B cells arise in the bone marrow and go on to produce antibodies that protect against microbial infection. Surprisingly, it seems that B-cell development also occurs in the gut, where it is stimulated by resident microbes. See Letter p.112
Mammalian intestines contain vast numbers of immune cells. This is not surprising, given that the gut lumen is home to a microbial ecosystem with cell numbers1 that exceed those of its human host by around 100 to 1. Microbial colonization of the gut begins at birth, and gut microbes make a crucial contribution to host metabolism. Appreciation is also growing of the role of the microbiome in shaping the host immune system2. For example, differentiation of a type of immune cell called TH17 cells requires the presence of certain bacteria in the gut3. Conversely, mice that have no resident microorganisms have perturbed immune responses and a diminished amount of lymphoid tissue, which is an integral part of the immune system. A study by Wesemann et al.4, reported on page 112 of this issue, extends the spectrum of these mutualistic interactions by reporting that B cells can undergo development in the mouse intestine, and that this is influenced by resident non-pathogenic bacteria*.
B cells produce antibodies. These bind specifically to a wide array of antigen molecules and flag them for elimination by other immune cells. Antibodies consist of a heterotetramer of proteins: two identical immunoglobulin heavy (IgH) chains and two identical immunoglobulin light (IgL) chains. The genes encoding these proteins exist as segments that are assembled into functional genes during B-cell development by a process known as V(D)J recombination5. Permutations in the combinations of these segments allow the genome to encode a vast repertoire of antigen-specific antibody molecules. The standard model of mammalian B-cell development holds that B cells arise from haematopoietic stem cells (which can give rise to all types of blood cell), in the liver during fetal life and in the bone marrow after birth. Once a B cell has produced a functional antibody molecule, this is transported to the cell membrane, where it serves a surveillance function as the cell's receptor for antigen.
The enormous diversity of antibodies generated by this process creates the problem of self tolerance. But several mechanisms exist to avoid the production of antibodies that attack normal host tissues, including clonal deletion, clonal anergy and receptor editing6. Receptor editing occurs in immature B cells (a late stage of B-cell development) and involves rearrangement of IgL-chain gene segments and IgL-chain replacement until a receptor is produced that does not recognize a self-antigen (Fig. 1a). In the bone marrow, which is sterile, receptor editing is stimulated by self-antigens. Wesemann and colleagues' results suggest that microbial antigens might also drive receptor editing in the lamina propria, an immune-cell-rich tissue layer in the mucous membranes that line body cavities, such as the gut and respiratory tract.
The authors studied mice that expressed a fluorescent-protein gene from within the Rag2 gene. The RAG2 protein, in a complex with RAG1, comprises the recombinase enzyme that is necessary for V(D)J recombination, and is expressed almost exclusively in developing T and B cells7. They observed a small but significant fraction of RAG2-expressing developing B cells in the intestinal lamina propria. Some of these cells also expressed the marker enzyme TdT, which suggested that they were at the pro-B-cell stage of development. But, perhaps more interestingly, most of the cells were at either the pre-B or immature developmental stages, meaning that they were undergoing IgL-chain gene-segment rearrangement or receptor editing (Fig. 1b). In addition to expressing the recombinase, these cells expressed reaction intermediates that also indicated ongoing recombination8. The IgH-chain repertoire of the cells resembled that of B cells elsewhere in the body, but their IgL-chain repertoire was distinctive.
In germ-free mice, which have extremely low bacterial loads in their guts, the authors found that gut-developing B cells were present, but that their numbers peaked when the mice were about 3 weeks old, suggesting that the process peters out in the absence of ongoing microbial stimulation. However, when the mice were colonized with gut microorganisms or stimulated with inflammatory molecules, the numbers of gut-developing B cells increased significantly.
The most straightforward interpretation of these results is that B-cell-receptor editing occurs in immature B cells developing in the gut, and that this is driven by interactions with antigens from resident microorganisms. Although the lamina propria is a sterile location, it is easy to imagine how either local physical trauma or inflammation might allow small numbers of microbes to penetrate the mucosal surface and induce these interactions. It is also possible that specialized mechanisms exist whereby the developing B cells could 'sample' the gut microbiome without breaching the mucosal barrier; such mechanisms are used by M cells that overlie specialized lymphoid structures known as Peyer's patches9.
That aspects of antibody-gene diversification occur outside the bone marrow is not without precedent. In rabbits, for example, V(D)J recombination during fetal life produces a repertoire of modest diversity that then undergoes significant diversification in the animals' appendix through a gene-conversion process that uses information stored in vast numbers of pseudogenes10.
Wesemann and colleagues' work raises a number of intriguing questions, most importantly, how this mode of B-cell development contributes to immune homeostasis. One possibility is that receptor editing driven by the vast array of microbial antigens in the gut substantially enhances overall antibody diversity. Alternatively, receptor editing in the gut might serve to decrease B-cell reactivity to the antigens of the most prevalent gut bacteria and thus reduce the possibility of inflammation in the bowel. Finally, B-cell development at this site might facilitate tolerance against self-antigens that are predominantly expressed there.
Other key questions include whether haematopoietic stem cells reside in the lamina propria and, if not, which B-cell precursors seed the gut, and when. Do these gut-derived B cells contribute to the pool of antibodies that is present in the body's serum, or do they exist predominantly in the gut lumen? Furthermore, this research was conducted in mice, so it remains to be determined whether B-cell development and receptor editing in the lamina propria also occur in humans. If they do, this surprising observation will have several implications for our understanding of human health and disease.
*This article and the paper under discussion4 were published online on 21 August 2013.