Cutting Edge: Transitional T3 B Cells Do Not Give Rise to Mature B Cells, Have Undergone Selection, and Are Reduced in Murine Lupus¹

Newly formed B cells that have passed through bone marrow (BM)³ tolerance checkpoints and express a functional BCR migrate to the spleen to complete development. These recent emigrants are known as transitional B cells. Of the ∼2 × 10⁷ surface (s) IgM⁺ cells produced in murine BM each day, only 10% enter the spleen and, furthermore, only 1–3% reach maturity (1). Transitional populations of B cells have been described previously (2, 3) and can be divided into three populations designated as T1 (CD93⁺sIgM^highCD23⁻), T2 (CD93⁺ IgM^highCD23⁺), and T3 (CD93⁺sIgM^lowCD23⁺) (4).

Continuous in vivo BrdU labeling further supports a developmental sequence in which T1 B cells give rise to T2 B cells, and T2 cells give rise to T3 B cells (4). DNA content analysis revealed that the transitional subsets do not undergo significant proliferation, suggesting that peripheral development is not associated with a proliferative burst (4). Adoptive transfer studies have shown that T2 cells can give rise to mature follicular B cells (MB) in wild-type recipient mice and marginal zone (MZ) B cells in lymphopenic recipients (5). CD21^int/highCD24^high cells, which likely encompass T2 cells, T3 cells, and MZ precursors, undergo maturation in vitro when given soluble B cell-activating factor (BAFF) in the presence of anti-IgM BCR stimulation (6). However, details concerning the role of the T3 subset in B cell development and tolerance are still unclear.

The T1 stage has been shown to be an important deletion checkpoint, as supported by studies showing that BCR cross-linking induces exaggerated apoptosis at this stage (2, 7, 8). T2 cells are also susceptible to anti-IgM-mediated apoptosis, consistent with susceptibility of T2 cells to deletional tolerance (4). Mice overexpressing or underexpressing BAFF, an important B cell survival factor (6), have also revealed the importance of transitional B cell tolerance checkpoints. BAFF-deficient mice have a developmental block at the T1 stage. In contrast, BAFF–transgenic mice have enhanced transitional and MZ B cell populations and develop a lupus-like disorder (9, 10). It has been demonstrated that excess soluble BAFF can rescue the lower affinity hen egg lysozyme (HEL)-reactive CD21^int/highCD23^high population, which may include T2 and T3 B cells in HEL-transgenic mice. These self-reactive B cells were allowed to enter forbidden splenic microenvironments from which they are normally excluded (10). BAFF did not rescue higher affinity HEL-reactive B cells deleted in the BM; thus, excess BAFF specifically perturbs transitional B cell tolerance (10).

Finally, a significant increase in the number of autoreactive B cells in the peripheral naive mature compartment compared with newly emigrated peripheral B cells in human systemic lupus erythematosus (SLE) further suggests that transitional B cell tolerance is faulty in human SLE (11). Together, these studies highlight the importance of transitional B cell tolerance in regulating autoimmunity. The lack of knowledge concerning the relative role that T3 B cells play in B cell development and tolerance led us to further characterize this population and to determine whether it can give rise to MB cells.

Four- to 5-wk-old C57BL/6 (B6) female mice were used as a source for splenic B cell populations for maturation, Ca²⁺ flux, and H chain usage studies. For quantitation of B cell subsets, 8- to 10-wk-old lupus-prone (BxSB/MpJ, NZM2410, NZBWF₁, and MRL/MpJ, MRL/^lpr/lpr) and control strains (BALB/c, B6, AKR/J, and 129/SvJ) were used. B6.SJL-Ptprc^aPepc^b/BoyJ (B6.Ly5^SJL) were used at the age of 4–5 wk. All mice were purchased from The Jackson Laboratory). All studies were approved by the Oklahoma Medical Research Foundation Institutional Animal Care and Use Committee.

Abs used were as follows: anti-mouse IgM F(ab′)₂-FITC or -Cy5 (Caltag Laboratories), anti-mouse CD23-PE (B3B4), anti-mouse B220 (RA3-6B2)-PerCP or PE-Cy5, and anti-mouse CD93 (493)-biotin. Biotinylated Abs were detected with secondary streptavidin (SA)-allophycocyanin, SA-allophycocyanin-Cy7, or SA-allophycocyanin-Cy5.5 (Molecular Probes). All Abs were purchased from BD Pharmingen, with the exceptions noted above. Anti-CD16/32 was used to block FcRs before staining 2 × 10⁶ cells, unless otherwise noted. Events were collected on FACSCalibur or LSRII instruments (BD Biosciences).

Splenic B cell subsets were isolated on a MoFlo cell sorter (DakoCytomation) or FACSAria (BD Biosciences). Purity of FACSAria-sorted cells ranged from 89–98%, and purity of MoFlo-sorted cells ranged between 50 and 70%. Gating strategy is shown in Fig. 1. An anti-IgM F(ab′)₂ was used during purification to avoid BCR cross-linking.

Maturation assays were performed as previously described (6). Sorted cells were incubated at 3 × 10⁶ cells/ml in 96-well microtiter plates in DMEM supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM glutamine, 2 mM sodium pyruvate, 0.1 mM Nonessential amino acids solution, and 50 μM 2-ME. Cells were stimulated for 72 h with 1.5 μg/ml recombinant soluble human myc-tagged BAFF, provided by Dr. S. Kalled (Biogen Idec, Cambridge, MA) (12). This treatment was shown to eliminate any contamination of MB cells by the 50-h time point (6). F(ab′)₂ rabbit anti-mouse Ig (Zymed Laboratories) was added to cultures at a final concentration of 10 μg/ml for the last 24 h of culture. At 72 h, cells were stained with mAbs to IgM, B220, and CD93 and analyzed by flow cytometry.

Sorted cells from B6 (CD45.2) donors were resuspended at 250,000 cells/250 μl of PBS containing 0.5% FCS per recipient. Cells were adoptively transferred into B6.Ly5^SJL (CD45.1) hosts by way of retro-orbital plexus. Ninety-six hours after transfer, B6.Ly5^SJL hosts were euthanized and the phenotype of CD45.1⁻CD45.2⁺ or CD45.1⁻ donor B cells was assessed using mAbs to IgM, B220, and CD93.

Sorted cells (1 × 10⁶) from each population (T1, T2, T3, and MB) were resuspended in 1 ml of HBSS with Ca²⁺ (Sigma-Aldrich) and loaded with 5 μM Indo-1 acetoxymethyl dye in the presence of 2% pluronic F-127 (Molecular Probes). Cells were washed twice, resuspended, and stimulated with 50 μg/ml F(ab′)₂ rabbit anti-mouse Ig (Zymed Laboratories). Ratiometric measurements of free intracellular Ca²⁺ concentrations were obtained by a Photon Technology International QuantaMaster spectrofluorometer equipped with an excitation monochromator set at 350 nm and two emission monochromators set at 405 and 485 nm.

Presorted cells were resorted into 96-well plates, subjected to RT-PCR, and sequenced using recently described primers and procedures (13).

In vitro studies have shown that immature B cells give rise to mature B cell populations; however, reports using similar strategies to determine the developmental fate of T3 B cells are not available. It was recently reported that BAFF delivered in combination with BCR stimulation enabled the maturation of CD21^int/highCD24^high immature B cells into a MB phenotype (6). We used a similar approach to determine whether the T2 and T3 populations can give rise to MB cells. As expected, BCR stimulation in the presence of BAFF led to maturation of a subset of the immature T2 population (Fig. 2). However, no maturation of the immature T3 population was observed as assessed by cell surface retention of the immature marker CD93 under similar conditions (Fig. 2). Although the purity (see Materials and Methods) and overall viability of isolated T2 and T3 cells varied with the sorter used, the degree of purity of the two subsets within a single experiment was always similar. Moreover, similar results were obtained in each experiment performed, as shown in Fig. 2 B, which depicts results from one experiment using lower purity/higher viability MoFlo-sorted cells and from two experiments using higher purity/lower viability FACSAria-sorted cells. Even though BAFF receptor levels were marginally lower in T3 cells compared with T2 B cell (data not shown), there was no significant difference in the number of live-gated cells recovered between T2 and T3 cultures under conditions of BAFF alone vs BAFF plus anti-sIgM (p = 0.127 for three independent experiments, Kruskal-Wallis test). This indicated that increases in the fraction of mature cells in T2 cultures reflected phenotypic changes and not differential cell death.

T2 B cells adoptively transferred into wild-type mice have been shown to acquire a MB phenotype within 4.5 days after transfer (5). We predicted that under similar conditions, T3 B cells would remain phenotypically immature. To address this question, we performed adoptive transfer experiments (5). Two hundred fifty thousand T2 and T3 B cells were sorted from B6 (CD45.2) donors and transferred into B6.Ly5^SJL (CD45.1) hosts. Four days after transfer, the hosts were euthanized and the phenotype of donor cells was assessed. Quantitation of the CD45.1⁻CD45.2⁺ donor cells positive for B220 expression revealed that a portion of donor T2 cells assume a mature phenotype as evidenced by loss of CD93 expression and down-regulation of cell surface IgM, whereas essentially all donor T3 cells remain CD93⁺. These results demonstrate that T2 cells but not T3 cells acquire a mature phenotype in vivo (Fig. 3). These results lead us to conclude that the T3 population makes at most a limited contribution to the mature B cell pool and does not represent a normal stage B cell development.

Studies have shown that T1 and T2 immature B cells mobilize Ca²⁺ to a greater extent than mature B cells following BCR cross-linking (14), whereas anergic B cells have a reduced mobilization of Ca²⁺ (15, 16). To examine the relative capacity of T3 B cells to respond to BCR cross-linking, we sorted the immature transitional cells into T1, T2, T3, and MB phenotypes, then measured Ca²⁺ mobilization following loading of cells with Indo-1. In response to BCR cross-linking, we found a reproducibly reduced Ca²⁺ flux to anti-IgM BCR stimulation in the T3 population when compared to T1, T2, and MB B cell populations (Fig. 4,A). T1 and T2 cells both mobilized Ca²⁺ more efficiently than mature B cells as predicted. All populations fluxed Ca²⁺ to similar extents with the addition of ionomycin, demonstrating efficient Indo-1 loading in all populations (Fig. 4 A). Similar results were observed in Fluo-4-labeled cells by flow cytometry (data not shown). Our results are consistent with the contention that the T3 population exhibits an anergic phenotype in response to BCR cross-linking (17).

To determine whether T3 cells show any evidence of selection away from the T1→T2→MB pathway, we performed single-cell RT-PCR on the T1, T2, T3, and MB populations to analyze their Ig V_H/J_H gene usage. Interestingly, we observed a significant enhancement of J_H3 segment usage in the T3 population (46 ± 13%), compared with T2 (17 ± 10%; p = 0.022) and mature cells having a follicular phenotype (17 ± 5%; p = 0.027; Fig. 4 B). A predominance of J_H6 usage has been observed within an edited autoreactive B cell subset in humans (18); however, no association between particular J_H segments and potentially autoreactive B cell subsets has been described in the mouse. To determine whether J_H3 might demonstrate an association with autoantibodies in the mouse, the Immunogenetics database was exhaustively searched for murine Abs directed to self-Ags or nonrodent Ags. Significantly more autoantibodies (38 of 110; 35%) than Abs to nonrodent Ags (64 of 262, 24%) used the J_H3 segment (p = 0.014; Fisher’s exact test), suggesting an association of J_H3 with autoreactivity (Supplemental Tables Ia and Ib). ⁴ Thus, biased J_H3 usage within the T3 subset might be explained by enrichment of this subset with autoreactive clones. These data are consistent with T3 B cells being anergic and autoreactive.

Given the impaired BCR responsiveness of the T3 population, reports of reversal of B cell anergy (19) led us to hypothesize that a tolerance defect at this selection point would most likely manifest itself as a significant reduction in the T3 population in autoimmune strains of mice. To explore this possibility, we enumerated T1, T2, T3, and MB cells in lupus-prone (BxSB/MpJ, NZBWF₁ NZM2410, MRL/^lprlpr, and MRL/MpJ) and normal inbred (BALB/c, B6, AKR/J, and 129/SvJ) strains of mice. Groups (n = 10) of mice were examined at 8–10 wk of age to reduce the influence of chronic inflammation on B cell numbers and to enhance the likelihood of detecting primary defects in B cell development, selection, or homeostasis. Examination of total spleen cellularity among the strains tested revealed two statistically separable groups (Fig. 5,A) that were consistent with strain-to-strain variation in spleen size. We analyzed the two groups separately and compared the lupus prone strains to the non-autoimmune control strain harboring the closest number of T3 B cells. In the large spleen group, the T3 population in MRL/MpJ and MRL/^lprlpr mice was significantly reduced compared to the T3 population in the AKR/J strain (closest normal control) (Fig. 5,B). In the small spleen group, significantly reduced numbers of T3 B cells were also observed in both NZM2410 and NZBWF₁ strains compared with 129/SvJ mice (Fig. 5 B). No reduction in T3 B cell numbers was observed in BXSB males.

To detect potential selection defects at earlier developmental stages, the numbers of T1 and T2 B cells among lupus-prone and control mice were also examined (Fig. 5, C and D). Both MRL/MpJ and MRL/^lprlpr mice exhibited reductions in the T1 and T2 populations in addition to the T3 population. Interestingly, there were no significant differences in the sizes of the T1 and T2 populations in NZBWF₁ and NZM2410 strains compared with normal strains, indicating specific reductions in the T3 compartment in these two genetically related strains. These data suggest a possible genetic etiology of reduced T3 B cell numbers in NZB/W-related strains consistent with either impaired T3 cell development or anergy. No significant differences in the numbers of MB cells were detected among the strains tested and no significant differences in MZ B cell numbers were observed in NZBWF₁ or NZM2410 strains compared with controls (data not shown). However, it is likely that the MB compartments of the autoimmune-prone strains harbor enhanced numbers of autoreactive cells.

Cumulatively, these data strongly suggest that the T3 transitional B cell population is not a normal stage of B cell development and is unlikely to give rise to mature B cells under normal conditions. The BAFF in vitro maturation assays and in vivo adoptive transfer assays demonstrate that T3 B cells, contrary to the T2 population, remain phenotypically immature when given survival factors plus BCR stimulation or when placed into a normal peripheral environment in the mouse. Furthermore, defective mobilization of Ca²⁺ in response to BCR stimulation and a biased J_H gene usage provide evidence that the T3 population harbors cells of low responsiveness that have been selected away from the T2→MB pathway. The observations that J_H3 usage is associated with autoreactivity and T3 B cells are selectively reduced in two genetically related lupus-prone models, along with recent data highlighting defective peripheral tolerance checkpoints in human SLE patients, suggest that T3 B cells are likely to be important in the regulation of autoimmunity.

We thank Dr. Susan Kalled for providing BAFF, the staff of the Oklahoma Medical Research Foundation Flow Cytometry Core, Andy Duty for support with calcium signaling, and Karen Schwarz for mouse husbandry. We also thank Drs. Mark Coggeshall, Carol Webb, and Shannon Maier for helpful comments on this manuscript.

The authors have no financial conflict of interest.