biomaterial combinations for stem cell-based tissue engineering

ARTICLE IN PRESS

Biomaterials 29 (2008) 302–313 www.elsevier.com/locate/biomaterials

Assessment of stem cell/biomaterial combinations for stem cell-based tissue engineering Sabine Neussa,b,, Christian Apelc, Patricia Buttlerc, Bernd Deneckea, Anandhan Dhanasinghd, Xiaolei Dinge,f, Dirk Grafahrendd, Andreas Grogerg, Karsten Hemmrichg, Alexander Herrh, Willi Jahnen-Dechenta,f,i, Svetlana Mastitskayaa, Alberto Perez-Bouzab, Stephanie Rosewicka,b, Jochen Salberd, Michael Wo¨ltjea, Martin Zenkee,f a

Interdisciplinary Centre for Clinical Research, IZKF ‘‘BIOMAT.’’, RWTH Aachen University, Pauwelsstrasse 30, 52074 Aachen, Germany b Institute of Pathology, RWTH Aachen University, Aachen, Germany c Department of Conservative Dentistry, Periodontology and Preventive Dentistry, RWTH Aachen University, Aachen, Germany d DWI e.V. and Institute of Technical and Macromolecular Chemistry, RWTH Aachen University, Aachen, Germany e Department of Cell Biology, Institute for Biomedical Engineering, RWTH Aachen University, Aachen, Germany f Helmholtz Institute for Biomedical Engineering, RWTH Aachen University, Aachen, Germany g Department of Plastic Surgery, Hand Surgery, Burn Unit, RWTH Aachen University, Aachen, Germany h Institute of Clinical Genetics, Medical Faculty Carl Gustav Carus, Dresden, Germany i Institute for Biomedical Engineering, Biointerface Group, RWTH Aachen University, Aachen, Germany Received 11 July 2007; accepted 18 September 2007

Abstract Biomaterials are used in tissue engineering with the aim to repair or reconstruct tissues and organs. Frequently, the identification and development of biomaterials is an iterative process with biomaterials being designed and then individually tested for their properties in combination with one specific cell type. However, recent efforts have been devoted to systematic, combinatorial and parallel approaches to identify biomaterials, suitable for specific applications. Embryonic and adult stem cells represent an ideal cell source for tissue engineering. Since stem cells can be readily isolated, expanded and transplanted, their application in cell-based therapies has become a major focus of research. Biomaterials can potentially influence e.g. stem cell proliferation and differentiation in both, positive or negative ways and biomaterial characteristics have been applied to repel or attract stem cells in a niche-like microenvironment. Our consortium has now established a grid-based platform to investigate stem cell/biomaterial interactions. So far, we have assessed 140 combinations of seven different stem cell types and 19 different polymers performing systematic screening assays to analyse parameters such as morphology, vitality, cytotoxicity, apoptosis, and proliferation. We thus can suggest and advise for and against special combinations for stem cell-based tissue engineering. r 2007 Elsevier Ltd. All rights reserved. Keywords: Mesenchymal stem cells; Preadipocytes; Endothelial progenitor cells; Dental pulp stem cells; Hematopoietic stem cells; Embryonic stem cells

1. Introduction In modern medicine, natural and synthetic biomaterials play an increasingly important role in the treatment of Corresponding author. Interdisciplinary Centre for Clinical Research, IZKF ‘‘BIOMAT.’’ RWTH Aachen University, Pauwelsstrasse 30, 52074 Aachen, Germany. Tel.: +49 241 8080622; fax: +49 241 8082439. E-mail address: [email protected] (S. Neuss).

0142-9612/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2007.09.022

diseases and the improvement of health care [1]. To date, many biomaterials, such as titanium, polyetherurethane or polydimethylsiloxane are used routinely. The development of novel ‘‘smart’’ biomaterials with optimized characteristics for very specific applications has become a main research focus [1–4]. For tissue-engineering applications, biomaterials often serve as scaffold for a specific cell type. An ideal scaffold should provide chemical stability or degradability and physical properties matching the surrounding tissue to

ARTICLE IN PRESS S. Neuss et al. / Biomaterials 29 (2008) 302–313

provide cytocompatibility, support adhesion, proliferation, stability, and mechanical strength. The adaptation of biomaterials for tissue-engineering applications is an iterative process: Usually a biomaterial is tested in combination with only one specific cell type. More recently, combinatorial approaches have been employed to identify biomaterials suitable for specific applications. In a landmark study, Langer et al. [5] produced a biomaterial array consisting of 576 individual polyacrylate spots. This array allowed for the simultaneous analysis of hundreds of cell–polymer combinations on a single microscopic slide. However, the robotic synthesis strategy limited the choice of biomaterials to relatively innoxious reacting solvents around ambient temperature. The parallel analysis of thermoplastic polymers, metals or ceramics that make up the majority of biomaterials currently in medical use was precluded. Further studies described the use of biomaterial arrays consisting of polymers or extracellular-matrix molecules [6,7]. Such arrays can be used for a high throughput screening of cell–biomaterial interactions and thus to identify materials supporting a specific cell function. Traditionally, cell–material studies are limited to few materials and an established cell line or a single-cell type. A relatively novel tissue-engineering concept advocates the use of scaffolds specifically designed to differentiate precursor cells or even stem cells into a defined phenotype in situ at the implantation site. To identify scaffolds with such innovative properties requires the testing of a maximum number of cell–material combinations with subsequent unbiased evaluation of cell proliferation and differentiation. Nowadays, stem cells represent a particularly attractive cell type for tissue-engineering applications. Stem cells are characterized by two unique properties in one cell: their high self-renewal activity and their multilineage differentiation potential, which make them an ideal source for cellular therapy and regenerative medicine. These cells can be expanded in vitro and differentiated into diverse cell types, processes that can be supported or induced by biomaterials [8]. Parameters such as surface topography, chemistry (physicochemical property) including surface wettability (surface energy) and surface charge strongly influence cell–material interactions [9]. So far, no general principles are known that allow a prediction of the extent of cellular behaviour on a given biomaterial [10]. Therefore, cell adhesion, morphology, vitality, proliferation, cytotoxicity, and apoptosis have to be analysed and matched into a basic assessment. We here introduce a grid-based platform for the assessment of stem cell–biomaterial interactions. We chose several stem cell types and thus compared pluripotent embryonic vs. multipotent adult stem cells (mesenchymal stem cells, preadipocytes, dental pulp stem cells, hematopoietic stem cells, and endothelial progenitor cells). All adult stem cell types are of mesodermal origin, but are precursors for different specialized cell types. We established a biomaterial

303

bank, comprising established and newly developed polymers, but also allowing testing of ceramics and ceramic polymer blends. We report the systematic screening of 140 different combinations of stem cells and polymers and demonstrate the usefulness of multifactorial analyses in the testing of cell–material combinations. 2. Materials and methods 2.1. Materials 2.1.1. Biopolymers, degradable 2.1.1.1. Alginate. Alginic acid sodium salt (Algin, sodium alginate) from brown algae (Macrocystis pyrifera) was purchased in BioChemika quality from Sigma-Aldrich Chemie GmbH (Germany). Alginate films were produced by adding 20 ml of an aqueous 1% (w/v) sodium alginate solution to 0.5 ml of 0.01 M CaCl2. Films were then stabilized by crosslinking alginate molecules with six-arm star-shaped isocyanateterminated poly(ethylene glycol) (IPDI-starPEG, Mw: 18,000 g/mol, SusTech GmbH & Co. KG Darmstadt, Germany) [27]. Alginate films were disinfected by spraying with 70% ethanol, washed six times with aqua bidest and finally six times with phosphate buffered saline (PBS, pH 7.4) before being placed in cell culture dishes of tissue culture polystyrene (TCPS, Greiner bio-one, Germany). Unless otherwise noted, this disinfection procedure was used for all other materials. 2.1.1.2. Collagen. The constant quality collagen matrices used in this study are large scale commercial products by Dr. Suwelack Skin & Health Care AG (Billerbeck, Germany). A collagen suspension containing collagen types I, III, and V was isolated from bovine skin (bovine spongiform encephalopathy tested, BSE tested). The dermal collagen was extensively highly purified and freeze-dried resulting in a sponge-like matrix structure. Collagen samples were taken from collagen sheets (1 mm thick) using a hole punch. For cell culture experiments, porous collagen scaffolds were transferred into non-porous sheets by compression between two stainless steel plungers and 10 tonne pressure at ambient temperature. 2.1.1.3. Fibrin. One hundred and eighty microlitre of a sterile fibrinogen-suspension consisting of 50 ml CaCl2 (50 mM, Roche, Mannheim, Germany), 120 ml GBSH5 buffer (without glucose and Ca2+), and 830 ml fibrinogen (20 mg/ml, Sigma, Steinheim, Germany) was mixed with 20 ml thrombin (10 units/ml, Sigma, Steinheim, Germany) and poured in cell culture plates to polymerize [15]. 2.1.1.4. Hyaluronic acid. Potassium hyaluronan (hyaluronic acid potassium salt, potassium hyaluronate, HA) from human umbilical cord with a protein content less than 2% (w/v) was purchased by Sigma-Aldrich Chemie GmbH (Germany). HA films were produced by mixing a 0.7% (w/v) solution of potassium hyaluronan (Mw: 750,000 g/mol) with a 15% (w/v) solution of IPDI-starPEG (Mw: 18,000 g/mol, SusTech GmbH & Co. KG Darmstadt, Germany) [27]. Freshly prepared solutions were mixed and directly casted onto a glass plate. Foil thickness was adjusted by using a scraper to approximately 600 mm. 2.1.2. Degradable synthetic polymers 2.1.2.1. BAK 1095. Polyesteramide type BAK 1095 was a gift from Bayer MaterialsScience AG (Germany). The degradable polymer was synthesized in an industrial plant, extruded and transferred into granules successively. The polymer was prepared by a two-step (ring-opening polymerization (ROP) and polycondensation) one-batch reaction [28]. Foil production was established by a melt-press process. Polymer granules were ground to powder in a cryo-mill. 1.2 g polymer powder per foil was placed between Teflon-covered metal plates, temperature was raised to 180 1C and maintained for 5 min. A load of 1 tonne was applied for 9 min. After cooling to room temperature the prepared foil was taken out of the

ARTICLE IN PRESS 304

S. Neuss et al. / Biomaterials 29 (2008) 302–313

metal plates. Blankets of different diameters were prepared by using a hole punch. Before disinfection procedure, BAK 1095 foils were cleaned by rinsing with hexane. Afterwards, foils were first dried in a slight nitrogen stream and finally for 2 h under vacuum. 2.1.2.2. Polyesteramide type C (PEA C). Synthesis and characterization of PEA C developed by DWI (German Wool Research Institute at RWTH Aachen University) is described elsewhere [28]. Melt-pressed foils were prepared and cleaned as described for BAK 1095. 2.1.2.3. Poly(e-caprolactone) (PCL). PCL (Mw: 80,000 g/mol) was purchased from Sigma-Aldrich GmbH (Germany). For each foil 3 g PCL granules were used. The PCL was placed on Teflon-covered metal plates, the temperature of the plates was raised to 85 1C and maintained for 5 min. A load of 1 tonne was applied for 1 min at the same temperature. After cooling to room temperature the prepared foil was taken out of the metal plates. Cleaned foils of different size were prepared as described for BAK 1095. 2.1.2.4. Resomers types. All Resomers types poly(L-lactic acid) (L209S, 2.6–3.2 dL/g), poly(D,L-lactic acid) (R203S, 0.25–0.35 dL/g), poly(D,L-lactic-co-glycolic acid) (lactic-glycolic acid ratio 50:50, RG503, 0.32–0.44 dL/g), poly(L-lactic-co-D,L-lactic) (L-lactic-D,L-lactic acid ratio 70:30, LR705, 2.0–2.8 dL/g), and poly(L-lactic acid-co-trimethylene carbonate) (lactic acid-trimethylene carbonate ratio 70:30, LT706, 1.2–1.6 dL/g) were purchased from Boehringer Ingelheim Pharma GmbH & Co. KG (Germany). L209S, LR705, RG503, and LT706 foils were prepared by melt-pressing technique (see above) and R203S foils by film casting. Ten grams of the polymer were dissolved in 60 ml chloroform and filled into a PTFE mould. Film formation was performed under controlled evaporation conditions for 24 h. All Resomers-type foils were cleaned by rinsing with isopropanol. Afterwards the materials were dried first in a nitrogen stream and then for 2 h under vacuum before disinfection protocol (see above). 2.1.2.5. PEO10-b-PCL60. Methoxy-terminated poly(ethylene oxide) (MPEO, Mw: 10,000 g/mol) was purchased from Iris Biotech GmbH (Germany). e-Caprolactone (e-CL) and stannous octoate (Sn(Oct)2) were obtained from Sigma-Aldrich GmbH (Germany) and used without further purification. MPEO10,000 was reacted with e-CL in the presence of Sn(Oct)2 as a catalyst at 130 1C for 20 h under a nitrogen atmosphere. Synthesized block copolymer was dissolved in dichloromethane and precipitated into cold n-hexane three times to remove unreacted monomers. The final product was dried under high vacuum. Foils were prepared by melt-pressing technique (see above). 0.5 g polymer powder per foil was used. Polymer powder was placed between PTFE covered metal plates, temperature was raised to 60 1C and maintained for 3 min. A load of 1 tonne was applied for 1 min at the same temperature. 2.1.2.6. PEO10-b-PDLLA25. PEO10-PDLLA25 was synthesized by ROP as described above. Dimeric D,L-lactic acid was obtained from Boehringer Ingelheim Pharma GmbH & Co. KG (Germany). Foils were prepared by melt-pressing technique (see above) and 1 g polymer powder per foil was used. Polymer powder was placed between PTFE covered metal plates, temperature was raised to 65–70 1C and maintained for 2.5 min. A load of 0.5 tonne was applied for 30 s at the same temperature. 2.1.3. Non-degradable synthetic polymers 2.1.3.1. PDMS. Poly(dimethyl siloxane) (PDMS) foils were prepared by film casting method from Sylgards 184 (Dow Corning, USA). Base and curing agent were mixed 10:1 (w/w). Bubble formation was prevented by evaporation. The mixture was casted on a cleaned glass plate and distributed by a scraper with a distance of 1 mm. Subsequently, the glass plate with this homogeneously distributed mixture was incubated for 1 h at 150 1C. Samples were cleaned by ultrasonication using ethanol for 2 times per 10 min and then dried in a dissicator over night.

2.1.3.2. Texins 950. Aromatic polyether-based thermoplastic polyurethane (PEU) Texins 950 was purchased from Bayer MaterialScience AG (Germany). Three grams of granules per foil were pre-heated at 120 1C for 1 h. The metal plates of the melt press were covered by aluminium foils and the granules were placed at the centre of the metal plates. Then, the temperature was raised to 200 1C and maintained for 5 min. A load of 5 tonne was applied for 10 s at the same temperature. The prepared foil was taken out after cooling down the metal plates to room temperature. The thickness of the foils were found to be approx. 230 mm. Extraction of PEU samples with hexane/ethanol mixture (79:21; v/v) was carried out in a Soxhlet. The samples were dried for 2 h at 80 1C and then for 2 days in a dissicator under vacuum. 2.1.3.3. PET, PTFE, and PVDF. Poly(ethylene terephthalate) (PET), poly(tetrafluor ethylene) (PTFE), and poly(vinylidene fluoride) (PVDF) were purchased from Goodfellow GmbH (Germany). PET and PVDF foils with a thickness of 50 mm and PTFE foils of 300 mm were used. The polymer blankets were extracted for 2 h in a Soxhlet apparatus using a hexane/ethanol mixture (79:21; v/v) and then dried at 80 1C for 1 h. Afterwards the samples were evacuated for 24 h.

2.2. Stem cells 2.2.1. Isolation and expansion of stem cells All cells were cultured in a 20% O2 and 5% CO2 humidified atmosphere at 37 1C. Human tissue donations were received with informed consent. Procedures were approved by the local University committee on Ethics in Medicine. 2.2.2. Human mesenchymal stem cells Human mesenchymal stem cells (hMSC) were isolated according to protocols from Haynesworth [11] and Pittenger [12], as previously described [13–15]. In brief, bone marrow spongiosa of patients with total hip joint endoprosthesis (TEP) was rinsed with stem cell medium several times. Spongiosa was removed and the remaining cell suspension was centrifuged for 10 min at 500g. Thereafter, the cell pellet was resuspended in stem cell medium and cells were seeded in a T75 culture flask. After 24 h, non-adherent (hematopoietic) cells were removed by medium change. Mesenchymal stem cells were expanded in medium consisting of 60% Dulbecco’s modified Eagle’s medium (DMEM) low glucose (PAA, Co¨lbe, Germany) and 40% MCDB-201 (Sigma, Steinheim, Germany) including 2% foetal calf serum (FCS) (Hyclone, Perbio Science, ErembodegemAalst, Belgium), 1  ITS-plus (insulin–transferrin–selenic acid+bovine serum albumin (BSA)–linoleic acid), 1 nM dexamethasone, 100 mM ascorbic-acid-2-phosphate, and 10 ng/ml epidermal growth factor (EGF) (all from Sigma, Steinheim, Germany). Medium was changed every 3–4 days. At 80–90% confluence, stem cells were trypsinized with stem cell trypsin (CellSystems, St. Katharinen, Germany) and reseeded in a density of 5  103 cells/cm2 for optimum proliferation. Cells were characterized by flow cytometry and multipotency was tested using standard protocols as previously described [13]. 2.2.3. Human preadipocytes Preadipocytes were isolated from freshly excised human subcutaneous abdominal or mammary fat tissue at the Department of Plastic Surgery and Hand Surgery, Burn Centre, from patients who underwent elective operations. Harvesting of adipose tissue by liposuction was performed according to Sydney Coleman (manually applied negative pressure using a 10-cc-syringe with a blunt tip cannula, no addition of any solution) from abdomen and breast [16]. Adipose tissue from aspiration was digested by collagenase treatment after washing with 0.9% NaCl. Fat lobules from excised adipose tissue were separated from capillaries and connective tissue, minced into small pieces of 2–6 mm, washed with 0.9% NaCl and digested with collagenase type I (0.494 U/ml, Sigma, Steinheim, Germany) and 1.5% BSA dissolved in collagenase buffer (100 mM HEPES, 120 mM NaCl, 50 mM KCl, 1 mM CaCl2, 50 mM Glucose, pH 7.4) for 60 min at 37 1C under constant shaking. The cell suspension was filtered through a

ARTICLE IN PRESS S. Neuss et al. / Biomaterials 29 (2008) 302–313 250 mm filter and centrifuged at 700g for 7 min at room temperature. Preadipocytes were separated from mature adipocytes by discharging the fat layer on top. Cells were seeded on tissue culture flasks in DMEM/ Ham’s F12 (1:1). After 24 h, cells were washed twice with 0.9% NaCl. 2.2.4. Human dental pulp stem cells Normal human impacted third molars were collected. Tooth surfaces were cleaned and cut around the cementum–enamel junction by using sterilized dental burs to reveal the pulp chamber. The pulp tissue was gently separated from the crown and roots and digested in a solution of 3 mg/ml collagenase type I and 4 mg/ml dispase (both from Sigma, Steinheim, Germany) for 1 h at 37 1C. Single-cell suspensions were obtained by passing the cells through a 70-mm strainer (BD, Heidelberg, Germany). Then, human dental pulp stem cells (DPSC) were seeded in culture flasks containing an a-modification of Eagle’s medium (Gibco, Karlsruhe, Germany) supplemented with 20% FCS (PAA, Co¨lbe, Germany), 100 mM L-ascorbic-acid2-phosphate (Sigma, Steinheim, Germany), and 50 mg/ml Gentamycine (Gibco, Karlsruhe, Germany) following a published protocol [17]. 2.2.5. Human endothelial progenitor cells Human peripheral blood of healthy donors was obtained from the Department of Transfusion Medicine of RWTH Aachen University. Mononuclear blood cells of the peripheral heparinized blood were isolated by density-gradient centrifugation with Biocoll (1.083 g/ml; Biochrom, Berlin, Germany) and washing with PBS (Biochrom, Berlin, Germany). CD34-positive selection of the mononuclear bloodcells (2  108 cells/ml) was performed by an EasySeps kit (Stemcell Technologie, Vancouver, Canada), using magnetic beads coated with anti-CD34 antibodies. After separation, the CD34-positive cells were resuspended in endothelial growth medium (EGM2-MV, Clonetics, Walkersville, USA). 2.2.6. Mouse mesenchymal stem cells Bone marrow was obtained from 8–9-week-old female Swiss Webster mice. The animals were sacrificed by cervical dislocation and bone marrow was flushed out of tibias and femurs. Cells were washed and the pellet was resupended in culture medium consisting of DMEM with 4.5 g/l glucose supplemented with 15% (v/v) heat inactivated FCS (both from Sigma, Steinheim, Germany), 2 mM L-glutamine, 100 U/ml penicillin, and 100 mg/ ml streptomycin (all from PAA, Co¨lbe, Germany Laboratories). Then, cells were plated in 10 cm tissue culture dishes. Non-adherent cells were removed by changing the medium. Medium was changed every 3 days until the monolayer reached 90% confluency. Then, mouse mesenchymal stem cells (mMSC) were harvested for reseeding or experimentation. 2.2.7. Mouse embryonic stem cells Mouse embryonic stem cell (ES) line R1 established from a 129/ Sv  129/Sv-CP F1 3.5-day blastocyst was obtained from Nagy [18]. Cells were cultivated on gelatine-coated dishes in DMEM with 4.5 g/l glucose containing 15% heat inactivated foetal calf serum (both Sigma, Steinheim, Germany), 2 mM L-glutamine (PAA, Co¨lbe, Germany), 1% MEM nonessential amino acids (Gibco, Karlsruhe, Germany), 100 U/ml penicillin, 100 mg/ml streptomycin (both PAA, Co¨lbe, Germany), 1000 U/ml recombinant mouse leukaemia inhibitory factor (LIF, ESGROs, Chemicon, Temecula, USA), and 0.1 mM b-mercaptoethanol (Sigma, Steinheim, Germany). The medium was changed every day. Cells at passage number 20–25 were used for experiments. 2.2.8. Mouse hematopoietic stem cells Flt3+ stem/progenitor cells were obtained from mouse bone marrow and amplified in vitro with specific cytokines as described [19,20]. Briefly, bone marrow suspensions were prepared from C57BL/6 or Balb/c mice (Charles River, Sulzfeld, Germany) and cells were seeded at 2  106 cells/ ml in Rosewell Park Memorial Institute (RPMI) 1640 medium supplemented with 10% FCS, 2 mM L-glutamine, 100 U/ml penicillin/streptomycin (all from Gibco, Karlsruhe, Germany) and 50 mM b-mercaptoethanol (Sigma, Steinheim, Germany) containing stem cell factor (SCF; 100 ng/ml), Flt3 ligand (Flt3L; 25 ng/m; PeproTech, London, UK), long-range insulin-like

305

growth factor-1 (IGF-1; 40 ng/ml; Sigma, Steinheim, Germany), IL-6/soluble IL-6 receptor fusion protein (hyper-IL-6; 5 ng/ml), granulocyte macrophage colony-stimulating factor (GM-CSF; 20 U/ml) and 106 M dexamethasone. After 3 days of culture, cells were subjected to Ficoll-Hypaque density gradient centrifugation (density 1.077 g/ml, Eurobio, Paris, France) and medium was replenished every 2 days. Cell numbers were determined with an electronic cell counter device (CASY1, Scha¨rfe Systems, Reutlingen, Germany) and hematopoietic stem cells (HSC) were maintained at 2  106 cells/ml and used for experiments at day 6–8 of culture.

2.3. Experimental procedures 2.3.1. Cytotoxicity/viability test after ISO 10993-5 To detect cytocompatible or cytotoxic polymers, a live/dead staining of stem cells on biomaterials was performed 24 h after seeding, according to the protocols 10993–5 of the International Standardization Organization (ISO). Briefly, 1.28 ml Ringer solution (Delta-Pharma, Pfullingen, Germany) containing 20 ml of fluoresceindiacetate (Sigma, Steinheim, Germany, 0.1 mg in 20 ml acetone) and 20 ml propidiumiodide (Sigma, Steinheim, Germany, 0.01 mg in 20 ml PBS) was added to the stem cell/ polymer constructs. After an incubation of 20 s, cells were analysed using fluorescence microscopy (Polyvar microscope, Leica, Bensheim, Germany). Green fluorescence indicates viable cells and red fluorescence dead cells. Quantification was done by analySIS software. The initial cell seeding density was cell type dependent, due to different cell sizes: 5  103 cells/cm2 for hMSC, DPSC, and preadipoytes, 1.25  104 cells/cm2 for mMSC, 18  104 cells/cm2 for mES, 1  105 cells/cm2 for EPC, and 2.5  105 cells/cm2 for HSC. 2.3.2. Cell viability, cytotoxicity, and apoptosis (multiplex assay) By combining three independent test kits for cell viability (CellTiterBlue), cytotoxicity (CytotoxOne), and apoptosis (ApoOne, all Promega, Mannheim, Germany), we analysed three different parameters from one single sample, consecutively. Briefly, stem cells were seeded on all polymers of our biomaterial bank in 96-well plates (2  10,000 cells/ 100 ml). After 24 h, the supernatant was transferred into a black 96-well plate (Greiner, Frickenhausen, Germany) and mixed with 100 ml of cytotoxicity reagent. After an incubation of 10 min at room temperature (RT), lactate-dehydrogenase (LDH) was detected for quantification of cytotoxicity. One hundred and twenty microliter culture medium containing 20 ml of viability reagent was added to the cell/polymer constructs and incubated for 1 h. The supernatant was transferred into a black 96-well plate to detect cell metabolism. Then, 200 ml culture medium containing 100 ml of apoptosis reagent was added to the cell/polymer constructs. After 2 h of incubation, the supernatant was transferred into a black 96-well plate to detect caspase-3/7 activity. Fluorescence intensity was measured using the fluorometer FLUOstar OPTIMA (BMG Labtech, Jena, Germany). For viability and cytotoxicity, the excitation and emission was 560 and 590 nm, respectively. For apoptosis, excitation and emission was 485 and 520 nm, respectively. As a positive control for the cytotoxicity assay, 2 ml lysis buffer was used to permeabilize cell membranes and thus release LDH. A 5 h incubation with 2 mM staurosporine served as a positive control for apoptosis. Polymers without stem cells were parallel incubated to serve as negative controls (background fluorescence). Each assay was performed in quadruplicates. 2.3.3. Proliferation of stem cells on polymers For the detection of proliferation, we performed the viability assay (CellTiterBlue, Promega, Mannheim, Germany), described above for the multiplex assay, at 1 and 3 days after cell seeding. 2.3.4. Scanning electron microscopy (SEM) HMSC/polymer biohybrids were fixed in 3% glutaraldehyde for at least 24 h, rinsed with sodium phosphate buffer (0.2 M, pH 7.39, MERCK, Darmstadt, Germany) and dehydrated by incubating consecutively in 30%, 50%, 70%, 90% acetone and then three times in 100% acetone for 10 min. The constructs were then critical-point-dried in liquid CO2, and

ARTICLE IN PRESS 306

S. Neuss et al. / Biomaterials 29 (2008) 302–313

then sputter-coated with a 30 nm gold layer. Samples were analysed using an environmental scanning electron microscope (ESEM XL 30 FEG, FEI, PHILIPS, Eindhoven, The Netherlands) in a high vacuum environment. 2.3.5. Principal component analysis and hierarchical clustering Datasets of multiplex assays were pseudonormalized by using the logarithmic values. The normalization step included mean centring and additional division by the standard deviation for rows and columns. Hierarchical clustering was performed using Pearson’s correlation as dissimilarity matrix and average linkage as agglomeration rule. All described calculations, including principal component analysis (PCA), were performed with the Genesis Software Package (v1.72) [21]. 2.3.6. XPS X-ray photoelectron spectroscopy (XPS) analysis was performed using an X-ProbeTM 206 spectrometer (Surface Science Instruments, USA). An alumina anode was used as X-ray source. The binding energies were referenced to hydrocarbon at 285.0 eV. The emission angle of electrons was set at 351 with respect to the sample normal, which results in an information depth of about 6 nm. 2.3.7. SFM Scanning force microscopy (SFM) investigations were performed with an Extended Multimode NanoScope IIIa (Digital Instruments, USA) operating in tapping mode TM. The oscillation frequency for tapping mode TM was set in the range of 320–360 kHz depending on the Si cantilever (k50 N/m, Nanosensors). 2.3.8. Contact angle measurements Static contact angles were measured on surfaces in ultrapure water with a goniometer microscope G40 (Kru¨ss GmbH, Germany) using the captive bubble method. At least 10 bubbles were measured per sample.

3. Results 3.1. Grid platform A grid-based platform was designed to assess cell–biomaterial interactions testing 7 different stem cells types for their behaviour on 19 different biomaterials. We used 24and 96-well fitted polymers for a standardized, parallel analysis of the five parameters morphology, vitality, cytotoxicity, apoptosis, and proliferation characterizing the stem cell–material interactions. 3.2. Stem cell morphology We selected two fluorine containing non-degradable polymers PVDF and PTFE, and two degradable polymers, the Resomerss RG503, poly(D,L-lactic-co-glycolic acid) with a lactic acid to glycolic acid ratio of 50:50, and LT706, poly(L-lactic acid-co-trimethylene carbonate) with a lactic acid to trimethylene carbonate ratio of 70:30. The materials were manufactured as flat polymer discs with roughness values below 250 nm, root mean square roughness (rms) was measured between 5 nm for PVDF and 83 nm for PTFE, respectively (Fig. 1). Wettability was analysed by the captive bubble method and average contact angle values differ between 1121 for PTFE and 731 for LT706. Average contact angles of PVDF and RG503 were estimated to be 791 (Fig. 1). The elemental composition

of the pre-conditioned biomaterials measured by means of XPS demonstrated no significant differences neither for both fluorine containing non-resorbable polymers, nor for the two biodegradable amorphous Resomers types. The PVDF spectrum only shows minor impurities (less than 1 at%) of inorganic origin. Fig. 1 clearly demonstrates different morphologies of hMSC on four different polymers. On PVDF and on RG503, two polymers of different chemical composition, hMSC morphology is identical to that on hMSC cultured on TCPS (not shown). Cells displayed a flat, fibroblast-like phenotype. In contrast, PTFE and LT706 yield in a round morphology of hMSC. These results indicate that cell morphology depends on topographical, as well as on chemical characteristis of the biomaterial. Thus, neither the chemical composition, nor the surface topography of a given biomaterial can definitively predict the cell morphology and behaviour. 3.3. Assessment of cytocompatiblity Live/dead stainings were performed to assess cytocompatibility for all combinations of stem cells and polymers contained in the grid platform. Fig. 2 shows representative images of live/dead stainings with viable cells (green) and dead cells (red). Visual inspection readily suggests impression that polymers sustained attachment and growth of specific stem cell types. As shown in Fig. 2, some polymers impaired the attachment of all stem cells, e.g. PEO10–PDLLA25 or BAK1095, while others only impaired the attachment of specific types of stem cells. For example PDLLA (Resomer R203S) supports the attachment of all stem cell types except mouse embryonic stem cells. Certain polymers preferrentially allowed attachment by only one or two stem cell types, e.g. only mMSC adhere to PEA-C, and only DPSC and HSC adhere to PDMS. To separate adhesion characteristics from cytocompatibility issues, we investigated the LDH secretion as a proxy of cell death. 3.4. Vitality, cytotoxicity, and apoptosis We measured cell viability, LDH secretion (as a measure for cytotoxicity), and caspase 3/7 activity (as a measure for apoptosis) in identical cell–material combinations. Results of this multivariate analysis were categorized and plotted in a heat map. Fig. 3 thus allows an assessment of suitable and unsuitable stem cell–polymer combinations. A high number of viable cells with a low cytotoxic and apoptotic signal is desirable. Such a combination was determined e.g. for EPC on PTFE, or for the DPSC on PDLLA (Resomer R203S). In contrast a low number of viable cells with a high cytotoxic and apoptotic signal was detected e.g. for HSC on PDLLA. 3.5. Stem cell proliferation Fig. 4a illustrates the fluorescence intensity of cellular resorufin as a proxy of proliferation of DPSC, hMSC, and

ARTICLE IN PRESS S. Neuss et al. / Biomaterials 29 (2008) 302–313

307

Fig. 1. Biomaterial quality control and stem cell morphology. XPS, SFM, and SEM analysis of exemplarily selected polymers. XPS analysis of PVDF, PTFE, resomer RG503, and resomer LT706 LT706 demonstrates the elemental composition and binding states (chemical functional groups) of the material surfaces. SFM analysis serves as a quality control and to estimate roughness parameters. The morphology of hMSC on PVDF, PTFE, Resomers RG503, and Resomers LT706 was analysed by SEM. The cells display the characteristic flat, fibroblast-like morphology on PVDF and on resomer RG503, and a rather round, spherical morphology on PTFE and on resomer LT706. For better visualizing the cell mophology, SEM view cell contours are separately shown. Scale bar: 50 mm.

mMSC at day 1 and 3 after cell seeding. Proliferation on identical polymers varied between stem cells of different tissues and between the same type of stem cells from different species. For example, BAK 1095 supported the proliferation of human DPSC, but not of human MSC ( ¼ cell type specific influence), although the two polymers are very similar in chemistry. Further, PTFE allowed mouse MSC proliferation, but not human MSC proliferation ( ¼ species-specific influence). For the sake of completeness, proliferation behaviour of EPC, preadipocytes, mES, and HSC on the polymers is shown in Fig. 4b. 3.6. PCA and cluster analysis Fig. 5a represents the PCA of vitality, cytotoxicity and apoptosis of the different stem cell types identifying patterns within a three-dimensional space. Interestingly,

human MSC behaved similar to human DPSC but not to human preadipocytes despite the fact that hMSC, DPSC, and preadipocytes are often described as very similar types of stromal stem cells. Furthermore, both bone-marrow derived adult mouse stem cell types, MSC and HSC, clustered in close proximity, while mouse ES cells obviously behaved different regarding vitality, cytotoxicity, and apoptosis. Our biomaterials belong to different polymer classes. Synthetic biodegradable and non-degradable polymers were investigated as well as degradable biopolymers. Cluster analysis shown in Fig. 5b clearly demonstrated that cell behaviour failed to cluster according to materials classes. PCA of the same data set revealed surprisingly similarities and differences in cell–material interactions (Fig. 5c). Although most biopolymers clustered in one quadrant, the fibril-forming protein collagen was located in

ARTICLE IN PRESS 308

S. Neuss et al. / Biomaterials 29 (2008) 302–313

an upper region of the quadrant, while fibrin was located together with the two polysaccharides alginate and hyaluronan. Furthermore, the two polyesteramides PEA-C and BAKs 1095 indeed located in quadrant I, but not in

immediate neighbourhood, although they are of similar chemical composition and are provided with similar physicochemical properties (surface topography [SEM, SFM] and wettability [sessile drop and captive bubble method]). At the moment we assume, that different surface crystallinities could lead to this distance in quadrant I and finally to the different cell behaviour. PCL, PEO10-b-PCL60 and poly(L-lactide-co-trimethylene carbonate) (PLLAco-TMC) clustered in quadrant II. Although the three Resomers types: PDLLA(R203S), PDLLA-co-GA (RG503), and PLLA-co-DLLA (LR704) are of related chemical composition, they were separately distributed but still in quadrant III. In the upper region of quadrant IV, the two elastomers Texins 950 and PDMS were located, but with a certain distance to each other. Interestingly, in the lower part of quadrant IV most of the semi-crystalline polymers are concentrated independently from their chemical composition. 3.7. Suitable and unsuitable combinations for stem cellbased tissue engineering Depending on the tissue-engineering application, a scaffold has to meet specific criteria. Results of our multiplex and proliferation assays suggest e.g. the following cell–material combinations with scaffolds supporting cell adhesion and proliferation while inhibiting apoptosis and cytotoxicity: DPSC on PDLLA (R203S); EPC on PVDF, PTFE, Texin, PLLA (L209S), Resomer LR705, PEO10-b-PDLLA25, and collagen; preadipocytes on texin; mMSC on PVDF, and texin; mES on fibrin. No materials meet these criteria for hMSC and for HSC. Further we can advice against the following combinations resulting in only weak cell adherence, but high apoptotic and necrotic signals: DPSC on PET and hyaluronic acid; preadipocytes on resomer LT706; mMSC on PEO10-b-PCL60 and PEA-C; mES on BAK; HSC on PET, Texin, PDMS, PLLA (L209S), PDLLA (R203S), PCL, PEA-C, fibrin, and collagen. 4. Discussion Array technologies for parallel analyses of large amounts of information are daily routine in modern medicine, biotechnology, and basic research. For simultaneous investigations, several kinds of ‘‘chip’’-based arrays with Fig. 2. Live/dead stainings of the stem cells on the polymers for the identification of cytocompatible combinations. (a) Cytocompatible biomaterials can be detected using a live/dead staining with fluoresceindiacetate (FDA) and propidium-iodide (PI). This staining results in greenfluorescent viable cells and red-fluorescent dead cells. Some polymers selectively prevent stem cell attachment (no cells). Some polymers are clearly cytocompatible, because all adherend stem cells are viable, e.g. PVDF for human preadipocytes (1), other polymers are cytotoxic, resulting in exclusive dead cells e.g. PDMS for human mesenchymal stem cells (2). On some polymers both, viable and dead cells are present (3). (b) Magnification of three selected combinations. Scale bars: A ¼ 500 mm; B ¼ 200 mm.

ARTICLE IN PRESS S. Neuss et al. / Biomaterials 29 (2008) 302–313

309

Fig. 3. Heat map summarising the multiplex assay results. The parameters cell vitality (metabolic activity), cytotoxicity (LDH secretion), and apoptosis (caspase 3/7 activity) were analysed in a multiplex assay for all stem cell–biomaterial combinations. Compared with stem cells on TCPS, resulting raw data were divided into data groups with corresponding colour pattern. Values differing only up to 10% to the control cells are colourless. This representation allows to find stem cell–polymer combinations with a high amount of viable cells, a low cytotoxic signal and a low apoptotic signal, which demonstrates the optimum, as well as combinations with a low amount of viable cells, but a high cytotoxic and apoptotic signal, which is not desirable.

hundreds or thousands of micro- or nano-scaled spots are developed, in particular gene chips, protein chips, and tissue chips [22]. The emerging field of biomaterial chips/ arrays allows a fast and simultaneous analysis of cell–biomaterial interactions [5–7,23]. However, the usage of biomaterial arrays to identify suitable combinations of cells and biomaterials for cellular therapy requires simultaneous assessment of cell morphology, vitality, cytotoxicity, apoptosis, and proliferation. In systematic material roughness gradient studies, human fibroblasts preferred rough surfaces for attachment and proliferation, accompanied by differences in morphology, due to the rough or smooth surface [24]. To exclude such topography-dependent changes in cell morphology, we decided to produce all our polymer discs of different materials as flat as technically possible independent from the desired surface area. However, the two amorphous Resomers types RG503 and LT706 with the same surface roughness and elemental composition (Fig. 1) and only a slight difference in their captive bubble values, 791 and 731, respectively, result in different morphologies of the same cell type. Biomaterials of the same category, e.g. non-degradable, synthetic polymers, do not essentially result in the same cell response. The fact that biomaterial categories cannot be used for the prediction of a scaffold capability for a specific cell-based application is verified by our PCA analysis. Fig. 5 demonstrates a rather randomized distribution of

polymers of the same category, although some polymers, which are chemically similar, cluster as expected. Further molecular studies on the cell surface receptors and the signalling cascades involved upon biomaterial contact should help to decipher the complex cell–biomaterial interactions. Our screening assays demonstrate that cell viability on one and the same biomaterial differs cell type dependent (Fig. 2), and biomaterials supporting cell viability can result in both, the enhancement and inhibition of proliferation. These results are in accord with results from Itthichaisri et al. [25]. The group investigated growth and proliferation of human osteoblast-like cells on seven different biomaterials. Although initial cell seeding numbers were identical and viable cells were detectable on all tested biomaterials after 1 week of culture, cell amount and thus cell proliferation was extremely different [25]. The cytotoxicity of biomaterials can be investigated in several (indirect) ways, e.g. by measuring proliferation, cell viability, or cell metabolism. For our cytotoxicity tests we first choosed a live/dead staining proposed by the national guidelines for cytotoxicity tests, ISO 10993, an internationally accepted standard, which harmonizes well with the European standard EN30993 and the American FDA’s blue book memorandum G95 [26]. These live/dead stainings have the advantage that the cells can be observed and examined microscopically, thus we can receive an impression of the cell behaviour on the biomaterial. Indeed,

ARTICLE IN PRESS 310

S. Neuss et al. / Biomaterials 29 (2008) 302–313

methods for quantification exist, but non-adherent cells are disregarded by this method. Therefore, we recommend an additional detection of LDH, which is released by cells undergoing necrosis as result of cytotoxicity and this assay includes the non-adherent cells in the supernatant. As described above, cell–biomaterial interactions are very complex mechanisms and the investigation of one or two parameters regarding the biomaterial characteristics or the cell behaviour is far from being adequate to assess the compabtility for a specific application. Thus, multifactorial analyses are essential. This fact is confirmed by the following example: Fig. 1 clearly shows a similar adherent and spreaded morphology of hMSC on PVDF and on Resomer RG503,

two polymers of absolute different chemical compositions: One semi-fluorinated, non-degradable homo-polymer (PVDF) and one degradable co-polymer without any fluorinated component (RG503). Further, the two polymers differ in their topographical characteristics, but this difference does not result in a varying cell behaviour regarding adherence and spreading. The wettability of these polymers is nearly identical, a characteristic that is very important for adhesion and spreading [10]. Both materials are semi-crystalline, a bulk characteristic with effects on the surface characteristics resulting in surfacelocated crystalline domains. Although these two polymers result in a very similar morphology of adherent hMSC, PVDF, and resomer RG503 are located distant in different

Fig. 4. Stem cell proliferation can be inhibited or supported by different polymers. Compared with control stem cells on TCPS, polymers can inhibit or enhance the proliferation of the stem cells. (a) The influence of the polymers on the proliferation can be cell type specific, but also species specific. BAK1095 for example supports the proliferation of human DPSC, but not of human MSC ( ¼ cell type specific influence), while PTFE allows mouse MSC proliferation, but not human MSC proliferation ( ¼ species-specific influence). (b) Proliferation analysis for the human EPCs and preadipocytes, as well as for the murine ES cells and HSCs on all polymers of our biomaterial bank.

ARTICLE IN PRESS S. Neuss et al. / Biomaterials 29 (2008) 302–313

311

Fig. 4. (Continued)

quadrants in the PCA analysis (Fig. 5c). Thus, the two materials interact absolute different with the stem cells, but show the same morphology. We here present the standardized investigation of various stem cell–biomaterial combinations to develop

innovative concepts for cell-based therapies and tissueengineering applications. The importance of parameters, such as cell attachment, proliferation or differentiation depends on the specific application, while the parameters apoptosis and necrosis are always unwanted. Our platform

ARTICLE IN PRESS 312

S. Neuss et al. / Biomaterials 29 (2008) 302–313

Fig. 5. Data analysis to find relationships in cell behaviour between different stem cell types and different types of polymers. The data set of our multiplex assay, which combines the three parameters vitality, apoptosis, and cytotoxicity was analysed in a PCA analysis (a) regarding the different stem cell types, and (c) regarding the different polymers. Biopolymers are green, the non-degradable synthetic polymers are blue, and the degradable synthetic polymers are red. (b) shows three cluster analyses for the parameters vitality, apoptosis, and cytotoxicity.

is able to analyse all necessary standard parameters for the assessment of cell–biomaterial combinations, and can therefore suggest or advise for and against specific stem cell–biomaterial combinations for tissue-engineering applications.

5. Conclusion This study demonstrates that the assessment of stem cell–biomaterial combinations for cell-based therapies is multifactorial. Besides material and cell characteristics,

ARTICLE IN PRESS S. Neuss et al. / Biomaterials 29 (2008) 302–313

cell–biomaterial interactions have to be investigated. For this, the consideration of one or two parameters only is definitely not sufficient as basic evaluation. Because of cell type specific and species-specific differences, factors, such as cell viability and proliferation have to be investigated for each single combination of cells and biomaterials. In conclusion, stem cell-based therapies using biomaterial scaffolds strictly require a stringent assessment of parameters such as material topography, cell adhesion, morphology, viability, proliferation, cytotoxicity, and apoptosis. All these parameters have to be analysed and matched into a basic assessment. We now can suggest the following combinations of stem cells and polymers for tissue-engineering applications, where cell adhesion and proliferation are supported, while apoptosis and necrosis are inhibited: human DPSC on PDLLA (R203S); human EPC on PVDF, PTFE, Texin, PLLA (L209S), Resomer LR705, PEO10-b-PDLLA25, and collagen; human preadipocytes on texin, mouse MSC on PVDF and texin; and mouse ES cells on fibrin. Further, we can advice against the following combinations resulting in only weak cell adherence, but high apoptotic and necrotic signals: human DPSC on PET and hyaluronic acid; human preadipocytes on resomer LT706; mouse MSC on PEO10-b-PCL60 and PEA-C, mouse ES cells on BAK; and mouse HSC on PET, Texin, PDMS, PLLA (L209S), PDLLA (R203S), PCL, PEA-C, fibrin, and collagen. Acknowledgements Thanks to Manfred Bovi (Electron Microscopic Facility, RWTH Aachen University) for SEM. We thank Dr. Andras Nagy, Reka Nagy, Dr. Janet Rossant, and Dr. Wanda Abramow-Newerly for providing the murine embryonic stem cell line R1. We thank the Department of Orthopedic Surgery (RWTH Aachen University) for providing bone spongiosa. We thank Dr. Suwelack Skin & Health Care AG (Billerbeck, Germany) for providing collagen samples. This work was supported by a grant from the Interdisciplinary Centre for Clinical Research ‘‘BIOMAT.’’ within the Faculty of Medicine at the RWTH Aachen University (VVB-110). References [1] Langer R, Tirrell DA. Designing materials for biology and medicine. Nature 2004;428(6982):487–92. [2] Rosso F, Marino G, Giordano A, Barbarisi M, Parmeggiani D, Barbarisi A. Smart materials as scaffolds for tissue engineering. J Cell Physiol 2005;203(3):465–70. [3] Zhang S. Fabrication of novel biomaterials through molecular selfassembly. Nat Biotechnol 2003;21(10):1171–8. [4] Boontheekul T, Hill EE, Kong HJ, Mooney DJ. Regulating myoblast phenotype through controlled gel stiffness and degradation. Tissue Eng 2007;13(7):1431–42. [5] Anderson DG, Levenberg S, Langer R. Nanoliter-scale synthesis of arrayed biomaterials and application to human embryonic stem cells. Nat Biotechnol 2004;22(7):863–6. [6] Anderson DG, Putnam D, Lavik EB, Mahmood TA, Langer R. Biomaterial microarrays: rapid, microscale screening of polymer–cell interaction. Biomaterials 2005;26(23):4892–7.

313

[7] Flaim CJ, Chien S, Bhatia SN. An extracellular matrix microarray for probing cellular differentiation. Nat Methods 2005;2(2):119–25. [8] Chai C, Leong KW. Biomaterials approach to expand and direct differentiation of stem cells. Mol Ther 2007;15(3):467–80. [9] Wei J, Yoshinari M, Takemoto S, Hattori M, Kawada E, Liu B, et al. Adhesion of mouse fibroblasts on hexamethyldisiloxane surfaces with wide range of wettability. J Biomed Mater Res B Appl Biomater 2007;81B:66–75. [10] Saltzman WM. Cell interactions with polymers. In: Lanza RP, Langer RS, Vacanti J, editors. Principles of tissue engineering. 2nd ed. San Diego: Academic Press; 2000. p. 221–35. [11] Haynesworth S, Goshima J, Goldberg VM, Caplan AI. Characterization of cells with osteogenic potential from human marrow. Bone 1992;13(1):81–8. [12] Pittenger MF, Mackay AM, Beck SC, Jaiswal RK, Douglas R, Mosca JD, et al. Multilineage potential of adult human mesenchymal stem cells. Science 1999;284(5411):143–7. [13] Neuss S, Becher E, Woltje M, Tietze L, Jahnen-Dechent W. Functional expression of HGF and HGF receptor/c-met in adult human mesenchymal stem cells suggests a role in cell mobilization, tissue repair, and wound healing. Stem Cells 2004;22:405–14. [14] Neuss S, Stainforth R, Salber J, Schenck P, Bovi M, Knu¨chel R, et al. Long-term survival and bipotent terminal differentiation of human mesenchymal stem cells (hMSC) in combination with a commercially available three-dimensional collagen scaffold. Cell Transplant [in revision]. [15] Neuss S, Schneider R, Tietze L, Jahnen-Dechent W, Knu¨chel R. Secretion of fibrinolytic enzymes enables human mesenchymal stem cell invasion into wounded tissue. [in preparation]. [16] Coleman SR. Facial recontouring with lipostructure. Clin Plast Surg 1997;24(2):347–67. [17] Gronthos S, Mankani M, Brahim J, Robey PG, Shi S. Postnatal human dental pulp stem cells (DPSCs) in vitro and in vivo. PNAS 2000;97:13625–30. [18] Nagy A, Rossant J, Nagy R, Abramow-Newerly W, Roder JC. Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. PNAS 1993;90:8424–8. [19] Hacker C, Kirsch RD, Ju X-S, Hieronymus T, Gust T C, Kuhl C, et al. Transcriptional profiling identifies Id2 function in dendritic cell development. Nat Immunol 2003;4:380–6. [20] Hieronymus T, Gust TC, Kirsch RD, Jorgas T, Blendinger G, Goncharenko M, et al. Progressive and controlled development of mouse dendritic cells from Flt3+CD11b+ progenitors in vitro. J. Immunol 2005;174:2552–62. [21] Sturn A, Quackenbush J, Trajanoski Z. Genesis: cluster analysis of microarray data. Bioinformatics 2002;18(1):207–8. [22] Kozarova A, Petrinac S, Ali A, Hudson JW. Array of informatics: applications in modern research. J Proteome Res 2006;5(5):1051–9. [23] Nakajima M, Ishimuro T, Kato K, Ko I-K, Hirata I, Arima Y, et al. Combinatorial protein display for the cell-based screening of biomaterials that direct neural stem cell differentiation. Biomaterials 2007;28:1048–60. [24] Kunzler TP, Drobek T, Schuler M, Spencer ND. Systematic study of osteoblast and fibroblast response to roughness by means of surface–morphology gradients. Biomaterials 2007;28:2175–82. [25] Itthichaisri C, Wiedmann-Al-Ahmad M, Huebner U, Al-Ahmad A, Schoen R, Schmelzeisen R, et al. Comparative in vitro study of the proliferation and growth of human osteoblast-like cells on various biomaterials. J Biomed Mater Res A 2007;82(4):777–87. [26] Es-Souni M, Es-Souni M, Fischer-Brandies H. Assessing the biocompatibility of NiTi shape memory alloys used for medical applications. Anal Bioanal Chem 2005;381:557–67. [27] Groll J, Reska A, Schulte P, Salber J, Offenha¨user A, Moeller M, et al. Surface grafting of PEO-based star shaped molecules for bioanalytical and biomedical applications. Macromol Biosci 2007;7(8):1010–23. [28] Hemmrich K, Salber J, Meersch M, Wiesemann U, Gries Th, Pallua N, et al. Three-Dimensional nonwoven scaffolds from a novel biodegradable poly(ester amide) for tissue engineering applications. JMSM [in press].