Tumour necrosis factor alpha promotes secretion of 14-3-3η by inducing necroptosis in macrophages

Background 14-3-3η is an intracellular protein also detected in the serum and synovial fluid of patients with rheumatoid arthritis (RA). It is closely related to disease activity and anti-cyclic citrullinated peptide antibody levels. However, the main source of 14-3-3η and the mechanism of its release into the extracellular space remain unclear. Addressing these two points was the main goal of the current study. Methods The source of 14-3-3η was investigated by immunostaining RA synovial tissue. Fibroblast-like synoviocytes, CD4+ cells, and macrophages were selected as candidates among the various cell types in the synovial tissue. Phosphorylation of mixed-lineage kinase domain-like pseudokinase (MLKL) and cell death of macrophages were studied by phalloidin staining and electron microscopy after stimulation with an oxidative stress inducer (diamide) or tumour necrosis factor (TNF)-α. Extracellular 14-3-3η protein levels were examined by western blotting. Results Macrophages from the synovial tissue from RA, but not osteoarthritis, showed dense and widespread cytoplasmic staining for the 14-3-3η protein, co-localized with peptidylarginine deiminase 4. Swelling and membrane rupture of macrophages were induced by treatment with TNF-α, but not interleukin (IL) 6/soluble IL-6 receptor (sIL-6R). Increased MLKL phosphorylation followed by necroptosis was also induced in TNF-α-stimulated macrophages. Necrostatin-1, a necroptosis inhibitor, antagonized MLKL phosphorylation. High levels of 14-3-3η were detected in the culture supernatants of macrophages stimulated with diamide and TNF-α, but not IL-6/sIL-6R. Conclusions Macrophages that highly express 14-3-3η undergo TNF-α-induced necroptosis with damage to the cellular structure, resulting in the secretion of 14-3-3η into the extracellular space. The current study provides a novel mechanism for 14-3-3η level increase in the RA synovial fluid.


Background
Rheumatoid arthritis (RA) is a systemic autoimmune disease characterized by chronic inflammation of the synovium and joint destruction [1][2][3][4][5]. Initially, the levels of inflammatory cytokines, such as tumour necrosis factor (TNF)-α and interleukin (IL) 6, and proteases, such as matrix metalloproteinase (MMP) 3, are increased in the synovial fluid of patients with RA. Lymphocyte infiltration of the synovial tissue plays an important role in the pathogenesis of the disease [6,7]. Therefore, TNF inhibitors, IL-6 receptor (IL-6R) inhibitors, cytotoxic T lymphocyte-associated protein 4, and low-molecular compound Janus kinase (JAK) inhibitors have been developed and used clinically [8][9][10][11][12]. However, not all patients with RA achieve remission, and some may relapse. To address this problem, it is important to find new molecules that play a crucial role in the pathogenesis of RA and could constitute new therapeutic targets. Chaperone 14-3-3η molecules, which are usually located intracellularly, are unexpectedly detected in the synovial fluid of RA patients and thus attracted scientific attention [13]. 14-3-3η molecules have been extracted from the brain of cattle and are detected in many other species [14]. 14-3-3 proteins are generally localized in the nucleus, cytoplasm, and cell membrane [15]. Seven isoforms of these proteins have been identified (β, ε, η, γ, τ, σ, and ζ), encoded by various genes in mammals. Further, the 14-3-3 protein binds to target proteins and regulates their modification, cellular localization, and enzymatic activity. In many cases, 14-3-3 appears to interact with a specific phosphorylated threonine-serine motif [15]. 14-3-3 proteins are involved in diverse biological events, such as cell cycle progression, intracellular protein trafficking, apoptosis, DNA damage repair, DNA replication, and transcriptional regulation [16]. It has been reported that siRNA-mediated suppression of 14-3-3 expression in a glioma cell line results in the suppression of cell proliferation, followed by apoptosis [17]. These observations indicate that the 14-3-3 protein plays an important role in the regulation of various intracellular events.
The importance of 14-3-3η in pathological processes relevant to RA was first proposed when 14-3-3η was detected in the serum of patients with early RA. Subsequently, the protein was found to be an early diagnostic marker of RA, showing complementarity with anticitrullinated protein antibodies (ACPA) and rheumatoid factor [13,[18][19][20]. We have reported that the serum levels of 14-3-3η correlate with disease activity markers (DAS28-ESR, CDAI, and SDAI scores) in patients with RA and also with the serum levels of ACPA. Adalimumab (ADA), an anti-TNF antibody, but not tocilizumab (TCZ), an anti-IL-6R antibody, reduces serum levels of 14-3-3η and ACPA [21]. Since ACPA is an autoantibody that targets proteins citrullinated by peptidylarginine deiminase 4 (PAD4), PAD4 may induce 14-3-3η citrullination. Furthermore, when the human monocytic cell line THP1 is stimulated with recombinant 14-3-3η, the expression of genes such as ones encoding TNF-α, IL-6, MMP-3, and RANKL is induced [22]. In addition, 14-3-3η appears to be a mediator of the RA pathogenesis as well as a diagnostic biomarker.
Of note, 14-3-3η is detected in the central nervous system and synovial tissue [18,23]. However, the cell types that are the primary source of 14-3-3η in patients with RA and the mechanism of 14-3-3η secretion into the joint fluid remain unknown. 14-3-3η may occur in the RA synovial fluid as a result of exocytosis or cell death. In RA, apoptosis of inflammatory cells is reduced, and the expression of necrosis markers in a collagen-induced mouse model of arthritis is increased in cells proximal to the joints [24]. The pro-inflammatory cytokine TNF-α, which is considered a key cytokine for the pathogenesis of RA, causes death of various cell types. In particular, its involvement in necroptosis attracts much attention [25][26][27]. Therefore, we hypothesized that necroptosis may be involved in the secretion of 14-3-3η. In the current study, we aimed to define the source of the 14-3-3η protein by immunostaining the synovial tissue and to define the mechanism of cell death of macrophages and 14-3-3η secretion in patients with RA.

Cell culture
Monocytes were purified from Ficoll-isolated peripheral blood mononuclear cells and cultured in 6-well plates at a density of 100,000 cells/cm 2 in RPMI-1640 medium (Wako, Wakayama, Japan). They were differentiated into macrophages by treating with 100 ng/ml macrophage colony-stimulating factor (Peprotech, Rocky Hill, NJ). Macrophages were cultured in RPMI-1640 medium containing 10% (v/v) foetal bovine serum (Tissue Culture Biological, Tulare, CA) and antibiotics (1% (v/v) penicillin and streptomycin, Life Technologies, Carlsbad, CA), seeded at a density of 100,000 cells/cm 2 in 6-well plates and incubated at 37°C under 5% CO 2 atmosphere, as described previously [28].

Histological examination
The synovial tissue of patients (ethical approval was obtained from the University of Occupational and Environmental Health, Japan Ethics Committee) with RA and osteoarthritis (OA) was obtained during operation, fixed in 10% formaldehyde, and embedded in paraffin. Samples were subjected to histological examinations, including IHC and IF. IHC was performed as described previously, with slight modifications [29]. Briefly, antigen retrieval was performed by soaking tissue specimens from surgery on slides in 5 mM sodium citrate solution in phosphatebuffered saline (PBS) containing 0.05% (v/v) Tween 20 (PBST) (pH 6.0). Slides were blocked with serum-free protein block (Dako, 2016-08) for 60 min at 25°C and then incubated for 2 h with mouse monoclonal antibodies specific for 14-3-3η and 14-3-3γ, which were diluted 1:200 in Can Get Signal Immunostain Solution A (NKB-501; Toyobo, Osaka, Japan). The slides were washed three times with PBST, incubated with antimouse IgG secondary antibodies conjugated with HRP polymer (DakoCytomation, Glostrup, Denmark) for 1 h, and then visualized by treatment with the DAB chromogen (3,3′ diaminobenzidine, DakoCytomation, #K3465) for 8 min, according to the manufacturer's instructions. The nucleus was visualized using Mayer's haematoxylin (1:1000 dilution in PBST; Merck, Darmstadt, Germany). For mounting, the sections were rinsed in water, dehydrated in graded ethanol series (90% ethanol for 30 s × 3 and 100% ethanol for 30 s × 3), cleared in xylene (for 30 s × 2), and sealed using Multi mount 480 (FM48001; Matsunami, Kishiwada, Japan). IF was performed as described previously [30]. Biopsy specimens were incubated with the respective antibodies specific against 14-3-3η, PAD4, CD4, CD55, and CD68. The nucleus was stained with DAPI (5 μg/ml).

Transmission electron microscopy
The experiment was performed as described previously [31]. Macrophages were fixed with PBS containing 2% glutaraldehyde and 2% paraformaldehyde, followed by incubation in 1% OsO 4 , and dehydrated in a graded alcohol series. After freeze-drying in t-butyl-alcohol, the samples were examined under a transmission electron microscope (TEM) (TM3000; Hitachi, Tokyo, Japan).
Proteins in the gel were stained with Coomassie Brilliant Blue (CBB) R-350 (GE Healthcare, Chicago, IL) according to the manufacturer's instruction.
WB protein bands were quantified by using ImageJ software as described previously [32].

Statistical analysis
All quantitative data are expressed as the mean ± standard deviation (SD). Differences between the two groups were tested for statistical significance by Student's unpaired two-tailed t test; P value < 0.05 was considered significant. Statistical analyses were performed using SPSS statistical software (version 21.0; SPSS, Inc., Chicago, IL).

Results
The synovial tissue specimens from RA and OA individuals were analysed by IHC using an antibody against 14-3-3η (Fig. 1). Patient information is provided in Additional file 1: Table S1. 14-3-3η was detected in the surface layer as well as inside the RA synovial tissue. By contrast, it was detected only in the surface layer in OA synovial tissue. Further, 14-3-3γ was only weakly detected in the RA synovial tissue, although it was detected in both the surface layer and inside the OA synovial tissue. The IgG isotype control did not react with the synovial tissue.
Macrophages derived from peripheral blood mononuclear cells of healthy donors (HD) were next treated with an oxidative stress inducer diamide. The cytoskeletal protein phalloidin was then detected using a fluorescent-labelled antibody (Fig. 4a). Diffuse staining was observed at 0 min and 20 min; by contrast, after 24h treatment, the staining was only apparent surrounding the nucleus. When the cells were stimulated with IL-6/ sIL-6R, phalloidin was detected in the cytoplasm; however, it was detected only surrounding the nucleus after TNF-α stimulation. Pre-treatment of macrophages with nec-1, a necroptosis inhibitor, blocked the effect of TNF-α. However, the non-selective JAK inhibitor TOF did not antagonize necroptosis of macrophages, indicating that the JAK-STAT pathway was not secondarily involved in necroptosis induced by TNF-α. Next, macrophages were stimulated with TNF-α or IL-6/sIL-6R for 24 h and probed by double-fluorescent immunostaining to detect PAD4 and 14-3-3η (Fig. 4b). In unstimulated macrophages from HD, PAD4 and 14-3-3η were clearly detected in both the nucleus and cytoplasm, and the isotype control IgG did not stain the cells. Stimulation of macrophages with IL-6/sIL-6R resulted in cells with similar staining features as those of cells without stimulation. By contrast, PAD4 and 14-3-3η were detected mainly in the nucleus of macrophages stimulated with TNF-α. This suggested a role of TNF-α in the cellular distribution of PAD4 and 14-3-3η. The effect was not associated with the upregulation of 14-3-3η Fig. 2 14-3-3η is expressed in CD68 + cells but not in CD4 + T cells from the synovium of RA patients. Synovium specimens of patients with RA (n = 3) were analysed by IF using specific antibodies against 14-3-3η and CD68 (a) or CD4 (b). c Lung tissue specimens of patients with RA (n = 3) were analysed using specific antibodies against 14-3-3η and CD68. Representative images of three independent experiments are shown. Scale bar, 50 μm protein levels, as TNF-α, IL-6/sIL-6R, IL-1β, and IL-21 did not affect the expression of endogenous 14-3-3η ( Fig. 4c and Additional file 2: Figure S1).
Next, the morphology of macrophages stimulated with various cytokines was observed by TEM (Fig. 5a). Unstimulated macrophages showed normal cell structure with a small circular nucleus. Similar cell morphology was observed after stimulation with IL-6/sIL-6R. By contrast, chromatin aggregation, nuclear fragmentation, plasma membrane blebbing, and cell shrinkage, consistent with apoptosis, were observed in actinomycin D-treated macrophages (Fig. 5a). Further, organelle and cell swelling was apparent after diamide treatment (upper panel in Fig. 5a), consistent with necroptosis, with a rupture of the cell membrane (lower panel in Fig. 5a). Organelles of macrophages treated with TNF-α were tumefied, similar to those treated with diamide (upper panel in Fig. 5a). The macrophage cell and intracellular organelles were enlarged, and further membrane rupture led to the release of the nucleus extracellularly and extracellular compartments intracellularly (lower panel in Fig. 5a). Nec-1, a necroptosis inhibitor, blocked the necroptotic effect of TNF-α (Additional file 2: Figure S2). TNF inhibitors ETN and ADA, and the JAK inhibitor TOF did not improve the cell morphology (organelle swelling) after diamide treatment of HD-derived macrophages (Additional file 2: Figure S3 and Figure S2) and RA-derived macrophages (data not shown, Additional file 2: Figure S4). These observations suggest that diamide and TNF-α both induced necroptosis in macrophages, via different pathways. Blocking TNF-α by the TNF inhibitor, but not JAK inhibitor, had a little effect on necroptosis triggered by TNF-α.
Next, we examined whether the signalling pathway involved in necroptosis was related to one of the key necroptosis proteins, receptor-interacting serine/threonine-protein kinase (RIP) 3. We analysed the phosphorylation of RIP3 (p-RIP3) in macrophages stimulated with TNF-α, diamide, and LPS/zVAD-FMK by ICC. No p-Akt (a control, serine/threonine-specific protein kinase) was detected in macrophages treated with TNF-α (Additional file 2: Figure S5). MLKL phosphorylation downstream of RIP3 (Fig. 5b) was next investigated by WB. Whereas no p-MLKL was detected in unstimulated macrophages, p-MLKL was observed from 2 to 24 h after TNF-α stimulation, with the p-MLKL signal reaching the maximum intensity after 8 h. Although the total amount of MLKL remained unchanged, MLKL phosphorylation was not detected in macrophages stimulated with IL-1β, IL-6/sIL-6R, or IL-21 (Additional file 2: Figure S6), or in cells treated with nec-1 prior to TNF-α  Figure S7). In addition, some CD68 staining of the RA-derived synovial tissue was detected. While the membranes of some cells showed widespread and strong CD68 staining, others showed intermittent positive CD68 staining, indicating the possible cell death (data not shown).

Discussion
In the current study, we examined the source of 14-3-3η in the RA synovial tissue by immunostaining and attempted to clarify the mechanism of macrophage cell death and 14-3-3η secretion into the inflamed synovia from patients with RA. We predicted two possible 14-3-3η secretion mechanisms: (1) exocytosis and (2) spontaneous secretion of the 14-3-3η protein after cell death. First, macrophages derived from HD monocytes were cultured in vitro. No 14-3-3η was detected in the culture supernatants in the presence or absence of various cytokine stimuli, such as IL-1β, IL-6/sIL-6R, and IL-21, except for TNF-α. Therefore, we concluded that the mechanism of 14-3-3η secretion is independent of exocytosis, since exocytosis does not cause secretion of 14-3-3η into the extracellular space. Next, we focused on TNF-α, which is important for the RA pathogenesis and elicits different types of cell death (apoptosis, necrosis, and necroptosis) depending on the cellular circumstances. TEM analysis revealed that resting macrophages or macrophages stimulated with IL-6/sIL-6R or TNF-α were not apoptotic. Furthermore, p-RIP3 and p-MLKL, markers of necroptosis, were detected in macrophages stimulated with TNF-α, indicating the role of TNF-α in macrophage necroptosis. Collectively, macrophages stimulated with TNF-α may suffer from necroptosis. It has been reported that the level of glutathione is reduced in various cell types in patients with RA [33,34].  , n = 3), or TNF-α (100 ng/ml; n = 3) for 24 h and then analysed by TEM. b WCLs prepared from macrophages cultured with or without TNF-α (100 ng/ml; n = 3) were analysed by IB using specific antibodies against the total or phosphorylation form of MLKL or β-actin. Quantification data for representative images from three independent experiments (n = 3) are shown. Scale bar, 5 μm (upper panel), 2 μm (lower panel) Fig. 6 14-3-3η is detectable in culture supernatants of macrophages derived from HD and treated with TNF-α. The culture supernatants of macrophages cultured with or without diamide, TNF-α (100 ng/ml; n = 3), IL-1β (10 ng/ml; n = 3), IL-6/sIL-6R (10 ng/ml; n = 3), or IL-21 (10 ng/ml; n = 3) were subjected to WB. Recombinant 14-3-3η was used as a positive control. BSA was used as a loading control and was stained with CBB-R350. Representative images from three independent experiments are shown We found that cell death induced in macrophages using diamide (which specifically oxidizes glutathione), LPS, or TNF-α was characterized by nuclear and organelle swelling, which was very similar to the events in cells undergoing necroptosis [35]. Based on the presented results, it appears that the cell death of macrophages in RA synovium occurs through necroptosis.
In addition, in a previous clinical study, we showed that the TNF-α inhibitor ADA strongly reduces the 14-3-3η levels in the serum of RA patients even before the patients achieve remission [21]. This suggested that the TNF-α pathway is closely related to the extracellular presence of 14-3-3η. To check the effect of TNF-α on 14-3-3η secretion, we here performed an ex vivo experiment using macrophages derived from HD monocytes. We observed that TNF-α stimulation induced necroptosis-like death of macrophages and 14-3-3η secretion into the culture supernatant. Taking together, we propose that in vivo, 14-3-3η is secreted by macrophages via necroptosis elicited by TNF-α.
It has been reported that a subset of genes, such as ones encoding TNF-α, IL-6, and MMP-3, is induced in the human monocytic cell line THP1 upon stimulation with recombinant 14-3-3η, resulting in inflammation [22]. The extracellular 14-3-3η may undergo citrullination, but the significance of this modification remains unknown. In a recent study, an immune complex consisting of extracellular citrullinated fibrinogen and ACPA was shown to activate macrophages by cross-linking with the TLR4 receptor and Fcγ on macrophages, which led to a strong induction of TNF-α [36]. Citrullination is an irreversible modification and is considered a crucial event that triggers the production of ACPA in RA. Although high 14-3-3η levels are observed in the RA synovial fluid, the mechanism of citrullinated 14-3-3η generation remains unknown. It has been reported that TNF-α activates PAD4 in the human oligodendroglial cell line, leading to the translocation of the latter to the nucleus [24]. We have previously shown that the serum levels of 14-3-3η are associated with the ACPA levels in patients with RA [21]. In the current study, we observed the co-localization of PAD4 and 14-3-3η in the nucleus of TNF-α-stimulated macrophages, indicating that 14-3-3η might be subjected to citrullination by TNF-α. 14-3-3 proteins play various roles in the cells, e.g., may be involved in gene expression and proteolysis. Hence, if the 14-3-3 protein undergoes citrullination, new target molecules of a possible citrullinated 14-3-3η could be proposed, which would attract the interest of researchers studying the 14-3-3 protein or ACPA production [16].

Conclusions
The intracellular protein 14-3-3η is detected in the extracellular milieu in the synovial fluid of patients with RA. The goal of the current study was to identify the main source of 14-3-3η and the mechanism of its secretion into the extracellular space. Macrophages in the synovial tissue from RA, but not OA, exhibited widespread cytoplasmic staining for the 14-3-3η protein.
Treatment with TNF-α, but not IL-6/sIL-6R, induced macrophage swelling and membrane rupture. This was followed by necroptosis. Further, high levels of 14-3-3η were detected in the culture supernatants of macrophages stimulated with TNF-α. These observations indicate that TNF-α promotes secretion of 14-3-3η by inducing necroptosis in macrophage, which constitutes a novel mechanism for 14-3-3η level increase in the RA synovial fluid.