Abstract
B lymphopoiesis in aged mice is characterized by reduced B cell precursors and an altered Ab repertoire. This likely results, in part, from reduced surrogate L chains in senescent B cell precursors and compromised pre-BCR checkpoints. Herein, we show that aged mice maintain an ordinarily minor pool of early c-kit+ pre-B cells, indicative of poor pre-BCR expression, even as pre-BCR competent early pre-B cells are significantly reduced. Therefore, in aged mice, B2 B lymphopoiesis shifts from dependency on pre-BCR expansion and selection to more pre-BCR-deficient pathways. B2 c-kit+ B cell precursors, from either young or aged mice, generate new B cells in vitro that are biased to larger size, higher levels of CD43, and decreased κ L chain expression. Notably, immature B cells in aged bone marrow exhibit a similar phenotype in vivo. We hypothesize that reduced surrogate L chain expression contributes to decreased pre-B cells in aged mice. The B2 pathway is partially blocked with limited B cell development and reduced pre-BCR expression and signaling. In old age, B2 pathways have limited surrogate L chain and increasingly generate new B cells with altered phenotype and L chain expression.
Bone marrow B lymphopoiesis is compromised, to variable extents, in aged mice of several inbred strains (1, 2, 3, 4, 5, 6). Multiple developmental stages show deficits, including pre-B cells, pro-B cells, and common lymphoid progenitors (1, 2, 3, 4, 5, 6). While the mechanisms responsible remain to be fully characterized, diminished responses of B cell precursors and progenitors to IL-7, increased apoptosis, and reduced capacity for V gene recombination have been described in aged mice (5, 7, 8, 9). We have shown that the transcriptional program underlying B lineage specification and commitment is altered in aged mice (10, 11, 12, 13). Aged B cell precursors are variably deficient in E2A expression; this results from increased turnover of E2A-encoded proteins (13, 14). E2A, together with early B cell factor, regulates transcription of the surrogate L chain genes λ5 and VpreB (15). Given the reductions in E2A, it is not surprising that B cell precursors from aged mice are often deficient in surrogate L chains (3, 11). We hypothesize that reductions in surrogate L chain level may limit formation and function of the pre-BCR in nascent pre-B cells in aged mice (10, 16). This would be expected to affect the capacities for pre-B cell expansion and further maturation in senescence.
As shown by Kawano et al. (17), different μ H chain V regions show varying capacities for binding surrogate L chains and this correlates well with pre-BCR signaling and further differentiation. Therefore, each pre-B cell, individually, undergoes further proliferation and differentiation dictated by its particular capacity for expressing the pre-BCR. In previous studies, lack of pre-BCR expression in surrogate L chain knockout (KO)3 mice does not preclude μ H chain rearrangements, but does dampen pre-B cell proliferation (18, 19, 20, 21). In λ5 gene KO mice, where pre-BCR expression cannot occur, the remaining pre-B cells have a unique phenotype and retain surface expression of the tyrosine kinase c-kit and fail to up-regulate CD25 and CD2 (19, 20). In studies by Kawano et al. (17), the extent of pre-BCR signaling coincided with down-regulation of c-kit in pre-B cells. As shown by ten Boekel et al. (22), c-kit+ pre-B cells, from either λ5 KO or wild-type (WT) mice, differ from c-kit− pre-B cells in their Vh repertoires, the former being enriched in rearranged 3′ Vh gene families. Notably, the majority of μ-chains cloned from c-kit+ pre-B cells failed to associate with surrogate L chains while μ-chains derived from c-kit− pre-B cells bound surrogate L chains and were competent to form the pre-BCR (22). Therefore, the expression (or lack thereof) of c-kit serves as a phenotypic marker of two distinct pre-B subsets that likely differ in their capacities for pre-BCR expression and signaling and, consequently, expansion and differentiation to the B cell stage.
B cell development in the bone marrow is dominated by the B2 B cell lineage, the precursors of typical follicular B cells. In aged mice, we would anticipate that poor surrogate L chain expression, by compromising the pre-BCR checkpoint, would diminish the numbers of pre-B cells generated in the bone marrow, but would allow continued, albeit less effective, B cell differentiation along a more “pre-BCR-deficient” path. Since pre-BCR-mediated selection, which is dependent upon Vh sequence, would be minimized, newly generated B cells in aged mice would be expected to have altered specificity and possibly function.
Consistent with this hypothesis, we have previously reported that aged mice show an increased proportion of immature B cells within the bone marrow that appear to have undergone activation as evidenced by altered surface phenotype (e.g., expression of CD43, CD5, CD11b, PD-1 Ags), dependence on the BCR pathway kinase Btk, increased size, and an altered Vh repertoire as evidenced by increased usage of the VhS107 family (23, 24). In this report, we show that B cells with a comparable phenotype are preferentially derived from pre-BCR deficient c-kit+ pre-B cells within the B2 lineage. Consequently, we hypothesize that the quality and quantity of new B cells derived from senescent bone marrow is impacted by reduced pre-BCR-dependent B2 B lymphopoiesis, subsequent to low surrogate L chain expression.
Materials and Methods
Mice
Young (2–4 mo) and aged (21–26 mo) BALB/c mice were purchased from the National Institutes of Aging colony at Harlan Sprague Dawley. Mice with obvious abdominal tumors and/or splenomegaly in the thoracic or abdominal cavities were eliminated from the studies. The λ5 gene KO mice were originally constructed by Kitamura et al. (18), and provided by W. Haas and K. Rajewsky, Cologne, Germany, and they and their normal littermate controls were bred in our colony.
Cell cultures
Femur and tibia pairs were flushed to harvest cells from the bone marrow as previously described (1). RBC were removed by treatment with ACK (0.15 M NH4CL, 1 mM KHCO3, and 0.1 mM EDTA) for 5 min at room temperature followed by centrifugation to remove red cell debris. Bone marrow cells were counted and used for cell sorting, flow cytometry, or cell culture. For unfractionated bone marrow culture, cells were resuspended at 1 × 106/ml in RPMI 1640 (Invitrogen), supplemented with 10% FCS (Sigma- Aldrich) plus 1% penicillin-streptomycin, 1% l-glutamine, and 2-ME at 2 × 10−5 M. Purified recombinant mouse IL-7 (BioSource International) was added at 5 ng/ml and remained in culture for 5–7 days, after which nonadherent cells were harvested and used for cell sorting and further analysis. For B cell precursors used in in vitro studies, the different subsets were sorted by fluorescence flow cytometry or magnetic bead sorting and cultured as above with recombinant mouse IL-7 (5 ng/ml) and recombinant mouse stem cell factor (SCF) (BioSource International) at 50 ng/ml for 4 days.
Fluorescence flow cytometry
Mouse bone marrow cells freshly harvested or cultured were stained with the following Abs that were labeled with an appropriate fluorochrome: IgM (II/41), CD43 (S7), B220 (RA3-6B2), CD19 (1D3), κ (187.1), λ (1-3) (R26-46) (BD Biosciences), and CD93 (AA4.1) (eBioscience). For the cytoplasmic staining of μ-chain, goat anti-mouse μ (Jackson ImmunoResearch Laboratories) and, for λ5, LM34 (BD Biosciences) Abs were used. Cells were initially stained for surface markers, permeablized using BD Cytofix/Cytoperm (BD Biosciences), washed with PermWash (BD Biosciences), and followed by cytoplasmic stain addition. Cells were analyzed within 30 min of staining. Analysis was performed on an LSR II fluorescence flow cytometer (BD Biosciences), and at least 5 × 105 events were acquired.
Cell sorting
B cell precursors (surface IgM−CD19+) from C57BL/6 (B6) and λ5 KO mice were isolated as follows: mouse bone marrow cells were first surface stained with anti-IgM allophycocyanin (II/41) followed by addition of anti-allophycocyanin magnetic microbeads according to the MiniMACS protocol (Miltenyi Biotec). IgM-negative cells were separated from IgM-positive cells, and post sorts revealed 90–95% purity. From the IgM− cells, CD19+ cells were further magnetically sorted first using anti-CD19 PE (1D3) and then anti-PE magnetic beads with a purity of >90%.
In some experiments, IL-7 cultured B cell precursors were sorted for CD2− pro-B cells by magnetic bead separation before analysis. Cells were first surface stained with anti-CD2 PE (BD Biosciences) and then stained using anti-PE microbeads according to the MiniMACS protocol. Purity of CD2− cells was >98%.
Bone marrow cells isolated from tibia and femur pairs of young and aged mice were sorted for c-kit+ and c-kit− precursors. Cells were stained for surface IgM (II/41), CD43 (S7), B220 (RA3-6B2), CD19 (1D3) (BD Biosciences), CD93 (AA4.1), and c-kit (2B8) (eBioscience). The c-kit+ precursors were defined as IgM−B220+CD43+CD19+AA4.1+c-kit+; c-kit− precursors were defined as IgM−B220+CD43+CD19+AA4.1+c-kit−. Cells were sorted with a FACSAria (BD Immunocytometry Systems) with purity ranging between 94 and 99%.
Western blotting
Cells were lysed with Mammalian Protein Extraction Reagent (Pierce) at 1 × 106 cells/10 μl. Protein Extraction reagent was supplemented with Halt Protease Inhibitor Cocktail (Pierce) at 10 μl/ml. Samples were denatured by boiling for 4 min in sample buffer, subjected to reducing conditions, and electrophoresed using SDS-PAGE 4–12% polyacrylamide gels for 50 min at 200 V. Proteins were run out on gels and then transferred onto nitrocellulose membranes for 90 min at 100 V. Nonspecific sites were blocked by incubation of the membranes with PBS-Tween 20 (1× PBS/0.05% Tween 20) containing 10% milk for 2 h at room temperature. Membranes were incubated as required with mouse monoclonal anti-actin (C-2; Santa Cruz Biotechnology) or hamster anti-λ5 mAb FS1 (3, 11, 25). Following overnight incubation with the primary Ab, immunoblots were incubated with the appropriate HRP-labeled secondary Abs for 2 h at room temperature, developed by enzyme chemiluminescence, and analyzed via an Alpha Innotech FluorChem Gel-doc system.
Statistical analysis
Student’s t test (two-tailed), either paired or unpaired, or the nonparametric Mann-Whitney U test were used as appropriate.
Results
Subsets of pre-B cells are differentially affected by senescence
Individual pre-B cells within the bone marrow differ in their capacities to assemble the pre-BCR and signal at the pre-BCR checkpoint. This is dependent upon the unique abilities of each μ H chain, inherent in its variable region, to associate with the surrogate L chain. The assembly, expression, and signaling via the pre-BCR results in a phenotypic change in newly formed pre-B cells in which the c-kit surface protein is down-regulated progressively in proportion to pre-BCR levels (17, 22, 26). This provides a well-characterized means to identify early pre-B cells within the bone marrow that are pre-BCR competent (e.g., c-kit−) and progress through the pre-BCR checkpoint and pre-B cells that likely are less effective in pre-BCR signaling (e.g., c-kitlow/−). Consistent with this, early pre-B cells present in the bone marrow of λ5 KO mice, which are deficient in the pre-BCR, are uniformly c-kit+ (Fig. 1) (19).
λ5 KO mice show loss of c-kit− but not c-kit+ early pre-B cells in vivo. A, Representative dot plots of c-kit+ and c-kit− precursors in WT (B6) and λ5 KO mice. B220+CD43+ are pro-B/early pre-B cells and late pre-B cells are B220+CD43− with percentages out of total bone marrow shown. Pro-B/early pre-B cells were further gated as surface IgM−CD19+ with percent cells present in total bone marrow indicated. Surface IgM−CD19+B220+CD43+ B cell precursors were gated based on c-kit and cytoplasmic μ (cμ) with proportion indicated for c-kit+cμ− pro-B, c-kit+cμ+ early pre-B, and c-kit−cμ+ early pre-B subsets. B, Analysis of c-kit+ pro-B, c-kit+ and c-kit− early pre-B, and late pre-B cell numbers for seven pairs of WT and λ5 KO mice. ∗, p < 0.02.
λ5 KO mice show loss of c-kit− but not c-kit+ early pre-B cells in vivo. A, Representative dot plots of c-kit+ and c-kit− precursors in WT (B6) and λ5 KO mice. B220+CD43+ are pro-B/early pre-B cells and late pre-B cells are B220+CD43− with percentages out of total bone marrow shown. Pro-B/early pre-B cells were further gated as surface IgM−CD19+ with percent cells present in total bone marrow indicated. Surface IgM−CD19+B220+CD43+ B cell precursors were gated based on c-kit and cytoplasmic μ (cμ) with proportion indicated for c-kit+cμ− pro-B, c-kit+cμ+ early pre-B, and c-kit−cμ+ early pre-B subsets. B, Analysis of c-kit+ pro-B, c-kit+ and c-kit− early pre-B, and late pre-B cell numbers for seven pairs of WT and λ5 KO mice. ∗, p < 0.02.
We and others (1, 2, 3, 4, 5, 6) have previously reported that pre-B cells are reduced in senescence; however, it is not clear whether this applies to all or only some of the identifiable subsets of pre-B cells within aged bone marrow. That aged mice might be anticipated to lose c-kit− pre-B cells, which are dependent upon pre-BCR signaling, but possibly retain c-kit+ pre-B cells is suggested by our previous observations that B cell precursors from aged mice often express lower levels of both λ5 and VpreB components of the surrogate L chain (3, 11, 13). This is confirmed and extended in Fig. 3 as discussed below.
As shown in Fig. 2, A and B, ∼30–40% of IgM−CD19+B220+CD43+cμ+ early pre-B cells in young adult BALB/c mice show expression of c-kit. In contrast, also as shown in Fig. 2, A and B, in BALB/c mice ∼20–22 mo of age, late-stage (Hardy Fraction D) (27) (CD43−B220+) pre-B cells are markedly reduced (∼33% of young controls). IgM−CD19+B220+CD43+c-kit−cμ+ (“c-kit−”) early pre-B cells are also reduced (Fig. 2, A and B); however, the numbers of IgM−CD19+B220+CD43+c-kit+cμ+ (“c-kit+”) pre-B cells are better maintained in aged bone marrow. These experiments indicate that, while late stage pre-B cells and c-kit− early pre-B cells are decreased in aged mice compared with young adult mice, c-kit+ early pre-B cells are less affected in aged bone marrow. In addition to pre-B cells, pro-B cells (IgM−CD19+B220+CD43+c-kit+cμ−) are variably reduced in aged mice (Fig. 2 B) in accordance with previous reports (3, 4, 5).
Aged mice show loss of c-kit− but not c-kit+ early pre-B cells in vivo. A, Representative dot plots of c-kit+ and c-kit− precursors in young (Yg) (2–4 mo) and aged BALB/c mice (21–26 mo). B220+CD43+ are pro-B/early pre-B cells and late pre-B cells are B220+CD43− with percentages out of total bone marrow shown. Pro-B/early pre-B cells were further gated as surface IgM−CD19+ with percent cells present in total bone marrow indicated. Surface IgM−CD19+B220+CD43+ B cell precursors were gated based on c-kit and cytoplasmic μ (cμ) with proportion indicated for c-kit+cμ− pro-B, c-kit+cμ+ early pre-B, and c-kit−cμ+ early pre-B subsets. B, Analysis of c-kit+ pro-B, c-kit+ and c-kit− early pre-B, and late pre-B cell numbers for six pairs of young and old mice. ∗, p < 0.02.
Aged mice show loss of c-kit− but not c-kit+ early pre-B cells in vivo. A, Representative dot plots of c-kit+ and c-kit− precursors in young (Yg) (2–4 mo) and aged BALB/c mice (21–26 mo). B220+CD43+ are pro-B/early pre-B cells and late pre-B cells are B220+CD43− with percentages out of total bone marrow shown. Pro-B/early pre-B cells were further gated as surface IgM−CD19+ with percent cells present in total bone marrow indicated. Surface IgM−CD19+B220+CD43+ B cell precursors were gated based on c-kit and cytoplasmic μ (cμ) with proportion indicated for c-kit+cμ− pro-B, c-kit+cμ+ early pre-B, and c-kit−cμ+ early pre-B subsets. B, Analysis of c-kit+ pro-B, c-kit+ and c-kit− early pre-B, and late pre-B cell numbers for six pairs of young and old mice. ∗, p < 0.02.
In this regard, aged bone marrow, depleted of pre-BCR-dependent c-kit− pre-B cells, begins to resemble young adult λ5 KO bone marrow where c-kit+ pre-B cells are predominant within the early pre-B pool. This is shown diagrammatically in Fig. 8 where pro-B cells in aged mice, low in surrogate L chain, generate more frequently new pre-B cells bearing c-kit at the expense of normal pre-BCR driven c-kit− pre-B cell development.
B cell precursors in aged bone marrow exhibit reduced levels of surrogate L chain protein
As shown in λ5 KO mice, deficiencies in surrogate L chain expression led to loss of c-kit− pre-B cells with retention of c-kit+ pre-B cells. We hypothesized that aged mice with increased representation of c-kit+ early pre-B cells and reduced numbers of c-kit− pre-B cells would have reduced levels of surrogate L chain. We have previously reported that B cell precursors in aged mice often show a decline in surrogate L chain expression and that this coincides with reductions in the E47 transcription factor, a product of the E2A gene necessary for optimal surrogate L chain transcription (11, 12, 13, 14). Moreover, the level of λ5 surrogate L chain expression in aged B cell precursors was significantly correlated with the extent of late stage pre-B cell loss in the bone marrow (11).
It was confirmed in the present study that reductions in surrogate L chain λ5 protein were observed, by fluorescent staining, in aged B cell precursors as early as the sIgM−CD19+B220+CD43+cμ− pro-B cell stage in vivo (Fig. 3,A). The reduced λ5 protein expression was observed in both c-kit+ and c-kit− early pre-B cells from aged mice (Fig. 3,A). These in vivo results were replicated in pro-B cell precursors grown from aged mice in the presence of IL-7 in vitro (Fig. 3 B). Therefore, abnormal decline in surrogate L chain expression is likely a relatively early event during B cell development in aged mice, occurring at or before the pro-B cell stage and extending through the pre-B cell stages.
Loss of λ5 protein in aged mice occurs at pro-B and pre-B cell stages. A, Cytoplasmic λ5 protein was stained and analyzed by flow cytometry in bone marrow c-kit+cμ− pro-B, c-kit+cμ+ early pre-B, and c-kit−cμ+ early pre-B cells in young and old BALB/c mice. Data shown are representative of three mice. Gray line, isotype control. Percent positive cells was determined with a cut-off as <2% positives in isotype controls. The change in mean fluorescent intensity (ΔMFI) was defined as (MFI experimental − MFI isotype control) and is shown in the upper left. B, λ5 protein expression was determined in CD2−cμ− pro-B cells generated from young and old BALB/c mice upon bone marrow cell culture with IL-7 for 5–7 days. CD2− pro-B cells were separated by MACS sorting before lysate preparation. Proteins were assayed by Western blot at 2.5, 5, and 10 × 105 cell equivalents. Data are representative of three experiments.
Loss of λ5 protein in aged mice occurs at pro-B and pre-B cell stages. A, Cytoplasmic λ5 protein was stained and analyzed by flow cytometry in bone marrow c-kit+cμ− pro-B, c-kit+cμ+ early pre-B, and c-kit−cμ+ early pre-B cells in young and old BALB/c mice. Data shown are representative of three mice. Gray line, isotype control. Percent positive cells was determined with a cut-off as <2% positives in isotype controls. The change in mean fluorescent intensity (ΔMFI) was defined as (MFI experimental − MFI isotype control) and is shown in the upper left. B, λ5 protein expression was determined in CD2−cμ− pro-B cells generated from young and old BALB/c mice upon bone marrow cell culture with IL-7 for 5–7 days. CD2− pro-B cells were separated by MACS sorting before lysate preparation. Proteins were assayed by Western blot at 2.5, 5, and 10 × 105 cell equivalents. Data are representative of three experiments.
Both c-kit− and c-kit+ early pre-B cells generate immature B cells in vitro but with different efficiencies
The above results indicate that low surrogate L chain levels, as are apparent in young adult λ5 gene KO mice and in WT senescent mice, are associated with altered representation of c-kit+ vs c-kit− early pre-B cells. In our previous studies, we have noted that those individual aged mice whose pre-B cell numbers are severely depleted (e.g., <20% that of young adults) also have reduced numbers of immature B cells (6, 24). Therefore, we next determined the relative capabilities of c-kit+ and c-kit− B cell precursors to generate immature B cells in vitro to determine whether differences in immature B cell production might also result from changes in pre-B cell composition in old age.
The λ5 KO mouse, with highly enriched c-kit+ pre-B cells, provides a model to determine the behavior of this pre-B subset. B cells are capable of being generated in vivo from the limited pool of predominantly c-kit+ pre-B cells available in surrogate L chain KO mice, albeit the efficiency of B cell population in the periphery is considerably reduced when compared with WT mice (18). Therefore, it was expected that the c-kit+, surrogate L chain-deficient pre-B cells in λ5 gene KO mice would generate B cells in vitro.
Surface IgM−CD19+ bone marrow cells were isolated from WT mice (∼70% of early pre-B cells were c-kit−) and from λ5 gene KO mice (>70% of early pre-B cells were c-kit+). These WT and λ5 KO precursors were cultured in the presence of IL-7 and SCF (c-kit ligand) for 4 days; during culture, proliferation of the B cell precursors was greater with WT as opposed to λ5 KO precursor cells (Fig. 4,A). Cultures of WT precursors yielded a greater percentage of B cells than did those from λ5 KO precursors (Fig. 4,B). Upon taking differences in cell growth into account, production of new B cells from WT B cell precursors was ∼20-fold more effective than from λ5 gene KO B cell precursors (Fig. 4 C). The B cells produced from both WT and λ5 gene KO precursors were of recent origin as shown by similar levels of the AA4.1 Ag, a marker of immature B cells and present on >80% of B cells in our cultures (data not shown). In culture, the IgM−CD19+ cells grown from both WT and λ5 gene KO mice had similar proportions of cμ+ pre-B cells (∼12–15%; data not shown). As shown by Milne et al. (28), pro-B cells proliferate extensively, generating pre-B and new B cell formation continuously in IL-7 containing cultures. Therefore, it is unlikely that the different efficacies of B cell generation resulted from any disparities in the number of pre-B cells available for further differentiation in vitro.
The c-kit+ and c-kit− precursors from λ5 KO and WT mice proliferate and differentiate into B cells in vitro. A–C, Bone marrow cells from WT (B6) and λ5 KO mice were MACS sorted as IgM−CD19+ B cell precursors and cultured for 4 days in the presence of IL-7 (5 ng/ml) and SCF (50 ng/ml). D–F, Bone marrow from young BALB/c mice were pooled and IgM−CD19+AA4.1+B220+CD43+c-kit+ and IgM−CD19+AA4.1+B220+CD43+c-kit− B cell precursors were FACS sorted and cultured for 4 days in the presence of IL-7 (5 ng/ml) and SCF (50 ng/ml). A and D, Stimulation index (SI) (output CD19+ cells/input CD19+ cells) of cultures initiated with precursors from either WT or λ5 KO mice (A) or BALB/c precursor populations (D). B and E, Representative dot plot of B cells generated from either WT or λ5 KO precursors (B) or BALB/c precursor populations (E). C and F, Relative B cell output from precursors of WT and λ5 KO mice (C) or BALB/c precursor populations (F). Relative B cell output was determined as SI × % B cells in culture. Data are summarized for six to eight experiments. ∗, p < 0.02.
The c-kit+ and c-kit− precursors from λ5 KO and WT mice proliferate and differentiate into B cells in vitro. A–C, Bone marrow cells from WT (B6) and λ5 KO mice were MACS sorted as IgM−CD19+ B cell precursors and cultured for 4 days in the presence of IL-7 (5 ng/ml) and SCF (50 ng/ml). D–F, Bone marrow from young BALB/c mice were pooled and IgM−CD19+AA4.1+B220+CD43+c-kit+ and IgM−CD19+AA4.1+B220+CD43+c-kit− B cell precursors were FACS sorted and cultured for 4 days in the presence of IL-7 (5 ng/ml) and SCF (50 ng/ml). A and D, Stimulation index (SI) (output CD19+ cells/input CD19+ cells) of cultures initiated with precursors from either WT or λ5 KO mice (A) or BALB/c precursor populations (D). B and E, Representative dot plot of B cells generated from either WT or λ5 KO precursors (B) or BALB/c precursor populations (E). C and F, Relative B cell output from precursors of WT and λ5 KO mice (C) or BALB/c precursor populations (F). Relative B cell output was determined as SI × % B cells in culture. Data are summarized for six to eight experiments. ∗, p < 0.02.
Similar results were seen when surface IgM−CD19+B220+AA4.1+CD43+c-kit+ and IgM−CD19+B220+AA4.1+CD43+c-kit− precursors were isolated by cell sorting from BALB/c mice and cultured with IL-7 and SCF for 4 days. Again, c-kit− precursors demonstrated more robust growth in vitro compared with c-kit+ precursors (Fig. 4,D). While both c-kit+ and c-kit− precursors generated AA4.1+ immature B cells in vitro, c-kit− precursors were considerably more effective in producing new B cells (Fig. 4,E). When differences in growth were also taken into account, the relative efficacy of B cell production from c-kit− precursors was ∼10-fold greater than for c-kit+ precursors (Fig. 4 F). In vitro, both c-kit+ and c-kit− precursor cells that expanded in response to cytokine showed similar composition with ∼30% of cells at day 4 expressing cμ-chain but not surface IgM and, therefore, were pre-B cells (data not shown).
Immature B cells derived from c-kit+ B cell precursors in young and aged mice have altered CD43 expression
We have previously reported that, in vivo, aged mice often have an increased frequency of immature bone marrow B cells characterized by higher surface expression of CD43 (recognized by the S7 mAb) and increased cell size (23, 24). To determine whether this phenotype was associated with the origin of the B cell precursors (e.g., c-kit− vs c-kit+), and noting the increased proportion of c-kit+ vs c-kit− pre-B cells in aged mice, we compared immature B cells derived from c-kit+ WT and λ5 KO precursor cells with those derived from c-kit− B cell precursors in vitro.
Immature B cells derived from either c-kit+ B cell precursors isolated from young adult BALB/c mice or present in λ5 KO bone marrow exhibited altered surface phenotype with higher levels of CD43 than were seen on B cells generated from c-kit− B cell precursors in vitro (Fig. 5). In addition, immature B cells from c-kit+ precursors were generally larger in size, as assessed by forward angle scatter, than was seen for B cells derived from c-kit− precursors (data not shown). Although CD43 levels were increased on B cells generated from c-kit+ precursors, little or no detectable CD23, CD5, or CD11b was seen on these B cells during the 4-day culture period (data not shown).
B cells derived from young and aged c-kit+ precursors show increased CD43 in vitro. A, Representative histograms of CD43 expression on B cells derived from WT (B6) and λ5 KO precursors (upper panels), as described in Fig. 4, and from young and aged BALB/c c-kit+ and c-kit− precursors (middle and lower panels) on day 4 of IL-7/SCF culture (gray line = isotype control). B, Comparison of percent CD43high B cells derived from WT (B6) and λ5 KO precursors and isolated young and aged BALB/c c-kit+ and c-kit− precursors. CD43 “high” B cells (CD43high) were determined as those with fluorescence greater than isotype controls as indicated in A. ∗, p < 0.05 for each pair.
B cells derived from young and aged c-kit+ precursors show increased CD43 in vitro. A, Representative histograms of CD43 expression on B cells derived from WT (B6) and λ5 KO precursors (upper panels), as described in Fig. 4, and from young and aged BALB/c c-kit+ and c-kit− precursors (middle and lower panels) on day 4 of IL-7/SCF culture (gray line = isotype control). B, Comparison of percent CD43high B cells derived from WT (B6) and λ5 KO precursors and isolated young and aged BALB/c c-kit+ and c-kit− precursors. CD43 “high” B cells (CD43high) were determined as those with fluorescence greater than isotype controls as indicated in A. ∗, p < 0.05 for each pair.
Like young adult c-kit+ precursors, c-kit+ precursors in aged mice also yielded new B cells with altered phenotype in culture exemplified by increased CD43 expression (Fig. 5). In addition, immature B cells from c-kit+ precursors from aged mice were also larger in size compared with B cells derived from c-kit− precursors (data not shown). These studies indicated that generation of new B cells characterized by increased cell size and enhanced CD43 expression was a property of c-kit+ B cell precursors in both young and aged mice.
B cells derived from c-kit− and c-kit+ precursors from both young and aged mice display different expression of Ig L chain isotypes
B cells derived in vitro from either c-kit− young adult BALB/c precursors or from IgM−CD19+ cells from WT (B6) mice showed predominant usage of the κ L chain (∼80–90%) (Fig. 6,A). In contrast, in vitro-derived B cells from both c-kit+ young BALB/c precursors and IgM−CD19+ cells from λ5 KO mice showed significant decreases in levels of κ expression (∼40–60%) (Fig. 6,A). The reduction in κ L chains among B cells generated in these instances reflected, in part, increases in B cells expressing the λ1, λ2, and/or λ3 isotypes (λ[1–3]); B cells with κ L chains in addition to λ (1–3) expression (κ+λ[1–3]+), but most importantly, B cells that were low in κ or negative for both κ and λ (1–3) L chains (κlow/−λ[1–3]−) (Fig. 6,B). Isolated populations of c-kit− and c-kit+ precursors from aged BALB/c mice yielded new B cells with different expression of the κ/λ isotypes similar to that seen with the same populations obtained from young adult mice (Fig. 6, A and B).
B cells derived from c-kit+ precursors from young, aged, or λ5 KO mice show altered L chain expression in vitro. A, Representative histograms show κ L chain expressing B cells derived in culture from WT (B6) and λ5 KO sIgM−CD19+ bone marrow, and young (Yg) and aged BALB/c isolated c-kit+ and c-kit− precursors cultured as indicated in Fig. 4. Data are representative of three to six experiments. B, Summary of L chain expression on B cells derived in culture for three to six mice per group. ∗, p < 0.05.
B cells derived from c-kit+ precursors from young, aged, or λ5 KO mice show altered L chain expression in vitro. A, Representative histograms show κ L chain expressing B cells derived in culture from WT (B6) and λ5 KO sIgM−CD19+ bone marrow, and young (Yg) and aged BALB/c isolated c-kit+ and c-kit− precursors cultured as indicated in Fig. 4. Data are representative of three to six experiments. B, Summary of L chain expression on B cells derived in culture for three to six mice per group. ∗, p < 0.05.
In vivo, aged WT mice and young λ5 gene KO mice exhibit alterations in immature B cell L chain expression
Comparable to the results obtained in analysis of B cell development from λ5 KO precursors in vitro, IgM+CD19+AA4.1+ immature B cells from λ5 KO mice in vivo showed a lower proportion expressing normal levels of κ L chains (Fig. 7, A and B). In λ5 KO mice, increased λ (1–3) usage on immature B cells was observed, often with coexpression of κ L chain; additionally, a greater incidence of immature B cells without detectable λ (1–3) and low to negligible κ L chains was also observed. In addition, the immature B cells characterized as κlow/−λ (1–3)− had lower levels of surface IgM than did either κ+ or κ+λ (1–3)+ immature B cells (data not shown). Hence, in the absence of surrogate L chain and pre-BCR signaling where c-kit+ pre-B cells are prevalent, immature B cells show skewed usage of L chain isotypes in vivo similar to that seen in vitro.
Immature B cells in aged BALB/c and λ5 KO bone marrow exhibit altered L chain expression in vivo. A, Representative histograms show κ L chain expressing immature (AA4.1+) bone marrow B cells from young and old BALB/c mice and WT (B6) and λ5 KO mice in vivo. B, Summary of L chain expression on immature bone marrow B cells from young and aged BALB/c mice and WT (B6) and λ5 KO mice. C, Relative incidences of L chain expression among CD43− and CD43+ immature B cells in young and aged BALB/c bone marrow. D, Bone marrow (BM) immature B cells (sIgM+AA4.1+CD19+) and splenic (SP) transitional B cells (sIgM+AA4.1+CD19+) from λ5 KO and aged mice were further evaluated for κ and λ dual L chain expression. Aged mice were chosen that had substantial increases in dual L chain isotype expressing immature B cells in their bone marrow for comparison with those in spleen. Data are cumulative for three to eight mice. ∗, p < 0.05.
Immature B cells in aged BALB/c and λ5 KO bone marrow exhibit altered L chain expression in vivo. A, Representative histograms show κ L chain expressing immature (AA4.1+) bone marrow B cells from young and old BALB/c mice and WT (B6) and λ5 KO mice in vivo. B, Summary of L chain expression on immature bone marrow B cells from young and aged BALB/c mice and WT (B6) and λ5 KO mice. C, Relative incidences of L chain expression among CD43− and CD43+ immature B cells in young and aged BALB/c bone marrow. D, Bone marrow (BM) immature B cells (sIgM+AA4.1+CD19+) and splenic (SP) transitional B cells (sIgM+AA4.1+CD19+) from λ5 KO and aged mice were further evaluated for κ and λ dual L chain expression. Aged mice were chosen that had substantial increases in dual L chain isotype expressing immature B cells in their bone marrow for comparison with those in spleen. Data are cumulative for three to eight mice. ∗, p < 0.05.
In aged bone marrow, alterations in L chain usage were also readily detected among AA4.1+ immature B cells, in particular among aged mice that were highly depleted (>90%) in c-kit− pre-B cells. These “severely pre-B cell depleted” (6) aged mice had reductions in κ L chain expression compared with young controls (Fig. 7, A and B). This occurred coincidentally with both increased κ+λ (1–3)+ and κlow/−λ (1–3)− immature B cells in aged bone marrow (Fig. 7 B).
Since our previous results have shown that B cells expressing CD43 increased in proportion within the immature B cell pool of aged bone marrow (24), we next asked whether CD43+ immature B cells preferentially exhibited altered L chain isotype expression. In both young adult and aged bone marrow, CD43+ immature B cells were enriched in κ/λ (1–3) dual expression (20–40% κ/λ dual expression in CD43+ compared with 5–10% in CD43− immature B cells) (Fig. 7,C). In contrast, immature B cells with low κ expression (κlow/−λ−) were roughly equivalent within the immature B cell pools; however, their levels increased in both CD43− and CD43+ immature B cell compartments of aged mice (Fig. 7 C). Therefore, altered L chain isotype expression was seen in both CD43− and CD43+ immature B cell subsets within aged bone marrow.
It remains to be seen whether those immature B cells with unusual phenotype (e.g., those with high CD43, dual L chain isotypes, and/or low κ L chain levels) are capable of migration to the spleen for further development and function. Initial experiments indicate that ∼10–30% of immature/transitional splenic B cells (T2/T3; IgM+AA4.1+CD23+) express CD43 while CD23− T1 transitional B cells do not (data not shown). While B cell maturation has been regarded as a progression from T1→T2→T3 in spleen, it has been demonstrated that this process is asynchronous and CD23 expression and progression into T2/T3 subsets occurs in bone marrow (29). These bone marrow T2/T3 cells may then transit to the spleen. While CD43+ immature B cells may transit from bone marrow to the spleen, experiments shown in Fig. 7,D indicate that κ+λ+ dual isotype B cells, enriched in the CD43+ subset (Fig. 7 C), are increased among immature B cells in bone marrow of both young adult λ5 KO mice and in aged BALB/c mice, but are only rarely observed in the transitional AA4.1+ immature B cell compartments of the spleen. Therefore, phenotypically defined subsets of immature bone marrow B cells exhibit different capacities to migrate to and/or populate the spleen.
Discussion
It is clear that senescent mice, of various inbred strains, have reduced numbers of bone marrow B cell precursors (1, 2, 3, 4, 5, 6). This is readily observed as a reduction in late stage pre-B cells (e.g., Hardy Fraction D), although Miller and Allman (4), Labrie et al. (5), and we (3, 6) have reported variable losses of pro-B cells and earlier B lineage progenitors in old mice as well. The loss of late stage pre-B cells may be due to multiple mechanisms, including increased apoptotic death (8), poor proliferation to IL-7 (7), decreased Ig H chain rearrangements (5, 9), and, as we have suggested, diminished expression of the pre-BCR surrogate L chains and reduced pre-BCR signaling at the pro-B to pre-B cell transition (3, 10, 11, 12, 13, 14).
In aged mice, dampening of the pre-BCR checkpoint leads to partial arrest of the B2 B cell pathway but allows for residual pre-BCR-deficient B cell development
The B2 pathway of B lymphopoiesis involves the sequential development of pro-B and pre-B cells with expansion at the early pre-B cell stage due to pre-BCR signaling. The pre-BCR-mediated transition of pro-B to pre-B cells is characterized by changes in several surface Ags, including down-regulation of c-kit (20, 22). Diminished expression of surrogate L chains in aged B cell precursors, as shown herein (and in Refs. 3, 11), likely reduces production of pre-B cells in aged bone marrow. This is seen in our studies as a loss of early and late stage pre-B cells, in particular those that down-regulate c-kit expression. Approximately, 20% of early pre-B cells in young adult bone marrow retain c-kit and, as shown by others (20, 22, 26), fail to express the pre-BCR; comparable numbers are also detected in aged bone marrow, even as total pre-B cell numbers are reduced. This indicates that, overall, pre-B cell generation is partially blocked in aged mice with an increasing proportion of early pre-B cells in aged bone marrow that likely have not transited the pre-BCR checkpoint. These c-kit+ B cell precursors retain the capacity to generate new B cells, albeit with reduced efficiency. This is not surprising since pre-BCR signaling has been implicated in the induction of L chain rearrangements (30).
We propose that, in aged mice, the well-described B2 pathway reliant on pre-BCR assembly and signaling, is constrained by loss of surrogate L chain and diminished pre-BCR expression at the pro-B to pre-B cell transition. Although generally a minor component of normal B lymphopoiesis, pre-B cell maturation and B cell formation in a manner less dependent on pre-BCR signaling can occur and this deviation of normal B lymphopoiesis is more apparent in aged bone marrow as summarized in Fig. 8.
Surrogate L chain down-regulation in aged B cell precursors alters B2 B lymphopoiesis and promotes development of new B cells with distinct surface phenotype and L chain isotypes. Young mice, Pro-B cells undergo Vh-Dh-Jh recombination (step 1); pre-BCR competency is dictated by Vh-Dh-Jh sequence. Pre-BCR signaling results in down-regulation of c-kit and expansion and population at the late pre-B cell stage (step 2). A minor proportion of early pre-B cells retains c-kit and express μ H chains that assemble poorly with surrogate L chains (SL). Step 3, Late pre-B cells differentiate into immature B cells with most derived from pre-BCR competent precursors. Immature B cells with increased CD43 and altered L chain expression derived from c-kit+ pre-B cells are a minor contributor to the immature B cell pool. Aged mice, SL and pre-BCR expression is low and overall pre-B cell development is reduced (step 4). c-kit+ pre-B cells (step 5) generate immature B cells with increased CD43+ and altered L chain patterns at increased frequencies (step 6).
Surrogate L chain down-regulation in aged B cell precursors alters B2 B lymphopoiesis and promotes development of new B cells with distinct surface phenotype and L chain isotypes. Young mice, Pro-B cells undergo Vh-Dh-Jh recombination (step 1); pre-BCR competency is dictated by Vh-Dh-Jh sequence. Pre-BCR signaling results in down-regulation of c-kit and expansion and population at the late pre-B cell stage (step 2). A minor proportion of early pre-B cells retains c-kit and express μ H chains that assemble poorly with surrogate L chains (SL). Step 3, Late pre-B cells differentiate into immature B cells with most derived from pre-BCR competent precursors. Immature B cells with increased CD43 and altered L chain expression derived from c-kit+ pre-B cells are a minor contributor to the immature B cell pool. Aged mice, SL and pre-BCR expression is low and overall pre-B cell development is reduced (step 4). c-kit+ pre-B cells (step 5) generate immature B cells with increased CD43+ and altered L chain patterns at increased frequencies (step 6).
Alterations in immature B cell phenotypes in aged bone marrow
The immature B cell pools of aged mice differ significantly in lifespan, surface markers, and Ab repertoire from that of young adults (24, 31, 32, 33, 34). In our previous studies, focusing mainly on aged mice with severe deficits in B cell precursors, newly formed B cells were often decreased, consistent with their extensive loss of pre-B cells (6, 24). However, these studies revealed that at least two subsets of immature B cells were present in the bone marrow: those characterized by relatively high CD43 levels and those with low or absent CD43 (23, 24). Those B cells expressing CD43 also often expressed other Ags associated with activation including CD5, CD11b, and/or PD-1 (23, 24).
Moreover, those immature B cells characterized by CD43 expression were maintained in aged bone marrow and comprised an increased percentage of the total immature B cell pool (24). As shown previously, optimal generation of CD43+ immature B cells required Btk and likely required BCR signaling (23). This is analogous to requisites for CD23 expression reported for a subset of bone marrow immature B cells as described by Lindsley et al. (29). Indeed, our findings suggest that the CD43+ immature B cells comprise a subset of the CD23+ immature B cells in bone marrow (T. Williamson-Leon and R. L. Riley, unpublished results). As we have previously shown, immature B cells with this activation phenotype are not undergoing significant levels of proliferation in vivo (23, 24).
Of importance, as shown herein, generation of B cells that express relatively high levels of CD43 is associated with differentiation from c-kit+ B2 lineage precursors rather than from c-kit− B2 pre-B cells. Taken together with our previous results, this suggests that c-kit+ pre-B cells may preferentially develop into B cells that more readily undergo activation. We speculate that such B cells have undergone partial activation in response to self-Ags within the bone marrow microenvironment. This is supported by the increased incidences of CD43+ immature B cells that coexpress dual κ and λ L chains. This may reflect receptor editing/dilution as seen in both normal immature B cells and in transgenic mice with anti-phosphorylcholine or anti-DNA specific B cells (35, 36, 37, 38). The increase in such immature B cells as well as increased levels of newly produced B cells, both CD43+ and CD43−, with reduced levels of κ L chain/surface Ig, may also reflect increased tolerance (receptor editing/anergy) in aged bone marrow (38). Whether immature B cells with these unusual phenotypes are capable of exiting the bone marrow and populating the periphery as functional mature B lymphocytes or instead remain in the bone marrow, possibly contributing to the increasingly “stagnant” pool of immature B cells described by Johnson et al. (33), is not known. However, since increased proportions of dual L chain isotype expressing immature B cells in the bone marrow of both λ5 KO and aged mice are not seen among splenic transitional B cells, this suggests that these B cells may not undergo typical patterns of migration to the spleen and further maturation along normal pathways in the periphery.
However, the question remains as to whether these alterations in B lymphopoiesis affect B cell development in old age. We speculate, based on our in vitro experiments, that in aged mice with highly reduced B2 B lymphopoiesis (<10–20% of normal), pre-BCR compromised pre-B cells (e.g., c-kit+) could account for up to 20% of immature B cell generation in the bone marrow. As discussed below, such a shift in relative usage of B lymphopoietic pathways has significant importance in determining the phenotype and L chain usage of immature B cells in aged mice.
We suspect that the alterations in B cell repertoire previously reported in aged mice (24, 31), particularly within the bone marrow, have their basis, in part, in the relative use of distinct B cell developmental pathways characterized by differences in pre-BCR signaling. While a variety of deficits may contribute to changes in B cell precursor numbers and B cell functions in aged mice, it is remarkable that particular changes in B cell precursor phenotype (e.g., loss of c-kit− but retention of c-kit+ pre-B cells) as well as alterations seen in the properties of new B cells (e.g., CD43 and L chain expression) are mimicked in the λ5 gene KO mouse. This suggests that some, but likely not all, defects seen in B lymphopoiesis in aged mice, particularly at the pre-BCR selection checkpoint, may be derived from reduced surrogate L chain expression in B cell precursors. In aged mice, down-regulation of surrogate L chain (likely resulting from compromised E2A expression [3,11]), results in truncation of the B2 pathway and its diversion along a pre-BCR-compromised track.
That differences in development of B cell precursors underlie changes in new B cell formation and Ab repertoire in old age is consistent with the results obtained by Klinman and colleagues (31, 34) indicating that alterations in the emerging B cell repertoire in aged bone marrow and spleen were dictated by the developmental potential of bone marrow surface Ig− precursor cells (e.g., pre-B cells). Our findings indicate that alterations in the pre-B cell compartment in aged mice, likely due to poor pre-BCR function, result in changes in phenotype and L chain expression by immature B cells (summarized in Fig. 8). The functions of these distinct B cell subsets, their roles in peripheral immunity, and consequences of their alterations in old age remain issues to be further explored.
Acknowledgments
We gratefully acknowledge the assistance of Jim Phillips and the Flow Cytometry Core Facility at the Sylvester Comprehensive Cancer Center. We thank Ana Marie Landon for assistance in maintaining and characterizing the λ5 KO mice. We thank all members of the Riley and Blomberg laboratories for support in the performance of these studies.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health grants (to R.L.R. and to B.B.B.).
Abbreviations used in this paper: KO, knockout; SCF, stem cell factor; WT, wild type; MFI, mean fluorescent intensity.