NRDR inhibits estradiol synthesis and is associated with changes in reproductive traits in pigs

Cumulus cells secreting steroid hormones have important functions in oocyte development. Several members of the short‐chain dehydrogenase/reductase (SDR) family are critical to the biosynthesis of steroid hormones. NADPH‐dependent retinol dehydrogenase/reductase ( NRDR), a member of the SDR superfamily, is overexpressed in pig breeds that also show high levels of androstenone. However, the potential functions and regulatory mechanisms of NRDR in pig ovaries have not been reported to date. The present study demonstrated that NRDR is highly expressed in pig ovaries and is specifically located in cumulus granulosa cells. Functional studies showed that NRDR inhibition increased estradiol synthesis. Both pregnant mare serum gonadotropin and human chorionic gonadotropin downregulated the expression of NRDR in pig cumulus granulosa cells. When the relationship between reproductive traits and single‐nucleotide polymorphisms (SNPs) of the NRDR gene was examined, we found that two SNPs affected reproductive traits. SNP rs701332503 was significantly associated with a decrease in the total number of piglets born during multiparity, and rs326982309 was significantly associated with an increase in the average birth weight during primiparity. Thus, NRDR has an important role in steroid hormone biosynthesis in cumulus granulosa cells, and NRDR SNPs are associated with changes in porcine reproduction traits.

The NADPH-dependent retinol dehydrogenase/reductase (NRDR) gene, also known as dehydrogenase/reductase short-chain alcohol dehydrogenase/reductase family member 4 (DHRS4), is located on the 14q11.2 chromosome (GenBank accession number AB045131) in the human genome, and it was initially purified from rabbit liver (Huang & Ichikawa, 1997). This gene is a member of the short-chain dehydrogenase/reductase (SDR) superfamily, and it shares the basic structural and functional features of the SDR superfamily (Du et al., 2004;Usami et al., 2003). The SDR superfamily is one of the largest and oldest protein families, and it is involved in the metabolism of a large variety of compounds, including steroid hormones, prostaglandins, retinoids, lipids, and xenobiotics (Bray, Marsden, & Oppermann, 2009;Persson et al., 2009).
NRDR is widely distributed in human organs and tissues. Because of the high activity of retinal reductase at physiological pH compared to other enzymes in the SDR superfamily (G. Lee & Bendayan, 2004), NRDR plays an important role in maintaining the homeostasis of retinal and retinoic acid. The NRDR gene cluster is located in an area of segmental duplications, and the expression of this cluster is downregulated in some common human cancers (Lu et al., 2007).
Human DHRS4 may have a novel role in human endocrinology.
Expression of DHRS4 is induced via peroxisome proliferatoractivated receptor (PPARα) activation. PPARα regulates various genes that control fatty acid catabolism, gluconeogenesis, ketone body synthesis, hormone synthesis, and cholesterol metabolism (Michalik et al., 2006). In other species, including pig, mouse, and dog, NRDR protein is also efficient at reducing retinal metabolism (Endo et al., 2007;Lei, Chen, Zhang, & Napoli, 2003;Usami et al., 2003). The human ortholog of NRDR, initially termed peroxisomal 2,4-dienoyl CoA reductase-related protein (Fransen, Van Veldhoven, & Subramani, 1999), is significantly active in reducing various aromatic ketones and alpha-dicarbonyl compounds, including 3-ketosteroids and cytotoxic 9,10-phenanthrenequinone (Endo et al., 2009;Kisiela, El-Hawari, Martin, & Maser, 2011;Matsunaga et al., 2008). NRDR is an important enzyme involved in the reductive metabolism of carbonyl groups, and it contributes to essential physiological roles in the metabolism of steroid hormones and retinoids in different species (M. S. Song, Chen, Zhang, & Napoli, 2003). NRDR is highly expressed in the pig testis, and it has been implicated in the production of steroid hormones (Moe et al., 2007;Rolland et al., 2009). Thus far, no studies have been reported on the expression and function of NRDR in the ovary. Our previous transcriptome sequencing revealed that NRDR expression varies in different stages of pregnancy in porcine ovaries (unpublished).
The primary aim of the present study was to determine NRDR expression and related functions in pig ovary. The results first demonstrated that NRDR is highly expressed in pig ovary and that it downregulates estradiol synthesis in pig cumulus granulosa cells. A relationship between reproductive traits and polymorphisms of the NRDR gene (DHRS4) was also demonstrated. One SNP was significantly associated with the total number of piglets born during multiparity, and a second SNP was significantly associated with changes in the average birth weight (ABW) during primiparity. The results describing the NRDR and E2 signaling pathway in cumulus cells increases the understanding of ovary physiology and reveals potential targets for pharmacological interventions. The results pertaining to SNPs may also be useful for marker-assisted selection and genomic selection strategies for genetic improvement programs in pigs.

| Animals
All procedures performed on animals followed the guidelines of the

| Cell culture
The ovaries used for in vitro experiments were obtained from Yorkshire pigs from a local slaughterhouse and transported to the laboratory within 2 hr of harvest. The ovaries were maintained at 37°C in a sterile physiological saline solution (0.9% NaCl, 100 IU/mL penicillin, and 100 IU/mL streptomycin). For cell culture, the follicular fluid was aspirated from 3 to 5 ml follicles using a 20 ml syringe and centrifuged at 500 g for 5 min to collect the COC. The COCs surrounded by a compact cumulus mass with evenly granulated cytoplasms were picked out using a mouth pipette under a microscope. Serum-free Burlington,ON,Canada) buffered with 10 mM HEPES and 26 mM bicarbonate was used for washing COCs. COCs were pipetted into 1.5 ml tubes, and 150 IU hyaluronidase was added and incubated for 3 min at 37°C. Oocytes were then picked out using a mouth pipette, and cumulus cells were collected by centrifugation at 1,000 g for 3 min at room temperature. Cumulus cells were then collected from the bottom of the centrifuge tube. The cumulus cells were washed three times with serum-free DMEM/F12 culture medium (Invitrogen, Carlsbad, CA). Cells were dispersed by pipetting up and down several times. Cell viability was determined using Trypan blue dye (Sigma-Aldrich, MO).
Cells were seeded at a concentration of 1 × 10 5 cells per well in 12well plates containing 1 mL of DMEM/F12 supplemented with 10% fetal bovine serum (Invitrogen, CA). Cells were cultured in a highly humidified atmosphere of 95% air and 5% CO 2 at 37°C. Porcine cumulus cells were cultured in DMEM/F12 medium with 10% fetal calf serum for 72 hr before further treatment, and the culture medium was refreshed every 24 hr.

| Cells transient transfection and treatment
An NRDR siRNA kit was purchased from RiboBio (Guangzhou, China), which contained three siRNAs for NRDR and an NC-siRNA.

| Cell viability assay and statistical analysis
Twenty-four hours after transfection, cells were cultured in 96-well plates (1 × 10 4 cells/well). Desired drugs and compounds were added to the wells separately immediately after cells were seeded and cultured with the cells for 24 hr. Cell viability was measured by fluorescence chemistry using a CCK8 kit (Lianke Bio, Beijing, China) with a spectrophotometer (Multiskan MK3; Thermo Fisher Scientific, Rockford, IL) at an optical density (OD) of 450 nm.
Each experimental condition was repeated three times with up to five multiple wells each time. Data are presented as the mean ± standard error, and statistical significance was calculated by Student's t test in Excel. P < 0.05 was considered significantly different.

| Isolation of RNA and real-time quantitative PCR (RT-qPCR)
Tissue samples (50-100 mg), which were stored at −80°C, were homogenized in 1 ml of TRIZOL (Invitrogen, CA). Ample volume did not exceed 10% of the volume of TRIZOL reagent. After homogenization, total RNA was isolated, treated with DNase I, and quantified by spectrophotometry according to the manufacturers' protocols. Purified total RNA (1 µg) was used as a template for complementary DNA (cDNA) synthesis using Moloney murine leukemia virus (M-MLV) reverse transcriptase (Thermo Fisher Scientific, CA) according to the manufacturer's instructions.
Total RNA from cumulus cells was isolated using TRIZOL reagent according to the manufacturer's directions. Purified RNA was treated with DNase I and quantified by spectrophotometry. Purified total RNA (1 µg) was used as a template for cDNA synthesis using M-MLV (Promega, CA) according to the manufacturer's instructions.
All reverse transcriptase reactions included no-template controls.
RT-qPCR was performed using an SYBR Green master mix (DRR420A, TaKaRa, Dalian, China) and an ABI PRISM 7500 Sequence Detection System (Applied Biosystems, CA). All reactions were performed in triplicate. RT-qPCR conditions were as follows: 95°C for 2 min; and 40 cycles of 95°C for 15 s and 60°C for 1 min. The −▵▵ 2 Ct method was used to determine the gene expression level (Livak & Schmittgen, 2001).
Porcine β-actin was selected as an internal control for mRNA. T tests were used to evaluate expression differences. All primers were designed using Primer 5.0 and are described in Table 1.

| Western blot
Ovaries were lysed with radioimmunoprecipitation buffer (

| Immunofluorescence assay
Adult pig ovaries were embedded in paraffin and cut into 5-μm sections. IFA of ovary sections was performed using methods described previously (Li et al., 2014). Antibodies against NRDR LIU ET AL.

| Statistical analysis
All expression and hormone experiments were independently performed three or more times. All data were analyzed using oneway analysis of variance followed by Student's t test. The values are presented as the mean ± SEM. Statistical analysis was performed DHRS47-reverse TGCAGTCAGAGTGGAAACCC using SPSS 10.0 (SPSS Inc., IL). A value of p < 0.05 was considered to be statistically significant.
The genotypic and allelic frequencies of each SNP were calculated using PopGene 3.2 software. Association analysis between SNPs and reproduction traits was performed using the general linear model (GLM) procedure of SAS 9.2 statistical software with the following fixed effect model: where y represents the phenotype records of reproduction traits; μ is the overall mean; gi is the fixed effect of the genotype; mk is the fixed effect of month-old; and e is the random error (Zhou et al., 2017).
Statistics are presented as probability values and least squares means ± standard error. The thresholds for statistical significance were *p < 0.05, **p < 0.01, and ***p < 0.001.

| NRDR expression in pig cumulus granulosa cells
Messenger RNA (mRNA) expression levels for NRDR were initially detected in different pig tissues using real-time quantitative polymerase chain reaction (RT-qPCR). The NRDR mRNA level was highest in the liver. In the reproductive system, NRDR expression was high in the ovaries and uterus as shown in Figure 1a. Furthermore, NRDR expression was confirmed in the ovary by immunofluorescence assay (IFA). As shown in Figure 1b, NRDR was present in all granulosa cells, and no NRDR signal was observed in the corpus luteum or stromal cells. These results suggested that NRDR is involved in the regulation of steroid hormone synthesis and cell proliferation in cumulus granulosa cells.

| High NRDR expression is present in pig breeds that have low levels of E2
Previous studies have shown that a significant difference exists in the reproductive ability between Meishan and Yorkshire pigs (Hunter et al., 1996;Hunter, Faillace, & Picton, 1994;Miller, Picton, Craigon, & Hunter, 1998;Sun et al., 2011). In addition, the level of E2 in the follicular fluid of Meishan pigs is significantly higher than that in Yorkshire pigs (Hunter, Biggs, & Faillace, 1993;Miller et al., 1998). Therefore, the difference of NRDR expression in Yorkshire and Meishan pig ovaries in estrus at two time points after sexual maturity (postnatal Days 180 and 300) was detected by RT-qPCR and western blot to elucidate the relationship between NRDR and reproductive ability. Higher levels of both NRDR mRNA and protein were expressed in Yorkshire pig ovaries compared with Meishan pig ovaries (Figure 2a-c), that is, NRDR was overexpressed in pig breeds with lower levels of E2 and reproductive ability. The above results suggested that NRDR may inhibit E2 synthesis in ovaries and that NRDR has a relationship with the reproductive ability of pigs.

| NRDR inhibits E2 synthesis in pig cumulus granulosa cells
The effects of NRDR on steroid hormone production were determined in cultured pig cumulus granulosa cells using NRDR small-interfering RNAs (siRNAs). The expression of NRDR was inhibited with siRNA, and β-actin siRNA was used as a positive control (Figure 3a). NRDR-siRNA2 and NRDR-siRNA3 decreased NRDR mRNA levels by approximately 50-60% in cultured cells 24 hr after transfection similar to the levels of the positive control (Figure 3b). Inhibition of NRDR expression caused a twofold increase of E2 levels in the medium compared to the negative control (NC)-siRNA ( Figure 3c). However, P4 levels in the medium did not change significantly after NRDR inhibition (Figure 3d).
To assay the effect of NRDR on E2 synthesis in cultured cells, E2 levels in the cultured cells were measured after NRDR knockdown. NRDR siRNAs had a significant effect on E2 synthesis in cultured cells ( Figure   3e), but no significant changes in P4 levels were observed in cells after NRDR inhibition (Figure 3f). Summation of the total levels of E2 and P4 in the media and cells showed that both siRNAs significantly affected E2 level (Figure 3g) but did not affect P4 level (Figure 3h).

| NRDR affects enzymes involved in E2 synthesis
The influence of NRDR inhibition on the mRNA levels of the following enzymes involved in E2 synthesis was determined: steroidogenic acute regulatory protein (StAR), cytochrome P450 family 11 subfamily A member 1 (cyp11A1), 3B-hydroxysteroid dehydrogenase (3B-HSD), and with siRNAs significantly downregulated HSD17B4 mRNA levels ( Figure 4d) but did not significantly affect StAR, cyp11A1, and 3β-HSD mRNA levels (Figure 4a-c). These results showed that NRDR affects E2 synthesis in pig cumulus granulosa cells.

| NRDR does not affect viability and proliferation of pig cumulus granulosa cells
To determine if changes in estradiol levels were due to changes in cell viability or proliferation, the influence of NRDR siRNAs on viability and proliferation of pig cumulus granulosa cells was determined using a cell counting kit-8 (CCK8) kit and flow cytometry.
Inhibition of NRDR with siRNA had no effect on viability and proliferation of pig cumulus granulosa cells as shown in Figure 5a,b.
Human CG and PMSG treatment had no effect on NRDR expression after 12 and 24 hr of treatment.

| Relationship between reproductive traits and the NRDR gene (DHRS4) polymorphisms
The relationship between reproductive traits and the NRDR gene SNPs were successfully genotyped using Sequenom MassARRAY (Table 3). Among these three SNPs, two were significantly associated with changes in reproductive traits (Table 4).
SNP rs701332503 (chr7:75245594 C > T) was a variant located in the coding region of DHRS4. This SNP was significantly associated with changes in the total number of piglets born (p < 0.01; Table 4) during multiparity. Animals with mutation genotype CC in rs341891833 had more piglets than those with genotypes TT and CT.
SNP rs326982309 (chr7: 75253401 A > T) was an intron variant located in the intron region of NM_214019.2 of DHRS4. This SNP was significantly associated with changes in ABW (p < 0.05; Table 4) during primiparity. Animals with mutation genotype TA in rs326982309 had higher birth weights than those with genotypes TT and AA.

| DISCUSSION
The present work showed that NRDR is expressed in pig cumulus granulosa cells and is involved in the regulation of E2 synthesis.
Previous studies have reported that NRDR is related to human cervical cancer, breast cancer, and other cancer tissues (Korkola et al., 2007;X. H. Song et al., 2007), but studies showing the systematic expression of NRDR in different tissues of the pig have not been published. NRDR was expressed at higher levels in the ovary compared to other tissues, except the liver. NRDR was specifically expressed in cumulus granulosa cells (shown by immunofluorescence assay), which corresponds to the area of steroid hormone synthesis in pig ovary (Yamashita et al., 2003). Studies have shown that the level of E2 in Meishan pig follicular fluid is significantly higher than that in Yorkshire pigs (Miller et al., 1998). Different expression levels of NRDR exist between Yorkshire and Meishan pig ovaries. The F I G U R E 6 PMSG and hCG inhibit NRDR expression in pig cumulus granulosa cells. Changes in NRDR mRNA levels over time after treatment with (a) 10 IU/ml hCG or (b) 10 IU/ml PMSG. Results are shown as the means ± SEM of three independent experiments conducted using triplicates and normalized to control treatment, *p < 0.05 (t test). Con: control; hCG: human chorionic gonadotropin; mRNA: messenger RNA; NRDR: NADPH-dependent retinol dehydrogenase/reductase; PMSG: pregnant mare serum gonadotropin; SEM: standard error of mean higher expression of NRDR in pig breeds with low levels of E2 is related to the NRDR function of inhibiting E2 synthesis.
NRDR is overexpressed in swine breeds with high androstenone levels (Grindflek, Berget, Moe, Oeth, & Lien, 2010;Leung, Bowley, & Squires, 2010;Moe et al., 2007). Several members of the SDR family are important in catalyzing an essential step in the biosynthesis of all classes of active steroid hormones (Penning, 1997). The present results showing that NRDR inhibits E2 synthesis in pig cumulus granulosa cells are consistent with the above reports. NRDR catalyzes the reduction of 3-keto-C19/C21-steroids into corresponding 3B-hydroxysteroids (Matsunaga et al., 2008). In a rabbit, pig, dog, and human, DHRS4 exhibits high reductase activity towards aromatic ketones and α-dicarbonyl compounds as well as low dehydrogenase activity towards some alcohols (Endo et al., 2007;Usami et al., 2003).
Studies in hamsters, pig, mice, and rats have demonstrated that estradiol, estrone (interconvertible metabolite of estradiol), and their catechol metabolites exist in the kidney, uterus, ovary, and mammary glands (Prater, Horton, & Thompson, 2015;Yager, 2015). According to these results, NRDR may play an important role as a reductase in the interconversion between estrone and estradiol.
StAR, cyp11A1, 3β-HSD, and HSD17Β4 are key enzymes in steroid hormone biosynthesis (Fukami, Homma, Hasegawa, & Ogata, 2013;Robic, Faraut, & Prunier, 2014). The present study evaluated the expression of these enzymes after NRDR was inhibited and found that only HSD17Β4 expression was downregulated. Moreover, HSD17Β4 catalyzes the last steps in the formation of androgens and estrogens (Payne & Hales, 2004). NRDR may affect E2 biosynthesis at the last steps by inhibiting HSD17B4 expression without affecting upstream enzymes. HSD17Β4 is expressed as the predominant dehydrogenase in several pig tissues (Kaufmann, Carstensen, Husen, & Adamski, 1995). HSD17Β4 has also been shown to efficiently inactivate estrogens in several tissues due to the preference for steroid oxidation Breitling, Marijanović, Perović, & Adamski, 2001). Similarly, when NRDR was inhibited in the present study, HSD17Β4 expression was downregulated, and E2 biosynthesis was upregulated. The change in HSD17B4 expression after NRDR inhibition indirectly suggested that NRDR affects E2 biosynthesis. In addition, both NRDR and HSD17Β4 are expressed at higher levels in high androgen pigs, supporting the idea that NRDR participates in steroid hormone biosynthesis (Grindflek et al., 2010;Leung et al., 2010;Moe et al., 2007). showing that the level of steroid hormones has no direct relationship on the proliferation of cumulus granulosa cells (Elis et al., 2015). The effects of hCG and PMSG on NRDR expression were also examined.
Previous studies have shown that differential expression of alleles is quite common in mammals and that variations may contribute to phenotypic variability (Lo et al., 2003;Yan, Yuan, Velculescu, Vogelstein, & Kinzler, 2002). SNPs may affect the expression of NRDR. Because inhibition of NRDR increased E2 levels, we speculated that the level of E2 was higher in pigs with NRDR mutations. According to previous reports, high levels of E2 are unfavorable for embryo implantation and may lead to reduced number of piglets (Burghardt, Bowen, Newton, & Bazer, 1997;Geisert, Renegar, Thatcher, Roberts, & Bazer, 1982). The present results showed that pigs with SNPs in the DHRS4 gene had fewer piglets born, which was in agreement with previous reports. Furthermore, the present SNP results may be useful for marker-assisted and genomic selection strategies for genetic improvement programs in pigs. However, further studies using methods, such as Cas9 editing and gene silencing analysis, are needed to determine the biological functions of these significant SNPs.
In summary, the present study demonstrated that NRDR is highly expressed in pig ovaries, specifically in cumulus granulosa cells.
Functional studies showed that NRDR is involved in regulating E2 synthesis in pig cumulus granulosa cells but has no effect on the proliferation of pig cumulus granulosa cells. PMSG and hCG both downregulated the expression of NRDR in pig cumulus granulosa cells. The relationship between reproductive traits and polymorphisms in the NRDR gene (DHRS4) was also analyzed. The rs701332503 SNP was significantly associated with changes in the total number of piglets born during multiparity, and the rs326982309 SNP was significantly associated with changes in ABW during primiparity.
Thus, these findings demonstrated that NRDR has an important role in pig steroid hormone biosynthesis. SNP analysis and further verification of the role of these polymorphisms may lead to the use of NRDR as a selective marker to improve pig reproduction.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.

AUTHOR CONTRIBUTIONS
K. L., R. Z., Z. G., H. A., and Y. L. conceived the project and designed experiments; Y. L. and Y. Z. analyzed data; Y. Z. and Y. Y. collected samples; Y. L., W. L., and Y. Z. performed experiments; and Y. L. and Y. Y.
wrote the manuscript. All authors read and approved the final manuscript.