Animals

C57BL6 female mice and SJl/J males were purchased from Charles River Laboratories, Wilmington, MA and maintained as a breeding colony. Animals were housed at the Cleveland Clinic’s Animal Care facility within an environment of controlled temperature and lighting (12 h dark:12 hr light). Animals were handled according to NIH guidelines and following IACUC protocols of the Cleveland Clinic. Mice were sacrificed by cervical dislocation. Ovaries were dissected from 10–12-day old female mice for follicle isolation. Each experimental run consisted of ovaries from 3–6 pups. Data from two independent experiments was pooled.

Follicle isolation and culture

Ovaries were placed in an organ culture dish containing pre-warmed Leibovitz-15 medium (Invitrogen; Carlsbad, CA) with 0.1% collagenase (Type I; Worthington Biochemical, Lakewood, NJ). The dish was placed on a laminar flow bench top at 37°C for 90 minutes. At 30 minute intervals, ovaries were transferred to a new dish containing L-15 medium and gently pipetted using Eppendorf pipette tips of decreasing bore size to facilitate release of follicles from underlying stroma. Undigested ovarian tissue was returned to the collagenase dish for further enzymatic digestion. The ovarian digest was visualized under a dissecting microscope at 40X magnification, and isolated follicles were collected with a finely drawn glass micropipette. Follicles were washed three times in culture medium to remove any traces of L-15 and enzyme. Culture was performed in α-modified Minimum Essential Medium (α-MEM; Invitrogen) supplemented with 5% fetal bovine serum, 100 mIU/ml FSH,10 mIU/ml LH, ITS (10 μg/ml insulin, 5 μg/ml transferrin and 5 ng/ml selenium; Invitrogen). Secondary follicles (100–130 μm, oocyte <65 μm) with a centrally located oocyte, surrounded by 2 or more layers of granulosa cells and enclosed within an intact basement membrane were selected for experiments.

Isolated follicles were used for fresh experiments and also cryopreserved.

Follicle vitrification and warming

The protocol for vitrification and warming of isolated pre-antral follicles has been previously described in detail [19]. The basal medium for all vitrification solutions was L-15 with 20% synthetic serum substitute (SSS; Irvine Scientific; Irvine, CA). Follicles were equilibrated for 5 min in 2 M ethylene glycol (EG; Sigma, St. Louis, MO) followed by a 30–60 second incubation in vitrification solution containing 6 M EG with 0.3 M raffinose (Sigma). All steps were performed at room temperature (22–25°C)

Vitrified follicles were warmed by immersion in L-15 medium containing 1.0 M sucrose. Follicles were held at room temperature for 10 minutes. Follicles were then quickly transferred to L-15 medium without sucrose and incubated at 37°C for another 10 minutes. The final wash step was in follicle culture medium (supplemented α-MEM) pre-equilibrated at 37°C with 6% CO₂. Intact follicles were pooled in a culture dish containing fresh medium. Follicles were embedded in 3 mg/ml of HA in groups of 5–7 follicles per gel as described below. Growth of vitrified-warmed follicles (n = 74) after HA- embedding or in standard culture wells was tested.

Follicle encapsulation

A novel tyramine-based hyaluronan (HA) hydrogel (provided by A. Calabro, Cleveland Clinic, Dept. of Biomedical Engineering) was tested for its biocompatibility with ovarian follicles. The HA was prepared at concentrations ranging from 2 to 5 mg/ml in either phosphate-buffered saline or Global medium (Life Global; Guilford, CT). HA was also formulated with extracellular matrix proteins (ECM) to try to further enhance follicle growth. Matrigel, an ECM preparation extracted from Engelbreth-Holm-Swarm mouse sarcoma (10 μg/ml; BD Biosciences; San Jose, CA) was added to HA in a ratio of 1:9. This was done on ice to prevent gelation of the ECM.

Aliquots of HA and ECM modified-HA (ECM-HA) were stored at −20°C and thawed just before use. The HA was activated for follicle embedding by addition of 5 μl of horse radish peroxidase (HRP) to 500 ul aliquots of HA. Cross-linking of HA was initiated by exposure of 25 μl of activated HA to one microliter of 0.03% hydrogen peroxide. Gels were rinsed in pre-equilibrated follicle culture medium before final plating.

We tested three different methods for embedding the pre-antral follicles in the HA: (1) Drops; (2); Microcapillary plugs; (3) Cylindrical beads. In Method 1, a one μl drop of 0.03% hydrogen peroxide and a 25 μl drop of activated HA were placed next to each other on a round glass coverslip. Follicles in groups of 7–10 were deposited into the HA drop, being careful not to track in too much medium. The HA drop with follicles was picked up with a microcapillary tube attached to a mouth pipet and then slowly expelled on top of the hydrogen peroxide. Gelation of the HA was generally completed within 3–4 minutes. The coverslip with HA gel drop was then placed in a 24-well dish with follicle culture medium pre-equilibrated at 37°C with 6% CO₂ in air. In Method 2, the HA drop with 1–3 follicles was drawn up into a glass capillary tube, with the follicles in the middle of the HA fluid column. The tip of the capillary tube was then positioned over the hydrogen peroxide drop. A small amount of HA was expelled into the hydrogen peroxide, and then quickly drawn back in to the capillary tube to initiate gelling of the HA-follicle column. It was critical to avoid drawing air bubbles back up in to the capillary tube, since this interfered with gelling. The HA-gel plug with follicles was then gently expelled into a 24-well dish and pre-equilibrated medium was quickly added. Method 3 was similar, to the first method described, except that the HA was shaped into a bead as it gelled. This was done using two 21 G tuberculin syringes with the beveled needles bent at a 90 angle. Two minutes into the gelling process, the needles were used to repeatedly “push” the edges of HA drop until it formed a cylindrical bead with encapsulated follicles.

In vitro follicle growth and oocyte meiotic competence

HA-gels containing follicles were placed individually in a 24-well plate with 800 μl of medium. Control wells were plated with a similar number of follicles per well. For each treatment, 6–18 wells were seeded, depending on total follicles available from the ovarian digest. After plating, the follicles were cultured at 37°C with 6% CO₂ in air for 12 days. The medium was exchanged every other day by replacing one half of the culture volume with fresh equilibrated medium. During this in vitro culture (IVC) interval, follicles were imaged at each medium exchange.

At the end of the follicular growth phase, in vitro maturation was triggered by exposure of the follicles to culture medium supplemented with 1.5 IU/ml of human chorionic gonadotrophin (hCG), and 5 ng/ml of epidermal growth factor (EGF) (R&D Systems, Minneapolis, MN). Maturation was allowed to proceed overnight for 18–20 hours. Follicles were then exposed briefly to hyaluronidase (80 U/ml) to assist in removal from the gel. Encapsulated follicles were then dislodged from the HA gels using tuberculin syringes. Oocytes were denuded of granulosa cells by pipetting with finely drawn glass micropipets of decreasing diameter. After rinsing to remove residual enzyme, the oocytes were placed in 5 μl drops of fresh medium, separated according to initial well number and treatment group. Oocytes were assessed using an inverted microscope at 300X magnification. Each oocyte was photographed and its meiotic status documented. Live imaging of metaphase II oocytes for the presence of a meiotic spindle was performed using the Oosight Imaging system (CRI; Hopkinton, MA).

Estradiol secretion

Levels of 17ß-estradiol were measured in conditioned media from encapsulated and control follicle cultures collected at various time points during IVC. For these measurements, conditioned medium was collected for each treatment at designated timepoints from three replicate wells containing 7–10 follicles. Estradiol was measured in duplicate using the Beckman Coulter Access 2 chemoluminescent immunoassay. Data was adjusted for number of follicles in each well by dividing the total measured estradiol secretion by the number of follicles at culture initiation. The sensitivity of the assay was 20 pg/ml.

Follicle assessment and outcome parameters

Light microscopic evaluation of follicles was carried out using an Olympus IX-70 light microscope with Hoffman modulation contrast optics and imaging software. Culture wells were monitored for continued follicular growth and antrum formation. Images were captured for morphometric analysis on Day 1 (24 hours after plating) and at 2–3 day intervals until final maturation of oocytes to the metaphase II stage. Care was taken to image each of the individual follicles embedded in the gel. Follicle diameters were measured in duplicate from the outer edge of the thecal cell layer. Oocyte diameter was measured (excluding the zona) at the outset of the experiment and after the final maturation step. Additional measurements were taken during IVC if the oocyte could be visualized. Follicles in which the oocyte was no longer surrounded by granulosa cells or where the granulosa cells appeared dark and degenerative were classsified as non-viable. Calculation of antrum formation, GVBD and MII formation was based on viable follicles present at the end of the in vitro culture interval.

Outcome parameters used to evaluate the biocompatibility of the HA-hydrogel with ovarian tissue were estradiol secretion, follicle survival to the end of the in vitro culture interval, progressive increase in follicle size, ability to form an antrum, resumption of meiosis (GVBD) and oocyte maturation to the metaphase II stage.

Statistical analysis

Statistical calculations were performed using Stats Direct software (Cheshire,UK). Statistical differences in follicle/oocyte diameters and estradiol secretion were analyzed using a two way ANOVA with repeated measures or a one-way ANOVA followed by the Tukey test for multiple comparisons. Categorical data was analyzed using the χ² test for independent samples. Differences were considered to be significant if P values were <0.05.

Biocompatibility and testing of HA concentrations

We wanted to first determine if HA was biocompatible with reproductive cells and the concentration range that might be most effective. The endpoint was ability to resume meiosis after 12 days of in vitro culture. HA was tested at four concentrations, 2, 3, 4 and 5 mg/ml.

In this experiment, Method 1 was used for embedding follicles. After 24 hours in culture, overt indications of cytotoxicity such as cellular degeneration and oocyte extrusion were absent. All four concentration of HA tested supported granulosa cell proliferation and in vitro follicular growth for 12 days. We observed that HA drops containing follicles had a tendency to flatten, especially at the lower HA concentrations. This affected follicle retention during IVC. Follicles located on the edges or sinking to the bottom of the gel, eventually ended up being extruded. Clearly, improvements to the encapsulation process will be needed to reduce follicle loss from the gel and assure maintenance of 3-D configuration in all embedded follicles.

Table 1 summarizes our results with the different HA concentrations tested. Amongst the follicles that remained embedded in the different HA treatments, 41–65% survived in vitro culture with an enclosed intact oocyte. Percent survival was not statistically different between the different HA gel concentrations. There was however a linear trend towards better survival with the lower gel concentrations (P = 0.039). The encapsulated follicles showed a continued increase in follicular diameter during the growth phase and retained their spherical shape. The resumption of meiosis after exposure to hCG and EGF indicate that HA was able to support follicular development in vitro. The proportion of oocytes capable of undergoing final maturation to metaphase II was not statistically different between HA concentrations. Polarized light imaging of metaphase II oocytes derived from HA gels revealed the presence of normal-looking meiotic spindles (Figure 1).

	2 mg/ml	3 mg/ml	4 mg/ml	5 mg/ml	P-value
Follicles initially plated	60	60	60	60
Follicles retained in gel	34	48	36	44
Survival * (%)	22 (65%)	24 (50%)	16 (44%)	18 (41%)	NS **
GVBD (%)	20 (91%)	24 (100%)	11 (69%)	15 (83%)	P < 0.05
MII (%)	12 (55%)	14 (58%)	7 (44%)	8 (44%)	NS

*Follicles with intact oocytes surviving 12 days in culture. Remaining follicles failed to develop and underwent atresia.

**Linear trend towards better survival with lower HA concentrations (P = 0.03). Follicles embedding in drop format.

Testing of alternative follicle encapsulation methods

Microcapillary gel plugs

Based on the above results, we decided the 2 or 3 mg/ml HA might be best suited for follicle growth. We then explored alternate methods of follicle encapsulation to achieve better follicle retention within the hydrogel and prevent flattening. In the first experiment, follicles were embedded in a thin column of HA (10 ul) drawn into a microcapillary tube to shape the hydrogel (Method 2). The inner diameter of the microcapillary tube measured 1142 μm. We found that at the two extreme ends of the HA column, gelling was incomplete. Reducing the HA volume to 5 ul and the peroxide concentration to 0.015%, gave the most consistent results. The 2 mg/ml HA did not have enough rigidity for this format.

Figure 2 (A-D) illustrates a single follicle in a microcapillary plug. The 3-D architecture of follicles embedded in 3 mg/ml HA was well maintained through the culture interval. Antrum formation was clearly visible by Day 6–7 in growing follicles and oocytes resumed meiosis after hCG trigger. Figure 2E depicts the growth pattern of encapsulated follicles and controls plated in culture wells. Follicle viability was maintained in HA gels. No difference was noted in percent follicles surviving IVC. Antral cavity formation and oocyte progression to the GVBD stage stage was in fact significantly higher in the HA gels (69% and 96%, respectively) as compared to control cultures (41% and 78%, respectively). Despite these positive indicators and the resumption of meiosis, final oocyte maturation to metaphase II was lower after growth in the HA microcapillary gel plugs. The reduced HA volume and closer proximity of the cross-linking agent to the follicles during gelling within the HA column may have caused subtle damage.

Although the data was still encouraging, this method of embedding was not practical for handling large numbers of follicles. At best 1–3 follicles could be handled per microcapillary plug. Technically it was challenging to keep all the loaded follicles centered in the 5 ul HA column. If during cross-linking they shifted downwards, the potential of follicle damage as a result of inadvertent exposure to peroxide was high. Failure to gel was always an issue if mixing between the microliter of peroxide and the edge of the HA gel in the microcapillary tube was not adequate.

Characterization of follicle growth in HA and ECM-modified HA

In this series of experiments, we characterized the growth pattern of secondary preantral follicles embedded in this novel tyramine-based hyaluronan (HA) hydrogel. We also looked at addition of ECM components to HA as a means to potentially improve follicle development and oocyte maturation. Control follicles were plated in a 24-well dish. Cylindrical beads containing 8–10 follicles were used and data was pooled from two independent runs. At the start of culture, the average diameter of follicles being distributed across the three treatment groups was 121.5 ± 12.9 μm. Average oocyte diameter was 57.2 ± 2.8 μm.

Typical morphology of preantral follicles and oocytes after 3-D culture in HA gels is shown in Figure 3. HA- embedded secondary follicles maintained their 3-D spheroid architecture throughout IVC. Multiple layers of granulosa cells formed around the oocyte. The diameter of HA-embedded follicles increased over the culture interval, measuring 385.6 ± 36.7 μm by Day 11 (Figure 4). Between Day 1 and 4, there was a 1.5 fold increase in diameter and mean follicle size was 183.4 ± 46.9 μm. Growth was accelerated from Day 4 through 11, resulting in a 2.1 fold increase in follicular diameter. An intact basal lamina with a thecal cell layer was typically observed.

In contrast, follicles in ECM-HA gels lost much of their spheroid symmetry by the fifth day of culture, as granulosa cells rapidly proliferated into the supporting matrix. A sharply delineated basal lamina and thecal cell layer was often difficult to see but follicular growth still continued in a 3-D fashion. Follicles in conventional 2-D culture flattened and lost their spherical shape within the first 3–5 days of culture. Granulosa cells proliferated and spread across the well surface. Oocytes became atretic and were spontaneously released from 25% of the follicles plated. By the 12th day of culture, granulosa:oocyte connections were tenuous, requiring very little to induce ovulation of the oocyte.

The pattern of estradiol secretion by the preantral follicles in the three different culture systems was compared (Figure 5). There was progressive increase in estradiol secretion amongst the encapsulated follicles over time in culture, suggesting continued growth. HA and ECM-HA embedded follicles secreted half as much estradiol by Day 5 as control follicles (P < 0.001). Thereafter estradiol levels increased quite rapidly. HA follicle cultures showed an eight fold increase in estradiol levels by Day 12 (703.6 ± 19.5 pg/ml) as compared to the five-fold increase observed in controls (1060.5 ± 15.8 pg/ml). Follicles in the ECM-modified HA exhibited the highest secretion of estradiol (4119.1 ± 161.6; P < 0.001). This indicated a 37-fold increase in secretion of estradiol between Day 5 and Day 12.

The developmental performance of follicles embedded in HA or ECM-HA is shown in Figure 6. Antral cavity formation was similar between treatments. A significantly higher percentage of oocytes from HA and ECM-HA encapsulated follicles resumed meiosis and underwent GVBD when compared to the control. The transition from GVBD to the metaphase II stage was however somewhat lower in the HA treatments (HA 47%, ECM-HA 48%, CT 67%; P < 0.05). Final oocyte diameter after IVC was comparable in all three culture systems (Figure 7).

In a follow up experiment, we individually cultured and tracked HA-embedded follicles for antrum formation and correlated this to oocyte nuclear status after hCG/EGF trigger and meiotic spindle organization . The rate of antrum formation was 65% (26/40) . Oocytes derived from follicles with antrums completed meiosis and progressed to the MII stage (88%; 23/26). Spindles were observed with polarized light in 20 of the 23 MII oocytes examined.

Development of cryopreserved follicles

Our last goal in this initial trial of HA embedding was to determine if the same methodology would support development of cryopreserved follicles. In this test run we assessed the developmental competence of cryopreserved follicles in this new 3-D culture system. Overall follicle survival after vitrification and warming was 92%. Warmed follicles were encapsulated in HA (n = 45) in groups of 5–7 and monitored for survival after IVC, GV breakdown and meiotic progression to the MII stage. This was contrasted to control wells with vitrified-warmed but non-embedded follicles (n = 29). No statistical difference was detected in any of the three parameters measured. Survival rate of HA-embedded follicles and the percentage of oocytes undergoing GVBD was 58% and 100%, respectively as compared to the control (79% and 96%, respectively). Oocytes derived from vitrified-warmed follicles embedded in HA were clearly able to mature to the metaphase II stage, much like control non-embedded follicles (54% vs 57%, respectively).