Identifying pancreatic cancer‑associated miRNAs using weighted gene co‑expression network analysis

  • Authors:
    • Pengfei Lyu
    • Zhengwen Hao
    • Haoruo Zhang
    • Jun Li
  • View Affiliations

  • Published online on: July 5, 2022     https://doi.org/10.3892/ol.2022.13417
  • Article Number: 297
Metrics: Total Views: 0 (Spandidos Publications: | PMC Statistics: )
Total PDF Downloads: 0 (Spandidos Publications: | PMC Statistics: )


Abstract

Pancreatic cancer is a common type of gastrointestinal tumour throughout the world and is characterised by high malignancy rates and poor prognosis. Studies indicated that early and effective diagnosis is key to prolonging patients' overall survival, particularly in the case of fluid biopsy. Given this, the present study was designed to evaluate the expression profile arrays of patients with pancreatic cancer from the Gene Expression Omnibus database in an effort to identify differentially expressed microRNAs (miRNAs/miRs) that may be suitable for application in liquid biopsy‑based diagnostics. Suitable miRNA candidates were identified using a weighted correlation network analysis (WGCNA) and key differentially expressed miRNAs were verified using reverse transcription‑quantitative PCR. WGCNA identified 11 differentially expressed miRNAs (miR‑155‑5p, miR‑4668‑5p, miR‑3613‑3p, miR‑3201, miR‑548ac, miR‑486‑5p, miR‑548a‑3p, miR‑8084, miR‑455‑3p, miR‑6068 and miR‑1246). Of these, miR‑4668‑5p was indicated to have the highest number of associated modules, making it most likely to be of diagnostic value. Thus, the present analysis identified 11 miRNAs associated with pancreatic cancer and further identified miR‑4668‑5p as a potential biomarker for pancreatic cancer diagnosis.

Introduction

An aging population means that cancers continue to grow as an overall health concern for numerous nations, including China. Pancreatic cancer is rapidly becoming one of the most common malignant tumour types in China and patients usually have poor prognosis (1). The most profound issues surrounding the prognosis of these patients include the lack of any early diagnostic methods and effective treatments for this condition (2).

This means there is a global focus on improving patient survival while improving overall quality of life, with the majority of studies clearly prioritising diagnosis, therapy and risk stratification as areas of importance. This means that the number of novel molecular biomarkers associated with pancreatic cancer continues to increase with the family of biomarkers expanding to include specific mutations and abnormal gene expression (3). DNA methylation (4), cell-free DNA (5) and exosomes (6) have also all been identified as potential research hotspots; however, despite this growth, there is still only a small number of markers that may be used in the clinical setting to support pancreatic cancer management. This means that the current clinical diagnosis and treatment continues to rely on imaging and tumour biomarker assays for identification. However, these examinations fail to achieve early diagnosis of pancreatic cancer, making the discovery of novel biomarkers with high sensitivity and specificity (7) a top priority for the field as a whole. This means that it is important to explore all novel biomarkers associated with the occurrence and development of pancreatic cancer. This has been aided by the continuous development of bioinformatics and the establishment of several public databases, which enable an increasing number of researchers to use bioinformatics analysis to evaluate the molecular genetic data from this disease and identify potential molecular markers for clinical application.

One example for this is the application of weighted gene co-expression network analysis (WGCNA) across several cancer types to identify potential molecular drivers in these conditions (7). In the present study, the expression profile array data and WGCNA were combined to identify potential molecular biomarkers for pancreatic cancer in an effort to provide novel diagnostic targets. Furthermore, the differentially expressed microRNAs (miRNAs/miRs) in clinical samples were experimentally verified, allowing for an added layer of validation, unique from similar gene expression omnibus (GEO)-based evaluations (7).

Materials and methods

Data retrieval and analysis

The miRNA expression profile array dataset GSE85589 was downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/gds/) and serum miRNA expression data were obtained for 88 patients with pancreatic cancer and 19 healthy controls. Differentially expressed miRNAs were screened using R software and differential expression was considered significant when the absolute value of log2FoldChange and false discovery rate (FDR) reached ≥1 and ≤0.05, respectively.

WGCNA

Serum miRNA expression data from 88 patients with pancreatic cancer were obtained from GSE85589 and WGCNA analysis was performed as previously described (8).

cDNA synthesis and quantitative PCR (qPCR)

A total of 2 ml serum was obtained from each of the 10 healthy controls (6 males, 4 females) and 16 patients with pancreatic ductal adenocarcinoma (PDAC) (8 males, 8 females) undergoing treatment at the Shanxi Tumor Hospital (Taiyuan, China) between January 2018 and January 2020. The serum samples of healthy control subjects without tumors were collected from the Health Examination Center, Shanxi Tumor Hospital (Taiyuan, China). Serum samples of patients were collected from patients newly diagnosed with PDAC and with no history of other malignancies. The healthy controls and patients with pancreatic cancer were matched in terms of sex and age. Total RNA was extracted according to a previous report (9). RNA was then reverse-transcribed using a First-Strand cDNA synthesis kit (Agilent Technologies, Inc.) according to the manufacturer's protocol. This cDNA was then amplified by qPCR using SYBR® Green PCR master mix (Applied Biosystems; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. The experiment was performed in an ABI 7500 PCR machine (Applied Biosystems; Thermo Fisher Scientific, Inc.) and the thermocycling conditions as follows: 95°C for 10 min, and 40 cycles of 95°C for 15 sec and 60°C for 1 min. U6 was used as an internal control. The relative miRNA expression levels were then determined by comparing the fold change between patients with pancreatic cancer and healthy controls using the 2−ΔΔCq method (10). These experiments were performed in triplicate and the primer sequences were as follows: miR-4668-5p forward, 5′-TCGGCAGGAGGGAAAAAAAAAA-3′ and reverse, 5′-CTCAACTGGTGTCGTGGA-3′; U6 forward, 5′-CTCGCTTCGGCAGCACAT-3′ and reverse, 5′-AACGCTTCACGAATTTGCGT-3′.

Statistical analysis

SPSS software (version 21.0; IBM Corporation) was used in each of the statistical evaluations and miR-4668-5p expression was compared between patients with pancreatic cancer and healthy individuals using an unpaired t-test. Differentially expressed miRNAs were screened using R software via the ‘impute’ and ‘limma’ packages. P<0.05 was considered to indicate a statistically significant difference.

Results

Screening identified 11 dysregulated miRNAs in patients with pancreatic cancer

Serum miRNA expression levels in patients with pancreatic cancer (n=88) were compared with those of a healthy control cohort (n=19) in the GEO dataset (Fig. 1A). The comparisons identified 11 clearly dysregulated miRNAs in these samples, including miR-155-5p, miR-4668-5p, miR-3613-3p, miR-3201, miR-548ac, miR-486-5p, miR-548a-3p, miR-8084, miR-455-3p, miR-6068 and miR-1246. Of these, miR-4668-5p, miR-3613-3p, miR-3201, miR-548ac, miR-486-5p, miR-548a-3p, miR-8084, miR-6068 and miR-1246 were upregulated in the serum of patients with pancreatic cancer, whereas miR-155-5p and miR-455-3p were downregulated (FDR ≤0.05; Table I).

Table I.

Differentially expressed miRNAs from the gene expression omnibus database.

Table I.

Differentially expressed miRNAs from the gene expression omnibus database.

miRNA |Log2FC|Average expressiontP-valueAdjusted P-value
miR-155-5pa1.080.858.86 1.55×10−14 4.00×10−11
miR-4668-5pb3.466.48−8.15 6.05×10−13 7.80×10−10
miR-3613-3pb3.636.58−8.07 9.27×10−13 7.96×10−10
miR-3201b2.302.59−7.78 4.14×10−12 2.67×10−9
miR-548acb1.101.91−6.94 2.76×10−10 1.19×10−7
miR-486-5pb1.046.62−6.39 4.08×10−9 1.17×10−6
miR-548a-3pb1.022.09−6.21 9.44×10−9 2.18×10−6
miR-8084b1.592.14−5.95 3.19×10−8 6.32×10−6
miR-455-3pa1.104.715.31 5.67×10−7 7.39×10−5
miR-6068b1.003.02−4.78 5.44×10−6 4.38×10−4
miR-1246b1.004.863.70 3.39×10−4 1.07×10−2

a Downregulation;

b upregulation. FC, fold change; miRNA/miR, microRNA.

miR-4668-5p is associated with a multitude of modules during WGCNA

The WGCNA package provides R functions for weighted correlation network analysis, which allowed the investigation of the co-expression patterns of the various miRNAs identified in the 88 patient samples. These evaluations identified 11 clusters (modules) of highly correlated genes, which were highlighted as follows: Dark orange, salmon, dark magenta, dark grey, light green, dark green, grey 60, cyan, midnight blue, green yellow and light cyan. A heat map of these miRNAs was also generated (Fig. 1B-D) and most of the previously identified miRNAs were grouped into the dark grey, dark magenta and light green modules (Table II). The module membership value for each of these miRNAs was then calculated and they were used to identify the most significantly linked modules in this data set (P≤0.05; Table II). Of note, these evaluations identified miR-4668-5p as the transcript with the highest number of associations, with this transcript linked to various modules including the dark magenta, dark grey, light green, grey 60 and green yellow modules. Therefore, it was suggested that miR-4668-5p may serve as a potential molecular biomarker for pancreatic cancer.

Table II.

Models and MM value of these differentially expressed miRNAs.

Table II.

Models and MM value of these differentially expressed miRNAs.

Model/miRNAs MM P<0.05
Darkgrey
  hsa-miR-4668-5pDarkmagentaDarkgreyLightgreenGrey60Greenyellow
  hsa-miR-3613-3pDarkmagentaDarkgreyLightgreenGreenyellow
  hsa-miR-6068Greenyellowgrey60LightgreenDarkgrey
Darkmagenta
  hsa-miR-548acDarkmagentaDarkgreyCyan
  hsa-miR-3201DarkmagentaCyan
  hsa-miR-548a-3pDarkmagentaDarkgrey
  hsa-miR-8084Darkmagenta
Lightgreen
  hsa-miR-155-5pLightgreen

[i] miRNA/miR, microRNA; MM, module membership.

miR-4668-5p is consistently upregulated in serum of patients with pancreatic cancer

The GEO and WGCNA analysis of the present study suggested that miR-4668-5p may be a key marker for pancreatic cancer. And evaluations of GSE85589 revealed a significant upregulation of miR-4668-5p in patients with pancreatic cancer (n=88) when compared to the healthy controls (n=19) (P≤0.0001; Fig. 2A). Given this, the study went on to validate the expression levels of miR-4668-5p using an RT-qPCR assay. The results confirmed that miR-4668-5p expression was consistently upregulated in the serum of patients with pancreatic cancer (n=16) when compared with that of healthy controls (n=10) (P≤0.05; Fig. 2B).

Discussion

The present study was designed to identify a subset of co-expressed miRNAs that may help to differentiate between pancreatic cancer and healthy tissues when applied in a liquid biopsy setting. The study relied on the WGCNA-based evaluation of GEO data to identify these differentially expressed transcripts and facilitate their further evaluation. These analyses identified 11 differentially expressed miRNAs in patients with pancreatic cancer and revealed that miR-4668-5p was likely a central regulator/effector, as it presented with the most associations and modular links to other miRNAs. This suggests that miR-4668-5p may be a potential molecular biomarker for pancreatic cancer. Given this, its upregulation was confirmed in patients with pancreatic cancer using RT-qPCR, which validated the initial identification supporting the notion that this transcript may be a valuable marker for diagnosis of pancreatic cancer.

It has been indicated that patients with incidentally diagnosed pancreatic cancer have a better prognosis than those who are diagnosed upon developing symptoms (11), which explains the clear benefits of early diagnosis in improving prognosis. Although there are numerous novel diagnostic methods, liquid biopsy continues to be a firm favourite, as this method is largely non-invasive and is designed to be convenient, economical and minimally traumatic (12). However, liquid biopsy relies on the identification of disease-specific biomarkers. To date, no clinically useful biomarkers have been reported for pancreatic cancer. Given this, in the present study the non-coding RNA profile array for patients with pancreatic cancer was downloaded from the GEO database and a weighted screening of the various differentially expressed miRNAs was performed in this dataset. The present evaluation identified 11 differentially expressed miRNAs in these samples, which were then further narrowed down to the key transcripts using WGCNA. Of note, the present evaluations included the construction of a co-expression network and modules, designed to increase the current understanding of the host-pathogen relationship. WGCNA has gradually been used to construct various co-expression networks (13), particularly for miRNAs. One example of this is the identification of several hub miRNAs associated with prognosis in colorectal cancer (14,15). Here, WGCNA was used to construct various modules and co-expression networks with the aim of identifying particularly well-connected transcripts, as a proxy for their importance in this pathology. These evaluations reduced the target miRNAs from 11 to just one transcript miR-4668-5p. This suggests that this transcript is likely key to this pathology and thus is a solid target for liquid biopsy development in the future.

Dysregulated miR-4668-5p expression has been observed in the peripheral blood and tissues of numerous diseases, including hepatocellular carcinoma, where it is indicated that this miRNA may have a role in cancer progression, vascular invasion and immune surveillance, as well as the potential coordination of other upregulated miRNAs (16). miR-4668-5p has also been observed to be dysregulated in gastric cancer (17) and dedifferentiated liposarcoma tissues (18) when compared to healthy controls. Although abnormal miR-4668-5p expression has been observed in several malignant tumour types, the molecular mechanism involved in the occurrence and progression of these tumours remains elusive. Relevant research is currently limited and more work is required in the future. From a clinical perspective, the relationship between miRNA expression, clinical stage and survival time of patients with pancreatic cancer will be investigated in a future study. A study of the effects of miRNA overexpression on the biological behaviour of malignant tumour cells in vivo and in vitro may also be performed with the aim to uncover the target genes and regulatory pathways impacted by this transcript with a focus on understanding its role in tumour progression. It is also worth noting that miR-4668-5p is also dysregulated in various diseases of the nervous system, including Alzheimer's disease (10) and mesial temporal lobe epilepsy with hippocampal sclerosis (19). However, studies evaluating miR-4668-5p expression in pancreatic cancer remain limited.

Despite its clear contribution, it is important to note that the present study does have certain limitations, including the fact that it is better to combine miRNA expression and clinical data, such as those on overall survival, clinical stage, etc., when evaluating the diagnostic/prognostic value. In addition, the potential mechanism underlying miR-4668-5p-mediated regulation of its target genes and their functions remain undefined and should be prioritised moving forward. Finally, these observations require to be confirmed in larger disease cohorts to determine the diagnostic capacity.

In conclusion, the present study was the first to indicate that miR-4668-5p is a differentially expressed miRNA consistently upregulated in the serum of patients with pancreatic cancer and the first to suggest that this transcript may be a potential diagnostic marker for pancreatic cancer. The present data may help facilitate the development of novel diagnostics for this condition in the future.

Acknowledgements

Not applicable.

Funding

Funding: No funding was received.

Availability of data and materials

The public datasets analysed are available from the sources stated above. Apart from that, data sharing is not applicable to this article, as no datasets were generated during the current study.

Authors' contributions

PL performed the primary bioinformatics analysis and experiments and wrote the manuscript; ZH and HZ made substantial contributions to data analysis and technical support. JL made substantial contributions to conception and design, and revising the manuscript critically for important intellectual content, and given final approval of the version to be published, and agreed to be accountable for all aspects of the work. PL and JL checked and confirmed the authenticity of the raw data. All authors contributed to the article and read and approved the final version of the manuscript.

Ethics approval and consent to participate

The study was approved by the ethics committee of the Shanxi Tumour Hospital (Taiyuan, China; no. 2022JC19) and the patients provided written informed consent.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References

1 

Jia X, Du P, Wu K, Xu Z, Fang J, Xu X and Lin K: Pancreatic cancer mortality in China: Characteristics and prediction. Pancreas. 47:233–237. 2018. View Article : Google Scholar : PubMed/NCBI

2 

McGuigan A, Kelly P, Turkington RC, Jones C, Coleman HG and McCain RS: Pancreatic cancer: A review of clinical diagnosis, epidemiology, treatment and outcomes. World J Gastroenterol. 24:4846–4861. 2018. View Article : Google Scholar : PubMed/NCBI

3 

Previdi MC, Carotenuto P, Zito D, Pandolfo R and Braconi C: Noncoding RNAs as novel biomarkers in pancreatic cancer: What do we know? Future Oncol. 13:443–453. 2017. View Article : Google Scholar : PubMed/NCBI

4 

Mishra NK and Guda C: Genome-wide DNA methylation analysis reveals molecular subtypes of pancreatic cancer. Oncotarget. 8:28990–29012. 2017. View Article : Google Scholar : PubMed/NCBI

5 

Vietsch EE, Graham GT, McCutcheon JN, Javaid A, Giaccone G, Marshall JL and Wellstein A: Circulating cell-free DNA mutation patterns in early and late stage colon and pancreatic cancer. Cancer Genet. 218–219. 39–50. 2017.PubMed/NCBI

6 

Zhao C, Gao F, Weng S and Liu Q: Pancreatic cancer and associated exosomes. Cancer Biomark. 20:357–367. 2017. View Article : Google Scholar : PubMed/NCBI

7 

Chu LC, Goggins MG and Fishman EK: Diagnosis and detection of pancreatic cancer. Cancer J. 23:333–342. 2017. View Article : Google Scholar : PubMed/NCBI

8 

Langfelder P and Horvath S: WGCNA: An R package for weighted correlation network analysis. BMC Bioinformatics. 9:5592008. View Article : Google Scholar : PubMed/NCBI

9 

Ruan QF, Jiang MJ, Ye ZQ, Zhao CL and Xie WG: Analysis of microRNA expression profile in serum of patients with electrical burn or thermal burn. Zhonghua Shao Shang Za Zhi. 33:37–42. 2017.(In Chinese). PubMed/NCBI

10 

Schmittgen TD and Livak KJ: Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc. 3:1101–1108. 2008. View Article : Google Scholar : PubMed/NCBI

11 

Zhou B, Xu JW, Cheng YG, Gao JY, Hu SY, Wang L and Zhan HX: Early detection of pancreatic cancer: Where are we now and where are we going? Int J Cancer. 141:231–241. 2017. View Article : Google Scholar : PubMed/NCBI

12 

Chu D and Park BH: Liquid biopsy: Unlocking the potentials of cell-free DNA. Virchows Arch. 471:147–154. 2017. View Article : Google Scholar : PubMed/NCBI

13 

Zhai X, Xue Q, Liu Q, Guo Y and Chen Z: Colon cancer recurrence-associated genes revealed by WGCNA co-expression network analysis. Mol Med Rep. 16:6499–6505. 2017. View Article : Google Scholar : PubMed/NCBI

14 

Yepes S, López R, Andrade RE, Rodriguez-Urrego PA, Lopez-Kleine L and Torres MM: Co-expressed miRNAs in gastric adenocarcinoma. Genomics. 108:93–101. 2016. View Article : Google Scholar : PubMed/NCBI

15 

Zhou XG, Huang XL, Liang SY, Tang SM, Wu SK, Huang TT, Mo ZN and Wang QY: Identifying miRNA and gene modules of colon cancer associated with pathological stage by weighted gene co-expression network analysis. Onco Targets Ther. 11:2815–2830. 2018. View Article : Google Scholar : PubMed/NCBI

16 

Pascut D, Krmac H, Gilardi F, Patti R, Calligaris R, Croce LS and Tiribelli C: A comparative characterization of the circulating miRNome in whole blood and serum of HCC patients. Sci Rep. 9:82652019. View Article : Google Scholar : PubMed/NCBI

17 

Bibi F, Naseer MI, Alvi SA, Yasir M, Jiman-Fatani AA, Sawan A, Abuzenadah AM, Al-Qahtani MH and Azhar EI: microRNA analysis of gastric cancer patients from Saudi Arabian population. BMC Genomics. 17 (Suppl 9):S7512016. View Article : Google Scholar : PubMed/NCBI

18 

Fricke A, Cimniak AFV, Ullrich PV, Becherer C, Bickert C, Pfeifer D, Heinz J, Stark GB, Bannasch H, Braig D and Eisenhardt SU: Whole blood miRNA expression analysis reveals miR-3613-3p as a potential biomarker for dedifferentiated liposarcoma. Cancer Biomark. 22:199–207. 2018. View Article : Google Scholar : PubMed/NCBI

19 

Yan S, Zhang H, Xie W, Meng F, Zhang K, Jiang Y, Zhang X and Zhang J: Altered microRNA profiles in plasma exosomes from mesial temporal lobe epilepsy with hippocampal sclerosis. Oncotarget. 8:4136–4146. 2017. View Article : Google Scholar : PubMed/NCBI

Related Articles

Journal Cover

September-2022
Volume 24 Issue 3

Print ISSN: 1792-1074
Online ISSN:1792-1082

Sign up for eToc alerts

Recommend to Library

Copy and paste a formatted citation
x
Spandidos Publications style
Lyu P, Hao Z, Zhang H and Li J: Identifying pancreatic cancer‑associated miRNAs using weighted gene co‑expression network analysis. Oncol Lett 24: 297, 2022
APA
Lyu, P., Hao, Z., Zhang, H., & Li, J. (2022). Identifying pancreatic cancer‑associated miRNAs using weighted gene co‑expression network analysis. Oncology Letters, 24, 297. https://doi.org/10.3892/ol.2022.13417
MLA
Lyu, P., Hao, Z., Zhang, H., Li, J."Identifying pancreatic cancer‑associated miRNAs using weighted gene co‑expression network analysis". Oncology Letters 24.3 (2022): 297.
Chicago
Lyu, P., Hao, Z., Zhang, H., Li, J."Identifying pancreatic cancer‑associated miRNAs using weighted gene co‑expression network analysis". Oncology Letters 24, no. 3 (2022): 297. https://doi.org/10.3892/ol.2022.13417