Ergo, also the increased glycolysis, IDH2 diverts the new glycolytic intermediates so you’re able to serine and you will glycine synthesis, that is needed for the fresh new bioenergetic demand and you will precursors one to service cancer tumors gains
To predict the potential metabolic vulnerability of IDH2 we applied GSMM, incorporating the proteomic data of the IDH2-perturbed cell lines. Constraint-based modeling with the iMAT algorithm computed the most likely metabolic activity in the cells, and the minimization of metabolic adjustment (MOMA) algorithm was then used to predict the protein essentiality only in IDH2-high cells. The MOMA analysis predicted PHGDH, PSAT1, and S-formyl glutathione hydrolase (ESD) as essential proteins for IDH2-induced biomass production, thus forming SDL interactions with IDH2 (Fig. 4A). PHGDH is the first enzyme in the serine biosynthesis pathway; it diverts 3-phosphoglycerate (3-PG) from glycolytic pathway to 3-phospho-hydroxypyruvate (3-PHP) and PSAT1 transaminates 3-PHP to phosphoserine and subsequently to serine and glycine (Fig. 4B). ESD converts S-formyl-glutathione to glutathione and formate, which is further fed into the one-carbon metabolism pathway (Fig. 4B). Examination of the clinical significance of the SDL interactions using mRNA expression data from the METABRIC project showed better survival of patients with IDH2-high/PHGDH-low or PSAT1-low than patients with high PHGDH or PSAT (Fig. 4C). There were no significant differences related to PHGDH and PSAT, in the group of IDH2-low patients (Fig. 4D). Further supporting the MOMA predictions, we found a positive correlation between the expression of IDH2 and PHGDH and PSAT in breast tumor transcriptomics data (from the TCGA) and in the clinical proteomics data (Fig. 4E and F). In agreement, expression of PHGDH, PSAT1, and PSPH was higher in IDH2-high cells HCC1599, HCC1143 under basal conditions, and in addition, increased with IDH2 overexpression in HCC38 and reduced upon IDH2 knockout in HCC1143, yet it was inconsistent in the IDH2 knockdown in HCC1599 cells and in the transient overexpression systems (Supplementary Fig. S5A and S5B). In accordance, 13 Csix-glucose tracing showed that serine (m3) and glycine (m2) levels were higher in IDH2-high cells (Fig. 4G and H; Supplementary Fig. S5C and S5D). The fractional contribution from 13 C6-glucose to serine (m3) and glycine (m2) increased with IDH2 overexpression and reduced with IDH2 knockdown and knockout (Supplementary Fig. S5E). Serine (m3) levels also correlated with basal IDH2 in the different cell lines (Supplementary Fig. S5F).
As expected, loss of sometimes PHGDH or PSAT1 during the IDH2-higher muscle resulted in faster entry away from glutamine-derived carbons with the TCA stage intermediates (Fig
In cancer cells with high PHGDH expression, serine biosynthesis pathway has been shown to contribute 50% of the aKG through PSAT1 transamination reaction to the TCA cycle (Fig. 5A; ref. 7). Hence, we hypothesized that IDH2-high cells will be more sensitive to PHGDH or PSAT knockout, due to their higher metabolic activity. To that end we knocked out PHGDH and PSAT1 in HCC38 cells, and transiently overexpressed IDH2 (Supplementary Fig. S6A–S6C). Isotope tracing with 13 C6-glucose or 15 N2-glutamine showed reduced incorporation of 13 C or 15 N to serine (m3 or m1, respectively) and glycine (m2 or m1, respectively; Supplementary Fig. S6D-S6I). The fraction of m1 glutamate did not change significantly between the control and knockout cells, presumably due to the activity of alternative pathway (e.g., via GLUD1; Supplementary Fig. S6G-S6I). 5B–G; Supplementary Fig. S7A–S7F), resulting in decreased oxygen consumption in the control cells with IDH2 high (active PHGDH and PSAT1) compared with the IDH2 low (PHGDH and PSAT1 knockout; Fig. 5H and I; Supplementary Fig. S7G). Our analyses showed that 20%–30% of the glutamine to ?KG through PHGDH and PSAT1 contributed to enhanced TCA cycle activity and mitochondrial respiration in IDH2-high cells (Supplementary Fig. S7H).