Tumor lipid status and the responses to therapy in neuroblastoma : with emphasis on treatment monitoring by proton magnetic resonance spectroscopy

Abstract

In this thesis, the aberrant lipid and phospholipid metabolism of neuroblastoma was exploited for tumor monitoring and as target of experimental therapy. Neuroblastoma is the most common and deadly solid tumor of childhood. Since neuroblastoma cells are highly proliferative, aberrantly regulated metabolic pathways would be expected. Lipids and phospholipids were considered relevant to monitor and target, since these biomolecules are required for membrane synthesis and signal transduction. Human neuroblastoma xenografts in athymic rats were analyzed in vivo by proton magnetic resonance spectroscopy (1H-MRS), a clinically available method previously found useful for assessment of tumor (phospho)lipid metabolism.

A significant association between the in vivo content of choline-containing species (total choline), MRSdetectable lipids, and neuroblastoma tissue viability was demonstrated. The in vivo lipid/choline ratio was significantly inversely correlated with the viable tumor fraction. Histological response of neuroblastoma xenografts to the angiogenesis inhibitor TNP-470 was associated with an enhanced lipid/choline ratio.

Neuroblastoma growth arrest induced by serum starvation in vitro was associated with a decreased intensity of most 1H-MRS-detectable metabolites, in particular total choline. The mobile lipid/choline 1H-MRS ratio was validated in vitro and in rats in vivo as an accurate predictor of neuroblastoma chemotherapy response. The lipid/choline ratio increase was observed in drug sensitive xenografts in vivo within two to three days of treatment with irinotecan, predicting and preceding tumor regression. No corresponding metabolic alterations were detected in saline treated tumors or in multidrug resistant irinotecan-treated tumors.

Neuroblastoma tissue from patients was analyzed for the expression of cyclooxygenase-2 (COX-2), an enzyme that catalyzes prostaglandin formation from the n-6 polyunsaturated fatty acid arachidonic acid, abundant in cell membranes. COX-2 was expressed in neuroblastoma tumors and cell lines, but not in non-transformed childhood adrenal medulla. Treatment with COX-inhibiting non-steroidal antiinflammatory drugs (NSAIDs) induced apoptosis of neuroblastoma cells, a process associated with mitochondrial depolarization, activation of caspase-9 and caspase-3, and an enhanced 1H-MRS lipid/choline ratio. The NSAIDs diclofenac and celecoxib induced caspase-3 activation in vivo, and significantly inhibited neuroblastoma xenograft growth.

Supplementing cell lines with the polyunsaturated fatty acid docosahexaenoic acid (DHA, n-3), as opposed to the monounsaturated oleic acid (OA, n-9), induced neuroblastoma cell death by mechanisms involving enhanced susceptibility to oxidative stress. DHA, but not OA, caused mitochondrial membrane depolarization in a cyclosporine-sensitive mariner, and significantly enhanced neuroblastoma response to treatment with chemotherapy, NSAIDs, or a clinically relevant concentration of arsenic trioxide. Fibroblasts were not substantially affected by DHA. In conclusion, based on aberrant tumor lipid metabolism, MRS-based surrogate markers of neuroblastoma viability and treatment response can be monitored non-invasively in animal models. Targeting of neuroblastoma fatty acid homeostasis by NSAIDs or DHA is feasible in preclinical neuroblastoma models, and warrants clinical testing as potential novel modes of therapy with improved neuroblastoma selectivity.