Receptor editing and genetic variability in human autoreactive B cells

The mechanisms by which B cells undergo tolerance, such as receptor editing, clonal deletion, and anergy, have been established in mice. However, corroborating these mechanisms in humans remains challenging. To study how autoreactive human B cells undergo tolerance, we developed a novel humanized mouse model. Mice expressing an anti–human Igκ membrane protein to serve as a ubiquitous neo self-antigen (Ag) were transplanted with a human immune system. By following the fate of self-reactive human κ⁺ B cells relative to nonautoreactive λ⁺ cells, we show that tolerance of human B cells occurs at the first site of self-Ag encounter, the bone marrow, via a combination of receptor editing and clonal deletion. Moreover, the amount of available self-Ag and the genetics of the cord blood donor dictate the levels of central tolerance and autoreactive B cells in the periphery. Thus, this model can be useful for studying specific mechanisms of human B cell tolerance and to reveal differences in the extent of this process among human populations.B lymphocytes are essential cells in establishing immunity, yet are also known contributors to autoimmune diseases. At least half of newly generated B cells are self-reactive (Grandien et al., 1994; Wardemann et al., 2003), and various selection checkpoints are enforced along B cell development and maturation pathways to increase immune function in host defense while preserving self-integrity (Shlomchik, 2008; Goodnow et al., 2010). Over the past several decades, we have acquired a greater understanding of how this selection operates, but more so in mice than in humans. BCR transgenic (Tg) or knock-in mouse models, in which the majority of the B cells harbor a single specificity that can be traced, have greatly aided in elucidating mechanisms of murine B cell selection (reviewed in Goodnow et al., 1995, 2010; Aït-Azzouzene et al., 2004; Pelanda and Torres, 2006, 2012; Kumar and Mohan, 2008; Shlomchik, 2008). These studies have shown that developing, self-reactive mouse B cells have several potential fates: one is to ignore antigen (Ag) if it is either sequestered or at a concentration too low for reactivity, a second is to become anergic (i.e., nonfunctional), a third is to undergo receptor editing, and a fourth is to undergo apoptosis. A fifth fate is to undergo positive selection to low-avidity self-Ags, an outcome accompanied by the differentiation into marginal zone or B1 B cells (Hayakawa et al., 1999; Martin and Kearney, 2000; Wen et al., 2005). Which particular mechanism is invoked depends on both the strength of the signal the self-reactive BCR receives and the developmental state of the cell (Goodnow et al., 1995; Kouskoff et al., 2000; Qian et al., 2001; Aït-Azzouzene et al., 2004; Hippen et al., 2005; Wen et al., 2005; Diz et al., 2008; Andrews et al., 2013). Moreover, depending on the location of the self-Ag, tolerance is defined as central (i.e., in the bone marrow) or peripheral (i.e., in other tissues).A criticism of using BCR Tg or knock-in mice for studying B cell selection is that these models hasten B cell development, restrict the B cell repertoire, and, sometimes (e.g., in some conventional Ig Tgs), express nonphysiological levels of BCR. These issues have been addressed by creating mice that express an Igκ reactive self-Ag, enabling studies of tolerance in B cells developing with a wild-type antibody (Ab) repertoire (Ait-Azzouzene et al., 2005). This and other similar Tg models have confirmed that even wild-type murine B cells use deletion, anergy, and receptor editing for the establishment of tolerance (Ait-Azzouzene et al., 2005; Aït-Azzouzene et al., 2006; Duong et al., 2010, 2011; Ota et al., 2011).The mechanisms that operate in humans to implement B cell tolerance have been more difficult to dissect, as human bone marrow tissue is less readily accessible, and determining the fate of any particular B cell with its own unique specificity is quite challenging. Therefore, human B cell tolerance studies have focused on measuring frequencies of a panel of defined autoreactive or polyreactive B cell specificities mainly in the blood and in few bone marrow samples of healthy individuals or patients with autoimmunity (reviewed in Meffre and Wardemann, 2008; Meffre, 2011). Although these studies confirm that selection processes occur during human B cell development and with checkpoints similar to those established in mice, they have done little to determine the exact mechanisms of tolerance induction. This is particularly true for mechanisms of central B cell tolerance.Immunodeficient mice transplanted with human hematopoietic stem cells (HSCs) provide a tool to study the human immune system in greater depth (Manz and Di Santo, 2009; Ito et al., 2012; Shultz et al., 2012). By using immunodeficient mice of the BALB/c-Rag2^nullIL2Rγ^null strain (BRG or BALB/c-DKO), we have previously established a robust humanized mouse (hu-mouse) model for the analysis of human B cells and their development (Lang et al., 2011, 2013). Aiming to investigate mechanisms of human B cell tolerance, in this study we modified the BRG strain by introducing a ubiquitous synthetic neo self-Ag reactive with the Igκ⁺ fraction of human B cells. We then followed the fate of the “self-reactive” human κ⁺ cells relative to the nonautoreactive λ⁺ cells and measured the level of tolerance induction in animals reconstituted with a human immune system.The results reveal the phenotype of human B cells while they undergo central tolerance and the mode of tolerance induction. They also indicate that our model can be used to explore differences in tolerance sensitivity among the human population.