Deciphering Cancer Metabolic Zonation: Studying the Vasculature-Dependent Lipid Content of Tumor Cells

Introduction

Understanding tumor heterogeneity is a major goal in cancer science. Metabolite composition offers a powerful tool for understanding gene function and regulatory processes involved in such heterogeneity, and metabolic reprogramming has profound effects on tumor progression. Metabolomics studies on cancer have thus far been performed primarily on whole organisms, organs, or cell lines, losing information about individual cell types within a tissue.

Here, we use a new protocol for the profiling of metabolite content in different cell populations within glioblastoma (GBM), according to their distance from vasculature, and study their metabolic content.

Material and Method

Tumor induction in NodScid mice by intracranial or subcutaneous injections of GFP labeled U87MG (human glioma) cells was followed by injection of dye into tail vein staining specifically perivascular cells. Cells were FACS sorted, and their metabolic content analyzed by our LC-MS metabolomics platform. Exact mass, isotope pattern and fragmentation pattern will be compared to metabolite libraries. We performed RNA-seq, and Enzymes of interest were pharmacologically inhibited and their relevance tested with functional assays. The functional assays we performed to test the effect of metabolite inhibited are migration assay with Transwell inserts, wound healing, and membrane fluidity assay using fluorescence recovery after photobleaching (FRAP).

Results and Discussion

Metabolic analysis demonstrated that the metabolome of the perivascular cells is distinctly separate from that of the distal cells. A significant and striking difference between the populations was found in the membrane phospholipid content. The membrane of perivascular cells was comprised of lipids with significantly more polyunsaturated and longer lipids. RNA-seq results showed no significant difference in expression of elongases or desaturases between the populations, while Phosphatidylethanolamine N-methyltransferase (PEMT) was significantly upregulated in perivascular cells. Pharmacological inhibition of PEMT resulted in attenuation of the kinetics of diffusion in membrane, suggesting lower membrane fluidity. PEMT inhibition significantly reduced the cells’ migration capacity and wound healing rate. In vivo, pharmacological inhibition of PEMT resulted in the expected reduction of the degree of unsaturation of PLs. Surprisingly, no effect was noted on tumor size.

Conclusions

Intra-tumor metabolic heterogeneity has been suggested to have important implications for cancer therapeutics, especially for hypoxic cells, highlighting the importance of a cell-specific methodology for the study of cancer metabolism. The suggested study hence makes important contributions to our understanding of the fundamental question of metabolic tumor heterogeneity, as well as provides means for a cell-based therapeutic approach.