Abstract
B cell development past the pro-B cell stage in mice requires the Cul4-Roc1-DDB1 E3 ubiquitin ligase substrate recognition subunit VprBP. Enforced Bcl2 expression overcomes defects in distal VH-DJH and secondary Vκ-Jκ rearrangement associated with VprBP insufficiency in B cells and substantially rescues maturation of marginal zone and Igλ+ B cells, but not Igκ+ B cells. In this background, expression of a site-directed Igκ L chain transgene increases Igκ+ B cell frequency, suggesting VprBP does not regulate L chain expression from a productively rearranged Igk allele. In site-directed anti-dsDNA H chain transgenic mice, loss of VprBP function in B cells impairs selection of Igκ editor L chains typically arising through secondary Igk rearrangement, but not selection of Igλ editor L chains. Both H and L chain site-directed transgenic mice show increased B cell anergy when VprBP is inactivated in B cells. Taken together, these data argue that VprBP is required for the efficient receptor editing and selection of Igκ+ B cells, but is largely dispensable for Igλ+ B cell development and selection, and that VprBP is necessary to rescue autoreactive B cells from anergy induction.
Introduction
During B cell development, the Ig gene loci undergo an ordered series of DNA rearrangements to assemble functional Ig H and L chain genes. This process, called V(D)J recombination, is initiated by proteins encoded by RAG1 and RAG2, which introduce DNA double-strand breaks in these loci that are subsequently repaired via nonhomologous end joining. Structure–function studies have established that the N-terminal third of RAG1 is dispensable for the catalytic activity of the RAG protein complex, yet this region is evolutionarily highly conserved and is required for efficient and high-fidelity V(D)J recombination (1). How the RAG1 N terminus promotes this outcome remains unclear, but evidence from our laboratory and others suggests it functions at least in part as a protein interaction domain to recruit factors to facilitate chromosomal V(D)J recombination. We recently identified that Vpr-binding protein (VprBP; DCAF1), a substrate adaptor molecule for the Cul4-Roc1-DDB1 (CRL4) and EDD/UBR5 E3 ubiquitin ligase complexes (2), associates with the N-terminal region of RAG1, and that VprBP is required for B cell development and plays a role in V(D)J recombination (3). Specifically, we found that conditional disruption of Vprbp early in B cell development arrests B cell maturation at the pro-B-to-pre-B cell transition, but this developmental block is partially rescued by expressing functionally rearranged Ig transgenes. Loss of VprBP expression in B cells is associated with impaired VH-DJH gene rearrangement, reduced fidelity of VH-DJH joining, defects in cell cycle progression, and increased apoptosis (3).
Given the elevated levels of apoptosis observed in VprBP-deficient B cells, in this study we investigated whether enforced expression of the prosurvival factor Bcl2 can compensate for the loss of VprBP during B cell development, as has been observed in other cases of genetic insufficiency manifesting impaired B cell development (4–7). As in those cases, we find that Bcl2 expression partially rescues B cell development, substantially reconstituting marginal zone, but not follicular, B cell populations. Unexpectedly, however, most B cells maturing under this program express Igλ rather than Igκ. The loss of Igκ+ B cells in this context can be partially rescued in mice bearing a site-directed Igκ L chain transgene, suggesting VprBP does not regulate L chain expression from a productively rearranged Igk allele. More detailed analysis of V(D)J rearrangement patterns in pre-B cells and rare Igκ+ B cells isolated from VprBP-deficient mice provides evidence for inefficient distal VH-DJH gene rearrangement and secondary Igk rearrangements associated with receptor editing in these animals. However, the apparent V(D)J recombination defects are substantially rescued by enforced Bcl2 expression, ruling out a direct role for VprBP in mediating the V(D)J rearrangement process itself.
As an alternative, we speculated that VprBP functions indirectly to regulate the efficiency of BCR editing and selection of Igκ+ B cells. To test this possibility, we analyzed how the loss of VprBP function affects B cell development and selection in mice harboring the site-directed VH3H9/56R (56R) anti-DNA H chain transgene, which is used as a model of VH gene replacement as well as L chain receptor editing and selection (8). Our results suggest that VprBP insufficiency impairs VH gene replacement and selection of Igκ editor L chains, but does not interfere with the selection of Igλ editor L chains. Interestingly, both H and L chain site-directed transgenic mice show an increased frequency of phenotypically anergic B cells when VprBP is inactivated. Taken together, these data argue that VprBP is required for the efficient editing and selection of Igκ+ B cells, but is largely dispensable for Igλ+ B cell development and selection, and is necessary to salvage B cells from potential anergy induction.
Materials and Methods
Mice
Mice with the following conditional alleles or transgenes have been previously described, as follows: Vprbpfl (9), mb1-Cre (10), Eμ-2-22 Bcl2 (11), site-directed VH3H9/56R H chain (56R) (8), and site-directed 3-83 κ-chain lacking all 3′ Jκ segments (3-83κiJκ−) (12). The latter mouse strain was derived by breeding mice harboring both site-directed 3-83 H and L chains (The Jackson Laboratory strain C.129P2(B6)-Igktm2Rsky Ightm2Rsky/J) to C57BL/6 mice and screening for offspring heterozygous for only the 3-83κiJκ− allele. All mice were bred on a C57BL/6 background and maintained in individually ventilated microisolator cages in an American Association for the Accreditation of Laboratory Animal Care–certified animal facility at Creighton University. All experimental procedures were reviewed and approved by the Creighton University Institutional Animal Care and Use Committee.
Flow cytometry
For detecting surface Ags, single-cell suspensions were prepared for bone marrow and spleen and stained with cocktails of biotin- or fluorochrome-conjugated Abs, as described previously (13). The following Abs and clones were used: BD Biosciences (San Jose, CA) anti–B220-PE, anti–PE-Texas Red, or anti-CF594 (all clone RA3-6B2), anti–CD19-APC-Cy7 (1D3), anti–CD23-biotin (B3B4), anti–CD43-APC (S7), anti–CD43-biotin (S7), anti–CD93-APC (AA4.1), anti–IgMa-biotin (DS-1), anti–IgMb-PE (AF6-78), anti–κ-PerCP-Cy5.5 (187.1), anti–λ-FITC (R26-46), anti–Ly-51-PE (BP.1), and anti–Ly6C-PerCP-Cy5.5 (AL-21); BioLegend (San Diego, CA) anti–CD21/CD35-PE-Cy7 (7E9) and anti–IgD-APC-Cy7 (11-26c); eBioscience (San Diego, CA) anti–CD19-A700 (1D3), anti–CD4-A700 (GK1.5), anti–CD49b-PE-Cy7 (DX5), anti–CD93-PE (AA4.1), anti–CD93-PE-Cy7 (AA4.1), anti–IgM-FITC (II/41), anti–IgM-APC (II/41), anti–IgM-PE-Cy5 (II/41), and anti–IgD-FITC (11-26c); and Southern Biotech (Birmingham, AL) anti–CD24/heat-stable Ag–spectral red (30-F1). Samples stained with biotinylated Abs were detected using streptavidin-Qdot585 (Invitrogen, Carlsbad, CA) or streptavidin-BUV737 (BD Biosciences). To detect Vλx by flow cytometry, a protein G–purified Vλx-specific mAb, called clone 10C5 (14) (gift of E. L. Prak, Department of Pathology and Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA), was conjugated to Alexa Fluor647 using a kit, according to manufacturer’s instructions (Life Technologies), and stained as described above, except that Fc blocking reagent (anti-mouse CD16/32 Ab; BD Biosciences) was added prior to and during the staining procedure.
To detect intracellular proteins, cells were Fc blocked and stained first with Abs to detect surface Ags, washed, and then incubated with streptavidin Q585 or BUV737 to detect biotinylated Abs. A second round of surface staining with 10-fold excess anti–IgM-FITC (II/41) and unconjugated anti-κ Abs was done to block all surface IgM and Igκ. Next, cells were fixed and permeabilized for 20 min on ice (Cytofix/CytoPerm kit; BD Biosciences). After washing, cells were stained with eBioscience anti–Pax5-APC (1H9) and anti–Irf4-PerCP-eFluor710 (3H4), or eBioscience anti–IgM-APC (II/41) and BD Biosciences anti–κ-PerCP-Cy5.5 (187.1), or with control Abs conjugated to APC (eBR2a; eBioscience), PerCP-eFluor710 (eBRG1; eBioscience), or PerCP-Cy5.5 (R3-34; BD Biosciences).
Data collection and cell sorting were performed using a FACSAria flow cytometer (BD Biosciences). Data were analyzed using the FlowJo software (Tree Star, Ashland, OR).
V(D)J and intron recombining sequence–recombining sequence rearrangement assays
Genomic DNA was prepared from sorted bone marrow pre-B cells (Ly6C−CD4−DX5−B220+CD43−IgM−) and splenic CD19+ Igκ+ B cells, as described (15). Briefly, 106 cells were lysed in 200 μL PCR lysis buffer (10 mM Tris [pH 8.4], 50 mM KCl, 2 mM MgCl2, 0.45% Nonidet P-40, 0.45% Tween 20, and 60 μg/ml proteinase K), incubated at 55°C for 1 h, and then heated to 95°C for 10 min to inactivate proteinase K. This procedure yielded DNA template equivalent to 5000 genomes/μl that was used directly for PCR analysis. Igh, Igk, Igl, and intron recombining sequence–recombining sequence (IRS-RS) rearrangements were amplified by PCR from template DNA (10,000, 2,500, and 625 genome equivalents). Briefly, PCRs (50 μl) containing template DNA and 0.5 μM each primer (see Table I) in sample buffer (0.2 mM dNTPs, 20 mM Tris-HCl [pH 8.4], 50 mM KCl, 1.5 mM MgCl2, and 2.5 U Taq polymerase [Promega, Madison, WI]) were subjected to initial denaturation (Igh, Igk, IgλR1, and IRS-RS rearrangements: 94°C for 1 min; Igλx rearrangements: 94°C for 4 min), 30–40 cycles of amplification (Igh and IRS-RS rearrangements: 94°C for 30 s, 59°C for 1 min, 72°C for 2 min; IgVλx rearrangements: 94°C for 20 s, 60°C for 30 s, 72°C for 1.5 min; IgλR1 rearrangements: 94°C for 30 s, 48°C for 1 min, 72°C for 2 min; Vκ1 rearrangements: 94°C for 30 s, 60°C for 1 min, 72°C for 2 min; Vκ21 rearrangements: 94°C for 30 s, 55°C for 1 min, 72°C for 2 min), and then a final extension (Igh, Igk, IgλR1, and IRS-RS rearrangements: 72°C for 4 min; IgVλx rearrangements: 72°C for 5 min). Primers flanking a fragment from the CD14 locus were used as a loading control. PCR products were separated by agarose gel electrophoresis and visualized by ethidium bromide staining or by Southern blotting using nested 32P-labeled oligonucleotide probes (see Table I) and autoradiographic imaging with a Typhoon 9410 Variable Mode Imager.
Designation . | Sequence 5′-3′ . | Reference . |
---|---|---|
Normalization | ||
5′CD14L | GCTCAAACTTTCAGAATCTACCGAC | (34) |
5′CD14R | AGTCAGTTCGTGGAGGCCGGAAATC | (34) |
IgH | ||
5′DHL deg | GGAATTCGMTTTTTGTSAAGGGATCTACTACTGTG | (15) |
5′VHJ558 | CGAGCTCTCCARCACAGCCTWCATGCARCTCARC | (15) |
5′VHQ52 | CGGTACCAGACTGARCATCASCAAGGACAAYTCC | (15) |
5′VH7183 | CGGTACCAAGAASAMCCTGTWCCTGCAAATGASC | (15) |
3′JH4 | GGGGAATTCCTGAGGAGACGGTGACT | (35) |
JH4 probe | ACCCCAGTAGTCCATAGCATAGTAAT | (35) |
Igκ | ||
5′Vκ21 proximal | YBWGCTSACYCARTCTCCWRC | (36) |
5′Vκ1 distal | CARACTCCACTCTCCCTGCC | (36) |
5′IRS (κ deletion) | CTGACTGCAGGTAGCGTGGTCTTCTAG | (37) |
3′RS (κ deletion) | CTCAAATCTGAGCTCAACTGC | (12) |
3′Jκ5 | CCAAGCTTGTACTTACGTTTCAGCT | (38) |
Jκ5 probe | GCTCACGTTCGGTGCTGGGACCAAGCTGGAGCTGAAACGTAAGTAC | (39) |
IRS-RS probe | CTAGTGGCAGCCCAGGGTGGATCTCCCTAGGACTGCAGTTGAGCTC | This study |
Igλ | ||
5′VλR1 | ATGAATTCACTGGTCTAATAGGTGGTACCA | (40) |
3′JλR1 | TAGAATTCACTYACCTAGGACAG | (40) |
VλR probe | CTGTGCTCTATGGTACAGCACCC | (40) |
5′Vλx | GAGCTTAAGAAAGATGGAAGCCA | (41) |
3′VλxR2 | GTTCCACCGCCGAAAACATA | (42) |
Vλx probe | TGCTGATCGCTACCTTAGCATTTCCAACATCCAGCCTGAAGATGAAGCAATATACATCTG | This study |
Designation . | Sequence 5′-3′ . | Reference . |
---|---|---|
Normalization | ||
5′CD14L | GCTCAAACTTTCAGAATCTACCGAC | (34) |
5′CD14R | AGTCAGTTCGTGGAGGCCGGAAATC | (34) |
IgH | ||
5′DHL deg | GGAATTCGMTTTTTGTSAAGGGATCTACTACTGTG | (15) |
5′VHJ558 | CGAGCTCTCCARCACAGCCTWCATGCARCTCARC | (15) |
5′VHQ52 | CGGTACCAGACTGARCATCASCAAGGACAAYTCC | (15) |
5′VH7183 | CGGTACCAAGAASAMCCTGTWCCTGCAAATGASC | (15) |
3′JH4 | GGGGAATTCCTGAGGAGACGGTGACT | (35) |
JH4 probe | ACCCCAGTAGTCCATAGCATAGTAAT | (35) |
Igκ | ||
5′Vκ21 proximal | YBWGCTSACYCARTCTCCWRC | (36) |
5′Vκ1 distal | CARACTCCACTCTCCCTGCC | (36) |
5′IRS (κ deletion) | CTGACTGCAGGTAGCGTGGTCTTCTAG | (37) |
3′RS (κ deletion) | CTCAAATCTGAGCTCAACTGC | (12) |
3′Jκ5 | CCAAGCTTGTACTTACGTTTCAGCT | (38) |
Jκ5 probe | GCTCACGTTCGGTGCTGGGACCAAGCTGGAGCTGAAACGTAAGTAC | (39) |
IRS-RS probe | CTAGTGGCAGCCCAGGGTGGATCTCCCTAGGACTGCAGTTGAGCTC | This study |
Igλ | ||
5′VλR1 | ATGAATTCACTGGTCTAATAGGTGGTACCA | (40) |
3′JλR1 | TAGAATTCACTYACCTAGGACAG | (40) |
VλR probe | CTGTGCTCTATGGTACAGCACCC | (40) |
5′Vλx | GAGCTTAAGAAAGATGGAAGCCA | (41) |
3′VλxR2 | GTTCCACCGCCGAAAACATA | (42) |
Vλx probe | TGCTGATCGCTACCTTAGCATTTCCAACATCCAGCCTGAAGATGAAGCAATATACATCTG | This study |
Statistics
Data are presented as mean values ± SEMs. Collected data were subjected to ANOVA and post hoc testing using the PASW Statistics 22.0 software package (SPSS, Chicago, IL). Differences with a p value ≤0.05 are considered statistically significant.
Results
Enforced Bcl2 expression partially rescues B cell development in Vprbpfl/fl Cre+ mice, but most B cells developing in this background are Igλ+
We previously used a Cre-loxP approach to conditionally disrupt Vprbp expression in the B lineage by breeding the mb1-Cre transgene onto a strain background in which both Vprbp alleles contain loxP sites flanking exons 7 and 8 (Vprbpfl/fl-Cre+) (see Fig. 1A), and showed that B cell development was arrested at the pro-B-to-pre-B cell transition in these animals (3). Concomitant expression of functionally rearranged anti-HEL Ig H and L chain transgenes could partially bypass the developmental block, arguing that VprBP plays a role in V(D)J recombination, but evidence for defects in cell cycle progression and increased apoptosis in pre-B cells from these animals pointed to potential additional roles for VprBP in cell proliferation and survival (3). In several other genetic models of B cell developmental arrest, enforced Bcl2 expression has been shown to enable some maturing B cells to bypass the developmental block (4–7). To determine whether Bcl2 expression can similarly rescue B cell development in Vprbpfl/fl Cre+ mice, we bred the Eμ-2-22 Bcl2 transgene (Bcl2+) (11) onto this strain background and analyzed the frequency of developing and mature B cell subsets in the bone marrow and spleen using flow cytometry.
Enforced Bcl2 expression partially rescues lymphocyte cellularity in Vprbpfl/fl Cre+ mice. (A) Diagram of wild-type and conditional Vprbp alleles; mb1-Cre expression deletes exons 7 and 8 in mice homozygous for the conditional Vprbp alleles (Vprbpfl/fl) and causes B cell developmental arrest at the pro-B-to-pre-B cell transition. (B) Analysis of wild-type (WT) mice and Vprbpfl/fl mice lacking or containing the mb1-Cre (Cre+) and/or Eμ-2-22 Bcl2 (Bcl2+) transgenes. For each genotype, the total number of cells and lymphocytes was determined in the bone marrow and spleen. Data are represented as mean ± SEM. Statistically significant differences are indicated for selected group comparisons.
Enforced Bcl2 expression partially rescues lymphocyte cellularity in Vprbpfl/fl Cre+ mice. (A) Diagram of wild-type and conditional Vprbp alleles; mb1-Cre expression deletes exons 7 and 8 in mice homozygous for the conditional Vprbp alleles (Vprbpfl/fl) and causes B cell developmental arrest at the pro-B-to-pre-B cell transition. (B) Analysis of wild-type (WT) mice and Vprbpfl/fl mice lacking or containing the mb1-Cre (Cre+) and/or Eμ-2-22 Bcl2 (Bcl2+) transgenes. For each genotype, the total number of cells and lymphocytes was determined in the bone marrow and spleen. Data are represented as mean ± SEM. Statistically significant differences are indicated for selected group comparisons.
Consistent with our previous work, total bone marrow and spleen cellularities and the absolute number of lymphocytes in these organs in wild-type and Vprbpfl/fl mice were not significantly different, but Vprbpfl/fl Cre+ mice showed a significant decrease in total and lymphocyte cellularities in the spleen, but not in bone marrow (Fig. 1B) and a strong developmental arrest at the pro-B cell stage (B220+CD43+IgM−), with numbers of pre-B cells (B220+CD43−IgM−) reduced ∼80% in bone marrow, and B cells at later developmental stages reduced to <1% of normal levels in bone marrow and spleen (Fig. 2, Supplemental Table I). For this analysis, cell populations found to contaminate the gates used to identify the early B cell subsets in the bone marrow were specifically excluded, including DX5+ NK cells, CD4+ dendritic cell progenitors, and Ly6C+ myeloid and plasma cells (16). As expected from earlier studies (4, 11), Vprbpfl/fl Bcl2+ mice showed significantly higher spleen cellularities and increased absolute numbers of total lymphocytes and all B cell developmental subsets in the spleen compared with their Vprbpfl/fl counterparts (Figs. 1B, 2, Supplemental Table I). We also detected a small splenic pro-B cell-like population (B220lowCD43+AA4.1+) that has been observed in Bcl2-transgenic mice by others (17) (Fig. 2A). Notably, when Bcl2 was expressed in the Vprbpfl/fl Cre+ background, pre-B cell numbers were restored to levels similar to Vprbpfl/fl mice, but still remained well below levels observed in Vprbpfl/fl Bcl2+ mice, whereas the abundance of later B cell developmental subsets was increased 10- to 100-fold compared with Vprbpfl/fl Cre+ mice, but still remained well below levels found in Vprbpfl/fl mice (Fig. 2B, 2C, Supplemental Table I). Unexpectedly, when B cells were interrogated for κ and λ L chain expression, we found that ∼60–75% of the L chain–expressing CD19+ B cells in the bone marrow and spleen of Vprbpfl/fl Cre+ Bcl2+ mice were Igλ+, whereas the proportion was closer to 10–20% in wild-type, Vprbpfl/fl, and Vprbpfl/fl Bcl2+ mice (Fig. 3A, 3B). This outcome is primarily due to a significant reduction of Igκ+ B cells, rather than an increase in absolute numbers of Igλ+ B cells in Vprbpfl/fl Cre+ Bcl2+ mice compared with Vprbpfl/fl Bcl2+ mice (Fig. 3C). One additional noteworthy observation is that Bcl2 expression in Vprbpfl/fl Cre+ mice restored marginal zone (B220+AA4.1−CD21+CD23−) B cell numbers to levels observed in Vprbpfl/fl Bcl2+ mice, whereas transitional and follicular mature populations only recovered to ∼10% of the levels detected in Vprbpfl/fl Bcl2+ mice (Fig. 2C, Supplemental Table I). Taken together, these results show that enforced Bcl2 expression in Vprbpfl/fl Cre+ mice partially rescues B cell development, enabling substantial reconstitution of Igλ+, but not Igκ+ B cells, and marginal zone, but not follicular, B cells.
Enforced Bcl2 expression partially rescues B cell development in Vprbpfl/fl Cre+ mice. (A) Lymphocytes from mice with the indicated genotype were analyzed by flow cytometry for the expression of different surface markers using gating strategies defined under each row. Developmental subsets identified by the staining pattern are shown at right with corresponding gates. The percentage of cells in each gate is shown for representative animals. A splenic pro-B cell-like population detected in Vprbpfl/fl Bcl2+ mice is identified with an asterisk. (B and C) The absolute number of cells in various B cell developmental subsets in the bone marrow (B) or spleen (C) defined by flow cytometry in (A) was determined for each of the indicated mouse genotypes. Data are represented as mean ± SEM and are summarized in Supplemental Table I. Statistically significant differences are indicated for selected group comparisons.
Enforced Bcl2 expression partially rescues B cell development in Vprbpfl/fl Cre+ mice. (A) Lymphocytes from mice with the indicated genotype were analyzed by flow cytometry for the expression of different surface markers using gating strategies defined under each row. Developmental subsets identified by the staining pattern are shown at right with corresponding gates. The percentage of cells in each gate is shown for representative animals. A splenic pro-B cell-like population detected in Vprbpfl/fl Bcl2+ mice is identified with an asterisk. (B and C) The absolute number of cells in various B cell developmental subsets in the bone marrow (B) or spleen (C) defined by flow cytometry in (A) was determined for each of the indicated mouse genotypes. Data are represented as mean ± SEM and are summarized in Supplemental Table I. Statistically significant differences are indicated for selected group comparisons.
B cells developing in Vprbp fl/fl Cre+ Bcl2+ mice are mostly Igλ+. (A) Gated splenic CD19+ lymphocytes from mice with the genotypes shown in Fig. 2 were analyzed for Igκ and Igλ expression by flow cytometry. (B and C) The proportion (B) and absolute number (C) of Igκ+ and Igλ+ cells in the bone marrow and spleen for each of the indicated mouse genotypes were determined using flow cytometric data shown in (A). Proportions were calculated as the percentage of Igκ+ cells (black bar) or Igλ+ cells (gray bar) among total CD19+ L chain–expressing lymphocytes (i.e., Igκ+ + Igλ+). Data in (C) are represented as mean ± SEM (see also Supplemental Table I). Statistically significant differences are indicated for selected group comparisons.
B cells developing in Vprbp fl/fl Cre+ Bcl2+ mice are mostly Igλ+. (A) Gated splenic CD19+ lymphocytes from mice with the genotypes shown in Fig. 2 were analyzed for Igκ and Igλ expression by flow cytometry. (B and C) The proportion (B) and absolute number (C) of Igκ+ and Igλ+ cells in the bone marrow and spleen for each of the indicated mouse genotypes were determined using flow cytometric data shown in (A). Proportions were calculated as the percentage of Igκ+ cells (black bar) or Igλ+ cells (gray bar) among total CD19+ L chain–expressing lymphocytes (i.e., Igκ+ + Igλ+). Data in (C) are represented as mean ± SEM (see also Supplemental Table I). Statistically significant differences are indicated for selected group comparisons.
Loss of VprBP function in B cells is associated with impaired distal VH-DJH rearrangement and a bias toward proximal Jκ rearrangements, but both are normalized by enforced Bcl2 expression
Using a PCR–Southern blotting approach to analyze V(D)J rearrangement patterns in total bone marrow of Vprbpfl/fl Cre+ mice, we previously showed that VH-DJH and Vκ-Jκ rearrangement is impaired in these animals (3). To explain the selective deficit in Igκ+ B cells in Vprbpfl/fl Cre+ Bcl2+ mice, we considered the fact that the Igl locus is approximately one-tenth the size of the Igh and Igk loci in mice, and therefore hypothesized that VprBP is required for efficient V(D)J recombination of the large Igh and Igk loci, but is dispensable for V(D)J rearrangement involving the smaller Igl locus. To test this hypothesis, we extended our previous studies of V(D)J rearrangement patterns in Vprbpfl/fl Cre+ mice to Vprbpfl/fl Cre+ Bcl2+ mice and included additional primer sets (see Table I) to evaluate V(D)J rearrangements involving Igh and Igk V genes that are proximal or distal to the J segments, those occurring in the Igl locus, and those involving IRS-RS recombination (18), a form of secondary V(D)J rearrangement that generally occurs after exhaustive Vκ-Jκ rearrangement and results in the excision of Cκ from the Igk locus (κ deletion).
Because the B cell developmental block in Vprbpfl/fl Cre+ mice occurs at the pro-B to pre-B cell transition, we focused our analysis on the pre-B cell subset. Among sorted pre-B cells (B220+CD43−IgM−), the pattern of D-JH rearrangement or VH-DJH rearrangement involving proximal (VHQ52 and VH7183) or distal (VHJ558) gene clusters was observed to be quite similar among Vprbpfl/fl, Vprbpfl/fl Bcl2+, and Vprbpfl/fl Cre+Bcl2+ mice, but in Vprbpfl/fl Cre+ mice, levels of D-JH rearrangement were consistently enriched and rearrangements to the distal VHJ558 gene cluster were consistently diminished (Fig. 4A). By contrast, VH-DJH rearrangements involving the proximal gene VH clusters (VHQ52 and VH7183) were reduced only modestly, if at all, in Vprbpfl/fl Cre+ mice. Elevated D-JH rearrangements detected in Vprbpfl/fl Cre+ mice most likely reflect poor ongoing VH gene rearrangement that would normally diminish the pool of template DNA that had not yet undergone VH-DJH rearrangement.
Vprbp fl/fl Cre+ mice show inefficient distal V(D)J rearrangement in the IgH and Igκ L chain loci that are rescued by enforced Bcl2 expression. (A and B) Genomic DNA prepared from bone marrow pre-B cells (A) (B220+CD43−IgM−) or splenic Igκ+ B cells (B) (10,000, 2,500, or 625 cell equivalents) sorted from mice with the indicated genotypes was subjected to PCR and Southern hybridization to detect the V(D)J rearrangements shown at right. Amplification of the nonrearranging CD14 locus was performed as a loading control.
Vprbp fl/fl Cre+ mice show inefficient distal V(D)J rearrangement in the IgH and Igκ L chain loci that are rescued by enforced Bcl2 expression. (A and B) Genomic DNA prepared from bone marrow pre-B cells (A) (B220+CD43−IgM−) or splenic Igκ+ B cells (B) (10,000, 2,500, or 625 cell equivalents) sorted from mice with the indicated genotypes was subjected to PCR and Southern hybridization to detect the V(D)J rearrangements shown at right. Amplification of the nonrearranging CD14 locus was performed as a loading control.
When Igk and Igl rearrangements in sorted pre-B cells were analyzed, we found these levels to be slightly reduced in Vprbpfl/fl Cre+ mice compared with Vprbpfl/fl mice (Fig. 4A). By contrast, however, both Vprbpfl/fl Bcl2+ and Vprbpfl/fl Cre+ Bcl2+ mice showed levels of these rearrangement products that were similar to each other, but were consistently elevated compared with Vprbpfl/fl and Vprbpfl/fl Cre+ mice, particularly for IRS-RS and Vλ1-Jλ1 rearrangements, suggesting this effect is attributed mainly to enforced Bcl2 expression.
To determine whether the distribution of Igk and Igl rearrangements observed in pre-B cells differs from those having undergone selection and migration out of the bone marrow, we analyzed the pattern of these rearrangements in splenic Igκ+ B cells sorted from the four mouse genotypes (Fig. 4B). Interestingly, the rare Igκ+ B cells isolated from Vprbpfl/fl Cre+ mice show predominant rearrangement to Jκ2 by the proximal Vκ21 segment, whereas rearrangements involving the distal Vκ1 segment, IRS-RS, and Vλ1-Jλ1 are at or below levels detected in Vprbpfl/fl mice. However, when Bcl2 is expressed in the Vprbpfl/fl Cre+ background, proximal and distal Vκ rearrangement patterns are normalized to those detected in Vprbpfl/fl Bcl2+ mice, whereas IRS-RS and Vλ1-Jλ1 rearrangement levels are slightly higher than those observed in Vprbpfl/fl Bcl2+ mice.
VprBP does not regulate Pax5 and IRF4 expression
The strong bias of rearrangements toward Jκ2 observed in Igκ+ B cells from Vprbpfl/fl Cre+ mice, taken together with the apparent defect in distal VH-DJH rearrangement in these animals and the striking reduction of Igκ+ B cells of Vprbpfl/fl Cre+ Bcl2+ mice compared with Vprbpfl/fl Bcl2+ mice, raised the possibility that loss of VprBP function adversely affects factors or processes that regulate distal VH-DJH rearrangement and/or secondary V(D)J rearrangements initiated in response to self-reactivity (receptor editing). The transcription factors Pax5 and Irf4 have been previously implicated in regulating distal VH-DJH rearrangement and L chain receptor editing, respectively (19, 20). To determine whether loss of VprBP function affects Pax5 and Irf4 expression, we compared the intracellular levels of these proteins among Vprbpfl/fl, Vprbpfl/fl Cre+, Vprbpfl/fl Bcl2+, and Vprbpfl/fl Cre+Bcl2+ mice by flow cytometry (Fig. 5A). We found that the levels of these proteins in pro-B and pre-B cells were not significantly different among the four mouse genotypes, suggesting VprBP does not directly regulate the expression of these proteins.
Loss of VprBP function does not alter intracellular Pax5 or Irf4 levels, but reduces cytoplasmic μ- and κ-chain levels. (A and B) Gated bone marrow pro-B (B220+CD43+sIgM−) and pre-B cells (B220+CD43−sIgM−) from mice with the genotypes shown in Fig. 2 were analyzed for the expression of (A) intracellular Pax5 and Irf4 or (B) cytoplasmic μ (cμ)- and κ (cκ)-chains by flow cytometry. Representative staining by Ag-specific and isotype control Abs (solid and dashed lines, respectively) is shown for animals of each genotype. In (B), the additional shaded histogram represents Ag-specific staining of cells that were not permeabilized. Note that pro-B cells with a Pax5low phenotype are CD19−, whereas the Pax5high cells are CD19+. The mean fluorescence intensity was calculated for n = 5 animals/genotype, and the data presented as the mean ± SEM.
Loss of VprBP function does not alter intracellular Pax5 or Irf4 levels, but reduces cytoplasmic μ- and κ-chain levels. (A and B) Gated bone marrow pro-B (B220+CD43+sIgM−) and pre-B cells (B220+CD43−sIgM−) from mice with the genotypes shown in Fig. 2 were analyzed for the expression of (A) intracellular Pax5 and Irf4 or (B) cytoplasmic μ (cμ)- and κ (cκ)-chains by flow cytometry. Representative staining by Ag-specific and isotype control Abs (solid and dashed lines, respectively) is shown for animals of each genotype. In (B), the additional shaded histogram represents Ag-specific staining of cells that were not permeabilized. Note that pro-B cells with a Pax5low phenotype are CD19−, whereas the Pax5high cells are CD19+. The mean fluorescence intensity was calculated for n = 5 animals/genotype, and the data presented as the mean ± SEM.
Absence of surface Igκ+ B cells in Vprbpfl/fl Cre+Bcl2+ mice is not due to cytoplasmic sequestration of Igκ chain
If loss of VprBP function impairs V(D)J recombination efficiency, one might expect that intracellular levels of Igμ and Igκ chains may be reduced in early stages of B cell development. Alternatively, if loss of VprBP function impairs surface expression of these molecules, they may accumulate inside the cell. To address these possibilities, we compared the intracellular levels of Igμ and Igκ chains in pro-B and pre-B cells among Vprbpfl/fl, Vprbpfl/fl Cre+, Vprbpfl/fl Bcl2+, and Vprbpfl/fl Cre+Bcl2+ mice by flow cytometry (Fig. 5B). Consistent with the former possibility, levels of cytoplasmic Igμ and Igκ (cμ and cκ) chains were significantly reduced in both pro-B and pre-B cell subsets in Vprbpfl/fl Cre+ mice compared with Vprbpfl/fl mice. Enforced Bcl2 expression in Vprbpfl/fl Cre+ mice substantially rescued cμ levels in pro-B cells, but a substantial fraction of pre-B cells lacked cμ expression. Notably, cκ levels were not rescued by enforced Bcl2 expression in pre-B cells in Vprbpfl/fl Cre+ mice, despite the apparent ability of these cells to undergo Igk rearrangement (see Fig. 4).
Loss of VprBP function in anti-dsDNA transgenic mice impairs receptor editing and Igκ chain selection, but not Igλ chain selection
To more specifically test whether loss of VprBP function impairs receptor editing in B cells, we used a well-established mouse model of receptor editing in which a transgene expressing an IgH specific for dsDNA, called VH3H9/56R (hereafter called 56R), is knocked into the endogenous locus (8). In 56R mice, most developing B cells undergo receptor editing to remove or silence dsDNA autoreactivity. Two major mechanisms of receptor editing have been described in the 56R model, as follows: 1) H chain gene replacement, in which an endogenous upstream VH and/or DH gene segment replaces most of the 56R transgene via rearrangement with a cryptic recombination signal sequence at the 3′ end of the transgene (21); and 2) successive L chain rearrangement, which enables selection of one of a restricted set of editor L chains that neutralize the dsDNA-binding activity of the 56R H chain (8). Certain editor L chains paired with 56R, notably Vκ38c, also confer reactivity toward a Golgi-associated Ag, causing the B cell to adopt a marginal zone phenotype and allowing it to escape negative selection (22).
In the experiments described in this work, we bred the 56R transgene onto the Vprbpfl/fl, Vprbpfl/fl Cre+, Vprbpfl/fl Bcl2+, Vprbpfl/fl Cre+ Bcl2+ strain backgrounds and compared the frequency of B cells that express μ H chain from the 56R transgene or the nontargeted allele (IgMa or IgMb, respectively), or that express the Igκ or Igλ L chain. The abundance of other B cell developmental subsets was also analyzed. One important initial finding from these studies was that, compared with Vprbpfl/fl Cre+ mice, Vprbpfl/fl Cre+56R+ mice had over twice the number of bone marrow pre-B cells, and 10- to 20-fold more splenic transitional (B220+AA4.1+) and mature (B220+AA4.1−) B cells (Supplemental Table I). Thus, the developmental block observed in Vprbpfl/fl Cre+ mice is partially overcome by 56R transgene expression. Consistent with this finding, cμ was detected at similar levels in bone marrow pro-B and pre-B cells from Vprbpfl/fl 56R+ mice as compared with Vprbpfl/fl Cre+ 56R+ mice (Fig. 6A). This result suggests that reduced cμ expression detected in pro-B and pre-B cells in Vprbpfl/fl Cre+ mice is unlikely to be attributed to impaired expression from a functionally rearranged allele.
56R mice lacking functional VprBP express IgH, but not Ig κ-chain, through early B cell developmental stages. (A and B) Bone marrow (BM) B cell developmental subsets identified by the staining pattern at right were analyzed for the expression of cytoplasmic μ H chain (cyto-μ) (A) or κ L chain (cyto-κ) (B) and presented as in Fig. 5B. The mean fluorescence intensity of Ag-specific staining is summarized in Supplemental Fig. 1C.
56R mice lacking functional VprBP express IgH, but not Ig κ-chain, through early B cell developmental stages. (A and B) Bone marrow (BM) B cell developmental subsets identified by the staining pattern at right were analyzed for the expression of cytoplasmic μ H chain (cyto-μ) (A) or κ L chain (cyto-κ) (B) and presented as in Fig. 5B. The mean fluorescence intensity of Ag-specific staining is summarized in Supplemental Fig. 1C.
When splenic B cells expressing IgMa were analyzed in the different 56R transgenic mouse strains, we found that the proportion of total CD19+IgM+ B cells expressing IgMa increased from ∼89–90% in Vprbpfl/fl 56R+ and Vprbpfl/fl 56R+ Bcl2+ mice, to 94% in Vprbpfl/fl Cre+ 56R+ and Vprbpfl/fl Cre+ 56R+ Bcl2+ mice (Fig. 7). These data suggest the efficiency of 56R VH gene replacement is impaired by loss of VprBP function. When the proportion of Igλ-expressing B cells was analyzed in the spleens from the different 56R transgenic mouse strains, we found that ∼10% of the L chain–expressing cells in Vprbpfl/fl 56R+ mice were Igλ+, a value that is close to that observed in Vprbpfl/fl mice. This proportion increased to ∼25% in Vprbpfl/fl 56R+ Bcl2+ mice, and was further elevated to 45 and 75% in Vprbpfl/fl Cre+ 56R+ and Vprbpfl/fl Cre+ 56R+ Bcl2+ mice, respectively (Fig. 7). Like the comparison between Vprbpfl/fl Bcl2+ and Vprbpfl/fl Cre+ Bcl2+ mice, the basis for the skewing toward Igλ-expressing B cells in Vprbpfl/fl Cre+ 56R+ Bcl2+ mice relative to Vprbpfl/fl 56R+ Bcl2+ mice is attributed to a >10-fold reduction in the absolute number of Igκ+ cells in the former strain, whereas the abundance of Igλ+ cells differs between these strains by only ∼2-fold (Supplemental Fig. 1A, Supplemental Table I). We also detected low cκ levels in developing and mature B cell subsets from Vprbpfl/fl Cre+ 56R+ and Vprbpfl/fl Cre+ 56R+ Bcl2+ mice (Fig. 6B, Supplemental Fig. 1C), suggesting that loss of Igκ+ cells in these animals is not caused by intracellular retention of the κ-chain. Interestingly, we note that the mean fluorescence intensity of Igλ staining is markedly reduced in Vprbpfl/fl Cre+ 56R+ Bcl2+ mice (as well as other 56R-transgenic strains) compared with Vprbpfl/fl Cre+ Bcl2+ mice (Fig. 7A, Supplemental Fig. 2A). We speculated that this phenomenon is due to the prevalent pairing of the 56R H chain with Vλx as an editor L chain, which is stained poorly, if at all, by commercially available Igλ-specific Abs that detect Vλ1/2 (23). Consistent with this possibility, among the Igλ+ cells in all the 56R+ strains, ∼90% were stained a Vλx-specific Ab (13) (Fig. 7). Few Vλx+ B cells were costained with the Vλ1/2-specific Ab, but most were surface Ig (sIg)κdim (Fig. 7A). Dual κ/λ expression may be a means to dilute the density of autoreactive BCR expressing Igκ (24). Further characterization of the splenic CD19+ Vλx+ B cells in Vprbpfl/fl Cre+ 56R+ Bcl2+ mice shows that most of these cells have a B220highsIgMintsIgD+CD21+CD23+AA4.1− phenotype (data not shown).
Loss of VprBP function in 56R anti-dsDNA transgenic mice impairs receptor editing and selection of Ig κ-chains, but is dispensable for Ig λ chain selection. (A) Splenocytes isolated from Vprbpfl/fl mice lacking or carrying the 56R, Cre, and/or Bcl2 transgenes in various combinations were analyzed for the expression of sIgMa (expressed from the 56R transgene) and sIgMb (expressed from the endogenous allele), Igκ and Igλ1/2, Igλx and Igλ1/2, or Igκ and Igλx on CD19+ lymphocytes using flow cytometry. The absolute numbers of CD19+ B cells expressing these markers are shown for each genotype in Supplemental Fig. 1A and Supplemental Table I. (B) The proportion of CD19+ B cells expressing the markers examined in (A) was determined and presented as in Fig. 3B. Statistically significant differences are indicated for selected group comparisons.
Loss of VprBP function in 56R anti-dsDNA transgenic mice impairs receptor editing and selection of Ig κ-chains, but is dispensable for Ig λ chain selection. (A) Splenocytes isolated from Vprbpfl/fl mice lacking or carrying the 56R, Cre, and/or Bcl2 transgenes in various combinations were analyzed for the expression of sIgMa (expressed from the 56R transgene) and sIgMb (expressed from the endogenous allele), Igκ and Igλ1/2, Igλx and Igλ1/2, or Igκ and Igλx on CD19+ lymphocytes using flow cytometry. The absolute numbers of CD19+ B cells expressing these markers are shown for each genotype in Supplemental Fig. 1A and Supplemental Table I. (B) The proportion of CD19+ B cells expressing the markers examined in (A) was determined and presented as in Fig. 3B. Statistically significant differences are indicated for selected group comparisons.
Loss of VprBP function in anti-dsDNA transgenic mice increases the frequency of phenotypically anergic B cells
Analysis of other splenic B cell populations also revealed several interesting effects of 56R transgene expression on Bcl2 transgenic and/or Vprbpfl/fl Cre+ backgrounds. First, the splenic pro-B cell-like (B220lowAA4.1+) population observed in Vprbpfl/fl Bcl2+ mice is not detected when the 56R transgene is concomitantly expressed (Fig. 8A). Second, among the three transitional (CD19+B220+AA4.1+) B populations that have been defined (25), the transitional (T)3 (sIgMlow/− CD23+) subset, which is thought to be functionally anergic (26), represents the vast majority of transitional B cells in both Vprbpfl/fl Cre+ 56R+ mice and Vprbpfl/fl Cre+ 56R+ Bcl2+ mice, whereas this population is much less predominant relative to T1 and T2 populations in the other mouse strains tested (Fig. 8A, Supplemental Table I). An expanded population of sIgMlow/− CD23− cells is also detected in these animals (Fig. 8A). Further phenotypic characterization of the T3 subset and the sIgMlow/−CD23− population reveals they share a similar phenotype aside from differential CD23 expression: mostly sIgDdimCD21dimsIgκ−sIgλ−, but cμ+ and cκ− (Fig. 8B, and data not shown). Third, in the mature (CD19+B220+AA4.1−) B cell compartment, Vprbpfl/fl Cre+56R+ mice and Vprbpfl/fl Cre+56R+ Bcl2+ mice consistently show a smaller proportion of marginal zone B cells relative to follicular B cells compared with their counterparts lacking the Cre transgene (Fig. 8C). Vprbpfl/fl 56R+ Bcl2+ (−/+ Cre) mice also show an expanded population of mature B cells that are CD21−CD23+ (Fig. 8A). Further characterization of these cells indicates that they phenotypically resemble T3 B cells: mostly sIgMdimsIgDdimIgκ−Igλ− and cμ+, but, whereas this population contains a large fraction of cκ+ cells in Vprbpfl/fl 56R+ Bcl2+ mice, this population is uniformly cκ− in Vprbpfl/fl Cre+ 56R+ Bcl2+ (Fig. 8B, and data not shown). Taken together, these data suggest that loss of VprBP function in 56R mice reduces the efficiency of receptor editing by both H chain gene replacement and sequential L chain gene rearrangement, and impairs selection of 56R+ B cells into the marginal zone compartment. As a result, most B cells developing in 56R mice in the absence of functional VprBP undergo rearrangement and selection of λ L chain–expressing Vλx and/or acquire an anergic phenotype.
Loss of VprBP function in 56R mice increases anergic B cell populations and impairs B cell selection into the marginal zone compartment. (A) Flow cytometry was used to identify various splenic transitional and mature B cell subsets as in Fig. 2A; the absolute numbers of cells in each subset are summarized in Supplemental Fig. 1B and Supplemental Table I. A splenic pro-B cell-like population detected in Vprbpfl/fl Bcl2+ mice (identified with an asterisk) is notably absent in Vprbpfl/fl 56R+Bcl2+ mice. (B) Cytoplasmic μ H chain (cyto-μ) or κ L chain (cyto-κ) was analyzed by flow cytometry in the populations identified in (A) and presented as in Fig. 5B. Examples of cyto-μ and cyto-κ staining profiles are shown for nonconventional B cell populations identified by an asterisk in (A) (transitional sIgM−CD23− and mature CD21−CD23+, respectively). The mean fluorescence intensity of Ag-specific staining is summarized in Supplemental Fig. 1C. (C) The proportion of marginal zone (MZ) and follicular mature (FM) B cells in the spleen for each of the indicated mouse genotypes was determined using flow cytometric data shown in (A). Proportions were calculated as the percentage of FM cells (black bar) or MZ cells (gray bar) among total gated FM and MZ cells (i.e., FM + MZ). Statistically significant differences are indicated for selected groups expressing the 56R transgene.
Loss of VprBP function in 56R mice increases anergic B cell populations and impairs B cell selection into the marginal zone compartment. (A) Flow cytometry was used to identify various splenic transitional and mature B cell subsets as in Fig. 2A; the absolute numbers of cells in each subset are summarized in Supplemental Fig. 1B and Supplemental Table I. A splenic pro-B cell-like population detected in Vprbpfl/fl Bcl2+ mice (identified with an asterisk) is notably absent in Vprbpfl/fl 56R+Bcl2+ mice. (B) Cytoplasmic μ H chain (cyto-μ) or κ L chain (cyto-κ) was analyzed by flow cytometry in the populations identified in (A) and presented as in Fig. 5B. Examples of cyto-μ and cyto-κ staining profiles are shown for nonconventional B cell populations identified by an asterisk in (A) (transitional sIgM−CD23− and mature CD21−CD23+, respectively). The mean fluorescence intensity of Ag-specific staining is summarized in Supplemental Fig. 1C. (C) The proportion of marginal zone (MZ) and follicular mature (FM) B cells in the spleen for each of the indicated mouse genotypes was determined using flow cytometric data shown in (A). Proportions were calculated as the percentage of FM cells (black bar) or MZ cells (gray bar) among total gated FM and MZ cells (i.e., FM + MZ). Statistically significant differences are indicated for selected groups expressing the 56R transgene.
VprBP does not regulate Igκ expression from a functionally rearranged allele
Although the data presented to this point suggest VprBP is required for efficient receptor editing, we wished to exclude the possibility that Igκ expression, rather than Igk locus recombination, is adversely affected by loss of VprBP function. One piece of evidence arguing against this possibility is that the mean fluorescence intensity of Igκ staining among CD19+Igκ+ B cells in Vprbpfl/fl Cre+ mice is at or above that observed in Vprbpfl/fl mice (Supplemental Fig. 2A), suggesting that, once rearranged, the Igk locus can be expressed at normal levels. If the frequency of Igκ+ B cells emerging in Vprbpfl/fl Cre+ mice is limited by Igk rearrangement efficiency, then enforced expression of a functionally rearranged L chain transgene in Vprbpfl/fl Cre+ mice should increase the number of Igκ+ B cells in these animals. To test this possibility, we obtained mice harboring a site-directed Igκ-expressing transgene, called 3-83κiJκ− (hereafter termed 3-83κ), which lacks endogenous 3′ Jκ segments to prevent transgene excision by secondary Vκ-Jκ rearrangement (12). We then bred the 3-83κ transgene onto the Vprbpfl/fl, Vprbpfl/fl Cre+, Vprbpfl/fl Bcl2+, and Vprbpfl/fl Cre+ Bcl2+ strain backgrounds, and compared the abundance and distribution of B cells expressing the Igκ or Igλ L chain, as well as other B cell developmental subsets (Fig. 9A, Supplemental Table II). We find that the number of splenic Igκ+ B cells in Vprbpfl/fl Cre+3-83κ+ Bcl2+ mice is ∼3-fold higher than in their Vprbpfl/fl Cre+ Bcl2+ counterparts, with the proportion of Igκ+ cells increased by ∼20% (Fig. 9B, Supplemental Table II). Moreover, surface Igκ expression in B cells from Vprbpfl/fl 3-83κ+ mice (−/+ Bcl2) is similar to their Vprbpfl/fl Cre+3-83κ+ (−/+ Bcl2) counterparts, as judged by comparing mean fluorescence intensity (Supplemental Fig. 2A), and levels of cκ staining are significantly higher in all splenic transitional and mature B cell subsets in Vprbpfl/fl Cre+3-83κ+ (−/+ Bcl2) mice compared with their Vprbpfl/fl Cre+ (−/+ Bcl2) counterparts (Fig. 10B, 10C, Supplemental Fig. 2E). Taken together, these data suggest that a functionally rearranged κ L chain gene can be expressed at normal levels from the endogenous locus in the absence of functional VprBP, and that inefficient primary and secondary rearrangement of the Igk locus is the most likely contributing factor for the inability of Igκ+ B cells to develop under these conditions. Nevertheless, it is somewhat surprising that such a large proportion of L chain–expressing B cells in Vprbpfl/fl Cre+3-83κ+ Bcl2+ mice is Igλ+, considering these cells encode a functionally rearranged κ L chain transgene. One potential explanation is that the site-directed 3-83κ allele is still susceptible to IRS-RS rearrangement that could delete Cκ and trigger endogenous Igl rearrangement. In support of this possibility, PCR–Southern blotting experiments show higher levels of IRS-RS rearrangement in the bone marrow of Vprbpfl/fl Cre+3-83κ+Bcl2+ mice compared with Vprbpfl/fl Cre+3-83κ+ mice (Fig. 9C), and concomitantly lower levels of cκ staining in all splenic transitional and mature B cell subsets (Supplemental Fig. 2E).
Enforced 3-83κ L chain expression promotes development of Igκ+ B cells in Vprbpfl/fl Cre+ Bcl2+ mice. (A) Splenocytes isolated from Vprbpfl/fl mice lacking or carrying the 3-83κ, Cre, and/or Bcl2 transgenes in various combinations were analyzed for the expression of Igκ and Igλ on CD19+ lymphocytes using flow cytometry. (B) The proportion and absolute number of CD19+ B cells expressing Igκ or Igλ were determined and presented as in Fig. 3B and 3C. Statistically significant differences are indicated for selected groups expressing the 3-83κ transgene. (C) Genomic DNA prepared from total bone marrow (10,000, 2,500, or 625 cell equivalents) with the indicated genotypes was subjected to PCR and Southern hybridization to detect IRS-RS and Vλ1-to-Jλ1 rearrangements. Amplification of the nonrearranging CD14 locus was performed as a loading control.
Enforced 3-83κ L chain expression promotes development of Igκ+ B cells in Vprbpfl/fl Cre+ Bcl2+ mice. (A) Splenocytes isolated from Vprbpfl/fl mice lacking or carrying the 3-83κ, Cre, and/or Bcl2 transgenes in various combinations were analyzed for the expression of Igκ and Igλ on CD19+ lymphocytes using flow cytometry. (B) The proportion and absolute number of CD19+ B cells expressing Igκ or Igλ were determined and presented as in Fig. 3B and 3C. Statistically significant differences are indicated for selected groups expressing the 3-83κ transgene. (C) Genomic DNA prepared from total bone marrow (10,000, 2,500, or 625 cell equivalents) with the indicated genotypes was subjected to PCR and Southern hybridization to detect IRS-RS and Vλ1-to-Jλ1 rearrangements. Amplification of the nonrearranging CD14 locus was performed as a loading control.
Enforced 3-83κ L chain expression promotes development of Igκ+ B cells in Vprbpfl/fl Cre+ Bcl2+ mice. (A–C) Flow cytometry was used to identify various splenic transitional (A and B) and mature (A and C) B cell subsets, as in Fig. 2A; the absolute numbers of cells in each subset are summarized in Supplemental Fig. 2C and Supplemental Table II. A splenic pro-B cell-like population detected in Vprbpfl/fl Bcl2+ mice (identified with an asterisk) is notably present in Vprbpfl/fl 3-83κ+Bcl2+ mice. Conventional and nonconventional splenic transitional (B) and mature (C) B cell subsets were analyzed for cytoplasmic μ H chain (cyto-μ) or κ L chain (cyto-κ) using flow cytometry, as in Fig. 5B.
Enforced 3-83κ L chain expression promotes development of Igκ+ B cells in Vprbpfl/fl Cre+ Bcl2+ mice. (A–C) Flow cytometry was used to identify various splenic transitional (A and B) and mature (A and C) B cell subsets, as in Fig. 2A; the absolute numbers of cells in each subset are summarized in Supplemental Fig. 2C and Supplemental Table II. A splenic pro-B cell-like population detected in Vprbpfl/fl Bcl2+ mice (identified with an asterisk) is notably present in Vprbpfl/fl 3-83κ+Bcl2+ mice. Conventional and nonconventional splenic transitional (B) and mature (C) B cell subsets were analyzed for cytoplasmic μ H chain (cyto-μ) or κ L chain (cyto-κ) using flow cytometry, as in Fig. 5B.
Two other interesting observations were also gleaned from the flow cytometric analysis of 3-83κ transgenic mice. First, we find that 3-83κ transgene expression in Vprbpfl/fl Bcl2+ mice does not perturb the population of splenic pro-B–like cells detected in these animals, which stands in contrast to the effect of 56R expression in the same background (compare Figs. 8A and 10A). Second, whereas 3-83κ expression is less efficient than 56R expression in rescuing B cell development in Vprbpfl/fl Cre+ mice (as assessed by comparing the absolute numbers of cells in each developmental subset in bone marrow and spleen), the opposite is true when the transgenes are expressed in Vprbpfl/fl Cre+ Bcl2+ mice (compare Supplemental Tables 1, 2). Nevertheless, for both Vprbpfl/fl Cre+ 3-83κ and Vprbpfl/fl Cre+ 56R mice, enforced Bcl2 expression further restores B cell development (Supplemental Tables 1, 2).
Discussion
How the N-terminal third of RAG1 helps promote efficient and high-fidelity Ag receptor gene rearrangement remains ambiguous. At least part of its role appears likely to involve recruiting other cellular factors to regulate or direct the V(D)J recombination machinery or facilitate the repair of RAG-induced DNA double-strand breaks, but few reported RAG1-interacting factors have been tested to formally establish their requirement and role in V(D)J recombination. Previously, we discovered that VprBP and its associated CRL4 E3 ubiquitin ligase complex associate with the RAG1 N terminus. Using conditional gene disruption and complementation strategies in mice, we further established that VprBP is required for normal B cell maturation, and that the developmental arrest induced by loss of VprBP function can be partially rescued by the expression of functionally rearranged Ig transgenes (3). Further insight into how VprBP promotes V(D)J rearrangement of endogenous Ig loci is limited by the fact that the cells undergo apoptosis at the pro-B to pre-B cell transition. However, the events leading to apoptosis upon loss of VprBP function may be separable from VprBP’s potential role in V(D)J recombination, because conditional disruption of VprBP expression also causes apoptosis in nonlymphoid cells (9). Therefore, we tested whether enforced expression of the prosurvival gene Bcl2 might suppress apoptosis triggered by factors extrinsic to V(D)J recombination, and allow B cells to develop beyond the pro-B to pre-B cell transition, in which the requirement of VprBP in processes occurring at later stages of B cell development could be evaluated.
Our finding that loss of VprBP function is associated with impaired distal VH and Vκ rearrangement indicates that VprBP function is not required for V(D)J recombination per se, but is rather required to support long-distance V(D)J recombination events. However, this defect can be overcome by enforced Bcl2 expression, arguing that VprBP is not directly involved in rendering the distal regions of the Igh and Igk loci accessible to V(D)J recombination. Rather, the observation that enforced Bcl2 expression partially restores B cell development in Vprbpfl/fl Cre+ mice, yet only selectively reconstitutes maturation of Igλ+ B cells, but not Igκ+ B cells, argues that VprBP plays a role in supporting rearrangement and/or selection processes that operate intrinsically at the Igk locus. This contention is made plausible by previous studies showing that rearrangement and selection processes at the Igk and Igl loci are independently regulated, with the latter relying heavily on NF-κB–dependent signaling that can be functionally substituted by transgenic expression of Bcl2 (27). One attractive possibility is that VprBP supports receptor-editing events initiated in the Igk locus to rescue autoreactivity. In this scenario, loss of VprBP function would inhibit secondary V(D)J rearrangements in the Igk locus. This possibility is supported by the finding that rare Igκ+ B cells recovered from Vprbpfl/fl Cre+ mice mostly harbor an Igk rearrangement involving Vκ and Jκ segments that lie in close proximity to one another (Vκ21-Jκ2), which would most likely represent a primary Vκ-Jκ rearrangement in this locus (Fig. 4B). Further evidence for this contention is that rearrangements involving Vκ1, although only detected at low levels in these Igκ+ B cells, are also skewed toward rearrangement with Jκ1 (Fig. 4B).
A plausible alternative explanation for the presence of Igλ+ B cells in Vprbpfl/fl Cre+ mice is that VprBP normally functions to suppress receptor-editing events in the Igk locus. In this scenario, loss of VprBP function might then enable exhaustive rearrangement of the Igk locus, activation of Igl rearrangement, and selection of Igλ+ B cells. However, if this were the case, one might have expected the pattern of Vκ rearrangement in Igκ+ B cells from Vprbpfl/fl Cre+ mice to be heavily skewed toward usage of Jκ5 as an indicator of exhaustive Igk rearrangement, which is not observed. One might argue, however, that because the Vκ21-Jκ2 L chain is effective at vetoing H chain–associated anti-DNA autoreactivity (28), and because most of the rare Igκ+ B cells from Vprbpfl/fl Cre+ mice harbor rearrangements involving 3′ VH7183 and VHQ52 gene family members that tend to be expressed early in ontogeny (29) and may be prone to autoreactivity (30), the overrepresentation of the Vκ21-Jκ2 L chain gene rearrangement may simply reflect a selection bias. Although we cannot entirely exclude this possibility, two lines of evidence argue that this is not the case. First, V(D)J rearrangements in the rare Igκ+ B cells from Vprbpfl/fl Cre+ mice involving the distal Vκ1 segment, although reduced compared with Vprbpfl/fl mice, are also mostly rearranged to the proximal Jκ1 segment (Fig. 4B). Second, one might predict that if H chain autoreactivity were further enforced, and Igk rearrangement was not affected by loss of VprBP function, selection of specific Igκ editors should be increased. However, this outcome is not observed in our experiments with the 56R mice, because Vprbpfl/fl Cre+ 56R+ mice have a larger proportion of B cells expressing H chain from the site-directed 56R allele than Vprbpfl/fl 56R+ mice, yet the proportion of Igκ+ cells is nevertheless diminished in these animals (Fig. 7). Concomitantly, selection of B cells into the marginal zone compartment, which is driven by 56R H chain pairing with the Vκ38c-bearing editor L chain (22), is also diminished in 56R mice lacking VprBP (regardless of Bcl2 transgene status) (Fig. 8C). Taken together, these data suggest that, in 56R mice, loss of VprBP function limits the ability of B cells to undergo sequential V(D)J rearrangements in the Igk locus to obtain an editor L chain capable of neutralizing the anti-DNA reactivity of the 56R H chain.
One consequence of a failure to effectively edit the Igκ chain when H chain anti-dsDNA autoreactivity is enforced by transgene expression is an increase in B cells with anergic phenotypes. For example, in transgenic mice that cannot rescue anti-dsDNA B cell specificity by receptor editing due to concomitant RAG2 deficiency, most developing B cells acquire an Ag-experienced IgMdimCD21−CD23+ phenotype (31), which is similar to the phenotype of the T3 B cell population reported to be functionally anergic (26). Our finding that B cells with a T3-like phenotype represented the largest subset of transitional B cells in Vprbpfl/fl Cre+56R+ mice (regardless of Bcl2 expression), and were also enriched in the mature (AA4.1−) B cell compartment, suggests that loss of VprBP impairs the ability of autoreactive B cells to be rescued by receptor editing, leading them to acquire an anergic phenotype.
Taken together, the data suggest VprBP plays a key role in the editing and selection of Igκ L chains in response to B cell autoreactivity. Loss of VprBP function impairs receptor editing, which may trigger apoptosis of self-reactive B cells. This would be the fate for most developing B cells, because estimates suggest 25–75% of B cells in a primary repertoire are autoreactive (32). In Vprbpfl/fl Cre+ mice, this frequency may be higher due to impaired distal VH gene rearrangement and skewing of the primary H chain repertoire toward 3′ VH gene families enriched in autoreactive specificities. The rare Igκ+ B cells detected in Vprbpfl/fl Cre+ mice may represent examples of B cells in which the primary Vκ rearrangement successfully neutralized the autoreactivity of the H chain. Enforced Bcl2 expression allows developing B cells in Vprbpfl/fl Cre+ mice to bypass the restriction on the usage of distal VH and Vκ gene segments, supports cellular signals that promote the generation of Igλ+ B cells, and also promotes survival of autoreactive B cells that would otherwise undergo apoptosis due to a failure to successfully edit BCR specificity away from autoreactivity. Under these conditions, cells may acquire an Igλ+ or an anergic phenotype, thereby avoiding clonal deletion. This possibility could explain the overrepresentation of B cells with these phenotypes in Vprbpfl/flCre+Bcl2+ mice compared with Vprbpfl/fl Bcl2+ mice, particularly when 56R is concomitantly expressed. We note, however, that, whereas loss of VprBP function impairs selection of Igκ editor L chains in 56R mice, it does not appear to affect Igλ L chain selection, as most B cells developing in Vprbpfl/flCre+Bcl2+56R+ mice express the effective Vλx editor L chain, rather than the poor Vλ1/2 L chain editors that predominate in Vprbpfl/flCre+Bcl2+ mice. Tethering VprBP to the V(D)J recombinase through its association with full-length RAG1 provides an attractive and convenient means to rapidly communicate signals to reinitiate V(D)J recombination in the Igk locus when the BCR is subjected to autoantigenic stimulation. The underlying mechanism(s) by which VprBP regulates this process remains to be determined, but could involve positive or negative feedback through its function as a substrate receptor for the CRL4 E3 ubiquitin ligase, through the actions of one of its associating factors, or through its recently described activity as serine/threonine kinase (2, 33).
Acknowledgements
We thank Dr. Luning Prak for the gift of anti-Vλx Ab and Krista Namminga for technical support.
Footnotes
This work was supported by National Institutes of Health Grants R56AI091748-01A1 and GM102487 (to P.C.S.). The National Center for Research Resources provided support for research laboratory construction (C06 RR17417-01), the Creighton University Animal Resource Facility (G20RR024001), and the Typhoon variable mode imaging system (1S10RR027352).
The online version of this article contains supplemental material.
References
Disclosures
The authors have no financial conflicts of interest.