination.Subsequently, this same plane and the remaining chamber volume was imaged via fluorescence microscopy, to determine the position of the liposomes.An overlay of the two channels provides spatial confirmation that the liposomes are coupled to the magnetic micropropellers (Figure 5a).A control sample without the use of silane-PEG-NHS led to little to no co-localization, as shown in Figure S5 (Supporting Information), indicating that unspecific, charge-based interactions are not enough to facilitate reliable functionalization.
To investigate, if the degradation of the microswimmers leads to a release of the liposomes, another chamber was prepared of the same sample after 24 h and imaged in the same way.The resulting overlay image, depicted in Figure 5b, shows that the microswimmers and liposomes are no longer co-localized.Freelyfloating liposomes could be found in the entire volume of the imaging chamber, while the propellers remain sedimented at the bottom of the observation chamber.The time scale of liposome release was determined by imaging the MgZn 70:30 and 80:20 samples in the same manner 0, 2, 6, and 24 h after functionalization in PBS.The number of functionalized particles (n > 100) at each time point was counted and is shown in Figure S6 (Supporting Information).The MgZn ratio affects the liposome release rate only slightly in the observed range, but follows the expected trend that a slower degradation leads to a slower release.

Propulsion
Characterization of the propulsion capabilities of helical microswimmers fabricated by GLAD using FePt has been investigated before. [8,73]However, the new material composition (MgZn) influences the final helical shape that can be obtained, and the coupled liposomes may also affect the hydrodynamics given the small size of the micropropellers we have fabricated.Further, the propulsion serves as an additional control concerning structural integrity as the propellers are designed to degrade over time.Using a 3-axis Helmholtz coil fitted inside an inverted fluorescence microscope, the propulsion behavior of the propellers was characterized while they are actuated by a 3D rotating magnetic field (1-5 mT).Both brightfield and fluorescence imaging modes are recorded.The analysis of the motion capabilities is shown in Figure 5c, where a propeller was actuated to swim on a pre-programmed L-shaped trajectory with a magnetic field of 1 mT and a rotation frequency of 100 Hz.Both recordings were taken consecutively of the same propeller, while switching the illumination source.Analysis of the propulsion behavior and particle tracking give speeds of ≈8 μm s −1 or ≈7 body lenghts s −1 , which is in agreement with previous studies on similar micropropellers. [8,73]For the utilization in transfection experiments, the velocity and steerability of the propellers need to persist over time.An additional observation after 28 h of incubation in PBS buffer of MgZn 70:30 propellers, navigating a manually controlled L-shape, was found to exhibit similar speeds and characteristics and thus confirm that the structures are functional over pre-determined time scales (Figure S7, Supporting Information).The fluorescence intensity was observed to be lower than before, suggesting that some liposomes had already detached from the propellers.Nonetheless, the available time is long enough to functionalize the micropropellers and perform experiments and is thus in the time span expected to be useful for applications.
While the application of rotating magnetic fields for propulsion has many advantages, primarily as it permits the use of weak fields, magnetic manipulations based on gradient fields are also interesting to investigate.The latter for instance finds application in magnetofection experiments.We show that the structures can also be manipulated using gradient magnetic fields, as expected.Here, a disc-shaped permanent magnet was mounted on a separate stage and moved toward the sample up to a distance of ≈10 mm to pull the micropropellers with a field gradient that is estimated to be ≈1 T m −1 and a speed of ≈5 μm s −1 in the direction, which is indicated by the position of the magnet in Figure 5d.This could be observed under brightfield and fluorescence illumination, as shown previously.After the magnet was removed the particles resumed their regular Brownian motion.

Conclusion
The successful fabrication of biocompatible and degradable MgZn helical micropropellers has been demonstrated with precise control over shape, dimensions, and composition.The microstructures can be grown in large quantities of several billion structures per deposition even on a two inch wafer substrate.Shadow-growth and cooling allowed for the realization of uniform complex shapes, despite high adatom mobilities.The corrosion rate could be tuned between a few hours and several weeks by varying the Mg:Zn ratio.The degradation behavior depends on multiple factors, including the ion concentration of the surrounding medium, local pH values, and flow conditions, and this can be accommodated by varying the Zn content from under 10% to 50%.The MgZn material system that is successfully used for degradable large-scale medical implants can thus also be utilized for applications in nanomedicine.
For propulsion, a thin layer of the hard-magnetic, biocompatible FePt in its L1 0 phase was incorporated, which is stable and could be actuated with weak magnetic fields in the low mT range including after immersion in PBS buffer for 24 h.The amount of magnetic material for propulsion can be altered for different applications, including high viscosity environments, where greater torque (magnetic moment) is required.The propulsion mechanism based on rotating magnetic fields is notably different from those in magnetic transfection (magnetofection) experiments where a magnetic field gradient "pulls" magnetic, often superparamagnetic iron oxide nanoparticles (SPION), toward a cell culture.We note that the microcarriers we describe here can also be pulled by gradient fields, but show much stronger magnetic moments than SPIONS.The fabrication method we describe is general, and other materials may in addition be included.For instance, it is possible to include plasmonic materials, having potential uses in heating and spectroscopy. [74]inally, we also show that the surface of the MgZn structures can be functionalized with liposomes, rendering the structures microcarriers that because of their degradation in water allow for a time-dependent release of their cargo.This suggests a powerful platform for targeted drug or gene delivery that can be integrated with established systems for magnetic actuation and transfection.Crucially, the propeller decomposition did not negatively affect cell viability of clinically relevant primary retinal pigment epithelium (RPE) cells, even at propeller-to-cell ratios as high as 20:1.While this design is shown to work in an aqueous buffer in vitro, the next step will be to test and adapt to in vivo conditions.Here, it is necessary to address the challenges presented by the in vivo environment, such as the propulsion in highly viscous media, tissue penetration, non-specific interactions, e.g., protein coronas, and fluid flows.To this end, our study presents a first step toward enabling the biomedical application of inorganic helical magnetic micropropellers that can be fabricated with biocompatible and -as we show here -degradable MgZn body.This material system is of interest as it suggests a solution to concerns regarding the long-term biosafety in the use of non-viral vectors for potential applications in gene therapy.Pursuing their application could offer a unique treatment outlook for diseases in hard-to-reach tissues that require precision delivery.Biocompatible, degradable MgZn structures for nanomedicine and in particular, magnetically actuated liposome microcarriers, offer the opportunity for targeted and programmed release and could thus minimize the need for certain invasive procedures.

Experimental Section
GLAD: For the fabrication of the helical micropropellers, a Si(100) substrate (Silicon Materials) was patterned with 1 μm polystyrene beads (Polysciences) in a hexagonal close-packed monolayer by Langmuir-Blodgett deposition, in which the beads were carefully placed on the water/air interface in a Langmuir-Blodgett trough.A substrate submerged in the trough was slowly raised and the beads transfer and to the substrate.
The beads were then reduced in diameter to an average of 560 nm by O 2 plasma treatment for 8 min at 300 W and 0.7 mbar, which also increased the spacing between the individual beads.PS beads etched with an O 2 plasma remain circular in the plane of the substrate (Figure S8, Supporting Information), while they become anisotropic orthogonal to the substrate surface plane as had been previously discussed by Lotito and Zambeline. [75]The substrate was placed in a HV-chamber with a motorized substrate manipulator for glancing angle deposition (GLAD).Initially, a 10 nm MgO (Santa Cruz Biotechnology Inc., >90%) adhesion layer was deposited on the polystyrene, and followed by the co-deposition of 40 nm Fe (Kurt J. Lesker Company Limited, 99.95%) and Pt (Kurt J. Lesker Company Limited, 99.99%) at an atomic ratio of 1:1.The deposition rate during the deposition process was observed via quartz crystal microbalances, that were calibrated for the respective deposited material.The rates in the co-deposition were adjusted continuously during the deposition to realize the target ratios.The FePt layer was then followed by a 10 nm Ti (Kurt J. Lesker Company Limited, 99.995%) adhesion layer.All depositions were at an angle of 65°.The sample was then temporarily removed from the vacuum chamber and annealed at 700 °C for 1 h to induce the formation of the L1 0 phase in FePt.The high temperature and a subsequent air plasma treatment for 10 min at 100 W and 0.4 mbar ensured the removal of the polystyrene.Given the PS depolymerization processes start at ≈390 °C it was expected that the PS was removed during the annealing step at 700 °C for 1 h.Additionally, a subsequent O 2 plasma etching step was applied to fully remove any possible residues of pyrolysis products. [76]The sample was then re-introduced to the vacuum chamber for the co-deposition of Mg (Kurt J. Lesker Company Limited, 99.95%) and Zn (Thermo Scientific, 99.999%) at the desired atomic ratio under substrate rotation to form lefthanded helices with 1.3 turns.The Mg content was adjusted between 90, 80, 70, or 50 atomic% (measured by QCM) for different samples and deposition runs.The Zn content was accordingly 10, 20, 30, or 50%.The Mg and Zn was deposited at an incident angle of 87°and at ≈90 K to facilitate the growth of the helical propeller structures.The properties of the microstructures could be adjusted by changing tilt angle, rotation speed and deposition length.The pitch was determined by the rotation speed and the width by the bead size.Furthermore, the length of the helix will depend on the amount of material deposited, while the diameter was adjusted by the tilt angle.
SQUID Measurements: The in-plane magnetization was determined using a Quantum Design MPMS3 SQUID at 300 K.A full hysteresis loop in the range of 3 T to −3 T at a rate of 10 mT s −1 was measured and can be seen in Figure S2 (Supporting Information).
Inductive Coupled Plasma-Optical Emission Spectroscopy (ICP-OES): The ICP-OES measurements were performed using a SPECTRO CIROS.The water used for all experiments was filtered and ion exchanged with a resistivity of 18 Ohm at 25 °C, unless stated otherwise.The fabricated structures were immersed in water and the supernatant was introduced into the inductively generated argon plasma via an atomizer system and excited to determine the content of Mg and Zn.A calibration was carried out beforehand on the same day of the measurement with a multi-element solution mixed from standard solutions.The intensities of measured spectral lines were compared to previously measured intensities of known concentrations for all studied elements.
Cell Culture: The human Müller cell line (MIO-M1) was obtained from the UCL Institute of Ophthalmology, London, UK. [77] ARPE-19 cells were purchased commercially from American Type Culture Collection, Manassas, VA, USA. [78]Since primary cells were more sensitive than immortalized cell lines, primary RPE cells were isolated from the porcine retina as previously described. [79]ropellers were dispersed directly in cell culture medium by ultrasonication of the substrate.This mixture was then vortexed and was serially diluted immediately to obtain 1:5; 1:10 and 1:20 ratio (cell:propeller) containing culture medium.Cells were incubated for 48 h.Untreated cells served as a control.
Cell Viability Assay: A total of 15000 MIO-M1 cells, ARPE-19 cells and 20000 primary porcine RPE cells per well were seeded the day before the experiment in transparent 96-well plates (Merck).CellTiter 96 Aqueous One Solution Cell Proliferation Assay (MTS, Promega) was used according to the manufacturer's protocol.Twenty microliters of One Solution Reagent was directly added to each culture well.MioM1, ARPE19 and primary RPE cells were incubated for 90 min.Absorbance was measured (490/690 nm ratio) using a NanoQuant infinite M200 Reader (Tecan Group AG).The results were presented in % in relation to the control (=100 %). [80]rystal Violet Staining: To determine cell density, some of the cells in 4.5 were fixed with 4% paraformaldehyde (Merck) for 15 min, washed, and incubated with crystal violet solution (Sigma-Aldrich).After a 30 min incubation period, cells were carefully washed until the water remained clear; then, 1% sodium dodecyl sulfate (Applichem) was added before further incubating for 1 h at RT. Absorbance was then measured using NanoQuant infinite M200 Reader at 595 nm.The results were presented in % in relation to the control (=100 %). [80]aspase 3/7 Activity Assay: Cells were seeded as described in section Cell Viability assay.After incubation with micropropellers, the caspase 3/7activity was determined using the CaspaseGlo 3/7 activity kit (Promega) according to the manufacturer's protocol: 100 μl of the CaspaseGlo 3/7 reagent was added directly to each well (96-well plate).After 60 min of incubation at room temperature, luminescence was measured with a luminescence reader.For this assay white 96-well plates with a clear bottom (Merck) were used.The amount of luminescence was proportional to the amount of caspase activity in the sample.The results were presented in relative light units (RLUs). [80]luorescent Liposome Synthesis: Fluorescent liposomes were synthesized using the ethanol injection method.[81] A lipid mixture of hydrogenated soybean phosphatidylcholine (HSPC, Lipoid, MW 762 g mol −1 ), cholesterol (Sigma-Aldrich, MW 386.65 g mol −1 ), 1,2-distearoyl-snglycero-3-phosphoethanolamine-N-methoxy-polyethylene glycol 2000 (DSPE-PEG-2000-NH2, Biochempeg, MW 2790 g mol −1 ), and DSPE-Cyanine 5 (synthesized as described previously, [64] MW 1213 g mol −1 ), in molar ratios of 54.6:40:5:0.4was dissolved in absolute ethanol at 65 °C.Once all the lipids were completely dissolved, the lipid suspension was added into PBS (Sigma-Aldrich) buffer solution to a final concentration of 100 mm total lipids. To otain homogenous 100-nm liposomes, the liposome mixture was extruded through 400-, 200-, 100-, and 80-nm pore-size polycarbonate membranes (Whatman) using a LIPEX extruder (Northern Lipids) above the melting temperature of 65°.In order to eliminate the ethanol and residues from the synthesis, the liposome solution was dialyzed overnight against PBS (1:1000 volume ratio) at 4 °C using a 12-to 14-kDa dialysis membrane (Spectrum Laboratories Inc.).The average hydrodynamic diameter of the liposomes was determined to 95.7 ± 3.8 nm with a polydispersity index (PDI) of 0.081 ± 0.034, and a liposome concentration of 1.16E + 14 particles per ml using a Zetasizer Ultra (Malvern Panalytical).
Liposome Functionalization: The propellers were activated on the substrate by an air plasma treatment for 5 min at 100 W and 0.4 mbar and then functionalized with Silane-PEG-SC (Biopharma PEG) in a solution of 1:1 ethanol to water mixture in a plastic container for at least 30 min.The substrate was subsequently rinsed with water and immersed in a PBS buffer.The silane-functionalized propellers were removed from the substrate by short sonication for a few seconds.The liposomes were added at a 10 4 :1 ratio to the propeller dispersion and incubated for 2 h.The unused solution was covered with aluminum foil to limit its exposure to light.Excess liposomes were removed by three washing steps, using gentle centrifugation (30 rcf, 10 min) and re-dispersion in PBS buffer.
Propulsion through Rotating Magnetic Fields: Before the propulsion experiments, the micropropellers were magnetized in-plane by an electromagnet (Walker Scientific Inc.) at an external field of 1.6 T for 5 min.
The custom-built 3-axis Helmholtz coils were operated at frequencies between 50-100 Hz and with magnetic field strengths of 1-5 mT.For the actuation along the pre-programmed L-shape trajectory (Figure 5c), each swimming direction was held for 5 s.Due to the lower brightness of the propeller after 24 h in solution, the propulsion times were set manually during the L-shape trajectory to improve downstream tracking analysis (Figure S4, Supporting Information).
Manipulation with Permanent Magnet: A disc-shaped permanent magnet (NdFeB, Type N45, Supermagnete) with a diameter of 25 mm and a height of 3 mm was brought within ≈10 mm of the sample.The magnetic gradient in that range was measured to be ≈1 T m −1 .Due to geometric restrictions, the magnet was positioned slightly above the sample, resulting in a slight reduction of the in-plane gradient field and an additional small upward pull.
Image Analysis: For microscopy imaging a Zeiss Observer Z1 inverted microscope was used.The particle dispersion was enclosed in a 10 × 10 mm Gene Frame, sandwiched between two untreated 24 × 24 mm coverslips.To minimize adhesion, the glass slides were pre-treated with Bovine Serum Albumin (Carl Roth).Videos S1 and S2 (Supporting Information) were recorded using the Andor SOLIS software with an Andor Zyla 4.2P sCMOS camera and a 63X objective (Carl Zeiss AG) with 0.85 NA.Recordings were usually performed at 10 frames per second (2 frames per second for Figure S4, Supporting Information) during fluorescence imaging and 30 frames per second under brightfield illumination.The videos were analyzed using Fiji [82] and tracked over a range of at least 200 frames via the Mosaic Suite [83] plugin utilizing inverted brightfield videos to obtain the depicted trajectories and particle movement.
SEM: Scanning electron microscope images were taken on a Zeiss Gemini Ultra 55, operating at 5 kV acceleration voltage and using InLens and SE2 detectors.

Figure 1 .
Figure 1.Schematic overview of a) the fabrication process for different shapes using specialized physical vapor deposition.b)The ratio of Mg to Zn determines the degradation rate, while incorporation of magnetic material allows for c) magnetic manipulation and propulsion.d) Surface functionalization allows for the covalent attachment of cargo molecules, which are e released during the degradation process.Created with BioRender.com.

Figure 3 .
Figure 3. Degradation time series of MgZn propellers in H 2 O for different ratios of Mg and Zn, observed with ICP-OES.The corresponding solid lines serve as guides to the eye.a) Measurement of dissolved Mg and Zn, for each ratio, normalized to the total amount of Mg and Zn (dissolved and undissolved) in each sample.b) Measurement of dissolved Mg and Zn over time, separated by ion type, and normalized by the total ion content.Side view electron micrographs of MgZn.The insets indicate the Mg:Zn ratios in the same order of the curves (top to bottom).c) 70:30 and d) 80:20 samples on the substrate after 11 d in H 2 O. Scale bars are 500 nm.

Figure 4 .
Figure 4. Incubation of ocular cells with MgZn 80:20 micropropellers in ratios of 1:5, 1:10, and 1:20 for 48 h.Bar graphs represent the mean values and standard error of mean, n = 8.The experiments were repeated three times with similar results.Statistical differences are indicated as follows: *p < 0.05, and ***p < 0.001 compared to control, according to one-way ANOVA.

Figure 5 .
Figure 5. Inverted brightfield (blue) and fluorescence (red) image overlays with zoomed-in inserts of MgZn 70:30 micropropellers on a glass slide.Image a) directly after functionalization and b) after 24 h in PBS solution.Scale bars are 25 μm.c) Tracking of a Mg:Zn 70:30 micropropeller functionalized with liposomes following a predetermined L-shaped trajectory, observed in brightfield and fluorescence microscopy.Scale bars are 10 μm.d) Tracking of Mg:Zn 70:30 micropropellers pulled by a permanent magnet and schematic illustration of the experiment.Scale bar is 50 μm.Created with BioRender.com.