Contribution of m5C RNA Modification-Related Genes to Prognosis and Immunotherapy Prediction in Patients with Ovarian Cancer

Liu, Yibin; Liu, Shouze; Yan, Lu; Zhang, Qianqian; Liu, Wenhua; Huang, Xianghua; Liu, Shikai

doi:https://doi.org/10.1155/2023/1400267

Mediators of Inflammation

On this page

Abstract Introduction Materials and Methods Results Discussion Conclusion Data Availability Conflicts of Interest Authors’ Contributions Supplementary Materials References Copyright Related Articles

Research Article | Open Access

Volume 2023 | Article ID 1400267 | https://doi.org/10.1155/2023/1400267

Contribution of m5C RNA Modification-Related Genes to Prognosis and Immunotherapy Prediction in Patients with Ovarian Cancer

Yibin Liu,¹Shouze Liu,^1,2Lu Yan,¹Qianqian Zhang,³Wenhua Liu,⁴Xianghua Huang,¹and Shikai Liu²

Academic Editor: Helen C. Steel

Received12 May 2023

Revised03 Sept 2023

Accepted16 Oct 2023

Published13 Nov 2023

Abstract

Background. 5-Methylcytosine (m5C) RNA modification is closely implicated in the occurrence of a variety of cancers. Here, we established a novel prognostic signature for ovarian cancer (OC) patients based on m5C RNA modification-related genes and explored the correlation between these genes with the tumor immune microenvironment. Methods. Methylated-RNA immunoprecipitation sequencing helped us to identify candidate genes related to m5C RNA modification at first. Based on TCGA database, we screened the differentially expressed candidate genes related to the prognosis and constructed a prognostic model using LASSO Cox regression analyses. Notably, the accuracy of the model was evaluated by Kaplan–Meier analysis and receiver operator characteristic curves. Independent prognostic risk factors were investigated by Cox proportional hazard model. Furthermore, we also analyzed the biological functions and pathways involved in the signature. Finally, the immune response of the model was visualized in great detail. Results. Totally, 2,493 candidate genes proved to be involved in m5C modification of RNA for OC. We developed a signature with prognostic value consisting of six m5C RNA modification-related genes. Specially, samples have been split into two cohorts with low- and high-risk scores according to the model, in which the low-risk OC patients exhibited dramatically better overall survival time than those with high-risk scores. Besides, not only was this model a prognostic factor independent of other clinical characteristics but it predicted the intensity of the immune response in OC. Significantly, the accuracy and availability of the signature were verified by ICGC database. Conclusions. Our study bridged the gap between m5C RNA modification and the prognosis of OC and was expected to provide an effective breakthrough for immunotherapy in OC patients.

1. Introduction

As one of the most common gynecological malignant tumors, ovarian cancer (OC) kills more than 200,000 women annually [1]. The high-mortality rate of OC is closely related to its asymptomatic nature, with a 5-year survival rate of less than 50% [2]. Cytoreductive surgery combined with chemotherapy remains the gold standard of treatment for OC. However, chemotherapy resistance followed by intraperitoneal dissemination still leads to unpredictable deaths, possibly due to the neglect of tumor cell heterogeneity, thus treating it as a single disease [3]. Therefore, personalized treatment seems to be a feasible strategy for improving the quality of prognosis in the patients with OC.

Being considered as an emerging gene expression regulation method, posttranscriptional RNA modifications are involved in the development of a range of human diseases, including cancer, which may develop to be an ideal target for cancer therapy [4, 5]. RNA epigenetic modifications mainly involve N1-methyladenosine (m1A), pseudouridine (ψ), 5-methylcytosine (m5C), 5-hydroxymethylcytosine (hm5C), N6-methyladenosine (m6A), and N7-methylguanosine (m7G) [5]. Characterized by the attachment of a methyl group to the fifth carbon atom in cytosine residues, RNA m5C participates in the metabolism and structural stability of RNA and the recognition of transfer RNAs (tRNA), which allows dynamic regulation of cellular responses to adapt to rapidly changing microenvironments. For example, adaptation to the chemotherapeutic drugs plays a vital role in ensuring the survival of tumor cells [6, 7], suggesting an affinity between RNA m5C and cancer cells. Recent studies have shown that RNA m5C performs well in predicting prognosis and immune infiltration characteristics in triple-negative breast cancer [8], cervical cancer [9], hepatocellular carcinoma [10], colorectal cancer [11, 12], clear cell renal cell carcinoma [13], and lung squamous cell carcinoma and adenocarcinoma [14, 15], the role of RNA m5C in OC prognosis remains to be elucidated.

In the present study, we attempted to expand our understanding of the effects of RNA m5C-related genes in the affecting of overall survival (OS) in OC. In addition, efforts have been made to reveal the potential role of these genes in the changes in the immune microenvironment and chemosensitivity of OC, which might help in improving treatment strategies.

2. Materials and Methods

2.1. Sample Collection and Ethical Approval

Three pairs of high-grade serous ovarian cancer (HGSOC) samples and corresponding adjacent tissues were obtained from the Second Hospital of Hebei Medical University, and all tissues were histopathologically confirmed by pathologists. After surgery, the HGSOC samples and adjacent normal tissues were immediately stored at −80°C for further use. This study was approved by the Ethics Committee and Institutional Review Board of the Second Hospital of Hebei Medical University. Written informed consent was obtained from all participants before the study began. This study was conducted in accordance with the Declaration of Helsinki.

2.2. Detection of m5C RNA Modification-Related Genes

First, we used TRIzol reagent (Invitrogen Corporation, Carlsbad, CA) to extract RNA. The m5C-methylated mRNA fragments were enriched by immunoprecipitation with an anti-m5C antibody (Synaptic Systems). The NEBNext Ultra II Directional RNA Library Prep Kit (New England Biolabs, Inc.) was used to construct the cDNA library, whose quality and purity were evaluated using the Agilent Bioanalyzer 2100 system (Agilent Technologies, Inc., USA). Sequentially, DCC and MACS (v1.4.2) were applied to identify circular RNA (circRNA) peaks and methylated m5C peaks on the circRNAs. Differentially, m5C methylation peaks were filtered with a fold change >2 or < 0.5 (p-value < 0.00001) using diffReps software. The circRNA region where the m5C peak was located was calculated for the two cohorts. Statistical significance was set at . Finally, we combined transcriptome and methylation analyses to identify the effects of methylation on transcriptional expression.

2.3. Original Data Acquisition

The RNA-seq data from 88 normal samples and 379 OC tissues were extracted from the Genotype-Tissue Expression (GTEx) database (https://xenabrowser.net/) and The Cancer Genome Atlas (TCGA) (https://portal.gdc.cancer.gov/), respectively. The corresponding clinical information was also downloaded. Then we unified the two sets of expression data from different sources using “limma” package in R 4.1.1 and got normalized datasets [16]. We collected the patients’ informed consent for the publication of the data used in the experiment. Additionally, the public materials involved in this study were downloaded free of charge from the public database; therefore, ethical review and approval are optional from the Ethics Committee.

2.4. Data Normalization

For the purpose of integrating the expression data from GTEx and TCGA database, we utlized the batch normalization to correct unwanted technical variation. At first, log2(x + 0.001) transformed RNA-seq FPKM values of ovary tissue were extracted from GTEx database. In the meanwhile, we obtained the RNA-seq FPKM data of TCGA-OV patients. Then, these expression data from both databases were log2(x + 1) transformed. Finally, batch normalization was performed across abovementioned data by the function named normalizeBetweenArrays in “limma” package in R 4.1.1 [17].

2.5. Calculation of Differentially Expressed Genes (DEGs) Related to Survival Time

We performed the DEGs screening process using R with the filtering criteria of false discovery rate (FDR) < 0.01 and |log2 FC|≥ 2. p-Values were adjusted to control for FDR using the Benjamini–Hochberg method, which can enhance the reliability of the results [18]. At the same time, univariate Cox regression analysis was applied to determine the target genes valuable for prognosis. Finally, the core genes in this signature were generated after intersection of the two gene sets.

2.6. Development of the Prognostic Model

With the aid of Lasso-penalized Cox regression analysis, we calculated the risk coefficient (coef) of each m5C-related gene and determined each sample’s risk score combined with gene expression. Each OC patient was assigned a risk score, which = gene1 (coef expression) + gene2 (coef expression) + .. gene7 (coef expression). As such, we built an m5C-related model in the TCGA cohort to predict the risk characteristics of OC patients, in which those with a risk score less than the median value were recognized as the low-risk group, whereas the others were recognized as the high-risk group. To evaluate the accuracy of this signature in prognostic value, we plotted Kaplan–Meier (K–M) survival curves and time-dependent receptor operating characteristic (ROC) curves using “survival” and “timeROC” package in R. Similarly, the median score in TCGA was used as a cutoff value to distinguish high- and low-risk groups in The International Cancer Genome Consortium (ICGC) (https://dcc.icgc.org/), and survival differences between two risk statuses were explored further following above steps.

2.7. Evaluation for the Performance of the Signature

Principal component analysis (PCA) and t-distributed Stochastic Neighbor Embedding (t-SNE) were employed to investigate whether OC patients could be significantly distinguished based on the m5C-related risk scores. Multivariate Cox analysis was used to determine whether m5C-related risk scores could be an independent risk factor affecting the prognosis of patients was figured out using multivariate Cox analysis. After that, we constructed a nomogram to predict the 1-, 3-, and 5-year OS utilizing “rms” package in R.

2.8. Biological Function and Pathway Enrichment Analysis

According to our previous method [19], the gene sets of differentially expressed genes between the two risk cohorts were prepared for the subsequent steps followed by the thresholds: and logFC = 1. We used the Kyoto Encyclopedia of Genes and Genomes (KEGG) and gene ontology (GO) analyses to label different signaling pathways and functions, respectively. The annotations were revealed using package “ggplot2” package in R.

2.9. Immune Infiltration Exploration

Discrepancies in the abundance of immunocytes and immune-related functions between low- and high-risk populations were assessed using single-sample gene set enrichment analysis (ssGSEA). Of note, 13 immunocyte types and 13 immunological functions were involved in the quantification of immune infiltration. And we controlled the entire process via R package “limma,” “ggpubr,” and “reshape2.”

2.10. Drug Sensitivity Prediction

Combining the m5C RNA modification-related genes with drug sensitivity, we assessed the therapeutic response to chemotherapy for OC patients in our signature, which was reflected by half-maximal inhibitory concentration (IC₅₀) values. To complete the whole procedure, “pRRophetic” package in R was applied in this prediction.

3. Results

3.1. Generation of Model Genes in OC

Conjoint analysis of transcriptomics and methylome data captured 2,493 m5C RNA modification-related genes from the prophase experiment (Supplementary 1), of which 897 were calculated as DEGs between tumor and normal samples in TCGA (Figure 1(a)) and 52 were identified to be significantly correlated with survival time (Supplementary 2). Finally, six genes remained (ATP1A3, GRIN2D, PLA2G2D, GALNT10, LAMP3, and GALNT6) after the intersection of the two gene sets (Figure 1(b)). The details are provided in a forest plot and heat map (Figures 1(c) and 1(d)).

(a)

(b)

(c)

(d)

(e)

Figure 1

Generation and characteristics of model genes. (a) Green and red dots: genes of significant differentially expressed (|log2 FC|≥ 2 with false discovery rate (FDR) < 0.01); black dots: normally expressed genes. (b) The Venn plot for core genes. (c) The forest plot of core genes. (d) The heat map for expression levels of core genes. (e) The correlation network between core genes and RNA m5C regulators. Blue lines represent negative correlation; red lines represent positive correlation.

3.2. Association Analysis between Hub Genes and m5C RNA Methylation Regulators

By means of “igraph” and “reshape2” packages in R, the correlation network captured the features of intimate connectivity among the candidate genes (Figure 1(e)). Significantly, ATP1A3, GRIN2D, LAMP3, and GALNT6 were positively correlated with m5C writers. GALNT10 was positively correlated with an eraser but negatively associated with the writer and reader genes. As shown in Supplementary 3, there is a weak positive correlation between PLA2G2D and the regulator of writers [20].

3.3. Construction and Evaluation of m5C RNA Modification-Related Signature in OC

By listing the coef values of the six genes (Table 1), we can use the formula mentioned above to compute the risk score of each sample in both training TCGA (Figure 2(a)) and testing ICGC (Figure 2(b)). We noticed distributional differences in OC patients, characterized by the risk scores in PCA and t-NSE (Figure 2(c)–2(f)). In addition, as the risk score increased, the spread of patients shifted toward a decreased survival time and higher mortality (Figures 2(g) and 2(h)). And we got to know from K–M survival curves that patients with low-risk scores took on significantly better survival outcomes than the others: training cohort (p = 2.633e⁻¹⁰) (Figure 2(i)) and testing cohort (p = 2.001e⁻⁰²) (Figure 2(j)). Crucially, it was deniable that the ROC curves showed an accurate predictive capability of this signature in 1-, 2- and 3- years OS with high-area under the curve (AUC) values. The AUC were 0.635, 0.693, and 0.670 in the training group and 0.721, 0.700, and 0.661 in the testing group, respectively (Figures 2(k) and 2(l)). Based on these results, our model demonstrated robust predictive performance for OS in general OC patients.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

(k)

(l)

Figure 2

Prognostic signature model structuring and validating. (a, b) The risk scores and distribution of patients from TCGA (a) and ICGC (b). (c, d) PCA plots of the training (c) and testing cohorts (d). (e, f) Results of t-SNE analysis in the training (e) and testing cohorts (f). (g, h) Survival status distribution of patients from TCGA (g) and ICGC (h). (i, j) Kaplan–Meier curves for the training (i) and testing groups (j). (k, l) the ROC curves in the training (k) and the testing sets (l).

3.4. Comprehensive Investigation of Risk Factors Affecting Survival in OC Patients

Taking age, risk score, and grade into univariate and multivariate analyses using Cox regression, we recognized age and risk score as independent factors for OS in TCGA cohort (Figures 3(a) and 3(b)). In the ICGC cohort, although age showed little significance for prognosis, the RNA modification-related signature still displayed superior value for survival prediction in both univariate (hazard ratio (HR) : 1.938; 95% confidence interval (CI) : 1.193 − 3.148) and multivariate (HR : 1.884; 95% CI : 1.159 − 3.062) Cox regression analyses (Figures 3(c) and 3(d)). Furthermore, aware that a single risk score would contribute to the insufficient prediction; we devised a nomogram plot composed of risk scores and age. In the training cohort, there was a high degree of agreement between the observed and nomogram-predicted survival rates at 1, 3, and 5-year (Figure 3(e)). Similarly, the calibration curves of the nomogram showed good congruence between the x- and y-axes in the validation cohort (Figure 3(f)).

(a)

(b)

(c)

(d)

(e)

(f)

3.5. Molecular Mechanisms of the Difference between Two Risk Status

A total of 82 DEGs were used for GO and KEGG analyses (Supplementary 4). As shown in Figure 4(a), they are of paramount importance for immunological biological processes, such as “humoral immune response,” “immunoglobulin production,” and “production of molecular mediators of immune response”. And the cellular compounds results indicated that these genes participated in the formation of “immunoglobulin complex,” “blood microparticle,” and “immunoglobulin complex, circulating”. While “antigen binding,” “receptor ligand activity,” and “signaling receptor activator activity” were significantly enriched for the category of molecular functions. Essentially, KEGG results emphasized the connection between the signature and immune-related signaling pathways, including the IL-17 signaling pathway, chemokine signaling pathway, toll-like receptor signaling pathway, and NF-κB signaling pathway (Figure 4(b)).

(a)

(b)

3.6. The Landscape of Immune Infiltration in Two OC Subgroups

Considering that the risk score was close to immune signatures, we compared the immune diversity between the two risk states (Figure 5). Patients with high-risk scores tended to have reduced levels of tumor-infiltrating immune cells, involving dendritic cells (DCs), activated DCs, plasmacytoid DCs, CD8⁺ T cells, tumor-infiltrating lymphocytes (TILs), type 1/2 T helper (Th1/2) cells, T follicular helper (Tfh) cells, B cells, and NK cells (Figure 5(a)). In other words, these high-abundance immune cells might account for the favorable prognosis in OC patients. Moreover, all immune functions, which were significantly different between the two risk subtypes, exhibited more active patterns in low-risk patients, as expected (Figure 5(b)). In any case, the ssGSEA results demonstrated that low-risk scores were associated with enhanced immune profiles.

(a)

(b)

3.7. Chemotherapy Response Identification

To further explore antitumor drugs, we collated IC₅₀ values of several compounds based on the Genomics of Drug Sensitivity in Cancer database (https://www.cancerrxgene.org/). As a consequence, we found multiple medicaments more effective in high-risk people (Figure 6), which included ponatinib, axitinib, AZ628 (a pan-Raf kinase inhibitor), saracatinib, bexarotene, bicalutamide, BMS-754807 (an insulin-like growth factor-1R/IR inhibitor), bryostatin 1, CHIR-99021 (a potent GSK-3 inhibitor), dasatinib, embelin, erlotinib, FH535 (a PPAR and Wnt/β-catenin inhibitor), FTI-277 (an inhibitor of Farnesyltransferase), GNF-2 (a BCR-ABL inhibitor), GSK269962A (an inhibitor of ROCK), GW-441756 (a TrkA inhibitor), imatinib, midostaurin, NVP-TAE684 (an ALK inhibitor), linsitinib, palbociclib, PF-562271 (an inhibitor of FAK), refametinib, WH-4-023 (a dual LCK/SRC inhibitor), tanespimycin, XMD8.85 (an inhibitor of ERK5 and LRRK2). Admittedly, these discoveries extended future prospects of OC therapy, in which the needs of prolonging survival of high-risk patients may be satisfied.

4. Discussion

Over the years, m5C RNA modification has gradually emerged as a prominent channel that influences tumor initiation and progression. Simultaneously, efforts have been made to improve the feasibility of m5C RNA modification in predicting the prognosis of clear cell renal cell carcinoma [21], lung squamous cell carcinoma [15], triple-negative breast cancer [8], and prostate adenocarcinoma [22]. However, few studies have focused on its role as an exclusive biomarker for OC on the one hand [23]. However, there are also limitations to access m5C RNA modification-related genes, which are confined to writers, erasers, and readers [23]. In comparison, we performed methylated-RNA immunoprecipitation sequencing (MeRIP-seq) to identify related genes in our research, other than, as in the similar studies, the involved genes were directly obtained from the previous findings. Therefore, we have provided more reliable and comprehensive strategies. We attached great importance to m5C RNA modification, which has not yet been reported. In the current study, we identified differentially expressed m5C RNA modification-related genes between human OC and adjacent non-tumor tissues, which was devised for the first time using the MeRIP-seq method. By making full use of TCGA and experimental data, we established a prognostic model encompassing six novel RNA m5C-related genes: GALNT10, ATP1A3, PLA3G2D, GRIN2D, LAMP3, and GALNT6. Subsequently, we categorized the OC patients into low- and high-risk subtypes in terms of coef values and the corresponding gene expression levels in TCGA. Subsequent analyses reached an absolute consensus on the significantly negative correlation between risk scores and survival advantage, which was validated using the ICGC dataset. In addition, the immune profile was dramatically different between the low- and high-risk populations. In conclusion, our results not only lay a solid foundation for the refinement of prognostic stratification but also exert positive effects on the immunotherapy in OC patients.

GALNT10 belongs to the N-acetylgalactosaminyltransferase (GalNAc-T) family, the encoded protein of which plays a catalytic role in O-glycan synthesis [24]. Subtle alterations in GALNT10 give rise to aberrant O-glycosylation, thus facilitating tumor cell proliferation in hepatocellular carcinoma and gastric cancer [24, 25]. In vitro cell assays have shown that the stem cell-like characteristics of OC cells are regulated by GALNT10, hinging on its glycosyltransferase property [26]. In addition, it has been reported that OC patients are more likely to experience dismal prognosis and immune suppression with elevated GALNT10 expression [27]. Moreover, it dovetails precisely with our findings. Regrettably, the role of GALNT10 in the regulation of RNA modification remains unclear. However, it must be noted that GALNT10 represents an adverse factor for OC patients’ survival, according to our data. Similarly, it is commonly recognized that GALNT6, another GalNAc-T family member, impedes tumor growth and progression when losing expression [28, 29]. Interestingly, previous studies have emphasized that OC patients tend to experience adverse outcomes upon increased GALNT6 expression [30, 31], which is in contrast to our present report. Undoubtedly, a completely different argument does cause a controversial issue about the GALNT6 functions of in OC prognosis, awaiting further exploration. As a widespread transmembrane protein, Na⁺/K⁺-ATPase (NKA) is responsible for the maintenance of the electrochemical gradient across plasma membranes, of which ATP1A3 encodes the α3 subunit [32]. It is commonly believed that pathogenic ATP1A3 variants have resulted in a variety of neurological disorders in the last 20 years, such as rapid-onset dystonia parkinsonism, alternating hemiplegia of childhood, and cerebellar ataxia [32]. By targeting ATP1A3, Y-box binding protein 1 (YB-1), and RNA-binding protein Human Antigen R (HuR) participate in the formation of RNA-dependent complexes on m5C modified nucleotides [33]. In addition, recent studies have confirmed that ATP1A3 activation not only enhances temozolomide sensitivity but also inhibits the proliferation of glioma cells immediately [34, 35]. In contrast, overexpressed ATP1A3, which promotes tumor invasion, has been identified in clinical specimens of gastric cancer [36]. More importantly, a recent study indicated that OC patients were more likely to suffer from poor OS when ATP1A3 was upregulated, which agrees well with our findings [37]. Phospholipase A2 (PLA2) proteins are a group of lipid metabolism-related molecules with catalytic properties of that hydrolyze membrane glycerol phospholipids and are therefore essential for various biochemical processes, such as the response to inflammation [38]. On the one hand, PLA2G2D is responsible for encoding a member of secreted PLA2, group IID PLA2, which is associated with resistance to immune checkpoint blockade as well as limitation of the immune system in primary solid tumors [39, 40]. In contrast, a positive correlation between immune infiltration and PLA2G2D expression has been identified in cervical cancer [38]. In addition, upregulation of PLA2G2D improves the prognosis of cutaneous melanoma [41], breast cancer [42], and head and neck squamous cell carcinoma [43]. Together with our results, PLA2G2D represents a good chance of acting as a beneficial regulator rather than an immunosuppressive role in OC. Encoding a subunit of the N-methyl D-aspartate receptor, GRIN2D is highly expressed in hepatocellular cancer and can exert an obstacle effect on the immune response, thus promoting tumor growth [44]. Additionally, GRIN2D can be used as an executor controlled by miR-129-1-3p to regulate breast cancer cell infiltration and migration [45]. Herein, elevated GRIN2D expression was dramatically detected in neoplastic tissues along with a potential marker for poor therapeutic effects in OC. LAMP3, a lysosome-associated membrane protein, is indispensable for sustaining the integrity of lysosome structure and related functions [46]. A recent study indicated that excessive LAMP3 activation could induce the degradation of other transmembrane proteins of the lysosome, which in turn increases membrane permeability, leading to caspase activation and autophagy inhibition [47]. More importantly, LAMP‑3(+) DCs stimulate the proliferation of tumor infiltrating cells CD8 which symbolizes a favorable prognosis in esophageal squamous cell carcinoma [48]. Conversely, they also cause tumor aggressiveness and resistance to endocrine therapy in breast cancer, indicating the opposite effects on cancer types [49]. Although the specific mechanism of LAPM3 in OC cells remains unclear, we preliminarily conclude that LAMP3 serves as a beneficial gene for prolonging the survival of OC patients.

Recently, m5C RNA modification has been recognized to play a role in immune cell biological processes [50]. First, GO analysis showed that the DEGs between the two risk groups objectively participated in the process of canonical immune responses and the formation of prominent immunoglobulins, which also have an impact on antigen and cytokine binding. The signaling pathways, “Chemokine signaling pathway,” “IL-17 signaling pathway,” “Toll-like receptor signaling pathway,” and “NF-kappa B signaling pathway” are all associated with immunomodulation and tumor metabolism. To some extent, functional annotation suggests that immune status can be essentially distinguished between the two risk groups in an m5C RNA modification manner.

In this study, we also noticed that immune cells and related functions were widely retarded in the high-risk group displayed on GSEA results, implying that antitumor immunity is perturbed in OC patients with high-risk scores. As specialized antigen-presenting cells, mature DCs, when stimulated by antigens, activate CD8⁺ T cells to function [51]. The density of CD8⁺ T cells is well known as the most essential antitumor effector associated with superior survival in OC patients [52]. It should be noted that type I IFN secreted by pDCs is a vital anti-cancer cytokine that activates and recruits NK and B cells [53, 54], although pDCs are generally considered intruders in the regulation of T cell responses to cancers [55]. In addition, Tfh cells have been positively correlated with prognosis owing to the promotion of B cell maturation and maintenance of CD8 T cell-dependent antitumor activity. The characteristics of Tfh cells in OC patients remain further refined, but we found that the group with prolonged survival appeared to carry more conspicuous Tfh cell infiltration in OC. Th 1 cells, one of the Th cell subsets, participate in antitumor immunity through the production of inflammatory mediators to eradicate cancer cells directly, while simultaneously promoting the activation and recruitment of NK cells [56, 57]. Th1 cells are particularly relevant for extending recurrence-free survival in OC patients [58]. In addition, a recent update revealed a distinct effector mechanism in that Th 2 cells induce terminal differentiation to osteoblasts and initiate breast cancer [59].

Over the past several years, blocking immune checkpoints has reframed the treatment of a variety of malignancies and has led to remarkable improvements in OS [60]. Accordingly, it seems to be more efficient for the low-risk population to benefit from immune checkpoint inhibitors, as they harbor a greater density of immune checkpoints that mediate immune escape. The immune escape mechanism of multiple solid tumors is mainly due to the loss of antigenicity induced by the deficiency of major histocompatibility complex class I (MHC I)/human leukocyte antigen (HLA) [61–63]. In contrast, an increase in tumor-specific antigen presentation recognized by T cells on HLA/MHC I molecules has been deemed to improve survival outcomes [62]. Our signature revealed impaired HLA/MHC I functions in high-risk OC patients. Moreover, MHC I can hardly be upregulated in poorly differentiated tumor cells under corresponding conditions for activation, a signature that is likely unique to OC [63]. Hence, there is presumably stratification of differentiation between our risk groups to some extent. Alongside these, the higher portion of antitumor-associated immune responses has been observed in the low-risk cohort, such as “APC co inhibition,” “cytolytic activity,” “CCR,” “parainflammation,” “flammation-promoting,” “T cell co-stimulation/inhibition,” and “Type I IFN Reponse.” Overall, the immune profile revealed bridges the gap between immunotherapy and m5C RNA modification, which is of clinical significance in strengthening the applications of immunotherapy in OC.

In addition, we have identified more effective chemotherapeutic drugs for high-risk patients, which may lay a solid foundation for prolonging the survival of the overall population with OC.

5. Conclusion

In conclusion, we have described the survival prediction patterns of six m5C RNA methylation-related genes in OC derived from MeRIP-seq. Based on TCGA, our signature could predict the survival profiles and immune infiltration status of OC patients independently, which has been significantly reproduced in a separate database. More importantly, our findings are expected to provide a novel horizon for stratified management and brilliant immunotherapeutic strategies as well as potential adjuvant chemotherapy for high-risk patients. However, the detailed regulatory mechanism of the candidate genes in m5C RNA methylation remains unclear, and further experimental studies are required.

Data Availability

The data supporting the findings of this study are openly available online. Transcriptome profiling and clinical data of OC patients were downloaded from TCGA (https://portal.gdc.cancer.gov/), the Genotype-Tissue Expression (GTEx) database (https://xenabrowser.net/), and ICGC (https://dcc.icgc.org/). In total, 88 normal ovary tissues in the GTEx, 379 OC samples in the TCGA-OV cohort, and 93 samples in the ICGC-OC-AU cohort were used for further analysis.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Authors’ Contributions

Liu and Liu designed the study, wrote the initial draft, and analyzed the databases. Qianqian Zhang and Wenhua Liu contributed to critical data and edited the language of the manuscript. Lu Yan and Xianghua Huang performed the analyses with constructive discussion. Shikai Liu revised the manuscript critically for important content. All authors approved the final draft of the manuscript. Yibin Liu and Shouze Liu contributed equally to this work.

Supplementary Materials

Supplementary 1. m5C RNA modification-related genes derived from the transcriptomic and methylome data.

Supplementary 2. The forest plot of m5C RNA modification-related genes correlated with survival time.

Supplementary 3. The slightly positive correlation between PLA2G2D and m5C RNA modification regulators.

Supplementary 4. The DEGs were used for functional enrichment analysis.

References

H. Sung, J. Ferlay, R. L. Siegel et al., “Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries,” CA: A Cancer Journal for Clinicians, vol. 71, no. 3, pp. 209–249, 2021.
View at: Publisher Site | Google Scholar
L. Han, S. Husaiyin, J. Liu, M. Maimaiti, M. Niyazi, and L. Li, “Period analysis of intraracial differences in incidence and survival rates in epithelial ovarian cancer,” Computational and Mathematical Methods in Medicine, vol. 2021, Article ID 8032209, 10 pages, 2021.
View at: Publisher Site | Google Scholar
F. Ravindran and B. Choudhary, “Ovarian cancer: molecular classification and targeted therapy,” in Ovarian Cancer, G.-Y. Ho and K. Webber, Eds., IntechOpen, Rijeka, 2021.
View at: Google Scholar
I. A. Roundtree, M. E. Evans, T. Pan, and C. He, “Dynamic RNA modifications in gene expression regulation,” Cell, vol. 169, no. 7, pp. 1187–1200, 2017.
View at: Publisher Site | Google Scholar
I. Barbieri and T. Kouzarides, “Role of RNA modifications in cancer,” Nature Reviews Cancer, vol. 20, no. 6, pp. 303–322, 2020.
View at: Publisher Site | Google Scholar
L. Dou, X. Li, H. Ding, L. Xu, and H. Xiang, “Prediction of m5C modifications in RNA sequences by combining multiple sequence features,” Molecular Therapy–Nucleic Acids, vol. 21, pp. 332–342, 2020.
View at: Publisher Site | Google Scholar
P. Nombela, B. Miguel-López, and S. Blanco, “The role of m(6)A, m(5)C and Ψ RNA modifications in cancer: novel therapeutic opportunities,” Molecular Cancer, vol. 20, no. 1, Article ID 18, 2021.
View at: Publisher Site | Google Scholar
Z. Huang, J. Pan, H. Wang et al., “Prognostic significance and tumor immune microenvironment heterogenicity of m5C RNA methylation regulators in triple-negative breast cancer,” Frontiers in Cell and Developmental Biology, vol. 9, Article ID 657547, 2021.
View at: Publisher Site | Google Scholar
J. Yu, L.-L. Liang, J. Liu et al., “Development and validation of a novel gene signature for predicting the prognosis by identifying m5C modification subtypes of cervical cancer,” Frontiers in Genetics, vol. 12, Article ID 733715, 2021.
View at: Publisher Site | Google Scholar
Y. Liu, S. Zheng, T. Wang, Z. Fang, J. Kong, and J. Liu, “Identification of the expression patterns and potential prognostic role of 5-methylcytosine regulators in hepatocellular carcinoma,” Frontiers in Cell and Developmental Biology, vol. 10, Article ID 842220, 2022.
View at: Publisher Site | Google Scholar
Q. Geng, Q. Wei, Z. Shen et al., “Comprehensive analysis of the prognostic value and immune infiltrates of the three-m5c signature in colon carcinoma,” Cancer Management and Research, vol. 13, pp. 7989–8002, 2021.
View at: Publisher Site | Google Scholar
X. Fang, C. Miao, T. Zeng et al., “Role of m(5) C RNA methylation regulators in colorectal cancer prognosis and immune microenvironment,” Journal of Clinical Laboratory Analysis, vol. 36, no. 4, Article ID e24303, 2022.
View at: Publisher Site | Google Scholar
J. Wu, C. Hou, Y. Wang, Z. Wang, P. Li, and Z. Wang, “Comprehensive analysis of m5C RNA methylation regulator genes in clear cell renal cell carcinoma,” International Journal of Genomics, vol. 2021, Article ID 3803724, 24 pages, 2021.
View at: Publisher Site | Google Scholar
H. Chen, X. L. Ge, Z. Y. Zhang et al., “M(5)C regulator-mediated methylation modification patterns and tumor microenvironment infiltration characterization in lung adenocarcinoma,” Translational Lung Cancer Research, vol. 10, pp. 2172–2192, 2021.
View at: Publisher Site | Google Scholar
J. Pan, Z. Huang, and Y. Xu, “m5C RNA methylation regulators predict prognosis and regulate the immune microenvironment in lung squamous cell carcinoma,” Frontiers in Oncology, vol. 11, Article ID 657466, 2021.
View at: Publisher Site | Google Scholar
Q. Wang, J. Armenia, C. Zhang et al., “Unifying cancer and normal RNA sequencing data from different sources,” Scientific Data, vol. 5, no. 1, Article ID 180061, 2018.
View at: Publisher Site | Google Scholar
M. E. Ritchie, B. Phipson, D. Wu et al., “limma powers differential expression analyses for RNA-sequencing and microarray studies,” Nucleic Acids Research, vol. 43, no. 7, Article ID e47, 2015.
View at: Publisher Site | Google Scholar
S. R. Shuken and M. W. McNerney, “Costs and benefits of popular p-value correction methods in three models of quantitative omic experiments,” Analytical Chemistry, vol. 95, no. 5, pp. 2732–2740, 2023.
View at: Publisher Site | Google Scholar
S. Liu, Q. Zhang, W. Liu, and X. Huang, “Prediction of prognosis in patients with endometrial carcinoma and immune microenvironment estimation based on ferroptosis-related genes,” Frontiers in Molecular Biosciences, vol. 9, Article ID 916689, 2022.
View at: Publisher Site | Google Scholar
Y.-S. Chen, W.-L. Yang, Y.-L. Zhao, and Y.-G. Yang, “Dynamic transcriptomic m(5) C and its regulatory role in RNA processing,” WIREs RNA, vol. 12, no. 4, Article ID e1639, 2021.
View at: Publisher Site | Google Scholar
H. Li, H. Jiang, Z. Huang, Z. Chen, and N. Chen, “Prognostic value of an m(5)C RNA methylation regulator-related signature for clear cell renal cell carcinoma,” Cancer Management and Research, vol. 13, pp. 6673–6687, 2021.
View at: Publisher Site | Google Scholar
Z. Xu, S. Chen, Y. Zhang, R. Liu, and M. Chen, “Roles of m5C RNA modification patterns in biochemical recurrence and tumor microenvironment characterization of prostate adenocarcinoma,” Frontiers in Immunology, vol. 13, Article ID 869759, 2022.
View at: Publisher Site | Google Scholar
P. Zheng, N. Li, and X. Zhan, “Ovarian cancer subtypes based on the regulatory genes of RNA modifications: novel prediction model of prognosis,” Frontiers in Endocrinology, vol. 13, Article ID 972341, 2022.
View at: Publisher Site | Google Scholar
G. Xu, Y.-L. Wu, N. Li et al., “GALNT10 promotes the proliferation and metastatic ability of gastric cancer and reduces 5-fluorouracil sensitivity by activating HOXD13,” European Review for Medical and Pharmacological Sciences, vol. 24, no. 22, pp. 11610–11619, 2020.
View at: Publisher Site | Google Scholar
Q. Wu, H.-O. Liu, Y.-D. Liu et al., “Decreased expression of hepatocyte nuclear factor 4α (Hnf4α)/microRNA-122 (miR-122) axis in hepatitis B virus-associated hepatocellular carcinoma enhances potential oncogenic GALNT10 protein activity,” Journal of Biological Chemistry, vol. 290, no. 2, pp. 1170–1185, 2015.
View at: Publisher Site | Google Scholar
F.-Y. Zhao, Q. Zhang, J.-M. Wang et al., “BAG3 epigenetically regulates GALNT10 expression via WDR5 and facilitates the stem cell-like properties of platin-resistant ovarian cancer cells,” Biochimica et Biophysica Acta (BBA)–Molecular Cell Research, vol. 1868, no. 9, Article ID 119077, 2021.
View at: Publisher Site | Google Scholar
G. Zhang, J. Lu, M. Yang, Y. Wang, H. Liu, and C. Xu, “Elevated GALNT10 expression identifies immunosuppressive microenvironment and dismal prognosis of patients with high grade serous ovarian cancer,” Cancer Immunology, Immunotherapy, vol. 69, no. 2, pp. 175–187, 2020.
View at: Publisher Site | Google Scholar
R. Kimura, T. Yoshimaru, Y. Matsushita et al., “The GALNT6‐LGALS3BP axis promotes breast cancer cell growth,” International Journal of Oncology, vol. 56, no. 2, pp. 581–595, 2020.
View at: Publisher Site | Google Scholar
F. Gao and G. Zheng, “RUNX3-regulated GALNT6 promotes the migration and invasion of hepatocellular carcinoma cells by mediating O-glycosylation of MUC1,” Disease Markers, vol. 2022, Article ID 2959846, 11 pages, 2022.
View at: Publisher Site | Google Scholar
R. Sheta, M. Bachvarova, M. Plante et al., “Altered expression of different GalNAc‐transferases is associated with disease progression and poor prognosis in women with high-grade serous ovarian cancer,” International Journal of Oncology, vol. 51, no. 6, pp. 1887–1897, 2017.
View at: Publisher Site | Google Scholar
R. Sheta, M. Bachvarova, E. Macdonald, S. Gobeil, B. Vanderhyden, and D. Bachvarov, “The polypeptide GALNT6 displays redundant functions upon suppression of its closest homolog GALNT3 in mediating aberrant O-glycosylation, associated with ovarian cancer progression,” International Journal of Molecular Sciences, vol. 20, no. 9, Article ID 2264, 2019.
View at: Publisher Site | Google Scholar
A. Vezyroglou, R. Akilapa, K. Barwick et al., “The phenotypic continuum of ATP1A3-related disorders,” Neurology, vol. 99, no. 14, pp. e1511–e1526, 2022.
View at: Publisher Site | Google Scholar
S. Busque, The HuR/YB1 Complex Stabilizes Pro-myogenic mRNAs by Interacting with a Consensus U-rich Motif in the 3′UTR, McGill University, Montreal, 2022.
Y.-L. Lan, C. Chen, X. Wang et al., “Gamabufotalin induces a negative feedback loop connecting ATP1A3 expression and the AQP4 pathway to promote temozolomide sensitivity in glioblastoma cells by targeting the amino acid Thr794,” Cell Proliferation, vol. 53, no. 1, Article ID e12732, 2020.
View at: Publisher Site | Google Scholar
Q. Ji, Z. Tu, J. Liu, K. Huang, X. Zhu, and J. Li, “Identification of a robust scoring system based on metabolic genes followed by in-depth validation of ATP1A3 in glioma,” Life Sciences, vol. 315, Article ID 121377, 2023.
View at: Publisher Site | Google Scholar
L. Li, R. Feng, Q. Xu et al., “Expression of the β3 subunit of Na+/K+-ATPase is increased in gastric cancer and regulates gastric cancer cell progression and prognosis via the PI3/AKT pathway,” Oncotarget, vol. 8, no. 48, pp. 84285–84299, 2017.
View at: Publisher Site | Google Scholar
W. Huang, Y. Zhang, Y. Xu et al., “Comprehensive analysis of the expression of sodium/potassium-ATPase α subunits and prognosis of ovarian serous cystadenocarcinoma,” Cancer Cell International, vol. 20, no. 1, Article ID 309, 2020.
View at: Publisher Site | Google Scholar
H. Liu, R. Xu, C. Gao et al., “Metabolic molecule PLA2G2D is a potential prognostic biomarker correlating with immune cell infiltration and the expression of immune checkpoint genes in cervical squamous cell carcinoma,” Frontiers in Oncology, vol. 11, Article ID 755668, 2021.
View at: Publisher Site | Google Scholar
M. S. Rooney, S. A. Shukla, C. J. Wu, G. Getz, and N. Hacohen, “Molecular and genetic properties of tumors associated with local immune cytolytic activity,” Cell, vol. 160, no. 1-2, pp. 48–61, 2015.
View at: Publisher Site | Google Scholar
S. Y. Cindy Yang, S. C. Lien, B. X. Wang et al., “Pan-cancer analysis of longitudinal metastatic tumors reveals genomic alterations and immune landscape dynamics associated with pembrolizumab sensitivity,” Nature Communications, vol. 12, no. 1, Article ID 5137, 2021.
View at: Publisher Site | Google Scholar
Y. Li, S. Lyu, Z. Gao et al., “Identification of potential prognostic biomarkers associated with cancerometastasis in skin cutaneous melanoma,” Frontiers in Genetics, vol. 12, Article ID 687979, 2021.
View at: Publisher Site | Google Scholar
Z. Ye, S. Zou, Z. Niu, Z. Xu, and Y. Hu, “A novel risk model based on lipid metabolism-associated genes predicts prognosis and indicates immune microenvironment in breast cancer,” Frontiers in Cell and Developmental Biology, vol. 9, Article ID 691676, 2021.
View at: Publisher Site | Google Scholar
K. Du, J. Zou, B. Wang et al., “A metabolism-related gene prognostic index bridging metabolic signatures and antitumor immune cycling in head and neck squamous cell carcinoma,” Frontiers in Immunology, vol. 13, Article ID 857934, 2022.
View at: Publisher Site | Google Scholar
Y.-G. Zhang, M.-Z. Jin, X.-R. Zhu, and W.-L. Jin, “Reclassification of hepatocellular cancer with neural-related genes,” Frontiers in Oncology, vol. 12, Article ID 877657, 2022.
View at: Publisher Site | Google Scholar
Q. Li, Z. Gu, Q. Tan, L. Ren, and S. Chen, “MicroRNA-129-1-3p represses the progression of triple-negative breast cancer by targeting the GRIN2D gene,” BioMed Research International, vol. 2022, Article ID 1549357, 9 pages, 2022.
View at: Publisher Site | Google Scholar
L. P. Lunding, D. Krause, G. Stichtenoth et al., “LAMP3 deficiency affects surfactant homeostasis in mice,” PLoS Genetics, vol. 17, no. 6, Article ID e1009619, 2021.
View at: Publisher Site | Google Scholar
T. Tanaka, B. M. Warner, D. G. Michael et al., “LAMP3 inhibits autophagy and contributes to cell death by lysosomal membrane permeabilization,” Autophagy, vol. 18, no. 7, pp. 1629–1647, 2022.
View at: Publisher Site | Google Scholar
J. Nishimura, H. Tanaka, Y. Yamakoshi et al., “Impact of tumor-infiltrating LAMP-3 dendritic cells on the prognosis of esophageal squamous cell carcinoma,” Esophagus, vol. 16, no. 4, pp. 333–344, 2019.
View at: Publisher Site | Google Scholar
A. Nagelkerke, A. M. Sieuwerts, J. Bussink et al., “LAMP3 is involved in tamoxifen resistance in breast cancer cells through the modulation of autophagy,” Endocrine-Related Cancer, vol. 21, no. 1, pp. 101–112, 2014.
View at: Publisher Site | Google Scholar
L. Cui, R. Ma, J. Cai et al., “RNA modifications: importance in immune cell biology and related diseases,” Signal Transduction and Targeted Therapy, vol. 7, no. 1, Article ID 334, 2022.
View at: Publisher Site | Google Scholar
X. Zhang, T. He, Y. Li et al., “Dendritic cell vaccines in ovarian cancer,” Frontiers in Immunology, vol. 11, Article ID 613773, 2021.
View at: Publisher Site | Google Scholar
G. Zhang, Q. Xu, X. Zhang et al., “Spatial cytotoxic and memory T cells in tumor predict superior survival outcomes in patients with high-grade serous ovarian cancer,” Cancer Medicine, vol. 10, no. 12, pp. 3905–3918, 2021.
View at: Publisher Site | Google Scholar
L. Guéry, J. Dubrot, C. Lippens et al., “Ag-presenting CpG-activated pDCs prime Th17 cells that induce tumor regression,” Cancer Research, vol. 74, no. 22, pp. 6430–6440, 2014.
View at: Publisher Site | Google Scholar
J. Fucikova, L. Palova-Jelinkova, J. Bartunkova, and R. Spisek, “Induction of tolerance and immunity by dendritic cells: mechanisms and clinical applications,” Frontiers in Immunology, vol. 10, Article ID 2393, 2019.
View at: Publisher Site | Google Scholar
S. K. Wculek, F. J. Cueto, A. M. Mujal, I. Melero, M. F. Krummel, and D. Sancho, “Dendritic cells in cancer immunology and immunotherapy,” Nature Reviews Immunology, vol. 20, no. 1, pp. 7–24, 2020.
View at: Publisher Site | Google Scholar
H. Braumüller, T. Wieder, E. Brenner et al., “T-helper-1-cell cytokines drive cancer into senescence,” Nature, vol. 494, no. 7437, pp. 361–365, 2013.
View at: Publisher Site | Google Scholar
H.-J. Kim and H. Cantor, “CD4 T-cell subsets and tumor immunity: the helpful and the not-so-helpful,” Cancer Immunology Research, vol. 2, no. 2, pp. 91–98, 2014.
View at: Publisher Site | Google Scholar
M. S. Block, A. B. Dietz, M. P. Gustafson et al., “Th17-inducing autologous dendritic cell vaccination promotes antigen-specific cellular and humoral immunity in ovarian cancer patients,” Nature Communications, vol. 11, no. 1, Article ID 5173, 2020.
View at: Publisher Site | Google Scholar
M. Boieri, A. Malishkevich, R. Guennoun et al., “CD4+ T helper 2 cells suppress breast cancer by inducing terminal differentiation,” Journal of Experimental Medicine, vol. 219, no. 7, 2022.
View at: Publisher Site | Google Scholar
E. Chang, L. Pelosof, S. Lemery et al., “Systematic review of PD-1/PD-L1 inhibitors in oncology: from personalized medicine to public health,” The Oncologist, vol. 26, no. 10, pp. e1786–e1799, 2021.
View at: Publisher Site | Google Scholar
K. Dhatchinamoorthy, J. D. Colbert, and K. L. Rock, “Cancer immune evasion through loss of MHC Class I antigen presentation,” Frontiers in Immunology, vol. 12, Article ID 636568, 2021.
View at: Publisher Site | Google Scholar
A. Hazini, K. Fisher, and L. Seymour, “Deregulation of HLA-I in cancer and its central importance for immunotherapy,” Journal for Immunology Therapy of Cancer, vol. 9, no. 8, Article ID e002899, 2021.
View at: Publisher Site | Google Scholar
N. Chovatiya, K. Kaur, S. Huerta-Yepez et al., “Inability of ovarian cancers to upregulate their MHC-class I surface expression marks their aggressiveness and increased susceptibility to NK cell-mediated cytotoxicity,” Cancer Immunology, Immunotherapy, vol. 71, no. 12, pp. 2929–2941, 2022.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2023 Yibin Liu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

208

Downloads

255

Citations