Fig 2: N-AS-acetylated COX2 produces SPMs.a Human recombinant COX2 treated in the presence or absence of 500 µM N-AS or aspirin was incubated with AA, EPA, or DHA in presence of human 5-LOX, and then SPM precursors were identified using systematic LC-MS/MS. Representative chromatogram showing the SPM precursors (Top, 15-HETE; Middle, 18-HEPE; Bottom, 17-HDHA). b Human recombinant COX2 treated in the presence of 500 µM N-AS or aspirin was incubated with AA, EPA, or DHA in presence of human 5-LOX. Related MS/MS spectra employed for identification of SPMs. 15R-LXA4, RvE1, and RvD1. c Quantification of 15R-LXA4, RvE1, and RvD1 in COX2 WT and COX2 S565A treated with 5-LOX in presence of N-AS, aspirin or not (n = 5–7 independent experiments per group). c One-way analysis of variance, Tukey’s post hoc test. All error bars indicate s.e.m. Source data are provided as a Source data file.

Fig 3: Loss of N-AS generation by Aß reduces COX2 acetylation and SPMs production in neurons and microglia, except astrocyte.a The primary culture of neuron, microglia, and astrocyte was prepared from C57BL/6 mice, and N-AS were detected by LC-MS/MS in these cells. Representative chromatograms of N-AS and quantification in neuron, microglia, and astrocyte treated 10 µM Aß or not (n = 6 per group). b Western blotting for ac-S565 and total COX2 in neuron, microglia, and astrocyte treated 10 µM Aß or not (n = 6 per group). c Immunofluorescence images and quantification of neuron (NeuN, red), microglia (Iba1, red), or astrocyte (GFAP, red) with ac-S565 (green) and COX2 (blue) (n = 6 per group, scale bars, 50 µm). d SphK activity in neuron and microglia treated 10 µM Aß or not (n = 6 per group). e Analysis of acetyl-CoA in neuron and microglia treated 10 µM Aß or not using assay kit (n = 6 per group). f Detection of sphingosine in neuron and microglia treated 10 µM Aß or not using UPLC (n = 6 per group). g Quantification of N-AS by LC-MS/MS in neuron and microglia treated 10 µM Aß or not in presence of acetyl-CoA, sphingosine, or SphK1 each (n = 6 per group). h Western blotting for ac-S565 in microglia and neuron treated 10 µM Aß or not in presence of N-AS, acetyl-CoA, sphingosine, or SphK1 each (n = 6 per group). i Quantification of 15R-LXA4, RvE1, and RvD1 using systematic LC-MS/MS in microglia and neuron treated 10 µM Aß or not in presence of N-AS, acetyl-CoA, sphingosine, or SphK1 each (n = 6 per group). a–f Student’s t-test. g–i One-way analysis of variance, Tukey’s post hoc test. All error bars indicate s.e.m. Source data are provided as a Source data file.

Fig 4: N-AS acetylates S565 in COX2.a N-AS binding activity of COX2 WT, COX2 S565A, COX2 N181A, COX2 T564A, and COX2 S567A was analyzed by filter binding assay. The binding velocity (Vbinding) of [14C] N-AS to COX2 was plotted to the N-AS concentration and the nonlinear regression analysis of the saturated plot yielded the kinetic parameters such as Kcat (catalytic constant) and KM (Michaelis–Menten constant) for N-AS and COX2 binding activity (n = 3 independent experiments per group). b Acetylation assay of purified COX2 protein treated with [14C] N-AS. The purified COX2 protein incubated in the presence of [14C] N-AS for 2 h at 37 °C and then COX2 was analyzed on scintillation counter. [14C] aspirin-treated COX2 protein and [14C] acetyl-CoA treated COX2 protein with SphK1 and sphingosine were positive control (n = 6–9 per group). c LC-MS spectra of peptide 560-GCPFTSFSVPDPELIK-575 (m/z = 918.94) of COX2 acetylated by N-AS. d LC-MS/MS spectra of ac-S565 in 560-GCPFTSFSVPDPELIK-575 of COX2. e Recombinant COX2 protein was incubated in the presence or absence of 2 mM N-AS or aspirin as indicated and S565 acetylation was determined by western blotting using anti-ac-S565 and COX2 antibodies. Data were replicated in six independent experiments with similar results. f The theoretical crystallographic model of human COX2 structure was taken from Protein Data Bank file 1V0X. S565 and residues (N181, T564, and S567) involved in the catalysis of COX2 were projected on the crystal structure of COX2 using PyMOL. Residues are shown as well: (1) S565 (red); (2) N181, T564, and S567 (green). g Proposed mechanism for the acetylation of S565 of COX2 by N-AS. h Acetylation assay of COX2 WT, COX2 S565A, COX2 N181A, COX2 T564A, and COX2 S567A treated with [14C] N-AS. (n = 6 independent experiments per group). b, h One-way analysis of variance, Tukey’s post hoc test. All error bars indicate s.e.m. Source data are provided as a Source data file.

Fig 5: Decreased N-AS in APP/PS1 microglia is restored with N-AS treatment, converting neurodegenerative microglia to phagocytic microglia via N-AS-triggered SPMs.a Scheme of experimental procedure. Daily administration of N-AS or vehicle subcutaneously to 7-months-old APP/PS1 mice for 8 weeks. b Representative chromatograms of N-AS and quantification in microglia and neuron derived from WT, APP/PS1 and APP/PS1 mice treated by N-AS (n = 6 per group). c Representative photomicrograph and quantification and quantification of microglia (Iba1, red) with ac-S565 (green) and COX2 (blue) in cortex of WT, APP/PS1, and APP/PS1 treated with N-AS mice. Scale bars, 20 µm (n = 6 per group). d Western blot analysis for ac-S565 and total COX2 in microglia and neuron derived from WT, APP/PS1, and APP/PS1 injected with N-AS mice (n = 6 per group). e Top, gating strategy for detection of CD86+CD206-(pro-inflammatory microglia), CD86+CD206+, and CD86-CD206+ (anti-inflammatory microglia) cells within brain cell populations. Bottom, graph displaying the calculated percentage of CD86+CD206-(pro-inflammatory microglia), CD86+CD206+, and CD86-CD206+ (anti-inflammatory microglia) cells in WT, APP/PS1, and APP/PS1 treated with N-AS mice (n = 6–7 mice per group). f Heatmap analysis of RNA seq expression data of microglia isolated WT, APP/PS1, and APP/PS1 injected with N-AS mice (n = 3–4 per group; each sample is a pool of two mice.). g–i GO term enrichment analysis of biological process (GOTERM_BP) (g), molecular function (GOTERM_MF) (h), and KEGG pathway enrichment analysis (i) was performed using DAVID Bioinformatics Resources 6.8. All data analysis was done at 9-months-old mice. b–e One-way analysis of variance. All error bars indicate s.e.m. Source data are provided as a Source data file.