BAFF Induces Tertiary Lymphoid Structures and Positions T Cells within the Glomeruli during Lupus Nephritis

Abstract

Tissue-specific immune responses play an important role in the pathology of autoimmune diseases. In systemic lupus erythematosus, deposits of IgG-immune complexes and the activation of complement in the kidney have long been thought to promote inflammation and lupus nephritis. However, the events that localize cells in non-lymphoid tertiary organs and sustain tissue-specific immune responses remain undefined. In this manuscript, we show that BAFF promotes events leading to lupus nephritis. Using an inducible model of systemic lupus erythematosus, we found that passive transfer of antinucleosome IgG into AID^-/-MRL/lpr mice elevated autoantibody levels and promoted lupus nephritis by inducing BAFF production in the kidneys, and the formation of renal tertiary lymphoid structures (TLSs). Reducing BAFF in vivo prevented the formation of TLSs and lupus nephritis; however, it did not reduce immune cell infiltrates, or the deposits of IgG and complement in the kidney. Mechanistically, lowering BAFF levels also diminished the number of T cells positioned inside the glomeruli and reduced inflammation. Thus, BAFF plays a previously unappreciated role in lupus nephritis by inducing renal TLSs and regulating the position of T cells within the glomeruli.

Figure 1. Treatment of AID^−/−MRL/lpr mice with anti-nucleosome IgG induces the infiltration of hematopoietic cells into the kidneys and renal TLSs

(A) AID^−/−MRL/lpr mice were treated with PBS, 500 μg of PL2-3 (anti-nucleosome) or control F(ab′)₂ of PL2-3 i.v. weekly for 2 or 5 weeks. (B) B cells (CD19⁺), T cells (CD3⁺), DCs (CD19^negCD3^negCD11b⁺CD11c^hi), and CD11b+ myeloid cells (CD19^negCD3^negCD11b^hiCD11c^neg) were enumerated by flow cytometry from total kidney cells. The cell numbers presented are from one kidney from each mouse (n= total of 10 to >25 mice per treatment from >5 separate experiments). (C) Kidneys from PL2-3 treated AID^−/−MRL/lpr mice for 5 weeks were stained for CD19⁺ B cells (Green), CD4⁺ and CD8⁺ T cells (Red), CD11b⁺ myeloid cells (Blue) (left panel); or CD19⁺ B cells (Green), CD3⁺ T cells (Red), CD21⁺ FDCs (Blue) (right panel). Images were taken at 12x magnification. Scale bar = 50 μm. Representative images from 3 experiments (n = 9 mice). In (B) results are ± SEM. *p<0.05, **p<0.01, ***p<0.001 by Kruskal-Wallis test with Dunn’s multiple comparison post-hoc procedure.

Figure 2. Treatment of AID^−/−MRL/lpr mice with anti-nucleosome IgG promotes heightened BAFF production

(A) Serum BAFF levels were analyzed by ELISA after 5 weeks of treatment. (B) BAFF secreting kidney cells were enumerated by ELISpot after 2 or 5 weeks of treatment. Fold changes over PBS treated AID^−/−MRL/lpr mice were graphed. Mean numbers of BAFF secreting cells per 10⁶ cells are presented below the graph (Cell number ranges: B6 87–661, MRL/lpr 267–1825, AID^−/−MRL/lpr+PBS 181–622, AID^−/−MRL/lpr+F(ab′)₂ 115–276, AID^−/−MRL/lpr+Hy1.2 105–574, AID^−/−MRL/lpr+PL2-3 2wk 225–310, AID^−/−MRL/lpr+PL2-3 5wk 375–3028). Results are ± SEM from >5 experiments, n=2–5 mice per treatment per experiment. *p<0.05, **p<0.01, ***p<0.001 by Kruskal-Wallis test with Dunn’s multiple comparison post-hoc procedure.

Figure 3. Reducing BAFF levels prevents autoimmunity

(A) AID^−/−MRL/lpr mice were treated with PL2-3 (500 μg/mouse; black arrows) weekly in combination with 3.3 mg/kg of BR3-Fc or control antibody (IgG1) given every other week (gray arrows), for 5 weeks. Separate cohorts of mice were treated with PBS. (B) Serum anti-nucleosome IgM, (C) anti-dsDNA IgM, and (D) total IgM levels were measured by ELISA. (E) Splenic B cells were enumerated by flow cytometry after 5 weeks of treatment. In (B to E), results are ± SEM from 4 experiments, n=2–5 mice per treatment per experiment. *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant by Kruskal-Wallis test with Dunn’s multiple comparison post-hoc procedure.

Figure 4. Reducing BAFF levels prevents lupus nephritis

(A) H&E stained kidney sections from AID^−/−MRL/lpr mice treated for 5 weeks. Sections from B6 and MRL/lpr mice were shown as controls. Arrow indicates fibrocellular crescents formation. Scale bar = 25 μm. (B) Proteinuria, (C) glomerular (left panel) and tubulointerstitial (right panel) scores following 5 weeks of treatments. Each circle depicts an individual animal. (D) IgG and C3 deposits in the kidneys from AID^−/−MRL/lpr mice following 5 weeks of treatments, B6, and MRL/lpr mice were assessed by immunofluorescence (upper panels). Scale bar = 140 μm. The intensity of IgG (lower left) and C3 (lower right) deposits on all the glomeruli from 2–5 sections from each mouse were measured and shown as normalized values over area. Each dot represents each glomerulus. Bars represent median and each circle represents each glomerulus. In (A and D) representative images are from >3 experiments. n=3–7 mice per treatment. In (B and C) results are from 3 experiments, n=2–5 mice per treatment per experiment, bars represent median, each circle represents each mouse. *p<0.05, **p<0.01, ***p<0.00, n.s. = not significant by Kruskal-Wallis test with Dunn’s multiple comparison post-hoc procedure.

Figure 5. Lowering BAFF levels prevents renal TLSs and alters the positioning of T cells

(A) TLSs from AID^−/−MRL/lpr mice following 5 weeks of treatments were stained and imaged at 12x magnification. Scale bar = 50 μm. CD19⁺ B cells (Green), CD4⁺ and CD8⁺ T cells (Red), CD11b⁺ myeloid cells (Blue). (B) Enumeration of B cells, T cells, DCs, and CD11b+ myeloid cells that infiltrated one kidney of each mouse after 5 weeks of treatment. 4 experiments, n=2–5 mice per treatment per experiment. (C) Localization of B cells (Green), T cells (Red), and CD11b⁺ myeloid cells (Blue) in kidneys from AID^−/−MRL/lpr mice following 5 weeks of treatments or from age-matched B6, MRL/lpr mice. Gray is for nuclear stain. Images were taken at 30x magnification. Scale bar = 50 μm. The images of glomeruli were randomly captured on multiple sections of any kidneys. (D) Frequency of glomeruli containing T cells, B cells, or CD11b⁺ myeloid cells over total number of glomeruli counted per mouse were graphed (total of 10–40 glomeruli per mouse were randomly counted from 2–4 mice per treatment from 2 experiments). Sectioning and staining between experiments were done on different days. Each circle depicts an individual animal. In (A and C) data are from 2–3 experiments with 2–4 mice per treatment. In (B) results are ± SEM, in (D) bars represent median. *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant by Kruskal-Wallis test with Dunn’s multiple comparison post-hoc procedure.