Salmonella Typhimurium biofilm disruption by a human antibody that binds a pan-amyloid epitope on curli

Biofilms are multicellular communities that adhere to medical devices and confer resistance to antibiotics and immune clearance, often necessitating surgical removal. Enteric pathogens such as Salmonella enterica and Escherichia coli cause bloodstream and device-associated infections where biofilms are central to persistence. Biofilms limit antibiotic penetration, protect against host defenses, and foster persister cells. Curli amyloid fibers (CsgA/CsgB) constitute the major extracellular matrix component in Enterobacteriaceae biofilms, mediating adhesion and surface attachment. Amyloids share cross-β-sheet conformational epitopes that can be recognized by monoclonal antibodies with pan-amyloid activity. The human mAb 3H3 preferentially binds oligomeric/fibrillar amyloid-β and other human amyloids and reduces amyloid deposition in vivo. Research question: does 3H3 similarly bind the bacterial amyloid curli to inhibit fibrillization, disrupt Salmonella Typhimurium biofilms, and enhance their susceptibility to antibiotics and immune clearance in vitro and in vivo?

Biofilm-associated infections are refractory to antibiotics and contribute to antimicrobial tolerance and resistance. Curli fibrils make up ~85% of the Enterobacteriaceae biofilm matrix and are essential for biofilm architecture and adhesion. Curli biogenesis involves CsgB-mediated nucleation of CsgA polymerization; curli fibrils share structural features with human amyloids, including cross-β architecture. Pan-amyloid-binding antibodies and other anti-amyloid agents can inhibit fibrillization of diverse amyloids and have shown in vivo activity against eukaryotic amyloid diseases. Structural studies suggest shared conformational epitopes across amyloids that can be therapeutically targeted. Prior work showed 3H3 binds multiple pathogenic human amyloids, inhibits Aβ polymerization in vitro, and reduces amyloid deposition in mouse models, motivating investigation of its activity against prokaryotic amyloid curli.

- Bacterial strains: Salmonella Typhimurium IR715 (nalidixic acid-resistant, derived from ATCC 14028), isogenic mutants csgBA (curli-deficient) and msbB; clinical isolates of S. enteritidis and S. pullorum; uropathogenic E. coli UT189 and E. coli MC4100. Cultured in LB-based low/no-salt media; biofilms induced statically at 28 °C for up to 72 h. - Antibodies: Human mAb 3H3 (pan-amyloid-binding); additional anti-Aβ mAbs (4A6, 4G1, 2C10) cloned from an AD patient; isotype control human IgG 6A; rabbit anti-CsgA serum. Purified by standard methods. - In vitro biofilm assays: Biofilms grown on glass coverslips or as pellicles; exposed to 0.5 mg/ml antibodies (unless titrated) during growth. Staining with Syto9 (bacteria) and Congo Red (amyloid). Confocal laser scanning microscopy (Leica TCS, ×63). ImageJ used for 3D reconstructions, biofilm thickness quantification, fluorescence quantification, and 3D surface plot analyses. Crystal violet pellicle biomass quantification at OD570. - Bead penetration/topology assay: 1 µm Crimson FluoSpheres applied to biofilm surface; time-lapse confocal imaging to assess vertical and horizontal movement and bead capture within biofilms; quantification by ImageJ. - Pre-established biofilms: Biofilms formed for 24 or 48 h, then incubated ±0.5 mg/ml 3H3 for 24 h; confocal imaging; analysis of surface coverage and particles above mean biofilm mass. - Curli fibrillization (Thioflavin T assay): Synthetic peptides CsgA_R4-5 (fibrillizing) and CsgA_R4-5 N122A (non-fibrillizing control) incubated with ThT ±0.5 mg/ml 3H3 or control 6A; fluorescence monitored every 8 min for 36 h at 37 °C; lag time (t0) and polymerization assessed. Dynamic light scattering for amyloid size (for Aβ in separate binding context). - Antibiotic combination studies in vitro: Biofilms formed ±0.5 mg/ml 3H3, then treated with ampicillin (30 µg/ml), ciprofloxacin (0.125 µg/ml), or streptomycin (12.5 µg/ml) for 24 h; confocal imaging and thickness analysis. - Catheter biofilm models: In vitro—1 cm IV catheter segments incubated with S. Typhimurium under biofilm-inducing conditions ±3H3; staining and confocal imaging. In vivo—2-mm catheter pieces pre-colonized with S. Typhimurium for 24 h implanted subcutaneously into female BALB/c mice; percutaneous injections into catheter lumen of 100 µg 3H3 or buffer at 24 and 48 h post-implantation. Some mice received ampicillin in drinking water (1 mg/ml) starting 24 h before implantation. Catheters explanted at 72 h, stained, and imaged. - Macrophage uptake assays: Bone marrow-derived macrophages (BMDMs) and immortalized macrophages cultured. Supernatants from biofilms grown ±3H3/6A/anti-CsgA applied to BMDMs; after 1 h, gentamicin protection assay performed; intracellular bacteria enumerated as CFU. Supernatant OD600 adjusted to 0.5 in a subset. Curli uptake: Congo Red-labeled purified curli ±3H3 incubated with macrophages; internalization visualized by confocal microscopy and quantified by fluorescence. - Statistics: GraphPad Prism; two-tailed Student’s t-tests; p<0.05 considered significant. Animal studies conducted under IACUC-approved protocols.

- Anti-amyloid mAbs reduced S. Typhimurium biofilm thickness and curli content in vitro. Mean thicknesses (µm): untreated ~20; 6A control 18.21; 4A6 10.9; 4G1 7.3; 2C10 9.5; 3H3 7.5; curli-deficient csgBA 5.5. Congo Red fluorescence (mean RFU) significantly reduced with 4G1 (23.9), 2C10 (22.1), 3H3 (23.2) vs control/6A (47.1/48.1), but above csgBA (7.91). p<0.05 or p<0.01. - Pellicle biofilm biomass (crystal violet OD570) was significantly decreased by all four anti-Aβ mAbs; 3H3 nearly abolished pellicle formation, similar to csgBA. 3H3 caused dose-dependent pellicle reduction in S. enteritidis and E. coli UT189/MC4100, but not S. pullorum. - 3H3 altered biofilm architecture to a loose, porous topology with increased cells above the mean biofilm mass and increased release of bacteria into supernatants (Supplementary Fig. 4). Bead assay: untreated biofilms captured an average of 17 beads, whereas 3H3- or anti-CsgA-exposed biofilms captured 68 and 75 beads, respectively (p<0.05/0.01), with both vertical and horizontal bead movement. - 3H3 destabilized pre-established biofilms (24 h and 48 h), reducing surface coverage and increasing bacteria above the mean biofilm mass, indicating inhibition of maturation and promotion of dissociation. - Curli fibrillization was inhibited by 3H3 in Thioflavin T assays: significant reduction in CsgA_R4-5 polymerization vs control and increased lag time for fibrillization; no effect on non-fibrillizing CsgA_R4-5 N122A or BSA controls. - Antibiotic synergy: Ampicillin alone did not reduce biofilm thickness (untreated 14.8 µm vs ampicillin 16.2 µm). 3H3 alone reduced thickness to 10.07 µm. 3H3 plus ampicillin decreased thickness to 7.4 µm and markedly reduced surface coverage. Ciprofloxacin or streptomycin alone were ineffective; 3H3 plus ciprofloxacin significantly decreased thickness and coverage; 3H3 plus streptomycin showed a non-significant trend toward reduction. - Catheter models: In vitro, 3H3 produced loose biofilm architecture lacking curli within catheter lumens. In vivo, percutaneous 3H3 injections into pre-colonized catheter lumens resulted in loose, weakly adherent biofilms; combination of systemic ampicillin (1 mg/ml in water) and 3H3 rendered catheters virtually biofilm-free with minimal Congo Red staining. - Immune interactions: 3H3 exposure increased macrophage phagocytosis of biofilm-released bacteria (gentamicin protection assays), even when supernatant bacterial densities were normalized (OD600=0.5). 3H3 enhanced internalization of Congo Red-labeled curli by macrophages; Congo Red-positive, cell-sized structures were observed on 3H3-treated catheters in vivo, consistent with phagocytic uptake.

The study demonstrates that targeting a pan-amyloid conformational epitope on curli with the human mAb 3H3 inhibits curli fibrillization and profoundly alters Salmonella Typhimurium biofilm architecture. Disruption of amyloid incorporation prevented formation of a dense, impermeable matrix, yielding a porous, loosely packed biofilm from which bacteria dissociated more readily. These structural changes increased penetration and trapping of particles within biofilms and sensitized biofilm-embedded bacteria to multiple antibiotic classes. 3H3 also destabilized pre-established biofilms, indicating activity beyond prophylaxis. In vivo, 3H3, particularly in combination with ampicillin, enabled clearance of catheter-associated biofilms, and promoted uptake of curli and biofilm-released bacteria by macrophages, suggesting an immuno-opsonic contribution to biofilm eradication. Together, the findings address the research hypothesis that a pan-amyloid-binding antibody can target bacterial amyloids to disrupt biofilms and enhance therapeutic clearance, extending anti-amyloid immunotherapy principles from eukaryotic to prokaryotic amyloids.

This work shows that the human monoclonal antibody 3H3, which recognizes a pan-amyloid epitope, binds bacterial curli, inhibits curli fibrillization, disrupts Salmonella Typhimurium biofilm structure, enhances antibiotic efficacy, and promotes innate immune uptake, enabling clearance of catheter-associated biofilms in mice. The results support a therapeutic strategy of targeting amyloid components of bacterial biofilms with monoclonal antibodies, either alone or in combination with antibiotics, and suggest broader applicability to other curli-containing and amyloid-based biofilms. Future research should define the precise epitopes and mechanisms of action, optimize antibody combinations and dosing, evaluate efficacy across diverse pathogens and device types, and systematically characterize interactions with different antibiotic classes in vitro and in vivo.

The study primarily uses in vitro systems and a mouse subcutaneous catheter model; translation to human clinical settings will require further validation. Antibiotic interaction effects were demonstrated for selected agents and doses; comprehensive pharmacodynamic studies are needed to generalize synergy across drug classes and regimens. While activity was demonstrated in multiple strains, one tested species (S. pullorum) did not show 3H3 effects on pellicles, indicating species/strain variability and potential alternative biofilm matrices. Mechanistic details of 3H3-curli binding (epitope mapping, effects on surface adhesion domains) and the relative contributions of opsonization versus matrix disruption to in vivo clearance remain to be elucidated.