• Users Online: 203
  • Print this page
  • Email this page

 Table of Contents  
Year : 2021  |  Volume : 5  |  Issue : 3  |  Page : 286-294

Identification of potential common molecular factors of pancreatic cancer and diabetes mellitus using microarray data analysis combined with bioinformatics techniques and experimental validation

1 Department of Biotechnology, School of Advanced Technologies in Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran
2 Liver and Pancreatobiliary Diseases Research Center, Digestive Diseases Research Institute, Tehran University of Medical Sciences, Tehran, Iran
3 Department of Medicine, University of Minnesota Medical School, Minneapolis, USA
4 Proteomics Research Center; Department of Tissue Engineering and Applied Cell Sciences, School of Advanced Technologies in Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran
5 Cellular and Molecular Biology Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran
6 Department of Hematology, Faculty of Medical Sciences, Iran University of Medical Sciences, Tehran, Iran
7 Digestive Oncology Research Center, Digestive Diseases Research Institute, Tehran University of Medical Sciences, Tehran, Iran
8 Department of Biology and Biochemistry, University of Bath, Bath, UK
9 Department of Biotechnology, School of Advanced Technologies in Medicine; Cellular and Molecular Biology Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran
10 Oncopathology Research Center, Iran University of Medical Sciences, Tehran, Iran

Date of Submission01-Jun-2021
Date of Acceptance05-Aug-2021
Date of Web Publication7-Sep-2021

Correspondence Address:
Bahram Kazemi
School of Advanced Technologies in Medicine, Shahid Beheshti University of Medical Sciences, Tehran
Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/bbrj.bbrj_122_21

Rights and Permissions

Background and Aim: Detection of early-stage pancreatic cancer (PaC) patients can decrease the mortality rate through surgical resection thanks to a screening approach of high-risk and asymptomatic individuals. Up to 80% of PaC patients are either hyperglycemic or diabetic at the time of diagnosis. Diabetes mellitus (DM) identification as an early manifestation of PaC may lead to detection of this malignancy at early and presymptomatic stages. Therefore, the current study aims to identify common molecular factors between DM and PaC to shed light on a potential screening for DM as a diagnostic approach for PaC. Methods: The study was divided into three steps including dataset selection and identification of key genes, quantitative real-time polymerase chain reaction (qRT-PCR) to validate these genes, and enrichment analysis of their target genes. Results: Using GEO2R analysis, conducted on microarray dataset (GSE15932), SPI1 and Yin Yang 1 (YY1) genes were found to be expressed in peripheral blood mononuclear cells of PaC, DM, and PaC + DM patients. Moreover, qRT-PCR results in validation of microarray data showed a significant increment of these two genes among all patient groups. Furthermore, enrichment analyses of SPI1 and YY1 target genes revealed that they are mainly expressed in hematopoietic cells and associated with immune responses as well as immune signaling pathways. Discussion: We speculate that this study on SPI1 and YY1 genes and their targets can result in a successful strategy to investigate diabetes as a screening step for better management of PaC screening using standard serologic tests.

Keywords: Blood test, diabetes mellitus, pancreatic cancer, screening

How to cite this article:
Kalantari S, Pourshams A, Roudi R, Zali H, Bandehpour M, Kalantari A, Ghanbari R, D'Angelo A, Kazemi B, Madjd Z. Identification of potential common molecular factors of pancreatic cancer and diabetes mellitus using microarray data analysis combined with bioinformatics techniques and experimental validation. Biomed Biotechnol Res J 2021;5:286-94

How to cite this URL:
Kalantari S, Pourshams A, Roudi R, Zali H, Bandehpour M, Kalantari A, Ghanbari R, D'Angelo A, Kazemi B, Madjd Z. Identification of potential common molecular factors of pancreatic cancer and diabetes mellitus using microarray data analysis combined with bioinformatics techniques and experimental validation. Biomed Biotechnol Res J [serial online] 2021 [cited 2021 Nov 27];5:286-94. Available from: https://www.bmbtrj.org/text.asp?2021/5/3/286/325605

  Introduction Top

Pancreatic cancer (PaC), one of the most aggressive solid tumors, accounts for the lowest overall survival rate among cancers.[1] This lethal malignancy is the seventh leading cause of cancer-related deaths worldwide[2] and the eighth and ninth leading causes of cancer death in males and females, respectively.[3] Clinical manifestation of PaC includes local invasion, early metastasis, and resistance to standard chemotherapy.[1] The majority of PaC originates from the ductal epithelium (adenocarcinoma) while, infrequently, from neuroendocrine cells as well as other types such as acinar, anaplastic, and adenosquamous cells (5%). Pancreatic ductal adenocarcinoma (PDAC) has a high mortality, which stems from the rapid dissemination of tumor cells and leads to widespread metastasis. Nowadays, there is some evidence that PDAC might evolve from noninvasive duct lesions as pancreatic intraepithelial neoplasia (PanIN).[4],[5] Local tissue infiltration and lymphatic invasion are evident in early PDAC, and circulating tumor cells in the bloodstream ultimately are hypothesized to result in cancer dissemination and metastasis.[6]

The majority of PaC patients have a median survival rate of 6 months whereas, in some cases, the survival rate can reach approximately 5 years (3%). This poor outcome is due to the lack of early diagnostic markers and efficient therapeutic tools as well as the aggressive nature of the disease.[3],[7] The treatment options available are chemotherapy and surgery although the former is often either ineffective or effective for a limited time only. Since PaC is generally diagnosed at an advance stage (locally advance or metastatic stage), only a few patients are considered eligible for surgical resection.[7],[8],[9] Therefore, the improvement of long-term survival by early detection of PaC would be the most beneficial and effective strategy for cancer management.

Various factors including genetic susceptibility and environmental factors are thought to contribute to PaC development and progression. Family history of PaC and cigarette smoking are the most significant and well-known risk factors.[10] Molecular and genetic alterations are also involved in the progression of PaC.[8] Several reports have demonstrated that up to 80% of PaC patients are either hyperglycemic or diabetic at the time of diagnosis.[11] A recent case–control study conducted by Sharma et al. revealed that the levels of blood sugar in PaC patients were elevated for up to 3 years before PaC diagnosis.[12] The risk of PaC in individuals in their fifties (>50 years) with new-onset diabetes (NOD) mellitus is eight times higher than the general population.[13],[14] A population-based study demonstrated that nearly 1% of NOD patients (≥50 years old) will be diagnosed with PaC within 3 years of the first met criteria for NOD and 56% of these within 6 months of meeting the criteria for NOD.[15] These studies suggest a close relationship between diabetes mellitus (DM) and PaC although it is still poorly understood whether DM is either a predisposing factor for PaC or the outcome of cancer progression. Therefore, recognition of a potential molecular relationship between PaC and DM may help screening of high-risk PaC individuals.

The blood gene expression profiling may offer a reliable and less invasive strategy to identify tumor biomarkers with high specificity and sensitivity. Currently, the best blood tumor biomarker for PaC is carbohydrate antigen 19-9 (CA19-9), with an overall 80% sensitivity and 82% specificity; CA19-9 is a sialylated Lewis antigen of MUC1 (mucin 1), and patients with the blood type of Lewis a- and b-genotype are unable to synthesize the CA19-9 epitope.[16] Furthermore, among patients with resectable PaC, CA19-9 levels are not elevated in up to 35% of the patients.[17] The increment of CA19-9 levels is also observed in cirrhosis, chronic pancreatitis, cholangitis, and obstructive jaundice as well as in other digestive tract cancers.[18] Therefore, all these issues limit the reliability of CA19-9 as a specific PaC biomarker. In the absence of effective biomarkers, the main goal for the identification of risk factors associated with PaC would be the screening of high-risk PaC individuals and the consequent improvement of the survival rate of PaC patients. As mentioned earlier, because of its potential association with PaC, DM can be considered as a risk factor for this lethal malignancy. Hence, in the current study, the identification of common molecular factors between PaC and DM has been investigated, with the ultimate goal of early-stage PaC screening and diagnosis using blood tests.

  Methods Top

Study design

The study was divided into three main steps. First, appropriate GSE of microarray data – based on the study's criteria (a high throughput study to compare blood gene expression profile between PaC, DM, and Pac + DM patients) – was selected and preprocessed to identify key genes that were common among the three groups. Then, quantitative real-time polymerase chain reaction (qRT-PCR) was used to validate identified genes. Finally, the target genes of validated genes obtained from the previous step were further investigated and enrichment analysis of target genes was performed. The workflow diagram of the current study is shown in [Figure 1].
Figure 1: The workflow of study. Degree cutoff: 2, node score cutoff: 0.2, K-core: 2

Click here to view

Gene expression omnibus data

Preprocessing of gene expression data

The transcription dataset deposited by Huang et al.[19] was retrieved from the Gene Expression Omnibus DataSets database (www.ncbi.nlm.nih.gov/gds) (Series GSE15932) in which the Platform Affymetrix Human Genome U133 Plus 2.0 Array (GPL570) was used. A total of 32 human blood samples including 8 patients with DM, 8 patients with PaC, 8 patients with PaC plus DM, and 8 paired controls without any history of diseases were used for processing the microarray data to identify differential expression genes (DEGs).

Identification of differential expression genes

DEG analysis of the dataset was performed using the online GEO engine, GEO2R (www.ncbi.nlm. nih.gov/geo/geo2r/), which normalized the data using the GEOquery and limma R packages. Top 250 ranked significantly differentially expressed genes were determined through an agreed P ≤ 0.05 and fold change cutoff of ≥2 as a threshold. The correction test of Benjamini and Hochberg (false discovery rate) also was considered for the result assessment.

Construction of gene network

Identified DEGs (obtained from DM, PaC, and PaC + DM groups) were pooled and uploaded into STRING database (http://string-db.org) for the interaction network assembly. All disconnected nodes were excluded while the experimentally validated interactions were loaded into Cytoscape v3.7.0 (http://cytoscape.org/). Subsequently, initial investigation was followed by the application of Network Analyzer (plugged in Cytoscape) to analyze the network in terms of well-known key features such as degree and betweenness centrality, to examine the potential most significative genes. Furthermore, the ClusterViz application (plugged in Cytoscape) was included in the network analyses for the screening of modules and subnetworks based on Molecular Complex Detection (MCODE) algorithm.

Experimental validation

Study population

To verify the potentially effective genes identified from the analyzed network, the blood samples of patients admitted in the gastroenterology department of Shariati Hospital (Tehran, Iran) were analyzed. As was reported in our previous study,[10] to identify PaC cases, conventional methods such as endoscopic ultrasonography, computed tomography scan, fine-needle aspiration, hematoxylin/eosin as well as immunohistochemistry staining were performed. Detection of pancreatic mass and tumor staging were reconfirmed by a second opinion from an expert gastroenterologist and a pathologist. Then, PaC cases at early stage (I/II) before surgery or any other medical intervention were selected. For DM cases, diagnostic tests were performed according to the American Diabetes Association criteria for the diagnosis including fasting plasma glucose ≥126 mg/dL (7.0 mmol/L), 2-h plasma glucose (2-h PG) ≥200 mg/dL (11.1 mmol/L) during oral glucose tolerance test, and A1C ≥6.5% (48 mmol/mol).[20] Participants who met the aforementioned criteria plus diabetes duration of <10 years were considered as DM patients. Individuals whose PaC and DM conditions were diagnosed and reconfirmed by a specialist were classified as PaC + DM cases. PaC cases were also tested for DM and confirmed to be nondiabetic. For comparative analysis, healthy individuals with similar age and gender were also included in the study.

In addition, a total of 48 blood samples from 24 male and 24 female volunteers, divided into four groups (DM, PaC, PaC + DM, and healthy individuals), were collected. The demographic information and clinical characteristics for each group are reported in [Table 1]. An informed consent form was signed by all donors, and it was approved by the ethical committees of Shahid Beheshti University of Medical Sciences with the number IR.SBMU.RETECH. REC.1395.398. Sample collection procedures performed in this study involving human participants were in full compliance with the institutional and/or national research committee of the 2013 Helsinki Declaration.
Table 1: Demographic information and clinical characteristics of patients and healthy individuals

Click here to view

RNA isolation and cDNA synthesis

A volume of 300 μl fresh peripheral whole blood was collected into EDTA tube, and total RNA was isolated using QIAzol Lysis Reagent (Qiagen, Hilden, Germany) by standard protocol. Potential DNA contamination of extracted RNA was removed by administering DNase I, RNase-free Kit (Thermo Scientific, Waltham, Massachusetts, US). cDNA synthesis was immediately carried out using AccuPower RocketScript RT PreMix (Bioneer Inc., South Korea), according to the manufacturer's structure. NCBI BLAST and Oligo7 software (Molecular Biology Insights, Inc., Cascade, CO), as well as AlleleID7.5 (PREMIER Biosoft, San Diego, CA), were employed to design and validate primer sets. The amplification conditions were optimized for each primer set, and β-actin was used as our internal control. The sequences of the primer sets are listed in [Table 2].
Table 2: List of primers used for real-time quantitative reverse transcription-polymerase chain reaction

Click here to view

Quantitative real-time polymerase chain reaction

The RealQ Plus 2x Master Mix Green (Ampliqon Co., Denmark) was used for qRT-PCR, and reactions were prepared in a total volume of 23 μl. The qRT-PCR program was carried out according to the following program: 15 min at 95°C, followed by 40 cycles (20 s each cycle) at 95°C for denaturation, and 1 min at 60°C for the annealing and extension steps. Eventually, to validate the specificity of the expected PCR product and nonoccurrence of primer-dimer formation, a melting curve analysis was performed. All samples were carried out in triplicate, and real-time data were analyzed by threshold cycle[21] values.

Gene ontology enrichment and pathway analysis

The DEG transcription factors (TFs) identified by GEO2R analysis and further validated by qRT-PCR were additionally investigated using TRRUST database (https://www.grnpedia.org/trrust) for the screening of TF target genes. Afterward, target genes were pooled and construction of protein–protein interaction (PPI) network and GO enrichment were carried out using FunRich v3.1.3 (http://www.funrich.org) and Metascape (http://metascape.org) tools with default parameters. Moreover, MCODE algorithm was employed to detect functional modules, highly interconnected regions, and protein complexes for these target genes with default parameters in Metascape. The MCODE cluster with the highest score was further investigated with GO enrichment analyses using the WebGestalt database (http://www.webgestalt.org), to explore biological processes and pathways analysis.


The RT-PCR quantification was carried out in triplicate for each sample. RT-PCR data analysis was conducted using Relative Expression Software Tool 2009 for Ct values and GraphPad Prism 8 for analysis of variance test. The numerical results were reported as the mean ± standard deviation, and significant P values were indicated as *P < 0.05, **P < 0.01, and ***P < 0.001.

  Results Top

SPI1 and Yin Yang 1 genes identified as common differential expression genes across diabetes mellitus, pancreatic cancer, and pancreatic cancer + diabetes mellitus patients

The microarray dataset GSE15932 obtained from the GEO database was analyzed by GEO2R, and the 250 top DEGs from each group – DM, PaC, and PaC + DM – were identified and combined. A detailed list of DEGs from DM, PaC, and PaC + DM groups are reported in [Supplementary Table 1]. Identified DEGs, following the exclusion of repetitive genes, were transferred to STRING platform for the PPI network assembly. Then, the PPI network was further analyzed by Network Analyzer (plugged in Cytoscape), and potential top effective genes – according to degree and betweenness centrality – were eventually identified. Among 40 hub genes with a high degree of connectivity (connection degree value higher than 10), only two genes (SPI1, degree = 22 and Yin Yang 1 [YY1], degree = 10) were shared between the three groups under evaluation. The 40 top DEGs identified with statistically significant values are reported in [Table 3]. In addition, ClusterViz (plugged in Cytoscape) based on MCODE algorithm identified 15 gene clusters from the PPI network based on node score cutoff = 0.2, degree cutoff = 2, and K-core = 2. Among these clusters, cluster 3 had 22 nodes including SPI1 and YY1 and 69 edges where YY1 was identified as a seed node. Consequently, we selected SPI1 and YY1 genes for the experimental validation stage. The gene list of each cluster is reported in [Table 4], and cluster 3 is illustrated in [Supplementary Figure 1].
Table 3: Forty top identified differential expression genes with significant values, analyzed by Cytoscape, based on the degree (connection degree > 10) and betweenness centrality parameters

Click here to view
Table 4: All clusters data obtained by ClusterViz (using Molecular Complex Detection algorithm)

Click here to view

Quantitative real-time polymerase chain reaction validated SPI1 and Yin Yang 1 gene expression increment in blood samples of patient groups

Bioinformatics analysis of microarray dataset (GSE15932) showed that SPI1 and YY1 genes were upregulated in DM, PaC, and PaC + DM samples in comparison to normal samples [Supplementary Table 2]. To confirm these preliminary results, qRT-PCR experiments were carried out to evaluate the mRNA expression of SPI1 and YY1 genes for patients included in our study in comparison with healthy controls. As shown in [Figure 2], transcript levels of SPI1 and YY1 genes were found to be significantly (P < 0.05) increased within three patient groups (DM, PaC, and PaC + DM) compared with normal controls.
Figure 2: Differentially expression levels of SPI1 and YY1 blood samples of patient groups versus healthy samples. Their transcript levels were found to be significantly increased in DM, PaC, and PaC + DM groups compared with normal group. *P < 0.05. DM: Diabetes mellitus, PaC: Pancreatic cancer, PaC + DM: Pancreatic cancer and diabetes mellitus

Click here to view

SPI1 and Yin Yang 1: Transcription factors with different target genes

Because both SPI1 and YY1 genes are TFs with different target genes, we first looked for these target genes. TRRUST database identified 61 and 91 target genes for SPI1 and YY1, respectively. These targets were eventually pooled and used for GO enrichment and pathway analyses, to identify their biological role. The list of target genes is reported in [Supplementary Table 3].

SPI1 and Yin Yang 1 target genes have a crucial role in immune response

The main biological pathways identified by FunRich tool were the tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) signaling pathway, GMCSF-mediated signaling events, IFN-γ pathway, α9β1 integrin signaling events as well as endothelin signaling with a similar percentage of target gene contribution. [Figure 3] illustrates the percentage of target genes for each biological pathway. Catalog Of Somatic Mutations In Cancer (COSMIC), the site of expression, and clinical phenotype of target genes were also found to be enriched using FunRich, as reported by the pie chart in [Figure 4].
Figure 3: The percent of target gene contribution in biological pathways

Click here to view
Figure 4: The percent of target gene contribution in clinical phenotype, site of expression, and Catalog of Somatic Mutations in Cancer target genes

Click here to view

PPI network assembly of target genes was performed using BioGrid database in Metascape. Enrichment analysis of this PPI network revealed that these target genes were mainly enriched in interleukins (R-HAS-449147), myeloid cell differentiation (GO: 0030099), and cytokine–cytokine receptor interaction (has04060) pathways, respectively. The top 20 enriched pathways of target genes are shown with bar graph in [Figure 5]. Using the MCODE algorithm, densely connected network components were identified and four MCODE clusters with 29 hub genes were observed. Pathway analysis of these 29 hub genes revealed that the three best-scoring pathways were those associated with interleukins (R-HSA-449147), cytokine-mediated signaling pathway (GO: 0019221), and interleukin-4 and interleukin-13 signaling (R-HSA-6785807). [Figure 6] shows these MCODE clusters as subnetworks. Of these four MCODE clusters, MCODE 1 (GATA1, SMAD7, MYC, HDAC1, HDAC2, STAT3, TP53, CREBBP, PARP1, TP73, HIF1A, and BRCA1) reported the highest MCODE score. Therefore, the hub genes of MCODE 1 cluster – through WebGestalt tool – were interpreted, and the positive transcription regulation by RNA polymerase II (GO: 0045944) showed the highest enrichment ratio (13.784). The results of WebGestalt tool analysis are shown as a volcano plot in [Figure 7].
Figure 5: Twenty top enriched clusters of target genes, analyzed by Metascape, visualized and colored by P-values, and illustrated as a bar graph

Click here to view
Figure 6: Four MCODE clusters with 29 hub genes obtained from target gene analysis using MCODE algorithm. MCODE 1 cluster has the highest MCODE score

Click here to view
Figure 7: Enrichment results of MCODE 1 cluster hub genes, visualized and colored by false discovery rate and illustrated as a volcano plot

Click here to view

  Discussion Top

In recent years, peripheral blood mononuclear cells (PBMCs) have emerged as substitute markers of several diseases including inflammatory and malignant diseases.[21],[22],[23],[24],[25] Many tumor markers, such as CA19-9, are released from tumor tissue and increase with greater tumor burden.[26] PBMCs, when in contact with the microenvironment of this altered tissue, can induce a complex transcriptional response and modify their gene expression profile.[23],[24],[25] Hence, gene expression profiling of peripheral blood may offer a reliable tool to investigate cancer markers. However, the role of peripheral blood as an accessible source for the diagnosis and prognosis of solid tumors remains limited and poorly investigated.[23]

Some studies investigated the gene expression profile of PBMCs in patients with PaC using microarray technique[19],[21],[27] to identify novel gene subsets potentially useful for a fast diagnosis of this malignancy. Huang et al. performed microarray-based gene expression profiling of PBMCs in patients with PaC-associated DM as well as patients with DM type 2 to explore specific blood biomarkers for early detection of PaC.[19] DM, as a considerable risk factor and early manifestation of PaC, has been widely studied[12],[28] to identify molecular biomarkers via screening procedures. Therefore, we employed high throughput experiment data and system biology tools as well as experimental validation to compare the blood gene expression profile of PaC and DM patients. Our study is a secondary study of data already published by Huang et al. in 2010[19] as the original research article “Novel blood biomarkers of pancreatic cancer-associated diabetes mellitus identified by peripheral blood-based gene expression profiles.” The purpose of our study is to identify potential common molecular factors between PaC and DM using microarray data analysis combined with bioinformatics techniques and experimental validation. The results from our study may provide the groundwork to identify molecular factors for PaC screening in high-risk individuals.

Huang et al. aimed to differentiate between PaC-associated DM and DM type 2; they explored specific biomarkers to achieve this goal. Eventually, they concluded that a panel of VNN1 and MMP9 genes could discriminate between PaC-associated DM from DM type 2.[19] In contrast, we aimed to explore the potential molecular similarity between PaC and DM, considering previous studies that claimed DM to be a risk factor and/or consequence of PaC.[11],[29],[30],[31],[32],[33],[34],[35],[36],[37],[38],[39] Initially, a list of 40 top identified DEGs using pooled data from DM, PaC, and PaC + DM groups from Huang et al. study were obtained [Table 3]. Among these 40 hub genes with high connection degree, only SPI1 and YY1 genes were shared between the three groups under evaluation. Furthermore, clustering analysis using the MCODE algorithm identified YY1 as a seed node for cluster 3, which is composed of 22 nodes including SPI1 and YY1 [Table 2] and [Supplementary Figure 1]. Thus, these two hub genes were selected for experimental validation.

SPI1 gene encodes PU.1 protein which is a TF of the ETS family with a critical role in hematopoietic differentiation.[40] SPI1/PU.1 axis has a pivotal effect on the limitation of hematopoietic stem cells and progenitors self-renewal;[41] also, it is a key regulator in the commitment and maturation of myeloid and B lymphoid lineage.[42] However, overexpression of SPI1 has an oncogenic role in proerythroblasts[42] and its expression or deregulation contributes to leukemia generation, where it can exert either an oncogene or tumor suppressor role.[41] On the contrary, in some cases, the downregulation of SPI1 in children and adolescents patients might result in lymphomas as reported by Özdemir et al. in 2018.[40] Another study supposed that senescence is an antiproliferative mechanism induced by SPI1 that could be protective against the development of acute myeloid leukemia.[41] Furthermore, more recent research has been carried out explaining the role of SPI1 in the progression of glioma via its target genes through multiple signaling pathways.[43] Overall, SPI1 is not only an effective gene in hematopoietic linage development and leukemia, but it might be also involved in other malignant cancer progressions such as glioma.[40]

Yin Yang 1 is a TF of the GLI-Kruppel family of nuclear proteins with an important role in various biological activities such as cell proliferation, angiogenesis, and metastasis.[44] The role of YY1 as a critical regulator of the development of early B-cell,[45] as well as in the progression of a variety of cancers including prostate, ovarian, and colorectal cancer, have been reported in several studies.[44] The overexpression of YY1 TF in pancreatic tumor can inhibit invasion and migration of PaC. Indeed, YY1 is highly expressed in PaC with an inhibitory effect.[44] Participation of YY1 in hepatic gluconeogenesis regulation was uncovered by Lu et al. in 2013, suggesting that YY1 might be a therapeutic target for hyperglycemia in diabetes.[46] Furthermore, another study revealed that YY1 upregulation plays an essential role in renal fibrosis of diabetic nephropathy, suggesting that the decrement of YY1 expression may represent a new therapeutic target for renal fibrosis induced by diabetic nephropathy.[45]

Investigating GEO data analysis and several reports about the role of SPI1 and YY1 in various biological activities (mentioned earlier), we found that they play a significant role in both PaC and DM. Hence, the expression level of these two potentially effective TFs (SPI1 and YY1) in blood samples from DM, PaC, and PaC + DM patients, as well as healthy controls, was investigated. The qRT-PCR technique was employed for the validation of achieved results using both GAPDH and β-actin genes as internal controls. Our results showed significant overexpression of GADPH gene in PaC patients' blood, supporting the results from other studies reporting upregulation of GAPDH in PaC.[27],[47],[48] Therefore, the qRT-PCR method was performed using only β-actin as an internal control. The mRNA expression of SPI1 and YY1 genes in all patients (DM, PaC, and PaC + DM) – in comparison with healthy controls – were found to be significantly (P < 0.05) increased [Figure 2], corroborating the microarray data in the literature by Huang et al. [Supplementary File 2].

In the last step of our study, enrichment analyses were performed on target genes of SPI1 and YY1 to comprehensively explore biological pathways associated with these genes. Our FunRich enrichment analysis of target genes identified TRAIL signaling as one of the most relevant pathways [Figure 3]. TRAIL is a cytokine with an anticancer role, capable of initiating apoptotic pathway by binding to its receptors (TRAILR1 and TRAILR2). The cell-killing effect of TRAIL in pancreatic tumor cells has been explored in numerous studies.[49],[50] It has also been shown that pancreatic tumor cells are responsive to the anticancer effect of TRAIL by activating nonapoptotic and proinflammatory signaling pathways and via NF-κB, PKC, and ERK1/ERK2 activation.[49] Enrichment analysis of target genes also indicated that blood cells are the main expression site of these genes. Moreover, immune cells, as well as hematopoietic and lymphoid tissues, were considered in clinical phenotype and COSMIC enrichment, respectively [Figure 4], suggesting that SPI1 and YY1 can be effective TFs for the identification of blood diagnostic markers. The PPI network analysis of target genes in the Metascape database confirmed these results identifying interleukins, myeloid cell differentiation, and cytokine–cytokine receptor interaction as main enriched and activated signaling pathways [Figure 5]. As shown in [Figure 6], four MCODE clusters were obtained through PPI network analyses using MCODE algorithm, in which MCODE cluster 1 had the highest MCODE score. GATA1, SMAD7, MYC, HDAC1, HDAC2, STAT3, TP53, CREBBP, PARP1, TP73, HIF1A, and BRCA1 genes were included in MCODE cluster 1. WebGestalt tool analysis of these hub genes indicated that the most enrichment ratio includes the positive regulation of transcription by RNA polymerase II [Figure 7]. Inactivating mutations of p53, a tumor suppressor gene, occur in 50%–75% of PaC cases[8] while inhibition of STAT3 signaling in PaC was reported to inhibit growth and to induce apoptosis in tumor cells.[8],[51] Inhibition of the SMAD signaling pathway prevents TGFβ-mediated epithelial-to-mesenchymal transition (EMT) in PaC cells, which indicates that SMAD is a critical component of TGFβ signaling, with a potential role in transcriptional regulation of EMT genes. Normally, TGFβ is a strong inhibitor of proliferation of epithelial cell whereas, at advanced stages of cancer, it enhances cellular motility, migration, and metastasis of PaC cells (EMT).[8] These reports support our results, also showing that some of identified hub genes are involved in various processes of PaC, including proliferation, angiogenesis, invasion, and apoptosis with a clinical diagnostic value.

  Conclusion Top

We hypothesized that the occurrence of alteration in blood transcription profile of PaC and DM patients versus normal individuals could be considered a valid strategy for the identification of similar gene subsets and common molecular factors between PaC and DM. GEO2R analysis, conducted on microarray data from Huang et al., found two genes (SPI1 and YY1) to be expressed in PBMCs from PaC, DM, and PaC + DM patients. qRT-PCR results validated the microarray data and showed the significant increment of these two genes in all patient groups. Moreover, enrichment analyses on SPI1 and YY1 target genes revealed that target genes are mainly expressed in hematopoietic cells and are in association with immune responses as well as immune signaling pathways. However, the small number of samples for the experimental validation step – mainly due to the relatively low frequency of PaC – and the lack of experimental study on target genes are the main limitations of the current study. Consequently, additional validation – ideally on a prospective cohort study with matched age, gender, and clinical features – is needed for the next studies. In summary, we propose that SPI1 and YY1 and also their target genes could be useful to investigate the relationship between PaC and DM, providing opportunities for management of PaC screening using blood tests.


This study was conducted in the Cellular and Molecular Biology Research Center and supported by the Biotechnology Department, School of Advanced Technologies in Medicine, Shahid Beheshti University of Medical Sciences (Tehran, Iran) with grant number 8701. We are grateful to Dr. A. Mohamadkhani for sample collecting, Dr. M. Moradi Chalashtari and A. Kouchaki for qRT-PCR technique consulting, and Dr. S. Bahrami and Dr. H. Rezaee for their critical suggestions regarding improving the manuscript.

Ethical approval

All procedures performed in this study involving human participants were in accordance with the ethical standards of the local Ethics Committee of Shahid Beheshti University of Medical Science with ethical number IR.SBMU.RETECH.REC.1395.398 and with institutional and/or national research committee of the 2013 Helsinki Declaration.

Financial support and sponsorship

This paper was extracted from Sima Kalantari's PhD thesis and supported by the School of advanced technologies in Medicine , Shahid Beheshti University of Medical Sciences, Tehran, Iran, through grant NO 8701.

Conflicts of interest

The authors declare that none of the authors have any competing interests.

  References Top

Siegel R, Ward E, Brawley O, Jemal A. Cancer statistics, 2011: The impact of eliminating socioeconomic and racial disparities on premature cancer deaths. CA Cancer J Clin 2011;61:212-36.  Back to cited text no. 1
Rawla P, Sunkara T, Gaduputi V. Epidemiology of pancreatic cancer: Global trends, etiology and risk factors. World J Oncol 2019;10:10-27.  Back to cited text no. 2
Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011;61:69-90.  Back to cited text no. 3
De Wilde RF, Hruban RH, Maitra A, Offerhaus GJ. Reporting precursors to invasive pancreatic cancer: Pancreatic intraepithelial neoplasia, intraductal neoplasms and mucinous cystic neoplasm. Diagn Histopathol 2012;18:17-30.  Back to cited text no. 4
Yonezawa S, Higashi M, Yamada N, Goto M. Precursor lesions of pancreatic cancer. Gut Liver 2008;2:137-54.  Back to cited text no. 5
Ting DT, Wittner BS, Ligorio M, Vincent Jordan N, Shah AM, Miyamoto DT, et al. Single-cell RNA sequencing identifies extracellular matrix gene expression by pancreatic circulating tumor cells. Cell Rep 2014;8:1905-18.  Back to cited text no. 6
Zhang JJ, Zhu Y, Xie KL, Peng YP, Tao JQ, Tang J, et al. Yin Yang-1 suppresses invasion and metastasis of pancreatic ductal adenocarcinoma by downregulating MMP10 in a MUC4/ErbB2/p38/MEF2C-dependent mechanism. Mol Cancer 2014;13:130.  Back to cited text no. 7
Sankpal UT, Maliakal P, Bose D, Kayaleh O, Buchholz D, Basha R. Expression of specificity protein transcription factors in pancreatic cancer and their association in prognosis and therapy. Curr Med Chem 2012;19:3779-86.  Back to cited text no. 8
Wagner M, Redaelli C, Lietz M, Seiler CA, Friess H, Büchler MW. Curative resection is the single most important factor determining outcome in patients with pancreatic adenocarcinoma. Br J Surg 2004;91:586-94.  Back to cited text no. 9
Sheikh M, Masoudi S, Bakhshandeh R, Moayyedkazemi A, Zamani F, Nikfam S, et al. Survival features, prognostic factors, and determinants of diagnosis and treatment among Iranian patients with pancreatic cancer, a prospective study. PLoS One 2020;15:e0243511.  Back to cited text no. 10
Chari ST, Leibson CL, Rabe KG, Timmons LJ, Ransom J, de Andrade M, et al. Pancreatic cancer-associated diabetes mellitus: Prevalence and temporal association with diagnosis of cancer. Gastroenterology 2008;134:95-101.  Back to cited text no. 11
Menini S, Iacobini C, Vitale M, Pesce C, Pugliese G. Diabetes and pancreatic cancer – A dangerous liaison relying on carbonyl stress. Cancers (Basel) 2021;13:313.  Back to cited text no. 12
Chari ST, Leibson CL, Rabe KG, Ransom J, de Andrade M, Petersen GM. Probability of pancreatic cancer following diabetes: A population-based study. Gastroenterology 2005;129:504-11.  Back to cited text no. 13
Pannala R, Basu A, Petersen GM, Chari ST. New-onset diabetes: A potential clue to the early diagnosis of pancreatic cancer. Lancet Oncol 2009;10:88-95.  Back to cited text no. 14
Illés D, Ivány E, Holzinger G, Kosár K, Adam MG, Kamlage B, et al. New Onset of DiabetEs in aSsociation with pancreatic ductal adenocarcinoma (NODES Trial): Protocol of a prospective, multicentre observational trial. BMJ Open 2020;10:e037267.  Back to cited text no. 15
Cen P, Ni X, Yang J, Graham DY, Li M. Circulating tumor cells in the diagnosis and management of pancreatic cancer. Biochim Biophys Acta 2012;1826:350-6.  Back to cited text no. 16
Goggins M. Molecular markers of early pancreatic cancer. J Clin Oncol 2005;23:4524-31.  Back to cited text no. 17
Duffy MJ, Sturgeon C, Lamerz R, Haglund C, Holubec VL, Klapdor R, et al. Tumor markers in pancreatic cancer: A European Group on Tumor Markers (EGTM) status report. Ann Oncol 2010;21:441-7.  Back to cited text no. 18
Huang H, Dong X, Kang MX, Xu B, Chen Y, Zhang B, et al. Novel blood biomarkers of pancreatic cancer-associated diabetes mellitus identified by peripheral blood-based gene expression profiles. Am J Gastroenterol 2010;105:1661-9.  Back to cited text no. 19
American Diabetes Association. 2. Classification and Diagnosis of Diabetes: Standards of Medical Care in Diabetes-2020. Diabetes Care 2020;43:S14-31.  Back to cited text no. 20
Caba O, Prados J, Ortiz R, Jiménez-Luna C, Melguizo C, Alvarez PJ, et al. Transcriptional profiling of peripheral blood in pancreatic adenocarcinoma patients identifies diagnostic biomarkers. Dig Dis Sci 2014;59:2714-20.  Back to cited text no. 21
Steer HJ, Lake RA, Nowak AK, Robinson BW. Harnessing the immune response to treat cancer. Oncogene 2010;29:6301-13.  Back to cited text no. 22
Whitney AR, Diehn M, Popper SJ, Alizadeh AA, Boldrick JC, Relman DA, et al. Individuality and variation in gene expression patterns in human blood. Proc Natl Acad Sci U S A 2003;100:1896-901.  Back to cited text no. 23
Liotta LA, Ferrari M, Petricoin E. Clinical proteomics: Written in blood. Nature 2003;425:905.  Back to cited text no. 24
Twine NC, Stover JA, Marshall B, Dukart G, Hidalgo M, Stadler W, et al. Disease-associated expression profiles in peripheral blood mononuclear cells from patients with advanced renal cell carcinoma. Cancer Res 2003;63:6069-75.  Back to cited text no. 25
Lee T, Teng TZJ, Shelat VG. Carbohydrate antigen 19-9-tumor marker: Past, present, and future. World J Gastrointest Surg 2020;12:468-90.  Back to cited text no. 26
Baine MJ, Chakraborty S, Smith LM, Mallya K, Sasson AR, Brand RE, et al. Transcriptional profiling of peripheral blood mononuclear cells in pancreatic cancer patients identifies novel genes with potential diagnostic utility. PLoS One 2011;6:e17014.  Back to cited text no. 27
Iacobini C, Vitale M, Pesce C, Pugliese G, Menini S . Diabetic Complications and Oxidative Stress: A 20-Year Voyage Back in Time and Back to the Future. Antioxidants 2021; 10: 727.  Back to cited text no. 28
Blot WJ, Fraumeni JF Jr., Stone BJ. Geographic correlates of pancreas cancer in the United States. Cancer 1978;42:373-80.  Back to cited text no. 29
Cuzick J, Babiker AG. Pancreatic cancer, alcohol, diabetes mellitus and gall-bladder disease. Int J Cancer 1989;43:415-21.  Back to cited text no. 30
La Vecchia C, Negri E, Franceschi S, D'Avanzo B, Boyle P. A case-control study of diabetes mellitus and cancer risk. Br J Cancer 1994;70:950-3.  Back to cited text no. 31
Gullo L, Pezzilli R, Morselli-Labate AM, Italian Pancreatic Cancer Study Group. Diabetes and the risk of pancreatic cancer. N Engl J Med 1994;331:81-4.  Back to cited text no. 32
Lee CT, Chang FY, Lee SD. Risk factors for pancreatic cancer in orientals. J Gastroenterol Hepatol 1996;11:491-5.  Back to cited text no. 33
Calle EE, Murphy TK, Rodriguez C, Thun MJ, Heath CW. Diabetes mellitus and pancreaticcancer mortality in a prospectivecohort of United States adults. Cancer Causes Control 1998;9:403-10.  Back to cited text no. 34
Silverman DT, Schiffman M, Everhart J, Goldstein A, Lillemoe KD, Swanson GM, et al. Diabetes mellitus, other medical conditions and familial history of cancer as risk factors for pancreatic cancer. Br J Cancer 1999;80:1830-7.  Back to cited text no. 35
Gupta S, Vittinghoff E, Bertenthal D, Corley D, Shen H, Walter LC, et al. New-onset diabetes and pancreatic cancer. Clin Gastroenterol Hepatol 2006;4:1366-72.  Back to cited text no. 36
Pannala R, Leirness JB, Bamlet WR, Basu A, Petersen GM, Chari ST. Prevalence and clinical profile of pancreatic cancer-associated diabetes mellitus. Gastroenterology 2008;134:981-7.  Back to cited text no. 37
Li D, Tang H, Hassan MM, Holly EA, Bracci PM, Silverman DT. Diabetes and risk of pancreatic cancer: A pooled analysis of three large case-control studies. Cancer Causes Control 2011;22:189-97.  Back to cited text no. 38
Ben Q, Xu M, Ning X, Liu J, Hong S, Huang W, et al. Diabetes mellitus and risk of pancreatic cancer: A meta-analysis of cohort studies. Eur J Cancer 2011;47:1928-37.  Back to cited text no. 39
Özdemir İ, Pınarlı FG, Pınarlı FA, Aksakal FNB, Okur A, Uyar Göçün P, et al. Epigenetic silencing of the tumor suppressor genes SPI1, PRDX2, KLF4, DLEC1, and DAPK1 in childhood and adolescent lymphomas. Pediatr Hematol Oncol 2018;35:131-44.  Back to cited text no. 40
Delestré L, Cui H, Esposito M, Quiveron C, Mylonas E, Penard-Lacronique V, et al. Senescence is a Spi1-induced anti-proliferative mechanism in primary hematopoietic cells. Haematologica 2017;102:1850-60.  Back to cited text no. 41
Gullo L. Chronic pancreatitis in Italy. Dig Liver Dis 2010;42:156.  Back to cited text no. 42
Xu Y, Gu S, Bi Y, Qi X, Yan Y, Lou M. Transcription factor PU.1 is involved in the progression of glioma. Oncol Lett 2018;15:3753-9.  Back to cited text no. 43
Chen Q, Yang C, Chen L, Zhang JJ, Ge WL, Yuan H, et al. YY1 targets tubulin polymerisation-promoting protein to inhibit migration, invasion and angiogenesis in pancreatic cancer via p38/MAPK and PI3K/AKT pathways. Br J Cancer 2019;121:912-21.  Back to cited text no. 44
Klöting N, Klöting I. Genetic variation in the multifunctional transcription factor Yy1 and type 1 diabetes mellitus in the BB rat. Mol Genet Metab 2004;82:255-9.  Back to cited text no. 45
Lu Y, Xiong X, Wang X, Zhang Z, Li J, Shi G, et al. Yin Yang 1 promotes hepatic gluconeogenesis through upregulation of glucocorticoid receptor. Diabetes 2013;62:1064-73.  Back to cited text no. 46
Hansen CN, Ketabi Z, Rosenstierne MW, Palle C, Boesen HC, Norrild B. Expression of CPEB, GAPDH and U6snRNA in cervical and ovarian tissue during cancer development. APMIS 2009;117:53-9.  Back to cited text no. 47
Giusti L, Iacconi P, Ciregia F, Giannaccini G, Donatini GL, Basolo F, et al. Fine-needle aspiration of thyroid nodules: Proteomic analysis to identify cancer biomarkers. J Proteome Res 2008;7:4079-88.  Back to cited text no. 48
Arlt A, Müerköster SS, Schäfer H. Targeting apoptosis pathways in pancreatic cancer. Cancer Lett 2013;332:346-58.  Back to cited text no. 49
Trauzold A, Röder C, Sipos B, Karsten K, Arlt A, Jiang P, et al. CD95 and TRAF2 promote invasiveness of pancreatic cancer cells. FASEB J 2005;19:620-2.  Back to cited text no. 50
Sahu RP, Srivastava SK. The role of STAT-3 in the induction of apoptosis in pancreatic cancer cells by benzyl isothiocyanate. J Natl Cancer Inst 2009;101:176-93.  Back to cited text no. 51


  [Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]

  [Table 1], [Table 2], [Table 3], [Table 4]


Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

  In this article
Article Figures
Article Tables

 Article Access Statistics
    PDF Downloaded44    
    Comments [Add]    

Recommend this journal