O-GlcNAc Transferase Suppresses Inflammation and Necroptosis by Targeting Receptor-Interacting Serine/Threonine-Protein Kinase 3

Abstract

Elevated glucose metabolism in immune cells represents a hallmark feature of many inflammatory diseases, such as sepsis. However, the role of individual glucose metabolic pathways during immune cell activation and inflammation remains incompletely understood. Here, we demonstrate a previously unrecognized anti-inflammatory function of the O-linked β-N-acetylglucosamine (O-GlcNAc) signaling associated with the hexosamine biosynthesis pathway (HBP). Despite elevated activities of glycolysis and the pentose phosphate pathway, activation of macrophages with lipopolysaccharide (LPS) resulted in attenuated HBP activity and protein O-GlcNAcylation. Deletion of O-GlcNAc transferase (OGT), a key enzyme for protein O-GlcNAcylation, led to enhanced innate immune activation and exacerbated septic inflammation. Mechanistically, OGT-mediated O-GlcNAcylation of the serine-threonine kinase RIPK3 on threonine 467 (T467) prevented RIPK3-RIPK1 hetero- and RIPK3-RIPK3 homo-interaction and inhibited downstream innate immunity and necroptosis signaling. Thus, our study identifies an immuno-metabolic crosstalk essential for fine-tuning innate immune cell activation and highlights the importance of glucose metabolism in septic inflammation.

Keywords: HBP; O-GlcNAc; OGT; RIPK3; inflammation; necroptosis.

Figure 1.. LPS stimulation affects glucose metabolism in mouse macrophages.

(A and B) Total metabolite profiling in mouse BMMs stimulated with or without LPS (200 ng/ml) for 6 h was determined by the liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based metabolomics assay, assessed by principle component analysis (A) and pathway-enrichment analysis (B). (C) A schematic of three glucose metabolic pathways, including the glycolysis (middle), PPP (upper) and HBP (lower). (D–F) LPS-induced fold changes in intermediate metabolites of the glycolysis (D), PPP (E), and HBP (F) in mouse BMMs. (G and H) LPS-induced fold changes in ¹³C-intermediate metabolites of the glycolysis (G) and HBP (H) in mouse BMMs in the presence of ¹³C₆-glucose. (I and J) Immunoblotting of OGT and total O-GlcNAc in mouse BMMs (I), or peritoneal macrophages isolated from mice 24 h after sham or cecal ligation and puncture (CLP) procedure (J). * P < 0.05, versus controls (two-tailed Student’s t-test (E–H)). Data are from one experiment representative of three experiments (A, B, D–F; mean ± s.d. of four biological replicates) or two experiments (G and H; mean ± s.d. of three biological replicates) or four experiments (I and J). Please also see Figure S1.

Figure 2.. OGT deficiency enhances activation of the innate immune responses.

(A-F) BMMs generated from Ogt^fl/fl and Ogt^fl/flxLyz2-cre mice were left untreated or stimulated with LPS (A–D, G–I) or Pam3Cys or CpG (E and F) for indicated periods. Transcripts of inflammatory genes (A and E), IL-6 and TNF-α proteins (B and F), and nitrite concentrations (D) in the supernatants were measured with RT-PCR, ELISA and Griess assay, respectively. Nos2 protein was assayed by immunoblotting (C) (G and H) Immunoblotting for NF-κB (G, left), and MAPK (H, left) signaling molecules and densitometric analysis (G and H, right). (I) Immunoblotting of NF-κB p65, RelB and p50 in the cytosolic (left) and nuclear (right) compartments. * P < 0.05, versus controls (two-tailed Student’s t-test). Data are from one experiment representative of five experiments (A, B, D–F; mean ± s.d. of four biological replicates) or four experiments (C, G, H and I). Please also see Figure S2 and S3.

Figure 3.. Myeloid-derived OG protects mice from experimental sepsis.

Ogt^fl/fl and Ogt^fl/flxLyz2-cre mice were injected intraperitoneally with 15 mg/kg body weight LPS (n=10 for each group) (A–C) or were subjected to sham or CLP procedure (n=8 to10 for each group) (D–G). Survival were recorded (A and D). IL-6, TNF-α, and IL-1β protein concentrations in peritoneal lavage (B and E), serum (C and F), and lungs homogenates (G) were measured by ELISA 24 h after LPS injection or CLP procedure. * P < 0.05, versus controls (two-tailed Student’s t-test). Data are from one experiment representative of three experiments (B, C, E–G; mean ± s.d.) or two experiments (A and F). Please also see Figure S4.

Figure 4.. OGT inhibits the innate immune responses through RIPK3.

(A) NF-κB-driven luciferase activities by co-expression of MYD88, TRAF6, RIPK1, RIPK3, IKK1, IKK2, or p65, in the presence or absence of the expression plasmid for OGT. (B and C) IL-6 and TNF-α proteins (B) and nitrite (C) produced by Ogt^fl/fl, Ogt^fl/flxLyz2-cre, Ripk3^−/− or Ogt^fl/flxLyz2-creRipk3^−/− BMMs stimulated with or without LPS. (D−G) Survival rate (D), IL-6, TNF-α and IL-1β proteins in peritoneal lavage (E), serum (F), or lungs (G) in Ogt^fl/fl (n=14), Ogt^fl/flxLyz2-cre (n=10), Ripk3^−/− (n=12) or Ogt^fl/flxLyz2-creRipk3^−/− mice (n=8) subjected to CLP procedure. (H and I) IL-6 and TNF-α proteins (H) and nitrite (I) produced by Ogt^fl/fl, Ogt^fl/flxLyz2-cre, Mlkl^−/− or Ogt^fl/flxLyz2-creMlkl^−/− BMMs stimulated with or without LPS. (J–L) Survival rate (J), IL-6, TNF-α and IL-1β proteins in serum (K) or peritoneal lavage in Ogt^fl/fl (n=12), Ogt^fl/flxLyz2-cre (n=10), Mlkl^−/− (n=15) or Ogt^fl/flxLyz2-creMlkl^−/− mice (n=13) mice subjected to CLP procedure. * P < 0.05, versus controls (two-tailed Student’s t-test (A–C, E–I, K and L)). The results shown are representative of four independent experiments (A–C; mean ± s.d. of four biological replicates) or two experiments (D–L). Please also see Figure S5.

Figure 5.. OGT inhibits necroptosis through RIPK3.

(A–C) Cell death assessed by LDH release (A) or Sytox Green staining (C), and phosphorylation of necroptosis signaling molecules (B) in Ogt^fl/fl or Ogt^fl/flxLyz2-cre BMMs left untreated or stimulated with LPS (200 ng/ml) plus zVAD (10 µM) for indicated periods. (D) Immunoblotting of p-RIPK3 and p-MLKL in Ogt^fl/fl or Ogt^fl/flxLyz2-cre BMMs placed in a transwell system stimulated with LPS and zVAD. (E) Immunoblotting of HMGB1 and IL-1α in the peritoneal lavage fluids of Ogt^fl/fl and Ogt^fl/flxLyz2-cre mice 24 h after sham or CLP procedure. (F–H) Necroptosis signaling molecules in the NP-40-insoluble fractions (F), total or phosphorylated RIPK1 in RIPK3 immunoprecipitates (G), and cell death assessed by LDH release in the absence or presents of RIPK1 inhibitor Nec-1 (20 µM) or RIPK3 inhibitor GSK-872 (10 µM) (H). (I–M) LDH release (I and K) and phosphorylation of necroptosis signaling molecules (J and L) in Ogt^fl/fl, Ogt^fl/flxLyz2-cre, Ripk3^−/− , Ogt^fl/flxLyz2-creRipk3^−/− (I and J), Mlkl^−/− or Ogt^fl/flxLyz2-creMlkl^−/− (K and L) BMMs left untreated or stimulated with LPS plus zVAD, or placed in culture medium with 2 mM glucose (M). * P < 0.05, versus controls (two-tailed Student’s t-test (a, g, h)). Data are from represent of four independent experiments (A, H, I, K and M; mean ± s.d. of four biological replicates) or three experiments (B–G, J and L). Please also see Figure S6.

Figure 6.. O-GlcNAcylation of RIPK3 on T467 suppresses inflammation and necroptosis.

(A–G) Total cell lysates of Ogt^fl/fl or Ogt^fl/flxLyz2-cre BMMs left untreated (A and B) or stimulated with LPS alone (C) or LPS plus zVAD (D) with or without PUGNAc pretreatment (E). Immunoprecipitation with anti-RIPK3 antibody (A, C–E) or succinylated wheat germ agglutinin (sWGA) beads to pull down O-GlcNAcylated proteins (B). (H) FLAG-tagged RIPK3 overexpressed in 293T cells together with OGT WT or enzyme-dead K908A mutant was immunoprecipitated with anti-FLAG beads, followed by immunoblotting with anti-O-GlcNAc antibody. (I) Domain organization of human RIPK3 and the RHIM sequence. (J) O-GlcNAcylation of the full-length, N-terminal or C-terminal fragments of RIPK3 overexpressed in 293T cells in the presence or absence of WT OGT. (K and L) O-GlcNAcylation of FLAG-tagged RIPK3 WT or mutants overexpressed in 293T cells together with OGT was analyzed as described in (H). (M–P) RIPK3-silenced THP-1 cells were virally transfected with RIPK3 WT or T467A mutant, followed by LPS (M and N) or LPS plus zVAD (O and P) stimulation. Transcripts of RIPK3, IL6, TNFA, and IL1B (M), total and phosphorylated RIPK3, Erk, IKKα/β, IκBα and p65 (N) or necroptosis signaling molecules (P), and LDH release (O) was assayed. *P < 0.05, versus controls (two-tailed Student’s t-test (L and N)). Data are from one experiment representative of three independent experiments (A–G, I, J, K, M and O) or four experiments (L and N; mean ± s.d. of three biological replicates). Please also see Figure S7.

Figure 7.. O-GlcNAcylation of RIPK3 inhibits RHIM-mediated RIPK3-RIPK1 interaction.

(A–D) FLAG-tagged RIPK3 was overexpressed in 293T cells together with indicated expression plasmids. Immunoprecipitation with anti-FLAG beads, followed by immunoblotting with specific antibodies against O-GlcNAc, p-RIPK3, RIPK1 and GFP. (E–G) GFP-tagged RIPK3 WT or mutants was overexpressed in 293T cells. Immunoprecipitation with anti-GFP beads, followed by immunoblotting with specific antibodies against O-GlcNAc, p-RIPK3 or RIPK1. (H and I) Transcripts of IL6 and TNFA (H) and LDH release (I) from RIPK3-silenced THP-1 cells virally transfected with RIPK3 WT or various mutants, followed by LPS (H) or LPS plus zVAD (I) stimulation. (J) Structural modeling of RIPK3 by using the I-TASSER program. O-GlcNAcylated T467 is red and the amino acids VQVG of the RHIM motif is highlighted as pink sphere. * P < 0.05, versus controls (two-tailed Student’s t-test (g, h)). Data are from one experiment representative of three experiments (A–F) or four experiments (H and I; mean ± s.d. of four biological replicates).