Background: Enrichment and purification of hematopoietic stem and progenitor cells (HSPCs) is important in transplantation therapies for hematologic disorders and in basic stem cell research. Primitive CD34+ HSPCs have demonstrated stronger rolling adhesion on selectins than mature CD34− mononuclear cells (MNCs). We have exploited this differential rolling behavior to capture and purify HSPCs from bone marrow by perfusing MNCs through selectin-coated microtubes.
Methods: Bone marrow MNCs were perfused through the cell-capture microtubes coated with adhesion molecules. We washed the device lumen and visualized and estimated captured cells by video microscopy. Adherent cells were eluted by high shear, calcium-free buffer, and air embolism. We used immunofluorescence staining followed by flow cytometry to analyze CD34+ HSPCs.
Results: CD34+ HSPC purity of cells captured in adhesion molecule–coated devices was significantly higher than the fraction of CD34+ cells found in bone marrow MNCs [mean (SE) 2.5% (0.8%)]. P-selectin–coated surfaces yielded 16% to 20% CD34+ cell purity, whereas antibody-coated surfaces yielded 12% to 18%. Although CD34+ cell purity was comparable between selectin and antibody surfaces, the total number of CD34+ HSPCs captured was significantly higher in P-selectin devices (approximately 5.7 × 104 to 7.1 × 104) than antibody devices (approximately 1.74 × 104 to 2.61 × 104).
Conclusions: P-selectin can be used in a compact flow device to capture HSPCs. Selectin-mediated capture of CD34+ HSPCs resulted in enrichment approximately 8-fold higher than the CD34+ cell population from bone marrow MNCs. This study supports the hypothesis that flow-based, adhesion molecule–mediated capture may be a viable alternative approach to the capture and purification of HSPCs.
The use of adult stem cells for disease therapies has many distinct advantages over the use of embryonic stem cells. These advantages include a decreased likelihood of immunologic rejection, a well-understood history of proven therapeutic benefit spanning decades, and a failure to spawn spontaneous teratomas, all while avoiding complex ethical barriers to research and clinical use. One population of adult stem cells, called hematopoietic stem and progenitor cells (HSPCs),1 has the ability to repopulate the hematopoietic system by proliferating and differentiating into all of the necessary lymphoid, myeloid, and erythroid lineages (1)(2). Because of this rare ability, HSPCs are used as transplantation agents in the treatment of blood- and immune-related malignant and benign diseases (3)(4)(5)(6).
HSPCs commonly express CD34 antigen, a heavily glycosylated transmembrane protein, as the cell-surface marker (4)(7). Bone marrow is the richest source of CD34+ cells, with 1% to 4% of mononuclear cell (MNC) population, compared with cord blood (approximately 1% of MNCs) and mobilized peripheral blood (0.1% to 1% of MNCs). Currently, CD34+ HSPCs are isolated by FACS or immunomagnetic bead separation. Although FACS offers a relatively pure population, it introduces potential sources of contamination. Another important drawback is that flow cytometers with high-speed sorting ability are prohibitively expensive for smaller laboratories. Immunomagnetic bead separation, which uses monoclonal antibodies to CD34 or CD133 antigens conjugated to magnetic beads, involves several steps, including incubation of the cells with antibody, cell selection, elution, collection, and release of conjugated antibodies. Evidence suggests that use of cell-surface antibodies may prohibit cell proliferation and differentiation (8). Furthermore, the method requires large amounts of starting material—for instance, the Dynal progenitor cell selection system (Invitrogen, prod. no. 113.01 and 113.02) requires at least 4 × 107 to 4 × 108 MNCs (corresponding to 40–50 mL bone marrow sample or at least 200 mL mobilized peripheral blood). From a practical point of view, obtaining such volumes of starting material can be uncomfortable for the patient and require multiple sessions. Development of simpler methods for the capture and purification of HSPCs would be a significant innovation.
Selectins, type I membrane proteins, have an N-terminal C-type lectin domain followed by an epidermal growth factor–like motif, a series of short consensus repeats, a transmembrane domain, and a cytoplasmic tail (9). During inflammation, selectins play a key role in the tethering and rolling of leukocytes on endothelium (10)(11). Selectins have been implicated in the homing of HSPCs to the bone marrow (12)(13)(14). HSPCs express at least 1 P-selectin (sP) ligand, sP glycoprotein ligand-1 (PSGL-1/CD162), and another selectin ligand, HCELL (15)(16) Primitive CD34+ HSPCs have been demonstrated to exhibit stronger rolling adhesion on selectins than mature CD34– MNCs (14)(17). We have exploited this differential rolling characteristic of CD34+ HSPCs on selectins to capture and enrich them in flow. In this study, to determine whether CD34+ cells could be specifically captured, we perfused bone marrow MNCs over functionalized microtubes that were coated internally with adhesion molecules consisting of sP and CD34 antibody.
Materials and Methods
extraction of MNCs from bone marrow
We obtained bone marrow samples from individuals who gave written informed consent in accordance with policies of the Research Subjects Review Board of the University of Rochester. Under sterile conditions, samples were diluted 3-fold (1:3 vol/vol) with PBS (calcium and magnesium free), and 35 mL of diluted sample was slowly layered on 15 mL Ficoll (GE Health Care) in 50-mL Falcon tubes. The tubes were centrifuged at 800g for 20 min at room temperature. The thin buffy coat containing MNCs was carefully collected in separate tubes, washed twice in PBS, passed through a cell strainer (40 μm nylon; BD Biosciences) to remove aggregated cells, counted, and suspended in PBS enriched with calcium at a concentration of 107 viable cells/mL. To count white cells, we diluted a fraction 3-fold with Turks solution (0.01% crystal violet and 3% acetic acid in distilled water) and counted by use of a hemocytometer. We quantified cell viability by trypan blue dye exclusion and staining with 7-amino-actinomycin D and measuring the fluorescence at 655 nm on a FACSCalibur flow cytometer (BD Biosciences).
CD34+ HSPC isolation
We suspended bone marrow MNCs in PBS to a final concentration of approximately 108 cells/mL and used the Dynal CD34 progenitor cell selection system for isolation of CD34+ HSPCs as per the manufacturer’s instructions. CD34+ cells were reconstituted to the desired final concentration in Hanks balanced salt solution (HBSS) supplemented with 0.5% (wt/vol) human serum albumin, 10 mmol/L HEPES, and 2 mmol/L CaCl2. Mean CD34+ cell purity was 73% as determined by flow cytometry.
We performed rolling experiments using a parallel plate flow chamber (Glycotech) that fits into a 35-mm tissue culture-grade polystyrene culture dish with a gasket defining a rectangular channel (20 by 25 by 0.127 mm) when the chamber is fully assembled. We positioned the flow chamber with the rectangular channel over the area of the dish where sP had been adsorbed with laboratory tubing (Dow Corning) of appropriate lengths attached to the inlet, outlet, and vacuum ports. The fully assembled apparatus was secured to the stage of an IX81-epifluorescent, motorized, inverted microscope (Olympus America Inc.) equipped with a CCD camera (Hitachi) connected to an S-VHS videocassette recorder (Sony Electronics) to facilitate image capture. We used a syringe pump (New Era Pump Systems) to control the flow rate of the cell suspension.
Rolling cells were defined as cells translating at <50% of the calculated hydrodynamic free stream velocity (18); cells that remained stationary for more than 10 s or rolled <4 cell diameters were not classified as rolling. We determined velocities of single cells by use of a Matlab program designed to measure the change in position of the cell centroid in a given time period. Experiments were performed in triplicate to account for donor variability of bone marrow, and rolling velocities were reported as mean (SE).
cell capture microtube
Recombinant human sP/Fc chimera (rhP/Fc) (R&D Systems), anti-CD34 monoclonal antibody (ICO115; Santa Cruz Biotechnology), and normal IgGs (His-Tag monoclonal IgG1 from EMD Biosciences and monoclonal antihuman IgG from Sigma) at various concentrations in PBS were adsorbed on to the inside surface of blood-compatible microrenathane tubing (MRE-025, inside diameter 300 μm, length 50 cm; Braintree Scientific Inc.) with a 2-h incubation at room temperature in sterile conditions. After a gentle wash with PBS, nonspecific blocking of the lumen surface was accomplished with incubation of milk protein (5% in PBS) for 1 h. After another gentle PBS wash, the tube was filled with calcium-enriched HBSS (HBSS+, pH 7.4) to activate sP and maintained at room temperature for 30 min or stored overnight at 4 °C before cell perfusion. We prepared control tubes identically, with adhesion molecules replaced by PBS.
rhP/Fc chimera adsorption
Various concentrations of rhP/Fc were adsorbed onto the inner surface of the MRE tube for 2 h as described above without secondary blocking. We collected unbound rhP/Fc fractions in separate vials by gently displacing them with an equal volume of PBS. We measured concentrations of rhP/Fc in the initial and unbound samples by use of an easy-titer human IgG (γ chain) assay set (Pierce). First, a standard curve was prepared by using various known concentrations of rhP/Fc as per the set instructions; then the percentage of rhP/Fc bound on the surface was calculated using the equation [(rhP/Fcinitial − rhP/Fcnonadsorbed) × 100] ÷ rhP/Fcinitial.
capture, collection, and analysis of cells
After surface coating, cell-capture tubes were positioned on the stage of the IX-81 microscope coupled to a CCD camera for direct visualization of the adherent cells in the tube lumen. MNCs from bone marrow (107/mL in PBS enriched with calcium) were perfused through the tube at a rate of 8 μL/min (wall shear stress 0.5 dyn/cm2) for 1 h or at a rate of 40 μL/min (wall shear stress 2.5 dyn/cm2) for 12 min using a syringe pump system. We then washed the tubes with HBSS+ for 20 min to remove nonadherent cells and erythrocytes and eluted adherent cells using a combination of high shear (flow rate 40 μL/min), calcium-free PBS, and air embolism.
For quantitative flow cytometric analysis of CD34+ and CD133+ cells, we divided the samples into 3 fractions. The 1st fraction was incubated with antibodies to CD45-Alexa 488 (AbD Serotec) and phycoerythrin (PE)-conjugated CD34 (Santa Cruz Biotechnology), the 2nd fraction was incubated with CD45-Alexa 488 and isotype-PE for CD34 antibody, and the 3rd fraction was incubated with CD45-Alexa 488 and CD133-PE (Miltenyi Biotec) for 10–20 min at 4 °C. Cells were washed twice with PBS and resuspended in 250 μL PBS, and fluorescence data was acquired. MNCs were also analyzed via flow cytometry to determine CD34+ and CD133+ cell concentrations in the initial samples. FlowJo software was used to analyze the data.
To quantify the number of cells captured inside the entire tube, we perfused known numbers of KG-1a cells (CCL-246.1™, a CD34+ acute myeloid leukemic cell line; ATCC) or bone marrow MNCs into a sP-coated MRE tube on the microscope stage. Cells captured in 20 randomly selected regions were manually counted by brightfield video microscopy. Because the net volume of the liquid contained in a 50-cm MRE tube is 40 μL, we derived a constant factor by dividing the total number of cells in 40 μL of sample with the total number of cells in 20 regions. Experiments were done at least 6 times independently, and the average constant factor was found to be 72. Therefore, to obtain the total number of cells captured in the adhesion molecule coated flow device, we counted the number of cells in 20 randomly selected regions and multiplied by 72.
colony-forming cell assay
We collected captured cells on sP surfaces from 3 tubes that had at least 5-fold CD34+ HSPC enrichment. We also collected the elutant or flow-through sample containing nonadherent cells from the sP tubes. We mixed 100 μL of 5 × 104 viable cells with 1 mL human methyl cellulose–enriched media (R&D Systems), plated the mixture in 35-mm tissue culture plates, and incubated them at 37 °C, 5% CO2. Equal numbers of bone marrow MNCs were plated in the same manner, as were 10 000 CD34+ cells isolated using Dynabead extraction. After 14 d of incubation, colony-forming cells (CFCs) were counted under an inverted microscope and photographed.
We used paired t-test to analyze the results wherever necessary, at the α = 0.05 level of significance.
rolling of CD34+ HSPCs on sP
To compare the rolling and tethering efficiency of CD34+ HSPCs on rhP/Fc and recombinant human sP, we performed typical rolling experiments using highly enriched CD34+ HSPCs (73% pure) in a parallel plate flow chamber. The studies indicated a significantly slower rolling velocity for CD34+ HSPCs on rhP/Fc in comparison with sP at a concentration of rhP/Fc (0.5 mg/L), which is 20-fold lower than the concentration of sP (10 mg/L) (Fig. 1⇓ ). In light of these results, the rhP/Fc chimeric form was used in all subsequent cell-capture experiments.
adsorption of rhP/Fc on the MRE tube inner surface
Incubation of sP on the polystyrene surface of a flow chamber is known to cause effective immobilization by simple nonspecific physisorption. We therefore expected that similar adsorption would occur on the polyurethane surface of MRE tubing, but the efficacy of such an immobilization was not known. Therefore, as a 1st step, we sought to understand the efficacy of such a binding process by quantifying the amount of rhP/Fc immobilized on the MRE surface. Because rhP/Fc is conjugated with an Fc fragment of IgG, to evaluate the concentrations of rhP/Fc, we used a colorimetric Fc quantification assay set generally used to estimate IgG concentration in serum. The R2 value for the slope of the rhP/Fc standard curve (range 15–500 μg/L) was 0.9967, suggesting that the assay works well for this molecule. The percentage of rhP/Fc adsorbed on the surface was found to decrease with an increase in the concentration; however, the total amount of rhP/Fc adsorbed on the surface increased with an increase in concentration (Fig. 2⇓ ). We chose 40 mg/L, at which approximately 55% of rhP/Fc was immobilized, as the working concentration to achieve HSPC capture.
capture and enrichment of HSPCs
Whereas the number of MNCs captured on the control surface was negligible, significant numbers were captured on surfaces coated with CD34 antibody and sP (Fig. 3A⇓ and B), a result indicating that tethering of MNCs on the surface is mediated by adhesion molecules and not due to nonspecific binding on the surface. As expected, the total MNCs captured increased with an increase in sP concentration (Fig. 3C⇓ ). It is also clear from the figure that the antibody surface captured a consistently lower number of cells compared with sP surfaces. On average, the surfaces coated with antibody (100 and 200 mg/L) captured approximately 68 400 (5405) and 128 333 (4509) MNCs, respectively, whereas surfaces coated with rhP/Fc (40 mg/L) captured approximately 356 160 (27 085) MNCs (Fig. 3C⇓ ). Cell adhesion as a function of surface area was measured to be 4.2 (0.4) cells/mm2 for control tubes, whereas tubes coated with cell adhesion molecules exhibited adhesion rates of 142.9 (11.3) cells/mm2 for CD34 antibody (100 mg/L), 302.8 (31.3) for CD34 antibody (200 mg/L), 346.2 (15.5) cells/mm2 for rhP/Fc (20 mg/L), and 743.9 (56.6) cells/mm2 for rhP/Fc (40 mg/L). Furthermore, to completely eliminate the possibility of nonspecific adhesion of cells to Fc receptor, and to test whether the device works at higher physiological shear stresses, we conducted similar adhesion experiments in the tubes coated with 2 different IgG controls (100 mg/L), along with rhP/Fc-coated tubes (40 mg/L) at sample perfusion of 2.5 dyn/cm2. Results indicate that the cells fail to bind to the IgG controls (thus, nonspecific binding to Fc is minimal) and adhesion to sP chimera is unchanged from 0.5–2.5 dyn/cm2 (Fig. 3D⇓ ). In addition, incubating MNCs with anti–PSGL-1 antibody to block PSGL-1 receptor before perfusion completely abolished the ability of sP-coated tubes to capture cells, suggesting that the interaction of sP with PSGL-1 is crucial for cell capture (data not shown).
Flow cytometric analysis indicated that CD34+ HSPC purity in adhesion molecule–coated devices was significantly higher compared with the fraction of CD34+ cells in Ficoll-extracted MNCs [2.93% (0.86%)]. sP-coated surfaces yielded 17.8% (2.2%) purity, and the antibody-coated surfaces yielded 15% (3%) purity (Fig. 4A⇓ ). Fig. 4B⇓ and C represent typical dot plots for CD34-PE vs CD45-Alexa 488. Region 1 (R1), the upper and lower right quadrants, includes all mononucleated cells positive for CD45 (leukocytes). Analysis of R1 for CD34 positivity depicted 2% purity for the control MNCs and 17% enrichment for the sample isolated from sP flow devices. Preliminary analysis for CD45+/CD133+ cells indicated a approximately 11-fold enrichment in the sP-extracted sample [6.7% (2.8%) purity] compared with control mononuclear cells [0.59% (0.16%) purity].
viability and phenotypic characteristics of HSPCs
More than 90% of the MNCs from initial sample, sP-coated, and antibody-coated devices were viable, suggesting that the entire procedure is gentle enough to keep the cells viable and that the short time interaction of the cells with the adhesion molecules does not cause deleterious effects. Furthermore, the CFC assay using equal number of cells resulted in 1.9 times more numerous burst-forming units erythroid (BFU-E) and 1.4 times more numerous colony-forming units granulocyte-macrophage (CFU-GM) for the cells extracted from the sP device [BFU-E 37 (7) and CFU-GM 53 (4)] compared with bone marrow MNCs [BFU-E 19 (2) and CFU-GM 37.5 (4.5)] (Fig. 5⇓ ). An increase in the numbers of CFUs is evident, although this difference did not reach statistical significance (P >0.05). Surprisingly, CFC assay of the elutant or flow-through sample containing nonadherent cells from sP tubes resulted in very few CFCs, suggesting that most of the HSPCs are captured during perfusion [average BFU-E 1 (0) and CFU-GM 12 (2)]. In a separate experiment to compare the colony-forming potential, we plated approximately equal numbers of CD34+ cells from P-selectin–enriched samples and Dynabead–isolated samples. We found that the 2 samples produced similar numbers of CFUs [average BFU-E 81 (3) and CFU-GM 78 (2) from sP-enriched and BFU-E 91 (10) and CFU-GM 88 (14) from Dynabead samples]. Taken together, these results imply that the P-selectin–mediated enrichment of HSPCs not only maintains but also enhances the short-term hematopoietic functional phenotypes in vitro.
HSPCs have broad developmental potential and display plasticity (19)(20); their purification/enrichment is a prerequisite for hematologic transplantation therapies and basic stem cell research. Distinctive rolling behavior of CD34+ HSPCs (stronger adhesion and slower rolling) than mature CD34− leukocytes on selectin surfaces indicated a possibility that this phenomenon could be exploited for the selective capture of HSPCs from blood flow by simply implanting selectin-coated stents into circulation (14)(17). Supporting this hypothesis are the observations that selectins are implicated in the homing of HSPCs to bone marrow (12)(13)(14) and that sP ligand, PSGL-1, is expressed on HSPCs (15).
A commercially available recombinant chimeric extracellular sP conjugated with the Fc region (rhP/Fc) exhibited better performance in terms of stability and affinity toward CD34+ cell tethering than a recombinant nonchimeric extracellular sP (Fig. 1⇑ ), most likely because of a combination of increased valency and an increased fraction of rhP/Fc adsorbing in the proper orientation to expose the lectin-binding domains upward into the flow field. Hence, we used rhP/Fc as the coating material in our cell-capture studies. A 40 mg/L incubation concentration produces a surface adsorption of approximately 0.9 μg per tube, which corresponds to a mass per area deposition of 0.19 μg/cm2. Thus, multilayer adsorption is not believed to be an issue, as our deposition mass per area is <1 μg/cm2.
To develop implantable versions of the cell-capture device, it is imperative to use blood-compatible materials such as microrenathane tubing (www.BraintreeSci.com) (21). Adsorption of sP onto the inner surface of the tubing was found to be concentration dependent. Also, the number of MNCs captured on sP increased with an increase in sP concentration. Interpretation of these 2 results allowed us to decide on 40 mg/L as the reasonable working concentration for rhP/Fc in cell-capture studies. Note that the control surface coated with PBS alone or normal IgGs failed to capture MNCs, suggesting that the cells were captured in the sP- or antibody-coated devices primarily owing to the interaction of adhesion molecules with the cells and not because of nonspecific binding to the renathane surface or Fc receptor.
We were able to achieve 6- to 8-fold enrichment of HSPCs using adhesion molecule–coated flow devices. Enriched cells were viable and exhibited improved expansion ability in culture to produce both types of CFCs. Capture efficiency was significantly higher with sP than CD34ab, owing to the natural ability of selectins to tether and induce rolling of leukocytes in circulation (13)(14). On the other hand, antibody–antigen bonds typically require much longer binding times (approximately 30 min) than convection (flow) time scales. Nevertheless, the results were sufficient to confirm our goal of employing antibody-coated surfaces as a positive control flow device to capture HSPCs. We believe that in future studies it will be possible to improve the purity of HSPCs by engineering the sP molecule so that it will have higher affinity toward HSPCs or lower affinity toward CD34-MNCs. Alternatively, one could also achieve higher purities by identifying better elution methods for captured cells specific to CD34+ HSPCs or CD34− MNCs. Multiple cycles of enrichment could also be employed to improve purity and yield. In future work, it will also be important to characterize the engraftment potential of selectin-enriched stem cells in the appropriate in vivo model.
Currently, methods to purify or enrich CD34+ stem and progenitor cells from marrow or mobilized peripheral blood are costly or time-consuming; the less cumbersome panning methods presently available sacrifice purity or efficiency of capture. Explorations of direct methods of stem cell isolation that rely on physiological properties of the cell therefore have interest for the fields of transplantation, regenerative medicine, and gene modifications. Identifying means to temporarily isolate these cells without alteration of their surface phenotype or cell-cycle status has potential implications for exposure of these cells to therapeutic agents or directing them to injured or disrupted tissues.
sP can be used in a compact flow device to isolate HSPCs. To the best of our knowledge, this study represents the 1st time that an adhesion molecule has been used to capture adult stem cells directly from MNC fraction, rather than through immunological or density gradient processing of samples. This flow-based system mimics physiological events involved in normal cellular trafficking and supports the hypothesis that flow-based, adhesion molecule–mediated capture may be a viable alternative approach to the capture and purification of HSPCs.
Grant/funding Support: This work was supported by grants from the New York State Office of Science, Technology, and Academic Research (NYSTAR) to M.R.K.
Financial Disclosures: M.R.K. serves on the scientific advisory board of CellTraffix Inc., a company in which he holds financial interest.
Acknowledgments: The authors gratefully acknowledge Dooyoung Lee, Karen Rosell, Kuldeep Rana, Brian Duffy, Bryce Allio, and the flow cytometry core facility for contributing to this study.
↵1 Nonstandard abbreviations: HSPC, hematopoietic stem and progenitor cell; MNC, mononuclear cell; rhP/Fc, recombinant human P-selectin/Fc chimera; HBSS, Hanks balanced salt solution; MRE, microrenathane tube; PE, phycoerythrin; sP, P-selectin; CFC, colony-forming cell; BFU-E, burst-forming unit, erythroid; CFU-GM, colony-forming unit, granulocyte-macrophage.
- © 2008 The American Association for Clinical Chemistry