High throughput lipid profiling for subtype classification of hepatocellular carcinoma cell lines and tumor tissues

Cell heterogeneity of tumor tissues is one of the causes of cancer recurrence after chemotherapy. Cell subtype identification in tumor tissues of specific cancer is critical for precision medicine and prognosis. As the structural and functional components of cells, lipids are closely related to the apparent morphology of cells. They are potential biomarkers of species of cancers and can be used to classify different cancer cell types, but it remains a challenge to establish a stable cell differentiation model and extend it to tumor tissue cell subtype differentiation. Here we describe a lipid profiling method based on nanostructure assisted laser desorption/ionization mass spectrometry (NALDI-MS), which could classify five hepatocellular carcinoma (HCC) cell lines and discriminate subtype of mixed cells and tumor tissues. The NALDI target was patterned with array of sample spots containing vertical silicon nanowires (Si NWs). Owing to its high ability to absorb laser energy, the vertical Si NWs can help to generate abundant lipid ions of cell extracts without need of organic matrix. Combined with statistical analysis methods, twenty-two ion peaks distributed in four MS peak clusters were selected as potential biomarkers to distinguish the subtype of the five HCC cell lines. Peak normalization was performed within each MS peak cluster to reduce the variation of peak intensity in batch to batch analysis. Compared to full-spectrum normalization method, the inner-cluster normalization method could help to distinguish cell subtype more stably and accurately. The molecular structure of these biomarkers was identified and sorted into two classes including phosphatidylcholine (PE, PI, PG, PA, PS) and glycosphingolipid (LacCer, ST). Furthermore, the established method was successfully applied to identify the major HCC cell subtype in mixed cell samples and xenograft tumor tissues as well as drug response test, showing great potential in precision medicine and prognosis.