Blood Journal
Leading the way in experimental and clinical research in hematology

Rapid and efficient homing of human CD34+CD38−/lowCXCR4+stem and progenitor cells to the bone marrow and spleen of NOD/SCID and NOD/SCID/B2m null mice

  1. Orit Kollet,
  2. Asaf Spiegel,
  3. Amnon Peled,
  4. Isabelle Petit,
  5. Tamara Byk,
  6. Rami Hershkoviz,
  7. Esther Guetta,
  8. Gad Barkai,
  9. Arnon Nagler, and
  10. Tsvee Lapidot
  1. 1 From the Department of Immunology, The Weizmann Institute of Science, Rehovot, Israel; Department of Internal Medicine, Meir Hospital, Kfar Saba, Israel; Department of Obstetrics and Gynecology and Danek Gertner Institute of Human Genetics, Sheba Medical Center Tel-Hashomer, and Department of Human Genetics and Molecular Medicine, Sackler Faculty of Medicine, Tel-Aviv University, Ramat-Aviv, Tel-Aviv, Israel; and Hadassah University Hospital, Jerusalem, Israel.


Stem cell homing into the bone microenvironment is the first step in the initiation of marrow-derived blood cells. It is reported that human severe combined immunodeficient (SCID) repopulating cells home and accumulate rapidly, within a few hours, in the bone marrow and spleen of immunodeficient mice previously conditioned with total body irradiation. Primitive CD34+CD38−/lowCXCR4+ cells capable of engrafting primary and secondary recipient mice selectively homed to the bone marrow and spleen, whereas CD34CD38−/lowLin cells were not detected. Moreover, whereas freshly isolated CD34+CD38+/high cells did not home, in vivo stimulation with granulocyte colony-stimulating factor as part of the mobilization process, or in vitro stem cell factor stimulation for 2 to 4 days, potentiated the homing capabilities of cytokine-stimulated CD34+CD38+ cells. Homing of enriched human CD34+ cells was inhibited by pretreatment with anti-CXCR4 antibodies. Moreover, primitive CD34+CD38−/lowCXCR4+cells also homed in response to a gradient of human stromal cell-derived factor 1 (SDF-1), directly injected into the bone marrow or spleen of nonirradiated NOD/SCID mice. Homing was also inhibited by pretreatment of CD34+ cells with antibodies for the major integrins VLA-4, VLA-5, and LFA-1. Pertussis toxin, an inhibitor of signals mediated by Gαiproteins, inhibited SDF-1–mediated in vitro transwell migration but not adhesion or in vivo homing of CD34+ cells. Homing of human CD34+ cells was also blocked by chelerythrine chloride, a broad-range protein kinase C inhibitor. This study reveals rapid and efficient homing to the murine bone marrow by primitive human CD34+CD38−/ lowCXCR4+cells that is integrin mediated and depends on activation of the protein kinase C signal transduction pathway by SDF-1.


During development hematopoietic stem cells migrate from the fetal liver into the bone marrow (BM) and continuously produce maturing hematopoietic cells that are released into the blood circulation. Hematopoietic stem cells are functionally defined, based on their ability to home to the BM microenvironment and to durably repopulate transplanted recipients with both myeloid and lymphoid cells.1-3 In vivo repopulating assays for human stem cells have been developed by several groups, including ours.4-7In these assays human severe combined immunodeficient (SCID) repopulating cells (SRCs), characterized as CD34+CD38 or as CD34+CD38−/low cells, engraft the BM of sublethally irradiated immune-deficient NOD/SCID and NOD/SCID/B2mnull mice with high levels of myeloid and lymphoid cells.7 8 Other primitive cells such as CD34+CD38+ or CD34CD38 cells were also found to have limited repopulation potential.9 10 The chemokine stromal cell–derived factor 1 (SDF-1), which is also secreted by stromal cells in the BM microenvironment, was shown to attract human CD34+ cells in vitro.11-13 Recently we reported that engraftment of human SRC is dependent on the expression of SDF-1 and its receptor CXCR4, recharacterizing SRCs as CD34+CD38−/lowCXCR4+ cells with major stem cell properties.14

To home from the blood circulation into the BM microenvironment, hematopoietic stem and progenitor cells must first roll on E and P selectins, which are expressed on the BM vascular endothelial cells.15-17 Following firm arrest and adhesion to the vessel wall under shear flow, a process mediated by the major integrins (VLA-4, VLA-5, and LFA-1) and their vascular ligands (VCAM-1 and ICAM-1), the cells extravasate through the endothelium into the hematopoietic compartment.18-22 SDF-1 is constitutively expressed by human and murine BM endothelial cells,23 24and it activates multiple adhesive processes such as LFA-1/ICAM-1 interactions.15-17 25-27 Human CD34+ cells require surface-bound SDF-1 on human endothelial cells for the development of integrin-mediated firm adhesion to the vascular endothelium under physiologic shear flow.24 In addition, SDF-1 regulates interactions between immature human CD34+cells and the BM microenvironment, ie, stromal cells and extracellular matrix, by activating the major integrins LFA-1, VLA-4, and VLA-5 that are crucial for engraftment of SRCs.28 Homing of human CD34+ cells was shown to be dependent on VLA-4, using the fetal sheep model.29

In the murine system, stem cell homing was detected in both the BM and spleen of transplanted recipients.30-32 Szilvassy et al30 found that, when transplanted into secondary recipients, homing cells recovered from the spleen produced higher numbers of colony-forming cells and also generated circulating leukocytes faster than did homing cells that were recovered from the BM. Different results were found by Lanzkron et al31, who found that only cells recovered from the BM, and not cells recovered from the spleen, were capable of secondary engraftment.

The chemokine receptor CXCR4 is a G-protein–coupled receptor.33 Pertussis toxin (PTX), an inhibitor of signal transduction mediated by the Gαi subunit, almost completely abrogated in vitro migration of human CD34+cells toward a gradient of SDF-1 in transwells.11 In contrast, in vivo experiments performed with PTX-pretreated murine stem and progenitor cells showed only delayed engraftment of the spleen and no change in BM repopulation.34 In this report we studied mechanisms that regulate homing of human SRC/stem cells in transplanted immune-deficient NOD/SCID or NOD/SCID/B2mnullmice.8

Materials and methods

Human cells

Human cord blood (CB) samples were obtained from full-term deliveries after informed consent and were used in accordance with the procedures approved by the human experimentation and ethics committees of the Weizmann Institute. The blood samples were diluted 1:1 in phosphate buffered saline (PBS) without Mg+2/Ca+2. Low-density mononuclear cells were collected after standard separation on Ficoll-Paque (Pharmacia Biotech, Uppsala, Sweden) and washed in PBS. CD34+ cells were purified, using the MACS cell isolation kit and MidiMacs columns (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's instructions, obtaining purity of more than 95%. Isolated CD34+ cells were either used immediately for homing experiments or after overnight incubation with RPMI supplemented with 10% fetal calf serum (FCS) and stem cell factor (SCF) (50 ng/mL). In both cases only primitive CD34+CD38−/low cells homed in vivo. Enriched CD34+ cells were further labeled with human specific monoclonal antibody (mAb) anti-CD34 FITC (Becton Dickinson, San Jose, CA) and anti-CD38 PE (Coulter, Miami, FL) and sorted for CD34+CD38−/low- or CD34+CD38+-purified subpopulations by FACStar+ (Becton Dickinson), obtaining purity of 97% to 99%. For CD34CD38Lin cell experiments, the primitive progenitor enrichment cocktail of StemSep was used according to the manufacturer's instructions (StemCell Technologies, Vancouver, Canada). CD34+ depletion was first carried out (unless indicated otherwise) before the remaining cells were incubated with StemSep cocktail. Leftover, mobilized peripheral blood leukocyte cells were obtained from healthy donors on days 5 and 6 following daily injections of human granulocyte colony-stimulating factor (10 μg/kg) (Neupogen, Hoffman-La Roche, Nutley, NJ) stimulation for allogeneic transplantation after informed consent. CD34+ cell enrichment was performed as described above. Human CB and MPB CD34+ cells were cultured for 2 or 3.5 days with SCF alone or with SCF+IL6 as indicated (50 ng/mL SCF [R&D] and IL-6 [Interpharm Laboratories, the Ares-Serono Group, Ness Ziona, Israel]).


NOD/LtSz-Prkdcscid (NOD/SCID) mice and β2 microglobulin knockout NOD/LtSz-scid B2mnull mice (NOD/SCID/B2mnull)8 35 were bred and maintained under defined flora conditions at the Weizmann Institute in sterile micro-isolator cages. All the experiments were approved by the animal care committee of the Weizmann Institute. Eight-week-old mice were sublethally irradiated (at 375 and 350 cGy for NOD/SCID and NOD/SCID/B2mnull, respectively, from a60Co source) and transplanted with 0.5-1 × 106 human CD34+ cells/mouse 24 hours later (unless indicated otherwise). In some experiments, nonirradiated NOD/SCID mice were anesthetized with a mixture containing 15% xylazin (20 mg/mL; Vitamed, Benyamina, Israel) and 85% ketaset (100 mg/mL; Fort Dodge Animal Health, Overland Park, KS) (15 μL/mouse, IP). Human SDF-1 (1 μg/mouse; R&D Systems, Minneapolis, MN) was injected directly into the spleen or into the BM of one femur. Human CD34+ cells were injected into the tail vein in 0.5 mL RPMI supplemented with 10% FCS. Mice were killed at different time points after transplantation as indicated; BM, spleen, and lung cells were harvested and suspended into single cell suspension.

Homing assay

Human CD34+-enriched cells were either labeled prior to transplantation with the fluorescent dye PKH26-GL (Sigma)30-32 (2 μL PKH26-GL were added to 1-10 × 106 CD34+ cells in a total volume of 1 mL Diluent C) or transplanted without prelabeling. Unlabeled cells were detected in the murine tissues, using human-specific anti-CD34-FITC (Becton Dickinson) and anti-CD38-PE (Coulter) antibodies. Transplantation cell dose of CD34+-enriched cells was: 0.5-1 × 106cells/mouse; sorted CD34+CD38−/low cells: 2 × 105 cells/mouse (Figure 1B, R2); sorted CD34+CD38+/high cells: 8 × 105 cells/mouse (Figure 1B, R1). Cells were recovered from the BM, spleen, or lungs of transplanted mice at time points as indicated and were analyzed for the presence of either PKH26+ or human cells by flow cytometry acquiring 106 cells per sample (FACScalibur). Mouse immunoglobulin G (IgG) and human plasma were used to block Fc receptors. Cells obtained from nontransplanted mice and isotope control antibodies were used to exclude false positive cells. Propidium iodide staining was used to exclude dead cells. Where indicated, human CD34+-enriched cells were incubated prior to injection with either 10 μg/106 cells of a blocking mouse antihuman CXCR4 mAb (clone 12G5; Pharmingen, San Diego, CA), anti-VLA-4 (MCA697), anti-VLA-5 (MCA1187), or anti-LFA-1 (MCA1149) (Serotec, Oxford, United Kingdom). Pretreatment of human CD34+ cells with PTX (100 ng/mL, 60 minutes, 37°C; CalBiochem, La Jolla, CA) or chelerythrine chloride (5 or 10 μM, 30-60 minutes, 37°C; CalBiochem) was carried out prior to injection. Cell viability was 95%.

Fig. 1.

Rapid and selective homing of primitive CD34+CD38−/low cells.

(A) Human PKH26+ cells in the BM, spleen, and lungs of transplanted NOD/SCID mice recovered at indicated time points after transplantation. Data represent mean ± SE from 6 experiments, n = 5 mice/time point. (B) Enriched human CD34+ cells before transplantation. Gate R2 represents CD34+CD38−/low cells. (C,D) CD34/CD38 staining of cells recovered from the BM and spleen of mice that were transplanted with unlabeled CD34+-enriched cells. (C) NOD/SCID mice: (i) noninjected mouse, (ii) 2 hours after transplantation. (D) NOD/SCID/B2mnull mice: (i) noninjected mouse, (ii) 1 and 2 hours after transplantation. (E) CD34/CD38 staining of cells recovered from NOD/SCID mice, 16 hours after transplantation with sorted 2 × 105CD34+CD38−/low or 8 × 105 CD34+CD38+ cells. Data shown represent the number of human CD34+CD38−/low cells per 106 total acquired cells. The experiment was repeated 3 times, showing similar results (P < .05); a representative flow cytometry analysis is shown. Homing cell numbers are presented per 106 acquired cells.

Secondary transplantation

Irradiated NOD/SCID mice, used as primary recipients, were transplanted with 106 human CD34+ cells/mouse 24 hours after total body irradiation. Sixteen hours later, mice were killed, and BM and spleen were harvested. Secondary recipients, irradiated NOD/SCID/B2mnull mice, were transplanted either with BM (2 × 106 cells/mouse) or spleen cells (4.5 × 106 cells/mouse) from one primary recipient and killed 1 month later. Human cell engraftment was assayed by flow cytometry, using human-specific anti–CD45-FITC (Immuno Quality Products, Groningen, The Netherlands) and anti–CD19-PE (Coulter, Miami, FL) for detection of lymphoid cell differentiation. In addition, cells recovered from the BM or spleen of secondary recipients were assayed for human progenitors and the levels of human DNA.

Colony-forming unit assay

To detect human progenitors in the BM and spleen of transplanted mice, semisolid cultures were performed as previously described.36 Briefly, 5 × 105 cells/mL, harvested from mice at the indicated time points after transplantation, were plated in 0.9% methyl cellulose (Sigma), 15% FCS, 15% human plasma, 5 × 10−5M 2ME, 50 ng/mL SCF, 5 ng/mL interleukin 3 (IL-3), 5 ng/mL granulocyte macrophage colony-stimulating factor (GM-CSF; R&D), and 2 U/mL erythropoietin (Orto Bio Tech, Don Mills, Canada). These conditions are selective for human colonies. The semisolid cultures were incubated at 37°C in a humidified atmosphere containing 5% CO2 and scored 14 days later for myeloid, erythroid, and mixed colonies by morphologic criteria. Total colonies per murine BM were calculated, based on the assumption that tibias, femurs, humeri, and pelvis bones represent 30% of the total marrow cellularity.37

Human DNA analysis

The levels of human cell engraftment were detected as previously described.14 36 Briefly, high molecular weight DNA was obtained from the BM of transplanted mice by phenol/chloroform extraction. DNA (5 μg) was digested with EcoRI, subjected to electrophoresis on 0.6% agarose gel, blotted onto a nylon membrane, and hybridized with a human chromosome 17-specific α-satellite probe (p17H8) labeled with 32P. After digestion with EcoRI, this probe hybridizes a characteristic multisize band pattern, specific for human DNA. For quantification of human DNA in the samples, the intensity was compared with that in artificial mixtures of human and mouse DNA (0%, 0.1%, and 1% human DNA) run in parallel lanes. Multiple exposures of the autoradiographs were taken to ensure sensitivity down to 0.01% human DNA.

Migration assay

A total of 600 μL RPMI supplemented with 10% FCS containing 125 ng/mL SDF-1 was added to the lower chamber of a Costar 24-well transwell (Corning, NY). CD34+ cells (1-2 × 105) in 100 μL medium with and without PTX pretreatment or CD34+CD38−/low and CD34CD38−/low were loaded to the upper chamber (pore size 5 μm) and were allowed to migrate for 4 hours at 37°C. Migrating cells were collected from the lower chamber and counted for 30 seconds, using a FACSort (Becton Dickinson). Control spontaneous migration was performed without SDF-1 (in the lower chamber).

Adhesion assay

As previously described with minor modifications,38microplates (96 wells) were coated with 50 μL/well PBS containing 50 μg/mL human fibronectin (Chemicon International, Temecula, CA) and incubated overnight at 4°C. Wells were washed 3 times with PBS, blocked with 100 μL 1% bovine serum albumin in PBS for 3 hours at room temperature, and then washed with PBS. Human CD34+-enriched cells were labeled with 51Cr for 1 hour in adhesion medium (RPMI supplemented with 0.5% bovine serum albumin), washed twice, and then were incubated with 100 ng/mL PTX for 1 hour at 37°C or left untreated. Cells were added to the precoated wells, 100 × 103 cells/well in 100 μL adhesion medium. SDF-1 (250 ng/mL) or phorbol-12-myristate-13-acetate (PMA) (100 ng/mL) that was used as a positive control were added with the cells into the triplicate wells. Enriched CD34+ cells were allowed to adhere for 45 minutes at 37°C in a humidified atmosphere containing 5% CO2. The cells were washed 3 times with PBS to remove nonadherent cells. Adherent cells were removed with NaOH 1M + 0.1% Triton ×100 and were counted for radioactivity levels.


Rapid and efficient homing of human CD34+CD38−/low cells to the BM and spleen of immune deficient mice

The homing kinetics of human CB CD34+-enriched cells in transplanted NOD/SCID mice were studied. Mice injected with PKH26-labeled CD34+ cells (106 cells/mouse) were killed at different time points during the first 24 hours following transplantation (Figure 1A). PKH26+ cells were observed 4 hours after transplantation in the BM, spleen, and lungs of transplanted mice (Figure 1A). A slight increase in the number of cells accumulating in the BM and spleen was observed 24 hours after transplantation. In contrast, the number of PKH26+ cells accumulating in the lung rapidly declined (Figure 1A).

Recently we showed that the newly developed immune deficient NOD/SCID/B2mnull mice are better recipients for human SRC/stem cells due to the absence of β2 microglobulin, which results in lack of murine natural killer cell activity, leading to reduced rejection of the human cells.35 These mice were engrafted up to 11 times higher when low doses of human CB mononuclear cells were transplanted compared to NOD/SCID mice.8 To determine the homing pattern during the first 1-2 hours, we used NOD/SCID/B2mnull mice as recipients since the levels of homing CD34+-enriched cells in the BM of NOD/SCID mice during the first 2 hours were very low (Figure 1C). Mice were transplanted with 1 × 106 unlabeled CD34+cells and killed 1 to 2 hours later. BM and spleen cells were harvested and analyzed for the expression of the human cell surface markers CD34/CD38. Figure 1D shows the presence of human cells in murine tissues as early as 1 to 2 hours after transplantation. The levels of CD38 surface expression on CD34+ progenitors correlate well with their degree of differentiation. Primitive SRC/stem cells are either negative or express only low levels of CD38.8 14 In the BM of mice that were transplanted with the entire CD34+cell population expressing variable levels of CD38 (Figure 1B), the primitive human CD34+CD38−/low cells, which represent 20% of the total CB CD34+ population, were selectively accumulating (Figure 1Dii). Similar accumulation of CD34+CD38−/low cells was observed also in the spleen (Figure 1Dii). We further sorted the primitive CD34+CD38−/low (20% of the CB CD34+ population, Figure 1B, R2) and the more mature CD34+CD38+ (> 70% of the CD34+population, Figure 1B, R1) subpopulations and injected them into the mice. Although the small population of immature CD34+CD38−/low cells homed successfully to the BM and spleen (Figure 1E), the large population of differentiating CD34+CD38+ cells were not detected in the murine tissues (Figure 1E).

Other primitive hematopoietic populations such as CD34+CD38+ or CD34CD38Lin cells, which were previously shown to have limited repopulating potential,9 10 were also tested for their homing properties. Granulocyte colony-stimulating factor mobilized peripheral blood (MPB) CD34+ cells were incubated with SCF for 3.5 days to induce and enhance CXCR4 and CD38 expression. Expectedly, compared to day 0 cells, incubated cells demonstrated increased levels of CXCR4 and CD38 expression following 3.5 days of SCF stimulation (Figure 2Ai,ii). In correlation, the migration index to SDF-1 was also increased by 15-fold comparing day 0 cells to stimulated cells (3.4% and 51.5%, respectively;P < .05). Interestingly, in contrast to freshly isolated CB CD34+CD38+ cells that demonstrated poor homing capacities (Figure 1E, right panel), day 0 MPB CD34+CD38+ cells within freshly isolated MPB CD34+ cells homed successfully to the BM of NOD/SCID mice. Following SCF stimulation in vitro for 3.5 days, most of the homing cells express CD38 (2.6-fold increase; P < .05; Figure 2Aiii). In vivo and in vitro cytokine-stimulated CD34+CD38+ cells could only transiently engraft and were not detected 1 month after transplantation (data not shown). Together, these results suggest different homing and regulation patterns for nontreated versus cytokine-stimulated CD34+CD38+ cells.

Fig. 2.

Cytokine stimulation-induced homing of CD34+CD38+ cells.

(A) G-SCF MPB CD34+-enriched cells were incubated with SCF (50 ng/mL) for 3.5 days. Flow cytometry analyses of CD34/CXCR4 (i) and CD34/CD38 (ii) staining before and after 3.5 days of stimulation are shown. Homing CD34+CD38+ cells in the BM of transplanted mice (iii). A representative experiment of 3 performed is shown. (B) CD34/CXCR4 (i) and C34/CD38 (ii) status of purified CB CD34CD38Lin cells. Enriched CD34CD38−/low and CD34+CD38−/low cells before transwell migration (iii, see also Table 1).

With the use of the StemSep mAb cocktail for enrichment of primitive human progenitors, both CD34+CD38−/lowLin and CD34CD38−/lowLinsubpopulations were isolated together or separately (see “Materials and methods” section). Despite cell surface expression of CXCR4 on some purified CD34CD38 Lincells (Figure 2Bi), these cells have decreased migration potential toward SDF-1 (data not shown) and a relatively low repopulating potential.10 39 To compare the migration capacities of both subpopulations in parallel, enriched cells containing both fractions were prelabeled with anti-CD34 and anti-CD38 mAb prior to migration in transwells. Figure 2Biii shows a representative ratio of CD34CD38Lin and CD34+CD38Lin cells in the initial cell population before the migration assay was performed (12% and 36%, respectively). Although CD34CD38Lin cells showed higher spontaneous migration in the absence of SDF-1, they displayed significantly lower levels (6.5-fold decrease) of migration to a gradient of SDF-1 compared to CD34+CD38Lin cells (P < .05; Table 1). Homing CD34CD38Lin cells (50-500 × 103 cells/mouse) were not detected in the murine BM and spleen 24 hours after transplantation (by using anti-CD45/CD34 mAb labeling, data not shown). Together, the poor in vitro migration to SDF-1 and the lower homing capabilities can elucidate the low engraftment levels obtained by these cells.10 39

View this table:
Table 1.

CD34CD38Lincells demonstrate lower SDF-1–induced migration compared to CD34+CD38 cells

Homing CD34+CD38−/low cells contain SRCs with multilineage repopulation potential

Stem cells are defined by their multilineage repopulation and self-renewal capabilities in transplanted recipients. To assess the repopulation potential of CD34+CD38−/low cells that homed to the murine BM and spleen, human cells recovered from these organs were assayed for colony formation and repopulation of secondary recipients. Most of the colonies that developed from the immature homing cells were multilineage colony-forming unit (CFU)-granulocyte erythrocyte macrophage megakaryocyte (GEMM) and primitive blast colonies, whereas the more differentiated myeloid colonies were rare (Figure3A). This pattern of colony formation is indicative of the presence of very primitive progenitor cells. Primitive homing cells recovered from the BM or spleen of primary NOD/SCID recipients 16 hours after transplantation were further transplanted into secondary NOD/SCID/B2mnullrecipients. One month later, mice were killed, and the presence of human multilineage hematopoiesis was assayed by flow cytometry, using human-specific mAb for the pan-leukocyte marker CD45 and the pre–B-cell marker CD19, by Southern blot analysis, using a human-specific satellite probe, and in semisolid progenitor assays. Human lymphoid CD45+CD19+ cells were detected in the BM of secondary engrafted mice (Figure 3Bi,ii, gate R1). Furthermore, human cells recovered from these mice gave rise to both erythroid and myeloid colonies, demonstrating multilineage repopulation by human SRCs (Figure 3C) The engraftment of secondary recipients was also detected by testing the presence of human DNA in the BM of transplanted mice (Figure 3D).

Fig. 3.

Homing cells are SRC/stem cells.

(A) BM and spleen cells of mice that were transplanted with human CD34+-enriched cells were harvested at indicated time points and assayed for colony formation. Data represent mean ± SE from 6 experiments, n = 12 mice. (B) NOD/SCID mice used as primary recipients were transplanted with 106 CD34+cells/mouse. Secondary recipients NOD/SCID/B2mnull mice were transplanted either with BM (2 × 106 cells/mouse) or spleen cells (4.5 × 106 cells/mouse) from one primary recipient. BM cells of secondary-transplanted recipients were assayed for lymphoid (R1), nonlymphoid (R2) cell differentiation (B), erythroid and myeloid progenitor cells (C), and human DNA levels (D). A representative experiment is shown of 3 performed, n = 12 mice.

Human CD34+-enriched cells home to the murine BM and spleen in a CXCR4-dependent manner

Recently we reported that human SRC/stem cell engraftment of NOD/SCID mice is dependent on the expression of the chemokine SDF-1 by the transplanted mice and its receptor CXCR4 by human SRCs.14 To test the role of this chemokine and its receptor in the homing process, enriched CB CD34+ cells were treated with neutralizing mAb against human CXCR4. Anti-CXCR4 pretreatment significantly reduced the number of human cells in the BM or spleen 4 to 8 hours following transplantation (Figure4A). Significant reduction in the number of human progenitor cells was also observed (Figure 4B). However, no significant differences in the number of cells accumulating in the lungs were found following treatment with antibodies to CXCR4 (Figure 4A).

Fig. 4.

Selective homing of CD34+ cells in a CXCR4-dependent manner.

PKH26-labeled CD34+ cells were injected into NOD/SCID mice, either untreated or after blocking with 10 μg mAb for human CXCR4 per 106 cells, as indicated. (A) The number of homing PKH26+ cells in the murine tissues 4 and 8 hours after transplantation is shown. Cells from the same organs of noninjected mice were used as a negative control. (B) Human progenitor cells in the BM or spleen of transplanted mice 8 hours after transplantation. Number of human colonies per total BM of one mouse. Data represent mean ± SE from 6 experiments, n = 9 mice/group, *P < .05. (C) Nonirradiated NOD/SCID mice were injected with 1 μg human SDF-1 directly into the BM of the femur (upper panel) or into the spleen (lower panel). Mice were further transplanted immediately with 1 × 106 CD34+ cells/mouse either untreated (ii) or preincubated with antihuman CXCR4 mAb (iii). Noninjected organs were used as a negative control (i). Data indicate number of human cells per 106 acquired cells. A representative experiment from 3 performed is shown. Similar results (P < .05) were obtained in all 3 experiments performed, n = 12 mice. (D) Enriched CD34+ cells were incubated for 48 hours with 50 ng/mL SCF+IL-6 and assayed for migration levels (i). NOD/SCID mice were transplanted with stimulated or unstimulated cells. Murine BM and spleen were harvested 16 hours later and assayed for the growth of human colonies (ii). Data shown is from 4 experiments, n = 8 mice.

Recently we reported that total body irradiation increases SDF-1 expression by stromal cells, mostly by immature osteoblasts and endothelial cells in the BM.40 To directly test the potential of human SDF-1 to attract SRC/stem cells in vivo, nonirradiated NOD/SCID mice were injected with human SDF-1 directly into the BM of one femur or into the spleen. Noninjected organs were used as controls. Mice were transplanted with human CB-enriched CD34+ cells immediately after SDF-1 injection, and the level of homing cells to the SDF-1 gradient was quantified 4 hours later. Human SDF-1 increased the number of primitive CD34+CD38−/low cells homing into the BM (Figure 4C, upper panel, ii) compared to the control mice (i). Moreover, when CD34+-enriched cells were preincubated with anti-CXCR4 mAb, homing was reduced to the basal levels despite the local administration of human SDF-1 (iii). Similar results were obtained when SDF-1 was injected into the spleen (Figure 4C, lower panel). Short-term in vitro incubation of human CD34+ or CD34+CD38−/low cells with SCF together with IL-6 for 48 hours induced increased surface CXCR4 expression and migration toward SDF-1, as well as higher levels of engraftment compared to untreated cells.14 To test whether the homing pattern of CD34+ cells is also affected by cytokine stimulation, cells were incubated with SCF+IL-6 for 48 hours and were then assayed for their migration and homing potential. A 3.5-fold increase in the level of migration was induced following 48-hour stimulation with SCF+IL-6 compared to untreated cells (Figure 4Di;P < .05). In addition cells that were stimulated with SCF+IL6 were further transplanted into NOD/SCID mice. Cytokine stimulation increased the homing levels of primitive colony-forming cells by 4.6-fold in the BM and by 3-fold in the spleen 16 hours after transplantation (Figure 4Dii; P < .05). More importantly, there was also an increase in the relative number of primitive multilineage CFU-GEMM colonies (Figure 4Di-ii).

Major integrins facilitate the homing of CD34+cells

Previously we showed that engraftment of NOD/SCID mice by human SRCs requires activation of the major integrins VLA-4, VLA-5, and, to a lesser degree, LFA-1.28 The direct effect of these integrins on the homing process was, therefore, studied. Transplanted cells were preincubated with specific antibodies prior to transplantation as previously described.28 As expected, homing into the BM was significantly reduced by blocking VLA-4 (37% ± 10%), VLA-5 (51% ± 10%), and LFA-1 (58%±16%) compared to control nontreated cells (Figure5A;P < .05). Homing into the spleen showed similar results (Figure 5A; P < .05).

Fig. 5.

Homing of CD34+ cells is facilitated by the major integrins VLA-4, VLA-5, and LFA-1 and is PTX insensitive.

(A) Enriched CB CD34+ cells (0.5 × 106cells/mouse) were incubated with antibodies specific for the major integrins prior to transplantation into NOD/SCID/B2mnull mice. Data shown represent mean ± SE from 5 experiments, n = 22 mice,P < .05. (B) CD34+-enriched cells were incubated with PTX (100 ng/mL, 60 minutes, 37°C) and assayed for migration toward SDF-1 (Bi, control cells were not treated with PTX), or SDF-1–mediated adhesion to fibronectin (FN) (Bii, control = adhesion to FN in the absence of SDF-1), or homing levels in transplanted NOD/SCID mice (C, 1 × 106 cells/mouse). CD34+ cells were treated with PTX alone or followed by incubation with anti CXCR4 mAb; control cells were not treated. Data represent the number of human cells per 106 total acquired cells; a representative experiment from 3 performed is shown. Results from all experiments gave similar results (P < .05; n = 9 mice).

Homing of human CD34+ cells to the BM is PTX insensitive and depends on the activation of protein kinase C

PTX, an inhibitor of signal transduction mediated by the αi subunit of G proteins, almost completely abrogated in vitro migration of human CD34+ cells toward SDF-1.11 In contrast, in vivo experiments performed with PTX-pretreated murine stem and progenitor cells showed only a delayed engraftment of the spleen and no change in BM repopulation.34 In this study we confirmed that PTX can significantly inhibit transwell migration of human CD34+-enriched cells in response to SDF-1 (Figure 5Bi). Surprisingly, we found that instead of reducing homing of human CD34+-enriched cells to the BM and spleen of transplanted mice, pretreatment of these cells with PTX led to increased homing (Figure 5C). However, additional pretreatment of the cells with antibodies to CXCR4 prevented the homing of PTX-stimulated cells (Figure 5C). Unexpectedly, pretreatment of CD34+-enriched cells with PTX also did not inhibit the adhesion of these cells to fibronectin in response to SDF-1 (Figure 5Bii). Laudanna et al41 found that chelerythrine chloride, a broad-range protein kinase C (PKC) inhibitor, blocks both CXCR2-mediated adhesion and chemotaxis of neutrophils while studying another CXC chemokine, IL-8. The involvement of the PKC pathway in SDF-1/CXCR4 signaling is suggested since chelerythrine chloride significantly inhibited SDF-1–induced migration and adhesion.42 Moreover, pretreatment of human CD34+ cells with chelerythrine chloride efficiently inhibited the homing of about 73% of human cells into the BM and about 80% of the homing into the spleen (Figure6A,B; P < .05).

Fig. 6.

PKC-mediated homing of CD34+ cells.

CD34+ cells were incubated with 10 μM chelerythrine chloride for 30 minutes in 37°C and were transplanted into NOD/SCID mice (1 × 106/mouse) 4 hours after sublethal irradiation. Data represent number of human cells per 106acquired cells 16 hours after transplantation. A representative experiment (A) and the mean ± SEM of 3 experiments,P < .05 (B) are shown.


Homing of hematopoietic stem cells to the BM microenvironment is essential for the development of blood formation in the developing embryo. At present, homing of stem and progenitor cells, detected shortly after transplantation, was studied mainly in the murine system, but no specific mechanism for this selective process was described.30-32 Zanjani et al29 used the fetal sheep model to investigate VLA-4–dependent homing of human CD34+ cells. However, due to the lack of human cell detection within a few hours after transplantation, donor cells were monitored only after 24 and 48 hours. We demonstrate that the homing process in which human SRCs migrate from the blood circulation of sublethally irradiated immunodeficient NOD/SCID and NOD/SCID/B2mnull mice into the BM and spleen is very rapid. Primitive human CD34+CD38−/low cells were found in the BM and in the spleen of NOD/SCID/B2mnull mice, already 1 to 2 hours after transplantation. Interestingly, it appears that homing into the spleen is faster since only a small minority of the cells were obtained in the BM within 1 hour. Szilvassy et al30 have also found higher levels of progenitor and homing cells in the spleen than in the BM 1 to 3 hours after transplantation of murine stem cells. Observing the same pattern, Papayannopoulou et al43 found a corresponding decrease in donor progenitor numbers in both spleen and peripheral blood, accompanied by increased numbers in the BM, 40 hours after transplantation of murine BM cells. An interesting mechanism is suggested by Vermeulen et al19, showing that VLA-4/VCAM-1 is involved in the homing of CFU-S (spleen) only into the BM, whereas CD44 is involved in CFU-S lodging into both BM and spleen. Therefore, activation of different adhesion molecules by SDF-1 may explain the differential accumulation of human SRCs in the murine BM and spleen.

In the present study we demonstrate that within the nonstimulated immature human CB CD34+ cell population, homing cells are exclusively primitive CD34+CD38−/lowCXCR4+ cells. The early differentiation status of homing cells was demonstrated by transplanting as few as 2 × 105 highly purified CB CD34+CD38−/low cells, which homed successfully, in contrast to 8 × 105CD34+CD38+/high cells that could not be detected in the murine BM and spleen. Moreover, homing CD34+CD38−/low cells recovered from the BM and spleen of primary transplanted mice could repopulate secondary transplanted NOD/SCID/B2mnull recipients with both lymphoid and myeloid cells, demonstrating that they are true SRC/stem cells. Previous reports also determined SRCs as CD34+CD38 cells7 or as CD34+CD38−/low cells,14although in some studies it has been suggested that the more mature CD34+CD38+ cells also have low repopulating potential.9 Different migration and homing capacities were observed when CXCR4 and CD38 expression was induced by in vivo, or in vitro cytokine stimulation as shown in this report, using MPB-enriched CD34+ cells. In addition, we showed recently that ex vivo cytokine-stimulated CB CD34+CD38+ cells had increased CXCR4 expression as well as improved migration capacities toward a gradient of SDF-1 and also could transiently home to the BM of NOD/SCID mice but failed to durably repopulate it for 1 month.44 The chemokine SDF-1 was shown to attract immature human CD34+ cells in vitro.11-13 Primitive CD34+CD38−/low cells express higher levels of CXCR4 compared to CD34+CD38+/high cells.45 A positive correlation between CXCR4 expression and SDF-1–induced transendothelial migration of leukemic cells was demonstrated.12 In correlation with these results, we showed that freshly isolated CB CD34+CD38−/low cells migrate better to SDF-1 compared to more mature CD34+CD38+cells,14 which also secrete low levels of SDF-1.46 We further demonstrated that human SRC/stem cell engraftment of NOD/SCID mice is dependent on CXCR4/SDF-1 interactions.14 Therefore, the involvement of SDF-1/CXCR4 interactions in SRC/stem cell homing is of interest. Blocking of CXCR4 signaling on transplanted CD34+-enriched cells prevented homing, whereas pretreatment of cells with cytokines lead to up-regulation of CXCR4 expression, which increased both in vitro migration to SDF-1 and in vivo homing. CXCR4-dependent homing was observed in the BM and spleen of engrafted mice but not in the lungs. Several studies have shown that murine hematopoietic progenitors are found in many tissues (lung, liver, kidney, BM, and spleen) during the first few hours after transplantation. However, their presence in nonhematopoietic tissues is transient, and by 48 hours they are retained mostly in the hematopoietic microenvironment, ie, BM and spleen,43 suggesting a chemokine-mediated selective and specific homing process.

CXCR4 is a 7-transmembrane receptor coupled to a PTX-sensitive Gαi-protein.33 In vitro migration toward SDF-1 by human CD34+ cells was blocked by PTX pretreatment,11 whereas in vivo BM engraftment by murine stem and progenitor cells was not affected, and only delayed repopulation was observed in the spleen.34 We found similar results with human CD34+ cells: PTX inhibited in vitro migration toward SDF-1 and unexpectedly did not impair the homing of the cells. However, incubation of PTX-pretreated cells with anti-CXCR4 mAb prior to transplantation inhibited homing of these cells. Unexpectedly, SDF-1–induced adhesion of CD34+ cells to fibronectin was also not inhibited by PTX, suggesting a major role for SDF-1–induced adhesion in the homing process. In addition PTX failed to inhibit SDF-1–mediated internalization of CXCR4 on human T cells, demonstrating additional Gαi-independent signaling pathways by SDF-1.47

In the BM, SDF-1 is mainly produced by immature bone-forming osteoblast cells and is specifically expressed on human40 as well as on murine23 BM endothelium in vivo. We further found that SDF-1 can stimulate integrin-mediated arrest of immature human CD34+ cells on vascular endothelium under physiologic shear flow,24 suggesting interactions between BM endothelium-expressed SDF-1 and transplanted CD34+CXCR4+ cells, promoting adhesion to endothelial ligands such as VCAM-1 and activating the major integrins LFA-1, VLA-4, and VLA-5 present on CD34+ cells that are essential for engraftment by SRC/stem cells.28

Total body irradiation is widely used clinically as well as in experimental models as a crucial conditioning procedure preceding stem cell transplantation. We have found that the expression of SDF-1 increased following conditioning with DNA-damaging agents (ionizing irradiation and 5-fluorouracil) and correlated with an increase in CXCR4-dependent engraftment by human SRC/stem cells transplanted into NOD/SCID mice.40 To test the direct potential of human SDF-1 on homing of CD34+-enriched cells, nonirradiated NOD/SCID mice were injected with human SDF-1 directly into the spleen or into the BM of one femur. SDF-1 increased the number of homing CD38−/lowCXCR4+ SRCs to the spleen and BM in a CXCR4-dependent manner, suggesting that human SDF-1 can also attract SRC/stem cells in vivo. Homing requires integrin-mediated adhesion interactions that are inhibited by blocking antibodies specific for the major integrins VLA-4, VLA-5, and LFA-1 (Figure 5A). Recently, Frenette et al demonstrated that both E and P selectins, which are expressed on the BM endothelium, are essential for the homing of murine hematopoietic progenitor cells17 and for their retention within the BM.48 With the use of knockout mice and blocking antibodies, the activity of both selectins was shown to be essential for successful homing. Naiyer et al49 showed that SDF-1–induced transendothelial migration of CD34+ cells is mediated via E selectin. P-selectin glycoprotein ligand 1 (PSGL-1), the sole P selectin receptor expressed on hematopoietic progenitor cells,50 can also bind to E51 and L selectins.52 Providing a more complicated interplay, HCLL, another L selectin ligand that is distinct from PSGL-1, was identified on human hematopoietic progenitors.53 This network of redundant receptor/ligand interactions prevents direct inhibition of homing by blocking specific selectin ligands. Therefore, future studies are needed to determine the relative role of all selectin ligands in the homing process by creating selectin-deficient NOD/SCID mice. Greenberg et al54 found that human BM CD34+CD38 cells roll more efficiently on E, P, and L selectins than CD34+CD38+ cells under physiologic shear flow. That study characterized a correlation between maturity and rolling efficiency, which strongly supports the superior homing potential of primitive human CD34+CD38−/low cells, observed in our model.

The PKC signaling pathway was shown to be essential in the migration and adhesion of neutrophils induced by the CXC chemokine IL-8 and the chemoattractant formyl-Met-Leu-Phe (fMLP), since both activities were blocked in a dose-dependent manner by chelerythrine chloride, a broad-range PKC-specific inhibitor.41Therefore, the inhibitory effect of chelerythrine chloride on homing of human enriched CD34+ cells into the BM and spleen was studied. In addition to blockage of SDF-1–induced chemotaxis,42 chelerythrine chloride inhibited the homing of both human CD34+ cells (as reported in this study) and, moreover, murine Sca-1+Lin stem and progenitor cells,42 which represent a more physiologic relevant, syngeneic model for homing.

Our findings provide evidence for the involvement of the PKC signaling pathway in the homing process of both human and murine stem and progenitor cells. Our results demonstrate a major role for SDF-1 and CXCR4 in the homing of human SRC/stem cells into the BM and spleen of immunodeficient mice and suggest a novel approach to improve the outcome of clinical stem cell transplantation and to enhance homing and repopulation by prestimulation with cytokines.55 56


We thank Raanan Margalit for his professional assistance with direct in situ injections.


  • Tsvee Lapidot, Dept of Immunology, The Weizmann Institute of Science, Rehovot 76100, Israel; e-mail:Tsvee.Lapidot{at}

  • Supported in part by grants from the Israel Science Foundation, the Ares Serono group, the Rich Foundation, and CONCERN Foundation. T. Lapidot is Incumbent of the Pauline Recanati Career Development Chair of Immunology.

  • The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 U.S.C. section 1734.

  • Submitted August 9, 2000.
  • Accepted November 30, 2000.


View Abstract