Eefje Chavli

Chromosomal abnormalities are a major cause of implantation failure and early pregnancy loss. To improve in-vitro fertilisation (IVF) outcomes, preimplantation genetic testing for aneuploidy (PGT-A) was introduced to enable the transfer of embryos based on their chromosomal composition. In the late 2010s, PGT-A at the cleavage stage was largely abandoned because it was shown that, due to mosaicism, a blastomere biopsy is not chromosomally representative of the remaining embryo. Consequently, PGT-A shifted to the analysis of trophectoderm (TE) biopsies, and its use skyrocketed. However, little was known about the incidence of mosaicism at the blastocyst stage, the impact of mosaicism on PGT-A accuracy at this stage, and the effect of abnormal cells on embryo development. The background for this thesis is provided in Chapter 1. Chapter 2 describes that if not all blastocysts, then the majority presents some degree of chromosomal mosaicism. We were the first to disaggregate the inner cell mass (ICM) and the TE of blastocysts and perform single-cell DNA sequencing. Chromosomal mosaicism was observed in 82% of the analysed embryos. We assume that the true incidence is even higher, as only a limited number of single-cells were analysed for the embryos that were tested normal. Furthermore, 69% of the embryos contained more than one abnormality, and in most cases, each abnormality was present in a limited number of cells. Abnormalities present in a limited number of cells within the embryo remain probably undetected with bulk DNA analysis of multiple cells, which explains the lower incidence of mosaicism reported in studies using bulk DNA sequencing. The single-cell analysis revealed that structural abnormalities are as common as numerical abnormalities, suggesting that DNA damage and replication stress are likely mechanisms contributing to chromosomal mosaicism. Additionally, we show that chromosomal abnormalities can be restricted to either the TE or the ICM, indicating that new mitotic errors can occur after TE and ICM formation. In Chapter 3, we examined the implications of chromosomal mosaicism on PGT-A practices by assessing how representative a TE biopsy is for the corresponding ICM, which will form the fetus later in development. From a clinical point of view, we show that a TE biopsy generally predicts the proportion of abnormal cells within an embryo in most cases. Thus, a low mosaic or normal TE biopsy often indicates that the ICM is also low mosaic or normal. Similarly, a high mosaic or abnormal TE often indicates that the ICM is also high mosaic or abnormal. However, the abnormalities observed may differ between the two cell lineages. From a cytogenetic perspective, the results for TE and ICM were concordant in around half of the cases. In Chapter 4, we describe how the chromosomal constitution impacts developmental characteristics of embryos that were cultured until day 8. Bulk DNA sequencing was applied on the TE biopsy and the corresponding day 8 embryo, which first underwent immunostaining. The cytogenetic results were compared to the total number of cells on day 8, and to the number of epiblast and hypoblast cells, two cell lineages of significance at that developmental stage. Despite the abnormalities detected on D5, the majority of the embryos survived the extended culture until D8. However, even though they remained viable on D8, we observed characteristics of impaired development in all fully abnormal embryos, and in 41% and 18% of mosaic and normal embryos, respectively. In a clinical setting, we are limited to the genetic analysis of 5-10 TE cells, which are not always representative of the remaining embryo. We therefore evaluated whether the impact of abnormal cells within an embryo is reflected in its morphokinetics. All embryos cytogenetically tested in Chapters 2 and 4 were cultured in a time-lapse incubator, allowing us to correlate the chromosomal status with morphological features, focusing on the blastocyst expansion rate. In Chapter 2, we show that the proportion of abnormal cells is negatively correlated with the total number of cells within a blastocyst and with blastocyst expansion rate. In Chapter 4, we describe that chromosomally normal embryos show a higher average expansion rate than fully abnormal embryos, while mosaic embryos exhibit an intermediate expansion. These findings suggest that abnormal cells impact blastocyst expansion. To further evaluate the blastocyst expansion rate as a non-invasive marker for embryo viability, in Chapters 5 and 6 we retrospectively compare the expansion rate of blastocysts that resulted in biochemical or ongoing pregnancies with those that did not. In Chapter 5, manual blastocyst surface measurements resulted in curves showing that blastocysts leading to ongoing pregnancy demonstrated a faster expansion rate. As manual measurements are time consuming for daily practice, we developed an artificial intelligence (AI) tool that automatically performs blastocyst surface measurements, as described in Chapter 6. By applying this AI tool on a second cohort of blastocysts, we confirmed that blastocyst expansion rate has predictive value for clinical outcomes. Last, Chapter 7 includes the general discussion of this thesis. It discusses the most important findings, the implications for clinical practice, and future research.