Activation-induced cytidine deaminase is a possible regulator of cross-talk between oocytes and granulosa cells through GDF-9 and SCF feedback system

Activation-induced cytidine deaminase (AID, Aicda) is a master gene regulating class switching of immunoglobulin genes. In this study, we investigated the significance of AID expression in the ovary. Immunohistological study and RT-PCR showed that AID was expressed in murine granulosa cells and oocytes. However, using the Aicda-Cre/Rosa-tdRFP reporter mouse, its transcriptional history in oocytes was not detected, suggesting that AID mRNA in oocytes has an exogenous origin. Microarray and qPCR validation revealed that mRNA expressions of growth differentiation factor-9 (GDF-9) in oocytes and stem cell factor (SCF) in granulosa cells were significantly decreased in AID-knockout mice compared with wild-type mice. A 6-h incubation of primary granuloma cells markedly reduced AID expression, whereas it was maintained by recombinant GDF-9. In contrast, SCF expression was induced by more than threefold, whereas GDF-9 completely inhibited its increase. In the presence of GDF-9, knockdown of AID by siRNA further decreased SCF expression. However, in AID-suppressed granulosa cells and ovarian tissues of AID-knockout mice, there were no differences in the methylation of SCF and GDF-9. These findings suggest that AID is a novel candidate that regulates cross-talk between oocytes and granulosa cells through a GDF-9 and SCF feedback system, probably in a methylation-independent manner.


Scientific Reports
| (2021) 11:3833 | https://doi.org/10.1038/s41598-021-83529-x www.nature.com/scientificreports/ addition, exogenously labelled-AID microinjected into the cytoplasm of oocytes in the germinal vesicle stage was shown to enter the nucleus of oocytes just before germinal vesicle breakdown, suggesting that AID can directly access oocyte DNA 16 . Since AID may play a role in epigenetic reprogramming 6,17,18 , it is necessary to clarify the physiological roles of AID in the cross-talk between oocytes and cumulus granulosa cells. Thus, we examined the expression profiles of AID in the mouse ovary and its possible role in ovarian functions.

Results
AID is expressed in murine oocytes and granulosa cells. First, we immunohistochemically stained murine Peyer's patches and observed that AID was expressed on B lymphocytes in the Peyer's patches (Fig. 1A) 19 . In 8-week-old mouse ovaries, AID expression was detected in the oocytes and granulosa cells in the follicles, but not in the surface epithelial cells, interstitial-stromal cells, nor luteal cells in the corpus luteum (Fig. 1A,B). The staining profiles of AID using another anti-AID antibody (Merck Millipore MABF63, clone 328.8) raised against a different antigenic peptide were similar to those detected with the initial antibody (Fig. 1C). The expression of AID was not observed in the ovarian follicles of AID-KO mice, validating the specificity of this assay (Fig. 1C).
In accordance with the results of immunohistological study, mRNA expression of AID in granulosa cells and oocytes was confirmed by RT-PCR (Fig. 1D). RT-qPCR showed that the AID mRNA expression level in  Figure S1. (E) Relative expression of AID in murine oocytes, granulosa cells, and liver cells as a negative control. AID mRNA expression in granulosa cells was higher than that in oocytes. Gapdh glyceraldehyde-3-phosphate dehydrogenase. www.nature.com/scientificreports/ granulosa cells was higher than that in oocytes (Fig. 1E). These results suggest that AID is abundantly expressed in the murine ovary, especially granulosa cells.
For further verification, we employed Aicda-Cre/Rosa-tdRFP reporter mice to trace the history of AID expression by tdRFP 7 . However, the expression of RFP was observed in granulosa cells and luteal cells that are derived from granulosa cells ( Fig. 2A), but not in oocytes (Fig. 2B).

Down-regulated genes in the granulosa cells of AID-KO mice.
To identify the AID-regulated molecules in murine granulosa cells, microarray analysis was performed. Up-regulated genes were defined as those with a fold-change > 2 with highly ranked genes in WAD ranking (304 genes), whereas down-regulated genes were those with a fold-change < 0.5 (274 genes) out of a total of 27,281 genes. Gene ontology biological process term enrichment analyses of the identified up-and down-regulated genes were performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) with the threshold at a P-value < 0.05, and the top five groups are presented (Fig. 3A).
Since the gene ontology term of regulation of reproductive process was detected in the down-regulated group, we focused on the genes related to follicular formation. Among them, GDF-9 (a secreted member of the transforming growth factor-β superfamily) and SCF (c-Kit ligand) were significantly down-regulated in granulosa cells of AID-KO mice. This reduction was confirmed by RT-qPCR using samples different from those used for the microarray (Fig. 3B).
Since GDF-9 and BMP-15 were reported to be secreted by oocytes, we also examined their mRNA expression in oocytes. RT-qPCR showed that the mRNA expressions of GDF-9 and BMP-15 in oocytes were high and the expressions were decreased in oocytes derived from the AID-KO mouse (Fig. 3C). Immunohistological examination revealed the protein expression of GDF-9 in oocytes and granulosa cells, whereas it was reduced in the AID-KO mouse (Fig. 3D). Since these findings suggest the de novo synthesis of GDF-9 by granulosa cells, we further investigated the mRNA expression of GDF-9 and BMP-15 in granulosa cells. Although the expression rates were low, both mRNA expressions were detected in granulosa cells and their expressions were significantly reduced in the AID-KO mouse (Fig. 3E). These findings suggest that AID regulates folliculogenesis-related genes, in both oocytes and granulosa cells.
Evaluation of folliculogenesis in AID-KO mice. The ovarian morphology of AID-KO mice showed no apparent abnormality in the structure compared with the wild type at both 3 and 8 weeks of age (Fig. 4A,B). Since GDF-9 and SCF were reported to be involved in follicular development [20][21][22] , we counted the number of www.nature.com/scientificreports/ secondary follicles and early antral follicles, but there was no significant difference ( Fig. 4C,D). The average litter size of AID-KO mice (7.0 ± 0.8) was the same as that in the background C57/BL6 mice (data not shown).
GDF-9 promotes AID and attenuates SCF expression in cultured granulosa cells. We next examined how AID expression is regulated in the cross-talk between oocytes and granulosa cells. Since GDF-9 was reportedly a pivotal secretory factor 23,24 , we cultivated primary granulosa cells from 3-week-old wild-type mice. In the absence of GDF-9, there was a rapid and marked decrease in AID expression during a 6-h culture (Fig. 5A). In contrast, in the presence of GDF-9, AID expression was maintained (Fig. 5A). Another oocytederived factor, bone morphogenetic protein-15 (BMP-15), also increased AID expression, but recovery of the lost AID expression was slight (Fig. 5A).
On the other hand, the mRNA expression of SCF was more than 3-times increased during a 6-h culture (Fig. 5B). This increase was completely suppressed by GDF-9, as previously reported 25 . In contrast, BMP-15 showed no effects on SCF expression in cultured granulosa cells (Fig. 5B).

AID attenuates SCF expression in cultured granulosa cells. To further clarify whether intrinsic AID
is responsible for SCF expression in granulosa cells, we performed an RNA interference assay. We cultured granulosa cells in the presence of GDF-9, where AID expression was maintained and SCF expression was not increased under the influence of GDF-9. When we performed knockdown of AID mRNA, both siRNA-AID_1 and siRNA-AID_2 significantly suppressed the mRNA expression of AID (Fig. 5C). Under this condition, SCF expression was significantly suppressed (Fig. 5D), which is consistent with the altered gene expression observed in AID-KO mice (Fig. 3). The results suggest that AID expression in granulosa cells was responsible for SCF expression. www.nature.com/scientificreports/ AID did not affect DNA methylation profiles of SCF and GDF-9. AID has been suggested to be involved in epigenetic regulation through DNA methylation 9,26 . Thus, we examined the DNA methylation profiles of SCF and GDF-9 using cultured granulosa cells and ovarian tissues. In the granulosa cell culture, geneknockdown of AID did not affect methylation profiles of SCF and GDF-9 ( Fig. 6A,B). We also observed no differences in methylation profiles of SCF and GDF-9 in the ovarian tissues between wild-type and AID-KO mice (Fig. 6C,D).

Discussion
This study is the first report demonstrating that AID regulates folliculogenesis-related factors, including SCF. GDF-9 and BMP-15, oocyte-secreted soluble factors, were shown to maintain AID expression in granulosa cells, whereas AID is necessary for expression of SCF, which is important for the growth and survival of oocytes. These findings suggest that AID is involved in cross-talk between oocytes and granulosa cells during folliculogenesis. GDF-9 is a member of the transforming growth factor-beta (TGF-β) superfamily and was reported to be specifically expressed in oocytes 27 . GDF-9 is considered to control the rate-limiting step in the early development of follicles. GDF-9-KO mice showed decrease in proliferation of granulosa cells and arrest of follicle development in the primary stage, leading to failed ovulation and no pregnancies 23,24 . On the other hand, SCF is a ligand for C-kit that is expressed on oocytes and theca cells in murine follicles. SCF was proposed to regulate the growth of oocytes 28,29 and theca cells 30 . It was reported that murine oocytes developed two-fold in the presence of SCF on in vitro culture, which corresponds to the early stages of oocyte growth 28 . In addition, inhibition of the SCF/c-kit interaction by anti-c-kit antibody induced oocyte death on in vitro culture 29 . SCF was also demonstrated to prevent oocytes in the primordial follicles from undergoing apoptosis in ovarian organ culture 31 . Furthermore, GDF-9 reportedly suppresses SCF expression 32,33 , whereas SCF interacts with oocytes and increases GDF-9 expression as negative feedback 32 . In the GDF-9-KO mouse, SCF expression was increased as high as 32 fold, showing the continuous growth of oocytes 24 . The above reports collectively indicate that GDF-9 and SCF www.nature.com/scientificreports/ contribute to critical cross-talk between oocytes and granulosa cells during oocyte development and follicular growth (Fig. 7A). From the expression profiles of GDF-9 and SCF in AID-KO mice, we suggest that AID enhances GDF-9 and SCF expression in follicles. To investigate this possibility, we employed a primary culture of granulosa cells, and confirmed the suppressive effects of GDF-9 on SCF expression, as previously reported 32,33 (Fig. 5B). Then, we examined the effects of gene suppression of AID by siRNA on SCF expression. This down-regulation of AID reduced SCF expression, supporting the above suggestion that AID enhances SCF expression. This also suggests that AID is involved in critical cross-talk between oocytes and granulosa cells through the GDF-9 and SCF feedback system (Fig. 7A).
Since AID is not a growth factor, the direct mechanism of the decreased expression of SCF in AID-KO mice and AID-knockdown granulosa cells is unknown. To explain this mechanism, we focused on the roles of AID in epigenetic regulation of DNA demethylation. AID reportedly induces demethylation by deaminating 5-methylcytosine to thymine, followed by base excision repair replacing thymine with non-methylated cytosine 9,26 . Nonetheless, AID knockdown did not affect methylation of SCF or GDF-9 (Fig. 6). The mechanism by which AID positively regulates the expression of folliculogenesis-related genes remains to be elucidated. An intriguing possibility is that AID binds to and stabilizes the target mRNAs. APOBEC1 reportedly binds to the 3′ UTRs of mRNA and increases its stability 34,35 . In murine enterocytes, APOBEC1 was reported to bind to cyclooxygenase 2 mRNA, increasing its expression and prostaglandin E2 synthesis 35 . Consequently, AID may contribute to the stabilization of SCF mRNA, but this requires further study (Fig. 7B). www.nature.com/scientificreports/ This study showed that AID expression rapidly decreased in the absence of GDF-9 (Fig. 5A), suggesting that continuous stimulation of GDF-9 is necessary to maintain AID expression in granulosa cells. Since GDF-9 is a member of the TGF-β superfamily, the regulation of AID expression in granulosa cells by GDF-9 is supported by our previous findings that AID expression is induced by TGF-β in a murine B-cell lymphoma line 36 . GDF-9-induced AID may counteract the suppressive action of GDF-9 on SCF production. Since granulosa cells in the antral follicles have follicular fluid, the concentrations of GDF-9 and SCF in the follicular cavity are relatively stable. However, in smaller follicles without follicular fluid, changes in GDF-9 production by oocytes rapidly influence SCF production by adjacent granulosa cells. AID may modulate the GDF-9 and SCF feedback system as a buffering role.
Although we detected mRNA and protein of AID in oocytes by RT-PCR and immunohistological staining (Fig. 1), RFP expression was not observed in oocytes of AID-Cre/Rosa-tdRFP reporter mice (Fig. 2). This may be simply attributed to its lower expression level in oocytes, being unable to drive detectable tdRFP expression via Cre recombinase. Alternatively, AID mRNA in oocytes may have an exogenous origin. Since AID is a candidate to control epigenetic reprogramming 6,17,18 , the presence of AID in oocytes is an important issue from the perspective of genetic regulation. Indeed, as follicular EVs were reported to contain granulosa cell-derived mRNA 11 , mRNA of AID may be transported from granulosa cells through follicular EVs [12][13][14][15] . In addition, a recent study demonstrated that murine oocytes were directly connected to granulosa cells, by fusing their cellular membrane, which enables the transport of large molecules including GDF-9, from oocytes to granulosa cells 37 . This connection may transport AID protein from granulosa cells to oocytes. The exact origin of AID in oocytes remains to be determined (Fig. 7B).
The expression of BMP-15 in oocytes and granulosa cells was also reduced in AID-KO mice (Fig. 3C,E). In the rat oocyte-granulosa cell co-culture, SCF was reported to suppress BMP-15 expression 32 . This study found no effects of BMP-15 on SCF expression in cultured granulosa cells (Fig. 5B). However, although recovery of the lost AID expression was slight, the BMP-15 enhanced AID expression (Fig. 5A). Considering the promoting effects www.nature.com/scientificreports/ of AID in SCF mRNA expression in granulosa cells, these findings suggest the presence of a negative feedback system between BMP-15 and SCF via AID (Fig. 7A). This study had several limitations. First, although AID was shown to regulate folliculogenesis-related genes including GDF-9, BMP-15, and SCF, we could not find evidence that AID knockdown affected fertility under physiological conditions (Fig. 4). This suggests that the reduction of GDF9 and BMP-15 is compensated for by another system. However, we could not detect these candidates from the microarray data. Second, the mechanism of AID-induced changes in gene expression of SCF in granulosa cells and GDF-9 and BMP-15 in oocytes remains unclear. The possible involvement of epigenetic changes in this mechanism should be further investigated (Fig. 7B). Since SCF was reported to have different effects on GDF-9 and BMP-15 expression on oocytes, AIDinduced unknown factors other than SCF, which induce GDF-9 and BMP-15 expression in oocytes, should be clarified (Fig. 7B). Third, the phenotype of the ovarian function in the aged AID-KO mouse was not investigated. Since insufficient GDF-9 may promote follicle atresia 38 , additional investigation of the effect of AID deficiency on fertility in the state of aging or under stress conditions should be performed. Finally, the direct effects of AID on oocyte function should be clarified.
In conclusion, this study demonstrated the role of AID in regulating gene expression of follicular cells, suggesting its possible involvement in regulating cross-talk between oocytes and granulosa cells. Further studies are warranted to clarify its role in folliculogenesis.

Methods
Mice. C57BL/6 mice at the indicated ages and gestational stages were purchased from Japan Sankyo Lab Service Corporation. AID-KO mice were crossed with C57BL/6 mice for more than 10 generations 2 . The Aicdacre/Rosa-tdRFP mouse 7 was prepared by mating Aicda-cre 39 and Rosa-tdRFP 40 mice, which were backcrossed at least 10 times on a C57BL/6 background. This mouse showed the transcriptional history of Aicda mRNA as the intracellular expression of RFP (red fluorescent protein). The mice were bred under specific pathogen-free conditions with a natural day/night cycle, fed freely with food and water, and handled in accordance with the guidelines for animal experiments of Kanazawa University. All experimental procedures and housing conditions were approved by the Animal Care and Use Committee of the Kanazawa University Animal Experiment Committee (AP-153522). All studies involving animals are reported in accordance with the ARRIVE guidelines for reporting experiments involving animals 41 .
Immunostaining methods. The small intestines and ovaries of mice were fixed for 2 h at 4 °C in a fresh solution of 4% paraformaldehyde (Wako, Japan). The samples were then washed in PBS, incubated overnight at 4 °C in a solution of 30% sucrose, and embedded in OCT compound (Sakura Finetek). The tissue segments were sectioned with a cryostat at 8 μm and stained for AID using a tyramide signal amplification system kit with TSA Plus TMR (PerkinElmer) 42 . GDF-9 were assessed by indirect immunofluorescent staining. Immunohistochemical staining was performed using a standard avidin-biotin complex peroxidase method, as described It was reported that GDF-9 suppresses SCF expression in granulosa cells, whereas SCF promotes GDF-9 expression in oocytes. This study confirmed the suppressive effects of GDF-9 on SCF expression in granulosa cells. It also showed that GDF-9 promotes AID expression and AID enhances SCF expression in granulosa cells, suggesting that AID modulates the GDF-9 and SCF feedback system as a buffering role. On the other hand, SCF was reported to suppress BMP-15 expression in oocytes. In this study, although BMP-15 had no effects on SCF expression, it enhanced AID expression in cultured granulosa cells, suggesting the presence of a negative feedback system between BMP-15 and SCF via AID. Follicle count. The number of follicles was counted in three different cross-sections in six ovaries of three mice. A secondary follicle was defined by two or more layers of granulosa cells, and an early antral follicle was defined by two or more layers of granulosa cells with a small antral cavity. Only follicles containing an oocyte were counted. Follicle counts were reported as the average number of follicles per ovarian surface area (mm 2 ). The differences in follicle counts were analyzed by the Mann-Whitney U test.

Isolation of oocytes and granulosa cells.
Non-stimulated ovaries from 3-week-old female mice were prepared. We punctured ovaries in the DMEM/F12 culture medium with a 28-gauge needle under a stereomicroscope, and follicles, oocytes, and granulosa cells were released. Oocytes without granulosa cells were isolated manually by a glass micropapillary tube tapered with a burner. If a small number of granulosa cells were attached, they were removed by suction and flushing. Granulosa cells were separated from oocytes and follicles by filtering the suspension through a nylon mesh (40 μm; BD Falcon, Bedford, MA, USA), which allowed granulosa cells but not oocytes or follicles to pass through 45 .
Culture of granulosa cells. Non-stimulated ovarian follicles from 3-week-old female mice were punctured with a 28-gauge needle, and a mixture of granulosa cells and oocytes was isolated. Granulosa cells were separated from oocytes by filtering the suspension through a 40 μm nylon mesh. The granulosa cells were cultured at 37 °C in a humidified atmosphere (5% CO 2 and 95% air) in DMEM/F12 supplemented with 0.5% heat-inactivated Fetal Bovine Serum, 5 μg/mL of insulin, 5 μg/mL of transferrin, 5 ng/mL of sodium selenite, and 1% penicillin-streptomycin. Granulosa cells were treated for 6 h with medium containing the solvent or 300 ng/mL of recombinant mouse GDF-9 (739-G9, R&D Systems) or recombinant human BMP-15 (5096-BM, R&D Systems).
RT-qPCR. RNA extraction and RT-qPCR were performed according to the previously described protocols 43 .
Briefly, Total RNA of the mouse ovary was extracted using TRIsure reagent (Bioline, UK). Total RNA from mouse follicles, granulosa cells, and oocytes was extracted using a RNeasy micro kit (Qiagen). Extracted total RNA was treated with amplification-grade DNase I (Invitrogen) and reverse-transcribed using Prime Script reverse transcriptase (Takara Bio Inc., Shiga, Japan) with random primers (nonadeoxyribonucleotide mixture; pd(N) 9 ) and oligo dT primers (Takara Bio Inc.). The cDNA was amplified using specific primers (Table 1). Quantitative PCR analysis was performed using SYBR Premix Ex Taq (Takara) on an MX3000 thermo-cycler (Stratagene) following the PCR protocol: a single hot-start cycle at 95 °C for 30 s, followed by 36 cycles of denaturing at 95 °C for 5 s, annealing, and extension at 60 °C for 20 s. Amplification was completed by an additional cycle at 95 °C for 1 min, and PCR products were analyzed by 4% agarose gel electrophoresis and stained with ethidium bromide. Relative gene expression was analyzed using the 2 −ΔΔCT method. Results are expressed as the means ± SD of at least two separate experiments using two or more mice. Differences between groups were analyzed for significance by ANOVA or unpaired t-tests. P-values < 0.05 were accepted as significant.
Microarray analysis. Total RNA from mouse granulosa cells was extracted for microarray analysis. Using 3-week-old female wild-type and AID-KO mice, follicles were punctured with a 28-gauge needle and granulosa cells were separated from oocytes by filtering the suspension through a 40-μm nylon mesh. Microarray experiments were performed using SurePrint G3 Mouse 8 × 60 K ver. 2.0. Data analysis was performed using Gene-Spring GX software (Agilent Technologies). In brief, raw intensity data were calculated with Agilent Feature Extraction 12.0.3.1 and normalized as a 75 percentile shift according to the protocol. Differential Expressed Genes were extracted from normalized microarray intensity data using the weighted average difference (WAD) ranking method 46 . This method is a statistical approach based on the fold-change www.nature.com/scientificreports/ method that uses not only the difference in gene expression, but also the signal intensity in microarrays. Foldchange was calculated by dividing the AID-KO value by the wild-type value.
RNA interference assay. AID-siRNA (MSS235859-60, Thermo Fisher Scientific) and negative control siRNA (AM4635, Ambion) were used. siRNA transfections were carried out by the reverse transfection method with Lipofectamine RNAi MAX according to the manufacturer's instructions. In brief, 20 pmol siRNA and 1 μL of Lipofectamine RNAi MAX were incubated in 100 μL Opti-MEM medium (Thermo Fisher Scientific) for 20 min at room temperature in 24-well plates, and then 500μL of cell suspension (2.0 × 10 5 cells) was added to the siRNA-RNAi MAX complex. The cells were incubated for 48 h at 37 °C in a humidified atmosphere containing 5% CO2 and 95% air in the DMEM/F12 described above with 1000 ng/mL of recombinant mouse GDF-9.
DNA methylation assay. DNA methylation assay was performed according to the previously described protocols 47 . Briefly, DNA from individual mouse ovaries and granulosa cells was extracted and bisulfite converted using the EZ DNA Methylation Gold Kit (Zymo Research, Orange, CA USA) according to the manufacturer's instructions and was eluted in 20 μL of water. Pyrosequencing primers were designed for each target of interest using PyroMark Assay Design 2.0 software (Qiagen Inc., CA, USA) to amplify the bisulfite-modified target region. PCR reactions were carried out with 25 ng of bisulfite-converted DNA using the PyroMark PCR kit (Qiagen Inc., CA) in a final volume of 25 µL containing 12

Statistical analysis.
In RT-qPCR studies, results are expressed as the means ± SD of at least two independent experiments using two or more mice. The significant differences between groups were analyzed by ANOVA followed by Dunnett or unpaired t-tests using Microsoft Excel software. The numbers of follicles were counted in three different cross-sections in six ovaries of three mice. The differences were analyzed by the Mann-Whitney U test (SPSS Statistics version 25.0, IBM, USA). P-values < 0.05 were regarded as significant.