Abnormal Expression of the Steroid Hormone Synthesis Pathway Associates with Cattle Ovarian Cysts

By enzyme-linked immunosorbent assay (ELISA) of the bovine uid from , we found that cystic follicles were characterized by lower oestradiol (E 2 Insulin-like growth factor 1(IGF1), and insulin levels but elevated progesterone (P 4 ) compared with preovulatory follicles (p <0.05). Further gene expression proling of follicle walls by RNA sequencing (RNA-seq) showed that there are 356 differentially expressed genes between preovulatory follicles and corpus luteum cyst groups, and 582 differentially expressed genes between preovulatory follicles and follicular cyst groups. Kyoto Encyclopedia of Genes and analysis linked to the formation and acute delta-isomerase cytochrome family subfamily and cytochrome family 1, subfamily B, polypeptide 1 (CYP1B1) genes in the steroid hormone synthesis pathway play important roles in this process. microarray analysis investigated expression dominant cystic follicles ultrasound-guided aspiration, 163 up-regulated and down-regulated High-throughput RNA-seq is even more powerful for identifying DEGs. In the present study, large-scale gene expression patterns were determined to analyze the molecular mechanism of ovarian cysts. The results revealed that steroid hormone synthesis was linked to the formation of ovarian cysts. In particular, the expression of STAR, CYP11A1, HSD3B1 CYP17A1, and CYP1B1 genes in related to the steroid hormone synthesis pathway plays an important role in the formation of ovarian cysts. levels of messenger RNAs luteinizing hormone receptor and 3β-hydroxysteroid dehydrogenase Δ 4, Δ 5 isomerase to normal dominant follicles.

gene expression pro ling of follicle walls by RNA sequencing (RNA-seq) showed that there are 356 differentially expressed genes between preovulatory follicles and corpus luteum cyst groups, and 582 differentially expressed genes between preovulatory follicles and follicular cyst groups. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis linked steroid hormone synthesis pathway to the formation of ovarian cysts, and steroidogenic acute regulatory protein (STAR), cytochrome P450, family 11, subfamily A, polypeptide 1 (CYP11A1), hydroxy-delta-5-steroid dehydrogenase, 3-beta-and steroid delta-isomerase 1 (HSD3B1), cytochrome P450, family 17, subfamily A, polypeptide 1 (CYP17A1), and cytochrome P450, family 1, subfamily B, polypeptide 1 (CYP1B1) genes in the steroid hormone synthesis pathway play important roles in this process.

Conclusions
Abnormal hormone pro les and expression of steroid hormone synthesis pathway genes related to the formation of ovarian cysts. The ndings lay a theoretical foundation for the prevention and treatment of ovarian cysts.

Background
Ovarian cysts are the most common reproductive dysfunction in high-producing dairy cows, causing signi cant economic losses to the dairy industry by extending the calving stage, increasing days in the postpartum period, and lengthening the replacement rate due to infertility (1,2). Ovarian cysts are de ned as follicle-like structures present on one or both ovaries, with a diameter of at least 2.0 cm for a minimum of 10 days (3). Ovarian cysts can be classi ed functionally as follicular or luteal, and follicular cysts have a relatively thin wall (£3 mm) and lack any evidence of luteal tissue, while luteal cysts have thicker walls (>3 mm) (4).
Studies over many years have shown that the formation of ovarian cysts is mainly the result of hormonal imbalance within the hypothalamic-pituitary-gonadal axis caused by endogenous and/or exogenous factors (3,5,6). The most widely accepted hypothesis explaining the formation of a cyst is that luteinizing hormone (LH) release from the hypothalamus-pituitary is altered, and the pre-ovulatory LH surge is either absent, insu cient in magnitude, or occurs at the wrong time during dominant follicle maturation, leading to cyst formation (7,8,9). Meanwhile, progesterone (P 4 ) is involved in the formation of ovarian cysts, and there is a strong association between intermediate concentrations of P 4 in peripheral blood and the occurrence of ovarian follicular cysts (3). Most cysts are accompanied by a decrease of P 4 , which promotes the development of the cyst (10). Molecular analysis of bovine cystic ovarian disease pathogenesis has revealed that ovaries in animals with ovarian cysts exhibit disrupted steroid receptor patterns related to follicle-stimulating hormone receptor (FSHR), progesterone receptor (PGR), LH/choriogonadotropin receptor (LHCGR), and estrogen receptor (ESR) (1,11,12).
Negative energy balance (NEB) is another factor that can also lead to the formation of ovarian cysts (6). Imbalance between energy intake through feed and energy expenditure through milk yield during the early postpartum period causes NEB, which is usually accompanied by hormonal and metabolic adaptations that affect ovarian function (13,14). During NEB, circulating concentrations of insulin-like growth factor 1 (IGF1), insulin (13), and leptin (15,16) are reduced. Zulu et al. (17) reported that low systemic IGF-1 concentrations in early postpartum may contribute to anovulation and subsequent development of cystic follicles, while Vanholder et al. (18) reported that reduced circulating insulin concentrations in early postpartum may play a role in ovarian dysfunction (i.e. cyst formation). Spicer (2001) hypothesized that above a certain threshold level, leptin acts as a trigger to initiate hypothalamo-pituitary gonadotropin secretion. In a moderate to high leptin environment, as occurs in obesity, leptin limits ovarian steroidogenesis (19). These ndings indicate that the molecular mechanism of the formation of ovarian cysts in dairy cows is complicated.
With the development of advanced molecular genetics technologies, especially next-generation sequencing and bioinformatics, transcriptome sequencing (RNA-seq) provides a convenient platform for measuring large-scale gene expression patterns in organisms (20). The ne detail provided by sequencing-based transcriptome approaches indicates that RNA-seq is likely to become the platform of choice for interrogating steady-state RNA levels, and RNA-seq also enables the detection of differentially expressed genes (DEGs) with low expression levels (21,22). Some studies have analyzed the transcriptome pro les of liver samples from lactating dairy cows divergent in NEB (23), and the anterior pituitary of heifers before and after ovulation (24).
However, RNA-seq has not yet been widely employed for determination of large-scale gene expression patterns to explain the molecular mechanism of ovarian cyst formation. Therefore, the present study aimed to investigate hormonal and gene expression patterns in follicular and luteal cysts compared with normal preovulatory follicles (controls) via enzyme-linked immunosorbent assay (ELISA) and deep sequencing of the transcriptome to identify novel genes and their associated biological pathways that are important in ovarian cysts in cattle.

Material And Methods
Animals and clinical diagnosis of ovarian cysts All cows were from local dairy farms in Beijing. First, the ovarian conditions of nonpregnant lactating cows were monitored daily by ultrasonography using a Honda HS1600V real-time B-mode scanner equipped with a 7.5 MHz linear-array trans-rectal transducer (Honda, Toyohashi, Japan). Based on the ultrasound image, in the absence of active luteal tissue, a follicle structure diameter >20 mm for more than 15 days with a thin wall (≤3 mm) and uniformly anechogenic follicular uid was de ned as a follicular cyst, while a thicker wall (>3 mm) with a visible echogenic rim, spots, and web-like structures signi ed a corpus luteum cyst (25). Holstein cows with normal preovulatory follicles on day 18 after synchronization of estrus and without reproductive disorders were assigned to the control group.
Tissue sampling, follicular uid collection, and processing Following B-ultrasound diagnosis, cyst and preovulatory follicles were obtained from cows during slaughter at a nearby abattoir, and rinsed in ice-cold saline (0.9% NaCl). The follicular uid (one sample collected from 1 cow) of the upper follicles was extracted, centrifuged at 2000 rpm for 10 min, the supernatant was collected in sterilized Eppendorf tubes, the follicular wall was exfoliated and immediately preserved in liquid nitrogen, and samples were stored at -80°C until further analysis.

Measurement of hormone concentrations in follicular uid
ELISA was conducted to measure the concentrations of oestradiol, P 4 , insulin and IGF-1 levels in follicular uids. All ELISA kits were purchased from the Jiancheng Bioengineering Institute (Nanjing, China), and assays were performed according to the manufacturer's instructions. Brie y, 50 μL of standards or samples were added to the appropriate well of the microtiter plate pre-coated with antibody, gently mixed, and incubated for 60 min at 37°C. After washing, biotinylated anti-IgG and streptavidin-horseradish peroxidase (HRP) were added along with chromogen solutions A and B. Finally, the optical density (OD) at 450 nm was recorded using a Multiskan MK3 automatic microplate spectrophotometer (Thermo Fisher Scienti c, Waltham, MA, USA). Hormone concentrations were calculated according to standard curves, and each experiment was repeated independently at least three times.

RNA extraction, library preparation, and sequencing
Total RNA was extracted from follicular walls using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions, and RNA was puri ed using an miRNeasy kit (Qiagen, Hilden, Germany). Sequencing and bioinformatics analysis were conducted by Beijing Genomics Institute (BGI; Beijing, China). The quality of RNA was assessed using an Agilent 2100 Bioanalyzer system (Agilent Technologies, Palo Alto, CA, USA) and samples with a RNA integrity number (RIN) >8 were used for RNA library construction. Brie y, polyA+ mRNA was puri ed using oligo-dT-attached magnetic beads. Selected mRNAs were fragmented and reverse-transcribed to double-stranded cDNA (dscDNA) using N6 random primers. Ends of dscDNAs were repaired with phosphate at the 5' end and A at the 3' end to ligate with adaptors with T at the 3' end of dscDNAs, which were subjected to ampli cation (26). RNA-seq libraries were sequenced on a BGISEQ-500 instrument (BGI; www.genomics.org.cn). Detailed procedures have been published previously (27). Raw reads have been deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive database (https://www.ncbi.nlm.nih.gov/sra) under accession number PRJNA602176

Bioinformatics analysis
Raw RNA-seq data were ltered into clean reads, followed by mapping against the Bos taurus reference genome (mm10) using HISAT (28). Gene expression levels were quanti ed using the RSEM software package (29). DEGs between control and cyst groups were identi ed using fold change ≥2 and false discovery rate (FDR) ≤0.001 as criteria. Gene Ontology (GO) annotation was used to map all DEGs to GO terms in the database (http://www.geneontology.org/), and GO terms with Q-values (corrected p-value) ≤0.05 de ned DEGs as signi cantly enriched. The Kyoto Encyclopedia of Genes and Genomes (KEGG) database was used to perform pathway enrichment analysis of DEGs, and pathway terms with Q-values ≤0.05 were de ned as signi cantly enriched (30).

Statistical analysis
All data are presented as mean percentages ± standard error of the mean (SEM) from a minimum of three independent replicate experiments. Different groups were analyzed by SPSS version 12.0 (SPSS, Chicago, IL, USA) and signi cant differences between means were determined using the least signi cant difference (LSD) test for comparison of multiple means. Statistical signi cance was de ned at p <0.05.

Results
Hormonal and metabolic pro ling of follicular uid Different ovarian conditions were monitored daily by ultrasonography, and typical ovarian cysts and preovulatory follicles were obtained from Holstein cows during slaughter at a nearby abattoir. Ultrasound images and in vitro ovarian images are shown in Figure 1.

Sequencing data summary
In this study, nine RNA samples from three different follicle wall tissues, namely follicular cysts (F1-F3), corpus luteum cysts (L1-L3), and normal preovulatory follicle controls (C1-C3), were subjected to deep sequencing using an Illumina HiSeq1500 platform. For each sample, ~20 million raw reads were generated, yielding at least 15 million mapped reads ( Table 2). After being ltered, clean reads were mapped to the bovine reference genome using HISAT 10 and Bowtie2 (http://bowtie-bio.sourceforge.net/ Bowtie2 /index.shtml). The average mapping ratio with reference genes was 92.85%, and separate mapping rates for each sample are listed in Table 3.

Identi cation of expressed genes and DEGs
A total of 31097 expressed genes were detected in all three samples for preovulatory follicles and ovarian cyst groups. Gene expression levels were quanti ed by RSEM, and the number of identi ed expressed genes in each sample is shown in Figure 2. The Venn diagram shows gene expression levels among the different groups. The number of co-expressed genes was 38,466, and the number of speci cally expressed genes was 1254, 2520, and 1118 for preovulatory follicles, follicular cysts, and luteum cysts, respectively ( Figure 3).
Based on the fragments per kilobase of transcript per million mapped reads (FPKM) values of all expressed genes, Pearson's correlation values were calculated for pairwise comparisons among samples, and were >0.90 among the three biological replicates for control and cyst groups. A total of 356 genes displayed differential expression between control and corpus luteum cyst groups, among which 197 and 159 DEGs were up-and down-regulated, respectively (Figure 4). A total of 582 genes displayed differential expression between control and follicular cyst groups, among which 471 and 111 DEGs were up-and down-regulated, respectively ( Figure 4). The correlation coe cient between different samples was also investigated, and luteal cysts exhibited the lowest correlations with other groups ( Figure 5).

Clustering analysis of DEGs
In order to identify the functions of DEGs in cyst vs. control and corpus luteum cyst vs. control groups, GO analysis was performed based on three functional categories; biological process (BP), cellular component (CC), and molecular function (MF), and the results are shown in Figure 6A and 6B. Regarding BP classes, DEGs (149 genes for cyst vs. control and 97 genes for corpus luteum cyst vs. control groups, respectively) related to cellular process were the most enriched and DEGs (100 genes for cyst vs. control and 74 genes for luteum cyst vs. control groups, respectively) related to biological regulation constituted the second largest group. Regarding CC classes, DEGs associated with cell (166 genes and 99 genes), cell part (166 genes and 99 genes) and organelle (131 genes and 92 genes) were the top enriched. Finally, most MF classes were linked to binding activity (124 and 84 genes) and catalytic activity (66 genes and 47 genes).

Pathway enrichment analysis of DEGs
Genes often interact with each other to mediate speci c biological functions. Pathway enrichment analysis of DEGs was therefore performed on DEGs using the KEGG database, and DEGs identi ed in the follicular cyst vs. control and corpus luteum cyst vs. control group comparisons were mapped to KEGG metabolic and regulatory pathways with a correct p-value cutoff of p <0.05. The top 20 KEGG enrichment results are shown in Figure 7A and 7B. From the pathway analysis, steroid hormone synthesis was linked to the formation of ovarian cysts.

Key DEGs associated with the formation of ovarian cysts
Pathway analysis showed that the steroid hormone synthesis pathway is associate with the formation of both follicular and luteum cysts. Thus, the expression of key genes involved in this signaling pathway was further analyzed. The results showed that steroidogenic acute regulatory protein (STAR), cytochrome P450, family 11, subfamily A, polypeptide 1 (CYP11A1), and hydroxy-delta-5-steroid dehydrogenase, 3beta-and steroid delta-isomerase 1 (HSD3B1) were up-regulated signi cantly (p <0.05) while cytochrome P450, family 17, subfamily A, polypeptide 1 (CYP17A1) and cytochrome P450, family 1, subfamily B, polypeptide 1 (CYP1B1) were down-regulated (p <0.05).

Discussion
In the current study, we performed hormonal and metabolic pro ling of bovine ovarian cysts, and identi ed DEGs by comparing normal (control) tissues with follicular and luteum cyst samples. It is known that dominant follicle selection is followed by ovulation (31). Follicle-stimulating hormone (FSH) induces antral follicle growth, associated with elevated estradiol production, and high estradiol levels enhance hypothalamic gonadotropin-releasing hormone (GnRH) pulses, triggering a surge in LH surge (32). However, decreased estradiol levels and a failed pre-ovulatory LH surge at the appropriate time during maturation of the dominant follicle leads to the formation of ovarian cysts (1,33). Herein, hormonal pro ling showed that, compared with dominant follicles, the ratios of estradiol to P 4 intrafollicular steroid levels in ovarian cysts were decreased signi cantly (p <0.05). Braw-Tal et al. (34) reported that preovulatory follicles are characterized by high estradiol and low P 4 concentrations, and the estradiol-to-P 4 (E/P) ratio in these follicles was 42, but this ratio drops sharply to 0.91 in corpus luteum cysts. Follicular cysts contain greater P 4 (p <0.05) but lower estradiol (p <0.05) levels than noncystic follicles (35). An imbalance between estradiol and P 4 in intrafollicular uids leads to the formation of ovarian cysts.
Hypothalamic-pituitary function and follicular growth and development may be affected by NEB through metabolic and/or hormonal adaptations. Insulin and IGF-1 have been postulated as key mediators between nutritional status and ovarian function in cattle (36,37,38). In vitro and in vivo studies on cows indicate that insulin and IGF-1 stimulate both estradiol synthesis in granulosa cells and androgen synthesis in theca cells (39,40). Herein, the average IGF-1 concentration in follicular uids of follicular cysts was signi cantly lower than in uids from preovulatory follicles. Therefore, reduced insulin and IGF-1 levels may affect the follicular responsiveness to LH stimulation, which could lead to anovulation and cyst formation.
Abnormal secretion of steroid hormones and metabolic factors can explain the formation of ovarian cysts, but the speci c molecular mechanism remains unknown. Some researchers reported that follicular cysts appear to be associated with changes in the transcription of IRs, IGFRs (18), PAPP-A (41), and HSD3B1 and LH receptor genes (42), as well as reduced localization of estrogen receptor β protein and increased localization of estrogen receptor α protein (43). Kisspeptin may also be involved in the pathogenesis of bovine follicular cysts (44). Cows with ovarian cysts display abnormal steroidogenic markers (LHCGR, StAR, CYP11A1, 3β-HSD, CYP19A), immunological markers (IL-1β, IL-6, IL-8, TLR-4, TNF), and metabolic markers (IGF-1, IRS1) (9). However, high-throughput studies on the formation of ovarian cysts are lacking. A previous microarray analysis investigated gene expression in granulosa cells from dominant and cystic follicles collected from dairy cows by ultrasound-guided aspiration, revealing 163 DEGs (p <0.01), of which 19 were up-regulated and 144 were down-regulated (35). High-throughput RNAseq is even more powerful for identifying DEGs. In the present study, large-scale gene expression patterns were determined to analyze the molecular mechanism of ovarian cysts. The results revealed that steroid hormone synthesis was linked to the formation of ovarian cysts. In particular, the expression of STAR, CYP11A1, HSD3B1 CYP17A1, and CYP1B1 genes in related to the steroid hormone synthesis pathway plays an important role in the formation of ovarian cysts.

Conclusion
In conclusion, abnormal hormone pro le and gene expression of the steroid hormone synthesis pathway associates with cattle ovarian cysts. Coupled with previous studies, our current work comparing cystic and normal preovulatory follicles greatly expands our knowledge in this area, but further investigation is needed to determine cause-and-effect relationships. Therefore, future research should focus on changes occurring during follicle growth that may interfere with normal follicle development and steroidogenesis, and lead to the formation of cysts. Availability of data and materials The data supporting the conclusions of this article is included within the article and its additional le.