We studied adult Scottish Greyface ewes (Ovis aries) weighing 60-75 kg in their third to fifth breeding season. Ewes were housed together in spacious pens with ad libitum hay supplemented with Excel ewe nuts (0.5-1.0kg/day; Carrs Billington, Lancashire, UK) and Crystalayx extra high energy lick (Caltech Solway Mills, Cumbria, UK). All experiments were conducted under Project Licence (PPL60/4401; PCD686E93) from the UK Home Office and underwent institutional ethics review. This work was conducted in accordance with Animals (Scientific Procedures) Act 1986, Galvani Policy on the Care, Welfare and Treatment of Animals Policy 040 and approved by the Galvani Bioelectronics Animal Scientific Review Committee and the GSK Policy on the Care, Welfare and Treatment of Animals. Ovarian sections from an earlier study collected 11 months after whole ovarian cryopreservation and transplantation using Scottish Greyface ewes as described in detail previously (19) were available for analysis.

Ewes were killed using a schedule 1 method and ovaries and ovarian vascular pedicles were collected. The vascular pedicles were fixed in Bouin’s solution for 24 hours and transferred to 70% ethanol for subsequent paraffin wax embedding. The ovaries were either: 1) fixed in Bouin’s solution for 24 hours and embedded in paraffin wax for subsequent immunohistochemistry, 2) halved longitudinally and one half fixed in Bouin’s solution and embedded into paraffin wax and the other half snap frozen and stored at -80 °C for subsequent RNA extraction and measurement of intraovarian noradrenaline or 3) fixed in 4% paraformaldehyde for optical projection tomography, depending on the experiment.

Mid-ovarian tissue sections cut to 5μm were mounted on permafrost slides. Sections were dewaxed, rehydrated as described previously (20). Antigen retrieval was carried out by pressure cooking for 5 min in 0.01 M citrate buffer, pH 6.0. Sections were washed in water before peroxidase quenching and blocking steps were performed via incubation with 3% H₂O₂ for 10 minutes, blocking with avidin and biotin (Vector Laboratories Ltd., Peterborough, UK) and then serum blocking with 20% normal goat serum/5% bovine serum albumin (BSA) in Tris Buffered saline (TBS, 0.05 M Tris pH 7.4, 0.85% NaCl). Slides were washed in TBS between treatments, then in TBS containing 0.025% Triton X-100 (TBS-T) prior to serum block and antibody incubation.

The primary antibody diluted in serum block (mouse anti-tyrosine hydroxylase 1:1000 (Sigma-Aldrich Ltd, Dorset, UK), mouse monoclonal anti-Ki67 1:100 (Novocastra, Newcastle, UK) (1:1000) or rabbit polyclonal anti-caspase 3 1:100 (Cell signalling, MA, USA)) was applied to sections and incubated overnight at 4°C. Slides were washed in TBS-T and the secondary antibody (biotinylated goat anti-mouse or goat anti-rabbit (Vector Laboratories, Peterborough, UK) diluted 1:500 in serum block) was applied to slides for 1 hour. Slides were washed in TBS-T followed by Vectastain ABC Elite tertiary complex (PK-1600 series; Vector Laboratories) for 1 hour after which 3,3′-diaminobenzidine (Dako, Cambridge, UK) was applied for 3 minutes to visualise binding. Sections were then counterstained with haematoxylin and mounted using Pertex mounting medium (Cellpath, Newtown, Powys, UK). Negative controls were non-specific mouse or rabbit serum of equivalent immunoglobulin concentrations in place of the primary antibody.

Dual labelled tissue sections were prepared for confocal microscopy following the immunohistochemistry protocol described above, with the following adjustments. The peroxidase wash step was omitted and the slides were permeabilised normally through a series of two five-minute washes in TBS-T. After an incubation time of one hour in 10% normal goat serum, the endothelial antibody (Rabbit monoclonal anti-CD31, Vector Laboratories) and mouse anti-tyrosine hydroxylase antibody were diluted together at a concentration of 1:100 and 1:200 respectively in TBS, before being added to the slides and incubated overnight at 4°C.

After washing in TBS with 0.01% Tween 20 (Sigma-Aldrich, UK) the slides were incubated with biotinylated goat-anti-rabbit secondary antibody diluted in 10% NGS at a concentration of 1:500 for one hour at room temperature. After washing in TBS-Tween 20 (0.01%) the slides were incubated with Dylight^® 594 (Thermo-Fisher, UK) and goat-anti-mouse IgG secondary antibody conjugated to Alexaflour^® 488 (Invitrogen, UK), each at 1:100 in 10% NGS. Slides were incubated in the dark at room temperature for two hours in a humidity chamber. After washing with TBS-Tween 20 in the dark slides were mounted in an aqueous solution containing 4’6-diamidino-2-phenyllindole (DAPI), and stored for 12 hours at 4°C, prior to visulisation on a Zeiss LSM 880 AxioObserver Z1 confocal fluorescent microscope (wavelengths: 405 nm, 488 nm, 594 nm, laser power set at 2%).

Two examiners, blinded to treatment, graded the immunohistological staining (based on area of staining) of tyrosine hydroxylase independently and the scores were averaged for whole ovary sections. Each ovary section was examined and graded out of four, with zero indicating no staining present and four indicating abundant staining throughout the tissue. Spatiotemporal examinations of vessel and nerve relationships were made using a Zeiss LSM880 Confocal Microscope. Images were captured to illustrate this at 20x magnification, and 63x magnification with oil.

Two independent examiners, blinded to treatment, also independently counted the number of follicles from a standardised mid-section of the ovary (21) as well as the number of preantral follicles. Follicles were classified as preantral if they did not show any antral cavity and antral if they showed a clear fluid-filled antrum (>500 µm). Immunohistochemical staining of whole ovary sections stained for proliferation (Ki67) and atresia (activated caspase 3) were blindly examined by two independent expert examiners. Each antral follicle was examined and staining was divided into two classifications, positive (clearly positive immunostaining present in multiple cells) and negative (scant/absent immunopositive cells). Number of follicles per classification was used for proportional analysis as described previously (21).

A mini-laparotomy was performed with sterile technique under general anaesthesia, induced using isoflourane (Isoflo, Abbott Animal Health, Maidenhead, UK). A small paramedian incision exposed the ovaries. To avoid non-specific ovarian damage we specifically targeted the nerves in the ovarian neurovascular pedicle. Needlepoint diathermy using monopolar coagulation current (Surgitran, STW-100) (22) was used to coagulate around the ovarian vessels, in the regions where sympathetic nerves had been identified, leaving blood vessel integrity intact.

A 3 mm³ sample from the ovarian cortex at the lateral edge of the ovary was used to measure intraovarian noradrenaline (NA) concentrations. It was weighed and homogenised in lysis buffer (0.01N HCl, 1 mM EDTA, 4mM Na₂S₂O₅). Lysate was spun at 5000 rpm at 4 °C for 10 min and the supernatant analysed. Quantification was carried out using the competitive NA ELISA kit (IMMUSMOL, Pessac, France) as described previously (23) following the manufacturer’s instructions. NA was extracted using a cis-diol-specific affinity gel, acylated and then derivatised enzymatically. The antibody bound to the solid phase was detected using an anti-rabbit IgG-peroxidase conjugate and tetramethylbenzidine (TMB) as a substrate. The reaction is monitored at 450 nm. The sensitivity was 2 pg/ml, and the intra and interassay CVs were <9%. The cross reactivity found was 0.14% for adrenaline and 1.8% for dopamine.

Plasma estradiol and progesterone concentrations were measured using a commercial ELISA following the manufacturer’s instructions on a Cobas E411 immunoanalyser (Roche, Mannheim, Germany). The progesterone assay (Cobas progesterone II) has a sensitivity of 0.48 nmol/l. The cross reactivity with related steroids is <1%. The estradiol assay (Cobas Estradiol III) has a sensitivity of 11 pmol/l. Apart from 6α-OH estradiol the cross reactivity of related steroids is <1%. Both assays have CVs <10%.

RNA was extracted from tissue using RNeasy mini spin columns following manufacturer’s protocol and concentration measured using NanoDrop 1000 Spectrophotometer as described previously (20). Complimentary DNA (cDNA) was synthesised from 200 ng RNA in accordance with manufacturer’s protocol (Applied Biosystems, California, USA). Subsequently, qRT-PCR was performed using SYBR Green. Real-time PCR reactions were carried out in duplicate 10 µl reactions, negative controls consisted of cDNA reaction without reverse transcriptase and a reaction replacing cDNA with nuclease-free water. Melt curve analysis revealed a single amplicon in all cases. GAPDH has been reported as a suitable internal control for ovarian stromal gene expression (24) and target gene expression was analysed relative to GAPDH and quantified using the DCt method.

Primer3 Input version 0.4, online software, was used to design forward and reverse primers from DNA sequences obtained from Ensembl Genome Browser. Sequences were checked for specificity using Basic Local Alignment Search Tool and validity confirmed as previously described (25). The primers 5’-3’ were: CCN2: TGCCCTCGCAGCTTACC and CTTGGAACAGGCACTCCACT; VEGF: TCTTCAAGCCATCCTGTGTG and TGCATTCACATTTGTTGTGC; NGF: CTGGCCACACTAAGGTGCATA and GCTGCCTGTATGCCGATCAA; IGF1: CATCCTCCTCGCATCTCTTC and CTCCAGCCTCCTCAGATCAC; FGF2: ACTTTAAGGACCCCAAGCGG and AGTTTGATGTGAGGGTCGCT; GAPDH: GGCGTGAACCACGAGAAGTATAA and AAGCAGGGATGATGTTCTGG.

Intravaginal progestogen-impregnated sponges (60 mg medroxyprogesterone acetate per sponge; Intervet Laboratories Ltd, Cambridge, UK) were inserted into ewes (n=6) and then 3 days later gonadotrophin releasing hormone implants (GnRH; Suprelorin; 4.7 mg Deslorelin acetate; Virbac, UK) were inserted, through large bore needles, for pituitary suppression. Eleven days later the sponge was removed and the sheep were given an injection of prostaglandin F2α (PG; 100 mg Cloprostenol; Estrumate; Coopers Animal Health Ltd, Crewe, Cheshire, UK) to ensure luteolysis of any residual corpora lutea and prepare for an artificial follicular phase. Three days later the jugular vein of the sheep was canulated and infusions started. FSH (Folltropin; Vetoquinol UK Ltd, Buckinghamshire, UK) given at 1mg/hour via jugular catheter using Graseby MS 16A syringe drivers; Luteinizing Hormone (LH; ovine LH NIADDK-oLH-27; Dr. A.F. Parlow, Harbor-UCLA-Medical Center, Torrance, CA) given as 4 hourly pulses via the jugular catheter (18 µg/pulse) using Zyklomat pulse infusion pumps, for 6 days. Sheep were given either physiological FSH concentrations only with baseline endogenous LH (n=3) or physiological FSH + additional exogenous LH infusions (n=3) (Figure 1A). At end of infusions the left ovary from each animal was processed for optical projection tomography (OPT) scanning as described below.

Progesterone sponges were inserted in ewes (n=6) and 7 days later GnRH implants were inserted as described above. Five days later mini-laparotomy was performed as described above followed by unilateral diathermy needle denervation to the left ovary was performed, allowing the right ovary to serve as an internal control. Six days later sponges were removed and PG injections were given as described above. Sheep were cannulated 3 days later and infusions started as described above with FSH (1 mg/hour) (n=3) or FSH+LH (FSH 1 mg/hour; LH 4 hourly pulses of 18 µg/pulse) (n=3) for 6 days (Figure 1B). At the end of infusions both ovaries were processed for OPT scanning as described above.

Ovaries processed for OPT were fixed in 4% paraformaldehyde overnight, washed 4 x 30 minutes Phosphate Buffered Saline (PBS), then 30 minutes each in 30%, 70%, 90%, 100% Ethanol. They were transferred to Methanol for 2 hours, then into fresh Methanol and stored at 4°C until processed for scanning. Ovaries were attached to mounting blocks, and then immersed in BABB (2 parts Benzyl Benzoate, 1 part Benzyl Alcohol) until cleared sufficiently for scanning. Cleared ovaries were scanned in a calibrated Bioptonics 3001 tomograph (Bioptonics, UK). Dataviewer (Version 1.5.2.4 Release, July 2015) was used to combine the scans into 2D and 3D models to then quantify the size and number of each follicle throughout each ovary. After scanning ovaries were returned to methanol to remove BABB, then washed 30 min each in 100%, 90% and 70% Ethanol. They were stored in 70% Ethanol prior to embedding in paraffin wax for sectioning.

Statistical analysis was conducted using GraphPad Prism 8.0 (GraphPad Software, San Diego, CA) with P<0.05 considered statistically significant. Proportional contingency table analysis was measured by Fisher’s exact test for 2x2 tables and Chi squared for larger tables. After checking for normality and similar variances two column comparisons were examined using unpaired two-tailed t-tests if the data was parametric and Mann Whitney U tests if not parametric or normally distributed. Correlation was assessed using Spearman co-efficient of correlation.

The ovaries and ovarian pedicle of Scottish Greyface sheep (n=6) were examined macroscopically (Figure 2A). The vascular bundle, consisting of the ovarian artery closely intertwined with the utero-ovarian venous plexus enters the hilum of the ovary through the ovarian pedicle (Figure 2A). Microscopic examination of the vascular pedicle showed sympathetic nerves running with the ovarian vessels (Figure 2B). Sometimes there was one discrete nerve in the ovarian pedicle (Figure 2C) but commonly there were several smaller nerves (Figures 2D-F) from 50 µm to 350 µm in diameter. There were no other neurovascular entry points to the ovary outside the hilum and ovarian pedicle. In each case at least one nerve ≥50 µm could be identified consistently with the ovarian artery. Sympathetic innervation of the ovine ovary runs along the vascular bundle in the ovarian pedicle.

The sympathetic nerves enter the ovary at the hilum as discrete nerves next to blood vessels (Figure 3A) and can be seen as discrete nerve fibre bundles within the ovarian medulla (Figure 3B). Sympathetic nerves branch further within the ovary and are associated with arterioles and small blood vessels (Figures 3C-E). Nerve fibres are also seen in the cortical regions of the ovary, in the vicinity of primordial and primary follicles, that are not associated with blood vessels (Figure 3F). Overall, 40% of nerves identified within the ovary were not associated with blood vessels. There is sympathetic innervation that is independent from blood vessels around primordial, primary and secondary follicles in the ovarian cortex.

If sympathetic nerves have a physiological role in ovarian function it would be expected that they would regenerate after ovarian denervation. We examined sympathetic nerves in the ovine ovary after oophorectomy, whole ovary cryopreservation and ovarian transplantation (18). The ovaries were disconnected from the neurovascular bundle and thus denervated. After transplantation back onto the ovarian pedicle the ovaries became functional (18). Ten months after transplant histological analysis of ovaries (n=4) showed that all ovaries had discrete nerves (50 µm) at the hilum (Figure 4A) and a normal distribution of nerves throughout the ovarian stroma, including association with arterioles (Figure 4B) and cortical tissue independent of blood vessels (Figure 4C). After denervation the ovine ovary is reinnervated in situ.

In order to determine if we could acutely denervate the ovine ovary, ewes (n=3) underwent laparotomy and unilateral denervation in the mid-follicular phase using monopolar needle micro-diathermy of the putative ovarian nerves within the neurovascular bundle leaving the vasculature intact. After 21 days the sheep were killed and the ovaries examined. There was a loss of sympathetic nerve fibre immunostaining within the treated ovary compared to the contralateral control ovary (Figures 5A-C). In addition, there was reduction in intraovarian noradrenaline concentrations (Figure 5D) that correlated with tissue immunostaining score (r=0.8407; P<0.05; Figure 5E). Needle diathermy of the sympathetic nerves in the ovarian pedicle can be used to acutely denervate the ovary.

A separate cohort of ewes were randomised to either bilateral denervation using micro-diathermy (n=4) or a sham procedure without diathermy (n=4). The hormonal profiles of the ewes were then examined daily over an ovarian cycle and ovaries were collected 21 days later and at that stage we examined intraovarian noradrenaline concentrations to confirm ongoing denervation during the experiment. In the follicular phase there was no difference in estradiol concentrations (Figure 6A). There was no effect of denervation in the timing of ovulation, post-ovulatory progesterone concentrations and luteolysis (Figures 6B, C). Analysis of the tissue immunostaining score for sympathetic nerves (P<0.01; Figure 6D) and intraovarian noradrenaline concentrations (P<0.005; Figure 6E) confirmed ovarian denervation after micro-diathermy.

Follicles were counted in a representative mid-ovarian section from each ovary. There were no differences in antral follicle numbers (Figure 6F) although there was a strong trend to less preantral follicles after diathermy but this didn’t reach statistical significance (P=0.056; Figure 6G). There were no differences in proportion of atretic antral follicles, assessed by cleaved caspase 3 expression (Figures 6H, I) or growing antral follicles, assessed by Ki67 localisation (Figures 6J, K). In addition, there were no differences in the ovarian transcript abundance of CCN2, VEGFA, NGF, IGF1 and FGF2 (Figure 6L). Acute denervation does not have any impact on follicular growth, ovulation and luteolysis or on the survival of antral follicles.

To develop an acute ovine model of PCOS, in normal cycling sheep, we first synchronised the sheep with progesterone sponges then switched off the hypothalamic pituitary ovarian axis using a GnRH agonist (n=3). Sheep were given an infusion of FSH with baseline endogenous LH (Control) or with additional high dose pulsatile LH (PCOS-like). After seven days the ovaries were collected and analysed by optical tomography, which allows the whole ovary to be viewed in real-time digital sections to accurately count and measure all the antral follicles (Figures 7A, C). These follicles could be identified in subsequent tissue sections, and had a normal follicular structure including healthy granulosa cell and theca cell layers (Figures 7A, B), after further fixation and sectioning after OPT was complete. There was a different pattern of follicles in the PCO-like ovaries (P<0.0001) with an increased number of smaller and a reduced number of larger follicles, suggesting a polycystic morphology (Figures 7C, D, F). Gonadotrophin manipulation can facilitate the acute development of polycystic ovaries.

After bilateral denervation we then assessed the effects of gonadotrophin infusion to create the acute PCOS-like model. There was a difference in the pattern of follicles in the control high dose pulsatile LH PCOS-like sheep (Figure 8A) compared to the control low LH sheep (Figure 8B). However, denervation showed no difference in the pattern or number of follicles in the high LH PCOS-like sheep (Figure 8C) or the control low LH sheep (Figure 8D). Denervation had no effect on gonadotrophin action in the development of a PCO ovary.