Patients with CD and healthy donors (HDs) were recruited from 2018 to 2019 at Suzhou Affiliated Hospital of Nanjing Medical University (Suzhou, Jiangsu Province, China). Ethical approval was obtained from the Ethics Committee of Nanjing Medical University. Informed consent was obtained from all participants. Specimens were collected using the method described in our previous study (Yin et al., 2019). Information on all participants is listed in Table 1.

Total RNA was isolated from PBMCs using the TRIzol Regent (Invitrogen, United States). Complementary DNA (cDNA) was synthesized with PrimeScript RT Master Mix (TaKaRa, Shiga, Japan). The relative expression of hsa_circRNA_103124 in PBMCs was detected using RT-QPCR, with β-actin as an internal reference, according to the methods described in our previous studies (Yin et al., 2019; Yin et al., 2020). The relative expression of hsa-miR-650 in PBMCs was detected using RT-QPCR with U6 as an internal reference. In brief, cDNA was synthesized from total RNA with the aid of MMLV Reverse Transcriptase in Hairpin-it™ miRNAs RT-PCR Quantitation Kit (TaKaRa, Shiga, Japan). And RT-QPCR was performed using 2xReal-time PCR Master Mix (SYBR) in Hairpin-it™ miRNAs RT-PCR Quantitation Kit. 2^−△△Ct method was used to analyze the relative expression of hsa_circRNA_103124 or hsa-miR-650. The primer sequences are listed in Table 2.

Peripheral blood (1.5 ml) was collected from each patient with CD or healthy control, using a vacuum blood collection tube with EDTA-K2. The WBC count in peripheral blood was detected using an automated hematology analyzer (Sysmex XN-1000, Kobe, Japan). Fresh stool samples (500 mg) were collected from patients with CD. Fecal CALP levels were measured using enzyme-linked immuno sorbent assay (ELISA) (WIZ, BIOTECH, Xiamen, China). Please refer to the manual for specific operational procedures.

The interaction between circRNA and microRNA was predicted with the Arraystar homemade miRNA target prediction software based on the TargetScan and miRanda databases. The 5 most likely miRNAs to which hsa_circRNA_103124 binds were predicted based on the arrangement of free energy and number of seed sequences. The intersection of the downstream genes of the target miRNAs in the Targetscan, miRDB and miRWalk databases was obtained using Venny 2.1. Gene Ontology (GO) and KEGG pathway analyses were performed to predict the biological function of hsa_circRNA_103124 in DAVID 6.7. Subsequent pairing of target miRNAs and genes was predicted using TargetScan.

Caco2 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM), and human intestinal epithelial cells (HIECs) were cultured in Roswell Park Memorial Institute 1,640 medium (RPMI 1640 medium), at 37°C with 5% CO2. Over-expression of hsa_circRNA_103124 was induced by transiently transfecting overexpression plasmid pLC5-ciR-circRNA_103124 (Geneseed, Guangzhou, China) in Caco2 cells and HIECs. The downregulation of hsa_circRNA_103124 was induced by transfecting a siRNA (Table 2). The overexpression or downregulation of hsa-miR-650 was induced by RNA mimics or inhibitor (Table 2). Lipofectamine™3000 Reagent (Invitrogen, CA, United States) was used to for the transient transfection of plasmids, siRNA, RNA mimics or inhibitor. In addition, detailed procedures are described in the manual. Relative expression level of hsa_circ_103124 or hsa-miR-650 in Caco2 or HIECs with different treatments in this study were detected by RT-QPCR.

The cell cycle distribution was detected using flow cytometry (FACSCalibur, BD, San Jose, CA, United States) after propidium iodide (PI, Beyotime, Shanghai, China) staining. Cell proliferation was examined using Cell Counting Kit-8 assay (Dojindo, Shanghai, China) and 5-ethynyl-2′-deoxyuridine staining (RiboBio, Guangzhou, Guangdong, China). The detailed methods are described in our previous study (Yin et al., 2020).

Fluorescence in situ hybridization (FISH) was performed to detect the location and expression levels of hsa_circRNA_103124 and hsa-miR-650 using FAM- and CY3-labeled specific probes, respectively (GenePharma, Shanghai, China), in Caco2 cells or HIECs. Please refer to the manual for specific operational procedures. A confocal laser scanning microscope (Carl Zeiss, Göschwitzer Strasse, Jena, Germany) was used to capture the cellular fluorescence images. The sequences of probes used in this experiment are listed in Table 2.

HEK293T cells were cultured in 24-well plates with DMEM, at a density of 2 × 10⁴ cells/ml. Cells were incubated in a humidified atmosphere at 37°C with 5% CO2. Three hundred 1 bp of the hsa-miR-650 response element in hsa_circRNA_103124 and its adjacent sequences were recombined (GenScript, Nanjing, Jiangsu) with psiCHECK-2 (Promega, Madison, WI, United States). A control plasmid including mutated sequences of the hsa-miR-650 response element was inserted into psiCHECK-2. The plasmid and the hsa-miR-650 mimics were transfected into HEK293T cells using Lipofectamine 3,000 Reagent, according to the manual. The Dual-Luciferase® Reporter Assay System (Promega, Madison, WI, United States) was used to detect the activities of Renilla luciferase and firefly luciferase with Biotek Synergy H1(Biotek, Vermont, United States), at 48 h following transfection.

HIECs or Caco2 cells were grown on coverslips in 24-well plates with RPMI 1640 or DMEMmedium at a density of 2 × 10⁴ cells/ml for 24 h. Rapamycin (20 ng/ml) was supplemented to induce autophagy for 24 h. The cells were stained with acridine orange (AO, 2 μg/ml) for 30 min. Fluorescent images of the AO positive cells (orange) were acquired using a confocal laser scanning microscope. The proportion of positively stained cells that underwent autophagy was calculated by normalization to the cells stained green.

The western-blotting methods were described in our previously published paper (Yin et al., 2020). Primary antibodies against the following proteins were used in this study: AKT2 (CST-3063S, Cell Signaling Technology, Danvers, MA, United States), CDK2(CST-2546S, Cell Signaling Technology, Danvers, MA, United States), TSC1(CST-6935S, Cell Signaling Technology, Danvers, MA, United States), microtubule associated protein 1 light chain 3 beta (LC3B, CST-2775S, Cell Signaling Technology, Danvers, MA, United States), and β-actin (CST-8457S, Cell Signaling Technology, Danvers, MA, United States).

GraphPad Prism 5 (GraphPad, La Jolla, CA, United States) was used for data analysis in this study. p < 0.05 was considered as a standard to determine statistical significance. Differences between groups were determined using Student’s t-test (between groups) or one-way analysis of variance (ANOVA) (among groups). Pearson’s correlation coefficients were used to describe correlations between variables.

Hsa_circRNA_103124 was discovered as a potential diagnostic biomarker of CD in our previous study (Yin et al., 2019). The specific correlation between hsa_circRNA_103124 and clinical diagnostic markers of CD was further studied. Hsa_circRNA_103124 was upregulated in patients with CD, which was validated in PBMCs by RT-QPCR (Figure 1A). A correlation analysis was performed between hsa_circRNA_103124 and the WBC count (Figure 1B) or CALP level (Figure 1C) to investigate the possible mechanism underlying the upregulation of hsa_circRNA_103124 in patients with CD. The results showed positive correlations between the expression of hsa_circRNA_103124 in peripheral blood mononuclear cells and WBC counts (r = 0.3052, p = 0.0330) or CALP (r = 0.3557, p = 0.0243). The WBC count and CALP level are relatively preferred noninvasive biomarkers for CD diagnosis and treatment as we mentioned in the introduction. Based on this result, hsa_circRNA_103124 potentially served as a diagnostic biomarker for CD.

According to the MRE analysis, the 5 miRNAs (hsa-miR-650, hsa-miR-548a-5p, hsa-miR-501-5p, hsa-miR-369-5p and hsa-miR-188-5p) that most likely bind to hsa_circRNA_103124 are listed in Figure 2A. In addition, according to the arrangement of free energy and number of seed sequences, hsa-miR-650 is the most likely miRNA. Therefore, FISH was performed to reveal the location and expression of hsa_circRNA_103124 and hsa-miR-650 in HIECs and Caco2 cells. The results ( Figure 2B ) showed that hsa_circRNA_103124 and hsa-miR-650 localized in both the cytoplasm and nucleus. When we reduced the expression of hsa_circRNA_103124 transfecting a siRNA in cells, the localization of hsa-miR-650 in the cytoplasm was enhanced. Moreover, luciferase reporter assays (Figures 2C,D) indicated the specificity of the hsa-miR-650 mimics for binding to the hsa_miR_650 response elements in hsa_circRNA_103124. HEK293T cells cotransfected with psiCHECK-2 containing a wild-type hsa-miR-650 MRE of hsa_circRNA_103124 and hsa-miR-650 mimics showed lower relative luminescence intensity. However, cells cotransfected with psiCHECK-2, containing a mutated hsa-miR-650 MRE of hsa_circRNA_103124 and hsa-miR-650 mimics showed no significant change in the relative luminescence intensity. Therefore, we inferred that hsa-miR-650 bound to the MRE of hsa_circRNA_103124. In patients with CD, the relative expression of hsa-miR-650 was downregulated (p = 0.043) compared with that in healthy controls. Moreover, hsa_circRNA_103124 expression negatively correlated with hsa-miR-650 expression (Figure 2E). Therefore, we inferred that hsa_circRNA_103124 might play a negative regulatory role in hsa-miR-650 expression. Correspondingly, the function of downstream proteins of hsa-miR-650 could be regulated.

GO and KEGG pathway analyses indicated that hsa-miR-650 participates in pathways in cancer and TNF signaling (Figure 3A). The Venn diagram showed 551 common downstream genes of hsa-miR-650 identified in Targetscan, miRDB and miRWalk databases (Figure 3B), which contributed to the abovementioned pathways in which hsa-miR-650 played roles. Four binding sites for hsa-miR-650 in 3′UTR of AKT2 were predicted by TargetScan (Figure 3C). The protein expression levels of AKT2 and CDK2 were down-regulated in HIECs and Caco2 in response to interference with hsa_circRNA_103124 expression by siRNAs (Figure 3D). AKT2 and CDK2 were upregulated following hsa_circRNA_103124 overexpression (Figure 3E). TSC1 (a downstream protein of AKT2) was expressed at lower levels in patients with CD (Figure 3F). Considering the aforementioned results that hsa_circRNA_103124 targeted hsa-miR-650, we inferred that hsa_circRNA_103124 regulated AKT2 and its downstream proteins CDK2 and TSC1 by sponging hsa-miR-650. However, further studies should be performed on the specific roles that hsa-miR-650 may play in the functions of hsa_circRNA_103124.

One of the most important biological functions of AKT2 is regulation of proliferation. Therefore, 5-Ethynyl-2′-deoxyuridine (EdU) staining, Cell Counting Kit-8 (CCK-8) assay and the cell cycle distribution test were performed to study the status of cell proliferation following changes in hsa_circ_103124 expression. In EdU staining, thymine in the DNA of proliferating cells was replaced with EdU, which was stained by Apollo, as shown in red. The ratio of red to blue cells, represents the percentage of proliferating cells. In the CCK-8 assay, the OD450 absorbance value showed cell viability. The cell cycle distribution was analyzed using flow cytometry revealed the proportion of cells in G1, S or G2 phases.

The expression of hsa_circRNA_103124 was downregulated in Caco2 cells or HIECs transfected with hsa_circRNA_103124 siRNAs (Supplementary Figure S1A). The results of EdU staining (Figure 4A), the CCK-8 assay (Figure 4B), and cell cycle distribution detection using flow cytometry (Figure 4C) indicated that hsa_circRNA_103124 downregulation in Caco2 cells and HIECs inhibited cell proliferation. A lower ratio of red to blue cells was observed by EdU staining (Figure 4A). A smaller OD450 absorbance value was detected at each time point (Figure 4B). Notably, siRNA mediated depletion of hsa_circRNA_103124 led to an increase in the numbers of Caco2 and HIEC cells in G1 phase and a concomitant decrease in the numbers of cells in S and G2 phases (Figure 4C). Furthermore, western-blotting showed that the inhibition of hsa_circRNA_103124 was accompanied by downregulation of AKT2 and CDK2 (Figure 4D). Meanwhile, a rescue experiment was carried out by cotransfecting of a hsa-miR-650 inhibitor with the hsa_circRNA_103124 siRNA. Hsa-miR-650 inhibitor couldn’t rescue the expression of hsa_circRNA_103124 (Supplementary Figure S1A). And there was no stable significant change of relative expression of hsa-miR-650 in Caco2 cells or HIECs transfected with hsa_circRNA_103124 siRNAs (Supplementary Figure S1B). However, the results of EdU staining (Figure 4A), the CCK-8 assay (Figure 4B) and cell cycle distribution detection using flow cytometry (Figure 4C) suggested that the hsa-miR-650 inhibitor reversed the effect hsa_circRNA_103124 on restoring cell proliferation. Moreover, the expression of AKT2 and CDK2 was upregulated by the hsa-miR-650 inhibitor (Figure 4D), with decreased expression of hsa-miR-650 in Caco2 cells and HIECs (Supplementary Figure S1C). Nevertheless, there was no significant change of relative expression of hsa_circRNA_103124 in Caco2 cells and HIECs transfected with hsa-miR-650 inhibitor (Supplementary Figure S1D). Therefore, we considered that CDK2, one of the G1 phase regulators, was regulated by AKT2 and its upstream molecules hsa-miR-650/hsa_circRNA_103124. The hsa-miR-650 inhibitor rescued the inhibitory effect of the hsa_circRNA_103124 siRNA on cell proliferation, along with corresponding changes of AKT2 and CDK2 in expression levels.

We confirmed that hsa_circRNA_103124 overexpression in Caco2 cells and HIECs promoted proliferation by performing EdU staining (Figure 5A), the CCK-8 assay (Figure 5B) and determining the cell cycle distribution (Figure 5C). With hsa_circRNA_103124 overexpressed (Supplementary Figure S1E), a higher ratio of red to blue cells was observed by EdU staining (Figure 5A). A greater OD450 absorbance value was detected at each time point (Figure 5B). The relative cell numbers of Caco2 cells and HIECs in G1 phase were decreased. While the relative cell numbers of Caco2 cells and HIECs in S and G2 phases were increased (Figure 5C). Furthermore, the expression of AKT2 and CDK2 was upregulated following the overexpression of hsa_circRNA_103124 (Figure 5D). Correspondingly, hsa-miR-650 mimics were cotransfected with hsa_circRNA_103124 in Caco2 cells and HIECs in rescue experiments. Hsa-miR-650 mimics couldn’t reverse the upregulation of hsa_circRNA_103124 in cells transfected with plc-CIR-circRNA_103124 (Supplementary Figure S1E). And the expression of hsa-miR-650 wasn’t increased with hsa_circRNA_103124 overexpressed (Supplementary Figure S1F). However, the expression of hsa_circRNA_103124 was down-regulated in Caco2 cells and HIECs transfected with hsa-miR-650 mimics merely (Supplementary Figures S1G,H). The results of EdU staining (Figure 5A), the CCK-8 assay (Figure 5B) and cell cycle distribution detection using flow cytometry (Figure 5C) indicated that the hsa-miR-650 mimics reversed the increase in cell proliferation induced by the overexpression of hsa_circRNA_103124. In addition, the expression of AKT2 and CDK2 was downregulated by hsa-miR-650 mimics. Meanwhile, hsa-miR-650 mimics reversed the upregulation effect of hsa_circRNA_103124 on AKT2 and CDK2 (Figure 5D). Therefore, overexpression of hsa_circRNA_103124 promoted proliferation of Caco2 cells and HIECs. While hsa-miR-650 mimics could rescue this effect. And we inferred that hsa_circRNA_103124 regulated the functions of hsa-miR-650 in a ceRNA manner, rather than by regulating the expression of hsa-miR-650.

Impaired autophagy plays an important role in the pathogenesis of IBD (Larabi et al., 2020). AKT2 is one of the most important proteins that regulate autophagy (Xie et al., 2020). TSC1 and LC3B are downstream proteins of AKT2 (Satyavarapu et al., 2018; Xie et al., 2020). Therefore, autophagy was induced by rapamycin (20 ng/ml) in Caco2 cells and HIECs to study the function of hsa_circRNA_103124 on autophagy. In the present study, hsa_circRNA_103124 downregulation by siRNAs promoted rapamycin-induced autophagy, as shown by significantly increased numbers of AO-positive particles (orange), (Figure 6A). Moreover, AKT2 expression was downregulated. The expression of TSC1 and LC3B was upregulated (Figure 6B). However, the hsa-miR-650 inhibitor reversed these effects, leading to opposite effects on protein expression (Figure 6B) and decreased numbers of AO-positive particles (Figure 6A).

The overexpression of hsa_circRNA_103124 in Caco2 cells and HIECs inhibited rapamycin-induced autophagy and reduced the number of AO-positive particles (Figure 6C). The expression of AKT2 expression was upregulated, while the expression of TSC1 and LC3B was downregulated (Figure 6D). However, when hsa-miR-650 mimics were cotransfected, autophagy was restored, as shown by significantly increased numbers of AO-positive particles (orange), (Figure 6C). The expression of AKT2 was downregulated, and the expression of TSC1 and LC3B was upregulated (Figure 6D). Thus, hsa_circRNA_103124 inhibited autophagy in a hsa-miR-650 dependent way.