The glucocorticoid receptor is required for stress erythropoiesis

Abstract

The glucocorticoid receptor (GR) coordinates a multitude of physiological responses in vivo. In vitro, glucocorticoids are required for sustained proliferation of erythroid progenitors (ebls). Here, we analyze the impact of the GR on erythropoiesis in vivo, using GR-deficient mice or mice expressing a GR defective for transactivation. In vitro, sustained proliferation of primary ebls requires an intact GR. In vivo, the GR is required for rapid expansion of ebls under stress situations like erythrolysis or hypoxia. A particular, GR-sensitive progenitor could be identified as being responsible for the stress response. Thus, GR-mediated regulation of ebl proliferation is essential for stress erythropoiesis in vivo.

Figure 1

Proliferation of murine ebls in vitro requires ligand-activated GR. Fetal livers were isolated on day 14.5 from wild-type (wt, ●) and GR^null/null (□) littermate embryos. Cells were suspended and cultivated in media containing Dex, SCF, and Epo. (A) Cells were counted at daily intervals, and cumulative cell numbers were determined. (B) Aliquots from the cultures were cytocentrifuged onto slides on day 6 and processed for histological staining. Images were taken using a CCD camera and processed with Adobe Photoshop software. Hemoglobin positive cells appear grey to black (Wessely et al. 1997). Note the large ebls in wild-type cultures. This contrasts with the mostly small, highly hemoglobinized reticulocytes/erythrocytes or dead cells and granulocytes occurring in GR^null/null cultures.

Figure 1

Proliferation of murine ebls in vitro requires ligand-activated GR. Fetal livers were isolated on day 14.5 from wild-type (wt, ●) and GR^null/null (□) littermate embryos. Cells were suspended and cultivated in media containing Dex, SCF, and Epo. (A) Cells were counted at daily intervals, and cumulative cell numbers were determined. (B) Aliquots from the cultures were cytocentrifuged onto slides on day 6 and processed for histological staining. Images were taken using a CCD camera and processed with Adobe Photoshop software. Hemoglobin positive cells appear grey to black (Wessely et al. 1997). Note the large ebls in wild-type cultures. This contrasts with the mostly small, highly hemoglobinized reticulocytes/erythrocytes or dead cells and granulocytes occurring in GR^null/null cultures.

Figure 2

Rapid ebl up-regulation in spleen upon anemia induction requires DNA-binding competent GR. Hemolytic anemia was induced by PHZ injection of adult GR^dim/dim mice and their wild-type littermates. On day 3, the spleens of these mice, as well as those from untreated GR^dim/dim and wild-type mice (two mice from each group), were isolated, cell suspensions were prepared, and aliquots of equal amounts of cells were seeded in duplicate into semisolid media supporting CFU-E formation. (A) After 2 days the number of CFU-E was counted and mean values and s.d.s determined. (Solid bars) Untreated; (stippled bars) anemic. (B) Pictures of colonies in semisolid medium were taken, and images were processed as described in Fig. 1. The arrows point to CFU-Es. Similar results were obtained in a second, independent experiment (data not shown).

Figure 2

Rapid ebl up-regulation in spleen upon anemia induction requires DNA-binding competent GR. Hemolytic anemia was induced by PHZ injection of adult GR^dim/dim mice and their wild-type littermates. On day 3, the spleens of these mice, as well as those from untreated GR^dim/dim and wild-type mice (two mice from each group), were isolated, cell suspensions were prepared, and aliquots of equal amounts of cells were seeded in duplicate into semisolid media supporting CFU-E formation. (A) After 2 days the number of CFU-E was counted and mean values and s.d.s determined. (Solid bars) Untreated; (stippled bars) anemic. (B) Pictures of colonies in semisolid medium were taken, and images were processed as described in Fig. 1. The arrows point to CFU-Es. Similar results were obtained in a second, independent experiment (data not shown).

Figure 3

Anemia-induced accumulation of erythroid, c-Kit-positive progenitors in spleen is dependent on DNA-binding competent GR. Aliquots of spleen cell suspensions, prepared as described in Fig. 2, were immunostained with fluorochrome-labeled antibodies directed against CD34, c-Kit (CD117), and the erythroid marker Ter119. (A) Viable cells were gated and evaluated by FACS analysis. The result is illustrated as a contour plot for CD34 vs. CD117. (B) The fraction of c-Kit-positive cells (above the horizontal line indicated in A) was quantitated in untreated (solid bars) and anemic (stippled bars) wild-type and GR^dim/dim mice. (C) The outlined, CD117/CD34 double positive cell populations (highly enriched in the anemic wild-type animals; A) were gated and analyzed for expression of Ter119 in wild-type (solid line, filled in white) and GR^dim/dim mice (solid line, filled in black). The broken line shows background staining in the absence of Ter119 antibody.

Figure 4

Lack of rapid adaptation to hypoxia in GR^dim/dim mice. Adult GR^dim/dim and wild-type littermates were subjected to hypoxic conditions (11% O₂) for 2 days. (A) Blood samples from hypoxia-treated mice (hatched bars; four GR^dim/dim, six wild-type animals) and untreated animals (black bars, five animals each of GR^dim/dim and wild-type) were analyzed for blood parameters by an automated hematometer (Serano 9020). Mean values and s.d.s for RBC counts, hemoglobin concentrations, and hematocrit values were calculated. The range of values determined for hemoglobin and hematocrit values in a large number of normal animals are depicted as dotted areas (Harkness and Wagner 1995). (B) The spleens of hypoxic (three per group) and normoxic mice (two per group) were isolated, cell suspensions were prepared, and aliquots containing equal cell numbers were plated in duplicate in semisolid medium supporting CFU-E formation. After 2 days, the number of CFU-E was counted, and mean values and s.d.s were calculated.

Figure 4

Lack of rapid adaptation to hypoxia in GR^dim/dim mice. Adult GR^dim/dim and wild-type littermates were subjected to hypoxic conditions (11% O₂) for 2 days. (A) Blood samples from hypoxia-treated mice (hatched bars; four GR^dim/dim, six wild-type animals) and untreated animals (black bars, five animals each of GR^dim/dim and wild-type) were analyzed for blood parameters by an automated hematometer (Serano 9020). Mean values and s.d.s for RBC counts, hemoglobin concentrations, and hematocrit values were calculated. The range of values determined for hemoglobin and hematocrit values in a large number of normal animals are depicted as dotted areas (Harkness and Wagner 1995). (B) The spleens of hypoxic (three per group) and normoxic mice (two per group) were isolated, cell suspensions were prepared, and aliquots containing equal cell numbers were plated in duplicate in semisolid medium supporting CFU-E formation. After 2 days, the number of CFU-E was counted, and mean values and s.d.s were calculated.

Figure 5

GR expression in ebls is essential for erythropoietic stress response. Adult C57BL/6 mice were irradiated with 9.5 Gy and subsequently grafted with fetal liver suspensions prepared from either GR^null/null or wild-type littermate embryos. Ten weeks after grafting, the wild-type mice reconstituted with wild-type or GR^null/null fetal liver cells were stressed by hypoxia and analyzed for blood parameters as described in the legend to Fig. 4 (total of four mice per group, two independent experiments) (hatched bars). As a control, mice grafted with wild-type or GR^null/null fetal livers and kept under normoxic conditions (black bars) were analyzed. The changes were evaluated for their statistical significance by the Wilcoxon rank sum test (increase regarded as significant when P < 0.05 ).