Transcriptional profiling is essential in fundamental studies of pathogenesis. In particular, multiplexed analysis of gene expression enables the identification of cancer-specific expression signatures that correlate with clinical disease and which can be used for the prediction of tumor presence and disease progression. Microfabricated capillary electrophoresis (CE) devices feature reduced sample and reagent requirements, faster analysis times, and increased automation. These advantages make microdevices ideal analytical platforms for the quantitative monitoring of biomarkers, and their potential as point-of-care devices for facilitating cancer detection in the clinical setting is explored in this thesis.

First, an integrated microdevice capable of performing low-volume, rapid, and highly sensitive expression analysis was developed. To further the goal of quantitative measurements of transcript levels from a small amount of sample, an affinity capture gel approach was used to address the problem of inefficient CE sample injection which limits sensitivity. Photopolymerization protocols were developed to define a small plug of oligonucleotide-modified polyacrylamide gel inline with the CE channel in order to accomplish efficient capture and sample microinjection for quantitative analysis. This concentration and purification step also demonstrated increased detection sensitivity and improved separation resolution.

To perform multiplexed analysis of cancer genes, an optimized protocol for transcript analysis with aid of affinity capture was used in conjunction with the integrated CE microdevice. The expression of genes implicated in prostate cancer was assayed directly from cells by solution hybridization with complementary fluorescently-labeled detection probes, followed by affinity bead capture of mRNA-probe complexes. Released detection probes were then injected on-chip and captured on a photopolymerized capture gel for concentration prior to CE separation and detection. The ability of the assay and microdevice system to evaluate gene expression was demonstrated by the measurement of absolute transcript levels of ten genes, enabling the successful identification of distinct expression signatures for the human prostate cancer cell lines LNCaP, VCaP, 22Rv1, and PC-3 with high sensitivity.

Finally, a microdevice also including on-chip PCR amplification is presented for improving the sensitivity of detection to enable the analysis of clinical samples. The functionality of the proposed microsystem is further expanded by the integration of upstream sample processing steps on-chip. By performing several bioanalytical processes on a single microfabricated platform, high-sensitivity expression analysis from complex biological samples should be possible, all the while reducing cost and analysis time. Based on the technologies developed in this thesis, these fully integrated devices could be implemented as diagnostic tools and play a key role in the future of clinical detection.