For purposes of categorization, autoimmune diseases are often lumped together as being either systemic, as in lupus, or organ-specific, as in multiple sclerosis or insulin-dependent diabetes mellitus. These categorizations have allowed researchers to understand the common principles underlying each class of autoimmune disease. Upon closer examination, autoimmune diseases represent a far more nuanced continuum of systemic and site-targeted reactivities. One such disease, rheumatoid arthritis (RA), is site-specific in the sense that painful, chronic joint inflammation remains the common, defining pathology. However, autoimmune phenomena in RA can include responses not necessarily specific for the joint; RA patients can develop autoantibodies more reminiscent of systemic autoimmunities in that they are directed towards ubiquitous antigens.

The usefulness of understanding genetics in the development of autoimmune disease is amply demonstrated by recent discoveries by Wakeland et al. who have been studying a murine model of Systemic Lupus Erythematosus (SLE). The NZM2410 mouse is a recombinant inbred mouse derived from the lupus prone NZM and NZB lineages, and demonstrates early expression of anti-nuclear autoantibodies and severe glomerulonephritis which are features of human SLE. Over the past ten years, classic genetic analysis of this mouse has demonstrated three major loci, Sle1, Sle2, and Sle3, which are essential for the inheritance of this disease. In earlier work, Wakeland et al have demonstrated that disease can be reconstituted in the ordinarily healthy C57Bl6 mouse through the introduction of these three loci from the NZM2410 background, fulfilling the genetic equivalent of Koch's postulates for transmission of a disease. Furthermore, introduction of certain pairs of loci, such as Sle1 and Sle2, or Sle1 and Sle3, but not Sle2 and Sle3 into the C57Bl6 background can recapitulate some, albeit milder forms of the disease. Interestingly, although expression of Sle1, Sle2 or Sle3 individually in the C57Bl6 background results in no apparent disease, there are noticeable immunologic abnormalities. For example, B6 mice with Sle1 alone produce antinuclear autoantibodies, and mice with Sle2 alone have hyperreactive B Cells and similarly, mice with Sle3 alone demonstrate T cell hyperreactivity.

The laboratories of Diane Mathis and Christophe Benoist pioneered a model of RA in mice that has enhanced our understanding of RA pathogenesis. This mouse strain, the K/BxN, has an engineered receptor on its T lymphocytes that recognizes the enzyme glucose 6-phosphate isomerase (GPI). This mouse develops a severe, chronic, progressive joint inflammation that carries many features of human rheumatoid arthritis. Early work in this model described mechanisms whereby initial activation of T and B lymphocytes lead to the production of arthritogenic immunoglobulins, antibodies specific for GPI that can cause disease when transferred to non-arthritic mice.

However, a glaring question left unclear by this early work concerned the nature of the proposed autoantigen, GPI. GPI is found virtually everywhere in the mammalian body; its function is integral to the basic metabolism of nearly all cells. How can an immune response against GPI cause inflammatory disease in the joint, and nowhere else that GPI is found? Mathis and colleagues address this question in a publication in the April 2002 issue of Nature Immunology.

These investigators first determined that GPI in the joint is not different from GPI found elsewhere, nor is it more abundant. Using immunofluorescent staining techniques, which take advantage of labeling antibodies which recognize specific molecules, the authors find that in the joint, GPI is normally located along the surfaces of the joint. However, when they transfer disease-causing antibodies, they can find antibodies depositing on top of the GPI.

They also find molecules of the complement network depositing along the surface of the joint. The complement proteins circulate through a normal host and activate in response to infection by one of three well-described processes, known as the classical, alternate and lectin pathways of complement activation. Activation of complement leads to a wide range of immune activation responses, including the attraction of other immune cells to a site of inflammation. In a paper in Immunity from February of 2002, the investigators found that components of the alternate complement cascade are important to the progression of disease.

The investigators describe that the codeposition of IgG and C3 is specific for the joint in recipients of arthritogenic K/BxN serum; C3 does not deposit alongside GPI and IgG in kidney. This joint-specific deposition of GPI-specific antibodies and C3 suggest that immune complexes may be important to disease. Immune complexes are very large aggregates of immunoglobulin and antigen that have pathogenic consequences in other disease models, such as lupus. The authors' data suggest that immune complexes would be important only if formed in the joint; immune complexes from circulation do not carry any arthritogenic potential.

The authors posit a model in which spontaneous activation of complement in the joint, alongside local immune complex formation leads to disease pathogenesis.

Commentary prepared by Jobert Barin