Fig. 4: Eosinophil-dependent osteoclast supression is mediated by eosinophil peroxidase, reducing mitochondrial ROS levels and inhibiting RANKL-mediated signaling in osteoclast precursors.

a, b Proteome profiler array with supernatant from sorted eosinophils collected after 48 h culture (a) and heat map of relative levels (integrated pixel density) of the detected cytokines grouped into anti-osteoclastogenic, pro-osteoclastogenic, and undefined role in osteoclastogenesis (b). c Protein level of cystatin C (CysC), myeloperoxidase (MPO), and eosinophil peroxidase (EPO/EPX) in eosinophil supernatant (n = 3, 3, 5). d–f Quantification (d), viability (DAPI gated on BMMs) (e), and representative images (f) of TRAP-positive polynucleated (≥5 nuclei) WT osteoclasts supplemented with eosinophils (1:1 Eos/BMMs ratio), eosinophil supernatant (1:2 dilution), 0.1–0.8 µM of CysC, 5–10 µg/mL of MPO, and 0.05–0.1 µg/mL of EPX compared with unstimulated control (n = 10, 10, 10, 5, 5, 5, 5, 5, 5). Scale bar, 100 μm. g, h Quantification (g) and representative images (h) of TRAP-positive polynucleated (≥5 nuclei) WT osteoclasts supplemented with eosinophil supernatant (1:2 dilution) in the presence or absence of 10 µM of MPO/EPX inhibitor compared with unstimulated control (n = 10). Scale bar, 100 μm. i, j Quantification (i) and gating strategy (j) of mitochondrial (mt) ROS-positive osteoclast precursors (mtROS⁺CSF1R⁺CD11b⁺CD45⁺ cells) analyzed by FC at day 2 of differentiation after stimulation with eosinophils (1:1 Eos/BMMs ratio), eosinophil supernatant (1:2 dilution), 10 µg/mL of MPO, and 0.05–0.1 µg/mL of EPX compared with unstimulated control (n = 5). k, l Quantification (k) and representative signal (l) of phosphorylated p38 (Thr180/Tyr182), phosphorylated JNK (Thr183/Tyr185), and phosphorylated ERK1/2 (Thr202/Tyr204; Thr185/Tyr187) in whole cell lysates of osteoclasts supplemented with eosinophils (1:1 Eos/BMMs ratio), eosinophil supernatant (1:2 dilution), or 1 µg/mL of EPX compared with unstimulated control (n = 5). Data are shown as mean ± SEM. Symbols represent individual mice. P values were determined by Kruskal–Wallis Dunn’s test (4d, 4k phospho-p38), one-way ANOVA Tukey’s test (4g) or one-way ANOVA Dunnett’s test (4i, 4k phospho-JNK, phospho-ERK1/2) for multiple comparisons. Asterisks mark statistically significant difference (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Source data are provided as a Source data file.