Reagents

IFNγ (BioLegend), IL1, IL2, IL4, IL6, IL13, IL23, macrophage colony-stimulating factor (M-CSF), granulocyte-macrophage colony-stimulating factor (GM-CSF) and transforming growth factor beta (TGFβ) were all purchased from Peprotech, and ultra-pure lipopolysaccharide (LPS) from Escherichia coli O111:B4 was purchased from InvivoGen. RT-qPCR primers are detailed in Supplementary Tables S1 and S2, with antibodies used for flow cytometry shown in Supplementary Table S3.

Primary cell isolation and culture

All experiments were carried out in accordance with the United Kingdom Home Office Animals Scientific Procedures Act 1986 under relevant personal and project licenses.

Single cell suspensions were generated from bone marrow by flushing femurs and tibias with Roswell Park Memorial Institute 1640 medium (RPMI), and adding red blood cell lysis buffer (2 min, room temperature, Sigma). Cells were cultured (7 d) in complete medium [RPMI supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, 50 µM 2-mercaptoethanol] supplemented with M-CSF (100 ng/ml, Peprotech) to generate BMDMs), or with GM-CSF (20 ng/ml, Peprotech) to generate bone marrow-derived dendritic cells (BMDCs). BMDMs were replated on day 7, and stimulated for 18 h with IL4 (20 ng/ml, Peprotech) plus IL13 (20 ng/ml, Peprotech), or with LPS (100 ng/ml, InvivoGen) plus IFNγ (100 ng/ml, BioLegend).

Single cell suspensions were generated from joint synovium by digesting opened hindleg knee joints in liberase TL (Roche) and DNase I (Sigma) (30 min, 37 °C), and from lungs by digestion with collagenase IV (Sigma) and DNase I (Sigma) (40 min, 37 °C). Single cell suspensions were generated from spleen and lymph nodes by pressing tissues through a 70 μm cell strainer. All single cell suspensions, except those use for generating chimeras, were treated with red blood cell lysis buffer (2 min, room temperature) before further use.

Generation of Sulf2 +/− mice

Sulf2+/− mice were generated from the Sulf2tm1a(KOMP)Wtsi embryonic stem cell line (RRID: MMRRC_063144-UCD, obtained from the KOMP Repository at University of California, Davis), in which exon 4 of the Sulf2 gene is flanked by LoxP sites. Embryonic stem cells were injected into blastocysts and resulting chimeras bred to germline transmission of the Sulf2 tm1a allele, before crossing with Tg(Pgk1-cre)Lni mice [30], available from JAX, to facilitate recombination, and backcrossing to a C57BL/6 background to generate Sulf2 tm1b allele mice, with a reporter-tagged deletion allele, and missing critical exon 4 of the Sulf2 gene. The deletion of exon 4 is expected to generate a mis-sense mutation and to cause a premature stop codon in exon 5, which is anticipated to lead to nonsense-mediated decay of the mRNA and loss of protein expression.

RT-qPCR

RNA was reverse transcribed using a High Capacity Reverse Transcription cDNA kit (Applied Biosystems) after extraction from cultured cells (with RNeasy mini spin-columns, Qiagen), tissues (with TRIzol, Life Technologies), or joints [by homogenisation in a TissueLyser II using a Precellys hard tissue Lysing Kit (Bertin) followed by RNeasy mini spin-columns, Qiagen]. TaqMan Low-Density Array (TLDA) cards (Applied Biosystems) were custom designed using TaqMan primer/probes (Applied Biosystems, Supplementary Table S1), with data acquired on a ViiA Real-Time PCR System (Applied Biosystems). Manual RT-qPCR was carried out using TaqMan primer/probes (Applied Biosystems) with data acquired on a ViiA Real-Time PCR System (Applied Biosystems), or with KiCqStart SYBR Green Primers with data acquired and melt curves examined on a QuantStudio 3 (ThermoFisher Scientific, Supplementary Table S2). Fold change in expression was calculated using the ΔΔCt method.

Flow cytometry

Cells were stained with Zombie near-infra red (BioLegend) or aqua (ThermoFisher Scientific) fixable live/dead stains (10 min, room temperature), washed in PBS and blocked with anti-mouse CD16/CD32 (10 min, 4 °C, BD Biosciences). For extracellular staining, cells were then incubated with fluorochrome-conjugated antibodies (Supplementary Table S3) in PBS with 1% FBS (4 °C, 30 min), washed twice, and fixed in BD CellFIX (10 min, on ice, BD Biosciences). For intracellular staining, cells were fixed in Foxp3/Transcription factor staining buffer (4 °C, 30 min, Invitrogen), and incubated with fluorochrome-conjugated antibodies in Permeabilisation buffer (4 °C, 30 min, Invitrogen). To quantify intracellular cytokines, cells were stimulated (4 h) with phorbol 12-myristate 13-acetate (PMA, 20 ng/ml) plus ionomycin (1 µg/ml) in the presence of GolgiPlug/GolgiStop (BD Biosciences). Fluorescence was quantified using an LSR-II or Fortessa X20 flow cytometer (BD Biosciences).

Phagocytosis assays

BMDMs were incubated with 1 μm fluorescent microspheres [Fluoresbrite plain yellow-green particles, multiplicity of infection (MOI) 50, 30 min, Polysciences Inc], fluorescein-conjugated Escherichia coli K-12 BioParticles (MOI 5, 15 min, Invitrogen), or fluorescein-conjugated S. cerevisiae zymosan A Bioparticles (MOI 5, 20 min, Invitrogen), and phagocytosis quantified by flow cytometry. Efferocytosis was quantified by incubating BMDMs with Jurkat T cells (MOI 2, 45 min) which had been labelled with calcein-acetoxymethyl (2 h, 37 °C, ThermoFisher Scientific) and exposed to UV light (305 nm, 2.5 h) to induce apoptosis. As a negative control for all phagocytosis assays, particles were also incubated with BMDMs on ice.

Generation of bone marrow chimeras and antigen-induced arthritis (AIA)

Recipient 8-week old wild-type (WT) C57BL/6 mice were irradiated (two doses of 5.5 Gy at a 4 h interval in a Gulmay X-ray irradiator), and injected 2 h later in the tail vein with bone marrow cells (5 million/mouse) from 8-week old WT or Sulf2+/− donors. Seven weeks later, mice were subcutaneously injected with methylated bovine serum albumin (mBSA, 100 µg, Sigma) emulsified 1:1 with Complete Freund’s adjuvant (BD Biosciences). Arthritis was induced 3 weeks later by intra-articular tibiofemoral injection of mBSA (100 µg in PBS, right knee) or PBS control (left knees). Knee swelling was measured daily for 7 days after intra-articular injection using a calliper, and pain assessed by measuring weight bearing on mBSA-injected compared to PBS-injected knees using a dynamic weight-bearing 2.0 chamber (Bioseb). Mice were sacrificed after 7 days, and immune infiltration and total histological score, composed of the sub-synovium, synovium and bone marrow density scores, calculated from blinded scoring [31]. Immunohistochemical analysis of pSTAT1 (Y701) (D4A7, Cell Signaling, followed by goat anti-rabbit Alexa 488, ThermoFisher Scientific) and CD206 (C068C2, BioLegend, followed by goat anti-rat Alexa 568, ThermoFisher Scientific) abundance was done following antigen retrieval in Tris-EDTA. Fluorescent signal (mean fluorescence per cell cluster x area) was quantified for at least 700 cell clusters per genotype using a BioTek Cytation 7 imaging reader and Gen5 v 3.14 software. To analyse T cell subsets, single cell suspensions from knee joints and lymph nodes were stimulated for 4 h with PMA (20 ng/ml) and ionomycin (1 mg/ml) in the presence of protein transport inhibitors, before CD3+, CD4+ and CD8+ subsets were analysed by flow cytometry.

Bulk RNASeq

RNA was isolated (RNeasy Micro Kits, Qiagen) from single cell suspensions prepared from knee joints of WT and Sulf2-deficient bone marrow chimera mice sacrificed 7 days after induction of arthritis. Oxford Genomics Centre prepared bulk RNA libraries using an NEBNext Single Cell/Low Input RNA Library Prep Kit for Illumina (New England Biolabs) and sequenced the samples at a depth of 25 million reads per sample on a NovaSeq6000 (Illumina). PolyA transcripts were analysed with the Nextflow RNA-Seq pipeline [32], using STAR for alignment with the GRCm38 reference murine genome. Deconvolution of cell types was performed with MuSIC [33], utilising single cell data from Park et al. [34]. DeSeq2 was used for differential gene expression analysis, with correction for multiple comparisons carried out using the Benjamini and Hochberg method. Functional enrichment analysis was performed using gProfiler2 (v0.2.0) utilising the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database [35] for terms. The gSCS (Set Counts and Sizes) correction method was used to determine significantly enriched pathways with significance p < 0.05. Differentially regulated genes were manually compared with the Interferome database (interferome.org, [36]), and Homer was used to analyse promoter regions (8 bp motifs, -400 to -100 base pairs upstream of transcriptional start sites) of differentially expressed genes [37].

Analysis of type I interferon signaling

For analysis of STAT1 phosphorylation, BMDMs from WT and Sulf2+/− mice were stimulated with murine IFNβ (50 ng/ml, 30 min, Peprotech) and phospho-STAT1 (Y701) (D4A7, Cell Signaling) and STAT1 (#9172, Cell Signaling) quantified by immunoblotting analysis on an Odyssey CLx fluorescent imaging system (LI-COR Biosciences) using Image Studio software (ver 4.0, LI-COR Biosciences).

For analysis of interferon-regulated gene expression, BMDMs from WT and Sulf2+/− mice were stimulated with murine IFNβ (50 ng/ml, 4 h, Peprotech) and expression of Ccl5, Ccl7, Tlr3 and Il6 quantified by RT-qPCR relative to Gapdh.

Primeflow analysis of Sulf2 mRNA expression in blood, bone marrow, spleen, and lung

The PrimeFlow RNA assay (Invitrogen) was used according to the manufacturer’s protocol to assess Sulf2 mRNA expression in various cell types. Isolated cells were surface-stained, permeabilised, fixed, and hybridised with a target probe set (2 h, 40 °C). After hybridisation of pre-amplifying primers (1.5 h, 40 °C) and amplifying primers (1.5 h, 40 °C), samples were incubated with labelled probes (1 h, 40 °C), stained for immune cell markers (CD45, CD3, CD19, Ly6C, Ly6G in blood and bone marrow; CD45, CD3, CD19, F4/80, CD11b, Ly6C, Ly6G, CD11c, CD64, MHC-II in spleen and lung) and fluorescence quantified using a Fortessa X20 flow cytometer (BD Biosciences).

Toll-like receptor (TLR) signaling

Macrophages were stimulated with the TLR agonists IFNγ, LPS, IFNγ + LPS (all at 100 ng/ml), fibroblast-stimulating lipopeptide-1 (FSL-1, 100 ng/ml), 10 ng/ml polyinosinic: polycytidylic acid [poly (I: C)], or 100 ng/ml N-palmitoyl-S-[2,3-bis(palmitoyloxy)-(2RS)-propyl]-[R]-cysteinyl-[S]-seryl-[S]-lysyl-[S]-lysyl-[S]-lysyl-[S]-lysine (Pam3CSK4) for 6 h and IL6 secretion measured by ELISA (R&D Systems).

Inflammasome activation

Macrophages were stimulated with 100 ng/ml LPS (100 ng/ml, 6 h) with nigericin (5 µM) added for the last 2 h of culture. Cell death was assessed using the CytoTox 96 Cytotoxicity assay (Promega) and IL1β secretion measured by ELISA (Invitrogen).

Antigen uptake and presentation

Antigen uptake was measured by flow cytometry analysis of BMDMs incubated with Alexa647-labelled ovalbumin (OVA, 0.1 mg/ml, 1 h, 37 °C, Molecular Probes). To assess antigen presentation, CD4+ T-cells were enriched from the spleen of OT-II mice using the MojoSort mouse CD4+ T-cell isolation kit (BioLegend) and labelled with CellTrace Violet (5 µM, 20 min, 37 °C, Invitrogen). Labelled CD4+ T-cells were co-cultured (4 d) with WT or Sulf2+/− antigen-presenting cells (BMDMs or BMDCs) activated with LPS (100 ng/ml LPS, 4 h, InvivoGen). As a source of antigen, OVA (0.1 mg/ml, InvivoGen) was added to antigen-presenting cells (1 h, 37 °C) before LPS-activation, or OVA peptide 323–339 (10 µg/ml, Peptides International) was added during co-culture of antigen-presenting cells with CD4+ T-cells. CD4+ T-cell proliferation and expression of CD25 were analysed by flow cytometry, with division index, proliferation index and percent divided cells determined using the FloJo proliferation tool.

Cytotoxicity assay

A CytoTox 96 Cytotoxicity assay (Promega) was performed according to manufacturer’s instructions. This assay detects release of lactate dehydrogenase from damaged cells.

Statistical analysis

Generation of heatmaps and principal component analysis (PCA) analysis was performed in R (version 3.5.2) and R studio (version 1.1.453). For PCA analysis, the prcomp function was used, without any maximum number of ranks and without clustering. The samples were then plotted on the obtained and rotated principal components, together with ellipsoids indicating 95% confidence around the centroids of the data groups. All other data were analysed using GraphPad Prism (version 8.4.1). Data were analysed for normality using the D’Agostino-Pearson test, and for statistical significance using relevant tests as detailed in the figure legends. Data are presented as mean ± SEM or median ± IQR where appropriate.