Pathobiont-driven antibody sialylation through IL-10 undermines vaccination

Abstract

The pathobiont Staphylococcus aureus (Sa) induces nonprotective antibody imprints that underlie ineffective staphylococcal vaccination. However, the mechanism by which Sa modifies antibody activity is not clear. Herein, we demonstrate that IL-10 is the decisive factor that abrogates antibody protection in mice. Sa-induced B10 cells drive antigen-specific vaccine suppression that affects both recalled and de novo developed B cells. Released IL-10 promotes STAT3 binding upstream of the gene encoding sialyltransferase ST3gal4 and increases its expression by B cells, leading to hyper-α2,3sialylation of antibodies and loss of protective activity. IL-10 enhances α2,3sialylation on cell-wall-associated IsdB, IsdA, and MntC antibodies along with suppression of the respective Sa vaccines. Consistent with mouse findings, human anti-Sa antibodies as well as anti-pseudomonal antibodies from cystic fibrosis subjects (high IL-10) are hypersialylated, compared with anti-Streptococcus pyogenes and pseudomonal antibodies from normal individuals. Overall, we demonstrate a pathobiont-centric mechanism that modulates antibody glycosylation through IL-10, leading to loss of staphylococcal vaccine efficacy.

Keywords: Adaptive immunity; Immunology; Imprinting; Infectious disease.

Figure 1. IL-10 plays a critical role in retuning protective IsdB antibody function.

(A) Experimental setup. C57BL/6 mice were injected i.p. with Sa (LAC) or saline, immunized i.p., then challenged with LAC i.p. (B) Effect of IsdB vaccination on serum cytokines, measured by multiplex cytokine assay performed 1 day after Sa infection. IsdB, naive mice vaccinated with IsdB plus alum; mock, naive mice given alum; Sa/IsdB, LAC-infected mice vaccinated with IsdB/alum; Sa/mock, LAC-infected mice given alum alone. The heatmap graph represents mean values from n = 5 mice per group. (C) Serum IL-10 protein 24 hours after infection. Experiment performed as in Figure 1A (n = 10 per mouse group from 2 independent experiments). (D) IL-10–neutralizing antibody restores IsdB antibody protection. Mice were infected as in Figure 1A and vaccinated with IsdB with or without αIL-10 antibody. Seven days after, serum was adoptively transferred into naive mice followed by LAC challenge (n = 10 per mouse group from 2 independent experiments). (E) IsdB vaccination induced protective antibodies in IL-10^–/– mice. C57BL/6 mice or congenic IL-10^–/– mice were infected and vaccinated as in A. Serum was adoptively transferred 7 days after vaccination to assess for anti-Sa immunity by LAC challenge (n = 10 per mouse group from 3 independent experiments). (F) Opsonophagocytosis of Sa (LAC) by primary mouse neutrophils in the presence of immunized sera from E. Mean values ± SD from 3 independent experiments. (G) IL-10 limits IsdB vaccine efficacy. Naive mice were administered IL-10 or control with IsdB vaccination. Serum was assessed for anti-Sa immunity by adoptive transfer into naive mice followed by LAC challenge (n = 10 per mouse group from 2 independent experiments). Bars represent group median; dashed lines indicate the limit of detection (D, E and G). **P < 0.01; ***P < 0.001, 1-way ANOVA followed by Bonferroni’s multiple-comparison adjustment (C–G).

Figure 2. B10 cells abrogate humoral protection induced by IsdB vaccination.

(A) IL-10 in culture supernatant 16 hours after stimulation of naive or Sa-experienced splenic B cells with LPS or heat-killed Sa at indicated MOI.NT, nontreated. (B) αIL-10 antibody abrogates suppressive B cell effect on IsdB vaccination. Naive mice administered B cells from Sa-exposed mice were IsdB vaccinated with/without αIL-10 antibody, then challenged as per Figure 1A (n = 4–5 per mouse group). (C) αCD22 antibody depletion of B10 cells a day prior to immunizations restores IsdB vaccine efficacy. (n = 7–10 per mouse group from 3 independent experiments). (D) Experimental setup for E and F. Splenic B cells were transferred from naive/Sa-infected CD45.1 mice into naive CD45.2 mice. Recipients were IsdB immunized, and splenic B cells were analyzed after 7 days (E). CD45.1 or CD45.2 B cells were then transferred into naive C57BL/6 mice followed by Sa challenge (F). (E) No difference in percentage (F) but lack of protective function of de novo developed (CD45.2) splenic B cells in mice exposed to suppressive B cells. (E and F) CD45.1, n = 5 per group, CD45.2, n = 20 per group from 4 independent experiments. (G) Bacterial burden in WT CD19^cre/+ mice or CD19^cre/+ IL-10RA^fl/fl that were infected and IsdB vaccinated as per Figure 1A (n = 8–10 per mouse group from 3 independent experiments). (H) B cells do not suppress IsdB vaccination when IsdB/HarA mutant Becker strain is used in prior infection. Infection/vaccination as per Figure 1A followed by antibody transfer. WT Becker was used in final challenge. (n = 15 per mouse group from 3 independent experiments). Bars represent group means; error bars represent means ± SD (A). Dashed lines indicate the limit of detection (B to C and E to H). *P < 0.05; ***P < 0.001; ****P < 0.0001, 1-way ANOVA followed by Bonferroni’s multiple-comparison adjustment (A to C and E to H).

Figure 3. IL-10 modifies α2,3 sialylation on IsdB antibodies.

(A) Effect of αIL-10 antibody treatment on IsdB antibody glycosylation. IsdB antibodies purified from mice in Figure 1D (treated with Ctrl IgG or αIL-10) and assessed by lectin ELISA to determine antibody glycan content. (B and C) Effect of IL-23– or IL-6–neutralizing antibody on IsdB antibody sialylation. Serum IsdB antibodies from IL-6 (B) or IL-23 (C) antibody-treated, Sa/IsdB vaccinated mice, per Figure 1D, were assessed for α-2,6 or α-2,3 sialylation by SNA and MAA lectin ELISA. (D) UPLC-FL analysis of N-glycans released from PNGaseF treatment of purified IsdB antibodies from mice in Figure 1D (treated with Ctrl IgG or αIL-10). (E) Pie charts showing percentage of N-glycans in Supplemental Figure 5F Glycan schematics used here and in all other figures follow the recommended symbol nomenclature for glycans (SNFG). Glycan nomenclature: blue, N-acetylglucosamine (GlcNAc); yellow, galactose (Gal); green, mannose (Man); pale blue, sialic acid (Neu5Gc or Neu5Ac); red, fucose (Fuc); A, antennae; S, sialic acid; F, fucose; G, galactose. Bars represent group median; each point represents an individual mouse (A–C). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001, 1-way ANOVA followed by Bonferroni’s multiple-comparison adjustment (A–C).

Figure 4. IL-10 promotes STAT3 binding to St3gal4 promoter, which drives suppressive hypersialylation of IsdB antibodies.

(A) Number of putative STAT3 (IL-10) and NF-κB (IL-17A) binding sites on glycotransferase genes. (B) Effect of recombinant IL-10 on DNA-binding activity of STAT3 in naive splenic B cells, assessed by ChIP quantitative PCR (ChIP-qPCR) analysis. (C–F) Effect of αIL-10 antibody treatment on splenic B cell St3Gal (C), St6Gal (D), or Fut (E and F) expression. Experiment performed as in Figure 1D. (G) Effect of IsdB antibody desialylation on anti-Sa immunity in vivo. Naive mice were injected with α2-3 neuraminidase- or control- treated, purified Sa/IsdB antibody, then infected with LAC (n = 5 per mouse group). Bars represent group means; each point represents an individual mouse; error bars represent means ± SD (C–F). Bar represents group means; error bars represent means ± SD (B). Bar represents group median; each point represents an individual mouse; dashed lines indicate the limit of detection (G). *P < 0.05; **P < 0.01, 1-way ANOVA followed by Bonferroni’s multiple-comparison adjustment (C–G).

Figure 5. IL-10 promotes sialylation of anti-Sa antibodies and reduces anti-Sa vaccine efficacy.

(A and B) Effect of αIL-10 antibody treatment on anti-Sa immunity conferred by IsdA, FhuD2, and MntC vaccination in Sa-exposed mice, performed as in Figure 1D (n = 7–10 per mouse group from 2 independent experiments). (C–E) Effect of αIL-10 antibody treatment on serum antibody sialylation after IsdA (C), FhuD2 (D), or MntC (E) vaccination in Sa-exposed mice, performed as in A and B, assessed by MAA and SNA lectin binding. Bars represent group median; each point represents an individual mouse; dashed lines indicate the limit of detection (A and B). Bar represents group median; each point represents an individual mouse (C–E). *P < 0.05; **P < 0.01; ****P < 0.0001, 1-way ANOVA followed by Bonferroni’s multiple-comparison adjustment (A–E).

Figure 6. Hyper-α2,3 sialylation of human anti-Sa antibodies.

(A and B) α-2,6 and α-2,3 sialylation of purified human serum Sa antibodies (n = 9) as assessed by SNA and MAA lectin binding normalized to IgG titer. (C) Sialylation of purified human serum antibodies against M protein (M) or S protein (S) of Streptococcus pyogenes (n = 18) as assessed by MAA and SNA lectin binding normalized to IgG titer. Gray dashed line indicates median SNA or MAA level of IsdB antibody. (D) Sialylation of purified human serum antibodies against P. aeruginosa CbpD or FliC (n = 5) as assessed by MAA and SNA lectin binding normalized to IgG titer. H, healthy normal human; CF, subject with cystic fibrosis. Gray dashed line indicates median SNA or MAA level of IsdB antibody. (E) OPK of Sa (LAC) by primary mouse neutrophils in the presence of human IsdB antibodies treated with α2-3 neuraminidase or buffer control from human donors (n = 9). (F) Correlation between purified human serum IsdB antibody (n = 9) binding to MAA lectin and OPK of LAC. Green circle indicates IsdB antibodies treated with α-2,3 neuraminidase. Bars represent group median; each point represents an individual human donor; error bars represent means± SD (A–D). Each point represents an individual human donor (E and F). *P < 0.05; **P < 0.01; ***P < 0.001, Student’s t test (C–E), 1-way ANOVA followed by Bonferroni’s multiple-comparison adjustment (A and B) or linear regression (F).