Differential surrogate light chain expression governs B-cell differentiation

Since the discovery that B cells belong to a discrete lymphoid lineage,1 much has been learned about this pathway of cellular differentiation (reviewed in2-5). B lymphopoiesis occurs in hemopoietic tissues, primarily embryonic liver and bone marrow in mammals. In these sites, lymphoid progenitors lacking immunoglobulin (Ig) expression (pro-B cells) give rise to large B lymphocyte precursors (pre-B cells) containing μ heavy chains (HCs).6-9 The replicating pre-B cells later exit the cell cycle to generate small pre-B cells that in turn become IgM-bearing B cells.10-14 The Ig V(D)J gene rearrangement events underlying this progression follow an orderly sequence in which D-J_H rearrangement is followed by V-DJ_H rearrangement in pro-B cells.15-17 Pre-B cells generated through a productive VDJ_H rearrangement undergo several rounds of cell division before exiting the cell cycle. V-J_L rearrangement in the kappa or lambda light chain (LC) loci preferentially occurs in these small resting pre-B cells.18-20 A productive VJ_L rearrangement allows the assembly of B-cell receptors (BCRs), composed of μHC, LC, and the Igα/β heterodimeric signal-transducing elements and their expression on newly formed B cells.21 22 

Differential expression of a large cohort of genes is required for the progression of B lineage differentiation,5,17,23 and the nonrearranging VpreB and λ5/14.1genes24-27 are central figures in this differentiation process. Their expression begins in pro-B cells and is maintained into the pre-B cell stage.28-31 The VpreB and λ5 protein products are immediately assembled to form surrogate light chains (SLCs) having an overall structure resembling conventional LCs.32,33 This structural similarity allows the SLC to bind nascent μHC proteins, thereby liberating them from the BiP retention protein in the endoplasmic reticulum.29,34 The μHC homodimers then associate with Igα/β heterodimers to allow transit of the newly assembled pre-BCR through the Golgi apparatus en route to the pre-B cell surface.35,36 Pre-BCRs are concentrated within lipid rafts, where they are constitutively associated with syk, lyn, BLNK, and PI-3 kinase signaling elements.37 The importance of SLC during early B lineage differentiation is evidenced by a block in pro-B to pre-B cell differentiation in mice and humans with nonfunctionalλ538,39 or VpreBmutations.40 Deficiencies in any of the other pre-BCR components41-44 or in the essential downstream signaling elements (reviewed in 45) also lead to abortive pro-B cell differentiation. Pre-BCR and its signaling competency are thus essential for growth and survival of the pre-B cell population.

SLCs, the unique components of pre-BCR, are easily detected on pre-B cell lines of mouse and human origin46-49 and on murine pro-B cell lines,48 but their cell surface expression by primary B lineage cells has been remarkably difficult to demonstrate.49-57 This difficulty has clouded the issue of exactly when and where SLCs exert their function during B lineage differentiation. Even in studies in which pre-BCR expression has been demonstrated, only a few pre-B cells could be identified by cell surface staining with anti-SLC antibodies.49,50,54,56,58The most easily identifiable SLC-bearing cells in mice have in fact been pro-B cells,53 whereas in human B lineage cells the appearance of cell surface SLCs has not been detectable before the onset of μHC expression in most studies.49,56,59 With rare exceptions,59,60 human pro-B cell lines have also been reported to lack cell surface SLCs.49,55,56 58Species differences may thus exist regarding when cell surface SLC expression begins. Given the very low levels at which SLC-containing receptors are expressed on primary B lineage cells, technical differences and variability in the antibody reagents used for SLC detection could also account for some of the discordant results.

The paradox posed by our understanding of the importance of pre-BCR in B lineage differentiation versus the uncertainty about when pro-BCR and pre-BCR are expressed led us to use an amplified immunofluorescence assay that can detect as few as 50 to 100 molecules per cell61 and a panel of monoclonal anti-VpreB/λ5 antibodies to readdress the issue of SLC expression during B lineage differentiation. The comparative analysis of mice and humans clearly indicates species variability in pro-BCR expression and, more important, reveals a conserved mode of selective SLC down-regulation to permit pre-B cells to exit the cell cycle to undergo B-cell differentiation. Models for mouse and human B cell differentiation indicated by this analysis have interesting implications for normal and neoplastic B lymphopoiesis.

Human pro-B (Nalm16, RS4;11, REH, and JEA2), pre-B (697, Nalm6, and OB5), and B cell lines (Daudi and Ramos) were cultured in RPMI 1640 medium containing 100 U/mL penicillin, 100 μg/mL streptomycin, 10% heat-inactivated fetal calf serum, 50 μM 2-mercaptoethanol, and 2 mM L-glutamine (Life Technologies, Grand Island, NY). Murine pro-B cell lines (Raw8.1, SCID7, 38B9, 40E1, 63-12, and D1F9) and the 70Z/3 pre-B cell line were grown in supplemented Dulbecco modified Eagle medium. SCID7 cultures contained interleukin-7 (1 ng/mL). Human bone marrow cells were obtained from ribs resected from donors of renal transplants and from long bones of 13- to 19-week-old previable fetuses in accordance with policies established by the UAB Institutional Review Board. Mouse bone marrow was obtained from 4- to 12-week-old Balb/c, C57BL/6, RAG-1^−/−, RAG-2^−/−, and μMT mice. Mononuclear cells were isolated from human bone marrow by Ficoll-Hypaque gradient centrifugation and from murine bone marrow after erythrocyte lysis by treatment with 0.8% ammonium chloride.

The SA-DA4.4 anti-human μHC, anti-human VpreB8,56anti-human VpreB (HSL96, 4G7), anti-human λ5 (HSL11), and anti-human pre-BCR (HSL2),58,59 anti-mouse VpreB (R3 and R5),62 and anti-mouse λ5 (LM34)48monoclonal antibodies (mAbs) were digoxigenin-labeled following the manufacturer's recommendations (Boehringer Mannheim, Germany). Fab fragments of sheep anti-digoxigenin antibodies were obtained from Boehringer Mannheim; the FcR blocking reagent was from Miltenyi Biotec (Auburn, CA); phycoerythrin (PE)-conjugated anti-human CD19, CD79b, and allophycoerythrin (APC)-conjugated anti-CD34 antibodies were from Becton Dickinson (Mountain View, CA); streptavidin-APC and biotin-conjugated goat antibodies to human μHC, λLC, and κLC were from Southern Biotechnology Associates (Birmingham, AL); mouse B220, CD19, CD24, CD43, BP-1, CD25, CD2, κLC, λLC, CD79a, CD79b, and CD16/CD32 (Fc Block) monoclonal antibodies were from BD PharMingen (San Diego, CA); and goat antibodies to mouse μHC were from Southern Biotechnology Associates and Jackson Immunoresearch (West Grove, PA). Fluorescein isothiocyanate (FITC)–conjugated monoclonal antibodies to mouse κLC (187.1) and λLC (JC5.1) were gifts from Dr John Kearney (University of Alabama at Birmingham).

The enhanced indirect immunofluorescence system61used fluorochrome-filled liposomes conjugated to Fab fragments of sheep antidigoxigenin antibodies as a second-step reagent. Viable cells pre-incubated with 200 μg/mL IgG and 20 μg/mL anti-FcγIIR antibody in phosphate-buffered saline containing 0.5% bovine serum albumin for 10 minutes at 4°C were incubated with a digoxigenin-conjugated anti-VpreB, anti-λ5, or anti–pre-BCR antibodies for 15 minutes before washing and incubation with antidigoxigenin conjugated fluorescent liposomes for 1 hour on ice with agitation, followed by washing and analysis by flow cytometry. Staining specificity was assessed by pre-incubation with a 100- to 1000-fold excess of the unlabeled primary antibody. B lineage cells enriched by magnetic cell sorting (Miltenyi Biotec) of bone marrow using anti-human CD19 or anti-mouse B220 antibodies were sorted using FACStar (Becton Dickinson) or MoFlow (Cytomation, Fort Collins, CO) instruments. Cells fixed by incubation in 0.05% paraformaldehyde solution at 4°C for 1 hour were permeabilized by treatment with 0.2% Tween 20 or 0.1% saponin in phosphate-buffered saline at room temperature for 20 minutes and were blocked with serum for 10 minutes before antibodies were added for intracellular staining.

Cells purified by immunofluorescence cell sorting were fixed in 95% ice-cold ethanol for more than 30 minutes before treatment with RNase A (50 μg/mL) for 30 minutes at 37°C, followed by staining with FITC-conjugated anti–Ki-67 or control antibodies (BD PharMingen) for 30 minutes on ice. The cells were then washed before incubation, with 40 μg/mL propidium iodide for 15 minutes at room temperature, and flow immunocytometric analysis.

Subpopulations of pro-B and pre-B cells purified by 2 sequential fluorescence-activated cell sorting (FACS) sorts were lysed in TRIzol reagent (Gibco, Grand Island, NY) before preparation of total RNA, as recommended by the manufacturer (Gibco). First-strand cDNA synthesis was performed using the SuperScript preamplification system (Life Technologies) in parallel with a control synthesis reaction without reverse transcriptase (RT) to test for genomic DNA contamination. Protocols for polymerase chain reaction (PCR) of human gene products involved denaturing at 94°C for 3 minutes, amplification by 36 cycles of 94°C for 1 minute, 30 seconds for annealing at 55°C for RAG-1 and RAG-2, 60°C for TdT and B29, or 65°C forβ-actin and VDJ-Cμ, 72°C for 30 seconds, and extension at 72°C for 5 minutes. Primers for PCR amplification were as follows: TdT,5′-ACACGAATGCAGAAAGCAGGA-3′, 5′-AGGCAACCTGAGCTTTTCAAA-3′;RAG-1, 5′-ATGACAGCAGATGACCTCCTA-3′, 5′-TACCTCCAGAAGTTTATGAAT-3′; RAG-2,5′-TTCTTGGCATACCAGGAGACA-3′, 5′-CTATTTGCTTCTGCACTGAAA-3′;λ5/14.1, 5′-ACTGTCGGATCCTCGCAGAGCAGG-3′, 5′-CAGTCAAGCTTCTATGAACATTCT-3′; VpreB,5′-GTAGAGGCATGCCAGCCGGTGCTG-3′, 5′-CTTGAAGCTTTCGAGGGACACGTGT-3′;B29, 5′-GAATCTCTCGCCACCCTCACC-3′, 5′-CCTTGCTGTCATCCTTGTCCA-3′; VDJ-Cμ,5′-GGGTCGACACGGCCGTGTATTACTGT-3′, 5′-TGGTGGCAGCAAGTAGACATC-3′; andβ-actin, 5′-GCGGGAAATCGTGCGTGACAT-3′, 5′-GTGGACTTGGGAGAGGACTGG-3′. Primers used for Vλ-Jλ PCR amplification have been previously described.63 PCR protocols for mouse gene products involved 36 cycles of amplification with annealing at 58°C for β-actin, VDJ-Cμ, B29,and λ5; 60°C for RAG-2; and 65°C forTdT, RAG-1, and VpreB for 1 minute followed by extension at 72°C for 1 minute and 94°C denaturation for 1 minute. Primers for mouse PCR amplification were as follows: TdT,5′-GAAGATGGGAACAACTCGAAGAG-3′, 5′-CAGGTGCTGGAACATTCTGGGAG-3′;RAG-1, 5′-TGAAAAGGCACCCGAAGAAGC-3′, 5′-GGTGCCACTCCACGGTCACTT-3′; RAG-2,5′-CACATCCACAAGCAGGAAGTACAC-3′, 5′-GGTTCAGGGACATCTCCTACTAAG-3′;λ5, 5′-GTTGGGTCTAGTGGATGGTGT-3′, 5′-TTGGTCTGTTTGGAGGGTTGG-3′; VpreB,5′-GCCACCATCCGCCTCTCCTGT-3′, 5′-CCCCACGGCACAGTAATACAG-3′; B29,5′-TCAGAAGAGGGACGCATTGTG-3′, 5′-TTCAAGCCCTCATAGGTGTGA-3′;VDJ-Cμ, 5′-CGCGCGGCCGCTGCAGCAGCCTGGGGC TGAG-3′, 5′-GGAATGGGCACATGCAGATCTC-3′; Vk-Ck,5′-GGCTGCAGSTTCAGTGGCAGTGGRTCWGGRAC-3′, 5′-CTCATTCCTGTTGAAGCTCTTGACAATGGG-3′; and β-actin,5′-CGCAGCTCAGTAACAGTC-3′, 5′-TACGAGGGCTATGCTCTC-3′. Cellular cDNA was serially diluted for template use in semiquantitative RT-PCR assays of VpreB and λ5 transcripts.

Pro-B, pre-B, and B cell lines were examined with conventional and enhanced indirect immunofluorescence assays in a preliminary survey of SLC expression during human and mouse B lineage differentiation. In the enhanced immunofluorescence assay, cells were incubated first with anti-SLC antibodies conjugated with digoxigenin and then with fluorochrome-loaded liposomes bearing antidigoxigenin antibodies. This method yielded approximately a 2-log enhancement of the mean fluorescence intensity for anti-VpreB and anti-λ5 staining of human and mouse pre-B cells over that observed by conventional indirect immunofluorescence (Figure 1). Specificity was verified by the inhibition of staining with unconjugated anti-VpreB or anti-λ5 antibodies. SLC components were also detected on mouse pro-B cell lines, with similar enhancement seen with the fluorochrome-loaded liposomes. In contrast, SLC components could not be detected on human pro-B cell lines (Figure 1) despite their production of intracellular VpreB and λ5 SLC proteins.56 58 This species distinction held for all mouse (SCID7, 38B9, 40E1, 63-12, D1F9, Raw8.1) and human (Nalm16, RS4;11, JEA2, and REH) pro-B cell lines included in the analysis. Cell surface expression of SLC on mouse pro-B cell lines versus its absence on human pro-B cell lines was confirmed using the entire panel of monoclonal antibodies against VpreB (HSL96, 4G7, VP245) and λ5/14.1 epitopes (HSL11, LM34). When mouse and human pre-B cell lines were examined for cell surface reactivity with anti-μHC antibodies, they were universally found to express this pre-BCR component, whereas none of the pro-B cell lines produced μHC (data not shown).

Normal B lineage cells identifiable by their expression of the human CD19 or mouse B220 markers were examined for SLC expression using conventional and enhanced immunofluorescence staining methods. Although the usual difficulty was encountered in detecting SLC expression on bone marrow cells by conventional immunofluorescence, SLC-bearing B lineage cells were easily detected by the enhanced immunofluorescence method in human and mouse bone marrow samples (Figure2). Specificity of this immunofluorescence staining was validated by the results of blocking experiments with unlabeled antibody. The remarkable contrast between the staining results achieved with both methods was emphasized by the fact that we were unable to unambiguously detect SLC expression on mouse bone marrow cells by conventional indirect immunofluorescence (Figure 2B, lower panels). The frequency of SLC⁺ B lineage cells among bone marrow mononuclear cells identified by the enhanced immunofluorescence method was relatively low (range, 0.6%-4.8%).

Enhanced immunofluorescence was used for the detection of the VpreB and λ5/14.1 components, and it was used with conventional indirect immunofluorescence to detect other lineage markers to determine the stages during which primary B lineage cells express cell surface SLC. The SLC⁺ subpopulation of the CD19⁺ B lineage cells in human bone marrow varied in frequency from 18.1% ± 1.9% (mean ± SE, n = 4) in fetal samples to 4.4% ± 0.5% (n = 6) in adult samples (Figure 3). In the experiments illustrated here the less sensitive, conventional immunofluorescence method was used for μHC detection, but the modest upward deflection observed with anti-μHC staining, especially for cells with the highest VpreB expression levels, suggested coordinate μHC expression by the SLC⁺ cells. By comparison, a complete lack of staining of VpreB⁺ cells was observed with the anti-κ/λ light chain antibodies (Figure 3A). Coexpression of μHC and VpreB by pre-B cells was also indicated in parallel experiments in which equivalent percentages of μHC⁺ and VpreB⁺ cells (14.2%± 1.5% vs 14.9%± 1.6%, mean ± SE; n = 3) were identified among CD19⁺ κ/λLC⁻ fetal bone marrow cells using the enhanced liposome method. Although the SLC⁺ B lineage cells in human bone marrow were predominately CD34⁻, a subpopulation of the pre-B cells expressed this early hematopoietic cell marker at relatively low levels. The existence of a minor CD34⁺SLC⁺ subpopulation was confirmed by analysis of bone marrow cells with other anti-VpreB (HSL96, 4G7), anti-λ5 (HSL11), and anti-pre-BCR (HSL2) antibodies (data not shown). Notably, the proportion of SLC⁺CD19⁺ cells that retained CD34 expression was much greater in fetal bone marrow than in adult bone marrow (approximately 40% vs less than 5%).

In bone marrow samples from 6 juvenile mice, the SLC⁺subset comprised 6.3 ± 1.5% of the B220⁺ B lineage cells. As in humans, all SLC⁺ cells in mice were CD19⁺, and none expressed κ or λ LCs. The heterogeneous CD43 and BP-1 expression patterns observed for mouse SLC⁺cells (Figure 3B) suggested that primary pro-B and pre-B cells in mice can express SLC on their cell surface, since CD43 expression is restricted to the pro-B and early pre-B cell stages.17Given that the BP-1 antigen is selectively expressed by pre-B and immature B cells,64 the detection of cell surface SLC expression on BP-1⁻ cells also suggested that primary pro-B cells may express SLC-containing receptors. Conversely, the SLC expression on CD43⁻/BP-1⁺ cells is indicative of pre-B cell receptor expression. The composite results thus suggest that pre-B cells are the only SLC-bearing cells in human bone marrow, whereas in the mouse pro-B cells and pre-B cells may express cell surface SLC.

To test the implied species variability in SLC expression, SLC-bearing cells in human bone marrow were purified by cell sorting on the basis of cell surface CD19 and VpreB expression and the absence of κ or λ LC. When permeabilized before immunofluorescence analysis, the CD19⁺VpreB⁺LC⁻ cells were found to contain μHC (Figure 4A, top panel), thereby indicating that all SLC-bearing cells in human bone marrow samples are pre-B cells, whereas pro-B cells lack cell surface SLC.

Most of the B220⁺VpreB⁺ LC⁻ cells in mouse bone marrow also contained μHC. However, a discrete subpopulation of these cells (approximately 25%) did not contain μHC (Figure 4B, top right panel). This finding is concordant with the cell line data and with earlier reports indicating that primary pro-B cells in mice can express plasma membrane SLC in association with surrogate HC proteins.48 To verify that primary B lineage cells in mice can express cell surface SLC in the absence of μHC, bone marrow samples from RAG-1^−/− and RAG-2^−/− mice were examined. In these experiments, cell surface SLC expression was demonstrable for approximately 55% of the CD19⁺ cells, all of which were also CD43⁺ (Figure 3B, right panels). These results confirm the expression of SLC on a subpopulation of mouse pro-B cells.

Previous studies using conventional indirect immunofluorescence have identified pre-BCR components on few, if any, of the μHC⁺ pre-B cells.49-51,53 55-57 It seemed possible that this reflected the relative insensitivity of the methods used for detecting pre-BCR expression. To our surprise, even with the fluorochrome-filled liposomes, we were unable to detect SLC components on most pre-B cells in human and murine bone marrow samples. When the cell surface SLC⁻ subpopulation of CD19⁺CD34⁻ κ/λLC⁻ cells was isolated from human bone marrow samples and examined for intracellular μHC, they were found to contain μHC as a clear indication of their pre-B cell status (Figure 4A, lower panel). Correspondingly, when the subpopulation of mouse B220⁺LC⁻ bone marrow cells lacking cell surface VpreB was isolated, μHC expression was found in most, but not all, of these cells (Figure 4B, lower panel). This subpopulation of B220⁺VpreB⁻LC⁻μHC⁻cells represents pro-B and possibly non-B lineage cells that lack cell surface SLC. These composite results indicate that most of the pre-B cells in mice and humans (60%-80%) lack pre-BCR.

Human pre-B cell subpopulations, VpreB⁺CD19⁺κ/λLC⁻ and VpreB⁻CD34⁻CD19⁺κ/λLC⁻, were purified by 2 rounds of cell sorting. Analysis of these subpopulations indicated that most of the pre-BCR⁺subpopulation (ie, VpreB⁺) were relatively large cells, whereas the pre-BCR⁻ subpopulation of pre-B cells was composed primarily of relatively small cells (Figure5). In accordance with the cell size difference, the analysis of Ki-67 expression and DNA content indicated that a greater proportion of the pre-BCR⁺ pre-B cells were in the G₁/S/G₂/M stages of the cell cycle. The same trends were observed for the pre-BCR⁺ and pre-BCR⁻ subpopulations of mouse pre-B cells, though the strategy for identifying these subpopulations in mouse bone marrow samples was necessarily different. The mouse pre-B cell population of B220⁺ cells was identified by expression of the BP-1 antigen in the absence of κ/λLC expression. This strategy was used because all BP-1⁺/LC⁻ cells were shown to express intracellular μHC.64 Thirty percent of the latter subpopulation expressed cell surface IgM after overnight culture, thereby confirming that these cells are immediate B cell precursors. Pre-BCR expression is therefore restricted to a subpopulation of relatively large cycling pre-B cells, whereas receptor-negative pre-B cells are predominantly small, resting pre-B cells.

Pro-B cells in human bone marrow were identified in this analysis as CD19⁺CD34⁺SLC⁻ cells, and the 2 pre-B cell subpopulations were isolated as in the previous experiments. The mouse pro-B cell subpopulations, SLC⁺ and SLC⁻, were isolated from RAG-2^−/−mice to ensure the absence of more mature B lineage cells. Mouse pre-B cells were isolated from wild-type bone marrow as CD19⁺BP-1⁺κ/λLC⁻ cells, and the VpreB⁺ and Vpre-B⁻ subpopulations were then separated. Two rounds of cell sorting were conducted to ensure purity (more than 99.5%) of the subpopulations.

Several notable changes in gene expression were evident in this analysis (Figure 6). As anticipated, full-length μHC transcripts were either absent or were present only in trace levels in pro-B cells, and Tdt expression was detected exclusively in the pro-B subpopulations. RAG-1 andRAG-2 transcripts were down-regulated in the pre-BCR⁺ subpopulation, though RAG-2 transcripts could still be demonstrated, whereas both RAG-1 andRAG-2 expression were clearly evident in the VpreB⁻ subpopulation of pre-B cells. Correspondingly, the level of full-length light chain transcripts was dramatically increased in the small Vpre-B⁻ subpopulation of pre-B cells.B29, VpreB, and λ5 transcripts were expressed during all the pro-B and pre-B cell stages. However, VpreBand λ5 transcript levels were selectively down-regulated in the pre-BCR⁻ subpopulation of pre-B cells in both species (Figure 6), a finding that was verified by semiquantitative analysis (data not shown). Moreover, when the receptor-negative subpopulation of pre-B cells was permeabilized to allow examination of intracellular pre-BCR components, these cells were found to be virtually devoid of VpreB and λ5 proteins, whereas intracellular stores of μHC, Igα, and Igβ were all maintained (Figures 4, 6C, and data not shown). Selective loss of SLC expression is therefore a defining characteristic of the subpopulation of relatively small pre-B cells that no longer express pre-BCR.

Since discovery of the SLC components, VpreB and λ5/14.1, their expression as unique pre-BCR components has been shown to be essential for pre-B cell growth and survival. Nevertheless, their detection on primary B lineage cells has been extremely challenging, largely because of their low levels of expression. This has led to conflicting views about when and where pro-BCR and pre-BCR are expressed during B lineage differentiation in mice and humans.49,51,54-60 The current studies used a sensitive immunofluorescence assay capable of detecting fewer than 100 molecules per cell61 to confirm the existence of a species difference regarding when cell surface SLC expression begins. They also define a conserved pattern of extinguished SLC expression in pre-B cells that allows them to exit the cell cycle and undergo the light chain rearrangements necessary for B-cell differentiation.

SLC components could not be detected on human pro-B cells with monoclonal antibodies recognizing the VpreB and λ5 proteins even with the enhanced immunofluorescence method. Primary pro-B cells and pro-B cell lines failed to express detectable cell surface SLC components despite their presence within these cells. Previous reports suggesting that human pro-B cells can express cell surface SLC in the absence of μHC57,59,60 may reflect, in part, differences in the phenotypic characteristics used to identify pro-B cells. CD34-bearing B lineage cells are generally considered to represent the human pro-B cell fraction, but, as did LeBien,65 we observed that CD19⁺CD34⁺ cells may express cell surface SLC, especially early in ontogeny when B lymphopoiesis is most active. However, all these CD34⁺SLC⁺ cells were shown to express μHC, thereby indicating that CD34 may be transiently expressed after B lineage cells undergo productive VDJ_Hrearrangement to become pre-B cells.

SLC expression for mouse pro-B cells was reaffirmed in these studies, though a physiological role of this type of pro-BCR remains enigmatic. Cell surface expression of SLC is restricted to a subpopulation of the murine pro-B cells, and the SLC⁺ phenotype is faithfully reproduced by pro-B cell lines in which the SLC may be associated with surrogate HC proteins.48,51 One of the surrogate HC proteins has been identified as BILL-cadherin,66 a transmembrane glycoprotein that apparently lacks signal transducing capability. Igα/Igβ heterodimers are also expressed in low levels on mouse pro-B cells, and their ligation can induce progression of cellular differentiation.67 However, these Igα/Igβ heterodimers are associated with calnexin and have no apparent linkage with SLC components. The normal development of pro-B cells in VpreB- and λ5-deficient mice40 68 further argues against a functional role for SLC on mouse pro-B cells. The onset of intracellular SLC production before μHC production during B lineage development may simply reflect a preparatory step for subsequent pre- BCR assembly.

The pattern of pre-BCR expression conserved in mice and humans was characterized by the restriction of receptor expression to a limited subpopulation of the pre-B cells. Although the entire pre-B population exhibited the hallmark features of intracellular μHC presence and LC absence, the pre-BCR⁺ and pre-BCR⁻subpopulations of pre-B cells defined by the enhanced immunofluorescence method exhibited distinctive genotypic and phenotypic profiles. Receptor-positive pre-B cells were found to be relatively large, and many of these had entered the cell cycle. Consistent with the results of previous studies using other combinations of cell identification markers,53,69,70 this population of relatively large pre-B cells exhibited down-regulation of the recombination activating genes, RAG1 andRAG2. In contrast, pre-B cells that had no detectable pre-BCR were predominantly small, resting lymphocytes. These receptor-negative pre-B cells expressed RAG1 andRAG2, correspondingly exhibited enhanced levels of full-length LC transcripts, and were further distinguished by the selective down-regulation of SLC genes, VpreB andλ5/14.1. The latter feature was reflected by absence of the VpreB and λ5 proteins, although intracellular stores of the μHC, Igα and Igβ proteins were maintained in these cells. This selective loss of SLC production, without which μHC and Igα/Igβ cannot be expressed on the cell surface, thus terminates the expression of pre-BCR. In this regard, Igα/Igβ heterodimers have not been shown to bind to μHC before their association with SLC.36 Hence, selective down-regulation of SLC expression during pre-B cell differentiation leads inexorably to the loss of pre-BCR expression, an event that coincides with pre-B cell exit from the cell cycle.

Bone marrow DNA labeling studies initially revealed that large dividing pre-B cells give rise to small postmitotic pre-B cells that become cell surface IgM⁺ B cells after a 1- to 2-day interval.10,13,14 V-J_L rearrangements in the κ and λ LC loci then take place in the small, postmitotic pre-B cells.18,20,70 The V(D)J rearrangement process is quiescent during the earlier pre-B phase when receptor-positive pre-B cells undergo mitosis. In addition to the corresponding down-regulation of RAG1 and RAG2, both of which are required for V(D)J rearrangement,71 phosphorylation of RAG2 leads to its efficient degradation before the S phase.72 This combination of events ensures protection against the creation of double-stranded DNA breaks and inappropriate cell death of the replicating, receptor-positive pre-B cells.

This analysis of SLC expression supports a B-cell differentiation scheme that differs for mice and humans primarily in the earlier onset of cell surface SLC expression in mice (Figure7). The extinction of pre-BCR expression by selective down-regulation of SLC production appears to have a profound impact on the pre-B cell differentiation process. The most well-documented function of pre-BCR is the promotion of pre-B cell growth and survival in conjunction with costimulatory signals provided by stromal cells or interleukin-7.73 Although evidence for pre-BCR promotion of HC allelic exclusion and LC gene rearrangement has also been suggested,74-76 these events can proceed in the absence of pre-BCR.38 68 The implication that pre-BCR expression is not directly involved in these differentiation events is reinforced by our observation that the loss of pre-BCR expression and withdrawal from the cell cycle coincide with the reinstatement ofRAG-1 and RAG-2 expression and VJ_Lrearrangement. The selective SLC down-regulation to extinguish pre-BCR expression is thus an important prerequisite to B-cell differentiation. This sequence of events also ensures that SLC and conventional κ/λLC are not expressed simultaneously during primary B-cell differentiation, as indicated by the finding that SLC-bearing cells and κ/λLC-bearing cells belong to nonoverlapping populations in mouse and human bone marrow samples, irrespective of donor age.

The model of primary B-cell differentiation supported by this analysis of SLC expression begs the question of how the VpreB andλ5/14.1 genes are turned off in concert at a key point in the pre-B cell differentiation process. A complex locus control region for the VpreB/λ5 genes has been demonstrated,77 but its function during primary B lineage differentiation is not entirely clear. Although the VpreBand λ5 promoter regions contain binding sites for Ikaros, EBF, E2A, and PAX5,4 expression of these transcription factors is maintained throughout pre-B cell differentiation. A possible mechanism for SLC suppression is suggested by the Ikaros-mediated transcriptional silencing that has been demonstrated for theλ5 gene in B cells.78-80 

The current observations are also relevant to the RAG and SLC expression observed for B lineage cells in secondary lymphoid tissues.56,81-86 Initially considered to reflect a reactivation of the pre-B cell gene program in germinal center B cells, further analysis of this phenomenon instead suggests an efflux of bone marrow pre-B cells to the periphery in response to inflammatory stimuli.86 However, a subpopulation of peripheral B cells with a restricted V gene repertoire has been reported to express SLC and κ or λ LC.85 We have so far been unable to identify SLC-bearing cells in normal peripheral lymphoid tissues, but background staining problems encountered in the analysis of peripheral B cells (Y.-H.W. and R.P.S., unpublished observations, 2001) indicate further study may be needed for a definitive resolution of this issue. This technical difficulty could reflect the expression of a newly identified family of Fc receptor homologues that are differentially expressed by B cells in peripheral lymphoid tissues.87,88 Finally, the dual expression of SLC and κ/λLC on leukemic cell lines48,49 is another remarkable exception to the rule established for normal B lymphopoiesis in humans and mice. Although this dual SLC/LC expression was previously interpreted as a reflection of the normal differentiation process,49 the current analysis suggests that the leukemic transitional pre-B/B cells must have been diverted from the normal differentiation pathway. In addition, the current differentiation model predicts that failure to down-regulate SLC expression would promote excessive pre-B cell proliferation at the expense of B-cell differentiation, a hypothesis that is being tested.

We thank Dr Larry Gartland for help with flow cytometry; Drs Peter Burrows, Flavius Martin, and John Kearney for helpful discussions; and Marsha Flurry and Ann Brookshire for help in preparing the manuscript. We thank Dr Alan Fantel (University of Washington, Seattle) for providing fetal bone marrow tissue samples.