G-CSF Mediated Decrease in Bone-Marrow SDF-1 Level Is Neutrophil Dependent but CD26 Independent

Interaction of CXCR4 expressed on hematopoietic stem and progenitor cells (HSPC) with bone-marrow stromal SDF-1 is believed to play a central role in retention or mobilization of HSPC. Recently, a mobilization regimen of G-CSF was shown to decrease osteoblast number resulting in reduced levels of bone-marrow SDF-1, however the detailed mechanism leading to this reduction is currently unknown. It is unlikely that G-CSF directly regulates osteoblast SDF-1 production since osteoblasts do not express G-CSF receptor. Proteolytic cleavage of SDF-1 by peptidase CD26 in the bone-marrow may be an alternative mechanism responsible for reduction of SDF-1 level. Although CD26 can cleave SDF-1 in vitro, direct evidence of SDF-1 cleavage by CD26 in vivo during G-CSF induced HSPC mobilization has not been demonstrated. We previously demonstrated that neutrophils are required for G-CSF induced HSPC mobilization and that CD26 expression on neutrophils, rather than HSPC, is critical for mobilization. To more fully understand the role of CD26 in altering SDF-1 protein/activity during G-CSF induced HSPC mobilization, we quantitated bone-marrow SDF-1 levels in CD26−/− and wild-type CD26+/+ mice by ELISA during G-CSF administration. A standard 4 day G-CSF mobilization regimen (100 μg/kg bid, sc × 4 days) decreased bone-marrow total SDF-1 from 4.55±0.3 to 0.52±0.06 ng/femur in wild-type CD26+/+ mice (8.7-fold) and from 4.51±0.3 to 0.53±0.05 ng/femur (8.5-fold) in CD26−/− mice. However, despite an equivalent decrease in SDF-1, total CFU mobilization and the absolute number of mobilized SKL cells were decreased (3.1 and 2.0 fold lower, respectively) in CD26−/− mice compared to wild-type CD26+/+ controls. These results suggest that the decrease in total SDF-1 level in marrow seen following G-CSF treatment is independent of CD26. Cytological examination of bone-marrow smears showed that the reduction in SDF-1 levels in bone-marrow of both wild-type CD26+/+ and CD26−/− mice following G-CSF administration correlated with an increase in total absolute bone-marrow neutrophil cell number, suggesting a role for neutrophils in modulation of SDF-1 protein. To determine if neutrophils affect osteoblast SDF-1 production, bone marrow Gr-1+ neutrophils from wild-type CD26+/+ and CD26−/− mice were purified using anti-Ly6G magnetic beads and co-cultured with MC3T3-E1 preosteoblasts in vitro. Gr-1+ neutrophils from both wild-type and CD26−/− mice decreased pre-osteoblast SDF-1 production by similar amounts (15.4-fold vs 14.8-fold respectively), while Gr-1 neg cells from both wild-type CD26+/+ or CD26−/− were without effect on SDF-1 levels. Similarly, Gr-1+ neutrophils from both wild-type and CD26−/− mice decreased SDF-1 produced by MC3T3-E1-derived osteoblasts from 1.85±0.3 to 0.52±0.06 ng/ml (3.5 fold) and 0.56±0.07 ng/ml (3.3 fold) respectively, with Gr-1neg cells having no effect. Gr-1+ neutrophils either from wild-type or CD26−/− mice, but not Gr-1neg cells, significantly induced apoptosis of MC3T3-E1 cells as measured by Annexin-V staining (70.5%±10.2 vs 71.2%±12.5 for wild-type CD26+/+ and CD26−/− neutrophils respectively) and significantly inhibited osteoblast activity (20-fold vs 20.6-fold for CD26+/+ and CD26−/− neutrophils respectively) as measured by osteocalcin expression. Furthermore, irrespective of G-CSF treatment, an inverse correlation between absolute neutrophil number and SDF-1 protein levels was observed, suggesting that G-CSF induces neutrophil expansion but does not directly affect SDF-1 production. Collectively, these results provide additional support for the critical role of neutrophils in G-CSF induced mobilization and strongly suggested that neutrophils directly regulate bone-marrow SDF-1 levels independent of CD26 activity.