Surface-immobilization of adhesion peptides on substrate for ex vivo expansion of cryopreserved umbilical cord blood CD34⁺ cells

Abstract

The interaction between integrins and extracellular matrix proteins play an important role in the regulation of hematopoiesis. Human hematopoietic progenitor cells express very late antigen-4 (VLA-4) and VLA-5, which mediate their interaction with fibronectin by recognizing the connecting segment-1 (CS-1 and RGD motifs, respectively. In this study, we investigated the ex vivo expansion of human umbilical cord blood (UCB) CD34⁺ cells on synthetic substrates surface-immobilized with peptides containing the CS-1 binding motif (EILDVPST) and the RGD motif (GRGDSPC). These peptides were covalently conjugated to poly(ethylene terephthalate) (PET) film at a surface density of 2.0–2.3 nmol/cm². UCB CD34⁺ cells were cultured for 10 days in serum-free medium supplemented with recombinant human thrombopoietin, stem cell factor, flt3-ligand and interleukin 3. The highest cell expansion fold was observed on the CS-1 peptide-modified surface, where total nucleated cells, total colony forming unit, and long-term culture initiating cells were expanded by 589.6±58.6 (mean±s.d.), 76.5±8.8, and 3.2±0.9-fold, respectively, compared to unexpanded cells. All substrates surface-immobilized with peptides, including the control peptides, were more efficient in supporting the expansion of CD34⁺, CFU-GEMM and LTC-ICs than tissue culture polystyrene surface. Nevertheless, after 10-days of ex vivo expansion from 600 CD34⁺ cells, only cells cultured on CS-1-immobilized surface yielded positive engraftment, even though the frequency was low. PET surface immobilized with RGD peptide was less efficient than that with CS-1 peptide. Our results suggest that covalently immobilized adhesion peptides can significantly influence the proliferation characteristics of cultured UCB CD34⁺ cells.

1. Introduction

Umbilical cord blood (UCB) has been identified as a promising alternative source of hematopoietic stem cells (HSCs) for autologous and allogeneic hematopoietic stem cell transplantation [1]. However, the relatively low number of cells obtainable from a single umbilical cord restricts the widespread use of UCB as a viable source of transplantable hematopoietic cells in adults [2,3]. Ex vivo expansion of HSCs is a sensible strategy to produce sustainable supply of cells to overcome this limitation. To date, numerous protocols have been developed to expand UCB-derived CD34⁺ cells in culture. Many of these protocols involve suspension cultures in the presence of various combinations of early acting cytokines such as stem cell factor (SCF), Flt-3 ligand (FL), interleukin-3 (IL-3), as well as thrombopoietin (TPO), erythropoietin (EPO), and granulocyte macrophage-colony stimulating factor (GM-CSF). Only meager attention has been paid to the nature of the culturing surface that might influence the outcome of the expansion.

Accruing evidence indicates that HSC self-renewal and homing are adhesion dependent [4–8]. HSCs reside in specialized niches in bone marrow (BM) in vivo, in which they are sequestrated through multiple interactions with stromal cells and extracellular matrix (ECM) molecules [9]. This BM stem cell niche provides crucial signals that control HSC self-renewal, differentiation, migration and homing [10–15]. Adhesive interactions between ECM molecules and various integrin receptors on hematopoietic stem/progenitor cells (HS/PCs) constitute an important part of the signaling control [16]. Fibronectin (FN), as one of the most important ECM molecules, is directly involved in the adhesion and proliferation of HS/PCs [17,18]. More interestingly, this FN-HS/PC interaction is not constitutive—it can be modulated by a number of stimuli, including cytokines [19–21], chemokines [22], other adhesion molecules [23], and cell-cycle status—indicating the cross-talk between adhesive signaling and other cues in HSC niche.

Two of the most important adhesion domains in FN are the connecting segment-1 (CS-1) and the RGD motif. Both are recognized by surface receptors on the early hematopoietic progenitors. The CS-1 domain binds to VLA-4 integrin receptor (very late antigen-4, α₄β₁, CD49d/CD29) while the RGD sequence binds to VLA-5 integrin receptor (α₅β₁, CD49e/CD29) found on these early progenitor cells [24–26]. Evidence shows that FN-VLA-4 interaction plays an important role in hematopoiesis [27–30] and both CS-1 and RGD are structurally important for growth-supporting effects of FN [31]. Moreover, it has been demonstrated that the VLA-4 binding sequence of FN stimulates the proliferation of human cord blood CD34⁺ cells [32].

Several studies have investigated the effects of immobilized FN on hematopoiesis [4,20,33–35], and showed that the adhesion of HS/PCs to coated FN on substrates favors the maintenance of long-term reconstituting ability during ex vivo expansion. In these studies, immobilization of FN has been achieved by physical adsorption onto tissue culture surfaces. This adsorption is primarily driven by van der Waal interaction between hydrophobic regions of the protein and the surface. Such a non-specific adsorption may result in a poorly defined substrate due to the random orientation of the adsorbed FN molecule [36,37]. In addition, the relatively large molecular size of FN precludes the synthesis of a substrate bearing high density of active FN.

In an effort to create a defined bioadhesive substrate that can mimic the adhesive property of FN, we have covalently immobilized peptides representing CS-1 and RGD binding motifs of FN on a polymeric film. We hypothesized that such peptide-immobilized polymer substrates would be able to recapitulate partially the cell-adhesion interaction in the natural HSC niche. Moreover, by evaluating separately the effects of CS-1 and RGD peptide-immobilized substrates on cord blood HSC proliferation, we could dissect the relative contributions provided by these domains. In this study, we evaluated the ability of covalently attached CS-1 and GRGDSPC on polyethylene terephthalate (PET) film in supporting ex vivo expansion of cryopreserved human UCB CD34⁺ cells.

2. Material and methods

2.1. Materials

Cryopreserved human umbilical cord blood CD34⁺ progenitor cells were purchased from AllCell Inc. (San Mateo, CA, USA). The purity of CD34⁺ cell is 91%. SCF, FL, TPO and IL-3 were purchased from StemCell Technologies Inc. (Vancouver, BC, Canada). Poly (ethylene terephthalate) (PET) film with a thickness of 0.1 mm was purchased from Goodfellow (Cambridge, UK). Cell-adhesion peptides, CS-1 (ELIDVPST), CS-1i (ELIEVPST) were obtained from Synpep (Dublin, USA). GRGDSPC and GRGESPC were obtained from GL Biochem. (Shanghai, China). N-hydroxysuccinimide (NHS), dicyclohexylcarbodiimide (DCC) and other reagents were from Aldrich (Milwaukee, WI, USA), if not specified.

2.2. Surface immobilization of cell-adhesion peptides (Fig. 1)

Fig. 1.

Reaction scheme for surface immobilization of peptides on PET surfaces.

PET film was cut into circular discs of 15-mm in diameter and treated with ethylenediamine/ethanol mixture (1:1, v/v) at 50 °C for 40 min. Discs were then washed five times with copious amounts of distilled water and dried under vacuum. To introduce carboxyl groups to the polymer surface, 6 pieces of discs were immersed in 2 ml of 5 mm succinic anhydride and 1 mm 4-dimethylaminopyridine (DMAP) in N, N-dimethylformamide (DMF) in a glass vial, and incubated with gentle agitation for 8 h at room temperature. Discs were subsequently washed with ethanol and water, and then air-dried. The surface carboxyl groups were activated in 3 ml 0.1 m DCC and 0.1 mN-hydroxysuccinimide (NHS) in DMF with agitation at room temperature for 12 h. These discs were washed with 5 ml DMF twice. Peptide conjugation was carried out by incubating these discs in 2 ml of peptide solution (100 μg/ml in DMF) for 12 h at room temperature. The modified discs were rinsed thoroughly with DMF and water, and sterilized with 70% ethanol for 2 h. As negative controls, mutant form of CS-1 and RGD peptides, CS-1i (ELIEVPST) and GRGESPC, respectively, were immobilized onto PET film according to the same procedure. Immediately before cell seeding, discs were thoroughly rinsed with sterile phosphate buffered saline (PBS) and placed at the bottom of wells of a 24-well tissue culture plate.

2.3. Surface analysis

Surface amine density was quantified according to the method described by Kakabakos et al. [38]. Briefly, primary amino groups were first converted to sulfhydryl groups by reacting the modified films with excess amount of iminothiolane (ITL). The surface sulfhydryl groups were quantitated by the microBCA assay (Pierce, IL, USA) using l-cysteine to generate a standard curve. As a negative control surface for immobilization, amine groups were converted to carboxylic groups using succinic anhydride as described above. The presence of carboxylic groups was confirmed by TBO staining [39].

Surface wettability of modified PET was characterized by measuring the water contact angle at room temperature using a video contact angle goniometer (Advanced Surface Technology, Billerica, MA). The attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy was performed on a Nicolet 170SX apparatus (Thermo Electron Corp., Asheville, NC, USA).

2.4. Ex vivo expansion of cord blood CD34⁺ cells

CD34⁺ cells were cultured in 24-well tissue culture plates (NUNC, Denmark) at a seeding density of 1 × 10³ cells/ml in 600 μl serum-free medium (StemSpan SFEM, StemCell Technologies, Vancouver, BC, Canada) supplemented with 100 ng/ml of SCF, 50 ng/ml of FL, 50 ng/ml of TPO, 20 ng/ml of IL-3, and 40 μg/ml low-density lipoprotein (LDL, Academy Biomedical Company, Houston, USA) in the presence of 1% (v/v) antibiotic–antimycotic solution (Sigma, St Louis, MO). Cells were incubated at 37 °C in humidified 5% CO₂ in air for 10 days without medium change. Cells cultured on untreated tissue culture polystyrene (TCPS) surface without any disc were used as an internal control. Cells were harvested at day 10, washed in Hank's balance saline supplement with 2% FBS and counted. Cell viability was assessed by Trypan blue dye exclusion. The surface phenotype of expanded cell populations was analyzed by flow cytometry. Progenitors were quantitated using standard clonogenic assay and LTC-IC assay. For FAC immunofluorescence staining, cells were washed with HBSS containing 2% FBS and incubated for 30 min on ice with the appropriate antibody, and then washed with PBS containing 2% FBS. Cells were analyzed by dual-color or triple-color flow cytometry on a FACSCalibur analyzer (BD Biosciences, San Jose, CA). Antibodies used were fluorescein isothiocyanate (FITC)-conjugated anti-human CD34 and phycoerythrin (PE)-conjugated anti-human CD38 and PerCPCy5.5 conjugated anti-human CD45 or VLA-4PE, VLA-5PE (BD Biosciences). Relevant isotype controls were included to confirm specificity. At least 20,000 events were acquired. Data acquisition and analyses were performed using Cell Quest Pro software (BD Biosciences).

2.5. Colony-forming unit (CFU) assay

A known number of uncultured CD34⁺ cells or expanded cells were resuspended in IMDM containing 2% FBS, plated in semi-solid culture medium MethoCult GF⁺ H4435 (StemCell Technologies) at a final density of 1000 cells/ml, and aliquoted in 35 mm petri dishes (StemCell Technologies). Cells were then incubated at 37 °C in a humidified atmosphere of 5% CO₂. Erythroid colony-forming unit (CFU-E), colony-forming unit granulocyte/macrophage (CFU-GM), and granulocyte/erythroid/macrophage/megakaryocyte colony-forming unit (CFU-GEMM) were enumerated on day 14 under an inverted microscope. Colonies were scored according to standard morphologic criteria.

2.6. Long-term culture-initiating cell (LTC-IC) assay

LTC-IC content was determined by culturing the expanded cells on an irradiated, pre-established stromal layer. The murine fibroblasts, M2-10B4 (kindly provided by StemCell Technologies), were harvested by trypsinization, irradiated with 8000 cGy from a ⁶⁰Co source, and seeded at 3 × 10⁵ cells into 35-mm Petri dishes in MyeloCult H5100 (StemCell Technologies) supplemented with 1 μm hydrocortisone (Sigma). Cells were incubated at 37 °C for 5 weeks with a weekly half-medium change. At the end of the culture period, cells were harvested, washed, and plated directly for CFC assay as described above.

2.7. NOD/SCID mice engraftment assay

Nonobese diabetic/severe combined immunodeficient (NOD/ SCID) mice (Animal Resource Center, Perth, Australia) were maintained at the Satellite Animal Holding Unit of the National University of Singapore. All animals were handled according to Institutional regulations, under sterile conditions and maintained in microisolators in a pathogen-free environment. Mice at 6–8 weeks of age were irradiated with 350 cGy of whole-body irradiation from a ⁶⁰Co radiation source and within 24 h of irradiation, were given single injection of cells via the tail vein. Either freshly thawed cryopreserved human CD34⁺ UCB cells or total cells harvested from 10-day expansion cultures were mixed with 3 × 10⁵ CD34⁺-depleted human bone marrow cells pre-irradiated with 1500 cGy (carrier cells), and administered to mice. Two unmanipulated, non-irradiated mice, five irradiated uninjected mice, and five non-irradiated mice injected with 3 × 10⁵ carrier cells were used as negative controls. Following transplantation, mice were fed acidified water containing 1.1 g/l neomycin sulfate and 131 mg/l of polymixin B ad libitum. Mice were sacrificed 6–8 weeks post-transplantation.

After euthanasia, bilateral femora and tibia were harvested from each animal and bone marrow cells were flushed out with HBSS containing 2% FBS and 5% human serum using a syringe and a 27-gauge needle. Red blood cells were lysed by 0.16 m ammonium chloride and remaining cells were washed with HBSS supplemented with 2% FBS. Cells were then incubated for 30 min at 4 °C with PerCpCy5.5 and FITC conjugated monoclonal antibodies specific for human CD45 and CD41, respectively. The presence of human hematopoietic cells in mouse bone marrow was examined by flow cytometry to detect human CD45 antibody-labeled cells. Analysis was performed with a FACSCalibur system using CellQuest software and at least 40,000 gated events were acquired. Successful engraftment of human hematopoietic stem cell was defined by the presence of at least 0.1% of human CD45⁺ cells in NOD/SCID mouse bone marrow cells.

2.8. Statistics

All data are presented as mean±s.d. The statistical significance of the data obtained was analyzed by the Student's t-test. Probability values of p<0.05 were interpreted as denoting statistical significance.

3. Results

3.1. Peptides immobilization and surface characterization

The schematic representation for surface immobilization of peptides on PET film is shown in Fig. 1. Aminolysis of PET film with EDA introduced primary amino groups to the PET surface. The aminated PET film was treated with succinic anhydride to generate surface carboxyl groups, which were then activated with DCC and NHS in DMF to yield NHS esters that are reactive towards the N-terminal amino group of the peptide. Peptides were linked to the activated PET surfaces via stable covalent amide bond formation. Modified PET surfaces were characterized by ATR-FTIR analysis and water contact angle measurement. Similar ATR-FTIR spectra and water contact angles (Table 1) were recorded for all four modified films conjugated with different peptides. The ATR-FTIR spectra of unmodified and modified PET films suggested the presence of carboxyl groups on the carboxylated PET film and the amide groups (amide I and II peaks at 1650 and 1530 cm⁻¹) on peptide-conjugated PET films (data not shown), consistent with literature reports [40,41]. Water contact angle values increased significantly after peptide conjugation, suggesting that hydrophobicity of these surfaces has increased compared to the carboxylated PET surface. The amount of peptide conjugated to PET surface, quantified by BCA assay, was similar for all four peptide modified surfaces (2.01–2.29 nmol/cm²). These conjugation degrees are substantially higher than that obtained by other conjugation schemes [42,43]. This is likely due to the improved reaction efficiencies, particularly for the final coupling step, which was performed in organic solvent instead of aqueous buffer.

Table 1.

Characterization of PET surfaces immobilized with different peptides

Surface	Water contact anglea (degrees)	Density of peptideb (nmol/cm2)
TCPS	54.1±1.2
Carboxylated PET	50.7±1.0
CS-1-immobilized PET	64.1±3.2	2.10±0.20
CS-1i-immobilized PET	59.1±2.3	2.29±0.31
RGD-immobilized PET	59.2±4.1	2.01±0.10
RGE-immobilized PET	62.2±2.2	2.13±0.14

^aContact angles shown are mean±s.d from 10 measurements.

^bPeptide densities shown are mean±s.d. from three independently prepared samples.

3.2. Ex vivo expansion of CD34⁺ cells

After 10 days of expansion on all substrates under the described conditions, cells were over 95% viable as indicated by the trypan blue exclusion test. Cells cultured on peptide-immobilized surfaces assumed similar morphology: spherical and apparently adhesive to the surface (Fig. 2c, d). Nevertheless, this adhesion was weak as cells could easily detach from the substrates by gentle washing. The type of peptide conjugated to PET surface did not affect cell morphology significantly (data not shown). In contrast, cells cultured on TCPS frequently showed non-adherent cell clusters/clones (Fig. 2a).

Fig. 2.

Phase contrast micrographs of cultured cord blood CD34⁺ cells at day 10, on (a) and (b) TCPS; (c) CS1 peptide-immobilized PET; (d) RGD peptide-immobilized PET. (a) Shows the clusters of suspended cells on TCPS at day 10. (b)–(d) show cells with different morphologies observed adhering to the surfaces. Bar length: 50 μm.

The expansion efficiency was expressed as the fold expansion of total cells and CD34⁺ cells cultured on different surfaces over the input cells (Fig. 3). All surfaces supported the proliferation of all nucleated cells. PET substrate with CS-1 peptide showed comparable total fold of cell expansion as TCPS surface, whereas carboxylated PET surface yielded the lowest cell expansion (p<0.01). Interestingly, the addition of soluble CS-1 and/or RGD peptide in culture medium did not significantly change the total fold of cell expansion or fold of CD34⁺ cell expansion for cells cultured on CS-1 or RGD-immobilized PET surface (data not shown).

Fig. 3.

Expansion fold of total nucleated cells and CD34⁺ cells following a 10-day culture of human CD34⁺ UCB cells on different substrates. Cryopreserved human CD34⁺ UCB cells were seeded at 1000 cells/ml/well in a 24-well plate and cultured in the expansion medium (see Method section for details). CD34⁺ cells were determined by FACS analysis at the end of culture. Values are the mean±s.d. of 4–6 independent experiments, each conducted in triplicates.

All four peptide-immobilized substrates exhibited significant improved expansion of CD34⁺ population (Fig. 3). After 10 days of culture, CD34⁺ cell expansion on CS-1 and RGD peptide-immobilized PET surfaces were 82- (range: 49–110) and 75- (range: 40–96.4) fold, respectively, which were significantly higher compared to TCPS (14.0, range: 4–24). Cultures on PET surfaces immobilized with CS-1i and RGE peptides, which are mutant forms of CS-1 and RGD, respectively, resulted in slightly lower total cell expansion but no significant difference in CD34⁺ cell expansion.

3.3. Surface marker expression on expanded cells

The expression of cell surface markers on expanded cells was determined by flow cytometry. CD34⁺ cell fraction decreased to approximately 2% when cultured on TCPS for 10 days. Under the same conditions, however, cultures on peptide-immobilized PET surfaces yielded 10–15% CD34⁺ cells, a marked increase over TCPS. However, the values among the peptide groups were not statistically significant (p>0.05). There was no significant difference in the proportion of cells expressing CD38 (∼25%), or CD45 (∼75%) on all surfaces. A small percentage of the CD34⁺ cells cultured on peptide-immobilized surfaces acquired a CD38⁻ phenotype, leading to an apparent mean expansion of CD34⁺ CD38⁻ cells of around 2000 folds, compared to a corresponding mean expansion of 400 folds on TCPS. Lineage markers such as CD3 (lymphocytes), CD13, CD33 (granulocytes), CD14 (monocytes), CD15 (neutrophils), CD19 (Pre-B), CD41a (megakaryocytes), and glycophorin A (erythroids) were also evaluated (Table 2). A higher proportion of cells cultured on TCPS acquired CD13, CD14 and CD15 surface proteins, indicating maturation of cultured cells along the granulocytic and monocytic lineages. Interestingly, the 10-day ex vivo expansion of UCB CD34⁺ cells resulted in a significant decline in relative numbers of CD3, CD19 and CD33-expressing cells on all surfaces tested. Peptide-immobilized surfaces resulted in a higher expression level of the erythrocyte precursor marker α-GpA, but no significant difference was observed for CD41 expression compared to TCPS.

Table 2.

Surface marker expression after a 10-day ex vivo expansion on different surfaces^a

Day 0	CD34 91.0±1.0	CD34+ CD38− 3.5±0.4	CD13	CD14	CD15	CD41	CD45	αGpA	CD34+ VLV-5+	VLA-5
Cells expanded on different surfaces for 10 days:
TCPS-	2.3±2.1	1.1±1.6	61.4±6.7	46.1±4.1	44.7±5.5	7.3±1.3	80.5±8.9	12.8±2.0	2.2±2.3	64.7±3.1
COOH-PET	6.1±1.8	4.9±2.2	N.A.	N.A.	N.A.	N.A.	N.A.	N.A.	N.A.	N.A.
CS-1-b	11.1±4.5	9.0±2.3	43.1±5.5	35.2±5.0	29.0±4.3	8.8±0.9	87.8±7.7	17.2±3.1	10.9±4.7	81.5±4.4
CS-1i-b	11.9±3.6	8.1±1.1	N.A.	N.A.	N.A.	N.A.	N.A.	N.A.	N.A.	76.5±4.6
RGD-b	12.5±3.1	10.2±0.6	51.2±4.7	31.4±4.5	33.0±2.1	7.2±2.0	82.1±8.1	21.1±1.1	11.5±4.3	78.6±4.0
RGE-b	13.3±4.3	10.7±1.2	N.A.	N.A.	N.A.	N.A.	N.A.	N.A.	N.A.	80.0±6.7

^aCD49d expressed on all cell surfaces (>99%); CD3⁺, CD19⁺ and CD33⁺ cells were not found in all groups.

^bPET membranes immobilized with these peptides.

We also examined the expression of VLA-4 and VLA-5 on expanded cells. Cells expanded on all tested surfaces (>99%) showed high CD49d (α₄) expression. The mean CD49e (α₅) expression level was significantly higher (Table 3) on cells expanded on peptide-immobilized surfaces than on TCPS. It is worth noting that CD34⁺ cells that were CD45^bright also showed strong α₄ and α₅ expressions.

Table 3.

Effect of immobilized peptides on the ex vivo expansion of CFUs and LTC-ICs

Substrate	CFU per 100 initial cellsa				LTC-IC per 100 initial cellsb
	CFU-GM	CFU-E	CFU-GEMM	CFU total
Uncultured cell	25±9.6	4.1±2.5	9.0±1.2	38.0±7.1	8.0±5.9
TCPS	1606.5±186.7	418.5±93.3	123.8±34.5	2189.0±181.7	5.6±2.0
Carboxylated PET	1254.5±116.7c	737.8±58.3c	341.8±17.7c	2332.8±81.7	13.0±4.7c
CS-1-immobilized PET	1870.5±163.3	495.3±44.7	528.0±70.0c	2893.8±245.6	25.7±7.0c
CS-1i-immobilized PET	979.2±115.7c	1001.8±115.7c	209.8±58.3	2188.8±135.0	19.6±3.9c
RGD-immobilized PET	1331.2±85.0	421.0±77.3	226.8±85.0	1661.8±128.3	21.0±6.1c
RGE-immobilized PET	1596.4±123.3	836.0±124.0c	450.0±163.3c	2992.0±193.3	21.9±3.3c

^aValues shown are mean±s.d. of 3–4 experiments in 4–6 replicate.

^bValues shown are mean±s.d., n = 5–6.

^cIndicates p<0.05 in comparison with TCPS.

3.4. Clonogenic assay

CFU and LTC-IC assays were used to evaluate the fraction of primitive cells in the expanded cultures. CS-1-immobilized surface resulted in the highest number of CFU-GM with a ∼75-fold increase over unexpanded cells compared to a ∼39-fold and ∼64-fold increase for CS-1i-immobilized surface and TCPS, respectively (Table 3). RGD peptide-immobilized surface seemed to have a negative effect on CFU expansion compared with CS-1-immobilized surface and TCPS surface. The expansion fold of both total CFU and CFU-GM was substantially lower among cells expanded on RGD peptide-immobilized surface, compared with TCPS and CS-1-immobilized substrates. All modified surfaces, including the carboxylated PET, appeared to support the expansion of the more primitive population, CFU-GEMMs, more efficiently than the TCPS surface. Among them, the CS-1-immobilized surface was the most efficient, yielding a 58.7-fold expansion compared to a 13.8-fold on TCPS surface. The RGD-immobilized PET surface again showed a lower expansion of CFU-GEMMs than CS-1- and RGE peptide-immobilized surfaces (Table 4).

Table 4.

Survival of irradiated NOD/SCID mice in the different experimental groups

Mice group	Cells implanted	Mice number	Survival
1	300,000 carrier cells	5	0/5
2	600 unexpanded cells	5	3/5
3	20,000 unexpanded cells	6	3/6
4	Cells expended from 600 initial cells on TCPS	6	3/6
5	Cells expended from 600 initial cells on CS-1 immobilized PET	5	4/5
6	Cells expended from 600 initial cells on CS-1i immobilized PET	5	2/5
7	Cells expended from 600 initial cells on RGD immobilized PET	5	3/5
8	Cells expended from 600 initial cells on RGE immobilized PET	5	5/5
9	Non-irradiated, non-injected	2	2/2

Cultures on all four peptide-immobilized as well as carboxylated surfaces led to higher LTC-IC expansion than that on TCPS. Cells expanded on CS-1-immobilized surface showed more than 3-fold increase in LTC-IC counts, whereas expansion on TCPS resulted in a 30% decrease in LTC-IC counts compared to unexpanded cells (Table 3). These data showed that culture of cryopreserved UCB CD34⁺ cells in serum-free medium in the presence of SCF, TPO, FL3 and IL-3 induced extensive expansion of lineage-committed progenitor cells and expansion of LTC-ICs; and peptide-immobilized substrates were more efficient than TCPS. Among them, CS-1-immobilized PET surface was the most efficient in expanding CD34⁺ cells, CFU-GEMMs and LTC-ICs.

3.5. Engraftment of expanded human hematopoietic cells in NOD-SCID mice

The goal of this experiment was to determine whether cells expanded on peptide-immobilized PET surfaces retained their engraftment potential. Cells harvested from a 10-day expansion culture were intravenously injected into sub-lethally irradiated NOD/SCID mice together with 3 × 10⁵ carrier cells (irradiated, CD34⁺ depleted human bone marrow cells). Engraftment was assessed by the percentage of human CD45⁺ cells present in murine BM cells 6–8 weeks following implantation. As positive controls, 600 or 20,000 unexpanded CD34⁺ UCB cells were transplanted using the same protocol. Representative FACS analysis of human CD45⁺ cell subpopulation present in the murine bone marrow is showed in Fig. 4. The presence of 0.1% or higher human CD45⁺ cells was used as a criterion for successful engraftment in the bone marrow of NOD/SCID mice (Fig. 5).

Fig. 4.

Representative FASC profiles of marrow cells from NOD/SCID mice transplanted with unexpanded or expanded CB cells. (a) Gated region (R1) set for analysis; (b) isotype control; (c) uninjected mouse; (d) mouse transplanted with 600 unexpanded cells; (e) mouse transplanted with 20,000 unexpanded cells; mouse transplanted with 600 initial cells expanded for 10 days on (f) TCPS; (g) CS-1 peptide-immobilized PET surface; (h) GRGDSPC-immobilized PET surface.

Fig. 5.

Engraftment efficiency of human CD45⁺ cells in bone marrow of NOD/SCID mice transplanted with (1) irradiated carrier cells alone; (2) 600 CB CD34⁺ cells; (3) 20,000 CB CD34⁺ cells; total cells expanded from 600 CB CD34⁺ cells for 10 days on (4) TCPS; (5) CS-1-immobilized PET surface; (6) CS-1i-immobilized PET surface; (7) RGD-immobilized PET surface; (8) RGE-immobilized PET surface. Mice were sacrificed 6–8 weeks after transplantation, and cells harvested from mouse BM were analyzed by flow cytometry. Uninjected mice were sacrificed and their bone marrow cells were analyzed as control (Group 9).

The irradiated (350 cGy) and untreated NOD/SCID mice all died within 2 weeks after irradiation. Of the six NOD/SCID mice infused with 20,000 pristine or unexpanded UCB CD34⁺ cells, three survived. All three of the survived mice showed engraftment frequency of 0.2% and above. Among all other tested groups, only cells expanded on CS-1-immobilized PET surface resulted in successful engraftment—3 out of 4 survived mice had positive engraftment, although the frequency was only 0.1–0.2%. However, the sample size was not sufficient for statistical analysis.

4. Discussion

The success of hematopoietic stem cell transplantation is critically dependent not only on the quality of the cells, but also on the cell dose administered [44,45]. In addition, the time to engraftment is directly related to the phenotypic composition of cells in the donor material. Therefore, any strategy for short-term ex vivo expansion of UCB cells needs to take these issues into consideration. In designing an ex vivo expansion system, it would be beneficial to incorporate salient features of the natural HSC niche for the purpose of maintaining self-renewal and proliferation properties of HS/PCs in culture. As cell adhesion has been shown to play an important role in controlling the HS/PC self-renewal and differentiation, we hypothesized that immobilization of cell-adhesion peptides on substrates provides a structurally defined functional substrate that imparts VLA-4- and VLA-5-mediated adhesive signaling.

This study demonstrates the expansion of UCB CD34⁺ cells on biomimetic PET surfaces conjugated with peptides representing the fibronectin adhesion domains, CS-1 and RGD. These peptide-immobilized PET substrates support adhesion of HS/PCs. It was observed that cells cultured in normal liquid culture tended to aggregate whereas cells cultured on peptide-immobilized PET films were adherent to the surfaces. This may be suggestive of an obligatory requirement for HS/PCs to engage their integrin receptors. In the absence of a suitable substrate, the cells would seek to establish adhesive interactions with their neighbors. Alternatively or in combination with other cues, the absence of an adhesive substrate may permit rapid cell differentiation leading to the production of cells that form colony-like aggregates (e.g. erythrocytes).

In particular, the ability of surface-immobilized adhesion peptide to selectively expand CD34⁺ cells was revealed in this study. While total cell expansion was comparable to that of cultures on standard tissue culture surface, peptide-immobilized PET substrates support higher expansion of CD34⁺ cells, CFUs and LTC-ICs. Among all the peptides immobilized, CS-1 peptide immobilization provides the highest expansion fold of CFU-GM, CFU-GEMM, total CFU and LTC-ICs. Immobilization of an inactive version of the CS-1 peptide, CS-1i (ELIEVPST) [46] is slightly less effective. Lower expansion outcome was observed, when peptide representing the more ubiquitous adhesion domain of fibronectin, RGD, was immobilized. No significant difference was observed between cells cultured on surfaces modified with RGD or its inactive mutant form, RGE. This suggests that these two adhesive domains of fibronectin, CS-1 and RGD, may make unequal contribution towards self-renewal of HS/PCs.

Phenotypic evaluation demonstrated that lineage-associated surface markers, such as CD13, CD14 and CD15 were expressed on the expanded cells. It is known that besides progenitor cells, CD34⁺-selected cells from UCB contain cells at various stages of differentiation and CD34⁺ CD38⁻ phenotype has been suggested as a more accurate measure of primitive cell proportion. Interestingly, we observed that the expression of the CD38 marker declined over the 10-day culture period without clear benefit for engraftment activity. The lack of expression of CD38 on CD34⁺ cells may simply be the consequence of serum-free culture condition, a peculiarity reported by Donaldson et al. [47]. The lack of expression of CD3 or CD19 on expanded cells indicated that differentiation along the T- and B-lymphoid pathways are deficient. Conversely, myeloid marker expressing cells predominate in the expanded populations, implying that the modified surfaces have skewed the differentiation towards the myeloid lineage.

We have investigated whether ex vivo expansion of CD34⁺ cells on peptide-immobilized PET substrates affects the expression of VLA-4 and VLA-5 integrins, which are known to be important to HSC homing to bone marrow. We found that α₄ (CD49d) and α₅ (CD49e) are highly expressed on the expanded CD34⁺ cells. However, this does not provide any correlative indication of the engraftment capability of the expanded cells. It is possible that ex vivo expansion has altered the activities of the integrin receptors without affecting their expression level, a postulation propounded by Ramirez et al. [48]. Another possibility for low engraftment efficiency observed is that the cell number used for engraftment assay is too low, i.e. the expansion of primitive cells is less than the apparent expansion of CD34⁺ cells. These results also suggest that CD34⁺ cell number is not a good indicator for engraftment outcome, as the numbers of injected CD34⁺ cells for the groups of peptide-immobilized substrates range from 255,000 to 350,000, much higher than the positive control group (20,000 CD34⁺ cells).

5. Conclusion

This study indicates that cryopreserved UCB-derived CD34⁺ cells can be significantly expanded in our short-term culture system by exposing cells to surface-immobilized adhesive peptides. CS-1 peptide-immobilized PET surface mediates the highest expansion for CD34⁺ cells, CFU-GEMMs and LTC-ICs, which is 4.1-, 4.3- and 4.6-fold higher than TCPS, respectively. Cultures on other peptide-immobilized surfaces also yield significantly improved expansions of the above cells compared with TCPS, although slightly lower than the CS-1-immobilized surface. Despite the high CD34⁺ cell expansion, cells cultured on peptide-immobilized surfaces fail to engraft sub-lethally irradiated NOD/SCID mice, except those on CS-1-immobilized substrate. Evidence suggests that the immobilized peptides play a role in the expansion of HS/PCs in part by regulating their adhesion to modified PET surface, even though additional mediators may also be involved. The ability to influence UCB-derived HS/PC proliferation and differentiation by manipulating the culture surface would have important clinical implications.