This study was performed following the approval of the Ethical Committee of Southern Central Hospital of Yunnan Province and complied with the Declaration of Helsinki. All patients had signed the informed consent. Animal experiments were implemented based on the Guide for the Care and Use of Laboratory Animals (20).

The CC tissues and paracancerous tissues were collected from 70 CC patients (aged 38–65 years, at an average age of 51.40 years) who underwent surgical resection in Southern Central Hospital of Yunnan Province. These patients did not receive any radiotherapy or chemotherapy before the operation, with complete clinical data. The tissues were frozen in liquid nitrogen immediately after the operation and kept at -80°C for subsequent experimentation. These patients were followed up by means of telephone or return visit. The detailed clinicopathological features and 5-year overall survival rate were recorded.

CC cells (SiHa, HeLa, C-33A, MS751, and CaSki) and human normal cervical epithelial cells (Ect1/E6E7) (ATCC, Manassas, Virginia, USA) were cultured in RPMI-1640 medium (GIBCO, Grand Island, NY, USA) containing 10% fetal bovine serum (Beyotime, Beijing, China). When the cells attained 70-80% confluence, transfection experiments were conducted.

The lentiviral overexpression vectors of METTL3 (LV-oe-METTL3) and negative control (LV-oe-NC) were obtained from Genechem (Shanghai, China). HeLa cells were infected with 5 mg/mL polybrene and lentivirus for 48 h. The titer of lentivirus (3 × 10⁸ PFU/mL) was determined using the fluorescent activated cell sorting method. The stably transfected cells were screened using 5 μg/mL puromycin (Sigma-Aldrich, Merck KGaA, Darmstadt, Germany). Three short hairpin RNA (shRNA) sequences for LINC00958 (sh-LINC00958#1, sh-LINC00958#2, and sh-LINC00958#3), pcDNA3.1 LINC00958 (pc-LINC00958), pcDNA3.1 B-cell lymphoma 2 (Bcl-2)-associated athanogene 3 (BAG3) (pc-BAG3), three shRNA sequences for METTL3 (sh-METTL3#1, sh-METTL3#2, and sh-METTL3#3), three shRNA sequences for c-MYC (sh-c-MYC#1, sh-c-MYC#2, and sh-c-MYC#3), and their NCs obtained from GenePharma (Shanghai, China) were transfected into SiHa or HeLa cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). The subsequent experiments were conducted after 48 h.

The treated SiHa or HeLa cells were rinsed with phosphate-buffered saline (PBS). Then, cell proliferation was measured using the CCK-8 kit (Dojindo Laboratories, Kumamoto, Japan). The cells were seeded into 96-well plates (2000 cells/well) and cultured at 37°C with 5% CO₂ for 0, 24, 48, and 72 h. Subsequently, 10 μL CCK-8 solution was added for another 3 h incubation. The absorbance was examined at 450 nm using a microplate reader (Bio-Rad, Hercules, CA, USA).

The treated SiHa or HeLa cells were collected, resuspended in RPMI-1640 medium, and counted. Then, cells were seeded into 6-well plates (150 cells/well) and cultured for 14 d. Afterward, the medium was sucked off and cells were fixed with 4% paraformaldehyde (Beyotime) for 20 min. Finally, the cells were stained with 200 μL crystal violet solution (Beyotime) for 15 min. The number of staining colonies was observed and counted under an inverted microscope (Nikon, Tokyo, Japan).

The treated SiHa or HeLa cells were seeded into a 6-well plate (1 × 10⁶ cells/well) and cultured at 37°C with 5% CO₂ (21). After 24 h of incubation, the cells were harvested using trypsin and washed with pre-cooled PBS. The Annexin V-FITC/PI apoptosis detection kit (Solarbio, Beijing, China) was used for double staining. The cells were stained with Annexin V-FITC for 15 min and then propidium iodide (PI) in the dark for 5 min. Apoptosis analysis was performed with the flow cytometer (cytoFlexS, Beckman).

TRIzol (Invitrogen) was utilized for total RNA extraction, and RNA quantity was examined using NanoDrop ND-1000. m⁶A RNA methylation quantification kit (ab185912, Abcam) was employed to test the m⁶A quality in total RNA, followed by the evaluation of absorbance at 450 nm.

MeRIP-qPCR was conducted as described in the previous literature (22) to quantify the level of m⁶A-modified LINC00958 in HeLa cells. Briefly, total RNA was extracted from CC cells using TRIzol. Purified RNA (5 µg) was digested by DNase I (M0303, NEB, Ipswich, MA, USA) and then incubated at 95°C for 25 s in RNA Fragmentation Reagents (AM8740, Ambion, Austin, Texas, USA), followed by ethanol precipitation and collection. Anti-m6A antibody (ab208577, Abcam) or anti-IgG (ab170190, Abcam) was incubated overnight with Protein A/G beads in IP buffer (150 mM NaCl, 0.1% NP-40, 10 mM Tris HCl, pH 7.4) at room temperature. RNA and prepared antibody-bead mixture were incubated in IP buffer at 4°C for 4 h. After three washes, the bound RNA was eluted from the beads with 0.5 mg/mL N6 methyladenosine (P3732, Berry&Associates) in IP buffer. The eluted RNA was extracted with Enol: Chloroform: Isoamylol (pH<5.0, P1025-500, Solarbio), and then cDNA was generated using All In One RT MasterMix (G490, ABM). The enrichment level of m6A was detected by reverse transcription quantitative polymerase chain reaction (RT-qPCR), with the control group level as the relative value.

HeLa cells were treated with 1 μg/mL actinomycin D. RNA was extracted for RT-qPCR at different times (0, 3, and 6 h).

The Magna RIP RNA-binding protein immunoprecipitation kit was obtained from Millipore (Billerica, MA, USA). More than 10⁷ SiHa or HeLa cells (23) were re-suspended in RIP lysis buffer with anti-c-MYC (ab32072, Abcam) or anti-IgG (ab172730, Abcam). Then, the cells in each group were added with magnetic bead and proteinase K. Finally, the precipitated RNA was isolated and purified for RT-qPCR.

Two sets of luciferase reporter assay were performed. 1). To confirm whether LINC00958 regulated the transcriptional activity of c-MYC, we obtained c-MYC-responsive 4x-Ebox reporter vector from GenePharma and transfected it into SiHa or HeLa cells interfered with LINC00958. 2). To confirm whether BAG3 was the transcriptional target of c-MYC, we constructed the BAG3 promoter sequence fragments containing a c-MYC binding site (WT) or mutant site (MUT) and then inserted them into the pGL3 reporter vector respectively (Promega, Madison, WI, USA). The luciferase reporter plasmids were co-transfected with pcDNA3.1 c-MYC or pcDNA3.1 NC into SiHa or HeLa cells. The cells were lysed after 48 h. The luciferase activity was examined using the dual-luciferase reporter assay kit (Promega).

EZ ChIP kit (Millipore) was utilized for ChIP assay. Briefly, the treated SiHa or HeLa cells (1 × 10⁷) (24) were subjected to chromatin cross-linking, and DNA was cut into fragments by repeated ultrasound. Then, immunoprecipitation of DNA-protein complex was performed using the anti-c-MYC antibody (ab32072, Abcam) or IgG antibody (ab172730, Abcam). Afterward, the DNA of immunoprecipitated DNA-protein complex was extracted and purified using the DNA fragment purification kit (Intron Biotechnology, South Korea) for RT-qPCR, with GAPDH as NC.

BALB/c nude mice (5-week-old) were obtained from Vital River Laboratory Animal Technology Co., Ltd (Beijing, China) [SYXK (Beijing) 2017-0033]. The mice were acclimated for 7 days and maintained under a 12-h dark/light cycle with sufficient food and water. HeLa cells were infected with LINC00958 shRNA lentivirus (Genechem) and the stably transfected cells were screened by 5 μg/mL puromycin. HeLa cells with stable LINC00958 knockdown or LINC00958 knockdown + METTL3 overexpression were collected and injected subcutaneously into mice on the right side near the forelimb (3 × 10⁶). From the 7^th day, the tumor growth was monitored weekly (Volume = Length × Width²/2). Four weeks after the injection, the mice were euthanized by intraperitoneal injection of 100 mg/kg pentobarbital sodium. The tumor was excised for subsequent analysis.

The tumor tissues were embedded in paraffin and sliced (5 μm). After dewaxing and rehydration, the antigen was retrieved with citrate buffer for 10 min. Then, the sections were subjected to incubation with 3% hydrogen dioxide for 15 min and normal goat serum (Solarbio) for 15 min. After that, the sections were incubated with METTL3 (ab195352, Abcam), c-MYC (ab32072, Abcam), and ki67 (ab15580, Abcam) at 4°C overnight, followed by incubation with the secondary antibody (ab205718, Abcam) at 37°C for 1 h. Finally, diaminobenzidine (Solarbio) and hematoxylin staining were conducted, followed by observation under a microscope (Olympus, Tokyo, Japan).

The gene expression in CC cells or tissues was quantitatively determined using RT-qPCR. TRIzol (Invitrogen) was used for RNA isolation. RNA concentration was measured using the Nano-Drop ND-1000 spectrophotometer. M-MLV (Takara, Kyoto, Japan) was employed for reverse transcription, and cDNA amplification was conducted using the SYBR Green Master Mix kit (Takara). Table 1 exhibits the primers. The relative expression of genes was calculated using the 2^-ΔΔCt method (25), with GAPDH as the internal control.

The protein was extracted using radio-immunoprecipitation assay buffer (Beyotime), separated by SDS-PAGE, and transferred onto PVDF membranes (Millipore). The membranes were blocked with skim milk for 2 h and incubated with the primary antibodies METTL3 (1:1000, ab195352, Abcam), c-MYC (1:1000, ab32072, Abcam), and GAPDH (1:2500, ab9485, Abcam) at 4°C overnight, followed by 3 times of tris-buffered saline-tween buffer (Solarbio) washing. Afterward, the membranes were incubated with the secondary antibody (1:2000, ab205718, Abcam) for 2 h. The gray value was analyzed using Image J (NIH, Bethesda, Maryland, USA).

LINC00958 expression in CC, the relationships between c-MYC and LINC00958, and c-MYC and BAG3, and the correlation between c-MYC expression and the prognosis of CC patients were predicted through the GEPIA database (http://gepia.cancer-pku.cn/) (26). The prognosis of LINC00958 in CC and the expression of METTL3 and BAG3 in CC were predicted through the UALCAN database (http://ualcan.path.uab.edu/analysis.html) (27). The correlation between LINC00958 expression and the prognosis of CC patients was predicted through the Kaplan-Meier Plotter database (http://kmplot.com/analysis/index.php?p=service&cancer=liver_rnaseq) (28). The binding site of c-MYC and BAG3 promoter was predicted through the Jaspar website (http://jaspar.genereg.net/) (29). The downstream genes of c-MYC were predicted through the RNAInter database (http://www.rna-society.org/rnainter/) (30).

Data analysis and map plotting were performed using the SPSS 21.0 (IBM Corp., Armonk, NY, USA) and GraphPad Prism 8.0 (GraphPad Software Inc., San Diego, CA, USA). The data complied with the assumption of normality and homogeneity of variance. The measurement data are presented as mean ± standard deviation. The t test was used for comparisons between two groups. One-way or two-way analysis of variance (ANOVA) was employed for comparisons among multiple groups, followed by Tukey’s multiple comparison test or Sidak’s multiple comparison test. The enumeration counting data were presented as cases. Fisher exact test was utilized for comparisons between groups. Kaplan-Meier survival curve and Log-rank test were employed to determine the correlation between LINC00958 expression and the prognosis of CC patients. Pearson correlation analysis was utilized to determine the correlation between the factors. All p-values were two sided and a value of p < 0.05 was considered statistically significant.

Considering that the role of LINC00958 in CC growth has not been fully elucidated, we predicted through the GEPIA database that LINC00958 exhibited an upregulation of expression in CC (Figure 1A). Our finding demonstrated that LINC00958 expression was elevated in CC tissues and cells (p < 0.01, Figures 1B, C). Then, we assigned CC patients to the LINC00958 high-expression group and LINC00958 low-expression group, with the median LINC00958 expression as the critical threshold (31). LINC00958 expression was correlated with tumor size, lymph node metastasis, and Federation of Gynecology and Obstetrics (FIGO) stage (p < 0.05, Table 2). Kaplan-Meier Plotter and UALCAN databases predicted that the survival of CC patients with high LINC00958 expression was notably shorter than that of CC patients with low LINC00958 expression (Figures 1D, E). Kaplan-Meier survival analysis of CC patients showed that higher LINC00958 expression indicated shorter overall survival (p < 0.01, Figure 1F). Altogether, LINC00958 was highly expressed in CC and correlated with the prognosis and clinicopathological features of CC patients.

To explore the effect of LINC00958 on the proliferation and apoptosis of CC cells, we transfected sh-LINC00958 into HeLa cells with relatively high LINC00958 expression, and successfully downregulated LINC00958 expression in HeLa cells (p < 0.01, Figure 2A). pc-LINC00958 was transfected into SiHa cells with relatively low LINC00958 expression, and LINC00958 expression was successfully upregulated in SiHa cells (p < 0.01, Figure 2B). LINC00958 silencing reduced HeLa cell proliferation, while LINC00958 overexpression enhanced SiHa cell proliferation (p < 0.01, Figures 2C, D). LINC00958 silencing increased HeLa cell apoptosis, while LINC00958 overexpression decreased SiHa cell apoptosis (p < 0.01, Figure 2E). Briefly, LINC00958 silencing suppressed CC cell proliferation and facilitated apoptosis.

Subsequently, we established a nude mouse xenograft tumor model to evaluate the effect of LINC00958. LINC00958 silencing depressed tumor growth (p < 0.01, Figure 3A) and reduced tumor weight (p < 0.01, Figure 3B). Ki67 can reflect cell proliferation (28), so we employed immunohistochemistry to detect the positive rate of ki67 in tumors. LINC00958 silencing notably reduced the positive rate of ki67 (p < 0.01, Figure 3C). Compared with the LV-sh-NC group, the LV-sh-LINC00958 group showed reduced LINC00958 expression (p < 0.01, Figure 3D). These results indicated that LINC00958 silencing repressed the growth of CC cells in vivo.

Then, we focused on the downstream mechanism of LINC00958 in CC growth. METTL3-mediated m⁶A modification regulates LINC00958 expression in hepatocellular carcinoma (19). UALCAN database predicted the upregulation of METTL3 expression in CC (Figure 4A), and our results also confirmed the elevation of METTL3 expression in CC tissues and cells (p < 0.05, Figures 4B–E). m⁶A quantitative analysis showed that the m⁶A level in CC tissues and cells was consistent with the trend of METTL3 expression (p < 0.05, Figures 4F, G). METTL3 expression was positively correlated with LINC00958 expression in CC tissues (p < 0.01, Figure 4H). We speculated that the upregulation of LINC00958 expression in CC was related to METTL3-mediated m6A modification. We successfully downregulated METTL3 expression in HeLa cells by transfecting sh-METTL3 (p < 0.01, Figures 4I, J), and found that the m⁶A level was also decreased (p < 0.01, Figure 4K). Additionally, METTL3 silencing decreased m⁶A level in LINC00958 (p < 0.01, Figure 4L) and reduced LINC00958 expression (p < 0.01, Figure 4M). Moreover, we treated HeLa cells with actinomycin D to block transcription, and found that METTL3 silencing notably shortened the half-life of LINC00958 (p < 0.01, Figure 4N). Briefly, METTL3-mediated m⁶A modification elevated LINC00958 expression in CC by promoting its RNA stability.

To verify the role of METTL3/LINC00958 axis in CC, we designed a functional rescue experiment. HeLa cells were infected with LV-oe-METTL3 to upregulate METTL3 expression (p < 0.01, Figures 5A, B), and then the HeLa cells were further treated with sh-LINC00958#1. Compared with LINC00958 silencing alone, the combined treatment of METTL3 overexpression and LINC00958 silencing increased LINC00958 expression in cells (p < 0.01, Figure 5C), enhanced cell proliferation (p < 0.01, Figures 5D, E), and decreased apoptosis (p < 0.01, Figure 5F). Briefly, METTL3 overexpression offset the impact of LINC00958 silencing on CC growth.

Thereafter, the downstream mechanism of LINC00958 was investigated. LINC00958 can activate c-MYC transcription in head and neck squamous cell carcinoma (HNSCC) (32). GEPIA database predicted that LINC00958 expression was positively correlated with c-MYC expression in CC (Figure 6A), and patients with higher c-MYC expression had shorter survival (Figure 6B). Hence, we speculated whether LINC00958 played a role in CC by activating the transcriptional activity of c-MYC. Our results demonstrated that c-MYC expression was elevated in CC tissues and cells (p < 0.01, Figures 6C–F), and LINC00958 overexpression increased c-MYC expression, while LINC00958 silencing decreased c-MYC expression (p < 0.01, Figures 6G–J). c-MYC expression was positively correlated with LINC00958 expression in CC tissues (p < 0.01, Figure 6K). Moreover, RIP results showed that compared with IgG, LINC00958 was highly enriched in c-MYC complex (p < 0.01, Figure 6L). LINC00958 overexpression induced the luciferase activity of c-MYC responsive construct, while LINC00958 silencing showed an opposite trend in HeLa cells (p < 0.01, Figure 6M). Altogether, LINC00958 activated the transcriptional activity of c-MYC in CC.

As a transcription factor, c-MYC can activate mRNA (33). The downstream genes of c-MYC were predicted through the RNAInter database, among which the apoptosis-related factor BAG3 is highly expressed in CC (Figure 7A) and concerned with CC cell proliferation and apoptosis (32–34). GEPIA database predicted that c-MYC was positively correlated with BAG3 in CC (Figure 7B). UALCAN database predicted an elevation of BAG3 expression in CC (Figure 7C). We speculated that BAG3 was a downstream mechanism of cMYC. Our results also revealed that BAG3 expression was elevated in CC tissues and cells (p < 0.01, Figures 7D, E), and c-MYC expression was positively correlated with BAG3 expression in CC tissues (p < 0.01, Figure 7F). Additionally, ChIP and dual-luciferase assay were designed according to the binding site of c-MYC and BAG3 promoter obtained from the Jaspar website (p < 0.01, Figures 7G, H). The results confirmed that c-MYC could bind to BAG3 promoter (p < 0.01, Figures 7I, J). Then, c-MYC was silenced in HeLa cells to further explore the effect of c-MYC on BAG3 transcription. RT-qPCR and Western blot results confirmed that the three c-MYC shRNAs had high intervention efficiency (p < 0.01, Figures 7K, L). BAG3 mRNA expression was decreased after c-MYC silencing (p < 0.01, Figure 7M). Briefly, c-MYC bound to the BAG3 promoter to promote its transcription.

The effect of LINC00958/BAG3 axis on CC cell growth was verified. pc-BAG3 was transfected into cells to upregulate BGA3 mRNA expression (p < 0.01, Figure 8A), and then the cells were further treated with sh-LINC00958#1. Compared with LINC00958 silencing alone, the combined treatment of LINC00958 silencing and BAG3 overexpression enhanced cell proliferation (p < 0.01, Figures 8B, C) and reduced apoptosis (p < 0.01, Figure 8D). Briefly, BAG3 overexpression offset the impact of LINC00958 silencing on CC growth.

Finally, we verified the role of METTL3 in CC growth and its regulatory effect on the c-MYC/BAG3 axis in vivo. METTL3 overexpression reduced the inhibitory effect of LINC00958 silencing on tumor growth, manifested as notably increased tumor volume and weight (p < 0.01, Figures 9A, B) and elevated ki67-positive rate (p < 0.01, Figure 9C). Compared with the sh-LINC00958#1 + oe-NC group, the sh-LINC00958#1 + oe-METTL3 group showed increased expressions of METTL3, LINC00958, c-MYC, and BAG3 in tumor tissues (p < 0.01, Figures 9D–G), and elevated m⁶A level (p < 0.01, Figure 9H). Immunohistochemical results also exhibited the same trend (p < 0.01, Figure 9I). Briefly, METTL3 accelerated the growth of CC cells in vivo via the LINC00958/c-MYC/BAG3 axis.