Collective Magnetic Mesoporous Silica Nanorobots for Targeted Oral Capsaicin Delivery in Colitis Intervention

TL;DR

Researchers developed magnetic nanorobots that deliver capsaicin to the colon to treat inflammatory bowel disease more effectively.

Contribution

A novel magnetic nanorobot system for targeted oral capsaicin delivery in colitis treatment is introduced.

Findings

Capsaicin-loaded nanorobots achieved collective locomotion up to 180.7 μm/s under magnetic guidance.

The system enables targeted drug enrichment and sustained release at colonic lesion sites.

The nanorobots overcome millimeter-scale obstacles during movement in the gut.

Abstract

Magnetic nanoparticles, with their excellent biocompatibility and biodegradability, serve as ideal materials for constructing targeted drug delivery systems. Iron oxide (Fe3O4) nanoparticles, controllably prepared via methods such as solvothermal synthesis, can be combined with mesoporous silica to construct magnetically steerable nanorobots. Such robots enable efficient drug loading and precise delivery. To address challenges in the treatment of Inflammatory Bowel Disease (IBD), including the significant side effects of systemic drugs and the low oral bioavailability and poor colonic targeting of novel food-derived drugs (e.g., capsaicin with anti-inflammatory activity), this study designed capsaicin-loaded iron oxide-mesoporous silica composite nanorobots (Cap-M@mSbots). Driven by a rotating gradient magnetic field of up to 80 mT, Cap-M@mSbots achieve large-scale emergent collective locomotion, with a maximum collective locomotion velocity reaching 180.7 μm/s, and are capable of long-distance movement overcoming millimeter-scale obstacles. This system can be actively propelled to colonic lesion sites under magnetic guidance, achieving targeted drug enrichment and sustained release, thereby offering a novel strategy for the targeted therapy of IBD.

Figures

Full text

1. Introduction

Magnetic nanoparticles represent a class of nanomaterials with inherent biocompatibility. Living organisms themselves synthesize certain magnetic nanoparticles through biomineralization processes, such as magnetotactic bacteria containing natural magnetic particles [1,2] and sensory organs in migratory birds that rely on the Earth’s magnetic field for navigation [3]. In addition to naturally occurring magnetic nanoparticles, synthetic magnetic nanomaterials are also widely used in applications such as magnetic hyperthermia therapy for tumors and as components of composite materials for implantable devices. Experimentally validated magnetic nanoparticles demonstrate good cytocompatibility, extremely low hemotoxicity and genotoxicity, and a high median lethal dose (LD50), indicating that magnetic nanoparticles are promising materials with extensive biomedical application potential [4,5,6,7,8,9,10,11,12,13]. They can be injected into the bloodstream for magnetically guided drug delivery and perform tasks such as delivering targeted therapeutics for thrombolysis, after which they are degraded via metabolic pathways (e.g., in the liver) and excreted from the body, demonstrating their inherent biodegradability [4,14]. Regarding drug delivery carriers, numerous studies have shown that modified magnetic nanoparticles can be utilized for drug loading and delivery [15,16]. Concerning the synthesis of magnetic nanoparticles, particularly the widely used iron oxide (Fe_3_O_4_) nanoparticles, various methods exist, including coprecipitation, solvothermal synthesis, and thermal decomposition, et al. Among these, the solvothermal method is particularly suitable for large-scale synthesis of Fe_3_O_4_ nanoparticles due to its advantages in producing particles with controllable size, uniform distribution, and ease of surface modification [17].

“Will injectable, disease-fighting nanobots ever be a reality?” is one of the 125 scientific questions for the 21st century proposed by Science. Among them, the application of magnetic micro/nanorobots driven by alternating magnetic fields is inseparable from micro/nano structures primarily composed of magnetic nanoparticles. Based on magnetic materials of various morphologies and sizes, magnetic nanorobots can be constructed through methods such as layer-by-layer assembly [18], surface modification [19], and nanoparticle deposition [20]. Examples include helical magnetic nanorobots fabricated by templated deposition on natural spiral microalgae [20] and hematite nanorobots [21] with dual-spherical structures. The primary functions of magnetic nanorobots are realized by tuning the magnetic nanoparticles and the nanorobot structure, with magnetic nanoparticles being a crucial component enabling external field manipulation and active targeted drug delivery [22,23]. Among the technical routes for modifying magnetic nanoparticles to equip them with various functions as nanorobots, coating magnetic nanoparticles with a mesoporous silica (mSiO_2_) shell provides a substantial volume for drug loading. Simultaneously, the network of narrow and densely packed pores can retard drug release, enabling efficient drug encapsulation [24]. The magnetic nanorobots composed of magnetic nanoparticles and mesoporous silica possess multiple advantages, including magnetically driven motion, precise manipulation, and targeted delivery, making them suitable for targeted drug delivery and therapy for diseases.

Inflammatory bowel disease (IBD) is a condition characterized by chronic, recurrent intestinal inflammation. Its global incidence continues to rise, posing a significant public health challenge [25,26]. Current therapeutic agents, such as glucocorticoids, primarily function through broad immunosuppression. However, these systemic treatments have notable limitations, including potential increased risks of infection and malignancy, potential long-term organ toxicity, and inadequate or lost response in some patients [27,28]. Consequently, there is a pressing clinical need to develop novel therapeutic strategies that combine good safety profiles with clear anti-inflammatory activity. In this context, natural active components derived from daily diets are considered highly promising candidate drug sources due to their widespread anti-inflammatory and antioxidant properties and relatively high safety profiles. Developing these dietary anti-inflammatory substances as adjunctive or alternative therapies for IBD aligns with the urgent clinical demand for safer and more accessible treatment options [29,30]. Capsaicin has garnered attention for its demonstrated anti-colitis activity in experimental models [31]. Studies indicate that, within an appropriate dosage range, capsaicin can effectively ameliorate dextran sulfate sodium-induced colitis in mice. Its mechanisms involve downregulating the expression of pro-inflammatory cytokines such as IL-6 and IL-18, while enhancing the expression of intestinal tight junction proteins (e.g., ZO-1) and mucins, thereby remodeling the gut microbiota and repairing the intestinal mucosal barrier [32,33]. However, translating capsaicin into a clinically effective drug faces significant pharmacokinetic barriers: its high hydrophobicity leads to extremely low oral bioavailability; furthermore, its lack of specific distribution within the gastrointestinal tract results in most of an oral dose being absorbed or metabolized in the upper digestive tract, making it difficult to achieve and maintain effective therapeutic concentrations at colonic lesion sites. Particularly important is that capsaicin has a direct irritant effect on the gastrointestinal mucosa, especially in the anal region, leading IBD patients to often avoid consuming spicy products. This makes safe and effective administration in IBD patients extremely challenging.

Herein, to fundamentally address the critical limitations of conventional oral administration—specifically the severe mucosal irritation and low bioavailability of capsaicin—we developed a magnetic mesoporous silica nanorobot system (Cap-M@mSbots) designed for the active interventional prevention and targeted treatment of Inflammatory Bowel Disease (IBD), as schemed in Figure 1. By encapsulating superparamagnetic Fe_3_O_4_ cores within a mesoporous silica shell, this composite structure not only shields the upper gastrointestinal tract from direct irritation but also serves as a high-capacity reservoir for the sustained release of anti-inflammatory agents. Crucially, distinguishing this work from passive nanocarriers, we introduce an active propulsion strategy driven by a rotating gradient magnetic field. Under this external actuation, the Cap-M@mSbots overcome Brownian motion to self-assemble into large-scale, reconfigurable linear swarms, exhibiting emergent collective locomotion with velocities reaching 180 μm/s. This swarm intelligence enables the nanorobots to navigate complex physiological environments, effectively surmounting millimeter-scale obstacles to achieve precise directional enrichment and prolonged retention at colonic lesion sites. Consequently, this active targeting capability significantly amplifies the therapeutic efficacy of capsaicin by restoring the intestinal mucosal barrier and down-regulating the NF-κB signaling pathway, thereby establishing a novel, safe, and efficient paradigm for the clinical management of colitis.

2. Materials and Methods

2.1. Materials

Ethylene glycol (Macklin, Shanghai, China), Poly(4-styrenesulfonic acid-co-maleic acid) sodium salt (PSSMA 3:1, Mw 20,000) (Aladdin, Shanghai, China, P107092), ferric chloride hexahydrate (FeCl_3_·6H_2_O) (Macklin, Shanghai, China), sodium acetate (NaAc) (Macklin, Shanghai, China), (3-aminopropyl)triethoxysilane (APTES) (Aladdin, Shanghai, China), fluorescein isothiocyanate (FITC) (Macklin, Shanghai, China), ammonia solution (Macklin, Shanghai, China), hexadecyltrimethylammonium bromide (CTAB) (Aladdin, Shanghai, China), HCl (Macklin, Shanghai, China), Ethanol, and capsaicin (Cap) (MCE, Minneapolis, MN, USA) were used.

2.2. Methods

2.2.1. Synthesis of Fe3O4 Magnetic Nanoparticles

The superparamagnetic Fe_3_O_4_ nanoparticles (MNPs) were synthesized via a modified solvothermal reduction method using PSSMA as a structure-directing stabilizer. This method has been previously reported in the literature [34]. Specifically, 1.0 g of Poly(4-styrenesulfonic acid-co-maleic acid) sodium salt (PSSMA, monomer ratio 3:1, Mw ≈ 20,000) was completely dissolved in 40 mL of ethylene glycol (EG) under vigorous magnetic stirring until a clear, transparent solution was formed. Subsequently, 1.08 g of FeCl_3_·6H_2_O was added to the solution. Upon the addition of 3.0 g of anhydrous sodium acetate (NaAc), the mixture turned into a homogeneous red-brown solution under continuous stirring for 30 min. The resulting mixture was then transferred into a Teflon-lined stainless-steel autoclave (100 mL capacity) and heated at 200 °C for 10 h. During this process, the PSSMA chains coordinated with iron ions to form a restricted gel-like network, controlling the nucleation and aggregation of nanocrystals. After the reaction, the autoclave was cooled naturally to room temperature. The black magnetic products were collected using a permanent magnet and washed repeatedly with ethanol and deionized (DI) water (3 times each) to remove excess organic residues and unreacted ions. Finally, the purified Fe_3_O_4_ MNPs were redispersed in 20 mL of DI water and stored at 4 °C.

2.2.2. Synthesis of Fe3O4@mSiO2 Nanoparticles (M@mSiO2 NPs)

The mesoporous silica shell was coated onto the Fe_3_O_4_ cores using a surfactant-templated sol-gel process. Initially, 0.48 g of CTAB was dissolved in 40 mL of DI water in a 250 mL three-necked flask at 40 °C. To this solution, 2 mL of the previously prepared Fe_3_O_4_ MNP dispersion containing approx. 34.4 mg of Fe_3_O_4_ was added and treated with ultrasonication for 30 min to ensure monodispersity. The mixture was then mechanically stirred at 60 rpm in a water bath maintained at 40 °C. After stabilizing for 14 min, 2 mL of ammonia solution (28 wt%) was rapidly injected to initiate the hydrolysis condition. Subsequently, a mixture containing 200 μL of TEOS and 120 μL of the FITC-APTES fluorescent precursor (prepared by stirring 1 mg FITC and 1 mL APTES in 20 mL anhydrous ethanol overnight) was added dropwise to the reaction system. The co-condensation reaction was allowed to proceed for 40 min under continuous stirring. The resultant core-shell nanoparticles were collected magnetically and washed with anhydrous ethanol. To completely remove the CTAB template and generate mesopores, the particles were refluxed in an acidic ethanol solution (1.5 mL concentrated HCl in 200 mL ethanol) at 60 °C for 6 h. The final Fe_3_O_4_@mSiO_2_ nanoparticles (M@mSiO_2_ NPs) were washed with DI water until neutral and stored at 4 °C in the dark.

2.2.3. Preparation of Capsaicin-Loaded Nanobots (Cap-M@mSbots)

The loading of capsaicin (Cap) into the mesoporous channels of M@mSiO_2_ NPs was achieved via a physical adsorption method driven by the concentration gradient. Briefly, 10 mg of purified M@mSiO_2_ NPs were collected by centrifugation and dispersed in 1 mL of anhydrous ethanol containing capsaicin (10 mM). The use of ethanol ensured the high solubility of the hydrophobic capsaicin and facilitated its diffusion into the pores. The suspension was incubated on a rotary shaker at room temperature for 3 h to reach adsorption equilibrium. Afterward, the capsaicin-loaded nanorobots were separated by centrifugation at 8500 rpm for 3 min. To remove the loosely attached drug on the surface, the precipitate was quickly washed with DI water. The final Cap-M@mSbots were resuspended in DI water and stored at 4 °C for immediate use in subsequent experiments.

2.2.4. Characterization of Nanoparticles

The morphological details, elemental composition, and internal structure of the synthesized nanoparticles were systematically characterized using a field-emission transmission electron microscope (TEM, TALOS F200X, FEI, Hillsboro, OR, USA) operated at an accelerating voltage of 200 kV. Specifically, High-Angle Annular Dark-Field Scanning Transmission Electron Microscopy (HAADF-STEM) mode was employed to visualize the core-shell configuration, while the integrated Energy-Dispersive X-ray Spectroscopy (EDS, Oxford, UK) system was used for elemental mapping and line scan analysis to confirm the distribution of Fe and Si elements. For sample preparation, the nanoparticle dispersion was diluted with ethanol, dropped onto carbon-coated copper grids, and dried at room temperature. The hydrodynamic diameter and polydispersity index (PDI) of the nanoparticles in aqueous suspension were determined using a Zetasizer Nano ZS90 instrument (Malvern Panalytical, Malvern, Worcestershire, UK) at 25 °C. To evaluate the magnetic properties, the magnetic hysteresis loops were measured at room temperature (300 K) using a Vibrating Sample Magnetometer (VSM, Lakeshore 7404, Westerville, OH, USA) with a magnetic field sweep range of −20,000 to +20,000 Oe. The optical properties and drug loading efficiency were analyzed using a UV-Vis-NIR spectrophotometer (Lambda 950, PerkinElmer, Shelton, CT, USA). The drug loading content was quantified by measuring the characteristic absorbance of capsaicin at 280 nm. Additionally, the in vitro drug release profile was monitored by dialyzing the Cap-M@mSbots against different media (water, simulated gastric fluid, and simulated intestinal fluid) and quantifying the released capsaicin in the supernatant at predetermined time intervals using the same spectrophotometric method.

2.2.5. Equipment Setup for Alternating Magnetic Field Propulsion

The collective motion and swarm behavior of Cap-M@mSbots were investigated using a custom-built magnetic generation system integrated with an inverted optical microscope (GP-830T, KSGAOPIN, Shenzhen, China). The magnetic actuation system consists of a custom-designed electromagnetic coil system connected to a programmable power supply, capable of generating a rotating gradient magnetic field with adjustable frequency and intensity (up to 80 mT). For motion observation, the nanorobot suspension was placed in a glass-bottom dish or a microfluidic channel. The formation of linear swarms and their locomotion trajectories were recorded in real-time using a CCD camera attached to the microscope. The motion videos were subsequently analyzed using ImageJ software (version 1.54p) with the Manual Tracking plugin to calculate the velocity and analyze the collective behavior patterns. To evaluate the obstacle-crossing capability, agarose gel blocks (approx. 1 mm in height) were fabricated and placed in the petri dish to simulate intestinal folds, and the swarms were guided to traverse these barriers under the rotating magnetic field.

2.2.6. Animal Models and Experimental Protocol

All animal experiments were strictly conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of the Second Affiliated Hospital of Harbin Medical University (Approval No.: SYDW2025-109). Male C57BL/6J mice (8–10 weeks old) were purchased from the Specific Pathogen-Free (SPF) animal center of our institution. They were housed under standard conditions (12 h light/dark cycle, temperature 22 ± 1 °C, humidity 50 ± 10%) with free access to sterilized food and water.

To evaluate the prophylactic and therapeutic potential of the magnetic nanocarrier system, mice were randomly divided into four groups (N = 6 per group): (1) Healthy Control Group: Received normal drinking water throughout the study and was orally administered a blank solvent vehicle, equal in volume to the nano-formulation, daily for the 3 weeks prior to the experiment. (2) DSS Model Group: Underwent colitis induction and was orally administered the blank solvent vehicle (equal in volume to the nano-formulation) daily for the 3 weeks prior to the experiment. (3) DSS + Cap-M@mSbots Group: Received daily oral administration of a suspension of capsaicin-loaded Cap-M@mSbots for the 3 weeks prior to the experiment. (4) DSS + Cap-M@mSbots + MF Group: Received daily oral administration of an equivalent dose of Cap-M@mSbots suspension for the 3 weeks prior to the experiment. Immediately after each administration, an external gradient magnetic field (Intensity: 40 mT; Duration: 30 min) was locally applied to the lower abdominal colon projection area to achieve active targeting guidance.

The total experimental timeline spanned five weeks. During the initial three weeks (Days 1–21), all groups received the respective oral gavage interventions as per the protocol every other day. Subsequently, except for the Healthy Control group, acute colitis was induced in all other groups by supplementing their drinking water with 2.0% (w/v) dextran sulfate sodium (DSS) continuously for 10 days (Days 22–31). The Healthy Control group continued to receive normal drinking water. Following the DSS treatment period, all mice resumed normal drinking water for 2 days (Days 32–33). The general condition and body weight of the mice were monitored daily. Throughout the DSS treatment and recovery periods, the disease activity index (DAI) was calculated based on established criteria combining weight loss, stool consistency, and fecal occult/gross blood. At the study endpoint (Day 34), all mice were euthanized. The entire colon was excised, and its length was measured as a macroscopic indicator of inflammation severity. Tissue samples were collected for subsequent histological and biochemical analysis.

2.2.7. Histological Analysis

To evaluate the degree of pathological damage in the colon tissues across different groups, a systematic histological analysis was conducted. The intact colon tissues collected after the experiment, following length measurement, were immediately placed in 4% (w/v) paraformaldehyde phosphate buffer and fixed at 4 °C for 24 h. Subsequently, the tissues underwent dehydration through a graded ethanol series, clearing with xylene, and were routinely embedded in paraffin wax. Using a microtome, the embedded tissue blocks were serially sectioned into 4 μm thick slices, which were then mounted onto poly-L-lysine-coated glass slides. The slides were baked overnight in a 60 °C oven in preparation for staining. First, Hematoxylin and Eosin staining was performed on the tissue sections to assess overall histomorphological changes. The procedure is briefly described as follows: After dewaxing with xylene and rehydration through a graded ethanol series, the sections were stained with hematoxylin for 5 min, rinsed under running water for bluing, and then counterstained with eosin for 1 min. Finally, the sections were dehydrated through a graded ethanol series, cleared with xylene, and mounted with neutral balsam. All stained sections were scanned using a high-resolution slide scanner (Pannoramic MIDI, 3DHISTECH Ltd., Budapest, Hungary) for image acquisition.

2.2.8. Western Blot

Colon tissues were placed in pre-cooled RIPA lysis buffer containing protease inhibitors (e.g., PMSF) and phosphatase inhibitors (e.g., NaF, Na_3_VO_4_). The tissues were fully homogenized using a tissue homogenizer. After homogenization, the lysate was allowed to lyse on ice for 30 min, followed by centrifugation at 4 °C and 12,000× g for 15 min. The supernatant was collected for subsequent analysis. Protein concentration was determined using the BCA assay to ensure the accuracy of further analyses. During quantification, the protein concentration of all samples was precisely adjusted to uniformity using the lysis buffer. An equal amount of total protein sample (typically 30–50 μg) was mixed with 5× loading buffer and denatured by heating at 95 °C for 5 min. Subsequently, the samples were separated by electrophoresis using an SDS-PAGE gel (a 10–12% separating gel is recommended). After electrophoresis, the proteins were transferred onto a PVDF membrane under constant current. Following transfer, the membrane was blocked at room temperature for 1 h with TBST buffer containing 5% skim milk to reduce non-specific binding. The blocked PVDF membrane was then incubated overnight at 4 °C with rabbit anti-phospho-IκBα (Ser32/36) primary antibody and rabbit anti-phospho-p65 (Ser536) primary antibody. The next day, the membrane was thoroughly washed three times with TBST buffer for 10 min each to remove unbound primary antibodies. After washing, the membrane was incubated with an HRP-conjugated goat anti-rabbit secondary antibody at room temperature for 1 h, followed by another three washes with TBST. Finally, the membrane was incubated with enhanced chemiluminescence (ECL) substrate and exposed for development in a chemiluminescence imaging system. The obtained protein band images were quantified for optical density using image analysis software such as ImageJ. The intensity of all target protein bands was normalized to the band intensity of the corresponding internal reference proteins (GAPDH or β-Actin) in the same lane to correct for variations in loading amounts.

2.2.9. ELISA

At the experimental endpoint, whole blood samples were collected from mice in each group via retro-orbital bleeding. After allowing the blood samples to clot at room temperature for 30 min, they were centrifuged at 3000 rpm for 15 min. The upper serum layer was carefully aspirated, aliquoted, and immediately stored at −80 °C for subsequent analysis. The levels of tumor necrosis factor-α (TNF-α) among the different groups were measured using a commercially available mouse-specific enzyme-linked immunosorbent assay (ELISA) kit. The procedure is briefly outlined as follows: standards and test serum samples were added to a 96-well plate pre-coated with the corresponding capture antibody and incubated at room temperature for 2 h. After washing, a biotin-labeled detection antibody was added and incubated for 1 h. Following another wash, a horseradish peroxidase (HRP)-labeled streptavidin working solution was added and incubated for 30 min. After the final wash, a tetramethylbenzidine (TMB) substrate solution was added for color development, followed by incubation at room temperature in the dark for 15–20 min. Finally, a stop solution was added to terminate the reaction. The absorbance of each well was immediately measured at a wavelength of 450 nm using a microplate reader. The concentration of TNF-α was calculated based on a standard curve generated from the absorbance values of the standards, and the results are expressed in picograms per milliliter (pg/mL).

2.2.10. Serum FITC-Dextran Assay

All mice were orally administered a 4 kDa FITC-labeled dextran solution (at a dose of 0.6 mg per 10 g of body weight, dissolved in sterile PBS). After gavage, the mice were returned to their cages for normal housing. Four hours later, whole blood samples were collected via cardiac puncture or retro-orbital venous plexus bleeding. The blood samples were allowed to clot at room temperature for 30 min and then centrifuged at 4 °C and 3000× g for 15 min to separate the serum. The supernatant (serum) was carefully aspirated and transferred to new nuclease-free microcentrifuge tubes, avoiding hemolysis. The fluorescence intensity in the serum samples was measured using a microplate reader (SpectraMax i3x, Molecular Devices, San Jose, CA, USA). The specific steps are as follows: 50 µL of serum was mixed with 150 µL of PBS in a black 96-well plate (1:4 dilution). Detection was performed using preset excitation and emission wavelengths of 485 nm and 528 nm, respectively. For accurate quantification, a standard curve prepared from FITC-dextran of known concentrations was run simultaneously with each assay. The concentration of FITC-dextran in the serum of each sample was calculated based on the standard curve, and the results are expressed in ng/mL. The serum FITC-dextran concentration is positively correlated with intestinal permeability.

2.2.11. Cell Culture and Experimental Treatments

NCM460 human normal colon epithelial cells were routinely cultured in DMEM/F12 medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin and maintained at 37 °C in a 5% CO_2_ atmosphere. To establish an in vitro inflammation model, cells in the logarithmic growth phase were seeded into Transwell plates (for TEER measurement). When the cells reached 80–90% confluence, the medium was replaced with fresh medium containing tumor necrosis factor-α (TNF-α, 10 ng/mL) for 24 h to induce inflammation. Concurrently with TNF-α stimulation, the experimental groups were co-treated with 50 µM capsaicin loaded in Cap-M@mSbots. For the magnetically controlled group, a rotating gradient magnetic field was applied beneath the cell culture, with a measured intensity of 60–80 mT detected within the culture dish. Epithelial barrier function was assessed by measuring transepithelial electrical resistance (TEER). TEER values were measured using an epithelial volt-ohm meter before treatment (0 h) and at 6, 12, and 24 h post-treatment. The percentage change relative to the initial value (0 h) or to the control group was calculated. All experiments were independently repeated at least three times.

2.2.12. Transepithelial Electrical Resistance (TEER) Measurement

After the aforementioned treatments, cells from each group were digested and resuspended at a density of 5 × 10^4^ cells/cm^2^ and then seeded onto the polycarbonate membranes of 0.4 μm pore size Transwell chambers. The medium was changed every two days until the cells reached 80–90% confluence. The Transwell plate was then removed and equilibrated at room temperature in a biosafety cabinet for 15 min. All external magnetic field devices were removed prior to measurement. Measurements were performed using Millicell ERS-2 volt-ohm meter electrodes that had been sterilized with 75% ethanol and rinsed with sterile PBS. The short electrode was placed inside the Transwell chamber, and the long electrode was placed in the outer compartment, ensuring the electrode tips did not contact the membrane or the chamber walls. The resistance value was recorded once the reading stabilized. Each group included six technical replicates. The final transepithelial electrical resistance value was calculated using the formula: TEER (Ω·cm^2^) = (Resistance value of sample well—Average resistance value of blank membrane wells) × Effective membrane area (1.12 cm^2^). Data are presented as the mean of three independent experiments for analysis.

2.2.13. Statistical Methods

Data in biological experiments are expressed as the mean ± standard error of the mean (SEM). Statistical analysis was performed using GraphPad Prism software (version 10.1.2). Comparisons among multiple groups were conducted using one-way analysis of variance (ANOVA). If the ANOVA results were statistically significant, Tukey’s multiple comparison test was further employed for pairwise comparisons between groups. A value of p < 0.05 was considered statistically significant. The specific sample sizes (N), representing the number of independent biological replicates, are indicated in the respective figure legends.

3. Results

The synthesis process of Cap-M@mSbots is illustrated in the schematic diagram in Figure S1. In brief, magnetic nanoparticles are synthesized by the solvothermal method, a silicon source is introduced, and a silica shell layer is grown on the surface of the nanoparticles in situ. Finally, capsaicin is loaded into the interior. As shown in Figure 2A, the results of STEM imaging and EDS mapping characterization for Cap-M@mSbots are presented. In comparison with the Magnetic NPs in Figure S1, it can be observed that Cap-M@mSbots exhibit a core-shell structure. The EDS mapping shows distributions of iron and silicon elements consistent with a core-shell configuration, confirming that Cap-M@mSbots possess a core of magnetite (Fe_3_O_4_) nanoparticles surrounded by a silica shell. Figure 2B and the corresponding Figure S2 display a single Cap-M@mSbot imaged under conditions optimized for observation. Measurement of the shell thickness of Cap-M@mSbots indicates that the silica shell is approximately 20–30 nm thick. Furthermore, a line scan analysis was performed on the single Cap-M@mSbot shown in Figure 2B, with the results presented in Figure 2C. The line scan results suggest that the Cap-M@mSbot is fully encapsulated by the silica shell. The scanning data indicate an approximate atomic ratio of iron to silicon of about 3:1 in the sample, implying that the silica shell accounts for roughly 20% of the total mass of the Cap-M@mSbots. Additionally, the particle sizes of Magnetic NPs, M@mSiO_2_ NPs, and Cap-M@mSbots were measured using Dynamic Light Scattering (DLS), as shown in Figure 2D. The DLS results show average particle sizes of 222.6 nm for Magnetic NPs, 301.1 nm for M@mSiO_2_ NPs, and 298.4 nm for Cap-M@mSbots. Since DLS measures the hydrodynamic diameter of nanoparticles in solution, the obtained values are larger than the dry-state particle sizes obtained from STEM measurements. Furthermore, the zeta potential of the gavage formulation containing Cap-M@mSbots was −18.2 mV. Moreover, due to the stronger hydrodynamic effects of the silica shell, the differences observed in the DLS results between particles with and without the silica shell are more pronounced compared to those from STEM measurements. To verify the magnetic properties of Cap-M@mSbots, hysteresis loops of Magnetic NPs, M@mSiO_2_ NPs, and Cap-M@mSbots were measured using a vibrating sample magnetometer, with the results shown in Figure 2E. The hysteresis loop exhibits no remanence or coercivity, confirming the superparamagnetic behavior of the samples. The saturation magnetization (Ms) values for the three samples are 11, 5, and 7 emu/g, respectively, which are lower than the generally reported range of 30–80 emu/g in the literature [35,36]. Apart from the influence of the mesoporous silica shell, this may also be attributed to the relatively coarse crystal growth resulting from the solvothermal synthesis method, leading to inconsistent internal crystallization directions within the nanoparticles, where partial magnetic fields cancel each other out upon magnetization.

Absorption spectra of Magnetic NPs, M@mSiO_2_ NPs, and Cap-M@mSbots were measured using a UV-Vis-NIR spectrophotometer, as shown in Figure 2F. The spectra in Figure 2F are presented qualitatively, while the corresponding absorbance data are provided in Figure S3. The results reveal a characteristic absorption peak of Capsaicin at a wavelength of 280 nm, which can be used to verify drug loading and release efficiency. The absorption spectrum of Magnetic NPs shows continuous absorption, with intensity gradually increasing as the wavelength decreases. The absorption spectrum of M@mSiO_2_ NPs is similar to that of Magnetic NPs, with a slight enhancement in absorption in the 350–450 nm wavelength range, likely due to the difference in refractive index between the added silica shell and the Magnetic NPs. In the absorption spectrum of Cap-M@mSbots, the Capsaicin absorption peak is observable. Compared to the spectrum of pure Capsaicin, the peak is relatively attenuated, as the continuous absorption of Magnetic NPs masks a significant portion of the peak. The absorption spectroscopy confirms drug loading in Cap-M@mSbots. Next, the drug release performance of Cap-M@mSbots was evaluated. As shown in Figure 2G, the drug release profiles of Cap-M@mSbots under different conditions are presented. The data indicate that after 48 h in water, Cap-M@mSbots release 4.3% of the loaded Capsaicin. In simulated gastric and intestinal fluids, the release proportions after 48 h are 13.8% and 19.8%, respectively. Considering the preventive effect of Capsaicin against colonic inflammation, Cap-M@mSbots are expected to exhibit sustained-release properties. The hydrophobic nature of Capsaicin and the narrow, elongated pores of the mesoporous silica shell contribute to achieving this sustained-release effect.

The motion of micro-nanorobots under magnetic field control is typically achieved using alternating magnetic fields with periodically changing directions, in which the rotational movement of the micro-nanorobots creates a rotating fluid field in the surrounding local space. By leveraging the asymmetry of the fluid field generated by their own asymmetric structures, the micro-nanorobots can achieve net displacement, or they can achieve autonomous motion through interactions with external surfaces such as substrates [20]. For Cap-M@mSbots, due to their extremely small size, it may be challenging to achieve precessional motion similar to that of micron-scale robots under rotating magnetic fields. Therefore, by applying a rotating gradient magnetic field with a relatively high average magnetic field strength, under magnetic fields exceeding 20 mT, the individual Cap-M@mSbot is more likely to aggregate and couple their fluid fields, forming larger linear swarms under magnetic guidance. These linear swarms are more easily driven by magnetic fields and can disperse upon removal of the magnetic field after reaching the target location and simultaneously achieve the effect of phase separation, wherein the Cap-M@mSbot transitions from a uniformly distributed state to a non-uniformly distributed state within the dispersion system. Alternatively, by changing the magnetic field pattern, they can achieve long-term in situ residence. This driving mechanism is illustrated schematically in Figure 3A, where one of the swarms is shown as an aggregate formed by a large number of Cap-M@mSbots. Figure 3B and the corresponding Video S1 demonstrate the process by which Cap-M@mSbots gradually form swarms under the influence of a rotating gradient magnetic field with an average magnetic field strength of 80 mT. Microscopic observations reveal that before the magnetic field driving device is activated, the Cap-M@mSbots are uniformly distributed within the field of view, resulting in an even grayscale image with no noticeable aggregation of Cap-M@mSbots. Upon applying the magnetic field, fluctuations in grayscale appear across the entire microscopic field of view. This occurs because the initially uniformly distributed Cap-M@mSbots simultaneously begin to move and rapidly undergo localized aggregation. Before reaching a scale observable by optical microscopy, these dynamics manifest as apparent changes in the overall grayscale of the field of view. As the aggregation process accelerates, linear aggregates of Cap-M@mSbots of varying lengths emerge in the field of view. These aggregates move forward rapidly in a tumbling motion as swarms, with some swarms achieving speeds exceeding 150 μm/s. Regarding the phenomenon of Cap-M@mSbots aggregating into linear swarms and performing collective motion under a rotating gradient magnetic field, control experiments were designed to investigate the effect of magnetic field strength on the length of the formed swarms. The statistical results are shown in Figure 3C. Within 60 s of activating the magnetic field, the average length of swarms formed under an 80 mT magnetic field was 61.4 μm, while those under a 40 mT field averaged 51.2 μm in length, and those under a 20 mT field averaged 37.9 μm in length. These results indicate a positive correlation between magnetic field strength and the rate at which Cap-M@mSbots form swarms.

To observe the swarming locomotion of the Cap-M@mSbots more clearly, a portion of the field of view in Video S1 was magnified, and motion trajectories were added using ImageJ software to visually demonstrate the movement state of the swarms. The results are shown in Video S2. The motion trajectories were added using the Manuel tracking plugin, with the assistance of a mouse macro to determine the geometric center of the detected Cap-M@mSbot swarms. The results show that the movement speed of the counted swarms exceeds 100 μm per second, and the motion trajectory of the larger swarm exhibits a distinct zigzag pattern. It is suggested that this results from the asymmetric rotational effect of the rotating gradient magnetic field (RGMF). The gradient direction of the RGMF is not necessarily coplanar with the forward direction of the Cap-M@mSbot swarms, which may cause lateral traction during forward movement, resulting in periodic directional deviations. Furthermore, the relationship between the motion speed of swarms and the rotational frequency of the magnetic field under different magnetic field strengths was statistically analyzed, with the results shown in Figure 3D. The data indicate that the motion speed of swarms generated by Cap-M@mSbots under a rotating gradient magnetic field initially increases and then decreases as the magnetic field rotational frequency rises, exhibiting a maximum value in motion speed. When the magnetic field strength is 20 mT, the maximum speed of the swarms occurs at a frequency of 7 Hz, reaching 115.9 μm/s. At a magnetic field strength of 40 mT, the maximum speed is observed at a frequency of 8 Hz, with a speed of 141.2 μm/s. When the magnetic field strength is 20 mT, the maximum speed appears at a frequency of 10 Hz, measuring 180.7 μm/s. Research on magnetically driven micro/nanorobots has shown that for such robots driven by rotating magnetic fields, there is an upper limit to their rotational angular velocity. When this limit is reached, increasing the rotational frequency of the magnetic field does not enhance the rotational speed of the micro/nanorobots, as their rotation can no longer keep up with that of the driving magnetic field due to limitations imposed by factors such as material properties and structural constraints. Consequently, the translational speed of the micro/nanorobots either reaches a maximum or begins to decline. When the motion speed of magnetically driven micro/nanorobots reaches its maximum as a function of magnetic field frequency, the corresponding frequency is defined as the cut-off frequency under those conditions. Beyond this frequency, further increases in rotational speed become ineffective. This corresponds to the frequency values associated with the maximum speeds observed under the various magnetic field strengths in the data above.

As the swarms generated by Cap-M@mSbots continue to move under the driving force of a rotating gradient magnetic field, localized fluid field coupling also occurs between approaching swarms. Unlike the fluid fields that drive swarm formation, the coupling between swarms in this case is insufficient to induce further fusion. Instead, it leads to coordinated motion in chain-like or ribbon-like formations when swarms come into close proximity, as shown in Video S3. To more clearly demonstrate the ribbon-like formations formed by the swarms, segments of Video S3 were captured and the motion trajectories were annotated using ImageJ software. The results are presented in Video S4. Video S4 provides a representative depiction of this collective swarming behavior, which is widely observed in Video S3 as well as in the additional raw data referenced in this paper. To investigate the influence of inter-swarm coordination on motion speed, the effect of the number of swarms participating in coordinated motion on the ratio between the average translational speed of the system and the speed of a single swarm under the same conditions was analyzed. The results are presented in Figure 3E. The data indicate that in over 80% of cases, the speeds of coordinated swarms fall within 80% to 120% of the speed of a single swarm under corresponding conditions. Moreover, the data exhibit randomness, suggesting no strong correlation between the number of swarms participating in coordinated motion and changes in swarm speed. The failure of further collective behavior to produce stronger motion effects may be attributed to the larger size of swarms, which results in greater fluid resistance, thereby limiting further fluid field coupling between different swarms. After verifying the motion performance of the swarms generated by Cap-M@mSbots under magnetic field actuation, it is essential to further discuss the feasibility of swarm motion in intestinal environments. To this end, the motion speeds of Cap-M@mSbots in PBS and simulated intestinal fluid were tested, with the results shown in Figure 3F. The data indicate that at a magnetic field strength of 20 mT, the motion speed of the swarms in the simulated medium is 95.8% of their speed in water. At 40 mT, the speed in the simulated medium is 91.9% of that in water, while at 80 mT, it reaches 97.2% of the speed in water. These test results demonstrate that the motion of Cap-M@mSbots under alternating magnetic field actuation is largely unaffected by biological media. In addition to the influence of the medium, it is also necessary to consider the impact of intestinal geometric structures on the motion performance of swarms generated by Cap-M@mSbots. The intestinal lining contains numerous groove-like structures, and it is essential to verify whether the swarms can surmount such obstacles. As illustrated in Figure 3G, the following experiment was designed to test the obstacle-crossing ability of the swarms: a block of agarose gel with a specific shape was adhered to a culture dish as an obstacle, with the agarose gel approximately 1 mm in height, far exceeding the possible dimensions of swarms formed by Cap-M@mSbots. Figure 3H and the corresponding Video S5 demonstrate the process of swarms generated by Cap-M@mSbots crossing the agarose gel barrier. A specific Cap-M@mSbot swarm arranged in a ribbon-shaped composite from Video S5 was highlighted, with the results shown in Video S6. During the climbing process, most swarms encounter obstacles of varying scales, yet the majority are able to consistently generate net displacement and ultimately overcome the obstacles. The swarms lined up to traverse the obstacle and maintained a speed exceeding 80 μm/s, confirming the feasibility of Cap-M@mSbots’ motion in complex environments.

Next, we systematically evaluated the preventive and therapeutic effects of the magnetically controlled mesoporous silica nano-delivery system loaded with capsaicin (Cap-M@mSbots) on DSS-induced colitis in mice. The experimental flowchart in Figure 4A outlines the key time points: a three-week daily oral pretreatment with mesoporous silica-loaded capsaicin, during which the Cap-M@mSbots + MF group received local external magnetic field guidance after each administration, followed by a 10-day induction period with 2% DSS in drinking water and a 2-day recovery phase. In terms of clinical phenotypes, compared to the normal control group, the DSS model group exhibited significant progressive weight loss starting from the induction period (Figure 4B), a sharp increase in the Disease Activity Index (DAI) score (Figure 4C), and marked colon shortening (Figure 4D). Preventive administration of Cap-M@mSbots partially alleviated these injuries, with its weight loss curve, DAI score, and colon length all significantly better than those of the DSS model group, although the improvement was limited. Furthermore, the Cap-M@mSbots + MF group with added magnetic guidance demonstrated far superior protective effects compared to the Cap-M@mSbots group, exhibiting higher protective efficacy. Specifically, the rate of weight loss in mice showed recovery relative to the Cap-M@mSbots group, accompanied by a reduction in DAI score and further restoration of colon length. These macroscopic data indicate that magnetic field-mediated active targeting significantly enhances the preventive efficacy of the nano-system against DSS-induced colitis.

At the level of inflammation and barrier function, ELISA assays revealed that serum TNF-α levels were significantly elevated in the DSS model group compared to the normal control group, while Cap-M@mSbots treatment effectively reduced these levels. Compared to the Cap-M@mSbots-only group, the Cap-M@mSbots + MF group exhibited the most pronounced inhibitory effect (Figure 4E). Assessment of intestinal barrier integrity yielded consistent conclusions: mice in the DSS model group displayed markedly high intestinal permeability, as indicated by an increased FITC-Dextran ratio. Pretreatment with Cap-M@mSbots reversed the DSS-induced increase in intestinal permeability, and the permeability in the Cap-M@mSbots + MF group was further reduced, showing a significant difference compared to the Cap-M@mSbots-only group (Figure 4F). Furthermore, we validated this conclusion using the colon epithelial cell line NCM460: under TNF-α-induced inflammatory injury, the transepithelial electrical resistance (TEER) ratio increased significantly compared to the normal group, while treatment with Cap-M@mSbots reversed the TNF-α-induced increase in TEER. The addition of magnetic control resulted in a barrier-protective effect significantly stronger than that of the mesoporous silica-encapsulated capsaicin treatment group alone (Figure 4G).

Further, histopathological and molecular experiments provided direct evidence. Representative H&E staining images showed severe disruption of the colonic mucosal structure in the DSS model group, accompanied by epithelial loss and extensive inflammatory cell infiltration, while Cap-M@mSbots pretreatment alleviated the pathological damage. The mucosal structure in the Cap-M@mSbots + MF group was the most intact compared to the Cap-M@mSbots group, with regular crypt morphology and minimal inflammatory infiltration (Figure 5A). At the molecular level, DSS induction significantly activated the NF-κB signaling pathway in colon tissues, manifested by upregulated expression of phosphorylated IκBα and p65 proteins. Cap-M@mSbots treatment suppressed the activation of this pathway, with the Cap-M@mSbots + MF group showing the strongest inhibitory effect, as the expression levels of key proteins were maximally suppressed (Figure 5B,C). In summary, the data from this study, encompassing both in vivo and in vitro experiments, comprehensively demonstrate at macroscopic, molecular, and pathological levels that the mesoporous silica-loaded capsaicin nano-system (Cap-M@mSbots) exerts a clear preventive and protective effect against DSS-induced colitis. Moreover, with the addition of active magnetic field guidance, the therapeutic efficacy of this nano-system is significantly enhanced, as reflected in better weight maintenance, reduced inflammation levels, improved intestinal barrier integrity, and more pronounced histopathological improvement.

4. Discussion

In this study, addressing the critical pharmacological limitations of natural dietary active ingredients—specifically the poor oral bioavailability and mucosal irritation of capsaicin—we successfully engineered a magnetically actuated bio-hybrid nanorobot system, Cap-M@mSbots, for the active interventional prevention of colitis. The experimental results systematically validate the superior performance of this system from physicochemical properties to biological efficacy. Under the actuation of a rotating gradient magnetic field (80 mT), the Cap-M@mSbots overcame the limitation of individual Brownian motion to form reconfigurable, large-scale linear swarms. These swarms exhibited robust collective locomotion with a maximum velocity of 180.7 μm/s and demonstrated the capability to surmount millimeter-scale obstacles, ensuring precise navigation in the complex gastrointestinal environment. Biologically, in the DSS-induced colitis mouse model, the magnetically guided Cap-M@mSbots significantly mitigated pathological damage. Compared to the model group and passive administration, the active targeting strategy minimized weight loss, reduced the Disease Activity Index (DAI), and preserved colon length more effectively. Molecular mechanism analysis further confirmed that the system effectively inhibited the NF-κB signaling pathway, significantly downregulating the expression of pro-inflammatory cytokines (TNF-α) and phosphorylation levels of p65 and IκBα, while reversing the defects in intestinal permeability.

To highlight the innovation of this work, it is essential to compare it with existing therapeutic strategies. Unlike conventional mesoporous silica nanoparticles that rely on passive diffusion or the EPR effect, which are often limited by gastrointestinal peristalsis and mucus turnover, our Cap-M@mSbots utilize active magnetic propulsion to physically reside at the lesion site [37]. Moreover, in contrast to chemically fueled micromotors that typically depend on short-lived, consumable fuels, our magnetic actuation offers a fuel-free and continuously controllable propulsion mechanism suitable for long-term physiological applications. Furthermore, distinct from systemic immunosuppressants that carry risks of broad side effects, or free oral capsaicin that causes direct gastric irritation, the core-shell structure of Cap-M@mSbots acts as a protective shield [38,39,40]. This design avoids premature release in the stomach and ensures localized high-concentration release in the colon. Regarding the biocompatibility and metabolic fate of the nanorobots, a critical consideration is the particle dimension. Compared with magnetic nanomaterials reported in the literature for in vivo applications, Cap-M@mSbots possess a relatively large size. However, because Cap-M@mSbots move within the digestive system, do not actively enter the internal environment, and can be readily metabolized and excreted from the body, their larger size does not present a disadvantage in the context of prophylactic intervention for colitis [41,42,43].

In summary, we developed a novel “active prevention” paradigm for IBD management. By transforming capsaicin from a dietary irritant into a targeted therapeutic agent via swarm-intelligent nanorobots, this study provides a highly efficient strategy for colitis intervention. The integration of the mesoporous silica shell—which enables high loading capacity and sustained release—with magnetic actuation ensures precise accumulation and prolonged retention at colonic lesion sites. This study not only demonstrates the therapeutic potential of Cap-M@mSbots in repairing the mucosal barrier and inhibiting inflammatory pathways but also offers a generic platform for the precise oral delivery of other hydrophobic bioactive compounds, paving the way for future clinical translation in the treatment of gastrointestinal diseases.