Magnetically driven active topography for long-term biofilm control

Abstract

Microbial biofilm formation on indwelling medical devices causes persistent infections that cannot be cured with conventional antibiotics. To address this unmet challenge, we engineer tunable active surface topographies with micron-sized pillars that can beat at a programmable frequency and force level in an electromagnetic field. Compared to the flat and static controls, active topographies with the optimized design prevent biofilm formation and remove established biofilms of uropathogenic Escherichia coli (UPEC), Pseudomonas aeruginosa, and Staphylococcus aureus, with up to 3.7 logs of biomass reduction. In addition, the detached biofilm cells are found sensitized to bactericidal antibiotics to the level comparable to exponential-phase planktonic cells. Based on these findings, a prototype catheter is engineered and found to remain clean for at least 30 days under the flow of artificial urine medium, while the control catheters are blocked by UPEC biofilms within 5 days.

Fig. 1. Engineering magnetically responsive PDMS pillars.

a Schematic of the steps for fabricating active surface topographies. b Settlement of MNPs through silicone elastomer mixture (base: curing agent = 10:1) driven by a 1 T magnet. The images show the original MNPs in a PDMS mixture and the settling of MNPs with a magnet. Each condition was tested with three replicates (error bar = standard deviation; n = 3). Source data of Fig. 1a, b are provided in the Source Data file. c Schematic of pillar bending in response to an external electromagnetic field. The location of MNPs in the pillar tips is indicated in the schematic, which was verified by imaging pillars loaded with fluorescently labeled MNPs (green signals shown in Fig. 1c). The profile of the pillar is highlighted with white dotted lines.

Fig. 2. Active pillars exhibited profound antifouling effects against UPEC biofilms.

Representative fluorescence images of biofilms on flat controls, static controls, and PDMS surfaces with active topographies. The pillars were 10 µm tall with a diameter of 2 µm and inter-pillar distance of 2 µm. Active surface topographies were operated in three modes for biofilm prevention and removal, including continuous beating for biofilm prevention, on-demand removal of mature biofilms (only actuated for 3 min after 48 h of biofilm growth), and a sequential combination of these two treatments. Biofilm cells were labeled with STYO^®9. The biomass was quantified using COMSTAT. The samples in the bar graph are indicated with the pattern labels above the corresponding fluorescence images. Each condition was tested with at least three biological replicates (error bar = standard deviation; n = 3–5; ***p < 0.0001, one-way ANOVA adjusted by Tukey test), and five random images were taken from each sample. Source data of Fig. 2 are provided in the Source data file.

Fig. 3. Optimization of the pillar design.

a Bending of PDMS pillars with a height of 10 µm, diameter of 2 µm, inter-pillar distance of 5 µm, and Young’s modulus of 2.1 MPa in response to a 5 mT magnetic field. The pillars appeared green in fluorescence microscopy after being labeled with Acridine Orange. b Stress propagation in viscoelastic biofilms generated by the on-demand beating of active pillars with different inter-pillar distances (D = 2, 5, and 10 µm). Active pillars (in white color) were actuated using a 5 mT magnetic field. c Representative fluorescence images of biofilms on flat controls, static controls, and PDMS surfaces with active surface topographies. The pillars were 10 µm tall with a diameter of 2 µm and inter-pillar distance of 5 µm. Active surface topographies were operated in three modes, including continuous beating for biofilm prevention, on-demand removal of mature biofilms (only actuated for 3 min after 48 h of biofilm growth), and a sequential combination of these two treatments. Biofilm cells were labeled with STYO^®9. The biomass was quantified using COMSTAT. The samples in the bar graph are indicated with the pattern labels above the corresponding fluorescence images. Each condition was tested with at least three biological replicates (error bar = standard deviation; n = 3–4), and five random images were taken from each sample. Source data of Fig. 3c are provided in the Source Data file.

Fig. 4. On-demand actuation sensitized biofilm cells to bactericidal antibiotics.

a, b Antibiotic susceptibility of 48 h UPEC ATCC53505 (a) and P. aeruginosa PAO1 (b) biofilm cells after treatment for 1 h in 0.85% NaCl at 37 °C. UPEC ATCC53505 biofilm cells were treated with 20 µg mL⁻¹ Ofx (bactericidal) and 200 µg mL⁻¹ Amp (bacteriostatic). P. aeruginosa PAO1 cells were treated with 50 µg mL⁻¹ Tob (bactericidal) and 200 µg mL⁻¹ Min (bacteriostatic). For the samples that went through on-demand actuation, the remaining biofilm cells were detached by bead beating (25 Hz bead beating for 30 s). All of the dispersed cells also went through the same bead beating step to eliminate artifacts before they were treated with an antibiotic. The biofilm cells dispersed by beat beating and then treated with an antibiotic are referred to as Static controls, and biofilms treated directly with an antibiotic without dispersion are referred to as Intact biofilms. Each condition was tested with at least three biological replicates (error bar = standard deviation; n = 3–4; ***p < 0.0001 and *p < 0.05, two-way ANOVA adjusted by Tukey test). Source data of Fig. 4 are provided in the Source data file.

Fig. 5. Biocompatibility of active topographies.

a, b Schematic of the experimental setup (a) and a cross-sectional view of the catheter prototype (b). c Effects of insulated copper coils on transformed human urinary bladder T24 cells without or with actuation (1 mT for 3 h or 5 mT for 3 min; Bar = 50 µm). Each condition was tested with three biological replicates (n = 3), and five random images were taken from each sample.

Fig. 6. Long-term test of the prototype catheters.

a Schematic of the flow cell system for testing long-term antifouling activities of the prototype catheters. A zoom-in view shows the coil section. b Antifouling activities of the prototype catheters. The bar graph shows the CFU per unit interior lumen area. The representative images show the complete blockage (visible white substance) of static control (on day 3) and flat control (on day 5) catheters, while the prototype catheters remained clear on day 30. The coils for generating the magnetic field were removed before imaging in Fig. 6b. The catheters were incubated at 37 °C with a continuous flow of artificial urine medium at 10 mL h⁻¹. Each condition was tested with three biological replicates (error bar = standard deviation; n = 3). Source data of Fig. 6b are provided in the Source data file.