Integrated bioinformatics analysis of potential biomarkers for pancreatic cancer

Abstract

Background

Pancreatic cancer, particularly pancreatic ductal adenocarcinoma (PDA), is an aggressive malignancy associated with a low 5‐year survival rate. Poor outcomes associated with PDA are attributable to late detection and inoperability. Most patients with PDA are diagnosed with locally advanced and metastatic disease. Such cases are primarily treated with chemotherapy and radiotherapy. Because of the lack of effective molecular targets, early diagnosis and successful therapies are limited. The purpose of this study was to screen key candidate genes for PDA using a bioinformatic approach and to research their potential functional, pathway mechanisms associated with PDA progression. It may help to understand the role of associated genes in the development and progression of PDA and identify relevant molecular markers with value for early diagnosis and targeted therapy.

Materials and methods

To identify novel genes associated with carcinogenesis and progression of PDA, we analyzed the microarray datasets GSE62165, GSE125158, and GSE71989 from the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) were identified, and the Database for Annotation, Visualization, and Integrated Discovery (DAVID) was used for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses. A protein‐protein interaction (PPI) network was constructed using STRING, and module analysis was performed using Cytoscape. Gene Expression Profiling Interactive Analysis (GEPIA) was used to evaluate the differential expression of hub genes in patients with PDA. In addition, we verified the expression of these genes in PDA cell lines and normal pancreatic epithelial cells.

Results

A total of 202 DEGs were identified and these were found to be enriched for various functions and pathways, including cell adhesion, leukocyte migration, extracellular matrix organization, extracellular region, collagen trimer, membrane raft, fibronectin‐binding, integrin binding, protein digestion, and absorption, and focal adhesion. Among these DEGs, 12 hub genes with high degrees of connectivity were selected. Survival analysis showed that the hub genes (HMMR, CEP55, CDK1, UHRF1, ASPM, RAD51AP1, DLGAP5, KIF11, SHCBP1, PBK, and HMGB2) may be involved in the tumorigenesis and development of PDA, highlighting their potential as diagnostic and therapeutic factors in PDA.

Conclusions

In summary, the DEGs and hub genes identified in the present study not only contribute to a better understanding of the molecular mechanisms underlying the carcinogenesis and progression of PDA but may also serve as potential new biomarkers and targets for PDA.

Pancreatic cancer is highly malignant, and using bioinformatics analysis (procedure below), we identified atotal of 11 key genes associated with prognosis, which may help in the diagnosis and treatment of pancreaticcancer.

1. INTRODUCTION

Pancreatic cancer, especially pancreatic ductal adenocarcinoma (PDA), which accounts for >90% of pancreatic cancers, is associated with a poor prognosis. ¹ The 5‐year overall survival rate remains approximately 7%. ² , ³ Despite advancements in our understanding of pancreatic cancer, the precise mechanisms underlying tumorigenicity remain unclear.

Approximately 10% of all patients have an inherited predisposition to PDA development, including KRAS activation mutations, which are present in >90% of tumors ⁴ ; inactivating mutations affecting TP53, CDKN2A, and SMAD4, which occur in 50–80% of pancreatic cancers ⁵ ; and mutations affecting other genes, including ARID1A, MLL3, and TGFBR2, which affect approximately 10% of tumors. ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² The median survival is 6–9 months for locally advanced disease and 3 months for metastatic disease. ¹³

With in‐depth research on the molecular mechanism of PDA, the serum marker Carbohydrate antigen199 (CA199) has been widely used in clinical practice as an adjunctive diagnostic test for PDA and postoperative monitoring. Although CA199 has a high sensitivity for PDA diagnosis, its lack of specificity is also quite limited. ¹⁴ About 75% of patients with PDA have mutations in the p53 gene, which is associated with cell cycle arrest and apoptosis. ⁶ , ¹⁵ More than 90% of pancreatic cancer patients have vascular epidermal growth factor (VEGF) overexpression, ¹⁶ which has been used as a target for drug therapy to inhibit tumor angiogenesis and thus reduce tumor blood supply to inhibit tumor growth. However, in one study, treatment of PDA with bevacizumab in combination with gemcitabine did not significantly improve the survival prognosis of patients. ¹⁷ With the widespread use of next‐generation sequencing (NGS) technology, a reliable tool for early diagnosis and progression prediction of cancer has been provided. ¹⁸ And a large amount of relevant biological and clinical data has been generated by NGS analysis, which provides the possibility to study and analyze the bioinformatics of gene expression and molecular functional characteristics associated with PDA. ¹⁹

To continue research on the molecular mechanisms of PDA carcinogenesis and progression, we identified a total of 202 DEGs and 12 hub genes. The corresponding bioinformatics analysis revealed that HMMR, CEP55, CDK1, UHRF1, ASPM, RAD51AP1, DLGAP5, KIF11, SHCBP1, PBK, and HMGB2, in particular, may be potential biomarkers and therapeutic targets for PDA.

2. METHODS AND MATERIALS

2.1. Cell culture

Four PDA cell lines (AsPC‐1, SW1990, PANC‐1, and BxPC‐3) and a normal human pancreatic ductal epithelial cell line (HPDE) were used in the study. Cell lines were purchased from the Cell Bank of the Type Culture Collection of the Chinese Academy of Sciences. According to the supplier's protocol, SW1990, PANC‐1, and HPDE cell lines were cultured in Dulbecco's modified Eagle's medium (DMEM; VivaCell), and BxPC‐3 and AsPC‐1 cells were cultured in RPMI 1640 medium (VivaCell). All culture media were supplemented with 10% fetal bovine serum (FBS; VivaCell) and 100 U/ml penicillin and 100 µg/ml streptomycin, and the cells were maintained under standard culture conditions (37°C, 95% humidified air, and 5% CO₂).

2.2. Microarray data

PDA gene expression datasets (GSE62165, ²⁰ GSE125158, ²¹ and GSE71989 ²² ) were obtained from the NCBI Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/), a public functional genomics data repository for storing high‐throughput gene expression datasets, sequence‐based data, and microarrays. ²³ GSE62165 (GPL13667 [HG‐U219] Affymetrix Human Genome U219 Array) is based on 118 PDA tissue samples and 13 noncancerous samples, GSE125158 (GPL6480, Agilent‐014850 Whole Human Genome Microarray 4x44K G4112F) is based on 17 PDA and 13 noncancerous samples, GSE71989 (GPL570 [HG‐U133_Plus_2] Affymetrix Human Genome U133 Plus 2.0 Array) is based on 13 PDA and nine noncancer samples.

2.3. Identification of DEGs

A total of 148 tumors and 35 nontumor tissue datasets were identified. DEGs between PDA and noncancer samples were identified using GEO2R (http://www.ncbi.nlm.nih.gov/geo/geo2r). GEO2R is an interactive web tool used to identify DEGs between different datasets by comparing them in a GEO series. This tool was also used to calculate the p‐value and |logFC|, and the cutoff threshold for identifying DEGs was set at p < 0.05, and |logFC|≥1.0). Statistical analyses were performed on each dataset. Venn diagrams were constructed using a web tool (http://bioinformatics.psb.ugent.be/webtools/Venn/).

2.4. GO and KEGG pathway analysis of DEGs

GO and KEGG pathway enrichment analyses were performed to determine the biological functions of DEGs. ²⁴ Gene Ontology (GO) is widely used in functional annotation and enrichment analysis; biological process (BP), molecular function (MF), and cellular component (CC) are the three major components of gene function. KEGG is a database resource for collecting a large amount of data on molecular level information, biological pathways, and chemical substances that are generated by high‐throughput experimental technologies. The database for annotation, visualization, and integrated discovery (DAVID; http://david.ncifcrf.gov) (version 6.8) ²⁵ is an online biological information database that integrates biological data and analysis tools and provides a comprehensive set of functional annotation information of genes and proteins for users to extract biological information. We used DAVID for GO and KEGG pathway analyses of DEGs. Statistical significance was set at p < 0.05.

2.5. PPI network construction and hub gene identification

The PPI network was constructed using STRING (http://string‐db.org), which integrates both known and predicted PPIs. PPI pairs with combined scores of >0.4 were extracted and the PPI network was visualized using Cytoscape version 3.7.2. The most significant PPI network modules were identified using the Molecular Complex Detection (MCODE) plug‐in of Cytoscape (version 1.6.1) using the following cutoff thresholds: MCODE degree cutoff = 2, node score cutoff = 0.2, Max depth = 100, and k‐score = 2.

2.6. Selection and analysis of hub genes

The hub genes were selected with the hub gene module, and their co‐expression network was analyzed using Coexpedia (https://www.coexpedia.org/). Hierarchical clustering of the hub genes was performed using the UCSC Cancer Genomics Browser (http://genome‐cancer.ucsc.edu). Differential expression of hub genes in PAAD patients was evaluated using Gene Expression Profiling Interactive Analysis (GEPIA) (http://gepia.cancer‐pku.cn/), which is a newly developed interactive web server for analyzing the RNA sequencing expression data of 9736 tumors and 8587 normal samples from the TCGA and GTEx projects, using a standard processing pipeline. ²⁶ The association between hub genes and overall survival was evaluated using Kaplan‐Meier analysis.

2.7. RNA isolation and quantitative real‐time PCR

Total RNA was isolated using TRIzol reagent (TaKaRa) and transcribed into cDNA using the PrimeScript RT Reagent Kit (TaKaRa). Quantitative real‐time PCR was carried out using an RT‐PCR amplifier system (Bio‐Rad, USA) using an SYBR Green PCR Kit (Takara). Relative quantitation of mRNA expression was performed using the 2ΔΔCt method after normalization to the endogenous reference GAPDH. Table S1 S1 shows the primers used in this study.

2.8. Statistical analysis

In this study, the relevant data were analyzed using GraphPad Prism 8.0 and GEPIA online tools. The box diagram in GEPIA was used to analyze the expression of related genes, and Kaplan‐Meier survival analysis with log‐rank test was used to analyze the overall survival as well as to plot the survival curves. qRT‐PCR data are expressed as mean ± standard deviation (SD). Comparisons among multiple groups were performed using a one‐way analysis of variance. Statistical significance was set at p < 0.05.

3. RESULTS

3.1. Identification of DEGs

In this study, we selected three gene expression datasets (GSE62165, GSE125158, and GSE71989). As shown in Table 1, GSE62165 contained 118 PDA specimens and 13 normal pancreatic specimens, GSE125158 contained 17 PDA specimens and 13 normal pancreatic specimens, and GSE71989 contained 13 PDA samples and nine normal pancreatic samples. Using GEO2R, 6887, 1770, and 4367 DEGs were identified in datasets GSE62165, GSE125158, and GSE71989, respectively, with 202 overlapping DEGs between the three datasets (Figure 1).

TABLE 1.

A summary of pancreatic cancer microarray datasets from different gene expression omnibus datasets

	Series	Platform	affymetrix GeneChip	Samples
1	GSE62165	GPL13667	Affymetrix Human Genome U219 Array	131
2	GSE125158	GPL6480	Agilent−014850 Whole Human Genome Microarray 4x44K G4112F	30
3	GSE71989	GPL570	Affymetrix Human Genome U133 Plus 2.0 Array	22

FIGURE 1.

Volcano plots of genes that are different expressions between PDA tissues and noncancerous samples. Each symbol represents a different gene. The red color of the symbols means upregulated, whereas the green color of the symbols means downregulated, and dark color represents not significantly changed genes. X‐axis is log2‐fold changes of genes in PDA compared with normal pancreas samples. Y‐axis is‐log10 (p‐values). (A) Volcano plots of GSE125158. (B) Volcano plots of GSE71989. (C) Volcano plots of GSE62165. (D) Venn diagram. The DEGs were selected with a fold change ≥1 and p‐value <0.05 among the mRNA expression profile sets GSE125158, GSE71989, and GSE62165. The three datasets showed an overlap of 202 genes. DEGs, differentially expressed genes; PDA, pancreatic ductal adenocarcinoma

3.2. KEGG and GO enrichment analyses of DEGs

Next, we used DAVID to perform the GO and KEGG pathway analyses (Table 2). For BP, DEGs were enriched mainly in cell adhesion, collagen, leukocyte migration, endodermal cell differentiation, maintenance of gastrointestinal epithelium, regulation of glucose metabolic processes, positive regulation of transcription, DNA‐templated, and wound healing. MF was mainly enriched in calcium ion binding, fibronectin‐binding, translation repressor activity, nucleic acid binding, integrin binding, insulin‐like growth factor I binding, and serine‐type endopeptidase activity. For CC, DEGs were enriched in extracellular space, matrix, extracellular exosome, and membrane raft. KEGG pathway analysis revealed that the DEGs were associated with pathways related to cancer, extracellular matrix (ECM) receptor interaction, protein digestion, absorption, tumor necrosis factor (TNF) signaling pathway, and focal adhesion.

TABLE 2.

GO and KEGG pathway enrichment analysis of DEGs in PDA samples

Category	Term	Count	p‐Value
GOTERM_BP_DIRECT	GO:0007155~cell adhesion	19	3.61E−06
GOTERM_BP_DIRECT	GO:0030574~collagen catabolic process	8	6.66E−06
GOTERM_BP_DIRECT	GO:0050900~leukocyte migration	9	6.40E−05
GOTERM_BP_DIRECT	GO:0030199~collagen fibril organization	6	6.70E−05
GOTERM_BP_DIRECT	GO:0035987~endodermal cell differentiation	5	2.10E−04
GOTERM_BP_DIRECT	GO:0030277~maintenance of gastrointestinal epithelium	4	2.73E−04
GOTERM_BP_DIRECT	GO:0030198~extracellular matrix organization	10	3.42E−04
GOTERM_BP_DIRECT	GO:0010906~regulation of glucose metabolic process	4	0.001763481
GOTERM_BP_DIRECT	GO:0045893~positive regulation of transcription, DNA‐templated	15	0.001824127
GOTERM_BP_DIRECT	GO:0042060~wound healing	6	0.001941507
GOTERM_BP_DIRECT	GO:0008284~positive regulation of cell proliferation	14	0.002110314
GOTERM_BP_DIRECT	GO:0090090~negative regulation of canonical Wnt signaling pathway	8	0.00228459
GOTERM_BP_DIRECT	GO:0010628~positive regulation of gene expression	10	0.00262849
GOTERM_BP_DIRECT	GO:0042475~odontogenesis of dentin‐containing tooth	5	0.003200207
GOTERM_BP_DIRECT	GO:0006508~proteolysis	14	0.003853736
GOTERM_BP_DIRECT	GO:0071407~cellular response to organic cyclic compound	5	0.00412687
GOTERM_BP_DIRECT	GO:0043586~tongue development	3	0.005179886
GOTERM_BP_DIRECT	GO:1902042~negative regulation of extrinsic apoptotic signaling pathway via death domain receptors	4	0.005714744
GOTERM_BP_DIRECT	GO:0001649~osteoblast differentiation	6	0.006006712
GOTERM_BP_DIRECT	GO:0050919~negative chemotaxis	4	0.006217342
GOTERM_BP_DIRECT	GO:0097192~extrinsic apoptotic signaling pathway in absence of ligand	4	0.006217342
GOTERM_BP_DIRECT	GO:0060394~negative regulation of pathway‐restricted SMAD protein phosphorylation	3	0.007487389
GOTERM_CC_DIRECT	GO:0005615~extracellular space	35	1.41E−06
GOTERM_CC_DIRECT	GO:0005576~extracellular region	39	1.56E−06
GOTERM_CC_DIRECT	GO:0031012~extracellular matrix	15	3.04E−06
GOTERM_CC_DIRECT	GO:0005578~proteinaceous extracellular matrix	14	5.22E−06
GOTERM_CC_DIRECT	GO:0070062~extracellular exosome	52	3.99E−05
GOTERM_CC_DIRECT	GO:0005581~collagen trimer	8	5.14E−05
GOTERM_CC_DIRECT	GO:0009897~external side of plasma membrane	9	0.001863917
GOTERM_CC_DIRECT	GO:0045121~membrane raft	8	0.006117264
GOTERM_MF_DIRECT	GO:0005509~calcium ion binding	22	5.53E−05
GOTERM_MF_DIRECT	GO:0001968~fibronectin binding	4	0.002970899
GOTERM_MF_DIRECT	GO:0000900~translation repressor activity, nucleic acid binding	3	0.005289145
GOTERM_MF_DIRECT	GO:0005178~integrin binding	6	0.006536482
GOTERM_MF_DIRECT	GO:0031994~insulin‐like growth factor I binding	3	0.007644107
GOTERM_MF_DIRECT	GO:0004252~serine‐type endopeptidase activity	9	0.008138653
KEGG_PATHWAY	hsa05200:Pathways in cancer	15	1.11E−04
KEGG_PATHWAY	hsa04512:ECM‐receptor interaction	7	4.26E−04
KEGG_PATHWAY	hsa04974:Protein digestion and absorption	7	4.53E−04
KEGG_PATHWAY	hsa04668:TNF signaling pathway	6	0.007120596
KEGG_PATHWAY	hsa04510:Focal adhesion	8	0.008530044

Abbreviations: DEGs, differentially expressed genes; GO, gene ontology; KEGG, kyoto encyclopedia of genes and genomes; PDA, pancreatic ductal adenocarcinoma.

3.3. PPI network construction and module analysis

The probability of relationships between the pathways was evaluated using STRING and a PPI network was constructed. A PPI network of DEGs that included 138 nodes and 640 interactions was constructed to identify gene interactions, as shown in Figure 2. The most important module was confirmed using Cytoscape (Figure 3A), and 12 hub genes were identified (Table 3). This analysis revealed that HMGB2 had the lowest degree (8), whereas all the others had a degree of 10. These hub genes were upregulated in PDA.

FIGURE 2.

The PPI network of DEGs was constructed using STRING database. The PPI network of DEGs that included 138 nodes and 640 edges. DEGs, differentially expressed genes; PPI, protein‐protein interaction; STRING, search tool for the retrieval of interacting genes

FIGURE 3.

Hub genes module and co‐expression network. (A) The most significant module was obtained from PPI network with 12 nodes and 126 edges, including all the hub genes. (B) The co‐expression network of hub genes was analyzed using Coexpedia online platform

TABLE 3.

Functional roles of 12 hub genes with degree ≥5

No.	Gene symbol	Full name	Function	Degree
1	CDK1	Cyclin‐dependent kinase 1	CDK1 is essential for G1/S and G2/M phase transitions of eukaryotic cell cycle, and associated with breast cancer	32
2	ASPM	Assembly factor for spindle microtubules	ASPM is essential for normal mitotic spindle function in embryonic neuroblasts, poor prognosis, and regulates cell proliferation in bladder cancer	28
3	KIF11	Kinesin family member 11	KIF11 encodes a motor protein, and associated with colorectal cancer	28
4	HMMR	Hyaluronan‐mediated motility receptor	HMMR is involved in cell motility, potentially associated with higher risk of breast cancer	24
5	CEP55	Centrosomal protein 55	Plays a role in mitotic exit and cytokinesis, may regulate breast cancer spinal metastases	24
6	UHRF1	Ubiquitin like with PHD and ring finger domains 1	UHRF1 plays a major role in the G1/S transition and retinoblastoma gene expression, and is up‐regulated in various cancers	24
7	DLGAP5	DLG‐associated protein 5	DLGAP5 is a potential cell cycle regulator, and overexpressed in hepatocellular carcinoma	24
8	PBK	PDZ‐binding kinase	PBK overexpression in tumorigenesis, and drives androgen‐independent growth in prostate cancer	24
9	RAD51AP1	RAD51‐associated protein 1	Structure‐specific DNA‐binding protein involved in DNA repair, and promotes progression of ovarian cancer	22
10	SHCBP1	SHC binding and spindle‐associated 1	SHCBP1 promotes the progression of prostate cancer	20
11	DTL	Denticleless E3 ubiquitin protein ligase homolog	DTL is involved in cell cycle control, and associated with endometrial squamous cell carcinoma	20
12	HMGB2	High mobility group box 2	HMGB2 may promote proliferation and invasion of renal tumor	16

3.4. Hub gene selection and analysis

A total of 12 genes with degrees ≥5 were identified as hub genes, and hub gene networks and hub gene co‐expression were assessed using Coexpedia (Figure 3B). Hierarchical clustering showed that the hub genes could differentiate PDA samples from noncancerous samples (Figure 4). Evaluation of the expression of these hub genes in PDA and noncancerous samples using GEPIA revealed that the expression of 12 hub genes was higher in PDA samples than in noncancerous samples (Figure 5, p < 0.05). Next, we used GEPIA to evaluate the relationship between the 12 hub genes and PDA survival and found that upregulation of HMMR, CEP55, CDK1, UHRF1, ASPM, RAD51AP1, DLGAP5, KIF11, SHCBP1, PBK, and HMGB2 was correlated with poor PDA prognosis (Figure 6). However, there was no significant correlation between DTL expression and prognosis (p > 0.05). Collectively, these findings highlight the therapeutic and diagnostic potential of these hub genes.

FIGURE 4.

Hierarchical clustering of hub genes was constructed using UCSC Xena. The samples under the pink bar are noncancerous samples, and the samples under the blue bar are PDA samples. Upregulation of genes is marked in red; downregulation of genes is marked in blue. PDA, pancreatic ductal adenocarcinoma

FIGURE 5.

The expression of 12 hub genes in PDA samples and noncancerous samples through GAPIA database (*p < 0.01). All hub genes were upregulated in the tumors of patients with PDA. Red represents tumor tissue and black represents normal tissue. PAAD or PDA, pancreatic ductal adenocarcinoma; GAPIA, gene expression profiling interactive analysis

FIGURE 6.

Kaplan‐Meier survival curve analysis. The overall survival of twelve hub genes (HMMR, CEP55, CDK1, UHRF1, ASPM, RAD51AP1, DLGAP5, KIF11, SHCBP1, DTL, PBK and HMGB2) expression in PAD using GEPIA. Red line depicts genes with higher expression associated with poor overall survival and blue line depicts genes with lower expression associated with good survival. PDA, pancreatic ductal adenocarcinoma

qRT‐PCR revealed that contrary to bioinformatics predictions CDK1, ASPM, and SHCBP1 were expressed at low levels in pancreatic cancer cell lines (AsPC‐1, SW1990, PANC‐1, and BxPC‐3) compared to that in HPDE cells (Figure 7). KIF11 was highly expressed in PANC‐1 compared to HPDE (p < 0.05); HMMR was highly expressed in AsPC‐1 and PANC‐1; CEP55 was highly expressed (p < 0.05); UHRF1 was highly expressed in AsPC‐1 and PANC‐1 (p < 0.05); DLGAP5 was highly expressed in AsPC‐1, BxPC‐3, and PANC‐1 (p < 0.05); PBK was highly expressed in AsPC‐1 and BxPC‐3; RAD51AP1 was highly expressed in AsPC‐1; and HMGB2 was highly expressed in AsPC‐1 (p < 0.05).

FIGURE 7.

Results of quantitative real‐time PCR revealed that CDK1, ASPM, KIF11, HMMR, CEP55, UHRF1, DLGAP5, PBK, RAD51AP1, SHCBP1, and HMGB2 transcription levels in four cancerous cell lines (AsPC‐1, SW1990, PANC‐1, BxPC‐3) and benign cells (HPDE). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

4. DISCUSSION

Here, we examined three GEO datasets using gene expression and protein‐protein interaction analyses and identified 202 DEGs in PDA versus noncancerous samples. A PPI network was constructed to evaluate the relationship between the DEGs and HMMR, CEP55, CDK1, UHRF1, ASPM, RAD51AP1, DLGAP5, KIF11, SHCBP1, DTL, PBK, and HMGB2 that were identified as hub genes (degrees ≥5). HMMR is involved in cell motility and ECM‐receptor interactions. Intracellular HMMR is an actin‐ and microtubule‐associated protein that maintains spindle integrity. ²⁷ Increased HMMR expression has been associated with progression and poor prognosis in various cancer types. ²⁸ , ²⁹ CEP55 is involved in mitotic exit and cytokinesis, ³⁰ , ³¹ as well as in mitotic nuclear division. It is located on 10q23.33 and contains nine exons. ³⁰ CEP55 is upregulated in ovarian epithelial carcinoma, gastric cancer, and breast cancer, and influences cell proliferation, cancer aggressiveness, and prognosis. ³² , ³³ , ³⁴ CDK1 interacts with various cyclins to regulate the cell cycle by regulating the centrosome cycle, mitotic onset, G2‐M transition, G1 progression, and G1‐S transition. ³⁵ , ³⁶ It phosphorylates p53. ³⁷ , ³⁸ CDK1 can substitute other CDKs and is sufficient to drive the mammalian cell cycle. ³⁹ UHRF1 positively regulates gene expression and DNA topoisomerase (ATP hydrolyzing) activity, nuclear chromatin, and protein binding. Several studies have shown that UHRF1 influences apoptosis in cancer cells. ⁴⁰ , ⁴¹ , ⁴² ASPM is involved in forebrain neuroblast division, neuron migration, mitotic spindle regulation, coordination of mitotic processes, and maintenance of neural progenitor cells. ⁴³ ASPM mutations have been implicated in microcephaly. ⁴⁴

RAD51AP1 is a structure‐specific DNA‐binding protein involved in DNA repair via the promotion of RAD51‐mediated homologous recombination. ⁴⁵ , ⁴⁶ Upregulation of DNA repair genes is associated with metastatic cancer. ⁴⁷ RAD51AP1 is upregulated in advanced breast cancer. ⁴⁸ DLGAP5 plays a role in spindle assembly, kinetochore fiber stabilization, and chromosomal segregation. ⁴⁹ , ⁵⁰ , ⁵¹ It is upregulated in hepatocellular carcinoma, colorectal cancer, urinary bladder transitional cell carcinoma, prostate cancer, meningioma, and adrenocortical carcinoma. ⁴⁹ , ⁵² KIF11 is associated with mitotic nuclear division, and its upregulation is associated with premature separation of sister chromatids and unequal chromosome segregation, ⁵³ which contributes to invasion and metastasis. ⁵⁴ KIF11 is upregulated in various cancers with poor prognosis, including pancreatic, and gastric cancer. ⁵⁵ , ⁵⁶ SHCBP1 (located on 16q11.2) is a member of the Shc adaptor downstream protein family. ⁵⁷ SHCBP1 phosphorylation is required for cleavage furrow separation, ⁵⁸ and it promotes the growth of hepatocellular carcinoma via the MEK/ERK pathway. ⁵⁹ DTL, which is located on chromosome 1q32, encodes a putative 730 amino acid protein ⁶⁰ associated with intracellular membrane‐bound organelles and regulates p53 protein stability. ⁶¹ DTL promotes metastatic potential in HCC as well as tumorigenesis in breast cancer cells. ⁶⁰ , ⁶² PBK, a novel serine‐threonine kinase, is a member of the mitogen‐activated protein kinase (MAPK) family and a negative modulator of the inflammatory response and mitotic nuclear division. It influences cytokinesis, DNA damage, and DNA repair. ⁶³ , ⁶⁴ , ⁶⁵ PBK has been implicated in various cancers. ⁶⁶ HMGB2 is a positive modulator of transcription, DNA‐templated, and negative regulation of extrinsic apoptosis via death domain receptors, as well as DNA binding, bending, and transcription factor binding. HMGB2 promotes breast cancer progression by regulating proliferation and glycolysis. ⁶⁷

The expression of these genes was verified by qRT‐PCR in pancreatic cancer and normal pancreatic cell lines. The results showed that CDK1, ASPM, and SHCBP1 were expressed at low levels in pancreatic cancer cells relative to normal pancreatic cell lines. The remaining eight genes were all or partially highly expressed in pancreatic cancer cell lines, consistent with those predicted by bioinformatic methods. We found that the expression of some genes was contrary to the bioinformatics prediction; therefore, they need to be validated at the tissue and protein levels during the next stage of subsequent experiments.

By reviewing the relevant literature, we found that PIAO J et al. ⁶⁸ performed immunohistochemical staining of CDK1 in 99 PDA tissues and 71 normal pancreatic tissues to analyze its expression and found that the CDK1 positivity rate (97/99, 98%) was significantly higher than that of normal pancreatic tissues (40/71, 56.3%). HSU C‐C et al. ⁶⁹ found that ASPM isoform I (ASPM‐iI) and ASPM‐iII, were differentially and heterogeneously expressed in various PDAC cell lines, while ASPM‐iIII and ASPM‐iIV were hardly expressed. YANG C et al. ⁷⁰ performed immunohistochemical staining of SHCBP1 in 186 PDA specimens and paracancerous tissues and showed that SHCBP1 (115/186, 62%) was expressed at higher levels in PDA samples than in paracancerous tissues.

Here, we identified HMMR, CEP55, CDK1, UHRF1, ASPM, RAD51AP1, DLGAP5, KIF11, SHCBP1, PBK, and HMGB2 as potential biomarkers for early diagnosis and treatment of PDA.

5. CONCLUSION

Using bioinformatic analysis, we identified 202 DEGs and HMMR, CEP55, CDK1, UHRF1, ASPM, RAD51AP1, DLGAP5, KIF11, SHCBP1, PBK, and HMGB2 as hub genes with diagnostic and therapeutic potential in PDA. However, further studies are required to elucidate the biological functions of these genes in PDA.

CONFLICTS OF INTEREST

The authors declare they have no conflicting financial interests.

Supporting information

Table S1

Shi H, Xu H, Chai C, Qin Z, Zhou W. Integrated bioinformatics analysis of potential biomarkers for pancreatic cancer. J Clin Lab Anal. 2022;36:e24381. doi: 10.1002/jcla.24381

DATA AVAILABILITY STATEMENT

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.