Hindawi BioMed Research International Volume 2021, Article ID 4804694, 18 pages https://doi.org/10.1155/2021/4804694

Research Article Integrative Analysis of Differently Expressed Reveals a 17- Prognosis Signature for Endometrial Carcinoma

Anna Wang,1 Hongyan Guo,2 and Zaiqiu Long 1

1Department of Gynecology, Cancer Hospital of China Medical University, Liaoning Cancer Hospital & Institute, Liaoning, China 2Department of Information Engineering, Shenyang Polytechnic College, Liaoning, China

Correspondence should be addressed to Zaiqiu Long; [email protected]

Received 15 May 2021; Revised 28 June 2021; Accepted 3 July 2021; Published 15 July 2021

Academic Editor: Tao Huang

Copyright © 2021 Anna Wang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Endometrial carcinoma (EC) is the fifth widely occurring malignant neoplasm among women all over the world. However, there is still lacking efficacy indicators for EC’s prognosis. Here, we analyzed two databases including an RNA-sequencing-based TCGA dataset and a microarray-based GSE106191. After normalizing the raw data, we identified 114 common genes with upregulation and 308 common genes with downregulation in both the TCGA and GSE106191 databases. Bioinformatics analysis showed that the differently expressed genes in EC were related to the IL17 signaling pathway, PI3K-Akt signaling pathway, and cGMP-PKG signaling pathway. Furthermore, we performed the least absolute shrinkage and selection operator (LASSO) Cox regression analysis and generated a signature featuring 17 prognosis-related genes (MAL2, ANKRD22, METTL7B, IL32, ERFE, OAS1, TRPC1, SRPX, RAPGEF4, PSD3, SIMC1, TRPC6, WFS1, PGR, PAMR1, KCNK6, and FAM189A2) and found that it could predict OS in EC patients. The further analysis showed that OAS1, MAL2, ANKRD22, METTL7B, and IL32 were significantly upregulated in EC samples after comparison with normal samples. However, TRPC1, SRPX, RAPGEF4, PSD3, SIMC1, TRPC6, WFS1, PGR, PAMR1, KCNK6, and FAM189A2 were significantly downregulated in EC samples in comparison with normal samples. And correlation analysis showed that our results showed that the expressions of 17 prognosis-related hub genes were significantly correlated based on Pearson correlation. We here offer a newly genetic biomarker for the prediction of EC patients’ prognosis.

1. Introduction lenges [4]. To determine effective therapeutic strategies in ameliorating the prognostic status of EC patients is thus Endometrial carcinoma (EC) is the fifth commonly occurring essential. malignant neoplasm among women all over the world, with Presently, as per the World Health Organization (WHO) an estimated 382,000 new EC cases and nearly 90,000 deaths classification system classification, EC comprises two sorts in 2018 [1, 2]. Especially in the United States, it is speculated on the basis of histological features [5]. Endometrioid adeno- that the number of newly diagnosed ECs will be increasing carcinoma or well-differentiated endometrioid subtypes over time. It is estimated that the occurrence rate is still rising accounted for 80% over EC cases and was considered as with increasing risk factors for certain ECs, including obesity estrogen-dependent type I of EC [1, 4, 6]. Approximately rate and the aging of the US population [3]. The incidence 10% of EC cases were type II, manifested as nonendometrial rate of EC increases after the age of 30, and the peak inci- or poorly differentiated EC. Difference existed in the molec- dence is within 60 to 69 years. 20% to 30% of patients with ular changes of the two EC types [1, 4, 7]. In general, in the EC are diagnosed in the advanced stage during surgery. In activated oncogene and inactivated tumor suppressor gene, the clinic, the five-year survival rate of patients in stage III defective DNA repair contributed mainly to the occurrence ranged from 40% to 70% and in stage IV was within 0 to of neoplasms [1, 8]. For instance, the inactivated tumor sup- 10% [1, 4]. Despite therapeutic advances having been made, pressor gene PTEN accompanied by DNA mismatch repair high recurrence rate and metastasis remain to be big chal- gene defects manifested as the microsatellite instability 2 BioMed Research International

Group 3 300

2

1 200 p value

0 10 Log –1 100

–2 0 –3 Group –10 –5 –1 0 1 5 Tumor_G1 Log2 (fold change) Normal Down-regulation None Up-regulation PCI (13.3%)

Group 8 3 2 6 1 4 0 p value 10 –1 Log 2 –2 –3 0 –0.378511623253730 0.37851162325373 Group G1 Log2 (fold change) G2 Down-regulation None Up-regulation

GSE106191-up GSE106191-down

42 114 2004 108 308 3681

TCGA-up TCGA-down

Figure 1: Screening of DEGs in EC. (a, b) Heat map (a) and volcano map (b) identified 2118 genes with upregulation and 3989 genes with downregulation by analyzing the TCGA database. (c, d) Heat map (c) and volcano map (d) showed 156 genes with upregulation and 416 genes with downregulation by analyzing the GSE106191 database. (e, f) Venn map analysis of common upregulated and downregulated genes in EC by analyzing the TCGA and GSE106191 databases. phenotype, or activated KRAS2 and/or adhesion molecules needed to identify the profiles between dif- genes were detected in the early stage of type I EC [9]. Previ- ferent histological types of ECs that distinguish normal cells ous studies have shown that mutated TP53 and Her-2 from cancer cells. There have been public reports showing occurred in type II EC, which was probably caused by the that the differentially expressed genes existed in different his- background of the atrophic endometrium. It seems that these tological sorts of EC [12, 13]. However, a limited set of genes molecular changes were specific in type I and type II of ECs were reported in these studies. More and more researches are [10, 11]. Although many efforts have been made to set up a thus needed to characterize EC and unearth the genes func- molecularly based histological classification, it is still urgently tioning importantly in the mechanisms of EC. BioMed Research International 3

KEGG pathway (up) −Log10 (p adjust) KEGG pathway (Up) −Log10 (p adjust) p53 signaling pathway Viral protein interaction with viral protein interaction with 6 Cytokine and cytokine receptor cytokine and cytokine receptor vibrio cholerae infection Transcriptional misregulation in cancer T17 cell diferentiation Rheumatoid arthritis Toll−like receptor signaling pathway 4 Primary immunodefciency TNF signaling pathway Mucin type O−glycan biosynthesis Rheumatoid arthritis IL−17 signaling pathway Human T−cell leukemia virus 1 infection Pyrimidine metabolism Hematopoietic cell lineage Proteoglycans in cancer 3 Glutathione metabolism NOD−like receptor signaling pathway Gastric cancer 4 Fanconi anemia pathway NF−kappa B signaling pathway Epstein−Barr virus infection Malaria Cytokine−cytokine receptor interaction 2 Legionellosis Cysteine and methionine metabolism Infuenza A Cell cycle Cell adhesion molecules (CAMs) IL−17 signaling pathway Bladder cancer Hepatitis C Biosynthesis of amino acids 1 ECM−receptor interaction 0.02 0.04 0.06 Cytokine−cytokine receptor interaction 2

Enrichment ratio Chemokine signaling pathway Cell cycle Count Bladder cancer 10 40 20 50 Amoebiasis 30 0.06 0.09 0.12 0.15

Enrichment ratio

Count 4 8 6 10 (a) (b)

KEGG pathway (Down) −Log10 (p adjust) cGMP−PKG signaling pathway Vascular smooth muscle contraction Relaxin signaling pathway 10 Regulation of lipolysis in adipocytes Ras signaling pathway Rap1 signaling pathway TCGA-KEGG-up Proteoglycans in cancer Protein digestion and absorption Phospholipase D signaling pathway 8 Phosphatidylinositol signaling system PI3K−Akt signaling pathway Neurotrophin signaling pathway MAPK signaling pathway FoxO signaling pathway 14 7 14 Focal adhesion 6 ECM−receptor interaction Dilated cardiomyopathy (DCM) Calcium signaling pathway Apelin signaling pathway AGE−RAGE signaling pathway 4 In diabetic complications GSE106191-KEGG-up 0.02 0.03 0.04 0.05 0.06 0.07

Enrichment ratio Count 25 75 50 100 (c) (d)

Figure 2: Continued. 4 BioMed Research International

KEGG pathway (Down) −Log10 (p adjust)

cGMP−PKG signaling pathway 3 Wnt signaling pathway TGF−beta signaling pathway Relaxin signaling pathway Proteoglycans in cancer Protein digestion and absorption Phenylalanine metabolism PI3K−Akt signaling pathway 2 Neuroactive ligand−receptor interaction Hippo signaling pathway−multiple species Hedgehog signaling pathway TCGA-KEGG-down Growth hormone synthesis, secretion, and action Glycosaminoglycan biosynthesis−chondroitin Sulfate/dermatan sulfate Gap junction

Focal adhesion 1 EGFR inhibitor resistance 14 7 14 ECM−receptor interaction Complement and coagulation cascades Basal cell carcinoma Axon guidance

0.025 0.050 0.075 0.100 GSE106191-KEGG-down Enrichment ratio

Count 4 12 8 16 (e) (f)

Figure 2: KEGG pathway analysis of DEGs in EC. (a) KEGG pathway analysis of upregulated genes by analyzing the TCGA database. (b) KEGG pathway analysis of upregulated genes by analyzing the GSE106191 database. (c) Venn map analysis of upregulated gene-related pathways. (d) KEGG pathway analysis of downregulated genes by analyzing the TCGA database. (e) KEGG pathway analysis of downregulated genes by analyzing the GSE106191 database. (f) Venn map analysis of downregulated gene-related pathways.

Herein, we attempted to systematically screen more novel 33 hyperplasia samples). Then, Software R (version 3.5.1, differentially expressed genes and related molecular path- https://www.r-project.org) and “limma” packages (http:// ways and the clinical significance of the identified genes in www.bioconductor.org/) [16] were applied to select the EC. To sum up, identification of EC-related genes and gene DEGs existing in the EC samples and control samples. These pathways as well as the clinical implication is conducive to selection criteria were adjusted p value < 0.05 and fold understanding the pathophysiology of this cancer and unco- change ðFCÞ ≥ 2 or fold change ðFCÞ ≤ 0:5. vering the potential diagnostic biomarkers of EC. 2.3. Functional Enrichment Analysis. We usually applied 2. Materials and Methods (GO) functional enrichment analysis to describe gene functions, consisting of molecular function 2.1. Establishment and Verification of Gene Prognostic (MF), biological process (BP), or cellular component (CC) Model. We carried out the LASSO Cox regression model [17, 18]. And we then utilized the Kyoto Encyclopedia of (R package “glmnet”) to select the candidate genes and sub- Genes and Genomes (KEGG) pathway enrichment analysis sequently established the prognostic model [14, 15]. We ulti- to identify molecular interaction and relation networks. p mately retained 17 genes and their coefficients and utilized value < 0.05 was thought to be statistically significant. Based the minimum criteria to determine the penalty parameter on the DEGs in the GEO datasets, we conducted GO and (λ). We calculated the risk score after centralizing and stan- KEGG enrichment analyses by the online tool DAVID dardizing (applying the “scale” function in R) the TCGA (https://david. http://ncifcrf.gov/home.jsp) [19, 20]. The top expression data. The risk score formula was shown as fol- significantly enriched analysis results were shown. lows: risk score = ∑7iXi×Yi (X: coefficients, Y: gene expres- sion level). 2.4. Survival Analysis. For identification of prognosis- 2.2. Identification of Differentially Expressed Genes (DEGs). predicting gens, we integrated the clinical data of EC patients We obtained expression matrixes and platform information in The Cancer Genome Atlas (TCGA) and carried out from Gene Expression Omnibus (GEO) datasets. The dataset Kaplan-Meier curve analysis [21, 22]. The survival curves of GSE106191 was used, which includes the primary tumor of DEGs were drew by “survival” package in R. p value < 0.05 66 endometrial cancer patients (64 carcinoma samples and was denoted as a significantly statistical difference. BioMed Research International 5

GO (Up) −Log10 (p adjust)

Sister chromatid segregation Response to molecule of bacterial origin 6 Response to lipopolysaccharide Response to interferon−gamma

GO (Up) −Log10 (p adjust) Regulation of microvillus organization

Spindle organization Regulation of cell projection size Sister chromatid segregation Organelle fssion 5 Regulation of nuclear division 10 Nuclear division Regulation of mitotic nuclear division Nuclear segregation Regulation of chromosome segregation Positive regulation of cell−cell adhesion Myeloid leukocyte migration Positive regulation of cell adhesion 9 Mitotic sister chromatid segregation Organelle fssion 4 Mitotic nuclear division Nuclear division Microvillus organization Nuclear chromosome segregation 8 Neutrophil chemotaxis Meiotic nuclear division Mitotic sister chromatid segregation Mitotic nuclear division Meiotic cell cycle process Mitotic cell cycle checkpoint 7 Meiotic cell cycle microtubule cytoskeleton organization 3 Involved in mitosis Granulocyte chemotaxis leukocyte chemotaxis Chromosome separation Chromosome segregation Chromosome segregation 6 Cell cycle checkpoint Cellular response to interferon−gamma Cell chemotaxis DNA replication

0.02 0.03 0.04 0.05 0.025 0.050 0.075 0.100 0.125

Enrichment ratio Enrichment ratio

Count Count 40 4 16 60 8 20 80 12 (a) (b) p GO (Down) −Log10 ( adjust) Striated muscle tissue development regulation of small GT pase Mediated signal transduction regulation of cell−substrate junction o rganization Regulation of cell−substrate adhesion Regulation of GT pase activity positive regulation of protein 14 serine/threonine kinase activity positive regulation of MAP kinase activity Muscle tissue development TCGA-GO-up Muscle system process Muscle organ development Muscle contraction 12 Focal adhesion assembly Extracellular structure organization Extracellular matrix organization Cell−substrate junction organization 14 7 13 Cell−substrate junction assembly 10 Cell−substrate adhesion Cell−matrix adhesion Actomyosin structure organization Ras protein signal transduction 0.01 0.02 0.03 0.04 0.05 GSE106191-GO-up Enrichment ratio

Count 40 120 80 160 (c) (d)

Figure 3: Continued. 6 BioMed Research International

GO (Down) −Log10 (p adjust)

Visual system development

Urogenital system development 15.0 Sensory system development Renal system development Regulation of ossifcation Regulation of epithelial cell proliferation 12.5 Ossifcation Nephron development Mesenchyme development Mesenchymal cell diferentiation Kidney development 10.0 TCGA-GO-down Eye development Extracellular structure organization Extracellular matrix organization Epithelial cell proliferation 7.5 Connective tissue development 20 2 18 Collagen fbril organization Cardiac epithelial to mesenchymal transition Biomineralization Biomineral tissue development 5.0

0.05 0.10 GSE106191-GO-down Enrichment ratio

Count 10 30 20 40 (e) (f)

Figure 3: GO analysis of DEGs in EC. (a) GO analysis of upregulated genes by analyzing the TCGA database. (b) GO analysis of upregulated genes by analyzing the GSE106191 database. (c) Venn map analysis of upregulated gene-related pathways. (d) GO analysis of downregulated genes by analyzing the TCGA database. (e) GO analysis of downregulated genes by analyzing the GSE106191 database. (f) Venn map analysis of downregulated gene-related pathways.

2.5. Statistical Analysis. SPSS 22.0 software (Chicago, USA) cinoma using TCGA and GSE106191, respectively. As was employed to analyze the data. All representative data presented in Figure 2, the KEGG analysis showed that the were shown as the mean ± standard deviation (SD) [23–25]. pathways related to DEGs were similar by analyzing either We carried out the Students’ t-test and one-way ANOVA TCGA or GSE106191 (Figures 2(a), 2(b), 2(d), and 2(e)). to separately determine the difference existing in two groups The KEGG analyses of upregulated genes were related to and multiple groups. p value < 0.05 was denoted as a signifi- bladder cancer, cell cycle, cytokine-cytokine receptor interac- cantly statistical difference [26–28]. All experiments were tion, IL17 signaling pathway, cytokine, and cytokine receptor performed in three independent times in replicates at one (Figure 2(c)). The KEGG analyses of downregulated genes time. were related to ECM-receptor interaction, focal adhesion, PI3K-Akt signaling pathway, protein digestion and absorp- 3. Results tion, proteoglycans in carcinoma, relaxin signaling pathway, and cGMP-PKG signaling pathway (Figure 2(f)). 3.1. Screening of DEGs in EC. To identify the DEGs in EC, we Furthermore, the GO analysis also showed that the bio- analyzed two databases including the RNA-sequencing- logical processes related to DEGs were similar by analyzing based TCGA dataset and microarray-based GSE106191. either TCGA or GSE106191 (Figures 3(a), 3(b), 3(d), and After normalizing the raw data, we identified 2118 genes with 3(e)). The GO analyses of upregulated genes were related to upregulation and 3989 genes with downregulation by analyz- chromosome segregation, mitotic nuclear division, mitotic ing the TCGA database (Figures 1(a) and 1(b)). Meanwhile, sister chromatid segregation, nuclear chromosome segrega- we found 156 genes with upregulation and 416 genes with tion, nuclear division, organelle fission, and sister chromatid downregulation by analyzing the GSE106191 database segregation (Figure 3(c)). The GO analysis of downregulated (Figures 1(c) and 1(d)). Among the DEGs, 114 common genes exhibited a relationship to extracellular matrix organi- genes with upregulation and 308 common genes with down- zation and structure organization (Figure 3(f)). regulation were identified in both the TCGA and GSE106191 databases (Figures 1(e) and 1(f)). 3.3. Identification of Prognosis-Related DEGs in EC. In the above analysis, we identified 572 DEGs in EC. In order to 3.2. Bioinformatics Analysis of DEGs in EC. Next, we per- identify prognosis-related DEGs in ECs, we carried out formed KEGG and GO analyses of DEGs in endometrial car- the Kaplan-Meier Plotter to determine the correlation of BioMed Research International 7

ERFE FBXO17 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate

Log−rank P = 0.005 0.2 Log−rank P = 0.002 0.2 HR = 1.85 (95%CI, 1.2−2.83) HR = 1.98 (95%CI, 1.29−3.03) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (a) (b) GLDC SRPX 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate

0.2 Log−rank P = 0.009 0.2 Log−rank P = 0.001 HR = 1.78 (95%CI, 1.16−2.73) HR = 2.02 (95%CI, 1.32−3.1) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (c) (d) MAL2 MCM10 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate P P Log−rank = 0.007 0.2 Log−rank = 0.003 0.2 HR = 1.9 (95%CI, 1.24−2.9) HR = 1.79 (95%CI, 1.17−2.74) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (e) (f)

Figure 4: Continued. 8 BioMed Research International

METTL7B OAS1 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate Log−rank P = 0.006 0.2 Log−rank P = 0.005 0.2 HR = 1.84 (95%CI, 1.2−2.8) HR = 1.81 (95%CI, 1.19−2.75) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (g) (h) PSAT1 PSD3 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate Log−rank P = 0.005 P 0.2 HR = 1.83 (95%CI, 1.2−2.8) 0.2 Log−rank = 0.001 HR = 2.13 (95%CI, 1.38−3.26) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (i) (j) UST RAPGEF4 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate Log−rank P = 0.008 0.2 Log−rank P = 0.001 0.2 HR = 1.78 (95%CI, 1.16−2.71) HR = 2.04 (95%CI, 1.33−3.14) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (k) (l)

Figure 4: Continued. BioMed Research International 9

PTTG1 SIMC1 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate Log−rank P = 0.002 0.2 0.2 Log−rank P = 0.004 HR = 1.98 (95%CI, 1.29−3.03) HR = 1.87 (95%CI, 1.22−2.86) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (m) (n) TRPC1 TRO 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate Log−rank P = 0.001 0.2 0.2 Log−rank P = 0.006 HR = 2.01 (95%CI, 1.31−3.08) HR = 1.8 (95%CI, 1.18−2.74) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (o) (p) BEX4 POLQ 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate

0.2 Log−rank P = 0.005 0.2 Log−rank P = 0.004 HR = 1.82 (95%CI, 1.19−2.78) HR = 1.88 (95%CI, 1.23−2.89) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (q) (r)

Figure 4: Continued. 10 BioMed Research International

TIMP3 1.0

0.8

0.6

0.4 Survival rate Log−rank P = 0.005 0.2 HR = 1.84 (95%CI, 1.2−2.8)

0.0 0 5 10 15

Survival years (OS)

High exp (N = 271) Low exp (N = 271) (s)

Figure 4: Identification of DEGs related to poor prognosis in EC. (a–s) Among these genes, higher expressions of BEX4, ERFE, FBXO17, GLDC, MAL2, MCM10, METTL7B, OAS1, POLQ, PSAT1, PSD3, PTTG1, RAPGEF4, SIMC1, SRPX, TIMP3, TRO, TRPC1, and UST were correlated to shorter OS time in patients with EC. the DEG expression with the overall survival (OS) time in in the high-risk group of EC patients, in comparison with the EC. Finally, we identified 30 DEGs that were related to low-risk group of EC patients by Kaplan-Meier Plotter analysis the prognosis of EC, including KCNK6, IL32, FAM189A2, (Figure 6(d)). We applied time-dependent receiver operating WFS1, GREB1, WFDC1, PGR, PAMR1, TRPC6, ADAM28, characteristic (ROC) analysis to assess the sensitivity and spec- ANKRD22, GLDC, RAPGEF4, MCM10, TRO, OAS1, BEX4, ificity of the prognostic model. And our results indicated that PSAT1, METTL7B, TIMP3, FBXO17, PTTG1, POLQ, theareaundertheROCcurve(AUC)was0.757for1-year, MAL2, SIMC1, ERFE, TRPC1, SRPX, UST, and PSD3. 0.758 for 3-year, 0.798 for 5-year, and 0.735 for 10-year survival Among these genes, higher expressions of BEX4, ERFE, (Figure 6(e)). FBXO17, GLDC, MAL2, MCM10, METTL7B, OAS1, POLQ, PSAT1, PSD3, PTTG1, RAPGEF4, SIMC1, SRPX, TIMP3, 3.5. Genetic Alteration Differences of Prognostic Genes in EC TRO, TRPC1, and UST were correlated to shorter OS time Patients. Furthermore, genetic alteration of prognostic genes in patients with EC (Figures 4(a)–4(s)). However, higher in EC was analyzed using the cBioPortal database, which expressions of WFS1, GREB1, FAM189A2, ANKRD22, included 726 patients from seven related studies. We WFDC1, TRPC6, KCNK6, IL32, PGR, PAMR1, and observed that the mutation rates of prognostic genes for EC ADAM28 were correlated to longer OS time in patients with ranged from 0.8% to 10% for individual genes (MAL2, 5%; EC (Figures 5(a)–5(k)). ANKRD22, 4%; METTL7B, 2.5%; IL32, 2.6%; ERFE, 0.8%; OAS1, 3%; TRPC1, 9%; SRPX, 6%; RAPGEF4, 10%; PSD3, 3.4. Establishing a Prognostic Gene Model in the TCGA 10%; SIMC1, 7%; TRPC6, 7%; WFS1, 6%; PGR, 7%; PAMR1, Cohort. We utilized the least absolute shrinkage and selection 7%; KCNK6, 4%; and FAM189A2, 5%). Among these genes, operator (LASSO) Cox regression analysis to establish the RAPGEF4 and PSD3 were found to have the highest muta- prognostic gene model. Figures 6(a) and 6(b) revealed a 17- tion rate in EC, which are mutated in about 10% EC cases gene signature constructed in the light of the optimum λ value. (Figure 7). We calculated the risk score as follows: risk score = ð4e − 04Þ ∗ MAL2 + ð−0:0263Þ ∗ ANKRD22 + ð0:0493Þ ∗ METTL7B 3.6. Validation of 17 Prognosis-Related Hub Gene Expressions + ð0:0688Þ ∗ IL32 + ð0:0022Þ ∗ ERFE + ð0:01Þ ∗ OAS1 + in EC. For verification of the bioinformatics analysis data in- ð0:0745Þ ∗ TRPC1 + ð0:1564Þ ∗ SRPX + ð0:4778Þ ∗ RAPGEF depth, UALCAN databases were used to confirm our find- 4+ð0:0496Þ ∗ PSD3 + ð0:0383Þ ∗ SIMC1 + ð−0:3016Þ ∗ ings. As presented in Figure 7, compared to normal samples, TRPC6 + ð−0:2001Þ ∗ WFS1 + ð−0:0341Þ ∗ PGR + ð−0:0821Þ OAS1, MAL2, ANKRD22, METTL7B, and IL32 were dra- ∗ PAMR1 + ð−0:0912Þ ∗ KCNK6 + ð−0:0813Þ ∗ FAM189A2. matically upregulated in EC samples, whereas TRPC1, SRPX, 543 patients with EC were divided equally into the low-risk RAPGEF4, PSD3, SIMC1, TRPC6, WFS1, PGR, PAMR1, group and the high-risk group on the basis of the median score KCNK6, and FAM189A2 were greatly downregulated in EC calculated by the risk score formula. Compared to patients in samples (Figures 8(a)–8(o)). the low-risk group, those in the high-risk group displayed a We also analyzed the correlation among these 17 larger death toll and a shorter survival time (Figure 6(c)). Our prognosis-related hub genes in EC. Our results showed that data revealed that an obviously lower OS time was observed the expressions of the 17 prognosis-related hub genes were BioMed Research International 11

FAM189A2 GREB1 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate P Log−rank = 0.001 P 0.2 0.2 Log−rank = 0.002 HR = 0.482 (95%CI, 0.311−0.747) HR = 0.504 (95%CI, 0.325−0.781) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (a) (b) ANKRD22 WFS1 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate Log−rank P = 0.007 Log−rank P = 0.001 0.2 0.2 HR = 0.56 (95%CI, 0.366−0.856) HR = 0.493 (95%CI, 0.319−0.76) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (c) (d) TRPC6 WFDC1 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate

Log−rank P = 0.006 0.2 0.2 Log−rank P = 0.002 HR = 0.548 (95%CI, 0.357−0.844) HR = 0.505 (95%CI, 0.328−0.777) 0.0 0.0

0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (e) (f)

Figure 5: Continued. 12 BioMed Research International

IL32 KCNK6 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate Log−rank P = 0.001 0.2 Log−rank P = 0.001 0.2 HR = 0.473 (95%CI, 0.305−0.732) HR = 0.474 (95%CI, 0.307−0.731) 0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (g) (h) PAMR1 PGR 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 Survival rate Survival rate P Log−rank P = 0.006 Log−rank = 0.006 0.2 0.2 HR = 0.547 (95%CI, 0.355−0.845) HR = 0.541 (95%CI, 0.351−0.835)

0.0 0.0 0 5 10 15 0 5 10 15

Survival years (OS) Survival years (OS) (i) (j) ADAM28 1.0

0.8

0.6

0.4 Survival rate

0.2 Log−rank P = 0.007 HR = 0.55 (95%CI, 0.358−0.847) 0.0 0 5 10 15

Survival years (OS)

High exp (N = 271) Low exp (N = 271) (k)

Figure 5: Identification of DEGs related to good prognosis in EC. (a–k) Higher expressions of WFS1, GREB1, FAM189A2, ANKRD22, WFDC1, TRPC6, KCNK6, IL32, PGR, PAMR1, and ADAM28 were correlated to longer OS time in patients with EC. significantly correlated based on Pearson correlation. The OAS1-TRPC6, WFS1-ANKRD22, and ANKRD22-WFS1. most significantly negatively correlated gene pairs included And the most significantly positively correlated gene pairs FAM189A2-MAL2, MAL2-FAM189A2, TRPC6-OAS1, included SIMC1-MAL2, MAL2-SIMC1, PGR-WFS1, WFS1- BioMed Research International 13

0 1217212629 29 28 27 26 26 24 22 20 17 16 15 12 5 3 2 RAPGEF4 13.2 0.5 SRPXTRPC1 PTTG1PSD3 12.8 ADAM28OAS1ERFEMETTL7BTIMP3SIMC1 ADAMPSAT1POLQGLDCMAL2WFDC1 28 0.0 KCNK6USTBEX4TROPGRGREB1 IL32ANKRD22 FAM189A2FBXO17PAMR1

Coefcients MCM10 12.4 −0.5 WFS1 TRPC6 Partial likelihood deviance –7 −6 −5 −4 −3 012345 λ L1 Norm Log ( ) (a) (b)

1.00 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ +++++ ++++++++++++++++++++++++++++++++++++ +++++ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 0 ++++++++++++++++++++++++++++ +++++++++++++++++++++++++++++++++++++++++++++++++++++++ +++++++++++++++ ++++++++++++++++++ 0.75 +++++++++++++++++ +++++ + +++++++++++++++++++++++ ++++++++++++++ –1 +++++++++++++++++++++++++++++++++++ 0.50 + –2

Risk score +++ + + 0.25 p –3 < 0.0001

Overall survival probability Median time : 8.5 0.00 Groups = High exp 271 44 3 1 0 Risk type Groups = Low exp 270 68 4 1 0 0 5 10 15 20 High_risk Low_risk Time (years) + Groups = High exp + Groups = Low exp 15 (d) 10 1.00 Time

5 0.75

0 0.50

Status 0.25 Alive True positive fraction Dead z−score of expression 0.00 FAM189A2 0.00 0.25 0.50 0.75 1.00 KCNK6 PAMR1 PGR 2 WFS1 False positive fraction TRPC6 SIMC1 1 PSD3 Type RAPGEF4 0 SRPX TRPC1 1 year, AUC = 0.757, 95%CI (0.668–0.847) OAS1 –1 ERFE 10 years, AUC = 0.735, 95%CI (0.557–0.912) IL32 METTL7B –2 ANKRD22 3 years, AUC = 0.758, 95%CI (0.697–0.818) MAL2 5 years, AUC = 0.798, 95%CI (0.738–0.858) (c) (e)

Figure 6: Establishing a prognostic gene model in the TCGA cohort. (a) LASSO coefficients profiles of 30 prognostic genes in EC. (b) LASSO regression with tenfold cross-validation obtained 17 prognostic genes using minimum lambda value. (c) 543 patients with EC were divided equally into low-risk group and high-risk group on the basis of the median score calculated by the risk score formula. (d) Lower OS time was observed in the high-risk group of EC patients, in comparison with the low-risk group of EC patients by Kaplan-Meier Plotter analysis. (e) Time-dependent receiver operating characteristic (ROC) analysis to assess the sensitivity and specificity of the prognostic model.

PGR, RAPGEF4-TRPC1, TRPC1-RAPGEF4, RAPGEF4- SRPX, SRPX-TRPC6, PAMR1-RAPGEF4, RAPGEF4- SRPX, SRPX-RAPGEF4, IL32-ANKRD22, ANKRD22-IL32, PAMR1, ERFE-METTL7B, METTL7B-ERFE, FAM189A2- FAM189A2-PGR, PGR-FAM189A2, FAM189A2-KCNK6, WFS1, WFS1-FAM189A2, PSD3-RAPGEF4, RAPGEF4- KCNK6-FAM189A2, SRPX-TRPC1, TRPC1-SRPX, TRPC6- PSD3, OAS1-MAL2, and MAL2-OAS1 (Figure 9). 14 BioMed Research International

Study of origin # samples per patient Profled for copy number alterations Profled for mutations Profled for structural variants ⁎ MAL2 5% ⁎ ANKRD22 4% ⁎ METTL7B 2.5% ⁎ IL32 2.6% ⁎ ERFE 0.8% ⁎ OAS1 3% ⁎ TRPC1 9% ⁎ SRPX 6% ⁎ RAPGEF4 10% ⁎ PSD3 10% ⁎ SIMC1 7% ⁎ TRPC6 7% ⁎ WFS1 6% ⁎ PGR 7% ⁎ PAMR1 7% ⁎ KCNK6 4% ⁎ FAM189A2 5%

Genetic alteration Inframe mutation (unknown signifcance) Missense mutation (unknown signifcance) Splice mutation (unknown signifcance) Truncating mutation (unknown signifcance) Structural variant Amplifcation (unknown signifcance) Deep deletion (unknown signifcance) No alterations Not profled

Study of origin Endometrial cancer (MSK, 2018) Uterine Corpus Endometrial Carcinoma (TCGA, Pan-Cancer Atlas) # samples per patient 03 Profled for copy number Yes No alterations Profled for mutations Yes No Profled for structural variants Yes No

Figure 7: Genetic alteration differences of prognostic genes in EC patients. The mutation rates of prognostic genes for EC ranged from 0.8% to 10% for individual genes (MAL2, 5%; ANKRD22, 4%; METTL7B, 2.5%; IL32, 2.6%; ERFE, 0.8%; OAS1, 3%; TRPC1, 9%; SRPX, 6%; RAPGEF4, 10%; PSD3, 10%; SIMC1, 7%; TRPC6, 7%; WFS1, 6%; PGR, 7%; PAMR1, 7%; KCNK6, 4%; and FAM189A2, 5%).

4. Discussion cGMP signaling pathway participates in the relaxation of the smooth muscle. cGMP-dependent PKG, which is essen- Emerging studies revealed that the occurrence of EC resulted tial for reducing cytoplasmic calcium and muscle tension, from the abnormally expressed multiple carcinoma-related was the downstream target of the NO-cGMP pathway [35]. genes [29, 30], amid which have been shown to display a rela- PKG was responsible for controlling the uterine smooth tionship to EC’s susceptibility and progression [29, 30]. muscle tone which produced force near menstruation and Many molecular biology methods have been used to identify regulated blood flow to the endometrial lining. The above biomarkers of cancers [31–34]. Nevertheless, the majority of data together confirmed that PKG functioned crucially in them merely concentrated on a single genetic factor, limiting controlling the contraction of the uterine and vascular these biomarkers’ reliability. smooth muscle during the periodical menstruation [35]. Our current study discovered more EC-related genes with PI3K-AKT signaling is one of the most important pathways differential expression than previous researches, indicating in our study. PI3K-AKT signaling could be antagonized by they may play importantly in the mechanism of EC. Our data the tumor suppressor phosphatase and tensin homolog revealed 114 common genes with upregulation and 308 com- (PTEN) which was reported to be usually mutated in several mon genes with downregulation in EC samples in comparison sorts of neoplasms, such as the endometrium, skin, brain, with normal samples. Bioinformatics analysis showed that and prostate cancers [36–38]. PTEN has a powerful phospha- these genes were significantly correlated to multiple key sig- tase activity, which is the best characterized physiological naling in EC, such as the cGMP-PKG signaling pathway. Fur- function leading to the tumor suppressor function of PTEN. thermore, we identified that 30 DEGs were related to the The IL17 signaling pathway was also one of the impor- prognosis of EC, comprising KCNK6, IL32, FAM189A2, tant pathways detected here. In inflammatory mediators, WFS1, GREB1, WFDC1, PGR, PAMR1, TRPC6, ADAM28, more and more evidence emphasizes the role of the ANKRD22, GLDC, RAPGEF4, MCM10, TRO, OAS1, BEX4, interleukin-17 (IL17) cytokine family in malignant diseases. PSAT1, METTL7B, TIMP3, FBXO17, PTTG1, POLQ, IL17 is becoming a crucial cytokine to promote and develop MAL2, SIMC1, ERFE, TRPC1, SRPX, UST, and PSD3. carcinomas by maintaining a chronic inflammatory micro- Here, we conducted GO and KEGG analyses of the environment which is conducive to tumor formation [39, involved biological processes and pathways related to these 40]. While IL17 may regulate chemokines and cytokines in DEGs in EC’s progression. The pathway analysis of these gynecologic cancers, Toll-like receptors may function impor- DEGs showed that the interconnected network of genes par- tantly in the gynecologic carcinomas’ development via trig- ticipated in the cyclic guanosine monophosphate- (cGMP-) gering an inflammatory response and cell survival in the protein kinase G (PKG) signaling pathway. As previously microenvironment of the tumor [41]. described, the contractility of the uterine smooth muscle is Our study generated a signature featuring17 prognosis- of importance for the periodic shedding of the endometrial related genes (MAL2, ANKRD22, METTL7B, IL32, ERFE, lining and the expulsion of the fetus during parturition. OAS1, TRPC1, SRPX, RAPGEF4, PSD3, SIMC1, TRPC6, There was one study showing that the nitric oxide- (NO-) WFS1, PGR, PAMR1, KCNK6, and FAM189A2) and BioMed Research International 15

125 250 12 MAL2 OAS1 ANKRD22 100 200 10 8 75 150 6 50 100 4 25 50

Transcript per million 2 Transcript per million Transcript per million 0 0 0 Normal EC Normal EC Normal EC (a) (b) (c) 30 600 50 FAM189A2 METTL7B IL32 25 500 40 20 400 30 15 300 20 10 200 10 Transcript per million Transcript per million 5 Transcript per million 100 0 0 0 Normal EC Normal EC Normal EC (d) (e) (f) 250 250 200 WFS1 PAMR1 PGR 200 200 150 150 150 100 100 100 50 50 50 Transcript per million Transcript per million Transcript per million 0 0 0 Normal EC Normal EC Normal EC (g) (h) (i) 12 60 14 PSD3 KCNK6 RAPGEF4 10 50 12 10 8 40 8 6 30 6 4 20 4 Transcript per million Transcript per million 2 Transcript per million 10 2 0 0 0 Normal EC Normal EC Normal EC (j) (k) (l) 10 60 20 TRPC6 SRPX TRPC1 8 50 15 40 6 30 10 4 20 5 2 Transcript per million Transcript per million Transcript per million 10 0 0 0 Normal EC Normal EC Normal EC (m) (n) (o)

Figure 8: Validation of 17 prognosis-related hub gene expressions in EC. demonstrated that they were utilized as predictors of OS in lators in the endometrium. Previous studies have shown that EC patients. We obtained many genes that were previously TRPC1 and TRPC6 were highly expressed in the entire reported to be involved in endometriosis patients or endome- endometrium during the periodical menstruation. Addition- trial stromal cells. For instance, the members of the transient ally, TRPV2, TRPV4, TRPC1/4, and TRPC6 were found in receptor potential (TRP) ion channel superfamily, known as human endometrial stromal cells (hESCs) from patients having the calcium permeability, has become pivotal modu- with endometriosis [25]. Previous reports suggested that 16 BioMed Research International

Corr 1.0 OAS1 1

MAL2 1 0.73

SIMC1 1 0.51 0.25

ERFE 1 0.25 0.34 0.4 0.5 METTL7B 1 0.62 0.38 0.37 0.19

IL32 1 0.36 0.28 −0.45 0.06 0.11

ANKRD22 1 0.54 0.17 0.17 −0.05 0.29 −0.09

TRPC6 1 0.2 −0.13 0.05 −0.26 0.26 −0.07 −0.51 0.0 SRPX 1 0.6 −0.01 −0.18 0.33 0.14 0.45 0.08 −0.08

PSD3 1 0.43 0.48 0.32 −0.06 0 0.15 0.24 0.06 −0.18

RAPGEF4 1 0.69 0.52 0.33 0.22 −0.12 −0.01 0.21 0.3 0.29 0.18

PAMR1 1 0.62 0.28 0.28 0.11 0.18 0.3 0.07 0.13 −0.12 −0.19 −0.08

TRPC1 1 0.48 0.52 0.08 0.57 0.12 −0.27 −0.28 0.08 −0.02 0.08 0.02 0.12 −0.5

FAM189A2 1 0.26 0.33 −0.01 0.06 −0.03 0.04 −0.47 −0.12 −0.42 −0.46 −0.49 −0.75 −0.46

WFS1 1 0.64 0.16 0.11 0.2 0.05 −0.05 0 −0.51 −0.34 −0.44 −0.23 −0.23 −0.48 −0.12

PGR 1 0.51 0.54 −0.09 0.46 0.3 0.27 −0.22 0.11 0.04 −0.01 −0.4 −0.17 −0.26 −0.41 −0.3 −1.0 KCNK6 1 0.49 0.22 0.55 −0.07 0.2 −0.07 0.17 0.05 0.23 0.19 0.33 −0.05 −0.25 −0.33 −0.42 −0.31

PGR IL32 WFS1 PSD3 SRPX ERFE MAL2 OAS1 KCNK6 TRPC1 PAMR1 TRPC6 SIMC1 RAPGEF4 FAM189A2 ANKRD22 METTL7B

Figure 9: 17 prognosis-related hub gene expressions were correlated to each other in EC. The expressions of 17 prognosis-related hub genes were significantly correlated based on Pearson correlation. the cAMP2-activated exchange protein (EPAC2, RAP- TRPC6, WFS1, PGR, PAMR1, KCNK6, and FAM189A2. GEF4), another cAMP mediator, took part in endometrial And correlation analysis showed that our results showed that stromal cell differentiation via regulating calreticulin the expressions of 17 prognosis-related hub genes were sig- (CALR) expression [42]. Compared with the control group, nificantly correlated based on Pearson correlation. the level of interleukin-32 (IL32) in peritoneal fluid (PF) in However, there are still several limitations in our litera- women with endometriosis was significantly higher. The ture. Firstly, the number of samples is limited, which should endometrial cells treated with IL32 in vitro significantly be enlarged in the following study, and all the samples are enhanced cell viability, proliferation, and invasion capabili- from public datasets; our own data is also very important. ties [43]. In silico methods can distinguish many key genes Secondly, we need to conduct more researches to expound related to the maintenance of telomeres, which were the function and potential mechanisms of these promising unknown to the occurrence and prognosis of EC before, biomarkers in the progression of EC. including WFS1. Prognostic biomarkers of EC are essential for ameliorating risk assessment before and after surgery 5. Conclusion and making guided- treatment decisions. PGR and PTEN were one of the most clinically valuable EC prognostic bio- In summary, our findings revealed 114 common genes with markers. In our research, we showed that significant genes upregulation and 308 common genes with downregulation with upregulation in EC samples included OAS1, MAL2, in EC samples relative to normal ones. Bioinformatics analy- ANKRD22, METTL7B, and IL32 after comparison with nor- sis showed these genes exhibited a significant relationship to mal samples. Obvious genes with downregulation in EC sam- multiple signaling and biological processes, such as the ples comprised TRPC1, SRPX, RAPGEF4, PSD3, SIMC1, cGMP-PKG signaling pathway and PI3K-AKT signaling. BioMed Research International 17

Moreover, we constructed a 17-gene signature to make a pre- endometrial carcinoma are associated with histological sub- diction of OS in EC patients using the TCGA cohorts. We type,” Modern Pathology, vol. 15, no. 10, pp. 1032–1037, 2002. collectively supplied a potential gene signature for the predic- [12] S. A. Saidi, C. M. Holland, D. P. Kreil et al., “Independent com- tion of EC patients’ prognosis. ponent analysis of microarray data in the study of endometrial cancer,” Oncogene, vol. 23, no. 39, pp. 6677–6683, 2004. Data Availability [13] J. I. Risinger, G. L. Maxwell, G. V. Chandramouli et al., “Microarray analysis reveals distinct gene expression profiles ff ” All the data and material were presented in GSE106191 and among di erent histologic types of endometrial cancer, Can- cer Research, vol. 63, no. 1, pp. 6–11, 2003. TCGA (https://portal.gdc.cancer.gov/). [14] S. Engebretsen and J. Bohlin, “Statistical predictions with glmnet,” Clinical Epigenetics, vol. 11, no. 1, p. 123, 2019. Conflicts of Interest [15] C. Gu, X. Shi, W. Qiu et al., “Comprehensive analysis of the fl prognostic role and mutational characteristics of m6A- The authors declare that there is no con ict of interests. related genes in lung squamous cell carcinoma,” Frontiers in Cell and Development Biology, vol. 9, p. 661792, 2021. Authors’ Contributions [16] I. Diboun, L. Wernisch, C. A. Orengo, and M. Koltzenburg, “Microarray analysis after RNA amplification can detect pro- Anna Wang is a co-first author. nounced differences in gene expression using limma,” BMC Genomics, vol. 7, no. 1, p. 252, 2006. “ References [17] C. Gu, J. Chen, X. Dang et al., Hippo pathway core genes based prognostic signature and immune infiltration patterns in lung squamous cell carcinoma,” Frontiers in oncology, [1] D. W. Bell and L. H. Ellenson, “Molecular genetics of endome- vol. 11, no. 1486, 2021. trial carcinoma,” Annual Review of Pathology, vol. 14, no. 1, “ pp. 339–367, 2019. [18] C. Gu, X. Shi, Z. Huang et al., A comprehensive study of con- “ 6 fi struction and analysis of competitive endogenous RNA net- [2] C. Gu, X. Shi, C. Dai et al., RNA m A modi cation in cancers: ” molecular mechanisms and potential clinical applications,” works in lung adenocarcinoma, Biochimica et Biophysica The Innovation, vol. 1, no. 3, p. 100066, 2020. Acta (BBA)-Proteins and Proteomics, vol. 1868, no. 8, p. 140444, 2020. [3] C. Gaber, R. Meza, J. J. Ruterbusch, and M. L. Cote, “Endome- “ trial cancer trends by race and histology in the USA: projecting [19] B. T. Sherman and R. A. Lempicki, Systematic and integrative ” analysis of large gene lists using DAVID bioinformatics the number of new cases from 2015 to 2040, Journal of racial ” – and ethnic health disparities, vol. 4, no. 5, pp. 895–903, 2016. resources, Nature Protocols, vol. 4, no. 1, pp. 44 57, 2009. “ [4] F. S. Saghir, I. M. Rose, A. Z. Dali, Z. Shamsuddin, A. R. Jamal, [20] X. Shi, T. Huang, J. Wang et al., Next-generation sequencing fi and N. M. Mokhtar, “Gene expression profiling and cancer- identi es novel genes with rare variants in total anomalous ” related pathways in type I endometrial carcinoma,” Interna- pulmonary venous connection, eBioMedicine, vol. 38, – tional Journal of Gynecological Cancer, vol. 20, no. 5, pp. 217 227, 2018. pp. 724–731, 2010. [21] J. Liu, S. Nie, J. Liang et al., “Competing endogenous RNA net- ” [5] T. M. Kim, P. W. Laird, and P. J. Park, “The landscape of work of endometrial carcinoma: a comprehensive analysis, – microsatellite instability in colorectal and endometrial cancer Journal of Cellular Biochemistry, vol. 120, no. 9, pp. 15648 genomes,” Cell, vol. 155, no. 4, pp. 858–868, 2013. 15660, 2019. “ [6] F. Amant, P. Moerman, P. Neven, D. Timmerman, E. van Lim- [22] C. Gu, Z. Huang, X. Chen et al., TEAD4 promotes tumor bergen, and I. Vergote, “Endometrial cancer,” Lancet, vol. 366, development in patients with lung adenocarcinoma via ERK no. 9484, pp. 491–505, 2005. signaling pathway,” Biochimica et Biophysica Acta - Molecular [7] S. F. Lax, “Molecular genetic pathways in various types of Basis of Disease, vol. 1866, no. 12, p. 165921, 2020. endometrial carcinoma: from a phenotypical to a molecular- [23] X. Pan, C. Gu, R. Wang, H. Zhao, J. Yang, and J. Shi, “Trans- based classification,” Virchows Archiv, vol. 444, no. 3, manubrial osteomuscular sparing approach for resection of pp. 213–223, 2004. cervico-thoracic lesions,” Journal of Thoracic Disease, vol. 9, – [8] D. Black, R. A. Soslow, D. A. Levine et al., “Clinicopathologic no. 9, pp. 3062 3068, 2017. significance of defective DNA mismatch repair in endometrial [24] C. Gu, R. Wang, X. Pan et al., “Comprehensive study of prog- carcinoma,” Journal of Clinical Oncology, vol. 24, no. 11, nostic risk factors of patients underwent pneumonectomy,” pp. 1745–1753, 2006. Journal of Cancer, vol. 8, no. 11, pp. 2097–2103, 2017. [9] M. P. Hayes, H. Wang, R. Espinal-Witter et al., “PIK3CA and [25] Y. Sun, C. Gu, J. Shi et al., “Reconstruction of mediastinal ves- PTEN mutations in uterine endometrioid carcinoma and sels for invasive thymoma: a retrospective analysis of 25 cases,” complex atypical hyperplasia,” Clinical Cancer Research, Journal of Thoracic Disease, vol. 9, no. 3, pp. 725–733, 2017. vol. 12, no. 20, pp. 5932–5935, 2006. [26] C. Chen, X. Zhang, C. Gu et al., “Surgery performed at night by [10] S. F. Lax, B. Kendall, H. Tashiro, R. J. C. Slebos, and L. H. continuously working surgeons contributes to a higher inci- Ellenson, “The frequency of p53, K-ras mutations, and micro- dence of intraoperative complications in video-assisted thora- satellite instability differs in uterine endometrioid and serous coscopic pulmonary resection: a large monocentric carcinoma: evidence of distinct molecular genetic pathways,” retrospective study,” European Journal of Cardio-Thoracic Cancer, vol. 88, no. 4, pp. 814–824, 2000. Surgery, vol. 57, no. 3, pp. 447–454, 2020. [11] P. W. Schlosshauer, L. H. Ellenson, and R. A. Soslow, “β- [27] C. Gu, X. Pan, R. Wang et al., “Analysis of mutational and clin- Catenin and E-cadherin expression patterns in high-grade icopathologic characteristics of lung adenocarcinoma with 18 BioMed Research International

clear cell component,” Oncotarget, vol. 7, no. 17, pp. 24596– [42] K. Kusama, M. Yoshie, K. Tamura, K. Imakawa, and 24603, 2016. E. Tachikawa, “EPAC2-mediated calreticulin regulates LIF ” [28] J. Chen, C. Gu, X. Chen et al., “Clinicopathological and prog- and COX2 expression in human endometrial glandular cells, – nostic analyses of 86 resected pulmonary lymphoepithelioma- Journal of Molecular Endocrinology, vol. 54, no. 1, pp. 17 24, like carcinomas,” Journal of Surgical Oncology, vol. 123, no. 2, 2015. pp. 544–552, 2021. [43] M. Y. Lee, S. H. Kim, Y. S. Oh et al., “Role of interleukin-32 in [29] J. Liu, M. A. Eckert, B. T. Harada et al., “m6A mRNA methyl- the pathogenesis of endometriosis: in vitro, human and trans- ation regulates AKT activity to promote the proliferation and genic mouse data,” Human Reproduction, vol. 33, no. 5, tumorigenicity of endometrial cancer,” Nature Cell Biology, pp. 807–816, 2018. vol. 20, no. 9, pp. 1074–1083, 2018. [30] Y. Xiao, L. Jin, C. Deng et al., “Inhibition of PFKFB3 induces cell death and synergistically enhances chemosensitivity in endometrial cancer,” Oncogene, vol. 40, no. 8, pp. 1409–1424, 2021. [31] R. C. Arend, B. A. Jones, A. Martinez, and P. Goodfellow, “Endometrial cancer: molecular markers and management of advanced stage disease,” Gynecologic Oncology, vol. 150, no. 3, pp. 569–580, 2018. [32] C. Gu and C. Chen, “Methylation in lung cancer: a brief review,” Methods in Molecular Biology, vol. 2204, pp. 91–97, 2020. [33] Y. Bao, C. Gu, H. Xie et al., “Comprehensive study of neoadju- vant targeted therapy for resectable non-small cell lung can- cer,” Annals of Translational Medicine, vol. 9, no. 6, p. 493, 2021. [34] Y. Yuan, Q. Huang, C. Gu, and H. Chen, “Disease-free survival improved by use of adjuvant EGFR tyrosine kinase inhibitors in resectable non-small cell lung cancer: an updated meta- analysis,” Journal of Thoracic Disease, vol. 9, no. 12, pp. 5314–5321, 2017. [35] S. H. Francis, J. L. Busch, J. D. Corbin, and D. Sibley, “cGMP- dependent protein kinases and cGMP phosphodiesterases in nitric oxide and cGMP action,” Pharmacological Reviews, vol. 62, no. 3, pp. 525–563, 2010. [36] H. Koike, M. Nozawa, M. A. de Velasco et al., “Conditional PTEN-deficient mice as a prostate cancer chemoprevention model,” Asian Pacific Journal of Cancer Prevention, vol. 16, no. 5, pp. 1827–1831, 2015. [37] M. Shi, X. Zhou, Z. Zhang et al., “A novel PI3K inhibitor dis- plays potent preclinical activity against an androgen- independent and PTEN-deficient prostate cancer model estab- lished from the cell line PC3,” Toxicology Letters, vol. 228, no. 3, pp. 133–139, 2014. [38] X. Jiang, S. Chen, J. M. Asara, and S. P. Balk, “Phosphoinosi- tide 3-kinase pathway activation in phosphate and tensin homolog (PTEN)-deficient prostate cancer cells is indepen- dent of receptor tyrosine kinases and mediated by the p110β and p110δ catalytic subunits,” The Journal of Biological Chem- istry, vol. 285, no. 20, pp. 14980–14989, 2010. [39] L. Liu, F. Chen, A. Xiu, B. du, H. Ai, and W. Xie, “Identification of key candidate genes and pathways in endometrial cancer by integrated bioinformatical analysis,” Asian Pacific Journal of Cancer Prevention, vol. 19, no. 4, pp. 969–975, 2018. [40] H. A. Alshaker and K. Z. Matalka, “IFN-γ, IL-17 and TGF-β involvement in shaping the tumor microenvironment: the sig- nificance of modulating such cytokines in treating malignant solid tumors,” Cancer Cell International, vol. 11, no. 1, p. 33, 2011. [41] N. Husseinzadeh and S. M. Davenport, “Role of toll-like recep- tors in cervical, endometrial and ovarian cancers: a review,” Gynecologic Oncology, vol. 135, no. 2, pp. 359–363, 2014.