Aberrant regulation of hematopoiesis by T cells in BAZF-deficient mice

Abstract

The BAZF (BCL-6b) protein is highly similar to the BCL-6 transcriptional repressor. While BCL-6 has been characterized extensively, relatively little is known about the normal function of BAZF. In order to understand the physiological role of BAZF, we created BAZF-deficient mice. Unlike BCL-6-deficient mice, BAZF-deficient mice are healthy and normal in size. However, BAZF-deficient mice have a hematopoietic progenitor phenotype that is almost identical to that of BCL-6-deficient mice. Compared to wild-type mice, both BAZF-deficient and BCL-6-deficient mice have greatly reduced numbers of cycling hematopoietic progenitor cells (HPC) in the BM and greatly increased numbers of cycling HPC in the spleen. In contrast to HPC from wild-type mice, HPC from BAZF-deficient and BCL-6-deficient mice are resistant to chemokine-induced myelosuppression and do not show a synergistic growth response to granulocyte-macrophage colony-stimulating factor plus stem cell factor. Depletion of CD8 T cells in BAZF-deficient mice reverses several of the hematopoietic defects in these mice. Since both BAZF- and BCL-6-deficient mice have defects in CD8 T-cell differentiation, we hypothesize that both BCL-6 and BAZF regulate HPC homeostasis by an indirect pathway involving CD8 T cells.

FIG. 1.

Scheme for mutating the BAZF gene in the mouse germ line. The mutation introduced by our gene-targeting construct deletes from the genome the region of the BAZF gene encoding 124 C-terminal amino acids of the 494-amino-acid BAZF protein. A region of BAZF encoding the most C-terminal 25 amino acids is not deleted by our targeting construct. The 124-amino-acid region deleted includes all of the zinc finger amino acids. The targeted BAZF gene is thus incapable of producing a DNA binding protein by alternative splicing. (A) Structure of the mouse BAZF gene locus, the targeting construct, and the recombined BAZF locus. ZF, zinc finger; E, EcoRI sites. Note that the four EcoRI sites in the targeting construct are not present in the endogenous gene. (B) Southern blot showing genomic DNA for wild-type (+/+) mice and mice heterozygous (+/−) and homozygous (−/−) for the BAZF mutation (BAZF-deficient mice). Genomic DNA was digested with EcoRI, and the blot was probed with the DNA fragment indicated in panel A. (C) RT-PCR for BAZF expression analyzing RNA from the lungs of wild-type (+/+) mice and BAZF-deficient (−/−) mice. The RT-PCR primers were designed to amplify the zinc finger region of BAZF that is deleted in BAZF-deficient mice. β-Tubulin is a control for the cDNA loading.

FIG. 2.

White blood cell and red blood cell counts of BAZF-deficient mice. Average percentages of lymphoid cells and myeloid subsets are shown for BM, peripheral blood leukocytes (PBL), and spleen (SP). Numbers of red blood cells (RBC) and platelets from PBL are shown. Results shown are averaged from four mice of each type. Error bars show standard deviations. P values were calculated with Student's t test. *, P < 0.05.

FIG. 3.

Decreased hematopoietic progenitor activity in the BM and increased hematopoietic progenitor activity in the spleens of BAZF-deficient mice. Results are the averages ± standard errors of the means from eight different wild-type and BAZF-deficient mice, assessed in two different experiments. P values were calculated with Student's t test. ***, P < 0.001; **, P < 0.01; *, P < 0.05. (A) BM cells were analyzed for HPC numbers per femur. (B) Cycling status of the BM hematopoietic progenitors, expressed as the percentage of cells in the cell cycle. (C) Spleen cells were analyzed for HPC numbers per spleen. (D) Cycling status of the splenic hematopoietic progenitors, expressed as the percentage of cells in the cell cycle.

FIG. 4.

BAZF-heterozygous mice show different HPC numbers and cycling activity than BAZF-deficient mice. HPC numbers and cycling in the BM and spleen were calculated for wild type, BAZF^+/−, and BAZF^−/− mice. Results shown are averaged from four of each mouse type. Error bars show standard deviations. P values are relative to wild type and were calculated with Student's t test. ***, P < 0.0001; **, P < 0.001; *, P < 0.05. Only statistically significant comparisons are designated.

FIG. 5.

Decreased hematopoietic progenitor activity in the BM and increased hematopoietic progenitor activity in the spleens of BCL-6-deficient mice. Results are the averages ± standard errors of the means from six different wild-type and BCL-6-deficient mice, assessed in two different experiments. P values were calculated with Student's t test. ***, P < 0.001; **, P < 0.01; N.S., not significant (P > 0.05). The reason the HPC numbers for the wild-type mice in this experiment are different from the HPC numbers for the wild-type mice in the experiment shown in Fig. 3 is that the experiment shown here was performed at a different time with different lots of serum and growth factors. (A) BM cells were analyzed for HPC numbers per femur. (B) Cycling status of the BM hematopoietic progenitors, expressed as the percentage of cells in the cell cycle. (C) Spleen cells were analyzed for HPC numbers per spleen. (D) Cycling status of the splenic hematopoietic progenitors, expressed as the percentage of cells in the cell cycle.

FIG. 6.

HPC from BAZF-deficient mice and BCL-6-deficient mice do not mount a synergistic response with SCF plus GM-CSF. Wild-type, BAZF-deficient, and BCL-6-deficient BM cells were stimulated in vitro with the indicated growth factors and then cultured in a methylcellulose colony assay. CFU-GM colony formation was determined after 7 days. Results shown are the averages for three different mice ± standard errors of the means for each type. P values were calculated with Student's t test. **, P < 0.01; *, P < 0.05 (compared to wild-type values).

FIG. 7.

HPC from BAZF-deficient mice and BCL-6-deficient mice do not respond to chemokine-mediated inhibition. Wild-type, BAZF-deficient, and BCL-6-deficient BM cells were stimulated in vitro to activate cycling and then, after a washing, cultured poststimulation in the presence of the indicated chemokines in a methylcellulose colony assay. Fifty-seven percent of wild-type BM cells were in cycle and 56% of both BAZF-deficient and BCL-6-deficient cells were in cycle following in vitro stimulation. CFU-GM colony formation was determined after 7 days. Results shown are the averages for three different mice ± standard errors of the means for each type. Results are plotted as percent inhibition of colony formation with chemokines and TNF-α, compared to colony formation without chemokine/TNF-α addition. P values were calculated with Student's t test. ***, P < 0.001; **, P < 0.01; *, P < 0.05 (compared to wild-type values).

FIG. 8.

Deletion of mature lymphocytes from BAZF-deficient mice ablates the increased development of splenic HPC. Spleen cells from the indicated mice were analyzed for the numbers of HPC (A) and the percentage of HPC in cycle (B). Error bars show standard errors. P values were calculated with Student's t test. ***, P < 0.0001; **, P < 0.001; *, P < 0.05. Only statistically significant comparisons are designated.

FIG. 9.

Depletion of T-lymphocyte subsets reverses abnormal HPC cycling activity in the BM and spleens of BAZF-deficient mice. Wild-type and BAZF-deficient (BAZF KO) mice were treated with Ab to CD4 or CD8 to deplete different subsets of T cells. BM cells from the indicated mice were analyzed for the numbers of HPC in (A) BM and (B) spleen. Results shown are averages for 10 mice per treatment. Error bars show standard errors. P values are in comparison to wild-type control mice and were calculated with Student's t test. *, P < 0.05 compared to wild-type “PBS control” mice; @, P < 0.05 compared to BAZF-deficient “PBS control” mice. Only statistically significant comparisons are designated.

FIG. 10.

Depletion of CD8 T lymphocytes promotes a normal synergistic response of BAZF-deficient (BAZF KO) BM progenitors to SCF plus GM-CSF stimulation. Wild-type and BAZF-deficient mice were treated with PBS alone, Ab to CD4, or Ab to CD8. BM cells from the indicated mice were analyzed for the numbers of HPC following in vitro stimulation with SCF and GM-CSF and culture in a methylcellulose colony assay. CFU-GM colony formation was determined after 7 days. Results shown are the averages of total CFU-GM per femur from four different mice ± standard errors of the means. P values are in comparison to wild-type control mice and were calculated with Student's t test. *, P < 0.05 compared to wild-type values.

FIG. 11.

Depletion of CD8 T lymphocytes restores the sensitivity of BAZF-deficient BM progenitors to chemokine-mediated inhibition. Wild-type and BAZF-deficient mice were treated with Ab to CD4 or CD8 to deplete different subsets of T cells. BM cells were stimulated in vitro to activate cycling and then, after a washing, cultured poststimulation in the presence of the indicated chemokines in a methylcellulose colony assay. CFU-GM colony formation was determined after 7 days. Results shown are the averages for four different mice ± standard errors of the means for each type. Results are plotted as percent inhibition of colony formation with chemokines and TNF-α, compared to colony formation without chemokine/TNF-α addition.