Growth Induction and Low-Oxygen Apoptosis Inhibition of Human CD34+ Progenitors in Collagen Gels

Various reports have indicated low survival of injected progenitors into unfavorable environments such as the ischemic myocardium or lower limb tissues. This represents a major bottleneck in stem-cell-based cardiovascular regenerative medicine. Strategies to enhance survival of these cells in recipient tissues have been therefore sought to improve stem cell survival and ensure long-term engraftment. In the present contribution, we show that embedding human cord blood-derived CD34+ cells into a collagen I-based hydrogel containing cytokines is a suitable strategy to promote stem cell proliferation and protect these cells from anoxia-induced apoptosis.


Introduction
e success of progenitor cells' transplantation in cardiovascular regenerative medicine depends on efficient engrament into the hostile ischemic environment [1]. In this framework, cell survival is a relevant issue. In fact, several investigations have shown that injected (stem) cells undergo a relatively rapid clearance from ischemic heart and lower limbs, which may lead to dramatic efficacy reduction. In addition, the absence of suitable conditions to preserve innate repair stem cell function into recipient tissues may lead to adverse effects such as chronic in�ammatory responses [2,3].
e design of advanced biomaterials in cardiovascular regenerative medicine may be a solution to enhance stem cells' engrament efficiency. In fact, the combination of extracellular matrix components with factors promoting stem cell proliferation and survival may help reproduce a "niche-" like three-dimension (3D) environment allowing optimal engrament and differentiation, and enhancing their in vivo repair function [4][5][6]. In the present study, we show that addition of a cytokine cocktail to a collagen I-based hydrogel formulation induced proliferation and signi�cantly reduced death of human cord blood-derived CD34 + cells cultured in low oxygen, a typical condition encountered into ischemic tissues.

Materials and Methods
CD34 + cells were isolated from human cord blood samples as already established by us [7]; this was performed according to an institutional review board formal approval for cord blood collection (Melzo Hospital; December 12, 2008; authorization no 843). Brie�y, cord blood samples were diluted 1 : 3 with phosphate buffered saline (PBS); diluted blood was loaded onto a Ficoll gradient (Lymphoprep) for separating cord blood mononuclear cells. Isolation of CD34 + cells was performed using MINI-MACS system (Miltenyi). Collagen I solution was prepared as already described [8]. Brie�y, this was performed by neutralizing collagen type I solution (PureCol 3 mg/mL, Nutacon; dissoved in 0.01 N HCl) with 10X phosphate buffered saline (PBS) and 0.1 M NaOH. is liquid working solution (2.4 mg/mL collagen I, PBS 1X, 0.01 M NaOH pH 7.4) was kept on ice until use. Before adding cells, a further dilution of collagen working solution (to a �nal 1.5 mg/mL concentration) was performed by adding Stem Span culture medium (Stem Cell Technologies) with or without stem cell factor (SCF; 100 ng/mL), interleukin-6 (IL6; 20 ng/mL), interleukin-3 (IL3; 20 ng/mL), and fetal liver tyrosine kinase 3-ligand (Flt3-L; 100 ng/m), all from Stem Cell Technologies. CD34 + cells (2 × 10 7 cells/mL in 10 L PBS) were �nally added to the collagen-cytokines mixture and gently mixed to allow a homogeneous distribution. Matrix/cell suspensions were then incubated at 37 ∘ C in a humidi�ed chamber for 15 minutes, aer which they were transferred into 24 well plates (Costar); the formation of a three-dimensional semisolid collagen matrix was completed in these wells aer further 45 minutes at 37 ∘ C. Each well containing CD34 + cells seeded matrix (3D condition) was, �nally, covered with Stem Span alone or containing the cytokines mix at the same concentration used in the matrix. As a control, the same number of cells was cultured in normal liquid culture using the same medium and cytokines concentration (2D condition). To assess cell phenotype aer culturing in 3D conditions, collagen gel was digested using collagenase IV solution, before recovering cells by centrifugation, staining with monoclonal antibodies against CD14, CD31, CD34, and CD133 antigens (all from BD Biosciences), and analysis with a FACScalibur (BD Biosciences) �ow cytometer. To perform culture in hypoxia, culture plates were placed into a humidi�ed hypoxia chamber, hermetically sealed, and perfused with a 95% N 2 /5% CO 2 atmosphere. Aer 7 days, cells were recovered by collagenase IV digestion and immediately stained for annexin V (AnnV) and propidium Iodide (PI) to detect apoptosis.

Results and Discussion
Figure 1(a) shows the experimental scheme adopted to perform our tests. Aer an initial preexpansion for two days in liquid culture, cells were cultured for one week in 2D and 3D conditions. Figure 1(b) shows that the majority of cells recovered from culture wells by hydrogel digestion were entrapped in the collagen matrix. Moreover, presence of cytokines induced the formation of typical clusters of rapidly expanding cells, present at different focus planes in the matrix. To screen for differences in the growth of cells in the presence or the absence of cytokines in 2D and 3D conditions, a �ow cytometry analysis of cells recovered aer 7 days of culture was performed. Figure  1(c) shows the number of cells expressing typical "early" endothelial progenitor cells (EPCs) markers CD34, CD133, CD31, and CD14 [1,9], also in comparison with that of cells cultured in 2D conditions for 7 days in the absence of cytokines (2D CTR). e presence of cytokines in the matrix promoted a net increase in the number of immature cells, as revealed by the signi�cant elevation of CD34 + and CD133 + cells in 3D versus 2D. Interestingly, the number of cells expressing CD31, a typical endothelial cells marker, was similarly increased in 2D and 3D, while the number of CD14 + monocytes was only marginally changed. Taken together, these results suggest that presence of cytokines in the 3D environment favors the expansion of immature EPCs, without increasing their monocyte commitment.
Exposure to lower than 20% oxygen is a strategy to enhance formation of endothelial cells' colonies (CFU-EC) by immature bone-marrow-derived CD133 + cells [10] or to preserve clonogenic ability of human hematopoietic progenitors [11]. In fact, oxygen level around or below 1% is a concentration normally experienced by the most immature hematopoietic progenitors in the bone marrow niche, where it appears a critical component of stemness maintenance [12,13]. On the other hand, exposure of immature progenitors to anoxia, such as in the ischemic myocardium, has been also reported to enhance death due to induction of apoptosis [14]. To explore whether CD34 + cells undergo cell death when cultured in oxygen deprivation conditions, cultures in anoxia chambers were set up, followed by AnnV/PI staining and �ow cytometry [15]. As shown in Figure 2, cells cultured in 2D for 7 days in anoxia, either in the presence or the absence of cytokines, maintained a viability around 60%. In addition, they showed a similar percentage of cells in early apoptosis (AnnV + /PI − ) and late apoptosis (AnnV + /PI + ) stages. By contrast, culture in 3D conditions for the same time extent improved survival to about 80% and reduced the percentage of cells in late apoptosis. Interestingly, the effect of 3D culture was observed only in the cytokinescontaining condition, while embedding in the hydrogel alone was not sufficient to rescue cells from low-oxygeninduced cell death, indicating the 3D/cytokines combination as the most powerful treatment preserving CD34 + cells viability.

Conclusions
e results shown in the present study show that embedding of primary CD34 + progenitors into cytokines-containing 3D matrices is a suitable strategy to maximize expansion under normal oxygen content and to protect them from low-oxygen-induced apoptosis. ese effects may result from an optimal cytokine "presentation" in the 3D environment compared to liquid cultures [16] and/or from a synergistic action of cytokine-dependent prosurvival pathways, cell proliferation signaling, and matrix-dependent cues [17]. �hether or not culturing in 3D collagen gel has an in�uence on CD34 + cell functions aer culturing in low oxygen, is currently being investigated in our laboratory. However, the relative ease of the collagen I mixture preparation and its formulation with components which are readily available for human use suggests that the embedding method shown here may be used to maximize stem cell transplantation efficiency in patients with heart and peripheral ischemia or to optimize stem cells seeding into biomaterials for the production of next generation bioarti�cial tissues.  2D control 3D control 2D cytokines 3D cytokines 2D control 3D control 2D cytokines 3D cytokines * * (c) F 2: Increased survival of 3D versus 2D cultured CD34+ cells under severe hypoxia conditions. (a-b) Representative �ow cytometry contour plots showing the physical appearance (FSC-SSC plot) and the annexinV/PI staining of CD34+ cells cultured for 7 days in an atmosphere containing 95%N 2 /5%CO 2 . Multicolor staining of the different regions indicates, respectively, viable cells (green), early apoptotic cells (red), late apoptotic cells (blue), and necrotic cells (gray). (c) �uanti�cation of viable and late apoptosis cells. No signi�cant differences were observed in viable and late apoptotic cells' number in 2D cultures (±cytokines) and 3D culture without cytokines; by contrast, the number of viable cells was signi�cantly increased and that of late apoptotic cells was signi�cantly reduced, by cytokines-supplemented matrix culture. * indicates by paired Student's t-test (3D cytokines versus 3D control); .