Introduction: B cells in systemic lupus erythematosus

A central feature of systemic lupus erythematosus (SLE) is the loss of B-cell tolerance. At least some autoantibodies from a limited spectrum of reactivities against mainly intracellular antigens are usually present, and probably account for some of the pathological manifestations. Although the numbers of B cells in the peripheral blood are often decreased, those that are present have abnormal phenotypes indicative of activation [1].

In addition, substantial evidence from mouse models of systemic autoimmunity clearly implicates the central role of B cells [2]. In several spontaneous models, the genetic abnormalities that cause the loss of tolerance must be expressed in those B cells that become autoimmune [3]. A wide variety of single gene abnormalities that are largely or solely expressed in B cells also leads to lupus-like systemic autoimmunity, either by lack of function through spontaneous mutations or knockout transgenics, or through hyperexpression of exogenous transgenes [4]. If B cells are removed from lupus models by genetic manipulations or chronic antibody therapy, the syndrome is largely suppressed, including T-cell abnormalities [5].

Other studies in mice genetically without B cells also implicate B cells in a number of immunoregulatory interactions that go beyond their clear role as the precursor of antibody forming cells [6]. B cells can regulate T cells, dendritic cells and other B cells. They can produce a variety of cytokines, including IL-4 and IL-10, and even can differentiate into subtypes that secrete certain sets of cytokines, analogous to T helper type 1 and T helper type 2 cells [7]. B cells are superb antigen presenting cells, since they can express MHC class II as well as costimulatory molecules such as CD80 and CD86, and their cell surface immunuoglobulin antigen receptor is ideal for focusing and concentrating specific protein molecules [8].

Curiously, at present we do not know for certain what role B cells play in human SLE [9]. Some clinical manifestations appear to be antibody mediated, such as hemolytic anemia and glomerular inflammation, but the pathogenesis of many of the aspects of the disease remains obscure, and most of the disease-associated autoantibodies do not appear to have a direct pathogenic role. The potential immunopathogenic importance of B cells is implicated in the occasional case reports of SLE patients that developed common variable immunodeficiency and showed improvement in the manifestations of SLE concomitant with loss of B-cell function [10].

Rituximab and B-cell depletion

It was thus a reasonable hypothesis that removing B cells in SLE might have a positive therapeutic effect [11]. The availability of Rituxan® (rituximab) (Genentech, South San Francisco, CA, USA) made it possible to test this hypothesis [12]. Rituximab is a chimeric monoclonal antibody reagent consisting of human IgG1 and kappa constant regions, and of mouse variable regions from a hybridoma directed at human CD20. CD20 is a specific B-cell marker present in all stages of B-cell development except the earliest and the latest [13]. Its cell function is unknown (CD20 knockout mice have no obvious B-cell deficits [14]) but it is expressed at high levels, it does not shed or endocytose when exposed to antibody, and it does not exist in a soluble form [15]. These features predicted that CD20 might be an excellent target for therapy directed at B-cell malignancies. This in fact proved to be the case, and rituximab was approved in 1997 for treatment of non-Hodgkin B-cell lymphomas [12]. After four weekly intravenous doses, rituximab also depletes normal B cells from the peripheral blood almost completely in most patients, and this depletion persists for 6 months and more, well beyond the persistence of the rituximab itself. Importantly, the extent of depletion of B cells from peripheral lymphoid organs is not known. However, serum immunoglobulins do not fall substantially during treatment, and increased infections have not been found to be a complication.

After rituximab received Food and Drug Administration approval for lymphoma, several investigators began trying it in uncontrolled series of patients with a variety of autoimmune diseases. The hope was not only that the drug might be therapeutically effective, but also that through monitoring its use we would learn a great deal about the role of B cells in the pathogenesis of these conditions. Encouraging anecdotal reports have appeared for a potential response to rituximab of patients with rheumatoid arthritis, polymyositis/dermatomyositis, idiopathic thrombocytopenia purpura, essential mixed cryoglobulinemia, hemolytic anemia, myasthenia gravis, Wegener's granulomatosis, and IgM-mediated neuropathy, as well as patients with SLE [16–23]. This approach has recently received a major impetus from the preliminary report of substantial efficacy in a controlled trial in rheumatoid arthritis [24].

So what about SLE? A published experience with six patients looked promising, as did a few individual anecdotes [17]. A phase I trial from Looney and colleagues showed improvement in certain subgroups in a post hoc analysis [25]. Our own phase I trial also has examples of patients who have improved clinically and who have decreased steroid usage after treatment (unpublished data). The safety profile has so far been good, as would have been predicted from the extensive experience (over 300,000 patients) in individuals with B-cell malignancies, although it must not be forgotten that rare cases of severe, fatal infusion reactions have been reported [26]. Surprisingly, many of the SLE patients developed human antichimeric antibody responses, perhaps in part because they did not receive the full therapeutic dose in the early dose-escalation period of the protocols.

At this point, it is essential that controlled trials be conducted to determine for certain whether rituximab has a clinically useful therapeutic effect in SLE. If efficacy could be established for one particular manifestation, such as renal disease or skin disease, or on the basis of overall disease scores, it would add an important additional drug to our approach to SLE in general. If rituximab permitted substantially lower use of steroids or cytotoxics, particularly cyclophosphamide, it would be a significant advance in patient safety.

What can we expect with rituximab treatment of SLE? The data so far indicate that autoantibodies, such as anti-DNA, are not suppressed. This is in distinction to what has been seen in the rheumatoid population, where both rheumatoid factors and anticyclic citrullinated peptide antibodies decreased [27]. Anecdotally, skin and musculoskeletal manifestations may have been particularly responsive. All of the phase I trials and case reports have so far been short term. If rituximab can be shown to be effective, how will it be appropriately used? Perhaps in conjunction with cytotoxic or steroid therapy, as in the rheumatoid arthritis trial? What about repeat dosing? Will human antichimeric antibody (HACA) development be avoided by full dosing and by combination with cytotoxic agents, in parallel with the experience with infliximab [28]? If HACA do appear, will they either cause increased complications, such as infusion reactions, or will they dampen the effect of treatment or even retreatment by blocking binding of rituximab to CD20 or by changing pharmacokinetics? In most patients treated with rituximab, the B cells return to the peripheral blood starting about 6 months after treatment. Will the return of B cells signal a recrudescence of clinical disease? When should patients be retreated, and at what doses and for how long? If prolonged B-cell suppression is necessary to maintain clinical control, will this eventually lead to an immunosuppressed state with a high risk for pyogenic infections? Will it be possible to combine rituximab with other biologicals that interfere with, for example, T cell–B cell collaboration, in order to achieve greater clinical benefit with less risk? Once efficacy is established in a controlled setting, all of these questions will have to be addressed either by additional trials or by collective experience.

From a more theoretical point of view, a major issue revolves around the role of CD20+ B cells in the pathogenesis of disease. If the key cells to target are the autoantibody forming cells themselves, then the effectiveness of rituximab would depend on the extent of persistence of the CD20 marker in this population, and the longevity of those antibody forming cells that do not express CD20 [29]. If rituximab successfully could deplete the CD20+ precursors of CD20- antibody forming cells, then these cells would not be replaced when they die off. Another issue is what the B-cell repertoire will look like when it returns, and what aspects of the disease are indeed B-cell dependent.

Conclusion

Given the variety of mouse models of SLE, and the fidelity with which they reproduce the spectrum of autoantibodies seen in SLE, it is unfortunate that no method currently exists to deplete B cells from mature, diseased animals as efficiently as rituximab depletes B cells in humans. If it were possible to model this therapeutic approach in mice, then we could get preliminary answers to many of the questions raised, and design our therapeutic approaches more logically. Although our and other laboratories are trying to develop this modality in mice, for the present it remains important that the trials to demonstrate safety and efficacy of rituximab in SLE, as well as in other autoimmune diseases, be accompanied by incisive studies of B-cell function and T-cell function that will provide insights into the disease mechanism and into the therapeutic potential of B-cell depletion.