Expression and Function of Homing-Essential Molecules and Enhanced In Vivo Homing Ability of Human Peripheral Blood-Derived Hematopoietic Progenitor Cells after Stimulation with Stem Cell Factor



Hematopoietic stem cell (HSC) homing from blood to bone marrow is a multistep process involving rolling, extravasation, migration, and finally adhesion in the correct microenvironment. With view to the hematopoietic recovery after clinical stem cell transplantation, we investigated the effect of stem cell factor (SCF) on the expression and the adhesive function of the α4β1 and α5β1 integrins very-late antigen (VLA)-4 and VLA-5 on peripheral blood-derived hematopoietic progenitor cells. After SCF stimulation, the expression of VLA-4 and VLA-5 on CD34+/c-kit+ cells obtained from healthy donors increased from 54% to 90% and from 3% to 82%, respectively. For patient-derived cells, the increase was 67% to 90% and 12% to 46%. The proportion of mononuclear cells adhering to the fibronectin fragment CH296 increased by stimulation with SCF from 14% to 23%. Accordingly, functional studies showed an approximate 30% increase of adherent long-term culture-initiating cell. The improvement of the homing abilities of SCF-stimulated HSC was confirmed by transplantation into sublethally irradiated nonobese diabetic-scid/scid mice. Six weeks after the transplantation, in eight of eight animals receiving human HSC with the addition of SCF, a profound multilineage hematopoietic engraftment was detected, whereas in the control group receiving only HSC, none of eight animals engrafted. Our data provide the first in vivo evidence that stimulation with cytokines improves the homing ability of transplanted human hematopoietic progenitor cells.


The transplantation of peripheral blood-derived hematopoietic stem cells (HSCs) has become a routinely used procedure for the treatment of hematological diseases. In heavily pretreated patients, the harvest of sufficient numbers of CD34+ cells can be difficult even after mobilization with chemotherapy and granulocyte-colony stimulating factor (G-CSF) [1]. In vitro studies revealed that the retention of hematopoietic progenitor cells (HPCs) in the bone marrow is primarily mediated via adhesion to extracellular matrix proteins by the β1 integrins very-late antigen (VLA)-4 and VLA-5 [2]. Bone marrow CD34+ cells express under steady-state conditions a variety of adhesion molecules, such as the α4, α5, β1, CD11a/CD18, and CD11b/CD18 integrins, L-selectin, platelet and endothelial cell adhesion molecule (PECAM)-1, and CD44 [37]. It has been shown that the expression or function of the β1 integrins can be modulated by various cytokines; in HL60 cells, the expression of the α4 domain was reduced after incubation with interleukin (IL)-3, IL-6, IL-11, or stem cell factor (SCF) [8]. The adhesive function of VLA-4 and VLA-5 on TF-1 and Mo7e cells to fibronectin increased after stimulation with granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-3, or SCF, although the expression of these integrins on the cell membrane was not altered [9]. In vivo studies using HPCs treated with anti–VLA-4 antibodies before transplantation showed a significantly reduced engraftment ability of the transplanted cells [10]. Phenotypic analysis and transplantation of human HPCs in nonobese diabetic-severe combined immune deficiency (scid)/scid (NOD/SCID) mice indicated that in addition to VLA-4, VLA-5 is also expressed on repopulating cells [11]. In mice, it was shown that HPCs found in the blood after treatment with cyclophosphamide and G-CSF express significantly lower levels of α2, α4, and β1 integrins, correlating with a 50% decrease in their ability to home to the bone marrow [12].

Based on the hypothesis that retention of HPC in hematopoietic organs is controlled by adhesive interactions, we reasoned that increasing the adhesiveness of peripheral blood-derived HPCs might enhance the proportion of the cells reaching and residing within the bone marrow. In the setting of HSC transplantation, this approach might improve the homing ability of the transplanted cells into the bone marrow of the recipient. The experiments described here were designed to test this hypothesis. First, we sought to identify progenitor cell populations that possibly were sensitive to stimulation with SCF by assessing the expression of c-kit. Then CD34+ cell–enriched suspensions of the peripheral blood of G-CSF–treated healthy volunteers or chemotherapy-treated cancer patients were cultured in the presence of SCF, and the expression and function of VLA-4 and VLA-5 were monitored. We used the long-term culture on murine stroma cells to detect and enumerate early human HPCs, referred to as long-term culture-initiating cells (LTC-ICs), to test the effect of SCF on the adhesive capabilities of these cells. Finally, the improvement of the in vivo homing abilities of SCF-stimulated HSCs was confirmed by transplantation into sublethally irradiated NOD/SCID mice.

Materials and Methods

Progenitor Cell Preparation

Human HPCs were either aliquots of leukapheresis products obtained from healthy individuals donating G-CSF–mobilized CD34+ cells for allogeneic transplantation or were remaining backup samples of cryopreserved leukapheresis products from patients after autologous stem cell transplantation. For both, approved institutional procedures, including written informed consent from each patient, were followed. The light-density (<1.077 g/cm3) cells were isolated from the samples by centrifugation on Ficoll-Hypaque (Pharmacia, Uppsala, Sweden) and then cryopreserved in fetal calf serum (FCS; Gibco BRL, Invitrogen GmbH, Paisley, U.K.) with 10% DMSO (Sigma, St. Louis, MO) at −180°C. Cells were subsequently thawed, and cell populations expressing mature erythroid, granulopoietic, megakaryopoietic, and lymphoid markers were removed using a StemSepTM column (Stem Cell Technologies, Vancouver, Canada), resulting in a lineage-depleted (lin) cell fraction enriched for CD34+ cells. For some experiments, a positive selection of CD34+ cells using an IsolexTM 300i column (Nexell Therapeutics, Irvine, CA) was performed.

Progenitor Cell Assays

Colony-forming cell (CFC) assays were performed in methylcellulose cultures (MethoCultTM GF H4434, Stem Cell Technologies) as described [13]. LTC-ICs were assayed using a 6-week CFC end point after maintenance of the cultures on murine fibroblast feeders [14]. To determine the frequency of LTC-ICs, the cells were plated under limiting dilution conditions in 96-well plates. After 6 weeks with weekly half-medium changes, the cells were plated in methylcellulose using one individual 35-mm Petri dish for each well and assessed for their CFC content. Wells that contained at least one CFC were scored positive, and the frequency of LTC-ICs was calculated using the L-calc® program (Stem Cell Technologies).

Flow Cytometric Analysis and Cell Sorting

CD34+ and CD34+CD117+ cells were isolated from lin HPCs after staining with antihuman CD34-APC or CD34-FITC (8G12, Becton, Dickinson and Company, San Jose, CA) and CD117-PE (Dianova, Hants, U.K.), washed twice with phosphate-buffered saline (PBS; Gibco BRL, Invitrogen GmbH), and resuspended in PBS containing propidium iodine (PI, 0.5 μg/ml; Sigma) to allow for dead cell discrimination. The expression of VLA-4 and -5 was determined after staining with anti-CD49d (BU 49, Dianova) and anti-CD49e (SAM-1, Dianova). Fluorescence-activated cell sorter (FACS) analysis was performed using a FACS Calibur (Becton, Dickinson). Cell sorting was done on a FACStarPlus (Becton, Dickinson) equipped with two lasers (488 nm, UV), an automatic cell deposition unit, and a sort enhancement module. Region and gates were defined in the fluorescence channel 1 (FL1) through FL3 and FL1 through FL2 fluorescence dot blot diagrams defining viable CD34+-FITC/CD117+-PE and their CD117 counterparts for cell sorting.

Cell Culture Conditions and Cytokine Stimulation

After immunomagnetic enrichment of CD34+ cells, an aliquot of the cells was stained for flow cytometry. The remaining cells were plated in serum-free medium (Cell-GroTM SCGM, BioWhittaker Europe, Verviers, Belgium) containing 10% FCS into Petri dishes (Greiner BioOne, Longwood, FL) at 105 cells/ml. Then SCF (100 ng/ml; Pepro Tech, Northampton, U.K.) or SCF plus genistein (100 μmol; Sigma), an inhibitor of tyrosine kinases [9], or genistein alone was added. After 24 hours, the cells were harvested and analyzed by flow cytometry.

Cell Adhesion Assay

CD34+ cells were sorted and then stained with calcein by incubation for 45 minutes with PBS containing 2.5 μl calcein-AM per ml. The cells were then washed twice and plated in nontissue culture-treated 96-well Petri dishes coated with 4 μg/cm2 of the fibronectin fragment CH-296. Bovine serum albumin (BSA) was used as a negative control. After 2 hours at room temperature, the remaining fluid was removed and the plates were blocked with Roswell Park Memorial Institute (RPMI) medium containing 2% BSA for another 2 hours at 37°C. Before use, the plates were washed three times. After adding the cells, the plates were centrifuged for 5′ at 1,000 RPM, subsequently read on a fluorescence reader (Spectrafluor Plus, Tecan, Crailsheim, Germany), and incubated for 30′ at 37°C. Then the nonadherent cells were removed, the plates were rinsed, and the fluorescence emission was determined again. The proportion of adherent cells was calculated by linear regression analysis using a standard curve. For some experiments, 24-well plates were used, and the adherent and nonadherent cell fractions were counted after the harvest and plated into functional progenitor cell assays to determine the distribution of CFCs and LTC-ICs between the various cell fractions.

NOD/SCID Mouse Reconstitution Assay

Within 24 hours before transplantation, 6- to 8-week-old NOD/SCID mice were irradiated sublethally using a linear accelerator (Primus, Siemens, FRG), receiving a total dose of 200 cGy. Lin cells from healthy donors containing 3.0 × 105 and 3.6 × 105 CD34+ cells were injected intravenously in 0.2-ml volumes either with or without the addition of 10 μg human SCF (Pepro Tech) per mouse. Upon irradiation, ciprofloxacin (0.1 mg/ml) was added to the drinking water. Six weeks later, the bone marrow of the animals was analyzed. For the red blood cell lysis, the bone marrow cells were incubated on ice for 30 minutes with NH4CL-Puffer. To block human and mouse Fc receptors, aliquots of the cells (5 × 105) were treated with human serum and an anti-mouse immunoglobulin G (IgG) antibody (Becton, Dickinson). Then the cells were stained using combinations of fluorescence-conjugated antibodies that stained human lymphoid cells (CD19/20-PE-CD34-APC, Beckton, Dickinson), and human myeloid cells (CD15(IgM)/66b-FITC-CD45/71-PE, Beckton, Dickinson). As control, bone marrow cells of mice that did not receive human cells were stained with the identical antibodies. After incubation, the cells were washed twice with PBS, one washing step containing PI (Sigma). Engraftment was defined by the detection of at least five lymphoid and five myeloid human cells detectable within 2 × 104 cells analyzed. To control the staining, aliquots of the cells were incubated with irrelevant isotype-matched control antibodies, directly labeled with the identical fluorochrome. FACS analysis was performed using a FACS Calibur (Becton, Dickinson).

DNA Analysis

DNA of the bone marrow cells was isolated using a QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. DNA 12.5 ng was used for polymerase chain reaction (PCR) for human 5′ actin using the following primers: 5′-GCTCACCATGGATGATGATATCGC-3′, 5′-GGAGGAGCAATGATCTTGATCTTC-3′. The PCR was performed using a Hot Star Taq Master Mix Kit (Qiagen) according to the manufacturer's instructions and was carried out in a programmable PTC-200 Peltier Thermal Cycler (Biozyme, Oldendorf, Germany) with the following conditions: 94°C for 15 minutes and then 30 cycles each comprising 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 1 minute. After the PCR was completed, the tubes were kept for 10 minutes at 72°C and then 4°C. The negative controls included a control without template. The 5′ actin PCR product's size is 1 kB and was analyzed on 1% agarose gel. The size of the PCR fragments was estimated using a 100-bp DNA ladder (Gibco BRL).


Isolation and Functional Characterization of CD34+/c-kit+ and CD34+/c-kit Cells

The immunomagnetic enrichment of CD34+ cells from the leukapheresis products of chemotherapy and G-CSF–treated tumor patients resulted in a lineage-depleted cell fraction containing on average 80 ± 8% CD34+ cells with a 62% recovery of the CD34+ cells after the enrichment process (n = 10). A total of 36% of the CD34+ cells was additionally c-kit+ (n = 3). To assess the functional capacity of these cells, the CD34+/c-kit+ and the respective c-kit fraction obtained from seven leukapheresis products were sorted and plated into CFC and LTC-IC assays. As shown in Table 1, no colony growth was detected in three specimens. In two of those samples, the LTC-IC frequency was too low (<1 in 7,800 and <1 in 28,800 cells) to give rise to any colonies within the c-kit+ cell fraction, but LTC-ICs were detectable within the c-kit fraction. In four experiments, the CFC content of the c-kit+ and the c-kit cells could be compared and showed a 2.2-fold higher number of CFCs within the c-kit+ fraction (p = .014). The LTC-IC frequency was calculated from the pooled data of six experiments and was 2.8-times higher within the c-kit fraction compared with the c-kit+ fraction.

Table Table 1.. Functional characterization of mobilized tumor patient (MTP)–derived CD34+/c-kit+ and CD34+/c-kit cells
  • a

    The c-kit+ and c-kit fractions of CD34+ cells from patients treated with chemotherapy plus G-CSF were sorted and then analyzed for the content of CFCs and LTC-ICs. The number of CFCs was scored after 14–16 days in culture in semisolid medium as described and is given per 102 cells plated. Mean and SEM for CFCs are calculated from experiments 3 and 5 through 7 (p = .014). The LTC-IC frequency was calculated using the L-calc program.

  • a

    aMean of the pooled data from the experiments 1 through 5 and 8.

  • c

    Abbreviations: CFC, colony forming cells; G-CSF, granulocyte colony-stimulating factor; LTC-IC, long-term culture-initiating cell; MTP, mobilized tumor patient; SEM, standard error of the mean.

Experiment (n)c-kit+c-kitc-kit+c-kit
1<1<11 in >7,8001 in 924
2<1<11 in >28,8001 in 6,197
325151 in 40,5251 in 77
4<1<11 in 11,5121 in 23,288
5951 in 9741 in 200
6142Not doneNot done
7187.5Not doneNot done
8Not doneNot done1 in 3,7351 in 481
Mean ± SEM16.5 (± 3.4)7.4 (± 2.8)1 in 7,229a1 in 2,550a

Flow Cytometric Analysis of c-kit, VLA-4, and VLA-5 on CD34+ Cells

Flow cytometry was performed with freshly isolated or cultured cells. Before culture, 14 ± 7% and 36 ± 4% of the CD34+ cells from healthy donors and chemotherapy-treated tumor patients were c-kit+. After culture, c-kit expression was reduced for both (Fig. 1). More than 50% of the CD34+ cells were spontaneously VLA-4+, and the proportion increased approximately 1.5-fold to 90% VLA-4+ cells after 24-hour SCF stimulation. The spontaneous expression of VLA-5 on CD34+ cells was remarkably lower and increased by a factor of 27 for cells from healthy donors and by a factor of 4 for patient-derived cells (Table 2) after SCF stimulation. The effect of SCF could be blocked by the addition of the tyrosine kinase inhibitor genistein (Fig. 2).

Table Table 2.. Stimulation with stem cell factor (SCF) increases the surface expression very-late antigen (VLA)-4 and VLA-5 on CD34+ cells
  1. a

    The expression of c-kit, VLA-4, and VLA-5 was measured by flow cytometric analysis on G-CSF–mobilized CD34+ cells obtained either from healthy donors or patients with tumors after chemotherapy. Indicated is the proportion (% positive cells) of CD34+ cells expressing c-kit, VLA-4, and VLA-5 and the fraction of CD34+/c-kit+ and CD34+/c-kit cells additionally expressing VLA-4 and VLA-5, respectively. In addition, for each cell population, the MSI is shown. MSI of the isotype control was 2 and 1 for healthy donors and 8 and 4 for mobilized patients for the FITC and PE channel, respectively. For all antigens, the surface expression is documented before and after stimulation with 100 ng/ml SCF for 24 hours. The data are given as mean ± standard error of the mean from three (healthy donors) or four (mobilized patients) independent experiments.

  2. b

    Abbreviations: FITC, fluorescein isothiocyanate; G-CSF, granulocyte colony-stimulating factor; MSI, mean staining intensity; PE, phycoerythrin.

 Before SCF (% positive)After SCF (%positive)Before SCF (%positive)After SCF (%positive)Before SCF (%positive)After SCF (% postive)Before SCF (%positive)After SCF (%positive)
CD34+/c-kit+14 ± 73 ± 27 ± 14 ± 136 ± 420 ± 412 ± 17 ± 2
CD34+/VLA-4+29 ± 477 ± 216 ± 127 ± 457 ± 1265 ± 822 ± 228 ± 4
CD34+/VLA-5+1 ± 031 ± 127 ± 117 ± 311 ± 1022 ± 1010 ± 212 ± 1
c-kit+/VLA-4+54 ± 790 ± 520 ± 242 ± 767 ± 990 ± 324 ± 335 ± 7
c-kit+/VLA-5+3 ± 182 ± 18 ± 130 ± 412 ± 1046 ± 1510 ± 216 ± 5
c-kit/VLA-4+27 ± 577 ± 715 ± 127 ± 448 ± 1462 ± 820 ± 224 ± 3
c-kit/VLA-5+1 ± 130 ± 147 ± 116 ± 311 ± 1012 ± 310 ± 29 ± 1
Figure Figure 1..

Representative flow cytometric dot blot diagram showing the expression of c-kit, VLA-4, and VLA-5 on CD34+ cells from a healthy donor before and after stimulation with SCF for 24 hours. The percentage of the cells is indicated in the corner of the respective quadrants. Abbreviations: SCF, stem cell factor; VLA, very-late antigen.

Figure Figure 2..

CD34+ cells were stimulated with SCF with and without the addition of genistein for 24 hours and then analyzed by flow cytometry. Shown is the proportion of CD34+/CD117+ cells expressing VLA-4 and VLA-5. Mean values ± standard error of the mean from three experiments are documented. Abbreviations: SCF, stem cell factor; VLA, very-late antigen.

Effect of SCF on the Adhesiveness of Primary CD34+ Cells

The effect of stimulating sorted CD34+ cells for 30 minutes with SCF at a concentration of 100 ng/ml was evaluated. The proportion of cells adhering to Petri dishes coated with 4 μg/cm2 CH296 increased from 14.0% without the addition of SCF to 23.0% with SCF stimulation. Although the distribution of CFC of adherent and nonadherent cells did not differ, the frequency of LTC-ICs was approximately three times higher in the adherent cell population (39 versus 15 LTC-ICs per 105 cells plated without SCF stimulation). Stimulation with SCF did not alter the frequency of CFCs and LTC-ICs within the respective cell fractions. As shown in Table 3, subsequent limiting dilution analysis of the LTC-IC frequency within the adherent cell fractions revealed a net increase of adherent LTC-ICs of approximately 30% by stimulation with SCF.

Table Table 3.. Stem cell factor (SCF) stimulation enhances the number of adherent long-term culture-initiating cells (LTC-ICs)
  1. a

    The frequency of LTC-ICs within the adherent cell fraction of CD34+ cells is slightly reduced after stimulation with SCF, whereas the absolute number of the adherent cells increases. Shown is the absolute number of adherent cells after 30′ incubation at 37°C with or without SCF. The frequency is calculated from the pooled data of four independent experiments. The number of LTC-ICs is calculated by dividing the number of adherent cells by the LTC-IC frequency.

 With SCFWithout SCF
Number adherent cells per 1.8 × 106 cells plated2.36 × 1053.74 × 105
Frequency of LTC-ICs1 in 1,6921 in 2,071
Number LTC-ICs detected140181

SCF Stimulation Increases the Homing Ability after Transplantation in NOD/SCID Mice

By transplanting human HPCs into irradiated NOD/SCID mice, we sought to confirm the improvement of the in vivo homing abilities of SCF-stimulated HSCs compared with untreated control. In two independent experiments, a total of 16 animals were injected with lin cells with or without the addition of SCF. In both experiments, bone marrow cells from all animals could be analyzed by flow cytometry and PCR. Eight of eight mice in the SCF treatment group showed multilineage hematopoietic engraftment of the human cells as assessed by the detection of at least five human lymphocytes and five human granulocytes per 2 × 104 bone marrow cells analyzed by flow cytometry and the detection of human DNA by PCR. In the animals of the control group, no human cells could be detected by flow cytometry or by PCR (Fig. 3). The mean proportion of human CD45 and CD71+ cells as determined by flow cytometry was 8.5 ± 7.2% (mean ± standard deviation [SD]; range, 1.2–17.2) and 3.0 ± 2.3% (mean ± SD; range, 1.2–6.4) after the transplantation of 3.0 × 105 and 3.6 × 105 lin human cells, respectively (Table 4).

Table Table 4.. Effect of stem cell factor (SCF) on the hematopoietic engraftment of enriched human CD34+ cells transplanted in irradiated nonobese diabetic-scid/scid (NOD/SCID) mice
  1. a

    In two independent experiments, 16 irradiated NOD/SCID mice were injected with lin cells with (+SCF) or without (−SCF) the addition of 10 μg SCF per mouse. Multilineage hematopoietic engraftment of the human cells was assessed by the detection of at least five human lymphocytes and five human granulocytes per 2 × 104 bone marrow cells analyzed. Indicated is the number of animals that showed engraftment of the human cells and the mean proportion (± standard deviation) of human CD45/CD71+ cells detected 6 weeks after the transplantation in the bone marrow of the mice.

 Number of mice engrafted/transplantedProportion of human cellsNumber of mice engrafted/transplantedProportion of human cells
+SCF4/43.0 ± 2.3%4/48.5 ± 7.2%
Figure Figure 3..

Bone marrow cells isolated from nonobese diabetic-scid/scid mice transplanted 6 weeks previously with lin hematopoietic progenitor cells were analyzed by PCR for human 5′ actin. (A): PCR analysis of mouse bone marrow cells mixed with a varying proportion of human cells. (B), (C): PCR analysis of two independent experiments in which two groups of mice were transplanted with (+SCF) or without (−SCF) the addition of SCF. As a negative control, bone marrow cells of a mouse that did not receive human cells were analyzed (C, mouse 0). Abbreviations: PCR, polymerase chain reaction; SCF, stem cell factor.


Hematopoietic engraftment is thought to be a multistep process involving rolling, extravasation, migration, and finally adhesion in the correct microenvironment within the bone marrow [15]. The first in vivo evidence of a cytoadhesion molecule on HSCs to be implicated in progenitor homing was provided by Williams et al. [16]. Since then, a variety of adhesion mechanisms has been studied that may be operative to achieve engraftment and—in the setting of clinical stem cell transplantation—reconstitution of hematopoiesis. Among them, the β1 integrins VLA-4 and VLA-5 are of specific interest because they are involved not only in the transmigration process but also in progenitor adhesion and retention within the bone marrow [15,17]. Under steady-state conditions, the expression of both integrins is higher on bone marrow–derived CD34+ progenitor cells than on their peripheral blood-derived counterparts [18,19]. The adhesive function of VLA-4 and VLA-5 on TF-1 and Mo7e cells can be increased by stimulation with GM-CSF, IL-3, or SCF [9]. Therefore, we sought to explore the effect of cytokine stimulation on the adhesion of functionally defined early HPCs. We focused on SCF because interaction between SCF and its receptor c-kit plays a major role during the process of proliferation and differentiation of hematopoietic cells [20] and the expression of c-kit has been described in early stages of the hematopoietic differentiation. The few published data on the distribution of functionally defined progenitor cell populations indicate that the early HPCs comprising repopulating cells and LTC-ICs are enriched in the c-kit fraction of CD34+ cells. Using the transplantation into fetal sheep, Kawashima et al. [21] showed that pluripotent HPCs within bone marrow–derived human CD34+ cells were found only in the c-kit fraction. In cord blood, the frequency of LTC-ICs was 2.7-fold higher within the CD34+/c-kit fraction compared with the corresponding c-kit+ fraction [22]. Similarly, it was shown that the frequency of LTC-ICs in the c-kit fraction of G-CSF–mobilized CD34+ cells from chemotherapy-treated patients was sixfold higher compared with the c-kit+ cells [23]. Our data confirm these observations, although the plating efficiency of our experiments was somewhat lower than reported [23]. With regard to the stimulation of CD34+ cells with SCF, these data imply that the effect of SCF should be restricted to the functionally more differentiated cells, i.e., CFCs, and not include LTC-ICs. For confirmation, we studied the surface expression of VLA-4 and VLA-5 of CD34+ cells before and after SCF stimulation. It turned out that SCF not only acts on c-kit+ cells but also increases the integrin expression of the c-kit fraction. Obviously, the sensitivity of the flow cytometric analysis was not sufficient to identify all cells that were sensitive to stimulation with SCF. An additional observation in these experiments was the difference in the integrin expression on cells from healthy donors and chemotherapy-treated tumor patients. The two major differences between the two groups are the underlying malignant disease and the chemotherapy included in the mobilization regimen in the patient group. It was shown earlier that the expression of both integrins is higher in CD34+ cells from tumor patients than from healthy donors; patients with Hodgkin's disease presented the highest proportion of integrin-expressing hematopoietic progenitors [24]. This suggests that a different population of CD34+ cells is able to leave the bone marrow upon chemotherapy and G-CSF stimulation compared with G-CSF alone or that the expression of the respective integrins on CD34+ cells of tumor patients is constantly higher than in healthy donors, probably due to continuous activation. Indeed, both chemotherapy and G-CSF stimulation mobilize CD34+ progenitor cells from the bone marrow, which supports the idea of an additional recruitment of CD34+ cells by chemotherapy/G-CSF combination despite a higher expression of adhesion molecules. However, various tumor entities are associated with a different integrin expression profile [24], which points to an influence of the underlying disease on the bone marrow environment. Most likely, the increase of the cell-surface expression of VLA-4 and VLA-5 in CD34+ cells from tumor patients is a result of both effects.

With the next series of experiments, we aimed to explore the effect of SCF on the adhesive potential of the various functionally defined cell populations by using fibronectin fragment CH296, which comprises specific binding sites for both VLA-4 and VLA-5 [25]. As reported by Levesque et al. [9] for TF-1 and Mo7e cells, a short time period was chosen for cytokine stimulation to measure the affinity confirmation of the integrins rather than the level of surface expression. We demonstrate that short-term stimulation with SCF induces the adhesive capabilities of all progenitor populations studied. Furthermore, it was shown that LTC-ICs are enriched within the adherent cell fraction. The minor decrease in the LTC-IC frequency within the adherent cell fraction after SCF stimulation is more than compensated by the overall increasing adhesiveness of the CD34+ progenitors. Our data confirm previous studies showing that cytokine stimulation upregulates β1 integrin expression and function on CD34+ cells [9, 19, 26]. Functional studies of the integrin function on primary human HPCs are scarce. Our data extend the reported observations by a functional definition of the stimulated peripheral blood-derived progenitor cell populations, including early HPCs like LTC-ICs and in vivo repopulating cells.

Adhesion through the α4β1 integrin is thought to be responsible for securing hematopoietic progenitors in the bone marrow, whereas the initial steps of the homing process supposedly depend on other adhesion molecules, such as L-selectin, P-selectin, and PECAM-1 [6,27]. It has been hypothesized that upregulation of the β1 integrin expression and function in vivo may occur within the bone marrow environment once the cells have migrated into the bone marrow space [10]. One responsible mechanism for the increased engraftment could be that SCF induces the proliferation of the repopulating cells. However, given the evidence that the CD34+/38 cell fraction of umbilical cord blood and bone marrow–derived cells that comprise most of the CRU need at least 2 to 3 days for the first all division [13,28], and given that the known half-life of SCF is approximately 2 hours [29], such a possibility seems unlikely. A second mechanism could be the effect of human SCF on the mouse bone marrow cells, possibly inducing secondary mediators for the progenitor's cell engraftment. Although such a possibility cannot formally be ruled out, it seems rather unlikely because it is known that the impact of human SCF on murine cells is 800 times lower than that of rodent SCF [30]. A third way that increased numbers of CRU could have engrafted is the increased integrin function of the cells on SCF stimulation. Such an effect is known to exist, because several in vitro studies have been published demonstrating the increased adhesive function of the β1 integrins after cytokine stimulation [9] and the increased transmigratory potential of the cells after short-term stimulation with SCF [31]. Alternatively, SCF injected simultaneously might increase the response of the cells to stromal-cell-derived factors (SDF-1), a possibility that cannot be completely ruled out. However, the effect of SCF on the adhesive function is known to be quick, whereas the induction of the SDF-1 responsiveness and the increased transmigratory potential reach their maximum after 40 and 48 hours, respectively [31,32]. Therefore, we rather believe that the increased integrin-mediated adhesiveness is responsible for the improvement of the in vivo homing potential.

In summary, the data presented in this report support the concept of circulating CD34+ cells to obtain adhesive capabilities by increased integrin expression and function. This observation comprises various HPC populations, including LTC-ICs. In addition, we provide the first in vivo evidence that stimulation with cytokines improves the homing ability of transplanted human HPCs. Cytokine stimulation with SCF seems to re-establishment an adhesive phenotype of early HPCs that may enhance the engraftment of these cells after clinical stem cell transplantation.


The excellent technical assistance of Nadine Pißler and Marit Hoffmann is gratefully acknowledged. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (HE 3199/4-1; to B.H.).