Neutrophil extracellular traps exacerbate Th1-mediated autoimmune responses in rheumatoid arthritis by promoting DC maturation

Abstract

Aberrant formation of neutrophil extracellular traps (NETs) is a key feature in rheumatoid arthritis (RA) and plays a pivotal role in disease pathogenesis. However, the mechanism through which NETs shape the autoimmune response in RA remains elusive. In this study, we demonstrate that inhibition of peptidylarginine deiminases activity in collagen-induced arthritis (CIA) mouse model significantly reduces NET formation, attenuates clinical disease activity, and prevents joint destruction. Importantly, peptidylarginine deiminase 4 blocking markedly reduces the frequency of collagen-specific IFN-γ-producing T helper 1 (Th1) cells in the draining lymph nodes of immunized mice. Exposure of dendritic cells (DCs) to CIA-derived NETs induces DC maturation characterized by significant upregulation of costimulatory molecules, as well as elevated secretion of IL-6. Moreover, CIA-NET-treated DCs promote the induction of antigen-specific Th1 cells in vitro. Finally, NETs from RA patients show an increased potential to induce the maturation of DCs from healthy individuals, corroborating the findings obtained in CIA mouse model. Collectively, our findings delineate an important role of NETs in the induction and expansion of Th1 pathogenic cells in CIA through maturation of DCs and reveal a novel role of NETs in shaping the RA-autoimmune response that could be exploited therapeutically.

Keywords: Autoimmune responses; Cl-amidine; Collagen-induced arthritis; Neutrophil extracellular traps; Rheumatoid arthritis.

Figure 1

Enhanced spontaneous NET formation in RA peripheral blood PMNs and in CIA bone marrow. (A) Confocal microscopy images of RA (right panel) versus healthy PMNs (left panel). NETs were visualized as structures costained for NE (green) and DAPI (blue). Original magnification, 40×; scale bar, 25 μm. Images are from a single experiment representative of ten independent experiments with 2–3 donors per experiment. (B) Quantification of NET release by RA PMNs (n=30) versus healthy untreated PMNs (n=7) under microscopy observation. PMA was used as positive control. Results are expressed as mean ± SEM and are combined from five independent experiments with 2–5 donor samples per experiment. (Unpaired t-test, *p=0.018, ****p<0.0001). (C) Percentage of NET releasing cells upon treatment with RA serum (n=10) and RA synovial fluid (SF) (n=8). Healthy serum (HS) (n=7) was used as negative control. Results are expressed as mean ± SEM. Data are combined from three independent experiments with PMNs from 2–3 donors per experiment. (Unpaired t-test, *p=0.010, ****p<0.0001). (D) Confocal microscopy images of BM PMNs (Ly6G⁺) derived from diseased CIA and control (naïve) mice. Original magnification, 63×; digital zoom, 3×. Scale bar; 25μm. Quantification of NET release by BM PMNs derived from diseased CIA mice versus control mice (n=5 mice/group) under microscopy observation based on NE and DAPI costaining (lower panel). PMA was used as positive control. Results are expressed as mean ± SEM and are representative from four independent experiments with 5 mice/group per experiment. (Unpaired t-test, ****p<0.0001). (E) Confocal microscopy analysis for NET formation by naïve BM PMNs treated with CIA diseased and naïve plasma. DAPI (blue) and NE (green) stainings are shown. Original magnification, 63×; digital zoom, 3×. Scale bar; 11μm. Quantification of NETs released by naive BM PMNs treated with CIA diseased (n=5) or naïve plasma (n=5) under microscopy observation. Results are expressed as mean ± SEM and are representatives from four independent experiments using plasma isolated from five mice per group. (Unpaired t-test. ****p<0.0001).

Figure 2

Cl-amidine treatment during CIA reduces clinical disease activity and joint destruction. (A) Outline of the experimental setup. DBA/1J mice were injected intradermally (i.d.) at the base of the tail with bovine collagen type II (bCII) emulsified in complete Freund’s adjuvant (CFA) with the addition of inactivated Mycobacterium tuberculosis on d0, followed by a boost immunization of bCII emulsified in incomplete Freund’s adjuvant (IFA) on d21. Each experimental group consisted of eight mice that were treated daily intraperitoneally (i.p.) with Cl-amidine or PBS/DMSO (vehicle control) after the d21 booster injection through sacrifice on d35–45 (B) Quantification of NET release by BM PMNs (Ly6G⁺) derived from diseased CIA-DMSO mice (n=8) versus CIA-Cl-amidine administered mice (n=6) under microscopy observation based on NE and DAPI costaining. Results are expressed as mean ± SEM and are combined from three independent experiments with 2–3 mice/group per experiment. (Unpaired t-test, **p<0.0065). (C) Evaluation of arthritis based on redness and swelling of the paws (n=5/group; day 26: *p = 0.017, day 27: *p =0.007, day 29: *p =0.043 day 30: *p = 0.02 day 31: *p = 0.03, Multiple tests, unpaired t-test) and time (days) of CIA onset are shown. (Mann-Whitney test, *p =0.023). Maximum score per mouse is 38. Results are expressed as mean ± SEM and are representatives of three independent experiments with 5 mice/group per experiment. (D) Representative H&E-stained sections of inflamed knee joints (proximal) from CIA-DMSO and CIA-Cl-amidine treated mice, are shown, at the peak of disease activity (d32–39) (original magnification 400×). Two representative samples of CIA-DMSO and CIA-Cl-amidine administered mice, are shown. Data shown are from one experiment representative of three independent experiments with 2 mice/group per experiment. (E) Anti-bCII Ab production in the plasma of CIA-DMSO (n=9) or CIA-Cl-amidine (n=10) administered mice, as measured by ELISA. Results are expressed as absorption at 450nm. Each dot represents one mouse. Mean is depicted. Data are combined from three independent experiments with 3–4 mice/group per experiment. (Mann-Whitney test, *p =0.043).

Figure 3

Cl-amidine treatment attenuates the expression of MHC-II and CD86 costimulatory molecules on DCs derived from draining lymph nodes (dLNs). dLNCs from bCII-injected DMSO and bCII-injected Cl-amidine treated mice were collected 12 d post collagen injection and analyzed by flow cytometry. (A) Relative numbers of CD4⁺ T cells/5 × 10⁵ total LNCs (n=18–21/group) and the percentage of CD4⁺ IFN-γ⁺ and CD4⁺IL-17⁺ cells (n=14–17/group) between the two groups are shown. Results are expressed as mean ± SEM. Data are combined from four independent experiments with 4–5 mice/group per experiment. (Unpaired t-test, **p=0.0031, *p=0.04; ns=non significant). (B) Representative flow cytometry analysis and relative numbers of CD11c⁺ DCs/5 × 10⁵ total LNCs are shown. Numbers in FACS plot denote frequency. Representative FACS histograms depicting the geometric mean fluorescence intensity (MFI) of MHC-II and CD86 expression between the two groups, are shown, as examined by flow cytometry. Data shown are from one experiment representative of three independent experiments with 2–3 mice/group per experiment. Relative numbers of CD11c⁺ cells/5 × 10⁵ total dLNCs (n=19–21/group) between the two groups are shown. Results are expressed as mean ± SEM. Data are combined from four independent experiments with 4–5 mice/group per experiment. (Unpaired t-test, ns=non significant). Dot plots represent the MFI of dLNCs of each CIA-DMSO or CIA-Cl-amidine administered mouse (n=8–9) normalized to the average MFI of CIA group (MFI fold change). Mean is depicted. Data are combined from three independent experiments with 2–3 mice/group per experiment. (Unpaired t-test, ****p<0.0001).

Figure 4

CIA BM-derived NETs increase the inflammatory properties of myeloid dendritic cells. Bone marrow-derived dendritic cells (BMDCs) from naïve mice were treated with condensed supernatants containing NETs (CIA-NETs) and control supernatants (control). (A) Representative confocal microscopy image from CIA-NETs after costaining with DAPI and MPO. Original magnification; 63×, digital zoom; 3×. Scale bar; 3μm. Microscopy image is representative of five independent experiments. (B) Representative flow cytometry analysis of CD11c⁺ DCs. Numbers in FACS plot denote frequency. Representative FACS histograms are shown with the corresponding MFI of CD80 and CD86 expression levels on CD11c⁺ cells. Data shown are from one experiment representative of four independent experiments. Dot plots represent expression levels of CD80 and CD86 of BMDCs treated with CIA-NETs or control (n=9–12) normalized against the expression of untreated BMDCs (fold induction). Results are expressed as mean; data are combined from four independent experiments. (Unpaired t-test, **p=0.013, *p=0.04). (C) IL-6 secretion in culture supernatants of BMDCs treated with CIA-NETs (n=12) or control (n=10) expressed as fold increase based on the secretion of the untreated BMDCs. Results are expressed as mean ± SEM; data are combined from four independent experiments. (Unpaired t-test, *p=0.038).

Figure 5

CIA-NETs treated BMDCs augments antigen (Ag)-specific Th1 responses in vitro. (A) Outline of the coculture experimental setup. BM from C57BL/6 (B6) mice was differentiated into BMDCs (CD11c⁺) in the presence of GM-CSF. At day 9 the immature BMDCs were pulsed with ovalbumin (OVA) protein in the presence or absence of CIA-NETs or control supernatants for 18hrs and subsequently cocultured with LNCs from OT-II transgenic mice. 48hrs later IFN-γ and IL-17 production was assessed by intracellular staining. (B) Gating strategy of CD4⁺ T cells and CD4⁺IFN-γ⁺ cells are shown. Numbers on the gates denote frequencies. Gates were set as indicated. Percentages of CD4⁺ IFN-γ⁺ and CD4⁺ IL-17⁺ cells in total LNCs derived from OT-II mice pulsed with OVA and treated with CIA-NETs (n=4) or control (n=6), are depicted. Results are expressed as mean ± SEM; data are combined from two independent experiments. (Unpaired t-test, *p=0.02).

Figure 6

Human rheumatoid arthritis-NETs (RA-NETs) increase the expression of costimulatory molecules on human monocyte differentiated dendritic cells (moDCs) and the secretion of proinflammatory cytokines. (A) Representative FACs histograms are shown with the corresponding MFI of moDCs treated with RA NETs or control. Dot plots represent the MFI of HLA-DR and CD86 of moDCs treated with RA-NETs (n=7–8) or control supernatants (n=4–6) normalized against the average of MFI from untreated moDCs (MFI fold change). Results are expressed as mean; data are combined from three independent experiments. (Unpaired t-test, *p=0.04). (B) IL-6 and TNF-α secretion in culture supernatants of moDCs treated with RA-NETs (n=7–9) or control supernatants (n=6) expressed as fold increase based on the secretion of the untreated moDCs. Results are expressed as mean ± SEM; data are combined from three independent experiments. [Unpaired t-test, *p=0.01 (for TNF-α) and *p=0.04 (for IL-6)].