Identifying diagnostic markers and constructing a prognostic model for pancreatic cancer based on microarray and bioinformatic analysis

doi:10.21203/rs.3.rs-4989505/v1

Download PDF

Research Article

Identifying diagnostic markers and constructing a prognostic model for pancreatic cancer based on microarray and bioinformatic analysis

https://doi.org/10.21203/rs.3.rs-4989505/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Pancreatic cancer (PC) is one of the leading causes of cancer-related death worldwide. The lack of effective diagnostic biomarkers and therapeutic targets makes PC difficult to screen and treat. The aim of this study was to develop a diagnostic and survival-related gene signature for PC to construct a prognostic model. An Arraystar RNA microarray was used to identify differentially expressed genes (DEGs) in clinical plasma samples between the PC group and the control group. We performed weighted gene co-expression network analysis (WGCNA) to identify significant modules of DEGs in the Gene Expression Omnibus (GEO) cohort and to obtain potential diagnostic hub genes by intersecting the significant module genes with microarray-derived messenger RNA (mRNA). In addition, survival analysis and univariate and multivariate Cox regression analyses were performed on the hub genes to construct a prognostic model. Our microarray data revealed 228 significantly upregulated mRNA in the PC group compared with the control group. Moreover, we identified 5 feature mRNA (FERMT1, S100A14, KCNN4, PKM, and ITGA3) with good diagnostic performance. According to survival analysis based on The Cancer Genome Atlas (TCGA) dataset, higher expression of the hub genes was related to a poorer survival rate in patients with PC. Univariate and multivariate Cox proportional hazard analyses revealed that the expression of FERMT1, S100A14, and ITGA3 was anindependent risk factor for poor prognosis. Our results revealed the potential biomarkers for the prediction of PC prognosis in addition to clinicopathological factors. Moreover, this study provides new insights into the molecular mechanisms of PC.

Pancreatic cancer (PC)

microarray

prognosis

bioinformatics analysis

diagnostic markers

Pancreatic cancer (PC) remains a leading cause of cancer-related mortality worldwide (Klein 2021, Siegel et al. 2023). Pancreatic ductal adenocarcinoma (PDAC), the most common histological type of PC, progresses rapidly and has a dismal prognosis (Mizrahi et al. 2020). Due to the subtle and asymptomatic nature of localized disease and the paucity of diagnostic biomarkers for early-stage tumors, routine tumor office-based screening remains a challenge, leading to low surgical resection rates and distant metastasis. Chemotherapy is not effective for advanced PC, and the 5-year overall survival rate is less than 10% (Feng et al. 2019, Siegel et al. 2021). Although numerous treatments for PC have been proposed in recent years, the survival of these patients has not significantly improved. Therefore, a more in-depth study of the underlying molecular framework and alterations that drive PC oncogenesis is necessary to provide effective treatment strategies for patients. Moreover, the initial diagnostic stage of the tumor is significantly related to disease prognosis. Thus, further research on molecular markers for the early diagnosis, prognosis, and therapeutic targeting of PC is needed.

With the development of transcriptomic analysis and genomic microarrays (Dreyer et al. 2022, Pompella et al. 2020), bioinformatics analysis has provided an effective way to explore the etiology, diagnosis, and novel targets for clinical treatment. Previous studies demonstrating that PC arises through a multistep series of precursor lesions have generated substantial interest in the discovery of disease-related biomarkers for early detection, especially in high-risk cohorts. Numerous studies (Kishikawa et al. 2015, Kitagawa et al. 2019, Thompson et al. 2020) have shown the potential of circulating tumor-derived analytes as potential biomarkers for PC clinical applications. Notably, the potential role of messenger RNA (mRNA) in the prognosis and diagnosis of PC remains poorly understood.

In this study, we investigated differentially expressed mRNA from plasma samples of PC patients using Arraystar Microarray Technology (KANGCHEN, Shanghai, China). Moreover, we integrated the data with the Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/) and The Cancer Genome Atlas (TCGA) (https://portal.gdc.cancer.gov/) databases through systematic bioinformatics methods to explore hub genes that are closely associated with the pathogenesis and progression of PC. The findings of this study are expected to improve the diagnosis and prognosis of PC and provide a novel platform for molecular therapeutic strategies. The study workflow is shown in Fig. 1.

Data collection

All PC-related datasets, including the TCGA and GEO databases, were obtained from public databases. The RNA sequencing (RNA-Seq) expression pancreatic adenocarcinoma (PAAD) data of 169 PC tissue samples and 4 normal tissue samples with corresponding clinical information were downloaded from TCGA. In addition, GSE62452 (n = 69 pancreatic tumors, n = 61 adjacent nontumor tissues) and GSE71989 (n = 14 PDAC tissues, n = 8 normal pancreatic tissues) were retrieved from the GEO database and merged as training datasets for weighted gene co-expression network analysis (WGCNA). For the validation dataset, the gene expression matrix of the GSE16515 dataset (n = 36 pancreatic tumors, n = 16 non-tumor tissue) was obtained.

Clinical sample collection and array expression analysis

A total of 30 patients diagnosed with PC and 30 healthy volunteers were enrolled in the present study. The inclusion criteria for the PC cohort were as follows: (1) Patients prior to surgical resection in our hospital; (2) Patients who had been pathologically diagnosed with PC; (3) No previous chemotherapy, radiotherapy, and other special treatment. The exclusion criteria were as follows: (1) Patients younger than 18 years of age; (2) Unclear pathological diagnosis; (3) Combined with other types of tumors, chronic disease, and acute inflammation.

The plasma samples used in this study were obtained with informed consent from the participants, and the research procedures were approved by the Ethics Review Committee of the Second Affiliated Hospital of Army Medical University (Approval No.2024-154-01). The clinical features of the patients and healthy volunteers enrolled in this study are provided in Table 1.

Table 1

The clinical features of the patients and healthy volunteers enrolled in this study
Clinical features	PC cohorts (n = 30)	HC cohorts(n = 30)
Age
< 60	18 (60.0%)	20 (66.7%)
≥ 60	12 (40.0%)	10 (33.3%)
Sex
Male	21 (70.0%)	18 (60.0%)
Female	9 (30.0%)	12 (40.0%)
T stage		NA
T1	3 (10.0%)
T2	11 (36.7%)
T3	7 (23.3%)
T4	9 (30.0%)
Tx	0 (0.0%)
N stage		NA
N0	14 (46.7%)
N1	7 (23.3%)
N2	4 (13.3%)
Nx	5 (16.7%)
M stage		NA
M0	23 (76.7%)
M1	6 (20.0%)
Mx	1 (3.3%)

Values are expressed as n (%); PC, pancreatic cancer; HC, healthy control

Peripheral venous blood was collected from each patient and healthy volunteer. The red blood cells and plasma were separated by centrifuging peripheral blood samples at 3,000 rpm for 10 minutes at 4 ℃. The separated plasma was transferred to RNasefree tubes and stored at − 80 ℃. Before the microarray analysis, 30 plasma samples were mixed into one with an equal volume from the patient group, and the same was performed for the healthy group. A total of 3 replicate chips were set up for each group. Arraystar mRNA Microarray Technology was used to analyze the differential expression of mRNA between the PC group and the control group.

Co-expression analysis

WGCNA (Langfelder and Horvath 2008) was performed with the “WGCNA” R package (Version 4.3.1; The R Foundation for Statistical Computing, Vienna, Austria) to screen significant modules of differentially expressed genes (DEGs) in the training datasets. All paired genes were subjected to Pearson correlation analysis to construct a correlation matrix, and soft-threshold power values and the best parameter β were set to establish a scale-free network. A topological overlap matrix (TOM) was obtained from the adjacency matrix. Hierarchical clustering and dynamic tree-cutting algorithms were used to identify gene-network modules. Thus, the correlation between module signature genes and clinical features was calculated to identify modules of interest with similar gene expression patterns. The gene significance (GS) and module membership (MM) were also measured to evaluate the significance of genes and the correlation of gene expression profile with module eigengenes.

Enrichment analysis

To further investigate the potential biological functions of the merged significant module genes, the “cluster Profiler, enrichplot, DOSE, ggplot2” R package (Version 4.3.1) was utilized for Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (Yu et al. 2012). A P-value < 0.05 was considered statistically significant.

Screening and validation of critical markers

Genes in the module were selected as candidate hub markers using GS > 0.50 and MM > 0.80 as the screening criteria. The significant module genes were intersected with the differentially expressed Arraystar-derived mRNA. The overlapping DEGs were subjected to least absolute shrinkage and selection operator (LASSO) algorithm analysis with the “glmnet” package (Version 4.3.1) to screen the hub genes as critical markers. Finally, the expression levels and the receiver operating characteristic (ROC) analysis of the hub genes were used to evaluate the diagnostic value of the identified hub genes in the training dataset and validation dataset. The GSE16515 dataset was selected as the validation cohort to evaluate the expression level and diagnostic value of the pivotal genes by using ROC curves and area under the curve (AUC) values. A P-value < 0.05 was considered to indicate statistical significance.

Correlation analysis between the critical markers and clinical features

To further verify whether the survival of PC patients was related to the screened hub genes, the datasets downloaded from the TCGA, including the clinical information and transcriptome data of 169 PC samples, were used to estimate the relationships between the hub genes and survival rate. Survival curves of the hub genes were plotted with the “survival” R package (Version 4.3.1). A P-value < 0.05 for the survival-related hub genes was considered to indicate statistical significance. Correlations between candidate biomarker genes and clinical parameters were also evaluated. Univariate and multivariate Cox regression analyses were conducted with clinical features, including age, sex, grade, and stage, to analyze the independent prognostic value of the survival-related genes.

Expression profiles of mRNA in clinical samples

In total, 16,091 mRNAs were detected using Arraystar RNA microarrays in plasma samples. With a |logFC|≥1 and P < 0.05, 228 mRNA were upregulated, and 315 mRNA were downregulated in the PC patients compared with the healthy controls (HC). The top 20 significant genes are listed in Supplementary Table S1. The significantly DEGs are presented with volcano plots (Fig. 2).

Construction of a weighted gene co-expression network

The R package ‘WGCNA’ was utilized to construct weighted gene co-expression networks based on the merged GSE62452 and GSE71989 databases, and a soft threshold power of β = 12 (R² = 0.87) was set to preserve the scale-free network. A total of 3 co-expression modules of DEGs were obtained, of which the brown module (correlation coefficient = 0.84, P = 1.3e − 44) exhibited a significant biological association with PC (Fig. 3A-B). The key module containing a total of 71 DEGs was screened according to GS > 0.5 and MM > 0.8 (Figure. 3C-D) for subsequent biomarker screening analysis (Supplementary Table S2).

Functional enrichment analysis of significant module genes

To elucidate the functions and potential pathways associated with the brown module DEGs, functional enrichment analysis was conducted. GO analysis revealed that a total of 71 DEGs in the key module were involved mainly in chromosome segregation, nuclear division, organelle fission, and cell–substrate junctions (Fig. 4A). KEGG analysis indicated that these genes were also enriched in the extracellular matrix (ECM)–receptor interaction, focal adhesion, and PI3K-Akt signaling pathways (Fig. 4B).

Identification and validation of the diagnosis-related genes

The “Venn Diagram” R package (Version 4.3.1) was used to screen the potential hub genes between the significant co-expression module and the DEGs in our clinical microarray dataset. The DEGs were intersected with the brown module genes to obtain 8 intersecting genes, including FERMT1, S100A14, PLEK2, KCNN4, RHBDL2, PKM, UBE2T, ITGA3 (Fig. 5A). Least absolute shrinkage and selection operator (LASSO) Cox analysis was further applied to identify the most valuable differentially expressed mRNA (DEmRNA) from the intersection genes in the 2 datasets. In total, 5 possible diagnostic markers were identified, namely, FERMT1, S100A14, KCNN4, PKM, and ITGA3 (Fig. 5B).

Subsequently, we verified the expression levels and diagnostic value of the 5 possible diagnostic biomarkers in merged training datasets GSE62452 and GSE71989. As shown in Fig. 6A, FERMT1, S100A14, KCNN4, PKM, and ITGA3 were all overexpressed in tumor tissue (P < 0.05). According to the ROC analysis, the AUCs fluctuated between 0.806 and 0.892 (FERMT1: AUC = 0.865, S100A14: AUC = 0.806, KCNN4: AUC = 0.892, PKM: AUC = 0.855, ITGA3: AUC = 0.845), which represents effective diagnostic value (Fig. 6B).

Additionally, the GSE16515 dataset was used to further validate the expression levels and diagnostic value of these possible diagnostic biomarkers. The results indicated that the expression levels of FERMT1, S100A14, KCNN4, PKM, and ITGA3 were significantly increased in the PC group compared to adjacent non-tumor tissues (Fig. 7A). In addition, the AUC values of the ROC curves were 0.939 for FERMT1, 0.943 for ITGA3, 0.918 for S100A14, 0.939 for KCNN4, and 0.896 for PKM, which suggested that these genes have high diagnostic value for PC (Fig. 7B).

Clinical correlation analysis of survival-related genes

The survival R package was used to estimate the prognostic value of potential biomarkers in PC patients. Survival rate curves revealed that increased expression of FERMT1, S100A14, KCNN4, PKM, and ITGA3 was related to a poor survival rate (P < 0.05) (Fig. 8).

We performed the correlation analysis between the expression of the hub genes and clinical parameters from the TCGA database. Clinical information, including sex, age at initial diagnosis, tumor stage, tumor (T) stage, node (N) stage, and the role of related genes in prognosis, was evaluated by univariate and multivariate Cox proportional hazards regression.

Univariate Cox regression analysis revealed that the FERMT1, S100A14, and ITGA3 genes had significant effects on overall survival in PC patients (Table 2), the result indicated expression of FERMT1, S100A14, and ITGA3 was an independent risk factor for poor prognosis (hazard ratio [HR] > 1, P < 0.05). In addition, further multivariate Cox analysis indicated that lymph node metastasis was an independent prognostic factor for PC related to FERMT1 (HR) = 1.960, 95% confidence interval [CI]: 1.111–3.455, P < 0.05), S100A14 (HR = 2.159, 95% CI: 1.219–3.824, P < 0.05), and ITGA3 (HR = 2.099, 95% CI: 1.191–3.697, P < 0.05) (Fig. 9).

Table 2

Univariate proportional hazard regression
Genes	HR (95% CI)	P value
FERMT1	1.031 (1.011–1.051)	0.002
S100A14	1.001 (1.000–1.003)	0.034
KCNN4	1.006 (0.999–1.012)	0.093
PKM	1.001 (0.999–1.003)	0.341
ITGA3	1.004 (1.000–1.001)	0.038

HR, hazard ratio; CI, confidence interval.

PC is an aggressive malignant tumor that often lacks early symptoms and is associated with a poor prognosis (Hruban et al. 2019). Although significant progress has been achieved in surgical resection and chemotherapy, these advancements have not translated into effective therapies for the majority of patients. Consequently, there is an immediate need to explore new biomarkers and therapeutic strategies for PC patients. Currently, the mechanisms underlying gene regulation that contribute to disease progression are not fully understood, and prognostic evaluations for PC patients remain unsatisfactory.

With the development of bioinformatics, several studies have explored potential prognostic and diagnostic biomarkers related to PC (Muramatsu et al. 2024, Xu et al. 2023, Zhou et al. 2019). Furthermore, researchers identified a set of prognostic RNA expression signatures based on a large-scale meta-analysis (Kordshouli et al. 2024, Klett et al. 2018, Haider et al. 2014). Therefore, we performed an initial investigation aimed at identifying hub genes to construct specific multigene prognostic signatures that could enhance the prediction of clinical outcomes in PC. In this study, we analyzed transcriptome data related to PC from clinical plasma samples alongside public databases. Using WGCNA on the merged datasets GSE62452 and GSE71989, we identified significant gene modules associated with PC. Furthermore, KEGG pathway and GO enrichment analyses of the genes in the identified modules revealed their involvement in ECM–receptor interaction processes and the PI3K-Akt signaling pathway.

After overlapping the DEGs of our clinical microarray dataset and the significant modules, we obtained 8 genes that were highly expressed in patients with PC. Then, through LASSO regression, we identified 5 genes: FERMT1, S100A14, KCNN4, PKM, and ITGA3. The validation dataset GSE16515 confirmed the elevated expression of these 5 hub genes in tumors, suggesting their potential diagnostic utility for PC.

Several studies have sought to combine key gene expressions with clinical parameters to develop models that aid in predicting patient prognosis. Survival analyses of the TCGA database revealed that higher expression of these genes was related to a poor survival rate. Moreover, according to the multivariate model, the expression of FERMT1, S100A14, and ITGA3 was an independent risk factor for poor prognosis. In this study, we examined the correlations between the expression of prognosis-related genes and clinicopathological characteristics. The results indicated that lymph node metastasis was an independent prognostic factor in PC patients.

Our findings contribute further evidence for the importance of the key molecules currently under investigation for PC. Based on our data, several studies have reported associations between these hub genes and the progression of PC to a certain extent. Metastasis is the main cause of cancer-associated death. Activation of the epithelial–mesenchymal transition (EMT)-associated program is regarded as a central driver of tumorigenesis to metastasis (Krebs et al. 2017, Bhutia et al. 2020). Currently, the role of EMT in KRAS-driven PC development has been intensively studied (Bhutia et al. 2020, Wang et al. 2015, Paul et al. 2023). Previous studies have indicated that S100A14 and ITGA3 may be important determinants of EMT in PC. Most S100 proteins are considered EMT facilitators or EMT markers in carcinoma cells (Al-Ismaeel et al. 2019, Ou et al. 2021). The depletion of S100A14 strongly increased cell invasion (Low et al. 2023). Consistent with the EMT and mesenchymal–epithelial transition states, ITGA3 (Liu et al. 2021)was reported to upregulate the expression of ZIP4 in a PC cell line through the ZEB1/YAP1-ITGA3 signaling axis to promote EMT plasticity. Fermitin family member 1 (FERMT1) is associated with immune cell infiltration, and its N6-methyladenosine modification might be a marker of the treatment response to PC chemotherapy (Wu et al. 2023). In addition, studies have demonstrated that FERMT1 functions as a methylation-driven gene and participates in PC oncogenesis (Deng et al. 2021, Nie et al. 2023).

Potassium calcium-activated channel subfamily N member 4 (KCNN4) is a type of intermediate conductance calcium-activated potassium channel, and numerous studies have demonstrated that KCNN4 closely relates to the invasion and metastasis in many types of tumors (Li et al. 2020, Xu et al. 2021, Ibrahim et al. 2021)]. KCNN4 has been shown to induce Ca²⁺ entry, which leads to the activation of nuclear factor kappa B signaling and contributes to PDAC growth and metastasis. Mo et al. (Mo et al. 2022) illustrated that KCNN4 promotes PDAC cell proliferation, migration, and invasion through the Ca²⁺/MET/AKT axis and identified KCNN4 as a potential therapeutic target in PDAC. As a new potential therapeutic target, the splicing of pyruvate kinase (PKM) is correlated with drug resistance in PC patients (Calabretta et al. 2016).

Numerous studies have affirmed the significance of hub genes in the landscape of PC. When considered alongside our previous analyses, data suggest that the differential expression of FERMT1, S100A14, KCNN4, PKM, and ITGA3 in tumor and normal tissues might be related to the underlying mechanisms of PC, which must be further explored.

However, this study also has several limitations. The results of this study are mainly based on a microarray of our clinical samples and bioinformatics analysis without in vitro or in vivo experimental validation. Further gene expression validation could strengthen our conclusions, and the genes involved require further evaluation at the protein level. In addition, functional and molecular experiments are needed to investigate the underlying mechanism involved. Therefore, future work will involve further elucidation and multi-center verification of the gene signature with larger clinical sample sizes.

This study identified signature genes as potential biomarkers for the diagnosis of PC and showed great value for clinical prognosis. These results may provide new insight into the etiology and prognosis of PC and provide therapeutic targets for this disease.

Author Contributions: Conception and design: C Liu, LQ Zhang; Provision of study materials or patients: ML Li, YQ You; Collection and assembly of data: C Liu, D Qian, Y Yang; Data analysis and interpretation: C Liu, D Qian, Y Tian; Final approval of manuscript: All authors.

Acknowledgments:We would like to acknowledge the reviewers for their helpful comments on this study.

Funding: This work was supported by the National Natural Science Foundation of China (Grant No. 81873981), Chongqing Municipal Science and Health Joint Medical Research Project Key Project (2024C040), and Chongqing Talent Program (CQYC20220303658).

Data Availability: The results presented here are in part based upon data generated by the TCGA Research Network (https://www.cancer.gov/tcga) and GEO database (https://www.ncbi.nlm.nih.gov/geo/).The microarray dataset used in this study are available from the corresponding author on reasonable request.

Conflicts of Interest: The authors have no conflicts of interest to declare.

Ethical Approval

The research procedures were approved by the Ethics Review Committee of the Second Affiliated Hospital of Army Medical University (Approval No.2024-154-01). Written consent was obtained from all participants, including patient and control individuals.

Al-Ismaeel Q et al (2019). ZEB1 and IL-6/11-STAT3 signalling cooperate to define invasive potential of pancreatic cancer cells via differential regulation of the expression of S100 proteins. Br J Cancer 121 (1):65-75. https://doi.org/10.1038/s41416-019-0483-9.
Bhutia YD et al (2020). Chronic exposure to excess iron promotes EMT and cancer via p53 loss in pancreatic cancer. Asian J Pharm Sci 15 (2):237-251. https ://doi.org/10.1016/j.ajps.2020.02.003.
Calabretta S et al (2016) Modulation of PKM alternative splicing by PTBP1 promotes gemcitabine resistance in pancreatic cancer cells. Oncogene 35 (16):2031-9. https://doi.org/10.1038/onc.2015.270.
Deng GC et al (2021) Identification of DNA methylation-driven genes and construction of a nomogram to predict overall survival in pancreatic cancer. BMC Genomics 22 (1):791. https://doi.org/10.1186/s12864-021-08097-w.
Dreyer SB et al (2022) Genomic and Molecular Analyses Identify Molecular Subtypes of Pancreatic Cancer Recurrence. Gastroenterology 162 (1):320-324 e4. https://doi.org/10.1053/j.gastro.2021.09.022
Feng RM et al (2019) Current cancer situation in China: good or bad news from the 2018 Global Cancer Statistics? Cancer Commun (Lond) 39 (1):22. https://doi.org/10.1186/s40880-019-0368-6.
Haider SJ. et al (2014) A multi-gene signature predicts outcome in patients with pancreatic ductal adenocarcinoma. Genome Med 6 (12):105. https://doi.org/10.1186/s13073-014-0105-3.
Hruban RH et al (2019) Why is pancreatic cancer so deadly? The pathologist's view. J Pathol 248 (2):131-141. https://doi.org/10.1186/s13073-014-0105-3.
Ibrahim SJ et al (2021) SK4 oncochannels regulate calcium entry and promote cell migration in KRAS-mutated colorectal cancer. Cell Calcium 96:102384. https://doi.org/10.1016/ j.ceca.2021.102384.
Kishikawa TM et al (2015) Circulating RNAs as new biomarkers for detecting pancreatic cancer. World J Gastroenterol 21 (28):8527-40. https://doi.org/10.3748/wjg.v21.i28.8527.
Kitagawa TK et al (2019) Circulating pancreatic cancer exosomal RNAs for detection of pancreatic cancer. Mol Oncol 13 (2):212-227. https://doi.org/10.1002/1878-0261.12398.
Klein, AP (2021) Pancreatic cancer epidemiology: understanding the role of lifestyle and inherited risk factors. Nat Rev Gastroenterol Hepatol 18(7):493-502. https://doi.org/10.1038/s41575-021-00457-x.
Klett HH et al (2018) Identification and Validation of a Diagnostic and Prognostic Multi-Gene Biomarker Panel for Pancreatic Ductal Adenocarcinoma. Front Genet 9:108. https://doi.org/10.3389/fgene.2018.00108.
Kordshouli, SO et al (2024) A comprehensive meta-analysis of transcriptome data to identify signature genes associated with pancreatic ductal adenocarcinoma. PLoS One 19 (2):e0289561. https://doi.org/10.1371/journal.pone.0289561.
Krebs AM et al (2017) The EMT-activator Zeb1 is a key factor for cell plasticity and promotes metastasis in pancreatic cancer. Nat Cell Biol 19 (5):518-529. https://doi.org/10.1038/ncb3513.
Langfelder P, Horvath S (2008) WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics 9:559. https://doi.org/10.1186/1471-2105-9-559.
Li QT et al (2020) KCNN4 promotes invasion and metastasis through the MAPK/ERK pathway in hepatocellular carcinoma. J Investig Med 68 (1):68-74. https://doi.org/10.1136/jim-2019-001073.
Liu MY. et al (2021) Zinc-Dependent Regulation of ZEB1 and YAP1 Coactivation Promotes Epithelial-Mesenchymal Transition Plasticity and Metastasis in Pancreatic Cancer. Gastroenterology 160 (5):1771-1783 e1. https://doi.org/10.1053/j.gastro.2020.12.077.
Low RRJ et al (2023) Correction to: S100 family proteins are linked to organoid morphology and EMT in pancreatic cancer. Cell Death Differ 30 (5):1400. https://doi.org/10.1038/s41418-023-01151-y.
Mizrahi JD et al (2020) Pancreatic cancer. Lancet 395 (10242):2008-2020. https://doi.org/10.1016/S0140-6736(20)30974-0.
Mo X et al (2022) KCNN4-mediated Ca(2+)/MET/AKT axis is promising for targeted therapy of pancreatic ductal adenocarcinoma. Acta Pharmacol Sin 43 (3):735-746. https://doi.org/10.1038/s41401-021-00688-3.
Muramatsu J et al (2024) Gap junction beta-4 accelerates cell cycle progression and metastasis through MET-AKT activation in pancreatic cancer. Cancer Sci. https://doi.org/10.1111/cas.16101.
Nie Y et al (2023) Prognostic utility of TME-associated genes in pancreatic cancer. Front Genet 14:1218774. https://doi.org/10.3389/fgene.2023.1218774.
Ou S et al (2021) S100A16 suppresses the proliferation, migration and invasion of colorectal cancer cells in part via the JNK/p38 MAPK pathway. Mol Med Rep 23 (2). https://doi.org/10.3892/mmr.2020.11803.
Paul MC et al (2023) Non-canonical functions of SNAIL drive context-specific cancer progression. Nat Commun 14 (1):1201. https://doi.org/10.1038/s41467-023-36505-0.
Pompella L et al (2020) Pancreatic Cancer Molecular Classifications: From Bulk Genomics to Single Cell Analysis. Int J Mol Sci 21 (8). https://doi.org/10.3390/ijms21082814.
Siegel RL et al (2021) Cancer Statistics, 2021. CA Cancer J Clin 71 (1):7-33. https://doi.org/10.3322/caac.21654.
Siegel RL et al (2023) Cancer statistics, 2023. CA Cancer J Clin 73 (1):17-48. https://doi.org/10.3322/caac.21763.
Thompson ED et al (2020) The genetics of ductal adenocarcinoma of the pancreas in the year 2020: dramatic progress, but far to go. Mod Pathol 33 (12):2544-2563. https://doi.org/10.1038/s41379-020-0629-6.
Wang L et al (2015) ATDC induces an invasive switch in KRAS-induced pancreatic tumorigenesis. Genes Dev 29 (2):171-83. https://doi.org/10.1101/gad.253591.114.
Wu Q et al (2023) FERMT1 Is a Prognostic Marker Involved in Immune Infiltration of Pancreatic Adenocarcinoma Correlating with m(6)A Modification and Necroptosis. Genes (Basel) 14 (3). https://doi.org/10.3390/genes14030734.
Xu J et al (2023) VSIG2 promotes malignant progression of pancreatic ductal adenocarcinoma by enhancing LAMTOR2-mediated mTOR activation. Cell Commun Signal 21 (1):223. https://doi.org/10.1186/s12964-023-01209-x.
Xu P et al (2021) KCNN4 promotes the progression of lung adenocarcinoma by activating the AKT and ERK signaling pathways. Cancer Biomark 31 (2):187-201. https://doi.org/10.3233/CBM-201045.
Yu G et al (2012) clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 16 (5):284-7. https://doi.org/10.1089/omi.2011.0118.
Zhou Q et al (2019) Quantitative proteomics identifies brain acid soluble protein 1 (BASP1) as a prognostic biomarker candidate in pancreatic cancer tissue. EBioMedicine 43:282-294. https://doi.org/10.1016/j.ebiom.2019.04.008.

No competing interests reported.

SupplementaryMaterial.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Identifying diagnostic markers and constructing a prognostic model for pancreatic cancer based on microarray and bioinformatic analysis

Status:

Version 1

Abstract

Figures

Introduction

Methods

Data collection

Clinical sample collection and array expression analysis

Co-expression analysis

Enrichment analysis

Screening and validation of critical markers

Correlation analysis between the critical markers and clinical features

Results

Expression profiles of mRNA in clinical samples

Construction of a weighted gene co-expression network

Functional enrichment analysis of significant module genes

Identification and validation of the diagnosis-related genes

Clinical correlation analysis of survival-related genes

Discussion

Conclusions

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1