Hematopoietic stem cells (HSCs) lose marrow reconstitution potential during ex vivo culture. HSC migration to stromal cell–derived factor (SDF)-1 (CXCL12) correlates with CXC chemokine receptor 4 (CXCR4) expression and marrow engraftment. We demonstrate that mobilized human CD34+ peripheral blood stem cells (CD34+ PBSCs) lose CXCR4 expression during prolonged culture. We transduced CD34+ PBSCs with retrovirus vector encoding human CXCR4 and achieved 18-fold more CXCR4 expression in over 87% of CD34+ cells. CXCR4-transduced cells yielded increased calcium flux and up to a 10-fold increase in migration to SDF-1. Six-day cultured CXCR4-transduced cells demonstrated significant engraftment in nonobese diabetic/severe combined immunodeficient mice under conditions in which control transduced cells resulted in low or no engraftment. We conclude that transduction-mediated overexpression of CXCR4 significantly improves marrow engraftment of cultured PBSCs.
Hematopoietic stem cell (HSC) egress to the circulation and homing to the bone marrow (BM) are regulated in part by interactions between CXC chemokine receptor 4 (CXCR4) and stromal cell–derived factor (SDF)-1 [1, 2]. Mobilization of CD34+ peripheral blood stem cells (CD34+ PBSCs) from BM occurs following administration of either granulocyte colony-stimulating factor (G-CSF; inactivation of SDF-1 and cleavage of CXCR4) or AMD3100 (partial agonist of CXCR4) [3–5]. SDF-1 binding to CXCR4 on HSCs mediates chemotaxis, expression of adhesion molecules, proliferation, and survival [6–8].
Experimental approaches under development for transplantation or gene therapy with HSCs may involve ex vivo manipulation that decreases available HSCs or affects BM homing and engraftment. The level of CXCR4 expression on CD34+ HSCs decreases during prolonged ex vivo culture . Ex vivo cultured human HSCs progressively lose nonobese diabetic/severe combined immunodeficient (NOD/ SCID) mouse reconstitution potential that is improved by SDF-1 treatment . In the clinic, donor CD34+ HSCs that express higher levels of CXCR4 are statistically correlated with earlier reconstitution of hematopoiesis after transplantation . All these observations suggest that loss of CXCR4 expression on CD34+ HSCs during culture might precede other processes that diminish the potential to participate in long-term hematopoiesis. In the study reported here we show that transduction-mediated overexpression of CXCR4 during culture of G-CSF–mobilized CD34+ PBSCs significantly improves BM repopulating potential over that of similarly cultured normal PBSCs, as measured by transplantation into sublethally irradiated NOD/SCID mice.
Collection of Normal CD34+ PBSCs
CD34+ PBSCs from five healthy donors were used. After informed consent following International Review Board–approved clinical protocol 94-I-0073 of the National Institutes of Health, the normal volunteers were administered 10 μg/kg/day G-CSF, apheresed on the fifth day, and CD34+ cells selected from each apheresis product by immunomagnetic beads (Isolex 300i, Baxter Healthcare, Deerfield, IL, http://www.baxter.com). The purity of CD34+ PBSCs after selection was >94%. Vials of CD34+ PBSCs were stored in liquid nitrogen and freshly thawed for each experiment.
Generation of MFGS Vector Encoding Human CXCR4 and gp91phox
To generate murine onco-retrovirus MFGS-CXCR4 and MFGS-gp91phox vector, open reading frames of human CXCR4 and human gp91phox cDNA, respectively, were inserted directionally into the NcoI-BamHI cloning site of the MFGS vector. For each vector, a high-titer FLYRD18 producer clone was generated, and virus supernatant was collected as previously described .
Transduction of CD34+ PBSCs
CD34+ PBSC cultures were initiated in Retronectin (Takara Shuzo Ltd., Otsu, Japan, http://www.takara-bio.co.jp/english) pre-coated six-well plates in growth medium (X-VIV0 10/1% human serum albumin; 50 ng/ml fms-like tyro-sine kinase 3 (FLT3) ligand, 10 ng/ml stem cell factor, 10 ng/ml thrombopoietin, 10 ng/ml interleukin-3). CD34+ PBSCs were transduced overnight four times (days 2, 3, 4, and 5) using dilutions of concentrated virus vector equivalent to 1 × 107 infectious units and 5 μg/ml protamine. Naive non-transduced or gp91phox-transduced cultured CD34+ PBSCs served as negative controls for CXCR4 expression.
Calcium Flux Response and Transwell MigrationAssay
CD34+ PBSCs were loaded in Hank's Balanced Salt Solution (HBSS) supplemented with 20 mM HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid) with 2 μ1 of a 1:1 mixture of 1 μg/μ1 Fura-2 AM and 20% Pluronic F-127 (Molecular Probes, Eugene, OR, http://www.probes.com); then they were incubated, washed, and pipetted into a well of a poly-L-lysine (Sigma Chemical Corp., St. Louis, http://www.sigma-aldrich.com) coated 96-well plate (Greiner Bio-One, Longwood, FL, http://www.gbo.com/bioscience). The plate was loaded into a Flexstation (Molecular Devices, Sunnyvale, CA, http://www.moleculardevices.com) set at 37°C. Stimuli in HBSS/20mM HEPES/1% bovine serum albumin (Sigma), were robotically added; using excitation at 340 and 380 nm, the fluorescence intensities were detected at 510 nm. Data are expressed as change in relative fluorescence calculated as the ratio of the two intensities (SoftmaxPro Software, Molecular Devices).
For the transwell migration assay, 0.5 × 106 of cultured CD34+ cells were added in 0.1 ml growth medium to the upper chamber of a 5-μm polycarbonate transwell (Corning Costar Corporation, Cambridge, MA,http://www.corning.com). The bottom chamber contained 0.6 ml of growth medium with or without 10 nM SDF-1. After 30 minutes incubation, migrated cells in the bottom chamber were collected with two to three washes and counted three times. All runs were performed in duplicate.
Transplantation of CD34+ PBSCs into NOD/SCID Mice and Harvest of Bone Marrow
Ten × 106 normal and 10 × 106 MFGS-CXCR4–transduced 6-day cultured PBSCs (equivalent to 4 × 106 CD34+ PBSCs at initiation of culture) were injected via tail vein into eight and 10 sublethally irradiated NOD/SCID mice, respectively. Mice were 8–10 weeks old at day of transplantation. In a separate experiment, 6-day cultured human MFGS-CXCR4 and MFGS-gp91phox control-transduced PBSCs (4.5 × 106 MFGS-CXCR4–transduced cells per mouse and 5 × 106 MFGS-gp91phox–transduced cells per mouse; both equivalent to 1 × 106 cells at initiation of culture) were injected via tail vein into three and two sublethally irradiated (312 Rads) 13-week-old NOD/SCID mice (Jackson ImmunoResearch Laboratories, West Grove, PA, http://www.jax.org), respectively. For each experiment, an aliquot of cells was retained in ex vivo liquid culture. To analyze BM engraftment, mice were sacrificed 6 or 9 weeks post-transplant, and BM from tibias and femurs was flushed into X-VIV0 10/1% human serum albumin.
Analysis of Transgene Expression and Human Cell Engraftment by Flow Cytometry
Human CXCR4 expression was determined by anti-human fluorochrome-conjugated monoclonal antibody staining (CXCR4-CyChrome or CXCR4-PE; clone 12G5; BD Biosciences Pharmingen, San Diego, http://www.bdbiosciences.com) after permeabilization of the cell membrane to detect both surface and intracellular CXCR4. Human gp91phox expression was determined by indirect staining with murine monoclonal antibody 7D5 followed by fluorescein isothiocyanate–conjugated goat anti-mouse immunoglobulin G antibody. Anti-human fluorochrome-conjugated monoclonal antibodies were used to identify human hematopoietic cells (CD45-PerCP or CD45-APC, CD34-PE; BD Biosciences Pharmingen).
Results are reported as mean ± standard error. Differences of cell populations were analyzed using a one tailed Student's t-test.
Results and Discussion
We first investigated the expression of CD34 and CXCR4 on G-CSF–mobilized normal CD34+ PBSCs in ex vivo culture. We found that during prolonged culture, CD34 expression is progressively lost on PBSCs, such that by day 6, only 10%–38% (n = 4) of cells remained CD34+. Initial CXCR4 expression on CD34+ PBSCs varied depending on the donor (20%–52%, n = 4) and was upregulated during the first 48 hours by cytokine stimulation, as was reported before [1, 2]. During prolonged culture, however, CXCR4 expression was progressively lost on cells that remained CD34+ such that by day 6, only 1.1%–8% continued to coexpress CXCR4 (Fig. 1). Based on the loss of CXCR4 expression on CD34+ cells under prolonged culture conditions, we overexpressed the CXCR4 transgene in CD34+ cells.
Cultured CD34+ PBSCs transduced overnight on days 2–6 with murine retrovirus vector RD114-pseudotyped MFGS-CXCR4, expressed high levels of CXCR4 (Fig. 1). Either normal (nontransduced) CD34+ PBSCs or RD114-MFGS-gp91phox (phagocyte oxidase transmembrane subunit)–transduced CD34+ PBSCs were used as control (92% of cells transduced; not shown). CXCR4 transgene–positive hematopoietic progenitor cells expressed 11- to 18-fold more CXCR4 per cell (measured by mean fluorescence intensity) than normal or control-transduced cells (n = 5). CXCR4-transduced hematopoietic progenitors (five independent tests with CD34+ PBSCs from five different donors) sampled on culture day 6 demonstrated increased calcium flux SDF-1 response that was 160%–220% of naive cultured CD34+ PBSCs (Fig. 2, inset). Respective CXCR4-transduced PBSC migration over 30 minutes on day 6 in response to SDF-1 was on average 3.4-fold greater than migration of cultured control cell migration in response to SDF-1 (Fig. 2). These ex vivo results suggest CXCR4 transgene–enhanced migration and calcium-flux response to SDF-1. Since cells were not split into CD34+ and CD34− populations for the calcium flux and migration assays, it should be noted that upregulated naive CXCR4 on more differentiated cells in culture might have influenced the results.
In a NOD/SCID mouse study, we compared repopulation potential of CXCR4-transduced or normal (nontransduced) cells at day 6 of culture, in which each mouse was transplanted with the progeny from 4 × 106 CD34+ PBSCs seeded at culture initiation. At the day of transplantation, the CXCR4-transduced CD34+ cells expressed CXCR4 in >87%, compared with 8% naive CXCR4 expression in the normal (non-transduced) CD34+ population (Fig. 1; Table 1, column 5). After 6 weeks, BM of 10 mice transplanted with the CXCR4-transduced 6-day cultured CD34+ PBSCs demonstrated 19.9% ± 3.4% engraftment of human CD45+ cells, while BM of eight mice transplanted with the normal 6-day cultured CD34+ PBSCs demonstrated significantly lower (p < .01) engraftment (9.3% ± 1.3%; Fig. 3; Table 1, column 8).
Table Table 1.. Human cell engraftment in NOD/SCID mice, CXCR4-transgene expression and cell migration ex vivo
No. of CD34+PBSCs (×106)per mouse at initiation of culture
No. of CD34+CXCR4+cells (×106) transplanted per mouse, day 6
Mean human CD45+cell engraftment in bone marrow (%)
No. of mice transplanted
Naive CD34+ PBSCs
9.3 ± 1.3
19.9 ± 3.4
0.65 ± 0.1
Control-transduced CD34+ PBSCs
We previously reported accelerated loss of the NOD/ SCID mouse marrow repopulating potential of human CD34+ PBSCs between the 4th and 6th days of ex vivo culture . With CD34+ PBSCs from a different healthy volunteer, we compared NOD/SCID repopulation potential of CXCR4 and gp91phox control-transduced cells at day 6 of culture, in which each mouse was transplanted with the progeny from only 1 × 106 CD34+ PBSCs seeded at culture initiation. Under similar conditions, no engraftment of normal CD34+ PBSCs was seen in NOD/SCID mice in previous experiments (unpublished data). At the day of transplantation, the CXCR4-transduced PBSCs expressed CXCR4 in 95%, compared with <2% expression of naive CXCR4 in the control-transduced cell population. After 9 weeks, BM of the three mice transplanted with CXCR4-transduced hematopoietic progenitors demonstrated 0.65% ± 0.1% engraftment of human CD45+ cells, while no engraftment—as expected—was detected in the two mice transplanted with gp91phox control-transduced cells (Fig. 4 A, B; Table 1, column 8). Furthermore, all CD45+ human cells from NOD/SCID chimeric BM in the CXCR4-transduced group coexpressed high levels of CXCR4 transgene (Fig. 4C, D), which is characteristic of CXCR4-transduced cells ex vivo.
The studies demonstrate that, under the prolonged ex vivo culture conditions, the CXCR4 transduction of CD34+ PBSCs results in significantly improved long-term NOD/ SCID engraftment conferred by transduction-mediated forced overexpression of CXCR4, and this is consistent with recently published data by Kahn et al. . Our results indicate that loss of marrow reconstitution potential of cultured HSCs not only is due to differentiation (and thus loss of self-renewal potential) but also may be preceded by loss of CXCR4 expression crucial for homing. This represents direct evidence that CXCR4 mediates HSC homing and engraftment to the BM, and it provides incentive for development of clinically applicable methods for increasing levels of expression of CXCR4 to enhance engraftment of ex vivo manipulated HSCs.