Journal of Membrane Science 351 (2010) 104–111
Contents lists available at ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Direct ex vivo expansion of hematopoietic stem cells from umbilical cord blood on membranes Akon Higuchi a,b,∗ , Siou-Ting Yang a , Pei-Tsz Li a , Miho Tamai c , Yoh-ichi Tagawa c , Yung Chang d , Yu Chang e , Qing-Dong Ling f,g , Shih-Tien Hsu h a
Department of Chemical and Materials Engineering, National Central University, Jhongli, Taoyuan 32001, Taiwan Department of Reproduction, National Research Institute for Child Health and Development, 2-10-1 Okura, Setagaya-ku, Tokyo 157-8535, Japan Graduate School of Bioscience and Biotechnology, Tokyo Institute of Technology, B-51 4259 Nagatsuta-cho, Midori-ku, Yokohama, Kanagawa 226-8501, Japan d Department of Chemical Engineering, R&D Center for Membrane Technology, Chung Yuan Christian University, 200, Chung-Bei Rd., Chungli, Taoyuan 320, Taiwan e Kaohsiung Medical University Hospital, No. 100, Tzyou 1st Road, Kaohsiung 807, Taiwan f Cell Biology and Anatomy Laboratory, Cathay Medical Research Institute, Cathay General Hospital, No. 32, Ln 160, Jian-Cheng Road, Hsi-Chi City, Taipei 221, Taiwan g Institute of Systems Biology and Bioinformatics, National Central University, No. 300, Jhongda Rd., Jhongli, Taoyuan 32001, Taiwan h Department of Community Medicine, Li Shin Hospital, 77, Kuangtai Road, Pingjen City, Tao-Yuan County 32405, Taiwan b c
a r t i c l e
i n f o
Article history: Received 8 October 2009 Received in revised form 13 January 2010 Accepted 15 January 2010 Available online 22 January 2010 Keywords: Umbilical cord blood Surface-modified membranes Hematopoietic stem cells Polyurethane Cell culture
a b s t r a c t We have developed a direct ex vivo hematopoietic stem cell (HSC) expansion method involving filtration of umbilical cord blood (UCB) through polyurethane foaming membranes. Most of the red blood cells and mononuclear cells flow through the membranes, but HSCs remain on the membranes. The use of this method reduces the working time to less than 30 min before culture of HSCs, while conventional purification of HSCs from UCB that have been purified by the Ficoll–Paque procedure, followed by magnetic-activated cell sorting, results in loss of cells (i.e., a yield less than 15% of HSCs), damages HSCs, and takes a long time to perform (e.g., 5–8 h). Following filtration of the UCB and washing of the membranes, the membranes can be placed into culture medium where the HSCs can be expanded in a three-dimensional (3D) environment ex vivo, such as the bone marrow niche. Direct ex vivo expansion of HSCs from UCB resulted in an increased number of cells (i.e., 6.6- to 45.7-fold), and the expanded cell populations showed good hematopoietic ability in colony-forming unit assays. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Hematopoietic stem cells (HSCs) are multipotent cells that have the specific capacity to self-renew and differentiate into all mature blood cell types [1,2]. Umbilical cord blood (UCB) is a promising alternative to bone marrow as a source of HSCs for hematopoietic stem cell transplantation [3] for the treatment of a variety of hematological disorders and as a supportive therapy for malignant diseases [4]. However, the low number of HSCs that can be obtained from a single umbilical cord blood donor limits the feasibility of direct transplantation of umbilical cord blood for the treatment of pediatric patients. Numerous efforts have been devoted to ex vivo expansion of HSCs to improve engraftment time and reduce the graft failure rate; successful ex vivo expansion of HSCs is of particular importance in the development of therapies for adult patients
∗ Corresponding author at: Department of Chemical and Materials Engineering, National Central University, No. 300, Jhongda Rd., Jhongli, Taoyuan 32001, Taiwan. Tel.: +886 3 422 7151x34253; fax: +866 3 280 4271. E-mail address:
[email protected] (A. Higuchi). 0376-7388/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2010.01.034
[5–8]. It is necessary to develop appropriate strategies and methods to expand HSCs ex vivo so that UCB can serve as a source of transplantable HSCs for adult patients [4,9–12] suffering from a variety of hematological disorders. The native bone marrow microenvironment, a complex network of stromal cells and extracellular matrix (ECM), serves as a stem cell niche that regulates HSC functions such as self-renewal, proliferation, homing and fate choice [1,4,13–17]. The main research source of HSCs used for ex vivo expansion has been commercially available cryopreserved CD34+ cells [1,4,12], mononuclear cells and CD34+ cells purified from mononuclear cells followed by magneticactivated cell sorting (MACS) [18,19]. The cryopreserved CD34+ cells are also purified from mononuclear cells followed by MACS or flow-associated sorting method (FACS). The Ficoll–Paque method is generally used to obtain a mononuclear cell population from peripheral blood (PB), umbilical cord blood or a bone marrow (BM) solution [20]. This method involves several centrifugation procedures for density gravimetry, washing steps and an erythrocyte lysing step. Therefore, it takes 5–8 h to obtain purified CD34+ cells by the Ficoll–Paque and MACS procedures, and the yield of CD34+ cells is typically less than 15%.
A. Higuchi et al. / Journal of Membrane Science 351 (2010) 104–111
Here, we propose direct ex vivo expansion of HSCs on polyurethane (PU) foaming membranes from UCB without usage of centrifugation procedures such as Ficoll–Paque method, but by filtration of UCB through the membranes where most of red blood cells and mononuclear cells permeate through the membranes, and HSCs remain on the membranes. After permeation of UCB and washing solution, the membranes can be directly inserted into culture medium where HSCs can expand in three-dimensional (3D) environment ex vivo as bone marrow niche. 2. Materials and methods 2.1. Materials The base membranes used for chemical modification were polyurethane (PU) foaming membranes (Ruby Cell S, Toyo Polymer Co., Ltd.) and PU foaming membranes containing 0.61% epoxy (PU–epoxy), which were plasma-polymerized using glycidyl methacrylate [21–24]. The average pore size of the PU and PU–epoxy membranes, evaluated using Capillary Flow Porometer measurements (Porous Materials Inc.), was 5.2 m. The PU and PU–epoxy membranes had 86% porosity and were 1.2 mm thick. Leukocyte removal filters (Imguard III-RC, Termo Co.) were used as received. Serum-free medium (StemSpanTM SFEM), StemSpanTM CC110 cytokine cocktail and MethoCultTM GF H4434 were purchased from StemCell Technology Inc. (British Columbia, Canada). Low-density lipoprotein (LDL, L8292) was purchased from Sigma–Aldrich Co. (St. Louis, MO). Anti-glycophorin A antibody conjugated with PE (phycoerythrin) (IM2211, Beckman Coulter Co.), anti-CD34 antibody conjugated with PE (A07776, Beckman Coulter Co.) and anti-CD45 antibody conjugated with FITC (IM2653K, Beckman Coulter Co.) were used as received. 7-AAD (A07704, Beckman Coulter Co.), Optilyse C (IM1401, Beckman Coulter Co.) and Flow-count bead solutions (7547053, Beckman Coulter Co.) were also used as received. Other chemicals, purchased from Sigma–Aldrich, Co. (St. Louis, MO), were reagent grade and were used without further purification. Ultrapure water produced from a Milli-Q System (Millipore Corporation) was used throughout the experiments. 2.2. Preparation of surface-modified PU foaming membranes Carboxylic acid groups were introduced via opening of the epoxy groups on the PU–epoxy membranes [21–24]. PU–epoxy membranes were immersed in a 0.5-M glycine solution containing 0.1 mol/L NaOH at 80 ◦ C for 24 h. The anticipated product from the ring-opening reaction of the epoxy group is shown in Fig. 1. The resulting membranes are referred to as PU–COOH membranes. After the reaction, the membranes were rinsed in ultrapure water for 3 h and then stored in ultrapure water at 4 ◦ C.
105
2.3. Direct ex vivo expansion of HSCs UCB was collected into a blood bag with an informed consent. The blood was subsequently injected into Syringe 2 of the blood permeation apparatus described in a previous study [21–24]. Two syringes (Syringe 1 and Syringe 2) were attached head to head. Water was injected into Syringe 1 using a perister pump (MP3, Tokyo Rikakiki Co.) at a speed of 1 mL/min, causing the head of Syringe 1 to push the blood-containing Syringe 2 back. One or six mL of UCB in Syringe 2 was filtered through PU–COOH membranes, attached inside the membrane holder, at the filtration rate of 1 mL/min at 25 ◦ C. The numbers of cells in the permeate and in the UCB (Np and Nf , respectively) were counted using flow cytometry (Coulter EPICSTM XL, Beckman-Coulter Co.) as described in the following section. The permeation ratio is defined as: Permeation ratio (%) =
Np × 100 Nf
(1)
After the blood filtration, the membranes were placed upside down in the membrane holder, and 6 mL of a washing solution (serum-free medium [StemSpanTM SFEM] or platelet-poor plasma [PPP]) was permeated through the membranes at a filtration speed of 1.0 mL/min at 25 ◦ C to remove the adhered cells from the membranes. This procedure enabled us to purify HSCs on the membranes by removing other blood cells. PPP was prepared by centrifugation of UCB at 2800 rpm for 15 min followed by filtration through microfiltration membranes with a 0.22-m pore size (MillexTM GS, Millipore Corporation). The recovery ratio is defined as: Recovery ratio (%) =
Nr × 100 Nf
(2)
where Nr is the number of cells in the permeate solution after the permeation of the washing solution. HSCs attached to the membranes after permeation of the washing solution as described above were injected directly into StemSpanTM SFEM serum-free culture medium containing the StemSpanTM CC110 cytokine cocktail (100 ng/mL rhFlt-3 ligand, 100 ng/mL rhStem Cell Factor and 100 ng/mL rhThrombopoietin) supplemented with 0.1 wt% LDL following the manufacturer’s instructions. The HSCs were then cultured for 10 days in a CO2 incubator at 37 ◦ C. Expansion fold of HSCs after ex vivo expansion was defined as: Expansion fold of HSCs (%) =
Nexpansion NUCB
× 100
(3)
where NUCB is the total number of HSCs in the UCB used for ex vivo expansion of HSCs, and Nexpansion is the total number of HSCs in the culture medium after ex vivo expansion of HSCs.
Fig. 1. Reaction scheme for synthesizing PU–COOH membranes for direct ex vivo expansion of HSCs from UCB.
106
A. Higuchi et al. / Journal of Membrane Science 351 (2010) 104–111
2.4. Conventional culture of HSCs from UCB Mononuclear cells were isolated by the Ficoll–Paque method [18]. The most pure CD34+ cells were separated from the mononuclear cells using an immunomagnetic separation kit and a MiniMacs column (Miltenyi Biotec) following the manufacturer’s instructions. After elution from the MiniMacs column, the cells were washed four times with phosphate-buffered saline (PBS) containing 2 mM ethylene diamine tetraacetic acid (EDTA). The purity of the isolated CD34+ cells and of the CD34+ subpopulations was determined by flow cytometry. HSCs (600 or 3000) from purified CD34+ cells from the immunomagnetic separation (MACS) were seeded into tissue culture dishes (TCPS, 24-well plate (142475), Nunc A/S, Roskilde; 2D culture) or onto PU and PU–COOH membranes in TCPS dishes (3D culture) containing the same serum-free culture medium that was used in the direct ex vivo expansion of HSCs. The cells were cultured for 10 days. The cells of each specific cell type were counted using flow cytometry as described in the following section. The colonyforming unit assays of HSCs after expansion in 2D and 3D culture were performed as described in the following section, and the results were averaged. 2.5. Bioreactor for HSC separation and culture A schematic representation of the bioreactor for the single-step process of separation, ex vivo expansion and harvest of HSCs from UCB is shown in Fig. 2. Umbilical cord blood (20 g) was permeated through a leukocyte removal filter composed of polyurethane foaming membranes at 4 mL/min, with valves 2, 3 and 5 closed and valves 1 and 4 open. The blood was then stored in the drain chamber shown in Fig. 2. PPP (washing solution, 20 g) was subsequently permeated through the leukocyte removal filter at 4 mL/min, with valves 1, 3 and 5 closed and valves 2 and 4 open, as seen in Fig. 2. To rinse the leukocyte removal filter, the PPP wash was followed by the permeation of 60 g of StemSpan SFEM culture medium at 4 mL/min, with valves 1, 2, 5, 6 and 7 closed and valves 3 and 4 open (Fig. 2). With valves 1, 2, 4 and 7 closed and valves 3, 5 and 6 open (Fig. 2), 30 g of serum-free StemSpan SFEM medium, supplemented with the StemSpan CC110 cytokine cocktail and 5 mg/mL of LDL, was injected into the reservoir of culture medium (Fig. 2) and was circulated through the leukocyte removal filter using a perister pump at 0.5 mL/min. In this case, the tubes attached to the leukocyte removal filter were connected to the perister pump, and the culture medium was circulated through the leukocyte removal filter. Fresh culture medium (20 g) was exchanged with the culture medium contained in the reservoir every 2 days, and the HSCs were cultured
for 10 days. On the final day of culture, following the removal of the culture medium, a washing solution of PBS containing 2 mM EDTA was circulated through the filter, with valves 1, 2, 4 and 7 closed and valves 3, 5 and 6 open. The HSCs could be harvested from the culture medium and from the washing solution in the reservoir. The number of HSCs and other blood cells in the culture medium was analyzed using flow cytometry as described in the following section, and the values were averaged. The colony-forming units assays of HSCs in the culture medium were performed as described in the following section, and the results were averaged. 2.6. Flow cytometry analysis of blood cells The number of red blood cells, platelets and leukocytes in the blood and culture medium was analyzed using the following cellsurface markers: glycophorin A for red blood cells (RBC), CD41 for platelets and CD45 for white blood cells (WBC). The samples were diluted 10-fold. Anti-glycophorin A antibody (20 L), antiCD41 antibody (20 L) and anti-CD45 antibody (20 L) were added to 100 L of the diluted sample. The sample was then incubated in the dark for 20 min, after which it was agitated using a vortex mixer (VX-100, Montreal Biotech Inc.) for 1 min. The sample was then diluted 100 times, and 100 L of Flow-count bead solution was subsequently added to 500 L of the diluted sample. Finally, the sample was analyzed using flow cytometry, and the number of red blood cells, platelets and leukocytes was counted [21–24]. The number of HSCs was analyzed by counting the number of CD34+ cells following the ISHAGE (International Society of Hematotherapy and Graft Engineering) guidelines [25] using flow cytometry after staining the cells with anti-CD34 antibody, antiCD45 antibody and 7-AAD [21–24]. 2.7. Colony-forming units assay The colony-forming unit (CFU) assay was used for the enumeration of progenitor cells after harvesting the cells from 2D culture and 3D culture. A methylcellulose medium of MethoCultTM GF H4434, containing stem cell factor, granulocyte–macrophage colony-stimulating factor (GM-CSF), erythropoietin and interleukin-3 (IL-3), was used to perform colony assays on ex vivo expanded CD34+ cells following the manufacturer’s instructions [26]. Harvested cells were suspended in methylcellulose medium and plated into 35-mm culture dishes. The dishes were incubated for 14 days at 37 ◦ C in a humidified atmosphere of 5% CO2 in air. After incubation, the colonies were counted using an inverted microscope. The types of colonies counted in this study were human colony-forming unit-erythroid
Fig. 2. Schematic representation of the bioreactor used for single-step separation, ex vivo expansion and harvest of HSCs from UCB.
A. Higuchi et al. / Journal of Membrane Science 351 (2010) 104–111
107
(CFU-E), burst-forming unit-erythroid (BFU-E), colony-forming unit-granulocyte/macrophage (CFU-GM) and pluripotent mixed colonies (colony-forming unit-granulocyte, erythroid, macrophage and megakaryocyte, CFU-GEMM). 2.8. Statistical analysis Results are expressed as the mean of three experiments and the standard deviation, except for the data in Fig. 9. Error bars represent the standard deviation. Statistical significance was determined by the Student’s t-test for two-tailed distributions, assuming equal variances for both samples (homoskedastic analysis) using Excel software (Microsoft Corporation). Probability values less than 0.05 were considered statistically significant. 3. Results and discussion 3.1. Direct ex vivo expansion of HSCs on PU–COOH membranes When UCB was permeated through 5-m pore size polyurethane membranes having nano-segments containing carboxylic acid groups (PU–COOH, Fig. 1), the permeation ratio of HSCs was found to be less than 1% (data not shown). The recovery ratios of HSCs and red blood cells filtered through PU–COOH membranes were found to be 7–20% and 16–24%, respectively, when the recovery solution was permeated through the membranes after permeation of UCB; the percentage of cells recovered depended on the components of the recovery solution (data not shown). As seen in our previous study [24], we could not extensively purify HSCs from UCB in the recovery solution even by using the PU–COOH membrane filtration method described here, although HSCs were purified from peripheral blood through the membranes [21–23]. The reason for this is that HSCs continued to adhere to the membrane surfaces even after the permeation of the washing solution. HSCs in UCB seem to be extremely adhesive compared to other blood cells. Therefore, we developed a direct method of ex vivo expansion of HSCs from UCB as follows. Fig. 3 shows the fold expansion of HSCs after direct ex vivo expansion of HSCs on PU–COOH membranes. The effect of the washing solution, i.e., platelet-poor plasma (PPP) and serum-free culture medium, was examined on direct ex vivo expansion of HSCs. When 1 or 6 mL of UCB was permeated through PU–COOH membranes (PPP-1N and PPP-6N), the increase in the fold expansion of HSCs was found to be almost the same for the two blood volumes after 10 days of culture, 2.7 and 2.5, respectively. PPP was found to be more effective as a washing solution (PPP-1N, PPP-1C and PPP6N) than serum-free medium (StemSpan SFEM medium, StemS-1N and StemS-1C). When the culture medium was changed at day 5 (PPP-1C and Stem-1C), the ex vivo expansion of HSCs (i.e., 6.6-fold for PPP-1C) was found to be higher with a t < 0.05 than when the
Fig. 3. Fold expansion of HSCs after direct ex vivo expansion of HSCs on PU–COOH membranes from 1 mL (StemS-1N, StemS-1C, PPP-1N and PPP-1C) or 6 mL of UCB (PPP-6N) using a washing solution of serum-free medium (StemS-1N and StemS1C) or PPP (PPP-1N, PPP-1C and PPP-6N), and fold expansion of HSCs cultured on 2D (TCPS-1 and TCPS-2) and 3D materials (PU-1, PU-2 and PU–COOH) using HSCs purified from UCB by the conventional Ficoll–Paque and MACS method. For Ficoll–Paque/MACS-purified cells, cell inoculation densities of 600 HSCs/dish (TCPS1, PU-1 and PU–COOH) or 3000 HSCs/dish (TCPS-2 and PU-2) were used. The culture medium was changed (StemS-1C, PPP-1C) or not changed (StemS-1N, PPP-1N, PPP6N, TCPS-1, TCPS-2, PU-1, PU-2 and PU–COOH) at day 5 of cell culture. Data are expressed as the means ± S.D. of three independent measurements.
culture medium (PPP-1N and StemS-1N) was not changed. Changing the culture medium provides better nutrition to HSCs on the membranes. We also performed conventional ex vivo expansion of HSCs purified by the Ficoll–Paque method followed by MACS for comparison to the method of direct ex vivo expansion of HSCs used in this study. HSCs were cultured on different 2D (TCPS dishes) and 3D materials (PU and PU–COOH) using HSCs purified by the Ficoll–Paque and MACS method. Fig. 4 shows micrographs of cells purified by the Ficoll–Paque and MACS method and cells after ex vivo expansion of HSCs on TCPS dishes for 10 days of culture. The number of cells was found to increase after 10 days of culture. The fold increase in the expansion of HSCs cultured on 2D and 3D materials is also shown in Fig. 3. Ex vivo expansion of HSCs in 3D culture (PU-1, PU-2 and PU–COOH) was found to be worse than that in 2D culture (TCPS-1 and TCPS-2) using HSCs purified by the Ficoll–Paque and MACS method. This is because pluripotent HSCs adhered tightly to the PU membranes, and HSCs in the PU membranes of the 3D culture could not be detached even after pipetting with culture medium and washing solution (t < 0.05; HSC expansion
Fig. 4. Micrographs of cells purified by the Ficoll–Paque and MACS method (a) and of cells after ex vivo expansion of HSCs on TCPS dishes [TCPS-1 (b) and TCPS-2 (c)] after 10 days of culture.
108
A. Higuchi et al. / Journal of Membrane Science 351 (2010) 104–111
fold on TCPS-1 vs. that on PU-1, and HSC expansion fold on TCPS-2 vs. that on PU-2). The direct ex vivo expansion of HSCs on PU–COOH membranes after permeation of 1 mL of UCB using a washing solution of PPP (PPP-1C) was found to have a higher fold expansion of HSCs cultured on 2D and 3D materials than that purified by the conventional Ficoll–Paque and MACS method (t < 0.05). It should be mentioned that the fold expansion of HSCs on tissue culture dishes (TCPS-1 and TCPS-2) was calculated from the fold expansion based on the number of HSCs in the UCB used and not from the number of HSCs after purification by the Ficoll–Paque and MACS procedures, which has been reported by most researchers [1,4,7,12]. We used purified HSCs for ex vivo expansion on 2D and 3D materials, of which the numbers were decreased to 15.1% of the HSCs in the original UCB. That means that the increases in the fold ex vivo expansion of HSCs on TCPS-1 and TCPS-2 were calculated to be 32.5 (=4.91/0.151) and 22.3 (=3.36/0.151), respectively, based on the number of purified HSCs; these fold expansion values are similar to those reported on TCPS (e.g., 50-fold expansion for cryopreserved CD34+ cells of UCB having 98% purity [1,4]) or other 2D culture materials by other investigators [1,4,7,12]. 3.2. Analysis of HSCs by colony-forming unit assays The analysis of HSCs by colony-forming unit (CFU) assays was performed after HSCs were expanded by the batch type of direct ex vivo expansion using HSCs purified from UCB. The number of HSCs obtained was compared to the number of the initial unexpanded HSCs and the number of HSCs present after purification from UCB by the conventional Ficoll–Paque method followed by MACS and subsequently cultured on several 2D and 3D materials. Figs. 5 and 6 show both macroscopic and microscopic pictures of the colonies on methylcellulose gels at 14 days. The colonies that originated from HSCs could clearly be seen on each dish. The number of each type of colony (i.e., the numbers of CFU-E, BFU-E, CFU-GM, CFU-GEMM and total colonies) in the CFU assay of HSCs at 14 days was evalu-
ated on each dish; the results are shown in Fig. 7. HSCs expanded by the direct ex vivo expansion method (PPP-1C) showed a much higher number of CFU-GEMM colonies (i.e., a multi-potential progenitor having primitive nature) compared with HSCs cultured on 3D materials (PU-1, PU-2 and PU–COOH) (t < 0.05). HSCs expanded by the direct ex vivo expansion method (PPP-1C) showed a similar number of CFU-GEMM colonies compared to the HSCs cultured on 2D materials (TCPS-1 and TCPS-2) (t > 0.05) that had been purified by the conventional Ficoll–Paque method followed by MACS procedures from UCB. The total number of HSC colonies after ex vivo expansion was directly related to the fold increase in HSC expansion and to the number of CFUs per 450-HSC inoculation. CFU expansion fold is defined as: CFU expansion fold = HSC expansion fold ×
Y X
(4)
where X is the number of CFUs per 450-HSC inoculation (450 HSCs/dish) for the CFU assay before ex vivo expansion (i.e., in fresh UCB), and Y is the number of CFUs per 450-HSC inoculation (450 HSCs/dish) for the CFU assay after ex vivo expansion. If the CFU expansion fold is greater than one, then the HSCs obtained after ex vivo expansion have a higher ability to generate CFUs than the initial HSCs in the UCB before ex vivo expansion, indicating that the ex vivo expansion of HSCs is meaningful. Fig. 8 shows the CFU expansion fold after 14 days, as measured by the CFU assay, using HSCs expanded by the direct ex vivo expansion method and HSCs expanded on several 2D and 3D materials, and compared to freshly isolated HSCs. HSCs expanded by direct ex vivo expansion and expanded on 2D materials (TCPS dishes) show a fold CFU expansion greater than one for all colonies. HSCs expanded by the direct ex vivo expansion method (PPP-1C) showed the highest CFU expansion fold of CFU-GEMM colonies in this study (t < 0.05). Furthermore, HSCs expanded by the direct ex vivo expansion method (PPP-1C) showed a similar CFU expan-
Fig. 5. Photographs of colonies on the methylcellulose gels using freshly isolated HSCs purified from UCB by the Ficoll–Paque method followed by MACS (a), HSCs after direct ex vivo expansion through PU–COOH membranes from 1 mL of UCB using a washing solution of PPP (PPP-1C, b) and HSCs expanded on TCPS dishes [TCPS-1 (c), and TCPS-2 (d)] and on PU [PU-1 (e)] and PU–COOH membranes (f), using a CFU seeding number of 450 HSCs/dish, at 14 days.
A. Higuchi et al. / Journal of Membrane Science 351 (2010) 104–111
109
Fig. 6. Micrographs of colonies on the methylcellulose gels using freshly isolated HSCs purified from UCB using the Ficoll–Paque method followed by MACS (a), HSCs after direct ex vivo expansion through PU–COOH membranes from 1 mL of UCB using a washing solution of PPP (PPP-1C, b and c) and HSCs expanded on TCPS dishes [TCPS-1 (d), and TCPS-2 (e)] and on PU [PU-1 (f) and PU-2 (g)] and PU–COOH membranes (h), using a CFU seeding number of 450 HSCs/dish, at 14 days.
Fig. 7. Colony-forming unit (CFU) counts using freshly isolated HSCs purified from UCB using the Ficoll–Paque method followed by MACS (UCB), HSCs after direct ex vivo expansion through PU–COOH membranes from UCB using a washing solution of PPP (PPP-1C) and HSCs expanded on TCPS dishes (TCPS-1 and TCPS-2) and on PU (PU-1 and PU-2) and PU–COOH membranes, using a CFU seeding number of 450 HSCs/dish, at 14 days.
Fig. 8. CFU expansion fold using HSCs after direct ex vivo expansion through PU–COOH membranes from UCB using a washing solution of PPP (PPP-1C) and HSCs expanded on TCPS dishes (TCPS-1, and TCPS-2) and on PU (PU-1 and PU-2) and PU–COOH membranes, using a CFU seeding number of 450 HSCs/dish, at 14 days. Data are expressed as the means ± S.D. of three independent measurements.
110
A. Higuchi et al. / Journal of Membrane Science 351 (2010) 104–111
Table 1 Composition of blood cells in UCB, culture medium after direct ex vivo expansion, culture medium after ex vivo expansion on TCPS (TCPS-1), culture medium after ex vivo expansion using bioreactor, purified CD34+ cells by Ficoll–Paque and MACS method, and on membranes. In solution or on membrane
HSCs (%)
RBC (%)
WBC (%)
Platelet (%)
In peripheral blood In UCB In purified CD34+ by Ficoll–Paque and MACS On membrane (PPP-1N)a On membrane (PPP-6N)b In medium after direct ex vivo expansion (PPP-1N)c In medium after direct ex vivo expansion (PPP-6N) In culture medium after direct ex vivo expansion (bioreactor) In culture medium after ex vivo expansion of HSCs on TCPS (TCPS-1)
1.0 × 10−4 5.8 × 10−3 19.0 1.3 × 10−2 5.3 × 10−2 0.4 0.2 23.6 5.9
92.9 53.0 0.7 0 0 58.5 55.2 52.0 7.9
0.1 1.0 22.5 0.1 8.7 7.8 1.8 35.0 63.9
7.0 46.0 76.8 99.9 91.3 33.7 43.0 13.0 28.2
a
One mL of UCB permeated through PU–COOH membranes followed by the permeation of 6 mL of PPP. Six mL of UCB permeated through PU–COOH membranes followed by the permeation of 6 mL of PPP. c One mL of UCB permeated through PU–COOH membranes followed by the permeation of 6 mL of PPP. The membranes were incubated in serum-free medium for 10 days without culture medium exchange. d Six mL of UCB permeated through PU–COOH membranes followed by the permeation of 6 mL of PPP. The membranes were incubated in serum-free medium for 10 days without culture medium exchange. b
sion fold for the total number of colonies to those expanded on 2D materials (TCPS-1 and TCPS-2) and a much higher increase than those expanded on 3D materials (PU-1, PU-2 and PU–COOH). These results indicate that HSCs expanded by the direct ex vivo expansion method retain more pluripotency and that they show better hematopoietic abilities than those expanded on 2D and 3D culture materials. 3.3. Bioreactor for direct ex vivo expansion of HSCs When HSCs expanded by the present direct ex vivo expansion method through the membranes were considered for use in clinical applications, the process of the purification, washing and culture of HSCs has to be performed in a closed sterile system. Therefore, a bioreactor system using a perfusion method [27–29] was developed for the ex vivo expansion of HSCs from UCB. The bioreactor allowed permeation of UCB through membranes for the adhesion of HSCs on the membranes, followed by permeation of a washing solution and finally injection of cell culture medium into the membranes, as shown in Fig. 2. Leukocyte removal filters made of polyurethane were used in this study. Fig. 9 shows the time dependence of the fold expansion of HSCs of three different experiments using the same experimental conditions except for different UCB in
the perfusion type of bioreactor for the direct ex vivo expansion of HSCs from UCB through PU membranes. The culture medium was changed every 2 days. HSCs could be harvested from the culture medium and the washing solution (PBS with 2 mM EDTA), which were both circulated through the membranes. Total fold ex vivo HSC expansion values of 14.6, 31.0 and 45.7 (compared to the initial number of HSCs in UCB) were obtained because the fold ex vivo expansion of HSCs is dependent upon the source of the UCB. HSCs could be harvested from the culture medium when the culture medium was exchanged every 2 days. If we assume that the CFU ability of HSCs expanded from UCB by the perfusion type is the same as that of HSCs expanded by direct ex vivo expansion, the CFU expansion fold values of CFU-GEMM and total colonies were calculated to be 8.7–27.2 and 8.54–26.7, respectively, and are much higher than those for colonies expanded on 2D materials (TCPS dishes). These results suggest that the direct ex vivo expansion of HSCs from UCB by the perfusion method is mimicking the bone marrow niche and is highly effective for the clinical application of the ex vivo expansion of HSCs. 3.4. Purity assays of HSCs in blood, in culture medium and on membranes The purity of HSCs in blood, in culture medium and on membranes was investigated and is summarized in Table 1. The purity of HSCs in the culture medium after batch-type direct ex vivo expansion was found to be much higher than the purity of HSCs in UCB, while the purity of HSCs in the culture medium after perfusion-type direct ex vivo expansion was higher than that of HSCs expanded in the conventional 2D culture after purification by the Ficoll–Paque and MACS procedures. The purity of HSCs on membranes was one order of magnitude higher than that in UCB, with almost no red blood cells (approximately 0%) and few white blood cells (less than 9%) present. These results suggest that the high purity of HSCs and the low content of red blood cells on the membranes enable HSCs to expand on the membranes in the direct ex vivo expansion method used in this study. 4. Conclusion
Fig. 9. Time dependence of fold expansion of HSCs on the perfusion type of bioreactor in direct ex vivo expansion of HSCs. The culture medium was changed every 2 days. HSCs can be harvested from the culture medium and the washing solution (10th day-W, PBS with 2 mM EDTA), which was circulated through the membranes.
The direct ex vivo expansion of HSCs from UCB filtered through membranes is a simple procedure and provides a high fold ex vivo expansion of HSCs with a short working time (less than 30 min) compared with the Ficoll–Paque and MACS method. The direct ex vivo expansion by the perfusion method is mimicking the bone marrow niche and will be effective for the clinical application of the ex vivo expansion of HSCs.
A. Higuchi et al. / Journal of Membrane Science 351 (2010) 104–111
Acknowledgments This research was partially supported by the National Science Council of Taiwan under Grant No. NSC97-2221-E-008-011-MY3, NSC97-2120-M-008-002 and NSC98-2120-M-008-002. This work was also supported by the VGHUST Joint Research Program, Tsou’s Foundation (VGHUST97-P3-08 and VGHUST98-P3-11) and a Cathay General Hospital Project (98CGH-NCU-B1). Grants-in-Aid for Scientific Research (No. 21500436) from the Ministry of Education, Culture, Sports, Science and Technology of Japan are also acknowledged. References [1] K.-N. Chua, C. Chaib, P.-C. Lee, S. Ramakrishna, K.W. Leong, H.-Q. Mao, Functional nanofiber scaffolds with different spacers modulate adhesion and expansion of cryopreserved umbilical cord blood hematopoietic stem/progenitor cells, Exp. Hematol. 35 (2007) 771–781. [2] A. Czechowicz, D. Kraft, I.L. Weissman, D. Bhattacharya, Efficient transplantation via antibody-based clearance of hematopoietic stem cell niches, Science 318 (2007) 1296–1299. [3] Y. Cohen, A. Nagler, Umbilical cord blood transplantation—how, when and for whom? Blood Rev. 18 (2004) 167–179. [4] K.N. Chua, C. Chai, P.C. Lee, Y.-N. Tang, S. Ramakrishna, K.W. Leong, H.-Q. Mao, Surface-aminated electrospun nanofibers enhance adhesion and expansion of human umbilical cord blood hematopoietic stem/progenitor cells, Biomaterials 27 (2006) 6043–6051. [5] N. Fujimoto, S. Fujita, T. Tsuji, J. Toguchida, K. Ida, H. Suginami, H. Iwata, Microencapsulated feeder cells as a source of soluble factors for expansion of CD34(+) hematopoietic stem cells, Biomaterials 28 (2007) 4795–4805. [6] J.A. LaIuppa, T.A. McAdams, E.T. Papoutsakis, W.M. Miller, Culture materials affect ex vivo expansion of hematopoietic progenitor cells, J. Biomed. Mater. Res. 36 (1997) 347–359. [7] X.S. Jiang, C. Chai, Y. Zhang, R.-X. Zhuo, H.-Q. Mao, K.W. Leong, Surfaceimmobilization of adhesion peptides on substrate for ex vivo expansion of cryopreserved umbilical cord blood CD34(+) cells, Biomaterials 27 (2006) 2723–2732. [8] K. Franke, T. Pompe, M. Bornhauser, C. Werner, Engineered matrix coatings to modulate the adhesion of CD133+ human hematopoietic progenitor cells, Biomaterials 28 (2007) 836–843. [9] S. Eridani, U. Mazza, P. Massaro, M.L. La Targia, A.T. Maiolo, A. Mosca, Cytokine effect on ex vivo expansion of haemopoietic stem cells from different human sources, Biotherapy 11 (1998) 291–296. [10] D.A. Stewart, D. Guo, J. Luider, I. Auer, J. Klassen, E. Ching, D. Morris, A. Chaudhry, C. Brown, J.A. Russell, Factors predicting engraftment of autologous blood stem cells: CD34+ subsets inferior to the total CD34+ cell dose, Bone Marrow Transplant. 23 (1999) 1237–1243. [11] S.H. Lee, M.H. Lee, J.H. Lee, Y.H. Min, K.H. Lee, J.W. Cheong, J. Lee, K.W. Park, J.H. Kang, K. Kim, W.S. Kim, C.W. Jung, S.-J. Choi, J.-H. Lee, K. Park, Infused CD34+ cell dose predicts long-term survival in acute myelogenous leukemia patients who received allogeneic bone marrow transplantation from matched sibling donors in first complete remission, Biol. Blood Marrow Transplant. 11 (2005) 122–128.
111
[12] Q. Feng, C. Chai, X.S. Jiang, K.W. Leong, H.-Q. Mao, Expansion of engrafting human hematopoietic stem/progenitor cells in three-dimensional scaffolds with surface-immobilized fibronectin, J. Biomed. Mater. Res. 78A (2006) 781–791. [13] D. Bonnet, Biology of human bone marrow stem cells, Clin. Exp. Med. 3 (2003) 140–149. [14] M. Takagi, Cell processing engineering for ex-vivo expansion of hematopoietic cells, J. Biosci. Bioeng. 99 (2005) 189–196. [15] E. Fuchs, T. Tumbar, G. Guasch, Socializing with the neighbors: stem cells and their niche, Cell 116 (2004) 769–778. [16] I.R. Lemischka, K.A. Moore, Stem cells: interactive niches, Nature 425 (2003) 778–779. [17] A. Colmone, M. Amorim, A.L. Pontier, S. Wang, E. Jablonski, D.A. Sipkins, Leukemic cells create bone marrow niches that disrupt the behavior of normal hematopoietic progenitor cells, Science 322 (2008) 1861–1865. [18] C.H. Cho, J.F. Eliason, H.W. Matthew, Application of porous glycosaminoglycanbased scaffolds for expansion of human cord blood stem cells in perfusion culture, J. Biomed. Mater. Res. A 86A (2008) 98–107. [19] C. Lin, S.H. Chen, L.J. Yang, Y. Tan, X. Bai, Y. Li, Evaluation of TCR V beta subfamily T cell expansion in NOD/SCID mice transplanted with human cord blood hematopoietic stem cells, Hematology 12 (2007) 325–330. [20] D. Freund, J. Oswald, S. Feldmann, G. Ehninger, D. Corbeil, M. Bornhäuser, Comparative analysis of proliferative potential and clonogenicity of MACSimmunomagnetic isolated CD34(+) and CD133(+) blood stem cells derived from a single donor, Cell Prolif. 39 (2006) 325–332. [21] A. Higuchi, S. Yamamiya, B.O. Yoon, N. Sakurai, M. Hara, Peripheral blood cell separation through surface-modified polyurethane membranes, J. Biomed. Mater. Res. 68A (2004) 34–42. [22] A. Higuchi, A. Iizuka, Y. Gomei, T. Miyazaki, M. Sakurai, Y. Matsuoka, S.H. Natori, Separation of CD34+ cells from human peripheral blood through polyurethane foaming membranes, J. Biomed. Mater. Res. 78A (2006) 491–499. [23] A. Higuchi, M. Sekiya, Y. Gomei, M. Sakurai, W.-Y. Chen, S. Egashira, Y. Matsuoka, Separation of hematopoietic stem cells from human peripheral blood through modified polyurethane foaming membranes, J. Biomed. Mater. Res. 85A (2008) 853–861. [24] A. Higuchi, S.-T. Yang, P.T. Li, R.-C. Ruaan, W.-Y. Chen, Y. Chang, Y. Chang, E.M. Tsai, Y.H. Chen, H.-C. Wang, S.-T. Hsu, Q.-D. Ling, Permeation of blood cells from umbilical cord blood through surface-modified polyurethane foaming membranes, J. Membr. Sci. 339 (2009) 184–188. [25] M. Keeney, I. Chin-Yee, K. Weir, J. Popma, R. Nayar, D.R. Sutherland, Single platform flow cytometric absolute CD34+ cell counts based on the ISHAGE guidelines, Cytometry (Comm Clin Cyto) 34 (1998) 61–70. [26] R&D System, Procedure for the Human Colony Forming Cell (CFC) Assay Using Methylcellulose-based Media, http://www.rndsystems.com/stem cell protocol detail objectname CFC.aspx. [27] M.P. Flaherty, A. Abdel-Latif, Q. Li, G. Hunt, S. Ranjan, Q. Ou, X.-L. Tang, R.K. Johnson, R. Bolli, B. Dawn, Noncanonical Wnt11 signaling is sufficient to induce cardiomyogenic differentiation in unfractionated bone marrow mononuclear cells, Circulation 117 (2008) 2241–2252. [28] B.O. Palsson, S.-H. Paek, R.M. Schwartz, M. Palsson, G.-M. Lee, S. Silver, S.G. Emerson, Expansion of human bone marrow progenitor cells in a high cell density continuous perfusion system, Biotechnology (NY) 11 (1993) 368– 372. [29] M.R. Koller, J.G. Bender, W.M. Miller, E.T. Papoutsakis, Expansion of primitive human hematopoietic progenitors in a perfusion bioreactor system with IL-3, IL-6, and stem cell factor, Biotechnology (NY) 11 (2000) 358–363.
