Ginsenoside 20(S)-Rg3 Prevents PKM2-Targeting miR-324-5p from H19 Sponging to Antagonize the Warburg Effect in Ovarian Cancer Cells Open Access

Ovarian cancer is the most lethal gynecological tumor, existing predominantly in the form of epithelial ovarian cancer (EOC). The estimated new cases of ovarian cancer are 22, 240, and the estimated deaths are 14, 070, which is 5% of deaths of females in the United States [1]. The 5-year survival has long been less than 50% in most countries [2]. Chemotherapy is the main adjuvant therapy for EOC, however, its efficacy is frequently impaired by occurrence of chemoresistance and adverse reaction [3]. Development of novel therapeutics is an important strategy to improve the prognosis of EOC patients.

Ginsenoside 20(S)-Rg3 is an active saponin monomer extracted from red ginseng, exerting anti-tumor function in multiple cancers [4]. A series of study have shown that 20(S)-Rg3 significantly inhibits ovarian cancer cells growth and metastasis partly via antagonizing the Warburg effect [5, 6]. The Warburg effect, also termed as aerobic glycolysis, is the main metabolic feature of cancers sustaining cancer cell growth under sufficient oxygen via rapidly generating ATP and biomaterials [7]. The Warburg effect is controlled by key glycolytic enzymes including hexokinase (HK), phosphofructokinase (PFK), pyruvate kinase (PK), and lactate dehydrogenase (LDH) [8]. Lately, long non-coding RNAs have been identified as a class of crucial regulators of the Warburg effect [9, 10]. However, the detailed lncRNA regulatory network modulated by 20(S)-Rg3 to prevent the Warburg effect in ovarian cancer cells has not been fully explored.

LncRNAs are a class of non-coding RNAs that are longer than 200 nucleotides implicated in tumor development and progression via regulating multiple genes expression to participate in various biological processes including metabolism, proliferation, apoptosis, migration and invasion. LncRNAs play their regulation role partly through acting as microRNA sponges to antagonize the binding of microRNA to their targets at 3’ untranslated regions (3’UTRs) and therefore prevent microRNA-rendered promotion of RNA degradation or inhibition of translation. Increasing evidences have shown that lncRNAs regulate the Warburg effect in various cancers via sponging microRNAs [11, 12].

Here, we screen the differentially expressed lncRNAs between 20(S)-Rg3-treated and non-treated ovarian cancer cells via deep sequencing, and establish one lncRNA regulatory pathway modulated by 20(S)-Rg3 to prevent the Warburg effect in ovarian cancer cells. The results provide essential information for clinical translation of 20(S)-Rg3 for ovarian cancer therapy.

Ginsenoside 20(S)-Rg3 was obtained from Tasly Pharmaceutical Company (Tianjin, China). 20(S)-Rg3 was dissolved in DMSO, and then diluted in RPMI 1640 media at the concentration of 2 mg/ml as a stock solution and stored at -20℃ for later use. The primary antibodies against PKM2 (#4053) and β-actin (#3700) were from Cell Signaling Technology (Beverly, MA, USA), and the primary antibody against Ago2 (03-110) was from Millipore (Billerica, MA, USA).

The human ovarian cancer cell line SKOV3 was obtained from the Shanghai Cell Bank of Chinese Academy of Sciences (Shanghai, China), and A2780 was from the Shandong Academy of Medical Sciences (Jinan, China). Cells were cultured in RPMI 1640 with 10% fetal bovine serum (GIBCO; Grand Island, NY, USA) at 37℃, 5%CO₂ and incubated with 80 μg/mL (for SKOV3 cells) or 40 μg/mL (for A2780 cells) of 20(S)-Rg3 for 24 h or 48 h.

Deep sequencing was performed, and the data were analyzed by Novel Bioinformatics Ltd. Co. (Shanghai, China). Total RNAs were extracted from SKOV3 cells treated with 0 or 80 μg/mL 20(S)-Rg3 for 24 h by TRIzol reagent (Invitrogen; Carlsbad, CA, USA), and RNA with RIN>8.0 was utilized to construct rRNA depletion library (VAHTSTM Total RNA-seq (H/M/R)) according to the manufacturer’s instructions. Raw data sequenced on the Ion Torrent Proton platform was filtered (removing the adaptor sequences, reads with >5% ambiguous bases (noted as N) and low-quality reads containing more than 20 percent of bases with qualities of <20). Clean data was then mapped to Human genome (GRCh38) utilizing HISAT2 algorithm [13]. HTSeq [14] was used to calculate the gene counts of lncRNAs. Differentially expressed lncRNA analysis was performed with DESeq algorithm [15] under following criteria: Fold Change>2, p-value<0.05 and FDR<0.05. The Miranda package [16] was used to predict microRNA target on the full length sequence of differentially expressed lncRNA.

Total RNA was extracted using TRIzol (Invitrogen; Carlsbad, CA, USA) and the concentrations were quantified on a UV spectrophotometer (Bio-Rad Inc.; Hercules, CA, USA). cDNA was synthesized according to the manufacturer’s instructions of PrimeScript^TM RT reagent kit with gDNA Eraser (TaKaRa; Dalian, China). Quantitative real-time PCR was performed on a CFX 96 real-time PCR system (Bio-Rad Inc,; Hercules, CA, USA) with SYBR Premix Ex Taq^TM II kit (TaKaRa; Dalian, China). β-actin was used as internal control of lncRNA and mRNA, and U6 was the internal control of microRNA. Relative gene expression was calculated according to 2^-ΔΔCt method. The oligonucleotide primers were synthesized by BGI (Shenzhen, China) and microRNA primers were synthesized by RiboBio Co. Ltd. (Guangzhou, China). Sequences of primers were as follows: H19 F 5’-TGC TGC ACT TTA CAA CCA CTG-3’, R 5’-ATG GTG TCT TTG ATG TTG GGC-3’; PKM2 F 5’-TCC GGA TCT CTT CGT CTT TG-3’, R 5’-GTC TGA ATG AAG GCA GTC CC-3’; β-actin F 5’-TCC CTG GAG AAG AGC TAC GA-3’, R 5’-AGC ACT GTG TTG GCG TAC AG-3’.

Total protein was extracted using RIPA buffer containing protease inhibitors (Roche; Indianapolis, IN, USA). Protein concentrations were quantified with the BCA-200 Protein Assay kit (Pierce; Rockford, IL, USA). Proteins were separated on 10% SDS-polyacrylamide gels and transferred onto nitrocellulose membranes (Pall Life Science; Port Washington, NY, USA). Then the membranes were blocked in 5% non-fat milk at room temperature for 1 hour and incubated at 4℃ overnight with the primary antibodies (PKM2 1: 10000, β-actin 1: 1000, Ago2 1: 500). After washing with 0.5% TBST, the membranes were incubated with HRP-conjugated secondary antibodies, and the blots were developed using ECL reagents (Millipore; Billerica, MA, USA) by a chemiluminescence imaging system (Bio-Rad Inc., Hercules, CA, USA).

The supernatant of cell culture media was collected to detect the levels of glucose and lactate separately with Glucose Assay kit and Lactic Acid Assay kit (Nanjing Jiancheng Bioengineering Institute; Nanjing, China) according to the manufacturer’s instructions. The optical density at 505 nm for glucose measurement and 530 nm for lactate measurement were read on an EnSpire multimode plate reader (PerkinElmer; Waltham, MA, USA). Glucose consumption and lactate production were calculated based on the absorption values and normalized to the cell numbers.

hsa-miR-324-5p inhibitor and mimic were purchased from RiboBio Co. Ltd. (Guangzhou, China). H19 siRNAs were synthesized by GenePharma (Shanghai, China). Sequences of siRNAs were as follows: siH19-a5’-UAA GUC AUU UGC ACU GGU UTT-3’; siH19-b 5’-CCA ACA UCA AAG ACA CCA UTT-3’; negative control siRNA 5’-GCG ACG AUC UGC CUA AGA UTT-3’. The human H19-expressing plasmid was constructed by Genechem (Shanghai, China). Cells were seeded into 6-well plates. hsa-miR-324-5p inhibitor and mimic and H19 siRNAs were transfected at 30-50% confluency with X-treme GENE siRNA transfection reagent (Roche; Indianapolis, IN, USA) according to the manufacturer’s instructions. The H19-expressing plasmid was transfected at 70-90% confluency with X-treme GENE HP DNA transfection reagent (Roche; Indianapolis, IN, USA). After 48 h or 72 h incubation, cells were harvested for further study.

The human H19 gene and the 3’UTR of human PKM2 gene containing predicted miR-324-5p binding sites were amplified from genomic DNA of SKOV3 using the following primers: H19 F 5’-CCC AAG CTT GGG CAG CCA CCA CAT CAT CCC AG-3’, R 5’-GGA CTA GTC CCA GAG TCG TGG AGG CTT TGA-3’; PKM2 3’UTR F 5’-CCC AAG CTT GGG GGA GGA ATG CTG GAC TGG AG-3’, R 5’-CCC AAG CTT GGG GGA GGA ATG CTG GAC TGG AG-3’. The PCR products were recovered from running gels, digested with Hind III and Spe I endonuclease and inserted into pMIR-REPORT luciferase vector (Ambion; Austin, TX, USA) to construct H19-WT vector and PKM2-WT 3’UTR vector. The transversion mutants in the miR-324-5p binding sites were generated by overlap extension method to construct H19-MUT vector and PKM2-MUT 3’UTR vector. The primers were used as follows:

H19-MUT P1: 5’-CCCAAGCTTGGGCAGCCACCACATCATCCCAG-3’,

P2: 5’-AGCTCCTCCAGCCTTCATACGCCGTCCCCACCA-3’,

P3: 5’-TGGTGGGGACGGCGTATGAAGGCTGGAGGAGCT-3’,

P4: 5’-GGACTAGTCCCAGAGTCGTGGAGGCTTTGA-3’;

PKM2-MUT 3’UTR P1: 5’-CCCAAGCTTGGGGGAGGAATGCTGGACTGGAG-3’,

P2: 5’-CCTTTAGAAAAAATCCTACGCCAGAGGACTCCC-3’,

P3: 5’-GGGAGTCCTCTGGCGTAGGATTTTTTCTAAAGG-3’,

P4: 5’-GGACTAGTCCTGGCTGTTTCTTGACCCCAA-3’.

All the recombinant vectors were confirmed by sequencing. Cells were seeded into 24-well plates. pRL-TK vectors (20 ng) and wild-type or mutation vectors (180 ng) were co-transfected along with 20 nM of miR-324-5p mimic or negative control at 80-90% confluency with X-treme GENE siRNA transfection reagent (Roche; Indianapolis, IN, USA) according to the manufacturer’s instructions. 24 h after transfection, cells were harvested and the relative firefly luciferase activity (normalized to Renilla luciferase activity) was measured with Dual-Luciferase Reporter Assay System (Promega; Madison, WI, USA) according to the manufacturer’s instructions.

Cells were scraped off culture dishes and lysed using complete RIP lysis buffer at 80-90% confluency using Magna RIP kit (Millipore; Billerica, MA, USA). 100 μl of cell lysate was incubated with RIP immunoprecipitation buffer containing magnetic beads conjugated with anti-Ago2 antibody or normal mouse IgG at 4℃ overnight. The immunoprecipitated RNA was purified by digesting protein with proteinase K and analyzed by qRT-PCR.

hsa-miR-324-5p-up lentivirus was constructed by Genechem (Shanghai, China). The lentivirus infected cells with MOI 10 at 20-30% confluency. The cells were then used for cell viability assay, colony formation assay and xenograft tumor model.

1200 cells/well were seeded into 96-well plates. Cell viability was examined with Cell Counting Kit-8 (7seapharmtech Co. Ltd.; Shanghai, China) for 7 days. 10 μl of CCK-8 was added into 100 μl of culture media, and cells were incubated at 37℃ for 0.5 h. The 450 nm wave-length absorption values were measured on an EnSpire multimode plate reader (PerkinElmer; Waltham, MA, USA). The detections were performed in triplicate wells on the same plate in three independent experiments.

500 cells/well were seeded into 6-well plates and cultured for 10 days. Cells were fixed with methanol and stained with crystal violet. The images were captured and the numbers of colonies in which the number of cells was more than 50 were counted under microscope.

All animal experimental procedures were approved by the Ethics Committee of the First Affiliated Hospital, Xi’an Jiaotong University. Four-week old BALB/c female mice were randomly divided into 2 groups (miR-324-5p and negative control, n=7) and injected subcutaneously with 2× 10⁶ cells (100 μl) into the right axilla. Mice weight and the length (L) and width (W) of xenograft tumors were measured every 3 days. Xenograft volume (V) was calculated as V = (L ×W²)/2. The tumors were collected, weighed, and preserved in liquid nitrogen or formalin for further study at day 24.

Histologic slides were prepared from the paraffin-embedded subcutaneous xenograft tumor tissue blocks, and then dewaxed in xylene and rehydrated in a descending alcohol series. The specimens were heated in 0.01 M citrate buffer (pH 6.0) at 98℃ for 20 minutes in a microwave oven to retrieve antigenic binding sites and incubated with anti-PKM2 antibody (1: 800) at 4℃ overnight. After incubation with HRP-conjugated secondary antibody at room temperature for 15 minutes, color was developed with DAB. Slides were counterstained with hematoxylin, dehydrated in an ascending alcohol series and mounted for analysis. Images were captured with an Olympus BH-2 microscope (Tokyo, Japan) installed with the DeltaPix Camera and software (Maalov, Denmark).

All data were represented as mean ± standard deviation and analyzed with SPSS software (Chicago, IL, USA). Statistical differences were determined by two-tailed t-test and considered significant (*) when P<0.05.

LncRNA deep sequencing was performed to examine the 20(S)-Rg3-regulated lncRNAs in SKOV3 ovarian cancer cells. The raw lncRNA sequencing data was submitted to Gene Expression Omnibus (GEO) database and accessed with the GEO accession number GSE118216 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE118216). A total of 5305 lncRNAs were detected in 20(S)-Rg3-treated SKOV3 cells and negative control cells, including 1524 ncRNAs and 3781 pseudo genes. Compared to the control cells, 67 lncRNAs were significantly decreased (fold change<0.5, p<0.05) including 12 ncRNAs and 55 pseudo genes, and 18 lncRNAs were significantly increased (fold change>1.5, p<0.05) including 3 ncRNAs, 14 pseudo genes and 1 unknown gene in the 20(S)-Rg3-treated SKOV3 cells (Fig. 1A). The fold change (FC) and the corresponding false discovery rate (FDR) for the differentially expressed lncRNAs between 20(S)-Rg3-treated and non-treated SKOV3 cells were summarized in Table 1. Among the significantly differentially expressed lncRNAs, 4 lncRNAs presented the lowest FDR value (FDR=0) including downregulated pseudo gene RNU1-19P and ncRNA H19, and pseudo genes LOC93622 and CICP7.

H19 level was examined in 20(S)-Rg3-treated SKOV3 and A2780 ovarian cancer cells by qRT-PCR. Consistent with lncRNA deep sequencing results (Fig. 1B), H19 was found to be decreased by 50% in 20(S)-Rg3-treated cells compared to control cells (Fig. 1C). The role of H19 in the inhibition of 20(S)-Rg3 on the Warburg effect was then examined. Knocking down of H19 by siRNAs in SKOV3 and A2780 cells (Fig. 2A) reduced the glucose consumption (Fig. 2B) and lactate production (Fig. 2C), and decreased PKM2 rather than HK2 expression (Fig. 2D). Overexpression of H19 in 20(S)-Rg3-treated SKOV3 and A2780 cells (Fig. 2E) reversed the inhibitory effect of 20(S)-Rg3 on the glucose consumption (Fig. 2F), lactate production (Fig. 2G) and PKM2 expression (Fig. 2H), but had no effect on HK2 expression. These results indicated that 20(S)-Rg3 prevented the Warburg effect via suppressing H19 expression in ovarian cancer cells.

We then investigated the mechanism of the stimulation of H19 on PKM2 expression. Since H19 was found to act as the microRNAs sponge, we attempted to identify a PKM2-targeting microRNA sponged by H19. We previously reported that 11 microRNAs were significantly upregulated by 20(S)-Rg3, including miR-3163, miR-664a-5p, miR-6717-5p, miR-4329, miR-603, miR-324-5p, miR-1283, miR-532-3p, miR-33a-3p, miR-519a-5p and miR-486-3p. Analysis using RNAHybrid algorithm found a binding site of miR-324-5p in H19 sequence located in 1897-1902nt, and PKM2 was predicted as a potential target of miR-324-5p by microRNA.org database. Knocking down of H19 increased miR-324-5p level (Fig. 3A), while overexpression of H19 abrogated the 20(S)-Rg3-stimulated miR-324-5p (Fig. 3B). But inhibition of miR-324-5p in 20(S)-Rg3-treated cells or overexpression of miR-324-5p did not influence the level of H19 (Fig. 3C and 3D). These results indicated H19 negatively regulated miR-324-5p, but not vice versa.

The direct action of H19 on miR-324-5p was further verified by dual-luciferase reporter assay and RNA-binding protein immunoprecipitation (RIP). Dual-luciferase reporter assay showed that miR-324-5p mimic weakened the luciferase activity of H19 WT luciferase reporter vector containing the wild-type miR-324-5p-binding site in H19 sequence, while it had no effect on that of H19 MUT luciferase reporter vector containing the mutant miR-324-5p-binding site (Fig. 3E). RIP for Ago2 followed by qRT-PCR showed that endogenous H19 and miR-324-5p were specifically enriched in the component pulled down by anti-Ago2 antibody (Fig. 3F), manifesting that H19 was associated with RNA-induced silencing complex composed of miR-324-5p and Ago2 protein.

Next we knocked down both of H19 and miR-324-5p in SKOV3 and A2780 (Fig. 3G). The reduction of glucose consumption and lactate production caused by H19 interference was prevented by miR-324-5p inhibition (Fig. 3H and 3I). Meanwhile, the protein level of PKM2 in H19 and miR-324-5p co-inhibited cells was increased compared with that in H19-inhibited cells (Fig. 3J). These results illustrated that H19 enhanced PKM2 expression to promote the Warburg effect in ovarian cancer cells by sponging miR-324-5p.

The effect of miR-324-5p on the Warburg effect in ovarian cancer cells was then examined. Overexpression of miR-324-5p via transfecting miR-324-5p mimic into SKOV3 and A2780 cells (Fig. 4A) reduced the glucose consumption (Fig. 4B) and lactate production (Fig. 4C), and significantly decreased PKM2 protein level (Fig. 4D). Next miR-324-5p was suppressed via transfecting miR-324-5p inhibitor into 20(S)-Rg3-treated SKOV3 and A2780 cells. qRT-PCR results showed that miR-324-5p inhibitor antagonized 20(S)-Rg3-induced elevation of miR-324-5p (Fig. 4E), and reduction of glucose consumption (Fig. 4F), lactate production (Fig. 4G) and PKM2 protein level (Fig. 4H). These results indicated that 20(S)-Rg3 inhibited the Warburg effect in ovarian cancer cells by upregulating miR-324-5p.

Online algorithm (microRNA.org) predicted the miR-324-5p-binding site was located in 331-336nt of PKM2 3’UTR. We constructed PKM2 WT-3’UTR containing the miR-324-5p-binding site and PKM2 MUT-3’UTR containing transversion mutations of the miR-324-5p-binding site. Dual-luciferase reporter assay showed that compared with the negative control, miR-324-5p mimics significantly decreased the luciferase activity of PKM2 WT-3’UTR, but had no effect on the luciferase activity of PKM2 MUT-3’UTR (Fig. 4I). These data verified that miR-324-5p directly targeted PKM2.

To verify the influence of the anti-Warburg effect of miR-324-5p in ovarian cancer cells, SKOV3 cells were infected by hsa-miR-324-5p-up lentivirus to overexpress miR-324-5p (Fig. 5A), and cell proliferation was examined. Compared with the negative control, overexpression of miR-324-5p significantly retarded cell viability (Fig. 5B) and colony formation capability (Fig. 5C), illustrating that miR-324-5p inhibited ovarian cancer cell proliferation in vitro. We then checked the effect of miR-324-5p on the proliferation of ovarian cancer cells in vivo. Nude mice subcutaneous xenograft models were established using miR-324-5p-up SKOV3 cells and negative control cells, respectively. The mice body weights in negative control group and miR-324-5p-up group were comparable (Fig. 5D). But compared with the negative control group, the tumorigenicity of miR-324-5p-up cells was substantially impaired as illustrated by the fewer xenografts and smaller tumor volumes (Fig. 5E and 5F). qRT-PCR results of xenograft tissues showed that miR-324-5p was increased while H19 remained unchanged in miR-324-5p-up group relative to those of the negative control group (Fig. 5G). Immunohistochemistry results showed that PKM2 protein level was decreased in miR-324-5p-up group compared to that in negative control group (Fig. 5H).

The Warburg effect is one of the main metabolism ways for ovarian cancer. Research and development of novel therapeutics targeting the Warburg effect will be a feasible strategy to enhance the efficacy of first-line chemotherapeutic drugs and reduce adverse reaction in EOC patients [17]. 20(S)-Rg3 exhibits anti-tumor activity in various cancers via inducing apoptosis or weakening proliferation, migration and invasion of cancer cells [18-20]. Our previous studies have found that 20(S)-Rg3 antagonized the Warburg effect in ovarian cancer and downregulates the two key glycolytic enzymes HK2 and PKM2 [5, 6]. Mechanistic studies have revealed that 20(S)-Rg3 downregulates HK2 via STAT3 [5] or DNMT3A/miR-532-3p [6] pathways. But the mechanism of PKM2 downregulation induced by 20(S)-Rg3 remains unknown.

We proceeded to investigate the PKM2 downregulation pathway based on previously acquired microRNA sequencing data [6]. PKM2 was predicted and experimentally confirmed as the direct target of 20(S)-Rg3-upregulated miR-324-5p. In view of accumulating evidence proving lncRNAs as microRNA sponges to negatively regulate microRNA function [21-23], we identified the 20(S)-Rg3-influenced lncRNA expression profile with deep sequencing technique to investigate the lncRNA competitively binding miR-324-5p. Fortunately, H19 was screened out from 67 20(S)-Rg3-reduced lncRNAs, with the binding site of miR-324-5p in its sequence. The competitive binding of H19 to miR-324-5p rescued PKM2 from being translationally inhibited by miR-324-5p. H19/miR-324-5p/PKM2 pathway was therefore established as the anti-Warburg mechanism of 20(S)-Rg3.

H19 was overexpressed in various cancers including ovarian cancer, and capable of stimulating cancer cell cycle progression, proliferation, migration and invasion [24, 25]. H19 facilitated tumor development and progression partly via sponging microRNA to recover the expression of microRNA-targeted genes [26-28]. Recently H19 was found to promote glucose metabolism and cell growth in malignant melanoma via competitively binding miR-106a-5p to release E2F3 [29]. Here we identified PKM2-targeting miR-324-5p as another target microRNA sponged by H19, and H19 therefore was a pro-Warburg lncRNA while miR-324-5p was an anti-Warburg microRNA. miR-324-5p has been reported as both oncomiR and tumor repressor in cancers because it inhibited cell growth and invasion in multiple myeloma and hepatocellular cancer [30], while promoted proliferation, migration and invasion of papillary thyroid microcarcinoma cells and breast cancer cells [31, 32]. But most published results were from in vitro data. Here, miR-324-5p was verified in vitro and in vivo as a suppressor of ovarian cancer proliferation via negatively regulating the Warburg effect.

In this study, we first reported the 20(S)-Rg3-regulated lncRNA expression profile and identified H19/miR-324-5p/PKM2 pathway responsible for the anti-Warburg activity of 20(S)-Rg3. Further study of 20(S)-Rg3-regulated lncRNA and microRNA could contribute to the understanding of both anti-cancer action of 20(S)-Rg3 and progression mechanism of ovarian cancer.

Our results indicated that 20(S)-Rg3 inhibited the Warburg effect via H19/miR-324-5p/PKM2 pathway to retard growth of ovarian cancer cells.

This work was supported by the National Natural Science Foundation of China (No. 81702576 and No. 30973429), the Natural Science Foundation of Shaanxi Province (No. 2017JM8107) and Foundation of the First Affiliated Hospital of Xi’an Jiaotong University (No. 2016QN-10).

The authors declare no conflicts of interest.

X. Zheng and Y. Zhou contributed equally to this work.