Targeting the PHGDH-mTOR Metabolic Axis in Osteosarcoma

Altering cellular energy metabolism has been highlighted as one of the emerging hallmarks of cancer. The reprogramming of bioenergetic pathways towards enhanced glycolysis, rather than the mitochondrial oxidative phosphorylation indicative of normal cells, results in increased biomass production and is associated with the activation of various oncogenes. The increased or decreased expression of key metabolic enzymes has been identified as a potential family of biomarkers that could serve as the targets for novel metabolic-based therapies in cancer. The serine, glycine, and one-carbon (SGOC) metabolism pathway consists of a series of enzymes and metabolites that drive protein and lipid production, enhanced DNA synthesis, increased methylation capacity, and generate antioxidants that regulate redox homeostasis within the cell. The pathway includes serine and glycine biosynthesis, the folate cycle, the methionine cycle, and the trans-sulfuration pathways. The metabolite 3-phosphoglycerate (3PG) is synthesized from glucose through upper glycolysis, and is converted into serine through the de novo serine biosynthesis pathway. A methyl unit, consisting of “one carbon”, is removed from serine and shunted into the folate cycle, allowing the generation and salvage of metabolites that make up the folate cycle and the methionine cycle, as well as sulfur metabolism. As such, the metabolic enzymes associated with this pathways, including 3-phosphoglycerate dehydrogenase (PHGDH), pyruvate kinase M2 (PKM2), serine hydroxymethyltransferase 2 (SHMT2), and methylenetetrahydrofolate dehydrogenase 2 (MTHFD2), have been identified as being overexpressed in a variety of cancers. In particular, given that the de novo biosynthesis of serine contributes the critical one-carbon unit to the rest of the network, the rate-limiting enzyme of serine biosynthesis, PHGDH, has been identified as being overexpressed in a number of cancers, including breast cancer, melanoma, colorectal cancer, and non-small cell lung cancer. The first research aim was to identify the role of serine and glycine biology in osteosarcoma, the most common bone cancer in children and young adults. PHGDH was identified as being overexpressed in osteosarcoma and correlated with significantly poorer relapse-free survival and overall survival in osteosarcoma patients. NCT-503 and PKUMDL-WQ-2101, small molecule inhibitors of PHGDH, were used to interrogate the reliance of osteosarcoma on the SGOC metabolic network. Inhibition of PHGDH resulted in decreased cellular proliferation, but did not cause cell death, suggesting that while osteosarcoma cells rely on PHGDH for sustained cellular growth and proliferation, the survival of the cells was not solely reliant on serine metabolism. Further analysis of osteosarcoma cells treated with PHGDH inhibitors identified blocked mitochondrial oxidative phosphorylation, causing full shunting of glucose through glycolysis and into lactate. Furthermore, an increase in neutral fatty acids was observed, suggesting an altered capacity for fatty acids as a fuel source. Gene expression analysis with PHGDH inhibition identified that the subsequent metabolite accumulation resulted in an activation of the mTORC1 signaling pathway, likely as a cellular attempt to supplement the nutrients no longer provided by the SGOC pathways. The second research aim focused on targeting the metabolic pathways induced by PHGDH inhibition. Previous studies in cancers with elevated PHGDH have utilized PHGDH inhibitors in conjunction with removing serine and glycine from the environment to cause cell death. Osteosarcoma cells, possibly due to the bone microenvironment, do not exhibit a modified response as a result of serine and glycine deprivation, further demonstrating that they are not solely reliant on serine and glycine biology for survival. Furthermore, translational metabolic therapies cannot be single-agent treatments, as cells maintain an efficient and dynamic capacity to rewire metabolic pathways. Rapamycin-based mTORC1 inhibitors have been used in osteosarcoma with limited success, and the combination of PHGDH inhibition with rapamycin treatment did not result in any enhanced cell death. However, perhexiline, a non-rapalog mTORC1 inhibitor, was synergistic with PHGDH inhibition and able to cause cell death in osteosarcoma cells. Using cell line-derived xenografts in mice, single the combination of NCT-503 with perhexiline was able to significantly reduce tumor growth. The final research aim sought to elucidate the mechanism of activity of non-rapalog mTOR inhibitors in osteosarcoma. Accumulation of SGOC metabolites including S-adenosylmethionine (SAM) and branched chain amino acids suggested activation of the nutrient sensing components of the mTOR signaling pathway, including SAMTOR and GATOR1. Fluorescent lysosome localization studies suggested that there was elevated GATOR1 localization at the mitochondria with perhexiline and NCT-503 treatment, resulting in inhibition of mTOR. To further specify the mechanism of non-rapalog based mTORC1 inhibition, an unbiased multiplex protein assay was used to identify components of PI3K/mTOR metabolism that were not affected by rapamycin treatment, but were affected by treatment with perhexiline or ALPI3MT55, another non-rapalog mTORC1 inhibitor. Glycogen synthase kinase 3α (GSK3α) and glycogen synthase kinase 3β (GSK3β) were found to be activated with perhexiline and ALPI3MT55 treatment, but unaffected by rapamycin. The two isoforms of GSK3 inhibit glycogen synthase, the enzyme responsible for generation of glycogen from glucose in the cell. PHGDH inhibition resulted in increased flux of glucose into glucose-1-phosphate, the precursor to glycogen, suggesting that glycogen synthesis was a potential cellular survival mechanism in response to PHGDH inhibition. The combination of PHGDH inhibition by NCT-503 and ALPI3MT55 treatment was synergistic in causing cell death in vitro and in vivo, suggesting that GSK3α/β may be the avenue by which osteosarcomas are sensitized to non-rapalog mTORC1 inhibition. Taken together, these aims gave a robust understanding of osteosarcoma biology and reliance on the SGOC metabolic network and identified a novel avenue of treatment that would be applicable in a high-risk subset of osteosarcoma patients. This work gives the preclinical justification for the combination of PHGDH inhibition and non-rapalog mTORC1 inhibition in osteosarcoma patients with high tumoral PHGDH expression