Cell Interactions with Biomaterials Gradients and Arrays

544 Combinatorial Chemistry & High Throughput Screening, 2009, 12, 544-553 1386-2073/09 $55.00+.00 © 2009 Bentham Science Publishers Ltd. Cell Interactions with Biomaterials Gradients and Arrays § Carl G. Simon Jr. *,1 , Yanyin Yang 1 , Vinoy Thomas 2,3 , Shauna M. Dorsey 1 and Abby W. Morgan 1,4 1 Polymers Division, National Institute of Standards & Technology (NIST), 100 Bureau Drive, Gaithersburg, MD 20899, USA 2 Ceramics Division, National Institute of Standards & Technology (NIST), 100 Bureau Drive, Gaithersburg, MD 20899, USA 3 Current Address: Center for Nanoscale Materials and Biointegration, University of Alabama at Birmingham (UAB), 1300 University Boulevard, Birmingham, AL 35294, USA 4 Current Address: Materials Science and Engineering, Chemical Engineering, Virginia Tech., 213 Holden Hall, Blacksburg, VA 24061, USA Abstract: Gradients and arrays have become very useful to the fields of tissue engineering and biomaterials. Both gradients and arrays make efficient platforms for screening cell response to biomaterials. Graded biomaterials also have functional applications and make useful substrates for fundamental studies of cell phenomena such as migration. This article will review the use of gradients and arrays in tissue engineering and biomaterials research, with a focus on cellular and biologic responses. Keywords: Biomaterials, cell adhesion, cell-material interactions, combinatorial screening, gradient, hydrogel, microarray, polymer, tissue engineering. 1. INTRODUCTION Despite significant investment in tissue engineering research, few profitable products have come to market [1, 2]. Hence, there is a need to accelerate tissue engineering research. One approach to accelerating research is combinatorial and high-throughput screening (CHT) (see Table 1 for all abbreviations used herein). Traditional research involves preparing samples one at a time for characterization and testing. In contrast, with CHT approaches, libraries are fabricated that combine many samples into miniaturized specimens. These libraries lower the cost of research by reducing the amount of time and material required for experiments [3]. Combinatorial approaches are utilized extensively for pharmaceutical research [4, 5] and their utility in biomaterials research is becoming apparent [6]. CHT methods for biomaterials research use two types of specimens: continuous gradients and discrete arrays (Figs. 1, 2). Gradients involve specimens that have continuously changing properties (composition, ligand density) along one or more of their axes. Arrays involve small, discrete specimens placed closely together on the same substrate (96- well plate, glass slide). There are advantages and disadvantages to both gradients and arrays (Table 2). Arrays can be easier to characterize but only include selected compositions. Plus, individual sample handling is required for each composition. Gradients require more character- *Address correspondence to this author at the Polymers Division, National Institute of Standards & Technology, 100 Bureau Drive, Gaithersburg, MD 20899-8545, USA; Tel.: (301) 975-8574; Fax: (301) 975-4977; E-mail: carl.simon@nist.gov §This Article, a Contribution of National Institute of Standards and Technology, is Not Subject to US Copyright Table 1. Table of Abbreviations 2D two-dimensional 3D three-dimensional bFGF basic fibroblast growth factor bisGMA 2,2-bis[4-(2-hydroxy-3- methacryloxypropoxy)phenyl] propane BMP-2 bone morphogenetic protein-2 CBFA1 core binding factor alpha 1 CHT combinatorial and high-throughput ECM extracellular matrix EGMP ethylene glycol methacrylate phosphate FGF-2 fibroblast growth factor-2 IGF-II insulin-like growth factor-II IKVAV isoleucine-lysine-valine-alanine-valine peptide NGF nerve growth factor NT-3 neurotrophin-3 PCL poly(�-caprolactone) PDLLA poly(D,L-lactic acid) pDTEc poly(desaminotyrosyl-tyrosine ethyl ester carbonate) pDTOc poly(desaminotyrosyl-tyrosine octyl ester carbonate) pHEMA poly(hydroxyethyl methacrylate) PLGA poly(lactic-co-glycolic acid) PLLA poly(L-lactic acid) SEM scanning electron microscope TEGDMA triethylene glycol dimethacrylate Biomaterial Gradients and Arrays Combinatorial Chemistry & High Throughput Screening, 2009, Vol. 12, No. 6 545 ization, but contain all possible compositions in a single specimen. Sample handling is significantly reduced for gradients since only one specimen is required. Arrays are amenable to quantitative analyses since individual assays can be performed for each discrete composition. Gradients, on the other hand, are amenable to more rapid, qualitative assessment since an entire composition range can be tested in a single specimen. Lastly, data analysis is usually easier for arrays than for gradients. It is also important to keep in mind that cells rarely interact directly with a biomaterial. Proteins present in blood in vivo or serum in vitro immediately adsorb onto most materials. Thus, cell response to a biomaterial is strongly influenced (possibly dominated) by the species, amount and conformation of proteins that adsorb onto a biomaterial [7]. Note that the current review focuses on biomaterials and that a large literature exists where gradients of factors have been fabricated and examined for their roles in development and cell migration. Though this work is partially covered here, these topics are beyond the scope of the current review [8- 13]. 2. TYPES OF GRADIENTS AND ARRAYS: CLASSIFICATION One way to classify gradients and arrays that have been used to examine cell response to biomaterials is by physical structure. The simplest structural delineation is two- dimensional (2D) versus three-dimensional (3D) (Figs. 1, 2). The majority of published work with biomaterials gradients arrays and cell screening has been done in a 2D format, i.e. films and surfaces (Table 3). For arrays, an array of biomaterial “spots” or “wells” in the form of films or surfaces with varying properties are presented to cells and cell function is assayed. For gradients, a substrate containing a biomaterial film, hydrogel or surface containing a gradient in properties is presented to the cells and cell response is observed. For 3D systems, cells are seeded within a hydrogel (not on the hydrogel) or onto a large-pore scaffold (> 0.1 mm dia. pores) such that the cells can explore the test biomaterials in three dimensions (Table 3). For arrays, an array of miniature scaffolds with varying properties would comprise the library. For gradients, a hydrogel or large-pore scaffold containing a continuous gradient in properties would comprise the library. 3. 2D FILMS & SURFACES 3.a. 2D Gradients: Surface Energy Surface energy is a fundamental material property that can influence cell behavior. Substrates with gradients of surface energy were some of the first gradients used to probe cell- material interactions [14-28]. Several methods for fabricating surface energy gradients employing a variety of surface chemistries have been demonstrated (reviewed in [19]). In addition, many cell functions have been studied during culture on surface energy gradients: adhesion, adhesion strength, morphology, spread area, migration, proliferation and differentiation. Collectively, the results from these studies do not always agree with one another. For example, cell adhesion is enhanced on different surface energies in different reports. Cell adhesion is equal on all of the water contact angles in Ruardy et al. [14], adhesion is enhanced on the hydrophobic regions in Ueda- Yukoshi & Matsuda [15] while adhesion was highest in hydrophilic regions in Chaudhury et al. [27]. Each study used a different cell type which suggests that cell adhesion can vary between cells. In addition, the surface chemistries used to make the gradients are different from each report. Differences in cell types and surface chemistry may explain the different observations made in these studies. One firm conclusion from these investigations is that surface energy can influence cell function. Fig. (1). Diagrams of (a) 2D gradient, (b) 2D array, (c) 3D scaffold gradient and (d) 3D scaffold array. 2D means the materials are presented as 2D flat films, surfaces or spots. 3D means the materials are presented as 3D scaffolds. 546 Combinatorial Chemistry & High Throughput Screening, 2009, Vol. 12, No. 6 Simon Jr. et al. 3.b. 2D Gradients Cell response to biomaterial gradients has been most frequently studied in “2D” systems where cell response to a gradient is assessed during culture on 2D films and surfaces [29-64]. Many innovative approaches have been employed to fabricate these gradients, and a variety of material parameters have been explored including gradients in i) proteins (cell membrane preparations [30]; basal lamina extracts [31]; SemA and SemC [32]; collagen I [41]; fibronectin [35, 38, 40, 43, 48, 49, 63]; laminin [31, 33, 41, 42, 46]; FGF-2 [40, 62] (Fig. 2b); ephrinA5 [50]; IGF-II [62]; BMP-2 [62]), ii) peptides (RGD from fibronectin [37, 47, 51]; IKVAV from laminin [36]; B160 peptide from laminin [56]; A10 peptide from laminin [53]), iii) polymer composition (PCL:PDLLA [34, 59]; PDLLA:PLLA [44]; PCL:PGLA [45, 64]; dental resins [58, 60]; pDTEc:pDTOc [61]) and iv) material processing (polymer crystallinity [39]; dental resin conversion [57, 58, 60]; roughness [52, 55, 56, 61]). Fig. (2). Examples of different types of gradients and arrays from the literature. (a) 2D Surface Array: Fluorescently-labeled primary chicken hepatocytes cultured on glycan arrays are visualized by fluorescence microscopy (used with permission [67]). Each column of spots is a different glycan, red indicates enhanced cell adhesion and each spot is 1.8 mm in diameter. (b) 2D Surface Gradient: A surface gradient of biotinylated-FGF-2 was fabricated using an inkjet printer and visualized with streptavidin-conjugated-quantum dots (used with permission [40]). The gradient is 1.75 mm by 1.25 mm. (c) 3D Scaffold Array: An array of salt-leached polymer scaffolds with varying composition fabricated in a 96-well plate (used with permission [92]). (d) 3D Scaffold Gradient: Rod-shaped, salt-leached, polymer scaffolds with a gradient in composition [92]. Six gradients 75 mm long, 8 mm wide and 4 mm deep are shown. Controls are shown on the sides. (e) 3D Pore Size Gradient: Polymer scaffold with a gradient in pore size created by a centrifugation method (used with permission [105]). Average pore size is given in the SEM images. (f) Gradients Exist In Vivo: Gradient in netrin-1 protein in the developing spinal chord of a day 9 mouse embryo visualized by immunohistochemistry (used with permission [115]). Size bar is 0.05 mm. Biomaterial Gradients and Arrays Combinatorial Chemistry & High Throughput Screening, 2009, Vol. 12, No. 6 547 In many studies, one material parameter is varied across the gradients [29, 31, 35-40, 42, 46, 47, 49, 51-57, 62, 63]. However, counter gradients have been fabricated where two components are varied in opposite directions from one another (cell membrane preparations [30]; SemA and SemC [32]; laminin/albumin [33]; PCL/PDLLA [34, 59]; laminin/ collagen I [41]; pHEMA/fibronectin [43, 48]; PLLA/PDLLA [44]; PLGA/PCL [45, 64]; bisGMA:TEGDMA [58]; BMP- 2/FGF-2 [62]). In addition, libraries have been fabricated where two parameters are varied orthogonally from one another: polymer composition and annealing temperature [34, 45, 59, 61, 64]; polymer composition and degree of polymerization [58, 60]. Also, the effect of the slope of biomaterial gradients on cell function has been explored in a number of studies [33, 46, 49, 54]. Thus, one parameter gradients, 2 parameter counter gradients, 2 parameter orthogonal gradients and gradients with different slopes have been used to examine cell behavior. In general, increased surface concentration of extracellular matrix proteins (ECM) proteins (laminin, fibronectin), adhesive peptides (RGD, IKVAV) or growth factors (FGF-2, BMP-2) in the gradients correlates with enhanced cell functions such as adhesion, spreading, proliferation, migration and differentiation [33, 35-38, 40-42, 47, 49, 51, 62]. In some cases, there is a threshold for surface density above which there is no increased effect on cell response [46, 51] or the cell response is inhibited [43, 48, 53, 54]. Several different observations were made when the slopes of protein or peptide gradients were varied: 1) cell functions were enhanced with increasing slope [49, 54], 2) cell functions were not affected by slope [32, 46], 3) there was a threshold for the slope below which the cells did not respond to the gradient and above which the cell response was enhanced by the gradient [33, 50] or 4) cell functions were enhanced with decreasing slope [30, 50]. Several cell studies with gradients of polymer composition or processing variables have been reported. For PLLA films annealed on a temperature gradient stage, the cool end stays smooth and amorphous while the hot end crystallizes and becomes rough [39]. Cell proliferation increased linearly with increasing PLLA crystallinity and roughness on these gradients. For composition gradients of PDLLA and PLLA, cell proliferation was enhanced only on the most PDLLA-rich regions [44]. Orthogonal gradients of polymer composition and annealing temperature induced polymer phase separation that yielded libraries with wide variations in domain sizes, roughness and composition. The cell culture results for these systems were especially interesting because unique composition/temperature combinations were observed to enhance cell functions [34, 45, 59, 61, 64]. These unique combinations may not have been discovered if discrete specimens had been used instead of the gradient approach. For micron-scale roughness gradients made by sandblasting, opposite effects were found for different cell types. Osteoblasts showed increased proliferation rate with increasing surface roughness while fibroblasts had decreased proliferation with increasing roughness [55]. For nano- particle gradients made by adsorbing silica nanoparticles to substrates, increased particle surface coverage inhibited cell proliferation [52, 56]. Finally, the viability of cells cultured on conversion gradients of a photopolymerizable dental resin was enhanced with increasing resin conversion [57, 58, 60]. Table 3. Biomaterial Properties Varied in Gradients and Arrays 2D Gradients Surface energy [14-28], cell filtrate or serum gradient [29], counter gradients of anterior and posterior tectal cell membranes [30], basal lamina protein gradients & merosin gradient [31], Sema3A & Sema3C gradient [32], albumin/laminin gradient [33], PCL/PDLLA composition/annealing gradient [34, 59], fibronectin gradient [4, 35, 38, 43, 48, 63], IKVAV laminin peptide gradient [36], RGD gradient [37], PLLA crystallinity gradient [39], FGF-2 gradient [40], laminin/coll I gradient [41], laminin gradient [42, 46], PDLLA/PLLA gradient [44], PLGA/PCL composition/annealing gradient [45, 64], RGD gradient [47, 51], ephrinA5 gradients [50], silica nanoparticle density gradient [52, 56], laminin B160 peptide gradient [53], laminin A10 peptide gradient [54], roughness gradient [55], dental resin bisGMA/TEGDMA composition/cure gradients [57, 58, 60], pDTEc/pDTOc composition/annealing gradients [61], immobilized BMP-2, IGF-II and FGF-2 gradients [62] 2D Arrays Acrylate-based polymer arrays [65], carbohydrate arrays [66, 67], polyarylate polymer array [68], polymer array [69, 72, 74, 77], ECM protein array [70, 71], growth factor array [73], polymer blend array [75], polyanhydride array [76] 2D Hydrogel Gradients : Cells on Gels modulus gradient [78], RGD gradient [79], NGF gradient [80], RGD gradient [81], bFGF gradient [82] 2D Hydrogel Arrays : Cells on Gels fibronectin/modulus array [83], polymer array [84] 3D Scaffold Gradients scaffold surface treatment gradient [90], covalently-linked protein gradient [91], polymer composition gradient [92], nanofibers with laminin gradient [93] 3D Scaffold Arrays polymer composition array [92], pDTEc/pDTOc array [94] 3D Hydrogel Scaffold Gradients: Cells in Gels laminin gradient [95], NGF and NT-3 concentration gradient [96], RGD gradient [97], laminin-1 and NGF gradient [116] 3D Hydrogel Scaffold Arrays: Cells in Gels (none to our knowledge) 3D Scaffold Pore Size/Porosity Gradients porosity gradient [101, 102], pore size/porosity gradient [103, 105], pore size gradient [104] 3D Scaffold Pore Size/Porosity Arrays (none to our knowledge) 548 Combinatorial Chemistry & High Throughput Screening, 2009, Vol. 12, No. 6 Simon Jr. et al. Thus, many creative approaches for screening cell response to biomaterials using films or surfaces containing gradients have been developed. 3.c. 2D Arrays “Two-dimensional” biomaterial arrays, where materials are presented to cells as discrete, flat films, surfaces or spots, have also been used to test cell response [65-77]. Several of these platforms used well-plates as a platform for biomaterial films [68, 75-77] while others used robotics to “spot” the biomaterials onto flat substrates (i.e. glass slides, Petri dishes) [65-67, 69, 70, 71, 73] (Fig. 2a). These arrays have covered several types of biomaterials including polymers [65, 68, 69, 72, 74-77], proteins [70, 71, 73] and carbohydrates [66, 67]. The studies with 2D polymer arrays have taken a number of innovative perspectives. Smith et al. [68] observed significant differences in fibroblast proliferation and macrophage expression of inflammatory cytokines during culture on arrays of degradable polyarylates. They used the large volume of data collected to successfully validate an artificial neural network algorithm for predicting cell response to untested materials. Anderson et al. [65] synthesized a polymer library of 576 different materials by photopolymerizing spotted blends of various acrylate monomers in situ. Most of the polymers supported human embryonic stem cell adhesion and differentiation into epithelia, however, several did not. In a follow-up, Anderson et al. [69] spotted blends of 24 different polymers into 1152 different ratios and combinations. Human mesenchymal stem cells adhered and spread on most blends although some combinations, especially those containing poly(ethylene glycol), inhibited attachment. Mant et al. [72] and Tourniaire et al. [74] synthesized a library of 120 different polyurethanes and spotted them onto agarose coated glass slides. The arrays were used to screen cell adhesion proliferation, and several hits were identified using either primary mouse bone marrow dendritic cells or primary human renal tubular epithelial cells. All of the “hits” contained 4,4’-methylenebis (phenylisocyanate) indicating that inclusion of this diisocyanate can enhance cell adhesion. Two companies have used biomaterial arrays for hemocompatibility screening. Cawse et al. [75] exposed a biomaterials library array to blood and found significant differences in leukocyte activation and platelet adhesion. In addition, Hezi-Yamit et al. [77] screened arrays of polymers designed for stents and found significant differences in monocyte adhesion and activation. Adler et al. [76] synthesized a library of polyanhydrides by using robotics to mix monomers in different ratios and found that several of the polymers were cytotoxic and induced an inflammato