Micropatterns of cell adhesive proteins on a non-adhesive background have become an important tool in cell biology research (Ashton et al., 2007; Azioune et al., 2009; Falconnet et al., 2006; Fink et al., 2007; Weghuber et al., 2010). Control of cell shape and behavior through this micropattern technique by spatial immobilization of adhesive proteins on a surface has provided novel insights in several aspects of cell biology, such as tissue morphogenesis, cell growth and cell differentiation, and apoptosis (Chen et al., 1997; Kilian et al., 2010; Théry et al., 2005, 2006). Micropatterning of biomolecules on solid surfaces, such as glass coverslips or polystyrene cell culture flasks, has also helped address basic biological questions, like how cells interact with surfaces and how geometrical constraints affect functions (Geiger et al., 2009; Zhao et al., 2010). These studies have given rise to a broad spectrum of applications in areas such as tissue engineering, medical implants, or cell based drug screening assays (Ashton et al., 2007; Weghuber et al., 2010).
A variety of techniques and chemistries have been developed to create molecular micropatterns on surfaces (Azioune et al., 2009; Falconnet et al., 2004; Fink et al., 2007; Kane et al., 1999). In the initial stages of the development of micropatterning, the most prevalent method used in cell biology applications was microcontact printing, which relies on transferring the desired adhesive molecules to a surface by making use of a stamp with predefined features (Chen et al., 1997; Kane et al., 1999). Subsequently, the unmodified area is backfilled with non-adhesive molecules. However, this method suffers from a lack of pattern resolution and, most importantly, reproducibility. Recently, photolithographic methods have been developed that provide higher pattern transfer fidelity (Azioune et al., 2009; Falconnet et al., 2004; Fink et al., 2007). In the process defined by Falconnet et al. (2004), termed molecular-assembly patterning by lift-off (MAPL), a photoresist is used to define the geometry of the pattern's adhesive layers. The resist is then lifted off and the non-adhesive background backfilled. Azioune et al. (2009) reported recently a simple and rapid process to produce protein micropatterns on a substrate passivated by poly-L-lysine-grafted-polyethylene glycol (PLL-g-PEG). The method involves pattern transfer by UV exposure through a photomask. In this case, exposure of a selected area to UV irradiation results in a transformation of carboxylic groups in PLL-g-PEG, potentiating the strong adsorption of fibronectin. The method is easy and robust. However, alternative materials to PLL-g-PEG would be desirable to add flexibility to micropatterning techniques. In the present work, we propose such an alternative, poly(ethylene oxide)-block-poly(4-vinylpyridine) (PEO-b-P4VP). We describe a photolithographic method using the diblock copolymer, PEO-b-P4VP, to produce micropatterns of cell adhesive proteins on glass surfaces. PEO-b-P4VP has been used before as a coating reagent to avoid wall–protein interactions in capillary electrophoresis for separation of proteins (Liu et al., 2008). The polycationic PEO-grafted copolymer was adsorbed rapidly and strongly through electrostatic interactions to the negatively charged glass surface.
The block copolymer utilized in this study was synthesized by atom transfer radical polymerization (ATRP) using a brominated PEO macromer (Liu et al., 2008). The synthesized diblock copolymer has a PEO block of 2,000 Da average number molar mass, and a P4VP block of approx. 1,500 Da average number molar mass. The polydispersity index of the P4VP block is approximately 1.2.
In the patterning process developed in this work and shown in Figure 1, PEO-b-P4VP is first spin-coated onto a glass coverslip substrate. Then a photomask with the desired features is placed on the PEO-b-P4VP and the masked coverslips exposed to UV light (185 nm) for 30 s. This high energy UV irradiation produces radical centers on the P4VP chains in a plausible mechanism that involves the transformation of the carbon bearing the pyridine ring into a tertiary radical according to the equation: –CH2–CH(Py)–CH2– → –CH2–C°(Py)–CH2– (Grozea et al., 2009). Radical recombination gives rise to coupling between two neighboring macromolecules. The repetition of this mechanism on several monomer units of the same macromolecule leads to a crosslinked system. The crosslinking reactions causes the polymer to solidify and attach to the coverslip in the areas exposed. These will correspond later to the protein and cell repellent regions. The next step is to rinse the coverslip in water to remove the non-irradiated, non-crosslinked PEO-b-P4VP. Finally, a solution of fibronectin is spread onto the coverslips and incubated for 30 min. During this incubation, fibronectin attaches to the masked areas where the PEO-b-P4VP has been removed. After this incubation time, the coverslip is rinsed with PBS buffer and the micropattern substrates are then ready to use for the cell seeding. This protocol renders high contrast between the adhesive patterns and the non-adhesive PEO-b-P4VP background. The fabrication process is easy and straightforward and requires no specialized equipment. Furthermore, PEO-b-P4VP is insensitive to air or humidity, so the prepared patterned substrates can be stored at room temperature for months without loss of passivation capacity.
Fluorescence imaging of Cy3-tagged fibronectin was used to determine the quality of the patterning process. Figure 2 shows several substrates of FN-Cy3 micropatterns on a glass surface grafted with PEO-b-P4VP. The images show high resolution patterns without any apparent fibronectin molecules on the passivated PEO-b-P4VP regions. It can also be noted that the quality of the patterns did not vary substantially with general features, that is, shape and size of the patterns. The fluorescence intensity of the fibronectin patterns was measured along a straight line on the different substrates. The graphs of those tracings are shown in Figure 1 alone side the corresponding substrates. The intensity plots show a significant difference in fluorescence intensity between the passivated areas (black) and the FN-Cy3 patterned (bright), indicating a good pattern contrast. The smallest feature that we achieved using this technique is in the range of 2 µm. This limitation is given by the precision of the method and the resolution of the mask.
We found that the exposure time had a significant effect on the quality of the pattern. During experiments when UV exposure was lower than 30 s, the patterns were not formed and the non-crosslinked PEO-b-P4VP was removed from the surface during the rinsing step. Beyond 60 s of exposure time, the micropatterns were also not formed but in this case, a residue was visible under the microscope after the rinsing step. We took this residue to be molecular debris after breakage and degradation of the polymerized PEO-b-P4VP by high energy irradiation. Hence, we infer that an optimal UV exposure is required to crosslink the polymer on the exposed areas. Too little exposure yields inadequate crosslinking (and thus inadequate binding); excessive exposure degrades the polymer, which becomes soluble, degrading the patterns' resolution. As in other photocrosslinking processes, the exposure time is related to other parameters, such as power of the lamp and distance between the UV lamp and sample. All these parameters must be optimized to achieve the best possible pattern resolution and contrast.
The effectiveness of the patterned surfaces to confine cells was studied with three cell lines: S180, A549, and NBT2 cells. Cells were seeded in serum-containing medium onto the PEO-b-P4VP passivated fibronectin patterned substrates. Incubation periods were from 2 to 12 h. At the end, all three different cell lines seeding on the micropatterns were observed with a phase contrast microscope.
Typical short-term cell responses are shown in Figure 3. Our experiments indicate that cells are found primarily on the protein patterned regions regardless of the size of those regions. The cells are tightly constrained by the patterns and are not easily removed from them. By contrast, cells found on the passivated areas cells show a round shape and could be easily detached from the surface by gently rinsing. With incubation periods of 12 h, cells showed the same behavior and did not round up, detach, or escape from the pattern areas. Moreover, cells were able to divide on the pattern and show no sign of toxicity as measured by cell apoptosis.
These results underscore the efficient antifouling properties of PEO-b-P4VP. Furthermore, these antifouling properties were not affected by using a medium with a high content of fetal calf serum (FCS, 10%). Another benefit of using PEO-b-P4VP is that crosslinked PEO-b-P4VP attached to surfaces is highly stable for weeks at room temperature. Highly resistant to serum protein, the crosslinked PEO-b-P4VP can prevent cell attachment for extended time, allowing prolonged culture periods. This high stability is required for certain type of cell biology research, such as stem cell culture or studies of cell differentiation.
We have described here a process for creating cell micropatterns with PEO-b-P4VP that is practical, efficient, and reliable. The protocol is different from other reported methods in that the antifouling agent, PEO-b-P4VP is immobilized in situ on the glass surface as a crosslinked, polyvalent polymer rather than simply physisorbed to the surface. The microdomains are defined during the crosslinking reaction by irradiation through a mask with the desired pattern. A lateral resolution down to 2 µm and high contrast between adhesive and non-adhesive regions achieved is sufficient for most experiments in cell biology. The highly crosslinked nature of the copolymer and its great stability permit for prolonged periods of storage. In addition, crosslinked PEO-b-P4VP remains repellent for long incubation periods in medium with high concentration of serum.