Abstract
To evaluate the role of neutrophil extracellular traps (NETs) in the genesis of joint hyperalgesia using an experimental model of arthritis and transpose the findings to clinical investigation.
C57BL/6 mice were subjected to antigen-induced arthritis (AIA) and treated with Pulmozyme (PLZ) to degrade NETs or Cl-amidine to inhibit NET production. Oedema formation, the histopathological score and mechanical hyperalgesia were evaluated. NETs were injected intra-articularly in wild type (WT), Tlr4−/−, Tlr9−/−, Tnfr1−/− and Il1r−/− mice, and the levels of cytokines and Cox2 expression were quantified. NETs were also quantified from human neutrophils isolated from RA patients and individual controls.
AIA mice had increased NET concentration in joints, accompanied by increased Padi4 gene expression in the joint cells. Treatment of AIA mice with a peptidyl arginine deiminase 4 inhibitor or with PLZ inhibited the joint hyperalgesia. Moreover, the injection of NETs into joints of naïve animals generated a dose-dependent reduction of mechanical threshold, an increase of articular oedema, inflammatory cytokine production and cyclooxygenase-2 expression. In mice deficient for Tnfr1, Il1r, Tlr4 and Tlr9, joint hyperalgesia induced by NETs was prevented. Last, we found that neutrophils from RA patients were more likely to release NETs, and the increase in synovial fluid NET concentration correlated with an increase in joint pain.
The findings indicate that NETs cause hyperalgesia possibly through Toll-like receptor (TLR)-4 and TLR-9. These data support the idea that NETs contribute to articular pain, and this pathway can be an alternative target for the treatment of pain in RA.
Neutrophil extracellular traps (NETs) cause pain in antigen-induced arthritis.
Toll-like receptors are involved in hyperalgesia induced by NETs.
Introduction
RA is a chronic autoimmune disease that affects ∼0.5–1% of the world adult population [1] and is characterized by the presence of autoantibodies [2, 3] and symmetric inflammation of joints, leading to joint pain and cartilage and bone damage [4]. The aetiology of RA is not yet fully elucidated, but it is known that it involves a combination of genetic, environmental and hormonal factors [5–7]. One of the main symptoms of RA is increased sensitivity to joint pain, called hyperalgesia, which leads to loss of quality of life [8]. Despite advances in RA therapy, treatment for joint inflammation and consequent pain relief remains limited for many patients.
Several groups, including ours, have demonstrated the importance of pro-inflammatory cytokines triggering joint hyperalgesia in different animal models, including in the RA model [9, 10]. Cytokines, such as TNF-α and IL-1β, and chemokines released by immune cells stimulate the production of eicosanoids that sensitize the primary afferent neurons [11].
Cytokines also induce the recruitment of neutrophils to the joints, where they are also an important source of inflammatory mediators, including those involved in hyperalgesia. In fact, the blockage of neutrophil migration to inflamed joints prevents the majority of the inflammatory events, including joint pain [11]. In addition to interacting with other immune cells [12], neutrophils release enzymes, such as myeloperoxidase (MPO) and matrix metalloproteinases, and oxygen- and nitrogen-derived free radicals, which lead to joint tissue lesions, including cartilage and bone damage [13, 14]. They also release eicosanoids, which are involved in the sensitization of primary neurons [11]. In the past decade, neutrophils have been demonstrated to release neutrophil extracellular traps (NETs), which are composed of an extracellular DNA matrix containing cytotoxic enzymes and histones. During inflammation, NETs play a pivotal role in the host killing capacity. In fact, NETs kill different bacterial strains, fungi, protozoa and viruses [15]. However, NETs are also involved in other pathophysiological processes, including autoimmune diseases. Inhibition of peptidyl arginine deiminase 4 (PADI-4), an important enzyme involved in NETosis, reduces the severity of experimental RA [16]. It was demonstrated that NETs mediate tissue damage in different experimental models of autoimmune diseases [17, 18]. Furthermore, besides these effects, NET components are recognized as damage-associated molecular patterns (DAMPs), turning release of NETs into a magnifying process for the immune response in the development of autoimmunity [19].
Although NET production during experimental and clinical RA is well established [20], there is no evidence that NETs participate in the onset of joint hyperalgesia in RA. Therefore, in the present study, we address the possible participation of NETs in joint hyperalgesia using an experimental model of RA. We also investigate the possible association between the intensity of pain and the levels of NET production in clinical RA.
Methods
Patients
Peripheral blood samples were collected from 24 RA patients recruited from the Ribeirão Preto Medical School Hospital (University of São Paulo, Ribeirão Preto, Brazil). All patients fulfilled the 2010 ACR/EULAR classification criteria for RA [21] and had <2 years of symptoms. Thirteen healthy volunteers matched for age and gender were recruited as controls. The research was conducted according to the Human Ethics Committee of the University of São Paulo at Ribeirão Preto Medical School (document no. 2.724.944, Presentation Certificate for Ethical Appreciation no. 89092417.3.0000.5440). The visual analogue scale (VAS; 0–100 mm) was used to measure pain intensity in patients [22].
Human neutrophil isolation
Neutrophils were isolated from peripheral blood from RA patients and control individuals through Percoll gradients (72%, 63%, 54% and 45%). After centrifugation, polymorphonuclear leukocytes accumulated as a band between 72 and 63% Percoll. Total cell numbers were counted using light microscopy. The percentage of neutrophils was determined microscopically through haematoxylin and eosin staining. One million neutrophils were incubated in HBSS medium (Corning, Manassas, VA, USA) for 4 h, and the supernatants were used to determine NET concentration.
Animals
The experiments were performed on male C57BL/6 wild type (WT) mice and mice deficient in the genes for Toll-like receptor 4 (Tlr4; C57BL/6 background), Toll-like receptor 9 (Tlr9; C57BL/6 background), TNF receptor 1 (Tnfr1; C57BL/6 background) and IL-1 receptor (Il1r; C57BL/6 background), weighing between 20 and 25 g. They were housed in temperature-controlled rooms (22–25°C) and given water and food ad libitum at the animal facility in the Department of Pharmacology, School of Medicine of Ribeirão Preto, University of São Paulo, Brazil. Animal husbandry and procedures were in accordance with the guidelines of the Animal Ethics Committee of the School of Medicine of Ribeirão Preto (241/2018).
Induction of experimental antigen-induced arthritis
The induction of antigen-induced arthritis (AIA) was performed as previously described, with some alterations [23]. Briefly, the mice were sensitized with 500 µg of methylated bovine serum albumin (mBSA) in 0.2 ml of an emulsion containing 0.1 ml saline and 0.1 ml complete Freund’s adjuvant (1 mg/ml of Mycobacterium tuberculosis) and given by subcutaneous injection on day 0. The mice were boosted with the same preparation on day 7. Control mice received similar injections but without the mBSA. Twenty-one days after the initial injection, arthritis was induced in the immunized animals by intra-articular (i.a.) injection of mBSA (100 µg/cavity) into the right knee joint or into both knee joints. On day 26 after the first immunization, mice were challenged again the same way. Six, 12 and 24 h after the second challenge, the NET concentration was dosed. Twenty-four hours after the second challenge, the knee joint thicknesses were measured by calliper in millimetres. The results were expressed as the means and the difference between the diameter on day 0 (basal) and after second challenge (Δmm). The knee joints were opened and washed with PBS for quantification of NETs. The gene expression of Padi4 in these aspirated fluids was evaluated 24 h after the second challenge. The knees were fixed in 10% buffered formalin for histopathological analysis. An electronic von Frey test was performed as previously described [24] at 3, 6 and 24 h after the second challenge to evaluate hyperalgesia.
Purification and intra-articular injection of NETs
NETs were obtained as previously described [25]. Briefly, isolated bone marrow from WT mice was used to isolate neutrophils with Percoll gradients (72% and 65%). These neutrophils (98% purity) were challenged with 50 nM of phorbol 12-myristate 13-acetate for 3 h at 37°C in RPMI 1640. Supernatants were discarded, and NETs were removed through extensive pipetting with fresh PBS. The NETs were centrifuged to remove cellular components, and NETs-containing supernatants were collected and centrifuged. Supernatants were removed and the pellets resuspended in PBS. NETs were then quantified through GeneQuant (GE Healthcare Bio-Sciences Corp., Sheffield, UK). WT mice received three different doses of NETs (5, 10 and 20 ng) in the right joint or received 10 ng of NETs incubated in vitro with Pulmozyme® (Pulmozyme®, San Francisco, California, USA) (1:1) for 3 h at 37°C. Knockout mice for Tlr4, Tlr9, Tnfr1 and Il1r received 10 ng of NETs in the right joint.
Immunofluorescence of NETs
Purified NETs (50 µl) were spread on polylysine-treated glass slides and then fixed with paraformaldehyde 4%. Slides were blocked with a solution containing BSA 2% and 0.1% Triton X-100 and afterward incubated with a rabbit anti-citrulline H3 (1:1000) (Abcam, Cambridge, UK) and a mouse anti-MPO (Abcam) (1:1000). Detection was performed using an anti-rabbit (AlexaFluor 488) (Thermo Fisher Scientific, Waltham, MA, USA) and an anti-mouse antibodies (AlexaFluor 594) (Thermo Fisher Scientific) (1:2000 each). Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; Vectashield® Antifade Mounting Medium; Vector Laboratories, Burlingame, CA, USA). Fluorescence images were acquired on a confocal microscope (Carl Zeiss, Oberkochen, Germany) (Supplementary Figs S1a and b, available at Rheumatology online). Secondary antibody only and IgG controls were performed to guarantee that staining was not background (Supplementary Fig. S2a and b, available at Rheumatology online). Confirming that citrullination is expressed in NET, we assessed colocalization analysis between DAPI and H3 citrulline, and DAPI and MPO, in 10 different samples, by performed Pearson's correlation coefficient using ImageJ software for quantification. The co-localization of H3 citrulline was observed (Supplementary Fig. S1c, available at Rheumatology online).
Determination of articular hyperalgesia
Articular hyperalgesia of the femur–tibia joint was determined as previously described [24]. An increasing perpendicular force was applied to the central area of the plantar surface of the hind paw to induce flexion of the femur–tibia joint followed by paw withdrawal. The pressure of the force applied when the paw was withdrawn was recorded by a pressure meter. The test was performed by investigators blinded to the treatment and repeated until three consistent consecutive measurements (variation <1 g) were obtained. Mechanical threshold is expressed in grams, and hyperalgesia is equated to reduction of this threshold.
Cytokine measurements
Two hours after injection of NETs in WT, Tlr4−/− and Tlr9−/− knockout mice, the animals were terminally anaesthetized, and the knee joints were removed and homogenized in 500 µl of buffer containing protease inhibitors. TNF-α and IL-1β concentrations were determined as described previously [26] by ELISA using paired antibodies (R&D Systems, Minneapolis, MN, USA). The results are expressed as pg/mg of each cytokine.
Histological analysis
Femur–tibia joints were collected 24 h after the second challenge with mBSA, fixed in 10% buffered formalin for 2 days, decalcified in 14% ethylenediaminetetraacetic acid (pH 7.4) for 21 days and processed for histology. The paraffin blocks were cut into sagittal sections of 4 µm thickness. At least three serial vertical sections stained with haematoxylin and eosin were evaluated for each animal. Three independent observers who were blinded to the treatment graded the extent of synovitis (inflammatory cell influx, synovial hyperplasia and bone loss) as previously described [23]. The histological scoring system used was as follows: none (0), slight (1), mild (2), moderate (3) or severe (4) for inflammatory cell influx; synovial hyperplasia was scored 0–3, and bone loss was scored as present (1) or absent (0). The grades were summed to obtain an arthritis index (ranging from 0 to 8, with the results expressed as the mean histopathological score).
NET quantification
Articular cavity washes from mice and neutrophils isolated from blood RA patients and control individuals were used to quantify NETs. Briefly, an antibody bound to a 96-well clear-bottom black plate captured the enzyme MPO (PA5-16672, Thermo Fisher Scientific), and the amount of DNA bound to the enzyme was quantified using the Quant-iT™ PicoGreen® kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. The fluorescence intensity (excitation at 488 nm and emission at 525 nm wavelength) was determined by a FlexStation 3 microplate reader (Molecular Devices, San Jose, CA, USA).
Reverse transcription–polymerase chain reaction assays
For the evaluation of Padi4 gene expression, AIA animals and their controls were euthanized 24 h after the second challenge with mBSA, and the joint washes of these animals were collected for RNA extraction using a Qiagen kit (Qiagen, Hilden, Germany) according to the directions supplied by the manufacturer. For the evaluation of Cox2 gene expression, synovial membranes were collected from naïve animals, WT animals and Tlr4, Tlr9, Tnfr1 and Il1r knockout mice 6 h after receiving 10 ng of i.a. NETs. Total cellular RNA from the synovial membrane was extracted using Trizol reagent (Thermo Fisher Scientific) according to the directions supplied by the manufacturer. Real-time quantitative PCR mRNA analysis was performed on an ABI Prism 7500 Sequence Detection System using the SYBR-green fluorescence system (Applied Biosystems, Foster City, California, USA) for quantification of amplicons. The primers used were the following: glyceraldehyde 3-phosphate dehydrogenase (Gapdh) forward: 5′-GGGTGTGAACCACGAGAAAT-3′; Gapdh reverse: 5′-CCTTCCACAATGCCAAAGTT-3′; Padi4 forward: 5′-TGACCAATGGATGCAGGACG-3′; Padi4 reverse: 5′-CTCTGTCCCTCGGGGAGTC-3′; Cox2 forward: 5′-GTGGAAAAACCTCGTCCAGA-3′; Cox2 reverse: 5′-GCTCGGCTTCCAGTATTGAG-3′.
Statistical analysis
The data are reported as the means (s.e.m.) and are representative of two separate experiments. Two-way ANOVA was used to compare the groups and doses at all times when the hyperalgesic responses were measured at different times after the stimulus injection. The means from different treatments were compared by one-way ANOVA with a Bonferroni correction or by Student’s t-test. Spearman’s rank-order correlation (r) was calculated to describe correlations between NET concentration from human neutrophils from RA patients and pain. P values <0.05 were considered significant. The experiments were repeated three times.
Results
NETs mediate joint edema and mechanical hyperalgesia in antigen-induced arthritis
First, we induced experimental arthritis through the AIA model with two challenges of mBSA (100 µg/joint) (Fig. 1A). The administration of mBSA (100 µg/joint) into the articular joint of immunized C57BL/6 mice induced a significant production of NETs compared with the control (saline injected i.a.) group. mBSA was administered into the joint injected 5 days prior with the same dose of mBSA, and the NET concentration in synovial fluid was determined 6, 12 and 24 h after the joint challenge. Compared with the controls (animals treated with saline), significant increases in the concentration of NETs were observed 12 h after the joint challenge, which was further enhanced at 24 h. The levels of NETs determined at 24 h were prevented by treatment of AIA animals with Pulmozyme (PLZ, 10 µg/joint) or Cl-amidine (1 mg/kg i.p.) 1 h before and 1 h after the second mBSA challenge (Fig. 1B). In addition, AIA animals had increased Padi4 gene expression in cells that emigrated into the articular joint, indicating that the increase in NET production may be associated with Padi4 expression (Fig. 1C). The pretreatment of AIA animals with PLZ or Cl-amidine reduced the oedema (Fig. 1D) and reduced the average histopathological score (Fig. 1E and F). Furhermore, PLZ and Cl-amidine treatment also reduced the joint hyperalgesia induced by mBSA (Fig. 1G).
NETs participate in the AIA model
(A) AIA mice were immunized and challenged with mBSA twice, and the treatment with PLZ (10 µg/joint) was administrated 1 h before and 1 h after second challenge. (B) NET was quantified 6, 12 and 24 h after the second challenge with mBSA in AIA mice. Twenty-four hours after second challenge, NETs from AIA+PLZ mice were quantified. Mean (s.e.m.), n = 5/group. *P < 0.05 vs CT group; #P < 0.05 vs AIA group, one-way ANOVA followed by Bonferroni post hoc test. (C) Genic expression of Padi4 in joint neutrophils was evaluated 24 h after the second challenge. Mean (s.e.m.), n = 5/group. *P < 0.05 vs CT group, two-sided Student’s t-test. (D) Joint swelling was evaluated 24 h after second challenge. Mean (s.e.m.), n = 5/group. *P < 0.05 vs CT group; #P < 0.05 vs AIA group, one-way ANOVA followed by Bonferroni post hoc test. (E) Femur–tibia joints were collected 24 h after second challenge with mBSA, fixed and decalcified in EDTA (pH 7.4) for 21 days and processed. The paraffin blocks were cut into sagittal sections of 4 µm thickness. (F) Histopathological score representation. Mean (s.e.m.), n = 5/group. *P < 0.05 vs CT group; #P < 0.05 vs AIA group, two-sided Student’s t-test followed by Mann–Whitney test. (G) Articular hyperalgesia was evaluated 3, 6 and 24 h after the second challenge with mBSA. Mean (s.e.m.), n = 5/group. *P < 0.05 vs CT group; #P < 0.05 vs AIA group, two-way ANOVA followed by Bonferroni post hoc test. The experiments were repeated three times. AIA: antigen-induced arthritis; CT: control; EDTA: ethylenediaminetetraacetic acid; Gapdh: glyceraldehyde 3-phosphate dehydrogenase; mBSA: methylated bovine serum albumin; NETs: neutrophil extracellular traps; Padi4: peptidyl arginine deiminase 4; PLZ: Pulmozyme.
Participation of cytokines in NET-induced joint hyperalgesia
To add to the mechanism by which NETs mediate joint hyperalgesia, we investigated whether the administration of purified NETs into joints could induce hyperalgesia. Compared with saline, i.a. administration of NETs into naïve joints induced dose-dependent (5–20 ng/joint) mechanical hyperalgesia. The mechanical hyperalgesia was already significant 3 h after NET administration, reaching a peak at 6 h and maintaining significance until 24 h. The joint hyperalgesia induced by 10 and 20 ng of NETs was similar in intensity and duration to that induced by zymosan at 30 µg/joint (Fig. 2A). We also found that the ability of purified NETs to induce hyperalgesia depends on its structure, since i.a. administration of NETs previously incubated with PLZ for 3 h failed to induce hyperalgesic activity (Fig. 2B) or joint swelling (Fig. 2C).
NETs cause hyperalgesia dependent of TNF-α and IL-1β
(A) Purified NETs (5, 10 or 20 ng) were injected in WT mice and articular hyperalgesia was evaluated 3, 6 and 24 h later. Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group, two-way ANOVA followed by Bonferroni post hoc test. (B) Purified NETs (10 ng) were incubated in vitro with PLZ (1:1) for 3 h and injected into joint of WT mice. Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group; #P < 0.05 vs NETs 10 ng group, two-way ANOVA followed by Bonferroni post hoc test. (C) Joint swelling was evaluated 24 h after injection i.a. of purified NETs (10 ng). Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group; #P < 0.05 vs NET 10 ng group, one-way ANOVA followed by Bonferroni post hoc test. (D) TNF-α was quantified in joint of WT mice 2 h after injection i.a. of purified NETs (10 ng). Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group, two-sided Student’s t-test. (E) Tnfr1−/− mice received purified NETs (10 ng) i.a. and articular hyperalgesia was evaluated 3, 6 and 24 h later. Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group; #P < 0.05 vs NET 10 ng group, two-way ANOVA followed by Bonferroni post hoc test. (F) IL-1β was quantified in joint of WT mice 2 h after injection i.a. of NET purified (10 ng). Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group, two-sided Student’s t-test. (G) Il1r−/− mice received purified NETs (10 ng) i.a. and articular hyperalgesia was evaluated 3, 6 and 24 h later. Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group; #P < 0.05 vs NET 10 ng group, two-way ANOVA followed by Bonferroni post hoc test. The experiments were repeated three times. i.a.: intra-articular; IL-1R, Il1r: IL-1 receptor; NETs: neutrophil extracellular traps; PLZ: Pulmozyme; TNFR-1, Tnfr1: TNF receptor 1; WT: wild type.
TNF-α and IL-1β are involved in joint hyperalgesia induced by mBSA in the AIA model [27]. Therefore, we investigated the possible involvement of these cytokines in NET-induced hyperalgesia. Similar to the administration of mBSA, i.a. administration of NETs induced significant production of TNF-α and IL-1β (Fig. 2D and F). Confirming the participation of these cytokines in NET-induced hyperalgesia, the administration of NETs in animals genetically deficient for Tnfr1 or Il1r receptor genes reduced the hyperalgesia 3 h later and prevented this response 6 and 24 h after NET injection, as observed in WT mice (Fig. 2E and G).
Participation of Toll-like receptors 4 and 9 in NET-induced hyperalgesia and cytokine production
Previously, it was demonstrated that inflammatory hyperalgesia depends on early activation of Toll-like receptors (TLRs), which coordinate the release of hyperalgesic cytokines [28]. To investigate the mechanism by which NETs are recognized in the joint to trigger local hyperalgesia, we tested the role of these two TLRs by injecting NETs i.a. into Tlr4- and Tlr9-deficient animals. Histones and DNA, constituents of the structure of NETs, are recognized by these two TLRs [29, 30]. These mice did not exhibit a reduced nociceptive threshold when purified NETs were injected i.a. (Fig. 3A and B), indicating that both receptors play a role in recognition of NETs. In addition, compared with WT mice, Tlr4−/− and Tlr9−/− mice injected with NETs presented a significant reduction of TNF-α and IL-1β concentrations in the joint (Fig. 3C and D), indicating that the release of these cytokines during NET-induced hyperalgesia depends on the activation of these receptors. Importantly, Tlr4−/− and Tlr9−/− mice have mechanical thresholds similar to WT mice since, when prostaglandin E2 (PGE2) was injected into TLR KO, the hyperalgesic response was similar to that observed in WT animals [31].
Recognition of NETs is dependent of TLR-4 and TLR-9 to causes hyperalgesia
(A) Tlr4−/− mice received purified NETs (10 ng) i.a. and articular hyperalgesia was evaluated 3, 6 and 24 h later. Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group; #P < 0.05 vs NET 10 ng group, two-way ANOVA followed by Bonferroni post hoc test. (B) Tlr9−/− mice received purified NETs (10 ng) i.a. and articular hyperalgesia was evaluated 3, 6 and 24 h later. Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group; #P < 0.05 vs NETs 10 ng group, two-way ANOVA followed by Bonferroni post hoc test. (C) TNF-α was quantified in joint of Tlr4−/− and Tlr9−/− mice 2 h after injection i.a. of purified NETs. Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group; #P < 0.05 vs NETs 10 ng group, one-way ANOVA followed by Bonferroni post hoc test. (D) IL-1β was quantified in joint of Tlr4−/− and Tlr9−/− mice 2 h after injection i.a. of purified NETs. Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group; #P < 0.05 vs NET 10 ng group, one-way ANOVA followed by Bonferroni post hoc test. The experiments were repeated three times. i.a.: intraarticular; NETs: neutrophil extracellular traps; TLR-4, Tlr4: Toll-like receptor 4; TLR-9, Tlr9: Toll-like receptor 9.
Participation of cyclooxygenase in NET-induced hyperalgesia
The treatment of WT animals that received NETs i.a. with indomethacin, a non-selective cyclooxygenase (COX) inhibitor (5 mg/kg, i.p.), prevented the mechanical hyperalgesia determined 6 h after NET injection (Fig. 4A). Previously, it was demonstrated that, during inflammatory hyperalgesia, the production of eicosanoids depends on the induction of COX-2 expression by the hyperalgesic cytokines [11]. Taking these findings into account, we further investigated the expression of Cox2 in the joint tissues of WT, Tlr4-, Tlr9-, Tnfr1- and Il1r-deficient mice injected with NETs i.a. Significantly increased expression of Cox2 was observed in WT mice when compared with the control group (animals injected with saline). On the other hand, Cox2 expression was partially reduced in all knockout groups compared with WT animals administered NETs i.a. (Fig. 4B). Together, the results suggest that COXs products participate of NETs-induced hyperalgesia.
NETs cause hyperalgesia dependent of COX-2
(A) WT mice were treated with indomethacin (5 mg/kg, i.p.) 1 h before injection of purified NETs (10 ng) i.a. and articular hyperalgesia was evaluated 6 h after the stimuli. Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group; #P < 0.05 vs NETs 10 ng group, one-way ANOVA followed by Bonferroni post hoc test. (B) Genic expression of Cox2 in joint was evaluated 6 h after injection of purified NETs (10 ng). Mean (s.e.m.), n = 5/group. *P < 0.05 vs saline group; #P < 0.05 vs NET 10 ng group, two-way ANOVA followed by Bonferroni post hoc test. The experiments were repeated three times. Cox2, COX-2: cyclooxygenase-2; Gapdh: glyceraldehyde 3-phosphate dehydrogenase; i.a.: intraarticular; Il1r: IL-1 receptor; i.p.: intraperitoneal; NETs: neutrophil extracellular traps; Tlr4: Toll-like receptor 4; Tlr9: Toll-like receptor 9; Tnfr1: TNF receptor 1; WT: wild type.
The pain level in RA patients is related to the concentration of NET production by circulating neutrophils
To translate these results, peripheral neutrophils isolated from 24 RA patients and 13 healthy subjects (Table 1) were cultured for 4 h, and the NET concentration was determined in the supernatants. Compared with healthy subjects, the neutrophils from RA patients released spontaneously (without any stimulus) higher levels of NETs (Fig. 5A). This result suggests that inflammatory mediators released during RA are stimulating the NET production by circulating neutrophils. Surprisingly, the VAS exhibited a positive correlation between the intensity of pain and the increased production of NETs (Fig. 5B), indicating a possible association between joint pain and peripheral neutrophil NET release.
Neutrophils from patients are more capable of releasing NETs and their concentration correlates with pain
(A) Human neutrophils were isolated from peripheral blood from RA patients and individual controls. After 4 h the supernatants were used to quantify NETs. Mean (s.e.m.), two-sided Student’s t-test. (B) Concentration of NETs was correlated with pain scale. Spearman’s rank-order correlation (r) was calculated to describe correlations between the concentration of NETs released by human neutrophils and pain in RA patients. NETs: neutrophil extracellular traps.
Clinical and laboratory profile of RA patients
| Patient | Age, years | Sex | Rheumatoid factor, IU/ml | ACPA, IU/ml | DAS28 | VAS pain, 0–100 | NET, ng/ml |
|---|---|---|---|---|---|---|---|
| 1 | 71 | F | 14.2 | 224 | 3.29 | 40 | 69.15 |
| 2 | 74 | F | 0 | 278 | 1.53 | 0 | 33.47 |
| 3 | 50 | M | 0 | 18.2 | 1.18 | 50 | 12.2 |
| 4 | 47 | F | 0 | 6.2 | 3.92 | 65 | 25.74 |
| 5 | 66 | F | 40 | 287.1 | 7.84 | 90 | 97.24 |
| 6 | 60 | F | 328 | 208 | 3.05 | 10 | 43.58 |
| 7 | 51 | F | 33 | 271 | 7.18 | 90 | 43.58 |
| 8 | 71 | M | 68 | 7.3 | 2.09 | 0 | 26.54 |
| 9 | 53 | F | 281 | 200.9 | 5.8 | 100 | 156 |
| 10 | 71 | F | 18 | 107 | 6.35 | 80 | 189.61 |
| 11 | 53 | F | 82 | 202.8 | 3.48 | 20 | 15.36 |
| 12 | 64 | F | 130 | 221 | 1.53 | 20 | 67.32 |
| 13 | 39 | F | 21 | 4 | 4.31 | 80 | 33.88 |
| 14 | 80 | F | 0 | 109.7 | 2.1 | 90 | 112.58 |
| 15 | 66 | F | 40 | 287.1 | 3.15 | 10 | 33.80 |
| 16 | 49 | F | 94 | 213 | 5.91 | 80 | 161.6 |
| 17 | 66 | M | 0 | 3.9 | 2.13 | 20 | 42.78 |
| 18 | 66 | M | 0 | 16 | 4.62 | 80 | 150.03 |
| 19 | 54 | F | 0 | 4.5 | 6.19 | 20 | 12.53 |
| 20 | 67 | M | 0 | 256.6 | 6.51 | 80 | 101.23 |
| 21 | 58 | F | 0 | 3.7 | 5.3 | 70 | 20.48 |
| 22 | 58 | F | 0 | 2.2 | 2.48 | 80 | 59.6 |
| 23 | 24 | F | 221 | 209 | 7.26 | 90 | 131.08 |
| 24 | 57 | F | 0 | 2.9 | 4.52 | 70 | 58.42 |
| Patient | Age, years | Sex | Rheumatoid factor, IU/ml | ACPA, IU/ml | DAS28 | VAS pain, 0–100 | NET, ng/ml |
|---|---|---|---|---|---|---|---|
| 1 | 71 | F | 14.2 | 224 | 3.29 | 40 | 69.15 |
| 2 | 74 | F | 0 | 278 | 1.53 | 0 | 33.47 |
| 3 | 50 | M | 0 | 18.2 | 1.18 | 50 | 12.2 |
| 4 | 47 | F | 0 | 6.2 | 3.92 | 65 | 25.74 |
| 5 | 66 | F | 40 | 287.1 | 7.84 | 90 | 97.24 |
| 6 | 60 | F | 328 | 208 | 3.05 | 10 | 43.58 |
| 7 | 51 | F | 33 | 271 | 7.18 | 90 | 43.58 |
| 8 | 71 | M | 68 | 7.3 | 2.09 | 0 | 26.54 |
| 9 | 53 | F | 281 | 200.9 | 5.8 | 100 | 156 |
| 10 | 71 | F | 18 | 107 | 6.35 | 80 | 189.61 |
| 11 | 53 | F | 82 | 202.8 | 3.48 | 20 | 15.36 |
| 12 | 64 | F | 130 | 221 | 1.53 | 20 | 67.32 |
| 13 | 39 | F | 21 | 4 | 4.31 | 80 | 33.88 |
| 14 | 80 | F | 0 | 109.7 | 2.1 | 90 | 112.58 |
| 15 | 66 | F | 40 | 287.1 | 3.15 | 10 | 33.80 |
| 16 | 49 | F | 94 | 213 | 5.91 | 80 | 161.6 |
| 17 | 66 | M | 0 | 3.9 | 2.13 | 20 | 42.78 |
| 18 | 66 | M | 0 | 16 | 4.62 | 80 | 150.03 |
| 19 | 54 | F | 0 | 4.5 | 6.19 | 20 | 12.53 |
| 20 | 67 | M | 0 | 256.6 | 6.51 | 80 | 101.23 |
| 21 | 58 | F | 0 | 3.7 | 5.3 | 70 | 20.48 |
| 22 | 58 | F | 0 | 2.2 | 2.48 | 80 | 59.6 |
| 23 | 24 | F | 221 | 209 | 7.26 | 90 | 131.08 |
| 24 | 57 | F | 0 | 2.9 | 4.52 | 70 | 58.42 |
VASpain: visual analogue scale for pain; NETs: neutrophil extracellular traps; ACPA: anti-citrullinated protein antibody.
Clinical and laboratory profile of RA patients
| Patient | Age, years | Sex | Rheumatoid factor, IU/ml | ACPA, IU/ml | DAS28 | VAS pain, 0–100 | NET, ng/ml |
|---|---|---|---|---|---|---|---|
| 1 | 71 | F | 14.2 | 224 | 3.29 | 40 | 69.15 |
| 2 | 74 | F | 0 | 278 | 1.53 | 0 | 33.47 |
| 3 | 50 | M | 0 | 18.2 | 1.18 | 50 | 12.2 |
| 4 | 47 | F | 0 | 6.2 | 3.92 | 65 | 25.74 |
| 5 | 66 | F | 40 | 287.1 | 7.84 | 90 | 97.24 |
| 6 | 60 | F | 328 | 208 | 3.05 | 10 | 43.58 |
| 7 | 51 | F | 33 | 271 | 7.18 | 90 | 43.58 |
| 8 | 71 | M | 68 | 7.3 | 2.09 | 0 | 26.54 |
| 9 | 53 | F | 281 | 200.9 | 5.8 | 100 | 156 |
| 10 | 71 | F | 18 | 107 | 6.35 | 80 | 189.61 |
| 11 | 53 | F | 82 | 202.8 | 3.48 | 20 | 15.36 |
| 12 | 64 | F | 130 | 221 | 1.53 | 20 | 67.32 |
| 13 | 39 | F | 21 | 4 | 4.31 | 80 | 33.88 |
| 14 | 80 | F | 0 | 109.7 | 2.1 | 90 | 112.58 |
| 15 | 66 | F | 40 | 287.1 | 3.15 | 10 | 33.80 |
| 16 | 49 | F | 94 | 213 | 5.91 | 80 | 161.6 |
| 17 | 66 | M | 0 | 3.9 | 2.13 | 20 | 42.78 |
| 18 | 66 | M | 0 | 16 | 4.62 | 80 | 150.03 |
| 19 | 54 | F | 0 | 4.5 | 6.19 | 20 | 12.53 |
| 20 | 67 | M | 0 | 256.6 | 6.51 | 80 | 101.23 |
| 21 | 58 | F | 0 | 3.7 | 5.3 | 70 | 20.48 |
| 22 | 58 | F | 0 | 2.2 | 2.48 | 80 | 59.6 |
| 23 | 24 | F | 221 | 209 | 7.26 | 90 | 131.08 |
| 24 | 57 | F | 0 | 2.9 | 4.52 | 70 | 58.42 |
| Patient | Age, years | Sex | Rheumatoid factor, IU/ml | ACPA, IU/ml | DAS28 | VAS pain, 0–100 | NET, ng/ml |
|---|---|---|---|---|---|---|---|
| 1 | 71 | F | 14.2 | 224 | 3.29 | 40 | 69.15 |
| 2 | 74 | F | 0 | 278 | 1.53 | 0 | 33.47 |
| 3 | 50 | M | 0 | 18.2 | 1.18 | 50 | 12.2 |
| 4 | 47 | F | 0 | 6.2 | 3.92 | 65 | 25.74 |
| 5 | 66 | F | 40 | 287.1 | 7.84 | 90 | 97.24 |
| 6 | 60 | F | 328 | 208 | 3.05 | 10 | 43.58 |
| 7 | 51 | F | 33 | 271 | 7.18 | 90 | 43.58 |
| 8 | 71 | M | 68 | 7.3 | 2.09 | 0 | 26.54 |
| 9 | 53 | F | 281 | 200.9 | 5.8 | 100 | 156 |
| 10 | 71 | F | 18 | 107 | 6.35 | 80 | 189.61 |
| 11 | 53 | F | 82 | 202.8 | 3.48 | 20 | 15.36 |
| 12 | 64 | F | 130 | 221 | 1.53 | 20 | 67.32 |
| 13 | 39 | F | 21 | 4 | 4.31 | 80 | 33.88 |
| 14 | 80 | F | 0 | 109.7 | 2.1 | 90 | 112.58 |
| 15 | 66 | F | 40 | 287.1 | 3.15 | 10 | 33.80 |
| 16 | 49 | F | 94 | 213 | 5.91 | 80 | 161.6 |
| 17 | 66 | M | 0 | 3.9 | 2.13 | 20 | 42.78 |
| 18 | 66 | M | 0 | 16 | 4.62 | 80 | 150.03 |
| 19 | 54 | F | 0 | 4.5 | 6.19 | 20 | 12.53 |
| 20 | 67 | M | 0 | 256.6 | 6.51 | 80 | 101.23 |
| 21 | 58 | F | 0 | 3.7 | 5.3 | 70 | 20.48 |
| 22 | 58 | F | 0 | 2.2 | 2.48 | 80 | 59.6 |
| 23 | 24 | F | 221 | 209 | 7.26 | 90 | 131.08 |
| 24 | 57 | F | 0 | 2.9 | 4.52 | 70 | 58.42 |
VASpain: visual analogue scale for pain; NETs: neutrophil extracellular traps; ACPA: anti-citrullinated protein antibody.
Discussion
The participation of NETs in the pathology of autoimmune diseases, including RA, has already been described. During RA, NETs participate in joint tissue lesions [32]. However, the role of NETs in the pathophysiology of inflammatory pain, which is one of the major symptoms of RA and constitutes the highest complaint from patients [33], was not investigated. The data presented in this study demonstrated, for the first time, that NETs are released in the joints in the AIA model and mediate articular pain through activation of TLRs. Subsequently, cytokines are released that stimulate COX-2 expression. Moreover, in a translational manner, similarly to Khandpur and colleagues [32], we demonstrated that neutrophils from RA patients release more NETs than neutrophils from healthy individuals, and the levels of NETs in RA directly correlate with the clinical score of pain.
First, we first investigated if the mBSA challenge into the joints of AIA mice was capable of inducing NETs. We used this model because it has several similarities to RA, such as oedema formation, histopathological changes, increased proinflammatory cytokine production and joint hyperalgesia [24, 34, 35]. It has been shown that NET production increases in the AIA model, and NET levels are associated with increased expression of Padi4, the gene for an enzyme involved with the release of NETs [36]. Furthermore, the administration of purified NETs into naïve joints of mice induced dose- and time-dependent joint hyperalgesia and oedema. These events induced by the i.a. dose of 10 ng of NETs were similar to those induced by 30 µg of zymosan, a well-known inducer of joint pain and oedema [37]. The treatment of AIA mice with an inhibitor of PADI-4, Cl-amidine, reduced the joint oedema and hyperalgesia and the concentration of NETs. In addition to inhibiting PADI-4, NETs can be inhibited through DNases that degrade themselves, such as Pulmozyme, a deoxyribonuclease I used in the clinic to treat cystic fibrosis [38]. The Pulmozyme treatment reduced the concentration of NETs in the joints of AIA mice, and the animals showed reduced joint oedema, synovitis and joint hyperalgesia, events that have been associated with neutrophil migration [24, 39]. Furthermore, prior incubation of NETs with Pulmozyme inhibited their capacity to trigger oedema formation and joint hyperalgesia, indicating that the chemical structure of NETs is necessary for their action. These isolated NETs are purified and structurally have components like MPO and histone-H3 citrulline, and Hawez showed that phorbol 12-myristate 13-acetate can produce NETs with citrullination [40]. Thus, our data add knowledge of the mechanism involved in the arthritogenic and hyperalgesic effects of migrated neutrophils during AIA, events, at least in part, triggered by NET release. The Pulmozyme-treated animals showed a reduced nociceptive threshold fall (mechanical hyperalgesia) before the levels of NETs became statistically significant (3 h after the joint challenge), indicating that NETs are effective at mediating joint hyperalgesia, even at low concentrations. Corroborating our data, in the collagen-induced arthritis model, there was also an increase in NET production and the treatment of arthritic animals with a PADI-4 inhibitor reduced joint inflammation [16].
During inflammation, eicosanoids are produced mainly by COX induction [41]. Among the eicosanoids, PGE2 is one of the major mediators involved in the sensitization of primary sensitive neurons [11]. This explains the clinical effectiveness of COX inhibitors in the control of inflammatory hyperalgesia [42]. Our research group and others have demonstrated that the induction of COX-2 is mediated by cytokines (TNF-α and IL-1β) released at the inflammatory site, including inflamed joints [11, 43]. Considering these findings, we investigated a possible mechanism by which NETs could induce joint hyperalgesia. First, we demonstrated that purified NETs, but not degraded NETs, induce the production of TNF-α and IL-1β, which mediate joint hyperalgesia, since mice genetically deficient for Tnfr1 or Il1r did not develop NET-induced hyperalgesia, mainly in the latter phase of the response. These cytokines probably are involved with the maintenance of hyperalgesia, since our results show that at the beginning of pain response the animals deficient of these cytokines did not present significate reduction of NET-induced hyperalgesia. Consistently, AIA animals treated with etanercept, a TNF-blocker, or an antibody against IL-1β presented reduced hyperalgesia [44, 45].
TLRs play a central role in regulating innate immunity, recognizing the inflammatory stimuli and triggering the host innate response [46]. A previous study demonstrated that joint pain induced by different stimuli, including mBSA, is triggered by TLRs and nucleotide-binding oligomerization domain-like (NLRs) receptors [47, 48]. Histones, a critical constituent of NETs, can be recognized via TLR-4, and DNA, another important constituent of NETs, by TLR-9 [49, 50]. Accordingly, Tlr4- or Tlr9-deficient mice injected i.a. with purified NETs did not present a mechanical threshold fall as observed in WT mice; furthermore, no increases in articular concentrations of TNF-α or IL-1β were observed. We also observed that the Tlr4-, Tlr9-, Tnfr1- and Il1r-deficient mice had reduced expression of Cox2 upon i.a. injection of purified NETs. Confirming the participation of COX-2 in NET-induced hyperalgesia, treatment of mice with indomethacin inhibited the NET-induced joint hyperalgesia. From a clinical perspective, neutrophils from RA patients cultured in vitro were more prone to release of NETs than those of healthy controls, which in turn correlated positively with the VAS of joint pain, suggesting the same behaviour for synovial fluid neutrophils.
In conclusion, we describe for the first time that NETs released in joints in the AIA model of arthritis induce the release of TNF-α and IL-1β possibly via TLR-4 and TLR-9. Subsequently, these cytokines lead to increase of COX-2 expression, contributing to hyperalgesia. Thus, inhibition of NET production could be a new potential target for analgesic therapies for inflammatory pain.
Acknowledgements
We are grateful for the excellent technical assistance of Lívia Ambrósio, Marcella Daruge, Juliana Abumansur, Ieda Regina dos Santos, Sérgio Roberto Rosa, Ana Kátia dos Santos and Diva Montanha. All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. A.H.S. conceived and designed the study, planned and performed all the experiments, analysed the data and drafted the manuscript. C.C.M. and A.M.G.M. contributed to performing several experiments and helped revise the manuscript. F.F.L.S., L.C.B. and R.D.R.O. contributed to the human data and helped revise the manuscript. S.Y.F. and T.A.S. designed the study, planned the experiments, contributed to histological staining and analyses, and helped to revise the manuscript. J.C.A.F., T.M.C., P.L.J. and F.Q.C. designed the study, planned the experiments, analysed the data and drafted the manuscript.
Funding: The research leading to these results received funding from the São Paulo Research Foundation (FAPESP) under grant agreement no. 2013/08216-2 (Center for Research in Inflammatory Diseases—CRID). The project also had support from CAPES and CNPq, Brazil.
Disclosure statement: The authors have declared no conflicts of interest.
Data availability statement
The data under lying this article are available in the article and in its online supplementary material.
Supplementary data
Supplementary data are available at Rheumatology online.
Comments