NEPHRITOGENIC AUTOANTIBODIES

Although multiple factors influence the variability of disease expression, autoantibodies are commonly involved in initiating disease activity.^30,32–34 Anti-dsDNA antibodies have been the most extensively studied.^30,35,36 In general, there is correlation of serum anti-dsDNA antibody levels with nephritis, a temporal association of rising titers and increased disease activity, and anti-dsDNA antibodies are enriched in glomerular immune deposits.^37–41 Furthermore, administration of some anti-dsDNA antibodies to normal mice can induce nephritis. However, it is also established that not all autoantibodies are pathogenic.^31,36,42 Moreover, among those that are nephritogenic their pathogenic profiles differ considerably. For example, after administration of anti-dsDNA antibodies to normal mice there was variability in location of deposits (e.g., subendothelial, mesangial) and the type of nephritis that resulted also differed.^32,33,36 These observations are consistent with clinical observations of patients with variable expression of nephritis.

Pathogenic insights are also provided by elution studies.^43–45 Antibodies eluted from human or murine lupus nephritis kidneys are typically of IgG isotype. They are distinguished from serum autoantibodies by their antigen-binding properties and their capacity to engage FcR and fix complement. Nephritogenic antibodies are highly cross-reactive and react with multiple autoantigens when compared to nonpathogenic serum autoantibodies, that prefer one antigen (e.g., DNA). These observations raised the possibility that nephritogenic antibodies cross-react with glomerular antigens. Further insights were derived from administration of monoclonal lupus autoantibodies to normal mice. Consistent with the aforementioned clinical observations in lupus patients, not all autoantibodies were nephritogenic (i.e., form immune deposits). However, cross-reactive autoantibodies appear to be more pathogenic than those that react with only DNA.

Furthermore, there are subsets of autoantibodies that produce distinguishable histologic and clinical patterns of nephritis. One explanation for these observations is due to differences in both autoantibody binding properties and the mechanism by which they form immune deposits.^40,46,47 In this context, glomerular immune deposits can potentially form by direct binding of the autoantibody to glomerular antigens, by immune complex formation where circulating autoantigens become complexed to glomerular antigens (and thereby serve as planted antigens for circulating autoantibodies), or by the deposition of circulating preformed immune complexes. Most evidence supports the initiation of immune deposition by in situ mechanisms, with either antibody binding to intrinsic glomerular antigens [e.g., cell surface antigens (cross-reactive theory) or circulating autoantigens that initially localize within glomeruli (planted antigen theory)].^{35,39,47–49}

Passive trapping of preformed immune complexes within glomeruli is more likely to accentuate disease by activating resident glomerular cells or cells that have infiltrated the kidney (e.g., through FcR interaction). In this regard, transfer of preformed immune complexes has not convincingly been observed to recapitulate nephritis. Nevertheless, immune complexes have been shown to transiently localize in glomeruli, typically in the mesangium. In culture, immune complexes stimulate mesangial cells, and in kidneys of lupus patients this may lower their threshold for activation, cause an accentuation of cytokine release, and/or increase in matrix production. This situation may be amplified in lupus patients because clearance of immune complexes by phagocytic cells is often impaired. Thus, although immune complex deposition may not initiate inflammation it likely plays an important role.

Further support for direct binding to glomeruli comes from experiments in which antibodies eluted from both human and murine lupus kidneys bound to glomerular extracts.^4,45,50,51 Subsequently, multiple autoantibody-glomerular cell surface and matrix antigen interactions have been described. For example, anti-DNA antibodies have been shown to react with glomerular cell surface antigens, laminin, heparan sulfate, and α-actinin.^48,52–56 Differences in cross-reactivity among autoantibody populations likely contribute to the diversity of disease observed in lupus patients. For example, different autoantibody specificities that predominate in a given individual influence where the deposits form and how disease is expressed.

Evidence also supports the “planted antigen” mechanism of immune deposition in lupus. Intracellular antigens released into the circulation after cell death (e.g., nucleosomes) may deposit in various sites within the glomerulus, where they can then serve as a nidus for binding of circulating autoantibodies. In support of a central role for nucleosomes, anti-DNA, anti-histone, and anti-nucleosome antibodies have all been shown to bind to nucleosomes previously localized within glomeruli. In this case, the nucleosome-glomerular interaction is facilitated by the relatively cationic charge of the histones with the relative negative charge of the GBM. Studies in both murine and human lupus sera have identified negatively charged heparan-sulfated glycosaminoglycan as the candidate ligand for this initial nucleosome binding. The deposition then presumably exposes the DNA within the nucleosome, which serves as a planted antigen for circulating anti-DNA antibodies.

In most instances, deposition of antibody alone is insufficient for nephritis. For inflammation to develop, activation of effector mechanisms is required, and this is largely influenced by the Fc region of the complexed antibody and the location of the deposits. For many years, the accepted model of immune-mediated injury (based on studies done in the 1950s and 1960s involving the Arthus reaction) held that complement activation was responsible for all effector responses and consequences of immune complex deposition.¹ More recently, an essential role for complement independent mechanisms were derived from observations in mice with targeted disruptions of complement components C3 or C4.¹ They still mounted an Arthus reaction and other inflammatory responses. Interestingly, mice that are deficient in the activating FcγR receptors (FcγRIII) did not develop inflammation.

In elegant studies directly related to lupus nephritis, Ravetch and colleagues demonstrated the obligatory function of FcR by back-crossing lupus-prone (NZB/NZW F1) mice with a congenic strain that had a homozygous deletion for the activating Fc receptor [FcRγ (−/−)].^58–62 The FcRγ (−/−)-deficient NZB/NZW mice developed circulating autoantibodies and glomerular deposits of IgG and complement. However, they did not develop proteinuria or nephritis. The results support the conclusions that deposited immune complexes and C3 are insufficient to trigger effector cell activation and that functional activator FcRs play a pivotal role in mediating inflammation in autoimmune nephritis.

The role of FcRs in autoimmunity is not limited to its triggering of effector cells. The inhibitory FcγRIIB suppresses B-cell activation and proliferation by opposing activation through the B-cell receptor (BCR).¹ By contrast, activation of B cells and exaggerated antibody responses occur if FcγRIIB expression is decreased or signaling is impaired. This can result in loss of tolerance and the development of autoimmunity. In this regard, inactivation of FcγRIIb in normal mice (C57BL/6) leads to a spontaneous lupus-like disease with anti-dsDNA autoantibodies, glomerular immune complexes, and severe glomerulonephritis.¹

Some autoimmune-prone mice (such as NZB and MRL) show reduced surface expression of Fc γRIIB, which has been attributed to polymorphisms in the promoter region of this receptor gene.¹ Furthermore, genetic polymorphisms of Fc receptors have also been described in humans. Some allelic variants alter the binding of the FcR Ig and thereby affect immune complex clearance. In some populations, the low-affinity binding polymorphisms confer a higher risk for development of lupus nephritis in comparison to the higher binding variants. Thus, altering FcR expression is a rational therapeutic approach for study.

Nevertheless, in lupus nephritis (like other forms of nephritis) the anatomic location of the immune deposits influences both the predominant effector mechanism and the ultimate clinical and histologic manifestations of the disease.^33,36,40,48 For example, if immune deposits form in the subepithelial area (as in membranous lupus) the presence of the GBM prevents inflammatory cell recruitment to the site, and the resultant pathology is nonexudative. This presumably occurs because FcR on circulating cells are not engaged. However, in this setting the membrane attack complex (C5b-9, generated from complement activation) mediates sublytic injury to cells. Consequently there is effacement of podocytes. This leads to altered glomerular barrier function and heavy proteinuria. Subsequently, there may be podocyte loss that may be exacerbated by the formation of very large deposits, contributing to overproduction of extracellular matrix components and leading to scarring.

By contrast, if the immune deposits are accessible to the vascular space (such as in the subendothelial and mesangial regions) effector cells are recruited and inflammatory lesions dominate. In the latter situation, activation of resident glomerular cells through FcR (largely) and complement cascade contribute to the lesions described previously.

Although the relationship of autoantibodies to lupus and other autoimmune diseases is well accepted,36–38,^43,48,53,64 recent data suggest that other factors are required.^65–68 In this regard, Arbuckle and colleagues evaluated serum samples obtained from the Department of Defense Serum Repository in order to investigate the onset of autoantibody development prior to the clinical well diagnosis of lupus.^65–67 The investigators determined that autoantibodies were present well before a diagnosis of SLE in 88% of patients studied.

An important finding was that there was a temporal pattern of autoantibody specificity with respect to the appearance of clinical manifestations/diagnosis of SLE. Specifically, certain autoantibodies (ANA, anti-Ro, and anti-La) appeared years prior to the diagnosis of SLE whereas others (anti-Sm and anti-nuclear ribonucleoprotein antibodies) appeared only during the periclinical stages. These observations are consistent with the notions that there are specific properties of nephritogenic antibodies that initiate disease and/or that other factors (e.g., activation of B cells) are required.