Defective Apoptosis

Defective apoptosis can result in the generation of neoepitopes.¹³ Proteolytic cleavage of lupus-associated autoantigens, like poly (ADP-ribose) polymerase and a catalytic subunit of DNA-dependent protein kinase (DNA-PKCs), has been shown to disturb homeostasis and cause increased apoptosis. As a result of nuclear fragmentation and membrane blebbing in apoptosis, autoantigens that are targeted in SLE are reorganized and transported to cell surfaces.¹⁷ Secondary necrosis can also be an additional source of proteolytically modified forms of specific autoantigens.¹⁸ For example, during the initial apoptotic stages, several autoantigens, including poly ADP-ribose, are cleaved into apoptosis fragments. The apoptotic cells then undergo secondary necrosis in the absence of phagocytosis with additional modifications of autoantigens.¹⁸ A misguided immune response to these modified nuclear and cytoplasmic antigens is believed to be a major mechanism underlying autoantibody production in SLE.

Impaired Removal of Apoptotic Cells

Impaired removal of apoptotic cells could contribute to an overload of autoantigens (particularly nucleosomes) in circulation or in target tissues that could become available to initiate an autoimmune response. Nucleosomes are formed during apoptosis by organized cleavage of chromatin. These nucleosomes together with other autoantigens cluster in apoptotic bodies at the surfaces of apoptotic cells. Systemic release of these autoantigens is normally prevented by swift removal of apoptotic cells. However, if excessive apoptosis exceeds the rate of removal of apoptotic bodies, nucleosomes are released. A number of studies have identified abnormalities that lead to impairment of apoptotic debris removal in patients with SLE and in mouse models. These include deficiencies in complement components, particularly C1q, C2, and C4¹⁹ as well as macrophage proteins that are pertinent to clearance of debris, including scavenger receptor A (SR-A), macrophage receptor and collagenous structure (MARCO),²⁰ and mer tyrosine kinase.²¹

Mutations

Mutations in self-antigens, which may create a neoepitope, might trigger autoimmunity. For example, in a complementary DNA (cDNA) library made from peripheral blood lymphocytes (PBLs) of a patient with primary Sjögren syndrome, one study identified a deletion of an (A)-residue in a cDNA encoding for the nuclear autoantigen La (SSB). This leads to a frame shift mutation in one of the major autoepitope regions of the La antigen.^22,23 The modified La peptide shared homology with (1) La protein itself and (2) a series of DNA-binding proteins, including other autoantigens, and viral proteins such as topoisomerase I, RNA-dependent RNA polymerase of influenza virus, and reverse transcriptase. The mutant La peptide represents a putative neoepitope that could be involved in triggering of the autoimmune response.

Genetic Polymorphisms

Genetic polymorphisms may create autoantigens.²⁴ Analysis of sequence variability has revealed significantly more single-nucleotide polymorphisms (SNPs) within coding regions of known human autoantigens (n = 348) than of other human genes (n = 14,881). Autoantigens had 7.2 SNPs per gene, compared with 3.6 SNPs per control gene. As an example, human Ro52, a major autoantigen in rheumatic diseases, contains two synonymous and three nonsynonymous SNPs, and one of the nonsynonymous SNPs is located in the central immunodominant region of the autoantigen.²⁵ Further, an intronic SNP that leads to aberrant splicing of Ro52 messenger RNA (mRNA), resulting in the generation of a shortened version of the Ro52 protein, is strongly associated with anti-Ro52 autoantibodies in primary Sjögren syndrome.²⁶

Alternative Splicing

Alternative splicing can create autoantigens.¹⁶ A bioinformatics analysis revealed alternative splicing in 100% transcripts of 45 randomly selected autoantigens, which is significantly higher than the approximately 42% rate of alternative splicing observed in 9554 randomly selected human gene transcripts. Within the isoform-specific regions of the autoantigens, 92% and 88% encoded major histocompatibility complex (MHC) class I– and class II–restricted T-cell antigen epitopes, respectively, and 70% encoded antibody-binding domains. Furthermore, 80% of the autoantigen transcripts underwent noncanonical alternative splicing, a rate that is also significantly higher than the less than 1% rate observed in randomly selected gene transcripts.

Posttranslational Modifications

Posttranslational modifications in a protein could also act as a means to promote autoreactivity.^6,7,15 Between 50% and 90% of the proteins in the human body acquire posttranslational modification. Many of these modifications are necessary for the biological functions of proteins. Some posttranslational modifications, however, can create new self-antigens by altering immunologic processing and presentation. Because these modifications occur after the lymphocyte has undergone negative selection, the existing B and T lymphocytes can recognize the modified antigens, thus causing tolerance breakdown. For example, the spontaneous conversion of asparagine or aspartic acid residues to isoaspartyl residue renders cytochrome c and snRNP D peptides immunogenic in murine models of SLE. Mice develop T-cell responses to the isoaspartic acid–containing peptides but not to the native aspartic acid–containing peptides. Autoantibodies in these mice, however, recognize both the isoaspartic peptides and the native aspartic acid peptides. Isoaspartic acid residues have also been detected in histone H2B, a common autoantigen in spontaneous and drug-induced lupus.²⁷ In other examples of a likely role of posttranslational modification in autoimmunity, patients with SLE have been found to have autoantibodies against the C-terminus of snRNP that contains symmetrical dimethyl arginines,²⁸ and phosphorylated serine/arginine–rich residues of the SR protein (a family of pre-mRNA splicing factors). Interestingly, some autoantibodies are directed at dephosphorylated SR proteins that normally would exist in a phosphorylated state.²⁹

Altered Antigen Processing

Altered antigen processing can lead to generation of new autoantigens for which the immune system is not tolerized. In xenobiotic models of lupuslike autoimmunity, cell death following exposure to autoimmunity-inducing agents leads to generation of novel protein fragments that may activate self-reactive T lymphocytes.³⁰ During apoptosis, interaction of several autoantigens with granzyme B has been shown to generate unique protein fragments that are not observed during any other form of cell death. Interestingly, nonautoantigens are either not cleaved by granzyme B or are cleaved to generate fragments identical to those formed in other forms of apoptosis. Therefore the ability of granzyme B to generate unique fragments appears to be an exclusive property of autoantigens.^31,32

Molecular Mimicry

Molecular mimicry has been proposed to explain the role of microbial antigens in inducing and/or exacerbating autoimmune diseases. Association between the development of SLE and viruses such as Epstein-Barr virus (EBV), Coxsackie virus, and retroviruses like human T-lymphocyte virus (HTLV) has been described.³³ For example, analysis of autoantibody responses in patients with SLE prior to the onset of clinical disease led to identification of an initial autoantigenic epitope that appears in some patients who have antibodies to 60-kDa Ro antigen. This initial epitope cross-reacts with a peptide from the latent viral protein Epstein-Barr virus nuclear antigen 1 (EBNA-1). Animals immunized either with the initial epitope of 60-kDa Ro or with the cross-reactive EBNA-1 epitope progressively develop autoantibodies binding to multiple epitopes of Ro and spliceosomal autoantigens. The immunized animals eventually demonstrate clinical symptoms of lupus, thus providing a strong evidence for association of EBV infection and development of SLE.³⁴ One study has demonstrated molecular mimicry at the T-cell epitope level between lupus-associated autoantigen Sm-D and microbial peptides (Table 21-1). The researchers found that distinct autoreactive T-cell clones were activated by different microbial peptides, suggesting a role for molecular mimicry at the T-cell epitope level for activation of autoantibody augmenting autoreactive T cells.⁸

Defective Sensing and Uptake of Autoantigen

Accumulating evidence suggests that under autoimmune conditions, DNA fragments alone are able to induce signaling cascades that promote inflammation. Pathways through which self-DNA is able to induce proinflammatory reactions are distinct from those activated by microbial nucleic acids.¹³ The dsDNA-containing immune complexes undergo endocytosis after engaging the B-cell receptor (BCR) on B cells or Fc receptors (FcRs) on dendritic cells (DCs), macrophages, and glomerular cells. Additionally, dsDNA can be internalized after binding to LL-37 (cathelicidin), or through the HMGB1-RAGE (receptor for advanced glycation end-products) pathway. These routes result in localization of DNA to endosomes. The dsDNA-sensing pathways activated differ according to the structure of the DNA. CpG-rich dsDNA activates Toll-like receptor 9 (TLR9), whereas AT-rich dsDNA signals through DAI (DNA-dependent activator of interferons) or RNA polymerase III. These signaling pathways all lead to production of type I interferon (IFN) and inflammation. A fourth dsDNA-sensing pathway involves the AIM2 inflammasome and results in activation of interleukin-1 beta (IL-1β) and induction of pyroptosis. It is logical to hypothesize that defects in these pathways may trigger autoantigenicity of DNA,¹³ although little is known about the exact role these pathways play in the pathogenesis of SLE.

Chronic Inflammation as a Trigger of Autoantigenesis

Exposure to naturally occurring hydrocarbon oils, such as the medium-length alkane 2,6,10,14-tetramethyl pentadecane (TMPD, also known as pristane), is associated with the development of chronic inflammation and a variety of pathologic findings in humans and animal models.^9,35 Depending on the genetic background, persistent inflammation in otherwise normal strains of mice eventuates in a cascade of events leading to a plethora of autoimmune manifestations, including glomerulonephritis, arthritis, pulmonary vasculitis, and autoantibodies against a variety of nuclear and cytoplasmic antigens, which mimics human SLE syndrome more closely than the genetically susceptible strains of mice.^9–11

Data suggest that different autoantibody subsets and organ injury are mediated through different pathways, and both innate and adaptive immune responses participate in the development of full lupuslike syndrome in TMPD-injected mice.¹² The initial response to TMPD is orchestrated by major components of the innate immune system. It starts with the infiltration of neutrophils and Ly6C^hi inflammatory monocytes into the peritoneal cavity, which lasts for several months.³⁵ Type I IFN (IFN-I) production downstream of TLR7 signaling and CCR2 (chemokine [C-C motif] receptor 2) plays a role in the influx of monocytes, whereas IL-1α and CXCL5 (chemokine [C-X-C motif] ligand 5) play a role in neutrophil recruitment to the peritoneal cavity. The adaptor molecules MyD88, IL-1R–associated kinase 4 (IRAK-4), IRAK-1, and IRAK-2 play a role in the recruitment of both monocytes and neutrophils. Deficiency of IL-6, TLR9, and TLR4 attenuate organ injury and production of anti-dsDNA and/or anti-RNP autoantibodies. Although the exact cascade of events that lead to autoantigenicity of lupus autoantigens is not known, studies in this model suggest a role for chronic inflammation in autoantigenesis.

Induction of Effector T cells

Autoantigen-specific T cells can contribute to pathogenesis of SLE by helping B cells produce autoantibodies or by directly infiltrating the tissues. Ample evidence also supports the requirement of T-cell help for pathogenic autoantibody production.^36,37 This help is provided by T cells that react with peptides derived from various autoantigens.^37–43 One study has shown that MRL-lpr mice that have no secreted immunoglobulin (Ig) develop spontaneous T-cell activation and some disease,⁴⁴ suggesting a direct, antibody-independent role for T cells in the development of SLE-like disease. “Autoantigen-specific” T cells, however, have not yet been demonstrated in the target tissues of humans and mice with lupus. In one study, the use of in situ tetramer staining has allowed identification of human antigen–specific T cells in the affected organs, such as the pancreases of patients with autoimmune diabetes.⁵ Both single and multiple T-cell autoreactivities were detected within individual islets in a subset of patients up to 8 years after clinical diagnosis. Use of such technology has potential to identify tissue-infiltrating autoantigen-specific T cells that can be true targets for treatment in patients with SLE.

Reduced Activation of Regulatory, Inhibitory, or Suppressor T Cells

Immunization with self-Ig peptides induces T cells that can suppress anti-DNA antibodies in healthy strains of mice.⁴⁵ Induction of such “self-reactive” regulatory T cells is impaired in lupus-prone mice that mount mostly pathogenic T-helper (Th) cell response.

Activation of Toll-Like Receptors

Activation of TLRs by autoantigens can amplify the autoimmune response by activating the innate immune component. Chromatin-containing CpG motif–rich DNA or RNP antigens containing dsRNA can potentially trigger lupuslike autoimmune responses by providing accessory signals through TLR9 on DCs, macrophages, or B cells, or through TLR3 on DCs.^46,47 Further, immune complexes containing IgG bound to chromatin can activate murine DCs through both TLR9-dependent and TLR9-independent pathways,⁴⁸ a feature that may affect autoimmune responses. Indeed, TLR9 deficiency specifically reduces the generation of anti-dsDNA and antichromatin autoantibodies in MRL-lpr mice.⁴⁹ Viral dsRNA can also activate DCs via TLR3 to induce the production of IFN-I and cytokines associated with disease activity in SLE. Furthermore, TLR3 expression is increased in infiltrating APCs as well as in glomerular mesangial cells in kidney sections of MRL-lpr mice.⁵⁰

HMGB1, a nuclear DNA-binding protein, can trigger a proinflammatory response by interacting with receptors TLR2, TLR4, and RAGE on macrophages and DCs.⁵¹ HMGB1 can also stimulate innate immune responses by acting as a universal sentinel for nucleic acids and facilitating their interaction with a number of receptors, including TLR3, TLR7, and TLR9.⁵² Importantly, in patients with SLE, serum HMGB1 and anti-HMGB1 autoantibody concentrations are elevated and correlate with disease activity.⁵³ A number of reports have linked the activation of TLRs with autoimmune diseases primarily through their ability to drive the induction of autoreactive T and B cells.⁵⁴ Furthermore, there is evidence that TLR activation can block T-regulatory (Treg) cell responses, thereby breaking tolerance to self-antigens.

Autoantigens as Chemoattractants

Autoantigens may serve as chemoattractants that recruit innate immune cells to sites of tissue damage.⁵⁵ A variety of autoantigens has been shown to induce leukocyte migration by interacting with various chemoattractant Gi protein–coupled receptors (GiPCRs). For instance, myositis autoantigen, histidyl-tRNA synthetase, are chemotactic for the CCR5 and CCR3, thereby recruiting T lymphocytes and immature DCs. Fibrillarin (U3-RNP) and topoisomerase I, which are autoantigens associated with scleroderma, have been shown to serve as chemoattractants for monocytes. In some cases, such as in SLE, a complex of two autoantigens has been found to be chemotactic for immature DCs. Thus, autoantigens not only may attract immune cells to a given tissue but also can activate B cells,⁵⁶ DCs,^57,58 and neutrophils⁵⁹ via their ability to interact with cell membrane receptors.

Altered Recognition of Autoantigens

Altered recognition of autoantigens in endomembrane traffic might elicit autoimmunity.⁶⁰ The RNA transcription termination factor La, a frequent target of Sjögren autoantibodies, appears in the acinar cell cytoplasm and plasma membranes during viral infection and after in vitro exposure to cytokines. The endomembrane compartments where proteolysis occurs contain La, galactosyltransferase, cathepsin B, and cathepsin D. MHC class II molecules cycle through this compartment. This traffic may permit trilateral interactions in which B cells recognize autoantigens at the surface membranes, CD4⁺ T cells recognize peptides presented by MHC II, the B cells provide accessory signals to CD4⁺ T cells, and CD4⁺ T cells provide cytokines that activate B cells.

Autoantigen Ro52

Autoantigen Ro52 is an E3 ligase that may regulate proliferation and cell death. Increased apoptosis in patients with SLE may result in greater expression of intracellular autoantigen Ro52, which may mediate ubiquitination in survival genes induced during CD40-mediated activation.⁶¹ Ro52 may also enhance functions of genes mediating apoptosis and cell death by relieving them from endogenous repression. Intriguingly, Ro52-deficient mice develop autoimmunity, suggesting that this E3 ligase may also act as a negative regulator.

Studies in Animal Models

Work in the late 1980s suggested that autoantibody production in humans and mice with SLE is antigen-driven and depends on Th cells that are mostly CD4⁺.^36,83–85 To identify the nature and specificity of such autoreactive Th cells, several laboratories have used diverse approaches, including T-cell pepscanning of candidate autoantigens, isolating autoreactive T-cell clones and deducing potential autoantigens, screening phage display libraries, and eluting naturally processed self-peptides from MHC class II molecules. These approaches have led to identification of epitopes that activate autoreactive Th cells in humans and mice with lupus and modulate disease in lupus mice (see Table 21-1).

Identification of Self-Epitopes in Human SLE

Human T cells reactive with several lupus autoantigens, including DNA-histones, the snRNP antigenic proteins Sm-B, Sm-D, U1-70kD, and U1-A, and heterogeneous RNP (hnRNP) A2, have been isolated from the peripheral blood of patients with SLE.¹⁰² Datta’s group first described T-cell lines from patients with SLE, which augmented the production of IgG anti-DNA antibodies ex vivo¹⁰³ and antihistone.¹⁰⁴ These autoantibody-promoting T cells are usually CD4⁺ T cells that use restricted CDR3 characteristic of antigen selection.¹⁰⁴ To identify antigenic epitopes for these T cells, they used 154 peptides spanning the entire length of core histones of nucleosomes to stimulate an anti-DNA antibody–inducing Th clone, CD4⁺ T cell lines, and freshly isolated T cells in peripheral blood mononuclear cells (PBMCs) from 23 patients with SLE.¹⁰⁵ In contrast to normal T cells, lupus T cells responded vigorously to certain histone peptides, irrespective of the patient’s disease status (see Table 21-1). Interestingly, most of the peptides that activated human T cells from patients with lupus were previously identified as T-cell epitopes in lupus-prone mice (see Table 21-1). Several additional epitopes, including peptides 34 through 48 of H2A, 91 through 105 and 100 through 114 of H3, and 49 through 63 of H4, were also found to activate human T cells from patients with lupus. Most of these sequences are located in the regions of histones that are accessible at the surfaces of nucleosomes and that contain B-cell epitopes targeted by lupus autoantibodies. Importantly, most T-cell epitopes have multiple HLA-DR binding motifs, that is, they are promiscuous with regard to their binding to HLA molecules. Thus, peptides containing these epitopes could potentially be used to treat many patients, obviating the development of individualized therapy.

To determine whether Ig-derived peptides activate T cells from patients with SLE, Williams cultured PBMCs from 28 patients with lupus and 13 healthy individuals with selected 16-mer peptides from two anti-DNA autoantibodies, B3 and 9G4.¹⁰⁶ Three of the 13 healthy individuals (23%) versus 17 of the 28 patients with SLE (61%) had T cells that proliferated in response to at least one V region peptide. In another study, Linker-Israeli cultured 12-mer overlapping peptides from the V_H of two anti-dsDNA antibodies, B3 and F51, with PBMCs from patients with SLE, their first-degree relatives, or unrelated healthy individuals.¹⁰⁷ The expressions of the early T-cell activation markers CD25 and CD69 and of cytokines were determined by flow cytometry. Patients with SLE had significantly increased T-cell activation markers and IL-4–secreting cells than either first-degree relatives or unrelated controls. A subsequent study by these investigators in a larger cohort (31 patients and 20 matched healthy controls) analyzed cytokine release by PBMCs in response to seven peptides from the CDR1/FR2 to CDR2/FR3 V_H regions of human anti-DNA monoclonal antibodies.¹⁰⁸ PBMCs from significantly higher proportions of patients with SLE than controls responded to V_H peptides by generating IFN-γ and IL-10. Three peptides were more stimulatory in patients with SLE than in controls. There was a skewing of the immune response to Th2 bias as the disease progressed from early to later stage. Although none of the peptides was restricted by any particular MHC class II allele, among “responders” there was greater prevalence of HLA-DQB1*0201 and/or DRB1*0301, alleles known to predispose to SLE. Thus, responses to some V_H peptides are more common in SLE and vary with disease duration. Increased peptide presentation by SLE-predisposing HLA molecules might permit brisker increased T-cell responses to autoantibody peptides, thus increasing risk for disease.

Guided by observations in the 16/6 Id murine model described earlier, a previous study examined immune responses of patients with SLE to peptides encompassing complementarity-determining regions (CDRs) of a monoclonal anti-DNA antibody with a 16/6 Id.¹⁰⁹ In contrast to the preceding data showing increased responses to anti-DNA–derived peptides,^106–108 this group found that peripheral blood lymphocytes (PBLs) from significantly fewer patients (37%) than controls (59%) proliferated in response to one of the anti-DNA peptides.¹¹⁰ A subsequent study by the same group reported in vitro proliferation of PBLs from 24 of the 62 patients with SLE tested after stimulation with the human 16/6 Id.⁴ Interestingly, peptides from both the human and murine anti-DNA autoantibodies specifically inhibited the 16/6 Id–induced proliferation and IL-2 production. The latter inhibitions correlated with increased production of TGF-β. Findings of this study suggested that certain anti-DNA peptides may downregulate autoreactive T-cell responses in patients with SLE. Indeed, treatment of severe combined immunodeficient (SCID) mice engrafted with PBLs of patients with SLE by repeated intraperitoneal administration of a human monoclonal anti-DNA autoantibody peptide (hCDR1) suppressed human anti-dsDNA antibodies but not anti–tetanus toxoid antibodies.¹¹¹ Such treatment also reduced proteinuria and renal deposits of human IgG and murine complement C3 in the engrafted SCID mice.

Human T cells reactive with various snRNP antigens, including Sm-B, Sm-D, U1-70kD, and U1-A, have been identified and characterized.¹⁰² Subsequent studies on snRNP-reactive human T-cell clones showed that they typically exhibit T-cell receptor alpha/beta–positive (TCRαβ⁺) CD4⁺ CD45RO⁺ phenotype, recognize antigen in the context of HLA-DR, and produce substantial amounts of IFN-γ, moderate quantities of IL-2, and variable amounts of IL-4 and IL-10.¹¹² Further, these cells can also provide help for relevant autoantibody production in vitro.¹¹³ Talken established two sets of T-cell clones from patients with connective tissue diseases: one set reacted with Sm-B autoantigen and the others recognized U1-70kD polypeptide.^114,115 Both sets of T-cell clones had a highly restricted TCR CDR3 β- or α-chain gene usage, respectively. Further analysis revealed that the Sm-B–reactive T-cell clones recognized a peptide, Sm-B2_48-96, in the context of HLA-DR. Subsequent T-cell epitope mapping studies of human T-cell clones reactive with the snRNPs U1-70kD, Sm-B, and Sm-D revealed that there are limited T-cell epitopes on these proteins and that almost all reside within functional regions of the protein—either within the Sm motifs for Sm-B and Sm-D or within the RNA binding domain for U1-70kD.

In another study, synthetic 15-mer overlapping peptides spanning the entire La sequence were cultured with PBMCs from patients with SLE and controls with a goal to identify T-cell epitopes in the La antigen. The researchers found a significant, albeit low-level, T-cell proliferative response to a peptide (La 49-63) in HLA-DR3⁺ patients or healthy controls.¹¹⁶ This study highlights difficulties in identifying relevant pathogenic T-cell epitopes using PBMC-based T-cell proliferation readout experiments. The findings further suggest that the presence of self peptide–reactive T cells in the peripheral blood of healthy individuals is not uncommon. It is unclear why these Th cells promote autoantibody production only in certain individuals.

As described previously, Muller and colleagues identified a CD4⁺ T-cell epitope in peptide sequence encompassing residues 131 through 151 of the spliceosomal U1-70K snRNP protein (RIHMVYSKRSGKPRGYAFIEY) and its analog phosphorylated at Ser¹⁴⁰ (called P140; RIHMVYSKRS(P)GKPRGYAFIEY) in MRL-lpr and (NZB/NZW)F1 mice.^40,117 Importantly, administration of the phosphorylated peptide P140 ameliorates the clinical manifestations of treated MRL-lpr mice.^40,99 Because this peptide sequence, which is completely conserved in the mouse and human U1-70K protein, contains an RNA-binding motif often targeted by antibodies from patients with lupus and mice, the group investigated these peptides as potential candidates for the treatment of patients with lupus.¹¹⁷ Binding assays with soluble HLA class II molecules and molecular modeling experiments indicate that both peptides behave as promiscuous epitopes and bind to a large panel of human DR molecules. In contrast to normal T cells and T cells from patients with non-SLE autoimmune disease, PBMCs and/or CD4⁺ T cells from 40% of patients with SLE proliferate in response to peptide 131-151. Interestingly, the phosphorylated analog peptide P140 prevents CD4⁺ T-cell proliferation but not secretion of regulatory cytokine IL-10. Thus, P140 can serve as a “universal” immunomodulatory T-cell epitope.

In summary, patients with SLE have circulating T cells that recognize diverse sets of autoantigenic peptides, which include core histone peptides, Sm-B, U1-70kD, and peptides derived from the V region of autoantibodies. The significance of these T cells in the pathogenesis of SLE remains to be fully understood.