Endoscope-assisted magnetic helical micromachine delivery for biofilm eradication in tympanostomy tube

Abstract

Occlusion of the T-tube (tympanostomy tube) is a common postoperative sequela related to bacterial biofilms. Confronting biofilm-related infections of T-tubes, maneuverable and effective treatments are still challenging presently. Here, we propose an endoscopy-assisted treatment procedure based on the wobbling Fe₂O₃ helical micromachine (HMM) with peroxidase-mimicking activity. Different from the ideal corkscrew motion, the Fe₂O₃ HMM applies a wobbling motion in the tube, inducing stronger mechanical force and fluid convections, which not only damages the biofilm occlusion into debris quickly but also enhances the catalytic generation and diffusion of reactive oxygen species (ROS) for killing bacteria cells. Moreover, the treatment procedure, which integrated the delivery, actuation, and retrieval of Fe₂O₃ HMM, was validated in the T-tube implanted in a human cadaver ex vivo. It enables the visual operation with ease and is gentle to the tympanic membrane and ossicles, which is promising in the clinical application.

An endoscopy-assisted treatment procedure based on helical micromachine eradicates biofilm in T-tube.

INTRODUCTION

Bacterial biofilms are complex microbial communities mainly consisting of bacteria cells and the secreted extracellular polymeric substances (EPSs), readily colonizing the surfaces of organs or medical implants to cause intractable and recurring infections (1–3). Otitis media, as a common global health problem more commonly seen in children (4), has been reported to relate to biofilm infections closely (5, 6). Insertion of a tympanostomy tube (T-tube) in the tympanic membrane to drain the middle ear fluid and ventilate the middle ear space is a common ambulatory surgery for alleviating otitis media with effusion (7). However, recurrent infection of the ear refractory to antibiotic treatment and occlusion of the T-tube is among the most common postoperative sequelae, mainly because of the biofilm formation. The biofilm formation in the T-tube contributes to otorrhea and debris deposition, which, in turn, blocks the T-tube. In addition, the dense biofilms render the occlusions more challenging to clear (8, 9). More commonly, the T-tube needs to be removed, leaving the potential need for reinsertion of a new T-tube surgically in the case of recurrent otitis media with effusion. Therefore, prevention and elimination of biofilm are critical to maintain the patency of the T-tube for ventilating the middle ear space. Current interventions focus on preventing biofilm, including optimizing the size of the T-tube or endowing itself with inherent antimicrobial activities (using biofilm-resistant materials, coating with antimicrobial composition, etc.) (10). However, the effect is not evident since biofilms are still readily formed. Regarding the active elimination of biofilm-associated occlusion, ear drops containing antibiotics are commonly used (11). The EPS assists in protecting the bacteria cell and enhancing the mechanical strength of biofilm. These biofilms are 10 to 1000 times more resistant to antibiotics than planktonic bacteria (12), reducing the efficacy of the antibiotics. Aiming at the severely blocked T-tube, the surgeon uses a small tool to clear the occlusion under the endoscopy (13). While this maneuver requires delicate and expert skills to avoid ossicular chain and tympanic membrane injury, it cannot clear the biofilm completely, allowing the biofilm to regrow. Besides, the tethered tools would be obstructive endoscopic view, increasing the operative difficulty in the tortuous external auditory canal and stenotic T-tube. Hence, developing alternative strategies with high efficiency and simple operation to eradicate bacterial biofilms in T-tube are needed clinically.

In recent years, micro/nanomachines (MNMs) remotely actuated by various methods (magnetic, light, and chemical actuation) (14–19) have attracted widespread attention in medical applications due to their capabilities such as navigation in confined spaces, cargo delivery, and performing targeted tasks (20–23). These MNMs (tubular shape, sphere, helix, and reconfigurable shape of swarm) have been applied in the antibacterial field because of the target delivery of antibacterial agents, such as antibiotics (24), reactive oxygen species (ROS) (25), and mechanical disruption (26) for killing the bacteria cells. The plasma membrane of human platelets is also modified onto the MNMs to bind and remove the bacteria cells (27). In addressing the biofilm infections, the MNMs also show advantages of movement-induced penetration and disruption of the stubborn biofilm, promoting the antibacterial agents to act on deeper parts (24–26, 28, 29). The helical micromachines (HMMs), as a typical representative of MNMs, imitating the bacterial flagella (30), are driven by a low-strength rotating magnetic field in three-dimensional (3D) space using the corkscrew motion (31, 32). This motion enables driller-like penetration, which has been applied to remove the blood clots in vessels (33–35). Hence, it might be useful for treating biofilm-induced occlusion in T-tube. Moreover, tissue and cell structures are not damaged within the low-strength magnetic field (36, 37), which is suitable for use in the sensitive and tortuous external auditory canal. The precision movement of HMM enables itself to perform tasks harmlessly in a confined space with high accuracy (38–42). To date, the HMMs have been studied to apply in several organs, such as actuating in the eyeball (43), penetrating gastric mucosa in the stomach (44), and delivering cells to cerebral vessels (45). HMMs are also used in antibacterial or antibiofilm, but most research is limited to simple in vitro studies (46, 47) and lack of progress in practical application. Therefore, developing the HMMs for treating the biofilm infection of implants, such as T-tube in the tympanic membrane, is of great interest in practical applications.

In this work, to address the biofilm infection of T-tube in the tympanic membrane (as shown in Fig. 1A), we report an endoscopy-assisted treatment procedure integrating rapid delivery, actuation (biofilm eradication), and retrieval based on the Fe₂O₃ HMM with peroxidase-mimicking activity. First, we present an efficient fabrication strategy for metal oxide–based HMMs with high throughput and adjustable size based on the microfluidics technique and graphene oxide (GO) template method. The as-fabricated Fe₂O₃ HMM under magnetic actuation exhibits high H₂O₂ catalytic activity to generate ROS, which can kill bacteria cells and degrade the EPS matrix without raising antimicrobial resistance (48–50). Meanwhile, three different motion modes (i.e., straight mode, tilted mode, and wobbling mode) of Fe₂O₃ HMM under magnetic actuation were investigated. The Fe₂O₃ HMM with wobbling motion (32, 51) can induce stronger physical force on the biofilm occlusion and the curved inner surface of the tube and generate higher fluid convections surrounding the wobbling body. Therefore, the wobbling HMM is beneficial to damage the biofilm occlusion into debris with high efficiency and enhance the catalytic generation and diffusion of ROS for killing residual living bacteria cells inside the detached biofilm debris. When the treatment procedure was conducted, the H₂O₂ and HMM were successively delivered to the T-tube under the ENT endoscope. Then, a rotating magnetic field was applied to actuate the HMM for biofilm eradication (Fig. 1B). By the synergies of chemical and mechanical effects (Fig. 1C), the Fe₂O₃ HMM shows the capability of eradicating biofilm. Eventually, the total treatment procedure was verified in the patient’s T-tube implanted in a human cadaver’s tympanic membrane ex vivo. The untethered HMM under magnetic actuation is controllable and gentle, avoiding causing damage to the tympanic membrane and ossicles. Afterward, the Fe₂O₃ HMM can also be retrieved by a tiny permanent magnet with ease. This endoscopy-assisted treatment procedure based on untethered antibacterial Fe₂O₃ HMM accommodates in tortuous and stenotic external auditory canal and enables direct visualization independent of surgeon skill, which is promising against occlusion of T-tube implanted in the tympanic membrane. Moreover, this method could be useful for other implants with a narrow lumen drain or fluidic channel in clinical applications and paves a new route for medical translation of MNMs for biofilm-contaminated implants in vivo.

Fig. 1. The schematic of the treatment procedure.

(A) Schematic diagram of T-tube in the tympanic membrane. (B) The schematic of the treatment procedure for biofilm eradication of the T-tube in the tympanic membrane under the endoscope. (C) Schematic diagram of biofilm eradication with an actuated Fe₂O₃ HMM by the synergy of chemical ROS and mechanical disruption.

RESULTS

Batch fabrication of metal oxide–based HMMs

Figure 2A describes the fabrication route for metal oxide–based HMMs. First, the GO-based HMMs (GO HMMs) were fabricated by the microfluidics technique. There are abundant oxygen-containing functional groups on GO sheets. GO HMMs attract the metal ions via electrostatic interaction when immersed in the metal precursor solution. Afterward, freeze-drying of the GO HMMs absorbing metal ions and annealing them at high temperature flown by airs, the metal salt platelets changed into metal oxides reproducing the micro- and macrostructure of GO host to form metal oxide–based HMMs (52–54). The microfluidic device is designed as illustrated in Fig. 2B. When the strips of GO mixed solution (including GO solution and Fe₃O₄ nanoparticles) entered into the polyvinyl alcohol (PVA) solution, it would form the GO HMMs by the liquid rope-coil effect (55). Figure S1 describes the formation process of GO HMMs, showing that the straight GO strips gradually changed into helical strips until the formation of stable GO HMMs. Besides, the GO HMMs with different lengths and pitches have been fabricated by adjusting the velocity of three kinds of fluids (Fig. 2C). Figure 2D showed the influence of fluid concentrations on the fabrication of GO HMMs. The x-ray diffraction (XRD) curves of the GO HMMs are shown in Fig. 2H. The main crystal peak is due to the content of Fe₃O₄ nanoparticles, which can provide magnetic property for the HMM. To further synthesize the metal oxide–based HMMs, the GO HMMs were immersed into different metal precursor solutions to absorb the corresponding metal ions, including Fe³⁺, Zn²⁺, Cu²⁺, Co³⁺, and Ni²⁺ (Fig. 2E). At high temperatures, the metal ions transformed into metal oxide to form the metal oxide–based HMMs. Figure 2G shows scanning electron microscope (SEM) images of iron oxide–based HMM, in which iron oxide nanoparticles are stacked or nested to form films, and the wrinkle structure was derived from the GO sheets. The SEM mapping images (Fig. 2F) show the elemental distribution of C, O, and Fe in iron oxide–based HMM. The black vacancy of C shows few C contents on the surface of iron oxide–based HMM. The XRD curves in Fig. 2H further indicate the corresponding crystal structure of major Fe₂O₃ and slight Fe₃O₄ on the surface of iron oxide–based HMM. For convenience, the iron oxide–based HMM is defined as Fe₂O₃ HMM. Different metal oxide–based HMMs, including ZnO, CuO, Co₃O₄, and NiO (figs. S2 and S3), were fabricated by a similar method. The magnetic saturation of Fe₂O₃ HMM is up to 35 electromagnetic unit g⁻¹ (fig. S4). Besides, the Fe₂O₃ HMMs in water also were separated quickly in 6 s (inset image of fig. S4), indicating the stable magnetic property. In a word, we propose a universal strategy for fabricating metal oxide–based HMMs with high throughput and tunable sizes.

Fig. 2. Fabrication and characterization of metal oxide–based HMMs.

(A) Schematic diagram of fabrication procedure and mechanism for metal oxide–based HMMs. (B) Optical image of a microfluidic device for GO HMM fabrication. (C) Optical images of GO HMMs with different lengths and pitches. (D) The influence of the concentration of GO and PVA solution on the fabrication of GO HMMs (perfect helix has the whole helical structure; partial helix has a partial helical structure, and straight line has no helical structure). (E) Optical image of GO HMMs immersed in the different metal precursor solution, including Fe³⁺, Zn²⁺, Cu²⁺, Co³⁺, and Ni²⁺. (F) The SEM mapping images of elemental distribution in Fe₂O₃ HMM. (G) The SEM images with different magnifications of Fe₂O₃ HMM. (H) The XRD curves of Fe₂O₃ HMMs and GO HMMs. a.u., arbitrary units.

The motion behaviors and catalytic performance of Fe₂O₃ HMMs

The as-fabricated Fe₂O₃ HMMs are actuated by the rotating magnetic field (30, 56). Here, a rotating spherical magnet is applied to generate the rotating magnetic field to actuate the HMM in a tube (Fig. 3A). This permanent magnet is carried by a 6-DOF robotic arm, which is flexible in the adjustment of the workspace and magnetic field parameters (fig. S5). In this magnetic actuation system, the input parameters, i.e., rotating frequency, the magnetic field strength, and the magnetic field direction, are tuned in a controlled fashion. Figure 3B exhibits the simulated magnetic flux density distribution on the workspace at the working distance (2.5 cm). Figure 3C shows the simulated magnetic field strength with different working distances. Furthermore, different capillary tubes were designed to test the propulsion behavior of Fe₂O₃ HMM (Fig. 3D). The Fe₂O₃ HMMs with different sizes swim across straight capillary tubes with diameters ranging from 0.1 to 0.9 mm (Fig. 3E and movie S1). The translational motion velocity of Fe₂O₃ HMM increases with its size (Fig. 3F). Apart from the straight motion, the Fe₂O₃ HMMs also achieve movement in different curved capillary tubes (Fig. 3, G and H, and movie S2). The magnetic actuation system provides a rotating magnetic field and generates a gradient field to overcome the gravity of Fe₂O₃ HMM, enabling it to swim upward in the capillary tube (Fig. 3H). The movement speed of Fe₂O₃ HMMs is dependent on the magnetic field parameter and the structural parameters of HMMs. Figure 3I delineates that the movement speed of Fe₂O₃ HMMs increases with the input rotating frequency. The speed is up to 0.3 mm/s (0.32 body length) at 8 Hz. Besides, the other metal oxide HMMs containing some Fe₃O₄ nanoparticles also enable swimming under magnetic actuation. We exhibit the movement of Co₃O₄ HMM and CuO HMM in fig. S6. These results validate that a simple rotating magnet system performs remote actuation of these fabricated HMMs in the various straight or curved tiny tubes with adjustable translational speed.

Fig. 3. The motion behaviors of Fe₂O₃ HMMs.

(A) The schematic diagram for generating the rotating magnetic field on the workspace by a spherical magnet. (B) Magnetic flux density on the workspace when the distance between the workspace and magnet is 2.5 cm. (C) Magnetic field strength on the workspace of the different distances between the workspace and magnet. (D) Schematic diagram of the actuation of Fe₂O₃ HMMs in the different tubes. Optical images of the navigation process of Fe₂O₃ HMMs in (E) straight tubes with different sizes (the images come from movie S1, and d represents the diameter of the tube), (F and G) curved tubes (the images come from movie S2, and d represents the diameter of the tube). (H) The change of translational velocity with the increasing size of Fe₂O₃ HMMs at 8 Hz (the dates come from movie S1, and the number of parallel experiments is 3). (I) The change of translational velocity of Fe₂O₃ HMM with the magnetic field frequency (the number of parallel experiments is 3).

The Fe₂O₃ HMMs have peroxidase-mimicking performance, catalyzing the ROS generation for degrading the EPS and killing the bacteria cells. The hydroxyl radical (•OH), as the main ROS in this catalytic reaction, can oxidize the 3,3,5,5-tetramethylbenzidine to produce the blue 3,3,5,5-tetramethylbenzidine oxide. Therefore, the blue product in the tube indicates that Fe₂O₃ HMMs catalyze the generation of ROS when moved between P1 and P2 (Fig. 4A). Moreover, the ROS was detected by the fluorescence probe terephthalic acid. The emergence of the characteristic emission peak around 430 nm further confirmed the production of •OH (Fig. 4B). The catalytic performance of different HMMs was also investigated by measuring the content of 3,3,5,5-tetramethylbenzidine-oxide, which has a characteristic absorption peak at 650 nm. In Fig. 4C, the higher peak indicates better catalytic performance, and it is observed that the Fe₂O₃ HMMs under actuation exhibited a higher catalytic activity than that of the static one, which is ascribed to the fact that the motion of Fe₂O₃ HMMs accelerates the catalytic reaction by enhanced fluid convections. The catalytic performance of actuated Fe₂O₃ HMMs is related to the concentrations of HMMs and H₂O₂ and the pH (fig. S7). The catalytic performance of actuated Fe₂O₃ HMMs has a positive correlation with concentrations of HMMs and H₂O₂. In contrast, the lower pH value contributes to the catalytic reaction. Hence, the acidic microenvironment in the biofilm (57) will contribute to the catalytic reaction of Fe₂O₃ HMMs. Figure S8 illustrates that the HMMs still keep catalytic activity after cycling eight times, demonstrating the high reusability of Fe₂O₃ HMMs. Moreover, we quantitatively assess the degradation capability of Fe₂O₃ HMMs by degrading organic dye methyl blue, indicating that the degradation efficiency reaches 95% in 30 min (fig. S9). The catalytic activity of Fe₂O₃ HMMs was compared to other works in table S1, showing that the designed Fe₂O₃ HMM is an ideal catalyst. In conclusion, the actuated Fe₂O₃ HMMs show good performance in catalyzing the H₂O₂ to generate ROS, which has potential for biofilm eradication. Besides, referring to the concentration of medical H₂O₂ solution (≤3%), we will apply a lower concentration (1%) to conduct the biofilm eradication studies, which is safe for clinical use.

Fig. 4. The catalytic and biofilm eradication performance of Fe₂O₃ HMMs in well plate.

(A) The optical images of catalyzing 3,3,5,5-tetramethylbenzidine by Fe₂O₃ HMM in the tube filled with 1% H₂O₂. (B) The fluorescence emission spectrum of terephthalic acid catalyzed by Fe₂O₃ HMMs for a different time. (C) Absorbance spectra of 3,3,5,5-tetramethylbenzidine catalyzed by pure H₂O₂, static Fe₂O₃ HMMs, and Fe₂O₃ HMMs actuated by the magnetic field [the 3,3,5,5-tetramethylbenzidine solution (2 mg/ml) contains 100 μg of HMMs and 1% H₂O₂, and the catalytic time is 15 min]. (D) Schematic diagram and optical images of biofilm removal process on the plane along a circular path under the magnetic actuation (the diameter of the plate is 14 mm). (E) 3D fluorescent image of the methicillin-resistant S. aureus (MRSA) biofilm treated by different groups (green fluorescent signal represents living bacteria, red represents dead bacteria, and the size of the image is 500 μm by 350 μm by 40 μm). (F) The statistical thickness of MRSA biofilm from image (E) (the number of parallel experiments is 3). (G) The results of MRSA biofilm assessed by crystal violet quantification from different treatment groups (the number of parallel experiments is 3, and the diameter of the plate is 14 mm).

The performance of Fe₂O₃ HMMs for biofilm eradication

The biofilm shows mechanical stability and prevents the permeation of antimicrobials owing to the dense polymeric structure. It is common to use a high antibiotic concentration thousands of times to treat biofilm infection, contributing to serious antibiotic resistance (12). The designed Fe₂O₃ HMMs with good catalytic performance and controllable actuation are promising for the targeted biofilm treatment. Our performance evaluation focuses on the Gram-positive methicillin-resistant Staphylococcus aureus (MRSA). MRSA, a common bacteria with drug resistance that causes serious infections in the human body, is also reported to have a close relationship with otitis media and T-tube occlusion (58, 59). Therefore, it is important to develop an effective strategy to combat MRSA biofilm infection. The MRSA biofilm is cultured in the 24-well plate; then, the Fe₂O₃ HMM is steered for biofilm eradication. Fe₂O₃ HMM is actuated under the rotating magnetic field generated by the moving spherical permanent magnet (fig. S10). Then, the Fe₂O₃ HMM rolls on the planar surface for biofilm elimination along the defined route; for instance, the circle trace in Fig. 4D. Figure 4E shows the 3D fluorescent microscopy of biofilms treated for 30 min by different groups, including pure H₂O₂, the static Fe₂O₃ HMMs in H₂O₂, actuated Fe₂O₃ HMMs in water (without H₂O₂), and the actuated Fe₂O₃ HMMs in H₂O₂. Compared to the area and thickness of biofilms treated by different groups (Fig. 4, E and F), the pure chemical •OH (static Fe₂O₃ HMMs in H₂O₂) and pure mechanical force (actuated Fe₂O₃ HMMs in water) also can eliminate some parts of the biofilm. Besides, the actuated Fe₂O₃ HMMs in H₂O₂ show the best performance for biofilm eradication due to the fact that the motion of Fe₂O₃ HMMs brings mechanical disruption to biofilm and contributes to the diffusion of ROS. There is no green fluorescence of biofilm treated by the actuated Fe₂O₃ HMMs in H₂O₂, indicating that it can clear the MRSA biofilm in 30 min. Meanwhile, the change in area and thickness of biofilms with the treatment time further demonstrate the same result (fig. S11). Furthermore, the viability of the bacterial cells of residual biofilm treated by actuated Fe₂O₃ HMMs in H₂O₂ was characterized by transmission electron microscopy (TEM) micrograph (fig. S12), and the most damaged bacteria cells indicated that the bacteria cells were killed after treatment. To further investigate the biofilm eradication performance, the colony counting method and crystal violet quantification (fig. S13 and Fig. 4G) were applied to test the biofilm treated by different groups for 30 min. Both show that the actuated Fe₂O₃ HMMs in H₂O₂ have the best performance for biofilm eradication. Moreover, fig. S14 exhibited that the Fe₂O₃ HMMs perform similarly well in confronting the Gram-positive Escherichia coli biofilm. In conclusion, the actuated Fe₂O₃ HMMs in H₂O₂ demonstrate the high capability of eradicating MRSA and E. coli biofilms by the synergistic effect between ROS and mechanical disruption, acting as a promising broad spectrum antibacterial agent without leading to antibiotic resistance.

Biofilms tend to grow in hard-to-reach and confined spaces, such as a 3D narrow tube, leading to occlusion of the tube. Therefore, we cultured MRSA biofilm in the tube and steered the Fe₂O₃ HMM in 1% H₂O₂ to dredge the biofilm occlusion. We investigated the biofilm occlusion disruption by applying Fe₂O₃ HMM with three different motion modes (32, 51). Three motion modes of the HMMs are shown in the schematic diagram of Fig. 5 (A and B), including HMM in a tube without tilted angle along the axis of the tube (i.e., mode 1, namely, straight mode), with a constant tilted angle (i.e., mode 2, namely, tilted mode), and with wobbling motion (i.e., mode 3, namely, wobbling mode). The applied rotating spherical magnet and magnetic field waveforms are shown in figs. S15 and S16. Movie S3 shows the process of the HMMs with three different motion modes damaging the biofilm occlusion in a tube and the real-time magnetic field distributions. Among them, the HMM with a wobbling motion (i.e., mode 3) has the highest efficiency in disrupting the biofilm occlusion into debris, as shown in Fig. 5C and fig. S17. In addition, the biofilm removal efficiency is also related to the size ratio between Fe₂O₃ HMM and tube. Figure 5D and fig. S18 show that it is better to dredge biofilm occlusion in the tube (with an inner diameter of 0.9 mm) when the ratio ranges from 0.5 to 0.62. From their motion behavior under the magnetic field (Fig. 3F), the Fe₂O₃ HMM with a bigger size exhibits a higher velocity of translational movement. On the contrary, the smaller the size ratio between Fe₂O₃ HMM and tube, the greater the wobbling motion of Fe₂O₃ HMM produced. Therefore, optimizing the size ratio and motion mode of HMMs are both essential to improve the biofilm occlusion removal efficiency.

Fig. 5. The biofilm eradication performance of Fe₂O₃ HMMs with different motion modes in the tube.

(A) Schematic diagram and optical images of HMM with different motion modes in a tube, including (i) straight mode and (ii) titled mode. (B) Schematic diagram and optical images of HMM with (iii) wobbling mode. (The inner diameter of the tube is 0.9 mm, and the optical images come from movie S3. The applied rotating magnetic fields on the HMMs with three motion modes are illustrated in fig. S11.) (C) The relationship between the biofilm removal efficiency and the motion mode of the Fe₂O₃ HMM (the data are related to fig. S12, and the number of parallel experiments is 4). (D) The relationship between the biofilm removal efficiency and the size ratio of the Fe₂O₃ HMM and tube (the data are related to fig. S13, and the number of parallel experiments is 3). Asterisks denote the level of significance: *P < 0.05, ***P < 0.001, and ****P < 0.0001. Simulation of shear stress distribution on (E) the interface of biofilm occlusion and (F) in the tube when the wobbling Fe₂O₃ HMM approaches the biofilm occlusion. (G) Simulation of the fluid velocity while the wobbling Fe₂O₃ HMM approaches the biofilm occlusion in the tube. (All the simulating parameters and state are related to motion mode 3, and the inner diameter of the tube is 0.9 mm.)

To further analyze the biofilm occlusion removal effect induced by the motion behavior of Fe₂O₃ HMM in the tube, the spatial distribution of fluid velocity and fluid shear stress was simulated, while Fe₂O₃ HMM with modes 1 and 3 removed biofilm occlusion in the tube. The whole moving process of Fe₂O₃ HMM was conceived as steady motion. On the basis of the momentum equation, the fluid dynamics in the tube are analyzed using the finite element method, and all the parameters are related to motion modes 1 and 3 of the experimental test. The momentum equation in the analysis is shown as follows (60)

$$\mathrm{\rho } \frac{\mathrm{\partial }u}{\mathrm{\partial }t}=\mathbf{F}+\nabla ·(-p\mathbf{I}\mathit{+}\mathrm{\mu }(\nabla \mathbf{u}+{(\nabla \mathbf{u})}^{\mathrm{T}})),$$ (1)

where u represents the flow velocity in the tube, F represents the body force density of HMM, p represents the pressure, I represents the identity tensor, μ represents the dynamic viscosity, and ρ is the fluid density. As the agitation of local flow is induced by rotating HMM, the fluid velocity in the fluid-solid interface is governed by

$$\mathbf{u}=2\mathrm{\pi }f \mathbf{r}$$ (2)

where u represents the flow velocity, f represents the rotational frequency, and r represents the coordinate vector of HMM. In Fig. 5 (E and F), when the wobbling HMM approached the interface of biofilm occlusion, the obvious fluid shear stress was observed on the interface of biofilm occlusion and on the boundaries of the tube. On the interface of biofilm occlusion, it produces the strongest fluid shear stress (approximately 450 mPa) around the HMM head (the reddest circles in Fig. 5E). Besides, the increasing fluid convections surrounding the wobbling HMM (as the fluid velocity distribution in Fig. 5G) contribute ROS’s catalytic generation and diffusion for killing the bacteria cells. Figure S14 demonstrates the fluid shear stress and fluid velocity distribution of Fe₂O₃ HMM with motion mode 1. The strongest shear stress on the interface of biofilm occlusion and on boundaries of the tube is up to approximately 340 (fig. S19B) and 200 mPa (fig. S19A), respectively, and the strongest fluid velocity surrounding the HMM is up to approximately 10 mm/s (fig. S19C). Those of wobbling HMM in the corresponding region are up to approximately 450 mPa (Fig. 5E) and 480 mPa (Fig. 5F) and 14 mm/s (Fig. 5G), respectively. By comparison, the fluid convection and fluid shear stress in the space of the tube induced by wobbling Fe₂O₃ HMM are stronger than those in mode 1.

In addition to fluid shear stress, the applied pressure by Fe₂O₃ HMM contacting on the biofilm occlusion produced in the rotating and moving forward process is also important for biofilm removal. As shown in fig. S20, the applied pressure induced by Fe₂O₃ HMM contacting on the biofilm is divided into two directions: the forward pressure generated by forward motion of HMM and the lateral pressure generated in the rotating motion of Fe₂O₃ HMM. The forward pressure and lateral pressure generated by the Fe₂O₃ HMM are calculated approximately by the Euler-Bernoulli beam deflection (61). As shown in text S1 and fig. S20, the estimated forward pressure by Fe₂O₃ HMM is 1.5 Pa (straight mode) and 18.4 Pa (wobbling mode), and the estimated lateral pressure by Fe₂O₃ HMM is 17.6 Pa (straight mode) and 65.9 Pa (wobbling mode), respectively. The contacting pressure on biofilm induced by wobbling Fe₂O₃ HMM is also larger than that in straight mode. These results show that the contacting pressure and fluid shear stress induced by wobbling Fe₂O₃ HMM can act together to mechanically destroy the biofilm occlusion. It also explains why the wobbling Fe₂O₃ HMM performs better in disrupting the biofilm occlusion into debris.

To eradicate the biofilm occlusion completely, the viability of the bacteria cells in detached biofilm debris cannot be neglected. Here, the whole biofilm eradication process in the tube is divided into two steps: steering the Fe₂O₃ HMM in 1% H₂O₂ to disrupt the biofilm occlusion into debris in stage I and keeping the actuation of Fe₂O₃ HMM to kill the bacteria cells in the detached biofilm debris in stage II. In stage I, the Fe₂O₃ HMM swam to approach the occlusion and disrupt the biofilm into debris similar to a driller (Fig. 6A). The removal process for the biofilm occlusion with a length of 2.7 mm by the wobbling Fe₂O₃ HMM needs approximately 1 min (movie S4). The images in Fig. 6B show the biofilm occlusion before treatment and biofilm debris after treatment. Last, the biofilm debris can be rinsed easily with a gentle flow of water, and Fig. 6B shows the clean tube after rinsing. The result of colony counting in Fig. 6C further validates that there are no living bacteria cells on the inner surface of the clean tube. These results indicate that the biofilm occlusion was disrupted and eradicated from the inner surface of the tube in stage I. To further kill the bacteria cells inside the detached biofilm debris, Fe₂O₃ HMM was steered back and forth repeatedly for 15 min in the debris in stage II. The image in Fig. 6D shows the fluorescent image of biofilm debris after treatment by different motion modes. There is no green fluorescence after treatment by the wobbling Fe₂O₃ HMM (mode 3). The fluorescent microscopy in fig. S21 shows the change of biofilm debris with treatment time, and counting the green fluorescence in Fig. 6E also exhibited that the HMM with motion mode 3 has the best performance for killing bacteria cells in biofilm debris. The colony counting in Fig. 6F further validates the result. In conclusion, these results indicate the high efficiency of the wobbling Fe₂O₃ HMM for both dredging the biofilm occlusion and killing bacteria cells in detached biofilm debris. In actual application, the treatment containing stage I is applicable to dredge the biofilm occlusion in some industrial devices without considering the viability of biofilm debris. The treatment containing both stages I and II can be applied in some medical equipment with the goal of killing all the bacteria cells in the biofilm. For instance, aimed at the biofilm infection in T-tube that demands both dredging biofilm occlusion and eradicating bacteria cells, the wobbling Fe₂O₃ HMM is applicable.

Fig. 6. The analysis of the biofilm eradication process in the tube by Fe₂O₃ HMMs.

(A) Schematic diagram of biofilm disruption by wobbling Fe₂O₃ HMM in the tube. (B) Optical images of MRSA biofilm occlusion before damaging, biofilm debris after damaging (the images come from movie S4), and tube after rinsing with the gentle flow. (The inner diameter of the tube is 0.9 mm.) (C) The viability of bacteria cells on the inner wall of the tube after rinsing. It was assessed by the colony counting method (the diameter of the plate is 90 mm, and the number of parallel experiments is 3; ND, not detected). (D) Fluorescent images of biofilm debris treated by different motion modes for 15 min. (E) The change of green fluorescent intensity of the biofilm debris treated by different motion modes. (The green fluorescent intensity is calculated from the fluorescent image in fig. S16, and the number of parallel experiments is 3). (F) The viability of bacteria cells of debris treated for 15 min. It was assessed by the colony counting method. (The number of parallel experiments is 3; ND, not detected; asterisks denote the level of significance: *P < 0.05 and ***P < 0.001.).

The treatment procedure for biofilm eradication in T-tube ex vivo

Insertion of T-tube into the tympanic membrane is a procedure to treat otitis media with effusion, whereas recurrent infection of the ear refractory to antibiotic treatment and occlusion of the T-tube are the postoperative complications related to the biofilms. Even some patients suffered from the T-tube occlusion as early as a month after tube placement (7), which seriously affects the curative effect of the surgery that demands the T-tube to be present for over 6 months. Antibiotics are usually used to prevent biofilm occlusion after the surgery, but they cannot prevent infection in the longer term, and the use of antibiotics commonly leads to antimicrobial resistance. The established biofilm occlusion usually is resistant to local or systemic antibiotics. Sometimes, patients need to attend to an otolaryngologist to have the tube blockage clearance using specialized mechanical tools. Nonetheless, the mechanical tool is unable to clear the biofilm completely within the T-tube, resulting in the regrowth of biofilm soon after. Patients most likely need to undergo a surgical procedure to replace it under anesthesia. In our work, a simple office-based procedure based on the Fe₂O₃ HMM is proposed to address the biofilm infection in T-tube under the assistance of an ENT endoscope. The wobbling Fe₂O₃ HMM enables rotation to approach the entire inner curved wall of the T-tube, which is more appropriate for treating the biofilm in confined 3D space. On the other hand, the designed treatment procedure combines with the endoscope, enabling the endoscopy imaging–guided operation with ease in real time.

Figure 7A reveals the structure of the external auditory canal, the tympanic membrane, and the T-tube. The applied T-tube in our research has a length of 2.5 mm and an inner diameter of 1 mm (the top right image in Fig. 7B). Figure 7B and fig. S22 show the ex vivo experiment setup, including the endoscope, magnetic actuation system, and cadaver head in our study. The treatment procedure under the endoscope was demonstrated in Fig. 7C and movie S5. The endoscope was first inserted into the human cadaver’s external auditory canal to identify the position of the T-tube in the tympanic membrane. Then, the H₂O₂ and Fe₂O₃ HMM were delivered successively into the T-tube by a self-made catheter. Different from squeeze injection of H₂O₂, the Fe₂O₃ HMM swam through the catheter into the T-tube under the magnetic actuation, preventing the HMM from dropping into the middle ear. Here, the concentration of H₂O₂ (1%) used was lower than that of medical H₂O₂ (3%), which is considered safe for the external auditory canal. The wobbling Fe₂O₃ HMM swam in the T-tube under magnetic actuation, and the biofilm was detached from the inner surface of the T-tube and then damaged into tiny biofilm debris. Meanwhile, the generated ROS could kill the bacteria cells inside the biofilm debris. Eventually, the biofilm in the T-tube was eradicated through the synergistic effect of ROS and mechanical disruption. Here, the infected T-tube was collected from the patient. The biofilm in the patient’s T-tube mainly consisted of mainly S. aureus, which is consistent with the literature (58, 59). Last, retrieval of the HMM was done with ease by a tiny permanent magnet (Fig. 7C). Before treatment, there was visible living biofilm on the inner surface of the patient’s T-tube (Fig. 7D). After treatment, the T-tube was extracted from the tympanic membrane to examine the biofilm. There is no green fluorescence on the inner surface of the T-tube after treatment (Fig. 7E), indicating that the biofilm was eradicated from the T-tube. Moreover, the viability of detached biofilm debris was assessed by the live/dead staining and colony counting method (fig. S23), indicating that the bacteria cells inside biofilm debris were also killed. In addition, the T-tube was cultured for an additional 24 hours, and there was no observable regrowth of the biofilm. To further assess the performance of biofilm eradication by the treatment procedure, we cultured the MRSA on the T-tube to conduct in vitro experiments. The images of the inner surface of the T-tube are shown in fig. S24A. The results of the live/dead staining, colony counting, and crystal violet quantification (fig. S24, B to D) indicated that the MRSA biofilm also was eradicated. These results exhibited the high performance of the proposed treatment procedure for biofilm eradication in T-tube.

Fig. 7. The treatment procedure for biofilm eradication in T-tube on cadaver head ex vivo.

(A) The schematic diagram of the external auditory canal and T-tube in the tympanic membrane. (B) The optical images of ex vivo experiment setup, including the endoscope, magnetic actuation system, cadaver head, and T-tube. (C) The optical images of the treatment procedure for biofilm eradication of T-tube implanted in human cadaver’s tympanic membrane under the endoscope (the images come from movie S5). (D) The top view of the patient’s T-tube (i), the inner surface of the T-tube (ii), and the fluorescent image of the biofilm in the T-tube before treatment (iii and iv). (E) The cross-sectional view of the patient’s T-tube (i). the inner surface of the T-tube (ii), and the fluorescent image of the biofilm in T-tube after treatment (iii) and after an additional 24-hour culture (iv).

In fig. S25A, the Fe₂O₃ HMM is actuated and confined in the T-tube because of the constraints of liquid surface tension, indicating a low risk of dropping into the middle ear during normal use. In fig. S25B, the Fe₂O₃ HMM is attracted by the magnet to overcome the liquid tension and is retrieved successfully from the T-tube. Last, the cytotoxicity of the Fe₂O₃ HMMs with different concentrations was assessed on normal cells [human umbilical cord endothelial cells (HUVECs)], as shown in fig. S26. It was found that the cell viability decreases slightly when the concentration of Fe₂O₃ HMM is beyond 100 μg/ml, and the cell viability is still much larger than 70% when the concentration of Fe₂O₃ HMM is up to 400 μg/ml, demonstrating the negligible cytotoxicity according to ISO 10993-5. These results show that the cytotoxicity of the Fe₂O₃ HMM in our required dose can be neglected. To summarize, the designed treatment procedure based on Fe₂O₃ HMM exhibits biocompatibility and good performance of biofilm eradication inside the T-tube in the ex vivo experiments. This synergy between chemical ROS and mechanical disruption of Fe₂O₃ HMM avoids raising the antimicrobial resistance to bacteria. Compared to tethered mechanical tools, this mechanical force induced by the magnetic actuation of untethered HMM is controllable and gentle, avoiding causing damage to the tympanic membrane and ossicles, which is independent of surgeon skill. Besides, this method based on untethered HMM is a simple office-based endoscopic procedure, functional in the tortuous and stenotic external auditory canal, and enables visualization treatment with ease. Moreover, this proposed treatment procedure not only can eradicate biofilm in the T-tube but also is applicable to treat biofilm infection or perform some tasks in the other implants with the tiny tubular drain in clinical applications.

DISCUSSION

This work reports the Fe₂O₃ HMM with high catalytic activity to treat T-tube biofilm infection with a wobbling motion and endoscopic assistance. The fabricated Fe₂O₃ HMM has high performance in generating the ROS, which can kill the bacteria cells and degrade the EPS of biofilm. Compared to ideal corkscrew motion, the wobbling motion of Fe₂O₃ HMM induces stronger interaction force and fluid convection in the space of the tube, which enables the eradication of biofilm occlusion in two stages. i.e., (i) to damage the biofilm occlusion into debris and (ii) to enhance the catalytic generation and diffusion of ROS to kill residue living bacteria cells inside the detached biofilm debris with high efficiency. Consequently, the wobbling Fe₂O₃ HMM can eradicate MRSA, E. coli, and patient’s biofilm through the synergy of chemical ROS and mechanical disruption, promising as a broad-spectrum antibacterial agent without increasing antimicrobial resistance. Last, the treatment procedure is compatible with high maneuverability, integrating rapid delivery, actuation, and retrieval of Fe₂O₃ HMM under the assistance of an endoscope. This treatment procedure was validated in the T-tube implanted in the human cadaver’s tympanic membrane ex vivo. In addition, it is gentle to the tympanic membrane and ossicles, independent of surgeon skill. Overall, the endoscopy-assisted treatment procedure based on Fe₂O₃ HMM is a promising strategy to combat T-tube biofilm infection, which is of high potential in clinical translation.

The MNMs are of notable advantages in combating bacterial biofilm infection, which has attracted broad research. However, most research on antibacterial MNMs is limited to in vitro studies, and there is a lack of progress in practical application. The original intention of our work is expected to promote the reasonable clinical translation of antibiofilm MNMs. Thus, we cooperate deeply with otolaryngology doctors and learn the clinical problem and treatment requirements for the biofilm infection of T-tube. Then, we developed the endoscopy-assisted treatment procedure based on Fe₂O₃ HMM to solve this problem. The treatment procedure with high maneuverability can meet the requirements of otolaryngology doctors, which is validated in human cadaver ex vivo. The human cadaver provides the actual environment of the external auditory canal, further enhancing the reliability of our research results. We believe that our work advances the cooperation of experts between material, engineering, and medicine, paving the way for the clinical translation of MNMs.

In addition, some points should be further promoted and explored. The first point is the general approach to fabricating metal oxide HMMs. We believe that the other metal oxide HMMs (not limited to CO₃O₄, ZnO, CuO, and NiO HMMs) will also be fabricated by this general approach. Moreover, the different metal oxide–based HMMs also hold great potential in some specific applications. For example, ZnO piezoelectric biomaterial has good biocompatibility to generate piezopotential to regulate the behaviors and fate of stem cells (62). Co₃O₄, CuO, and NiO have been reported as important electrode materials (63), which may be applied to the micromotor-induced supercapacitor system. The second point is on the biofilm debris. In our work, we first disrupted the biofilm occlusion into biofilm debris and then kept the actuation of HMM to further kill the bacterial cells in the broken biofilm debris. However, if the broken/killed biofilm debris can be collected appropriately, it may reduce the treatment time and improve the removal efficiency. Catheter technology is a mature technology and has been successfully used to deliver and adsorb liquid (blood, water, and mucus) in the clinic. Moreover, the intravascular magnetic catheter has also been reported to have great potential in the retrieval of nanoparticles (64). We anticipate that catheter technology can be explored to collect broken/killed biofilm debris in the future, strengthening the maturity of the treatment procedure for biofilm.

MATERIALS AND METHODS

Synthesis of metal oxide–based HMMs

GO was synthesized by the modified Hummer’s method (65). The Fe₃O₄ nanoparticles were synthesized via the solvothermal method (26). Mix the GO solution (10 mg/ml) and Fe₃O₄ nanoparticles (10 to 40 weight %) to obtain GO mixed solution. Three kinds of fluids—GO mixed solution, n-hexadecane, and 10% PVA solution—were pumped into the capillary microfluidic device through the corresponding channel, respectively. GO HMMs were fabricated by this capillary microfluidic device, and the size of GO HMMs can be adjusted by changing the velocity of three kinds of fluids. Then, the GO HMMs were immersed into different metal precursor solutions to absorb the corresponding metal ions overnight. The concentration of FeCl₃, ZnCl₂, CuCl₂, CoCl₂, and NiCl₂ ranged from 0.5 to 2 M. Next, GO HMMs adsorbing metal ions were freeze-dried and further annealed at 200°C for 2 hours and 500°C for 0.5 hour flown by air to acquire different oxide-based HMMs.

Magnetic actuation of the Fe₂O₃ HMM

The magnetic actuation was conducted in the rotating magnetic field generated by a rotating spherical magnet. The Fe₂O₃ HMMs were put into the capillary tube with different shapes and sizes filled with deionized water. Apply the magnetic actuation system to actuate the Fe₂O₃ HMM with a constant working distance. An optical camera was mounted to track and record data visually.

Catalytic performance test

We added 3,3,5,5-tetramethylbenzidine (2 mg/ml) and H₂O₂ (0.1 to 3%) into the acetic acid buffer (pH ≈ 5.2) in the tank and then added 10 to 100 μg of Fe₂O₃ HMMs into the tank. Last, the HMMs were actuated to contribute to the catalytic reaction. Varying the contents of HMMs, H₂O₂, and pH, the corresponding absorption spectrum was obtained. The ROS was detected by using terephthalic acid as a probe. Terephthalic acid can react with ROS to generate 2-hydroxylterephthalic acid with a fluorescence emission peak of about 426 nm. HMMs (100 μg), H₂O₂ (1%), and terephthalic acid (10 mM) were reacted in a tank for 10 and 30 min, respectively. Next, the fluorescence emission spectrum of the resulting solution was tested.

The degradation capacity of Fe₂O₃ HMMs for methylene blue

Add methylene blue (10 mg/liter) and H₂O₂ (1%) into the acetic acid buffer (pH ≈ 5.2) in the tank and then add 100 μg of Fe₂O₃ HMMs into the tank. Last, the HMMs were actuated to contribute to the degradation reaction.

Biofilm eradication in 24-well plates

Culture the biofilm

MRSA or E. coli was dispersed in Luria-Bertani (LB) medium in Pyrex flasks; then, the bacteria cells were cultured overnight at 37°C under shaking at 300 rpm. Then, we obtained MRSA or E. coli bacteria cells by centrifuge separation (5000g for 3 min) and washed them with phosphate-buffered saline (PBS). Next, the bacteria cells (10⁷ colony-forming units/ml) were dispersed in LB media. Then, the dispersion solution was added into 24-well plates, and capillaries with different shapes were cultured for 48 hours at 37°C. The biofilm will grow on the surface of the corresponding container.

Biofilm eradication

The biofilms in 24-well plates were treated by Fe₂O₃ HMMs. The Fe₂O₃ HMMs and H₂O₂ (1%) were added into the corresponding container to be actuated by the magnetic actuation system. The biofilm removal efficiency was evaluated by fluorescence imaging of live/dead staining, colony counting method, and crystal violet quantification. All the experiments were repeated three times. To acquire fluorescence imaging, live/dead staining used the BacLight bacterial viability kits, consisting of SYTO-9 and propidium iodide. In the colony counting method, the residual biofilms after treatment were dispersed into suspensions, and suspension was spread on solid medium and cultured at 37°C for 12 hours to count the colony. In crystal violet quantification, the residual biofilms in 24-well plates after treatment were washed using PBS (pH ≈ 7.2), and then 200 μl of methyl alcohol was added to immobilize biofilms. After 15 min, dry the biofilm. Next, add 200 μl of 1% crystal violet solution to stain biofilms. After 5 min, remove the excess crystal violet and wash with deionized water for three times. Dry the biofilm again to obtain the corresponding image. Eventually, the immobilized crystal violet was released by 200 μl of 33% acetic acid for 30 min; then, the absorbance was measured at 590 nm. Obtain TEM images of bacteria cells; the bacteria cells were treated with 2.5% glutaraldehyde for 4 hours and then dehydrated by graded ethanol (25, 50, 75, 100, and 100%) for 10 to 15 min, respectively.

Biofilm eradication in tubes

Culture the MRSA biofilm according to the above method in different tubes. The Fe₂O₃ HMMs and H₂O₂ (1%) were added to the tube to be actuated. Investigate the HMM with different motion modes. Magnetic field strength and frequency remain constant, adjusting the field direction by changing the direction of the magnet to acquire different motion modes. To test the viability of bacteria on the inner surface of the tube after treatment, The tube was under ultrasonic washing for 3 to 5 min; then, the colony counting method was used to test the bacteria cell in the washing water. The viability of bacteria in biofilm debris was tested by fluorescence live/dead staining and colony counting method.

Biofilm eradication of T-tube in human cadaver ex vivo

The human cadaver and patient’s infected T-tube were provided by Prince of Wales Hospital. Then, the T-tube was implanted into the tympanic membrane of the human cadaver. First, the 3 mm 0° ENT endoscope in the external auditory canal was used to visualize the T-tube. Then, 1% H₂O₂ and Fe₂O₃ HMM were delivered into the T-tube by a self-made catheter under endoscopic guidance. The rotating magnet steered the Fe₂O₃ HMM in T-tube from outside the cadaver head. After treatment, the Fe₂O₃ HMM was retrieved by a tiny magnet bar with a diameter of 1 mm. All the procedures and video recordings were conducted through the endoscope. Last, the T-tube was extracted from the tympanic membrane to observe and test for the biofilm.

Cytotoxicity assay of Fe₂O₃ HMMs

The viability of HUVEC cells was evaluated by MTS assay after treating by Fe₂O₃ HMMs. Typically, a suspension of HUVECs was added into the 96-well plate and incubated at 37°C with 5% CO₂ incubator for 24 hours. Assessing the influence on the viability of the cell, Fe₂O₃ HMMs with different concentrations were added into 96-well plate, and HUVECs were treated for 30 min. The cells were incubated for 24 hours again. Afterward, MTS solution (10 μl, 5 mg/ml) was added to the treated cells, and the cells were further incubated for 2 hours. Next, the HMM samples were removed by the magnet. Last, the absorbance was measured at 490 nm by a microplate reader.

Supplementary Materials

This PDF file includes:

Supplementary Text S1

Figs. S1 to S26

Table S1

References

Other Supplementary Material for this manuscript includes the following:

Movies S1 to S5