Gastric cancer (GC) ranks in the top five in global morbidity and mortality.^1,2 Despite many advances have been made in science and technology, the 5-year overall survival rate in most countries is less than 30%.^3,4 Therefore, the mechanism of GC tumorigenesis is explored to develop more effective therapeutic strategies.

Circular RNAs (circRNAs) show higher conservation and resistance to RNase R than the corresponding linear RNAs.⁵ circRNAs play roles in multiple molecular mechanisms, such as sponge microRNAs (miRNAs/miRs), which regulate gene expression and binding of RNA-binding proteins.^6,7 circRNAs are aberrantly expressed in a variety of cancers. For example, circASAP1, circ-ERBIN, circSATB2, and cir_0000263 were upregulated in hepatocellular carcinoma, colorectal cancer, non-small cell lung cancer (NSCLC), and cervical cancer, and were involved in cancer development.^8-11 Knockdown of circNRIP1 successfully prevented the growth of GC cells.¹² Circ_0043947 is derived from the exons of breast cancer gene 1 (BRCA1) gene, and was discovered to be higher in both the tissues and serum of GC patients according to the GSE83521 and GSE93541 datasets, and might function in GC cells via circRNA-miRNA-mRNA network.¹³ Here, we probed the action of circ_0043947 on GC oncogenicity.

According to the bioinformatics analysis,^14,15 we found that miR-384 possessed complementary sequences on circ_0043947 and CAMP response element binding protein (CREB1). Previous studies showed that miR-384 acted as an anticancer RNA to impair the malignant phenotypes of GC cells.^16,17 CREB1 acted as a proto-oncogene and was involved in metabolism and DNA repair processes.¹⁸ According to previous reports, CREB1 deregulation is closely linked with many cancers, such as colorectal cancer, bladder cancer, and NSCLC.^19-21 Norepinephrine-induced phosphorylation of CREB1 promotes human colon cancer cells.²² In addition, CREB1 is highly and related to the distant metastasis and advanced tumor stage of GC patients.²³ Nevertheless, the upstream regulatory mechanism of miR-384 and CREB1 in GC remains vague.

In this study, we identified circ_0043947 sequestered miR-384 to regulate CREB1 expression, thereby promoting the progression of GC. In conclusion, our data showed that circ_0043947 regulated GC development through miR-384/CREB1 axis.

1. Tissue specimens and cell culture

In this study, 65 GC tissue samples and adjacent normal tissues were obtained from Xiangxi Autonomous Prefecture People’s Hospital. Normal GES1 cells, and GC cell lines (AGS and HGC27) were purchased from Shanghai Institutes for Biological Sciences (Shanghai, China), and maintained in DMEM (Invitrogen, Carlsbad, CA, USA) with 10% FBS and 1% penicillin streptomycin (Invitrogen) at 37°C with 5% CO₂. The present study was approved by the ethical review committee of the First People’s Hospital of Changzhou (No. 20220316). Written informed consent was obtained from all enrolled patients. The animal study was approved by the Animal Care Committee of Xiangxi Autonomous Prefecture People’s Hospital (No. 20220316).

2. Quantitative real-time polymerase chain reaction (qRT-PCR)

TRIzol reagent (Takara, Tokyo, Japan) was adopted for isolating RNAs, which were then reversely transcribed into complementary DNA (cDNA) by using SuperScript III^® (Invitrogen), followed by qRT-PCR using the SYBR Green PCR kit (Takara) with primers (Table 1). The relative expression was calculated by using the 2^−ΔΔCt method.

Table 1 . The Sequences of Primers Used for qRT-PCR in This Study.

Name		Primers for PCR (5’-3’)
hsa_circ_0043947	Forward	TCATCATTCACCCTTGGCAC
	Reverse	TACCCATTTTCCTCCCGCAA
BRCA1	Forward	TTCACCCTCTGCTCTGGGTA
	Reverse	TGGTCACACTTTGTGGAGACA
miR-384	Forward	GTATGAATTCCTAGAAATTG
	Reverse	CTCAACTGGTGTCGTGGAG
CREB1	Forward	GTGTGTTACGTGGGGGAGAG
	Reverse	GCATCTCCACTCTGCTGGTT
GAPDH	Forward	GACAGTCAGCCGCATCTTCT
	Reverse	GCGCCCAATACGACCAAATC
U6	Forward	CTCGCTTCGGCAGCACA
	Reverse	AACGCTTCACGAATTTGCGT

qRT-PCR, quantitative real-time polymerase chain reaction..

3. Actinomycin D and RNase R treatment

RNAs were treated with 3 U/μg RNase R (Invitrogen) for 30 minutes at 37°C. In addition, cells were incubated with actinomycin D (2 μg/mL; Invitrogen) for specific times. The contents of circ_0043947 and BRCA1 were measured by qRT-PCR.

4. Nucleic acid electrophoresis

The determination of the cDNA and genomic DNA PCR products was carried out using 2% agarose gel electrophoresis with Tris-acetate-EDTA running buffer (Invitrogen). The ultraviolet irradiation was applied for band examination.

5. Cell transfection

Small interfering RNAs (siRNAs) or short hairpin RNA (shRNA) targeting circ_0043947 (si-circ_0043947 and sh-circ_0043947), miR-384 mimics and inhibitors (miR-384 and anti-miR-384) and the corresponding controls were provided by RiboBio (Guangzhou, China). The CREB1 overexpression plasmid (pcDNA-CREB1) was constructed with pcDNA3.1 plasmids (GenePharma, Shanghai, China).

6. Cell counting kit-8

The medium containing 10 µL cell counting kit-8 reagent (Dojindo, Tokyo, Japan) was used to incubate for 4 hours with GC cells 37℃. Cell proliferation was observed by detecting the absorbance at 450 nm.

7. 5-Ethynyl-2’-deoxyuridine (EdU) detection

Cells were incubated with EdU (RiboBio) for 2 hours, then stained with Apollo. Cell nucleus was stained with Hoechst for another 30 minutes. Finally, microscope was applied to acquire the image and calculate EdU-positive cells.

8. Colony formation assay

About 800 cells were hatched in 6-well plates for 2 weeks. The colonies were then dyed with 0.1% crystal violet (Solarbio, Beijing, China). And the number of colonies was recorded.

9. Wound healing assay

The transfected HGC27 and AGS cells were seeded in 6-well plates overnight, the cell monolayer was then scratched by a pipette tip. After washing, cells were immediately photographed and marked, and then the field of view was photographed at the same position after 24 hour-incubation.

10. Transwell assay

The transfected cells were added to the upper chamber of Transwell chambers (Corning, Cambridge, MA, USA) with an 8-μm pore membrane (migration) or Matrigel-coated membrane (invasion) in a 24-well format. The lower chamber was added with FBS-medium. After 24 hours, cells that migrated or invaded through the membranes were photographed and counted after crystal violet staining.

11. Flow cytometry

Transfected HGC27 and AGS cells were stained with Annexin V-fluorescein isothiocyanate and propidium iodide solution (Beyotime, Shanghai, China). The transfected HGC27 and AGS cells were stained with propidium iodide solution (Beyotime) containing RNase A. Flow cytometry was used for cell apoptosis and cell cycle analyses.

12. Western blot

Proteins were obtained from HGC27 and AGS cells with RIPA lysis buffer (Beyotime). Subsequently, protein samples were subjected to 10% separating gel and then loaded onto nitrocellulose membranes (Beyotime). Then, blots were probed on a rocker with primary antibodies at 4°C overnight. Next, the secondary antibody was incubated with the membranes for 2 hours. The blots were visualized and analyzed using ECL-PLUS Substrate and Image J software. The antibodies at 1/2,000 dilution were used in this study: B-cell lymphoma-2 (Bcl-2) (ab32124), Bcl-2-associated X (Bax) (ab32503), cleaved caspase-3 (ab2302), CREB1 (ab32515), and GAPDH (ab9485) (Abcam, Waltham, MA, USA).

13. RNA pull-down assay

The biotinylated-labeled circ_0043947 probe and oligo probe, biotinylated (Bio)-labeled miR-384 mimic and control (Bio-NC) were synthesized by RiboBio and transfected into HGC27 and AGS cells. After 48 hours, cell lysate and streptavidin-coated magnetic beads were incubated overnight at 37°C. After incubation, the bound RNA was washed and purified. RNA was used for qRT-PCR analysis.

14. Dual-luciferase reporter assay

The wild-type (WT) or mutant-type (MUT) luciferase reporter vectors circ_0043947 WT, circ_0043947 MUT, CREB1 3'UTR WT and CREB1 3’UTR MUT were established using pmirG10 luciferase vector (Promega, Madison, WI, USA), and co-transfected with miR-NC or miR-384 into HGC27 and AGS cells. Finally, the luciferase activity was detected by the dual-luciferase reporter kit (Promega).

15. Animal experiments

In order to establish the xenograft GC mice model, AGS cells stably transfected with circ_0043947 shRNA lentiviral vector (sh-circ_0043947) or negative control (sh-NC) were then subcutaneously injected into the nude mice (n=5; BALB/c, Charles River Labs, Beijing, China). The volume of xenografts was evaluated every week for 35 days. The animals were sacrificed at end point, the weight of the resected tumor was measured.

16. Immunohistochemistry analysis

Sections were incubated with antibodies against ki-67 (ab15580), CREB1 (ab32515) at 4℃ overnight, and secondary antibodies at 37℃ for 2 hours. Then the slide was finally incubated with diaminobenzidine and counterstained with hematoxylin (Beyotime).

17. Statistical analysis

The experiment was performed at least 3 times, and each group was subjected to three biological replicates in parallel. The data was shown as mean±standard deviation. The difference and comparison were analyzed by the Student t test or analysis of variance. The Kaplan-Meier method was used to generate the survival curve and the significance was analyzed by log-rank test. p<0.05 was considered as statistically significant.

1. GC samples and cells showed high circ_0043947 expression level

Circ_0043947 was highly expressed in GC tissues (n=65) compared to paired normal tissues. Also, the overexpression of circ_0043947 was observed in AGS and HGC27 compared to the GES1 cells. Circ_0043947, but not linear BRCA1, could resist RNase R digestion. Moreover, we used actinomycin D to inhibit transcription and then detect the half-life of circ_0043947 and linear BRCA1, and found that circ_0043947 was more stable than BRCA1. In addiiton, circ_0043947 was only amplified in cDNA but not in genomic DNA (Fig. 1), further revealing the high stable of circ_0043947. In addition, it was found that higher circ_0043947 expression was correlated with advanced TNM stage, lymph node metastasis and distant metastasis (Table 2). Moreover, high expression of circ_0043947 was significantly related to the low overall survival of GC patients (Fig. 1).

Figure 1. Circ_0043947 expression was upregulated in gastric cancer (GC) tissues and its high expression predicted poor prognosis. (A) Circ_0043947 expression in 65 pairs of GC tissues and matched para-carcinoma tissues was measured by qRT-PCR. (B) Expression of circ_0043947 was determined in GC cell lines compared with GES1 cells. (C, D) The abundance of circ_0043947 and linear BRCA1 in AGS and HGC27 cells after treatment with actinomycin D at the indicated time points. (E, F) qRT-PCR analysis of the levels of circ_0043947 and linear BRCA1 in GC cells treated with or without RNase R. (G) Agarose gel electrophoresis and PCR were conducted to validate the existence of circ_0043947. Circ_0043947 can be amplified by divergent primers in cDNA but not in gDNA. (H) Kaplan-Meier survival analysis showed that patients with high circ_0043947 expression had a poorer prognosis. The data are presented as the mean±SD. qRT-PCR, quantitative real-time polymerase chain reaction; cDNA, complementary DNA; gDNA, genomic DNA. *p<0.05.

Table 2 . Association of Circ_0043947 Expression with Clinicopathological Features of 65 Gastric Cancer Patients.

Features	No. of patients	Circ_0043947 expression		p-value
		High (n=33)	Low (n=32)
Age
≥60 yr	41	19	22	0.351
<60 yr	24	14	10
Sex
Male	35	20	15	0.267
Female	30	13	17
TNM stage
I-II	38	13	25	0.002
III-IV	27	20	7
Lauren type
Intestinal	25	10	15	0.170
Diffuse	40	23	17
Tumor location
Gastric fundus	15	8	7	0.089
Gastric corpus	21	11	10
Pylorus	29	16	13
Lymph node metastasis
Yes	33	22	11	0.009
No	32	11	21
Distant metastasis
Yes	25	19	6	0.001
No	40	14	26

2. Knockdown of circ_0043947 inhibited the progression of GC cells

Following the transfection of siRNAs targeting circ_0043947, qRT-PCR assay showed high interference rates in siRNA#1 and #2 transfection, which were selected for subsequent experiments. Deletion of circ_0043947 repressed the viability of AGS and HGC27 cells. Consistently, the proportion of EdU-positive cells was decreased by circ_0043947 depletion. Moreover, knockdown of circ_0043947 significantly inhibited colony-forming ability of GC cell lines. Additionally, cell migration and invasion were inhibited by knockdown of circ_0043947 (Fig. 2). Furthermore, cell cycle of si-circ_0043947-transfected AGS and HGC27 cells was arrested at the G0/G1 phase. The flow cytometry results showed that circ_0043947 deletion significantly evoked the apoptosis in AGS and HGC27 cells. Besides, Western blot data demonstrated that suppression of circ_0043947 resulted in the decline in Bcl-2 expression and the increase in Bax and cleaved caspase-3 protein content (Fig. 3). Overall, circ_0043947 knockdown inhibited cell oncogenic phenotypes in vitro.

Figure 2. Knockdown of circ_0043947 inhibited the proliferation, migration and invasion of gastric cancer cells. (A) qRT-PCR was performed to verify the silencing efficiency of three independent siRNAs targeting circ_0043947 (si-circ_0043947#1, si-circ_0043947#2, and si-circ_0043947#3). The impacts of circ_0043947 knockdown on cell proliferation were checked by using CCK-8 (B, C), EdU staining (D) and colony formation assays (E) in AGS and HGC27 cells. Effects of circ_0043947 knockdown on invasion and migration were detected by wound healing (F) and transwell assays (G, H). The data are presented as the mean±SD. qRT-PCR, quantitative real-time polymerase chain reaction; siRNA, small interfering RNA; NC, negative control; CCK-8, cell counting kit-8; EdU, 5-ethynyl-2’-deoxyuridine. *p<0.05.

Figure 3. Knockdown of circ_0043947 promoted apoptosis and induced G0/G1 phase arrest in gastric cancer cells. AGS and HGC27 cells were transfected with si-circ_0043947#1, si-circ_0043947#2, or si-NC. (A, B) Flow cytometry was performed to evaluate the effect of circ_0043947 knockdown on cell cycle distribution and apoptosis. (C, D) Western blots were performed to test the levels of apoptosis-related proteins, including Bcl-2-associated X (Bax), B-cell lymphoma-2 (Bcl-2), and cleaved-caspase-3. The data are presented as the mean±SD. NC, negative control. *p<0.05.

3. Circ_0043947 acted as a sponge for miR-384

To explore whether circ_0043947 could function as a "miRNA sponge" in GC cells, biotinylated circ_0043947 probes were designed to identify which miRNAs might interact with circ_0043947. qRT-PCR analysis revealed that miR-384 was a miRNA that was significantly captured by circ_0043947 probe in AGS and HGC27 cells. The circinteractome website was employed to predict the interaction binding sites of miR-384 and circ_0043947. miR-384 mimic was constructed and its transfection efficiency was confirmed by qRT-PCR. Furthermore, it was discovered that cells co-transfected with circ_0043947 WT and miR-384 mimic showed significantly decreased luciferase activity. Further pull-down assay clarified that biotin-labeled miR-384 captured more circ_0043947 compared with control group. Thereafter, a low level of miR-384 in the 65 GC samples was observed. Furthermore, miR-384 expression was downregulated in GC cells compared to GES1 cells (Fig. 4). Collectively, these data demonstrated that circ_0043947 bound to miR-384 in GC.

Figure 4. miR-384 was verified as a direct target of circ_0043947. (A) An RNA pull-down assay was performed to detect the correlation between circ_0043947 and miR-384 in AGS and HGC27 cells. (B) The putative binding sites between miR-384 and circ_0043947 are shown. (C) qRT-PCR analysis for the transfection efficiency of the miR-384 mimic. (D, E) Relative luciferase activities were detected in GC cells transfected with WT or MUT circ_0043947 luciferase reporter plasmids and miR-384 or negative control. (F) RNA pull-down assay was executed in AGS and HGC27 cells, followed by qRT-PCR analysis of the enrichment of circ_0043947. (G, H) The expression levels of miR-384 in GC tissues and GC cells were determined by qRT-PCR. The data are presented as the mean±SD. qRT-PCR, quantitative real-time polymerase chain reaction; GC, gastric cancer; WT, wild type; MUT, mutant type; NC, negative control. *p<0.05.

4. miR-384 inhibitor effectively reversed the effect of circ_0043947 knockdown on GC progression

Then we further investigated whether circ_0043947 can affect the functions of AGS and HGC27 cells. Following indicated transfection, miR-384 content was notably downregulated in GC cells infected with anti-miR-384. Further functional studies demonstrated that circ_0043947 deletion could repress the proliferation, migration and invasion of AGS and HGC27 cells, while miR-384 inhibitor significantly abolished these effects. Flow cytometric analysis manifested that transfection of the miR-384 inhibitor abolished circ_0043947 deficiency-induced G0/G1 arrest and apoptosis enhancement. Western blot analysis further validated the downregulated expression of Bcl-2, and the upregulation of Bax and cleaved-caspase-3 levels mediated by si-circ_0043947 were markedly restored by miR-384 inhibition (Fig. 5).

Figure 5. Knockdown of miR-384 rescued the si-circ_0043947-mediated effects on GC cell proliferation, metastasis, cell cycle arrest and apoptosis. (A) miR-384 expression was monitored using qRT-PCR assay in GC cells transfected with si-NC or si-circ_0043947. (B-M) GC cells were co-transfected with si-NC, si-circ_0043947#2, si-circ_0043947#2 + anti-miR-NC, or si-circ_0043947#2 + anti-miR-384. (B-E) CCK-8, EdU assay and colony formation assays were performed to evaluate the proliferation ability of GC cells. (F-H) Wound healing and transwell experiments were conducted to assess GC cell migration and invasion abilities. (I-K) A flow cytometry assay was used to evaluate the effect of miR-384 silencing on the si-circ_0043947-mediated effects on the cell cycle distribution and apoptosis. (L, M) Western blot analysis of the protein levels of Bcl-2-associated X (Bax), B-cell lymphoma-2 (Bcl-2), and cleaved-caspase-3 in AGS and HGC27 cells. The data are presented as the mean±SD. GC, gastric cancer; qRT-PCR, quantitative real-time polymerase chain reaction; NC, negative control; CCK-8, cell counting kit-8; EdU, 5-ethynyl-2’-deoxyuridine. *p<0.05.

5. Circ_0043947/miR-384/CREB1 formed an axis

TargetScan bioinformatics predicted that CREB1 shared the same microRNA response elements with miR-384. Dual-luciferase reporter assays then demonstrated that the luciferase activity in CREB1 3’UTR WT group rather than the MUT group was remarkably reduced by miR-384 mimics. The RNA pull-down assay further showed that biotin-labeled miR-384 probe captured more CREB1 compared with the control probe. Furthermore, miR-384 mimics decreased CREB1 level in GC cells (Fig. 6). Thereafter, we found CREB1 expression was increased in stomach adenocarcinoma (Supplementary Fig. 1). We detected the expression profile of CREB1 in GC tissues and GC cell lines, it was found that the content of CREB1 in GC tissues and cells was significantly higher than that in non-tumor tissues and normal cell lines (Fig. 6). Moreover, the decline of CREB1 expression induced by the knockdown of circ_0043947 was reversed by the miR-384 inhibitor (Fig. 6).

Figure 6. Circ_0043947 positively regulated CREB1 expression in GC cells by sponging miR-384. (A) Potential binding sites between miR-384 and CREB1 were predicted by Targetscan database. (B, C) The targeting relationship between miR-384 and CREB1 was analyzed by luciferase reporter gene assay. (D) RNA pull-down assays and qRT-PCR assays were executed to determine the enrichment of CREB1. (E) The protein level of CREB1 in AGS and HGC27 cells transfected with miR-NC or miR-384 mimic was determined by Western blot. (F-I) The mRNA and protein levels of CREB1 were examined by qRT-PCR and Western blot assays in GC tissues and cells. (J, K) The mRNA and protein levels of CREB1 were measured by qRT-PCR and Western blot assays after transfection of si-NC, si-circ_0043947#2, si-circ_0043947#2 + anti-miR-NC, or si-circ_0043947#2 + anti-miR-384. The data are presented as the mean±SD. CREB1, CAMP response element binding protein; GC, gastric cancer; qRT-PCR, quantitative real-time polymerase chain reaction; NC, negative control; mRNA, messenger RNA. *p<0.05.

6. miR-384 inhibited GC cell progression in vitro through inhibiting CREB1

Next, we investigated whether miR-384/CREB1 axis was associated with GC development. CREB1 level was upregulated by pcDNA-CREB1 transfection in miR-384-infected GC cells. Functionally, miR-384 mimics resulted in a significant suppression on cell viability, EdU-positive cells, colony formation ability, cell migration and invasion. However, overexpression of CREB1 significantly abolished these effects. In addition, results of flow cytometry exhibited that CREB1 overexpression abolished miR-384 mimics-induced cell cycle arrest in G0/G1 phase and cell apoptosis enhancement. miR-384 reduced Bcl-2 level, and elevated Bax and cleaved-caspase-3 levels in GC cells, and these effects were also reverted by CREB1 overexpression (Fig. 7). Altogether, these findings indicated that miR-384 participated in GC progression by targeting CREB1.

Figure 7. miR-384 inhibited GC cell malignant behaviors by modulating CREB1 expression. (A) The protein level of CREB1 in AGS and HGC27 cells transfected with pcDNA-NC or pcDNA-CREB1 was determined by Western blot analysis. (B-E) CCK8 assays, EdU assays and colony formation assays were performed to analyze the cell proliferation ability. (F-H) Wound healing and transwell assays were utilized to determine cell migration and invasion. (I-K) Flow cytometry was used to evaluate cell cycle distribution and apoptosis. (L, M) Western blot analysis was used to detect the protein levels of Bcl-2-associated X (Bax), B-cell lymphoma-2 (Bcl-2), and cleaved-caspase-3 in AGS and HGC27 cells. The data are presented as the mean mean±SD. GC, gastric cancer; CREB1, CAMP response element binding protein; NC, negative control; CCK-8, cell counting kit-8; EdU, 5-ethynyl-2’-deoxyuridine. *p<0.05.

7. Silencing of circ_0043947 hindered tumor growth in vivo

Finally, in vivo assay was carried out. Circ_0043947 deletion notably decreased the volume and weight of tumors. As expected, circ_0043947 level was decreased in the resected xenografts of the sh-circ_0043947 group, while miR-384 level was increased. Furthermore, silencing of circ_0043947 resulted in CREB1 decline in xenografts. Immunohistochemistry assays also confirmed that silencing of circ_0043947 weakened CREB1 and Ki-67 protein content in xenografts (Fig. 8).

Figure 8. Knockdown of circ_0043947 inhibited GC tumor growth in vivo. AGS cells stably transfected with sh-NC or sh-circ_0043947 were subcutaneously injected into the nude mice to establish the xenograft mice model. Tumor growth curve (A), tumor weight (B), and schematic representation (C) of xenograft tumors are shown. (D) The levels of circ_0043947 and miR-384 were examined in xenograft tumor tissues via qRT-PCR. € The protein levels of CREB1 were detected in xenograft tumor tissues by Western blot analysis. (F) Representative images of CREB1 and Ki-67 immunohistochemistry staining (×100) staining in xenograft tumor tissues are shown. The data are presented as the mean±SD. GC, gastric cancer; NC, negative control; qRT-PCR, quantitative real-time polymerase chain reaction; CREB1, CAMP response element binding protein. *p<0.05.

GC is still one of the cancers with the highest incidence in the world. Invasion and metastasis were the biggest obstacles to successful surgical resection of tumors.²⁴ With the development of technology, circRNA has gradually become the main focus of tumor research due to their regulatory action in tumors.²⁵ In this study, we explored the existence and function of circ_0043947 in GC. GC tissues and cells manifested high circ_0043947 expression, and circ_0043947 deletion weakened cell proliferation, invasiveness and migration, and promoted apoptosis in vitro. Silencing of circ_0043947 curbed tumor growth in vivo. Currently, molecular targeted therapies are revolutionizing therapeutics, with many such treatments already approved by the Food and Drug Administration for a myriad of cancer types. Additionally, more than 50 RNA-based drugs are currently under clinical testing.^26,27 Therefore, circ_0043947 siRNAs or shRNAs-based drugs may possess enormous potential for clinical therapy in GC.

miRNA are small non-coding RNAs, with an average 22 nucleotides in length, which are involved in biological regulation.²⁸ miR-4651 inhibited the growth of NSCLC cells in human NSCLC tissues.²⁹ miR-144-3p promoted the progression of nasopharyngeal carcinoma by targeting phosphatase and tensin homolog.³⁰ In this study, the bioinformatics website predicted that miR-384 has the ability to bind to circ_0043947, and we verified that miR-384 was targeted by circ_0043947 in GC cells. The inhibitory effect of miR-384 on NSCLC,³¹ colorectal cancer³² and osteosarcoma³³ has been discovered. In our study, low level of miR-384 were observed in GC tissues and cells. In addition, it was confirmed that miR-384 was targeted by circ_0043947. In addition, silencing of miR-384 can partially restore the suppressing action of circ_0043947 deletion on GC cell progression.

circRNA has been considered as a sponge of miRNA, thereby increasing the level of target genes. This study displayed a high level of CREB1 in GC. And overexpression of miR-384 or silencing circ_0043947 inhibited the expression of CREB1. Through rescue experiments, CREB1 overexpression could counteract the anticancer action of miR-384 on GC progression. Finally, the anti-growth function of circ_0043947 knockdown on GC tumor growth has been confirmed by in vivo experiments. All these data revealed the significant influence of the circ_0043947/miR-384/CREB1 pathway on the regulation of GC progression.

In conclusion, we proved that circ_0043947 bind to miR-384 to regulate the expression of CREB1, thereby affecting cell growth, invasiveness, and migration. This study provided new evidence for our understanding of the function of circ_0043947 and provided a promising biomarker for GC treatment.

No potential conflict of interest relevant to this article was reported.

Study concept and design: F.Z., Y.L. Data acquisition: P.Y., Y.L., W.Y. Data analysis and interpretation: P.Y., Y.L., W.Y. Drafting of the manuscript: C.Z. Critical revision of the manuscript for important intellectual content: C.Z., F.Z., Y.L. Statistical analysis: P.Y., Y.L., W.Y. Administrative, technical, or material support: W.Y. Study supervision: W.Y. Approval of final manuscript: all authors.

Supplementary materials can be accessed at https://doi.org/10.5009/gnl230173.

The analyzed data sets generated during the present study are available from the corresponding author on reasonable request.