Inhibition of Phosphoglycerate Dehydrogenase Radiosensitizes Human Colorectal Cancer Cells under Hypoxic Conditions

Abstract

Augmented de novo serine synthesis activity is increasingly apparent in distinct types of cancers and has mainly sparked interest by investigation of phosphoglycerate dehydrogenase (PHGDH). Overexpression of PHGDH has been associated with higher tumor grade, shorter relapse time and decreased overall survival. It is well known that therapeutic outcomes in cancer patients can be improved by reprogramming metabolic pathways in combination with standard treatment options, for example, radiotherapy. In this study, possible metabolic changes related to radioresponse were explored upon PHGDH inhibition. Additionally, we evaluated whether PHGDH inhibition could improve radioresponse in human colorectal cancer cell lines in both aerobic and radiobiological relevant hypoxic conditions. Dysregulation of reactive oxygen species (ROS) homeostasis and dysfunction in mitochondrial energy metabolism and oxygen consumption rate were indicative of potential radiomodulatory effects. We demonstrated that PHGDH inhibition radiosensitized hypoxic human colorectal cancer cells while leaving intrinsic radiosensitivity unaffected. In a xenograft model, the first hints of additive effects between PHGDH inhibition and radiotherapy were demonstrated. In conclusion, this study is the first to show that modulation of de novo serine biosynthesis enhances radioresponse in hypoxic colorectal cancer cells, mainly mediated by increased levels of intracellular ROS.

Keywords: colorectal cancer; hypoxia; radiosensitivity; reactive oxygen species; serine synthesis pathway.

Figure 1

Serine synthesis pathway enzymes are expressed in human colorectal cancers. (A) PHGDH, PSAT1 and PSPH mRNA expression levels of GTEx normal tissues were compared to mRNA expression profile of CRC tumor tissues out of the TCGA database (RSEM values). (B) Kaplan–Meier curves showing overall survival in PHGDH, PSAT1 and PSPH low/high mRNA-expressing rectum cancer patients out of the TCGA database. (C) Representative Western blots of PHGDH, PSAT1 and PSPH with β-actin as loading control in HCT116 and DLD-1 cells under normoxic conditions. (D) Relative expression of PHGDH, PSAT1 and PSPH normalized to β-actin in HCT116 and DLD-1 cells under normoxic conditions. Data are shown from at least three replicates as mean ± SEM. Unpaired t-test with Mann–Whitney test was used for statistical analysis: *** p < 0.001, **** p < 0.0001.

Figure 2

PHGDH inhibition affects mitochondrial cellular respiration under normoxic conditions. (A) Representative experiment showing the oxygen consumption rates (OCR) of normoxic HCT116 (top) and DLD-1 (bottom) was measured upon treatment with NCT-503 (16 h) at indicated concentrations after consecutive injection at indicated time points of oligomycin, FCCP, rotenone and antimycin A using the Seahorse analyzer. The OCR was expressed as pmoles/min/µg protein. (B) Summary graphs showing basal respiration; (C) summary graphs showing ATP production; (D) summary graphs showing the levels of maximal respiration. Data are shown from at least three replicates as mean ± SEM. One-way ANOVA with Dunnett’s multiple comparison test was used to calculate statistics: * p < 0.05, ** p < 0.01, **** p < 0.0001.

Figure 3

PHGDH inhibition does affect end-metabolites aKG, GSH, NADPH and redox balance in normoxic CRC cells. HCT116 (top) and DLD-1 (bottom) cells were treated with NCT-503 for 5 h at indicated concentrations. (A) aKG levels—normalized to control; (B) GSH levels—normalized to control; (C) NADPH levels—normalized to control; (D) ROS levels—normalized to control. Data are shown from at least four replicates as mean ± SEM. One-way ANOVA with Dunnett’s multiple comparison test was used to calculate statistics: * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 4

Hypoxia induces the expression of PHGDH in human colorectal cancers. (A) Correlation between PHGDH overexpression (p ≥ 1.5) and hypoxia by assessing the Buffa, Winter and Ragnum hypoxia scores accessed through cBioPortal for Cancer Genomics. (B) Representative Western blots of PHGDH, PSAT1 and PSPH with β-actin as loading control in HCT116 and DLD-1 cells under hypoxic conditions (0.1% O₂). (C) Relative protein expression of PHGDH, PSAT1 and PSPH under hypoxic conditions (0.1% O₂) normalized to the expression under normoxic conditions. (D) D-serine levels were determined upon treatment with NCT-503 (5 h) at indicated concentrations under hypoxic conditions (0.1% O₂). Unpaired t-test (A) and ordinary one-way ANOVA with Dunnett’s multiple comparison test (D) were used to calculate statistics: * p < 0.05, *** p < 0.001, **** p < 0.0001.

Figure 5

PHGDH inhibition mainly disrupts ROS homeostasis in human hypoxic colorectal cancer cells. HCT116 (top) and DLD-1 (bottom) cells were treated with NCT-503 for 5 h at indicated concentrations under hypoxic conditions (0.1% O₂). (A) aKG levels—normalized to control; (B) GSH levels—normalized to control; (C) NADPH levels—normalized to control; (D) ROS levels—normalized to control. Data are shown from at least four replicates as mean ± SEM. One-way ANOVA with Dunnett’s multiple comparison test was used to calculate statistics: * p < 0.05, ** p < 0.01.

Figure 6

Inhibition of PHGDH radiosensitizes human hypoxic colorectal cancer cells through excessive ROS. HCT116 (top) and DLD-1 (bottom) cells were treated with NCT-503 (16 h) at indicated concentrations, irradiated and reseeded for colony formation assay. (A) Dose–response curves under normoxic (B) and hypoxic conditions. (C) Dose–response curve under hypoxic conditions with treatment of 60 μM NCT-503 with the addition of NAC.

Figure 7

Inhibition of PHGDH combined with fractionated radiation delays tumor growth in HCT116 xenografts. (A) Representative schedule of the experimental setup in vivo. (B) Tumor growth curve of mice with day 0 as the starting point of treatment. (C) Survival curve of mice with day 0 as the starting point of treatment. Data are shown as mean with SEM (n = 5).