Multiplexed Surrogate Analysis of Glycotransferase Activity in Whole Biospecimens

摘要

Dysregulated glycotransferase enzymes in cancer cells produce aberrant glycans—some of which can help facilitate metastases. Within a cell, individual glycotransferases promiscuously help construct dozens of unique glycan structures, making it difficult to comprehensively track their activity in biospecimens—especially where they are absent or inactive. Here we describe an approach to deconstruct glycans in whole biospecimens then analytically pool together resulting monosaccharide-and-linkage-specific degradation products (“glycan nodes”) that directly represent the activities of specific glycotransferases. To implement this concept a reproducible, relative quantitation-based glycan methylation analysis methodology was developed that simultaneously captures information from N−, O−, and lipid linked glycans and is compatible with whole biofluids and homogenized tissues; in total over 30 different glycan nodes are detectable per GC-MS run. Numerous non-liver organ cancers are known to induce production of abnormally glycosylated serum proteins. Thus following analytical validation in blood plasma the technique was applied to a cohort of 59 lung cancer patient plasma samples and age/gender/smoking-status-matched non-neoplastic controls from the Lung Cancer in Central and Eastern Europe Study to gauge the clinical utility of the approach towards detection of lung cancer. Ten smoking-independent glycan node ratios were found that detect lung cancer with individual ROC c-statistics ranging from 0.76–0.88. Two glycan nodes provided novel evidence for altered ST6Gal-I and GnT-IV glycotransferase activities in lung cancer patients. In summary, a conceptually novel approach to the analysis of glycans in unfractionated human biospecimens has been developed that, upon clinical validation for specific applications, may provide diagnostic and/or predictive information in glycan-altering diseases.