Single-Cell Patterning and Adhesion on Chemically Engineered Poly

Feb 27, 2009 - We demonstrate a new approach to achieve single cell arrays using chemically modified poly(dimethylsiloxane) (PDMS) substrates...
8 downloads 0 Views 3MB Size
Download PDF
http://pubs.acs.org/langmuir © 2009 American Chemical Society

Single-Cell Patterning and Adhesion on Chemically Engineered Poly (dimethylsiloxane) Surface Kirsty Leong,† Anna K. Boardman,† Hong Ma,‡ and Alex K.-Y. Jen*,†,‡ †

Department of Chemistry, ‡Department of Materials Science and Engineering, University of Washington, Seattle, Washington 98195-2120 Received November 10, 2008. Revised Manuscript Received January 21, 2009

We demonstrate a new approach to achieve single cell arrays using chemically modified poly(dimethylsiloxane) (PDMS) substrates. Four different microwell geometries (ranging from 10 to 50 μm in diameter) and interstitial spacing (ranging from 30 to 250 μm) were fabricated using soft lithography. The surface of each microwell was sputtered with 25 nm of gold and functionally engineered with a self-assembled monolayer (SAM) of (10-mercaptomethyl-9-anthyl)(4-aldehydephenyl)acetylene (MMAAPA), a fused-ring aromatic thiolated molecule. Collagen was covalently bound to the SAM of MMAAPA using Schiff base chemistry. Cells were found to be attracted and adherent to the chemically modified microwells. By tuning the structural parameters, microwells with a diameter of 20 μm and interstitial spacing of 250 μm resulted in single cell arrays. By combining soft lithography and surface engineering, a simple methodology produced single cell arrays on biocompatible substrates.

Introduction Patterning cells into an array plays a significant role in fundamental cell biology, medical diagnostics, tissue engineering, cell-based biosensor development, and bioelectronics.1-5 Recently, patterning single cells has received more attention because of its ability to analyze and study the life cycle of cells where cell-to-cell differences would otherwise be lost in bulk measurements.6 To be able to control cells onto a contoured surface and to produce single cell arrays have proven to be more challenging, because cells exhibit complex behaviors and interactions with their surrounding environment.7 This dynamic interaction between cells and surfaces plays a significant role in cell proliferation, spreading, differentiation, and apoptosis.8 Therefore, a highly controlled and compatible surface chemistry and environment are vital to the survival and functionality between cells and organic as well as inorganic materials. A number of techniques have been used to examine cellsubstrate interactions and to spatially pattern cells. These techniques include elastomeric stamps, microfluidic channels, metallic surfaces, and microcontact printing, all of which may include peptide mediation, adsorption of extracellular matrix (ECM) proteins, and/or nonfouling molecules such as poly (ethylene glycol) (PEG).9,10 In metallic and elastomeric techniques, a thin sheet of material with pore patterned holes *Corresponding author. E-mail: [email protected]. Tel: (206) 543-2626. Fax: (206) 543-3100. (1) Veiseh, M.; Zhang, Y.; Hinkley, K.; Zhang, M. Biomed. Microdevices 2001, 3, 45–51. (2) Ziauddin, J.; Sabatini, D. M. Nature 2001, 411, 107–110. (3) Lee, K.-B.; Park, S.-J.; Mirkin, C. A.; Smith, J. C.; Mrksich, M. Science 2002, 295, 1702–1705. (4) Veiseh, M.; Zareie, M. H.; Zhang, M. Langmuir 2002, 18, 6671–6678. (5) Chen, C. S.; Alonso, J. L.; Ostuni, E.; Whitesides, G. M.; Ingber, D. E. Biochem. Biophys. Res. Commun. 2003, 307, 355–361. (6) Teruel, M. N.; Meyer, T. Science 2002, 295, 1910–1912. (7) Curtis, A.; Wilkinson, C. Biomaterials 1997, 18, 1573–1583. (8) Neff, J. A.; Tresco, P. A.; Caldwell, K. D. Biomaterials 1999, 20, 2377– 2393. (9) Prime, K. L.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115, 10714– 10721. (10) Harris, J. M.; Zalipsky, S. Poly(ethylene glycol): Chemistry and Biological Applications; American Chemical Society: Washington, DC, 1997.

Langmuir 2009, 25(8), 4615–4620

is placed in contact with the substrate, and the cells are seeded through the holes. This technique creates cellular micropatterns on a variety of substrates such as gels, polymers, and metals.11 Another approach uses microfluidic channels to flow proteins and cells through specific pathways. This technique reduces the potential of damage to proteins and cells and can mediate cells onto patterns of varying materials. Microcontact printing uses poly(dimethylsiloxane) (PDMS) to “print” site-specific molecules onto metal surfaces such as thiol-containing compounds, proteins, or nonfouling molecules. This is a cost-effective and simple approach capable of patterning cells on planar and nonplanar surfaces. Typical surface engineering strategies for cell patterning aim to control protein adsorption on surfaces via adhesion promoting or inhibiting molecules. One common approach to attract and control the attachment of cells onto a surface is by means of self-assembled monolayers (SAMs) to adhere extracellular matrix proteins. PEG blocks protein attachments, while protein attachment is directed by a recognition motif of the arginine-glycine-aspartic acid (RGD) peptide sequence found on the external part of the cell. SAMs are commonly used as two-dimensional chemical patterning platforms. Recently, they have been amplified into a three-dimensional functional structure.12 They have been engineered with functional groups by covalent chemistry to tailor specific surface properties (e.g., thiol-gold). These individual surface properties enable dynamic control of the interfacial activity of proteins and cells onto surfaces. SAMs of alkane thiols with carboxylic acid and aldehyde endgroups have been shown to be promoting molecules that covalently bond to ECM proteins.13 The chemical diversity of functional groups on a (11) Veiseh, M.; Wickes, B. T.; Castner, D. G.; Zhang, M. Biomaterials 2004, 25, 3315–3324. (12) (a) Rozkiewicz, D. I.; Ravoo, B. J.; Reinhoudt, D. N. Langmuir 2005, 21, 6337–6343. (b) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Chem. Rev. 2005, 105, 1103–1170. (c) Ulman, A. Chem. Rev. 1996, 96, 1533–1554.dUlman, A. An Introduction to Ultrathin Organic Films; Academic Press: Boston, 1991. (13) Nishizawa, M.; Takahasi, A.; Kaji, H.; Matsue, T. Chem. Lett. 2002, 31, 904–905.

Published on Web 2/27/2009

DOI: 10.1021/la8037318

4615

Article

Leong et al.

surface can direct the attachment of macromolecules,14 nanoparticles,15 proteins and DNA,16 or cells.17 The use of SAMs has helped to further understand the behavior of anchoragedependent cells on controlled cell-substrate contact,18-21 cell attachment and growth,22-24 and cell-cell interaction.25 ECM protein surrounds a cell’s dynamic microenvironment and can govern a cell’s functionality, orientation, and adhesion onto a surface.26 Materials derived from natural ECM, such as collagen, provide natural adhesive ligands that promote cell attachment through integrins.27 At low concentration, a protein can maximize interactions with the surface both by its orientation or by unfolding itself, leading to denaturation and irreversible adsorption of the protein at the surface. At high concentration, proteins undergo fewer interactions and, therefore, have a more stable conformation and desorb more easily. In addition, protein adsorption increases with increasing hydrophobicity of the surface. Two commonly used ECM proteins are fibronectin and collagen, both known to interact with a cell’s integrins. These proteins guide the adhesion of cells to site specific locations onto a surface. Because cells will adhere to most proteins on a surface, PEG is commonly used to block protein adsorption for site-specific surface manipulation. The nonfouling behavior of PEG is a convenient platform to create molecular recognition interfaces that undergo specific cell-surface interaction. This behavior is attributed to its conformation, which results in a high dipole moment of the ethylene oxide repeats. This causes extensive hydration of the grafted polymer chains and thereby sterically repels proteins that approach the surface. The high density and disorder of the (EG)n headgroups inhibit protein adsorption. Aside from PEG, mannitol, bovine serum albumin,28 and oligosaccharide have also shown to be protein-resistant.29 (14) (a) Pathak, S.; Singh, A. K.; McElhanon, J. R.; Dentinger, P. M. Langmuir 2004, 20, 6075–6079. (b) Schmelmer, U.; Jordan, R.; Geyer, W.; :: :: Eck, W.; Golzhauser, A.; Grunze, M.; Ulman, A. Angew. Chem., Int. Ed. 2003, 42, 559–563. (c) Li, H.; Kang, D.-J.; Blamire, M. G.; Huck, W. T. S. Nano Lett. 2002, 2, 347–349. (d) Ghosh, P.; Lackowski, W. M.; Crooks, R. Macromolecules 2001, 34, 1230–1236. (e) Pfohl, T.; Kim, J. H.; Yasa, M.; Miller, H. P.; Wong, G. C. L.; Bringezu, F.; Wen, Z.; Wilson, L.; Kim, M. W.; Li, Y.; Safinya, C. R. Langmuir 2001, 17, 5343–5351. (15) (a) Chakrabarti, R.; Klibanov, A. M. J. Am. Chem. Soc. 2003, 125, 12531–12540. (b) Gates, B.; Qin, D.; Xia, Y. Adv. Mater. 1999, 11, 466–469. (16) Kung, L. A.; Kam, L.; Hovis, J. S.; Boxer, S. G. Langmuir 2000, 16, 6773–6776. (17) (a) Yeo, W.-S.; Yousaf, M. N.; Mrksich, M. J. Am. Chem. Soc. 2003, :: 125, 14994–14995. (b) Vogt, A. K.; Lauer, L.; Knoll, W.; Offenhausser, A. Biotechnol. Prog. 2003, 19, 1562–1568. (c) Zhu, Y.; Gao, C.; Liu, X.; Shen, J. Biomacromolecules 2002, 3, 1312–1319. (18) Liu, V. A.; Jastromb, W. E.; Bhatia, S. N. J. Biomed. Mater. Res. 2002, 60, 126–134. (19) Singhvi, R.; Kumar, A.; Lopez, G. P.; Stephanopoulos, G. N.; Wang, D. I. C.; Whitesides, G. M.; Ingber, D. E. Science 1994, 264, 696–698. (20) Bearinger, J. P.; Castner, D. G.; Golledge, S. L.; Rezania, A.; Hubchak, S.; Healy, K. E. Langmuir 1997, 13, 5175–5183. (21) Chen, C. S.; Mrksich, M.; Huang, S.; Whitesides, G. M.; Ingber, D. E. Science 1997, 276, 1425–1428. (22) Stenger, D. A.; Georger, J. H.; Dulcey, C. S.; Hickman, J. J.; Rudolph, A. S.; Nielsen, T. B.; McCort, S. M.; Calvert, J. M. J. Am. Chem. Soc. 1992, 114, 8435–8442. (23) Saneinejad, S.; Shoichet, M. S. J. Biomed. Mater. Res. 1998, 42 13–19. (24) Folch, A.; Toner, M. Ann. Rev. Biomed. Eng. 2000, 2, 227–256. (25) Bhatia, S. N.; Balis, U. J.; Yarmush, M. L.; Toner, M. Biotechnol. Prog. 1998, 14, 378–387. (26) Di Carlo, D.; Lee, L. P. Anal. Chem. 2006, 78, 7918–7925. (27) Liu, W. F.; Chen, C. S. Mater. Today 2005, 8, 28–35. (28) (a) Nakanishi, K.; Sakiyama, T.; Imamura, K. J. Biosci. Bioeng 2001, 91, 233–244. (29) (a) Luk, Y. Y.; Kato, M.; Mrksich, M. Langmuir 2000, 16, 9604–9608. (b) Holland, N. B.; Qiu, Y. X.; Ruegsegger, M.; Marchant, R. E. Nature 1998, 392, 799–801.

4616

DOI: 10.1021/la8037318

The use of SAMs, ECM proteins, and PEG on a flat metallic surface presents a challenge when attempting to achieve single cell arrays. The utility of these techniques falls short because they immobilize clusters of cells rather than single cells in a given location and entail multistep processing. However, single cell patterning has been attempted in the past by Lee et al.30 and Rettig and Folch.31 Lee et al. patterned cells onto glass substrates using biotinylated cells and streptavidin interaction with PEG. Rettig and Folch used PDMS microwells without functionalizing the surface with inhibiting and/ or promoting molecules. They investigated the geometric parameters for single cell occupancy and the settling time of cells into the PDMS microwells. In this approach, the cells were capable of moving in and out of the wells if perturbed because the cells were not bonded to the surface. Comparatively, we functionalized the microwells to chemically bond the cells onto the surface. Here, we employ a simplified approach using soft lithography to create flexible micrometer-sized wells of specific diameter and spacing to attain single cell arrays and demonstrate its application in analyzing cell-to-cell heterogeneity. This is accomplished using surface chemical reactions to mediate cell-surface recognition. The well-packed MMAAPA SAM promotes a densely packed collagen layer, which directs the adhesion of single cells into a patterned array. The added advantage to our approach is the ease of locating and imaging the cells at a fixed location, and of injecting the cells into highthroughput systems for cellular analysis.

Experimental Section and Methods Materials. Acetone, acetic acid, and ammonia hydroxide (28.0-30.0% NH3) were purchased from Aldrich (St. Louis, MO) and used as received unless stated otherwise. Absolute (200 proof) ethanol was purchased from Aaper Alcohol and Chemical Company and used in preparation of thiol solutions. (10-mercaptomethyl-9-anthryl)-(4-aldehyde-phenyl)acetylene (MMAAPA) was designed and synthesized in our laboratory.32 Collagen type I from calf skin and collagenfluorescein bovine from calf skin were purchased from Sigma (St. Louis, MO). Dulbecco’s phosphate buffered saline (PBS) without calcium and magnesium was purchased from Invitrogen (Carlsbad, CA). Deionized water for rinsing was produced with a NANOpure Diamond purification unit (Barnstead International, Dubuque, IA) and had a resistivity of ∼15 MΩ cm-1. Mouse pituitary tumor cells (AtT-20) were donated by the Dovichi laboratory from the Chemistry Department at the University of Washington (Seattle, WA). Substrates. Boron-doped Si(100) wafers of test grade (diameter, 100 mm; resistivity, 1-10 Ω/cm; thickness, 525 ( 50 μm) were purchased from Silicon Sense (Nashua, NH). Polycrystalline gold substrates were prepared by electron-beam evaporation of gold (23 nm thick; 99.99%, Kurt J. Lesker Company) onto Si (100) wafers that had been primed with a layer of titanium (2 nm thick; 99.995%, Kurt J. Lesker Company) to promote adhesion between silicon and gold. A chrome mask with 10  250 μm2 and 20  250 μm2 positive features were fabricated at the University of Minnesota Nanofabrication Center (Minneapolis, MN). All other Si wafers were fabricated at the Nanotechnology User Facility from thin plastic films purchased from Fineline Imaging (Colorado Springs, CO). (30) Lee, Z.-W.; Lee, K.-B.; Hong, J.-H.; Kim, J.-H.; Choi, I. Chem. Lett. 2005, 34, 648–649. (31) Rettig, J. R.; Folch, A. Anal. Chem. 2005, 77, 5628–5634. (32) (a) Zin, M. T.; Yip, H.-L.; Wong, N.-Y.; Ma, H.; Jen, A. K.-Y. Langmuir 2006, 22, 6346–6351. (b) Zin, M. T.; Ma, H.; Sarikaya, M.; Jen, A. K.-Y. Small 2005, 1, 698–702.

Langmuir 2009, 25(8), 4615–4620

Leong et al. Patterned PDMS slabs were prepared from Sylgard 184 Silicone Elastomer Kit purchased from Dow Corning (Midland, MI). A Denton Vacuum Desk II (Moorestown, NJ) was used to sputter gold into the PDMS micron wells. Substrate Preparation. For single cell patterning, a PDMS mold was generated from a patterned silicon wafer master with positive features. A mixture (10:1 v/v ratio) of PDMS and crosslinking agent was degassed to prevent the formation of bubbles before pouring over the master and curing under ambient conditions in a polystyrene Petri dish. Once the 3 mm thick patterned slab of PDMS was produced, a thin film of gold (25 nm in thickness) was deposited on the PDMS mold. After 20 min, the gold layer was removed from the top surface of the PDMS stamp using a self-adhesive Scotch tape (3M, St. Paul, MN). For AFM and fluorescence imaging, gold substrates were diced into 1 cm  1 cm squares and O2 plasma-treated for 20 min prior to self-assembly of MMAAPA and surface chemical reaction with collagen. Self-Assembly of MMAAPA. Solutions of 0.05 mM MMAAPA were prepared in absolute (200 proof) ethanol. The patterned gold substrates were immersed into 2 mL MMAAPA solution, and deprotection of the thiol moieties by deacylation of thioacetyl groups using 2 μL of NH4OH (28.030.0% NH3) permitted the formation of SAMs. After 1 h of self-assembly under nitrogen, the samples were removed from the solution, rinsed thoroughly in ethanol, and dried under nitrogen. Assembly of Collagen Layer. A 100 μg/mL solution of collagen and collagen-fluorescence bovine was prepared in PBS buffer with 1 wt % acetic acid. A 200 μL solution of collagen covered the surface of the substrates for 1 h. Then, the substrates were rinsed with PBS buffer and DI water for 3 min each. Fluorescence spectroscopy was used to characterize the collagen layer. Cell Preparation and Single Cell Patterning. In a stable environment of 37 °C with 5% CO2, the AtT-20 cells were grown in Dulbecco’s modified Eagle medium supplemented with 10% horse serum and antibiotics. The synthesis of tetramethylrhodamine-GM1 was described elsewhere.33 In brief, the tetramethylrhodamine-GM1 was added to the medium, which contained defatted bovine serum albumin. AtT-20 cells were incubated with this medium for 50 h. After the incubation, the cells were detached using 1 Trypsin/EDTA and washed five times with phosphate-buffered saline (PBS). Following the incubation, the cells were fixed with 4% formaldehyde in PBS for 12 min. The formalin solution was removed from the cell suspension. PBS, supplemented with 10 mM glycine, was added to quench the reaction. The cell suspension was rinsed five times with fresh PBS and stored at 4 °C. Prior to the experiment, the cells were rinsed twice with PBS. Both live and fixed AtT-20 cells (approximate size of ∼15 μm) were trypsinized and seeded at a concentration of 1.0  106 cells/ mL onto protein-patterned surface for 24 h at 37 °C under humidified 5% CO2. Optical microscopy was used to image the pattern cell arrays. Then, the patterned cells were aspirated into the capillaries and electrophoresis began. The separations were performed using 40 cm long, 25 μm i.d., 148 μm o.d. uncoated fused-silica capillaries. The running buffer contained 35 mM sodium deoxycholate, 10 mM sodium tetraborate, and 5 mM methyl-β-cyclodextrin, and the running voltage was 16 kV. Details of the custom-designed and -built instrumentation are described elsewhere.34 (33) Larsson, E. A.; Hindsgaul, O.; Whitmore, C. D.; Martins, R.; Tettamanti, G.; Schnaar, R. L.; Dovichi, N. J.; Palcic, M. M.; Hindsgaul, O. Carbohydr. Res. 2007, 342, 482–489. (34) Boardman, A. K.; McQuaide, S. C.; Zhu, C.; Whitmore, C. D.; Lidstrom, M. E.; Dovichi, N. J. Anal. Chem. 2008, accepted/in press.

Langmuir 2009, 25(8), 4615–4620

Article

Characterization of Surfaces. Atomic Force Microscopy (AFM). Images from experiments performed on gold substrates were carried out in air under ambient conditions (ca. 40-50% relative humidity, 25 °C temperatures) on a Digital Instruments Nanoscope III Multimode AFM (Veeco Inc., Santa Barbara, CA) using tapping mode at a scan rate of 0.51.0 Hz. Silicon cantilevers with spring constants ranging from 12 to 103 N/m were used. Image resolution was 512  512 pixels. Cross-sectional analysis was performed using an algorithm contained in the AFM software. Optical Microscopy. Optical spectroscopy/microscopy was performed on an inverted microscope (Eclipse TE2000-U, Nikon) equipped with a true-color digital camera and a spectrometer (USB2000, Dunedin, FL). Unpolarized white light illumination by a metal halide lamp (EXFO X-Cite 120) was used for epifluorescence illumination. The fluorescence images were obtained through a combination of a 490 nm short pass exciter, a 500 nm dichroic, and a 520 nm band-pass emitter. The filter set, purchased from Chroma Technology Corp (Rockingham, VT), reduced background signal to less than 1% of the sample signal. These images were acquired from PDMS-gold-coated microwell substrates. ATR-FTIR. The surface-modified gold substrates were studied using attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR). Polarized FTIR spectra were obtained using a Bruker Tensor spectrometer (Ettlingen, Germany) equipped with a nitrogen-cooled Harrick GATR single angle reflection accessory (Ossining, NY). Each spectrum was run with a minimum of 1056 scans at resolution of 4 cm-1 while being purged with dry air in between each data collection series to eliminate water vapor from the sample compartment. Contact Angles. Water contact angle measurements were performed on gold substrates using a Rame-Hart 100 goniometer under ambient conditions on gold substrates. A drop of distilled water (2 μL) was applied to the surface, and the contact angle was determined within 10 s.

Results and Discussion Utilizing photolithography and surface engineering, we patterned cells into single arrays in the gold-coated microwells of PDMS substrates. The PDMS substrate possessed a multifunctional, nonfouling surface to inhibit cell adhesion, and a barrier for cell migration. Using soft lithography illustrated in Scheme 1, microscale circular and square wells of 10 μm, 20 μm, and 50 μm in diameter were designed to investigate a cell’s migration, topography, and alignment in obtaining a single cell in each microwell. Substrate size was characterized using an optical microscope (Figure 1). The surfaces of the microwells were functionalized to attach cells using a thiolated SAM with an aldehyde terminal end group and type I collagen. The step-by-step self-assembly process to guide and pattern single cell arrays is illustrated in Scheme 2. A well-packed and highly selective protein pattern for cellular adhesion was accomplished by modifying the gold regions with a SAM of an aromatic thiol containing molecule, (10-mercaptomethyl-9-anthyl)(4-aldehydephenyl)acetylene (MMAAPA). The PDMS substrates were immersed in a 0.05 mM ethanolic solution of MMAAPA. The aldehyde endgroup (-CHO) on MMAAPA provided reactive sites for the covalent attachment of collagen through Schiff base formation (-HCdN-) between the amine groups (-NH2) on collagen and aldehyde endgroups by a one-step reaction. Water contact angle and AFM roughness data (Table 1) were collected to confirm the SAM formation. SAMs of MMAAPA and collagen were characterized using AFM. For ease of data collection, layer-by-layer assemblies were DOI: 10.1021/la8037318

4617

Article

Leong et al.

Scheme 1. Fabrication of PDMS Substrates Modified with Gold Microwellsa

Figure 1. Optical images of PDMS substrates modified with gold microwells. The following features (a) 50 μm in diameter  50 μm in interspacing, (b) 20 μm in diameter  250 μm in interspacing, and (c) 10 μm in diameter  250 μm in interspacing were fabricated.

Scheme 2. Surface Engineering to Guide Cellular Adhesiona

a

PDMS is poured onto a Si wafer master, cured for 24 hrs, and then released. A PDMS replica of the master is fabricated with positive features, 3 mm in thickness and diced into 1 cm  1 cm squares. After sputtering 25 nm of gold, gold on the top surface was removed using selfadhesive tape leaving gold in the microwells.

performed on bare gold. SAMs of MMAAPA formed an ordered monolayer (Figure 2a). The strong π-π interaction of a fused aromatic thiol enhanced the film intergrity and reduced disorder. The aldehyde functional group was favorable for subsequent layer-by-layer assembly of collagen in order to enhance the attraction and adhesion of cells to the microwells (Figure 2c).32 By comparison of the number of cells located in a single microwell, PDMS substrates modified with a SAM of MMAAPA was found to contain more single cell arrays than PDMS substrates unmodified with MMAAPA. The SAM of MMAAPA potentially enhanced a more favorable collagen orientation for the attraction of cells to substrate surface and to promote cell adhesion compared to the absence of MMAAPA. Characterization of Immobilized Collagen by Fluorescence Microscopy. Because of the morphology and large molecular weight of collagen, confinement and control of the selfassembly of collagen onto the surface is critical to achieve single cell arrays. The success of the protein immobilization relies on the accessibility of its primary amine groups to the terminal aldehyde groups of the MMAAPA monolayer immobilized on the gold wells. Corresponding fluorescence images of MMAAPA and Schiff base surface chemical reaction of MMAAPA and collagen conjugated with 4618

DOI: 10.1021/la8037318

a MMAAPA is self-assembled onto gold microwells as a robust foundation to support the covalent attachment of collagen. Schiff base chemistry was employed in the surface chemical reaction to selectively modify the microscale features and to control the collagen layer. Last, cells are seeded onto the stamps for 24 hrs and rinsed, leaving adherent cells in the microwells.

fluorescein, shown in Figure 2b,d, indicates the successful biomolecular assemblies. Fluorescein, chosen for its commercial availability, excites at 420 nm and emits at 520 nm. Langmuir 2009, 25(8), 4615–4620

Leong et al.

Article Table 1. Surface Characterization by Water Contact Angles and AFMa water contact angle (deg) advancing

receding

surface roughness (nm) Ra

Rq

Polycrystalline gold (O2 plasma treated) 0 0 0.57 ( 0.12 0.60 ( 0.20 SAM of MMAPA 87 ( 2 65 ( 2 0.45 ( 0.15 0.48 ( 0.30 a The surfaces include polycrystalline gold substrate (subjected to O2 plasma treatment) and SAM of MMAPA. Ra = average roughness, the arithmetic average of the deviations from the center plane; Rq = root mean square roughness, the standard deviation of the Z value within a given area.

Figure 3. FTIR spectrum of molecular assemblies on gold showing the -CdO stretching frequency at ∼1705 cm-1, indicating SAM of MMAAPA aldehyde group, and -CdN- stretching frequency of ∼1640 cm-1, indicating the covalent attachment of collagen to SAM of MMAAPA.

Figure 2. (a) AFM image of MMAAPA SAM layer with (b) corresponding fluorescence image. (c) AFM image of the FITC labeled collagen layer after its attachment on to SAM and (d) its corresponding fluorescence image indicating the attachment of the collagen layer in a network fashion. The SAM of MMAAPA does not exhibit fluorescence. After the surface chemical reaction, green fluorescence from collagen conjugated with fluorescein occurred. Characterization of Immobilized Collagen by FTIR. FTIR verified the Schiff base surface chemical reaction between the aldehyde-terminated SAM from MMAAPA and SAM modified with collagen. This reaction produces an imine bond where there is a loss of a water molecule. Figure 3 shows the FTIR spectrum of the biomolecular assemblies on gold; the CdO stretching frequency of ∼1750 cm-1 indicating the aldehyde group and -CdN- stretching frequency of ∼1640 cm-1 indicating the formation of an imine bond and collagen attachment. Single Cell Patterning on Chemically Modified PDMS Substrates. The appropriate geometric parameters such as diameter and interspacing between microwells were initially investigated in order to obtain single cell arrays using AtT-20 cells. Microwells with 50  50 μm2 diameter and interspacing were first chosen to determine the area needed to attract cells into the microwells. The depth of the wells remained constant at 25 μm, slightly larger than the diameter of the cell (∼20 μm). Deeper wells would have made it difficult to inject cells into an instrument for cellular analysis and wells that were too shallow could possibly result in clusters of cells. Substrates were rinsed with copious amounts of PBS to ensure the removal of nonadherent cells and the adhesion of cells to the chemically modified microwells and then imaged. Multiple cells were found inside the 50  50 μm2 microwells because of larger volume compared Langmuir 2009, 25(8), 4615–4620

to the cell size (∼20 μm) (Figure 4a). The morphology of the cells also appeared to be less circular. Cell-to-cell communication and cell spreading are typical cellular behaviors that may be responsible for the morphology change. In addition, control experiments were performed to prove that cells were adherent to the chemically modified microwells. After rinsing, cells were not found in the unmodified PDMS microwells, the PMDS-gold microwells, or the PDMSgold microwells modified with MMAAPA. This further explained that the observation of cells found in the microwells is due to chemical modifications that guide and trap the cells. Microwells with smaller diameters and larger interspacings were fabricated to achieve single cell arrays because of multicell occupancy, cell-to-cell communication, and cell spreading found in the previous microwell geometries. Circular microwells with a diameter of 10 μm were fabricated with larger interspacing, 250 μm. The 250 μm interspacing was chosen on the basis of capillary electrophoresis (CE) instrument injection block configuration used for cellular analysis. After seeding the cells onto the substrates, clusters of cells formed around the perimeter of the microwell. After rinsing, 30% of mirowells were found to contain a single cell; the rest were empty (Figure 4c). Of the cells found in the microwells, they remained round in morphology. Thus, the majority of the diameters of the cells were larger than 10 μm, and larger microwell spacings eliminated the morphological changes associated with cell-to-cell communication and spreading. The diameter of the microwell was increased to 20 μm, while the interspacing was constant at 250 μm. Figure 4b showed that single cells were found to be located in each microwell with a yield of ∼83%. The remaining microwells did not contain cells. The cells located in the microwells were often found to be closer to the walls. One explanation could DOI: 10.1021/la8037318

4619

Article

Leong et al.

Figure 4. Optical images of pattern cell arrays in various microwell diameter and spacing. (a) 50  50 μm. Multiple cells are located in each microwell. (b) 20  250 μm. Generation of single cell array. (c) 10  250 μm. Cells are not located in the microwell due to cell size (∼15 μm) be that some collagen-SAM-modified gold was present on the sides of the microwells, which would make the cells be close to there preferentially. These single cell arrays will be demonstrated in a high-throughput system for cellular analysis. Application. Analyzing a single cell plays a crucial role in high-throughput cellular analysis systems, sensors, and tissue engineering.35 Here, we utilize these patterned single cell arrays in the analysis of gangliosides using capillary electrophoresis (CE) with laser-induced fluorescence detection.36 Gangliosides are sphingolipids that have one to four sialic acids attached to the head and have a long-chain fatty acid attached to the sphingosine to create a ceramide tail. The naming of these lipids has been described in a previous publication.36 Because gangliosides are located primarily on the outer membrane of neurons, they are involved in a range of extracellular interactions, and thus, several genetic disorders are associated with the disruption of the metabolism of these compounds. TaySachs disease is a prominent example where large amounts of GM2 accumulate. For single cell analysis, cells are first incubated with tetramethylrhodamine-labeled GM1, fixed in formaldehyde, seeded on the substrates, and aspirated into the CE instrument where the cells are lysed, releasing their lipid contents. Once separated, the labeled metabolites are detected using laser-induced fluorescence using a sheath flow cuvette. The metabolic profile of one aspirated cell is illustrated in Figure 5a with a standard mix solution of six gangliosides, illustrated in Figure 5b. The relative differences in intensities of the gangliosides between cells will allow for a better understand(35) Lidstrom, M. E.; Meldrum, D. R. Nat. Rev. Microbiol. 2003, 1, 158–164. (36) Whitmore, C. D.; Hindsgaul, O.; Palcic, M. M.; Schnaar, R. L.; Dovichi, N. J. Anal. Chem. 2007, 79, 5139–5142.

4620

DOI: 10.1021/la8037318

Figure 5. Electropheretic monitoring of the metabolic pathways of single AtT-20 cells with GM1 using capillary electrophoresis. Electropherograms of (a) a single AtT-20 cell and (b) a standard mixture of six gangliosides identified in the selected glycosphingolipid pathway. The peaks indicate that those labeled compounds were present within the cell. ing of the biochemical processes involved in the anabolism and catabolism of the glycosphingolipids.

Conclusions Single cell arrays adherent to a chemically modified PDMS surface were successfully achieved using 20 μm  250 μm microwells. This surface engineering and soft lithography strategy generated single cell arrays with high selectivity, reproducibility, and coverage. This technique combined surface molecular engineering with soft lithography to selectively guide, trap, and attach cells into a microwell without damaging the cells. PDMS served as a nonfouling surface to render small molecule, protein, and cellular adhesion into the microwells, thereby eliminating a nonfouling processing step. Gold deposited in the microwells mediate a thiolated SAM layer and sequentially, the attachment of collagen, an adhesion receptor ligand. This method can be widely applicable to biomedical applications, cell-based sensors, and bioelectronics. Acknowledgment. Financial support from the Microscale Life Sciences Center (an NIH Center of Excellence), and the Boeing-Johnson Foundation is acknowledged. The authors thank Dr. Norman Dovichi of the University of Washington for their kind donation of AtT-20 cells. The authors thank Guy G. Ting for help in obtaining ATR-FTIR spectra and University of Washington Engineered Biomaterials Center for the use of the ATR-FTIR.

Langmuir 2009, 25(8), 4615–4620