ALKBH5-induced circRNA NRIP1 promotes glycolysis in thyroid cancer cells by targeting the miR-541-5p/PKM2 and miR-3064-5p/PKM2 axes

DOI: https://doi.org/10.21203/rs.3.rs-1493404/v1

Abstract

Background: Circular RNAs (circRNAs) play important regulatory roles in various cancers. During tumor progression, circRNAs are involved in cell proliferation, differentiation, apoptosis, and invasion. However, the underlying regulatory mechanisms of circRNAs in thyroid cancer have not been fully elucidated. This article aims to study the role of circRNA regulated by m6A modification in PTC, and to clarify the effect and mechanism of this circRNA on the biological behavior of PTC.

Methods: Quantitative real-time PCR (qRT-PCR), western blotting and immunohistochemistry were performed to investigate the expressions of circNRIP1 in PTC tissues and in the adjacent non-cancerous thyroid tissues. In vitro and in vivo experiments were performed to assess the effects of circNRIP1 on PTC glycolysis and growth. The m6A modification mechanisms of circNRIP1 were evaluated by MeRIP-qPCR, luciferase reporter gene and RNA stability analyses.

Results: CircNRIP1 levels in PTC tissues were significantly upregulated. In PTC patients, elevated circNRIP1 levels correlated with high tumor lymph node metastasis (TNM) stage and larger tumor sizes. Functionally, circNRIP1 significantly promoted glycolysis as well as PTC cell proliferation in vitro, and enhanced tumorigenesis in vivo. Mechanistically, circNRIP1 acted as a sponge for miR-541-5p and miR-3064-5p and jointly upregulated PKM2 expressions. Knockdown of the m6A demethylase (ALKBH5) significantly enhanced circNRIP1 m6A modification and upregulated its expressions.

Conclusion: ALKBH5 knockdown upregulates circNRIP1 to promote glycolysis in PTC cells, therefore, circNRIP1 is a potential prognostic biomarker and therapeutic target for PTC by acting as a sponge for oncogenic miR-541-5p and miR-3064-5p and upregulating PKM2 expressions.

Background

Thyroid cancer is the most common endocrine-related malignancy, of which papillary thyroid cancer is the most prevalent thyroid cancer subtype, accounting for approximately 90% of all thyroid cancer cases [1, 2]. Over the past two decades, there have been an increase in the prevalence of thyroid cancer, more than doubling every year [3]. The recommended option for PTC therapy is surgical resection combined with radioactive iodine (RAI) therapy and thyroid-stimulating hormone (TSH) suppression therapy [4]. Although prognostic outcomes after treatment are good, 10–15% of PTC patients present with post-treatment recurrence and metastasis [5]. For PTC patients with local recurrence, their overall survival rate is 70%-85% [6]. A proportion of advanced invasive and locally recurrent and/or metastatic thyroid cancers do not respond to the above treatments, and are often incurable [7, 8]. For patients with such poor prognostic outcomes, their 10-year survival rate is 40% [9].. Therefore, elucidation of the molecular mechanisms involved in PTC development and establishment of prognostic factors for PTC therapy are vital for informing therapy.

Originally, circRNAs were considered to occur due to splicing errors during RNA splicing and had no biological functions [10]. However, with advances in RNA sequencing (RNA-seq) technology and bioinformatics, a large number of circRNAs have become increasingly common in mammalian cells. The discovery disapproved this hypothesis. CircRNAs are major transcripts for multiple human cell types [11], implying that they are stable, conserved, and non-random products of RNA splicing [12]. Importantly, during tumor progression, circRNAs are involved in cell proliferation, differentiation, apoptosis and invasion [13, 14]. However, the potential regulatory mechanisms of circRNAs in PTC have not been fully elucidated.

Abnormal glucose metabolism is an important feature of tumor cells. Despite the presence of sufficient oxygen, the efficiency of aerobic glycolysis is low. Cancer cells tend to generate energy through glycolysis, a process referred to as the "Warburg effect" [15–17]. Therefore, abnormal expressions of some key glycolysis-related genes, such as hexokinase 2 (HK2), pyruvate kinase isoenzyme M2 and glucose transporter genes, in various tumors promotes cancer cell proliferation and invasion. [18, 19]. Pyruvate kinase (PK) is the last rate-limiting enzyme in glycolysis. Four PK isozymes have been identified: M, K, L and R types. Aberrantly expressed pyruvate kinase M2 (PKM2) is common in tumor cells [20]. Many dysregulated circRNAs in cancer regulate PKM2 expressions. Thus, elucidation of the mechanisms involved in PKM2 functions and dysregulation in thyroid cancer may provide new directions for PTC therapy [21, 22].Traditionally, epigenetic regulation refers to DNA- or histone-associated chemical modifications that regulate gene expressions, independently of changes in genome sequence [23]. While N6-methyladenosine (m6A), which was first reported in the 1970s is the most prevalent internal chemical modification associated with eukaryotic mRNAs and ncRNAs [24, 25], it affects many steps in mRNA metabolism, including RNA processing, nuclear export, translation, degradation and RNA-protein interactions [26, 27]. This is a reversible chemical process that is dynamically regulated by balanced activities of m6A methyltransferases and demethylases. The modification of m6a has an important role in regulating the onset of cancer cell progression [28–31]. However, the significance of m6a modification in PTC and its potential regulatory mechanisms for circRNAs have not been fully explored.

In this study, we stably downregulated circNRIP1 in PTC tissues as well as cell lines and analyzed the potential relationship between circNRIP levels and clinic-pathological features of PTC. CircNRIP enhanced glycolysis in PTC cells by upregulating PKM2 levels and via sponging miR-541-3p and miR-3064-5p. However, these effects were inhibited by the demethylase, ALKBH5. Our findings elucidate on the mechanisms through which m6a modifications regulate PTC glycolysis and ceRNAs.

Materials And Methods

Human specimens and cell lines

A total of 102 pairs of PTC tissues and adjacent non-cancerous tissues were obtained between 2019 and 2020 from patients undergoing surgery at the First Hospital of China Medical University. Patient tissue samples consisted of PTC tissues and adjacent non-cancerous tissues. Formalin-fixed and paraffin-embedded (FFPE) tissue samples were subjected to immunohistochemistry analysis, while fresh frozen tissue samples were used for qRT-PCR and western blotting. Samples were confirmed by independent examinations by two histopathologists. None of the patients had received preoperative local or systemic therapy. This study was performed in accordance with ethical standards of the Research Ethics Committee of the First Hospital of China Medical University and the 1964 Declaration of Helsinki and its subsequent amendments. All study participants provided written informed consents. The human thyroid follicular epithelial cell line (Nthy-ori3-1) and PTC cell lines (TPC1, K1, IHH4 and BCPAP) were used in this study. The sources and culture methods for these cell lines are described in Additional File 2: Supplementary Materials and Methods.

Cell transfection and lentivirus infection

CircNRIP1, ALKBH5, WTAP, YTHDF1, FTO, METTL3, METTL14 and NC siRNAs were obtained from GenePharma (Suzhou, China) and their sequences are shown in Additional file 1 (Table S2). Lipofectamine 3000 (Invitrogen, Waltham, MA, USA) was used to perform transfections, according to the manufacturer's instructions. A recombinant lentivirus containing sh-circNRIP1 or sh-ALKBH5 was established to stably knock out circNRIP1 or ALKBH5. A non-targeting shRNA (sh-NC) was used as the negative control. circNRIP1 or ALKBH5 were overexpressed using recombinant lentiviruses containing the complete coding sequences of these genes.

An empty lentivirus (vector) was used as the negative control. The lentiviral vectors were constructed by Obio Technologies (Shanghai, China). Infected cells were selected using puromycin. Then, qRT-PCR and immunoblotting were performed to assess the transfection and infection efficiencies.

Seahorse metabolic analysis

Respectively, the SeahorseXF Glycolytic Stress Test Kit and the SeahorseXF Cell Aqueous Stress Test Kit (Agilent Technologies, Palo Alto, CA) were used to determine the extracellular acidification rate (ECAR) and oxygen consumption rate (OCR). Transfected PTC cells were seeded in a 96-well cell culture plate dedicated to the hippocampus overnight. Then, glucose, oligomycin, and 2-deoxyglucose (2-DG) were sequentially added to the corresponding wells on the sensor cartridge for ECAR measurements while oligomycin, FCCP, antimycin A and rotenone were sequentially added to the corresponding wells for the OCR assay. The prepared cell plates were analyzed by the hippocampal XFe96 analyzer while the hippocampal wave software was used for data analysis.

Glucose uptake and lactate production

Glucose uptake was determined using a glucose colorimetric assay kit (Biovision, USA), according to the manufacturer's protocol. Briefly, transfected PTC cells were seeded in 6-well plates and incubated for 48 h. Specifically, the glucose enzyme mixture oxidizes glucose to produce a product (OD value 570 nm) that reacts with the dye that results in color generation. Glucose levels in the cell culture medium were evaluated after which cells in each well were counted to normalize glucose levels. Indirectly, glucose uptake was determined by measuring the amounts of glucose remaining in the cell culture medium. The lactate colorimetric assay kit (Biovision, USA) was used to determine lactate levels, according to the manufacturer's protocol. Exclusively, lactate reacts with the enzyme mixture to produce a product that interacts with the lactate probe to generate a color (OD value 570 nm). Lactate production in the cell culture medium was detected and cells in each well counted to normalize lactate concentration.

Luciferase reporter assay

The circNRIP1 fragment with putative binding sites for miR-541-5p and miR-3064-5p was PCR-amplified and cloned into a firefly luciferase expression vector, pmiR-REPORT (Obio Technol ogy), which was named circNRIP1-Wt. To mutate the miR-541-5p and miR-3064-5p putative binding sites in circNRIP1, putative binding site sequences were replaced and named circNRIP1-Mut1/Mut2. The day before transfection, cells were seeded in 96-well plates and thereafter transfected with pRL-TK (Promega, Madison) with pmiR-REPORT-circNRIP1-Wt, pmiR-REPORT-circNRIP1-Mut1/Mut2 reporter vector, and Renilla luciferase expression vector, WI, USA) and miR-541-5p mimic, miR-3064-5p mimic, or NC using the Lipofectamine 3000 reagent (Invitrogen). Moreover, PKM2-3'UTR-Wt and PKM2-3'UTR-Mut were established and cloned into the firefly luciferase expression vector, pmiR-REPORT (Obio Technology). The day before transfection, cells were seeded in 96-well plates and treated with the pmiR-REPORT-PKM2-30 UTR-WT and pmiR-REPORT-PKM2-30 UTR-Mut reporter vectors, as well as the Renilla fluorescein-erased expression vector pRL -TK and miR-541-5p mimic, miR-3064-5p mimic, or NC using the Lipofectamine 3000 reagent. After 48 h, cells were harvested and measured for firefly and Renilla luciferin enzymatic activities using a dual luciferase reporter assay system (Promega).

A circNRIP1 fragment with a wild-type m6a motif (circNRIP1-Wt) and a mutant m6A motif (m6A was replaced by C, circNRIP1-Mut) was subcloned into a pmir-reported firefly luciferase vector (Obio Technologies, Shanghai, China) Chinese companies). Using the kidney luciferase expression vector, pRL-TK (Promega, Madison, WI, USA), wild-type and mutant circNRIP1 luciferase plasmids were co-transfected with empty, sh-ALKBH5 and ALKBH5 plasmids. Then, the dual luciferase reporter gene assay system (Promega, Madison, USA) was used to determine relative luciferase activities. Firefly luciferase activities normalized to firefly luciferase activities were then determined.

MeRIP assays

The RNAisoPlus (Takara, Japan) was used for total RNA extraction, after which DNAse was added to remove DNA. The m6aRNA enrichment kit (Epigentek, USA) was used to detect MeRIP, according to the manufacturer's instructions. The m6a-containing target fragment was pulled down using a bead-bound m6a capture antibody, and the RNA sequence containing both ends of the m6a region cleaved using a lyase cocktail. Enriched RNA was released, purified and eluted. After MeRIP, qRT-PCR was conducted to quantify changes in target gene m6a methylation.

Nude mice xenograft models and 18FFDG PET imaging

To construct the xenograft tumor models, five-week-old female athymic BALB/c nude mice were purchased from Beijing HFK Bioscience Co., Ltd. (Beijing, China). Transfected TPC1 cells were injected into the subcutaneous tissues of mice. After 4 weeks, mice were euthanized to excise and measure the tumors. Tumor volumes were determined as: tumor volume (mm3) = long diameter × short diameter 2/2. The day before euthanasia, mice were fasted for 8 h, anesthetized using 1% sodium pentobarbital after which about 200–300 µCi18F-FDG was injected via the lateral tail vein. Mice were kept in cages for 1 h at room temperature. Then, micro-PET and micro-CT imaging were performed in the prone position on the examination bed. 18F-FDG uptake was quantified by plotting regions of interest (ROI) and plotting maximum normalized uptake values (SUVmax) using the MetisViewer software. All animal experiments were in accordance with the Guidelines of the Institutional Animal Care and Use Committee of China Medical University.

Statistical analysis

The SPSS 26.0 (IBM, Chicago, USA) and GraphPad Prism 8.3.0 (San Diego, USA) softwares were used for data analyses. Data are shown as mean ± standard deviation (SD) for n = 3. Between and among group comparisons of means were performed using the Student's t-test or analysis of variance, respectively. Wilcoxon signed-rank test was used to assess differences in relative expression levels of circNRIP1, hsa-miR-541-5p, hsa-miR-3064-5p, PKM2 and ALKBH5 in PTC tissues and in adjacent non-cancerous tissues. Correlations between circNRIP1 levels and clinic-pathological characteristics of patients were determined using the Chi-square test. p < 0.05 was set as the threshold for statistical significance.

Results

Elevated circNRIP1 levels in PTC correlated with higher TNM stage and tumor size in PTC patients

To investigate the role of circRNAs in PTC, differential expressions of circRNAs in PTC tissues and normal thyroid tissues were determined using the GEO database. Expressions of 14 circRNAs in the two GEO datasets (GSE93522, and GSE171011) significantly differed, when compared to normal tissues (Fig. 1a). Heat map analysis revealed that among these genes, hsa-circ-0004771 (circNRIP1) had the largest expression difference, Fold change = 3.13 p < 0.01 (Fig. 1b), while qRT-PCR analysis showed that circNRIP1 levels in 102 PTC tissues were higher, relative to those of circNRIP1. Expression levels in paired adjacent non-cancerous tissues are shown in Fig. 1c and Figure S1a. Elevated circNRIP1 levels correlated with higher tumor-node-metastasis (III/IV) (TNM) stage (p = 0.04) and tumor size (p = 0.04) in PTC patients (Table 1). qRT-PCR analysis revealed that circNRIP1 levels were significantly high in patients with higher TNM stages and larger tumor volumes (Figure S1c and Figure S1d). Consistent with the PTC tissue sample data, circNRIP1 levels in PTC cell lines (TPC1, BCPAP, K1, IHH4) were significantly higher than in the normal thyroid cell line, Nthy-ori-3-1 (Fig. 1d). circNRIP1 was characterized in TPC1 cells by RT-PCR, Sanger sequencing (Fig. 1e), and RNase R treatment (Fig. 1f). After RT-PCR analysis using different primers, sequenced PCR products corresponded to 5' exon 3 to 3' exon 2 (Fig. 1e). Resistance experiments to RNase R exon nuclease digestion confirmed that this RNA species has a circular RNA structure (Fig. 1f). After treatment with actinomycin D, a transcription inhibitor, qRT-PCR analysis revealed that the half-life of circNRIP1 was over 24 h, while the half-life of the related linear transcript was about 4 h (Fig. 1g), indicating that circNRIP1 was more stable in PTC cells. Nucleocytoplasmic cell isolation and fluorescence in situ hybridization (FISH) analysis revealed that circNRIP1 was mainly localized in the cytoplasms of TPC1 and BCPAP cells (Fig. 1h, i and Figure S1e, f). Collectively, these results suggest that circNRIP1 is a circRNA that is stably expressed in PTC cytoplasms.

Table 1

Correlation Between circNRIP1 Expression and Clinicopathological Features in PTC Tissues (n = 102)

Characteristic

n

circNRIP1

P Value

High Expression(%)

Low Expression(%)

Gender

        

Male

39

18(51.4)

21(48.6)

0.393

Female

63

33(49.3)

30(50.7)

  

Age, years

        

< 55

60

33(55.0)

27(45.0)

0.227

≥ 55

42

18(42.9)

24(57.1)

  

Tumor size, cm

        

< 2cm

55

22(40.0)

33(60.0)

0.03*

≥ 2cm

47

29(61.7)

18(38.3)

  

Extrathyroidal

extension

        

Yes

35

25(71.4)

10(28.6)

0.297

No

67

26(38.8)

41(61.2)

  

LNM

        

Yes

64

38(64.4)

26(35.6)

0.16

No

38

13(30.2)

25(69.8)

  

Multifocality

        

Yes

32

17(53.1)

15(46.9)

0.67

No

70

34(48.6)

36(51.4)

  

TNM staging

        

Ⅰ-Ⅱ

86

45(52.3)

41(47.7)

0.04*

Ⅲ-Ⅳ

16

6(37.5)

10(62.5)

  

Hashimoto

thyroiditis

        

Yes

27

13(48.1)

14(51.9)

0.822

No

75

38(50.7)

37(49.3)

  

 

CircNRIP1 promotes glycolysis in PTC cells

Cell Counting Kit-8 (CCK-8) experiments revealed that circNRIP1 knockdown in TPC1 and BCPAP cells inhibited proliferation (Fig. 1j and Figure S1g). To explain the effects of circNRIP1 on proliferation, extracellular glucose analysis and extracellular lactate level analysis experiments were conducted. Glucose uptake and lactate production in PTC cells were significantly reduced by circNRIP1 suppression (Fig. 1k). After inhibition of circNRIP1, the OCR and ECAR of cells were tested by hippocampal metabolic analysis experiments. The ECAR value was significantly reduced in TPC1 and BCPAP cells, implying that glycolysis in PTC was inhibited, indicating that circNRIP1 suppression decreased glycolysis levels in post-PTC cells (Fig. 1l and Figure Sh). The OCR value was significantly increased, indicating that PTC mitochondrial oxidative phosphorylation was inhibited, and that circNRIP1 has a promoting effect on glycolysis (Figure Si, j). To investigate the regulatory mechanisms of circNRIP1 on glycolysis, circNRIP1 was knocked out. Then, levels of key glycolysis-related enzymes were evaluated. qRT-PCR analysis showed that circNRIP1 silencing suppressed the molecular levels of PKM2 in TPC1 and BCPAP cell lines (Fig. 1m, n). Analysis of the TCGA dataset revealed that the levels of PKM2, a key enzyme in glycolysis, was significantly upregulated in PTC tissues, compared to normal tissues (Fig. 2o). The relative expressions of PKM2 in 102 pairs of PTC tissues and matched adjacent noncancerous tissues were verified by qRT-PCR. Compared with matched adjacent non-cancerous tissues, PKM2 levels were higher in PTC tissues (Fig. 2p). Further, PKM2 protein levels in PTC tissue samples were detected by immunohistochemistry. PKM2 protein levels were significantly up-regulated in PTC tissues and were mainly localized in the cytoplasms of PTC cells (Fig. 2q). Spearman's correlation analysis revealed a positive correlation between circNRIP1 levels and PKM2 levels in PTC tissues (Fig. 2r).

CircNRIP1 promoted glycolysis by regulating the expressions of PKM2 via competitive binding to miR-541-5p and miR-3064-5p

Subcellular localization of circRNAs is closely related to their biological functions and potential molecular roles. As an important circRNA regulatory mechanism, the competing endogenous RNA (ceRNA) theory holds that circRNAs highly expressed in the cytoplasm can compete with miRNAs to bind and regulate downstream target genes. Therefore, it has been postulated that circNRIP1 regulate PTC cell functions through a ceRNA mechanism. The miRWalk (http://mirwalk.umm.uni-heidelberg.de/) and ENCORI (http://starbase.sysu.edu.cn/) databases revealed 12 miRNAs that may bind circNRIP1 and PKM2 via complementary base pairing (Fig. 2a). Next, expressions of candidate miRNAs after downregulation of circNRIP1 in TPC1 and BCPAP cells were evaluated by qRT-PCR (Figs. 2b and 2c), respectively. Among the candidate miRNAs, miR-541-5p, miR-3064-5p and miR-3140-5p may regulate PKM2 expressions. To confirm this, we overexpressed candidate miRNAs and examined PKM2 expressions in PTC cells. Western blotting showed that only overexpressions of miR-541-5p and miR-3064-5p could inhibit PKM2 expressions (Fig. 2d), suggesting that PKM2 may be a common target gene for miR-541-5p and miR-3064-5p. Then, we investigated the expressions and functions of miR-541-5p and miR-3064-5p. qRT-PCR revealed that miR-541-5p and miR-3064-5p levels in 98 PTC tissues were suppressed, relative to paired adjacent non-cancerous tissues (Figs. 2e and 2f). In PTC tissues, miR-541-5p and miR-3064-5p were negatively correlated with circNRIP1 and PKM2 (Figs. 3g and 3h). To verify the effects of miR-541-5p and miR-3064-5p on cell proliferation and glycolysis, PTC cells with low expressions of miR-541-5p and miR-3064-5p were constructed. Western blotting revealed that miR-541-5p and miR-3064-5p knockdown promoted PKM2 expressions (Fig. 2i) while CCK-8 analysis showed that miR-541-5p and miR-3064-5p knockdown promoted PTC cell proliferations (Fig. 2j and Figure S2a). Extracellular glucose and extracellular lactate level analyses experiments showed that both glucose uptake and lactate production were significantly elevated when miR-541-5p and miR-3064-5p were downregulated in PTC cells (Fig. 2k). Then, we performed hippocampal metabolic analysis experiments to evaluate OCR and ECAR values of TPC1 cells after inhibiting miR-541-5p and miR-3064-5p. The ECAR values of PTC cells were significantly elevated (Fig. 2l and S2c), while OCR values were significantly suppressed (Figure S2d, 2e). These findings imply that miR-541-5p and miR-3064-5p knockdown enhanced the activation of glycolysis in PTC cells.

CircNRIP1 can directly bind miR-541-5p and miR-3064-5p to regulate PKM2 expressions

Then, we determined whether circNRIP1 can function as a ceRNA by directly sponging miR-541-5p and miR-3064-5p. Bioinformatic predictions indicated that miR-541-5p and miR-3064-5p could bind circNRIP1 and directly target the 3′UTR of PKM2. The dual-luciferase reporter gene assays showed that up-regulation of miR-541-5p significantly reduced the luciferase activities of PTC cells co-transfected with wild-type circNRIP1 (circNRIP1-Wt), while cells with mutant circNRIP1 (circNRIP1-Mut1) and upregulation of miR-541-5p did not affect changes in luciferase activities when co-transfected (Fig. 3a). Co-transfection of the miR-3064-5p mimic and circNRIP1-WT significantly reduced luciferase activities, whereas co-transfection of the miR-3064-5p mimic and circNRIP1-Mut2 did not alter luciferase activities (Fig. 3b), indicating that circNRIP1 directly binds miR-541-5p and miR-3064-5p. Then, a dual-luciferase reporter system was used to determine whether PKM2 is a direct target of miR-541-5p and miR-3064-5p. Compared to the NC group, miR-541-5p overexpressions significantly inhibited PKM2-3'UTR-WT luciferase activities, while the activity of PKM2-3'UTR-Mut remained unaffected (Fig. 3c). Overexpressions of miR-3064-5p significantly inhibited PKM2-3'UTR-Wt luciferase activities, but did not affect the activities of PKM2-3'UTR-Mut (Fig. 3d). Therefore, in PTC cells, miR-541-5p and miR-3064-5p specifically bound PKM2-3'UTR. Another dual-luciferase reporter assay was used to confirm whether circNRIP1 could regulate PKM2 by interacting with miR-541-5p and miR-3064-5p. It was found that circNRIP1-WT significantly increased the luciferase activities of WtPKM2; however, circNRIP1-Mut did not affect the luciferase activities of WtPKM2 (Figs. 3e and 3f). Our results suggest that circNRIP1 regulates the expressions of PKM2 as a ceRNA by sponging miR-541-5p and miR-3064-5p.

Overexpressions of miR-541-5p and miR-3064-5p inhibited circNRIP1-induced glycolysis and proliferation in PTC cells

Rescue experiments showed that miR-541-5p and miR-3064-5p could restore the effects of circNRIP1 on PTC cell proliferation and glycolysis. Western blotting revealed that miR-541-5p and miR-3064-5p mimics significantly reversed the glycolysis of circNRIP1 and inhibited PKM2 expressions (Figs. 4a and 4b). The CCK-8 assay showed that miR-541-5p and miR-3064-5p could attenuate the promotion effects of circNRIP1 on TPC1 and BCPAP cell proliferations (Figs. 4c and 4d and Figures S2h and 2i). In addition, extracellular glucose and extracellular lactate level assays showed that miR-541-5p and miR-3064-5p antagonized the promoting effects of circNRIP1 in TPC1 and BCPAP cells to induce glucose uptake and lactate production (Fig. 4e-4h). miR-541-5p and miR-3064-5p also attenuated the glycolytic effects of circNRIP1 in PTC cells, as detected by the hippocampal metabolic analysis experiments (Figs. 4i and 4j and Figures S2j and 2K).

CircNRIP1 promoted in vivo tumor growth

To clarify whether circNRIP1 affects tumor growth in vivo, a control vector (pcDNA3.1), a circNRIP1 overexpression vector, a mixed vector of circNRIP1 overexpression and miR-541-5p mimic, and a mixed vector of circNRIP1 overexpression and miR-3064-5p mimic were stably transfected. Stained TPC1 cells were injected into the back of mice. From the 10th day after injection, the growths of subcutaneous tumors were observed every 3 days. The long and short diameters of tumors were measured using a vernier caliper and tumor volumes calculated. The xenograft tumor assay showed that the circNRIP1 overexpressed ov-circNRIP1 group had larger xenograft tumor volumes than the pcDNA3.1 group, ov-circNRIP1 + miR-541-5p mimic group, and ov-circNRIP1 + miR-3064-5p mimic group (Fig. 5a). Compared to the other three groups, tumor volumes and weights of the ov-circNRIP1 group were significantly increased, however, overexpressions of miR-541-5p and miR-3064-5 suppressed the tumor volumes and weights (Figs. 5b and 5c). The resected tumor mass showed that circNRIP1 overexpression increased circNRIP1 and PKM2 levels while decreasing miR-541-5p and miR-3064-5p levels (Fig. 5d). Consistent with in vitro observations, histochemical detection revealed that PKM2 protein levels were significantly up-regulated in the circNRIP1 overexpression group, while overexpressions of miR-541-5p and miR-3064-5 antagonized its protein level (Fig. 5e). To determine glucose uptake in mice tumors, changes in glycolysis in vivo were measured using 18F-FDGPET ([18F]-fluoro-2-deoxyglucose positron emission tomography). Glucose uptake in xenograft mice tumor models were significantly increased after circNRIP1 overexpressions, compared to the control group, however, miR-541-5p and miR-3064-5p antagonized this effect (Fig. 5f).

ALKBH5 inhibited PTC development by suppressing the expressions of circNRIP1

m6A has an important role in RNA methylation modification of circRNAs. Therefore, we used bioinformatics to assess the potential roles of m6A in thyroid cancer. The mRNA levels of methylases (METTL3, METTL14, and WTAP), demethylases (FTO and ALKBH5), and methylation recognition proteins in PTC tissues were significantly different compared to normal tissues (Figure Annex S3a and 3b). Subsequently, the major m6A methyltransferases were disrupted in TPC1 and BCPAP cells, respectively, and the expressions of circNRIP1 detected. In both cell lines, only ALKBH5 knockdown had an effect on expressions of circNRIP1 (Figs. 6a and 6b). To explore the significance of ALKBH5 downregulation in PTC tissues, the tissue sample size was increased to 102 pairs. qRT-PCR analysis showed that ALKBH5 levels in PTC tissues were, relative to paired adjacent non-cancerous tissues (Fig. 6c). Correlation analysis showed that ALKBH5 levels negatively correlated with circNRIP1 (Fig. 6d). ALKBH5 knockdown promoted cell proliferation in both TPC1 and BCPAP cell (Figs. 6e and S3c) and significantly increased glucose uptake as well as lactate production in both TPC1 and BCPAP cells (Fig. 6f and 6g). Similarly, we assessed OCR and ECAR values of PTC cells after ALKBH5 suppression via hippocampal metabolic analysis experiments. In TPC1 and BCPAP cells, ECAR values were significantly increased (Figs. 6h and S3d) while OCR values were significantly decreased (Figures S3f and 3g). These findings indicate that ALKBH5 suppression promotes PTC mitochondrial oxidative phosphorylation and has an inhibitory effect on glycolysis. Subsequently, we enriched the m6A-modified fragments using the m6A antibodies through the MeRIP assay. Then, we performed real-time quantitative PCR assays using primers containing the m6A-modified sequence of circNRIP1, as predicted by MeRIP-seq data. Agarose electrophoresis showed that ALKBH5 indeed passes m6A. The modification binds circNRIP1. Through MeRIP-qPCR experiments, it was found that m6A enrichment levels of circNRIP1 were significantly increased after ALKBH5 suppressions (Figs. 6i and S3g). Then, we constructed a luciferase reporter gene containing wild-type and mutant circNRIP1 to detect the effects of m6A modification on circNRIP1 levels. When the A in m6A modification site of circNRIP1 was mutated to C, overexpressions of ALKBH5 decreased while luciferase activities of wild-type and mutant circNRIP1 did not change (Figs. 6j and S3h). Similarly, after interfering with ALKBH5, luciferase activities of wild-type and mutant circNRIP1 increased significantly (Figs. 6j and S3h). There were no changes in activities (Figs. 6k and S3i), implying that ALKBH5 affects circNRIP1 expressions through m6A modifications.

ALKBH5 promoted tumor growth in vivo

To clarify whether ALKBH5 affects tumor growth in vivo, TPC1 cells stably transfected with ALKBH5 knockout vector and control vector (pcDNA3.1) were respectively injected into the backs of mice. From the 10th day after injection, the growth of subcutaneous tumors was observed every 3 days. The long and short diameters of tumors were measured using a vernier caliper, and tumor volumes calculated. The sh-ALKBH5 group xenograft tumors were larger than those of the NC vector (Fig. 6l). Compared to the NC vehicle group, tumor volumes and weights were significantly increased in the ALKBH5 knockout group (Figures S3j and S3k). Consistent with in vitro observations, PKM2 protein levels were upregulated in the sh-ALKBH5 group (Fig. 6M). Furthermore, ALKBH5 knockdown triggered a decrease in ALKBH5 levels and an increase in circNRIP1 levels (Figure S3l). By performing 18F-FDG micro PET imaging on the tumor-bearing nude mice and measuring the SUVmax value of tumors, we found that the SUVmax value of the tumor in the sh-ALKBH5 group was significantly higher than that of the pcDNA3.1 group (Fig. 6n), indicating that ALKBH5 knockdown promotted the uptake of glucose in PTC xenografts. Finally, the results showed that ALKBH5 regulates PKM2 by modifying circNRIP1 via competing for the sponge actions of miR-541-5p and miR-3064-5p, thereby affecting the glycolytic functions of PTC cells (Fig. 6o).

Discussion

Circular RNAs (circRNAs) are a novel type of non-coding small RNAs (ncRNAs) with covalently closed loop structures [32]. It is generated by exon skipping or back splicing and has no 5'-3' polarity or polyadenylation tail [33, 34]. In the past 20 years, a large number of exonic and intronic circular RNAs have been found in eukaryotes [35]. This indicates that circular RNAs are not by-products of aberrant splicing, but have multiple potential biological functions. Further, circRNAs are more stable and durable as compared with linear RNAs because they lack free ends for RNase-mediated degradation [36]. In addition, accumulating evidence suggests that circRNAs can promote tumor progression in different types of cancers [37–39]. However, the progression of circRNAs in thyroid cancer is still not clear.

The present study identified circ-NRIP1 whose host gene is NRIP1, through GEO database data and RT-qPCR analysis. Further, the circ-NRIP1was up-regulated in both PTC tissues and cell lines, indicating that it may be involved in the occurrence and development of PTC. Previous studies have shown that circNRIP1 can promote migration and invasion of cancer cells by sponging miR-629-3p and regulating the PTP4A1/ERK1/2 pathway in cervical cancer [40], as well as promote the esophageal squamous cell through the miR-339-5p/CDC25A axis, progression of squamous cell carcinoma [41], or by targeting the ZEB2 signaling pathway to activate miR-653, and hence inhibit proliferation of breast cancer as well as induce apoptosis [42]. However, to date, there is no studies that have explored the function of CircNRIP1 in thyroid cancer.

Therefore, the present study examined the TCGA-THCA (TCGA Thyroid Cancer Data Collection) database and found that CircNRIP was significantly elevated in thyroid cancer tissues as compared with normal noncancerous tissues. Up-regulated expression of CircNRIP was also observed in 102 PTC tissues relative to adjacent noncancerous tissues. More significantly, it was demonstrated that circNRIP1 is a circular RNA and its higher relative expression is associated with poorer prognosis including higher TNM stage and larger tumor volume, suggesting that circNRIP has a promoting role in progression of PTC. Therefore, results of the previous studies have suggested that circNRIP1 may be involved in the malignant progression of PTC.

On the contrary to what is observed in the surrounding tissue, cancer cells absorb large amounts of glucose to produce lactate even under aerobic conditions. This phenomenon is known as "aerobic glycolysis" or the Warburg effect [43]. Glycolysis is increasingly used as a marker of tumor progression because during glucose metabolism, large amounts of lipids, proteins, and nucleotides are produced which help to accelerate the proliferation and division of cancer cells [44, 45].

The key to control of glycolysis or oxidative phosphorylation is regulation of glycolytic flux through glycolytic enzymes. Among several key enzymes in glycolysis, pyruvate kinase catalyzes the final reaction of glycolysis by transferring high-energy phosphate from phosphoenolpyruvate to ADP for generation of ATP and pyruvate. There are four isoforms of enzyme pyruvate kinase and the M2 isoform (PKM2) is the predominant type of the enzyme in proliferation of cancer cells [46]. When PKM2 is overexpressed, the rate of glycolysis is high, most of the glucose is converted to lactate and ATP is rapidly produced [47].

It has been previously demonstrated that PKM2 can be regulated by circRNAs to promote tumor glycolysis to accelerate tumor progression. For instance, exosome-delivered hsa_circ_0005963 promotes glycolysis through the miR-122-PKM2 axis to induce chemoresistance in colorectal cancer [48]. Circular RNA MAT2B promotes glycolysis and malignancy in hepatocellular carcinoma through the miR-338-3p/PKM2 axis under hypoxic stress [49]. Further, CircATP2B1 promotes aerobic glycolysis in gastric cancer cells by regulating PKM2 [50]. However, the current studies have not explored the regulation of PKM2 with circNRIP1.

Previous experiments have confirmed that circNRIP1 can promote the proliferation and glycolysis of PTC cells. To further study the downstream target proteins of circNRIP1, the present study focused on the detection of glycolysis-related genes and found that PKM2 was significantly decreased in the si-circNRIP1 group in both TPC1 and BCPAP cell lines. In addition, tissue expression assays have shown that PKM2 was expressed in PTC tissues, down-regulated and positively correlated with circNRIP1. The previous studies suggest that circNRIP1 is involved in tumor progression by regulating PKM2.

Furthermore, some previous studies have shown that circRNAs can regulate gene expression by acting as ceRNAs [51, 52]. In the present study, it was found that circNRIP1 is mainly located in the cytoplasm of PTC cells. Therefore, it is speculated that circNRIP1 may also function as a ceRNA to regulate PKM2 expression by sponging miRNAs in PTC. Bioinformatics analysis and target prediction tools in the present study identified miRNAs that could target not only PKM2 but also the circNRIP1 binding sites.

Initially, 12 miRNAs were predicted to interact with circNRIP1 and PKM2. Using qRT-PCR to detect the expression of miRNA in PTC cells with up- or down-regulation of circNRIP1, it was found that miR-541-5p, miR-3064-5p, and miR-3140-5p may be the candidate miRNAs. Results of Western blotting indicated that only miR-541-5p and miR-3064-5p could downregulate the levels of PKM2. Previous studies have also shown that miR-541-5p is involved in the progression of hepatocellular carcinoma and intrahepatic cholangiocarcinoma [53, 54], whereas miR-3064-5p is involved in the development of colorectal cancer [55], cervical cancer [56], and gastric cancer [57], Therefore, it was speculated that miR-541-5p and miR-3064-5p are also involved in thyroid cancer whereas PKM2 is a common target gene for both miR-541-5p and miR-3064-5p.

Subsequently, results of the present study found that miR-541-5p and miR-3064-5p were significantly under-expressed in 98 pairs of PTC tissues and matched normal tissues. Further, results of the correlation analysis showed that the levels of circNRIP1 were negatively correlated with miR-541-5p and miR-3064-5p.

Functional experiments in the current study showed that miR-541-5p and miR-3064-5p can promote proliferation of PTC cells and promote glycolysis. Finally, a rescue strategy was used to confirm that both miR-541-5p and miR-3064-5p can antagonize the proliferative and glycolytic effects of circNRIP1 on the PTC cells, both in vivo and in vitro.

The dual-luciferase reporter assays was then performed in the PTC cells to confirm whether miR-541-5p and miR-3064-5p could directly bind to circNRIP1 and PKM2 3’UTRs. Results of the assays showed that miR-541-5p and miR-3064-5p are important miRNAs that binds to circNRIP1 and PKM2 3'UTR whereas the circNRIP1 contains the binding sites for miR-541-5p and miR-3064-5p. In addition, the obtained results revealed that circNRIP1 promotes glycolysis in PTC cells through the miR-541-5p/PKM2 and miR-3064-5p/PKM2 axes, respectively.

The RNA n6-methyladenosine (m6A) is considered to be a new epigenetic regulatory layer. This biochemical process regulates cell growth, differentiation, and self-renewal by controlling RNA splicing, translation, and stability [58–60]. Previous studies have also found that m6a is an abundant co-transcriptional modification of mRNA [61–63] and is involved in many aspects of post-transcriptional mRNA metabolism [64–66]. However, little is known about biological impact of the m6a in modification of cellular circular RNA. However, recent studies have found that m6a is also present in the modification of ncRNAs, including circular RNAs [67]. In addition, biogenesis of different aspects of circular RNAs has been extensively studied. Further, it is thought to be a co-transcription product resulting from canonical linear mRNA splicing that occurs in the nucleus [68]. However, most circular RNAs are localized in the cytoplasm and hence it is crucial to study the underlying mechanisms that regulate their export from the nucleus to the cytoplasm.

Given that circNRIP1 which is localized in the cytoplasm can promote the growth of thyroid cancer and possesses an m6A site. Therefore, it was speculated that m6A may regulate the biological behavior of PTC cells by rationally inducing the demethylation of circNRIP1. To the best of our knowledge, the current study is the first to demonstrate that circNRIP1 is a direct downstream target of ALKBH5-mediated m6A modification, revealing the mechanism by which ALKBH5 can manipulate circNRIP1 to regulate cell proliferation, and glycolytic capacity.

In the present study, it was noted that among a variety of major M6a key genes in TPC1 and BCPAP cells, only ALKBH5 was down-regulated whereas circNRIP1 was significantly changed, and ALKBH5 was down-expressed in PTC relative to normal thyroid tissue. Results of the experiments conducted in the current study demonstrated that ALKBH5 can inhibit proliferation and glycolysis in PTC cells, suggesting an inhibitory role of ALKBH5 in PTC tumorigenesis. In addition, results of Spear man's correlation analysis showed that ALKBH5 was negatively correlated with circNRIP1. Furthermore, results of MeRIP experiments demonstrated that circNRIP1 contains an m6A site, whereas luciferase experiments demonstrated that ALKBH5 can modify the expression of circNRIP1, thereby affecting the tumor growth and glycolytic function of PTC in vitro and in vivo.

Conclusion

Based on the described experimental results of the present study, it is evident that down-regulation of ALKBH5 activated circNRIP1–promoted glycolytic function of PTC cells through sponging of oncogenic miR-541-5p and miR-3064-5p as well as upregulating the expression levels of PKM2. Therefore, it is evident that circNRIP1 can be a promising biomarker and therapeutic target in PTC tissues.

Abbreviations

circRNAs: circular RNAs; TNM:tumor lymph node metastasis; RAI: radioactive iodine; TSH: thyroid-stimulating hormone; DTC: Differentiated thyroid cancer; FTO: Fat-mass and obesity-associated protein; PKM2: pyruvate kinase M2; FFPE: Formalin-fixed and paraffin-embedded; FISH: fluorescence in situ hybridization; CCK-8: Cell Counting Kit-8 ; ceRNA: competing endogenous RNA; m6A: N6-methyladenosine; PTC: Papillary thyroid carcinoma; ALKBH5: alkylation repair homolog protein 5; qRT-PCR: Quantitative real-time PCR; RNA-seq: Transcriptome-sequencing; MeRIP-seq: Methylated RNA immunoprecipitation sequencing; APOE: Apolipoprotein E; METTL3: Methyltransferase-like 3; METTL14: Methyltransferase-like 14; GEO: Gene Expression Omnibus; WTAP: Wilms tumor 1 associated protein; IGF2BP: Insulin-like growth factor 2 mRNA-binding proteins; BAG3: Bcl-2-associated athanogene 3; HK-II: Hexokinase II; AD: Alzheimerʼs disease; ECAR: Extra cellular acidification rate; mRNA: Messenger RNA; OCR: Oxygen consumption rate; RIP: RNA immunoprecipitation; ROI: Region of interest; SUVmax: Maximum standardized uptake values; TCGA: The Cancer Genome Atlas; 2-DG: 2-deoxy-glucose; 18F-FDG PET: 18F-fluoro-2-deoxyglucose positron emission tomography; FDR: False discovery rate; siRNA: Short interfering RNA.

Declarations

Ethics approval and consent to participate

The research was approved by the Ethics Committee of the First Hospital of China Medical University. All animal studies were conducted in accordance with the principles and procedures outlined in the guidelines of the Institutional Animal Care and Use Committee (IACUC) of China Medical University

Consent for publication

All authors give consent for the publication of manuscript in Journal of Experimental & Clinical Cancer Research.

Availability of data and materials

The data used during the current study are available from the corresponding author on reasonable request. The RNA-seq and MeRIP-seq data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus and are accessible through GEO Series accession number GSE93522 and GSE171011.

Competing interests

The authors declare that they have no competing interests.

Funding

National Natural Science Foundation of China (81902726), the Natural Science Foundation of Education Bureau of Liaoning Province (QNZR2020002), Natural Science Foundation of Education Bureau of Liaoning Province (QNZR2020009), the Natural Science Foundation of Liaoning Province (grant number 2020-MS-143), and the Natural Science Foundation of Liaoning Province (2020-MS-186), the Natural Science Foundation of Liaoning Province (grant number 2021-MS-193) and Science and Technology Project of Shenyang City (21-173-9-31).

Authors' contributions

XJ and WS designed this study, performed the statistical analysis, and drafted the manuscript; XJ, WS, JH and CL carried out the experiments. JH and CL analyzed the data; XJ and WS provided clinical samples as well as clinical information; and JH and CL supervised the study. All authors read and gave final approval of the manuscript.

Acknowledgements

This project was funded by the National Natural Science Foundation of China (81902726), the Natural Science Foundation of Education Bureau of Liaoning Province (QNZR2020002), Natural Science Foundation of Education Bureau of Liaoning Province (QNZR2020009), the Natural Science Foundation of Liaoning Province (grant number 2020-MS-143), and the Natural Science Foundation of Liaoning Province (2020-MS-186), the Natural Science Foundation of Liaoning Province (grant number 2021-MS-193) and Science and Technology Project of Shenyang City (21-173-9-31). We thank Freescience for its linguistic assistance during the preparation of this manuscript.

References

  1. Zhu X, Yao J, Tian W. Microarray technology to investigate genes associated with papillary thyroid carcinoma. MOL MED REP. 2015;11(5):3729–33.
  2. Al-Salamah SM, Khalid K, Bismar HA. Incidence of differentiated cancer in nodular goiter. SAUDI MED J. 2002;23(8):947–52.
  3. Bray F, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2018;68(6):394–424.
  4. Brose MS, Cabanillas ME, Cohen EE, et al. Vemurafenib in patients with BRAF(V600E)-positive metastatic or unresectable papillary thyroid cancer refractory to radioactive iodine: a non-randomised, multicentre, open-label, phase 2 trial. LANCET ONCOL. 2016;17(9):1272–82.
  5. Mazzaferri EL, Kloos RT. Clinical review 128: Current approaches to primary therapy for papillary and follicular thyroid cancer. J Clin Endocrinol Metab. 2001;86(4):1447–63.
  6. Shaha AR. Recurrent differentiated thyroid cancer. ENDOCR PRACT. 2012;18(4):600–3.
  7. Haddad RI, Nasr C, Bischoff L, et al. NCCN Guidelines Insights: Thyroid Carcinoma, Version 2.2018. J Natl Compr Canc Netw. 2018;16(12):1429–40.
  8. Llamas-Olier AE, Cuéllar DI, Buitrago G. Intermediate-Risk Papillary Thyroid Cancer: Risk Factors for Early Recurrence in Patients with Excellent Response to Initial Therapy. THYROID. 2018;28(10):1311–7.
  9. Fröhlich E, Wahl R. The current role of targeted therapies to induce radioiodine uptake in thyroid cancer. CANCER TREAT REV. 2014;40(5):665–74.
  10. Cocquerelle C, Mascrez B, Hétuin D, et al. Mis-splicing yields circular RNA molecules. FASEB J. 1993;7(1):155–60.
  11. Salzman J, Gawad C, Wang PL, et al. Circular RNAs are the predominant transcript isoform from hundreds of human genes in diverse cell types. PLoS ONE. 2012;7(2):e30733.
  12. Jeck WR, Sorrentino JA, Wang K, et al. Circular RNAs are abundant, conserved, and associated with ALU repeats. RNA. 2013;19(2):141–57.
  13. Zhang L, Zhou Q, Qiu Q, et al. CircPLEKHM3 acts as a tumor suppressor through regulation of the miR-9/BRCA1/DNAJB6/KLF4/AKT1 axis in ovarian cancer. MOL CANCER. 2019;18(1):144.
  14. Li Z, Ruan Y, Zhang H, et al. Tumor-suppressive circular RNAs: Mechanisms underlying their suppression of tumor occurrence and use as therapeutic targets. CANCER SCI. 2019;110(12):3630–8.
  15. Stacpoole PW. Therapeutic Targeting of the Pyruvate Dehydrogenase Complex/Pyruvate Dehydrogenase Kinase (PDC/PDK) Axis in Cancer. J Natl Cancer Inst 2017; 109 (11).
  16. WARBURG O. On the origin of cancer cells. Science. 1956;123(3191):309–14.
  17. Warburg O, Wind F, Negelein E. THE METABOLISM OF TUMORS IN THE BODY. J GEN PHYSIOL. 1927;8(6):519–30.
  18. Reitman ZJ, Yan H. Isocitrate dehydrogenase 1 and 2 mutations in cancer: alterations at a crossroads of cellular metabolism. J Natl Cancer Inst. 2010;102(13):932–41.
  19. Teoh ST, Lunt SY. Metabolism in cancer metastasis: bioenergetics, biosynthesis, and beyond. Wiley Interdiscip Rev Syst Biol Med 2018; 10 (2).
  20. Li Q, Chen P, Zeng Z, et al. Yeast two-hybrid screening identified WDR77 as a novel interacting partner of TSC22D2. Tumour Biol. 2016;37(9):12503–12.
  21. Li Q, Pan X, Zhu D, et al. Circular RNA MAT2B Promotes Glycolysis and Malignancy of Hepatocellular Carcinoma Through the miR-338-3p/PKM2 Axis Under Hypoxic Stress. HEPATOLOGY 2019; 70 (4): 1298–316.
  22. Yook JI, Li XY, Ota I, et al. Wnt-dependent regulation of the E-cadherin repressor snail. J BIOL CHEM. 2005;280(12):11740–8.
  23. Suzuki MM, Bird A. DNA methylation landscapes: provocative insights from epigenomics. NAT REV GENET. 2008;9(6):465–76.
  24. Beemon K, Keith J. Localization of N6-methyladenosine in the Rous sarcoma virus genome. J MOL BIOL. 1977;113(1):165–79.
  25. Chen-Kiang S, Nevins JR, Darnell JJ. N-6-methyl-adenosine in adenovirus type 2 nuclear RNA is conserved in the formation of messenger RNA. J MOL BIOL. 1979;135(3):733–52.
  26. Fu Y, Dominissini D, Rechavi G, et al. Gene expression regulation mediated through reversible mâ࿽¶A RNA methylation. NAT REV GENET. 2014;15(5):293–306.
  27. Lee Y, Choe J, Park OH, et al. Molecular Mechanisms Driving mRNA Degradation by m(6)A Modification. TRENDS GENET. 2020;36(3):177–88.
  28. Tang B, Yang Y, Kang M, et al. m(6)A demethylase ALKBH5 inhibits pancreatic cancer tumorigenesis by decreasing WIF-1 RNA methylation and mediating Wnt signaling. MOL CANCER. 2020;19(1):3.
  29. Yue B, Song C, Yang L, et al. METTL3-mediated N6-methyladenosine modification is critical for epithelial-mesenchymal transition and metastasis of gastric cancer. MOL CANCER. 2019;18(1):142.
  30. Chen M, Wei L, Law CT, et al. RNA N6-methyladenosine methyltransferase-like 3 promotes liver cancer progression through YTHDF2-dependent posttranscriptional silencing of SOCS2. HEPATOLOGY 2018; 67 (6): 2254-70.
  31. Cai X, Wang X, Cao C, et al. HBXIP-elevated methyltransferase METTL3 promotes the progression of breast cancer via inhibiting tumor suppressor let-7g. CANCER LETT. 2018;415:11–9.
  32. Ashwal-Fluss R, Meyer M, Pamudurti NR, et al. circRNA biogenesis competes with pre-mRNA splicing. MOL CELL. 2014;56(1):55–66.
  33. Jeck WR, Sorrentino JA, Wang K, et al. Circular RNAs are abundant, conserved, and associated with ALU repeats. RNA. 2013;19(2):141–57.
  34. Salzman J, Chen RE, Olsen MN, et al. Cell-type specific features of circular RNA expression. PLOS GENET. 2013;9(9):e1003777.
  35. Memczak S, Jens M, Elefsinioti A, et al. Circular RNAs are a large class of animal RNAs with regulatory potency. Nature. 2013;495(7441):333–8.
  36. Wesselhoeft RA, Kowalski PS, Anderson DG. Engineering circular RNA for potent and stable translation in eukaryotic cells. NAT COMMUN. 2018;9(1):2629.
  37. Yang F, Fang E, Mei H, et al. Cis-Acting circ-CTNNB1 Promotes β-Catenin Signaling and Cancer Progression via DDX3-Mediated Transactivation of YY1. CANCER RES. 2019;79(3):557–71.
  38. Yang Z, Qu CB, Zhang Y, et al. Dysregulation of p53-RBM25-mediated circAMOTL1L biogenesis contributes to prostate cancer progression through the circAMOTL1L-miR-193a-5p-Pcdha pathway. Oncogene. 2019;38(14):2516–32.
  39. Wang R, Zhang S, Chen X, et al. EIF4A3-induced circular RNA MMP9 (circMMP9) acts as a sponge of miR-124 and promotes glioblastoma multiforme cell tumorigenesis. MOL CANCER. 2018;17(1):166.
  40. Li X, Ma N, Zhang Y, et al. Circular RNA circNRIP1 promotes migration and invasion in cervical cancer by sponging miR-629-3p and regulating the PTP4A1/ERK1/2 pathway. CELL DEATH DIS. 2020;11(5):399.
  41. Huang E, Fu J, Yu Q, et al. CircRNA hsa_circ_0004771 promotes esophageal squamous cell cancer progression via miR-339-5p/CDC25A axis. EPIGENOMICS-UK. 2020;12(7):587–603.
  42. Xie R, Tang J, Zhu X, et al. Silencing of hsa_circ_0004771 inhibits proliferation and induces apoptosis in breast cancer through activation of miR-653 by targeting ZEB2 signaling pathway. Biosci Rep 2019; 39 (5).
  43. Liberti MV, Locasale JW. The Warburg Effect: How Does it Benefit Cancer Cells? TRENDS BIOCHEM SCI. 2016;41(3):211–8.
  44. Gatenby RA, Gillies RJ. Why do cancers have high aerobic glycolysis? NAT REV CANCER. 2004;4(11):891–9.
  45. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646–74.
  46. Wong N, Ojo D, Yan J, et al. PKM2 contributes to cancer metabolism. CANCER LETT 2015; 356 (2 Pt A): 184 – 91.
  47. Chaneton B, Gottlieb E. Rocking cell metabolism: revised functions of the key glycolytic regulator PKM2 in cancer. TRENDS BIOCHEM SCI. 2012;37(8):309–16.
  48. Wang X, Zhang H, Yang H, et al. Exosome-delivered circRNA promotes glycolysis to induce chemoresistance through the miR-122-PKM2 axis in colorectal cancer. MOL ONCOL. 2020;14(3):539–55.
  49. Li Q, Pan X, Zhu D, et al. Circular RNA MAT2B Promotes Glycolysis and Malignancy of Hepatocellular Carcinoma Through the miR-338-3p/PKM2 Axis Under Hypoxic Stress. HEPATOLOGY 2019; 70 (4): 1298–316.
  50. Zhao X, Tian Z, Liu L. circATP2B1 Promotes Aerobic Glycolysis in Gastric Cancer Cells Through Regulation of the miR-326 Gene Cluster. FRONT ONCOL. 2021;11:628624.
  51. Shyu AB, Wilkinson MF, van Hoof A. Messenger. RNA regulation: to translate or to degrade. EMBO J. 2008;27(3):471–81.
  52. Fabian MR, Sonenberg N, Filipowicz W. Regulation of mRNA translation and stability by microRNAs. ANNU REV BIOCHEM. 2010;79:351–79.
  53. Li D, Zhang J, Yang J, et al. CircMTO1 suppresses hepatocellular carcinoma progression via the miR-541-5p/ZIC1 axis by regulating Wnt/β-catenin signaling pathway and epithelial-to-mesenchymal transition. CELL DEATH DIS. 2021;13(1):12.
  54. Tang J, Wang R, Tang R, et al. CircRTN4IP1 regulates the malignant progression of intrahepatic cholangiocarcinoma by sponging miR-541-5p to induce HIF1A production. PATHOL RES PRACT. 2022;230:153732.
  55. Luo Z, Hao S, Yuan J, et al. Long non-coding RNA LINC00958 promotes colorectal cancer progression by enhancing the expression of LEM domain containing 1 via microRNA miR-3064-5p. BIOENGINEERED 2021; 12 (1): 8100–15.
  56. Wei M, Chen Y, Du W. LncRNA LINC00858 enhances cervical cancer cell growth through miR-3064-5p/ VMA21 axis. CANCER BIOMARK. 2021;32(4):479–89.
  57. Luo Z, Hao S, Yuan J, et al. Long non-coding RNA LINC00958 promotes colorectal cancer progression by enhancing the expression of LEM domain containing 1 via microRNA miR-3064-5p. BIOENGINEERED 2021; 12 (1): 8100–15.
  58. Dai D, Wang H, Zhu L, et al. N6-methyladenosine links RNA metabolism to cancer progression. CELL DEATH DIS. 2018;9(2):124.
  59. Geula S, Moshitch-Moshkovitz S, Dominissini D, et al. Stem cells. m6A mRNA methylation facilitates resolution of naïve pluripotency toward differentiation. Science. 2015;347(6225):1002–6.
  60. Du Y, Hou G, Zhang H, et al. SUMOylation of the m6A-RNA methyltransferase METTL3 modulates its function. NUCLEIC ACIDS RES. 2018;46(10):5195–208.
  61. Dai D, Wang H, Zhu L, et al. N6-methyladenosine links RNA metabolism to cancer progression. CELL DEATH DIS. 2018;9(2):124.
  62. Gilbert WV, Bell TA, Schaening C. Messenger. RNA modifications: Form, distribution, and function. Science. 2016;352(6292):1408–12.
  63. Li S, Mason CE. The pivotal regulatory landscape of RNA modifications. Annu Rev Genomics Hum Genet. 2014;15:127–50.
  64. Xiao W, Adhikari S, Dahal U, et al. Nuclear m(6)A Reader YTHDC1 Regulates mRNA Splicing. MOL CELL. 2016;61(4):507–19.
  65. Wang X, Lu Z, Gomez A, et al. N6-methyladenosine-dependent regulation of messenger RNA stability. Nature. 2014;505(7481):117–20.
  66. Wang X, Zhao BS, Roundtree IA, et al. N(6)-methyladenosine Modulates Messenger RNA Translation Efficiency. Cell. 2015;161(6):1388–99.
  67. Zhou C, Molinie B, Daneshvar K, et al. Genome-Wide Maps of m6A circRNAs Identify Widespread and Cell-Type-Specific Methylation Patterns that Are Distinct from mRNAs. CELL REP. 2017;20(9):2262–76.
  68. Ashwal-Fluss R, Meyer M, Pamudurti NR, et al. circRNA biogenesis competes with pre-mRNA splicing. MOL CELL. 2014;56(1):55–66.