Chrysin Induced Cell Apoptosis Through H19/let-7a/COPB2 Axis in Gastric Cancer Cells and Inhibited Tumor Growth

Background Chrysin is a natural flavone that is present in honey and has exhibited anti-tumor properties. It has been widely studied as a therapeutic agent for the treatment of various types of cancers. The objectives of this present study were to elucidate how chrysin regulates non-coding RNA expression to exert anti-tumor effects in gastric cancer cells. Methods Through the use of RNA sequencing, we investigated the differential expression of mRNAs in gastric cancer cells treated with chrysin. Furthermore, COPB2, H19 and let-7a overexpression and knockdown were conducted. Other features, including cell growth, apoptosis, migration and invasion, were also analyzed. Knockout of the COPB2 gene was generated using the CRISPR/Cas9 system for tumor growth analysis in vivo. Results Our results identified COPB2 as a differentially expressed mRNA that is down-regulated following treatment with chrysin. Moreover, the results showed that chrysin can induce cellular apoptosis and inhibit cell migration and invasion. To further determine the underlying mechanism of COPB2 expression, we investigated the expression of the long non-coding RNA (lncRNA) H19 and microRNA let-7a. Our results showed that treatment with chrysin significantly increased let-7a expression and reduced the expression of H19 and COPB2. In addition, our results demonstrated that reduced expression of COPB2 markedly promotes cell apoptosis. Finally, in vivo data suggested that COPB2 expression is related to tumor growth. Conclusions This study suggests that chrysin exhibited anti-tumor effects through a H19/let-7a/COPB2 axis.


INTRODUCTION
Currently, gastric cancer (GC) is the third most common cause of human mortality among malignant cancers (1). Although surgical treatment for GC has led to increased survival rates, the diagnosis of GC needs to improve (2). There are several potential biomarkers of GC, including many genes and cell signaling pathways that are involved in GC development, such as BRCA2 and Ras-Raf-MAPK signaling (3,4). Coatomer protein complex subunit beta 2 (COPB2) is a protein that functions to transport other proteins as vesicles from the endoplasmic reticulum to Golgi apparatus (5). Recently, numerous reports have indicated that COPB2 is abnormally expressed in colorectal cancer (CRC), cholangiocellular carcinoma and lung cancer (6)(7)(8). Previous studies have indicated that the reduction of COPB2 expression inhibited cell growth and induced apoptosis through the JNK/c-Jun signaling pathway in RKO and HCT116 cells (9). Moreover, miR4461 and miR335 were found to regulate COPB2 expression, which subsequently inhibited cell growth in CRC and lung cancer cells (10,11).
Increasing evidence suggests that non-coding RNAs (ncRNAs), such as miRNAs, can be applied to the classification of GC (12). As tumor suppressors, the let-7 family is down-regulated in GC (13). Increased expression of let-7a has been shown to inhibit cell migration and invasion in prostate cancer (14). Compared to the loss of let-7a expression, the long non-coding RNA (lncRNA) H19 has been shown to be highly expressed in cancers, including GC (15). As a molecular sponge, H19 was found to be related to let-7 in the context of breast cancer stem cells (16). Previous reports have suggested that H19 expression suppressed endogenous let-7 while H19 mutant was not related to let-7 (17). Additionally, reduced expression of H19 induced cellular apoptosis and inhibited cell growth in HCC (18). However, there is little evidence to suggest that COPB2 expression is associated with lncRNAs and miRNAs in GC.
As a traditional Chinese medicine, chrysin is a natural flavone that has anti-cancer function (19). Previous reports have indicated that chrysin induces cellular apoptosis and inhibits tumor glycolysis in HCC (20). Moreover, chrysin has been shown to inhibit cell migration and invasion in melanoma cells (21). In this study, chrysin was used to treat GC cells and we screened differentially expressed genes using RNA-seq. Additionally, we created a COPB2 knockout (KO) cell line using the CRISPR/Cas9 system. Our findings indicate that chrysin can regulate COPB2 expression through let-7, which is antagonized by H19.

Cell Culture and Chrysin Treatment
Human GC cell lines (SGC7901, MKN45 and BGC823) and the human gastric epithelial cell line GES-1 were grown in Dulbecco's modified Eagle's (DMEM; Gibco) supplemented with 10% fetal bovine serum (Gibco), and cultured at 37°C in 5% CO 2 . The human gastric epithelial cell GES1 served as control. Experiments were conducted by treating GES1, SGC7901, MKN45 and BGC823 cells with 40 mM of Chrysin (Yuanye Bio-Technology, Shanghai) for 48 h when they reached 80 to 90% confluence.

RNA Isolation and RNA-Seq Analysis
GC cells were treated with 40 mM chrysin for 48 h, after which total RNA was extracted. To remove and purify ribosomal RNA (rRNA), we used the RiboZero Magnetic Gold Kit (Epidemiology, USA). Then, RNA-seq (Sangon Biotech, Shanghai, China) was carried out on HiSeq2500 (Illumina, USA) to analyze raw reads, which were quality controlled by FastQC. Using the HISAT2 software, the sequenced-reads were aligned to the reference sequence. The gene expression analysis and differential gene expression analysis were determined using DEGseq and DESeq program, respectively, in HISAT2 (qValue <0.05, Fold Change >2). Using clusterProfiler, the enrichment analysis, including Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway, of differential expressed genes was determined.
The COPB2-KO was generated using the CRISPR/Cas9 system (px459, Addgene, USA). The single guide RNAs (sgRNAs) were designed as previously reported (22). The sequences of sgRNAs are listed in Table S2. Transfection was conducting using COPB2-KO (5 mg) and Lipofectamine 2000 (Invitrogen, USA) for 48 h. Next, puromycin helped select the positive clones. After 14 days, the clones (COPB2-KO) were grown and picked for subsequent western blot, qPCR and sequencing analysis.

DNA Methylation Analysis
The Bisulfite Sequencing PCR (BSP) protocol was carried out as previously described (23). Using the TIANamp Genomic DNA Kit (TIANGEN, Beijing, China), the DNA of SGC7901 and BGC823 cells were extracted. The DNA was modified using CpGenome ™ Turbo Bisulfite Modification Kit (Millipore, USA). The differentially methylated regions (DMRs) of H19 were amplified using nested PCR. The products, which included 10 positive clones, were analyzed using the BiQ Analyzer software (http://biq-analyzer.bioinf.mpi-inf.mpg.de/tools/Methylation Diagrams/index.php). The primers of H19 DMR are listed in Table S3.

Gene Expression Analysis
The GC cells' RNA was extracted and cDNAs were generated using the cDNA first-strand synthesis kit (TIANGEN, China). Gene expression analysis was conducted using quantitative realtime PCR (qPCR). The conditions for qPCR included heating to 94°C for 3 min, and then denaturation at 94°C for 10 s after 35 cycles. The annealing was carried out at 59°C for 15 s. The products were extended at 72°C for 30 s. The internal controls included GAPDH and U6 for genes and miRNAs, respectively. The primers for qPCR are listed in Table S4. The sequences of COPB2 exon 5 and exon 22 are listed in Table S5.

Cell Migration and Invasion Analysis
Wound healing assay was conducted to analyze cell migration of GC cells. In brief, 5 × 10 5 cells were cultured and seeded before treatment with chrysin, siRNAs, overexpression vectors, let-7a-3p mimics or let-7a-3p inhibitor. The cells were cultured after a scratched line was created with culture medium without serum. Cell migration was measured using the scratched area at 12, 24 and 48 h.
For cell invasion assays, GC cells (3 × 10 4 ) were cultured and seeded with 20 ml Matrigel prior to treatment with chrysin, siRNAs, overexpression vectors, let-7a-3p mimics and let-7ainhibitor (BD Biosciences, USA). Next, 0.5 ml of medium, which contained 10% FBS, was added to the cells for 24 h. Then, 0.2% crystal violet dye (Solarbio, China) was used to stain cells after being fixed with 4% paraformaldehyde. The stained cells were assayed using the ImageJ software.

Cell Apoptosis Analysis
GC cells (1 × 10 6 ) were cultured prior to treatment with chrysin, siRNAs, overexpression vectors, let-7a-3p mimics and let-7ainhibitor for detection of cellular apoptosis, as previously described (25). Then, Annexin V-FITC/PI reagent was added to cell to react for 30 min and flow cytometry (BD Biosciences, Franklin Lakes, NJ, USA) was used to detect fluorescent cells.

Hematoxylin and Eosin (H&E) Staining
Tumor tissues from the control and chrysin groups were fixed in 4% paraformaldehyde for 48 h, embedded in paraffin wax and sliced into 5 mm sections. The slides were then stained with H&E and cancer cell infiltration was determined by observation under a light microscope.

Animals and Animal Care
For in vivo experiments, 17 female nude mice (6-8 weeks old) were utilized to determine the effect of chrysin treatment and COPB2 KO on tumor growth. The mice were acquired and grouped-housed in the Laboratory Animal Center of Jilin University. All mice were provided ad libitum access to standard rodent food and tap water within the laboratory cages, as well as under specific pathogen-free (SPF) conditions. The BGC823, pcDNA3.1-COPB2 and COPB2-KO cell lines (3 × 10 6 ) were subcutaneously injected into the left flank of each mouse, and tumors were observed after seven days. The tumor length (L) and width (W) were calculated as L × W 2 /2.

Statistical Analysis
An unpaired Student's t-test was utilized in the present study. The SPSS 16.0 software (SPSS Inc., Chicago, IL, USA) helped conduct statistical analysis. All data was expressed as mean ± SD. A p-value of <0.05 was considered to be statistically significant. The website of http://ualcan.path.uab.edu/index.html was used for The Cancer Genome Atlas (TCGA) analysis. The TargetScan database was used to predict miRNA.

Screen of Differentially Expressed Gene of Chrysin-Treated GC Cells
In order to analyze gene expression patterns of chrysin treatment in gastric cancer cells, we performed RNA-Seq. Overall, 20,010 genes were identified as core genes ( Figure 1A). Compared to the control group, 380 genes were significantly up-regulated while 2,071 were significantly down-regulated ( Figure 1B). Data from heatmap and KEGG pathway suggests that the differentially expressed genes have functions in cell death and growth ( Figures 1C, D). In order to confirm this data, six genes (CAPN2, MXI1, HSPA9, RHBDD2, COPB2 and GABAPAPL1), which were related to cell growth and death, were further validated ( Figure 1E). The qPCR results indicated that COPB2 expression was downregulated upon chrysin treatment in SGC7901 ( Figure 1F) and MKN45 ( Figure 1G) cells. These results indicate that COPB2 expression is regulated by chrysin in GC cells.

Chrysin Increased let-7a and Inhibited H19 and COPB2 Expression in GC Cells
In order to further verify the expression levels of COPB2 in GC, we utilized the TCGA database. Results indicated increased expression of COPB2 in primary tumor of the stomach adenocarcinoma (STAD) patients ( Figure S1A). Compared to GES1 cells, qPCR and western bolt results indicated increased expression of COPB2 in MKN45, SGC7901 and BGC823 cells (Figures 2A, B). To investigate which miRNAs were involved in COPB2 expression, bioinformatics analysis was performed. The database suggested that let-7a targets COPB2 (Figures 2C, S2). Furthermore, we analyzed let-7a levels in the TCGA database. Results indicated no differences between normal and tumor tissues ( Figure S1B). However, qPCR results suggested that let-7a levels were reduced in GC cells, compared to GES1 cells ( Figure 2D). Considering that let-7a is associated with expression of the lncRNA H19, we analyzed the expression pattern of H19. The TCGA database indicated increased expression of H19 among STAD patients ( Figure S1C). The qPCR results indicated increased expression of H19 in GC cells ( Figure 2E). DNA methylation results indicated the hypo-methylation profile of H19 DMR in GC cells ( Figure S3). The cell growth was analyzed after chrysin treatment. The CCK8 results indicated that 40 mM was the optimal dose for subsequent experiments ( Figure 2F). Moreover, qPCR results indicated increased expression of let-7a, as well as reduced expression of H19 after chrysin treatment in GC cells ( Figure 2G). Besides, chrysin was able to induce cell apoptosis, as well as inhibit cell migration and invasion in GC cells (Figures 3, S4). These results indicate that chrysin has an anti-cancer role and regulates expression of COPB2, H19 and let-7a in GC cells.

H19/let-7a Regulate COPB2 Expression
Considering that H19 has a role in let-7a expression, we analyzed the effect of H19 knockdown and overexpression in GC cells. The results indicated reduced expression of let-7a in the H19 overexpression group, as well as overexpression of let-7a in the H19 knockdown group (Figures 4A-C). In order to investigate the expression pattern of let-7a, we utilized miRNA mimics and inhibitors. The qPCR and western blot results demonstrated that let-7a mimics led to suppressed expression of COPB2 (Figures 4D-G). These results confirm that H19 acts as a sponge that competes with let-7a, thus regulating COPB2 expression.

Reduced Expression of COPB2 Induced Cell Apoptosis and Inhibited Cell Invasion
In order to analyze whether COPB2 expression has an effect on cell death and growth, we transfected knockdown and overexpression vector of COPB2 into GC cells. The CCK8 data demonstrates that reduced expression of COPB2 and increased expression of let-7a inhibits cell growth ( Figure S5). Moreover, our results suggested that reduced expression of COPB2 induced cellular apoptosis (Figures 5A, B). In order to validate this, we analyzed markers of cell apoptosis. The results showed that increased expression of p53 was observed in the COPB2 knockdown group ( Figure S6). In addition, COPB2 expression did not have an effect on cell migration in GC cells ( Figures  5C, D). However, our data shows that reduced expression of COPB2 inhibited cell invasion ( Figures 5E, F). Next, we investigated the effect of chrysin about cell migration and invasion in the COPB2 overexpression group. This result demonstrated that chrysin induced cell apoptosis and inhibited cell migration and invasion ( Figures S7, S8). In order to validate this finding, the overexpression and knockdown of H19 and let-7a were used to analyze cell apoptosis. The results demonstrated that reduced expression of H19 and increased expression of let-7a induced cell apoptosis ( Figure S9). Overall, these results suggest that COPB2 has a role in cell apoptosis and invasion.

Loss Expression of COPB2 Inhibited Tumor Growth In Vivo
In order to assess the anti-cancer effect of chrysin in vivo, we utilized nude mice. The BGC823 cells were injected into nude mice and after seven days, mice were treated with chrysin (20 mg/kg) for two weeks. The results showed that chrysin is able to inhibit tumor growth and COPB2 expression in vivo ( Figures  6A-D). The H&E staining results confirm this data ( Figure S10). Moreover, qPCR result indicated that chrysin inhibited the expression of H19, and increased let-7a in vivo ( Figure S11). In order to determine the effect of loss expression of COPB2 in vivo, CRISPR/Cas9 system was used to edit the COPB2 exon 5 ( Figure 6E). The qPCR data showed lower expression of COPB2 in the COPB2 KO group compared to the control group ( Figure 6F). Moreover, chrysin treatment reduced COPB2 expression in overexpression and KO of COPB2 cells ( Figure  S12). To further confirm the effect of COPB2 expression, COPB2 KO cells were injected into nude mice. Results suggested that loss of expression of COPB2 inhibited tumor growth Relative expression of let-7a, H19 and COPB2 in Nc-mimics, mimics-let-7a, Nc-inhibitor and inhibitor-let-7a group using qPCR (E-G). The data are represented as the mean ± SD (n = 3). * (p <0.05), ** (p <0.01) and *** (p <0.001) indicate statistically significant differences.
( Figures 6G, H). Overall, these results indicate that reduced expression of COPB2 leads to anti-tumor effects.

DISCUSSION
Chrysin, a natural medicine, has anti-inflammatory and anticancer function and has been used to treat degenerative disorders and cancers in several Asian countries (26,27). In this study, chrysin was used to treat GC cells in order to evaluate its effect on cellular apoptosis, growth, migration and invasion. Previous reports have indicated that chrysin induces cell apoptosis and inhibits cell growth, migration and invasion in glioblastoma cells (28), which was by validated by our results, which showed that chrysin has anti-cancer effects in SGC7901, MKN45 and BGC823 cells. Moreover, chrysin was found to increase expression of miR-9 and let-7a in GC cells, in accordance with previous data (29). Interestingly, our previous data suggested that chrysin inhibited cell migration and invasion in MKN45 cells through TET1, which regulates global DNA methylation (30). Previous reports have indicated abnormal DNA methylation in GC (31). Herein, our results suggested that H19 DMR is hypomethylated in SGC7901 and BGC823 cell lines, which is related to increased expression of H19. In addition, chrysin functions to regulate the expression of H19, let-7a and COPB2. Furthermore, chrysin inhibited cell invasion which was overexpression of COPB2.
Recently, lncRNAs and miRNAs have been demonstrated to have roles in development of different types of cancer, including HCC (18). There is evidence that reduced expression of lncRNA H19 leads to inhibition of tumor growth in breast cancer, bladder cancer and colorectal cancer (32). Moreover, H19, as a competitive endogenous RNA, is associated with miRNAs, such as miR-29 and let-7 (16,33). Our results indicated that H19 and let-7a have competitive regulation in GC cells, which has also been confirmed in a previous report (17). Emerging evidence suggested that loss of expression of let-7 correlates with poor prognosis in various cancer (34). Our data showed that chrysin increased expression of let-7a and inhibited cell migration and invasion. Previous reports indicated that silencing of let-7, which targets MDM4, promotes cell proliferation, migration and invasion (35). Further, reduced expression of H19 and increased let-7a expression induced cell apoptosis in GC cells, which validated the previous report (36). These results suggested that the expression of H19 and let-7a is involved in cell apoptosis, growth and invasion of GC cells. Previous reports indicated that let-7f targets HMGA2 in thyroid cancer (37). Compared to let-7f, IL-6 and CKIP-1 were reported to be targets of let-7a (38,39). Our results suggest that let-7a targets COPB2, which leads to differential expression after chrysin treatment in GC cells. A previous study indicated that the expression of COPB2 is associated with cell growth, apoptosis, migration and invasion, functioning through a miR-216a manner, in lung cancer (8). Our results suggested that  expression of COPB2 was regulated by H19/let-7 axis in GC cells ( Figure 7). Moreover, reduced expression of COPB2 induced cellular apoptosis and inhibited cell growth in prostate cancer (40). Our data showed reduced expression of COPB2 increased p53 and E-cadherin expression. These results indicated that reduced expression of COPB2 induced cell apoptosis and inhibited invasion through H19/let-7a.
In order to confirm the effect of COPB2 in vivo, a xenograft model using nude mice was established. The in vivo results suggested that chrysin led to reduced expression of COPB2, which further confirms our data in GC cells. Furthermore, chrysin inhibited tumor growth in vivo, which is in accordance with a previous report in melanomas (41). These data indicate that chrysin can regulate COPB2 expression, which inhibits tumor growth in vivo. In order to further analyze the putative effect of COPB2, the KO and overexpression COPB2 cell line were injected into nude mice. As a powerful gene editing tool, the CRISPR/Cas9 system was wildly used in vitro and in vivo. Our results indicated that COPB2 KO suppressed tumor growth.
In summary, COPB2 is a differentially expressed gene that was identified after chrysin treatment in GC cells. Our data indicated that COPB2 is regulated by let-7a, which acts as a molecular sponge of H19. Moreover, reduced expression of COPB2 induced cellular apoptosis and inhibited cell growth and invasion. Therefore, this present study revealed that COPB2 is a potential molecular targeted therapy in GC.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material.