The prognosis of patients with systemic lupus erythematosus (SLE) has improved significantly over the past two decades 1. Earlier diagnosis on the basis of better awareness, description of new autoantibody specificities, and improvement of serological techniques may have supported this development. However, the introduction of pulse cyclophosphamide therapy for lupus nephritis 2 as well as advances in hemodialysis techniques were pivotal for this improvement, since it was irreversible renal failure and its consequences that previously had a high impact on mortality 3. Nevertheless, almost 10% of SLE patients still die within the first 5 years of their disease and their mean life expectancy is significantly shorter than in the general population, due partly to relentlessly progressive lupus in some patients and partly to sequelae of treatment, particularly those of cytotoxic agents and glucocorticoids 45. This situation calls for the search for new therapeutic strategies with higher efficacy and lesser comorbidity.

SLE is the prototype non-organ-specific autoimmune disease. A multisystem disorder, it destroys cells and organs by means of autoantibodies and immune complexes. The mechanisms underlying the hyper-reactivity and autoreactivity of the immune system in SLE are unknown. A setting of genetic susceptibility involving multiple genes 6 in conjunction with environmental triggers constitutes the hypothetical etiopathogenic background. As long as the triggers of the disease are unknown, novel therapeutic approaches must be aimed mainly at interference with the generation of autoantibodies and immune complexes or with their consequences, namely cell destruction and inflammation. In this review, some of the many potential future therapeutic approaches are discussed. Further information is included in a recent textbook chapter 7.

The role of infectious triggers of autoimmune diseases has been debated for decades. The earliest evidence pointing to such associations stems from acute rheumatic fever induced by streptococcal infections and the subsequent generation of antistreptococcal antibodies, which cross-react with cardiac tissue and lead to rheumatic heart disease in susceptible individuals 8. Type I diabetes has been often linked to Coxsackie virus infections 9; some forms of vasculitis appear to be a consequence of infection with hepatitis B or C virus 10; and peptide sequences of a variety of SLE-related autoantigens are homologous to sequences of various viral proteins 11121314. In fact, a peptide of the Sm protein, which can elicit a variety of autoantibodies and experimental lupus in an immunized animal, has homologies with a protein present on Epstein–Barr virus (EBV) 1516.

In contrast to the earlier, unsupported hypotheses, there is at least some recent important epidemiological evidence that SLE may be associated with EBV infection 16. EBV is a common infection worldwide. In the African regions, EBV infection is commonly associated with a variety of malignancies 17. In these regions, SLE is rare 18. On the other hand, in the industrialized world, people of African origin have a high risk of developing SLE 18, while EBV-associated malignancies are rare in those regions. Thus, it is conceivable that under different environmental circumstances EBV may induce different diseases or be, at least, a cofactor in the pathogenesis of different disorders, one of them SLE. Importantly, as EBV is one cause of lymphomas in the industrialized world 17 and is usually contracted during adolescence, one wonders if the design of a vaccine protecting against EBV infection and given in early childhood would reduce the risk not only of such malignancies, but also of SLE.

Autoimmune diseases are often considered to be a consequence of lost tolerance to self-antigens. Whether this is truly the case or there are other pathways responsible for the evolution of a pathogenic autoimmune response, induction of unresponsiveness and reversal of the respective immune response might constitute an interesting and successful therapeutic approach.

Given that some of the presumably most pathogenic types of autoantibodies in SLE are directed to dsDNA, down-modulation of their production is one important therapeutic aim. In experimental animals, a compound containing four oligonucleotides on a triethylene glycol backbone (LJP 394) is capable of downmodulating anti-dsDNA production, presumably by cross-linking the specific antigen receptor on the surface of the B cell. This approach led to amelioration of disease and higher survival in mice with lupus 19. Anti-dsDNA was also reduced in patients with SLE who were treated with LJP 394 20. Phase II/III randomized controlled trials are now under way.

In patients with autoimmune diseases, autoantibodies are usually of the IgG class and have hypermutated V region genes in comparison with the germ line. This clearly suggests the involvement of T-cell help. In fact, T cells incubated with nucleosomes or histones in both experimental and human SLE support the production of anti-dsDNA by B cells 2122. Since autoimmunity directed towards histone H1 appears to be of pivotal importance in SLE 23, induction of tolerance to nucleosomal antigens may be an interesting approach; it has already been successfully applied in experimental models 24. Moreover, activation of 'suppressor' T cells, which more than two decades ago were found to be defective in SLE 25 and have conceptually reemerged more recently as 'regulatory' T cells 26, may be an interesting new therapeutic approach for the induction of unresponsiveness. Moreover, since interaction of CTLA-4 with its ligand CD80/86 interrupts the costimulatory pathways needed to activate T cells 27, application of a CTLA-4-IgG fusion protein may interfere with the immunologic processes involved in disease induction in mice and man 2829 and lead to tolerance. Similar effects may be seen with antibodies to CD80/86 (B7.1 and 2) 30.

Tolerance may also be achieved by active immunization with tolerizing peptides and a reduction of autoantibody production has been observed experimentally when peptides from anti-dsDNA antibodies were used 31.

The mere presence of autoantibodies is not necessarily associated with disease. On the one hand, nonpathogenic autoimmunity is part of our 'normal' immunologic repertoire 32; on the other hand, the pathogenicity of autoantibodies and the consequent immune complexes is mostly brought about by the activation of complement and the interaction with cell-membrane-bound Fc receptors. Thus, interference with the complement pathways, as in knockout mice or when specific antibodies are used, can prevent or ameliorate lupus 3334. Soluble complement receptors may also be beneficial 35. Likewise, interference with the IgG Fcγ receptor (FcγR) interaction, as in FcγR I/III knockout mice or when anti-CD16 antibodies are used, can prevent the evolution of clinical manifestations of the disease 3637. On the other hand, activation of inhibitory FcγRs which contain an immunoreceptor tyrosine inhibitory motif (ITIM), in contrast with the immunoreceptor tyrosine activation motif (ITAM) of other FcγRs 38, may downmodulate B-cell function when co-cross-linked with the B cell's antigen receptor. Such FcγR-mediated inhibition of B-cell activity may not only be induced by immune complexes that carry an antigen binding to the surface immunoglobulin of the B cell while the immunoglobulin moiety of the immune complex engages the FcγRIIb, but also by intravenous immunoglobulin 39. There are reports of the efficacy of intravenous immunoglobulin in SLE 40, although further confirmation is awaited. The importance of FcγRs as potential therapeutic targets is also supported by reports on genetic linkage of SLE with a region on chromosome 1 that encodes the FcγRs 41.

The interventions discussed above were all directed at the consequences of immune complex production. However, considering autoantibody production, pathogenicity may also be prevented by interfering with autoantibody binding to the (auto)antigen or by eliminating the already bound autoantigen. The latter approach was not blessed with clinical efficacy, since the application of recombinant DNase, aiming at eliminating DNA from the respective immune complexes, had no clinical effects 42. In contrast, the application of heparin, which prevents the binding of circulating charged nucleosomal antigens to the glomerular basement membrane, prevented the occurrence of nephritis in experimental lupus and possibly should constitute an adjunctive therapy in patients with lupus nephritis 43. Another interesting means is to displace the antigen in the pathogenic immune complex with cross-reactive peptides. In one study, such an approach using peptides containing D-amino acids prevented glomerular deposition 44.

Although the debate whether SLE is primarily a Th1- or a Th2-mediated disease is still unresolved, cytokines appear to play important roles both in human and murine lupus. Not only has IFN-γ been found to be highly increased in sera of patients with lupus 45, but therapy with this cytokine has led to activation and induction of SLE 4647. The value of IFN-γ as a therapeutic target is supported by the fact that IFN-γ knockout lupus-prone mice do not develop the disease; moreover, treatment of experimental SLE with IFN-γ receptors inhibits lupus nephritis 48495051. All these notions are further supported by the observation of an amelioration of experimental lupus by the prototypic Th2 cytokine IL-4 52.

While the lymphokines mentioned above play important roles in the generation of the primary immune response and its skewing towards specific reactivity patterns 53, the proinflammatory cytokines are significantly involved in tissue destruction. The central proinflammatory cytokines, tumor necrosis factor (TNF)-α and IL-1, are increased in SLE and can both be activated by immune complexes 545556. Moreover, we have recently observed significant amounts of TNF-α by immunohistochemistry in renal biopsies from patients with lupus nephritis (manuscript in preparation). Nevertheless, the role of TNF-α is currently under intensive discussion. On the one hand, in experimental animals, TNF can induce nephritis and TNF-α deficiency ameliorates nephritis 5758; on the other hand, injection of TNF-α can ameliorate murine SLE under certain circumstances 59. This latter observation in conjunction with an occasional appearance of a lupus-like syndrome in patients with rheumatoid arthritis (RA) who are treated with TNF blockers 60 has led to the suggestion that TNF may be protective in lupus and that inhibition of TNF may therefore be potentially detrimental. However, not only are these events rare and, as of now, no more commonly observed than similar drug-induced lupus syndromes during many other therapies used for RA 6162, but also the anti-dsDNA autoantibodies observed among patients treated with TNF blocker are not consistently observed and are usually of the IgM rather than a pathogenic IgG isotype 6364.

To account for all these findings, my colleagues and I have proposed that TNF may play a dual role in SLE. This cytokine could well interfere with the regulation of the immune response and lead to an increase of autoantibody production; however, it may also have a critical role in the final pathway of SLE disease, namely immunologically induced and inflammation-induced tissue destruction. Thus, inhibition of TNF-α may, in fact, be a highly valuable tool in patients with active SLE, while inhibition of a potential autoantibody-enhancing activity could be achieved by concomitant immunosuppressive agents. The rapid interference of TNF blockers with the inflammatory response 65 suggests that they may be very beneficial for patients with active lupus nephritis and possibly other SLE manifestations 6667. Support for the efficacy and safety of TNF blockade in connective tissue disease stems from observations in patients with RA/SLE overlap (D Furst, personal communication) and individual cases of patients with mixed connective tissue disease (68 and unpublished observations). My colleagues and I are currently embarking on a small clinical trial with Ethical Committee approval to address the potential of TNF blockade to ameliorate SLE.

Proinflammatory cytokines and lymphokines mediate their effects by activating transcription factors via diverse signal transduction mechanisms induced after receptor ligation. Among the most important pathways are those involving mitogen-activated protein kinases (MAPKs) and nuclear factor (NF)-κB as well as the Janus kinases (JAKs). Interference with these pathways can ameliorate inflammatory diseases. One compound already approved for RA, leflunomide, interferes with NFκB activation 69 and also has some beneficial effects on mild SLE 7071. Many other drugs are currently in development and may have a potential as future therapeutic agents 7273.

The use of myeloablative cytotoxic therapy to combat the immunoinflammatory insult, with subsequent stem cell rescue to replenish the hematopoietic system and reconstitute the immune system, may be an attractive way of treating aggressive forms of SLE. Autologous stem cell therapy has been performed in small series of patients in recent years with some success 74, and we ourselves also had successful results in relentlessly progressive, life-threatening SLE 75.

In summary, rescue from organ failure and survival of patients with SLE need to be further improved. Advances in immunology and molecular biology have provided new therapeutic targets and new tools for potential treatment success. It will be important to study such new therapies using thoroughly designed protocols 76, but clearly there is hope for even better remedies than are available today and possibly cure of the disease.

CTLA-4 = cytotoxic T-lymphocyte antigen 4.