A Novel Mitochondria-Related Gene Signature in Esophageal Carcinoma: Prognostic, Immune, and Therapeutic Features

doi:10.21203/rs.3.rs-2456179/v1

Download PDF

Research Article

A Novel Mitochondria-Related Gene Signature in Esophageal Carcinoma: Prognostic, Immune, and Therapeutic Features

https://doi.org/10.21203/rs.3.rs-2456179/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 29 Mar, 2023

Read the published version in Functional & Integrative Genomics →

You are reading this latest preprint version

Esophageal Carcinoma (ESCA) is a common and lethal malignant tumor worldwide. A role for mitochondria in tumorigenesis and progression has been proposed. The mitochondrial biomarkers were useful in finding significant prognostic gene modules associated with ESCA. In the present work, we obtained the transcriptome expression profiles and corresponding clinical information of ESCA from The Cancer Genome Atlas (TCGA). Differential expressed genes (DEGs) were overlapped with mitochondria related genes to obtain mitochondria related DEGs. The univariate cox regression, Least Absolute Shrinkage and Selection Operator (LASSO) regression, and multivariate cox regression was sequentially used to define the risk scoring model for mitochondria-related DEGs, and its prognostic value was verified in the external datasets GSE53624. Based on risk score, ESCA patients were divided into high and low risk groups. GO, KEGG and Gene Set Enrichment Analysis (GSEA) were performed to further investigate the difference between low and high risk groups in the gene pathway level. CIBERSORT was used to evaluate immune cell infiltration. The mutation difference between high and low risk groups was compared by the R package “Maftools”. Cellminer was used to assess the interactions of the risk scoring model and drug sensitivity. As the most important outcome of the study, we obtained 306 mitochondria related DEGs, and constructed a 6-gene risk scoring model (APOOL, HIGD1A, MAOB, BCAP31, SLC44A2 and CHPT1). Between high and low risk group, pathways including “hippo signaling pathway” and “cell-cell junction” was enriched. According to CIBERSORT, samples with high risk demonstrated higher abundance of CD4⁺ T cells, NK cells, M0 and M2 Macrophages, and lower abundance of M1 Macrophages. The immune cell marker genes were correlated with risk score. In mutation analysis, the mutation rate of TP53 was significantly different between the high and low risk groups. Drugs with strong correlation with model genes and risk score were selected. In conclusion, we focused on the role of mitochondria-related genes in cancer development, and proposed a prognostic signature for individualized integrative assessment.

Esophageal Carcinoma

TCGA

Prognosis

Tumor immunity

Esophageal Carcinoma (ESCA) is a common malignant tumor of digestive tract, with the seven highest incidence and the six mortality worldwide^[1]. Only about one-quarter of patients are candidates for attempted curative treatment, such as surgery, chemotherapy, radiation therapy, molecular targeting therapy, and combinations of these therapies, and the prognosis still remains poor, with a 5-year survival rate of about 30 percent^[2]. There are several factors that contribute to the development of ESCA, including environmental factors, living habits, microorganisms, genetic factors and differential gene expression^[3]. Since there is no sensitive method for detecting ESCA early, about half of the patients diagnosed with the tumors already have distal metastases^[4]. The Cancer Genome Atlas Project (TCGA) profiles at least 20 different types of cancer and makes raw and processed data available to all researchers using genomic platforms^[5]. The TCGA released a large number of mRNA sequencing data for ESCA patients, which will be crucial to uncovering more ESCA-associated genes.

The mitochondria produce adenosine triphosphate (ATP) for cells. Recent studies have shown that mitochondria regulate various physiological and pathological functions, such as the initiation, progression and metastasis of tumor cells^[6–7]. The mitochondria produce reactive oxygen species (ROS) in the cell. As evidenced by accumulating studies, dysregulation of ROS contributes to tumor malignancy^[8]. Moreover, as a key regulator of apoptosis, the mitochondrion can enhance the antiapoptotic ability of cancer cells^[8]. Apoptosis of mitochondria has been implicated in the action of triterpenoid saponins in other types of human cancer, including liver cancer, gastric cancer, ESCA, and colorectal cancer^[9]. Among the most common mechanisms of cancer cell death, the mitochondrial pathway has been widely investigated as a potential cancer treatment target^[10]. For instance, dioscin induces death through the mitochondrial pathway in ESCA cells^[11]. By causing mitochondrial dysfunction, propafenone suppresses ESCA proliferation^[12]. In cervical cancer cells, mitochondria were implicated in autophagic cell death induced by ionizing radiation (IR), but their role in ESCA is unknown^[13].

Considering mitochondria are crucial mediators of tumorigenesis, it remains to be identified how mitochondria affect the prognosis of patients. Besides, as cancer immunotherapy has advanced rapidly in recent years, it is now being explored as a treatment for ESCA^[14]. However, mitochondria and immune are rarely researched together in tumors. In this study, we explored mitochondria - related biomarkers and constructed Cox proportional - hazards model to predict patients’ prognosis of ESCA. Combining these biomarkers with immune infiltration analysis provided a more comprehensive view of the risk factors for ESCA as well as a prediction of patient prognosis. Infiltrating immune cells in cancer tumors have been shown to influence carcinogenesis and progression, and may also predict patient response to neoadjuvant chemotherapy^[15]. In summary, our findings highlight mitochondria-related ESCA signatures, explore their correlation with immunity to predict prognosis and provides new ideas for the diagnosis and treatment of ESCA.

1. Datasets and source

Gene expression RNAseq (HTseq) was downloaded from The Cancer Genome Atlas database (TCGA: https://portal.gdc.cancer.gov/). The University of California SANTA CRUZ (UCSC: https://xenabrowser.net/datapages/) provided phenotype and survival data. GEO datasets are all downloadable from the National Center for Biotechnology Information (NCBI) online Gene Expression Omnibus (GEO) database, which is publicly available at https://www.ncbi.nlm.nih.gov/geo.

Human mitochondrial related genes were retrieved from Gene Set Enrichment Analysis (http://www.gsea-msigdb.org/gsea/msigdb/index.jsp, Table S1). Figure 1 shows the workflow of this study.

2. Differential Expression Analysis

Based on the expression of genes in ESCA, differential expression analysis to identify DEGs was performed by Limma package of R (version 3.5.1)^[16]. “Adjusted P < 0.01 and |log (Fold Change)| > 0.5” was defined as the thresholds for screening DEGs.

3. Enrichment Analysis

Pathway enrichment analysis was conducted using Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Set Enrichment Analysis (GSEA). GO and KEGG enrichment were performed with “clusterProfiler” R package. All pathways that are available in the C2 curated gene set from GSEA were included in the pathway analysis. P-values less than 0.05 were considered statistically significant. Pathway-related genes were downloaded from GSEA (https://www.gsea-misgdb.org/gsea/index.jsp). Data visualization was performed using "ggplot2" in R. Lollipop chart of correlation coefficients were plotted in bioinformatics website (http://www.bioinformatics.com.cn/)

4. Establishment of Risk Scoring Model for Prognosis

Univariate Cox regression, LASSO regression, and multivariate Cox regression were applied, using “survival” package in R. A mathematical formula (\(\text{R}\text{i}\text{s}\text{k} \text{s}\text{c}\text{o}\text{r}\text{e} = {\sum }_{\text{i}=1}^{\text{n}}\text{E}\text{x}{\text{p}}_{\text{i}}\ast \text{C}\text{o}{\text{e}}_{\text{i}}\)) was developed to predict the risk score for each patient based on the multivariate Cox regression analysis. All patients are classified into high and low risk groups according to the median risk score.

5. Evaluation of the Value of the Risk Scoring Model

The Kaplan-Meier (K-M) survival analysis was conducted using the R package "survival". The time-dependent relative operating characteristic (ROC) curve was used to assess the accuracy of the established model, calculated with the R package “survival ROC”^[7]. Concordance index (C-index) values were calculated based on Harrell’s C-index^[17], to evaluate the difference between predicted value and actual value of Cox model in survival analysis and judge the prediction accuracy of the prognostic model of tumor patients. To verify the general applicability of the Risk Scoring Model, GSE53624 dataset was downloaded from Gene Expression Omnibus (GEO) database. Risk plot was generated using R package “pheatmap”. Other charts were created with “ggplot2”.

6. Construction of Nomogram Signature

Using univariate Cox regression analysis (p < 0.1) and multivariate Cox regression analysis (p < 0.05), proper prognostic factors were identified. Based on the results of multivariate Cox proportional hazards analysis, a nomogram was formulated to predict 1-, 3-, and 5-year overall survival. Each of the factors was represented by a graphical representation through R package “rms”, which could be used to calculate the risk of death for an individual ESCA patient based on points associated with each risk factor. Similarly, the Nomogram was evaluated by ROC curves and C-index. With the calibration plot, predicted probability and actual occurrence rate are compared. Decision curve analysis (DCA) was used to determine net clinical benefits through R package “rmda”.

7. Immune Analysis

Tumor-infiltrating immune cells were calculated using the CIBERSORT algorithm, with R package ”cibersort”. ESTIMATE (Estimation of Stromal and Immune cells in Malignant Tumor tissues using Expression) algorithm defined immune score, stromal score, Estimate score and tumor purity of ESCA patients via R package “estimate”. TIDE (Tumor Immune Dysfunction and Exclusion) was used to assess the potential response to immune checkpoint inhibitor (ICI) therapy (http://tide.dfci.harvard.edu/). Gene lists encoding immunomodulators and chemokines were downloaded from translation initiation site database (TISDB) website (http://cis.hku.hk/TISIDB/download.php). Lollipop chart of correlation coefficients were plotted in bioinformatics website (http://www.bioinformatics.com.cn/)

8. Mutation Analysis

Somatic mutation profiles were obtained from TCGA database in the form of mutation annotation format (MAF). The mutation profiles and frequencies of genes in high and low risk groups were analyzed and visualized using the R package "Maftools". Tumor mutation burden (TMB) was calculated as mutations per megabase (mut/Mb)^[18].

9. Drug Sensitivity Analysis

The same sample of gene expression and drug sensitivity data was downloaded from the cellminer website (https://discover.nci.nih.gov/cellminer/)^[19]. Drug sensitivity data after clinical laboratory verification and FDA standard certification was filtered. Pearson correlation analysis was conducted to determine the correlation between Risk Scoring Model and drug sensitivity. The drugs were ranked in order of the absolute value of correlation, and the first 2 with strong correlation with each gene were displayed.

10. Statistic Analysis

Data analysis and visualization was performed using R version 4.0.2. Chi-square test was executed for the comparison of categorical variables between groups. Student t test or Kruskal-Wallis test were used to estimate statistical significance between two groups for continuous variables. The correlation between two parameters was assessed through the Spearman correlation analysis. Two-tailed p values less than 0.05 were considered statistically significant.

1. Identification of DEGs and Enrichment Analysis in ESCA

MRNA sequencing results of ESCA patients collected from 150 tumor tissues and 11 adjacent normal tissues were downloaded from TCGA database. Three patients with OS fewer than 30 days were removed. A total of 2195 differentially expressed genes (DEGs) were obtained between the two groups with the screening criteria of p value < 0.01 and |log FC| > 0.5 (Fig. 2A). Among these DEGs, 306 genes were overlapped with mitochondria related genes(Fig. 2B).

Gene ontology (GO) pathway enrichment analysis of the 306 mitochondria related DEGs showed, for Biological Process (BP), DEGs were mainly enriched in the pathway of cofactor metabolic process, small molecule catabolc process, and fatty acid metabolic process; for Cellular Component (CC), DEGs were significantly enriched in the pathway of mitochondrial matrix, mitochondrial inner membrane, and mitochondrial outer membrane; for Molecular Function (MF), DEGs were associated with coenzyme binding, lyase activity, and flavin adenine dinucleotide binding (Fig. 2C).

Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis indicated that the pathways were mainly associated with neurodegeneration - multiple disease, carbon metabolism, and diabetic cardiomyopathy (Fig. 2D).

2. Risk Scoring Model Construction and Validation

Based on the gene expression and overall survival (OS) information of 147 tumor samples, 17 genes were screened out by univariate cox regression analysis of 306 mitochondrial related DEGs (Fig. 3A). Combining the results of LASSO regression analysis, it was considered that the model fit the best when the penalty coefficient was 12 (Fig. 3B), thus the corresponding 12 mitochondrial related DEGs including PDHA1, HSPE1, APOOL, H6PD, HIGD1A, MAOB, MTERF2, AASS, MT-ND1, BCAP31, SLC44A2, CHPT were selected. Then, multivariate cox regression analysis was performed on the 12 genes and six genes APOOL, HIGD1A, MAOB, BCAP31, SLC44A2 and CHPT1 were able to enter the equation as a prognostic predictor (Fig. 3C, Table 1). In ESCA, the expression level of BCAP31 was significantly up-regulated in tumor tissues, and the expression level of APOOL, HIGD1A, MAOB, SLC44A2 and CHPT1 were down-regulated in tumor tissues (Fig. 3D). A subsequent stratification analysis revealed these genes that were not differentially expressed by age, gender, grade, or stage (Fig. S1A). Kaplan-Meier survival curves (Table 2, Fig. S1B) confirmed that higher expression of MAOB and lower expression of BCAP31, CHPT1, and HIGD1A contributed to better survival. The difference of overall survival time between high and low expression level of APOOL and SLC44A2 had no statistical significance.

Table 1. The information of 6 prognosis-related genes.

Gene symbol	Gene ID	Full name	Location	Function of the encoded protein
APOOL	139322	apolipoprotein O like		The protein encoded by this gene contains an apolipoprotein O superfamily domain. This domain is found on proteins in circulating lipoprotein complexes.
HIGD1A	25994	HIG1 hypoxia inducible domain family member 1A	Mitochondrion and Nucleoplasm	Acts upstream of or within negative regulation of apoptotic process. Located in mitochondrion and nucleoplasm. Part of protein-containing complex.
MAOB	4129	monoamine oxidase B	Mitochondrion	The protein encoded by this gene belongs to the flavin monoamine oxidase family. It is a enzyme located in the mitochondrial outer membrane. It catalyzes the oxidative deamination of biogenic and xenobiotic amines and plays an important role in the metabolism of neuroactive and vasoactive amines in the central nervous sysytem and peripheral tissues. This protein preferentially degrades benzylamine and phenylethylamine.
BCAP31	10134	B cell receptor associated protein 31		The encoded protein is a multi-pass transmembrane protein of the endoplasmic reticulum that is involved in the anterograde transport of membrane proteins from the endoplasmic reticulum to the Golgi and in caspase 8-mediated apoptosis.
SLC44A2	57153	solute carrier family 44 member 2	mitochondrion and plasma membrane.	Enables choline transmembrane transporter activity. Involved in choline transport and transmembrane transport.
CHPT1	56994	choline phosphotransferase 1	Golgi membrane	Enables diacylglycerol cholinephosphotransferase activity. Involved in phosphatidylcholine biosynthetic process and platelet activating factor biosynthetic process.

Table 2. Clinical characteristics between low- and high-risk groups.

Variables		Low risk（N=74）	High risk（N=73）	P value
Age				0.939
	＜60	34	34
	≥60	40	39
Barretts Esophagus				0.302
	Yes	10	15
	No	48	45
	N/A	16	13
Columnar Metaplasia Present				0.166
	Yes	8	17
	No	29	31
	N/A	37	25
Esophageal Tumor Cental Location				0.000
	Distal	40	62
	Mid	29	9
	Proximal	4	2
	N/A	1	0
History of Esophageal Cancer				0.115
	Yes	2	6
	No	55	47
	N/A	17	20
Neoplasm Histologic Grade				0.161
	G1/2	42	33
	G3/X	32	40
M				0.026
	M0	66	52
	M1/X	5	13
	N/A	3	8
N				0.007
	N0	40	22
	N1/2/3/X	32	45
	N/A	2	6
T				0.496
	T0	0	1
	T1/2	34	28
	T3/4	38	38
	N/A	2	6
Reflux History				0.327
	Yes	20	26
	No	40	36
	N/A	14	11
Gender				0.176
	Male	60	65
	Female	14	8
Tumor Stage				0.030
	I/II	49	33
	III/IV	22	32
	N/A	3	8
Disease Type				0.000
	Adenomas and Adenocarcinomas	23	47
	Squamous Cell Neoplasms	51	25
	Cystic, Mucinous and Serous Neoplasms	0	1
Alcohol History				0.413
	Yes	50	53
	No	23	18
	N/A	1	2
BMI				0.200
	Underweight	3	5
	Normal	39	30
	Overweight	26	36
	N/A	6	2
Survival Status				0.001
	Alive	56	36
	Dead	18	37

Risk score of each patient was calculated according to the following formula.

Risk Score=APOOL×0.6955447་HIGD1A×0.4618985-MAOB×0.177846་BCAP31×0.5375771།SLC44A2×0.619591་CHPT1×0.3943502

Divided into high and low risk groups based on the median value of risk scores, Kaplan - Meier survival analysis confirmed that the prognosis of patients in high risk group was significantly worse than that in low risk group (p value = 8.8e-06) (Fig. 4A). Figures 4B, C show risk scores and survival status distributions. With the increase of risk score, the risk plot showed that the death toll increased and the patients’ survival time decreased. The heatmap showed the expression level of 6 modeling genes in the two groups (Fig. 4D). The AUC values for 1, 3, and 5 years were 0.738, 0.738, and 0.924, respectively, indicating that the model could be used to predict patient prognosis accurately (Fig. 4E). The C-index of the model was 0.723. As shown in Figs. 4F-J, the results of the validation set (GSE53624) were consistent with our previous research results.

3. Nomogram construction and validation

Clinical information of 147 ESCA patients was downloaded from the UCSC database. By univariate and multivariate Cox hazard regression analysis, Columnar metaplasia present, M stage, reflux history and risk score were selected to construct Nomogram as P value < 0.05 (Fig. S2A). The AUC values for 1, 3, and 5 years were 0.7, 0.722, and 0.835, respectively, indicating the accuracy of the model in differentiation (Fig. S2B). The C-index of Nomogram was 0.777. The calibration plot revealed excellent agreement between observations and predictions (Fig. S2C). Based on this result, Decision Curve Analysis (DCA) suggested the clinical effectiveness of risk score when combined with various clinical features and Nomogram (Fig. S2D). These results revealed that the risk signature could serve as an independent factor for predicting the prognosis of ESCA patients.

4. Pathway analysis in high and low risk groups

1188 DEGs were obtained between the high risk score and the low risk score groups with the screening criteria of p value < 0.01 and |log FC| > 0.5. The DEGs were then analyzed for GO and KEGG enrichment.

For GO BP, DEGs were mainly enriched in “regulation of neuron projection development, axonogenesis, and epidermis development”. For GO CC, DEGs were significantly enriched “cell-cell junction, collagen-containing extracellular matrix, and apical part of cell”. For GO MF, DEGs were associated with “cell adhesion molecule binding, actin binding, actin filament binding” (Fig. 5A).

KEGG enrichment analysis indicated that the DEGs were mainly enriched in “proteoglycans in cancer, axon guidance, hippo signaling pathway, and tight junction” (Fig. 5B).

GSEA was performed to investigate the potential functions and pathways associated with the high risk group and the low risk group. The top 20 enriched pathways in the high risk and low risk group were obtained (Fig. S3). The result demonstrated that “genes down-regulated in metastases from malignant melanoma, genes down-regulated in esophageal adenocarcinoma and Barret's esophagus” were abundantly enriched in low risk group (Fig. 5C). While “Genes up-regulated in early gastric cancer, cervical cancer proliferation cluster” potential were enriched in high risk group (Fig. 5D).

On the basis of KEGG enrichment analysis results, we explored the correlation between risk score and Hippo signaling pathway biomarker genes. The results of Spearman correlation analyses demonstrated that risk score negatively correlated with a majority of the Hippo signaling pathway biomarker genes in ESCA samples, including AMOTL1, WWTR1, NGFR, TEAD2 et, al. (Fig. 5E). Based on GO CC results, Spearman's correlation analysis revealed that risk score was negatively correlated with more cell-cell junction biomarker genes (Fig. 5F).

5. Immune analysis in high and low risk groups

According to CIBERSORT immune analysis, there were significant differences in the types of immune cells in the high and low risk groups (Fig. 6A). Samples with high risk demonstrated higher abundance of memory resting CD4⁺ T cells, resting NK cells, M0 and M2 Macrophages, and lower abundance of M1 Macrophages. Results based on CIBERSORT also suggested that samples in high risk group tended to possess higher infiltration levels of B cell, as well as lower infiltration levels of macrophage, and T cell (Fig. S4A, B). To determine whether there is a correlation between risk score and immune cells, including effector memory CD4⁺ T cell, central memory CD4⁺ T cell, activated CD4⁺ T cell, activated DC, immature DC, plaslacytoid DC, macrophage and NK cell, we conducted a correlation test between immune cell marker genes and risk score, as shown in Figure S4. Figure 6B-I showed marker genes connected with risk score, which was suggestive of a certain correlation between immune cells and risk score.

As a new weapon against cancer, immune checkpoint therapy, which controls the activity of T cells, improved anti-tumor immune responses^[20]. Therefore, we identified both high risk and low risk group immune checkpoint expression (Fig. S4C). The result indicated that there were significant differences in the expression of CD115, TIM1, CD30, CD112, HVEM, HHLA2, SEMA4A, and BTNL2 (Fig. 6J). Afterwards, we calculated the correlation between risk score and the expression of immune checkpoint (Fig. S4D). The results illustrated that BTNL2, SEMA4A, and CD30 were negatively correlated with risk score, while TIM1, CD155, HVEM, CD112, and HHLA2 were positively correlated with risk score (Fig. 6K).

ESTIMATE calculates the tumor purity, the ESTIMATE score, the immune score, and the stromal score of each ESCA sample. Spearman correlation analysis revealed risk score was negatively correlated with the stromal score, while not significantly associated with the tumor purity, the ESTIMATE score, and the immune score(Fig. 7A). And between high risk and low risk group, there was no significant difference in tumor purity, stromal score, immune score, and ESTIMATE score. (Fig. 7B).

ICI therapy has undoubtedly proven to be an effective strategy of immunotherapy, which led to an important breakthrough in the field of antitumor therapy^[21].Thus, the TIDE algorithm was used to predict the potentially different responses to ICB therapy. Patients with higher risk score showed lower level of TIDE score (Fig. 8A). The high risk group contained a higher proportion of responders to ICI therapy compared with low risk group (Fig. 8B). Samples were divided into the high score group and the low score group according to the median value of the stromal score. The low stromal score group contained a higher proportion of responders to ICI therapy compared with low risk group (Fig. 8C). Furthermore, ICI therapy was more effective in patients with high risk scores and low stromal scores as compared to other groups (Fig. 8D). Above all, we speculated that patients with a high risk and low stromal score could benefit more from ICI therapy than patients with a low risk and high stromal score.

6. Mutation analysis in high and low risk groups

Mutation data of the ESCA patients were obtained from TCGA by the maftools package. As shown in Fig. 9A, TP53, TNN, SYNE1, MUC16, DNAH5, CSMD3, HMCN1, KMT2D, LRP1B, PCLO, USH2A, and FLG were frequently mutated in both high risk and low risk group, excluding samples with no mutations called. Among these many mutant genes, the mutation rate of TP53 was significantly different between the high and low risk groups (Fig. 9B). Tumor mutation burden (TMB) scores were calculated for each ESCA patient. Spearman correlation analysis suggested a positive correlation between TMB score and risk score (Fig. 9C), though there was no significant difference in TMB score between high and low risk groups (Fig. 9D). The higher the TMB, the better the effect of tumor immunotherapy. The result is consistent with the result of the high risk group contained a higher proportion of responders to ICI therapy.

7. Drug Sensitivity Analysis of Risk Scoring Model

CellMiner was used to assess the interactions of model genes on drug sensitivity, in order to facilitate better precision treatment of ESCA. As shown in Fig. 10, drugs with the 2 highest correlation with model genes and riskcore were selected, including Simvastatin, Fluvastatin, SNS-314, Nitazoxanide, Everolimus, malacid, Tegafur, 4SC-202, PX-316, Artemether, Selumetinib, Acetalax, BMS-690514,and Sapitinib. A complete list of all the drugs related with model genes and risk score (|r|>0.2, P < 0.05) is provided in the Table S2.

As it is reported in Global Cancer Statistic 2020, ESCA is a cancer with high morbidity and mortality^[22]. There were over 477,900 new cases and 375,000 annual deaths occurring in China^[23]. Early symptoms of ESCA are not typical and easy to ignore. Hospitalized patients are often at a late stage, making treatment difficult. Although tremendous advances have been made in diagnostic technologies and management approaches, ESCA patients still have an unsatisfactory five-year survival rate^[24]. Functional mitochondria, which essential for the cancer cell, is able to alter the mitochondrial bioenergetic and biosynthetic state^[25]. For instance, Chiyo et al have reported that Gal-9 induces mitochondrial-mediated apoptosis in esophageal squamous cell carcinoma cells^[26]. Similarly, in the study of Ma and Ke et al, DS2 induces the mitochondria-mediated apoptosis in human ESCA cells^[27]. MiR-634 activates mitochondrial apoptosis in esophageal squamous cell carcinoma and enhances chemotherapy-induced cytotoxicity through direct targeting of genes associated with mitochondrial homeostasis, antiapoptosis, antioxidant ability, and autophagy^[28]. Therefore, to develop more effective treatment strategies, identifying the molecular etiology involved in ESCA development and progression is imperative.

To investigate the relationship between ESCA and mitochondria, we downloaded the mRNA sequencing results of ESCA patients from TCGA database. 306 mitochondria related DEGs were screened between the tumor and normal groups. GO and KEGG enrichment analysis of the mitochondria related DEGs indicated that molecular metabolic pathways were significantly enriched. In view of this, further research is needed on marker genes associated with diagnosis and prognosis.

Based on univariate cox regression analysis, LASSO regression analysis and multivariate cox regression analysis, six genes MAOB, APOOL, HIGD1A, BCAP31, SLC44A2 and CHPT1 were obtained for inclusion in the Cox proportional - hazards model as prognostic predictors. Notably, MAOB can be used as a novel indicator to predict the progression and prognosis of CRC patients^[29]. APOOL is a cardiolipin - binding component of the Mitofilin/MINOS complex that determines mitochondrial cristae shape^[30]. HIGD1A, a survival factor induced by Hypoxia - Inducible Factor − 1 (HIF1), decreases tumor growth while promotes tumor cell survival in vivo^[31]. Fu et al have reported that BCAP31 plays an essential role in the development of triple-negative breast cancer^[32]. BCAP31 was selected to construct a prognostic signature on ESCA associated with DNA repair in the study of Wang and Li et al, which was consistent with our research^[33]. SLC44A2, whose product is expressed by cells that are central to the pathophysiology of thrombosis, is a susceptibility loci for venous thromboembolism (VTE)^[34]. Downregulation of SLC44A2 was associated with poor outcome of neuroblastoma^[35]. CHPT1 is a tumor-promoting gene in castration-resistant prostate cancer that catalyzes phosphatidylcholine (PtdCho) synthesis and regulates choline metabolism^[36].

On the basis of the median value of the risk score calculated by these six genes, the tumor samples were divided into high and low risk groups. Kaplan-Meier survival analysis revealed a worse prognosis for high risk group than low risk group. This result was confirmed by GSE53624 and suggested the Cox proportional-hazards model was an effective biomarker for predicting the prognosis of ESCA. ROC curves and C-index demonstrated that the Cox proportional-hazards model could accurately predict prognosis.

Enrichment analysis manifested that DEGs between high and low risk groups were enriched in Hippo signaling pathway and cell-cell junction pathways. Hippo signaling pathway regulates cell proliferation, apoptosis, and stem cell renewal^[37]. Dysregulation of the Hippo signaling pathway leads to tumor progression and initiation^[38]. The cell-cell junction plays a crucial role in maintaining normal physiological functions, such as migration, proliferation, and paracellular transport. There are several diseases associated with altered cell-cell junctions.^[39]. ESCA samples showed a negative correlation between risk score and biomarkers of Hippo signaling pathway and cell-cell junction. This indicates that in ESCA the Hippo signaling pathway and cell-cell junctions may be suppressed. However, the molecular mechanisms should be further explored in the future.

Bioinformatics analyses have been conducted in various studies to explore potential pathological mechanisms in ESCA over the past decade. Nevertheless, the mechanisms of immune cell infiltration and the genes involved in immune infiltration in ESCA remain unknown. Hence, we calculated immune score, estimate score, stromal score, and tumor purity. High risk patients had lower levels of stromal score, while tumor purity, ESTIMATE score, and immune score were not significantly affected. Using CIBERSORT, a comparison of 22 immune cell subsets in high risk and low risk samples was conducted. It was found that the high risk group had higher levels of memory resting CD4 + T cells, resting NK cells, and M0 and M2 macrophages, and a lower level of M1 macrophages. Risk scores are moderately correlated with immune cells. Previous studies have reported high levels of CD4 + T cells in ESCAs^[40], which is in agreement with the findings of our study. Tumor-infiltrating innate immune cells, like NK cells and macrophages, are responsible for antitumor immunity^[41]. While M1 type macrophages can kill tumor cells, M2 type macrophages promote tumor growth^[42]. Previous research indicated that esophageal squamous cell carcinomas with high numbers of tumor-infiltrating CD4 + T cells have a favorable prognosis^[43]. It was suggested that increased infiltration of CD4 + T cells can potentially sensitize tumors to immune therapy. This hypothesis was confirmed by the subsequent prediction of the response to ICI therapy. Based on the TIDE algorithm, high risk group presented relatively lower TIDE score compared with low risk group. ICI therapy was more effective in patients with higher risk scores and lower stromal scores. As a consequence, high-risk group might benefit from immunotherapy.

CellMiner is a Web site that provides genomic and pharmacologic tools to identify drug patterns and transcripts in the NCI-60 cell line^[44], including 22,379 gene expression levels, and 20,503 compounds incorporating 102 Food and Drug Administration (FDA)-approved drugs^[45]. In our study, we analyzed the relationship between model genes with the drug sensitivity of anticancer drugs, offering a novel insight into treating tumors and avoiding resistance. According to some studies, the use of statins, including Simvastatin and Fluvastatin, was significantly associated with a decrease in all-cause and cancer-specific mortality in patients with ESCA^[46], this validates the results of our screening. Besides, Everolimusis used for treating breast cancer, neuroendocrine tumors, and kidney cancer as an immunosuppressive drug^[47]. A combination of tegafur and uracil is widely used for treatment of breast cancer, and other solid tumors like lung, gastric and rectal cancers^[48]. There is evidence that 4SC-202 inhibits colorectal cancer, hepatocellular carcinoma, and urothelial carcinoma^[48–50]. Selumetinib has been approved by the FDA for the treatment of adult melanoma, colon cancer, and non-small cell lung cancer, has a favorable toxicity profile, and has shown antitumor activity in preliminary trials^[51]. Colorectal cancer and non-small cell lung carcinoma are effectively suppressed by BMS-690514^[52–53]. In treating ESCA, however, these drugs have not been fully investigated. At present, only a few chemotherapeutic drugs are effective in the treatment of ESCA, and our study sheds a new light on the search for new anti-ESCA drugs.

Nonetheless, several limitations of this study should be addressed. First, all analyses were performed using retrospective data from public databases; the lacking prospective data from multi center cohorts would produce more reliable results. Second, according to previous report, there are regional differences in the mortality of ESCA because of different Racial characteristics, living habits, and other complex risk factors^[54–56]. By obtaining more detailed information about the clinical samples and by categorizing them more precisely, we can reach more accurate conclusions. Third, the detailed influence mechanisms of the risk model genes on immunity should be verified in vivo and in vitro experiments in the further.

Taken together, we focused on the role of mitochondria-related genes in esophageal carcinoma development, and proposed a prognostic signature for individualized integrative assessment. It may be possible to predict the prognosis of ESCA patients more accurately and choose better immunotherapy options in the future, with further research needed.

Authors' contributions

X.Z., H.W., J.Z. and M.L. designed the experiments. X.Z. and H.W. completed the experiments. X.Z., H.W., J.N., Y.H., W.Z. and J.C. analyzed the data. X.Z and M.L. drafted the manuscript. J.N., L.L, J.Z, C.Z., and M.L. edited the manuscript. All authors reviewed the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (No. 81902513) and Applied Basic Research Project of Shanxi Province (No.202103021224228 and No.20210302124376).

Availability of data and materials

Gene expression RNAseq (HTseq) and somatic mutation profiles was downloaded from The Cancer Genome Atlas database (TCGA: https://portal.gdc.cancer.gov/). Phenotype and survival data was downloaded from the University of California SANTA CRUZ (UCSC: https://xenabrowser.net/datapages/). GEO datasets are all downloadable from the National Center for Biotechnology Information (NCBI) online Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo). Human mitochondrial related genes were retrieved from Gene Set Enrichment Analysis (http://www.gsea-msigdb.org/gsea/msigdb/index.jsp). Pathway-related genes were downloaded from GSEA (https://www.gsea-misgdb.org/gsea/index.jsp). Gene lists encoding immunomodulators and chemokines were downloaded from translation initiation site database (TISDB) website (http://cis.hku.hk/TISIDB/download.php). The same sample of gene expression and drug sensitivity data was downloaded from the cellminer website (https://discover.nci.nih.gov/cellminer/).

Competing interests

The authors declare no competing interests.

Conflict of interest

The authors declare no competing interests.

Zhao Y, Wang Y, Shan L, Peng C, Zhang W, Zhao X. A network meta-analysis for neoadjuvant and adjuvant treatments for resectable squamous cell carcinoma of esophagus. Sci Rep, 2021, 11(1). DOI: 10.1038/s41598-021-86102-8.
Djärv T, Derogar M, Lagergren P. Influence of co-morbidity on long-term quality of life after oesophagectomy for cancer. Br J Surg, 2014, 101(5): 495-501. DOI: 10.1002/bjs.9417.
Arnal M J D. Esophageal cancer: risk factors, screening and endoscopic treatment in western and eastern countries. World J Gastroenterol, 2015, 21(26): 7933. DOI: 10.3748/wjg.v21.i26.7933.
Liang Z, Xu J, Gu C. Novel role of the sry-related high-mobility-group box d gene in cancer. Semin Cancer Biol, 2020, 67: 83-90. DOI: 10.1016/j.semcancer.2019.07.011.
Chandran U R, Medvedeva O P, Barmada M M, Blood P D, Chakka A, Luthra S, et al. Tcga expedition: a data acquisition and management system for tcga data. Plos One, 2016, 11(10): e165395. DOI: 10.1371/journal.pone.0165395.
Giampazolias E, Tait S W G. Mitochondria and the hallmarks of cancer. Febs J, 2016, 283(5): 803-814. DOI: 10.1111/febs.13603.
Wang Z, Zhang N, Lv J, Ma C, Gu J, Du Y, et al. A five-gene signature for recurrence prediction of hepatocellular carcinoma patients. Biomed Res Int, 2020, 2020: 1-13. DOI: 10.1155/2020/4037639.
Kudryavtseva A V, Krasnov G S, Dmitriev A A, Alekseev B Y, Kardymon O L, Sadritdinova A F, et al. Mitochondrial dysfunction and oxidative stress in aging and cancer. Oncotarget, 2016, 7(29): 44879-44905. DOI: 10.18632/oncotarget.9821.
Cheng L, Shi L, Wu J, Zhou X, Li X, Sun X, et al. A hederagenin saponin isolated from clematis ganpiniana induces apoptosis in breast cancer cells via the mitochondrial pathway. Oncol Lett, 2017. DOI: 10.3892/ol.2017.7494.
Lopez J, Tait S W G. Mitochondrial apoptosis: killing cancer using the enemy within. Br J Cancer, 2015, 112(6): 957-962. DOI: 10.1038/bjc.2015.85.
Wang Y C, Wu D W, Wu T C, Wang L, Chen C Y, Lee H. Dioscin overcome tki resistance in egfr-mutated lung adenocarcinoma cells via down-regulation of tyrosine phosphatase shp2 expression. Int J Biol Sci, 2018, 14(1): 47-56. DOI: 10.7150/ijbs.22209.
Zheng W B, Li Y J, Wang Y, Yang J, Zheng C C, Huang X H, et al. Propafenone suppresses esophageal cancer proliferation through inducing mitochondrial dysfunction. Am J Cancer Res, 2017, 7(11): 2245-2256.
Chen Z, Wang B, Yu F, Chen Q, Tian Y, Ma S, et al. The roles of mitochondria in radiation-induced autophagic cell death in cervical cancer cells. Tumor Biology, 2016, 37(3): 4083-4091. DOI: 10.1007/s13277-015-4190-8.
Toh U, Sasada T, Takahashi R, Itoh K, Akagi Y. Tumor immunotherapy of esophageal and gastric cancers. Berlin, Heidelberg: Springer Berlin Heidelberg: 185-197.
Tian M, Liu W, Tao C, Tang Z, Zhou Y, Song S, et al. Prediction of overall survival in resectable intrahepatic cholangiocarcinoma: isicc ‐applied prediction model. Cancer Sci, 2020, 111(4): 1084-1092. DOI: 10.1111/cas.14315.
Ritchie M E, Phipson B, Wu D, Hu Y, Law C W, Shi W, et al. Limma powers differential expression analyses for rna-sequencing and microarray studies. Nucleic Acids Res, 2015, 43(7): e47. DOI: 10.1093/nar/gkv007.
HARRELL F E, LEE K L, MARK D B. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med, 1996, 15(4): 361-387. DOI: 10.1002/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4.
Mayakonda A, Lin D, Assenov Y, Plass C, Koeffler H P. Maftools: efficient and comprehensive analysis of somatic variants in cancer. Genome Res, 2018, 28(11): 1747-1756. DOI: 10.1101/gr.239244.118.
Reinhold W C, Sunshine M, Varma S, Doroshow J H, Pommier Y. Using cellminer 1.6 for systems pharmacology and genomic analysis of the nci-60. Clin Cancer Res, 2015, 21(17): 3841-3852. DOI: 10.1158/1078-0432.CCR-15-0335.
Sharma P, Allison J P. The future of immune checkpoint therapy. Science, 2015, 348(6230): 56-61. DOI: 10.1126/science.aaa8172.
Wu Y, Zhang X, Feng H, Hu B, Deng Z, Wang C, et al. Exploration of redox-related molecular patterns and the redox score for prostate cancer. Oxid Med Cell Longev, 2021, 2021: 4548594. DOI: 10.1155/2021/4548594.
Sung H, Ferlay J, Siegel R L, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: a cancer journal for clinicians, 2021, 71(3): 209-249. DOI: 10.3322/caac.21660.
Chen W, Zheng R, Baade P D, Zhang S, Zeng H, Bray F, et al. Cancer statistics in china, 2015. CA: A Cancer Journal for Clinicians, 2016, 66(2): 115-132. DOI: 10.3322/caac.21338.
Yoshimizu S, Yoshio T, Ishiyama A, Tsuchida T, Horiuchi Y, Omae M, et al. Long-term outcomes of combined endoscopic resection and chemoradiotherapy for esophageal squamous cell carcinoma with submucosal invasion. Dig Liver Dis, 2018, 50(8): 833-838. DOI: 10.1016/j.dld.2018.01.138.
Wallace D C. Mitochondria and cancer. Nat Rev Cancer, 2012, 12(10): 685-698. DOI: 10.1038/nrc3365.
Chiyo T, Fujita K, Iwama H, Fujihara S, Tadokoro T, Ohura K, et al. Galectin-9 induces mitochondria-mediated apoptosis of esophageal cancer in vitro and in vivo in a xenograft mouse model. International Journal of Molecular Sciences, 2019, 20(11): 2634. DOI: 10.3390/ijms20112634.
Ma Y, Ke Y, Zi X, Zhao F, Yuan L, Zhu Y, et al. Induction of the mitochondria-mediated apoptosis in human esophageal cancer cells by ds2, a newly synthetic diterpenoid analog, is regulated by bax and caused by generation of reactive oxygen species. Oncotarget, 2016, 7(52): 86211-86224. DOI: 10.18632/oncotarget.13367.
Fujiwara N, Inoue J, Kawano T, Tanimoto K, Kozaki K, Inazawa J. Mir-634 activates the mitochondrial apoptosis pathway and enhances chemotherapy-induced cytotoxicity. Cancer Res, 2015, 75(18): 3890-3901. DOI: 10.1158/0008-5472.CAN-15-0257.
Yang Y, Chien M, Lai T, Su C, Jan Y, Hsiao M, et al. Monoamine oxidase b expression correlates with a poor prognosis in colorectal cancer patients and is significantly associated with epithelial-to-mesenchymal transition-related gene signatures. International Journal of Molecular Sciences, 2020, 21(8): 2813. DOI: 10.3390/ijms21082813.
Weber T A, Koob S, Heide H, Wittig I, Head B, van der Bliek A, et al. Apool is a cardiolipin-binding constituent of the mitofilin/minos protein complex determining cristae morphology in mammalian mitochondria. Plos One, 2013, 8(5): e63683. DOI: 10.1371/journal.pone.0063683.
Ameri K, Jahangiri A, Rajah A M, Tormos K V, Nagarajan R, Pekmezci M, et al. Higd1a regulates oxygen consumption, ros production, and ampk activity during glucose deprivation to modulate cell survival and tumor growth. Cell Rep, 2015, 10(6): 891-899. DOI: 10.1016/j.celrep.2015.01.020.
Fu W, Sun H, Zhao Y, Chen M, Yang X, Liu Y, et al. Bcap31 drives tnbc development by modulating ligand-independent egfr trafficking and spontaneous egfr phosphorylation. Theranostics, 2019, 9(22): 6468-6484. DOI: 10.7150/thno.35383.
Wang L, Li X, Zhao L, Jiang L, Song X, Qi A, et al. Identification of dna-repair-related five-gene signature to predict prognosis in patients with esophageal cancer. Pathology and Oncology Research, 2021, 27. DOI: 10.3389/pore.2021.596899.
Germain M, Chasman D I, de Haan H, Tang W, Lindström S, Weng L, et al. Meta-analysis of 65,734 individuals identifies tspan15 and slc44a2 as two susceptibility loci for venous thromboembolism. The American Journal of Human Genetics, 2015, 96(4): 532-542. DOI: 10.1016/j.ajhg.2015.01.019.
Lasorsa V A, Cimmino F, Ognibene M, Mazzocco K, Erminio G, Morini M, et al. 19p loss is significantly enriched in older age neuroblastoma patients and correlates with poor prognosis. Npj Genom Med, 2020, 5(1). DOI: 10.1038/s41525-020-0125-4.
Wen S, He Y, Wang L, Zhang J, Quan C, Niu Y, et al. Aberrant activation of super enhancer and choline metabolism drive antiandrogen therapy resistance in prostate cancer. Oncogene, 2020, 39(42): 6556-6571. DOI: 10.1038/s41388-020-01456-z.
Zhang C, Xiong J, Yang Q, Wang Y, Shi H, Tian Q, et al. Profiling and bioinformatics analyses of differential circular rna expression in prostate cancer cells. Future Sci Oa, 2018, 4(9): A340. DOI: 10.4155/fsoa-2018-0046.
Lee U, Cho E, Jho E. Regulation of hippo signaling by metabolic pathways in cancer. Biochimica et Biophysica Acta (BBA) - Molecular Cell Research, 2022, 1869(4): 119201. DOI: 10.1016/j.bbamcr.2021.119201.
Gray K M, Jung J W, Inglut C T, Huang H, Stroka K M. Quantitatively relating brain endothelial cell–cell junction phenotype to global and local barrier properties under varied culture conditions via the junction analyzer program. Fluids Barriers Cns, 2020, 17(1). DOI: 10.1186/s12987-020-0177-y.
Yang Z, Novak A J, Stenson M J, Witzig T E, Ansell S M. Intratumoral cd4+cd25+ regulatory t-cell-mediated suppression of infiltrating cd4+ t cells in b-cell non-hodgkin lymphoma. Blood, 2006, 107(9): 3639-3646. DOI: 10.1182/blood-2005-08-3376.
Vitale M, Cantoni C, Pietra G, Mingari M C, Moretta L. Effect of tumor cells and tumor microenvironment on nk-cell function. Eur J Immunol, 2014, 44(6): 1582-1592. DOI: 10.1002/eji.201344272.
Huang R, Liu J, Li H, Zheng L, Jin H, Zhang Y, et al. Identification of hub genes and their correlation with immune infiltration cells in hepatocellular carcinoma based on geo and tcga databases. Front Genet, 2021, 12: 647353. DOI: 10.3389/fgene.2021.647353.
Cho Y, Miyamoto M, Kato K, Fukunaga A, Shichinohe T, Kawarada Y, et al. Cd4+ and cd8+ t cells cooperate to improve prognosis of patients with esophageal squamous cell carcinoma. Cancer Res, 2003, 63(7): 1555-1559.
Reinhold W C, Sunshine M, Liu H, Varma S, Kohn K W, Morris J, et al. Cellminer: a web-based suite of genomic and pharmacologic tools to explore transcript and drug patterns in the nci-60 cell line set. Cancer Res, 2012, 72(14): 3499-3511. DOI: 10.1158/0008-5472.CAN-12-1370.
Shankavaram U T, Varma S, Kane D, Sunshine M, Chary K K, Reinhold W C, et al. Cellminer: a relational database and query tool for the nci-60 cancer cell lines. Bmc Genomics, 2009, 10(1): 277. DOI: 10.1186/1471-2164-10-277.
Nguyen T, Khan A, Liu Y, El-Serag H B, Thrift A P. The association between statin use after diagnosis and mortality risk in patients with esophageal cancer: a retrospective cohort study of united states veterans. Am J Gastroenterol, 2018, 113(9): 1310. DOI: 10.1038/s41395-018-0169-6.
Mine S, Hishima T, Suganuma A, Fukumoto H, Sato Y, Kataoka M, et al. Interleukin-6-dependent growth in a newly established plasmablastic lymphoma cell line and its therapeutic targets. Sci Rep, 2017, 7(1). DOI: 10.1038/s41598-017-10684-5.
Fu M, Wan F, Li Z, Zhang F. 4sc-202 activates ask1-dependent mitochondrial apoptosis pathway to inhibit hepatocellular carcinoma cells. Biochem Biophys Res Commun, 2016, 471(2): 267-273. DOI: 10.1016/j.bbrc.2016.01.030.
Zhijun H, Shusheng W, Han M, Jianping L, Li-sen Q, Dechun L. Pre-clinical characterization of 4sc-202, a novel class i hdac inhibitor, against colorectal cancer cells. Tumor Biology, 2016, 37(8): 10257-10267. DOI: 10.1007/s13277-016-4868-6.
Pinkerneil M, Hoffmann M J, Kohlhof H, Schulz W A, Niegisch G. Evaluation of the therapeutic potential of the novel isotype specific hdac inhibitor 4sc-202 in urothelial carcinoma cell lines. Target Oncol, 2016, 11(6): 783-798. DOI: 10.1007/s11523-016-0444-7.
Irving J, Matheson E, Minto L, Blair H, Case M, Halsey C, et al. Ras pathway mutations are prevalent in relapsed childhood acute lymphoblastic leukemia and confer sensitivity to mek inhibition. Blood, 2014, 124(23): 3420-3430. DOI: 10.1182/blood-2014-04-531871.
Yang S, Ren X, Liang Y, Yan Y, Zhou Y, Hu J, et al. Knk437 restricts the growth and metastasis of colorectal cancer via targeting dnaja1/cdc45 axis. Oncogene, 2020, 39(2): 249-261. DOI: 10.1038/s41388-019-0978-0.
de La Motte Rouge T, Galluzzi L, Olaussen K A, Zermati Y, Tasdemir E, Robert T, et al. A novel epidermal growth factor receptor inhibitor promotes apoptosis in non–small cell lung cancer cells resistant to erlotinib. Cancer Res, 2007, 67(13): 6253-6262. DOI: 10.1158/0008-5472.CAN-07-0538.
Ma C, Congly S E, Chyou D E, Ross-Driscoll K, Forbes N, Tsang E S, et al. Factors associated with geographic disparities in gastrointestinal cancer mortality in the united states. Gastroenterology, 2022. DOI: 10.1053/j.gastro.2022.04.019.
Middleton D R S, Mmbaga B T, Menya D, Dzamalala C, Nyakunga-Maro G, Finch P, et al. Alcohol consumption and oesophageal squamous cell cancer risk in east africa: findings from the large multicentre esccape case-control study in kenya, tanzania, and malawi. The Lancet Global Health, 2022, 10(2): e236-e245. DOI: 10.1016/S2214-109X(21)00506-4.
Liao Y, Li C, Xia C, Zheng R, Xu B, Zeng H, et al. Spatial distribution of esophageal cancer mortality in china: a machine learning approach. Int Health, 2021, 13(1): 70-79. DOI: 10.1093/inthealth/ihaa022.

No competing interests reported.

Supplementary.zip

Download PDF

Journal Publication

published 29 Mar, 2023

Read the published version in Functional & Integrative Genomics →

Editorial decision: Major revision
02 Feb, 2023
Reviews received at journal
12 Jan, 2023
Reviewers agreed at journal
12 Jan, 2023
Reviewers invited by journal
11 Jan, 2023
Editor assigned by journal
10 Jan, 2023
Submission checks completed at journal
10 Jan, 2023
First submitted to journal
08 Jan, 2023

You are reading this latest preprint version

A Novel Mitochondria-Related Gene Signature in Esophageal Carcinoma: Prognostic, Immune, and Therapeutic Features

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Materials And Methods

1. Datasets and source

2. Differential Expression Analysis

3. Enrichment Analysis

4. Establishment of Risk Scoring Model for Prognosis

5. Evaluation of the Value of the Risk Scoring Model

6. Construction of Nomogram Signature

7. Immune Analysis

8. Mutation Analysis

9. Drug Sensitivity Analysis

10. Statistic Analysis

Result

1. Identification of DEGs and Enrichment Analysis in ESCA

2. Risk Scoring Model Construction and Validation

3. Nomogram construction and validation

4. Pathway analysis in high and low risk groups

5. Immune analysis in high and low risk groups

6. Mutation analysis in high and low risk groups

7. Drug Sensitivity Analysis of Risk Scoring Model

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1