Human IgA binds a diverse array of commensal bacteria - PubMed (original) (raw)
. 2020 Mar 2;217(3):e20181635.
doi: 10.1084/jem.20181635.
Jehane Fadlallah 1, Olivia Adams 2, Claire Fieschi 3, Christophe Parizot 1, Karim Dorgham 1, Asok Rajkumar 1, Gaëlle Autaa 1, Hela El-Kafsi 1, Jean-Luc Charuel 1, Catherine Juste 4, Friederike Jönsson 5, Thomas Candela 6, Hedda Wardemann 7, Alexandra Aubry 1, Carmen Capito 6, Hélène Brisson 1, Christophe Tresallet 8, Richard D Cummings 9, Martin Larsen 1, Hans Yssel 1, Stephan von Gunten 2, Guy Gorochov 1
Affiliations
- PMID: 31891367
- PMCID: PMC7062531
- DOI: 10.1084/jem.20181635
Human IgA binds a diverse array of commensal bacteria
Delphine Sterlin et al. J Exp Med. 2020.
Erratum in
- Correction: Human IgA bind a diverse array of commensal bacteria.
Sterlin D, Fadlallah J, Adams O, Fieschi C, Parizot C, Dorgham K, Rajkumar A, Autaa G, El-Kafsi H, Charuel JL, Juste C, Jönsson F, Candela T, Wardemann H, Aubry A, Capito C, Brisson H, Tresallet C, Cummings RD, Larsen M, Yssel H, von Gunten S, Gorochov G. Sterlin D, et al. J Exp Med. 2020 Mar 2;217(3):e2018163501152020c. doi: 10.1084/jem.2018163501152020c. J Exp Med. 2020. PMID: 31990288 Free PMC article. No abstract available.
Abstract
In humans, several grams of IgA are secreted every day in the intestinal lumen. While only one IgA isotype exists in mice, humans secrete IgA1 and IgA2, whose respective relations with the microbiota remain elusive. We compared the binding patterns of both polyclonal IgA subclasses to commensals and glycan arrays and determined the reactivity profile of native human monoclonal IgA antibodies. While most commensals are dually targeted by IgA1 and IgA2 in the small intestine, IgA1+IgA2+ and IgA1-IgA2+ bacteria coexist in the colon lumen, where Bacteroidetes is preferentially targeted by IgA2. We also observed that galactose-α terminated glycans are almost exclusively recognized by IgA2. Although bearing signs of affinity maturation, gut-derived IgA monoclonal antibodies are cross-reactive in the sense that they bind to multiple bacterial targets. Private anticarbohydrate-binding patterns, observed at clonal level as well, could explain these apparently opposing features of IgA, being at the same time cross-reactive and selective in its interactions with the microbiota.
© 2019 Sterlin et al.
Conflict of interest statement
Disclosures: Drs. Sterlin, Fadlallah, Larsen, and Gorochov reported a patent to EP 18306006.0 pending. No other disclosures were reported.
Figures
Figure 1.
Carbohydrate-binding profile of polyclonal IgA1 and IgA2 antibodies. (A) Glycan reactivity of serum polyclonal IgA1 and IgA2 (n = 5 healthy donors pooled in one experiment). Each peak represents an individual glycan recognized by IgA1 (blue line) or IgA2 (red line). (B) Heatmap diagram depicting glycans recognized by IgA1 or IgA2. Each row represents an individual glycan. (C) Preferential recognition of distinct terminal carbohydrate moieties by serum polyclonal IgA1, IgA2, or both. Terminal carbohydrate moieties equally recognized by both IgA1 and IgA2 are depicted in white. Terminal moieties preferentially recognized by IgA1 or IgA2 are showed in gray or black, respectively. (D) Isotype-dependent recognition of Galα-terminal structures of bacterial origin and other bacterial antigens excluding Galα-terminal antigens (without Galα) by serum polyclonal IgA, as compared with entire glycan microarray dataset. Terminal carbohydrate moieties equally recognized by both IgA1 and IgA2 are depicted in white. Terminal moieties preferentially recognized by IgA1 or IgA2 are showed in gray or black, respectively. (E) Differential isotype binding to bacterial attachment sites.
Figure S1.
Gut bacteria segregate into IgAbright and IgAlow fractions in healthy humans. Related to Figs. 1 and 2. (A) The secretory component binds a modest range of carbohydrates. Glycan reactivity with the secretory component was assessed using glycan microarray technology (660 structures). Representative median RFUs are shown. Glycans specifically recognized by secretory component are colored in red. (B) Anti-IgA1 and anti-IgA2 antibodies do not cross-react. Flow cytometry analysis of IgA1 and IgA2 expression on peripheral B cells from a healthy donor. (C) IgA-coated bacteria split into IgAbright and IgAlow fractions depending on IgA1 and IgA2 coating. Representative flow cytometry analysis of IgA1 and IgA2 coating in IgAbright-coated bacteria (red lines), IgAlow-coated bacteria (blue lines), and IgA-unbound bacteria (gray lines). (D) Representative flow cytometric analysis of colon and ileum microbiota (left and central panels, respectively) with anti-IgA FITC or isotype-matched control antibody, as indicated. Numbers indicate percentage of positive cells. Data are cumulative from three independent experiments. Boxes extend from the 25th to the 75th percentiles. Error bars represent minimum and maximum values. P values were defined using the Mann-Whitney test. ***, P < 0.001. SSC-A, side scatter area. Quantification of IgA-coated bacteria in stool (n = 20) and ileum (n = 5).
Figure 2.
IgA1 targets IgA2-coated bacteria. (A) Representative flow cytometry analysis of endogenous IgA1 and IgA2 fecal or ileal microbiota coating. (B) Endogenous IgA1 and IgA2 microbiota coating levels in 20 fecal and 5 ileal healthy samples. Data are cumulative from three independent experiments. Error bars represent minimum and maximum values. P values were calculated using the Mann-Whitney test (colon vs. ileum) or Wilcoxon test (*, P < 0.05; ***, P < 0.001). (C) Endogenous IgM microbiota coating levels in 20 fecal and 5 ileal healthy samples (two independent experiments). Dark bars indicate medians. Red dotted line represent significant the positive cutoff. P values were calculated using the Mann-Whitney test (***, P < 0.001). (D) Left: Representative flow cytometric analysis of endogenous IgM-coated fecal or ileal microbiota. Right: Representative flow cytometric analysis of endogenous IgA1 and IgA2 binding in IgM-coated fecal or ileal microbiota (two independent experiments). SSC-A, side scatter area. (E) Endogenous IgM microbiota coating levels in 20 fecal and 5 ileal healthy samples (two independent experiments). Error bars represent minimum and maximum values. P values were calculated using the Mann-Whitney test (**, P < 0.01; *, P < 0.05). (F) Model Venn diagram showing overlap among endogenous IgA1, IgA2, and IgM binding in IgA-coated microbiota. (G) Endogenous IgA1 and IgA2 microbiota–coating levels in eight fecal samples from formula-fed neonates (two independent experiments). Error bars represent minimum and maximum values. P values were calculated using the Wilcoxon test (*, P < 0.05; **, P < 0.01). (H) Flow cytometric analysis of IgA1 and IgA2 expression on colonic lamina propria B cells from a 3-mo-old infant (one experiment).
Figure S2.
IgA1+IgA2+- and IgA2+-sorted fractions show distinct compositions within the same donor and between donors . Related to Fig. 4. (A) Sorting strategy of IgA1- and IgA2-coated bacteria (representative of five independent experiments). SSC-A, side scatter area. (B) Relative composition of phyla in fecal samples (input). Each column corresponds to one healthy donor (HD) sample out of five analysed (HD1 to HD5, as indicated). (C) Genera diversity of input, IgA1+, and IgA2+ fractions calculated using the Shannon index. Boxes extend from the 25th to the 75th percentiles. Error bars represent minimum and maximum values. (D) Relative composition of genera in fecal samples (input). Each column corresponds to one sample. The 25 most abundant genera are shown. (E) Relative abundance of families in input and sorted fractions from five healthy donors. The top 16 most abundant families are shown.
Figure 3.
IgA1 coordinates with IgA2 to coat distinct commensal bacteria. (A) Median relative abundance of the four most frequent families in sorted fractions from five healthy donors. Each dot represents one donor. Boxes extend from the 25th to the 75th percentiles. Error bars represent minimum and maximum values. (B) Median relative abundance of genera from IgA1+IgA2+ and IgA2+ fractions from five healthy donors. Each dot represents one donor. (C) Relative abundance of indicated top 19 most abundant genera in sorted fractions from one healthy donor. (D) Specificity of IgA subclass targeting for all individuals analyzed (n = 5). Number of samples in which a given genera had a positive indicated IgA subclass EI, divided by the total number of samples. The formula is detailed in Materials and methods. (E) Flavobacterium enrichment in IgA2+ as compared with IgA1+IgA2+ fractions. Each dot represents one donor. P values were defined using the Wilcoxon test (*, P < 0.05). (F) Binding of purified breast milk IgA to indicated bacterial strains. Monoclonal IgA isotype control (gray-filled histogram) was included as negative flow cytometry control. The same experiment was repeated twice. ns, not significant.
Figure S3.
IgA1- and IgA2-coated bacteria promote cytokine production by macrophages. (A) One out of three representative flow cytometric analyses of human gut microbiota purified from an IgA-deficient donor incubated with breast milk IgA and subsequently with anti-IgA FITC (three independent experiments). Numbers indicate percentage of positive cells. (B) Cytokine levels measured using Simoa technology in supernatants of macrophages incubated for 24 h with heat-killed S. haemolyticus opsonized with IgA1 (blue) or IgA2 (red) or without IgA (gray). The Mann-Whitney test was used to calculate P values; ns, not significant (n = 3 healthy donors, two independent measurements). Error bars indicate maximum values.
Figure S4.
Human monoclonal IgA bind a broad but nevertheless private pattern of commensals. Related to Fig. 4. (A) Flow cytometric sorting of intestinal IgA+ memory B cells (defined as CD19+, cell surface IgA+ IgD−). Sort gate among CD19+ B cells is shown. Doublets and dead cells were excluded before CD19 gating, CD19+ cells were gated among CD45+ cells (not shown). Representative images from three independent sorts are shown. (B) Transduced B cells exhibited a stable germinal center–like phenotype and maintained IgA expression. Transduced B cells were surface-labeled with anti-CD38, anti-CD95, or anti-IgA (orange lines) or appropriate isotype antibody controls (gray dotted lines). Representative images of eight monoclonal B cell lines, evaluated in three independent experiments, are shown. (C) Monoclonal B cell lines produced dimeric IgA. Representative immunoblotting showing high molecular weight dimeric mAb in nonreducing conditions for mAb 8. Representative image of eight mAbs and two independent experiments are shown. (D) Representative flow cytometric plot of microbiota B stained with anti-IgG Alexa Fluor 647 and anti-IgA FITC. The same experiment was repeated twice. (E) Heatmap diagram of EI of the 50 most frequent genera from microbiota A. Hierarchical clustering grouped mAb+ fractions and genera. (F) Heatmap diagram of EI of the 50 most frequent genera from microbiota B. Hierarchical clustering grouped mAb+ fractions and genera. (G) Flow cytometric analysis of mAb or negative control (mAb− supernatant [left] or irrelevant IgG [right, anti-TNFα IgG1]) staining of pure bacterial strains (two independent experiments).
Figure 4.
Human monoclonal IgA target highly diverse commensal bacteria. (A) Representative flow cytometry plot of microbiota reactivity for mAb#2 and human IgA2 anti-TNP. mAbs 1–8 are gut monoclonal IgA2 expressed as dimeric IgA. SSC-A, side scatter area. (B) mAb coating of IgA-free microbiota. mAbs are classified in increasing fluorescence intensity order (median fluorescence intensity [MFI], range 2,080–10,724). The same experiment was repeated twice. (C) Somatic mutations of mAbs are not correlated to IgA staining intensity. Somatic mutations in the V-region of IGH gene were analyzed. Nonparametric Spearman correlation was calculated. (D) Representative flow cytometry plot of microbiota reactivity for mAb 10 and human IgG1 anti-TNFα. mAbs 9–16 are gut monoclonal IgA2 expressed as IgG1 (Benckert et al., 2011). (E) mAb coating of IgA and IgG-free microbiota. MAbs are classified in increasing fluorescence intensity order (MFI, range 3,585–17,683). The same experiment was repeated twice. (F) Somatic mutations of mAbs are not correlated to mAb staining intensity. Somatic mutations in V-region of IGH gene were analyzed. Nonparametric Spearman correlation was calculated. (G) mAb+ and mAb− fractions of IgA-free gut microbiota were sorted by flow cytometry, and their composition was analyzed by 16S rRNA sequencing. (H) Relative abundance of phyla in whole microbiota (input) and mAb+ fractions. Microbiota A is IgA-free, while microbiota B is IgA- and IgG-free. 16S rRNA sequencing data were from two independent experiments. (I) Relative abundance of genera in whole microbiota A (input) and mAb+ fractions. (J) Relative abundance of genera in whole microbiota B (input) and mAb+ fractions. (K) Scatter dot plot of median relative abundance of genera from mAb+ and mAb− fractions. ns, not significant.
Figure 5.
Microbial surface glycans are common targets of gut human IgA. (A) Glycan reactivity for five mAbs (5 µg/ml) was assessed using glycan microarray technology (660 structures). Representative median RFUs of mAb1 are shown. Glycan ID numbers of top-bound glycans are indicated. (B) The median RFUs for top-bound glycans (above background) are shown for all five mAbs tested. (C) The ABR of glycans showing positive reactivity (above background, as in I) was calculated using an isotype control and visualized in an ordered (by dendrogram algorithm) reactivity matrix (heatmap). Each column represents a glycan. Color key is shown. (D) Representative flow cytometric dot plot of microbiota-reactivity for mAb 6 after whole-microbiota enzymatic deglycosylation (left). Decrease of mAb binding to microbiota after enzymatic deglycosylation (right). Columns and error bars represent median and maximum values, respectively. The same experiment was repeated twice. SSC-A, side scatter area. (E) Representative dot blots of purified PG and lipoteichoic acid (TA) from S. aureus and S. haemolyticus probed with irrelevant IgG1 (anti-TNFα IgG1), mAb 9, and mAb 10 (gray, red, and green circles, respectively). mAbs 9 and 10 were generated with IgG1 Fc domain. The same experiment was repeated twice.
Figure S5.
Self-reactivity and glycan reactivity of antigen-selected secretory IgA . (A) Median frequency of nonsilent (black) and silent (gray) somatic mutations in CDRs and VH FWRs in mAb IGH genes (n = 16 mAbs, four independent experiments). (B) Self-reactivity was tested by IFA with HEp-2000 cells: (1) negative control; (2) positive control: autoimmune human serum containing anti-DNA; (3) purified IgA from fecal water (20 µg/ml); (4) negative staining representative of 15 nonreactive mAbs; and (5) mAb 4. Representative images of three independent experiments are shown.
References
- Arbatsky N.P., Shashkov A.S., Literacka E., Widmalm G., Kaca W., and Knirel Y.A.. 2000. Structure of the O-specific polysaccharide of Proteus mirabilis O11, another Proteus O-antigen containing an amide of D-galacturonic acid with L-threonine. Carbohydr. Res. 323:81–86. 10.1016/S0008-6215(99)00257-8 - DOI - PubMed