Systemic lupus (SLE) is an autoimmune disease characterized by several autoantibodies directed against intracellular antigens. Kidney involvement is frequent and indeed constitutes one of the primary causes of morbidity and mortality. It is generally agreed that antibodies are the principal agents at work in lupus nephritis, forming immune deposits by different mechanisms, more than one of which may be involved. Their central role in causing renal damage has been convincingly demonstrated in a genetically manipulated autoimmune mouse strain. MRL/ lpr-lpr mice lacking Ig heavy chain Jh genes, and therefore lacking B cells and autoantibodies, do not develop glomerular, tubular, or interstitial damage, even in the presence of the lpr mutation (Shlomchik et al, 1994). While not excluding the contribution of T cells, soluble mediators, and B cell functions other than antibody formation (Chan et al., 1999; Tipping and Holdsworth, 2003), antibodies have been demonstrated to be a key factor in the pathogenesis of nephritis. Three mechanisms may lead to their targeting to the kidney: deposition of CICs, binding of antibodies to antigens previously complexed to glomerular structures, and binding of antibodies to intrinsic glomerular antigens.
Given the close relationship between lupus and anti-DNA antibodies, the latter immunoglobulins have been proposed to cause renal damage by the three mechanisms described above. As far as CIC deposition is concerned, experimental studies have yielded conflicting results: circulating DNA seemed to be rapidly cleared from the circulation (Emlen and Mannik, 1984), and, intravenous administration of preformed DNA-anti-DNA ICs to normal mice led only to a transient deposition in glomeruli without inflammation. Moreover, convincing evidence for the presence of DNA-anti-DNA complexes in the circulation of SLE patients is still lacking. Therefore, other pathogenetic factors in addition to CIC deposition are presumably at work. In fact, the DNA that circulates in the plasma of SLE patients is contained in nucleosomes (Rumore and Steinman, 1990), particles composed of DNA linked to histones and nonhistonic proteins. Since nucleosomes are exposed in clusters together with other internal antigens on the cell surface during apoptosis, they may act as antigenic stimuli. The histone components have a cationic charge, which allows them to bind to the glomerulus by interacting with its negatively charged components, mainly heparan sulphate in the basement membrane. Thus, nucleosomes can mediate the binding of circulating anti-DNA antibodies to glomeruli (Berden et al., 1999). The results of a number of studies in both murine and human lupus support this scenario. Nucleosomes have been found in immune deposits of nephritic kidneys in human SLE (van Bruggen et al., 1997a), and there is corroborative evidence for a link between antinucleosome antibodies and lupus nephritis. Antinucleosome antibodies have been eluted from nephritic kidneys of lupus-prone mice, where they are produced in the early disease stages, even before anti-DNA antibodies are produced (Amoura et al., 1994). When hybridomas synthesizing antinuclear antibodies of different specificities were inoculated to normal mice, anti-DNA and antinucleosomal antibodies, but not antihistone antibodies, deposited in the GBM (van Bruggen et al., 1997b).
In addition to DNA, anti-DNA antibodies can cross-react with a large number of intracellular and extracellular molecules, and this polyreactivity could play an important role in conferring a pathogenetic potential, allowing them to bind intrinsic renal antigens (Madaio and Shlomchik, 1996). This was demonstrated by the presence of anti-DNA antibodies in the eluates from nephritic kidneys in human SLE and in murine models. The immunoglobulins deposited in the kidney consist mainly of IgG, often able to bind several targets such as DNA, polynu-cleotides, ribonucleoproteins, and gp70 (Pankewycz et al., 1987). Similarly, elu-ates obtained from nephritic SLE kidneys were found to contain polyreactive antibodies. In addition to anti-DNA antibodies, some of the specificities highly enriched in the kidneys were anti-C1q and anti-Sm antibodies (Mannik et al., 2003).
The ability of anti-DNA antibodies to bind intrinsic renal antigens has been extensively investigated. In early studies, polyclonal or monoclonal anti-DNA antibodies were infused into isolated kidneys or injected into normal mice
(Madaio et al., 1987; Raz et al., 1989), resulting in glomerular deposition, and, in some cases, in renal damage (Ehrenstein et al., 1995). In fact, not all SLE patients with high anti-DNA antibody titers have clinically significant nephritis, while SLE patients low titers of anti-DNA antibodies may exhibit overt nephritis. Therefore it appears likely that some but not all anti-DNA antibodies have a nephritogenic potential.
More recently, a number of glomerular antigens targeted by anti-DNA antibodies have been isolated and characterized. Laminin, a constituent of the GBM, is an extracellular protein composed of several chains that regulates the kidney's filtering efficiency. Anti-laminin antibodies also display anti-DNA binding activity (Foster et al., 1993) and anti-DNA antibodies cross-react with laminin (Sabbaga et al., 1989). Spontaneously produced autoantibodies directed toward laminin were found in MRL/lpr-lpr mice (Foster et al, 1993) and in the Graft versus Host Disease (GvHD), another experimental model of lupus (Peutz-Kootstra et al., 2000). In GvHD mice, the glomerular expression of laminin chains alters as the disease progresses, and concomitant modifications of the fine specificity of antilaminin antibodies have been observed. Other proteins located in the kidney may constitute antigenic targets of nephritogenic antibodies in lupus nephritis. Murine monoclonal anti-DNA antibodies with nephritic potential (Mostoslavsky et al., 2001) demonstrate binding ability against a-actinin, a structural protein expressed in the cytoplasm but also on the cell membrane of various glomerular cells, especially podocytes (Deocharan et al., 2002). Furthermore, it has been suggested that cross-reactivity with a-actinin is a property of anti-DNA antibodies isolated from SLE patients with active nephritis (Mason et al., 2004). Since mice lacking a-actinin 4 (the isoform most frequently expressed in the kidney) develop severe nephritis (Kos et al., 2003) and because high titers of anti-a-actinin antibodies are present in the sera and in renal eluates of mice with nephritis (Deocharan et al., 2002), it is conceivable that alterations of this protein may lead to nephritis. In addition to a-actinin and laminin, other intrinsic glomerular antigens may be recognized by anti-DNA antibodies, but further studies are needed to ascertain the role of these antibodies in the pathogenesis of lupus nephritis.
Once the immune deposits are formed in the glomeruli, autoantibodies, for example, anti-C1q antibodies, may enhance the IC renal disease (Trouw et al., 2004). It is known that anti-C1q antibodies are associated with active SLE and in particular with active nephritis (Siegert et al., 1991). However, administration of anti-C1q antibodies is followed by glomerular deposition, but neither significant albuminuria nor proliferative lesions are detected in the injected animals (Trouw et al., 2003b). Experiments in Rag2-~ mice confirm that autologous immunoglobulins are needed for the renal deposition of anti-C1q antibodies, indicating that they can form in situ ICs reacting with preexisting immune deposits (Trouw et al., 2003a). Taken together, these data suggest that while anti-dsDNA antibodies, which are polyreactive and able to bind renal or planted antigens, may directly initiate the formation of renal immune deposits, anti-C1q antibodies may amplify a preexisting damage. It should be stressed that anti-C1q antibodies may also contribute to the development of nephritis by lowering the levels of C1q and thus impairing its protective role.
Pathogenic antibodies may cause tissue damage by activating complement, by recruiting effector cells via FcRs or, more directly, by penetrating into living cells. First, the complement activation induced by antigen-antibody complexes may impair glomerular structures through the membrane attack complex, and by means of mediators recruited by chemotactic components produced during activation of the cascade. While complement is a mediator of inflammation, it also plays a role in IC clearance by binding to CR1 receptors, as was demonstrated in a study using preformed DNA-anti-DNA ICs (Craig et al., 2000). Therefore, any impairment of this clearance ability—whether acquired or genetically determined— can amplify the pathogenetic potential of ICs. Likewise, a deficiency in one of the components of the classical complement pathway constitutes an important susceptibility factor for SLE. Recently, a more extended protective role has been postulated for the complement system, because C1q plays a fundamental role in the clearance of apoptotic cells, a mechanism that may explain the occurrence of glomerulonephritis in C1q-deficient mice (Botto and Walport, 2002).
Second, the binding of antibodies to FcRs located on various cell types may have a wide range of consequences, resulting in either an effector function for human FcyRI, FcyRIIa, and FcyRIII or in a downmodulation of the immune response, when FcyRIIb is recruited. The effector functions include antibody-dependent cell cytotoxicity, phagocytosis and release of inflammatory mediators, and probably also internalization of antigens for processing (Takai, 2002; Reefman et al., 2003). The ability of preformed ICs to elicit inflammation in C3-and C4-deficient mice supports the theory that FcRs are key mediators of the inflammatory reaction (Sylvestre et al., 1996). The importance of FcRs in nephritis has been clearly demonstrated in (NZB x NZW) F1 lupus-prone mice lacking the Fcy chain (Clynes et al., 1998). These mice lack FcyRI and FcyRIII, which are necessary for the clearance of nonsoluble ICs, but nevertheless express the inhibitory FcyRIIb. Like the parental strain, they develop autoantibodies, but their nephritis is delayed and milder. ICs and complement are deposited in the kidney without any sign of inflammation, indicating that Fc binding of IC, rather than complement fixation, contributes to the development of nephritis. Moreover, the ratio of inhibitor versus activating Fcy receptors in the kidney conditions the response elicited by the ICs that are deposited. This ratio is regulated by several factors, including the cytokines present in the microenvironment (Ravetch, 2002). Among the activating Fcy receptors, FcyRIIa is the most frequent isoform in inflammatory cells and it is only present in humans. Mice transgenic for this receptor seem to be particularly sensitive to antibody-mediated inflammation, and one of the clinical manifestations is glomerulonephritis with IC deposition (Tan Sardjono et al., 2003). Conversely, FcyRIIb-- mice develop inflammation with crescent glomerulonephritis and other characteristics of Goodpasture's syndrome (Nakamura et al., 2000). The fact that double-deficient FcyRI/III mice are protected from nephrotoxic nephritis suggests that both high-affinity FcyRI and low-affinity FcyRIII are significant mediators of nephritis (Tarzi et al., 2003). In the search for a polymorphism that confers disease susceptibility, genes encoding FcRs were investigated in SLE patients and epidemiological studies were performed in different populations. Two alleles of FcyIIRa that encode receptors with different affinities have been found. A higher frequency of the allele encoding the lower affinity receptor (possibly responsible for a delayed clearance of ICs) has been reported in various studies (Salmon et al., 1996; Michel et al., 2000), but it did not confer a higher risk for nephritis (Karassa et al., 2002).
Third, autoantibodies, and in particular anti-DNA antibodies, may also penetrate living cells. It is known that they can bind different antigens, including cell surface molecules (Raz et al., 1993; Puccetti et al., 1995). For example, riboso-mal P proteins, one of the antigens bound by some anti-DNA antibodies (Caponi et al., 2002), have been found on the surface of various cell types, including endothelial cells, fibroblasts, hepatocytes, and astrocytes (Reichlin, 1996; Sun et al., 1996). Similarly, a subset of anti-DNA antibodies bind a-enolase, another protein that is expressed also on the cell surface (Pratesi et al., 2000). The binding to particular membrane proteins may mediate antibody internalization, and, when internalized, the antibody may still bind to intracellular targets.
The observation that antibodies directed against intracellular antigens may be able to penetrate into the cell has recently found confirmation and attracted renewed attention, because it could explain the perturbation seen in the physiological functions of the cells (Madaio and Yanase, 1998; Putterman, 2004). One of the cell surface receptors that mediate binding and internalization of anti-DNA antibodies is brush border myosin 1 (Yanase et al., 1997), a protein widely expressed in tissues. Its binding seems to induce a vesicle-mediated transport of the antibody into the cell. Once internalized, the antibody may interact with other potential targets. For example, antibodies with double specificities for DNA and the enzyme DNase may interfere with its activity, making the cells resistant to apoptotic stimuli (Madaio et al., 1996).
It is generally held that the loss of tolerance to nuclear antigens such as dsDNA and nucleosomes can be a central step in the pathogenesis of lupus nephritis. However, severe glomerulonephritis with proteinuria can be present even in the absence of antinuclear antibodies, anti-dsDNA, and antinucleosome antibodies, suggesting that breaking tolerance to chromatin and dsDNA is not required for the pathogenesis of lupus nephritis (Waters et al., 2004). This raises the possibility that other autoantibodies are able to cause glomerulonephritis in SLE.
Was this article helpful?