G-CSF down-regulation of CXCR4 expression identified as a mechanism for mobilization of myeloid cells

Abstract

CXCR4 receptor expression is required for the retention of granulocyte precursors and mature neutrophils within the bone marrow, and disruption of the SDF-1/CXCR4 axis in the bone marrow results in the mobilization of myeloid lineage cells to the peripheral circulation. We report that G-CSF down-regulates CXCR4 expression in bone marrow-derived murine and human myeloid lineage cells. When exposed to G-CSF, murine Gr1(+) bone marrow myeloid cells display a time-dependent reduction of cell-surface CXCR4 and respond poorly to SDF-1 in attachment and migration assays. Bone marrow-derived cells of nonmyeloid lineage display no change in surface CXCR4 expression upon exposure to G-CSF. Compared with controls, mice treated with G-CSF for mobilization of hematopoietic progenitor cells display reduced levels of CXCR4 selectively in bone marrow Gr1(+) myeloid cells. Since bone marrow myeloid cells express G-CSF receptors and G-CSF rapidly reduces CXCR4 expression in purified Gr1(+) cells populations, these results provide evidence that G-CSF acts directly on myeloid lineage cells to reduce CXCR4 expression. By down-regulating CXCR4 expression in bone marrow myeloid cells and attenuating their responsiveness to SDF-1, G-CSF promotes their mobilization from the bone marrow to the peripheral blood.

Figure 1.

G-CSF selectively reduces levels of surface CXCR4 in myeloid lineage cells. (A) Flow cytometric analysis of surface CXCR4 expression in bone marrow Gr1⁺ myeloid cells. Unfractionated bone marrow cell populations were incubated (1 × 10⁶/mL) at 37°C for 6 hours in medium alone or with G-CSF at concentrations of 0.1, 10, and 100 ng/mL. Cells were double stained for Gr1 (APC-labeled) and CXCR4 (FITC-labeled 2B11 antibody) or with APC- and FITC-labeled control antibodies. Gr1⁺ cells (R1 gate) represented 40.1% of total nucleated cells. (B) Analysis of CXCR4 expression in CD19⁺ B-lineage (top panel) and CD3e⁺ T-lineage (bottom panel) cells. Unfractionated bone marrow cell populations were incubated at 37°C for 6 hours with or without G-CSF (100 ng/mL). Cells were double stained for CD19 (APC-labeled) and CXCR4 (FITC-labeled) or CD3e (APC-labeled) and CXCR4 (FITC-labeled) and gated on CD19⁺ cells (44% of total nucleated cells) or CD3e⁺ (5.2% of total nucleated cells) cells. (C) Expression of CXCR4 in Gr1⁺ cells as a function of culture time (0-18 hours) in medium alone or with G-CSF (100 ng/mL). The results are expressed as mean fluorescence intensity (MFI) of CXCR4 detected on Gr1⁺ cells (mean ± SEM of 3 experiments). (D) Time-dependent reduction of surface CXCR4 in Gr1⁺ cells induced by G-CSF (100 ng/mL). Unfractionated bone marrow cell populations were incubated at 37°C for 18 hours with or without G-CSF (100 ng/mL). The results are expressed as the mean (± SEM of 5 experiments) percent reduction of surface CXCR4 induced by G-CSF compared with medium alone. *P < .05 medium versus G-CSF. (E) G-CSF reduces surface CXCR4 in precultured bone marrow cells. Unfractionated bone marrow cell populations were preincubated (37°C, 18 hours) in medium alone to achieve high-level surface CXCR4 expression, and further incubated for 18 hours with or without G-CSF (100 ng/mL). The results reflect levels of surface CXCR4 expression (expressed as MFI) in Gr1⁺ cells cultured with or without G-CSF after the preculture (mean ± SEM of 3 experiments). (F) Unfractionated bone marrow cell populations were incubated (37°C) for 18 hours with or without G-CSF (100 ng/mL), or for 40 minutes with or without SDF-1α (400 ng/mL) to achieve a reduction of surface CXCR4. After washing (PBS containing 50 mM glycine), the cells were incubated for 2 hours at 37°C in medium containing cycloheximide (100 μg/mL). The results reflect surface CXCR4 expression in Gr1⁺ myeloid cells (representative experiment of 3 performed).

Figure 2.

Effect of G-CSF on bone marrow myeloid cell adhesion and migration to SDF-1. Unfractionated bone marrow cell populations (1 × 10⁶/mL) were preincubated at 37°C for 18 hours with or without G-CSF (100 ng/mL). (A) After washing, the cells were plated onto SDF-1–coated (4 μg/mL) or diluent only–coated (PBS with 0.1% BSA) microtiter wells and incubated at 37°C for 30 minutes. After removal of nonadherent cells, the remaining cells were detached with 5 mM EDTA, counted, and stained for surface Gr1. The results reflect the differential attachment of Gr1⁺ cells preincubated in medium only or with G-CSF and are expressed as the mean (± SE) percent Gr1⁺ cells attached of input Gr1⁺ cells (results from 5 experiments). (B) After preincubation with or without G-CSF and washing, cells were tested for trans-well migration to SDF-1 (2.4-300 ng/mL) or medium only. Cells recovered in the lower well after 2-hour incubation at 37°C were counted and stained for surface Gr1 expression. Results are expressed as the mean (± SE) percent of migrated Gr1⁺ cells (results reflect the means from 3 experiments). *P < .05 (cells preincubated with medium versus G-CSF).

Figure 3.

Analysis of the relative contribution of G-CSF and soluble mediators induced by G-CSF to the reduction of surface CXCR4 on bone marrow myeloid cells. Unfractionated bone marrow cell populations were preincubated (1 × 10⁶/mL) at 37°C for 18 hours in medium alone or with G-CSF (100 ng/mL). (A) Cell-free supernatants from these cultures were preincubated (1 hour, 37°C) with rabbit neutralizing antibody against human (final concentration, 5 μg/mL) and rat neutralizing antibodies against mouse (final concentration, 3 μg/mL) G-CSF or with PBS only, and then added undiluted (1 mL) to freshly isolated bone marrow cell populations (1 × 10⁶ cells); the mixture was further incubated (6 hours, 37°C). Levels of CXCR4 expression in Gr1⁺ cells were measured by flow cytometry. The results are expressed as MFI (± SE) of triplicate determinations. *P < .05 (medium versus G-CSF). (B) Detection of G-CSF bioactivity in culture supernatants from bone marrow cell populations incubated with or without neutralizing antibodies against mouse and human G-CSF (as described under “Preparation of murine and human bone marrow cells”). G-CSF bioactivity was measured by ³H thymidine incorporation in murine NFS-60 cells, and the results are expressed as mean (± SE) of triplicate cultures. *P < .05. (C-D) Neutrophil elastase reduces levels of surface CXCR4 in bone marrow myeloid (C) and lymphoid (D) cells. Unfractionated bone marrow cell populations were incubated (1 × 10⁶/mL in DMEM medium with 0.2% BSA for 4 hours at 37°C) with medium alone, leukocyte elastase (enzyme, 40 μg/mL), and with leukocyte elastase plus leukocyte elastase inhibitor III (10 μM). Levels of surface CXCR4 were measured by flow cytometry on Gr1⁺ cells and lymphoid cells. (E) Dose dependency of leukocyte elastase–induced reduction of surface CXCR4 on Gr1⁺ cells and lymphoid cells. Unfractionated bone marrow cell populations were incubated with leukocyte elastase (5-100 μg/mL) for 2 hours at 37°C. The results are expressed as percent CXCR4 expression compared with untreated cells (MFI of elastase-treated cells/MFI of untreated cells × 100). (F) Surface CXCR4 expression in the presence of specific enzyme inhibitors. Unfractionated bone marrow cell populations were preincubated (2 hours, 37°C) in medium only (DMEM with 0.2% BSA); or with leukocyte elastase inhibitor III (LEI 10 μM); cathepsin-G (CGI, 10 μM); LEI plus CGI (10 μM each); the CD26/dipeptdyl peptidase inhibitor AB192 (50 μM); and the metalloproteinase inhibitor BB94 (10 μM). Subsequently, cells were incubated with or without G-CSF (100 ng/mL, 4 hours, 37°C). Surface CXCR4 was measured on Gr1⁺ cells by flow cytometry. The results are expressed as percent MFI compared with cells incubated in medium alone without inhibitors.

Figure 4.

Effects of G-CSF on CXCR4 protein levels in murine bone marrow myeloid cells. (A) Typical representation of CXCR4 protein in cell lysates of freshly isolated Gr1⁺ cells purified from bone marrow cell populations detected by SDS-PAGE, transferred to nitrocellulose membranes, and immunoblotted with specific antibodies to CXCR4 (2B11 antibody). (B) Effect of the enzyme PNGase F (10 U/40 mg protein) on CXCR4 protein detected by immunoblotting (2B11 antibody). (C) Time-dependent reduction of Gr1⁺ cell–associated CXCR4 induced by G-CSF. Cells lysates were prepared from purified Gr1⁺ bone marrow cells either immediately after purification or after culture in medium alone (M) or with G-CSF (G, 100 ng/mL). Samples were immunoblotted with anti-CXCR4 antibodies (2B11), stripped, and reprobed with antibodies to beta-actin (representative experiment of 5 performed). (D) Appearance of CXCR4-related bands after 18-hour incubation of electronically sorted Gr1⁺ cells (> 95% purity) in medium only (M) or with G-CSF (G, 100 ng/mL). Bottom panel reflects actin-related bands upon reprobing with anti–beta actin antibodies. (E) Levels of CXCR4 protein in Gr1⁺ cells immediately after purification and after culture (1-18 hours) with medium only or with 100 ng/mL G-CSF measured by band intensities relative to actin. The results are expressed as percent CXCR4 measured in Gr1⁺ cells immediately after purification (mean ± SD of 3-6 separate experiments; 1-hour time point: mean of 6 experiments; all other time points: mean of 3 experiments). *P < .05; **P < .1 (medium vs G-CSF).

Figure 5.

G-CSF down-regulates CXCR4 mRNA expression in bone marrow myeloid cells. (A) Effects of cycloheximide (CHX) on CXCR4 protein levels in Gr1⁺ cells detected by immunoblotting with specific antibodies to CXCR4 and actin. Purified bone marrow Gr1⁺ cells were cultured (1 × 10⁶/mL) for 1 and 2 hours in medium alone (M) or with G-CSF (G, 100 ng/mL), with or without CHX (5 μg/mL). Total cell lysates were resolved by SDS-PAGE, transferred to nitrocellulose membrane, and immunoblotted (representative experiment of 3 performed). (B) Levels of CXCR4 mRNA in purified bone marrow Gr1⁺ cells immediately after purification and after culture with G-CSF (100 ng/mL, 30 minutes to 16 hours) measured by real-time RT-PCR analysis. The results are expressed as mean fold (± SE) mRNA change relative to mRNA levels detected before culture (results from 3 independent experiments each tested in triplicate).

Figure 6.

G-CSF–induced hematopoietic cell mobilization is associated with a reduction of CXCR4 in myeloid cells. Bone marrow Gr1⁺ cells were purified from the bone marrow of groups of mice (C57/BL6) either untreated or treated for 5 days with G-CSF (5 μg/mouse, daily intraperitoneally for 5 days). (A) Gr1⁺ cells and Gr1^– cells were obtained from groups of mice treated for 5 days with G-CSF or buffer only. CXCR4 content in Gr1⁺ cells and Gr1^– cell lysates evaluated by immunoblotting with specific CXCR4 antibodies, and reprobing with anti–beta actin antibodies. (B) Western blot analysis of CXCR4 and actin content in cell lysates from 2 groups of mice treated with G-CSF and a control group of untreated mice.

Figure 7.

Effects of G-CSF on CXCR4 expression in human bone marrow myeloid cells. (A-B) Unfractionated bone marrow cell populations were incubated (1 × 10⁶/mL) at 37°C for 18 hours in medium alone or with G-CSF (1 and 100 ng/mL). Cells were double stained for CD33 (PE-labeled) and CXCR4 (APC-labeled 12G5 antibody) or with PE- and APC-labeled control antibodies (isotype). (A) Surface CXCR4 expression in human bone marrow CD33⁺ myeloid cells cultured for 18 hours in medium alone (none) or with G-CSF (1 and 100 ng/mL). (B) Surface CXCR4 expression in human bone marrow CD33^– cell populations cultured for 18 hours in medium alone (none) or with G-CSF (1 and 100 ng/mL). (C) Expression of surface CXCR4 in human CD33⁺ bone marrow cells as a function of culture time (0-18 hours) in medium alone or with G-CSF (100 ng/mL). The results are expressed as mean fluorescence intensity (MFI) of CXCR4 detected on CD33⁺ cells (mean ± SEM of 3 experiments). (D) Time-dependent reduction of surface CXCR4 in CD33⁺ cells induced by G-CSF (100 ng/mL). Unfractionated bone marrow cell populations were incubated at 37°C for 18 hours with or without G-CSF (100 ng/mL). The results are expressed as the mean (± SEM of 3 experiments) percent reduction of surface CXCR4 induced by G-CSF compared with medium alone. *P < .05 medium versus G-CSF. (E) CXCR4 detected in cell lysates of purified bone marrow CD33⁺ cells (> 85% purity) immediately after separation (0h), and after 1.5-hour and 18-hour incubation in medium only (M) or with G-CSF (G, 100 ng/mL). Bottom panel reflects the results upon membrane reprobing with anti–beta actin antibodies.