Article Text

Download PDFPDF

Citrullination of autoantigens implicates NETosis in the induction of autoimmunity
  1. Nishant Dwivedi1,
  2. Marko Radic2
  1. 1Program in Cellular and Molecular Medicine, Boston Children's Hospital and Harvard Medical School, Boston, Massachusetts, USA
  2. 2Department of Microbiology, Immunology and Biochemistry, University of Tennessee Health Sciences Center, Memphis, Tennessee, USA
  1. Correspondence to Professor Marko Radic, Department of Microbiology, Immunology and Biochemistry, UT Health Science Center, 858 Madison Avenue, rm 211, Memphis, TN 38163, USA; mradic{at}


Tolerance blocks the expression of autoantibodies, whereas autoimmunity promotes it. How tolerance breaks and autoantibody production begins thus are crucial questions for understanding and treatment of autoimmune diseases. Evidence implicates cell death and autoantigen modifications in the initiation of autoimmune reactions. One form of neutrophil cell death called NETosis deserves attention because it requires the post-translational modification of histones and results in the extracellular release of chromatin. NETosis received its name from NET, the acronym given to Neutrophil Extracellular Trap. The extracellular chromatin incorporates histones in which arginines have been converted to citrullines by peptidylarginine deiminase IV (PAD4). The deiminated chromatin may function to capture or ‘trap’ bacterial pathogens, thus generating an extracellular complex of deiminated histones and bacterial cell adjuvants. The complex of bacterial antigens and deiminated chromatin may be internalised by host phagocytes during acute inflammatory conditions, as arise during bacterial infections or chronic autoinflammatory disorders. The uptake and processing of deiminated chromatin together with bacterial adjuvants by phagocytes may induce the presentation of modified histone epitopes and co-stimulation, thus yielding a powerful stimulus to break tolerance. Autoantibodies to deiminated histones are prevalent in Felty's syndrome patients and are present in systemic lupus erythematosus (SLE) and patients with rheumatoid arthritis (RA). These observations clearly implicate histone deimination as an epigenetic mark that can act as an autoantibody stimulant.

  • Rheumatoid Arthritis
  • Systemic Lupus Erythematosus
  • Autoimmunity
  • Ant-CCP

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Despite decades of intense research, we do not know the stimuli that trigger the production of autoantibodies in systemic autoimmune diseases.1 The autoantibodies themselves have been studied and found to be useful as diagnostics for different autoimmune diseases. For example, autoantibodies reacting against nuclear antigens, most notably DNA and histones, are a distinguishing feature of systemic lupus erythematosus (SLE), whereas autoantibodies to a citrullinated cyclic peptide are a reliable diagnostic for rheumatoid arthritis (RA). A plausible prediction is that the careful analysis of autoantibodies will lead to a better understanding of their origins.

The structure of autoantibodies argues that they are selected for improved binding to autoantigens. The genetic mechanisms that give rise to autoantibodies argue strongly in favour of this conclusion. Autoantibody sequences indicate that even though diverse V genes code for anti-DNA and rheumatoid factor autoantibodies in murine models of SLE and RA, junctional diversity and somatic mutations make essential contributions to binding.2–4 In consequence, only few B cell clones secure conditions for expansion. The expanded autoreactive B cell clones participate in antigen presentation, given that T cell help promotes antibody H chain isotype switching, and selects for mutations that improve antibody binding to autoantigens. As a result, autoantibodies acquire all the molecular hallmarks of affinity maturation, just as antibodies to foreign antigens do.5 Therefore, the identification of molecular features of autoantigens that drive systemic autoimmune diseases holds the key for understanding the pathways that stimulate the production of autoantibodies.

Cell death and autoimmunity

Every tissue in the body, including the central nervous system, generates new cells to replace cells that die, and most physiological cell death occurs by apoptosis.6 Apoptosis is an energy-demanding process in which specific enzymes are sequentially activated to cleave hundreds of defined substrates. As a result, large-scale morphologic transitions are induced that disrupt the structure of cell organelles and cell membranes. The condensation of nuclear chromatin and the fragmentation of the nucleus represent some of the most dramatic events of apoptosis. The nuclear fragments separate and punch through the plasma membrane, thus forming protruding blebs at the cell surface.7 Previous work has demonstrated that the exposure of nuclear fragments at the cell surface in apoptosis exposes histones and DNA to the extracellular milieu. The surface exposure of nuclear autoantigens is instrumental in providing access of autoantigens to B cells.8

Genetic evidence supports the role of apoptosis in the induction of autoimmunity. Various defects of the receptors that function in the clearance of apoptotic cells, or of serum proteins that facilitate the binding and uptake of apoptotic cell remnants, result in an increased risk for developing autoimmunity.9 It stands to reason that normal clearance pathways carried out by macrophage and mediated by innate receptors for apoptotic cells maintain tissue homeostasis in a non-inflammatory manner.

By contrast, the persistent exposure of nuclear autoantigens on apoptotic cells may stimulate the adaptive immune response to mobilise and clear the excess of apoptotic debris.10 It is interesting that apoptotic cell clearance depends on innate receptors that also function in the recognition of microbial molecular patterns. For example, apoptotic cell clearance is accomplished through binding of the C1q complement protein,11 the serum amyloid P component,12 C-reactive protein,13 milk fat globule protein E814 and pentraxins15 to the surface of apoptotic cells, and the recognition of binding by the Gas6-dependent tyrosine kinase receptors Mer, Axl and Tyro 3.16 ,17 The proper physiological clearance of apoptotic cells is essential for maintaining immunological tolerance. If, due to defects in the normal clearance pathways, B cells are recruited to assist in the clearance of apoptotic cells, the nuclear autoantigens may stimulate innate toll-like receptors expressed in endosomal compartments of B cells.18 ,19 These observations clearly support the view that inadequate clearance of apoptotic debris stimulates the production of antibodies to nuclear antigens. By extension, it could be argued that other forms of cell death also may supply nuclear autoantigens for an adaptive immune response to self. Different cell death pathways are induced by different stimuli and result in the precise execution of evolutionarily conserved steps, including the post-translational modification (PTM) of proteins.

Post-translational modifications of autoantigens

Autoantibodies frequently target specific modifications in self-antigens,20 ,21 including PTMs that are introduced during apoptotic cell death.22 ,23 Thus, PTMs may provide a mechanism for converting innocuous self-components into targets of an autoimmune response. PTMs generate ‘neo- epitopes’ (new binding determinants for B or T cell antigen receptors) and, therefore, alter antigen processing and presentation.24 That PTMs play an important role in the induction of autoimmunity is consistent with some of the possible PTMs in histones. Histones, a defining antigen in systemic autoimmune disorders, exhibit PTMs that autoantibodies target.25 Thus, specific histone modifications that arise in apoptosis are recognised by SLE sera.26 A different example is the conversion of arginines in histones to citrullines. The resulting citrullines become nestled in a thicket of other histone PTMs (figure 1). Citrullination, also called deimination, is the work of peptidylarginine deiminases (PADs). PADs are Ca++-dependent enzymes that are conserved in all vertebrates.27 ,28 The conserved nature of PADs implies an important physiological role. Indeed, citrullination regulates zygotic development, the myelination of neurons, and properties of the epidermis.

Figure 1

Post-translational modifications in histones. Sequences of the amino termini of the four core histones are shown to indicate the location of various amino acid modifications. Lysine residues can be acetylated or methylated and there is possibility of two acetyl groups or up to three methyl groups per residue. The trimethylated lysine 27 in H3 is a post-translational modification (PTM) that arises in apoptosis, and that is recognised by lupus autoantibodies.22 Additionally, lysine 15 of H2A can also carry a ubiquitin modification, a mark for DNA damage.118 Phosphorylation of serine and threonine residues is also observed and includes phosphorylated serine 15 in H2B which is another PTM that is induced during apoptosis but does not appear in NETosis.47 The most prominent marks of NETosis are multiple citrullines that replace arginines in H2A, H4 and, most prominently, in H3. The PTM are indicated by symbols that are explained below the diagram. Access the article online to view this figure in colour.

Notably, although the five PAD isoforms exhibit different tissue distribution, many of their substrates are prominent autoantigens. Table 1 lists autoantigens that are deiminated by PADs and that reside in different tissues or cell compartments. Many of the substrates are components of the extracellular matrix, but others are cytoplasmic or nuclear, depending on their function. Expression of PAD isoforms is similarly heterogeneous.26 Thus, PAD1 is expressed in epidermis where it deiminates keratin and the keratin-associated protein filagrin.29–31 Autoantibodies to citrullinated keratin and citrullinated filaggrin exhibit high specificity for RA.32 In general, anticitrullinated protein antibodies (ACPA) have over 90% specificity and 75% sensitivity for RA, although the percentage of patient sera that react with any particular citrullinated autoantigen ranges widely from over 70% to less than 20%. Therefore, depending on their specific target, ACPA can be used as reliable diagnostic markers for arthritis.33

Table 1

List of deiminated autoantigens

Myelin basic protein (MBP), vimentin and glial fibrillary acidic protein (GFAP) are substrates of PAD2, and autoantibodies to MBP arise in multiple sclerosis (MS) patients, and in experimental autoimmune encephalomyelitis, a mouse model of MS.34 ,35 By contrast, autoantibodies to citrullinated vimentin arise in RA36 and autoantibodies to keratin and trichohyalin, a structural protein of sheath cells in hair follicles, characterise alopecia areata, a disorder exhibiting sudden hair loss.37

PAD4 is the only PAD that has a nuclear localisation signal, and deiminates histones H2A, H3 and H4.38 ,39 A link exists between PAD4 haplotypes and risk of RA. Alleles linked to an increased susceptibility to RA encode a more stable PAD4 mRNA than alleles that are neutral for RA.40 The increased mRNA stability may lead to increased expression of PAD4 in individuals bearing the susceptible haplotype. An association between deiminated proteins and autoimmune diseases strongly suggests that deimination generates novel epitopes that are able to break tolerance and initiate autoimmune responses.

PAD4 deimination of nuclear autoantigens

PAD4 is most abundantly expressed in granulocytes,41 monocytes42 and mast cells.43 PAD4 may associate with cytoplasmic granules, but it also enters the nucleus where it may participate in gene regulation by deiminating histones.38 Deimination transforms a positively charged amino acid (arginine) into a neutral amino acid (citrulline), resulting in the net loss of one positive charge per modified residue.44 Therefore, deimination changes not only the primary sequence of a protein but also its net charge. During the reaction, the guanidino group of arginine is attacked by a Cys residue of PAD4, resulting in the formation of a tetrahedral adduct.45 A nucleophilic attack by a water molecule then cleaves the adduct to form the keto group of citrulline, while releasing ammonia and regenerating the Cys residue of the enzyme.44 ,46 The net result is the conversion of arginine to citrulline with a mass shift of +1 Da.

The activation of PAD4 in neutrophils is exquisitely sensitive to inflammatory stimuli, such as formylated Met-Leu-Phe peptide (f-MLP), LPS, lipoteichoic acid (LTA) or TNF.47 By contrast, the induction of neutrophil apoptosis blocks PAD4 activation.47 This discovery came as a surprise because numerous reviews and research reports prior to 200847 had proposed that PAD4 is active during neutrophil apoptosis.48 The misconception may, in part, be accounted for by the fact that neutrophils undergo two different forms of cell death. Whereas senescent neutrophils that were not needed to fight an infection die by apoptosis,49 ,50 neutrophils faced with micro-organisms in the course of an infection perform an elaborate form of cell suicide. Those neutrophils die by ejecting nuclear chromatin in order to capture, or ‘trap’, microbial pathogens.51 PAD4 activity precedes and facilitates the release of chromatin from neutrophils.52 In fact, PAD4 is essential for chromatin release and, in the absence of PAD4, the innate response to bacterial infections is impaired.53 Clearly, neutrophil extracellular traps (NET) deserve wider experimental scrutiny because, in infections,54 or in response to adjuvants,55 NETs interpose citrullinated histones with strong immune stimuli. This feature of NETs could be instrumental in breaking tolerance and inducing an autoimmune response to chromatin antigens. Nonetheless, work in NADPH oxidase-deficient MRL.Fas(lpr) mice suggests that, in that model of lupus, NETs are not required for autoimmunity.55a

Neutrophil extracelluar traps and autoimmunity

Remarkable progress has been achieved in understanding pathways that regulate NETosis, the unique form of cell death that leads to the release of chromatin NETs.56 ,57 Particular attention has centred on signalling pathways that carry out the massive cellular reorganisation that is required for NET release. Signals from the cell surface, originating with innate pattern recognition receptors58 and integrins,59 are transmitted via protein kinase C (PKC),60 ,61 and result in the generation of reactive oxygen species. Signals activate PAD4 in the nucleus and at cytoplasmic hotspots, leading to chromatin decondensation, nuclear envelope breakdown, and mixing of chromatin with cytoplasmic and granule contents.47 ,62

Figure 2 outlines proposed steps in the transmission of signals that regulate NETosis. Microbes, including bacteria, fungi or viruses interact with cell surface receptors, whose downstream signals lead to a transient increase of cytoplasmic calcium.63 This, in turn, activates PKC64 and extracellular signal-activated kinase 1/2 (ERK1/2).60 ERK1/2 leads to the activation of the p47 subunit of NADP oxidase,65 followed by the assembly of an active oxidase complex. The result is reactive oxygen production and perchlorate formation.66 Pathogen-associated molecules, such as lipopolysaccharide (LPS), may act via toll-like receptors and activate the nuclear factor kappa B (NFκB) pathways.67 Tumour progression locus 2 (TLP2) is a candidate kinase that could promote activation of ERK1/2 in this scenario.68 Converging pro-NETosis signals favour neutrophil granule and nuclear envelope disassembly, thus allowing association between granule components such as myeloperoxidase (MPO) and neutrophil elastase (NE) with nuclear chromatin.69 These enzymes, among other NET components, associate with the unravelling chromatin and acquire additional extracellular functions. PAD4, itself regulated by opposing PKC subunits,61 plays an essential role in the decondensation of chromatin and the subsequent release of NETs.53

Figure 2

Proposed steps in the transmission of signals from the cell surface to the nucleus in NETosis. Events that lead to the release of NET chromatin from the cell, likely induced by bacteria, fungi or viruses, may initiate with interactions with cell surface receptors that activate phospholipase C γ (PLCγ) and transiently elevate the cytoplasmic calcium levels. This, in turn, activates protein kinase C (PKC) isoforms that regulate peptidylarginine deiminase 4 (PAD4) and extracellular signal-activated kinase 1/2 (ERK1/2). ERK1/2 may phosphorylate and activate the p47 subunit of NADP oxidase, thus favouring the assembly of the active multisubunit complex whose reactive oxygen production and associated perchlorate formation exerts extracellular and intracellular functions. Soluble pathogen-associated molecules such as lipopolysaccharide (LPS) bind toll-like receptors (TLR) and activate the MyD88 signalling pathway. The tumour progression locus 2 (TLP2) may connect the TLR-NFkappaB pathway to the activation of ERK1/2. Neutrophil granules disintegrate and release myeloperoxidase (MPO) and neutrophil elastase (NE) into the cytoplasm where they associate with the unravelling chromatin. PAD4 appears to play an essential role in the decondensation of chromatin and the subsequent release of NETs. Access the article online to view this figure in colour.

Alternative forms of NETs may consist of mitochondrial DNA70 or be released from a novel type of vesicles containing nuclear chromatin.71 Notable differences may exist between these NET variations. Differences may include different speeds of NET deployment, directionality of NET release, and active neutrophil chemotaxis during—and after—NET expulsion. In the ‘classic’ form of NETosis, decondensed nuclear chromatin associates with MPO and elastase,69 along with defensin peptides, such as LL37,72 before it is released from the cell. The released chromatin brings benefits in responses to bacteria,51 ,73 ,74 fungi75 and viruses.76 Reports show that in vitro NETs are capable of attaching and damaging microbes, and studies in mouse models show that the deployment of NETs is beneficial to the organism.53 ,76 However, uncontrolled NET production that can also be induced by sterile inflammation in chronic inflammatory disorders has quite serious consequences.

That NETs may stimulate autoimmunity follows from the fact that autoantibodies target various NET components. Autoantibodies from patients with systemic autoimmunity recognise innate immune defence proteins, including the major NET components: elastase, cathepsin G and proteinase 3,77–79 as well as bactericidal peptides that are stored in granules and co-purify with NETs, such as neutrophil defensins and the LL37 peptide.80 ,81 The autoantibody targets that arise in NETosis were summarised by Darrah and Andrade.82 Khandpur and colleagues convincingly demonstrated that NETs are a primary source of deiminated autoantigens, and that RA neutrophils are more susceptible than osteoarthritis neutrophils to NETosis.83 NETs stimulate inflammatory responses, including the increased expression of adhesion molecules, cytokines and chemokines.83 Separate studies indicate an increased abundance of activated neutrophils and elevated concentrations of neutrophil granule components in the sera of patients with more advanced disease.77 ,84 Additionally, the severity of lupus also correlates with the titres of anti-NET autoantibodies.85 Thus, lupus autoantibodies bind NET-associated proteins, in addition to histones and DNA, the NETs’ structural scaffold.

Other studies directly point to neutrophil activation as a key factor in the stimulation of autoantibodies. As discovered in studies involving juvenile lupus participants, SLE neutrophils show an exacerbated response to lupus autoantibody complexes.86 In turn, NETs may persist for a longer time in autoimmunity, as SLE patients’ anti-DNA antibodies protect NETs from degradation by nucleases.85 Persistent NETs may aggravate inflammation by stimulating plasmacytoid dendritic cells (pDC) to release cytokines such as interferon type I.81 ,86 Recent data implicate the binding of anti-PAD4 autoantibodies with the enzymatic activation of the deiminase,87 thus arguing for a likely extracellular role of PAD4. In the extracellular milieu, PAD4 activation may yield citrullinated forms of fibrinogen, fibronectin, collagen and other matrix proteins that have been linked with manifestations of connective tissue inflammation (table 1). Deimination of myelin basic protein, most likely by PAD2 or PAD4, may participate in the pathogenesis of neuro-degenerative disorders.88

Studies that report a benefit from inhibitors of PAD4 in mouse models of systemic autoimmunity deserve special attention. First, in the collagen-induced arthritis model, inhibition of PAD4 by Cl-amidine improved quantifiable disease manifestations.89 Subsequently, PAD4 inhibition by Cl-amidine was used to treat NZM 2328 mice, a reliable mouse model of lupus.90 The drug was administered to mice over a period of 3 months without adverse effects, and the treatment inhibited NET formation and specifically decreased anti-DNA titres without measurably affecting serum antibody isotypes. Additionally, Cl-amidine reduced MPO and immune complex deposition in the kidneys of NZM mice and improved endothelial cell differentiation and vasorelaxation, while reducing the risk of arterial thrombosis.90 Similarly, amelioration of neurodegenerative symptoms was reported in a mouse model of MS, following administration of a novel deiminase inhibitor.91

Together, these findings suggest that, in autoimmune disorders, NETosis may be more easily triggered, NETs may induce the production of anti-NET antibodies, and the NET antigens may persist in tissues because they are protected by autoantibodies. Thus, NET antigens and anti-NET autoantibodies may be linked in a pernicious pathogenic cycle.

Induction of histone deimination by bacterial infections

Diverse bacterial and fungal infections trigger NET release.92 NETosis is induced by neutrophil encounter with Staphylococcus aureus51 or Mycobacterium tuberculosis93 and in response to exposure to Candida albicans75 ,94 and Aspergillus fumigatus.95 ,96 NETs function in trapping and damaging micro-organisms, and this explains how extracellular nucleases produced by bacteria function as virulence factors.97 The release of NETs is linked to various clinical complications of infectious diseases, including bacterial sepsis98 ,99 and the increased risk of thrombosis.100 Interestingly, more serious infections and blood clots are known as complications in a number of autoimmune disorders.101 These complications could be linked, in part, to the inopportune release of NETs from autoimmune patient neutrophils.

To explore the neutrophil response to infections, we used an animal model. The mouse peritoneum is a suitable environment for S aureus, as the bacterium rapidly divides in the murine peritoneal cavity. To test the induction of histone deimination in vivo, we injected 109 colony-forming units of S aureus in PBS or PBS alone into mice. After 1 h of incubation, the peritoneal fluid was collected and filtered through a 0.22 µm filter to remove intact bacteria and mouse cells. The filtrate contained NET fragments generated in vivo. Figure 3A shows that histone deimination and NET release had occurred in mice injected with the bacterial pathogen but not in those receiving vehicle alone. The rapid release of histones containing citrullines may occur via the recently described rapid release of vessiculated chromatin from neutrophils.71 Our results confirm that a wide range of infections elicit the production of deiminated histones and their release from the cell in the form of chromatin traps.

Figure 3

Neutrophil activation in response to infection and crystals. (A) Deiminated histones are present in peritoneal lavage fluids collected from Staphylococcus aureus injected mice. Peritoneum from mice injected with 109 CFU of S aureus was flushed and collected. After filtration through 0.22 um filter, the lavage fluid was run on a 12% SDS PAGE gel, blotted to a membrane and probed with rabbit antideiminated H3 antibody to detect the presence of free deiminated histones. (B) Quantitation of NETs induced by crystals. Human neutrophils (100 000/well) were distributed to a 96-well polypropylene plate and treated with monosodium urate (MSU) (200 ug/mL), Alum (150 µg/mL) and H2O2 (1 mM) for 2 h at 37°C. At the end of incubation sytox orange was added to a final concentration of 1.66 uM, and fluorescence was measured (excitation at 535 nm and emission at 595) on a fluorimeter. The experiment was done at least three times, and triplicate values from one experiment are plotted. Error bars indicate SD from mean values.

Sterile inflammation leads to histone deimination

A wide range of inflammatory stimuli induce histone deimination and NETosis,47 hence NETosis can proceed in the absence of an infection. To test additional sterile inflammatory conditions that lead to deimination of histones by PAD4, we activated PAD4 with compounds associated with sterile inflammation. Monosodium urate (MSU), a byproduct of nucleic acid metabolism, is associated with the chronic autoinflammatory disorder of gout.102 MSU elicits sterile inflammation by forming crystals in articular joints of individuals afflicted by gouty arthritis. The crystals activate the inflammasome and result in the release of interleukin 1β.103 Due to the persistent inflammation, bone damage, erosion and remodelling ensue.

Aluminium hydroxide, a substance that is widely used as an adjuvant in vaccines, provides an alternative stimulus that can induce NETosis and histone deimination in the absence of an infection.55 To test the production of NETs from neutrophils that were exposed to MSU or alum, we developed an in vitro assay in which released chromatin is quantified by using a cell-impermeable DNA dye, Sytox Orange, and we measured the chromatin in the supernatant of cells by fluorometry. As control, we used hydrogen peroxide, a strong inducer of NETs and histone deimination.47 Figure 3B illustrates that MSU and aluminium hydroxide were potent stimuli of NET release, suggesting that the pathogenic mechanism in gout involves the intra-articular activation of neutrophils, along with the release of NETs containing deiminated histones. It is therefore likely that sterile inflammatory disorders, such as gout, elicit the extracellular release of deiminated histones from neutrophils.104 ,105 Under these conditions, the epigenetic histone modification may break tolerance and lead to B cell autoreactivity.

Autoantibodies to deiminated histones

If NETs are the source of disease-associated autoantigens, then autoantibodies from patients with systemic autoimmunity should bind to NETs more avidly than control sera. In addition, NETs would be predicted to contain unique targets for autoantibodies that are absent from unstimulated neutrophils. Indeed, we found prominent NET autoreactivity (figure 4) in the sera from patients with SLE, RA and Felty's syndrome (FS). FS shares features of SLE and RA and is defined by the triad of severe arthritis, splenomegaly and neutropenia.106 All three disorders are characterised by the production of antihistone autoantibodies.107 ,108 FS arises in 1–3% of RA patients after an extended period of incompletely mitigated disease.109 ,110 FS patients experience severe destructive arthritis and develop extra-articular manifestations, including rheumatic nodules and vasculitis.111

Figure 4

Autoimmune sera react with neutrophil extracellular traps. Neutrophils from non-autoimmune human donors were purified by density gradient centrifugation and incubated in Hanks balanced salt solution (panels A and C), or stimulated with lipopolysaccharide (LPS) (panels B and D). Cells were incubated on microscope cover slips for 2 h and then processed for immunofluorescence confocal microscopy, as previously described.47 In panels A and B, the cells were incubated with a systemic lupus erythematosus (SLE) serum, whereas, the cells in panels C and D were incubated with Felty's syndrome (FS) serum. The serum reactivities with NETs and deiminated histones were characterised by Dwivedi et al.106 DNA is shown in blue, IgG binding in red, and the plasma membrane in green. Antibodies react preferentially with LPS-treated neutrophils that have released extracellular chromatin.

Experiments suggest that neutropenia in FS is due to autoimmune aetiology rather than a production defect. In mice infused with the sera or isolated immune complexes of FS patients, neutrophil counts drop precipitously, suggesting autoimmunity as the mechanism.112 Immune complexes and soluble antibodies from FS patients bind neutrophils and mediate their sequestration in the spleen.113 The spleen of FS patients exhibits an expansion of red pulp, sinus hyperplasia, macrophage infiltration and prominent germinal centres.114 These hallmarks of increased splenic activity indicate that clearance of activated neutrophils by the spleen may promote FS. Interruption of the self-sustaining autoreactivity by removal of the patient's spleen restores neutrophil counts, arguing that the spleen provides a sink for neutrophils in FS.114 Thus, it is reasonable to argue that autoantibodies against neutrophil antigens in FS drive neutropenia by splenic sequestration of neutrophils. To understand FS, it will be important to identify the neutrophil targets of autoantibodies.

There are three parameters that suggest a possible pathogenic mechanism. FS patients score positive for binding to the cyclic citrullinated peptide, a screening test that identifies 77% of FS patients.115 ,116 Additionally, a remarkably high percentage of FS patients have autoantibodies to histones, and reports indicate that over 80% of FS sera contain antihistone IgG.108 The sequestration of activated neutrophils in the spleen may, over time, induce autoantibody production. Because we hypothesised that stimulation of neutrophils in FS would lead to PAD4 activation and histone deimination, and because FS patients make antihistone antibodies, we decided to test FS patient sera for IgG to deiminated histones.106

To test for the presence of antibodies to deiminated histones, we reacted histones from calf thymus with recombinant PAD4 to completion. ELISA and immunoblotting with deiminated and non-deiminated histones revealed autoantibody preference for deiminated histones in a majority of FS patients and in a subset of SLE and RA patients.106 Deiminated H3 was the major target autoantigen, although binding to other histones, H4 and H1, was also observed. A control group of patients with antineutrophil cytoplasmic antibody-associated vasculitis-expressed antibodies that bound non-deiminated histones with preference over deiminated histones. This confirmed that the selection of autoantibody targets depends on the pathogenic process and may differ between autoimmune disorders. After our study was published, Pratesi et al117 identified citrullinated histone H4 as a specific autoantibody target in a majority of RA sera. In their study, deiminated histones were prepared from calcium ionophore-treated neutrophils, and reactivity was confirmed by immunoprecipitation and mass spectrometry.

Autoantibodies to deiminated histones, antigens that are produced during neutrophil activation, represent strong evidence that NETs contribute to autoimmunity in FS and SLE. Because neutrophils are involved in responses to infections, our findings also suggest a molecular link between chronic infections and autoimmunity. Neutrophil activation and NET formation are also induced by inflammatory cytokines, suggesting how one might arrive at a positive feedback loop that is the hallmark of autoimmune disorders.

Model of autoimmune pathogenesis in Felty's syndrome

Based on the biological properties of neutrophils, our data on the incidence of autoantibodies to deiminated histones, and the disease progression in FS, we propose the following model of FS pathogenesis (figure 5). The diagram connects the clinical observations that characterise FS, for example the pre-existing arthritis, neutropenia, splenomegaly and an increased incidence of infections, with the neutrophil response to infections and inflammation. In these circumstances, neutrophils induce histone deimination and expel chromatin from the cell. NETs containing deiminated histones, in complex with bacterial adjuvants, are the most likely antigenic trigger to account for the production of autoantibodies to the modified histones. These autoantibodies, similar to other antineutrophil cytoplasmic antibodies (ANCA), may further stimulate neutrophils, thus completing a cycle that is self-sustaining and drives depletion of mature neutrophils. What factors break tolerance to deiminated histones in FS remains to be determined. Yet, the increased severity of FS as compared to RA, with its rapidly progressing arthritis, extra-articular manifestations and life-threatening neutropenia, clearly indicate the need to identify the role of NETosis in the pathogenic mechanisms of FS.

Figure 5

A model of potential factors involved in autoimmune pathogenesis of Felty's syndrome. Factors that may contribute toward the development of anti-NET autoantibodies are discussed in the text, and hypothetical interactions are shown.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 55a.
  57. 56.
  58. 57.
  59. 58.
  60. 59.
  61. 60.
  62. 61.
  63. 62.
  64. 63.
  65. 64.
  66. 65.
  67. 66.
  68. 67.
  69. 68.
  70. 69.
  71. 70.
  72. 71.
  73. 72.
  74. 73.
  75. 74.
  76. 75.
  77. 76.
  78. 77.
  79. 78.
  80. 79.
  81. 80.
  82. 81.
  83. 82.
  84. 83.
  85. 84.
  86. 85.
  87. 86.
  88. 87.
  89. 88.
  90. 89.
  91. 90.
  92. 91.
  93. 92.
  94. 93.
  95. 94.
  96. 95.
  97. 96.
  98. 97.
  99. 98.
  100.  99.
  101. 100.
  102. 101.
  103. 102.
  104. 103.
  105. 104.
  106. 105.
  107. 106.
  108. 107.
  109. 108.
  110. 109.
  111. 110.
  112. 111.
  113. 112.
  114. 113.
  115. 114.
  116. 115.
  117. 116.
  118. 117.
  119. 118.
  120. 119.
  121. 120.
  122. 121.
  123. 122.
  124. 123.
  125. 124.
  126. 125.
  127. 126.
  128. 127.


  • Handling editor Tore K Kvien

  • Contributors Both authors contributed to writing and approving the submitted manuscript.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.