NETs Enhance STING To Promote Surgical Adhesion


 Background

Postoperative adhesion (PA) following abdominal surgery may cause bowel obstruction, chronic pain, infertility, or even death. Knowledge of adhesion biology is limited, and preventive agents in clinical trials have failed to achieve efficacy.
Results

In the present study, we showed that neutrophils accumulate in the injured peritoneum at early stage of PA, and neutrophils within the ischemic buttons undergo cell death and form neutrophil extracellular traps (NETs) that contribute to PA. Neutrophil depletion reduces adhesion burden at 7 days after adhesion induction. Peptidylarginine deiminase 4 (PAD4), an essential enzyme for NET formation, is increased in ischemic buttons. Degradation of NETs by DNase 1 and suppression of NET formation by pharmacologic inhibition of PAD4 alleviated adhesion burden, collogen deposition and fibrosis formation. Mechanistically, administration of DNase I and PAD inhibitor reduces STING-mediated inflammatory response. STING deficiency attenuates adhesion burden, collogen deposition, and α-SMA production in the adhesive tissues at 7 days after surgery.
Conclusions

Collectively, our findings reveal NETs/STING signaling as a therapeutic target to prevent PA.


Introduction
Postoperative adhesion (PA) is the most common complication of abdominal surgery, which can lead to intestinal obstruction, abdominal pain, female infertility, and even death [1]. Repeated surgery including adhesiolysis can cause large burden of morbidity [2]. However, effective strategy is unavailable to prevent or treat PA [3]. Deep understanding of molecular mechanisms of PA process allows the design of new anti-adhesion treatment.
Neutrophils are recognized as the earliest leukocytes in the response to tissue damage [4]. Adhesion tissue was obtained from adhesive small bowel obstruction (ASBO) patients, and pathological staining indicated that in ltrating neutrophils within thick peritoneum was accompanied by increased DNA bers structure ( Supplementary Fig. 1), which indicated NETs (neutrophil extracellular traps, NETs) formation. NETs contain double-stranded DNA, histone, and granule proteins including neutrophil elastase, cathepsin G, and myeloperoxidase (MPO) [5]. These NETs were associated with immune disorder, in ammation, and brosis [6]. However, whether NETs contribute to favorable or poor outcomes during PA formation remains unclear. A recent study found that delayed NETs formation was showed in the adhesive tissue in mouse [7]; However the underlying mechanism in NETs formation and their cascade effect remain largely unknown during PA development.
Mechanism during later stage of PA formation involving brin deposition and brinolysis has been wellstudied; In contrast, the cellular and molecular detail during initial stage requires extensive elucidation [8].
In this study, we discover early neutrophil-dependent NETs formation in injured peritoneum in a murine PA  Fig.1a). Informed consent was obtained from all subjects after surgery.

Adhesion induction
Adhesion-induction surgery was performed as previously described [8]. In brief, a small portion of the peritoneum was grasped and ligated at its base using 4-0 silk suture (Ethicon), creating a standardized peritoneal button. This was repeated for a total of four buttons (one in each quadrant). Light abrasion on each button and on the adjacent liver, cecum, and intestine was performed using surgical brash. Finally, peritoneum and skin were closed using 4-0 silk suture. Mouse ischemic button model is easily reproducible to induce intra-abdominal adhesions compared to other models as previously demonstrated [9]. The experimental procedure in this study was approved by the Animal Care and Use Committee of Nanjing Drum Tower Hospital.

Histology, IHC and immuno uorescence
Tissues were xed in 2% PFA overnight at 4°C and were embedded in para n. para n sections were cut at 5μm. H&E, Masson's trichome stains, and immunohistochemistry (IHC) were performed following standard protocols.
For IHC staining, slides were incubated with anti-APD4 antibody, anti-citH3 antibody, and anti-α-SMA antibodies according to the manufacture's instruction. Nuclei were stained with DAPI (Invitrogen).
Immuno uorescence experiment was performed on para n sections. NETs associated proteins were evaluated by immuno uorescent staining using rabbit mouse anti-mouse MPO, and rabbit anti-mouse citH3 (Abcam, Cambridge, UK). DAPI was use to counterstain DNA. Images were acquired using a Leica DMI 4000 B uorescence microscope.

ELISA assay
Enzyme-linked immunosorbent assays (ELISA) was performed to evaluate the production of in ammatory cytokines, including IL-1β, IL-6, and TNF-α according to the manufacturer's instructions.

Neutrophil depletion
Mice were injected with 100 µg intraperitoneal (i.p.) monoclonal anti-mouse Ly6G at 12h or 3d after induction of ischemic button. Rat IgG2a isotype control was administered in the same way. To verify neutrophil depletion, blood neutrophil levels were evaluated by ow cytometry and analyzed using FlowJo V10 software as previously described [10].

DNase I and PAD inhibitor administration
Mice received i.p. injection of 50 μg DNase I at 2h, 12 h or 3d after induction of ischemic button. Control counterparts were injected with PBS. PAD inhibitor Cl-amidine was dissolved in dimethyl sulfoxide (DMSO) as previously described. The stock solution was dissolved in saline and injected i.p. (10 mg/kg) at 2h, 12h and 3d after adhesion induction. Vehicle (DMSO) was administered as control (Fig.3a).

Western blot
The peritoneal ischemic buttons were dissected after the mice were sacri ced, and the protein from adhesive tissues and HPMCs were extracted using RIPA lysis buffer. Proteins were separated by 10% SDS-PAGE, and transferred to PVDF membrane. After blocking with 5% milk, the membranes were incubated with anti-Ly6G, anti-PAD4, anti-citH3, anti-p-TBK1 and anti-p-IRF3 at 4 °C overnight. GAPDH served as internal control.

Statistical analysis
All statistical analyses were evaluated using SPSS 19.0 and GraphPad Prism 6. All data were presented as mean ± SEM. Unpaired t test and one-way analysis were used to compare continuous variables between groups. P<0.05 was considered signi cant.

Results
Early neutrophil in ltration causes peritoneal structure damage in murine PA Mice were subjected to ischemic button and analyzed at 6h, 12h and 24h. The damaged peritoneum was markedly thickened and showed collagen deposition by H&E and Masson staining (Fig. 1a). Increased polymorphonuclear leukocyte were observed in the buttons compared with sham group. MPO staining showed that neutrophils were visualized within thicken peritoneum after adhesion induction (Fig.1b). A 5-fold increase in total content of the neutrophil enzyme MPO was found in the buttons at 24h after surgical trauma (Fig.1c). Using Ly6G antibody, the levels of neutrophil continued to increase within 24h after adhesion induction (Fig.1d).
To investigate the role of neutrophils in brosis development during delayed phases of PA, we depleted neutrophil using anti-Ly6G antibody. Peripheral white blood cell counts and neutrophils were signi cantly decreased by anti-Ly6G treatment ( Supplementary Fig. 2a,2b). Correspondingly, H&E and immunostaining showed that neutrophil recruitment in the ischemic buttons was signi cantly reduced in anti-Ly6G-treated mice compared to IgG-treated mice ( Supplementary Fig. 2c,2d). Mice receiving anti-Ly6G antibody after injury exhibited a signi cantly decreased adhesion burden compared with controls ( Supplementary  Fig.2e).

PA induces early NETs formation
We subsequently elucidate whether NETs formation occurs in ischemic buttons and affects adhesion outcomes. IHC and Western blot analysis showed an increased production of PAD4 and citH3 within the ischemic buttons at 24h, indicating that NETs may play a vital role at the early stage of PA (Fig.2a, 2b).
Microscopic analysis detected positive staining of citH3 and MPO, overlapping with diffused DNA scaffolds, that indicated increased formation of NETs structure in the buttons (Fig.2c). Interestingly, colocalization of citH3 and MPO was not observed at 6h but appeared at 12h following adhesion induction ( Supplementary Fig.3a).

Disruption of NETs structure and inhibition of NETs formation alleviate adhesion formation
We next investigate whether disruption of NETs could alleviated PA formation in mice. PAD4 acts as a driver of NETs formation. Treatment with DNase I and Cl-amidine (PAD4 inhibition) failed to affect the amount of neutrophils in the ischemic buttons ( Supplementary Fig. 4a, 4b) but signi cantly reduced citH3 levels con rmed by IHC staining (Fig. 3b). Immuno uorescence staining for citH3 and MPO in ischemic buttons at 24h found that administration of either DNase I or Cl-amidine could signi cantly prevent NETs formation in injured peritoneum. In addition, adhesion burden, collogen deposition and brosis formation were signi cantly reduced following DNase I and Cl-amidine treatment after 7 days compared with controls ( Supplementary Fig.4c and Fig.3d).
NETs are responsible for STING-mediated PA formation An increased production of in ammatory mediators in the early stage leads to extracellular matrix formation during PA, and NETs are associated with increased in ammation [11]. We next evaluated whether NETs regulate in ammatory response. In ammatory mediators (IL-1β, IL-6 and TNF-α) were signi cantly suppressed in the ischemic buttons following DNase I treatment at 24h after surgery ( Supplementary Fig.5a).
We next investigated whether STING-de cient mice were more resistant to PA formation. Adhesion burden, collogen deposition, and α-SMA production were signi cantly decreased in the adhesive tissues at 7d in STING-de cient mice after surgery ( Fig.4e and Supplementary Fig.6a). Interestingly, we unexpectedly found that STING de ciency was able to reduce NETs formation within the ischemic buttons at 24h, suggesting a contribution of STING signaling to NETosis (Supplementary Fig.7a).

Discussion
PA develops in most patients receiving abdominal surgery. Despite its clinical importance, cellular and molecular processes and preventive strategy of PA remains largely unknown [9]. Previous studies suggested that neutrophil chemotaxis played an important role in exacerbating adhesion [8,13]. A recent study suggested that delayed NETs formation occurred in adhesive tissues surface [7]. In our study, early neutrophil in ltration and NETs formation were showed in injured peritoneum. Collectively, neutrophil accumulation and NETs formation are essential for both early and late stages of PA formation.
Neutrophils are important in contributing to the transition from pro-in ammatory response to pro-brotic condition [14]. Our recent study showed accumulation of neutrophils in adhesive tissues during all stages of adhesion, and the mount of neutrophils reached to the peak point within 1-2 days [4]. Therefore, understanding the molecular function of neutrophils may help explain the formation of adhesions.
Indeed, neutrophil depletion reduced adhesion burden at 7 days after adhesion induction. Reduction in circulating neutrophils could possibly lead to immune compromise state during peri-operative period.
Further studies towards the consequence of neutrophil depletion on side effects after surgery will be essential.
Neutrophils release NETs-associated cytotoxic proteases such as histone, elastase and MPO, which was associated with in ammatory response, collagen production ECM deposition [14,15]. NETs could also promote differentiation and function of broblasts, that contribute to organ brosis [15,16]. Our data suggested that activated neutrophils release excessive NETs at injured peritoneum. Digestion of NETs with DNase I signi cantly decreased in ammatory response, collagen production and brosis formation.
These ndings strongly indicated that NETs could be critical for neutrophil-dependent in ammation and pro-brotic condition during early and late stage of PA process.
PAD4 is an essential enzyme for NETs formation [17]. We demonstred that PAD4 was signi cantly increased in the ischemic buttons after surgery. Consistently, we found that PAD4 inhibition suppressed early NETs formation, and subsequently alleviated collagen deposition, brosis formation, and adhesion burden. These data suggest that PAD4 depletion could ameliorate brosis development.
STING is an vital regulator of tissue in ammation and brosis in multiple organ, including lung [18], liver [19], kidney [20] and heart [21]. Our study suggested that PA activates the DNA sensor STING and induces the production of in ammatory cytokines. STING depletion in mice decreased collogen deposition and α-SMA production. In addition, degradation and inhibition of NETs suppressed the activation of STING signaling. Recent studies also highlight that NETs are responsible for STING activation during in ammatory or immune diseases [11,22]. Therefore, STING-mediated in ammatory response may link NETs and PA formation. Interestingly, we found that depletion of STING in turn inhibited the formation of NETs structure. A recent study also suggested that STING pathway may contribute to the induction of NETs by mtDNA stimulation [23]. Therefore, it possible that NETs/STING pathway may form a vicious cycle during early stage of adhesion, which eventually contribute to brosis and PA formation.

Conclusion
Our study revealed that early neutrophils recruited in injured peritoneum that further released NETs structure, and subsequently activated STING-associated in ammatory response. Furthermore, we demonstrated that disruption of NETs and inhibition of STING signaling could signi cantly decrease the adhesion burden. Our data suggested that NETs/STING pathways could serve as targets for PA     NETs are responsible for STING-mediated adhesion formation. (a) Immunoblot analysis of STING, pTBK1, and pIRF3 in the ischemic buttons 24 hours after surgery, and treatment with anti-Ly6G antibody (b) or Cl-amidine (c). H&E staining of adhesive tissue between WT mice and STING-/-mice.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.