Direct reduction of antigen receptor expression in polyclonal B cell populations developing in vivo results in light chain receptor editing

Abstract

Secondary antibody variable region gene segment rearrangement, termed receptor editing, is a major mechanism contributing to B lymphocyte self tolerance. However, the parameters that determine whether a B cell undergoes editing or not are a current subject of debate. We tested the role that level of BCR expression plays in regulation of receptor editing in a polyclonal population of B cells differentiating in vivo. Expression of an shRNA for κ light chain RNA in B cells resulted in reduction in levels of this RNA and surface BCRs. Strikingly, fully mature and functional B cells that developed in vivo and efficiently expressed the shRNA predominantly expressed BCRs containing λ light chains. This shift in L chain repertoire was accompanied by inhibition of development, increased Rag gene expression and increased λ V gene segment cleavage events at the immature B Cell stage. These data demonstrate that reducing the translation of BCRs that are members of the natural repertoire at the immature B Cell stage is sufficient to promote editing.

INTRODUCTION

Antigen receptors play pivotal roles in regulating lymphocyte activation, proliferation and developmental progression (1-3). Early in B Cell development, expression of an IgM heavy (H) chain and formation of the pre-BCR results in suppression of the expression of the Rag-1 and 2 proteins and, thus, VDJ recombination at the H chain loci (3-5). Pre-BCR signaling also promotes a phase of proliferation leading to the stage in which light (L) chain V gene segment recombination is initiated (5-7). Expression of a functional L chain leads to expression of Cell surface IgM (sIgM), suppression of Rag gene expression and thus L chain VJ rearrangement, as well as promotion of further development (8).

Superimposed on these processes is the tolerance mechanism termed receptor editing or revision (9-12). Development of B cells expressing sIgM with certain autoreactivities is blocked at an early immature stage and the expression of the Rag proteins is sustained. This leads to ongoing VJ recombination at the κ and then λ L chain loci. A developing B Cell whose autoreactivity is altered or reduced via editing then proceeds in development (11).

The role that BCR and pre-BCR ligands play in regulation of primary B Cell development and receptor editing is a current subject of controversy. While ligand(s) for the pre-BCR have been found, their engagement appears unnecessary for functional pre-BCR signaling (13, 14). This suggests that formation of a membrane bound pre-BCR complex is sufficient to provide the “tonic” or “basal” signaling required to suppress Rag expression and promote development.

That level of BCR expression on developing B cells and, hence amount of basal BCR signaling might regulate receptor editing was first suggested from the analysis of mice hemizygous for Ig transgenes encoding an anti-class I MHC BCR. In the apparent absence of BCR ligands, the B cells in these mice displayed extensive receptor editing, but B cells in mice homozygous for the Ig transgene locus, resulting in an approximately two-fold increase in sIgM expression did not (15, 16). Additional studies supportive of this concept showed that perturbation of BCR signaling pathways either by deletion of coreceptors or downstream effectors or treatment with pharmacologic agents targeting these pathways could influence editing (16-18). More recent studies have implicated modulation of the Syk-PI(3)K-Akt-Foxo pathway by basal BCR signaling as being crucial in the control of receptor editing (19, 20).

More direct evidence of a role for BCR expression and basal signaling in the regulation of receptor editing has been obtained by analysis of the behavior of immature B cells developing in in vitro cultures after expression of their transgenic BCRs are genetically ablated. (21, 22). This results in induction of Rag gene expression, endogenous VJ rearrangement, and “back” differentiation to an immature phenotype. This led to a model for the regulation of receptor editing in which basal signaling from sIgM played a central, suppressive role, and that autoantigen engagement by the BCR induced receptor editing not directly via BCR signaling, but due to stimulation of BCR internalization via endocytosis. This, in turn, led to reduced basal signaling from cell surface BCRs.

As has been pointed out, however, all of the above studies are subject to caveats due to the experimental systems employed (16, 23). For example, expression of Ig transgenes begins early in development, abnormally accelerating or preventing subsequent stages of differentiation. IL-7-supported bone marrow cultures undoubtedly do not accurately mimic the microenvironmental “niches” in which developing B Cell reside in vivo. Addition of exogenous antagonists or agonists of BCR signaling pathways to such cultures may result in additional non-physiological responses. As such, we have investigated the effects of direct reduction of BCR expression levels in developing B cells in vivo using a κ light chain RNA “knockdown” approach. This approach differs fundamentally from those of previously published studies in which expression of the κ locus was precluded by the introduction of germline deletions (24, 25), as developing B cells would first have to productively rearrange and express a κ allele to become susceptible to reduction of BCR levels via κ RNA knockdown.

MATERIALS AND METHODS

Ligation-mediated PCR (LM-PCR)

Genomic DNA for LM-PCR was prepared from sorted cells according to an established protocol (26). After ligation of oligonucleotide linkers, DNAs were amplified with either κ or λ locus specific primers to detect particular dsDNA breaks in these loci as described (27). The λ-specific signal end break primer (5′GGAGATGTAGCCACCTGTTAAG3′) was kindly provided by Dr. Pamela Nakajima.

RNA isolation, real-time RT-PCR and sequence analysis

Total mRNA was purified from sorted Cell populations using the RNeasy Mini Kit (Qiagen) according to the manufacturer. First strand cDNA was synthesized using TaqMan reverse transcription reagents (Applied Biosystems) with random hexamer primers. Relative-quantitative real-time RT-PCR was performed using TaqMan gene expression assays on an ABI Prism 7000 sequence detection system (Applied Biosystems). The primers used were obtained from Applied Biosystems: κ light chain (Mm01611305_g1), Rag-2 (Mm00501300_m1) and GAPDH (Mm99999915_g1, used as endogenous control). Real-time RT-PCR reactions were performed in triplicate and delta delta CT values were normalized to those obtained from GAPDH RNA amplification. The mean and standard deviations were calculated using the Microsoft Excel software. The synthesized cDNA was also used for normal PCR reactions. Variable regions of λ light chain cDNA were amplified with the primer pair 5′TGGAGACAAGGCTGCCCTCACCATCACAG3′ and 5′GAGCTCYTCAGRGGAAGGTGGAAACABGGT3′ (28). Purified PCR products were cloned into plasmid vectors using the pGEM-T Easy Vector System (Promega, Madison, WI). Sequencing reactions were performed on the plasmid inserts in the Kimmel Cancer Center genomics facility and the CLUSTAL W multiple sequence alignment program was utilized to align the V and J regions of λ light chains.

Hybridomas

Total splenocytes from chimeric mice were stimulated with LPS (20 μg/ml) and IL-4 (50 ng/ml) for 3 days. Hybridomas were constructed using the SP2/0 fusion partner as previously described (29).

Mice

C57BL/6 (CD45.2) and C57BL/6.SJL-Ptprca Pepcb/BoyJ (CD45.1) congenic mice were purchased from The Jackson Laboratory and maintained under specific pathogen-free conditions and received autoclaved food and water. Mice were 6-12 weeks old at initiation of experiments. The use of mice in these studies was approved by the Institutional Animal Care and Use Committee of Thomas Jefferson University under protocol 344A.

Antibodies

The following antibodies were from eBioscience: biotin-, APC- and Pacific Blue-anti-B220 (RA3-6B2), PE-Cy7-anti-IgM (11/41), PE-anti-TCRβ (H57-597), APC-anti-CD43 (S7, Ly48), biotin- and PE-Cy7-anti-CD23 (B3B4), APC-anti-C1qRp (AA4.1), biotin-anti-GL7 (Ly-77), PerCP-Cy5.5-SA, biotin-anti-CD45.1 (A20) and PE-anti-CD45.2 (104); from BD biosciences: PE-anti-BP-1 (6C3), APC-anti-CD21/CD35 (7G6), biotin-anti-HSA (30-F1), biotin-anti-CD5 (Ly-1) and biotin-anti-CD138 (281-2); from Southern Biotechnology Associates: biotin-anti-IgD (11-26), biotin-anti-κ and PE-anti-λ.. PE-anti-IgM was from Jackson ImmunoResearch Laboratories, and SA-Alexa 633 was from Molecular Probes.

B lymphocyte cultures

B lymphocyte cultures were as previously described (29). Briefly, BM cultures were performed with murine IL-7 (32 ng/ml, R&D Systems). Nal̈ve splenic B lymphocytes were enriched by negative selection with anti-CD43 MACS microbeads (Miltenyi Biotec) and cultured in RPMI 1640 medium (10% FBS, 2 mM L-glutamine, 100 U/ml penicillin, 100 ug/ml streptomycin, and 5 uM β-mercaptoethanol), LPS (20 ug/ml) and murine IL-4 (50 ng/ml, PeproTech Inc.).

Retrovirus constructs

The following shRNA-encoding 97mer oligonucleotide was designed using the online program RNAi Central-RNAi Oligo Retriever (http://katahdin.cshl.org/html/scripts/resources.pl) to target the constant region of κ light chain mRNA, and synthesized by Integrated DNA Technologies (IDT): TGCTGTTGACAGTGAGCGACTCAGTCGTGTGCTTCTTGAATAGTGAAGCCACAGATGT ATTCAAGAAGCACACGACTGAGGTGCCTACTGCCTCGGA. The region in bold is homologous to the Cκ coding region, beginning at nucleotide 486. This region was cloned into the MSCV/LTRmiR30-PIG (LMP) retroviral vector ((30) from Open Biosystems) according to the manufacturer’s protocol. Retroviral supernatants were harvested from 293T cells transfected using Fugene 6 (Roche) according to the manufacturer’s protocol. Cultured BM or spleen cells were infected with retroviral supernatant in six-well plates containing 4 μg/ml polybrene (Sigma-Aldrich) by spinning at 2500 rpm at RT for 1.5 hrs on each of the next two days.

Bone marrow transduction and mixed bone marrow chimeras

Bone marrow cells from 5-fluorouracil treated (5-FU, American Pharmaceutical Partners Inc., injected (i.p.)) CD45.2 B6 mice were cultured overnight in complete RPMI 1640 medium supplemented with stem Cell factor (100 ng/ml), IL-3 (10 ng/ml), and IL-6 (5 ng/ml) (PeproTech Inc.) according to an established protocol (31). The cells were then spin-infected as described above on each of the next two days. After washing in PBS, a 1:1 mixture of the infected cells and freshly isolated BM cells from CD45.2 B6 mice were adoptively transferred i.v. into irradiated (11Gy) CD45.1 congenic recipients. This BM Cell mixture was used to improve the level of T Cell reconstitution in the resulting chimeric mice. Chimeric mice were given antibiotics and analyzed or immunized at least eight weeks later.

Immunizations

Chimeric mice were immunized with 100 ug of NP16-CGG (Biosearch Technologies) in alum i.p. and were sacrificed at day nine for flow cytometric and histological analysis as previous described (32).

Flow cytometry and cell sorting

Up to six-color flow cytometric analyses were performed using Becton Dickinson LSRII or FACSCalibur flow cytometers and data analyzed using the FlowJo software (Tree star). Cell sorting was performed on a MoFlow high-performance cell sorter (DakoCytomation).

Immunofluorescence

Spleens were isolated and fixed in 4% paraformaldehyde and 10% sucrose for 2 hrs at 4°C and processed by snap-freezing, cryosectioning and immunostaining as previously described (29, 33). Stained sections were analyzed on a Leica DM5000B digital microscope using related software (Leica Microsystems).

Statistical analyses

Statistical significance was performed using the Microsoft Excel program by a two-tailed, unpaired Student’s t test.

RESULTS

We designed five shRNAs targeting the constant region exon of κ L chain mRNA. DNA encoding these shRNAs was introduced into the MSCV-LTRmiR30 PIG (LMP) retroviral vector that contains a gene encoding eGFP (30). C57BL/6 (B6) splenic B cells were activated with LPS in vitro and transduced with the κ shRNA vectors or the control LMP vector. One day later, surface expression of κ L chains was assayed by flow cytometry, and levels of κ RNA transcripts were evaluated by QRTPCR in FACS purified B220+, GFP+ cells. One shRNA construct, termed KKD68, reduced surface κ expression approximately two-fold and κ RNA to 40% of control levels (Figure 1A and B).

Figure 1. Knock down of κ light chain expression in cell cultures.

C57BL/6 (B6) splenic B lymphocytes were cultured with LPS (20 ug/ml) and IL-4 (50 ng/ml) and then infected with KKD68 or control vector retroviruses as described in Materials and Methods. A: Surface κ light chain levels were evaluated by flow cytometry. B: The relative expression levels of κ RNA were quantified by real time RT-PCR in FACS purified B220+GFP⁺ B cells from these cultures. C: IL-7 supplemented B6 BM cultures were infected with KKD68, KKD59 or control vector retroviruses as described in Materials and Methods. Surface λ and κ light chain levels were evaluated in the B220⁺IgM⁺ B cell subset by flow cytometry. Gates are set to show the percentages of λ⁺ and λ⁺, κ⁺ B cells. As in splenic B cells (panel A), average surface levels of κ expression were reduced approximately two-fold among GFP⁺, κ⁺ cells in the KKD68 cultures. D: Relative expression levels of Rag-2 RNA were quantified by real time RT-PCR in the FACS purified B220⁺GFP⁺ B cell subset from these BM cultures. Figures are representative of data obtained from at least two independent experiments.

Immature B cells developing in vitro that express the anti-kappa shRNA reveal evidence of receptor editing

To initially determine whether the reduction of κ expression mediated by the KKD68 shRNA might induce receptor editing, an IL-7 bone marrow (BM) culture system was employed. B6 BM cultures were infected with the KKD68 retrovirus, a κ knockdown retroviral vector that did not result in reduction in surface κ expression on LPS activated splenic B cells (KKD59), and the control vector for two days. Two days later, the frequency of B cells expressing surface λ and κ L chains were evaluated. KKD68-transduced, immature B cells contained three to four-times the frequency of λ-expressing cells as controls. In addition, a major subpopulation of KKD68-transduced cells that expressed both λ and κ L chains was detected (Figure 1C). The cultures transduced with the control, KKD59 and KKD68 vectors revealed an increased frequency of B cells expressing very low to undetectable levels of L chains, perhaps due to nonspecific effects of retroviral gene expression on development in vitro. FACS-purified B220+, GFP+ cells from KKD68-transduced BM cultures expressed more than two-fold more Rag-2 RNA than controls (Figure 1D). These data indicate that B cells in BM cultures that express the KKD68 shRNA undergo an increased frequency of L chain receptor editing.

ShRNA expressing B cells developing in bone marrow chimeras display altered primary development

To assess the effects of KKD68 shRNA expression on B cell development in vivo, we transduced B6 BM cells with the KKD68 or control vectors and then reconstituted lethally irradiated B6.CD45.1 recipients with a mixture of this BM and freshly isolated, untreated B6 BM. Ten to 12 weeks later, donor-derived lymphocyte development was assayed in the chimeric mice by histology and flow cytometry. The percentages of KKD68-transduced cells in the pro-B and early pre-B stages of development (B220^low, CD43+) appeared normal, but increases in the percentage of late pre-B cells (B220^low, CD43⁻, sIgM⁻/IgM^low) were observed relative to controls (Figure 2A). Also, the frequency of immature B cells (B220^low, CD43-, sIgM+) was increased in the KKD68-transduced compartment and these cells expressed lower levels of sIgM than controls, and the frequency of BM mature B cells (B220^high) transduced with the KKD68 vector was reduced.

Figure 2. Altered development of B lymphocytes transduced with the KKD68 retrovirus in bone marrow chimeras.

Bone marrow B lymphocytes from BM chimeras generated using KKD68 or control vector retrovirus-infected BM cells were stained with Abs specific for the indicated markers and analyzed by flow cytometry. A: Percentages of various fractions (45) are indicated next to the gates. Upper panels show late pre B cells (B220^lowCD43⁻ IgM7IgM^low); IgM^int and IgM^high immature B cells (B220^lowCD43⁻); and, mature B cells (B220^highCD43⁻gM⁺). Lower panels show Hardy fraction A (B220+CD43+BP-1⁻HSA⁻); fraction B (B220+CD43+BP-1⁻HSA⁺); and, fraction C (B220+CD43+BP-1⁺HSA⁺). B: lower panels, pro-/pre-B cells (B220^lowIgM⁻), immature B cells (B220^lowIgM⁺) and recirculating mature B cells (B220^highIgM⁺) were analyzed in the B220⁺GFP⁺ gated subsets illustrated in the upper panels. Percentages of cells in each fraction are shown next to the gates in the lower panels. The data were derived from at least two BM chimeras of each type.

Whereas the majority of BM B220^low pro, pre-B and immature B cells transduced with either the KKD68 or control viruses expressed high levels of GFP, many transduced B220^high cells expressed intermediate to low levels of GFP (Figure 2B). Since this phenomenon was observed in both vector and KKD68 transduced B cell populations it could not be attributed only to κ knockdown but may have resulted from dilution and silencing of retroviral genomes during pre-BCR-driven proliferation and subsequent differentiation. In agreement with the data in Figure 2A, the KKD68 shRNA had inhibitory effects on immature BM B cell development, as the frequency of B220^high, IgM^low, GFP+, mature B cells was reduced, and the frequency of B220^low, sIgM⁻, GFP+ B cells was increased compared to controls (Figure 2B). In total, these data suggest that the developmental progression of many KKD68-transduced B cells is inhibited at the pre-B to immature B cell transition, but a subpopulation escapes this fate and enters the mature B cell pool expressing normal sIgM levels.

KKD68 shRNA vector gene expression results in decreased levels of κ RNA, increased levels of Rag gene and λ L chain expression, and increased frequencies of λ V gene segment cleavage events in vivo

The above data suggested that a subpopulation of KKD68-transduced B cells might escape a developmental block at the immature B cell stage by undergoing L chain receptor editing, resulting in expression of λ L chains. To test this idea we assessed levels of κ and λ L chain surface expression in the B220^low, GFP^high and GFP⁻ BM subpopulations of the two types of chimeras. While κ-expressing cells were abundant and λ-expressing cells barely detectable in both compartments in control chimeras, in the GFP+ compartment of KKD68 chimeras few κ expressing cells could be detected and a small subpopulation of B cells that expressed both surface κ and λ L chains was present (Figure 3A).

Figure 3. Knock down of κ light chain in BM B cells in BM chimeric mice is associated with increased levels of λ light chain and Rag-2 gene expression and λ locus recombination.

A: Surface κ and λ light chain levels were evaluated in immature B220^low subsets in BM chimeras created with KKD68 or control vector retrovirus-transduced BM. Percentages of B cells in each gate in the lower panels are shown. B: BM B cells from B6 and BM chimeric mice created with KKD68 or control vector retrovirus-transduced BM were stained with anti-B220 and anti-κ light chain Abs, and B220^low, surface κ^neg cells that were either GFP⁺ or GFP⁻ were purified by FACS (upper panel). A representative flow cytometric plot is shown. Plots obtained from all BM samples were similar in terms of subset distribution. The relative expression levels of Rag-2 and κ light chain RNA were quantified by real time RT-PCR (upper and middle panels, respectively) in these fractions and the frequency of double stranded breaks in the λ locus quantified by LM-PCR (lower panel) in the GFP⁺ fractions. Figures are representative of data obtained from at least three independent experiments.

Next, we FACS purified surface B220^low, κ^neg BM B cells from various types of chimeric and control mice and measured κ and Rag-2 RNA levels, and frequency of cleavage events at V gene segments in the κ and λ loci in these subpopulations. Rag-2 RNA levels were approximately two-fold elevated, and κ RNA levels approximately 8-fold decreased in the KKD68-transduced subpopulation, compared to all control subpopulations (Figure 3B). Much reduced, and dramatically increased frequencies of DNA breaks at J gene segments in κ (Supplemental Figure 1) and λ loci (Figure 3B and Supplemental Figure 1), respectively, were detected by LM-PCR in the KKD68-transduced subpopulation compared to controls. These latter data indicate that most developing B cells that undergo only κ editing or rearrange and express both κ alleles are lost due to KKD68 shRNA knockdown. While alterations in development due to expression of the KKD68 shRNA might have influenced the data obtained in these PCR assays, it seems unlikely this could account for the approximately eight-fold decrease and 100-fold increase in κ RNA and RS-λ DNA breaks, respectively, in the KKD68-transduced B220^low, κ^neg BM B cells versus controls.

Among GFP+ B cells in the spleen, lymph nodes and peritoneal cavity those transduced with the KKD68 vector contained a dramatically increased frequency of surface λ+ cells relative to GFP⁻ B cells. Smaller subpopulations of GFP+ B cells in the spleen and LN of KKD68 chimeras were also present that did not detectably express λ L chains, and that expressed both surface κ and λ L chains (Figure 4A).

Figure 4. Knock down of κ light chain results in increased λ light chain expression among peripheral B cells.

A: Cells from peripheral lymphoid tissues (spleen - SP, lymph nodes - LN, and peritoneal cavity - PTC: upper, middle and lower groups of panels, respectively) from BM chimeras generated with KKD68 or control vector retrovirus-transduced BM were analyzed by flow cytometry. Percentages of cells in each fraction are shown in the gates. B: Spleen cryosections from BM chimeric mice were analyzed by immunohistology and fluorescence microscopy. Upper panels: Sections were stained with anti-TCRβ (red) and anti-B220 (blue). Middle panels: Sections were stained with anti-λ (red) and anti-κ (blue) light chain. Lower panels: Sections were stained with anti-IgM (red) and anti-IgD (blue). The data shown are representative of those obtained from at least three mice from each type of chimera.

While Figure 2 shows that the percentage of mature B cells in the BM within the KKD68-expressing lymphocyte subpopulation was somewhat reduced, Figure 4B illustrates that the spleens in all the chimeric mice had normal lymphoid architecture, with KKD68-transduced cells (GFP+) uniformly distributed in the red pulp, white pulp and MZ areas. In agreement with the flow cytometric analyses, KKD68 chimeras contained a substantially increased frequency of λ-expressing cells in B cell follicles. As expected, many of these cells appear yellow in the immuno fluorescent image shown in Figure 4B (middle right panel), due to staining of GFP+ (green) cells with anti-λ-PE (red).

KKD68 transduced B cells that develop in chimeric mice via the L chain editing pathway contribute to peripheral B cell compartments in a manner similar to λ-expressing B cells in normal mice

We next determined if developing B cells that expressed the κ shRNA contributed normally to the peripheral B cell compartment and its subsets. The percentage of KKD68-transduced splenic B cells that expressed high levels of the complement receptor C1qRp, detected by the AA4.1 mAb was increased relative to controls (Figure 5, middle panels). While high C1qRp expression is associated with the transitional stages of B cell development (34, 35), the frequency of AA4.1+ cells among normal λ-expressing splenic B cells in B6 mice was found to be elevated relative to κ-expressing cells (Figure 5, left panels), as previously reported (34). Moreover, the frequency of λ-expressing B cells in the T1 and T2 stages of development is reduced, and increased, respectively in normal B6, control and KKD68 chimeras as compared to λ⁻ B cells in the same mice (Figure 5, left and right panels). This may be a manifestation of developmental selection of B cells with particular specificities predominantly conferred by λ L chains.

Figure 5. Similar phenotypes of λ-expressing splenic B cells in C57BL/6 and KKD68 BM chimeric mice.

Splenic B cells from B6 and BM chimeras generated with KKD68 or control vector retrovirus-transduced BM were stained with Abs specific for the indicated markers and evaluated by flow cytometry. IgM and CD23 expression levels were analyzed on AA4. 1⁺ B cells and gates were set to show percentages of T1 (IgM^highCD23⁻), T2 (IgM^highCD23⁺), and T3 (IgM^lowCD23⁺) transitional subsets. Upper and lower panels show data from λ⁺ and λ⁻ gated cells, respectively. Figures are representative of data obtained from at least two independent experiments.

This notion was reinforced by the results of analysis of mature B cell subsets in the spleens and peritoneal cavity (PC) of these mice. Analysis of levels of CD21 and CD23 expression indicated similar ratios of follicular (FO) and MZ B cells among λ-expressing splenic B cells in B6 mice, and λ-expressing GFP^high B cells in vector control and KKD68 chimeric mice. However, in all three of these subpopulations a greater percentage of B cells are sIgM^high, sIgD^low as compared to λ⁻ B cells (Supplemental Figure 2). An increased frequency of sIgM^high, λ+ B cells were also observed in the PC. In addition, in the PC a greater percentage of λ-expressing B cells in B6 mice, and GFP^high, λ⁺ B cells in both types of chimeric mice are CD23^low, sIgM^high, CD5⁺, that is, are members of the B1a subset (Supplemental Figure 3). This is consistent with previous observations that λ-expressing B cells are enriched in the B1a compartment (36). This biased development of λ⁺ B cells to a B1a phenotype took place in the control and KKD68 chimeras despite the fact that normal adult BM does not efficiently reconstitute this B cell subset (37).

To determine if λ V gene diversity in control and KKD68-transduced peripheral B cell populations were similar, rearranged Vλl and Vλ2 genes were RT-PCR amplified from FACS purified GFP+ splenic B cells from KKD68 and control chimeric mice, and the products cloned and sequenced. Among 23 Vλ gene control sequences six were unique, with a predominant Vλ1-Jλ1 representation (18 clones with three different junction sequences). Similarly, among 27 such Vλ genes sequenced from KKD68 chimeras, six were unique, with a predominance of Vλ1-Jλ1 genes (21 clones with three different junction sequences, Table 1).

Table 1.

	Vλ1+Jλ1	Vλ1+Jλ2	Vλ1+Jλ3	Vλ2+Jλ1	Vλ2+Jλ2	Vλ2+Jλ3
Vector chimera B cells	13 CAT TGG  4 CAC TGG  1 CAT TTT TGG	0	0	1 CAT TTG 1 CAT TGG	3 CAT TAT	0
KKD68 chimera B cells	12 CAT TGG  7 CAT TTG  2 CAC TGG	0	1 CAT TTC	0	3 CAT TAT 2 CAT TTT	0

The number of Vλ gene clones with each type of gene segment combination, and sequence at the Vλ-Jλ junction are indicated. Junctional sequences are shown with contributions from Vλ and Jλ segments separated. Nucleotides in these junction regions that are not encoded in germline Vλ or Jλ segments are shown in bold type.

Frequency and level of surface λ L chain expression among peripheral B cells correlates with level of KKD68 vector expression

The above observations suggested that the fate of KKD68-transduced immature B cells originally expressing κ L chains and that expressed high levels of the κ shRNA was either elimination, or rearrangement and expression of λ L chain genes resulting in expression of normal levels of λ-containing sIgM, allowing developmental progression. If recovery of normal levels of sIgM expression was the driving force behind this pathway, developing B cell subpopulations expressing lower levels of the κ shRNA might follow alternative pathways. To test this idea, mature B cells in the PBL and spleens of chimeric mice that expressed low, intermediate and high levels of GFP, as an indicator of general retroviral gene expression level, were analyzed for frequency of surface λ⁺ cells. Among GFP⁻ and the three subpopulations of GFP⁺ splenic B cells in control chimeras, all contained similarly low frequencies of surface λ⁺ cells. In contrast, these three GFP⁺ subpopulations in KKD68 chimeras all contained increased frequencies of surface λ⁺ B cells as compared to GFP⁻ B cells, with these frequencies correlating with level of GFP expression (Figure 6A). Analogous results were obtained from PBL (data not shown).

Figure 6. Frequency and level of λ light chain and both κ and λ light chain expression on splenic B cells in KKD68 BM chimeras is correlated with level of KKD68 retroviral gene expression.

A: Splenic B cells from BM chimeras created with KKD68 or control vector retrovirus-transduced BM were stained with Abs specific for the indicated markers and evaluated by flow cytometry. Left panel: B220⁺ B cells were divided into different subsets according to GFP expression levels (GFP⁻, GFP^low, GFP^int and GFP^high) and percentages of cells in each subset are shown next to the gates. Right panels: Surface κ and λ light chains levels were evaluated in these subsets by flow cytometry. Splenic B cells from nal̈ve C57BL/6 mice served as a control. Vector-1 and Vector-2 are two BM chimeras created with control vector retrovirus-transduced BM, and KKD68-1 and KKD68-2 are two BM chimeras generated with KKD68 retrovirus-transduced BM. The percentages of cells in each subset are shown next to the gates. Figures are representative of data obtained from at least three independent experiments. B: Hybridomas were generated from the mitogen activated splenic B cells of a KKD68 chimeric mouse and a vector control chimeric mouse and screened for GFP expression. GFP+ hybridomas were then assayed for expression of κ and λ L chains by flow cytometry. All hybridomas from the vector control mouse expressed only κ L chain (example in left panel). Of eight GFP+ hybridomas from the KKD68 chimeric mouse, six expressed reduced levels of κ L chain only (example in middle left panel), and two expressed both κ and λ L chains (two right panels).

As mentioned above, among the λ-expressing B cells in KKD68 chimeras two subpopulations were observed, one exclusively expressing λ L chains and one that appeared to express both κ and λ L chains. The former cells constituted the predominant subpopulation of all λ-expressing cells among GFP^high KKD68-transduced B cells (Figure 6A), and expressed levels of sIgM and sIgD similar to λ⁺ B cells in B6 mice (Supplemental Figure 2). The latter cells made up a greater percentage of all λ-expressing B cells in the GFP^low and intermediate compartments as compared to the GFP^high compartment. These data indicate that in cells expressing high levels of the κ shRNA, the major developmental pathway that is followed is receptor editing resulting in expression of levels of λ L chain compatible with normal levels of expression of sBCR. However, when levels of κ shRNA are lower, other pathways may result in acquisition of normal levels of sBCR and developmental progression, such as co-expression of κ and λ L chains. Expression of both κ alleles by a developing B cell may also represent one of these pathways. However, as mentioned above the fact that we did not detect an increased frequency of κ locus rearrangements among KKD68 transduced B220^low, κ^neg BM B cells (Supplemental Figure 1) suggests that if a cell embarks on this pathway it may be rapidly lost due to κ knockdown.

To more rigorously determine if a subpopulation of KKD68-transduced splenic B cells indeed expressed both κ and λ L chains (i.e. was isotypically included), splenic B cells from a KKD68 chimeric mouse were polyclonally activated in vitro and hybridomas generated. GFP-expressing hybridomas were then analyzed for expression of κ and λ L chains. Of eight GFP+ hybridomas analyzed, two were found to express κ and λ L chains (Figure 6B), a frequency in good agreement with the analysis of splenic B cells by flow cytometry (Figure 6A).

GFP^high KKD68-transduced B cells also contain a subpopulation that did not detectably express λ L chains, and expressed surface levels of κ L chain that appeared approximately three-fold reduced (MFI) as compared to controls. Nonetheless, this subpopulation was found to express levels of sIgM and sIgD similar to λ-expressing B cells (Supplemental Figure 2). The nature of this subpopulation is currently unclear. We observed a higher percentage of B cells that stained with a mAb specific for the usually infrequently expressed Vλx L chain in the KKD68 GFP^high B cell compartment as compared to this compartment in control mice (Supplemental Figure 4). However, this increase could not account for all of the κ^low, λ⁻, GFP^high B cells in KKD68 chimeras.

Lambda L chain, KKD68-expressing B cells can mount normal T cell dependent immune responses

To further examine the differentiative capabilities of λ-expressing, KKD68-transduced B cells we exploited the fact that the primary T cell dependent (TD) immune B cell response to 4-hydroxy-3-nitrophenylacetyl (NP) in B6 mice contains a predominant λ-expressing component (38). BM chimeric mice were immunized with NP-chicken gamma globulin and nine days later the frequency of λ-expressing B cells in the GFP^neg and GFP^high responding germinal center (GC) and antibody forming cell (AFC) compartments was evaluated. As expected, among both nontransduced and transduced B cells in control chimeras, and nontransduced cells in KKD68 chimeras λ⁺ cells represented a major fraction of both GC B cells, and AFCs. However, such cells made up over 90% of both the GC B cell and AFC subpopulations in the GFP^high compartment of KKD68 chimeras (Figure 7). Histological analysis of the spleens of immunized KKD68 chimeric mice corroborated these results, showing that GFP⁺, λ⁺ B cells were prevalent within GCs in follicles and AFC foci located in bridging channels (data not shown). These data demonstrate that B cells that have developed via the KKD68 shRNA-induced λ L chain editing pathway are fully capable of participating in a primary TD immune response.

Figure 7. Receptor-edited λ⁺ B cells in KKD68 BM chimeras mount a normal TD immune response.

KKD68 BM chimeric mice were immunized (i. p. ) with 100 μg NP¹⁶-CGG in alum. Mice were sacrificed on day 9 after immunization and splenocytes were stained with Abs specific for the indicated markers and evaluated by flow cytometry. Upper panels: The frequency of λ⁺ B cells was evaluated in transduced splenic antibody forming cells (AFCs, B220^lowCD138⁺GFP⁺) and non transduced AFCs (B220^lowCD138+GFP⁻). Lower panels: The frequency of λ⁺ B cells were evaluated in transduced splenic germinal center B cells (GC, B220⁺GL7⁺GFP⁺) and non transduced GC B cells (B220+GL7+GFP⁻). The percentages of cells in each fraction are shown next to the gates. Figures are representative of data obtained from at least two independent experiments.

DISCUSSION

Our data demonstrate that direct inhibition of BCR expression promotes L chain receptor editing in developing polyclonal B cell populations in vivo, and that B cells that emerge from this pathway contribute normally to peripheral subsets and are functional. Studies on 3-83Igi anti-MHC I Ig knockin mice first suggested that level of BCR expression influenced receptor editing (15, 16). In mice hemizygous for the BCR transgenes and lacking expression of the cognate MHC I allele, endogenous L chains, including λ were predominant expressed among peripheral B cells. In contrast, such cells represented a minor fraction of peripheral B cells in mice homozygous for the BCR transgenes. Recent studies in this system have shown that an activated form of N-Ras can rescue the development of 3-83Igi B cells expressing low levels of BCR (39). Caveats in these studies included Ig transgene-driven perturbation of development and lack of clonal competition that might alter the BCR signaling requirements for regulation of editing, and that the transgenic BCR is cross reactive with a non-MHC ligand. In addition, these studies utilized a BCR derived from the memory, not the nal̈ve B cell compartment.

We previously described anti-nuclear antigen Ig knockin transgenic mice termed HKIR in which developing transgenic BCR-expressing B cells do not undergo receptor editing, and develop to mature FO B cells, but express low levels of sBCR (40, 41). We proposed that these cells avoid the editing pathway by down regulating levels of sIgM early in development and thus, avidity for autoantigen. In support of this idea, in mice in which the transgenic HKIR H chain locus was homozygous, extensive L chain editing took place, resulting in emergence of a major population of B cells expressing BCRs with reduced autoreactivity encoded by the transgenic H chain and endogenous λ L chains (29). These data were interpreted to support the idea that the avidity of developing HKIR B cells for endogenous autoantigen(s) directly influenced levels of autoantigen-induced BCR signaling, with low levels of such signaling being compatible with developmental progression, and high levels inducing L chain editing.

Apparently disparate data like these have continued to fuel the debate on the pathways that regulate receptor editing. Behrens and colleagues have attempted to reconcile such data by proposing that the role of autoantigen, and BCR autoantigen avidity in regulating editing is via the stimulation of endocytic removal of BCR from the cell surface, thus reducing levels of basal BCR signaling. Data supporting this hypothesis were obtained using a system allowing inducible deletion of an Igh-targeted transgenic H chain, resulting in ablation of BCR expression on immature B cells in BM cultures. This led to increased Rag expression and an overall gene expression profile consistent with “back” differentiation to a more immature stage (22). Subsequent studies using conventional Ig transgenic B cells developing in in vitro BM cultures also supported this hypothesis (21). However, it is unlikely that these experimental systems recapitulate the immature B cell developmental pathway in normal mice, as BM culture conditions are optimized to promote growth of B cell precursors, likely promoting Rag expression and reducing interclonal competition. In addition, as discussed above, the constitutive expression of transgenic Ig chains starting very early in B cell development accelerates or even precludes subsequent steps in this pathway.

The experimental approach use in the studies reported here did not utilize BCR transgenic B cells and so is not subject to the caveats mentioned above. However, this approach requires that developing, immature B cells express an endogenous κ L chain gene to become susceptible to KKD68 shRNA-mediated reduction of BCR levels. While κ recombination and expression precedes or precludes λ recombination and expression in most developing mouse B cells (42), the former is not a prerequisite for the latter (24, 25, 43). In combination with the fact that many immature B cells expressing high levels of the KKD68 shRNA appear to be lost, this raises the possibility that preferential outgrowth of B cells that first rearranged and expressed the λ locus might explain our results.

We view this interpretation are unlikely for the following reasons. Among KKD68-transduced B cells expressing different amounts of GFP, but similar levels of surface IgM and IgD there was a clear correlation between the frequencies of λ-expressing cells, levels of surface λ and κ chain and of GFP expression. This suggests that there are at least two “developmental solutions” of how to acquire levels of BCR expression compatible with developmental progression in the face of different extents of κ RNA reduction. In one, substantial elimination of κ RNA precludes developmental progression unless L chain editing takes place and the resulting B cell expresses only λ L chain. If early rearrangement and expression of λ L chains followed by preferential expansion accounted for the presence of this compartment, then exclusive expression of λ should not correlate directly with high levels of KKD68 vector expression. In addition, such preferential expansion might be expected to result in an oligoclonal λ-expressing B cell compartment, yet we found that λ V gene diversity in KKD68-transduced and control B cell compartments was similar.

Secondly, lower levels of KKD68 vector expression results in L chain isotype inclusion that, by definition, requires that both κ and λ loci undergo productive rearrangement and are expressed. Indeed, we isolated hybridomas derived from KKD68-transduced splenic B cells that co-expressed κ and λ L chains. The presence of a third population of KKD68-transduced B cells with undetectable and reduced levels of surface λ and κ L chains, respectively, indicates that a third pathway of developmental compensation for reduced levels of κ expression may also be taking place, as suggested by our previous studies of the HKIR knockin mice (40, 41). However, it is also possible that these cells express BCRs with particular H: L pairs that reduce or preclude detection with the anti-L chain reagents used in our studies. Indeed, using a mAb that detects normally rarely used Vλx L chains (44), we found an increased frequency of Vλx⁺ cells among KKD68, but not control vector transduced B cells.

Since we do not know the specificities of BCRs initially expressed by the members of the polyclonal population of immature B cells that subsequently gave rise to the λ-expressing mature B cells observed in our studies, we cannot formally rule out that autoantigen-induced BCR signaling is necessary, but not sufficient for induction of L chain editing. However, given the diversity of BCR V regions in the normal repertoire, this scenario would require that an equally diverse collection of autoantigen specificities and avidities all would fulfill this requirement. Thus, our data support the conclusion that level of BCR expression is the dominant parameter influencing receptor editing at the immature B cell stage. Future studies will be required to determine if this reduced expression, as expected, results in reduced levels of basal BCR signaling, and whether B cells that regain normal levels of BCR expression via receptor editing resulting in expression of λ light chains regain normal levels of basal signaling.