C57Bl/6 background WT mice were obtained from Jackson Laboratory, and Ripk3 ^–/– mice were kindly obtained from Dr. Dixit (Genentech, Inc.). Sex- and age-matched (8 – 12 weeks of age) mice were used. Animal protocols were approved by the Institutional Animal Care and Use Committee.

C57Bl/6 mice (male, 8 – 10 weeks) were i.p. injected with 10 or 20 mg/kg Mdivi-1 or diluted DMSO in PBS for 2 h, followed by i.p. injection of 15 mg/kg LPS. Serum samples were collected at 1 or 2 hours after LPS injection, and the survival of mice was monitored.

LPS from Escherichia coli O111:B4 was obtained from List Biolabs and Sigma. Mdivi-1 was obtained from Enzo Life Sciences. Actinomycin D was purchased from Calbiochem. Phospho-ASK1 was purchased from Santa Cruz Biotechnology. Anti-GAPDH antibody was obtained from Chemicon. IκBα, phospho-p38α, phospho-JNK, phospho-ERK, Drp1, and phospho-Drp1 (Ser 637) antibodies were obtained from Cell Signaling Technology.

Control, MyD88 inhibitory peptide, and TRIF inhibitory peptide were purchased from IMGENEX. The amino acid sequences are; control peptide, RQIKIWFQNRRMKWKK; MyD88 inhibitor peptide, RQIKIWFQNRRMKWKKRDVLPGT; TRIF inhibitor peptide, RQIKIWFQNRRMKWKKFCEEFQVPGRGELH.

HEK 293T cells and bone-marrow-derived macrophages (BMDM) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum and antibiotics. Peritoneal macrophages were obtained from the peritoneal cavities of mice 3 days after intraperitoneal injection of 4% thioglycollate. BM cells obtained from the femurs of mice were cultured in M-CSF (10 ng/ml). Immortalized macrophages (iBMDMs) were generated by infecting BMDMs with retroviruses encoding v-raf and v-myc as previously reported (21).

To assess the polarization of macrophages, LPS-treated macrophages were incubated for 24 hours. Cells were incubated with anti-mouse CD16/CD32 Ab, followed by cell surface staining with CD11b-FITC (clone M1/70, Thermo Fisher), F4/80-PerCPCy5.5 (clone BM8, Thermo Fisher), and CD11c-AF700 (clone N418, Thermo Fisher). After fixation and permeabilization, intracellular NOS2 was stained with NOS2-PE eF610 (clone CXNFT, Thermo Fisher).

To test the levels of mtROS, macrophages were incubated with MitoSOX (5 μM) and LPS for 12 hours and washed. Data from FACSCalibur (BD Biosciences) were analyzed with FlowJo software.

Preparation of plasmids for shRNA targeting PGAM5 or Drp1 and recombinant lentivirus packaging were conducted as previously reported (19).

Control iBMDMs were transfected with NF-AT-, AP-1-, C/EBP-, or CREB-driven firefly luciferase plasmids, and pTK-RL Renilla plasmid by using Lipofectamine 3000 transfection reagent following the manufacturer’s protocol (Thermo Fisher). After 48 hours, cells were treated with LPS for 24 hours to measure the luciferase activity using the Dual-Luciferase Reporter Assay System (Promega). Activity of the Renilla luciferase was measured as the internal control. The fold induction of luciferase activity was calculated to the untreated cells.

Total RNA was isolated using RNeasy kit (Qiagen) following the manufacture’s instruction. The purity and concentration of RNA samples were measured using Nanodrop Spectrophotomemter. Preparation of cDNA library and transcriptome sequencing was conducted by Novogene Co., LTD (Davis, CA). Differential gene expression was analyzed by iDEP (integrated Differential Expression and Pathway analysis) (22). Genes with |log2(FoldChange)| > 1 and adjusted p-value < 0.05 were considered as “differentially expressed”. RNA-Seq data have been deposited at GEO and are openly available in NCBI GEO at https://www.ncbi.nlm.nih.gov/geo, reference number GSE235306.

Using total RNAs from macrophages, cDNA was prepared and quantitative PCR (qPCR) analysis was performed using PowerUp SYBR Green Master Mix (Thermo Fisher). Gene expression level was calculated by normalizing to Gapdh mRNA level. The sequences of primer sets are shown in Table 1 .

To analyze Tnf or Il6 mRNA stability, macrophages were stimulated with LPS, followed by actinomycin D (ActD, 10 µg/ml) addition to block transcription. Total RNAs were extracted from cells for qRT-PCR. The mRNA level of cells without ActD was set as 100%. The percentage of mRNA abundance in ActD-treated cells at different times was calculated.

Macrophages were washed and incubated at 4°C for 30 min in lysis buffer containing protease inhibitor cocktail (Roche), 1 mM DTT, and 1 mM PMSF. After 30 min at 4°C, the cell suspensions were centrifuged at 12,000 rpm for 10 min at 4°C. Protein concentrations were determined by the BCA method (Thermo Fisher). Cell lysates were incubated with isotype (goat Ig) or anti-PGAM5 Abs, and further incubated with Protein G-agarose. Immunoprecipitates were washed with lysis buffer and subjected to SDS-PAGE.

Whole cell lysates were separated with SDS-PAGE, and proteins were transferred to membranes. Immunoblotting was performed with the appropriate primary and HRP-conjugated secondary antibodies and visualized by chemiluminescence.

TNF and IL-6 levels in sera or culture supernatants were measured by ELISA according to the manufacturer’s protocol (Thermo Fisher).

Succinate or triglyceride levels in macrophages were determined using Succinate Assay Kit or Triglyceride Assay Kit (Abcam).

Metabolic changes were monitored by an extracellular flux analyzer XFe96 (Seahorse Bioscience). Macrophages were resuspended in XF Base DMEM containing 2 mM glutamine, and seeded in a Seahorse Bioscience 96-well plate at the density of 4 10⁴ cells per well. After 3 h, cells were treated with medium or LPS + IFN-γ, and incubated for 24 h at 37°C in an atmosphere of 5% CO₂.

Glycolytic rate assay was performed. During extracellular flux analysis, cells were sequentially treated with rotenone + antimycin A (final concentration of 10 μM each), followed by 2-deoxy-D-glucose (50 μM, final concentration) and the extracellular acidification rate (ECAR, glycolysis indication) was measured.

Cellular mitochondrial stress test was performed to measure mitochondrial metabolism. Cells were sequentially treated with 2 μM oligomycin A (OA), 1 μM FCCP + 1 mM pyruvate, and 1 μM rotenone + antimycin A to measure the oxygen consumption rate (OCR, mitochondrial respiration indication) and to calculate ATP production (23).

Statistical significance was analyzed by a two-tailed unpaired t-test or one-way ANOVA followed by Tukey’s post hoc multiple comparison test using Prism software (version 8, GraphPad, San Diego, CA). The survival of mice was tested by Kaplan-Meier analysis. A p value <0.05 was considered statistically significant. Each experiment was repeated as indicated in the figure legends; representative results are shown.

First, we examined the role of Drp1 in regulating the proinflammatory responses in macrophages. Previously, we generated a macrophage cell line using bone marrow cells from WT mice (21). The immortalized WT bone marrow-derived macrophages (iBMDMs) were infected with lentiviruses encoding control or Drp1 shRNA sequences to knock down Drp1 ( Figure 1A ). Control or Drp1 knockdown (KD) iBMDMs were stimulated with TLR ligands, and the levels of cytokines in culture supernatants were measured after 24 hours. TLR-induced TNF and IL-6 production was reduced significantly in Drp1 KD iBMDMs compared with control cells, indicating that any TLR signaling is involved in the Drp1-mediated inflammatory responses in macrophages ( Figure 1B ). Similarly, the expression of Tnf, Il6, or Il1b mRNAs was induced by LPS treatment in control iBMDMs, whereas KD of Drp1 significantly reduced their expression ( Figure 1C ). We further evaluated the role of Drp1 in macrophages by inhibiting Drp1 activity with Mdivi-1. We found that the production of TNF and IL-6 was also downregulated by inhibition of Mdivi-1 ( Figure 1D ), supporting the role of Drp1 in the regulation of inflammatory responses in macrophages.

Next, we examined how Drp1 regulates the signaling pathways in the TLR-mediated induction of inflammatory responses in macrophages. Degradation of IκB-α, and phosphorylation of p38, JNK, and ERK by LPS treatment were significantly reduced or delayed in Drp1 KD iBMDMs compared with control cells ( Figure 1E ), indicating that Drp1 plays a role in regulating the TLR-proximal signaling pathway in macrophages.

We further examined the role of Drp1 in vivo using endotoxin-induced sepsis model in mice. Mice received PBS or Mdivi-1 intraperitoneally, followed by an i.p. injection of LPS. The survival of mice was significantly extended by Mdivi-1 administration, suggesting the proinflammatory role of Drp1 in sepsis ( Figure 1F ). We further observed that serum levels of TNF or IL-6 increased by LPS administration, while Mdivi-1 administration substantially lowered their levels ( Figure 1G ), which indicates the inhibition of Drp1 activity in macrophages reduced the production of proinflammatory cytokines, resulting in the extended survival of mice in sepsis.

Collectively, our results suggest that Drp1 regulates the TLR-mediated signaling pathways for the induction of macrophage inflammatory responses and further contributes to the pathology of endotoxin-induced septic shock in mice, as inhibition of Drp1 extended the survival of mice.

PGAM5 dephosphorylates and activates Drp1, which results in mitochondrial fission and ROS production that cause cellular damage during necrosis (24). Although previous studies showed that Drp1 regulates the expression of cytokines in macrophages and T cells (25, 26), it is still unclear whether PGAM5 directly regulates Drp1 activity in innate immunity. Since our previous study suggests a novel function of PGAM5-Drp1 signaling in the regulation of NKT cell activation during liver inflammation (19), we tested whether Drp1 activation is dependent on PGAM5 activity in innate immunity. To test this, we first examined whether PGAM5 regulates Drp1 phosphorylation. We observed that LPS treatment resulted in the dephosphorylation of Drp1 serine 637 (Ser 637) in control iBMDMs, whereas PGAM5 KD did not show Drp1 dephosphorylation ( Figure 2A ). However, phosphorylation of Drp1 Ser 616 was not detected in macrophages (data not shown). Thus, this result indicates that PGAM5 regulates the phosphorylation status of Drp1 for activation in macrophages.

We further examined the role of PGAM5 in regulating the inflammatory responses of macrophages. Control or PGAM5 KD iBMDMs were stimulated with different TLR ligands. TLR ligand-induced production of TNF or IL-6 was significantly reduced in PGAM5 KD iBMDMs compared with control cells ( Figure 2B ), indicating the role of PGAM5 in macrophage activation. Similarly, TNF or IL-6 production was lower at the early time points after LPS stimulation, and the induction of Tnf or Il6 mRNAs was reduced in PGAM5 KD cells ( Figures 2C, D ), indicating that PGAM5 signaling plays a crucial role in the regulation of inflammatory responses in innate immunity. We further examined whether PGAM5 regulated the stability of inflammatory cytokines in macrophages and found that the rate of Tnf or Il6 mRNA degradation is comparable between control and PGAM5 KD iBMDMs ( Figure 2E ), suggesting that PGAM5 regulates the signaling events that induce the expression of cytokines, while the mRNA stability is not affected by PGAM5.

Next, we compared the TLR-mediated signaling pathways that are crucial for the production of inflammatory cytokines. Degradation of IκB-α, and phosphorylation of p38α, JNK, and ERK were significantly reduced in LPS-treated PGAM5 KD iBMDMs ( Figure 2F ). Additionally, the activation of transcription factors such as NF-κB, AP-1, CREB, and C/EBP by LPS was substantially reduced in PGAM5 KD iBMDMs, while NFAT activation was comparable between control and PGAM5 KD cells ( Figure 2G ). Collectively, our results suggest that PGAM5 regulates TLR-mediated macrophage activation by controlling the activation of downstream signaling pathways.

Previous studies have shown that PGAM5 regulates ASK1 activity for the activation of p38α and JNK, and ASK1 induces the activation of the MKK3/6 → p38α pathway in innate immune cells (27–29). Thus, we tested whether PGAM5 regulates the phosphorylation of ASK1. Control or PGAM5 KD iBMDMs were treated with LPS, and ASK1 phosphorylation at Thr 845 was assessed by immunoblotting. ASK1 phosphorylation was increased by LPS treatment, but the level of phosphorylation was comparable between control and KD cells ( Figure 2F ). This result indicated that PGAM5 does not directly regulate ASK1 in the activation of MAP kinase signaling. Therefore, our results ruled out the involvement of ASK1 downstream of PGAM5.

Because PGAM5 is known to regulate the induction of necroptosis, we tested whether necroptosis was involved in PGAM5-mediated macrophage activation. Induction of cell death by inhibition of caspase activity by zVAD in LPS-treated macrophages (necroptosis) was comparable between control and PGAM5 KD iBMDMs ( Figure 2H ), which indicates that PGAM5 does not induce necroptosis during macrophage activation. Additionally, the viability of macrophages was also comparable between control and Drp1 KD cells ( Figure 2H ). Collectively, our results suggest that the regulation of PGAM5-Drp1 activity is an essential aspect of necroptosis-independent TLR-mediated inflammatory responses in innate immune cells.

Next, we tested whether PGAM5-mediated Drp1 activation is dependent on adaptor proteins such as myeloid differentiation primary response 88 (MyD88) and TIR domain-containing adapter-inducing interferon-β (TRIF) that interact with TLRs upon ligand recognition. To this end, WT peritoneal macrophages were incubated with control, MyD88- or TRIF-inhibitory peptides, followed by LPS or poly I:C stimulation. We observed that LPS- or poly I:C-induced Drp1 dephosphorylation was delayed or reduced by inhibiting MyD88 or TRIF oligomerization ( Figure 3A ), suggesting that TLR-mediated signaling pathways regulate the PGAM5-Drp1-mediated induction of proinflammatory responses in macrophages.

One of the essential functions of mitochondria in macrophages is the generation of ROS (30). In addition to the bactericidal role of mtROS (31, 32), mtROS can also act as a signal that induces the expression of proinflammatory genes by regulating NF-κB and MAPK signaling pathways (8, 33). Thus, we asked whether PGAM5 regulates mtROS generation in macrophages. Stimulation with LPS increased the level of mtROS in control iBMDMs; however, the mtROS level in PGAM5 KD cells was significantly reduced ( Figures 3B , S1 ), implicating the role of PGAM5 in the generation of mtROS by regulating mitochondrial metabolism in macrophages. Based on this, it is suggested that PGAM5 signaling regulates mtROS production for the expression of inflammatory cytokines.

Activation of TLR signaling regulated the phosphorylation status of Drp1 as PGAM5 dephosphorylated Drp1 by LPS treatment. Thus, we further examined whether PGAM5 interacted with Drp1 in activated macrophages. Association of PGAM5 and Drp1 was detected at a low level in resting macrophages, while LPS treatment increased the formation of PGAM5 and Drp1 ( Figure 3C ), which suggests that MyD88/TRIF-dependent TLR signaling facilitates the formation of PGAM5 and Drp1 complex, leading to the dephosphorylation of Drp1 in macrophages.

PGAM5 has been identified as a key regulator of RIPK3-mediated signaling since PGAM5 interacts with the RIPK1/RIPK3 complex after stimulation. We tested whether RIPK3 is involved in TLR-mediated inflammatory responses in innate immunity. Peritoneal macrophages from WT or RIPK3 KO mice were treated with different TLR ligands, and TNF and IL-6 levels were measured. Proinflammatory cytokine levels induced by TLR ligand were comparable between WT and RIPK3 KO macrophages, indicating that RIPK3 does not regulate PGAM5-mediated inflammatory responses in macrophages ( Figure 4A ). Similarly, RIPK3 deficiency did not affect the levels of LPS-induced Tnf or Il6 mRNAs in macrophages ( Figure 4B ).

TLR-mediated signaling pathways were examined. LPS-induced IκB-α degradation and p38α, JNK, and ERK phosphorylation were comparable between WT and RIPK3 KO peritoneal macrophages ( Figure 4C ). Collectively, our data indicates that RIPK3 does not play a role in regulating the induction of inflammatory responses, and PGAM5 is not regulated by RIPK3 activity in macrophages.

High plasticity and adaptation to various stimulatory conditions are the hallmark of macrophages. Emerging evidence shows that contrasted metabolic activities determine the proinflammatory or anti-inflammatory phenotypes of macrophages (4–6). We further examined whether PGAM5-Drp1 signaling plays a role in the determination of macrophage phenotype. Classically activated and pro-inflammatory M1 macrophages are induced by LPS and IFN-γ, while IL-4 treatment increases the induction of alternatively activated M2 macrophages that express the anti-inflammatory genes in vitro (34). To test the role of PGAM5 in the induction of M1 macrophages, control or PGAM5 KD iBMDMs were incubated with LPS + IFN-γ, and the expression of proinflammatory markers was examined. LPS-induced expression of Tnf, Nos2, Il23, or Il6 in control cells was reduced significantly by KD of PGAM5 ( Figure 5A ), indicating that PGAM5 regulates the promotion of the proinflammatory phenotype of macrophages. Moreover, flow cytometry analysis indicated that the expression levels of NOS2 and CD11c, which are markers of proinflammatory macrophages, were lower in PGAM5 KD iBMDMs compared with control cells ( Figures 5B , S1 ). Furthermore, Drp1 also regulated the induction of proinflammatory responses in macrophages, as the expression of proinflammatory genes was significantly reduced in LPS + IFN-γ-treated Drp1 KD iBMDMs compared with control cells ( Figures 5C, D , S1 ). Collectively, it is suggested that PGAM5-Drp1 signaling contributes to the promotion of the proinflammatory macrophage phenotype.

High plasticity and adaptation to various stimulatory conditions are the hallmarks of macrophages. Emerging evidence shows that contrasted metabolic activities determine the proinflammatory M1 macrophages or anti-inflammatory M2 phenotypes of macrophages (4–6). Therefore, we further examined whether PGAM5 regulates the metabolic reprogramming of macrophages that results in the induction of a proinflammatory phenotype.

First, we examined whether the expression of metabolism genes is regulated by PGAM5 in macrophages by RNA-seq analysis. By comparing the data from LPS-treated control or PGAM5 KD iBMDMs, we found that LPS-induced expression of glucose transporter 1 (Glut1, Slc2a1), lactate dehydrogenase A (Ldha) and fatty acid synthase (Fasn) was significantly reduced by KD of PGAM5, while that of ATP citrate synthase (Acly), and acetyl-CoA carboxylase (Acaca) was comparable between control and PGAM5 KD cells. However, pyruvate dehydrogenase E1 subunit α 1 (Pdha1) and aconitate decarboxylase 1 [Acod1, or immunoresponsive gene 1 (Irg1)] expression by LPS was substantially increased in PGAM5 KD macrophages ( Figures 6A, B ). We further confirmed the expression of metabolism genes between LPS-treated control, PGAM5 KD, and Drp1 KD iBMDMs was compared by qPCR analysis ( Figure 6C ). This result indicates that glycolysis and fatty acid synthesis are tightly regulated by the mitochondrial phosphatase PGAM5 for the metabolic reprogramming of macrophages to a proinflammatory phenotype.

Upon LPS stimulation of macrophages, mitochondrial TCA cycle remodeling is accompanied, which results in TCA metabolite accumulation, including succinate, which plays a critical role in inflammatory cytokine production (35–37). We found that LPS-induced succinate production was comparable between control and PGAM5 KD macrophages ( Figure 6D ). Although glycolysis was reduced, Acod1 expression was increased in PGAM5 KD macrophages. Another metabolic feature of LPS-stimulated proinflammatory macrophages is fatty acid synthesis, leading to the accumulation of lipids in the form of triacylglycerol (TG) (2, 38). We found that LPS stimulation increased the concentration of TG in control macrophages, but not in PGAM5 KD cells ( Figure 6E ), which is consistent with the finding that the expression of Fasn is reduced in PGAM5 KD macrophages ( Figure 6B ).

To further test the role of PGAM5 in regulating cell metabolism, glycolysis and mitochondrial metabolism were analyzed using a Seahorse XF analyzer. By comparing the extracellular acidification rates (ECAR) between LPS +IFN-γ-treated control and PGAM5 KD iBMDMs, we found that KD of PGAM5 reduced the rate of glycolysis in macrophages ( Figure 6F ). Additionally, we found that PGAM5 also alters mitochondrial oxidative phosphorylation (OXPHOS) as the oxygen consumption rates (OCR) were significantly lower in PGAM5 KD iBMDMs than control cells ( Figure 6G ). Furthermore, ATP production induced by LPS + IFN-γ was significantly reduced in PGAM5 KD macrophages compared with control cells ( Figure 6H ).

Glycolysis rate, mitochondrial respiration, and ATP production were reduced in Drp1 KD iBMDMs compared with control cells, although the degree of reduction is not as significant as that of PGAM5 KD cells ( Figures 7A–C ).

Collectively, these data indicate that PGAM5 promotes the metabolic reprogramming of proinflammatory macrophages by regulating glycolysis and mitochondrial metabolism, which further contribute to fatty acid synthesis in macrophages and the induction of proinflammatory responses in macrophages.