What Changed on the Folliculogenesis in the Process of Mouse Ovarian Aging?

There are about 1-2 million follicles presented in the ovary at birth, while only around 1000 primordial follicles are left at menopause. The ovarian function also decreases in parallel with aging. Folliculogenesis is vital for ovarian function, no matter the synthesis of female hormones or ovulation, yet the mechanisms for its changing with increasing age are not fully understood. Early follicle growth up to the large preantral stage is independent of gonadotropins in rodents and relies on intraovarian factors. To further understand the age-related molecular changes in the process of folliculogenesis, we performed microarray gene expression profile analysis using total RNA extracted from young (9 weeks old) and old (32 weeks old) mouse ovarian secondary follicles. The results of our current microarray study revealed that there were 371 (≥2-fold, q-value ≤0.05) genes differentially expressed in which 174 genes were upregulated and 197 genes were downregulated in old mouse ovarian secondary follicles compared to young mouse ovarian secondary follicles. The gene ontology and KEGG pathway analysis of differentially expressed genes uncovered critical biological functions such as immune system process, aging, transcription, DNA replication, DNA repair, protein stabilization, and apoptotic process were affected in the process of aging. The considerable changes in gene expression profile may have an adverse influence on follicle quality and folliculogenesis. Our study provided information on the processes that may contribute to age-related decline in ovarian function.


Introduction
Ovarian aging results in the cessation of ovarian function, that is, anovulation and a decrease in gonadal steroids secretion. The anovulation causes loss of fertility and reduced hormone production results in multiple health consequences, including vasomotor symptoms, cardiovascular disease, osteoporosis, cognitive dysfunction, depression, mood disorders, sexual dysfunction, vaginal atrophy, and even mortality [1,2]. The age at which natural menopause occurs may be a marker of ovarian aging which is considered to be the multiple pacemakers [3].
Ovarian follicle is the basic unit of ovarian physiological function. After puberty, the periodic development of the ovarian follicles enables the ovary to produce female hormones to maintain secondary sexual characteristics and ovulation. The reproductive aging process is considered to be dominated by the gradual decrease of both the quantity and the quality of the oocytes residing within the follicles present in the ovarian cortex [4]. Females have approximately 1-2 million primordial follicles at birth [5,6]. After birth, the number of follicles decreases gradually with increasing age through atresia with some 300,000 to 400,000 primordial follicles remaining at menarche [4,7]. During the reproductive years, the number of primordial follicles declines until a critical threshold when only about 1000 left at the time of menopause [8][9][10].
The information on the hormonal changes observed gradual decline of the follicle pool and the reduced oocyte quality during ovarian aging is quite a bit; however, the molecular mechanisms behind that are still not fully understood. Studies have shown that accumulation of reactive oxygen species (ROS) and free radicals and the action of environmental factors such as radiation and chemotherapeutic drugs used in cancer patients can cause DNA damage in the oocytes during long periods of dictyate arrest and without repairing, the extent of DNA damage may cause genomic abnormalities (chromosomal breakages and mutations) leading to cell death and follicle atresia [11,12]. Researches have revealed that the expression of BRCA1 (breast cancer type 1) related DNA repair decreased in the process of ovarian aging in rat and buffalo primordial follicles [13,14]. Laboratory and clinical studies also demonstrated that expression of BRCA1 declines in single mouse and human oocytes and BRCA1 mutation is associated with primary ovarian insufficiency [15][16][17].
A better understanding of follicle biology is essential to help make ovarian aging process explicit. Early follicle growth up to the large preantral stage is independent of gonadotropins in rodents and relies on intraovarian factors [6]. Thus in our present study, the secondary follicles from young and old mice ovaries were used to investigate the changes of expression profile during ovarian aging by genome-wide microarray analysis.

Isolation of Secondary Follicles from the Mouse Ovary.
The experimental animals were maintained as per the guidelines of the Animal Care Committee of Tongji Hospital within the Tongji Medical College at the Huazhong University of Science and Technology in China. The 9week old and 32-week old, specific pathogen-free (SPF), female C57BL/6J mice were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). All mice were killed by decapitation and ovaries were collected free of adhering tissue. Under the stereomicroscope, follicles with diameter of 120-140 m, an intact basal membrane, a central and spherical oocyte surrounded by granulosa cells were mechanically dissecting by 2 syringe needles. The ovarian secondary follicles were stored at −80 ∘ C.

RNA Isolation and Microarray
Analysis. The total RNA of the ovarian secondary follicles was extracted with RNAiso plus reagent according to the manufacturer's instructions (Takara, Japan). Of the total of 6 samples, 3 replicate samples were from young mouse ovarian secondary follicles and 3 replicate samples were from old mouse ovarian secondary follicles. All RNA samples were stored in DEPC in order to prevent RNA degeneration. GeneChip hybridization for each sample was examined on Affymetrix 3' IVT Expression Arrays (Mouse Genome 430 2.0 array) at Bioassay Laboratory of CapitalBio Corporation. The technical procedures and quality controls were performed at the CapitalBio Corporation. Hybridization assay procedures were as described in the GeneChip Expression Analysis Technical Manual (http://www.affymetrix.com).

Microarray Data Analysis.
The raw data from microarray analysis was normalized using robust multiarray average (RMA) algorithm. The differentially expressed genes with a fold change ≥2 and a q-value ≤0.05 were identified using Significant Analysis of Microarray (SAM) software. For visualization of differentially expressed genes, unsupervised hierarchical clustering was performed using HemI 1.0.3.7 software (http://hemi.biocuckoo.org/down.php) [18]. Gene Ontology (GO) consisting of three items: molecular functions, biological processes, and cellular components [19] and Kyoto Encyclopedia of Genes and Genomes (KEGG), a set of high-throughput genes and protein pathways [20], analyses of differentially expressed genes were performed using the DAVID online tools (https://david.ncifcrf.gov/) compared with the mouse whole genome [21]. Whole Mouse Genome was used as the reference group. Statistical significance was calculated with a standard hypergeometric equation corrected by a Benjamini Yekutelli correction for multiple testing, which takes into account the dependency among the GO categories. The minimal length of considered GO-paths was 2. Significance was set at corrected p-value < 0.05. The Search Tool for the Tetrieval of Interacting Genes (STRING) database (http://string-db.org/), an online software that provides comprehensive interactions of lists of proteins and genes, was used to build a PPI network of the differentially expressed genes [22]. The cut-off criteria of the minimum required interaction score were 0.7 for the PPI network. The visualizing of the PPI network was constructed using the Cytoscape software (version 3.6.1) [23]. The Clustering with Overlapping Neighborhood Expansion (ClusterONE) plug-in for Cytoscape was used to detect protein complexes in the PPI network [24]. The gene regulatory network modeling for selected differentially expressed genes was performed using Cytoscape software (version 3.6.1).

Global Gene Expression Analysis of Secondary Follicles from Mouse Ovaries.
To characterize the genes that are associated with mouse ovarian aging, we examined the gene expression profile of secondary follicles from young and old mouse ovaries. The expression values of all the six samples (three samples each from young and old mouse ovaries) were normalized using the robust multiarray average (RMA) method. The results of our microarray data were made available in the public domain NCBI-GEO repository (accession ID: GSE121493). The box-whisker plot analysis of normalized data showed uniform distribution of the expression levels in both intra-and intersample manner indicating reliable hybridization (see Figure 1). Summary statistics showed effectiveness of quantile normalization as 50th percentile values were close to 4.9. After normalization of raw data for all three biological replicates, the R package significance analysis of microarray (SAM) was used to identify genes that are differentially expressed in secondary follicles from young and old mouse ovaries (fold change ≥2 or ≤0.5 and q-value ≤0.05). And the results Robust multiarray average (RMA) algorithm was used to eliminate the variation in the arrays from noisy data. Box plot was constructed to illustrate the distribution of normalized probe hybridization signal intensities (log ratios) for all six arrays (young and old mouse ovarian secondary follicles). The probe distribution by 0-100% quantiles as whiskers, the 25-75% quantiles as different color boxes, and the 50% quantile as horizontal line within the box indicated a similar range of signal intensities and confirmed perfect hybridization.
revealed that 371 genes were differentially expressed between the two groups, while 174 genes were upregulated and 197 genes were downregulated in the secondary follicles from the old mouse ovaries compared to those from the young mouse ovaries. Further, unsupervised hierarchical clustering analysis using the HemI 1.0.3.7 software showed distinct patterns of up-and downregulated genes in the secondary follicles from young and old mouse ovaries (see Figure 2).

Functional Annotation for the Differentially Expressed
Genes. The identified differentially expressed genes in the secondary follicles from the old mouse ovaries compared to those from the young mouse ovaries were further analyzed via gene ontology (GO) and KEGG pathway analysis using the DAVID online tool. As shown in Table 1, GO term enrichment analysis showed that the upregulated genes were significantly enriched in immune system process in the biological processes category, cytoplasm, and nucleus in the cellular component category and RNA and DNA binding in the molecular function category. While listed in Table 2, the functional annotation for the downregulated genes revealed that the most significant categories of biological process were involved in transcription and its regulation, cellular component was involved in nucleus and cytoplasm, and molecular function was involved in protein, RNA, and DNA binding. Furthermore, KEGG pathway analysis showed that most of the upregulated genes took part in virus related and Toll-like receptor signaling pathways, whereas downregulated genes mainly participated in PI3K-Akt signaling pathway and Adherens junction in Table 3.

Discussion
Elucidating the mechanism of ovarian aging has significant meanings to female health. The gradual decrease of both the quantity and the quality of the oocytes surrounded by the granulosa cells in all stages of follicles dominates the reproductive aging [4]. In previous study, Govindaraj et al. revealed that gene expression patterns changed considerably in the rat primordial follicles in the process of ovarian aging [25]. Folliculogenesis in the ovary is a highly dynamic and periodic process regulated by both intra-and extraoocyte factors [26]. At each reproductive cycle, activated primordial follicles join the growing pool transiting to primary follicles [26]. Through further development, a primary follicle grows into a secondary follicle [26]. And these stages are gonadotropin independent, but depend on the complex bidirectional communication between the oocyte and the somatic cells [26]. In the subsequent stages of folliculogenesis, the presence of pituitary gonadotropins, follicle-stimulating hormone (FSH), and luteinizing hormone (LH) are required [26]. So the secondary follicles from the mouse ovaries were selected as research objective in our study. Gene expression profile of the secondary follicles from the young and old mouse ovaries was compared by microarray analysis to find what changed in the process of ovarian aging in the present study. The results of our research found that 174 genes were upregulated and 197 genes were downregulated in the secondary follicles from the old mouse ovaries compared to those from the young mouse ovaries.
Further GO and KEGG pathway analyses were performed to study the function of the differentially expressed genes. The result of GO analysis showed that the upregulated genes were mainly involved in biological process such as immune system process and defense response, while downregulated genes were closely related to gene transcription and cell apoptosis. However, there was an unexpected phenomenon in the results of our functional enrichment. Many upregulated genes were involved in response to virus and interferon in the biological process and took part in several virus related signal pathway. This phenomenon revealed that the SPF mice used in our study might infect some viruses. However, the certain thing is that the immune related genes can be expressed in ovarian granulosa cells, not only in immune cells. Several earlier studies indicated that viruses can induce innate immune response in granulosa cells and perturb ovarian function in mouse [27,28]. And immune response genes were overexpressed with increasing age as showed by several microarray studies of aging [29]. Recently the concept that innate immunity is an essential requisite in the ovulation process is forwarded [30]. The important role of innate immune cells in decreasing the senescence burden was also recognized [31]. There was a probability that innate immune related genes were upregulated in the process of aging and affected the progress of ovarian aging. Yet, the actual role of innate immunity in the process of ovarian aging or folliculogenesis needs to be further researched.
We conducted protein-protein interaction network analysis of differentially expressed genes. The nodes regarded as hub genes were mostly involved in innate immune system. As in the gene regulatory network, many differentially expressed genes between young and old mouse ovarian secondary follicles such as BRCA1, STAT3, JUN, AKT1, SEPRING1, TCF3, MAP3K7, and IRF7 showed their genetic interactions by various pathways. Thus, a number of pathways were interacted through many genes in the process of ovarian aging.

Conclusions
In conclusion, our data showed quite different gene expression patterns of the secondary follicles between young and old mouse ovaries. The differentially expressed genes involved in the process of ovarian aging are central to biological processes such as immune system process, aging, transcription, DNA replication, DNA repair, protein stabilization, and apoptotic process. However, many upregulated genes in the old mouse ovarian secondary follicles were innate immune related genes. We proposed that innate immune system may play a vital role in the process of ovarian aging. Our results of altered genes and related transcriptional networks may be helpful for understanding the mechanism of the folliculogenesis in the process of ovarian aging in mice.
There are however limitations in the present study. Findings of our present research were mainly based on the bioinformatics analysis and further experiments are needed to verify. Furthermore, these data were acquired with secondary follicles from mouse ovaries and needed to be confirmed with the samples from the human.

Data Availability
The results of our microarray data were made available in the public domain NCBI-GEO repository.  Figure 5: Gene regulatory network modeling for selected differentially expressed genes by using Cytoscape software (version 3.6.1). Gene regulatory network modeling of selected genes such as BRCA1, RPA2, STAT3, JUN, AKT1, STK3, SOX4, TCF3, S100A8, and IFIT2 showed their genetic interactions by various pathways. Circles indicate genes, red color indicates upregulation, and green color indicates downregulation.