Enhanced formation and impaired degradation of neutrophil extracellular traps in dermatomyositis and polymyositis: a potential contributor to interstitial lung disease complications

Abstract

Dermatomyositis (DM) and polymyosits (PM) are systemic autoimmune diseases whose pathogeneses remain unclear. Neutrophil extracellular traps (NETs) are reputed to play an important role in the pathogenesis of autoimmune diseases. This study tests the hypothesis that NETs may be pathogenic in DM/PM. Plasma samples from 97 DM/PM patients (72 DM, 25 PM) and 54 healthy controls were tested for the capacities to induce and degrade NETs. Plasma DNase I activity was tested to further explore possible reasons for the incomplete degradation of NETs. Results from 35 DM patients and seven PM patients with interstitial lung disease (ILD) were compared with results from DM/PM patients without ILD. Compared with control subjects, DM/PM patients exhibited a significantly enhanced capacity for inducing NETs, which was supported by elevated levels of plasma LL-37 and circulating cell-free DNA (cfDNA) in DM/PM. NETs degradation and DNase I activity were also decreased significantly in DM/PM patients and were correlated positively. Moreover, DM/PM patients with ILD exhibited the lowest NETs degradation in vitro due to the decrease in DNase I activity. DNase I activity in patients with anti-Jo-1 antibodies was significantly lower than in patients without. Glucocorticoid therapy seems to improve DNase I activity. Our findings demonstrate that excessively formed NETs cannot be degraded completely because of decreased DNase I activity in DM/PM patients, especially in patients with ILD, suggesting that abnormal regulation of NETs may be involved in the pathogenesis of DM/PM and could be one of the factors that initiate and aggravate ILD.

Introduction

Dermatomyositis (DM) and polymyositis (PM) are systemic autoimmune diseases clinically characterized by chronic proximal muscle inflammation, weakness and multi-organ involvement. The most frequently observed organ involvement is interstitial lung disease (ILD), which is associated with poor prognoses for these patients 1–4. However, the aetiology and pathogenesis of DM/PM and the complicated ILD are still unclear 3,4.

Recent studies have indicated that neutrophil extracellular traps (NETs) are involved in several autoimmune diseases and inflammatory lung diseases 5–15. Activated by microbes, neutrophils can release DNA, histones and anti-microbial peptides to form NETs to trap and kill invading microbes 16,17, which is the physiological function of NETs. However, in certain situations NETs can be pathogenic. NETs are composed largely of endocellular and intranuclear materials that could be potential autoantigens to the body's immune system. It has been established that NETs are related to the formation of autoantibodies in systemic lupus erythematosus (SLE), rheumatoid arthritis (RA) and small-vessel vasculitis 6,12,18. Moreover, some components of NETs – including extracellular histones and neutrophil granular proteins – can directly injure vascular endothelial cells 15,19,20.

Besides microbes, proinflammatory substances and activated platelets can also induce neutrophils to form NETs 7,15. DM/PM patients have abnormal cytokine environments 1,2, which may result in excessive NETs formation. These NETs can stimulate plasmacytoid dendritic cells (pDCs) to release type I interferons (IFNs) and further disturb the cytokine network 21,22, resulting in a vicious cycle of continued NETs formation. If these surplus NETs cannot be cleared in a timely manner to halt this cycle, they may injure lung endothelium and disrupt internal environment homeostasis 15,23. If other unknown factors aid this process, DM/PM may occur and progress to ILD.

However, whether or not NETs are involved in DM/PM and how NETs contribute to the initiation and progression of ILD are unknown. To address these questions, we compared NETs induction and degradation as well as plasma DNase I activity in DM/PM patients and control subjects. We further compared these findings between DM/PM patients with ILD (DML/PML) and DM/PM patients without ILD (DMNL/PMNL).

Methods

Subjects

Seventy-two DM patients and 25 PM patients diagnosed according the Bohan and Peter criteria 24,25 were recruited for this study from the Department of Rheumatology at China–Japan Friendship Hospital (January 2012–September 2013). General information about DM/PM groups can be found in Table 1. During the same period, 54 age- and sex-matched healthy Chinese volunteers were selected to be control subjects. The study was approved by the Ethics Committee of the China–Japan Friendship Hospital. All patients and controls gave written informed consent. Ethylenediamine tetraacetic acid (EDTA) anti-coagulant venous blood samples were separated by centrifugation at 400×g for 10 min. Plasma was divided into six parts and stored at −80°C.

Table 1.

Demographic and clinical features of dermatomyositis/polymyositis (DM/PM) patients

Group	No.	Sex (male/female)	Age (years)	ILD (with/without)	Anti-Jo-1 (+/−)	ANA (+/−)	Anti-Ro (+/−)	Duration (months)	MYOACT-p (mean ± s.d.)	MMT (mean ± s.d.)
PM	25	5/20	48·79 ± 15·07	(7/18)	(4/21)	(17/8)	(2/23)	44·8	1 ± 0·62	75·13 ± 5·17
DM	72	12/60	43·73 ± 16·11	(35/37)	(4/68)	(55/17)	(9/63)	26·4	1·21 ± 0·74	72·82 ± 14·82

DM = dermatomyositis; PM = polymyositis; ILD = interstitial lung disease; ANA = anti-nuclear antibodies; MYOACT-p = myositis disease activity assessment visual analog scales-pulmonary disease activity; MMT = manual muscle test; s.d. = standard deviation.

Patients' clinical data were obtained from the electronic medical record system of the China–Japan Friendship Hospital. Disease activity and muscle strength were assessed by two rheumatologists using the myositis disease activity assessment tool (MDAAT) and the manual muscle test (MMT) 26. ILD was diagnosed by two rheumatologists based on high-resolution computed tomography (HRCT), pulmonary function tests and clinical presentation. Thirty-five of the 72 DM patients and seven of the 25 PM patients had ILD complications. These patients were categorized into three subgroups; (i) acute or subacute types: patients who present with severe, rapidly progressive dyspnoea with progressive hypoxaemia within a month of the onset of lung involvement; (ii) chronic type: patients who present with more slowly progressing dyspnoea; and (iii) asymptomatic type: patients with ILD as demonstrated by either a chest radiograph or a pulmonary function test in the absence of clinically apparent signs or symptoms 4. According to our classification scheme, six DM patients belonged to the subacute subgroup and 29 DM patients and seven PM patients belonged to the chronic or asymptomatic subgroup.

Reagents

Histopaque-1119, Histopaque-1077, calf thymus DNA, Sybr Green, phorbol myristate acetate (PMA), polylysine and 4′,6-diamidino-2-phenylindole (DAPI) were purchased from Sigma-Aldrich (St Louis, MO, USA). Sytox Green, Quant-iT PicoGreen dsDNA reagent and kit and fetal bovine serum (FBS) were purchased from Invitrogen (Carlsbad, CA, USA). The LL-37 enzyme-linked immunosorbent assay (ELISA) kit was purchased from Hycult (Uden, the Netherlands). RPMI without phenol red was purchased from Gibco (Carlsbad, CA, USA). DNase I solution was purchased from Worthington (Lakewood, NJ, USA). Antibodies to LL-37 (ab87701) and neutrophil elastase (NE) (ab21595) were purchased from Abcam (Cambridge, MA, USA). Secondary antibodies coupled to Dylight 488 and Dylight 594 were purchased from ZSGB-BIO (Beijing, China). Anti-CD11b and anti-CD16 were purchased from Biolegend (San Diego, CA, USA).

Neutrophil isolation

For neutrophil purification, blood was collected from the healthy individuals using an EDTA vacutainer and then separated by density gradient centrifugation within 2 h of collection, according to the manufacturer's instructions (specific details can be found in the Supporting information). The purity of neutrophils (90%) was determined via flow cytometer (BD Accuri C6) using anti-CD11b and anti-CD16 staining.

Assessment of NETs formation

Neutrophils were resuspended in RPMI without phenol red, and 3 × 10⁴ cells were cultured with 10% plasma from different subjects in a 96-well dark plate with transparent bottoms. After 2 h of incubation in a CO₂ incubator at 37°C cells were fixed with 4% paraformaldehyde, and 0·2 μM of Sytox Green were added. Sytox Green can easily penetrate cells with compromised plasma membranes and yet will not cross the membranes of live cells. Therefore, using fluorescence microscopy, NETs formation can be observed directly. NETs formation was quantified by plate assays, in which fluorescence intensity (excitation 485 nm, emission 520 nm) was measured using Gemini EM (Molecular Devices, Sunnyvale, CA, USA) with a bottom-reading model. Results were reported as DNA relative fluorescence units (RFUs) 6. The capacity to induce NETs formation was tested in all participants and this test was performed three times.

Measurement of plasma circulating cell-free DNA (cfDNA) and LL-37 levels

Plasma samples were again centrifuged at 400×g for 10 min before cfDNA was measured. Plasma cfDNA was measured for all participants using the Quant-iT PicoGreen dsDNA reagent and kit, according to the manufacturer's instructions. Each sample was performed in triplicate. Plasma LL-37 concentrations were measured by using an LL-37 ELISA kit, according to the manufacturer's instructions. Samples from 36 DM patients, 12 PM patients and 38 healthy controls were tested and each sample was performed in duplicate.

Assessment of NETs degradation in vitro

PMA was used to induce maximum NETs formation in vitro. Sterile 13-mm round glass coverslips coated with polylysine were added to each well of a 24-well cell culture plate. In each well, 2 × 10⁵ neutrophils suspended in 500 μl of RPMI-1640 containing 4% FBS were added, and the plate was incubated for 1 h in a CO₂ incubator at 37°C. PMA was then added to each well to achieve a final concentration of 50 nM. After 4 h incubation, cells and NETs were fixed with 4% paraformaldehyde. These fixed NETs were used for in-vitro degradation tests.

Plasma was used to degrade NETs in vitro 9,11. DNase buffer (10 mM Tris-HCl pH 7·5, 10 mM MgCl₂, 2 mM CaCl₂ and 50 mM NaCl) was used to dilute plasma or exogenous DNase I. The fixed NETs were left untreated (negative control), incubated with 500 μl of exogenous DNase I solution (0·55 U/ml) (positive control) or incubated with 10% plasma for 2 h at 37°C. Each plasma sample was tested in duplicate. Then, the coverslips were washed three times with phosphate-buffered saline (PBS) in preparation for the next immunofluorescence detection. LL-37 and NE are specific to NETs, so primary antibodies to LL-37 and NE were used to detect NETs. After 2 h incubation at 37°C, bound antibodies were detected with secondary antibodies coupled to Dylight 488 and Dylight 594 for 1 h at 37°C. The DNA was stained with DAPI for 1 min. The typical images were projections of a confocal stack (Leica, Solms, Germany). NETs were identified as netlike structures positive for NE (green), LL-37 (red) and dsDNA (blue). Degradation of NETs was observed separately by two investigators using a fluorescence microscope, and the NETs were counted manually in three visual fields (original magnification ×200). The negative controls were considered to exhibit 0% degradation of NETs, and the positive controls were considered to exhibit 100% degradation of NETs. Using these controls, mean degradation percentages from the three fields were calculated. Higher percentages represented stronger capacities to degrade NETs.

Detection of DNase I activity

Based on the hydrolysis of substrate DNA in a DNA-agar plate, the radial enzyme-diffusion method quantifies the activity of DNase I by measuring the dark areas of the agar plate 27 (specific details can be found in the Supporting information). Plasma DNase I activity was measured in all participants and each sample was performed in duplicate.

Statistical analysis

GraphPad Prism version 5 was used to perform comparisons between different groups and to draw figures. Calculations were based on a 95% confidence interval (CI). P-values less than 0·05 were considered significant. Student's t-test was used for continuous variables that were normally distributed. The Mann–Whitney U-test was used for continuous variables that were not normally distributed. One-way analysis of variance (anova) was used for comparing continuous variables between three groups.

Ethics approval

The Ethics Committee of China–Japan Friendship Hospital approved this study.

Results

DM/PM plasma induced higher rates of NETs formation in vitro

NETs formation is a dynamic process that progresses through several stages, including the lobulated nucleus, delobulated nucleus, diffused NETs and spread NETs stages (Fig. 1a). Greater numbers of Sytox green-positive cells and bigger nuclei were considered to represent more and faster NETs formation. Plasma from both healthy controls (Fig. 1b) and DM/PM patients (Fig. 1c) can induce neutrophils to form NETs. Compared with control plasma, DM/PM plasma induced greater numbers of normal neutrophils to form NETs at faster speed. This result was evidenced by quantitative fluorescence intensity-based plate assays, in which DM/PM plasma showed statistically higher RFUs than did control plasma (246 ± 93·48 RFUs versus 191·6 ± 52·88 RFUs, P = 0·002, Fig. 1d). We further compared the capacity to induce NETs formation between PM and DM patients, and no significant difference was found. We also found no significant difference between DML/PML and DMNL/PMNL (data not shown).

Fig. 1.

Measurement of plasma-induced neutrophil extracellular traps (NETs) formation and NETs-related plasma biomarkers. NETs formation is a dynamic process that progresses through the lobulated nucleus, delobulated nucleus, diffused NETs and spread NETs stages (a). Scale bars represent 10 μm. Original magnification ×400. Dermatomyositis (DM) plasma (c) induces more neutrophils to form NETs than does control plasma (b). (d) Results from quantitative plate assays, in which fluorescence (excitation 485 nm, emission 520 nm) was measured by using Gemini EM and results were reported as DNA relative fluorescence units (RFUs). This experiment was performed three times. (e,f) Comparisons of plasma LL-37 and circulating cell-free DNA (cfDNA) concentrations between the control group and the DM/polymyositis (PM) group. cfDNA was measured for all participants, and each sample was performed in triplicate. LL-37 was measured for 36 DM patients, 12 PM patients and 38 healthy controls, and each sample was performed in duplicate. Statistical analysis was performed using the two-tailed Mann–Whitney U-test. **P < 0·01; ***P < 0·001.

PM/DM patients exhibited elevated plasma levels of LL-37 and cfDNA

We found that DM/PM plasma induced more normal neutrophils to form NETs in vitro, which may not represent in-vivo NETs formation. Therefore, we used other serological markers, LL-37 and cfDNA, to ascertain if NETs formation was enhanced in vivo. LL-37 is an important component of NETs 23, so we surmised that plasma concentrations of LL-37 may be related to NETs formation in vivo. Plasma LL-37 concentration in the DM/PM group was 48·7 ± 11·25 ng/ml, significantly higher than that in the control group (33·81 ± 13·08 ng/ml, P = 0·0031) (Fig. 1e). Residual NETs are a major source of cfDNA, and excessive NETs formation may lead to an increase in plasma cfDNA 12,28. Similarly, plasma cfDNA concentration in the DM/PM group was 264·9 ± 62·95 ng/ml, significantly higher than that in the control group (197·1 ± 31·36 ng/ml, P < 0·0001) (Fig. 1f). These elevated levels of LL-37 and cfDNA are indirect evidence of enhanced NETs formation in DM/PM patients.

DML/PML subjects failed to completely degrade NETs in vitro

Overproduction of NETs is harmful to the body's immune system, so we tested whether excessively formed NETs can be cleared normally by DM/PM plasma. We randomly selected 15 healthy subjects, 19 DML/PML patients and 17 DMNL/PMNL patients whose plasma samples were tested for NET degradation in vitro. PMA-induced NETs showed the representative fibriform structure, in which DNA was blue, LL-37 was red and NE was green (Fig. 2a). Compared with control plasma (Fig. 2b), plasma from DML/PML and DMNL/PMNL could not degrade NETs completely (Fig. 2c). Quantitative analysis indicated that the percentage of NETs degradation in DML/PML (58·58 ± 21·4%) was significantly lower than that in DMNL/PMNL (83·41 ± 12·64%, P = 0·0002) and that in the control group (95·07 ± 5·35%, P < 0·0001, Fig. 2d).

Fig. 2.

Dermatomyositis/polymyositis (DM/PM) patients with interstitial lung disease (ILD) (DML)/PML) plasma failed to degrade neutrophil extracellular traps (NETs) completely in vitro. The images are projections of a confocal stack (Leica). (a) Typical netlike structures of NETs, in which DNA is blue, LL-37 is red and NE is green. (b) NETs were completely degraded by exogenous DNase I and control plasma. In contrast, plasma from DML/PML patients did not degrade NETs completely (c). Scale bars represent 25 μm. Original magnification ×600. (d) Percentage of NET degradation in the three subsets. Samples from 15 healthy subjects, 19 DML/PML patients and 17 DM/PM patients without ILD (DMNL/PMNL) patients were tested and each sample was performed in duplicate. Statistical analysis was performed by using one-way analysis of variance (anova). **P < 0·01; ***P < 0·001.

Decrease in DNase I activity was the main reason for DML/PML failure to degrade NETs

DNase I is responsible for degrading NETs in vitro, so its activity was measured to analyse the reason for plasma incompletely degrading NETs in DML/PML. The results indicated that DNase I activity in the PM and the DM groups was 0·1822 ± 0·0940 U/ml and 0·2094 ± 0·1112 U/ml, respectively, significantly lower than that in the control group (0·3933 ± 0·1523 U/ml, P < 0·0001) (Fig. 3a). Furthermore, DML/PML patients exhibited lower DNase I activity than did DMNL/PMNL patients (0·1677 ± 0·077 U/ml versus 0·2301 ± 0·1187 U/ml, P = 0·0039) (Fig. 3b). These results indicate that DNase I activity was decreased in both the DM and PM groups, especially in DML/PML patients. We further analysed the correlation and found that plasma DNase I activity correlated significantly with the capacity of plasma-degrading NETs (r = 0·5549, P = 0·0004) (Fig. 3c), suggesting that the impairment of DNase I activity is responsible for plasma failing to degrade NETs in vitro.

Fig. 3.

Comparisons of DNase I activity in the paired subgroups. (a,b) The short lines represent the means. Statistical analysis was performed using Student's t-test. (c) Correlation analysis was performed on paired data from 36 dermatomyositis/polymyositis (DM/PM) patients. (d) Statistical analysis was performed using Student's t-test. (e,f) Statistical analysis was performed using the two-tailed Mann-Whitney U-test. DMLS = DM patients with subacute interstitial lung disease (ILD); DMLC = DM patients with chronic or asymptomatic ILD. Plasma DNase I activity was measured in all participants and each sample was performed in duplicate. *P < 0·05; **P < 0·01; ***P < 0·001.

DNase I activity was improved by glucocorticoid treatment

Among our DM/PM subjects, 12 patients had plasma samples from before and after glucocorticoid treatment. The mean treatment interval was 2 months, and the mean dosage of glucocorticoids was 35 mg per day. By comparing these 12 paired samples, we discovered that treatment significantly restored DNase I activity (Fig. 3d).

Patients with anti-Jo-1 antibodies exhibited lower DNase I activity

Anti-Jo-1 antibodies and anti-Ro antibodies are considered to be related to ILD, so relationships between autoantibodies and DNase I activity were analysed. Among our DM/PM subjects, eight patients were positive for anti-Jo-1 antibody. We compared these eight patients with other DM/PM patients; the eight patients exhibited significantly lower DNase I activity (0·1234 ± 0·1003 U/ml versus 0·21 ± 0·1044 U/ml, P = 0·0188) (Fig. 3e). Eleven patients were positive for anti-Ro antibody. DNase I activity was compared in patients with and without anti-Ro antibodies, but no significant difference was found (date not shown).

DM patients with subacute ILD exhibited lower DNase I activity

We hypothesized that the pathogenesis of ILD was different at different stages of the disease, so we compared DNase I activity in ILD patients at different stages. In our DM group, six patients had subacute ILD complications and other patients had chronic or asymptomatic ILD complications. The six subacute ILD patients exhibited significantly lower DNase I activity than did patients with chronic or asymptomatic ILD (0·1229 ± 0·0548 U/ml versus 0·1887 ± 0·0773 U/ml, P = 0·0416) (Fig. 3f).

Analysis of the relationship between DNase I activity and serological and clinical parameters

Correlation analysis was performed between DNase I activity and other parameters, including erythrocyte sedimentation rate, immunoglobulin, complements, rheumatoid factor, LL-37, cfDNA, MYOACT-pulmonary disease activity, MMT, creatine kinase MB (CK-MB), C-reactive protein (CRP), aspartate aminotransferase (AST) and pulmonary function test. DNase I activity was shown to correlate positively with CRP (P = 0·007) and neutrophil concentrations (P = 0·032), but no correlation was found with any other parameter (Table 2).

Table 2.

Correlation between DNase I activity and other serological markers in dermatomyositis (DM) patients

		LL-37	cfDNA	Neutrophil	CRP	CK	MYOACT-p	MMT
DNase I	r	−0·260	0·009	−0·237	−0·296	−0·114	0·207	−0·167
	P	0·440	0·959	0·032*	0·007**	0·304	0·195	0·446

^*Correlations were considered significant at the 0·05 level (two-tailed).

^**Correlations were considered significant at the 0.01 level (two-tailed). Statistical analysis was performed by using the two-tailed Pearson's correlation test. CRP = C-reactive protein; CK = creatine kinase; MYOACT-p = myositis disease activity assessment visual analog scales-pulmonary disease activity; MMT = manual muscle test; cfDNA = circulating cell-free DNA.

Discussion

In the present study, we have demonstrated that DM/PM plasma induced higher rates of NETs formation in vitro. Secondly, we found that PMA-induced NETs cannot be completely degraded by DM/PM plasma, and DML/PML plasma exhibited the weakest degradation capacity. Thirdly, we identified that DNase I activity was impaired in DM/PM patients, especially in DML/PML patients, and that the impairment of plasma DNase I activity was responsible for the failure to degrade NETs. Fourthly, patients with anti-Jo-1 antibody or subacute ILD exhibited lower DNase I activity, which could be increased by glucocorticoid treatment. These findings suggest that abnormal regulation of NETs is involved in the pathogenesis of DM/PM. NETs may be a potential contributor to the development of DM/PM-complicated ILD, and DNase I could be a possible therapeutic target.

Enhanced NETs formation has been found in SLE, RA and small-vessel vasculitis 6,10,12; to our knowledge, this is the first report concerning enhanced NETs formation in DM/PM. These autoimmune diseases share some common features, chiefly that the proinflammatory cytokines predominate. Proinflammatory substances can induce neutrophils to form NETs 5–7, and increased levels of interleukin (IL)-8, IL-17 and tumour necrosis factor (TNF)-α have been reported in DM/PM 1. Therefore, we infer that plasma proinflammatory cytokines play a key role in plasma-induced NETs formation. Although NETs formation is excessive in vitro, NETs formation in vivo remains unknown in DM/PM. Because plasma LL-37 and cfDNA can represent NETs formation indirectly in vivo, we tested these two markers 12,23,28. Significantly increased levels of LL-37 and cfDNA seem to suggest that NETs formation is also enhanced in vivo.

In DM/PM patients, excessive NET formation may damage lung tissue and trigger a series of immune responses. In the airway lumen, neutrophils can be stimulated by a variety of agents to form NETs 5,7. Infiltrating NETs have been found in the lung specimens from mouse models suffering from inflammatory lung diseases 5,15. Although NETs can trap and kill invading microbes, components of NETs – including citrullinated histone (CitH3), NE, myeloperoxidase (MPO), LL-37 and other neutrophil proteases – can contribute to lung damage 5,14,15. Moreover, extracellular histones can directly cause epithelial and endothelial cell death 19,20. C1q deposition on NETs can activate complements to further increase neutrophil recruitment 11, and these neutrophils can further form NETs in the proinflammatory enviroment, thus forming a vicious cycle and persistently injuring cells. Permanent cellular injury is accompanied by the continuous regeneration of lung tissue. In a healthy individual, type I pneumocytes are normally generated by type II pneumocytes, and injury does not result in fibrosis. However, when residual NETs cause severe injury in hosts who have an abnormal immunoregulation background, injured type I pneumocytes may be replaced by fibroblasts and then pulmonary fibrosis may occur and develop 3. Therefore, the timely removal of NETs is crucial for preventing and decreasing NETs-associated tissue damage and pulmonary fibrosis.

Besides being internalized and removed by macrophages, NETs are mainly degraded by DNase I 9,11, which was demonstrated in our study. We found that the capacity to degrade NETs correlated positively with plasma DNase I activity. Furthermore, DML/PML plasma cannot completely degrade NETs owing to the severe impairment of DNase I activity. Impairment of DNase I activity has been found in SLE 9,11. SLE patients were found to carry specific inhibitors of DNase I and anti-NETs antibodies, which can inhibit the degradation of NETs by plasma DNase I. We propose that similar inhibitors or autoantibodies that exist in DM/PM are key factors for improving DNase I activity after glucocorticoid treatment. A possible reason for this improvement is that glucocorticoid treatment inhibits the inflammatory response and decreases the production of autoantibodies.

Insufficient clearance of NETs has been implicated in autoantibody formation in several autoimmune diseases 6,12,18. In this study, eight patients with anti-Jo-1 antibodies exhibited significantly decreased DNase I activity, which suggests that impaired DNase I may be related to the formation of anti-Jo-1 antibody. Anti-Jo-1 antibody is a specific autoantibody of DM/PM and has a strong association with ILD. These eight patients all had ILD complications. We propose that the induction of autoantibodies is a possible mechanism for the DNase I contribution to the development of ILD in DM/PM patients.

In our study, six patients had subacute ILD complications, and they exhibited lower DNase I activity. These patients seem to have more proinflammatory cytokines and more severe autoimmune responses, which may inhibit DNase I activity. However, the mechanism by which ILD is triggered in DM/PM remains unknown. Based on current results, we infer that abnormal cytokine and immune environments not only lead to excessive NETs formation, they also inhibit DNase I activity, all of which exposes lung tissue to large amounts of NETs. If other unknown factors co-operate with NETs, ILD may occur and develop. Therefore, DNase I may be a crucial treatment target for decreasing NETs exposure and preventing ILD. Further experiments are needed to establish the therapeutic effect of DNase I on ILD in DM/PM.

We acknowledge one main limitation of our study: we did not have direct pathological evidence to show that NETs infiltrated lung tissue. Because ILD was diagnosed by using HRCT, pulmonary function tests and clinical presentations, lung biopsy specimens were not available for further histological testing.

In conclusion, DM/PM patients are potentially exposed to large amounts of NETs. More importantly, DM/PM patients – especially DML/PML patients – cannot adequately clear NETs because of low DNase I activity. These findings provide a novel insight into the molecular pathway of the pathogenesis of DM/PM-complicated ILD and suggest that the NETs pathway may play a key role in the development of DM/PM-complicated ILD.

Disclosure

The authors have no conflicts of interest.

Supporting information

Additional Supporting information may be found in the online version of this article at the publisher's web-site:

Fig. S1. Radial enzyme-diffusion method quantifying activity of DNase I in plasma. Figure S1 is the DNase I standard, with concentrations from 1·1 to 0·034 U/ml.

Fig. S2. Stability of plasma DNase I activity.