(219c) Informatics and Data Mining of Combinatorial Datasets for Cell-Material Interactions

In this talk we present recent innovations in integrating novel data mining strategies with experimental combinatorial screening of osteoblast (MC3T3-E1) response to micropatterned surfaces. Polymer surface-cell interactions are critical in tissue engineering, diagnostic materials, and other implantable biomaterials. In this study, we investigate the ability to use natural phase-separation in blends of biodegradable poly(D,L-lactide) and poly(caprolactone) as a micropatterning strategy to optimize osteoblast attachment and proliferation.

The complex nature of cells and polymer surface features, along with the statistical variance of cell behaviors, can be overwhelming to traditional experimental design methodology and global (summary statistic) data analysis methods. For this reason, we have previously reported a high-throughput screening system that integrates combinatorial surface pattern libraries, in-situ bioassays, and high-throughput data acquisition to identify qualitative cell-material interactions. In this talk, we describe the implementation of a database, novel cell-surface metrics, and data mining methods to allow quantitative analysis of the large data sets that results from such experiments.

We have developed a set of cell-based, localized metrics for describing correlations between material surface features and cell response assays. The separate effects of material surface lateral patterns, topography, and cell-cell interactions can be successfully decoupled. Application to osteoblast function on blends of biodegradable poly(D,L-lactide) and poly(caprolactone) suggest a "holder-shaper" model, in which cells tend to attach to specific surface patterns (holders) via focal adhesions, while avoiding other surface patterns (shapers). This has lead to the hypothesis that combinations of critical sizes of holders and shapers can optimize osteoblast attachment and proliferation.

To test this hypothesis, we designed a new type of combinatorial library for quantitative investigation of the effects of "holder" and "shaper" micropattern combinations on cell behaviors. Combinatorial libraries investigate different island sizes and island-to-island distances using soft lithography techniques. Effects of "holders" and "shapers" and their combinations on cell proliferation are analyzed with both traditional and individual-cell based methods.