circVEGFA inhibits apoptosis in porcine ovarian granulosa cells by binding to miR-21-3p and up-regulating TMX4 expression

Abstract

Background

Follicular atresia is a major determinant of ovarian failure in multiparous sows. Non-coding RNAs (ncRNAs) play an important role in the regulatory mechanisms controlling apoptosis within ovarian granulosa cells (GCs).

Methods

The circular structure of circVEGFA was validated by RNase R and actinomycin D treatments. The function of circVEGFA during apoptosis in GCs was investigated by si-RNA transfection. Furthermore, competitive binding of circVEGFA and TMX4 to miR-21-3p was confirmed by a dual-luciferase reporter gene assay and co-transfection with their inhibitors or siRNA. Results: In this study, we present a novel circular RNA (circRNA), circVEGFA, which shows significantly reduced expression in atretic follicles (AFs) compared to healthy follicles (HFs).

Conclusions

The study demonstrates that circVEGFA increases TMX4 expression and inhibits apoptosis in GCs through competitive binding to miR-21-3p. This study contributes to the understanding of circRNA regulation after follicular atresia.

Introduction

Follicular atresia is a phenomenon in which follicles stop proliferating during development and gradually degenerate. This process is a normal physiological event in mammalian reproduction, with a substantial number of follicles failing to mature and ovulate throughout their life cycle. Approximately 99% of follicles experience atresia and only about 1% are able to successfully ovulate [1]. Granulosa cells (GCs) are the primary component of the follicle and a previous study showed that follicular atresia occurs via apoptosis of GCs [2]. GCs are responsible for providing nutrients and support to oocytes and play a regulatory role in follicular growth and atresia by secreting a variety of hormones and growth factors during follicular development [3]. GCs interact with oocytes and other cells through intercellular signalling to influence the process of follicular growth and atresia [4].

The study of non-coding RNAs (ncRNAs) in reproductive biology has gradually gained attention, especially in the context of follicular development [5], spermatogenesis [6], and embryonic development [7]. Specific long non-coding RNAs (lncRNAs) and circular RNAs (circRNAs) have been reported to be associated with GC function and oocyte maturation [8, 9]. They affect follicular growth and apoptosis by regulating the expression of relevant genes. circRNAs are a special class of ncRNAs named for their unique ring-like structure [10]. This structural feature gives circRNAs increased stability and specificity within the cell, making them resistant to degradation by ribonucleases [11]. In recent years, circRNAs have attracted considerable attention in biological research, particularly in relation to the regulation of gene expression and their involvement in various biological processes. It has been shown that circRNAs are rich in microRNA (miRNA) binding sites and can act as sponges for miRNAs, thereby alleviating the inhibitory effect of miRNAs on their target genes and regulating the expression levels of target genes [12, 13]. Moreover, evidence has emerged that circRNAs are able to directly regulate the activity and function of proteins, thereby influencing cellular physiological and pathological processes by interacting with proteins [14]. Furthermore, research has demonstrated the ability of circRNAs to translate functional peptides [15, 16]. As circRNAs are differentially expressed in various diseases and are highly stable and conserved across species, they have great potential for clinical application as diagnostic markers for diseases such as tumours [17, 18]. miRNAs are 22–25 nucleotide ncRNAs that are relatively evolutionarily conserved and function as negative regulators of gene expression [19]. miRNAs induce the degradation of target mRNAs by binding to the 3’ untranslated region (UTR) of mRNAs, thereby down-regulating gene expression [20]. Studies on the mechanism of action of circRNAs have been reported that circRNAs act as ceRNAs (competing endogenous RNAs) and indirectly regulate the expression of their target genes by adsorbing and inhibiting the activity of miRNAs. Our previous research showed that circRNAs and miRNAs are differentially expressed in healthy follicles (HFs) and atretic follicles (AFs) [21, 22], and that a circRNA-miRNA regulatory network may be present during follicular atresia.

TMX4 (thioredoxin-like transmembrane 4) is a member of the PDI (protein disulfide isomerase) family of N-glycosylated type I membrane proteins localized to the endoplasmic reticulum (ER) [23]. The ER plays a central role in protein synthesis and folding as well as lipid secretion. Dysregulation of protein folding induces ER stress [24], while estradiol (E2) synthesis has been shown to play an essential role in the process of follicular atresia [25, 26]. TMX4 functions in the ER in a similar way to other members of the PDI family, participating in protein folding and redox reactions [27]. Given that TMX4 maintains endoplasmic ER homeostasis through its protein folding and redox functions, its dysregulation may affect endoplasmic ER stress pathways and endoplasmic ER-related processes that are critical for follicle survival, such as E2 synthesis. However, there are fewer studies on TMX4 function and its regulatory mechanisms in reproduction, particularly during apoptosis and follicular atresia in GCs.

In this study, we report the identification of a novel circRNA, circVEGFA, transcribed from the gene encoding vascular endothelial growth factor A (VEGFA). We hypothesized that circVEGFA regulates TMX4 expression by absorption of miR-21-3p, thereby resisting GCs apoptosis and follicular atresia. The aim of this study was to demonstrate the role of circVEGFA-miR-21-3p-TMX4 in GCs apoptosis and to identify the effect of TMX4 on apoptosis in porcine ovarian GCs. The study would improve our understanding of ncRNA in follicular atresia.

Materials and methods

Follicles collection and isolation

Ovaries were collected from adult sows of the Duroc × (Landrace × Yorkshire) three-way cross at a slaughterhouse in Huai’an, Jiangsu Province, China The isolated fresh ovaries were transferred to a saline solution containing 1% penicillin-streptomycin at 37 °C, and subsequently returned to the laboratory for further testing. The ovaries utilized for follicle isolation underwent a thorough cleansing process involving three washes with a saline solution containing 1% penicillin-streptomycin. The ovaries were then placed in petri dishes containing a Phosphate Buffered Saline (PBS) buffer (10099, Gibco, Carlsbad, CA, USA), and follicles measuring 3–5 mm in diameter were isolated and rinsed three times repeatedly with PBS. The classification of the follicles was determined based on their morphological characteristics, the density of GCs within the follicles, and the ratio of progesterone (P4) to E2 (P4/E2) in the follicular fluid. The follicles were classified into healthy (P4/E2 < 5) and early HFs (5 < P4/E2 < 20). The HF and AF follicles were classified as described in our previous study [28]. The sorted out HFs and early Afs samples were stored at -80 °C for subsequent experiments.

Cell culture and transfection

The ovaries used in this study were obtained from healthy, non-stimulated commercial reserve sows from abattoirs through routine slaughter procedures. Fresh ovaries collected on the same day were subjected to a cross-washing procedure using a solution of 1% penicillin-streptomycin saline and 75% alcohol. This process was repeated with two cycles of saline, one cycle of alcohol, and three cycles of washing. The follicular fluid from 3 to 5 mm diameter follicles in the washed ovaries was extracted using a syringe irradiated with ultraviolet light, collected into a centrifuge tube, and subjected to centrifugation at 1000 rpm/min for 5 min. The upper layer of follicular fluid was then discarded. The lower layer of GCs was washed with pre-warmed PBS containing 1% penicillin-streptomycin, washed twice with PBS, and then subjected to centrifugation at 1000 rpm/min for 5 min, and discarded the PBS. The cells were then resuspended in DMEM/F12, containing 15% fetal bovine serum (10100147 C, Gibco) and 1% penicillin-streptomycin. The cells were then blown up to homogeneity using a pipette, and then inoculated with 800µL of DMEM/F12 medium and 200µL of GCs suspension per well in a 12-well cell culture plate. The cells were then shaken back and forth and incubated in a constant temperature incubator at 37℃ with 5% CO₂. Following a 36 h incubation period, the GCs were observed to adhere to the wall, and the wells were washed twice with 1000 µL of PBS containing 1% penicillin-streptomycin each. The PBS was then discarded, and the cells were added to the DMEM/F12 medium. The density of the GCs was observed under a microscope. The transfection system was configured according to the instruction manual of the LipofectamineTM 3000 (L3000015, Invitrogen, Carlsbad, CA, USA) transfection reagent. The cells were found to be capable of transfection when the growth status was satisfactory and the density was above 70%. Inhibitors and mimics of microRNA and circular RNA were produced by GenePharma (Shanghai, China) (Table S1). According to the requirements of the subsequent experiments, the samples were cultured at a constant temperature for 24–48 h.

RNA preparation and qRT-PCR

Following a 24 h period of transfection, the GCs should be collected and total RNA extracted using the procedure outlined in the instruction manual for the Trizol reagent (R401-01, Vazyme, Nanjing, China). Genomic DNA should then be removed, after which the total RNA should be reverse transcribed into cDNA using HiScript III RT SuperMix (Q312, Vazyme, Nanjing, China) and prepared for use. An additional 1 µL of stem-loop primer was added for miRNA detection. Fluorescence quantification was performed on the StepOnePlus System (Applied Biosystems, Carlsbad, CA, USA) using the SYBR Green Master Mix Kit (Q711, Vazyme, Nanjing, China). U6 was used as an internal reference for miRNA quantification, and the relative expression was calculated using the 2^−ΔΔCT method. The primer sequences are presented in Table S2.

RNase R digestion and qRT-PCR

The RNA was extracted from the cultured GCs. Subsequently, half of the extracted RNA was directly reverse transcribed, while the other half was treated with RNase R (R7092S, Beyotime, Shanghai, China) reagent before reverse transcription. RNA was reversed to cDNA according to the kit instructions, and the products were subsequently verified by qRT-PCR experiments.

Actinomycin D treatment

GCs were cultivated in 12-well plates and exposed to 5 µg/mL actinomycin D at a cell density of 70% for the experiment. At predetermined time points, the cells were harvested for further analysis. To assess the RNA expression levels of circVEGFA and VEGFA mRNA, qRT-PCR was conducted.

Separation of nucleus and cytoplasm

Porcine ovarian GCs were cultured in vitro for 36 h, after which the cells were washed using a PBS solution containing 1% penicillin-streptomycin. Subsequently, trypsin solution at 37 °C and 0.5% EDTA (25200056, Thermo, Wilmington, DE, USA) was added to detach the cells from the culture dish. The trypsin-digested cells were then collected and 500 µL of pre-cooled PBS solution was added. The cell suspension was then incubated on ice for 10 min, after which 5 µL of NP-40 reagent was added to the cell suspension. The cell suspension was then homogeneously mixed for 3–5 h at 4 °C to fully separate the nucleus from the cytoplasm. Centrifugation at 12,000 rpm/min for three minutes was then used to separate the nuclei and cytoplasm. The upper cytoplasm-containing layer and the nucleus-containing layer were then collected separately for subsequent RNA extraction.

Fluorescence in situ hybridization (FISH)

In vitro cultures of GCs were subjected to a 36 h culturing period, after which the cells were transferred to coverslips and subsequently incubated in an incubator. Following cell stabilization, the cells were fixed with 4% paraformaldehyde dissolved in DEPC-treated, autoclaved PBS for 20 min. Thereafter, the fixed cells were washed thrice using a solution of PBS containing 1% penicillin-streptomycin. The cells were then digested for 5 min with proteinase K (20µL/mL), probes specifically labelled with circVEGFA and miR-21-3p were synthesized (Servicebio, Wuhan, China), and nuclei were stained with DAPI. Finally, fluorescence images were acquired using a Nikon orthogonal fluorescence microscope (Nikon DS-U3, Tochigi, Japan). The probe sequences are delineated in Table S3.

Apoptosis assay

GCs were subjected to transfection and culturing for a period of 48 h, after which the degree of apoptosis was measured using a membrane-linked V-FITC/ PI staining kit (Vazyme, Nanjing, China). The apoptosis rate was then detected through flow cytometry (Becton Dickinson FACS Calibur Franklin Lakes, NJ, USA) according to the principle of fluorescence-activated cell sorting (FACS). For each independent experiment, a minimum of 10,000 gated events per sample. Apoptosis rates were analyzed using FlowJo v7.6 software (Stanford University, Stanford, CA, USA).

Western blotting

Following a 48 h period of cell transfection, the cells were washed using PBS containing 1% penicillin-streptomycin. Thereafter, the cells were lysed by the addition of 150 µL of RIPA lysate (AKR-191, AnnoRon, Beijing, China). Following a 10 min, the total cellular proteins were collected with a spatula. The protein concentration was determined using a concentration assay kit (BL521A, Beyotime, Shanghai, China), and the same concentration of sample (20 µg per lane) was added to the protein uploading buffer for denaturation (99 °C, 10 min). Under 140 V conditions, polyacrylamide gel electrophoresis (SDS-PAGE, 15% polyacrylamide) was performed for 90 min to separate proteins of different molecular weights. After the electrophoresis was completed, the protein was transferred onto a PVDF membrane using the wet transfer method (110 V, 100 min). The PVDF membrane was blocked in 5% BSA for 2 h. The primary antibody was incubated at 4 °C for 14 h. Following the conclusion of the incubation period, the strips were washed with a TBST solution. Thereafter, the secondary antibody was allowed to incubate at room temperature for 1 h. The strips were then subjected to a subsequent wash. Ultimately, the PVDF membrane was imaged in an electrophoretic gel imaging system using the BCL chemical detection kit. The primary antibodies used are anti-TMX4 (1:1000, Abclonal, Wuhan, China) and anti‐GAPDH (1:3000, Origene, Rockville, MD, USA). GAPDH served as internal controls. An IgG antibody (Cell Signaling Technology, Beverly, MA, USA) was used as the secondary antibody.

Dual-luciferase activity assay

Following a 24 h transfection period, the luciferase activity of the cells was detected in accordance with the instructions provided by the luciferase reporter gene kit (ADD1205, Vazyme, Nanjing, China). The cells were lysed using 100 µL of lysate to collect the lysates within the cells, and subsequently, the samples were processed according to the instructions for the use of the kit. The resulting samples were then analyzed for firefly fluorescence versus sea kidney fluorescence in samples was detected using a luminescence detector (Promega, Glomax20/20, Madison, WI, USA). The binding sites and mutated sequences are shown in Table S4.

Bioinformatics analysis

The mature sequences of miRNA were obtained from the online database miRBase (https://www.mirbase.org/). Predictions of interactions between miRNA and their target genes were made using the online databases miRDB (http://mirdb.org/), and Targetscan (http://www.targetscan.org/vert_72/) and TarBase (http://www.microrna.gr/tarbase) websites for prediction. Predictions of binding sites for miRNAs were made using RNAhybrid (https://bibiserv.cebitec.uni-bielefeld.de/rnahybrid).

Statistical analysis

The statistical analyses and plots were performed using GraphPad Prism 8 software (GraphPad Software Inc.). Date was analyzed using t-test and one-way ANOVA analyses. If the analysis of variance indicates a significant effect, then the Tukey test will be used for the post-hoc pairwise comparisons. All experimental data are presented as mean average ± S.E.M. Each experiment was repeated three times. pvalue < 0.05 was indicative of significant differences, and pvalue < 0.01 was indicative highly significant differences.

Results

Identification and validation of circVEGFA

In this study, we utilized laboratory circRNA sequencing and biological analyses to identify a novel circular RNA, designated circVEGFA. circVEGFA is a circular RNA formed by back-splicing of VEGFA gene exon 3–5, with a length of 304 bp. Our findings demonstrate that circVEGFA expression is significantly reduced during early follicular atresia. To verify the accuracy of the splice site, PCR product sequencing was performed, with divergent primers designed based on the splice site of circVEGFA. The presence of a circular structure was confirmed for circVEGFA (Fig. 1A). Convergent and divergent primers were designed for expression in gDNA and cDNA, respectively. The agarose gel electrophoresis assay results demonstrated that the different primers could produce specific bands in cDNA. Conversely, the divergent primers failed to produce any bands in gDNA, thereby validating the existence of circVEGFA (Fig. 1B). Subsequent, RNase R experiments revealed that the addition of RNase R led to a substantial reduction in the expression of linear VEGFA, while having no significant impact on the expression of circVEGFA (Fig. 1C). Treatment with actinomycin D demonstrated that circVEGFA was more stable than linear VEGFA, thus confirming the stability of circVEGFA (Fig. 1D). In addition, circVEGFA expression was found to be significantly reduced in AFs compared to HFs by qRT-PCR (Fig. 1E). To further explore the localization of circVEGFA in GCs, experiments were conducted to separate the nucleus and cytoplasmic, and fluorescence in situ hybridization experiments were performed, which showed that circVEGFA was distributed in both the nucleus and cytoplasm (Fig. 1F and G).

Fig. 1

Identification and validation of circVEGFA. (A) Schematic representation of the circVEGFA structure, generated by the process of back splicing of Exon3-5 of the VEGFA gene. (B) Amplification products of circVEGFA and linear VEGFA mRNA using divergent and convergent primers in cDNA and gDNA. (C) VEGFA mRNA and circVEGFA expression levels following RNase R (+) and RNase R (-) digestion. (D) The stability of circVEGFA in comparison to linear VEGFA following actinomycin D treatment. (E) Differential expression of circVEGFA in HFs and AFs as detected by qRT-PCR (n = 6). (F) Nucleus and cytoplasmic separation assay will be used to determine the localization of circVEGFA. (G) The localization of circVEGFA in GCs detected by FISH. The circVEGFA was labelled with a red fluorescent probe, and the nuclei were stained with DAPI (blue). Scale bar: 20 μm. * p < 0.05, ** p < 0.01

circVEGFA inhibits GCs apoptosis

To further explore the function of circVEGFA in the apoptosis of porcine GCs, specific small interfering RNAs were designed based on the back shear site of circVEGFA. si-circVEGFA was transfected into GCs, and the qRT-PCR results showed a significant decrease in the expression level of circVEGFA, whereas that of linear VEGFA was unaffected (Fig. 2A). This finding suggests that si-circVGEFA is capable of specifically interfering with circVEGFA. Furthermore, qRT-PCR analysis of apoptotic genes BAX and BCL2 revealed a significant decrease in the ratio of BCL2/BAX. (Fig. 2B). After to a 48 h of transfection period, the apoptosis rate of GCs was detected by Annexin V-FITC/PI staining, revealing a significant increase in the apoptosis rate of GCs following transfection with si-circVEGFA (Fig. 2C). These results collectively indicate that circVEGFA functions as a suppressor of apoptosis in GCs.

Fig. 2

circVEGFA inhibits GCs apoptosis. (A) The expression levels of circVEGFA and its corresponding linear mRNAs were measured following the treatment of GCs with either the negative control (NC) or si-circVEGFA. (B) The expression ratio of BCL2/BAX was detected by qRT-PCR following interference with si-circVEGFA. (C) The apoptosis rate of porcine GCs was detected by Annexin V-FITC/PI following interference with circVEGFA. * p < 0.05, ** p < 0.01

circVEGFA as a sponge for miR-21-3p

Following the findings of previous studies, including those employing miRNA-seq and bioinformatics prediction analysis, the expression level of miR-21-3p is known to be elevated during the process of follicular atresia. Moreover, it has been hypothesized that miR-21-3p may bind to circVEGFA. Consequently, qRT-PCR was employed to detect the expression levels of miR-21-3p in both HFs and AFs. The results obtained demonstrated that the expression levels of miR-21-3p in AFs were significantly higher than those in HFs (Fig. 3A), thereby corroborating the findings of the preceding sequencing results. Furthermore, the expression levels of miR-21-3p were found to be significantly increased in response to the transfection of si-circVEGFA in comparison with the control (Fig. 3B). To further explore the correlation between circVEGFA and miR-21-3p in individual follicles, qRT-PCR results showed that the expression levels of circVEGFA and miR-21-3p in individual porcine follicles exhibited a significant negative correlation (Fig. 3C). To further investigate the role of miR-21-3p in the apoptosis of porcine ovarian GCs, miR-21-3p mimics and inhibitors were transfected into GCs to verify its expression and interference efficiency. Compared with the control group, the expression and interference efficiencies significantly increased and decreased levels (Fig. 3D and E). Subsequent apoptosis assay revealed that the apoptosis rate of GCs was significantly increased after transfection with miR-21-3p mimics in comparison with the control group. In contrast, the apoptosis rate of GCs was significantly decreased after transfection with miR-21-3p inhibitor (Fig. 3F and G). To further explore the binding between circVEGFA and miR-21-3p, wild-type and mutant vectors of circVEGFA containing the miR-21-3p binding site were constructed and subjected to a dual-luciferase activity assay (Fig. 3H). The results demonstrated that the expression of miR-21-3p mimics was observed to bind to circVEGFA wild-type vectors, but not to mutant vectors (Fig. 3I). To further verify the localization of circVEGFA and miR-21-3p in cells, the results of the FISH assay showed that circVEGFA and miR-21-3p co-localized in the cytoplasm of GCs (Fig. 3J). The results obtained thus far demonstrate that the expression of circVEGFA and miR-21-3p in GCs is capable of inhibiting apoptosis in GCs by adsorbing miR-21-3p.

Fig. 3

circVEGFA as a sponge for miR-21-3p. (A) qRT-PCR was utilized to detect differential expression of miR-21-3p in HFs and AFs. (B) qRT-PCR was employed to detect the expression levels of miR-21-3p after interference with circVEGFA. (C) The expression levels of circVEGFA and miR-21-3p were found to be negatively correlated in single follicles (n = 14). (D, E) Validation of the efficacy of miR-21-3p overexpressing and inhibitor techniques. (F, G) Changes in the apoptosis rate of GCs after transfection with miR-21-3p mimics and its inhibitor were detected by flow cytometry. (H) Binding sites of circVEGFA and miR-21-3p predicted by RNAhybrid and maps of pmirGLO vector constructs. (I) The validation of the binding of the miR-21-3p mimics to the circVEGFA-WT was conducted by dual-luciferase activity analysis. (J) The validation of the localization of the circVEGFA (red) and the miR-21-3p (green) in GCs was conducted by FISH with nuclei stained by DAPI (blue). Scale bar: 20 μm. * p < 0.05, ** p < 0.01

TMX4 inhibits GCs apoptosis

Following the findings of previous RNA-seq data and predictions derived from bioinformatics software, TMX4 was identified as a potential downstream target gene in the circVEGFA/miR-21-3p axis, exhibiting a high absolute value of binding energy (Fig. 4A). qRT-PCR analysis revealed a significant decrease in TMX4 expression levels in AFs compared to HFs (Fig. 4B). To further explore the function of TMX4 in GCs, specific interfering RNA was designed and si-TMX4 was transfected into GCs, resulting in a significant reduction in mRNA and protein levels of TMX4 (Fig. 4C and D). The results of the apoptosis assay demonstrated that the apoptosis rate of GCs increased significantly following the knockdown of TMX4 (Fig. 4E). The above results indicated that TMX4 could inhibit apoptosis of GCs.

Fig. 4

TMX4 inhibits GCs apoptosis. (A) Venn diagram illustrating the overlap between predicted miR-21-3p target genes from miRDB, targetscan and tarBase, and target genes that were found to be differentially expressed between HFs and AFs. (B) The expression of TMX4 in HFs and AFs was analyzed by qRT-PCR.qRT-PCR analysis was employed to assess the differential expression of TMX4 in HFs and AFs. (C) The efficiency of TMX4 mRNA interference was determined by qRT-PCR following the transfection of si-TMX4. (D) The interference efficiency of TMX4 protein was detected by Western blotting (WB) after transfection with si-TMX4. (E) The changes in the apoptosis rate of GCs after interference with TMX4. * p < 0.05, ** p < 0.01

miR-21-3p promotes GCs apoptosis by targeting TMX4

In this study, the interactions between miR-21-3p and TMX4 in GCs were explored through the construction of TMX4 wild-type and mutant vectors containing miR-21-3p binding sites. These vectors were constructed based on the mutation of miR-21-3p and TMX4 binding sites as predicted by the RNAhybird bioinformatics website (Fig. 5A). The wild-type and mutant vectors were then co-transfected with miR-21-3p mimics into cells, respectively. The results of the dual-luciferase activity assay indicated that miR-21-3p mimics could bind to the wild-type vector but not to the mutant vector (Fig. 5B). To further explore the correlation between circVEGFA and miR-21-3p, qRT-PCR results showed that circVEGFA exhibited a significant negative correlation with the expression level of miR-21-3p in single follicles (Fig. 5C). Transfection of miR-21-3p mimics into GCs resulted in a significant decrease in both mRNA and protein expression levels of TMX4 (Fig. 5D and E). Conversely, the transfection of the miR-21-3p inhibitor resulted in a significant increase in mRNA and protein expression levels of TMX4 (Fig. 5F and G). The miR-21-3p inhibitor and si-TMX4 were co-transfected into GCs, and the apoptosis level of GCs was detected using an apoptosis assay. The experimental results demonstrated that si-TMX4 could reverse the decrease in the apoptosis rate of GCs caused by the miR-21-3p inhibitor (Fig. 5H). The above results indicated that the expression of miR-21-3p and TMX4 in GCs was negatively correlated, and that miR-21-3p promoted apoptosis in porcine ovarian GCs by regulating the expression of TMX4.

Fig. 5

miR-21-3p promotes GCs apoptosis by targeting TMX4. (A) Prediction of the binding site of miR-21-3p and TMX4 by RNAhybrid, along with a map of the pmirGLO vector construct. (B) Validation of miR-21-3p mimics binding to TMX4-WT by dual-luciferase activity analysis. (C) Negative correlation between the expression levels of TMX4 and miR-21-3p in single follicle (n = 14). (D, E) Changes in TMX4 mRNA and protein levels after transfection with miR-21-3p mimics. (F, G) Changes in TMX4 mRNA and protein levels after transfection with miR-21-3p inhibitor. (H) Changes in the apoptosis rate of GCs after co-transfection with miR-21-3p inhibitor and si-TMX4 were detected by flow cytometry. * p < 0.05, ** p < 0.01

circVEGFA inhibits GCs apoptosis by upregulating TMX4 through sponging miR-21-3p

Correlation analysis of circVEGFA and TMX4 expression levels in individual follicles demonstrated a positive correlation between circVEGFA and TMX4 expression (Fig. 6A). To further explore the apoptosis inhibition effect of circVEGFA in porcine GCs by targeting TMX4 through adsorption of miR-21-3p, si-circVEGFA and miR-21-3p inhibitors were co-transfected into GCs. The qRT-PCR and WB experiments demonstrated that the miR-21-3p inhibitor could reverse the reduction of TMX4 mRNA and protein levels of TMX4 induced by the transfection of si-circVEGFA (Fig. 6B and C). Furthermore, the results of the apoptosis assay results showed that the miR-21-3p inhibitor could reverse the elevated apoptosis rate of GCs induced by si-circVEGFA (Fig. 6D). The results indicated that circVEGFA was able to upregulate the expression of TMX4 through adsorption of miR-21-3p, thereby inhibiting apoptosis of GCs.

Fig. 6

circVEGFA inhibits GCs apoptosis by upregulating TMX4 through sponging miR-21-3p. (A) The expression levels of TMX4 and miR-21-3p were found to be negatively correlated in single follicles (n = 14) (B) Changes in TMX4 mRNA levels after co-transfection with si-circVEGFA and miR-21-3p inhibitor were detected by qRT-PCR. (C) WB detection of changes in TMX4 protein levels after co-transfection with si-circVEGFA and miR-21-3p inhibitor. (D) Flow cytometry detection of changes in apoptosis rate of GCs after co-transfection with si-circVEGFA and miR-21-3p inhibitor. * p < 0.05, ** p < 0.01

Discussion

circRNAs are thought to play an essential role in the control of gene expression by acting as miRNA sponges, translational regulators and RNA-binding protein sponges [29]. This mechanism has been proven to activate in a variety of biogenesis processes, such as growth and development [30], cell cycle [31], and disease genesis [32]. In ovarian follicles, this mechanism has been demonstrated in hormone secretion [33], follicular development and GCs apoptosis [34, 35]. During human polycystic ovary syndrome (PCOS) ovaries, circ-FURIN has been shown to inhibit GCs apoptosis by binding to miR-195-5p and increasing BCL2 expression [36]. During porcine follicular atresia, circINHA has been shown be inhibit GCs apoptosis by binding to miR-10a-5p and increasing CTGF expression [37]. During poultry ovarian atresia, circRAB11A inhibited GCs apoptosis by binding miR-24-5p and increasing the expression of EGFR and RAB11A [38]. In our study, we proved that circVEGFA increased TMX4 gene level via sponging miR-21-3p, thus leading to a resist effect on GCs apoptosis. As a member of PDI family, TMX4 is involved in redox reactions in the ER by interacting with Erp57 to fulfill its function as a glycoprotein-specific oxidase [39]. Furthermore, TMX4 has also been reported to play an important role in protein folding and ER transport functions [27]. Therefore, our result is the first evidence that the circRNA-miRNA-mRNA axis may affect cell fait through ER function, and this result broadened our understanding of the role of circRNA-related mechanisms in GCs.

The influence of the ER on apoptosis may be achieved through two pathways. Firstly, the stable function of the ER plays a decisive role in maintaining the normal secretion of steroid hormones. It is clear that E2 plays a key role in follicle development and maturation, and its dysfunction may cause reproductive dysfunction such as PCOS [26, 40]. During the process of follicular atresia, E2 to P4 ratios were significantly higher in HFs than in AFs [41]. The ER plays an essential role in the synthesis of E2. In GCs, androstenedione and testosterone, which are formed by theca cells, form E2 by the action of aromatase (encoded by CYP19A1 gene) located on the ER, and are then secreted by free diffusion to play a role outside the cell [25, 42, 43].

Secondly, ER stress within an appropriate range is extremely important for maintaining the normal functions and survival of cells. ER stress is a cellular stress response that caused a result of the accumulation of unfolded or misfolded proteins in the ER [44]. In melanoma cell lines, the use of ER stress inducer thapsigargin (THG) has been shown to decrease the genetic expression of PDI family members, including TMX4 and to induce apoptosis [45]. Also, the use of the PDI gene inhibitor bacillin in melanoma cells has been shown to enhance ER stress response and result in apoptosis [46]. Moreover, TMX4 has been demonstrated to influence the structure and function of the nuclear membrane, which is connected to the ER, by reducing disulfide bonds and modulating protein interactions associated with the nuclear membrane [47]. In ovaries, ER stress has been suggested as a local factor affecting the follicular microenvironment and involved in the pathogenesis of PCOS [48, 49]. In the GCs of PCOS patients, ER stress has been shown to induce the expression of pro-apoptotic factors C/EBP homologous protein (CHOP) and death receptor 5 (DR5) to promote GCs apoptosis [50]. These related results suggest that TMX4 and its functions in follicular atresia through ER stress are also worthy of in-depth exploration.

Lastly, circRNAs are derived from their corresponding host genes, and the mechanisms of mutual regulation with their host genes have been extensively studied [51,52,53]. circRNAs can target transcriptional regulatory regions of host genes by binding to RNA polymerase II (Pol II) [54], by recruiting proteins [55], or by forming an R-loop [56] to positively or negatively regulate the transcription of host genes. As the host gene of circVEGFA, the function of VEGFA may be affected by circVEGFA transcription and thus cause inevitable consequences in follicular angiogenesis. Perifollicular angiogenesis has been demonstrated to play a crucial role in the development of ovarian follicles [57, 58]. As one of the most crucial angiogenesis-related cytokines, VEGFA is synthesized in GCs and theca cells and either acts on receptors in the endothelial cells of the membranous layer of the ovary to control follicular angiogenesis or acts directly on GCs and theca cells to promote follicular growth [59]. Therefore, the act of circVEGFA transcription itself may change the expression level of VEGF, thereby affecting its downstream functions. The transcriptional relationship between circVEGFA and its coding gene VEGF, as well as its impact on GCs, is also worthy of further exploration.

Conclusions

To sum up, we have identified a circRNA formed from the VEGFA gene, circVEGFA, which exhibits reduced expression in atretic ovaries compared to HFs. Our findings indicate that circVEGFA functions as an endogenous miR-21-3p sponge, thereby regulating the expression of the target gene TMX4 and inhibiting apoptosis of GCs, thus contributing to the suppression of porcine follicular atresia. The potential mechanism of TMX4 involves steroid hormones, which are closely associated with ER function, and consequently affects GCs apoptosis. This study provides new insights into the role of ncRNAs in regulating the process of porcine follicular atresia and apoptosis in GCs.