|Home | About | Journals | Submit | Contact Us | Français|
The present review focuses on pathogenic molecular and transcriptional events in patients with lupus nephritis. These factors are renal DNaseI, exposed chromatin fragments and the corresponding chromatin-reactive autoantibodies. Lupus nephritis is the most serious complication in human systemic lupus erythematosus, and is characterised by deposition of chromatin fragment-IgG complexes in the mesangial matrix and glomerular basement membranes. The latter deposition defines end-stage disease. This event is stringently linked to a renal-restricted shutdown of expression of the DNaseI gene, as determined by loss of DNaseI mRNA level and DNaseI enzyme activity. The major aim of the present review is to generate new therapeutic strategies based on new insight into the disease pathogenesis.
Shortly after their discovery in 1957 [1-3], antibodies to dsDNA were associated with renal manifestation of systemic lupus erythematosus (SLE). A prominent observation was that anti-dsDNA antibodies were eluted from affected glomeruli in the context of lupus nephritis [4-8]. At the time when the nephritogenic potential of anti-bodies to dsDNA was revealed, their binding in glomeruli was logically claimed to depend on exposed DNA. This DNA was thought to be bound in situ in glomeruli, where it was targeted by the antibodies. This assumption derived from two facts: DNA bound glomerular collagen [9,10], and the antibodies were specific for DNA [11,12].
One problem was linked with this model. Not all individuals with anti-dsDNA antibodies in their circulation developed nephritis. A convenient model to understand nephritogenicity of anti-dsDNA antibodies proposes that only those antibodies which cross-reacted with inherent renal antigens induced the organ disease. A nephritogenic potential of antibodies against DNA (or nucleosomes) is thus today critically challenged by alternative models implying that antibodies cross-react with glomerular antigens such as α-actinin, laminin, or cell surface structures [13-19]. Conflicting data from different types of analytical strategies have resulted in different models explaining how anti-DNA antibodies induce nephritis. Even though these models are attractive, none have been validated beyond any doubt, although the dominant specificity of nephritogenic antibodies for dsDNA may point to the most obvious target structures in nephritic kidneys-nucleosomes released from dead cells. An alternative model that may explain whether an anti-dsDNA antibody executes a nephritogenic potential might therefore be the availability of exposed chromatin particles within glomeruli. This hypothesis means that anti-dsDNA antibodies execute their pathogenic potential only in situations where chromatin fragments are exposed in glomeruli. In the absence of this target structure, the antibodies remain nonpathogenic epiphenomena despite their diagnostic potential.
The origin of renally exposed chromatin fragments has been difficult to assess. One general idea has been that they reach glomeruli through circulation. Taking into consideration that the target antigens for anti-dsDNA and anti-nucleosome antibodies appear by immune electron microscopy as large chromatin fragments , however, it is difficult to explain how these may reach and deposit in glomeruli.
A notable change in thinking entailed by our studies is rather that chromatin fragments exposed in glomeruli are released from dying renal cells, and that these fragments are not degraded during the cell death process because of an acquired loss of the dominant renal nuclease DNaseI . This model is the focus of the present review, and will be discussed in detail below.
SLE, as we understand the disease today, is linked to B-cell and T-cell autoimmunity to nucleosomes, and particularly to the individual components of nucleosomes-native (ds)DNA and histones. These are important diagnostic parameters for SLE [12,22]. Furthermore, sets of these autoantibodies possess the potential to induce nephritis, the most serious complication in SLE [23,24].
The aetiology of SLE is not fully understood, but there are recent advances in its understanding. For example, there is growing interest in regulatory RNA molecules in SLE. miRNAs belong to a family of short noncoding RNAs. These have been shown to play important roles in gene regulation. Recent data suggest that miR-126 regulates DNA methylation in CD4+ T cells and contributes to T-cell autoreactivity in SLE by directly targeting DNMT1 . Similarly, a recently published comprehensive analysis of miRNA expression patterns in renal biopsies of lupus nephritis patients further demonstrates that miRNAs are probable factors involved in the pathogenesis of lupus nephritis. We see now the contour of a new scientific field to understand elements of lupus nephritis; study of regulatory RNA in autoimmune syndromes such as SLE and lupus nephritis is a new and fast-growing field to analyse transcriptomics in SLE , and miRNA may have a strong impact on progressive kidney diseases as discussed by Kato and colleagues .
Another cascade of events that may relate to pathogenesis of SLE and lupus nephritis is linked to engagement of Toll-like receptors (TLRs) by exposed chromatin. Activation of TLRs induces upregulation of proinflammatory cytokines (TNFα, IFNγ) and interleukins . For example, IFNγ contributes directly to the progression of lupus nephritis . Furthermore, Rönnblom and colleagues discussed recently the increasing evidence that activated type I interferons in lupus are critical in the aetiopathogenesis of the disease and an important therapeutic target . Kidney sections from patients with SLE glomerulonephritis contain high amounts of TNFα, and expression levels correlated with local (histological) disease activity .
TNFα and IFNγ are important inducers of the matrix metalloproteases (MMPs) MMP2 and MMP9. These are collagenases that, when overexpressed, have the potential to disintegrate membranes [32,33]. Membrane disintegration may be the factor that promotes deposition of immune complexes in glomerular basement membranes (GBMs), as discussed recently . The engagement of TLRs is thus an event central in the pathogenesis and progression of SLE and lupus nephritis.
In the next sections, the current insight into murine forms of lupus nephritis will be discussed, with potential implications of data on the human form of this organ disease.
Substantial data have been provided during recent years related to why and how anti-dsDNA antibodies are produced (see, for example, [35-39]); to how they exert their clinical impact, whether through interaction with DNA or nucleosomes [5,40-42], or through cross-reaction with inherent renal antigens [13,15,17,43]; and to analyse whether the nature of their glomerular target structures are reflected by their specificity or cross-reactivity [12,40,44].
In recent studies, we and other workers have developed high-resolution techniques providing evidence that nephritogenic anti-dsDNA/nucleosome antibodies recognise selectively intraglomerular, extracellular chromatin structures in vivo [20,40,41,44]. These structures appear as electron-dense structures by transmission electron microscopy, and have been shown to be composed of chromatin fragments and IgG molecules by different forms of immune electron microscopy and by co-localisation terminal deoxynucleotidyltransferase biotin-dUTP nicked end-labelled immune electron microscopy assay . Autoantibody deposits in vivo are strictly localised to these structures, and co-localise with antibodies to DNA and histones added to the sections in vitro .
These data confirm the historical hypothesis that anti-dsDNA antibodies form complexes with nucleosomes and these immune complexes deposit in glomerular membranes (reviewed in ). This deposition does not exclude the involvement of other autoantibodies that may participate in the progression of lupus nephritis, such as antibodies specific for the membrane and matrix component , α-actinin [13,43], C1q  and, for example, renal cell membranes . The role of these latter anti-bodies in lupus nephritis, however, remains to be determined.
Recently, we demonstrated that anti-DNA antibodies, renal DNaseI and matrix MMP mRNA levels and enzyme activities are cooperative and instrumental in early and late events in murine lupus nephritis, as determined in (NZBxNZW)F1 mice . Early phases of nephritis were associated with chromatin-IgG complex deposition in the mesangial matrix, which correlated with appearance of anti-dsDNA antibodies. Subsequent to this event, we observed a dramatic downregulation of renal DNaseI mRNA level and enzyme activity, while MMP2 and MMP9 mRNA levels and enzyme activities increased. Reduced levels of renal DNaseI correlated with deficient renal fragmentation of chromatin from dead cells, and with accumulation of large chromatin fragments in GBMs. A similar downregulation of DNaseI was not observed in mesangial nephritis , or in nephritis in the context of Wegener's granulomatosis . In situ deposition of chromatin fragments has been described in several experimental nuclease deficiencies on nonautoimmune backgrounds (reviewed in ).
In contrast to the correlation of renal DNaseI shutdown, Martinez-Valle and colleagues did not observe any statistical relationship between serum DNaseI activity and disease evolution time, clinical and laboratory parameters including proteinuria and autoantibodies, or the treatment pattern received by the patients [50,51]. In agreement with this observation, increasing DNaseI activity in vivo by injecting recombinant human DNaseI intravenously and subcutaneously in patients with SLE failed to show any effect on serum markers of disease activity . Furthermore, mutations causing reduced DNaseI in lupus patients did not correlate with unique clinical symptoms . This lack of correlation may mean that extracellular DNaseI enzyme activity is not important in the context of lupus nephritis pathogenesis. Rather, DNaseI is important in the context of cell death, where DNaseI is in fact the initiator of chromatin fragmentation in order to fascilitate a silent removal to avoid, for example, inflammation [54,55]. Renal DNaseI gene shutdown may therefore impose chromatin exposure in situ because of inefficient enzymatic degradation. In this model, serum DNaseI may play an inferior role in extracellular chromatin degradation. It is questionable whether extracellular chromatin, when bound to membranes and covered by IgGs, will be degraded at all by DNaseI.
Recent data in murine lupus nephritis thus demonstrate that acquired loss of renal DNaseI enzyme activity is a dominant event responsible for the progression of mesangial nephritis into end-stage organ disease . However, exposed chromatin may not be pathogenic in the absence of antibodies to dsDNA or to nucleosomes . The principal cellular and molecular requirements needed to produce these autoantibodies have been explained experimentally [35-38], but the mechanism(s) accounting for them in vivo in the context of SLE and lupus nephritis has not yet been determined. Published data, however, indicate that defects in nucleases linked to apoptotic or necrotic cell death are not associated with induction of anti-dsDNA or anti-nucleosome autoantibodies (for review, see ). The data discussed here, nevertheless, explain how an unusual exposure of chromatin may be a central factor in the evolution of lupus nephritis, but not in promoting nephritogenic chromatin-specific autoimmunity.
With loss of renal DNaseI, the signalling pathway from chromatin fragment-stimulation of TLR to MMP expression has been described [58,59]. MMPs are collagenases with potential to disintegrate membranes [32,33]. Membrane disintegration may promote deposition of immune complexes in GBMs.
In murine lupus nephritis, anti-DNA antibodies gain their pathogenic potential when chromatin fragments are exposed in glomeruli. Chromatin fragments thus represent the axis in a circulus vitiosus, where chromatin - the inducer of nephritogenic autoimmunity - is the glomerular target for the autoantibodies, and thereby accounts for the organ disease (discussed in [60,61]).
Chromatin fragments exposed in the kidneys may derive from either increased apoptosis or deficient clearance of apoptotic or secondary necrotic material [61-63]. How tolerance against chromatin components is terminated is not fully understood. Chromatin undergoes alterations during apoptosis and is normally not exposed for the immune system. In the case of increased apoptosis or deficient clearance, however, these components may be exposed as secondary necrotic chromatin with the potential to induce an antigen-selective immune response [64-66]. For example, plasma chromatin found in SLE patients is hypomethylated , and hypomethylated DNA is more immunogenic and can induce maturation of dendritic cells and potentially activate autoimmune T cells and B cells . Furthermore, sera of SLE patients contain circulating chromatin fragments complexed with the DNA-binding protein HMGB1 . This protein is a proinflammatory mediator that binds chromatin of apoptotic cells. The HMGB1-nucleosome complexes may activate antigen-presenting cells, which have the potential to promote activation of relevant T-helper cells and then DNA-specific B cells, with production of chromatin-specific autoantibodies as a net result . Exposed and retained chromatin may therefore promote production of chromatin-specific autoantibodies.
In an infectious context, viruses such as polyomavirus BK may induce cell death as a consequence of virus replication. This process may be relevant to lupus nephritis, since there are several reports that demonstrate productive polyomavirus infection in human SLE (see  and references therein). Productive renal polyoma virus activation may be imposed by treatment of the disease with immunosuppressive drugs, and may not be specifically linked to the lupus pathogenesis [68,69]. This expression pattern is similar to what is seen in renal transplants during immunosuppresion [70,71].
Irrespective of the cause for polyomavirus replication, the viral transcription factor large T antigen forms complexes with the host cell chromatin. This complex may affect the immune system in analogy with a hapten-carrier complex, where B cells bind nucleosomal DNA (the hapten) through the DNA-specific antigen receptor and process and present T-antigen-derived peptides (the carrier) to nontolerant T cells (this model is extensively reviewed in ). Chromatin may thus generate autoimmunity by quite different pathways linked to modification of chromatin and various infections [36,72].
We recently performed in vitro studies demonstrating that nucleosomes and nucleosomes in complex with anti-DNA antibodies have high affinity for glomerular and epidermal basement membrane components such as laminin and collagen . This affinity may be a major factor that explains why chromatin-containing immune complexes associate with membrane and matrix structures in human nephritis .
One factor that may contribute to deviation in chromatin composition and size is DNaseI, which is the major nuclease in kidneys  but also in serum, where it may participate in chromatin degradation in context of necrosis [55,74]. Several studies have demonstrated reduced levels of serum DNaseI in SLE patients [50,51,74-76]. This reduction could provisionally explain why the chromatin concentration in circulation of lupus patients with nephritis is reported to be higher than in control individuals [77,78]. Another reason for this reduction could be that immune complexes in SLE patients are protected against nuclease attacks by DNA-binding proteins and immunoglobulins, present in sera. A problem that needs to be solved is therefore whether circulating chromatin-containing immune complexes in lupus nephritis patients are less sensitive to DNaseI than chromatin fragments in anti-chromatin antibody-negative healthy donors.
Most of the data on lupus nephritis so far relate to studies of murine models of the disease. In the next section, we will translate available basic data into a detailed evidence-based model for human lupus nephritis. We shall subsequently convert this information into new and rational treatment modalities.
Irrespective of the complexity of potentially nephritogenic autoantibodies associated with SLE, a consensus has evolved that antibodies to dsDNA and nucleosomes are central pathogenic factors involved in development of human lupus nephritis [12,23,24,79-81]. The divergent models to explain the basic processes in human lupus nephritis may have evolved simply because we still lack data that provide definitive insight into the nephritic process(es).
In a pilot study, data demonstrate that human nephritogenic anti-DNA antibodies bind chromatin-like structures in GBMs and the mesangial matrix , similar to what we have observed in murine lupus nephritis . In that pilot study it became evident that in-vivo-bound GBM-associated autoantibodies co-localised in electron-dense structures with experimental antibodies to histone H1, histone H3 and transcription factor TBP, and with nicked DNA ; that is, results identical to those observed in murine lupus nephritis. In a recent study, we also demonstrated in advanced stages of human lupus nephritis that the DNaseI protein was nearly absent in the nephritic kidneys compared with non-nephritic kidneys and nonaffected tissue of kidneys extirpated due to cancer . The nephritic processes such as those determined in murine lupus nephritis thus seem highly relevant to understand the process in the human form of the disease. The disease process as outlined in Figure Figure11 is therefore most probably relevant to understand both forms of lupus nephritis.
Based on the results discussed above, we propose the following model to understand initiation and progression of lupus nephritis in both mice and humans. The data from murine lupus nephritis are summarised as follows (see Figure Figure11 for details). The impact of antibodies to dsDNA is crucial for early deposition of chromatin fragments in the mesangial matrix. Linked to progression of the disease, secondary necrotic chromatin fragments are generated and retained in kidneys when the renal nuclease DNaseI mRNA level and DNaseI enzyme activity are downregulated. Secondary to this process, chromatin is not degraded appropriately, and instead large chromatin fragments are retained in glomerular capillary membranes in association with chromatin-reactive IgG autoantibodies. In this situation, chromatin fragments are also exposed to macrophages and dendritic cells in which they stimulate TLRs, which may explain the increased expression of MMPs. The increase of MMPs may further impose deposition of chromatin fragments in glomerular membranes because of capillary membrane disintegration [32-34]. Our conclusion is that human lupus nephritis is dependent on the same distinct processes.
Acquired loss of DNaseI in both murine and human lupus nephritis may be controlled at different levels, including activation of convergently encoded genes using sequence elements from the DNaseI gene, methylation of DNaseI encoding elements and the promoter, or interference with miRNA targeting DNaseI mRNA or other mRNAs involved in expression of DNaseI. This hypothesis is currently under investigation in our laboratory.
Despite improvements in outcomes of immunosuppressive treatment of patients with lupus nephritis, renal remission is obtained in less than 50% of cases within 2 years; approximately 10% may progress to end-stage renal disease [83,84], which is associated with significantly increased rates of cardiovascular mortality . In general, patients with SLE and renal involvement have more cardiovascular disease than the remaining SLE patients [86,87]. These findings call for improved treatment regimes for patients with lupus nephritis, not only in terms of improved renal outcome but also with regard to cardiovascular outcome.
The results discussed above on the aetiology of lupus nephritis demonstrate that DNaseI, the major renal nuclease, is profoundly downregulated during development of severe membrano-proliferative nephritis. Considering this single information, it may be sound to conclude that lupus nephritis is a disease entity that depends on processes that are unique to the kidney, and that an acquired shutdown of renal DNaseI is the factor that determines the disease process and clinical outcome, as outlined in Figure Figure1.1. This opens the way for new therapeutic directions
There are strong data allowing us to assume that the two-stepped process accounting for murine lupus nephritis is also relevant in human lupus nephritis. If this assumption is correct then we can introduce new types of treatment focusing on disruption of chromatin structures in vivo by chaperone molecules that open the compact and nuclease/protease-resistant chromatin structure. Such molecules may make the chromatin structure more susceptible for proteases and nucleases [88-91]. On the contrary, certain chaperone molecules may prevent binding of nucleosomes to glomerular membranes by altering the net charge of chromatin fragments, as demonstrated in vitro by surface plasmon resonance , and also potentially in vivo in the context of permanent infusion of such chaperone molecules . Therefore, it is important to determine whether processes that account for the potentially fatal human lupus nephritis can be avoided without influence on the immune system. In the next section, a possible strategy and relevant experiments will be discussed.
One possible approach is to use molecules that are involved in chromatin assembly, disassembly or remodelling. Such molecules have the ability to alter the conformation of the chromatin structure, which may result in increased sensitivity for both nucleases and proteases. This could lead to increased degradation of the potentially immunogenic chromatin fragments [61,93] - fragments that otherwise would be presented to the immune system - thereby inducing pathogenic anti-dsDNA/anti-nucleosome antibody responses.
This idea derives from the described effects of chaperone molecules such as nucleosome assembly protein 1, a histone chaperone molecule that modulates binding of the linker DNA-associated histone H1 to chromatin and induces an extended and open chromatin fibre conformation [94,95]. Nucleoplasmin is also a histone chaperone that binds and exchanges histones to re-establish the chromatin structure, and is involved in opening and relaxation of the chromatin structures [91,96]. The heat shock protein HSP90 has a similar effect on chromatin structure [97,98].
These are examples of molecules that induce changes in chromatin conformation which may result in increased accessibility for proteases and nucleases and in increased degradation of nucleosomes. Whether chaperone molecules are tolerated in vivo in doses necessary for therapeutic effect has not been determined. One chaperone molecule that may be used in a therapeutic context is heparin, a negatively charged molecule that is well tolerated in vivo and has similar effects on chromatin structure as the molecules mentioned above. Heparin derivatives have been evaluated for their effect on nucleosome and chromatin structure. Common for these studies are data demonstrating that heparin makes nucleosomes more accessible to nucleases [74,99] by binding the trypsin-sensitive solvent-phase tails of the core histones . Heparin also increases enhancer-promoter communication  by disassembling the chromatin structure [101,102]. Typical for heparin-induced structural changes is increased fragmentation of the nucleosomal structure by nucleases ubiquitously present in biological fluids . As anionic heparin binds tightly to histone tails, and potentially changes the net charge of the nucleosome, heparin may in fact also inhibit binding of nucleosome-containing immune complexes to components of the GBM, like laminins and collagens. Heparin may thus have a two-sided effect on the role of chromatin fragments in lupus nephritis; increased enzymatic degradation of chromatin fragments, and inhibition of their binding to glomerular membranes.
Interfering with chromatin-IgG complex binding to glomerular extracellular membranes could be a new treatment strategy. Negatively charged heparin binds to positively charged histones in the nucleosome complex and opens their architecture .
A pilot study has demonstrated that chromatin is more sensitive to both DNaseI and proteases in the presence of low molecular weight heparin, indicating changes in chromatin structure. Highly promising was the observation that heparin inhibited binding of chromatin-IgG complexes to glomerular laminin and collagen in vitro, as demonstrated by surface plasmon resonance (Figure (Figure1)1) . There is thus a strong indication that heparin derivatives (or other nucleosome re-modelling proteins such as nucleoplasmin ) indeed exert a two-sided therapeutic effect on lupus nephritis: heparin alters chromatin structures and allows a nearby complete degradation of B-cell-recognising DNA in chromatin, thus preventing production of nephritogenic anti-DNA antibodies; and heparin prevents binding of chromatin-IgG fragments that escape enzymatic degradation in vivo . Continuous infusion of heparin delayed production of anti-dsDNA antibodies and development of lupus nephritis in (NZBxNZW)F1 hybrid mice. Previous studies have revealed that heparin/heparinoid treatment has a therapeutic effect on the activity of lupus nephritis in MRL-lpr/lpr mice . The mechanism(s) for this therapeutic effect, however, was not determined - but the anticoagulant effect does not seem to be essential . In another study, Naparstek and colleagues indicated that the binding of antibodies to dsDNA could be inhibited by heparin . This potentially important observation, however, has not been followed up by further studies.
The aim of current experiments in our laboratory is to determine an epigenetic mechanism(s) for renal DNaseI shutdown, and to analyse whether DNaseI sensitivity of immune complexes purified from patients with lupus nephritis is increased by heparin at concentrations tolerated in a clinical context. Highly relevant also are the planned experiments to analyse whether low molecular weight heparin interferes with processing and presentation of chromatin fragments by antigen-presenting cells.
Detailed studies have offered new insight into molecular and transcriptional events that explain processes contributing to lupus nephritis. This insight has provided us with new therapeutic ideas and possibilities. Analysing chemical compounds that inhibit binding of chromatin-IgG complexes to components of the extracellular matrices and membranes, in combination with alteration of extracellular chromatin structure to make them more sensitive to proteases and nucleases, is the focus for our investigation. In the future it may be possible to prevent both autoimmunity to DNA and chromatin fragments on one side, and to inhibit binding of chromatin fragments to the mesangial matrix and the GBMs on the other side.
dsDNA: double-strand DNA; GBM: glomerular basement membrane; IFN: interferon; miRNA: microRNA; MMP: matrix metalloprotease; SLE: systemic lupus erythematosus; TLR: Toll-like receptor; TNF: tumour necrosis factor.
The authors declare that they have no competing interests.
Autoimmune Basis of Rheumatic Diseases
This article is part of a series on Systemic lupus erythematosus, edited by David Pisetsky, which can be found online at http://arthritis-research.com/series/lupus
This series forms part of a special collection of reviews covering major autoimmune rheumatic diseases, available at: http://arthritis-research.com/series/abrd
The authors thank Søren Jacobsen for help with the manuscript. The present study was supported by grants from the Northern Norway Regional Health Authority Medical Research Program (grant numbers SFP-100-04 and SFP-101-04), and from the University of Tromsø as Milieu support given to OPR.