Abstract
Pancreatic adenocarcinoma (PAAD) is a malignancy with the highest mortality rate worldwide. There is a pressing need for novel biomarkers that can facilitate early detection and serve as targets for therapeutic interventions beyond the commonly utilized CA199 marker. This study utilized microarray datasets (GSE15471, GSE62165, and GSE28735) from the Gene Expression Omnibus (GEO) to identify differentially expressed genes (DEGs) and construct a protein-protein interaction network using STRING and Cytoscape. Hub genes were selected using BiNGO. Expression profiles and clinical data from the Cancer Genome Atlas (TCGA) were then used to compare the expression levels of CTSK and PLAU in pancreatic cancer and healthy pancreatic tissues via the Wilcoxon rank-sum test, with further validation using qPCR. Functional enrichment analysis was conducted to explore potential signaling pathways and biological functions. Prognostic values were assessed by the Kaplan-Meier and Cox regression analyses, and an overall survival (OS) nomogram was created to predict 1-, 2-, and 3-year survival after cancer diagnosis. The infiltration of immune cells was evaluated by single-sample gene set enrichment analysis. The methylation status of both genes was analyzed using the UALCAN and MethSurv databases. The results demonstrated that CTSK and PLAU were overexpressed in pancreatic cancer and that the hypomethylation status of both genes was associated with a poor prognosis. The overexpression of both genes was positively correlated with various immune cells, and functional enrichment analysis revealed that they were associated with immune cell infiltration. Besides, the effects of PLAU on the migration and invasion of pancreatic cancer cells were also verified by scratch and transwell experiments. Consequently, CTSK and PLAU have potential as prognostic biomarkers for pancreatic cancer.
1. Introduction
Pancreatic adenocarcinoma (PAAD) is an increasingly prevalent malignancy and a leading cause of cancer-related mortality, exhibiting a low 5-year survival rate of approximately 10% and a high mortality rate [1]. The incidence and mortality of PAAD are also increasing annually in comparison to some developed countries [2]. CA19-9 is a well-established circulating biomarker for pancreatic cancer [3], but its sensitivity as a diagnostic biomarker is limited to approximately 80% [4]. Consequently, there is a need for new biomarkers that can facilitate early detection. Differential gene expression analysis, carried out using bioinformatic and microarray techniques, has identified numerous differentially expressed genes (DEGs). However, discerning credible results from independent microarray analyses’ false-positive rates can be challenging. As such, we utilized the Gene Expression Omnibus to choose and download three microarray datasets (GEO) comparing pancreatic cancer tissues and noncancerous tissues to acquire DEGs.
Notably, KRAS and TP53 mutations are two significant drivers of pancreatic cancer and play essential roles in tumor invasion and transformation [5]. KRAS, which can be regarded as a rheostat in cells due to its dependence on the number of active molecules [6], can be utilized as a biomarker to predict the prognosis of PAAD or as a target for tumor therapy [7]. TP53, on the other hand, is a tumor suppressor whose mutations suppress DNA repair-stimulatory and apoptosis-inducing functions, leading to the development of PAAD [8]. Its mutation enhances the cancer’s activities by up to 70% [5]. ARF6-AMAP1 pathways, empowered by KRAS/TP53 mutations, may foster fibrosis in PAAD and lead to immune evasion [9].
So, while the molecular biology of KRAS and TP53 has been sufficiently explored, other genes are also critical to PAAD. Subsequently, using TCGA database, we performed Gene Ontology (GO), GSEA from the Kyoto Encyclopedia, and protein-protein interaction network (PPI) to gain a better understanding of the molecular mechanism of PADD development. We employed PCR to demonstrate the differential expression of both genes in the PAAD cell line, and we studied the immune infiltration and methylation-related indicators of the target genes. This study is aimed at employing bioinformatics to investigate and comprehend the expression of PLAU and CTSK in pancreatic cancer, as well as their relationship with clinicopathological and prognostic significance, underlying molecular mechanisms, and immune cell infiltration, to aid physicians.
To summarize, our findings suggest that CTSK and PLAU may serve as novel potential biomarkers for pancreatic cancer. The materials and methods used are detailed below.
2. Methods and Materials
2.1. Data Collection and Processing
We obtained three datasets of gene expression: GSE15471, GSE62165, and GSE28735 from the Gene Expression Omnibus (GEO) database, which were generated using the Affymetrix Human Genome U133 Plus 2.0 Array. We converted the probe data to the corresponding gene symbols using the annotation information on the platform. The GSE15471 dataset comprised of 39 pairs of pancreatic adenocarcinoma (PAAD) tumors and normal pancreatic tissues. The GSE62165 dataset included 118 PDAC samples and 13 control samples, while the GSE28735 dataset contained 45 PDAC tissues and 45 adjacent nontumorous counterparts. Additionally, we downloaded 33 kinds of tumor project STAR process RNAseq data from TCGA database (https://portal.gdc.cancer.gov). The relevant data on normal tissues and cells were downloaded from the Genotype-Tissue Expression (GTEx) database. Transcripts per million reads (TPM) are used to standardize the HTSeq FPKM Level 3 data. R software v4.2.1 was used for statistical analysis. Wilcoxon’s rank-sum test was used to detect the data of the two groups (ns, ; ; ; ).
2.2. Differentially Expressed Gene
We utilized the GEO2R tool to determine the differentially expressed genes (DEGs) between PAAD tissues and matched normal pancreatic tissue samples. GEO2R is an interactive web-based tool that allows for the comparison of two or more sets of samples from the GEO series to identify genes that exhibit significant expression changes in different experimental conditions (https://www.ncbi.nlm.nih.gov/geo/geo2r/). To minimize false positives and identify statistically significant genes, we used a combination of values and the Benjamini and Hochberg false discovery rate. We excluded probe sets with mismatched gene symbols or genes with multiple probe sets. logFC (fold change) >1 or <-1 and value < 0.05 were considered statistically significant.
2.3. PPI Network Construction and Module Analysis
We employed an online database and interactive gene search tool, STRING (http://string-db.org) (version 11.5) [10], to explore the molecular interactions among DEGs. We selected interactions with a composite score of >0.4, and the resulting molecular interaction networks were visualized using Cytoscape (version 3.9.1) [11]. To identify the most significant modules in the protein-protein interaction (PPI) network, we used the Molecular Complex Detection (MCODE) algorithm [12]. We employed the following selection criteria: MCODE , degree , node score , max , and -.
2.4. Hub Genes Selection and Analysis
Genes with a were considered hub genes. Coexpressed genes were analyzed through the cBioPortal online platform (http://www.cbioportal.org) [12, 13]. To identify the biological processes associated with the hub genes, we used the Biology Network Gene Ontology tool (BiNGO) plugin (version 3.0.5) for Cytoscape [14]. The hierarchical clustering of hub genes was constructed using the UCSC Cancer Genomics Browser (http://genome-cancer.ucsc.edu) [15]. The overall and disease-free survival of hub genes were analyzed using the Kaplan-Meier curves on cBioPortal. Furthermore, we assessed the correlation between the expression of PLAU and CTSK and their seed genes in PAAD with Spearman’s correlation analysis.
2.5. Functional Enrichment Analysis
We utilized the R package clusterProfiler (4.4.4) and GOplot (1.0.2) to conduct Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes analyses (KEGG), and gene set enrichment analysis (GSEA) of the differentially expressed genes (DEGs). We considered values < 0.05 and false detection rates (FDR) <0.25 to identify statistically significant functions or signal pathways (Subramanian et al., 2005; Yu et al., 2012).
2.6. Diagnostic Value Analysis
The receiver operating characteristic (ROC) and time-dependent receiver operating characteristic curve were used to evaluate the diagnostic value of PLAU and CTSK in PAAD. The pROC package (version 1.18.0) was used to analyze the data, and the results were visualized by ggplot2 (version 3.3.6). By default, the pROC package corrects the outcome order of the data, and is considered statistically significant.
2.7. Clinical Significance of PLAU and CTSK in PAAD
Calibration, nomogram, and forest map were used for further clinical significance analysis of PLAU and CTSK in PAAD. The survival package (version 3.3.1) was used for proportional hazard hypothesis testing and Cox regression analysis, and the rms package (version 6.3-0) was used for calibration analysis and visualization. The survival package was used for proportional hazard hypothesis testing and Cox regression analysis, and the rms package was used to construct and visualize the nomogram correlation model. Forest map visualization was performed using ggplot2 (version 3.3.6).
2.8. DNA Methylation Analysis
To explore the potential mechanisms of CTSK and PLAU in pancreatic cancer, we assessed the methylation status of their promoters using the UALCAN database (Chandrashekar et al., 2017). Furthermore, we used the MethSurv database, an online tool for multivariate survival analysis based on DNA methylation data, to evaluate the predictive value of the methylation levels of PLAU and CTSK (Modhukur et al., 2018).
2.9. Immune Infiltration Analysis
The abundance of 24 immune cells was quantified to evaluate the level of immune infiltration in pancreatic cancer. The GSEA algorithm was applied to identify the relative enrichment of immune cells using a single probe utilizing the GSV R package (Bindea et al., 2013). Spearman’s correlation analysis was used to examine the association between the expression levels of PLAU and CTSK and immune cells, while the Wilcoxon rank-sum test was used to compare the differences in the level of immune infiltration between the high and low expression groups of PLAU and CTSK.
2.10. Cell Line and Cell Culture
Pancreatic cancer cell lines BxPc-3, PaTu8988, PANC-1, and mia-Paca2 and normal pancreatic cell line HPDE6-C7 were stored by the Central Laboratory of the Affiliated Hospital of Jiangsu University and were kindly provided by Second Military Medical University in Shanghai. All cell lines were cultured in DMEM (Hyclone, Beijing, China) supplemented with 10% fetal bovine serum (Gibco, Carlsbad, CA, USA) at 37°C in a humidified incubator with a 5% CO2 supply.
2.11. RNA Extraction and Real-Time PCR
Trizo Plus (Takara, Shiga, Japan) was used to extract the total RNA. According to the manufacturer’s specification, reverse transcription was performed using the RevertAid First Strand cDNA Synthesis Kit (Thermo, Waltham, MA, USA). Real-time PCR was performed in triplicate in 20 ml reactions with iQ SYBR Premix Ex Taq Perfect Real Time (Bio-Rad Laboratories, Inc., Hercules, CA, USA), 50 ng of first-strand cDNA, and 0.4 mg of each primer. Actin was selected as a housekeeping gene. The primer pairs used for the amplification of the CTSK, PLAU, and human action were as follows: CTSK forward primer: 5- GTGTCTGAGAATGATGGCTGTGGAG-3, and reserve primer: 5- TGCCTTGCCTGTTGGGTTGTAC-3, and PLAU: forward primer:5- TCGCTCAAGGCTTAACTCCAACAC-3, and reserve primer: 5- ACGGATCTTCAGCAAGGCAATGTC-3, actin forward primer:5- CACGAAACTACCTTCAACTCC-3, and reserve primer: 5- CATACTCCTGCTTGCTGATC-3. Samples were cycled once at 95°C for 2 min and then subjected to 35 cycles of 95°C, 56°C, and 72°C for 30 s each. The relative mRNA content was calculated using the method with GAPDH as an endogenous control.
2.12. Cell Migration and Invasion Assay
A scratch test was used to detect cell migration ability according to the reagent manufacturer’s protocol. Transwell assays were performed using transwell inserts (Corning, Corning, New York, USA) containing 8 m permeable wells according to the manufacturer’s protocol. Transfected PaTu8988 and BxPc-3 cells were harvested, resuspended in serum-free medium, and transferred to 8 m permeable wells (B100 000 cells per well). The cells were then incubated with a culture medium containing 10% FBS for 24 h before detection. The cells on the upper surface were scraped off, and the migrating cells on the lower surface were fixed and stained with 0.05% crystal violet for 30 min. Finally, five independent fields per transwell were counted, and the average number of cells per field is represented in the figure. To assess cell invasion, 100,000 cells were seeded in Matrigel-coated transwell inserts (BD Bioscience, Corning, NY, USA) in a serum-free medium. Cells were then treated similarly to cell migration assays.
2.13. Statistical Analysis
All statistical analyses were performed using R (version 3.6.3). Wilcoxon’s rank-sum test and paired sample -test were used to assess the statistical significance of PALU and CTSk expression in unpaired and paired tissues. Associations between clinical features and their expression were assessed using the Wilcoxon rank-sum test and logistic regression. All tests were two-tailed, and values < 0.05 were considered statistically significant.
3. Results
3.1. Identification of DEGs in PAAD
After analysing the microarray results, DEGs (375 in GSE15471, 2611 in GSE62165, and 56 in GSE28735) were identified. Among the 3 datasets were 249 genes shown in the Venn diagram (Figure 1(a)), including 183 downregulated genes and 66 upregulated genes among pancreatic ductal adenocarcinoma tumors that matched standard pancreatic tissue samples.
(a)
(b)
(c)
3.2. PPI Network Construction and Module Analysis
Using Cytoscape, we drew the PPI network of DEGs and identified the most important module (Figures 1(a) and 1(c)). KEGG and GO’s functional analyses of genes involved in this module were analyzed using the R package cluster profile. Results showed that genes in this module were mainly enriched in an extracellular matrix organisation, laminin complex collagen binding, and the PI3K-Akt signaling pathway (Table 1).
3.3. KEGG and GO Enrichment Analyses of DEGs
To analyze the biological classification of DEGs, GO analysis showed that for all upregulated DEGs, they were mainly enriched in toxic substance binding, exopeptidase activity, serine-type endopeptidase activity, serine hydrolase activity, serine-type peptidase activity in cell component (CC), asymmetric synapse, platelet alpha granule, postsynaptic density, platelet alpha granule lumen, and neuron-to-neuron synapse in molecular function (MF). For downregulated DEGs, they were mainly enriched in metalloendopeptidase activity, collagen binding, integrin binding, and extracellular matrix structural constituent in molecular function (MF). In the cell component (CC), they are mainly enriched in the basement membrane, endoplasmic reticulum lumen, extracellular matrix component, and the collagen-containing extracellular. And in biological processes (BP), they are mainly enriched in the formation of the primary germ layer, collagen fibril organisation, extracellular structure organization, and extracellular matrix organization. KEGG pathway analysis revealed that the downregulated DEGs were mainly enriched in protein digestion and absorption and pancreatic secretion. In contrast, the majority of the PI3K-Akt signaling pathway, amoebiasis, focal adhesion, protein digestion and absorption, and ECM-receptor interaction were concentrated in the elevated DEGs (Figures 2(a)–2(d)).
(a)
(b)
(c)
(d)
(e)
(f)
(g)
3.4. Hub Gene Selection and Analysis
We select 12 genes with as hub genes. In Table 2, we make a note of their names and functions. According to TCGA database, the correlation heat map was made by the supermen correlation coefficient, showing that those hub genes were correlated (Figure 2(e)). The biological process analysis of the hub genes is shown in Figure 2(f). We analyze the biological process of the hub genes. Hub genes could be clearly distinguished by hierarchical clustering between pancreatic ductal adenocarcinoma tumors and matching normal pancreatic tissue samples (Figure 2(g)). The cBioPortal online database was used explore overall survival and disease-free survival by using the Kaplan-Meier curve. Results showed that the expression of ASPM, OAS1, PLAU, and CTSK may lead to a poor prognosis for overall survival (Figures 3(a)–3(d)). We can also find that the expression of ASPM, OAS1, PLAU, and ITGA3 may lead to a poor prognosis for disease-free survival (Figures 3(e)–3(i)). Among these genes, COL1A1, PLAU, and CTSK showed the highest node genes, which may play important roles in the carcinogenesis or progression of PAAD. As COL1A1 has been well studied in pancreatic cancer, we selected PLAU and CTSK, which have shown reductions in overall and disease-free survival. The hub gene PLAU whose overall and disease-free survival is statistically significant ( for disease-free survival and for overall survival) (Figures 3(c) and 3(g)); however, the hub gene CTSK whose overall and disease-free survival is not statistically significant ( for disease-free survival and ) for overall survival. (Figures 3(d) and 3(i)). Further confirmation of CTSK clinical significance in overall survival and disease-free survival by obtaining patient samples is needed in follow-up studies.
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
3.5. Elevated Expression of CTSK and PLAU in PAAD for Bioinformatics Analysis
Pancancer analysis suggested that PLAU and CTSK were differentially expressed in pancreatic cancer and other cancers such as kidney chromophobe, stomach adenocarcinoma, and v = ovarian serous cystadenocarcinoma (Figures 4(a) and 4(b)). The expression of PLAU and CTSK was significantly higher in pancreatic cancer samples than in normal breast tissues () (Figures 4(c) and 4(d)).
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
3.6. Diagnostic Value Analysis
In addition, the ROC curve showed that PLAU and CTSK had excellent predictive ability to distinguish pancreatic cancer, with an area under the curve of 0.968 (95% -0.988) for CTSK and an area under the curve of 0.976 (95% -0.991) for PLAU (Figures 4(e) and 4(g)). Additionally, we created time-dependent ROC curves to assess further and compare the predictive performance of PLAU and CTSK, with the area under the ROC curve (AUROC) as the criterion. Results demonstrated that CTSK and PLAU both had excellent prediction for PAAD, with an area under the curve of 0.499, 0.662, and 0.773 in 1, 3, and 5 years for CTSK and an area under the curve of 0.672, 0.744, and 0.792 in 1, 3, and 5 years for PLAU (Figures 4(f) and 4(h)).
3.7. Prognostic Value of CTSK and PLAU in PAAD
To find prognostic factors, univariate and multivariate Cox regression analyses were performed.
The results of the univariate analysis demonstrated that for PLAU, T4&T3 stage (adjusted , 95% –3.816, ), N1 (adjusted , 95% –3.618, ), and the expression of PLAU (adjusted , 95% –1.576, ) were independent factors of OS in patients with PAAD, while for CTSK, T4&T3 stage (adjusted , 95% –3.850, ), N1 (adjusted , 95% –3.539, ), and the expression of CTSK (adjusted , 95% –2.326, ) were independent factors of OS in patients with PAAD (Figures 5(a) and 5(b)). While the results of the multivariate analysis demonstrated that for PLAU, N1 (adjusted , 95% –3.187, ) and the expression of PLAU (adjusted , 95% –1.470, ) were multivariate factors of OS in patients with PAAD, while for CTSK, N1 (adjusted , 95% -3.295, ) was independent factors of OS in patients with PAAD (Figures 5(c) and 5(d)).
(a)
(b)
(c)
(d)
3.8. Construction and Validation of the Nomogram
We constructed a nomenclature map based on independent OS factors to predict the (b–d, f–h) prognosis of pancreatic cancer patients. The higher the average score, the worse the prognosis (Figures 6(a) and 6(e)). Furthermore, the predictive performance of the nomogram was assessed using standard curves (Figures 6(b)–6(d) and 6(f)–6(h)). These results suggest that the nomogram is suitable.
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
3.9. Analysis of Related Differentially Expressed Genes of PLAU and CTSK in PAAD
To further explore the functional enrichment of the two target molecules in pancreatic cancer, the differentially expressed genes (DEGs) of PLAU and CTSK in PAAD were analyzed by the Dseq2 R package. The results showed 125 differentially expressed genes between the high-expression and low-expression CTSK groups, including 20 upregulated genes and 105 downregulated genes ( value < 0.05, ). Moreover, in PLAU, there were 339 gene expression differences between the high- and low-expression CTSK groups, of which 73 were upregulated and 266 were downregulated ( value < 0.05, ) (Figures 7(a) and 7(b)). And then, the relationship between the top 5 high and top 5 low DEGs in CTSK (including EPYC, MAGEA4, CASP14, MAB21L2, COL11A1, APELA, LINC01929, LRRC15, MAGEA11, and IGLV1−41) were presented in Figure 7(c), and the relationship between the top 5 high and top 5 low DEGs (including CASP14, MIR205HG, CGB5, S100A2, AC034223.2, KRT14, LINC01929, KRT5, MAGEB2, and MUC21) in PLAU were presented in Figure 7(d).
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
3.10. Function Enrichment for DEGs of CTSK and PLAU in PAAD
For GO enrichment analysis, DEGs of CTSK were mainly enriched in digestion, response to food, serine-type endopeptidase activity, serine-type peptidase activity, serine hydrolase activity, and hormone activity (Figure 7(f)), while PLAU is mainly enriched in cornification, modulation of chemical synaptic transmission, regulation of trans-synaptic signaling, cation channel complex, ion channel complex, transmembrane transporter complex, cation channel activity, and syntaxin-1 binding (Figure 7(h)). For KEGG analysis, CTSK is mainly enriched in pancreatic secretion, starch and sucrose metabolism, protein digestion and absorption, carbohydrate digestion and absorption, salivary secretion, fat digestion and absorption, and glycerolipid metabolism (Figure 7(e)), while PLAU is mainly enriched in pancreatic secretion, protein digestion and absorption, dopaminergic synapse, starch and sucrose metabolism, cholinergic synapse, nicotine addiction, fat digestion and absorption, and synaptic vesicle cycle (Figure 7(g)). In Figures 7(i) and 7(h), we analyzed all DEGs for their GSEA which showed that PLAU and CTSK are connected with Reactome GPCR ligand binding, Reactome G alpha-1 signaling events, Reactome neuronal system, Reactome metabolism of amino acids and derivatives, Reactome class A 1 rhodopsin-like receptors, Reactome metabolism of amino acids and derivatives, Reactome translation, Reactome rRNA processing, KEGG Huntington’s disease, and Reactome regulation of expression of Slits and Robos in PAAD. These results suggest that there may be a relationship between these two genes and immune infiltration.
3.11. Correlation between Methylation and Expression of PLAU and CTSK
To further characterise the mechanism of PLAU and CTSK overexpression in PAAD, we examined the correlation between PLAU and CTSK expression levels and methylation status. Figures 8(a) and 8(b) clearly show that the DNA methylation of CTSK and PLAU at the promoters is significantly lower than that of normal pancreatic tissue from the UALCAN database. In addition, several methylated sites were associated with poor prognosis in CTSK, including cg1946165 and cg26574794 (Figures 8(c) and 8(d)), while for PLAU, it was cg04939496 (Figure 8(e)).
(a)
(b)
(c)
(d)
(e)
3.12. Correlation between PALU and CTSK Expression and Immune Infiltration
CTSK expression was significantly positively linked with Th1 immune cell, macrophage, and neutrophil infiltration levels (Figures 9(a) and 9(c)). Moreover, the enrichment scores of Th1 cells, macrophages, and neutrophils in the CTSK high expression group were higher than those in the CTSK low expression group (all ) (Figures 9(e)–9(g)). While for PLAU, Th1 cells, Th2 cells, and macrophage immune cell infiltration were positively correlated with the levels of PLAU expression (Figures 9(b) and 9(d)). Moreover, the enrichment scores of Th1 cells, Th2 cells, and macrophages in the PLAU high expression group were markedly higher than those in the PLAU low expression group (all ) (Figures 9(h)–9(j)).
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
3.13. Relative Expression of PLAU in Pancreatic Cancer Cell Lines
To further identify the expression of PLAU between pancreatic cell lines and normal pancreatic cell lines, we examined the expression level of PLAU in HPDE6-C7, PaTu8988, PANC-1, BxPc-3, and mia-Paca2 cells using real-time PCR. The expression of PaTu8988, PANC-1, and BxPc-3 was significantly higher than that of human normal pancreatic ductal epithelial cell HPDE6-C7. In different pancreatic cancer cells, the expression level was also different. The expression level of the BxPc-3 cell line was relatively low, the PaTu8988 cell line was relatively high, and PANC-1 and mia-Paca2 were between the two, and the difference was statistically significant () (Figure 10(a)).
(a)
(b)
(c)
(d)
3.14. PLAU Improve the Migration and Invasion of Pancreatic Cancer Cells
Scratch experiments showed that PaTu8988 cell lines had stronger migration ability than BxPc3 cell lines (Figure 10(b)). Transwell assay showed that the migration and invasion abilities of PaTu8988 were stronger than those of BxPc3, suggesting that PLAU plays a great role in the migration and invasion of pancreatic cancer (Figures 10(c) and 10(d)).
4. Discussion
Pancreatic cancer global incidence has significantly contributed to cancer-related deaths due to challenges in early detection and limited treatment options for advanced-stage cancers. Although carbohydrate antigen 19-9 has been FDA-approved as a serum biomarker, its limited specificity precludes its use for early detection of pancreatic cancer [16]. Thus, there is an urgent need for reliable biomarkers to facilitate the early detection of pancreatic cancer. TCGA [17] and GEO [18] databases contain vast amounts of clinical data that can be used for further data analysis. In this study, we analyzed three microarray databases to identify differentially expressed genes between pancreatic cancer tissues and noncancerous tissues. The results showed that COL1A1, PLAU, and CTSK were highly expressed in the three datasets, with COL1A1 having been shown to play a significant role in pancreatic cancer [19]. Therefore, we selected PLAU and CTSK as target genes for further research.
PLAU and CTSK have been shown to play a role in various tumors, with PLAU being considered as a potential biomarker in head and neck squamous cell carcinoma [10], promoting the EMT process [11], and tumor development in gastric cancer by cooperating with FOXM1 [12]. In pancreatic cancer, existing studies have shown that PLAU may be related to lymphatic metastasis [13], but its specific molecular mechanism is yet to be elucidated. While in pulmonary lymphangioleiomyomatosis, CTSK is a better prognostic marker [14], in colorectal cancer, it acts as a mediator between gut microbiota imbalance and tumor metastasis [20], and in prostate cancer, it affects tumor migration and invasion through the IL-17/CTSK axis, making it a potential therapeutic target [15]. However, the molecular mechanism and clinical prediction of CTSK in pancreatic cancer have not yet been studied.
Therefore, we used TCGA database to confirm the clinical value of these two genes including the KM curve, ROC curve, and Cox regression analysis. And our parts of methods mainly refer to similar literature on colon cancer [21]. As the molecular mechanism of CTSK and PLAU in the development of pancreatic cancer has not been studied, we identified the DEGs of these two genes in pancreatic cancer through the DsEq2 package and conducted functional studies of KEGG, GO, and GSEA enrichment. The results suggest that CTSK and PLAU are involved in the related molecular mechanism of immune infiltration.
Methylation is involved in a variety of cellular processes and human diseases, and methylation-related genes have been confirmed as potential biomarkers for early diagnosis of colon cancer [22]. M1a methylation modification has been confirmed to play an important role in the prognosis and shaping of the immune microenvironment of ovarian cancer [23]. Therefore, we performed methylation analysis of these two genes simultaneously to further validate their clinical value and predict their potential molecular mechanisms in pancreatic cancer.
Tumor microenvironments are characterized by immune cells, stromal cells, blood vessels, and the extracellular matrix, with tumors infiltrated by a variety of adaptive and innate immune cells that exert both protumor and antitumor functions [24]. Molecules that constitute the immune microenvironment, such as pyroptosis-related lncRNAs, can predict patient prognosis and immunotherapy effects [25]. Immune infiltration provides new ideas for tumor therapy [3, 26], as confirmed in malignant tumors such as breast cancer [27]. Immunotherapy, as a new tumor treatment method, is based on the principle of providing passive or active immunity for malignant tumors by using the immune system to target them, which is closely related to immune infiltration and has been widely used clinically [19], but it also faces problems such as drug resistance, similar to traditional chemotherapy. Differentially expressed genes change the composition of the microenvironment by recruiting different types of immune cells, thereby affecting the efficacy of immunotherapy. In epithelial ovarian cancer, immune-related genes screened using the CIBERSORT algorithm proved useful for the development of novel immune biomarkers and targeted therapy for epithelial ovarian cancer [23]. In pancreatic cancer, where immunotherapy has implications in it. Further studies are needed to validate the clinical significance of these biomarkers in larger cohorts of patients and to elucidate their molecular mechanisms in immune infiltration and other aspects of pancreatic cancer.
In conclusion, our study identified PLAU and CTSK as potential biomarkers for pancreatic cancer through bioinformatics analysis of microarray data. We validated their clinical significance through methylation analysis, immunohistochemistry, and survival analysis using TCGA data. Furthermore, we provided insights into the potential molecular mechanisms of these biomarkers in immune infiltration and other aspects of pancreatic cancer. Our findings contribute to the development of new therapeutic strategies for pancreatic cancer and lay the foundation for further studies to elucidate the role of PLAU and CTSK in this disease.
Data Availability
The datasets generated and/or analyzed during the current study are available in TCGA database (https://www.cancer.gov/ccg/research/genome-sequencing/tcga) and the GEO database (https://www.ncbi.nlm.nih.gov/geo/).
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Authors’ Contributions
Yuntao Ding, Zhangzuo Li, and Huizhi Wang contributed equally to this work.
Acknowledgments
This study was supported by the National Natural Science Foundation of China (Nos. 82072754 and 32170910), Jiangsu Provincial Key Research and Development Program (No. BE2018689), Natural Science Foundation of Jiangsu Province (No. BK20211124), and Zhenjiang Key Research and Development Program (Nos. SH2021037 and SH2018033).