Abstract
The exact mechanisms mediating G-CSF-induced hematopoietic progenitor cell (HPC) egress from the bone marrow (BM) are incompletely understood. Recent studies have suggested that the degradation of SDF-1 in the BM by G-CSF-induced proteolysis may play an important role. We previously hypothesized that endogenous galactocerebrosides (GCs) might be involved in HPC trafficking since certain sulfogalactolipids share biological properties with fucoidan, a sulfated fucose polymer endowed with mobilization activity, and showed that G-CSF fails to induce HPC mobilization in mice lacking UDP-galactose:ceramide galactosyl transferase (CGT) (Blood 2001 98:811a), the enzyme necessary for GC synthesis. To gain further mechanistic insights, we assessed protease activity and found no difference in elastase release from BM cells and in the degradation of exogenous SDF-1 in BM extracellular fluid (BMEF) between CGT−/− and +/+ littermates. Furthermore, endogenous SDF-1 levels in BMEF of CGT−/− and +/+ mice showed a similar reduction after G-CSF stimulation (>50% in CGT−/− mice, n=7–9, p<0.05) despite a virtual absence of mobilization. These data suggest that the reduction of SDF-1 in bone marrow is not essential for G-CSF-induced mobilization. To evaluate the spacial distribution of SDF-1 expression in mouse BM, we stained SDF-1 using the tyramide amplification system. We found that SDF-1 staining was sparsely distributed in the BM but, surprisingly, strong homogenous staining was observed in the surrounding bone. Staining specificity was confirmed by ELISA (2.6±0.5 vs 5.8±1.0 ng SDF-1 per femur for BMEF and bone protein extracts, respectively, n=8, p<0.05). Following G-CSF stimulation, SDF-1 protein levels were significantly decreased in bone extracts from CGT+/+ littermates (53% reduction, n=4–5, p<0.05), but were virtually unchanged in CGT−/− mice. Quantitative real-time RT-PCR analyses revealed that SDF-1 was transcriptionally downregulated by G-CSF in both BM and bone in CGT+/+ mice but there was no significant reduction in CGT−/− bone. Since osteoblasts may represent a major source of SDF-1, we suspected that osteoblast activity might be altered in CGT−/− mice. We thus measured plasma osteocalcin levels by ELISA and found a significant reduction in CGT−/− mice compared to CGT+/+ littermates (39% reduction, n=6–9, p<0.001). Immunofluorescence experiments revealed that bone lining osteoblasts in CGT−/− mice were flattened and small while in CGT+/+ littermates, these cells displayed a healthy cobblestone-like appearance. Furthermore, there was a trend toward reduction of gene expression in Runx2, a critical transcription factor in osteoblasts, and α1(I) collagen, an osteoblast-specific bone matrix protein, in CGT−/− BM by real-time RT-PCR. These data suggest that dysregulation of bone SDF-1 in CGT−/− mice may be due to constitutive downregulation of osteoblastic activity. Strikingly, Runx2 and α1(I) collagen were dramatically downregulated by G-CSF in CGT+/+ BM (Runx2; 65% reduction, n=4, p<0.001, α1(I) collagen; 92% reduction, n=4, p<0.05 by real-time RT-PCR). G-CSF does not appear to act directly on osteoblasts since G-CSF receptor mRNA was not detectable in primary osteoblast and 4 different osteoblast lineage cell lines. In conclusion, bone SDF-1, rather than that of BM, may regulate HPC mobilization. The abnormal regulation of bone SDF-1 and reduced osteoblastic activity in CGT−/− mice strongly suggest that bone SDF-1 originates from osteoblasts and that a rapid downregulation of osteoblastic activity may play a key role in the egress of stem cells from BM.
Author notes
Corresponding author
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal