We next asked how M-B cells develop in the BM. After LPS treatment, the frequency of CD11b+ B cells significantly increased early in the BM but not in the blood, spleen, or peritoneal cavity (Fig. 3 a), suggesting that M-B cells are derived from B cell precursors that represent about half of CD19+B220+ cells in the BM of naive mice (Fig. 1, e and f; and Fig. 3 b). Upon LPS treatment, the number of B cell precursors dramatically decreased and the number of mature B cells conversely increased (Fig. 1, e and f), implying the immediate differentiation of B cell precursors into mature B cells. Indeed, pre-B cells that had been transplanted into the BM differentiated into mature B cells within 36 h after LPS treatment (Fig. S2, a and b). To identify the main source of M-B cells, mature, immature, pre-, or pro-B cells isolated from CD45.2+, WT mice were transplanted into the BM of CD45.1+ B6.SJL-ptprca (B6.SJL) mice, and CD11b expression by the donor-derived B cells was examined 48 h after PBS or LPS injection (Fig. 3, c and d; and Fig. S2 c). The expression of CD11b was low after PBS treatment of mice transplanted with any of the donor cells (Fig. 3, c and d). However, upon LPS treatment, CD11b-expressing M-B cells emerged at similar frequencies in each population of donor cell–derived B cells (Fig. 3, c and d), which suggests that M-B cells are derived from all stages of BM-B cell lineages after LPS treatment. We next examined the retention capacity of mature, immature, and pre-B cells after LPS treatment. A 1:1 mixture of pre-B cells from CD45.2+ WT mice and pre-B cells, immature B cells, or mature B cells from CD45.2+ CAG-EGFP mice was injected into the BM of CD45.1+ B6.SJL mice (intra BM injection, IBI), and the ratio of EGFP+ and WT donor cells in the BM was compared 20 h after LPS treatment (Fig. 3 e). As a control, a 1:1 mixture of EGFP+ pre-B cells and EGFP− pre-B cells was equally retained in the BM after IBI and LPS treatment (Fig. 3, f and g). In contrast, the frequency of EGFP+ immature or mature B cells was dramatically decreased in the BM compared with EGFP− pre-B cells (Fig. 3, f and g). In contrast, when EGFP− pre-B cells and EGFP+ mature B cells were mixed at 1:1 and transplanted into CD45.1+ B6.SJL mice (Fig. S2, d and e), the frequency of EGFP+ mature B cells became significantly higher than that of EGFP− pre-B cells in the spleen after LPS treatment (Fig. S2, f and g), which suggests that the capacity of pre-B cells to stay in the BM was much higher than mature and immature B cells after LPS treatment. Considering that immature and mature B cells rapidly leave the BM after LPS administration, B cell precursors containing pre- and pro-B cells seem to be a major source of M-B cells in the BM.

To examine the mechanism of M-B cell induction, we tested whether TLR4-mediated signaling directly induces M-B cell differentiation from B cell precursors. We found that both TLR4-sufficient and TLR4-deficient pre-/pro-B cells transplanted into the BM of recipient mice also clearly express CD11b after LPS treatment (Fig. S2, h–j), suggesting that TLR4-mediated signaling on B cells is not essential for M-B cell generation. Rather, inflammatory cytokines indirectly upregulated in other cells by LPS likely induce M-B cells.