RORα inhibits gastric cancer proliferation through attenuating G6PD and PFKFB3 induced glycolytic activity

Abstract

Background: Glycolysis is critical for harvesting abundant energy to maintain the tumor microenvironment in malignant tumors. Retinoic acid-related orphan receptor α (RORα) has been identified as a circadian gene. However, the association of glycolysis with RORα in regulating gastric cancer (GC) proliferation remains poorly understood.

Methods: Bioinformatic analysis and retrospective study were utilized to explore the role of RORα in cell cycle and glycolysis in GC. The mechanisms were performed in vitro and in vivo including colony formation, Cell Counting Kit-8 (CCK-8), Epithelial- mesenchymal transition (EMT) and subcutaneous tumors of mice model assays. The key drives between RORα and glycolysis were verified through western blot and chip assays. Moreover, we constructed models of high proliferation and high glucose environments to verify a negative feedback and chemoresistance through a series of functional experiments in vitro and in vivo.

Results: RORα was found to be involved in the cell cycle and glycolysis through a gene set enrichment analysis (GSEA) algorithm. GC patients with low RORα expression were not only associated with high circulating tumor cells (CTC) and high vascular endothelial growth factor (VEGF) levels. However, it also presented a positive correlation with the standard uptake value (SUV) level. Moreover, the SUV_max levels showed a positive linear relation with CTC and VEGF levels. In addition, RORα expression levels were associated with glucose 6 phosphate dehydrogenase (G6PD) and phosphofructokinase-2/fructose-2,6-bisphosphatase (PFKFB3) expression levels, and GC patients with low RORα and high G6PD or low RORα and high PFKFB3 expression patterns had poorest disease-free survival (DFS). Functionally, RORα deletion promoted GC proliferation and drove glycolysis in vitro and in vivo. These phenomena were reversed by the RORα activator SR1078. Moreover, RORα deletion promoted GC proliferation through attenuating G6PD and PFKFB3 induced glycolytic activity in vitro and in vivo. Mechanistically, RORα was recruited to the G6PD and PFKFB3 promoters to modulate their transcription. Next, high proliferation and high glucose inhibited RORα expression, which indicated that negative feedback exists in GC. Moreover, RORα deletion improved fluorouracil chemoresistance through inhibition of glucose uptake.

Conclusion: RORα might be a novel biomarker and therapeutic target for GC through attenuating glycolysis.

Keywords: G6PD; Gastric cancer; Glycolysis; PFKFB3; Proliferation; RORα.

Fig. 1

RORα is associated with GC proliferation and glycolysis. A GSEA algorithm analysis according to Reactome and Wikipathways database, respectively. B The difference analysis of CTC number and VEGF levels in RORα-high and RORα-low groups. C Left: Representative 18F-FDG PET/CT images of GC patients with maximum and minimum SUV_max levels. Up: The correlation analysis of SUV_max levels with CTC number and VEGF levels in GC patients, respectively. Down: The difference analysis of SUV_max levels in RORα-high and RORα-low groups. D G6PD and PFKFB3 mRNA expression levels in stomach adenocarcinoma-tumor (STAD-tumor) tissues and stomach adenocarcinoma-normal (STAD-normal) tissues from TCGA database were analyzed through TIMER 2.0. F Up: RORα, G6PD and PFKFB3 expression levels were detected through immunohistochemistry stain in normal gastric and GC patients with pT2N+, pT3N+ and pT4N+ stages tissues. Original 200 and 400 magnification. Scale bar =100 μm. Down: The correlation analysis of RORα expression levels with G6PD and PFKFB3 expression levels in GC patients, respectively. G The DFS time of GC patients with different RORα, G6PD and PFKFB3 expression patterns. All data are represented as the mean ± standard deviation. *P < 0.05, **P < 0.01, and ***P < 0.001

Fig. 2

RORα deletion promotes GC proliferation and is reversed by SR1078. A RORα expression levels were detected by western blot in lentiviral transfected AGS, MKN-74 and MFC cells. β-actin as a loading control. B The colonies of GC cells were performed with colony formation assay. N = 3. C The absorbance value of RORα-KO GC cells or GC cells treated with SR1078(5 µM) according to time gradient through CCK-8 assay. N = 3. D The relative mRNA levels of E-cadherin, N-cadherin and Vimentin were detected by qPCR assay in RORα-KO GC cells or GC cells treated with SR1078(5 µM). N=3. E The expression levels of Ki-67 and PCNA were detected through immunohistochemistry stain and MOD method in subcutaneous tumor. Original 200 and 400 magnification. Scale bar = 100 μm. N = 3. The RORα-KO-MFC cells were injected into subcutaneous flank of mice until the tumor volume reached 100 to 300 mm³. F The expression levels of Ki-67 and PCNA were detected through immunohistochemistry stain and MOD method in subcutaneous tumor. Original 200 and 400 magnification. Scale bar = 100 μm. N = 3. The MFC cells treated with SR1078 (5 µM) were injected into subcutaneous flank of mice until the tumor volume reached 100 to 300 mm³. All data are represented as the mean ± standard deviation. Vehicle was PBS. *P < 0.05, **P < 0.01, and ***P < 0.001

Fig. 3

RORα deletion promotes glycolysis and is reversed by SR1078 in GC cells. A The changes of ECAR levels in RORα-KO GC cells. The glycolysis and glycolytic capacity were measured in the treatment of glucose and oligomycin, respectively. N = 3. B The changes of ECAR levels in GC cells treated with SR1078(5 µM). The glycolysis and glycolytic capacity were measured in the treatment of glucose and oligomycin, respectively. N = 3. All data are represented as the mean ± standard deviation. Vehicle was PBS. *P < 0.05, **P < 0.01, and ***P < 0.001

Fig. 4

RORα deletion promotes glycolysis induced by 3PO and DHEA in GC cells The changes of ECAR levels in RORα-KO GC cells treated with 3PO (25 µM) or DHEA (250 µM). The glycolysis and glycolytic capacity were measured in the treatment of glucose and oligomycin, respectively. N = 3. All data are represented as the mean ± standard deviation. Vehicle was DMSO. *P < 0.05, **P < 0.01, and ***P < 0.001. Vs. RORα-KO+Vehicle group. ⁺P < 0.05, ⁺⁺P < 0.01, and ⁺⁺⁺P < 0.001. Vs. CON+Vehicle group

Fig. 5

RORα deletion promotes GC proliferation induced by 3PO and DHEA in vitro. A The colonies of GC cells were performed with colony formation assay. N=3. ^#P > 0.05. Vs. CON+Vehicle group in AGS cell. ^##P > 0.05. Vs. RORα-KO+Vehicle group in MKN-74 cell. B The absorbance value of RORα-KO GC cells treated with 3PO (25 µM) or DHEA (250 µM) according to time gradient through CCK-8 assay. N = 3. C The relative mRNA levels of E-cadherin, N-cadherin and Vimentin were detected by qPCR assay in RORα-KO GC cells treated with 3PO (25 µM) or DHEA (250 µM). N = 3. All data are represented as the mean ± standard deviation. Vehicle was DMSO. *P < 0.05, **P < 0.01, and ***P < 0.001. Vs. RORα-KO+Vehicle group. ⁺P < 0.05, ⁺⁺P < 0.01, and ⁺⁺⁺P < 0.001. Vs. CON+Vehicle group

Fig. 6

RORα inhibits G6PD and PFKFB3 gene expression through the recruitment of promoter in GC cells. A G6PD and PFKFB3 protein expression levels were detected by western blot in RORα-KO, CON, RORα-KO plus 3PO and RORα-KO plus DHEA GC cells. β-actin as a loading control. B G6PD and PFKFB3 mRNA levels were detected by q-PCR in lentiviral transfected GC cells. N = 3. C RORα was recruited to G6PD and PFKFB3 gene promoter in GC cells through Chip assay. N=3. All data are represented as the mean ± standard deviation. *P < 0.05, **P < 0.01, and ***P < 0.001

Fig. 7

High proliferation and high glucose inhibit RORα expression in GC cells. A The colonies of GC cells treated with TGF-β1 (10ng/ml) were performed with colony formation assay. N = 3. B The absorbance value of GC cells treated with TGF-β1 (10 ng/ml) according to time gradient through CCK-8 assay. N = 3. C The relative mRNA levels of E-cadherin, N-cadherin and Vimentin were detected by qPCR assay in GC cells treated with TGF-β1 (10 ng/ml). N = 3. D RORα protein expression levels were detected by western blot in GC cells treated with TGF-β1 (10 ng/ml). E RORα protein expression levels were detected by western blot in GC cells treated with glucose in concentration gradient. All data are represented as the mean ± standard deviation. Vehicle was DMSO. β-actin as a loading control. *P < 0.05, **P < 0.01, and ***P < 0.001

Fig. 8

Inhibition of glucose uptake improved fluorouracil chemoresistance in RORα-KO GC cells. A The GC cells were pretreated with 5mM or 50mM glucose for 24h, and then were treated with fluorouracil according to 0, 50, 100, 150 and 200 µg/ml concentration gradient. N = 3. B The GC cells were pretreated with 5 mM or 50 mM glucose for 24 h, and then were treated with fluorouracil (50 and 200 µg/ml concentration) according to time gradient. N = 3. The absorbance value was measured through CCK-8 assay. All data are represented as the mean ± standard deviation. Vehicle was DMSO. *P < 0.05, **P < 0.01, and ***P < 0.001