Colorectal cancer (CRC) is a prevalent aggressive neoplasm and an important reason of cancer-associated death globally.^1,2 Despite considerable advances have been acquired in CRC systemic therapy, most metastatic CRC patients die of their disease.^3,4 Hence, research on the mechanisms of CRC progression is essential for the pursuit of new therapeutic strategies.

Circular RNAs (circRNAs) are a cluster of non-coding RNAs with covalent circular enclosed loops.^5,6 Plentiful studies have expounded that circRNAs with abnormal expression exert a regulatory effect via targeting microRNAs (miRNAs).^7-9 For example, circ_001971 aggravated the vicious phenotypes of CRC via miR-29c-3p/VEGFA axis.¹⁰ Moreover, Chen et al.¹¹ declared that Homo sapiens circRNA_0003602 (hsa_circ_0003602; also termed as circSMARCC1) was strikingly overexpressed in CRC, and it could facilitate the progression of CRC via segregating miR-140-3p. Nonetheless, the support about the biological role of circ_0003602 in CRC pathogenesis is far from enough.

miRNAs are one category of small non-coding RNAs that serve as pivotal moderators in various malignancies by regulating their downstream targets.^12,13 Also, miRNAs serve a significant part in the progression of CRC and may work as effective diagnostic and therapeutic biomarkers for CRC.^14,15 Previous study had found the downregulation of miR-149-5p in CRC and it could inhibit CRC cell motility through targeting BGN.¹⁶ Besides, SLC38A1 was highly expressed in CRC and its silencing blocked CRC cell migration and proliferation.¹⁷ However, the relations among circ_0003602, miR-149-5p, and SLC38A1 are not identified as yet.

We aimed to assess the action of circ_0003602 in CRC. Moreover, the possible modulatory network of circ_0003602 in CRC was further probed. Hence, this study might offer novel curative target for CRC patients.

1. Collection of clinical samples

Clinical samples were harvested from CRC patients undergoing surgical resection at Affiliated Hospital of Chengdu University of Traditional Chinese Medicine, including 23 CRC samples and matching healthy samples. None of the participating patients received any pre-operative therapies. The permission was obtained from the Ethics Committee of Affiliated Hospital of Chengdu University of Traditional Chinese Medicine (approval number: 20210318), and written informed consents were signed by each participant. The analyzed data sets generated during the present study are available from the corresponding author on reasonable request.

2. Cell culture

CRC cells SW620 (#CL-0225; Procell, Wuhan, China), LoVo (#CL-0144, Procell), SW480 (#CL-0223, Procell), HCT-116 (#CL-0096, Procell) and a normal cell line NCM460 (#BFN608006385; Bluefbio, Shanghai, China) were fostered in DMEM medium (Procell) plus 10% fetal bovine serum (Procell) and 1% antibiotics (Procell). All cells were maintained at 37^℃ in a humid 5% CO2 atmosphere.

3. Cell transfection

siRNA or shRNA against circ_0003602 was used for circ_0003602 knockdown and homologous contrasts si-NC or sh-NC. MiR-149-5p inhibitor and mimic (anti-miR-149-5p or miR-149-5p), as well as matched contrasts anti-miR-NC or miR-NC were utilized for the silencing and overexpression of miR-149-5p. Overexpression vector of circ_0003602 and SLC38A1 (circ_0003602 and SLC38A1) and corresponding contrasts pCD5-ciR or pcDNA were used for circ_0003602 or SLC38A1 overexpression. siRNA against SLC38A1 (si-SLC38A1) and corresponding reference si-NC were utilized for SLC38A1 silencing. Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA) was utilized to transduce all these designated plasmids or oligonucleotides into CRC cells.

4. Quantitative real-time polymerase chain reaction

Total RNA segregation was done utilizing TRIzol (Invitrogen). Subsequently, PrimeScript RT Reagent Kit (Takara, Dalian, China) for circ_0003602 and SLC38A1 or PrimeScript miRNA real-time polymerase chain reaction (RT-PCR) Kit (Takara) for miR-149-5p was employed to product cDNA, followed by conducting quantitative RT-PCR (qRT-PCR) analysis with the BeyoFast™ SYBR Green qPCR Mix (Beyotime, Shanghai, China). The sequences of primers were provided in Table 1. Abundances of circ_0003602, miR-149-5p and SLC38A1 were analyzed using 2^-ΔΔCt strategy. GAPDH or U6 was regarded as inner contrast.

Table 1 . List of Primers Sequences for qRT-PCR.

Primers for qRT-PCR (5’-3’)
circ_0003602	Forward	ACCTACGGTCTGGAGAGCAG
	Reverse	GTCAAGTTCCTCCGACAAGC
miR-149-5p	Forward	GGCTCTGGCTCCGTGTCTT
	Reverse	CAGTGCAGGGTCCGAGGTATT
SLC38A1	Forward	CTTTGGAGCCACCTCTCTACAG
	Reverse	ACCAGGCTGAAAATGTCTCTTCC
GAPDH	Forward	GACAGTCAGCCGCATCTTCT
	Reverse	GCGCCCAATACGACCAAATC
U6	Forward	CTCGCTTCGGCAGCACA
	Reverse	AACGCTTCACGAATTTGCGT

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

5. RNase R assay

RNase R test was done via digestion of total RNA (20 μg) with RNase R (3 U/μg; Geneseed, Guangzhou, China). Half an hour later, the abundances of linear SLC38A1 and circ_0003602 were estimated.

6. Cell viability assay

At 48 hours after transfection, cells were co-incubated with CCK-8 solution (Beyotime) at 37^℃. After 4 hours of reaction, a microplate reader was employed to inspect the absorption at 450 nm.

7. Flow cytometry assay

CRC cells were harvested after 48 hours of transfection. To assess the apoptotic capacity of CRC cells, Annexin V-FITC Apoptosis Detection kit (Beyotime) was utilized following the operation manuals. Briefly, transduced CRC cells were suspended in phosphate-buffered saline and subsequently stained using Annexin V-FITC and propidium iodide for 15 minutes in darkness. The apoptotic proportion of CRC cells was monitored by a flow cytometer.

8. Transwell invasion assay

The transwell chamber (8 μm; Corning Costar, Corning, NY, USA) covered with Matrigel (Solarbio, Beijing, China) were employed for monitoring the invasive capacities of CRC cells. Briefly, transduced cells in non-serum medium were suspended and then seeded into the top surface of the chamber. While the bottom compartment was replenished with complete medium plus fetal bovine serum. Following 24 hours of culture, the cells invaded into the bottom compartment were subjected to the fixation of 4% paraformaldehyde (Beyotime). Following dying with crystal violet (0.1%, Solarbio), the invasive cells were counted using an inverted microscope (100×) from five randomly chosen regions.

9. Wound healing assay

In order to inspect the migratory capacity of CRC cells, wound healing assay was carried out. Transfected cells were kept in 24-well plates for 24 hours. Next, scratches were gently made on cell monolayers with sterilized pipette tips. Following removing the debris, the cells were cultivated for 24 hours. ImageJ software was used to calculate the widths of wounds (migrated distance) with 40× magnification after the images were obtained at 0 and 24 hours using an optical microscope.

10. Tube formation assay

To test angiogenic property of human umbilical vein endothelial cells, the culture supernatants of transfected CRC cells were gathered, and then co-incubated with human umbilical vein endothelial cells in a 96-well plate covered with Matrigel (Solarbio) for 6 hours. Finally, the generated tubular and branched structures were observed under an optical microscope (100×).

11. Western blot

Protein extraction was implemented employing RIPA buffer (Solarbio). The protein was segregated using SDS-PAGE gel (10%, Beyotime) and subsequently received the transference to PVDF membrane (Beyotime). Following sealing with 5% slim milk, the membrane was reacted with primary antibodies against c-Myc (1:1000, ab32072; Abcam, Cambridge, UK), SLC38A1 (ab272910, 1:2000, Abcam), matrix metalloproteinase 9 (MMP9) (ab137867, 1:1000, Abcam) and β-actin (1:200, ab115777, Abcam). Visualizing of the protein signals was made using the BeyoECL Star Kit (Beyotime) after incubation with a secondary antibody (ab205718, 1:30000, Abcam).

12. The determination of glutamine metabolism

The glutamine metabolism was assessed by checking the levels of glutamine consumption, α-ketoglutarate and glutamate production. The levels of glutamine, α-ketoglutarate production and glutamate production in the culture medium were measured with glutamine, α-ketoglutarate and glutamate assay kits (Abcam) following manufacturer’s instructions, respectively. Last, the supernatant from cells was analyzed at 565 nm for glutamine metabolism, at 450 nm for glutamine uptake or at 570 nm for α-ketoglutarate production using a microplate reader.

13. Bioinformatics analysis

The possible miRNA targets of circ_0003602 were predicted by starBase 2.0 (http://starbase.sysu.edu.cn/starbase2), Circinteractome (https://circinteractome.irp.nia.nih.gov), and circBank (http://www.circbank.cn) databases, while starBase 2.0 was used to predict the possible mRNA targets of miR-149-5p.

14. Dual-luciferase reporter assay

The pmirGLO vector (Promega, Madison, WI, USA) was utilized to form wild-type (WT)-circ_0003602, WT-SLC38A1 3’ untranslated region (3’UTR), mutant (MUT)-circ_0003602, and MUT-SLC38A1 3’UTR reporters, respectively. Next, the above reporters and miR-NC or miR-149-5p were transduced into CRC cells. After 48 hours of co-transfection, the luciferase intensity was inspected by Dual-Lucy Assay Kit (Solarbio).

15. Xenograft tumor experiment

HCT-116 cells (3×10⁶) in 200 μL of phosphate-buffered saline (Solarbio) stably transduced with sh-NC or sh-circ_0003602 were inoculated hypodermically into the right flanks of BALB/c nude mice (male, 4-week-old) (Vital River, Beijing, China) to establish the xenograft models in vivo. The volume of xenografts was computed every third day by the formula: length×width²/2. The mice were euthanatized at day 23 post-inoculation, and the transplanted neoplasms were excised and weighed. The animal research got the empowerment from the Animal Care and Use Committee of Chengdu University of Traditional Chinese Medicine.

16. Immunohistochemistry assay

The paraffin-embedded sections of internal tissues from transplanted mice were separated into slices (4 µm), followed by antigen retrieval and the reaction with primary antibodies against c-Myc with a dilution at 5 µg/mL (ab32072, Abcam), MMP9 with a dilution at 1:1000 (ab137867, Abcam), and Ki67 with a dilution at 1:200 (ab16667, Abcam). Then, a secondary antibody with a dilution at 1:2000 (ab205718, Abcam) was then used, and the slices subsequently were stained with 3, 3’-diaminobenzidine solution (Beyotime). Simultaneously, counterstaining of nuclei was done with hematoxylin (Beyotime). Finally, the representational areas were photographed using a light microscope with appropriate magnification.

17. Statistical analysis

The experiments were conducted with at least three replicates. All data were displayed by mean±standard deviation and processed by GraphPad Prism 6 software. The differences were compared via Student t-test (for two groups) or one-way analysis of variance (for more than two groups). Differences were considered statistically significant at p<0.05.

1. Circ_0003602 was overexpressed in CRC samples and cells

To inquire the action of circ_0003602 in CRC development, circ_0003602 expression was examined. The qRT-PCR results suggested that circ_0003602 level was distinctly raised in CRC tissues (tumor, n=23) (Fig. 1A). Also, the patients with high circ_0003602 level had the lower overall survival (p=0.0466) (Fig. 1B). CRC patients were divided into I-II group (n=12) and III group (n=11) based on TNM stage. We found that circ_0003602 expression was positively correlated with the TNM stage of CRC patients (Fig. 1C). Furthermore, the abundance of circ_0003602 in CRC cells (SW620, LoVo, SW480, and HCT-116) was overtly elevated (Fig. 1D). RNase R digestion assay exhibited that linear GAPDH was dramatically reduced after RNase R treatment, whereas circ_0003602 level was not affected (Fig. 1E and F), indicating that circ_0003602 was a stable circRNA with a loop structure. All these findings manifested that dysregulation of circ_0003602 might be concerned with CRC progression.

Figure 1. Circ_0003602 levels were raised in colorectal cancer (CRC) tissues and cells. (A) The expression of circ_0003602 in CRC tissues (tumor, n=23) and normal tissues (normal, n=23) was tested by quantitative real-time polymerase chain reaction (qRT-PCR). (B) The overall survival of CRC patients was analyzed in the circ_0003602 high-expression group (n=12) and circ_0003602 low-expression group (n=11). (C) CRC patients were divided into group I-II (n=12) and group III (n=11) based on TNM stage. The level of circ_0003602 in the two groups was shown. (D) qRT-PCR was performed to analyze the expression of circ_0003602 in CRC cells (SW620, LoVo, SW480, and HCT-116) and normal NCM460 cells. (E, F) The relative expression of circ_0003602 and linear GAPDH in SW480 and HCT-116 cells were detected by qRT-PCR after RNase R treatment. *p<0.01, ^†p<0.001, ^‡p<0.0001.

2. Circ_0003602 knockdown suppressed CRC cell progression, whereas circ_0003602 overexpression did the opposite

Subsequently, loss-of-function experiments were conducted to illuminate the action of circ_0003602. Expectedly, circ_0003602 level was remarkably declined in CRC cells after si-circ_0003602 transfection (Fig. 2A), demonstrating the high transfection efficacy of si-circ_0003602. As presented by CCK-8 assay and flow cytometry, cell viability was obviously inhibited and cell apoptosis was notably promoted by the downregulation of circ_0003602 in CRC cells (Fig. 2B and C). Besides, the indicators related to glutamine metabolism were detected. The results showed that glutamine consumption and the production of glutamate and α-ketoglutarate were all evidently downregulated by circ_0003602 silencing in CRC cells (Fig. 2D-F).

Figure 2. Circ_0003602 knockdown suppressed colorectal cancer cell viability and glutamine metabolism and induced cell apoptosis. SW480 and HCT-116 cells were transfected with si-NC or si-circ_0003602. (A) The expression of circ_0003602 in SW480 and HCT-116 cells after si-circ_0003602 transfection was determined by quantitative real-time polymerase chain reaction. (B) The cell viability was examined using a CCK-8 assay. (C) Cell apoptosis was assessed using flow cytometry. The glutamine metabolism was monitored by measuring glutamine consumption (D), α-ketoglutarate (E), and glutamate production (F) using the corresponding kits. *p<0.01, ^†p<0.001, ^‡p<0.0001.

In addition, interference of circ_0003602 curbed CRC cell invasion and migration, respectively (Supplementary Fig. 1A and B). Human umbilical vein endothelial cell tube formation assay exhibited that the number of tubules induced by circ_0003602-knockdown CRC cells was conspicuously lessened (Supplementary Fig. 1C). Moreover, the levels of proliferation-related protein c-Myc and motility-related protein MMP9 in CRC cells were noticeably declined after si-circ_0003602 introduction (Supplementary Fig. 1D and E). Taken together, these data illustrated that circ_0003602 downregulation impeded CRC cell the malignant behaviors.

On the contrary, circ_0003602 overexpression elevated CRC cell viability and promoted cell invasion, migration, angiogenesis, and glutamine metabolism (Supplementary Fig. 2), which further highlighted that circ_0003602 played an oncogenic role in CRC cells.

3. Knockdown of circ_0003602 inhibited tumor growth of CRC in vivo

Xenograft mice models were established to clarify the action of circ_0003602 in CRC tumorigenesis in vivo. As illustrated in Fig. 3A and B, tumor volume and weight were lessened in the sh-circ_0003602 group compared to the control group. Furthermore, circ_0003602 expression was apparently downregulated in harvested tumor tissues of sh-circ_0003602 group (Fig. 3C). Besides, immunohistochemistry assay revealed that the expression of proliferation-related marker Ki-67 and c-Myc and motility-related marker MMP9 were markedly reduced after circ_0003602 deficiency (Fig. 3D). These evidences certified that circ_0003602 knockdown lowered CRC growth in vivo.

Figure 3. Circ_0003602 knockdown restrained tumor growth of colorectal cancer in vivo. HCT-116 cells stably introduced with sh-circ_0003602 or sh-NC were subcutaneously inoculated into nude mice. (A) The tumor volume was monitored every 3 days from day 8 after inoculation. (B) The tumor weight was examined after 23 days when the tumor tissues were excised. (C) The level of circ_0003602 in the harvested tumor tissues was determined by quantitative real-time polymerase chain reaction. (D) An immunohistochemistry assay for the expression of Ki-67, c-Myc and MMP9 was conducted (×200). *p<0.01, ^†p<0.001, ^‡p<0.0001.

4. MiR-149-5p was sponged by circ_0003602

Next, the downstream miRNAs targeted by circ_0003602 were explored in CRC cells. The miRNAs that possessed the possible binding sites for circ_0003602 were predicted through Circinteractome, circBank and starBase 2.0 databases. From Venn diagram analysis of three overlapping circles, it was found that miR-942-5p and miR-149-5p were the shared miRNAs that had the potential complementary sequence for circ_0003602 (Fig. 4A). Inhibition of circ_0003602 led to the upregulation of miR-149-5p (Fig. 4B and C), so miR-149-5p was selected for subsequent study. The potential binding sites between circ_0003602 and miR-149-5p were exhibited in Fig. 4D. Meanwhile, dual-luciferase reporter assay attested that miR-149-5p overexpression overtly suppressed the luciferase intensity of WT-circ_0003602 reporter (Fig. 4E and F). Furthermore, an overt reduction in miR-149-5p expression was got in CRC tissues (Fig. 4G). Interestingly, the circ_0003602 level was negatively correlated with miR-149-5p level in CRC tissues (Fig. 4H). Consistent with the result in tissues, miR-149-5p was downregulated in CRC cells relative to that in normal NCM460 cells (Fig. 4I). In addition, the circ_0003602 overexpression vector was used for upregulation of circ_0003602 (Fig. 4J). Expectedly, circ_0003602 overexpression could obviously inhibit the expression of miR-149-5p in CRC cells (Fig. 4K). Furthermore, we found that circ_0003602 knockdown-induced upregulation of miR-149-5p could be restored by the addition of circ_0003602 plasmid in CRC cells (Supplementary Fig. 3). Collectively, above data manifested that circ_0003602 could modulate miR-149-5p expression.

Figure 4. MiR-149-5p directly interacted with circ_0003602 in colorectal cancer (CRC) cells. (A) Venn diagram representing the potential miRNAs with possible binding sites of circ_0003602 predicted by the Circinteractome, circBank and starBase 2.0 databases. (B, C) The relative expression of miR-942-5p and miR-149-5p in SW480 and HCT-116 cells transfected with si-circ_0003602 was tested by quantitative real-time polymerase chain reaction (qRT-PCR) analysis. (D) Schematic diagram exhibiting the target binding sites between circ_0003602 and miR-149-5p. (E, F) A dual-luciferase reporter assay was conducted to examine the luciferase activity in SW480 and HCT-116 cells co-transfected with WT-circ_0003602 or MUT-circ_0003602 and miR-149-5p or miR-NC. (G) The expression of miR-149-5p in CRC tissues and normal tissues was detected by qRT-PCR assay. (H) The correlation between circ_0003602 and miR-149-5p in CRC tissues was analyzed by Spearman correlation coefficient analysis. (I) The abundance of miR-149-5p in NCM460, SW480, and HCT-116 cells was evaluated by qRT-PCR. (J) The overexpression efficiency of circ_0003602 in SW480 and HCT-116 cells was determined by qRT-PCR assay. (K) The expression of miR-149-5p in SW480 and HCT-116 cells introduced with pCD5-ciR or circ_0003602 was detected by qRT-PCR assay. *p<0.05, ^†p<0.01, ^‡p<0.001, ^§p<0.0001.

5. Circ_0003602 knockdown suppressed CRC cell progression by binding to miR-149-5p

Rescue experiments were performed to investigate whether circ_0003602 could regulate CRC cell progression by targeting miR-149-5p. Circ_0003602 interference dramatically raised miR-149-5p expression, whereas the introduction of anti-miR-149-5p partly reversed the effect (Fig. 5A). Functionally, circ_0003602 knockdown-mediated impacts on CRC cell viability (Fig. 5B), apoptosis (Fig. 5C), invasion (Fig. 5D), migration (Fig. 5E), and tube formation capacity (Fig. 5F) were impaired by miR-149-5p inhibition. Consistently, circ_0003602 knockdown reduced c-Myc and MMP9 protein levels in CRC cells, which were largely overturned by silencing miR-149-5p (Fig. 5G and H). Additionally, circ_0003602 downregulation-induced suppression on glutamine metabolism (Fig. 5I-5K) were effectively weakened by anti-miR-149-5p introduction. To sum up, circ_0003602 adsorbed miR-149-5p to regulate CRC cell malignant phenotypes.

Figure 5. Circ_0003602 knockdown retarded colorectal cancer cell progression by binding to miR-149-5p. SW480 and HCT-116 cells were transfected with si-NC, si-circ_0003602, si-circ_0003602+anti-miR-NC or si-circ_0003602+anti-miR-149-5p. (A) The level of miR-149-5p was examined by quantitative real-time polymerase chain reaction. Cell viability (B), apoptosis (C), invasion (D), migration (E), and angiogenesis (F) were, respectively examined by CCK-8 assay, flow cytometry, transwell invasion assay, wound healing assay and tube formation assay. (G, H) Western blot was executed for detecting the expression of c-Myc and MMP9 in SW480 and HCT-116 cells. (I-K) The glutamine metabolism was monitored by corresponding glutamine, α-ketoglutarate and glutamate assay kits. *p<0.05, ^†p<0.01, ^‡p<0.001, ^§p<0.0001.

6. SLC38A1 was targeted by miR-149-5p

SLC38A1 was found to have complementary sequence with miR-149-5p using online tool starBase 2.0 (Fig. 6A). Subsequently, transient transfection of WT-SLC38A1 3’UTR in CRC cells in the presence of miR-149-5p visibly inhibited the luciferase activity, whereas this effect was dramatically abrogated in MUT-SLC38A1 3’UTR group (Fig. 6B and C). Likewise, SLC38A1 mRNA expression levels were overtly higher in CRC tissues (Fig. 6D), and it was inversely correlated with miR-149-5p level in CRC tissues (Fig. 6E). Also, SLC38A1 protein levels were higher in CRC tissues than that in adjacent normal tissues (Fig. 6F). Also, SLC38A1 mRNA expression in CRC tissues was positively correlated with TNM stage (Fig. 6G), implying the critical role of SLC38A1 in CRC progression. Consistently, SLC38A1 protein levels were higher in two CRC cell lines than that in the NCM460 cell line (Fig. 6H). In addition, miR-149-5p or anti-miR-149-5p was used for overexpression or silencing of miR-149-5p (Fig. 6I). Moreover, miR-149-5p upregulation resulted in a decrease in the SLC38A1 protein level, but miR-149-5p inhibition had a reverse function (Fig. 6J). Therefore, miR-149-5p regulated SLC38A1 expression via sponging SLC38A1 in CRC cells.

Figure 6. SLC38A1 was a direct target gene of miR-149-5p. (A) The putative binding sites between miR-149-5p and SLC38A1 3’ untranslated region (3'UTR) are shown. (B, C) The combination between miR-149-5p and SLC38A1 was verified by dual-luciferase reporter assay. (D) The mRNA level of SLC38A1 in colorectal cancer (CRC) tissues and normal tissues was determined by quantitative real-time polymerase chain reaction (qRT-PCR) assay. (E) The correlation between the levels of miR-149-5p and SLC38A1 mRNA was analyzed by Spearman correlation coefficient analysis. (F) The protein level of SLC38A1 in CRC tissues and normal tissues was measured by Western blot assay. (G) CRC patients were divided into group I-II (n=12) and group III (n=11) based on TNM stage. The mRNA expression of SLC38A1 in two groups was shown. (H) Western blot assays were performed to measure the protein level of SLC38A1 in NCM460, SW480, and HCT-116 cells. The levels of miR-149-5p (I) and SLC38A1 protein (J) in SW480 and HCT-116 cells transfected with miR-NC, miR-149-5p, anti-miR-NC or anti-miR-149-5p were measured by qRT-PCR assay and Western blot assay, respectively. *p<0.01, ^†p<0.001, ^‡p<0.0001.

7. SLC38A1 knockdown restrained CRC cell malignant phenotypes

Subsequently, the function of SLC38A1 in CRC cell development was surveyed. As depicted in Fig. 7A, the obvious reduction of SLC38A1 expression in CRC cells after si-SLC38A1 transfection was observed, indicating the successful knockdown of SLC38A1. SLC38A1 knockdown restrained CRC cell viability, invasion, migration, angiogenesis, and glutamine metabolism and induced CRC cell apoptosis (Fig. 7B-K), indicating that SLC38A1 absence reduced the malignant potential of CRC cells.

Figure 7. SLC38A1 silence impeded the malignancy of colorectal cancer cells. SW480 and HCT-116 cells were transduced with si-NC or si-SLC38A1. (A) A Western blot assay was adopted to measure the protein level of SLC38A1. (B, C) Cell viability and apoptosis were, respectively examined by CCK-8 assay and flow cytometry. (D, E) Cell invasion and migration capacities were respectively checked by transwell invasion assay and wound healing assay. (F) The tube formation rate was tested by tube formation assay. (G, H) The expression of c-Myc and MMP9 was assessed by Western blot. (I-K) Glutamine metabolism was monitored by the corresponding glutamine, α-ketoglutarate, and glutamate assay kits. *p<0.01, ^†p<0.001, ^‡p<0.0001.

8. miR-149-5p impeded CRC progression largely by targeting SLC38A1 expression

Considering the direct target relationship between miR-149-5p and SLC38A1, we conducted rescue experiments to analyze their functional relationship. miR-149-5p overexpression reduced SLC38A1 protein levels, which were largely recovered by the addition of SLC38A1 (Fig. 8A). Forced miR-149-5p expression suppressed CRC cell viability and promoted CRC cell apoptosis, and these effects could be largely alleviated by SLC38A1 overexpression (Fig. 8B and C). Upregulated miR-149-5p restrained CRC cell invasion, migration, and angiogenesis abilities, but these impacts were largely counteracted after SLC38A1 overexpression (Fig. 8D-F). Increased miR-149-5p expression reduced c-Myc and MMP9 protein levels, while SLC38A1 upregulation largely rescued the changed protein levels (Fig. 8G and H). miR-149-5p overexpression-induced inhibitory effect on the glutamine metabolism of CRC cells was largely overturned by SLC38A1 overexpression (Fig. 8I-K). These results showed that miR-149-5p targeted SLC38A1 to curb CRC cell malignant behaviors. Furthermore, circ_0003602 silencing reduced SLC38A1 protein levels, whereas repressed miR-149-5p expression impaired this impact (Fig. 8L), suggesting that circ_0003602 interacted with miR-149-5p to mediate SLC38A1 expression.

Figure 8. SLC38A1 addition attenuated miR-149-5p overexpression-mediated inhibitory influences on the progression of colorectal cancer cells. SW480 and HCT-116 cells were transfected with miR-NC, miR-149-5p, miR-149-5p+pcDNA, or miR-149-5p+SLC38A1. The protein expression of SLC38A1 (A), cell viability (B), apoptosis (C), invasion (D), migration (E), and angiogenesis (F) were respectively analyzed by Western blot, CCK-8 assay, flow cytometry, transwell invasion assay, wound healing assay and tube formation assay. (G, H) The protein expression of c-Myc and MMP9 was examined by Western blot. (I-K) Glutamine metabolism was assessed by using the corresponding glutamine, α-ketoglutarate, or glutamate assay kits. (L) The protein abundance of SLC38A1 in SW480 and HCT-116 cells transfected with si-NC, si-circ_0003602, si-circ_0003602+anti-miR-NC, or si-circ_0003602 + anti-miR-149-5p was checked by Western blot. *p<0.05, ^†p<0.01, ^‡p<0.001, ^§p<0.0001.

Presently, emerging studies have exposed that circRNAs serve significant roles in the biologic courses of diversiform malignancies via modulating multiple pathways,^18,19 including CRC.²⁰ Here, we confirmed the enhanced level of circ_0003602 in CRC and circ_0003602 silencing restrained CRC advancement by modulating the miR-149-5p/SLC38A1 axis.

An increasing body of proofs have attested that dysregulated circRNAs are connected with the advancement of CRC.²¹ For instance, circASS1 overexpression inhibited the CRC cell malignant phenotypes through regulating the miR-1269a/VASH1 pathway.²² And circ_0007334 deficiency restrained CRC cell growth, motility and angiogenesis via regulation of the miR-577/KLF12 axis.²³ However, the study about the function of circ_0003602 in CRC etiology is still restricted and requires further investigation. Previous research has reported that circ_0003602 was upregulated in acute myeloid leukemia.²⁴ Similarly, circ_0003602 abundance was raised in CRC, and it could regulate CRC cell advancement via sponging miR-140-3p.¹⁷ Here, we expounded that circ_0003602 had been elevated in CRC. Furthermore, we unmasked that circ_0003602 knockdown repressed CRC cell viability, migration, invasion, angiogenesis and glutamine metabolism but accelerated CRC cell apoptosis. Moreover, circ_0003602 downregulation also lessened tumor growth in vivo. All finding urged us to speculate that circ_0003602 served as a carcinogenic factor in CRC.

CircRNAs have been proved to modulate targeted gene expression by serving as the sponges of miRNAs.²⁵ Therefore, the potential downstream miRNAs of circ_0003602 were probed. Through Venn diagram analysis of Circinteractome, circBank and starBase 2.0, miR-942-5p and miR-149-5p were verified as the common miRNA candidates that had the potential complementary sites with circ_0003602. Therein, miR-149-5p was chosen for further research for the higher level of miR-149-5p in CRC cells after circ_0003602 deficiency. As reported, miR-149-5p exerted a pivotal anti-cancer factor in diverse malignancy, such as cervical cancer,²⁶ gastric cancer²⁷ and papillary thyroid cancer.²⁸ Likewise, miR-149-5p expression was lower in CRC, and forced miR-149-5p expression played a suppressive role in CRC progression.²⁹ In keeping with this study, we confirmed that miR-149-5p abundance was decreased in CRC tissues and cells. Also, introduction of miR-149-5p had repressive impacts on CRC cell vicious evolution. Besides, circ_0003602 downregulation-mediated repression of CRC cell development was effectively overturned by miR-149-5p silencing. These findings highlighted that circ_0003602 interacted with miR-149-5p to mediated CRC progression.

Next, we further probed into the latent mechanism of circ_0003602. Our results identified miR-149-5p as a molecular decoy for SLC38A1. SLC38A1 has been testified to have elevated expression and play an important part in multifarious malignancies.^30,31 Also, Yu et al.³² claimed that circRUNX1 deficiency suppressed the growth, motility, and glutaminolysis and induced apoptosis through the miR-485-5p/SLC38A1 axis. Here, we discovered that SLC38A1 was raised in CRC. Moreover, SLC38A1 deficiency impeded CRC cell malignant development. Additionally, the repressive influence of miR-149-5p addition on CRC cell malignant characteristics were offset by SLC38A1 reintroduction, hinting that miR-149-5p restrained CRC cell malignancy via absorbing SLC38A1. Besides, circ_0003602 could control SLC38A1 level through competitively combining with miR-149-5p. Collectively, circ_0003602 knockdown blocked CRC development by targeting miR-149-5p and suppressing SLC38A1 expression.

In conclusion, circ_0003602 was elevated in CRC, and it could facilitate the viability, migration, invasion, angiogenesis, and glutamine metabolism and hamper the apoptosis of CRC cells via administering the miR-149-5p/SLC38A1 axis. This novel regulatory axis might supply a hopeful curative strategy for CRC remedy.

Supplementary materials can be accessed at https://doi.org/10.5009/gnl210542.

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

Study concept and design: S.T. Data acquisition: Q.W. Data analysis and interpretation: P.K. Drafting of the manuscript: F.L. Critical revision of the manuscript for important intellectual content: R.W., S.T. Statistical analysis: Q.W. Administrative, technical, or material support; study supervision: R.W., S.T. Approval of final manuscript: all authors.