Fabrication of anisotropic magnetic helical microswimmers utilizing Spirulina platensis templates and their integration with Janus PCL/Chitosan nanoparticles

Abstract

Magnetically-actuated microswimmers are promising candidates for controlled cargo transport and microscale manipulation; however, achieving efficient, fuel-free propulsion with precisely defined anisotropic structures remains a significant fabrication challenge. We present a novel bio-hybrid helical microswimmer architecture that utilizes the bio-template of Spirulina platensis. This design incorporates anisotropic magnetite Janus nanoparticle (MJNP) heads made from hydrophilic chitosan and hydrophobic polycaprolactone (PCL). This architecture is created through a two-step dip-coating and anisotropic linking process, which ensures stability, magnetic responsiveness, and a defined structural asymmetry. To assess therapeutic potential, Doxorubicin (DOX) was loaded onto the MJNPs, and the biocompatibility of the microswimmers was confirmed in vitro. The microswimmers exhibited efficient corkscrew-like propulsion when exposed to rotating magnetic fields. By systematically tracking their trajectories in biofluids, we observed that increasing the number of helical turns significantly enhances both their forward velocity and translational diffusion. This directly correlates the length of the helices with the optimized conversion of rotational torque into translational motion. This study establishes a robust, bio-templated platform that provides valuable design insights for enhancing the functional versatility of helical microswimmers in advanced biomedical applications.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36118-9.

Introduction

In recent years, a wide range of micro- and nanomotors (MNMs) have been developed to function under both fuel-based and fuel-free conditions. Among these, magnetically actuated MNMs have attracted significant attention due to their ability to be wirelessly controlled in terms of motion, direction, and speed, while also ensuring safe interaction with living cells and tissues^1–3. Within this category, helical MNMs (H-MNMs), inspired by the flagellar propulsion of Escherichia coli, have been designed to achieve efficient corkscrew-like motion. This motion is generated by the rotation of their helical tails when exposed to a rotating magnetic field^4–6. Owing to their torque-driven locomotion, these magnetic MNMs require relatively low magnetic field strengths for precise maneuverability, making them safer and more suitable for biomedical and therapeutic applications⁵.

Various fabrication strategies have been proposed for creating helical microswimmers, including self-scrolling⁷, 3D direct laser writing (DLW)^8–10, glancing angle deposition (GLAD)¹¹, electrochemical processing¹², and bio-templated synthesis^13,14. Among these methods, bio-templated techniques present a particularly cost-effective and biocompatible alternative, which is crucial for medical use¹³. A prominent example of this is the cyanobacterium Arthrospira platensis (commonly known as Spirulina), a commercially available microorganism with a naturally intact, three-dimensional helical structure.

The application of Spirulina in robotics has evolved significantly, transitioning from static bio-templates to multifunctional active swimmers. Initially, the microalgae served primarily used as templates for synthesizing functional microstructures, such as TiO₂ or metal microcoils for photocatalytic and electromagnetic applications^15–17. A significant shift occurred in 2015 when Yan et al. transformed these templates into active magnetic swimmers, laying the groundwork for targeted delivery¹⁸. Since then, functionalization strategies have advanced rapidly. Researchers have coated Spirulina with nickel or hybrid nanoparticles to enhance propulsion velocity and enable synergistic chemo-photothermal therapies^19–21. Recently, the field has expanded into complex biohybrid systems capable of wireless stem cell stimulation²², tumor hypoxia alleviation via photosynthesis²³, and advanced imaging-guided therapies using photoacoustic or DSA modalities^24,25.

Despite these advancements, critical limitations remain. Notably, there is a lack of a defined anisotropic head structure that could enhance payload capacity and allow for selective surface functionalization, such as controlled wettability or drug loading. Moreover, the influence of the Spirulina’s natural helical geometry - specifically how variations in the number of helical turns affect propulsion efficiency - has not been systematically investigated. Understanding this complex interplay between external magnetic field parameters, including field strength⁵ and rotational frequency^26,27, alongside the geometrical properties of the H-MNMs - such as length, diameter, helical pitch, helicity angle, and chirality^28,29- is crucial for optimizing motion. This investigation will be the primary focus of this work.

In summary, this research aims to address some current limitations in the design of bio-templated microswimmer by establishing a robust platform that combines the natural helical motility of Spirulina with chemically-defined anisotropic Janus heads. We will first describe the fabrication process and thoroughly characterize the novel bio-hybrid microswimmer architecture. Subsequently, we load Doxorubicin onto magnetite Janus nanoparticles and assess the biocompatibility of the microswimmers. Furthermore, we systematically investigate how the intrinsic geometrical properties of Spirulina, specifically the number of helical turns, influence propulsion dynamics and diffusion behavior under diverse magnetic field conditions in water and biofluids. Our findings will provide valuable design principles for optimizing propulsion efficiency in bio-templated helical microswimmers, ultimately facilitating their application in precise in vivo sensing and drug delivery. The fabricated microswimmers and prepared nanoparticles were characterized using various techniques, including optical and fluorescence microscopy, scanning electron microscopy (SEM), transmission electron microscopy (TEM), proton nuclear magnetic resonance (¹H NMR), Fourier-transform infrared spectroscopy (FT-IR), thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), dynamic light scattering (DLS), X-ray diffraction (XRD), vibrating-sample magnetometer (VSM), and ultraviolet-visible spectroscopy (UV-Vis). Furthermore, their movements in four fluids were analyzed across three sizes and three propulsion modes under the influence of a magnetic field.

Experimental section

Materials

Iron(II) chloride tetrahydrate, iron(III) chloride hexahydrate, ammonium hydroxide (25% NH₃ in water), (3-aminopropyl)triethoxysilane (APTES), succinic anhydride, triethylamine (TEA), N, N′-dicyclohexylcarbodiimide (DCC), 4-dimethylaminopyridine (DMAP), ε-Caprolactone, chitosan (medium molecular weight), tin(II) 2-ethylhexanoate, sodium hydroxide, dimethylformamide (DMF), hydrochloric acid (37%), ethanol, methanol, water (distilled and deionized (DI)), dichloromethane (DCM), N-hydroxysuccinimide (NHS), methyl acrylate (MA), 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC), fluorescein, acridine orange, dimethyl sulfoxide (DMSO), cetrimonium bromide (CTAB), tetraethyl orthosilicate (TEOS), ammonium nitrate, poly(vinyl alcohol) (PVA), were purchased from Merck Chemical Co. Bovine serum albumin (BSA), human serum albumin (HAS), fetal bovine serum (FBS), 3-(4,5-dimethylth iazol-2-yl)−2,5-diphenyltetrazolium bromide (MTT), and dialysis bag with the cut-off of 12 kDa were purchase from Sigma-Aldrich. Dichloromethane was obtained from Saba Company (Iran), and was distilled before use. DMF was dried using barium oxide. Spirulina and Zarrouk’s standard media were obtained from the Iranian National Algae Culture Collection (INACC).

Synthesis of anisotropic magnetite microhelix (An-SP@SiO₂)

Synthesis of the head of the anisotropic magnetite microhelix (Si-MJNPs)

The head of the anisotropic microhelix was prepared via a three-step method. First, magnetite Janus nanoparticles (MJNPs) based on poly(ε-caprolactone) and chitosan were synthesized according to a previously reported procedure with some modifications (see section S1.1)³⁰.

The second step involved the methylacrylation of the amine groups of chitosan, followed by modification with TEOS. To achieve this, 60 mg of MJNPs were dispersed in 5 mL of methanol using an ultrasonic probe for 5 min. The mixture’s temperature was raised to 50 °C while stirring with a magnetic stirrer and purging with N₂. Then, 0.5 mL of methyl acrylate (5.55 mmol) was added, and the mixture was stirred for 48 h. The resulting ester-decorated Janus nanoparticles (MA-MJNPs) were magnetically separated, washed three times with methanol to remove any unbounded materials, and then dried overnight in an oven at 60 °C (The reaction yield was 82%).

Finally, 13 mg of MA-MJNPs were dispersed in anhydrous DMF (15 mL) under sonication for 10 min. While stirring, 2 µL of APTES (8.55 µmol) was added to the reaction mixture and stirred for 72 h. Si-MJNPs (with 85% yield) were magnetically separated and dried in an oven at 40 °C.

Synthesis of the tail of the anisotropic magnetite microhelix

• Cultivation of spirulina - Microalgae were cultivated in 1 L Erlenmeyer flasks, each containing 250 mL of culture medium and 10 mL of stock spirulina sample. The cultivation was carried out with a photoperiod of 12 h light followed by 12 h of darkness, using a lamp that provided a light intensity of 60 µmol photons.m²/s. The temperature was maintained at 26 ± 1 °C. The cultivation process started after autoclaving for 20 min at 121 °C and continued for 12 days while shaking at 120 rpm. At the end of the cultivation, the biomass was centrifuged at 3500 rpm for 15 min and then dried in an oven at 40 °C for 24 h³¹.

• Fabrication of magnetite-coated microhelices (SP@Fe ₃ O 4 ) – IONPs were prepared via co-precipitation method (see section S1.1). Lastly, 100 µL of HCl solution (5 M) was added, followed by another 10 min of sonication, to yield positive-charged IONP suspensions (ca. pH 2.1) with a 86.4% yield¹. For deposition, Spirulina samples were washed three times with DI water, to remove non-deposited particles, and sonicated for 5 s in DI water (5 mL) before being added to the IONPs suspension. 1 g of sonicated Spirulina was mixed with 50 mL of the IONP suspension (containing ⁓ 40 mg, 0.17 mmol), and shaken on an orbital shaker (120 rpm) for 66 h to prevent agglomeration The final products were separated by centrifugation, washed three times with DI water, or until the suspension became clear, and stored in DI water¹. The efficiency of this reaction was 91%.

• Coating of SP@Fe 3O ₄ with silica layer (SP@SiO2) − 20 mg of SP@Fe₃O₄ was added to a solution of 32.8 mg of CTAB in 10 mL distilled water (9 mM). After 30 min of mixing by mechanical stirring (300 rpm), 0.15 ml of NaOH solution (0.1 M) was slowly added to the mixture to adjust the pH to 10–11. The reaction system was transferred to an oil bath set at 60 °C. Then, 1 mL of TEOS/ethanol (50 µL/950 µL) was added slowly at a drip rate of 330 µL/30 min, and the mixture was stirred at 60 °C for 24 h. The crude product was collected by a magnet and then washed several times with water and ethanol. Finally, to eliminate the CTAB, the crude product underwent reflux at 80 °C for 24 h in 50 mL of ethanol containing 300 mg of ammonium nitrate. The final product was then washed with ethanol, to remove unreacted precursors, and preserved in it. The overall yield of synthesized SP@SiO₂ was ~ 85%.

Connecting the tail of the microswimmer to its head

One end of the isotropic SP@SiO₂ microhelices was immobilized vertically within a PVA film, allowing only one end of these helical microstructures to protrude from the film. To prepare An-SP@SiO₂, 1 mg of the SP@SiO₂ was dispersed in a solution of 0.25 mg PVA in 5 mL of H₂O, placed on a substrate under a magnetic field, and allowed to dry. Then, the resulting film was immersed in 5 mL of ethanol under a nitrogen atmosphere while magnetically stirring, and 2 mg of Si-MJNP was added. After a reaction period of 6 h, the film was separated and washed with water to yield the final structures.

Characterization techniques

The morphology of the resulting microstructures and nanoparticles was characterized using a scanning electron microscope (SEM, Hitachi S-4160, Japan) and transmission electron microscopy (TEM, Zeiss LEO 906, Germany). Energy Dispersive X-ray (EDX) analysis was performed using a TESCAN MIRA3 SEM at an accelerating voltage of 20 kV and a beam current of 10 nA.

The ¹H NMR spectra were recorded in CDCl₃ using a Bruker DRX 500 spectrometer (500 MHz, Germany) to determine the number average molecular weight (Mn) of PCL. FT-IR measurements were conducted using a Thermo Scientific Nicolet 800 FT-IR spectrometer. The hydrodynamic size and zeta potential of the particles were measured using Dynamic Light Scattering (DLS, Horiba SZ-100) in an aqueous solution at 25 °C.

The surface polymer content on the magnetic nanoparticles was evaluated by thermogravimetric analysis (TGA) using a Q50 V6.3 Build 189 instrument, where samples were heated at a rate of 20 °C/min under a nitrogen atmosphere, and data were collected from 30 to 600 °C. Differential Scanning Calorimetry (DSC) analysis was carried out to assess the thermal properties using a Pishtaz Engineering Co. Model TA-2 instrument, with samples heated at a rate of 10 °C/min under an inert nitrogen atmosphere. X-ray powder diffraction (XRD) was conducted using a DX-2700BH diffractometer (40 kV, 30 mA) with the monochromator engaged. Samples were scanned over a 2θ range of 5° to 80° using a slit configuration of 1°−1°−0.2 mm. The magnetization of nanoparticles was measured using vibrating sample magnetometer at room temperature (VSM, MDKB model. Magnetic Daneshpajoh Co. Applied field ranged from − 14000 mT to 14000 mT). A SONOPULS ultrasonic homogenizer (20 kHz ± 500 Hz ultrasonic generator), model HF-GM 3200 (BANDELIN Electronic GmbH & Co. KG), equipped with a titanium microtip MS-72 probe (3 mm diameter), was used to disperse the particles. Sample drying was performed using a Christ Alpha 1–2 LSCbasic freeze dryer (Germany), with a condenser temperature of −55 °C, enabling the drying of aqueous samples.

Fluorescent labeling

A Zeiss light microscope equipped with fluorescence capabilities was used to analyze the morphology of the Janus nanoparticles and to observed the distribution of loaded dyes. For this purpose, 1 mg of fluorescein (a hydrophobic dye), 1 mg of acridine orange (a hydrophilic dye), and 2 drops of DMSO were added to 1 mL of acetone. Separately, 3 mg of Janus nanoparticles were dispersed in 1 mL of acetone, and this suspension was then added to the dye solution and stirred at 400 rpm for 1 h. The resulting mixture was added dropwise into deionized water, and the organic solvent was removed under reduced pressure. A few drops of the final suspension were transferred onto a slide and allowed to dry.

Preparation of drug-loaded nanoparticles

• Loading of Doxorubicin on MJNPs - Doxorubicin-loaded magnetite Janus nanoparticles were prepared using the nanoprecipitation method. First, 10 mg of MJNPs was dispersed in 4 mL of phosphate buffer saline solution (PBS) for 10 min. Subsequently, 1 mg of Doxorubicin, was added to the mixture, which was then stirred at 300 rpm for 24 h in a light-protected vial at room temperature. Then, DOX-loaded MJNPs were separated via magnetic decantation. The prepared nanoparticles were washed with deionized water once to remove the unloaded drug and then remove from aqueous medium using freeze-drying.

• Measurement of drug loading content and encapsulation efficiency of MJNPs - The drug loading content (DLC) was quantified by measuring the absorbance value of unloaded Doxorubicin in the supernatant with a UV-spectrophotometer (Jasco Inc V-750, Japan) at its maximum absorption wavelength (λ_max = 481 nm). To determine the encapsulation efficiency (EE), a predetermined amount of freeze-dried DOX-loaded nanoparticles was dispersed in 4 mL of PBS. The absorbance of Doxorubicin was determined at 481 nm. Quantification of Doxorubicin was performed by generating a calibration curve with known drug concentrations. Based on the measurements from this curve, DLC and EE were calculated using Eqs. 1 and 2:

1

2

In vitro drug release assay

The drug release behavior of the final nanoparticles was assessed in PBS at pH 7.4 and pH 5.8, using a diffusion technique. Initially, 1.0 mg of MJNPs was placed in a dialysis bag (with a molecular weight cut-off of 12 kDa) containing 0.5 mL PBS, and was submerged in 10 mL of PBS at 37 °C under mild stirring conditions (100 rpm). At regular intervals ranging from a few hours to a few weeks, 1 mL of the release medium was sampled and replaced with 1 mL of fresh buffer. The concentration of DOX in the medium was measured using the UV-Vis spectrophotometer at 481 nm, and the drug release percentage was determined by Eq. 3:

3

In vitro cytotoxicity assay

The cytotoxicity of Doxorubicin, MJNP, and MJNP/Doxorubicin against B16-F10 melanoma cells was evaluated using the conventional 3-(4,5-dimethylthiazol-2-yl)−2,5-diphenyltetrazolium bromide (MTT) assay. Cells were seeded at a density of 8000 cells per well in 96-well plates containing 100 µL of complete medium. After incubating for 24 h at 37 °C in a humidified atmosphere with 5% CO₂, 100 µL of nanoparticle suspension was added to achieve the desired final concentrations. Following another 24 h incubation, the treatment medium was removed, and the wells were washed twice with PBS. Then, 100 µL of MTT solution (0.5 mg/mL) was added to each well and incubated for 4 h in the dark. Afterward, the MTT solution was discarded, and 200 µL of DMSO was added to dissolve the formazan crystals. The plates were shaken for 15 min, and the absorbance was measured at 560–570 nm.

Dynamic motion setup and analysis

Dynamic observation of particle motion was performed using a National Optical 131-RLED student microscope equipped with a digital camera. Motion was induced using a custom-built setup featuring a precision-engineered motor that rotates a permanent magnet to generate a consistent and uniform magnetic field. The sample motion was recorded at 30 frames per second. A magnetic field (7 mT, 3 Hz) was applied.

Particle trajectories were tracked using MTrackJ tool in Fiji and analyzed via mean squared displacement (MSD) calculations. The MSD was calculated using the following Equation (Eq. 4):

4

Where N is the number of frames; P_x(i) and P_y(i) are the x and y positions of a tracked particle at frame i. The diffusion coefficient (D) and the average particle velocity (v) were obtained from the plot of MSD vs. Δt by fitting the data to the parabolic component of the Stokes-Einstein Equation (Eq. 5):

5

Results and discussion

We created Janus magnetite nanoparticles decorated with PCL and CS and attached them to helical structures from their hydrophilic side. For this purpose, hydroxyl-functionalized poly(ɛ-caprolactone) was synthesized by ring-opening polymerization (ROP). MNPs were synthesized by co-precipitation method, followed by functionalization with carboxylic acid. Then, Janus NPs were prepared through an O/W emulsion (Fig. 1, Step 1). Spirulina platensis (SP), a spiral-shape microalgae subspecies, was chosen as a bio-template for the fabrication of magnetic microhelix. The fabrication scheme for the core-shell Fe₃O₄@SiO₂ microrohelix is shown in Fig. 1, Step 2. The Fe₃O₄ core was formed by submerging the bio-templates in a suspension of IONPs, which bestowed magnetite properties to the microhelix. Subsequently, the magnetite microhelixes were coated with a SiO₂ layer; the silica shell ensured the stability of the microhelixes. Lastly, the two parts were linked using the masking method (Fig. 1, Step 3).

Fig. 1.

A diagram showing the steps used to synthesize An-SP@SiO₂: Synthesis of Si-MJNPs (Step 1). Synthesis of silica-coated magnetic microhelix (Step 2). Integration of magnetic Janus nanoparticles with silica-coated Spirulina to form the final composite structure (Step 3).

Preparation and characterization of anisotropic helical microswimmers (An-SP@SiO₂)

In the initial phase of our research, we focused on the synthesis of the head component of the anisotropic helix. This was achieved by utilizing Janus nanospheres, which are unique particles characterized by their dual-surface properties. These nanospheres were meticulously coated with two different polymers: polycaprolactone (PCL), known for its biodegradable qualities and flexibility, and chitosan, a biocompatible material. For this purpose, hydroxyl-terminated poly(ε-caprolactone) was synthesized by ring-opening polymerization of ɛ-caprolactone with ethanol as initiator and Sn(Oct)₂ as catalyst at 120 °C (Figure S1). The structure was confirmed by FT-IR spectroscopy, as shown in Figure S2. The peaks at 1731 and 1172 cm^− 1 are identified as characteristic peaks of PCL and are attributed to the stretching vibration frequency of C = O carbonyl and C-O esteric groups, respectively. The peaks at 2867 and 2946 cm^− 1 indicate the presence of C-H stretching bonds in PCL, and the O-H stretching vibrations of the terminal alcoholic group were located at 3441 cm^− 1.

¹H NMR was employed to analyze the chemical composition of PCL. The backbone protons featured characteristic peaks at 2.28 (c), 1.62 (d, f), 1.37 (e), and 4.05 (g) ppm. The methyl group of ethanol as the reaction initiator was observed at 1.24 (a) ppm, and its methylene group appeared at 4.18 (b) ppm. The triplet peak at 3.63 (g^/) ppm corresponds to the last methylene monomer of the terminal polymer, which has an alcohol bond instead of an ether bond (Figure S3). The end groups of the PCL chains made it possible to calculate the average molecular weights of the PCL chains. The integrated peak areas at 1.20 and 2.31 ppm (based on the ethylenic and PCL chain regions, respectively) in the¹H NMR spectra were used to determine the number average molecular weight (). Accordingly, the molecular weight of the resulting product was about 11,446 g/mol.

Iron oxide nanoparticles were synthesized using Fe (II) and Fe (III) in the appropriate ratio via the co-precipitation method under a nitrogen atmosphere (Figure S4). Figure S5 shows the FT-IR spectrum of the bare iron oxide nanoparticles (IONPs). The broad peak at 3400 cm^− 1 is due to the stretching vibration of Fe-OH functional groups present on the surface of magnetite NP. Additionally, the peak at 583 cm^− 1 is related to the Fe-O bonds of Fe₃O₄.

The synthesized nanoparticles were modified by (3-aminopropyl) thriethoxysilane (Am-IONPs) (Figure S6). FT-IR spectrum of Am-IONPs (Figure S7) compared to the spectrum of Fe₃O₄ nanoparticles shows the appearance of 1103 and 1160 cm^− 1 peaks related to the Si-O-H and Si-O-Si bonds. The spectrum also shows peaks at 1630 and 3433 cm^− 1 respectively, related to the bending and stretching vibrations of the N-H bonds, in addition to the strong peak at 2920 cm^− 1 related to the stretching vibration of the C-H bonds. These results confirm the binding of APTES to the surface of iron oxide.

Subsequent modification of magnetite nanoparticles was performed through a condensation reaction between Am-IONPs and succinic anhydride, as shown in Figure S8 (Ac-IONPs). Compared to the Am-IONPs FT-IR spectrum, the peaks at 1632 and 1727 cm^− 1 were assigned to the C = O stretching vibrations of the amide and carboxylic acid functions, respectively. The bond at 3426 cm^− 1 was assigned to the N-H stretching vibration, which interfered with the broad peak of the O-H bonds of the carboxylic acid groups formed on the surface (Figure S9).

The surface charge measurements of iron oxide nanoparticles (IONPs), aminated iron oxide nanoparticles (Am-IONPs), and carboxylated iron oxide nanoparticles (Ac-IONPs) revealed distinct charge characteristics: IONPs exhibited a negative surface charge (−10.2 mV), while Am-IONPs displayed a positive surface charge (+ 37.2 mV), and Ac-IONPs again showed a negative surface charge (−17.5 mV). These variations in surface charge can be directly linked to the types of functional groups present on the surface of the nanoparticles. Specifically, the negative charge on IONPs and Ac-IONPs is likely due to the presence of hydroxyl groups and carboxylic acid functionalities, which contribute to a net negative charge. In contrast, the positively charged Am-IONPs can be attributed to the presence of amine functional groups, which can accept protons, leading to an overall positive surface charge. This functionalization not only affects the surface charge but can also influence nanoparticle sizes. After modifying the nanoparticles with APTES, a notable reduction in the nanoparticle size was observed, bringing it down to 48.1 nm. Conversely, when the nanoparticles were further modified with succinic anhydride, the zeta potential absolute value decreased. Despite this reduction in charge, the size of the nanoparticles experienced only a modest increase of 5.4 nm, ultimately reaching 53.5 nm. This relatively small change in size highlights the stability of the nanoparticles, as they maintained stability with a zeta potential of −17.5 mV.

Magnetite Janus nanoparticles with isolated PCL and chitosan domains (MJNPs) were synthesized through an oil-in-water emulsion technique. These nanoparticles were further modified to form silanated magnetic Janus nanoparticles through a two-step reaction. MA-MJNPs were first prepared by Michael addition between the amino group of chitosan and the double bond of methyl acrylate structure (Figure S10) and then Si-MJNPs were obtained by a nucleophilic substitution reaction (SN2) between MA-MJNP and APTES (Figure S11). The product was subsequently characterized by FT-IR, DLS, and SEM.

The FT-IR spectrum of MJNPs illustrated a peak at 1728 cm^− 1 attributed to the C = O of the ester group, and C-O-C stretching vibrations appeared at 1100 and 1160 cm^− 1 (Fig. 2a). The stretching vibrations of C-N were at 1240 cm^− 1 and C-H bonds at 641, 1418, and 2926 cm^− 1. Two peaks at 1575 and 1640 cm^− 1 were attributed to the stretching vibrations of N-H in the amine and amide groups, respectively. The strong absorption peak at 3472 cm^− 1 was attributed to N-H bonds and O-H of alcohol, which overlapped. SEM, Fluorescence microscopy, and DSC analysis were used to investigate the morphology of the JNP (Fig. 3a and c).

Fig. 2.

FT-IR spectra of MJNP (a), MA-MJNP (b), and Si-MJNP (c).

Fig. 3.

SEM image of MJNPs (a), fluorescence microscopy image of MJNPs loaded with fluorescein (green) and acridine orange (red) dyes (b), comparative DSC thermograms of Ac-IONPs and MJNPs (c), SEM image of MA-MJNPs (d), SEM images of Si-MJNPs with two magnifications (e), and TGA/DTG thermograms of Si-MJNPs in comparison to MJNPs (f).

SEM analysis was performed to identify the morphology of the Janus NPs (Fig. 3a). According to this image, the Janus nanoparticles exhibit two distinct sides, and their shapes are almost spherical, with a diameter of about 60 nanometers. Nanoprecipitation was employed to prepare the fluorescence-labeled Janus nanoparticles for subsequent fluorescence microscopy analysis. Fluorescence microscopy images of the Janus nanoparticles revealed distinct emission patterns from each hemisphere (Fig. 3b), with the green fluorescence corresponding to the fluorescein-labeled PCL side and the orange fluorescence arising from the acridine orange-functionalized chitosan hemisphere. The clear boundary between the two regions confirms the successful synthesis of Janus nanoparticles with well-defined anisotropy. Figure 3c illustrates the thermal behavior of the synthesized MJNPs through DSC. This analysis aims to validate the successful conjugation of polycaprolactone (PCL) and chitosan (CS) onto the surface of Ac-IONPs. The DSC thermogram reveals a prominent endothermic peak at 60 °C, indicative of the melting point (Tm) of PCL. Additionally, a notable transition is observed around 165 °C, corresponding to the glass transition temperature (Tg) of CS. Further examination shows two peaks at approximately 266 °C and 320 °C, likely representing the thermal decomposition of CS and PCL, respectively. These specific thermal transitions provide strong evidence for the presence of both PCL and CS in the MJNPs’ composition.

The magnetization of Janus nanoparticles was measured using a VSM to assess their magnetic properties at room temperature. The saturation value was magnetization was determined to be approximately 7.2 emu/g (Figure S12). Although this value is lower than some might expect, previous studies have shown that saturation magnetization in this range is adequate for applications in magnetic drug targeting³².

The synthesized MJNPs were further modified with methyl acrylate (MA-MJNP) (Figure S10). FT-IR spectrum of this compound compared to the spectrum of MJNP shows that the peaks associated with N-H bonds disappeared in the regions of 1575 and 1640 cm^− 1 (Fig. 2b). SEM image of MA-MJNPs (Fig. 3d) illustrates that modifying magnetite nanoparticles (MJNPs) with methyl acrylate resulted in noticeable particle aggregation, likely attributed to a reduction in surface charge. In order to investigate this hypothesis, we assessed the size distribution and zeta potential of both the unmodified and modified MJNPs in aqueous media using dynamic light scattering (DLS) analysis. The zeta potential of the MJNPs was observed to decrease significantly from + 14.9 mV to + 2.7 mV. This considerable reduction in zeta potential can be attributed to the reaction between the amine groups of the nanoparticles and the methyl acrylate. This reaction facilitated the formation of amide bonds, which in turn reduced the overall positive surface charge of the chitosan. This substantial change correspondingly influenced the hydrodynamic diameter, which increased from 138.3 ± 7.2 nm to 876 ± 79.9 nm, confirming that aggregation occurred. This trend further supports the notion that particle aggregation is inversely correlated with surface charge.

The final modification involved the introduction of APTES to the methyl acrylate-modified magnetic Janus nanoparticles (MA-MJNP), which are now referred to as Si-MJNP (Figure S11). A comparative analysis of the FT-IR spectra between Si-MJNP and MA-MJNP revealed several significant changes. Notably, three distinct peaks emerged in the spectrum of Si-MJNP at wavenumbers of 1021, 1080, and 1150 cm^− 1. These peaks are associated with the stretching vibrations of the Si-O-Si bonds, indicating successful incorporation of the silane group into the nanoparticle structure. Additionally, a prominent absorption peak at 467 cm^− 1 was detected, which is attributed to the presence of Si-O bonds (Fig. 2c). Following the additional modification with APTES, the zeta potential increased to + 17.8 mV. This change suggests that the introduction of APTES resulted in the formation of new amide groups. These new groups contributed to the increase in the positive surface charge, effectively decreasing the particle size once more (253.4 59 nm), which can influence their stability and reactivity (Table 1).

Table 1.

Particle size and zeta potential of IONPs, MJNPs, and their derivatives.

Nanoparticles	Hydrodynamic size (nm)	Zeta potential (mV)
IONPs	123.924.9	−10.2
Am-IONPs	48.1 24.4	37.2
Ac-IONPs	53.5 30.6	−17.5
MJNPs	138.3 7.2	14.9
MA-MJNPs	876.2 79.9	2.7
Si-MJNPs	253.4 59	17.8

SEM analysis was conducted to elucidate the morphology of the Si-MJNPs (Fig. 3e). The micrograph reveals two detectable domains in each nanoparticle, which predominantly exhibit a nearly spherical morphology with an approximate diameter of 230 nanometers.

TGA was utilized to quantify the weight% of the coated polymers on the surface of the iron oxide nanoparticles. The TGA and DTG thermograms of MJNP and Si-MJNP are shown in Fig. 3f. The TGA curve indicated a significant weight loss of 73% for MJNP, compared to 62% for Si-MJNP, attributed to the decomposition of the organic layer on the surface. A weight loss of about 1.2% at 200 °C indicated the evaporation of adsorbed moisture and volatile components. Notably, the weight loss of approximately 38.5% at 255 °C belonged to chitosan, while the weight loss of approximately 28.35% at 302.2 °C was attributed to PCL. At 600 °C, Si-MJNP exhibited a higher residual weight of roughly 10%, likely due to the presence of increased mineral content (silica residue).

Following the previous step, we entered the second phase, where we concentrated on the development of the tail component of the anisotropic helix. This tail was crafted with a sophisticated helical architecture designed to enhance functionality. At its core, we incorporated magnetite, a magnetic iron oxide known for its ferromagnetic properties, which was seamlessly encapsulated within a silica shell. The silica not only provides structural stability but also protects the magnetite core from environmental interactions, thereby preserving its magnetic characteristics. Spirulina was employed as a template to fabricate microswimmer tails. The morphology of the prepared spirulina template was characterized using optical microscopy. The initial structural lengths were approximately 100 μm (Fig. 4a), which needed to be reduced through ultrasound irradiation for effective application as potential microswimmers in biological environments. The optimal ultrasonic exposure time was determined by systematically varying the irradiation duration from 2 to 60 s. Optical microscopy analysis of the original spirulina and SP@SiO₂ before irradiation (Fig. 4a and b) and after 2 s of ultrasonication (Fig. 4c and d) showed that their lengths decreased to approximately 60 μm. Further ultrasonication for 5, 30, and 60 s resulted in a gradual reduction in size to 50 μm, 30 μm and 20 μm, respectively (Figure S13).

Fig. 4.

Optical microscopic images of (a) original spirulina (SP) and (b) SP@SiO₂ in distilled water, compared to (c) SP and (d) SP@SiO₂ after 2 s of ultrasound sonication treatment.

Our primary objective is to synthesize helical magnetite nanoparticles (SP@Fe₃O₄). To achieve this, we utilized Spirulina as a natural template to construct the tails of our microswimmers. The process begins with the synthesis of iron oxide nanoparticles (IONPs), which are created while integrating Spirulina into the mixture. This integration allows for the efficient deposition of the iron oxide nanoparticles onto the Spirulina structure, facilitating the formation of the desired helical shape Iron oxide nanoparticles (IONPs) around Spirulina platensis were confirmed using FT-IR spectroscopy. Figure 5 compares the FT-IR spectra for pure Spirulina platensis and Fe₃O₄-coated Spirulina platensis (SP@Fe₃O₄). The FT-IR spectrum of pure Spirulina platensis reveals various functional groups present on its surface. It exhibits absorption peaks corresponding to N-H and O-H bonds (~ 3200–3400 cm^− 1), aliphatic C-H (2850–2920 cm^− 1), C = O (amide I, 1651 cm^− 1), and C-H bending (1451 cm^− 1), along with N-H bending (amide II, 1539 cm^− 1). Additionally, a characteristic band within the 1000–1400 cm^− 1 range suggests the presence of vibrations related to various functional groups, such as S = O, C = C, C-O, and C-OH^33,34. SEM images were used to investigate the morphology of raw and modified spirulina and to measure their respective filament diameters using ImageJ (Fig. 5). The raw form of Spirulina displayed its characteristic helical structure, with lengths of approximately 60 μm, a helix diameter of approximately 18 μm, and a filament diameter of around 1.83 μm. As illustrated in Fig. 5, the FT-IR spectrum of Spirulina Platensis, upon being coated with Fe₃O₄ nanoparticles, retained all of the primary absorption bands that correlate with the core components of Spirulina. Notably, the spectrum revealed two significant absorption peaks at 579 cm^− 1 and 618 cm^− 1. These peaks correspond to the stretching vibrations of iron-oxide (Fe-O) bonds, which serve as strong indicators of the successful immobilization of iron oxide nanoparticles (IONPs) onto the Spirulina platensis template. Further examination using SEM provided visual confirmation of these changes. After the coating process with Fe₃O₄ nanoparticles, the overall helical morphology of the Spirulina was found to be well-preserved. However, there was a marked decrease in surface roughness, which strongly suggests the successful and relatively uniform deposition of Fe₃O₄ nanoparticles onto the Spirulina surface (Figure S14). Additionally, it was observed that the filament diameter increased to approximately 2.8 μm (calculated from Fig. 5), a measurement that indicates the presence of the nanoparticle layer enveloping the Spirulina filaments and further illustrates the effectiveness of the coating process.

Fig. 5.

FT-IR spectra and SEM images of Spirulina platensis, SP@Fe₃O₄, and SP@SiO₂.

Silica coating is used to protect iron oxide nanoparticles (IONPs) and improve their performance in biomedical applications. Bare magnetic nanoparticles have high surface energy, leading to agglomeration and rapid clearance from the body. They also risk oxidation, which can degrade their magnetic properties and release free iron ions, potentially promoting bacterial growth. Thus, surface modification is crucial for safe use in biomedical settings. The silica coating enhances the stability, biocompatibility, and functionality of IONPs by preventing aggregation and allowing for easy functionalization due to the hydroxyl groups on its surface. This method is effective for creating stable and biocompatible nanostructures, as silica offers optical transparency, chemical inertness, thermal stability, and controlled porosity, which are essential for minimizing toxicity and enhancing stability in biomedical applications.

The FT-IR analysis provided additional validation of the silica enveloping the microhelices. In Fig. 5, a direct comparison is made between the spectral characteristics of SP@Fe₃O₄ particles and those of SP@SiO₂ particles. Notably, the spectrum of SP@SiO₂ displays a prominent and broad absorption band centered around 1086 and 1146 cm^− 1. This distinct feature is indicative of the Si-O-Si asymmetric stretching vibrations, suggesting a robust silica network surrounding the microstructures. The SEM image of encapsulated SP@Fe₃O₄ with a silica layer (SP@SiO₂) shows a porous surface topology (Fig. 5), and surface roughness further confirms this, likely due to the mesoporous nature of the silica layer (Figure S14). Additionally, the filament diameter was measured to be 4.6 μm (calculated from Fig. 5), confirming the presence of the silica coating. A comparison of the SEM images illustrates the changes in surface morphology and the gradual increase in filament diameter with each functionalization step, transitioning from raw Spirulina to SP@Fe₃O₄ and finally to SP@SiO₂.

In our quantitative investigation of the silica surface, we have systematically characterized pore size dimensions using critical parameters extracted from SEM imagery. Binarization is the process of dividing the grayscale values in an image into two categories by setting a specific threshold. This creates a binary image with only two values (black and white), which makes the analysis faster. Choosing the best threshold is crucial because it affects the accuracy of any further analysis. Figure S15 demonstrates the binarized image of the silica surface obtained using the Renyi entropy thresholding method, where black regions correspond to the pores, while white areas indicate the surface and background. The pore size distribution ranges from 35 to 180 nm, with a mode size of approximately 87 nm, indicating both mesoporous and macroporous characteristics (Fig. 6a).

Fig. 6.

SEM image-based frequency distribution of pore sizes (a), EDX spectrum (b), XRD pattern (c), and TEM image (d) of SP@SiO₂.

The EDX spectrum confirms the presence of Si, Fe, O, and C in the synthesized SP@SiO₂, indicating the successful silica coating on the SP@Fe₃O₄ (Fig. 6b). The crystallographic structure of the SP@SiO₂ microhelices was analyzed using X-ray diffraction (Fig. 6c). A prominent broad peak observed between 2θ values of 20° and 30° (specifically centered at 2θ = 22°) corresponds to the characteristic peak of amorphous SiO₂, as indicated by JCPDS Card No. 29–0085, thus confirming the presence of silica in the samples in the form of mesoporous. Additionally, distinct diffraction peaks at 30°, 35.3°, 38°, 43°, 53.7°, 57°, and 62.9° correspond to the (220), (311), (222), (400), (511), (440), and (533) crystal planes of the cubic Fe₃O₄ phase (JCPDS No. 75–0449). The TEM image of the synthesized SP@SiO₂ (Fig. 6d) demonstrates a mesoporous silica layer with a thickness of approximately 200–300 nm. This observed porous architecture appears to provide an optimal matrix for the adsorption of water pollutants, making it a promising candidate for remediation applications. Furthermore, this structure may serve as an effective vehicle for the delivery of pharmaceuticals or catalysts. However, further investigations will be essential to fully explore these potentials.

The final step of our synthesis process involved the precise conjugation of the head and tail components (Fig. 1, Step 3). This critical assembly stage was essential for creating the complete anisotropic helix structure, ensuring that both parts were effectively integrated to achieve the desired anisotropic properties and functionalities of the final product. To fabricate anisotropic helices, it is essential to ensure that only one end of the SP@SiO₂ particles is connected with spherical particles. To achieve this, we prepared a mixture of SP@SiO₂ suspended in H₂O/PVA, which was subsequently cast onto a freshly cleaned glass substrate. A neodymium plate magnet was employed to cover the entire surface during this process (Figure S16). PVA serves a dual purpose in this setup. First, it modified the surface tension of the aqueous solution, depending on its concentration. Previous studies have demonstrated that in diluted regimes, PVA acts similarly to a surfactant, reducing the surface tension of water as the polymer concentration increases, reaching a minimum at approximately 4 wt% PVA³⁵. Beyond this concentration, surface tension begins to rise again until around 10 wt%. We opted for a 5% PVA solution to achieve an optimal balance between reduced surface tension and sufficient polymer concentration for influencing object buoyancy. Under these conditions, horizontal helices are unable to float towards the water’s surface, while vertical helices can do so. This behavior is analogous to the case of a needle in water: a larger contact area results in reduced pressure exerted on the liquid surface, facilitating buoyancy. When a steel needle is placed vertically, its small contact area results in greater pressure on the water surface, causing it to submerge. Similarly, when objects attempt to breach the liquid surface, surface tension effectively inhibits horizontal helices from emerging, while allowing vertical helices to exit with minimal applied magnetic field strength (Fig. 7a). The second advantage of PVA lies in its function as a soft substrate. During the drying process, facilitated by evaporation under ambient conditions, PVA is solidified, effectively encapsulating the helices. This solidification occurs as the viscosity of the PVA solution increases fourfold compared to pure water³⁶, a change that mitigates delocalization of the helices. Furthermore, the addition of modified magnetic Janus nanoparticles (Si-MJNPs) to a PVA film, which contains microhelices structures suspended in a solvent within a nitrogen atmosphere, leads to reactive interactions. Specifically, the active sites of Si-MJNPs engage with the exposed ends of the helices protruding from the PVA matrix (Figure S16).

Fig. 7.

Optical microscopy image of a cross-section of a PVA film containing SP@SiO₂, where one end of the helices protrudes from the film surface while the rest remains embedded within the matrix (a), FT-IR spectrum (b) and SEM images (c) of An-SP@SiO₂.

The final architecture of the composite was validated through FT-IR and SEM analysis. The FT-IR spectrum of helix microswimmers connected to MJNPs exhibited several prominent bands at approximately 3434, 2925, 1633, 1729, and 1236, and within the range of 1060–1535 cm^− 1 (Fig. 7b). The peak at 3434 cm^− 1 is attributed to –OH and –NH groups. The peak around 2925 cm^− 1 corresponds to –CH stretching vibrations. The bands in the 1390–1730 cm^− 1 region can be associated with –C = O stretches from aldehydes, ketones, and carboxylates. Additionally, stretching vibrations of –C–O, –C–C, and –C–OH are observed in the 1000–1300 cm^− 1 region. Figure 7c presents SEM images of two fabricated microswimmers featuring Janus nanoparticle heads. These images emphasize the morphological characteristics and surface architecture of the final constructs. Notably, despite variations in shape and pitch distances, the intrinsic structure of the final microparticles exhibited a consistent pattern, characterized by the anisotropic deposition of the Janus head onto the helical tail.

Evaluation of the properties of DOX-loaded MJNPs

Drug loading and encapsulation efficiency were determined by UV-Vis spectroscopy, resulting in a loading value of 6.98% loading and an encapsulation efficiency of 75%. The in vitro release of doxorubicin (DOX) from the MJNPs was assessed in PBS at 37 °C under two pH conditions: pH 7.4, which simulates physiological conditions, and pH 5.8, designed to emulate the tumor-microenvironment. As shown in Fig. 8a, both release profiles exhibited an initial burst phase during the first 6 h, with approximately 60% of DOX was released at pH 5.8, compared to 55% at pH 7.4. Following this phase, the release rate decreased, transitioning to a slower, more sustained release. The faster release observed under acidic conditions can be attributed to the protonation of DOX’s amino groups at a lower pH increases its solubility, thereby facilitating its diffusion out of the nanoparticle matrix.

Fig. 8.

In vitro release profile of DOX-loaded NPs at two different pH (a) Viability of B16-F10 cells at various concentrations of free DOX, DOX-loaded NPs, and drug-free NPs (b).

Biocompatibility and cytotoxicity were evaluated by MTT assay on rat B16-F10 melanoma cells after a 24 h treatment. Figure 8b shows cell viability in response to free DOX, DOX-loaded MJNPs, and drug-free MJNPs across a range of 0.001–200 µg/mL at 37 °C. The drug-free nanoparticles did not induce significant cytotoxicity, indicating that the MJNPs themselves are biocompatible and that any observed toxicity is due to released DOX. Toxicity of both free DOX and DOX-loaded MJNPs was dose- and time-dependent; the half-maximal inhibitory concentration (IC₅₀) values were calculated to be 6.5 µg/mL and 18.7 µg/mL, respectively. The higher IC₅₀ value of the DOX-loaded MJNPs compared to free DOX suggests a sustained drug release profile, where the nanoparticle matrix delays the immediate intracellular accumulation of the drug during the first 24 h.

Motion characterization of An-SP@SiO₂ microswimmers

The engineered microswimmers we developed can be controlled and maneuvered wirelessly through a carefully configured magnetic field that takes into account their unique shapes and inherent magnetic characteristics. An experimental setup developed by our research group includes a precision-controlled rotating permanent magnet system. This setup was specifically designed to generate a uniform rotating magnetic field and is coupled with a microscope and a digital camera, enabling real-time video capture of microswimmer motion under various field conditions. The experimental setup is illustrated in Figure S17. The microswimmer exhibited precise movement control in distilled water, as confirmed by time-lapse imaging (Figure S18). The typical system employed on the anisotropic microhelices under three different conditions, included: without any external magnetic field (Brownian motion, Video S1), in the presence of a fixed magnet (Video S2), and in the presence of a rotational magnet (Video S3). To evaluate the effects of particle size and medium, we tracked the swimming behavior of large (approximately 70 μm), medium (approximately 40 μm), and small (approximately 20 μm) microswimmers in four media: distilled water (Video S4), HSA (1%) (Video S5), BSA (1%) (Video S6), and FBS (10%) (Video S7). The protein-based solutions served as biologically relevant environments that better mimic in vivo conditions. The resulting pathways traced by helical microswimmers demonstrate their distinct movement dynamics at the microscale. These diminutive swimmers, often analogous to miniature screws, maneuver through diverse environments by converting energy into rotational motion. The helical trajectories were characterized by spiraling paths, facilitating efficient propulsion within fluidic media.

Given the inherent size heterogeneity of the biotemplating method, a quantitative geometric analysis was performed to ensure a justifiable categorization for our propulsion tests. Based on the number of turns, the particles were categorized as Large (H ≈ 70 μm, ~ 4 turns), Medium (H ≈ 40 μm, ~ 1 turn), and Small (H ≈ 20 μm, 0.5 turn). Critically, we observed that while the helix height (and thus the number of turns) varied significantly, the helix radius (D_helix/2) and helical pitch (P_helix) remained statistically consistent across the groups (e.g., D_helix was 18 ± 2 μm, and P_helix was 22 ± 2 μm). Therefore, the performance differences discussed here are primarily attributable to the number of helical turns.

The trajectories of microswimmers and the corresponding MSD analysis for three particle sizes under three propulsion modes in four environments over a 5-second observation period are presented in Figures S19 and 9a, respectively. As expected, the swimming behavior was strongly influenced by particle size, which is directly related to the number of turns in the helix. Larger microswimmers, which contain more magnetic material and have a greater number of helical turns, generate higher magnetic torque when exposed to an applied field. However, their increased drag, resulting from a larger mass and surface area, limits their net propulsion when the field was not rotating. In contrast, under a rotating magnetic field, the helical structure of these swimmers enables them to convert rotational torque into forward motion more effectively. This efficiency tends to improve with an increasing number of turns and a more favorable helix geometry. Additionally, the higher viscosity of the protein-based media used to mimic biological environments reduces swimming speeds compared to water. This reduction is particularly pronounced for large swimmers due to their higher drag. In the smallest size group, the difference in swimming behavior between rotating and static field conditions was less noticeable. This discrepancy likely arises from the incomplete formation of a fully developed helical structure at this scale, resulting in a reduced ability to convert rotation motion into effective forward propulsion.

Figure 9b presents the average velocities (µm/s) and diffusion coefficients (µm²/s) of the microswimmers, which were extracted from MSD analyses. The data demonstrates similar trends for both parameters when tested under various actuation modes and media. Notably, in more viscous fluids, large-sized particles continue to maintain relatively high propulsion speeds when subjected to rotating magnetic fields. This observation supports our hypothesis that helical architectures, particularly those with multiple turns, remain effective in converting torque-to-thrust, even in increased viscous conditions. Consequently, medium-sized microswimmers with only a single helical turn, as well as smaller particles that have largely lost their helical morphology, exhibit markedly reduced propulsion performance compared to their behavior under constant magnetic fields. It is worth noting that in the rotational mode depicted in Figure S19, a lateral drift (orthogonal to the intended swimming direction) is observed. This deviation can be attributed to hydrodynamic interactions between the microswimmers and the nearby substrate^37,38. When helical swimmers operate near a solid boundary, the apparent viscosity of the surrounding fluid effectively increases as the distance to the wall decreases. This viscosity gradient induces an imbalance in the local drag forces, causing the microrobots to roll or tumble along the surface. As a result, the interaction between rotation and surface-induced drag leads to a sideways drifting motion. This effect becomes more pronounced for larger microswimmers or in low-viscosity media, where enhanced sedimentation brings the swimmers closer to the substrate (Table S1).

Fig. 9.

Typical single-particle tracking MSD values of microswimmers (a), along with their velocity and diffusion coefficient (b) profiles for three different-sized samples under three motion modes in four environments are presented (field parameters: B = 7 mT, f = 3 Hz, with a time scale of Δt = 5 s) (obtained from Supplementary Videos S4-S7).

The swimming performance of the helical microswimmers is limited by two specific frequency thresholds. The break-away frequency marks the point at which sustained propulsion begins, where the generated magnetic torque and hydrodynamic thrust overcome adhesive and gravitational forces. In contrast, the step-out frequency defines the upper limit for synchronous rotation with the magnetic field; beyond this frequency, the swimmer loses phase alignment, leading to a rapid decline in propulsion efficiency⁵. For larger microswimmers in water, these thresholds were measured as 1.5 Hz for the break-away frequency and 18 Hz for the step-out frequency, establishing an effective operational frequency window.

Conclusion

We successfully developed a robust fabrication method for anisotropic helical microswimmers, which consist of a magnetite Janus PCL/Chitosan polymer head coupled to a Spirulina-derived helical tail coated with Fe₃O₄ and SiO₂. Subsequent biological evaluations on B16-F10 melanoma cells revealed a crucial dual behavior, while the drug-free microswimmers demonstrated high biocompatibility with negligible toxicity, their DOX-loaded counterparts exhibited significant therapeutic efficacy. This distinct contrast validates the bio-hybrid system as a safe yet potent platform for targeted drug delivery. Furthermore, the motion analysis showed that while the diffusion coefficient and velocity predictably decreased with increasing viscosity, the rotational-to-translational motion was significantly enhanced with an increase in the number of helices. The highest observed speed was approximately 65 μm/s for large particles in water under a rotating magnetic field, validating the effectiveness of the propulsion mechanism.

The successful integration of magnetic steering and biocompatible drug loading in this platform represents a promising path for actively-guided therapeutic delivery. Preliminary results further demonstrated that the silica-coated helical tails could be fractured upon brief exposure to ultrasound. Future work will build on our current capabilities and focus on three key directions: First, we will perform systematic testing of the microswimmers’ velocity across a wide range of magnetic field rotation frequencies to fully characterize the kinematics. Second, we will investigate the potential for dual drug loading using the distinct compartments of the Janus head structure. Finally, we aim to complete the in vitro evaluation of localized, ultrasound-triggered drug release from both the Janus head and the helical tail for a precise, sequential therapeutic strategy. This comprehensive approach, which will eventually include in vivo testing, aims to achieve the full therapeutic potential of these microswimmers.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Mahdieh Jahani : Investigation, Data curation, Formal analysis, Writing – original draft & editing. Sepideh Khoee: Supervision, Resources, Funding acquisition Conceptualization, Project administration, Data curation, Formal analysis, Validation, Writing – review & editing. Masoud Mirmasoumi : Cultivation of spirulina.

Data availability

The authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Information files.

Declarations

Competing interests

The authors declare no competing interests.