Abstract

   The cytokine interleukin-10 (IL-10) limits the immune response and
   promotes resolution of acute inflammation. Because of its
   immunosuppressive effects, IL-10 up-regulation is a common feature of
   tumor progression and metastasis. Recently, IL-10 regulation has been
   shown to depend on mitochondria and redox-sensitive signals. We have
   found that Suppressor of site III[Qo] Electron Leak 1.2 (S3QEL 1.2), a
   specific inhibitor of reactive oxygen species (ROS) production from
   mitochondrial complex III, and myxothiazol, a complex III inhibitor,
   decrease IL-10 in lipopolysaccharide (LPS)–activated macrophages. IL-10
   down-regulation is likely to be mediated by suppression of c-Fos, which
   is a subunit of activator protein 1 (AP1), a transcription factor
   required for IL-10 gene expression. S3QEL 1.2 impairs IL-10 production
   in vivo after LPS challenge and promotes the survival of mice bearing
   B16F10 melanoma by lowering tumor growth. Our data identify a link
   between complex III–dependent ROS generation and IL-10 production in
   macrophages, the targeting of which could have potential in boosting
   antitumor immunity.
     __________________________________________________________________

   Mitochondrial respiratory complex III sustains IL-10 production in
   macrophages and promotes tumor-mediated immune evasion.

INTRODUCTION

   Immune cells undergo metabolic reprogramming during activation ([58]1),
   and the intricacies of the metabolic processes occurring are still
   being unraveled. Macrophages have been a key focus with profound
   changes in glycolysis and mitochondrial metabolism occurring after
   stimulation with activators such as lipopolysaccharide (LPS) from
   Gram-negative bacteria ([59]2).

   They produce, among many cytokines, the anti-inflammatory cytokine
   interleukin-10 (IL-10). Initially found as a T helper 2
   (T[H]2)–specific cytokine ([60]3), IL-10 is a soluble mediator whose
   production is shared by innate and adaptive immune cells such as
   dendritic cells, monocytes, B cells, and regulatory T cells (T[regs])
   ([61]4). IL-10 can elicit immune suppression by limiting antibody
   production in B cells and costimulatory molecules in T cells ([62]5).
   Considering the further literature on mitochondrial biology, IL-10 has
   been shown to decrease reactive oxygen species (ROS) production, drive
   controlled mitophagy, and promote mitochondrial fitness ([63]6). IL-10
   deficiency leads to dysfunctional mitochondria and impaired oxidative
   phosphorylation (OXPHOS) with an increase in glycolysis in macrophages
   ([64]6). More recently, the redox protein thioredoxin has been shown to
   be necessary for the differentiation of IL-10^+ regulatory B cells
   through a mechanism that includes controlling ROS generation ([65]7),
   providing a link between oxidative stress and IL-10.

   Mitochondrial respiratory complex III is part of the electron transport
   chain (ETC), where it oxidizes coenzyme Q and reduces cytochrome c to
   generate adenosine triphosphate (ATP) during oxidative phosphorylation.
   It is also the main source of superoxide anion (O[2]^·−) and subsequent
   ROS from mitochondria, together with complex I, during forward and
   reverse electron transport (FET and RET, respectively) ([66]8). In
   humans, mutations in three complex III–specific genes lead to
   mitochondrial disorders with different onset and clinical
   manifestations ([67]9), making their characterization and treatment a
   challenge. The contribution of complex III to immune cell function
   requires further exploration. Complex III has been shown to be
   essential for hematopoietic stem cell stability by regulating
   epigenetic changes in histones ([68]10) and for the immune suppressive
   functions of T[regs] ([69]11). Moreover, complex III is required for
   antigen-specific activation of cytotoxic T cells via ROS ([70]12) and
   for complex I stability and maintenance along the ETC ([71]13). The use
   of complex III–specific inhibitors, such as myxothiazol (MYX), has
   helped in deciphering the multifaceted nature of complex III. In
   addition, the small molecule Suppressor of site III[Qo] Electron Leak
   (S3QEL) has been developed to stop electron leak from the ETC to
   oxygen, thus blocking ROS production specifically from complex III
   ([72]14, [73]15).

   Here, we have examined the role of complex III in macrophages using
   S3QEL 1.2 and MYX. We have found that inhibition of complex III
   activity and subsequent ROS generation lead to a decrease in IL-10
   production in vitro and in vivo and that the down-regulation of the
   transcription factor activator protein 1 (AP1) might contribute to this
   event. Moreover, we show that S3QEL 1.2 treatment decreases melanoma
   progression and improves the survival of mice in vivo, impairing the
   production of IL-10 by macrophages in the tumor microenvironment (TME).
   Our results indicate a role for ROS from complex III in the induction
   of IL-10 in macrophages, the limiting of which could have utility in
   the promotion of antitumor immunity.

RESULTS

S3QEL 1.2 and MYX block ROS production by affecting OXPHOS in macrophages

   The previous literature has described S3QEL 1.2 ([74]Fig. 1A),
   developed by Brand and colleagues ([75]15), as a specific inhibitor of
   complex III–derived ROS production ([76]15), but its effect in immune
   cells has not been established. Here, we sought to investigate the
   effects of S3QEL 1.2 on activated bone marrow–derived macrophages
   (BMDMs) and to compare it to MYX, a specific inhibitor for complex III
   activity. First, we confirmed that LPS increased ROS production in
   macrophages by measuring H[2]O[2] release and demonstrated that S3QEL
   1.2 and MYX inhibited this response ([77]Fig. 1B) The same result was
   obtained with the mitochondria-targeted antioxidant MitoTEMPO at 5 mM,
   which decreased LPS-induced H[2]O[2] production ([78]Fig. 1C). We then
   performed metabolic flux analysis to assess whether complex III
   inhibitors were perturbing respiration of macrophages. As shown in
   [79]Fig. 1 (D to F), S3QEL 1.2 was marginally inhibitory against OXPHOS
   in macrophages at 10 μM with 500 nM MYX abolishing respiration. LPS
   inhibited OXPHOS as shown previously ([80]16). To confirm this, we
   performed metabolic flux analysis with different concentrations of
   S3QEL 1.2, MYX, and AA, another complex III inhibitor, injected into
   BMDMs in real time. Also here, S3QEL 1.2 did not affect respiration at
   any concentration used ([81]Fig. 1G), while MYX and AA impaired it
   ([82]Fig. 1, H and I). Moreover, S3QEL 1.2 did not change the activity
   of complex III, while MYX completely disrupted it ([83]Fig. 1, J and
   K), as shown in a previous study ([84]17). The two inhibitors were also
   not toxic to the cells, as shown by
   3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium (MTT) assay to
   examine the viability of cells (fig. S1A), and did not change the
   protein abundance of the ETC complexes (fig. S1B), analyzed by Western
   blotting. S3QEL 1.2 and MYX also slightly boosted glycolysis (fig. S1,
   C and D).

Fig. 1. Characterization of complex III inhibitors in macrophages.

   [85]Fig. 1.
   [86]Open in a new tab

   (A) S3QEL 1.2 chemical structure. (B and C) Rate of relative resorufin
   fluorescence units (rRFUs) directly proportional to H[2]O[2] production
   in the presence of DMSO, S3QEL 1.2 (10 μM), and MYX (500 nM) (B) or
   DMSO, MitoTempo (5 mM), and LPS (100 ng/ml) (C) for 48 hours in BMDMs.
   NS, not stimulated. (D to F) BMDMs were pretreated with DMSO, MYX (500
   nM), or S3QEL 1.2 (10 μM) for 3 hours followed by LPS (100 ng/ml) for
   24 hours. Mito Stress Test showing OCR of cells after addition of
   oligo, FCCP, and rotenone (Rot)/AA (D). Maximal respiration (E) and
   spare respiratory capacity (F) calculated from the OCR values. (G to I)
   Mito Stress Test showing OCR of cells after treatment with S3QEL 1.2
   (0.1, 1, or 10 μM) (G), MYX (50, 500, or 5000 nM) (H), or AA (50, 500,
   or 5000 nM) (I) and addition of oligo, FCCP, and rotenone/AA. UNTR,
   untreated. (J to K) BMDMs were pretreated with DMSO, MYX (500 nM), or
   S3QEL 1.2 (10 μM) for 3 hours followed by LPS (100 ng/ml) for 24 hours.
   Metabolic flux analysis showing complex III (CIII)–specific OCR of
   cells over time after addition of ADP + duroquinol, oligo, and AA. (K)
   Quantification of complex III–specific OCR. Data in (B) are means ± SEM
   for n = 3 from three independent experiments. Data in (C) are means ±
   SEM for n = 5 from five independent experiments. Data from (D) to (F)
   are means ± SEM from three independent experiments, with n = 7 to 12
   technical replicates for each condition. Data from (G) to (I) are means
   ± SEM from two independent experiments, with n = 9 technical replicates
   for each condition. Data from (J) to (K) are means ± SEM from three
   independent experiments, with n = 12 technical replicates for each
   condition. P values were calculated using one-way ANOVA for multiple
   comparisons. Differences were statistically significant at **P < 0.01,
   ***P < 0.001, and ****P < 0.0001.

Complex III inhibition provokes specific transcriptional changes in activated
macrophages

   We then performed RNA sequencing (RNA-seq) analysis on BMDMs treated
   with S3QEL 1.2 and MYX and activated with LPS to investigate the
   contribution of complex III to the transcriptomic profile of
   macrophages. Differential analysis identified 146 genes, which were
   commonly impaired by both compounds in LPS-treated macrophages
   ([87]Fig. 2A). Particularly, genes coding for protein that are
   essential for inflammatory macrophage function, such as cluster of
   differentiation 14 (CD14), the transcription factors Ets2 and Fosl2,
   the receptors for interferon-α/β IFNAR1 and IFNAR2, the ROS-activated
   Fgr kinase, NFAT-activating protein with ITAM motif 1 (Nfam1), IL4r,
   and suppressor of cytokine signaling 3 (Socs3), showed lowered
   expression, as indicated by arrows ([88]Fig. 2B). Inhibition of complex
   III by MYX and S3QEL 1.2 boosted certain genes, including that encoding
   growth differentiation factor 15 (GDF15) (fig. S2, A and B). We focused
   our attention on the cytokine profile of activated macrophages and
   performed immune-responsive enrichment analysis (IREA) using our
   RNA-seq dataset. IREA is an approach based on a dictionary of
   single-cell transcriptomic profiles of 17 immune cell types in response
   to 86 different cytokines in vivo (publicly available data) ([89]18).
   The analysis revealed that IL-10 was among the cytokines whose
   down-regulation by S3QEL 1.2 and MYX after LPS stimulation was
   prominent ([90]Fig. 2C). In addition, gene set enrichment analysis
   (GSEA) revealed that genes involved in IL-10 receptor signaling target
   genes were among the 20 most down-regulated terms ([91]Fig. 1D) and
   genes contributing to tumor necrosis factor–α (TNF-α) and nuclear
   factor κB (NF-κB) signaling were among the 20 most up-regulated terms
   ([92]Fig. 1E). IL-10 is well known to limit TNF-α production ([93]19),
   hence this regulation found in our results.

Fig. 2. Complex III inhibition reveals a specific transcriptional signature
in activated macrophages.

   [94]Fig. 2.
   [95]Open in a new tab

   (A to E) RNA sequencing of LPS-primed mouse macrophages (BMDMs) and
   pretreated (3 hours) with DMSO, S3QEL 1.2 (10 μM), and MYX (500 nM).
   (A) Venn diagram of commonly differentially expressed genes (146)
   between S3QEL 1.2 + LPS versus LPS and MYX + LPS versus LPS [adjusted P
   (P[adj]) < 0.1]. (B) Heatmap of the top 100 commonly up-regulated genes
   by LPS alone (red column) and down-regulated by both MYX and S3QEL 1.2
   after LPS priming (second and third columns; in blue). hrs, hours. (C)
   IREA of differentially down-regulated genes by S3QEL 1.2 (top) and MYX
   (down) after LPS stimulation [adjusted P (P[adj]) < 0.05]. (D) and (E)
   GSEA of differentially regulated genes. (D) Overrepresentation analysis
   of RNA-seq data of BMDMs pretreated with S3QEL 1.2 or MYX (n = 3 from
   one independent experiment; LPS, 4 hours). Reactome pathway enrichment
   analysis of differentially down-regulated genes. (E) Overrepresentation
   analysis of RNA-seq data of BMDMs pretreated with S3QEL 1.2 or MYX (n =
   3 from one independent experiment; LPS, 4 hours). Reactome pathway
   enrichment analysis of differentially up-regulated genes.

The inhibition of complex III decreases IL-10 production in activated
macrophages

   Because of the prominence of IL-10 in our analysis and the reported
   role of ROS in its regulation, we focused our attention on IL-10. We
   analyzed IL-10 transcripts by reverse transcription quantitative
   polymerase chain reaction (RT-qPCR) and released form by enzyme-linked
   immunosorbent assay (ELISA), and we found that S3QEL 1.2 decreased
   LPS-induced IL-10 production in a dose-dependent manner ([96]Fig. 3, A
   and B) and increased TNF-α ([97]Fig. 3, C and D), as we had observed in
   the RNA-seq analysis. MYX had the same effects on IL-10 mRNA and
   protein expression ([98]Fig. 3, E and F) but curiously did not increase
   TNF-α ([99]Fig. 3, G and H). We confirmed the redox-sensitive
   regulation of IL-10 with the pretreatment of macrophages with the
   mitochondria-targeted antioxidants MitoTEMPO and MitoQ, which, in the
   case of MitoQ, lowered IL-10 expression in response to LPS (fig. S3, A
   to D). We also tested the glutathione precursor N-acetyl cysteine
   (NAC), which impaired LPS-induced IL-10 expression especially at 1, 5,
   and 10 mM (fig. S3, E and F). Moreover, IL-10 regulation appeared to be
   complex I independent because pretreatment of activated BMDMs with
   rotenone, an inhibitor of complex I, and S1QEL 1.1, which impairs ROS
   production from complex I ([100]20), did not substantially decrease
   IL-10 mRNA or protein production (fig. S3, A, B, G, and H).

Fig. 3. The inhibition of complex III decreases LPS-induced IL-10 expression
in macrophages.

   [101]Fig. 3.
   [102]Open in a new tab

   (A to D) BMDMs were pretreated with DMSO or S3QEL 1.2 (0.1 to 10 μM)
   for 3 hours before LPS (100 ng/ml) stimulation for 4 hours, and cell
   lysates and supernatants were harvested. (A) and (B) Quantification of
   IL-10 mRNA (A) and released protein (B) by RT-qPCR, relative to the
   Rps18 housekeeping gene, and ELISA. (C) and (D) Quantification of TNF-α
   mRNA (C) and released protein (D) by RT-qPCR, relative to the Rps18
   housekeeping gene, and ELISA. FC, fold change. (E to H) BMDMs were
   pretreated with DMSO and MYX (500 nM) for 3 hours before LPS (100
   ng/ml) stimulation for 4 hours, and cell lysates and supernatants were
   harvested. (E) and (F) Quantification of IL-10 mRNA (E) and released
   protein (F) by RT-qPCR, relative to the Rps18 housekeeping gene, and
   ELISA. (G) and (H) Quantification of TNF-α mRNA (G) and released
   protein (H) by RT-qPCR, relative to the Rps18 housekeeping gene, and
   ELISA. (I) Schematic diagram of in vivo procedure and sample
   collection. Mice were intraperitoneally (ip) injected with PBS or S3QEL
   1.2 (1 mg/kg) for 2 hours, followed by LPS (2.5 mg/kg) for 2 hours.
   Blood (then serum) was collected. hrs, hours. (J) Quantification of
   IL-10 by ELISA in the serum. Data from (A) to (H) are expressed as
   means ± SEM for n = 5 to 9 from three independent experiments. Data
   from (J) are expressed as means ± SEM (n = 10 per group within two
   independent in vivo experiments). P values were calculated using
   one-way ANOVA for multiple comparisons. Differences were considered
   statistically significant at *P < 0.05, **P < 0.01, ***P < 0.001, and
   ****P < 0.0001.

   We next tested S3QEL 1.2 in vivo. S3QEL 1.2 has been used in a few
   studies where it was administered to mice in drinking water ([103]21).
   Here, mice were injected intraperitoneally with S3QEL 1.2 and then
   injected with LPS after 2 hours ([104]Fig. 3I). LPS-induced IL-10 in
   the serum was inhibited in the S3QEL 1.2–treated animals ([105]Fig.
   3J). Also, TNF-α was increased in serum after LPS injection, but S3QEL
   1.2 did not have an inhibitory effect on it (fig. S4A).

   We also analyzed S3QEL 1.2 effects on IL-10 produced by T[H]1
   lymphocytes. S3QEL 1.2 decreased IL-10 release after T[H]1 cell
   polarization (fig. S5, A and B). Furthermore, MYX also suppressed IL10
   expression and release in LPS-stimulated macrophages derived from human
   peripheral blood mononuclear cells (PBMCs) (fig. S6, A and C). S3QEL
   1.2 showed a trend toward a decrease in IL-10 at 0.1 and 0.5 μM (fig.
   S6, B and D). These results indicate that complex III–regulated IL-10
   production is also evident in human cells.

IL-10 suppression depends on inhibition of the AP1 subunit Fos

   From the RNA-seq dataset previously analyzed, we performed
   transcription factor enrichment analysis, which revealed that the
   expression of Fosl1 and Fosl2, canonical subunits of the transcription
   factor AP1, was inhibited by S3QEL 1.2 and MYX ([106]Fig. 4A).
   Furthermore, we found that many of the genes that were down-regulated
   by S3QEL 1.2 and MYX were Fos-dependent target genes ([107]Fig. 4B). We
   next analyzed the transcriptional levels of c-Fos by RT-qPCR, and we
   found that S3QEL 1.2 and MYX also diminished c-Fos levels after 30 min
   of LPS stimulation ([108]Fig. 4, C and D). We also analyzed c-Jun,
   which is another subunit of AP1 and works with c-Fos to form a dimer,
   moves to the nucleus, and binds to the DNA. c-Jun mRNA expression after
   30 min of LPS stimulation was decreased upon S3QEL 1.2 (fig. S7A) and
   MYX (fig. S7B) pretreatment. Subsequent confocal microscopy, performed
   to assess c-Fos activation and subsequent nuclear translocation,
   confirmed that the pretreatment with S3QEL 1.2 and MYX diminished the
   LPS-induced presence of phospho-c-Fos in the nucleus ([109]Fig. 4, E
   and F). Together, these results suggest that the imperfect electron
   transport–derived ROS perturb AP1, which could likely block IL-10
   production by LPS-activated macrophages. Since AP1 is being inhibited
   by S3QEL 1.2, and since IL-10 is an AP1-dependent gene, this may be
   part of the mechanism of IL-10 repression, particularly as AP1 is known
   to be ROS sensitive ([110]22).

Fig. 4. Complex III–dependent IL-10 regulation is mediated by the c-Fos
subunit of the AP1 transcription factor.

   [111]Fig. 4.
   [112]Open in a new tab

   (A and B) Overrepresentation analysis of RNA-seq data of BMDMs
   pretreated with SQ3EL 1.2 or MYX (data are from n = 3 from one
   independent experiment; LPS, 4 hours). (A) EnrichR transcription factor
   enrichment analysis of differentially down-regulated genes. (B) Heatmap
   of significantly down-regulated AP1 target genes with adjusted P
   (P[adj]) < 0.05. (C and D) BMDMs were pretreated with DMSO, MYX (500
   nM), or S3QEL 1.2 (10 μM) for 1 hour before LPS (100 ng/ml) stimulation
   for 0 to 120 min, and cell lysates were harvested. Quantification of
   c-Fos mRNA by RT-qPCR, relative to the Rps18 housekeeping gene, in
   BMDMs pretreated with S3QEL 1.2 (C) or MYX (D) and stimulated with LPS.
   (E and F) Confocal microscopy images (E) and quantification (F) of
   BMDMs pretreated with DMSO, MYX (500 nM), or S3QEL 1.2 (10 μM) for 1
   hour before LPS (100 ng/ml) stimulation for 2 hours. After that, cells
   were fixed and stained for phospho-c-Fos to assess cellular
   localization. Merge depicts phospho-c-Fos (red) together with nuclear
   staining (DAPI). Data in (C) and (D) are means ± SEM for n = 5 from
   three independent experiments. Data in (F) are means ± SD from one of
   the experiments in (E), showing the activated cells giving a
   fluorescence signal. Confocal images are representative of two
   independent experiments. P values were calculated using one-way ANOVA
   for multiple comparisons. Differences were considered statistically
   significant at **P < 0.01, ***P < 0.001, and ****P < 0.0001.

S3QEL 1.2 impairs tumor growth and elicit an immune cell response in mice in
a B16F10 in vivo model

   IL-10 has been shown to be immunosuppressive in several tumor models
   and promoting tumor survival, notably in the B16F10 melanoma model
   ([113]23). We therefore tested S3QEL 1.2 in this model alone (by
   intraperitoneal injection) or in combination with the immunostimulatory
   agent CpG ([114]24). We first confirmed that similar to LPS, S3QEL 1.2
   would inhibit IL-10 induction by CpG DNA in BMDMs, as shown in
   [115]Fig. 5A. MYX was also inhibitory ([116]Fig. 5A). The B16F10 model
   involved treating tumor-bearing mice with CpG DNA intratumorally and
   repeated intraperitoneal injections of S3QEL 1.2 ([117]Fig. 5B). As
   shown in [118]Fig. 5C, mice treated with S3QEL 1.2 or CpG alone limited
   tumor growth but had the most marked effect when given in combination.
   Similarly, as shown in [119]Fig. 5D, both agents promoted survival with
   the combination, resulting in the longest survival time. Last, we
   analyzed by flow cytometry the immune cell subsets infiltrating the
   tumor [the gating strategy for myeloid cell and non–myeloid cell
   subsets is shown in fig. S8 (A to D)]. Tumor burden was decreased in
   the S3QEL 1.2 alone–treated group ([120]Fig. 5E). Cell counts and tumor
   mass were also lowered (fig. S9, A to C). M1-like macrophages
   infiltrating the tumor showed increased expression of major
   histocompatibility complex class II (MHCII) molecules ([121]Fig. 5F),
   relevant for the presentation of tumor antigens to CD4^+ T cells, after
   S3QEL 1.2 administration while M2-like macrophages presence showed a
   moderate decrease in the same group ([122]Fig. 5G). IL-10 production by
   tumor-infiltrating macrophages was impaired by S3QEL 1.2 ([123]Fig.
   5H). The number of tumor-infiltrated leukocytes was increased after
   S3QEL 1.2 administration (fig. S9D), while the number of myeloid cells,
   neutrophils, and monocyte-macrophages remained the same (fig. S9, E to
   G). The amounts of dendritic cells were lower in the S3QEL 1.2–treated
   group (fig. S9H). This last subset of cells did not show any augmented
   expression of MHCII molecules in mice treated with S3QEL 1.2 (fig.
   S9I). Last, we also investigated the effect of S3QEL 1.2 on B16F10
   melanoma cells in vitro. The highest concentration of S3QEL 1.2
   impaired the proliferation of B16F10 cells but did not block OXPHOS and
   complex III–specific activity, in contrast to MYX, which impaired all
   these processes (fig. S10, A to F), as assessed by metabolic flux
   analysis and crystal violet assay. The effect of S3QEL 1.2 on tumor
   growth may therefore be a combination of a dual effect on tumor cell
   proliferation and inhibition of IL-10 to promote antitumor immunity.

Fig. 5. S3QEL 1.2 administration ameliorates survival and inhibits melanoma
progression in mice in vivo.

   [124]Fig. 5.
   [125]Open in a new tab

   (A) BMDMs were pretreated with DMSO, MYX (500 nM), or S3QEL 1.2 (10 μM)
   for 3 hours, stimulated with CpG (1 μg/ml) for 4 hours, and cell
   lysates and supernatants were harvested. Quantification of IL-10 mRNA
   by RT-qPCR, relative to the Rps18 housekeeping gene, and of IL-10
   levels by ELISA. (B) Schematic diagram of the melanoma in vivo model.
   Mice were challenged with 5 × 10^6 B16F10 cells subcutaneously and
   administered with PBS, S3QEL 1.2 (1 mg/kg), and/or CpG (2.5 mg/kg).
   Tumors were measured daily. D, day. (C) Mean initial tumor growth curve
   relative to tumor volume from day −1 to day 8. (D) Kaplan-Meier
   survival graph up to day 65 (end of experiment). (E) Tumor growth rate
   curve relative to tumor area from day 0 to day 17. (F to H) Immune
   cells infiltrating the tumor were analyzed by flow cytometry. Cells
   expressing the surface markers Ly6C, F4/80, and CD11b showing the
   percentage of MHCII^+ macrophages and mean fluorescence intensity (MFI)
   of MHCII (F), the percentage of CD206^+ macrophages and MFI of CD206
   (G), and the percentage of IL-10^+ macrophages and MFI of IL-10 (H).
   Data from (A) are means ± SEM for n = 5 from three independent
   experiments. Data from (C) and (D) are means ± SEM for n = 5 for the
   PBS group, n = 6 for the S3QEL 1.2 group, n = 7 for the CpG group, and
   n = 6 for the S3QEL 1.2 + CpG group. Statistical significance for
   survival analysis was determined by Mantel-Cox test. Statistical
   significance for tumor growth analysis was determined by two-way ANOVA;
   asterisks are for P < 0.05 in the group’s comparison. Data from (E) to
   (H) are means ± SD for n = 9 from one independent in vivo experiment. P
   values were calculated using two-tailed unpaired Student’s t test.
   Differences were statistically significant at *P < 0.05, **P < 0.01,
   ***P < 0.001, and ****P < 0.0001.

DISCUSSION

   Mitochondrial complex III and its involvement in regulating immune
   function have only recently begun to be examined. Mice lacking the
   Rieske iron sulfur protein subunit of complex III showed the diminished
   suppressive capacity of T[regs], together with decreased expression of
   markers of T[regs] such as EOS and CTLA-4, without changing their Foxp3
   and HELIOS expression ([126]11). In another study, complex III knockout
   mice had lower numbers of CD4^+ and CD8^+ T cells in the periphery,
   with less IL-2 expression ([127]12).

   Here, we have deployed S3QEL 1.2, a highly selective suppressor of
   complex III–derived ROS, which has been developed by Brand and
   colleagues ([128]15). S3QELs were identified through an Amplex
   UltraRed–based detection system used to screen 635,000 small molecules
   against H[2]O[2] production caused by electron leak at sites III[Qo],
   I[Q], and II[F] in isolated muscle mitochondria. After the elimination
   of compounds that were unselective for site III[Qo] or that inhibited
   energy metabolism, S3QELs were confirmed to potently and selectively
   suppress site III[Qo] superoxide production without impairing any
   bioenergetic function, including mitochondrial membrane potential
   (ΔΨ[m]) ([129]15). S3QEL 1.2 has been shown to prevent the accumulation
   of ROS produced by complex III and IL-6 production from alveolar
   macrophages exposed to particulate matter ([130]25). In addition, in
   another study involving BMDMs activated by LPS and extracellular ATP,
   S3QEL 1.2 did not affect IL-1β production, suggesting a mechanism for
   NLRP3 inflammasome activation that is independent of complex
   III–derived ROS ([131]26). However, a role for complex III–derived ROS
   in the production of other key cytokines from macrophages was not
   investigated. IL-10 appeared to be strongly affected by complex III
   inhibition in our data, and increasing literature is highlighting its
   dependency on mitochondrial ETC, although precise details are lacking
   ([132]27–[133]29).

   Our findings show that complex III sustains IL-10 generation through
   intact respiration and production of ROS. Inhibition of complex II has
   been shown to block IL-10 production ([134]27), possibly because of
   electron flow to complex III, preventing it from generating ROS. In
   addition, our bioinformatics analysis suggests that complex III can
   regulate the cellular antioxidant response and drive cytokine
   production via a plethora of pathways. From our findings, it is clear
   that complex III impairment up-regulates glycolysis as a possible
   compensatory mechanism to fuel ATP production in the absence of
   O[2]-dependent respiration.

   The mechanism that we propose for the decrease in IL-10 expression by
   complex III impairment is down-regulation of the transcription factor
   AP1 and, especially, its subunit c-Fos (fig. S11). We hypothesize that
   ROS produced by complex III would oxidize specific cysteine residues on
   AP1 or its upstream regulators and subsequently activate them. The
   redox-sensitive nature of c-Fos has already been described by Wang and
   colleagues, who showed that the mitochondria-targeted antioxidant
   mitoTEMPO decreased the oxLDL-mediated abundance of c-Fos and
   phosphorylated c-Fos in smooth muscle cells ([135]30). In another
   study, mitoTEMPO abolished the binding of AP1 to DNA in endothelial
   cells activated by lysophosphatidylcholine ([136]31). As c-Fos and
   c-Jun transcripts are lowered by S3QEL 1.2 and MYX, we do not exclude
   the possibility of the regulation of proteins that control the
   stability of mRNAs here ([137]32). However, further investigation is
   necessary to understand the exact mechanism for c-Fos regulation by
   ROS. Peroxiredoxins, which are antioxidant enzymes responsible for the
   maintenance of redox balance within the cell, and particularly
   peroxiredoxin-6, have been detected as interactors of AP1 in mice
   ([138]33).

   IL-10 overexpression is a key feature of aggressive and highly
   proliferative tumors, such as melanoma and colorectal cancer ([139]34).
   Because tumor-associated macrophages (TAMs), together with
   myeloid-derived suppressor cells (MDSCs) and T[regs] and B cells, are
   the main sources of IL-10 in the tumor microenvironment and contribute
   to the abolishment of the cytotoxic T cell response ([140]35), we
   hypothesized that S3QEL 1.2 could have a beneficial effect through the
   dampening of tumor-driven immunotolerance mediated by IL-10. We
   therefore tested S3QEL 1.2 in the B16F10 melanoma in vivo model, where
   neutralization of the IL-10 receptor with an antibody has been shown to
   boost antitumor immunity, especially in combination with CpG DNA
   ([141]36). The beneficial effect of S3QEL 1.2 here is likely to be via
   inhibition of IL-10 production by macrophages. However, other
   pro-inflammatory cytokines and chemokines involved in the cross-talk
   between tumor and immune cells are differentially regulated after
   complex III impairment. For example, CXCL9 is up-regulated by S3QEL 1.2
   pretreatment (fig. S2B) and it has been shown to define macrophage
   polarization, thus an antitumor immune response, in human colorectal
   cancers ([142]37). In addition, IL-7 and IL-15 are other two cytokines
   that are down-regulated after complex III impairment ([143]Fig. 2C).
   Previous findings have described IL-7 as fundamental for the activation
   of myeloid cell populations in mice bearing CT26 tumors ([144]38) and
   for tumor invasiveness in prostate cancer ([145]39), while
   IL-15–dependent signaling in tumor-associated macrophages inhibited
   CD8^+ T cell recruitment in breast cancer ([146]40). Taking into
   consideration the complexity of the tumor microenvironment, we cannot
   rule out possible effects of S3QEL 1.2 on other cell types and we
   believe that S3QEL 1.2 might also inhibit proliferation of the B16F10
   melanoma cells themselves. Therefore, the treatment of macrophages
   before in vivo injection with tumor cells, as previously deployed
   ([147]41), or the use of IL-10 conditional knockout models may
   represent other options required to further elucidate the specific
   mechanism of action and target of S3QEL 1.2.

   In conclusion, our work provides evidence for complex III–derived ROS
   in regulating IL-10 production in macrophages, broadening the existing
   knowledge about ETC-mediated rewiring of myeloid cells ([148]42).
   Moreover, S3QEL 1.2 could be a useful tool for dissecting
   immunotolerance in the context of tumor development and progression.
   Recently, a complex I inhibitor was tested in acute myeloid leukemia
   and advanced solid tumors but proved to be neurotoxic ([149]43). The
   advantage of S3QEL 1.2 is its limited effect on respiration, and
   therefore, it may be less toxic in the treatment of cancer.

MATERIALS AND METHODS

Animal details

   All mice used for the in vitro experiments were on a C57BL/6JOlaHsd
   background, unless stated below. Wild-type mice, when needed, were
   purchased from Envigo. In vivo experiments were performed with
   8-week-old female mice bred under specific pathogen–free conditions,
   under the license and approval of the local animal research ethics
   committee (Health Products Regulatory Authority) under European Union
   regulations. The first two in vivo LPS models were performed with
   6-to-12-week-old C57BL/6JOlaHsd mice, and littermates were randomly
   assigned to experimental groups. All animal procedures were ethically
   approved by the Trinity College Dublin Animal Research Ethics Committee
   before experimentation and conformed to the directive 2010/63/EU of the
   European Parliament (authorization AE19136/P148).

   The first B16F10 melanoma in vivo model was performed with
   6-to-12-week-old C57BL/6J mice, which were randomly assigned to
   experimental groups. All animal procedures were ethically approved by
   the Trinity College Dublin Animal Research Ethics Committee before
   experimentation and conformed to the directive 2010/63/EU of the
   European Parliament (authorization AE19136/P172).

   The second B16F10 melanoma in vivo model was performed in the animal
   facilities of the Medical University of Graz. Wild-type C57BL/6J mice
   were purchased from Charles River Laboratories. Ethical approval for
   animal experimental protocols was granted by the Austrian Federal
   Ministry of Science and Research (protocol number:
   2022-0.748.851/2024-0.783.500).

B16F10 in vivo model

   B16F10 cells (American Type Culture Collection) were maintained in
   Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) containing 10% fetal
   bovine serum (FBS) and 1% penicillin-streptomycin at 37°C and 5% CO[2]
   in a humified incubator. Mice were injected subcutaneously with 3.5 ×
   10^5 B16F10 cells on the right flank, which has been previously shaved.
   Tumor growth was measured daily using a digital caliper, and the tumor
   volume was calculated as [Length × (Width)^2]/2. Upon reaching 30 to 60
   mm^3, mice were randomly assigned to experimental treatment groups and
   received intratumoral injections of phosphate-buffered saline (PBS) or
   CpG ODN 1826 (2.5 mg/kg each treatment), with treatments repeated 4
   days later on day 4 and on day 8 for a total of three intratumoral
   treatments. In one experiment, two groups of mice received S3QEL 1.2 (1
   mg/kg; 5% in PBS) via the intraperitoneal route every 2 days from day
   one before the first intratumoral treatment until day 19. Mice were
   culled upon reaching the predetermined humane end point of tumor
   diameter >15 mm or, in the case of ulceration, with survival analyzed
   using Kaplan-Meier analysis.

   In the second in vivo experiment, B16F10 cells (American Type Culture
   Collection) were maintained in DMEM (Gibco) containing 10% FBS and 1%
   penicillin-streptomycin at 37°C and 5% CO[2] in a humified incubator.
   Mice were injected subcutaneously with 5 × 10^6 B16F10 cells on the
   right flank, which has been previously shaved. Tumor growth was
   measured daily using a digital caliper, and the tumor volume was
   calculated as [Length × (Width)^2]/2. Upon reaching 30 to 60 mm^3, mice
   were randomly assigned to experimental treatment groups and received
   S3QEL 1.2 or PBS intraperitoneally. S3QEL 1.2 was dissolved in a
   vehicle containing dimethyl sulfoxide (DMSO):ethanol:Kolliphor EL
   (Sigma-Aldrich) at a 1:1:2 ratio and divided into aliquots; a vehicle
   control containing no drug was also prepared. Before injection, an
   aliquot of S3QEL 1.2 or vehicle in DMSO/ethanol/Kolliphor was added to
   pre-prepared 5% (w/v) glucose in water at a 1:6.66667 ratio. Two
   hundred microliters of this mixture per mouse was injected
   intraperitoneally every other day to give a final S3QEL 1.2
   concentration of 1 mg/kg per mouse.

Flow cytometry of tumor single-cell suspensions

   B16F10 tumors were removed, mechanically minced using surgical
   scissors, and subsequently digested with deoxyribonuclease I (160 U/ml;
   Worthington) and collagenase (4.5 U/ml; Worthington) for 30 min at
   37°C, with vortexing and shaking. Samples were then passed through a
   40-μm cell strainer, washed in PBS and 2% FBS, and resuspended in flow
   staining buffer (PBS containing 2% FBS and 2 mM EDTA). Cells were
   counted on the EVE automated cell counter (NanoEntek).

   A total of 3 × 10^6 cells per sample in 100 μl of staining buffer were
   used for flow cytometry staining. Single-cell suspensions were
   incubated for 20 min in fixable viability dye eFluor 780 (eBioscience)
   according to the manufacturer’s instructions. Fc receptor blocking was
   then performed using TruStain FcX (BioLegend) according to the
   manufacturer’s instructions. For surface staining, cells were incubated
   for 30 min at 4°C in the dark in flow staining buffer with the
   following antibodies (all purchased from BD/BioLegend): CD45-AF700
   (clone 30-F11, 560510), CD3-BV421 (clone 17A2, 564008), CD4-BUV496
   (clone GK1.5, 612952), CD8-PE-Cy7 (clone 53-6.7, 561097), NKp46-BV510
   (clone 29A1.4, 563455), CD19-BV711 (clone 1D3, 563157), CD11b-BUV737
   (clone M1/70, 612801), Ly6C-APC (clone AL21, 560595), Ly6G-PE/Dazzle594
   (clone 1A8, 127648), CD11c-BV605 (clone N418, 117333), F4/80-BUV395
   (clone T45-2342, 565614), SiglecF-PE (clone E50-2440, 552126),
   CD206-FITC (clone C068C2, 141704), and MHCII-PerCP-Cy5.5 (clone
   M5/114.15.2, 107626). Single stains and fluorescence minus one (FMO)
   controls were also used. Cells were then washed twice in flow staining
   buffer, fixed using IC Fixation Buffer (eBioscience) for 10 min, washed
   again twice, and resuspended in flow staining buffer.

   For intracellular staining of IL-10, 3 × 10^6 cells of each single-cell
   suspension sample were seeded into a 96-well U-bottom plate with RPMI
   containing GolgiStop (1.5 μl/ml, BD Biosciences) and incubated for 1
   hour in a 37°C cell culture incubator. Fixable viability dye staining,
   Fc receptor blocking, and surface antibody staining were performed as
   above. Cells were then permeabilized and fixed using Perm/Fix Buffer
   (BD Biosciences) for 30 min, followed by subsequent staining for IL-10
   using an IL-10-PE antibody (BioLegend) for 30 min. Single stains and
   FMO controls were also used. Cells were then washed twice in Perm/Fix
   buffer (prepared according to the manufacturer’s instructions; BD
   Biosciences), and resuspended in flow staining buffer.

   Samples were analyzed on a BD LSRFortessa flow cytometer using FACSDiva
   software (BD Biosciences). Analyses were performed using FlowJo version
   10.10 software, and FMO and single stains were used to compensate and
   define gates.

Flow cytometry of T[H]1 lymphocytes

   After activation, the cells were then washed and stained for 10 min at
   room temperature with fixable viability dye (Thermo Fisher Scientific)
   for dead cell exclusion and fluorochrome-labeled monoclonal antibodies
   against surface cell markers (CD4-PerCPCy5.5, clone RM4-4, 116011).
   Afterward, cells were washed and fixed for 15 min at room temperature
   using the eBioscience Foxp3 Fixation/Permeabilization Kit. After other
   washing steps, cells were permeabilized using the same kit and stained
   intracellularly with monoclonal antibodies (IFN-γ-APC, clone XMG1.2,
   505810; and IL-10-APCCy7, JES5-16E3, 505035). Data were acquired on
   FACSCanto II machines (BD Biosciences) and analyzed using FlowJo
   software (Tree Star), and unstained single-stained controls were used
   to compensate and define gates.

Generation of murine BMDMs

   Mice (6- to 12-week-old) were euthanized in a CO[2] chamber, and death
   was confirmed by cervical dislocation [procedures previously described
   in ([150]44, [151]45)]. Bone marrow was subsequently harvested from the
   tibia, femur, and ilium. For monocyte/macrophage isolation, cells were
   differentiated in DMEM containing L929 supernatant (20%), FBS (10%),
   and penicillin-streptomycin (1%) for 6 days, after which cells were
   counted and plated at 0.5 × 10^6 cells/ml in 10% L929 supernatant,
   unless otherwise stated.

Isolation of human PBMCs

   Human blood samples from healthy donors were collected and processed at
   the School of Biochemistry and Immunology at the Trinity Biomedical
   Sciences Institute (TCD). Blood samples were obtained anonymously, and
   written informed consent for the use of blood for research purposes was
   obtained from the donors. All the procedures involving experiments on
   human samples were approved by the School of Biochemistry and
   Immunology Research Ethics Committee (TCD). 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 of 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 layer of mononuclear cells at the
   plasma-density gradient medium interface was retained, and 20 ml of 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, counted, and plated at 1 × 10^6
   cells/ml in RPMI supplemented with FBS (10%) and
   penicillin-streptomycin (1%).

Generation of human macrophages

   PBMCs were obtained, and CD14^+ monocytes were isolated using a
   MagniSort Human CD14 Positive Selection kit (Thermo Fisher Scientific)
   according to the manufacturer’s protocol. CD14 monocytes were then
   differentiated in T-175 flasks in RPMI containing FBS (10%),
   penicillin-streptomycin (1%), and recombinant human macrophage
   colony-stimulating factor (1:10,000). After 6 days, the supernatant was
   discarded, cells were scraped and counted, and human monocyte-derived
   macrophages were plated in 12-well plates at 1 × 10^6 cells/ml in RPMI
   containing FBS (10%) and penicillin-streptomycin (1%).

In vitro T[H]1 cell differentiation

   Naive CD4^+CD25^− T cells were purified from spleens of wild-type
   C57BL/6OlaHsd mice with the naïve CD4 T cell isolation kit (Miltenyi
   Biotec) by using an AutoMACS magnetic cell sorter (Miltenyi Biotec)
   according to the manufacturer’s protocol. Purified cells were activated
   with anti-CD3ε:CD28 (3 and 2 μg/ml, respectively; BD Biosciences)
   coated on an F-bottomed plate (10^5 per well). Skewing conditions were
   rmIL-12 and rmIL-2 (both 20 ng/ml; R&D Systems) for T[H]1 cells for 72
   hours. When indicated, MYX (50, 500, or 5000 nM) and S3QEL 1.2 (0.1, 1,
   or 10 μM) were added to the polarization media. Subsequently, T[H]1
   cells were activated by phorbol 12-myristate 13-acetate/ionomycin
   stimulation (50 and 500 ng/ml, respectively) in the presence of
   StopGolgi solution (1.5 μl/ml, BD Biosciences) at 37°C in a humidified
   5% CO[2] chamber for 3 hours.

Reagents

   MYX was dissolved in DMSO, both purchased from Sigma-Aldrich. S3QEL 1.2
   and S1QEL 1.1 were provided by M. D. Brand (Buck Institute for Research
   on Aging, Novato, CA, US). LPS from Escherichia coli (ALX-581-010-L002)
   was purchased from Enzo Life Sciences for in vitro experiments. NAC and
   MitoTEMPO were purchased by Sigma-Aldrich. CpG ODN 1826
   (130-109-373/374) was purchased from Miltenyi Biotec for in vitro and
   in vivo experiments. LPS from E. coli (L4524) was purchased from
   Sigma-Aldrich for in vivo experiments. MitoTEMPO was purchased from
   Sigma-Aldrich (0737), and MitoQ was purchased from Abcam (ab285406) for
   in vitro expperiments. Rotenone (Sigma-Aldrich, 8875) and piericidin A
   (MedChemExpress, HY-114936) were also used for in vitro experiments.
   NAC was purchased from Merck (A7250) for in vitro experiments.

RNA isolation and real-time qPCR

   Total RNA was isolated using the PureLink RNA Mini Kit (Invitrogen) and
   quantified using a Nanodrop 2000 UV-visible spectrophotometer. cDNA was
   prepared using total RNA (20 to 100 ng/μl) by RT-PCR using a
   high-capacity cDNA reverse transcription kit (Applied Biosystems)
   according to the manufacturer’s instructions. Real-time qPCR was
   performed with an ABI 7500 Fast real-time PCR system (Applied
   Biosystems) on cDNA using SYBR Green (Invitrogen). Data were normalized
   to murine RPS18 as endogenous controls, and the mRNA expression fold
   change relative to controls was calculated using the 2^−ΔΔCt method.
   All fold changes are expressed normalized to the untreated control. The
   following murine primers were used: IL-1β, 5′-GAGGACATGAGCACCTTCTTT-3′,
   3′-GCCTGTAGTGCAGTTGTCTAA-5′; TNF-α, 5′-GCCTCTTCT-CATTCCTGCTT-3′,
   3′-TGGGAACTTCTCATCCCTTTG-5′; IL-10, 5′-AGGCGCTGTCATCGATTT-3′,
   3′-CACCTTGGTCTTGGAGCTTAT-5′; IL-6, 5′-CCACAGTCCTTCAGAGAGATACA-3′,
   3′-CCTTCTGTGACTCCAGCTTATC-5′; c-Fos, 5′-GGGAATGGTGAAGACCGTGTCA-3′,
   3′-GCAGCCATCTTATTCCGTTCCC-5′; c-Jun, 5′-CAGTCCAGCAATGGGCACATCA-3′,
   3′-GGAAGCGTGTTCTGGCTATGCA-5′; Actβ, 5′-GCCTTCCTTCTTGGGTATGG-3′, 3′-A
   GCACTGTGTTGGCATAGAG-5′; RPS18, 5′-GCGAGTACTCAACACCAACA-3′,
   3′-CCTCAACACCACATGAGCATA-5′. The following human primers were used:
   IL-10, 5′-GTCCTTGCTGGAGGACTTTA-3′, 3′-ATGTCTGGGTCTTGGTTCTC-5′; RPS13,
   5′-CGAAAGCATCTTGAGAGGAACA-3′, 3′-TCGAGCCAAACGGTGAATC-5′.

RNA sequencing

   BMDMs (three independent mice) were treated as indicated, and RNA was
   extracted as previously detailed. mRNA was extracted from total RNA
   using poly-T-oligo–attached magnetic beads. After fragmentation, the
   first-strand cDNA was synthesized using random hexamer primers,
   followed by second-strand cDNA synthesis using either deoxyuridine
   triphosphate for directional library or deoxythymidine triphosphate for
   nondirectional library. The library was checked with Qubit and
   real-time PCR for quantification and a bioanalyzer for size
   distribution detection. Quantified libraries were pooled and sequenced
   on the NovaSeq 6000 S4 (Illumina).

   Raw data (raw reads) in fastq format were first processed through
   in-house (Novogene) perl scripts. In this step, clean data (clean
   reads) were obtained by removing reads containing an adapter, reads
   containing ploy-N, and low-quality reads from raw data. At the same
   time, Q20, Q30, and GC content were calculated. All the downstream
   analyses were based on the clean data with high quality. The index of
   the reference genome was built usingHisat2 version 2.0.5, and
   paired-end clean reads were aligned to the reference genome using
   Hisat2 version 2.0.5.

   FeatureCounts version 1.5.0-p3 was used to count the read numbers
   mapped to each gene. Then, fragments per kilobase of transcript per
   million mapped reads of each gene were calculated on the basis of the
   length of the gene and read counts mapped to this gene.

   Differential expression was analyzed using the DESeq2R package
   (1.20.0). The resulting P values were adjusted using Benjamini and
   Hochberg’s approach for controlling the false discovery rate. Genes
   with an adjusted P ≤ 0.05 found by DESeq2 were assigned as
   differentially expressed. All the analyses and heatmaps were performed
   using custom scripts in R version 4.3.1. GSEA was performed using the
   online EnrichR tool.

ELISA

   A DuoSet ELISA kit for IL-10 and TNF-α (R&D Systems) in the supernatant
   of BMDMs was purchased and used according to the manufacturer’s
   instructions with appropriately diluted cell supernatants added to each
   plate in duplicate or triplicate. For in vivo samples, a Quantikine
   ELISA kit for mouse IL-10 (R&D Systems) was used according to the
   manufacturer’s instructions with appropriately diluted serum added to
   each plate. Absorbance was read at 450 nm and 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 on GraphPad Prism 10.1.1.

MTT assay

   BMDMs at a concentration of 1 × 10^5 cells were seeded in each well of
   a 96-well F-bottom plate in duplicates or triplicates and treated
   accordingly. MTT was purchased from Sigma-Aldrich and reconstituted in
   PBS to a concentration of 1 mg/ml. On each plate, three wells had no
   cells but the medium only to serve as a blank. At the end of the
   treatment, cells were washed with 100 μl per well and 100 μl of MTT
   solution was added in each well. The plate was incubated for 2 hours at
   37°C wrapped in tin foil. At the end of the incubation, the MTT
   solution was discarded, and the crystals were dissolved in 100 μl of
   DMSO solution in each well. The plate was incubated for 20 min at 37°C
   in the dark and read on a spectrophotometer at 595 nm of wavelength.
   Data were normalized to untreated and unstimulated cells.

Extracellular flux analysis

   A Seahorse XFe-96 extracellular flux analyzer (Seahorse Bioscience,
   Agilent Technologies) was used to define oxygen consumption rates
   (OCRs) and extracellular acidification rates. For monitoring OXPHOS,
   cells were plated at 1 × 10^5 cells per well in XF DMEM (pH 7.4)
   supplemented with pyruvate, glutamine, and glucose. Mitochondrial
   perturbation experiments were carried out by sequential addition of
   oligomycin (oligo), carbonyl cyanide 4-(trifluoromethoxy)
   phenylhydrazone (FCCP), antimycin A (AA), and rotenone according to the
   manufacturer’s instructions. For monitoring glycolysis, cells were
   plated in XF DMEM (pH 7.4) supplemented with glutamine. Glycolysis
   perturbation experiments were carried out by sequential addition of
   glucose, oligo, and 2-deoxyglucose according to the manufacturer’s
   instructions. For monitoring the OCR of only intact mitochondria, cells
   were plated in MAS buffer (70 mM sucrose, 220 mM mannitol, 10 mM
   KH[2]PO[4], 5 mM MgCl[2], 2 mM Hepes, and 1 mM EGTA) and then treated
   with the XF plasma membrane permeabilizer (Seahorse, Agilent
   Technologies), followed by addition of 0.5 mM duroquinol, 1 mM
   adenosine 5′-diphosphate (ADP), 15 μM oligo, and 0.02 mM AA for
   monitoring complex III–driven respiration. In this case, the assay
   includes a cycle of mixing for 2′, waiting for 2′, and measuring for 3′
   (repeated three times); duroquinol and ADP injections, followed by the
   same cycle repeated two times; then oligomycin injection, followed by
   the same cycle repeated two times; and AA injection, followed by the
   same cycle repeated two times. Changes in OCR on substrate addition
   were calculated relative to the preinjection rate.

Western blotting

   The supernatant was removed from cells after stimulation, and lysates
   were harvested in 30 to 50 μl of lysis buffer [0.125 M Tris (pH 6.8),
   10% glycerol, 0.02% SDS, and 5% dithiothreitol]. Lysates were
   subsequently heated to 95°C for 5 min to denature proteins.
   SDS–polyacrylamide gel electrophoresis was used to resolve proteins by
   molecular weight. Samples were boiled at 95°C for 5 min before loading
   into a 5% stacking gel. The percentage resolving gel (8 to 12%)
   depended on the molecular weight of the given protein. The Bio-Rad gel
   running system was used to resolve proteins, and the Bio-Rad wet
   transfer system was used for the electrophoretic transfer of proteins
   onto the polyvinylidene difluoride membrane. After transfer, the
   membrane was incubated in blocking solution [5% milk powder or 5%
   bovine serum albumin (BSA) in Tris-buffered saline with Tween 20
   (TBST)] for 1 hour and subsequently incubated in a primary antibody (5%
   milk powder or 5% BSA in TBST) rolling overnight at 4°C. The membrane
   was incubated for 1 hour with a secondary antibody (diluted in 5% milk
   powder or 5% BSA in TBST) at room temperature. Before visualization,
   the membrane was immersed in WesternBright ECL Spray (Advansta).
   Proteins were visualized on a ChemiDoc MPTM Imaging System (Bio-Rad),
   and both chemiluminescent and white light images were taken.

   Quantification of Western blot images was performed using Image Lab
   Software (Bio-Rad). The adjusted band volume was calculated for each
   band, and for each experimental condition, this was presented as the
   target protein/housekeeping protein.

   The following antibodies were used: antimouse c-Fos (2250),
   phospho-c-Fos (5348), c-Jun (9165), phospho-c-Jun (3270), GAPDH
   (glyceraldehyde-3-phosphate dehydrogenase; 2118), and α-tubulin (2144);
   they were purchased from Cell Signaling Technology. Total OXPHOS rodent
   Western blot antibody cocktail was purchased from Abcam (ab110413).
   Antimouse β-actin (A5316) was purchased from Sigma-Aldrich.

Immunofluorescence and confocal microscopy

   Cells (2.5 × 10^5) were plated on 12-mm coverslips in 24-well plates.
   Cells were treated as required and washed three times with warm PBS.
   Cells were subsequently fixed for 10 min with 2% paraformaldehyde in
   PBS at 37°C. Cells were washed three times with PBS and permeabilized
   for 1 hour in block solution [1% BSA, glycine (22.52 mg/ml), and 0.1%
   Tween 20 in PBS]. Anti-phospho-c-Fos antibody (Cell Signaling
   Technology, 5348) was diluted 1:800 in block solution and incubated
   with cells overnight at 4°C. Cells were washed three times with PBS for
   5 min per wash. A mix containing Alexa 647–conjugated goat antirabbit
   IgG (H + L) antibody (1:1000) and DAPI (4′,6-diamidino-2-phenylindole;
   1:1000, Thermo Fisher Scientific, 62248) was subsequently added to
   cells for 90 min at room temperature in the dark. Cells were
   subsequently washed three times with PBS for 5 min per wash. Coverslips
   were mounted onto microscope slides using 10 to 20 μl of ProLong Gold
   antifade reagent (Thermo Fisher Scientific, [152]P36930). Slides were
   imaged using a Leica SP8 scanning confocal microscope with a ×40.0
   objective. Images were analyzed using 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 of phospho-c-Fos signal intensity was
   performed using Imaris software 10.1.1. The mean signal intensity was
   calculated for individual cells in single-color images and displayed
   relative to the signal intensity of control cells.

Amplex Red H[2]O[2] measurement

   BMDMs were seeded in a 96-well plate with black opaque inserts with a
   concentration of 1 × 10^5 cells per well in duplicates. Cells were
   treated accordingly and stimulated with LPS for 48 hours (100 ng/ml)
   and with MitoTEMPO for 48 hours (5 mM). Cells were treated with S3QEL
   1.2 (10 μM) and MYX (500νM) on the day of the assay, and media were
   changed to KRPG Buffer. The H[2]O[2] standard was prepared as per the
   manufacturer’s instructions. Fifty microliters of 100 μM Amplex Red and
   HRP working solution (0.2 U/ml) were added to the wells of the H[2]O[2]
   standard. One hundred microliters of 50 μM Amplex Red and HRP working
   solution (0.1 U/ml) were added to the wells containing the samples.
   Fluorescence with excitation between 450 and 490 nm and emission of 590
   nm was read at the beginning of the reaction, after 30 min, and then
   every hour for maximum 3 hours. The fluorescence of the samples was
   extrapolated from the H[2]O[2] standard curve and normalized to the
   untreated sample.

LPS-induced model of inflammation

   C57BL/6OlaHsd mice were pretreated intraperitoneally with S3QEL 1.2 or
   DMSO vehicle control at the final concentration of 1 mg/kg (1:10 in
   PBS) for 2 hours and then injected intraperitoneally with LPS or PBS
   vehicle control at the final concentration of 2.5 mg/kg for 2 hours.
   After 4 hours, mice were euthanized and blood was collected
   retro-orbitally. Peritoneal exudate cells were harvested in PBS. Blood
   was centrifuged at 4°C for 10 min at 5000 rcf, and then the supernatant
   (serum) was kept at −80°C for further analysis by ELISA. Peritoneal
   exudate cells were pelleted and lysed in RNA lysis buffer (Invitrogen)
   for further analysis by qPCR.

Quantification and statistical analysis

   Details of all statistical analyses performed can be found in the
   figure legends. Data were expressed as means ± SEM, unless stated
   otherwise. P values were calculated using two-tailed Student’s t test
   for pairwise comparison of variables and one-way analysis of variance
   (ANOVA) for multiple comparisons of variables. A Sidak’s multiple
   comparisons test was used as a post test when performing an ANOVA. A
   Mantel-Cox test was used for log-rank analysis of the survival
   Kaplan-Meier curve. A confidence interval of 95% was used for all
   statistical tests. For statistical significance, exact P values are
   included within each panel. Sample sizes were determined on the basis
   of previous experiments using similar methodologies. All depicted data
   points are biological replicates taken from distinct samples. Each
   figure consists of a minimum of three independent experiments from
   multiple biological replicates, unless stated otherwise in the figure
   legends. n refers to the number of animals or the number of independent
   experiments with cell lines. For in vivo studies, age-matched mice were
   randomly assigned to treatment groups.

Acknowledgments