Potential roles of microRNAs and ROS in colorectal cancer: diagnostic biomarkers and therapeutic targets

As one of the most commonly diagnosed cancers worldwide, colorectal adenocarcinoma often occurs sporadically in individuals aged 50 or above and there is an increase among younger patients under 50. Routine screenings are recommended for this age group to improve early detection. The multifactorial etiology of colorectal cancer consists of both genetic and epigenetic factors. Recently, studies have shown that the development and progression of colorectal cancer can be attributed to aberrant expression of microRNA. Reactive oxygen species (ROS) that play a key role in cancer cell survival, can also lead to carcinogenesis and cancer exacerbations. Given the rapid accumulating knowledge in the field, an updated review regarding microRNA and ROS in colorectal cancer is necessary. An extensive literature search has been conducted in PubMed/Medline databases to review the roles of microRNAs and ROS in colorectal cancer. Unique microRNA expression in tumor tissue, peripheral blood, and fecal samples from patients with colorectal cancer is outlined. Therapeutic approaches focusing on microRNA and ROS in colorectal cancer treatment is also delineated. This review aims to summarize the newest knowledge on the pathogenesis of colorectal cancer in the hopes of discovering novel diagnostic biomarkers and therapeutic techniques.


INTRODUCTION
Over one million new cases of colorectal cancer are identified annually, making it the third most common cancer worldwide [1]. It is largely diagnosed in individuals older than 50 years of age. Thus, standard colonoscopy is recommended starting at the age of 50 to improve early screening and detection [2]. The multifactorial etiology of colorectal cancer consists of both genetic and epigenetic factors. The hereditary syndromes comprise familial adenomatous polyposis, Lynch syndrome (hereditary nonpolyposis colon cancer), MYH-associated polyposis, and juvenile polyposis. Inflammatory bowel disease is also considered a risk factor [3,4]. Moreover, inappropriate diets and lifestyles are known to associate with increased colorectal cancer incidence. Some risk factors include high intake of red meat/processed food, heavy alcohol consumption, obesity, smoking, and physical inactivity. It is suggested that the development of colorectal cancer can be prevented by regular exercise and proper diets [5,6]. The progression of normal colonic mucosa to invasive colorectal cancer requires multiple steps of molecular alterations. The estimated time interval of malignant transformation from normal mucosa to adenomatous polyp and invasive adenocarcinoma is 5-10 years [7]. Five year survival rates for Tumor Node Metastasis (TNM) stages I to IV colorectal cancer are 90%, 80%, 60%, and 8%, respectively [8].

MicroRNA profiles in colorectal tissues, plasma, and stool samples
Nearly 400 dysregulated microRNAs have been identified in colorectal cancer in the past decade, yet minimal consistency of microRNA expression profile is reported. Despite the striking potential of microRNAs as biomarkers of cancer, the transition of microRNAs from bench to clinical use remains challenging as the detection techniques including the commonly used qRT-PCR are needed to be optimized. For example, the selection of suitable reference genes for data normalization in the qRT-PCR analysis is highly subjective, which may lead to inconsistency among different studies. The intrinsic properties of microRNAs, such as a high degree of sequence similarity within the same family, tissuespecific expression, and small-quantity, also cause certain limitations of these detection methods [24]. In addition, by far a considerable amount of results in the literature is derived from retrospective cohorts, thereby limiting the prognostic significance of microRNAs and possibly contributing to some inconsistent results.

ROS in colorectal cancer
ROS, including superoxide (O 2 •-), hydroxyl radical ( • OH) and hydrogen peroxide (H 2 O 2 ), are generated under physiological conditions and serve as important mediators in multiple cell signaling pathways. Despite their importance, excessive ROS can oxidize major cellular components (e.g., DNA, lipids, and proteins), resulting in irreversible damages [33]. Normally, the cellular levels of ROS are carefully monitored by the body's natural antioxidant defense system in order to maintain redox homeostasis. When such homeostasis is disrupted (termed oxidative stress), either due to ROS overproduction or compromised antioxidant function, it can give rise to pathological conditions that ultimately leads to diseases [34,35]. Compared to normal cells, the basal level of ROS has been shown to elevate in cancer cells, which is mainly attributed to increased metabolic activity and altered cellular signaling [26]. Originating from the epithelium in the intestine, colorectal cancer cells have a high metabolic rate and often divide rapidly, potentially causing DNA oxidation [27]. These ROS-induced genetic mutations as well as transcription factor modulations (e.g., hypoxia inducible factor-1) are crucial in the regulation of gene expression relative to cancer cell survival, growth, invasion, and metastasis, contributing to all three stages of carcinogenesis (initiation, promotion and progression) [25,27]. Cancer cells are normally accompanied with strong antioxidant defenses, generating a powerful ROS scavenging capacity that can adapt to a highly oxidized environment and avoid apoptosis [25,26]. Sustained and excessive ROS promote oncogenic activity and genomic instability, contributing to carcinogenesis [28]. The association between colorectal cancer and oxidative stress has been identified in the past decades. Additionally, the increased levels of oxidative stress biomarkers, such as 8-oxodG in DNA, suggest that the ROS are markedly elevated in the whole blood of patients with colorectal cancer [27].

Interaction of microRNAs and ROS
Since the dysregulation of microRNAs and ROS are both observed in colorectal cancer, it is essential to understand their potential role of interaction in relation to colorectal carcinogenesis and progression. Among those microRNAs presented in our tables, miR-210 overexpression has been shown to increase ROS production in colorectal cancer cells [ . Although miR-21-ROS interaction has not been documented in colorectal cancer yet, further research may be necessary to explore towards this direction, and perhaps with other microRNAs, given their tight association with the colorectal cancer.
Certain microRNAs, such as miR-34a, have been shown to inhibit ROS synthesis by silencing the genes that code for mitochondrial complexes and other ROSproducing enzymes, contributing to apoptosis resistance. The restoration of these microRNAs is suggested to sensitize the tumor in response to IR-induced oxidative effects [28]. Clearly, ROS and microRNAs are capable of interacting synergistically or antagonistically to influence the complex and multiphase development of cancer. However, limited effects have been observed in the application of antioxidants or microRNAs in cancer treatment, and the functional consequences of individual microRNA in colorectal cancer patients are still largely unknown. Further exploration on the ROS-microRNAs network may provide powerful therapeutic potential for colorectal cancer as microRNAs can be utilized to enhance ROS-induced apoptosis or alleviate ROS-mediated oxidative stress [28].

Oncogenes miR-21
Oncogenic miR-21 is one of the most extensively studied microRNAs. Its expression is often upregulated in tumor tissue, sera, and stool samples from patients with colorectal cancer [44-46]. Remarkably, serum miR-21 has been shown to be a promising biomarker of colorectal cancer for early detection and prognosis [30]. The expression level of miR-21 is associated with TNM staging, recurrent-free cancer-specific survival, and overall survival [46,47]. In addition, the expression of miR-21 is decreased after chemotherapy, which is related to tumor response [48]. Several studies have shown that miR-21 overexpression significantly increases the resistance of tumor cells to 5-fluorouracil and radiation in colon cancer cells [49]. Furthermore, the knockdown of miR-21 reversed these effects on tumor cells by increasing the sensitivity to 5-fluorouracil chemotherapy [49, 50]. One study suggests that miR-21 mediates resistance through the regulation of Sprouty 2 protein, a tumor suppressor, which enhances the cytotoxic effect of 5-fluorouracil in colon cancer cells [51]. These findings indicate that targeting miR-21 could enhance the sensitivity of cancer cells to chemoradiotherapy.

miR-31
Uniformly elevated expression of miR-31 is also observed in tumor tissue of colorectal cancer patients [64][65][66][67][68][69][70]. The expression level of miR-31 is positively correlated with the tumor TNM staging [65,71]. MiR-31 is involved in cell proliferation and apoptosis by activating RAS signaling pathway through the inhibition of its target gene, RASA1, thereby enhancing cancer cell growth ( Figure 1) [72,73]. Functional analysis has demonstrated that an inhibitor of miR-31 has an anti-tumor effect [74]. Thus, miR-31 may serve as a diagnostic biomarker and a promising therapeutic target in colon cancers.

TUMOR SUPPRESSOR GENES miR-143
As a tumor suppressor, miR-143 is downregulated in tumor tissue, sera, and fecal samples of patients with colorectal cancer [82][83][84][85][86][87]. Its expression is also decreased in the front-specific tumor invasion in liver metastasis [88]. Reduced expression of miR-143 is associated with aggressive mucinous phenotype and is strongly correlated with clinical stage and nodal metastasis [71,89]. The augmented postchemotherapy level of miR-143 is thought to be associated with a better prognosis [48].
Hexokinase 2 (HK2), metastasis-associated in colon cancer-1 (MACC1), insulin-like growth factor-I receptor (IGF-IR), KRAS, and DNA methyltransferases 3A (DNMT3A) are the confirmed target genes of miR-143 [89][90][91][92]. Studies have shown that HK2 is involved in glucose metabolism in colon cancer cells [90]. MACC1 has been identified to express highly in colorectal cancer cells and promotes tumor metastasis through activating a metastasis-inducing HGF/MET signaling pathway [91]. MiR-143 plays a role in suppressing colorectal cancer cell growth via directly inhibiting KRAS [69]. IGF-IR, a known oncogene, has expression levels that are inversely correlated with miR-143 expression in human tumor tissues (Figure 2). In an IGF-IR-dependent manner, miR-143 overexpression has been shown to boost colorectal cancer chemosensitivity to oxaliplatin treatment [89]. In addition, specific changes in DNA methylation patterns are associated with human cancers. Methylation changes to the genome are controlled by DNA methyltransferases (DNMT). MiR-143 specifically regulates DNMT3A that reflects its role in the regulation of DNA methylation (Figure 2) [92]. Taken together, through modulating multiple targets, miR-143 provokes potent effects on cancer cell growth and tumorigenesis.
Reduced expression of miR-145 has been observed in the non-mutated adenomatous polyposis coli, suggesting its role in the initiation of colorectal tumor development [97]. PAK4, N-RAS, IRS1, paxillin, FLI1, DFF45, and p70S6K1 are the known recognized targets of miR-145 ( Figure 2) [96,[98][99][100][101][102][103][104][105]. PAK4, a subfamily of serine/threonine kinases linked to cell growth, motility, and cytoskeletal dynamics, is greatly involved in oncogenic signaling pathways [101]. Overexpressed miR-145 exerts its anti-tumor function by modulating a target gene, PAK4, which blocks the activation of AKT and ERK1/2 pathways, thus inhibiting tumor growth [101]. In addition, miR-145 also blocks the activation of AKT and ERK1/2 pathways and the expression of HIF-1 and VEGF via directly targeting N-RAS and IRS1, leading to the inhibition of tumor growth [96]. Meanwhile, miR-145 is involved in inhibiting cell proliferation, migration, and invasion by targeting paxillin [99]. MiR-145 also regulates the 3′-UTR of Fli-1 mRNA. The downregulation of Fli-1 has a profound effect on the growth of colon cancer [102]. DNA fragmentation factor-45, a direct target by miR-145, plays an essential role in apoptosis, a crucial aspect of tumorigenesis [106]. MiR-145 is a key regulator of intestinal cell differentiation by directly targeting SOX9, a marker of undifferentiated progenitors in the colonic crypts [100]. However, whether SOX9 is oncogenic remains controversial as inactivating mutation of this gene is frequent in colorectal cancer [107]. Indeed, Bastide et al. observed the presence of microadenomas in the intestinal epithelium of SOX9-knockout mice. SOX9 acts a regulator of Wnt/beta-catenin pathway and a transcriptional target to maintain intestinal epithelium homeostasis [108]. In addition, SOX9 expression is an important biomarker for the prediction of colorectal cancer relapse [109]. MiR-145 also directly targets catenin δ-1 and contributes largely to oncogenic Wnt/beta-catenin signaling in human colon cancer cells (Figure 2) [104]. Moreover, miR-145 targets p70S6K1 and downregulates HIF-1 and VEGF expression thereby inhibiting tumor growth and angiogenesis [103]. These findings support miR-145 as an important mediator in tumorigenesis and indicate the potential development of the miR-145-based targeted approach for the treatment of colorectal cancer.

MICRORNAS AS PROMISING THERAPEUTIC TARGETS FOR COLORECTAL CANCER
Exploring the underlying mechanisms that regulate gene expression and the complex signaling pathways is essential in developing novel therapeutics in colorectal cancer treatment. The distinctive ability of microRNAs to target multiple genes and signaling pathways has drawn great attention to their roles as potential innovative therapeutic agents. Generally, microRNA-based therapies include the restoration of downregulated/tumor suppressor microRNA expression or the inhibition of overexpressed/ oncogenic microRNAs [110]. For example, the silencing of miR-135b, in which its upregulation corresponds to colorectal tumor progression, reduces the number and size of tumors in a mouse model with no obvious signs of toxicity [111]. In a mouse model of colon carcinoma, intact tumor suppressor miR-145 molecules that are successfully delivered into xenograft tumors result in profound antitumor effects by increasing apoptosis and reducing tumor size [98]. Despite functioning as oncogenes or tumor suppressors, microRNAs are also involved in immune responses and their aberrant expressions in the immune system are observed in cancers. Indeed, microRNA-related immunotherapy has recently been employed in colorectal cancer treatment.
More than half of the patients with colorectal cancer experience recurrence and metastasis after surgical resection. Sadly, conventional non-surgical treatment such as chemotherapy is ineffective against metastasis, and the poor prognosis further indicates the urge of developing novel therapeutics that improve clinical outcomes [112]. The immunomodulatory role of microRNAs is gradually recognized and investigated by multiple studies. MicroRNAs have been shown to mediate tumor immune escape by indirectly suppressing T cell activation and proliferation. For example, upregulated MiR-21 and miR-130b in advanced colorectal cancer can inhibit phosphatase and tensin homolog (PTEN; a tumor suppressor gene) expression, leading to programmed death ligand 1 (PD-L1) overexpression and immune evasion of colorectal cancer [112,113]. MiR-124 has been shown to posttranscriptionally target signal transducer and activator of transcription 3 (STAT3), which is overly activated in each stage of colorectal cancer development and is associated with tumor-mediated immunosuppression [112,114]. Accordingly, Zhang et al. observed reduced colorectal cancer cell survival and tumor growth in vitro and in vivo, respectively, when miR-124, a microRNA known to be downregulated in colorectal cancer, was re-introduced [114]. It is plausible to control inflammatory signaling pathways and immune responses via microRNA therapy, thereby decreasing the risk of colorectal tumorigenesis.
The development of microRNA-based treatments, including immunotherapy, remains challenging [110,112,115]. Although one single microRNA can regulate a broad set of genes simultaneously, allowing for effective targeting of heterogeneous cancer cells, it is likely to induce off-target side effects due to lack of specificity . The safe delivery and retention of exogenous microRNAs in vivo also pose difficulty in microRNA-based treatments [110,115]. To date, no therapeutic manipulation of microRNAs has been applied in human patients with colorectal cancer; research is only at an experimental stage in either cell lines or animal models. Given the great significance of microRNAs as biomarkers that can be manipulated to reverse tumor progression, we cannot help speculating that a new therapeutic concept--a microRNAbased therapy--might be used in colorectal cancer in the near future. However, standardized microRNA measurement and analysis should be utilized to obtain valid prognostic and diagnostic microRNA profiles in colorectal cancer before their application in the clinical setting.

CONCLUSIONS
Accumulating evidence has significantly expanded our understanding of microRNA in the pathogenesis of colorectal cancer in all critical progressions of tumorigenesis including cell proliferation, apoptosis, metastasis, and chemoresistance. However, we do not yet completely understand the biology of microRNAs in colorectal cancer development, as well as its association with oxidative stress. In addition, varied methods of handing and processing of microRNA yield inconsistent or incomparable results. It is therefore necessary to establish standardized microRNA protocols and analysis to obtain reliable data and accurately quantify microRNAs by developing a set of stable reference genes for each type of cancer. Besides all the unknowns and difficulties, microRNAs have demonstrated their potential as diagnostic biomarkers. Along with redox therapies targeting cancer cells, the development of personalized therapeutic targets utilizing microRNA for colorectal cancer is feasible. However, the aforementioned challenges should be addressed for the successful transition of microRNAs as prospective biomarkers from bench to clinical setting.

ACKNOWLEDGMENTS
We acknowledge Tracey Bender, Tingyang Zhou, Andrew Graef and Alex Ziegler for their excellent assistance during the preparation of this manuscript. We would like to thank all colleagues who have done research in the field of microRNA and colorectal cancer. We apologize to the colleagues whose work was not cited in the review due to space limitation.

CONFLICTS OF INTEREST
No conflicts of interest, financial or otherwise, are declared by the authors.