Aspartate has a central role in cancer cell metabolism. Aspartate cytosolic availability is crucial for protein and nucleotide biosynthesis as well as for redox homeostasis. Since tumor cells display poor aspartate uptake from the external environment, most of the cellular pool of aspartate derives from mitochondrial catabolism of glutamine. At least four transporters are involved in this metabolic pathway: the glutamine (SLC1A5_var), the aspartate/glutamate (AGC), the aspartate/phosphate (uncoupling protein 2, UCP2), and the glutamate (GC) carriers, the last three belonging to the mitochondrial carrier family (MCF). The loss of one of these transporters causes a paucity of cytosolic aspartate and an arrest of cell proliferation in many different cancer types. The aim of this review is to clarify why different cancers have varying dependencies on metabolite transporters to support cytosolic glutamine-derived aspartate availability. Dissecting the precise metabolic routes that glutamine undergoes in specific tumor types is of upmost importance as it promises to unveil the best metabolic target for therapeutic intervention.

Uncoupling proteins (UCPs) form a distinct subfamily of the mitochondrial carrier family (MCF) SLC25. Four UCPs, DmUCP4A-C and DmUCP5, have been identified in Drosophila melanogaster on the basis of their sequence homology with mammalian UCP4 and UCP5. In a Parkinson’s disease model, DmUCP4A showed a protective role against mitochondrial dysfunction, by increasing mitochondrial membrane potential and ATP synthesis. To date, DmUCP4A is still an orphan of a biochemical function, although its possible involvement in mitochondrial uncoupling has been ruled out. Here, we show that DmUCP4A expressed in bacteria and reconstituted in phospholipid vesicles catalyzes a unidirectional transport of aspartate, which is saturable and inhibited by mercurials and other mitochondrial carrier inhibitors to various degrees. Swelling experiments carried out in yeast mitochondria have demonstrated that the unidirectional transport of aspartate catalyzed by DmUCP4 is not proton-coupled. The biochemical function of DmUCP4A has been further confirmed in a yeast cell model, in which growth has required an efflux of aspartate from mitochondria. Notably, DmUCP4A is the first UCP4 homolog from any species to be biochemically characterized. In Drosophila melanogaster, DmUCP4A could be involved in the transport of aspartate from mitochondria to the cytosol, in which it could be used for protein and nucleotide synthesis, as well as in the biosynthesis of ß-alanine and N-acetylaspartate, which play key roles in signal transmission in the central nervous system.

Glioblastoma (GBM), the most prevalent primary malignant brain tumor, remains challenging to treat due to extensive inter- and intra-tumor heterogeneity. This variability demands combination treatments to improve therapeutic outcomes. A significant obstacle in treating GBM is the expression of O⁶-methylguanine-DNA methyltransferase, a DNA repair enzyme that reduces the efficacy of the standard alkylating agent, temozolomide, in about 50% of patients. This underscores the need for novel, more targeted therapies. Our study investigates the metabolic–epigenetic impact of combining SN-38, a novel topoisomerase inhibitor inducing DNA double-strand breaks, with rabusertib, a checkpoint kinase 1 inhibitor. We identified this synergistic combination through high-throughput drug screening across a panel of GBM cell lines using a cancer drug library combined with SN-38. A secondary metabolic screening with the PEDS algorithm demonstrated a synergistic modulation of purine, one-carbon, and redox metabolism. Furthermore, the combined treatment led to the significant depletion of epigenetically relevant metabolites such as 5-methyl-cytosine, acetyl-lysine, and trimethyl-lysine. Reduced intermediates of the glutathione cycle indicated increased cellular stress following combinatorial treatment. Overall, the combination of SN-38 and rabusertib synergistically disrupts metabolites associated with epigenetic adaptations, leading to cytotoxicity independent of O⁶-methylguanine-DNA methyltransferase status, thereby underpinning this combination as a promising candidate for combinatorial therapy in GBM.

The human mitochondrial carrier family (MCF) consists of 53 members. Approximately one-fifth of them are still orphans of a function. Most mitochondrial transporters have been functionally characterized by reconstituting the bacterially expressed protein into liposomes and transport assays with radiolabeled compounds. The efficacy of this experimental approach is constrained to the commercial availability of the radiolabeled substrate to be used in the transport assays. A striking example is that of N-acetylglutamate (NAG), an essential regulator of the carbamoyl synthetase I activity and the entire urea cycle. Mammals cannot modulate mitochondrial NAG synthesis but can regulate the levels of NAG in the matrix by exporting it to the cytosol, where it is degraded. The mitochondrial NAG transporter is still unknown. Here, we report the generation of a yeast cell model suitable for identifying the putative mammalian mitochondrial NAG transporter. In yeast, the arginine biosynthesis starts in the mitochondria from NAG which is converted to ornithine that, once transported into cytosol, is metabolized to arginine. The deletion of ARG8 makes yeast cells unable to grow in the absence of arginine since they cannot synthetize ornithine but can still produce NAG. To make yeast cells dependent on a mitochondrial NAG exporter, we moved most of the yeast mitochondrial biosynthetic pathway to the cytosol by expressing four E. coli enzymes, argB-E, able to convert cytosolic NAG to ornithine. Although argB-E rescued the arginine auxotrophy of arg8∆ strain very poorly, the expression of the bacterial NAG synthase (argA), which would mimic the function of a putative NAG transporter increasing the cytosolic levels of NAG, fully rescued the growth defect of arg8∆ strain in the absence of arginine, demonstrating the potential suitability of the model generated.