Comparative profiling by microchip electrophoresis of serum N-glycans from cancer patients

Glycosylation of various soluble and membrane-bound proteins is impacted by the onset and progression of cancer. Quantitative analysis of these changes in the glycome may provide valuable insight for clinical diagnosis and prognosis. We are developing methods to detect cancer and pre-malignant conditions by microchip electrophoresis and laser-induced fluorescence detection of N-glycans derived from human blood serum. We designed microfluidic devices with serpentine separation channels that are 22-cm long and have optimized turn geometries to minimize turn-induced band broadening. These devices separate glycans with efficiencies and reproducibilities comparable to capillary-based platforms, but have significantly reduced analysis times and sample consumption. We used this platform to analyze N-glycans derived from the serum of control individuals (age and sex matched, disease-free volunteers) and patients with Barrett's esophagus, high-grade dysplasia, and esophageal adenocarcinoma. Separations at 1250 V/cm yielded efficiencies up to 700,000 plates and resolved high- and low-abundance N-glycans and their structural isomers within 100 s. Statistical analysis determined quantitative differences among the four sample groups.;For ovarian cancer, we determined quantitative differences among the N-glycan profiles of control individuals and patients with late-stage recurrent cancer prior to and after an experimental drug treatment. Here, we identified several N-glycan peaks and peak combinations that distinguished the control and cancer samples with high accuracy. To improve glycan identification, we developed a strategy to neutralize the charges on sialylated N-glycans by covalent derivatization. Peaks from the electropherograms are assigned specific N-glycan structures through a correlation of migration time, molar mass, relative peak abundance, and isomeric composition. Because the neutralization reaction and fluorescent labeling are both quantitative, we are able to compare derivatized samples from ovarian cancer patients and control individuals and identify N-glycan structures that have quantitative differences among these groups.