Identification of a Glycolysis-Related Gene Signature for Survival Prediction of Ovarian Cancer Patients

doi:10.21203/rs.3.rs-133263/v1

Download PDF

Primary research

Identification of a Glycolysis-Related Gene Signature for Survival Prediction of Ovarian Cancer Patients

https://doi.org/10.21203/rs.3.rs-133263/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 05 Oct, 2021

Read the published version in Cancer Medicine →

Version 1

posted

You are reading this latest preprint version

Background: Ovarian cancer (OV) is deemed as the most lethal gynecological cancer in women. The aim of this study was construct an effective gene prognostic model for OV patients.

Methods: The expression profiles of glycolysis-related genes (GRGs) and clinical data of patients with OV were extracted from The Cancer Genome Atlas (TCGA) database. Univariate, multivariate, and least absolute shrinkage and selection operator Cox regression analyses were conducted, and a prognostic signature based on GRGs was constructed. The predictive ability of the signature was analyzed in training and test sets.

Results: Based on nine GRGs (ISG20, CITED2, PYGB, IRS2, ANGPTL4, TGFBI, LHX9, PC, and DDIT4), a gene risk signature was identified to predict the outcome of patients with OV. The signature showed a good prognostic ability for OV, particularly high-grade OV, in the TCGA dataset, with areas under the curve of 0.709, 0.762, and 0.808 for 3-, 5- and 10-year survival, respectively. Similar results were found in the test sets, and the signature was also an independent prognostic factor. Moreover, a nomogram combining the prediction model and clinical factors was constructed.

Conclusion: Our study established a nine-GRG risk model and a nomogram to better perform on OV patients’ survival prediction. The risk model represents a promising and independent prognostic predictor for OV patients. Moreover, our study of GRGs could offer guidance for underlying mechanisms explorations in the future.

Cancer Biology

Bioinformatics

Ovarian cancer

Glycolysis

Prognostic signature

Among gynecological cancers, ovarian cancer (OV) is considered the most fatal. And it is estimated that the respective numbers of new cases and new deaths were 22,530 and 13,980 in the United States in 2019.[1] That the main challenges in effective methods for screening and predicting prognosis are attributed to the significant heterogeneity at clinical, histopathological and molecular levels of this disease.[2] And clinical and pathologic factors are not sufficient to predict long-term survival.[3] More and more opportunities for exploring tumor prognostic markers emerged, and that attributed to the establishment and development of public biological databases, which provides available gene expression data and clinical data of cancers. Many biomarkers, including the EN2 and HE4 genes, which are associated with the prognosis and survival of OV, have been identified currently.[4–7] With the rapid development of high-throughput sequencing, a variety of patient genome databases have been constructed so that researches can get more systematic understanding of genomic changes. By mining the database, thousands of prognostic biomarkers have been identified. [8, 9] Additionally, studies found that the genetic model constructed by multiple genes has better prediction performance for cancer prognosis than the single gene.[9, 10] Gene model constructed based on tumor biopsy has practical significance for the guidance of targeted therapy. Currently, several studies have explored to do establish multigene signatures for assessing OV patients’ survival risk and predicting clinical outcomes.[8–11]

Glycolysis takes place in all cells of the body.[12] A previous study reported that genes involved in glycolysis are ubiquitously overexpressed in 24 cancer classes.[13] So far, the relationships between glycolysis and the processes of cancers’ oncogenesis, development, proliferation and invasion have been the focus of many studies.[14–16] The results from previous studies provide compelling evidence of new glycolysis-related biomarkers for the prediction of cancer patients’ survival. Poor prognosis occurs in pancreatic cancer patients with high expression of TCF7L2 compared to those with low expression levels, and the involved mechanism behind it was that the positive regulation of TCF7L2 on aerobic glycolysis through the axis of EGLN2/HIF-1α.[17] Four glycolysis-related genes (GRGs; AGRN, AKR1A1, DDIT4, and HMMR) were also identified in a previous study and are strongly associated with clinical outcome of lung adenocarcinoma patients with lung adenocarcinoma.[18] The combination model of nine GRGs were identified to effectively predict overall survival (OS) of endometrial cancer.[19] In addition, a glycolytic gene expression signature score, established on the basis of 10 glycolytic genes (HK2, HK3, LDHA, PKM2, GAPDH, ENO1, LDHB, PKLR, ALDOB, and GALM), predicts unfavorable clinical outcomes of patients with glioblastoma and is closely associated with the mesenchymal subtype.[14] However, no signature based on GRGs has been established to predict survival in patients with OV.

In this study, we aimed to investigate specific GRG markers that had close associations with the survival of patients with OV by using data from The Cancer Genome Atlas (TCGA; https://portal.gdc.cancer.gov/) database and to evaluate the prognostic value of the biomarkers for the prediction of patient survival in OV. An effective 9-GRG risk predictive model was constructed for predicting outcomes of OV patients. Noteworthily, the GRG risk model enabled identification of patients with poor prognoses in the high-risk group. The results acquired from the multivariate Cox regression analyses implied that our risk model was a predictor independent of clinical factors, and did better in the prediction of OV patients’ survival.

2.1. Data collection

We extracted clinical and RNA sequencing data of OV patients from TCGA (https://portal.gdc.cancer.gov/). The exclusion criteria were set as follows: (1) confirmed non-OV pathological diagnosis; and (2) OV patients with incomplete information regarding clinical characteristics (age, tumor stage, and histological grade, survival time and status). Finally, the covered clinical information of 583 patients was complete integrated. Other datasets (GSE63885, GSE26193, and GSE30161) from the Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/) were selected as independent validation sets for validation of the robustness of the eight-DRG prognostic model. The GRG sets were obtained from the Molecular Signatures Database (MSigDB).[20] The gene list was obtained by synthesizing the genes of five related gene sets after excluding duplicate genes.

2.2. Construction and evaluation of the nine-GRG prediction model

We applied log2 transformation to standardize the expression of each gene. We used p value of < 0.05 as the screening criterion, and performed univariate Cox analysis, the least absolute shrinkage and selection operator (LASSO) method lasso, [21, 22] as well as multivariate Cox regression analysis in order. Follow the criterion of Akaike information, the final selected GRGs related to prognosis are used to construct gene prediction models. We calculated the risk score using the following formula: Risk score .[23] The Kaplan-Meier survival curve constructed using the R package “survival” [24, 25] demonstrated the OS of the high- and low-risk groups, which were stratified according to risk signature. The predictive performance and validity of the model was assessed by a time-dependent receiver-operating characteristic (ROC) curve, which was conducted by using the R package “survivalROC.”[24] Univariate and multivariate Cox regression analyses were performed to determine the prognostic values of the signature and some clinic-pathological features.

2.3. Statistical analyses

We compared the distribution of the clinical features using the chi-square tests, which included age, tumor stage, and histological grade, between the different subgroups. The alteration in selected genes is obtained from the cBioPortal website (http://www.cbioportal.org/). We made use of the R version 3.6.2 software to conduct the statistical analyses. When p value is less than 0.05, the statistical differences is deemed significant.

3.1. Patient characteristics and collection of GRGs

From TCGA database, we downloaded the complete data of 583 OV patients which contained clinical information and the expression profiles of RNA sequencing. We manually searched for glycolysis-related gene sets from MSigDB version 6.2 and referenced relevant literature. Five related gene sets (REACTOME_GLYCOLYSIS, HALLMARK_GLYCOLYSIS, GO_GLYCOLYTIC_PROCESS, KEGG_GLYCOLYSIS_GLUCONEOGENESIS, and BIOCARTA_GLYCOLYSIS_ PATHWAY) were downloaded, and 443 genes were obtained. After excluding duplicate genes, the corresponding 386 genes remained for the subsequent analysis. The integrated clinical data and list of the GRGs are shown in Tables 1 and S1, respectively.

Table 1

Clinic pathological characteristics of extracted patients with ovarian cancer.
Characteristic	Group	No. of cases (%)
Age (years)	≤ 65	403(69.13)
	> 65	180(30.87)
TNM stage	Stage I	33(5.66)
	Stage II	41(7.03)
	Stage III	389(66.7)
	Stage IV	102(17.50)
	Stage X	18(3.08)
Histologic grade	G1	11(1.89)
	G2	101(17.32)
	G3	456(78.21)
	G4	1(0.17)
	GX	14(2.40)
Vital status	Alive	241(41.34)
	Dead	342(58.66)
Abbreviation: Stage X: unknown pathological stage; GX: unknown histological grade.

3.2. Construction of the glycolysis-related risk signature

To examine the prognostic value of the 386 GRGs, 11 genes were obtained by preliminary screening, which was performed by univariate Cox regression analysis with the criterion of adjusted p < 0.05. Finally, nine genes (ISG20, CITED2, PYGB, IRS2, LHX9, PC, ANGPTL4, TGFBI, and DDIT4) significantly correlated with OS (adjusted p < 0.05) after filtration using LASSO and multivariable Cox regression analyses (Fig. 1A, B and C, Table 2). And then a predictive signature based on GRGs was constructed and the formula to assess the risk of every patients survival was showed as follows: risk score = (− 0.25414) × ISG20 expression level + 0.07897 × CITED2 expression level + 0.11769 × PYGB expression level + 0.09112 × IRS2 expression level + 0.06399 × ANGPTL4 expression level + 0.04811 × TGFBI expression level + 0.03555 × LHX9 expression level + 0.05593 × PC expression level + 0.05907 × DDIT4 expression level. Additionally, Fig. 1D showed the mutation of the nine selected GRGs of 583 OV patients in the cBioPortal database. And we used the gene model calculated every patient’s risk score in the training set on the basis of the nine-GRG expression levels). The results suggested that patients with high risk scores were associated with a worse prognosis, when compare to patients with low risk scores (P < 0.0001, log-rank test; Fig. 2A and 2B). The areas under the curve (AUC) for 3-, 5-, and 10-year OS were 0.709, 0.762, and 0.808, respectively (Fig. 2C). The heatmap revealed the expression patterns of the 9 GRGs between the two different prognostic patient groups (Fig. 2D).

Table 2

Coefficients and multivariable Cox model results of each gene in 9-GRG risk signature.
Gene	Ensemble ID	Coefficient	HR (95%CIs)	P value
ISG20	ENSG00000172183	-0.25414	0.78 (0.70–0.87)	5.22E-06
CITED2	ENSG00000164442	0.078975	1.08 (0.92–1.28)	0.356061
PYGB	ENSG00000100994	0.117691	1.12 (0.94–1.34)	0.186993
IRS2	ENSG00000185950	0.091117	1.10 (0.97–1.23)	0.133015
ANGPTL4	ENSG00000167772	0.063993	1.07 (0.96–1.18)	0.218495
TGFBI	ENSG00000120708	0.048112	1.05 (0.94–1.17)	0.393959
LHX9	ENSG00000143355	0.035546	1.04 (0.97–1.11)	0.326248
PC	ENSG00000173599	0.055931	1.05 (0.91–1.23)	0.47917
DDIT4	ENSG00000168209	0.059074	1.06 (0.94–1.19)	0.328672
Abbreviation: HR: hazard ratio.

3.3. Evaluation of the predictive capability of the nine-GRG risk signature

After constructing the GRG predictive model, we selected three datasets to verify its prediction performance. The sets for validation including GSE63885, GSE26193, and GSE30161 datasets with 101, 107, and 58 patients with OV, respectively (Fig. 3). The AUCs of 3-, 5-, and 10-year OS were respectively 0.716, 0.767, and 0.819 in the GSE26193 dataset; 0.808, 0.8, and 0.853 in the GSE30161 dataset; and 0.636, 0.722, and 0.815 in the GSE63885 dataset. Survival analysis revealed that our risk signature performed well in the validation sets. And the survival differences between the satisfied high-risk and low-risk groups were all statistically significant (P value of < 0.05). The distribution of risk scores and survival statuses of the patients with OV in the three sets are also displayed in Fig. 3.

3.4. Risk score generated from the nine-gene signature as an independent prognostic indicator

The exploration of independent predictive factors is through univariate analysis of clinical factors and risk models combined with multivariate regression analysis. Table 3 showed that, in addition to age and tumor stage, our GRG risk model could independently predict the OS according to the results from univariate analysis (HR [95% CI], 2.334 [1.817 − 2.997]; P < 0.001) and multivariate analysis (HR [95% CI], 2.361 [1.830 − 3.047]; P < 0.001), referring to the statistical standard of adjusted P value of < 0.05. Furthermore, the AUC of the nine-gene signature was higher than that of any single clinicopathological variable (Fig. 4A). The findings in the present study suggest that the gene model had independently and effectively predictive ability in the survival prediction for OV patients. To develop a quantitative method that can predict the OS of patients with OV, a nomogram was constructed. The predictors included risk score and age (Fig. 4B).

Table 3

The risk score generated from the nine-GRG signature as an independent indicator according to Cox proportional hazards regression model.
Variable	Univariate analysis			Multivariate analysis
Variable	HR (95%CI)	P-value		HR (95%CI)	P-value
Age (≤ 65 / >65)	1.023(1.010 − 1.036)		< 0.001	1.024(1.011 − 1.037)		< 0.001
TNM stage ( I/II / III/IV )	1.643(1.028 − 1.752)		< 0.039	1.198(0.885 − 1.622)		0.177
Histologic grade(G_1/2 / G_3/4)	1.213(0.816 − 1.801)		0.340	1.308(0.871 − 1.963)		0.196
Risk score (H / L)	2.334(1.817 − 2.997)		< 0.001	2.361(1.830 − 3.047)		< 0.001
Abbreviation: GRGs, glycolysis-related genes; H, High; L, low.

3.5. Validation of the nine-GRG signature in predicting survival using Kaplan-Meier curves

The clinical features of age, histological grade, and tumor stage represent predictive prognostic factors of OV after the performance of univariate Cox regression analysis of OS. Kaplan-Meier curves revealed that clinical features that showed consistent results, namely patient age of > 65 years and disease stages III and IV, were associated with poor prognosis (Figs. 4C–4D).

For the purpose of testing whether our nine-GRG signature can play a role in different TNM stages, histological grades, and ages, a subgroup analysis was performed for each clinical feature. After the analyses of Kaplan-Meier survival, the risk signature was demonstrated to be stably prognostic power and applicable to the OV patients when they were stratified in to different age and TNM stage groups (Figs. 5A and 5B). However, when the patients were stratified into high-grade (grades 3 and 4) and low-grade (grades 1 and 2) subgroups, the risk score of the eight-gene signature remained an independent prognostic indicator in the high-grade subgroup (P < 0.001), but not in the low-grade subgroup (P = 0.067; Fig. 5C). The risk model showed more effective prediction in the patients groups of high-grade OV.

More and more attentions to the globe burden ovarian cancer have been paid. Despite current advances in surgery and chemotherapy, its poor prognosis remains big challenges.[3] Because of its heterogeneity and the lack of convenient and accurate biomarkers, the current prognostic tools of OV patients has limited clinically predictive abilities.[2, 26] Subtype identification, risk stratification, and characterization of the underlying mechanism are critical for improvement of the existing treatment methods, development of more precise and personalized therapies, and prolongation of survival time. Thus, a predictive model with a broad scope of application is needed for accurate prediction of OS in patients with OV and for guiding clinicians in targeted treatment and better prognosis. With the popular application of large databases, more and more prognostic markers are recognized.[8, 9, 27] Many studies have explored the biological function of GRGs in cancers.[14–16] In addition, considering that an increasing number of studies have discovered prognostic markers of glycolysis-related genes, the establishment of GRG-based risk signature to predict the survival of OV patients is necessary.

In our predictive model, this study consisted of a training set and 3 validation cohorts, which included 813 patients with OV. Nine genes with prognostic value for patients with OV were identified using univariate, multivariate, and LASSO Cox regression analyses. The study results indicate that the nine-GRG signature developed herein significantly correlates with poor prognosis in OV. In addition, this risk signature was still an independent prognostic factor in the multivariate Cox analyses. Results of survival analysis suggested that patients with higher risk score tend to have worse clinical outcomes. The 8-gene model showed better predictive ability than any single gene and clinicopathological factors. Nevertheless, due to the lack of relevant data in TCGA database, one limitation is that our model cannot predict recurrence and distant metastasis of OV patients. The model established in present study de well in the OS prediction. And a novel nomogram was established in our study by combing the prediction model and clinical characteristics. The nomogram used the complementary values of clinical characteristics and the prediction model and provided superior estimation of OS. Moreover, the gene signature could further assessed the survival risk when the patients have different clinical features (age, TNM stage, and histological grade). The risk model have effective prediction power for patients with diverse clinical characteristics, but its predictive power showed limited when patents with low grade which should be explored in depth in the future. This result implies that the clinical application of genetic models is far-reaching and the means for patients to predict prognosis in clinic will become more diverse, which can guide clinicians to treat more accurately.

The 9 GRGs we identified include ISG20, CITED2, PYGB, IRS2, ANGPTL4, TGFBI, LHX9, PC, and DDIT4. Among the 9 biomarker genes identified in the present study, DDIT4 (DNA damage-inducible transcript 4) with high expression level actively responded to hypoxia-inducible factor 1 and act together to regulated the generation of cell reactive oxygen species.[28] As an oncogene, [28, 29] its overexpression correlates with tumor progression and worse outcomes in several human cancers, including OV.[18, 30–32] PYGB (brain-type glycogen phosphorylase) could regulate multiple biological characteristics of cancer cells, such as proliferation, invasion, and apoptosis and metastatic phenotypes of several cancers.[33–38] PYGB regulates the Wnt/β-catenin signaling pathway to achieve cancer-promoting effects in OV[39], non-small cell lung cancer (NSCLC),[40] and gastric cancer.[41] IRS2 (insulin receptor substrates 2) mediates mitogenic and antiapoptotic signaling of insulin-like growth factor 1 receptor (IGF-IR), insulin receptor (IR), and other oncoproteins[42, 43] and is essential for cancer cell motility and metastasis.[44–46] IRS2 acted as an oncogene in OV and was involved in cell proliferation and ascites migration effect in OV progression.[47, 48] ANGPTL4 (angiopoietin-like 4) has been reported to be involved in ferroptotic cell death and chemoresistance of epithelial OV.[49] Moreover, large amounts of it has been detected in malignant ascites of serous OV patients.[50] High ANGPTL4 levels predict shorter relapse-free survival in serous OV.[50, 51] Studies have found that high promoter hypermethylation of TGFBI (transforming growth factor-beta-inducible gene) is involved in chemotherapy resistance of paclitaxel in OV.[52, 53] A study showed that TGFBI and periostin, predict poor prognosis in serous epithelial OV.[54] PC (pyruvate carboxylase) is a biotin-containing enzyme that converts pyruvate to oxaloacetate and has been implicated in cancer progression. PC is strongly involved in tumorigenesis in several cancers, such as breast cancer, NSCLC, glioblastoma, renal carcinoma, and gallbladder cancer.[55–58] Moreover, PC may mediate the regulation of TNKS (tankyrase) in aerobic glycolysis and may be involved in the TNKS-regulated development of OV, as its oncogenic activity is induced by TNKS activating Wnt/β-catenin/snail signaling.[59] Not much evidence has been accumulated on the following genes from OV basic research: ISG20 (interferon-stimulated gene 20), an RNA exonuclease,[60] stimulates tumor progression in hepatocellular carcinoma, clear cell renal cell carcinoma, and glioma.[61–63] The high expression level of ISG20 has association with the poor clinical outcome of OV patients.[63] CITED2 (Cbp/p300-interacting transactivator 2), a pleiotropic protein, has been reported to participate in several biological functions of cells, which included transcription and differentiation. High CITED2 expression level is correlated with poor patient survival in breast [64] and prostate cancers.[65] CITED2 participated in the regulation of the cell cycle, promoted cell proliferation and then played active role in progression of lung cancer.[66, 67] and supports gastric cancer cell colony formation and proliferation.[68] CITED2 was involved in resistance to platinum-based chemotherapy in OV.[69] LHX9 (LIM homeobox 9) is a developmentally expressed transcription factor[70] and is strongly expressed in the ovarian surface epithelium.[71] Previous research has the development of childhood malignant gliomas involved LHX9 abnormal methylation and silencing [72] The relationship of ISG20, CITED2, and LHX9 with OV and its molecular mechanism must be examined in depth in future studies. We integrated the 9 GRGs into a panel and established a novel multigene signature for predicting prognosis in OV. This signature showed a strong predictive ability and acted as an independent prognostic molecular factor in patients with OV.

To our knowledge, our study firstly identified the GRG risk predictive signature using the data from TCGA public database. The nine-GRG risk model showed a promising survival prediction ability for the prognosis of OV. We also analyzed mutations in the 9 selected genes in the cBioPortal database. Despite these promising results, questions remain. First, our study was not a prospective study, and all the patients with OV were identified from public databases. Second, large-scale multicenter cohorts are necessary to verify our findings, and further studies are need to further explored the functional roles of the GRGs involved in the initiation and development of OV. Moreover, the gene signature performed more effectively in high-grade OV patients than in low-grade patients, and the mechanism of this observation is should be fully investigated in the future. To further validate the utility of this risk model, we the works of clinical data and specimens collection have been undertaken.

We constructed a valid, innovative, reliable nine-GRG prognostic model (ISG20, CITED2, PYGB, IRS2, ANGPTL4, TGFBI, LHX9, PC, and DDIT4) for predicting patient outcomes in OV. Moreover, our signature is an independent and important risk factor of OV. The predictive capability of this model in OV requires further testing to improve prognostic stratification and treatment management.

ROC curve: receiver operating characteristic curve; AUC: areas under the curve; OV: ovarian cancer; OS: overall survival; GRG: glycolysis-related gene; HR: hazard ratio; TCGA: The Cancer Genome Atlas database; GEO: the Gene Expression Omnibus database; LASSO: the least absolute shrinkage and selection operator cox; GX: unknown histological grade; Stage X: unknown pathological stage.

Acknowledgments

We thank The Cancer Genome Atlas (TCGA) project, NCBI GEO and cBioPortal database and their contributors for these valuable public data sets.

Authors’ contributions

Dai Zhang, Si Yang and Yiche Li: Data curation, Writing - Original Draft, Methodology, Software. Meng Wang, Jia Yao and Yi Zheng: Conceptualization, Methodology, Software, Validation. Yujiao Deng and Na Li: Visualization, Investigation. Zhen Zhai, Ying Wu and Bajin Wei: Software, Validation. ZJD and HFK: Supervision, Project administration. All the authors read and approved the final manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Ethics approval and consent to participate

This study has been proved by the Institutional Review Board of the First Affiliated Hospital of Zhejiang University in Zhejiang Province (Hangzhou, China).

Availability of data and materials

The datasets generated and analyzed during the current study are available in the TCGA (http://cancergenome.nih.gov/abouttcga), GEO (https://www.ncbi.nlm.nih.gov/geo/) and cBioPortal (http://www.cbioportal.org) databases. ZJD and HFK had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Conflicts of interest

The authors declare that there are no conflicts of interest.

Consent for publication

Written informed consent for publication was obtained from all participants.

Siegel RL, Miller KD, Jemal A: Cancer statistics, 2019. CA: a cancer journal for clinicians 2019, 69(1):7-34.
Au KK, Josahkian JA, Francis JA, Squire JA, Koti M: Current state of biomarkers in ovarian cancer prognosis. Future oncology (London, England) 2015, 11(23):3187-3195.
Hoppenot C, Eckert MA, Tienda SM, Lengyel E: Who are the long-term survivors of high grade serous ovarian cancer?Gynecologic oncology 2018, 148(1):204-212.
Kang S, Kim EH, Hwang JE, Shin JH, Jeong YS, Yim SY, Joo EW, Eun YG, Lee DJ, Sohn BH et al: Prognostic significance of high metabolic activity in breast cancer: PET signature in breast cancer. Biochem Biophys Res Commun 2019, 511(1):185-191.
Zhou Z, Huang R, Chai R, Zhou X, Hu Z, Wang W, Chen B, Deng L, Liu Y, Wu F: Identification of an energy metabolism-related signature associated with clinical prognosis in diffuse glioma. Aging (Albany NY) 2018, 10(11):3185-3209.
McGrath SE, Annels N, Madhuri TK, Tailor A, Butler-Manuel SA, Morgan R, Pandha H, Michael A: Engrailed-2 (EN2) - a novel biomarker in epithelial ovarian cancer. BMC cancer 2018, 18(1):943.
Yang WL, Lu Z, Bast RC, Jr.: The role of biomarkers in the management of epithelial ovarian cancer. Expert review of molecular diagnostics 2017, 17(6):577-591.
Shen S, Wang G, Zhang R, Zhao Y, Yu H, Wei Y, Chen F: Development and validation of an immune gene-set based Prognostic signature in ovarian cancer. EBioMedicine 2019, 40:318-326.
Liu G, Chen L, Ren H, Liu F, Dong C, Wu A, Liu Z, Zheng Y, Cheng X, Liu L: Seven Genes Based Novel Signature Predicts Clinical Outcome and Platinum Sensitivity of High Grade IIIc Serous Ovarian Carcinoma. International journal of biological sciences 2018, 14(14):2012-2022.
Wang R, Ye XH, Zhao XL, Liu JL, Zhang CY: Development of a five-gene signature as a novel prognostic marker in ovarian cancer. Neoplasma 2019, 66(3):343-349.
Sun H, Cao D, Ma X, Yang J, Peng P, Yu M, Zhou H, Zhang Y, Li L, Huo X et al: Identification of a Prognostic Signature Associated With DNA Repair Genes in Ovarian Cancer. Frontiers in genetics 2019, 10:839.
Akram M: Mini-review on glycolysis and cancer. Journal of cancer education : the official journal of the American Association for Cancer Education 2013, 28(3):454-457.
Altenberg B, Greulich KO: Genes of glycolysis are ubiquitously overexpressed in 24 cancer classes. Genomics 2004, 84(6):1014-1020.
Chen C, Shi Y, Li Y, He ZC, Zhou K, Zhang XN, Yang KD, Wu JR, Kung HF, Ping YF et al: A glycolysis-based ten-gene signature correlates with the clinical outcome, molecular subtype and IDH1 mutation in glioblastoma. Journal of genetics and genomics = Yi chuan xue bao 2017, 44(11):519-530.
Nahm JH, Kim HM, Koo JS: Glycolysis-related protein expression in thyroid cancer. Tumour Biol 2017, 39(3):1010428317695922.
Ganapathy-Kanniappan S, Geschwind JF: Tumor glycolysis as a target for cancer therapy: progress and prospects. Molecular cancer 2013, 12:152.
Xiang J, Hu Q, Qin Y, Ji S, Xu W, Liu W, Shi S, Liang C, Liu J, Meng Q et al: TCF7L2 positively regulates aerobic glycolysis via the EGLN2/HIF-1α axis and indicates prognosis in pancreatic cancer. Cell death & disease 2018, 9(3):321.
Liu C, Li Y, Wei M, Zhao L, Yu Y, Li G: Identification of a novel glycolysis-related gene signature that can predict the survival of patients with lung adenocarcinoma. Cell cycle (Georgetown, Tex) 2019, 18(5):568-579.
Wang ZH, Zhang YZ, Wang YS, Ma XX: Identification of novel cell glycolysis related gene signature predicting survival in patients with endometrial cancer. Cancer cell international 2019, 19:296.
Liberzon A, Subramanian A, Pinchback R, Thorvaldsdóttir H, Tamayo P, Mesirov JP: Molecular signatures database (MSigDB) 3.0. Bioinformatics (Oxford, England) 2011, 27(12):1739-1740.
TIBSHIRANI R: THE LASSO METHOD FOR VARIABLE SELECTION IN THE COX MODEL. Statistics in Medicine 1997, 16(4):385-395.
Wang H, Lengerich BJ, Aragam B, Xing EP: Precision Lasso: accounting for correlations and linear dependencies in high-dimensional genomic data. Bioinformatics (Oxford, England) 2019, 35(7):1181-1187.
Chen HY, Yu SL, Chen CH, Chang GC, Chen CY, Yuan A, Cheng CL, Wang CH, Terng HJ, Kao SF et al: A five-gene signature and clinical outcome in non-small-cell lung cancer. The New England journal of medicine 2007, 356(1):11-20.
Lorent M, Giral M, Foucher Y: Net time-dependent ROC curves: a solution for evaluating the accuracy of a marker to predict disease-related mortality. Stat Med 2014, 33(14):2379-2389.
Harrell FE, Jr., Lee KL, Mark DB: Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med 1996, 15(4):361-387.
SC M, T B, V V, A B, ME J, KK W, DC P, K H, DK Y, H D et al: A gene signature predictive for outcome in advanced ovarian cancer identifies a survival factor: microfibril-associated glycoprotein 2. Cancer cell 2009, 16(6):521-532.
Yang S, Wu Y, Deng Y, Zhou L, Yang P, Zheng Y, Zhang D, Zhai Z, Li N, Hao Q et al: Identification of a prognostic immune signature for cervical cancer to predict survival and response to immune checkpoint inhibitors. Oncoimmunology 2019, 8(12):e1659094.
Shoshani T, Faerman A, Mett I, Zelin E, Tenne T, Gorodin S, Moshel Y, Elbaz S, Budanov A, Chajut A et al: Identification of a novel hypoxia-inducible factor 1-responsive gene, RTP801, involved in apoptosis. Molecular and cellular biology 2002, 22(7):2283-2293.
Smith ER, Xu XX: REDD1, a new Ras oncogenic effector. Cell cycle (Georgetown, Tex) 2009, 8(5):675-676.
Tirado-Hurtado I, Fajardo W, Pinto JA: DNA Damage Inducible Transcript 4 Gene: The Switch of the Metabolism as Potential Target in Cancer. Frontiers in oncology 2018, 8:106.
Pinto JA, Rolfo C, Raez LE, Prado A, Araujo JM, Bravo L, Fajardo W, Morante ZD, Aguilar A, Neciosup SP et al: In silico evaluation of DNA Damage Inducible Transcript 4 gene (DDIT4) as prognostic biomarker in several malignancies. Sci Rep 2017, 7(1):1526.
Chang B, Meng J, Zhu H, Du X, Sun L, Wang L, Li S, Yang G: Overexpression of the recently identified oncogene REDD1 correlates with tumor progression and is an independent unfavorable prognostic factor for ovarian carcinoma. Diagnostic pathology 2018, 13(1):87.
Altemus MA, Goo LE, Little AC, Yates JA, Cheriyan HG, Wu ZF, Merajver SD: Breast cancers utilize hypoxic glycogen stores via PYGB, the brain isoform of glycogen phosphorylase, to promote metastatic phenotypes. PloS one 2019, 14(9):e0220973.
Wang Z, Han G, Liu Q, Zhang W, Wang J: Silencing of PYGB suppresses growth and promotes the apoptosis of prostate cancer cells via the NF‑κB/Nrf2 signaling pathway. Molecular medicine reports 2018, 18(4):3800-3808.
Shimada S, Shiomori K, Tashima S, Tsuruta J, Ogawa M: Frequent p53 mutation in brain (fetal)-type glycogen phosphorylase positive foci adjacent to human 'de novo' colorectal carcinomas. Br J Cancer 2001, 84(11):1497-1504.
Tashima S, Shimada S, Yamaguchi K, Tsuruta J, Ogawa M: Expression of brain-type glycogen phosphorylase is a potentially novel early biomarker in the carcinogenesis of human colorectal carcinomas. The American journal of gastroenterology 2000, 95(1):255-263.
Takashi M, Koshikawa T, Kurobe N, Kato K: Elevated concentrations of brain-type glycogen phosphorylase in renal cell carcinoma. Japanese journal of cancer research : Gann 1989, 80(10):975-980.
Shimada S, Tashima S, Yamaguchi K, Matsuzaki H, Ogawa M: Carcinogenesis of intestinal-type gastric cancer and colorectal cancer is commonly accompanied by expression of brain (fetal)-type glycogen phosphorylase. Journal of experimental & clinical cancer research : CR 1999, 18(1):111-118.
Zhou Y, Jin Z, Wang C: Glycogen phosphorylase B promotes ovarian cancer progression via Wnt/β-catenin signaling and is regulated by miR-133a-3p. Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie 2019, 120:109449.
Xiao L, Wang W, Huangfu Q, Tao H, Zhang J: PYGB facilitates cell proliferation and invasion in non-small cell lung cancer through activating Wnt/β-catenin signaling. Biochemistry and cell biology = Biochimie et biologie cellulaire 2020.
Xia B, Zhang K, Liu C: PYGB Promoted Tumor Progression by Regulating Wnt/β-Catenin Pathway in Gastric Cancer. Technology in cancer research & treatment 2020, 19:1533033820926592.
Dearth RK, Cui X, Kim HJ, Kuiatse I, Lawrence NA, Zhang X, Divisova J, Britton OL, Mohsin S, Allred DC et al: Mammary tumorigenesis and metastasis caused by overexpression of insulin receptor substrate 1 (IRS-1) or IRS-2. Molecular and cellular biology 2006, 26(24):9302-9314.
DeAngelis T, Chen J, Wu A, Prisco M, Baserga R: Transformation by the simian virus 40 T antigen is regulated by IGF-I receptor and IRS-1 signaling. Oncogene 2006, 25(1):32-42.
Dalmizrak O, Wu A, Chen J, Sun H, Utama FE, Zambelli D, Tran TH, Rui H, Baserga R: Insulin receptor substrate-1 regulates the transformed phenotype of BT-20 human mammary cancer cells. Cancer Res 2007, 67(5):2124-2130.
Chan BT, Lee AV: Insulin receptor substrates (IRSs) and breast tumorigenesis. Journal of mammary gland biology and neoplasia 2008, 13(4):415-422.
Hoang CD, Zhang X, Scott PD, Guillaume TJ, Maddaus MA, Yee D, Kratzke RA: Selective activation of insulin receptor substrate-1 and -2 in pleural mesothelioma cells: association with distinct malignant phenotypes. Cancer Res 2004, 64(20):7479-7485.
Meunier L, Puiffe ML, Le Page C, Filali-Mouhim A, Chevrette M, Tonin PN, Provencher DM, Mes-Masson AM: Effect of ovarian cancer ascites on cell migration and gene expression in an epithelial ovarian cancer in vitro model. Translational oncology 2010, 3(4):230-238.
Tan Y, Cheung M, Pei J, Menges CW, Godwin AK, Testa JR: Upregulation of DLX5 promotes ovarian cancer cell proliferation by enhancing IRS-2-AKT signaling. Cancer Res 2010, 70(22):9197-9206.
Yang WH, Huang Z, Wu J, Ding CC, Murphy SK, Chi JT: A TAZ-ANGPTL4-NOX2 Axis Regulates Ferroptotic Cell Death and Chemoresistance in Epithelial Ovarian Cancer. Molecular cancer research : MCR 2020, 18(1):79-90.
Conway GD, Buzza MS, Martin EW, Duru N, Johnson TA, Peroutka RJ, Pawar NR, Antalis TM: PRSS21/testisin inhibits ovarian tumor metastasis and antagonizes proangiogenic angiopoietins ANG2 and ANGPTL4. Journal of molecular medicine (Berlin, Germany) 2019, 97(5):691-709.
Schumann T, Adhikary T, Wortmann A, Finkernagel F, Lieber S, Schnitzer E, Legrand N, Schober Y, Nockher WA, Toth PM et al: Deregulation of PPARβ/δ target genes in tumor-associated macrophages by fatty acid ligands in the ovarian cancer microenvironment. Oncotarget 2015, 6(15):13416-13433.
Wang N, Zhang H, Yao Q, Wang Y, Dai S, Yang X: TGFBI promoter hypermethylation correlating with paclitaxel chemoresistance in ovarian cancer. Journal of experimental & clinical cancer research : CR 2012, 31(1):6.
Kang S, Dong SM, Park NH: Frequent promoter hypermethylation of TGFBI in epithelial ovarian cancer. Gynecologic oncology 2010, 118(1):58-63.
Karlan BY, Dering J, Walsh C, Orsulic S, Lester J, Anderson LA, Ginther CL, Fejzo M, Slamon D: POSTN/TGFBI-associated stromal signature predicts poor prognosis in serous epithelial ovarian cancer. Gynecologic oncology 2014, 132(2):334-342.
P P, CA P, PA G, X L, P C, P M, GJ G, S J: MicroRNA-143-3p targets pyruvate carboxylase expression and controls proliferation and migration of MDA-MB-231 cells. Archives of biochemistry and biophysics 2019, 677:108169.
Phannasil P, Thuwajit C, Warnnissorn M, Wallace JC, MacDonald MJ, Jitrapakdee S: Pyruvate Carboxylase Is Up-Regulated in Breast Cancer and Essential to Support Growth and Invasion of MDA-MB-231 Cells. PloS one 2015, 10(6):e0129848.
U L-O, PV A, S J: Roles of pyruvate carboxylase in human diseases: from diabetes to cancers and infection. Journal of molecular medicine (Berlin, Germany) 2018, 96:237-247.
Q L, Y H, X W, Y Z, M H, W G, Y H, T Z, L L, Z S et al: Targeting Pyruvate Carboxylase by a Small Molecule Suppresses Breast Cancer Progression. Advanced science (Weinheim, Baden-Wurttemberg, Germany) 2020, 7(9):1903483.
Yang HY, Shen JX, Wang Y, Liu Y, Shen DY, Quan S: Tankyrase Promotes Aerobic Glycolysis and Proliferation of Ovarian Cancer through Activation of Wnt/β-Catenin Signaling. BioMed research international 2019, 2019:2686340.
Zheng Z, Wang L, Pan J: Interferon-stimulated gene 20-kDa protein (ISG20) in infection and disease: Review and outlook. Intractable & rare diseases research 2017, 6(1):35-40.
SL L, SM W, IH C, YH L, CY C, HC C, TK L, CT Y, KH L: Stimulation of Interferon-Stimulated Gene 20 by Thyroid Hormone Enhances Angiogenesis in Liver Cancer. Neoplasia (New York, NY) 2018, 20(1):57-68.
Xu T, Ruan H, Gao S, Liu J, Liu Y, Song Z, Cao Q, Wang K, Bao L, Liu D et al: ISG20 serves as a potential biomarker and drives tumor progression in clear cell renal cell carcinoma. Aging (Albany NY) 2020, 12(2):1808-1827.
Gao M, Lin Y, Liu X, Li Y, Zhang C, Wang Z, Wang Z, Wang Y, Guo Z: ISG20 promotes local tumor immunity and contributes to poor survival in human glioma. Oncoimmunology 2019, 8(2):e1534038.
Jayaraman S, Doucet M, Kominsky SL: Down-regulation of CITED2 attenuates breast tumor growth, vessel formation and TGF-β-induced expression of VEGFA. Oncotarget 2017, 8(4):6169-6178.
Shin SH, Lee GY, Lee M, Kang J, Shin HW, Chun YS, Park JW: Aberrant expression of CITED2 promotes prostate cancer metastasis by activating the nucleolin-AKT pathway. Nature communications 2018, 9(1):4113.
Bai L, Merchant JL: A role for CITED2, a CBP/p300 interacting protein, in colon cancer cell invasion. FEBS letters 2007, 581(30):5904-5910.
Yao Y, Zhang T, Qi L, Liu R, Liu G, Wang X, Li J, Li J, Sun C: Competitive Endogenous RNA Network Construction and Comparison of Lung Squamous Cell Carcinoma in Smokers and Nonsmokers. Disease markers 2019, 2019:5292787.
Tang Z, He G, Xu J, Zhongfu L: Knockdown of Cbp/P300-interacting transactivator with Glu/Asp-rich carboxy-terminal domain 2 inhibits cell division and increases apoptosis in gastric cancer. The Journal of surgical research 2017, 211:1-7.
Ju W, Yoo BC, Kim IJ, Kim JW, Kim SC, Lee HP: Identification of genes with differential expression in chemoresistant epithelial ovarian cancer using high-density oligonucleotide microarrays. Oncology research 2009, 18(2-3):47-56.
S B, SP K, VK V, S C, SP M, H R, D P, P K, K S: Aberrant gene-specific DNA methylation signature analysis in cervical cancer. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2017, 39(3):1010428317694573.
Mullin NP, Yates A, Rowe AJ, Nijmeijer B, Colby D, Barlow PN, Walkinshaw MD, Chambers I: The pluripotency rheostat Nanog functions as a dimer. The Biochemical journal 2008, 411(2):227-231.
V V, T M, A W, N S, J X, PC R, T P: Aberrant methylation and reduced expression of LHX9 in malignant gliomas of childhood. Neoplasia (New York, NY) 2009, 11(7):700-711.

TableS1.docx
Supplementary Materials Table S1. The list of the glycolysis-related genes.

Download PDF

Journal Publication

published 05 Oct, 2021

Read the published version in Cancer Medicine →

Version 1

posted

You are reading this latest preprint version

Identification of a Glycolysis-Related Gene Signature for Survival Prediction of Ovarian Cancer Patients

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Materials And Methods

2.3. Statistical analyses

3. Results

3.1. Patient characteristics and collection of GRGs

3.2. Construction of the glycolysis-related risk signature

3.3. Evaluation of the predictive capability of the nine-GRG risk signature

3.4. Risk score generated from the nine-gene signature as an independent prognostic indicator

3.5. Validation of the nine-GRG signature in predicting survival using Kaplan-Meier curves

4. Discussion

Conclusion

Abbreviations

Declarations

References

Supplementary Files

Status:

Journal Publication

Version 1