(40) Maternal-Effect Genes Altered In ART-produced Oocytes and Embryos

Camilo Pena-Bello^1,2; Ben Shaffer^1,2; Qidan Hu³; Eldin Jašarević^1,2,4; Mellissa R.W. Mann^1,2

1. Department of Obstetrics, Gynecology and Reproductive Sciences, University of Pittsburgh School of Medicine, Pittsburgh, USA

2. Magee-Womens Research Institute, Pittsburgh, USA

3. School of Medicine, Tsinghua University, Beijing, China

4. Department of Computational and Systems Biology, University of Pittsburgh School of Medicine, Pittsburgh, USA

 

Abstract Text:

Alarmingly, ~50 million couples worldwide are unable to conceive after 5 years of unprotected sex, with infertility rates still rising. Medically assisted reproductive technologies (ARTs) constitute important treatment modalities for infertile couples and represent their best chance to conceive. An important ART employed in humans and mice is ovarian stimulation (superovulation, SO), which is a drug regimen to recover large numbers of mature oocytes for generating diploid embryos. Previously, we found that SO resulted in a loss of maternal Kcnq1ot1, Snrpn and Peg3 methylation as well as a loss of paternal H19 methylation in blastocyst embryos. This was not due to an imprint acquisition error but rather an embryonic imprint maintenance error. We hypothesize that ART-produced embryos are predisposed to imprinted methylation errors because ARTs disrupt crucial maternal-effect transcripts in oocytes, which are required to maintain post-fertilization imprinted methylation. To test this hypothesis, we collected oocytes/embryos, processed RNA, made cDNA libraries and sequenced a total of 132 individual oocytes and embryos (NoART controls: 19 oocytes, 24 1-cell and 23 2-cell embryos; SO samples: 20 oocytes, 29 1-cell and 17 2-cell embryos). Comparison across developmental time course (NoART oocyte, 1-cell and 2-cell vs SO oocyte, 1-cell and 2-cell embryos) revealed striking differences in gene expression dynamics. Using a novel machine learning approach to capture complex temporal patterns in transcriptome data (splineTimeR), we identified 3,595 genes that showed significantly different expression trajectories between NoART and SO groups (FDR < 0.001). This temporal analysis showed dramatic shifts in key developmental programs, including gene expression regulation (751 genes), cell survival pathways (1,134 genes), cellular function (1,246 genes), developmental processes (1,202 genes), and growth/proliferation networks (1,190 genes). When we examined genes known to be critical for early development, we found that ART procedures altered the expression dynamics of 268 maternally expressed genes (Database of Transcriptome in Mouse Early Embryos) and 29 well-characterized maternal effect genes (recently updated and reviewed). Further examination of one candidate maternal-effect gene, Zfp57, revealed significantly lower Zfp57 mRNA levels in ART-derived embryos compared to NoART controls at the 2-cell stage. While ZFP57 translocated from the cytoplasm to the nucleus at the 4-cell stage, ART-derived embryos had delayed ZFP57 nuclear import until the 8-cell stage, which could contribute to the loss of imprinted methylation in these embryos. This systematic perturbation of maternal factors suggests that ART procedures may compromise the molecular machinery of the early embryo, including factors involved in genomic imprinting. This work was supported by NIH/NIHCD, 5R01HD101574, to MRWM.