Enforced expression of EBF in hematopoietic stem cells restricts lymphopoiesis to the B cell lineage

Abstract

Mice deficient in early B cell factor (EBF) are blocked at the progenitor B cell stage prior to immunoglobulin gene rearrangement. The EBF-dependent block in B cell development occurs near the onset of B-lineage commitment, which raises the possibility that EBF may act instructively to specify the B cell fate from uncommitted, multipotential progenitor cells. To test this hypothesis, we transduced enriched hematopoietic progenitor cells with a retroviral vector that coexpressed EBF and the green fluorescent protein (GFP). Mice reconstituted with EBF-expressing cells showed a near complete absence of T lymphocytes. Spleen and peripheral blood samples were >95 and 90% GFP+EBF+ mature B cells, respectively. Both NK and lymphoid-derived dendritic cells were also significantly reduced compared with control-transplanted mice. These data suggest that EBF can restrict lymphopoiesis to the B cell lineage by blocking development of other lymphoid-derived cell pathways.

Fig. 1. Retroviral expression of EBF in reconstituted animals. (A) Murine stem cell retroviral (MSCV) constructs. LTR, long terminal repeat; IRES, internal ribosomal entry site; GFP, green fluorescent protein; EBF, early B cell factor. (B) RT–PCR analysis demonstrates coexpression of EBF with GFP in transplanted cells. GFP⁺ myeloid (M: Mac-1⁺Gr-1⁺) or B-lineage (B: B220⁺) cells were sorted from transplanted GFP (control) and EBF- reconstituted mice. The lane designated ‘N’ has all RT–PCR reagents but no template as a control. (C) Western blot analysis of FACS-sorted, splenic B cells from control and EBF-reconstituted mice using an EBF polyclonal and a β-actin monoclonal antibody. (D) Analysis of the c-kit⁺Lin^–Sca-1⁺ stem cell compartment in an animal reconstituted 12 weeks previously with EBF-expressing cells.

Fig. 2. EBF-expressing cells in the periphery are predominantly B cells. (A) Flow cytometric analysis of peripheral blood at 8 weeks post-transplant. GFP versus B220⁺ dot plots of representative GFP (control) and EBF-reconstituted animals were gated on donor Ly-5.2 cells. Bar graphs indicating the percentages of various blood cell lineages among the GFP⁺ cells are shown for 15 GFP (control) and 15 EBF mice at 12 weeks post-transplant. Error bars represent one standard deviation (P < 0.01). Samples containing an asterisk indicate statistically significant differences between control GFP and EBF-expressing populations. (B) Flow cytometric analysis of spleen cells from representative GFP control- and EBF-reconstituted animals. The B220 histograms were first gated on GFP⁺ cells in the spleen. The bar graph indicates the percentage of B220⁺GFP⁺ cells among the total GFP⁺ cells in the spleens from nine GFP control and 12 EBF-reconstituted animals at 12 weeks post-transplant (P < 0.01).

Fig. 3. B cell development in bone marrow and spleen. (A) Bone marrow GFP⁺B220⁺ cells from representative animals were first gated and then analyzed for expression of CD43 and IgM. Fraction A–C cells are B220⁺CD43⁺IgM^–, Fraction D cells are B220⁺CD43^–IgM^–, Fraction E–F cells are B220⁺CD43^–IgM⁺. The results are representative of nine GFP (control) mice and 13 EBF mice generated from three independent transplant experiments. (B) Immunoglobulin light chain expression was analyzed on cells gated for GFP expression. κ chain versus CD19 and λ chain versus CD19 dot plots were from representative GFP control (n = 9) and EBF (n = 13) spleen samples. The normal κ:λ ratio in C57B/6 animals is 10:1. (C) B cell development and maturation in spleen of 12 week post-transplant animals: GFP control (n = 9) and EBF (n = 13). Cells gated as GFP⁺B220⁺ were analyzed for IgD versus IgM expression. Splenocytes gated as GFP⁺ were analyzed for syndecan-1 versus IgM expression by FACS. Gates on the syndecan-1 plots represent percentages of terminally differentiated plasma cells. (D) EBF-expressing B cells respond to T-dependent (TD) and polyclonal (LPS) antigen stimulation and become plasma cells. The response to LPS injection was representative of four GFP- and four EBF-reconstituted mice in two independent experiments. The TD response to sheep red blood cell injection was representative of two GFP and four EBF mice in two independent experiments. Cells gated for GFP expression were analyzed for syndecan-1 versus IgM expression. Gating on syndecan^hi plasma cells is indicated.

Fig. 4. EBF-expressing T cells are severely reduced in thymus. (A) FACS analysis of GFP⁺ cells in the thymi of animals reconstituted with GFP control (n = 10) or EBF-expressing (n = 15) cells at 12 weeks post-transplant. (B) T cell development in the thymus is blocked at the TN2 stage in cells that express EBF. Cells were first gated as GFP⁺CD3^–CD4^–CD8^– and then were analyzed for CD44 versus CD25 expression in animals transplanted with GFP control- or EBF-expressing cells. The various TN subsets are indicated. Results are representative of six GFP and six EBF mice. (C) Severe reduction of EBF⁺ T-lineage cells in the spleen at 12–16 weeks post-transplant. Splenocytes were gated for GFP expression and then analyzed for CD4 versus CD8 expression (n = 10 each). (D) EBF induces apoptosis in transduced EL4 cells. The percentage of GFP⁺TUNEL⁺ cells from control or EBF-transduced EL4 cells grown for 48 h is indicated for three independent experiments (red line, GFP control; blue line, EBF-expressing EL4 cells).

Fig. 5. EBF expression leads to a partial block in dendritic and NK cell development. (A) Splenocytes were gated for GFP expression and then analyzed for Mac-1 and CD11c expression from animals transplanted with GFP control- and EBF-expressing cells. Immature DCs are gated as Mac-1⁺CD11c^int, myeloid dendritic cells (MDC) are gated as Mac-1⁺CD11c⁺, and lymphoid dendritic cells (LDC) are Mac-1^–CD11c⁺. The results represented by the bar graph are from seven GFP control and eight EBF-transplanted mice (standard deviation = 1). Asterisks indicate statistically significant differences between control GFP and EBF-expressing populations. (B) EBF expression partially blocks NK cell development. Spleen (SP) and peripheral blood (PB) cells from animals transplanted with GFP control- (n = 10) and EBF-expressing cells (n = 10) 12–16 weeks previously were first gated for GFP⁺ cells and then for mature NK cells (NK1.1⁺CD3^–).

Fig. 6. Expression of EBF in the myeloid lineage results in a reduction of myeloid cells in the periphery. (A) The percentage of GFP⁺ granulocytes among the total GFP⁺ white blood cells is indicated for peripheral blood at 4, 8 and 12 weeks post-transplant. Data are representative of 12 GFP control and 12 EBF animals (standard deviation = 1). (B) Wright–Giemsa stain of cells isolated from either methylcellulose colonies or bone marrow cells sorted as GFP⁺Gr-1⁺Mac-1⁺ from animals reconstituted with GFP control- and EBF-expressing cells. Magnification is 1000×. (C) The bar graph indicates the percentage of GFP⁺Gr-1⁺ cells in bone marrow of animals 10 weeks post-transplant. Data are from six GFP control and ten EBF animals.

Fig. 7. Endogenous EBF expression in hematopoietic progenitor cells. (A) Bone marrow cells were double sorted for long-term self-renewing HSC as Lin^–c-kit⁺Sca-1⁺Thy-1.1^lo; common lymphoid progenitors (CLP) as Lin^–c-kit^intSca-1^intThy-1.1^loIL-7R⁺; Fraction A as B220⁺CD19^–NK1.1^–. Five cells per well were sorted directly into lysis buffer for the RT reaction followed by two rounds of nested PCR using primers for EBF, Pax5 and β2 microglobulin. C1 represents a control reaction without sorted cells while C2 represents a control reaction that contains five sorted cells without the addition of RT. (B) Real-time PCR was used to evaluate the relative expression of VpreB, Pax5 and E47 in sorted bone marrow populations from EBF- transduced cells. Each data point represents an average of 2–5 separate reactions containing 1000 cells each that were first normalized to an internal actin control in all reactions. Two independent experiments were performed for each sorted subset.