The expression landscape and clinical significance of methyltransferase-like 17 in human cancer and hepatocellular carcinoma: a pan-cancer analysis using multiple databases

Abstract

Background

Methyltransferase-like (METTL) family protein plays a crucial role in the progression of malignancies. However, the function of METTL17 across pan-cancers, especially in hepatocellular carcinoma (HCC) is still poorly understood.

Methods

All original data were downloaded from TCGA, GTEx, HPA, UCSC databases and various data portals. First, we comprehensively analyzed RNA-seq data from the HPA database of 25 human tissues. An array of bioinformatics methods was employed to explore the potential oncogenic roles of METTL17, including analyzing its related prognosis, mutation, landscapes, tumor stemness index, immune cell infiltration, and other factors among different tumors. Additionally, gene set enrichment analysis (GSEA) was used to analyze pathways associated with METTL17 in HCC. Immunohistochemistry (IHC) was performed on clinical samples to validate the differential expression of METTL17 in HCC and normal tissues. Ultimately, we constructed a METTL17-related risk-score model of HCC and validated its prognostic classification efficiency. Survival rates were calculated using the Kaplan-Meier method. Statistical significance was defined as P < 0.05.

Results

METTL17 was differentially expressed in various cancers. METTL17 maintained strong correlations with the cancer patient’s prognosis, genetic alterations, tumor stemness index, and immune-infiltrated cells, etc. In addition, IHC experiments verified that METTL expression was significantly decreased in liver tissues of HCC patients compared to normal liver tissue. GESA analysis indicated METTL17 mainly involves oncogenic and immune-related pathways among HCC. MRPS5, CHCHD2, NCBP1, LRPPRC, DAP3, and BMS1 were included in a prognostic model based on METTL17’s interaction networks. Kaplan–Meier survival analysis of the prognostic model showed that the overall survival (OS) of the low-risk group was significantly better than that of the high-risk group (P < 0.001). The area under the receiver operating characteristic (ROC) curve (AUC) of the 1-year, 3-year, and 5-year OS were 0.747, 0.671, and 0.631, respectively.

Conclusions

METTL17 may serve as a novel prognostic marker and therapeutic target for human tumors, offering a theoretical foundation for formulating more effective and tailored clinical treatment options for cancers, particularly HCC.

Background

The methyltransferase-like (METTL) proteins are characterized by their seven-beta-strand (7BS) methyltransferases [1,2,3,4]. METTL17, a member of the METTLs, is a mitochondrial protein that functions as an N4-methylcytidine (m⁴C) methyltransferase within the mitochondria. It increases the stability of 12 S mt-rRNA by selectively methylating C 840 and C 842 [2]. Lysine methylation is a crucial epigenetic alteration that has a significant impact on controlling gene expression and perhaps influences the development of breast tumors [3]. METTL17 is specifically involved in the methylation of lysine residues on histones. As an important epigenetic modification, lysine methylation may regulate mammary tumorigenesis [2]. Previous studies [4] have indicated that the absence of METTL17 leads to impaired formation of mitoribosomes and reduced protein synthesis. Moreover, METTL17 is necessary for the differentiation of embryonic stem cell (ESC) differentiation. Nevertheless, there are only a few researches investigating the potential function of METTL17 [3]. Notably, tumor cells, especially tumor stem cells and ESCs, share many similarities in unlimited proliferation, invasion, migration, escaping from immune surveillance, and certain signaling pathways [2]. Interestingly, METTL17 plays an important role in oxidative phosphorylation in K562 cells, a type of human myeloid leukemia. Additionally, the knockout of METTL17 inhibits breast cancer cell proliferation [3]. However, there is a lack of comprehensive analysis of the relationship between METTL17 and human cancer.

Liver cancer is the second most common cause of cancer-related deaths worldwide [5]. Hepatocellular carcinoma (HCC) is the primary histological subtype of liver cancer, characterized by numerous genetic and epigenetic alterations that contribute to its initiation and / or progression [5]. The dysregulation of gene transcription, which includes gene mutations and the activation of oncogenes, is widely recognized to promote tumor metastasis and eventually impact the survival of HCC patients. Thus, it is crucial to understand the molecular mechanism to create innovative diagnostic, therapeutic solutions and eventually decrease the mortality rates of HCC. METTL proteins exhibit a wide array of substrates and numerous members are strongly linked to tumorigenesis [2]. For instance, high METTL1 expression in HCC is associated with large tumor size, high alpha-fetoprotein (AFP) levels, vessel invasion, and unfavorable prognosis [6]. Furthermore, upregulated METTL1 levels promote hepatoma cell proliferation and migration in vitro [6]. Xu H et al. [7] discovered that the augmentation of small ubiquitin-like modifier (SUMOylation) of METTL3 in response to mitogen-stimulation was directly correlated with the elevated metastatic capacity of HCC. The progression of HCC is regulated by the process of SUMOylation in METTL3 which depends on the activity of m6A methyltransferase. This regulation is achieved by modulating the balance of Snail mRNA [7]. Currently, there is little research concerning the function of METTL17 in human cancers.

This study aimed to examine the role of METTL17 in pan-cancers and establish a connection between the expression of METTL17 and the prognosis of different malignancies. In addition, we analyzed the genetic changes associated with METTL17, performed a genomic heterogeneity study, assessed the tumor stemness index, examined immune cell infiltration, constructed a protein interaction network, and identified critical signaling pathways using various algorithms. Furthermore, a comprehensive analysis was conducted to examine the involvement of METTL17 in the progression of HCC. Ultimately, our research established a risk model related to METTL17 and utilized it to conduct a stratification analysis on the prognosis of HCC.

Methods

Single-cell data acquisition and bulk expression in tissues

Clinical data of patients with different tumor types were obtained from the Cancer Genome Atlas (TCGA) (https://xenabrowser.net/datapages/). The gene expression data were downloaded from the Genotype-Tissue Expression (GTEx) (https://www.gtexportal.org/home/) and TCGA databases. These data were combined using the Perl package in R to generate a matrix file for comparing METTL17 expression in normal and tumor samples. The ConsensusPathDB (https://cpdb.molgen.mpg.de/), Human Protein Atlas (HPA) (https://www.proteinatlas.org/), and GTEx datasets were utilized to examine the METTL17 RNA expression and the level of METTL17 protein expression was analyzed using the FANTOM5 dataset [8].

Single-cell expression analysis

We measured single-cell RNA-seq gene expression(scRNAseq) of 25 human tissues including analysis of all protein-coding genes in 444 different cell clusters across 15 separate cell groups (https://www.proteinatlas.org/single+cell+type) [9].

The gene selection process was based on the following criteria: (1) Cells were not previously sorted or enriched for any specific cell types. (2) The number of read counts obtained from the sequencing data was greater than 20 million. (3) The total number of sequenced cells equals to 4,000. (4) The association between the patterns of pseudo-expression and real-expression was statistically significant. The gene expression in each of the cell types was displayed using UMAP plots and bar charts. The level of confidence was determined by calculating the fraction of times a gene was allocated to the cluster during repeated clustering. This measure accurately indicated the strength of the gene’s association with the cluster. A confidence level of 1 was used to imply the gene was consistently assigned to a specific cluster across all repeated clusters.

Immune cell type specificity analysis

The specificity analysis of immune cell type was conducted utilizing the immune-cell part of the HPA (https://www.proteinatlas.org/immune+cell). The immune cell lineages analyzed in this study encompassed T-cells, granulocytes, NK-cells, dendritic cells, B-cells, monocytes, progenitors, and total peripheral-blood-mononuclear-cell (PBMC). The expression of METTL17 in several blood cell lineages was investigated using the HPA and Monaco datasets.

Subcellular localization of METTL17

Immunofluorescence microscopy analysis of METTL17 revealed the specific subcellular localization (https://www.proteinatlas.org/subcellular).

Pan-cancer data acquisition and METTL17’s prognosis analysis

We extracted the expression data of the ENSG00000165792 (METTL17) gene in each sample from the University of California Santa Cruz (UCSC) (https://xenabrowser.net/) database. This data was downloaded from the uniform standardized pan-cancer dataset: the Cancer Genome Atlas (TCGA) (https://xenabrowser.net/datapages/), Therapeutically Applicable Research To Generate Effective Treatments (TARGET) (https://ocg.cancer.gov/programs/target), and Genotype-Tissue Expression (GTEx) (https://www.gtexportal.org/home/). The dataset consists of 19,131 samples and 60,499 genes. Tumor types are displayed in Table 1.

Table 1 Tumor abbreviation

Then, we analyzed METTL17 expression levels in both tumor and non-tumor groups across several types of cancer using the “limma” package in R software (version 4.0.3). MaxStat in R software was used to calculate the best cut-off value of METTL17 expression by setting 25–75% as the grouping number range. After dividing each cancer type into high and low METTL17 expression groups, we used Kaplan-Meier survival analysis to predict the differences in overall survival (OS), disease-specific survival (DSS), and disease-free interval (DFI) between the two groups. Statistical analysis was conducted using R software and ggplot2 (version 3.3.2). P < 0.05 was considered statistically significant.

Tumor stemness index analysis

Our group analyzed the METTL17 expression data downloaded from the UCSC database. We specifically focused on samples from Primary Blood Derived Cancer (Peripheral Blood, Primary Tumor). Additionally, we calculated the tumor stemness score for each tumor using methylation profiling data from previous studies [9]. Then, we used the Spearman correlation test to conduct the correlation analysis between the expression of METTL17 and DNA stemness score (DNAss), RNA stemness score (RNAss), differentially methylated probes-based (DMPss), enhancer elements/DNA methylation-based (ENHss), DNA methylation-based (EREG-METHss), RNA expression-based (EREG.EXPss) for 37 different cancer types.

Analysis of immune cell infiltration

The deconvo_CIBERSOR method [10], TIMER method [11] from the R package IOBR (version 0.99.9) [12] was used to quantitatively evaluate tumor immune cell infiltration based on 44 different types of tumor data from the pan-cancer cohort. Correlations between METTL17 expression and infiltrating cells of interest were investigated using Spearman’s rank correlation analysis. The immune, stromal, and ESTIMATE scores of METTL17 expression were evaluated using the ESTIMATE algorithm [13].

METTL17-related protein-protein interaction (PPI) network and key signaling pathway predictions

We performed the analysis of PPI networks using the STRING website (https://string-db.org/). The criteria used to identify the interacting proteins of METTL17 were defined as follows: minimum required interaction score [“Medium confidence (0.400)”], the meaning of network edges (“evidence”), and active interaction sources (“experiments”).

Based on the clinical expression of the METTL17 data, the gene set variation analysis (GSVA) score was compiled and we explored associations between pathway activity and expression score of METTL17 and then compared the GSVA score of METTL17 genes in different cancers. The expression and pathway activity modules calculate the variation in gene expression among distinct groups of pathway activity. These groups are determined by the median pathway scores (http://bioinfo.life.hust.edu.cn/GSCA/). Table 2 presents essential information pertaining to the signalling pathway.

Table 2 Key signaling pathway-related abbreviations

Immunohistochemistry (IHC)

IHC was conducted on 30 pairs of HCC tumor tissues and normal liver tissues. The tissues were treated with 1% formaldehyde at room temperature for fixation. Subsequently, the tissue slices were exposed to METTL17 antibody (ThermoFisher, PA5-26973, 1:200) and incubated overnight at 4 °C. Next, the slides were exposed to the secondary antibody, and photographs were captured. The intensity of IHC staining was calculated from the intensity and number of stained cellular sections. The evaluation criteria for staining intensity were as follows: 0, 1, 2, and 3 represented negative, weak, moderate, and strong staining, respectively. The evaluation criteria for the number of stained cells were 0, 1, 2, 3, and 4, representing the percentages of stained cells as < 10%, 10 ∼ 25%, 25 ∼ 50%, 50 ∼ 75%, and > 75%, respectively. IHC score = staining intensity × staining number. A score ≥ 6 indicates a high expression; and a score < 6 represents a low expression [14].

Gene set enrichment analysis (GSEA)

GSEA was performed to explore the possible mechanisms or biological processes involving METTL17. This analysis was based on the Kyoto Encyclopedia of Genes and Genomes (KEGG), Gene Ontology (GO) and HALLMARK gene sets from the molecular signature database (https://www.gsea-msigdb.org/gsea/msigdb) as reference sources. The R package “clusterProfler” is utilized to perform GSEA using data obtained from differential expression analysis. For each cancer type, false discovery rate (FDR) and normalized enrichment score (NES) values are calculated for every hallmark [15]. The “ggplot2” R tool was used to show the METTL17 enrichment data in different pathways among HCC.

The construction, validation, and clinical use of METTL17-related prognostic model

Univariate Cox regression analysis assessed the relationship between risk factors and prognosis. Subsequently, a multivariate Cox regression analysis was conducted to determine if the risk score of METTL17 and clinical features served as independent predictors of OS. A nomogram was constructed utilizing the risk score and additional clinical indications to predict the 1-year, 3-year, and 5-year OS. Then, we employed data calibration curves to assess the prediction accuracy of the nomogram and to compare the predicted probability of OS with the observed probability of OS for consistency. Consequently, we compared the nomogram with all the independent prognostic factors to those with only one independent prognostic factor using decision curve analysis (DCA) [16]. DCA was used to calculate the clinical net benefit of each model compared with all or none of the strategies. The best model is the one with the highest calculated net benefit. We utilized the R package-ggDCA to evaluate the model associated with METTL17.

Statistical analysis

The study was conducted using the R language program (version 4.0.3) (https://www.r-project.org/). The default Wilcoxon’s test was used to assess the differences between the two groups, while one-way analysis of variance (ANOVA) was used to analyze the differences across several groups. The Kaplan-Meier method was used to estimate survival rates, and the differences were assessed using the log-rank test. The ggplot2 program in R was implemented to calculate all correlation coefficients and visualize the results. Results were considered statistically significant when the P < 0.05.

Results

Bulk and Single-cell expression of METTL17

To investigate the basic expression levels of METTL17 in human normal tissues, we employed the TCGA and the GTEx database. Our findings indicate that METTL17 was most highly expressed in the endocrine tissues, proximal digestive tract, and bone marrow & lymphoid tissues (Fig. 1A). The RNA expression of METTL17 was confirmed in the ConsensusPathDB, HPA, and GTEx datasets. These data revealed that METTL17 expression was highest in the adrenal gland, tongue, and bone marrow (Fig. 1B). The FANTOM5 dataset was used to analyze the expression of METTL17 protein in various human tissues and organs. The results showed that the salivary gland, seminal vesicle, and ductus deferens had the highest level of METTL17 expression among all the tissues and organs examined (Fig. 1C).

Fig. 1

METTL17 expression in normal tissues. (A) METTL17 RNA expression in healthy tissues or organs from the HPA database. (Different colors represent different tissues and organs.) (B) RNA expression of METTL17 in healthy tissues or organs from the ConsensusPathDB database. (C) METTL17 protein expression in healthy tissues or organs from the FANTOM5 dataset. (D) Single-cell analysis of METTL17 expression in the bone marrow. (E) Single-cell analysis of METTL17 expression in the spleen. (F) Single-cell analysis of METTL17 expression in the lymph node. (G) Single-cell analysis of METTL17 expression in the adipose tissue. (H) METTL17 expression in all cells at the single-cell level

Next, we evaluated the expression of METTL17 in hematopoietic cells, scRNAseq data from hematopoietic tissues were analyzed. The results indicated that T cells and plasma cells expressed higher levels of METTL17 compared to other immune cells in the bone marrow and spleen, respectively (Fig. 1D-E). METTL17 had elevated expression in B cells within lymph nodes, as depicted in Fig. 1F. METTL17 exhibited elevated expression in T cells within adipose tissues, as indicated by Fig. 1G. Next, RNA Single-cell type specifications analysis revealed that METTL17 expression was increased in early spermatids, rod photoreceptor cells, and basal squamous epithelial cells (Fig. 1H).

The 49 gene clusters obtained from the Louvain clustering of gene expression across all single-cell types were shown using UMAP (Fig. 2A). Figure 2B displayed the top 15 nearest neighbors based on single-cell type RNA expression. The specificities of immune cell types in hematopoietic tissues were confirmed using the HPA and Monaco datasets. According to the HPA dataset, basophils, myeloid dendritic cells (DC) and memory B-cells exhibited significantly high levels of METTL17 expression, as shown in Fig. 2C. The findings of the analysis of Monaco database analysis were consistent with those of the HPA dataset. Plasmablast, plasmacytoid DC, and terminal effector memory CD4⁺ T cells exhibited the highest level of METTL17 expression among the hematopoietic immune cells (Fig. 2D). Based on data downloaded from Immunofluorescence microscopy, we discovered that METTL17 localized in the nucleoplasm (Fig. 2E).

Fig. 2

METTL17 expression in tumors based on Single-cells dataset. (A) UMAP analysis of METTL17 expression using single-cell RNA sequencing data from the HPA database. (B) The top 15 nearest neighbors are based on single-cell type RNA expression. (C) METTL17 expression in immune cells from the HPA database. (D) METTL17 expression in immune cells from the Monaco database. (E) Subcellular localization of METTL17 by immunofluorescence microscopy

METTL17 expression and its prognostic significance in pan-cancer

We analyzed the mRNA expression of METTL17 expression in tumors and corresponding adjacent tissues to explore the significance of METTL17 in cancers (tumor types are displayed in Table 1). The results from the UCSC database showed that GBM, GBMLGG, LGG, UCEC, KIRP, KIPAN, HNSC, WT, ALL, LAML, KICH, and CHOL expressed higher levels of METTL17 expression. In contrast, BRCA, CESC, LUAD, ESCA, STES, COAD, COADREAD, PRAD, STAD, LUSC, LIHC, SKCM, THCA, OV, TGCT, UCS, PCPG, and ACC showed lower expression of METTL17 (Fig. 3A) (Table S1). Figure 3B demonstrated significant gender disparities in METTL17 levels across ARC, KIRP, HNSC, KIRC, and LIHC. Our analysis revealed a positive correlation between age and METTL17 expression in LAML and THCA, but a negative correlation was found in COAD and COADREAD (Fig. 3C).

Fig. 3

Different expression and prognostic value of METTL17 in pan-cancer. (A) METTL17 expression in different cancer and paired normal tissues. (B) Correlation of METTL17 expression with gender. (C) Correlation of METTL17 expression with patient age. (D) The expression of METTL17 using the immunohistochemical images abtained from the HPA database. (E) Kaplan-Meier curves for OS

In addition, we evaluated the protein expression of METTL17 by utilizing immunohistochemical images obtained from the HPA database. Figure 3D provided a visual representation of the protein expression of METTL17 in COAD, PRAD, BRCA, and LUAD, by IHC images. This demonstrates both weak and moderate staining intensities of diaminobenzidine staining within the target cell region.

We obtained and examined curated survival data from the UCSC database to delve into the predictive significance of METTL17 mRNA levels in human malignancies. The OS analysis revealed that METTL17 was a protective factor for patients with THYM, but it was a risk factor for patients with LAML, KIRC, LIHC, and UVM (Fig. 3E) (Table S2). The findings of the DSS analyses showed that METTL17 had a protective effect in patients with KIRP, but had a detrimental effect in patients with LIHC, THCA, UVM, and PCPG (Fig. 4A-D) (Table S3). The DFI analysis found that METTL17 functioned as a risk factor for patients with CESC, PRAD, and LIHC (Fig. 4E) (Table S4).

Fig. 4

Prognostic analysis and immune cell infiltration of METTL17 in pan-cancer. (A-D) Kaplan-Meier curves for DSS. (E) Kaplan-Meier curves for DFI. (F) The correlation between METTL17 expression and infiltration level of immune cells using CIBORSORT score

Based on tumor node metastasis (TNM) classifications, we observed notable variations in METTL17 expression among different T grades in patients with PRAD, LUAD, and KIRP (P < 0.01) (Figure S1A). There were significant variations in the levels of METTL17 among patients with LUAD, KIRP, PRAD, KIRC, SKCM, BLCA, and ACC, depending on their N grades (P < 0.01) (Figure S1B). The expression of METTL17 was significantly higher in M1 compared to M0 in MESO (P < 0.01) (Figure S1C).

Tumor stemness index analysis

The tumor stemness index was employed to examine the resemblance between tumor cells and stem cells, primarily by assessing DNAss, RNAss, and other related factors. The analysis found a substantial association between METTL17 levels and DMPss, ENHss (P < 0.05) (Figure S4C-D). Figures S4A-B and S4E-F demonstrated a strong positive correlation between the expressions of METTL17 and RNAss, and a negative correlation between METTL17 expression and the DNAss, EREG-METHss, EREG.EXPss in different types of malignancies (P < 0.05).

Analysis of immune cell infiltration

To evaluate the link between METTL17 expression and immune cell infiltration, we utilized two distinct sources of immune cell infiltration data for the correlation study. We performed correlation analysis using the CIBERSORT algorithm and evaluated 26 immune cells in 44 types of cancer. These data indicated that the expression of METTL17 was mainly significantly associated with follicular helper T cells, Tregs, and CD4⁺ T cells memory activated in the majority of tumors (43 cancer cases) (P < 0.05) (Fig. 4F). The TIMER database analysis showed an intense relationship between the expression of METTL17 and CD4⁺ T cells, DC, and B cells in most tumors (38 cancer cases) (P < 0.05) (Figure S5).

METTL17-related key signaling pathway predictions

We identified the mechanisms and relevant pathways of possible members of the METTL family by calculating the enrichment scores of canonical pathways, as shown in Fig. 5A; Table 2. The findings demonstrated that METTL17 has the ability to stimulate several cellular pathways including the cell cycle, androgen signaling, DNA damage, apoptosis, PI3K/AKT, and TSC/mTOR pathways. Conversely, METTL17 exhibited a negative association with the EMT, RAS/MAPK, RTK, and estrogen signaling pathways.

Protein-protein interaction network of METTL17

We conducted a mapping of the PPI networks involving METTL17 and resented the visualization of the interaction between METTL 17 and its related molecules using the STRING (Fig. 5B) (Table 2). METTL17 exhibited a significant association with MRPS5, DAP3, LRPPRC, LIPT2, NCBP1, MRPS10, N6AMT1, KIAA0020, CHCHD1, and BMS1.

Fig. 5

Key signaling pathway predictions, protein-protein interaction network and immunological values of METTL17 . (A) The correlation between METTL17 expression and related key pathways. (B) METTL17-related PPI network. (C) METTL17 expression in HCC and paired normal tumor tissues using immunohistochemical techniques. (D) Differences in immune cell scores in different tissues in HCC. (E) Proportional graph of 22 immune cells in HCC

Immunological values and functional analysis of METTL17 in hepatocellular carcinoma

The pan-cancer prognostic analysis demonstrated a substantial correlation between METTL17 and the prognosis of numerous kinds of cancer. More precisely, there was a significant association between increased expression of METTL17 and reduced OS, DSS, and DFI in HCC as shown in Figs. 3E， 4Aand 4E. Furthermore, the examination of IHC confirmed that the decrease in METTL17 expression was significantly correlated with the progression of HCC (P < 0.05) (Fig. 5C).

The CIBERSORT scores for various pathologic grades of HCC showed significant differences in the following immune cell populations: B cell naïve, B cell memory, B cell plasma, CD8⁺ T cell, CD4⁺ T cell memory resting & activated, T cell follicular helper, T cell regulatory (Tregs), T cell gamma delta, and NK cell resting (P < 0.01) (Fig. 5D). In addition, we performed a visual examination of the ratios of 22 different types of immune cells using the CIBERSORT algorithm (Fig. 5E). The METTL17 expression showed a statistically significant association with the abundance of infiltrating CD4⁺ T cells, neutrophils, macrophages, and myeloid dendritic cells, as determined by the TIMER algorithm (P < 0.001) (Fig. 6A).

Fig. 6

Relationship between METTL17 and immune cells and signaling pathways of hepatocellular carcinoma. (A) Correlation scatterplot of immune cells associated with METTL17 expression. (B-C) GSEA analysis of METTL17-related pathway based on the KEGG database. (D-E) GSEA analysis of METTL17-related pathway based on the GO database. (F-G) GSEA analysis of METTL17-related pathway based on the HALLMARK database

The KEGG pathway enrichment analysis revealed that dysregulation of METTL17 was predominantly enriched in the primary bile acid biosynthesis and cell cycle pathway (Fig. 6B-C). The GO pathway enrichment analysis discovered that METTL17 played a crucial role in maintaining a robust association between nucleosome binding and the tRNA binding pathway (Fig. 6D-E). Based on the HALLMARK pathway enrichment analysis, it was shown that METTL17 presented an intense association with E2F targets and the mitotic spindle signal pathway (Fig. 6F-G). Afterwards, we obtained the top 7 genes (MRPS5, LRPPRC, LIPT2, NCBP1, N6AMT1, KIAA0020, and BMS1) that linked to METTL17 expression in HCC data. These genes were then showed in the scatter plots in Fig. 7A.

Fig. 7

Construction and validation of the METTL17-related predictive model. (A) The top 7 genes associated with METTL17 expression. (B-C) The result of the LASSO regression. (D) The heatmap of METTL17-related genes in the high-risk and low-risk groups. (E) The distribution of risk scores. (F) The distribution of survival statuses. (G) Kaplan–Meier curve of patients. (H) The ROC curve of the METTL17-related genes model in predicting 1-, 3-, and 5-year OS

Construction, validation, and clinical use of METTL17-related prognostic model in hepatocellular carcinoma

A univariate Cox regression analysis was employed to identify 7 METTL17-related genes associated with patient prognosis. Based on these 7 genes, a least absolute shrinkage and selection operator (LASSO) regression and stepwise Cox regression analysis were conducted to construct a METTL17-based prognostic model. Finally, 6 key genes (MRPS5, CHCHD2, NCBP1, LRPPRC, DAP3, and BMS1) were selected to construct models (Fig. 7B-C). The risk score is calculated by multiplying the respective coefficients with the corresponding variables and summing them up. The coefficients are as follows: risk score = (0.1746*MRPS5) + (0.1534*DAP3) + (0.0656*LRPPRC) + (0.1313*NCBP1) + (0.0982*CHCHD1) + (0.2508*BMS1). The risk score for each HCC patient was determined based on the expression levels of above 6 genes. Subsequently, the patients were categorized into high-risk and low-risk groups, depending on the median risk score. Figure 7D utilized a heat map to visually represent the gene expression profiles of both the high-risk group and the low-risk group. The risk score distribution of the HCC patients was indicated in Fig. 7E, which showed a progressive increase in the curve from left to right, thus dividing the patients into two distinct groups. Figure 7F displayed the distribution of survival status and survival time across patients with varying risk ratings. The OS of the high-risk group was shown to be significantly lower than that of the low-risk group (P < 0.05) (Fig. 7G). The area under the curve (AUC) values for the METTL17-based model were 0.747 at 1-year, 0.671 at 3-years and 0.631 at 5-years, respectively (Fig. 7H). In addition, the results from DCA demonstrated that the predictive nomogram had superior net benefits for forecasting tumor progression in HCC patients (Figure S10A-B).

Discussion

METTL family proteins are linked to tumor progression, metastasis, and prognosis suggesting that members of the METTL family have the potential to serve as valuable indicators and targets for therapeutic intervention against tumors [17,18,19,20,21]. The investigation of METTL family members provides novel avenues for biomarker discovery [2]. Hence, it is imperative to examine the function of METTL17 in human tumors. Our study conducted a comprehensive analysis of the expression levels of METTL17 in pan-cancers based on several different databases. These results revealed a significant increase of METTL17 expression in 22 tumor tissues compared to their corresponding normal tissues. Furthermore, we found a correlation between the expression level of METTL17 and the prognosis of multiple types of cancer, as well as a significant association between aberrant METTL17 expression and tumor immune cell infiltration. We also observed the distinct expression of METTL17 in HCC tissues using IHC experiments and developed a predictive model based on METTL17. Thus, our research team discovered for the first time that METTL17 played a tumor-promoting role in the HCC.

The current investigation involved the examination of METTL17 expression in various tissues, with a particular focus on hematopoietic tissues, through the analysis of single-cell sequencing data. Our study revealed that METTL17 exhibited predominantly high levels of expression in early spermatids, rod photoreceptor cells, and basal squamous epithelial cells. These results further confirm the important immune function of METTL17 in the regulation of the immune system.

Multiple studies [18,19,20] have documented varying levels of abnormal expression of the METTL family, including METTL16, METTL7B, and METTL18 in ovarian epithelial carcinoma, non-small cell lung cancer, and HCC for members, respectively. These findings indicated that METTL family members might serve as valuable indicators of tumors. In this study, the expression of METTL17 was shown to be upregulated in 12 tumors and downregulated in 18 tumors. We also revealed the significant differences in the expression of METTL17 among patients with different gender, age, tumor stages, and types. Additonally, we discovered a positive correlation between age and METTL17 expression in LAML and THCA, and a negative association in COAD and COADREAD. These results could serve as valuable references for immunological and targeted therapy in tumor patients across various age groups and genders. Furthermore, the OS analysis demonstrated that METTL17 exhibited a protective effect in patients with THYM, but it posed a risk for patients with HCC based on the results of DSS and DFI studies. These findings led us to propose that METTL17 could serve as a biomarker for assessing the prognosis of patients with these tumors.

The METTL17 mutation was involved in the post-translational modification of mitochondrial proteins, emphasizing the critical function of METTL17 in maintaining human health [4]. The development of malignancy is also strongly linked to CNVs [22] and SNVs [23]. Previous studies have reported that CNV data could be utilized to categorize tumors as either malignant or benign [22, 24, 25]. Our group examined the CNV data to calculate the frequency and patterns of mutations. It revealed a close correlation between CNV, SNV, and aberrant expression of METTL17 in the majority of malignancies. In addition, the genetic mutation patterns of METTL17 are primarily characterized by missense mutations. Hence, we speculate that regular monitoring of METTL17 expression, along with the identification of CNV or SNV in precancerous lesions could potentially predict the occurrence of METTL17-related tumors.

Tumor heterogeneity is a universal and essential feature in the formation of tumors and has a strong connection to the clinical diagnosis and treatment of patients [26]. Our research revealed a close relationship between METTL17 expression and genomic heterogeneity-related factors, including TMB, MATH, MSI, NEO, purity, ploidy, HRD, and LOH in tumors. High TMB (defined as ≥ 10 mutations/Mb) is a useful predictor for the response to immune checkpoint inhibitors (ICI) treatment in solid tumors [27]. Furthermore, a high TMB is indicative of more efficient suppression of cervical squamous cell carcinoma by ICI therapy [28]. According to previous reports [29], patients with high levels of MSI exhibited stronger anti-tumor immune responses and had a better prognosis compared to patients with low MSI levels or stable microsatellites. Taken together, we hypothesize that tumor patients who exhibit abnormal expression of METTL17, combined with high levels of TMB and / or MSI, etc., may experience improved outcomes following ICI therapy. Nevertheless, further experimental research is needed to prove this hypothesis.

The tumor microenvironment (TME) comprises immune cells and stromal components that play a critical role in malignancy initiation, progression, and chemoresistance. In addition, the tumor stemness index is correlated with the purity and patient prognosis [30, 31]. Our work showed a substantial correlation between the expressions of METTL17 and the levels of DNAss, EREG-METHss, RNAss, DMPss, ENHss, and EREG.EXPss in different types of malignancies. Therefore, METTL17 might represent an immunotherapeutic target for augmenting immune responses in various human cancers. In terms of TME patterns, advances in the recognition of interactions between METTL17 and tumor stemness index could provide novel insights for the understanding of METTL17-mediated immunomodulation and METTL17-based stem cell immunotherapy.

The characteristics of TME, specifically the presence of tumor-infiltrating immune cells, can serve as a biomarker for evaluating the effectiveness of immunotherapy on tumor cells [32]. Furthermore, tumor-infiltrating immune cells play critical roles in the initiation and progression of cancers and are linked to better prognosis [33, 34]. Our research demonstrated a close association between the expression of METTL17 and tumor immune cell infiltration in nearly all types of tumor cases, including the infiltration of different lymphocytes. It is reported [35] that the particular advantages of tumor-infiltrating lymphocyte (TIL) treatment in treating solid tumors, owing to its diversified T-cell antigen receptor clonality, greater tumor-homing capacity, and less off-target damage. Given the close relationships of METTL17 with TME, this makes METTL17 an effective immunomodulatory variable.

The investigation of the PPI networks of METTL17 discovered a substantial association with MRPS5, DAP3, LRPPRC, LIPT2, NCBP1, MRPS10, N6AMT1, KIAA0020, CHCHD1, and BMS1. In addition, we have identified significant processes and pathways indicate that the involvement of METTL17 in the tumorigenesis of human malignancies. This is achieved by influencing the cell cycle and metabolism, making METTL17 a prospective target for therapeutic intervention in tumor therapy. Then, a robust correlation between METTL17 and its associated genes/pathways has been identified, establishing a solid foundation for the potential integration of molecular targeted immunotherapy in the future. In summary, these findings suggest that METTL17 has the potential to be used as a target in immunotherapy of cancer.

HCC remains one of the most fatal malignant tumors globally, with highly intricate underlying molecular mechanisms. Hence, it is crucial to discover biomarkers to predict the prognosis and formulate personalized treatment strategies for HCC patients. Recent developments in gene sequencing technology and bioinformatics have enabled the identification of predictive models [36]. Our study has shown the immunological value, related signaling pathways, and protein interaction targets of METTL17 in pan-cancer and we further explored the impacts of METTL17 on HCC progression.

Distinct molecular mechanisms involved in hepatocarcinogenesis elicit various particular populations of immune cells, leading to either inflammation or the suppression of anti-tumor immune responses [37]. By examining the transcriptome and clinical data of HCC patients in the dataset, we found that the elevated METTL17 expression was associated with poorer prognosis of HCC patients. Moreover, the expression of METTL17 showed a positive association with the number of infiltrating CD4⁺ T cells, neutrophils, macrophages, and myeloid dendritic cells in HCC. In the context of cellular immunity, B cells activate NK cells to directly eliminate tumor cells. Furthermore, B cells release pro-inflammatory substances, which in turn stimulate T cells [38]. B cells produce antibodies that specifically target tumor cells and contribute to humoral immunity in HCC. The number of NK cells is decreased in peripheral blood and HCC tumors, together with reduced levels of interferon-gamma (IFN-γ), indicating a compromised function of NK cells in the tumor microenvironment [34]. Chew V et al. [39] found a significant correlation between the frequency of peripheral and intrahepatic NK cells and the survival time of HCC patients [39], suggesting that these immune cells may play an essential function in monitoring and detecting tumors. Moreover, activation of immune cells such as CD4⁺ T cells, CD8⁺ T cells, NK cells, and monocytes mediates hepatic inflammation and ultimately leads to the progression of HCC [40]. Thus, considering the strong correlation between METTL17 expression with the TME, we propose that METTL17 could serve as a prognostic indicator and a promising target for immunotherapy in HCC.

Subsequently, we conducted GESA analysis by KEGG, GO, HALLMARK data. Our findings revealed that dysregulation of METTL17 was predominantly enriched in the primary bile acid biosynthesis, nucleosome binding pathway, and E2F targets signal pathway, respectively. Similar to our results, it has been observed that bile acids, specifically ursodeoxycholic acid, and diethylstilbestrol have the ability to slow down the progression of liver injury and carcinogenesis [41]. The vital roles of nucleosome binding and the E2F targets pathway in malignancies have been firmly established [42, 43]. These findings indicate that METTL17 may influence the onset and development of HCC through the signaling pathways outlined above. Nevertheless, further research is required to explore the underlying mechanisms by which METTL17 is involved in the above pathways.

The prognostic model proposed in the present study was composed of 6 METTL17-related genes that included BMS1, CHCHD2, NCBP1, LRPPRC, DAP3, and MRPS5. The 1-year, 3-year, and 5-year AUCs of the METTL17-related genes model were 0.747, 0.671 and 0.631, respectively. In addition, we proved this model’s optimal utility in forecasting HCC progression by DCA analysis. All these results indicated that this model was effective in predicting the prognosis of HCC patients. According to Yang et al., [44], Cuproptosis-Associated Long non-coding RNA can be a prognostic predictor in HCC. The ROC curve demonstrated an AUC of 0.723 at 1 year, 0.711 at 2 years, and 0.700 at 3 years, indicating a strong predictive ability of the signature. A recent study was done to investigate newly found genes associated with cuproptosis and their potential impact on the OS of HCC patients. The study found that the 1-year AUC was 0.683, 0.652, and 0.614 in the training set, validation set, and external validation set, respectively, at 1 year [45]. Compared with these published prognostic models, our model showed better performance in predicting the prognosis of HCC. In future research, we will perform external validation of our model to further prove its clinical significance.

The research had some limitations. First, functional experiments, such as CRISPR/Cas9 knockout or overexpression of METTL17 are needed to verify the critical role of METTL17 in various types of cancers, especially in HCC. Second, the mechanisms underlying the impacts of METTL17 in the modulation of immunological systems remain unclear. Furthermore, our population-level investigation utilized a pan-cancer database, which did not consider individual variances that could influence clinical outcomes and potentially introduce analytical bias. Finally, a METTL17-related prognostic model for HCC was established in our study; however, external validation of this model is necessary to further ascertain its clinical significance.

Conclusions

This study examined the aberrant expression of METTL17 across diverse cancers and its strong association with the patient’s prognosis. Furthermore, the expression of METTL17 was tightly linked to the tumor immune microenvironment, tumor stemness index, and various other characteristics, suggesting the importance of METTL17 in tumor immune regulation. Finally, we revealed that METTL17-related prognostic models could be utilized to evaluate outcomes in HCC patients.