Acquisition of a multifunctional IgA+ plasma cell phenotype in the gut

Abstract

The largest mucosal surface in the body is in the gastrointestinal tract, a location that is heavily colonized by microbes that are normally harmless. A key mechanism required for maintaining a homeostatic balance between this microbial burden and the lymphocytes that densely populate the gastrointestinal tract is the production and transepithelial transport of poly-reactive IgA (ref. 1). Within the mucosal tissues, B cells respond to cytokines, sometimes in the absence of T-cell help, undergo class switch recombination of their immunoglobulin receptor to IgA, and differentiate to become plasma cells. However, IgA-secreting plasma cells probably have additional attributes that are needed for coping with the tremendous bacterial load in the gastrointestinal tract. Here we report that mouse IgA(+) plasma cells also produce the antimicrobial mediators tumour-necrosis factor-α (TNF-α) and inducible nitric oxide synthase (iNOS), and express many molecules that are commonly associated with monocyte/granulocytic cell types. The development of iNOS-producing IgA(+) plasma cells can be recapitulated in vitro in the presence of gut stroma, and the acquisition of this multifunctional phenotype in vivo and in vitro relies on microbial co-stimulation. Deletion of TNF-α and iNOS in B-lineage cells resulted in a reduction in IgA production, altered diversification of the gut microbiota and poor clearance of a gut-tropic pathogen. These findings reveal a novel adaptation to maintaining homeostasis in the gut, and extend the repertoire of protective responses exhibited by some B-lineage cells.

Figure 1. IgA⁺ plasma cells in the small intestinal lamina propria can produce iNOS and TNFα

a,b, Small intestinal lamina propria cells (LPC) from LTβ^−/− and LTβR^−/−, J_H^−/− and WT mice were isolated and viable cells were analyzed for CD11c and iNOS expression by flow cytometry. Please see quantification in Supplemental Table I. iNOS⁺ cells were gated and analysed for IgA expression. c, Small intestinal LPC cells were isolated from WT and TNFα^−/−iNOS^−/− double-deficient (dKO) mice and analysed for IgA, TNFα and iNOS expression (note a WT example here for IgA/B220 staining is shown. Although IgA-expressing cells are reduced in dKO mice as per Fig. 3c, it is possible to still gate on IgA-expressing cells within the LP of dKO mice). Frequencies of TNFα⁺ and iNOS⁺ populations can be found in Suppl. Table II. d, Intestinal LPC from AID-YFP animals were isolated and viable cells were analyzed for their expression of IgA and YFP. Specifically, YFP⁻IgA⁻ = grey rectangle, YFP⁺IgA⁻ = blue rectangle and YFP⁺IgA⁺ = red rectangle, please see Suppl. Table III for relative frequency of each population. These gated populations were further analyzed for their expression of iNOS and TNFα (note that cross-hairs were added based on isotype control staining; see Suppl. Table IV for average frequencies of each population). Expression of other lineage-specific markers for each of the three populations are denoted as histograms (YFP⁻IgA⁻ = grey trace, YFP⁺IgA⁻ = blue trace, YFP⁺IgA⁺ = red trace). Note that for CD11c expression, YFP⁻IgA⁻ dendritic cells were used as a positive control (grey trace). Representative plots are shown from n=9 mice. Sections of small intestines from WT versus dKO mice (e) and AID⁺YFP⁺ mice (f) were stained with specific fluorochrome-tagged antibodies for iNOS, TNFα and IgA (or visualized for YFP) as indicated. Stained sections were then analyzed by fluorescence microscopy at 200x. Representative pictures are shown from at least 3 separate experiments. Arrows denote areas of co-localization and the rectangle indicates a villus that was enlarged to visualize simultaneous expression of each iNOS, TNFα and YFP.

Figure 2. TNFα and iNOS-expression in IgA⁺ plasma cells requires microbial exposure

a, Sections of small intestines of germ-free (GF) and GF mice that were re-colonized with either a defined Altered Schaedler Flora (ASF), or reversibly re-colonized with HA107 E. coli by continual gavage administration for 14 days followed by 14 days without gavage returning them to a GF status. All sections were stained with specific fluorochrome-tagged antibodies for IgA (green), iNOS (red) and EpCAM (blue) and analyzed by fluorescence microscopy. Representative pictures are shown from 6 mice per group. b, B220⁺ bone marrow (BM) cells from CD45.1⁺ wild type mice were co-cultured for 7 days with CD45.2⁺ intestinal lamina propria cells (referred to as gut stroma) from either SPF or GF animals in the presence of IL-7, TGFβ, IL-21 and αCD40 antibody. In some cases faecal matter (intestinal wash) was added on the first day of the culture. The expression of IgA and iNOS was analyzed by flow cytometry, and the black boxes that indicate cells that are considered iNOS⁺ were based on the absence of staining derived from isotype controls (not shown). To ensure selective analysis of BM-derived precursors, cells were pre-gated on the CD45.2⁻ population. Representative flow cytometry plots of cells are depicted. Data are representative of at least two independent experiments with 3–4 mice per group per experiment.

Figure 3. Reduced IgA production and altered commensal flora composition in iNOS/TNFα double-deficient mixed chimeric mice

a–b, Steady-state serum IgA, but not IgG₁ levels are markedly decreased in dKO and dKO mixed chimeras. Serum IgA and IgG₁ levels were measured by ELISA in WT (n = 15) TNF^+/−iNOS^+/− (n = 12) and TNF^−/−iNOS^−/− (dKO, n = 7) non-chimeric mice as well as in WT + dKO (n = 8), J_H^−/− + WT (n = 5) and J_H^−/− + dKO (n = 6) mixed chimeric mice. Similar results were obtained with two additional batches of mixed bone marrow chimeric mice (not shown). Data represent means and SD values. c, The numbers of small intestinal LP IgA⁺ plasma cells and CD8α⁺ cells were quantified from IF microscopy images utilizing ImageJ. d, Representative images of frozen small intestinal tissue sections derived from the groups described in (a–c) stained for CD8α (red) and IgA (green) and DAPI (blue). A total of 5 different sections from 6 mice per group were analyzed. e, Seven weeks after BM reconstitution, the small intestines of WT and dKO mixed chimeras (n = 4) were analyzed for their commensal bacteria composition by quantitative real-time PCR amplification of 16S rRNA isolated from small intestinal preparations. Colored pie graphs represent the percentages of the indicated bacteria species. A significant difference in the relative representation of segmented filamentous bacteria (SFB) was observed between WT and dKO mice. Similar results for SFB were obtained in small intestinal “scrapes” of the epithelium (data not shown). Data represent means and SD values. *** p = 0.001, ** p = 0.01, * p < 0.05.

Figure 4. iNOS/TNFα double-deficient mixed chimeras are more susceptible to infection with Citrobacter rodentium

AID-YFP animals were infected by oral gavage with 1×10⁹ colony-forming units (CFU) of a nalidixic-acid resistant strain of Citrobacter rodentium DBS100. a–c, The kinetics of accumulation of viable YFP⁺IgA⁻CD11c⁻B220⁺ (black bars), YFP⁻CD11c⁺ (blue bars) and YFP⁺IgA⁺ (red bars) cells in the small intestine, the caecum (C) and the large intestine (LI) as analyzed by flow cytometry (day 0 (D0), 36 hours (36hrs), day 4 (D4) and day 7 (D7)). The percentages of viable cells as mean values and SD are shown (n = 4). A representative example of two individual experiments is shown. d, The colonization by C. rodentium was determined 9 days after infection by homogenizing spleens (Sp), large intestines (LI) and caecums (C) followed by serial dilution plating on nalidixic acid-containing LB plates and counting of colonies. C. rodentium colonization of spleen (Sp) is significantly enhanced in dKO mice at day 9 post- and at D4 post-infection (data not shown). Data are representative of two individual experiments. e, The percentage of body WT + J_H ^−/− →RAG^−/− and dKO + J_H ^−/− →RAG^−/− mixed chimeras after C. rodentium weight loss in infection over time is depicted. Significantly increased body weight loss was observed in dKO mice from day 6 to 9 post-infection (n=5–12 per group). NB: dKO + J_H^−/− mixed chimeric mice were sacrificed 9-days post infection for humane reasons as weight loss exceeded 20% of their original body weight. f, Large intestines (LI) from WT + J_H^−/− and dKO + J_H^−/− mixed chimeric mice were harvested 9 days post-infection and the pathological scores were analyzed by standard histological staining procedures using hematoxylin and eosin (H&E). LI’s from dKO + J_H^−/− mice show more severe LI pathology. H&E stainings of two representative dKO mice with a moderate and a more severe condition, as well as one representative WT + J_H^−/− chimeric mouse are shown. The panel shows the original magnification of 100 x. Scale bars represent 100 mm. * p < 0.05.