Cell adhesion to extracellular matrices (ECM) is critical to differentiation, proliferation, migration, and apoptosis. Alterations in adhesive mechanisms are central to the behavior of cells in pathological conditions and aging, including cancer, atherosclerosis, and defects in wound healing. Cell adhesion is a significant consideration in biomedical and biotechnology applications including biomaterials, tissue engineering, and cell culture supports.

This research project focused on quantitatively analyzing the adhesive responses while systematically modulating the adhesive interface. The objective of this project was to analyze the role of nanoscale geometry of the adhesive interface in regulating integrin recruitment to adhesive contacts and modulating cell adhesion strengthening to ECM. Our central hypothesis was that the size and location of clusters of recruited integrin modulates cell adhesion strengthening in response to nanoscale organization of the adhesive interface.

Technical limitations have made it challenging to analyze the role of nanoscale spatial geometry in biological functionality. To overcome this limitation, we developed an experimental platform that provides control over the nanoscale geometry of the adhesive interface on samples that match the needs of cell biology studies. We first focused on developing a technique for producing high resolution patterns of proteins in biologically relevant geometries. To this end, the subtractive patterning technique was used to produce a pattern of proteins on a flat elastomer using a silicon nanotemplate which was then transferred to a final substrate by contact and release. Atomic force microscopy and fluorescence microscopy analysis demonstrated that this technique can produce patterns of antibodies with sizes as small as 90 nm with high contrast and high reproducibility. A wide range of pattern geometries were demonstrated by printing lines, linelets, and squares with spacing between features ranging from hundreds of nanometers to 64 µm. Patterns comprising two types of antibodies with intrinsic self-alignment were produced by the successive inking, subtraction, and printing of antibodies. Our results introduce a facile, high-throughput technique for patterning proteins on surfaces that enables the production of arrays of multiple types of proteins with high resolution, high contrast, and self-alignment in geometries that are relevant to cell adhesion studies.

In order to use the subtractive patterning technique for cell adhesion experiments, a robust immobilization strategy was required that maintained the original geometry of the protein patterns under extended cell culture conditions. The objective of our next study was to develop a method for producing cell adhesion arrays that constrain adhesion to nanoscale patterns of protein that are surrounded by a non-fouling background. To this end, we combined the subtractive patterning technique with mixed self-assembled monolayers. A mixed self-assembled monolayer was produced by assembling mixed carboxylic acid- and tri(ethylene glycol)- terminated alkanethiols into self-assembled monolayers on gold-coated substrates. The carboxylic acid-terminated alkanethiol component provided an anchoring point for immobilization of proteins into patterns. The tri(ethylene glycol)-terminated alkanethiol component provided a protein-resistant background when setup into a self-assembled monolayer. The subtractive patterning technique was used to produce complex patterns with multi-length scale dimensions. Following activation of COOH-end groups of the self-assembled monolayer, proteins transferred from the elastomer to the substrate during printing and tethered by coupling of protein primary amines to surface groups presenting NHS-esters.

Patterns of the cell adhesion protein fibronectin were produced to demonstrate the ability of the technique to immobilize proteins in controlled geometries while maintaining protein activity. Activity of the tethered FN was verified by binding of the FN-specific HFN7.1 monoclonal antibody which is receptor-mimetic. The background regions between FN-tethered regions remained devoid of antibody indicating that the non-fouling background effectively resists protein adsorption. Taken together, these results verify that protein activity is maintained during patterning and immobilization and that the non-patterned areas are resistant to protein adsorption.

Complex patterns of proteins with spacing and sizes varying across multiple length scales are desirable for studies of biological processes whose functionality requires coordination across the same scales. Previous experimental techniques typically achieve either micro- or nano-meter features but not both or are not able to maintain high-throughput or large sample areas that are needed for many biology experiments. In order to demonstrate the ability of our strategy to overcome these limitations, arrays of patterns of proteins were produced over large areas (~500 mm²) in geometries that include feature dimensions at both micro- and nanometer length scales. Features measuring as small as several hundred nanometers were simultaneously patterned and printed with micron feature sizes of 2 × 2 µm². These results demonstrate the ability of this technique to overcome previous technical limitations by producing patterns with dimensions across multiple length scales.

Covalent immobilization of proteins and a protein-resistant background of alkanethiols were used to ensure robust arrays of protein that could maintain a controlled adhesive interface during extended periods of cell culture. We verified that proteins were tethered to the carboxylic acid-terminated alkanethiol and that the mixed-SAM background would resist cell adhesion by plating cells on substrates with various alkanethiol treatments. These results confirm that protein tethering occurs through carboxylic acid-terminated alkanethiols that have been activated by NHS/EDC chemistry. Further, the non-adhesive character of mixed SAMs including >95% tri(ethylene glycol)-terminated alkanethiols was maintained during printing and was able to resist deposition of protein from solution and/or cells. Taken together, these results demonstrate that printing to mixed-SAMs is applicable to cell studies that require the ability to control the location of cell adhesion.

Our patterning technique was shown to be a useful approach to controlling cellular processes by using FN patterns to direct the formation of focal adhesions in adherent cells. Staining of the focal adhesion component vinculin in cells spread on non-patterned FN showed areas of high intensity at sites of vinculin localization, indicating the formation of elongated focal adhesions that are typical of spread cells. Vinculin staining of cells adhered to patterns consisting of eight squares with dimensions of 1 × 1 µm² showed constrained localization of focal adhesions to the patterned region. These results demonstrate that focal adhesion formation can be directed with high precision by modulating the geometry of the adhesive region. Combined, these results demonstrate that the combination of the subtractive patterning technique with mixed self-assembled monolayers produces robust cell adhesion arrays in which the geometry of the adhesion region can be used to direct cellular processes.

The objective of our next study was to analyze the recruitment of integrins into adhesive clusters in response to nanoscale geometry of the adhesive interface (adhesion area, spacing, and clustering) and determine the functional implications of integrin recruitment by quantifying variations in adhesion strength. Patterns of FN consisting of features with a range of nanoscale geometries (adhesion islands with dimensions of 1000, 500, 333, and 250 nm in clusters of 1, 2, 4, or 9) were produced using the subtractive patterning technique to directly immobilize proteins by covalent tethering onto surfaces presenting mixed self-assembled monolayers of alkanethiols. Cells seeded on the arrays were limited to one cell per pattern and adhesion was constrained to the adhesion region presented by the protein pattern. Spreading in between the patterns is prevented by the non-fouling background. Patterned substrates were shown to maintain the original pattern dimensions and resist FN deposition from cells using immunostaining with FN antibodies. These results demonstrate that the patterned arrays constrain cell adhesion to defined regions and that the adhesion patterns maintain their original design throughout the experiment.

Integrin recruitment was assessed using two metrics:​ 1) pad occupancy, which is defined as the number of adhesion pad locations that have integrin recruitment, and 2) integrin clustering characteristics, which includes the quantity of integrins that are recruited to a cluster, the localization of integrins within a cluster, and the area of the cluster. Three patterns were used with the same total area (12 µm²) but in different area splitting configurations of 1000 nm × 1, 500 nm × 4, and 333 nm × 9 (square island edge dimension in nm × number of islands in cluster). A smaller total area (6 µm²) was achieved with the pattern configuration of 250 nm × 4. Integrin clustering characteristics were analyzed by creating heat map images of α5 integrin recruitment by averaging individual images of integrin staining in cells on patterns. Results show that as the size of adhesion islands decreases, integrin recruitment is reduced until reaching a pattern size that is too small to support integrin clustering. These results establish a threshold size between 333 nm and 250 nm at which an adhesion area-dependent transition that occurs from adhesion areas that support formation of clusters with high levels of integrin binding to adhesion areas that result in low level integrin binding at low frequency. These results demonstrate that the recruitment of bound integrins into clusters is directed by the nanoscale geometry of the adhesive interface.

The second metric of integrin recruitment that we analyzed was pad occupancy which describes the extent of cell adhesion that is supported on different pattern geometries. Cells on 1000 nm × 1 patterns predominantly occupied three or more pads (out of eight total pads) with over half of the locations showing pad occupancy on all adhesive pads. In contrast, cells on 250 nm × 9 patterns showed low pad occupancy with over 90% of locations having two or less pads occupied. Patterns 500 nm × 4 and 333 nm × 9 generated pad occupancies that were equally distributed from partial to full occupancy. These results indicate a range of adhesion that occurs on patterns with nanoscale geometries. Larger adhesion patterns provide the highest level of pad occupancy and therefore a greater extent of adhesion. Decreased adhesion occurs with decreased pattern size until limited cell adhesion occurs at pattern sizes below the threshold for recruitment of integrin clusters. Interestingly, no difference in pad occupancy occurs between the 500 nm and 333 nm pattern. These results establish a relationship between geometry of available adhesion areas and the ability of cells to generate integrin clusters that are required for adhesion and spreading.

Contractile forces in adherent cells are known to be important in the formation and maintenance of adhesive structures. In order to determine the effect of contractile forces on adhesion to nanopatterns, integrin recruitment was analyzed in cells on 500 nm × 4 patterns after treatment with an inhibitor of Rho-kinase. The inhibitor Y-27632 has been shown to reduce contractility and focal adhesion assembly. Cells treated with inhibitor showed no difference in pad occupancy compared to control cells. However, cells treated with inhibitor exhibited a decrease in the level of integrin recruitment. These results indicate that cells with inhibited contractile forces are still able to recruit integrins to nanoscale adhesive contacts in a similar response as control cells. However, a decrease in the level of integrin recruitment occurs due to an inhibition of contractile forces.

Adhesion strength on nanoscale patterns was quantified in order to assess the functional dependence of cell adhesion on nanoscale geometry of the adhesion interface. Adhesion strength was analyzed as a function of total adhesion area, spacing between adhesion points, and size of individual adhesion points at nanoscale dimensions in order to uncover their roles in generation of adhesive force. Analysis of adhesion strength on various patterns uncovered several unexpected roles for nanoscale geometry in modulation of adhesion strength. The importance of nanoscale area was determined by results showing that adhesion strength decreased with a decrease in total pad area. Adhesion strength was also shown to depend on area splitting. When total pad area was kept constant but adhesive pads were broken down into multiple islands of smaller dimensions, adhesion strength decreased. This area splitting effect occurred at pattern dimensions of both 1000 nm and 500 nm indicating a range of sizes over which this effect can occur. In another set of experiments, no differences in adhesion strength occurred with changes to the space between adhesion islands. Further analysis determined a relationship between adhesion strength and the size of individual adhesion islands independent of the number of islands per pad. No difference in adhesion strength occurred on patterns of 500 nm × 4 and 500 nm × 1. Combined with results from integrin recruitment analysis, these results suggest that pad occupancy plays a dominant role in generation of adhesion strength and that a integrin clusters with sizes ranging between 0.25 µm² to 1 µm² can produce similar adhesion forces.

This thesis project has developed a unique experimental approach to analyze recruitment of bound integrins into clusters and quantify modulation of adhesion strength in response to systematic variation of the area, spacing, and clustering of adhesion areas. We determined that integrin recruitment is directed by changes in the size, clustering, and orientation of adhesion regions. We established a threshold pattern area between 333 × 333 nm² (0.11 µm²) and 250 × 250 nm² (0.06 µm²) below which integrin recruitment switches from robust integrin clusters to low frequency punctate formations. The role of area splitting in adhesion strengthening was established where adhesion strength changes despite no change to the total available adhesion area. A relationship was established between adhesion strength and area of individual adhesion islands. Patterns with adhesion areas below the threshold were unable to generate adhesion strength. Adhesion strength is seen to vary with integrin pad occupancy and not with the level of integrin clustering at adhesion regions. Furthermore, our results suggest that integrin clusters with areas between 0.25 µm² and 1 µm² generate equal adhesion strengths. As a whole, this project provides new insights on the role of size and location of clusters of recruited integrin in the modulation of adhesion strength in response to nanoscale geometry of the adhesive interface.