IRF5 Controls Plasma Cell Generation and Antibody Production via Distinct Mechanisms Depending on the Antigenic Trigger

Abstract

Elevated levels of serum autoantibodies are a hallmark of systemic lupus erythematosus (SLE) and are produced by plasma cells in response to a variety of antigenic triggers. In SLE, the triggers are complex and may include both T cell-dependent/-independent and TLR-dependent/-independent mechanisms of immune activation, which ultimately contributes to the significant immune dysregulation seen in patients at the level of cytokine production and cellular activation (B cells, T cells, dendritic cells, neutrophils and macrophages). Interferon regulatory factor 5 (IRF5) has been identified as an autoimmune susceptibility gene and polymorphisms in IRF5 associate with altered expression and hyper-activation in distinct SLE immune cell subsets. To gain further insight into the mechanisms that drive IRF5-mediated SLE immune activation, we characterised wild-type (WT) and Irf5 ^-/- Balb/c mice in response to immunisation. WT and Irf5 ^-/- Balb/c mice were immunised to activate various signalling pathways in vivo followed by systemic immunophenotyping and detection of antibody production by multi-colour flow cytometry and ELISPOT. We identified two pathways, TLR9-dependent and T cell-dependent that resulted in IRF5 cell type-specific function. Immunisation with either CpG-B + Alum or NP-KLH + Alum but not with R848 + Alum, NP-LPS + Alum or NP-Ficoll+Alum resulted in decreased plasma cell generation and reduced antibody production in Irf5 ^-/- mice. Notably, the mechanism(s) leading to this downstream phenotype was distinct. In CpG-B + Alum immunised mice, we found reduced activation of plasmacytoid dendritic cells, resulting in reduced IFNα and IL6 production in Irf5 ^-/- mice. Conversely, mice immunised with NP-KLH + Alum had reduced numbers of T follicular helper cells and germinal centre B cells with reduced expression of Bcl6 in Irf5 ^-/- mice. Moreover, T follicular helper cells from Irf5 ^-/- mice were functionally defective. Even though the downstream phenotype of reduced antibody production in Irf5 ^-/- mice was conserved between T cell-dependent and TLR9-dependent immunisation, the mechanisms leading to this phenotype were antigen- and cell type-specific.

FIGURE 1 |

IRF5-mediated in vivo PC differentiation is dependent on the antigenic trigger. Mice were immunised with antigens intra-peritoneally, spleens collected on day 7 and cells stained with live/dead, CD45, IgD, B220, CD138 and IgG2a for flow cytometric analysis, PC (B220hi/lo CD138+ cells out of CD45+) and IgG2a + B cells (B220hi/lo IgG2a + cells out of CD45+) were analysed. (A) Graphical summary of flow data showing reduced PC and IgG2a + B cells in SRBC-injected Irf5^−/− mice (n = 6/genotype). (B) Representative flow plots showing reduced PC and IgG2a + B cells in NP-KLH + Alum-injected Irf5^−/− mice. (C–G) Graphical summary of flow data for PC and IgG2a + B cells from NP-KLH + Alum (C; n = 11/genotype), CPG-B + Alum (D; n = 7–12/genotype), NP-LPS + Alum (E; n = 8/genotype), NP-Ficoll+Alum (F; n = 8/genotype) and R848 + Alum (G; n = 9–11/ genotype). Alum is designated by A. (H) Representative ELISPOT images showing reduced IgM producing B cells from Irf5^−/− mice in response to CpG-B + Alum. (I, J) Graphical summary of dots counted from ELISPOT from CPG-B + Alum (I) and NP-KLH + Alum (J); n = 6/genotype. Data are presented as mean ± SD. p values are reported after unpaired parametric T-test was performed.

FIGURE 2 |

Irf5^−/− splenic B cells have defective in vitro PB differentiation and proliferation in response to C4. Sorted naïve B cells were cultured with C4 cocktail (CPG-B + IL-21 + anti IgM + anti CD40) to access PC differentiation and proliferation. (A–D) Representative flow plots and graphical summaries of B220hi/lo CD138+ cells out of CD45+ PC (A, B) and B220hi/lo IgG2a + cells out of CD45+ IgG2a+ B cells (C, D) from 5-day in vitro culture with C4 (n = 9/genotype). (E, F) Sorted naïve B cells were labelled with Violet cell trace dye and cultured with CPG-B, IL-21, anti-IgM, anti CD40 individually and with C4 cocktail for 5 days. The cells were then collected, labelled with L/D stain, CD45, CD19 and B220 and raw flow data analysed using ModFit Proliferation Wizard. Representative plots are shown (E) with graphical summary of proliferation indices of B cells from ModFit (F); n = 10/genotype. Data are presented as mean ± SD. p values are reported after unpaired parametric T-test (B, D). p values are reported after one way ANOVA analysis with Tukey’s multiple comparison tests (F).

FIGURE 3 |

Intrinsic defects in Irf5^−/−B and T cells is pathway-dependent. (A) Schematic representation of adoptive transfers of B cells from Irf5^−/− mice and WT mice to RAG2^−/− mice followed by CPG-Alum immunisation and detection of B220hi/lo CD138+ PC and B220hi/lo IgG2a positive cells. (B) Graphical summary of PCs and IgG2a + B cells detected in the spleens of RAG2^−/− mice at day 7 post CPG-Alum immunisation (n = 8/genotype). (C) Schematic representation of adoptive transfer of purified T and B cells to RAG2^−/− mice followed by NP-KLH immunisation and detection of PC and IgG2a positive cells. (D) Graphical summary of B220hi/lo CD138+ PCs and B220hi/lo IgG2a + B cells detected in the spleens of RAG2^−/− mice at day 7 post NP-KLH immunisation. n = 8/genotype. Data are presented as mean ± SD. p values are reported after unpaired parametric T-test was performed (B). p values are reported after one way ANOVA analysis with Tukey’s multiple comparison tests (D).

FIGURE 4 |

Defective pDC activation, IFNα and IL-6 production in Irf5^−/− mice in response to TLR9 immunisation. CPG-B + Alum was injected at day 0, spleens collected and stained for flow cytometric analysis on the indicated days post-immunisation; naïve (N), Day 1 (D1), Day 3 (D3). (A) Representative flow plot of CD45 + B220 + PDCA1+ pDC numbers on D3. (B–D) Graphical summary of the percentage of CD45 + B220 + PDCA1+ pDCs out of CD45+ from spleens of naïve mice (B; n = 9/genotype), at D1 (C; n = 13/genotype) and D3 (D; n = 16–18/genotype). (E) Representative flow plot of CD45+ B220 + PDCA1 + CD86+ pDC activation on D3. (F–H) Graphical summary of the percentage of CD45 + B220 + PDCA1 + CD86+ activated pDCs from spleens of naïve mice (F; n = 9/genotype), at D1 (G; n = 11/genotype) and D3 (H; n = 14/genotype). (I) Representative flow plot of IFNα production by CD45 + B220 + PDCA1+ pDCs on D3. (J–K) Graphical summary of the percentage of IFNα producing pDCs on D3 (J; n = 14/genotype), IFNα MFI from pDCs on D3 (K; n = 9/genotype). (L) Representative flow plot of IFNα production by activated CD45 + B220 + PDCA1 + CD86+ pDCs on D3. (M, N) Graphical summary of the percentage of activated (CD86) pDCs producing IFNα on D3 (M; n = 14/genotype), and MFI of IFNα from activated pDCs on D3 (N; n = 9/genotype). (O) Representative flow plot of IL-6 production by pDC on D3. (P, Q) Graphical summary of IL-6 production by pDC on D3 (P) and IL-6 MFI from pDCs on D3 (Q); n = 13/genotype. Data are presented as mean ± SD. p values are reported after unpaired parametric T-test.

FIGURE 5 |

Reduced Irf5^−/− Tfh cells in response to TD immunisation. NP-KLH + Alum was injected at day 0, spleens collected and stained for flow cytometric analysis on the indicated days post-i mmunisation; naïve (N), Day 1 (D1), Day 3 (D3). (A) Representative flow plot of CD45 + CD4 + PD1 + CXCR5+ TfH numbers out of CD4+ cells on D1. (B–D) Graphical summary of the percentage of CD45 + CD4 + PD1 + CXCR5+ TfH cells out of CD4+ in the spleen of naïve mice (B; n = 11–12/genotype), D1 (C; n = 9/genotype) and D3 (D; n = 8/genotype). (E) Representative flow plot of the percentage of CD45 + CD4 + PD1 + Bcl6+ TfH cells on D1. (F–H) Graphical summary of the percentage of CD45 + CD4 + PD1 + Bcl6+ TfH cells in the spleen of naïve mice (F), D1 (G) and D3 (H); n = 8–10/genotype. (I) Representative flow plot of activated CD45 + CD4 + PD1 + BCl6 + CD69+ TfH numbers on D3. (J, K) Graphical summary of activated CD45 + CD4 + PD1 + CXCR5 + CD69+ TfH cells on D3 (J) and activated CD45 + CD4 + PD1 + BCl6+ CD69+ TfH on D3 (K); n = 12/genotype. (L) Representative flow plot of GL-7+ GC B cells out of B220+ on day 5 (D5). (M–O) Graphical summary of the percentage of GL-7+ GC out of NP+B220+ B cells from spleens of naïve mice (M; n = 9/genotype), at D3, D5 and Day 7 (D7) post-immunisation (N; n = 8/genotype), percentage of Bcl6+ GL-7+ GC B cells out of B220 + GL7+ at D3, D5, D7 (O; n = 5–6/genotype). (P, Q) Graphical summary of Bcl6 MFI from GL-7 + B220+ GC B cells from spleens of naïve mice (P; n = 9/genotype), at D3, D5, D7 (Q; n = 5–9/ genotype). (R) Representative histogram from (Q) comparing MFI of Bcl6 of GL7 + B220+ cells between KO, WT and FMO on D5. Data are presented as mean ± SD. p values are reported after unpaired parametric T-test (B–D, F–H, J, K, M, P). p values are reported after one way ANOVA analysis with Tukey’s multiple comparison tests (N, O, Q).

FIGURE 6 |

Functional defect in Irf5^−/− Tfh cells. NP-KLH + Alum was injected at day 0, spleens collected on day 4 post-immunisation and CD4 + PD1 + CXCR5+ Tfh cells and CD19 + GL-7-B cells sorted. Cells were then co-cultured for 5 days and stained for GL-7. (A) Representative flow plot of GL-7+ GC cells out of B220+ generated in response to NP-KLH in culture for 5 days. (B) Graphical summary of percentage of B220 + GL-7+ GC B cells out of B220+ from (A); n = 7–8/genotype. (C–G) Graphical summary of MFI of GL-7 out of B220+ cells (C), MFI of Bcl6 out of B220+ cells (D), percentage of GL-7 + Bcl6+ GC B cells (E), percentage of Bcl6+ GC B cells (F); n = 7–9/genotype. (G) MFI of Bcl6 out of B220 + GL-7+ GC B cells from a single experiment (n = 4/genotype). Data are presented as mean ± SD. p values are reported after one way ANOVA analysis with Tukey’s multiple comparison tests.

FIGURE 7 |

IRF5 plays a critical role in TLR9 (CPG-B) and T cell-dependent (NP-KLH) antibody production. (A) In the presence of IRF5, PC-mediated antibody production downstream of TLR9 is dependent on pDC activation of IFNα and IL-6 production and does not require the GC response. Conversely, T cell-dependent PC antibody production requires the activation/function of Tfh cells and GC formation via IRF5-m ediated Bcl6 expression. (B) In the absence of IRF5, neither of these pathways are functional leading to defective antibody production.