Biological and clinical implications of hsa_circ_0086720 in gastric cancer and its clinical application

Abstract Background Circular RNAs (circRNAs) are thought to be vital participants in carcinogenesis and have the characteristics of being stable, specific, and well conserved. However, their clinical significance and application value in gastric cancer (GC) are still poorly understood. Hsa_circ_0086720 was found to be a dysregulated circRNA in GC by microarray screening and was further explored for its clinical significance and application. Methods Hsa_circ_0086720 was detected in GC cell lines, tissues, and plasma, and the clinicopathological correlations were investigated. The existence, stability, origin, and change in the plasma hsa_circ_0086720 level were verified in early GC patients. Moreover, receiver operating characteristic and Kaplan–Meier survival curves were constructed to analyze the diagnostic and prognostic values, and bioinformatics analysis was used to identify the potential functions. Finally, risk factors and nomogram predicting were established. Results Hsa_circ_0086720 was found to be downregulated in gastric carcinogenesis, and tissue hsa_circ_0086720 was negatively associated with perineural invasion, Borrmann type, disease‐free survival, and overall survival. Hsa_circ_0086720 was stable in circulating plasma and was actively secreted by cells in gastric carcinogenesis. As a biomarker for early GC screening, plasma hsa_circ_0086720 had good sensitivity and specificity, and its stability met the clinical application requirements. Bioinformatics analysis suggested that dysregulated hsa_circ_0086720 has important functions in gastric carcinogenesis. Univariate Cox regression analysis identified factors associated with overall survival time and disease‐free survival time. The nomograms showed good accuracy of predicting survival time. Conclusion Hsa_circ_0086720 is a novel biomarker for screening early GC and predicting the prognosis of advanced‐stage patients.


| INTRODUC TI ON
Gastric cancer (GC) is the fifth most common cancer and the third leading cause of cancer-related death. 1 Although diagnostic and treatment technologies have greatly improved, the prognosis of GC patients is still unsatisfactory because of the high recurrence rate and low 5-year overall survival (OS) rate. 2,3 Possible explanations for this phenomenon may include indistinct molecular mechanisms of GC progression and an extremely low early diagnostic rate. 4 To date, early and advanced GC has still been mainly diagnosed by gastroscopy for symptomatic patients, and screening of clinically asymptomatic patients has been limited largely due to the absence of desirable biomarkers. Thus, the identification of new biomolecules for GC screening and treatment will be advantageous for reversing the current situation.
Distinctly different from linear RNAs, circular RNAs (circRNAs) have a special structure that lacks a 5′-terminal cap and 3′-terminal polyA tail, and the ends are joined together. [5][6][7] Although circRNAs were previously deemed to be byproducts of the RNA splicing process, they have gradually been reclassified from "junk" to "treasure" with the help of next-generation sequencing. 8,9 Recently, the critical functions of circRNAs have mainly been positioned to (1) act as miRNA sponges or transcriptional controllers to participate in gene regulation 8 ; (2) be translated to proteins closely related to carcinogenesis 9 ; (3) interact with RNA-binding proteins to regulate cell function 8,10 ; and (4) stably exist in body fluids and meet clinical testing requirements. 11,12 These findings provide a new direction for circRNA research in carcinogenesis.
Abnormal expression profiles of circRNAs in GC tissues have been covered by circRNA microarrays in our previous studies. 13 According to the microarray analysis results, we found that hsa_circ_0086720 is a significantly dysregulated circRNA in GC. Hsa_circ_0086720 contained 714 nucleotides in the final spliced sequence. Sequence analysis showed that its gene is located at human chr9:33953282-33996331 with a length of 43,049 nt, and the associated symbol is UBAP2 (ubiquitinassociated protein 2).
In this study, hsa_circ_0086720 was chosen as a target to explore its clinical significance and potential applications in gastric carcinogenesis. Thus, the existence, stability, origin, change, clinical values, and potential functions of hsa_circ_0086720 were verified in GC patients. Our study indicates that hsa_circ_0086720 is a novel biomarker for early GC screening and prognostic estimation in advanced-stage patients.

| Sample materials
A total of 96 GC tissues and paired adjacent nontumor tissues were obtained by surgical resection from GC patients, while 24 early GC (EGC) tissues and paired adjacent nontumor tissues were obtained by endoscopic submucosal dissection (ESD). 14 EGC is generally defined as cancer cells confined to the mucosa or submucosa, regardless of tumor size or local lymph node metastasis. Peripheral blood was collected from 40 healthy volunteers and 42 EGC patients who were finally pathologically diagnosed with intramucosal adenocarcinoma after ESD. Blood was collected before ESD treatment. Plasma and tissues were stored at −80°C. 15 The diagnosis of the patients was ultimately confirmed by pathology. The tumor-node-metastasis (TNM) staging system (7 th ed.) and National Comprehensive Cancer Network clinical practice guideline of oncology (V.1.2012) were used to determine the clinical stages and histological grades of tumors, respectively. 16 Borrmann type is classified according to the tumor appearance. According to Borrmann type, advanced GCs were classified into the following four types: nodular fungus type (Ⅰ), localized ulcer type (Ⅱ), ulcer infiltration type (Ⅲ), and diffuse infiltration type (Ⅳ). Written informed consent was obtained from all patients.

| Cell culture
Cell lines were purchased from the Chinese Academy of Sciences

| Risk factors and nomogram predicting
Univariate Cox regression analyses were constructed to estimate the potential risk factors based on hsa_circ_0086720 levels and the clinicopathological parameters. Positive risk factors were selected and incorporated into prognostic nomogram predicting models. Calibration curves were used to validate the predictive performance of nomograms. Data were analyzed by R software package including "survival" and "rms."

| Statistical analysis
Statistical analysis was performed with Statistical Program for Social Sciences (SPSS) 20.0 Software (SPSS). Data were analyzed by Student's t test, one-way ANOVA, or Kaplan-Meier analysis according to actual conditions. The statistical significance level was regarded as p < 0.05.

| Hsa_circ_0086720 is decreased during carcinogenesis
To verify the authenticity of hsa_circ_0086720 expression in the cir-cRNA microarray, we detected expression in cell lines and tissues.
Compared to expression in the normal gastric epithelial cell line GES-1, hsa_circ_0086720 expression in GC cells was significantly downregulated ( Figure 1A). Consistent with the cell results, compared to expression in control tissues, hsa_circ_0086720 expression was significantly decreased in 82.3% (79/96) of GC tissues (p < 0.001, Figure 1B,C).
Next, we explored hsa_circ_0086720 expression in 24 paired EGC tissues collected from ESD patients. Similar to the GC results, hsa_circ_0086720 was also downregulated in the very early stage of EGC (p < 0.001, Figure 1D).

| Clinicopathological correlation analysis
Hsa_circ_0086720 in GC tissue was significantly negatively associated with Borrmann type (p < 0.001) and perineural invasion F I G U R E 1 Hsa_circ_0086720 expression is downregulated during gastric carcinogenesis. (A) Hsa_circ_0086720 expression levels in four gastric cancer cell lines (AGS, BGC-823, HGC-27, and MGC-803) and a normal gastric epithelial cell line (GES-1) were determined by qRT-PCR. (B) The expression levels of hsa_circ_0086720 in cancer tissues (n = 96) and adjacent normal tissues (n = 96). (C) Compared to expression in adjacent normal tissues, Hsa_circ_0086720 expression was significantly downregulated in 82.3% (79/96) of gastric cancer tissues. (D) Hsa_circ_0086720 expression was significantly downregulated in early-stage EGC (n = 24). Data are shown as the means ± SDs of two independent experiments. A higher ΔC t value means a lower hsa_circ_0086720 level. Asterisks represent significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) TA B L E 1 Relationship of hsa_circ_0086720 expression levels (ΔC t ) in cancer tissues with clinicopathological factors of GC patients

| Prognostic value of hsa_circ_0086720 in GC
Surgical patients were divided into "low" or "high" groups according to their hsa_circ_0086720 levels in GC tissues compared to the levels in paired adjacent nontumor tissues. The ΔΔCt method (ΔΔCt = ΔCt GC tissue -ΔCt adjacent nontumor tissues ) was used to compare the hsa_circ_0086720 levels in GC tissues and paired adjacent nontumor tissues. The actual cutoff point was 0. If the patient's ΔΔCt value was greater than 0, they were assigned to the low expression group; otherwise, they were assigned to the high expression group.
Then, Kaplan-Meier analysis was performed to distinguish the differences in OS and disease-free survival (DFS) between the "low" and "high" hsa_circ_0086720 expression level groups. Interestingly, GC patients in the low group had longer OS and longer DFS times than those in the high group, and this trend was not only confined to advanced cancer stages (stages III and IV; Figure 2B,D) but also applied to all clinical stages (Figure 2A,C).

| Clinical diagnostic values as a biomarker
Circulating plasma is a suitable material for early cancer screening in clinical practice. We wondered whether hsa_circ_0086720 might exist in circulating plasma. Thus, the plasma PCR products were sequenced. As expected, the sequencing results were consistent with the sequence in circBase (http://circr na.org/; Figure 3A). Moreover, primer amplification sequences contain the back-splice junction site ( Figure 3A). The DNA sequence confirmed the existence of circulating plasma hsa_circ_0086720.
Then, circulating plasma from 40 healthy volunteers and 42 EGC patients was obtained and quantified by qRT-PCR. As shown in Figure 3B, the plasma hsa_circ_0086720 levels were significantly increased in the EGC stage (p < 0.001). EGC patients were finally pathologically diagnosed with intramucosal adenocarcinoma and were designated as locally cancerous ( Figure 3C-D). Moreover, it is noteworthy that the changes in hsa_circ_0086720 in plasma and cancer tissue were completely opposite.
Next, an ROC curve was constructed to investigate the potential value of plasma hsa_circ_0086720 in EGC screening. We found that the area under the ROC curve (AUC) was up to 0.771  Figure 3E).
The cutoff value, sensitivity, and specificity were 3.77, 67.4%, and 87.2%, respectively ( Figure 3E). However, the AUC of hsa_ circ_0086720 in cancer tissues was up to only 0.670 (95% CI, 0.594-0.746; p < 0.001; Figure 3F), which was lower than that in plasma. In cancer tissue, the sensitivity and specificity of hsa_ circ_0086720 were 41.67% and 86.46%, respectively. Compared to tissue hsa_circ_0086720, plasma hsa_circ_0086720 had a higher diagnostic value.

| Plasma hsa_circ_0086720 is mainly derived from cells
Gastric cells were cultured in serum-free medium, and the super-

| The stability of hsa_circ_0086720 in plasma
Good stability is a prerequisite for clinical detection. To test whether the stability of plasma hsa_circ_0086720 meets clinical requirements, freeze-thaw and incubation experiments were performed.
Our results confirmed that plasma hsa_circ_0086720 had good stability within a limited number of freeze-thaw and incubation times, which meets the needs of routine clinical detection ( Figure 4D,E).

| Predictors selection and nomogram model development
Univariate Cox regression analysis were constructed to estimate the potential risk factors associated with OS time and DFS time based on hsa_circ_0086720 levels and the clinicopathological parameters.
As shown in Table 2, Age, CEA, lymphatic metastasis, perineural invasion, and hsa_circ_0086720 level were associated with OS, whereas CEA, lymphatic metastasis, perineural invasion, and hsa_ circ_0086720 level were independently related with DFS ( Table 2).
Positive risk factors were selected and incorporated into prognostic nomogram predicting models. The nomograms showed good accuracy of predicting OS (C-index, 0.822; Figure 6A) and DFS (C-index, 0.772; Figure 6B). Calibration curves reflected good discriminative ability of prognosis ( Figure 6C,D).

| DISCUSS ION
The rapid development of biological technologies has contributed to the in-depth exploration of functional circRNAs. 17 CircRNAs are characterized as being stable, specific, and well conserved and have been suggested to be vital participants in carcinogenesis, making them possible candidates for cancer screening. [18][19][20] Accordingly, circRNAs have attracted more attention in tumor pathogenesis and biomarker research in recent years. In the field of GC, a number of abnormally expressed circRNAs have been discovered, and their functions and molecular mechanisms have been revealed. 13,21 For example, Yu et al. 22 reported that circ-TNPO3 acts as a protein decoy for IGF2BP3 to regulate the MYC-SNAIL axis suppressing the proliferation and metastasis of GC. Zhang et al. 23 identified that circLARP4 could be an miR-424-5p sponge and regulate large tumor suppressor kinase 1 (LATS1) expression to affect cell proliferation and invasion of GC. All this evidence suggests that circRNAs are emerging molecular targets for GC screening, diagnosis, and treatment.
Hsa_circ_0086720 is a dysregulated circRNA measured by microarray analysis in our previous studies. 13 In the current study, we found that compared to expression in controls, hsa_circ_0086720 expression was decreased in all four GC cell lines and in 82.3% of GC tissues ( Figure 1A-C). More importantly, hsa_circ_0086720 was also significantly downregulated at the very early stage of EGC ( Figure 1D). Bioinformatics analysis revealed that hsa_ circ_0086720 harbors some miRNA seed sequences and that the hsa_circ_0086720-miRNA axis is involved in various biological functions and signaling pathways ( Figure 5). This indicated that low hsa_ circ_0086720 expression in gastric tissues is closely associated with the initiation and development of gastric carcinogenesis. Luo et al. 24 conducted a meta-analysis to obtain the results that Borrmann type IV GC was associated with lymph node metastases, poor cell differentiation, and poor prognosis. Li et al. 25     for the combination of all three biomarkers in EGC. 27 Early cancer screening by blood biomarkers still faces challenges. In this study, we wondered whether hsa_circ_0086720 could exist in human circulating plasma and be used as an early GC screening biomarker. As expected, we confirmed its existence in plasma by cloning and sequencing methods ( Figure 3A). PCR results identified that the levels of plasma hsa_circ_0086720 were significantly increased in the EGC stage ( Figure 3B). The ROC curve further confirmed the potential values of plasma hsa_circ_0086720 in EGC screening ( Figure 3E). As a screening biomarker, the sensitivity and specificity of hsa_circ_0086720 for EGC were 67.4% and 87.2%, respectively. This means that compared to biomarkers currently used in the clinic, hsa_circ_0086720 has better sensitivity and specificity. Combined with the existing markers, endoscopy and hsa_circ_008672 may greatly improve the detection rate of EGC and make up for the diagnostic deficiencies of a single marker. Moreover, freeze-thaw and incubation experiments confirmed the stability of plasma hsa_circ_0086720, which implies that it meets the requirements needed for routine clinical detection ( Figure 4D,E). Therefore, hsa_circ_0086720 is a potential marker for early GC screening.
Exosomes are 30 to 150 nm endocytic membrane-derived vesicles actively secreted by cells. 28 Tumor-derived exosomes have reportedly been involved in carcinogenesis and can function as diagnostic biomarkers for specific cancers. 28,29 Recent research has shown that nucleic acids can be highly enriched and selectively released by exosomes to participate in cell communication. [30][31][32] Currently, in this study, the abnormal phenomenon of the opposite changes in the plasma and cancer tissue hsa_circ_0086720 levels has aroused our great interest. Thus, we speculated that plasma hsa_circ_0086720 is actively secreted by exosomes during GC carcinogenesis. As expected, we preliminarily verified that hsa_circ_0086720 in the supernatant tended to increase with time ( Figure 4A-C). This result indicates that plasma hsa_circ_0086720 mainly exists in exosomes and is actively secreted by cells during gastric carcinogenesis. In addition, the existing form of hsa_circ_0086720 also gives it better stability in plasma.
In this study, we chose hsa_circ_0086720 as a targeted cir-cRNA to explore its clinical significance in gastric carcinogenesis.
Our study validated the existence of hsa_circ_0086720 and analyzed the potential relationship between this circRNA and GC clinicopathological factors in 96 patients. However, we did not find other clinicopathological correlations, such as distal metastasis, differentiation, and TNM stage. This does not mean that our data completely negate the potential relationship between hsa_ circ_0086720 and other pathological factors. If the sample size is increased, other potential correlations may appear. Therefore, more samples are needed in future studies, and the biological function of hsa_circ_0086720 in cancer cells needs further experimental verification.
In conclusion, our study indicates that hsa_circ_0086720 is a novel biomarker for early GC screening and prognostic estimation in advanced-stage patients.