Skip to main content
NCBI home page
Science Progress logoLink to Science Progress
. 2021 Mar 4;104(1):0036850421997276. doi: 10.1177/0036850421997276

Identification of key genes associated with lung adenocarcinoma by bioinformatics analysis

Xinyu Wang 1,*, Jiaojiao Yang 2,*, Xueren Gao 1,✉
PMCID: PMC10454774  PMID: 33661044

Abstract

Lung adenocarcinoma (LUAD) is the most common histological type of lung cancer, comprising around 40% of all lung cancer. Until now, the pathogenesis of LUAD has not been fully elucidated. In the current study, we comprehensively analyzed the dysregulated genes in lung adenocarcinoma by mining public datasets. Two sets of gene expression datasets were obtained from the Gene Expression Omnibus (GEO) database. The dysregulated genes were identified by using the GEO2R online tool, and analyzed by R packages, Cytoscape software, STRING, and GPEIA online tools. A total of 275 common dysregulated genes were identified in two independent datasets, including 54 common up-regulated and 221 common down-regulated genes in LUAD. Gene Ontology (GO) enrichment analysis showed that these dysregulated genes were significantly enriched in 258 biological processes (BPs), 27 cellular components (CCs), and 21 molecular functions (MFs). Furthermore, protein-protein interaction (PPI) network analysis showed that PECAM1, ENG, KLF4, CDH5, and VWF were key genes. Survival analysis indicated that the low expression of ENG was associated with poor overall survival (OS) of LUAD patients. The low expression of PECAM1 was associated with poor OS and recurrence-free survival of LUAD patients. The cox regression model developed based on age, tumor stage, ENG, PECAM1 could effectively predict 5-year survival of LUAD patients. This study revealed some key genes, BPs, CCs, and MFs involved in LUAD, which would provide new insights into understanding the pathogenesis of LUAD. In addition, ENG and PECAM1 might serve as promising prognostic markers in LUAD.

Keywords: Gene, lung adenocarcinoma, bioinformatics, biomarker

Introduction

Lung adenocarcinoma (LUAD) is the most common histological type of lung cancer, comprising around 40% of all lung cancer. 1 Although progress has been made in the diagnosis and treatment of LUAD, the 5-year survival rate for LUAD patients is still very poor, which is partly due to the fact that LUAD is often diagnosed at advanced stages involving disseminated metastatic tumors. 2 In addition, the pathogenesis of LUAD has not been completely elucidated until now. Therefore, it is necessary to have a better understanding of the molecular mechanism of occurrence and development of LUAD and find specific biomarkers with prognostic significance, which will help to improve the treatment and survival rate of LUAD.

In recent years, the combination of high-throughput gene expression detection technologies (such as RNA-seq and microarray) and bioinformatics has become an effective measure to provide new insights into the occurrence and development of diseases, including LUAD.3–5 Zheng et al. 3 found that the expression level of SUV39H2 was up-regulated in LUAD tissues through bioinformatics analyses. SUV39H2 overexpression was significantly associated with its amplification and with shorter overall survival of LUAD patients. Subsequently, RNA-seq demonstrated that SUV39H2 might mediate tumorigenesis and metastasis of LUAD by regulating TPM4, OPTN, and STOM. Shi et al. 4 developed a survival prediction scoring model based on LUAD RNA-seq data in The Cancer Genome Atlas (TCGA) database. The model including 31 lncRNAs could predict overall survival (OS) of LUAD patients with high accuracy. Su et al. 5 identified key genes (such as CCL2, LYZ, and MMP2) and the essential miRNA, hsa-let-7d, related to LUAD brain metastases via microarray analysis of cDNA expression profiles. Chen et al. 6 provided profiles of the tumor immune microenvironment that had prognostic value for patients with locally advanced LUAD based on targeted RNA-Seq data and bioinformatics methods. Zhu et al. 7 revealed an important candidate gene (RN7SL494P) involved in both nodal metastasis and prognosis in LUAD by analyzing RNA-Seq data. These findings enhanced our understanding of genes involved in LUAD, but there were still hidden genes waiting to be revealed because of the complexity and heterogeneity of LUAD.

In the current study, we identified the dysregulated genes in lung adenocarcinoma based on gene expression profiles of 64 LUAD and 64 non-tumor lung tissues. Subsequent bioinformatic analyses were performed to explore biological processes (BPs), cellular components (CCs), molecular functions (MFs) and pathways enriched by the dysregulated genes and key genes invloved in LUAD. The related datasets from TCGA and Genotype-Tissue Expression (GTEx) databases were used to confirm the expression levels of key genes in LUAD and assess the relationships between the expression levels of key genes and the survival of LUAD patients. Taken together, this study will provide useful guidance for further research on various genes associated with the occurrence and development of LUAD.

Materials and methods

Microarray datasets

Two sets of gene expression datasets (GSE32863 and GSE118370) were obtained from the Gene Expression Omnibus (GEO) database. Samples from GSE32863 included 13 male and 45 female. Thereinto nine patients were younger than 60 years old. Samples from GSE118370 included three male and three female. Thereinto three patients were younger than 60 years old. GSE32863 including 58 LUAD and 58 adjacent non-tumor lung tissues, were generated based on Illumina HumanWG-6 v3.0 expression beadchip (GPL6884) and contributed by Ite Laird-Offringa. 8 The data had been processed by log2-transformation and Robust Spline Normalization (RSN) using the lumi package in R. GSE118370 including six invasive LUAD tissues and paired normal lung tissues, were generated based on Affymetrix Human Genome U133 Plus 2.0 Array (GPL570) and contributed by Xu et al. 9 The data had been were normalized by MAS 5.0 algorithm.

Identification of the dysregulated genes

GEO2R online tool (https://www.ncbi.nlm.nih.gov/geo/geo2r/) was used to identify the dysregulated genes in LUAD. The t test and Benjamini-Hochberg method were used to calculate the p-value and false discovery rate (FDR), respectively. The dysregulated genes were defined according to adjusted p-value (adj.P.Val) <0.05 and |log2FC| >1. The common dysregulated genes in GSE32863 and GSE118370 datasets were screened out by Venny 2.1 online tool (http://bioinfogp.cnb.csic.es/tools/venny/index.html).

Enrichment analysis of the dysregulated genes

To identify BPs, CCs, MFs and pathways closely related to LUAD, Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the dysregulated genes were performed using clusterProfiler package in R. 10 The adjusted p-value <0.05 was considered as significant enrichment.

Protein-protein interaction (PPI) network analysis of the dysregulated genes

The PPI network of the common dysregulated genes in the two datasets was constructed using the STRING database (https://string-db.org/) with default parameter values (Evidence was selected as meaning of network edges; Textmining, experiments, databases, co-expression, neighborhood, gene fusion and co-occurrence were selected as active interaction sources; Medium confidence (0.4) was selected as minimum required interaction score) and visualized using Cytoscape software.11,12 The degree of a node in PPI network was regarded as the number of interactions with other nodes. Key nodes (proteins/genes) in PPI network were selected according to degree >15.

Validation of the expression levels of key genes

To confirm the expression levels of key genes in LUAD, gene expression profiling interactive analysis (GEPIA) online tool (http://gepia.cancer-pku.cn/) was used to explore the related datasets in TCGA and GTEx databases, and analyze the expression levels of key genes in LUAD tissues compared with normal tissues. The method for differential analysis is one-way ANOVA. P-value <0.05 was considered as significant difference.

The relationships between the expression levels of key genes and the survival of LUAD patients

Based on TCGA database, GEPIA online tool was further used to analyze the relationships between the expression levels of key genes and the survival of LUAD patients. In survival analysis, the cut-off value of high/low expression groups was determined according to median value. P-value <0.05 was considered significant.

Development and evaluation of prognosis prediction model for LUAD patients

RNA-seq and survival data of LUAD patients were downloaded from UCSC Xena (http://xena.ucsc.edu/). Survival package in R was used to develop prognosis prediction model for LUAD patients. A nomogram predicting 3- and 5-year survival of LUAD patients was drawn by rms package. Receiver operating characteristic (ROC) curves was used to assess the discriminatory power of model and drawn by timeROC package. Area under curve (AUC) ≥0.7 was considered as an effective model.

Results

Identification and analysis of the dysregulated genes

In GSE118370 datasets, a total of 893 dysregulated genes including 226 up-regulated and 667 down- regulated genes in LUAD tissues were identified (Figure 1). In GSE32863 datasets, a total of 1263 dysregulated genes including 511 up-regulated and 752 down-regulated genes in LUAD tissues were identified (Figure 1). Venn diagrams showed that 275 dysregulated genes including 54 up-regulated and 221 down-regulated genes in LUAD tissues were overlapped in the two datasets (Figure 2 and Table 1).

Figure 1.

Figure 1.

Open in a new tab

Volcano plot comparing all of the differentially expressed genes in two different datasets (The differentially expressed genes were selected according to adjusted p-value <0.05 and |log2FC| > 1).

Figure 2.

Figure 2.

Open in a new tab

Venn diagrams showing the dysregulated genes in two sets of gene expression datasets: (a) the upregulated genes in two datasets and (b) the down-regulated genes in two datasets.

Table 1.

Common dysregulated genes in GSE32863 and GSE118370 datasets.

Gene ID log2FC(GSE32863) log2FC(GSE118370)
CDH5 1003 −2.42 −2.50
RTKN2 219790 −1.11 −3.80
SLC6A4 6532 −1.92 −6.18
EDNRB 1910 −2.38 −3.38
CCM2L 140706 −1.94 −2.80
SEMA6A 57556 −2.18 −3.39
PTPRB 5787 −1.73 −2.85
CLEC1A 51267 −1.6 −2.58
GIMAP8 155038 −1.64 −3.19
RHOJ 57381 −1.45 −2.44
SDPR 8436 −2.72 −2.67
VIPR1 7433 −1.99 −3.04
ADARB1 104 −1.66 −2.68
PECAM1 5175 −2.27 −2.23
PRX 57716 −1.11 −3.62
TBX3 6926 −1.13 −3.43
CLDN18 51208 −4.79 −3.18
NOTCH4 4855 −1.15 −2.20
MME 4311 −2.27 −4.43
NDRG4 65009 −1.59 −2.83
TIMP3 7078 −2.37 −2.40
SPOCK2 9806 −2.93 −2.02
KANK3 256949 −1.51 −3.32
FHL1 2273 −2.46 −2.96
SASH1 23328 −1.68 −1.83
TCF21 6943 −2.67 −2.71
GRK5 2869 −1.99 −1.91
JAM2 58494 −1.88 −2.53
STXBP6 29091 −1.47 −3.78
GSTM5 2949 −1.06 −2.53
SH2D3C 10044 −1.93 −2.50
CD36 948 −2.9 −2.57
EMCN 51705 −1.7 −3.45
FOXF1 2294 −1.62 −2.71
TMEM74B 55321 −1.73 −2.62
LIMS2 55679 −1.54 −1.59
CAV2 858 −2.83 −1.95
SPTBN1 6711 −1.68 −2.35
TEK 7010 −2.71 −2.44
TMEM100 55273 −3.05 −4.47
FEZ1 9638 −2.17 −1.94
FAM107A 11170 −3.87 −2.91
CAV1 857 −3.55 −2.37
FGFR4 2264 −1.67 −2.37
IL7R 3575 −1.98 −2.23
EPB41L2 2037 −1.28 −1.70
NDRG2 57447 −1.58 −2.00
ADCY4 196883 −1.32 −2.08
TBX2 6909 −1.18 −2.05
FAM65A 79567 −1.44 −1.47
ARHGEF6 9459 −1.6 −1.58
AGER 177 −4.31 −2.81
ANKRD29 147463 −1.62 −2.99
SEMA5A 9037 −1.26 −2.24
LAMA3 3909 −1.01 −2.68
CALCRL 10203 −1.96 −2.97
ACVRL1 94 −1.72 −2.22
HEG1 57493 −1.73 −2.07
SVEP1 79987 −2.44 −3.97
S1PR1 1901 −1.71 −2.07
RASIP1 54922 −1.93 −2.52
ADRB2 154 −2 −2.27
PTRF 284119 −1.66 −1.91
SLIT2 9353 −1.75 −2.66
CCDC50 152137 −1.33 −1.81
GBP4 115361 −1.43 −2.35
ABI3BP 25890 −2.21 −2.06
LAMA4 3910 −1.31 −3.03
FERMT2 10979 −1.36 −1.67
ECSCR 641700 −1.14 −1.97
C10orf67 256815 −1.21 −1.50
KLF4 9314 −2.23 −2.24
MSRB3 253827 −1.47 −1.90
GPX3 2878 −2.11 −2.23
NKG7 4818 −1.38 −1.92
TMEM47 83604 −1.64 −2.12
PDK4 5166 −2.2 −2.21
ID2 3398 −1.95 −1.65
GPC3 2719 −2.32 −2.18
MYL9 10398 −1.15 −1.76
CAMK2N1 55450 −1.15 −1.79
CA2 760 −2.24 −2.18
SLCO2A1 6578 −1.41 −2.31
LRRC32 2615 −1.75 −1.60
FGD5 152273 −1.81 −2.14
EPAS1 2034 −2.15 −2.19
ZNF331 55422 −1.03 −1.63
HHIP 64399 −1.1 −2.83
TGFBR3 7049 −2.2 −2.99
HSPC324 101928612 −1.43 −2.05
EML1 2009 −1.16 −2.22
FABP4 2167 −3.68 −4.82
CASP1 834 −1.24 −1.60
RGCC 28984 −2.74 −2.03
PDE5A 8654 −1.8 −2.06
SMAD6 4091 −2.21 −2.68
CLEC14A 161198 −2.35 −1.86
DACH1 1602 −1.62 −2.57
UPK3B 80761 −1.54 −1.43
LTBP4 8425 −1.75 −2.08
SH3GL3 6457 −1.16 −3.36
B3GALNT1 8706 −1.21 −1.52
COL6A6 131873 −1.67 −2.32
PDLIM3 27295 −1.32 −2.41
STARD13 90627 −1.02 −2.11
STARD8 9754 −1.07 −1.80
VWF 7450 −2.23 −2.24
CDO1 1036 −1.54 −2.71
LMCD1 29995 −1.63 −1.92
MCEMP1 199675 −4.24 −2.23
RAMP3 10268 −2.27 −2.52
HSPB6 126393 −1.45 −2.38
CD93 22918 −2.3 −1.43
FMO2 2327 −3.12 −2.51
ESAM 90952 −1.83 −1.86
ARAP3 64411 −1.11 −1.31
LDB2 9079 −2.71 −2.35
TACC1 6867 −1.74 −1.39
LEPR 3953 −2.19 −1.96
SOX7 83595 −1.43 −2.58
ITM2A 9452 −2.37 −1.83
KLF13 51621 −1.2 −2.47
GHR 2690 −1.37 −3.10
DUOXA1 90527 −1.36 −1.93
LIMCH1 22998 −1.57 −1.27
AQP4 361 −2.52 −2.01
STX11 8676 −2.36 −2.44
GJA4 2701 −1.31 −1.22
PLPP3 8613 −2.02 −1.71
MYH11 4629 −1.99 −1.74
TM6SF1 53346 −1.14 −1.40
ANGPT1 284 −1.92 −2.03
CYYR1 116159 −2.01 −2.07
TCEAL2 140597 −1.61 −2.24
HOXA5 3202 −2.09 −1.83
LYVE1 10894 −2.69 −3.27
ZEB2 9839 −1.1 −1.47
ADGRL2 23266 −1.17 −2.84
SLIT3 6586 −1.62 −3.98
CPED1 79974 −1.73 −1.96
FGR 2268 −2.06 −1.66
GNLY 10578 −1.27 −3.07
ETS1 2113 −1.27 −1.12
MYH10 4628 −1.48 −1.63
AHNAK 79026 −1.34 −1.06
GNG11 2791 −2.63 −2.62
NTNG1 22854 −1.14 −4.37
FCN3 8547 −3.91 −4.35
IL1RL1 9173 −1.13 −2.66
MFAP4 4239 −3.28 −2.36
COX4I2 84701 −1.41 −1.42
PPP1R14A 94274 −1.93 −2.29
ADH1B 125 −2.67 −3.25
PRKCH 5583 −1.24 −1.32
SPARCL1 8404 −2.66 −1.45
SCARA5 286133 −1.83 −2.86
CRYAB 1410 −2.21 −1.74
FXYD6 53826 −1.69 −1.27
KLF9 687 −1.54 −1.91
LRRN3 54674 −1.24 −1.83
COX7A1 1346 −2.21 −1.71
FZD4 8322 −1.67 −1.23
SOX17 64321 −1.14 −3.61
ENG 2022 −1.2 −1.75
APOL3 80833 −1.21 −2.58
VGLL3 389136 −1.04 −2.50
CD300LG 146894 −2.34 −4.98
TMEM204 79652 −1.54 −1.26
CELF2 10659 −1.86 −1.23
MAOB 4129 −1.25 −2.36
GFOD1 54438 −1.14 −1.56
MAL 4118 −1.95 −1.75
PODXL 5420 −1.08 −1.12
FAM189A2 9413 −1.84 −1.44
TNNC1 7134 −2.61 −3.16
SYNM 23336 −1.07 −1.32
CLIC5 53405 −2.39 −2.22
SLC19A3 80704 −1.73 −4.92
AXIN2 8313 −1.02 −2.83
GATA2 2624 −1.15 −1.40
ZNF366 167465 −1.12 −2.34
AOC3 8639 −1.1 −2.15
OLFML2A 169611 −1.21 −1.31
TSPAN18 90139 −1.11 −1.39
PIK3R1 5295 −1.09 −1.17
SBSPON 157869 −1.44 −1.48
ACADL 33 −1.18 −3.47
SLC39A8 64116 −1.96 −1.53
JAML 120425 −1.68 −1.86
ALOX5 240 −2.3 −1.75
SFTPC 6440 −3.68 −3.57
CD52 1043 −2.71 −1.27
DKK3 27122 −1.51 −1.92
LHFP 10186 −1.67 −1.47
SOSTDC1 25928 −2.12 −1.97
ACSL4 2182 −1.18 −1.53
WFDC1 58189 −1.23 −1.65
ITPRIP 85450 −1.2 −1.23
PDZD2 23037 −1.18 −1.72
HIGD1B 51751 −2.52 −1.65
PPP1R3C 5507 −1.2 −1.78
PMP22 5376 −1.8 −1.43
FPR1 2357 −2.22 −3.29
EPB41L3 23136 −1.15 −1.30
OLFML1 283298 −1.58 −1.55
KANK2 25959 −1.53 −1.30
WISP2 8839 −1.78 −2.98
C1orf162 128346 −2.2 −1.29
GIMAP5 55340 −1.82 −1.31
HBEGF 1839 −2.1 −1.44
HYAL1 3373 −2.29 −1.83
DUOX1 53905 −2.07 −2.32
PLAC9 219348 −3.07 −1.77
HBB 3043 −4.15 −3.54
COL13A1 1305 −1.33 −2.06
DENND2A 27147 −1.22 −1.68
FBLN1 2192 −2.01 −1.72
C10orf54 64115 −1.57 −1.47
RGS5 8490 −1.09 −1.29
DPEP2 64174 −1.7 −1.74
ZAK 51776 −1.25 −1.15
BLACAT1 101669762 1.06 2.22
PROM2 150696 2.34 2.35
SPINK1 6690 3.16 4.75
COL10A1 1300 1.38 2.23
LAPTM4B 55353 1.23 2.41
TIMP1 7076 1.6 1.77
RASEF 158158 1.25 2.00
GOLM1 51280 1.82 1.53
SGPP2 130367 1.94 2.37
NQO1 1728 1.69 2.21
TMEM45B 120224 1.76 2.01
SULF1 23213 1.27 1.75
SPSB1 80176 1.11 1.44
AGR2 10551 1.26 2.40
SOX4 6659 1.29 1.36
FAM83H 286077 1.44 2.83
ZNF750 79755 1.22 1.98
HOOK1 51361 1.79 1.51
NME1 4830 1.47 1.75
HABP2 3026 1.17 2.49
UBE2T 29089 1.3 1.96
ST14 6768 1.58 1.59
SMPDL3B 27293 1.32 2.08
NET1 10276 1.1 1.40
KCNK5 8645 1.24 1.27
CD24 100133941 1.03 2.58
KCNQ3 3786 1.23 1.62
ZDHHC9 51114 1.25 1.88
TLCD1 116238 1.3 1.91
AGRN 375790 1.64 1.07
KRT6A 3853 1.07 3.63
ALDH18A1 5832 1.01 1.48
SLC39A11 201266 1.35 1.60
SEMA4B 10509 1.85 1.75
TOP2A 7153 2.52 1.78
NEK2 4751 1.13 2.04
CEACAM6 4680 1.3 2.52
GJB2 2706 1.06 3.06
GYG2 8908 1.1 4.05
GFPT1 2673 1.26 1.62
P4HB 5034 1.09 1.55
SPDEF 25803 1.94 3.06
TXNDC17 84817 1.23 1.61
GDF15 9518 1.15 3.61
TFAP2A 7020 1.84 2.57
RAB25 57111 1.12 1.65
HMGB3 3149 2.13 1.94
FUT2 2524 1.36 1.08
GGCT 79017 1.2 1.61
FAM83A 84985 1.98 3.26
ELF3 1999 1.3 1.68
SERINC2 347735 2.27 1.95
TMPRSS4 56649 1.46 3.90
EPCAM 4072 1.27 1.37
Open in a new tab

The common dysregulated genes in LUAD tissues were further explored by GO and KEGG pathway enrichment analysis. The results based on GO enrichment analysis showed that these genes were markedly enriched in 27 CCs, 258 BPs, and 21 MFs. The top five enriched GO terms were as follows: CCs (cell-cell junction, extracellular matrix, collagen-containing extracellular matrix, membrane raft, membrane microdomain) (Figure 3(a)), BPs (extracellular structure organization, heart morphogenesis, morphogenesis of an epithelium, cardiac chamber morphogenesis, extracellular matrix organization) (Figure 3(b)) and MFs (transforming growth factor beta binding, glycosaminoglycan binding, extracellular matrix structural constituent, cytokine binding, growth factor binding) (Figure 3(c)). However, no pathway was significantly enriched by these common dysregulated genes.

Figure 3.

Figure 3.

Open in a new tab

Bubble plots showing top five GO terms enriched by the common dysregulated genes: (a) cellular components (CCs), (b) biological processes (BPs), and (c) molecular functions (MFs).

As shown in Figure 4, there were 226 nodes and 463 edges in the PPI network. The average number of neighbors in the network was 4.097. Node degree analysis indicated that PECAM1 (platelet and endothelial cell adhesion molecule 1), CDH5 (cadherin 5), VWF (von Willebrand factor), ENG (endoglin), and KLF4 (Kruppel like factor 4) with degree >15 were key genes in the network.

Figure 4.

Figure 4.

Open in a new tab

PPI network of the common dysregulated genes (The interaction relationship was constructed based on textmining, experiments, databases, co-expression, neighborhood, gene fusion and co-occurrence. The minimum required interaction score is 0.4).

Validation of the expression levels of key genes

By merging the related data in TCGA and GTEx databases, we found that the expression levels of PECAM1, CDH5, VWF, ENG, and KLF4 were significantly down-regulated in LUAD tissues compared with normal tissues, which was consistent with the current microarray results (Figure 5).

Figure 5.

Figure 5.

Open in a new tab

The expression levels of key genes in LUAD tissues (TCGA tumors vs (TCGA normal + GTEx normal); num: Number; T: tumor; N: normal; Ordinate values represent log2 (TPM + 1)).

The relationship between the expression levels of key genes and the survival of LUAD patients

The relationship between the expression levels of PECAM1, CDH5, VWF, and ENG and the survival of LUAD patients was analyzed based on 478 patients. The relationship between the expression levels of KLF4 and the survival of LUAD patients was analyzed based on 477 patients. Results showed that low expression of ENG was associated with poor OS in LUAD patients. Low expression of PECAM1 was associated not only with poor OS but also with poor disease free survival (DFS) in LUAD patients (Figure 6).

Figure 6.

Figure 6.

Open in a new tab

The relationship between the expression levels of key genes and the survival of LUAD patients (The hazards ratio was calculated based on Cox PH Model. The 95% confidence interval was added as dotted line).

Development and evaluation of prognosis prediction model for LUAD patients

Potential prognostic factors including age, tumor stage, ENG, and PECAM1 were used to develop prognosis prediction model for LUAD patients (Figure 7(a)). ROC curve showed that the model could effectively predict 5-year survival of LUAD patients (AUC = 0.7) (Figure 7(b)).

Figure 7.

Figure 7.

Open in a new tab

A nomogram predicting 3- and 5-year survival of LUAD patients and the receiver operating characteristic (ROC) curve analysis for the survival prediction model: (a) nomogram and (b) ROC curve.

Discussion

With the development of high-throughput technologies, numerous datasets including mRNA, miRNA, and lncRNA expression profiles were generated and stored in public databases. Reanalysis of these datasets could obtain novel potential biomarkers of diseases. In the present study, we used bioinformatics methods and tools to retrieve the related datasets in GEO database and identify genes, BPs, CCs, and MFs closely related to LUAD. A total of two independent datasets (GSE32863 and GSE118370) were included in the current study. Differential gene expression analysis showed 275 common dysregulated genes (54 up-regulated and 221 down-regulated genes in LUAD tissues) in the two datasets. Using GO enrichment analysis, we found that these common dysregulated genes were significantly enriched in 27 CCs, 258 BPs, and 21 MFs, such as morphogenesis of an epithelium, transforming growth factor beta binding, cytokine binding, and glycosaminoglycan binding. By constructing the PPI network, a number of hub genes involved in LUAD were identified, including PECAM1, CDH5, VWF, ENG, and KLF4. PECAM1 gene located on chromosome 17q23.3 and encoded the protein which was found on the surface of monocytes, platelets, neutrophils, and some types of T-cells, and made up a large portion of endothelial cell intercellular junctions.13,14 Furthermore, the protein was a member of the immunoglobulin superfamily and likely participated in leukocyte migration, angiogenesis, and integrin activation.15–17 CDH5 gene encoded a classical cadherin of the cadherin superfamily and played a role in endothelial adherens junction assembly and maintenance. The aberrant CDH5 expression had been observed in tumor cells of various malignancies, including lung cancer.18–21 VWF gene was located on chromosome 12p13.31 and encoded a plasma glycoprotein that played a critical role in primary hemostasis. 22 In addition, several studies had also identified a series of additional biological functions for VWF. For instance, Starke et al. 23 described a novel role for VWF as a negative regulator of angiogenesis. Inhibition of VWF expression in endothelial cells increased vascular endothelial growth factor (VEGF) receptor-2 (VEGFR-2)-dependent proliferation and migration. VWF was important in modulating inflammatory responses, which was supported by data from several different animal models.24,25

ENG gene encoded a homodimeric transmembrane protein which was a major glycoprotein of the vascular endothelium. O’Leary et al. 26 identified ENG as an epigenetically regulated tumor-suppressor gene in lung cancer. KLF4 gene encoded a protein that belonged to the Kruppel family of transcription factors. The encoded protein was thought to control the G1-to-S transition of the cell cycle following DNA damage by mediating the tumor suppressor gene p53. Li et al. 27 found that KLF4 was dramatically down-regulated in lung adenocarcinoma tissue and cell lines. Restoration of KLF4 inhibits migration and invasion of LUAD cells through suppressing MMP2.

In order to further confirm the expression levels of hub genes in the current PPI network, we analyzed the related data in TCGA and GTEx databases and found that these genes were also significantly down-regulated in LUAD tissues. Furthermore, survival analysis based on TCGA data suggested that the expression levels of PECAM1 and ENG were associated with survival of LUAD patients. Specifically, low expression of ENG was associated with poor OS of LUAD patients. Low expression of PECAM1 was associated not only with poor OS but also with poor DFS of LUAD patients. The prognosis prediction model based on age, tumor stage, ENG and PECAM1 could effectively predict 5-year survival of LUAD patients.

Although some key genes associated with lung adenocarcinoma had been identified and the prognosis prediction model had been constructed in the current study, these findings based on bioinformatics analysis should be further investigated experimentally.

Conclusions

The present study demonstrated that PECAM1, CDH5, VWF, ENG, and KLF4 were key genes in LUAD, and the expression levels of PECAM1 and ENG were associated with survival of LUAD patients, which would provide the foundation for the development of novel methods for the diagnosis and therapy of LUAD.

Author biographies

Xinyu Wang is a College Student. Her research interest focuses on bioinformatics analysis and the pathogenesis of lung adenocarcinoma.

Jiaojiao Yang, Master of Medicine, is mainly engaged in research into the pathogenesis of cancer.

Xueren Gao, Doctor of Medicine, is mainly engaged in research into the pathogenesis of cancer.

Footnotes

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

  • 1.Zappa C, Mousa SA.Non-small cell lung cancer: current treatment and future advances. Transl Lung Cancer Res 2016; 5(3): 288–300. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Gao L, Qiu H, Liu J, et al. KLF15 promotes the proliferation and metastasis of lung adenocarcinoma cells and has potential as a cancer prognostic marker. Oncotarget 2017; 8(66): 109952–109961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Zheng Y, Li B, Wang J, et al. Identification of SUV39H2 as a potential oncogene in lung adenocarcinoma. Clin Epigenetics 2018; 10(1): 129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Shi X, Tan H, Le X, et al. An expression signature model to predict lung adenocarcinoma-specific survival. Cancer Manag Res 2018; 10: 3717–3732. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Su H, Lin Z, Peng W, et al. Identification of potential biomarkers of lung adenocarcinoma brain metastases via microarray analysis of cDNA expression profiles. Oncol Lett 2019; 17(2): 2228–2236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Chen Y, Chen H, Mao B, et al. Transcriptional characterization of the tumor Immune microenvironment and its prognostic value for locally advanced lung adenocarcinoma in a Chinese population. Cancer Manag Res 2019; 11: 9165–9173. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Zhu X, Luo H, Xu Y.Transcriptome analysis reveals an important candidate gene involved in both nodal metastasis and prognosis in lung adenocarcinoma. Cell Biosci 2019; 9: 92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Selamat SA, Chung BS, Girard L, et al. Genome-scale analysis of DNA methylation in lung adenocarcinoma and integration with mRNA expression. Genome Res 2012; 22(7): 1197–1211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Xu L, Lu C, Huang Y, et al. SPINK1 promotes cell growth and metastasis of lung adenocarcinoma and acts as a novel prognostic biomarker. BMB Rep 2018; 51(12): 648–653. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Yu G, Wang LG, Han Y, et al. ClusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012; 16(5): 284–287. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Szklarczyk D, Morris JH, Cook H, et al. The STRING database in 2017: quality-controlled protein-protein association networks, made broadly accessible. Nucleic Acids Res 2017; 45(D1): D362–D368. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Shannon P, Markiel A, Ozier O, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 2003; 13: 2498–2504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Woodfin A, Voisin MB, Nourshargh S.PECAM-1: a multi-functional molecule in inflammation and vascular biology. Arterioscler Thromb Vasc Biol 2007; 27(12): 2514–2523. [DOI] [PubMed] [Google Scholar]
  • 14.Lertkiatmongkol P, Liao D, Mei H, et al. Endothelial functions of platelet/endothelial cell adhesion molecule-1 (CD31). Curr Opin Hematol 2016; 23(3): 253–259. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Nakada MT, Amin K, Christofidou-Solomidou M, et al. Antibodies against the first Ig-like domain of human platelet endothelial cell adhesion molecule-1 (PECAM-1) that inhibit PECAM-1-dependent homophilic adhesion block in vivo neutrophil recruitment. J Immunol 2000; 164(1): 452–462. [DOI] [PubMed] [Google Scholar]
  • 16.DeLisser HM, Christofidou-Solomidou M, Strieter RM, et al. Involvement of endothelial PECAM-1/CD31 in angiogenesis. Am J Pathol 1997; 151(3): 671–677. [PMC free article] [PubMed] [Google Scholar]
  • 17.Cao G, O’Brien CD, Zhou Z, et al. Involvement of human PECAM-1 in angiogenesis and in vitro endothelial cell migration. Am J Physiol Cell Physiol 2002; 282(5): C1181–C1190. [DOI] [PubMed] [Google Scholar]
  • 18.Hung MS, Chen IC, Lung JH, et al. Epidermal growth factor receptor mutation enhances expression of cadherin-5 in lung cancer cells. PLoS One 2016; 11(6): e0158395. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Mao XG, Xue XY, Wang L, et al. CDH5 is specifically activated in glioblastoma stemlike cells and contributes to vasculogenic mimicry induced by hypoxia. Neuro Oncol 2013; 15(7): 865–879. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Fujita T, Igarashi J, Okawa ER, et al. CHD5, a tumor suppressor gene deleted from 1p36.31 in neuroblastomas. J Natl Cancer Inst 2008; 100(13): 940–949. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Metodieva SN, Nikolova DN, Cherneva RV, et al. Expression analysis of angiogenesis-related genes in Bulgarian patients with early-stage non-small cell lung cancer. Tumori 2011; 97(1): 86–94. [DOI] [PubMed] [Google Scholar]
  • 22.Sadler JE.Biochemistry and genetics of von Willebrand factor. Annu Rev Biochem 1998; 67: 395–424. [DOI] [PubMed] [Google Scholar]
  • 23.Starke RD, Ferraro F, Paschalaki KE, et al. Endothelial von Willebrand factor regulates angiogenesis. Blood 2011; 117(3): 1071–1080. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Hillgruber C, Steingräber AK, Pöppelmann B, et al. Blocking von Willebrand factor for treatment of cutaneous inflammation. J Invest Dermatol 2014; 134(1): 77–86. [DOI] [PubMed] [Google Scholar]
  • 25.Methia N, André P, Denis CV, et al. Localized reduction of atherosclerosis in von Willebrand factor-deficient mice. Blood 2001; 98(5): 1424–1428. [DOI] [PubMed] [Google Scholar]
  • 26.O’Leary K, Shia A, Cavicchioli F, et al. Identification of Endoglin as an epigenetically regulated tumour-suppressor gene in lung cancer. Br J Cancer 2015; 113(6): 970–978. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Li S, Huang L, Gu J, et al. Restoration of KLF4 inhibits invasion and metastases of lung adenocarcinoma through suppressing MMP2. J Cancer 2017; 8(17): 3480–3489. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Science Progress are provided here courtesy of SAGE Publications

RESOURCES