Selecting embryos with optimal cleavage kinetics improves ongoing pregnancy rate following blastocyst transfer in a mouse model

To develop a model predicting blastocyst (blast) formation based on early cleavage parameters in the mouse and to determine if the model predicts pregnancy outcome following blast transfer. Microarray analysis was performed to identify genes associated with early embryo development. Laboratory research. 313 2-pronuclear (PN) embryos were collected from superovulated (SO) female mice and were cultured to blast in 5% (n=180) and 20% (n=133) oxygen (O2). Time-lapse videos were collected in the Eeva™ System (Progyny). Classification and regression tree analysis (CART) was used to predict blast formation based on 4 cleavage parameters: 1st cytokinesis duration, and time from 2-3 cell, 3-4 cell, and 4-5 cell stages. The model was built in the 20% O2 cohort and validated in the 5% O2 cohort. For the transfer experiments, 2PN embryos were collected following mating of SO females to transgenic male mice heterozygous for green-fluorescent protein (GFP) and cultured in the Eeva™ system. Blasts were rated as having “optimal” (Opt) or “suboptimal” (Subopt) timing based on the CART model. 10 blasts were transferred into each of 10 pseudo-pregnant recipients, using GFP status to tag the embryos as Opt or Subopt. Pregnancy outcomes were assessed at mid-gestation (E10.5). Differences in outcomes were assessed using Fisher’s exact test. Single-embryo microarray analysis was performed using the Affymetrix GeneChip Mouse Gene ST 1.0 Array on Opt and Subopt blasts. The final model utilized 2 parameters, cleavage times from 2-3 cell and 3-4 cell. The model predicted blast formation with a 97.5% sensitivity and 62.5% specificity in 5% O2. Following transfers, 8 mice became pregnant and were included for analysis. The overall implantation rate per embryo transferred did not differ significantly between groups (86% Opt vs. 77% Subopt, p=0.3). However, the ongoing pregnancy rate was higher for the Opt compared to Subopt embryos [60% (35/58) vs. 32% (7/22) (p=0.03)]. The resorption rate (implantation sites with placental tissue only) among implanted embryos was significantly higher in the Subopt compared to Opt group [59% (10/17) vs. 30% (15/50)(p=0.04)]. Gene expression differences were seen in 74 genes between the Opt and Subopt groups (FD<25%). Pathway analysis revealed enrichment for genes encoding membrane-bound proteins including amino acid and ion transporters. We have validated a mouse model of time-lapse microscopy with morphokinetic parameters that predicts blast formation and improved ongoing pregnancy rates following blast transfer. We have also shown that cleavage kinetics are correlated with gene expression changes. The data from this model suggest that time-lapse may allow selection of blasts with the highest probability of live birth.