Conserved natural IgM antibodies mediate innate and adaptive immunity against the opportunistic fungus Pneumocystis murina

Abstract

Host defense against opportunistic fungi requires coordination between innate and adaptive immunity for resolution of infection. Antibodies generated in mice vaccinated with the fungus Pneumocystis prevent growth of Pneumocystis organisms within the lungs, but the mechanisms whereby antibodies enhance antifungal host defense are poorly defined. Nearly all species of fungi contain the conserved carbohydrates β-glucan and chitin within their cell walls, which may be targets of innate and adaptive immunity. In this study, we show that natural IgM antibodies targeting these fungal cell wall carbohydrates are conserved across many species, including fish and mammals. Natural antibodies bind fungal organisms and enhance host defense against Pneumocystis in early stages of infection. IgM antibodies influence recognition of fungal antigen by dendritic cells, increasing their migration to draining pulmonary lymph nodes. IgM antibodies are required for adaptive T helper type 2 (Th2) and Th17 cell differentiation and guide B cell isotype class-switch recombination during host defense against Pneumocystis. These experiments suggest a novel role for the IgM isotype in shaping the earliest steps in recognition and clearance of this fungus. We outline a mechanism whereby serum IgM, containing ancient specificities against conserved fungal antigens, bridges innate and adaptive immunity against fungal organisms.

Figure 1.

nAbs against conserved fungal cell wall carbohydrates are produced in diverse species, without requirement for exogenous antigenic stimulation. (A) Characterization of serum IgM antibodies in BALB/c mice before and after intratracheal challenge with Pneumocystis (PC) against fungal cell wall carbohydrates laminarin (primarily β-1,3 linked glucan), chitosan/chitin, and mannan by ELISA (n = 5–6 mice per time point from two independent experiments). (B) Relative quantity of IgM specific for β-glucan and chitosan/chitin in C57BL/6 mice reared under conventional SPF or germ-free (GF) conditions was measured by ELISA (n = 6 pooled from two independent experiments). Data are shown as OD₄₅₀ with the background of secondary antibody alone subtracted. In all cases, the secondary antibody alone yielded an OD₄₅₀ value of <0.050. (C) ELISPOT detection of individual anti–β-glucan IgM-producing B cells at various tissue sites in BALB/c mice (n = 6 per condition pooled from two independent experiments). (D) Anti–β-glucan IgM-producing CD5⁺ and CD5⁻ B cells sorted from the spleen of BALB/c mice and detected by ELISPOT (n = 6 per condition pooled from two independent experiments; *, P < 0.05 by Mann-Whitney test). SFU, spot-forming unit. (E–G) Reactivity of tetrameric IgM homologue in catfish sera (E), human cord blood IgM (F), or rhesus macaque cord blood IgM (G) with β-glucan and chitosan/chitin was measured by ELISA (n = 6–8 per group pooled from two independent experiments). All error bars represent SEM.

Figure 2.

Natural IgM antibodies bind fungal cell walls, and transfer of nAb-containing serum from WT SPF mice enhances immune responses and impairs growth of Pneumocystis in the lungs of susceptible SCID mice. (A) Representative histograms of IgM binding. Fungal cells or fungal cell walls were incubated with serum diluted 1:4 in 2% BSA/PBS, washed, and then probed with anti-Ig Cy3 conjugates and studied by flow cytometry. Serum from BALB/c SPF mice (I) or human cord blood (II) was incubated with zymosan particles and then probed with anti–mouse IgM or anti–human IgM conjugate antibodies, respectively (red line). Black line indicates primary incubation with SCID sera (I) or anti–human IgM conjugate antibody alone (II; the dotted line in II indicates unstained control). (III) Cyst-enriched Pneumocystis preparation incubated with BALB/c SPF sera followed by anti–mouse IgM conjugate antibody. (IV, top) A. fumigatus resting (solid lines) and swollen conidia (dotted lines) incubated with serum from SCID (black lines) or BALB/c SPF mice (red lines), followed by anti–mouse IgM conjugate antibody. (bottom) Relative binding of mouse sera natural IgG (green line) versus natural IgM (red line) from WT SPF mice to A. fumigatus swollen conidia. Secondary antibody staining with anti–mouse IgG conjugate antibody was not different than when SCID sera control was used as primary (not depicted). FACS plots are representative of three independent experiments. (B) BALB/c SPF serum diluted 1:4 in 2% BSA/PBS was preincubated with laminarin before exposure to zymosan (laminarin final concentration is expressed). Particles were washed and then probed with anti–mouse IgM conjugate antibody. (C and D) Cohorts of SCID mice received a total of 400 µl of serum (100 µl of serum i.v. and 300 µl of serum i.p.), from either SPF WT mice or SCID mice, and 1 h thereafter were challenged with Pneumocystis intratracheally. Pathogen burden was determined at 5 (C) and 14 d (D) by assessing lungs for total Pneumocystis (PC) mitochondrial large subunit ribosomal RNA (mtLSU) subunit copy numbers by real-time PCR. Data are plotted as mean ± SEM (n = 6–8 pooled from three independent experiments; *, P < 0.05; **, P < 0.01). (E and F) Kinetics of IL-1β (E) and IL-6 (F) production in the lung homogenate of infected mice were determined by Luminex assay (n = 6 pooled from two independent experiments; *, P < 0.05). Error bars represent SEM.

Figure 3.

Mice unable to produce sIgM have impaired pulmonary clearance of Pneumocystis and altered inflammatory responses in local DLNs and in lung tissue. (A and B) C57BL/6J WT and sIgM-deficient (KO) mice were challenged with Pneumocystis intratracheally, and 14 d thereafter, pulmonary Pneumocystis (PC) burden (A) was assessed by real-time PCR. Symbols represent individual mice; horizontal lines represent the mean (n = 8–9 pooled from three independent experiments; **, P < 0.01 by Mann-Whitney test). mtLSU, mitochondrial large subunit ribosomal RNA. (B) IL-1β, IL-6, and IL-10 in lung homogenate were measured at 14 d (n = 6–8 pooled from two independent experiments; *, P < 0.05 by Mann-Whitney test). Cytokine concentrations in uninfected lungs were all <10 pg/ml. Error bars represent SEM. (C) Mediastinal DLNs isolated from WT or sIgM KO mice 14 d after intratracheal Pneumocystis challenge were gently teased apart, and cells from each LN were individually restimulated with or without Pneumocystis antigen and cultured for 3 d. Supernatants were assessed for production of IL-5, IL-9, IL-17, and IFN-γ by Luminex assay (n = 6 mice per group pooled from two independent experiments). Data are mean ± SEM (*, P < 0.05 by Mann-Whitney test). Cytokine responses without antigen stimulation were <25 pg/ml in these experiments.

Figure 4.

nAbs influence the earliest aspects of fungal antigen presentation and infiltration of inflammatory cells into the lungs. (A) C57BL/6J WT or sIgM^−/− (KO) mice were intratracheally challenged with 500 µg FITC-zymosan delivered in 50 µl PBS. 18 h later, mediastinal DLNs were harvested and dissociated into single-cell suspensions. Cells were stained with CD11c-APC, and FITC association was assessed by flow cytometry. Representative WT or sIgM KO individual LNs are presented. (B) Recovery of CD11c⁺ FITC^hi cells in LNs per cohort (n = 5–9 mice per group pooled from two independent experiments; *, P < 0.05 by Mann-Whitney test). (C–H) Bone marrow–derived DCs were grown in RPMI 1640 +10% FCS + GM-CSF + IL-4. After 6 d, DCs were collected, washed, and cultured in serum-free RPMI 1640. Pneumocystis cysts obtained from chronically infected SCID mice were unopsonized or opsonized with WT serum or serum from sIgM KO mice and added to DCs at a dose of 10 cysts per cell. Cell supernatants were assayed for TNF, IL-1β, IL-6, or IL-12p40 by Luminex or IL-23 by ELISA (n = 6 mice per group pooled from two independent experiments; *, P < 0.05). In H, total RNA was extracted, and relative expression of Ccr7 was assayed by quantitative real-time PCR normalized to GAPDH (n = 6 mice per group pooled from two independent experiments; *, P < 0.05). Error bars represent SEM.

Figure 5.

Mice unable to secrete IgM have diminished Th2 and Th17 responses in mediastinal LNs after Pneumocystis challenge and impaired Th2-type adaptive B cell responses. C57BL/6J WT and sIgM KO mice were challenged with Pneumocystis intratracheally. (A) 14 d after challenge, mediastinal LNs were harvested, and cells were dissociated into single-cell suspensions, enumerated, and assessed for production of IL-5, IL-17, and IFN-γ after 2-d culture by ELISPOT. In the absence of infection, spot frequencies are <5 spot-forming units (SFUs) per million cells (not depicted; n = 6 mice per group pooled from two independent experiments). (B) Before challenge, Pneumocystis inoculum was preopsonized with serum from either C57BL/6J WT or sIgM KO mice and thereafter administered to sIgM KO mice intratracheally. 14 d after infection, individual LNs were collected, and cells were dissociated into single-cell suspensions, enumerated, and stimulated with Pneumocystis antigen in culture for 2 d. IL-5 and IL-17 production was measured in tissue culture supernatants by Luminex assay (n = 5–6 pooled from two independent experiments). Data are mean ± SEM. (C) Serum was collected serially from WT and sIgM KO mice challenged intratracheally with Pneumocystis and probed for adaptive IgG1 and IgG2c responses against Pneumocystis antigen. BALF was collected at 28 d after infection and probed for total IgG and IgA responses against Pneumocystis antigen (n = 5–6 mice/condition/time point pooled from two independent experiments). Error bars represent SEM (*, P < 0.05; **, P < 0.01 by Mann-Whitney test).