Ovarian culture with mouse serum improves follicle development compared with fetal bovine serum, showing the importance of as yet unidentified factors in follicle growth

Collection of ovaries

Mouse pups were sacrificed by cervical dislocation at postnatal Day 3 (P3) and the ovaries and associated tissues were removed. The ovaries were dissected and isolated in Dulbecco’s phosphate-buffered saline (DPBS; Sigma-Aldrich, UK) containing 1 mg/mL bovine serum albumin (BSA; Fisher Scientific, UK), by using a stereomicroscope. Each ovary used in the study was collected from a different mouse.

Ovarian tissue culture

P3 ovary culture was performed as previously described (Lo et al. 2019). Briefly, whole ovaries were cultured in 300 μL of media (Waymouth-based culture media (Merck, Germany) supplemented with 1% insulin–transferrin–selenium (ITS-G; Sigma-Aldrich, UK); 100 IU/mL penicillin and 0.1 mg/mL streptomycin (Sigma-Aldrich, UK), 2.5 IU/mL recombinant follicle-stimulating hormone (FSH; Gonal-F; Merck-Serono, Feltham, UK), 10 μg/mL BSA (Factor V heat-shock treated; Fisher Scientific, UK), and 25 μg/mL ascorbic acid (Acros Organics, Belgium) in a 24-well tissue culture treated plate (Corning CoStar, Fisher Scientific, UK) on a polycarbonate Transwell membrane (6.5 mm diameter, 0.4 mm pore size; Corning Incorporated, USA). Media was supplemented with either 5% fetal bovine serum (FBS; Biosera, USA) or 5% mouse serum (MS; Biosera). All cultures used serum from a single manufacturing lot for both FBS and MS to minimise batch-to-batch variability. The culture media were replaced every 2 days during the 14-day culture period. The plates were incubated at 37°C with 5% carbon dioxide and 95% air.

Tissue processing

Histological staining with periodic acid schiff (PAS) and haematoxylin

Every 20^th section was stained with periodic acid schiff (PAS) and haematoxylin. Sections were dewaxed in three washes of xylene for 20 min and rehydrated in graded concentrations of ethanol (100% × 2, 90%, 70%, 50%), and then placed in tap water. Sections were then incubated in periodic acid stock (Sigma-Aldrich, UK) for 5 min, rinsed in dH_²O, immersed in Schiff’s reagent (Merck, Germany) for 15 min, followed by washing in tap water for 5 mins, and then immersion in haematoxylin (Gills no2 Haematoxylin; Sigma-Aldrich, UK) for 90s. Sections were then dehydrated in 80%, 95% and 100% ethanol, cleared in xylene and then mounted using DPX mounting media (Sigma-Aldrich).

Immunohistochemistry

Immunohistochemistry was performed to detect AMH as a marker of follicular function and laminin to assess extracellular matrix (ECM) integrity as previously described (Adeniran et al. 2021). We define functional follicles as those with intact GC layers and detectable AMH, indicative of active growth and developmental potential. In contrast, atretic follicles show disrupted morphology with no detectable AMH. In brief, sections were dewaxed in xylene and rehydrated in graded concentrations of ethanol (100%, 90%, 70%, 50%) then washed in double-distilled water (ddH₂O). Antigen retrieval was performed by placing the slides in a 1:10 dilution of 100 × Vector® antigen unmasking solution (Vector Labs, UK; according to manufacturer’s instructions) in dH₂O, and microwaving for 1 min on full power and 9 min at low power. To block endogenous peroxidase activity, sections were treated with 3% hydrogen peroxide (Sigma-Aldrich, 31642) in phosphate-buffered saline (PBS; Sigma-Aldrich) for 5 min, then washed in ddH₂0 for 3 min, then PBS with 0.05% Tween 20 (Fisher Scientific, Loughborough, UK) (PBS-T) for 3 min. To prevent non-specific antibody binding, sections were blocked with 5% normal goat serum (NGS; Vector Laboratories, PK-6101) in PBS-T for 1 h at room temperature in a humidifying chamber. To detect AMH, sections were incubated with mouse monoclonal anti-human AMH (5 μg/mL, MCA2246, Bio-Rad UK) in a 1:50 dilution in 5% NGS in PBS-T. For laminin detection, sections were incubated with rabbit anti-laminin (1:30; Sigma-Aldrich UK, 030M4798). Negative control sections were incubated in 5% NGS in PBS-T, with the primary antibody being omitted. For both antibodies, sections were then incubated at 4°C overnight. Sections were then washed in PBS-T and all sections (including negative controls) were incubated with secondary antibodies for 1 h at room temperature: biotinylated goat anti-mouse immunoglobulin G (IgG) (1:100; Vector Laboratories, UK, BA-9200) for AMH and goat anti-rabbit IgG (1:100; Vector Laboratories, BA-1000) for laminin. Sections were washed and incubated in Vectastain Elite ABC kit (Vector Laboratories, PK-6101; according to manufacturer’s instructions) in a humidifying chamber at room temperature for 30 min. Sections were then washed in PBS-T followed by exposure to 3,3′-diaminobenzidine (DAB) peroxidase (Vector Laboratories, SK-4100; prepared according to manufacturer’s instructions). All slides in each experiment were exposed to DAB for the same amount of time. The sections were then dehydrated in graded concentrations of ethanol (50%, 70%, 90%, 2 × 100%) and cleared in xylene. Slides were mounted using DPX mounting media (Sigma-Aldrich). Slides were examined and imaged using a Leica DM2500 light microscope (Microscope Services Ltd., Woodstock, UK) and MicroPublisher 5.0 RTV camera (Qimaging; Microscope Service Ltd.). Images were captured using the Infinity analyse software (ver. 6.5.5; see https://www.teledynevisionsolutions.com/products/infinity-analyze/?model=infinityanalyze&vertical=tvs-lumenera&segment=tvs; Lumenera Corporation (now part of Teledyne Lumenera), Canada) to obtain pre-counterstained images. Following imaging, the AMH immunohistochemistry slides were counterstained with haematoxylin. First, the slides were transferred to xylene to loosen and remove the coverslips, followed by graded concentrations of ethanol before counterstaining with haematoxylin as described above.

Follicle analysis

Every 20^th section stained with PAS and haematoxylin was used for follicle analyses. Images were obtained using a Leica DM2500 light microscope (Microscope Services Ltd., Woodstock, UK), and a MicroPublisher 5.0 RTV camera (Qimaging; Microscope Services Ltd.). Follicles with a visible oocyte were classified as primordial, transitional, primary, secondary, and preantral according to established criteria (Pedersen and Peters 1968; Lo et al. 2019; Adeniran et al. 2021). Follicles were classed as atretic if they had a pyknotic nucleus, shrunken oocyte or ≥10% pyknotic granulosa cells, or a combination of the above factors (Winship et al. 2019; Walker et al. 2021). Ovarian follicles as well as their components were analysed, namely, oocyte, nucleus, follicle basal lamina (FBL), granulosa cells (GCs), and theca cells (TCs). All image analysis measurements were performed blind in ImageJ with the FIJI® software (ver. 2.1.0/1.53c; National Institute of Health), with 4.39 pixels/μm as a scale. The total number of ovarian follicles at each follicle stage was calculated using the Abercrombie correction factor (Abercrombie 1946), by using the nuclear diameter (assessed using ImageJ in these ovaries). GC and TC numbers were manually counted; GCs were identified as cuboidal cells found within the FBL, whereas TCs were identified by their squamous appearance and their extension from the outer wall of the FBL to the visibly different stromal cells (Chiti et al. 2017). Follicle area (μm²) was determined by measuring the area from the outer edge of the FBL or the outer wall of the TCs by using ImageJ. Oocyte diameter (μm) was determined by averaging two measurements at the widest cross-section of follicles (Griffin et al. 2006). To analyse AMH-positive follicles, the counterstained IHC images were analysed in ImageJ to assess the number of primary, secondary and preantral follicles and cells that were AMH-positive (AMH+ve; i.e. stained brown with DAB); all other cells and follicles were considered AMH-negative. The number of AMH+ve follicles is expressed as a percentage of all follicles counted in the section, this is then broken down into follicle stages. To determine the percentage of GCs that were AMH+ve, the non-counterstained IHC images were used.

Statistical analysis

GraphPad Prism (ver. 9.2.0; GraphPad Software Inc., USA) was used for statistical analysis and creating graphs. For all results, a Shapiro–Wilk test and the associated Q–Q plot was used to determine whether data were normally or abnormally distributed. Normally distributed data were compared using two-sample Student’s t-tests; this was performed for atretic counts and AMH+ve follicle counts. Abnormally distributed data were analysed using the non-parametric Mann–Whitney test; this was performed for total follicle counts, somatic cell counts, and proportions including follicles without a defined FBL and AMH+ve GCs. Linear regressions were also performed and plotted to correlate follicle area and oocyte diameter to GC number. Results are presented as mean ± standard deviation (STDEV), with P < 0.05 being considered statistically significant.

Ethical approval

All animal studies and experimental procedures using mice were conducted with approval from the Local Ethical Review Panel at the University of Oxford, in accordance with the UK Animals (Scientific Procedures) Act 1986, under a project licence held by Professor Suzannah Williams (Project Licence number: PP1638762). Male C57BL/6J mice, purchased from the Biomedical Services at the University of Oxford, UK, were mated in-house with female CD-1 mice purchased from Charles River, UK.

Follicle health and development is improved in mouse ovarian tissue cultured in MS versus FBS

Follicles were analysed in PAS and haematoxylin-stained sections (FBS, n = 3; MS, n = 3). Follicles at different stages of development were found in ovarian tissue cultured in both FBS-supplemented media and MS-supplemented media (Fig. 1a); however, there was no difference in the total number of follicles between the two groups (Fig. 1b; P = 0.700).

Fig. 1.

Analysis of follicle number, follicular somatic cells, and atretic follicles in neonatal mouse ovaries cultured in fetal bovine serum (FBS)- and mouse serum (MS)-supplemented culture media. (a) P3 mice ovaries were cultured for 14 days in Waymouth’s media supplemented with either FBS (n = 3) or MS (n = 3), and sections were stained in PAS and haematoxylin. Scale bar = 100 μm. (b) The total number of follicles in FBS- and MS-cultured ovaries was not significantly (P > 0.05) different; n.s., non-significant. (c) Representative images of each follicle type. No secondary or preantral follicles were found in the FBS group. Scale bar = 50 μm. (d) The number of follicles at each developmental stage did not differ although secondary and preantral follicles were only in the MS group. (e) The proportion of atretic follicles was lower in the MS-supplemented culture group (*P < 0.05); however, the stage of follicle development when atresia occurred did not differ (f). Data are means ± std.

To evaluate the development and growth of follicles in FBS- versus MS-supplemented media, the number of follicles at each stage of development was determined. Ovaries cultured with FBS-supplemented media were present only at the early stages of development, with no secondary or preantral follicles present, whereas ovaries cultured with MS-supplemented media contained follicles at all stages of development; primordial to preantral (Fig. 1c). Therefore, although there was a clear difference in the more developed follicles, there were no differences in earlier-stage follicles (P > 0.05) (Fig. 1d). Focusing on later follicle development, all three MS-treated ovaries contained secondary follicles, whereas none of the FBS-treated ovaries contained secondary follicles. Moreover, although preantral follicles were found only in the MS-treated group, these were restricted to one individual.

To compare follicle health of mouse ovaries cultured in FBS- and MS-supplemented media, the number of atretic follicles was determined; follicle atresia was defined as follicles exhibiting a pyknotic nucleus, a notably shrunken oocyte, or ≥10% pyknotic granulosa cells. The proportion of atretic follicles in the MS group was significantly lower than in the FBS group (P < 0.02; Fig. 1e). However, of the atretic follicles, there was no difference in the number of atretic follicles at each stage of development between the two culture media supplement groups (Fig. 1f). Because no secondary or preantral follicles were observed in the FBS group, the percentage of atresia for these stages could not be compared with the MS group.

Analysis of follicular cells and follicle size in mouse ovaries cultured in FBS- versus MS-supplemented culture media

The number of GCs and TCs were determined to assess how MS affected somatic cells (Fig. 2a). However, since ovaries cultured in FBS did not contain any secondary (FBS, n = 0; MS, n = 11) or preantral (FBS, n = 0; MS, n = 5) follicles, only primary follicles were compared (FBS, n = 39; MS, n = 58). Despite MS clearly supporting later stages of development evidenced by the presence of secondary and preantral follicles, there was no difference in the number of GCs and TCs in primary follicles, although this is unsurprising because primary follicles are categorised on the basis of the number of GCs (Fig. 2b, c).

Fig. 2.

Evaluation of follicular somatic cells and follicle development in mouse ovaries cultured in FBS- versus MS-supplemented culture media. (a) Representative images of the granulosa and theca cells in a late-stage follicle. Sections were stained in PAS and haematoxylin. The black arrow indicates GCs, and the yellow arrow indicates TCs. Scale bar = 50 μm. The number of (b) granulosa and (c) theca cells around follicles from the primary stage onwards were counted. n.s., non-significant. Primary follicles: FBS, n = 39; MS, n = 58. Secondary follicles: MS, n = 11. Preantral follicles: MS, n = 5. Data are means ± std. (d) Follicle area and (e) oocyte diameter were plotted against GC number for primary follicles and linear regressions were performed: FBS, n = 39 follicles, and MS, n = 58 follicles. No significant difference was found between the gradient of the slopes for either analysis.

Another parameter measured to compare follicle development between ovaries cultured in FBS and those cultured in MS-supplemented media was the growth rate of follicles and oocytes. As the number of GCs increases, it is expected that the GC area, and thus follicle area, will also increase (Lo et al. 2019). Therefore, the rate of area increase per GC can reflect follicle growth rate. To investigate this, follicle area and oocyte diameter were assessed relative to GC number for primary follicles in both groups. Primary follicles were chosen because they represent the initial stage of follicle development where a rapid increase in follicle area, GC count, and oocyte size occurs; secondary and preantral follicles were not assessed as none were observed in the FBS group. Linear regression lines were plotted (Fig. 2d, e) to analyse the relationships between follicle area and GC count, as well as oocyte diameter and GC count in both groups. The linear regressions showed weak but positive correlations between follicle area and GC count, as well as oocyte diameter and GC count (follicle area: MS vs FBS, r² = 0.314 and r² = 0.185; Oocyte diameter: MS vs FBS, r² = 0.02 and r² = 0.135). No significant difference was found in the slope gradient, and thus growth trajectories, between the MS and FBS groups for either follicle area or oocyte diameter.

Analysis of follicle basal lamina in mouse ovaries cultured in FBS- versus MS-supplemented culture media

During follicle analysis, some follicles were found that lacked a well-defined FBL and thus the stage of development could not be classified according to established follicle classification criteria (Fig. 3a). To investigate whether the occurrence of these was improved by the addition of MS, we analysed the proportion of these in the two conditions. Although there was no significant difference in the incidence of follicles with an undefined FBL (FBS vs MS, P = 0.100), the mean proportion in MS was half the proportion found in FBS-supplemented media (Fig. 3b).

Fig. 3.

Analysis of follicle basal lamina in mouse ovaries cultured for 14 days in FBS- and MS-supplemented culture media. (a) A representative image of a follicle-like structure with a defined FBL and an indistinguishable FBL; black solid arrows indicate a clear FBL surrounding the follicle, and broken black arrows indicate follicles without a defined FBL. O, oocyte; ZP, zona pellucida. Sections were stained in PAS and haematoxylin. (b) The percentage of follicles without a defined FBL (P = 0.1). n.s., non-significant. Data are means ± std.

Assessment of follicle function in ovaries cultured in FBS-supplemented media compared with MS-supplemented media

To assess follicle function, we detected AMH by using immunohistochemistry. Ovaries cultured in either FBS or MS both contained follicles that were clearly positive for AMH (AMH+ve; Fig. 4a). However, ovaries cultured in media supplemented with MS had more AMH+ve follicles than did those in FBS-supplemented media (P < 0.05; Fig. 4b). To determine whether this reflected the increased number of later-stage follicles in MS or a difference in AMH at a specific stage of development, the percentage of AMH+ve follicles per stage was analysed (Fig. 4c, d). Interestingly, the proportion of AMH+ve primary follicles was higher in the MS-supplemented group compared with the FBS-supplemented group (P < 0.05). Since no secondary or preantral follicles were detected in ovaries cultured with FBS-supplemented media, no comparison could be made at the later stages of development. All GCs within AMH-positive (AMH+ve) follicles were not equally stained (Fig. 4e). With the aim of distinguishing between them, the percentage of GCs that were AMH+ve was calculated for each AMH+ve follicle to determine whether ovaries cultured in MS-supplemented media had a higher proportion of AMH+ve GCs than those in FBS-supplemented media (Fig. 4f). This analysis showed that a higher percentage of GCs was AMH+ve in the primary follicles of ovaries cultured in MS-supplemented media compared to those in FBS-supplemented media (P < 0.05).

Fig. 4.

Analysis of anti-Müllerian hormone in mouse ovaries cultured in fetal bovine serum (FBS)- or mouse serum (MS)-supplemented media for 14 days. (a) Representative images of ovarian sections using immunohistochemistry to detect anti-Müllerian hormone (AMH). Images are from the same experiment with the same DAB exposure to allow for direct comparison. Scale bar = 100 μm. (b) The percentage of AMH+ve follicles in the MS group was higher than in the FBS group (P < 0.05). (c) Representative images of AMH+ve primary, secondary and preantral follicles detected using immunohistochemistry with a DAB chromogen. Scale bar = 50 μm. (d) The percentage of AMH+ve follicles per stage was analysed and the proportion of primary AMH+ve follicles was higher in the MS group at the primary stage than in the FBS group (P < 0.05). There were no secondary and preantral follicles in the FBS group and thus no comparisons could be made. (e) Representative sections of AMH+ve follicles with varying proportions of AMH+ve GCs. The red arrows point to an example of AMH+ve GCs, whereas the black arrows point to AMH-negative GCs. Scale bar = 50 μm. (f) The percentage of AMH+ve per follicle at each stage of development. (Primary follicles: FBS, n = 15; MS, n = 26; secondary follicles: FBS, n = 0; MS, n = 12; preantral follicles: FBS, n = 0; MS, n = 3). There was a higher proportion of AMH+ve in primary follicles in the MS group than in the FBS group (*P < 0.05). There were no secondary and preantral follicles in the FBS group and thus no comparison could be made. Data are means ± std.

ECM evaluation via laminin in MS- and FBS-cultured mouse ovarian tissue

With the aim of identifying the FBL, TCs and exploring the composition of the ECM, laminin was detected in ovarian sections by immunohistochemistry (Fig. 5). In ovarian tissue cultured within both FBS- and MS-supplemented media, the majority of the section was positive for laminin, with the exception of the GCs within follicles. Visible laminin staining suggests the presence of the FBL around follicles in both conditions, appearing more prominent in later-stage follicles that were cultured in MS. Thus, both MS- and FBS-supplemented culture media supported the expression of laminin in ovarian tissue, indicating their potential to maintain essential ECM integrity for follicle development.

Fig. 5.

Detection of laminin using immunohistochemistry in mouse ovarian tissue cultured in FBS- versus MS-supplemented culture media. P3 mice ovaries were cultured for 14 days in Waymouth’s media supplemented with either FBS or MS. Control sections were not exposed to the primary antibody. Images are representative of the same experiment with the same DAB exposure to allow for direct comparison. Scale bar = 100 μm.