Article Cite This: Langmuir XXXX, XXX, XXX−XXX
pubs.acs.org/Langmuir
A Versatile Protein and Cell Patterning Method Suitable for LongTerm Neural Cultures Serge Weydert,†,⊥ Sophie Girardin,†,⊥ Xinnan Cui,‡ Stefan Zürcher,§ Thomas Peter,† Ronny Wirz,∥ Olof Sterner,§ Flurin Stauffer,† Mathias J. Aebersold,† Stefanie Tanner,† Greta Thompson-Steckel,† Csaba Forro,́ † Samuele Tosatti,§ and Jań os Vörös*,† †
Laboratory of Biosensors and Bioelectronics, ETH Zurich, Gloriastrasse 35, 8092 Zurich, Switzerland Department of Chemical Engineering, Graduate School of Engineering, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 819-0395, Japan § SuSoS AG, Lagerstrasse 14, 8600 Dübendorf, Switzerland ∥ Bruker Optics GmbH, Industriestrasse 26, 8117 Fällanden, Switzerland
Downloaded via TULANE UNIV on February 18, 2019 at 02:30:15 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
‡
S Supporting Information *
ABSTRACT: Herein, we present an easy-to-use protein and cell patterning method relying solely on pipetting, rinsing steps and illumination with a desktop lamp, which does not require any expensive laboratory equipment, custom-built hardware or delicate chemistry. This method is based on the adhesion promoter poly(allylamine)-grafted perfluorophenyl azide, which allows UV-induced cross-linking with proteins and the antifouling molecule poly(vinylpyrrolidone). Versatility is demonstrated by creating patterns with two different proteins and a polysaccharide directly on plastic well plates and on glass slides, and by subsequently seeding primary neurons and C2C12 myoblasts on the patterns to form islands and mini-networks. Patterning characterization is done via immunohistochemistry, Congo red staining, ellipsometry, and infrared spectroscopy. Using a pragmatic setup, patterning contrasts down to 5 μm and statistically significant long-term stability superior to the gold standard poly(L-lysine)-grafted poly(ethylene glycol) could be obtained. This simple method can be used in any laboratory or even in classrooms and its outstanding stability is especially interesting for long-term cell experiments, e.g., for bottom-up neuroscience, where well-defined microislands and microcircuits of primary neurons are studied over weeks.
■
INTRODUCTION
Many surface modification techniques have been developed for patterning culture substrates with proteins, polymers, and adhesion promoting molecules. Surface coatings can be patterned with the help of photolithography,23,24 microcontact printing with poly(dimethylsiloxane) (PDMS)22,25−27 or polyolefin28 stamps, photobleaching-based cross-linking,29 masked plasma-etching,18 UV-based photoscission,30 laserinterference ablation,31 polymer brushes,32−34 polymer and biomolecule printing,20,35,36 and many similar methods.37−43 However, despite the wide range of available techniques, they all present important shortcomings. The first issue with the mentioned techniques is that they are neither very accessible to the general user nor very practical. For example, photolithography-based approaches usually require a cleanroom facility and are based on wafers instead of substrates suitable for cell culture, such as well
Surfaces coated with micrometer-scale geometrical patterns have been a tool for half a century to arrange and examine cultured cells.1−3 Coating patterns have, for example, been used to investigate the role of tissue geometry in mono and cocultures4−7 as well as the effect of cell−cell contacts on cellular functions such as tissue formation,8 differentiation,9 division,10 and survival or death.11 Morphological control over cell cultures is still an important tool today, especially for neuroscience, where the so-called bottom-up approach emerged as a new discipline in the last decade.12 Many researchers have concluded that to understand information processing in the brain, one eventually has to reengineer small functional components of it. Thus, living neurons, mainly cortical and hippocampal cells from rat embryos, have been arranged and cultured in patterns representing distinct network topologies and analyzed via activity-dependent calcium imaging or multielectrode arrays.13−22 © XXXX American Chemical Society
Received: November 5, 2018 Revised: February 1, 2019
A
DOI: 10.1021/acs.langmuir.8b03730 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir plates. Other laser-activation/ablation or inkjet printing-based methods require a custom-made hardware. Additionally, nonstraightforward chemical reactions and specialized lab equipment are often involved. Another shortcoming is the long-term stability of the different methods. Limiting cells to adhere and grow only on selected areas requires a stable adhesive pattern and a longlasting antifouling background. Stability is particularly important for neural cultures, where cells start to be electrically active only after a certain time in culture, e.g., roughly 2 weeks for embryonic rat neurons.44 However, over time, axons and dendrites growing from neurons tend to spread over imperfect nonfouling coatings.45 Two-dimensional (2D) patterns are thus usually invaded with neurites even before they become active, making it impossible to use the pattern for topologydependent analysis. However, apart from a few exceptions,21,46 long-term stability of the antifouling coating is seldom quantified in 2D neuronal cell patterning methods, that is, characterization of cell patterning is often only done after less than a week in vitro,14,16,18,24,26,33,47 whereas the studies that do discuss long-term stability usually only show representative images after a few weeks in vitro22,23,32,41,48 or quantify soma compliance to patterns, but not the extent of neurite overgrowth on the antifouling background.42 Recently, our laboratory demonstrated truly long-term stable patterns with neurons for the first time using microcontact printing and with the help of a novel poly(2-methyl-2-oxazoline) (PMOXA)based ultra-antifouling coating.13 Microcontact printing19,49−51 is a fairly easy process, but stamps have to be preproduced in the cleanroom and surface printing has to be done manually for every sample, thereby strongly limiting the throughput. Herein, we report a surface patterning method based on a perfluorophenyl azide (PFPA) adhesion promoter,52,53 which does not require complicated chemical synthesis, expensive or custom-made equipment or cleanroom facilities, and is stable in the long-term even for neural cultures. In this patterning protocol, PFPA is grafted on poly(allylamine) (PAAm) by mixing perfluorophenyl azide Nhydroxysuccinimidyl ester (PFPA-NHS) with PAAm. The amines in the PAAm are positively charged around physiological pH and will thus form a molecular monolayer on diverse surfaces, such as glass, polystyrene, silicon wafers, or titanium oxide, because these are negatively charged after chemical, plasma, or UV/ozone cleaning (Figure 1). The azide PFPA, when exposed to UV light in the 250 nm range, initiates addition-reactions with double bonds, insertion into C−H and N−H sites, or ring expansion to react with nucleophiles.54 This allows the PAAm-grafted-PFPA (PAAm-g-PFPA) to act as an easily applicable first adhesion promoting monolayer, which is able to bind proteins or organic polymers to the surface upon UV exposure. Well-defined patterns of proteins can be obtained by using masks transparent to 250 nm light, such as cyclo-olefin polymer (COP) foil masks.55 As a nonfouling background, 1.3 MDa poly(vinylpyrrolidone) (PVP) chains were bound to the adhesion promoter. PVP has been demonstrated to be more stable and better antifouling polymer in comparison to alternative polymers with the PFPA-system.56 As an example of the versatility of the system, patterns with proteins (laminin, fibronectin) and a polysaccharide (heparan sulfate) were created on glass and polystyrene, which are common culture surfaces, and primary hippocampal rat neurons as well as C2C12 mouse myoblasts were cultured
Figure 1. (A) Adhesion agent PAAm-g-PFPA adsorbed to a negatively charged substrate with a protein or organic polymer of interest in solution. (B) UV exposure initiates the local insertion of the azide N3 into the molecule of interest. (C) Poly(vinylpyrrolidone) (PVP) is a polymer with excellent nonfouling properties.
on the patterns. Laminin and PVP-patterned surface areas were characterized by measuring the accessibility of the protein via immunohistochemistry and PVP via Congo red staining. Furthermore, the amount of bound proteins and PVP on the patterns and the background were quantified via ellipsometry and qualitatively verified with infrared spectroscopy (IRspectroscopy). Long-term culture stability was demonstrated for primary hippocampal neurons and compared to a standard microcontact printing protocol using poly(L-lysine)-grafted poly(ethylene glycol) (PLL-g-PEG). PLL-g-PEG has been extensively studied, it is the most used antifouling polymer and is often considered to be the gold standard for simple and versatile surface modification.13,57−59 Overall, the presented method allows building stable and electrically active neural circuits with primary neurons, making long-term stable 2D cell patterning easy enough for a classroom experiment. The two most delicate steps of the process are the UV illumination, during which exposure of eyes and skin to the UV light must be avoided, and the cell culture, which must be kept sterile.
■
EXPERIMENTAL SECTION
Substrates. For cell culture experiments and surface accessibility characterization, PFPA-based protein coating was performed on the following substrates: 12- and 24-well polystyrene tissue culture plates (TPP Techno Plastic Products) with or without glass coverslips inside the wells (Menzel Gläser, borosilicate, 15 and 18 mm, CB00150RA1 and CB00180RA1, thickness 1). Well plates were blow-dried with nitrogen prior to use. Preparing the Adhesion Promoter. Before starting the coating process, the adhesion promoter PAAm-g-PFPA was freshly prepared through an N-hydroxysuccinimide (NHS) conjugation. Using PAAmg-PFPA prepared within two days of the start of the experiment gave consistent results and no protocol was developed for freezing or longterm storage of the promoter.
B
DOI: 10.1021/acs.langmuir.8b03730 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Figure 2. Complete scheme for the PFPA-based coating process. (A) Coating a plasma or UV ozone activated polystyrene well plate (alternatively, glass well plates or glass slides can be used) with the adhesion agent PAAm-g-PFPA. (B) Addition of a solution with the polymer or protein of interest. (C) Placement of the wax-printed foil masks and UV exposure with a desktop lamp equipped with a UV-C bulb. (D) Washing with SDSPBS. (E) Addition of PVP as antifouling background coating. (F) Second UV exposure without mask. (G) Washing with MeOH (compatible with well plates). (H) The surface pattern is ready and can be used for cell culturing. For a stock solution of 3 mg/mL PAAm-g-PFPA and a grafting ratio of roughly 1:4, 13 mg PAAm (Sigma-Aldrich, 283215) were dissolved in 2.6 mL ultrapure water together with 31.8 mg potassium carbonate (Sigma-Aldrich, 791776) to adjust the pH, heated up to the boiling point for better dissolution, and quickly left to cool back down. 11.2 mg of PFPA-NHS (SuSoS AG or Iris Biotech, N3-TFBANHS) was dissolved in 4.135 mL pure ethanol (EtOH) in a nontransparent beaker, quickly ultrasonicated (about 10 s), and slowly added to the PAAm solution on a magnetic stirrer. The mixture was left on the stirrer for at least 3 h to obtain a clear solution. Occasionally, precipitation was observed. In these cases, the mixture was discarded and the synthesis was repeated. Prior to use, the prepared PAAm-g-PFPA was diluted to 0.1 mg/mL in 2:3 HEPES1/ EtOH, leading to a slightly turbid colloidal solution. HEPES1 buffer consists of 10 mM HEPES in ultrapure water at pH 7.4. Coating Protocol. Safety note in the case of this protocol is used for a classroom experiment: UV-C light has to be handled with care. Exposing eyes or unprotected skin to UV-C light must be strictly avoided. We recommend using UV-C eye protection, gloves, and long sleeves when handling a UV-C lamp. The first part of the protocol consists of patterning the protein or polymer of interest (laminin, fibronectin, and heparan sulfate, SigmaAldrich L2020, F4759, and H7640). Substrates were activated and cleaned with oxygen plasma (100-E Plasma Systems, Technics Plasma GmbH, 2 min, 100 W, oxygen) and coated with the adhesion agent PAAm-g-PFPA for 30 min (Figure 2A). After rinsing the substrates with 2:3 HEPES1/EtOH as well as ultrapure water, a 50 μg/mL solution (in PBS, Thermo Fisher Scientific 10010023) with the protein or polymer of interest was added (Figure 2B) and left for 10 min. The substrates were then blow-dried with N2 and COP foil masks were placed on the sample surface. To avoid unwanted illumination, the samples were placed onto black nonreflective paper and positioned in a handmade darkroom box (cardboard with black paper on the walls) below a desktop lamp (MAULatlantic 82136) equipped with a UV-C light bulb (Philips TUV PL-S 11 W/2P, max
emission at 250 nm) and a 10° honeycomb disk grid for collimation (Jin-bei Photographic Equipment Co. Ltd). If well plates were used, precut stripes of nonreflective black paper were additionally added into the wells against the walls to block sideways reflection. The lamp was warmed up for 3 min and the samples were exposed to UV (Figure 2C) for 3 min at 2 cm distance of the lamp (0.4 mW/cm2). Afterwards the samples were washed (Figure 2D) with an SDS-PBS solution (PBS with 5 mg/mL SDS, Sigma-Aldrich, L3771) and adjusted to pH 10. The washing protocol consisted of 5 min ultrasonication (Cleanet Ultraschall 120, 70% power), at least 1 h incubation in fresh PBS−SDS solution, and again 5 min sonication in fresh solution, rinsing with ultrapure water, and blow drying with N2. The second part of the protocol consists of binding PVP to the areas where the PFPA is still unreacted, i.e., the ones which were covered by the mask in the first illumination step. A droplet of PVP (1.3 MDa, Sigma-Aldrich, 437190, 10 mg/mL in EtOH) was added to cover the whole sample (Figure 2E) and directly blow-dried with N2. A second illumination step of 3 min was performed without masks (Figure 2F). Unbound PVP was rinsed off with methanol (MeOH) for at least 1 h with changing to fresh solution every 15 min, followed by an additional 5 min sonication in fresh MeOH (Figure 2G). At the end, the samples were again rinsed in ultrapure water and blow-dried (Figure 2H). Until the second illumination was completed, the samples were kept in aluminum foil for light protection. Prior to cell seeding, all samples were sterilized with 70% ethanol for 30 min and rinsed with ultrapure water. Immunohistochemistry and Congo Red Staining. Prior to staining, patterned samples were blocked with 1% BSA (AlbuMAX, Gibco, Life Technologies, 1100-021, in PBS) for 45 min. Then, either 20 μg/mL of anti-laminin antibody (Sigma-Aldrich, L-9393, produced in rabbit, in 1% BSA in PBS) or 10 μg/mL Congo red (Sigma-Aldrich, C6277, BioXtra) in ultrapure water were added on top of the samples and left overnight at room temperature. All samples were rinsed with PBS. 20 μg/mL anti-rabbit Alexa Fluor 488 secondary antibodies C
DOI: 10.1021/acs.langmuir.8b03730 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Figure 3. Test patterns, masks and examples for surface contrast and cell culturing results. (A) Test patterns used in this study. (B) COP foil mask with test patterns printed using wax ink. (C) Laminin accessibility tested via immunohistochemistry with anti-laminin. (D) Examples with primary hippocampal rat neuron cultures at 2 weeks in vitro. (produced in goat, Invitrogen, Thermo Fisher Scientific, A11034) in PBS were added to samples stained with anti-laminin, which were then left for 2.5 h at room temperature and rinsed again with PBS. All samples were left in PBS for imaging. Fluorescence imaging was performed with an Olympus FluoView FV3000 confocal laser scanning microscope. For anti-laminin, measurements were referenced to a laminin photobleached sample and to a layer of pure laminin physisorbed on a plasma-treated polystyrene surface. For Congo red, a layer of laminin cross-linked to PAAm-g-PFPA and a layer of pure PAAm-g-PFPA-PVP on plasma-treated polystyrene were used as lower and upper references. COP Mask Printing. Masks were printed on UV transparent COP foil (Denz Bio-Medical, cyclic olefin polymer foil, 100 μm thick, A4, 1000023) with a standard wax printer (Xerox ColorCube 8570N, pure black, photo quality). Occasional warping of the foil caused by pressure and heat in the printer could be avoided by printing on 5 cm × 10 cm foil pieces scotch-taped onto paper instead of A4 size foils. The foil with printed patterns was then cut into pieces small enough to fit onto the desired substrate. Standard liquid inkjet printing onto the COP foil was also possible, but solid wax ink adhered better to the substrate. Laser printing is not adapted, because the COP foil overheats and buckles. Culturing of Primary Neurons. Cortical and hippocampal cells were dissociated from E18 rat embryo brain tissue (from dissected time-mated pregnant Wistar rats from Harlan Laboratories or alternatively from BrainBits LLC). Animal experiments were approved
by the Cantonal Veterinary Office Zurich. The dissociation was performed according to Weydert et al.13 Prior to starting the experiments, glass samples were immersed in serum-free B27/ Neurobasal medium (Thermo Fisher, Gibco, 21103049 and 17504001, 1:50 (v/v) B27 in Neurobasal), supplemented with 1:100 (v/v) GlutaMax (Thermo Fisher, Gibco, 35050061), and 1:100 (v/v) penicillin−streptomycin (Thermo Fisher, Gibco, 15140122). The dissociated neurons were seeded onto the samples at a density of 30 000 cells/cm2 and kept at 37 °C and 5% CO2 in a humidified incubator. Two-thirds of the cell medium was replaced every week with fresh medium. Culturing of C2C12 Myoblasts. Before the cells were seeded, substrates were immersed in Dulbecco’s modified Eagle’s medium (Thermo Fischer, Gibco, 41965) with 1:10 (v/v) FBS (Thermo Fischer, Gibco, 10270) and 1:100 (v/v) penicillin−streptomycin (Thermo Fischer, Gibco, 15140). C2C12 cells (Mouse myoblast C2C12, American type cell collection, ATCC CRL-1772) were harvested at passage 5 by trypsinization (Thermo Fisher, Gibco, R001100, trypsin/EDTA) and seeded at a density of 5000 cells/cm2. Substrates were washed with medium before imaging. Neurite Outgrowth Measurement. Test patterns were made via the PFPA-based patterning on polystyrene and glass substrates. For comparison, patterns with microcontact printing were produced on glass with poly-D-Lysine (PDL) as adhesive ink and PLL-g-PEG as an antifouling background, following the protocol by Weydert et al.13 Primary cortical neurons were seeded on the patterns and imaged D
DOI: 10.1021/acs.langmuir.8b03730 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Figure 4. Accessibility of laminin and PVP patterns obtained with PFPA-based coating on polystyrene well plates, measured with immunohistochemistry for laminin and with the Congo red method for PVP. For laminin, 100% refers to a fully covered layer of pure laminin electrostatically adhered to a plasma-treated polystyrene surface right after coating and 0% refers to the signal of the fluorescent secondary antibody after photobleaching. For PVP, 100% refers to a pure PAAm-PFPA-PVP layer and 0% refers to the Congo red signal from a layer of laminin crosslinked to PAAm-g-PFPA. (A) Staining average intensity profile curves for an anti-laminin and a Congo red stained squares, obtained with the PFPA-based coating protocol, with a wax-printed mask. The curves are the average of profile lines taken every 5 pixels across a 400 μm square. (B) Medians and quartiles for a test round with 3 samples patterned with square and Ψ-shaped networks, 10−18 images per sample. Patt.: intensity of pattern, Bckg.: intensity of background. Significance: ****
