Persistent inflammatory response in the diabetic wound impairs the healing process, resulting in significant morbidity and mortality. Mounting evidence indicate that the activation of Nod-like receptor protein (NLRP) 3 inflammasome in macrophages (MΦ) contributes to the sustained inflammatory response and impaired wound healing associated with diabetes. However, the main trigger of NLRP3 inflammasome in the wounds is not known. Neutrophils, as sentinels of the innate immune system and key stimulators of MΦ, are immune cells that play the main role in the early phase of healing. Neutrophils release extracellular traps (NETs) as defense against pathogens. On the other hand, NETs induce tissue damage. NETs have been detected in the diabetic wound and implicated in the impaired healing process, but the mechanism of NETs suspend wound healing and its role in fostering inflammatory dysregulation are elusive. Here, we report that NLRP3 and NETs production are elevated in human and rat diabetic wounds. NETs overproduced in the diabetic wounds triggered NLRP3 inflammasome activation and IL-1β release in MΦ. Furthermore, NETs up-regulated NLRP3 and pro-IL-1β levels via the TLR-4/TLR-9/NF-κB signaling pathway. They also elicited the generation of reactive oxygen species, which facilitated the association between NLRP3 and thioredoxin-interacting protein, and activated the NLRP3 inflammasome. In addition, NET digestion by DNase I alleviated the activation of NLRP3 inflammasome, regulated the immune cell infiltration, and accelerated wound healing in diabetic rat model. These findings illustrate a new mechanism by which NETs contribute to the activation of NLRP3 inflammasome and sustained inflammatory response in the diabetic wound.
Diabetic foot ulcer (DFU), one of the most complex, costly, and frequent complications of diabetes mellitus, results in one patient losing a lower limb or part of a lower limb for amputation every 30s globally . Amputation increases the 5-year death risk of patients with DFU 2.5 times . More research into DFU pathogenesis and treatment strategies is needed because of the patients’ suffering, huge associated costs, and limited treatment options [3,4].
The wound healing process involves integrated and overlapping phases of hemostasis, inflammation, proliferation, and remodeling . Chronic wounds, such as DFU, often fail to follow such precise controlled stages and remain unresolved in the inflammatory phase [6–8]. Macrophages (MΦ) are involved in each phase of the healing process and have been demonstrated to play a critical role in the repair of wound tissue, especially at the inflammatory stage [9,10]. In the early stages of inflammation, MΦ excrete inflammatory cytokines, clearing pathogens and necrotic cells. In the late stages of inflammation, MΦ contribute to growth factor production, tissue repair, and proliferation . In the pathogenesis of the nonhealing diabetic wound, MΦ exhibit prolonged accumulation in tissues and a predominance of the proinflammatory phenotype, resulting in elevated proinflammatory cytokine levels and reduced levels of various growth factors [12–14].
Emerging evidence suggest that activation of a multiprotein complex in MΦ, called the Nod-like receptor protein (NLRP) 3 inflammasome, together with elevated IL-1β levels contributes to the diabetes-associated impaired wound healing in human and mouse [15–17]. During inflammasome activation, NLRP3 recruits an adaptor molecule, apoptosis-associated speck-like protein (ASC), and caspase-1. ASC contains a caspase recruitment domain that activates caspase-1. Subsequently, the activated caspase-1 cleaves the pro-IL-1β and pro-IL-18 molecules. Since active caspase-1 can release proinflammatory cytokines and IL-1β and induce MΦ death, NLRP3 inflammasome is considered to play a critical role in sustained inflammation in the diabetic wound [18,19]. However, the upstream triggers of the NLRP3 inflammasome remain elusive.
Neutrophils, as sentinels of the innate immune system, were rapidly and substantially recruited to the wounds . These cells debride the necrotic tissue and pathogens by phagocytosis and intracellular destruction, and then undergo apoptosis. Neutrophils were mopped up by MΦ, which serving as key regulated stimuli stimulators of the MΦ phenotype and inflammation statement . Some inflammatory diseases, such as atherosclerosis, asthma, and systemic lupus erythematosus (SLE), diabetic wound, exhibit the accumulation of neutrophils and macrophages in lesions, which implied that their interaction between neutrophils and macrophages may play an important role in inflammatory cascade [21–23]. Recently, much interest has focused on NETosis, a novel death strategy of neutrophils, whereby the cells expel decondensed chromatin decorated with cytotoxic and antimicrobial agents. The reticular structure of the decondensed chromatin led to the term ‘neutrophil extracellular traps’ (NETs). The decondensation of chromatin is one of the most crucial processes in NET extrusion . Histone3 were citrullinated by peptidyl arginine deiminase-4 (PAD4), which causes the loss of electrostatic attraction . Thus, the hypercitrullinated histone3 was considered the symbol and the reliable indicator of NETs. NETs play a role in cellular defenses by immobilizing and killing pathogens. Meanwhile, NETs also induced host tissue damage [26,27]. NETs are frequently associated with infection, inflammatory, ischemic organ damage, thrombosis, and autoimmunity disease [28–30]. The DNA scaffold of NETs can be hydrolyzed by deoxyribonuclease I (DNase I), the first well known extracellular nonrestriction endonuclease, to short oligonucleotides . Earlier in vitro and in vivo works has shown that plasmic DNase1 is required to digest the DNA backbone of NETs [32,33]. The therapeutic administration of DNase I is capable to alleviate host injury in inflammatory or immunological conditions in various animal models . DNase I has been approved by FDA for the treatment of cystic fibrosis (https://www.drugs.com/pro/pulmozyme.html), and it will be easily translated into clinical practice as a potent stimulator of diabetic wound healing
Since NETs were detected to impair healing of diabetic wound . We hypothesized that NETs, primed by diabetes, can trigger NLRP3 expression and activation of MΦ, amplifying the inflammatory response in the diabetic wound. In the present study, we demonstrated that NETs induce NLRP3 inflammasome formation and activation in MΦ through the TLR/NF-κB and reactive oxygen species (ROS)/thioredoxin-interacting protein (TXNIP) pathways, respectively. We also showed that such NETs may be degraded by DNase I, which might regulate the inflammatory statement and constitutes a possible treatment for diabetic wound healing.
Materials and methods
The diabetic wound tissue was from debridement or amputation of unhealed lower lamb wounds that persisted for at least 3 months. Ethical approval was from the independent ethics committee of Shanghai Ninth People’s Hospital affiliated to Shanghai Jiao Tong University, School of Medicine (Number: 2016-105-T54). The tissue was removed from the margin of the wound. The healthy patients’ wound tissue was obtained during debridement of acute trauma injury or plastic surgery. All patients provided written informed consent prior to the study.
Culture media/supplements were from GIBCO; streptozocin and thioglycolate (STZ), Phorbol 12-myristate-13-acetate (PMA), Deoxyribonuclease I from bovine pancreas (Dnase I), antioxidant N-acetylcysteine (NAC) were from Sigma-Aldrich (St Louis, MO); JSH-23 was from Selleck Chemicals (Houston, TX, U.S.A.); TLR4 inhibitor (CLI-095) and TLR9 inhibitor (ODN 2088) were from InvivoGen (San Diego, California, U.S.A.); NLRP3 inhibitor (MCC950) were from Selleck (Selleckchem, Houston, U.S.A.); 4’,6-diamidino-2-phenylindole (DAPI) were from Life Technologies (Carlsbad, CA). The following antibodies were obtained from various supplies: anti-Histone H3 (citrulline R2+R8+R17), anti-IL-1β, anti-TXNIP were from Abcam (Cambridge Science Park, Cambridge); Anti-Histone 3 (H3), CD66b were from ABclonal Biotech (Woburn MA), anti-NLRP3 and anti-Caspase-1 (Millipore, Merck KGaA, Darmstadt); anti-CD68 (Bio-Rad, Philadelphia, PA); anti-TLR4, anti-TLR9 (Santa Cruz Biotechnology, Santa Cruz, CA); anti-NF-κB, anti-Phospho-NF-κB, anti-β-actin and HRP-linked antibody (Cell Signaling Technology, Danvers, MA), CellROX™ Deep Red Reagent, donkey anti-rabbit 132 alexa flour 488, anti-rabbit alexa flour 594, anti-mouse alexa flour 488 (Thermo fisher, Carlsbad, CA).
Sixty male Sprague Dawley (SD) rats (6–8 weeks old, 150–180 g) were purchased from the Shanghai Laboratory Animal Center and housed at the Animal Science Center of the Shanghai Jiao Tong University, School of Medicine (SJTUSM, rats were maintained under a 12-h light/dark cycle at 22°C, two rats per cage). Experiments involving animals were performed in agreement with the rules of the Animal Care Committee of SJTUSM. All experimental protocols were approved by the SJTUSM Institutional Animal Care and Use Committee. Animals were randomly assigned to two groups (diabetic group, n=40; normal group, n=20). Type 1 diabetes was induced in rats by a single I.P. injection of STZ (60 mg/kg body weight) in citrate buffer (0.1 M, pH 4.5). The control group received I.P. injection of saline. Random blood glucose level >16.6 mmol/l 8 weeks after the STZ injection was considered as an indicator of diabetes.
Wound procedure and wound healing examination
Before wounding day, the back of all rats was shaved and hairs were thoroughly removed with depilatory after anesthetization (I.P. Thiopental sodium, 40 mg/kg BW). The shaved area was cleaned with sterile water and disinfected with 75% ethanol twice. The positions of wound were pointed by surgical marker pen and the spine was set as the long axis. Mice received four full thickness excision dorsal wounds by using 0.9 cm corneal trephine to outline the wound area, and the epidermal and dermal layer were excised using scissors. Sterile gauze covered the wounds. The rats were maintained separately, then supplied with unrestricted food and water. The diabetic group were assigned into two groups randomly (diabetic group, n=20; Dnase I treatment group, n=20). The wounds of control, diabetic group were dealt with saline every 24 h for 8 consecutive days. And the wounds of DNase group were administrated by DNase I (1 mg/kg, dissolved in saline, 10 mg/ml). At day 1, 2, 5, 9, 12, and 14 after injury, rats were anesthetized, the wounds were photographed and wound closures were measured. All images were analyzed with ImageJ software.
Skin wound tissues and additional 5 mm of the surrounding normal skin after anesthesia were harvested and fixed in 10% polyformaldehyde solution on days 2, 5, 9, and 14 after wounding. The fixed tissues were embedded in paraffin and sectioned to a thickness of 6 to 7 μm. The sections were stained with hematoxylin/eosin (H&E) for histological and morphometric evaluation.
Isolation and purification of rat neutrophils and MΦ
Rat neutrophils were isolated as previously described . Briefly, sterile peritonitis were induced in rats by I.P. injection of 3% thioglycolate medium (15 ml/kg); 4–6 h after thioglycolate injection, rats were killed, sterilized with 75% ethanol, and the intact peritoneal wall exposed. Then, 15 ml of PBS was I.P. injected as a sterile harvest solution for neutrophil isolation. The entire peritoneal fluid was next withdrawn. The neutrophils were separated with PolymorphPrep gradient solution at 450 × g for 30 min, and washed by PBS. Then hypotonic lysis was performed for 10 min to eliminate red blood cells. After centrifugation at 250 × g for 10 min, the cells were resuspended in 1% FBS RPMI 1640, counted, and used for NET production and visualization.
Rat MΦ were also isolated from the peritoneal fluid. Briefly, sterile peritonitis was induced in rats by I.P. injection of 3% thioglycolate medium (15 ml/kg). Three days after the injection, 15 ml of DMEM was I.P. injected as a sterile harvest solution for MΦ isolation. The entire peritoneal fluid was then withdrawn. Cells within the peritoneal fluid were placed in a flask for 2 h, and then washed with PBS to remove non-adherent cells. Cells were cultured in DMEM with 10% FBS, counted, and used for the procedures described below.
NETs induction and detection
Isolated neutrophils were seeded and cultured in 6-cm dishes or eight-well EZ chamber slides (5 × 105 cells/ml), for 1 h at 37°C, under 5% CO2. PMA (100 ng/ml) was then added to stimulate NET formation for 3 to 5 h. Unstimulated neutrophils were used as a control group. NETs eliminated by DNase I (0.1 mg/ml) digestion for 3 h were used as the DNase group. Cells were fixed with 10% polyformaldehyde for immunofluorescence analysis. Cell lysates were harvested for Western blotting (details of immunofluorescence and Western blotting analyses are provided below).
NETs were prepared as mentioned above. To determine the effect of PMA-induced NETs on primary MΦ, MΦ were harvested by trypsin treatment, centrifuged at 800 rpm for 5 min, and re-suspended in DMEM (1 × 105/ml). The cells were then incubated with neutrophils (control group), NETs (NET group), or NETs treated with DNase I (DNase group) for 12 h at 37°C, under 5% CO2. Subsequently, the cells were fixed with 10% polyformaldehyde for immunofluorescence analysis. Cell lysates were collected for Western blotting. MΦ were preincubated with various compounds, such as NF-κB transcription activity inhibitor (JSH-23, 10 μM), TLR-4 inhibitor (CLI-095, 3 μM), TLR-9 inhibitor (ODN 2088, 5 μM), NLRP3 inhibitor (MCC950, 10nM), or N-acetylcysteine (NAC, 10 mM) for 1 h.
For tissue analysis, the wound tissue was harvested 2, 5, 9, and 14 d after injury, and stored at −80°C. The tissue was homogenized in tissue lysis buffer with a protease inhibitor cocktail. For cell analysis, neutrophils or MΦ were lysed in lysis buffer with a protease inhibitor cocktail. All lysates were processed as follows. DNA was sheared by sonication and the lysates were incubated on ice for additional 30 min before centrifugation at 12000 rpm for 15 min. The supernatants were collected and stored at —80°C. Equal amounts of protein were mixed with sample buffer, boiled, and analyzed using SDS-PAGE. The membrane was blocked and then incubated with the appropriate primary antibodies, followed by incubation with HRP-conjugated secondary antibodies. Protein bands were visualized by using enhanced chemiluminescence (Millipore, GER). Relative protein quantities were determined using a densitometer and are presented in comparison with β-actin or H3 protein levels.
Cell ROX detection
MΦ were treated as ‘cell culture’ for 6 h. Then, they were incubated with the oxidative stress detection at 37°C for 30 min. After washing, ROS were observed under a fluorescence microscope.
Coimmunoprecipitation (Co-IP) assays were performed using primary MΦ. The experiments were performed according to the manufacturer’s instructions for protein A/G magnetic beads for IP (Biotool, U.S.A.). Briefly, the antibodies (5–10 μg) were incubated with protein A/G magnetic beads (50 μl) for 2 h at room temperature (RT). Cell lysis was achieved by incubating in cell lysis buffer for Western and IP (Beyotime, China). Equal amounts of TXNIP protein (200 μg) in equal volume (100 μl) of the lysis buffer were incubated with the protein A/G magnetic bead-antibody complex for 30 min at RT. Before each experiment, protein A/G magnetic beads or protein A/G magnetic bead–antibody complexes were washed three times in washing buffer (50 mM Tris, 150 mM NaCl, and 0.4% Tween 20, pH 7.5). The protein A/G magnetic bead–antibody complex was then washed using elution buffer (0.1–0.2 M glycine and 0.4% Tween 20, pH 2.5). Protein complexes were collected and submitted to immunoblot analysis.
Wound tissue sections for immunofluorescence analysis were prepared as mentioned above. The sections were then deparaffinized, rehydrated, and washed in distilled water. Cell slides for immunofluorescence were fixed with paraformaldehyde for 20 min. Slides were washed with PBS three times, permeabilized with 0.1% Triton X-100, and incubated with 10% BSA in PBS to minimize nonspecific binding of the primary antibody. They were then incubated with the appropriate primary antibodies overnight at 4°C. After washing, the molecules of interest were visualized by treatment with a secondary antibody and observed under a fluorescence microscope. The nuclei were counterstained with DAPI. All experiments were performed in triplicate.
Statistical analysis was performed using the Student’s t-test or one-way analysis of variance (ANOVA). Data were analyzed by using SPSS 19 software (SPSS Inc., Chicago, IL, U.S.A.). All statistical analysis was performed using the GraphPad Prism software (GraphPad 243 Software Inc.). The value of P<0.05 was considered statistically significant. Results are expressed as the mean ± SD.
Diabetes primes human and rat NET formation, and NLRP3 inflammasome activation in the wound
To determine NET performance and NLRP3 inflammasome levels in the diabetic wound, and their possible pathological impact on diabetic unhealed wound, we compared the relative protein levels in skin wound samples from healthy and diabetic patients. In contrast with the elevated levels of hyper-citrullinated histone 3 (Cit-H3, NET marker) and NLRP3 in the diabetic ulcer wound, only low-levels of Cit-H3 (NETs) and NLRP3 were detected in wounds from healthy patients (Figure 1A). Similar to human, NETs was barely exist in the healthy rat wound, while high amounts of NETs were observed in wound of STZ-induced type 1 diabetic rats. Meanwhile, the NLRP3 inflammasome-related proteins, such as NLRP3, pro-IL-1β, and IL-1β, were significantly higher in diabetic rats wound than that in healthy rats and showing an increasing trend along with the progress of healing (Figure 1B). Considering the fact that neutrophils are first rapidly recruited to the wound after injury and then mopped up by MΦ, these findings suggested that NETs might serve as regulated stimuli to promote NLRP3 inflammasome formation and activation in MΦ.
Diabetes primes human and rat NETs formation, and NLRP3 inflammasome activation in the wound
NETs stimulate NLRP3 production and IL-1β release in MΦ
To investigate whether NETs were able to promote NLRP3 production and IL-1β release in MΦ, we extracted intraperitoneal-derived primary neutrophils and MΦ from SD rats. CD68 staining patterns were then analyzed. Immunofluorescence evaluation indicated that both markers were present in over 90% of cells extracted from the peritoneal cavity (data not shown). Next, we induced NETs by exposing neutrophils to PMA (0.1 μg/ml) for 3 to 5 h and then degraded these NETs over a 3-h DNase I treatment . Immunofluorescence staining of Cit-H3 and CD66b clearly revealed the wire-mesh structure of NETs, and the structure was no longer observed after the DNase I treatment (Figure 2A). Western blotting analysis confirmed the presence and dissolution of NETs, following NET induction and DNase I treatment (Figure 2B). Primary rat MΦ were then exposed to control neutrophils, NETs, or DNase I-treated NETs, for 12 h. Western blotting revealed the most robust increase of NLRP3, pro-IL-1β, IL-1β, ASC, and active caspase-1 levels in MΦ treated by NETs (compared with the other treatments), suggesting enhanced inflammasome production and activation induced by NETs (Figure 2C). NLRP3 production was confirmed by immunofluorescence, with most NLRP3-positive cells and strong fluorescence intensity in MΦ stimulated by NETs (Figure 2D). Considering that most of inflammasome sensor molecules, such as NLRP3, AIM2, NLRC4, can activate caspase-1 and subsequently cleave the proinflammatory IL-1 family, NLRP3 specific inhibitor was exerted to explore the specificity that NETs activate NLRP3. The diaryl-sulfonylurea-containing chemical compound, MCC950, exhibits potent and selective inhibition for NLRP3 inflammasome, which is utilized as a tool for the NLRP3 inflammasome study and a potential therapeutic for NLRP3-associated autoinflammatory and autoimmune syndromes [37–39]. Pretreated with MCC950 blunt the substantial increases of NLRP3, active caspase-1 and IL-1β expression (except pro-IL-1β) induced by NETs on macrophage (Figure 2E), which implies the activation of NLRP3 inflammasome under the stimulation of NETs. Thus, we determined that NETs stimulated NLRP3 activation and IL-1β release in MΦ.
NETs stimulate NLRP3 production and IL-1β release in MΦ
TLR/NF-κB signaling pathway contributes to the up-regulation of NLRP3 and pro-IL-1β induced by NETs
The activation of NLRP3 inflammasome need the prime signal . TLR/NF-κB is one of the most important signaling pathways, which up-regulates the expression of NLRP3 and pro-IL-1β protein . To investigate whether NF-κB was involved in the relative elevation of NLRP3 protein levels induced by NETs, we examined the degree of NF-κB phosphorylation in MΦ incubated with control neutrophils, NETs, or DNase I-treated NETs. The fraction of phosphorylated NF-κB (p-p65) was higher in MΦF treated with NETs than in MΦ incubated with the control or DNase I-treated NETs (Figure 3A,B). To verify the role of NF-κB pathway in the relative elevation of the levels of NLRP3 inflammasome-related proteins induced by NETs, we incubated MΦ with a NF-κB nuclear translocation blocker, JSH-23. JSH-23 treatment resulted in reduced protein levels of NLRP3, pro-IL-1β, and IL-1β, but not those of active caspase-1 (Figure 3C, red box). This was confirmed by immunofluorescence staining (Figure 3C,D). These findings demonstrated that NF-κB pathway was involved in the increase of NLRP3 and pro-IL-1β levels induced by NETs. However, activation of the inflammasome might not have entirely depended on the elevated NLRP3 and pro-IL-1β levels.
TLR/NF-κB signaling pathway contributes to the up-regulation of NLRP3 and pro-IL-1β induced by NETs
Extracellular histones and DNA are widely recognized as endogenous damage-related molecular patterns that can be recognized by TLR receptors and exert a proinflammatory effect [41,42]. Histones may be recognized by TLR-4 and TLR-9. Unmethylated CpG motifs and phosphodiester DNA backbone efficiently dimerize TLR-9, following which downstream inflammatory pathways, such as NF-κB, are activated [43,44]. Since histones and DNA are the major components of NETs, we sought to determine whether TLR-4 and TLR9 acted as intermediaries between NETs and the activation of the NF-κB pathway. We observed that NETs enhanced the TLR-4 and TLR-9 levels (Figure 3E). When MΦ were costimulated with NETs and TLR antagonists, such as CLI-095 (TLR-4 antagonist) or ODN2088 (TLR-9 antagonist), together and separately, the TLR-4 and TLR-9 levels were reduced, indicating that the antagonists were active (Figure 3E). The antagonist treatment also resulted in a reduction of NLRP3, pro-IL-1β, and IL-1β levels (Figure 3E). We further noticed that the degree of reduction of NLRP3 and pro-IL-1β levels by TLR-4 and TLR-9 antagonists was similar. By contrast, the TLR-4 antagonist more efficiently reduced IL-1β levels than the TLR-9 antagonist (Figure 3E). Collectively, these observations suggested that the TLR-4/TLR-9/NF-κB signaling pathway contributed to the induction of the NLRP3 inflammasome-related proteins by NETs. TLR-4 antagonist more efficiently inhibited the IL-1β release than TLR-9 antagonist.
NETs maintain NLRP3 inflammasome activation by the reactive oxygen species/TXNIP signaling pathway
In addition to the priming signal, NLRP3 inflammasome activation requires a second stimulus . The data presented above indicated that the levels of active caspase-1, the main effector protein of inflammasome activation, were not reduced by an NF-κB blocker. Since the production of ROS is induced or required by nearly all NLRP3 agonists for NLRP3 inflammasome activation , we hypothesized that NETs may trigger NLRP3 inflammasome activation, which might not be entirely dependent on the elevated production of NLRP3 and pro-IL-1β. Indeed, compared with the control and DNase I-treatment groups, ROS levels were elevated in MΦ incubated with NETs (Figure 4A). Neutralization of ROS by a pretreatment with the antioxidant N-acetylcysteine (NAC) dramatically reduced the level of activated caspase-1 and the secretion of IL-1β in MΦ (Figure 4B). This suggested that ROS participated in the activation of NLRP3 inflammasome in MΦ stimulated by NETs. Consistently, intracellular NLRP3 and ASC colocalization was apparent after MΦ stimulation by NETs, while the colocalization was less pronounced in the control, DNase I-treated, and NAC-treated groups (Figure 4C).
NETs maintain NLRP3 inflammasome activation by the ROS/TXNIP signaling pathway
NLRP3 inflammasome will be activated by combining TXNIP . Generally, TXNIP interacts with thioredoxin (TRX). ROS can oxidize TRX and release TXNIP from TRX [46,47]. We sought to determine whether TXNIP participated in the NLRP3 inflammasome activation. We observed that NETs caused more interaction between TXNIPand NLRP3, which could be blocked by a ROS inhibitor (NAC), as well as DNase I treatment (Figure 4D). Collectively, these observations suggested that NETs triggered ROS generation in MΦ, allowing TXNIP to bind NLRP3 and activate the NLRP3 inflammasome.
Elimination of NETs by a topical DNase I treatment promotes wound healing by regulating the NLRP3 inflammasome and inflammatory cell infiltration
To verify the role of NETs in the inflammatory cascade in diabetic wound, we detected the presence of NETs in normal wounds, diabetic wounds, and DNase I-treated diabetic wounds of rats on day 2 and 5 after injury. NETs appeared and subsided quickly between day 2 and 5 after injury in the normal rat wound, while high amounts of NET were sustained in diabetic wounds. Exogenously supplemented DNase I efficiently degraded NETs (Figure 5A). By day 9 after the injury, Cit-H3 levels in wounds of normal rats, diabetic rats, and DNase I-treated diabetic rats were almost vanished and had not significantly difference between different groups (data not shown). The presence of NETs (DNA colocalized with Cit-H3) were demonstrated by intense staining of CD66b (neutrophil marker) using immunofluorescence staining. The expression of colocalized Cit-H3/CD66b in healthy and DNase I treated diabetic wounds was not as obvious as it in the untreated diabetic wound (Figure 5B). We compared protein levels of NLRP3 inflammasome components in normal wounds, diabetic wounds, and DNase I-treated diabetic wounds of rats on day 5, 9, and 14 after injury. Prominent levels of NLRP3, pro-IL-1β, IL-1β, ASC, and active caspase-1 were detected by immunoblotting in the diabetic rat wound (Figure 5C). The protein levels were much lower in the other wound types, implying that the elimination of NETs by DNase I salvaged the activation of NLRP3 inflammasome in the diabetic wound. Further, the levels of NLRP3 inflammasome-related proteins were generally reduced as the healing progressed (Figure 5C). These observations were verified by immunofluorescence staining of NLRP3 and CD68, with most CD68- and NLRP3-positive cells found in the diabetic wound on day 9 after injury (Figure 5D).
Elimination of NETs by a topical DNase I treatment regulates the NLRP3 inflammasome
We then compared the time-course of inflammatory cell infiltration by assessing the presence of CD66b (neutrophil maker) and CD68 (MΦ maker) in the wounds. All, CD66b-positive, and CD68-positive cells were counted, and average percentages of the stained cells were calculated. In healthy rat wounds, most neutrophils and MΦ were observed on day 2 and 5, respectively, and then the cell numbers quickly subsided (Figure 6A). By contrast, in the diabetic wounds, the neutrophils accounted for most cells and their numbers subsided more slowly than in healthy wounds; MΦ reacted to injury less robustly, reached a small peak on day 9, and subsided slowly, with a relatively higher proportion on day 14 than in healthy wounds. Notably, in DNase I-treated wounds, the proportion of neutrophils was lower and their numbers subsided more rapidly than in untreated diabetic wounds; MΦ were more numerous than in the healthy group on day 2 and 5, and then their numbers were rapidly reduced (Figure 6A). This indicated that DNase I might regulate the inflammatory response in the diabetic wound.
Elimination of NETs by a topical DNase I treatment promotes wound healing and inflammatory cell infiltration
Furthermore, we evaluated the healing process of healthy and diabetic rats, and investigated whether impaired healing in diabetic rats would be accelerated by the administration of DNase I. As shown by photographs of the representative healing wounds of healthy, diabetic, and DNase I-treated diabetic rats, DNase I treatment efficiently accelerated diabetic wound healing, alleviated inflammatory swelling, and reduced inflammatory cell infiltration on day 5 after injury (Figure 6B,D). Accelerated wound healing was also verified by rapid re-epithelialization on day 9 after wounding (Figure 6C). Collectively, diabetes primes NEs formation and NLRP3 inflammasome activation in the rat wound. Eliminating NETs by a topical DNase I administration reduced the NLRP3 inflammasome activation, reduced the inflammatory cell infiltration, and promoted diabetic wound healing.
Considering the extremely high prevalence of diabetes mellitus, chronic diabetic wounds contribute to high morbidity and mortality . It has been established that a sustained inflammatory response plays a key role in the pathogenesis of chronic diabetic ulcer . The sustained inflammatory response involves the accumulation of neutrophils and MΦ, and the release of proinflammatory products. The presence of NETs and the activation of NLRP3 inflammasome in macrophages are involved the pathological process of diabetic wounds [15,19,35]. However, whether NETs and NLRP3 inflammasome in MΦ act together and fuel the inflammatory cascade in the diabetic wound remain unclear. The purpose of the present study was to determine whether NETs trigger NLRP3 inflammasome activation and result in sustained inflammatory response in the diabetic wound. The major and novel findings of the current investigation are as follows: (a) NETs serve as activators of the NLRP3 inflammasome in MΦ; (b) induction of the NLRP3 and pro-IL-1β levels by NETs is dependent on the TLR-4/TLR-9/NF-κB signaling pathways; (c) ROS/TXNIP plays an essential role in the activation of NLRP3 inflammasome by NETs; and (d) elimination of NETs promotes wound healing by reducing the activation of NLRP3 inflammasome and mediating the infiltration of innate immune cells into the wound.
We detected elevated levels of Cit-H3 and NLRP3 inflammasome-related proteins in both human and rat diabetic wounds. We hypothesized that NETs play a role of regulated stimuli that promote NLRP3 inflammasome activation in MΦ. According to a recent in vitro study, NETs significantly induce the production of IL-1β in mouse MΦ-like J774 cells in the presence of LPS . The results presented in the present study are consistent with the conclusion that NETs promote IL-1β release. However, the data presented herein indicated that NETs alone are able to efficiently induce pro-IL-1β levels and elicit IL-1β production in MΦ. We believe that in addition to the different animal species and cell treatments investigated, this discrepancy may be cell source related. In the present study, we used primary MΦ from the peritoneal cavity, which might better reflect intrinsic MΦ properties than MΦ-like J774 cells. It has been also suggested that the AIM2 inflammasome, instead of NLRP3, is a source of IL-1β . In the present study, the utilization of NLRP3 specific inhibitor, MCC950, suggested specificity of NLRP3 inflammasome activation under the stimuli of NETs. On the other hand, reduction of IL-1β levels after a knockdown of NLRP3 was reported in a SLE study . Here, we demonstrated the NLRP3 inflammasome activation contributed to the IL-1β release induced by NETs.
A priming signal of NLRP3 inflammasome activation in MΦ induces the NLRP3 and pro-IL-1b levels through the NF-κB pathway [40,50]. Consistently, data presented in the present study indicated that the NF-κB signaling pathway participated in the induction of NLRP3 and pro-IL-1β levels by NETs in MΦ. As the major NET components, histones and DNA are widely identified as endogenous damage-related molecular patterns that are recognized by TLR receptors . Thus, we hypothesized that the induction of NLRP3 and pro-IL-1β was dependent on TLR-4 and TLR-9. Indeed, we presented evidence that the incubation of NETs with MΦ led to elevated TLR-4 and TLR-9 levels therein, whereas, specific antagonists reduced TLR-4 and TLR-9 levels, as well as those of NLRP3, pro-IL-1β, and IL-1β. TLR-4 antagonist inhibited IL-1β release more efficiently than TLR-9 antagonist. It was shown that TLR-4 is able to bind to the NADPH subunit of NOX4, which transduces proinflammatory signals by producing ROS . ROS, in turn, are at the center of the NLRP3 inflammasome activation and IL-1β release . This may explain why TLR-4 antagonist appeared to more efficiently reduce IL-1β levels than the other compound tested.
Further, the role of histone and free DNA in TLR and inflammasome activation remains unanswered. The effects of histones, free DNA, and their combination in nucleosomes on these processes have been examined by several studies. For example, intravenous injection of histones is lethal to mouse within minutes, while an injection of antihistone antibodies or Tlr4 knockout counteracts high levels of TNF-α, IL-6, and IL-10 and reduces mortality [53,54]. Lupus DNA-containing immune complexes (SLE-ICs) purified from the serum of patients were shown to play important roles in the pathogenesis of lupus. Under SLE-ICs stimulation, TLR-9 was found to be responsible for the high levels of IFN-α, TNF-α, and IL-18 . Furthermore, Tlr9 knockout mouse is protected from histone-mediated NLRP3 inflammasome activation during liver ischemia/reperfusion injury . That may be attributed to histone binding the DNA to facilitate its uptake, amplified by TLR-9 stimulation. These observations imply that histones and DNA act together to activate TLR and the inflammasome.
As shown in the present study, an NF-κB blocker did not reduce the levels of active caspase-1, the main effector protein of inflammasome activation. We hypothesized that NETs trigger the assembly and activation of NLRP3 inflammasome, which might not be entirely dependent on the elevated NLRP3 and pro-IL-1β levels, considering that ROS production is induced and required by nearly all NLRP3 agonists for NLRP3 inflammasome activation . NAC, ROS scavenger, notably reduced the levels of activated caspase-1 and IL-1β, and also resulted in less pronounced colocalization of NLRP3 and ASC in MΦ stimulated by NETs, verifying that ROS were indeed involved in NLRP3 inflammasome activation.
ROS oxidizes TRX and causes TXNIP to separate from TRX, upon which TXNIP combines with NLRP3 and activates the NLRP3 inflammasome. The opinions about the role of TXNIP in NLRP3 inflammasome activation in MΦ are diverse. According to one study, IL-1β secretion by bone marrow-derived MΦ isolated from TXNIP-deficient mouse is not different from that of wild-type MΦ . By contrast, other studies indicate that the TXNIP/NLRP3 association is linked to IL-1β secretion in MΦ. For example, dissociation of TXNIP and TRX is evident after treating MΦ with MSU or R-837 and is ROS-dependent. Simultaneously with TXNIP/TRX dissociation, the TXNIP/NLRP3 interaction was detected with a subsequent release of IL-1β . Further, TXNIP siRNA alleviates the IL-1β and caspase-1 levels in MΦ stimulated by fructose . More recently, increased expression of TXNIP, NLRP3, and IL-1β was observed in MΦ of individuals with type 2 diabetes . Similarly, in the present study, we observed increasing association of TXNIP and NLRP3 in MΦ during incubation with NETs, which could be alleviated by NET digestion or ROS neutralization.
Finally, we confirmed that eliminating NETs by DNase I treatment accelerated healing of the diabetic wound whose healing process in impaired. This was in agreement with a previous report of Wong et al. . We first observed that NET elimination by DNase I administration alleviated the elevated levels of NLRP3, pro-IL-1β, IL-1β, ASC, and active caspase-1 in the diabetic wound in rats, which verified that the activation of NLRP3 inflammasome was stimulated by NETs in vivo. Moreover, we observed that NETs mediated the inflammatory response by inducing the infiltration of innate immune cells into the diabetic wound. Upon DNase I administration, the neutrophils infiltrated the tissue to a lower extent, with a rapid regression. Further, after DNase I treatment, MΦ infiltrate the tissue earlier to a higher extent and their numbers fell more rapidly than in untreated diabetic wounds. Whether the effects of NETs on the innate immune cell infiltration are associated with a direct NET-mediated toxicity or via enhanced NLRP3 inflammasome activation remains unknown. However, reduced infiltration of dendritic cells, neutrophils, and inflammatory monocytes in Nlrp3 knockout mouse was reported in studies into liver ischemia/reperfusion injury [56,60]. Similarly, reduction of the levels of proinflammatory factors, neutrophils, and MΦ in Nlrp3 or Il1R knockout mice was reported in studies into diabetic wound healing [19,61]. Collectively, the published and new observations imply that NETs recruit innate immune cells by activating NLRP3 inflammasome. However, the exact underpinning mechanism requires further study.
In conclusion, in the present study, we first showed that the NLRP3 inflammasome was activated by NETs in the chronic diabetic wound (Figure 7). Furthermore, we demonstrated that NETs induced the NLRP3 and pro-IL-1β levels via the TLR-4/TLR-9/NF-κB pathway, and activated NLRP3 inflammasome via the ROS/TXNIP pathway. NET activity and MΦ activation were ultimately attributed to the infiltration of innate immune cell into the diabetic wound. These findings illustrate a novel mechanism whereby NETs may lead to unhealed wound formation and suggest potential therapeutic targets for accelerating diabetic wound healing.
Hypothetical model of NET participation in sustained inflammation in the diabetic wound by activation of NLRP3 inflammasome in MΦ
Diabetic foot ulcer results in one patient losing a lower limb or part of a lower limb for amputation every 30 s globally. NETs and NLRP3 inflammasome were involved in diabetic foot ulcer, while interaction between them is elusive. We investigated the role of NETs and NLRP3 inflammasome in the sustained inflammatory response of the diabetic wound.
The present study show that NETs overproduced in the diabetic wounds. NETs induce NLRP3 inflammasome activation in MF through TLR/NF-kB and ROS/TXNIP pathways.
DNase I alleviated the activation of NLRP3 inflammasome, and accelerated wound healing in diabetic rat model. These suggest a potential therapeutic target for accelerating diabetic wound healing and will be easily translated into clinical practice.
This work was supported by the Doctorial Innovation Fund of Shanghai Jiao Tong University School of Medicine [BXJ201813]; National Natural Science Foundation of China ; Shanghai Committee of Science and Technology, China [18ZR1423800]; Research Grant from Shanghai Hospital Development Center [SHDC12014117]; Shanghai Collaborative Innovation Center for Translational Medical [TM201705]; and the National Natural Science Foundation of China . The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
The authors declare that there are no competing interests associated with the manuscript.
Dan Liu designed the research study and analyzed most of the data and wrote the draft of the paper; Peilang Yang and Dan Liu performed Western blot and IF experiment; Min Gao performed animal experiment, including wound procedure and healing examination; Tianyi Yu neutrophils and macrophages isolation; Yan Shi and Min Yao collected the wound tissue of the health and diabetic patients; Meng Zhang prepared tissue section; Xiong Zhang and Yan Liu performed the microscopy, revised the manuscript and contributed to the overall conclusions.
apoptosis-associated speck-like protein
diabetic foot ulcer
neutrophils release extracellular trap
Nod-like receptor protein
peptidyl arginine deiminase-4
reactive oxygen species
systemic lupus erythematosus
This author contributed equally to this work.