Clinical significances of hsa_circ_0067582 and hsa_circ_0005758 in gastric cancer tissues

Abstract Background Circular RNAs (circRNAs) are a special class of endogenous noncoding RNAs that have numerous biological functions in normal situation and diseases including cancers. However, the clinical significance of circRNAs in gastric cancer (GC) remains largely unknown. Here, we chose two representative circRNAs, hsa_circ_0067582 and hsa_circ_0005758, to investigate their clinical significance in GC patients. Methods Using real‐time quantitative reverse transcription‐polymerase chain reaction (qRT‐PCR), we explored the expression levels of hsa_circ_0067582 and hsa_circ_0005758 in tissues with different stages of gastric tumorigenesis. Then, the relationships between their expression levels and GC patients' clinicopathological factors were further investigated. Receiver operating characteristic (ROC) curves were established for evaluating diagnostic values of hsa_circ_0067582 and hsa_circ_0005758. Results Compared with healthy control tissues, both hsa_circ_0067582 and hsa_circ_0005758 were significantly decreased in GC tissues. Besides, hsa_circ_0067582 expression was associated with GC patients' tissue CEA level (P <.001) and stages (P = .037); and hsa_circ_0005758 expression was relevant to tissue CEA level (P < .001) and perineural invasion (P = .048). The area under the ROC curve (AUC) of hsa_circ_0067582 was up to 0.671. The cutoff value was set at 10.61, with which the sensitivity and specificity were 55.2% and 75.0%, respectively. Similar to hsa_circ_0005758, the AUC of hsa_circ_0005758 was 0.721. The cutoff value was set at 10.20, with which the sensitivity and specificity were 75.0% and 67.7%, respectively. Conclusion These results showed that both hsa_circ_0067582 and hsa_circ_0005758 may play an important role in gastric carcinogenesis; and they may be potential indicators for GC diagnosis.

as tumor suppressors and oncogenes. 6 CircRNAs appear to be more often down-regulated in tumor tissues compared to normal tissues. 8 In our previous study, we analyzed the gene expression profiles of GC and paired normal tissue samples, and identified a number of genes that are significantly up-regulated or down-regulated in cancer tissues compared with their adjacent normal tissues. 2 Hsa_circ_0067582 and hsa_circ_0005758 are gastric cancer-associated circRNAs based on bioinformatics analysis. Hsa_circ_0067582, 394 nt in length, is transcripted from chr3:141231004-141259451. Its associated gene symbol is RASA2 (RAS p21 protein activator 2). Hsa_ circ_0005758 is a circRNA with 373 nt in length. Its gene is located at chr1:155891165-155893478, with associated gene symbol KIAA0907.
They are both among the most deregulated circRNAs in gastric cancer.
In this study, the expression of hsa_circ_0067582 and hsa_ circ_0005758 in GC tissue specimens and paired normal tissues were first explored. Then, their expression levels in tissues with different stages of gastric tumorigenesis were measured. The potential relationships between circRNAs' expression levels and patients' clinicopathological factors were further investigated to gain a better understanding of their biological roles in gastric cancer. Finally, ROC curves were established for evaluating their diagnostic values. Our results indicated that both hsa_ circ_0067582 and hsa_circ_0005758 may play an important role in gastric carcinogenesis, and they may be potential indicators for GC diagnosis.

| Total RNA preparation and qRT-PCR detection
Total RNA was extracted from specimens by TRIzol reagent for 30 seconds, and 72°C for 30 seconds. The cycle threshold (C t ) values were recorded for hsa_circ_0067582, hsa_circ_0005758, and GAPDH. The data were analyzed through the ΔC t method. All results were expressed as mean ± SD of three independent experiments. Larger ΔC t value indicates lower expression. The PCR products of both hsa_circ_0067582 and hsa_circ_0005758 were confirmed by sequencing ( Figure S1).

| Immunohistochemical analysis of tissue CEA and CA19-9
We incubated the paraffin tissue sections in primary anti-carcinoembryonic antigen (CEA) or anti-carbohydrate antigen 19-9 (CA19-9) (DAKO) for 1 hour at room temperature; then we incubated the tissues in diaminobenzidine (DAB; DAKO) for color development after incubation with broad spectrum second antibody K5007 (DAKO). The results were classified as negative or positive ( Figure S2).

| Statistical analysis
All statistical analyses were performed by Statistical Program for Social Sciences 20.0 software (SPSS), GraphPad Prism 5.0 (GraphPad Software), and SigmaPlot 10.0 (SigmaPlot Software). We used Student's t test, one-way analysis of variance (ANOVA) test, and rank-sum test as appropriate. Statistical significance was accepted at P < .05.

| Hsa_circ_0067582 and hsa_circ_0005758 were down-regulated in GC tissues
After the verification by qRT-PCR method, hsa_circ_0067582 and hsa_circ_0005758 were both found markedly down-regulated in GC tissues compared with the adjacent non-tumorous tissues ( Figure 1A,  Figure 2B). Furthermore, the expression of them were both lower in the early GC tissues than the matched adjacent tissues ( Figure 1C, Figure 2C). Compared with healthy control group, hsa_circ_0067582 and hsa_circ_0005758 expression levels were significantly decreased in GC tissues ( Figure 1D, Figure 2D).
Besides, hsa_circ_0005758 expression level was remarkably decreased in GIM tissues compared with gastritis group ( Figure 2D). There was no significant difference in the level of hsa_circ_0067582 between GIM group and gastritis group ( Figure 1D).

| Relationship between hsa_circ_0067582, hsa_ circ_0005758 levels, and clinicopathological factors
The correlations between hsa_circ_0067582, hsa_circ_0005758 expression, and clinicopathological features of GC patients were further analyzed. The results indicated that low expression of hsa_circ_0067582 was associated with GC patients' tissue CEA level (P < .001) and stages (P = .037). Although low expression of hsa_circ_0005758 was relevant to tissue CEA level (P < .001) and perineural invasion (P = .048) ( Table 1), both hsa_circ_0067582 and hsa_circ_0005758 expression level were not significantly correlated with age, gender, diameter, tissue CA19-9 level, and so on.

| ROC curve of hsa_circ_0067582 and hsa_ circ_0005758
Receiver operating characteristic (ROC) curves were generated to

| D ISCUSS I ON
CircRNAs, form a covalently closed continuous loop through connecting 5′ splice site to 3′ splice site, are previous regarded as F I G U R E 1 Hsa_circ_0067582 expression levels in gastric cancer tissues. A, The expression levels of hsa_ circ_0067582 in cancer tissues (n = 96) and adjacent normal tissues (n = 96). B, The expression level of hsa_circ_0067582 was significantly down-regulated in 80.2% (77/96) gastric cancer tissues compared with the adjacent normal tissues. C, The hsa_circ_0067582 expression levels in early gastric cancer (n = 24) compared with the adjacent normal (n = 24) tissues. D, The hsa_circ_0067582 expression levels in healthy gastric mucosa (n = 29), gastritis mucosa (n = 29), gastric intestinal metaplasia (n = 13), and gastric cancer (n = 96) tissues were determined by qRT-PCR. Higher Δ Ct value indicates lower expression (*P < .05, **P < .01, ***P < .001) functionless by-products and have become a new hot topic in very recent years with the RNA-seq technology of next-generation sequencing development. [9][10][11][12] CircRNAs can affect the regulation of gene expression through acting as competing endogenous RNA (ceRNAs) with the characteristics of stable structure, abundance, and tissue/developmental-stage-specific expression. 10,11 Previous studies demonstrated that circRNAs existed widely in all kinds of organizations and exhibit abnormal expression levels in digestive system cancers. 13 Guo et al 14  are gastric cancer-associated circRNAs based on our bioinformatics analysis. We first verify the expression levels of hsa_circ_0067582 and hsa_circ_0005758 in GC tissues and found that both of them were significantly down-regulated in GC tissues compared with the paired non-tumorous tissues ( Figure 1A and 2A). Furthermore, the expression of them were also both lower in the early GC tissues than the matched adjacent tissues (Figures 1C and 2C). As the prognosis of advanced GC is poor, the 5-year survival rate of early gastric cancer (EGC) that being performed by radical operations can be higher than 90%, even close to 100% for intramucosal invasive EGC. 17 The research showed that these two circRNAs may be used as potential indicators for early gastric cancer.
As it is a gradual progression from inflammation to atrophic gastritis, metaplasia, dysplasia, and finally to adenocarcinoma, 18 gastric intestinal metaplasia (GIM) is a premalignant stage in the Correa's cascade. Pittayanon et al 19 found that incomplete GIM is an important risk factor in predicting the development of high-grade dysplasia and/or gastric cancer. We also explored hsa_circ_0067582 and hsa_ circ_0005758 expression levels in different stages of gastric carcinogenesis. Compared with healthy control group, hsa_circ_0067582 and hsa_circ_0005758 expression levels were significantly decreased in GC tissues, but there was no significant difference between healthy control group and gastritis group (Figures 1D and 2D). These results showed that low expression levels of hsa_circ_0067582 and hsa_ circ_0005758 in tissues are closely related to gastric cancerogenesis.
Tumor markers, such as carcinoembryonic antigen (CEA), carbohydrate antigen 19-9 (CA 19-9), and carbohydrate antigen 72-4 (CA 72-4), which are simple and easy for screening tumors, have been widely used for the diagnosis of different types of cancers, including gastric cancer. 20 However, these markers have low sensitivity, the sensitivity in patients with recurrences was 44% for CEA and 56% for CA 19-9. 21 In our study, we constructed ROC curves to evaluate the clinical diagnostic value of hsa_circ_0067582 and hsa_circ_0005758. For hsa_circ_0067582, the AUC was up to 0.671, with sensitivity and specificity 55.2% and 75.0%, respectively ( Figure 3A). While for hsa_circ_0005758, the AUC was 0.721, with the sensitivity and specificity 75.0% and 67.7%, respectively ( Figure 3B). The results showed that hsa_circ_0067582 and hsa_ circ_0005758 were better than the traditional biomarkers of gastric cancer, and hsa_circ_0005758 had a better distinguishing value than hsa_circ_0067582.
Tumor progression is not only influenced by the reaction of the host to malignancy, but also depends on tumor clinicopathological features. A study conducted by Wang et al 22 revealed that tissue F I G U R E 2 Hsa_circ_0005758 expression levels in gastric cancer tissues. A, The expression levels of hsa_ circ_0005758 in cancer tissues (n = 96) and adjacent normal tissues (n = 96). B, The expression level of hsa_circ_0005758 was significantly down-regulated in 81.3% (78/96) gastric cancer tissues compared with the adjacent normal tissues. C, The hsa_circ_0005758 expression levels in early gastric cancer (n = 24) compared with the adjacent normal (n = 24) tissues. D, The hsa_circ_0005758 expression levels in healthy gastric mucosa (n = 29), gastritis mucosa (n = 29), gastric intestinal metaplasia (n = 13), and gastric cancer (n = 96) tissues were determined by qRT-PCR. Higher ΔCt value indicates lower expression (*P < .05, **P < .01, ***P < .001)