EBI2-mediated bridging channel positioning supports splenic dendritic cell homeostasis and particulate antigen capture

Abstract

Splenic dendritic cells (DCs) present blood-borne antigens to lymphocytes to promote T cell and antibody responses. The cues involved in positioning DCs in areas of antigen exposure in the spleen are undefined. Here we show that CD4(+) DCs highly express EBI2 and migrate to its oxysterol ligand, 7α,25-OHC. In mice lacking EBI2 or the enzymes needed for generating normal distributions of 7α,25-OHC, CD4(+) DCs are reduced in frequency and the remaining cells fail to situate in marginal zone bridging channels. The CD4(+) DC deficiency can be rescued by LTβR agonism. EBI2-mediated positioning in bridging channels promotes DC encounter with blood-borne particulate antigen. Upon exposure to antigen, CD4(+) DCs move rapidly to the T-B zone interface and promote induction of helper T cell and antibody responses. These findings establish an essential role for EBI2 in CD4(+) DC positioning and homeostasis and in facilitating capture and presentation of blood-borne particulate antigens. DOI:http://dx.doi.org/10.7554/eLife.00757.001.

Keywords: Dendritic cells; EBI2; Mouse; Oxysterols; antigen capture; lymphotoxin; spleen bridging channel.

Figure 1.. EBI2 expression in DCs and deficiency of CD4⁺ DCs in mice lacking EBI2 or correct amounts of EBI2 ligand.

(A) Flow cytometric detection of GFP fluorescence in gated splenic CD11c⁺MHCII⁺ cells from Ebi2^GFP/+ mice. (B) Quantitative PCR analysis of Ebi2 transcript abundance in sorted splenic CD11c⁺MHCII⁺CD4⁺ cells (CD4⁺ DCs) and CD11c⁺MHCII⁺CD8⁺ cells (CD8⁺ DCs). Expression is shown relative to Hprt (n = 4 mice). (C) EBI2 surface staining of gated splenic CD4⁺ and CD8⁺ DCs from Ebi2^−/− or Ebi2^+/+ mice (one representative of four experiments). (D) Migration of CD4⁺ DCs and CD8⁺ DCs in response to 7α,25-OHC (mean from four mice ± SE, combined from two experiments). (E)–(I) Flow cytometry and enumeration of splenocytes (E, F, and G), inguinal LN cells (H), and mesenteric LN cells (I) from Ebi2^−/−, Ch25h^−/−, Cyp7b1^−/−, Hsd3b7^−/− mice, and their matched littermate controls. Numbers adjacent to outlined areas in E indicate percent cells in each gate. DN DCs are defined as CD4^-CD8^-CD11c⁺MHCII⁺ cells. Each dot in F–I represents an individual mouse and error bars indicate mean ± SE of samples combined from three to five independent experiments. Lymph node migratory DCs are defined as MHCII^hiCD11c^int and resident DCs, including the CD8⁺ and 33D1⁺ DCs, as MHCII^intCD11c^hi. *p<0.05, **p<0.01, ***p<0.001, Student’s T-test. DOI: http://dx.doi.org/10.7554/eLife.00757.003

Figure 1—figure supplement 1.. DC properties in EBI2-deficient mice and in mice lacking CYP7B1 in radiation resistant cells.

(A) Quantitative PCR analysis of Ebi2, Ccr7, Cxcr4, and Cxcr5 transcript abundance in sorted CD11c⁺MHCII⁺CD4⁺ cells (CD4⁺ DCs), CD11c⁺MHCII⁺CD8⁺ cells (CD8⁺ DCs), and CD4^-CD8^-CD11c⁺MHCII⁺ cells (DN DCs). Expression is shown relative to Hprt. (B) Cyp7b1^+/- or Cyp7b1^-/- recipients were reconstituted with WT bone marrow cells and analyzed by flow cytometry for the indicated DC markers (n = 8, combined from three experiments). (C) Flow cytometric detection of 33D1 expression in gated splenic DC subsets. One representative of three replicated experiments is shown. (D) MHCII, CD86, CD83, and CD80 expression on gated CD11c⁺MHCII⁺ cells from Ebi2^+/− and Ebi2^−/− mice. Gating is based on isotype control and mean ± SE frequencies are shown for positive staining cells (n = 4). (E) Sorted CD4⁺ and CD8⁺ DCs from Ebi2^+/− and Ebi2^−/− mice were cultured with purified Cell Trace violet labeled allogeneic BALB/c CD4⁺ T cells. After three days of culture, T cell proliferation was visualized by flow cytometry. Gated percentage of proliferating cells is shown as mean ± SE (n = 3). (F) Flow cytometric detection of GFP fluorescence in gated lymph node DCs from Ebi2^GFP/+ mice. Gating was performed as described for Figure 1H. DOI: http://dx.doi.org/10.7554/eLife.00757.004

Figure 2.. Intrinsic requirement for EBI2 in CD4⁺ DCs.

(A) and (B) Wild-type CD45.2⁺ mice were lethally irradiated and reconstituted with mixed BM cells (1:1 ratio) from CD45.1⁺ Ebi2^+/+ mice and CD45.1⁺CD45.2⁺ Ebi2^−/− mice or their wild-type littermate controls. Eight weeks after reconstitution, splenocytes were analyzed by flow cytometry for CD4⁺ DCs, CD8⁺ DCs, DN DCs (gated as in Figure 1), and pDCs (defined as CD11c⁺siglecH⁺MHCII⁺). (A) Representative flow cytometric data. (B) Summary of CD45.2⁺ cell frequencies for data of the type in (A) (n = 5 mice). (C) and (D) Mice were reconstituted with BM transduced with a retroviral construct encoding EBI2 or truncated neural growth factor receptor (NGFR), with an IRES–truncated human CD4 cassette as a reporter. Flow cytometric analysis of splenic DCs of chimeras six weeks after reconstitution (n = 5–7 mice). Plots in (C) are shown gated on hCD4⁺ cells and right plots are further gated on MHCII⁺CD11c⁺ cells. Plots in (D) are summary of frequency of 33D1⁺ DCs among total CD11c⁺MHCII⁺ cells that were non-transduced (hCD4⁻) and transduced (hCD4⁺). ***p<0.001, Student’s T-test. DOI: http://dx.doi.org/10.7554/eLife.00757.005

Figure 3.. EBI2 is required for DC positioning in MZ bridging channels and LN interfollicular regions.

Immunohistochemistry of spleen (A and B) and lymph node (C) sections from the indicated mice were stained for IgD and 33D1 or DEC205, as indicated. Data are representative of at least six mice of each type. DOI: http://dx.doi.org/10.7554/eLife.00757.006

Figure 4.. Normal mature DC turnover and pre-DC function in EBI2-deficient mice.

(A) BrdU positive cells as percent of CD4⁺ and CD8⁺ DCs from Ebi2^+/− or Ebi2^−/− mice given BrdU in drinking water for 1, 2, or 4 days. Data are pooled from two to three experiments with five mice in each group. (B) Cyp7b1^+/− or Cyp7b1^−/− recipients were reconstituted with equal number of BM cells from Bim^−/− (CD45.2) and Bim^+/+ (CD45.1) mice and DCs enumerated by flow cytometry (n = 4). (C) Flow cytometric detection of GFP fluorescence in gated splenic pre-DCs or DCs from Ebi2^GFP/+ or WT mice (negative control). Pre-DCs are defined as Lin^- (CD19, B220, CD3, NK1.1, Ter119) CD11c⁺MHCII^-CD172a^intCD135⁺. (D) Number of pre-DCs from spleen and BM of Ebi2^+/− and Ebi2^−/− mice (n = 5 mice). (E) Sorted pre-DCs (1.0 × 10⁵) from CD45.1⁺ WT mice were adoptively transferred into CD45.2⁺ Cyp7b1^+/− or Cyp7b1^−/− recipients. (F) Sorted pre-DCs (1.0 × 10⁵) from CD45.2⁺ Cyp7b1^+/− or Cyp7b1^−/− mice were adoptively transferred into CD45.1⁺ WT recipients. 6 days after transfer, different subsets of CD45.1⁺(E) or CD45.2⁺ (F) pre-DC derived DCs were quantified by flow cytometric analysis, with the right plots in each pair being pre-gated as shown in each left plot (n=3 mice, one representative of two experiments). Graphs on right show a summary of data from three recipients of each type. (G) and (H) 1:1 Ratio mixed Ebi2^−/− (CD45.2⁺) and Ebi2^+/− (CD45.1⁺CD45.2⁺) pre-DCs were labeled with violet tracer and transferred into CD45.2⁺ WT recipients. 6 days after transfer, the appearance of DC subsets (G) and the frequency of each DC type that had divided three or more times (H) were analyzed by flow cytometry (n = 6). A description of the experimental scheme and an example of the flow cytometric data is included in the figure supplement. DOI: http://dx.doi.org/10.7554/eLife.00757.007

Figure 4—figure supplement 1.. PreDC transfer strategy, Notch and IRF4-induced gene expression and effects of Flt3L on DC frequencies.

(A) 1:1 Ratio mixed sorted Ebi2^−/− (CD45.2⁺) and Ebi2^+/− (CD45.1⁺CD45.2⁺) pre-DCs were labeled with violet tracer and transferred into CD45.2⁺ WT recipients. 6 days after transfer, the appearance of DC subsets and their extent of proliferation were analyzed by flow cytometry (n = 6 mice). Representative patterns were gated on CD11c⁺MHCII⁺CD45.2⁺Ebi2^GFP/+ cells. Numbers in plots indicate mean (±SE) frequency of cells in each gate. (B) Flow cytometric analysis of ESAM expression on gated CD4⁺ and CD8⁺ DCs from the indicated mice. Percentage of ESAM⁺ cells is shown as mean ± SE (n = 7 mice). (C) Quantitative PCR analysis of Dtx1 and Mmp12 transcripts in sorted CD4⁺ DCs, CD8⁺ DCs and DN DCs. Expression is shown relative to Hprt (n = 6 mice). (D) Ebi2^+/− and Ebi2^−/− mice were inoculated 1 × 10⁵ Flt3-ligand expressing or control B16 melanoma cells. 7 days later, splenocytes from the indicated mice were analyzed for DC subsets. One representative of four replicated mice is shown. DOI: http://dx.doi.org/10.7554/eLife.00757.008

Figure 5.. Rescue of EBI2-deficient CD4⁺ DCs by increased LTβR agonsim.

(A) Surface staining of LTβR in indicated DC subsets from LTβR^−/−, WT, and Ebi2^−/− mice (n=7 mice). Right panel shows the Ebi2^−/−/WT LTβR staining median fluorescence intensity (MFI) ratio. (B) and (C) 1:1 Mixed Ebi2^−/− (CD45.2) and Ebi2^+/+ chimeric mice were treated with LTβR agonist antibody or saline control. (B) Shows flow cytometric analysis for CD4 and CD8 on CD11c⁺ cells and (C) shows percentage of CD4⁺ DCs among total DCs from three to four mixed chimeras of each type. (D) Ebi2^+/− or Ebi2^−/− mice were treated with LTβR agonist antibody for 6 days. Immunohistochemistry staining for 33D1 and IgD in splenic sections from the indicated mice treated with LTβR agonist antibody. One representative picture of three replicated mice is shown. (E) Cyp7b1^+/− or Cyp7b1^−/− recipients were reconstituted with WT or lymphotoxin (LTβ10) transgenic (tg) BM cells and analyzed by flow cytometry for the indicated markers. Plots on right are pre-gated on MHCII⁺CD11c⁺ cells (n = 4 mice). *p<0.05, **p<0.01, ***p<0.001 by Student’s T-test. DOI: http://dx.doi.org/10.7554/eLife.00757.009

Figure 6.. MZ bridging channel positioning facilitates DC capture of blood-borne particulate antigen and rapid movement to T zone.

(A) Mice were i.v. immunized with SRBC-PKH26 or PBS control and analyzed 24 hr later by flow cytometry for PKH26 in gated DC subsets. Numbers indicate the frequency of cells within gate (mean ± SE, n = 4 mice). (B) Ebi2^+/+ or Ebi2^−/− mice were immunized with SRBC-PKH26 and 24 hr later the number of PKH26⁺ total DCs or CD4⁺ DCs were quantified as in (A) (n = 4 mice of each type). (C) Ebi2^+/− and Ebi2^−/− mice were immunized with SRBCs and spleens isolated at the indicated time points and stained by immunohistochemistry for 33D1 and IgD. (D) Mice were pre-immunized with SRBCs or PBS and 6 hr later, immunized with SRBC-PKH26. 1 day later, the percentage of PKH26⁺ DCs was quantified by flow cytometry (n = 3 mice). (E) CCR7 surface staining of CD4⁺ DCs from Ccr7^−/− mice or WT mice immunized with or without SRBC for 1.5 hr. One representative flow cytometric pattern of four mice is shown. (F) Migration of CD4⁺ DCs from PBS or SRBC immunized mice (1.5 hr) in response to medium only (nil), CCL19 (0.1 μg/ml), CCL21 (0.1 μg/ml), or 7α,25-OHC (0.5 nM). Bars show mean ± SE (n = 4 mice). *p<0.05, **p<0.01, ***p<0.001 by Student’s T-test. DOI: http://dx.doi.org/10.7554/eLife.00757.010

Figure 6—figure supplement 1.. Splenic distribution and DC capture of PKH26-labeled SRBC antigen and induction of DC repositioning by SRBC and LPS.

(A) and (B) Flow cytometric detection of PKH26 signal on gated CD4⁺ DCs in spleens taken from mice injected with SRBC-PKH26 the indicated number of hours before. Data are representative of at least three mice from two experiments. (C) Immunofluorescence picture of mice injected with SRBC-PKH26 and anti-CD11c for 0.5 hr (n = 3). (D) Flow cytometric analysis of PKH26 signal on DCs in splenocytes co-isolated from a mouse injected with SRBC-PKH26 1 day earlier and a control GFP transgenic mouse that did not receive PKH26-SRBC. Numbers in gates indicate percent cells in each. Data are representative of six mice from two experiments. (E) Immunohistochemistry staining of 33D1 and IgD on splenic sections of WT mice immunized with SRBC or LPS at the indicated time points prior to analysis. (F) Spleen sections from chimeric mice reconstituted with CCR7^+/+ or CCR7^−/− BM mixed with Ebi2^−/− BM (1:1 ratio), immunized 1 day earlier with SRBCs or saline (PBS), stained for CD11c and IgD. DOI: http://dx.doi.org/10.7554/eLife.00757.011

Figure 7.. DC deficiency in Ebi2^−/− mice is associated with a reduced ability to support CD4 T cell and B cell responses to particulate antigen.

(A) CD45.1⁺ OTII splenocytes were labeled with cell trace violet and adoptively transferred into Ebi2^+/− and Ebi2^−/− mice. 1 day after transfer, mice were immunized with SRBC-OVA conjugate. 3 days post immunization, OTII T cell proliferation and expression of ICOS and PD1 was examined by flow cytometry. Left panels are pre-gated on CD45.1⁺CD45.2⁻ cells and right three panels are further gated on TCR Vα2⁺ OTII cells. Numbers on gates indicate mean (±SE) frequencies for seven mice combined from two experiments. (B) and (C) Cell trace violet labeled Hy10 splenocytes were transferred into Ebi2^+/− or Ebi2^−/− recipients (B) or 1:1 mixed CD11cDTR and Ebi2^+/− or Ebi2^−/− BM chimeras (C). 1 day after cell transfer, mice were immunized with SRBC-HEL^2x and flow cytometric analysis of Hy10 B cell frequency and proliferation was conducted at day 3 after immunization. Hy10 B cells were identified as intracellular HEL-binding cells. For (C), mice were treated with DT at the day of Hy10 B cell transfer and again 2 days after cell transfer. Mean (±SE) cell frequency is shown next to each gate (n = 5–9 mice combined from two to three replicated experiments). (D) and (E) CD45.1⁺ Hy10 splenocytes were transferred into Ebi2^+/− or Ebi2^−/− recipients and mice were immunized with SRBC-HEL^2x. 5 days after immunization, HEL-binding Hy10 B cells and HEL-specific plasma cells of different Ig isotypes were quantified by flow cytometric analysis. Frequencies of GC B cells and total plasma cells are shown in plots (middle two panels) pre-gated on Hy10 B cells (gate shown in left panels). Frequencies of IgG1⁺ plasma cell are shown in plots (right panels) pre-gated on Hy10 plasma cells. Numbers on plots in D indicate mean (±SE) cell frequencies in the indicated gates and graph in (E) provides a summary of total HEL-binding cell numbers of the indicated types. Data are representative of three replicated experiments with four to six mice in each experiment. Each dot represents an individual mouse. *p<0.05, **p<0.01, ***p<0.001, by Student’s T-test. DOI: http://dx.doi.org/10.7554/eLife.00757.012

Figure 7—figure supplement 1.. EBI2 deficiency on DCs results in defective CD4+ T cell response to particulate antigen but not soluble and 33D1-coupled OVA.

(A) Cell trace violet labeled CD2-dsRed OTII splenocytes were transferred into 1:1 mixed CD11cDTR and Ebi2^+/− or Ebi2^−/− BM chimeras. 1 day after cell transfer, mice were immunized with SRBC-HEL^2x and flow cytometric analysis of OTII cell proliferation and ICOS expression was conducted at day 3 after immunization. Patterns are pre-gated on OTII cells identified as dsRed⁺TCRVα2⁺CD4⁺ cells. Recipients were treated with DT at the day of Hy10 B cell transfer and again 2 days after cell transfer. Mean (±SE) frequencies are shown next to each gate (n = 3 mice). (B) and (C) Ebi2^+/− and Ebi2^−/− mice (CD45.2) were transferred with cell trace violet labeled OTII splenocytes (CD45.1) and immunized with soluble OVA (A) or 33D1-OVA conjugate (B). 3 days post immunization, flow cytometric analysis was performed to determine the OTII cell frequency (defined as TCRVα2⁺CD45.2⁻) and proliferation in splenocytes from the indicated mice. Numbers next to gates indicate mean (±SE) frequency of cells in each (n > 6 mice). DOI: http://dx.doi.org/10.7554/eLife.00757.013

Figure 8.. Model for the role of EBI2 in mediating marginal zone (MZ) bridging channel positioning of CD4⁺ DCs in the spleen.

The major zones and cell types are labeled. EBI2 ligand is suggested to be produced in high amounts in the MZ bridging channel by stromal cells (not depicted) and acts to promote positioning of EBI2^hiCD4⁺ DCs (that are also 33D1⁺) in this region. This in turn juxtaposes them to MZ and follicular B cells, resulting in exposure to B cell-derived LTα1β2 (not depicted). LTβR engagement on the DC transmits a homeostasis-promoting signal. This positioning also exposes the cells to particulate antigens traveling with blood flow through the marginal sinus and the MZ. Following particulate antigen exposure, DCs migrate in a CCR7-dependent manner from the bridging channel in to the T zone where they present antigen to helper T cells. DOI: http://dx.doi.org/10.7554/eLife.00757.014

Author response image 1

(A, B) Mixed (1:1) CD11cDTR (CD45.1⁺) and Ebi2^-/- or
Ebi2^+/+ (CD45.1ʱCD45.2⁺) BM chimeric mice were immunized with SRBCs and flow cytometric analysis of B cell responses performed five days later. DT was administrated day -1, day 1, and day 3 after immunization. (A) Left panel shows a representative example of the CD45.1 and CD45.2 staining profile of total splenocytes. Right panels show representative flow cytometric pattern of gated CD45.1⁺CD45.2⁺ cells and CD45.1⁺ for plasma cells (defined as B220^lowCD138⁺) and germinal center cells (defined as 
GL7⁺Fas⁺). (B) Representative flow cytometric pattern of gated CD45.1⁺CD45.2⁺ Ebi2^+/+ (upper) or Ebi2^-/- (lower) cells. Mean (±SE) frequency of plasma cells and germinal center B cells are indicated next to each gate (n=3 mice). (C) Mixed Ebi2^+/+ (CD45.1+) and Ebi2^-/- (CD45.1⁺CD45.2⁺) BM chimeric mice were immunized with SRBCs and flow cytometric analysis was performed on B cell responses five days after immunization. Mean (±SE) frequency of plasma cells and germinal center B cells (GL7⁺B220⁺) are indicated next to each gate (n=4 mice, representative of 3 experiments).