Mineral particles stimulate innate immunity through neutrophil extracellular traps containing HMGB1

Abstract

Calcium phosphate-based mineralo-organic particles form spontaneously in the body and may represent precursors of ectopic calcification. We have shown earlier that these particles induce activation of caspase-1 and secretion of IL-1β by macrophages. However, whether the particles may produce other effects on immune cells is unclear. Here, we show that these particles induce the release of neutrophil extracellular traps (NETs) in a size-dependent manner by human neutrophils. Intracellular production of reactive oxygen species is required for particle-induced NET release by neutrophils. NETs contain the high-mobility group protein B1 (HMGB1), a DNA-binding protein capable of inducing secretion of TNF-α by a monocyte/macrophage cell line and primary macrophages. HMGB1 functions as a ligand of Toll-like receptors 2 and 4 on macrophages, leading to activation of the MyD88 pathway and TNF-α production. Furthermore, HMGB1 is critical to activate the particle-induced pro-inflammatory cascade in the peritoneum of mice. These results indicate that mineral particles promote pro-inflammatory responses by engaging neutrophils and macrophages via signaling of danger signals through NETs.

Figure 1

Mineralo-organic particles induce NET release by neutrophils. Mineralo-organic particles were prepared by adding 3 mM of CaCl₂ and NaH₂PO₄ each in DMEM containing (A) 0.1% or (B) 3% FBS, prior to incubation and preparation for TEM without staining as described in Methods. Data are representative of three independent experiments. Scale bars: 200 nm. (C) Live cell imaging of human neutrophils stained with Hoechst 33342 (blue) and treated at time 0 with mineral particles (labeled with FITC-bovine serum albumin; FITC-BSA; green). NET-associated DNA was stained with Sytox (red). Data are representative of three independent experiments. Scale bars: 10 μm. Neutrophils were incubated for 2 hours in the absence (D) or presence of particles in 0.1% FBS (E), 0.12 mg/ml bovine serum fetuin-A (BSF) (F), or 40 mg/ml BSF (G). NET-associated DNA was stained green by Sytox. Data are representative of three independent experiments. Scale bars: 20 μm. (H) Percentage of neutrophils forming NETs 2 hours after addition of particles. Data are shown as means ± standard errors of the mean (SEM) and the results of at least three independent experiments. **p < 0.005, vs. untreated (control) neutrophils. (I) Mineral particles were prepared with BSA or BSF before incubation with neutrophils. Data are shown as means ± SEM and the results of at least three independent experiments. *p < 0.05 and **p < 0.005, vs. untreated (control) neutrophils.

Figure 2

NETs released from particle-treated neutrophils induce secretion of TNF-α by macrophages. (A) Human neutrophils were treated with mineralo-organic particles for 30 min, followed by incubation with PMA-primed macrophages for 8 hours. TNF-α secretion was measured using ELISA. Data are shown as means ± SEM and the results of three independent experiments. *p < 0.05, as compared to untreated neutrophils. (B) Experimental design used to investigate the neutrophil-macrophage interaction. THP-1 cells were either directly co-cultured with neutrophils (top panel) or grown in the upper chamber of the Transwell system, separated from the neutrophils in the lower chamber by a semi-permeable membrane, which prevented cell-to-cell contact but allowed diffusion of soluble material (bottom panel). (C) Co-culture of THP-1 cells with particle-treated neutrophils enhanced TNF-α secretion. Data are shown as means ± SEM and the results of three independent experiments. *p < 0.05 and **p < 0.005, vs. untreated neutrophils. (D) Indirect co-culture of THP-1 cells and particle-treated neutrophils in the Transwell system failed to induce TNF-α secretion, but treatment of NETs with DNase enhanced TNF-α secretion. Data are shown as means ± SEM and the results of three independent experiments. **p < 0.005, vs. untreated neutrophils. (E) THP-1 cells or (F) neutrophils in the Transwell system were subjected to immunoblotting to detect intracellular TNF-α. Data are representative of three independent experiments.

Figure 3

ROS production is critical for particle-induced pro-inflammatory activation. (A) ROS production was determined by loading neutrophils with CM-H₂DCFDA and subjecting the cells to flow cytometry after co-culture with particles. Treatment with mineral particles enhanced intracellular ROS production. Data are shown as means ± SEM and the results of at least three independent experiments. (B) Neutrophils were incubated for 2 hours in the absence (left panel) or presence (middle panel) of particles in 0.1% FBS. In addition, concurrent treatment with apocynin (APO) was also carried out in the presence of particles in 0.1% FBS (right panel). NET-associated DNA was stained green by Sytox. Data are representative of three independent experiments. Scale bars: 20 μm. (C) Proportion of neutrophils forming NETs in the presence of particles with or without APO in panel (B) was quantified as percentage of total neutrophils. Data are shown as means ± SEM and the results of at least three independent experiments. **p < 0.005, vs. untreated neutrophils. ^## p < 0.005, vs. particle-stimulated neutrophils. (D) TNF-α production by NET-stimulated macrophages was determined by ELISA. Data are shown as means ± SEM and the results of at least three independent experiments. *p < 0.05, vs. untreated neutrophils.

Figure 4

HMGB1 is associated with NETs. (A) Cross-section (top panel) and 3D-reconstruction (bottom panel) of NETs interacting with THP-1 cells. NETs were stained green; chromatin DNA, blue. Plasma membrane of THP-1 cells was stained red to outline the cells. THP-1 macrophages closely interacted with NETs. Data are representative of three independent experiments. (B) NET-associated HMGB1 was released along with NETs after addition of mineral particles, and digestion of NETs removed the presence of HMGB1. Data are representative of three independent experiments. (C) Percentage of neutrophils forming NETs in response to particle stimuli in presence or absence of the neutrophil elastase inhibitor GW311616A for 2 hours. Data are shown as means ± SEM and the results of at least three independent experiments. **p < 0.005, vs. resting (control) neutrophils. ^## p < 0.005, as compared to particle-stimulated neutrophils. (D) Detection of freed NET-bound HMGB1 in culture medium by ELISA after DNase treatment. Data are shown as means ± SEM and the results of at least three independent experiments. **p < 0.005, vs. particle-stimulated, non-DNase-treated neutrophils without GW311616A treatment. ^## p < 0.005, vs. particle-stimulated, DNase-treated neutrophils without GW311616A treatment. (E) Proteins in culture supernatant of particle-treated neutrophils or THP-1 cells were precipitated and subjected to immunoblotting. HMGB1 was detected only in culture medium of particle-treated neutrophils upon DNase treatment to solubilize extracellular NETs. Data are representative of three independent experiments.

Figure 5

HMGB1 induces macrophage activation via the MyD88-TLR2/4 pathway. (A) THP-1 macrophages were co-cultured with particle-treated neutrophils in direct contact in the presence or absence of HMGB1-neutralizing antibodies (HMGB1 Ab). “% extent of activation” represents the extent of activation as indicated by TNF-α release, relative to the full response to particle induction, expressed in percentage (%). Data are shown as means ± SEM for the results of at least three independent experiments. ^## p < 0.005, vs. particle-stimulated neutrophils. (B) Partial inhibition of TNF-α production by anti-HMGB1 antibodies was observed for THP-1 cells in a non-contact system and exposed to DNase-digested NETs. Data are shown as means ± SEM and the results of at least three independent experiments. ^## p < 0.005, vs. particle-stimulated neutrophils. (C) Neutralizing antibodies targeting TLR2, TLR4, or receptor for advanced glycation end products (RAGE) were added to the co-culture system. Data are shown as means ± SEM and the results of at least three independent experiments. *p < 0.05, vs. resting (control) neutrophils. ^# p < 0.05, vs. particle-stimulated neutrophils. (D) THP-1 cells with expression of MyD88 (MyD88 KD) or RAGE stably knocked down (RAGE KD) by shRNA were subjected to the direct co-culture system. Data are shown as means ± SEM and the results of at least three independent experiments. *p < 0.05 and **p < 0.005, vs. resting (control) neutrophils. ^# p < 0.05, vs. particle-stimulated neutrophils.

Figure 6

Neutralization of HMGB1 inhibits primary macrophage activation by particle-induced neutrophils. Primary macrophages were co-cultured with particle-treated neutrophils in the presence or absence of HMGB1-neutralizing antibodies. Both neutrophils and macrophages were isolated from the blood of healthy donors. Macrophages (10⁶) were co-cultured with neutrophils at the indicated number and were pre-activated or not by particles. Data are shown as means ± SEM and the results of at least three independent experiments. **p < 0.005, vs. resting (control) neutrophils. ^# p < 0.05, vs. particle-stimulated neutrophils.

Figure 7

Intraperitoneal mineralo-organic particles promote pro-inflammatory activities in C57BL/6 mice. (A) DMEM without (“−”) or with mineralo-organic particles (“P”) were injected into the peritoneum of C57BL/6 mice. Eight hours after injection, intraperitoneal lavage was performed on the treated animals to access intraperitoneal cell population profiles. Data are representative of three independent experiments. Scale bars: 10 μm. (B) Mineralo-organic particles were intraperitoneally injected into C57BL/6 mice either together with anti-HMGB1-neutralizing antibodies (“Anti-HMGB1”) or isotype control antibody of chicken IgY (“Control IgY”). A vehicle control without the addition of mineralo-organic particles and antibodies also was included in this study. Two hours after injection of mineralo-organic particles, the RAW264.7 cells previously primed with IFN-γ were intraperitoneally injected into the mice. The inflammatory response elicited after six hours post-injection of mineralo-organic particles was assessed by detecting TNF-α via ELISA. Data are shown as means ± SEM and the results of at least three independent experiments. **p < 0.005, vs. vehicle control. ^# p < 0.05, vs. Control IgY. (C) Particles were intraperitoneally injected into C57BL/6 mice alone or together with GW311616, followed by injection of interferon (IFN)-γ-primed RAW264.7 cells two hours later. The inflammatory response elicited after 6-hour post-injection of particles was assessed by ELISA. Data are shown as means ± SEM and the results of at least three independent experiments. *p < 0.05 and **p < 0.001, vs. vehicle control. ^## p < 0.005, vs. GW311616-free control.

Figure 8

Working model of neutrophil-macrophage crosstalk via NET-bound HMGB1 during activation of neutrophils by mineralo-organic particles. When exposed to mineral particles, neutrophils interact with the particulate matters, which in turn lead to ROS production and NETosis. The released NETs deliver danger signals to bystander macrophages via NET-bound HMGB1. The NET-bound HMGB1 signals via a TLR2/TLR4-dependent pathway and initiates pro-inflammatory signaling cascades in the macrophages, resulting in inflammation via secretion of TNF-α.