Potent efficacy of an IgG-specific endoglycosidase against IgG-mediated pathologies

Abstract

Endo-β-N-acetylglucosaminidases (ENGases) that specifically hydrolyze the Asn297-linked glycan on immunoglobulin G (IgG) antibodies, the major molecular determinant of fragment crystallizable (Fc) γ receptor (FcγR) binding, are exceedingly rare. All previously characterized IgG-specific ENGases are multi-domain proteins secreted as an immune evasion strategy by Streptococcus pyogenes strains. Here, using in silico analysis and mass spectrometry techniques, we identified a family of single-domain ENGases secreted by pathogenic corynebacterial species that exhibit strict specificity for IgG antibodies. By X-ray crystallographic and surface plasmon resonance analyses, we found that the most catalytically efficient IgG-specific ENGase family member recognizes both protein and glycan components of IgG. Employing in vivo models, we demonstrated the remarkable efficacy of this IgG-specific ENGase in mitigating numerous pathologies that rely on FcγR-mediated effector functions, including T and B lymphocyte depletion, autoimmune hemolytic anemia, and antibody-dependent enhancement of dengue disease, revealing its potential for treating and/or preventing a wide range of IgG-mediated diseases in humans.

Keywords: Corynebacterium ulcerans; ENGases; Fc γ receptor-mediated effector functions; FcRn blockers; GH18; IgG specificity; T cell and B cell depletion; antibody-dependent enhancement of dengue disease; autoimmune diseases; bacterial immune evasion; corynebacterial endoglycosidases.

Figure 1.. Members of a Corynebacterial pathogen clade encode single-domain ENGases with sequence similarity to known multi-domain IgG-specific ENGases.

A) SSN analysis clustering GH18 family ENGases into three sub-clusters: I (multi-domain IgG-specific), II (non-IgG-specific), III (Corynebacterial ENGases). B) Neighbor-joining phylogenetic tree of Corynebacterial ENGases with bootstrap values (1000 replicates). C) Schematic representation of IgG-specific ENGases (EndoS, EndoS2) and Corynebacterial ENGase domain architecture showing canonical active site. Scissors indicate the regions removed from the proteins for expression. D) Sequence identity matrix between Corynebacterial and IgG-specific ENGases (EndoS, EndoS2) displayed with a color scale. See also Figure S1.

Figure 2.. Corynebacterial ENGases exhibit strict specificity for N-glycans attached to Asn297 on native human IgG antibodies.

A) Competitive deglycosylation assay representing the % of glycosylated fraction of glycoprotein substrates indicated, determined by intact glycoprotein LC-MS. Enzymes were added at 10 nM final concentration in a reaction tube containing 5 μM of each glycoprotein (n=3). Data is presented as mean values ± SD. B) Heatmap representing Fc glycosylated fraction indicative of enzymatic activity of selected IgG-specific ENGases against different subtypes of IgG antibodies from human (IgG1, IgG2, IgG3 and IgG4) or murine origin (IgG1, IgG1(D265A) and IgG2a). Data is presented as mean values (n=3). C) Melting temperature (Tm) of Corynebacterial IgG-specific ENGases at pH 7.4 determined by DSF assays. Tm is presented as mean values. D) Percentage of completely hydrolyzed CT N-glycans on human IgG1 Fc by Corynebacterial IgG-specifc ENGases after 5 h incubation.at 23 °C. Activity values were normalized to CU43 activity values. Data is presented as mean values ± SD (n=3). E-H) Kinetic analysis of E) CU43 F) CM49 G) EndoS and H) EndoS2 hydrolyzing CT N-glycan from human IgG1 at 37 °C. 10 nM of each enzyme was incubated with 5 μM hIgG1 CT substrate in buffer PBS pH 7.4. Diglycosylated, monoglycosyalted and deglycosylated Fc ratio curves are represented in purple, green and orange, respectively. Curves are representative of independent triplicates. See also Figure S2

Figure 3.. CU43 recognizes both glycan and protein components of its glycoprotein substrate.

A) Two views of cartoon representation of the crystal structure of CU43_D187A-E189A (CU43i) at 2.3 Å resolution. B) CU43 structure with G0F glycan docked representing B-factor values. Range color B-factor key from 15 less flexible to 80 more flexible. C) Surface representation of CU43 crystal structure with docked G0F glycan (upper panel), compared with EndoS structure co-crystallized with G0F (bottom panel). Deep view of glycan binding pocket is illustrated. D) Affinities determined by SPR of CU43i binding to IgG bearing different CT N-glycans. CU43i analyte concentration range plotted (serial dilution 1:2): 39 nM to 1.25 μM for ligands 1–3 (IgG CT diglycosylated and IgG CT monoglycosyalted R and L), 156 nM to 5 μM for ligand 4 and (IgG aglycosylated) and 312 nM to 10 μM for ligand 5 (IgG diglycosylated HM-Man9). SPR assays were run in independent duplicates. See also Figure S3.

Figure 4:. A CU43-Fc fusion protein prevents cytotoxic mAb-mediated T cell depletion in FcγR humanized mice at a significantly lower dose than does efgartigimod.

A) FcγR humanized mice (n=2–4/group) were treated (i.v.) with 10 μg anti-CD4 mAb (clone YTS191) expressed as afucosylated hIgG1, followed by the indicated doses of CU43 Fc fusion constructs. B) FcγR humanized mice (n=3–4/group) were treated (i.v.) with 10 μg afucosylated hIgG1 YTS191 (anti-mouse CD4) followed by 0.1 μg CU43-Fc (active or inactive) or CM49-Fc (active). Abundance of CD4+ T cells in peripheral blood was assessed one day before injection with YTS191 and Fc-fusion enzymes (baseline) and three days post-injection and expressed the change in CD4+ abundance over time as (%CD4+ of CD3+, day 3)/((%CD4+ of CD3+, baseline). Comparisons between groups were made using one-way ANOVA (a=0.05; Bonferroni post-hoc). C) FcγR humanized mice (n=4–6/group i.v.) were treated with 10 μg afucosylated human IgG1 YTS191 mAb followed by the indicated dose of CU43-Fc fusion construct or efgartigimod. D) Plot displays a representative gating of CD4+ cells in peripheral blood one day post-injection with YTS191 and 0.05μg of either CU43-Fc or efgatigimod. In all experiments, the abundance of CD4+ T cells in peripheral blood was assessed by flow cytometry (expressed as CD4+/CD3+ and normalized to baseline). Error bars shown represent SEM. See also Figure S4.

Figure 5:. In vivo activity of CU43-Fc prevents cytotoxic anti-human CD20 mAb-mediated depletion of B cells in FcγR humanized mice.

A-B) FcγR/CD20 humanized mice (n=4 mice/group, i.v.) were co-treated with the B cell depleting anti-human CD20 antibody 2B8 (expressed as the Fc-optimized human IgG1 variant GAALIE (G236A/A330L/I332E) at a dose of 1 mg/kg followed by 0.5 μg CU-43 Fc (active or inactive) or PBS. A) B cell counts in peripheral blood (B220+ cells) were assessed three days post-injection by flow cytometry. B) Representative gating of B220+ cells in peripheral blood taken from mice treated with either CU43-inactive or CU43-active. Error bars shown represent SEM.

Figure 6. CU43-Fc impedes mAb-mediated induction of autoimmune haemolytic anemia in mice expressing human CD47, SIRPa, and FcγRs.

(A-B) Mice humanized for CD47, SIRPa, and FcγRs (n=3–4/group) were injected i.v. with 5F9 at a dose of 5 mg/kg followed by CU-43 Fc (active or inactive) at a dose of 0.1 μg/mouse. A) RBC counts were determined one day preinjection (d-1) and one day post-injection (d1) and expressed as the % of RBC d1/RBC d-1 (one-way ANOVA (a=0.05; Bonferroni post-hoc)). B) Comparison of mouse survival among treatment groups (Mantel Cox log rank test; p=0.0404 CU43i vs CU43 active). Error bars shown represent SEM.

Figure 7.. CU43-Fc fusion abrogates antibody-dependent enhancement (ADE) of dengue disease in humanized FcγR mice.

(A-C) Eight hours before infection with DENV2 virus (3.5×108 GE, i.v.), type I interferon-knockout mice humanized for FcγRs (Ifnar1−/− /hFcγR mice; n=6–7 mice/group) were pre-treated with 20 μg (i.v.) of the anti-DENV2 E protein antibody C10 (hIgG1) modified to either enhance (afucosylated) binding to FcγRIIIa or abrogate (GRLR variant, G236R/L328R) binding to all FcγRs followed by either 5 μg (i.v.) CU-43 Fc or an equivalent volume of PBS. (A) Changes in body weight from baseline. The dotted line represents the weight loss threshold (80%) at which death is defined. B) Comparison of areas under the curve (AUC) shown in (A) ((one-way ANOVA (a=0.05; Boferroni post-hoc)). Error bars shown represent SEM. (C) Survival (death defined as loss of >20% baseline weight) (Mantel Cox log rank test; p=0.008 αDENV IgG afuc + CU43 vs αDENV afuc).