Follicle morphology during 3-D culture in HA gel
Intact secondary follicles with centrally located oocytes surrounded by 2-3 layers of granulosa cells were selected for encapsulation in HA gel. Pre-antral follicles measured 140.4 ± 29.1 µM with oocyte diameters of 63.5± 4.0 µM at time of seeding. Figure 1 depicts follicle morphology and morphological changes during in vitro growth. The HA-embedded preantral follicles maintained their 3-D spheroid architecture throughout the in vitro culture interval. Between Day 1 and 4, there was a 1.4 fold increase in follicle size with a mean diameter measuring 201.6 ± 28.5 µm. The basal lamina with thecal cell layer was evident in most follicles. Antral cavity formation by encapsulated follicles was generally initiated by day 7 or 8 of culture. Ovulation was typically triggered between days 9-12, when ~50% of cultured follicles displayed an antrum. The average oocyte diameter in collected metaphase II oocytes was 90.2 ± 8.0 µM.
HA gel also supported the culture of ovarian tissue fragments. Growth of follicle clusters (FL-cluster) containing 6-8 intact preantral follicles from fresh and vitrified ovaries is also shown in Figure 1. Follicular architecture was preserved within these FL-clusters. Granulosa cells around the oocytes proliferated extensively. Within four days of culture it was difficult to discern individual follicles within the cultured fragment. Antrum formation was however still easily visualized. Mature oocytes could be retrieved from this “organoid” like follicle construct after hCG trigger. No differences were observed in oocyte morphology or final diameter between oocytes derived from individually isolated follicles or FL-clusters. Enzymatic treatment of ovaries followed by mechanical manipulation to obtain either individual follicles or FL- clusters was the most efficient means for handling fresh ovaries. In contrast, with vitrified ovaries, attempts to isolate intact individual follicles using collagenase were not satisfactory. All too often the oocytes were released or damaged. We found the best approach was to handle the tissue minimally. Mechanical dissection of the vitrified/warmed ovaries to obtain FL-clusters that could subsequently be encapsulated in HA-gel appeared to be the optimal method to derive mature oocytes. Embedding the FL-clusters also helped to keep the follicles intact and prevented the premature ovulation of oocytes before the end of the in vitro culture interval.
Table 1 compares in vitro follicle growth and maturation data for oocytes from fresh and vitrified ovaries. With fresh ovaries, the oocyte ovulation and the final rate of oocyte maturation were similar regardless of whether follicles were cultured individually or as a cluster. In contrast, meiotic maturation to the metaphase II stage was significantly lower with the FL-clusters isolated from vitrified ovarian tissue (34%) as compared to fresh FL-cluster and FL-Individual (55%and 59%, respectively, p=0.007).
Morphologic assessment of oocyte competence
Changes in GV oocyte chromatin organization from the preantral to the antral follicle stage were evaluated. At outset of follicle culture, all of the isolated GV oocytes exhibited the NSN staining pattern, with the nucleolus none surrounded by chromatin. Chromatin pattern was re-assessed when antrums were observed in 40-50% of follicles. GV oocytes were extracted by pipetting the HA encapsulated follicles. We then classified the oocytes into two groups SN and NSN based on chromatin pattern after Hoechst staining (Figure 2 a,b). Oocyte nuclear chromatin in the GV oocytes had reorganized during the in vitro culture interval such that 83% of oocytes (38/46) recovered from follicles now exhibited the SN configuration. Distinct perinuclear rings were observed surrounding the nucleolus.
We also examined meiotic spindle morphology in the in vitro matured metaphase II oocytes. Birefringent spindles were easily visualized upon live imaging of ovulated oocytes with polarized light (Figure 2c). Immunofluorescent staining was performed on fixed metaphase II oocytes to further assess spindles. Figure 2d shows an example of an oocytes with a normal spindle configuration. Normal meiotic spindles were detected in 72% (28/39) of IVM eggs.
Functional competence of oocytes to fertilize and develop
IVM data from three experiments were pooled. (Table 2). A total of 402 pre-antral follicles were embedded in HA gel. Antrums were observed in 55% of follicles by day 9. The morning after hCG trigger, 314 oocyte-cumulus complexes ovulating from the cultured follicles were collected for IVF. The rate of GVBD to metaphase I was 84% (264/314) and rate of maturation to metaphase II was 72.6%. (228/314). The fertilization rate, based on observation of two pronuclear zygotes or 2-cell embryos at fertilization check was 82.5% (188/228). The subsequent blastulation rate with these IVM derived oocytes was 46.3% (87/188). Blastocysts of good morphology were observed with a well organized trophectoderm and a distinct compact inner cell mass. A total of 61 blastocysts were cryopreserved for future transfer experiments to determine their ability to produce live offspring. Hoechst staining was performed to approximate blastomere number (Figure 3). The mean cell count in the assessed blastocysts was 55 ± 7.5. However since the more advanced blastocysts were cryopreserved, this somewhat low overall cell count was not unexpected.
Vitrified blastocysts were sent to Charles River laboratories for transfer experiments as it was not possible to house post-surgical animals at our facility. A total of 46 thawed blastocysts were transferred into three recipients. Two recipients each received 16 embryos via bilateral uterine transfer and the third received 14 embryos. One of the three recipient mice achieved a pregnancy. A single pup was delivered. Health of the pup could not be determined as it was cannibalized shortly after birth. Logistics and cost of using an outside laboratory made it difficult to expand on these transfer attempts with IVM blastocysts.
Follicle encapsulation and effect of extracellular matrix proteins
In this series of experiments, we investigated modifying the HA gel with extracellular matrix components to determine if this might further enhance IVM outcomes. To this end, we tested laminin as well as two extracellular matrix complexes, one from mouse tumor cells (Matrigel), the other from human placenta. The basic composition of both ECM matrices consisted of collage IV, laminin, glycosaminoglycans, chondroitin sulfate and growth factors reflective of their tissue of origin.
Follicles encapsulated in HA-LM, HA-MG and HA-PM exhibited more rapid proliferation of granulosa cells and within a few days of culture it became difficulty to visualize the oocytes. Table 3 compares follicular growth and oocyte maturation in HA alone versus in HA modified with LM or ECMs. Significant differences were noted between the treatments in terms of follicle retention in the gel, percent antrum formation and ultimately ovulated complexes after hCG trigger. Preantral follicles encapsulated within HA gel with either type of ECM matrix were better retained in 3-D culture to the end of the growth interval (HA-MG 88 %, HA-PM 94% vs HA alone 69%, p<0.005). Significantly lower antrum formation was noted amongst follicles in HA-LM (29%), HA-MG (18%) and HA-PM (26%) as compared to HA alone, 48%, p=0.006). The HA-Matrigel encapsulated follicle treatment group had a lower rate of ovulation (37%) as compared to all three other treatment groups ( 65-67%, P=0.001). The rate of maturation to metaphase II oocytes was 74% for HA gel alone as compared to HA-LM ( 67%), HA-MG (56%) and HA-PM (58% ) but differences were not statistically significant. Inclusion of matrix components in the HA gel did however significantly increase overall estradiol secretion measured on day 8 and 10 of culture (Figure 4). The highest level of estradiol per encapsulated follicle was with HA-LM (929 ± 104 pg/ml) as compared to HA alone (309 ± 15 pg/ml, p<0.001).
HA-LM appeared to be the most promising amongst the ECM products tested. In a subsequent experiment, we tested modifying the HA gel by addition of laminin at varying concentrations (Figure 5). HA-LM at either 50 or 100 ug /ml worked well in conjunction with the HA gel in supporting follicle growth, oocyte ovulation and maturation. However, no advantage was observed with HA-LM over HA alone. Hoechst staining of GV oocytes before hCG trigger revealed very similar chromatin patterns between oocytes derived from the three culture models. The percentage of GV oocytes with the SN staining pattern was 93.8%, 73.3% and 77.8% for HA, HA-LM50 and HA-LM 100, respectively. Normal chromosome alignment and spindle configuration was observed in 64.1%, 54.8% and 57.7 % of immunostained metaphase II oocytes from HA, HA-LM50 and HA-LM100, respectively. We also used polarized light to view meiotic spindles in living oocytes (unfixed). Retardance values measured for meiotic spindles in the groups did not differ (HA 2.03 ± 0.30, HA-LM50 2.05 ± 0.34, HA-LM100 1.96 ± 0.04), suggesting similar spindle birefringent characteristics. Estradiol secretion with different concentrations of laminin are shown in Figure 6. Estradiol secretions in the HA-LM50 treatment were higher than HA alone, most likely reflecting more proliferation of the granulosa cell compartment.