Animals and disease models

Mice were bred and housed in a specific pathogen-free facility at the Medical University of Vienna and kept in a 12 hours light cycle at 21°C–23°C. Wildtype animals (C57BL/6J, RRID: IMSR_JAX:000664) for BMCs extraction and in vivo arthritis models were purchased from Charles River Laboratories or used from in-house breedings. K/BxN serum transfer arthritis was induced by intraperitoneal injection of 100 µL serum of K/BxN mice on day 0 and day 2 in C57BL/6J mice or Irg1^−/− and Irg1^+/+ mice (C57BL/6NJ-Acod1em1(IMPC)J/J, RRID:IMSR_JAX:029340) from The Jackson Laboratory. Animals were harvested 12 days after the first injection of serum.44 Irg1^−/− and Irg1^+/+ mice were kept in the same mouse facility and under the same conditions and identified by PCR using the following primers: 5′-GTGGGGAGGGGAACTATGAG-3′ (common), 5′-ATTTGGAGGAACCCCATGAC-3′ (WT) and 5′-CAGCCTCTAAGCCAGACAGC-3′ (Mutant). Human tumour necrosis factor (hTNFTg/+, RRID: MGI:3053718) transgenic animals were identified by PCR using hTNFTg primers (5′-TACCCCCTCCTTCAGACACC-3′ and 5′-GCCCTTCATAATATCCCCCA-3′) and were harvested at 8 weeks of age. hTNFTg/+ mice were a kind gift of George Kollias20 and crossed with Irg1^−/− to create hTNFTg/+/Irg1^−/− and hTNFTg/+/Irg1^+/+ mice. CD11c-DTR mice (B6.FVB-1700016L21Rik^{Tg(Itgax-HBEGF/EGFP)57Lan}/J, RRID:IMSR_JAX:004509) and B6 CD45.1 mice (B6.SJL-Ptprc^a Pepc^b/BoyJ, RRID:IMSR_JAX:002014) were bought from The Jackson Laboratory. Ovariectomy of 10-weeks-old female Irg1^−/− and Irg1^+/+ animals was performed as described previously.45 Briefly, ovaries were removed for ovariectomy while skin and peritoneum of sham-operated female animals was incised, and ovaries were left intact. After 4 weeks of recovery mice were sacrificed. Bone loss was evaluated in the tibiae using μCT image analysis. For itaconate injections, mice were intraperitoneally injected with 400 µL saline or pH-adjusted itaconate (Sigma, #I29204) solution (50 mg/mL) for 7 days on a daily basis and harvested at day 8. For RANKL injections, mice were intravenously injected with 50 µL PBS or RANKL (Abcam #ab129136) dissolved in PBS (0.4 µg/µL). For depletion of CD11c cells in CD11c-DTR mice during K/BxN serum transfer arthritis, 200 µL of diphtheria toxin dissolved in PBS (150 ng/mL) was intraperitoneally injected every 3 days starting on the day of arthritis induction. For depletion of CD11c cells in CD11c-DTR mice during homeostasis, mice were intraperitoneally injected once with 200 µL diphtheria toxin dissolved in PBS (150 ng/mL). All animal procedures were approved by the local ethics committee of the Medical University Vienna and were conducted in accordance with Austrian law.

Clinical assessment of murine arthritis

Clinical signs of arthritis were assessed by a routinely used semiquantitative double-blind score system.46 Swelling per paw was recorded: 0–3 (0=no swelling; 1=mild swelling of toes/ankle; 2=moderate swelling of toes/ankle; 3=severe swelling of toes/ankle). Grip strength per paw was assessed on a wire mesh: 0–3 (0=normal grip strength; 1=mildly reduced grip strength; 2=moderately reduced grip strength; 3=severely reduced grip strength). Total score was calculated by combining scores of swelling and grip strength of all four paws.

Histological analyses

Histological assessment was performed as previously described.7 Briefly, hind paws were fixed in 7.5% buffered formalin solution for 6 hours and then decalcified in 14% pH 7.2 EDTA/ammonium hydroxide buffer (Sigma, #ED-1kg, #318604) at 4°C for at least 7 days. Two micrometre decalcified paraffin-embedded sections were prepared and stained with H&E and TRAP (Sigma, #387A, Leukocyte Phosphatase Staining Kit). After deparaffinisation and rehydration, slides were stained 10 min with 1:5 diluted Meyer’s hemalum (Merck, #1.09249.0500), rinsed with distilled water, differentiated in 1% HCl ethanol and rinsed 10 min in tap water. Afterwards, slides were stained in eosin working solution (300 mL 1% Eosin, Sigma, #318906, 600 mL distilled water, 0.1 mL acetic acid 100%) for 15 s. Slides were then rinsed with distilled water, 96% ethanol, 100% ethanol, n-Butyl-acetate and mounted with Neo-Mount (Merck, #1.09016). For TRAP staining, slides were preincubated 1 hour at 37°C with the TRAP staining solution (250 µL naphthol AS-BI phosphoric acid, 1000 µL acetate solution, 500 µL tartrate solution, 45 mL distilled water; Sigma, #387A) protected from light. Afterwards, slides were developed for 2 min at 37°C with a mix of 250 µL of Fast garnet GBC base solution and 250 µL of sodium nitrite solution. Nuclei were stained with Meyer’s hemalum as described above (without HCl) and slides mounted in Aquatex (Merck, #108562). H&E and tartrate-resistant acid phosphatase stained sections were analysed using an Axioskop 2 microscope (Carl Zeiss MicroImaging) and Osteomeasure Analysis System (OsteoMetrics) to quantify the areas of joint inflammation, erosion, osteoclast numbers and osteoproliferative lesions according to the SMASH recommendations.47

RNA in situ hybridisation

For RNA in situ hybridisation, bones and paws were fixed in 7.5% buffered formalin solution at 4°C overnight and then decalcified in 14% pH 7.2 EDTA/ammonium hydroxide buffer (Sigma, #ED-1kg, #318604) at 4°C for at least 7 days. Two micrometre decalcified paraffin-embedded sections were deparaffinised by incubating the slides at 60°C for 1 hour and rinsed three times with Neo-clear solution (Sigma # 1098435000) for 7 min. Slides were moved into 100% EtOH and afterwards air-dried at room temperature. After drying, RNAscope Hydrogen Peroxide (#322300) was added on the slides and incubated for 10 min at room temperature. Slides were washed with ddH₂O for three times and then transferred into a staining dish filled with 95°C ddH₂O for 10 s. Then the slides were transferred into a staining dish filled with 1× RNAscope target retrieval solution (#322000) and incubated at 95°C for 15 min. After target retrieval, slides were washed with ddH₂O and dehydrated with 100% ethanol for 3 min. Slides were then air dried at room temperature. After creating a hydrophobic barrier two drops of RNAscope Protease Plus (#322300) were added and incubated for 30 min at 40°C in the HybEZ oven (#321720). Slides were then washed three times with ddH₂O. Afterwards slides were incubated with the RNAscope probes Mm-Irg1-Mus musculus immunoresponsive gene 1 (Irg1) (#450241) or Negative Control (dapB) (#310043) and incubated for 2 hours at 40°C in the HybEZ oven. Slides were then washed two times with 1× RNAscope washing buffer (#310091). Slides were then incubated with the RNAscope HD detection reagents AMP 1–6 (#322310) for 30 min at 40°C (AMP1 and 3) or 15 min at 40°C (AMP2 and 4) or 30 min at room temperature (AMP5) or 15 min at room temperature (AMP6) with two washing steps with 1× RNAscope washing buffer between each AMP reagent solution. After the last washing step slides were incubated with RNAscope DAB solution (equal volumes of DAB-A and DAB-B) and incubated for 10 min. Then slides were washed with ddH₂O and stained with Meyer’s hemalum (Merck, #1.09249.0500) as described above.

Analysis of bone parameters

Histomorphometry and bone formation analysis of mouse tibia was analysed as described previously.45 For µCT analysis the SCANCO Medical µCT 35 was used to produce images from trabecular bone and analysed with SCANCO evaluation software for segmentation, three-dimensional morphometric analysis, density and distance parameters.

Dynamic labelling of bone

Calcein labelling was performed and analysed as previously described.46 In short, mice were subcutaneously injected with calcein green (Sigma, #C0875-5G) (30 mg/kg) at day 7 and day 1 before harvest. Left tibial bones were embedded in methoxymethylmetacrylate. Measurements were performed on the entire marrow region within the cortical shell using OsteoMeasure.

Bone marrow transplantations

Bone marrow from Irg1 ^−/−, Irg1 ^+/+ and B6 CD45.1 animals was isolated by flushing out cells from the femur and tibia with sterile PBS. Lethally irradiated (9 Gy) 8–12-week-old Irg1 ^−/−, Irg1 ^+/+ and B6 CD45.1-recipient mice were injected retro-orbitally with 3×10⁶ cells under anaesthesia and were allowed to recover for 8 weeks. Chimerism was confirmed by flow cytometry of the blood. After recovery, arthritis was induced by intraperitoneal injection of 150 µL serum of K/BxN mice on day 0 and day 2.

Osteoclastogenesis and staining

Osteoclast assays were performed as previously described.7 In brief, bone marrow cells were isolated and non-adhering haematopoietic stem cells were cultured in MEMα (Gibco, #32561037) completed with 1% Pen-Strep (Gibco, #15140122) and 10% fetal calf serum (FCS, Gibco, #10082147) supplemented with 100 ng/mL M-CSF (R&D Systems, #416). After 3 days, cells were washed with PBS, scraped, replated and cultured in complete MEMα supplemented with 30 ng/mL M-CSF and/or 50 ng/mL RANKL (R&D Systems, #462) for another 3–4 days including one medium change on day 6. Osteoclasts were defined as TRAP positive (TRAP staining kit, Sigma, #387A) multinucleated cells (≥3 nuclei). For glutamine-free and glucose-free media, MEMα powder without nucleosides, glucose, pyruvate, vitamin C, amino acids (Genaxxon, #C4002.0010) was dissolved in ddH₂O and supplemented with all amino acids used in the formulation of Gibco MEMa except from glutamine. Dialysed FCS (Gibco, #26400036) was used for the assays with glutamine-free and glucose-free medium. For osteoclast assays showing the influence of α-KG addition to glutamine-free medium, dimethyl 2-oxolglutarate (Sigma, #349631) was used. For RANKL stimulation of HoxB8 precursor cells,17 HoxB8 cells were cultured in MEMα (Gibco, #32561037) completed with 1% Pen-Strep (Gibco, #15140122) and 10% FCS (Gibco, #10082147) supplemented with 100 ng/mL M-CSF (R&D Systems, #416). After 3 days, cells were washed with PBS, and medium was changed with complete MEMα supplemented with 30 ng/mL M-CSF and/or 50 ng/mL RANKL (R&D Systems, #462). For RANKL stimulation of sorted human CD14⁺ monocytes, classical monocytes (CD14⁺ CD16⁻) were sorted from peripheral blood mononuclear cells of healthy donors after density gradient centrifugation and cultured in MEMα (Gibco, #32561037) completed with 1% Pen-Strep (Gibco, #15140122) and 10% FCS (Gibco, #10082147) supplemented with human 50 ng/mL M-CSF (R&D systems, #216-MC). After 3 days, medium was changed with complete MEMα supplemented with 30 ng/mL human M-CSF and/or 50 ng/mL human RANKL (R&D Systems, #6449-TEC). Human tissue collection was conducted in accordance with the Medical University of Vienna Review Board ID 1075/2021.

Osteoblast cultures

Osteoblast cultures of murine cells were performed as previously described.45 In brief, osteogenic precursor cells were isolated from calvariae of neonatal mice. Calvariae were digested for 10 min per fraction and a total of five fractions in MEMα (Gibco, #32561037) containing 0.1% Collagenase (Sigma, #10103586001) and 0.2% Dispase II were collected (Sigma, #D4693). Osteogenic precursors were expanded in MEMα containing 10% FCS (Gibco) and 1% Pen-Strep (Gibco, #15140122). Differentiation of osteoblasts was achieved using mineralisation medium by adding 0.2 mM l-ascorbate (Sigma, #A5960) and 5 mM β-glycerophosphate (Sigma, #G9422). Medium was replaced every 2–3 days. Itaconate (Sigma, #I29204) was dissolved in 1M NaOH solution, pH-adjusted to 7.4 and added at a final concentration of 1 mM or 5 mM. After 26 days, cells were fixed with 7.5% buffered formalin solution, following staining with ddH₂O containing 8% Alizarin Red (Sigma, #A5533). Cell number was assessed by detaching cells with trypsin and counting them with a TC20 automated cell counter (Bio-Rad). For diethylsuccinate (Sigma #112402) treatment of osteoblasts, mineralisation medium was supplemented with 100 µM diethylsuccinate. For coculture experiments, osteogenic precursors were isolated and expanded as described above. Bone marrow cells from Irg1^+/+ and Irg1^−/− mice were isolated as described above and cultured in MEMα completed with 1% Pen-Strep and 10% FCS supplemented with 100 ng/mL M-CSF. After 3 days, cells were washed with PBS, scraped and seeded onto the osteogenic precursors. Differentiation of osteoblasts was achieved using mineralisation medium by adding 0.2 mM l-ascorbate and 5 mM β-glycerophosphate and osteoclastic differentiation was induced by addition of 30 ng/mL M-CSF and/or 50 ng/mL RANKL.

For the human osteoblast cultures, commercially available human bone marrow mononuclear cells (Lonza, #LON2M-125C) were expanded in standard growth medium (DMEM high glucose (HG) supplemented with 10% FCS, 1% L-glutamine, 1% Pen-Strep and 1 ng/mL basic fibroblast growth factor (Fisher Scientific #100-18B)). Aliquots were frozen in cryopreservation medium (standard growth medium containing 10% DMSO) in liquid nitrogen. After cultivation, cells were switched to osteogenic differentiation medium (DMEM HG containing 10% FCS, 1% L-glutamine, 1% Pen-Strep, 10 nM dexamethasone, 50 µM ascorbic acid-2-phosphate, 10 mM ß-glycerophosphate) (Sigma #A8960, #G9422, #D4902) supplemented with or without 1 mM or 5 mM itaconate. Medium change was performed twice per week. After 3 weeks, samples for ALP enzyme activity and mature osteoblast marker genes were harvested. After one additional week, samples for Alizarin Red S (Sigma #A5533) staining as well as for calcium deposition determination were harvested.

Real-time PCR

Total RNA was extracted from cells using TRIzol reagent (Invitrogen, #12034977) and Monarch RNA Clean-up kit (NEB, #T2030L). Reverse-transcription was performed using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, #4368814). Luna Universal qPCR Master Mix (NEB, #M3003E) was used for the quantitative PCR reaction. For data with murine cells to obtain sample-specific ΔCt values, normalisation to hypoxanthine phosphoribosyltransferase 1 (Hprt) (for data with murine BMCs, murine calvarial osteoblasts, HoxB8 cells, cross-organ Irg1 comparison), glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (for data with tibial bone and paws) and beta-actin (Actb) (for data with diethylsuccinate treatment) within each sample was performed. Data are shown as fold change, where 2−ΔΔCt values were calculated (ΔΔCt=ΔCt treatment−ΔCt control). For data with human cells, normalisation to GAPDH (for human bone marrow stromal cells) or to beta-actin (ACTB) (for human sorted CD14⁺ monocytes) within each sample was performed. Real-time PCR was performed using the following primers for murine genes: Hprt: 5′-CGCAGTCCCAGCGTCGTG-3′ and 5′-CCATCTCCTTCATGACATC TCGAG-3′; Actinb: 5'-TGTCCACCTTCCAGCAGATGT-3' and 5'-AGCTCAGTAACAGTCCGCCTAGA-3', Gapdh: 5'TCGTCCCGTAGACAAAATGG-3' and 5'TTGAGGTCAATGAAGGGGTC-'3; Nfatc1: 5′-GACAGACATCGGGAGGAAGA-3′ and 5′-AGCCTTCT CCACGAAAATGA-3′; Jdp2: 5′-CTCCTCCTGCTATGATGCCT-3′ and 5′-CTCTTGCCCAGTTTCACCTC-3′; Ocstamp: 5′-ATGAGGACCATCAGGGCAGCCACG-3′ and 5′-GGAGAAGCTGGGTCAGTAGTTCGT-3′; Dcstamp: 5′-TCCTCCATGAACAAACAGTTCCAA-3′ and 5′-AGACGTGGTTTAGGAATGCAGCTC-3′; Fos: 5′-AGCCCAGACCT GCAGTGGCT-3′ and 5′-GCGCTCTGCCTCCTGACACG-3′, Irg1: 5′-GCGAACGCTGCCACTCA-3′ and 5′-ATCCCAGGCTTGGAAGGTC-3′, Runx2: 5′-CGGAGCGGACGAGGCAAGAGTTTC-3′ and 5′-AGACAGCGGCGTGGTGGAGTGGAT-3′, Ocn: 5′-ACCCTGGCTGCGCTCTGTCTCT-3′, Col1a1: 5′-CTGACTGGAAGAGCGGAGAG-3′ and 5′-GCACAGACGGCTGAGTAGG-3′, Alp1: 5′-GCTGATCATTCCCACGTTTT-3′ and 5′-CTGGGCCTGGTAGTTGTTGT-3′, Opn: 5′-CTCCTTGCGCCACAGAATG-3′ and 5′-TGGGCAACAGGGATGACA-3′, Osx: 5′-AGCGACCACTTGAGCAAACAT-3′ and 5′-GCGGCTGATTGGCTTCTTCT-3′, Itgam: 5′-CATCAAGGGCAGCCAGATTG-3' and 5'-GAGGCAAGGGACACACTGAC-3', Emr1: 5'-CTTTGGCTATGGGCTTCCAGTC-3' and 5'-GCAAGGAGGACAGAGTTTATCGTG-3'. Real-time PCR was performed using the following primers for human genes: ACTB: 5'-ATTGCCGACAGGATGCAGAA-3' and 5'- GCTGATCCACATCTGCTGGAA-3'; IRG1: 5'-ATGCTGCTTTTGTGAACGGTG-3' and 5'-CTACCACGGAAGGGGGATGGA-3'. For human bone marrow stromal cells TaqMan universal PCR master mix and the following TaqMan Gene Expression Assays (Thermo Fisher Scientific) were used: GAPDH (Hs02786624_g1), ALP (Hs01029144_m1), OPN (Hs00959010_m1), BSP (Hs00173720_m1), COL1A1 (Hs00164004_m1) and RunX2 (Hs04940094_m1).

Western blot

Cells were washed with ice-cold PBS and lysed with RIPA lysis buffer system (Santa Cruz, #24948). The homogenate was centrifuged at 4°C for 15 min at 16 000 g and the supernatant containing the protein fraction recovered. Protein concentration in the supernatant was determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, #23225). A total of 10 µg of proteins were resolved by SDS-PAGE and transferred to PVDF membranes (GE Healthcare, #10600023). Membranes were blocked with TBS-Tween (0.1%) containing 5% skim milk powder (Sigma, #70166) and incubated with constant agitation with primary antibody solution at 4°C overnight. The following antibodies were used: IRG1 (Abcam, #ab222411, 1:1000) and beta-ACTIN (Cell Signaling #4970, 1:2000). Incubated membranes were washed three times for 5 min with TBS-Tween (0.1%) and probed with a goat antirabbit horseradish peroxidase-linked secondary antibody (Cell Signaling, #7074, 1:10,000). Antigen-specific binding of antibodies was detected with WesternBright Sirius HRP substrate (Advansta, #541020). Western blots were assessed by an area density analysis using the Vision Works Software (Analytik Jena).

RNA sequencing

For RNA-seq data of murine BMCs stimulated with RANKL, total RNA was extracted from cells using TRIzol reagent (Invitrogen, #12034977) and Monarch RNA Clean-up kit (NEB, #T2030L). NGS libraries were prepared using the QuantSeq 3′ mRNA-Seq FWD library preparation protocol (Lexogen GmbH, Vienna, Austria) and sequenced on a HiSeq 3000 instrument (Illumina, San Diego, California, USA) following a 50-base-pair, single-end recipe. HiSeq Control Software (HCS, HD V.3.4.0.38) and Real-Time Analysis Software (RTA, V.2.7.7) were used for raw data acquisition and base calling, respectively. Raw fastq files were quality and adapter trimmed using Trim galore! (V.0.4.4) followed by alignment to the mouse genome GRCm38 using STAR (V.2.5.248). R software (V.4.0.4) and DESeq2 package (V.1.30.1) was used to calculate differential gene expression and a |shrunkFC|>1.5 and a FDR<0.05 were defined as cut-offs.49 Functional enrichment analyses were performed using the Curated.WikiPathways database implemented in ShinyGO (V.0.7650). Only pathways containing at least five genes were considered.

For RNA seq data of human bone marrow stromal cells, RNA was extracted using TRIzol reagent and a standard protocol for acid-guanidium-phenol-chloroform extraction with RNA precipitation by isopropanol and washing with ethanol. RNA pellet was resuspended in nuclease-free water and stored at −80°C until further use. Libraries were prepared with the QuantSeq 3′ mRNA-Seq V2 Library Prep Kit with UDI (Lexogen, Austria) according to the manufacturer’s instructions. A 250 ng total RNA was used as input and final amplification of libraries was performed with 17 PCR cycles. Libraries were pooled in equimolar ratio and sequenced on Illumina NovaSeq SP Flowcell in SR100 mode. Overall quality of the next-generation sequencing data was evaluated automatically and manually with fastQC V.0.11.8 and multiQC V.1.7. Reads from all passing samples were adapter trimmed and quality filtered using bbduk from the bbmap package V.38.69 and filtered for a minimum length of 17nt and phred quality of 30. Alignment steps were performed with STAR V.2.7 using samtools V.1.9 for indexing, whereas reads were mapped against the genomic reference GRCh38.p12 provided by Ensembl. Assignment of features to the mapped reads was done with htseq-count v0.13. The human gene-level read counts were further analysed in R V.4.2.1. The read counts were normalised by total number of all mappable reads (library size) for each gene. The limma voom (V.3.52.4) results in a matrix of normalised gene expression values on log2 scale. The counts and normalised log2 limma voom expression values were used as a raw input for all the analysis. Outlier samples were checked by principal component analysis. The genes that showed expression below one count per million (cpm<1) in the group of replicates were excluded from downstream analysis. To calculate differential expression for human data we used |logFC|>0.5 and FDR<0.1. Functional enrichment analysis were performed using ShinyGO V.0.77 and GSEA using fgsea R package (v1.25.1).

Data availability and code availability

Raw RNA-seq data were deposited at the Gene Expression Omnibus (GEO) database (GSE208359, for mouse data) and (GSE236452, for human data). The data code is available from the corresponding authors on reasonable request.

U-¹³C glucose and U-¹³C glutamine isotope tracing

Cells were cultured for the indicated time points in the presence of fully labelled glucose (Cambridge Isotope Laboratories, #CLM-1396–1) and/or fully labelled glutamine (Cambridge Isotope Laboratories, CLM-1822-H-0.1). Metabolites were extracted using methanol:water (80:20, v/v) and extracts were centrifuged for 10 min at 5000 g. The supernatant was collected and dried using a nitrogen evaporator. The samples were reconstituted in 50 µL of methanol, centrifuged for 10 min at 1000 g and supernatant was used for LC-MS analysis. For AA tracing, a Vanquish UHPLC system (Thermo Scientific) coupled to an Orbitrap Fusion Lumos (Thermo Scientific) mass spectrometer was used for the LC–MS analysis. The chromatographic separation was carried out on an ACQUITY UPLC BEH Amide, 1.7 µm, 2.1×100 mm analytical column (Waters) equipped with a VanGuard: BEH Amide, 2.1×5 mm pre-column (Waters). The column was maintained at a temperature of 40°C and 2 µL sample were injected per run. The mobile phase A was 0.15% formic acid (v/v) in water with 10 mM ammonium formate and mobile phase B was 0.15% formic acid (v/v) in 85% ACN (v/v) with 10 mM ammonium formate. The flow rate was 0.4 mL/min and gradient elution was performed with a total analysis time of 17 min. The mass spectrometer was operated in a positive electrospray ionisation mode: spray voltage 3.5kV; sheath gas flow rate 60arb; auxiliary gas flow rate 20arb; capillary temperature 285°C. For the analysis a full MS scan mode with a scan range m/z 50–250, resolution 500 000; AGC target 2e5 and a maximum injection time 50 ms was applied. The data processing was performed with the TraceFinder 4.1 software (Thermo Scientific). For TCA cycle tracing, a Vanquish UHPLC system (Thermo Scientific) coupled to an Orbitrap Fusion Lumos (Thermo Scientific) mass spectrometer was used for the LC-MS analysis. The chromatographic separation was carried out on an ACQUITY HSS T3, 1.8 µm, 2.1×100 mm analytical column (Waters) equipped with a VanGuard HSS T3, 2.1×5 mm pre-column (Waters). The column was maintained at a temperature of 40°C and 2 µL of sample was injected per run. The mobile phase A was 0.1% formic acid (v/v) in water and mobile phase B was 0.1% formic acid (v/v) in methanol. The flow rate was 0.5 mL/min and gradient elution was performed with a total analysis time of 10 min. The mass spectrometer was operated both in positive and negative electrospray ionisation mode: spray voltage was 3.5kV for positive mode and 3.0kV for negative mode; sheath gas flow rate 60arb; auxillary gas flow rate 20arb; capillary temperature 285°C. For the analysis a full MS scan mode with a scan range m/z 80–400, resolution 500 000; AGC target 2e5 and a maximum injection time 50 ms was applied. The data processing was performed with the TraceFinder 4.1 software (Thermo Scientific).

Metabolomics

For determining TCA cycle intermediates from Irg1^+/+ and Irg1^−/− cells, metabolite extracts were generated by removing medium and washing once with cold PBS. After addition of −20°C quenching solution (methanol:acetonitrile:water, 2:2:1 v/v) plates were covered with dry ice for 15 min. After thawing on wet ice, cells were scraped and transferred to a microcentrifuge tube. After three cycles of freeze and thaw with liquid nitrogen, extracts were left on −20°C for 1 hour and spun down with 15 000 g for 15 min. Supernatant was transferred to a new tube and stored at −80°C. For determining itaconate levels from tibial bone or paws, tissue was snap frozen in liquid nitrogen and grinded with liquid nitrogen in a metal morter. Grinded tissue powder was transferred to microcentrifuge tubes and 500 µL of addition of −20°C quenching solution (methanol:acetonitrile:water, 2:2:1 v/v) was added. After three freeze and thaw cycles with liquid nitrogen, extracts were left on −20°C for 1 hour and spun down with 15 000 g for 15 min. For determining itaconate in serum samples, blood was collected in microcentrifuge tubes and stored at room temperature for 1 hour. Blood was then spun down with 5000 g for 10 min and serum was transferred to new microcentrifuge tubes. Proteins were removed by adding 400 µL of a methanol/ethanol mixture (4:1, v/v) to 100 µL of plasma in a microcentrifuge tube followed by vigorous vortex shaking for 5 min at room temperature and centrifugation at 4000 g for 10 min at 4°C. The supernatant was collected, transferred to another microcentrifuge tube, shock frozen with liquid nitrogen and stored at −80°C until analysis. Metabolite extracts were analysed by hydrophilic interaction chromatography (HILIC), directly coupled to tandem mass spectrometry (LC-MS/MS). In brief, 1 µL of the original sample was directly injected onto a polymeric iHILIC-(P) Classic HPLC column (HILICON, 100×2.1 mm; 5 µm) and the respective guard column, operated at a flow rate of 100 µL/min. The HPLC (Ultimate 3000 HPLC system; Dionex, Thermo Fisher Scientific) was directly coupled via electrospray ionisation to a TSQ Quantiva mass spectrometer (Thermo Fisher Scientific). A linear gradient (A: 95% acetonitrile 5%, 10 mM aqueous ammonium acetate; B: 5 mM aqueous ammonium bicarbonate, both supplemented with 0.1 µg/mL medronic acid) starting with 15% B and ramping up to 60% B in 9 min was used for separation. Detection and quantification have been done by LC-MS/MS, employing the selected reaction monitoring (SRM) mode of the instrument. Authentic standards were used for determining collision energies and optimal transitions of the SRM and for validating experimental retention times via standard addition. The following SRM transitions were used for quantitation in the negative ion mode: m/z 115 to m/z 71 (fumarate), m/z 117 to m/z 73 (succinate), m/z 129 to m/z 85 (itaconate), m/z 133 to m/z 115 (malate), m/z 145 to m/z 101 (α-KG), m/z 173 to m/z 85 (aconitate), m/z 191 to m/z 111 (citrate) and m/z 808 to m/z 408 (acetyl-CoA). Data interpretation was performed using TraceFinder (Thermo Fisher Scientific).

Real-time cell metabolic analysis

Oxygen consumption rate and extracellular acidification rate were measured on a Seahorse XF HS mini Analyzer (Agilent) using the Seahorse XF Real-Time ATP Rate Assay Kit (Agilent, #1 03 591-100) according to the manufacturer’s instructions. In brief, 5000 MC3T3-E1 osteogenic precursor cells (ATCC, #CRL-2593) were seeded per well in a XF HS mini plate using MEMα containing 10% FCS (Gibco) and 1% Pen-Strep (Gibco #15140122). After 3 days when cells reached full confluency, medium was replaced with mineralisation medium by adding 0.2 mM l-ascorbate (Sigma # A5960) and 2 mM β-glycerophosphate (Sigma #G9422) to complete MEMα with or without 1 mM or 5 mM itaconate. On the next day, medium was changed to XF DMEM medium (Agilent, #1 03 575–100) supplemented with or without itaconate (1 mM or 5 mM) and plates were kept in a non-CO₂ incubator for 45 min for equilibration before measurement. During the measurements, oligomycin (1.5 µM) and rotenone/antimycin A (500 nM) were subsequently injected. Raw data were analysed using Seahorse Analytics Online Software system, exported as a Microsoft Excel file and graphed in GraphPad Prism V.9 (GraphPad Software).

ALP enzyme activity assay

To determine alkaline phosphatase (ALP) activity, the samples were lysed in a solution containing 0.5% Triton X-100 in 0.5M 2-amino-2-methyl-1-propanol buffer with 2 mM MgCl₂ (with pH 10.3) for 1 hour at 4°C. The lysate was centrifuged for 5 min at 300 g and ALP activity of the supernatant was determined by incubation with 0.02M p-nitrophenyl phosphate substrate solution at 37°C. The reaction time until sufficient development of colour change was recorded and the reaction was stopped by adding 0.2M NaOH stop solution. Absorbance was measured at 405 nm, and the ALP activity was determined using a standard curve constructed with p-nitrophenol solutions of known concentrations.

Calcium concentration assay

For calcium content quantification, samples were extracted using 5% trichloroacetic acid at RT for 30 min. The samples were collected, centrifuged for 10 min at 4°C and the calcium content of the supernatant was determined using the calcium (CPC) LiquiColor test (Stanbio Laboratory, Boerne, USA) according to manufacturer’s instructions.

ELISA

ELISA-based measurements of CTX-1 and P1NP levels were performed according to the manufacture’s protocol using the Cross Linked C-Telopeptide Of Type I Collagen (CTXI) ELISA kit (Cloud-Cone Corp., #CEA665Mu) and the Mouse P1NP (Procollagen I N-Terminal Propeptide) ELISA Kit (Elabscience, # E-EL-M0233).

Statistics

Statistical analysis was performed by using a two-tailed t-test, an ordinary one-way ANOVA followed by Tukey’s, Šidàk’s or Dunnet’s multiple comparison tests if comparing selected pairs of means or a two-way ANOVA followed by Tukey’s or Šidàk’s multiple comparisons test with Prism V.9 software (GraphPad, La Jolla, California). All error bars indicate±SD