To the Editor:
In a recent issue of The Journal of Experimental Medicine, Thomas Rothstein and colleagues, a group with long-standing expertise in the field of mouse B1 cells, reported the description of a B1 B cell subset in human blood, a population that has thus far eluded identification (Griffin et al., 2011).
Mouse B1 cells are the main constituents of the B lymphocyte pool in pleural and peritoneal cavities, and are characterized as CD5+ and/or CD11b+ cells (Hardy, 2006; Baumgarth, 2011). However, CD5, which was initially identified on chronic lymphocytic leukemia tumors (Boumsell et al., 1978), is not a B1 marker in humans. Human CD5 marks immature/transitional B cells in bone marrow, blood, and spleen (Sims et al., 2005; Cuss et al., 2006), as well as in T cells. Mouse B1 cells are also described as IgMhighIgDlowCD43+. The function of CD43 in B1 cells is unknown, and it is also expressed by T lymphocytes, B cell precursors, and plasma cells.
Griffin et al. (2011) described a CD20+CD27+CD43+CD70− subset present in adult and human cord blood with functional characteristics that they describe as typical B1 cell attributes: spontaneous IgM secretion, constitutive BCR signaling, and ability to drive allogeneic T cell proliferation. It should be noted that this last feature has been shown to be displayed by switched memory B cells in humans, likely because of the high expression of CD80 and/or CD86 on these cells (Liu et al., 1995; Good et al., 2009).
A striking point of the observations of Griffin et al. (2011) was their quantitative aspect. Despite considerable variations between individual blood donors, the proportion of CD43+ cells among CD27+ B cells averaged 40–50% in adults, with a higher frequency in young individuals, and a lower frequency in the elderly. We find this puzzling, as these quantitative figures closely match the frequency of marginal zone–like (or IgM memory) B cells (Weill et al., 2009), with the difference being that marginal zone–like B cells are IgDint whereas CD43+ B cells are mainly IgDhigh (Griffin et al., 2011). Therefore, we analyzed to what extent these two populations may superimpose.
CD43+ B cells appear as large cells that require a wide lymphocyte gate to be detected. In so doing, the risk of inclusion of cell doublets in the analysis/isolation is high. Usually, specific selective criteria on the cell flow (FSC-W, SSC-W) are applied to remove doublets, unless the morphological characteristics of the cell population justify such an omission. In such cases, careful controls are obviously required to avoid the confusion between cell doublets and large cells.
We added anti-CD3 antibodies to the staining reaction, as T cells are the major lymphocyte subset in human blood compared with B cells (95:5). A large fraction of CD20+CD27+CD43+ cells stained positive for CD3 (Fig. 1 A). We believe that these CD20+CD27+CD43+CD3+ cells are doublets involving T cells that account for the CD43 and CD27 labeling, and (mainly) naive B cells that account for the IgDhigh phenotype (unpublished data). Pre-enrichment of peripheral blood B cells through CD19+ selection strongly reduced the proportion of CD20+CD27+CD43+ cells (Fig. 1 A).
Gating on CD20+ cells excluded plasma cells, identified as CD43++CD27++CD38++. The remaining CD20+CD3−CD27+CD43+ cells accounted for 2–3% of the total B cell pool (2.2% for adult samples, 2.8% for child samples; Fig. 1, B and C). These cells harbored either an IgD+ or an IgD− phenotype and a somewhat heterogeneous CD38 intensity (Fig. 1 B). Interestingly, the IgD+ over IgD− ratio among CD43+ cells varied with age; IgD+ cells dominated in children <5 yr, and IgD− cells dominated in adults (Fig. 1 D). IgD− cells include cells expressing IgG or IgA, as well as a minority of IgM-only cells (unpublished data). The presence of CD20+CD43+ cells displaying IgD, IgG, or IgA makes their possible B1 equivalence a more complex issue, even though B1 cells can give rise to IgA-producing plasma cells in the lamina propria (Suzuki et al., 2010).
We therefore propose that the quantitative variations observed when counting CD43+ B cells may be largely contributed by staining artifacts. The potential presence of doublet events in the CD20+CD43+ population analyzed by Griffin et al. (2011) obviously questions the in vitro functional characteristics described for this putative B1 subset. Whether CD20+CD43+CD27+IgD+ or IgD− B cells are activated cells on their way to plasma cell differentiation or a new innatelike subset with B1 functional features remains to be seen.
We thank Dr. Chantal Brouzes (Necker Hospital, Paris) for the child blood samples, and Maya Chrabieh and Mélanie Migaud (both from INSERM U980, Paris) for the adult samples. INSERM U783 is supported by the Ligue contre le Cancer (équipe labellisée) and by an ERC Advanced Grant to Jean-Claude Weill.
The authors declare they have no financial conflicts of interest.