Authors: Whitney JB, CM Rios, K Balloch, RE Anderson, MC Schiewe

Study question: Can trophectoderm biopsy technique induce mosaicism using a technician-induced, manual cell lysis research model?

Summary answer: Prospective induction of cellular damage on research aneuploid embryos revealed a positive correlation to mosaicism, suggesting that mosaicism can be a laboratory created artifact.

What is known already: NextGenertion sequencing (NGS) has enhanced the reporting of mosaicism. The importance of mosaic profiles on PGS results is not well defined, but viable pregnancies have been produced. As increased mosaicism has been associated with decreased implantation potential, it is relatively unknown how accurately mosaic profiles predict the final ploidy of the fetus. It has been proposed that mosaicism may be linked to biopsy technique, but prospective attempts have not been performed. With the desire to offer a transfer for all patients and encouragement of mosaic embryo transfers, the reproducibility and accuracy of mosaic reports per entire embryo is not well studied.

Study design, size, duration: Nine research consented embryos were thawed as part of a sister concordance study. Upon completion of two standard biopsy procedures for the original study, the technician attempted to perform poor biopsies with emphasis on cellular lysis and increased laser ablations. A total of 50 samples were tubed, 9-original biopsy, 26-sister concordance study, 10-poor technique, 5-remaining ICM’s and 2-media blanks. All samples were blindly analyzed (NGS) by two geneticists and reported according to lab standards.

Participants/materials, setting, methods: Research consented aneuploid embryos were selected based on single/dual chromosomal aneuploidies without knowledge of original NGS profiles. Vitrified blastocysts were warmed and cultured in LifeGlobal medium under tri-gas, humidified atmospheric conditions at 37°C (≤4 hours). Each blastocyst was biopsied according to standard protocols and video recorded. Samples were blindly assessed using the MiSeq (Illumia) NGS platform. Standard lab reporting procedures were applied with a threshold of 30% gain/loss being considered euploid and 30-80% considered mosaic.

Main results and the role of chance: Evaluating full gains and losses, 49 of 50 samples (98%) achieved concordance. Of the 10 samples with technician induced lysis, 3 of 10 (30%) reported the full aneuploidy and an additional mosaic profile. The video review of biopsies showed that as cellular lysis increased so did the level of genetic noise and presence of mosaicism. Of the 2 biopsies which had >30 additional zona ablations, mosaicism was not observed but unique low-level noise was observed being strikingly different from all other samples profiles from the same embryo. Most importantly, spent media from a biopsy which had significant lysis was tubed as a “blank”. The latter blank tube showed amplification portraying a chaotic profile, whereas spent biopsy from a clean biopsy (second blank control) exhibited no amplification. The concordance arm of this study, including ICM analysis, showed extreme repeatability. The efforts made to induce mosaic profiles was successful. Of the 3 samples showing mosaicism, none of the other samples from the embryo, including the ICM, showed any of the same mosaic aneuploidies.

Limitations, reasons for caution: Mosaicism, and its importance on NGS calling policies, ultimately impact patient transfer decisions. Mosaic embryos graded as aneuploid can produce healthy live births, revealing a potential technical artifact. Upmost importance should focus on proper biopsy technique if we are to better understand and reduce possible sources of mosaicism.

Wider implications of the findings: Mosaic profiles can be made from known cell lines within genetic laboratories. Yet, mosaic profiles themselves are not in question, but instead the possible loss of DNA by cellular lysis as a source of low level mosaicism. Further scrutiny on laboratory procedures is needed to properly identify true mosaic embryos.