Metabolite profiling identifies a signature of tumorigenicity in hepatocellular carcinoma

Abstract

HCC (Hepatocellular carcinoma) cells exhibit greater metabolic plasticity than normal hepatocytes since they must survive in a dynamic microenvironment where nutrients and oxygen are often scarce. Using a metabolomic approach combined with functional in vitro and in vivo assays, we aimed to identify an HCC metabolic signature associated with increased tumorigenicity and patient mortality. Metabolite profiling of HCC Dt81Hepa1-6 cells revealed that their increased tumorigenicity was associated with elevated levels of glycolytic metabolites. Tumorigenic Dt81Hepa1-6 also had an increased ability to uptake glucose leading to a higher glycolytic flux that stemmed from an increased expression of glucose transporter GLUT-1. Dt81Hepa1-6-derived tumors displayed increased mRNA expressions of glycolytic genes, Hypoxia-inducible factor-1alpha and of Cyclin D1. HCC tumors also displayed increased energy charge, reduced antioxidative metabolites and similar fatty acid biosynthesis compared to healthy liver. Increased tumoral expression of glycolytic and hypoxia signaling pathway mRNAs was associated with decreased survival in HCC patients. In conclusion, HCC cells can rapidly alter their metabolism according to their environment and switch to the use of glucose through aerobic glycolysis to sustain their tumorigenicity and proliferative ability. Therefore, cancer metabolic reprogramming could be essential for the tumorigenicity of HCC cells during cancer initiation and invasion.

Keywords: glucose; hepatocellular carcinoma; liver; metabolic signature; tumorigenicity.

Figure 1. Metabolomic and glucose metabolism analysis of Dt81Hepa1-6 cells

(A) Heatmap depicting the metabolomic analysis of 29 metabolites in Dt81Hepa1-6 cells and primary hepatocytes both cultured in 25 mM glucose for 48 hrs. (B) Median fluorescence intensity analysis of glucose uptake with increasing doses of glucose fluorescent analog 2-NBDG [0–100 mM] in glucose-free DMEM. (C) Fluorescent signal quantification and representative microphotographs of 2-NBDG-labeled primary hepatocytes and Dt81Hepa1-6 cells [50 mM of 2-NBDG]. (D–E) Extracellular acidification rate (ECAR) measurements using Seahorse XF24 Extracellular Flux analyzer. Primary hepatocytes and Dt81Hepa1-6 cells were cultured in 25 mM glucose DMEM for 48 hrs. Glycolytic capacity and glycolytic reserve were calculated based on the increase in ECAR after injection of oligomycin. Values are ± SEM of at least 3 independent experiments. (^**P < 0.01, ^***P < 0.001).

Figure 2. Increased glucose uptake by Dt81Hepa1-6 is mediated by a rearrangement of glucose transporters

(A–B) Protein levels of Glucose transporter 1 (GLUT-1) and Glucose transporter 2 (GLUT-2) in primary hepatocytes and Dt81Hepa1-6 cells after 48 hrs incubation in 25 mM glucose DMEM (In vitro) and in healthy liver, non-tumoral and tumoral liver specimens (In vivo). Values are ± SEM of at least 3 independent experiments. (^***P < 0.001).

Figure 3. Expression of glycolysis-related genes by Dt81Hepa1-6 cells and Dt81Hepa1-6-derived tumors

mRNA gene expression of (A) Hexokinase II (Hk II), Phosphofructokinase liver (Pfkl), Pyruvate dehydrogenase (Pdh), Pyruvate dehydrogenase kinase 1 (Pdk1), Peroxisome proliferator-activated receptor gamma coactivator 1-alpha (Pgc-1α), (B) Hypoxia inducible factor-1alpha (Hif-1α) and (C) Cyclin D1, in primary hepatocytes and Dt81Hepa1-6 cells after a 48 hrs incubation in 25 mM glucose DMEM (In vitro) and in healthy liver, non-tumoral and tumoral liver specimens (In vivo). Values are ± SEM of at least 3 independent experiments. (^*P < 0.05, ^**P < 0.01, ^***P < 0.001).

Figure 4. Energetic profile of Dt81Hepa1-6-derived tumors greatly differs from that of healthy liver and non-tumoral samples

(A) Evaluation of total intracellular ATP and ATP/ADP ratio, (B) calculated Energy charge values, (C) NADH/NAD and (D) Lactate/Pyruvate ratios in healthy liver, non-tumoral and tumoral liver specimens. Values are ± SEM of at least 3 independent experiments. (^*P < 0.05, ^***P < 0.001).

Figure 5. Dt81Hepa1-6-derived tumors show lower contents of antioxidative-related metabolites

(A–B) Total intracellular NADP, NADPH and GSH, GSSG levels in healthy liver, non-tumoral and tumoral liver specimens. Values are ± SEM of at least 3 independent experiments. (^**P < 0.01, ^***P < 0.001).

Figure 6. Fatty acid biosynthesis in HCC cells

(A) mRNA relative gene expression of ATP citrate lyase (Acly), Acetyl-CoA carboxylase (Acc), Fatty acid synthase (Fasn), (B) assessment of triglyceride (TG) content, respectively in primary hepatocytes and Dt81Hepa1-6 cells after a 48 hrs incubation in 25 mM glucose DMEM (In vitro) and in healthy liver, non-tumoral and tumoral liver specimens (In vivo). (C) Kaplan–Meier (KM) plots of Overall survival probability of HCC cancer patients (TCGA data). Patients have been stratified into high (red lines) or low (green lines) expression-based ‘risk-groups’ by their mean of median transcript-expressions of fatty acid biosynthesis related genes. The patient follow-up is indicated in days. Respective Log-rank test p-values and Hazard Ratio (HR) are shown. The numbers of patients for each group are indicated below the respective KM plots. Studied genes are described in Supplementary Table 2.

Figure 7. High expressions of glycolysis and hypoxia-induced response genes are associated with poor prognosis in HCC patients

Kaplan–Meier (KM) plots of Overall survival probability of HCC cancer patients (TCGA data). Patients have been stratified into high (red lines) or low (green lines) expression-based ‘risk-groups’ by their mean of median transcript-expressions of (A) glycolytic and (B) hypoxia-induced response genes. Patient follow-up is indicated in days. Respective Log-rank test p-values and Hazard Ratio (HR) are shown. The numbers of patients for each group are indicated below the respective KM plots. Studied genes are described in Supplementary Table 2.