Isolation of Mesenchymal Stem Cells from G-CSF-Mobilized Human Peripheral Blood Using ﬁbrin Microbeads

ORIGINAL ARTICLE Isolation of mesenchymal stem cells from G-CSF-mobilized human peripheral blood using ﬁbrin microbeads

I Kassis1, L Zangi1, R Rivkin1, L Levdansky1, S Samuel2, G Marx3 and R Gorodetsky1,3

1Radiobiology and Biotechnology Laboratory, Sharett Institute of Oncology, Hadassah-Hebrew University Medical Center, Jerusalem, Israel; 2Department of Bone Marrow Transplantation, Hadassah-Hebrew University Medical Center, Jerusalem, Israel and 3HAPTO Biotech Ltd, Jerusalem, Israel

Adult mesenchymal stem cells (MSC) that are able to Introduction differentiate into various mesenchymal cell types are typically isolated from bone marrow, but their significant The presence of non-hematopoietic stem cells in the bone presence in human peripheral blood (PB) is controversial. marrow (BM) was first suggested by Fridenshtein.1,2 These Fibrin microbeads (FMB) that bind matrix-dependent cells were shown to be able to differentiate into a number of cells were used to isolate MSC from the mononuclear mesenchymal phenotypes, including osteocytic and chon- fraction of mobilized PB of adult healthy human donors drocytic cells. Subsequently, these cells were isolated and treated with a granulocyte colony-stimulating factor. defined as mesenchymal stem cells (MSC). They consisted Isolation by plastic adherence resulted in a negligible of fibroblast-like adherent cells with high proliferative number of MSC in all samples tested, whereas FMB- capacity.3–5 Other compartments in different tissues were based isolation yielded spindle-shaped cell samples that also proposed as possible sources for MSC, including could further expand on plastic or on FMB in eight out of adipose tissue,6,7 liver,8 first-trimester blood8 and fetal the 11 samples. The yield of these cells at days 17–18 after pancreas.9 The ability to separate a significant number of the harvest was B0.5% of the initial cell number. The adult MSC from other sources, such as cord blood (CB) isolated cells were grown on plastic and characterized by and peripheral blood (PB), is controversial.10–14 The FACS analysis and immunohistochemistry for specific isolation of a high number of potent MSC sets the basis markers. Following culturing and first passage, the FMB- for new methods for tissue regeneration and cell therapy. isolated cells stained positive for mesenchymal stromal Nevertheless, the procedure of BM extraction is traumatic cell markers CD90 and CD105, expressed vimentin and and the amount of material extracted is limited. Therefore, fibronectin and were negative for hematopoietic markers exploring new sources and isolation techniques for obtain- CD45 and CD34. These cells could differentiate into ing such cells is of great interest. osteoblasts, adipocytes and chondrocytes. This study Stem cell mobilization is a procedure that induces rapid indicates that FMB may have special advantage in proliferation of hematopoietic stem cells in the BM and isolating MSC from sources such as mobilized PB, where release of high number of immature stem cells into the the number of such cells is scarce. bloodstream.15 The presence of a small number of MSC Bone Marrow Transplantation (2006) 37, 967–976. was detected in mobilized PB of healthy patients as well as doi:10.1038/sj.bmt.1705358 those with malignancies.13,16–23 Most of the studies Keywords: mesenchymal stem cells (MSC); granulocyte suggested that some MSC may also be present in non- colony-stimulating factor (G-CSF); mobilization; fibrin mobilized PB in healthy donors or in patients with microbeads (FMB); differentiation; human peripheral malignant diseases, but they are too few to be detected blood and cultured for a long term.13,16,22 Fibrinogen, the major component of the blood coagula- tion process, also seems to play a crucial role in tissue repair by providing an interim matrix that supports cell adhesion, proliferation and migration involved in wound repair.23,24 Based on these observations, three-dimensional cultures based on fibrin matrices were proposed for cell culture for cellular tissue repair therapy.23–33 Fibrin microbeads (FMB) were proposed as a particulate form of fibrin more suited to cell culture conditions. Correspondence: Dr R Gorodetsky, Radiobiology and Biotechnology Essentially, the FMB were formed from fibrin in a heated Laboratory, Sharett Institute of Oncology, Hadassah-Hebrew University Medical Center, POB 12000, Jerusalem 91120, Israel. oil emulsion system that permitted crosslinking by endo- E-mail: rafi@hadassah.org.il genous factor XIII. The heat was kept relatively low (below Received 28 November 2005; revised and accepted 27 February 2006 801C) to retain the cell attachment epitopes of FMB for MSC isolation from mobilized blood I Kassis et al 968 native fibrinogen.25 Fibrin microbeads of 50–250 mm PBSC software, Gambro BCT, Lakewood, CO, USA). diameter were formed.25,27,29 Based on their high binding Basically, anticoagulant heparin solution was added to the efficiency of matrix-dependent cells, the FMB were used to blood to keep it from clotting during the procedure and the selectively isolate and expand MSC derived from mouse blood/anticoagulant mix cycled through a centrifuge to and rat BM. The isolated cells could be driven to separate the mononuclear and stem cells from the other differentiate on the slowly biodegradable FMB27,28 and components and plasma of the PB. The separated cell delivered with the FMB that also serve as vehicles for system pumped the separated cells into a collection bag, carrying the cell into the target tissues.27 These qualities of whereas the other blood components and plasma were the FMB make them the candidates of choice for various returned to the donor. The collection of cells was carried tissue engineering applications. In a preliminary study, we out at the interface of 1–2% hematopoietic cell transplant- showed that FMB were able to separate MSC from a ation. The main final output was composed of 70–90% population of rat BM at a higher yield than the conven- mononuclear and B10–15% polymorphonuclear cells, tional isolation method based on plastic adherence.26,27 which were used for transfusion to donors. Furthermore, we showed that mesenchymal cells isolated The samples for our experiments were obtained from by FMB from rat BM could form bone structures both minute residual waste materials remaining after such PB- in vitro and in vivo.27 These qualities make the FMB a based BMT procedure. The use of this source of material choice matrix for various tissue engineering applications. was approved by Hadassah Helsinki Ethics Committee as a In the present study, we evaluated the ability of FMB to use of waste samples from human sources for research separate human MSC from different sources other than purposes, with no tracking of the donor’s identity and with BM, with special emphasis on granulocyte colony-stimulat- no DNA examinations or manipulations. ing factor (G-CSF)-mobilized PB of healthy individuals. We showed that FMB were able to isolate significant Cell culture media numbers of MSC from the mobilized PB progenitor cells The main media used were either ‘poor’ or ‘rich media’ (PM (PBPC) of adult healthy individuals, with low contamina- and RM, respectively). Poor media were based on tion by other cell types. The cells that were isolated and Dulbecco’s modified Eagle’s medium with low glucose expanded on FMB proved their pluripotency by their (1.0 g/ml) þ 10% fetal calf serum (FCS) þ 1% vitamins þ ability to differentiate into various phenotypes from the 1% non-essential amino acids þ 1% glutamine þ 1% Pen/ mesodermal lineage. Strep (all from Biological Industries, Bet-Haemek, Israel). Rich media consisted of Dulbecco’s modified Eagle’s medium with high glucose (4.5 g/ml) þ 20% FCS þ 1% Materials and methods vitamins þ 1% non-essential amino acids þ 1% glutamine þ 1% Pen/Strep (all from Biological Industries, Bet-Haemek, Fibrin microbeads preparation and use Israel). Fibrin microbeads were produced from concentrated fibrinogen by a stirred heated oil emulsion technique as previously described.26–29 The diameter of FMB used for Preparation of peripheral blood progenitor cells for this study was in the range of 105–180 mm. After produc- isolation of mesenchymal stem cells tion, the FMB were stored dried and sterilized by The PBPC samples were washed ( Â 3) with PBS by 1 immersion in sterile 70% ethanol for at least 8 h before centrifugation for 6min at 1100 r.p.m. (24 C). Cell debris use, or by g-irradiation (15 kGy). The dried FMB have was removed and the samples of PBPC were re-suspended prolonged shelf life at room temperature (RT) without with fresh medium. The cells were then subjected to the two losing their activity. Before their use for cell separation, the MSC isolation methods, by plastic adherence or by FMB, FMB were washed and hydrated three times with sterile as described below. Dulbecco’s phosphate-buffered saline (PBS Biological Industries, Beit-Haemek, Israel). The hydrated FMB Mesenchymal stem cells isolation by adherence to plastic increased in size by about 30% but were resistant to Peripheral blood progenitor cells were seeded on plastic degradation in ethanol, organic solvents and aqueous plates in ‘rich’ or ‘poor’ medium supplemented with solutions of neutral pH for prolonged periods of at least 20 U/ml heparin (Sigma-Aldrich, Rehovot, Israel). Cultures a few weeks. were maintained at 371C in a humidified atmosphere

containing 7% CO2 for 48 h, then the plates were washed with PBS to remove non-adhered cells and the medium was Peripheral blood progenitor cell samples derived from exchanged. The cultures were maintained for an additional mobilized peripheral blood 18–19 days with twice weekly medium exchange. The blood mobilization process was performed to prime the BM of normal donors. It consisted of two daily injections for 5 days 10 mg/kg G-CSF – Neupogens (Amgen Inc., Fibrin microbeads-based mesenchymal stem cells Thousand Oaks, CA, USA). Following the G-CSF treat- isolation method ment, the mononuclear polymorphonuclear cells were For the isolation of MSC by FMB, a total number of separated and collected from the fraction rich in mono- B100 Â 106 PBPC were mixed with B100 ml packed nuclear cells of the PB of the donor, using a cell separator volume of hydrated FMB in 50 ml tubes with perforated (COBE Spectra Apheresis System operated with version 6.1 cap to allow gas exchange. The perforated cap was

Bone Marrow Transplantation FMB for MSC isolation from mobilized blood I Kassis et al 969 wrapped loosely with a sheet of aluminum foil to allow gas 5 ml FACS tubes. From each sample, 10 000 cells were exchange, as previously described.25 Poor or rich medium analyzed with FACScan (BD Biosciences, Franklin Lakes, was supplemented with 20 U/ml heparin. The tubes were NJ, USA). placed in an angle of B201 in a slowly rotating device 1 placed in a cell culture incubator at 37 C with 7% CO2 for Immunostaining 27 48 h. For collection of cells loaded on the FMB, the Adhered cells were trypsinized for 7–10 min with 0.25% medium was removed after 48 h of incubation with the non- trypsin/EDTA (Biological Industries, Bet-Haemek, Israel), attached cells and the FMB were washed with PBS. The washed and resuspended in 24-well plates (B5 Â 104 cells/ washed cells loaded on FMB were mixed with 2–3 mg/ml well). Cells were incubated at 371C, 7% CO2 for 24 h. diluted fibrinogen (Vitex, New York, NY, USA, or Omrix, Medium was discarded and cells were washed gently with Tel Aviv, Israel, or equivalent source). The FMB were 0.05% Tween 20 diluted in PBS ( Â 3). Cells were fixed with dispersed on plastic dishes and extra fibrinogen was fresh paraformaldehyde (PFA, 4%) for 20 min at RT. For pumped out. Thrombin (10 U/ml; Vitex, New York, NY, staining of intracellular components, cells were permeabi- USA, or Omrix, Tel Aviv, Israel) was then sprayed on the lized with 0.5% Triton100 Â (Sigma-Aldrich, Rehovot, plate to form a thin fibrin film that immobilized the FMB Israel) added for 7 min. For blocking unspecific binding, loaded with cells to the plastic surface by polymerizing the cells were rinsed with 5% BSA (Biological Industries, Bet- minute residual fibrinogen in the plate. Medium (RM or Haemek, Israel) in PBS for 60 min at RT on a slowly plane- PM) was then carefully added and the FMB attached to rotated plate. Then, cells were washed by 0.05% Tween 20 the plastic dishes were incubated for an additional 18–19 diluted in PBS ( Â 3) and were incubated with the following days. Cultures were maintained at 371C in a humidified primary antibodies: mouse anti-human CD45 (1:10), mouse atmosphere containing 7% CO2 with medium exchange anti-human CD34 (1:10), mouse anti-human CD90 (1:10), twice weekly. mouse anti-human CD105 (1:10), mouse anti-human fibronectin (1:500) and rabbit anti-human vimentin (1:50) diluted Cell nuclei staining to required concentrations with FACS buffer containing 1% To indicate the density of cells loaded on FMB, their nuclei BSA. Cells were washed gently ( Â 3) in PBS with 0.05% were stained with propidium iodide (PI) (Sigma-Aldrich, Tween 20. For detection, cells were incubated with the Rehovot, Israel). A few microliters of suspended FMB were following secondary antibodies: goat anti-mouse-Cy3-con- fixed with 500 ml ethanol (70%) for 10 min at RT, then 5 ml jugated (1:100), goat anti-rabbit-Cy2-conjugated (1:100) of PI (5 mg/ml) added for 7 min in darkness. The stained both from Jackson Immunoresearch Laboratories, West FMB þ cells were then mounted on a microscope slide with Grove, PA, USA, diluted in 1% BSA buffer, on a slowly mounting solution. The red-stained nuclei of cells attached plane-rotated plate for 45 min in darkness at RT. Cells were to FMB were visualized by fluorescence microscopy. mounted with mounting solution and viewed by fluorescence and light by Nomarsky’s optics microscopy. Flow cytometry for cell antigens Cells cultured on plastics were trypsinized and fixed Assays for mesenchymal stem cells differentiation À5 with fresh paraformaldehyde (4%) for 20 min at RT. For osteogenic differentiation, PM was used with 10 M For staining of intracellular components, the cells were dexamethasone, 50 mg/ml ascorbic acid and 10 mM b- permeabilized by incubation with 0.5% Triton100 Â glycerophosphate (Sigma, Israel). Confluent MSC were (Sigma-Aldrich, Rehovot, Israel) for 7 min in RT. For cultured with differentiation factors for 21 days with two blocking non-specific binding, cells were rinsed with 5% medium exchanges per week. Adipogenic differentiation bovine serum albumin (BSA) in PBS for 60 min at RT on a was induced using MesenCultt basal medium (Stem Cell slowly plane-rotating plate. The cell suspension was then Technologies, Vancouver, BC, Canada) supplemented incubated with the following primary antibodies to the with adipose stimulatory supplements: 1-methyl-3-isobu- À9 different markers : mouse anti-human CD45 (1:10), mouse tylxanthine, 10 M dexamethasone, 5 mM insulin and 5 mM anti-human CD34 (1:10), mouse anti-human CD90 (1:10), indomethacin (Stem Cell Technologies, Vancouver, BC, mouse anti-human CD105 (1:10), all from Serotec, Oxford, Canada). Confluent MSC were cultured with differentia- UK, mouse anti-human fibronectin (1:500) and rabbit anti- tion factors for 21 days with three weekly medium human vimentin (1:50) both from Sigma-Aldrich, Rehovot, exchange. Chondrogenic differentiation was induced by Israel. The antibodies were diluted to required concentra- culturing the cells with PM containing 10 ng/ml TGF-b3 tions with FACS buffer containing 1% BSA þ 0.01% (Sigma-Aldrich, Rehovot, Israel). Confluent MSC were sodium azide (Sigma-Aldrich, Rehovot, Israel) in PBS. cultured with differentiation factors for 14 days with For their assay, the cells were incubated with the following medium exchange three times a week. secondary antibodies: donkey anti-mouse-FITC-conjugated (1:100) or donkey anti-rabbit-FITC-conjugated Cytochemical analysis of the differentiated cells (1:100) both from Jackson Labs, Maine, USA, diluted in Alizarin red staining was used to stain mineralized matrices FACS buffer on a slowly plane-rotated plate for 45 min at for the detection of bone-like structures formed by RT in darkness. Rinsing–washing of the cells between monolayer cell culture. Differentiated cells were fixed with different steps was carried out by short microfuge a solution of 4% formaldehyde–methanol–water (in the centrifugation. Cells were suspended in 400 ml FACS buffer ratio of 1:1:1.5), then stained with saturated solution of and then filtered through a mesh sieve of 50–90 mm into alizarin red (pH 4.0) (Sigma-Aldrich, Rehovot, Israel) for

Bone Marrow Transplantation FMB for MSC isolation from mobilized blood I Kassis et al 970 15 min at RT, washed with water and air dried. The surface The cells on FMB were downloaded to plastic dishes areas covered with red stain represented mineralized from FMB from the 2nd day with adhesion of the beads to nodules. Nitro-blue tetrazolium/indolylphosphate (NBT/ the plastics by a dilute fibrin clot. The cells expanded on the BCIP) staining was also used to detect expression of plastics and covered its surface with fibroblast-like cells alkaline phosphatase, which indicated osteoblast activity. within the next 18–19 days. With time, colonies of cells The NBT/BCIP staining was carried out with substrates of from the FMB expanded to form eventually a monolayer nitro-blue tetrazolium (NBT, non-colored) mixed with on the plastic surface as shown in Figure 1b–d. Then, the indolylphosphate (BCIP) yielding a blue color owing to cells were harvested and counted to evaluate their yield. the enzyme activity. Before staining, the cells were washed For the further expansion of these cells, they were re- with PBS, 0.5 ml of NBT/BCIP (ICN Biomedicals, Irvine, cultured in RM with regular passages when 80–90% CA, USA) was added and they were then incubated at 371C confluence was reached (Figure 1e).

in a humidified atmosphere containing 6% CO2 for 30 min. Cells with fibroblast-like morphology that could be The cells were washed with PBS and fixed with 4% PFA. further expanded were obtained by the FMB isolation The surface areas that stained blue represented alkaline method in eight of 11 PBPC samples examined. The ratio phosphatase activity. For the detection of adipogenic of the FMB-isolated cells after 18–19 days in culture differentiation, accumulated intracellular lipid droplets relative to the total number of PBPC in the initial sample were detected by oil-red-O staining. Cells were washed from which the MSC were isolated ranged between ( Â 3) with PBS. A 0.25 ml portion of oil-red-O solution B0.3570.077% in PM and B0.5070.16% in RM. On (10 mg/ml) (Sigma-Aldrich, Rehovot, Israel) was added to the other hand, the number of surviving and proliferating the cells for 20 min at RT. Cells were washed with PBS, and cells by the conventional plastic adherence method in fixed with 4% PFA. Under the light microscope, round both culturing media was too low to allow a significant intracellular lipid drops were stained orange. For the calculation of a yield and in none, a culture of ‘MSC-like’ detection of chondrogenic differentiation, alcian blue cells could be established. staining was used. It stained mucopolysaccharides or In the plastic adhesion isolation technique, the glucosaminoglycans as markers of secreted chondrocytes PBPC that remained on the plate after 2 days of incuba- matrix. The cells were then washed with PBS ( Â 3). For tion did not reveal the fibroblast-like shape as seen fixation, fresh 4% PFA was added for 20 min at RT. The with the FMB isolation method. Rather, most cells were cells were rinsed with alcian blue solution, pH 2.5 (Sigma, larger and rounder and resembled leukocytes (Figure 1f). Israel) for 30 min at RT. They were then washed in running Following 18 days of incubation, most of these cells died in tap water for 2 min and in distilled water for another 2 min. culture, yielding a negligible number of cells with mesen- Extracellular matrix components secreted by chondrocytes chymal phenotypes that could not be further expanded in were then stained blue. culture.

Photomicrographs Immunophenotypic characterization of peripheral blood The samples were mounted on glass with mounting progenitor cell-derived mesenchymal stem cells solution (Sigma, Israel). Pictures from microscopy were FACS analysis (Figure 2a and b) and immunohistochem- taken with a Nikon Coolpix camera with a Nikon Eclipse istry (Figure 2c) were performed in parallel to identify the TE200 microscope with Nomarsky optics and fluorescence types of cells isolated by different methods. The cell set-up. The pictures were taken with objective Â 10, Â 20, markers used were as follows: CD45, a common leukocyte Â 40 and a zoom set-up to yield a full field width of 1, 0.5 marker; CD34, hematopoietic stem cell marker; CD105 and and 0.25 mm, respectively. Fluorescence pictures were CD90, vimentin and fibronectin as putative markers for taken by manual setting (typically at an exposure time of MSC. Initially, before the isolation procedure, the vast 0.1–1 s, aperture of 6.4). In some cases, low-light exposure population of PBPC was found to stain positive for CD45 was superimposed on the fluorescence image to show the and negative for the mesenchymal antigens CD90 and outlines of the structures. CD105 (Figure 2a). However, after the isolation procedure with FMB, harvesting (passage 0) and after 2–3 passages, the residual cells were typically found to be negative Results for hematopoietic CD45 and CD34, indicating the non- hematopoietic origin of these cells, whereas the positive Isolation of mesenchymal stem cells using fibrin microbeads staining for CD90, CD105, vimentin and fibronectin The FMB-based method for isolating MSC was based (Figure 2b) indicated their similarity to MSC derived from on adding hydrated FMB to the suspension of PBPC BM. It is noteworthy that the degree of hematopoietic cell for 2 days in slow rotation in culture conditions in either contamination (detected by CD45 þ antigen) of the isolated PM or RM. Then, the FMB with attached cells were cells by FMB adhesion after their harvest at 18–19 days allowed to settle and the non-adhered cells in the (passage 0) may vary. In some samples, initially they suspension were evacuated. The FMB-attached cells reached up to B20% of the total cells (not shown). were downloaded and further grown on plastic dishes in However, this residual fraction of CD45 þ cells was no PM or RM for an additional 18–19 days. The density of longer present after 2–3 passages, as shown by FACS cells attached to the FMB after 2 days could be visualized analysis (Figure 2b) and by immunohistochemistry by fluorescence microscopy with PI nuclei staining (Figure 2c), whereas the residual adhered cells were (Figure 1a). recorded as being both CD90 and CD105 positive. In a

Bone Marrow Transplantation FMB for MSC isolation from mobilized blood I Kassis et al 971

a b

100m 200m

c d

200m 200m

e f

200m 200m

Figure 1 Isolation of human mesenchymal stem cells (MSC) on fibrin microbeads (FMB) and their downloading on plastic dishes. (a) Fluorescence micrograph of cells loaded on FMB after 48 h of incubation with peripheral blood progenitor cells from mobilized blood in suspension. The FMB have faint self-fluorescence. The nuclei of cells bound to FMB are visualized by intense propidium iodide staining of the cell nuclei. (b–d) Downloading cells from FMB (with the aid of low concentrations of fibrin) after 7, 14 and 18 days in culture, respectively. (e) Near-confluent MSC population isolated with FMB following two passages on plastic dishes. (f) Typical phenotype of the cell population isolated by the conventional plastic adhesion method after 18 days of culture. These cells were easily detached and their survival was reduced so that after several passages they were eliminated. few instances, cells isolated by the conventional plastic In vitro differentiation of osteocytes, adipocytes and adherence method could be collected to allow FACS chondrocytes from peripheral blood progenitor cell-derived analysis before they died or detached. Most of these cells mesenchymal stem cells were found to be positive for CD45 and negative for CD90 To investigate the differentiation potential of the cells and CD105 (not shown). isolated by FMB to mesenchymal phenotypes, cells from

Bone Marrow Transplantation FMB for MSC isolation from mobilized blood I Kassis et al 972 a CD45 CD105 CD90 100 80 100

50 40 50 Counts x 1000 0 0 0 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 FL1

b CD45 CD34 CD90

0 100 101 102 103 104 100 101 102 103 104100 101 102 103 104

CD105 Fibronectin Vimentin Counts x 100 80

0 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 FL1 c CD45 CD34 Vim

FN CD105 CD90

100m

Figure 2 Immunophenotype of mesenchymal stem cells isolated by fibrin microbeads (FMB). (a) FACS analysis of the immunophenotype profile for hematopoietic marker CD45 and for mesenchymal antigens CD90 and CD105 of the initial peripheral blood progenitor cell sample before isolation. Cells isolated by FMB technique were harvested 18–19 days following harvesting and cultured for further passages. Shaded histograms represent the fluorescence from control cells stained with only second antibody; non-filled histograms represent positive staining of cells with the indicated antibody. (b) Cells from passages 2–3 were harvested and labeled with antibodies against CD45, CD34, CD90, CD105, fibronectin (FN) and vimentin (Vim) and analyzed by FACS. Shaded histograms represent the fluorescence from control cells stained with only second antibody; non-filled histograms represent positive staining of cells with the indicated antibody. (c) Cells from passage 5 were harvested and immunostained with antibodies to CD45 (Cy3 labeled), CD34 (Cy3 labeled), CD90 (Cy3 labeled), CD105 (Cy3 labeled), FN (Cy3 labeled) and Vim (Cy2 labeled). The fluorescence images were merged with dim-light images of the same field to show cell contours.

early passage (passages 3–4) were cultured under conditions shaped ﬁbroblast-like cells were transformed to rounded appropriate for inducing differentiation to osteogenic, cells that contained lipid vesicles, as visualized by the oil- adipogenic and chondrogenic lineages. The adipogenic red-O staining (Figure 3a). Chondrogenic differentiation potential was evaluated morphologically, where spindle- was detected by positive staining of alcian blue, which

Bone Marrow Transplantation FMB for MSC isolation from mobilized blood I Kassis et al 973 stains mucopolysaccharides or glycosaminoglycans of foci FMB. Propidium iodide staining for nuclei showed the of chondrogenic cells (Figure 3b). When induced to density of all the cells on the FMB as a reference. A blue differentiate to osteoblasts, the cells formed mineralized staining was observed on most of the FMB loaded with structures, as shown by alizarin red staining of calcium cells, indicating osteogenic alkaline phosphatase activity deposits (Figure 3d). Nitro-blue tetrazolium/indolylphos- and differentiation into osteogenic cells on most cells that phate staining also showed high activity of alkaline loaded FMB (Figure 4b, black arrow). On some beads, all phosphatase, typical of the activity of osteoblasts in the cells were stained, whereas in others, only part of them foci of differentiated cells (Figure 3c). differentiated. This hints at the possibility that once a few cells initiate differentiation, they induce similar differentiation in neighboring cells that reside on the same bead Differentiation of mesenchymal stem cells to bone-forming (Figure 4a, white arrow). On plastic two-dimensional cells on ﬁbrin microbeads culture, cell differentiation was also evident in the forma- To test the ability of the isolated FMB-loaded cells to tion of foci of high number of differentiated cells. differentiate into bone-forming cells, the MSC isolated from PBPC were cultured in suspension on the beads for 5 days in RM. Then, they were supplemented with bone Discussion differentiation medium for further 10 days. Nitro-blue tetrazolium/indolylphosphate staining was used to observe Major efforts have been made to examine techniques for alkaline phosphatase activity of the osteogenic cells on the isolation of MSC from PB with methods adapted from the

a b

50m 200

c d

100 100

Figure 3 In vitro differentiation of mesenchymal stem cells isolated by ﬁbrin microbeads from mobilized peripheral blood. Mesenchymal stem cells, expanded by 3–4 passages, were driven to differentiate into (a) adipocytes, as visualized by oil-red-O staining that stains cellular oil droplets, and (b) chondrocytes that were detected by alcian blue staining, which stains matrix secreted by these cells. Differentiation to osteoblasts was demonstrated by the alkaline phosphatase activity detected by nitro-blue tetrazolium/indolylphosphate (c) and by secreted calciﬁed matrices that were stained positive by alizarin red (d).

Bone Marrow Transplantation FMB for MSC isolation from mobilized blood I Kassis et al 974 a Table 1 Reports on isolation circulating human MSC/stromal cells Study Blood MSC/stromal Notes (ref) fraction cells isolated

Fernandez PB, PBPC PB | PBPC Â Samples collected from et al.16 breast cancer patients after chemotherapy Lazarus PBPC Â Samples collected from et al.21 cancer patients undergoing autologous infusion of PBPC Zvaifler Elutriated | Samples from healthy non- et al.22 PB mobilized blood Initial volume of each blood sample was 50 ml and 7–8 steps of eluatriation was needed to 200 m reach a yield of 0.05% Kuznetsov PB | Only two colonies of et al.17 adherent fibroblast-like b cells were observed The isolated cells were CD45, CD14 and CD34 negative; however, they did not express mesenchymal markers such as CD105 and Stro-1 In vivo bone formation by the isolated cells was observed Wexler PBPC Â Cultures from PBPC et al.13 showed a minimal sparse fibroblast-like cells that could not be passaged, and expressed CD45+ and CD14+

200 m Abbreviations: MSC ¼ mesenchymal stem cells; PB ¼ peripheral blood; PBPC ¼ peripheral blood progenitor cell. | MSC/stromal–mesenchymal cells detected. Figure 4 Differentiation of bone-like structures on fibrin microbeads Â MSC/stromal–mesenchymal cells not detected. (FMB). (a) Cells were loaded and proliferated on FMB for a prolonged period of about 2 weeks with osteogenic differentiation medium. The cell density on the FMB is shown by propidium iodide red fluorescence of cell nuclei (white arrow). (b) Staining with nitro-blue tetrazolium/indolylphosphate showing alkaline phosphatase activity of the cells that were driven to differentiate into bone-forming cells (black arrow). transiently attach to the FMB. Macrophages that may bind to FMB are not tightly bound and are detached within the isolation period of 48 h. Circulating endothelial cells may techniques for their isolation from BM.5 However, the also attach FMB, but in the process of expansion, a success of this process was controversial.13,16–23 Some negative selection leaves on the FMB only cells with groups looked for MSC in samples of cord blood, in fibroblastic phenotype. In previous reports, we showed that normal PB of adult donors or in G-CSF-mobilized PBPC FMB-mediated MSC isolation from rat BM was 3–5 times in the PB of adult volunteers,10–13,16,22 as summarized in higher in comparison to the conventional isolation techni- Table 1. But the results so far were controversial. que of adherence to plastic.26 Fibrin microbeads are not identical to native fibrin gels. The current study shows the ability of FMB to isolate They are composed of highly crosslinked, dense, stable and MSC from G-CSF of mobilized blood of healthy donors partially denatured three-dimensional fibrin matrix. The collected by apheresis. This procedure separates efficiently FMB do not exhibit procoagulant activity but retain the from the donor’s mobilized blood circulation a higher high cell-binding properties of native fibrin toward non- number of cells from the mononuclear fraction, which hematopoietic cells, including mesenchymal cells.29–32 The contain mostly hematopoietic stem cells for transplantation mechanism of cells binding to this fibrin-based matrix 24–25 with relatively low contamination by other cell types. may be at least partially attributed to the exposure of Mobilization may also increase the release into circulation haptides, new cell-binding epitopes on fibrinogen that of other BM-derived stem cells, including MSC, which may selectively adhere to matrix-dependent cells.34 Fibrin also be found in the mononuclear fraction. microbeads could be used for the isolation of MSC, for In eight out of 11 samples of mobilized blood, we their expansion and differentiation. We previously reported isolated a significant number of MSC, that could be further that FMB could separate MSC from a suspension of mixed expanded. The yield of isolated MSC from mobilized PB cell types with high efficiency.27 Other cell types may also downloaded and expanded on the flasks at days 17–18 by

Bone Marrow Transplantation FMB for MSC isolation from mobilized blood I Kassis et al 975 the FMB technique was only B0.5%. This yield was lower contamination with rounded non-mesenchymal cells could than that we have previously reported for whole rat BM.26 be observed (Figure 2c). In order to get a clearer Nevertheless, this small cell number was still sufficient to characterization of the differentiation potential of the allow separation and expansion of such cells from most of isolated cells, they were driven to differentiate into different the samples. By contrast, the number of cells isolated by mesodermal cell types, which included osteocytes, chon- conventional plastic adherence that could be further drocytes and adipocytes (Figure 3). Indeed, the isolated expanded was negligible in all the samples tested, as progenitors showed indications for their mesenchymal previously reported by other groups using the plastic nature, both morphologically and immunophenotypically. separation methods.13,21,22 Mesenchymal stem cells isolated from human mobilized The fact that not all samples of mobilized blood could PB were able to expand and differentiate into bone-forming yield MSC, even with our FMB technique, may be cells while still attached to the FMB (Figure 4). The attributed to different factors. It may be associated with implications of these findings are that the cells on the FMB the variability in the number of progenitors secreted upon can be driven to differentiate into the desired phenotype G-CSF treatment. It may also be associated with the age and then to be injected into a target organ while still on the and health condition of the donor and the exact time of biodegradable FMB. This may have major applications for aphaeresis after G-CSF administration. These factors regenerative medicine. For example, FMB could be used to should be further explored in order to improve the chances separate and expand MSC from autologous mobilized of recruiting MSC from mobilized blood. blood obtained from the pheresis of mobilized PB. FACS analysis showed that most cells in the fresh PBPC Following MSC expansion and differentiation into the samples were of hematopoietic origin, with positive desired phenotype, the same FMB loaded with cells could expression for CD45 and negative expression for CD90 then be used to transplant these autologous cells into the and CD105 (Figure 2a). Cells isolated by the conventional target organ, without the need to trypsinize the cells from plastic adherence method yielded mostly non-fibroblast- the FMB, which serve as their supporting biodegradable like cells resembling the morphology of monocytes and matrix. This may permit simpler delivery of the cells in macrophages (Figure 1f). Rare fibroblast-like cells were conditions that allow for better survival and better chances observed on the plastic plates, and they did not expand to regenerate the wounded tissue. further. Culturing in different media (RM or PM) had no significant effect on the outcome. However, cells isolated by FMB and downloaded onto Acknowledgements plastic were mostly spindle-shaped, with fibroblast-like morphology (Figure 1a–e). In some cases, these cells were We thank Anna Hotovely-Solomon, Irena Shimeliovich and initially contaminated with small population of round cells Elena Gaberman for their technical help and advice. This work and monocytes. The best yield with highest cell purity was was partially supported by the Israel Science Foundation Grant achieved when the cell isolation by FMB was performed in #697/001 to RG and by HAPTO Biotech Ltd. RM. The majority of cells isolated by FMB and further expanded in culture for 2–3 passages stained poorly for CD45 and CD34, which served as hematopoietic References cell markers, and stained positive for mesenchymal and stromal markers CD105, CD90, vimentin and fibronectin 1 Fridenshtein A, Piatetzky-Shapiro I, Petrakova K. Osteo- (Figure 2b). The moderate expression of CD34 þ in the genesis in transplants of bone marrow cells. J Embryol fraction of the FMB-isolated fibroblast-like cells in some of Exp Morphol 1966; 16: 381–390. the samples is of interest (not shown), as BM-derived MSC 2 Fridenshtein A. Stromal bone marrow cells and the hemato- were previously described as CD34 negative.5 Similar poietic microenviroment. Arkh Patol 1982; 44: 3–11. observations related to progenitor cells in PB of cancer 3 Caplan AI. Mesenchymal stem cell. J Orthop Res 1991; 9: patients were previously reported.16 This cell fraction may 641–650. Ann NY represent fibrocytes in mobilized blood that display 4 Clark B, Keating A. Biology of bone marrow stroma. Acad Sci 1995; 770: 70–78. adherent, spindle-shaped morphology when cultured in 5 Pittenger M, Mackay A, Beck SC, Jaiswal R, Douglas R, 35,36 vitro and still retain leukocytic properties. These cells Moorman M et al. Multilineage potential of adult human were reported to be able to migrate into sites of tissue injury mesenchymal stem cells. Science 1999; 284: 143–147. and possibly play a role in the inflammatory process.35,36 6Zuk PA, Zhu M, Mizuno H, Huang J, Futrell JW, Katz AJ Sub-populations of these cells were reported to be able to et al. Multilineage cells from human adipose tissue: implica- secrete collagen and fibronectin and still be positive for the tions for cell-based therapies. Tissue Eng 2001; 7: 211–228. hematopoietic markers CD45, CD13 and CD34.37,38 The 7 Zuk PA, Zhu M, Ashjian P, De Ugarte DA, Huang JI, Mizuno variation in the proportion of the CD34 þ CD45 þ cells H et al. Human adipose tissue is a source of multipotent stem within the FMB-isolated cell populations may derive from cells. Mol Biol Cell 2002; 13: 4279–4295. difference in the number of fibrocytes in the bloodstream in 8 Campagnoli C, Roberts I, Kumar S, Bennett P, Bellantuono I, Fisk NM. Identification of mesenchymal stem/progenitor cells different donors, although the mesenchymal differentiation in human first-trimester fetal blood, liver, and bone marrow. of such cells as reported here should not be ruled out. Blood 2001; 98: 2396–2402. The FMB-isolated MSC, which were downloaded onto 9 Hu Y, Liao L, Wang Q, Ma L, Ma G, Jiang X et al. Isolation plastic, could be further expanded. After 4–5 passages, the and identification of mesenchymal stem cells from human fetal cell population appeared homogeneous. No significant pancreas. J Lab Clin Med 2003; 141: 342–349.

Bone Marrow Transplantation FMB for MSC isolation from mobilized blood I Kassis et al 976 10 Erices A, Conget P, Minguell J. Mesenchymal progenitor cells in and for accelerating wound healing. Methods Med Biol 2004; human umbilical cord blood. Br J Haematol 2000; 109: 235–242. 238: 11–24. 11 Romanov Y, Svintsitskaya V, Smirnov V. Searching for 26Zangi L, Rivkin R, Kassis I, Samuel S, Levdansky L, alternative sources of postnatal human mesenchymal stem Marx G et al. High yield isolation, expansion and cells: candidate MSC-like cells from umbilical cord. Stem Cells differentiation of rat bone-marrow-derived mesenchymal 2003; 21: 105–110. stem-cells with fibrin-microbeads (FMB). Tissue Engineering 12 Gang EJ, Hong SH, Jeong JA, Hwang SH, Kim SW, Yang IH 2006(in press). et al. In vitro mesengenic potential of human umbilical cord 27 Gurevich O, Vexler A, Marx G, Prigozhina T, Levdansky L, blood-derived mesenchymal stem cells. Biochem Biophys Res Slavin S et al. Fibrin microbeads for isolating and growing Commun 2004; 321: 102–108. bone marrow-derived progenitor cells capable of forming bone 13 Wexler SA, Donaldson C, Denning-Kendall P, Rice C, Bradley tissue. Tissue Eng 2002; 8: 661–672. B, Hows JM. Adult bone marrow is a rich source of human 28 Gorodetsky R, Clark R, An J, Gailit J, Levdansky L, Vexler A mesenchymal ‘stem’ cells but umbilical cord and mobilized et al. Fibrin microbeads (FMB) as biodegradable barriers for adult blood are not. Br J Haematol 2003; 121: 368–374. culturing cell and for accelerating wound healing. J Invest 14 Mareschi K, Biasin E, Piacibello W, Aglietta M, Madon E, Dermatol 1999; 112: 866–872. Fagioli F. Isolation of human mesenchymal stem cells: bone 29 Naito M, Funaki C, Hayashi T, Yamada K, Asai K, marrow versus umbilical cord blood. Haematologica 2001; 86: Yoshimine N et al. Substrate-bound fibrinogen, fibrin and 1099–1100. other cell attachment-promoting proteins as a scaffold for 15 Lapidot T, Petit I. Current understanding of stem cell cultured vascular smooth muscle cell. Atherosclerosis 1992; 96: mobilization: the roles of chemokines, proteolytic enzymes, 227–234. adhesion molecules, cytokines, and stromal cells. Exp Hematol 30 Fang H, Peng S, Chen A, Li F, Ren K, Hu N. Biocompatibility 2002; 30: 973–981. studies on fibrin glue cultured with bone marrow mesenchymal 16Fernandez M, Simon V, Herrera G, Cao C, Del Favero H, stem cells in vitro. J Huazhong Univ Sci Technol Med Sci 2004; Minguell JJ. Detection of stromal cells in peripheral blood 24: 272–274. progenitor cell collections from breast cancer patients. Bone 31 Bensaid W, Triffitt JT, Blanchat C, Oudina K, Sedel L, Petite Marrow Transplant 1997; 20: 265–271. H. A biodegradable fibrin scaffold for mesenchymal stem cell 17 Kuznetsov SA, Mankani MH, Gronthos S, Satomura K, transplantation. Biomaterials 2003; 24: 2497–2502. Bianco P, Robey PG. Circulating skeletal stem cells. J Cell Biol 32 Ye Q, Zund G, Benedikt P, Jockenhoevel S, Hoerstrup SP, 2001; 153: 1133–1140. Sakyama S et al. Fibrin gel as a three dimensional matrix in 18 Huss R, Lange C, Weissinger EM, Kolb HJ, Thalmeier K. cardiovascular tissue engineering. Eur J Cardiothorac Surg Evidence of peripheral blood-derived, plastic-adherent 2000; 17: 587–591. CD34(À/low) hematopoietic stem cell clones with mesenchymal 33 Karp JM, Sarraf F, Shoichet MS, Davies JE. Fibrin-filled stem cell characteristics. Stem Cells 2000; 18: 252–260. scaffolds for bone-tissue engineering: an in vivo study. J 19 Wu GD, Nolta JA, Jin YS, Barr ML, Yu H, Starnes VA et al. Biomed Mater Res 2004; 71A: 162–171. Migration of mesenchymal stem cells to heart allografts during 34 Gorodetsky R, Vexler A, Shamir M, An J, Levdansky L, chronic rejection. Transplantation 2003; 75: 679–685. Shimeliovich I et al. New cell attachment peptide sequences 20 Ojeda-Uribe M, Brunot A, Lenat A, Legros M. Failure to from conserved epitopes in the carboxy termini of fibrinogen. detect spindle-shaped fibroblastoid cell progenitors in PBPC Exp Cell Res 2003; 287: 116–129. collections. Acta Haematol 1993; 90: 139–143. 35 Bucala R, Spiegel LA, Chesney J, Hogan M, Cerami A. 21 Lazarus HM, Haynesworth SE, Gerson SL, Caplan AI. Circulating fibrocytes define a new leukocyte subpopulation Human bone marrow-derived mesenchymal stromal (progeni- that mediates tissue repair. Mol Med 1994; 1: 71–81. tor cells MPCs) cannot be recovered from peripheral blood 36Yang L, Scott PG, Giuffre J, Shankowsky HA, Ghahary A, progenitor cell collections. J Hematother 1997; 6: 447–455. Tredget EE. Peripheral blood fibrocytes from burn patients: 22 Zvaifler NJ, Marinova-Mutafchieva L, Adams G, Edwards CJ, identification and quantification of fibrocytes in adherent cells Moss J, Burger JA et al. Mesenchymal precursor cells in the cultured from peripheral blood mononuclear cells. Lab Invest blood of normal individuals. Arthritis Res 2000; 2: 477–488. 2002; 82: 1183–1192. 23 Forsyth CB, Solovjov DA, Ugarova TP, Plow EF. Integrin 37 Simmons P, Torok-Storb B. CD34 expression by stromal aMb(2)-mediated cell migration to fibrinogen and its recogni- precursors in normal human adult bone marrow. Blood 1991; tion peptides. J Exp Med 2001; 193: 1123–1133. 78: 2848–2953. 24 Gorodetsky R, Vexler A, An J, Mou X, Marx G. 38 Siena S, Bregni M, Brando B, Ravagnani F, Bonadonna G, Haptotactic and growth stimulatory effects of fibrin(ogen) Gianni AM. Circulation of CD34+ hematopoietic stem cells in and thrombin on cultured fibroblasts. J Lab Clin Med 1998; the peripheral blood of high-dose cyclophosphamide-treated 131: 269–280. patients: enhancement by intravenous recombinant human 25 Gorodetsky R, Vexler A, Levdansky L, Marx G. Fibrin granulocyte–macrophage colony-stimulating factor. Blood microbeads (FMB) as biodegradable carriers for culturing cells 1989; 74: 1905–1914.

Bone Marrow Transplantation