Activated Gs signaling in osteoblastic cells alters the hematopoietic stem cell niche in mice

Abstract

Adult hematopoiesis occurs primarily in the BM space where hematopoietic cells interact with stromal niche cells. Despite this close association, little is known about the specific roles of osteoblastic lineage cells (OBCs) in maintaining hematopoietic stem cells (HSCs), and how conditions affecting bone formation influence HSC function. Here we use a transgenic mouse model with the ColI(2.3) promoter driving a ligand-independent, constitutively active 5HT4 serotonin receptor (Rs1) to address how the massive increase in trabecular bone formation resulting from increased G(s) signaling in OBCs impacts HSC function and blood production. Rs1 mice display fibrous dysplasia, BM aplasia, progressive loss of HSC numbers, and impaired megakaryocyte/erythrocyte development with defective recovery after hematopoietic injury. These hematopoietic defects develop without compensatory extramedullary hematopoiesis, and the loss of HSCs occurs despite a paradoxical expansion of stromal niche cells with putative HSC-supportive activity (ie, endothelial, mesenchymal, and osteoblastic cells). However, Rs1-expressing OBCs show decreased expression of key HSC-supportive factors and impaired ability to maintain HSCs. Our findings indicate that long-term activation of G(s) signaling in OBCs leads to contextual changes in the BM niche that adversely affect HSC maintenance and blood homeostasis.

Figure 1

Changes in the BM cavity of Rs1 mice. (A) Representative photographs of femurs from 6-week-old (6 wk) and 12-week-old (12 wk) control (Ctrl) and ColI(2.3)⁺/Rs1⁺ (Rs1) mice. Scale bar represents 1 cm. (B) Dual-energy x-ray absorptiometry measurements of bone mineral density (BMD) in age-matched Ctrl and Rs1 littermates (n = 6-15 mice per group). (C-F) Hematoxylin and eosin staining of decalcified femurs from 12-weeks-old Ctrl and Rs1 mice. Low magnification (C) showing loss of the normal BM canal and gross morphologic bone changes. Scale bar represents 2 mm. High magnification showing hematopoietic BM cells and cortical bone in Ctrl (D) and Rs1 bone containing fibrous infiltrate and disorganized trabeculae (E) and hematopoietic BM cells (F) intermingled with fibrocellular infiltrate. c indicates cortical bone; bm, BM space; t, trabecular bone; and f, fibrous stromal cells. Scale bar represents 100 μm. (G) Micro-CT analysis on femurs of 15-week-old (15 wk) Ctrl (n = 4) and Rs1 (n = 5) mice. Scale bar represents 1 mm. TV and mineralized BV were determined on 50 mid-femur slices. (H) Enumeration of BM hematopoietic cell numbers released by crushing and extensive washing of bones isolated from 6-week-old and 12-week-old Ctrl and Rs1 mice (n = 6-13 per group). (I) Enumeration of BM Lin⁻/CD45⁻ endosteal stromal cells released by collagenase-digestion from the crushed bones of 12-week-old control and Rs1 mice. Data are mean ± SD. *P ≤ .05. **P ≤ .01. ***P ≤ .001.

Figure 2

Lineage-specific hematopoietic defects in the BM of Rs1 mice. Twelve-week-old (12 wk) control (Ctrl) and ColI(2.3)⁺/Rs1⁺ (Rs1) mice (n = 8-13 per group) were used for these analyses. (A) Schematic overview and (B) gating strategy for the different hematopoietic lineages investigated (gray represents populations decreased in frequency in Rs1 mice). CLP indicates common lymphoid progenitor; B, B-cell; CMP, common myeloid progenitor; GMP, granulocyte/macrophage progenitor; MEP, megakaryocyte/erythrocyte progenitor; E, erythrocyte; Plt, platelet; Gr, granulocyte; and M, macrophage. The myeloid progenitor (MP) population containing CMPs, GMPs, and MEPs and the LSK population containing HSCs and MPPs are boxed. (C) Frequencies of the indicated populations in the BM. (D) Methylcellulose read-out for myeloid CFU activity in 12 500 unfractionated Ctrl and Rs1 BM cells. (E) Schematic overview and (F) gating strategy for detailed analyses of megakaryopoiesis and erythropoiesis (gray represents populations decreased in frequency in Rs1 mice). Pre-MegE indicates megakaryocyte/erythrocyte precursor; MkP, megakaryocyte precursor; Mk, megakaryocyte; CFU-E, colony-forming unit-erythrocyte; I, pro-erythroblast; II, basophilic erythroblast; III, late basophilic/polychromatophilic erythroblast; and IV, orthochromatic erythroblast. (G) Frequencies of the indicated populations in the BM. Data are mean ± SD. *P ≤ .05. **P ≤ .01. ***P ≤ .001.

Figure 3

Impaired recovery of Rs1 mice after acute hematopoietic injury. Nine-week-old control (Ctrl) and ColI(2.3)⁺/Rs1⁺ (Rs1) mice (n = 6 per group) were used to start these experiments. (A) Hematopoietic recovery after a single injection (150 mg/kg) of 5-FU. Blood was sampled and analyzed by automated complete blood count for the indicated populations. Plt indicates platelets; and myeloid subsets, eosinophil/basophil/neutrophil/monocyte. (B) Forty days after the first 5-FU injection, mice were injected every 7 days with escalating doses of 5-FU (ranging from 180 mg/kg to 250 mg/kg, as indicated) and followed for survival. Data are mean ± SD. *P ≤ .05. **P ≤ .01. ***P ≤ .001.

Figure 4

Loss of HSC numbers in Rs1 mice. Six-week-old (6 wk) and 12-week-old (12 wk) control (Ctrl) and ColI(2.3)⁺/Rs1⁺ (Rs1) mice (n = 6-13 per group) were used for these analyses. (A) Gating strategy, (B) frequencies, and (C) total cell numbers for the indicated BM stem and multipotent progenitor cells. (D) Cell cycle analyses. Gating strategy (left) and frequencies (right) of HSCs in G₀, G₁, and S-G₂/M phases of the cell cycle. Lethally irradiated FVB/N-CD45.2 congenic recipients (n = 4 or 5 per group) were transplanted with either (E) 1 × 10⁶ CD45.1 BM cells or (F) 150 CD45.1 purified HSCs together with 300 000 Sca-1–depleted CD45.2 helper BM cells. Transplanted mice were bled every 4 weeks and analyzed for the percentage of donor (CD45.1⁺) chimerism in the peripheral blood. Data are mean ± SD. *P ≤ .05. **P ≤ .01. ***P ≤ .001.

Figure 5

Expansion of BM stromal cells in Rs1 mice. Nine-week-old (9 wk) and 12-week-old (12 wk) control (Ctrl) and ColI(2.3)⁺/Rs1⁺ (Rs1) mice (n = 4-12 per group) were used for these analyses. (A) Gating strategy used to identify subpopulations of endosteal BM stromal cells (top row) and to monitor ColI(2.3) promoter activity in ColI(2.3)-tTA/TetO-H2B-GFP double-transgenic reporter mice (bottom row). (B) Frequencies of stromal BM cell subpopulations expressing GFP in ColI(2.3)-tTA/TetO-H2B-GFP mice (n = 4). (C) Total cells for the indicated endosteal BM stromal cells. (D) Quantitative RT-PCR analysis of purified OBCs for expression of the indicated osteoblastic lineage genes. Data are expressed as log₂ fold relative to the average of the Ctrl samples (set to 0). Averages are shown as black bars. (E-F) Frequencies of CFU-F, CFU-Alk, and CFU-OB formed by OBCs (E) and MSCs (F) isolated from 12-week-old Ctrl and Rs1 mice. Data are mean ± SD. *P ≤ .05. **P ≤ .01. ***P ≤ .001.

Figure 6

Defective HSC-supportive function by Rs1 OBCs. Twelve-week-old control (Ctrl) and ColI(2.3)⁺/Rs1⁺ (Rs1) mice (n = 5-12 per group) were used for these analyses. (A) Quantitative RT-PCR analysis of purified OBCs for expression of the indicated HSC-supporting genes. Data are expressed as log₂ fold relative to the average of the Ctrl samples (set to 0). Averages are shown as black bars. (B) Schematic of the coculture experiment of wild-type (WT) and β-actin-GFP C57BL/6-CD45.2 HSCs (500: counting and methylcellulose; 300: transplantation) with or without Ctrl or Rs1 OBCs. (C) Enumeration of CD45⁺ hematopoietic cells per well. (D) Methylcellulose read-out for myeloid CFU activity per well. (E) Competitive transplantations. A 1:1 ratio of freshly isolated HSCs (150 GFP⁺:150 GFP⁻) or mixed wells containing the progeny of HSCs cocultured on Ctrl (GFP⁺) or Rs1 (GFP⁻) OBCs (300 HSC-derived cell equivalent) were transplanted into lethally irradiated C57BL/6-CD45.1 recipient mice together with 300 000 Sca-1–depleted CD45.1 helper BM cells (n = 3-5 per group). Transplanted mice were bled every 4 weeks and analyzed for the percentage of GFP contribution to donor (CD45.2⁺) chimerism in the peripheral blood. Data are mean ± SD. *P ≤ .05.

Figure 7

Effect of constitutive Rs1-mediated G_s signaling in OBCs on hematopoiesis. Model summarizing the extensive changes in the bone and BM stromal of ColI(2.3)⁺/Rs1⁺ (Rs1) mice and their consequences for HSC maintenance and blood development. OBC-specific expression of Rs1 increases G_s signaling and expands the number of immature osteoblasts leading to increased trabecular bone formation. This also results indirectly in increased numbers of MSCs and ECs. Together, these expanded BM stromal cells restrict the space available for hematopoietic cells and lead to BM aplasia, which is not compensated by extramedullary hematopoiesis. Increased G_s signaling in Rs1-expressing OBCs decreases their expression of key HSC-maintenance genes, including Cxcl12, Angpt1, and Vcam1, and impairs their HSC-supportive function leading to a severe loss of HSC numbers. Rs1 mice also show diminished production of megakaryocyte and erythrocyte progenitors through a mechanism that still largely remains to be elucidated but involves decreased levels of circulating Tpo. As a consequence, Rs1 mice have impaired regenerative potential after acute injury but overall preserved blood function probably because of the contribution of splenic hematopoiesis.