ARTICLE pubs.acs.org/Langmuir
Cell Fluidics: Producing Cellular Streams on Micropatterned Synthetic Surfaces Maurizio Ventre,† Francesco Valle,‡ Michele Bianchi,‡ Fabio Biscarini,‡ and Paolo A. Netti*,†,§ †
Istituto Italiano di Tecnologia, Center for Advanced Biomaterials for Health Care @CRIB Istituto per i Materiali Nanostrutturati, Consiglio Nazionale delle Ricerche § Interdisciplinary Research Center on Biomaterials, University of Naples Federico II ‡
bS Supporting Information ABSTRACT: Patterning cell-adhesive molecules on material surfaces provides a powerful tool for controlling and guiding cell locomotion and migration. Here we report fast, reliable, easy to implement methods to fabricate large patterns of proteins on synthetic substrates to control the direction and speed of cells. Two common materials exhibiting very different protein adsorption capacities, namely, polystyrene and Teflon, were functionalized with micrometric stripes of laminin. The protein was noncovalently immobilized onto the surface by following either lithographically controlled wetting (LCW) or micromolding in capillaries (MIMIC). These techniques proved to be sufficiently mild so as not to interfere with the protein adhesion capability. Cells adhered onto the functionalized stripes and remained viable for more than 20 h. During this time frame, cells migrated along the lanes and the dynamics of motion was strongly affected by the substrate surface chemistry and culturing conditions. In particular, enhanced mismatches of cell adhesive properties obtained by the juxtaposition of bare and laminin-functionalized Teflon caused cells to move slowly and their movement to be highly confined. The patterning procedure was also effective at guiding migration on conventional cell culture dishes that were functionalized with laminin patterns, even in the presence of serum proteins, although to a lesser extent compared to that for Teflon. This work demonstrates the possibility of creating well-defined, long-range cellular streams on synthetic substrates by pursuing straightforward functionalizing techniques that can be implemented for a broad class of materials under conventional, long-time cell-culturing conditions. The procedure effectively confines cells to migrate along predefined patterns and can be implemented in different biomedical and biotechnological applications.
1. INTRODUCTION The ligand density and the type of adhesive molecules and their spatial arrangement may affect different aspects of cell behavior such as adhesion migration and differentiation. In particular, Chen et al.1 demonstrated that the regulation of the shape of primary cells by controlling the planar arrangement of adhesive islands directly influences cell growth and death, regardless of the ligands used to mediate the adhesion. Using a similar experimental setup, McBeath et al.2 reported that the shape of adhesive islands and consequently the shape of cells regulate the commitment of human mesenchymal stem cells. Therefore, the control of cell adhesion was sought with great interest because it has the potential to control and guide the biological processes that govern cell fate. For this reason, several papers have focused on developing technologies to fabricate synthetic substrates displaying adhesive spots with defined shapes and patterns. Among these, the microcontact printing (μCP) of self-assembled monolayers (SAMs) proved to be a versatile and effective method of creating adhesive patterns of predetermined shape and position with micrometric spatial resolution.3 Owing to these characteristics, it is widely used for r 2011 American Chemical Society
accurately positioning ligands on surfaces and controlling cell adhesion and behavior.4,5 Alternatively, direct printing and transfer were successfully implemented,6,7 but these involve drying or heating steps in the process that might not be compatible with many proteins. Other methods exploited diffusion chambers and microfluidic networks that enable the realization of patterned surfaces in the form of stripes and gradients of proteins despite using small quantities of chemicals.8,9 Patterns of improved spatial resolution were realized by means of AFM-based nanolithographic techniques.10,11 These studies were instrumental in enriching our knowledge of basic cell biology and cellmaterial interactions, and thanks to them, novel, robust tools such as DNA arrays and biochips were successfully created. However, these techniques often require complex, time-consuming chemical reactions and expensive equipment and might not be practical for patterning large areas. Received: October 22, 2011 Revised: November 23, 2011 Published: November 28, 2011 714
dx.doi.org/10.1021/la204144k | Langmuir 2012, 28, 714–721
Langmuir
ARTICLE
Table 1. Material Surfaces, Culturing Conditions, and Nomenclature of the Four Types of Experiments substrate
patterning method
suspension and culture medium
Laminin pattern on the polystyrene cell suspension dish
MIMIC
FBS supplemented
PS+
Laminin pattern on the polystyrene cell suspension dish
MIMIC
FBS deprived
PS-
Laminin pattern on the Teflon AF-coated glass
LCW
FBS supplemented
T-AF+
Laminin pattern on the Teflon AF-coated glass
LCW
FBS deprived
T-AF-
Concerning the studies on cellpattern interactions, a majority of these papers were mainly focused on developing methods and pattern features to gain tight control over cell positioning or to study the effect that cell shape has on cell fate. In comparison, only a few of them have investigated the effects of adhesive patterns on cell migration under conventional serumsupplemented culturing conditions in a systematic way. This is particularly relevant in vivo because most of the motile cells are sensitive to localized ligands and redirect their movement in response to the ligands' density and spatial arrangement. Adhesion patterns are physiologically present in vivo and are necessary for cell migration to occur. In particular, directional motility is an essential feature of tissue development, angiogenesis, and wound healing, which are those events that tissue engineering aims to recapitulate to generate functional tissues. Interestingly, Doyle et al.12 have recently reported that cell moving on micropatterned surfaces exhibit features that are strikingly closer to those observed in vivo as opposed to those on flat 2D surfaces. Therefore, the fabrication of stable, extended adhesion patterns might be useful for studying not only the dynamics of individual cell motion but also the formation of supracellular structures such as capillary and neuron networks in a systematic way. However, many existing cell-patterning techniques suffer from poor stability and are subject to cell-dependent and cellindependent remodeling13 and might not be suitable for characterize these aspects. Therefore, the design of novel platforms with tailor-made patterns of selected molecules that are stable under long-term and conventional cell cultures may pave the way to controlling cell migration, promoting regenerative events, and mimicking specific biological functions in vitro. Recently, Junkin and Wong 14 reported a very elegant method of patterning conventional cell culture dishes by plasma lithography. Their patterned surfaces were stable for long-time experiments; however, cell confinement is due to hydrophilic/hydrophobic surface mismatches rather than a predetermined deposition of specific adhesive proteins. In this work, we generated and characterized long-range streams of cells resembling those that might be experienced in microfluidic devices by encoding patterns of adhesive proteins on large synthetic surfaces. The patterning procedures are simple, generally applicable, and very effective even under conventional culturing conditions. The surfaces that were chosen, polystyrene and Teflon, display very different affinities toward protein transfer. This is due to the requirement of generating different local contrast in the cell adhesion propensity to understand how this may affect the cell fate. Yet the procedure that we developed strongly affects cell migration in terms of both the directional motion and velocity.
experiment name
unconventional lithographic methods were used, namely, lithographically controlled wetting (LCW) and micromolding in capillaries (MIMIC).15,16 Both techniques require an elastomeric stamp obtained by the replica molding of an appropriate master. The masters were always fabricated by contact photolithography with a mask aligner (MJB4, λ = 365 nm, Karl Suss, Garching, Germany) on negative photoresist AR P3210 (All Resist, Strausberg, Germany) films of variable thickness. The replica were obtained using poly(dimethylsiloxane) (PDMS, Sylgard 184, Dow Corning, Midland, MI) and its curing agent mixed in a 10:1 ratio. For the experiments reported, the patterned array features channels whose width ranges from 17 to 70 μm and are alternatively separated by 160 or 230 μm spacing. LCW was applied to the fabrication of patterns on Teflon-AF (DuPont, Wilmington, DE) substrates, and the whole procedure was reported in detail in a previous work.17 Briefly, the Teflon-AF substrates were realized by spin coating glass slides with a solution of Teflon-AF (16.5 mg/mL) in Fluorinert FC-40 (3M, Maplewood, MN). The films were approximately 200 nm thick. The pattern was then fabricated by placing a droplet of human laminin solution (Sigma-Aldrich, St. Louis, MO, 10 mg/mL in DMEM) on the Teflon-AF surface, and then the PDMS stamp was placed on top of it. Upon drying, the solution is pinned under the protrusions or the recesses of the stamp, and the solute is then deposited just under the corresponding regions. The surface tension of the PDMS stamp plays an important role in the process, and for the present application, an oxygen plasma treatment (O2 = 40 mTorr, 40 mW power) is required to achieving proper Teflon functionalization. MIMIC was used to fabricate the pattern on the polystyrene (PS) culture dishes (35 mm, Corning, Corning, NY). This technique is based on filling the channels formed by the recesses of an elastomeric stamp placed in contact with a surface with a solution. The capillary effects allow the solution to enter the channels, and then, upon drying, the solute is deposited in the regions corresponding to the channels. The same elastomeric stamp used for LCW on Teflon was employed for the MIMIC on the culture dishes, thus the geometrical parameters are the same as reported above. Also in this case the surface tension of PDMS is required to be changed by oxygen plasma treatment to obtain more hydrophilic channels. The functionalization was carried out with human laminin solution in DMEM, as described previously.
2.2. Contact Angle and Scanning Force Microscopy Measurements. The substrates prior to and upon patterning were characterized in terms of the contact angle with water. The measurements were performed using a Digidrop DS (GBX Instruments, Bourg de Peage, France) and Milli-Q water. Several measurements in different areas of each sample were collected to obtain statistical significance. Under some conditions, such as the laminin pattern over PS, the contact angle was so small that it could not be accurately determined. Scanning force microscopy images were acquired with a Smena microscope (NT-MDT, Moskow, Russia) operated in semicontact mode under ambient conditions. The cantilever employed was an NSG (NT-MDT, Moscow, Russia) with a nominal tip radius of curvature of 10 nm and a resonance frequency of between 90 and 230 kHz. Images were analyzed by using free Image Analysis software (NT-MDT, Moscow, Russia); the size and the height of the pattern features were measured by the section analysis tool.18
2. MATERIALS AND METHODS 2.1. Substrate Preparation. The substrates employed for the experiments were fabricated following a soft lithography approach. Two 715
dx.doi.org/10.1021/la204144k |Langmuir 2012, 28, 714–721
Langmuir
ARTICLE
2.3. Cell Culture. Adult bovine dermal fibroblasts were obtained from Coriell (Coriell Institute, Camden, NJ). Cell cultures were grown in minimum essential medium supplemented with 10% fetal bovine serum (ThermoFisher Scientific Hyclone, Waltham, MA), 1000 U/L penicillin, 100 mg/L streptomycin (Sigma, St. Louis, MO) and 2 nonessential amino acids (Invitrogen Gibco, Carlsbad, CA). Cell cultures were kept in a humidified atmosphere of 5% CO2 at 37 C, and the medium was replaced every 3 days. Confluent cultures were detached with trypsin/ EDTA, concentrated by centrifugation (1500 rpm for 5 min), and then resuspended in DMEM either with or without FBS, according to specific experimental requirements (Table 1). Drops of the cell suspension, containing 6000 cells, were placed on the patterned region of the substrates. These were incubated for 2 h under standard culturing condition. Subsequently, the medium was aspirated in order to remove nonadherent cells and replaced with 1 mL of fresh medium. 2.4. Cell Migration Experiments. Cell-cultured substrates were placed in a mini-incubator connected to an automated stage (Prior, Rockland, MA) on an Olympus IX 50 optical microscope (Olympus Co., Tokyo, Japan). Images of selected regions of the substrate were recorded with a CoolSnap Camera (Photometrics, Tucson, AZ) every 10 min for 16 h. The images were sequentially mounted, thus obtaining a 96-frame time-lapse video. The videos were then rotated in order to align the pattern with the horizontal axis. Cell trajectories (i.e., the sequence of cell centroid coordinates) were reconstructed from time-lapse videos using Metamorph software (Molecular Devices, Sunnyvale, CA) and subsequently filtered with a three data span moving-average filter with Matlab (MathWorks, Natick, MA) in order to reduce the fluctuations arising from the uncertainty in the cell centroid determination. Cells undergoing division or exhibiting extended contacts with other cells were not taken into account in the tracking process. The mean square displacement (MSD), root-mean-square speed (S), and persistence time (P) were chosen as relevant parameter in order to describe the macroscopic features of cell migration on the different substrates. These parameters were calculated according to the procedure reported by Dunn.19 Briefly, the mean square displacement Æd2æt of each cell and its components along and across the pattern, namely, Æd//2æt andÆd^2æt, are calculated according to the overlapping time intervals method.20 The operator Æ•æt denotes the average over time intervals. Subsequently, the speed and persistence time of each cell along and across the pattern, S//, P// and S^, P^, respectively, are estimated by a nonlinear fit to the equations Æd== 2 æt ¼ S== 2 P== ½t P== ð1 et=P== Þ Æd^ 2 æt ¼ S^ 2 P^ ½t P^ ð1 et=P^ Þ
Figure 1. Bright-field images of fibroblasts cultivated on patterned substrates. (A) PS+ substrate. Note cells between adjacent lanes of laminin (arrowheads) and cells in the reservoir zone (asterisks). (B) T-AF+ substrate. The bar is 100 μm. displacement vectors have a different length and orientation according to the time interval in which they are evaluated, Q is also a function of t. Chosen for a specific time interval, operator Æ•æ is the average of all possible displacements of different cells. The statistical significance among MSD, S, P, and J values was assessed by performing a nonparametric KruskalWallis analysis using Matlab. p values of
