Investigating the biomarker value of circRNAs in the diagnosis of colorectal cancer: a systematic review

Abstract

Circular RNAs (circRNAs) has recently been considered a class of endogenous RNAs that form a continuous closed loop with an ability to cancer development. Due to its properties, circRNAs has promising potential to be considered as non-invasive cancer biomarkers. The present review comprehensively and systematically assessed the role of circRNAs as diagnostic markers in blood or tissue samples of colorectal cancer (CRC) patients. Articles published until September 2022 were searched across Scopus, Web of Science (WoS), and PubMed databases to screen and find suitable circRNAs as diagnostic markers in CRC, Based on inclusion and exclusion criteria, 55 articles were selected as the final included articles. The sample size of the patients group ranged from 12 to 212. Among the circRNAs investigated in this study, 47 circRNAs were increased in patients and probably act as oncogenes and activate the downstream pathways in the initiation and exacerbation of cancer; 28 circRNAs were decreased, which probably acted as tumor suppressors and their decrease played a role in the progression of CRC, indicating that the aberrant expression of circRNAs are involved in the promotion of CRC and may associated with its clinicopathological features. Many pathological processes of cancer are regulated by the altered expression of circRNAs through the regulation of different signaling pathways. According to the results of this study, while circRNAs demonstrates promise as a biomarker for CRC diagnosis, it is necessary to conduct evaluative studies on a larger scale in future.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12672-024-01602-z.

Introduction

Colorectal cancer (CRC), one of the common malignant neoplasms in the digestive tract, originating from the non-cancerous proliferation of mucosal epithelial cells [1]. CRC is a severe peril for global health with nearly 2 million new cases and over 900,000 deaths worldwide in 2021 [2]. It is estimated that in the year 2024, in excess of 150,000 novel cases of colon and rectal cancer will be identified within the United States, and approximately 53,000 people will die from this cancer. Colorectal cancer was historically the fourth predominant cause of cancer-related mortality among both male and female individuals under the age of 50 during the late 1990s; however, it has since ascended to the foremost position among men and the second position among women [3].

The National Central Cancer Registry of China's data collection demonstrates the enormous rise in the incidence and mortality of CRC [4]. In the next few years, modern lifestyles and society's age structure will increase the new cases of CRC and even death. Due to the lack of early cancer diagnosis and moreover absence of efficient and standard treatment, the mortality rate of CRC has been more than 80% in the past three decades [5]. It was shown that the 5-year survival rate of CRC patients is related to TNM stages. Some studies reported nearly 20% overall 5-year survival rate in the last stages of the disease and about 90% in the early stages [6, 7]. Detecting lesions with typical cancer diagnostic clinical methods including, CT, MRI, endoscopy, ultrasound, X-ray, and nuclear imaging is possible when asymptomatic masses grow enough observable. Some tumors are in the middle or final stages or have metastasized, therefore, developing and improving biomarker(s) and other diagnostic strategies is a necessity. The types of cancer biomarkers are DNA, RNA, enzymes, metabolites or transcription factors, which are useful in the fields of patient classification, prediction, early diagnosis and prognosis of the disease. These biomaterials may be produced by the tumor itself or as a reaction to the tumor from the host body [8]. Among these biomarkers, the ideal one for CRC diagnosis is the one with following features: It can be easily obtained and quantitatively measured by a non-invasive and inexpensive method using readily available biological samples such as serum, breath, urine, or feces, very specific and sensitive, reliable and repeatable. This biomarker should also differentiate between patient populations and select patients who require second-line testing (endoscopic and radiological investigations) [9]. Non-coding RNAs (ncRNAs) constitute more than 90% of the transcripted RNAs from the genome of mammals. The critical molecular role of RNAs has been revealed after understanding complicated gene expression dynamics in various types of cancers [10, 11]. Circular RNAs (circRNAs), a special subgroup of lncRNAs vastly existing in mammalian cells, has recently been considered a class of endogenous RNAs that form a continuous closed loop [12, 13]. In 1976, in a plant-based virus was observed the first instance of circRNA, which were then referred to as “covalently closed circRNAs molecules” [14]. However, circRNAs were initially considered to be the byproducts of abnormal RNA splicing and did not lure much attention from researchers during the next decades [15]. Advances in bioinformatics increasingly have led to identifying a large amount of circRNAs, and some of their features [16].

CircRNAs are new and beneficial genetic tools for examining physiology and pathology in human disorders, due to their wide expression in a diverse range of human cell types. They are more stable than the normal liner RNAs, because of their specific circular structure, which provides the absence of a 3′ overhang in the RNA. Furthermore, they can conserve under RNase R treatment [17]. In addition, circRNA expression has been found to be cell-type and stage-specific [18–20]. Genome-wide analysis has realized that circRNAs show higher sequence conservation and more abundance than their linear counterparts [21].

The ceRNA mechanism or RNA sponge is the most prominent function noted for circular RNAs. circRNAs act as miRNA sponges and competitively bind to miRNAs and regulate target gene expression by inhibiting miRNA's translational and degradational functions. Although circRNAs are classified as long non coding RNAs (lncRNAs), there is evidence that circRNAs contain N6-methyladenosine (m6A) modifications or internal ribosome entry site (IRES) or start codon can be translated into protein. However, Lacking the 5′-cap due to the circular structure of circRNAs causes the translation to be done through cap-independent manner. In addition, studies have shown that circRNA exert their biological functions through CircRNA-protein interactions including: altering interactions between proteins, Blocking proteins from DNA, RNA, and other proteins; recruiting chromatin remodelers, transcription factors and modifying enzymes to chromatin; forming circRNA-protein-mRNA ternary complexes regulating translation or RNA stability; and translocating proteins to the nucleus or cytoplasm (Fig. 1) [22–24].

Fig. 1.

Main functions of circRNAs, A miRNA sponges, B interacting with proteins, C transcriptional regulators, D translation template

The hypothesis of circRNAs dysregulation expected in CRC and connected to CRC pathogenesis, on the other hand, it is supported by the different circRNA expression profiles between non-cancerous and cancerous colorectal cells [25, 26]. These features indicate that circRNAs may be able to play a role in pathological cellular processes and create a potential for their use as biomarkers with diagnostic value. Therefore, the present review comprehensively and systematically assessed the circRNAs literatures related to diagnostic biomarkers in blood or tissue.

Materials and methods

Literature search strategy

In the present study, electronic databases including, Scopus, Web of Science (WoS), and PubMed databases were searched to screen and find suitable circRNAs as diagnostic markers in CRC. The search was conducted by two researchers separately on September 22, 2022 and at the same time. Search strategy in each database was (colorectal cancer) AND (Circular RNA) AND (sensitivity) OR (specificity) OR ("positive predictive value") OR ("negative predictive value") OR ("positive likelihood ratio") OR ("negative likelihood ratio") OR (AUC) OR ("Area under the ROC Curve") OR (ROC) OR ("Receiver operating characteristics"). TITLE-ABS-KEY was used in Scopus database, Topic and all Fields were used in Web of Science and PubMed databases, respectively.

Study selection criteria

Inclusion criteria were all case–control, cross-sectional, or cohort studies that investigated circRNAs in order to early diagnosis of CRC. Studies are reviewed, regardless of the year of publication and the country of origin. The language of the reviewed articles was only English. The exclusion criteria were as follows: (1) qualitative studies, (2) letters to the editor, (3) review studies, meta-analysis and pure bioinformatics; (4) case report; (5) no clinical sample; (6) in vitro and animal studies, and (7) in general studies that were not in line with the purpose of the study or with insufficient data. Two researchers (AK and TL) separately reviewed all studies. If there was any disagreement about any specific issue, it was resolved after discussion by two of the authors (AM and SA).

Data extraction

Two researchers, AK and TL, independently extracted the data. The extracted data included the following information: the name of the first author, year of publication, country, number of participants in each group (patients and controls), and characteristics of the patients, such as age, gender, and race. Additionally, the type of biological samples, the technique used, the names of circRNAs and their alterations, the degree of sensitivity and specificity, and the area under the curve were also recorded.

After data extraction, the studies were classified and summarized based on the screened circRNAs, their sensitivity and specificity, and their role in the early diagnosis of colorectal cancer (CRC).

The flow chart of this study is shown in Fig. 2.

Fig. 2.

Flowchart of systematic study progress and identification of studies

The quality of the included studies was assessed in according to the Newcastle–Ottawa scale (NOS).

Results

After searching the databases, a total of 405 studies were found, of which 151, 92, and 162 were related to Scopus, WoS, and PubMed, respectively. Then, using EndNote program to remove duplicates, 255 unique articles were found. Then, by reading the articles by two researchers separately, with the reasons of animal and in vitro study, review studies, meta-analysis, pure bioinformatics, 8, 77, 13 and 20 other studies were eliminated. Other reasons also led to the exclusion of a number of studies that were outside the main objective of our study (early detection of CRC by circRNAs). The number of these studies was 39 and a total of 74 studies remained. Unfortunately, 3 of the articles lacked full text and the correspondence with the authors did not yield results. Among the 71 articles found and after reading their full text, 16 articles were removed due to various reasons such as not having a control group, not performing statistical analysis, etc. and 55 articles were selected as the final articles (Fig. 2).

The highest number of articles belonged to China with 45, followed by Iran with 6 articles in this field. The sample size of the patients group ranged from 12 to 212. Table 1 shows the main characteristics of the selected case–control manuscripts, including author and year, investigated circRNA(s), biological samples, sensitivity (95% CI), specificity (95% CI) and area under the curve (AUC). This table shows the distinction between CRC and non-CRC patient samples and the diagnostic role of circRNAs investigated in the studies. According to the studies, it was observed that 47 circRNAs are increased in patients and probably act as oncogenes and activate the downstream pathways in the initiation and exacerbation of cancer. 28 circRNAs were decreased, which probably acted as tumor suppressors and their decrease played a role in the progression of CRC.

Table 1.

Characteristics of the reviewed studies and diagnostic value of circRNAs in CRC

Authors/year (References)	Study no.	circRNA profile (fold change/up or down)	Sample size (CRC/non-CRC)		Biological specimens (serum, plasma, stool, saliva, tissue)	Sensitivity (95% CI)	Specificity (95% CI)	AUC	Biomarker categorization
Chen et al. 2022 [27]	1	Hsa_circ_0084927/Up	96	96	Tissue/plasma			0.78 (CI 0.65–0.87)	Prognostic
Mohammadi et al. 2021 [28]	2	Hsa_circ_0006282/Up	100	100	Tissue/plasma	71.2	80.1	0.812 (CI 0.625–0.801)	Diagnostic, monitoring
Zheng et al. 2022 [29]	3	Hsa_circ_0087960 (circLPAR1)/Down	112	60	tumor/normal tissues and peripheral blood samples	87.3	76.3	0.858 (CI 0.786–0.930)	Diagnostic
Zhang et al. 2019 [30]	4	Hsa_circ_0000615 (circZNF609)/Down	45/ 46	45/ 46	tumor tissue and adjacent normal colon tissue/peripheral blood	65.2	80.4	0.767 (CI 0.670–0.864)	Diagnostic
Ji et al. 2018 [31]	5	Has_circ_0001649/Down	64/18	64/18	Tissue/peripheral blood	82.8	78.1	0.831	Diagnostic
He et al. 2021 [32]	6	Circ_001659/Up	120	98	Peripheral blood – > serum	67.7	91.5	0.872	Diagnostic, prognostic
Yang et al. 2020 [33]	7	Hsa_circ_0002320/Down	50	100	Peripheral blood – > plasma			0.823 (CI 0.097–0.258)	Prognostic
Zhang et al. 2018 [34]	8	Hsa_circ_0007534/Up	112	112/46	Tumor/normal tissues and peripheral blood samples	92	52.2	0.78 (CI 0.703–0.857)	Prognostic
Liang et al. 2020 [35]	9	Hsa_circ_0026416/Up	169/212	169/183	Tumor/normal tissues and peripheral blood samples			0.767	Diagnostic, prognostic
Wang et al. 2021 [36]	10	Hsa_circ_0043278/Down	43	43	Tumor/normal tissues	72	70	0.71	Diagnostic
Yin et al. 2022 [37]	11	Circ_0000375 and circ_0011536/Down	34/56	24/28	Tumor/normal tissues and peripheral blood samples			0.997	Diagnostic
Zhang et al. 2017 [38]	12	Hsa_circRNA_103809 and hsa_circRNA_104700/Down	170	170	Tumor/normal tissues			0.699, 0.616	Diagnostic
Pan et al. 2019 [39]	13	Hsa-circ-0004771/Up	110	35	Tumor/normal tissues and peripheral blood samples	80.91	82.86	0.88 (CI 0.815–0.940)	Diagnostic
Kadkhoda et al. 2021 [40]	14	CircRNA0009910/Up	50	50/10	Tumor/normal tissues	72	84	0.82	Diagnostic, prognostic
Chalbatani et al. 2021 [41]	15	Hsa_circ_001787/Down	43	43	Tumor/normal tissues	69.77	95.35	0.83	Diagnostic, prognostic
Liu et al. 2021 [42]	16	CircNRIP1/Up	72	72	Tumor/normal tissues			0.814	Diagnostic
Barbagallo et al. 2018 [43]	17	CircHIPK3/Up	20/20	20/20	Tumor/normal tissues and peripheral blood samples	71	80	0.771 (CI 0.508–0.936)	Diagnostic, prognostic
Radanova et al. 2021 [44]	18	Hsa_circ_ 0001445 (circSMARcA5), hsa_circ_0003028 (circFUT8), hsa_circ_0007915 (circIPO11) and hsa_circ_0008717 (circABCB10)/Up	Stage IV (n = 122) Stage III (n = 28) 150	90	Peripheral blood – > serum	67.79, 61.54, 61.33, 41.22	68.18, 73.2, 85.06, 90.00	0.739 (CI 0.67–0.81), 0.693 (CI 0.62–0.76), 0.776 (CI 0.71–0.84), 0.626 (CI 0.56–0.70)	Diagnostic, prognostic
Hsiao et al. 2017 [45]	19	CircCCDC66/Up	12	12	Tumor/normal tissues			0.88	Diagnostic, prognostic
Song et al. 2021 [46]	20	Hsa_circ_0001821/Up	102	80	Peripheral blood – > plasma			0.815 (CI 0.751–0.869)	Diagnostic
Li et al. 2020 [47]	21	Hsa_ circ_0001900, hsa_circ_0001178, and hsa_circ_0005927/Up	102	80	Tumor/normal tissues and peripheral blood samples	72.55	82.73	0.859 (CI 0.805–0.903)	Diagnostic
Lin et al. 2019 [48]	22	Circ-CCDC66, circ-ABCC1 and circ-STIL/Down	75	107	Peripheral blood – > plasma	64.4	85.2	0.780	Diagnostic, predictive
Yuan Liu et al. 2022 [49]	23	Hsa_circRNA_100833, hsa_circRNA_103828, hsa_circRNA_103831/Up and hsa_circRNA_103752, hsa_circRNA_071106, hsa_circRNA_102293/Down	33	33	Tumor/normal tissues			0.6860, 0.8127, 0.7502, 0.9945, 0.9642, 0.9486	Diagnostic, predictive
Xie et al. 2020 [50]	24	CircPNN (hsa_circ_0101802)/Up	Training set 88… Validation set 58	Training set 88… Validation set 58	Peripheral blood – > serum	Training set 89.8 … Validation set 89.7	Training set 73.9… Validation set 69.0	Training set 0.855 (CI 0.794–0.904) … Validation set 0.826 (CI 0.745–0.890)	Diagnostic
Mohammadi et al. 2020 [51]	25	Hsa_circ_000425/Down	100	100 + 33	Tumor/normal tissues	80.3	76.6	0.839 (CI 0.789–0.889)	Diagnostic, prognostic
Li et al. 2018 [52]	26	Hsa_circ_0000711/Down	101	101	Tumor/normal tissues	91	58	0.81	Diagnostic, prognostic
Zhuo et al. 2017 [53]	27	CircRNA0003906/Down	122	122 + 40	Tumor/normal tissues	80.3	72.5	0.818 (CI 0.749–0.888)	Diagnostic, monitoring
Alkhizzi et al. 2021 [54]	28	CircMETTL3, circUSP3/Up	42	32	Peripheral blood – > plasma	45.2, 26.2	93.8, 96.9	0.6946, 0.6280	Diagnostic
Ye et al. 2019 [55]	29	Hsa_circ_0000370/Up Hsa_circ_0082182/Up Hsa_circ_0035445/Down	156	66	Plasma			0.8152 (CI 0.7647–0.8903). 0.7371 (CI 0.6807–0.8236), 0.7028 (CI 0.6344–0.8013)………mixture of these three circRNAs = 0.8347	Diagnostic
Sadeghi et al. 2020 [56]	30	Hsa_circ_0060927/ Up	83	83	Tissues	68	83	0.78	Diagnostic
Duan et al. 2022 [57]	31	Hsa_circ_0001944/Up	133	133	Tissues	87.97	53.38	0.732	Diagnostic, prognostic
Lai et al. 2021 [58]	32	Circ 0000317/Down	45	45	Tissues				Diagnostic
Xiao et al. 2020 [59]	33	Hsa_circ_022382 (circFADS2)… Up	200	200	Tissue	73.6	77.5	0.803	Diagnostic, prognostic
Peng et al. 2021 [60]	34	Circ-GALNT16 (circBase ID: hsa_circ_0102495)/Down	100	100	Tissue				Diagnostic, prognostic
Jiang et al. 2021 [61]	35	CircIL4R/Up	120	40	Tumor/normal tissues and peripheral blood samples	60	87.50	0.718 (CI 0.599–0.837)	Diagnostic, prognostic
Chi et al. 2021 [62]	36	CircNSUN2… Up	32		Tissue				Diagnostic
Yang et al. 2020 [63]	37	CircPTK2 (hsa_circ_0005273) …Up	131	131	Tissue, peripheral blood (serum)			Serum = 0.05295, tissue = 0.7865	Diagnostic, prognostic, predictive
Wang et al. 2021 [64]	38	CircDUSP16 …Up	46	46	Tissue				Diagnostic, prognostic, predictive
Tian et al. 2019 [65]	39	Hsa_circ_0004585 …Up	Tissue 50, peripheral blood 142	Tissue 50, peripheral blood 142	Tissue, peripheral blood (plasma)	Tissue 0.851 Peripheral blood 0.908	Tissue 0.511 Peripheral blood 0.408	Tissue 0.731 Peripheral samples 0.707	Diagnostic
Liu et al. 2021 [66]	40	Circ_0084927 …Up	30	30	Tissue			0.806 (CI 0.683–0.896)	Diagnostic, prognostic
Yu et al. 2020 [67]	41	CircRUNX1 …Up	31	31	Tissue				Diagnostic, prognostic, predictive
Tang et al. 2020 [68]	42	CircMBOAT2 … Up	169/107	169/100	Tissue, peripheral blood (plasma)			0.750 (CI 0.694–0.805)	Diagnostic, prognostic, predictive
Wang et al. 2020 [69]	43	Circ_0060745 …Up	28/60	28/60	Tissues and paraffin-embedded tissue			0.8442 (CI 0.7737–0.9147)	Diagnostic, predictive
Ruan et al. 2019 [70]	44	Circ_0002138/Down	52	37	Tissue	62.86	74.29	0.7249 (CI 0.6065—0.8433)	Diagnostic, predictive
Chen et al. 2020 [69]	45	CircHUWE1 …. Up	58	58	Tissue			0.732	Diagnostic, predictive
Li et al. 2019 [71]	46	CircVAPA… Up	60	43	Tissue, plasma			0.724	Diagnostic, predictive
Wang et al. 2017 [72]	47	Hsa_circ_0000567 ……Down	102	102	Tissue	0.8333	0.7647	0.8653 (CI 0.8166–0.9141; P < 0.0001)	Diagnostic
Zhang et al. 2022 [73]	48	CircRNA_0006401 ……Up	12	12	Tissue			0.770 (CI 0.549–0.991, P = 0.041)	Diagnostic, predictive
Qi et al. 2022 [74]	49	Three circRNA (hsa_circ_001978, hsa_circ_105039, and hsa_circ_103627) …Up	100	100	Plasma			Training set (n = 20) = 1.000 Validation set (n = 80) = 0.969	Diagnostic
AmeliMojarad et al. 2022 [75]	50	CircADAM9 …Up	60	60	Tissue	0.59	0.55	0.77	Diagnostic
Guo et al. 2016 [76]	51	Hsa_circ_0000069…. Up	30	30	Tissue				Diagnostic, predictive
Zhu et al. 2019 [77]	52	CircRNAs (hsa_circ_0049487, hsa_circ_0066875, and hsa_circ_0007444)… Down	30	30				0.806	Diagnostic, prognostic
Wang et al. 2015 [78]	53	Circ_001988 …Down	31	31	Tissue	0.68	0.73	0.788 (95% CI 0.68–0.90, P < 0.001)	Diagnostic, predictive
Huang et al. 2021 [79]	54	CircMYLK … Up	90t…40p	90t…40p	Tissue, plasma			0.863	Diagnostic, predictive
Mai et al. 2021 [80]	55	Circ_PVT1 and circ_001569 … Up	148	148	Plasma				Diagnostic, prognostic

Empty cells indicate that the case is not mentioned in the study

According to the NOS criteria, out of 55 studies, 52 studies had score 7 or more and 3 studies had score 5–6. None of them were included in the meta-analysis stage because of the lack of at least three articles for the same circRNAs (Additional file 1 appendix).

Additional file 1 in the appendix also summarizes the number of people, age, gender in the CRC and non-CRC patient groups and the clinical characteristics of CRC patients. This table shows the relationship between the mentioned cases and the diagnostic role of the examined circRNAs in the diagnosis of CRC.

Then we investigated the effect of studied circRNAs (Table 1) and their related pathways in previous studies, which can be seen in Fig. 3.

Fig. 3.

The effect of circRNAs and their associated pathways in previous studies and expression changes

Chen et al. observed that hsa_circ_0084927 increases VEGFA in CRC patients with a sponge effect on miR-106b-5p. Then KEGG enrichment analyzes in competing endogenous RNAs (ceRNA) networks showed that target genes are involved in Golgi vesicle transport, organization of membrane system and cytosolic transport [27]. Zheng et al. showed that exosomal hsa_circ_0087960 (circLPAR1) suppresses CRC tumorigenesis by inhibiting the expression of BRD4 in interaction with eIF3h [29]. Zhang et al. found that the overexpression of hsa_circ_0000615 (circZNF609) as a tumor suppressor increases the expression of p53 protein in cells [30]. Another study using function analysis and signal enrichment based on DEGs showed that circ_001659 was significantly correlated with genes involved in EMT, metastasis, and cytoskeletal changes [32]. Liang et al. demonstrated that hsa_circ_0026416 may act as a ceRNA through competitive recruitment of miR-346 to upregulate NFIB expression and oncogenic effect [35]. Another study, through KEGG analysis, showed that circ_0000375 and circ_0011536, through their effect on miR-1182 and miR-1246, respectively, caused expression changes in genes in some pathways related to tumorigenesis, including signaling pathways regulating stem cell pluripotency, RIG-I receptor signaling pathway, Hippo and Wnt signaling pathway [37]. Zhang et al. found that hsa_circRNA_103809 has putative miRNA binding sites including miR-511-5p, miR-130b-5p, miR-642a-5p, miR-532-3p, and miR-329-5p and hsa_circRNA_104700 has putative miRNA binding sites including miR-141-5p, miR-500a-5p, miR-509-3p, miR-619-3p, and miR-578 [38]. The reduction of hsa_circ_001787 in CRC patients causes the disappearance of the sponge effect on miR-1204, miR-450b-3p, miR-1270 and miR-769-3p and through the activation of FGF13, KCNB1, RNF217, and SARGAP2 is involved in the activation of MAP Kinases and the activation family of AP-1 and other pathways [41]. Barbagallo et al. showed that LncRNA UCA1 is involved in cell migration through its effect on downstream pathways and ANLN, BIRC5, IPO7, KIF2A, and KIF23 genes [43]. Another study showed that circON (hsa_circ_0101802) is involved in the thyroid hormone signaling pathway, serotonergic synapse, NFkB signaling pathway, chemokine signaling pathway through the sponge of 5 different microRNAs [50]. Circ 0000317 has an inhibitory role by targeting miR-520g in cancer-related pathways, including TNF, VEGF, mTOR signaling pathways [58]. Jiang et al. reported that TFAP2C-induced circIL4R can act as a sponge for miR-761 and increase the expression of TRIM29. TRIM29 overexpression then induces ubiquitin-mediated degradation of PHLPP1 to stimulate the activation of PI3K/AKT signaling pathway and promote CRC cell proliferation and metastasis [61]. Significant increase in circ_0084927 in CRC patients compared to controls. Also, this value is higher in CRC patients in advanced stages than in early stages. This circRNA by affecting miRNA-20b-3p and then glutathione S-transferasemu 5 (GSTM5) potentially regulates the migration and invasion of CRC cells through the AKT/mTOR pathway [81]. It was also reported that circHUWE1 affects miR-486 and then this miR causes CRC to progress by targeting PLAGL2 and subsequently inhibiting PLAGL2/IGF2/β-catenin signal pathways [69].

Discussion

Considering that CRC is the second cause of cancer-related deaths in the world and the current methods of diagnosis are facing challenges, it is imperative to seek a non-invasive biomarker with acceptable sensitivity and specificity.

Some patients are excluded from biopsy due to coagulation defects or other underlying medical issues [82]. On the other hand, for imaging, the size of the tumor must reach the diagnostic limit [83]. Besides being efficient, a biomarker should be economical and capable of detecting cancer in its early stages, thereby reducing the financial burden of advanced stages of the disease and its recurrence [84].

circRNAs are a type of lncRNAs without 5 and 3 ends, which have a circular structure, so they are abundant, stable and resistant to exonucleolytic degradation by RNAse [85] In such a way that, the half-life of circRNAs is more than 48 h, while mRNA has a half-life of about 10 h [86].

The presence of circRNAs has been reported in various body fluids such as blood and saliva, and due to altered circRNA expression in numerous cancers; its conserved sequence and stable structure, circRNAs has promising potential to be considered as non-invasive cancer biomarkers.

Considering that previous reviews on circRNAs in CRC are small in size and lack a comprehensive systematic review, we conducted a systematic study of the role of circRNAs as diagnostic markers in CRC. The list of included articles in our study are shown in Table 1.

According to recent studies, it has been found that the expression profile of circRNAs are significantly different in cancerous tissues and their adjacent normal tissues. Up-regulated circRNAs act as oncogenes and those with down-regulated expression are considered as tumor suppressor genes.

Many pathological processes of cancer are regulated by the altered expression of circRNAs through the regulation of different signaling pathways. Among the circRNAs investigated in this study, 47 circRNAs increased and 28 circRNAs showed decreased expression, which indicates that the aberrant expression of circRNAs are involved in the promotion of CRC and may associated with its clinicopathological features.

Investigating the altered expression of circRNAs in the hepatocellular carcinoma and lung cancer has shown that higher levels of circRNAs are associated with a worse prognosis of the disease [87, 88].

In the last decade, it was found that circRNAs are present in secreted extracellular vesicles and are enriched and stable in exosomes, so that circRNAs are more abundant in exosomes than cell lines [89, 90]. Exosomes are nano scale bioactive extracellular vesicles secreted by cells into biofluids, containing proteins, nucleic acids, lipids and so on [91, 92]. Li et al. showed that circRNAs enter the circulation and can be measured in serum, and exosomal circRNAs can be used to distinguish between control samples and patients with colon cancer [90].

There were only a few exosome-derived circRNAs in the present study. CircLPAR1 was found to be packaged within exosomes with a notable level of stability and detectability. A significant reduction in its expression within plasma exosomes was observed during the progression of CRC, followed by a subsequent recovery post-surgery.

The levels of circLPAR1 were diminished in CRC tissues, showing a correlation with the overall survival of patients. The mechanism involved the uptake of exosomal circLPAR1 by CRC cells, leading to the inhibition of tumor growth. This effect could be attributed to the direct binding of exosomal circLPAR1 with eIF3h, thereby specifically hindering the interaction between METTL3 and eIF3h, consequently reducing the translation of the oncogene BRD4 [29].

In other studies, the exosomal circRNAs including circ-PNN and hsa-circ-0004771 were investigated and introduced as potential biomarkers in CRC diagnosis [39, 50]. All three of these exosomal circRNAs were derived from body fluids such as serum and plasma.

In a study by Han et al., it was found that CircLONP2 is involved in enhancing cancer invasion and metastasis through exosome. CircLONP2 functions as a crucial initiator of metastasis in CRC progression by controlling the intracellular maturation and intercellular transportation of miR-17, leading to the spread of metastasis-initiating capabilities at the primary site and the hastening of metastasis development in distant organs.

Mechanically, circLONP2 engages in a direct interaction and facilitates the processing of primary microRNA-17 (pri-miR-17) by recruiting DiGeorge syndrome critical region gene 8 (DGCR8) and the Drosha complex in a DDX1-dependent mode. Upregulated miR-17-5p can be packaged into exosomes and taken up by neighboring cells, thereby augmenting their aggressiveness [93]. Therefore, it seems that due to the efficiency of exosomes, their role in cell to cell communication, ease of access and less invasiveness of collecting body fluids, this type of samples should be given more attention in future studies.

Various functions are ascribed for circRNAs. They can act as sponges of miRNAs in ceRNA network [94], interact With RNA binding proteins as protein sponges [95, 96], translate to small peptide [97, 98], act as dynamic protein scaffolds [99] and Recruit Proteins such as chromatin modifiers [100].

Most of the circRNAs involved in this study played the role of miRNA Sponge. These circRNAs act as ceRNAs, forming circRNA-miRNA-mRNA axis, moderate gene expression as a post-transcriptional regulator, which summerized in Fig. 3.

Only one small peptides derived from circRNAs was reported in these studies. Zheng et al. showed that Hsa_circ_0006401 encodes a small peptide (198-aa peptide) that regulate COL6A3 mRNA expression and promotes CRC growth and migration [73].

And finally, Chaofan Peng reported circ-GALNT16 recruits the KH3 domain of hnRNPK, which prevents SENP2-mediated deSUMOylation of hnRNPK [60].

Despite our efforts to conduct a comprehensive research; there were some limitations in our study. Due to the lack of access to full text of some article, lack of control sample and statistical data we had to exclude a number of studies and despite requests from the authors, we could not access the details of the data.

We only included articles in English, which may be a language bias. On the other hand, As mentioned above most of investigated article and included patients in this study was from Asia, predominantly china, it means there may be population bias; so more studies with diversity in genetic, ethnic and geographic background are needed in other regions, to obtain more practical results that can be generalized to the entire world population.

Most of the reports based on the altered expression of circRNAs have a small sample size, and each of the circRNAs has been investigated in only one study. In order to avoid the small-study effect and to further validating the findings, it is necessary to conduct more extensive studies with a larger sample size.

In this research, most of the authors used tissue samples for their study, some used serum and plasma samples, and some performed their study with a combination of tissue and blood samples. Therefore, the non-uniformity of the types of samples may cause heterogeneity in the studies and affect their findings.

Overall, with all the advantages that circRNAs have as biomarkers, there are also some limitations. Initially, there should be a standard circRNA nomenclature system. Currently, the same circRNA is named variously, which may be confusing for researchers. On the other hand, to confirm that a circRNA is CRC specific, it is necessary to investigate the expression of circRNAs in other cancers. Furthermore, to have more effective evaluated studies, it is recommended to investigate the signaling pathways and the function of circRNAs and its clinical significance. Along with developing high throughputs methods in detection of miRNA and circRNAs, these methods should be upgraded in an easy, precise, fast and low-cost way to use in clinic. By overcoming these challenges, we can hope that circRNAs will be used as a promising biomarker in the early detection of cancers.

Conclusion

circRNAs have been studied due to their potential in progressive and recurrent disease monitoring, chemotherapy progress following, vaccine and drug development, and application in personalized medicine.in addition, several clinical trials and studies are currently underway to evaluate the use of circular RNAs as cancer biomarkers. Research is being conducted to explore their potential in various aspects, including early diagnosis, prognosis, and therapeutic monitoring across different types of cancers.

Due to the increase in cancer related deaths and the necessity of detecting patients in early stages we conducted a systematic analysis of the role of circRNAs as diagnostic markers in CRC. As we have deduced from the results of this study, with all the advantages that circRNAs have as biomarkers, there are also some limitations.

By standardizing detection protocols, it is possible to prevent the variability of results in different studies. On the other hand, factors such as age, sex, genetics, lifestyle, environment can cause changes in the expression level of circRNAs and cause biological variability that can overshadow its biomarker potential. Also, the selected circRNAs should have a good sensitivity and specificity so that it can be considered as a marker of a specific disease or a specific type of cancer. In addition, the function of many circRNAs has not yet been accurately identified. On the other hand, there is a lack of a comprehensive database on circRNAs, and studies should be done on a larger scale in different populations in order to achieve a universal biomarker. In addition to overcoming these vallidation challenges to achieve a reliable biomarker, for the clinical use of circRNAs as biomarkers, circuses-based tests should be cost-effective and ease-of-use, and ethical considerations should also be considered.

Author contributions

T.L.P. Initial writing and preparation of tables, A.Kh. Initial writing and preparation of figures and tables, A.D.I. Primary text reading and revision, S.A. Conceptualization, study of the initial text and revision, A.MA. Conceptualization, study of the final text and final revision, All authors have read and approved the manuscript.

Funding

Not applicable.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors of this article declare no competing interests that might influence the results and/or discussion reported in this article.