Comprehensive analysis of ceRNA network of ERCC4 in colorectal cancer

Introduction

DNA damage caused by endogenous or exogenous genetic toxicants which contribute to genomic instability may lead to a variety of cancers (Iyama & Wilson, 2013). DNA repair systems play a critical role in maintaining the integrity and stability of the genome, of which nucleotide excision repair (NER) can identify and amend multiple types of potential DNA damage. The NER process consists of four main steps namely damage identification, damage partitioning and unwinding, damage incision and new strand synthesis (de Laat, Jaspers & Hoeijmakers, 1999; Liu et al., 2014; Slyskova et al., 2012). Different key proteins are involved in different NER steps. ERCC4 (XPF) is one of the indispensable molecules of NER, which can determine the activity of NER and is responsible for the removal of UV-C photoproducts and large volume adducts from DNA (Enzlin & Scharer, 2002). Besides, the heterodimer of ERCC4-ERCC1 is involved with the 5′ incision step of the NER pathway (Enzlin & Scharer, 2002). ERCC4-ERCC1 can repair damaged DNA in both replicating and non-replicating cells as a structure-specific endonuclease (Evans et al., 1997; Mu, Hsu & Sancar, 1996; Sijbers et al., 1996). Therefore, the abnormal expression of ERCC4 could seriously affect the NER pathway.

Many researches have mentioned the role of ERCC4 in cancers, such as bladder cancer, gastric cancer and oral cancer (Li & Ma, 2018; Qiu et al., 2014; Sa et al., 2020; Wang et al., 2010; Wei et al., 2014). Colorectal cancer (CRC) is a malignant tumor which is the third cause of cancer death in China (Zhang et al., 2021). Previous studies also focused on the relationship between ERCC4 polymorphisms and the risk of CRC (He et al., 2014; Kabzinski et al., 2015; Yang, Li & Li, 2015). In recent studies, the post-transcriptional regulation of non-coding RNA plays an important role in the development of cancer (Anastasiadou, Jacob & Slack, 2018; Wei et al., 2017). Salmena et al. (2011) put forward the hypothesis of ceRNA, which expands our insight into the pathogenesis of CRC (Wei et al., 2017; Yuan et al., 2019). For example, Zhu et al. (2019) suggested that LINC00365 is an oncogene in CRC, which can regulate the expression of several mRNAs by sponging miRNAs. H19 was found to regulate PI3K-Akt signaling pathway through ceRNA network, and predict the poor prognosis of CRC (Zhong et al., 2019). Since ERCC4 plays a significant role in CRC tumorigenesis, it is necessary to explore the related ceRNA regulatory network to better understand its mechanism in the development and progression of CRC. Although there have been studies investigating the relationship between ERCC4 expression and the risk of CRC based on small sample size (Zhang et al., 2019), the ceRNA network regulation of ERCC4 in CRC has not been identified.

In this study, we screened DElncRNAs and DEmiRNAs in two groups of ERCC4^high and ERCC4^low expression in CRC. Then a lncRNA-miRNA-ERCC4 regulatory network was constructed based on DElncRNAs and DEmiRNAs using bioinformatics methods. Further, RT-PCR was performed to validate the expression of the representative molecules in the ceRNA network in CRC and normal tissues. The relationship between drug sensitivity and these molecules were also evaluated. Our research might be useful for further studies of the molecular mechanism of ERCC4 in CRC.

Materials and Methods

Results

The expression and prognostic value of ERCC4 in CRC

To investigate the possible role of ERCC4 in CRC, based on TCGA data, we found that ERCC4 was overexpressed in many kinds of tumors on mRNA level, including CRC (Fig. 1A). In addition, based on the HPA database, high expression of ERCC4 was also observed on protein level in CRC (Fig. 1B). Since ERCC4 was up-regulated in tumor specimens, we then investigated the clinical significance of ERCC4 expression in CRC patients. According to the Kaplan-Meier survival curves, as shown in Fig. 1C, our results showed that higher expression of ERCC4 was correlated with poor overall survival (OS) in colon cancer patients.

Identification of DEmiRNAs and DElncRNAs

In total, 1,885 DElncRNAs (412 upregulated and 1,473 downregulated), 68 DEmiRNAs (7 upregulated and 61 downregulated) were identified from CRC samples in ERCC4^high and ERCC4^low expression groups (Figs. 2A, 2B, File S2).

Survival analysis of the molecules in the ceRNA network

To investigate the clinical values of the molecules in the ceRNA network in CRC, Kaplan-Meier survival curves were used to evaluate the prognostic significance. The results showed that miR-200c-3p (hazard ratio [HR] = 0.62, P = 0.032), MALAT1 (hazard ratio [HR] = 1.54, P = 0.016), and AC005520.2 (hazard ratio [HR] = 1.75, P = 0.002) were significantly associated with the prognosis of CRC (Figs. 3A–3F).

RT-PCR verification of the molecules in the ceRNA network

To further verify our bioinformatic analysis results, we performed RT-PCR experiment to explore the expression of the molecules in the ceRNA network in CRC and normal tissues. Because MALAT1 showed a prognostic value in CRC, and it was a well-studied lncRNA in cancers, we selected ERCC4, miR-200c-3p and MALAT1 for verification. Through differential expression analysis, it could be found out that ERCC4 and MALAT1 were up-regulated in CRC compared with normal tissues, while miR-200c-3p was down-regulated in CRC compared with normal tissues, which were all with statistical significance (Fig. 4).

In addition, we validated the association of MALAT1-miR-200c-3p-ERCC4 by correlation analysis (Fig. 5). A strong negative correlation (r = −0.613) was found between MALAT1 and miR-200c-3p (P < 0.0001), while the correlation between miR-200c-3p and ERCC4 was not statistically significant (P = 0.723).

The correlation between ERCC4, miR-200c-3p, MALAT1 and drug sensitivity

We separately explored the relationship between drug sensitivity and ERCC4, miR-200c and MALAT1 using RNAactDrug database. The results were shown in File S3. In addition, we found that ERCC4, miR-200c and MALAT1 were all associated with Cisplatin (Fig. 6).

Discussion

NER pathway plays an important role in identifying and repairing DNA damage, in which ERCC4 is one of the unnegligible molecules. Abnormal expression of ERCC4 can lead the imbalance between DNA damage and repair (Neven et al., 2018). ERCC4 has been found with high-expression in cancer tissues, including CRC (McDaniel & Schultz, 2008). Although there have been a few studies on ERCC4 in several cancers, its epigenetic regulation is still not fully understood. Therefore, we attempted to establish a ERCC4-related ceRNA triple network in CRC to provide a better understanding of the biological mechanisms of ERCC4 in CRC.

Firstly, we evaluated the expression and prognostic value of ERCC4 in CRC using TCGA data. ERCC4 was found to be upregulated in CRC tissues compared with normal tissues both on mRNA and protein levels. It was also associated with prognosis of colon cancer patients. Then 1885 DElncRNAs and 68 DEmiRNAs were sorted out from CRC samples in ERCC4^high and ERCC4^low expression groups. One miRNA and 10 lncRNAs were then identified from the DElncRNAs and DEmiRNAs through bioinformatics prediction using Starbase database. Based on these results, a lncRNA-miRNA-ERCC4 regulatory network was constructed. We also investigated the clinical values of the molecules in the ceRNA network in CRC, and found that miR-200c-3p, MALAT1 and AC005520.2 were significantly associated with the prognosis of CRC.

Further, we performed RT-PCR experiment to verify the expression of the molecules in the ceRNA network in CRC and normal tissues. By comprehensively considering the prognostic value and research status, we selected ERCC4, miR-200c-3p and MALAT1 for verification. Our results showed that ERCC4 and MALAT1 were up-regulated in CRC, while miR-200c-3p was down-regulated in CRC. In addition, we validated the association of MALAT1-miR-200c-3p-ERCC4 by correlation analysis. A strong negative correlation was observed between MALAT1 and miR-200c-3p. The drug sensitivity analysis showed that ERCC4, miR-200c and MALAT1 were all associated with Cisplatin. Therefore, the MALAT1-miR-200c-3p-ERCC4 axis may be closely associated with the development, prognosis and chemotherapy sensitivity of CRC and deserve further study.

ERCC4, as a key gene in NER, is mainly involved in the removal of damage fragments (Friedberg, 2001). ERCC4 domain is also necessary to form a tight complex with ERCC1, a structure specific DNA repair endonuclease responsible for 5’-primer cleavages during DNA excision and repair (Tripsianes et al., 2005; Tsodikov et al., 2005). MALAT1 is located on human chromosome 11q13 and can generate an 8 kb length transcript of long non-coding and nuclear-retained (Ji et al., 2003; Lin et al., 2007; Zhang, Hamblin & Yin, 2017). It can act at the transcriptional, post-transcriptional and translational levels, thus it may be involved in various biological processes such as DNA damage repair, metabolism, and cell signaling (Castro-Oropeza et al., 2018). MiR-200c is highly studied because it has an impact on development, stemness, proliferation, epithelial–mesenchymal transition (EMT), metastasis and therapy resistance (Mutlu et al., 2016). MiR-200c is predicted to target SIRT1 which was found to be negatively regulated by Poly polymerase-2, an enzyme that plays significant roles in the DNA damage response (Nabih, 2020). Therefore, these non-coding RNAs in ceRNA network may competitively regulate ERCC4 to affect the balance between DNA damage and repair, and participate in the development of CRC.

MALAT1 as a highly conserved lncRNA, has been found to be overexpressed in several human neoplasms and to promote tumor cell invasion and metastasis (Castro-Oropeza et al., 2018). It has been reported regulating mRNA expression and then contributing to CRC progression by sponging miRNA in a few studies, like miR-26a/26b (Xu et al., 2020), miR-106b-5p (Zhuang et al., 2019), and miR-203 (Wu et al., 2018). The miR-200 family played an important role in EMT. It was mentioned that miR-200c-3p was not only differentially expressed in CRC, but also may inhibit CRC migration and invasion, and promote apoptosis after LPS stimulation (Huang et al., 2020; Jiang et al., 2020). The miR-200c was directly target ZEB1 and ZEB2 which are the CDH1 transcription inhibitors. Inhibition of miR-200c reduces the expression of CDH1 and therefore induces EMT (Gregory et al., 2008; Park et al., 2008). MALAT1 has been shown to induce EMT during endometriosis (Du et al., 2019) and metastasis in clear cell kidney carcinoma mouse models (Xiao et al., 2015) via the miR200/ZEB2 axis by acting as a sponge for the miR200 family. Sreekumar R found that ZEB2-dependent EMT transcriptional programme drives therapy resistance by activating nucleotide excision repair genes ERCC1 and ERCC4 in CRC (Sreekumar et al., 2021). These results demonstrated that MALAT1, miR-200 and ERCC4 were all connected to the EMT. Therefore, the MALAT1-miR-200c-3p-ERCC4 axis may be involved in the malignant behaviour of CRC by affecting EMT.

DNA repair has an essential role in resistance to platinum agents. Platinum agents inhibit DNA replication and transcription by binding to DNA molecules in the form of platinum DNA adducts (Altaha et al., 2004). ERCC4 is an important protein in the NER pathway, which is a major part of the DNA repair pathway (Fuss & Cooper, 2006; Gillet & Scharer, 2006). CRC cells expressing ZEB2 developed resistance to oxaliplatin while ERCC4 is induced upon ZEB2 expression (Sreekumar et al., 2021). Forced expression of miR-192-5p in SGC7901/DDP cells can significantly inhibit the expression of ERCC4, making the gastric cancer cells more sensitive to cisplatin in vitro and in vivo (Xie et al., 2019). MiR-138-5p regulated the cisplatin resistance by modulating the expression of the DNA repair proteins ERCC4 and ERCC1 in gastric cancer cells (Ning et al., 2019). During repeated treatments of oxaliplatin, the down-regulated expression of miR-200c could lead the resistance to oxaliplatin (Tanaka et al., 2015). It is worth noting that miR-200c and miR-200b can reverse the resistance of epithelial ovarian cancer to cisplatin by targeting DNA methyltransferase (DNMT) (Liu et al., 2019a). Zhang, Li & Zhang (2020) found that MALAT1 modulated ZFP91 through sponging miR-22-3p to promote oxaliplatin resistance in gastric cancer cells. MALAT1 was significantly up-regulated in CRC cells and tissues which wre resistant to oxaliplatin. MALAT1 modulated ADAM17 via sponging miR-324-3p to enhance oxaliplatin resistance in CRC cells (Fan et al., 2020). Liu et al. (2019b) has fingered out that the down-regulated of MALAT1 improved the chemotherapeutic drug sensitivity and inhibited the cisplatin resistance of the bladder cancer. Interestingly, our results showed that ERCC4, miR-200c and MALAT1 were all associated with Cisplatin, which suggested that MALAT1-miR-200c-3p-ERCC4 axis may play a role in the Platinum-based chemotherapy for CRC patients.

However, the limitation of this study should be acknowledged. The sample size is relatively small and a larger sample verification is still needed. The biological relationship of MALAT1-miR-200c-3p-ERCC4 axis and the biological functions they participate in are still unclear, and further in vivo and in vitro studies are needed. It would make more sense if we could obtain the chemotherapy information of patients in TCGA. We will further explore the underlying association between platinum-based chemotherapy and the network in future research.

Conclusion

In summary, we constructed a ceRNA network of ERCC4 in CRC, of which the MALAT1-miR-200c-3p-ERCC4 axis may be involved in the development, prognosis and chemotherapy sensitivity of CRC. This study will help to further reveal the molecular mechanism of ERCC4 in CRC and provide new clues and insights for the further study of NER pathway. However, our results are only based on bioinformatic analysis and tissue validation of small samples, further in vivo and in vitro studies are needed to investigate the biological relationship of MALAT1-miR-200c-3p-ERCC4 axis and its biological function.

Supplemental Information