Immature B cells that encounter self-antigen are eliminated from the immune repertoire by negative selection. Negative selection has been proposed to take place by two distinct mechanisms: deletion by apoptosis or alteration of the antigen receptor specificity by receptor editing. While convincing evidence exists for each, the two models are inherently contradictory. In this paper, we propose a resolution to this contradiction by demonstrating that the site of first antigen encounter dictates which mechanism of negative selection is utilized. We demonstrate that the bone marrow microenvironment provides signals that block antigen-induced deletion and promote RAG reinduction. In the periphery, the absence of these signals allows the immature B cell to default to apoptosis as a result of BCR engagement.

Self-reactive immature B cells may be eliminated in the bone marrow (BM) after B cell receptor (BCR) engagement in a process known as negative selection. Immature B cells emigrating from the BM, the so-called transitional cells, remain sensitive to negative selection and are likely to be important targets of tolerance towards peripheral antigens. Transitional cells are deleted through apoptosis after BCR cross-linking in vitro. Using anti-Ig as a surrogate antigen, we determined the signaling requirements for the induction of apoptosis in transitional cells. Treatment with anti-Ig for only 20 min causes most cells to be apoptotic 16 h later. Furthermore, apoptosis of transitional cells is induced with low doses of anti-Ig while mature cell proliferation requires extended culture at 30-fold higher concentrations. For both populations of B cells, total surface Ig expression is equivalent, therefore indicating that the threshold of BCR signaling required to elicit these responses is different. T cell help can modulate B cell tolerance. However, specific help may not be available when apoptosis is triggered by a peripheral antigen. The opportunity to reverse apoptosis of transitional cells is surprisingly long. Even 8 h after anti-Ig treatment, IL-4 or anti-CD40 antibody can block apoptosis. The upper time limit of protection is concurrent with irreversibility of apoptosis as measured by DNA fragmentation. These findings indicate that B cell negative selection is more easily triggered than activation, and that the induction of apoptosis and its reversal by T cell help can be events that occur in distinct microenvironments.

Previously we defined a Thy1dull bone marrow-derived cell population that regulated fate decisions by immature B cells after Ag receptor signaling. The microenvironmental signals provided by this cell population were shown to redirect the B cell Ag receptor -induced apoptotic response of immature B cells toward continued recombination-activating gene (RAG) expression and secondary light chain recombination (receptor editing). Neither the identity of the cell responsible for this activity nor its role in immature B cell development in vivo were addressed by these previous studies. Here we show that this protective microenvironmental niche is defined by the presence of a novel Thy1dull, DX5pos cell that can be found in close association with immature B cells in vivo. Depletion of this cell eliminates the anti-apoptotic effect of bone marrow in vitro and leads to a significant decrease in the number and frequency of bone marrow immature B cells in vivo. We propose that, just as the bone marrow environment is essential for the survival and progression of pro-B and pre-B cells through their respective developmental checkpoints, this cellular niche regulates the progression of immature stage B cells through negative selection.

Our laboratory is interested in a variety of issues related to lymphocyte development. More specifically, we have focused on the processes that regulate the decision to commit to the B lymphocyte (B cell) lineage, then the subsequent signals that are involved in maintaining this commitment to the B cell lineage. These signals result in the positive selection of those B cells that properly execute the complex genetic changes associated with B cell development, then trigger the elimination of B cells that are responsive to self-antigens and, therefore, possess the potential to mediate autoimmune disease. Our general experimental approach has been to address these issues from the perspective of signal transduction. Our goal is to define the biochemical and genetic processes that are integrated in order to accomplish these selection processes. To do so, we employ in vivo animal models as well as more defined in vitro studies, using both primary and transformed cell lines. For the past several years, we have been primarily interested in the precise mechanisms by which the B cell antigen receptor (BCR), and intermediate forms of this receptor, regulate these complex developmental processes. We have used the ongoing studies described below as two representative examples of how we are approaching these issues and some of the insights that we have made. To place both of these studies in context, we will begin with a brief introduction into B cell development.

Antibodies (Ab) to acetylcholine receptors ( AchR) are instrumental in bringing about the neuromuscular defect in myasthenia gravis (MG) . The demonstration of Ab by immunoprecipitation assay proved to be highly specific, of great diagnostic value, and in good correlation to clinical improvement in MG patients on plasma exchange therapy.l, Only little information is available in the long term correlation between intraindividual AB titers and clinical course.2 This study presents a group of 99 MG patients (41 male, 58 female; mean age 45, five years) followed up for 36 months, using a new, simple and reproducible MG-Score (see TABLE 1 ) and serial measurements of Ab titers by a modified immunoprecipitation assay.3To calculate the correlation of Ab titer and clinical course, relative changes of MG-Score in individual patients and the differences in normalized Ab titers of the identical interval were plotted against each other. Each interval of usually one to two months was analyzed separately. (Bivariate plotting program BMDP6D by H.C.S.F. U.C.L.A. 1977, adapted to a Siemens 4004/151 computer.) The correlation is calculated in different groups of MG patients, referring to age, sex and duration as well as to MG patients with mild (MG-Score <1.0) and marked (MG-Score >1 .O j disease.We found a high correlation between Ab titer and clinical course in serial measurements of patients with moderate and marked MG. In contrast there was no correlation in patients with mild disease. No influence of age and sex of MG-patients or duration of the disease upon this correlation could be demonstrated. Long-term observations of individual patients generally showed a close correlation between MG-Score and serially determined Ab titer. In about 10% of the severely affected MG patients however rising Ab titers could be demonstrated without immediate increase in MG-Score. In those patients deterioration could be detected within three weeks to three months. The MG-Score used in this study is a simple and reliable tool for the clinical follow-up of MG patient as well as for long term investigations. From our * Correspondence address for U. A. Besinger includes Moehlstrasse 28 8000 Munich