Monoclonal Antibody Disrupts Biofilm Structure and Restores Antibiotic Susceptibility in an Orthopedic Implant Infection Model

Abstract

Bacterial biofilms on orthopedic implants are resistant to the host immune response and to traditional systemic antibiotics. Novel therapies are needed to improve patient outcomes. TRL1068 is a human monoclonal antibody (mAb) against a biofilm anchoring protein. For assessment of this agent in an orthopedic implant infection model, efficacy was measured by reduction in bacterial burden of Staphylococcus aureus, the most common pathogen for prosthetic joint infections (PJI). Systemic treatment with the biofilm disrupting mAb TRL1068 in conjunction with vancomycin eradicated S. aureus from steel pins implanted in the spine for 26 of 27 mice, significantly more than for vancomycin alone. The mechanism of action was elucidated by two microscopy studies. First, TRL1068 was localized to biofilm using a fluorescent antibody tag. Second, a qualitative effect on biofilm structure was observed using scanning electron microscopy (SEM) to examine steel pins that had been treated in vivo. SEM images of implants retrieved from control mice showed abundant three-dimensional biofilms, whereas those from mice treated with TRL1068 did not. Clinical Significance: TRL1068 binds at high affinity to S. aureus biofilms, thereby disrupting the three-dimensional structure and significantly reducing implant CFUs in a well-characterized orthopedic model for which prior tested agents have shown only partial efficacy. TRL1068 represents a promising systemic treatment for orthopedic implant infection.

Keywords: S. aureus; biofilm; monoclonal antibody; orthopedic implant infection; prosthetic joint infection; spinal implant infection.

Figure 1

X-ray of spinal implant in situ. A 0.1 mm diameter, 10 mm long stainless-steel pin, with a 90-degree bend 2 mm from the proximal end, was implanted into the L4 spinous process.

Figure 2

Ex vivo localization of fluorescently tagged TRL1068. The in vivo spinal implant infection model was used. Isotype IgG control antibody and TRL1068 were tagged with a fluorescent dye using the VivoTag^® 680XL kit. Animals were implanted and inoculated with S. aureus Xen36 on POD0 and received fluorescently tagged TRL1068 (Fl-TRL1068), fluorescently tagged isotype IgG control antibody (Fl-CAb), fluorescent tag alone, or vehicle control on POD4. Twenty-four hours after antibody dosing, the animals were sacrificed, and the implants were harvested and stained with TOTO^®-1 eDNA stain to co-localize bacterial biofilm. Images were taken with confocal microscopy. CONTROL (bacterial biofilm alone, vehicle control, no dyes), FL-TAG ONLY (Infected control + dyes alone (VivoTag^® 680XL alone and TOTO^®-1 eDNA stain)), FL-CAb (biofilm-infected implant + IgG isotype control antibody tagged with VivoTag^® 680XL + TOTO^®-1 eDNA stain), and FL-TRL1068 (biofilm infected implant + TRL1068 antibody tagged with VivoTag^® 680XL + TOTO^®-1 eDNA stain) groups are shown here. There was no fluorescent signal in group A. Biofilm staining as measured by TOTO^®-1 signal intensity was similar in groups B, C, and D (1.56 × 10⁷ lumens/m², 2.32 × 10⁷ lumens/m², and 1.07 × 10⁷ lumens/m², respectively). By comparison, there was a seven-times greater signal intensity within the VivoTag^® 680XL wavelength in group D (FL-TRL1068) relative to control groups B and C (7.19 × 10⁷ lumens/m² vs. 1.45 × 10⁷ lumens/m² and 1.48 ×10⁷ lumens/m², respectively). This confirms localization of FL-TRL1068 to biofilm in vivo.

Figure 3

Ex vivo scanning electron microscopy (SEM) evaluation of three-dimensional biofilm structure. Stainless steel pins were implanted into the L4 spinous process of mice on Day 0 and inoculated with saline (sterile control) or 1× 10³ S. aureus Xen36. On Day 8, the animals were sacrificed (n = 3 per group), and the pins were imaged according to a standard SEM protocol. Experimental groups: (Sterile) sterile control showing a smooth metal pin; (Infected Control) infected control (bacteria alone) showing a thick 3-dimensional biofilm structure; (CAb) animals exposed to IgG isotype control antibody (CAb) on Days 4 and 7, showing 3-D biofilm, bacterial cocci (yellow arrowheads), and host cells (green arrows); (TRL1068) animals exposed to TRL1068 on Days 4 and 7, showing flattened biofilm remnant with no bacteria or host cells present. Note: no antibiotics were given for this experiment. Magnification was the same for all images: width = 75 µm; scale bars: 2 µm.

Figure 4

Day 35 implant CFUs. On POD 35, animals were sacrificed, implants were harvested, and CFUs were enumerated. (A) POD 35 ex vivo mean implant CFUs. There was a 4-log reduction in average implant CFUs in the TRL1068+vancomycin group relative to the infected control group (6.17 × 10⁻¹ vs. 1.03 × 10⁴, respectively; p < 0.0001), and a 3.3-log reduction in average implant CFUs in the TRL1068+vancomycin group relative to the vancomycin only group (6.17 × 10⁻¹ vs. 1.13 × 10⁴, respectively; p = 0.003). (B) Implant infection. Pearson’s Chi-square test showed a statistical relationship between treatment and infection status for the implants (χ² [df = 3, N = 78] = 18.73; p < 0.001). The TRL1068+Vanc group had significantly lower proportion of infected implants (1/27, 3.7%; p < 0.05) compared to the infected control (8/12, 66.7%) and the Vanc (5/12, 41.7%) groups. ns = not significant, * p < 0.05, ** p < 0.005, **** p < 0.0001.

Figure 5

Day 35 peri-implant tissue CFUs. On POD 35, animals were sacrificed, peri-implant tissues were harvested, and CFUs per gram of tissue were enumerated. (A) POD 35 ex vivo mean peri-implant tissue CFUs/g. There was no significant difference in the mean peri-implant tissue CFUs/g between any of the experimental groups. (B) Pearson’s Chi-square test showed a statistical relationship between treatment and infection status for the implants: χ² [df = 3, N = 78] = 8.30; p = 0.03. In the peri-implant tissues, the infected control group had a significantly higher proportion of infected tissues (13/13, 100%; p < 0.05) compared to TRL1068+Vanc (16/27, 59.3%) and Vanc (10/12, 83.3%) groups. No significant difference in the proportion of infected tissues was noted among the treatment groups.