Neutrophil extracellular trap-induced intermediate monocytes trigger macrophage activation syndrome in adult-onset Still’s disease

All AOSD patients fulfilled Yamaguchi’s criteria after exclusion of those with infectious, neoplastic, and autoimmune disorders. Patients were considered as having active AOSD if they had fever and/or arthralgia/arthritis and/or any suggestive skin lesions and/or sore throat. All HC subjects were recruited from age- and sex-matched volunteers with no history of autoimmune, rheumatic, or other diseases. Information on demographic and clinical data was entered into a database together with the laboratory results. The AOSD disease activity of each patient was assessed using a modified Pouchot score [21]. AOSD complicated with MAS (AOSD-MAS) was defined using the 2016 EULAR/ACR/PRINTO classification criteria for MAS complicating systemic JIA (sJIA-MAS criteria) [22]. The experimental design was approved by the Ethics Committee of Shanghai Jiao Tong University (identifier 2016–62), and all the participants provided informed consent.

A total of 37 AOSD patients (27 active and 10 inactive AOSD patients) and 12 HCs were included for flow cytometry. The clinical information of AOSD patients was provided in the Additional file 1: Table S1 and Additional file 1: Table S2. Heparinized whole blood from AOSD patients and HCs were stained using following antibodies: anti-human CD14 (Biolegend, San Diego, CA, USA, catalog: 325,604), anti-human CD16 (Biolegend, clone:3G8, catalog: 302,012), anti-human CD80 (Biolegend, catalog: 305,221), anti-human CD86 (Biolegend, catalog: 305,431), anti-human CD163 (Biolegend, catalog: 333,608), anti-human HLA-DR (Biolegend, catalog: 307,653). All assays were performed by a FACS Canto II cytometer (BD). Data were analyzed using FlowJo software (Tree Star, Inc., Ashland, OR).

Briefly, heparinized blood from AOSD patients and healthy controls was isolated by density gradient centrifugation on Polymorphprep (Serumwerk Bernburg AG) according to the manufacturer’s instructions. Neutrophils (1 × 10⁶ cells/mL) were cultured in a total volume of 1 ml RPMI 1640 supplemented with 10% FBS for 3.5 h at 37℃. Then, the cells were washed twice with fresh PRMI 1640, and the NETs were collected by extensively pipetting with 1 ml RPMI 1640. Thereafter, the NETs were collected by centrifugation at 400 × g and 17,000 × g. The DNA concentrations of NETs were determined using the Quant-iT PicoGreen double-stranded DNA (dsDNA) assay kit (Invitrogen) according to the manufacturer’s instructions. To remove DNA components in NETs, isolated NETs were treated with DNase I (Sigma Aldrich) to degrade DNA for 15 min.

Peripheral blood mononuclear cells (PBMCs) were isolated from patients with AOSD and HC using Lymphoprep (Serumwerk Bernburg AG) under sterile conditions following the manufacturer’s instructions. Monocytes were isolated with CD14 positive magnetic sorting (Miltenyi Biotec, Bergisch Gladbach, Germany). Total RNA was extracted using TRIzol reagent following the manufacturer’s instructions (Takara, Japan), then qualified by Nano Drop and Agilent 2100 bioanalyzer (Thermo Fisher Scientific, MA, USA). For monocyte subset isolation, CMs and IMs were purified by flow cytometry (BD FACSAriaIII) based on surface CD14 and CD16 staining.

To assess the impact of monocyte subsets on T cell activation, autologous CD4 + T cells were cocultured with either CMs or IMs (5:1 ratio) in the presence of anti-CD3 (1 μg/mL) (BD Biosciences), anti-CD28 (1 μg/mL) (BD Biosciences) antibodies, and macrophage colony-stimulating factor (M-CSF) (50 ng/mL) (MCE, China). After a 5-day incubation, T cells were stimulated for 5 h with phorbol 12-myristate 13-acetate (PMA) (50 ng/mL) (Sigma Aldrich), ionomycin (1 μg/mL) (MCE, China), and brefeldin A (BFA) (10 μg/mL) (MCE, China). Subsequently, T cells were analyzed through intracellular staining using anti-IFN-γ (Th1), anti-17A (Th17) and anti-Foxp3 (Treg) antibodies.

Sorted CMs or IMs were seeded at 5 × 10⁴ cells/well in a 96-well plate, and cell supernatants were harvested after 24 h. Detection of the IL-1β, IL-6 and TNF were measured with Cytokine Bead Array (catalog: 551,811, BD Biosciences, USA). Fifty microliters of the standard dilutions or samples was incubated with capture beads and detection reagent for 3 h at room temperature. Assays were performed by flow cytometry.

Sorted CMs or IMs were seeded at 5 × 10⁴ cells/well in a 96-well plate. pHrodo Green E. coli BioParticles (catalog: P35366, Thermo Fisher Scientific, MA, USA) were added to cells in a 1:10 dilution and incubated for 1 h. The images were visualized using an Olympus microscope (IX73, Tokyo, Japan) and the fluorescence intensity was acquired using a fluorescence plate reader (Biotek Synergy Neo2).

PBMCs were isolated using Lymphoprep (Serumwerk Bernburg AG) under sterile conditions following the manufacturer’s instructions. PBMCs (5 × 10⁶ cells/mL) were cultured in RPMI 1640 supplemented with 10% FBS and were stimulated with NETs isolated from AOSD (AOSD-NETs) or isolated from HCs (HC-NETs) (equal concentration of 100 ng/ml NET-DNA) for 24 h. The PBMCs were collected for quantitative real-time PCR (qRT-PCR) and flow cytometry.

A total of 60 AOSD patients (including 19 AOSD-MAS patients) and 20 HCs were included for plasma CCL8 and CXCL10 measurement. The clinical information of AOSD patients is provided in the Additional file 1: Table S3. CCL8 and CXCL10 levels in plasma and cell supernatants were measured by commercial sandwich enzyme-linked immunosorbent assay (ELISA, Cusabio, China) following the manufacturer’s instructions.

Total RNA was extracted using Trizol reagent following manufacturer’s instructions (Takara, Japan) and reverse-transcribed into cDNA using PrimeScript™ RT Reagent Kit (Takara). qRT-PCR was performed with SYBR Green (Takara). The relative expression levels of mRNA were normalized against GAPDH. Specific primers of human TLR9, MRE11, DDX41, PQBP1, DHX36, CGAS, DHX9, ZBP1, DDX60, AIM2, and IFI16 were used. Primer sequences are listed in Additional file 1: Table S4.

Purified IMs were primed with 100 ng/mL of LPS (Sigma) for 4 h and then stimulated with NETs or 5 mM ATP (Sigma) for 2 h. Following treatment, the media was collected for the quantification of IL-1β (catalog: DY201) and IL-18 (catalog: DY318) using ELISA (R&D, Minneapolis, USA) according to the manufacturer’s instructions. The caspase-1 activity was measured using Caspase-1 Colorimetric Assay Kit (catalog: K111, BioVision, USA). The expression of pro-caspase-1, caspase-1 p20, pro-IL-1β, and IL-1β was examined using Western blotting.

IMs were stimulated as described above and subsequently lysed in RIPA lysis buffer (Beyotime Institute of Biotechnology, Shanghai, China) containing protease inhibitor cocktails (Roche Diagnostics, Mannheim, Germany). Equal quantities of protein (20 μg) were separated using 10% SDS-PAGE and transferred to a PVDF membrane (Millipore, USA). The membranes were incubated with the following primary antibodies: rabbit polyclonal anti-caspase1 (catalog:2225S, 1:1000, CST, USA), rabbit polyclonal anti-IL-1β (catalog: 5128, 1:1000, BioVision, USA), and anti-GAPDH (1:1000, AF1186, Beyotime Institute of Biotechnology, Shanghai, China). Peroxidase was visualized using an enhanced chemiluminescence system (ECL) (Millipore).

All data were statistically analyzed using the Graphpad Prism v8.0 software or R v4.2. Quantitative data are expressed as the means ± SD (standard deviation). Data with a Gaussian distribution was analyzed using an unpaired two-sided t-test, one-way analysis of variance (ANOVA), while nonparametric data were assessed using the Mann–Whitney U test. Spearman’s correlation analysis was used to test the correlations. Receiver operating characteristics (ROC) analysis and least absolute shrinkage and selection operator (LASSO) analysis were used to assess the diagnostic performance. Statistical significance was defined as p < 0.05.

In order to determine the monocyte subsets in AOSD, we employed flow cytometry to test three monocyte subsets and several surface markers including costimulatory receptors (CD80 and CD86), MHC molecules (HLA-DR), and scavenger receptor (CD163) in 37 AOSD patients and 12 HCs. The clinical characteristics of these subjects in each group are detailed in Additional file 1: Table S1 and Additional file 1: Table S2.

As shown in Fig. 1, the frequencies of IMs (CD14^brightCD16 + monocytes) were higher in patients with AOSD than in HCs (AOSD: 22.47 ± 10.89% vs. HCs: 10.36 ± 5.74%, p = 0.0002). Conversely, CMs (CD14^brightCD16 − monocytes) demonstrated a decrease in AOSD patients (AOSD: 68.82 ± 13.16% vs. HCs: 81.67 ± 8.65%, p = 0.0015), while NCMs (CD14^dimCD16 + monocytes) exhibited no significant difference (AOSD: 8.39 ± 6.78% vs. HCs: 7.72 ± 3.76%, p = 0.7050). We then determined the proportions of monocyte subsets in AOSD patients with diverse disease activity. The frequencies of IMs were higher in active AOSD patients compared to those with inactive disease (p = 0.0176). The proportions of CMs were reduced in patients with active AOSD while NCMs did not differ (CMs, p = 0.0134; NCMs, p = 0.5429).

Furthermore, we examined the expression of several cell-surface markers across different monocyte subsets. IMs displayed elevated expression levels of CD80, CD86, HLA-DR, and CD163 relative to CMs and NCMs, indicating a more activated phenotype for IMs (Fig. 2A). Next, we compared these cell-surface markers in AOSD patients and HCs. The expressions of CD80, CD86, and CD163 on total monocytes were indistinguishable between AOSD patients and HCs. The level of HLA-DR on total monocytes was reduced in AOSD patients (Fig. 2B). Regarding IMs, the expression level of CD163 was markedly higher, but the level of HLA-DR was lower in AOSD compared to HCs. No differences in CD80 and HLA-DR were observed on IMs (Fig. 2C).

We also assessed functional distinctions between CMs and IMs from AOSD patients and HCs. Notably, both AOSD CMs and HC CMs exhibited higher phagocytic activity compared to IMs. However, AOSD IMs displayed enhanced phagocytic activity compared to HC IMs, indicating the potential phagocytic ability of activated IMs in AOSD (Fig. 2D). Regarding T cell activation, we observed that HC IMs induced both Th1 cell and Th17 cell differentiation, while AOSD IMs significantly increased Th17 cell frequencies and reduced Th1 cell frequencies. Both CMs and IMs from AOSD patients and HCs did not influence the differentiation of Treg cells (Fig. 2E). Additionally, we found that AOSD CMs produced higher levels of IL-1β, IL-6, and CCL8 than HC CMs. AOSD IMs produced increased levels of IL-1β, IL-6, CCL8, and CXCL10 than HC IMs. Furthermore, we observed that CXCL10 levels were higher in the cell supernatants of AOSD IMs than in those of AOSD CMs (Fig. 2F). These results demonstrate the distinct phenotypes and functions of different monocyte subsets in AOSD patients and HCs.

To assess the associations between IMs and clinical manifestations, AOSD patients were divided into two groups: patients with a high proportion of IMs and those with a low proportion of IMs. Increased frequencies of fever, rash, sore throat, and lymphadenopathy were found in the patients with high IM proportions (Fig. 3A, Additional file 1: Fig. S1A). We also measured the correlation between monocyte subsets and routine clinical parameters of AOSD. We found no correlation between the frequencies of IMs and routine blood tests, as well as liver function tests (Fig. 3B). We subsequently analyzed the correlation between IMs and Pouchot systemic disease activity score, as well as routine laboratory inflammatory markers, such as ferritin levels, C-reactive protein (CRP), and erythrocyte sedimentation rate (ESR). The results showed the proportions of IMs were positively correlated with disease activity score (r = 0.3669, p = 0.0255), ferritin (r = 0.4008, p = 0.0140), and CRP (r = 0.3260, p = 0.0490) (Fig. 3B). The proportions of CMs exhibited a negative correlation with disease activity score (r =  − 0.4856, p = 0.0023), ferritin (r =  − 0.4953, p = 0.0018), and CRP (r =  − 0.4464, p = 0.0056) (Fig. 3B). The proportions of NCMs were positively correlated with ESR (r = 0.3290, p = 0.0468) (Fig. 3B).

We then analyzed the correlation of monocyte subsets with cytokine levels. We found the proportions of IMs were positively correlated with soluble IL-2R levels (r = 0.5309, p = 0.0193). In view of the elevation of sIL-2R in AOSD-MAS, we next investigated the association between monocyte subsets and MAS. Among 37 AOSD patients, 10 patients could be diagnosed as AOSD-MAS according to the 2014 systemic sJIA-MAS criteria. The frequencies of IMs were significantly elevated in AOSD-MAS compared to those without MAS (Fig. 3C, p < 0.0001). Moreover, HScore, a scoring system for estimating secondary hemophagocytic syndrome [23], was correlated with the proportions of IMs (Fig. 3D). ROC plot of the frequency of IMs, as a predictor of MAS in AOSD patients, had an area under the curve (AUC) of 0.9111 (p = 0.0001), indicating a critical role of IMs in the development of MAS in AOSD patients (Fig. 3E).

In order to explore the potential mechanism that contributes to the phenotype change of monocytes in AOSD-MAS, we conducted RNA-seq on blood monocytes from 10 HCs and 12 patients with AOSD (9 with AOSD-MAS, 3 with AOSD without MAS). A total of 578 upregulated DEGs were found in AOSD-MAS compared to the HC group, with only 19 of these genes were upregulated in the AOSD group (Fig. 4A). Gene ontology (GO) enrichment analysis revealed pathway enrichment in the AOSD-MAS group, particularly in genes involved in defense response to viruses and cytokine-mediated signaling pathways (Fig. 4B). To investigate the characteristics of monocyte during the development of AOSD to AOSD-MAS, we employed Gene Set Enrichment Analysis (GSEA) to compare the upregulated marker genes of CMs, IMs, and NCMs as previously reported. Notably, monocytes from the AOSD-MAS group showed a higher normalized enrichment score (NES) of “Intermediate monocyte” gene set than “Classical monocyte” and “Non-classical monocyte”. This finding suggests that monocytes from AOSD-MAS patients exhibited a stronger IM signature (Fig. 4C). Furthermore, KEGG analysis revealed upregulation of pathways related to viral infections. GO analysis also revealed an increase in viral immune response in the AOSD-MAS group compared to the AOSD group (Fig. 4D). Furthermore, we identified CCL8 and CXCL10 as potential biomarkers for AOSD-MAS because they were the most upregulated secretory molecules in AOSD-MAS patients in the response to virus pathway (Fig. 4E).

To determine whether CCL8 and CXCL10 could be biomarkers in distinguishing MAS in AOSD patients, we then measured plasma levels of CCL8 and CXCL10 from the second cohort. The characteristics of these subjects are shown in Additional file 1: Table S3. Compared with HCs, we detected higher plasma levels of CCL8 and CXCL10 in AOSD patients (Fig. 5A,B). Moreover, plasma CXCL10 levels were significantly higher in AOSD-MAS patients than those in AOSD patients but plasma CCL8 levels did not (Fig. 5B). And plasma CXCL10 levels were positively correlated with ferritin, sIL2R, and Hscore (Fig. 5C). We also found that CXCL10 levels were correlated with hemoglobin (Hb), platelet (PLT), alanine transaminase (ALT), aspartate transaminase (AST), and fibrinogen (Fg) levels (Fig. 5D-E). Besides, plasma CXCL10 levels were higher in patients with fever, splenomegaly, lymphadenopathy, and pleuritis (Additional file 1: Table S5). These data suggest a close relationship between plasma CXCL10 levels and AOSD-MAS. The ROC of plasma CXCL10 for predicting MAS had an AUC of 0.9101 (Fig. 5F). The AUC of ferritin, sIL2R, and Hscore were 0.8723, 0.9127, and 0.9588, respectively (Fig. 5F). By applying the LASSO algorithm to the CCL8, CXCL10, and 8 clinical variables (including WBC, Hb, PLT, ALT, AST, TG, Fg, ferritin and sIL2R), 7 variables (including CXCL10, Hb, PLT, AST, TG, ferritin and sIL2R) were selected by LASSO for predicting MAS, with an AUC of 0.9615 (Fig. 5G), highlighting the value of CXCL10 in the prediction and evaluation of AOSD-MAS.

Next, we explored the regulatory mechanism underlying monocyte alteration. GSEA identified an upregulation of the cytosolic DNA-sensing pathway in the AOSD-MAS group compared to the AOSD group (Fig. 6A). Through gene set variation analysis (GSVA), we discovered a strong correlation between the DNA-sensing pathway and the IM signature (Fig. 6B). Single-gene analysis of DNA sensors revealed that DDX60 and IFI16 were significantly elevated in the AOSD-MAS group compared to the AOSD group, while ZBP1 and AIM2 demonstrated an upward trend (Fig. 6C,D). Correlation analysis also revealed a positive correlation between IM signature and AIM2, DDX60, and IFI16 (Additional file 1: Fig. S2A), suggesting that the DNA-sensing pathway may play a role in the expansion of IMs in AOSD-MAS.

Finally, in order to investigate the effects of NETs on monocytes, we stimulated PBMCs from healthy donors with NETs isolated from active AOSD patients (AOSD-NETs) and HCs (HC-NETs) in vitro. Remarkably, AOSD-NETs induced surface CD16 expression and IM expansion, whereas HC-NETs did not (Fig. 6E-G). Upon AOSD-NETs stimulation, the human leukemia monocytic cell line THP1 showed enhanced expression of IFI16 and DDX60 relative to HC-NETs, which was consistent with the expression pattern observed in AOSD patients (Additional file 1: Fig. S3A-B). We also confirmed the increased expression of IFI16 and DDX60 following AOSD-NETs stimulation (Fig. 6H). These findings suggest that the DNA components present in AOSD-NETs may induce CD16 expression. We treated AOSD-NETs and HC-NETs with DNase I to remove DNA components. AOSD-NETs without DNA displayed reduced capacity to induce the expression of IFI16 and DDX60, surface CD16 expression, and IM expansion (Fig. 6I–K). These results highlight the contribution of DNA components contained in AOSD-NETs to the regulation of monocyte subsets in AOSD patients. Given the crucial role of NET-DNA in activating the inflammasome pathway, we also examined the levels of inflammasome pathways in IMs stimulated by NETs. We exposed sorted IMs to AOSD-NETs, AOSD-NETs without DNA, HC-NETs, HC-NETs without DNA, and a positive control, ATP. Following stimulation with AOSD-NETs, IMs displayed elevated levels of caspase-1 p20 and mature IL-1β, as detected by Western blot (Fig. 6L). Moreover, AOSD-NETs significantly increased caspase-1 activity levels, while AOSD-NETs without DNA did not (Fig. 6M). Furthermore, we found that the levels of IL-1β in the cell supernatants were elevated in IMs with AOSD-NETs, although IL-18 levels did not reach statistical significance (Fig. 6N). These findings suggest that DNA components in AOSD-NETs activate the inflammasome pathway in IMs.