Expression of RAGs in Peripheral B Cells outside Germinal Centers Is Associated with the Expression of CD5
http://www.100md.com
免疫学杂志 2005年第9期
Abstract
Previous studies have indicated that mature B cells reactivate secondary V(D)J recombination inside and outside the germinal center (GC) of peripheral lymphoid organs. The nature of the B cells undergoing Ig rearrangement before they enter GC is unknown. In this study, we present evidence that activated mature CD5-positive human tonsil B cells coexpress both RAG1 and RAG2 mRNA and protein, and display DNA cleavage resulting from their recombinase activity. Furthermore, in vitro activation of CD5-negative naive mature B cells by IgR and CD40 cross-linking induces expression of CD5 on a subset of cells, and leads to the up-regulation of RAG1 and RAG2 only in cells turned positive for CD5. Thus, RAG gene expression is closely related to CD5 expression outside GCs. These data suggest that CD5 is associated with receptor revision in activated mature B cells and likely to promote expression of suitable IgR capable of initiating the GC reaction.
Introduction
Secondary V(D)J rearrangement of Ig genes has been extensively described in immature bone marrow B cells. This process is referred to as receptor editing, contributes to the maintenance of immunological tolerance (1), and could rescue potentially autoreactive B cells from apoptosis (2). Receptor editing is initiated upon BCR engagement by Ag, resulting in the up-regulation of RAG gene expression, which then participates in editing the autoreactive BCR-encoding Ig V(D)J rearrangements. Secondary V(D)J recombination has been shown to also occur in murine spleen and lymph node B cells in response to immunization (3, 4, 5). This peripheral V(D)J rearrangement, termed receptor revision, might contribute to the generation of high-affinity Abs in germinal centers (GCs)3 following the process of somatic hypermutation in response to stimulation by T-dependent Ags (6, 7).
There is evidence that human B cells also undergo secondary Ig recombination in the periphery (8, 9). Seven peripheral subpopulations of B cells have been identified in human tonsils based on the surface expression of CD38 and IgD in conjunction with other markers (reviewed in Ref. 10), which have led to the proposition of a model of T cell-dependent mature B cell differentiation. The naive mature B (Bm) cells Bm1 (CD38–IgD+CD23–) and Bm2 (CD38+IgD+CD23+) would be activated in extrafollicular areas through interaction with interdigitating cells and Ag-specific T cells. Activated blasts may either terminally differentiate into low-affinity Ab-forming cells or become GC founder cells Bm2' (CD38++IgD+). In GCs, Bm2' cells would differentiate into centroblast Bm3 (CD38++IgD–CD77+), in which somatic hypermutation in V gene region might take place during proliferation. This process may generate cells bearing BCRs with various affinities. The mutant cells will be selected during their differentiation into centrocytes Bm4 (CD38++IgD–CD77–), depending on their BCR affinity for the Ag that is trapped as an immune complex on the surface of follicular dendritic cells. High-affinity Bm4 may interact strongly with the Ag, process, and present it to GC T cells. These T cells will be induced to express CD40L, to secrete cytokines, and to promote survival, proliferation, and isotype switching of the B cells. These activated B cells will thus terminally differentiate either into early memory (CD38+IgD–) and memory Bm5 (CD38–IgD–) cells or into high-affinity Ab-forming cells. In contrast, autoreactive Bm4 cells would be deleted, because they do not receive survival signal from T cells (11). Mutant Bm4 cells with low affinity for Ag may have the potential to revise their receptor due to an increased RAG gene transcription (12). This could raise the affinity of the BCR for the Ag (8, 9). In this setting, strong BCR engagement would switch off RAG gene expression (13), suggesting that receptor revision, which might be involved in affinity maturation of Abs, will be terminated when BCR is strongly cross-linked. This process will permit the positive selection of high-affinity B cells to terminally differentiate.
Recently, RAG expression in peripheral B cells has been observed outside the GC in a number of mouse models (14, 15) and in humans (16). In the human model, RAG-positive B cells were found in the follicular mantle zone (FMZ) of the GCs where naive Bm1 and Bm2 cells are positioned (10). The T cell marker CD5 is expressed by <10% of the B cells in the adult human spleen and <30% in lymph nodes (17). It is interesting that the CD5+ B cell population is also enriched in the FMZ (18, 19).
Interestingly, in hen egg lysosyme (HEL)/anti-HEL transgenic mice, CD5 expression is observed on mature anergic B cells (20). Moreover, immature anergic B cells may activate the machinery for V(D)J rearrangement to maintain B cell tolerance (21). Although CD5+ B cells have been shown (22) to express RAG in the peritoneal cavity, it is unknown whether CD5 is involved in inducing the recombination process. Mature anergic B cells that fail to rearrange may express CD5 to negatively regulate the BCR and avoid elimination, yet maintaining tolerance to self-Ag. However, by raising the threshold required for activation (23), CD5 may also contribute to the triggering of the BCR revision. This view implies that CD5 expression in the FMZ could be induced on mature B cells with BCR affinity insufficient to enter GC. RAG expression would, therefore, be found in activated cells expressing CD5. The present study was designed to test this hypothesis.
Materials and Methods
Preparation of B cells
Tonsils were obtained from 5- to 18-year-old children undergoing routine tonsillectomy. Tissues were minced up, diluted in PBS, and filtered to deplete larger cells and clumps. Filtered cells were layered onto Ficoll-Hypaque density medium and centrifuged for 30 min at 450 x g. Cells were then incubated with neuraminidase-treated SRBC for 1 h at 4°C. Depletion of T cells was achieved by a second round of centrifugation on Ficoll-Hypaque for 25 min.
Flow cytometry analysis
FITC-mAbs to human CD19 (clone J4.119), CD21 (clone BL13), CD10 (clone ALB1), CD95 (clone UB-2), CD27 (clone MP271) and CD23 (clone 9P25), rat unlabeled Ab to CD77 (clone 38-13) visualized with FITC-goat anti-rat, PE-linked to cyanin 5 (PC5)-Abs to CD38 (clone LS198), and CD5 (clone BL1a), and biotinylated (biot) anti-CD5 (clone UCHT2) visualized with PE-Texas Red energy-coupled dye (ECD)-streptavidin, were from Beckman Coulter. FITC-anti-IgD (clone IA6-2), PE-anti-IgD, PE-anti-IgM (clone G20127), and biot-anti-IgM visualized with ECD-conjugated streptavidin were from BD Biosciences. Four-color analysis and sorting were performed on an Epics Elite flow cytometer (Beckman Coulter).
RNA isolation and RT-PCR amplification
Total RNA was extracted from 105 B cells using the RNable reagent (Eurobio) and reverse transcribed in 20 μl with Superscript II RNase H-reverse transcriptase (Invitrogen Life Technologies). For RT-PCR amplifications, nested reactions were performed with 1 μl of cDNA using the primer pairs listed in Table I with TaqDNA polymerase (Invitrogen Life Technologies) as follows: in the first round of PCR, cDNA was amplified for 25 cycles of 30 s at 94°C, 1 min at 56°C, and 1 min at 72°C with a final 10-min extension at 72°C. The second PCR round was for 35 cycles of 30 s at 94°C, 1 min at 56°C, and 1 min at 72°C with a final extension step at 72°C for 10 min. For the GAPDH RT-PCR amplification, only one round of PCR was conducted for 40 cycles. Because the GAPDH primers spanned a short intron of 100 bp, genomic DNA could be amplified, but easily distinguishable from the specific cDNA product. In such cases, samples were excluded from the RAG studies. Moreover, RAG1 and RAG2 first-round PCR primers spanned an intron at 5168 and 1174 bp, respectively, which thus could not amplify contaminating genomic DNA. Taken together, these precautions certify that the PCR products originate specifically from mRNA. RT-PCR products were analyzed on 2% agarose gels stained with ethidium bromide. The expected size for each specific product is indicated in Table I.
Table I. Sequences of oligonucleotides used as primers for the amplification of cDNA in RT-PCR and nested RT-PCR, and of DNA in LM PCR
Single-cell protocol
Individual IgD+CD38+CD5+ Bm2 cells were sorted into PCR tubes containing 10 μl of reverse transcriptase buffer (1x first-strand buffer (Invitrogen Life Technologies), 5 μM random hexamer, 0.5x reverse transcriptase buffer, 0.01 M DTT, 0.5 mM dNTP (all from Promega)), using the flow cytometer outfitted with the Autoclone single-cell deposition unit (Beckman Coulter). mRNA conversion from a single cell was conducted with Superscript II RNase H-reverse transcriptase (Invitrogen Life Technologies) for 2 h at 42°C. A nested RT-PCR for RAG1, RAG2, and GAPDH was performed with 25 cycles of 30 s at 94°C, 1 min at 56°C, and 1 min at 72°C with a final 10-min extension at 72°C in the first round of PCR. The second PCR round was for 40 cycles of 30 s at 94°C, 1 min at 56°C, and 1 min at 72°C with a final extension at 72°C for 10 min. Frequencies of cells positive for one transcript (RAG1+2– or RAG1–2+), negative for both (RAG1–2–), or positive for both (RAG1+2+) were calculated.
Ligation-mediated PCR (LM-PCR)
DNA was extracted from 105 cells using the DNAzol kit (Invitrogen Life Technologies), and 500 ng ligated to the BW1/2 linker at 20 pmol with T4 DNA ligase (Promega) in ligation buffer (66 mM Tris (pH 7.5), 5 mM MgCl2, 1 mM DTT, and 1 mM ATP) for 16 h at 16°C (24). The BW1/2 linker was obtained by hybridization of 2 nmol BW1 primer with BW2 primer (Table I) for 5 min at 60°C. PCR was performed using TaqDNA polymerase following a denaturation step at 95°C for 15 min. In the first PCR round, linker primer BW1 was used with either J4-1, J5-1, J6-1, or J7-1 primer. Samples were denaturated for 3 min at 94°C and amplified for 15 cycles of 40 s at 94°C, 40 s at 65°C, and 1 min at 72°C, followed by a final 10-min extension at 72°C. After 3 min at 94°C for denaturation, an additional 40 cycles of PCR were performed using the BW3 linker primer with either J4-2, J5-2, J6-2, or J7-2 nested locus-specific primer. The expected size for each specific nested product is indicated in Table I. Similar PCR, with 40 s at 55°C for the hybridization step, were also performed with the Jk locus-specific primers. PCR products were analyzed on 2% agarose gels stained with ethidium bromide. The identity of the amplified products were verified by digestion with BamHI restriction enzyme for the J7 signal end, and by sequence analysis of all the PCR products.
Indirect immunofluorescence
Sorted cells were washed in PBS, centrifuged onto glass slides, fixed in 4% paraformaldehyde, and further prepared as described, with modifications (25). First, to avoid nonspecific fixation, the cells were incubated in PBS containing 0.1% Nonidet P-40 (Sigma-Aldrich), 2% nonfat dry milk, 5% FBS, and 0.02% sodium azide, for 20 min, followed by an additional treatment for 1 h with 50 μg/ml Cohn fragment (Sigma-Aldrich). Cells were then left overnight at 4°C with goat anti-RAG2 Ab (Santa Cruz), incubated with biot donkey anti-goat IgG (Jackson ImmunoResearch Laboratories) and visualized with tetramethylrhodamine isothiocyanate-streptavidin (Jackson ImmunoResearch Laboratories). Finally, the cells were left overnight at 4°C with rabbit anti-RAG1 Ab (Santa Cruz), followed by FITC- donkey anti-rabbit IgG (Jackson ImmunoResearch Laboratories). Cells were examined immediately by fluorescence microscopy using an oil immersion lens (Zeiss; Axioplan), and enumerated as positive for one protein (RAG1+2– or RAG1–2+), negative for both (RAG1–2–), or positive for both (RAG1+2+). Examination by confocal microscopy (TCS NT; Leica) was also performed following staining of the cells for RAG1 or RAG2 by FITC labeling, using FITC-donkey anti-goat (Jackson ImmunoResearch Laboratories) to visualize anti-RAG2 Ab, and propidium iodide nucleus counterstaining, to evaluate the localization of each protein. The human 697 pre-B cell line and Hep-2 epithelial carcinoma cell line were used as positive and negative controls for the RAG protein detection, respectively (data not shown).
Cell culture
Cells were cultured in RPMI 1640 medium (Invitrogen Life Technologies) supplemented with 10% heat-inactivated FCS (Invitrogen Life Technologies), 2 mM L-glutamine (Invitrogen Life Technologies), 200 U/ml penicillin (Rh?ne-Poulenc), and 100 μg/ml streptomycin (Diamant). A total of 105 murine NIH3T3 fibroblasts untransfected (control), or transfected with human CD40L gene, were incubated with mitomycin C (Sigma-Aldrich) and used for stimulation of B cells (26). Transfectants were grown in DMEM medium (Invitrogen Life Technologies) supplemented with 10% FCS, 2 mM L-glutamine, and 10 μg/ml gentamicin (Rh?ne-Poulenc). Isolated B cells were dispensed at 106 cells per well in a final volume of 1 ml with, or without, anti-IgM Ab-coated beads at 1 μg/ml. Cells were also stimulated with anti-IgM Ab-coated beads alone, or with 30 ng/ml PMA.
Results
Based on IgD and CD38 expression, all mature B cell subpopulations were sorted (Fig. 1A), and RAG mRNA was analyzed by nested RT-PCR. As expected, we found that Bm3/4 GC cells contained RAG1 and RAG2 mRNA. Furthermore, all but naive Bm1 cells and Bm5 memory cells expressed both RAG transcripts (Fig. 1B), supporting the notion that some B cells outside GC may also undergo secondary rearrangement. Specifically activated naive Bm2 cells, pre-GC Bm2' cells, and early Bm5 memory cells seemed to have the machinery to rearrange.
FIGURE 1. RAG1, RAG2, and CD5 expression in tonsil B cell subsets. A, Tonsillar B cell subsets were defined by the surface expression of CD19, IgD, and CD38. The gates for each subset are shown. B, Specific RAG1 and RAG2 mRNA of sorted subsets were amplified by nested RT-PCR and that of GAPDH by RT-PCR. C, Tonsillar B cell subsets were analyzed for the membrane expression of CD5. D, The CD5 expression was evaluated in the total B cells, and the frequency of each B cell subset in the CD5-positive population was determined. E, The CD5+ and CD5– fractions of each B cell subset were sorted, and the RAG1, RAG2, and GAPDH mRNA were amplified.
To determine whether the presence of CD5 influences RAG transcription, we first examined each subpopulation for the expression of CD5 (Fig. 1C). The frequency of CD5+ B lymphocytes increased in the extrafollicular areas from 16.2 ± 4.7% in Bm1 to 18.1 ± 7.4% in Bm2, and to 22.7 ± 10.9% in Bm2' subpopulations (Table II). Remarkably, the frequency decreased once the cells entered the GC, that is, as few as 3.4 ± 1.9% of Bm3/4 cells express CD5. When B cells left the GC, they did not re-express CD5, because there were only 9.2 ± 5.3% of early Bm5 and 9.8 ± 6.4% of Bm5 cells to carry CD5+. We reasoned that, because it is transiently expressed on extrafollicular lymphocytes, CD5 may influence B cell behavior before their entry into the GC.
Table II. Distribution of the different B cell subpopulation in human tonsil B lymphocytes and frequency of the CD5+ B cell subset
Next, we sorted the B cell subpopulations according to their CD5 expression (Fig. 1D) and analyzed the presence of RAG1 and RAG2 mRNA by nested RT-PCR. Interestingly, from the B lymphocytes located in the extrafollicular areas, only the activated Bm2 cells expressing CD5 displayed RAG1 and RAG2 transcripts (Fig. 1E). These observations suggest that the presence of CD5 at this stage of maturation may be associated with the recombination machinery to be operational. In contrast, because both CD5+ and CD5– early Bm5 displayed RAG1 and RAG2 mRNA, CD5 expression does not seems to be influential on V(D)J rearrangement at the memory stage. RAG1 and RAG2 mRNA detection was performed on different tonsil samples. In five of eight, the CD5+ Bm2 lymphocytes expressed RAG1 and RAG2 transcripts. Among them, one sample showed the CD5– Bm2 cells also expressing both RAG gene mRNA.
For the V(D)J recombination machinery to be operational, both RAG1 and RAG2 mRNA must be coordinately expressed in a cell at the same time (27). With this requirement in view, we used single-cell RT-PCR to assay for RAG1 and RAG2 transcript expression. In four different tonsil samples, single CD5+ Bm2 lymphocytes were sorted, and nested RT-PCR was performed to evaluate the transcription of RAG1 and RAG2 mRNA in the same cell. This analysis was restricted to cells expressing GAPDH mRNA to make sure cDNA was present in the samples after sorting, whereas samples with contaminated genomic DNA were excluded (Fig. 2A). We found that 23.5 ± 8.4% of the cells coexpressed RAG1 and RAG2 mRNA, 37.3 ± 10.7 and 14.7 ± 5.8% expressed either RAG1 or RAG2, respectively, whereas 22.9 ± 13.9% expressed neither (Fig. 2B). We noticed that almost one-fourth of the CD5+ Bm2 cells expressed both mRNA, whereas 37.1% of the Bm3/4 cells, 6 and 16.6% of the CD5+ and CD5– early Bm5, respectively, expressed also RAG1 and RAG2 transcripts (data not shown). Given that CD5+ Bm2 cells represented 2.6–13.9% of the total B cells, the CD5+ Bm2 lymphocytes that have a complete machinery for V(D)J rearrangement accounted for 0.7–2% of the total tonsil B lymphocytes (Table III).
FIGURE 2. Single-cell mRNA and protein expression of RAG1 and RAG2 in individual tonsil Bm2 cell subset. A, Individual IgD+CD38+CD5+ tonsil Bm2 cells were sorted with the Autoclone single-cell deposition unit of the Epics Elite flow cytometer. RAG1 and RAG2 mRNA expression were analyzed by nested RT-PCR. Among 40 cells sorted in each experiment, only GAPDH mRNA-positive cells were considered for analysis. A representative example is shown where crosses indicate cells excluded from the RAG1 and RAG2 analysis. B, The frequencies of single RAG1- or RAG2-positive cells as well as the frequencies of RAG1- and RAG2-double-negative or -positive cells are shown. Mean ± SD of four experiments. C, Tonsil Bm2 cells identified by the expression of IgD and CD38 were sorted according to the presence of CD5. Purified B cells were cytocentrifuged and stained for RAG1 and RAG2 protein detection by immunofluorescence. Cells stained without anti-RAG1 and anti-RAG2 Abs were used as controls. D, Among the IgD+CD38+CD5+ tonsil Bm2 cell subpopulation, the frequencies of single RAG1+ or RAG2+ cells, as well as the frequencies of RAG1- and RAG2-double-negative or -positive cells were established by direct enumeration in a randomly selected subset. Mean ± SD of three experiments. E, IgD+CD38+ CD5+ and CD5– tonsil Bm2 cell subsets were sorted, and RAG1 and RAG2 enzymatic activity was evaluated. The J signal breaks were detected by LM-PCR using different sets of specific primers. LM-PCR products were visualized on ethidium bromide-stained gel. The identity of the J7 amplified product was confirmed by BamHI enzymatic digestion.
Table III. Proportion of RAG-expressing CD5+ Bm2 cells in human tonsils
Synthesis of RAG1 and RAG2 proteins was then established to confirm that RAG1 and RAG2 mRNA transcription resulted in the expression of the complete enzymatic complex in the same cell. CD5+ and CD5– Bm2 lymphocytes were sorted, and immunofluorescence analysis was performed (Fig. 2C) on cytocentrifuged cells. Although (Fig. 2D) most of these cells were negative for both proteins (25.5 ± 12.7% RAG1–RAG2–), or positive for only one (28.1 ± 3% RAG1+RAG2–, and 13 ± 1.8% RAG1–RAG2+), we observed that one-third of them coexpressed both proteins (33.2 ± 12.9% RAG1+RAG2+). Although RAG1 and RAG2 were detectable in the cytoplasm, both proteins were predominantly colocalized to the nucleus when examined by confocal microscopy (data not shown). Because CD5+ Bm2 cells represented 5.4–9.1% of the total B cells, 1.6–2.1% of the B lymphocytes would thus be CD5+ Bm2 lymphocytes expressing the whole enzymatic machinery for V(D)J recombination (Table IV), confirming the mRNA data. These findings indicate that very few B cells may rearrange outside the GCs. To definitively underscore the functionality of RAG1 and RAG2, LM-PCR was performed according to previously described procedure (24) on DNA isolated from sorted CD5+ and CD5– Bm2 lymphocytes. Putative DNA breaks on recombination signal sequences due to recombinase activity were determined using primers upstream the 5' end of each J3 to J7 and J2 to J5 genes paired with primer encompassing the LM-PCR linker and the recombination signal sequences. The expected product lengths are indicated in Table I. We found a specific PCR product corresponding to a recombination of the J7 gene in the CD5+ but not in the CD5– Bm2 population (Fig. 2E). The PCR-amplified fragment contained a BamHI cleavage site. Digestion of the PCR product with BamHI produced two fragments at the expected size, thus confirming the identity of the amplified product (Fig. 2E). This latter was also sequenced. The alignment with the 3' end upstream the J7 gene of the genomic DNA further confirmed the specificity of the product (data not shown). Recombination of the J4 and J5 genes were also found in some CD5+ Bm2 cell samples. Both products were also verified as specific by sequencing (data not shown). These data demonstrate the presence of active V(D)J recombination enzyme in the CD5+ subset of the Bm2 lymphocyte subpopulation.
Table IV. Expression of RAG1 and RAG2 proteins in CD5+ Bm2 lymphocytes
It may be argued that RAG-positive cells in peripheral lymphoid organs could in fact be immature B cells (28) and, specifically, transitional B cells (29). Moreover, transitional type 2 cells are located in the follicles (30). It was thus important to ascertain whether the IgD+CD5+ Bm2 cells were premature or mature B cells. Based on surface expression of IgD and IgM, discrete mouse B cell subpopulation may be identified. Transitional type 2 B lymphocytes are IgD positive and IgM bright, and express a high level of CD21. They can be distinguished from mature B cells which are IgD bright, IgM dull, and CD21 positive (30). Although transitional B cells have not been evidenced in human, we performed fine flow cytometric studies on human tonsil B lymphocytes using these definitions. When gated on IgD-positive cells, we identified two discrete subpopulations in the CD21/IgM bi-parametric analysis (Fig. 3A). The expression of CD5 was determined on their surface. We found that the IgD+IgM+CD21+ mature B cells contained a CD5+ subset, in contrast to the IgD+IgMbrightCD21bright transitional type 2-like B cells, which lack CD5 expression (Fig. 3A). These phenotypical analyses indicate that premature transitional type 2 B lymphocytes might be present in human tonsils but do not express CD5. When gated on IgD-negative cells, B lymphocytes displayed a marginal zone phenotype (IgD–IgMbrightCD21bright) without expression of CD5, whereas transitional type 1 population (IgD–IgMbrightCD21–) was undetected (our unpublished observations).
FIGURE 3. Mature differentiation state of the CD5+ Bm2 cell subset. A, Anti-IgD, anti-IgM, anti-CD21, and anti-CD5 Abs were used to distinguish the Bm cell population (IgD+IgM+CD21+) from the premature type 2 transitional B cell (BT2) population (IgD+IgMbrightCD21bright) by flow cytometry. CD5 expression was analyzed on both populations. B, The maturation status of the IgD+CD38+CD5+ Bm2 cells were further evidenced by flow cytometric analysis using a combination of anti-IgD, anti-CD38, and anti-CD5 Abs associated with anti-IgM, anti-CD23, anti-CD44, anti-CD10, anti-CD77, anti-CD27, or anti-CD95 Abs.
To establish the mature status of the CD5+ B cell subpopulation, we studied the expression of the surface markers that identify the seven B cell subpopulations in human tonsils (10). When gated on CD5+ lymphocytes, we found that the IgD+CD38+ B cell subpopulation appeared to be IgM+, CD23+, and CD44+, but CD10, CD77, CD27, and CD95 negative (Fig. 3B). Such observations confirm the mature status of the CD5+ subpopulation in which RAG1 and RAG2 mRNA are inducible. Although CD44 is expressed on naive and memory cells, the absence of CD27 demonstrates that they do not belong to the memory pool. They do not belong either to the centroblast pool, because they do not express CD77. Furthermore, expression of IgM and CD23 by the IgD+CD38+CD5+ B cells is evidence for activation of a preswitch naive subset. Finally, the absence of CD95 suggests that they are not prone to apoptosis through interaction with CD95L on T cells.
To directly ascertain whether CD5 could be associated with RAG expression and to determine how this might work, we conducted in vitro stimulation on mature tonsil B lymphocytes. Our aim was to assess whether induction of CD5 correlated with RAG1 and RAG2 mRNA transcription. It has long been known that membrane CD5 can be induced on human B cell following stimulation with PMA (17, 31, 32). PMA stimulation of total tonsil B lymphocytes indeed generated an increase in CD5+ B cell number that plateaued after 30 h (Fig. 4A). However, although RAG1 and RAG2 mRNA were detected in isolated cells before activation, mRNA for the two genes was down-regulated after stimulation when measured by nested RT-PCR (Fig. 4B). This suggests that specific signals might be required for the cells to induce CD5, and to permit the transcription of RAG1 and RAG2 mRNA. The effect of direct antigenic stimulation was, therefore, investigated, and tonsillar B cells activated with anti-IgM Abs. Because this activation leads to a high rate of apoptosis over 30 h (33, 34), hampering analysis of RAG expression, the investigation was limited to the first 30 h of stimulation. The frequency of CD5+ B lymphocytes increased until 30 h (Fig. 4A), consistent with the detection of RAG1, but not RAG2 mRNA, at that time (B). Inasmuch as CD40/CD40L interaction has also been shown to induce CD5 on B cell surface (35), we cultured tonsillar B lymphocytes on CD40L-transfected fibroblasts. The experiments confirmed that, following 3 days of stimulation, CD5 was induced on a subpopulation of cells for up to 6 days of culture (Fig. 4A), but without up-regulation of RAG1 and RAG2 mRNA transcription (B). These results indicate that specific conditions may be combined for CD5 to be induced and to permit V(D)J rearrangement. From the data reported with mouse spleen B cells (3), but in contrast to human GC B cells (13), RAG transcription may be up-regulated in mature B cells in vitro following costimulation. We thus looked for RAG transcripts in tonsillar B lymphocytes following costimulation with anti-IgM and CD40L, known to trigger survival of anti-IgM-induced apoptosis (36). As may be predicted from the previous experiments, CD5 expression was induced increasingly, for up to 5 days of culture (Fig. 4A). Interestingly, at day 4, this costimulation system extended the induction of both RAG1 and RAG2 mRNA transcripts (Fig. 4B).
FIGURE 4. Regulated expression of RAG1, RAG2, and CD5. Tonsillar B cells were isolated and cultured with PMA at 30 ng/ml or with anti-IgM-coated beads at 1 μg/ml for 30 h, or on CD40L-transfected fibroblasts alone or together with 1 μg/ml anti-IgM-coated beads for 6 days. A, Frequencies of CD5+ B cells were determined by flow cytometry. Mean ± SD of three experiments. B, RAG1 and RAG2 mRNA expression of cultured B cells was analyzed by nested RT-PCR, and that of GAPDH by RT-PCR. A representative data for each culture condition is shown.
To more definitively elucidate the association between CD5 expression and RAG transcription, CD5-nonexpressing Bm1 and Bm2 lymphocytes were sorted and stimulated with anti-IgM plus CD40L. A CD5+ subpopulation was generated anew, and, at days 4 and 5, RAG1 and RAG2 mRNA were coexpressed (Fig. 5A). These data demonstrate that RAG1 and RAG2 can be concomitantly up-regulated with the induction of CD5 expression. To identify those cells expressing RAG, we repeated the experiments by sorting CD5+ and CD5– B cells after 5 days of culture, and analyzed RAG transcripts in both populations. As shown in Fig. 5B, cells turned positive for CD5 expressed both RAG1 and RAG2 transcripts, whereas cells that remained CD5 negative expressed RAG1 only. These data demonstrate that, when activated, only CD5-induced IgD-positive cells were capable of rearrange their BCR-encoding genes.
FIGURE 5. Induced expression of RAG1 and RAG2 in CD5+ acquired Bm2 cells. A, IgD+CD38–/+CD5– tonsil B cells were sorted and cultured on CD40L-transfected fibroblasts with anti-IgM-coated beads at 1 μg/ml for up to 5 days. Frequency of CD5+ induced B cells was evaluated by flow cytometry, and RAG1 and RAG2 mRNA expression was analyzed by nested RT-PCR and that of GAPDH by RT-PCR. B, IgD+CD38–/+CD5– tonsil B cells were sorted and cultured as above for 5 days. Then, cells turned positive for CD5 and cells still negative were sorted again, and expression of RAG1, RAG2, and GAPDH mRNA was analyzed as above.
Discussion
The origin and functions of the CD5+ B cells have fueled extensive studies and discussions. Two mutually exclusive hypotheses have been proposed as to the origin of CD5+ B cells. The first is the lineage hypothesis, which contends that some B cell precursors are destined to become CD5+ B cells (reviewed in Ref. 37). In contrast, the differentiation hypothesis implies that all B cells have the potential to acquire CD5+ B lymphocyte characteristics (reviewed in Ref. 38). A new paradigm has recently emerged that reconciles both hypotheses, proposing that two subsets of CD5 B cells may exist in mice (39) as well as in humans (17). One subset of B cells would constitutively express CD5, whereas the other may be induced to do so in response to specific activation. The present work extends the knowledge of the latter subpopulation by demonstrating that, under specific stimulation conditions, some activated tonsillar B cells may be induced to express CD5 and, at the same time, gain the capacity to revise their BCR.
Based on IgD and CD38 expression, we have identified different subpopulations in human tonsils and established that RAG expression was not restricted to GC cells. Our study revealed that activated Bm2 cells, pre-GC Bm2' cells, and early Bm5 memory cells all displayed RAG1 and RAG2 mRNA. However, when subdivided according to the expression of CD5, the Bm2 CD5+ cells was identified as a subpopulation that subsists outside GC and has the potential to undergo secondary Ig gene recombination. Indeed, the results have shown that, for this B cell subset, RAG1 and RAG2 transcripts and proteins are expressed at the single-cell level. Frequencies of double-positive cells in the different B cell subsets are in the same range as those determined by others (16). Cells negative for both enzymes or positive for only one were also found. This might be attributable to the different means of RAG1 and RAG2 regulation. Although RAG1 expression mainly depends upon the cell activation status (40), that of RAG2 is linked to the cell cycle phase leading to its accumulation during the G1 phase (41). Because isolated B cells were not synchronized, all of the patterns of RAG1 and RAG2 expressions could be observed during the kinetics study. Furthermore, we have confirmed their functional recombinase activity. Thus, freshly isolated CD5+ Bm2 cells had J7, J4, and/or J5 signal breaks. Because detection of L chain recombination intermediates as well as increased gene usage have been correlated to receptor revision (13), our results indicate that CD5+ Bm2 cells have a functional V(D)J recombination machinery. From these combined data, we conclude that, before the cells enter the GC in vivo, naive activated Bm2 lymphocytes may initiate V(D)J rearrangement when they express the CD5 molecule. This conclusion is strengthened by our in vitro experiments. They showed that, when stimulated with anti-IgM plus CD40L, some IgD+CD38+CD5– Bm2 cells were encouraged to express RAG1 and RAG2, but this behavior was restricted to those turned into CD5 positive.
It has been shown that a subset of transitional B cells retains the capacity to up-regulate RAG genes when stimulated via their BCR in an appropriate environment (42). Although transitional B cells have not been underscored in humans, and do not express CD5 in mice (37), it was of importance to prove that the presence of RAG+ B cells in human tonsils could not be ascribed to newly immigrated immature-type B lymphocytes. According to the description by Loder et al. (30), we found transitional type-2-like B cells (IgD+IgMbrightCD21bright) in the IgD+ population, but their functional characteristics remain to be determined. However, we observed that these lymphocytes do not express CD5 and that all the CD5+ cells belong to the mature population (IgD+IgM+CD21+). Furthermore, expression of IgM, CD23, and CD44, and lack of CD10, CD77, and CD27 expression strongly confirm the preswitch characteristics of the IgD+CD38+CD5+ B cells, i.e., before they enter the GC (10). Taken together, these data demonstrate the mature status of the CD5+ B cells expressing RAG1 and RAG2 genes and thus undergoing V(D)J recombination outside GCs. However, it should be noted that, in one case, the CD5– Bm2 lymphocytes also expressed RAG1 and RAG2 mRNA. Because CD5 is transiently induced, it is likely that these latter cells may be CD5-induced Bm2 lymphocytes that have achieved V(D)J rearrangement, and for which CD5 expression has been terminated, but not that of RAG1 and RAG2. Moreover, the presence of RAG1 and RAG2 transcripts in both CD5+ and CD5– early Bm5 cells indicates that some postswitch B lymphocytes may revise their BCR independently of CD5 expression. However, the recombinase activity of the RAG protein in these cells remains to be demonstrated.
In contrast to activated naive Bm2 lymphocytes, resting naive Bm1 cells were negative for RAG1 and RAG2 mRNA expression. This indicates that V(D)J recombination might be induced by activation. We found that stimulation with PMA, known to be a potent cell activator and inducer of CD5 expression (31, 32), reduces RAG expression. This is consistent with previous work (43) and supports the notion that specific activation, such as Ag encounter, might be required for RAG up-regulation, as recently suggested (16). However, from our in vitro studies, BCR engagement alone cannot promote RAG mRNA up-regulation. In RAG1–/– mice reconstituted with RAG1-GFP splenic B cells, in vitro stimulation with either IL-1 and anti-IgM or IL-7 with anti-CD40 failed to induce GFP signal, suggesting that other mechanisms might contribute to receptor revision (15). A second signal delivered by T lymphocytes seems to be necessary. Thus, BCR and CD40 coligation triggers expression of RAG1 and RAG2. Interestingly, this was induced only in the naive mature B cells turned positive for CD5, indicating that CD5+ B cells expressing RAG mRNA are likely to belong to the inducible CD5 B cell subpopulation, although the involvement of the "constitutive" CD5+ B cell pool (17) cannot be totally discounted. Whether RAG+ B1 cells found in the peritoneal cavity of mice are part of the "inducible" CD5+ B cells remains also to be determined (22). It is now clear that induction of CD5 expression on human mature B cells represents part of the differentiation pathway and is a particular feature of certain B cells that are susceptible to special inductive signals (44). The decreased frequency of CD5+ B cells observed once the cells entered the GC lends support to this hypothesis, indicating that CD5 expression might be lost at this stage of development. However, we cannot rule out the possibility that a selection may also operate resulting in a diminished expression of CD5.
The CD5 expression, associated with RAG1 and RAG2 up-regulation, raised the question as to how appearance of CD5 may influence RAG1 and RAG2 expression. Other questions are the nature of the B cell subpopulation affected and the physiological role this process plays in vivo. It has been shown that receptor editing in immature B cells is induced because of distinct BCR signaling capacities in these cells compared with mature B cells. In immature B cells, engagement of the BCR leads to small increases in intracellular free Ca2+ and to the induction of receptor editing. In contrast, BCR engagement in mature B cells results in high elevation in intracellular Ca2+ concentration leading to the up-regulation of CD86, CD69, and MHC class II, but not V(D)J recombination (45). CD5+ B lymphocytes in mice have been shown to be hyporesponsive to BCR ligation compared with their CD5– B cell counterpart. They have a reduced capacity for intracellular Ca2+ mobilization, diminished proliferation, and increased apoptosis after BCR cross-linking due to a negative role for CD5 in the BCR signaling (23, 46). CD5 reduces strong BCR-mediated signaling and permits only limited increases in intracellular free Ca2+ (23). Similarly, CD5 expression may increase BCR threshold in human Bm2 lymphocytes. In this setting, such activated cells would respond like immature B cells with respect to Ca2+ mobilization. Low intracellular Ca2+ mobilization would thus promote the induction of RAG expression, leading ultimately to receptor revision. Therefore, the physiological relevance of this process may be as follows: whereas clones of B cells with high BCR affinity would be permitted to migrate into the dark zone and proliferate as centroblasts, B cells with very low BCR affinity would be deprived of rescuing signals, and thus deleted by apoptosis. We hypothesize that B cell clones with intermediate BCR affinity receive signals not strong enough to induce maturation into centroblast cells but sufficient enough to avoid deletion and trigger CD5 expression. Together with CD40 engagement, this will promote V(D)J rearrangement. When high-affinity revised BCR is expressed, BCR engagement would trigger maturation of the cells into centroblasts and in the process turn off RAG expression (13), and possibly CD5, which is known to be transiently induced under T-dependent activation (47). The relevance of this process may be to raise the affinity of the BCR up to the level required for maturation and differentiation. Our data suggest that BCR revision is not confined to a specific subpopulation but rather may involve a broad spectrum of B cells. Any B cell may express CD5, depending on its BCR affinity and/or specificity and, thereby, acquire the potential to revise its BCR. This is consistent with the hypothesis that CD5+ B cells may generate CD5– B cells and contribute to the heterogeneity of cell populations that proliferate in the GC (48).
This phenomenon may also have pathophysiological consequences. Modification of the BCR in anergic CD5+ B cells in autoimmune transgenic HEL/anti-HEL model has been observed (20). BCR engagement through low-affinity auto-Ags generates weak mobilization and reduced proliferation (49), which are ameliorated when the mice are bred onto a CD5–/– background. These CD5+ B cells manifest similarities with anergic cells. For example, both populations are stimulated by Ag (50, 51), respond suboptimally to IgM cross-linking (49, 52), show evidence of secondary receptor recombination (22, 53), and are excluded from the GCs (reviewed in Ref. 54). An obvious relationship links anergy and V(D)J rearrangement in maintaining B cell tolerance, as well as anergy-like state and CD5+ B cell development (20). Interestingly, we observed that the CD5+ Bm2 cells lack CD95 expression and display lower levels of IgM compared with the CD5– counterpart (our unpublished observations). These cells, which are protected from T-dependent apoptosis, might be in an anergic-like state.
CD5+ B cells constitute a major source of natural auto-Abs, although they require exposure to Ag for maturation into Ab-secreting cells (51). Furthermore, anergy-inducing BCR engagement is often associated with receptor editing (53, 55), suggesting that a hierarchy of decreasing signaling thresholds is required in BCR/Ag interactions for the induction of editing, anergy, and selection of CD5+ or CD5– B cells (56). Therefore, the accumulation of CD5+ B cells in human autoimmune diseases (57, 58) may reflect Ag-driven activation and subsequent anergy-like state induction of autoreactive B cells. From observations described herein, mechanisms contributing to the maintenance of tolerance in anergic CD5+ B cells may include the control of V(D)J recombination at the CD5+ Bm2 state. The control of CD5 expression and/or receptor revision before the cells enter the GC may be faulty in autoimmune disorders and leads to the emergence of auto-Ab-secreting cells.
Acknowledgments
We acknowledge Laetitia Le Pottier for her skillful and cheerful assistance. The editorial assistance of Prof. Rizgar A. Mageed is greatly appreciated Thanks are also due to Simone Forest and Cindy Séné for secretarial assistance.
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.
1 This work was supported by grants from Ministère de l’Enseignement Supérieur et de la Recherche and from Académie Nationale Fran?aise de Médecine.
2 Address correspondence and reprint requests to Prof. Pierre Youinou, Laboratory of Immunology, Brest University Medical School Hospital, BP 824, F 29609 Brest, France. E-mail address: youinou@univ-brest.fr
3 Abbreviations used in this paper: GC, germinal center; Bm, mature B; FMZ, follicular mantle zone; HEL, hen egg lysosyme; biot, biotinylated; LM-PCR, ligation-mediated PCR; ECD, energy-coupled dye.
Received for publication August 25, 2004. Accepted for publication February 16, 2005.
References
Hertz, M., D. Nemazee. 1997. BCR ligation induces receptor editing in IgM+IgD– bone marrow B cells in vitro. Immunity 6:429.
Radic, M. Z., M. Zouali. 1996. Receptor editing, immune diversification, and self-tolerance. Immunity 5:505.
Hikida, M., M. Mori, T. Takai, K. Tomochika, K. Hamatani, H. Ohmori. 1996. Reexpression of RAG-1 and RAG-2 genes in activated mature mouse B cells. Science 274:2092.
Han, S., B. Zheng, D. G. Schatz, E. Spanopoulou, G. Kelsoe. 1996. Neoteny in lymphocytes: Rag1 and Rag2 expression in germinal center B cells. Science 274:2094
Papavasiliou, F., R. Casellas, H. Suh, X. F. Qin, E. Besmer, R. Pelanda, D. Nemazee, K. Rajewsky, M. C. Nussenzweig. 1997. V(D)J recombination in mature B cells: a mechanism for altering antibody responses. Science 278:298.
Kelsoe, G.. 1999. V(D)J hypermutation and receptor revision: coloring outside the lines. Curr. Opin. Immunol. 11:70.
Nemazee, D., M. Weigert. 2000. Revising B cell receptors. J. Exp. Med. 191:1813.
de Wildt, R. M., R. M. Hoet, W. J. van Venrooij, I. M. Tomlinson, G. Winter. 1999. Analysis of heavy and light chain pairings indicates that receptor editing shapes the human antibody repertoire. J. Mol. Biol. 285:895.
Wilson, P. C., K. Wilson, Y. J. Liu, J. Banchereau, V. Pascual, J. D. Capra. 2000. Receptor revision of immunoglobulin heavy chain variable region genes in normal human B lymphocytes. J. Exp. Med. 191:1881.
Liu, Y. J., C. Arpin. 1997. Germinal center development. Immunol. Rev. 156:111.
Lindhout, E., G. Koopman, S. T. Pals, C. de Groot. 1997. Triple check for antigen specificity of B cells during germinal centre reactions. Immunol. Today 18:573.
Giachino, C., E. Padovan, A. Lanzavecchia. 1998. Re-expression of RAG-1 and RAG-2 genes and evidence for secondary rearrangements in human germinal center B lymphocytes. Eur. J. Immunol. 28:3506.
Meffre, E., F. Papavasiliou, P. Cohen, O. de Bouteiller, D. Bell, H. Karasuyama, C. Schiff, J. Banchereau, Y. J. Liu, M. C. Nussenzweig. 1998. Antigen receptor engagement turns off the V(D)J recombination machinery in human tonsil B cells. J. Exp. Med. 188:765.
Hertz, M., V. Kouskoff, T. Nakamura, D. Nemazee. 1998. V(D)J recombinase induction in splenic B lymphocytes is inhibited by antigen-receptor signalling. Nature 394:292.
Igarashi, H., N. Kuwata, K. Kiyota, K. Sumita, T. Suda, S. Ono, S. R. Bauer, N. Sakaguchi. 2001. Localization of recombination activating gene 1/green fluorescent protein (RAG1/GFP) expression in secondary lymphoid organs after immunization with T-dependent antigens in rag1/gfp knock-in mice. Blood 97:2680.
Girschick, H. J., A. C. Grammer, T. Nanki, M. Mayo, P. E. Lipsky. 2001. RAG1 and RAG2 expression by B cell subsets from human tonsil and peripheral blood. J. Immunol. 166:377.
Youinou, P., C. Jamin, P. M. Lydyard. 1999. CD5 expression in human B-cell populations. Immunol. Today 20:312.
Dono, M., V. L. Burgio, C. Tacchetti, A. Favre, A. Augliera, S. Zupo, G. Taborelli, N. Chiorazzi, C. E. Grossi, M. Ferrarini. 1996.
Subepithelial B cells in the human palatine tonsil. I. Morphologic, cytochemical and phenotypic characterization. Eur. J. Immunol. 26:2035.
Abe, M., K. Tominaga, H. Wakasa. 1994. Phenotypic characterization of human B-lymphocyte subpopulations, particularly human CD5+ B-lymphocyte subpopulation within the mantle zones of secondary follicles. Leukemia 8:1039.
Hippen, K. L., L. E. Tze, T. W. Behrens. 2000. CD5 maintains tolerance in anergic B cells. J. Exp. Med. 191:883.
Tze, L. E., E. A. Baness, K. L. Hippen, T. W. Behrens. 2000. Ig light chain receptor editing in anergic B cells. J. Immunol. 165:6796.
Qin, X. F., S. Schwers, W. Yu, F. Papavasiliou, H. Suh, A. Nussenzweig, K. Rajewsky, M. C. Nussenzweig. 1999. Secondary V(D)J recombination in B-1 cells. Nature 397:355.(Sophie Hillion, Alain Sar)
Previous studies have indicated that mature B cells reactivate secondary V(D)J recombination inside and outside the germinal center (GC) of peripheral lymphoid organs. The nature of the B cells undergoing Ig rearrangement before they enter GC is unknown. In this study, we present evidence that activated mature CD5-positive human tonsil B cells coexpress both RAG1 and RAG2 mRNA and protein, and display DNA cleavage resulting from their recombinase activity. Furthermore, in vitro activation of CD5-negative naive mature B cells by IgR and CD40 cross-linking induces expression of CD5 on a subset of cells, and leads to the up-regulation of RAG1 and RAG2 only in cells turned positive for CD5. Thus, RAG gene expression is closely related to CD5 expression outside GCs. These data suggest that CD5 is associated with receptor revision in activated mature B cells and likely to promote expression of suitable IgR capable of initiating the GC reaction.
Introduction
Secondary V(D)J rearrangement of Ig genes has been extensively described in immature bone marrow B cells. This process is referred to as receptor editing, contributes to the maintenance of immunological tolerance (1), and could rescue potentially autoreactive B cells from apoptosis (2). Receptor editing is initiated upon BCR engagement by Ag, resulting in the up-regulation of RAG gene expression, which then participates in editing the autoreactive BCR-encoding Ig V(D)J rearrangements. Secondary V(D)J recombination has been shown to also occur in murine spleen and lymph node B cells in response to immunization (3, 4, 5). This peripheral V(D)J rearrangement, termed receptor revision, might contribute to the generation of high-affinity Abs in germinal centers (GCs)3 following the process of somatic hypermutation in response to stimulation by T-dependent Ags (6, 7).
There is evidence that human B cells also undergo secondary Ig recombination in the periphery (8, 9). Seven peripheral subpopulations of B cells have been identified in human tonsils based on the surface expression of CD38 and IgD in conjunction with other markers (reviewed in Ref. 10), which have led to the proposition of a model of T cell-dependent mature B cell differentiation. The naive mature B (Bm) cells Bm1 (CD38–IgD+CD23–) and Bm2 (CD38+IgD+CD23+) would be activated in extrafollicular areas through interaction with interdigitating cells and Ag-specific T cells. Activated blasts may either terminally differentiate into low-affinity Ab-forming cells or become GC founder cells Bm2' (CD38++IgD+). In GCs, Bm2' cells would differentiate into centroblast Bm3 (CD38++IgD–CD77+), in which somatic hypermutation in V gene region might take place during proliferation. This process may generate cells bearing BCRs with various affinities. The mutant cells will be selected during their differentiation into centrocytes Bm4 (CD38++IgD–CD77–), depending on their BCR affinity for the Ag that is trapped as an immune complex on the surface of follicular dendritic cells. High-affinity Bm4 may interact strongly with the Ag, process, and present it to GC T cells. These T cells will be induced to express CD40L, to secrete cytokines, and to promote survival, proliferation, and isotype switching of the B cells. These activated B cells will thus terminally differentiate either into early memory (CD38+IgD–) and memory Bm5 (CD38–IgD–) cells or into high-affinity Ab-forming cells. In contrast, autoreactive Bm4 cells would be deleted, because they do not receive survival signal from T cells (11). Mutant Bm4 cells with low affinity for Ag may have the potential to revise their receptor due to an increased RAG gene transcription (12). This could raise the affinity of the BCR for the Ag (8, 9). In this setting, strong BCR engagement would switch off RAG gene expression (13), suggesting that receptor revision, which might be involved in affinity maturation of Abs, will be terminated when BCR is strongly cross-linked. This process will permit the positive selection of high-affinity B cells to terminally differentiate.
Recently, RAG expression in peripheral B cells has been observed outside the GC in a number of mouse models (14, 15) and in humans (16). In the human model, RAG-positive B cells were found in the follicular mantle zone (FMZ) of the GCs where naive Bm1 and Bm2 cells are positioned (10). The T cell marker CD5 is expressed by <10% of the B cells in the adult human spleen and <30% in lymph nodes (17). It is interesting that the CD5+ B cell population is also enriched in the FMZ (18, 19).
Interestingly, in hen egg lysosyme (HEL)/anti-HEL transgenic mice, CD5 expression is observed on mature anergic B cells (20). Moreover, immature anergic B cells may activate the machinery for V(D)J rearrangement to maintain B cell tolerance (21). Although CD5+ B cells have been shown (22) to express RAG in the peritoneal cavity, it is unknown whether CD5 is involved in inducing the recombination process. Mature anergic B cells that fail to rearrange may express CD5 to negatively regulate the BCR and avoid elimination, yet maintaining tolerance to self-Ag. However, by raising the threshold required for activation (23), CD5 may also contribute to the triggering of the BCR revision. This view implies that CD5 expression in the FMZ could be induced on mature B cells with BCR affinity insufficient to enter GC. RAG expression would, therefore, be found in activated cells expressing CD5. The present study was designed to test this hypothesis.
Materials and Methods
Preparation of B cells
Tonsils were obtained from 5- to 18-year-old children undergoing routine tonsillectomy. Tissues were minced up, diluted in PBS, and filtered to deplete larger cells and clumps. Filtered cells were layered onto Ficoll-Hypaque density medium and centrifuged for 30 min at 450 x g. Cells were then incubated with neuraminidase-treated SRBC for 1 h at 4°C. Depletion of T cells was achieved by a second round of centrifugation on Ficoll-Hypaque for 25 min.
Flow cytometry analysis
FITC-mAbs to human CD19 (clone J4.119), CD21 (clone BL13), CD10 (clone ALB1), CD95 (clone UB-2), CD27 (clone MP271) and CD23 (clone 9P25), rat unlabeled Ab to CD77 (clone 38-13) visualized with FITC-goat anti-rat, PE-linked to cyanin 5 (PC5)-Abs to CD38 (clone LS198), and CD5 (clone BL1a), and biotinylated (biot) anti-CD5 (clone UCHT2) visualized with PE-Texas Red energy-coupled dye (ECD)-streptavidin, were from Beckman Coulter. FITC-anti-IgD (clone IA6-2), PE-anti-IgD, PE-anti-IgM (clone G20127), and biot-anti-IgM visualized with ECD-conjugated streptavidin were from BD Biosciences. Four-color analysis and sorting were performed on an Epics Elite flow cytometer (Beckman Coulter).
RNA isolation and RT-PCR amplification
Total RNA was extracted from 105 B cells using the RNable reagent (Eurobio) and reverse transcribed in 20 μl with Superscript II RNase H-reverse transcriptase (Invitrogen Life Technologies). For RT-PCR amplifications, nested reactions were performed with 1 μl of cDNA using the primer pairs listed in Table I with TaqDNA polymerase (Invitrogen Life Technologies) as follows: in the first round of PCR, cDNA was amplified for 25 cycles of 30 s at 94°C, 1 min at 56°C, and 1 min at 72°C with a final 10-min extension at 72°C. The second PCR round was for 35 cycles of 30 s at 94°C, 1 min at 56°C, and 1 min at 72°C with a final extension step at 72°C for 10 min. For the GAPDH RT-PCR amplification, only one round of PCR was conducted for 40 cycles. Because the GAPDH primers spanned a short intron of 100 bp, genomic DNA could be amplified, but easily distinguishable from the specific cDNA product. In such cases, samples were excluded from the RAG studies. Moreover, RAG1 and RAG2 first-round PCR primers spanned an intron at 5168 and 1174 bp, respectively, which thus could not amplify contaminating genomic DNA. Taken together, these precautions certify that the PCR products originate specifically from mRNA. RT-PCR products were analyzed on 2% agarose gels stained with ethidium bromide. The expected size for each specific product is indicated in Table I.
Table I. Sequences of oligonucleotides used as primers for the amplification of cDNA in RT-PCR and nested RT-PCR, and of DNA in LM PCR
Single-cell protocol
Individual IgD+CD38+CD5+ Bm2 cells were sorted into PCR tubes containing 10 μl of reverse transcriptase buffer (1x first-strand buffer (Invitrogen Life Technologies), 5 μM random hexamer, 0.5x reverse transcriptase buffer, 0.01 M DTT, 0.5 mM dNTP (all from Promega)), using the flow cytometer outfitted with the Autoclone single-cell deposition unit (Beckman Coulter). mRNA conversion from a single cell was conducted with Superscript II RNase H-reverse transcriptase (Invitrogen Life Technologies) for 2 h at 42°C. A nested RT-PCR for RAG1, RAG2, and GAPDH was performed with 25 cycles of 30 s at 94°C, 1 min at 56°C, and 1 min at 72°C with a final 10-min extension at 72°C in the first round of PCR. The second PCR round was for 40 cycles of 30 s at 94°C, 1 min at 56°C, and 1 min at 72°C with a final extension at 72°C for 10 min. Frequencies of cells positive for one transcript (RAG1+2– or RAG1–2+), negative for both (RAG1–2–), or positive for both (RAG1+2+) were calculated.
Ligation-mediated PCR (LM-PCR)
DNA was extracted from 105 cells using the DNAzol kit (Invitrogen Life Technologies), and 500 ng ligated to the BW1/2 linker at 20 pmol with T4 DNA ligase (Promega) in ligation buffer (66 mM Tris (pH 7.5), 5 mM MgCl2, 1 mM DTT, and 1 mM ATP) for 16 h at 16°C (24). The BW1/2 linker was obtained by hybridization of 2 nmol BW1 primer with BW2 primer (Table I) for 5 min at 60°C. PCR was performed using TaqDNA polymerase following a denaturation step at 95°C for 15 min. In the first PCR round, linker primer BW1 was used with either J4-1, J5-1, J6-1, or J7-1 primer. Samples were denaturated for 3 min at 94°C and amplified for 15 cycles of 40 s at 94°C, 40 s at 65°C, and 1 min at 72°C, followed by a final 10-min extension at 72°C. After 3 min at 94°C for denaturation, an additional 40 cycles of PCR were performed using the BW3 linker primer with either J4-2, J5-2, J6-2, or J7-2 nested locus-specific primer. The expected size for each specific nested product is indicated in Table I. Similar PCR, with 40 s at 55°C for the hybridization step, were also performed with the Jk locus-specific primers. PCR products were analyzed on 2% agarose gels stained with ethidium bromide. The identity of the amplified products were verified by digestion with BamHI restriction enzyme for the J7 signal end, and by sequence analysis of all the PCR products.
Indirect immunofluorescence
Sorted cells were washed in PBS, centrifuged onto glass slides, fixed in 4% paraformaldehyde, and further prepared as described, with modifications (25). First, to avoid nonspecific fixation, the cells were incubated in PBS containing 0.1% Nonidet P-40 (Sigma-Aldrich), 2% nonfat dry milk, 5% FBS, and 0.02% sodium azide, for 20 min, followed by an additional treatment for 1 h with 50 μg/ml Cohn fragment (Sigma-Aldrich). Cells were then left overnight at 4°C with goat anti-RAG2 Ab (Santa Cruz), incubated with biot donkey anti-goat IgG (Jackson ImmunoResearch Laboratories) and visualized with tetramethylrhodamine isothiocyanate-streptavidin (Jackson ImmunoResearch Laboratories). Finally, the cells were left overnight at 4°C with rabbit anti-RAG1 Ab (Santa Cruz), followed by FITC- donkey anti-rabbit IgG (Jackson ImmunoResearch Laboratories). Cells were examined immediately by fluorescence microscopy using an oil immersion lens (Zeiss; Axioplan), and enumerated as positive for one protein (RAG1+2– or RAG1–2+), negative for both (RAG1–2–), or positive for both (RAG1+2+). Examination by confocal microscopy (TCS NT; Leica) was also performed following staining of the cells for RAG1 or RAG2 by FITC labeling, using FITC-donkey anti-goat (Jackson ImmunoResearch Laboratories) to visualize anti-RAG2 Ab, and propidium iodide nucleus counterstaining, to evaluate the localization of each protein. The human 697 pre-B cell line and Hep-2 epithelial carcinoma cell line were used as positive and negative controls for the RAG protein detection, respectively (data not shown).
Cell culture
Cells were cultured in RPMI 1640 medium (Invitrogen Life Technologies) supplemented with 10% heat-inactivated FCS (Invitrogen Life Technologies), 2 mM L-glutamine (Invitrogen Life Technologies), 200 U/ml penicillin (Rh?ne-Poulenc), and 100 μg/ml streptomycin (Diamant). A total of 105 murine NIH3T3 fibroblasts untransfected (control), or transfected with human CD40L gene, were incubated with mitomycin C (Sigma-Aldrich) and used for stimulation of B cells (26). Transfectants were grown in DMEM medium (Invitrogen Life Technologies) supplemented with 10% FCS, 2 mM L-glutamine, and 10 μg/ml gentamicin (Rh?ne-Poulenc). Isolated B cells were dispensed at 106 cells per well in a final volume of 1 ml with, or without, anti-IgM Ab-coated beads at 1 μg/ml. Cells were also stimulated with anti-IgM Ab-coated beads alone, or with 30 ng/ml PMA.
Results
Based on IgD and CD38 expression, all mature B cell subpopulations were sorted (Fig. 1A), and RAG mRNA was analyzed by nested RT-PCR. As expected, we found that Bm3/4 GC cells contained RAG1 and RAG2 mRNA. Furthermore, all but naive Bm1 cells and Bm5 memory cells expressed both RAG transcripts (Fig. 1B), supporting the notion that some B cells outside GC may also undergo secondary rearrangement. Specifically activated naive Bm2 cells, pre-GC Bm2' cells, and early Bm5 memory cells seemed to have the machinery to rearrange.
FIGURE 1. RAG1, RAG2, and CD5 expression in tonsil B cell subsets. A, Tonsillar B cell subsets were defined by the surface expression of CD19, IgD, and CD38. The gates for each subset are shown. B, Specific RAG1 and RAG2 mRNA of sorted subsets were amplified by nested RT-PCR and that of GAPDH by RT-PCR. C, Tonsillar B cell subsets were analyzed for the membrane expression of CD5. D, The CD5 expression was evaluated in the total B cells, and the frequency of each B cell subset in the CD5-positive population was determined. E, The CD5+ and CD5– fractions of each B cell subset were sorted, and the RAG1, RAG2, and GAPDH mRNA were amplified.
To determine whether the presence of CD5 influences RAG transcription, we first examined each subpopulation for the expression of CD5 (Fig. 1C). The frequency of CD5+ B lymphocytes increased in the extrafollicular areas from 16.2 ± 4.7% in Bm1 to 18.1 ± 7.4% in Bm2, and to 22.7 ± 10.9% in Bm2' subpopulations (Table II). Remarkably, the frequency decreased once the cells entered the GC, that is, as few as 3.4 ± 1.9% of Bm3/4 cells express CD5. When B cells left the GC, they did not re-express CD5, because there were only 9.2 ± 5.3% of early Bm5 and 9.8 ± 6.4% of Bm5 cells to carry CD5+. We reasoned that, because it is transiently expressed on extrafollicular lymphocytes, CD5 may influence B cell behavior before their entry into the GC.
Table II. Distribution of the different B cell subpopulation in human tonsil B lymphocytes and frequency of the CD5+ B cell subset
Next, we sorted the B cell subpopulations according to their CD5 expression (Fig. 1D) and analyzed the presence of RAG1 and RAG2 mRNA by nested RT-PCR. Interestingly, from the B lymphocytes located in the extrafollicular areas, only the activated Bm2 cells expressing CD5 displayed RAG1 and RAG2 transcripts (Fig. 1E). These observations suggest that the presence of CD5 at this stage of maturation may be associated with the recombination machinery to be operational. In contrast, because both CD5+ and CD5– early Bm5 displayed RAG1 and RAG2 mRNA, CD5 expression does not seems to be influential on V(D)J rearrangement at the memory stage. RAG1 and RAG2 mRNA detection was performed on different tonsil samples. In five of eight, the CD5+ Bm2 lymphocytes expressed RAG1 and RAG2 transcripts. Among them, one sample showed the CD5– Bm2 cells also expressing both RAG gene mRNA.
For the V(D)J recombination machinery to be operational, both RAG1 and RAG2 mRNA must be coordinately expressed in a cell at the same time (27). With this requirement in view, we used single-cell RT-PCR to assay for RAG1 and RAG2 transcript expression. In four different tonsil samples, single CD5+ Bm2 lymphocytes were sorted, and nested RT-PCR was performed to evaluate the transcription of RAG1 and RAG2 mRNA in the same cell. This analysis was restricted to cells expressing GAPDH mRNA to make sure cDNA was present in the samples after sorting, whereas samples with contaminated genomic DNA were excluded (Fig. 2A). We found that 23.5 ± 8.4% of the cells coexpressed RAG1 and RAG2 mRNA, 37.3 ± 10.7 and 14.7 ± 5.8% expressed either RAG1 or RAG2, respectively, whereas 22.9 ± 13.9% expressed neither (Fig. 2B). We noticed that almost one-fourth of the CD5+ Bm2 cells expressed both mRNA, whereas 37.1% of the Bm3/4 cells, 6 and 16.6% of the CD5+ and CD5– early Bm5, respectively, expressed also RAG1 and RAG2 transcripts (data not shown). Given that CD5+ Bm2 cells represented 2.6–13.9% of the total B cells, the CD5+ Bm2 lymphocytes that have a complete machinery for V(D)J rearrangement accounted for 0.7–2% of the total tonsil B lymphocytes (Table III).
FIGURE 2. Single-cell mRNA and protein expression of RAG1 and RAG2 in individual tonsil Bm2 cell subset. A, Individual IgD+CD38+CD5+ tonsil Bm2 cells were sorted with the Autoclone single-cell deposition unit of the Epics Elite flow cytometer. RAG1 and RAG2 mRNA expression were analyzed by nested RT-PCR. Among 40 cells sorted in each experiment, only GAPDH mRNA-positive cells were considered for analysis. A representative example is shown where crosses indicate cells excluded from the RAG1 and RAG2 analysis. B, The frequencies of single RAG1- or RAG2-positive cells as well as the frequencies of RAG1- and RAG2-double-negative or -positive cells are shown. Mean ± SD of four experiments. C, Tonsil Bm2 cells identified by the expression of IgD and CD38 were sorted according to the presence of CD5. Purified B cells were cytocentrifuged and stained for RAG1 and RAG2 protein detection by immunofluorescence. Cells stained without anti-RAG1 and anti-RAG2 Abs were used as controls. D, Among the IgD+CD38+CD5+ tonsil Bm2 cell subpopulation, the frequencies of single RAG1+ or RAG2+ cells, as well as the frequencies of RAG1- and RAG2-double-negative or -positive cells were established by direct enumeration in a randomly selected subset. Mean ± SD of three experiments. E, IgD+CD38+ CD5+ and CD5– tonsil Bm2 cell subsets were sorted, and RAG1 and RAG2 enzymatic activity was evaluated. The J signal breaks were detected by LM-PCR using different sets of specific primers. LM-PCR products were visualized on ethidium bromide-stained gel. The identity of the J7 amplified product was confirmed by BamHI enzymatic digestion.
Table III. Proportion of RAG-expressing CD5+ Bm2 cells in human tonsils
Synthesis of RAG1 and RAG2 proteins was then established to confirm that RAG1 and RAG2 mRNA transcription resulted in the expression of the complete enzymatic complex in the same cell. CD5+ and CD5– Bm2 lymphocytes were sorted, and immunofluorescence analysis was performed (Fig. 2C) on cytocentrifuged cells. Although (Fig. 2D) most of these cells were negative for both proteins (25.5 ± 12.7% RAG1–RAG2–), or positive for only one (28.1 ± 3% RAG1+RAG2–, and 13 ± 1.8% RAG1–RAG2+), we observed that one-third of them coexpressed both proteins (33.2 ± 12.9% RAG1+RAG2+). Although RAG1 and RAG2 were detectable in the cytoplasm, both proteins were predominantly colocalized to the nucleus when examined by confocal microscopy (data not shown). Because CD5+ Bm2 cells represented 5.4–9.1% of the total B cells, 1.6–2.1% of the B lymphocytes would thus be CD5+ Bm2 lymphocytes expressing the whole enzymatic machinery for V(D)J recombination (Table IV), confirming the mRNA data. These findings indicate that very few B cells may rearrange outside the GCs. To definitively underscore the functionality of RAG1 and RAG2, LM-PCR was performed according to previously described procedure (24) on DNA isolated from sorted CD5+ and CD5– Bm2 lymphocytes. Putative DNA breaks on recombination signal sequences due to recombinase activity were determined using primers upstream the 5' end of each J3 to J7 and J2 to J5 genes paired with primer encompassing the LM-PCR linker and the recombination signal sequences. The expected product lengths are indicated in Table I. We found a specific PCR product corresponding to a recombination of the J7 gene in the CD5+ but not in the CD5– Bm2 population (Fig. 2E). The PCR-amplified fragment contained a BamHI cleavage site. Digestion of the PCR product with BamHI produced two fragments at the expected size, thus confirming the identity of the amplified product (Fig. 2E). This latter was also sequenced. The alignment with the 3' end upstream the J7 gene of the genomic DNA further confirmed the specificity of the product (data not shown). Recombination of the J4 and J5 genes were also found in some CD5+ Bm2 cell samples. Both products were also verified as specific by sequencing (data not shown). These data demonstrate the presence of active V(D)J recombination enzyme in the CD5+ subset of the Bm2 lymphocyte subpopulation.
Table IV. Expression of RAG1 and RAG2 proteins in CD5+ Bm2 lymphocytes
It may be argued that RAG-positive cells in peripheral lymphoid organs could in fact be immature B cells (28) and, specifically, transitional B cells (29). Moreover, transitional type 2 cells are located in the follicles (30). It was thus important to ascertain whether the IgD+CD5+ Bm2 cells were premature or mature B cells. Based on surface expression of IgD and IgM, discrete mouse B cell subpopulation may be identified. Transitional type 2 B lymphocytes are IgD positive and IgM bright, and express a high level of CD21. They can be distinguished from mature B cells which are IgD bright, IgM dull, and CD21 positive (30). Although transitional B cells have not been evidenced in human, we performed fine flow cytometric studies on human tonsil B lymphocytes using these definitions. When gated on IgD-positive cells, we identified two discrete subpopulations in the CD21/IgM bi-parametric analysis (Fig. 3A). The expression of CD5 was determined on their surface. We found that the IgD+IgM+CD21+ mature B cells contained a CD5+ subset, in contrast to the IgD+IgMbrightCD21bright transitional type 2-like B cells, which lack CD5 expression (Fig. 3A). These phenotypical analyses indicate that premature transitional type 2 B lymphocytes might be present in human tonsils but do not express CD5. When gated on IgD-negative cells, B lymphocytes displayed a marginal zone phenotype (IgD–IgMbrightCD21bright) without expression of CD5, whereas transitional type 1 population (IgD–IgMbrightCD21–) was undetected (our unpublished observations).
FIGURE 3. Mature differentiation state of the CD5+ Bm2 cell subset. A, Anti-IgD, anti-IgM, anti-CD21, and anti-CD5 Abs were used to distinguish the Bm cell population (IgD+IgM+CD21+) from the premature type 2 transitional B cell (BT2) population (IgD+IgMbrightCD21bright) by flow cytometry. CD5 expression was analyzed on both populations. B, The maturation status of the IgD+CD38+CD5+ Bm2 cells were further evidenced by flow cytometric analysis using a combination of anti-IgD, anti-CD38, and anti-CD5 Abs associated with anti-IgM, anti-CD23, anti-CD44, anti-CD10, anti-CD77, anti-CD27, or anti-CD95 Abs.
To establish the mature status of the CD5+ B cell subpopulation, we studied the expression of the surface markers that identify the seven B cell subpopulations in human tonsils (10). When gated on CD5+ lymphocytes, we found that the IgD+CD38+ B cell subpopulation appeared to be IgM+, CD23+, and CD44+, but CD10, CD77, CD27, and CD95 negative (Fig. 3B). Such observations confirm the mature status of the CD5+ subpopulation in which RAG1 and RAG2 mRNA are inducible. Although CD44 is expressed on naive and memory cells, the absence of CD27 demonstrates that they do not belong to the memory pool. They do not belong either to the centroblast pool, because they do not express CD77. Furthermore, expression of IgM and CD23 by the IgD+CD38+CD5+ B cells is evidence for activation of a preswitch naive subset. Finally, the absence of CD95 suggests that they are not prone to apoptosis through interaction with CD95L on T cells.
To directly ascertain whether CD5 could be associated with RAG expression and to determine how this might work, we conducted in vitro stimulation on mature tonsil B lymphocytes. Our aim was to assess whether induction of CD5 correlated with RAG1 and RAG2 mRNA transcription. It has long been known that membrane CD5 can be induced on human B cell following stimulation with PMA (17, 31, 32). PMA stimulation of total tonsil B lymphocytes indeed generated an increase in CD5+ B cell number that plateaued after 30 h (Fig. 4A). However, although RAG1 and RAG2 mRNA were detected in isolated cells before activation, mRNA for the two genes was down-regulated after stimulation when measured by nested RT-PCR (Fig. 4B). This suggests that specific signals might be required for the cells to induce CD5, and to permit the transcription of RAG1 and RAG2 mRNA. The effect of direct antigenic stimulation was, therefore, investigated, and tonsillar B cells activated with anti-IgM Abs. Because this activation leads to a high rate of apoptosis over 30 h (33, 34), hampering analysis of RAG expression, the investigation was limited to the first 30 h of stimulation. The frequency of CD5+ B lymphocytes increased until 30 h (Fig. 4A), consistent with the detection of RAG1, but not RAG2 mRNA, at that time (B). Inasmuch as CD40/CD40L interaction has also been shown to induce CD5 on B cell surface (35), we cultured tonsillar B lymphocytes on CD40L-transfected fibroblasts. The experiments confirmed that, following 3 days of stimulation, CD5 was induced on a subpopulation of cells for up to 6 days of culture (Fig. 4A), but without up-regulation of RAG1 and RAG2 mRNA transcription (B). These results indicate that specific conditions may be combined for CD5 to be induced and to permit V(D)J rearrangement. From the data reported with mouse spleen B cells (3), but in contrast to human GC B cells (13), RAG transcription may be up-regulated in mature B cells in vitro following costimulation. We thus looked for RAG transcripts in tonsillar B lymphocytes following costimulation with anti-IgM and CD40L, known to trigger survival of anti-IgM-induced apoptosis (36). As may be predicted from the previous experiments, CD5 expression was induced increasingly, for up to 5 days of culture (Fig. 4A). Interestingly, at day 4, this costimulation system extended the induction of both RAG1 and RAG2 mRNA transcripts (Fig. 4B).
FIGURE 4. Regulated expression of RAG1, RAG2, and CD5. Tonsillar B cells were isolated and cultured with PMA at 30 ng/ml or with anti-IgM-coated beads at 1 μg/ml for 30 h, or on CD40L-transfected fibroblasts alone or together with 1 μg/ml anti-IgM-coated beads for 6 days. A, Frequencies of CD5+ B cells were determined by flow cytometry. Mean ± SD of three experiments. B, RAG1 and RAG2 mRNA expression of cultured B cells was analyzed by nested RT-PCR, and that of GAPDH by RT-PCR. A representative data for each culture condition is shown.
To more definitively elucidate the association between CD5 expression and RAG transcription, CD5-nonexpressing Bm1 and Bm2 lymphocytes were sorted and stimulated with anti-IgM plus CD40L. A CD5+ subpopulation was generated anew, and, at days 4 and 5, RAG1 and RAG2 mRNA were coexpressed (Fig. 5A). These data demonstrate that RAG1 and RAG2 can be concomitantly up-regulated with the induction of CD5 expression. To identify those cells expressing RAG, we repeated the experiments by sorting CD5+ and CD5– B cells after 5 days of culture, and analyzed RAG transcripts in both populations. As shown in Fig. 5B, cells turned positive for CD5 expressed both RAG1 and RAG2 transcripts, whereas cells that remained CD5 negative expressed RAG1 only. These data demonstrate that, when activated, only CD5-induced IgD-positive cells were capable of rearrange their BCR-encoding genes.
FIGURE 5. Induced expression of RAG1 and RAG2 in CD5+ acquired Bm2 cells. A, IgD+CD38–/+CD5– tonsil B cells were sorted and cultured on CD40L-transfected fibroblasts with anti-IgM-coated beads at 1 μg/ml for up to 5 days. Frequency of CD5+ induced B cells was evaluated by flow cytometry, and RAG1 and RAG2 mRNA expression was analyzed by nested RT-PCR and that of GAPDH by RT-PCR. B, IgD+CD38–/+CD5– tonsil B cells were sorted and cultured as above for 5 days. Then, cells turned positive for CD5 and cells still negative were sorted again, and expression of RAG1, RAG2, and GAPDH mRNA was analyzed as above.
Discussion
The origin and functions of the CD5+ B cells have fueled extensive studies and discussions. Two mutually exclusive hypotheses have been proposed as to the origin of CD5+ B cells. The first is the lineage hypothesis, which contends that some B cell precursors are destined to become CD5+ B cells (reviewed in Ref. 37). In contrast, the differentiation hypothesis implies that all B cells have the potential to acquire CD5+ B lymphocyte characteristics (reviewed in Ref. 38). A new paradigm has recently emerged that reconciles both hypotheses, proposing that two subsets of CD5 B cells may exist in mice (39) as well as in humans (17). One subset of B cells would constitutively express CD5, whereas the other may be induced to do so in response to specific activation. The present work extends the knowledge of the latter subpopulation by demonstrating that, under specific stimulation conditions, some activated tonsillar B cells may be induced to express CD5 and, at the same time, gain the capacity to revise their BCR.
Based on IgD and CD38 expression, we have identified different subpopulations in human tonsils and established that RAG expression was not restricted to GC cells. Our study revealed that activated Bm2 cells, pre-GC Bm2' cells, and early Bm5 memory cells all displayed RAG1 and RAG2 mRNA. However, when subdivided according to the expression of CD5, the Bm2 CD5+ cells was identified as a subpopulation that subsists outside GC and has the potential to undergo secondary Ig gene recombination. Indeed, the results have shown that, for this B cell subset, RAG1 and RAG2 transcripts and proteins are expressed at the single-cell level. Frequencies of double-positive cells in the different B cell subsets are in the same range as those determined by others (16). Cells negative for both enzymes or positive for only one were also found. This might be attributable to the different means of RAG1 and RAG2 regulation. Although RAG1 expression mainly depends upon the cell activation status (40), that of RAG2 is linked to the cell cycle phase leading to its accumulation during the G1 phase (41). Because isolated B cells were not synchronized, all of the patterns of RAG1 and RAG2 expressions could be observed during the kinetics study. Furthermore, we have confirmed their functional recombinase activity. Thus, freshly isolated CD5+ Bm2 cells had J7, J4, and/or J5 signal breaks. Because detection of L chain recombination intermediates as well as increased gene usage have been correlated to receptor revision (13), our results indicate that CD5+ Bm2 cells have a functional V(D)J recombination machinery. From these combined data, we conclude that, before the cells enter the GC in vivo, naive activated Bm2 lymphocytes may initiate V(D)J rearrangement when they express the CD5 molecule. This conclusion is strengthened by our in vitro experiments. They showed that, when stimulated with anti-IgM plus CD40L, some IgD+CD38+CD5– Bm2 cells were encouraged to express RAG1 and RAG2, but this behavior was restricted to those turned into CD5 positive.
It has been shown that a subset of transitional B cells retains the capacity to up-regulate RAG genes when stimulated via their BCR in an appropriate environment (42). Although transitional B cells have not been underscored in humans, and do not express CD5 in mice (37), it was of importance to prove that the presence of RAG+ B cells in human tonsils could not be ascribed to newly immigrated immature-type B lymphocytes. According to the description by Loder et al. (30), we found transitional type-2-like B cells (IgD+IgMbrightCD21bright) in the IgD+ population, but their functional characteristics remain to be determined. However, we observed that these lymphocytes do not express CD5 and that all the CD5+ cells belong to the mature population (IgD+IgM+CD21+). Furthermore, expression of IgM, CD23, and CD44, and lack of CD10, CD77, and CD27 expression strongly confirm the preswitch characteristics of the IgD+CD38+CD5+ B cells, i.e., before they enter the GC (10). Taken together, these data demonstrate the mature status of the CD5+ B cells expressing RAG1 and RAG2 genes and thus undergoing V(D)J recombination outside GCs. However, it should be noted that, in one case, the CD5– Bm2 lymphocytes also expressed RAG1 and RAG2 mRNA. Because CD5 is transiently induced, it is likely that these latter cells may be CD5-induced Bm2 lymphocytes that have achieved V(D)J rearrangement, and for which CD5 expression has been terminated, but not that of RAG1 and RAG2. Moreover, the presence of RAG1 and RAG2 transcripts in both CD5+ and CD5– early Bm5 cells indicates that some postswitch B lymphocytes may revise their BCR independently of CD5 expression. However, the recombinase activity of the RAG protein in these cells remains to be demonstrated.
In contrast to activated naive Bm2 lymphocytes, resting naive Bm1 cells were negative for RAG1 and RAG2 mRNA expression. This indicates that V(D)J recombination might be induced by activation. We found that stimulation with PMA, known to be a potent cell activator and inducer of CD5 expression (31, 32), reduces RAG expression. This is consistent with previous work (43) and supports the notion that specific activation, such as Ag encounter, might be required for RAG up-regulation, as recently suggested (16). However, from our in vitro studies, BCR engagement alone cannot promote RAG mRNA up-regulation. In RAG1–/– mice reconstituted with RAG1-GFP splenic B cells, in vitro stimulation with either IL-1 and anti-IgM or IL-7 with anti-CD40 failed to induce GFP signal, suggesting that other mechanisms might contribute to receptor revision (15). A second signal delivered by T lymphocytes seems to be necessary. Thus, BCR and CD40 coligation triggers expression of RAG1 and RAG2. Interestingly, this was induced only in the naive mature B cells turned positive for CD5, indicating that CD5+ B cells expressing RAG mRNA are likely to belong to the inducible CD5 B cell subpopulation, although the involvement of the "constitutive" CD5+ B cell pool (17) cannot be totally discounted. Whether RAG+ B1 cells found in the peritoneal cavity of mice are part of the "inducible" CD5+ B cells remains also to be determined (22). It is now clear that induction of CD5 expression on human mature B cells represents part of the differentiation pathway and is a particular feature of certain B cells that are susceptible to special inductive signals (44). The decreased frequency of CD5+ B cells observed once the cells entered the GC lends support to this hypothesis, indicating that CD5 expression might be lost at this stage of development. However, we cannot rule out the possibility that a selection may also operate resulting in a diminished expression of CD5.
The CD5 expression, associated with RAG1 and RAG2 up-regulation, raised the question as to how appearance of CD5 may influence RAG1 and RAG2 expression. Other questions are the nature of the B cell subpopulation affected and the physiological role this process plays in vivo. It has been shown that receptor editing in immature B cells is induced because of distinct BCR signaling capacities in these cells compared with mature B cells. In immature B cells, engagement of the BCR leads to small increases in intracellular free Ca2+ and to the induction of receptor editing. In contrast, BCR engagement in mature B cells results in high elevation in intracellular Ca2+ concentration leading to the up-regulation of CD86, CD69, and MHC class II, but not V(D)J recombination (45). CD5+ B lymphocytes in mice have been shown to be hyporesponsive to BCR ligation compared with their CD5– B cell counterpart. They have a reduced capacity for intracellular Ca2+ mobilization, diminished proliferation, and increased apoptosis after BCR cross-linking due to a negative role for CD5 in the BCR signaling (23, 46). CD5 reduces strong BCR-mediated signaling and permits only limited increases in intracellular free Ca2+ (23). Similarly, CD5 expression may increase BCR threshold in human Bm2 lymphocytes. In this setting, such activated cells would respond like immature B cells with respect to Ca2+ mobilization. Low intracellular Ca2+ mobilization would thus promote the induction of RAG expression, leading ultimately to receptor revision. Therefore, the physiological relevance of this process may be as follows: whereas clones of B cells with high BCR affinity would be permitted to migrate into the dark zone and proliferate as centroblasts, B cells with very low BCR affinity would be deprived of rescuing signals, and thus deleted by apoptosis. We hypothesize that B cell clones with intermediate BCR affinity receive signals not strong enough to induce maturation into centroblast cells but sufficient enough to avoid deletion and trigger CD5 expression. Together with CD40 engagement, this will promote V(D)J rearrangement. When high-affinity revised BCR is expressed, BCR engagement would trigger maturation of the cells into centroblasts and in the process turn off RAG expression (13), and possibly CD5, which is known to be transiently induced under T-dependent activation (47). The relevance of this process may be to raise the affinity of the BCR up to the level required for maturation and differentiation. Our data suggest that BCR revision is not confined to a specific subpopulation but rather may involve a broad spectrum of B cells. Any B cell may express CD5, depending on its BCR affinity and/or specificity and, thereby, acquire the potential to revise its BCR. This is consistent with the hypothesis that CD5+ B cells may generate CD5– B cells and contribute to the heterogeneity of cell populations that proliferate in the GC (48).
This phenomenon may also have pathophysiological consequences. Modification of the BCR in anergic CD5+ B cells in autoimmune transgenic HEL/anti-HEL model has been observed (20). BCR engagement through low-affinity auto-Ags generates weak mobilization and reduced proliferation (49), which are ameliorated when the mice are bred onto a CD5–/– background. These CD5+ B cells manifest similarities with anergic cells. For example, both populations are stimulated by Ag (50, 51), respond suboptimally to IgM cross-linking (49, 52), show evidence of secondary receptor recombination (22, 53), and are excluded from the GCs (reviewed in Ref. 54). An obvious relationship links anergy and V(D)J rearrangement in maintaining B cell tolerance, as well as anergy-like state and CD5+ B cell development (20). Interestingly, we observed that the CD5+ Bm2 cells lack CD95 expression and display lower levels of IgM compared with the CD5– counterpart (our unpublished observations). These cells, which are protected from T-dependent apoptosis, might be in an anergic-like state.
CD5+ B cells constitute a major source of natural auto-Abs, although they require exposure to Ag for maturation into Ab-secreting cells (51). Furthermore, anergy-inducing BCR engagement is often associated with receptor editing (53, 55), suggesting that a hierarchy of decreasing signaling thresholds is required in BCR/Ag interactions for the induction of editing, anergy, and selection of CD5+ or CD5– B cells (56). Therefore, the accumulation of CD5+ B cells in human autoimmune diseases (57, 58) may reflect Ag-driven activation and subsequent anergy-like state induction of autoreactive B cells. From observations described herein, mechanisms contributing to the maintenance of tolerance in anergic CD5+ B cells may include the control of V(D)J recombination at the CD5+ Bm2 state. The control of CD5 expression and/or receptor revision before the cells enter the GC may be faulty in autoimmune disorders and leads to the emergence of auto-Ab-secreting cells.
Acknowledgments
We acknowledge Laetitia Le Pottier for her skillful and cheerful assistance. The editorial assistance of Prof. Rizgar A. Mageed is greatly appreciated Thanks are also due to Simone Forest and Cindy Séné for secretarial assistance.
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.
1 This work was supported by grants from Ministère de l’Enseignement Supérieur et de la Recherche and from Académie Nationale Fran?aise de Médecine.
2 Address correspondence and reprint requests to Prof. Pierre Youinou, Laboratory of Immunology, Brest University Medical School Hospital, BP 824, F 29609 Brest, France. E-mail address: youinou@univ-brest.fr
3 Abbreviations used in this paper: GC, germinal center; Bm, mature B; FMZ, follicular mantle zone; HEL, hen egg lysosyme; biot, biotinylated; LM-PCR, ligation-mediated PCR; ECD, energy-coupled dye.
Received for publication August 25, 2004. Accepted for publication February 16, 2005.
References
Hertz, M., D. Nemazee. 1997. BCR ligation induces receptor editing in IgM+IgD– bone marrow B cells in vitro. Immunity 6:429.
Radic, M. Z., M. Zouali. 1996. Receptor editing, immune diversification, and self-tolerance. Immunity 5:505.
Hikida, M., M. Mori, T. Takai, K. Tomochika, K. Hamatani, H. Ohmori. 1996. Reexpression of RAG-1 and RAG-2 genes in activated mature mouse B cells. Science 274:2092.
Han, S., B. Zheng, D. G. Schatz, E. Spanopoulou, G. Kelsoe. 1996. Neoteny in lymphocytes: Rag1 and Rag2 expression in germinal center B cells. Science 274:2094
Papavasiliou, F., R. Casellas, H. Suh, X. F. Qin, E. Besmer, R. Pelanda, D. Nemazee, K. Rajewsky, M. C. Nussenzweig. 1997. V(D)J recombination in mature B cells: a mechanism for altering antibody responses. Science 278:298.
Kelsoe, G.. 1999. V(D)J hypermutation and receptor revision: coloring outside the lines. Curr. Opin. Immunol. 11:70.
Nemazee, D., M. Weigert. 2000. Revising B cell receptors. J. Exp. Med. 191:1813.
de Wildt, R. M., R. M. Hoet, W. J. van Venrooij, I. M. Tomlinson, G. Winter. 1999. Analysis of heavy and light chain pairings indicates that receptor editing shapes the human antibody repertoire. J. Mol. Biol. 285:895.
Wilson, P. C., K. Wilson, Y. J. Liu, J. Banchereau, V. Pascual, J. D. Capra. 2000. Receptor revision of immunoglobulin heavy chain variable region genes in normal human B lymphocytes. J. Exp. Med. 191:1881.
Liu, Y. J., C. Arpin. 1997. Germinal center development. Immunol. Rev. 156:111.
Lindhout, E., G. Koopman, S. T. Pals, C. de Groot. 1997. Triple check for antigen specificity of B cells during germinal centre reactions. Immunol. Today 18:573.
Giachino, C., E. Padovan, A. Lanzavecchia. 1998. Re-expression of RAG-1 and RAG-2 genes and evidence for secondary rearrangements in human germinal center B lymphocytes. Eur. J. Immunol. 28:3506.
Meffre, E., F. Papavasiliou, P. Cohen, O. de Bouteiller, D. Bell, H. Karasuyama, C. Schiff, J. Banchereau, Y. J. Liu, M. C. Nussenzweig. 1998. Antigen receptor engagement turns off the V(D)J recombination machinery in human tonsil B cells. J. Exp. Med. 188:765.
Hertz, M., V. Kouskoff, T. Nakamura, D. Nemazee. 1998. V(D)J recombinase induction in splenic B lymphocytes is inhibited by antigen-receptor signalling. Nature 394:292.
Igarashi, H., N. Kuwata, K. Kiyota, K. Sumita, T. Suda, S. Ono, S. R. Bauer, N. Sakaguchi. 2001. Localization of recombination activating gene 1/green fluorescent protein (RAG1/GFP) expression in secondary lymphoid organs after immunization with T-dependent antigens in rag1/gfp knock-in mice. Blood 97:2680.
Girschick, H. J., A. C. Grammer, T. Nanki, M. Mayo, P. E. Lipsky. 2001. RAG1 and RAG2 expression by B cell subsets from human tonsil and peripheral blood. J. Immunol. 166:377.
Youinou, P., C. Jamin, P. M. Lydyard. 1999. CD5 expression in human B-cell populations. Immunol. Today 20:312.
Dono, M., V. L. Burgio, C. Tacchetti, A. Favre, A. Augliera, S. Zupo, G. Taborelli, N. Chiorazzi, C. E. Grossi, M. Ferrarini. 1996.
Subepithelial B cells in the human palatine tonsil. I. Morphologic, cytochemical and phenotypic characterization. Eur. J. Immunol. 26:2035.
Abe, M., K. Tominaga, H. Wakasa. 1994. Phenotypic characterization of human B-lymphocyte subpopulations, particularly human CD5+ B-lymphocyte subpopulation within the mantle zones of secondary follicles. Leukemia 8:1039.
Hippen, K. L., L. E. Tze, T. W. Behrens. 2000. CD5 maintains tolerance in anergic B cells. J. Exp. Med. 191:883.
Tze, L. E., E. A. Baness, K. L. Hippen, T. W. Behrens. 2000. Ig light chain receptor editing in anergic B cells. J. Immunol. 165:6796.
Qin, X. F., S. Schwers, W. Yu, F. Papavasiliou, H. Suh, A. Nussenzweig, K. Rajewsky, M. C. Nussenzweig. 1999. Secondary V(D)J recombination in B-1 cells. Nature 397:355.(Sophie Hillion, Alain Sar)