LAUSR

Does conventional ovarian stimulation with gonadotropins alter the molecular composition, size, and developmental fitness of superovulated mouse oocytes? Ovarian stimulation perturbs 31% of developmental proteome vs. 2% of transcriptome, yielding smaller oocytes that form blastocysts with less primitive endoderm and diminished fetal yield. Prior mouse studies aimed to assess the impact of superovulation on oocyte and embryo quality provided variable results. Aberrations were observed in some studies but not in others, and were associated with a variable time spent in the oviduct until retrieval for in vitro culture. Although the natural ovarian cycle lasts 4 days in mice, the conventional stimulation protocol time spans 2 days. No genome-wide studies have been devoted yet to the global gene expression of superovulated mouse oocytes and derivative embryos, to determine if gene products accumulate to the same extent as in natural cycles. Approx. 1100 female mice were injected, half with serial equine and human chorionic gonadotropin, eCG and hCG, to induce superovulation; the other half were injected with saline as control. Both groups were mated to vasectomized or fertile males to obtain, respectively, metaphase II and fertilized oocytes. These were removed from the oviducts, and followed up in vitro to blastocyst, or in vivo to term after surgical transfer to naturally cycling females. B6C3F1 oocytes (n= ∼16000) from superovulation (10 I.U. eCG+hCG) or natural ovulation were measured for diameter, and upon fertilization, they were cultured in KSOM(aa) to collect preimplantation stages for analyses (mass spectrometry; RNA sequencing; immunofluorescence for counting of trophectoderm, epiblast, and primitive endoderm cells). Embryos at the 4-cell stage were transplanted to naturally cycling females (8/female, 45 recipients). Results were compared between natural ovulation and superovulation with 48h (conventional) or 72h interval of eCG-hCG stimulation. Preimplantation embryos of superovulated oocytes were affected in 31% of the proteins (893 / 2855) vs. 2% of the transcripts (482 / 21784), compared to natural counterparts (adj.P<0.05, Wilcoxon test). Gene set enrichment analysis of the perturbed proteome returned the top-terms ‘thin zona pellucida’ (ZP1, ZP2, ZP3) and ‘abnormal inner cell mass apoptosis’ (DAB2, STAT3). Microscope measurements verified a thinner zona pellucida (p = 0.077, Wilcoxon test) along with a smaller diameter (p < 0.0001, Wilcoxon test) of superovulated oocytes, which gave rise to blastocysts deficient in primitive endoderm (p < 0.013, Fisher’s exact test). Since 529 of the 893 differently expressed proteins were depleted, we considered that ovarian stimulation provided insufficient time for protein accumulation. Increasing the eCG-hCG interval from 48 h (conventional) to 72 h restored oocytes’ diameters, and improved their fetal output from 25% to 59%, compared to 54% of natural ovulation (15 embryo transfers per group). Conversely, oocytes lost part of their developmental potential to the micromanipulation-assisted reduction of their volume. This study provides evidence of an additional novel effect of exposure to gonadotropins on mouse oocyte quality, whose mechanism is mediated not by the stimulated genital tract, but by the time-dependent accumulation of proteins in oocytes. This is an animal model study based on one mouse strain. Ovarian stimulation protocols differ between mice and humans. There can be yet other, more subtle or long-term differences between superovulated and naturally ovulated oocytes, than those described here. Proteome and transcriptome analysis cover much, but not everything. There is a trade-off between oocyte quantity and quality in mice subjected to superovulation. Cytological and molecular deficits define a ‘small oocyte syndrome’. Problematic is not so much the gonadotropin treatment, rather its timing. An evidence-based protocol for superovulation may be different from that used currently in mice. Not applicable