AUTOIMMUNITY

Summary Viral infections can be fatal because of the direct cytopathic effects of the virus or the induction of a strong, uncontrolled inflammatory response. Virus and host intrinsic characteristics strongly modulate the outcome of viral infections. Recently we determined the circumstances under which enhanced replication of virus within the lymphoid tissue is beneficial for the outcome of a disease. This enforced viral replication promotes anti‐viral immune activation and, counterintuitively, accelerates virus control. In this review we summarize the mechanisms that contribute to enforced viral replication. Antigen‐presenting cells and CD169+ macrophages exhibit enforced viral replication after infection with the model viruses lymphocytic choriomeningitis virus (LCMV) and vesicular stomatitis virus (VSV). Ubiquitin‐specific peptidase 18 (Usp18), an endogenous type I interferon blocker in CD169+ macrophages, has been identified as a proviral gene, as are B cell activating factor (BAFF) and carcinoembryonic antigen‐related cell adhesion molecule 1 (CEACAM1). Lymphotoxins (LT) strongly enhance viral replication in the spleen and lymph nodes. All these factors modulate splenic architecture and thereby promote the development of CD169+ macrophages. Tumor necrosis factor alpha (TNF‐α) and nuclear factor kappa‐light‐chain‐enhancer of activated B cell signaling (NF‐κB) have been found to promote the survival of infected CD169+ macrophages, thereby similarly promoting enforced viral replication. Association of autoimmune disease with infections is evident from (1) autoimmune phenomena described during a chronic virus infection; (2) onset of autoimmune disease simultaneous to viral infections; and (3) experimental evidence. Involvement of virus infection during onset of type I diabetes is strongly evident. Epstein–Bar virus (EBV) infection was discussed to be involved in the pathogenesis of systemic lupus erythematosus. In conclusion, several mechanisms promote viral replication in secondary lymphatic organs. Identifying such factors in humans is a challenge for future studies. Introduction Rapid spread of virus to a susceptible organ causes serious tissue damage, which can lead to fatal disease because of the direct cytopathic effects of the virus. A typically infectious dose of one to a few infectious particles must be replicated several times before it reaches such harmful levels. During this time window, the immune system must be activated so that viral replication can be limited and the spread of virus to the susceptible organ can be avoided. Replication of virus in secondary lymphatic organs accelerates immune activation, and thereby helps to control the virus. Dendritic cells and CD169+ macrophages are important initiators of the anti‐viral immune response The importance of the splenic cellular architecture on viral replication has been recognized for decades 1. Various cell types influence efficient virus replication. In particular, antigen‐presenting cells (APCs), such as specialized macrophages and dendritic cells (DCs), exhibit strong immune‐activating ability. Therefore, viral replication in these cells exerts an important effect on the outcome of an infection. Viral infection leads to antigen presentation on major histocompatibility complex (MHC) class I molecules of DCs, which contributes strongly to the priming of cytotoxic T lymphocytes (CTLs). Even in a situation where viral replication is strongly limited (i.e. infection of mice, which lack the entry receptor of human poliovirus, with human poliovirus) DCs manage to replicate sufficient amounts of virus to induce priming of CD8+ T cells 2. Therefore, it has been suggested that DCs have a special ability to be infected by virus. This ability, as well as the expression of co‐stimulatory molecules would explain the strong immune‐activating ability of DCs during viral infection. However, infectivity of virus may also be dependent on the virus type and the DC subset. Indeed, other studies have indicated that some DC subsets are not susceptible to viral infection 3, 4. The enhanced viral replication in CD169+ macrophages is obviously even more important in those instances where DCs themselves cannot be infected. In such cases, antigen transfer and cross‐presentation by DCs would become important for CD8+ T cell activation 5, 6. Furthermore, we draw attention to the importance of metallophilic macrophages and subcapsular sinus macrophages that express a C‐type lectin, CD169 (Siglec‐1). During virus infection, CD169+ macrophages play an important role in activating the innate and adaptive immune systems by capturing virus and promoting viral replication 7-10. CD169+ macrophages can directly activate CD8+ T cells, but can also transfer antigen to DCs 11. In line with this, CD169+ macrophages can activate CD8+ T cells against non‐replicating antigen 11-13. The absence of CD169+ macrophages leads to a fatal disease outcome because of reduced interferon (IFN) production and limited activation of adaptive immune cells 8, 14, 15. Although we found that the replication ability of these macrophages is an important factor explaining the strong effect of CD169+ macrophages on the anti‐viral immune response, additional mechanisms by which CD169+ macrophages contribute to anti‐viral immune activation must be determined. Mechanisms of splenic architecture maintenance Overview The proper development and function of CD169+ macrophages rely on an intact splenic architecture. Several cytokines and chemokines (i.e. RANK, LXRa), maintain the splenic architecture and thereby influence the development of CD169+ macrophages 16-18. Similarly, the proper development of CD169+ macrophages relies on a balanced B cell function 1, 19. Factors that influence B cell development have a strong effect on viral replication in the spleen of mice. The general absence of B cells or the absence of tumor necrosis factor (TNF)‐α or lymphotoxin (LT) production, which both is linked to the normal function of B cells, influences splenic maintenance and virus replication during lymphocytic choriomeningitis virus (LCMV) infection 20. The promotion of virus replication in CD169+ macrophages relies solely on the presence of B cells rather than on the antibody response 19. In addition, certain members of the TNF superfamily 21, i.e. LTα, LTβ, TNF‐α, B cell‐activating factor (BAFF) and carcinoembryonic antigen‐related cell adhesion molecule 1 (CEACAM1) 22, 23, have been found to be similarly important for the maintenance of splenic architecture because of their influence on B cell development 21-23. Although B cells and members of the TNF superfamily play an important role in the function of macrophages, their impairment of the immune response of DCs is also important. Lymphotoxins are cytokines that consist of LTα and LTβ. Surface‐bound LT expression has been detected on innate lymphoid cells (ILCs), DCs, natural killer (NK) cells and B and T lymphocytes 24. LTs serve several functions in vivo, including organ maintenance, lymphoid organogenesis and immune response regulation 25, 26. Deficiency in LTα, LTβ or lymphotoxin‐beta receptor (LTβR) can manipulate the development of not only Peyer’s patches and lymph nodes, but also of CD169+ macrophages 20. Interestingly, despite a reduction in CD169+ macrophages, LTβ‐deficient mice exhibit reduced viral replication 20. Mice lacking LTβ fail to restrict the replication of vesicular stomatitis virus (VSV) to the marginal zone of the spleen, and this failure leads to fatal outcome of the infection 20. The impact of LTβ deficiency is restricted not only to macrophages, but also to DCs. Mechanistically, LTβ expression influences maturation, activation and the homing ability of lymphocytes. Detailed mechanisms by which LT influences the development of CD169+ macrophages remain to be determined. LT not only influences the development of CD169+ macrophages but also contributes to the generation of splenic conduits, whose function is essential for early virus distribution in the spleen and thereby influences innate immune activation during virus infection 15. Although the absence of LT is clearly linked to impaired organogenesis and reduced numbers of CD169+ macrophages, the role of TNF‐α in the development of CD169+ macrophages is controversial. The impact of TNF‐α on the formation of primary B cells and germinal centers, as well as on follicular DC (FDC) networks, has been studied intensively 27-29, but its role in the maintenance of CD169+ macrophages remains controversial 27-29. On one hand, some reports state that mice deficient in TNF‐α and mice deficient in p55 tumor necrosis factor receptor 1 (TNF‐R1) exhibit severe impairment in marginal zone development 29; on the other hand, some reports hint at the depletion activity of TNF‐α after infection. Excessive production of TNF‐α in mice infected with Leishmania donovani leads to rapid remodeling of splenic structure and loss of marginal zone macrophages (MZMs). In line with this finding, other researchers have found that the pathogenic burden during L. donovani infection can be reversed by TNF depletion 30. In contrast, a recent study found reductions in the numbers of CD169+ cells in the absence of TNF‐α or TNF‐R1. Consequently, reduced production of type I IFN (IFN‐I) can be detected during VSV infection and leads to a severe disease outcome. At the molecular level, the absence of TNF‐α expression impedes RelA translocation into the nuclei of CD169+ macrophages, which is necessary for early IFN‐I expression after viral infection 31. B cell activating factor (BAFF), another TNF superfamily‐related factor, is crucial for B cell development and humoral immunity in mice and humans 32. BAFF binds to B lymphocytes via one of three receptors: B cell maturation antigen (BCMA), transmembrane activator and calcium‐modulating cyclophilin ligand interactor (TACI) or BAFF‐R 33. BAFF specifically co‐stimulates B cell proliferation, promotes splenic B cell survival in vitro and is a key regulator of peripheral B cell populations. BAFF‐R signaling has been found to be crucial for B cell formation. In the absence of BAFF‐R signaling, the development of CD169+ macrophages is hampered. Transfer of BAFF‐R‐competent B cells partially restores the number of CD169+ macrophages 23. Therefore, BAFF‐R contributes to enforced viral replication during viral infection. Limited induction of a functional innate immune response by congenital IFN‐I after viral infection in Baffr−/− mice is due to reduced viral replication. Decreased virus presentation consequently results in greatly reduced anti‐viral adaptive immune responses 23. CEACAM1, in addition to TNF family‐related molecules, is a member of the carcinoembryonic antigen and immunoglobulin families and influences viral replication in a B cell‐dependent manner. CEACAM1 exerts limited influence on the proliferation of B cells but, rather, induces the survival of proliferating B cells via the Bruton’s tyrosine kinase/spleen tyrosine kinase/nuclear factor kappa‐light‐chain‐enhancer of activated B cells (BTK/Syk/NF‐κB) axis. During viral infection, CEACAM1‐deficient mice can barely induce an anti‐viral B cell response. In addition, because of the lack of CD169+ macrophages, CEACAM1‐deficient mice exhibit a reduction in innate immune activation. Therefore, CEACAM1‐deficient mice die after VSV infection 22. Molecules influencing viral replication The influence on viral replication in CD169+ macrophages and DCs is not exclusively restricted to the splenic architecture. Cell‐intrinsic molecules have a large effect on viral replication. Deubiquitinating enzymes are involved in several essential biological processes and pathways, such as transcriptional silencing, growth control and regulation of virus infection 34. The family of ubiquitin‐specific peptidase (USPs) contains more than 100 proteins, all of which share homolog sequences. Ubiquitin‐specific peptidase 18 (Usp18) is a ubiquitin‐specific peptidase that cleaves ISG15 from Isg15ylated proteins 35. In addition to this function, Usp18 binds directly to the Janus kinase (JAK) binding site of the IFN‐α receptor and blocks IFN‐I signaling. At a molecular level, Usp18 represses signal transducer and activator of transcription 1 (STAT‐1) and STAT‐2 heterodimer formation and IFN regulatory factor 9 (IRF9)‐dependent activation of IFN‐stimulated gene (ISG) transcription. CD169+ macrophages and DCs strongly express Usp18 14, 36. Therefore, over‐expression of Usp18 by CD169+ macrophages leads to IFN‐I unresponsiveness and enhanced viral replication within these cells (Fig. 1). The absence of Usp18 in mice leads to rapid death due to defective activation of innate and adaptive immune responses 7. In addition to Usp18, the paracaspase mucosa‐associated lymphoid tissue lymphoma translocation protein 1 (MALT1) is an important factor in viral replication. MALT1 is widely expressed on various immune cells and plays an important role in immunity, inflammation and cancer by fine‐tuning immune responses 37. After VSV infection, the absence of MALT1 inhibits RelA translocation into nuclei in a TNF‐dependent manner 31. Ablation of MALT1 increases the accumulation of RelB, which regulates RelA signaling by competitive binding to DNA 38. NF‐κB subunit RelA is crucial for regulating early IFN‐β expression after virus infection. Consequently, MALT1 deficiency reduces VSV replication and causes inefficient activation of the innate immune response; this lower immune response leads to a severe outcome of the infection 39. We consider that additional pathways influence viral replication in APCs. Additional screening must be performed so that researchers can identify more players that specifically promote viral replication within APCs in mice and humans. Acute cytopathic infection VSV has a tropism for neuronal tissue. Once VSV reaches the brain, its rapid propagation within the brain tissue usually leads to death. If this lethal outcome is to be avoided, the spread of VSV to the brain must be prevented by the innate and adaptive immune systems 40, 41. The two most important arms of VSV control are IFN‐I and neutralizing antibodies 40. IFN‐I reduces VSV replication in all organs to undetectable levels. Induction of neutralizing immunoglobulin (Ig)G allows complete control of all viral particles, even those that can escape the innate immune response. The absence of IFN‐I leads to overwhelming replication of VSV in several organs, including the liver, spleen, kidney and brain. Therefore, IFN‐I‐deficient mice die quickly after VSV infection 40, 42. In the absence of neutralizing antibodies, some viral particles manage to persist or even exhibit low‐level replication in the presence of IFN‐I 40. From such peripheral niches, virus can spread into the brain and cause death 40. Because of these specific features of VSV, rapid activation of the innate and adaptive immune systems is required for virus control. Viral replication within the lymph nodes and the spleen accelerates the activation of the innate and adaptive immune systems. This activation is mainly explained by increasing concentrations of viral RNA and proteins. Indeed, factors that promote viral replication within the lymphatic system (i.e. Usp18, BAFF, TNF, CEACAM1) accelerate the activation of the innate and adaptive immune systems after VSV infection and thereby promote rapid control of VSV in peripheral niches 14, 22, 23, 39. In this situation, the replication of VSV is beneficial or even essential during infection. VSV is a highly cytopathic virus that responds strongly to IFN‐I. Therefore, early viral replication in the spleen may be especially important. Whether viral replication in the spleen is also important for other virus infections remains to be determined. So far, it is known that several viruses, including murine cytomegalovirus (mCMV), mouse hepatitis virus, LCMV and vaccinia virus, replicate in the spleen, in addition to their typical tissue tropism 42-45. Vaccination Another example of the overall beneficial effect of viral replication occurs during vaccination with live vaccines. In fact, earlier studies indicated that replicating viruses elicit a stronger neutralizing antibody response than do inactivated virus particles 46, 47. In some cases, immunization with a dead virus is inefficient. A number of live attenuated vaccines provide sufficient protection, whereas inactivated vaccines fail to induce a neutralizing antibody response (e.g. rubella, measles, mumps, yellow fever, varicella, chickenpox) 48. Replication of active influenza virus leads to better immune activation than inactive virus does 49, 50. Besides the strength of the immune response, its quality can also be influenced by viral replication. For example, active rubella virus or poliovirus induces the production of IgA in addition to IgG 51-53. Therefore, immunization with active virus is beneficial in disrupting the infection chain, because IgA can neutralize poliovirus in the gut. After exposure to measles virus, prophylactic administration of live attenuated measles, mumps and rubella (MMR) vaccine leads to a beneficial outcome of the infection in some cases if the vaccine is administered within hours or days after exposure 54. In a separate study, we determined whether replication can contribute to anti‐viral immunity during challenge infection with LCMV. Interestingly, virus‐specific antibodies specifically prevent the replication of LCMV in peripheral organs except for the spleen and lymph nodes. The presence of virus‐specific antibodies allows rapid boosting of the existing and new immune components. This boosting results in rapid control of the virus, even in the case of infection with the highly persistent LCMV strain Docile. In this situation, virus control is dependent on CD169+ macrophages and on the priming of new virus‐specific CD8+ T cells 55. Exhaustion and viral replication in APCs During infection with the persistent LCMV strain Docile or clone 13, exhaustion of CD8+ T cells can prevent immunopathology 56. Although this process leads to viral persistence, it can be beneficial for survival. However, partial reduction of exhaustion by depletion of interleukin (IL)‐10, programmed cell death protein 1 (PD‐1) or CEACAM1 can result in virus control with limited adverse effects 57-60. Therefore, exhaustion is a double‐edged sword, with both benefits and disadvantages. Exhaustion is induced by overwhelming amounts of antigen presented together with exhaustive signals 61. IFN‐I strongly influences exhaustion 62. On one hand, IFN‐I reduces viral replication, and this reduction limits the number of antigens and thereby prevents exhaustion 63, 64. On the other hand, IFN‐I induces programmed death ligand 1 (PD‐L1) and thereby promotes exhaustion 62. We found that the absence of CD169+ macrophages or of Usp18 limits the induction of IFN‐I and thereby prevents the expression of PD‐L1 65. This factor also limits the exhaustion of CD8+ T cells; therefore, mice died because of severe immunopathology. Also, timewise‐limited depletion of CD169+ macrophages worsens the outcome of virus infection. Therefore, we conclude that, during infection with an exhaustive virus strain, limitation of replication in the spleen may prevent exhaustion of CD8+ T cells but increases the risk of severe immunopathology. Entry receptor and enforced viral replication To date, researchers have identified several factors that promote replication of virus in APCs. One important question that has not yet been addressed is the role of the entry receptor during the infection of APCs. One important bottleneck in the life cycle of a virus is entry into cells in which the virus can replicate. To enter cells viruses express glycoproteins, which allow cell entry via a specific receptor. Therefore, the specificity of the glycoprotein is a crucial factor determining organ‐specific replication. We used primarily LCMV for our studies. LCMV uses alpha‐dystroglycan as its main entry receptor. Because of the wide expression of alpha‐dystroglycan, LCMV can infect several types of cells in almost all organs 66-68. Because of this feature, LCMV can easily infect APCs. For viruses with a more specific entry receptor, the role of enforced viral replication can be questioned. For example, human poliovirus uses human CD155 as an entry receptor. Wild‐type (WT) mice, which do not express human CD155, can normally not be infected by poliovirus. However, mice expressing human CD155 are susceptible to poliovirus infection 69, 70. Although the role of Usp18 has not been studied during poliovirus infection, Freigang et al. have shown that murine APCs can be infected with poliovirus even in the absence of human CD155 2. This finding suggests that APCs can be infected via a receptor‐independent pathway. Hepatitis C virus (HCV) replicates almost exclusively in hepatocytes. The narrow tropism of HCV can be explained by its requirement for cell entry. The entry of HCV into hepatocytes is synchronized via various receptors: glycosaminoglycans (GAGs), the low‐density lipoprotein receptor (LDL‐R), tetraspanin CD81, the high‐density lipoprotein receptor scavenger receptor class B type I and two tight junction proteins, claudin‐1 (CLDN1) and occludin (OCLN) 71. The summation of these factors is essential for sufficient entry of HCV into cells and explains the liver specificity. Therefore, it is unthinkable that HCV could replicate in APCs. Nevertheless, HCV infection can induce sufficient immune activation (at least in some people). These facts clearly show that anti‐viral immune activation is possible even in the absence of viral contribution to enforced viral replication. Autoimmune disease and viral replication Autoimmune disease is often associated with infections. This evidence results from three facts: (1) autoimmune phenomena are described during a chronic virus infection 72-74; (2) evidence for a virus infection in patients with autoimmune disease 75, 76; and (3) experimental evidence (cross‐reactive epitopes, mouse models) 77-80. Well‐studied examples are multiple sclerosis and type I diabetes. Viruses can contribute to pathogenesis of autoimmune disease in several ways 77, 81. In type I diabetes, one well‐studied mechanism how viruses can induce diabetes is molecular mimicry 82, 83. In fact, Coxsackie virus, which infection is associated with type I diabetes, carries cross‐reactive antigens with beta islet cells 84. In line with these data, cross‐reactive antigens between a virus and the beta islet cells can be well modeled in the autoimmune diabetes model 85. Beside adaptive immune activation, local and systemic innate immune activation is usually occurring during a virus infection. Obviously, over‐activation of these innate signals can contribute to autoreactivity 86, 87. Especially in the presence of autoreactive CD8+ T cells, a strong induction by pattern recognition receptors (i.e. Toll‐like receptor (TLR)‐3, TLR‐7), which is usually occurring during virus infection, can turn such an autoimmunity into an autoimmune disease 88, 89. Further evidence for the role of innate immune activation in type I diabetes comes from genetic screens showing that a mutation in the genes for MDA5 (IFN‐induced helicase 1, IFIH1), which is one of the most important pattern recognition receptors for viruses, are associated with type I diabetes 90. Further, a rare mutation in IFIH1 can protect from type I diabetes 91, 92. In a recent study we asked whether enforced viral replication, can contribute to autoimmune diabetes. Therefore, we used the rat insulin promoter‐glycoprotein (RIP‐GP) model, which allows the study of virus‐induced diabetes. In this model, induction of diabetes depends on CD8+ T cells and interferons 85, 88, 93. In fact, deletion of Usp18, which limits replication of LCMV in lymphatic tissue, limited the innate immune activation and the induction of autoreactive CD8+ T cells 36. These data were confirmed using a second type I diabetes model, the RIP‐nucleoprotein (NP) model 36. Systemic lupus erythematosus (SLE) is a heterogeneous chronic autoimmune disease affecting diverse organ systems and tissues 94. Uncontrolled inflammatory response results in various clinical manifestations, including skin rash, arthritis, cardiovascular disease, hemolytic anemia and glomerulonephritis 95. While the pathogenesis of SLE is complex and complicated, the pivotal characteristic of SLE disease is an impaired removal of dead cells, with accumulation of apoptotic debris providing the main source for autoantigens 96. Exposition of the nuclear materials derived from apoptotic cells driven by APCs, such as myeloid DCs and macrophages, prime autoreactive T cells 97. Subsequently, activated autoreactive T cells interact with B cells causing production of autoantibodies which form immune complexes. Finally, deposits of immune complexes on tissues activate the complement system, resulting in severe inflammation and tissue damage. Additionally, plasmacytoid DCs triggered by immune complexes up‐regulate the release of type I IFN‐inducing IFN signature as a major pathological indicator of SLE 98. Several environmental factors were suggested to contribute toward the pathogenesis of SLE. Among them, viruses seem to be involved in the establishment of SLE disease in genetically susceptible individuals 99, 100. There is growing evidence suggesting a relationship between infection with Epstein–Barr virus (EBV) and SLE. Recent studies revealed an increase prevalence of EBV infection as well as higher titers of anti‐EBV‐antibodies in SLE patients compared to healthy subjects 101-103. Furthermore, structural similarities between components of EBV protein, in particular Epstein–Barr nuclear antigen 1 (EBNA1) and autoantigens, were described in SLE, suggesting molecular mimicry as a potential mechanism to promote production of cross‐reactive autoantibodies 104, 105. Enhanced expression of type I interferon genes due to viral infection may also mediate IFN signature, allowing the abnormal immune response and as a consequence the onset of SLE. Loss of tolerance to self‐antigens is thought to be additionally responsible for the induction of SLE 106, 107. Fcgr2b contributes to lack of tolerance in SLE 108. Therefore, mice lacking Fcgr2b develop a lupus‐like syndrome 109. In fact, we and others found that during age Fcgr2b knock‐out (KO) mice developed anti‐dsDNA antibodies 110. Almost all mice older than 21 weeks showed elevated anti‐nuclear and anti‐dsDNA autoantibodies. Despite elevated levels of autoantibodies, we did not observe strong disease onset in Fcgr2b KO mice in the absence of further stimuli. However, upon virus infection with LCMV, mice died quickly during infection 110. The death of the mice depended on IFN‐γ and CD8+ T cells. Mechanistically, IFN‐γ derived from virus‐specific CD8+ T cells strongly increased expression of Fcgr1 and Fcgr3 on myeloid cells and thereby made self‐reactive antibodies visible to these immune cells. Together, these data clearly show that activation of virus‐specific CD8+ T cells can also lead to serious harm to the body if they are not cross‐reactive for an autoantigen. Especially in a systemic autoimmune disease, systemic activation of virus‐specific CD8+ T cells will strongly influence myeloid cells, but perhaps also other adaptive immune cells, which then impact on the disease onset. Concluding remarks In some circumstances, replication in immunological niches can be beneficial for the control of virus; however, it carries the risk to induce autoimmune disease. More studies are necessary before final conclusions can be reached. Disclosures The authors declare no competing interests. References Citing Literature Number of times cited according to CrossRef: 1 U. Christen, Pathogen infection and autoimmune disease, Clinical & Experimental Immunology, 195, 1, (10-14), (2018). Wiley Online Library

The disease systemic lupus erythematosus (SLE)—marked by the attack on joints, skin, and kidneys by the body's immune system—is linked to an abnormal mix of bacteria in the gut. This is according to a new study led by scientists at NYU School of Medicine.

Pathogens in autoimmune diseases