IFN-γ licenses CD11b+ cells to induce progression of systemic lupus erythematosus

Journal of Autoimmunity 62 (2015) 11e21

Contents lists available at ScienceDirect

Journal of Autoimmunity journal homepage: www.elsevier.com/locate/jautimm

IFN-g licenses CD11bþ cells to induce progression of systemic lupus erythematosus € rg b, Vishal Khairnar a, Namir Shaabani a, b, 1, Nadine Honke a, b, 1, Sebastian Dolff c, Boris Go Katja Merches a, Vikas Duhan a, Sabine Metzger d, Mike Recher e, Carmen Barthuber f, €ussinger b, Oliver Witzke c, Cornelia Hardt a, Peter Proksch g, Dieter Ha b, h, 1 a, b, *, 1 , Karl S. Lang Philipp A. Lang a

Institute of Immunology, Medical Faculty, University Duisburg-Essen, Essen, Germany Department of Gastroenterology, Hepatology and Infectious Diseases, Heinrich-Heine-University Düsseldorf, Düsseldorf, Germany c Department for Nephrology, Medical Faculty, University Duisburg-Essen, Essen, Germany d Metabolomics Facility, Cologne Biocenter, University Cologne, Cologne, Germany e Clinic for Primary Immunodeficiency, Medical Outpatient Unit and Immunodeficiency Lab, Department of Biomedicine, University Hospital Basel, Basel, Switzerland f Department of Laboratory Medicine, Heinrich-Heine-University Düsseldorf, Düsseldorf, Germany g Institute of Pharmaceutical Biology and Biotechnology, Heinrich-Heine-University Düsseldorf, Düsseldorf, Germany h Department of Molecular Medicine II, Heinrich-Heine-University Düssledorf, Düsseldorf, Germany b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 February 2015 Received in revised form 11 May 2015 Accepted 18 May 2015 Available online 19 June 2015

Autoantibodies are a hallmark of autoimmune diseases, such as rheumatoid arthritis, autoimmune hepatitis, and systemic lupus erythematosus (SLE). High titers of anti-nuclear antibodies are used as surrogate marker for SLE, however their contribution to pathogenesis remains unclear. Using murine model of SLE and human samples, we studied the effect of immune stimulation on relapsing of SLE. Although autoantibodies bound to target cells in vivo, only additional activation of CD8þ T cells converted this silent autoimmunity into overt disease. In mice as well as in humans CD8þ T cells derived IFN-g enhanced expression of Fc-receptors on CD11bþ cells. High expression of Fc-receptors allowed CD11bþ cells to bind to antibody covered target cells and to destroy them in vivo. We found that autoantibodies induce clinically relevant disease when adaptive immunity, specific for disease non-related antigen, is activated. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Fcgr2b Autoantibodies Systemic lupus erythematosus Inflammation Infection Virus IFN-g

1. Introduction Autoantibodies are a hallmark of autoimmune diseases, including rheumatoid arthritis, autoimmune hepatitis, and systemic lupus erythematosus (SLE) [1e6]. SLE is characterized by multisystemic autoimmunity, including autoimmune manifestations in skin, joints, the hematopoietic system, kidneys, and the central nervous system [7]. The disease is associated with a polyclonal hypergammaglobulinemia and high titers of antibodies against

* Corresponding author. Institute of Immunology, Medical Faculty, University Duisburg-Essen, Essen, Germany. E-mail address: [email protected] (K.S. Lang). 1 These authors contributed equally to this manuscript. http://dx.doi.org/10.1016/j.jaut.2015.05.007 0896-8411/© 2015 Elsevier Ltd. All rights reserved.

numerous self-antigens [8]. Therefore, loss of tolerance to selfantigens is thought to be mainly responsible for the induction of SLE [9,10]. Fcgr2b contributes to lack of tolerance in SLE [11]. Fcgr2b is expressed by follicular dendritic cells, monocytes, macrophages, neutrophils and B cells. Co-activation of Fcgr2b together with the Bcell receptor (BCR) can lead to apoptosis of the activated B cell [12]. Therefore, Fcgr2b is important for maintaining peripheral B-cell tolerance [11,13]. In humans, the Fcgr2bT232 polymorphism exhibits the strongest association with SLE, and is associated with loss of tolerance and generation of autoantibodies [14]. However, presence of autoantibodies alone is not sufficient to induce SLE. This is seen in individuals, who have high titers of autoantibodies, but do not show active disease. In line with that finding autoantibodies are often induced years before the first clinical manifestation of SLE [15].

12

N. Shaabani et al. / Journal of Autoimmunity 62 (2015) 11e21

N. Shaabani et al. / Journal of Autoimmunity 62 (2015) 11e21

Therefore additional signals, besides autoantibodies, seem to be essential for induction of active SLE. However, underlying mechanisms which are involved in conversion of non-pathogenic autoantibodies into active disease remains unknown. In recent years it has become clear that in addition to the loss of self-tolerance, SLE is associated with the over expression of type I interferon (IFN-I)einduced genes, known as the ‘IFN-I signature’. In particular, patients with active disease exhibit enhanced activation of IFN-Ieinduced genes [16,17]. This finding is in line with observations that treatment with recombinant IFN-I can induce symptoms of SLE [18]. In addition, mouse strains with lupus-like disease, such as New Zealand Black mice, lose their autoimmune phenotype when the IFN-I receptor is lacking [19]. Moreover, it has been shown that DNA bound by antinuclear antibodies (ANAs) may activate Toll-like receptor 9 (TLR 9), and this activation contributes to disease [20]. Several immunological processes are induced synergistically by IFN-I and IFN-II (IFN-g) [21,22]. This includes T cell activation, antigen presentation, induction of chemokines and inflammatory processes [23e25]. Therefore strong induction of IFN-g might -similar to IFN-I- contribute to pathogenesis of SLE. How the IFN-I signature is linked with autoantibodies and whether IFN-II can contribute to disease onset is still unknown. Using lupus-prone Fcgr2b/ mice and human patient samples, we attempted to determine the circumstances under which autoantibodies may become pathogenic. In our mouse model, we found that autoantibodies are not associated with clinically significant autoimmunity in vivo. However, infection with the lymphocytic choriomeningitis virus (LCMV) converted “silent” autoimmunity into overt autoimmune disease. Disease onset was linked to the activation of virus specific CD8þ T cells. This activation led to production of IFN-g which upregulated Fcgr1 and Fcgr3 on CD11bþ cells, leading them to attack antibody covered tissue cells. Similarly, serum from human SLE patients activated CD11bþ cells in the presence of IFN-g and frequencies of IFN-gþ CD8þ T cells correlated with SLE activity. In conclusion, we found that, in the presence of strong activation of adaptive immune cells, tissue-binding antibodies can induce clinically relevant disease. 2. Results 2.1. Autoantibodies do not induce disease despite binding to tissue in vivo In order to study the pathogenicity of autoantibodies, we used a murine lupus model, Fcgr2b/ mice. Similar to humans [26,27], genetic variation in Fcgr2b is significantly associated with SLE [28]. Fcgr2b deficient mice develop anti-dsDNA antibodies (Fig. 1A). Almost all mice older than 21 weeks showed elevated anti-nuclear and anti-dsDNA antibodies (Fig. 1A). Consistently, sera from Fcgr2b/ mice which were older than 21 weeks showed antibodiesbinding to liver and kidney as tested by immune fluorescence analysis (Fig. 1B). Wild-type (WT) and young Fcgr2b/ mice (8e20 weeks old) did not show such autoreactivity (Fig. 1B and Supplementary Fig. 1). The induction of autoantibodies was not

13

gender specific, since both male and female Fcgr2b/ mice older than 20 weeks showed induction of autoantibodies (Fig. 1C). In order to use non-genetic lupus mouse model, we treated C57BL/6 mice with Pristane. Administration of Pristane leads to induction of autoantibodies [29]. Serum of Pristane-treated mice showed also tissue binding autoantibodies (Fig. 1D). Next we wondered whether these autoantibodies bind to tissue in vivo. To answer this question, we perfused WT and 26-week-old Fcgr2b/ mice with 100 ml phosphate-buffered saline (PBS) to wash out circulating antibodies and detected tissue binding antibodies by anti-IgG antibody. This analysis showed that autoantibodies bound to several tissues in 26week-old Fcgr2b/ mice, whereas WT mice exhibited no tissue binding (Fig. 1E). Mass spectral analysis revealed that some of those tissue binding antibodies bound to actin and myosin, which explains the binding to a broad range of different tissues (Supplementary Fig. 2). This finding suggested that IgG autoantibodies obtained from autoimmunity prone Fcgr2b/ mice bind to tissue in vivo. To see whether autoantibodies are pathogenic in Fcgr2b/ and Pristane-treated mice we analyzed markers for tissue damage. ALT and AST markers for liver cell damage and LDH marker for general tissue damage were comparable between control and lupus model mice (Fig. 1F). Histology of kidney did not show signs of inflammation (Fig. 1G). Mice remained clinically healthy, no premature death was observed (Fig. 1H). This suggested that presence of autoantibodies did not directly correlate with tissue damage and raised the question, how autoantibodies contribute to autoimmune disease. 2.2. Infection converts autoimmunity into overt SLE We wondered under which circumstances tissue-binding antibodies could support the onset of overt autoimmune disease. We speculated that inflammatory signals in addition to the presence of those antibodies might result in overt autoimmune disease. To test this theory we infected Fcgr2b/ mice with the poorly-cytopathic arenavirus LCMV. This infection is associated with an antiviral inflammatory response within liver, lung, kidney and heart tissues [30e32]. LCMV infection of Fcgr2b/ mice, but not WT mice, with 2  106 PFU led to higher ROS production in splenic CD11bþ cells (Fig. 2A). ROS production was associated with increased tissue destruction three days after infection, as indicated by elevated levels of ALT, AST and LDH (Fig. 2B). Similarly, infecting Pristanetreated mice with LCMV led to tissue destruction relative to control mice (Fig. 2B). Such signs of tissue destruction were not found after LCMV infection of wild-type (WT) mice (Fig. 2B). The destruction was stronger in Fcgr2b/ mice than Pristane-treated mice which may be due to the types and/or the concentration of autoantibodies. Histology revealed hypercellularity in kidney glomeruli in Fcgr2b/ mice (Fig. 2C) and Fcgr2b/ mice showed higher glomerulonephritis score compared to WT mice (Fig. 2D). In line with this, the albumin/creatinine ratio is increased in the urine of Fcgr2b/ mice related to WT mice (Fig. 2E). Tissue destruction was associated with infiltration of CD11bþ cells in liver and kidney of Fcgr2b/ mice (Abþ) (Fig. 2F). The destruction was followed by the rapid death of the majority of Fcgr2b/ (Abþ) mice, whereas

Fig. 1. Autoantibodies do not induce disease. (A) Antiedouble-stranded DNA (anti-dsDNA) antibodies, anti-nuclear antibodies (ANAs), as detected by enzyme-linked immunosorbent assay (ELISA) in serum from naïve wild-type (WT), Fcgr2b/ (8e20 weeks old) and Fcgr2b/ (21e32 weeks old) mice (n ¼ 7e14). (B) Immunofluorescence analysis of tissue sections from naïve Jh/ mice incubated with serum from naïve WT mice and Fcgr2b/ mice (21e32 weeks old). Immunoglobulin G (IgG) binding was detected by antimouse IgG antibody (green), DAPI (blue) (n ¼ 3). Scale bar, 50 mm. (C) Percentage of the both genders of Fcgr2b/ (Abþ) older than 20 weeks (n ¼ 28). (D) Immunofluorescence analysis of tissue sections from naïve Jh/ mice stained with serum from naïve untreated mice or Pristane-treated mice. Immunoglobulin G (IgG) binding was detected by antimouse IgG antibody (green), DAPI (blue) (n ¼ 3). Scale bar, 100 mm. (E) Immunofluorescence of tissue sections from naïve perfused WT mice and Fcgr2b/ (Abþ) mice stained with anti-mouse IgG (green) (n ¼ 4). Scale bars: 100 mm (main images), 20 mm (insets). (F) Indicated parameters were measured in the serum of naïve WT mice and Fcgr2b/ mice (Abþ) (n ¼ 10) or untreated mice and Pristane-treated mice (n ¼ 4). (G) Kidney sections from naïve WT, Fcgr2b/ (Ab) and Fcgr2b/ (Abþ) mice were stained with hematoxylin and eosin (n ¼ 3). Scale bar, 50 mm. (H) Survival of naïve WT and Fcgr2b/ mice (n ¼ 24e50). Data are shown as mean ± s.e.m in (A and F).

14

N. Shaabani et al. / Journal of Autoimmunity 62 (2015) 11e21

N. Shaabani et al. / Journal of Autoimmunity 62 (2015) 11e21

WT mice survived LCMV infection (Fig. 2G). Death of Fcgr2b/ mice was associated with presence of autoantibodies but not with lack of Fcgr2b per se since only Fcgr2b/ mice which developed autoantibodies (Abþ) died during LCMV infection (Fig. 2G and Supplementary Fig. 3). This strong pathogenicity could not be explained by the triggering of auto-reactive B cells during LCMV infection, since autoantibody titers did not increase during LCMV infection (Supplementary Fig. 4). To check whether the pathogenicity is restricted to LCMV infection or other pathogens may trigger relapse of disease, we infected Fcgr2b/ mice with Listeria monocytogenes (L.m.). L.m. infection lead to activation of immune system and high production of IFN-g [33]. Interestingly, Fcgr2b/ mice showed higher susceptibility to L.m. infection relative to WT mice (Fig. 2H). In conclusion, infecting Fcgr2b/ mice converted a state of clinically silent autoimmunity into an overt autoimmune disease. 2.3. IFN-g derived CD8þ T cells trigger disease in SLE mice Next we aimed to analyze the mechanism of virus-dependent pathogenicity in vivo. Virus specific CD8þ T cells are strongly activated during LCMV infection [34,35] and their activation is associated with high production of IFN-g directly after infection. Depletion of CD8þ T cells (a-CD8) or genetic ablation of CD8þ T cells in CD8a/ mice blunted induction of IFN-g in serum (Fig. 3A). In line with these results, CD8þ T cell depleted mice and CD8a/ mice showed reduced mRNA level of IFN-g in peripheral organs comparing to WT mice (Fig. 3B). However, in both wild-type (WT) and Fcgr2b/ mice virus-specific CD8þ T cells were similarly induced upon LCMV infection (Fig. 3C). Re-stimulation with virus-antigen induced IFN-g production in CD8þ T cells derived from WT and Fcgr2b/ mice (Fig. 3D). The replication of LCMV was similar between WT and Fcgr2b/ mice in various tested organs (Supplementary Fig. 5) which suggests that death of Fcgr2b/ mice was not related to viral control. To test whether indeed CD8þ T cells or IFN-g derived CD8þ T cells contributed to disease development, we depleted CD8þ T cells or IFN-g in Fcgr2b/ mice before infection with LCMV. In the absence of either CD8þ T cells or IFN-g ,Fcgr2b/ (Abþ) mice survived LCMV infection compared to none treated Fcgr2b/ mice (Fig. 3E). In order to check whether the depletion of CD8þ T cells or IFN-g may influence the concentration of IFN-a, we measured the IFN-a levels in the serum of untreated mice, a-CD8 and a-IFN-g treated mice after LCMV infection. We found that the concentrations of IFN-a were comparable in all groups (Fig. 3F). We conclude that enhanced pathogenicity of LCMV infected autoantibody-positive Fcgr2b/ mice is CD8þ T cell-dependent and correlated with the production of T cell-derived IFN-g. 2.4. IFN-g license CD11bþ activation and relapse of SLE IFN-g is a strong activator of CD11bþ cells, we speculate that CD8þ T cell- derived IFN-g may activate CD11bþ cells to attack autoantibodies-opsonized tissues. To study the role of IFN-g derived CD8þ T cells in the presence of autoantibodies, we first

15

measured the influence of IFN-g on the function of CD11bþ cells. The upregulation of Fc receptors during virus infection has been suggested to be critical for enhancing the effector function of innate immune cells [36,37]. We therefore analyzed Fcgr upregulation. We treated bone marrow- derived CD11bþ cells with IFN-g. IFN-g treatment led to upregulation of Fcgr1 and Fcgr3 (Fig. 4A). Next we wanted to investigate whether the upregulation of Fc receptors in the presence of autoantibodies led to strong signaling through Fc receptors. We incubated CD11bþ cells with serum of WT or Fcgr2b/  (Abþ) mice in the presence or absence of IFN-g. Interestingly, we found higher phosphorylation of Syk only in the presence of both autoantibodies and IFN-g (Fig. 4B). In order to test whether strong signaling through Fc receptors overactivated CD11bþ cells, we coincubated serum-labeled MC57 cells with bone marrow isolated CD11bþ cells in the presence or absence of IFN-g. In the absence of IFN-g, serum from Fcgr2b/ (Abþ) mice, but not WT controls showed activation of CD11bþ cells and ROS production (Fig. 4C). The addition of recombinant IFN-g further increased the ROS activity of CD11bþ cells (Fig. 4C). ROS production was due to signaling through Fc receptors because treating cells with Piceatannol (PC) which inhibits P-Syk blunted ROS production (Fig. 4C). Interestingly, IFN-a has no impact on stimulation of CD11bþ cells (Fig. 4D). However, CD11bþ cells isolated from bone marrow of Fcgr2b/ showed higher ROS activity compared to WT mice (Supplementary Fig. 6) which may be due to the absence of inhibitory receptor Fcgr2b. To test if IFN-g can activate CD11bþ cells in vivo in the presence of autoantibodies, we injected WT and Fcgr2b/ mice with recombinant IFN-g and measured ROS production in granulocytes. Only autoantibody-positive Fcgr2b/ mice showed increased production of ROS while WT and autoantibody-negative Fcgr2b/ mice did not (Fig. 4E). From those data we concluded that CD8þ T cell-derived IFN-g upregulated Fc receptors on CD11bþ cells. The upregulation of Fc receptors subsequently led to recognition of autoantibody-bounded tissue and activation of CD11bþ cells. 2.5. Frequency of IFN-gþ CD8þ T cells correlates with disease activity in SLE patients Next we wanted to analyze whether our finding in the Fcgr2b/ murine model has an impact in the pathogenesis of human SLE. As expected, SLE patients had strongly enhanced anti-dsDNA antibody titers as compared to healthy controls (Fig. 5A). Analysis of the antidsDNA antibody titer with the SLEDAI disease activity index score did not show correlation with disease severity (Fig. 5B). Therefore, similar to the mouse system, presence of autoantibodies did not directly correlate with disease activity. Next we tested whether human autoantibody-containing serum had the ability to activate CD11bþ cells in the presence of IFN-g. Therefore we incubated CD11bþ cells from healthy donors (blood group O) with serum from healthy individuals or from SLE patients. We found that serum from SLE patients led to activation of CD11bþ cells (Fig. 5C). In the presence of IFN-g this activation was enhanced (Fig. 5C). From those data we hypothesized that enhancement of IFN-g in SLE patients would probably enhance disease activity. We found no significant

Fig. 2. Infection converts autoimmunity into overt autoimmune disease. (A) FACS analysis of ROS production measured in CD11bþ cells from WT or Fcgr2b/ (Abþ) mice on day 4 after infection with 2  106 PFU LCMV-WE or left untreated (n ¼ 3-8). (B) Indicated parameters were measured in the serum of wild-type (WT) mice and autoantibodies tissuebindingepositive Fcgr2b/ mice (Abþ) on day 3 after infection or wild-type (WT) mice and Pristane-treated mice infected with 2  106 plaque-forming units (PFU) of LCMV strain WE on day 6 after infection (LCMV, n ¼ 3e7; Pristane, n ¼ 4e8). (C) Kidney sections of WT, Fcgr2b/ (Ab) and Fcgr2b/ (Abþ) 3 days after intravenous infection with 2  106 PFU of LCMV-WE were stained with hematoxylin and eosin (n ¼ 3). Scale bar, 50 mm. (D) Glomerulonephritis score of WT and Fcgr2b/ (Abþ) 3 days after intravenous infection with 2  106 PFU of LCMV-WE were stained with hematoxylin and eosin (n ¼ 31e36 histological area from 3 mice per group). (E) Albumin/creatinine ratio in the urine of WT mice and Fcgr2b/ (Abþ) mice 8 days after infection with 2  106 (PFU) of LCMV strain WE (n ¼ 3e4). (F) Immunofluorescence of tissue sections from WT, Fcgr2b/ (Ab) and Fcgr2b/ (Abþ) mice infected intravenously with 2  106 (PFU) of LCMV strain WE stained with anti-CD11bþ (red) (n ¼ 3). (G) Survival of WT mice and Fcgr2b/ (Abþ) mice infected intravenously with 2  106 (PFU) of LCMV strain WE (n ¼ 6). (H) Survival of WT mice and Fcgr2b/ (Abþ) mice infected intravenously with 4000 colony forming units (CFU) of Listeria monocytogenes L.m (n ¼ 5e6). Data are shown as mean and s.e.m in (A, B, D and E).

16

N. Shaabani et al. / Journal of Autoimmunity 62 (2015) 11e21

Fig. 3. IFN-g-derived CD8þ T cells trigger disease in SLE mice (A) IFN-g concentration in the serum was measured by ELISA in WT mice left untreated, WT mice treated with antiCD8 (a-CD8) and CD8a/ mice following intravenous LCMV infection with 2  106 PFU at the indicated time points (n ¼ 5). (B) Untreated WT mice, anti-CD8 (a-CD8) treated WT mice and CD8a/ mice were infected intravenously with 2  106 PFU LCMV. On day 6, IFN-g gene expression in different organs was measured (n ¼ 4e5). (C) Wild-type (WT) and Fcgr2b/ (Abþ) mice were infected intravenously with 2  106 plaque-forming units (PFU) of LCMV strain WE. After 8 days, total number of CD8þ T cells and percentage of specific Tet-Gp33þ CD8þ T cells in the blood were measured by tetramer staining and flow-cytometric analysis in WT and Fcgr2b/ (Abþ) mice (n ¼ 5e6). (D) Percentage of IFN-gþ CD8þ T cells was measured by intracellular cytokine staining and flow-cytometric analysis after following in vitro re-stimulation with Gp33 peptide for 6 h from WT and Fcgr2b/ (Abþ) mice (n ¼ 5). (E) Survival of LCMV infected Fcgr2b/ mice (Abþ) treated with a-CD8 depleting antibody, a-IFN-g or control Fcgr2b/ mice (Abþ) left untreated (n ¼ 4e10). (F) IFN-a levels in serum of LCMV (2  106 PFU) infected WT mice treated with a-CD8 depleting antibody, a-IFN-g or left untreated (n ¼ 3). Data are shown mean ± s.e.m in (A, C and F) and mean and s.e.m in (B and D).

difference in the percentage of IFN-gþ CD8þ T cells between healthy individuals and SLE patients (Fig. 5D). However, in line with our mouse data, frequencies of IFN-gþ CD8þ T cells in human SLE patients positively correlated with enhanced disease activity (Fig. 5E). Interestingly, frequencies of IFN-gþ CD8þ T cells correlated better with SLEDAI than anti-dsDNA antibodies. Therefore we concluded that activation of IFN-g producing CD8þ T cells indeed are one pathogenic mechanism triggering disease activity in SLE.

3. Discussion The results of this study suggest that tissue-binding autoantibodies alone are not sufficient to induce overt autoimmune disease. However, further immune activation (induced e.g. by viral infection) induces the production of T cellederived IFN-g, which might attract CD11bþ cells to autoantibody covered tissue and licence them to induce autoimmune tissue damage.

N. Shaabani et al. / Journal of Autoimmunity 62 (2015) 11e21

17

Fig. 4. IFN-g license CD11bþ activation and relapse of SLE. (A) CD11bþ cells were isolated from bone marrow of WT mice and treated with IFN-g or left untreated. After 6 h, Fcgr1 and Fcgr3 gene expression was measured (n ¼ 4e5). (B) Representative immunoblot of MACS sorted CD11bþ cells from bone marrow of WT mice, treated with IFN-g or left untreated for 6 h, followed by challenge with serum from WT or Fcgr2b/ (Abþ) mice for 10 min. Blots were probed with antibodies to Phospo-Syk, Syk, and GAPDH (n ¼ 3). (C) FACS analysis of ROS production of WT CD11bþ cells which were sorted by magnetic-activated cell sorting (MACS) and then co-incubated with MC57 cells and WT serum or coincubated with MC57 cells and Fcgr2b/ serum (Abþ) in the presence or absence of recombinant IFN-g (100 ng/ml) and in the presence or absence of Piceatannol (PC, 2 mM) (n ¼ 6). (D) FACS analysis of ROS production of WT CD11bþ cells which were sorted by magnetic-activated cell sorting (MACS) and then co-incubated with MC57 cells and WT serum or coincubated with MC57 cells and Fcgr2b/ serum (Abþ) in the presence or absence of recombinant IFN-g (100 ng/ml) and in the presence or absence of IFN-a (50 U/ml) (n ¼ 4) (E) FACS analysis of ROS production of granulocytes derived from WT, Fcgr2b/ (Ab), and Fcgr2b/ (Abþ) mice treated in vivo with 2 mg IFN-g intravenously for 12 h or left untreated (n ¼ 6e9). Data are shown mean and s.e.m in (A, C, D and E).

These findings demonstrate that autoantibody-associated disease may progress in the presence of additional adaptive immune activation, and this additional immune activation might explain the occurrence of phases of relapse and remission in human autoimmune disease such as SLE. In SLE, disease activity is associated with the onset of infection [38e41] with viruses such as Epstein-Barr virus, Measles virus, Myxovirus, and endogenous retroviruses [42e46]. Our results indicate that IFN-g induced by viral infection might activate CD11bþ cells, which then are further activated in the presence of tissue-binding antibodies. In the past, the association of SLE with virus infections has often been explained by cross-reactive viral-derived antigens and autoantigens [39,47]. The fact that several of the viruses which are famous to associate with SLE do not have known cross-reactive antigen suggests that other factors are important for inducing disease in lupus-prone individuals [39]. Our study shows that autoantibodies are not directly affected by viral infection; however, the inflammation induced by the virus, especially the virus-specific IFN-gþ CD8þ T cell leads to a strong

activation of granulocytes that then attack tissue in the presence of tissue-binding autoantibodies. One of CD11bþ cells are granulocytes which were linked to the pathogenesis of SLE by having the capacity to synthesize neutrophil-derived extracellular traps (NETs) [48], which cause more tissue infiltration and endothelial damage [49,50]. On the other hand, SLE is associated with increased levels of neutrophil apoptosis [51]. The connection between autoantibodies, inflammatory signals and granulocytes has so far not been addressed. From our data we suggest that presence of autoantibodies and IFNg producing CD8þ T cells will lead to sensitization of granulocytes to antibody covered tissue. There is a close relation between the T cell stimulating cytokine interleukin-2 (IL-2) and induction of SLE [52]. In vitro study of T cells from SLE patients showed defect in the ability to produce IL-2 [52]. However, T cells from SLE patients show upregulation of IL-2R and this upregulation correlated with the activity of the disease [53]. In line with that, an increase in the CD8þ lymphocytes expressing

18

N. Shaabani et al. / Journal of Autoimmunity 62 (2015) 11e21

Fig. 5. SLE patients show enhanced activated CD8þ T cells.(AeC) For correlative studies of SLEDAI with anti-dsDNA healthy individuals and patients were recruited. (A) Anti-doublestranded DNA (anti-dsDNA) antibodies, in serum of healthy individuals and SLE patients (n ¼ 10e19). (B) Correlation plot of SLE Disease Activity Index (SLEDAI) score versus antidouble-stranded DNA autoantibody titers in serum of SLE patients (n ¼ 19). (C) Percentage of ROS producing CD11bþ cells derived from a healthy donor (O) which were coincubated with serum from healthy individuals or co-incubated with serum from SLE patients in the presence or absence of IFN-g (100 ng/ml) analyzed by flow cytometry 90 min after incubation (n ¼ 13e19). (D-E) To another time point, healthy and SLE patients were again recruited and SLEDAI scores were re-evaluated for correlative studies of SLEDAI with IFN-gþ CD8þ T cells. (D) Percentage of IFN-gþ CD8þ T cells from blood samples of healthy individuals or SLE patients after re-stimulation with Phorbol 12-myrstate 13-acetate (PMA) and Ionomycin measured with flow cytometry 6 h after re-stimulation (n ¼ 10e18). (E) Correlation of percentage of IFN-gþ CD8þ T cells (as measured in D) with SLEDAI (score) from the tested SLE patients (n ¼ 18). (Student's t-test; untreated samples, Paired t-test; SLE samples). Data are shown mean ± s.e.m. in (A and D) and mean and s.e.m in (C).

perforin and granzyme B was correlated with enhanced disease activity in SLE patients [54]. Additionally, an increase in the number of urinary effector T cells was noticed in active lupus nephritis [55,56]. We show here that frequencies of IFN-gþ producing CD8þ T cells are comparable between SLE patients and healthy controls. However high frequencies of IFN-gþ CD8þ T cells correlated with disease activity in the presence of autoantibodies. This suggests that antiviral IFN-gþ producing CD8þ T cells participate in the pathogenesis of autoantibodies. Changes of the numbers and/or activity of IFN-gþ producing CD8þ T cells can therefore explain the phases of flares and remission of SLE disease. In summary, we propose that SLE may be separated in two pathogenetic steps in vivo. First, loss of tolerance leads to the activation of autoreactive B cells and the production of autoantibodies. Second activation of CD8þ T cells by disease non-related

antigens might licence CD11bþ cells to migrate to autoantibody covered tissue and to induce tissue damage. 4. Methods 4.1. Mice, treatment, viruses and bacteria All of the mice strains were maintained at university hospital Essen animal house facility (Essen, Germany). Animal experiments were carried out in Germany under the authorization of €ramt of Düsseldorf, in accordance with the German the Veterina laws for animal protection. All experiments were performed with the animals housed in single ventilated cages. Jh/ mice and Fcgr2b/ mice were kept on C57BL/6 background. Jh/ mice were 6e8 weeks old. Fcgr2b/ (Ab) mice were 8e20 weeks old.

N. Shaabani et al. / Journal of Autoimmunity 62 (2015) 11e21

Fcgr2b/ (Abþ) mice were 21e32 weeks old. Autoantibodies in serum were proved before starting experiments. For Pristane treatment, 8 weeks old mice were injected intraperitoneally with 134,26 mg Pristane per mouse and autoantibodies were detected after 6 weeks. Cell depletions in vivo were performed with rat monoclonal antibodies specific for CD8þ T cells (YTS 169.4), as previously described [57]. The hybridomas were initially obtained from H. Waldmann. We injected 200 mL of purified CD8þ T celledepleting antibody intraperitoneally on days 3 and 1 before infecting mice with the lymphocytic choriomeningitis virus (LCMV). The efficiency of depletion was always confirmed by flow-cytometric analysis of peripheral blood. LCMV strain WE (LCMV-WE) was originally obtained from F. Lehmann-Grube (Heinrich Pette Institute, Hamburg, Germany) and was propagated in L929 cells. Virus titers were measured by focusforming assay, as previously described [58]. L. monocytogenes were grown in brain heart infusion overnight. 4.2. Recruitment of patients For correlative study of SLEDAI with anti-dsDNA, this study enrolled systemic lupus erythematosus (SLE) patients (n ¼ 19) aged 42.4 ± 14.2 (mean ± SD) years who attended the outpatient clinic, and healthy control subjects (n ¼ 10 for anti-dsDNA aged 40.9 ± 16.5 and n ¼ 13 for measurement of ROS production aged 40.4 ± 14.3 years). All patients fulfilled at least four of the American College of Rheumatology's revised criteria for a diagnosis of SLE [59]. All patients included were anti-dsDNA antibody positive at time of recruitment or in the past. Disease activity was assessed by the SLE Disease Activity Index (SLEDAI). Eleven patients had inactive disease (SLEDAI score  4) and eight patients had active SLE (defined as SLEDAI score > 4). Median disease activity for all patients was 4.0 (range 0e12). Sixteen patients had a current or former renal biopsy consistent with lupus nephritis (LN) while three had no history of renal involvement. All patients were on a constant dose of immunomodulating drugs (azathioprine (n ¼ 2), mycophenolate mofetil (n ¼ 9), cyclosporine A (n ¼ 2), hydroxychloroquine (n ¼ 7). Sixteen patients receive additionally a median dose of prednisone of 6.25 mg/d (5e50 mg/ d). For stimulation experiments we included to another time point subsequently eighteen SLE patients (n ¼ 18) aged 42.4 ± 15.9 (mean ± SD) years who attended the outpatient clinic, five with active disease and thirteen with inactive disease. Healthy control subjects (n ¼ 10) aged 39.4 ± 12.5 years. Ten patients had a current or former renal biopsy consistent with LN while eight had no history of renal involvement. All patients were on a constant dose of immunomodulating drugs (azathioprine (n ¼ 2), mycophenolate mofetil (n ¼ 10), cyclosporine A (n ¼ 4), hydroxychloroquine (n ¼ 10). Thirteen patients receive additionally a median dose of prednisone of 5 mg/d (2.5e100 mg/d). 4.3. Histology Histological analyses were performed on snap-frozen or formalin-fixed tissues. Serum was diluted 1:100 unless stated otherwise in the figure legends. CD45R (B220), Gr-1 (Ly6G), CD11b, antibodies were purchased from eBioscience (San Diego, CA, USA). Fluorescein isothiocyanate (FITC)-AffiniPureF(ab')2 Fragment Goat anti-Mouse IgG (H þ L) was purchased from Jackson Immuno Research (West Grove, PA, USA). Hematoxylin and Eosin for H&E staining were purchased from Merck (Darmstadt, Germany) and Sigma-Aldrich (Steinheim, Germany). Bone marrow histology was done by cytospin of cell suspension.

19

For glomerulonephritis score, glomeruli were assessed to determine the crescent infiltration of H&E kidney sections. The crescent formation in each glomerulus was scored from 0 to 3 as follows: 0, no crescent present; 1, if crescent is occupying up to 1/3 of glomerulus; 2, if crescent is occupying up to 2/3 of glomerulus; and 3, crescent is occupying more than 2/3 of glomerulus. 4.4. Enzyme-linked immunosorbent assay ELISAs for interferon gamma (IFN-g; eBioscience, San Diego, CA, USA), anti-nuclear antibodies (ANAs), and antiedouble-stranded DNA (anti-dsDNA) (Alpha Diagnostic, San Antonio, USA) were performed according to the manufacturer's instructions. For actin and myosin ELISAs 96 well plates were coated with anti-actin (Sigma-Aldrich, Steinheim, Germany) and anti-myosin (Abcam, Cambridge, UK) antibodies respectively. Plates were then incubated with lysates from HEK cells. Incubation with lysates from HEK cells did not lead to detectable signals and was considered as background control. 4.5. Perfusion Mice were anesthesized, and a catheter was placed into the portal vein. The heart and aorta were opened, and the liver was perfused with 100 ml phosphate-buffered saline (PBS). 4.6. Flowcytometric analysis Tetramer staining and surface FACS staining were performed as previously described [58]. 4.7. ROS production measurement For mouse samples CD11bþ cells were incubated with Dihydrodamine (DHR) (10 mg) (Enzo Life Sciences Inc; NY, USA) for 30 min at 37  C, in the presence or absence of mouse recombinant IFN-g (100 ng) (Life Technologies; CA, USA), in the presence or absence of mouse recombinant IFN-a (50 U) (PBL Assay Science; NJ, USA) and Piceatannol (2 mM) (Sigma; St Louis, MO, USA). CD11bþ cells were stained for additional 30 min with anti-mouse CD11b (eBioscience; San Diego, CA, USA). For human samples, blood from healthy donor type (Oþ) was incubated with serum from SLE patients or healthy individuals, Dihydrodamine (DHR) (10 mg) was added for 30 min at 37  C in the presence or absence of human recombinant IFN-g (100 ng) (R&D, Emeryville, CA, USA). Afterward, granulocytes were stained for additional 30 min with anti-human CD11b (eBioscience; San Diego, CA, USA). 4.8. CD8þ T cell stimulation Blood from SLE patients or healthy individuals was stimulated with Ionomycin 2.5 mg (Sigma; Saint Louis, MO, USA) and Phorbol 12 myristate 13-acetate (PMA) 20 mg (Sigma; Saint Louis, MO, USA) and Brefeldin-A 10 mg (Sigma; Saint Louis, MO, USA) for 4 h at 37  C. CD8þ T cells were stained with anti-human CD8 (eBioscience; San Diego, CA, USA). The intracellular cytokine staining was done after fixation with Formaldehyde and permeabilization with Saponin. IFN-g was stained with anti-human IFN-g (eBioscience; San Diego, CA, USA). 4.9. ALT, AST and LDH measurement Biochemical analyses were done by the central laboratory, Heinrich- Heine-University Düsseldorf, Düsseldorf, Germany.

20

N. Shaabani et al. / Journal of Autoimmunity 62 (2015) 11e21

4.9.1. Immunoblotting CD11bþ cells were isolated from bone marrow by positive selection with CD11bþ-conjugated magnetic beads (MACS, Miltenyi Biotech). 2  106 cells were cultured with or without 1 mM/ml IFN-g (Sigma-Aldrich). After 6 h incubation at 37  C, cells were challenged with Fcgr2b/ serum for 10 min. Cells were lysed with boiling sodium dodecyl sulfate (SDS) buffer (1.1% SDS, 11% glycerol, 0.1 M Tris; pH 6.8) with 10% 2-Mercaptoethanol. Total cell extracts were examined by 10% SDS-PAGE gels and transferred onto Whatman nitrocellulose membrane (GE Healthcare) by standard techniques. Membranes were blocked for 1 h in 5% Nonfat dried milk powder (AppliChem) in TBS supplemented with 1% Tween-20 and incubated with the following antibodies: anti-Phospho-Syk/ZAP (Y352/ Y319); anti-Syk (from Cell Signaling Technologies). Secondary antibody anti-GAPDH (Meridian Life Science) was detected by horse radish peroxidase (HRP)-conjugated anti-mouse IgG (BIO RAD) or anti-rabbit IgG (GE Healthcare) antibodies, or both. Signals were detected with the BIO RAD ChemiDoc imaging system and analyzed with the manufacturer's software. 4.10. Statistical analyses When appropriate, data are expressed as mean ± S.E.M or mean and SEMs. Statistically significant differences between two groups were determined with the unpaired Student's t-test or Paired t-test between lupus samples (Fig. 5C). Statistically significant differences between experimental groups over multiple time points were determined with two-way ANOVA (repeated measurements). Statistical significance was set at the level of * P < 0.05, ** P < 0.01 and *** P < 0.00. Ethical aspects All animal experiments were approved by the Landesamt für Umwelt und Verbraucherschutz (LANUV), NRW, Germany (Az.: 8402.04.2014.A076). All human samples were collected with approval of Ethik-Kommission, NRW, Germany (12-5149-BO). Acknowledgments €ttel and Patricia Spieker for technical We thank Konstanze Scha support. This study was supported by the Alexander von Humboldt Foundation (SKA-2008 to K.S.L. and SKA-2010 to P.A.L.), and the Deutsche Forschungsgemeinschaft (DFG) with CRC974, CRC/TRR60 and LA1419/5-1. National Institutes of Health (NIH, USA) Tetramer Core Facility provided the tetramers. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jaut.2015.05.007. References [1] A. Wiik, Autoantibodies in vasculitis, Arthritis Res. Ther. 5 (2003) 147e152. [2] B. Baslund, A. Wiik, Anti-neutrophil cytoplasmic autoantibodies (ANCA) and vasculitis, Clin. Rev. Allergy 12 (1994) 297e304. [3] P.E. Lipsky, Systemic lupus erythematosus: an autoimmune disease of B cell hyperactivity, Nat. Immunol. 2 (2001) 764e766. [4] M. Eggert, U.K. Zettl, G. Neeck, Autoantibodies in autoimmune diseases, Curr. Pharm. Des. 16 (2010) 1634e1643. [5] D.P. Bogdanos, G. Mieli-Vergani, D. Vergani, Autoantibodies and their antigens in autoimmune hepatitis, Semin. Liver Dis. 29 (2009) 241e253. [6] S.J. Lee, A. Kavanaugh, 4. Autoimmunity, vasculitis, and autoantibodies, J. Allergy Clin. Immunol. 117 (2006) S445eS450. [7] G.C. Tsokos, Systemic lupus erythematosus, N. Engl. J. Med. 365 (2011) 2110e2121.

[8] O.T. Preble, R.J. Black, R.M. Friedman, J.H. Klippel, J. Vilcek, Systemic lupus erythematosus: presence in human serum of an unusual acid-labile leukocyte interferon, Science 216 (1982) 429e431. [9] S. Nagata, R. Hanayama, K. Kawane, Autoimmunity and the clearance of dead cells, Cell 140 (2010) 619e630. [10] P.M. Rumore, C.R. Steinman, Endogenous circulating DNA in systemic lupus erythematosus. Occurrence as multimeric complexes bound to histone, J. Clin. Invest. 86 (1990) 69e74. [11] T.L. McGaha, B. Sorrentino, J.V. Ravetch, Restoration of tolerance in lupus by targeted inhibitory receptor expression, Sci. (New York, N. Y. ) 307 (2005) 590e593. [12] V. Paunovic, N.A. Carter, T. Thalhamer, D. Blair, B. Gordon, E. Lacey, A.M. Michie, M.M. Harnett, Immune complex-mediated co-ligation of the BCR with FcgammaRIIB results in homeostatic apoptosis of B cells involving Fas signalling that is defective in the MRL/Lpr model of systemic lupus erythematosus, J. Autoimmun. 39 (2012) 332e346. [13] F. Nimmerjahn, J.V. Ravetch, Fcgamma receptors as regulators of immune responses, Nat. Rev. Immunol. 8 (2008) 34e47. [14] C. Kyogoku, H.M. Dijstelbloem, N. Tsuchiya, Y. Hatta, H. Kato, A. Yamaguchi, T. Fukazawa, M.D. Jansen, H. Hashimoto, J.G. van de Winkel, et al., Fcgamma receptor gene polymorphisms in Japanese patients with systemic lupus erythematosus: contribution of FCGR2B to genetic susceptibility, Arthritis Rheum. 46 (2002) 1242e1254. [15] M.R. Arbuckle, M.T. McClain, M.V. Rubertone, R.H. Scofield, G.J. Dennis, J.A. James, J.B. Harley, Development of autoantibodies before the clinical onset of systemic lupus erythematosus, N. Engl. J. Med. 349 (2003) 1526e1533. [16] L. Bennett, A.K. Palucka, E. Arce, V. Cantrell, J. Borvak, J. Banchereau, V. Pascual, Interferon and granulopoiesis signatures in systemic lupus erythematosus blood, J. Exp. Med. 197 (2003) 711e723. [17] A. Marshak-Rothstein, Toll-like receptors in systemic autoimmune disease, Nat. Rev. Immunol. 6 (2006) 823e835. [18] C. Gota, L. Calabrese, Induction of clinical autoimmune disease by therapeutic interferon-alpha, Autoimmunity 36 (2003) 511e518. [19] M.L. Santiago-Raber, R. Baccala, K.M. Haraldsson, D. Choubey, T.A. Stewart, D.H. Kono, A.N. Theofilopoulos, Type-I interferon receptor deficiency reduces lupus-like disease in NZB mice, J. Exp. Med. 197 (2003) 777e788. [20] S.R. Christensen, J. Shupe, K. Nickerson, M. Kashgarian, R.A. Flavell, M.J. Shlomchik, Toll-like receptor 7 and TLR9 dictate autoantibody specificity and have opposing inflammatory and regulatory roles in a murine model of lupus, Immunity 25 (2006) 417e428. [21] K. Schroder, P.J. Hertzog, T. Ravasi, D.A. Hume, Interferon-gamma: an overview of signals, mechanisms and functions, J. Leukoc. Biol. 75 (2004) 163e189. [22] S. Hervas-Stubbs, J.L. Perez-Gracia, A. Rouzaut, M.F. Sanmamed, A. Le Bon, I. Melero, Direct effects of type I interferons on cells of the immune system, Clin. Cancer Res. 17 (2011) 2619e2627. [23] U. Boehm, T. Klamp, M. Groot, J.C. Howard, Cellular responses to interferongamma, Annu. Rev. Immunol. 15 (1997) 749e795. [24] X. Zhang, S. Sun, I. Hwang, D.F. Tough, J. Sprent, Potent and selective stimulation of memory-phenotype CD8þ T cells in vivo by IL-15, Immunity 8 (1998) 591e599. [25] M. Montoya, G. Schiavoni, F. Mattei, I. Gresser, F. Belardelli, P. Borrow, D.F. Tough, Type I interferons produced by dendritic cells promote their phenotypic and functional activation, Blood 99 (2002) 3263e3271. [26] A. Rahman, D.A. Isenberg, Systemic lupus erythematosus, N. Engl. J. Med. 358 (2008) 929e939. [27] L.C. Willcocks, E.J. Carr, H.A. Niederer, T.F. Rayner, T.N. Williams, W. Yang, J.A. Scott, B.C. Urban, N. Peshu, T.J. Vyse, et al., A defunctioning polymorphism in FCGR2B is associated with protection against malaria but susceptibility to systemic lupus erythematosus, Proc. Natl. Acad. Sci. U S A 107 (2010) 7881e7885. [28] H. Fukuyama, F. Nimmerjahn, J.V. Ravetch, The inhibitory Fcgamma receptor modulates autoimmunity by limiting the accumulation of immunoglobulin Gþ anti-DNA plasma cells, Nat. Immunol. 6 (2005) 99e106. [29] H. Leiss, B. Niederreiter, T. Bandur, B. Schwarzecker, S. Bluml, G. Steiner, W. Ulrich, J.S. Smolen, G.H. Stummvoll, Pristane-induced lupus as a model of human lupus arthritis: evolvement of autoantibodies, internal organ and joint inflammation, Lupus 22 (2013) 778e792. [30] P.A. Lang, M. Recher, D. Haussinger, K.S. Lang, Genes determining the course of virus persistence in the liver: lessons from murine infection with lymphocytic choriomeningitis virus, Cell. Physiol. Biochem. 26 (2010) 263e272. [31] D. Kagi, B. Ledermann, K. Burki, R.M. Zinkernagel, H. Hengartner, Molecular mechanisms of lymphocyte-mediated cytotoxicity and their role in immunological protection and pathogenesis in vivo, Annu. Rev. Immunol. 14 (1996) 207e232. [32] R.M. Zinkernagel, Lymphocytic choriomeningitis virus and immunology, Curr. Top. Microbiol. Immunol. 263 (2002) 1e5. [33] C. Thale, A.F. Kiderlen, Sources of interferon-gamma (IFN-gamma) in early immune response to Listeria monocytogenes, Immunobiology 210 (2005) 673e683. [34] P.A. Lang, C. Contaldo, P. Georgiev, A.M. El-Badry, M. Recher, M. Kurrer, L. Cervantes-Barragan, B. Ludewig, T. Calzascia, B. Bolinger, et al., Aggravation of viral hepatitis by platelet-derived serotonin, Nat. Med. 14 (2008) 756e761. [35] M. Recher, K.S. Lang, A. Navarini, L. Hunziker, P.A. Lang, K. Fink, S. Freigang, P. Georgiev, L. Hangartner, R. Zellweger, et al., Extralymphatic virus

N. Shaabani et al. / Journal of Autoimmunity 62 (2015) 11e21

[36]

[37]

[38] [39] [40]

[41] [42] [43] [44] [45] [46] [47] [48] [49]

sanctuaries as a consequence of potent T-cell activation, Nat. Med. 13 (2007) 1316e1323. H.H. Radeke, I. Janssen-Graalfs, E.N. Sowa, N. Chouchakova, J. Skokowa, F. Loscher, R.E. Schmidt, P. Heeringa, J.E. Gessner, Opposite regulation of type II and III receptors for immunoglobulin G in mouse glomerular mesangial cells and in the induction of anti-glomerular basement membrane (GBM) nephritis, J. Biol. Chem. 277 (2002) 27535e27544. N. Joller, S.S. Weber, A.J. Muller, R. Sporri, P. Selchow, P. Sander, H. Hilbi, A. Oxenius, Antibodies protect against intracellular bacteria by Fc receptormediated lysosomal targeting, Proc. Natl. Acad. Sci. U S A 107 (2010) 20441e20446. E. Toussirot, J. Roudier, Epstein-Barr virus in autoimmune diseases, Best. Pract. Res. Clin. Rheumatol. 22 (2008) 883e896. J.B. Harley, J.A. James, Epstein-Barr virus infection induces lupus autoimmunity, Bull. NYU Hosp. Jt. Dis. 64 (2006) 45e50. M. Ramos-Casals, A. Pares, L.J. Jara, R. Solans, O. Vinas, P. Vazquez, J.M. Sanchez-Tapias, J. Rodes, J. Font, Antimitochondrial antibodies in patients with chronic hepatitis C virus infection: description of 18 cases and review of the literature, J. Viral Hepat. 12 (2005) 648e654. S. Ijichi, M. Osame, Human T-cell lymphotropic virus type I and systemic lupus erythematosus, Intern. Med. 38 (1999) 459e460. E.M. Morgan, F. Rapp, Measles virus and its associated diseases, Bacteriol. Rev. 41 (1977) 636e666. F. Gyorkey, K.W. Min, J.G. Sincovics, P. Gyorkey, Systemic lupus erythematosus and myxovirus, N. Engl. J. Med. 280 (1969) 333. M. Ziff, Viruses and the connective tissue diseases, Ann. Intern. Med. 75 (1971) 951e958. C.L. Christian, P.E. Phillips, Viruses and autoimmunity, Am. J. Med. 54 (1973) 611e620. A. Perl, D. Fernandez, T. Telarico, P.E. Phillips, Endogenous retroviral pathogenesis in lupus, Curr. Opin. Rheumatol. 22 (2010) 483e492. N. Agmon-Levin, M. Blank, Z. Paz, Y. Shoenfeld, Molecular mimicry in systemic lupus erythematosus, Lupus 18 (2009) 1181e1185. M.J. Kaplan, Neutrophils in the pathogenesis and manifestations of SLE, Nat. Rev. Rheumatol. 7 (2011) 691e699. E. Villanueva, S. Yalavarthi, C.C. Berthier, J.B. Hodgin, R. Khandpur, A.M. Lin, C.J. Rubin, W. Zhao, S.H. Olsen, M. Klinker, et al., Netting neutrophils induce

[50]

[51]

[52]

[53]

[54]

[55]

[56]

[57]

[58]

[59]

21

endothelial damage, infiltrate tissues, and expose immunostimulatory molecules in systemic lupus erythematosus, J. Immunol. 187 (2011) 538e552. C. Carmona-Rivera, M.J. Kaplan, Low-density granulocytes: a distinct class of neutrophils in systemic autoimmunity, Semin. Immunopathol. 35 (2013) 455e463. Y. Ren, J. Tang, M.Y. Mok, A.W. Chan, A. Wu, C.S. Lau, Increased apoptotic neutrophils and macrophages and impaired macrophage phagocytic clearance of apoptotic neutrophils in systemic lupus erythematosus, Arthritis Rheum. 48 (2003) 2888e2897. L.A. Lieberman, G.C. Tsokos, The IL-2 defect in systemic lupus erythematosus disease has an expansive effect on host immunity, J. Biomed. Biotechnol. (2010), 740619. D.R. Wigfall, R.S. Sakai, D.J. Wallace, S.C. Jordan, Interleukin-2 receptor expression in peripheral blood lymphocytes from systemic lupus erythematosus patients: relationship to clinical activity, Clin. Immunol. Immunopathol. 47 (1988) 354e362. P. Blanco, V. Pitard, J.F. Viallard, J.L. Taupin, J.L. Pellegrin, J.F. Moreau, Increase in activated CD8þ T lymphocytes expressing perforin and granzyme B correlates with disease activity in patients with systemic lupus erythematosus, Arthritis Rheum. 52 (2005) 201e211. S. Dolff, W.H. Abdulahad, S. Arends, M.C. van Dijk, P.C. Limburg, C.G. Kallenberg, M. Bijl, Urinary CD8þ T-cell counts discriminate between active and inactive lupus nephritis, Arthritis Res. Ther. 15 (2013) R36. S. Dolff, W.H. Abdulahad, M.C. van Dijk, P.C. Limburg, C.G. Kallenberg, M. Bijl, Urinary T cells in active lupus nephritis show an effector memory phenotype, Ann. Rheum. Dis. 69 (2010) 2034e2041. M. Recher, K.S. Lang, L. Hunziker, S. Freigang, B. Eschli, N.L. Harris, A. Navarini, B.M. Senn, K. Fink, M. Lotscher, et al., Deliberate removal of T cell help improves virus-neutralizing antibody production, Nat. Immunol. 5 (2004) 934e942. K.S. Lang, P. Georgiev, M. Recher, A.A. Navarini, A. Bergthaler, M. Heikenwalder, N.L. Harris, T. Junt, B. Odermatt, P.-A. Clavien, et al., Immunoprivileged status of the liver is controlled by Toll-like receptor 3 signaling, J. Clin. Invest. 116 (2006) 2456e2463. M.C. Hochberg, Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus, Arthritis Rheum. 40 (1997) 1725.