EPB41L4A-AS1 regulates cervical cancer by proliferative cells: mendelian randomization and single-cell transcriptomics analyses

Introduction

Cervical cancer is the second most prevalent gynecological malignancy. Even though benefiting from vaccine-driven prevention and cytological screening programs, approximately six hundred thousand cervical cancer cases and 340 thousand deaths occurred globally in 2020 (1). The prognosis of recurrent or metastatic cervical cancer has significantly improved in recent years due to advancements in surgical techniques, radiochemotherapy, and drug therapy. Nevertheless, these improvements are limited (2). Accordingly, a comprehensive grasp of preventive strategies, particularly insights into the mechanisms of tumors and the incidence of cervical cancer, can significantly contribute to the diminution of both the incidence and mortality rates within the general population (3). Abnormal cell cycle progression not only contributes to tumorigenesis, but also integrates with other hallmarks of cancer, such as immune evasion (4) and metabolism remodeling (5), ultimately facilitating tumor development. Although it is widely accepted that cervical cancer cells exhibit rapid division and unrestricted proliferation, the underlying mechanisms of abnormal cell proliferation is still not very clear. The identification and characterization of the proliferative cell population in breast cancer have been achieved through the utilization of spatial transcriptomics and single cell RNA sequencing (scRNA-seq) techniques (6). However, there is a dearth of studies focusing on the proliferative cell population in cervical cancer. Consequently, the role of proliferative cells in cervical cancer and the mechanisms underlying their generation remain unknown.

Previous studies indicate that long non-coding RNAs (lncRNAs) are involved in cell cycle and cell proliferation (7). EPB41L4A-AS1 is a P53 inducible lncRNA. Evidences declare that EPB41L4A-AS1 regulates various carcinogenesis-related activities, including metabolic remodeling (8), cell proliferation, migration, and invasion (9). Although many studies have proved an association between EPB41L4A-AS1 expression and cell proliferation, the reports on its role in cell growth are contradictory. Studies declare that EPB41L4A-AS1 inhibits cell proliferation in breast and lung cancer. However, the evidence argues that EPB41L4A-AS1 enhances cell proliferation in osteosarcoma, colorectal cancer, and bone marrow-derived mesenchymal stem cells. These controversial data should be further discussed.

Mendelian randomization (MR) emerges as a potent methodology for inferring causal relationships between exposure and clinical outcomes (10). Moreover, MR is widely utilized in the examination of causal connections between gene expression and clinical outcomes. For example, Chignon et al. employed a two-sample MR approach to identify 16 blood eGenes that are causally associated with the lifespan (11). In this study, we studied the causal relationship of proliferative cells in cervical cancer using MR. Analysis of scRNA-seq data revealed a high proportion of proliferative cells in cervical cancer tissues. Genes that were specifically expressed in various cell types were subjected to MR and meta-analysis, which yielded results indicating a causal association between cervical cancer and proliferative cells, as opposed to columnar or squamous epithelial cells. The knockdown of EPB41L4A-AS1 resulted in enhanced cell proliferation. The genes regulated by EPB41L4A-AS1 were predominantly enriched in proliferative cells rather than columnar or squamous epithelial cells. Consequently, our study offered novel perspectives on the regulation of cervical cancer by EPB41L4A-AS1 through its impact on proliferative cells. We present this article in accordance with the MDAR and STROBE-MR reporting checklists (available at https://tcr.amegroups.com/article/view/10.21037/tcr-24-949/rc).

Methods

Cell culture and transfected

HeLa (SCSP-504) cells were purchased from the cell bank of the Chinese academy of science (Shanghai, China). The cells were cultured in the Dulbecco’s modified Eagle’s medium (DMEM, Bio-Channel, Nanjing, China; Cat. No. BC-M-005) containing with 10% fetal bovine serum (FBS, ExCell Bio, Suzhou, China; Cat. No. FSP500) in a 5% CO₂-humidified incubator at 37 ℃. SiEPB41L4A-AS1 was transiently transfected using siLentFect lipid reagent (Bio-Rad, Cat. No. 1703361) as described by the manufacturer’s protocol. HeLa cells were infected with lentivirus containing shEPB41L4A-AS1 and then selected with 1 µg/mL puromycin (Thermo Fisher Scientific, Waltham, USA; Cat. No. A1113803) to obtain an EPB41L4A-AS1 stable knockdown cell line. The sequences of SiRNAs showed as following: SiEPB41L4A-AS1#1 5'-GGAUGUCCUUGGUGAGGAU-3'; SiEPB41L4A-AS1#2 5'-GGCCUUACCGUAUAACUGA-3'.

Datasets

UCSC Xena (http://xena.ucsc.edu/) is an online platform storing multiomics data (12). The gene expression profiles of EPB41L4A-AS1 and DNA methyltransferases, DNA methylation levels, and PARADIGM integrated pathway levels (IPLs) of pathway features on cervical cancer in the The Cancer Genome Atlas (TCGA) cohort were downloaded from the UCSC Xena database. The Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/) is a public functional genomics data repository. The scRNA-seq data, EPB41L4A-AS1 expression, and clinical information of cervical cancer were downloaded from the GEO database as well. The association between DNA methylation levels and cervical cancer patient prognosis was analyzed using the EWAS data hub database (https://ngdc.cncb.ac.cn/ewas/datahub) (13). The expression profiles of EPB41L4A-AS1 across different types of healthy tissues were analyzed using the UCSC genome browser database (http://www.genome.ucsc.edu/). The EPB41L4A-AS1 expression between cancer tissues and paired healthy tissues was analyzed using the GEPIA database (http://gepia.cancer-pku.cn/) (14). Online tool of Human Cell Atlas (HCA, https://data.humancellatlas.org/) (15) and CellMarker database (http://bio-bigdata.hrbmu.edu.cn/CellMarker/) (16) were used to study the expression of EPB41L4A-AS1 in scRNA-seq data of cervical cancer. The prognosis values of marker genes specifically expressed in proliferative cells, columnar and squamous epithelial cells were predicted using Home for Researchers database (https://www.home-for-researchers.com/static/index.html#/). Comprehensive pathways of EPB41L4A-AS1-regulated cells were analyzed using Metascape database (https://metascape.org/gp/index.html#/main/step1). This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

ScRNA-seq analysis

The scRNA-seq data for cervical cancer (GSE208653, GSE168652) were obtained from GEO database and subsequently analyzed using the Seurat package (Appendix 1: Code of scRNA-seq). Cells with nFeature_RNA count larger than 200 and smaller than 3,000 were selected to further study. Genes exhibiting high expression across different cells were analyzed using FindAllMarkers function. The identification of cell types was accomplished by referencing the CellMarker database and considering the highly expressed genes. To map EPB41L4A-AS1-regulated genes onto cell types, the multilevel function within the fast gene set enrichment analysis (fgsea) package was employed, and the resulting data were visualized using a scatter plot. The enrichment of EPB41L4A-AS1-regulated pathways was displayed using heatmap.

MR and meta-analysis

Single nucleotide polymorphisms (SNPs) that were independent (r²<0.01) and exhibited strong association with the expression of marker genes (P<1×10⁻⁵, F statistics >10) were chosen as instrumental variables, and were downloaded from www.eqtlgen.org/cis-eqtls.html. With these stringent criteria, only 9 out of the 15 marker genes in proliferative cells had suitable SNPs for conducting the MR study. Similarly, only 7 and 3 marker genes in columnar and squamous epithelial cells, respectively, were included in the study. The effect size (beta) and standard error (SE) were calculated based on the Z-score, sample size, and allele frequency of each SNP, using the equation previously reported by Chignon et al. (11). SNPs associated with cervical cancer were obtained from the MRC-IEU consortium, which comprises 462,933 samples and 9,851,867 SNPs. MR analysis (Appendix 1: Code of MR) was conducted using the two sample MR package. Heterogeneity was assessed using the MR_heterogeneity function, and all SNPs included in this study exhibited no significant heterogeneity (P>0.05). The presence of horizontal pleiotropy was evaluated using the global test of the MR-PRESSO package, and all SNPs in this study showed no evidence of horizontal pleiotropy (P>0.05). The causal relationship between genes and cervical cancer was visualized using forest plots. The meta-analysis was performed using Review Manager 5 software, employing a random-effects model.

Gene set enrichment analysis (GSEA)

The GSEA analysis was developed using GSEA software (version 4.3.2). Briefly, the cervical cancer tissues of TCGA cohort were split into two groups according to the median expression of EPB41L4A-AS1. Gene sets of “h.all.v2023.1.Hs.symbols.gmt” and “c5.all.v2023.1.Hs.symbols” were downloaded from MSigDB database (https://www.gsea-msigdb.org/gsea/msigdb/index.jsp) and then performed the GSEA analysis between these two group samples. Because several pathways in the gene set of “c5.all.v2023.1.Hs.symbols” were enriched with P<0.05, these results were displayed using a bubble chart.

fgsea

fgsea analysis (Appendix 1: Code of fgsea) was developed using the fgsea package (version 1.16.0) (17). Briefly, cell proliferation-related gene sets were downloaded from the MSigDB database and then performed the fgsea analysis. The cervical cancer tissues of TCGA cohort were divided into two groups base on the median level of EPB41L4A-AS1. The cervical cancer tissues with EPB41L4A-AS1 expression level of greater than or equal to the median were considered as high expression groups. On the contrary, cervical cancer tissues were considered as low expression groups. Differentially expressed genes between these two groups were analyzed using the edgeR package (Appendix 1: Code of edgeR) and then arranged according to the fold-change. The fgsea was finished using fgseaLabel function, and the minSize, nperm, and maxSize parameters were set to 15, 10,000, and 5,000, respectively. Gene sets with P<0.05 were considered significant enrichment and subsequently displayed using the plotGseaTable function

Weighted gene co-expression network analysis (WGCNA)

WGCNA analysis (Appendix 1: Code of WGCNA) was developed using the WGCNA package (version 1.67) (18). Briefly, 1,000 differentially expressed genes with the highest absolute fold-change value were selected to calculate co-expression modules using plotDendroAndColors function of WGCNA package. Following, genes in each module were extracted to perform Gene Ontology (GO) analysis, which was finished using the enrichGO function in the clusterProfiler package (version 4.3) (19). The result of GO items with P<0.05 was considered significantly enriched and subsequently displayed using a bubble chart.

Quantitative real-time polymerase chain reaction (qRT-PCR)

The total RNA of HeLa cells was extracted using RNA iso Plus (Takara, Okinawa, Japan; Cat. No. D9108B). Next, reverse transcription of RNA was exploited using SweScript RT I First Strand cDNA Synthesis Kit (Servicebio, Wuhan, China; Cat. No. G3330-50). The qRT-PCR was performed using 2× SYBR Green qPCR Master Mix (High ROX) (Servicebio, Cat. No. G332205). Finally, data were normalized using GAPDH, and 2^−ΔΔCt was used to calculate relative fold-change. These processes were repeated in triplicate independently.

The sequences of primers showed as following: EPB41L4A-AS1, forward 5'-ACTGGCACTTCTCCCTCCG-3', reverse 5'-ACAGGCTTCCGTCCCACAA-3'; E2F2, forward 5'-CGTCCCTGAGTTCCCAACC-3', reverse 5'-GCGAAGTGTCATACCGAGTCTT-3'; CDK6, forward 5'-GCTGACCAGCAGTACGAATG-3', reverse 5'-GCACACATCAAACAACCTGACC-3'; GAPDH, forward 5'-GAAGGTGAAGGTCGGAGTC-3', reverse 5'-GAAGATGGTGATGGGATTTC-3'.

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay

HeLa cells were instantaneously transfected with SiNC or SiEPB41L4A-AS1 for 6 h. Then, cells were transferred into a 96-well, with 5,000 cells per well. After 62 h, 10 µL of MTT (Merck, Darmstadt, Germany; 57360-69-7) with 5 mg/mL of final concentration was added into the culture medium and cultured for another 4 h. Following, 100 µL of DMSO was added into each well and incubated for 10 min at room temperature. Finally, the absorbance of cells was measured using a microplate reader.

Flow cytometry

HeLa cells were instantaneously transfected with SiNC or SiEPB41L4A-AS1 for 72 h. Then, cells were fixed using cold 1 mL PBS containing 70% ethanol at 4 ℃ overnight. The second day, cells were harvested using a cell scraper and then centrifuged for 1,500 g at 4 ℃. The supernatant was removed and the cells were stained with 20 µL propidium iodide solution (Merck, 25535-16-4) for 10 min on ice. Finally, the DNA content of cells was analyzed using BD Influx™ Flow Cytometer (BD Biosciences, New Jersey, USA).

Statistical analysis

Statistical analysis was performed using GraphPad Prism software (version 8.0). Differences between the two groups were analyzed using the Mann-Whitney U test, unpaired two-tailed Student’s t-test. The association between the two groups was analyzed using Spearman correlation analyses. The result with P<0.05 was considered statistically significant.

Results

The proportion of proliferative cells in cervical cancer tissue was high

To investigate the contribution of distinct cell types in cervical cancer, the analysis of single-cell transcriptomics was conducted. A total of 9,329 cells obtained from 3 cancerous tissues and 6 normal tissues (GSE208653) led to the identification of 16 major cell types based on their marker genes (Figure 1A, Figure S1A). Following this, a comparative examination of the cell types between tumor and healthy tissue unveiled an elevated proportion of vascular endothelial cells, mesenchymal cells, Treg cells, tumor-associated macrophage, squamous epithelial cells, epithelial cells, and proliferative cells in cervical cancer tissues. On the contrary, a higher proportion of Th2 cells, pan-macrophage, vascular smooth muscle cells, erythrocyte, CD4⁺ T cells, and columnar epithelial cells was observed in healthy tissue (Figure 1B, Figure S1B). A similar analysis was performed on the GSE168652 dataset, encompassing one cervical cancer tissue and one normal adjacent tissue (Figure 1C, Figure S2A). The findings indicated a heightened proportion of columnar epithelial cells, follicular B cells, proliferative cells, epithelial cells, monocyte, plasma cells, mesenchymal progenitor cells, CD4⁺ T cells, and cervical squamous cells in cervical cancer tissue. Conversely, a greater proportion of tumor-associated macrophage, pro-tumor type 2 pericyte, naive CD4 T cells, vascular endothelial cells, mesenchymal cells, and vascular smooth muscle cells was observed in the adjacent normal tissue (Figure 1D, Figure S2B). Given the observed increase in the proportion of proliferative cells in cancer tissues across both datasets, further attention was directed towards these cells. A set of genes that exhibited high expression in proliferative cells was designated as the proliferative cell signature. Subsequent GSEA analysis demonstrated a significant enrichment of the proliferative cell signature in cancer tissues (Figure 1E), aligning with the notion that cancer cells possess sustained proliferative signaling.

Figure 1 The proportion of proliferative cells in cervical cancer tissue was high. (A) and (C) 2D-tSNE graphs comparatively displayed the distributions of different cells (left panel) and tissues (right panel) in scRNA-seq data (A: GSE208653, C: GSE168652) of cervical cancer. (B,D) The proportion of different cell types in scRNA-seq data (B: GSE208653, D: GSE168652) of cervical cancer. (E) GSEA enrichment score curve showing that the proliferative cells signature was significantly enriched in cervical cancer tissues. The green curve represents the enrichment score. NES, normalized enrichment score; 2D-tSNE, two dimensional-t-distributed stochastic neighbor embedding; scRNA-seq, single cell RNA sequencing; GSEA, gene set enrichment analysis; TEMRA, memory cells with terminal differentiation effect reexpressing CD45RA.

Proliferative cells causally associated with cervical cancer

To delve deeper into this phenomenon, a subset of fifteen genes that were specifically expressed in proliferative cells were extracted for further investigation (Figure 2A). A multivariate Cox proportional model was employed to examine the prognostic significance of these fifteen genes in cervical cancer patients. The analysis revealed that patients with a higher risk score exhibited a poorer survival rate [hazard ratio (HR) =1.893, P=0.009] (Figure 2B), suggesting that proliferative cells may serve as a potential biomarker for predicting the prognosis of cervical cancer patients. The similar study was performed in columnar and squamous epithelial cells. The result uncovered that marker genes of squamous epithelial cells rather than columnar epithelial cells could predict the prognosis of cancer patients (Figure 2A, Figure S3A,S3B). Consequently, these specifically expressed genes were designated as exposure factor in order to investigate the causal association between their expression and cervical cancer using MR (Figure 2C). Specifically, DLGAP5, TYMS, BIRC5, and RRM2 were considered causally associated with cervical cancer. A comprehensive meta-analysis of all genes supported the causal relationship between these genes and cervical cancer (Figure 2D), indicating a potential causal association between proliferative cells and cervical cancer. Additionally, the analysis was extended to include squamous and columnar epithelial cells, but neither cell type demonstrated a causal association (Figure S3C,S3D).

Figure 2 Proliferative cells were considered causally associated with cervical cancer. (A) Heatmap showing genes specifically expressed in proliferative cells, squamous or columnar epithelial cells. The expression levels of genes were displayed as color intensity. (B) The Kaplan-Meier survival curve showing the prognostic value of genes specifically expressed in proliferative cells. (C) The scheme of Mendelian randomization analysis. (D) Forest plots indicating the effects on cervical cancer of genes specifically expressed in proliferative cells. IVW, inverse-variance weighted; HR, hazard ratio; CI, confidence interval; OR, odds ratio.

EPB41L4A-AS1 regulated genes were enriched predominantly in proliferative cells

Previous research has demonstrated that lncRNAs play a crucial role in the regulation of tumor occurrence and development through cell proliferation. One such lncRNA, EPB41L4A-AS1, is induced by the P53 protein. However, there are conflicting reports regarding its impact on cell proliferation, necessitating further investigation. In order to explore the relationship between EPB41L4A-AS1 and proliferative cells, a cDNA microarray analysis was conducted to identify genes that were regulated by EPB41L4A-AS1. The result revealed that EPB41L4A-AS1 influenced the expression of 1,416 genes (|fold change| >2 and P<0.05), with 937 genes being upregulated and 479 genes being downregulated genes (Figure 3A, table available at https://cdn.amegroups.cn/static/public/tcr-24-949-1.xls). Subsequently, the biological functions of these 1,416 genes were annotated using GenMAPP and KEGG analyses. The result indicated that several pathways, including cell cycle, were enriched (Figure 3B,3C). A total of 19 genes involved in cell cycle pathway were regulated by EPB41L4A-AS1, with eight genes being upregulated and 11 genes being downregulated in EPB41L4A-AS1 knockdown cells (Figure 3D). To validate the results obtained from the cDNA microarray analysis, qRT-PCR assay was performed to detect the expression levels of two cell cycle-related genes, E2F2 and CDK6, in HeLa cells transfected with SiEPB41L4A-AS1 for 48 h. The result revealed a significant increase in CDK6 and E2F2 levels upon EPB41L4A-AS1 knockdown (Figure 3E), suggesting that EPB41L4A-AS1 might regulate cell proliferation via cell cycle related genes.

Figure 3 EPB41L4A-AS1 regulated expression levels of cell cycles related genes. (A) Scatter plot showing the result of cDNA microarray in HeLa cells transfected with SiEPB41L4A-AS1 for 48 h. The red and green dots represent significantly upregulated and downregulated genes, respectively. The black dots represent no significance. (B,C) Pie chart presenting the top 15 pathways enriched in EPB41L4A-AS1-regulated genes, analyzed by GenMAPP pathway (B) and KEGG analyses (C). (D) Heatmap showing 19 genes involved in cell cycle pathway by cDNA microarray analysis. The color intensity represents base 10 logarithms of expression value. The red and green colors indicate upregulated and downregulated, respectively. (E) The qRT-PCR analysis of E2F2 and CDK6 expression in HeLa cells transfected with SiEPB41L4A-AS1 for 48 h. The data are represented as means ± SD. **, P<0.01; ***, P<0.001, unpaired, two-tailed, Student’s t-test. cDNA, complementary DNA; GenMAPP, gene microarray pathway profiler; KEGG, Kyoto Encyclopedia of Genes and Genomes; qRT-PCR, quantitative real-time polymerase chain reaction; SD, standard deviation; MAPK, mitogen-activated protein kinase; siNC, small interfering RNA negative control; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; mRNA, messenger RNA; GPCRDB, G protein-coupled receptor database; TGF, transforming growth factor.

Next, an over-representation analysis (ORA) was conducted on a total of 937 upregulated genes using Metascape tool. The results of this analysis indicated a significant enrichment of synapse-related pathways, such as synaptic signaling, inorganic ion transmembrane transport, regulation of neuron projection development, neuronal system, brain development, modulation of chemical synaptic transmission, cell surface interactions at the vascular wall, regulation of synapse organization, and nervous system development (Figure 4A). On the contrary, the 479 downregulated genes exhibited a strong association with mitosis-associated pathways, such as mitotic cell cycle, microtubule cytoskeleton organization, chromatin organization (Figure 4B). Furthermore, an GSEA enrichment analysis was performed on the scRNA-seq data (GSE208653) pertaining to the aforementioned 937 upregulated and 479 downregulated genes. This analysis revealed a predominant enrichment of upregulated genes in proliferative cells and squamous epithelial cells within cancerous tissues. In contrast, the downregulated genes exhibited a predominant enrichment in proliferative cells and CD4⁺ T cells within normal and cancerous tissues, respectively (Figure 4C). Subsequently, the ORA pathways associated with EPB41L4A-AS1-regulated genes were mapped to proliferative cells. The result observed that pathways of mitotic cell cycle, microtubule cytoskeleton organization and chromatin organization were enriched in proliferative cells (Figure 4D). In summary, our findings demonstrated that EPB41L4A-AS1-regulated genes were significantly enriched in proliferative cells, suggesting a regulatory role of EPB41L4A-AS1 in proliferative cells.

Figure 4 EPB41L4A-AS1-regulated genes were enriched in proliferative cells. (A,B) Metascape database analyzing pathways of upregulated (A) and downregulated (B) genes in HeLa cells with EPB41L4A-AS1 knockdown. (C) Enrichment of EPB41L4A-AS1-regulated genes across various types of cells in normal (left panel) and cervical cancer tissues (right panel) (GSE208653). The significance and normalized enrichment score were assessed using GSEA. (D) Mapping pathways of EPB41L4A-AS1-regulated genes in proliferative cells. The color intensity represents the normalized enrichment score. The blue and red texts represent pathways of downregulated and upregulated genes, respectively. GSEA, gene set enrichment analysis; NES, normalized enrichment score; IL, interleukin; TGF, transforming growth factor; PIP, phosphatidylinositol; SUMO, small ubiquitin-related modifier.

EPB41L4A-AS1 expression decreased in cervix cancer

The scRNA-seq data of cervical cancer (GSE168652) revealed that EPB41L4A-AS1 expression in single cell of cancer tissue was higher than that of normal tissue (Figure 5A). CellMarkrer database revealed that EPB41L4A-AS1 expression in tumor cells was lower than that in other cells (Figure 5B). Similarly, the scRNA-seq data of cervical cancer reported by Liu et al. (20) also showed similar trend, low expressed in malignant cells (Figure 5C). The expression of EPB41L4A-AS1 in cancer tissues was further explored using GEPIA database. EPB41L4A-AS1 level was significantly downregulated in the cervix, breast, ovary, and stomach cancer tissues, and significantly upregulated in kidney cancer, thymoma, and large B-cell lymphoma tissues (Figure 5D,5E). The same analysis was performed in cervical cancer tissues from the GEO database. The EPB41L4A-AS1 expression showed a similar trend, downregulated in cervix cancer tissues (Figure 5F). In summary, we observed a phenomenon that EPB41L4A-AS1 expression was low in cervical cancer.

Figure 5 EPB41L4A-AS1 expression was downregulated in cervical cancer. (A) The expression level of EPB41L4A-AS1 in scRNA-seq data (GSE168652) of cervical cancer. Each point indicates one cell. (B,C) The 2D-umap (B) and 2D-tSNE (C) graphs comparatively displayed the distributions of different cells (left panel) and EPB41L4A-AS1 expression (right panel) in cervical cancer tissues. The cell cluster in red circle indicates cancer cells. The color in left panel represents cell subpopulation. The intensity of red color in right panel represents the level of EPB41L4A-AS1. (D) EPB41L4A-AS1 expression profiles across different tumors and paired normal tissues in the GEPIA database. Each dot represents EPB41L4A-AS1 expression in one tissue. The red and green dots represent cancer and healthy tissue, respectively. The black text indicates that EPB41L4A-AS1 is not significantly expressed between cancer and normal tissues. The green and red texts indicate that EPB41L4A-AS1 expression is significantly low and high in cancer tissues, respectively. (E,F) Mann-Whitney U analysis of EPB41L4A-AS1 expression between cervical cancer tissues and matched healthy tissues in the GEPIA (E) and GEO database (F). *, P<0.05; ***, P<0.001. CESC, cervical cancer; scRNA-seq, single cell RNA sequencing; 2D-umap, two dimensional-uniform manifold approximation and projection; 2D-tSNE, two dimensional-t-distributed stochastic neighbor embedding; GEPIA, Gene Expression Profiling Interactive Analysis; GEO, Gene Expression Omnibus; Umap, uniform manifold approximation and projection; TPM, transcripts per million.

We also examined the EPB41L4A-AS1 level across different types of tissues. The ovary had the highest level of EPB41L4A-AS1 expression in the healthy tissues of the GTEx cohort, followed by the brain and the cervix (Figure S4A). The EPB41L4A-AS1 level in ocular melanomas was the highest among 30 types of cancer from TCGA database (Figure S4B). The EPB41L4A-AS1 expression was highest in tumor cells derived from autonomic ganglia among all cancer cell lines from CCLE cohort (Figure S4C).

EPB41L4A-AS1 knockdown promoted cell proliferation in vitro

The association between EPB41L4A-AS1 expression and tumor sizes of cervical cancer was investigated. The result observed that EPB41L4A-AS1 expression in cervical cancer tissues with large tumor sizes was low (Figure 6A). Next, whether EPB41L4A-AS1 regulated cell proliferation in cultured cells was studied. HeLa cells were transiently knocked down using siEPB41L4A-AS1 for 48 h and then analyzed by MTT and flow cytometry assays (Figure 6B). The MTT result revealed that EPB41L4A-AS1 suppression increased cellular viability (Figure 6C). The flow cytometry assay observed that EPB41L4A-AS1 knockdown cells revealed an increase in the G2/M phase (Figure 6D), suggesting EPB41L4A-AS1 knockdown promoted cell proliferation. To avoid off-target effects of siRNA, we developed EPB41L4A-AS1 stable knockdown cells in HeLa cells using EPB41L4A-AS1 shRNA (Figure 6E). Stable knockdown of EPB41L4A-AS1 triggered an increase of cell number during M phase and a decrease of cell number during G1/G0 phase (Figure 6F). In conclusion, EPB41L4A-AS1 knockdown promoted cell proliferation.

Figure 6 EPB41L4A-AS1 knockdown promoted cell proliferation in vitro. (A) Mann-Whitney U analysis of EPB41L4A-AS1 expression in cervical cancer tissues with large and small tumor sizes. (B) The qRT-PCR analysis of EPB41L4A-AS1 expression in HeLa cells transfected with SiEPB41L4A-AS1 for 48 h. MTT assay (C) and flow cytometry assay (D) analysis of cell proliferation and cell cycle in HeLa cells transfected with SiEPB41L4A-AS1 for 48 h, respectively. The purple, green, and pink indicate the cells during G1, S, and G2 phases. (E) The qRT-PCR analysis of EPB41L4A-AS1 expression in HeLa cells with EPB41L4A-AS1 stable knockdown. (F) The flow cytometry assay analysis of cell cycle in HeLa cells with EPB41L4A-AS1 stable knockdown. The purple, green, and pink indicate the cells during G1, S, and G2 phases. The data are represented as means ± SD. *, P<0.05; **, P<0.01; ***, P<0.001, unpaired, two-tailed, Student’s t-test. qRT-PCR, quantitative real-time polymerase chain reaction; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; G1, Gap1 phase; S, DNA synthesis phase; G2, Gap2 phase; SD, standard deviation; siNC, small interfering RNA negative control; sh-NC, short hairpin RNA negative control; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Association between EPB41L4A-AS1 level and cell cycle-related pathways

To further confirm function of EPB41L4A-AS1, GSEA analysis was performed in TCGA cervical cancer. The result presented that mitotic-related pathways were enriched in tissues with low EPB41L4A-AS1 expression (Figure 7A,7B). Additionally, fgsea was applied to compare EPB41L4A-AS1 expression to cell growth-related gene sets in the same sample cohort. The result revealed that cell proliferation-related gene sets were significantly enriched in EPB41L4A-AS1 (Figure 7C), suggesting an association between EPB41L4A-AS1 and cell proliferation. We next performed the WGCNA analysis in the same sample cohort to confirm our conclusion. Initially, we analyzed differentially expressed genes between cervical cancer tissues with low and high EPB41L4A-AS1 expression. Following, 1,000 genes with the highest absolute value of fold-change were selected to calculate gene co-expression modules by WGCNA analysis. The result identified four different co-expression modules (Figure 7D). Analyzing the biological pathways of genes in each module revealed that cell proliferation-related pathways were specially enriched in genes of blue module (Figure 7E), demonstrating an association between EPB41L4A-AS1 expression and cell proliferation-related pathways.

Figure 7 The association between EPB41L4A-AS1 expression and cell cycle-related pathways. (A) GSEA enrichment score curve showing that the mitotic spindle pathway was significantly enriched in cervical cancer patients with low EPB41L4A-AS1 expression. The green curve represents the enrichment score. (B) GSEA analysis of several cell cycle-related gene sets were enriched in cervical cancer patients with low EPB41L4A-AS1 expression. The circle size indicates NES. The color intensity represents the negative base 10 logarithms of the P. (C) fgsea displaying EPB41L4A-AS1 expression evaluated cell proliferation-related gene sets in cervical cancer tissues. (D) WGCNA analyzing co-expression modules of genes differently expressed between cervical cancer tissues with high and low EPB41L4A-AS1 expression. (E) Bubble plot showing the biological pathways of genes enriched in the blue module by GO analysis. The red text indicates cell proliferation-related pathways. The circle size indicates the gene count of each pathway. The color intensity represents the negative base 10 logarithms of the P. GSEA, gene set enrichment analysis; NES, normalized enrichment score; fgsea, fast gene set enrichment analysis; WGCNA, weighted gene co-expression network analysis; GO, Gene Ontology; FDR, false positive rate.

PARADIGM IPLs could be used to infer the activation of pathway features (21). Therefore, we investigated association between EPB41L4A-AS1 expression and PARADIGM IPLs of cell cycle-related pathways in TCGA cervical cancer tissues. The result observed that PARADIGM IPLs of four cell cycle-related pathways were low in cervical cancer tissues with high EPB41L4A-AS1 expression (Figure S5A). Moreover, their IPLs were negatively correlated with EPB41L4A-AS1 expression (Figure S5B). Collectedly, our findings suggested that EPB41L4A-AS1 expression was negatively correlated with activation of cell cycle-related pathways. We also studied the association between EPB41L4A-AS1 expression and cell cycle-related genes in the same sample cohort. The result discovered that CDK6, LONP2, and ESPL1 expressions were high in cervical cancer tissues with low EPB41L4A-AS1 expression (Figure S5C). Their expression was negatively correlated with EPB41L4A-AS1 expression as well (Figure S5D). Analysis of the association between EPB41L4A-AS1 expression and their activation discovered that PARADIGM IPLs of CDK6 and LONP2 were high in cervical cancer tissues with low EPB41L4A-AS1 expression (Figure S5E). Besides, PARADIGM IPLs of CDK6 and LONP2 were negatively correlated with EPB41L4A-AS1 expression (Figure S5F). Our findings confirmed that EPB41L4A-AS1 level was negatively correlated with expression and activation of cell cycle-related genes, suggesting that low EPB41L4A-AS1 level might promote cell proliferation.

The negative association of DNA methylation values with EPB41L4A-AS1 level

Next, we questioned why EPB41L4A-AS1 expression was low in cervical cancer. DNA methylation is widely known to inhibit gene expression at the transcriptional level. Therefore, we investigated the relationship between DNA methylation and EPB41L4A-AS1 expression in TCGA cervical cancer tissues. The result discovered that DNA methylation values, described as beta values of four array probes, were lower in cervical cancer tissues with high EPB41L4A-AS1 expression (Figure S6A,S6B). Additionally, DNA methylation values negatively correlated with EPB41L4A-AS1 expression (Figure S6C). Subsequently, we analyzed the association between DNA methyltransferases and EPB41L4A-AS1 expression in the same sample cohort. The result observed that DNMT1 and DNMT3A were lower in cervical cancer tissues with high EPB41L4A-AS1 expression (Figure S6D). Figure S6E revealed a negative association of EPB41L4A-AS1 level with DNMT1 or DNMT3B expression. Cervical cancer patients with higher DNA methylation level at EPB41L4A-AS1 gene had a poor prognosis (Figure S6F). Collectedly, our result confirmed a negative association between DNA methylation and EPB41L4A-AS1 level.

Discussion

Proliferative cells are a class of cells that exhibit positive expression of the cell proliferation marker MKI67. In breast cancer, the presence of a “proliferative” cell population is associated with the advancement of the disease (6). In heart tissue, proliferative cell population can be generated from fibroblasts (22). While numerous studies have established a correlation between cell proliferative activity and cervical cancer (23), limited research has specifically investigated the role and generation of proliferative cell populations in this type of cancer. This study identified proliferative cell population that exhibited positive expression of MKI67 in cervical cancer using scRNA-seq data. Given that cancer cells possess the ability to evade growth suppressors and sustain proliferative signaling, it is unsurprising that the proportion of proliferative cells within tumor tissue was elevated. Survival analysis demonstrated that proliferative cells could serve as a biomarker for predicting the prognosis of cervical cancer patients. Furthermore, causal relationship analysis indicated a potential causal link between proliferative cells and the development of cervical cancer. All of these findings provided evidence supporting the notion that proliferative cells play crucial roles in the development of cervical cancer and patient survival (Figure 8).

Figure 8 The workflow of this study. GSEA, gene set enrichment analysis; fgsea, fast gene set enrichment analysis; WGCNA, weighted gene co-expression network analysis; cDNA, complementary DNA; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; scRNA-seq, single-cell RNA sequencing.

EPB41L4A-AS1 is a P53 induced lncRNA and contributes to many biological processes. The relationship between EPB41L4A-AS1 and cell proliferation is conflicting. EPB41L4A-AS1 overexpression significantly reduced the cancer cell proliferation in breast and lung cancer (9,24). In contrast, EPB41L4A-AS1 knockdown inhibited cell proliferation in osteosarcoma (25) and colorectal cancer (26). EPB41L4A-AS1 overexpression enhanced cell proliferation in bone marrow-derived mesenchymal stem cells (27). In cervical cancer, evidence demonstrated that EPB41L4A-AS1 knockdown plus with glutaminase inhibitor treatment suppressed tumor progression in vivo (8). This study indicated that the suppression of EPB41L4A-AS1 led to an increase in the number of cells during the M phase, thereby promoting cell proliferation. EPB41L4A-AS1 knockdown regulated the expressions of cell cycle-related genes, such as CDK6 and E2F2. Furthermore, both upregulated and downregulated genes by EPB41L4A-AS1 were found to be enriched in proliferative cells. The comparison of ORA pathways of these genes and proliferative cells indicated an enrichment of mitotic-related pathways in the proliferative cells, thereby suggesting a regulatory role of EPB41L4A-AS1 in proliferative cells. Given the low expression of EPB41L4A-AS1 in cancer cells and tumor tissues, it was plausible that the reduced levels of EPB41L4A-AS1 in tumor cells might induce differential gene expression, ultimately resulting in the generation of proliferative cells. Nevertheless, further research is required to elucidate the underlying mechanism of proliferative cell generation.

Previous studies indicated that EPB41L4A-AS1 regulates cell metabolism, including glycolysis, glutaminolysis, glucose uptake, and fatty acid oxidation, via histone acetylation, crotonylation, and methylation modification (8,28-30). Additionally, EPB41L4A-AS1 can influence the migration and invasion of cancer cells (9,26). Besides, EPB41L4A-AS1 has been associated with inflammatory response and apoptosis (9,31). In this study, the result of cDNA microarray annotation indicated that EPB41L4A-AS1-regulated genes were involved in metabolic and synthesis pathways, electron transport chain, amino acid metabolism, carbohydrate metabolism, matrix metalloproteinases, cell adhesion molecules (CAMs), apoptosis, and inflammatory response pathways, consistent with previous studies.

EPB41L4A-AS1 has low expression across several cancers. Previous studies considered that is probably because the EPB41L4A-AS1 chromatin region is frequently deleted. Besides, the p53 gene regulates EPB41L4A-AS1 expression. P53 is frequently mutated in cancers, which may lead to EPB41L4A-AS1 downregulation. Additionally, miR-146a inhibits the endogenous EPB41L4A-AS1 expression (27). DNA methylation, an epigenetic signature, has been hypothesized to suppress gene expression by recruiting gene repression-related proteins to DNA or by inhibiting the interaction between transcription factors and DNA (32). In this study, we uncovered that the levels of DNA methyltransferases and DNA methylation were lower in cervical cancer tissues with high EPB41L4A-AS1 expression than those with low EPB41L4A-AS1 expression. We observed negative associations between EPB41L4A-AS1 expression and DNA methylation level or methyltransferase expression. These findings suggested that DNA methylation may inhibit EPB41L4A-AS1 expression. Changes in DNA methylation serve as cancer diagnostic, prognostic, and predictive biomarkers (33). This study discovered that cervical cancer patients with high DNA methylation levels at EPB41L4A-AS1 gene had poor prognoses. This result is consistent with the hypothesis that DNA methylation represses EPB41L4A-AS1 expression because low EPB41L4A-AS1 expression is correlated with poor prognosis in cervical cancer patients.

There are some limitations in this study. The precise mechanism underlying the generation of proliferative cells within tumor cells and tissues, as well as the mechanism of the low expression of EPB41L4A-AS1 in several cancers, warrant further investigation by seasoned researchers.

Conclusions

In summary, we reported a causal relationship between proliferative cells and cervical cancer, and highlighted a key role of EPB41L4A-AS1 in proliferative cells generation, which might provide a new perspective for the occurrence of cervical cancer.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the MDAR and STROBE-MR reporting checklists. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-24-949/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-24-949/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-24-949/prf

Funding: This work was supported by the Youth Program of National Natural Science Foundation of China (No. 31900540) and Jiangsu Training Program of Innovation and Entrepreneurship for Undergraduates (No. 202310313059Z).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-24-949/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Singh D, Vignat J, Lorenzoni V, et al. Global estimates of incidence and mortality of cervical cancer in 2020: a baseline analysis of the WHO Global Cervical Cancer Elimination Initiative. Lancet Glob Health 2023;11:e197-206. [Crossref] [PubMed]
2. D'Oria O, Bogani G, Cuccu I, et al. Pharmacotherapy for the treatment of recurrent cervical cancer: an update of the literature. Expert Opin Pharmacother 2024;25:55-65. [Crossref] [PubMed]
3. Ferrari F, Giannini A. Approaches to prevention of gynecological malignancies. BMC Womens Health 2024;24:254. [Crossref] [PubMed]
4. Li J, Stanger BZ. Cell Cycle Regulation Meets Tumor Immunosuppression. Trends Immunol 2020;41:859-63. [Crossref] [PubMed]
5. Liu J, Peng Y, Shi L, et al. Author Correction: Skp2 dictates cell cycle-dependent metabolic oscillation between glycolysis and TCA cycle. Cell Res 2021;31:104. [Crossref] [PubMed]
6. YoshitakeRMoriHHaDIdentification and characterization of a proliferative cell population in estrogen receptor-positive metastatic breast cancer through spatial and single-cell transcriptomics.bioRxiv 2023. doi: .
7. Zhu Q, Zhang C, Qu T, et al. MNX1-AS1 Promotes Phase Separation of IGF2BP1 to Drive c-Myc-Mediated Cell-Cycle Progression and Proliferation in Lung Cancer. Cancer Res 2022;82:4340-58. [Crossref] [PubMed]
8. Liao M, Liao W, Xu N, et al. LncRNA EPB41L4A-AS1 regulates glycolysis and glutaminolysis by mediating nucleolar translocation of HDAC2. EBioMedicine 2019;41:200-13. [Crossref] [PubMed]
9. Yang F, Lv S. LncRNA EPB41L4A-AS1 Regulates Cell Proliferation, Apoptosis and Metastasis in Breast Cancer. Ann Clin Lab Sci 2022;52:3-11.
10. Holmes MV, Ala-Korpela M, Smith GD. Mendelian randomization in cardiometabolic disease: challenges in evaluating causality. Nat Rev Cardiol 2017;14:577-90. [Crossref] [PubMed]
11. Chignon A, Bon-Baret V, Boulanger MC, et al. Single-cell expression and Mendelian randomization analyses identify blood genes associated with lifespan and chronic diseases. Commun Biol 2020;3:206. [Crossref] [PubMed]
12. Goldman MJ, Craft B, Hastie M, et al. Visualizing and interpreting cancer genomics data via the Xena platform. Nat Biotechnol 2020;38:675-8. [Crossref] [PubMed]
13. Xiong Z, Li M, Yang F, et al. EWAS Data Hub: a resource of DNA methylation array data and metadata. Nucleic Acids Res 2020;48:D890-5. [Crossref] [PubMed]
14. Li C, Tang Z, Zhang W, et al. GEPIA2021: integrating multiple deconvolution-based analysis into GEPIA. Nucleic Acids Res 2021;49:W242-6. [Crossref] [PubMed]
15. Mamanova L, Miao Z, Jinat A, et al. High-throughput full-length single-cell RNA-seq automation. Nat Protoc 2021;16:2886-915. [Crossref] [PubMed]
16. Hu C, Li T, Xu Y, et al. CellMarker 2.0: an updated database of manually curated cell markers in human/mouse and web tools based on scRNA-seq data. Nucleic Acids Res 2023;51:D870-6. [Crossref] [PubMed]
17. KorotkevichGSukhovVBudinNFast gene set enrichment analysis.bioRxiv 2021. doi: .
18. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics 2008;9:559. [Crossref] [PubMed]
19. Yu G, Wang LG, Han Y, et al. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012;16:284-7. [Crossref] [PubMed]
20. Liu X, Ni G, Zhang P, et al. Single-nucleus RNA sequencing and deep tissue proteomics reveal distinct tumour microenvironment in stage-I and II cervical cancer. J Exp Clin Cancer Res 2023;42:28. [Crossref] [PubMed]
21. Hoadley KA, Yau C, Hinoue T, et al. Cell-of-Origin Patterns Dominate the Molecular Classification of 10,000 Tumors from 33 Types of Cancer. Cell 2018;173:291-304.e6. [Crossref] [PubMed]
22. Wang J, Gu S, Liu F, et al. Reprogramming of fibroblasts into expandable cardiovascular progenitor cells via small molecules in xeno-free conditions. Nat Biomed Eng 2022;6:403-20. [Crossref] [PubMed]
23. Xu Y, Chen X, Pan S, et al. TM7SF2 regulates cell proliferation and apoptosis by activation of C-Raf/ERK pathway in cervical cancer. Cell Death Discov 2021;7:299. [Crossref] [PubMed]
24. Dong B, Zhang F, Zhang W, et al. IncRNA EPB41L4A-AS1 Mitigates the Proliferation of Non-Small-Cell Lung Cancer Cells through the miR-105-5p/GIMAP6 Axis. Crit Rev Eukaryot Gene Expr 2023;33:27-40. [Crossref] [PubMed]
25. Yan Y, Liu X, Li Y, et al. EPB41L4A-AS1 and UNC5B-AS1 have diagnostic and prognostic significance in osteosarcoma. J Orthop Surg Res 2023;18:261. [Crossref] [PubMed]
26. Bin J, Nie S, Tang Z, et al. Long noncoding RNA EPB41L4A-AS1 functions as an oncogene by regulating the Rho/ROCK pathway in colorectal cancer. J Cell Physiol 2021;236:523-35. [Crossref] [PubMed]
27. Cui P, Zhao X, Liu J, et al. miR-146a interacting with lncRNA EPB41L4A-AS1 and lncRNA SNHG7 inhibits proliferation of bone marrow-derived mesenchymal stem cells. J Cell Physiol 2020;235:3292-308. [Crossref] [PubMed]
28. Zhu Y, Liu Q, Liao M, et al. Overexpression of lncRNA EPB41L4A-AS1 Induces Metabolic Reprogramming in Trophoblast Cells and Placenta Tissue of Miscarriage. Mol Ther Nucleic Acids 2019;18:518-32. [Crossref] [PubMed]
29. Liao W, Xu N, Zhang H, et al. Persistent high glucose induced EPB41L4A-AS1 inhibits glucose uptake via GCN5 mediating crotonylation and acetylation of histones and non-histones. Clin Transl Med 2022;12:e699. [Crossref] [PubMed]
30. Chai W, Wang X, Lu Z, et al. Elevated exosome-transferrable lncRNA EPB41L4A-AS1 in CD56(bright) NK cells is responsible for the impaired NK function in neuroblastoma patients by suppressing cell glycolysis. Clin Immunol 2023;250:109322. [Crossref] [PubMed]
31. Wang Z, Liao W, Liu F, et al. Downregulation of lncRNA EPB41L4A-AS1 Mediates Activation of MYD88-Dependent NF-κB Pathway in Diabetes-Related Inflammation. Diabetes Metab Syndr Obes 2021;14:265-77. [Crossref] [PubMed]
32. Moore LD, Le T, Fan G. DNA methylation and its basic function. Neuropsychopharmacology 2013;38:23-38. [Crossref] [PubMed]
33. Müller D, Győrffy B. DNA methylation-based diagnostic, prognostic, and predictive biomarkers in colorectal cancer. Biochim Biophys Acta Rev Cancer 2022;1877:188722. [Crossref] [PubMed]