Report

Inflammation-induced disruption of SCS macrophages impairs B cell responses to secondary infection

See allHide authors and affiliations

Science  06 Feb 2015:
Vol. 347, Issue 6222, pp. 667-672
DOI: 10.1126/science.aaa1300

Bacterial infection breaks the lymph node barrier

During infections, lymph nodes are command central. Fragments from invading pathogens enter lymph nodes through the lymph. There, specialized cells called subcapsular sinus (SCS) macrophages capture these antigens and use them to initiate humoral immunity. Despite being such important players, Gaya et al. report that in mice, infection throws these organized sentinels into disarray (see the Perspective by Buzsaki). Disrupting SCS macrophages had important consequences: Bacterially infected mice could not respond as efficiently to a subsequent viral infection.

Science, this issue p. 667; see also p. 612

Abstract

The layer of macrophages at the subcapsular sinus (SCS) captures pathogens entering the lymph node, preventing their global dissemination and triggering an immune response. However, how infection affects SCS macrophages remains largely unexplored. Here we show that infection and inflammation disrupt the organization of SCS macrophages in a manner that involves the migration of mature dendritic cells to the lymph node. This disrupted organization reduces the capacity of SCS macrophages to retain and present antigen in a subsequent secondary infection, resulting in diminished B cell responses. Thus, the SCS macrophage layer may act as a sensor or valve during infection to temporarily shut down the lymph node to further antigenic challenge. This shutdown may increase an organism’s susceptibility to secondary infections.

The highly organized architecture of the lymph node (LN) is critical for mounting effective immune responses against pathogens. One particular facet of this organization is the layer of CD169+ macrophages at the subcapsular sinus (SCS) floor; strategically positioned at the lymph-tissue interface to capture pathogens as they enter the LN (1). This prevents systemic dissemination of pathogens (25) and allows presentation of intact antigen in the form of immune complexes, viruses, and bacteria to cognate B cells for the initiation of humoral immune responses (2, 68).

Infection causes a remodeling of the global architecture of the LN (9, 10). However, how this process affects the organization of the SCS macrophage layer is not well defined. To address this, we visualized the distribution of SCS macrophages in draining LNs of C57BL/6 mice after ear skin infection with Staphylococcus aureus, the most common etiological organism of skin and soft tissue infection. Cryosections of superficial cervical LNs were immunostained and examined 7 days after infection by confocal microscopy. The overall dimensions of the LNs were increased around fourfold with considerable enlargement of the B220+ follicular regions and a marked disruption of the CD169+ population at the SCS and interfollicular regions (Fig. 1A). We observed a similar disruption in the SCS macrophage organization after localized infection with influenza A virus, Vaccinia virus (VACV), and group B Streptococcus (GBS) but not after administration of ultraviolet (UV)–inactivated virus or inert beads (fig. S1, A to C). Notably, mice receiving either the Toll-like receptor 9 (TLR9) agonist CpG or the TLR4 agonist lipopolysaccharide (LPS) in the footpad display similar changes in the LN organization (Fig. 1B and figs. S1D and S2). This is a temporary process as the compact layer of SCS macrophages is reestablished after 28 days (fig. S1E). Overall, these data show that disruption of SCS macrophages in draining LNs is a frequent feature associated with inflammation and viral or bacterial infections.

Fig. 1 Infection and inflammation disrupt SCS macrophage organization in draining LNs.

(A and B) Draining LN cryosections stained for CD169 (green) and B220 (red) derived from mice administered either PBS (control), (A) 107 colony-forming units (CFU) of S. aureus in the ear 7 days previously, or (B) 10 μg of CpG or 50 μg of LPS in the footpad 4 days previously. Scale bars, 300 μm (top); 60 μm (bottom). Bar charts show the quantification of SCS macrophage disruption and distance of macrophages to LN border in each condition for an individual experiment (see materials and methods). Each dot represents the analysis of a distinct follicle. Data are shown as mean ± SEM and are representative of at least three independent experiments. (C) Three-dimensional multiphoton microscopy of explanted popliteal LNs from animals injected in the footpad with either PBS or CpG (4 days) and anti-CD169 (green) 10 min before dissection. Second harmonic signal generated by collagen fibrils is shown (cyan). Scale bars, 500 μm (left); 40 μm (middle and right). Bar charts show the number and density of CD169+ macrophages in the SCS in each condition from three independent experiments. Each dot represents an individual LN. Data are shown as mean ± SEM. (D) Representative Z sections (right) and schematics (left) of 3View electron microscopy analysis of popliteal LNs after 4 days of PBS or CpG administration. Yellow stars indicate macrophages. Scale bar, 5 μm. Bar chart represents the distance of macrophages to LN SCS in an individual experiment. Each dot indicates a single macrophage. Data are shown as mean ± SEM and are representative of two independent experiments. Student’s t test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

We further investigated this loss of integrity in the SCS macrophage layer following inflammation by using two high-resolution technologies. Three-dimensional (3D) multiphoton imaging of whole explanted LNs showed a decrease in the number and density of CD169+ macrophages accompanied with a change in morphology and retraction of these cells from the SCS (Fig. 1C and movie S1). Three-dimensional electron tomography, scanning 100-nm LN sections over a distance of 100 μm, revealed that macrophages are located toward the follicular interior during inflammation, whereas they are positioned longitudinally on the inner wall of the SCS on steady state (Fig. 1D and movie S2). Together, this shows that the characteristic disruption of SCS macrophages is a result of both cell loss and displacement of cells toward inner follicular areas.

To analyze the mechanism by which inflammation triggers alterations in SCS macrophage organization, we used the CpG-mediated inflammation model and a series of mouse strains in which different components of the TLR signaling pathway have been genetically ablated. Mice deficient in TLR9, MyD88-TRIF adaptors, or MyD88 alone did not exhibit SCS macrophage disruption, concordant with the notion that responses to CpG are mediated by TLR9 signaling via MyD88 (Fig. 2A and fig. S3). To determine which MyD88-expressing immune cell population is required for this process, we deleted MyD88 in B cells (Myd88flox/flox Cd19 Cre+ mice), dendritic cells (DCs) (Myd88flox/flox Cd11c Cre+ mice), and neutrophils (Myd88flox/flox Lyz2 Cre+ mice) (fig. S4). As revealed by flow cytometry, MyD88 abundance in SCS macrophages remains unchanged in the different transgenic models (fig. S4). B cells are not required for SCS macrophage disruption, as this process was observed in Myd88flox/flox Cd19 Cre+ and Rag2-deficient mice (Fig. 2A and fig. S5A). Furthermore, the disruption was not prevented by lack of expression of MyD88 in neutrophils or their in vivo depletion with an antibody against Ly6G (anti-Ly6G) (Fig. 2A and fig. S5, B and C). In contrast, whereas draining LNs were enlarged in Myd88flox/flox Cd11c Cre+ mice following inflammation, disruption of the SCS macrophage layer was significantly reduced (Fig. 2A and fig. S3B). These observations demonstrate that SCS macrophage disruption during inflammation is not simply due to associated enlargement of draining LNs but requires the expression of MyD88 in the DC population.

Fig. 2 DC arrival at draining LN during inflammation is necessary for SCS macrophage disruption.

(A) Confocal microscopy images of popliteal LNs from WT, Myd88−/−, Myd88flox/flox Cd19 Cre+, Myd88flox/flox Lyz2 Cre+, and Myd88flox/flox Cd11c Cre+ mice collected 4 days after footpad administration of PBS or CpG. Cryosections were stained with anti-mouse monoclonal antibodies (mAbs) to CD169 (green) and B220 (red). Scale bars, 300 μm (top); 60 μm (bottom). Bar charts represent quantification of SCS macrophage disruption and distance of macrophages to LN border in each condition for an individual experiment. Each dot represents a distinct follicle. Data are shown as mean ± SEM and are representative of three independent experiments. (B and C) Confocal microscopy images of popliteal LN sections stained with mAbs to CD169 (green) and (B) Langerin (magenta) or (C) CD11c (magenta) derived from mice that were injected with PBS or CpG 4 days previously. Scale bars, 300 μm (top); 60 μm (bottom). Histograms represent CD169, Langerin, and CD11c fluorescence measured from the outer edge of the LN to the inner areas. Stars indicate DCs at the SCS. (D) Confocal images of popliteal LNs cryosections stained as in (A) from WT and Ccr7−/− mice 4 days after footpad administration of PBS or CpG. Scale bars, 300 μm (top); 60 μm (bottom). Quantifications were performed as in (A). Data are representative of three independent experiments. (E) Confocal images of popliteal LN cryosections 4 days after footpad injection with 3 × 106 carboxyfluorescein diacetate succinimidyl ester (CFSE)–labeled (magenta) control BMDCs, CpG-treated WT BMDCs, or CpG-treated Tlr9−/− BMDCs. Sections were stained with mAb to CD169 (green). Scale bars, 300 μm (top); 60 μm (bottom). Quantifications were performed as in (A). Data are representative of three independent experiments. Student t test, *P < 0.05, ***P < 0.001, ****P < 0.0001.

In line with these observations, we found a significant accumulation of DCs in the SCS following inflammation (Fig. 2, B and C, and fig. S6, A and B). Therefore, to determine if DC arrival is necessary for the disruption of the SCS macrophage layer, we injected phosphate-buffered saline (PBS) or CpG into either wild-type mice or mice lacking CCR7, the main LN-homing chemokine receptor for DCs (fig. S6, C and D) (11). Indeed, disruption of the macrophage layer was significantly reduced in CCR7-deficient animals (Fig. 2D). To evaluate whether DC arrival to the LN is sufficient for this disruption, we adoptively transferred into the footpad bone marrow–derived DCs (BM-DCs) that were previously treated with CpG in vitro. We observed disruption of the SCS macrophage layer in draining LNs of animals receiving CpG-treated BM-DCs but not control DCs; neither PBS-treated nor TLR9-deficient DCs (Fig. 2E and fig. S6E). Noticeably, a marked disruption was also observed when BM-DC migration was induced independently of TLR signaling with prostaglandin E2 (12), suggesting that DC migration per se can alter SCS macrophage organization (fig. S6F). Our observations suggest that the arrival of mature DCs during inflammation plays an important role in the disruption of SCS macrophages

We next examined the functional implications of the impaired integrity of SCS macrophages for the retention and presentation of antigen arriving in a subsequent wave. We treated mice with PBS (control), CpG, VACV, or GBS and then challenged them with fluorescent microspheres (0.2 μm) or fluorescently labeled GBS. As described previously (6), antigen particles localized at the SCS of LNs from control mice after 6 hours (Fig. 3A). However, this accumulation was diminished in mice that had previously received CpG, VACV, or GBS (Fig. 3, A and B, and fig. S7, A to D). Antigen retention was not impaired owing to disruption in lymphatic flow or antigen transport (fig. S7, E and F) and was restored by 4 weeks after initiation of inflammation (fig. S7G). Thus, infection or inflammation has a temporary, functional impact on the capacity of SCS macrophages to retain antigen from a subsequent wave.

Fig. 3 Inflammation impedes acquisition of subsequent antigen by SCS macrophages and cognate B cells.

(A) Confocal microscopy images of popliteal LNs from mice injected in the footpad with PBS or CpG and 4 days later injected again with 8 × 108 fluorescent microspheres (0.2 μm, white). Sections were labeled with antibodies to B220 (red) and CD169 (green). Scale bar, 70 μm. Bar charts represent beads fluorescence quantification in SCS and follicles from three independent experiments. (B) Flow cytometry analysis of popliteal LNs from animals treated as in (A). Representative dot plots depict beads acquisition by B220CD3 cells. Quantification of beads-positive cells from three independent experiments is shown in the bar chart. (C) Confocal microscopy images of LNs from mice that were (i) adoptively transferred with 5 × 106 SNARF-labeled MD4+ Tlr9−/− B cells (red) on day 0, (ii) administered PBS or CpG in the footpad on day 1, (iii) injected again with 8 × 108 avidin fluorescent particles (green) coated or not with HEL on day 5, and (iv) killed after 6 hours. Sections were stained with B220 antibody (blue). Scale bar, 20 μm. Bar charts represent the proportion of MD4 B cells loaded with particles for an individual experiment. Data are representative of three independent experiments. (D and E) ELISPOT analysis of HEL-specific ASCs (day 14) in popliteal LNs of animals that were (i) adoptively transferred with (D) 5 × 106 MD4 Tlr9−/− B cells and 5 × 106 OT-II T cells or (E) 2 × 106 MD4 Tlr9−/− B cells, (ii) injected with PBS or CpG in the footpad on day 1, and (iii) injected again with (D) OVA-HEL beads or (E) αGalCer-HEL beads on day 7. Bar charts represent the number of HEL-specific ASCs in each condition for an individual experiment. Data are representative of three independent experiments. In all panels, each dot in bar charts represents a single mouse. Data are shown as mean ± SEM. Student’s t test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Antigen retained by SCS macrophages can be presented to follicular B cells, so we examined the effect of inflammation on the ability of cognate B cells to acquire antigen arriving at the LN in a subsequent wave. SNARF-labeled Tlr9−/− MD4 B cells, expressing a transgenic B cell receptor specific for hen egg lysozyme (HEL) but unable to respond to CpG, were adoptively transferred into recipient animals. These mice were treated with either PBS (control) or CpG and later received either uncoated or HEL-coated fluorescent microspheres. In control mice, ≈50% of MD4 B cells acquired one to five HEL-coated microspheres, whereas in CpG-treated mice, only 3% of B cells acquire HEL-coated microspheres and never more than one per cell (Fig. 3C and fig. S7, H and I). Thus, concordant with the reduction in antigen retention by SCS macrophages, the acquisition of cognate antigen by B cells is reduced in draining LNs following inflammation.

Does the observed reduction in antigen acquisition during inflammation affect the capacity of B cells to respond to subsequent antigen challenge in vivo? To address this, wild-type mice were adoptively transferred with Tlr9−/− MD4 B cells and OT-II T cells (ovalbumin-specific T cell receptor). Afterwards, they were treated with PBS or CpG followed by administration of HEL-ovalbumin–coated microspheres. LNs from control mice contained ≈103 HEL-specific antibody-secreting cells (ASCs), although this was significantly lower in LNs from CpG-treated mice (Fig. 3D). We observed a similar reduction in HEL-specific ASC formation when mice were adoptively transferred with Tlr9−/− MD4 B cells followed by injection of CpG and HEL-αGalCer–coated microspheres (13) (Fig. 3E). These approaches demonstrate that inflammatory signals affect the extent to which B cells can acquire and respond to further antigenic challenge in vivo.

We then examined the potential impact that inflammation-mediated disruption of the SCS layer has on B cell responses to a subsequent viral infection. Wild-type mice were treated with PBS (control) or CpG and then infected with VACV. Control LNs exhibited an expansion of GL7+Fas+Bcl6+ germinal center (GC) B cells and CD138+IgDlo virus-specific plasma cells (PCs) (Fig. 4, A to D). Instead, both GC and PC formation were impaired when infection was initiated after inflammation induction (Fig. 4, A to D). B cell responses to VACV were restored to control levels when infection occurred 4 weeks after CpG administration (fig. S8), a time frame consistent with the recovery of the structural integrity of the SCS macrophage layer (fig. S1E). Similar reductions in B cell responses were observed when Diphtheria toxin or clodronate liposomes were used to deplete SCS macrophages (14, 15) before VACV infection (fig. S9) or when mice received CpG or clodronate in the ear before infection with S. aureus (fig. S10). Therefore, it appears that the SCS macrophage disruption triggered by inflammation affects the ability of B cells to mount responses to viral or bacterial antigen arriving in a secondary wave.

Fig. 4 Inflammation and primary infection shut down B cell responses to subsequent pathogens.

(A) Flow cytometry analysis of GC formation (day 14) in popliteal LNs from mice that were administered PBS or CpG in the footpad and 7 days later infected with 104 plaque-forming units (PFU) of VACV. Representative contour plots display the percentage of B220+ cells that are GL-7+Fas+. Bar charts display the quantification of GL-7+Fas+ B cells in the different conditions for a single experiment. (B) Confocal microscopy analysis of popliteal LNs from mice treated as in (A). Sections were stained with mAbs to B220 (red) and Bcl-6 (cyan). Scale bar, 60 μm. Quantification of the GC area for an individual experiment is depicted on the right bar chart. (C) Flow cytometry analysis of PC formation in mice treated as in (A). Representative contour plots show the percentage of B220+ cells that are also CD138+IgDlow. The quantification of CD138+IgDlow B cells for a single experiment is shown on the right bar charts. (D) ELISPOT analysis of total, immunoglobulin M (IgM) and IgG VACV-specific ASCs in popliteal LNs from mice that were treated as in (A). Bar charts represent the number of VACV-specific ASCs for an individual experiment. (E) ELISPOT analysis of total, IgM, and IgG VACV-specific ASCs (day 14) from mice that were administered PBS or 106 CFU of GBS in the footpad and then infected with 104 PFU of VACV on day 7. Bar charts represent the number of VACV-specific ASCs from three independent experiments. In all panels, experiments were performed at least three times and each dot represents a different mouse. Data are shown as mean ± SEM. Student’s t test, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Finally, we asked whether SCS macrophage disruption during a primary infection impedes the ability of B cells to respond to a secondary pathogen. After PBS (control) or GBS administration, mice were infected with VACV. The formation of VACV-specific PCs was significantly reduced in GBS-infected mice (Fig. 4E). This is consistent with the notion that loss of integrity of the SCS macrophage layer during primary infections severely affects the capacity of B cells to respond to secondary pathogens.

SCS macrophages have been placed at the heart of antipathogen responses because of their key role in antigen trapping and immune response initiation (18, 1620). We found that infection or inflammation leads to a significant loss of CD169+ macrophages at the SCS and to a displacement of these cells toward inner follicular areas. Regardless of whether this is due to cell death (5) and/or macrophage redistribution, these observations raise the important question of the potential benefit of such a phenomenon. We speculate that this marked architectural reorganization might be beneficial in allowing the entry of afferent lymph-derived immune cells directly through the SCS floor (21); it may facilitate antigen relay to follicular DCs (7) or it might maximize presentation of antigen to B cells.

However, although these scenarios would be likely to enhance immune responses in a primary infection, the disruption of SCS macrophages would also render draining LNs temporarily refractory to newly arriving pathogens. This inability to respond to subsequent pathogens parallels recent findings that the failure of host defenses to counteract secondary infections results from loss of lymphoid tissue integrity or compromised innate host defense (22, 23). Here, we propose a model in which SCS macrophages function as a valve that senses inflammation within draining LNs, triggering the temporary shutdown of humoral responses to secondary infections to prioritize the effective control of contemporaneous lymph-borne infecting pathogens (fig. S11).

Supplementary Materials

www.sciencemag.org/content/347/6222/667/suppl/DC1

Materials and Methods

Figs. S1 to S11

References (2429)

Movies S1 and S2

References and Notes

  1. Acknowledgments: We thank A. Rot and D. Withers for the CCR7 knockout mouse strain. We thank L. Collinson, H. Armer, and C. Peddie from the electron microscopy facility for electron microscopy and 3View analysis of lymph nodes; the experimental histopathology facility for the initial preparation of tissue cryosections; and S. Lutter for assistance with whole-body imaging. We thank N. Harwood and J. Coleman for editing of the manuscript, and P. Barral J. Caamaño, R. Germain, and the members of the Lymphocyte Interaction Laboratory for critical reading of the manuscript. We thank F. Sallusto for feedback and suggestion of the term “shut down.” The data presented in this manuscript are tabulated in the main paper and in the supplementary materials. Supported by Cancer Research UK.
View Abstract

Navigate This Article