Comprehensive Analysis of N6-Methylandenosine-Related lncRNAs in Clear Cell Renal Cell Carcinoma: A Correlation With Prognosis, Tumor Progression, and Therapeutic Response

Abstract This study aims to develop a prognostic signature based on m6A-related lncRNAs for clear cell renal cell carcinoma (ccRCC). Differential expression analysis and Pearson correlation analysis were used to identify m6A-related lncRNAs associated with patient outcomes in The Cancer Genome Atlas (TCGA) database. Our approach led to the development of an m6A-related lncRNA risk score (MRLrisk), formulated using six identified lncRNAs: NFE4, AL008729.2, AL139123.1, LINC02154, AC124854.1 and ARHGAP31-AS1. Higher MRLrisk was identified as a risk factor for patients’ prognosis in ccRCC. Furthermore, an MRLrisk-based nomogram was developed and demonstrated as a reliable tool for prognosis prediction in ccRCC. Enrichment analysis and tumor mutation signature studies were conducted to investigate MRLrisk-related biological phenotypes. The tumor immune dysfunction and exclusion (TIDE) score was employed to infer patients’ response to immunotherapy, indicating a negative correlation between high MRLrisk and immunotherapy response. Our focus then shifted to LINC02154 for deeper exploration. We assessed LINC02154 expression in 28 ccRCC/normal tissue pairs and 3 ccRCC cell lines through quantitative real-time polymerase chain reaction (qRT-PCR). Functional experiments, including EdU incorporation, flow cytometry and transwell assays, were performed to assess the role of LINC02154 in ccRCC cell functions, discovering that its downregulation hinders cancer cell proliferation and migration. Furthermore, the influence of LINC02154 on ccRCC cells’ sensitivity to Sunitinib was explored using CCK-8 assays, demonstrating that decreased LINC02154 expression increases Sunitinib sensitivity. In summary, this study successfully developed an MRLrisk model with significant prognostic value for ccRCC and established LINC02154 as a critical biomarker and prospective therapeutic target in ccRCC management.


Introduction
Renal cell carcinoma (RCC), the second most prevalent urological malignancy in males and the foremost in females (1), predominantly presents as clear cell renal cell carcinoma (ccRCC), accounting for approximately 70% of kidney cancers (2).Conventional imaging approaches remain the mainstays in RCC diagnosis, while emerging techniques like radiomics and radiogenomics showed promise in enhancing presurgical assessments (3,4).Partial or radical nephrectomy remains the golden standard in the management of localized RCC (2).However, about 35% of RCC patients present advanced or metastatic disease initially, necessitating systemic treatments, which are sometimes augmented with cytoreductive nephrectomy (5,6).Owing to RCC's relative resistance to radiotherapy and standard chemotherapy, the current systemic treatment spectrum primarily comprises cytokines, vascular endothelial growth factor (VEGF) receptor inhibitors, mTOR inhibitors and the emerging immune checkpoint blockers (2,5,7).Despite the advancements in treatment strategies, the prognosis of late-stage RCC remains bleak, with a five-year disease-specific survival rate below 10% in stage IV patients (2,8).The treatment of RCC is further complicated by tumor heterogeneity among individuals and intra-tumoral changes over time, which significantly impact treatment responses and challenge the feasibility of a universal therapy.To improve therapeutic precision, a better understanding of ccRCC's molecular mechanisms is in demand.The identification of novel biomarkers is pivotal in enhancing early diagnosis, tailoring personalized treatment, improving prognosis prediction, and potentially overcoming drug resistance in ccRCC patients.
Early discovered in 1970s, N6-methyladenosine (m6A) is now recognized as the most abundant post-translational modification of mRNA and non-coding RNA in eukaryotes (9,10).M6A modification plays an important role in almost every stage of RNA metabolism, including splicing, export, translation, decay, etc (10).As a highly dynamic and reversible process in human cells, m6A modification is mainly regulated by three types of m6A regulators, "writers," "erasers," and "readers," corresponding to methyltransferases, signal transducers and demethylases (11).Accumulating evidence demonstrate the potential links between m6A modification and human cancers, suggesting dysregulated expression of m6A enzymes could affect the progression of tumors (10).Li et al. reported that m6A methyltransferase METTL3 might act as a tumor suppressor in the development and biological progress of RCC through EMT and PI3K-Akt-mTOR pathways (12).Gu et al. reported that DMDRMR exerts an essential oncogenic role in RCC by co-acting with m6A transducer IGF2BP3 in an m6A-dependent manner (13).These findings underscore the growing understanding of m6A's role in cancer biology, particularly in the context of RCC, emphasizing the need for further investigation into its underlying mechanisms.
Long non-coding RNAs (lncRNAs), defined as non-protein-coding RNA transcripts exceeding 200 nucleotides in length, are engaged in a variety of biological processes within eukaryotes, mainly by modulating gene expressions and functions at different levels (14).The dysregulation of lncRNA profile has been implicated in the pathogenesis of malignant tumors, exerting influences on cell proliferation, migration, invasion and drug resistance (15).In the realm of cancer research, certain lncRNAs have been identified as diagnostic markers and therapeutic targets across various cancer types.For example, LINC00887 and GIHCG were found to play oncogenic roles in RCC by facilitating cell proliferation, thus positioning them as promising biomarkers for RCC diagnosis (16,17).LncARSR, a mediator of poor Sunitinib response in RCC, functions via competitively binding miR-34/miR-449 and consequently upregulating AXL and c-MET expression.This insight proposed lncARSR as a possible predictor and therapeutic target for Sunitinib resistance in RCC patients (18), highlighting the significant impact of lncRNAs in cancer therapeutics.These results offered new avenues for RCC diagnosis and treatment, underscoring the importance of integrating molecular insights into clinical practice.
In response to these challenges and opportunities, we constructed a prognostic signature based on m6A-related lncRNAs for ccRCC patients, utilizing large-scale RNA-seq data from The Cancer Genome Atlas (TCGA) database.To explore more potential clinical applications of m6Arelated lncRNA risk score (MRLrisk), we investigated its potency in predicting immunotherapeutic efficacy in ccRCC patients.Additionally, we validated the role of LINC02154 in promoting ccRCC cell growth and reducing sensitivity to Sunitinib through a series of experimental approaches, adding depth to our understanding of ccRCC pathology and treatment.

Datasets and preparation
The transcriptome profiling data of ccRCC patients were downloaded from TCGA database (https://portal.gdc.cancer.gov/),including 541 tumor tissues and 76 normal tissues.The data was meticulously annotated and disintegrated into messenger RNA and lncRNA using the Ensembl human annotation file (https://asia.ensembl.org/info/data/index.html).From this process, a total of 16,773 lncRNAs were obtained.The clinicopathological data pertaining to these cases were also extracted from the TCGA database.
After excluding those lacking complete survival data, a total of 532 cases were included and randomly allocated into two sets, a training set (268 cases) and a testing set (264 cases).We conducted a comparative analysis of baseline demographic and clinical characteristics between the two sets (Table 1).The training set served as the basis for constructing a prognostic model, which was subsequently validated using the testing set.Univariate Cox regression analysis was performed to identify m6A-related lncRNAs associated with cancer prognosis (p < 0.05).To streamline the number of variables and prevent model overfitting, the least absolute shrinkage and selection operator (LASSO) Cox regression analysis was used and a more refined model was generated, leading to the identification of six independent prognostic-related lncRNAs (19).Based on the analysis results, a m6A-related lncRNA risk score (MRLrisk) formula was constructed: MRLrisk ¼ P n i coef ðiÞ � xðiÞ, where n, coef and x represent the number of lncRNAs, the coefficient and the lncRNA expression, respectively.The median MRLrisk was chosen as the cutoff value.Cases with a risk score below the cutoff were categorized as low-risk subgroup, while those exceeding the cutoff were deemed high-risk.Principal component analysis (PCA) was applied to obtain the validity of consensus clusters.

Validation of m6A-related lncRNAs prognostic signature
The predictive capability of the signature was rigorously validated.We employed Kaplan-Meier survival analysis to ascertain the significance of survivorship differences between the defined subgroups.The prognostic ability of MRLrisk was assessed using receiver operating characteristic

Functional enrichment analysis
Based on MRLrisk, differentially-expressed genes between low-risk subgroup and high-risk subgroup (log 2 jfold changej >1 and false discovery rate (FDR) <0.05) were determined using limma R package.The identified genes were then submitted to gene ontology (GO) enrichment analysis to elucidate their functional distribution.

Analysis of tumor mutation burden
The sample mutation dataset was acquired from the TCGA database.Tumor mutation burden (TMB), defined as the total count of non-synonymous somatic mutations per megabase in examined regions, was calculated via Perl script.The association between TMB and MRLrisk was assessed.We then divided patients into two subgroups using the median TMB count as threshold, and Kaplan-Meier survival analysis was conducted accordingly.

Prediction of immunotherapeutic efficacy
We further explored the ability of MRLrisk to predict patient's response to immunotherapy.The potential response of ccRCC patients to immunotherapy was inferred using the tumor immune dysfunction and exclusion (TIDE) score (http:// tide.dfci.harvard.edu/).Results were compared between the low-risk subgroup and high-risk subgroups, with p value <.05 considered as statistically significant.

Patient samples
28 paired tumor and adjacent normal tissues were obtained from ccRCC patients at the Department of Urology, Peking University First Hospital.The fresh tissue samples were immediately preserved in liquid nitrogen and subsequently stored at −80 � C until required for RNA extraction.Ethical approval was granted from the Ethics Committee of Peking University First Hospital.All participants provided written informed consent, authorizing the use of their tissue samples for scientific research purposes.

Cell lines and cell culture
Three human ccRCC cell lines, 786-O, OS-RC-2 and A498, along with an immortalized human normal renal tubular epithelial cell line (HK-2), were procured from the American Type Culture Collection (Rockville, MD, USA).A498 and HK-2 were cultured in Dulbecco's Modified Eagle's Medium (Invitrogen, Carlsbad, CA, USA), supplemented with 10% Fetal Bovine Serum (FBS).786-O and OS-RC-2 were cultured in PRIM-1640 (Invitrogen, Carlsbad, CA, USA), enriched with 10% FBS.All cells were maintained in a sterile environment at 37 � C with 5% CO 2 to ensure optimal growth conditions.

RNA extraction and quantitative real-time polymerase chain reaction (qRT-PCR)
Total RNA was extracted from cells or tissues using Trizol reagent (Invitrogen, Carlsbad, CA, USA).The concentration and quality of the RNA were determined using a Nanodrop ND-2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).cDNA was synthesized through reverse transcription (TansGEN, Beijing, China).qRT-PCR was conducted on the ABI PRISM 7000 Fluorescent Quantitative PCR System (Applied Biosystems, Waltham, MA, USA), with ACTB served as the endogenous control.The relative RNA expression levels were quantified using the 2 −DDCT method.

Infection with lentivirus
Recombinant lentiviruses designed to knock down LINC02154 (sh-LINC02154) and the corresponding control viruses (sh-NC) were sourced from HANBIO (Shanghai, China).Two distinct sh-LINC02154 variants (shRNA1 and shRNA2) were employed.The 786-O and OS-RC-2 cell lines were plated in six-well plates and infected with the lentiviruses.Post-infection, the 786-O and OS-RC-2 cells underwent treatment with puromycin at concentrations of 4 lg/ml and 6 lg/ ml, respectively, for a duration exceeding seven days.

RNA fluorescence in situ hybridization (FISH)
Cy3-fluorescence-labeled probes targeting LINC02154 were supplied by RiboBio (Guangzhou, China).The RNA FISH assay was conducted using a fluorescent in situ hybridization kit provided by RiboBio (Guangzhou, China), with procedures strictly adhering to the manufacturer's instructions.The process entailed fixing the 786-O and OS-RC-2 cells in 4% formaldehyde and permeabilization with 0.5% Triton X-100.Hybridization with the probes targeting LINC02154 was conducted at 37 � C for a duration of 12 h, followed by DNA staining with DAPI.For the visualization, a laser scanning confocal microscope (Leica, Solms, Germany) was employed.

Ethynyl-2-deoxyuridine (EdU) incorporation assay
Cell proliferation was assessed through EdU incorporation assay, using an EdU Apollo DNA in vitro kit (RiboBio, Guangzhou, China).Stably infected cells and control cells were cultured in 12-well plates until adherence.EdU working solutions were applied to each well, followed by 2 h's incubation.Subsequently, the cells were fixed using 4% paraformaldehyde, rinsed three times with PBS, and treated with 0.3% Triton X-100.Later, EdU reaction buffer was added to each well to facilitate the detection process.The final phase involved staining the cells with Hoechst 33342 and conducting analyses using a confocal laser scanning microscope (Leica, Solms, Germany).Each experiment was repeated at least three times to ensure accuracy and repeatability of results.

Cell counting kit-8 (CCK-8) assay
Cell proliferation following the application of Sunitinib was quantitatively assessed using the Cell Counting Kit-8 (Beyotime Inst Biotech, China), according to the manufacturer's instructions.The absorbance values, indicative of cell viability, were measured at a wavelength of 450 nm using a microplate reader (Bio-Rad, Hercules, CA, USA).

Cell migration assay
Cell migration was measured using a transwell assay.A total of 5 � 10 4 cells, suspended in 100 lL of serum-free PRIM-1640 medium, were seeded into a 24-well plate equipped with 8-lm pore transwell inserts (Corning, NY, USA).The lower chambers of the wells were filled with 500 lL of PRIM-1640 medium enriched with 10% FBS.Following a 24-h incubation period, cells that successfully migrated to the bottom surface of the filter membrane were fixed with 4% paraformaldehyde, stained with a 0.5% crystal violet solution for 30 min, and then imaged using a Leica DM IL microscope.

Flow cytometric analysis
In the cell apoptosis assay, cells were first stained with Annexin V-APC and Propidium Iodide (PI, KeyGEN BioTECH), in strict accordance with the manufacturer's guidelines.The detection and analysis of apoptotic cells were carried out with a flow cytometer (FACSCalibur, Becton Dickinson, New Jersey, USA).For the cell cycle assay, infected cells were collected and fixed with 70% pre-chilled ethanol at 4 � C overnight.Subsequently, these cells were stained with PI and incubated in the dark at 37 � C for 30 min.The cell cycle stages were then accurately analyzed using the flow cytometer.

Construction of an m6A-related lncRNAs prognostic signature
We identified 16,773 lncRNAs in the TCGA dataset, of which 2307 were identified as m6A-related through Pearson correlation analysis.A Sankey diagram was used to visualize the results (Figure 1(A)).To filter lncRNAs associated with overall survival, univariate Cox regression analysis was performed on the training set.LASSO Cox regression analysis was subsequently applied (Figure 1(B,C)).We further developed a prognostic signature consisting of six m6A-related lncRNAs, complete with their respective coefficients, via multivariate Cox regression analysis.
The m6A-related lncRNA risk score (MRLrisk) was calculated for each patient using the following formula: Patients were assigned to low-risk and high-risk subgroups based on the median MRLrisk.Of the six m6A-related lncRNAs, NFE4, AL008729.2,AL139123.1 and LINC02154 were identified as risk factors, while AC124854.1 and ARHGAP31-AS1 were protective (Figure 1(D)).A heatmap was generated to show the correlations between these six m6A-related lncRNAs and 23 m6A regulators (Figure 1(E)).

Evaluation of the m6A-related lncRNAs prognostic signature
The prognostic signature of m6A-related lncRNAs was evaluated using the training dataset.Kaplan-Meier survival analysis revealed poorer survival outcomes for patients with higher MRLrisk scores (Figure 2 for MRLrisk, predicting one, three, and five-year overall survival, are presented in Figure 2(B).The corresponding AUC values were 0.818, 0.825 and 0.827 respectively, demonstrating the strong predictive ability of the m6A-related lncRNAs prognostic signature in the training set.Moreover, the association between MRLrisk and overall survival was established through univariate Cox regression, and the independent predictive value of MRLrisk in ccRCC patients was confirmed by multivariate Cox regression analyses (Supplementary Figure S1(A,B)).Scatter plots illustrating the MRLrisk and survival status of each ccRCC patient in the training set revealed that higher MRLrisk scores were associated with shorter survival times (Figure 2(C)).A heatmap depicted the significant differential expression profiles of the six m6A-related lncRNAs between the low-risk and high-risk subgroups (Figure 2(C)).These findings underscore the reliability and sensitivity of the m6A-related lncRNAs prognostic signature in predicting the prognosis of ccRCC patients in the training set.

Validation of the m6A-related lncRNAs prognostic signature
The prognostic signature of m6A-related lncRNAs was further validated using both the testing dataset and the overall dataset.In these datasets, patients classified in the high-risk subgroup exhibited a poorer prognosis (Figure 2(D,G)).The ROC curves and their corresponding AUC values displayed an optimistic predictive performance of the m6A-related lncRNAs prognostic signature in both the testing set (one-year AUC ¼ 0.664, three-year AUC ¼ 0.666, five-year AUC ¼ 0.708) and across the entire cohort (oneyear AUC ¼ 0.752, three-year AUC ¼ 0.729, five-year AUC ¼ 0.741) (Figure 2(E,H)).Subsequent univariate and multivariate Cox regression analyses established MRLrisk as an independent prognostic factor in both the testing set and the entire cohort (Supplementary Figure S1(C-F)).Scatter plots representing the MRLrisk and survival status of each ccRCC patient were presented (Figure 2(F,I)).The expression profiles of the six m6A-related lncRNAs and their correlations with MRLrisks were summarized in heatmaps (Figure 2(F,I)).The consistency of these results with the findings from the training set reinforces the conclusion that the m6A-related lncRNAs prognostic signature we developed possesses a stable and reliable predictive capability.

Estimation of the clinical value of the m6A-related lncRNAs prognostic signature
The clinical significance of the m6A-related lncRNAs prognostic signature was assessed through multivariate ROC curve analysis.As depicted in Figure 3(A), the AUC value of MRLrisk is 0.752, surpassing those of age (0.653), gender (0.504), and grade (0.721) in predicting clinical outcomes.To evaluate the discriminatory power for overall survival among various variables, we computed concordance index (C-index) scores for time-dependent AUC.The results, visualized in Figure 3(B), demonstrated a consistent and robust discriminative capability of MRLrisk over a span exceeding 10 years.To provide a more quantitative method to predict patients' prognosis, a comprehensive nomogram integrating MRLrisk with age, gender, and stage was constructed (Figure 3(C)).Calibration curves at one-, three-and five-year intervals confirmed the strong agreement between predicted and observed probabilities (Figure 3(D)), suggesting the MRLrisk-based nomogram as a dependable tool for predicting prognosis in ccRCC.Furthermore, stratification analyses based on age (� 65 years vs. > 65 years), gender (male vs. female), grade (G1-2 vs. G3-4) and stage (stage I-II vs. stage III-IV) were conducted.Kaplan-Meier survival analyses were then employed to examine the clinical outcomes across these stratifications.The results indicated that MRLrisk consistently maintained its predictive accuracy across different patient characteristics within the ccRCC cohort (Supplementary Figure S2(A-H)).Furthermore, to determine if MRLrisk offer greater accuracy compared to existing prognostic signatures, we conducted a multivariate ROC curve analysis.The results revealed that MRLrisk significantly outperforms other models in predicting the prognosis of ccRCC patients, as illustrated in Figure 3(E-G) (20)(21)(22)(23)(24).

Functional enrichment analysis associated with MRLrisk
To explore the potential biological processes related to MRLrisk, functional enrichment analysis was performed.Differential expression genes (DEGs) were identified between the low-risk and high-risk subgroups and subjected to GO enrichment analysis.Within each of the three main categories of the GO classification (molecular function, cellular component, and biological process), the primarily enriched subcategories were identified (Figure 4(A)).The findings revealed that the functions of DEGs predominantly centered on immune-related aspects.This included antigen binding within the molecular function category, immunoglobulin complex in the cellular component category, and both defense response to bacterium and humoral immune response within the biological process category.These results highlighted a strong emphasis on immune-related functions among the DEGs.

Analysis of tumor mutation burden and its correlation with survival
The somatic mutation profiles of ccRCC patients were sourced from TCGA database.The top 20 genes with the highest mutation frequency were identified and visualized in waterfall plots (Figure 4(B,C)).We set out to investigate the association between TMB and MRLrisk.Figure 4(D) illustrated a trend toward higher TMB in the highrisk subgroup, though this trend did not reach statistical significance.Intriguingly, high TMB has been proposed as a predictive biomarker for better response to immunotherapy, mainly due to its association with increased production of immunogenic neoantigens (25,26).However, recent studies suggested that its predictive reliability is limited to certain cancer types.In cancers like ccRCC, where neoantigen load does not necessarily correlate with CD8 T-cell infiltration, high TMB may correlate with reduced immunotherapy response and poorer outcomes (27).To evaluate the potential implication of TMB on the prognosis of ccRCC patients, we divided patients into low-TMB subgroup and high-TMB subgroup, with the median TMB count as the threshold.Survival differences between these subgroups were analyzed using the Kaplan-Meier method (Figure 4(E)).In accordance with previous studies, ccRCC patients with higher TMB exhibited poorer survival outcomes.We further stratified patients into four subgroups based on their TMB and MRLrisk: high TMB/high MRLrisk, high TMB/low MRLrisk, low TMB/high MRLrisk, and low TMB/low MRLrisk.Kaplan-Meier survival analysis for these subgroups revealed that patients with high TMB and high MRLrisk had the poorest survival, whereas those with low TMB and low MRLrisk had the best survival outcomes (Figure 4(F)).

Immunotherapy response prediction using MRLrisk
The efficacy of MRLrisk in predicting the response of ccRCC patients to immunotherapy was evaluated through the calculation of TIDE scores.Notably, the high-risk subgroup exhibited elevated TIDE scores (Figure 4(G)), suggesting a reduced likelihood of positive response to immunotherapy.This observation aligns with the earlier hypothesis based on the TMB data, further substantiating the relevance of MRLrisk in predicting the efficacy of immunotherapeutic interventions in ccRCC patients.

Confirmation of LINC02154 expression levels and localization in ccRCC
Previous studies have reported that LINC02154 was associated with poor prognosis in hepatocellular carcinoma and ccRCC (22,28).Building on this understanding, our study focused on LINC02154 to delve deeper into its function in ccRCC.We first conducted qRT-PCR to verify the expression levels of LINC02154 in both ccRCC cell lines and clinical tissue samples.The findings demonstrated a significant upregulation of LINC02154 expression in the 786-O cell line compared to the HK-2 cell line (Figure 5(A)).Correspondingly, LINC02154 expression was significantly higher in ccRCC tumor tissues relative to adjacent normal kidney tissues (Figure 5(B)).These observations led us to select the 786-O and OS-RC-2 cell lines, both exhibiting high levels of LINC02154, for further investigation.To elucidate the functional mechanism of LINC02154 in ccRCC cells, RNA FISH was utilized to determine its cellular distribution.The analysis indicated a predominant localization of LINC02154 within the cytoplasm of 786-O and OS-RC-2 cells (Figure 5(C)), providing valuable insights into the potential molecular pathways through which LINC02154 may exert its effects in ccRCC.

Exploration of LINC02154's biological functions in ccRCC cells
In order to elucidate the biological roles of LINC02154 in ccRCC, we employed lentivirusmediated stable knockdown of LINC02154 in 786-O and OS-RC-2 cell lines.The successful downregulation of LINC02154 was validated through qRT-PCR, as shown in Figure 5(D).We examined the impact of LINC02154 knockdown on cell proliferation using EdU incorporation assays.The results indicated that LINC02154 knockdown significantly inhibited cell proliferation of both 786-O and OS-RC-2 cells in comparison to the control cells (Figure 5(E,F)).To assess the influence of LINC02154 expression on cell migration, transwell migration assays were conducted on 786-O and OS-RC-2 cells.These assays demonstrated a marked reduction in migration potential for cells with LINC02154 downregulation compared to the controls (Figure 5(G,H)).Furthermore, we examined the potential alterations in cell cycle progression and apoptosis resulting from the knockdown of LINC02154.Flow cytometry analysis was conducted to assess the cell cycle distribution in 786-O and OS-RC-2 cells with reduced LINC02154 expression.Results showed a notable increase in the G0/G1 phase population and a corresponding decrease in the S and G2 phase populations among the groups treated with sh-RNAs (Figure 6(A-D)).Additionally, we observed a connection between LINC02154 expression and apoptosis, noting that LINC02154 downregulation led to significantly increased apoptotic activity in these cell lines (Figure 6(E-H)).

Assessment of LINC02154's influence on sunitinib sensitivity in ccRCC cells
Sunitinib, one of the firstly used VEGF receptor inhibitors in ccRCC, is widely utilized in the treatment of advanced-stage ccRCC patients.While our research previously linked LINC02154 with a worse prognosis and an ability to enhance cancer cell growth in vitro, contrasting findings have emerged from drug sensitivity analyses using the pRRophetic R package, suggesting that high LINC02154 expression might increase Sunitinib sensitivity in ccRCC (22).Considering these seemingly contradictory results, our study further investigated the influence of LINC02154 expression on the growth dynamics of ccRCC cells in the presence of Sunitinib.The 786-O and OS-RC-2 cell lines, both with and without reduced LINC02154 expression, were exposed to a gradient of Sunitinib concentrations over a 48h period.For the groups with LINC02154 downregulation, we utilized cells treated with shRNA2, as it showed superior efficacy compared to shRNA1.Cell viability was then assessed using CCK-8 assays.The results revealed that cells with reduced LINC02154 expression exhibited lower viability and a decreased IC50 value in comparison to the control groups, indicating an increased sensitivity to Sunitinib (Figure 6(I,J)).This difference highlights LINC02154's potential role in modulating the sensitivity of ccRCC cells to Sunitinib, underscoring its potential importance in the development of ccRCC therapeutic strategies.

Discussion
Accumulating evidence underscores that m6A modifications, accounting for the majority of RNA methylation, are pivotal in tumorigenesis.As has been reported, m6A-modified lncRNAs can affect the biological functions of tumor cells in a large amount of tumor entities (29)(30)(31)(32)(33)(34).With lncRNA signatures being increasingly adopted to assess prognosis of cancer patients (35), m6A-related lncRNA prognostic models are attracting considerable interest due to their promising efficacy.For instance, Tu et al. developed a prognostic signature comprising nine m6A-related lncRNAs to predict survival in patients with lower-grade glioma (36).Xu's research highlighted the value of m6A-related lncRNAs in prognostic predictions for lung adenocarcinoma patients (37).Similarly, Wang et al. demonstrated that the m6A-related lncRNA signature could serve as a potential prognostic factor in gastric cancer (38).Despite these advancements, there is a notable gap in research regarding the role of m6A-related lncRNAs in ccRCC.The full scope of their potential as prognostic biomarkers, along with their underlying biological mechanisms in ccRCC, are yet to be fully elucidated.In this study, we constructed a prognostic model based on m6A-related lncRNAs and delved into its utility in predicting immunotherapeutic responses in ccRCC.Additionally, experimental validations were conducted to elucidate the specific role of a key m6A-related lncRNA identified in our research.
Firstly, the expression profile data of ccRCC cohort was imported to yield a coexpression network, through which a total of 2307 m6A-related lncRNAs were identified.Notably, among all the 23 m6A regulators involved, m6A writer RBM15 showed correlations with the highest number of lncRNAs.Through stepwise regression analyses, we confirmed the prognostic value of six m6Arelated lncRNAs (NFE4, AC124854.1,AL008729.2,ARHGAP31-AS1, AL139123.1,LINC02154), forming a corresponding prognostic signature.Among these, NFE4 has garnered attention for its association with human hemoglobin switching (39), and has recently been implicated as a critical component in prognostic models for both ccRCC and acute myeloid leukemia (21,40,41).AC124854.1, on the other hand, has established its relevance in prognostic predictions for ccRCC (20,(42)(43)(44), while ARHGAP31-AS1 has been linked to liver cancer prognosis (45).However, the precise biological mechanisms by which these three lncRNAs influence tumor behavior in ccRCC are yet to be fully unraveled.Intriguingly, LINC02154 stands out among these lncRNAs, with a more extensively characterized role in cancer progression.Notably, high expression levels of LINC02154 have been associated with poorer prognoses in laryngeal squamous cell carcinoma (46)(47)(48)(49).Additionally, LINC02154 is known to enhance SPC24 promotor activity, thereby contributing to cell proliferation, migration and invasion in hepatocellular carcinoma (28).The functions of the other identified lncRNAs are being recognized for the first time in this context.
Subsequently, ccRCC patients were stratified into high-and low-risk subgroups based on the median MRLrisk score, with a worse prognosis confirmed in the high-risk subgroup.Our analyses indicated that the MRLrisk score outperformed conventional clinical features in predictive capability.Additionally, we developed a nomogram for quantitative prognosis prediction in ccRCC patients.GO enrichment analysis highlighted significant involvement in immune-related processes, such as antigen binding, immunoglobulin complex, and humoral immune response, offering new perspectives on the functions of m6A-related lncRNAs in ccRCC.In clinical practice, employing the MRLrisk model can significantly improve patient stratification into defined risk categories, thereby optimizing patient management.The nomogram, integrating the MRLrisk score with standard clinical indicators, offers a user-friendly and precise method for individual risk assessment.Such advancements promise to refine clinical decision-making, fostering personalized patient care by aligning treatment strategies with prognostic expectations.Collectively, our findings validate the m6Arelated lncRNA signature as both a robust and accurate predictive model, and an effective method for discovering novel biomarkers for future research.
Immunotherapy has emerged as a pioneering treatment strategy in various malignancies, including ccRCC, sparking significant interest in identifying patients who may derive benefit from these therapies.TMB, commonly defined as the total number of somatic coding mutations, is recognized for its role in generating neoantigens that trigger anti-tumor immunity, thereby serving as a predictive marker for patient response to immune checkpoint inhibitor (ICI) therapy (50).Research across different cancer types, such as non-small cell lung cancer and urothelial cancer (51)(52)(53), has established a notable link between TMB and ICI efficacy.Notably, high TMB has been shown to predict favorable immune therapy responses, irrespective of PD-L1 expression (54,55).However, the relationship between TMB and ICI efficacy is discrepant across tumor types (56).Labriola et al. observed no significant association between TMB and ICI response in patients with metastatic renal cell carcinoma (mRCC), possibly due to mRCC's typically low TMB and patient-specific variations in antigen presentation (57).McGrail et al. suggested that the disconnect between neoantigen load and CD8 T-cell infiltration might impede TMB's predictive reliability for ICI treatment in cancers like ccRCC, prostate cancer and breast cancer (27).Intriguingly, high TMB in ccRCC is associated with lower response rates and poorer clinical outcomes compared to low TMB tumors (27).In our study, we noted a trend toward higher TMB in the high-risk subgroup, suggesting a potential correlation between elevated MRLrisk and reduced immunotherapy response, although this did not reach statistical significance.Kaplan-Meier analysis further revealed poorer survival outcomes in ccRCC patients with higher TMB counts.
To better understand the link between MRLrisk and immunotherapy response, we utilized the TIDE prediction score.The TIDE algorithm considers both mechanisms of tumor immune evasion, namely T cell exclusion and dysfunction, which enhanced the prediction of cancer immunotherapy response (58).Given the complexity of ccRCC's tumor immune microenvironment, TIDE may have a better predictive performance than TMB.In our study, patients in the high-risk subgroup presented higher TIDE scores, suggesting a likely diminished response to immunotherapy.Consequently, our study posits that the m6A-related lncRNA signature could provide dependable biomarkers for predicting ICI response in ccRCC patients.
Of the six m6A-related lncRNAs identified in our study, LINC02154 has emerged as the most extensively validated in terms of its biological role.Existing literature has highlighted its oncogenic role in both laryngeal squamous cell carcinoma and hepatocellular carcinoma (28,(46)(47)(48)(49). Building on this understanding, we directed our focus toward a detailed analysis of LINC02154 in ccRCC.Initial bioinformatic analyses had already indicated its overexpression in ccRCC compared to normal tissues, linking this heightened expression with poorer prognostic outcomes.Our further investigation, utilizing primary patient samples and ccRCC cell lines, confirmed its elevated expression in ccRCC.Moreover, loss of function experiments revealed that a reduction in LINC02154 expression could effectively suppress cell proliferation, impede cell cycle progression, and reduce cell migration in vitro.Taken together, these results underscore LINC02154's pivotal involvement in ccRCC tumorigenesis and position it as a potential therapeutic target for the disease.
Targeted therapy has remained the cornerstone for treating advanced ccRCC, with numerous novel agents introduced in recent decades.Currently, three major categories of targeted drugs are employed for ccRCC treatment: cytokines, VEGF receptor inhibitors and mTOR inhibitors.These therapies have significantly extended patient survival and enhanced their quality of life.However, the inherent tumor heterogeneity among patients and changes over the course of the disease present substantial challenges in formulating a universal therapeutic strategy.Our study intends to investigate the role of LINC02154 in guiding treatment strategies for ccRCC.While prior bioinformatic analysis proposed a positive correlation between heightened LINC02154 expression and increased sensitivity to Sunitinib in ccRCC (22), our results revealed that LINC02154 knockdown leads to reduced cell viability under Sunitinib treatment, implying an elevated sensitivity to the drug.This finding prompts a reevaluation of the conclusions drawn from bioinformatic analyses, underscoring their potential limitations.However, it is also crucial to acknowledge the constraints of our own study.The knockdown of LINC02154 was observed to induce apoptosis in ccRCC cells, which could confound the interpretation of data pertaining to cell proliferation under drug treatment, indicating the complexity of assessing therapeutic responses in ccRCC and underscoring the need for additional validation of these findings.
The current study, while comprehensive, is not without its limitations.Firstly, the m6A-related lncRNA prognostic signature was established based solely on the TCGA dataset.Owing to the lack of other publicly available ccRCC lncRNAseq datasets, it is imperative to validate our model through larger clinical cohorts, thereby bolstering its reliability and applicability.Secondly, our study primarily relies on bioinformatics analysis and in vitro experiments, which may not fully capture the complexity of the tumor microenvironment in vivo.Thereby, additional in vivo experiments are needed to validate our findings.Moreover, exploring the interaction of these lncRNAs with other components of the tumor microenvironment, such as immune cells, may provide insights into their role in modulating immune responses.
In summary, we developed a MRLrisk model that demonstrates robust prognostic value and the capacity to predict immunotherapy response in ccRCC patients, offering a valuable tool for patient stratification and treatment guidance.Moreover, our results identify LINC02154 as a key regulator in ccRCC, impacting cell proliferation and migration, and modulating response to Sunitinib, underscoring its therapeutic relevance in ccRCC management.

Figure 1 .
Figure 1.Construction of a m6A-related lncRNAs prognostic signature.(A) Sankey diagram of m6A regulators and m6A-related lncRNAs.(B, C) The tuning parameter (logk) was controlled to select relevant covariates and estimate coefficients in the least absolute shrinkage and selection operator (LASSO) Cox regression analysis.According to the minimal criterion and 1-se criterion, perpendicular imaginary lines were drawn at the optimal value.(D) Multivariate Cox regression analysis revealed six independent prognostic m6A-related lncRNAs (E) Heatmap depicts the correlations between 23 m6A regulators and 6 m6A-related lncRNAs.

Figure 2 .
Figure 2. Evaluation and validation of the m6A-related lncRNAs prognostic signature.(A) Kaplan-Meier curves of overall survival (OS) of patients in the high-and low-risk subgroups in training set.(B) Receiver operating characteristic (ROC) curves of the m6Arelated lncRNA risk score (MRLrisk) for predicting the one/three/five-year overall survival in training set.(C) Distributions of MRLrisk, survival status and the expression levels of six prognostic m6A-related lncRNAs for each patient in training set.(D) Kaplan-Meier curves of OS of patients in the high-and low-risk subgroups in testinIet.(E) ROC curves of the MRLrisk for predicting the one/three/five-year overall survival in testing set.(F) Distributions of MRLrisk, survival status and the expression levels of six prognostic m6A-related lncRNAs for each patient in testing set.(G) Kaplan-Meier curves of OS of patients in the high-and lowrisk subgroups in overall set.(H) ROC curves of the MRLrisk for predicting the one/three/five-year overall survival in overall set.(I) Distributions of MRLrisk, survival status and the expression levels of six prognostic m6A-related lncRNAs for each patient in overall set.

Figure 3 .
Figure 3. Estimation of the clinical value of the m6A-related lncRNAs prognostic signature.(A) ROC curves of MRLrisk and clinicopathological characteristics.(B) Concordance indexes of MRLrisk and clinicopathological characteristics.(C) A prognostic nomogram constructed based on MRLrisk and clinicopathological characteristics for predicting patients' overall survival at one, three and five years.(D) Calibration curves of the nomogram.(E) ROC curves comparing one-year prognostic accuracy of MRLrisk with previously published ccRCC signatures.(F) ROC curves comparing three-year prognostic accuracy of MRLrisk with previously published ccRCC signatures.(G) ROC curves comparing five-year prognostic accuracy of MRLrisk with previously published ccRCC signatures.

Figure 4 .
Figure 4. Estimation of tumor immune environment and immunotherapy response using the m6A-related lncRNAs prognostic signature.(A) GO enrichment analysis.(B) Waterfall plot displays mutation information of genes with high mutation frequencies in the low-risk subgroup.(C) Waterfall plot displays mutation information of genes with high mutation frequencies in the high-risk subgroup.(D) Difference of tumor mutation burden (TMB) in the high-and low-risIubgroups.(E) Kaplan-Meier curves of the OS of patients in the high-and low-TMB subgroups.(F) Kaplan-Meier curves of the OS of patients classified by MRLrisk and TMB.(G) Difference of tumor immune dysfunction and exclusion (TIDE) score in the high-and low-risk subgroups.

Figure 6 .
Figure 6.LINC02154 enhances proliferation of ccRCC cells and their sensitivity to Sunitinib in vitro.(A, B) Evaluation of the cell cycle distribution in 786-O cells using flow cytometry.(C, D) Evaluation of the cell cycle distribution in OS-RC-2 cells using flow cytometry.(E, F) Examination of cell apoptosis in 786-O cells using flow cytometry.(G, H) Examination of cell apoptosis in OS-RC-2 cells using flow cytometry.(I) Assessment of cell viability in 786-O cells under Sunitinib treatment using CCK8 assay.(J) Assessment of cell viability in OS-RC-2 cells under Sunitinib treatment using CCK8 assay.� p < .05,�� p < .01,��� p < .001,���� p < .0001.

Table 1 .
Comparison of clinical characteristics of clear cell renal cell carcinoma (ccRCC) patients in training set and testing set.