Abstract Itaconate is one of the most highly upregulated metabolites in inflammatory macrophages and has been shown to have immunomodulatory properties. Here, we show that itaconate promotes type I interferon production through inhibition of succinate dehydrogenase (SDH). Using pharmacological and genetic approaches, we show that SDH inhibition by endogenous or exogenous itaconate leads to double-stranded mitochondrial RNA (mtRNA) release, which is dependent on the mitochondrial pore formed by VDAC1. In addition, the double-stranded RNA sensors MDA5 and RIG-I are required for IFNβ production in response to SDH inhibition by itaconate. Collectively, our data indicate that inhibition of SDH by itaconate links TCA cycle modulation to type I interferon production through mtRNA release. __________________________________________________________________ Itaconate is one of the most upregulated metabolites in inflammatory macrophages^[54]1. Itaconate is synthesized through the decarboxylation of the tricarboxylic acid (TCA) cycle intermediate cis-aconitate by the enzyme aconitate decarboxylase 1 (ACOD1). ACOD1 is encoded by immune-responsive gene 1 (Irg1), expression of which is induced by several inflammatory stimuli^[55]2. Although many studies have identified mechanisms by which itaconate regulates macrophage function^[56]3, how it exerts its immunomodulatory effects remains to be fully elucidated. Two such mechanisms of action of itaconate in cells include the inhibition of mitochondrial complex II, also known as succinate dehydrogenase (SDH)^[57]1,[58]4–[59]7, and the post-translational modification of cysteine thiols through a process termed 2,3-dicarboxypropylation. Itaconate has been found to regulate processes such as Nrf2 activation^[60]1, inflammasome activation^[61]8, aerobic glycolysis^[62]9,[63]10, alternative macrophage activation^[64]11 and lysosomal biogenesis^[65]12. Because it had been unclear whether macrophages could take up extracellular itaconate, esterified derivatives with increased membrane permeability were synthesized to elucidate the intracellular mechanisms of action of itaconate. These derivatives include dimethyl-itaconate (DI) and 4-octyl-itaconate (4-OI), which are more electrophilic than underivatized itaconate. It has since been demonstrated that itaconate can exit macrophages through the ATP-binding cassette transporter G2 (ABCG2)^[66]13 and can subsequently be taken up by macrophages^[67]5. Itaconate competitively inhibits succinate dehydrogenase (SDH), preventing oxidation of succinate to fumarate^[68]4. In previous work, inhibition of SDH by itaconate has led to succinate accumulation and, importantly, the inhibition of Il1b transcription^[69]5,[70]8. Mechanistically, inhibiting SDH activity prevents reverse electron transport (RET) and reduces production of reactive oxygen species (ROS)^[71]14, thereby decreasing hypoxia-inducible factor 1-α (HIF1α)-mediated Il1b transcription^[72]15. However, the full consequences of SDH inhibition by itaconate are yet to be explored. Here, we provide evidence that SDH inhibition by itaconate is required for type I interferon (type I IFN) production in macrophages in response to lipopolysaccharide (LPS), a product of Gram-negative bacteria. We show that this is due to mtRNA release through VDAC1 pores on the mitochondrial membrane, which signals through RIG-I and MDA5, to upregulate type I IFN production. Our work establishes a link from the TCA cycle to IFNs through SDH inhibition by itaconate. Results Itaconate boosts LPS-driven IFNβ production We first aimed to uncover the main pathways regulated by itaconate in inflammatory macrophages by treating LPS-stimulated murine bone-marrow-derived macrophages (BMDMs) with underivatized itaconate, followed by RNA sequencing (RNA-seq). Using the recently developed bioinformatic tool Immune Response Enrichment Analysis (IREA)^[73]16, which has been refined for studying immune processes, we employed RNA sequencing to determine which cytokine signaling pathways are regulated by itaconate in BMDMs. We found that IFNα and IFNβ were among the most actively signaling cytokines in itaconate-treated macrophages ([74]Fig. 1a). We also observed widespread transcriptomic effects, demonstrating that itaconate has multifaceted roles as an immunometabolite ([75]Extended Data Fig. 1a,[76]b). Consistent with previous literature^[77]1, we found that itaconate suppressed transcription of inflammatory genes ([78]Extended Data Fig. 1c); we also found an upregulation of Nrf2-dependent genes ([79]Extended Data Fig. 1d), illustrating the immunoregulatory properties of itaconate. Fig. 1 |. Itaconate boosts LPS-mediated IFNβ expression in macrophages. Fig. 1 | [80]Open in a new tab a, IREA plot of RNA-seq data of significantly (adjusted P (P[adj]) < 0.05) differentially expressed genes in BMDMs pretreated with PBS or itaconate (10 mM, 3 h) (n = 3 mice; LPS 4 h). Data are from three independent experiments. b, IFNβ release from BMDMs pretreated with PBS itaconate (n = 8; LPS 16 h). Data are from three independent experiments. c, Ifnb1 expression, determined through RNA-seq, in BMDMs pretreated with PBS or itaconate (n = 3 mice, LPS 4 h). Data are from three independent experiments. d, IFNB expression, assessed by quantitative PCR (qPCR), in human monocyte-derived macrophages (MDMs) pretreated with PBS or itaconate (ITA) (10 mM, 3 h) (n = 4 donors; LPS 4 h). Data are from two independent experiments e, IFNβ release from Irg1^+/+ versus Irg1^−/− BMDMs (n = 6 mice; LPS 4 h). Data are from three independent experiments. f, IFNβ release (P = 0.0023) from human PBMCs with IRG1 knockdown (n = 5 donors; LPS/IFNγ 24 h). NC siRNA, negative control siRNA. Data are from three independent experiments. In the graphs, data are shown as mean ± s.e.m. P values were calculated using one- or two-way analysis of variance (ANOVA). Additionally, itaconate treatment boosted LPS-induced IFNβ release ([81]Fig. 1b) and transcription ([82]Fig. 1c), consistent with previous observations by Swain et al.^[83]5. This effect was conserved in human macrophages, which also produced higher levels of IFNB in response to itaconate ([84]Fig. 1d). The effect of itaconate contrasts with the more electrophilic compound 4-octyl itaconate (4-OI) which, as we have shown previously, inhibits the production of IFNβ^[85]1,[86]5 ([87]Extended Data Fig. 2a,[88]b) in a similar manner to the Nrf2 activators diethyl maleate (DEM) ([89]Extended Data Fig. 2c) and dimethyl fumarate (DMF) ([90]Extended Data Fig. 2d). Any effect of 4-OI, or 4-OI-derived itaconate, on SDH is likely to be overcome by 4-OI being more reactive to cysteines, for example those in JAK1; 4-OI and itaconate might also block IFNs through activation of Nrf2 (ref. [91]11). Furthermore, Irg1^−/− BMDMs, which lack itaconate, produced less IFNβ than did Irg1^+/+ BMDMs in response to LPS ([92]Fig. 1e and [93]Extended Data Fig. 1e). Genetic ablation of IRG1 in human peripheral blood mononuclear cells (PBMCs) also caused a reduction in IFNβ production in response to a combination of LPS and IFNγ ([94]Fig. 1f and [95]Extended Data Fig. 1f). SDH activity regulates production of type I interferon Because itaconate is a well-characterized SDH inhibitor, we hypothesized that it could underlie its regulation of type I IFN through SDH inhibition; we have previously observed that inhibition of fumarate hydratase (FH) had this effect^[96]17,[97]18. We first examined SDH activity in inflammatory macrophages. We found that LPS suppressed SDH activity in Irg1^+/+ BMDMs, but not in Irg1^−/−BMDMs ([98]Fig. 2a), confirming the inhibition of SDH by endogenous itaconate. Furthermore, succinate failed to accumulate in Irg1^−/− BMDMs, and levels of metabolites directly downstream of SDH were increased ([99]Fig. 2b). Fig. 2 |. SDH inhibition boosts LPS-mediated IFNβ production in macrophages. Fig. 2 | [100]Open in a new tab a, Measurement of the complex II (CII)-specific oxygen consumption rate (OCR), assessed through a Seahorse assay, of Irg1^+/+ and Irg1^−/− BMDMs (n = 5 technical replicates; LPS 24 h). Data are from one experiment. b, Metabolomics data showing relative expression of itaconate (P = 4.15 × 10^−15), succinate (P = 2.854.15 × 10^−13), fumarate (P = 1.00 × 10^−13) and malate (P = 3.38 × 10^−5) in Irg1^+/+ and Irg1^−/− BMDMs. Data are from a publicly available dataset^[101]17 (n = 3 mice; LPS 24 h). c,d, qPCR data on Ifnb1 from BMDMs pretreated with AA5 (c), TTFA (d) or DMSO (3 h) (n = 3 mice, LPS 4 h). Data are from three independent experiments. e, IFNB expression, assessed by qPCR, in human MDMs pretreated with DMSO or TTFA (0.5 mM; 3 h) (n = 4 donors; LPS 4 h). Data are from two independent experiments f, IFNβ release in Sdha^+/+ and Sdha^−/− RAW.264.7 macrophages (n = 2 mice; LPS). Data are from one experiment. g,h, IFNβ release (g) and Ifnb1 expression (h) in BMDMs from Sdhb^fl/fl and Sdhb^−/− cells (n = 2 mice, LPS 4 h). Data are from one experiment. i, Heatmap showing relative expression (z score) of interferon-stimulated genes (ISGs) commonly upregulated in BMDMs following TTFA and itaconate treatment (P[adj] < 0.05). (n = 3 mice, LPS 4 h). Data are from three independent experiments. j, IFNB and IL1B expression in SDH-deficient (n = 5 patients) and SDH-competent (n = 5 patients) tumor samples. In all graphs, data are presented as mean ± s.e.m. P values were calculated using a one- or two-way ANOVA. We then used 2-theoyltrifluoroacetone (TTFA) and Atpenin A5 (AA5), two specific and well-characterized pharmacological SDH inhibitors^[102]19,[103]20, to explore whether SDH inhibition could drive IFNβ expression. Both compounds enhanced Ifnb1 transcription in LPS-activated macrophages ([104]Fig. 2c,[105]d). TTFA also increased IFNβ mRNA levels in human macrophages ([106]Fig. 2e). Using a RAW.264 Sdha^−/− macrophage cell line, we observed a major enhancement in IFNβ production over the course of LPS stimulation ([107]Fig. 2f). Additionally, we generated a mouse model of tamoxifen-inducible Sdhb loss, and we generated BMDMs from these animals ([108]Extended Data Fig. 3a). We found that BMDMs lacking Sdhb had decreased levels of Il1b expression, as expected ([109]Extended Data Fig. 3b), as well as augmented levels of IFNβ release and transcription ([110]Fig. 2g,[111]h). To further characterize this phenotype, we performed RNA-seq using TTFA-treated BMDMs, in which we found profound transcriptomic changes ([112]Extended Data Fig. 1a). We compared itaconate and TTFA-treated BMDM transcriptomic data and found many commonly upregulated IFN-dependent genes ([113]Fig. 2i), further indicating that SDH inhibition by itaconate was responsible for the induction of type I IFN. TTFA also increased IFNβ levels in Irg1^−/− BMDMs, further supporting a role for this pathway in a model of increased SDH activity ([114]Extended Data Fig. 3c). Considering that mesaconate is a derivative of itaconate, we ruled out any potential effects of mesaconate itself, because it does not affect IFNβ expression ([115]Extended Data Fig. 3d). Last, we examined SDH-deficient tumors versus site- and age-matched SDH-competent tumors. We found in these SDH-deficient samples an increase in IFNB expression but no change in IL1B expression, indicating that a specific IFN response was upregulated ([116]Fig. 2j). SDH inhibition activates cytosolic-RNA-sensing pathways Next, we aimed to investigate how SDH inhibition by itaconate increases type I IFN responses. We performed pathway enrichment analysis of the commonly upregulated differentially expressed genes in itaconate- and TTFA-treated macrophages ([117]Fig. 3a). Our results show, within the REACTOME analysis, that the ‘DDX58/IFIH1-mediated Induction of IFN-α/β’ stood out as a possible candidate pathway that could account for the upregulation of type I IFN ([118]Fig. 3b). Additionally, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis revealed a notable upregulation of the ‘RIG-I-like receptor signalling pathway’ ([119]Fig. 3c), suggesting that RIG-I and MDA5 were possibly involved. Fig. 3 |. SDH inhibition increases IFNβ production through RIG-I- and MDA5-dependent signaling. Fig. 3 | [120]Open in a new tab a, Venn diagram showing the number of shared upregulated differentially expressed genes, determined by RNA-seq, between TTFA- and itaconate-treated BMDMs (n = 3 mice; LPS 4 h, P[adj] < 0.05). b,c, REACTOME pathway enrichment analysis (b) and KEGG pathway enrichment analysis (c) of commonly upregulated differentially expressed genes between TTFA- and itaconate-treated BMDMs (n = 3 mice; LPS 4 h). Data are from three independent experiments. FDR, false discovery rate. d, Ifnb1 expression after siRNA knockdown of Mb21d1 (P = 8.42 × 10^−10) and Ddx58 in BMDMs pretreated with DMSO or TTFA (n = 5 mice; LPS 4 h). Data are from two independent experiments. e, Ifnb1 expression after siRNA knockdown of Tlr9 and Ifih1 in BMDMs pretreated with DMSO or TTFA (n = 3 mice; LPS 4 h). Data are from three independent experiments. f, IFNβ release after treatment with NC siRNA (n = 7 mice) or siRNA targeting Mb21d1 (n = 4 mice), Ifih1 (n = 6) or Ddx58 (n = 6 mice) in BMDMs pretreated with PBS or ITA (LPS 4 h). Data are from three independent experiments. In all graphs, data are presented as mean ± s.e.m. P values were calculated using two-way ANOVA. siRNAs targeting Ddx58 ([121]Extended Data Fig. 4a) and Ifih1 ([122]Extended Data Fig. 4b) led to inhibition of the TTFA-dependent increase of LPS-induced IFNβ production ([123]Fig. 3d,[124]e). We found that knockdown of cGAS (encoded by Mb21d1) ([125]Extended Data Fig. 4c) or TLR9 ([126]Extended Data Fig. 4d) did not affect IFNβ production, indicating that DNA sensing is unlikely to be involved in this process ([127]Fig. 3d,[128]e). Similarly, we showed that itaconate-mediated IFNβ release was also dependent on MDA5 and RIG-I and independent of cGAS ([129]Fig. 3f and [130]Extended Data Fig. 4e–[131]g). We then confirmed that our siRNA approach successfully impaired the ability of the classical stimuli–cGAS, MDA5 and RIG-I–to induce IFN production, validating our method ([132]Extended Data Fig. 4h–[133]j). Further ruling out the role of DNA sensing in the system, we also found that knockdown of STING (encoded by Tmem173) had no effect on TTFA-induced IFNβ production ([134]Extended Data Fig. 4k,[135]l). SDH inhibition drives mtRNA release in a VDAC1-dependent manner Recent research has shown that mitochondrial nucleic acids regulate type I IFN production by activating both RNA- and DNA-sensing pathways^[136]21–[137]23. To confirm that mitochondrial nucleic acids are required for the increased IFNβ production with SDH inhibition, we used ethidium bromide to deplete mtRNA levels ([138]Extended Data Fig. 5a). We found that, in the presence of ethidium bromide, the increase in IFNβ production due to TTFA was reduced ([139]Fig. 4a), suggesting that mitochondrial nucleic acids are involved in the increase in type I IFN signaling with SDH inhibition. We then treated cells with IMT1, an inhibitor of the mitochondrial transcription factor POLRMT, which led to reduced expression of mtDNA-encoded genes ([140]Extended Data Fig. 5b). In the presence of IMT1, itaconate could not increase Ifnb1 levels ([141]Fig. 4b), showing that the effect of itaconate on type I IFN requires mtRNA. Fig. 4 |. SDH inhibition drives mtRNA release in a VDAC1-dependent manner. Fig. 4 | [142]Open in a new tab a, Ifnb1 expression in BMDMs in control or ethidium bromide (EtBr)-containing medium pretreated with DMSO or TTFA (n = 3 mice; LPS 4 h). Data are from three independent experiments. b, Ifnb1 expression in BMDMs pretreated with DMSO, IMT1, PBS or ITA (n = 4 mice; LPS 4 h). Data are from two independent experiments. c, mt-ND4 expression in the isolated cytosolic fraction of BMDMs pretreated with PBS or ITA (n = 3 mice; LPS 4 h). Data are from three independent experiments. d,e, Expression of the D-loop region of mtDNA in the isolated cytosolic fraction of BMDMs pretreated with DMSO, TTFA (d) or DMM (e) (n = 8 mice; LPS 4 h). Data are from three independent experiments. f, Heatmap of RNA-seq data for mitochondrial genes in BMDMs pretreated with ITA or TTFA (n = 3 mice; LPS 4 h). Data are from three independent experiments. g, Ifnb1 expression in the presence of NC siRNA or siRNA targeting Snx9 or Vdac1 in BMDMs pretreated with DMSO (n = 6 mice; LPS 4 h). Data are from three independent experiments. h,i, IFNβ release (h) and Ifnb1 expression (P = 2.13 × 10^−11) (i) in BMDMs pretreated with VBIT-4 (16 h), DMSO or TTFA. h, (n = 3 mice; LPS 4 h); i, (n = 6 mice; LPS 4 h). Data are from three independent experiments. j, Ifnb1 expression in BMDMs pretreated with DMSO or VBIT-4 (16 h) followed by PBS or ITA pretreatment (n = 3 mice; LPS 4 h). k, D-loop, mt-ND4 and mt-ND6 expression in the isolated cytosolic fraction of BMDMs pretreated with DMSO or VBIT-4 (16 h) followed by PBS or ITA pretreatment (n = 4; LPS 4 h). Data are from three independent experiments. l, D-loop expression in the cytosolic fraction of BMDMs (n = 6; LPS 48 h). Data are from three independent experiments. m, Expression of D-loop (P = 2.32 × 10^−5), mt-Nd4 (P = 4.22 × 10^−6), mt-Nd5 (P = 1.10 × 10^−5) and mt-ND6 (P = 2.03 × 10^−5) in the cytosolic fraction of Irg1^+/+ and Irg1^−/− BMDMs (n = 5 mice; LPS 24 h). Data are from three independent experiments. n, Schematic of the mechanism underlying how SDH impairment leads to increased IFNβ production. The figure was created with [143]BioRender.com. Data in graphs are presented as mean ± s.e.m. P values were calculated using one or two-way ANOVA for multiple comparisons or two-tailed Student’s t-test for pairwise comparisons. Considering that the increase in IFNβ expression in response to TTFA and itaconate seems to be dependent on cytosolic double-stranded RNA (dsRNA) sensors, we next investigated how these sensors were activated. Using digitonin to isolate the cytosolic from the organellar fraction within BMDMs ([144]Extended Data Fig. 5c), we found that both itaconate and the SDH inhibitors TTFA and dimethyl malonate (DMM) enriched cytosolic mtRNA levels derived from the heavy (ND4 and ND5) and light (ND6) chains ([145]Fig. 4c–[146]e). Strikingly, we found an overall increase in the transcriptional levels of mtDNA-encoded RNAs in inflammatory macrophages, which could serve as a priming signal required to release mtRNA upon SDH inhibition ([147]Fig. 4f). We then examined how mtRNA is released from mitochondria into the cytosol. We first knocked down Bax and Bak1 expression and found that TTFA could still increase IFNβ levels ([148]Extended Data Fig. 5d). We also found that TTFA could increase IFNβ production after Snx9 knockdown, whereas Vdac1 knockdown abrogated the increase in IFNβ production ([149]Fig. 4g and [150]Extended Data Fig. 5e,[151]f). This suggests that VDAC1 pores mediate mtRNA release upon SDH inhibition. To further examine this pathway, we used a well-characterized inhibitor of VDAC1 oligomerization, VBIT-4 (ref. [152]24), which also reduced IFNβ production in the presence of TTFA ([153]Fig. 4h,[154]i) and itaconate ([155]Fig. 4j). Interestingly, we also found that SDH inhibition strongly increases mitochondrial membrane potential (MMP), as assessed by TMRM staining in immunofluorescence ([156]Extended Data Fig. 5g). This process might be responsible for mtRNA release and VDAC1 oligomerization, given that previous reports have found that other methods of increasing MMP levels, such as FH inhibition and ATP synthase inhibition, resulted in mtRNA release and type I IFN production^[157]17. This also demonstrates that VDAC1 is directly required for the release of mtRNA into the cytosol by itaconate, as VBIT-4 impaired the accumulation of cytosolic mtRNAs in response to itaconate ([158]Fig. 4k). We next examined whether the release of mtRNA was regulated by endogenously produced itaconate. Treatment of BMDMs with LPS for 48 h induced the release of mtRNA into the cytosol ([159]Fig. 4l). We also found that cytosolic mtRNA levels were reduced in LPS-activated Irg1^−/− BMDMs ([160]Fig. 4m), suggesting that endogenously produced itaconate is sufficient to activate this pathway and regulate IFN production. Overall, we found that inhibition of SDH by itaconate promotes mtRNA release through VDAC1 to drive type I IFN through RIG-I and MDA5, as represented in [161]Figure 4n. Discussion It is now widely understood that TCA cycle metabolites can act as immunomodulatory signals that regulate cytokine production, reclassifying them as ‘immunometabolites.’ Specifically, several reports have explored the TCA cycle and the regulation of IFN production^[162]25. Focusing on further understanding the role of itaconate in macrophage biology, we provide evidence that itaconate is an endogenous regulator of mtRNA release and type I IFN production. In our unbiased approach, production of type I IFN stood out as a pathway regulated by itaconate. Itaconate boosted LPS-mediated IFNβ expression in murine and human macrophages. Additionally, genetic ablation of the IRG1–itaconate pathway abrogated type I IFN production in LPS-stimulated macrophages and PBMCs. Pharmacological and genetic approaches targeting SDH indicated that SDH inhibition by itaconate is responsible for this effect. Mitochondria are currently being repositioned as more than just organelles involved in bioenergetics, but as dynamic and multifaceted organelles of endosymbiotic origin with multiple behaviors, functions and activities^[163]26,[164]27. Mitochondrial release of nucleic acids has been shown to occur in numerous contexts, including apoptosis, inflammatory diseases, mtDNA stress through TFAM deficiency, inflammasome activation and infection^[165]28. They consist of mtRNA and mtDNA. mtdsDNA can activate AIM2, NLRP3, cGAS and TLR9 DNA sensors, whereas mtRNA can activate RNA sensors, such as RIG-I and MDA-5, to boost IFNβ and cytokine production^[166]29. Dysregulation of mitochondrial metabolic enzymes, such as FH, has also been shown to cause the release of mtRNA into the cytosol to boost IFNβ production^[167]17,[168]30. This led us to examine whether SDH inhibition by itaconate drives mtRNA release, which could be responsible for increased levels of IFNβ. Initially, data suggested that itaconate was a negative regulator of type I IFN production^[169]1,[170]31. However, this was based on evidence from derivatized itaconate (4-OI and DI), which is more electrophilic and cysteine-reactive than endogenous itaconate^[171]1,[172]5. Although 4-OI can be directly converted to itaconate once inside macrophages^[173]8, a large proportion of 4-OI likely remains esterified, given that many mass-spectrometry studies have identified modification of cysteine residues by 4-OI itself^[174]1,[175]8. Furthermore, 4-OI is used in the μM range in vitro, which is insufficient to outcompete physiological levels of succinate^[176]6. Mechanistically, the suppression of type I IFN production by derivatives is not fully understood, although evidence suggests the activation of Nrf2 is involved^[177]32. Notably, at concentrations typically used, a crucial discrepancy between underivatized itaconate and 4-OI and DI is that the derivatives failed to increase intracellular levels of succinate and inhibit SDH^[178]5. Our observation in Irg1^−/− BMDMs, consistent with data presented in Swain et al.^[179]5, shows decreased IFNβ levels in cells lacking endogenous itaconate. Therefore, inhibition of SDH by itaconate, a property first described more than six decades ago^[180]33, regulates production of type I IFN. Mechanistically, we suggest that SDH inhibition by itaconate increases production of type I IFN through the activation of cytosolic dsRNA-sensing pathways, which were revealed by Gene Set Enrichment Analysis to be activated by TTFA and itaconate treatment. Furthermore, genetic ablation of RIG-I and MDA5 blocked the ability of TTFA and itaconate to upregulate type I IFN production, whereas DNA-sensing pathways were dispensable. As we see more of a dependency on MDA5 signaling, this suggests the role for longer strand dsRNA in response to itaconate. We found increased mtRNA present in the cytosol when SDH was impaired. Strikingly, Irg1^−/− BMDMs show a decrease in cytosolic mtRNA levels, indicating that this is an endogenous process mediated by itaconate. To investigate how mtRNA is released, genetic knockdown of the mitochondrial transporter VDAC1, as well as pharmacological inhibition of VDAC pore formation with VBIT-4, blocked the release of mtRNA and the induction of IFNβ expression driven by SDH inhibition and itaconate. This demonstrates that itaconate induces the release of mtRNA by triggering the formation of VDAC1 pores. Our work therefore adds to the understanding of itaconate as an immunomodulatory metabolite ([181]Fig. 4n). Inhibition of SDH by itaconate is an endogenous process that is essential for metabolic reprogramming in inflammatory macrophages, leading to type I IFN production. This process could play a role in infection or inflammatory diseases, whereby itaconate and type I IFNs are dysregulated. Indeed, previous work has shown that itaconate is an anti-viral metabolite and that this process might involve SDH inhibition^[182]34–[183]36. Our work also extends to other disease contexts in which SDH is inhibited, such as cancer, and could reveal a new therapeutic target for SDH-deficient tumors. Although our study uncovers a new pathway by which itaconate regulates type I IFN production, further studies are required to investigate the potential impacts of this pathway in the context of infection or inflammation wherein itaconate is highly upregulated. Methods Animal details Wild-type (WT) mice for most in vitro experiments were on a C57BL/6JOlaHsd background. WT mice were bred in house or purchased from Harlan laboratories. Irg1^−/− mice (named C57BL/6N-Acod1^em1(IMPC)J/J) were generated through CRISPR-targeted deletion of exon 4 of Acod1. These mice were purchased from The Jackson Laboratory. The Sdhb^fl/fl mice were donated by N. Armstrong, and R26-Cre-ERT2 mice were purchased from The Jackson Laboratory. Experimental mice were homozygous for the conditional LoxP Sdhb allele and expressed the Cre recombinase–ERT2 fusion under control of the Rosa26 promoter (Sdhb^fl/fl; R26 CreERT2/CreERT2). In vitro experiments, which were performed on BMDMs, were done in 6- to 18-week-old female mice. Generation of BMDMs Mice (6–18 weeks old) were euthanized using CO[2], and cervical dislocation was used to confirm death; immediately afterwards, bone marrow was flushed from the tibia, femur and ilium, and collected cells were differentiated in DMEM medium containing L929 supernatant (20%), fetal calf serum (FCS) (10%) and antibiotics (penicillin– streptomycin) (1%) for 6 days. The differentiated BMDMs were then counted and plated at 0.5 × 10^6 cells ml^−1, unless stated otherwise. BMDMs were plated in 12- or 6-well cell culture plates and left to adhere for at least 12 hours before experiments. After replating BMDMs from Sdhb^fl/fl; R26 CreERT2/CreERT2 mice, cells were cultured in DMEM containing 4-hydroxytamoxifen (100 uM) for 72 h before LPS stimulation. Generation of human PBMCs and macrophages Blood samples from healthy donors were collected and processed at the School of Biochemistry and Immunology at the Trinity Biomedical Sciences Institute (TCD). Blood was donated and obtained anonymously. Written informed consent for the use of blood for research purposes was given from the donors. Procedures for experiments using human blood were approved by the School of Biochemistry and Immunology Research Ethics Committee (TCD). All experiments were conducted according to the TCD guide on good research practice, which follows the guidelines detailed in the National Institutes of Health Belmont Report (1978) and the Declaration of Helsinki. Whole blood (30 ml) was layered on 20 ml Lymphoprep (Axis-Shield) and centrifuged for 20 min at 400g without the brake on. The upper plasma layer was then discarded, and the white layer of PBMCs at the plasma-density-gradient medium interface was retained, and 20 ml PBS was subsequently added. Cells were centrifuged for 8 min at 300g, after which the supernatant was discarded. The pelleted PBMCs were then resuspended in RPMII, counted using an automated cell counter and plated at 1 x 10^6 cells ml^−1 in RPMI supplemented with FCS (10%) and antibiotics (penicillin– streptomycin) (1%). Once PBMCs were obtained, CD14^+ monocytes were isolated using a MagniSort Human CD14 Positive Selection kit (Thermo Fisher), according to the manufacturer’s protocol. Human monocytes were then differentiated in T175 flasks in RPMI containing FCS (10%), antibiotics (penicillin–streptomycin) (1%) and recombinant human macrophage colony-stimulating factor (1:1,000). After 6 days of differentiation, the supernatant was removed, and remaining monocyte-derived macrophages were scraped and counted, and human monocyte-derived macrophages were plated in 12-well plates at 1 × 10^6 cells ml^−1 in RPMI containing FCS (10%) and antibiotics (penicillin– streptomycin) (1%). SDH-deficient tumor samples Tissue excess to diagnostic requirements, from SDH-deficient (n = 5) and sex-, age- and anatomical-site-matched SDH-competent (n = 5) tumors, were used with consent or assent from the donor or their guardian. Consent or assent was obtained following approval of the study by the Medical and Research Ethics Committee at Children’s Health Ireland. RNA was extracted in the diagnostic Histology Laboratory at Children’s Health Ireland from these tumors for use in qPCR, as described. SDHB immunohistochemistry was performed on the diagnostic tissue, prompting confirmatory germline genetic testing and review by the clinical geneticist, where the tumor immunoreactivity for SDHB was lost. Cell-line details Generation of macrophages from THP1 cell lines. THP1 cell lines were donated by F. Dorsay at Eli Lilly and Company. Cells were maintained at a concentration range between 5 × 10^5 and 1.0 × 10^6 cells ml^−1 in RPMI–glutamax medium with 10% FCS, 1% penicillin–streptomycin and 55 μM β-mercaptoethanol. Cells were plated at 0.5 × 10^6 cells ml^−1 and treated with 100 ng ml^−1 PMA for 48 h for macrophage differentiation. Cells were then treated. SDHA RAW 264.7 cells. Cell lines were donated by M. Simarro, as described in ref. [184]37. SDHA RAW 264.7 cell lines were maintained in DMEM, 10% FCS and 1% penicillin–streptomycin. Cells were scraped in PBS and centrifuged at 500g for 5 min, and were then resuspended and counted. Cells were plated at 0.5 × 10^6 cells ml^−1 and left overnight. Cells were then treated. Reagents. LPS was obtained from Escherichia coli serotype EH100 (ALX-581-010-L001), which was acquired from Enzo Life Sciences. High-molecular-weight (HMW) poly(I:C) (tlrl-pic), G3-YSDctrl (tlrl-ydnac), G3-YSD (tlrl-ydna), 5′ppp-dsRNA-control (tlrl-3prnac) and 5′ppp-dsRNA (tlrl-3prna) were purchased from Invivogen. DMSO (D8418), itaconate (I29204-100G), DMM (136441-250g) and DEM (D97703-100G) were all purchased from Sigma-Aldrich. DMF (HY-17363), TTFA (HY-D0190), IMT1 (HY-134539), VBIT-4 (HY-129122) and AA5 (HY-126653) were purchased from MedChemExpress. IFN was purchased from ImmunoTools (11343536). Compound treatments. All compounds used DMSO as a vehicle, unless otherwise stated. DMM was diluted in ethanol, and itaconate was prepared in PBS. LPS was used at a concentration of 100 ng ml^−1, unless otherwise specified, for the indicated time points (in figure legends). HMW Poly(I:C), 5′-PPP-ctrl, 5′-PPP-dsRNA, G3-YSD-ctrl and G3-YSD were transfected at the indicated concentrations using lipofectamine RNAiMax for 6 h. TTFA (500 μM), IMT1 (10 μM), DMM (5 mM), DEM (100 uM), ITA (1-10 mM), 4-OI (62.5–250 μM), DMF (25 μM) and AA5 (1 and 2 μM) pretreatments were performed for 3 h before the addition of LPS. VBIT-4 (10 uM) was used for 16 h before treatment with other compounds. RT–qPCR. RNA extraction from cells was performed using the Purelink RNA kit (Invitrogen), according to the manufacturer’s instructions. BMDMs were treated as required and immediately lysed in 350 μl of RNA lysis buffer. Isolated RNA was quantified using a NanoDrop 2000 spectrophotometer, and the RNA concentration was normalized to the lowest concentration across all samples using RNase-free water. If needed, samples were treated with DNase I (Thermo Fisher) after quantification, following the manufacturer’s guidelines. Isolated RNA was then normalized and converted into cDNA using the High-Capacity cDNA Reverse Transcription kit (Thermo Fisher), as directed. Next, 10 μl of RNA (maximum concentration of 100 ng μl^−1) was mixed with 10 μl of reverse transcription master mix to complete the reaction. Real-time qPCR (RT–qPCR) was conducted on the resulting cDNA using primers designed in-house and ordered from Eurofins Genomics, as detailed in [185]Supplementary Table 1. The reaction was performed in a 96-well qPCR plate using the 7500 Fast Real-Time PCR machine (Thermo Fisher). Relative expression (2^−ΔΔCT) was calculated from the CT values for each sample and gene of interest. RNA interference in BMDMs. Pre-designed silencer select siRNAs were used, targeting Cgas (s103166), Tmem173 (s91058), Tlr9 (s96268), Ddx58 (s106376), Ifih1 (s89787), Vdac1 (s75920), Snx9 (s83576), Bak1 (s62859) or Bax (s62874); and NC siRNA (4390843) was also used. These were ordered from Thermo Fisher. BMDMs were transfected with 50 nM siRNA using 5 μl Lipofectamine RNAiMAX, according to the manufacturer’s instructions (Thermo Fisher). BMDMs were transfected in DMEM without serum and antibiotics, which was replaced with a complete medium after 8 h. BMDMs were then left for at least a further 12 h before experiments. RNA interference in PBMCs. Anonymized untransfused blood samples were obtained from the Irish Blood Transfusion Board. The School of Biochemistry and Immunology Research Ethics Committee (TCD) approved all the procedures involving experiments on human samples, which were conducted according to the TCD guide on good research practice. Thirty milliliters of whole blood was layered on 20 ml Lymphoprep (Axis-Shield), followed by centrifugation for 20 min at 400g with the brake off, after which the upper plasma layer was removed and discarded. The mononuclear layer at the plasma-density gradient medium interface was retained, and 20 ml PBS was added. Cells were centrifuged for 8 min at 300g, and the resulting supernatant was removed and discarded. The remaining pellet of mononuclear cells was resuspended and counted, and 1 × 10^6 cells per point were used for siRNA transfection. The cells were resuspended in nucleofector buffer along with five pmol siRNA (IRG1 (s60727)) and pulsed using the Y-010 program on the AMAXA nucleofector machine. The cells were plated in 1 ml of RPMI supplemented with FCS (10%) and placed in a CO[2] incubator for 48 h. Ethidium bromide treatment. Cells were replated in the presence or absence of ultrapure ethidium bromide (100 ng ml^−1) and incubated in the medium for 6 days before treatment to generate mtDNA-depleted cells. Depletion of mtRNA was determined by RT–qPCR of mtDNA-encoded genes (mt-ND4, mt-ND5, mt-ND6). Digitonin fractionation. BMDMs were plated at 0.5 × 10^6 cells per well prior to treatment. Following treatment, cells were washed once with room-temperature PBS, scraped on ice into ice-cold PBS and pelleted at 500g for 5 min at 4 °C. The supernatant was discarded, and the pellet was resuspended in 400 μl of extraction buffer (150 mM NaCl, 50 mM HEPES pH 7.4, and 25 μg ml^−1 digitonin). Samples were rotated at 4 °C for 10 min before being centrifuged at 2,000g at 4 °C for 5 min. The resulting supernatant constituted the cytosolic fraction, from which RNA and DNA were isolated using the All Prep DNA/RNA Mini kit (Qiagen). Alternatively, the cytosolic fraction was concentrated with Strataclean resin (Agilent) and analyzed by western blot. The pellet, containing membrane-bound organelles, was lysed in RNA lysis buffer for RNA isolation or in western blot lysis buffer for analysis by western blotting. To detect mtRNA and mtDNA in the cytosol, qPCR was performed using primers specific for the mitochondrial D-loop on cDNA reverse-transcribed from cytosolic RNA (mtRNA) and on DNA isolated from the cytosolic fraction (mtDNA). RNA-seq. BMDMs from three mice were treated as indicated by TTFA and ITA, and RNA was extracted as detailed above using the Purelink RNA extraction kit. The mRNA was isolated from total RNA using poly-T-oligo-attached magnetic beads. After fragmentation, the first strand of complementary DNA was synthesized using random hexamer primers, followed by a second strand of cDNA synthesis. The final library was then checked with Qubit and real-time PCR for quantification, and a bioanalyzer was used to determine the size distribution. The quantified libraries were then pooled and sequenced using a NovaSeq 6000 S4 (Illumina). Differential expression analysis of the conditions per group was then performed using normalized read counts and the DESeq2 R package. Gene set and pathway enrichment analyses were performed as described below in the ‘[186]Quantification and statistical analysis’ section. Seahorse XF complex II. BMDMs were plated at 100,000 cells per well in 100 μl of DMEM and left overnight to adhere. The protocol was carried out as previously described^[187]38. In brief, BMDMs were treated as indicated, and the medium was then replaced with a mitochondrial assay solution (MAS) and the manufacturer’s plasma membrane permeabilizer (Agilent). BMDMs were then kept in a CO[2]-free incubator for 40 min. The cell plate was then transferred to the Seahorse XFe/XF96 Analyzer, and the experiment was performed. After the experiment, Seahorse Wave software (Agilent) was used for analysis. Western blotting. After treatment, the cell supernatant was then removed. Cell lysates were then collected in 30 μl Laemmli buffer (0.125 M Tris pH 6.8, 10% glycerol, 0.02% SDS and 5% DTT) and they were subsequently heated to 95 °C for 5 min to denature proteins. SDS–PAGE was then used to resolve proteins according to their molecular weight. The samples were also heated at 95 °C for 5 min before loading into a 5% stacking gel. The percentage resolving gel used depended on the molecular weight of the given protein of interest and was typically between 10% and 15%. The Bio-Rad gel running system was used to separate proteins, and the Bio-Rad wet transfer system was used to perform the electrophoretic transfer of proteins on PVDF membrane. The membrane was then incubated in powdered milk (5% in TBST) for 1 h and subsequently incubated in primary antibody (IRG1 (Cell Signalling Technologies, 17805), Lamin b1 (Cell Signalling Technologies, 12586), alpha-tublin (Cell Signalling Technologies, 2144) or β-actin (Sigma-Aldrich (A5316)), rolling overnight at 4 °C. The membrane was then washed three times for 5 min each, and incubated for 1 h with a horseradish-peroxidase-conjugated secondary antibody (diluted in 5% milk powder) at room temperature. Before development, the PVDF membrane was sprayed with Western Bright ECL Spray (Advansta). Protein visualization was performed using a ChemiDoc MPTM imaging system (Bio-Rad), and both chemiluminescent and white light images were taken. Images were analyzed using Image Lab 6.0.1 (Bio-Rad). ELISA. DuoSet ELISA kits for IFN-β were purchased from R&D Systems and were carried out according to the manufacturer’s instructions with appropriately diluted cell supernatants added to each plate either in technical duplicates or triplicates. Absorbance at 450 nm was quantified using a FLUOstar Optima plate reader. Corrected absorbance values were calculated by subtracting the background absorbance, and cytokine concentrations were subsequently obtained by extrapolation from a standard curve plotted using GraphPad Prism 9.2.0. Immunofluorescence imaging. BMDMs were plated in 35 mm dishes at 0.5 × 10^6 (cells ml^−1) in 2 ml DMEM. After treatments, TMRM (25 nm) and HOESCHT blue (0.5 μg ml^−1) were added to the culture medium for 30 min at 37 °C. Cells were then washed three times with warm PBS, and fresh DMEM (10% FCS) containing TMRM (1.25 nM) was added. Slides were imaged using a Leica SP8x scanning confocal microscope with a ×100.0 objective. Images were analyzed using the LAS X Life Science Microscope Software Platform (Leica). The same microscope instrument settings were used for all samples, and all images were analyzed using the same settings. Quantification and statistical analyses. Details of all statistical analyses are provided in the figure legends. Representative western blots are shown, and uncropped images are provided as source data. One-way ANOVA corrected for multiple comparisons using the Tukey statistical test was used, and P[adj] < 0.05 was set as the significance cut-off. The Bioconductor package limma was used for analysis using an R-based online tool. Data were visualized using a heatmap with autoscaled features (genes). IREA was used to uncover important cytokines in itaconate-treated cells, as displayed in [188]Figure 1a. Enrichr was used to analyze TTFA- and itaconate-regulated gene sets, as displayed in [189]Figure 3a–[190]c. GraphPad Prism v.9.2.0 was used to calculate statistics in bar charts using the appropriate statistical tests depending on the data described in the figure legends, including one-way ANOVA and two-tailed unpaired t-test. P[adj] values were assessed using appropriate correction methods, such as Tukey, Kruskal–Wallis and Holm–Sidak tests. Sample sizes were determined on the basis of those in previous experiments using similar methodologies. Unless stated otherwise, all depicted data points are biological replicates taken from distinct samples. Each figure consists of at least three independent experiments from multiple biological replicates, unless stated otherwise. Extended Data Extended Data Fig. 1 |. Immunoregulatory effects of itaconate. Extended Data Fig. 1 | [191]Open in a new tab a, PCA plot of RNA-Seq analysis. b, Differential expression gene clustering heatmap with Z-score. c, MsigDB Hallmark pathway enrichment analysis of differentially expressed genes from itaconate pretreated BMDMs followed by 3 h LPS treatment. d, NRF2 pathway enrichment plot of itaconate-treated BMDMs. e) Western blot of IRG1 in Irg1^+/+ and Irg1^−/− BMDMs treated with LPS for 24 h and 48 h. f, qPCR of IRG1 (p = 1.017e-11) in human PBMCs that were harvested in the presence of Irg1 siRNA (n=3 donors; LPS 24 h). Data are presented as mean ± SEM. P values were calculated using two-way ANOVA for multiple comparisons. Extended Data Fig. 2 |. Effect of electrophilic compounds on IFNβ. Extended Data Fig. 2 | [192]Open in a new tab a, IFNβ release and b, Ifnb1 expression in BMDMs pretreated with DMSO or 4-OI a (n=6 mice; LPS 4 h) b, (n=9; LPS 4 h). c, Ifnb1 expression of BMDMs pretreated with DMSO or DEM (n=3 mice; LPS 4 h) from 3 independent experiments. d, Normalised counts from RNA-seq data in BMDMs pretreated with DMSO or DMF (n=3 mice; LPS 4 h) from 3 independent experiments Data are presented as mean ± SEM. P values were calculated using one-way ANOVA for multiple comparisons or two-tailed student’s t-test for pairwise comparisons. Extended Data Fig. 3 |. SDH-deficiency is involved in the itaconate-interferon phenotype. Extended Data Fig. 3 | [193]Open in a new tab a, Sdhb and b, Il1b expression in Sdhb^fl/fl and Sdhb^−/− BMDMs (n=2 mice; LPS 4 h) from 2 independent experiments. c, IFNB expression in IRG1^+/+ and IRG1^−/− THP-1 cell lines pretreated with DMSO or TTFA (n=4 replicates; LPS/IFNγ 4 h) from 2 independent experiments d, IFNB expression in THP-1 derived macrophages pretreated with PBS or Mesaconate (n=4 replicates; LPS/IFNγ 4 h) from 2 independent experiments. Data are presented as mean ± SEM. P values were calculated using two-way ANOVA for multiple comparisons or two-tailed student’s t-test for pairwise comparisons. Extended Data Fig. 4 |. Knockdown of nucleic acid sensors in BMDMs. Extended Data Fig. 4 | [194]Open in a new tab a-d, Expression of a, Ddx58, b, Ifih1, c, Mb21d1, and d, Tlr9 in BMDMs in the presence of a,c, Mb21d1, Ddx58, b,d, Tlr9 and Ifih1 siRNA pretreated with DMSO or TTFA (n=3 mice; LPS 4 h) from 3 independent experiments. e-g, Expression of e, Mb21d1 f, Ifih1, and g, Ddx58 in the presence of Mb21d1, Ifih1, and Ddx58 siRNA in BMDMs pretreated with PBS or ITA (n=2-3 mice; LPS 4 h) from 2 or 3 independent experiments. h-j IFNβ release in BMDMs transfected with h, G3-YSD, i, Poly(I:C), or j, 5′PPP-dsRNA in the presence of h, Mb21d1, i, Ifih1, or j, Ddx58 siRNA (n=3 mice; 4 h) from 3 independent experiments. k, Ifnb1 and l, Tmem173 expression in BMDMs pretreated with DMSO or TTFA in the presence of Tmem173 siRNA (n=3 mice; LPS 4 h). Data are presented as mean ± SEM. P values were calculated using two-way ANOVA for multiple comparisons. Extended Data Fig. 5 |. mtRNA is released in a VDAC1-dependent manner. Extended Data Fig. 5 | [195]Open in a new tab a, qPCR of mitochondrial DNA-encoded genes in BMDMs after mtDNA depletion with EtBr (6 days) followed by TTFA pretreatment (0.5 mM, 3 h) and LPS stimulation (n = 3 mice, 4 h) from 3 independent experiments. b, qPCR of mitochondrial DNA-encoded gene in BMDMs after POLRMT inhibition with IMT1 followed by itaconate pretreatment (3 h) and LPS stimulation (n = 3 mice, 4 h) from 3 independent experiments. c, Western blot of organellar and cytosolic fractions of BMDMs after subcellular digitonin fractionation (n=2 mice); Images representative of two independent experiments. d, qPCR of Ifnb1, Bak1, and Bax, after Bak1 or Bax siRNA knockdown followed by TTFA pretreatment (3 h) and LPS stimulation (n = 3 mice, 4 h) from 3 independent experiments. qPCR data of e, Vdac1 or f, Snx9 in BMDMs after Vdac1 or Snx9 siRNA knockdown followed by pretreated with TTFA (0.5 mM, 3 h) followed by LPS stimulation (n = 5 mice, 4 h) from 3 independent experiments. g, Immunofluorescence images of BMDMs pretreated with DMSO or TTFA (0.5 mM, 3 h) followed by LPS (n = 2 mice, 4 h). Images representative of 2 independent experiments. Data are presented as mean ± SEM. P values were calculated using two-way ANOVA for multiple comparisons. Supplementary Material 42255_2024_1145_MOESM7_ESM [196]NIHMS2079310-supplement-42255_2024_1145_MOESM7_ESM.xlsx^ (9KB, xlsx) 42255_2024_1145_MOESM8_ESM [197]NIHMS2079310-supplement-42255_2024_1145_MOESM8_ESM.xlsx^ (11.9KB, xlsx) 42255_2024_1145_MOESM9_ESM [198]NIHMS2079310-supplement-42255_2024_1145_MOESM9_ESM.xlsx^ (11.7KB, xlsx) 42255_2024_1145_MOESM10_ESM [199]NIHMS2079310-supplement-42255_2024_1145_MOESM10_ESM.xlsx^ (21.4KB, xlsx) 42255_2024_1145_MOESM11_ESM. [200]NIHMS2079310-supplement-42255_2024_1145_MOESM11_ESM_.xlsx^ (14.2KB, xlsx) 42255_2024_1145_MOESM12_ESM [201]NIHMS2079310-supplement-42255_2024_1145_MOESM12_ESM.pdf^ (154.4KB, pdf) Supplementary Table 1 [202]NIHMS2079310-supplement-Supplementary_Table_1.xlsx^ (22.8KB, xlsx) The online version contains [203]supplementary material available at [204]https://doi.org/10.1038/s42255-024-01145-1. Acknowledgements