For flow cytometry assays, the sample sizes were of at least three individuals per experiment. For enzyme-linked immunosorbent assay (ELISA), sample sizes were of at least three per condition. Sample size was determined on the basis of experimental feasibility and statistical significance. The experiments were not randomized. The investigators were not blinded to the allocation during experiments and analyses.

B6.129S-IL1rntm1Dih/J (JAX stock #004754), B6.Cg-Tg(Cd4-cre)1Cwi/Bflu/J (JAX stock #022071), B6.129(Cg)-Il1r1tm1.1Rbl/J (JAX stock #028398), B6.129(Cg)-Foxp3tm4(YFP/icre)Ayr/J (JAX stock #016959) mice were provided by the Jackson Laboratory. C57BL/6N-Atm1BrdIl1r2tm1a(EUCOMM)Wtsi/WtsiPh (MGI:4842437) were provided by Institute of Molecular Genetics of the ASCR. Confirmed genotyping was performed in accordance with the supplier recommendations. All transgenic mice were in a C57BL/6J background. C57BL/6 Foxp3-GFP mice expressing GFP under the control of the promoter of Foxp3 gene were provided by B. Malissen of the Centre d’Immunologie de Marseille-Luminy (France). To generate CD4^creIL-1R1^lox mice, Il1r1tm1.1Rbl/J mice were mated with B6.Cg-Tg(Cd4-cre)1Cwi/Bflu/J mice. To generate FoxP3^creYFPIL-1R1^lox mice Il1r1tm1.1Rbl/J mice were mated with B6.129(Cg)-Foxp3^{tm4(YFP/icre)Ayr}/J mice. To generate Foxp3^GFPCD4^creIL-1R1^lox mice, allowing to distinguish Tfh and Tfr cells, CD4^creIL-1R1^lox mice were crossed with Foxp3^GFP mice. To generate Foxp3^GFPIL-1Ra^-/- mice, B6.129S-IL1rntm1Dih/J mice were crossed with Foxp3^GFP mice. To generate FoxP3^creYFPIL-1R2^lox mice, C57BL/6N-A^tm1BrdIl1r2^{tm1a(EUCOMM)Wtsi/Wtsi} mice were mated with B6.129(Cg)-Foxp3^{tm4(YFP/icre)Ayr}/J mice. All animals were maintained at the Sorbonne University (SU), Centre d’Expérimentation Fonctionnelle animal facility (Paris, France) under specific pathogen–free conditions in agreement with current European legislation on animal care, housing, and scientific experimentation (agreement number A751315). All procedures were approved by the local animal ethics committee.

Splenocytes from immunized mice were stained with Ter-119–biotin, CD8-biotin, CD11c-biotin and B220-biotin antibodies for 20 min at 4°C and labeled with anti-biotin magnetic beads (Miltenyi Biotec) for 15 min at 4°C. Biotinylated cells were depleted on an autoMACS separator (Miltenyi Biotec), following the manufacturer’s procedure. Enriched T cells were stained as described in the “Flow cytometry analysis of human cells” section, and the following subsets were sorted on BD FACSAria II (BD Biosciences), with a purity of >98%: CD4⁺CD8⁻CXCR5^hiPD1^hiFoxp3⁻ Tfh cells, CD4⁺CD8⁻CXCR5^hiPD^hiFoxp3⁺ Tfr cells. Inguinal lymph nodes were comminuted mechanically and filtered through a cell strainer (35 μm, StemCell). The diluent was stored at -80°C to perform IL-1β ELISA. The cell suspension was transferred into a 5 mL tube (Falcon) containing PBS, washed twice with PBS medium and stained.

One kidney/mouse was embedded in OCT Tissue-Tek compound (CellPath), snap-frozen, and stored at -80˚C until use. Kidneys were frozen with optimal cutting temperature (OCT) compound, cut into 7 μm sections and fixed with acetone. Sections will then be stained with FITC-conjugated anti-C3 and Alexa Fluor 647-conjugated anti-mouse IgG (Abcam, Cambridge, UK) antibodies. DNA was visualized using DAPI (Sigma-Aldrich). Quantification of fluorescence will be performed at ×20 magnification in one random microscopic field of each kidney, followed by binary analysis using Image J 1.49v software. Scores on a scale of 0–3 were estimated for both C3 and IgG deposition, based on the fluorescence intensity (0 = no deposits, 1 = low, 2 = moderate, 3 = high). An intensity threshold of 2 was considered positive for IgG or D3 deposits. Immune complex deposits positivity was assessed by a score above 2 for IgG and C3.

Fresh total cells from spleens were isolated in PBS1×–3% fetal bovine serum (FBS) and stained for 20 min at 4°C with the following monoclonal antibodies at predetermined optimal dilutions: CD121b-BV421, CD19-PeCF594, CD4-V500, CD8a-AF700, Bcl6-APC, CXCR5-Biotin, GL7- e450, CD95 PE, Foxp3-AF488, PD-1–PE (BD Biosciences) or PE-Texas Red (PETR), streptavidin-APC or streptavidin–APC-Cy7 (BD Biosciences), GITR PETR (Miltenyi). CXCR5 staining was performed using biotinylated anti-CXCR5 for 30 min at 20°C followed by APC- or APC-Cy7–labeled streptavidin at 4°C. Intracellular detection of Foxp3 was performed on fixed and permeabilized cells using appropriate buffer (eBioscience), following the manufacturer’s recommendations. Stained cells were run on CytoFLEX S cytometer (Beckman - Coulter) and analyzed using FlowJo software (TreeStar Inc.). Dead cells were excluded by forward/side scatter gating.

Two million cells per sample were stained. Cell viability was assessed by Viability UV T (Thermofisher). Cells were washed with PBS. After washing, anti-chemokine antibody (CXCR5 (BD Biosciences)) and brilliant stain buffer (used in complement of Brilliant fluorescent polymer dyes) were added in PBS 1 × at room temperature for 20 minutes. Other membrane markers were added next (CD3 (Biolegend), CD4 (Biolegend), CD8 (Biolegend) CD11b (BD Biosciences), PD1 (BD Biosciences), ICOS (BioLegend), Fas (BD Biosciences), GL7 (BioLegend), CTLA4 (BD Biosciences), GITR (BD Biosciences), CD25 (BD Biosciences), IgD (BD Biosciences), IgG (BD Biosciences), CD38 (BD Biosciences), B220 (Biolegend), CD19 (Biolegend)) and lymph node cells were incubated for 30 additional minutes at room temperature. After washing, the cells were washed and then resuspended in a residual volume of 100uL and acquired on an AURORA. Supervised analyses, conducted using FlowJo software, were used to study the phenotype and function of some specific T and B cell populations (Supplementary Figure S2).

For unsupervised analysis, batch-corrected expression values for all markers and cell clusters analyses were performed using Imubac R package (20, 21). The FlowSOM-guided clustering initially identified 40 clusters. Clusters were then manually analyzed and annotated. Five cell populations have been identified and annotated as follows: CD19⁺ (CD19⁺CD4^-CD8^-CD11b^-), CD4⁺ (CD19^-CD4⁺CD8^-CD11b^-), CD8⁺ (CD19^-CD4^-CD8⁺CD11b^-), and CD11b⁺ (CD19^-CD4^-CD8^-CD11b⁺). Additionally, 9 cell subpopulations have been identified and annotated as follows: CD8⁺, CD11b⁺, Tfr (CD4⁺CD8^-CD19^-CD11b^-Foxp3⁺CXCR5⁺PD1⁺), Treg (CD4⁺CD8^-CD19^-CD11b^-Foxp3⁺CXCR5^-PD1^-), Tfh (CD4⁺CD8^-CD19^-CD11b^-Foxp3^-CXCR5⁺PD1⁺), Tconv (CD4⁺CD8^-CD19^-CD11b^-Foxp3⁺CXCR5^-PD1^-), Plasmablast (CD19⁺CD38⁺B220^lowIgG⁺IgD^-), GCB (CD19⁺GL7⁺Fas⁺IgD^-), and Transitional B cell (CD19⁺CD38^-/lowB220^lowIgD^-GL7⁺Fas⁺). A cell count for each cell population and subpopulation was obtained for each sample. The proportion of each cell subpopulation was then obtained by computing the proportions of cell subpopulations among cell population.

Sorted cells were washed in PBS1× and stored in RNAqueous kit lysis buffer (Ambion Inc./Life Technologies) at -80°C. Total RNA was extracted according to the manufacturer’s instructions, and quality was assessed on a bioanalyzer using the Pico RNA Reagent Kit (Agilent Technologies). For quantitative PCR, we analyzed 25 ng of cDNA for expression of a range of genes using the GeneAmp 5700 Sequence Detection System (Applied Biosystems). The samples were normalized to the GADPH protein as control mRNA, by change in cycling threshold (ΔCT) method and calculated based on 2-ΔΔCT.

Tfr cells were sorted from D-8 OVA-immunized mice, following a previously described “Immunization” protocol. For all the conditions, 1 × 10⁴ Tfr cells were cultured for 12 hours in 96-well plates (Nunc) in complete RPMI 1640 (Thermo Scientific) and three CD3/CD28 beads for one T cell (Dynabeads Mouse T- Activator, Thermo Scientific). We then added or not recombinant mouse 1 μg of IL-1β (Milteniy Biotec) to Tfr cell cultures. IL-10 production by Tfr cells was measured by ELISA (eBioscience) in supernatants of cultured Tfr cells. Expression of activation marker (GITR) was analyzed by flow cytometry.

The concentrations of mouse IL-10 (eBiosciences), IL-1Ra (eBiosciences), IL-1β (eBiosciences) in the culture supernatant, lymph node lavage fluid and serum were detected using ELISA kits. The serum levels of anti-OVA, anti-DNA and anti-RNP antibodies were detected with in-house ELISA. Briefly, 5 ng/ml OVA proteins were coated on 96-well high-binding EIA/RIA plates overnight. After blocking, mouse serum that had been diluted at different concentration (1 : 4000, 1:12000, 1: 36000, 1:108000, 1:324000, 1:972000) was added to the plates and incubated for 2 h; then HRP- conjugated anti-mouse IgG (Southern Biotech) was added at a 1 : 2500 dilution followed by Streptavidin revelation. For ELISA quantifying anti-DNA and anti-RNP antibodies, 200 µg/ml DNA or 1,33 µg/ml RNP proteins were coated on 96-well high-binding EIA/RIA plates overnight. After blocking, mouse serum that had been diluted at different concentration (1:100, 1:300, 1:900, 1:2700) was added to the plates and incubated for 2 h; then HRP- conjugated anti-mouse IgG (Southern Biotech) was added at a 1: 2500 dilution followed by Streptavidin revelation. For total IgG levels in serum, Maxisorp (Nunc) plates were coated with anti-mouse Ig (ref: 0107-01, SouthernBiotech) on 96-well high-binding EIA/RIA plates overnight. After blocking, mouse serum that had been diluted at different concentration was added to the plates and incubated for 2 h; then HRP- conjugated anti-mouse IgG (ref: 1030-01, Southern Biotech) was added at a 1: 2500 dilution followed by Streptavidin revelation.

All statistical tests were performed by using GraphPad PRISM 6.0 (GraphPad Software, Inc, La Jolla, Calif). For all experiments throughout this article, a p-value threshold of 0.05 and statistical significance are displayed as ∗P less than or equal to.05; ∗∗P less than or equal to.01; ∗∗∗P less than or equal to.001; and ∗∗∗∗P less than or equal to.0001. Two groups were analyzed by non-parametric Mann-Whitney U test for unpaired data and paired Wilcoxon for paired data.

In CD4⁺ T cells, IL-1R1 is mostly expressed by Tfh cells and at a lower level on Tfr cells (15). We crossed CD4^cre mice with IL-1R1^lox mice and observed a proper deletion of IL-1R1 on CD4⁺ cells (Supplementary Figure S1A). At steady state, CD4^creIL-1R1^lox mice had normal numbers of Tfh, Tfr or Treg cells, and normal numbers of transitional B cells (CD19⁺CD38^-/lowB220^lowIgD^-GL7^-), GC B cells (GCBs) (CD19⁺GL7⁺Fas⁺IgD^-) or plasmablasts (CD19⁺CD38⁺B220^lowIgG⁺IgD^-). However, they had significantly reduced serum IgG Ab levels (Figure 1A). Four weeks after their immunization with ovalbumin (OVA), we observed a significant reduction of the proportion of Tfh but not of Tfr cells, leading to a significantly increased Tfr/Tfh ratio in comparison with control mice (Figure 1B). This was accompanied by a markedly decreased production of GCBs (CD19⁺GL7⁺Bcl6⁺IgD^-) (Figure 1B). These effects translated into a significant decrease of anti-OVA Ab production (Figure 1C), whereas the ratio of OVA-specific Ab to total Ab was not affected (Figure 1D). These results indicate an effect of IL-1R1 on the Tfh/Tfr control of humoral responses.

We confirmed the functionality of IL-1R1 on Tfr cells in vitro. The addition of IL-1β to purified Tfr from OVA-immunized mice significantly increased IL-10 production and GITR expression (Figures 1E, F), and such activation could no longer be detected with IL-1R1-deleted Tfr (Figures 1E, F). Collectively, these results suggest a functional role of IL-1R1 on Tfr cells, which calls for study of the impact of specific depletion of IL-1R1 on Tfr cells.

We thus crossed Foxp3^creYFP mice with IL-1R1^lox mice. We confirmed that Foxp3^creYFPIL-1R1^lox mice had a proper deletion of IL-1R1 expression on Tfr but not on Tfh cells (Supplementary Figure S1B). At steady state, compared to control mice, Foxp3^creYFPIL-1R1^lox mice presented (i) no significant modifications of Tfh, Tfr or Treg cells, nor the B cell subsets (Figure 2A) and (ii) a significantly lower level of circulating IgG Ab (Figure 2A). Four weeks after immunization, compared to controls, they had a significantly decreased Tfr/Tfh ratio (Figure 2B) and an increase of specific anti-OVA Ab (Figure 2C), with a significantly increased ratio of specific to total Ab (Figure 2D). It should be noted that the proportion of Treg in CD4^creIL-1R1^lox and Foxp3^creYFPIL-1R1^lox mice is not changed after immunization (Supplementary Figure S3), which supports the idea that IL-1R1 deletion mainly affects Tfr cells. Altogether, these results indicate that the expression of IL-1R1 on Tfr cells is involved in their own activation.

To investigate the role of IL-1R2 in Tfr cell activation, we crossed Foxp3^creYFP mice with IL-1R2^lox mice. At steady state, we observed a non-significant increase of follicular T cells including both Tfh and Tfr in Foxp3^creYFPIL-1R2^lox mice compared to control mice (Figure 3A). However, Foxp3^creYFPIL-1R2^lox mice showed a significant decrease in the proportion of transitional B cells, with a significant increase in the proportion of GCBs and plasmablasts (Figure 3A). This suggests that deletion of IL-1R2 on Tfr specifically impacts B cell maturation, even though circulating IgGs were not affected in naïve mice (Figure 3A). Four weeks after OVA-immunization, we observed a significant increase of the Tfr/Tfh ratio in FoxP3^creYFPIL-1R2^lox mice compared to control mice, whereas the GCB/B ratio remained similar (Figure 3B). Interestingly, FoxP3^creYFPIL-1R2^lox mice produced greatly reduced levels of OVA-specific IgG compared to control mice (Figure 3C).

The role of IL-1Ra on Tfr cells was investigated in IL-1Ra^-/- mice in the absence of a direct means of eliminating IL-1Ra only in Tfr cells (Figures 3D-F). At steady state, IL-1Ra deletion did not affect the proportion of Tfh and Tfr cells (Figure 3D). In addition, circulating IgG were not affected, whereas GCB cells were significantly decreased (Figure 3D). Four weeks after OVA-immunization, and similarly to Foxp3^creYFPIL-1R2^lox mice, we observed a significant decrease in OVA-specific IgG production, whereas the proportion of GCBs was not significantly altered (Figures 3E, F) Of note, the proportion of Treg in Foxp3^creYFPIL-1R2^lox and Il-1Ra^-/- mice is not altered after immunization (Supplementary Figure S3), suggesting that IL-1R2 and IL-1Ra deletion mainly affects Tfr cells. Overall, our data indicate that IL-1R2 and IL-1Ra on Tfr cells are critical for the production of antigen-specific antibodies.

To go further into the characterization of IL-1R2 and IL-1Ra in Tfr cell functions, we developed functional assays where Tfr cells were cultured in the presence of IL-1β. Compared with control Tfr cells, Tfr cells derived from Foxp3^creYFPIL-1R2^lox or Foxp3^GFPIL-1Ra^-/- mice showed increased expression of GITR without IL-1β stimulation (Figures 4A, B). Upon IL-1β stimulation, expression of GITR was even more enhanced in Tfr from Foxp3^creYFPIL-1R2^lox mice, whereas it remained stable in Tfr from Foxp3^GFPIL-1Ra^-/- mice. Of note, in the absence of IL-1β, IL-1Ra-deleted Tfr cells produced more IL-10 than control Tfr cells (Figure 4B), whereas production of IL-10 by Tfr from Foxp3^creYFPIL-1R2^lox mice was equivalent to controls (Figure 4A). IL-10 production was enhanced by the addition of IL-1β in control Tfr cells, whereas it did not affect IL-10 production by Tfr from Foxp3^creYFPIL-1R2^lox and Foxp3^GFPIL-1Ra^-/- mice. Noteworthily, IL-1R2 but not IL-1Ra deletion on Tfr cells enhanced their expression of IL-1R1, suggesting a feedback loop of regulation in between IL-1R2 and IL-1R1 (Figures 4A, B).

Finally, we assessed the dynamics of IL-1β concentrations in the draining lymph nodes of OVA-immunized mice. IL-1β was detected as early as day 0, but its concentration peaked between D6 and D10 (Figure 4C). This increase preceded the proliferation of Tfr, which appeared after D9 (Figure 4D). Based on these results, we propose that IL-1 inhibitory receptors are expressed by Tfr to prevent their early own activation by IL-1, which is present at low levels early after immunization.

We then evaluated whether this control of humoral immunity could also affect the autoimmune humoral response. We used pristane-induced lupus in which pristane injection induces mainly anti-RNP antibodies in C57BL/6 mice (22), but also anti-DNA antibodies and renal lesions with glomerular immune complex deposits at 6 months (22, 23).

Remarkably, while both Foxp3^creYFPIL-1R1^lox mice and control mice developed anti-RNP antibodies (Figure 5A), Foxp3^creYFPIL-1R2^lox and IL-1Ra^-/- mice did not (Figure 5A). In contrast, the level of anti-DNA antibodies produced by control and mutant mice was similar (Figure 5B).

Foxp3^creYFPIL-1R1^lox mice showed more glomerular C3 and immune complex depositions than control mice, while the proportion of glomeruli affected by IgG deposition was not significantly increased (Figure 5C). Foxp3^creYFPIL-1R2^lox mice showed similar rates of C3, IgG, or immune complex deposition as controls (Figure 5B), despite the absence of anti-RNP production. Similarly, IL-1Ra^-/- mice showed a non-significant decrease of immune complex deposition. Altogether, results indicate that deletion of IL-1R2 and IL-1Ra on Tfr abolishes the production of anti-RNP antibodies, whereas IL-R1 deletion on Tfr enhances renal damage.