RETRACTED ARTICLE: MIR600HG sponges miR-125a-5p to regulate glycometabolism and cisplatin resistance of oral squamous cell carcinoma cells via mediating RNF44

There is increasing evidence that dysregulated long non-coding RNA (lncRNA) is implicated in tumorigenesis and progression. We aim to explore the role of lncRNA MIR600HG in glycometabolism and cisplatin (DDP) resistance of oral squamous cell carcinoma (OSCC) cells via regulating microRNA-125a-5p (miR-125a-5p) and RING finger 44 (RNF44). Expression of MIR600HG, miR-125a-5p, and RNF44 in OSCC clinical samples, cell lines, and DDP-resistant OSCC cells (SCC-9/DDP) was determined. In SCC-9 cells, proliferation, IC50 value of DDP, migration, invasion, and apoptosis were detected; in SCC-9/DDP cells, proliferation, IC50 value of DDP, apoptosis, glucose consumption, and production of lactic acid and ATP were evaluated. The interaction of MR600HG, miR-125a-5p, and RNF44 was verified. MIR600HG and RNF44 were upregulated while miR-125a-5p was downregulated in OSCC tissues and cell lines, and also in SCC-9/DDP cells. In SCC-9 cells, MIR600HG overexpression improved cell growth, metastasis, and inhibited cell susceptibility to DDP; in SCC-9/DDP cells, silencing of MIR600HG promoted apoptosis, improved DDP sensitivity, and inhibited cell glycolysis. Downregulation of miR-125a-5p showed the opposite effect to downregulation of MIR600HG. MIR600HG bound to miR-125a-5p and miR-125a-5p targeted RNF44. Downregulation of miR-125a-5p reversed the improvement of DDP sensitivity and the inhibition of cell glycolysis by downregulated MIR600HG on SCC-9/DDP cells. Downregulating RNF44 reversed the promotion of DDP resistance and cell glycolysis of SCC-9/DDP cells mediated by downregulation of miR-125a-5p. Collectively, our study addresses that MIR600HG downregulation elevates miR-125a-5p and reduces RNF44 expression, thereby improving DDP sensitivity and inhibiting glycolysis in DDP-resistant OSCC cells.


INTRODUCTION
Oral squamous cell carcinoma (OSCC) is a metastatic and aggressive malignant oral malignancy with an increasing incidence [1]. Currently, although progresses have been made in the administration and diagnosis of OSCC, the survival rate of OSCC patients is still disappointing due to the high recurrence rate, later diagnosis at advanced stages, and a lack of early diagnostic biomarkers [2]. As a gene-related disease, the detailed molecular mechanism of OSCC tumorigenesis still remains to be further explored. Therefore, it is of importance to investigate diagnostic strategies and find therapeutic targets to improve the treatment of OSCC. Cisplatin (DDP) is an effective first-line chemotherapy drug for OSCC treatment, while its therapeutic effect often fails as a result of the drug resistance [3]. Tumor cells can take up more glucose, and glucose undergo aerobic glycolytic synthesis to generate more intermediate substances and eventually to generate lactic acid. These intermediate products such as NADPH and 5-phosphoribose are often the raw materials for the synthesis of nucleic acids and proteins [4], and these products can promote the development of tumor resistance [5]. Related reports have shown that tumor cells have highly active glucose metabolism levels which is related to chemotherapeutic drug resistance [6][7][8]. Thus, it is vital to identify the underlying mechanisms of chemoresistance to promote DDP sensitivity.
Long non-coding RNAs (lncRNAs) are important regulatory transcripts lacking protein-coding functions [9]. Previous studies have reported that dysregulated lncRNAs are involved in the occurrence and development of various tumors, including OSCC [10,11]. MIR600HG has been reported to suppress metastasis and enhance oxaliplatin chemosensitivity in colorectal cancer (CRC) [12]. Moreover, it has been demonstrated that MIR600HG is related to autophagy in OSCC [13] and acts as a prognostic lncRNA in pancreatic ductal adenocarcinoma (PDAC) [14]. We found through the bioinformatic analysis that there existed binding sites between MIR600HG and microRNA (miR)-125a-5p. miRs are small non-coding RNA molecules that modulate the genetic expression at the post-transcriptional level [15]. Importantly, miR-125a-5p has been revealed to affect the chemosensitivity to DDP in osteosarcoma cells [16], and it has also been unraveled that miR-125a-5p participates in the progression of head and neck squamous cell carcinoma [17] and esophageal squamous cell carcinoma [18]. Nevertheless, the role of miR-125a-5p in OSCC remains unknown. Through the bioinformatic analysis, we also found the binding sites between miR-125a-5p and RING finger 44 (RNF44). RNF44 is a novel E3 ligase that has been reported to be responsible for AMP-activated protein kinase-α1 degradation [19]. Its role in OSCC is still unknown.
We speculated that MIR600HG may sponge miR-125a-5p to affect DDP resistance in OSCC via regulating RNF44.

MIR600HG is upregulated in OSCC
To investigate whether MIR600HG participates in OSCC progression, we detected the expression of MIR600HG in OSCC tissues and normal tissues from 85 patients using quantitative reverse trabscription PCR (RT-qPCR). It was found that versus the normal tissues, MIR600HG was upregulated in OSCC tissues (Fig. 1A). MIR600HG expression in OSCC cell lines and normal human oral keratinocyte (NHOK) was assessed. The results indicated that it was upregulated in OSCC cell lines as well, compared to NHOK (Fig. 1B). MIR600HG expression in SCC-9/DDP cells and the DDP sensitivity were assessed. As expected, MIR600HG expression was increased (Fig. 1C) and DDP sensitivity was reduced in SCC-9/DDP cells (Fig. 1D). These data suggested that MIR600HG may be related to DDP resistance of OSCC cells.
MIR600HG affects growth, DDP sensitivity, and glycometabolism of OSCC cells MIR600HG was overexpressed through oe-MIR600HG transfection in SCC-9 cells while it was knocked down in SCC-9/DDP cells through si-MIR600HG transfection ( Fig. 2A, B). After transfection, colony formation assay and MTT assay were performed and it was found that overexpression of MIR600HG induced cell growth (Fig. 2C, D) and DDP resistance of SCC-9 cells (Fig. 2E). Migration and invasion were determined using Transwell assay and the results showed that MIR600HG overexpression facilitated migration and invasion of SCC-9 cells (Fig. 2F). Moreover, flow cytometry was utilized to assess cell apoptosis and it was discovered that the apoptosis rate of SCC-9 cells was reduced when MIR600HG was overexpressed (Fig. 2G).
For SCC-9/DDP cells, reduction of MIR600HG reduced cell viability and elevated apoptosis rate ( Fig. 2H-J). Furthermore, metabolism of SCC-9/DDP cells was determined, including production of lactic acid and adenosine triphosphate (ATP), as well as glucose consumption. The results indicated that silencing of MIR600HG reduced glucose consumption and production of lactic acid and ATP (Fig. 2K). Thus, it was concluded that MIR600HG could affect growth, DDP sensitivity, and glycometabolism of OSCC cells.
SCC-9/DDP cells were transfected with miR-125a-5p inhibitor, and the transfection efficacy was confirmed using RT-qPCR (Fig. 3M). Then, we found through a series of assays that inhibition of miR- Fig. 1 MIR600HG is upregulated in OSCC tissues, cells, and DDP-resistant cells. A MIR600HG expression in clinical samples was assessed using RT-qPCR (n = 85); B MIR600HG expression in NHOK and OSCC cell lines was assessed using RT-qPCR; C MIR600HG expression in SCC-9 and SCC-9/DDP cells was assessed using RT-qPCR; D IC 50 of SCC-9 and SCC-9/DDP cells was detected using MTT assay; N = 3; *P < 0.05; the results were exhibited as mean ± standard deviation; t-test or ANOVA was used to analyze data. 125a-5p promoted cell viability (Fig. 3N), inhibited DDP sensitivity (Fig. 3O), repressed apoptosis (Fig. 3P), increased the production of lactic acid and ATP, and the amount of glucose consumption (Fig. 3Q).
MIR600HG/miR-125a-5p axis regulates the proliferation, apoptosis, DDP sensitivity, and glycometabolism of DDPresistant OSCC cells To verify whether the functions of MIR600HG on DDP-resistant OSCC cells were correlated with miR-125a-5p, si-MIR600HG and miR-125a-5p inhibitor were transfected into SCC-9/DDP cells, and the transfection efficiency was determined using RT-qPCR ( RNF44 is upregulated in OSCC, and RNF44 affects growth, DDP sensitivity, and glycometabolism of OSCC cells RNF44 expression in OSCC clinical samples and cell lines was determined and we found that it was upregulated in OSCC tissues and cell lines (Fig. 5A, B), especially in SCC-9/DDP cells (Fig. 5C).

MIR600HG knockdown reduces the growth of DDP-resistant OSCC cells in vivo
The in vivo assay was performed to observe the role of MIR600HG in tumor growth. SCC-9/DDP cells that had been transfected with sh-MIR600HG were injected into the nude mice, and then the tumor growth on days 7, 14, 21, and 28 was assessed. The results suggested that DDP reduced the tumor weigh and volume, which were further inhibited by MIR600HG knockdown (Fig. 7A-C). Thus, we demonstrated that MIR600HG inhibition strengthens DDP sensitivity of OSCC cells in vivo.

DISCUSSION
OSCC is the sixth h most prevalent malignancy globally. It occurs due to the alteration in gene expression affected by specific Fig. 2 Effects of MIR600HG on growth, DDP sensitivity, and glycometabolism of OSCC cells. A MIR600HG expression in SCC-9 cells after transfection of oe-NC or oe-MIR600HG was assessed using RT-qPCR. B The knockdown effect after MIR600HG siRNA transfection was confirmed by RT-qPCR in SCC-9/DDP cells. C Colony formation ability of SCC-9 cells was determined using colony formation assay. D viability of SCC-9 cells was determined using MTT assay. E IC 50 of SCC-9 cells was determined using MTT assay. F Migration and invasion of SCC-9 cells was detected using Transwell assay. G Apoptosis of SCC-9 cells was detected using flow cytometry. H Viability of SCC-9/DDP cells was determined using MTT assay. I IC 50 of SCC-9/DDP cells was determined using MTT assay. J Apoptosis of SCC-9/DDP cells was detected using flow cytometry. K Glycometabolism of SCC-9/DDP cells was analyzed; N = 3; *P < 0.05; the data were expressed as mean ± standard deviation and t-test was used to analyze data.
MIR600HG is a lncRNA that has been reported to be related to autophagy in OSCC [13]. Here we found that MIR600HG was upregulated in OSCC tissues and cells. Similar to our result, Song et al. [14] have demonstrated that MIR600HG acted as a prognostic target for predicting PDCA and was significantly associated with the overall survival of PDCA patients. In this study, it was found that in SCC-9 cells, MIR600HG overexpression promoted proliferation, migration, invasion and DDP resistance, and inhibited apoptosis of cells. In SCC-9/DDP cells, silencing of MIR600HG suppressed proliferation and glycometabolism, and enhanced apoptosis and DDP sensitivity of cells. These findings X. Liu et al. Fig. 3 MIR600HG binds to miR-125a-5p. A Binding sites between miR-125a-5p and MIR600HG were predicted using bioinformatic websites; B targeting relationship between miR-125a-5p and MIR600HG was confirmed using dual luciferase reporter gene assay; C miR-215a-5p expression in SCC-9 cells after transfection of oe-NC or oe-MIR600HG was assessed using RT-qPCR; D miR-125a-5p expression in clinical samples was assessed using RT-qPCR (n = 85); E miR-125a-5p expression in NHOK and OSCC cell lines was assessed using RT-qPCR; F miR-125a-5p expression in SCC-9 and SCC-9/DDP cells was assessed using RT-qPCR; G miR-215a-5p expression in SCC-9 cells after transfection of mimic NC or miR-125a-5p mimic was assessed using RT-qPCR; H colony formation ability of SCC-9 cells was determined using colony formation assay; I viability of SCC-9 cells was determined using MTT assay; J IC 50 of SCC-9 cells was determined using MTT assay; K migration and invasion of SCC-9 cells was detected using Transwell assay; L apoptosis of SCC-9 cells was detected using flow cytometry; M miR-215a-5p expression in SCC-9/DDP cells after transfection of inhibitor NC or miR-125a-5p inhibitor was assessed using RT-qPCR; N viability of SCC-9/DDP cells was determined using MTT assay; O IC 50 of SCC-9/DDP cells was determined using MTT assay; P apoptosis of SCC-9/DDP cells was detected using flow cytometry; Q glycometabolism of SCC-9/DDP cells was analyzed; N = 3; *P < 0.05; Data were expressed as mean ± standard deviation; t-test or ANOVA was used to analyze data. Fig. 4 miR-125a-5p inhibition reverses the role of MIR600HG silencing on DDP-resistant OSCC cells. A miR-215a-5p expression in SCC-9/ DDP cells after transfection of si-MIR600HG+miR-125a-5p inhibitor or si-MIR600HG+inhibitor NC was assessed using RT-qPCR; B viability of SCC-9/DDP cells was determined using MTT assay; C IC 50 of SCC-9/DDP cells was determined using MTT assay; D, E apoptosis of SCC-9/DDP cells was detected using flow cytometry; F glycometabolism of SCC-9/DDP cells was analyzed; N = 3; *P < 0.05; data were expressed as mean ± standard deviation; and ANOVA was used to analyze data.  85); B RNF44 mRNA expression in NHOK and OSCC cell lines was assessed using RT-qPCR; C RNF44 mRNA expression in SCC-9 and SCC-9/DDP cells was assessed using RT-qPCR; D two effective si-RNF44 were designed to knockdown the expression of RNF44 and its interference efficiency was detected by RT-qPCR and western blot analysis in SCC-9 cells; E colony formation ability of SCC-9 cells was determined using colony formation assay after transfection of si-RNF44; F viability of SCC-9 cells was determined using MTT assay after transfection of si-RNF44; G IC 50 of SCC-9 cells was determined using MTT assay after transfection of si-RNF44; H migration and invasion of SCC-9 cells was detected using Transwell assay after transfection of si-RNF44; I apoptosis of SCC-9 cells was detected using flow cytometry; J glycometabolism of SCC-9/DDP cells was analyzed after transfection of si-RNF44; N = 3; *P < 0.05; Data were expressed as mean ± standard deviation; t-test or ANOVA was used to analyze data.
indicated an oncogenic role of MIR600HG in the development of OSCC. Yao et al. [12] have reported that MIR600HG suppressed metastasis and enhanced oxaliplatin chemosensitivity by targeting ALDH1A3 in CRC. Due to the limited literature, the accurate mechanism of MIR600HG needs further explorations in human cancers.
LncRNAs are known to regulate the biological function of miRNAs by acting as a competitive endogenous RNA (ceRNA) [21]. In the present study, the interaction between MIR600HG and miR-125a-5p was confirmed. Furthermore, our results suggested that miR-125a-5p was expressed at a low level in OSCC cell lines and tissues, as well as in DDP-resistant OSCC cells. In accordance with our finding, Cao   Fig. 6 MiR-125a-5p targets RNF44. A Binding sites between miR-125a-5p and RNF44 were predicted using Targetscan website; B Targeting relationship between miR-125a-5p and RNF44 was confirmed using dual luciferase reporter gene assay; C RNF44 expression in SCC-9/DDP cells after transfection of miR-125a-5p inhibitor+si-NC or miR-125a-5p inhibitor+si-RNF44 was detected using RT-qPCR; D Viability of SCC-9/ DDP cells was determined using MTT assay; E IC 50 of SCC-9/DDP cells was determined using MTT assay; F apoptosis of SCC-9/DDP cells was detected using flow cytometry; G glycometabolism of SCC-9/DDP cells was analyzed; N = 3; *P < 0.05; Data were expressed as mean ± standard deviation; t-test or ANOVA was used to analyze data. et al. [22] have clarified that miR-125a-5p was downregulated in both cervical cancer tissues and cell lines, and miR-125a-5p levels shows a marked reduction in patients with prostate cancer in comparison to healthy controls [23]. Also, our results implied that miR-125a-5p elevation inhibited malignant phenotypes and DDP resistance of OSCC cells, and reduction of miR-125a-5p promoted proliferation and glycometabolism, and reduced apoptosis and DDP sensitivity of DDP-resistant OSCC cells. Consistently, it has been reported that miR-125a-5p inhibition attenuated ANRIL knockdowninduced chemosensitivity to DDP in osteosarcoma cells [16]. Moreover, a recent study has indicated that miR-125a-5p inhibited cell proliferation, cell cycle progression, and migration whereas promoted apoptosis of head and neck cancer cells [24]. It has also been revealed that miR-125a-5p reduced glycolysis to improve pulmonary arterial hypertension [25] and ameliorated hepatic glycolipid metabolism disorder in type 2 diabetes mellitus [26].
Furthermore, RNF44 was confirmed as a target gene of miR-125a-5p in this study, and RNF44 was found to be upregulated in OSCC. For OSCC cells, RNF44 silencing reduced the growth and enhanced DDP sensitivity; for DDP-resistant OSCC cells, RNF44 reduction could reverse the role of miR-125a-5p knockdown. As previously reported, RNF44 expression in BRAF inhibitors (BRAFi)-resistant melanoma (BR) cells was increased, and high levels of RNF44 also appeared in BR xenografts and melanoma tumor tissues from BR and BRAFi/MEK inhibitor-resistant melanoma patients [19].
Taken together, this research indicated that MIR600HG depletion reduces glycometabolism and DDP resistance of OSCC cells through acting as a ceRNA of miR-125a-5p to regulate RNF44. Our study may provide novel biomarkers for OSCC diagnosis and treatment. However, the effects of the MIR600HG/miR-125a-5p/ RNF44 are scarcely studied; thus, more efforts are needed for further explorations.

MATERIALS AND METHODS Ethics statement
Written informed consents were acquired from all patients before this study. The protocol of this study was confirmed by the Ethic Committee of

Study subjects
Eighty-five cases of OSCC tissues and the matched normal tissues were harvested from OSCC patients accepted treatment in School and Hospital of Stomatology, Cheeloo College of Medicine, Shandong University. These tissues were immediately frozen in liquid nitrogen and then preserved at −80°C. Patients' information is detailed in Supplementary Table 1.
The target plasmid (2 μg), siRNA, miRNA mimic or inhibitor (50 nM) was transfected into cells using Lipofectamine 3000 (Invitrogen) [22]. After the OSCC cells were transfected onto six-well plates for 48 h, the transfection efficiency was evaluated and subsequent experiments were performed.

Colony formation assay
Proliferation and colony formation of transfected SCC-9 cells were measured by colony formation assay. At 48 h post-transfection, cells were seeded into six-well plates at a density of 400 cells/well, cultured for 15 days, and then were fixed and stained with crystal violet (Sigma-Aldrich, CA, USA) for 5 min. The colonies were counted under a microscope (Olympus, Tokyo, Japan) [28].

Flow cytometry
Transfected SCC-9 cells or SCC-9/DDP cells were seeded in a six-well plate at 2 × 10 5 cells/well. Cell apoptosis was determined using Annexin V-FITC/ Fig. 7 Silencing of MIR600HG strengthens the DDP sensitivity of OSCC cells in vivo. A Representative images of xenografts from nude mice; B tumor volume of each group; C tumor weight of each group; n = 6; *P < 0.05; the data were expressed as mean ± standard deviation and ANOVA was used to analyze data. PI Apoptosis Detection kit (BD Biosciences, San Diego, CA, USA) and analyzed on a flow cytometer (FACSCalibur; BD Biosciences) [30].

Transwell assay
The migration and invasion of transfected SCC-9 cells were determined using Transwell chambers (Costar, NY, USA) as previously described [31]. Matrigel (BD Bioscience, CA, USA) was used in the invasion assay but not migration assay.

Metabolism parameters of glucose utilization and lactic acid production
The culture medium was collected after SCC-9/DDP cells were incubated for 48 h in the six-well plate. Using a Glucose Assay Kit (Sigma-Aldrich) and a L-Lactate Assay Kit (Eton Bioscience, CA, USA), the concentration of glucose and lactate in spent medium was quantitatively measured. Glucose consumption of SCC-9/DDP cells was calculated as the glucose concentration of the unused cell culture medium minus the glucose concentration in the culture medium [32].

ATP level evaluation
ATP level in cells was assessed using the luminometer (Promega, WI, USA) and CellTiter-Glo 2.0 Assay kits. The total protein in relative cell lysate was used to normalize the relative cellular ATP level [33].

Dual-luciferase reporter gene assay
The amplified DNA sequence of MIR600HG containing wild type (WT) or mutant (MUT) miR-125a-5p-binding sites were amplified by PCR and then cloned into a pmirGLO dual-luciferase plasmid (Promega, WI, USA) to produce MIR600HG-WT and MIR600HG-MUT reporter plasmids. SCC-9 cells were co-transfected with MIR600HG-WT or MIR600HG-MUT and miR-125a-5p mimic or its NC using Lipofectamine 3000 (Invitrogen). At 48 h posttransfection, dual-luciferase reporter assay was performed using the Dual-Glo® Luciferase Assay System (Promega) according to the manufacturer's instructions. Firefly luciferase activity of the pmirGLO vector was measured to normalize Renilla luciferase activity. The targeting relationship between miR-125a-5p and RNF44 were verified in the same way [34,35].

Subcutaneous tumorigenesis in nude mice
Twenty-four BALB/c nude mice (Laboratory Animal Center of Shandong University, Shandong, China) were fed under specific pathogen-free conditions. SCC-9/DDP cells with sh-MIR600HG or sh-NC were suspended in 100 μL medium, and cells (2 × 10 5 ) were resuspended in 50 μL normal saline, added with 50 μL Matrigel Matrix and subcutaneously injected into the mouse right abdomen. Seven days after the injection, mice were injected with DDP (4 mg/kg) or phosphate-buffered saline (PBS) every 4 days. The tumor volume was measured at the day 7, 14, 21, and 28. (a × b 2 )/2 (a: tumor length; b: tumor width). The allowed maximum diameter of the tumors was 1.5 cm. After the experiment, mice were euthanized with their xenografts resected and weighed [37].

RT-qPCR
Total RNA in tissues and cells was extracted using Trizol kits (Invitrogen) and reversely transcribed into cDNA using reverse transcriptase (Epicentre, WI, USA) or the miS-cript Reverse Transcription Kit (Qiagen). cDNA amplification was performed using the SYBR Premix Ex Taq (Takara, Dalian, China) and the PCR was conducted using Applied Biosystems 7900HT PCR system with β-actin as the internal reference. Primers are shown in Supplementary Table 2 and data were analyzed using the 2 -ΔΔCt method.

Western blot analysis
Total protein was extracted using extraction kits (KeyGen) with the concentration determined using bicinchoninic acid protein detection kits (Thermo Fisher). Proteins were performed with gel electrophoresis, transferred onto membranes, blocked with 5% skim milk powder, and incubated with primary antibodies RNF44 (1:1000; Proteintech, Chicago, USA), and β-actin (1:10,000; Abcam) at 4°C overnight. Then, the proteins were incubated with relative secondary antibody for 1 h and the bands were analyzed using the Multi Gauge ver 3.0 software (Fuji Photo Film Co., Ltd, Tokyo, Japan) with β-actin as the internal reference.

Statistical analysis
All data analyses were conducted using GraphPad Prism 6.0 software (GraphPadSoftware, CA, USA) and SPSS 20.0 software (IBM, NY, USA). Data were presented as mean ± standard deviation. t-test was performed for comparisons between two groups and analysis of variance was used for comparisons among multiple groups, followed by Tukey's post hoc test. P value <0.05 was indicative of statistically significant difference.