Introduction
Familial Mediterranean fever (FMF) is a prototypical autoinflammatory syndrome associated with phagocytic cell activation. FMF is traditionally seen as an autosomal-recessively inherited disorder caused by mutations in the MEFV gene located on chromosome 16p that encodes a 781 amino acid protein known as pyrin (or marenostrin) ([
1] International FMF Consortium, Cell, 1997 [
2]; French FMF Consortium, Nat Gen, 1997). Pyrin mutations are the genetic basis of this disease, which is clinically characterized by self-limited episodes of fever and inflammation and can trigger amyloidosis as a long-term consequence, leading to organ damage. Expression of pyrin has been described in monocytes, granulocytes, dendritic cells, and synovial fibroblasts [
3]. The NH2-terminal pyrin domain (PYD) can interact with apoptosis associated speck-like protein (ASC), a protein complex involved in IL-1β activation after intracellular sensing of pathogens and danger signals, leading to the formation of a distinct pyrin inflammasome [
4,
5]. Pyrin functions here as a cytosolic pattern recognition receptor and triggers the formation of a capase-1 inflammasome in response to bacterial toxins [
6].
The phagocyte-specific protein S100A12, which belongs to the group of damage associated molecular patterns (DAMP), can be detected at massively elevated levels in the serum of FMF patients [
7]. This molecule exerts pro-inflammatory effects via interaction with pattern recognition receptors (PRRs), specifically Toll-like receptor 4 (TLR-4) [
8], at concentrations that are found in FMF patients in vivo during active disease. Co-localization with the cytoskeleton and a Golgi-independent but tubulin-dependent release has been described for phagocyte-specific S100 proteins [
9,
10]. This so-called alternative secretory pathway is also involved in IL-1α/β and IL-18 secretion; however, the exact mechanisms of cellular release of S100A12 are still elusive.
Recently, we have shown that FMF patients demonstrate a gene-dose effect of the pyrin mutations based on S100A12 and IL-18 serum levels [
11]. Furthermore, ex vivo cultured granulocytes from FMF patients exhibit a unique phenotype with spontaneous release of high levels of IL-18, S100A12, MPO (myeloperoxidase), caspase-1, and proteinase [
12], as well as activation quantified by spontaneous shedding of CD62L from the cell surface. Neutrophil activation appears to be independent of IL-1 activation and exhibits a gene-dose effect that may be responsible for the genotype-dependent phenotypes in FMF [
12].
In this context, it is noteworthy that FMF patients with inactive disease—i.e., without clinical signs of disease and without elevated classical inflammatory markers (CRP and SAA)—have consistently elevated S100A12 levels [
11]. Even heterozygous MEFV mutation carriers appear to have elevated S100A12 levels compared with healthy controls [
13]. Whether this biochemical pattern is crucial for FMF pathogenesis, especially in the long term with respect to developing amyloidosis, is as yet unknown and should be monitored in these patients. We therefore investigated the mechanisms of S100A12 release from granulocytes of FMF patients with inactive disease.
Patients and methods
Patients and patient material
Patients presenting to our outpatient clinic (FMF patients and healthy controls, patient details see Table
1) routinely had clinical disease activity and parameters of inflammation (S100A8/A9, CRP, SAA, ESR, leukocytes, neutrophils) measured and retrospectively analyzed from patient charts (ethical approval by University Muenster, Ref.: 2022-703-f-S). Inflammatory diseases were excluded in the group of healthy controls. Median age of these controls was 8.8 years (4.5–16.1). Biosamples were prospectively collected from patients presenting to our outpatient clinic (FMF patients and chronic granulomatous disease (CGD) patients, patient details see Table
2) and processed as described below (ethical approval by University Muenster, Ref.: 2015-670-f-S). FMF patients that donated blood for the experiments involving granulocytes all were homozygous (see Table
2). Additionally FMF patients enrolled in the German autoinflammatory registry AID-Net were analyzed for clinical activity and S100 protein levels (ethical approval by University Muenster, Ref.: 2009-031-f-S) [
11,
14].
Table 1
Patient demographics of the FMF cohorts and healthy controls
Gender (female/male) | 9/9 | 10/9 | 21/18 | 8/11 |
Mutations | M694V/M694V | n/a | M694V/M694V | M694V |
Age in years at diagnosis (median, 25th–75th percentile) | 4.9 (3.4–10.9) | n/a | 5.3 (3.2–7.6) | 5.6 (3.6–12.0) |
Medication | Colchicine | n/a | Colchicine | Colchicine |
CRP (mg/dl) | <0.5 | <0.5 | <0.5 | <0.5 |
SAA (mg/l) | 9.9 (6.5–12.2) 8 from 18 neg. | 5.7 (5.2–7.6) 16 from 19 neg. | 3 (2–5) | 2 (1–4) |
ESR (mm/h) | 9 (6–14) | 5 (3–6) | 12 (8–15) | 8 (5–13) |
Leukocytes (per nl) | 7.7 (5.8–8.5) | 6.5 (5.5–7.7) | 7.2 (6.2–8.65) | 6.850 (5.1–9.6) |
S10012 (ng/ml) | n/a | n/a | 170 (80–280) | 81 (43–101) |
S100A8/A9 (ng/ml) | 15,730 (8975–18,730) | 2040 (1400–3820) | 11,225 (3896–28,627) | 3938 (1650–6703) |
Table 2
Patient demographics of the FMF cohort, CGD patients, and healthy controls (HC) serving for granulocyte isolation and analyses
Gender (female/male) | 8/7 | 2/3 | 4/6 |
Mutations | 11× M694V/M694V 2× M694I/M694I 1× V726A/M680I 1× M680I/M694V | 3× CYBB heterozygous, 2× NCF4 homozygous | n/a |
Age in years | 11 (6–13) | 17.5 (7.5–18) | 39 (29.3–48) |
Medication | 13× colchicine, 2× colchicine + canakinumab | Antibiotic and antimycotic prophylaxis | n/a |
CRP (mg/dl) | 0.2 (0.2–1) | <0.5 | n/a |
SAA (mg/l) | 4.2 (2–38.7) | n/a | n/a |
ESR (mm/h) | 7 (5–15) | n/a | n/a |
Leukocytes (per nl) | 6.09 (5.38–8.34) | 6.91 | n/a |
S10012 (ng/ml) | 56 (23–1108) | 56 (32–71) | n/a |
S100A8/A9 (ng/ml) | 3380 (2320–43,480) | 1540 | n/a |
Isolation of primary neutrophils and subsequent stimulation and inhibition
Neutrophils were isolated from whole blood (EDTA) using MACSxpress Whole Blood Neutrophil Isolation Kit (Miltenyi Biotec, Bergisch-Gladbach, Germany) according to the manufacturer’s instructions. Purity of neutrophils was controlled by FACS staining of CD66 on isolated cells and was routinely 95% or above.
Neutrophils were either left untreated or were stimulated with PMA (100 nM, Sigma-Aldrich, Taufkirchen, Germany), with MSU crystals (200 μg/ml, InvivoGen, San Diego, CA, USA), or with
Clostridium difficile toxin A (TcdA, 0.5 μg/ml, Sigma-Aldrich, Taufkirchen, Germany) for the indicated times and subsequently analyzed. When inhibitors were used, neutrophils were pre-incubated for 1 h before introducing neutrophils into subsequent assays or before further stimulation. ROS inhibitor DPI (diphenyleneiodonium chloride, used at 10 μM) and inhibitor of gasdermin D pore formation C23 (disulfiram—BMS-986165, used at 30 μM) were purchased from Selleckchem.com via Biozol, München, Germany. Bafilomycin A1 (1 μM) was from InvivoGen, Toulouse, France. IL-1β (5 ng/ml) was purchased from PeproTech, Hamburg, Germany. Anti-IL-1β antibody canakinumab (10 μg/ml) was from Novartis and IL-1RA anakinra (IL-1 receptor antagonist, 200 ng/ml) was from Sobi, Germany. Anti-CD66b antibody (clone g10F5) was from BioLegend, San Diego, USA. For immunoblots, anti-SQSTM1 1:2000 (Enzo, PW9860), rabbit polyclonal anti-LC3 1:2000 (GeneTex, GTX82986), anti-β-actin, and anti-GAPDH from Cell Signaling Technology, Leiden, Netherlands, were used. Immunoblots were performed as described in Koenig et al. [
15].
Cell death assay (Sytox™Green)
We performed cell death assays using SytoxGreen, a fluorescent dye that is naturally intercalating DNA, but only accessing DNA when the nuclear membrane is disintegrated.
For quantification of cell death, either 1 × 106 cells/ml or 5 × 106 cells/ml were plated in a 96-well plate (2 × 105/well or 1 × 106/well, respectively). Cells were incubated in the presence of 1.5 μM Sytox Green (Sytox™Green, ThermoFisher Scientific, Dreieich, Germany) and subsequently stimulated with the indicated activators and inhibitors. Cells were incubated for 5 h and fluorescence at 523 nm was assessed every hour using a fluorescence reader (TECAN infinite M200 pro, Tecan, Crailsheim, Germany). Supernatants were stored for S100A12 and LDH measurement.
S100A8/A9 and S100A12 ELISA
S100A8/A9 levels were measured with Bühlmann MRP8/14 Calprotectin ELISA (Bühlmann Laboratories AG) according to manufacturer’s instructions. S100A12 in supernatants and serum was analyzed using an in-house ELISA (normal level < 150 ng/ml) as previously reported [
16].
LDH measurement
LDH measurement was done from supernatants of cells using the Pierce LDH Cytotoxicity Assay Kit (ThermoFisher Scientific, Dreieich, Germany) according to the manufacturer’s instructions.
ROS (reactive oxygen species) measurement in flow cytometer
Ten microliters of DHR (dihydrorhodamine 123, ThermoFisher Scientific, Dreieich, Germany) was added to the cells 30 min prior to the measurement. DHR is passively transported through membranes and upon oxidation, e.g., by ROS, it becomes fluorescent and can be measured by flow cytometry (488 nm).
Fluorescence microscopy
For SytoxGreen fluorescence, 0.5 × 106/ml neutrophils were seeded in Lab-Tek chambers and then were either left untreated or were stimulated as indicated. After 4 h, supernatant was removed and NETs were then stained with 2.5 μM SytoxGreen. Stained NETs were analyzed using a fluorescence microscope at a wavelength of 488 nm emission filter with magnification 100× and 200× (Zeiss AxioVert, Zeiss, Jena, Germany).
For detection of NE (neutrophilic elastase) and histone H1 within NETs, Lab-Tek chambers were fixed with 1% paraformaldehyde for 15 min, then washed with PBS and blocked with Fc-blocking reagent. Subsequently NETs were incubated with either isotype control antibody (AF488- and AF647-conjugated) or anti-NE antibody (1 μg AF488-conjugated; Biozol, Eching, Germany) for 30 min and 1 μg anti-histone H1-AF647 antibody was used (Biozol, Eching, Germany). After 2× washing, mounting medium was added to the chambers. Glass cover slips were used on top of the mounting medium. Fluorescence images were acquired using a fluorescence microscope from Zeiss, Jena, Germany (AxioVert with Apotome). Lenses with 100× and 400× magnification were used.
Statistical analysis
Data were analyzed with GraphPad Prism software (version 9.0 for Mac OS X, GraphPad Software, La Jolla, CA, USA), and tests applied as indicated in figure legends. Significance of differences in levels of SytoxGreen, LDH, and S100A12 content were analyzed by Brown-Forsythe and Welch’s ANOVA test followed by Dunn’s multi-comparison test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001, and p ≤ 0.05 were considered statistically significant.
The reference values for normal controls for the “Patient cohort Muenster” in Table
1 and for patients in Table
2 are for CRP < 0.5 mg/dl and for SAA < 6.4 mg/l.
Discussion
In our study, we describe for the first time that cell death mechanisms like NETosis and pyroptosis can intrinsically occur in neutrophils from inactive FMF patients. ROS-mediated cell death is the canonical way described in neutrophils to lead into netotic cell death [
21,
28]. Cell death pathway as detected by SytoxGreen assay and LDH strongly depended on ROS in neutrophils from inactive FMF patients. Both, SytoxGreen and LDH, could be inhibited completely using DPI, an inhibitor of intracellular NADPH oxidase activity [
22]. In addition, cell death also depended on pore forming protein GSDMD that originally was described to be triggered by NLRP3 inflammasome to drive pyroptosis and to be independent of NETosis [
29]. Recently, it became more and more evident that GSDMD also plays a role in other cell death pathways than pyroptosis, namely, NETosis and even apoptosis [
20,
25].
Apparently, neutrophils from inactive FMF patients are readily activated by intrinsic pyrin activity that presumably lowers activation threshold for subsequent mechanisms like cell death pathways. In monocytes from FMF patients, it was demonstrated that stimulation via pyrin leads to enhanced inflammation and cytokine responses compared to monocytes from HC. However, when other inflammasomes were addressed, e.g., NLRP3 and NLRC4, the inflammatory response of the monocytes did not differ. The authors conclude that FMF-associated MEFV mutations decrease the activation threshold of the pyrin inflammasome without altering other inflammasomes [
30]. So, it is conceivable that pre-activated pyrin in neutrophils from FMF patients could accumulate pyrin-inflammasome-dependent signals, e.g., with rising cell-to-cell contact that leads into overactivation and cell death. It is, however, not known to date whether these mediators are surface molecules that profit from cell-to-cell contact in rising cell densities or soluble factors that profit from short distances. Cell density seems to play a role for initiation of NETosis as demonstrated in gout-induced aggregated NETs [
19]. Recently, we described neutrophils from also clinically inactive FMF patients already displaying increased activation measured by ex vivo shedding of L-selectin without any additional external stimulation, and this was following a gene-dose effect [
12]. Additionally we already described the release of S100A12, IL-18, caspase-1, and MPO from these granulocytes without further stimulus and we can now add to this the probable cell death mechanism responsible for at least S100A12 release.
Apostolidou et al. [
17] and Skendros et al. [
18] described netotic cell death in neutrophils from FMF patients in disease attack that was dependent on autophagy and led to release and display of IL-1β to the extracellular space and the released IL-1β was able to drive inflammation. By using inhibitor of autophagy in our SytoxGreen assay, we could not detect any contribution of autophagy to the ROS-induced cell death.
Kanneganti et al. [
26] demonstrated in a mouse model of FMF that gasdermin D was required for autoinflammatory pathology. Here, one major player was IL-1β that was shown to be secreted through GSDMD pores and contributed to pathology. The whole process was driven by infection of mice with
Clostridium difficile (
C. difficile) to induce NLRP3 inflammasome leading to pyroptosis and IL-1β release through pore forming GSDMD. In their study, they utilized macrophages as major source of Il-1β. Also Evavold et al. [
31] demonstrated that GSDMD regulates IL-1β secretion in living macrophages. When we used neutrophils from inactive FMF patients, we could not detect a direct influence of IL-1β leading to auto-driven cell death, as shown by the use of anti-IL-1 antibody (canakinumab) and IL-1RA (anakinra) in our model. Neither canakinumab nor anakinra had an influence on cell death in FMF neutrophils (Sytox assay, LDH). We also used recombinant IL-1β to induce cell death and S100A12 release from neutrophils but were unable to detect cell death and protein release (Supp. Fig.
2).
However, release of DAMP S100A12 was significantly blocked when inhibitor of GSDMD cleavage and pore forming, C23 (disulfiram), was used in cell culture. This directly points to the importance of GSDMD activation and cleavage as a major mechanism for “alternative” release of cytokines and proteins into the extracellular space. We recently [
11] showed that neutrophils from FMF patients already released S100A12 (and also IL-18 and Casp-1) spontaneously, presumably through the GSDMD cleavage and pore formation. In addition, we could already demonstrate that S100A12 acts as a pro-inflammatory DAMP on human cells [
32]. Even more, Jorch et al. [
33] demonstrated the importance of S100A8/9 release through GSDMD pores for FMF pathology in a mouse model of FMF and in cell lines. We expand this knowledge by adding S1000A12 to the list of GSDMD pore-regulated proteins in neutrophils. Additionally, S100A12 release from neutrophils is also dependent on ROS, as neutrophils from CGD patients, which do not build active ROS intrinsically, do not release S100A12, not even when stimulated with PMA (Fig.
4). Semino et al. [
27] demonstrated that blocking ROS production also prevents GSDMD cleavage claiming the importance of oxidative stress in GSDMD-mediated secretion. We confirm here this finding by examining S100A12 release from human neutrophils. When we stimulate neutrophils from HC with PMA, S100A12 is released. This can be blocked by both, ROS inhibitor and by GSDMD inhibitor, suggesting that ROS can also drive GSDMD cleavage. Sollberger [
25] already found GSDMD to play a vital role in the generation of extracellular traps. Especially, when we induce NETosis by using PMA, the role of GSDMD becomes evident, because in this case GSDM inhibitor not only blocked S100A12 release but also cell death.
We used TCDA to trigger cell death and S100A12 release to demonstrate that pyrin inflammasome is involved in cell death and ROS production. We could show that TCDA induced stronger S100A12 release and cell death when neutrophils were from FMF patients, although inactive clinically, they apparently present with pre-activated pyrin inflammasome. However, TCDA was shown to also activate NLRP3 inflammasome [
34,
35]. Thus, there is additional experimentation necessary to detect contribution of either pyrin and/or NLRP3 inflammasome to cell death and GSDMD activation, and subsequent protein release. Limitations of our study are certainly limited number of patients and missing age matched controls from pediatric individuals.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.