Identification and validation of an anoikis-related gene signature based on machine learning algorithms in esophageal squamous cell carcinoma

doi:10.21203/rs.3.rs-2350536/v1

Download PDF

Research Article

Identification and validation of an anoikis-related gene signature based on machine learning algorithms in esophageal squamous cell carcinoma

https://doi.org/10.21203/rs.3.rs-2350536/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Purpose

Anoikis plays a key role in the process of tumor metastasis. This study aims to investigate the characteristic of anoikis-related genes in esophageal squamous cell carcinoma (ESCC).

Methods

The differentially expressed anoikis-related genes in the TCGA-ESCC cohort were identified. The key anoikis-related genes were selected by least absolute shrinkage and selection operator (LASSO) regression and support vector machine-recursive feature elimination (SVM-RFE), and their expression were verified by qRT-PCR. Multivariate Cox regression was applied to construct the anoikis-related gene signature (ARGS). GSEA was applied to investigate the differences of biological function and pathway in different ARGS subgroups. Immune cell analysis was analyzed by ssGSEA. The expression characteristics of key anoikis-related genes in the single-cell dataset derived from the TISCH database. Immunotherapy response prediction was performed by the TIDE algorithm.

Results

The signature containing 5 key anoikis-related genes (CLDN1, EGFR, PLK1, SATB1, and TNFSF10) was constructed. CLDN1, PLK1, SATB1, and TNFSF10 were shown to be highly expressed in ESCC by qRT-PCR. The ARGS-high group had enriched in more abundant cells and immune-related pathways. Additionally, the ARGS-high group benefited well from immunotherapy, while the ARGS-low group was more sensitive to chemotherapy.

Conclusion

This study identified 5 key anoikis-related genes and conducted the ARGS, which can help predict prognosis and may guide treatment strategies for ESCC patients.

Anoikis

esophageal squamous cell carcinoma

machine learning

tumor microenvironment

immunotherapy

Esophageal cancer is seventh cancer with the incidence rate in the world. In 2020, 540000 people died of esophageal cancer worldwide[1]. And the pathological type of about 90% of patients is esophageal squamous cell carcinoma (ESCC)[2]. Although the comprehensive treatment of esophageal cancer has made significant progress, many patients have already distant metastasis at the time of initial diagnosis, thus losing the opportunity for surgical treatment. The 5-year survival rate of esophageal squamous cell carcinoma is only 15–20%[3]. In view of the excellent progress made in immune checkpoint inhibitors (PD-1, PD-L1, and CTLA-4), immunotherapy has become one of the important ways to prolong the survival of ESCC patients[4]. However, most patients have no obvious response to immunotherapy[5], it is of great practical significance to identify which patients are sensitive to immunotherapy to reduce the economic and physical burden of patients with advanced ESCC.

Anoikis was first described in 1994[6], it was defined as apoptosis caused by the interaction disorder between epithelial cells and extracellular matrix (ECM). In essence, anoikis is an apoptosis process that performs a vital role in maintaining normal cell development and inhibiting tumor metastasis[7]. Under physiological conditions, exfoliated epithelial cells maintain tissue homeostasis by activating apoptosis. While in tumors, cancer cells inhibit anoikis through multiple mechanisms, leading to their continued proliferation in other tissues or organs after shedding from the primary site[8]. The reactivation of anoikis can limit the metastasis of cancer cells, which may become a new target for cancer treatment. Previous studies have shown that dysregulation of anoikis was closely related to poor prognosis in prostate cancer and colon cancer[9, 10]. However, in ESCC, research involving anoikis has not been fully developed.

This study focused on exploring the characteristics of anoikis-related genes in ESCC and developed a new prognostic signature. The expression of key genes was verified by qRT-PCR. In addition, based on this signature typing, we explored the differences in biological pathways, tumor microenvironment (TME), and the prediction of the immune response.

2.1 Data collection

The flowchart of this study was shown in Fig. 1. ESCC transcriptome profiles and clinical information are downloaded from The Cancer Genome Atlas (TCGA). After excluding the data of missing key clinical information, 78 ESCC and 11 normal cases were collected. Anoikis-related genes were obtained from GeneCard database[11] (https://www.genecards.org/). According to the criterion relevance scores > 0.4, we got 473 anoikis-related genes (Table S1).

2.2 Identification and enrichment analysis of differentially expressed genes

The 'limma' package was used to analyze differentially expressed anoikis-related genes between ESCC and normal tissues[12]. The threshold was defined as false discovery rate (FDR) < 0.05 and log2 |fold change| ≥ 1. Heatmap and volcano plot showed these results. Then, the "clusterProfiler" package was used for enrichment analysis of GO and KEGG[13].

2.3 Construction of the anoikis-related gene signature

Univariate cox regression analysis identified prognostic anoikis-related genes(p < 0.05). LASSO regression is a generalized linear model, which can reduce the overfitting of variables. Support vector machine-recursive feature elimination (SVM-RFE) is a powerful feature selection algorithm, widely used in clustering and regression[14]. We used these two machine learning algorithms for analysis, and then obtained the key anoikis-related genes by taking their intersection. Then, the ARGS was constructed through multivariate cox regression analysis, and the area under the curve (AUC) was used for evaluating the prediction ability of ARGS.

2.4 Gene set enrichment analysis

In order to explore the biological characteristics and function in different ARGS subgroups, we downloaded 'c5.go.v7.5.1.symbols' from the molecular signature database to execute GSEA (version: 4.2.3; Broad Institute, Inc, USA)[15].

2.5 Evaluation of tumor microenvironment

The scores of 16 immune cells and 13 immune-related pathways in ESCC were calculated using single set gene set enrichment analysis (ssGSEA) based on “GSVA” package[16]. TISCH is a single-cell database focusing on TME[17]. We explored the TME characteristics of hub anoikis-related genes at the single-cell level through TISCH.

2.6 Prediction of immunotherapy response

Tumor Immune Dysfunction and Exclusion (TIDE, http://tide.dfci.harvard.edu/) is an online analysis website that simulates tumor immune escape to predict cancer immunotherapy response[18]. We calculated the TIDE score of each patient on the ground of the expression profiles of ESCC. The higher the TIDE score, the more likely immune escape will occur during immunotherapy.

2.7 Drug sensitivity analysis

We calculated the half-maximal inhibitory concentration (IC50) of common chemotherapy drugs by using the “pRRophic” package[19], and analyzed their differences in different subgroups.

2.8 Quantitative Real-time PCR

This study was approved by the Sun Yat-sen University Cancer Center ethics committee (GZR2018-120). The samples were derived from patients who had undergone radical surgery at our center. Total RNA was extracted from 8 pairs of ccRCC tissues and their adjacent non-tumor tissues by TRIzol reagent (TIANGEN, Beijing, China). A UV spectrophotometer was used to determine the concentration and purity of RNA. cDNA was synthesized by reverse transcriptase and amplified. Each sample was measured in triplicate. The PCR primers used for amplification were as follows: SATB1, 5′- ACAGAGGTGTCTTCCGAA-3′ (forward), 5′- TCTGAAAGCAAGCCCTGAGT-3′ (reverse); PLK1, 5′-CCAGGATCACACCAAGCT-3′ (forward), 5′- CCAGGATCACACCAAGCT − 3′ (reverse); CLDN1, 5′-GGCTGAGACACTGAAGAAGT-3′ (forward), 5‘- AGGCACTGAACCACATGA-3′ (reverse); TNFSF10, 5′- TGCTGATCGTGATCTTCACAG-3′ (forward), 5′- GGAGTCTTTCTAACGAGCTGAC-3′ (reverse); GAPDH as an endogenous control, 5′-ATCAAGAAGGTGGTGAAGCAGG-3′ (forward), 5′-CGTCAAAGGTGGAGGAGTGG-3′ (reverse). The relative expression levels of UPF1 were calculated using the 2 − ΔΔCT method.

2.9 Statistical analysis

All charts and statistical analyses in this study were performed using R software (version: 4.1.3) and SPSS (version 26.0; IBM, Chicago, USA). The difference between different subgroups was analyzed by Wilcoxon rank sum test. Student t-test was used to analyze the expression of key genes in ESCC and normal tissues. Correlation analysis was used Pearson method. P < 0.05 was considered significant.

3.1 Identification and enrichment analysis of differentially expressed anoikis-related genes

We identified 165 differentially expressed anoikis-related genes from ESCC and normal tissues (Fig. 2A, B). GO enrichment analysis showed that these genes were mainly centered on multiple apoptosis-related pathways (Fig. 2C). KEGG enrichment analysis indicated that differentially expressed anoikis-related genes were mainly enriched in p53 signal pathway, PI3K-Akt signal pathway, cell cycle, and ECM receptor interaction (Fig. 2D).

3.2 Construction of anoikis-related gene signature

Univariate cox regression analysis found that 21 genes were related to overall survival (OS) (Fig. 3A). Subsequently, we included these 21 genes into LASSO regression (Fig. 3B, C) and SVM-RFE (Fig. 3D, E) for analysis, and obtained 7 overlapping genes (Fig. 3F). After stepwise multivariate cox regression analysis, we finally identified five key anoikis-related genes (CLDN1, EGFR, PLK1, SATB1, and TNFSF10) and their corresponding coefficients (Table 1). Risk score = CLDN1exp*0.638104-EGFRexp*1.2036-PLK1exp*0.96132-SATB1exp*1.56641 + TNFSF10exp*0.999589. ESCC patients were divided into ARGS-high (n = 39) and ARGS-low (n = 39) group according to the median risk score. The prognosis of the two subgroups was discernibly different (p < 0.0001, Fig. 3G). The distribution of risk score, survival status, and key anoikis-related genes expression between the two subgroups were shown in Figure S1. Subsequently, we randomly divided ESCC patients into training and test datasets using the “caret” package and conducted KM survival analysis respectively. The results showed that in the training/test group, the ARGS-low group showed a better prognosis (Figure S2). Additionally, we also used AUC to evaluate the ability of signature to predict prognosis, and the results showed that signature has excellent predictive efficacy (AUC = 0.774, 95% CI 0.665–0.884, Fig. 3H).

3.3 Verification of key anoikis-related genes expression

Comparing with the normal tissues, CLDN1, EGFR, PLK1, and TNFSF10 were highly expressed in ESCC, while SATB1 had higher expression in normal tissues (Fig. 4A). Then, the qRT-PCR results from 8 paired ESCC samples obtained from our Cancer Center revealed that the expression of CLDN1, PLK1, and TNFSF10 in cancer tissues was apparently higher than that in normal esophageal tissues (p < 0.05) (Fig. 4B-D), which agreed with the results from bioinformatics analysis. However, the expression of SATB1 was contrary (Fig. 4E).

3.4 Analysis of biological function and pathways

GSEA enrichment analysis showed that the ARGS-high group primarily concentrated on multiple immune cell-mediated pathways, such as antigen presentation and processing, T cells, natural killer, macrophages, and neutrophils (Fig. 5A). In the ARGS-low group, it was mainly enriched in the pathways related to cell aging, and epithelial-mesenchymal transformation (EMT) (Fig. 5B).

3.5 Differences of tumor microenvironment

Based on ssGSEA, we analyzed the differences of immune cells and function between the two subgroups. A variety of immune cells were significantly infiltrated in the ARGS-high group, including CD8 + T cells, DC, NK cells, neutrophils, and macrophages (Fig. 6A). Similarly, scores of various immune-related pathways were higher in the ARGS-high group (Fig. 6B). Furthermore, we utilized the single-cell dataset (GSE160269) from the TISCH database to analyze the differences of 5 key anoikis-related genes expression in the TME. The distribution of various types of cells was shown in Fig. 6C. CLDN1 was mainly expressed in malignant cells (Fig. 6D) and EGFR was mainly expressed in malignant cells and fibroblasts (Fig. 6E). PLK1 was mainly expressed in proliferating T Cells and malignant cells (Fig. 6F). SATB1 is mainly expressed in a variety of immune cells, such as CD4Tconv (Conventional CD4 T Cells), Treg (Regulatory T Cells), CD8Tex (Exhausted CD8 T Cells), and Mono/Macro (Monocells or Macrophages) (Fig. 6G). TNFSF10 was expressed in malignant cells, stromal cells, and various immune cells (Fig. 6H). Then, we also quantified the expression of 5 genes in the TME using the bar graphs (Figure S3).

3.6 Prediction of immunotherapy response

Based on the TIDE database, we calculated the TIDE score of each ESCC patient. We found that the ARGS-low group had higher TIDE score and T cell exclusion score, while the ARGS-high group has higher T cell dysfunction score, IFNG (interferon gamma), and Merck18 (T cell-inflamed subset) (Fig. 7A). These results indicated that the ARGS-low group may be more likely to have immune escape during immunotherapy. In addition, a higher proportion of patients in the ARGS-high group responded to immunotherapy (Fig. 7B).

3.7 Prediction of chemotherapy drug sensitivity

Chemotherapy is still the first-line treatment for advanced ESCC. We analyzed the sensitivity of multiple chemotherapy drugs between two subgroups (Fig. 8A-I), and found that the ARGS-low group was more sensitive to multiple chemotherapy drugs, such as cisplatin, doxorubicin, etoposide, gemcitabine, and vinorelbine. While camptothecin, paclitaxel, vinblastine, and docetaxel had no significant discrepancy in sensitivity between the two subgroups.

Although great changes have taken place in cancer diagnosis methods and treatment strategies in the past decades, the prognosis of ESCC patients is still poor. Additionally, because of tumor heterogeneity, TNM stage cannot fully reflect the molecular characteristics of ESCC patients[20], and often the same TNM staging has different prognosis. Therefore, it is urgent to find new biomarkers. Previous studies have found that anoikis is crucial for tumor metastasis. Cells shed from the original site will activate anoikis to maintain tissue homeostasis[21]. However, since cancer cells have lost their sensitivity to anoikis, they can continue to proliferate without adhering to ECM and metastasize to other organs[22]. Overcoming the resistance of cancer cells to anoikis may become a new therapeutic direction.

In this study, we identified key anoikis-related genes through LASSO regression and SVM-RFE machine learning algorithms and constructed prognosis-related ARGS, including CLDN1, EGFR, PLK1, SATB1 and TNFSF10. The expression of these five genes in ESCC and normal tissues was significantly different, which was confirmed by qRT-PCR. CLDN1 is a membrane protein that constitutes the tight junction of cells and participates in the invasion and metastasis of cancer. The expression of CLDN1 in advanced and metastatic colon cancer lesions was significantly higher than that in normal cells[23]. In ESCC, EGFR is overexpressed in about half of the patients[24]. Cetuximab, which targets EGFR overexpression in ESCC, was carrying out several related studies. Although it has not significantly improved the prognosis of ESCC patients, the study of the population with significant expression of this molecular target will eventually bring dawn to ESCC patients[25]. PLK1 is a serine/threonine kinase that regulates cell cycle and is highly expressed in many cancers. PLK1 inhibitors combined with other chemotherapy drugs have been proven to reduce drug resistance[26]. High expression of SATB1 promotes the occurrence and metastasis of multiple cancers by affecting cell apoptosis and the EMT process[27]. However, in this study, SATB1 was low expressed in ESCC compared with normal tissues, while qRT-PCR showed the opposite result. The possible reason was that the sample size was too small. TNFSF10 enhances tumor metastasis by promoting EMT. In vitro experiments showed that knockout of TNFSF10 can reduce ESCC cell migration[28].

ESCC was divided into two subgroups according to ARGS, and there was significant discrepancy in biological function and pathways between the two groups. The ARGS-high group was mainly related to immune-related pathways, and ssGSEA analysis also confirmed this result. The infiltration of immune cells in the TME is closely related to the potential response to immunotherapy[29]. Therefore, we can speculate that there is a positive correlation between immune cell infiltration and the effect of immunotherapy. EMT plays a key role in promoting tumor metastasis, and it is also a necessary process for tumor cells to obtain anti-anoikis property[30]. The results of GSEA indicated that the ARGS-low group was mainly related to the negative regulation of cell aging and negative regulation of EMT, which may be the main reasons for the better prognosis of such patients.

Given the correlation between ARGS and immune pathways, we further predicted immunotherapy efficacy based on ARGS. It is found that the ARGS-low group has higher the TIDE score and T cell exclusion score. Higher TIDE score means a greater likelihood of immune escape of immunotherapy. Many chemotherapy drugs inhibit tumor growth by inducing apoptosis, and tumor cells will increase their drug resistance due to resistance to anoikis[31]. We analyzed IC50 of multiple chemotherapy drugs and found that the ARGS-low group was more sensitive to multiple chemotherapy drugs. These results may provide more appropriate diagnosis and treatment strategies for ESCC patients.

This study has some limitations. First of all, the transcriptome data in this study was from public databases, and its credibility still needs to be verified by real world data. In addition, the effects of key anoikis-related genes alone or in combination need to be further analyzed through in vitro and in vivo experiments.

In summary, we analyzed the characteristics of anoikis-related genes in ESCC and verified the expression of them. The signature composed of 5 key genes can well reflect the characteristics of the TME and provide proper therapeutic strategies for ESCC patients to improve their prognosis.

Authors contribution

Wuguang Chang, Hongmu Li, Chun Wu: Conceptualization, data curation, formal analysis, writing–original draft, writing–review and editing. Zenghao Chang, Leqi Zhong: data-collecting, analysis. Wei Ou: data-collecting, writing–review and editing. Siyu Wang: Conceptualization, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing. All authors reviewed and approved the manuscript.

Funding

This study was supported by the Guangdong Association Study of Thoracic Oncology (GASTO-202202).

Ethical Approval

All procedures performed in this study involving human material were in accordance with the ethical standards of the Sun Yat-sen University Cancer Center ethics committee (GZR2018-120) and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. For this type of study, formal consent is not required, as many of them were dead when the study was initiated. This article does not contain any studies with animals performed by any of the authors.

Competing interests

The authors declare no conflicts of interest in this work.

Availability of data and materials

We thank TCGA database for providing their platforms and meaningful datasets (https://portal.gdc.cancer.gov/, accessed on 8 August 2022).

Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, Bray F: Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021, 71(3):209-249.
Rustgi AK, El-Serag HB: Esophageal carcinoma. N Engl J Med 2014, 371(26):2499-2509.
Reichenbach ZW, Murray MG, Saxena R, Farkas D, Karassik EG, Klochkova A, Patel K, Tice C, Hall TM, Gang J et al: Clinical and translational advances in esophageal squamous cell carcinoma. Adv Cancer Res 2019, 144:95-135.
Kojima T, Doi T: Immunotherapy for Esophageal Squamous Cell Carcinoma. Curr Oncol Rep 2017, 19(5):33.
Baba Y, Nomoto D, Okadome K, Ishimoto T, Iwatsuki M, Miyamoto Y, Yoshida N, Baba H: Tumor immune microenvironment and immune checkpoint inhibitors in esophageal squamous cell carcinoma. Cancer Sci 2020, 111(9):3132-3141.
Frisch SM, Francis H: Disruption of epithelial cell-matrix interactions induces apoptosis. J Cell Biol 1994, 124(4):619-626.
Taddei ML, Giannoni E, Fiaschi T, Chiarugi P: Anoikis: an emerging hallmark in health and diseases. J Pathol 2012, 226(2):380-393.
Simpson CD, Anyiwe K, Schimmer AD: Anoikis resistance and tumor metastasis. Cancer Lett 2008, 272(2):177-185.
Sakamoto S, Kyprianou N: Targeting anoikis resistance in prostate cancer metastasis. Mol Aspects Med 2010, 31(2):205-214.
Shanmugathasan M, Jothy S: Apoptosis, anoikis and their relevance to the pathobiology of colon cancer. Pathol Int 2000, 50(4):273-279.
Rebhan M, Chalifa-Caspi V, Prilusky J, Lancet D: GeneCards: integrating information about genes, proteins and diseases. Trends Genet 1997, 13(4):163.
Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK: limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 2015, 43(7):e47.
Yu G, Wang LG, Han Y, He QY: clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012, 16(5):284-287.
Sanz H, Valim C, Vegas E, Oller JM, Reverter F: SVM-RFE: selection and visualization of the most relevant features through non-linear kernels. BMC Bioinformatics 2018, 19(1):432.
Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES et al: Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005, 102(43):15545-15550.
Hanzelmann S, Castelo R, Guinney J: GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics 2013, 14:7.
Sun D, Wang J, Han Y, Dong X, Ge J, Zheng R, Shi X, Wang B, Li Z, Ren P et al: TISCH: a comprehensive web resource enabling interactive single-cell transcriptome visualization of tumor microenvironment. Nucleic Acids Res 2021, 49(D1):D1420-D1430.
Jiang P, Gu S, Pan D, Fu J, Sahu A, Hu X, Li Z, Traugh N, Bu X, Li B et al: Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med 2018, 24(10):1550-1558.
Geeleher P, Cox N, Huang RS: pRRophetic: an R package for prediction of clinical chemotherapeutic response from tumor gene expression levels. PLoS One 2014, 9(9):e107468.
Balachandran VP, Gonen M, Smith JJ, DeMatteo RP: Nomograms in oncology: more than meets the eye. Lancet Oncol 2015, 16(4):e173-180.
Gilmore AP: Anoikis. Cell Death Differ 2005, 12 Suppl 2:1473-1477.
Zhong X, Rescorla FJ: Cell surface adhesion molecules and adhesion-initiated signaling: understanding of anoikis resistance mechanisms and therapeutic opportunities. Cell Signal 2012, 24(2):393-401.
Dhawan P, Singh AB, Deane NG, No Y, Shiou SR, Schmidt C, Neff J, Washington MK, Beauchamp RD: Claudin-1 regulates cellular transformation and metastatic behavior in colon cancer. J Clin Invest 2005, 115(7):1765-1776.
Hanawa M, Suzuki S, Dobashi Y, Yamane T, Kono K, Enomoto N, Ooi A: EGFR protein overexpression and gene amplification in squamous cell carcinomas of the esophagus. Int J Cancer 2006, 118(5):1173-1180.
Suntharalingam M, Winter K, Ilson D, Dicker AP, Kachnic L, Konski A, Chakravarthy AB, Anker CJ, Thakrar H, Horiba N et al: Effect of the Addition of Cetuximab to Paclitaxel, Cisplatin, and Radiation Therapy for Patients With Esophageal Cancer: The NRG Oncology RTOG 0436 Phase 3 Randomized Clinical Trial. JAMA Oncol 2017, 3(11):1520-1528.
Gutteridge RE, Ndiaye MA, Liu X, Ahmad N: Plk1 Inhibitors in Cancer Therapy: From Laboratory to Clinics. Mol Cancer Ther 2016, 15(7):1427-1435.
Ding M, Pan J, Guo Z, Liu Q, Yang C, Mao L: SATB1 is a Novel Molecular Target for Cancer Therapy. Cancer Invest 2018, 36(1):28-36.
Zhang H, Qin G, Zhang C, Yang H, Liu J, Hu H, Wu P, Liu S, Yang L, Chen X et al: TRAIL promotes epithelial-to-mesenchymal transition by inducing PD-L1 expression in esophageal squamous cell carcinomas. J Exp Clin Cancer Res 2021, 40(1):209.
Zhang Y, Zhang Z: The history and advances in cancer immunotherapy: understanding the characteristics of tumor-infiltrating immune cells and their therapeutic implications. Cell Mol Immunol 2020, 17(8):807-821.
Frisch SM, Schaller M, Cieply B: Mechanisms that link the oncogenic epithelial-mesenchymal transition to suppression of anoikis. J Cell Sci 2013, 126(Pt 1):21-29.
Pistritto G, Trisciuoglio D, Ceci C, Garufi A, D'Orazi G: Apoptosis as anticancer mechanism: function and dysfunction of its modulators and targeted therapeutic strategies. Aging (Albany NY) 2016, 8(4):603-619.

Table 1 Multivariate cox regression analysis of anoikis-related genes

Gene	Coef	p-value	HR	95.L	95.H
CLDN1	0.638104	0.008893	1.892889	1.173567	3.053109
EGFR	-1.2036	0.002751	0.300113	0.1365	0.659835
PLK1	-0.96132	0.018678	0.382389	0.171626	0.851977
SATB1	-1.56641	0.000237	0.208794	0.090573	0.481323
TNFSF10	0.999589	0.000051	2.717166	1.675271	4.407043

L: low; H: high.

No competing interests reported.

FigureS1.pdf
Figure S1 | Prognostic value of ARGS. Distribution of risk score (A) and survival status (B) in TCGA-ESCC cohort. (C) Heatmap displaying five anoikis-related genes.
FigureS2.pdf
Figure S2 | Kaplan-Meier survival analysis in training (A) and test (B) dataset.
FigureS3.pdf
Figure S3 | The expression of five key anoikis-related genes in single cell dataset (GSE160269)
TableS1.xlsx
Table S1: Anoikis-related genes.
TableS2.xlsx
Table S2: Univariate Cox regression analysis of differentially expressed anoikis-related genes.

Download PDF

Version 1

posted

You are reading this latest preprint version

Identification and validation of an anoikis-related gene signature based on machine learning algorithms in esophageal squamous cell carcinoma

Status:

Version 1

Abstract

Purpose

Methods

Results

Conclusion

Figures

1. Introduction

2. Material And Methods

2.1 Data collection

2.2 Identification and enrichment analysis of differentially expressed genes

2.3 Construction of the anoikis-related gene signature

2.4 Gene set enrichment analysis

2.5 Evaluation of tumor microenvironment

2.6 Prediction of immunotherapy response

2.7 Drug sensitivity analysis

2.8 Quantitative Real-time PCR

2.9 Statistical analysis

3. Results

3.1 Identification and enrichment analysis of differentially expressed anoikis-related genes

3.2 Construction of anoikis-related gene signature

3.3 Verification of key anoikis-related genes expression

3.4 Analysis of biological function and pathways

3.5 Differences of tumor microenvironment

3.6 Prediction of immunotherapy response

3.7 Prediction of chemotherapy drug sensitivity

4. Discussion

Conclusion

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1