Hierarchy of molecular-pathway usage in bone marrow homing and its shift by cytokines

Abstract

Efficient bone marrow (BM) homing is a prerequisite for successful engraftment of transplanted hematopoietic cells (HPCs). Contradictory conclusions about the contribution of SDF-1/CXCR4 have clouded our understanding of its role within the molecular pathway cooperation needed for BM homing, particularly with the well-defined hierarchic network of adhesion molecules. In the present study we sought to unravel cooperative and compensatory molecular pathways guiding BM homing. Fresh BM-HPCs, rendered either SDF-1 unresponsive or Gi-signaling refractory, homed quite efficiently, because of compensation by alpha4-integrin interacting with VCAM-1. The contribution of SDF-1/CXCR4- or Gi-protein-mediated signals to BM homing became apparent after their blockade was combined with deletion of alpha4-integrin, leading to dramatic reduction in BM homing. Similar conclusions were revealed when VCAM-1-deficient hosts were used. Cytokine incubation changed the functional properties of BM-HPCs and hierarchy of molecular pathway usage in homing, by shifting the dominance among the homing mediators: loss of CXCR4 or Gi-signaling now significantly reduced BM homing, with only partial compensation through alpha4/VCAM-1 and endothelial selectins. These studies depict a flexible hierarchy of cooperating homing pathways, in which dominant players are repositioned with changing cytokine milieu, and possibly source of HPCs.

Figure 1.

Dominant role of α4-integrin/VCAM-1 interaction in BM homing of fresh BM-HPCs. (A) AMD3100 or PTX block SDF-1-induced chemotaxis. Chemokinetic () or chemotactic (SDF-1, ▪) migration of fresh c-kit⁺ WT BM-HPCs through 5-μm transwells was tested in untreated control BM cells, in the presence of AMD3100 (100 μg/mL) or after PTX treatment (100 ng/mL, 4 hours). AMD3100 (left) significantly attenuated SDF-1-directed migration. PTX incubation (right) completely blocked SDF-1-directed migration, but low-level spontaneous migration was maintained in the presence of PTX (^*P < .05). (B-D) BM homing of fresh BM-HPCs was tested 3 hours (B) or 18 hours (C,D) after transplantation of lethally irradiated recipients. Donor-cell strain/treatment and recipient strain are indicated below the respective bars. BM homing is given as the percentage of injected CFU-Cs recovered per femur. Depicted are mean plus SEM of all mice tested with this modality. (B) Three-hour BM homing of fresh BM cells treated with or without AMD3100. BM homing of fresh WT BM cells was not affected by AMD3100 incubation/coinjection. BM homing of α4^-/- BM cells in WT recipients was significantly decreased compared with WT-to-WT (^*P < .05). AMD3100 treatment of α4^-/- BM cells additionally reduced BM homing (^*P < .05 compared with untreated α4^-/--to-WT transplantation). (C) Eighteen-hour BM homing of fresh BM cells. BM homing of α4^-/- cells in WT recipients, or WT cells in VCAM-1^-/- recipients was significantly reduced compared with WT-to-WT (^*P < .05). BM homing of WT BM in CD62^-/- recipients, which are deficient in endothelial selectins, was no different from WT-to-WT. BM homing of α4^-/- BM-HPCs in CD62^-/- recipients was significantly less efficient than that of α4^-/- in WT (^*P < .05). (D) Eighteen-hour BM homing of fresh PTX-treated BM cells. BM homing of PTX-treated WT cells in WT recipients was no different from untreated WT-to-WT, but BM homing of PTX-treated α4^-/- cells in WT recipients, or BM homing of PTX-treated WT cells in VCAM-1^-/- hosts were almost completely abrogated (^*P < .05 compared with untreated α4^-/--to-WT or WT-to-VCAM-1 transplantation). In contrast, PTX-treated WT cells homed normally in CD62^-/- hosts.

Figure 2.

Dominant role of CXCR4/SDF-1 and Gi-signals in BM homing of ex vivo cytokine-incubated BM-HPCs. (A) SCF incubation increases cell motility. Spontaneous migration () and SDF-1-directed chemotactic migration (▪) of fresh or overnight SCF-incubated cells, with or without PTX, was quantified after 4 hours. Depicted are mean plus SEM. Spontaneous and SDF-1-directed migration were both significantly increased by SCF incubation. This increase in migration was completely blocked by PTX, to levels of spontaneous migration of fresh BM, which apparently represents Gi-protein-independent migration (^*P < .05). (B-D) BM homing of overnight SCF-incubated BM-HPCs. BM homing of SCF-incubated BM-HPCs was tested 3 hours (B) or 18 hours (C,D) after transplantation of lethally irradiated recipients. Donor-cell source/treatment and recipient strain are indicated below the respective bars. BM homing is given as percentage of injected CFU-Cs recovered per femur. Depicted are mean plus SEM of all mice tested with this donor/host constellation. (B) Three-hour BM homing of SCF-incubated cells treated with or without AMD3100. BM homing of fresh WT BM cells was significantly reduced by AMD3100 incubation/coinjection (^*P < .05). (C) Eighteen-hour BM homing of SCF-incubated BM cells. BM homing of α4^-/- cells in WT recipients was significantly reduced compared with WT-to-WT (^*P < .05). BM homing of WT cells in CD62^-/- recipients or of β2^-/- cells in WT recipients were no different from WT-to-WT. (D) Eighteen-hour BM homing of SCF plus PTX-treated BM cells. BM homing of SCF plus PTX-treated WT cells in WT recipients was 75% reduced compared with SCF-treated WT-to-WT without PTX (P < .05). BM homing of SCF plus PTX-treated α4^-/- cells in WT recipients or BM homing of SCF plus PTX-treated WT cells in VCAM-1^-/- hosts was almost completely abrogated, significantly more strongly than SCF plus PTX-treated WT-to-WT (^*P < .05). The inhibition of homing of SCF plus PTX-incubated β2^-/- cells in WT recipients was similar to that of SCF plus PTX-treated WT donor cells.

Figure 3.

Gi-protein signals in BM homing and short-term engraftment. (A) PTX-sensitive signals guide BM homing of TPO-incubated murine CFU-Cs. Murine WT BM cells, incubated overnight in TPO ± PTX, was transplanted and analyzed as described above. As with SCF plus PTX-treated BM cells, 18-hour homing of TPO plus PTX-incubated CFU-Cs was significantly decreased compared with TPO alone (^*P < .05). BM homing is given as percentage of injected CFU-Cs recovered per femur (mean plus SEM). (B) Effect of PTX on lin^-c-kit⁺ BM-cell homing. BM homing of SCF ± PTX-incubated lin^-c-kit⁺ cells, a population enriched in HSCs, was tested. The efficiency of 18-hour BM homing was significantly reduced by PTX (^*P < .05), similar to identically treated CFU-Cs. The mean of cells recovered from BM of recipients of untreated control cells was considered 100%; BM homing is given as percentage thereof (percentage of mean of control plus SEM). (C) Equal numbers of BM cells, which contained similar numbers of CFU-Cs, either incubated with cytokine-free ± PTX for 4 hours (left) or with SCF ± PTX overnight (right), were injected into lethally irradiated recipients. Mean plus SEM of recovered CFU-C/femur 8 days after transplantation of fresh (left) or SCF-incubated (right) BM cells are depicted. Short-term engraftment of PTX-treated samples was significantly poorer than that of PTX untreated samples (^*P < .05), in excess of the effect of PTX on BM homing under the respective conditions. (D) Schematic representation of the cooperative model of homing molecules between α4β1-integrin/VCAM-1, SDF-1/CXCR4/Gi-proteins, endothelial selectins, and β2-integrins, suggested by the data presented in this study. Cytokine incubation shifts the dominance among the molecular homing pathways from α4β1/VCAM-1 (top) toward SDF-1/CXCR4 (bottom).