Recent advances microRNAs and metabolic reprogramming in colorectal cancer research

Colorectal cancer (CRC) is a cancer with the highest incidence and mortality. Alteration of gene expression is the main pathophysiological mechanism of CRC, which results in disturbed signaling pathways and cellular metabolic processes. MicroRNAs are involved in almost all pathophysiological processes and are correlative with colorectal cancer metabolism, proliferation, and chemotherapy resistance. Metabolic reprogramming, an important feature of cancer, is strongly correlative with the development and prognosis of cancers, including colorectal cancer. MicroRNAs can target enzymes involved in metabolic processes, thus playing a regulatory role in tumor metabolism. The disorder of the signaling pathway is another characteristic of tumor, which induces the occurrence and proliferation of tumors, and is closely correlative with the prognosis and chemotherapy resistance of tumor patients. MicroRNAs can target the components of the signaling pathways to regulate their transduction. Understanding the function of microRNAs in the occurrence and proliferation of CRC provides novel insights into the optimal treatment strategies, prognosis, and development of diagnosis in CRC. This article reviews the relationship between CRC and microRNA expression and hopes to provide new options for the diagnosis and treatment of CRC.


MicroRNAs and metabolic reprogramming
Metabolism reprogramming is a critical hallmark of CRC. It refers to the alteration of metabolism that occurs as CRC cells adapt to the tumor microenvironment (TME). These changes satisfy the energy and nutrient demands of tumor cells and sustain the rapid proliferation rate of CRC (6). MiRNAs are important regulators of metabolic reprogramming. They regulate metabolic processes and reshape TME by promoting or inhibiting biomolecules (enzymes or transporters) involved in metabolic processes. The crosstalk between tumor metabolites and TME components is tightly related to proliferation, metastasis, and drug resistance in CRC. In this review, we mainly focus on recent studies of miRNAinduced metabolic reprogramming in glucose, lipid, and amino acid metabolism (Table 1 and Figure 1).

Glucose metabolism
Glucose is a vital energy source for the human body. Therefore, to meet the energy needs of tumor growth, abnormal glucose metabolism often occurs in CRC cells. Unlike most normal tissues, tumor cells can obtain ATP through glycolysis, even in the presence of oxygen (44). This phenomenon is termed aerobic glycolysis, which is also known as the Warburg effect ( Figure 2).
Whereas oxidative phosphorylation of a glucose molecule in normal tissues produces 36 ATP molecules, aerobic glycolysis produces only two (44). However, aerobic glycolysis in tumor cells may occur at much higher rates than oxidative phosphorylation (45). Therefore, it can meet the enormous energy demands of tumor proliferation.
The first step in glucose metabolism is transporting glucose from the extracellular to the intracellular space through glucose transporter 1 (GLUT1) (46). GLUT1 is upregulated in tumor tissues to meet the demand for the large amount of glucose required for aerobic glycolysis (46). During this process, some miRNAs can directly or indirectly act on GLUT1. One study showed that the expression level of miRNA-143 in CRC was significantly related to tumor size and that its overexpression 143 could inhibit the function of GLUT1, thereby inhibiting the proliferation of CRC cells (14). Santasusagna et al. (16) demonstrated that miRNA-328 could inhibit the solute carrier family 2 member 1 (SLC2A1), which encodes the GLUT1 protein, thus downregulating the expression of GLUT1 in tumors and inhibiting tumor proliferation. Increased circRNA DENND4C expression in CRC downregulates the expression of its target miRNA-760, leading to the overexpression of GLUT1, which would accelerate glycolysis and ultimately promote the proliferation of CRC (17). The expression of several of the abovementioned miRNAs was negatively correlated with that of GLUT1; its downregulation could promote CRC proliferation.
The second step in glucose metabolism is hexose phosphorylation, where glucose is converted to glucose-6-phosphate (G-6-P) by hexokinase (HK). Four isoenzymes of hexokinase, HK-1 to HK-4, have been found in mammals, of which HK-1 and HK-2 are found on the outer mitochondrial membrane. In various cancers, including CRC, HK-1 and HK-2 are upregulated to meet the high metabolic state of tumor cells and accelerate tumor cell proliferation, migration, and invasion (47). Many miRNAs target HK1 or HK2 to affect glucose metabolism in CRC and interfere with tumor proliferation. Li et al. (18) found that lncARSR sponged miRNA-34a-5p to promote the expression of HK-1. MiRNA-4458 targets and inhibits HK2 to suppress glycolysis and lactate production (19). I_circ_0000231 functioned as a sponge of miR-502-5p and promoted HK2 protein expression; this indicated that miRNA-502-5p repressed HK2 by inhibiting the expression IHsa_circ_0000231 and ultimately inhibited aerobic glycolysis (20). MiRNA-143, miRNA-513a-3p, and miRNA-98 have low expression in CRC; these miRNAs could target HK2 and downregulate its expression, thereby inhibiting aerobic glycolysis in CRC (15,21,22).
G-6-P is catalyzed by hexose phosphate isomerase to form fructose-6-phosphate (F-6-P), which becomes fructose-1,6bisphosphate (F-1,6-P 2 ) via phosphofructokinase 1 (PFK1). PFK1 is the main rate-limiting enzyme in glycolysis and is involved in the second phosphorylation reaction during glycolysis (48). F-2,6-2P, an isomer of F-1,6-2P, activates PFK1 and accelerates the production of F-1,6-2P (48). The intracellular concentration of F-2,6-2P is regulated by 6-phosphofructo-2-kinase/fructose-2,6biphosphatase 3 (PFKFB3) (49). Cui et al. (23) found that  Overview of miRNAs in metabolism reprogramming of CRC. Schematic illustration shows microRNAs that regulate metabolic reprogramming in colorectal cancer by targeting three metabolically related enzymes. The black arrows depict the transformation of substances during metabolism, and the red arrows depict the inhibitory effect of miRNAs on enzymes involved in metabolism. The expression of miRNAs in red squares decreased in colorectal cancer. The illustration was generated by Figdraw.
miRNA-147b was downregulated in CRC, resulting in the upregulation of PFK1 and PKM2. Deng et al. (24) reported that miRNA-488 targeted PFKFB3 in CRC and that miRNA-488 and PFKFB3 inhibited each other. They suggested that miRNA-488 negatively regulates glycolysis in CRC, which may be related to the regulation of F-2,6-2P by miRNA-488 and PFKFB3. The third step in glycolysis is the production of pyruvate. One molecule of F-1,6-2P generates two molecules of phosphoenolpyruvate (PEP), which then generate two molecules of pyruvate through pyruvate kinase (PK). PK is a key enzyme of glucose metabolism and can be divided into PKM1 and PKM2 (50). PKM2 is expressed in tissues with abnormal differentiation or proliferation. In tumor cells, PKM2 can catalyze the conversion of PEP to pyruvate under aerobic conditions, which plays a vital role in aerobic glycolysis (50). Ren et al. (25) found that miRNA-142-3p is lowly expressed in CRC; however, its overexpression inhibited PKM2 by binding to the 3′ UTR of PKM2 mRNA, thus inhibiting proliferation, migration, and invasion of CRC. Wang et al. (51) reported that the circulaINA hsa_circ_0005963 could enhance the expression of PKM2 by sponging miRNA-122, thereby promoting aerobic glycolysis and chemotherapy resistance. Zheng et al. (26) found that the lncRNA XIST-miRNA-137-PKM axis could control the ratio of PKM1 and PKM2 in CRC. After silencing XIST, the expression of miRNA-137 was upregulated, and the percentage of PKM2/PKM1 was decreased, inhibiting glycolysis.
In the fourth step of aerobic glycolysis, pyruvate is catalyzed by lactate dehydrogenase (LDH) to produce lactic acid in tumor cells, rather being converted to acetyl-CoA in normal tissues under aerobic conditions. The tetramer proteins of LDHA and LDHB include five LDH isoenzymes (LDH1-5). LDHA is upregulated in several cancers and is close to aerobic glycolysis (52). One study reported that the expression of miRNA-34a, miRNA-34c, miRNA-369-3p, miRNA-374a, and miRNA-4524a/b was negatively correlated with that of LDHA in CRC (32). Only the LDHA knockout cannot prevent pyruvate conversion to lactic acid because LDHB can replace the function of LDHA in the absence of LDHA (53). This suggests that simultaneous inhibition of LDHA and LDHB could attenuate tumor aerobic glycolysis. miRNA-335-5p could downregulate the expression of LDHB, thereby inhibiting the proliferation and metastasis of CRC cells (54). After the knockout of both LDHA and LDHB, aerobic glycolysis was disrupted, tumor growth was delayed but not abolished, and glucose metabolism was transformed to oxidative phosphorylation, indicating that tumor metabolism is a complex and flexible process and the Warburg effect is not irreplaceable.
In addition, miRNAs can regulate glucose metabolism in CRC by affecting the expression of certain genes. For example, the forkhead Box M1 (FOXM1) is a protein-coding gene that controls cell proliferation. Both miRNA-1224-5p and miRNA-874-3p have low expression in CRC and inhibit CRC progression and cell glycolysis by targeting the 3′ UTR of FOXM1 (33, 34). Wang et al. (35) found that the CDC28 protein kinase regulatory subunit 1B (CKS1B) is a direct target of miRNA-485-5p. The Effect of miRNAs on glycolysis.
interaction between CKS1B and miRNA-485-5p regulates the expression of GLUT1 and LDHA. MiRNA-485-5p inhibits the development of CRC by reducing the expression of CKS1B.
Some miRNAs are also upregulated in CRC and contributing to promote tumor proliferation, metastasis, and invasion. The levels of miRNA-181a/d are abnormally elevated in CRC, promoting aerobic glycolysis (37,38). Pyruvate dehydrogenase complex X (PDHX) is a protein-coding gene that catalyzes pyruvate conversion to acetyl-CoA. MiRNA-26a suppresses the expression of PDHX by targeting the 3′ UTR mRNA of PDHX and promotes aerobic glycolysis to meet the increased energy demands in CRC (39). Liu et al. (40) found that abnormally upregulated miRNA-142-5p in CRC decreased the expression of SDHB and promoted aerobic glycolysis by increasing glucose consumption and lactate production.

Lipid metabolism
Abnormal lipid metabolism is one of the fundamental metabolic features of cancer. Lipids participate in regulating tumor proliferation and metastasis. b-oxidation is one of the critical steps in lipid metabolism, which produces acetyl-CoA and provides energy for tumor growth and proliferation (55). Some miRNAs can regulate tumor cell proliferation and progression by influencing the enzymes involved in b-oxidation. b-oxidation first requires the activation of fatty acids, which generates fatty acid CoA (FA-CoA) under the catalysis of acyl-CoA synthase (ACS). ACSL5, a subtype of ACS, is upregulated in CRC and related to the poor prognosis of CRC (41). One study showed that miRNA-497-5p negatively regulated ACSL5 (41). They proposed that miRNA-497-5p may act as a therapeutic strategy for regulating lipid metabolism in CRC; however, further research is needed. Cruz-Gil et al. (42) found that miRNA-142, miRNA-544a, and miRNA-19b-1 downregulated the expressions of ACSL1, ACSL4, and SCD. Furthermore, high expression of miRNA-19b-1 was related to a better prognosis in patients with stage II and stage III CRC. One study found that miRNA-146-5p was upregulated in CRC, targeting the 3′ UTR mRNA of HOXC10, which stimulated WAT browning and cachexia, maintaining the hypermetabolic state of CRC (43). In addition, cholesterol accumulation is also a feature of altered lipid metabolism in CRC. Sharma et al. (56) found that miR-18a-5p, miR-144-3p, and miR-663b play roles in regulating cholesterol homeostasis in CRC.

Amino acid metabolism
Glutamine is one of the important molecules in tumor metabolic reprogramming; its metabolite a-ketoglutarate (a-KG) is the key metabolic intermediate in the TCA cycle. Alanine, serine, and cysteine transporter 2 (ASCT2), also known as SLC1A5, is an amino acid transporter that is upregulated in many cancers, including KRAS-mutant CRC (57). ASCT2 can transport glutamine into the cytoplasm for catabolism. Glutamine is deaminated by the catalysis of glutaminase (GLS) to generate glutamate, which is transferred to mitochondria and converted to a-KG. a-KG accesses the TCA cycle to provide energy and participates in the synthesis of amino acids.
MiRNAs can affect glutamine metabolism by regulating glutamine transporters and metabolically related enzymes. MiRNA-137 could downregulate ASCT2 and GLS1, reduce glutamine uptake by tumor tissues, and inhibit glutamine metabolism (27,28). SLC38A1 is a sodium-dependent amino acid transporter that was found to accelerate CRC cell proliferation and metastasis and promote glutamine metabolism while inhibiting apoptosis. However, when miRNA-485-5p was overexpressed, the expression of SLC38A1 was inhibited, and the malignant progression of CRC cells was prevented (36).

MicroRNAs and signaling pathways
The occurrence of CRC is typically accompanied by the dysregulation of signaling pathways. Multiple signaling pathways jointly induce the occurrence, metabolic disturbance, chemotherapy resistance, and metastasis of CRC. Some miRNAs can regulate this process.

Wnt/b-catenin pathways
The Wnt/b-catenin pathways regulate pluripotent stem cell differentiation, organ development and regeneration, and epithelial mesenchymal transformation (EMT) (58). The dysregulation of the Wnt/b-catenin pathways is the prevalent feature of CRC, which facilitates tumor growth, differentiation, and metabolism in early-stage CRC developments (59). More than 90% of CRC patients have Wnt/b-catenin pathway-related gene mutations, among which APC or CTNNB1 mutations lead to the abnormal activation of the pathway (60). In the absence of the activation of the Wnt/b-catenin signaling pathways, b-catenin is degraded by a protein complex composed of adenomatous polyposis coli (APC), axis inhibitory protein (AXIN), glycogen synthase kinase 3 (GSK3), and casein kinase 1 (CK1), which make b-catenin cannot accumulate in the cytoplasm (61). The activated Wnt/b-catenin signaling pathway inhibits the decomposition of b-catenin. The accumulated b-catenin initiates the transcription of downstream target genes, leading to tumor proliferation and metabolic abnormalities in CRC (62). The Wnt/bcatenin signaling pathway activates downstream target genes to enhance aerobic glycolysis and participates in metabolic reprogramming of tumors. PDK1 is a downstream gene of the Wnt/b-catenin signaling pathway; it is also a key metabolic regulator of glycolysis and inhibits the transformation of pyruvate to acetyl-CoA and promotes lactate production ( Figure 3) (63).
Several miRNAs take part in regulating the Wnt/b-catenin signaling pathway. Some miRNAs regulate Wnt/b-catenin signaling by affecting b-catenin expression, some target ligands and receptors in the cell membrane or extracellularly, some regulate the expression of the GSK3b/APC/AXIN/CK1a/PP2A protein complex, while others silence the inhibitors (Table 2) APC is a kind of tumor-inhibiting factor that antagonizes the Wnt/b-catenin signaling pathway. Some miRNAs were upregulated in CRC and activated the Wnt/b-catenin signaling pathway by affecting the expression of APC or APC2 genes. Previous studies have showed that miRNA-501-3p, miRNA-942, miRNA-494, miRNA-582, and miRNA-135a/b are negatively correlated with APC expression, which induces activation of the Wnt/b-catenin signaling pathway and accelerates CRC development (70, 71,[73][74][75]. Similarly, miRNA-663b activates the Wnt/b-catenin signaling pathway by inhibiting APC2 expression (76). In addition, miRNA-103a and miRNA-1827 are highly expressed in CRC; they downregulate the expression of APC/APC2, upregulate the expression of Wnt3a and b-catenin, and then Wnt/b-catenin signaling is enhanced (77). On the contrary, miRNA-137 downregulated in CRC, but it can inhibit the expression of Wnt3a and b-catenin and lead to the reduction in Wnt/b-catenin signaling. The first two promote CRC proliferation and the last one inhibits it by modulating Wnt/b-catenin signaling pathways.
A few miRNAs act on cell membranes or extracellular elements, indirectly affecting the Wnt/b-catenin signaling pathway. Li et al. (83) found that miRNA-135b-5p downregulates ZNRF3, then activates the Wnt/b-catenin signaling pathway, leading to invasion and metastasis of CRC. Si et al. (84) revealed that miRNA-1246 is highly expressed in CRC and activates the Wnt/ b-catenin signaling pathway to promote migration of CRC. Sun et al. (82) found that miRNA-34a binds to the 3′ UTR of WNT1 and inhibits the expression of WNT1.

EGFR signaling pathways
EGFR is essentially a transmembrane glycoprotein that is located on the surface of epithelial cells and is involved in gene regulation of cell proliferation, differentiation, and apoptosis. EGFR is upregulated in various tumors and is closely correlated with development, metastasis, and chemotherapy resistance of CRC. EGFR is activated to regulate downstream RAS/RAF/MAPK, PI3K/Akt, and JAK/STAT pathways, which transmit signals from the cytoplasm to the nucleus, perform gene regulation, and play a role in promoting tumor cell proliferation and anti-apoptosis (85) (Figure 4).
KRAS is an essential molecule in the EGFR signaling pathway. Approximately 40% of CRC cases have KRAS mutations (86). Mutated KRAS does not require ligands to bind to EGFR receptors and directly activate downstream signaling pathways, FIGURE 3 MicroRNAs involved in Wnt/b-catenin signaling pathways. Schematic illustration shows microRNAs that target Wnt/b-catenin signaling pathways (left) and Wnt/b-catenin signaling pathways regulated Warburg effect via PDK1 (right). Pyruvate dehydrogenase (PDH) catalyzes the oxidative decarboxylation of pyruvate. Phosphorylation of PDH by pyruvate dehydrogenase kinase (PDK) results in inactivation and thereby downregulates aerobic respiration and inhibits the formation of acetyl-coenzyme A from pyruvate. As a result, lactic acid production was increased, and Warburg effect was enhanced. The illustration was generated by Figdraw.
leading to resistance to anti-EGFR-targeted drugs, such as cetuximab and panitumumab (87). Activation of EGFR signaling pathways increases GLUT1 expression in CRC epithelial cells, inducing the Warburg effect (88).
Some others have also found that mutated KRAS activates the PI3K-AKT-mTOR pathway, leading to metabolic reprogramming in CRC (98,99). In addition, Xu et al. (100) reported that circRNA_0000392 regulates the PI3K-AKT pathway through sponge miRNA-193a-5p; overexpression of miRNA-193a-5p can significantly reduce the expression level of PIK3R3. Jiang et al. (101) found that circIL4R activates the PI3K-AKT signaling pathway through sponge miR-761. Tang et al. (102) found that miRNA-19a target the PI3K-AKT-mTOR signaling pathway; overexpressed miRNA-19a activates the PI3K-AKT-mTOR signaling pathway and affects the development of CRC. The activation of STAT3 in the JAK/STAT pathway, a oncogenic transcription factor, is often associated with CRC progression and poor prognosis. Wang et al. (103) found that STAT3 upregulated miRNA-572 expression in CRC cell lines. Modulator of apoptosis 1 (MOAP-1) is a pro-apoptotic protein, and the expression of MOAP-1 is inhibited by miRNA-572. STAT3 induced CRC cell growth, migration, and invasion through miR-572-MOAP-1 pathway. In addition, downregulation of PIAS3, an inhibitor of STAT, and high activation of NF-kB and STAT3 have been observed in CRC patients. Ma et al. (104) found that the activation of STAT3 and NF-kB led to a significant increase in miRNA-18a levels in the colon epithelium, and that overexpressed miRNA-18a promoted CRC growth by inhibiting PIAS3.

TGF-b signaling pathway
The TGF-b signaling pathway has the function of immune supervision. Its main components are TGF-b r1, TGF-b r2, and downstream SMAD molecules ( Figure 5). TGF-b ligands bind to receptors on the cell membrane, activate downstream SMAD, induce SMAD accumulation in the nucleus, and participate in transcriptional regulation. In the early stage of tumorigenesis, the TGF-b signaling pathway can inhibit tumor proliferation, while in the late stage of the tumor, a persistently elevated or abnormally transduced TGF-b signaling pathway may lead to tumor proliferation and metastasis (105). There are eight subtypes of SMAD. MiRNA-27a targets and inhibits SMAD2 and SMAD4, then inhibits the development and metastasis of CRC (106,107). Zhai et al. (108) found that miRNA-140-5p is low expressed in CRC; when miRNA-140-5p is overexpressed, it can inhibit the expression of SMAD, resulting in reduced tumor proliferation and invasion. In addition, miRNA-34a and miRNA-18 are negatively correlated with the expression of SMAD4 (109, 110). SMAD6 and SMAD7 are inhibitory molecules that inhibit TGF-b signaling. MiRNA-581 targets SMAD7 and negatively correlates with its expression level. Similarly, miRNA-4775 negatively correlates with the expression level of SMAD7 (111). Both miRNA-581 and miRNA-4775 can promote the invasion and migration of CRC through the SMAD7-TGF-b pathway and induce the epithelialmesenchymal transformation (EMT) of CRC. In addition, Bu et al. (112) found that miRNA-1269a activates the TGF-b signaling pathway by inhibiting SMAD7, which, in turn, upregulates the expression level of miRNA-1269a, forming a positive feedback loop, thereby promoting the recurrence and metastasis of CRC.

NOTCH pathway
Notch signaling pathway regulates cell proliferation, differentiation, apoptosis, enhanced stem-like properties, and chemoresistance in CRC cells. Notch signaling pathway consists of four parts: Notch receptor, Notch ligand, CSLDNA-binding protein, and the downstream target gene. Notch receptor binding with ligand triggers the activation of Notch signal. Notch receptor proteolysis occurs twice and is transferred to the nucleus, thus activating target gene transcription and regulating protein expression. There are four Notch receptors (Notch1, 2, 3, and 4) and five Notch ligands (Jagged 1, Jagged 2, Delta 1, Delta 3, and Delta 4) in the human body.
MiRNA-34a negatively regulates Notch1 and is downregulated in multiple cancers, including CRC. Zhang et al. (113) found that miRNA-34a could predict patients' chemotherapy response to 5-FU and was promising in combination with 5-FU for CRC. MiRNA- Overview of EGFR signaling pathways.
195-5p inhibits the expression of Notch2 and recombination signal binding protein for immunoglobulin kappa J region (RBPJ) by targeting and binding to their 3′ UTRs, thereby inhibiting cell stemness and chemotherapy resistance of CRC (114). Upregulation of miRNA-206 inhibited the expression of NOTCH3 in SW480 and SW620 cells, leading to the inhibition of CRC proliferation and migration and activation of apoptosis (115). Chen et al. (116) found that miRNA-598 inhibited metastasis in CRC by inactivating the JAG1/Notch2 pathway to inhibit the EMT of CRC cells.
3 Clinical diagnosis and treatment strategies based on miRNAs

Early diagnosis of CRC
Patients with early-stage CRC detected through screening have more opportunities for treatment and a better prognosis. Although colonoscopy is the gold standard for CRC diagnosis, it is not widely used owing to its invasiveness and high cost. Some miRNAs are maladjusted in CRC and can serve as biomarkers for diagnosis. A meta-analysis showed that serum miRNA-21 had a sensitivity and specificity of 77% and 83%, respectively, for CRC, showing its potential diagnostic value (117). Furthermore, miRNA-92a, miRNA-18a, miRNA-144, and miRNA-29a have been studied and showed a good sensitivity and specificity, thus indicating that they are possible biomarkers for CRC diagnosis (118)(119)(120)(121). Some miRNAs derived from the serum or stool may be used as novel biomarkers for screening assays.

Prognosis of CRC
The 5-year relative survival of CRC remains at only 68%. The main reasons for the poor prognosis of patients with advanced CRC are tumor metastasis and recurrence. Approximately 50% of advanced CRC will metastasize, accounting for the high mortality rate of CRC patients. After radical surgery, approximately a third of patients experience tumor recurrence (124). Therefore, it is necessary to develop new therapeutic modalities and prognostic biomarkers. MiRNAs are involved in tumorigenesis, progression, and metastasis (125). Many researchers have found various miRNAs as prognostic biomarkers and therapeutic targets for CRC.

Metastasis of CRC
EMT is a biological process where epithelial cells derived from malignant tumor cells acquire the ability to migrate and invade. EMT is related to the metastasis of CRC. Many miRNAs interact directly or indirectly with EMT-related molecules to induce or inhibit the progression of EMTs, thereby affecting tumor metastasis.  Therefore, prevention of EMT may be an effective method to inhibit tumor metastasis. SNAIL, ZEB1, and ZEB2 are transcription factors of EMT that regulate epithelial and mesenchymal markers and induce EMT in CRC. Members of the miRNA-200 family (miRNA-200a, miRNA-200b, miRNA-200c, miRNA-141, and miRNA-429) reduce the migration and invasion of CRC cells by inhibiting ZEB1 and ZEB2 mRNA translation. MiRNA-132, miRNA-92, and miRNA-335 also directly target ZEB2. Members of the miRNA-34 family suppress SNAIL and inhibit EMT, reducing the migration of CRC cells. Downregulation of these miRNAs is associated with distant metastasis and advanced tumors (126). In addition, miRNA-4775 and miRNA-496 promote EMT and migration in CRC through the TGF-b and WNT/b signaling pathways, respectively (127,128). miRNA-192 and miRNA-194, which are downstream molecules of SNAIL, inhibit EMT in CRC (129). MiRNA-612, miRNA-219-5p, miRNA-185, miRNA-296, and miRNA-421 target carcinogenic factors and inhibit metastasis of EMT and CRC (130).
Metastasis of colorectal cancer is a complex process that involves EMT, angiogenesis, and remodeling of TME. Prevention of these processes is an effective strategy to inhibit tumor metastasis, and miRNAs are involved in the regulation of these processes. Therefore, miRNAs can serve as biomarkers to determine the risk of CRC metastasis.

Recurrence of CRC
The 5-year overall recurrence rate was 9.3% for stage I CRC, 27.2% for Stage II CRC, and 56.1% for Stage III CRC. Tumor TMN stage is considered a main risk factor and the only independent predictor of recurrence (138). However, recently, miRNAs have been demonstrated as potential biomarkers to predict CRC recurrence.

Chemotherapy resistance of CRC
Surgery is the most effective treatment for CRC; however, when the patient is in the advanced stage of the disease, surgical resection alone does not significantly prolong overall survival. In these cases, adjuvant therapy, such as chemotherapy, radiotherapy, and targeted drug therapy, can effectively reduce tumor volume and prolong patient survival. Chemotherapy is an effective treatment for most CRC patients, but resistance to chemotherapy can reduce its effectiveness. The miRNA expression of patients receiving chemotherapy differ significantly. First-line chemotherapy drugs for CRC include 5-FU, capecitabine, oxaliplatin, irinotecan, cetuximab, and bevacizumab. miRNA levels can affect the targets of chemotherapeutic drugs or related signaling pathways. Therefore, the miRNA level of patients' serum can be used to predict the efficacy of chemotherapy and provide a reference for clinical decision-making. These mechanisms can guide clinicians in choosing the most appropriate chemotherapy regimen.

Capecitabine and 5-FU
Capecitabine is an oral fluorouracil that is converted to 5-FU in tumor tissues, and 5-FU inhibits tumor growth by interfering with DNA synthesis. MiRNAs can regulate the transduction of signaling pathways to regulate chemotherapy resistance in tumors. MiRNA-149 and miRNA-320 target FoxM1 and affect the Wnt/b-catenin signaling pathway and glucose metabolism, thereby increasing the sensitivity of 5-FU (143,144). MiRNA-135b, miRNA-182, and miRNA-587 can promote 5-FU resistance in CRC by activating the PI3K-AKT pathway (145,146). In contrast, miRNA-302 and miRNA-20b inhibit the EGFR-AKT pathway and increase the sensitivity of 5-FU in CRC (147, 148).

Oxaliplatin
Oxaliplatin acts on DNA and inhibits tumor growth and proliferation by blocking DNA replication and transcription. It is often used in combination with other anticancer agents. The FOLFOX (5-FU combined with oxaliplatin) and CapeOX (capecitabine combined with oxaliplatin) regimens showed better response rates, progression-free survival, and overall survival in first-line treatment of advanced or metastatic CRC. Activation of the Wnt/b-catenin signaling pathway is correlated with oxaliplatin resistance in CRC. MiRNA-103 and miRNA-107 promote the Wnt/ b-catenin signaling pathway to induce oxaliplatin resistance in CRC (74). In addition, miRNA-506 overexpression inhibits the Wnt/bcatenin signaling pathway and enhances sensitivity to oxaliplatin (149). In the EGFR signaling pathway, miRNA-17 and miRNA-19a can downregulate PTEN and reduce oxaliplatin resistance in CRC (150, 151). Furthermore, Zhou et al. (152) found that miRNA-203 induced oxaliplatin resistance in CRC by inhibiting ataxia telangiectasia mutated (ATM) kinase expression.

Irinotecan
The FOLFIRI regimen (irinotecan combined with 5-FU) is not suitable for postoperative adjuvant chemotherapy for CRC but can be used for palliative treatment of advanced or metastatic CRC. Bitarte et al. (153) found that miRNA-451 expression levels were lower in patients who did not respond to first-line irinotecan therapy, which was associated with downregulation of miRNA-451 expression in the intestinal nodal globule. The expression of miRNA-451 led to the self-renewal of intestinal nodal globules, reducing their oncogenicity and resistance to irinotecan chemotherapy. In addition, the overexpression of miRNA-146a, miRNA-194, miRNA-514b-3p, and miRNA-519c can increase the sensitivity of CRC to irinotecan (154-157).

Cetuximab
Cetuximab is an EGFR-targeting monoclonal antibody commonly used in wild-type KRAS advanced or metastatic CRC patients. The upregulation of miRNA-100/125b promoted Wnt/bcatenin signaling and mediated resistance to cetuximab in CRC (158).

Treatment of CRC based on miRNAs
In addition to the increased sensitivity of chemotherapeutic agents mentioned above, there are two miRNAs-based therapies: (1) miRNAs used in immunotherapy (2) miRNAs as therapeutic targets.
Tumor-related miRNAs can be divided into two categories: oncomiRNAs are overexpressed in tumors, whereas tumorsuppressor miRNAs are underexpressed. Two treatment strategies have been proposed for treating patients with cancer using miRNAbased therapy: suppression of oncomiRNAs and upregulation of tumor suppressor miRNAs (162). These methods involve the artificially regulation of miRNA expression. The former usually involves the use of miRNA mimics to artificially increase the expression of target miRNA; hence, that it can exert tumor inhibition. Meanwhile, the latter can be achieved using miRNA antagonists, including antisense oligonucleotides, antagomirs, and miRNA sponges. Wu et al. (163) developed a novel miRNA-129 mimic (Mimic-1) that can be delivered to cancer cells without the need for any delivery vehicle. Mimic-1 showed potent efficacy in eliminating resistant colon cancer stem cells both in vitro and in vivo.
Although miRNAs have shown great potential in the diagnosis, prognosis, and treatment of CRC, their application in clinical practice requires more extensive and in-depth research.

Conclusion and prospective
MiRNAs are involved in almost all pathophysiological processes in CRC and play an essential role in its occurrence and development, metabolic reprogramming, abnormal signaling pathways, treatment, and prognosis. The changes in some miRNAs, in combination with conventional detection methods, can improve the diagnosis of CRC at an early stage. The expression levels of some miRNAs can also predict the efficacy of chemotherapy and help choose the best chemotherapy regimen. Therapeutic strategies that target the correction of maladjusted miRNAs are promising. In the future, we can develop screening assays with better sensitivity and specificity based on relevant miRNAs, suppress tumor growth by correcting abnormally elevated or suppressed miRNA expression, predict chemotherapy efficacy based on miRNA expression, and synergistically treat tumors using chemotherapy drugs. Our understanding of some miRNAs is not perfect, clinical translation of theoretical knowledge remains limited, and further research is needed. Investigating the role of miRNAs in the occurrence and treatment of CRC is of great significance for improving its diagnosis and prognosis.

Author contributions
All authors have read and agreed to the published version of the manuscript. BX and QH drafted the first version of this manuscript. HZ collected literature. SL provided critical revisions and edited the manuscript. JX revised the manuscript. All authors contributed to the article and approved the submitted version.

Funding
This work was jointly supported by the 2021 joint fund project for science and technology innovation of Fujian province (2021Y9029) and Science and Technology Bureau of QuanZhou (Grant number 2020CT003).