Inactivation of growth differentiation factor 9 blocks folliculogenesis in pigs

Abstract

Growth differentiation factor 9 (GDF9) is a secreted protein belonging to the transforming growth factor beta superfamily and has been well characterized for its role during folliculogenesis in the ovary. Although previous studies in mice and sheep have shown that mutations in GDF9 disrupt follicular progression, the exact role of GDF9 in pigs has yet to be elucidated. The objective of this study was to understand the role of GDF9 in ovarian function by rapidly generating GDF9 knockout (GDF9^−/−) pigs by using the CRISPR/Cas9 system. Three single-guide RNAs designed to disrupt porcine GDF9 were injected with Cas9 mRNA into zygotes, and blastocyst-stage embryos were transferred into surrogates. One pregnancy was sacrificed on day 100 of gestation to investigate the role of GDF9 during oogenesis. Four female fetuses were recovered with one predicted to be GDF9^−/− and the others with in-frame mutations. All four had fully formed oocytes within primordial follicles, confirming that knockout of GDF9 does not disrupt oogenesis. Four GDF9 mutant gilts were generated and were grown past puberty. One gilt was predicted to completely lack functional GDF9 (GDF9^−/−), and the gilt never demonstrated standing estrus and had a severely underdeveloped reproductive tract with large ovarian cysts. Further examination revealed that the follicles from the GDF9^−/− gilt did not progress past preantral stages, and the uterine vasculature was less extensive than the control pigs. By using the CRISPR/Cas9 system, we demonstrated that GDF9 is a critical growth factor for proper ovarian development and function in pigs.

Inactivation of growth differentiation factor 9 revealed its critical role for proper ovarian development and function in pigs.

Introduction

Folliculogenesis is the orchestrated progression of ovarian follicles by which the oocyte within the follicle continues to grow and develop by communicating with the surrounding somatic cells. As the communication between oocytes and somatic cells is bidirectional, the oocyte secretes growth factors that promote proliferation, differentiation, and survival of the somatic cells for overall follicular progression. One crucial signaling network known to be essential for folliculogenesis is the transforming growth factor beta (TGF-β) superfamily [1–3]. Importantly, growth differentiation factor 9 (GDF9), a member of the TGF-β superfamily, was the first oocyte-derived growth factor determined to be necessary for follicular somatic cell function [4]. Growth differentiation factor 9 is expressed from the primary follicle stage onwards and is required for acquisition of follicle-stimulating hormone (FSH) receptors by the granulosa cells [5–7]. During antrum formation, GDF9 promotes synthesis of hyaluronic acid and expansion of the cumulus cells in preparation for ovulation [8]. Moreover, GDF9 has antiapoptotic effects on granulosa cells through activation of the phosphatidylinositol 3-kinase/Akt pathway, thereby preventing follicular atresia [6]. Activation of GDF9 accompanies other growth factors in the TGF-β signaling pathway for folliculogenesis to proceed. A second TGF-β member, bone morphogenic protein 15 (BMP15), is another oocyte-secreted factor that has an important role in this process. Mature GDF9 or BMP15 ligands derived from the C-terminus of the proteins can either homodimerize or heterodimerize to induce TGF-β signaling [9]. Heterodimers exhibit strong activity across species, whereas homodimers of GDF9 have been shown to be active in rodents’ ovarian follicles with no activity in humans or sheep, and BMP15 homodimers show no activity in rodents and low activity in humans and sheep [9]. However, GDF9 and BMP15 dimer activity is unknown in pigs that are a litter-bearing species.

Mutations in GDF9 have demonstrated the critical role of this gene in ovarian function and fertility. In mice, knockout of Gdf9 results in female infertility with no progression past the primary follicular stage as well as the formation of single or bilateral ovarian cysts, whereas knockout of Bmp15 results in subfertility with decreased ovulation rates [4, 8]. Oocytes from Gdf9 knockout mice were shown to grow faster than control oocytes; however, the number of transzonal projections was decreased in the Gdf9 knockout oocytes, ultimately resulting in follicular atresia [10]. Furthermore, a naturally occurring single amino acid change in the C-terminus of GDF9 in sheep resulted in no follicular progression past the primary stage, numerous collapsing follicles, and a grossly underdeveloped reproductive tract [11]. In humans, natural mutations in GDF9 were shown to contribute to diminished ovarian reserve through decreased granulosa cell proliferation [12]. The phenotypes of Gdf9 knockout mice and naturally occurring polymorphisms in sheep and humans indicate that the role of GDF9 should be widely conserved in mammals. However, because of technical challenges, no other GDF9 knockout model has been generated beyond the mouse.

Pigs are a highly important agricultural species and have emerged as a leading biomedical model for genetic diseases in humans. Because of the prolonged gestation period and length of time to sexual maturity, genetically engineered pig models are seldom produced to elucidate biological pathways. In this study, the first genetically engineered livestock model for GDF9 was generated in a rapid timeframe (<6 months) by delivering the CRISPR/Cas9 system into one-cell stage embryos, and GDF9 knockout pigs were produced to understand the role of GDF9 during oogenesis and folliculogenesis.

Materials and methods

Ethics statement

Use of animals was in accordance with approved protocol and standard operating procedures by the Animal Care and Use Committee of the University of Missouri (Protocol 13520).

Guide design for targeting GDF9

To disrupt the porcine GDF9 gene, three single-guide RNAs (sgRNAs) were designed (two within exon 1 and one within exon 2) by using the CRISPOR web-based program (http://crispor.tefor.net/). Each designed sgRNA was chosen based on criteria, such as specificity scores, predicted efficiency, and number of predicted off-targets. Sequences of sgRNAs are presented in Figure 1. In vitro transcription of the sgRNAs was conducted as previously described [13, 14].

Figure 1.

Schematic of the generation of GDF9-edited fetuses and pigs. Two sgRNAs were designed to target exon 1, and one sgRNA was designed to target exon 2 of the porcine GDF9 gene. Single-guide RNAs and Cas9 mRNA were injected into zygotes. After culture, blastocyst-stage embryos were transferred into surrogates to establish pregnancies for fetal collections or farrowing.

In vitro fertilization

In vitro fertilization (IVF) was conducted as previously described [15]. Briefly, oocytes from terminal cross gilts or sows were collected from either abattoir-derived ovaries or purchased from DeSoto Biosciences, Inc. Cumulus oocyte complexes (COCs) were placed in maturation medium (TCM-199 medium containing 0.1% polyvinyl alcohol, 3.05 mM D-glucose, 0.91 mM sodium pyruvate, 10 μg/mL gentamicin, 0.57 mM cysteine, 10 ng/mL EGF, 0.5 μg/mL FSH, 0.5 μg/mL LH, 40 ng/mL FGF2, 20 ng/mL LIF, 20 ng/mL IGF1) at 38.5°C in 5% CO₂ for 42–44 h. Following maturation, COCs were denuded, and oocytes with a visibly extruded polar body were selected for IVF. Groups of 25–30 mature oocytes were placed in 50 μL droplets of IVF medium (modified Tris-buffered medium with 2 mg/mL BSA and 2 mM caffeine; mTBM) in a mineral oil overlay. Sperm from one Large White × Yorkshire boar was washed by centrifugation in 60% Percoll solution followed by mTBM. The spermatozoa pellet was resuspended in mTBM medium to 0.5 × 10⁶ cells/mL. Then, 50 μL of the sperm suspension was added to the droplets to obtain a final concentration of 0.25 × 10⁶ cells/mL. Gametes were incubated together in a humidified incubator with an atmosphere of 5% CO₂ in air at 38.5°C for 4 h. Before microinjection, presumptive zygotes were transferred into MU4 medium and incubated at 38.5°C and 5% CO₂ in air [16].

Microinjection of the CRISPR/Cas9 system and embryo transfer

Microinjection was conducted as previously described [17]. Specifically, after IVF, presumptive zygotes were injected with ~10–20 pL of the CRISPR/Cas9 system to disrupt GDF9. The sgRNA and Cas9 mRNAs were mixed and diluted in water prior to the microinjection; the concentration of each sgRNA was 10 ng/μL and Cas9 mRNA was 20 ng/μL, and embryos were held in manipulation medium (TCM199 with 0.6 mM NaHCO₃, 2.9 mM HEPES, 30 mM NaCl, 10 ng/mL gentamicin, and 3 mg/mL BSA) covered by mineral oil. Groups of 50 injected zygotes were transferred into 500 μL of MU4 [16] and cultured for 5–6 days at 38.5°C, 5% CO₂, and 5% O₂ before transfer into a surrogate gilt. Blastocyst- and morula-stage embryos (40–50) were surgically transferred into the ampullary–isthmic junction of a cycling gilt on day 4 of her estrous cycle. Pregnancy was determined by heat checking and monitoring by ultrasound after day 25. Two of four surrogates became pregnant after transfer of GDF9-edited embryos. One surrogate was euthanized on day 100 of gestation to determine the effect of editing GDF9 on oogenesis, and the other surrogate was allowed to farrow (Figure 1).

Genotyping of embryos, fetuses, and piglets

To determine the efficiency of each sgRNA, genomic DNA (gDNA) was isolated from individual embryos at the blastocyst stage. A single embryo was placed in lysis buffer (50 mM KCl, 1.5 mM MgCl₂, 10 mM Tris–HCl pH 8.5, 0.5% Nonidet P40, 0.5% Tween-20, and 200 μg/mL Proteinase K) and incubated at 65°C for 30 min followed by 85°C for 10 min. Target regions were amplified by using DreamTaq DNA polymerase (Thermo Fisher Scientific, Waltham, MA, USA), and primers for exon 1 were Forward 5′-GGGTGTGAAGAGCCGGACAG and Reverse 5’-CAGCTGTGTGTCGGCTATAACACC and for exon 2 were Forward 5′-CATGTAACATGACTCTTCTGGCAGCCC and Reverse 5′-GAACATCATCCACGAGAAGCTCGACTC. PCR conditions were as follows, 95°C for 3 min, 33 cycles of 95°C for 30 s, 60°C for 30 s, 72°C for 30 s, and 72°C for 5 min. Amplicons were inserted into the pCR™2.1 vector for TOPO cloning and sent for Sanger sequencing to identify types of alleles introduced by the CRISPR/Cas9 system. After validation of sgRNAs and embryo transfers, fetal or piglet gDNA was isolated from ear samples by using the PureLink Genomic DNA kit (Thermo Fisher Scientific) following the manufacturer’s protocol and subjected to the same genotyping protocol.

Tissue collection

On day 100 of gestation, one surrogate pig was euthanized, and the fetuses were manually removed from the uterine horns. Wild-type (GDF9^+/+) fetuses at the same age were used as controls. Reproductive tracts of the females were exposed and weighed. The left ovary was fixed in 10% neutral buffered formalin, and the right ovary was snap-frozen for RNA extraction. Once the other surrogate farrowed, females were maintained until 7 months of age, ensuring that the age of puberty was surpassed (5–6 months for domestic pigs). Similarly, reproductive tracts were weighed, and sections of the left ovary, uterus, cervix, and vagina were fixed in 10% neutral buffered formalin. The right ovary was snap-frozen for RNA extraction.

RNA extraction

RNA was extracted from snap-frozen ovarian tissues by using the RNeasy plus mini kit (Qiagen) and reverse transcribed into complementary DNA (cDNA) by using the SuperScript VILO cDNA Synthesis Kit (Thermo Fisher Scientific) following the manufacturer’s protocol. Genotyping by using ovarian cDNA was used to confirm results with gDNA. The coding region encompassing all three sgRNAs was amplified by using Forward 5′-ATGGCGCTTCCCAGAAA and Reverse 5′-AATACGAAGACATGATCGCAAC with conditions of 95°C for 3 min, 37 cycles of 95°C for 30 s, 58°C for 30 s, 72°C for 1 min, and 72°C for 5 min followed by Sanger sequencing.

Histological analysis and immunofluorescence

Samples were submitted to the University of Missouri Veterinary Medical Diagnostic Laboratory for embedding and hematoxylin and eosin (H&E) staining. Fixed tissues were trimmed, processed by using a Sakura Tissue-Tek VIP 6 tissue processor (Torrance, CA, USA), embedded in paraffin, sectioned at 4 μm, and stained with H&E to observe architecture of the ovary, uterus, cervix, and vagina between GDF9^+/+ and GDF9-edited fetuses and gilts.

For immunofluorescence, unstained ovarian tissues were rehydrated, and antigen retrieval was performed by incubating in diluted Reveal Decloaker (Biocare Medical, Pacheco, CA, USA) for 30 min at 100°C. Tissues were permeabilized by using PBST (PBS with 0.1% Tween 20) for 20 min at room temperature (RT) and washed with PBS. After blocking with 2% goat serum, the primary antibody, anti-mouse GDF-9 (Novus Biologicals, Littleton, CO, USA), was diluted 1:500 in 2% goat serum, applied to the sections, and incubated at 4°C overnight. Sections were washed twice with PBS, and the Alexa 647-conjugated goat anti-mouse secondary antibody (Thermo Fisher Scientific) was diluted 1:400 in 2% goat serum and incubated on the sections for 1 h at RT. Sections were stained with Hoechst 33342 (1:2000 in MilliQ H₂O) for 5 min, washed, and mounted for visualization. The negative control received the secondary antibody only. Fluorescent images were collected with a Leica DM5500 B upright microscope by using the Leica Application Suite X.

Results

Generation of GDF-edited fetuses and pigs

Two surrogates became pregnant from the four embryo transfers. One pregnancy was used to collect fetuses on day 100 of gestation to determine the role of GDF9 on oogenesis, and the other pregnancy was used to understand the importance of GDF9 on folliculogenesis. The pregnancy collected on day 100 of gestation contained four female and two male fetuses that all appeared to be normal and carried targeted modification of GDF9. Fetus 3 (female) was predicted to be a GDF9 knockout (GDF9^−/−) as all other female fetuses possessed genome edits that were multiples of three, thus leading to in-frame mutations (Supplementary Table S1). The two male fetuses also had genome edits in GDF9 (data not shown). The pregnancy that farrowed resulted in four females and six males with one male being stillborn. The four gilts were maintained until 7 months of age to ensure that they reached puberty (Supplementary Figure S1). Genotyping results varied between gDNA from the ear and cDNA from the ovary for gilts B1, B3, and B4; however, gilt B1 was predicted to demonstrate a GDF9^−/− phenotype based upon both genotyping assays (Supplementary Tables S2 and S3). All six males possessed genome edits in GDF9 (data not shown), and one male that was predicted to be GDF9^−/− was maintained until sexual maturity and bred to a wild-type gilt. The genotype of the GDF9^−/− boar is presented in Supplementary Table S2. The genome editing efficiency of GDF9 for fetuses and pigs was 100%, which allows for rapid generation of founders for functional studies.

Fetal ovary development nor oogenesis is impacted by disruption of GDF9

Gross ovarian morphology was not observed to be different between day 100 GDF9-edited fetuses and GDF9^+/+ fetuses (Figure 2A), and the ratio of ovary weights per body weight was not noticeably different among the fetuses (Table 1). Examination of the histological architecture of the ovaries from GDF9^+/+ and GDF9-edited fetuses revealed fully formed oocytes within primordial follicles in the cortex (Figure 2B). Specifically, fetus 3 (GDF9^−/−) did not present any abnormalities in oogenesis and follicular formation, which is supported by the fact that GDF9 is not secreted from the oocyte until the primary follicular stage [18].

Figure 2.

Analysis of ovaries from GDF9-edited fetuses. (A) Images of reproductive tract development in GDF9^+/+ and GDF9^−/− fetuses. Ovaries are denoted by white arrows in the top pictures. (B) Histological examination of the ovarian cortex from GDF9^+/+ and GDF9-edited fetuses. Scale bars equal 100 μm.

Table 1.

Body and ovary weights of GDF9-edited and WT fetuses

Fetus	Body weight (g)	Left ovary weight (g)	Right ovary weight (g)	Combined ovary weight/body weight (%)
1	1040	0.12	0.10	0.021
2	833	0.04	0.04	0.010
3	1180	0.06	0.06	0.010
4	1200	0.07	0.07	0.012
Day 91 GDF9+/+	679	0.06	0.05	0.016

Abnormal development of reproductive tracts and lack of follicular progression in the absence of GDF9

The four GDF9-edited gilts were maintained until 7 months of age to ensure that they reached sexual maturity. Between six and 7 months of age, gilts B2, B3, and B4 showed signs of standing estrus; however, gilt B1 (GDF9^−/−) did not demonstrate standing estrus. After excision of reproductive tracts at 7 months of age, gilt B1 exhibited an underdeveloped reproductive tract compared with the other three gilts or the age matched wild-type (GDF9^+/+) gilt, and the reproductive tract weight and appearance was more similar to one from a GDF9^+/+ prepubertal 4-month-old gilt (Table 2 and Figure 3). However, the ovaries from the GDF9^+/+ prepubertal gilt have visible follicles on the surfaces, whereas the ovaries from B1 did not have visible follicles but two cysts (Figure 3; black arrows). Gilts B2, B3, and B4 had corpora lutea, indicating ovulations, but gilt B4 also demonstrated abnormal size differences between the left and right ovaries and the presence of cysts (Table 2 and Figure 3; black arrows).

Table 2.

Reproductive tract and ovary weights of GDF9-edited and WT gilts

Gilt ID	Reproductive tract weight (g)	Left ovary weight (g)	Right ovary weight (g)
B1	157	1.17	0.98
B2	817	6.22	6.69
B3	734	7.26	5.02
B4	988	6.06	24.2
Prepubertal GDF9+/+	162	1.87	2.13
Adult GDF9+/+	822	5.89	5.35

Figure 3.

Gross morphology of reproductive tracts and ovaries from GDF9^+/+ prepubertal, GDF9^+/+ postpubertal, and GDF9-edited gilts.

Expression of GDF9 has been detected at the primary follicle stage and all subsequent stages [5], and as expected, ovarian sections from GDF9^+/+ and GDF9-edited fetuses did not exhibit signal for GDF9 by immunofluorescence (data not shown). In the sexually mature gilts, GDF9 protein could not be detected in the ovarian sections from gilt B1, confirming the knockout of the GDF9 gene, whereas GDF9 was detected in the follicles and surrounding tissues of ovaries from B2, B3, and B4 (Figure 4). Gilt B4 had some follicles that did not have a strong GDF9 signal (Supplementary Figure S2), which may correspond to mosaicism detected in her ovary by genotyping. Histological analysis demonstrated that gilt B1 (GDF9^−/−) did not have follicular progression past the preantral stages as opposed to the other gilts (Figure 5), and several follicles were undergoing atresia with collapsing oocytes (Supplementary Figure S3; black arrows). Furthermore, vasculature within the uterus of B1 was underdeveloped and was comparable to the GDF9^+/+ prepubertal gilt. Although smaller in size, the cervix and vagina of B1 appeared to be morphologically normal (Figures 3 and 5).

Figure 4.

Immunofluorescence staining of ovarian sections from GDF9^+/+ postpubertal and GDF9-edited gilts. Scale bars equal 200 μm.

Figure 5.

Representative histological images of ovarian, uterine, cervical, and vaginal sections from GDF9^+/+ prepubertal, GDF9^+/+ postpubertal, and GDF9-edited gilts. Scale bars equal 100 μm.

Disruption of GDF9 does not impact fertility of the male

One GDF9^−/− boar was maintained until sexual maturity and was bred to a wild-type gilt. Fifteen fetuses were collected on day 36 of gestation, indicating that the GDF9^−/− boar was fertile. All fetuses except two appeared to be healthy and normal (Figure 6A and B). Genotypes of the fetuses were expected from the boar and are presented in Supplementary Table S4.

Figure 6.

Images of fetuses from breeding the GDF9^−/− boar. (A) Two healthy fetuses. (B) Two fetuses exhibiting signs of degeneration.

Discussion

The necessity of GDF9 during folliculogenesis has been demonstrated in multiple species, including mice, sheep, and humans [4, 11, 12]. However, the role of GDF9 during oogenesis or folliculogenesis has not been experimentally demonstrated beyond mouse models because of technical challenges of genetic engineering and the prolonged length of time to follow the entire process of folliculogenesis, i.e. reaching puberty. Pigs are an agriculturally important animal and are emerging as a preferred model for genetic diseases in humans [19–21]. Importantly, the litter size of pigs is highly determined by proper growth factor signaling within the ovarian follicles [22]. In the current study, ablation of functional GDF9 was experimentally performed, and its roles in ovarian and uterine development were investigated for pigs.

Prior studies have shown that GDF9 is not expressed in oocytes until the primary stage, at which it promotes granulosa cell proliferation, FSH signaling, and antrum formation [5, 7, 23]. This is consistent in pigs as the GDF9^−/− fetus in the current study had fully formed oocytes within primordial follicles throughout the ovarian cortex, similar to the GDF9^+/+ controls. After confirming the involvement of GDF9 during oogenesis, subsequent investigations were focused on investigating the role of GDF9 during folliculogenesis. The GDF9-edited gilts were grown past puberty, and only one gilt (B1) had a GDF9^−/− genotype. Strikingly, the GDF9^−/− gilt had a severely underdeveloped reproductive tract that was comparable to a GDF9^+/+ prepubertal gilt at 4 months of age. The lack of follicular progression and presence of atresia in the GDF9^−/− gilt confirmed that GDF9 is essential for this process. The gilt did not show any signs of standing estrus, further indicating that she was not cycling and infertile.

Infertility observed in the GDF9^−/− gilt uncovers species-specific actions of TGF-β signaling on folliculogenesis. After proteolytic cleavage from the C-terminus, mature GDF9 or BMP15 ligands can homodimerize or heterodimerize. Homodimers of GDF9 have been shown to demonstrate moderate activity in the follicles of rodents but no activity in humans or sheep, whereas BMP15 homodimers have no activity in rodents and low activity in humans and sheep, pointing to these species-specific regulation mechanisms [9]. In all species studied, heterodimers of GDF9 and BMP15 have the most potent activity for activation of TGF-β signaling to stimulate granulosa cell proliferation and steroidogenesis during folliculogenesis [9, 24]. Contrary to Gdf9^−/− mice, Bmp15^−/− knockout mice demonstrated subfertility with decreased ovulation rates, but overall follicular and corpora lutea development was normal [8]. However, analysis of BMP15^−/− cloned pigs revealed underdeveloped reproductive tracts and ovaries [25], which was similar to the GDF9^−/− gilt in the current study. This underdeveloped phenotype is likely the result of reduced estradiol production by granulosa cells, which was noted in pigs with a knockdown of BMP15 [26]. Moreover, oogenesis was not perturbed after histological examination of ovaries from 1-day-old BMP15^−/− pigs, but several abnormal and atretic preantral follicles were observed in the ovaries of BMP15^−/− postpubertal gilts. These functional studies revealed that GDF9 may have a more conserved role across species than BMP15, although both are essential for folliculogenesis in pigs.

Understanding the role of key players during germ cell development is important for production of genetically engineered pigs and preservation of oocytes or sperm. Disruption of the nanos C2HC-type zinc finger 2 (NANOS2) gene in pigs blocked germ cell development in the male, but females were normal and fertile [27]. Although the role of GDF9 in female reproduction has been investigated in several species, its function has not been fully elucidated for male reproduction. In rats, Gdf9 was shown to be produced in round spermatids and regulated formation of tight junctions between Sertoli cells by inhibiting expression of tight junction-associated proteins through TGF-β signaling [28]. Similar results were observed in cattle Sertoli cells in vitro; however, exogenous GDF9 was shown to promote Sertoli cell proliferation and inhibit apoptosis [29]. Moreover, Gdf9^−/− male mice were observed to be normal and fertile [4], but mutations in Gdf9 have not been investigated in males for other species. One GDF9^−/− boar was maintained until sexual maturity, and breeding to a wild-type gilt confirmed that he was fertile with the development of 15 fetuses by day 36 of gestation. The genotypes of the boar and fetuses demonstrate that the edits in GDF9 cause frameshift mutations and are present in a region that would disrupt function of the protein; thus, knockout of GDF9 does not impact male fertility in pigs.

This study utilized genome editing technology to rapidly and efficiently generate GDF9-edited pigs. The timeframe was ~6 months from designing the sgRNAs to farrowing GDF9-edited pigs. Microinjection of the CRISPR/Cas9 system into porcine zygotes instead of other laborious methods, such as somatic cell nuclear transfer, allowed for the rapid production of gene edited pigs with phenotypes observed in the founder animals. It is important to note that all the fetuses and pigs possessed edits in GDF9 without carrying detectable wild-type sequences on both alleles of GDF9, indicating a highly efficient system. However, one limitation of this strategy was that several edits in GDF9 were multiples of three, which produce in-frame mutations and do not necessarily impair protein function. As a typical characteristic of most secreted proteins, the catalytic domain of GDF9 is at the C-terminal end [30]; therefore, genome edits in multiples of three may have little impact on function of GDF9, particularly those edits near the N-terminus or middle of the protein. Moreover, the genotypes between gDNA from the ear and cDNA from the ovary did not match for every gilt, which may be the result of mosaicism of edits by the CRISPR/Cas9 system. Specifically, gilt B4 demonstrated mosaicism, and the immunofluorescence analysis revealed strong and weak GDF9 signals in different follicles. Nevertheless, the secretion of GDF9 by some oocytes presumably allowed for follicular progression and ovulations.

In summary, GDF9 was efficiently edited by using the CRISPR/Cas9 system in pigs to determine the effect on ovarian development and function. Similar to mice and sheep, the reproductive tract and ovaries of the GDF9^−/− gilt were underdeveloped with no follicular progression past the primary stage. Therefore, GDF9 has a key role in the reproductive potential of gilts and sows, which has important implications for animal agriculture and creating biomedical models.

Conflict of interest

The authors have declared that no conflict of interest exists.

Author contributions

P.R.C. and K.L. conceived and designed the experiments. K.U. designed sgRNAs and performed microinjections. L.D.S. and K.U. generated porcine embryos. P.R.C., K.M., and E.D.R. performed fetal and gilt collections. P.R.C. and K.L. analyzed the data. P.R.C. wrote the manuscript. K.L and R.S.P supervised the work. All authors revised and accepted the manuscript.

Data availability

The data that support the findings of this study are available upon request from the corresponding author.