Development of a two-circular RNA panel as potential prognostic biomarker for gastric cancer

doi:10.21203/rs.3.rs-457318/v1

Background

Circular RNAs (circRNAs) have attracted increasing attention in recent years for their potential application as disease biomarkers due to their high abundance and stability. In this study, we attempted to screen circRNAs that can be used to predict postoperative recurrence and survival in patients with gastric cancer (GC).

Methods

High-throughput RNA sequencing was used to identify differentially expressed circRNAs in GC patients with different prognoses. The expression level of circRNAs in the training set (n = 136) and validation set (n = 167) was detected by quantitative real-time PCR (qRT-PCR). Kaplan–Meier estimator, receiver operating characteristic (ROC) curve and cox regression analysis were used to evaluate the prognostic value of circRNAs on recurrence-free survival (RFS) and overall survival (OS) in GC patients. CeRNA network prediction, gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed for the circRNAs with prognostic significance.

Results

A total of 259 differentially expressed circRNAs were identified in GC patients with different RFS. We found two circRNAs (hsa_circ_0005092 and hsa_circ_0002647) that highly expressed in GC patients with good prognoses, and subsequently established a predictive model for postoperative recurrence and prognosis evaluation, named circPanel. Patients with circPanel^low might have shorter recurrence-free survival (RFS) and overall survival (OS). We also performed circRNA-miRNA-mRNA network prediction and functional analysis for hsa_circ_0005092 and hsa_circ_0002647.

Conclusions

CircPanel has the potential to be a prognostic biomarker in GC patients with greater accuracy than a single circRNA and certain traditional tumor markers (e.g., CEA, CA19-9 and CA724).

Translational Medicine

Immunology

Cancer Biology

Gastrointestinal Surgery

gastric cancer

circular RNA

biomarker

prognostic assessment

circPanel

Gastric cancer (GC) is the fifth most common cancer and the third leading cause of cancer deaths worldwide (1). Due to GC’s high heterogeneity and the absence of specific symptoms, GC patients are often found at an advanced stage. Surgical resection, combined with adjuvant radiotherapy and chemotherapy, remains the major treatment strategy for GC. Nevertheless, advanced GC patients are still at risk for postoperative recurrence, often accompanied by rapid disease progression and a poor survival rate (2). Therefore, it is particularly necessary to stratify the postoperative recurrence risk of GC patients and adjust the treatment strategy accordingly. The current prognostic assessment of GC in the clinic mainly depends on pathological classification and TNM staging (3). Some traditional tumor biomarkers, including carcinoembryonic antigen (CEA), carbohydrate antigen 19 − 9 (CA19-9) and carbohydrate antigen 724 (CA724), have shown limited accuracy in prognostic evaluations (4, 5). Therefore, developing novel and effective biomarkers for the prognostic assessment of GC patients has always been the goal of the majority of researchers.

Circular RNA (circRNA) is a new type of RNA molecule with a covalently closed structure. Since their discovery in 1976 (6), the formation mechanism, expression characteristics and biological functions of circRNAs have been gradually revealed. Unlike linear mRNAs, circRNAs are mostly derived from black-splicing of precursor mRNAs, and they are widely expressed in mammals with conserved sequences (7). The circular structure of circRNAs causes them to lack 5' and 3' polarity, preventing them from being degraded by exonuclease and resulting in a longer half-life. Moreover, the expression of circRNAs is generally tissue-, cell- and developmental stage-specific. Therefore, circRNAs have been considered as potential disease biomarkers (8–10). In addition, circRNAs have been found to regulate gene expression at the transcriptional, posttranscriptional and translational levels through a variety of pathways. They can competitively bind microRNAs (miRNAs) to further inhibit their regulation of downstream target genes, and can be used as protein scaffolds to participate in the activation or segregation of proteins (11). A few circRNAs also have the potential to be translated into polypeptides or proteins(12). Abnormal expression of circRNAs has been found in a variety of cancers (13), including lung cancer, liver cancer, breast cancer, colorectal cancer and gastric cancer.

Considering the correlation between circRNAs and gastric cancer, we speculated that circRNAs could serve as biomarkers for prognostic evaluation of GC. In this study, we revealed the circRNA expression profiles of GC patients with different prognoses and established a postoperative recurrence model (circPanel), based on hsa_circ_0005092 and hsa_circ_0002647 to evaluate the prognoses of GC patients.

Patients and samples

A total of 313 samples from GC patients were collected from January 1, 2010, to December 31, 2015, at Tianjin Medical University Cancer Institute and Hospital for discovery (n = 10), training (n = 136) and validation (n = 167). All of the patients met the following criteria: (1) pathologically diagnosed with gastric adenocarcinoma; (2) without preoperative radiotherapy or chemotherapy; (3) without a history of other cancers; and (4) without severe postoperative complications (including infection, bleeding and anastomotic fistula). Tumors were staged according to the 8th edition of the American Joint Committee on Cancer tumor-node-metastasis (TNM) staging system. The detailed clinicopathological characteristics of the training and validation sets are summarized in Appendix Table S1.

RNA sequencing and bioinformatics analysis

Tumor tissue-derived RNAs from 5 paired gastric cancer patients with different prognoses (RFS ≤ 1 year vs. RFS ≥ 3 years) were used for circRNA sequencing through the Illmina HiSeq ™2500 platform. The original image data obtained from sequencing were translated into sequenced reads through base calling analysis. The sequencing error rate of each sample was less than 0.1%. circRNAs were identified by the find_circ_ tool (14). The circRNAs currently known were named with circBase ID, while novel circRNAs were annotated as hg19_circ_xxxxxxx with chromosome position and splicing sequences. The expression of circRNAs in each sample was counted and normalized by the transcripts per million (TPM). Volcano plots and heat maps were used to describe the distribution of differentially expressed circRNAs (padj < 0.05) between patients with different prognoses.

Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) assay

Total RNA was extracted from GC tissues with TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. RNA integrity was verified by agarose gel electrophoresis, and the concentration and purity were measured with a NanoDrop spectrophotometer. A total of 500ng of total RNA was reverse-transcribed with random primers using the PrimeScritTM RT reagent kit (Takara, Dalian, Liaoning, China). The obtained cDNA was analyzed by qRT-PCR with TB Green Premix Ex Taq II (Takara) on the QuantStudio 5 Real-Time PCR System (Thermo Fisher Scientific). The PCR conditions were as follows: 95°C for 30 s, 40 cycles at 95°C for 5 s and 62°C for 34 s. Melting curves were generated at the end of amplification. β-actin was used as an internal reference gene, and the relative expression level of circRNAs was calculated by the -ΔCT method. Divergent primers for circRNAs were designed for their back-splicing sites and synthesized by Sangon (Shanghai, China). Primer sequences were as follows: hsa_circ_0005092 forward 5´-GGCCAGATGAAGAAGGTAGTGAT-3´, hsa_circ_0005092 reverse 5´-ACAGGTCTGATGAATGGTGTG‐3´, hsa_circ_0002647 forward 5´‐TGACCTGAGACACCTATGGC‐3´, reverse 5´‐TAGTGTGTTGGTGCCATCCT‐3´. Other primers are shown in Appendix Table S2. The accuracy of the amplified product was confirmed by agarose gel electrophoresis and Sanger sequencing.

Actinomycin D treatment

GC cells (5x10⁴, SNU-1) were seeded in 24-well plates and cultured with 5% CO₂ at 37°C. After 24h, actinomycin D was added to the wells at a final concentration of 2ug/mL for 0h, 3h, 6h, 9h, and 12h. The cells at each time point were harvested and extracted for RNA. qRT-PCR was used to detect the relative expression of circRNAs and their host genes. Primer sequences were as follows: IPO7 (the host gene of hsa_circ_0005092) forward 5´-ATCGAGAAACAGCACCAGGG-3´, reverse 5´-CTACCAATGGACTCCGCTCC-3´; EHMT1 (the host gene of hsa_circ_0002647) forward 5´-GCTGTGTGAAAACCGAGCTG-3´, reverse 5´-TCCGCTATCCGAGTTAGTGTG-3´.

CeRNA network analysis and function annotation

The potential interactions between circRNAs and miRNAs were predicted by the CircInteractome (https://circinteractome.nia.nih.gov/index.html) website. For each circRNA, 6 miRNAs with more binding sites and lower context scores were selected for further analysis. Target mRNAs of these miRNAs were predicted on the TargetScan (http://www.targetscan.org/vert_72/) and miRDB (http://mirdb.org/) websites and intersected using Venn diagrams. Cytoscape was used to delineate the circRNA-miRNA-mRNA network. To further understand the potential functions of circRNAs, gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed on the ceRNA regulatory network of circRNAs via the DAVID website (https://david.ncifcrf.gov/).

Statistical analysis

All of the statistical data were analyzed using SPSS software, version 21.0 (SPSS, Chicago, IL, USA), and GraphPad software, verson 7.0 (GraphPad Software, San Diego, CA, USA). Kaplan–Meier estimator was performed to evaluate the prognostic value of circRNAs for RFS and OS in GC patients. Cox regression analysis was used to screen the prognostic risk factors for GC and establish a risk assessment model for postoperative recurrence. Receiver operating characteristic (ROC) curves were generated, and the area under the ROC curve (AUC) was measured to reflect the accuracy of the model with the Youden index (specificity + sensitivity-1). The chi-square test was used to analyze the relationships between circRNA levels and clinicopathological factors in GC patients. p values < 0.05 were considered to be significant.

circRNA profiles in GC patients with different prognoses

The flowchart for circRNA screening in this study is shown in Fig. 1a. We performed RNA sequencing to characterize the circRNA expression profiles of GC patients with different RFS. A total of 23,995 circRNAs were detected in the discovery set, among which 9,443 circRNAs were included in the circBase database and annotated (Fig. 1b). Consistent with previous studies, circRNAs were distributed in all human chromosomes, and their splicing length was mostly less than 1,000 nucleotides (Fig. 1c-d). Differential analysis was performed on the circRNA expression profile of GC patients with different prognoses. A total of 259 differentially expressed circRNAs were identified with the threshold of padj < 0.05, of which 192 circRNAs were upregulated, and 67 circRNAs were downregulated in GC patients with good prognoses (Fig. 1e).

Screening and validation of prognostic circRNAs

Then, we evaluated whether circRNAs could be used as prognostic biomarkers in GC patients. The top 20 up- and downregulated circRNAs with extremely significant differences obtained by RNA-seq were included in the preliminary screening (Fig. 2a). Considering the contingency of RNA sequencing samples, 12 circRNAs with significant expression differences in at least 4 of 5 pairs of samples were selected for further validation. The circRNA ID, genome position, spliced length, and gene symbols of these 12 candidate circRNAs are shown in Table 1.

Table 1

The information of 12 candidate circRNAs
circRNA ID	Genome position	Strand	Spliced length	Gene name	Log2FC	P Value	Regulation
Hsa_circ_0005729	chr18:29691716–29693823	+	282	RNF138	5.974	9.58E-05	Up
Hsa_circ_0005092	chr11:9451220–9452550	+	290	IPO7	4.9603	0.002134	Up
Hsa_circ_0001965	chr3:169854206–169863309	-	346	PHC3	4.7984	0.003229	Up
Hsa_circ_0002647	chr9:140611077–140646860	+	1163	EHMT1	4.7371	0.003767	Up
Hsa_circ_0063604	chr22:42204878–42206004	+	241	CCDC134	4.6347	0.00505	Up
Hsa_circ_0105599	chr16:53967896–53971745	+	3849	FTO	4.5821	0.005785	Up
Hsa_circ_0008197	chr1:51032748–51061888	-	524	FAF1	4.5784	0.005578	Up
Hsa_circ_0099642	chr12:9833518–9847565	+	692	CLEC2D	-4.8723	0.001811	Down
Hsa_circ_0006847	chr16:29916172–29917280	+	286	ASPHD1	-4.8243	0.002375	Down
Hsa_circ_0072697	chr5:64863339–64868113	+	773	PPWD1	-4.6578	0.003623	Down
Hsa_circ_0015491	chr1:180047597–180049796	+	357	CEP350	-4.4914	0.005258	Down
Hsa_circ_0008403	chr5:141732790–141733148	+	358	TCONS_l2_00023557	-4.268	0.008983	Down

FC, fold change.

The relative expression levels of 12 circRNAs were detected by qRT-PCR assay in the training set (n = 136). The median circRNA expression level was used as the cutoff to divide GC patients into two groups with high and low expression. Kaplan-Meier survival analysis showed that four circRNAs (hsa_circ_0005092, hsa_circ_0002647, hsa_circ_0008197 and hsa_circ_0105599) were significantly associated with RFS in GC patients and were upregulated in patients with good prognoses, consistent with the sequencing results (Fig. 2b-e, Appendix Figure S1). Among them, hsa_circ_0005092 and hsa_circ_0002647 were confirmed to be independent prognostic factors of RFS in GC patients by univariate and multivariate Cox regression analysis (Table 2).

Table 2

Univariate and multivariate Cox regression analysis of prognostic factors for RFS in the training set
Factors	Univariate analysis		Multivariate analysis
Factors	HR (95% CI)	P value	HR (95% CI)	P value
Age	0.997 (0.976–1.018)	0.774	1.010 (0.987–1.034)	0.405
Sex	1.221 (0.730–2.043)	0.446	1.095 (0.637–1.885)	0.742
Tumor size	1.024 (0.940–1.114)	0.589	0.998 (0.901–1.105)	0.967
Hsa_circ_0005092	1.531 (1.219–1.922)	0.000***	1.784 (1.256–2.534)	0.001**
Hsa_circ_0008197	1.121 (0.942–1.334)	0.198	0.960 (0.811–1.137)	0.637
Hsa_circ_0002647	1.345 (1.139–1.589)	0.001**	1.294 (1.072–1.563)	0.007**
Hsa_circ_0105599	1.176 (1.003–1.378)	0.045*	0.872 (0.706–1.078)	0.206
RFS, recurrence-free survival; HR, Hazard ratio; CI, confidence interval; p༜0.05, p༜0.01, **p༜0.001.

Construction of prognostic model based on hsa_circ_0005092 and hsa_circ_0002647

Next, the training set and validation set were combined to construct a postoperative recurrence model based on the regression coefficients of hsa_circ_0005092 and hsa_circ_0002647 for statistical significance (Appendix Table S3):

A total of 303 gastric cancer patients were included in the statistics, and were divided into high- and low-risk groups according to the median. Survival analysis showed that patients with circPanel^low had a shorter RFS than those with circPanel^high (hazard ratio [HR]: 2.229, 95% confidence interval [CI]: 1.662-2.989, p<0.0001, Figure 4a). The ROC curve and the area under the ROC curve (AUC) were further used to analyze the prognostic value of circPanel, as shown in Figure 4c. Compared with the hsa_circ_0005092 and hsa_circ_0002647, circPanel had a larger AUC (0.709, 95% CI: 0.607-0.742), with sensitivity of 69.9% and specificity of 58.1%.

Moreover, we observed a similar effect of circPanel on OS in GC patients. Patients with circPanel^low had a shorter OS (HR: 2.025, 95% CI: 1.362-3.010, p=0.0004, Figure 4b). As shown in Figure 4d, the AUCs of hsa_circ_0002647, hsa_circ_0005092 and circPanel were 0.612 (95% CI: 0.533-0.691), 0.631 (95% CI: 0.553-0.709) and 0.638 (95% CI: 0.56-0.716), respectively. The sensitivity and specificity of circPanel for predicting OS were 51.0% and 75.5%, respectively. In short, circPanel could be used as a biomarker for the prognostic evaluation of GC patients, with better predictive performance than single circRNA.

Comparative analysis of circRNA and other clinical indicators

We also analyzed the correlation between circPanel and the clinicopathological features of GC patients. The results suggested that circPanel was related to the tumor differentiation and T stage in GC patients (Table 4). Patients with circPanel^high might have higher tumor differentiation and lower T stages. No significant association was observed between circPanel and other characteristics (including sex, age, tumor size, N stage and TNM stage).

In addition, we compared the prognostic value of circPanel with some traditional tumor markers (including CEA, CA19-9 and CA724) by ROC curve analysis. As a result, the AUCs of circPanel, CEA, CA19-9 and CA724 were 0.67, 0.508, 0.56, and 0.493 for RFS, respectively, and 0.638, 0.485, 0.566, and 0.524 for OS. CircPanel has greater prognostic value than CEA, CA19-9 and CA724 (Figure 4e-f). In summary, these results strongly confirmed the clinical application value of circPanel in the prognostic assessment of GC patients.

Table 4. Correlation between circRNAs and clinicopathologic features of GC patients

			circPanel
Variables	Group	Cases	High expression	Low expression	P value
		n=303	n = 151	n = 152
Age	≤60 years	155	73	88	0.0958
	＞60 years	148	78	64
Sex	Male	222	111	111	0.9242
	Female	81	40	41
Tumor size	≤5 cm	143	80	73	0.3885
	＞5 cm	150	71	79
T stage	T1-3	41	26	15	0.0359*
	T4a	181	93	88
	T4b	82	32	49
N stage	N0	73	36	37	0.1369
	N1	73	41	32
	N2	76	42	34
	N3	81	32	49
TNM stage	I-II	75	40	35	0.1041
	IIIA	113	64	49
	IIIB	63	26	38
	IIIC	52	21	30
Differentiation	Poor-undifferentation	158	69	89	0.0251*
	Well-moderate differentation	145	82	63

*p＜0.05.

CeRNA network of hsa_circ_0005092 and hsa_circ_0002647

Furthermore, we attempted to explore the biological function of these two selected circRNAs. CeRNA regulation is currently considered one of the main mechanisms for circRNA functioning. Through the prediction of the CircInteractome, 6 miRNAs (miR-616，miR-599，miR-409-3p，miR-217，miR-513a-5p and miR-890) targeted by hsa_circ_0005092 and 6 miRNAs (miR-370, miR-626, miR-637, miR-648, miR-326 and miR-574-5p) targeted by hsa_circ_0002647, with more binding sites and lower context scores, were screened. Then, we predicted the target mRNAs of these miRNAs by TargetScan (cumulative weighted context score ≤-0.3) and miRDB (target score ≥70) and obtained 259 and 434 mRNAs for hsa_circ_0005092 and hsa_circ_0002647, respectively. Cytoscape was further used to describe the regulatory network of circRNA-miRNA-mRNA (Figure 5a-b).

GO and KEGG analysis of hsa_circ_0005092 and hsa_circ_0002647

To predict the potential function of circRNAs, GO and KEGG enrichment analyses were performed on the ceRNA regulatory networks of these two circRNAs. The significantly enriched biological processes (BPs), cellular components (CCs) and molecular functions (MFs) are shown in Figure 6a-b. The main annotations of the hsa_circ_0005092 network included protein processing in the endoplasmic reticulum, protein binding and regulation of epidermal growth factor receptor signaling, while hsa_circ_0002647 was mainly involved in RNA transcriptional regulation. The pathways enriched for the two circRNAs based on KEGG analysis are shown in Figure 6c-d. The metabolic pathways, chemokine signaling pathways, and proteoglycans in cancer for hsa_circ_0005092 and the cell adhesion molecules, cAMP signaling pathway, and transcriptional misregulation in cancer for hsa_circ_0002647 are all cancer-related.

Although screening methods and treatment strategies have made great progress in recent years, the mortality rate of GC patients remains high. Radical surgical resection remains the only possible way to cure GC. However, unfortunately, postoperative recurrence of GC patients is very common. Generally, recurrence within 2 years after surgery is defined as early recurrence, and after more than 2 years, it is defined as late recurrence. In China, approximately 42.5% of GC patients experience early recurrence (15), indicating a poor prognosis. Postoperative recurrence assessment plays an important role in guiding clinical medication and improving prognosis in gastric cancer (16, 17). Currently, tumor size, invasion depth, pathological type, number of metastatic lymph nodes and some tumor markers (CEA, CA19-9 and CA72-4) are used as predictors of GC recurrence clinically. However, it is difficult to accurately predict early recurrence by relying on these indicators.

The development of molecular biology has provided more possibilities for disease assessment. circRNAs have become potential biomarkers due to their high stability, specificity and conservation (18). Research on circRNAs as prognostic markers for GC patients is gradually being performed. For example, Tang et al. found that the expression of circ-KIAA1244 in plasma was negatively correlated with TNM staging and lymph node metastasis, and GC patients with low circ-KIAA1244 expression often had shorter overall survival (19). Chen et al. also considered circPVT1 to be a novel proliferative factor and prognostic biomarker in GC (20).

In our study, we identified the differential expression profiles of circRNAs in GC patients with different prognoses by RNA high-throughput sequencing, and we confirmed that hsa_circ_0005092 and hsa_circ_0002647, which were downregulated in patients with shorter RFS, are independent prognostic factors in GC patients. Based on these two circRNAs, we constructed a postoperative recurrence model, named circPanel. CircPanel was correlated with the degree of tumor differentiation and T stage, which are important prognostic factors for GC. Via validation in 303 samples, circPanel was proved to be able to distinguish GC patients with different prognoses with greater accuracy than some tumor markers (CEA, CA19-9 and CA72-4). Patients with circPanel^low had shorter RFS and OS. Therefore, circPanel is an effective and promising biomarker for the assessment of postoperative recurrence and the prognoses of GC patients. Timely and effective intervention in patients at risk of early recurrence predicted by circPanel is expected to greatly improve the prognoses of patients.

However, our study also has some limitations. First, due to the difficulty in obtaining tissue samples, we were unable to monitor the circRNA levels during treatment dynamically, preventing further study. Second, the sample size in our study could not meet the requirements of grouping patients according to their treatment strategies for statistics, making it impossible to eliminate interference from the treatment strategies of patients. Despite these limitations, the clinical significance of circPanel in the prediction of postoperative recurrence and the prognoses of gastric cancer cannot be denied. Further multicenter prospective studies with larger sample sizes are needed to verify the effectiveness of circPanel and find the optimal cutoff values for clinical use.

Furthermore, we also performed an in-depth study of the two screened circRNAs. Hsa_circ_0005092 and hsa_circ_0002647 were derived from the back-splicing of importin 7 (IPO7) and euchromatic histone lysine methyltransferase 1 (EHMT1) respectively, with lengths of 290 bp and 1163 bp. Considering that miRNA sponges are among the main mechanisms for circRNAs (21), we predicted their ceRNA regulatory networks. The ceRNA network of hsa_circ_0005092 contains 6 miRNAs and 259 mRNAs. Functional analysis suggested that hsa_circ_0005092 might participate in the EGFR signaling pathway and in various metabolic processes associated with tumors. The ceRNA network of hsa_circ_0002647 includes 6 miRNAs and 434 mRNAs, some of which are involved in RNA transcriptional regulation. These two circRNAs could have a wide range of biological effects through the above mechanisms. In our study, both hsa_circ_0005092 and hsa_circ_0002647 were highly expressed in GC patients with longer RFS and OS and were correlated with tumor differentiation and T stage, suggesting that these two circRNAs might be negatively associated with the malignant behavior of GC. Generally, circRNA levels decline with cell division, leading to the possibility that cells with higher circRNA levels could have a lower proliferation rate. Considering the relationship between drug resistance and tumor recurrence (22, 23), we speculated that these two circRNAs might be involved in the regulation of cell proliferation, cell invasion and drug resistance. However, the detailed functions and mechanisms of hsa_circ_0005092 and hsa_circ_0002647 must still be further studied.

In conclusion, we described the circRNA expression profiles of GC and identified 2 upregulated circRNAs (hsa_circ_0005092 and hsa_circ_0002647) in patients with good prognoses. The circPanel, calculated based on these two circRNAs, was confirmed to be positively correlated with RFS and OS in GC patients. CircPanel could serve as a potential biomarker for prognostic evaluation in GC patients.

circRNA: circular RNA; RFS:recurrence-free survival; OS:overall survival; CEA:carcinoembryonic antigen; CA19-9:carbohydrate antigen 19 − 9; CA724:carbohydrate antigen 724; miRNA:microRNA; TPM:transcripts per million; qRT-PCR:quantitative reverse transcription-polymerase chain reaction; ceRNA:competing endogenous RNA; GO:gene ontology; KEGG:Kyoto Encyclopedia of Genes and Genomes; ROC:receiver operating characteristic; AUC:area under the ROC curve; BP:biological process; CC:cellular component; MF:molecular function; IPO7:importin 7; EHMT1:euchromatic histone lysine methyltransferase 1.

Ethics approval and consent to participate

All of the patients provided written informed consent, and this study was approved by the Ethics Committee of Tianjin Medical University Cancer Institute and Hospital (Ek2017002).

Consent for publication

Not applicable.

Availability of data and materials

All data generated or analysed during this study are included in this published article.

Competing interests

The authors declare that they have no competing interests.

Funding

This work was funded by the National Natural Science Foundation of China (81772620), National Key R&D Program of China (2018ZX09201015) and Tianjin Science and Technology Major Project of Chronic Diseases Prevention and Control (17ZXMFSY00130).

Authors' contributions

J. Liu and H. Li designed the research; J. Liu and M. Yan carried out data collection and experimental work; J. Liu performed data analysis and wrote the paper; H. Li contributed new reagents or analytic tools and revised the paper. All authors read and approved the final manuscript.

Acknowledgements

We thank the Biobank of Tianjin Medical University Cancer Institute and Hospital for their support and reviewers for their valuable advice.

Authors' information

¹Department of Gastrointestinal Cancer Biology, Tianjin Medical University Cancer Institute and Hospital, Tianjin, China. ²Key Laboratory of Cancer Immunology and Biotherapy, Tianjin, China. ³National Clinical Research Center for Cancer, Tianjin, China

Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: a cancer journal for clinicians. 2021.
Hartgrink HH, Jansen EP, van Grieken NC, van de Velde CJ. Gastric cancer. Lancet. 2009;374(9688):477–90.
Reim D, Loos M, Vogl F, Novotny A, Schuster T, Langer R, et al. Prognostic implications of the seventh edition of the international union against cancer classification for patients with gastric cancer: the Western experience of patients treated in a single-center European institution. J Clin Oncol. 2013;31(2):263–71.
Matsuoka T, Yashiro M. Biomarkers of gastric cancer: Current topics and future perspective. World journal of gastroenterology. 2018;24(26):2818–32.
Qiu FM, Yu JK, Chen YD, Jin QF, Sui MH, Huang J. Mining novel biomarkers for prognosis of gastric cancer with serum proteomics. Journal of experimental clinical cancer research: CR. 2009;28:126.
Sanger HL, Klotz G, Riesner D, Gross HJ, Kleinschmidt AK. Viroids are single-stranded covalently closed circular RNA molecules existing as highly base-paired rod-like structures. Proc Natl Acad Sci U S A. 1976;73(11):3852–6.
Chen LL. The biogenesis and emerging roles of circular RNAs. Nat Rev Mol Cell Biol. 2016;17(4):205–11.
Zhang H, Zhu L, Bai M, Liu Y, Zhan Y, Deng T, et al. Exosomal circRNA derived from gastric tumor promotes white adipose browning by targeting the miR-133/PRDM16 pathway. International journal of cancer. 2018.
Bahn JH, Zhang Q, Li F, Chan TM, Lin X, Kim Y, et al. The landscape of microRNA, Piwi-interacting RNA, and circular RNA in human saliva. Clinical chemistry. 2015;61(1):221–30.
Shao Y, Li J, Lu R, Li T, Yang Y, Xiao B, et al. Global circular RNA expression profile of human gastric cancer and its clinical significance. Cancer medicine. 2017;6(6):1173–80.
Li X, Yang L, Chen LL. The Biogenesis, Functions, and Challenges of Circular RNAs. Mol Cell. 2018;71(3):428–42.
Yang Y, Gao X, Zhang M, Yan S, Sun C, Xiao F, et al. Novel Role of FBXW7 Circular RNA in Repressing Glioma Tumorigenesis. Journal of the National Cancer Institute. 2018;110(3).
Kristensen LS, Hansen TB, Veno MT, Kjems J. Circular RNAs in cancer: opportunities and challenges in the field. Oncogene. 2018;37(5):555–65.
Memczak S, Jens M, Elefsinioti A, Torti F, Krueger J, Rybak A, et al. Circular RNAs are a large class of animal RNAs with regulatory potency. Nature. 2013;495(7441):333–8.
Liu D, Lu M, Li J, Yang Z, Feng Q, Zhou M, et al. The patterns and timing of recurrence after curative resection for gastric cancer in China. World J Surg Oncol. 2016;14(1):305.
Whiting J, Sano T, Saka M, Fukagawa T, Katai H, Sasako M. Follow-up of gastric cancer: a review. Gastric Cancer. 2006;9(2):74–81.
Jiao X, Wang Y, Wang F, Wang X. Recurrence pattern and its predictors for advanced gastric cancer after total gastrectomy. Medicine. 2020;99(51):e23795.
Arnaiz E, Sole C, Manterola L, Iparraguirre L, Otaegui D, Lawrie CH. CircRNAs and cancer: Biomarkers and master regulators. Seminars in cancer biology. 2018.
Tang W, Fu K, Sun H, Rong D, Wang H, Cao H. CircRNA microarray profiling identifies a novel circulating biomarker for detection of gastric cancer. Mol Cancer. 2018;17(1):137.
Chen J, Li Y, Zheng Q, Bao C, He J, Chen B, et al. Circular RNA profile identifies circPVT1 as a proliferative factor and prognostic marker in gastric cancer. Cancer letters. 2017;388:208–19.
Han B, Chao J, Yao H. Circular RNA and its mechanisms in disease: From the bench to the clinic. Pharmacol Ther. 2018;187:31–44.
Marusyk A, Janiszewska M, Polyak K. Intratumor Heterogeneity: The Rosetta Stone of Therapy Resistance. Cancer Cell. 2020;37(4):471–84.
Viale A, Draetta GF. Metabolic Features of Cancer Treatment Resistance. Recent Results Cancer Res. 2016;207:135–56.

AppendixFigureS1.tif
Selection of candidate circRNAs Kaplan-Meier survival analysis of hsa_circ_0006847 (a), hsa_circ_0008403 (b), hsa_circ_0015491 (c), hsa_circ_0072697 (d), hsa_circ_0063604 (e), hsa_circ_0001965 (f), hsa_circ_0099642 (g) and hsa_circ_0005729 (h) in the training set (n=136). Blue line: high expression; red line: low expression.
AppendixTable.xlsx

Development of a two-circular RNA panel as potential prognostic biomarker for gastric cancer

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Background

Methods

Statistical analysis

Results

Discussion

Conclusions

Abbreviations

Declarations

References

Supplementary Files

Status:

Version 1