Editor's Choice Platform for identifying human glycan-specific antibodies against bacterial pathogens using synthetic glycan fragments

Abstract

Bacterial infections represent a substantial global health challenge, impacting both human and veterinary health. The ongoing evolution of antibiotic-resistant pathogens, coupled with limited new antibiotic discoveries, urges the need for alternative strategies to treat and prevent these infections. Passive immunization with monoclonal antibodies (mAbs) is gaining interest as a promising alternative. Here, we report an experimental pipeline for generating human mAbs from healthy donor B cells using synthetic mimics of complex bacterial glycans. We identified functional mAbs recognizing discrete and unique epitopes on the surface glycans of two bacterial priority pathogens; Staphylococcus aureus and Streptococcus pyogenes. The use of chemically-defined synthetic glycans was critical for the discovery and systematic characterization of mAbs. From a heterogeneous mix of B cell specificities, antibody sequences were identified, leading to the production of mAbs with distinct reactivities against immunodominant but also to less common or even masked epitopes. The pipeline can be adapted to different glycan targets, donor material or specific antibody isotypes. This work thereby paves the way for the discovery of glycan-specific mAbs with clinical relevance to treat, prevent or diagnose infections with S. aureus, S. pyogenes or other bacterial pathogens.

Keywords: Staphylococcus aureus; Streptococcus pyogenes; antibody discovery; bacterial glycan; single cell B cell receptor sequencing.

Fig. 1

sWTA probe generation and validation. A) schematic representation of sWTA probe generation for all three glycoforms, i.e. RboP +β-1,4-GlcNAc, +α-1,4-GlcNAc, or + β-1,3-GlcNAc. All sWTA probes were made for detection in two different fluorescence channels using streptavidin conjugated to AF647 or BB515. B) dual sWTA probe labeling of protein A-coated beads coated with anti-β-GlcNAc RboP (clone 4497), anti-α-GlcNAc RboP (clone 4461), and isotype IgG1. Data in dot plots represent geometric mean fluorescence intensity (gMFI) signals (fluorophores: AF647 and BB515) on the beads. Q2 and Q4 comprise, respectively, double positive (dual sWTA probe labeling) and double negative (no sWTA probe binding) beads. Signals within Q1 and Q3 represent aspecific binding of, respectively, AF647 and BB515 streptavidin to beads. Histograms are included on the sides of the dot plots to visualize relative amounts of different bead populations within a fluorescent channel.

Fig. 2

sWTA specificity screening of pilot-scale produced mAbs. Binding profiles of 15 B cell-derived mAbs, expressed by HEK293T cells (production levels in Fig. S3) as human IgG1, to streptavidin-coated beads immobilized with biotinylated sWTA to determine clone reactivity as measured by flow cytometry. Fluorescent signals are depicted in this figure as gMFI fold changes (mean + s.d. of three independent experiments) relative to the condition without antibodies to compensate for technical variation. One-way ANOVA was performed to determine significant binding to glycan-coated beads compared to empty beads. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001.

Fig. 3

Specificity verification of sWTA-reactive mAbs at equimolar level. Clones that displayed sWTA reactivity in the pilot screening (main Fig. 2) were selected for large-scale production in HEK293 freestyle cells and purified through protein G agarose. Selected clones were categorized based on their α-GlcNAc A) or β-GlcNAc B) reactivity and relative binding capacities to sWTA beads were determined at a concentration of 3 μg/ml. Beads coated with polyrhamnose (PR) + β-1,3-GlcNAc and empty beads were used as controls for cross-reactivity and background, respectively. IgG1 binding to sWTA beads was measured by flow cytometry and data represent the gMFI mean + s.d. of three independent experiments. One-way ANOVA with Dunnett’s multiple comparisons test was performed to determine significant binding of sWTA-reactive clones to glycan-coated beads compared to empty beads. ns not significant ^**P < 0.01, ^****P < 0.0001. Index sort data of all sorted clones can be found in Fig. S4.

Fig. 4

Binding of sWTA-reactive mAbs to WTA on S. aureus surface. For each glycoform specificity, one mAb clone was selected to assess bacterial opsonization. A-D) binding of W1C11 (anti-α-GlcNAc), W1F10 (anti-β-GlcNAc, with preference for β-1,3-GlcNAc), W1G7 (anti-β-GlcNAc), and B12 (isotype control) to S. aureus strains N315 Δspa A), N315 Δspa ΔtarSP B), Newman Δspa Δsbi C), and Streptococcus pyogenes strain 5448 ΔgacH D). IgG1 binding to bacteria was measured using flow cytometry and data represent normalized mean gMFI + s.d. (isotype signals set to 1) of three independent experiments. N315 Δspa ΔtarSP and S. pyogenes 5448 ΔgacH were included as controls for WTA GlcNAc (species) specificity.

Fig. 5

Effector functions of anti-WTA clones towards S. aureus. A-C) C3b deposition by sWTA-reactive mAbs onto S. aureus strains N315 Δspa A), N315 Δspa ΔtarSP B), and Newman Δspa Δsbi C). Data represent C3b binding (normalized gMFI + s.d.) of three independent experiments as measured by flow cytometry. Fluorescent signals are depicted as a fold change relative to the condition without antibodies to compensate for variation in background signals between biological replicates. D) neutrophil-mediated phagocytosis of GFP-expressing S. aureus Newman Δspa Δsbi by anti-WTA mAbs. Displayed data represent percentages of GFP-positive neutrophils and are representative of three biological replicates (individual replicates can be found in Fig. S5). Curves were generated using nonlinear dose–response fitting model. E) relative phagocytic capacities of the anti-WTA mAbs. Absolute IC₅₀ values were determined for each replicate individually using nonlinear dose–response fitting model. Black lines represent means of the IC₅₀ values which are depicted as data points. Assays (A-D) were performed in the presence of 1% IgG-/IgM-depleted human serum as complement source. Statistical differences compared to isotype were determined by one-way ANOVA. ^**P < 0.01, ^***P < 0.001.

Fig. 6

Discovery and characterization of GAC-specific mAbs. A) schematic representation of sGAC probe generation for two glycoforms, i.e. PR and PR + β-1,3-GlcNAc. All sGAC probes were made for detection in two different fluorescence channels using streptavidin conjugated to Pe-Cy7 or BV421. B) sGAC probe binding to protein beads immobilized with goat polyclonal anti-GAC GlcNAc (Ab9191). N.B. no bead coat option was available to test PR specificity. Data in dot plots represent fluorescence signals (fluorophores: PE-Cy7 and BV421) on the beads. Q2 and Q4 comprise, respectively, double positive (dual sGAC probe labeling) and double negative (no sGAC probe binding) beads. Signals within Q1 and Q3 represent aspecific binding of, respectively, Pe-Cy7 and BV421 streptavidin to beads. C) specificity verification of sGAC-reactive mAbs at equimolar level. Clones were produced in HEK293 freestyle cells and purified through protein G agarose. Relative binding capacities to sGAC beads were determined at a concentration of 3 μg/mL. Beads coated with RboP +β-1,3-GlcNAc and empty beads were used as controls for cross-reactivity and background, respectively. IgG1 binding to sGAC beads was measured by flow cytometry and data represent the mean gMFI ± s.d. of three independent experiments. One-way ANOVA with Dunnett’s multiple comparisons test was performed to determine significant binding of sGAC-reactive clones to glycan-coated beads compared to empty beads. ^****P < 0.0001. Index sort data of all sorted clones can be found in Fig. S8. D) binding of sGAC-reactive mAbs to natural GAC on Streptococcus pyogenes. For each glycoform specificity, one mAb clone was selected that showed evident binding to sGAC beads (panel C) and tested for bacterial opsonization. Titration of G1E8 (anti-PR), G1C4 (anti-β-1,3-GlcNAc PR), and B12 (isotype control) to S. pyogenes 5448 Δemm1. Bacterial opsonization was determined by measuring IgG1 binding to bacteria using flow cytometry and data represent normalized mean gMFI + s.d. (isotype signals set to 1) of three independent experiments. E) effector functions of anti-GAC clones towards S. pyogenes. C3b deposition by sGAC-reactive mAbs onto S. pyogenes 5448 Δemm1. Data represent C3b binding (normalized gMFI + s.d.) of three independent experiments and was measured by flow cytometry. Fluorescent signals are depicted as a fold change relative to the condition without antibodies to compensate for variation in background signals between biological replicates.

Fig. 7

Schematic overview of experimental workflow for identifying glycan-specific antibodies against bacterial pathogens using synthetic glycan fragments.