Metabolic effect of TAp63α: enhanced glycolysis and pentose phosphate pathway, resulting in increased antioxidant defense

Abstract

TAp63α is a member of the p53 family, which plays a central role in epithelial cancers. Recently, a role has emerged for p53 family members in cancer metabolic modulation. In order to assess whether TAp63α plays a role in cancer metabolism, we exploited p53-null osteosarcoma Tet-On Saos-2 cells, in which the expression of TAp63α was dependent on doxycycline supplementation to the medium. Metabolomics labeling experiments were performed by incubating the cells in 13C-glucose or 13C15N-glutamine-labeled culture media, as to monitor metabolic fluxes upon induced expression of TAp63α. Induced expression of TAp63α resulted in cell cycle arrest at the G1 phase. From a metabolic standpoint, expression of Tap63α promoted glycolysis and the pentose phosphate pathway, which was uncoupled from nucleotide biosynthesis, albeit prevented oxidative stress in the form of oxidized glutathione. Double 13C-glucose and 13C15N-glutamine metabolic labeling confirmed that induced expression of TAp63α corresponded to a decreased flux of pyruvate to the Krebs cycle and decreased utilization of glutamine for catabolic purposes in the TCA cycle. Results were not conclusive in relation to anabolic utilization of labeled glutamine, since it is unclear to what extent the observed minor TAp63α-dependent increases of glutamine-derived labeling in palmitate could be tied to increased rates of reductive carboxylation and de novo synthesis of fatty acids. Finally, bioinformatics elaborations highlighted a link between patient survival rates and the co-expression of p63 and rate limiting enzymes of the pentose phosphate pathway, G6PD and PGD.

Figure 1. Effect on glycolysis

(A). A simplified overview of glycolysis, pentose phosphate pathway and hexosamine pathway. Abbreviated metabolites and enzymes (Uniprot names) are reported in black and gray, respectively. (B) Isotopomer distribution of metabolites of glycolysis, graphed from left to right, from top to bottom in their order of appearance in the Embden-Meyerhoff pathway. Grouped columns in the left indicate TAp63α-expressing (doxycyline supplemented cells – Dox +), while columns in the right refer to non-induced controls (Dox -). Isotopomers are indicated as M+0 (only light unlabeled ¹³C atoms), M+6 (for hexoses deriving all six carbon atoms from heavy glucose) and M+3 (heavy trioses). Kinetics assays have been performed by harvesting cells at 1, 6, 12 and 24h, from medium replacement (which in turn was performed at 24h from doxycyline supplementation). * and # indicate statistical significance for inter-group (Dox + vs Dox -) or intra-group variations (time-course analyses). The number of * and # is related to p-values < 0.05; 0.01 or 0.001, respectively.

Figure 2. Effect on Pentose Phosphate Pathway

Isotopomer distribution of metabolites of the Pentose Phosphate Pathway, graphed from A to D in their order of appearance (oxidative and non-oxidative arms). Grouped columns in the left indicate TAp63α-expressing (doxycyline supplemented cells – Dox +), while columns in the right refer to non-induced controls (Dox -). Isotopomers are indicated as M+0 (light, no heavy ¹³C atoms incorporated), M+6 or M+5 (for hexoses and pentoses deriving all their carbon atoms from heavy glucose). Kinetics assays have been performed by harvesting cells at 1, 6, 12 and 24h from medium replacement (which in turn was performed at 24h from doxycyline supplementation). * and # indicate statistical significance for inter-group (Dox + vs Dox -) or intra-group variations (time-course analyses). The number of * and # is related to p-values < 0.05; 0.01 or 0.001, respectively.

Figure 3. Effect on GSH homeostasis

(A-F) Heavy and total isotopomers of reduced and oxidized glutathione (GSH and GSSG, respectively), as representative metabolites for oxidative stress accumulation. In A, heavy isotopomers are reported on the basis of the presence of heavy glucose derived ¹³C atoms, such as M+2, M+3 and M+4, M+5 and M+7 (sum of the previous). While glycine and cysteine labeling is expected to be low in the light of their relative abundance in comparison to the light M+0 isotopomers (Suppl. Fig. 4), heavy C atoms might be incorporated from glutamate derived via transamination of labeled ketoglutarate from acetyl-CoA produced via glycolytic consumption of labeled glucose (Suppl. Fig. 6A) (J-L). Isotopomer distribution of glutathione in glutamine labeling experiments. Grouped columns in the left indicate TAp63α-expressing (doxycyline supplemented cells – Dox +), while columns in the right refer to non-induced controls (Dox -). Isotopomers are distributed as follows: M+5+1 (fully labeled glutamate in both carbon and nitrogen atoms), M+5 (glutamate labeled only in carbon atoms). M+2 and M+3 (fast equilibrium with ketoglutarate; M+4 and 4+1 isotopomers were not detected). Kinetics assays have been performed by harvesting cells at 1, 6, 12 and 24h from medium replacement (which in turn was performed at 24h from doxycyline supplementation). From left to right, the graphs represent only heavy isotopomers (J), all isotopomers (K), and only heavy isotopomers in Dox + against Dox – cells at 1h after medium replacement (L). In L, differential results were statistically significant (p < 0.01 Student's t-test). Kinetics assays have been performed by harvesting cells at 1, 6, 12 and 24h from medium replacement (which in turn was performed at 24h from doxycyline supplementation). * and # indicate statistical significance for inter-group (Dox + vs Dox -) or intra-group variations (time-course analyses). The number of * and # is related to p-values < 0.05; 0.01 or 0.001, respectively.

Figure 4. Effect on Krebs cycle (heavy labeled isotopomers only)

(A) Distribution of heavy isotopomers of metabolites of the Krebs cycle from ¹³C-glucose labeling experiments, graphed from left to right, from top to bottom in their order of appearance in the oxidative reactions of the TCA cycle pathway. Isotopomers are indicated as M+5 (incorporating five ¹³C atoms) or M+10 (for nucleotides deriving one or two ribose moieties from heavy glucose), M+2, M+3, M+4, M+5, M+6 (progressively heavy labeled Krebs cycle intermediates from glucose-derived acetyl-CoA). (B). Distribution of heavy isotopomers of metabolites of the Krebs cycle from ¹³C¹⁵N-glutamine labeling experiments. Isotopomers are distributed as follows: M+4 (incorporating four ¹³C atoms, i.e. fully labeled from glutamine-derived ketoglutarate), M+2 and +1 (subsequent oxidative cycles of the TCA supplied by other unlabeled carbon sources than labeled glutamine). M+3 isotopomers of succinate, fumarate and malate would derive from reductive carboxylation of ketoglutarate. Grouped columns in the left indicate TAp63α-expressing (doxycyline supplemented cells – Dox +), while columns in the right refer to non-induced controls (Dox -). Kinetics assays have been performed by harvesting cells at 1, 6, 12 and 24h from medium replacement (which in turn was performed at 24h from doxycyline supplementation). * and # indicate statistical significance for inter-group (Dox + vs Dox -) or intra-group variations (time-course analyses). The number of * and # is related to p-values < 0.05; 0.01 or 0.001, respectively.

Figure 5. Bioinformatic validation of p63 in cancer metabolism

(A) Survival estimation curves of top 10% (good prognosis) and low 10% (bad prognosis) survival outcome from Metabric dataset. Each cohort includes approximately 200 samples. (B) Table showing correlation factors between p63 and metabolic enzymes of interest in the two cohorts “good prognosis” and “bad prognosis”. In dark green positive correlation, in red negative correlation. (C-D) Positive p63/G6PD and p63/PDG correlations (“Correlation” subgroups, in green) represent a negative prognostic factor for survival of breast cancer patients. Breast cancer datasets (GSE3494) was split in two cohorts ‘p63/metabolic enzyme Interaction’ and ‘p63/metabolic enzyme NO Interaction’ on the basis of existence or not of their correlation. Survival estimation was compute to show prediction of survival outcome for the two cohorts.