This work was supported by an operating grant from the Canadian Blood Services/Canadian Institutes for Health Research Blood Utilization and Conservation Initiative to AJW.
CD34+ cell responsiveness to stromal cell–derived factor-1α underlies rate of engraftment after peripheral blood stem cell transplantation
Article first published online: 2 OCT 2008
© 2008 American Association of Blood Banks
Volume 49, Issue 1, pages 161–169, January 2009
How to Cite
Marquez-Curtis, L. A., Turner, A. R., Larratt, L. M., Letcher, B., Lee, S. F. and Janowska-Wieczorek, A. (2009), CD34+ cell responsiveness to stromal cell–derived factor-1α underlies rate of engraftment after peripheral blood stem cell transplantation. Transfusion, 49: 161–169. doi: 10.1111/j.1537-2995.2008.01937.x
This work was awarded the special recognition of outstanding poster in the cellular therapy category at the 2007 AABB Annual Meeting and TXPO held in Anaheim, CA.
- Issue published online: 23 DEC 2008
- Article first published online: 2 OCT 2008
- Received for publication June 5, 2008; revision received July 30, 2008; and accepted August 5, 2008.
BACKGROUND: Stromal cell–derived factor (SDF)-1, a chemokine produced in the bone marrow (BM), is essential for the homing of hematopoietic stem/progenitor cells (HSPCs) to the BM after transplantation. This study examines whether there is a correlation between the in vitro chemotaxis of CD34+ HSPC toward an SDF-1 gradient and in vivo hematopoietic engraftment.
STUDY DESIGN AND METHODS: Thirty-five patients underwent granulocyte–colony-stimulating factor HSPC collection and autologous transplant with a median dose of 7.7 (range, 3.9-41.5) × 106 CD34+ cells per kg body weight. The chemotactic index (CI) of CD34+ cells isolated from leukapheresis products collected from these patients was calculated as the ratio of the percentages of cells migrating toward an SDF-1 gradient to cells migrating to media alone. Expression of the SDF-1 receptor CXCR4 on CD34+ cells was measured by flow cytometry.
RESULTS: Spontaneous cell migration (range, 3.1 ± 0.6 to 26.5 ± 7.7%) and SDF-1–directed chemotaxis (11.1 ± 0.7 to 54.9 ± 8.3%) of CD34+ cells did not correlate with time to neutrophil engraftment, which occurred at a median of 10 days (range, 8-16 days). Nonparametric tests showed a negative correlation (r = −0.434) between CI and CD34+ cell dose such that neutrophil recovery occurred within the same period in patients transplanted with a lower dose of CD34+ cells but having a high CI as in those transplanted with a higher dose of CD34+ cells but having a low CI. Moreover, CI correlated (r = 0.8) with surface CXCR4 expression on CD34+ cells.
CONCLUSION: In patients transplanted with a relatively lower CD34+ cell dose who achieved fast engraftment, a higher responsiveness to SDF-1 and high CI could have compensated for the lower cell dose. However, to apply the CI as a prognostic factor of the rate of engraftment requires validation in a larger number of patients.
hematopoietic stem/progenitor cell(s)
peripheral blood stem cell transplant
stromal cell–derived factor
Transplantation of hematopoietic stem/progenitor cells (HSPCs) collected from peripheral blood (PB) after their mobilization by chemotherapy and/or granulocyte–colony-stimulating factor (G-CSF) has become a common treatment for various hematologic and nonhematologic disorders.1 Hematopoietic engraftment after transplantation is dependent on the successful homing of HSPCs to the bone marrow (BM) and their repopulation.2-5 Homing occurs within the first few hours after transplantation and involves a complex series of events including integrin-mediated adhesion of HSPCs on the endothelial cells, transendothelial migration, chemotaxis, and extracellular matrix degradation.6-8 The α chemokine stromal cell–derived factor (SDF)-1/CXCR4 receptor axis has been identified as a primary axis governing HSPC homing and engraftment in the BM.9,10 SDF-1 (also termed CXCL12) is produced by BM stromal and endothelial cells as well as osteoblasts, and the CXCR4 receptor (also known as CD184) is expressed by HSPCs. Knock-outs of SDF-1 or CXCR4 resulted in a significant defect of colonization of embryonic BM by HSPCs and lethal defects in the development of various organs.11,12 On the other hand, overexpression of CXCR4 by human CD34+ cells increased their migration and repopulation of BM in nonobese diabetic (NOD)/SCID mice.13 The realization that HSPC homing and retention in the BM is regulated by the SDF-1/CXCR4 axis has led to the development of strategies that use interference with this interaction as the basis for mobilization of HSPC from the BM to the circulation (e.g., the clinical application of the CXCR4 antagonist AMD3100).14-16 However, protocols to improve the homing of HSPC to BM have not yet been developed. We have proposed, for example, that in vitro priming of CD34+ cells with various factors to increase CXCR4 expression could speed up the process of homing and hematopoietic engraftment. We showed that upregulation of the incorporation of the CXCR4 receptor into the membrane lipid rafts of HSPCs by these factors improves engraftment in a murine transplant model.17
We sought to establish that in vitro chemotaxis toward an SDF-1 gradient of HSPC from mobilized patients correlates with hematopoietic recovery after a peripheral blood stem cell transplant (PBSCT). We hypothesized that HSPCs that have increased chemotaxis and high CXCR4 expression would engraft faster.
MATERIALS AND METHODS
Patients, mobilization, and engraftment
Forty-one patients were recruited for this study; however, 6 were excluded because 2 died before PBSCT and 4 did not require a transplant. The clinical characteristics of the remaining 35 patients are summarized in Table 1. The patients were diagnosed with relapsed, chemotherapy-sensitive non-Hodgkin's and Hodgkin's lymphomas, solid tumors, or amyloidosis. Patients were mobilized with cyclophosphamide (Day 1) and etoposide (Days 1-3) and, starting on Day 4, G-CSF (filgrastim, Amgen, Thousand Oaks, CA; 5 µg/kg, subcutaneously twice daily, for 5 to 10 days), followed by apheresis. Patients were apheresed once the CD34+ cell count in the PB reached a minimum of 15 per µL (see below). In our transplant center, patients who do not attain this level are not collected.
|Mean||42 ± 14.3 (n = 35)|
|Number of patients <40 years||15|
|Number of patients ≥40 years||20|
|Salvage chemotherapy (number of cycles)*|
|Chemotherapy plus G-CSF||33|
|CD34+ cell dose transplanted (×106/kg)|
|Mean||10.6 ± 1.5 days|
|Median||10 (8-16) days|
|Mean||11.5 ± 4.7 days|
|Median||10 (4-26) days|
PB on the day of leukapheresis as well as an aliquot from leukapheresis product (LP) was obtained with the patients' informed consent (in accordance with the institutional guidelines approved by the Human Research Ethics Board of the University of Alberta) before cryopreservation. Patients underwent autologous PBSCT at the Cross Cancer Institute (Edmonton, Alberta, Canada) after they had undergone a conditioning regimen. Neutrophil engraftment was defined as the number of days from the day after transplant up to and including the first day when neutrophils and bands (absolute neutrophil count) were greater than 0.5 × 109 per L for 3 consecutive days. Platelet (PLT) recovery was defined as the number of days when the PLT count was greater than 20 × 109 per L for 3 consecutive days without PLT transfusions.
CD34+ cell analysis, isolation, and clonogenic assay
The absolute number of CD34+ cells in the PB and in the LP was evaluated using single-platform flow cytometric analysis, according to a modified version of the International Society of Hematotherapy and Graft Engineering protocol,18 using a reagent with a lyse–no wash protocol (Stem-kit, Beckman Coulter, Hialeah, FL). The cell viability was evaluated using 7-aminoactinomycin D to identify and exclude 7-aminoactinomycin D–positive nonviable cells. Flow cytometric analysis was performed on a flow cytometry system (EPICS XL-MCL or FC500, Beckman Coulter) after proper instrument alignment, standardization, color compensation, and validation. Moreover, cells from LP were dispensed into methylcellulose complete medium containing cytokines (Methocult GF H4434, Stem Cell Technologies, Vancouver, British Columbia, Canada), plated, and incubated at 37°C in a humidified incubator with 5 percent CO2 in air. After 14 days colony-forming unit–granulocyte-macrophage (CFU-GM) colonies were scored using an inverted microscope.
Another aliquot of LP was used to isolate CD34+ cells using immunomagnetic beads (Miltenyi-Biotec, Auburn, CA) according to the manufacturer's instructions and as previously described by us.19 Briefly, light-density mononuclear cells were separated by density gradient centrifugation over 60 percent Percoll (GE Healthcare, Quebec, Canada), labeled with a CD34+ antibody (QBEND 10) and passed through a positive selection column. The isolated CD34+ cells were further analyzed for surface CXCR4 expression and in vitro migration (as described below).
The CD34+ cells isolated from LP were labeled with fluorescein isothiocyanate (FITC)-conjugated anti-human CD34 antibody (IM1870, Beckman Coulter Canada, Mississauga, Ontario, Canada), phycoerythrin (PE)-conjugated anti-human CD184 (CXCR4, clone 12G5, BD Biosciences Canada, Mississauga, Ontario, Canada) or FITC/PE-labeled mouse immunoglobulin (Ig)G1 isotypic controls (IM0639 and IM0670, Beckman Coulter) for 30 minutes at room temperature. Similarly, PB obtained on the day of leukapheresis collection was stained and the red blood cells were lysed. The remaining cells were washed and fixed with 1 percent paraformaldehyde. Surface CXCR4 expression on CD34+ cells separated from the LP and the CD34+-gated population in PB collected on the day of leukapheresis was measured as median fluorescence intensity (MFI) by flow cytometry (BD FACSCalibur system and CellQuest software, Becton Dickinson, Oakville, Ontario, Canada) after standardizing with QC3 microbeads (Bangs Laboratories, Inc., Fishers, IN), which allowed direct comparison of results acquired on different days.
Spontaneous migration and chemotaxis assays
The cell migration assay was carried out using modified Boyden chambers (Neuro Probe, Inc., Gaithersburg, MD) and polycarbonate filters (13-mm diameter, 8-µm pore size, Nucleopore, Toronto, Ontario, Canada). The lower compartments were filled with serum-free Iscove's modified Dulbecco's medium (Invitrogen, Burlington, Ontario, Canada) supplemented with 0.1 percent bovine serum albumin (BSA; Sigma-Aldrich Canada, Oakville, Ontario, Canada) with or without 200 ng per mL recombinant human SDF-1α (Biomedical Research Centre, UBC, Vancouver, British Columbia, Canada). CD34+ cells (1.5 × 106 cells suspended/mL of Iscove's modified Dulbecco's medium/0.1% BSA) were loaded in the upper compartments and incubated (at 37°C and 5% CO2) for 5 hours. For some experiments CD34+ cells were also incubated (for 30 min at 37°C and 5% CO2) with 10 µg per mL AMD3100 (Sigma). Cells were recovered from the lower compartment and counted using a hemacytometer. Percentage migration was calculated from the ratio of the number of cells that migrated to the total number of cells loaded. The chemotactic index (CI) was defined as the ratio of the percentage of cells that migrated towards SDF-1α to the percentage of cells that migrated toward medium alone.
The correlations between CD34+ cell dose, CFU-GM cell dose, percentage spontaneous migration, percentage SDF-1α–induced migration and CI, and number of days to neutrophil recovery were analyzed using linear regression analysis. Since the CI was not normally distributed, nonparametric correlations (Spearman's rho and Kendall's tau tests) were conducted to study the relationship between CI and CD34+ cell dose. Correlation coefficient (r) values of greater than 0.5 were considered significant. Significance was defined as a p value of less than 0.05. Mean and standard deviation were calculated for continuous variables, and frequencies (percentages) for categorical variables.
CD34+ cell dose and engraftment
All 35 patients were transplanted with relatively high CD34+ cell doses as patients with a PB CD34+ cell count of less than 15 per µL are not apheresed in our center. Table 1 shows that the median dose of CD34+ cells was 7.7 × 106 per kg and it ranged from 3.9 × 106 to 41.5 × 106 per kg body weight. Engraftment was rapid with patients achieving an absolute neutrophil count of more than 0.5 × 109 per L within a mean of 10.6 ± 1.5 days (median, 10 days; range, 8-16 days). No correlation (r = −0.13) between the number of days to neutrophil recovery and CD34+ cell dose could be established (Fig. 1A). The mean CFU-GM cell dose (×105/kg) transplanted was 15.5 ± 12.0 (median, 11.54; range, 4.7-53.1) and did not correlate with days to neutrophil recovery either (r = 0.07, data not shown). However, a significant negative correlation (r = −0.54) between days to PLT recovery and CD34+ cell dose could be established (Fig. 1B).
Spontaneous migratory potential of CD34+ cells and neutrophil recovery
To explain why patients transplanted with very high doses of CD34+ cells do not engraft faster, we evaluated the inherent migratory ability of CD34+ cells isolated from LP. Using a spontaneous migration assay, we observed broad interpatient differences in the spontaneous migratory ability of CD34+ cells, ranging from 3.1 ± 0.6 to 26.5 ± 7.7 percent. A positive correlation (r = 0.6) between spontaneous migration and the percentage of CD34+ cells in the LP could be established (Fig. 2A), suggesting that cells with inherently greater migratory propensity preferentially leave the BM and enter the PB. To determine whether more motile cells would also have a greater ability to home to the BM after transplantation, we evaluated the relationship between spontaneous migration and the number of days to neutrophil recovery. We observed a negative correlation, although it did not reach significance (r = −0.4), suggesting that the more motile the cells, the faster engraftment is achieved (Fig. 2B).
CD34+ cells have variable responsiveness to an SDF-1α gradient, which did not correlate with neutrophil recovery
Because SDF-1α produced in the BM plays a crucial role in the homing of HSPCs after transplantation, we evaluated SDF-1α–induced migration of CD34+ cells. Again, we found a wide range of variability, from 11.1 ± 0.7 to 54.9 ± 8.3 percent, in SDF-1α–induced migration of CD34+ cells from LP. We then evaluated whether CD34+ cells with higher responsiveness in vitro to an SDF-1α gradient would home more efficiently in vivo to the BM and engraft faster. Surprisingly, SDF-1α–induced migration did not correlate with neutrophil recovery (r = 0.1, data not shown). However, because the SDF-1α–induced migration did not increase proportionately with spontaneous migration, we normalized responsiveness to the SDF-1α gradient against passive migration by calculating the CI (ratio of percentage migration toward SDF-1α and percentage migration toward media alone). The CI values ranged from 1.1 to 5.4, with the highest value corresponding to the greatest responsiveness to the SDF-1α gradient. There was no correlation (r = 0.18) between the CI and time to neutrophil recovery either (Fig. 3A). This lack of correlation could be explained by the fact that in our patient population neutrophil engraftment occurred within a very narrow margin of time (mean, 10.6 ± 1.5 days) owing to the rather high doses of CD34+ cells that were transplanted, although they ranged from 3.9 × 106 to 41.5 × 106 per kg. We then sought to determine whether there is a relationship between the CI and the dose of CD34+ cells transplanted. Because the CI was not normally distributed but slightly skewed to the left, nonparametric tests were conducted. Both Spearman's rho and Kendall's tau tests showed that there is a negative correlation (r = −0.434, p = 0.009) between CI and CD34+ cell dose (Fig. 3B). This suggests that, in many cases, rapid hematopoietic recovery was achieved because a higher CI likely compensated for the lower dose of CD34+ cells transplanted.
CXCR4 expression correlates with the CI
Next we looked at the relationship between the CI and CXCR4 expression. Since patient samples were collected over an extended period of time, to ensure meaningful comparison of fluorescence intensity acquired on different days, standardization of flow cytometric data was carried out by using QC3 microbeads every time flow cytometric data acquisition was performed. Representative histograms from patients demonstrate that, like the CI, CXCR4 expression varied from patient to patient (median, 302; range, 244 to 399). We found that a high CI was associated with strong surface CXCR4 expression (r = 0.8). Moreover, the CXCR4 expression (MFI) was not significantly different on CD34+ cells isolated from LP in comparison to the CD34+-gated population in the PB examined on the day of leukapheresis (Fig. 4A). Chemotaxis of CD34+ cells toward an SDF-1α gradient was blocked by the CXCR4 antagonist AMD3100 (Fig. 4B). Interestingly, the CI correlated positively with the percentage inhibition of chemotaxis by AMD3100 (r = 0.51), suggesting that CD34+ cells that are more strongly retained in the BM (i.e., having high CI and strong CXCR4 expression) respond better to AMD3100 (Fig. 4C).
The dose of CD34+ cells transplanted has been generally regarded as having the greatest influence on hematopoietic recovery, and in many studies a correlation between the number of HSPCs infused and the rate of engraftment has been shown.20-26 Thus, it has been the practice of HSPC transplant centers to collect a sufficient number of CD34+ cells to attain a favorable clinical outcome. However, in autologous transplant settings there is frequently little control over the number of HSPCs available for grafting. Hence, a better understanding of the factors that impact engraftment efficiency would be helpful for designing protocols to hasten hematopoietic recovery and influence clinical outcomes. This is important because delayed engraftment can result in greater transfusion requirements, increased antibiotic use and growth factor support, longer hospitalization, and greater risk of mortality. An early analysis of engraftment kinetics in 692 patients as a function of CD34+ cell dose provided evidence that 5.0 × 106 or more CD34+ cells per kg is an optimal dose for transplantation.24 Later it was suggested that a threshold dose of 2.0 × 106 or more per kg would ensure engraftment;27 however, some multiple myeloma patients who received well above this threshold dose exhibited delayed hematopoietic engraftment.25 Interestingly, in a more recent study of 225 patients with amyloidosis who underwent autologous PBSCT, no differences in the rate of neutrophil engraftment were observed when CD34+ cell doses of more than 5.0 × 106 per kg were used.23 Consistent with this latter report we found that in our study of 35 patients transplanted with a relatively high CD34+ cell dose ranging from 3.9 × 106 to 41 × 106 per kg (median, 7.7 × 106), there was no correlation between cell dose and neutrophil recovery and only a weak correlation between CD34+ cell dose and PLT recovery. In a study of 508 multiple myeloma patients undergoing PBSCT, it was found that high doses of CD34+ cells transplanted (≥6.5 × 106/kg) shortened the number of days to hematopoietic reconstitution, but had no significant impact on transplant-related mortality and overall survival within 100 days.26 In the further search for predictive factors of engraftment, Jansen and coworkers28 retrospectively studied 323 autologous PBSCT patients and found that while both CD34+ and CFU-GM cell dose correlated with the variables of engraftment, the relationship with CD34+ cell dose was stronger and that the CFU-GM cell dose did not add any predictive value. In our study, CFU-GM cell dose, as evaluated by the colony-forming cell assay, did not correlate with neutrophil recovery. Interestingly, this assay has been criticized because it is based on a subjective visual assessment that cannot be calibrated or standardized and does not directly measure proliferation. Recently, a new assay that relies on instrument-based measurement of intracellular ATP concentration has been shown to be directly proportional to the proliferation status of cells29 and to predict engraftment outcome.30 It may replace the colony-forming cell assay in the future but awaits full validation.
Because migration is a normal physiologic function that also enables HSPCs to return to the BM after transplantation, we investigated whether in vitro HSPC migration potential correlates with rate of their engraftment. We found that the spontaneous migratory potential of CD34+ cells correlated strongly with the number of CD34+ cells in the LP, but had only a weak negative correlation with days to neutrophil recovery. This is consistent with another study of 22 patients who received autologous PBSCT, in which a negative correlation was reported between spontaneous migration of CD34+ cells in vitro and neutrophil recovery in vivo (r = −0.5; p < 0.05).31 Surprisingly, we did not find a correlation between neutrophil recovery and the migratory potential of CD34+ cells toward an SDF-1α gradient. Instead, we found that the CD34+ cell dose was inversely related to the chemotactic response toward SDF-1α. Our results suggest that a higher responsiveness of CD34+ cells to SDF-1α, i.e., a high CI, could compensate for a lower CD34+ cell dose in achieving a fast hematopoietic engraftment. Thus, patients who are considered poor mobilizers, from whom only a lower number of CD34+ cells can be collected for transplantation, could still benefit from autologous PBSCT as long as their CD34+ cells have a high CI. On the other hand, patients who are good mobilizers and from whom much more than 5 × 106 CD34+ cells per kg are collected and transplanted do not necessarily have an additional advantage in the rate of engraftment, especially if the CI is low. However, whether the CI could become a predictor of the rate of hematopoietic engraftment requires further investigation using a larger number of PBSCT patients.
Previously, it was shown that CD34+ cells mobilized to PB are characterized by a low CXCR4 membrane expression.32 Here we show that CXCR4 expression on CD34+ cells isolated from LP correlates with CI (r = 0.8; Fig. 4A, left panel) and does not differ significantly between CD34+ cells from LP and those of the CD34+-gated population in mobilized PB (Fig. 4A, right panel). Based on this observation we suggest that CXCR4 expression be evaluated alongside the flow cytometric enumeration of CD34+ cells in mobilized PB. Strong CXCR4 expression would predict a high CI and a higher homing potential. In addition to the SDF-1/CXCR4 axis, several other factors or mechanisms have been identified as contributing to homing and engraftment of HSPCs. Among these are cell motility, actin polymerization, stress signals, increased production of proteolytic enzymes by neutrophils, and CD34+ cells and adhesion molecules (CD44 and VLA-4) activated by SDF-1 which synergize with other molecules and potentiate CD34+ cell adhesion and motility.2-7,33
In summary, we found that CD34+ cells that responded better to an SDF-1α gradient had high CI and CXCR4 expression. These cells may also have a greater propensity to be retained in the BM and this could explain why they were mobilized and collected in lower numbers. However, their high CI may eventually compensate for the lower number of CD34+ cells transplanted. This could explain why neutrophil recovery in patients transplanted with a lower dose of CD34+ cells having a high CI was achieved within the same period of time as in those patients in whom a higher number of CD34+ cells with a low CI was transplanted. We suggest that, in addition to the CD34+ cell dose, CI and the CXCR4 expression may allow us to better predict the rate of hematopoietic recovery after PBSCT, if investigations on a larger number of patients support this.
We thank the technical staff, Doris Quinn and her team of nurses from the CBS HPC Transplant Program, and Laurena Beirnes and Nanette Cox-kennett from the Cross Cancer Institute, Edmonton, for their assistance with patient samples and clinical data. We appreciate the technical help of Jencet Montaño (CBS R&D) and the statistical analysis carried out by Sunita Ghosh (Cross Cancer Institute). There are no conflicts of interest to disclose.
- 5Homing and engraftment of progenitor cells: a prerequisite for cell therapy. J Mol Cell Cardiol 2008 Jan 18. [Epub ahead of print]., , .
- 26Effect of CD34+ cell dose on hematopoietic reconstitution and outcome in 508 patients with multiple myeloma undergoing autologous peripheral blood stem cell transplantation. Eur J Haematol 2006;78:21-8., , , , , , , , , , , , .
- 30A novel stem cell potency and release assay for umbilical cord blood. Cytotherapy 2008;10:168., .