Fli-1 belongs to the Ets transcription factor family and is expressed primarily in hematopoietic cells, including most cells active in immunity. To assess the role of Fli-1 in lymphocyte development in vivo, we generated mice that express a truncated Fli-1 protein, lacking the C-terminal transcriptional activation domain (Fli-1ΔCTA). Fli-1ΔCTA/Fli-1ΔCTA mice had significantly fewer splenic follicular B cells, and an increased number of transitional and marginal zone B cells, compared with wild-type controls. Bone marrow reconstitution studies demonstrated that this phenotype is the result of lymphocyte intrinsic effects. Expression of Igα and other genes implicated in B cell development, including Pax-5, E2A, and Egr-1, are reduced, while Id1 and Id2 are increased in Fli-1ΔCTA/Fli-1ΔCTA mice. Proliferation of B cells from Fli-1ΔCTA/Fli-1ΔCTA mice was diminished, although intracellular Ca2+ flux in B cells from Fli-1ΔCTA/Fli-1ΔCTA mice was similar to that of wild-type controls after anti-IgM stimulation. Immune responses and in vitro class switch recombination were also altered in Fli-1ΔCTA/Fli-1ΔCTA mice. Thus, Fli-1 modulates B cell development both centrally and peripherally, resulting in a significant impact on the in vivo immune response.
In mice, B cells are generated in the bone marrow after birth. Their stages of development are divided into pro-B, pre-B, and immature B cells by the sequential expression of cell surface markers and the ordered rearrangement of Ig H and L chain segments (1, 2). Progenitor B cells undergo rearrangement of the Ig H chain locus to become pro-B cells. Subsequently, expressed mu H chains associate with surrogate L chains and the signaling molecules Igα and Igβ, forming the pre-BCR complex (2). The pre-BCR complex plays a critical role in the clonal expansion of mu+ pro-B cells and differentiation to the pre-B cell stage (3, 4). Pre-B cells undergo Ig L chain rearrangement and the resultant L chains associate with the mu H chain to form membrane-bound IgM (3). After successful rearrangement of both H and L chain genes, the BCR, which includes Igα and Igβ, is formed and expressed on the surface of B cells. The BCR functions as an Ag-binding and signal transduction molecule during further B cell development (5).
Immature B cells emigrate from the bone marrow to the spleen, and there progress through transitional states. Transitional B cells enter splenic follicles, further differentiating to follicular (FO)3 B cells, which comprise the majority of the mature splenic B cell population and marginal zone (MZ) B cells (5, 6). The differentiation and maintenance of these mature B cells require signals derived from the BCR (7). Maintenance of transitional, FO, and MZ B cell subsets are controlled by their microenvironment and interactions with other cell types (5, 8). In mice, MZ B cells, together with metallophilic and marginal-sinus-associated macrophages, initiate a rapid first line of defense against blood-borne particulate Ags. MZ B cell responses were first thought to be primarily directed against T-independent Ags; however, recent reports indicate that MZ B cells play roles in both T-independent and -dependent immune responses. FO cells participate primarily in T-dependent Ab responses (8, 9). Progression through B cell development depends on the combined effects of multiple regulatory genes, including Pax-5, early B cell factor (EBF), the E2A factors E12 and E47, and the Ets transcription factor, PU.1 (10). Members of the Ets gene family are found in genomes of diverse organisms, including Drosophila, Xenopus, sea urchin, chicken, mouse, and human (11, 12). The transcription factor Fli-1 is a member of the Ets family and is expressed in hematopoietic cells and a variety of other tissues (13, 14). Like all members of the Ets gene family, Fli-1 has the conserved DNA-binding domain, the Ets domain. The Fli-1 gene encodes two protein products, the larger comprised of 452 aa (14, 15). The Ets (DNA-binding) domain of Fli-1 is located between amino acids 277 and 361. Deletion analysis has identified two transcriptional regulatory domains, designated ATA and CTA, for N-terminal transcriptional and C-terminal transcriptional activation domains, respectively (16). Ets proteins bind to DNA sequences that contain a consensus GGA(A/T) core motif (Ets-binding site) and can function as either transcriptional activators or repressors (11, 12). Ets proteins regulate the expression of genes that are critical for the control of cellular proliferation, differentiation, and programmed cell death. Previous studies indicate that Fli-1 plays an important role in the regulation of megakaryocyte development (17, 18, 19, 20).
Studies regarding the role of Fli-1 in immune function or dysfunction are limited. Several reports indicate that Fli-1 can bind specific Ets-binding sites and transcriptionally activate a number of genes including those encoding Egr-1, multiple megakaryocytic genes, bcl-2, mb-1, and hTERT (21, 22, 23, 24, 25). Fli-1 can also act as a repressor of Rb and collagen 1 expression (26, 27). Targeted constitutive disruption of the Fli-1 gene resulted in hemorrhage into the neural tube and embryonic death due in part to thrombocytopenia (28, 29). Two-fold overexpression of the Fli-1 protein in transgenic mice resulted in the development of a lupus-like disease, including progressive immune complex-mediated renal disease and ultimately premature death from renal failure. Hypergammaglobulinemia, splenomegaly, B cell peripheral lymphocytosis, and autoantibody production (antinuclear Ab and anti-dsDNA) were prominent in these transgenic mice (30). We previously reported that reduced expression of the Fli-1 protein in MRL/lpr mice, a murine model of lupus, significantly increased survival and decreased renal disease compared with wild-type littermates (31). MRL/lpr mice with reduced Fli-1 expression also had decreased total serum Ig and anti-dsDNA Abs. Taken together, these studies suggested that Fli-1 plays an important role in the immune system, the B cell being an apparent target.
To assess the role of Fli-1 in lymphocyte development in vivo, we generated C57BL/6 mice that expressed a truncated Fli-1 protein (amino acids 1–384) lacking the CTA domain (Fli-1ΔCTA). Initial examination of Fli-1ΔCTA/Fli-1ΔCTA revealed a statistically significant reduction in total peripheral blood B220+ B cells. Examination of spleens from Fli-1ΔCTA/Fli-1ΔCTA mice demonstrated that the percentage of FO B cells was significantly decreased, whereas transitional B cells and MZ B cells were significantly increased. Furthermore, pre-B cells were also significantly decreased in the bone marrow of Fli-1ΔCTA/Fli-1ΔCTA mice. Thus, Fli-1 appears to be a negative regulator of MZ B cell development and a positive regulator of FO B cell development, possibly via, among other mechanisms, regulating surface expression of Igα. In addition, altered expression of E2A and Id mRNAs may also contribute to the observed increase in MZ B to FO B cell ratio. These findings also highlight a unique in vivo function for the CTA domain of Fli-1 in the regulation of B cell development and function.
Materials and Methods
Generation of the Fli-1 allele (Fli-1ΔCTA) that encodes truncated Fli-1 protein (amino acids 1–384) mice has been described in detail (O. Moussa, A. LaRue, R. Abangan, Jr., X. Zhang, M. Masuya, Y. Gong, D. Spyropoulos, M. Ogawa, G. Gilkeson, and D. Watson, manuscript in preparation). The mice were backcrossed with C57BL/6 mice for at least eight generations and then used in this study. Ly5.1 mice (B6.SJL-Ptprca Pep3b/BoyJ) were purchased from The Jackson Laboratory for the bone marrow transplantation study. All mice were maintained in the specific-pathogen-free animal facilities of the Ralph H. Johnson Veterans Affairs Medical Center, and all animal procedures were approved by the institutional animal care and use committee.
Splenocytes, thymocytes, or purified B cells with negative selection kits (Miltenyi Biotec) from spleen cells were lysed by radioimmunoprecipitation assay buffer. Lysates were analyzed by 10% SDS-PAGE and transferred to a polyvinylidene fluoride membrane. The membrane was incubated with specific Fli-1 Ab and revealed by horseradish-conjugated goat-anti-rabbit Abs, following by the ECL detection system.
ELISA were used to detect the Ig concentrations from mice as described previously (31).
Chromatin immunoprecipitation (ChIP) assays
Total bone marrow (4 × 106 cells) was isolated from wild-type and Fli-1ΔCTA/Fli-1ΔCTA mice. A ChIP assay was performed as described previously using anti-Fli1 rabbit polyclonal Ab (21). The −160 to +34 region of the mouse mb-1 promoter was amplified by real-time PCR using primers 5′-GGCAGGGCTCTGAGGGCTTCT-3′ and 5′-TCTCCACTCAGAGCCCACACA-3′, specific for the proximal promoter region. Real-time PCR was conducted using a LightCycler (Roche) with the Platinum SYBR Green qPCR SuperMix UDG (Invitrogen) according to the manufacturer’s instructions. PCR primers were used at a concentration of 250 nM. The cycling conditions for mb-1 were: preincubation at 50°C for 2 min, 95°C for 2 min, followed by 55 cycles of denaturation at 94°C for 10 s, annealing 57°C for 10 s, and extension at 72°C for 25 s, with a single data acquisition at the end of each extension. All ramping was done at 20°C per second. Relative expression analysis was conducted using the program LinRegPCR according to the suggested specifications.
Flow cytometry analysis and cell sorting
Single-cell suspensions were prepared from spleen, bone marrow, or thymus from mice at the age of 6–12 wk. The cells were stained with fluorochrome- or biotin-conjugated Abs and analyzed on a FACSCalibur flow cytometer. Data were analyzed using CellQuest (BD Immunocytometry Systems) software. Transitional, FO, or MZ B cells were sorted by the MoFlo High-Performance Cell Sorter (DakoCytomation) after staining with Abs and used for RNA preparation in real-time PCR. All Abs were purchased from BD Pharmingen.
Immunization and determination of Ig titers
Four 8-wk-old mice were immunized by i.p. injection of 50 μg of 2,4,6-trinitrophenyl (TNP) Ficoll or 50 μg of TNP-keyhole limpet hemocyanin (KLH; Biosearch Technologies) mixed with CFA. The mice immunized with TNP-KLH were given 50 μg of TNP-KLH mixed with IFA booster 1 wk after first immunization. Anti-TNP Abs were determined by ELISA with TNP-BSA as described previously (32).
B cell purification, culture, and stimulation
B cells were purified with B cell-negative selection kits from Invitrogen. The purity of B cells (over 90%) was confirmed by flow cytometry analysis. For CD23 induction, B cells were cultured in RPMI 1640 medium with 10% FBS, 2 mM l-glutamine, and 50 μM 2-ME at the concentration of 106 cells/ml. Cells were stimulated with 50 ng/ml IL-4 (R&D Systems) and expression of CD23 was measured after 24 h by flow cytometry after staining with anti-CD23-PE Abs. For in vitro class switch assays, B cells were cultured at 106 cells/ml with 50 μg of LPS (Sigma-Aldrich) and recombinant murine IL-4 at 50 ng/ml to induce switching to IgG1, with LPS alone to induce switching to IgG3 and IgG2b, or with LPS and IFN-γ at 10 ng/ml (eBioscience) to induce switching IgG2a. The percentage of IgG1, IgG3, IgG2a, and IgG2b expression was measured by flow cytometric analysis after 4 days of stimulation and staining with FITC-conjugated anti-mouse IgG1, IgG3, IgG2a, or IgG2b and PE-conjugated anti-B220.
In vivo proliferation assays
A total of 105 purified B cells were dispensed in a 96-well plate at 100 μl/well and cultured with RPMI 1640 medium with 10% FBS, 2 mM l-glutamine, and 50 μM 2-ME. Cells were stimulated with 20 or 40 μg/ml anti-mouse IgM (goat F(ab′)2; Jackson ImmunoResearch Laboratories) for 72 h, and [3H]thymidine (1 μCi/well) was added to each well 24 h before harvest.
Ca2+ mobilization assay
B cells were purified as described above. B cells were incubated at 30°C for 30 min with Fluo-3 in RPMI 1640 with 10% FBS. Cells were washed before stimulation. Cells were warmed at 30°C for 5 min before stimulation and the basal Ca2+ concentration was measured for ∼30 s. Stimulation was performed with anti-mouse IgM F(ab′)2 (20 μg/ml; Jackson ImmunoResearch Laboratories) and the events were recorded for 220 s (33).
Immunofluorescent staining of spleen
Spleens were removed from the mice and frozen in liquid nitrogen. A section of spleen was cut at a thickness of 5 μm. After air-drying, tissue sections were fixed in 4% paraformaldehyde for 20 min. Immunohistochemical staining was done with FITC-conjugated rat anti-mouse metallophilic macrophage (MOMA-1; monoclonal; MCA947F; Serotec) and biotin-conjugated goat anti-mouse IgM (Southern Biotechnology Associates) after blocking with 10% rat serum for 1 h. Next, the sections were stained with Texas-Red streptavidin (Vector Laboratories). After being washed, sections were examined with an Olympus Fluoview IX70 confocal laser scanning microscope.
Bone marrow transplantation
Bone marrow cells were isolated from the femurs of wild-type or Fli-1ΔCTA/Fli-1ΔCTA B/6 mice and were injected i.v. into sublethally irradiated recipient Ly5.1 mice (B6.SJL-Ptprca Pep3b/BoyJ) mice (600 Gy). The mice were sacrificed 12 wk later and splenic B cells were examined.
Total RNA was prepared from bone marrow, splenic cells, or purified B cells for real-time PCR analysis. A total of 2 μg of RNA was used to synthesize cDNA (SuperScript First-Strand Synthesis System; Invitrogen). Real-time PCR was performed in duplicate using Platinum SYBR Green qPCR SuperMix UDG (Invitrogen) according to the manufacturer’s instructions, with three independent RNA preparations. PCR was done using the LightCycler (Roche) and relative expression analysis was conducted using the program LinRegPCR according to the suggested specifications (34). The cycling conditions for all genes were: preincubation at 50°C for 2 min, 95°C for 2 min, followed by 30–50 cycles of denaturation at 94°C for 5 s, annealing at 1° below the lowest temperature for a given primer pair (Table I) for 5 s and extension for 45 s at 72°C, with a single data acquisition at the end of each extension. All ramping was done at 20°C/s.
|Primer Name .||Forward Primer .||Reverse Primer .||Amplicon Length .||Accession No. .||Taa (°C) .|
|Primer Name .||Forward Primer .||Reverse Primer .||Amplicon Length .||Accession No. .||Taa (°C) .|
Ta, Annealing temperature.
Quantitative data are expressed as mean ± SD. Statistical analysis was performed using the two-tailed Student t test.
Generation of Fli-1ΔCTA/Fli-1ΔCTA mice
Because complete disruption of the Fli-1 gene is embryonic lethal, we generated a new Fli-1 allele (Fli-1ΔCTA) that encodes a truncated Fli-1 protein (amino acids 1–384) lacking the CTA domain and used homozygous mutant mice to evaluate the role of Fli-1 in normal lymphocyte development. Removal of this domain of Fli-1 reduces overall in vitro transcriptional activation activity by 40–50% (16). We backcrossed Fli-1ΔCTA/Fli-1+ mice (B6 × 129 heterozygotes) with C57BL/6 (B6) mice for eight generations. Fli-1ΔCTA/Fli-1+ B6 mice were mated to generate two groups of mice, Fli-1ΔCTA/Fli-1ΔCTA and wild-type (Fli-1+/Fli-1+) littermate controls for this study. A stable truncated Fli-1 protein was expressed in splenocytes, thymocytes, and isolated splenic B cells from Fli-1ΔCTA/Fli-1ΔCTA mice (Fig. 1,A). Peripheral blood analysis demonstrated that Fli-1ΔCTA/Fli-1ΔCTA mice had a significant decrease in B220+ B cells (Table II). The difference in B cell number was dose-dependent, as Fli-1ΔCTA/Fli-1+ mice had fewer B cells than wild-type controls, but more than Fli-1ΔCTA/Fli-1ΔCTA mice (Table II).
|.||Wild Type (n = 20) .||Fli-1ΔCTA/+ (n = 7) .||Fli-1ΔCTA/Fli-1ΔCTA (n = 20) .|
|B220+||52.36 ± 4.82||41.40 ± 5.99bc||32.58 ± 8.19b|
|.||Wild Type (n = 20) .||Fli-1ΔCTA/+ (n = 7) .||Fli-1ΔCTA/Fli-1ΔCTA (n = 20) .|
|B220+||52.36 ± 4.82||41.40 ± 5.99bc||32.58 ± 8.19b|
The percentages of B220+ cells were assessed by flow cytometry. Values represent mean ± SD.
Value of p < 0.0001 when compared to wild type.
Value of p < 0.05 when compared to Fli-1ΔCTA/Fli-1ΔCTA homozygous mice.
The IgG1 and IgG3 concentrations in serum were decreased >50% in Fli-1ΔCTA/Fli-1ΔCTA mice compared with controls (for IgG1, wild type, 408.6 ± 75.0 μg/ml vs Fli-1ΔCTA/Fli-1ΔCTA mice, 188.3 ± 41.2 μg/ml, n = 14–15 in each group, p = 0.014; for IgG3, wild type, 97.2 ± 22.6 μg/ml vs Fli-1ΔCTA/Fli-1ΔCTA mice, 44.3 ± 12.2 μg/ml, n = 14–15 in each group, p = 0.045, Fig. 1,B). Although the concentrations of serum IgM, IgG2a, and IgG2b in Fli-1ΔCTA/Fli-1ΔCTA mice were slightly lower than in wild-type mice, the difference was not statistically significant (Fig. 1 B).
Fli-1 deficiency affects pre-B cell development in bone marrow
To further characterize the impact of Fli-1 CTA deficiency on B cell development, bone marrow cells were isolated from the femurs of wild-type and Fli-1ΔCTA/Fli-1ΔCTA mice and analyzed by flow cytometry. Pre-B cells and immature B cells (B220lowCD43−) were significantly reduced in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type mice as shown in Fig. 2 (wild type, 17.6 ± 2.0% vs Fli-1ΔCTA/Fli-1ΔCTA mice, 10.3 ± 1.1%, n = 5 in each group, p = 0.02). B220lowCD43− cells were further subdivided into fractions D–F based on surface expression of IgM and IgD (35). Populations of fraction D (B220lowIgM−IgD−), fraction E (B220lowIgM+IgD−), and fraction F (B220lowIgM+IgD+) were all significantly lower in Fli-1ΔCTA/Fli-1ΔCTA mice than in wild-type controls (Fig. 2). There was no statistically significant difference in the number of pro-B cells (B220lowCD43+) from Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type mice (Fig. 2).
Fli-1 deficiency impacts FO and MZ B cell development
Next, we examined whether Fli-1 deficiency affected splenic B cell populations. The total number of spleen cells was significantly higher in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls (wild type, 74.8 × 106 ± 10.4 × 106 cells, vs Fli-1ΔCTA/Fli-1ΔCTA, 105.7 × 106 ± 4.9 × 106, n = 8 in each group, p = 0.035, Fig. 3,A) though spleen weights were similar (data not shown). The percentage of CD3+ T cells in spleens from Fli-1ΔCTA/Fli-1ΔCTA mice tended to be higher compared with wild-type control mice, but did not reach statistical significance (wild type, 28.1 ± 0.9% vs Fli-1ΔCTA/Fli-1ΔCTA, 32.3 ± 2.7%, n = 8, p = 0.18). The percentage of splenic B cells, however, was significantly decreased in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type littermate controls (wild-type mice, 40.9 ± 1.7% vs Fli-1ΔCTA/Fli-1ΔCTA mice, 30.2 ± 2.6%, n = 8, p = 0.0014), though due to the overall increase in spleen cell number, the total number of splenic B cells in the Fli-1ΔCTA/Fli-1ΔCTA mice was slightly greater than wild-type mice (Fig. 3 A).
Next, we analyzed the splenic B cells for surface expression of CD21 and CD23 to discriminate between FO B cells (B220+CD21intCD23high) and MZ B cells (B220+CD21highCD23−/low). The percentage of FO B cells were significantly decreased in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls (wild-type, 62.3 ± 0.7% vs Fli-1ΔCTA/Fli-1ΔCTA mice, 43.8 ± 0.8%, n = 8, p < 0.0001, Fig. 3,B). The total number of FO B cells in the spleens from Fli-1ΔCTA/Fli-1ΔCTA mice was significantly less than in wild-type controls (Fig. 3,A). Overall, the intensity of cell surface CD23 expression on all splenic B cells was reduced in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type mice (Fig. 3 C). It has been reported that CD23 is up-regulated by IL-4 (36). We found that expression of CD23 in B cells from both wild-type and Fli-1ΔCTA/Fli-1ΔCTA mice was up-regulated similarly after IL-4 stimulation (data not shown).
In contrast to the decreased percentage of FO B cells, MZ B cell percentage was significantly increased in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls (Fig. 3,B, wild type, 5.5 ± 0.5% vs Fli-1ΔCTA/Fli-1ΔCTA, 8.3 ± 0.9%, n = 8; p = 0.0371). The significant increase in MZ B cells from Fli-1ΔCTA/Fli-1ΔCTA mice was further confirmed using additional markers for MZ B cells, including CD1d and CD9 (Fig. 3,D). The absolute number of MZ B cells in spleens from Fli-1ΔCTA/Fli-1ΔCTA mice was also significantly higher than in wild-type mice (Fig. 3 A). Histological examination of the spleens showed an increase of IgM+ B cells outside the MOMA-1+ metallophilic macrophage region, which defines the boundary between the FO and MZ B cell zones (data not shown). These data, consistent with the flow cytometric data, demonstrate that Fli-1ΔCTA/Fli-1ΔCTA mice have an increased number of MZ B cells.
Increased number of transitional B cells in Fli-1ΔCTA/Fli-1ΔCTA mice
Upon arrival in the spleen, immature B cells emerging from the bone marrow are designated as transitional B cells and are divided into transitional stage 1 (T1) and transitional stage 2 (T2) B cells by Loder et al. (37). Recently, several studies demonstrated that these T2 cells are actually MZ precursors (MZP) cells (8, 38, 39). To further delineate possible B cell developmental alterations in the spleens of Fli-1ΔCTA/Fli-1ΔCTA mice, the transitional B cell populations were examined. T1 B cells were characterized by B220+IgMhighCD21−/lowCD23− surface marker expression, and MZP B cells by the expression of B220+IgMhighCD21highCD23+. As shown in Fig. 4, significant increases in splenic percentage of T1 and MZP B cells were observed in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type mice (for T1 B cells, wild type, 4.7 ± 0.2%, vs Fli-1ΔCTA/Fli-1ΔCTA, 11.2 ± 0.9%, n = 5, p = 0.0003; for MZP B cells, wild type, 6.8 ± 0.3%, vs Fli-1ΔCTA/Fli-1ΔCTA, 8.2 ± 0.5%, n = 5, p = 0.0442, Fig. 4). The absolute number of T1 B cells in Fli-1ΔCTA/Fli-1ΔCTA mice is 10-fold higher than in wild-type controls (40.3 × 105 vs 4.0 × 105, respectively, Fig. 3 A). However, the absolute number of MZP B cells in the spleen is similar between Fli-1ΔCTA/Fli-1ΔCTA and wild-type control mice, despite the differences in percentage. These data suggest that efficient progression of T1 B cells to the FO B cell stage was decreased in Fli-1-deficient mice. The increase in MZ B cells may reflect preferential development of MZ B cells in Fli-1 deficiency due to a block in FO B cell development rather than a positive effect on MZ B cell development.
B1 cell development is not affected in Fli-1ΔCTA/Fli-1ΔCTA mice
B1 cells are self-renewing, residing mainly in the peritoneal and pleural cavities (9). To examine whether Fli-1 deficiency also affected B1 cell development, we collected peritoneal lymphocytes from Fli-1ΔCTA/Fli-1ΔCTA and wild-type control mice. The cells were stained with anti-CD5 and -IgM. There was no significant difference in peritoneal B1 cell numbers between Fli-1ΔCTA/Fli-1ΔCTA and wild-type control mice (wild type, 14.7 ± 1.2%, vs Fli-1ΔCTA/Fli-1ΔCTA, 14.2 ± 3.4%, n = 5, p = 0.8792).
B cell-autonomous Fli-1 defect affects B cell development in Fli-1ΔCTA/Fli-1ΔCTA mice
The reduced number of FO B cells and increased number of transitional and MZ B cells could be the result of an intrinsic Fli-1 defect in the B cells themselves or due to influences of Fli-1 deficiency on the splenic microenvironment (non-B cells). To distinguish between these two possibilities, 1 × 106 bone marrow cells from wild-type or Fli-1ΔCTA/Fli-1ΔCTA mice with the Ly5.2 (CD45.2) genotype were transferred into sublethally irradiated B6 mice with the Ly5.1 (CD45.1) genotype. The mice were sacrificed 12 wk after transplantation and the spleen cells were analyzed by flow cytometry. As shown in Fig. 5,A, over 94% of spleen cells from the recipients were CD45.2+ indicating that the reconstituting B cells in the recipients were derived from donor bone marrow. B6 Ly5.1 recipients that received wild-type bone marrow had numbers of MZ B cells and FO B cells similar to those of wild-type mice (Fig. 5 B). In contrast, the B6 Ly5.1 mice that received Fli-1ΔCTA/Fli-1ΔCTA bone marrow cells had an increased number of MZ B cells with a decreased number of FO B cells, similar to unmanipulated Fli-1ΔCTA/Fli-1ΔCTA mice (for FO B cell, wild-type bone marrow recipients, 77.0 ± 0.5% vs Fli-1ΔCTA/Fli-1ΔCTA bone marrow recipients, 49.5 ± 1.6%, n = 5; p < 0.0001, for MZ B cells, wild-type bone marrow recipients, 4.5 ± 0.5% vs Fli-1ΔCTA/Fli-1ΔCTA bone marrow recipients, 14.0 ± 1.1%, n = 5, p < 0.0001). These data indicated that the reduction of FO B cells and increase of transitional B cells and MZ B cells results from lymphocyte/bone marrow-intrinsic Fli-1 deficiency and not stromal cell effects.
Altered B cell gene expression in Fli-1ΔCTA/Fli-1ΔCTA mice
Because previous reports suggested Fli-1 may be involved in regulation of Igα (mb-1) expression (23), we first examined the expression of Igα on B cells from Fli-1ΔCTA/Fli-1ΔCTA mice. As shown in Fig. 6,A, the surface expression of Igα on B cells from Fli-1ΔCTA/Fli-1ΔCTA mice was decreased compared with wild-type mice. Real-time RT-PCR analysis of mb-1 mRNA demonstrated significantly reduced expression in bone marrow from Fli-1ΔCTA/Fli-1ΔCTA mice compared with that from wild-type mice. (Fig. 6 B). Thus, both at the protein and transcript level, Igα expression was decreased in Fli-1ΔCTA/Fli-1ΔCTA mice.
A limited set of transcription factors (e.g., ikaros, PU.1, E2A, EBF, Pax-5) regulate expression of B-lineage-specific genes. B cell development is dependent upon proper expression of multiple transcription factors, including E2A (E12 and E47), Id1, and Id2 (10, 40, 41). Real-time RT-PCR was used to examine the expression of multiple transcripts in mRNA prepared from total bone marrow and purified B cells (Fig. 6,B). Id proteins are negative regulators of transcription factors E2A and Pax-5 (40, 41). Although PU.1 and EBF were not significantly altered in the Fli-1-deficient mice, the expression of E2A and Pax-5 mRNA was significantly reduced and Id1 and Id2 mRNA was significantly increased in RNA prepared from bone marrow of Fli-1-deficient vs wild-type mice (Fig. 6,B). E2A transcript levels were reduced in RNA prepared from splenic B cells of the Fli-1-deficient mice, although the values did not reach statistical significance. Id2 transcripts were statistically significantly increased in total RNA prepared from bone marrow and purified splenic B cells (Fig. 6,B) from Fli-1ΔCTA/Fli-1ΔCTA compared with wild-type mice. Egr-1 is required for B cell maturation and is regulated by Fli-1 (25, 42). Quantitative real-time RT-PCR assessment of Egr-1 mRNA levels in RNA prepared from bone marrow and purified B cells demonstrated a consistent reduction in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type mice. Transitional, FO, and MZ B cells were isolated with a FACSVantage cell sorter for analysis of gene expression. Reduced expression of mb1 and Pax-5 in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls was observed in these isolated populations (Fig. 6,C). Transitional B cells demonstrated higher expression of Fli-1 than MZ and FO B cells (Fig. 6 C).
Programmed cell death is a critical process in B lymphocyte development. We found that Bcl-xL and Bcl-2 expression was decreased in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls (Fig. 6 B). We did not, however, detect differences in in vivo or in vitro apoptosis between wild-type and Fli-1ΔCTA/Fli-1ΔCTA mice (data not shown).
As noted above (Fig. 3,C), cell surface CD23 protein expression was reduced in Fli-1ΔCTA/Fli-1ΔCTA mice. Consistent with this observation, CD23 mRNA was also reduced in Fli-1ΔCTA/Fli-1ΔCTA bone marrow and isolated B cells. The CD23 mRNA reduction demonstrated in the spleen was over 2-fold (Fig. 6,B). Furthermore, CD19 transcripts were reduced in Fli-1ΔCTA/Fli-1ΔCTA mice (Fig. 6 B). At this time, it is unclear whether the reduction in CD23 and CD19 is a primary effect of Fli-1 deficiency or is secondary to the impact of Fli-1 on B cell development.
Next, we asked whether the reduced gene expression observed reflects decreased binding of Fli-1ΔCTA to Fli-1-responsive promoters. For this, we examined whether truncated Fli-1 protein binds the mb1 promoter in vivo in B cells. After immunoprecipitation with a Fli-1-specific Ab that binds both native Fli-1 and truncated Fli-1, enrichment of the endogenous promoter fragments in each sample was detected by PCR amplification. As a positive control, 5% of the input chromatin was used as a template. Normal rabbit IgG was used as a negative control (Fig. 6 D). These results indicate that while native Fli-1 and truncated Fli-1 each bound to the mb-1 promoter, Fli-1ΔCTA appears to have reduced affinity.
BCR signaling and proliferation of B cell in Fli-1ΔCTA/Fli-1ΔCTA mice
Because Igα expression was decreased in Fli-1ΔCTA/Fli-1ΔCTA mice, we isolated B cells and measured proliferation after cross-linking the BCR with anti-IgM. [3H]Thymidine uptake by B cells from Fli-1ΔCTA/Fli-1ΔCTA mice following anti-IgM cross-linking was significantly less than that from wild-type controls (Fig. 7,A) indicating that B cells from Fli-1ΔCTA/Fli-1ΔCTA mice have decreased proliferative capacity compared with wild-type B cells. To determine whether the decreased proliferation was due to early components of BCR signaling, we assessed intracellular Ca2+ levels following anti-IgM cross-linking. Despite the clear impact of Fli-1ΔCTA/Fli-1ΔCTA on B cell proliferation, intracellular Ca2+ increase after anti-IgM stimulation of B cells from Fli-1ΔCTA/Fli-1ΔCTA mice was similar to that of wild-type controls (Fig. 7 B). Thus, the decreased proliferative capacity following BCR cross-linking of Fli-1ΔCTA/Fli-1ΔCTA B cells was not due to early defects in BCR signaling, despite the decrease in Igα expression.
Immune responses of Fli-1ΔCTA/Fli-1ΔCTA mice
FO and MZ B cells have different functions in immune responses (6, 8). A decrease in FO B cells and an increase of MZ B cells in Fli-1ΔCTA/Fli-1ΔCTA mice may affect in vivo immune responses. Thus, to determine whether the lack of Fli-1 CTA had a significant in vivo effect, we examined the humoral immune responses to T cell-independent and -dependent Ags in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls. Mice were injected with 50 μg of T cell-independent Ag TNP-Ficoll or T cell-dependent Ag TNP-KLH mixed with CFA i.p. Mice immunized with T cell-dependent TNP-KLH were boosted with 50 μg of TNP-KLH mixed with IFA 1 wk after the first immunization. Induced TNP-specific Abs were measured 7, 14, and 21 days after immunization. Fli-1ΔCTA/Fli-1ΔCTA mice produced significantly lower levels of IgG1 Abs against the T cell-dependent Ag compared with wild-type control mice. Fli-1ΔCTA/Fli-1ΔCTA mice, however, produced higher levels of IgM Abs against the T cell-independent Ag, and higher levels of IgG3 Abs against both T cell-dependent and -independent Ags (Fig. 8). These results are consistent with effect of Fli-1ΔCTA on B cell development, given the increased MZ B cells (increased IgM and IgG3 response) and decreased FO B cells (reduced IgG1 response) observed in Fli-1ΔCTA/Fli-1ΔCTA mice.
Reduced in vitro IgG1 class switch of B cells from Fli-1ΔCTA/Fli-1ΔCTA mice
Because the level of IgG1 Abs was lower in Fli-1ΔCTA/Fli-1ΔCTA mice after immunization compared with wild-type controls, next we investigated whether the IgG1 class switch of B cells from Fli-1ΔCTA/Fli-1ΔCTA mice was alternated. B cells were purified from spleen and stimulated with 50 μg of LPS and recombinant murine IL-4 at 50 ng/ml to induce switching to IgG1, with LPS alone to induce switching to IgG3 and IgG2b, or with LPS and IFN-γ at 10 ng/ml to induce switching IgG2a. The percentage of IgG1-expressing B cells was significantly lower in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type mice (Fig. 9, wild type, 26.9 ± 1.6%, vs Fli-1ΔCTA/Fli-1ΔCTA, 20.9 ± 1.0%, n = 5, p = 0.029). There was no significant difference in IgG3 expression between B cells from Fli-1ΔCTA/Fli-1ΔCTA mice and that from wild-type controls (Fig. 9, wild type, 11.4 ± 1.2%, vs Fli-1ΔCTA/Fli-1ΔCTA, 10.2 ± 0.4%, n = 5, p = 0.1806). Percentages of IgG2a- and IgG2b-expressing B cells were also not significantly different between B cells from wild-type mice and Fli-1ΔCTA/Fli-1ΔCTA mice after stimulation (data not shown).
T cell development was not affected in thymus in Fli-1ΔCTA/Fli-1ΔCTA mice
Finally, we investigated whether T cell development was affected in Fli-1ΔCTA/Fli-1ΔCTA mice. Thymocyte populations were isolated from Fli-1ΔCTA/Fli-1ΔCTA and wild-type mice and analyzed by flow cytometry. There were no significant differences in mature CD4+CD8− and CD4−CD8+ thymocytes between Fli-1ΔCTA/Fli-1ΔCTA and wild-type mice (data not shown). Furthermore, the numbers of CD4+CD8+ and CD4− CD8− thymocytes in Fli-1ΔCTA/Fli-1ΔCTA mice were similar to those found in wild-type mice (data not shown).
Prior experiments in Fli-1-transgenic and heterozygous knockout mice indicated that Fli-1 has significant effects on immune function, specifically B cell function (30, 31). Because complete deficiency of Fli-1 is embryonic lethal, to determine the impact of Fli-1 deficiency on B cell development, we developed a mouse that expressed a truncated Fli-1 protein and demonstrated conclusively that Fli-1 plays a critical role in B cell development. Specifically, we found a significant decrease in splenic FO B cells and an increase in transitional and MZ B cells in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls. Although bone marrow pro-B cell number was not affected, the reduction of bone marrow pre-B cells observed in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type mice supports the proposed model that wild-type Fli-1 function is required for maximum efficiency of early B cell differentiation.
The pre-B cell population was significantly lower in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type mice, suggesting that Fli-1 plays a critical role in the differentiation of pro- to pre-B cells. Formation of the pre-BCR is required for the clonal expansion of pro-B cells and differentiation to the pre-B cell stage (3, 4). Disruption of pre-BCR signaling resulted in a significant reduction of pre-B cells in mice (43). Complete knockout of the Pax-5 gene resulted in arrest at the early pro-B cell stage and lack of pro- to pre-B cell differentiation (44, 45). Elevated Id1 contributed to a reduced pre-B cell population (46, 47). Disruption of Egr-1 and E2A genes also significantly reduced pre-B cell populations (42, 48, 49). We found that expression of mb-1, Pax-5, E2A, and Egr-1 was significantly reduced and expression of Id1 was increased in bone marrow-derived cells from Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls. The decrease in pre-B cells in Fli-1ΔCTA/Fli-1ΔCTA mice is likely due to the combination of decreased expression of Igα (pre-BCR), Erg-1, Pax-5, and E2A or increased expression of Id (Fig. 10). The decreased proliferative response of Fli-1-deficient B cells to BCR cross-linking indicates that a Fli-1-regulated factor is required for full BCR responsiveness. We believed that the decreased expression of Igα would result in decreased intracellular Ca2+ flux following BCR cross-linking. The similar response of Fli-1-deficient B cells and wild-type B cells to calcium indicates that the impact of Fli-1 expression is not at the earliest phases of BCR signaling, but further down the signal transduction pathway. Thus, it is unlikely that the decreased expression of Igα alone is sufficient to explain the B cell phenotype observed in the Fli-1- deficient mice.
The significant increase of T1 B cells in Fli-1ΔCTA/Fli-1ΔCTA mice indicated that Fli-1 is critical in the progression of B cells through the transitional stages to mature B cell subsets. A truncated cytoplasmic tail of Igα or Igβ resulted in failure of T1 B cells to progress to mature B cells. These results suggested that the BCR is also important in the transition of T1 to mature B cells (37, 50). In addition to being an early response gene following AgR engagement on mature B cells, a role for Egr-1 in promotion of transitional B cells to mature B cell subsets was demonstrated (42). We have found that expression of Egr-1 is reduced in splenic B cells in Fli-1ΔCTA/Fli-1ΔCTA mice. The decreased expressions of Igα and Egr-1 in Fli-1-deficient mice likely contribute to the decreased transition of T1 B cells to mature B cells (Fig. 10).
In this study, we found a significant decrease in FO B cells and increased MZ B cells in Fli-1–1ΔCTA/Fli-1ΔCTA mice. The BCR-signaling pathway is a major factor in determining the fate of mature B cells (5, 8, 51). Disruption of BCR-signaling components blocks transition of T2 B cells to FO B cells (43, 52, 53). Mice in which BCR signaling was impaired because of the presence of a knocked-in L chain mutation had increased numbers of MZ B cells (51). A model was proposed to explain the defects observed in MZ and FO B cell development in these gene-disrupted animals (5, 8). In this model, strong BCR signaling leads immature B cells to differentiate to FO B cells, while weak BCR signaling leads immature B cells to differentiate to MZ B cells. Igα is one of the signaling components of the BCR and plays a very important role in the BCR-signaling pathway as noted above. Disruption of the Igα cytoplasmic tail blocked generation of mature B cells (53). Mice in which the tyrosine residues in the Igα ITAM were mutated to phenylalanine exhibited enhanced Ca+2 signaling downstream of the BCR and resultant decreased number of MZ B cells (52). Fli-1 has been reported to coregulate Igα expression with Pax-5 (23). In the current study, we demonstrated that Fli-1 protein can directly bind to the promoter of the mb-1 gene. Although the truncated Fli-1 protein can also bind to the promoter of mb1, it does so to a reduced extent, perhaps accounting, in part, for the decreased expression of surface Igα protein and mRNA in B cells from Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls. This reduced expression may contribute to “weak” BCR signaling and concomitant decreased number of FO B cells and increased number of MZ B cells observed in these mice. We also found that expression of Id genes were higher in B cells from Fli-1ΔCTA/Fli-1ΔCTA mice. Relative to BCR signaling, Id1 expression is negatively correlated with mb-1 expression (54). Thus, elevated Id1 may contribute to the reduced FO B cell population. Notch-RBP-J signaling was reported to be involved in the fate determination of MZ B cells (55, 56). Because we did not find significant differences in the expression of Notch 1 or its negative regulators, NUMB, NrarP, MINT, and sel-10 (data not shown), it is unlikely that Fli-1 affects MZ B cell development via aberrant Notch signaling. B cell-activating factor (BAFF) plays an important role in B cell development. BAFF-deficient mice lost all FO and MZ B cells, but B1 cell development was not affected (57, 58). We found that there was no difference in the BAFF concentrations in the sera from Fli-1ΔCTA/Fli-1ΔCTA mice compared with that from wild-type mice (data no shown). However, in this study, we did not investigate whether the expression of BAFF receptors in Fli-1ΔCTA/Fli-1ΔCTA mice was affected, nor did we investigate signaling through the BAFF receptor. Such an alternation of BAFF receptor signaling could contribute to the lower numbers of FO B cells in Fli-1ΔCTA/Fli-1ΔCTA mice.
Fli-1 is also reported to regulate the expression of Bcl-2 (22, 59). Bcl-2 is involved in the development of FO and MZ B cells (60). Bcl-2-transgenic mice have no MZ B cells, whereas Bcl-2-knockout mice have an increased number of MZ B cells (60). We found that Bcl-2 and Bcl-xL expression was decreased in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls. Thus, decreased expression of these antiapoptotic genes may partially contribute to the increase of MZ B cells, although we could not demonstrate a significant impact of Fli-1 genotype on measures of apoptosis in the spleen (data not shown).
We and others have demonstrated that Fli-1 is highly expressed in T cells, and expression of Fli-1 is decreased after T cell activation (61, 62). Although the T cell population was higher in the spleen, we did not find significant differences in the stages of T cell development in the thymus in Fli-1ΔCTA/Fli-1ΔCTA mice. Because only the CTA domain was removed in Fli-1ΔCTA/Fli-1ΔCTA mice, it remains possible that, while Fli-1 is involved in the development of T cells, the CTA domain is not critical for these functions. In addition, these results suggest that this allele does not seem to regulate T cell lymphocyte lineage choice. These results also suggest that the B cell lineage differences are not secondary to T cell effects, but are rather intrinsic to B cells.
Several studies have demonstrated that stromal cells play an important role in lymphocyte development, and that gene-deficient stromal cells affect normal lymphocyte development (63, 64). Our bone marrow transplantation study clearly demonstrated that the observed reduction of FO B cells and increase of MZ and transitional B cells resulted from bone marrow cell-intrinsic Fli-1 deficiency, not stromal cell Fli-1 deficiency.
Fli-1 deficiency impacts in vivo immune response phenotypes. Serum IgG1 concentrations were significantly lower in Fli-1ΔCTA/Fli-1ΔCTA mice compared with wild-type controls, probably due to the decreased number of FO B cells and their decreased abilities to switch IgG1 in these mice. It is also notable that the serum IgG3 concentrations were also significantly lower in Fli-1ΔCTA/Fli-1ΔCTA mice, even though the total number of MZ B cells was higher in these mice. MZ B cells participate in early T cell-independent immune responses and interact with blood-borne Ags by production of IgM and IgG3. MZ B cells also play a role in T cell-dependent immune responses, but mature FO B cells are the principal effectors of these responses via IgG1 production (6, 8). We found that Fli-1ΔCTA/Fli-1ΔCTA mice had significantly less IgG1, but increased IgM Ab production, following challenge with T cell-dependent and -independent Ags, compared with wild-type control mice. These data are consistent with Fli-1ΔCTA/Fli-1ΔCTA resulting in increased MZ B cells and decreased FO B cells and indicate that Fli-1 induced changes in splenic B cells are of in vivo physiologic relevance.
In summary, the lack of a CTA domain of Fli-1 resulted in a novel profound impact on B cell development. The transitions from pro- to pre-B cell, pre-B to immature B cells, and T1 B cells to FO or MZ B cells appear primarily affected. A number of genes known to affect B cell development showed altered expression in Fli-1ΔCTA/Fli-1ΔCTA mice (Fig. 10). The impact of Fli-1 on B cell development is, thus, likely multifactorial, implicating BCR signaling and expression of transcription factors important for B cell lineage development/differentiation. Further studies, such as genetic reconstitution and generation of B cell-specific Fli-1-null mice, are needed to dissect the distinct mechanisms underlying the B cell effects noted. The in vivo importance of these findings in normal immunity is evidenced by the altered response to immunization of mice deficient in Fli-1 CTA expression. Furthermore, our previous report of a profound effect of Fli-1 heterozygosity on lupus development in MRL/lpr mice underscores the in vivo implications of these studies.
We thank Jeremy Mathenia for excellent technical support, Dr. David Turner for assistance with laser scanning microscopy, and Dr. Haiqun Zeng for help with cell sorting.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Abbreviations used in this paper: FO, follicular; MZ, marginal zone; CTA, C-terminal transcriptional activation domain; ChIP, chromatin immunoprecipitation; MZP, MZ precursor; KLH, keyhole limpet hemocyanin; BAFF, B cell-activating factor; TNP, 2,4,6-trinitrophenyl.