The Transcription Factor Early Growth Response 1 (Egr-1) Advances Differentiation of Pre-B and Immature B Cells

Fetal liver cells of day 15–18 embryos were removed and plated onto irradiated ST-2 feeder cells in Iscove's medium containing IL-7 and 10% FCS. Cells were cultured as described previously (36). Cells from transgenic lines were identified by PCR. For further analyses, nonadherent cells were collected and washed twice in ice-cold PBS. Samples from wells containing only ST-2 feeder cells were treated in parallel and served as controls.

The detailed description of the generation of Egr-1 transgenic mice using the BALB/c embryonic stem cell line BALB/c-I will be described elsewhere. Egr-1 transgenic mice of the IA7 line were transferred to a special pathogen-free unit by implanting transgenic one-cell embryos into C57BL/6 foster animals kept under specific pathogen–free conditions. The IA7 line was then bred further by mating with wild-type BALB/c mice. RAG-2^−/− mice expressing transgenic Egr-1 were obtained by backcrossing female Egr-1 transgenic IA7 mice twice with C57BL/6 RAG-2^−/− males. Animals were tested by PCR for the Egr-1 transgene using genomic DNA, 5′-CTTTCGGTTTGGGGCTGGACA-3′ and 5′-CGCTGCTGGTGCTGCTGCTGCTAT-3′, as transgene-specific primer pair. The RAG-2^−/− phenotype was verified by FACS^® analysis of peripheral blood cells.

RNA was extracted using the guanidinium isothiocyanate method as described (37). For Northern blotting, 10 μg of total RNA was separated in a 1% agarose gel containing 7% formaldehyde, transferred onto nylon filters, and fixed by UV cross-linking. Filters were prehybridized (50% deionized formamide, 5× SSC, 5× Denhardt's solution, 50 mM NaH₂PO₄, pH 7, 10 mM Na₄P₂O₇, 0.1% SDS, 0.1 mg/ml denatured salmon sperm DNA) for 2 h at 42°C. For detection of Egr-1–specific transcripts, [α-³²P]dATP-labeled probes were prepared from a 1.6-kb EcoRI-HindIII fragment from plasmid 533 (a gift from V. Sukhatme) containing the Egr-1 cDNA by the oligonucleotide priming method (38). The probe was added to the prehybridization and filters were incubated overnight, washed with 0.2× SSC, 0.1% SDS at 42°C, and exposed to X-ray films. Egr-1 expression was analyzed by PCR using cDNA reverse transcribed from total RNA with Superscript II (GIBCO BRL, Eggenstein, Germany) and the Egr-1– specific primers 5′-GCAGATCTCTGACCCGTTCGG-3′ and 5′-CCGAGCGTTTGGCTRGGGATA-3′ as described by T. Miyazaki (35). PCR was performed using Taq polymerase (MBI Fermentas, Inc., Amherst, NY) using 1/25 of the cDNA reaction as template at an annealing temperature of 54°C.

Bone marrow cells from six femurs were isolated and resuspended in FACS buffer (0.1% sodium azide, 3% FCS in PBS). B220-specific biotin-labeled antibody RA3.3A1 (39) was added and incubated for 30 min on ice. Cells were washed, magnetic streptavidin-labeled beads (Dynal, Oslo, Norway) were added, and B cells were isolated. Quality of the sorting process was verified by flow cytometric analysis. The B cells were resuspended in 30 μl lysis buffer (1% NP-40, 150 mM NaCl, 10 mM Tris-HCl, pH 7.0, 0.1 mM PMSF) and incubated on ice for 10 min. Cell debris was removed by centrifugation (10 min, 4°C, 22,000 g), and the extract was separated by SDS-PAGE (8%) and transferred onto nitrocellulose membrane (Hybond C extra; Amersham Pharmacia Biotech, Uppsala, Sweden). Egr-1 was detected using the antiserum C19 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) at 25 ng/ml followed by peroxidase-conjugated goat anti–rabbit IgG F(ab′)₂ (Dianova GmbH, Hamburg, Germany) at 200 ng/ml. Expression of nur77 was analyzed using a mouse IgG anti-nur77 mAb (a gift of B. Osborne, University of Massachusetts, Amherst, MA) followed by peroxidase-conjugated goat anti–mouse IgG (Southern Biotechnology Associates, Inc., Birmingham, AL). IgM was detected by a goat anti–mouse IgM peroxidase-labeled serum (Southern Biotechnology Associates, Inc.). Signals were visualized using an enhanced chemiluminescence (ECL) detection system (Amersham Pharmacia Biotech).

Flow cytometry was carried out as described previously (40) using the following antibodies: RS3.1-biotin specific for murine IgM^a (41), 6C3-biotin for BP-1, 7D4-biotin for IL-2Rα chain, 2B8-biotin for c-kit, S7-biotin for leukosialin, AMS9.1-biotin for IgD^a, RA3-6B2–PE for B220, IM7-biotin for Pgp-1, 3E2-PE for intercellular adhesion molecule 1 (ICAM-1) (all from PharMingen Europe, Hamburg, Germany), and biotinylated PB493 (42) to stain immature B lymphocytes. Cells were counterstained using PE- or APC-conjugated streptavidin (PharMingen Europe). Unspecific binding to Fc receptors was blocked by adding unlabeled mouse FcγR-specific mAb 2.4G2. Dead cells were excluded by staining with propidium iodide. Using a FACSCalibur^® and CellQuest^® software (Becton Dickinson, San Jose, CA), 3–5 × 10⁴ cells were acquired according to their forward/side scatter pattern and analyzed. To analyze Egr-1 expression in bone marrow B cell subsets, 3.4 × 10⁵ pre-B and 7 × 10⁴ immature B cells were isolated from both femurs of a 5-wk-old BALB/c mouse by cell sorting at 4°C according to their IgM^a/ PB493 staining pattern using a FACStar^® cell sorter and LYSIS II^® software (Becton Dickinson).

Bone marrow cells were labeled with bromodeoxyuridine (BrdU; Sigma, Deisenhofen, Germany) starting with a single injection of 1 mg/ml i.p. BrdU and feeding mice continuously with drinking water containing 1 mg/ml BrdU for 48 h as described (43). During the labeling period, the drinking water was protected from light. Simultaneous detection of surface staining and BrdU labeling was done as described (44). After surface marker staining, cells were resuspended in 500 μl 0.15 M NaCl, 1.2 ml ice-cold 95% ethanol was added, and the cells were incubated 30 min on ice. Cells were washed and resuspended in 1 ml fixation buffer (1% paraformaldehyde and 0.01% Tween in PBS). After incubation for 30 min at room temperature, the cells were incubated for 30 min in DNase I solution (50 KU DNase I in 4.2 mM MgCl₂/ 0.15 M NaCl, pH 5). Cells were washed, 10 μl anti-BrdU antibody (Becton Dickinson) was added, and the cells were then incubated for 30 min and washed.

Gel shift was carried out using recombinant Egr-1 as described (45) with double-stranded radiolabeled oligonucleotides from the nur77 and BP-1 promoter regions carrying putative Egr-1 binding sites (bold): 5′-TTCCAAAGTTCCCCCTCAACCCCTC-3′ for BP-1 (position −753 to −729), 5′-GTCAGTGGCGCCCCCGCCCCTCTCCAA-3′ for nur77 (position −66 to −50), and 5′-GGATCCAGCGGGGGCGAGCGGGGGCG-3′ for Egr-1.

BCR cross-linking has previously been reported to induce Egr-1 expression in mature B cells, but not in immature B lymphocytes or in immature B cell lines (5, 46, 47). We addressed the question of whether unstimulated pre-B and immature B cells express Egr-1 by analyzing sorted B cell subsets from murine bone marrow. Transcription of the Egr-1 gene was found by PCR in both sIgM⁻ (pre-B) and sIgM⁺ PB493⁺ immature B cells (Fig. 1,A). Likewise, Egr-1 protein was detected by immunoblotting in sIgM⁻ pre-B cells isolated from fetal liver and expanded in culture on ST-2 stroma cells in the presence of IL-7 (Fig. 1 B). Both results show transcription of the Egr-1 gene and translation of Egr-1 mRNA into detectable amounts of protein as early as the pre-B cell stage before BCR surface expression.

These results suggested that Egr-1 function might also be important during early stages of B lymphopoiesis. To test this hypothesis, we generated transgenic mice expressing Egr-1 specifically in B lymphocytes using an Ig heavy chain promoter/enhancer construct. Four different founder mice showing Egr-1 germline transmission were obtained. By breeding to BALB/c mice, we established the Egr-1 transgenic lines IA7, IB10, IC4, and ID4 (to be published elsewhere). At first, we compared Egr-1 expression between transgenic and BALB/c control mice by Northern and immunoblotting (Fig. 2). Spleen cells of the line IA7 expressed 10-fold more Egr-1 mRNA than the control littermates, whereas the other lines showed Egr-1 expression levels of about two- to threefold above unstimulated spleen cells (Fig. 2,A). Since transgenic IC4 mice expressed only low levels of Egr-1 they were abandoned. Carrying out most of the experiments with mice from lines IA7, IB10, and ID4, we found only small variations between these transgenic lines. Testing Egr-1 protein expression, we found high levels in purified B220⁺ bone marrow B cells as well as in cultivated pre-B cells isolated from fetal liver (Fig. 2, B and C, respectively).

These results show that transgenic Egr-1 is expressed during similar stages of B cell maturation, but at far higher levels than endogenous Egr-1.

To examine whether enhanced Egr-1 expression has an effect on early stages of B cell development, we backcrossed the IA7 transgenic mice to a RAG-2–deficient background. The RAG-2 mutation prevents rearrangement of the Ig genes (32) and therefore blocks B cell maturation at the pro/pre-B–I cell stage (16, 34). These cells carry the surface markers c-kit and CD43; <5– 15% express BP-1 and <1% express the IL-2Rα chain (data not shown). Phenotypically, these pro/pre-B–I cells correspond to fraction B as classified by Hardy et al. (13). FACS^® analysis of control and Egr-1 transgenic RAG-2– deficient mice revealed an unchanged expression pattern for c-kit and CD43, but a three- to fourfold increase in the fraction of BP-1⁺ cells, compared with control littermates (Fig. 3). Phenotypically, these BP-1⁺ c-kit⁺ CD43⁺ pre-B lymphocytes are defined as fraction C cells (13), and progression into this stage normally requires RAG-2 expression and Ig heavy chain gene rearrangement (16, 34), suggesting that Egr-1 might support the maturation of fraction B pre-B cells even in the absence of RAG-2 activity. The reduced cell size of transgenic BP-1⁺ B220⁺ lymphocytes as reflected by changes in forward/side scatter (Fig. 3 A, c and d) also supports the Egr-1–induced progression of B cell development. Since we did not detect cell surface markers characteristic for pre-B–II/fraction D cells, i.e., IL-2Rα and high levels of heat-stable antigen, the expression of the Egr-1 transgene seems to promote the transition of pre-B–I cells from fraction B to fraction C but not further.

These results indicated that Egr-1 expression might also change maturation of later stage bone marrow B cell subsets. According to the expression pattern of B220 and IgM, three main subsets corresponding to three consecutive stages of maturation can be distinguished: IgM⁻ B220^low pre-B cells, B220^low IgM⁺ IgD⁻ immature B cells, and finally the B220^high IgM⁺ IgD⁺ mature B cells (16, 48). The proportion of mature B cells in the bone marrow increases with the age of the animal. Therefore, mice at 4 wk of age have fewer mature B cells than older animals. Comparing 23 control littermates with 23 IA7 transgenic mice between 4 and 40 wk of age, we found in general fewer immature and more mature B cells in the IA7 bone marrow than in age-matched controls (Fig. 4). These results suggest that Egr-1 also promotes the differentiation of immature B cells into mature B cells.

The enlarged population of BP-1⁺ pre-B cells in transgenic RAG-2–deficient mice and the higher percentage of mature cells in transgenic RAG-2⁺ animals could result from an increased proliferative potential of B cell precursors. Because dividing B lymphocytes can be traced by incorporation of the nucleotide analogue BrdU (49), we compared the proportion of BrdU⁺ cells in the pre-B, immature B, and mature B cell subsets between control littermates and transgenic mice after 2 d of in vivo labeling. The similar percentages of BrdU⁺ cells in all B cell subsets of both groups of mice suggest that the expression of transgenic Egr-1 does not change the proliferative potential of B cells (Fig. 5).

The increased maturation of transgenic B cells would be expected to correlate with changes in the expression of genes controlled directly or indirectly by Egr-1. Using a panel of antibodies, we screened transgenic bone marrow B cells for alterations in the expression pattern of B cell–specific surface markers by flow cytometry. Changes were not observed except for a slight increase in the percentage of BP-1⁺ cells and in BP-1 expression levels in all transgenic lines (Fig. 6,A). As discussed above, the proportion of BP-1⁺ cells might relate to the Egr-1–enhanced transition of fraction B pre-B cells into the fraction C stage, but it would not explain higher BP-1 surface levels. Therefore, we speculated that Egr-1 might regulate the transcription of the BP-1 gene and screened the BP-1 promoter for potential Egr-1 binding sites. Finding a (5′-GAGGGGGAA) sequence ∼1.6 kb upstream of the mRNA start (50) resembling an Egr-1 binding site (5′-GCGGGGGCG), we analyzed by an electrophoretic mobility shift assay (EMSA) if recombinant Egr-1 binds to an oligonucleotide containing the putative Egr-1 recognition site from the BP-1 promoter. As shown in Fig. 6 C, labeled oligonucleotides containing a cognate Egr-1 binding site (lane 1) or the binding site from the BP-1 promoter (lane 15) produced a shifted DNA–protein complex with identical electrophoretic mobility. Their intensities were reduced only by adding an excess of unlabeled oligonucleotides with an Egr-1 binding site but not by competing with an Sp-1 binding site (lanes 2–5 and 16–19, respectively). Likewise, only the addition of Egr-1– but not of Sp-1–specific antibodies retarded the migration of the complex (lanes 6 and 7, 20 and 21). Therefore, the forced expression of Egr-1 in transgenic B cells may not only help pre-B cells to proceed from fraction B into fraction C but may also enhance the expression of BP-1, which is normally upregulated during this transition.

The nur77 gene (also called NGFI-B [3] or N10 [51]) encoding an orphan nuclear receptor represents one of the transcription factors found to be induced by Egr-1 (52). Analyzing nur77 expression in purified B220⁺ bone marrow cells of transgenic (ID4) and control littermates by Western blotting, we found nur77 to be expressed only by transgenic but not by control B cells (Fig. 6,B). Since upregulated nur77 expression was not found in cultivated transgenic sIg⁻ pre-B cells (data not shown), the induction of the nur77 gene seems to be confined to pre-B–II or sIgM⁺ bone marrow cells. To analyze if Egr-1 could directly induce nur77 transcription, we tested by EMSA the binding of recombinant Egr-1 to an oligonucleotide of the nur77 promoter (position −74 to −50) containing a cognate Egr-1 binding site. As shown in Fig. 6 C, binding of recombinant Egr-1 produced a DNA–protein complex (lane 8). The specificity of Egr-1 binding was demonstrated by a competition assay using an excess of an oligonucleotide carrying an Egr-1 binding site (lanes 11 and 12) and by the decreased mobility of the complex upon the addition of an Egr-1–specific antibody (lane 13).

Since it was shown recently that nur77 activity is involved in the induction of apoptosis during negative selection of thymocytes (53–56) and in the upregulation of CD95L expression (57), we looked for enhanced CD95L expression in Egr-1 transgenic mice compared with control littermates. In contrast to the results reported for nur77– expressing T cells, we did not find increased CD95L expression in Egr-1 transgenic B cells (data not shown).

In response to BCR-derived signals, Egr-1 is thought to modulate the expression pattern of downstream genes that promote further activation and differentiation of B lymphocytes. Using variants of the B cell line WEHI-231, it was shown that Egr-1 induces the expression of CD44 and of intracellular adhesion molecule 1 (ICAM-1 [58, 59]). Therefore, we tested the expression pattern of both surface markers in our transgenic mice, but did not detect differences when compared with BALB/c controls (data not shown).

Mature B cells respond to signals resulting from antigen receptor engagement by immediately inducing Egr-1 transcription (5), but the role of Egr-1 in earlier stages of B cell development has not been defined. The different stages and the order of B cell development are well characterized, allowing the precise typing of bone marrow B cells according to the expression of characteristic cell surface markers, the rearrangement of Ig genes, and the proliferative and differentiation potential of B cell precursors (13, 14, 16, 60, 61). By analyzing Egr-1 expression in bone marrow–derived B lymphocyte subsets and by testing Egr-1 expression in cultivated, fetal liver–derived pre-B cells, we have shown that Egr-1 is also expressed in pre-B cells lacking sIgM as well as in immature sIgM⁺ B cells in the absence of sIgM-induced signals. These observations suggest that Egr-1 might also have a regulatory function in pre-B cell development. By studying transgenic mice overexpressing Egr-1 from the pre-B stage on, we have found higher proportions of mature B cells and fewer immature B cells in transgenic animals than in control littermates. To identify if early stages of B lymphopoiesis are sensitive to Egr-1 activity, we arrested B cell development at the stage of pro/pre-B–I cells by backcrossing the Egr-1 transgenic line IA7 to mice deficient in RAG-2. Since the null mutation in the RAG-2 gene prevents rearrangement of Ig genes (32), B cell precursors do not receive stimulating signals required for developmental progression beyond the stage of B220^low CD43⁺ BP-1⁻ pro/pre-B cells (16, 62), also defined as fraction B (13). Comparing the phenotype of bone marrow pro/pre-B cells from transgenic and control mice, we found a three- to fourfold increased population of BP-1⁺ pre-B cells in Egr-1 transgenic mice. Since the transcription activation function of Egr-1 seems to enhance BP-1 expression in more mature B cell subsets, the increase in BP-1⁺ pre-B cells could also reflect the induction of BP-1 expression only and not Egr-1–induced differentiation. However, this seems to be less likely because transgenic BP-1⁺ cells were found to be smaller than BP-1⁻ cells, consistent with further maturation. Therefore, these results suggest that forced expression of Egr-1 in BP-1⁻ pro/pre-B cells induces progression into the stage of BP-1⁺ pre-B cells (fraction C). Since these cells failed to upregulate the IL-2Rα chain and heat-stable antigen, two markers characteristic for pre-B–II cells (fraction C′ and D [13, 16]), overexpression of Egr-1 in pro/pre-B–I cells seems to be sufficient to induce differentiation to fraction C, but not to more mature stages of B lymphopoiesis.

Progression of pro/pre-B cells developmentally arrested by a mutation in the RAG-2 gene into more mature pre-B cell stages is also induced by in vivo cross-linking of the Ig-α/Ig-β heterodimer using Ig-β–specific mAbs (63). Under those conditions, anti–Ig-β–treated pro/pre-B–I cells become smaller in size and acquire IL-2Rα expression in addition to BP-1. Since they also downregulate c-kit (CD117) and CD43, they are considered as small pre-B–II cells. In the same report, it was shown that Ig-β cross-linking stimulates tyrosine phosphorylation of several substrate proteins, including Ig-α, Syk, and Vav, and the activation of mitogen-activated protein kinase extracellular signal– regulated kinase (ERK)1. Based on these results, Nagata et al. (63) proposed that the signal cascade initiated by Ig-β activation evokes differentiation signals similar to those delivered by the pre-BCR in normal B cell development. For mature B cells it is known that BCR engagement upregulates Egr-1 transcription through a signal cascade including p21/ras and mitogen-activated protein kinase (ERK [10, 64]), and for other cell types it has been shown that ERK activation induces Egr-1 transcription (65). Since RAG-2– deficient pro/pre-B–I cells overexpressing Egr-1 do not reach the same developmental stage as anti–Ig-β–stimulated cells, it seems likely that Egr-1 activity substitutes only part of the differentiation signal originating from the pre-BCR.

Analyzing later stages of B cell development in RAG-2^+/+ Egr-1 transgenic mice, we observed lower proportions of immature and increased proportions of mature bone marrow B cells compared with their wild-type littermates, whereas there was no increased proliferation of transgenic pre-B or immature B cells detectable. These findings are consistent with the current model of the development from immature to mature B cells (66, 67). Immature B cells leave the bone marrow and enter the spleen where about half of them reach the mature stage (42). Mature bone marrow B cells are thought to be part of the recirculating pool. This would suggest that Egr-1 influences this migration at one or several steps.

Egr-1 expression was also found in CD4⁻CD8⁻ double negative thymocytes by Miyazaki (35). Overexpression of transgenic Egr-1 in a RAG-2–deficient background allowed thymocytes to bypass the RAG-2–dependent block at the IL-2R⁺ Pgp-1⁻ double negative stage and develop into immature CD8 single-positive cells, but not further to the CD4⁺CD8⁺ double-positive cell stage. In cortical CD4⁺CD8⁺ thymocytes, Egr-1 expression was reported by Shao et al. (68) to be dependent on TCR engagement, suggesting that high level expression of Egr-1 in the thymus might be a consequence of thymocyte selection. The high coincidence of Egr-1 expression in analogous B and T cell precursor subsets and the increased differentiation of pro/pre-B–I cells and thymocytes in Egr-1 transgenic mice suggest that Egr-1 activity regulates similar functions in both types of lymphocytes.

Searching for potential downstream target genes responding to Egr-1, we found increased expression of the nuclear orphan receptor nur77 in bone marrow B cells from transgenic mice but not in cultivated transgenic pro/pre-B–I cells. Since we also could demonstrate binding of recombinant Egr-1 to a cognate Egr-1 binding site present in the nur77 promoter, it seems likely that Egr-1 directly induces nur77 expression in B lymphocytes before the mature B cell stage. It was reported that nur77 activity is involved in the regulation of thymocyte apoptosis (53, 54) by inducing CD95L expression (57). However, in the Egr-1 transgenic mice, we could detect neither upregulation of CD95L expression by bone marrow B cells nor an increased frequency of apoptotic cells (Warnatz, K., unpublished results). On the other hand, cellular responses other than apoptosis may be linked to nur77 function, since it is also induced upon antigen receptor ligation in B and T cells during proliferative responses (52, 69). Besides nur77 promoter, we also found enhanced expression of BP-1 in pre-B–II and in IgM⁺ bone marrow cells. Similar to the nur77, we could also demonstrate binding of recombinant Egr-1 to a sequence from the BP-1 promoter resembling an Egr-1 binding site. Therefore, it seems that Egr-1 activity promotes not only development to the stage of BP-1⁺ cells (fraction C), but also an increased surface expression of BP-1 on BP-1⁺ B cells. Although it is known that BP-1 acts as an aminopeptidase catalyzing the hydrolysis of acidic amino acid residues from the NH₂ termini of proteins (70), its role in B lymphopoiesis remains to be clarified (71, 72).

Here we provide evidence that Egr-1 supports at least two distinct steps of B cell maturation, the progression into the pre-B and into the mature B cell stage. Since Egr-1 activity is also sufficient to promote the development of double negative thymocytes into immature single-positive CD8^low cells (35), as well as macrophage in vitro differentiation (73, 74), this transcription factor seems to play an important role in the differentiation of three major hematopoietic cell types.

We are grateful to Petra Fiedler for excellent technical assistance. We thank Barbara Osborne (University of Massachusetts, Amherst, MA) for generously providing nur77-specific antibodies and probes, Toru Miyazaki (Basel Institute for Immunology) for helpful discussions and for communicating unpublished results, and Harald Illges (University of Konstanz, Germany) and Mary O'Riordan (University of California, San Francisco, CA) for critically reading the manuscript.

BCR

B cell antigen receptor

BrdU

bromodeoxyuridine

Egr-1

early growth response 1

EMSA

electrophoretic mobility shift assay

ERK

extracellular signal–regulated kinase

RAG

recombinase-activating gene