Enzymatic isolation of porcine preantral follicles impairs oocyte viability and long-term in vitro growth

Abstract

The enzymatic isolation of preantral follicles (PAFs) is considered the most efficient method for retrieving a large number of intact follicles, offering significant advantages in terms of yield and processing time. However, the low success rate of enzymatically isolated follicles in long-term culture raises concerns regarding their impact on oocyte quality and developmental potential. This study addresses a critical gap in understanding how enzymatic retrieval of PAFs affects the oocyte-granulosa cell connection and its relationship with high mortality and culture failure observed during in vitro growth (IVG). By systematically comparing crude collagenases (IA and IV) and purified collagenases (Liberase TM and DH) with a mechanical isolation protocol, we identified the optimal enzyme concentrations that maximize follicle yield while minimizing cellular damage. Our results reveal that the enzymatic retrieval of PAFs corresponds to the loss of transzonal projections (TZPs) post-isolation, as well as premature oocyte extrusion and follicle deformities during IVG. Our findings also highlight the differential apoptotic responses in oocytes and granulosa cells. Although these enzymes sustain follicle cell integrity, they compromise oocyte viability during isolation. Notably, crude collagenases impair oocyte growth during prolonged culture, whereas purified collagenases preserve the developmental potential of oocytes. This study also provides the first evidence that enzymatic isolation of PAFs adversely affects TZPs. Overall, our study highlights the importance of selecting an appropriate method, enzyme type, and concentration for preserving the integrity of oocytes, follicles, and their connections, thereby supporting successful in vitro culture. Additionally, our results suggest that mechanical protocols and high-purity enzymes are preferred for maintaining oocyte competence.

Ovarian tissue transplantation has emerged as a promising method for preserving female fertility and protecting endangered species. However, this technique is not preferred for patients with acute lymphoblastic or myeloid leukemia because of the risk of malignant cell re-implantation [1]. Therefore, preantral follicles (PAFs) are considered a safer alternative for preserving fertility [2]. Moreover, techniques for PAF isolation and culture have allowed scientists to study the detailed mechanisms of follicle development, hormonal regulation, oocyte maturation, and artificial ovary development [3, 4]. This can provide an in-depth understanding of reproductive biology and potential treatments for infertility.

Developing effective protocols for retrieving large quantities of highly viable PAFs from fresh or cryopreserved ovarian tissues is crucial for both clinical applications and fundamental research. The main techniques used to isolate PAFs involve mechanical and enzymatic methods. During mechanical isolation, the follicles are microdissected from the surrounding stroma and extracellular matrix (ECM). Notably, homogenizers have been used to isolate numerous follicles in three ruminant species [5], along with hand blenders [6], or some methods combining with tissue choppers and homogenization followed by filtration [7]. However, these techniques are not suitable for human follicles. Microdissection using needles has also been used to isolate follicles from the ECM of the ovarian cortex tissues of various species, including mouse [8], human [9], dog [10], cows [11], cat [12], and pigs [13]. Moreover, this method is time-consuming, labor-intensive, and requires careful handling during follicle collection, despite its low yield.

Enzymatic digestion involves the proteolytic digestion of ovarian stroma, increasing the recovery rate and follicle yield, while being less labor-intensive than mechanical isolation [14]. Several studies have employed different types of collagenases to isolate partial or complete human PAFs [15,16,17,18,19]. Collagenase type IA and IV have been shown to achieve high oocyte viability in porcine models [20,21,22]. However, the culture outcomes remain extremely low owing to their detrimental effects on theca cells, degradation of the basement membrane [14, 23], and excess lipid accumulation in granulosa cells (GCs) [17]. Although collagenase contains a large amount of endotoxin and exhibits batch-to-batch variability [24], purified collagenase has fewer bacterial byproducts and endotoxins. Liberase, a mixture of purified collagenases, is widely used to isolate PAFs. Liberase TM and Liberase DH have been shown to protect the ultrastructure and morphology of follicles better compared to those with collagenase [6, 25,26,27,28,29]. However, these effects cannot be fully assessed without evaluating long-term cultures. Recent studies comparing mechanical isolation followed by either Liberase TM or Liberase DH treatment during long-term culture have demonstrated persistently low efficiency [6, 12]. However, the reasons underlying this diminished efficiency in the long-term culture of PAFs remain insufficiently understood.

Recent studies on enzymatically isolated follicles have speculated that failure during in vitro culture is attributed to the loss of the theca/interstitial cell layer [30] and hormonal function [3, 12, 31]. Additionally, direct pressure from the alginate gel on the ‘exposed’ granulosa cell layers may contribute to this issue. However, even when enzymatically isolated human PAFs are cultured on the ECM layer, high mortality is observed beyond the first day [17], and high rates of premature oocyte extrusion are observed on Matrigel II [16]. Recent studies have also indicated that GCs can differentiate into androgen-producing cells in the absence of theca-interstitial cells [8]. Therefore, the failure of PAF cultures to utilize enzymatic retrieval methods can be attributed to factors other than theca cell deficiency.

Enzymatic isolation of PAFs involves the degradation of collagen, which interacts with cell surface receptors known as integrins. These integrins are crucial for cell adhesion to the ECM and for mediating cell-cell adhesion [32, 33]. Transzonal projections (TZPs) are bidirectional, contact-dependent communication points between oocytes and adjacent GCs. Maintenance of the follicular structure with minimal damage to GCs and oocytes is crucial for any isolation protocol [26], and disruption of TZPs has been associated with impaired oocyte development, indicating their critical role in maintaining oocyte integrity and functionality through intercellular interaction [34,35,36,37]. However, the correlation between the enzymatic isolation of PAFs and their impact on TZPs remains to be elucidated.

This is the first study to utilize a porcine in vitro culture model to systematically compare mechanical isolation with prevalent enzymatic isolation methods, including collagenase type IA, collagenase type IV, Liberase TM, and Liberase DH. This study aims to assess TZPs and evaluate the detrimental effects of TZP loss on oocyte development. Additionally, it investigates the potential effect of enzymatic treatment on the growth of oocytes and follicles under extended culture conditions. The present study also evaluates the impact of different isolation protocols on the apoptotic status of oocytes and follicular cells separately. These findings have significant implications for human in vitro follicle development based on the biological similarity of the porcine model, in which extensive enzymatic digestion is necessary to extract follicles from the dense ovarian cortex.

Materials and Methods

Chemicals and reagents

All reagents and chemicals used in this study were purchased from Sigma Chemical Co. (St. Louis, MO, USA), unless otherwise stated.

Collection of preantral follicles for in vitro growth (IVG)

Mechanical isolation of PAFs has been described previously [13]. Briefly, oocyte-granulosa complexes (OGCs) from the final stage of PAFs (0.3–0.4 mm) were torn off using an 18-gauge needle under a stereomicroscope and stored in MEM-HEPES (M2645), 0.1% (w/v) polyvinyl alcohol (PVA) supplemented with 10% (v/v) fetal bovine serum (FBS-F2424).

For enzymatic isolation, thin ovarian cortical fragments containing PAFs were selected and washed several times with Hanks′ Balanced Salt solution (HBSS-H6648). Then, they were randomly divided into groups of different enzyme types and concentrations including collagenase type IV-C5138 (0.05, 0.1, 0.5, and 1 mg/ml), collagenase type IA-C9891 (0.05, 0.1, 0.5, and 1 mg/ml), Liberase TM–5401127001 (0.035, 0.07, 0.7, and 1.4 Wünsch Units/ml), and Liberase DH–5401054001 (0.035, 0.07, 0.7, and 1.4 Wünsch Units/ml) and incubated for 1 h at 37^oC in a humidified atmosphere. After incubation, enzyme activity was inhibited by adding FBS to a final concentration of 20% (v/v) and transferred to MEM-HEPES-0.1% PVA supplemented with 5% (v/v) FBS to gently pick up the PAFs using an 18-gauge needle.

Enzymatic treatments in denuded oocytes

After mechanical isolation from 0.3–0.4 mm PAFs, oocytes were denuded of GCs by repeated pipetting with a fine-bore mouth-controlled micropipette. The cells were then recovered in drops of α-MEM supplemented with 10% (v/v) FBS for 1 h. The denuded oocytes were randomly divided into drops of HBSS containing different enzyme types and concentrations, followed by incubation for 1 h at 37°C in a humidified atmosphere. Oocyte morphology was observed at 0, 15, 30, and 60 min.

In vitro growth of preantral follicles

The culture medium for OGCs was based on our previous study [38] with minor modifications, consisting of α‐MEM (12571063, Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10 mM taurine, 2% (w/v) polyvinylpyrrolidone‐360.000, 3 mM hypoxanthine, 1% (w/v) insulin-transferrin-selenium, 1 µg/ml 17β-estradiol, 5% (v/v) heat-inactivated FBS, 7.5% (v/v) porcine follicular fluid, 50 µg/ml L-ascorbic acids, and 0.08 mg/ml kanamycin sulfate.

In this study, an agarose-3D culture system was used to facilitate unrestricted follicle growth while minimizing external influences on morphology. This approach capitalizes on the advantage of easy access to the culture medium, facilitating meticulous assessment of follicle and oocyte morphologies [13, 39]. Briefly, 20–40 follicles were distributed on agarose-coated 24-well plates containing 1 ml culture medium, and overlaid with 200 µl of mineral oil (Fertipro, Beernem, Belgium). The culture was maintained at 38.5°C with 5% CO₂ in a humidified atmosphere. The culture medium was replaced every 2 days, and dead follicles or oocytes were removed during medium changes.

F-actin, Nile-red, and Immunofluorescence staining

Denuded oocytes or empty zona pellucida (ZP) were prepared for TZP visualization by F-actin staining as previously described [40,41,42,43,44], with some modifications. Briefly, the ZP of intact oocytes or empty ZPs were fixed in 4% (w/v) paraformaldehyde in phosphate-buffered saline (PBS) for 5 min and blocked with 1% (w/v) bovine serum albumin in PBS (PBS-BSA). Fixed oocytes or empty ZPs were then incubated with 200 nmol TRITC-Phalloidin (PF7551; ECM Biosciences, Versailles, KY, USA) for 90 min and examined under an optical sectioning microscope (ZEISS-Apotome-3; Carl Zeiss Microscopy GmbH, Jena, Germany). Nile-red staining was performed by permeabilizing oocytes with 0.1% Triton X-100 for 30 min, followed by incubation with 10 μg/ml Nile Red (72485) for 40 min. Nuclear membrane staining was performed by incubating permeabilized oocytes with the primary goat polyclonal mouse anti-lamin B antibody (1:100, c-6217, Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 2 h at RT. After washing three times (15 min each) in PBS-BSA, the oocytes were incubated with Alexa-Fluor-568-labeled goat anti-mouse secondary antibodies (1:200, A11004, Molecular Probes, Eugene, OR, USA). DNA was counterstained with 2 μg/ml 4,6-diamidino-2-phenylindole (DAPI; Molecular Probes). Finally, after washing three times with PBS-BSA, the oocytes and ZP were mounted on glass slides and examined under an optical sectioning microscope (ZEISS-Apotome-3). The number of TZPs were counted in 0.73 µm scanning sections at 20 × magnitude. Quantitative intensity analysis was performed as previously described [45]. Briefly, signals were quantified using Zeiss-ZEN software; TZP intensity was measured based on the total F-actin signal in the ZP, whereas lipid accumulation was assessed using the total Nile Red signal in the oocyte. The total measured signal was corrected by subtracting the background signal from an equal area.

Differential staining

Follicular cell viability was assessed in 1-day and 18-day cultured follicles as described previously [46]. Briefly, follicles were stained with 50 µg/ml propidium iodide (PI) and Hoechst-33342 for 10 min each, washed, mounted in glycerol, and examined using a Nikon-Intensilight-C-HGFI fluorescence microscope (Nikon, Melville, NY, USA).

mRNA extraction and qPCR analysis

For each biological replicate, 25–30 oocytes cultured for 1 d or 7–10 oocytes cultured for 18 days were subjected to mRNA extraction using the Dynabeads™-mRNA-DIRECT™ Kit (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). For follicular cell RNA extraction, total RNA was isolated from 20 –25 1-day cultured or 5–10 18-day cultured follicles per replicate using PureLink-RNA-Micro-Kit (Invitrogen) according to the manufacturer’s protocol.

RT-qPCR was performed using Power SYBR™ Green RNA-to-CT™ 1-Step-Kit (Applied Biosystems, Foster City, CA, USA). RT-qPCR was performed in a thermocycler (QuantStudio™ 5 Real-Time-PCR System, Applied Biosystems). The running methods and primer sequences are listed in Supplementary Table 1. The reaction was tracked using the QuantStudio™Design&Analysis software v1.5.2 on a connected computer. Each sample was normalized to the expression of the endogenous gene Actin-β for subsequent analysis.

Statistical analysis

Descriptive statistics were performed to determine the normality and homogeneity of variances before parametric analyses, with skewness values between –1 and +1. One-way analysis of variance (ANOVA) was conducted using the Statistical Package for Social Statistics version 20, with a significance level of P-value < 0.05. Data are presented as the mean ± standard error of the mean (SEM). The Log2FC method was used to analyze the PCR data, and the abundance of mRNA was expressed as fold change relative to the mean value of the mechanical groups.

Results

Efficiency of PAF retrieval using different enzymes and concentrations

To compare the potential effects of enzyme concentration and type on the retrieval of PAFs based on follicle integrity and oocyte quality, we classified the retrieved follicles into four grades (1–4), representing high to low integrity. Grade 1 included intact follicles or follicles with thin residual theca cell layers (Fig. 1, Aa-a1). Grade 2 consisted of partially chipped follicles or oocytes completely covered with multiple layers of GCs (Fig. 1, Ab-b1). Grade 3 oocytes were not fully covered by granulosa cell layers (Fig. 1, Ac-c1). Grade 4 oocytes were nearly or completely denuded of granulosa cell layers (Fig. 1, Ad-d1). We were able to culture only grade 1 and grade 2 follicles because of the high oocyte extrusion in grade 3 and grade 4 follicles during IVG, making the selection of grade 1 and grade 2 follicles important for further culture. The quality of collected oocytes was assessed based on their morphology, with normal oocytes maintaining a round shape (Fig. 1, Ae-e1). In contrast, dysmorphic oocytes were characterized by an unrounded zona pellucida (Fig. 1, Af-f1), enlarged cytoplasm with dispersed lipid droplets (Fig. 1, Ag-g1), and dense cytoplasm (Fig. 1, Ah-h1). Representative images of grade 1 and grade 2 follicles retrieved using different methods were quantified for IVG, as shown in Fig. 1B.

Fig. 1.

Efficiency of preantral follicle retrieval using various enzymatic and mechanical methods. (A) Classification of follicular grades - Grade 1: Intact follicles or follicles with thin residual theca cell layers (a-a1); Grade 2: Partially chipped follicles or oocytes completely covered by multiple layers of granulosa cells (b-b1); Grade 3: Oocytes partially covered by multiple granulosa cell layers (c-c1); Grade 4: Oocytes nearly or completely denuded of granulosa cell layers (d-d1). Normal oocytes retain a round shape, with lipid droplets showing relatively even distribution (e-e1). In contrast, oocyte abnormalities are characterized by a dysmorphic zona pellucida (f-f1), enlarged cytoplasm with dispersed lipid droplets (g-g1), and dense-looking cytoplasm (h-h1). Scale bar = 200 µm. (B) Images of collected preantral follicles retrieved using different enzyme types and concentrations. Scale bar = 500 µm. (C) Percentages of oocytes maintaining a round shape within the collected preantral follicles. (D) Matrix heatmap illustrating the classification of follicles after isolation using collagenase IA and IV at concentrations of 1 mg, 0.5 mg, 0.1 mg, and 0.05 mg (IA: n = 1025, 775, 625, and 439; IV: n = 678, 692, 656, and 483, respectively); Liberase TM and DH at concentrations of 1.4 W, 0.7 W, 0.07 W, and 0.035 W (TM: n = 391, 650, 643, and 571; DH: n = 535, 491, 586, and 482, respectively); and mechanical methods (n = 559). Data were obtained from five independent experiments. Data are expressed as mean ± SEM. Different letters (a, b) indicate significant differences between groups (P < 0.05). Retrieval methods for preantral follicles, i.e., Collagenase IA, Collagenase IV, Liberase TM, Liberase DH, and Mechanical methods are abbreviated as IA, IV, TM, DH, and Mech, respectively. Wünsch unit (W).

We observed that increasing the enzyme concentrations gradually increased the number of grade 1 follicles. However, the number of normal morphological oocytes decreased at high enzyme concentrations (0.5, 1 mg/ml of collagenase IA and IV; 0.7, 1.4 W/ml of Liberase TM and DH) compared with those at low enzyme concentrations (0.05, 0.1 mg/ml of collagenase IA and IV; 0.035, 0.07 W/ml of Liberase TM and DH). Therefore, concentrations of 0.1 mg/ml of collagenase IA or IV and 0.07 W/ml of Liberase TM or DH were found to provide a sufficient number of grade 1 and grade 2 follicles as well as an adequate percentage of normal oocytes (Fig. 1, C, D).

Effect of different enzymes and concentrations in enzymatic isolation on oocytes derived from PAFs

The oocytes of PAFs are densely packed with multiple compact layers of GCs, making it difficult to observe the detailed changes in both the ZP and oocyte membranes during enzyme exposure. Additionally, oocytes with a highly corroded, thinned, or transformed zona pellucida or those with a damaged membrane are unable to undergo mechanical granulosa cell removal, often leading to oocyte disruption before complete denudation.

Therefore, to investigate the primary effects of enzymatic isolation on the oocytes of PAFs and explain the high mortality and dysmorphology observed post-isolation, we exposed mechanically denuded oocytes to different types of enzymes at varying concentrations for 1 h, as shown in Figs. 2A–D. We observed that high concentrations (0.5–1 mg/ml or 0.7–1.4 W) of all enzyme types rapidly damaged the ZP within 15 min. Specifically, collagenases IA and IV tended to deform both the ZP and the oocyte basal membrane, whereas Liberase TM and DH showed higher efficacy in dissolving the ZP but caused less membrane damage. Lower concentrations (0.05–0.1 mg/ml or 0.035–0.07 W) did not cause any observable changes, suggesting that these concentrations may be more suitable for isolating PAFs as they minimize possible damage to the oocytes, thereby supporting further culture of PAFs.

Fig. 2.

Representative images of enzyme effects on the zona pellucida (ZP), oocyte membrane, and transzonal projections (TZPs). All oocytes were collected from preantral follicles using mechanical methods and were mechanically denuded. (A) Oocytes were exposed to collagenase IA at concentrations of 0.05, 0.1, 0.5, and 1 mg/ml. (B) Oocytes were exposed to collagenase IV at concentrations of 0.05, 0.1, 0.5, and 1 mg/ml. (C) Oocytes were exposed to Liberase TM at concentrations of 0.035, 0.07, 0.7, and 1.4 W/ml. (D) Oocytes were exposed to Liberase DH at concentrations of 0.035, 0.07, 0.7, and 1.4 W/ml. Scale bar = 50 µm. Wünsch unit (W). (E) Phalloidin staining images in oocytes with a single layer (scanning thickness: 0.73 µm) (a-a4), and in the entire ZP (b-b4). (F) Number of TZPs in a single layer of the oocyte population (IA: n = 57, IV: n = 58, TM: n = 58, DH: n = 56, Mech: n = 53). (G) Total intensity of TZPs located in the ZP of oocytes obtained from different retrieval methods (IA: n = 58, IV: n = 59, TM: n = 61, DH: n = 61, Mech: n = 60). Data are expressed as the mean ± SEM. Different letters (a, b) indicate significant differences between groups (P < 0.05). Retrieval methods for preantral follicles, i.e., Collagenase IA, Collagenase IV, Liberase TM, Liberase DH, and Mechanical methods are abbreviated as IA, IV, TM, DH, and Mech, respectively.

Enzymatic isolation method induces TZP loss

TZPs are integral to the complex interplay between oocytes and GCs, ensuring that the oocyte develops properly and acquires the competence necessary for successful fertilization and embryonic development. F-actin was visible between oocytes and GCs in all groups. However, the data showed that the enzymatic groups exhibited a significantly lower number of TZPs than those in the mechanical groups (P < 0.0001) (Fig. 2, Ea-a4, F), and reduced the total fluorescence intensity of F-actin in TZPs (IA, IV: P < 0.01; TM, DH: P < 0.001) (Fig. 2, Eb-b4, G). Therefore, the enzymatic isolation resulted in a significant loss of TZPs immediately after PAF retrieval.

Effect of enzymatic isolation on PAF culture

We confirmed that low concentrations of enzymes (0.1 mg/ml of collagenase IA and IV, 0.07 W of Liberase TM and DH) were suitable for further culture, maintaining an acceptable quantity and quality of PAFs. In vitro culture of PAFs is a crucial step in optimizing collection methods. The typical morphology of follicles during the 18-day culture period is shown in Fig. 3A. Viability was assessed based on the oocyte and follicle status, including dark and dense-looking oocytes, extruded oocytes, and degenerated follicles, which were considered non-viable during culture. Follicle viability gradually decreased during the culture period (Fig. 3B), with a notable decline during the first six days. In the enzymatic groups, the survival rates dropped by > 45% after two days and by > 70% after six days of IVG, whereas the survival rates of the mechanical groups decreased by less than 4% and 30%, respectively. On day 2 of IVG, the oocytes were denuded to assess the high mortality rate in the enzymatic groups, which revealed observable deformation of the ZP and breakdown of the oocyte membrane (Fig. 3E). On day 18, the terminal survival rate of the enzymatic group was significantly lower than that of the mechanical group (P < 0.0001). Among the enzymatic groups, collagenase IA resulted in the lowest viability (IV: P < 0.01; TM and DH: P < 0.001) (Fig. 3B). All tested enzymatic isolation methods resulted in comparable outcomes for antrum formation (Fig. 3C) and follicle diameter (Fig. 3D) compared to those of the mechanical group (P > 0.05). These results indicate that collagenase-based PAF isolation critically damages oocytes, reducing their viability during IVG, whereas the formation and growth of follicles remain unaffected.

Fig. 3.

Development of preantral follicles retrieved using different methods during in vitro culture. (A) Representative follicle morphology in each group on day 0 (a-a4), day 6 (b-b4), day 12 (c-c4), and day 18 (d-d4, e-e4) of in vitro culture. Scale bars: a-d4 = 200 µm, e-e4 = 500 µm. (B) Oocyte/follicle viability over the 18-day culture period (IA: n = 1454, IV: n = 984, TM: n = 782, DH: n = 918, Mech: n = 314). (C) Percentage of antrum formation after 18 days of culture. (D) Diameters of follicle populations on day 18 (IA: n = 138, IV: n = 168, TM: n = 154, DH: n = 172, Mech: n = 208). (E) Denuded oocytes derived from follicles after 2 days of culture (a-e), with enlarged representative images of dead oocytes (red arrow) (a1-e1) and deformed zona pellucida (ZP) in the enzymatic groups. Scale bar = 200 µm. Data were obtained from 12 independent experiments. All data are expressed as mean ± SEM. Different letters (a, b, c) indicate significant differences between groups (P < 0.05). Retrieval methods for preantral follicles, i.e., Collagenase IA, Collagenase IV, Liberase TM, Liberase DH, and Mechanical methods are abbreviated as IA, IV, TM, DH, and Mech, respectively.

Enzymatic isolations induce the extrusion of oocytes and cumulus-oocyte complexes (COCs)

The connection between follicular cells and oocytes plays a crucial role in regulating oocyte development. During PAF culture, we observed that the oocytes extruded from the follicle and subsequently degenerated (Fig. 4A). Oocyte extrusion was particularly pronounced in the enzyme treatment groups (Figs. 4B, C). Extrusion events predominantly occurred between days 4 and 6, with significantly higher rates in the enzymatic groups (IA, IV: P < 0.01; TM, DH: P < 0.001) than in the mechanical group, in which extrusion primarily occurred between days 2 and 4. Furthermore, oocyte extrusion appeared to correlate with the loss of TZPs, as TZPs were present only at the remaining connection sites between the extruding oocyte and follicle (Fig. 4D). Additionally, the enzymatic groups exhibited a significantly higher rate of COC extrusion than in the mechanical group (P < 0.0001) (Fig. 4E, F). This indicates that enzymatic isolation methods leads to a significantly higher rate of oocyte and COC extrusion, correlating with the loss of transzonal projections, compared to that with mechanical isolation techniques.

Fig. 4.

Extrusion of oocyte and cumulus-oocyte complex (COC) during in vitro culture. (A) Oocyte extruded from the follicle. Scale bar = 100 µm. (B) Dynamics of oocyte extrusion throughout the culture period at 2-day intervals. (C) Percentage of oocyte extrusion during 18 days of in vitro growth (IVG). (D) Transzonal projections (TZPs) at the remaining connection site between the extruding oocyte and the follicle (white arrow). (E) Extrusion of the COC after 18 days of culture. The normal COC is located inside the antrum, whereas the extruded COC is typically located outside the antrum. Scale bar = 200 µm. (F) Percentage of COC extrusion on day 18. Data were obtained from 12 independent experiments. All data are expressed as mean ± SEM. Different letters (a, b, c) indicate significant differences between groups (P < 0.05). Retrieval methods for preantral follicles, i.e., Collagenase IA, Collagenase IV, Liberase TM, Liberase DH, and Mechanical methods are abbreviated as IA, IV, TM, DH, and Mech, respectively.

Impact of enzymatic isolations on oocyte quality during the long-term culture of PAFs

After 18 days of culture, surviving oocytes in all groups were typically enclosed by multiple layers of cumulus cells, forming COCs (Fig. 5, Aa-a4), and exhibited a normal cytoplasmic distribution with a rounded morphology (Fig. 5, Ab-b4). In contrast, extruded oocytes failed to survive beyond a few days after extrusion. Further, there were no significant differences in the average oocyte diameters among the enzymatic groups (P > 0.05). However, the collagenase IA group exhibited a significant reduction in oocyte diameter compared to that in the mechanical group (P < 0.05) (Fig. 5B).

Fig. 5.

Developmental competence of oocytes and statistics of intercellular transzonal projections (TZP) derived from preantral follicles retrieved using different enzymes after 18 days of in vitro growth (IVG). (A) Cumulus-oocyte complexes (COCs) (a-a4) and their oocytes (b-b4) were isolated from follicles after 18 days of culture. Scale bar = 300 µm. (B) The diameter of oocytes on day 18 derived from different retrieval methods (IA: n = 131, IV: n = 149, TM: n = 134, DH: n = 167, Mech: n = 198). (C, D) Meiotic competence of oocytes based on nuclear status with filamentous chromatin/stringy chromatin (FC/SC) or germinal vesicle (GV) stage (nuclei stained with DAPI, shown in blue; nuclear membrane stained with Lamin B, shown in red). Data were generated from three independent experiments. (E, F) Lipid accumulation in oocytes derived from different retrieval methods (IA: n = 24, IV: n = 30, TM: n = 31, DH: n = 30, Mech: n = 31), measured by the fluorescence intensity of Nile red. The nuclei were stained with DAPI (blue), and lipids were stained with Nile red (red). Data were obtained from three independent experiments. (G) Phalloidin staining images in COCs (a-a4), in oocytes with a single layer (scanning thickness: 0.73 µm) (b-b4), and in the entire zona pellucida (ZP) (c-c4). (H) The number of TZPs in the single layer of various oocyte populations (IA: n = 18, IV: n = 24, TM: n = 26, DH: n = 26, Mech: n = 29). (I) Total intensity of TZPs located in the zona pellucida of oocytes derived from different retrieval methods (IA: n = 17, IV: n = 17, TM: n = 19, DH: n = 17, Mech: n = 20). Data are expressed as mean ± SEM. Different letters (a, b, c) indicate significant differences between groups (P < 0.05). Retrieval methods for preantral follicles, i.e., Collagenase IA, Collagenase IV, Liberase TM, Liberase DH, and Mechanical methods are abbreviated as IA, IV, TM, DH, and Mech, respectively.

The growing oocytes collected from PAFs were at the filamentous chromatin (FC) and stringy chromatin (SC) stages, whereas in vivo fully grown oocytes collected from large antral follicles were at the germinal vesicle (GV) stage [47]. Representative images of the FC/SC and GV stages are shown in Fig. 5C. The percentage of oocytes reaching the GV stage was significantly lower in the collagenase IA group than in the other groups (IV, TM, DH: P < 0.05; Mechanical: P < 0.01), and the collagenase IV group showed a significantly reduced percentage compared to those in the mechanical group (P < 0.05) (Fig. 5D). Lipid accumulation, as indicated by Nile Red staining of oocytes on day 18, is shown in Fig. 5E. The collagenase IA group exhibited a significantly lower level of lipid accumulation than that in the Liberase TM and mechanical groups (P < 0.05), whereas the remaining groups showed no significant difference (P > 0.05) (Fig. 5F).

The connection between the oocytes and nearby cells was assessed using F-actin staining. The COCs exhibited a dense F-actin signal at the contact sites between the oocytes and cumulus cells across all groups (Fig. 5, Ga-a4). No significant differences (P > 0.05) were observed in the number of TZPs (Fig. 5, Gb-b4, H) or the total intensity (Fig. 5, Gc-c4, I) between the enzymatic and mechanical groups.

These results indicate that enzymatic isolation methods do not interfere with COC formation or TZP establishment. The use of Liberase TM and DH for isolation has been shown to facilitate oocyte growth, GV-stage attainment, and lipid accumulation in a manner comparable to that of mechanical isolation methods. In contrast, the use of collagenases IA and IV significantly reduces these parameters.

Effect of enzymatic isolations on oocyte apoptosis and follicular cells

RT-qPCR was performed to assess oocytes and their associated follicular cells across the different enzymatic isolation groups at the molecular level. The mRNA levels of representative genes were analyzed on days 1 and 18, including anti-apoptotic BCL2, pro-apoptotic P53 and BAX, and oocyte-secreted factors BMP15 and GDF9.

The results indicated that the expression levels of all three genes in oocytes and follicle cells did not significantly differ among the enzyme-treated groups on day 1 of culture. However, BCL2 was significantly downregulated in the oocytes from the collagenase IA and IV groups, whereas P53 was upregulated in all enzyme-treated groups compared to that in the mechanical groups (P < 0.05) (Fig. 6A). In contrast, the opposite expression pattern was observed in surrounding follicular cells. While the anti-apoptotic gene BCL2 was upregulated in the follicular cells in the collagenase IA and IV groups, the pro-apoptotic genes P53 and BAX were downregulated in the follicular cells of all the enzyme-treated groups (P < 0.05) (Fig. 6B). These molecular findings on follicle cell apoptosis aligned with the observations of dead cells per follicle, where the number of nonviable cells did not significantly differ among the enzyme-treated groups (P > 0.05), but was significantly lower than that in the mechanical group (P < 0.0001) (Figs. 6C, D). These results indicate that enzymatic isolation induces apoptosis in oocytes while minimizing apoptosis in follicular cells during the retrieval process, whereas mechanical isolation has the opposite effect.

Fig. 6.

Relative expression of apoptosis-related genes in follicular cells and oocytes retrieved using different enzymatic methods after isolation and 18 days of in vitro culture (IVG). (A, B) Relative mRNA expression levels of apoptosis-related genes (BCL2, BAX, P53) in oocytes and follicular cells on day 1. (C) The average number of dead follicular cells on day 1, derived from different retrieval methods (IA: n = 43, IV: n = 44, TM: n = 44, DH: n = 42, Mech: n = 31). (D) Representative images of dead follicular cells on day 1, visualized by differential staining methods. All nuclei were stained with Hoechst (gray), and nuclei of dead cells were stained with PI (red). (E, F) Relative mRNA expression of apoptosis-related genes and oocyte-secreted factors (GDF9, BMP15) in oocytes, and of apoptosis-related genes in follicular cells on day 18. (G) The average number of dead follicular cells on day 18 derived from different retrieval methods (IA: n = 14, IV: n = 24, TM: n = 29, DH: n = 26, Mech: n = 27). (H) Representative images of dead follicular cells on day 18. Relative mRNA expression is reported as a fold change from the mechanical groups (log2 scale); a value of 0 represents the baseline of target gene expression in the mechanical groups. Data of differential staining were obtained from three independent experiments, and relative mRNA expression data were obtained from four independent experiments. Data are expressed as mean ± SEM. For relative mRNA expression: a indicates a significant difference compared with the mechanical group; b denotes significant differences in the IA group; c indicates significant differences in the IV group; d shows significant differences in the TM group; and e highlights significant differences in the DH group. For the average number of dead cells, different letters (a, b) indicate significant differences between groups (P < 0.05). Retrieval methods for preantral follicles, i.e., Collagenase IA, Collagenase IV, Liberase TM, Liberase DH, and Mechanical methods are abbreviated as IA, IV, TM, DH, and Mech, respectively.

On day 18, BCL2 expression was downregulated in oocytes from the collagenase IA and IV groups (P < 0.05), whereas BAX was upregulated in the collagenase IA group (P < 0.05) compared to that in the mechanical groups. However, no significant differences were observed in the expression of these apoptosis-related genes in oocytes retrieved using Liberase TM and DH. To further assess the impact of these enzymes on oocyte quality, we analyzed the expression of oocyte-secreted factors GDF9 and BMP15. GDF9 expression was downregulated in the collagenase IA group (P < 0.01), whereas BMP15 expression was reduced in the collagenase IA, IV (P < 0.01) and Liberase DH (P < 0.05) groups (Fig. 6E). Despite the differences in apoptotic gene expression in oocytes, no significant variations were observed in follicular cells (Fig. 6F), consistent with a comparable average number of dead cells across the groups (P > 0.05) (Figs. 6G, H). This result demonstrates that collagenases IA and IV have detrimental effects on oocyte quality by inducing apoptosis, whereas none of the enzymatic isolation methods adversely affect follicular cells in long-term culture.

Discussion

The use of different enzyme types, along with variations in concentrations, culture systems, and species, complicates the assessment of their impact on long-term cultures. This study compared the potential effects of mechanical techniques and commonly used enzymes (collagenase IA, collagenase IV, Liberase TM, and Liberase DH), along with their concentrations, on the retrieval of PAFs and their subsequent culture. We determined the optimal enzymatic concentrations for retrieving porcine PAFs across all enzyme types. Additionally, TZP loss, premature oocyte extrusion, and COC extrusion were found to be significantly influenced by enzymatic treatment. Further, the apoptotic status of oocytes provides an appropriate evaluation mechanism for the isolation methods.

To achieve our first objective, we aimed to optimize the concentrations of the four enzyme types, ensuring uniformity for the comparison of retrieval efficiency. The optimal concentrations achieved for collagenase IA and IV were 0.1 mg/ml, and Liberase TM and DH were 0.07 Wünsch/ml. This allowed the collection of the maximum number of oocytes with minimal damage. However, after isolation, we were unable to classify the oocytes in the enzymatic isolated follicles as dead or alive based on live-cell quantification, as reported in previous studies [48,49,50]. The enzyme treatments in our experiments resulted in completely dysmorphic oocytes in the denuded samples, which were still identified as live oocytes based on negative trypan blue and positive fluorescein diacetate staining (Supplementary Fig. 1). This result suggests that these live-cell quantification methods are not reliable for evaluating the efficiency of the enzymatic isolation of PAFs. Consequently, we used morphological assessment as a reliable alternative method for evaluating oocyte viability.

Collagenases do not seem to interact directly with the cell membrane [51]. However, we observed significant morphological changes in the oocyte membrane when exposed to high concentrations of collagenases IA and IV, which were not observed with the highly purified enzymes Liberase TM and DH. Previous studies have shown that non-purified collagenase IV induces apoptosis in chondrocytes [52]. This finding aligns with previous suggestions that isolated oocytes retrieved using crude collagenase have lower viability than that of oocytes retrieved with Liberase-grade enzymes owing to the higher endotoxin content in non-purified enzymes [25, 53]. The observed effects can also be explained by the fact that all the assessed enzymes contain both collagenase and neutral protease. Collagenase has been found to be particularly efficient in removing the ZP. Studies have shown that low concentrations of enzymes, such as 0.05 mg/ml, can partially dissolve the mouse ZP within 5–10 min [54, 55]. Although our results showed that denuded porcine oocytes exposed to low concentrations of these enzymes did not initially exhibit noticeable changes in the ZP and oocyte membranes, subsequent observations after two days of culture revealed deformation of the ZP and disruption of the membrane. This suggests that even at low enzyme concentrations, the ZP structure and oocyte membranes are affected during culture.

Collagens I and IV are present in the stromal and theca cell compartments, as well as in the granulosa cell compartment of PAFs, with a notably high concentration in the regions surrounding the oocyte [56]. Collagenases break the peptide bonds in collagen, a key component of the extracellular matrix (ECM), resulting in cell loosening. Additionally, highly purified collagenase VII-S, a member of the collagenase family, has been used to digest connective tissue components and liberate individual cells, similar to collagenase IA, collagenase IV, and Liberase grades. It induces damage to tight junction proteins and VE-cadherin, the key components of cell-to-cell adherent junctions in endothelial cells [57]. However, the effects of other collagenases on these junctions remain unknown. In our study, we observed variability in the location of PAFs within ovarian cortex fragments and in the number of granulosa cell layers. This inconsistency may lead to differential digestive activity during collagenase exposure across follicles. We suggest that COC extrusion is related to the disruption of collagen in the outer layer of GCs, whereas oocyte extrusion may result from the complete degradation of the collagen-rich area surrounding the oocyte. This process likely corresponds to a loss of adherence between the oocyte and its surrounding GCs. Furthermore, the disruption of collagen is known to facilitate cell migration [51, 58]. Consequently, the complex of the oocyte and inner granulosa cell layer may be pulled outward, leading to its extrusion during the reassembly of follicle cells during IVG.

Several previous studies have indicated that TZPs are recoverable [43, 44, 59,60,61]. However, in our study, we observed a significant percentage of oocyte extrusion in PAFs derived from enzymatic isolation during IVG, coinciding with the complete loss of TZPs. This suggests that extrusion may also be attributed to the loss of structural integrity in the ZP induced by the proteolytic activity of collagenase. The structural integrity of the ZP is maintained through peptide bonds that link amino acids within glycoproteins. Furthermore, the ZP comprises two subdomains, the ZP-N and ZP-C, which are connected by a protease-sensitive linker region [62]. However, proteases can also cleave these linkers [63,64,65]. In addition, evidence suggests that acidic and enzymatic conditions have different effects on the ZP. Acids induce gradual thinning of the ZP [66], whereas collagenase deforms the ZP and causes its subsequent dissolution, as observed in our results. Consequently, the loss of TZPs could be the result of multiple concurrent factors, including the loosening of cell-cell junctions and structural deficiencies of the ZP.

In cell isolation, gentle enzymatic dissociation has been shown to consistently result in lower levels of intracellular reactive oxygen species than with mechanical isolation methods [67]. In the present study, we observed a high rate of apoptosis and an increased number of dead GCs in PAFs cultured for 24 h after collection using mechanical methods. This increased cell death may have resulted from the GCs undergoing multiple dissection events and prolonged contact with the sharp needle surfaces, leading to significant damage. After 18 days of culture, the expression of apoptotic genes was similar in GCs cultured using both mechanical and enzymatic methods, consistent with a previous study on the culture of cat PAFs [12]. However, the expression of apoptotic and oocyte-specific growth factors in oocytes subjected to 18 days of IVG was significantly reduced in oocytes treated with crude collagenases (IA and IV). This finding suggests that, although enzymatic treatment exerts no long-term effects on GCs, crude collagenases adversely affect oocytes. In addition, oocytes and GCs respond differentially to apoptosis; therefore, they must be evaluated separately to objectively assess the effectiveness of isolation and culture methods.

In conclusion, the use of collagenase-based enzymes for retrieving PAFs has the potential to facilitate their development to the fully grown stage. However, even low concentrations of these enzymes can lead to significant detrimental effects on oocyte viability, apoptotic status, and the integrity of connections between the oocyte and surrounding GCs following isolation. The use of highly purified collagenase in culture may significantly enhance viability and promote the long-term culture of oocytes, whereas the use of crude collagenase is associated with adverse long-term effects. Thus, selecting appropriate collection methods is essential for successful in vitro oocyte culture, particularly when integrating extrusion-blocking systems to ensure optimal development. Addressing these challenges can significantly enhance the process of human follicle isolation, as follicles are embedded in a dense and resilient matrix, rendering mechanical isolation both challenging and labor-intensive.

Conflict of interests

The authors declare no conflicts of interest in the research.