A cancer stemness-related eight-gene signature for predicting prognosis and malignant degree of lung adenocarcinoma

doi:10.21203/rs.3.rs-1442582/v1

Download PDF

Research Article

A cancer stemness-related eight-gene signature for predicting prognosis and malignant degree of lung adenocarcinoma

https://doi.org/10.21203/rs.3.rs-1442582/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

To identify a cancer stemness-related prognostic signature, we analyzed RNA-seq data of patients with lung adenocarcinoma from the TCGA and GEO cohort using univariate Cox and LASSO analysis. Eight genes (PLEK2, GPX8, TCN1, PERP, CD79A, PAX5, FSCN1, and CNNM1) were eventually identified to develop a cancer stemness-related signature. AUCs for 5-year overall survival were 0.681, 0.673, and 0.774 in the training, validation, and test sets, respectively. Kaplan-Meier curve analysis showed that the high-risk group has a poor survival as compared with the low-risk group in the training set, validation set, and test set, respectively (P < 0.05). Six genes (PLEK2, GPX8, TCN1, PERP, FSCN1, and CNNM1) were significantly upregulated in the cancerous tissue compared with in the normal tissue (P < 0.001), and correlated with an unwanted prognosis (P < 0.05). We constructed a prognostic nomogram combining the cancer stemness-related signature with TNM staging to precisely assess the prognosis of the patient with lung adenocarcinoma. Finally, we observed that the stemness-related gene signature also could predict the malignant levels of lung adenocarcinoma possibly via hypoxia, epithelial-mesenchymal transition and angiogenesis. Conclusively, this study developed a novel cancer stemness-related gene signature, and identified six pivotal genes which can serve as potential therapeutic targets in the treatment of lung cancer.

cancer stemness

lung adenocarcinoma

prognostic signature

CNNM1

GPX8

Pulmonary carcinoma is the leading cause of cancer-related death worldwide. Lung adenocarcinoma (LUAD) accounts for approximately half of all types of lung cancer and has been increasing annually [1]. Despite recent advances in the diagnosis and treatment of LUAD, the 5-year survival rate for patients with LUAD is still relatively low, less than 20% [2]. Moreover, there is a lack of satisfactory prognostic markers for predicting survival in lung adenocarcinoma. Therefore, the identification of a comparatively reliable and applicable prognostic signature for LUAD in order to guide clinical decision-making is required.

Cancer stemness is primarily constituted of chemo-resistance, radio-resistance, tumorigenesis, tumor self-renewal, cancer metabolism reprogramming, and tumor immune microenvironment remodeling [3-5], and has an important role in the treatment and prognosis of cancer. Multiple signaling pathways are reported to be involved in cancer stemness, including Wnt/β-catenin signaling pathway. Wnt/β-catenin signaling is a key driver of cancer stem, and contribute to the development and progression of cancers, such as colorectal cancer[6], hepatocellular carcinoma [7], and pancreas cancer [8]. MAPK signaling is also reported to be related to cancer stemness [9, 10]. The role of hippo signaling and JAK/STAT3 signaling in cancer stemness have already been demonstrated [11-13]. While cancer stemness has been one hotpot of cancer research field in recent years, the molecular mechanisms underlying its function are incompletely understood. Moreover, it is unknown whether cancer stemness-related genes could serve as an independent prognostic factor for patients with LUAD.

To identify a cancer stemness-related prognostic signature and develop a new nomogram predictive model for patients with LUAD, we analyzed the RNA sequencing (RNA-seq) data and the corresponding clinical data from the TCGA and GEO databases using bioinformatic tools. The findings would provide a useful prognostic gene signature and a practical nomogram for predicting survival for patients with LUAD. Researchers, clinicians and patients would handily forecast the survival probability for each individual patient using this nomogram.

Public datasets

RNA-seq data and survival data for the TCGA cohort of patients with LUAD were downloaded from UCSC Xena on Sep 9, 2021 (https://xena.ucsc.edu/), including gene expression profile and the corresponding clinical data for 585 samples including 526 lung adenocarcinoma samples and 59 normal lung samples. The microarray expression profile and the corresponding survival data of 138 lung cancer samples was obtained from the GSE8894 cohort in the Gene Expression Omnibus (GEO). The GSE8894 dataset was based on the platform of GPL570 (Affymetrix Human Genome U133 Plus 2.0 Array).

Identification of differentially expressed genes

Differentially expressed genes (DEGs) were identified with original read counts using R package "edgeR" [14]. Note that "edgeR" is designed to work with actual read counts. Normalized data, including FPKM, RPKM and TMP, is not recommended to be used in place of actual counts in edgeR. R package "edgeR" implements novel statistical methods based on the negative binomial distribution as a model for count variability, including empirical Bayes methods, exact tests, and generalized linear models, and quasi-likelihood tests to calculate DEGs.

Prior to further analysis, "edgeR" provides a procedure to filter out low expressed genes and normalize the raw data into TMM data which is subsequently used to calculate DEGs. The R function of filterByExpr is used to filter out low expressed genes. The filterByExpr function keeps rows that have worthwhile counts in a minimum number of samples.

Here, DEGs between 57 cancerous tissue and 57 matched normal tissue from the TCGA cohort of patients with LUAD were calculated using a paired design, which can be performed using a formula of "design = model.matrix(~ patient + group)" in the R package “edgeR”. Selection criteria for DEGs were as follows: |logFC| > 2 and FDR < 0.001. Benjamini-Hochberg method was used to adjust P values.

Single-sample gene set enrichment analysis

To identify the cancer stemness-associated genes in LUAD, we first calculated the enrichment scores of the gene set of cancer stemness for 526 tumorous samples from the TCGA cohort of LUAD patients using single-sample gene set enrichment analysis (ssGSEA) based on R package “GSVA” [15]. ssGSEA is a non-parametric, unsupervised method to calculate variation of gene set enrichment through the samples from an expression dataset. Each ssGSEA enrichment score represents the degree to which the genes in a particular gene set are coordinately up- or down-regulated within a sample. The key parameters were as follows: kcdf = “Gaussian”, min.sz = 1, max.sz = Inf, tau = 0.25, abs.ranking = TRUE.

Establishment of the cancer stemness-associated gene signature

Among 526 LUAD patients from the TCGA cohort, there are 500 LUAD patients with a completed survival data. These 500 patients were randomly divided into the training set (n = 400) and the validation set (n = 100).

The cancer stemness-associated genes were first analyzed using the univariate Cox regression analysis in the training set. The variables with a P < 0.05 in the univariate Cox regression analysis were then used as input in the least absolute shrinkage and selection operator (LASSO) regression analysis in the training set. Finally, genes with a P < 0.05 in LASSO analysis were used for developing a prognostic cancer stemness-associated gene signature based on the multivariate Cox regression mode.

Validation of the cancer stemness-associated gene signature

The prognostic gene signature was used to calculate the risk score of each patient with LUAD based on multivariate Cox regression model. The predictive performance of the prognostic gene signature was assessed using area under the curve (AUC), C-index, Kaplan-Meier curve in the training set, the validation set, and an independent test set (GSE8894).

Functional annotation

Functional annotation of DEGs was conducted using the R package “clusterProfiler” (version: 3.18.1) [16], which provides a comprehensive set of functional annotation tools for researchers to comprehend the biological meaning behind specific gene sets. The clusterProfiler package depends on the Bioconductor annotation data GO.db and KEGG.db to obtain the maps of the entire gene oncology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) corpus. Bioconductor annotation packages org.Hs.eg.db, org.Mm.eg.db, and org.Sc.sgd.db were imported for genome-wide annotation of mapping Entrez gene identifiers or ORF identifiers for humans, mice, and yeast, respectively. Functional annotation consists of the GO analysis and KEGG pathways, which allows one can investigate what biological functions and signaling pathways a given gene set is involved in. We also conducted gene set enrichment analysis (GSEA) based on a ranked gene set using “clusterProfiler”. The GSEA analysis could reveal some enriched signaling pathways missed in the GO and KEGG analysis. Key parameters were as follows: pAdjustMethod = 'BH', pvalueCutoff = 0.05, qvalueCutoff = 0.2, nPerm = 1000, minGSSize = 10, maxGSSize = 500.

Development and validation of a nomogram for precisely predicting prognosis

To investigate the prognostic significance of the cancer stemness-related gene signature in combination with common clinical characteristics, we developed a predictive nomogram combining the prognostic gene signature and clinical characteristics for LUAD patients. The nomogram can be used to precisely estimate the survival risk of each LUAD patient and facilitate clinical decision-making in the treatment of LUAD.

Statistics

All statistical analyses were completed using R software (Version 4.0.1). Based on the data homogeneity of variance and normal distribution, either the independent sample t test or Wilcoxon signed rank test was chosen. The log-rank test and Kaplan-Meier curve were used to evaluate survival significance. Spearman’s correlation coefficient was used to assess the correlation between two continuous variables. P < 0.05 was considered statistically significant.

Construction of the cancer stemness-related gene signature

First, to identify the cancer stemness-related genes in LUAD, we calculated the enrichment scores of the gene set of cancer stemness for 526 tumorous samples from the TCGA cohort of LUAD patients using the ssGSEA algorithm, and obtained 5253 cancer stemness-related genes (P < 0.05, R > 0; Figure 1 A).

Next, we performed a paired differential expression analysis using R package “edgeR” based on RNA-seq data of 57 tumorous samples and 57 matched normal tissues we previously selected from the TCGA cohort of LUAD patients; 3393 DEGs were obtained, including 1224 upregulated DEGs and 2169 downregulated DEGs (|logFC| > 1, FDR < 0.05; Figure 1 B). We also investigated the prognostic value of all genes using RNA-seq data and survival data for the TCGA cohort of LUAD patients, and obtained 4237 relevant genes (P < 0.05; log-rank test). Finally, we intersected 5253 cancer stemness-related genes, 1224 upregulated DEGs and 4237 prognostically relevant genes, and obtained 47 overlapping genes (Figure 1 C), which were defined as qualified cancer stemness-related genes for subsequent construction of the stemness-related gene signature.

The 47 qualified cancer stemness-related genes were analyzed as the input of the univariate Cox regression analysis for the training set. Then, 35 eligible genes in the univariate Cox regression analysis (P < 0.05) were further screened using LASSO regression analysis. Eventually, eight genes (PLEK2, GPX8, TCN1, PERP, CD79A, PAX5, FSCN1, and CNNM1) were selected to develop a stemness-related gene signature using the multivariate Cox regression model (Figure 1 D and 1 E).

Evaluation of predictive capability of the cancer stemness-related gene signature

The predictive performance of the cancer stemness-related gene signature was assessed in the training set (n = 400), the validation set (n = 100) and an external test set (GSE8894; n = 61). AUCs for 5-year overall survival were 0.681, 0.673 and 0.774 in the training, validation and test sets, respectively (Figures 2 A – 2 C), suggesting its robust predictive ability for predicting prognosis.

Next, we calculated risk score of each LUAD patient from the TCGA cohort based on RNA-seq data of the eight genes using multivariate Cox regression model, and classified all patients into the low-risk and high-risk groups based on the median of risk scores of all patients. Principal component analysis of genes consisting the cancer stemness-related signature revealed a distinct expression pattern between the low-risk and the high-risk groups in dimensionality 1 (Figure 2 D – 2 F), indicating the cancer stemness-related gene signature has a good distinguishing ability.

To demonstrate the relationship between the risk score and the expression of eight genes comprising the cancer stemness-related signature and the survival of patients with LUAD, we displayed the risk score, survival status, and heatmap for gene expression in a picture with the same abscissa. The findings showed that the high-risk scores were associated with poor survival outcomes and high expression levels of six genes (PLEK2, GPX8, TCN1, PERP, FSCN1, and CNNM1) in the training set (Figure 2 G), the validation set (Figure 2 H) and the external test set (GSE8894; Figure 2 I).

Clinical significance of the cancer stemness-related gene signature

Since the gene signature manifested a robust predictive ability, we next investigated its survival value in the training set, the validation set, and the external test set. As previously described, we calculated risk score of each LUAD patient from the TCGA cohort based on RNA-seq data of the eight genes using multivariate Cox regression model, and classified all patients into the low-risk and high-risk groups based on the median of risk scores of all patients. In line with the results generated from the previous step, the low-risk group had a better survival outcome as compared with the high-risk group in the training set, the validation set, and the external test set (P < 0.001; Figure 3 A – 3 C).

To further investigate the association between the risk score and the TNM classification of LUAD, we compared the risk score across different TNM stages based on the TCGA cohort. Consistent with the findings of the survival analysis, we found that the risk score was significantly augmented in patients with higher TNM staging (Figure 3 D – 3 F), indicating a significant impact of the cancer stemness-related gene signature on TNM staging. Moreover, we investigated the association between the risk score and the immune score in LUAD. The immune score was computed using R package “estimate” based on RNA-seq data of patients with LUAD. The results showed that risk score was negatively correlated with immune score (P = 0.0005, R = -0.1552; Figure 3 G).

To explore the biological processes associated with the cancer stemness-related gene signature, we performed functional annotation of genes that were correlated with the cancer stemness -related gene signature. We first investigated the relationship of the gene signature with all genes using RNA-seq data from the TCGA cohort of patients with LUAD. 332 genes were selected as the gene signature-related genes (P < 0.01, R > 0.3). Afterwards, these 553 genes were functional annotated using R package “clusterProfiler”. The findings showed that several biological processes related to cancer invasion and metastasis were significantly enriched, including cell adhesion molecules (CAMs), extracellular structure organization, focal adhesion (Figure 3 H – 3 I)

Identification of the hub genes related with cancer stemness

Considering the previously established eight-gene signature not only can reflect overall survival, but also was associated with TNM staging and immune score, we speculated that the eight genes consisting of the gene signature were potential critical genes in the development and progression of LUAD. To identify potential hub genes, we investigated their survival significance in two independent cohorts (TCGA and GSE8894). The results demonstrated that six genes were associated with poor survival (PLEK2, GPX8, TCN1, PERP, FSCN1, and CNNM1; log-rank test; P < 0.05; Figure 4 B – 4 D and Figure 4 F – 4 H; Figure S1), while two genes were linked to better survival (CD79A and PAX5; log-rank test; P < 0.05; Figure 4 A and Figure 4 E; Figure S1).

To evaluate the expression levels of these eight genes in LUAD, we compared their expression levels between LUAD tissue and normal lung tissue using RNA-seq data from the TCGA cohort. As expected, the results showed that the expression levels of the eight genes were significantly enhanced in tumor tissue compared with normal tissue (P < 0.01; Figure 4 I – 4 L).

Establishment of a prognostic nomogram for precision medicine

Considering the prognostic significance of the eight-gene signature, we sought to combine it with several common clinical characteristics to better predict survival of LUAD patients. We first conducted a multivariate Cox regression analysis to examine the prognostic significance of the risk score based on the gene signature, age, immune score, pathological T, pathological N, and pathological M. The results showed that the risk score, pathological T, and pathological N could be used as effective prognostic characteristics for LUAD (P < 0.01; Figure 5 A). We noticed that pathological M showed no significant prognostic value (P = 0.172), which seems to be contrary to the current knowledge that patients with pathological M have a poor survival. Thus, we further analyzed the clinical data and found that there were 229 patients with M0 while 14 patients with M1. In other words, the sample size between the two groups was extremely imbalanced, which possibly affected the accuracy of the statistical results.

Afterwards, we constructed a nomogram combining the risk score, pathological N, and pathological T (Figure 5 B), which can be used to accurately predict the patients’ survival risk for precision medicine.

Impact of the stemness-related gene signature on cancer behaviors

Considering the robust predictive ability of the stemness-related gene signature, we wondered whether this gene signature could reflect the malignant degree of lung cancer. To this end we estimated the enrichment score of proliferation, invasion and metastasis of lung adenocarcinoma using ssGSEA algorithm based on RNA-seq data of patients with lung adenocarcinoma from the TCGA cohort. Then we analyzed the relationship of the enrichment score of proliferation, invasion and metastasis with the risk scores based on the stemness-related gene signature; the findings showed that the risk scores were significantly correlated with these malignant tumor behaviors including proliferation, invasion and metastasis (P < 0.05, Figure 6 A – 6 C).

To investigate the underlying mechanisms of the effects of the stemness-related gene signature on cancer behaviors, we further explored the relationship between the risk score and hypoxia, epithelial-mesenchymal transition (EMT) and angiogenesis, which are all the potential causes of tumor metastasis. As expected, the risk scores were positively correlated with the levels of hypoxia, EMT and angiogenesis, strongly suggesting that the stemness-related gene signature could predict the malignant degree of lung cancer (P < 0.05, Figure 6 D – 6 F).

We developed and validated an eight-gene signature for predicting overall survival in the patients with LUAD, constructed a nomogram combining a cancer stemness-related gene signature and TNM staging for precision medicine. The cancer stemness-related gene signature was significantly associated with TNM staging and negatively correlated with immune response. Several biological processes related to cancer invasion and metastasis were significantly enriched, including cell adhesion molecules (CAMs), extracellular structure organization, focal adhesion. These findings provided a robust predictor for predicting prognosis of patients with LUAD, and identified several potential therapeutic targets for the treatment of LUAD.

One of the main contributions of this study is the establishment of a cancer stemness-related prognostic signature comprising eight genes (PLEK2, GPX8, TCN1, PERP, FSCN1, and CNNM1), for predicting survival of patients with LUAD. The robustness of the model has been verified in an external data set (the GSE8894 cohort), which makes the gene signature highly reliable and widely applicable. The predictive power of this gene signature is superior to other known gene signatures. Although the performance of the eight-gene signature was not statistically compared with other similar gene signatures for LUAD, we searched the PubMed database and observed that the predictive power of this cancer stemness-related gene signature is better than other similar markers [17-20].

We also revealed that six cancer stemness-related oncogenes (PLEK2, GPX8, TCN1, PERP, FSCN1, and CNNM1), which could be used as potential therapeutic targets. We revealed that these six genes were overexpressed in cancerous tissue compared with norm lung tissue and linked to a worse prognosis in patients with LUAD. The protein encoded by PLEK2 is pleckstrin-2, which is implicated in actin rearrangement in lymphocyte spreading and immune synapse formation [21, 22]. Moreover, PLEK2 was identified as the strongest gene to classify melanoma patients[23], and upregulated during TGF-β induced epithelial-mesenchymal transition and related to poor overall survival in lung adenocarcinoma patients [24]. Loss of pleckstrin-2 reverts lethality and vascular occlusions in JAK2V617F-positive myeloproliferative neoplasms [25]. PLEK2 is also involved in cancer invasion and metastasis in gallbladder cancer [26], pancreatic cancer [27], non-small cell lung cancer [28]. Consistent with our findings, GPX8 is also identified as a prognostic biomarker of gastric cancer [29], and is essential for tumor aggressiveness in breast cancer [30], non-small cell lung cancer [31], and gastric cancer [32]. FSCN1 was implicated in nasopharyngeal carcinoma [33] and non-small cell lung cancer [34]. Although CNNM1 has been suggested to be related to prostate cancer [35] and hepatocellular carcinoma [36], its association with lung cancer is unknown. Here, we revealed its overexpression and correlation with poor survival outcomes in lung adenocarcinoma, pointing to the importance of further investigating its role in the pathogenesis of lung cancer.

Another important finding in this study is that several important biological processes, which are involved in cancer stemness and poor prognosis, were identified, including cell adhesion molecules (CAMs), extracellular structure organization, and focal adhesion. These functions are all involved in extracellular matrix (ECM), which has been found to cause regrowth and healing of tissue and the malignancy and growth of the tumor [37]. In many tumors, the ECM compromises up to 60% of the tumor mass. Source of these ECM molecules are the tumor cells themselves, but to an even larger degree cancer-associated fibroblasts (CAFs) [38, 39]. Like other components of tumor microenvironment, ECM diverges significantly between tumor and normal tissue. Intratumoral signaling pathways, metabolisms, oxygenation, and immunogenicity are critically influenced by ECM. Exerting this regulatory control, the ECM does not only influence malignancy and growth of the tumor but also its response toward therapy [37]. Targeting the tumor extracellular matrix could be an alternative cancer therapy strategy [40, 41].

This study has important implications for the treatment as well as prognosis of melanoma. First, our study provides a novel prognostic signature and a new nomogram that could aid clinical treatment strategies for LUAD. Second, we revealed several critical oncogenes and pathways that could serve as promising therapeutic targets for the treatment of LUAD. Nevertheless, further in vitro and in vivo investigations are warranted to study the role of these pivotal genes in LUAD and their precise mechanisms of action.

This study has a few limitations that warrant further research. First, as the four key genes exhibit critical significance in LUAD prognoses, further in vitro and in vivo studies are required to explore their physiological mechanisms of actions. In addition, cell cycle, DNA replication, chromosome segregation, and nuclear division are indicated to therapeutic targets for the treatment of LUAD, warranting further investigation regarding the precise underlying mechanism.

In conclusion, we successfully constructed and validated a novel cancer stemness-related signature and a nomogram precisely predicting survival of patients with LUAD, by integrative analysis of RNA-seq from TCGA and GEO cohorts using bioinformatic tools. Our study also revealed the important function of several critical genes and ECM in LUAD that could serve as potential therapeutic targets. These findings provide the rationale for further investigation and would aid clinical decision-making in LUAD immunotherapy.

PLEK2, Pleckstrin-2; GPX8, Glutathione peroxidase 8; TCN1, Transcobalamin 1; PERP, P53 apoptosis effector related to PMP22; CD79A, CD79a molecule; PAX5, Paired box 5; FSCN1, Fascin actin-bundling protein 1; CNNM1, Cyclin and CBS domain divalent metal cation transport mediator 1.

Ethics approval and consent to participate

There is no ethics in this article.

Consent to publish

I agree to be published in this magazine.

Availability of data and materials

All data generated and described in this article are available from the corresponding web servers, and are freely available to any scientist wishing to use them for noncommercial purposes, without breaching participant confidentiality. Further information is available from the corresponding author on reasonable request.

Competing interests

The authors declare that they have no competing interests.

Authors’ Contributions

Yuanjun Cheng, Jie Yao and Qianru Fang contributed equally to this work. Chen Bin and Guohui Zang collected the data. All authors reviewed and approved the final manuscript.

Acknowledgments

This research was supported by grants from Key research and Development Program of Anhui Province in 2021 (202104j07020055).

Thakur, S.K., D.P. Singh, and J. Choudhary, Lung cancer identification: a review on detection and classification. Cancer Metastasis Rev, 2020. 39(3): p. 989-998.
Siegel, R.L., et al., Cancer Statistics, 2021. CA Cancer J Clin, 2021. 71(1): p. 7-33.
Zhou, H.M., et al., Targeting cancer stem cells for reversing therapy resistance: mechanism, signaling, and prospective agents. Signal Transduct Target Ther, 2021. 6(1): p. 62.
Ayob, A.Z. and T.S. Ramasamy, Cancer stem cells as key drivers of tumour progression. J Biomed Sci, 2018. 25(1): p. 20.
Huang, T., et al., Stem cell programs in cancer initiation, progression, and therapy resistance. Theranostics, 2020. 10(19): p. 8721-8743.
Tang, Q., et al., TM4SF1 promotes EMT and cancer stemness via the Wnt/β-catenin/SOX2 pathway in colorectal cancer. J Exp Clin Cancer Res, 2020. 39(1): p. 232.
Wei, S., et al., KIF2C: a novel link between Wnt/β-catenin and mTORC1 signaling in the pathogenesis of hepatocellular carcinoma. Protein Cell, 2021. 12(10): p. 788-809.
Nusse, R. and H. Clevers, Wnt/β-Catenin Signaling, Disease, and Emerging Therapeutic Modalities. Cell, 2017. 169(6): p. 985-999.
He, F., et al., The protective role of microRNA-21 against coxsackievirus B3 infection through targeting the MAP2K3/P38 MAPK signaling pathway. J Transl Med, 2019. 17(1): p. 335.
He, J., et al., Circular RNA MAPK4 (circ-MAPK4) inhibits cell apoptosis via MAPK signaling pathway by sponging miR-125a-3p in gliomas. Mol Cancer, 2020. 19(1): p. 17.
Ni, W., et al., Long noncoding RNA GAS5 inhibits progression of colorectal cancer by interacting with and triggering YAP phosphorylation and degradation and is negatively regulated by the m(6)A reader YTHDF3. Mol Cancer, 2019. 18(1): p. 143.
Liang, C., et al., Long Non-coding RNA ITIH4-AS1 Accelerates the Proliferation and Metastasis of Colorectal Cancer by Activating JAK/STAT3 Signaling. Mol Ther Nucleic Acids, 2019. 18: p. 183-193.
Wu, R., et al., m(6)A methylation controls pluripotency of porcine induced pluripotent stem cells by targeting SOCS3/JAK2/STAT3 pathway in a YTHDF1/YTHDF2-orchestrated manner. Cell Death Dis, 2019. 10(3): p. 171.
McCarthy, D.J., Y. Chen, and G.K. Smyth, Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation. Nucleic Acids Res, 2012. 40(10): p. 4288-97.
Hänzelmann, S., R. Castelo, and J. Guinney, GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics, 2013. 14: p. 7.
Yu, G., et al., clusterProfiler: an R package for comparing biological themes among gene clusters. Omics, 2012. 16(5): p. 284-7.
Zhang, L., et al., Identification of a panel of mitotic spindle-related genes as a signature predicting survival in lung adenocarcinoma. J Cell Physiol, 2020. 235(5): p. 4361-4375.
Zhang, Y.Q., et al., Evaluation of the roles and regulatory mechanisms of PD-1 target molecules in NSCLC progression. Ann Transl Med, 2021. 9(14): p. 1168.
Jin, D., et al., Identification of Three lncRNAs as Potential Predictive Biomarkers of Lung Adenocarcinoma. Biomed Res Int, 2020. 2020: p. 7573689.
He, R. and S. Zuo, A Robust 8-Gene Prognostic Signature for Early-Stage Non-small Cell Lung Cancer. Front Oncol, 2019. 9: p. 693.
Bach, T.L., et al., PI3K regulates pleckstrin-2 in T-cell cytoskeletal reorganization. Blood, 2007. 109(3): p. 1147-55.
Hamaguchi, N., et al., Pleckstrin-2 selectively interacts with phosphatidylinositol 3-kinase lipid products and regulates actin organization and cell spreading. Biochem Biophys Res Commun, 2007. 361(2): p. 270-5.
Luo, Y., et al., Transcriptome profiling of whole blood cells identifies PLEK2 and C1QB in human melanoma. PLoS One, 2011. 6(6): p. e20971.
Yin, H., et al., Drug-resistant CXCR4-positive cells have the molecular characteristics of EMT in NSCLC. Gene, 2016. 594(1): p. 23-29.
Zhao, B., et al., Loss of pleckstrin-2 reverts lethality and vascular occlusions in JAK2V617F-positive myeloproliferative neoplasms. J Clin Invest, 2018. 128(1): p. 125-140.
Shen, H., et al., PLEK2 promotes gallbladder cancer invasion and metastasis through EGFR/CCL2 pathway. J Exp Clin Cancer Res, 2019. 38(1): p. 247.
Dahlem, C., et al., The Insulin-Like Growth Factor 2 mRNA Binding Protein IMP2/IGF2BP2 is Overexpressed and Correlates with Poor Survival in Pancreatic Cancer. Int J Mol Sci, 2019. 20(13).
Wu, D.M., et al., PLEK2 mediates metastasis and vascular invasion via the ubiquitin-dependent degradation of SHIP2 in non-small cell lung cancer. Int J Cancer, 2020. 146(9): p. 2563-2575.
Zhang, X., et al., Glutathione Peroxidase 8 as a Prognostic Biomarker of Gastric Cancer: An Analysis of The Cancer Genome Atlas (TCGA) Data. Med Sci Monit, 2020. 26: p. e921775.
Khatib, A., et al., The glutathione peroxidase 8 (GPX8)/IL-6/STAT3 axis is essential in maintaining an aggressive breast cancer phenotype. Proc Natl Acad Sci U S A, 2020. 117(35): p. 21420-21431.
Zhang, J., et al., GPX8 promotes migration and invasion by regulating epithelial characteristics in non-small cell lung cancer. Thorac Cancer, 2020. 11(11): p. 3299-3308.
Chen, H., et al., GPX8 is transcriptionally regulated by FOXC1 and promotes the growth of gastric cancer cells through activating the Wnt signaling pathway. Cancer Cell Int, 2020. 20(1): p. 596.
Li, Y.Q., et al., MiR-24 functions as a tumor suppressor in nasopharyngeal carcinoma through targeting FSCN1. J Exp Clin Cancer Res, 2015. 34: p. 130.
Liang, Z., et al., Fascin 1 promoted the growth and migration of non-small cell lung cancer cells by activating YAP/TEAD signaling. Tumour Biol, 2016. 37(8): p. 10909-15.
Huang, Y., et al., Ginsenoside Rh2 Inhibits Angiogenesis in Prostate Cancer by Targeting CNNM1. J Nanosci Nanotechnol, 2019. 19(4): p. 1942-1950.
Xie, Y., et al., SNHG7 Facilitates Hepatocellular Carcinoma Occurrence by Sequestering miR-9-5p to Upregulate CNNM1 Expression. Cancer Biother Radiopharm, 2020. 35(10): p. 731-740.
Henke, E., R. Nandigama, and S. Ergün, Extracellular Matrix in the Tumor Microenvironment and Its Impact on Cancer Therapy. Front Mol Biosci, 2019. 6: p. 160.
Castermans, K. and A.W. Griffioen, Tumor blood vessels, a difficult hurdle for infiltrating leukocytes. Biochim Biophys Acta, 2007. 1776(2): p. 160-74.
Naba, A., et al., The matrisome: in silico definition and in vivo characterization by proteomics of normal and tumor extracellular matrices. Mol Cell Proteomics, 2012. 11(4): p. M111.014647.
Vitale, D.L., et al., Targeting the Tumor Extracellular Matrix by the Natural Molecule 4-Methylumbelliferone: A Complementary and Alternative Cancer Therapeutic Strategy. Front Oncol, 2021. 11: p. 710061.
Belhabib, I., et al., Extracellular Matrices and Cancer-Associated Fibroblasts: Targets for Cancer Diagnosis and Therapy? Cancers (Basel), 2021. 13(14).

No competing interests reported.

FigureS1.tif

Download PDF

Version 1

posted

You are reading this latest preprint version

A cancer stemness-related eight-gene signature for predicting prognosis and malignant degree of lung adenocarcinoma

Status:

Version 1

Abstract

Figures

Introduction

Materials and methods

Public datasets

Identification of differentially expressed genes

Single-sample gene set enrichment analysis

Establishment of the cancer stemness-associated gene signature

Validation of the cancer stemness-associated gene signature

Functional annotation

Development and validation of a nomogram for precisely predicting prognosis

Statistics

Results

Construction of the cancer stemness-related gene signature

Evaluation of predictive capability of the cancer stemness-related gene signature

Clinical significance of the cancer stemness-related gene signature

Identification of the hub genes related with cancer stemness

Establishment of a prognostic nomogram for precision medicine

Impact of the stemness-related gene signature on cancer behaviors

Discussion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1