The present study was performed to analyze whether marginal zone B (MZ-B) cells in nondeliberately immunized adult rats are selected on basis of the specificity of their B cell receptor, and to determine to what extent memory B cells contribute to the MZ-B cell subset. To this end, the Ig PC7183 VH gene repertoire was studied among VHDJH-μ transcripts expressed in four sequential stages of B cell development, of two individual untreated adult rats. B cell subsets, i.e., pro/pre-B cells and newly formed B (NF-B) cells from bone marrow, and recirculating follicular B cells and MZ-B cells from spleen were sorted by flow cytometry. In addition, from one these rats, cells were microdissected from follicular and MZ areas of the spleen and productive PC7183 VH gene rearrangements were analyzed for the presence of somatic mutations. Sequence analysis reveals that most MZ-B cells in the adult rat, either defined by flow cytometry or by their anatomical location in the spleen, express germline encoded VH genes (naive MZ-B cells) and a minor fraction (about 20%) of the MZ-B cells carry somatic mutations (memory MZ-B cells). In addition, we show that naive MZ-B cells are a selected population of cells, both based on PC7183 VH gene repertoire and on the length of the Ig heavy (H) chain complementarity-determining region 3 (H-CDR3) region, i.e., PC7183 VHDJH-μ transcripts of MZ-B cells carry significantly shorter H-CDR3 regions than other B cell subsets.

Marginal zone B (MZ-B)2 cells are a particular B cell subset that differs in many respects from the predominating subset of mature recirculating follicular B (RF-B) cells, which are small surface (s)IgMlowsIgDhigh B cells. In contrast to RF-B cells, MZ-B cells are intermediate-sized sIgMhighsIgDlow cells, which do not recirculate and are exclusively found in the spleen, where they are located in a distinct area surrounding the B cell follicles and the periarteriolar lymphoid sheath (1, 2, 3). Increasing (but still circumstantial) evidence indicates that MZ-B cells are involved in T cell-independent type 2 (TI-2) responses to (lipo-)polysaccharide Ags. These Ags form the major constituent of cell walls of encapsulated bacteria like Streptococcus pneumoniae, Neisseria meningitidis, and Haemophilus influenzae (2, 3). The phenotype of MZ-B cells (pyroninophilic cytoplasm, less condensed nuclear chromatin, CD21high, CD23low, CD35high, and higher basal levels of CD80 and CD86) (4, 5, 6, 7) suggests that MZ-B cells are in a somewhat activated state that might result from antigenic experience. Experiments in mice, using ex vivo stimulation of purified B cell subsets, revealed that MZ-B cells respond generally more rapidly to mitogens like LPS, dextran-conjugated anti-IgM or IgD Abs, or CD40 ligation than RF-B cells (7, 8, 9). Taken together, the (activated) phenotype of MZ-B cells in combination with their unique anatomical location in the spleen may reflect their immunological function (i.e., to respond quickly to bloodborn infections with encapsulated bacteria).

MZ-B cells may be of mixed origin. Part of the MZ-B cells are presumably naïve (i.e., nonmemory) cells that develop in the absence of T cells and stimulation with exogenous Ags. This is suggested by the presence of MZ-B cells in spleen of Ag-free (10), germfree (11, 12), nude (13, 14), TCR−/− (7), or MHC class II-deficient animals (15). We and others showed that (naive) MZ-B cells in rat can develop directly from mature RF-B cells, a maturation step that does not require cell proliferation (2, 16). Kinetic studies imply that only a small proportion of RF-B cells eventually becomes MZ-B cell (17). B cells that become incorporated into the pool of MZ-B cells are most likely selected on the basis of their B cell receptor (BCR) specificity. First, as mentioned before, MZ-B cells are thought to be involved in Ab responses against TI-2 Ags and therefore carry BCRs specific for these types of Ags. Second, Chen et al. (18) demonstrated that, in VH81X Ig heavy (H) chain transgenic mice, the splenic MZ is selectively populated by a (polyreactive) B cell clone. Recently, Martin and Kearney (19) showed that the (positive) selection of this B cell clone into the MZ-B cell compartment is presumably due to its BCR specificity.

In addition to the presence of these so-called naïve MZ-B cells, also Ag-experienced memory cells may contribute to the MZ-B cell pool. Hapten-binding (memory) B cells reside in the splenic MZ of previously immunized rats (2, 20) and most MZ-B cells taken from human spleen carry somatically mutated Ig VH genes (21, 22). Moreover, virtually all MZ-B cells in man express CD27 (23), which is selectively expressed by somatically mutated memory B cells, but not by naive B cells (24). Memory MZ-B cells are most likely derived from germinal centers (25, 26).

In this study, we directly aimed at establishing the relative contribution of naive vs memory type of MZ-B cells and to explore whether (naive) MZ-B cells are indeed a selected population of cells. For this reason, we analyzed the relative representation and nucleotide sequences of VH genes belonging to the PC7183 VH gene family in several stages of B cell development in rats. Like in mice and man, Ig H chains and light (L) chains in rats are produced after sequential rearrangement of VH, D, and JH region genes, and VL and JL region genes, respectively (27). We recently showed that VH genes in rat can be subdivided into VH gene families similar to those of mice (17, 27, 28). At present, sequence analysis has revealed the existence of at least 10 VH gene families in rat (PC7183, X24, Vh11, S107, J606, Vh10, Q52, 3609, J558, and VGAM3.8) (17, 27, 28). Here, we analyzed the relative representation of VH genes belonging to the PC7183 VH gene family in four flow cytometry-defined stages of B cell development. In individual rats, the PC7183 VH gene repertoire expressed by MZ-B cells was compared with the repertoire expressed by mature splenic RF-B cells and that of the subsets of pro/pre-B cells and newly formed B cells (NF-B) from bone marrow. In addition, by analyzing the PC7183 VH genes for the presence of somatic mutations, we also determined to what extent memory B cells contribute to the MZ-B cell subset. Because representation of memory type of MZ-B cells may differ between flow cytometry-defined MZ-B cells and MZ-B cells defined on basis of their anatomical location in the spleen, we also analyzed productive PC7183 VH genes microdissected from follicular and MZ areas of rat spleen. Together, our data show that, in nonintentionally immunized rats, most MZ-B cells (defined either by flow cytometry or histology) express germline encoded PC7183 VH genes and possibly represent naive cells. Only a minor fraction of MZ-B cells appear to be memory B cells, as revealed by the presence of somatic mutations in their VH genes. The PC7183 VH gene repertoire of MZ-B cells differs from that of RF-B cells (and other B cells) and carry significantly shorter Ig H chain complementarity-determining region 3 (H-CDR3) regions, indicating that MZ-B cells are a selected population of cells.

Bone marrow cells and spleens were taken from two 6-mo-old male PVG rats (R1 and R2, respectively), which were raised under clean conventional conditions. The rats were obtained from and housed at the Central Animal Facility of the Faculty of Medical Sciences (University of Groningen, Groningen, The Netherlands). The preparation of cell suspensions used for flow cytometry and cell sorting has been described elsewhere (16). Bone marrow cells were collected from either femur. Bone marrow and spleen cell suspensions were stained by a combination of FITC-conjugated mouse anti-rat IgM (HIS40) (29), PE-conjugated mouse anti-rat CD90 (HIS51; anti-Thy-1) (30), and biotinylated mouse anti-rat IgD (MaRD3; a generous gift from Dr. H. Bazin, University of Louvain, Brussels, Belgium). Bone marrow cells were additionally incubated with biotinylated mouse anti-rat TCRαβ (R73; PharMingen, San Diego, CA) (31). Biotinylated mAbs were revealed by streptavidin-allophycocyanin (PharMingen). Cell analysis and sorting were performed on a dual laser Coulter Epics-Elite flow cytometer with enhanced system performance upgrade (Coulter, Hialeah, FL). Dead cell, plasma cell, monocyte/macrophage, and erythrocyte contamination was excluded from the sorting by using forward and side scatter profiles. Sorted cells were collected in sterile tubes containing ∼500 μl newborn calf serum (Life Technologies, Glasgow, Scotland). FlowJo software (version 2.7; Tree Star, San Carlos, CA) was used for flow cytometry data analysis.

Sorted cells were pelleted by 300 × g centrifugation for 10 min at 4°C. Equivalents of 2–4 × 105 cells were lysed in 0.5 ml TRIzol Reagent (Life Technologies) and stored at −80°C. Total RNA was isolated according to instructions of the manufacturer (Life Technologies). First-strand cDNA synthesis was performed on total RNA isolated from 2–4 × 105 sorted cells. Briefly, RNA was taken up in a 19-μl reaction volume containing 1,6 μg Oligo d(T)12–18 primer (Amersham-Pharmacia Biotech, Uppsala, Sweden) and 1.5 μl 10 mM each of dGTP, dATP, dTTP, and dCTP. This mixture was heated to 70°C for 10 min and quickly chilled on ice. First-strand cDNA reaction was performed in a final reaction volume of 30 μl, by adding 1 μl RNAquard (28 U/μl; Amersham-Pharmacia Biotech), 6 μl 5× first strand buffer (250 mM Tris-HCl (pH8.3), 375 mM KCl, 15 mM MgCl2), 3 μl 0.1 mM DTT, and 1 μl SuperScriptII reverse transcriptase (200 U/μl; Life Technologies). The cDNA reaction was conducted at 42°C for 50 min and finally inactivated by heating at 70°C for 15 min.

For construction of Ig VHDJH-μ region cDNA libraries, 2 μl of cDNA was amplified by PCR in a 50-μl reaction mixture containing 0.6 μM VH PC7183 family-specific primer (VH PC7183; for primer sequences see Table I), 0.6 μM μ constant-region primer (Cμ1.3), 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 1.5 mM MgCl2, 2.5 U Taq DNA polymerase (Life Technologies), and 0.2 mM each of dGTP, dATP, dTTP, and dCTP. The PCR amplification program consisted of 35 cycles of 1 min at 94°C (2 min in first cycle), 1 min at 65°C, and 1 min at 72°C, followed by an incubation at 72°C for 10 min. The size of the PCR products was checked by agarose gel electrophoresis. PCR products were subsequently cloned into pCR2.1-TOPO vector by using the TOPO TA cloning kit (Invitrogen BV, Groningen, The Netherlands). Plasmid was obtained from single white colonies, randomly picked, and grown in overnight cultures (High Pure plasmid isolation kit; Boehringer Mannheim GmbH, Mannheim, Germany). Aliquots of the isolated plasmids were digested by EcoRI and subjected to agarose gel electrophoresis. Clones containing a ±500-bp insert were subsequently sequenced at our local sequence facility (Department of Pathology and Laboratory Medicine, Division of Medical Biology, University of Groningen, Groningen, The Netherlands). Sequence reactions were performed by cycle sequencing using Thermo Sequenase primer cycle sequencing kit 7-deaza dGTP or MegaBACE DYEnamic ET dye terminator kit (Amersham-Pharmacia Biotech). Samples were run on the ALFexpress (Amersham-Pharmacia Biotech) or MegaBASE-1000 system (Molecular Dynamics and Amersham-Pharmacia Biotech, Sunnyvale, CA).

Table I.

PCR primers used for amplification of rat rearranged Ig VHDJH regions

Primer NameNucleotide Sequence (5′→3′)Locationa
VH PC7183 CTT AGT GCA GCC TGG AAG GT VH IMGT-FR1 pos. 33–52 
Cμ1.3 CCC TGG ATG ACT TCA GTG TTG μ-chain exon 1 pos. 138–158 
RJH-A TCC CCT AGG CAG TTT AGT CC pos. 1617–1636 
RJH-B CGA ATC TTG GCT CCC ATT TG pos. 1010–1029 
RJH-C CCT CCC CGA CAA ATG CAG TA pos. 597–616 
RJH-D GCA GGA TGT GGT GTC CAG AT pos. 317–336 
RJH-1 GAA AGG TCT TAC CTG AGG ACA C pos. 235–256 
RJH-2 AAG GAC TTA CCT GAG GAG ACT G pos. 540–561 
RJH-3 GAC TCA CCT GAA GAG ACA GTG AC pos. 924–946 
RJH-4 GCC ATT CTT ACC TGA GGA GAC A pos. 1501–1522 
Primer NameNucleotide Sequence (5′→3′)Locationa
VH PC7183 CTT AGT GCA GCC TGG AAG GT VH IMGT-FR1 pos. 33–52 
Cμ1.3 CCC TGG ATG ACT TCA GTG TTG μ-chain exon 1 pos. 138–158 
RJH-A TCC CCT AGG CAG TTT AGT CC pos. 1617–1636 
RJH-B CGA ATC TTG GCT CCC ATT TG pos. 1010–1029 
RJH-C CCT CCC CGA CAA ATG CAG TA pos. 597–616 
RJH-D GCA GGA TGT GGT GTC CAG AT pos. 317–336 
RJH-1 GAA AGG TCT TAC CTG AGG ACA C pos. 235–256 
RJH-2 AAG GAC TTA CCT GAG GAG ACT G pos. 540–561 
RJH-3 GAC TCA CCT GAA GAG ACA GTG AC pos. 924–946 
RJH-4 GCC ATT CTT ACC TGA GGA GAC A pos. 1501–1522 
a

pos., Start and end nucleotide position of primer. For VH gene-specific primers, the position is given according to the IMGT nomenclature (32 ), where the first nucleotide of a codon = IMGT codon number × 3–2. The last nucleotide of IMGT-FR3 is 312. For Ig μ constant region, the position is mentioned relative to the first nucleotide of the first codon of μ exon 1 (GenBank accession numbers: X68782 and X78895). Positions of the RJH primer series is relative to the first nucleotide of the genomic DNA sequence of the rat JH gene locus published by Lang and Mocikat (33 ).

Cryostat sections of 7 μm were made from snap frozen spleen tissue (liquid freon at −80°C) of rat number 1 (R1), air dried and fixed in acetone for 10 min at room temperature. Sections were stained for 1 h with mouse anti-rat IgM mAb (HIS40). After rinsing in PBS, the slides were incubated with polyclonal rabbit anti-mouse Ig conjugated to HRP (DAKO, Glostrup, Denmark). Peroxidase activity was revealed by 3,3′-diaminobenzidine tetra-HCl (Sigma, St. Louis, MO), containing 0.01% (v/v) H2O2. Sections were counterstained with hematoxylin according to Mayer and stored overnight in PBS at 4°C. In two separate experiments, three clusters of about 50 cells were microdissected from 6 different primary B cell follicles as well as from the MZ areas surrounding these follicles. Thus in total, 18 follicular and 18 MZ samples were taken. Microdissection was performed with a hydraulic micromanipulator (MMO-202; Narishige, Tokyo, Japan), as described by Küppers et al. (34). Microdissected cells were taken in 20 μl of PCR buffer (20 mM Tris-HCl (pH 8.4) and 50 mM KCl) and stored at −20°C. Microdissected cells were digested by 250 ng proteinase K (Boehringer Mannheim) by incubating for 60 min at 50°C, followed by a proteinase K inactivation step at 95°C for 10 min (reactions were covered with mineral oil). Primers used for PCR amplification were all HPLC-purified, and stock solutions were stored in aliquots at −20°C. In the first round of PCR amplification, the samples were taken in a 50-μl reaction volume containing 0.2 mM primer VH PC7183, 0.05 mM each of RJH-A,-B,-C,-D primer, 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 2.5 mM MgCl2, 2.5 U Taq DNA polymerase (Life Technologies), and 0.2 mM each of dGTP, dATP, dTTP, and dCTP. Samples were heated to 95°C before adding 5 μl Taq DNA polymerase (diluted to 0.5 U/μl in deionized water) to the reaction. The first round PCR amplification program consisted of 35 cycles of 1.5 min at 95°C (2 min in the first cycle), 1 min at 65°C (4 min in first cycle), and 1.5 min at 72°C, followed by an incubation at 72°C for 10 min. In the second round, PCR amplification reactions were conducted with each of the four nested 3′ JH primers (RHJ-1, RHJ-2, RJH-3, and RJH-4) in combination with the VH PC7183 family-specific primer VH PC7183. Second round of PCR amplification was conducted in 50 μl containing 0.6 mM primer VH PC7183, 0.6 mM of either primer RJH-1, RJH-2, RJH-3, or RJH-4, 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 1.5 mM MgCl2, 2.5 U Taq DNA polymerase (Life Technologies), and 0.2 mM each of dGTP, dATP, dTTP, and dCTP. Hot start was performed by heating the reactions to 95°C before adding 5 μl Taq DNA polymerase (diluted to 0.5 U/μl in deionized water). Water overlaying the tissue section during the microdissection served as negative control for the PCR. Ten-microliter aliquots of the PCR samples were analyzed by agarose gel electrophoresis. Cloning, insert screening, and sequencing were conducted as described above.

Statistical analysis was performed using SPSS 9.0 for Windows (SPSS, Chicago, IL). Frequency distributions of nominal variables such as VH gene and JH gene usage, and numbers of mutations per VH gene, were compared using the likelihood ratio test for contingency tables. In many cells, the frequencies were below 5%. Therefore, we calculated the exact p values or, when this was not feasible, we estimated them by a Monte Carlo method (±99% confidence interval (CI)). Differences in PC7183 VH gene repertoire were evaluated by taking the most frequently used members (sum of these particular members over the B cell subsets compared (row count) > 5) and “all others” (row count ≤ 5) as the categories. The difference between mean H-CDR3 length and mean loss of JH nucleotides between the B cell subsets was statistically evaluated by a multiple comparison test (general linear model) based on Bonferroni statistics. H-CDR3 length and loss of JH nucleotides are normally distributed variables (Kolmogorov-Smirnov test). For comparison of the frequency distributions of H-CDR3 length and loss of JH nucleotides, the Mann-Whitney U nonparametric rank-order test was applied (p = Monte Carlo significance ± 99% CI). Differences between groups were considered significant when p < 0.05.

In Fig. 1, we show the immunoperoxidase staining pattern of normal rat spleen for IgM, IgD, and CD90 (Thy-1). B cells in MZs and follicles stain brightly for IgM, whereas (bright) IgD staining is restricted to B cells located in the follicles. Staining for CD90 (Thy-1), which is expressed in the B cell lineage by immature B cells (35), clearly demonstrates that virtually all cells in the MZ and most cells in the follicles are CD90-negative. Only in follicles, low numbers of CD90+ cells are present, thus most likely representing immature B cells (NF-B cells and/or early recirculating follicular B cells) (35).

FIGURE 1.

Serial sections of the spleen of a normal untreated 6-mo-old rat stained with mouse anti-rat mAbs directed to IgM (HIS40), IgD (MaRD3), and CD90 (HIS51, anti-Thy-1). MZ, Marginal zone; F, follicle; P, periarteriolar lymphocyte sheath; RP, red pulp.

FIGURE 1.

Serial sections of the spleen of a normal untreated 6-mo-old rat stained with mouse anti-rat mAbs directed to IgM (HIS40), IgD (MaRD3), and CD90 (HIS51, anti-Thy-1). MZ, Marginal zone; F, follicle; P, periarteriolar lymphocyte sheath; RP, red pulp.

Close modal

With this panel of Abs (anti-IgM, anti-IgD, and anti-CD90), MZ-B cells and RF-B cells can unequivocally be distinguished in spleen cell suspensions: MZ-B cells are sIgMhighsIgDlowCD90, whereas RF-B cells are sIgMlowsIgDhighCD90. This is in contrast to the two immature and recently bone marrow-derived B cell subsets, NF-B cells (sIgMhighsIgDlow/−CD90+) and early recirculating follicular B cells (sIgMlowsIgDhighCD90+). Pro-/pre-B cells in bone marrow lack sIgM and sIgD, but express high levels of CD90. Fig. 2 shows the flow cytometry profiles of these B cell subsets in bone marrow and spleen.

FIGURE 2.

Three-color flow cytometry procedure for sorting the subsets of RF-B cells and MZ-B cells from spleen (A) and the subsets of pro/pre-B cells and NF-B cells from bone marrow (B). Spleen and bone marrow cells were stained with mouse anti-rat mAbs directed to IgM (HIS40), IgD (MaRD3), and CD90 (HIS51, anti-Thy-1). A, Gate settings for sorting RF-B cells and MZ-B cells from spleen. Gates were set according to relative expression levels of sIgM and sIgD. From both subsets, the CD90 cells were sorted, revealing RF-B cells (sIgMlowsIgDhighCD90) and MZ-B cells (sIgMhighsIgDlowCD90). B, Gate settings for sorting pro/pre-B cells (CD90+sIgMsIgD) and NF-B cells (CD90+sIgMhighsIgDlow/−) from bone marrow. To exclude T cell contamination within the pro/pre-B cell sorts, bone marrow cells were also stained by mouse anti-rat TCRαβ (R73). Except for the reanalysis, all plots are gated for lymphoid cells using forward-sideward scatter profile. Reanalysis was performed on ±1000 sorted cells.

FIGURE 2.

Three-color flow cytometry procedure for sorting the subsets of RF-B cells and MZ-B cells from spleen (A) and the subsets of pro/pre-B cells and NF-B cells from bone marrow (B). Spleen and bone marrow cells were stained with mouse anti-rat mAbs directed to IgM (HIS40), IgD (MaRD3), and CD90 (HIS51, anti-Thy-1). A, Gate settings for sorting RF-B cells and MZ-B cells from spleen. Gates were set according to relative expression levels of sIgM and sIgD. From both subsets, the CD90 cells were sorted, revealing RF-B cells (sIgMlowsIgDhighCD90) and MZ-B cells (sIgMhighsIgDlowCD90). B, Gate settings for sorting pro/pre-B cells (CD90+sIgMsIgD) and NF-B cells (CD90+sIgMhighsIgDlow/−) from bone marrow. To exclude T cell contamination within the pro/pre-B cell sorts, bone marrow cells were also stained by mouse anti-rat TCRαβ (R73). Except for the reanalysis, all plots are gated for lymphoid cells using forward-sideward scatter profile. Reanalysis was performed on ±1000 sorted cells.

Close modal

In the first experiment, MZ-B and RF-B cells were isolated from spleen, and pro/pre-B cells and NF-B cells from bone marrow from two individual rats (R1 and R2) by three-color flow cytometry cell sorting using Abs to IgM, IgD, and CD90. Gates set for sorting and reanalysis of the sorted B cell subsets are shown in Fig. 2. All sorted B cell fractions were obtained with purities exceeding 98%.

VHDJH-μ transcripts expressed in sorted B cell fractions were amplified by RT-PCR using a PC7183 VH gene family-specific primer (VH PC7183) and a μ constant-region primer (Cμ1.3) (Table I). The VH PC7183 primer sequence was deduced from the PC7183 VH genes described previously (27). The VH PC7183 primer is located in Ig H chain framework 1, at a position which is highly conserved among members of the PC7183 VH gene family, but more diverged in VH gene sequences from members of other families. The position of the VH PC7183 primer was chosen in such a way that no degenerated sites were necessary. This is of advantage because competition between different VH gene primers is now prevented, and therefore no preferential amplification based on primer sequence constitution will occur during the PCR. From either rat, about 20 VHDJH region sequences were analyzed from each B cell subset, resulting in a total of 173 PC7183 VHDJH region sequences.

Previously, we obtained a number of germline PC7183 VH gene sequences from rat genomic liver DNA (27, 28). The large number of PC7183 VH gene sequences obtained in the present study offered the opportunity to identify more germline PC7183 VH genes of rat origin. New germline PC7183 VH gene sequences were established by comparing the VH gene sequences obtained from the present study with each other and with already established rat germline PC7183 VH gene sequences. Rearranged VH gene sequences were considered to be germline when two or more independently sampled, rearranged or genomic, VH gene sequence(s) were 100% identical. These rat VH PC7183 reference germline sequences are designated as “PC-n ” (e.g., PC-9, PC-15). In a few cases, we established germline consensus sequences, which were derived from sequences containing Taq DNA polymerase errors and/or sequence ambiguities. These nontrivial VH gene sequences are followed by a “w” after their designation (e.g., PC-10w). Currently, we have established 28 individual germline VH gene sequences of the PC7183 VH gene family in rat (sequences are available at the IMGT database: http://imgt.cnusc.fr:8104 (32)). We assume that the PC7183 VH gene family in rat consists of more than 28 members, because a few VH gene sequences obtained from genomic liver DNA could not be confirmed at least twice.

The VH gene of each sequenced VHDJH region was assigned to the most homologous germline PC7183 VH gene. Similarly, the JH genes of the rearrangements could be established after comparison to published genomic JH gene sequences of the rat (33). In contrast to VH and JH genes, very little sequence information is available on rat Ig H-chain diversity (D) genes; so far only one genomic D gene has been determined (DQ52) in this species (36). However, probably more D genes exist in rat, because many of the VH-D-JH junction sequences (H-CDR3) obtained in the present study contain identical stretches of nucleotides (data not shown). These identical stretches likely represent various D gene segments. In a previous study, we sequenced several D to JH rearrangements in which the 5′-terminal-end of the D gene was flanked by D gene specific recombination recognition sequences (nonamer, 12 bp spacer, heptamer) (28). Also these data revealed the existence of different D region-specific sequences. Knowledge of genomic D gene sequences is necessary to identify the number of N nucleotide additions between VH-D and D-JH boundaries. Because to date DQ52 is the only complete D segment known for rat, we could only establish N nucleotide additions for VH-D-JH junctions that carry the DQ52 gene.

In Fig. 3, we show the VHDJH-μ transcripts expressed in the flow cytometry-defined subsets of bone marrow pro/pre-B cells and NF-B cells, and splenic RF-B cells, and MZ-B cells. The vast majority (>90%) of the VH-D-JH junction sequences from R1 and all VH-D-JH junction sequences from R2 represent unique rearrangements, as indicated by their unique H-CDR3 region. In R1, four identical VHDJH region sequences were recovered from the MZ-B cell subset (MZ1801 series, clone 1, 16, 18, and 20). These clonally related sequences were most probably sampled from separate MZ-B cells, because two of these sequences carry more mutations in the VH region than that can be explained by Taq DNA polymerase (see further). Alignment of these four clonally related sequences reveals that most mutations occurred independently from each other (Fig. 4). Only clones MZ1801-1 and MZ1801-16 share one identical nucleotide exchange (G → C mutation in codon 57; Fig. 4 A), which is indicative for their genealogical relationship. Thus, these four VHDJH region sequences are most likely derived from distinct B cells that belong to a clonally expanded B cell. We speculate that this clonal expansion took place within a germinal center. Also, Tierens et al. (22) observed clonally related B cells in the MZ of the human spleen. Two pairs of identical VHDJH region sequences were recovered from the R1 RF-B cell subset. However, these sequences display no evidence for somatic mutations. Therefore, we cannot determine whether these two pairs were sampled from one cell, or represent independent, identical rearrangements of two different cells. Furthermore, because we found no evidence for the existence of dominant B cell clones within the RF-B cell subset (denaturing gradient gel electrophoresis analysis of Ig H-CDR3 regions; data not shown), we used only one of the two sequences for further (statistical) analysis.

FIGURE 3.

Nucleotide sequences of PC7183 VH-D-JH junctions of Ig H chain μ transcripts expressed in the flow cytometry-defined subsets of bone marrow pro/pre-B cells and NF-B cells, and splenic RF-B cells and MZ-B cells. For each sequence the most homologous germline PC7183 VH gene and JH gene is given. Sequences span the region from the most 3′-terminal-end of the VH gene (TGT = conserved Cys on position 92) to the conserved JH TGG (Trp on position 103). Codon numbering and nomenclature is according to Kabat et al. (37 ). Nucleotides encoding the H-CDR3 region are boxed. In some rearrangements, codon 95 (start of H-CDR3) and 96 may be encoded by the most 3′ VH nucleotides. Because the 3′ terminal-end of germline PC7183 VH genes have not been resolved at the genomic level, the 3′-terminal-end of the each PC7183 VH member is determined by sequence alignment. Regions encoded by the DQ52 gene segment are marked bold. N nucleotide additions in VH-DQ52-JH junctions are shown in lowercase characters. Lowercase and underscored characters represent P nucleotides (38 ), and are shown for DQ52 and JH segments only. ∗, Nonproductive rearrangement. 1, 4-bp deletion in VH gene of PR0804-18 (nucleotide position 196) and NF1401-14 (nucleotide position 53) (IMGT nomenclature (32 )). 2, VHDJH region sequence identical with RF1501-14. 3, VHDJH region sequence identical with RF1501-1. 4, Clonal relationship with MZ1801-1. Full length VHDJH region sequences are available at the EMBL database (European Bioinformatics Institute (EBI), Cambridge, U.K.) under accession numbers AJ286144-AJ286315.

FIGURE 3.

Nucleotide sequences of PC7183 VH-D-JH junctions of Ig H chain μ transcripts expressed in the flow cytometry-defined subsets of bone marrow pro/pre-B cells and NF-B cells, and splenic RF-B cells and MZ-B cells. For each sequence the most homologous germline PC7183 VH gene and JH gene is given. Sequences span the region from the most 3′-terminal-end of the VH gene (TGT = conserved Cys on position 92) to the conserved JH TGG (Trp on position 103). Codon numbering and nomenclature is according to Kabat et al. (37 ). Nucleotides encoding the H-CDR3 region are boxed. In some rearrangements, codon 95 (start of H-CDR3) and 96 may be encoded by the most 3′ VH nucleotides. Because the 3′ terminal-end of germline PC7183 VH genes have not been resolved at the genomic level, the 3′-terminal-end of the each PC7183 VH member is determined by sequence alignment. Regions encoded by the DQ52 gene segment are marked bold. N nucleotide additions in VH-DQ52-JH junctions are shown in lowercase characters. Lowercase and underscored characters represent P nucleotides (38 ), and are shown for DQ52 and JH segments only. ∗, Nonproductive rearrangement. 1, 4-bp deletion in VH gene of PR0804-18 (nucleotide position 196) and NF1401-14 (nucleotide position 53) (IMGT nomenclature (32 )). 2, VHDJH region sequence identical with RF1501-14. 3, VHDJH region sequence identical with RF1501-1. 4, Clonal relationship with MZ1801-1. Full length VHDJH region sequences are available at the EMBL database (European Bioinformatics Institute (EBI), Cambridge, U.K.) under accession numbers AJ286144-AJ286315.

Close modal
FIGURE 4.

Alignment of four clonally related VHDJH sequences recovered from the MZ-B cell subset of R1 (A). Sequences of the germline PC7183 VH gene (PC-25) and JH2 gene are shown on top. Codon numbering and nucleotide subdivisions are made according to IMGT nomenclature (32 ). Dashes indicate identical nucleotides and gaps resulting from the IMGT subdivision are marked by dots. Replacement mutations are shown in bold/underscored characters. (B) Genealogical tree deduced from the mutations observed in the VH genes of these four clones. For each clone the replacement over silent mutation (R/S) ratio is depicted and mentioned for both H-FR and H-CDR regions of the genes (R/S ratio is shown in respect to the immediate ancestral gene).

FIGURE 4.

Alignment of four clonally related VHDJH sequences recovered from the MZ-B cell subset of R1 (A). Sequences of the germline PC7183 VH gene (PC-25) and JH2 gene are shown on top. Codon numbering and nucleotide subdivisions are made according to IMGT nomenclature (32 ). Dashes indicate identical nucleotides and gaps resulting from the IMGT subdivision are marked by dots. Replacement mutations are shown in bold/underscored characters. (B) Genealogical tree deduced from the mutations observed in the VH genes of these four clones. For each clone the replacement over silent mutation (R/S) ratio is depicted and mentioned for both H-FR and H-CDR regions of the genes (R/S ratio is shown in respect to the immediate ancestral gene).

Close modal

Table II summarizes the characteristics of the productive VH-DJH-μ transcripts expressed in the various B cell subsets. The Taq DNA polymerase mutation rate was about 0.1%, as established from the ±100-bp sequence of the μ constant-region (exon 1) adjacent to the 3′ end of every VHDJH region. Based on this Taq error rate, VH genes were considered to be unmutated when no more than two mutations were observed upon comparison with the germline gene (chance of more than two mutations caused by Taq in a 239 bp VH region by an error rate of 0.1% is <6% (=(239 bp × 0.001)2). By using this mutation criterion, we found that the VH genes expressed by all pro/pre-B cells and almost all NF-B cells and RF-B cells are unmutated (Fig. 5). Only two VH genes recovered from the NF-B cell fraction of R1 carry three to four mutations per gene. We cannot, however, completely exclude the possibility that this low number of mutations is still the result of Taq errors, especially because the number of sequences with this mutation rate is rather low. One VH gene obtained from the RF-B cell subset of R1 reveals eight mutations when compared with the most homologous PC7183 germline gene.

Table II.

Characteristics of productively rearranged PC7183-encoded Ig VHDJH regions expressed in flow cytometry-defined B cell subsets, or isolated from follicular and MZ areas of the spleen, of adult conventionally raised rats

SourceaRatnbVH Gene Mutation AnalysiscH-CDR3 Length (±SD)gLoss of JH Nucleotides (±SD)h
Mutated ClonesdMutations/Gene (±SD)eMutation Frequency (±SD)f
Flow cytometry        
Pro/pre-B R1 20 0.3 ± 0.6 0.1 ± 0.2 11.5 ± 2.2 5.7 ± 4.7 
 R2 20 0.2 ± 0.4 0.1 ± 0.2 12.1 ± 3.0 5.6 ± 3.0 
 Total 40 0.2 ± 0.5 0.1 ± 0.2 11.8 ± 2.6 5.6 ± 3.9 
NF-B R1 21 0.8 ± 1.2 0.3 ± 0.5 10.6 ± 3.1 5.1 ± 3.2 
 R2 23 0.5 ± 0.6 0.2 ± 0.3 11.7 ± 2.5 3.6 ± 3.0 
 Total 44 0.6 ± 0.9 0.3 ± 0.4 11.2 ± 2.8 4.3 ± 3.2 
RF-B R1 19 0.7 ± 1.9 0.3 ± 0.8 11.3 ± 2.3 5.1 ± 3.3 
 R2 24 0.5 ± 0.6 0.2 ± 0.3 10.0 ± 2.9 5.5 ± 4.2 
 Total 43 0.6 ± 1.3 0.2 ± 0.5 10.6 ± 2.7 5.3 ± 3.8 
MZ-B R1 20 2.0 ± 3.0 0.8 ± 1.2 9.2 ± 3.2 5.7 ± 3.9 
 R2 20 1.3 ± 2.5 0.5 ± 1.0 8.8 ± 2.4 5.0 ± 3.5 
 Total 40 1.6 ± 2.7 0.7 ± 1.1 9.0 ± 2.8i 5.3 ± 3.7 
Microdissection        
FO R1 13 1.1 ± 1.0 0.5 ± 0.4 9.7 ± 2.9 4.5 ± 4.4 
MZ R1 10 1.0 ± 2.5 0.4 ± 1.1 9.0 ± 3.4 5.7 ± 3.7 
SourceaRatnbVH Gene Mutation AnalysiscH-CDR3 Length (±SD)gLoss of JH Nucleotides (±SD)h
Mutated ClonesdMutations/Gene (±SD)eMutation Frequency (±SD)f
Flow cytometry        
Pro/pre-B R1 20 0.3 ± 0.6 0.1 ± 0.2 11.5 ± 2.2 5.7 ± 4.7 
 R2 20 0.2 ± 0.4 0.1 ± 0.2 12.1 ± 3.0 5.6 ± 3.0 
 Total 40 0.2 ± 0.5 0.1 ± 0.2 11.8 ± 2.6 5.6 ± 3.9 
NF-B R1 21 0.8 ± 1.2 0.3 ± 0.5 10.6 ± 3.1 5.1 ± 3.2 
 R2 23 0.5 ± 0.6 0.2 ± 0.3 11.7 ± 2.5 3.6 ± 3.0 
 Total 44 0.6 ± 0.9 0.3 ± 0.4 11.2 ± 2.8 4.3 ± 3.2 
RF-B R1 19 0.7 ± 1.9 0.3 ± 0.8 11.3 ± 2.3 5.1 ± 3.3 
 R2 24 0.5 ± 0.6 0.2 ± 0.3 10.0 ± 2.9 5.5 ± 4.2 
 Total 43 0.6 ± 1.3 0.2 ± 0.5 10.6 ± 2.7 5.3 ± 3.8 
MZ-B R1 20 2.0 ± 3.0 0.8 ± 1.2 9.2 ± 3.2 5.7 ± 3.9 
 R2 20 1.3 ± 2.5 0.5 ± 1.0 8.8 ± 2.4 5.0 ± 3.5 
 Total 40 1.6 ± 2.7 0.7 ± 1.1 9.0 ± 2.8i 5.3 ± 3.7 
Microdissection        
FO R1 13 1.1 ± 1.0 0.5 ± 0.4 9.7 ± 2.9 4.5 ± 4.4 
MZ R1 10 1.0 ± 2.5 0.4 ± 1.1 9.0 ± 3.4 5.7 ± 3.7 
a

B cell subsets were determined by three-color flow cytometry using anti-IgM, anti-IgD, and anti-CD90 staining. FO, Follicle.

b

Number of productive VHDJH region sequences.

c

Revealed from nucleotide position 52 (codon 18) up to and including 312 (codon 104) according to IMGT Ig VH gene nomenclature (32 ).

d

VH gene sequences with >2 mutations (>3 mutations for FO and MZ sequences). Number of mutated VH genes within the MZ-B cell subset is significantly different from pro/pre-B, NF-B, and RF-B cells (pexact (2-sided): p = 0.005, p = 0.042, and p = 0.013, respectively). Number of mutated VH genes between pro/pre-B, NF-B, and RF-B is not significantly different (pexact (2-sided) = 0.773).

e

Mean number of mutations per VH gene.

f

Mean mutation frequency (percent). Mutation frequency for a sequence = (number of mutations/number of nucleotides overlap) × 100%. Taq DNA polymerase error rate not subtracted from the values.

g

Mean H-CDR3 length in codons, calculated from codon 95 up to and including codon 102 according to the nomenclature of Kabat et al. (37 ) for compatibility to mouse.

h

Mean number of nucleotides deleted at the 5′ terminal end of the JH gene. No statistically significant difference between B cell subsets (Bonferroni, α = 0.05).

i

, Mean H-CDR3 length in VHDJH-μ transcripts expressed by MZ-B cells is significantly different from the mean H-CDR3 length expressed in the other B cell subsets (Bonferroni, α = 0.05). No significant difference in mean H-CDR3 length is observed between pro/pre-B, NF-B, and RF-B (Bonferroni, α = 0.05).

FIGURE 5.

Frequency distribution of the number of mutations in Ig PC7183 VH genes expressed in the flow cytometry-defined subsets of bone marrow pro/pre-B cells (n = 40) and NF-B cells (n = 44), and splenic RF-B cells (n = 43) and MZ-B cells (n = 40). The graph is produced using data from R1 and R2. Sequences are grouped according to the number of mutations per VH gene and for each group the frequency of sequences within the four B cell subsets is shown. Number of mutations is revealed from nucleotide position 52 (codon 18) up to including 312 (codon 104) according to IMGT nomenclature (32 ). Due to Taq DNA polymerase errors, the zero to two mutations per VH gene group can be considered as unmutated (see text).

FIGURE 5.

Frequency distribution of the number of mutations in Ig PC7183 VH genes expressed in the flow cytometry-defined subsets of bone marrow pro/pre-B cells (n = 40) and NF-B cells (n = 44), and splenic RF-B cells (n = 43) and MZ-B cells (n = 40). The graph is produced using data from R1 and R2. Sequences are grouped according to the number of mutations per VH gene and for each group the frequency of sequences within the four B cell subsets is shown. Number of mutations is revealed from nucleotide position 52 (codon 18) up to including 312 (codon 104) according to IMGT nomenclature (32 ). Due to Taq DNA polymerase errors, the zero to two mutations per VH gene group can be considered as unmutated (see text).

Close modal

There was no statistical difference in number of mutated VH between the pro/pre-B, NF-B, and RF-B cell subsets; the mutation frequency calculated for each of these B cell subsets is almost similar to the mutation frequency expected from Taq (Table II). In contrast, the number of mutated VH genes expressed in MZ-B cells is significantly higher compared to other B cell subsets (Table II). About 20% of the VH genes recovered from MZ-B cells of either rat carry mutations, ranging from 3 to 11 nucleotide exchanges per gene (Fig. 5). The higher number of mutated VH genes expressed in this subset results in a mutation frequency which is more than three times higher than calculated for the other B cell subsets (Table II). Two of eight mutated MZ-B cell VH gene sequences have only three nucleotide exchanges with the germline, and thus could have been possibly caused by Taq; the other 6 VH genes carry 5 to 11 mutations per gene, and are therefore very likely the result of somatic mutations.

An Ag-driven response results in Abs with higher affinity. This is generally reflected by a higher proportion of amino acid exchanges in the CDR regions (Ag-binding sites) than in the FR regions of the VH genes. For this reason, we calculated the replacement over silent mutation (R/S) ratio of CDR and FR regions of VH genes carrying more than four mutations. The observed R/S mutation ratios were compared with the R/S mutation ratios theoretically expected from a random distribution of mutations and the chance (p) was calculated that the number (either an excess or scarcity) of replacement mutations in the CDR and H-FR regions was just the result of coincidence (39). Table III shows the observed and expected R/S mutation ratios of seven VH genes (one from the RF-B cell subset and six from the MZ-B cell subset) carrying more than four mutations. In all these VH genes, mutations in the H-CDR regions resulted in amino acid replacements. In five of seven cases, the replacement mutations in H-CDR and H-FR regions are randomly distributed (p > 0.05). However, in VH genes MZ1801-17 and MZ1801-18, we observed an excess of replacement mutations in the H-CDR regions (p = 0.027 and p = 0.00015, respectively) and a scarcity of replacement mutations in the H-FR regions (p = 0.045 and p = 0.012, respectively) which cannot solely be explained by coincidence. Therefore, the pattern of mutations observed in these two MZ-B cell derived VH gene sequences resembles that of VH genes encoding high affinity mAbs and provide evidence that these MZ-B cells have undergone Ag-driven selection.

Table III.

Observed and expected R/S mutation ratios of mutated VH genes expressed in the flow cytometry-defined B cell subsets and of a mutated VH gene recovered from the MZ area of the spleen by microdissection

CloneaSourcebRatGermlineMutationscMutation Frequencyc,dR/S Mutation Ratioe
Observed (R/S)Expectedp
Flow cytometry         
RF1501–19 RF-B R1 PC-23 3.6 FR 0.67 (2/3) 3.08 0.045 
      CDR ∞ (3/0) 3.55 0.11 
MZ1801–13 MZ-B R1 PC-35 10 4.5 FR 2.00 (6/3) 3.07 0.25 
      CDR ∞ (1/0) 3.79 0.32 
MZ1801–17 MZ-B R1 PC-1 3.6 FR 1.00 (2/2) 3.08 0.045 
      CDR ∞ (4/0) 3.48 0.027 
MZ1801–18 MZ-B R1 PC-25 2.25 FR 0 (0/0) 2.92 0.012 
      CDR ∞ (5/0) 3.86 0.00015 
MZ2803–5 MZ-B R2 PC-5 3.6 FR ∞ (6/0) 2.95 0.19 
      CDR ∞ (2/0) 4.48 0.27 
MZ2803–9 MZ-B R2 PC-28 3.2 FR 2.00 (4/2) 2.95 0.29 
      CDR ∞ (1/0) 3.55 0.39 
MZ2803–16 MZ-B R2 PC-15 2.3 FR 1.00 (2/2) 3.01 0.24 
      CDR ∞ (1/0) 3.63 0.40 
Microdissection         
M11–4 MZ R1 PC-5 3.6 FR 2.00 (2/1) 2.95 0.049 
      CDR ∞ (5/0) 4.48 0.0054 
CloneaSourcebRatGermlineMutationscMutation Frequencyc,dR/S Mutation Ratioe
Observed (R/S)Expectedp
Flow cytometry         
RF1501–19 RF-B R1 PC-23 3.6 FR 0.67 (2/3) 3.08 0.045 
      CDR ∞ (3/0) 3.55 0.11 
MZ1801–13 MZ-B R1 PC-35 10 4.5 FR 2.00 (6/3) 3.07 0.25 
      CDR ∞ (1/0) 3.79 0.32 
MZ1801–17 MZ-B R1 PC-1 3.6 FR 1.00 (2/2) 3.08 0.045 
      CDR ∞ (4/0) 3.48 0.027 
MZ1801–18 MZ-B R1 PC-25 2.25 FR 0 (0/0) 2.92 0.012 
      CDR ∞ (5/0) 3.86 0.00015 
MZ2803–5 MZ-B R2 PC-5 3.6 FR ∞ (6/0) 2.95 0.19 
      CDR ∞ (2/0) 4.48 0.27 
MZ2803–9 MZ-B R2 PC-28 3.2 FR 2.00 (4/2) 2.95 0.29 
      CDR ∞ (1/0) 3.55 0.39 
MZ2803–16 MZ-B R2 PC-15 2.3 FR 1.00 (2/2) 3.01 0.24 
      CDR ∞ (1/0) 3.63 0.40 
Microdissection         
M11–4 MZ R1 PC-5 3.6 FR 2.00 (2/1) 2.95 0.049 
      CDR ∞ (5/0) 4.48 0.0054 
a

VH gene sequences carrying more than four mutations.

b

See legend to Table II.

c

Revealed from nucleotide position 55 (codon 19) up to and including 297 (codon 99) according to IMGT Ig VH gene nomenclature (32 ).

d

Mutation frequency (percent), for calculation see Table II.

e

Observed R/S mutation ratio is the quotient of observed R and observed S mutations in H-FR and H-CDR regions, respectively. The actual number of R and S mutations is given in parentheses. The theoretical expected (inherent) R/S mutation ratio is the quotient of total possible R and total possible S mutations in the germline gene, and is calculated according to Chang and Casali (39 ). The possibility (p) that an excess or scarcity of replacement mutations in H-CDR or H-FR regions results solely from chance is negated by the significantly low probability values (p < 0.05) calculated according to the binomial distribution (39 ).

In man, nearly all MZ-B cells are somatically mutated, as was revealed by analysis of microdissected MZ B cells (21, 22). This is in marked contrast to our aforementioned findings in flow cytometry-defined MZ-B cells in the rat. To exclude the possibility that these differences can be explained by the technique used to collect MZ-B cells (microdissection vs flow cytometry), we also analyzed productively rearranged PC7183 VH genes from cells microdissected from follicular and MZ B cell areas of the spleen of R1. An example of a frozen section of the spleen after microdissecting cells from the follicular and MZ B cell areas is shown in Fig. 6. Cells were scraped from six different follicles and from the MZs surrounding these follicles. From these areas, we analyzed 13 and 10 productive PC7183 VHDJH sequences, respectively. All VHDJH regions carry unique VH-D-JH junctions and are thus derived from individual B cells (Fig. 7). In addition to the productive VHDJH region sequences, two sequences isolated from follicular areas and four sequences from the MZ areas were nonproductive due to a frameshift and/or a stop codon in the VH-D-JH junctional area (Fig. 7). These nonproductive VHDJH region sequences were excluded from further analysis. The characteristics of the productive sequences from these microdissected cells are shown in Table II. Except for one sequence, all follicular PC7183 VH gene sequences contain less than two mutations when compared with their most homologous germline gene. One follicular VH gene (F32-3) carries three mutations. Given the higher Taq error rate (doubled number of PCR cycles), we assume that these three mutations are most likely Taq errors (chance of three mutations in 239-bp VH region with Taq error rate of 0.2% = (239 × 0.002)3 = 0.11)) rather than somatic mutations. Among the productive VHDJH region sequences recovered from the microdissected MZ B cell areas, only 1 of 10 VH gene sequences carries more mutations than could be expected from Taq (M11-4, eight mutations upon comparison to germline PC-5). Because of the excess of R mutations in H-CDR regions (and their scarcity in the H-FR regions) in the VH gene of M11-4 (Table III), we conclude that this sequence is derived from an Ag-selected and somatically mutated MZ-B cell. The frequency of mutated PC7183 VH gene sequences from microdissected MZ B cells thus closely resembles that of flow cytometry-sorted MZ-B cells, and confirms that the difference in frequency of somatically mutated MZ-B cells observed between man and rat, are not enforced by the technique used to isolate the cells (microdissection vs flow cytometry).

FIGURE 6.

Example of a spleen section showing the location of the cells microdissected from follicles and MZ. Each microdissected area consisted of about 50 cells. Section was stained for IgM (HIS40).

FIGURE 6.

Example of a spleen section showing the location of the cells microdissected from follicles and MZ. Each microdissected area consisted of about 50 cells. Section was stained for IgM (HIS40).

Close modal
FIGURE 7.

Genomic DNA sequences of PC7183 VH-D-JH junctions of Ig H-chain variable regions recovered from B cells microdissected from follicular (n = 13) and MZ (n = 10) areas of the spleen of R1. Gaps are marked by dots. Full length VHDJH region sequences are available at the EMBL database under accession numbers AJ391285-AJ391313. See legend of Fig. 3 for further explanation.

FIGURE 7.

Genomic DNA sequences of PC7183 VH-D-JH junctions of Ig H-chain variable regions recovered from B cells microdissected from follicular (n = 13) and MZ (n = 10) areas of the spleen of R1. Gaps are marked by dots. Full length VHDJH region sequences are available at the EMBL database under accession numbers AJ391285-AJ391313. See legend of Fig. 3 for further explanation.

Close modal

To establish whether the VH gene repertoire expressed by MZ-B cells is the result of positive or negative selection of certain Ig specificity’s, we determined the PC7183 VH gene usage in MZ-B cells and their precursors (pro/pre-B cells, NF-B cells, and RF-B cells). To this end, we counted the frequency of individual PC7183 VH family members productively expressed in each B cell subset in either rat (Fig. 8). In all four B cell subsets, generally the same PC7183 VH members are used. In addition, we observed no substantial difference in PC7183 VH gene usage between the two rats. However, the PC7183 VH gene repertoire within the various B cell subsets is not randomly expressed and certain members are more frequently used than others. For example, PC7183 member PC-4 is relatively frequently used by pro/pre-B cells and NF-B cells in bone marrow and splenic RF-B cells (12.5%, 20.5%, and 9.3%, respectively), whereas this gene is not observed among MZ-B cells. PC-15 is also frequently expressed by pro/pre-B cells and NF-B cells (20% and 11.4%, respectively) and less frequent by RF-B cells and MZ-B cells (2.3% and 7.5%, respectively). Finally, VHDJH-μ transcripts encoded by PC7183 VH gene member PC-1 are predominantly present in the RF-B cell subset (20.9%), whereas this particular gene is used by 5.0% of both pro/pre-B cells as well as MZ-B cells, and by 9.1% of the NF-B cells. Memory (MZ)-B cells are (by definition) Ag-selected cells. To reveal whether the repertoire expressed by naïve MZ-B cells is also selected, we statistically evaluated the usage of unmutated PC7183 VH genes among the various B cells subsets. A statistically significant difference (p = 0.040 ± 0.005) in PC7183 VH gene repertoire is found between the B cell subsets when both rats are taken together. When individual rats are analyzed, a difference in PC7183 VH gene usage is observed in R1 (p = 0.040 ± 0.005) and a nearly significant difference in R2 (p = 0.102 ± 0.007). Given the linear developmental relationship between the four B cell subsets (pro/pre-B cells → NF-B cells → RF-B cells → MZ-B cells), we further analyzed the repertoires between two subsequent B cell stages. The difference in PC7183 VH gene usage is most prominent between RF-B cells and MZ-B cells (R1 + R2, pRF-B vs MZ-B = 0.020 ± 0.004). Our data also indicate that the PC7183 VH gene repertoire expressed by NF-B cells and RF-B cells might be different, although confidence is no more than 85% (R1 + R2, pNF-B vs RF-B = 0.146 ± 0.009). The PC7183 VH gene repertoire expressed in pro/pre-B cells does not differ from that of NF-B cells (R1 + R2, ppro/pre-B vs NF-B = 0.655 ± 0.012). Together, the data indicate that during B cell development in rats, selection not only occurs between the stages of NF-B cells in the bone marrow and RF-B cells in the periphery, but also between RF-B cells and naïve MZ-B cells in spleen. In a previous report, studying the kinetics of B cell subsets, we already proposed that significant selection of cells might take place between NF-B cells and RF-B cells (35, 40).

FIGURE 8.

VH gene utilization of individual members of Ig PC7183 VH gene family in VHDJH-μ transcripts expressed in the flow cytometry-defined subsets of bone marrow pro/pre-B cells and NF-B cells, and splenic RF-B cells and MZ-B cells. Each individual sequence is marked by a dot and located near the most homologous PC7183 germline gene (PC-n). The source of the sequence (R1 or R2) is given by the number in the dot (❶ = unmutated VH gene from R1 (0–2 mutations/gene); ① = mutated VH gene from R1). The PC7183 germline genes are arranged in descending order of homology with PC-1 arbitrarily put on top. The subgroups (I, II, and III) are formed by PC7183 members with more than 90% sequence identity. A significant difference is revealed between RF-B and MZ-B (p = 0.020 ± 0.004) and a near significant difference between NF-B and RF-B (p = 0.146 ± 0.009), using the most frequently used members (sum of particular member over B cell subsets (row count) >5) and “all others” (row count ≤5) as categories. There is no significant difference in PC7183 VH gene usage between pro/pre-B and NF-B (p = 0.655 ± 0.012).

FIGURE 8.

VH gene utilization of individual members of Ig PC7183 VH gene family in VHDJH-μ transcripts expressed in the flow cytometry-defined subsets of bone marrow pro/pre-B cells and NF-B cells, and splenic RF-B cells and MZ-B cells. Each individual sequence is marked by a dot and located near the most homologous PC7183 germline gene (PC-n). The source of the sequence (R1 or R2) is given by the number in the dot (❶ = unmutated VH gene from R1 (0–2 mutations/gene); ① = mutated VH gene from R1). The PC7183 germline genes are arranged in descending order of homology with PC-1 arbitrarily put on top. The subgroups (I, II, and III) are formed by PC7183 members with more than 90% sequence identity. A significant difference is revealed between RF-B and MZ-B (p = 0.020 ± 0.004) and a near significant difference between NF-B and RF-B (p = 0.146 ± 0.009), using the most frequently used members (sum of particular member over B cell subsets (row count) >5) and “all others” (row count ≤5) as categories. There is no significant difference in PC7183 VH gene usage between pro/pre-B and NF-B (p = 0.655 ± 0.012).

Close modal

The frequencies of the four JH genes used in productively rearranged VHDJH-μ transcripts of bone marrow pro/pre-B cells and NF-B cells, and splenic RF-B cells and MZ-B cells is shown in Fig. 9,A. JH2 and JH3 are the most predominantly used JH gene segments. In all B cell subsets (of the two rats together), JH2 and JH3 together encode for more than 70% of the JH region in the VHDJH-μ transcripts. In pro/pre-B cells, this percentage is even 90%. Most likely, the predominance of these two JH genes reflects the preference of the recombination machinery for JH-RSS (recombination signal sequence; nonamer-spacer-heptamer) with a 23-bp spacer sequence over JH-RSS carrying a 22-bp spacer (41). In rats, both JH2 and JH3 have a RSS with a 23-bp spacer sequence, whereas JH1 and JH4 carry a RSS with a 22-bp spacer (28, 33). Interestingly, in both rats, VHDJH-μ transcripts expressed by MZ-B cells are most frequently encoded by the JH2 gene (R1, 60% and R2: 80%), which is at least 1.5 times higher compared with other B cell subsets (Fig. 9 A). The p value for differences in JH gene usage between the four B cell subsets is 0.068 ± 0.06.

FIGURE 9.

A, JH gene usage in VHDJH-μ transcripts from flow cytometry-defined pro/pre-B cells, NF-B cells, RF-B cells, and MZ-B cells. The difference in JH gene usage between the four B cell subsets is nearly significant (R1 + R2, p = 0.068 ± 0.06). No significant difference is observed between pro/pre-B vs NF-B (p = 0.306 ± 0.012), NF-B vs RF-B (p = 0.355 ± 0.012), and RF-B vs MZ-B (p = 0.108 ± 0.008). B, Frequency distributions of H-CDR3 codon lengths and JH nucleotide loss in productive VHDJH-μ transcripts isolated form pro/pre-B cells, NF-B cells, RF-B cells, and MZ-B cells. H-CDR3 length was measured from position 95 (VH Cys = position 92) to the TGG (Trp at position 103) of the JH, corresponding to the H-CDR3 region according to Kabat et al. (37 ). There is no significant difference in H-CDR3 length distributions among pro/pre-B, NF-B, and RF-B as computed by the Mann-Whitney U test (R1 plus R2: p = 0.244 ± 0.011). However, there is significant difference between MZ-B vs RF-B (p = 0.009 ± 0.003), MZ-B vs NF-B (p = 0.001 ± 0.000), and MZ-B vs pro/pre-B (p < 0.001). Loss of JH nucleotides in VHDJH sequences was determined by counting the number of 5′ JH nucleotides that were not present in the rearrangement, starting from the nucleotide immediate 3′ of the respective JH-RSS heptamer sequence as indicated by Lang and Mocikat (33 ). No significant difference is revealed in the distributions of JH nucleotide loss between the B cell subsets (R1 plus R2, p = 0.431 ± 0.013).

FIGURE 9.

A, JH gene usage in VHDJH-μ transcripts from flow cytometry-defined pro/pre-B cells, NF-B cells, RF-B cells, and MZ-B cells. The difference in JH gene usage between the four B cell subsets is nearly significant (R1 + R2, p = 0.068 ± 0.06). No significant difference is observed between pro/pre-B vs NF-B (p = 0.306 ± 0.012), NF-B vs RF-B (p = 0.355 ± 0.012), and RF-B vs MZ-B (p = 0.108 ± 0.008). B, Frequency distributions of H-CDR3 codon lengths and JH nucleotide loss in productive VHDJH-μ transcripts isolated form pro/pre-B cells, NF-B cells, RF-B cells, and MZ-B cells. H-CDR3 length was measured from position 95 (VH Cys = position 92) to the TGG (Trp at position 103) of the JH, corresponding to the H-CDR3 region according to Kabat et al. (37 ). There is no significant difference in H-CDR3 length distributions among pro/pre-B, NF-B, and RF-B as computed by the Mann-Whitney U test (R1 plus R2: p = 0.244 ± 0.011). However, there is significant difference between MZ-B vs RF-B (p = 0.009 ± 0.003), MZ-B vs NF-B (p = 0.001 ± 0.000), and MZ-B vs pro/pre-B (p < 0.001). Loss of JH nucleotides in VHDJH sequences was determined by counting the number of 5′ JH nucleotides that were not present in the rearrangement, starting from the nucleotide immediate 3′ of the respective JH-RSS heptamer sequence as indicated by Lang and Mocikat (33 ). No significant difference is revealed in the distributions of JH nucleotide loss between the B cell subsets (R1 plus R2, p = 0.431 ± 0.013).

Close modal

The H-CDR3 region is the most diverse region of an Ig molecule, both in length and amino acid constitution. Amino acids encoded by the H-CDR3 play an important role in binding of the BCR with Ag. The length of the H-CDR3 region becomes fixed during VH-D-JH rearrangement and is determined by the length of the D segment used, the number of N nucleotide additions (by the enzyme TdT), and the number of nucleotides removed at the site of recombination by exonuclease activity. As shown in Table II, the mean H-CDR3 length and the distribution of the number of H-CDR3 codons does not differ between pro/pre-B cells, NF-B cells, and RF-B cells. However, the mean H-CDR3 length of MZ-B cells is 9.0 ± 2.8 codons and significantly shorter than that observed in pro/pre-B cells (11.8 ± 2.6 codons), NF-B cells (11.2 ± 2.8 codons), and RF-B cells (10.6 ± 2.7 codons) (Bonferonni, α = 0.05). Also the distribution of the number of H-CDR3 codons expressed by MZ-B cells differs from the other B cell subsets (Fig. 9,B). Similar to flow cytometry-defined MZ-B cells, also the VHDJH regions of MZ-B cells microdissected from the MZ area of the spleen carry short H-CDR3 regions (9.0 ± 3.4 codons; Table II). The mean H-CDR3 length of B cells microdissected from the follicular areas (9.7 ± 2.9 codons; Table II) is, however, only slightly higher compared with MZ-B cells. This might be explained by the fact that follicles do not consist exclusively of RF-B cells, but probably also contain MZ-B like cells (sIgMhigh, HIS24low, HIS57high) ( Ref. 16 , and P.M.D. and F.G.M.K., unpublished observation).

The shorter H-CDR3 regions observed in MZ-B cells are not due to total lack of N nucleotide additions or excess in exonuclease activity at VH-D-JH junctions in these cells. We show that MZ-B carrying VH-DQ52-JH rearrangements do contain N nucleotide additions (Fig. 3). Furthermore, many homopolymer tracts of dG and dC residues are observed at the VH-D and D-JH junctions, which are typical for N nucleotide additions (42, 43). The frequency of nucleotide deletions caused by exonuclease activity at the VH-D-JH junctions was measured by the number of nucleotides deleted at the 5′-terminal-end of the JH segments. No difference is observed in the distribution of 5′-JH nucleotide loss (Fig. 9,B) and no significant difference is revealed between the average number of deleted nucleotides at this site (Table II), between all the investigated B cell subsets. Consequently, the use of shorter H-CDR3 regions in VHDJH-μ transcripts of MZ-B cells, compared with pro/pre-B cells, NF-B cells, and RF-B cells, is most likely the result of a selection event rather than a fundamental difference in recombination machinery.

In this study, we provide evidence that splenic MZ-B cells in the rat are a heterogeneous population of cells. In normal untreated—nonintentionally immunized—rats, the majority (around 80%) of PC7183 VH genes expressed by MZ-B cells, either defined by flow cytometry or by microdissection, are encoded by germline VH genes and thus represent naive B cells. A minor fraction (20%) of the MZ-B cell pool, however, is constituted by B cells that express somatically mutated VH genes, whereas several of these mutated sequences (three of seven; see Table III) exhibit signs of Ag-selection. Somatic mutations of VH genes in conjunction with Ag-selection appear to be confined to germinal centers (44). For this reason, we think that these somatically mutated MZ-B cells are germinal center-derived memory B cells. The presence of mutated MZ-B cells in our study confirms previous observations that a proportion of B cells in the splenic MZ might be germinal center-derived cells (2, 20, 25, 26). We like to emphasize here that, in our experiments, the memory MZ-B cells, that were collected by flow cytometry, are not (yet) isotype switched, because 1) they were sorted on the basis of sIgM and sIgD expression and 2) were amplified using a μ constant-region specific primer. As shown recently, the frequency of memory MZ-B cells in mice also seems to be rather low. Makowska et al. (15) reported that in C57BL/6 mice nearly all J558 VH expressing CD1high splenic B cells (mainly consisting of MZ-B cells) were unmutated (zero to two mutations per VH gene). These observations in rats and mice are in striking contrast to the situation in human spleen where the majority (more than 85%) of splenic MZ- B cells are somatically mutated (21, 22). The difference in the proportion of memory MZ-B cells between man and rodents is probably not explained by differences in the technique used to isolate MZ-B cells. We did not observe a difference in frequency of mutated VH genes in MZ-B cells either collected by microdissection or sorted by flow cytometry. In the studies performed by Dunn-Walters et al. (21) and Tierens et al. (22), MZ-B cells were exclusively collected from human spleen by microdissection. However, one should realize that the microanatomical structure of the human spleen differs from that of rats (and mice) with respect to organization of the MZ and the presence of a so-called perifollicular zone surrounding the MZ (45). The splenic MZ in man is composed of an inner and an outer area and one cannot rule out the possibility that these areas differ in composition of memory and naive MZ-B cells. In the aforementioned studies, MZ-B cells could have been scraped from only selected (inner or outer) areas of the MZ. Another, probably more likely, explanation for the lower frequency of memory B cells among MZ-B cells in rodents, is the lower antigenic exposure of these animals, caused by the fact that they are housed in controlled laboratory animal facilities, and the much shorter life history of rodents in comparison to adult man.

Although most MZ-B cells in adult rat spleen are naive B cells, our results support the hypothesis that MZ-B cells are selected on basis of the specificity of their BCR. Evidence for this follows from our observations that, in comparison to RF-B cells, NF-B cells and pro/pre-B cells 1) naive rat MZ-B cells express a different PC7183 VH gene family repertoire, 2) they tend to use more frequently JH2 genes, and 3) excitingly, MZ-B cells carry significantly shorter H-CDR3 regions. In rats, MZ-B cells most likely originate from the pool of RF-B cells (16). However, we cannot exclude the possibility that MZ-B cells can also develop directly from other B cell subsets. For example, Martin et al. (19) have shown in mice that MZ-B cells are recruited from (recirculating) IgMhighCD21int B cells rather than from the bulk of long-lived recirculating IgMlowCD21int B cells. Little is known about the nature of the ligands involved in the recruitment of naive MZ-B cells. MZ-B cells are thought to be involved in TI-2 Ab responses against (lipo-)polysaccharide Ags from encapsulated microorganisms (2, 3). Evidence for this is, however, largely circumstantial and remains to be established at the single cell level. On the other hand, recently, Kearney and coworkers showed that polyreactive B cells in Ig H chain transgenic mice, i.e., expressing a BCR that reacts to phosphorylcholine (PC) in addition to numerous of self-Ags, are positively selected into the MZ-B cell pool, whereas B cells highly specific for hen egg lysozyme (HEL) are not (18, 19). Possibly, MZ-B cells are primarily polyreactive B cells with Ig that bind with low affinities to a broad range of Ags, including bacterial-derived (TI-2) Ags as well as self-Ags. This notion suggests that self-Ags probably play an important role in establishing the B cell repertoire of the MZ-B cell subset, and thus also explains why in animals deprived of exogenous Ags (i.e., germfree and Ag free animals) MZ-B cells still develop (10, 11, 12, 15).

The hypothesis that MZ-B cells express primarily polyreactive Ig is consistent with another intriguing finding of the present study, i.e., that MZ-B cells express PC7183 VHDJH-μ transcripts with significantly shorter mean H-CDR3 length in comparison to other B cell subsets, such as RF-B cells in spleen, and NF-B cells and pro/pre-B cells in bone marrow. The mean number of amino acids in the H-CDR3 region of MZ-B cells (either defined by flow cytometry or microdissection) is 9 amino acids, which is 2–3 amino acids shorter than found among other B cell subsets. Several groups have reported the importance of the H-CDR3 region in determining the polyreactive behavior of Abs (46, 47, 48, 49, 50). By gene-shuffling experiments using combinations of mono- and polyreactive mAbs that are encoded by highly similar Ig H and L chain V genes, Ichiyoshi and Casali (48) and Crouzier et al. (47) showed that mAbs lose polyreactivity when their H-CDR3 region is replaced by the corresponding region of their monoreactive counterpart, but not when the polyreactive mAbs were grafted with either the H chain framework 1-framework 3 (VH) region or with the Vκ L chain region of that monoreactive counterpart. Although the H-CDR3 length of polyreactive Abs is variable, there are several examples showing that polyreactive mAbs carry relatively short H-CDR3 regions (46, 51). Furthermore, from the data published by Tornberg and Holmberg (52) we calculated that PC7183 VHDJH-μ transcripts expressed by peritoneal B-1a and B-1b cells in mice also have relatively short H-CDR3 lengths (mean ± SD: 8.6 ± 3.4 and 8.7 ± 2.9, respectively) in comparison to conventional B-2 cells from spleen (11.3 ± 2.6). Similar to MZ-B cells, B-1 cells are preferentially directed against bacterial coat Ags, and produce poly- and self-reactive Abs (53, 54). Schroeder et al. (55) reasoned that relatively short H-CDR3 regions would probably generate “flat” Ab-binding sites, which would allow more interactions possible between residues of the H-CDR3 and potential Ags, resulting in polyreactivity. We propose that, at least for PC7183 VH gene encoded Ig H chains, polyreactive Ig is frequently encoded by short H-CDR3 regions.

Studies reported over the past years have provided substantial evidence that BCR signaling is a prerequisite for proper development and maintenance of B cells. BCR signaling does not only play a role in early B cell differentiation (56) but does also determine the fate of B cells in later (more mature) stages of their development (19, 57, 58, 59, 60). The signals delivered by the BCR are subjected to positive and negative feedback regulation by accessory receptors like CD45, CD19, and CD22, and their secondary signal transducing molecules (61, 62). In this context, it is relevant to mention that Mason et al. (63) demonstrated in HEL/anti-HEL IgM/IgD double transgenic mice that anergic B cells cannot be recruited into the MZ-B cell pool, but persist in the lymphoid follicles. Possibly, B cells may become MZ-B cells when their BCR delivers a signal that in strength exceeds a certain threshold. By lowering the strength of BCR signaling (by diminishing the positive feedback of the BCR) this threshold is increased, resulting in reduced numbers or complete absence of MZ-B cells, as is the case in Xid (Bruton’s tyrosine kinase mutant) mice and in CD19- deficient mice, respectively (15, 19). We like to speculate here that selection of MZ-B cells is presumably based upon the degree of polyreactivity of their BCR, because B cells expressing a polyreactive BCR may reach the threshold levels required for MZ-B cell maturation more easily due to binding to multiple (self-)Ags.

In summary, data shown here clearly demonstrate that in rats the majority of cells populating the MZ area of the spleen are naïve B cells, which are most probably recruited from recirculating B cells by means of positive selection. Self-Ags may somehow be involved in the selection process of MZ-B cells. This process presumably favors the selection of polyreactive B cells, resulting in MZs populated with MZ-B cells expressing BCRs that bind with moderate affinity to various self-Ags, but in addition also exhibit reactivity toward (lipo-)polysaccharide (TI-2) Ags. In vivo, the polyreactive Abs produced by MZ-B cells may be of crucial importance in the first-line of defense against bloodborn infections with encapsulated bacteria. Polyreactive naturally occurring Abs have been shown to be essential in the primary immune response against pathogens, largely because of their neutralizing effect (64, 65, 66). Herewith naive MZ-B cells form an integral part of (innate) immunity by providing a strategically located buffer of polyreactive B cells in the spleen that can rapidly produce neutralizing Abs (mainly IgM) upon bloodborn and life threatening infections with encapsulated bacteria or other microorganisms.

We thank Geert Mesander for flow cytometry assistance, Dr. Vlavel Fidler for helping with statistics, Alice Arendzen, RenseVeenstra, Marcel Bruinenberg, and Dr. Peter Terpstra for DNA sequencing, and Bert Hellinga for photography and DTP.

2

Abbreviations used in this paper: MZ-B, marginal zone B cell; BCR, B cell receptor for Ag; CDR, complementarity-determining region; NF-B, newly formed B cell; R, replacement mutation; RF-B, recirculating follicular B cell; S, silent mutation; TI, T cell independent; HEL, hen egg lysozyme.

1
Kraal, G..
1992
. Cells in the marginal zone of the spleen.
Int. Rev. Cytol.
132
:
31
2
MacLennan, I. C. M., Y.-J. Liu.
1991
. Marginal zone B cells respond to both polysaccharide antigens and protein antigens.
Res. Immunol.
142
:
346
3
Spencer, J., M. E. Perry, D. K. Dunn-Walters.
1998
. Human marginal-zone B cells.
Immunol. Today
19
:
421
4
Nieuwenhuis, P. 1971. On the origin and fate of immunologically competent cells. Ph. D. thesis, University of Groningen, Groningen, The Netherlands.
5
Gray, D., I. McConnell, D. S. Kumararatne, I. C. M. MacLennan, J. H. Humphrey, H. Bazin.
1984
. Marginal zone B cells express CR1 and CR2 receptors.
Eur. J. Immunol.
14
:
47
6
Waldschmidt, T. J., F. G. M. Kroese, L. T. Tygrett, D. H. Conrad, R. G. Lynch.
1991
. The expression of B cell surface receptors. III. The murine low-affinity IgE Fc receptor is not expressed on Ly 1 or “Ly 1-like” B cells.
Int. Immunol.
3
:
305
7
Oliver, A. M., F. Martin, J. F. Kearney.
1999
. IgMhighCD21high lymphocytes enriched in the splenic marginal zone generate effector cells more rapidly than the bulk of follicular B cells.
J. Immunol.
162
:
7198
8
Oliver, A. M., F. Martin, G. L. Gartland, R. H. Carter, J. F. Kearney.
1997
. Marginal zone B cells exhibit unique activation, proliferative and immunoglobulin secretory responses.
Eur. J. Immunol.
27
:
2366
9
Snapper, C. M., H. Yamada, D. Smoot, R. Sneed, A. Lees, J. J. Mond.
1993
. Comparative in vitro analysis of proliferation, Ig secretion, and Ig class switching by murine marginal zone and follicular B cells.
J. Immunol.
150
:
2737
10
Bos, N. A., J. C. A. M. Bun, B. Meedendorp, A. S. Wubbena, F. G. M. Kroese, V. A. Ploplis, and J. J. Cebra. 2000. B cell populations in antigen-free mice. Old Herborn University Seminar Monograph. In press.
11
Iijima, S., T. Yamane.
1968
. The spleen of germ free animals. M. Miyakawa, and T. D. Luckey, eds.
Advances in Germfree Research and Gnotobiology
139
Iliffe Books, London.
12
Pollard, M..
1967
. Germinal centers in germfree animals. H. Cottier, and N. Odartschenko, and R. Schindler, and C. C. Congdon, eds.
Germinal Centers in Immune Responses
343
Springer, Berlin.
13
Bazin, H., D. Gray, B. Platteau, I. C. M. MacLennan.
1982
. Distinct δ+ and δ B lymphocyte lineages in the rat.
Ann. N. Y. Acad. Sci.
399
:
157
14
Kumararatne, D. S., H. Bazin, I. C. M. MacLennan.
1981
. Marginal zones: the major B cell compartment of the rat spleens.
Eur. J. Immunol.
11
:
858
15
Makowska, A., N. N. Faizunnessa, P. Anderson, T. Midtvedt, S. Cardell.
1999
. CD1high B cells: a population of mixed origin.
Eur. J. Immunol.
29
:
3285
16
Dammers, P. M., N. K. de Boer, G. J. Deenen, P. Nieuwenhuis, F. G. M. Kroese.
1999
. The origin of marginal zone B cells in the rat.
Eur. J. Immunol.
29
:
1522
17
Dammers, P. M., A. Visser, E. R. Popa, P. Nieuwenhuis, and F. G. M. Kroese. 2000. Immunoglobulin VH gene analysis in rat: most marginal zone B cells express germline encoded VH genes and are ligand selected. In Lymphocytes in B Cell Neoplasia. P. Potter and F. Melchers, eds. Curr. Top. Microbiol. Immunol. 252:107.
18
Chen, X., F. Martin, K. A. Forbush, R. M. Perlmutter, J. F. Kearney.
1997
. Evidence for selection of a population of multi-reactive B cells into the splenic marginal zone.
Int. Immunol.
9
:
27
19
Martin, F., J. F. Kearney.
2000
. Positive selection from newly formed to marginal zone B cells depends on the rate of clonal production, CD19, and Btk.
Immunity
12
:
39
20
Liu, Y.-J., S. Oldfield, I. C. M. MacLennan.
1988
. Memory B cells in T cell-dependent antibody responses colonize the splenic marginal zone.
Eur. J. Immunol.
18
:
355
21
Dunn-Walters, D. K., P. G. Isaacson, J. Spencer.
1995
. Analysis of mutations in immunoglobulin heavy chain variable region genes of microdissected marginal zone (MGZ) B cells suggests that the MGZ of human spleen is a reservoir of memory B cells.
J. Exp. Med.
182
:
559
22
Tierens, A., J. Delabie, L. Michiels, P. Vandenberghe, C. De Wolf-Peters.
1999
. Marginal-zone B cells in the human lymph node and spleen show somatic hypermutations and display clonal expansion.
Blood
93
:
226
23
Tangye, S. G., Y.-J. Liu, G. Aversa, J. H. Phillips, J. E. de Vries.
1998
. Identification of functional human splenic memory B cells by expression of CD148 and CD27.
J. Exp. Med.
188
:
1691
24
Klein, U., K. Rajewsky, R. Küppers.
1998
. Human immunoglobulin (Ig)M+IgD+ peripheral blood B cells expressing the CD27 cell surface antigen carry somatically mutated variable region genes: CD27 as a general marker for somatically mutated (memory) B cells.
J. Exp. Med.
188
:
1679
25
Nieuwenhuis, P., F. G. M. Kroese, D. Opstelten, H. G. Seijen.
1992
. De novo germinal center formation.
Immunol. Rev.
126
:
77
26
Nieuwenhuis, P., F. J. Keuning.
1974
. Germinal centres and the origin of the B-cell system. II. Germinal centres in the rabbit spleen and popliteal lymph nodes.
Immunology
26
:
509
27
Kroese, F. G. M., ed. 1998. Immunology of the rat. In Handbook of Vertebrate Immunology. P.-P. Pastoret, H. Bazin, A. Govaerts, and P. J. Griebel, eds. Academic, London, p. 137.
28
Vermeer, L. A. 1995. Molecular analysis of rat B cell differentiation. Ph. D. thesis, University of Groningen, Groningen, The Netherlands.
29
Kroese, F. G. M., E. C. Butcher, P. A. Lalor, A. M. Stall, L. A. Herzenberg.
1990
. The rat B cell system: the anatomical localization of flow cytometry-defined B cell subpopulations.
Eur. J. Immunol.
20
:
1527
30
Hermans, M. H. A., D. Opstelten.
1991
. In situ visualisation of hematopoietic cell subsets and stromal elements in the rat and mouse bone marrow by immunostaining of frozen sections.
J. Histochem. Cytochem.
39
:
1627
31
Hünig, T., H. J. Wallny, J. K. Hartley, A. Lawetzky, G. Tiefenthaler.
1989
. A monoclonal antibody to a constant determinant of the rat T cell antigen receptor that induces T cell activation. Differential reactivity with subsets of immature and mature T lymphocytes.
J. Exp. Med.
169
:
73
32
Lefranc, M. P., V. Giudicelli, C. Ginestoux, J. Bodmer, W. Muller, R. Bontrop, M. Lemaitre, A. Malik, V. Barbie, D. Chaume.
1999
. IMGT, the international ImMunoGeneTics database.
Nucleic Acids Res.
27
:
209
33
Lang, P., R. Mocikat.
1991
. Immunoglobulin heavy-chain joining genes in the rat: comparison with mouse and human.
Gene
102
:
261
34
Küppers, R., M. Zhao, M.-L. Hansmann, K. Rajewsky.
1993
. Tracing B cell development in human germinal centres by molecular analysis of single cells picked from histological sections.
EMBO J.
12
:
4955
35
Kroese, F. G. M., N. K. de Boer, T. de Boer, P. Nieuwenhuis, A. B. Kantor, G. J. Deenen.
1995
. Identification and kinetics of two recently bone marrow-derived B cell populations in peripheral lymphoid tissues.
Cell. Immunol.
162
:
185
36
Brüggemann, M., J. Free, A. Diamond, J. Howard, S. Cobbold, H. Waldmann.
1986
. Immunoglobulin heavy chain locus of the rat: striking homology to mouse antibody genes.
Proc. Natl. Acad. Sci. USA
83
:
6075
37
Kabat, E. A., T. T. Wu, H. M. Perry, K. S. Gottesman, and C. Foeller. 1991. Sequences of Proteins of Immunological Interest. National Institutes of Health (publication no.91-3242), Bethesda, MD.
38
Lafaille, J. J., A. DeCloux, M. Bonneville, Y. Takagaki, S. Tonegawa.
1989
. Junctional sequences of T cell receptor γδ genes: implications for γδ T cell lineages and for a novel intermediate of V-(D)-J joining.
Cell
59
:
859
39
Chang, B., P. Casali.
1994
. The CDR1 sequences of a major proportion of human germline Ig VH genes are inherently susceptible to amino acid replacement.
Immunol. Today
15
:
367
40
Deenen, G. J., F. G. M. Kroese.
1993
. Kinetics of B cell subpopulations in peripheral lymphoid tissues: evidence for the presence of phenotypically distinct short-lived and long-lived B cell subsets.
Int. Immunol.
5
:
735
41
Suzuki, H., H. Shiku.
1992
. Preferential usage of JH2 in D-J joinings with DQ52 is determined by the primary DNA sequence and is largely dependent on recombination signal sequences.
Eur. J. Immunol.
22
:
2225
42
Alt, F. W., D. Baltimore.
1982
. Joining of immunoglobulin heavy chain gene segments: Implications from a chromosome with evidence of three D-JH fusions.
Proc. Natl. Acad. Sci. USA
79
:
4118
43
Gauss, G. H., M. R. Lieber.
1996
. Mechanistic constrains on diversity in human V(D)J recombination.
Mol. Cell. Biol.
16
:
258
44
Berek, C., M. Ziegner.
1993
. Somatic hypermutation and affinity maturation.
Immunol. Today
14
:
400
45
Steiniger, B., P. Barth, B. Herbst, A. Hartnell, P. R. Crocker.
1997
. The species-specific structure of microanatomical compartments in the human spleen: strongly sialoadhesin-positive macrophages occur in the perifollicular zone, but not in the marginal zone.
Immunology
92
:
307
46
Chen, C., P. M. Stenzel, M. B. Rittenberg.
1991
. Natural auto- and polyreactive antibodies differing from antigen-induced antibodies in the H chain CDR3.
J. Immunol.
147
:
2359
47
Crouzier, R., T. Martin, J. L. Pasquali.
1995
. Heavy chain variable region, light chain variable region, and heavy chain CDR3 influences on the mono- and polyreactivity and on the affinity of human monoclonal rheumatoid factors.
J. Immunol.
154
:
4526
48
Ichiyoshi, Y., P. Casali.
1994
. Analysis of the structural correlates for antibody polyreactivity by multiple reassortments of chimeric human immunoglobulin heavy and light chain V segments.
J. Exp. Med.
180
:
885
49
Ditzel, H. J., K. Itoh, D. R. Burton.
1996
. Determinants of polyreactivity in a large panel of recombinant human antibodies from HIV-1 infection.
J. Immunol.
157
:
739
50
Adib-Conquy, M., M. Gilbert, S. Avrameas.
1998
. Effect of amino acid substitutions in the heavy chain CDR3 of an autoantibody on its reactivity.
Int. Immunol.
10
:
341
51
Lipsanen, V., B. Walter, M. Emara, K. Siminovitch, J. Lam, A. Kaushik.
1997
. Restricted CDR3 length of the heavy chain is characteristic of six randomly isolated disease-associated VH J558+ IgM autoantibodies in lupus prone motheaten mice.
Int. Immunol.
9
:
655
52
Tornberg, U. C., D. Holmberg.
1995
. B-1a, B-1b and B-2 B cells display unique VHDJH repertoires formed at different stages of ontogeny and under different selection pressures.
EMBO J.
14
:
1680
53
Hardy, R. R., K. Hayakawa.
1992
. Developmental origins, specificities and immunoglobulin gene biases of murine Ly-1 B cells.
Int. Rev. Immunol.
8
:
189
54
Kroese, F. G. M., R. de Waard, N. A. Bos.
1996
. B-1 cells and their reactivity with the murine intestinal microflora.
Semin. Immunol.
8
:
11
55
Schroeder, H. W., F. Mortari, S. Shiokawa, P. M. Kirkham, R. A. Elgavish, F. E. Bertrand.
1995
. Developmental regulation of the human antibody repertoire.
Ann. N. Y. Acad. Sci.
764
:
242
56
Rajewsky, K..
1996
. Clonal selection and learning in the antibody system.
Nature
381
:
751
57
Hayakawa, K., M. Asano, S. A. Shinton, M. Gui, D. Allman, C. L. Stewart, J. Silver, R. R. Hardy.
1999
. Positive selection of natural autoreactive B cells.
Science
285
:
113
58
Loder, F., B. Mutschler, R. J. Ray, C. J. Paige, P. Sideras, R. Torres, M. C. Lamers, R. Carsetti.
1999
. B cell development in the spleen takes place in discrete steps and is determined by the quality of B cell receptor-derived signals.
J. Exp. Med.
190
:
75
59
Neuberger, M. S..
1997
. Antigen receptor signaling gives lymphocytes a long life.
Cell
90
:
971
60
MacLennan, I. C. M..
1998
. B-cell receptor regulation of peripheral B cells.
Curr. Opin. Immunol.
10
:
220
61
Cyster, J. G..
1997
. Signaling threshold and interclonal competition in preimmune B-cell selection.
Immunol. Rev.
156
:
87
62
Healy, J. I., C. C. Goodnow.
1998
. Positive versus negative signaling by lymphocyte antigen receptors.
Annu. Rev. Immunol.
16
:
645
63
Mason, D. Y., M. Jones, C. C. Goodnow.
1992
. Development and follicular localization of tolerant B lymphocytes in lysozyme/anti-lysozyme IgM/IgD transgenic mice.
Int. Immunol.
4
:
163
64
Briles, D. E., M. Nahm, K. Schroer, J. Davie, P. Baker, J. Kearney, R. Barletta.
1981
. Antiphosphocholine antibodies found in normal mouse serum are protective against intravenous infection with type 3 Streptococcus pneumoniae.
J. Exp. Med.
153
:
694
65
Ehrenstein, M. R., T. L. O’Keefe, S. L. Davies, M. S. Neuberger.
1998
. Targeted gene disruption reveals a role for natural secretory IgM in the maturation of the primary immune response.
Proc. Natl. Acad. Sci. USA
95
:
10089
66
Ochsenbein, A. F., T. Fehr, C. Lutz, M. Suter, F. Brombacher, H. Hengartner, R. M. Zinkernagel.
1999
. Control of early viral and bacterial distribution and disease by natural antibodies.
Science
286
:
2156