Chromosomal mosaicism in human blastocysts: the ultimate challenge of preimplantation genetic testing?

To what extent does a trophectoderm (TE) biopsy reliably reflect the chromosomal constitution of the inner cell mass (ICM) in human blastocysts?

Concordance between TE and ICM was established in 62.1% of the embryos analysed.

Next generation sequencing (NGS) platforms have recently been optimised for preimplantation genetic testing for aneuploidies (PGT-A). However, higher sensitivity has led to an increase in reports of chromosomal mosaicism within a single TE biopsy. This has raised substantial controversy surrounding the prevalence of mosaicism in human blastocysts and the clinical implications of heterogeneity between the TE and ICM.

To define the distribution and rate of mosaicism in human blastocysts, we assessed chromosomal profiles of the ICM and multiple TE portions obtained from the same embryo. We evaluated donated embryos with an unknown chromosomal profile (n = 34), as well as PGT-A blastocysts, previously diagnosed as abnormal or mosaic (n = 24). Our intra-embryo comparison included a total of 232 samples, obtained from 58 embryos.

Four embryo samples, including the ICM and three distinct TE portions, were acquired from good quality blastocysts by micromanipulation. Whole genome amplification (WGA), followed by NGS was performed on all embryo segments. Profiles were compared between samples from the same embryo, while the results from pretested blastocysts were further correlated to the original report. The embryos investigated in our untested group were obtained from good prognosis patients (n = 25), with maternal age ranging from 23 to 39 years. For the pretested embryo group, maternal age ranged from 23 to 40 years (n = 18).

We uncover chromosomal mosaicism, involving both numerical and structural aberrations, in up to 37.9% of the blastocysts analysed. Within the untested group, the overall concordance between the ICM and all TE portions was 55.9%. A normal ICM was detected in 20.6% of blastocysts for which at least one TE portion showed a chromosomal aberration. Conversely, 17.6% of embryos presented with mosaic or uniform abnormalities within the ICM, while showing normal or mosaic TE profiles. For the pretested blastocysts, the overall concordance between the ICM and all TE samples was 70.8%. However, 50% of embryos previously diagnosed with mosaicism did not confirm the original diagnosis. Notably, 31.3% of embryos with a mosaic aberration reported in the original TE biopsy, revealed a euploid profile in the ICM and all three TE samples. Taken together, concordance between the ICM and all TE portions was established in 62.1% of blastocysts, across both embryo groups. Finally, we could not observe a significant effect of age on embryo mosaicism (P = 0.101 untested group; P = 0.7309 pretested group). Similarly, ICM and TE quality were not found to affect the occurrence of chromosomal mosaicism (P = 0.718 and P = 0.462 untested group; P = 1.000 and P = 0.2885 pretested group).

All data that support the findings of this study are available online in Vivar (http://cmgg.be/vivar) upon request.

Evaluating biological variation in some instances remains challenging. The technological limitations of sampling mitotic errors that lead to mosaicism, as well as WGA artefacts, warrant careful interpretation.

Our results highlight the complex nature of genetic (in)stability during early ontogenesis and indicate that blastocysts harbour a higher rate of chromosomal mosaicism than may have been anticipated. Moreover, our findings reveal an overall high diagnostic sensitivity and relatively low specificity in the context of PGT-A. This suggests that a considerable proportion of embryos are potentially being classified as clinically unsuitable. Ultimately, more precise quantification will benefit the clinical management of embryo mosaicism.

M.P. is supported by the Special Research Fund, Bijzonder Onderzoeksfonds (BOF01D08114). J.T. and L.D. are supported by the agency for innovation through science (131673, 141441). B.H. and this research are supported by the Special Research Fund, Bijzonder Onderzoeksfonds (BOF15/GOA/011). The authors declare no competing interests.

Not applicable.