Integrative analyses of molecular pathways and key candidate biomarkers associated with colorectal cancer

1. Introduction

Colorectal cancer (CRC) is one of the most common devastating cancer that leads to high mortality and morbidity in both women and men worldwide [1, 2]. However, the incidence of CRC was found to continue rising in developed countries especially in Europe along with the decline in mortality rates [3]. CRC is a complex heterogeneous disease and genetic, epigenetic and environmental risk factors such as diet, body fat distribution, and microbiome composition are known to contribute to the disease pathogenesis [4, 5, 6]. Continuing efforts to understand the genetic epidemiology of CRC have been successfully revealed several loci associated with the disease risk, yet early molecular diagnosis and accurate assessment of prognosis of CRC are still challenging [7, 8, 9]. Thus, in recent years, researchers have begun to investigate the missing heritability underlying the etiology of CRC by exploring the possible contributions of molecular regulators of the gene expression such as microRNAs (miRNAs) and transcription factors (TFs).

OPEN IN VIEWER

Dysregulation of the genes involved in the critical pathways in colorectal carcinogenesis is known to be associated with CRC. Hence, it is highly imperative to unveil the key mediators and their molecular function influencing the expressions of the genes dysregulated in CRC. Recently, miRNAs were proposed to have functional importance in carcinogenesis and also accumulating evidence indicates their utilization as a diagnostic and prognostic biomarker for CRC [10, 11, 12]. Also, dysregulated expression of some TFs was shown to be a major impact on colorectal tumor growth, CRC-associated oncogenic process as well as CRC metastasis [13, 14, 15, 16, 17]. Nowadays, researchers shed light on the interactions of the miRNAs, TFs and their target genes and suggested to consider them as a network to fully uncover their roles in CRC pathogenesis [18, 19].

Our study was designed as an “in silico” approach to generating novel knowledge related the missing genetic determinants and key regulators of the CRC as well as pinpoint candidate critical molecular pathways which may be targeted for therapeutic approaches and utilized for CRC management.

3. Results

The etiology of colorectal cancer is still not fully understood for efficiently predicting disease development and designing tailored treatment options. Genetic risk factors are known to be associated with disease pathogenesis, prognosis and survival however there is a lack of knowledge regarding the CRC genetics and this needs to be comprehensively elucidated.

In this work, we used the transcriptomic data of TCGA-COAD samples and GTEx normal samples in an attempt to comprehensively investigate the key molecular pathways and mediators related CRC.

We initially identified the 1556 DEGs and constructed the PPI network by STRING to compute their topological properties and define the hub genes in the network. Ten overexpressed genes (AURKA, CDK1, MYC, CDH1, CCNB1, CDC20, UBE2C, PLK1, KIF11, CCNA2) were prioritized as hub (essential) genes based on the degree centrality in Cytohubba analysis. The degree centrality measure for a protein suggests its interaction level with other proteins in the network and the largest degree pinpoints the most connected and essential protein in the network [28]. Thus, these hub genes are likely to be facilitated in therapeutic approaches for CRC.

Functional enrichment of the DEGs in KEGG pathways and GO biological process highlight their role in the cell cycle and cell division. The IPA results were in concordance with these results and indicate the significant associations of hub genes with cell cycle progression. Moreover, the 5 of the hub genes including the 3 important genes (CDK1, CCNB1, PLK1) of kinase gene family were found to be involved in “Mitotic roles of Polo-like kinase” pathway (Fig. 2). Mitotic kinases are known to have critical functions for proper cell division. Specifically, PLK1 was shown to be a key regulator of mitotic cell division and have tumor suppressor properties by influencing the tumor formation [29, 30]. In recent years, upregulation of PLK1 was reported to be associated with many cancers including colon cancer, thus therapeutic approaches using anti-PLK1 have been exciting research topics for cancer researchers [31, 32]. Furthermore, CDK1 and CCNB1 were also primary cell cycle regulators and overexpressions of these genes were previously shown to contribute to the CRC pathogenesis [33, 34, 35].

In addition to examining the pathways associated with CRC, we mined the key transcriptional and post-transcriptional regulators of the CRC development using in silico tools. MicroRNAs are key mediators of the gene regulation and miRNAs targeting the genes involving in the fundamental pathways were highly important for unraveling the miRNA-mediated post-transcriptional mechanisms in CRC pathogenesis.

Recent research also sheds light on the miRNA-TF-target gene interactions and TFs tuning both miRNAs and their target genes may be considered as key regulators of the particular network and promising target for translational research applications [10, 18, 19, 36]. Considering the TFs as crucial mediators of the regulation of the miRNA expressions, uncovering the TF-miRNA interactions may have a considerable impact on developing novel drug targets and biomarkers. Our analyses revealed 3 key TFs (TP53, EZH2, KLF4) that were predicted to affect the regulations of hub genes and also the miRNAs targeting these genes (Fig. 3). The abnormal function of the tumor suppressor TP53 has been associated with various types of cancers including colorectal cancer and cause poor prognosis [37]. The role of EZH2 in CRC was also shown in several studies and an increased level of EZH2 was suggested to affect CRC development [38, 39]. Our result highlighting the imperative role of KLF4 is consistent with the literature which accumulating research indicates KLF4 as a tumor suppressor gene for colorectal cancer and overexpression of KLF4 was suggested to be a used in novel therapeutic and prognostic applications to alleviate the burden of CRC [40, 41]. Of note, all of the three TFs were predicted to be the common regulators affect 4 miRNA (miR-200b, miR-200c, miR-145, miR-143) (Table 6) which were previously reported to be dysregulated in CRC [12, 42, 43, 44, 45].

The results of our study depend solely on in silico analyses thus creates some limitations. Identified hub genes and key regulators of the gene regulations need to be further confirmed by biological experiments to evaluate their roles in the utilization of novel translational applications. However, our results provide additional insight for the interplay between TFs-miRNAs that were likely to involve in CRC development. It is also worth mentioning that our findings are eminently in accord with previous experimental research which strengths the accuracy of our analyses and research methodology.

Our study suggests key genes in dysregulated pathways implicated in CRC and shows in silico evidence for candidate key components of the transcriptional and post-transcriptional regulatory network in CRC. The results of this study can further enhance the molecular etiology of the CRC as well as can be employed in future research for biomarker and drug development related CRC.