Title Page Title: Chromosomal Concordance between the Baby Produced by Preimplantation Genetic Testing Embryo and the Trophectoderm Biopsy: a Retrospective Study

Background: Chromosomal mosaicism and aneuploidies are routine phenomena throughout human preand post-implantation development. The benefit of implanting mosaicism or aneuploidies is still controversial. The purposes of the study are to investigate the developmental potential of embryos with chromosomally segmental or mosaic abnormalities, and whether precise Next Generation Sequencing (NGS) resolution would reduce the development of an abnormal embryo in preimplantation genetic testing (PGT) cycles. Methods: The peripheral blood of 17 PGT babies were collected for single nucleotide polymorphism (SNP) array and were compared with trophectoderm (TE) biopsy results at different NGS resolutions. Results: 76.5% (13/17) of babies’ peripheral blood chromosome analysis was consistent with 10Mb TE biopsies and 58.8% (10/17) of babies’ analysis was consistent with 4Mb TE biopsies. 2 babies who had euploid TE showed abnormal peripheral blood chromosome analysis. 17.6% (3/17) embryos with aberrant TE biopsies produced healthy babies. Although the sensitivity of 10Mb was lower than 4Mb (25% vs. 50%), the specificity (100% vs. 76.9%), PPV (100% vs. 40%) and diagnostic accuracy (82.4% vs. 70.6%) of 10Mb showed better results than 4Mb. Conclusion(s): The chromosomal results between peripheral blood samples and TE biopsies of born babies are not completely congruent. Aneuploid and mosaic embryos have potential to produce healthy babies, whereas normal embryos also have chance to produce babies with chromosomal abnormalities. In spite of low sensitivity of both resolutions, 10Mb has higher specificity, PPV and diagnostic accuracy than 4Mb. It is suggested that TE biopsy be analyzed in both 10Mb and 4Mb resolutions to uncover severely adverse chromosomal aberrations but use 10Mb resolution to guide transfer. Trial registration: The study was retrospectively registered in the Chinese Clinical Trial Registry (ChiCTR2100042522).


Background
Since the first preimplantation genetic testing (PGT) baby was born in 1989 (1), more patients with chromosomal diseases, monogenetic diseases, or even recurrent miscarriages have opportunities to have their own babies. However, not all patients will obtain normal embryos through PGT cycles. Sometimes, after comprehensive considerations and adequately counseling with patients, embryos with partial genetic problems have been transferred into uterus. Strange and interesting phenomena during the followup that those embryos have potential to produce healthy children were reported (2)(3)(4). Of note, TE biopsies were used for PGT in those studies for not damaging embryos, which were not direct to reflect the inner cell mass (ICM). These discoveries lead us to think about a question: is PGT method using TE biopsy accurate enough to predict the developmental potential of an embryo? To date, there is no consensus whether it is beneficial to implant embryos with chromosomal abnormalities in the clinic.
Furthermore, most studies focused on the chromosomal pattern of the babies born by implanting mosaic or aneuploid embryos, not euploid embryos (2,4). There were few studies on chromosomal copy number variations (CNVs) of born babies after implanting euploid embryos. Thus, it is also worth discussing the CNV concordance between normal PGT embryos and born babies.
PGT techniques have evolved quickly, from fluorescence in situ hybridization (FISH), to array comparative genomic hybridization(aCGH), single nucleotide polymorphism (SNP) array, and more recently, next generation sequencing (NGS) (5). Among them, NGS is an emerging technology that provides unprecedented high-throughput, highly parallel, and base-pair resolution data for genetic analysis (6). Obviously, it shows better application than other technologies with superior sensitivity and higher precision for chromosomal analysis, and now it is possible to detect chromosomal mosaicism at levels between 20%-80%(4, 7-10). However, the benefits of such a high-resolution technique that detects more embryos with chromosomal aberrations which perhaps result in the discard of some potentially viable blastocytes remains unclear.
We hypothesized there possibly existed allegedly euploid embryos in rough NGS resolution which were implanted and gave birth to babies in our department. Moreover, we hypothesized that finer NGS resolution would reduce the chance of some mosaic and aneuploid embryos to develop healthy babies.
Thus, the present retrospective study presented and estimated the chromosomal concordance between peripheral blood samples of born PGT babies and NGS resolutions in 4Mb and 10Mb to explore the accuracy of TE biopsy and whether precise NGS resolution should be applied into the clinical practice.

Study Design and Ethical Permission
The study was approved by the Institutional Review Board, Department of Reproductive Center, Xiangya Hospital, Central South University (No.202006) and registered in the Chinese Clinical Trial Registry (ChiCTR2100042522). In our department, 10Mb NGS resolution was conducted to analyze TE biopsy for PGT before 2019, but now the NGS resolution could reach 4Mb or even more precise. Not regarding normal embryos or abnormal embryos implanted, a comprehensive search was conducted in the database of from January 2017 to December 2018. The inclusion criteria: (1) couples were treated by PGT; (2) couples gave birth to a baby. All couples that met the criteria were called to ask about their wills to participate in the study. After receiving their approval and written informed consent forms, the preserved DNA amplified samples of their implanted embryos were sent to retested by NGS at 4Mb resolution and their children's peripherical blood samples were collected for SNP array analysis to confirm CNVs.

Embryo Culture and Biopsy
For patients required with PGT, controlled ovarian stimulation was performed according to conventional protocols. Transvaginal oocyte retrieval was then scheduled for 36 hours later after hCG was administrated to induce final oocyte maturation. Embryos were all fertilized by intracytoplasmic sperm injection (ICSI) in order to eliminate the risk of contamination from sperm DNA, and then cultured in cleavage embryo culture medium (COOK, K-SICM-50) and subsequently blastocyte culture medium (COOK, K-SIBM-50). Until Day 5 or Day 6 of development, blastocytes were scored based on the Gardner system(11) and then a single laser pulse using Saturn™ Lasers Systems (CooperSurgical Inc., USA) was carefully applied in assistance to blastocoel hatching, providing a safe distance from the ICM. 5-8 cells would be sucked for TE biopsy was performed in the expanded blastocytes after 5-6 hours, and all biopsied embryos were then vitrified.

Frozen embryo transfer
A selected embryo was transferred in the next frozen cycle after careful considerations by reproductive clinicians and embryologists until the NGS results came out. Blastocyst vitrification and warming methods were described as before (12). During the frozen embryo transfer (FET) cycle, hormone replacement therapy (HRT) or natural cycle (NC) were chosen for personalized treatment depending on patients. In HRT-FET cycle, estrogen pills were used to prepare endometrium for approximately 12 days.
Then progesterone pills were administered for endometrial transformation. In NC-FET, clinicians monitored development of the dominant follicle until it was discharged and the progesterone pills were applied for corpus luteum support.

Sample Processing and Testing
All trophectoderm (TE) samples used in the study had been processed by multiple annealing and loopingbased amplification cycles. After approval, the saved TE amplified samples belonging to participants were sent for NGS to be retested at resolutions of 4Mb. Additionally, the peripheral blood samples were collected from participants for SNP array. CNVs was detected in both kinds of samples. And all testing was conducted by Yikon Genomics according to the methods as they previously described (13)(14)(15). 30%-70% abnormal cells were classified as mosaicism. Otherwise, euploidy or aneuploidy was reported. For 10Mb resolution, 1M reads were obtained from one reaction in high-throughput sequencing, the sequencing length was approximately 50bp, and the depth was around 0.025X. As to 4Mb resolution, 1.5M reads could be obtained in one reaction, the sequencing length was around 50bp, and the depth reached 0.025X-0.040X.

NGS Protocol Validation
Before initiation of our study, the validation of NGS protocol was performed for accuracy. 3 Preimplantation Genetic Testing -Aneuploidy Kits (Semiconductor sequencing) (Yikon, H170601, H170602, H170603) were used in 36 positive reference materials (Yikon, H170301) for library construction. Then libraries were sequenced with a DA8600 sequencer. Every sample was tested 3 times. negative control: 46, XN; blank control) were required to be detected when all reference samples were detected. The total library construction failure rate should not be more than 3%, the valid data should not be lower than 1Mb and the genomic coverage should be not lower than 4%. All tested PGT-A kits met the accuracy standards and could be used for clinical sample sequencing (data were provided by Yikon Genomics and not shown here).

Statistical Analysis
GraphPad Prism 8 was applied for statistical analysis and graph preparation.

Trophectoderm Biopsy and Peripheral Blood Chromosome Analysis of Born Babies
After research and inquiry, 17 families were recruited into the study. The results were described in the

Sensitivity and specificity of 10Mb and 4Mb resolution
For the purpose of assessing diagnostic accuracy of NGS in the 10Mb and 4Mb resolutions, sensitivity, specificity, PPV, NPV and diagnostic accuracy was calculated (Table 2). We analyzed them by classifying concordance of the TE and peripheral blood samples as true positive (abnormal TE, abnormal blood sample), true negative (normal TE, abnormal blood sample), false negative (normal TE, abnormal blood sample), and false positive (abnormal TE, normal blood sample). Based on this estimate, sensitivity (i.e., the probability of an abnormal blood sample being diagnosed from a TE biopsy) of 10Mb and 4Mb resolution was low, but specificity (i.e., the probability of a normal blood sample being diagnosed form a TE biopsy), PPV and diagnostic accuracy of 10Mb resolution were higher than 4Mb. No statistical significance was found in sensitivity and specificity between 4Mb and 10Mb resolutions. Furthermore, the consistency between peripheral blood samples and TE biopsies were quite weak (Kappa value≤0.4).

Discussion
First of all, our results manifested that only 58.8% and 76.5% of babies had congruent peripheral chromosomal results with TE biopsies in 4Mb and 10Mb resolutions, respectively. 11.8% (2/17) of babies whose TE biopsies were euploid in both 4Mb and 10Mb resolutions showed abnormal A number of previous studies adopted TE biopsies as a suitable source for PGT (10,16,17). However, the accuracy of TE biopsy for PGT to determine an euploid embryo is still a controversial issue. It is because TE biopsy is not a direct way to reflect ICM and just obtains 5-10 TE cells for subsequently diagnosing an embryo. For example, if certain abnormal chromosomal disjunction occurred in the TE cells, the TE biopsy might diagnose an euploid ICM as mosaicism or aneuploidy. It was reported that the concordance between TE and ICM was established in 62.1% of the embryos analyzed (8). And in Victor's research, there were 5 out of 100 aneuploidy embryos which had euploid ICM (7). In addition, the limited number of biopsied cells from a blastocyte that contains a few hundred TE cells makes the rate of mosaicism and aneuploidies vary widely (3,18).
Another important reason for the inaccuracy of the TE biopsy is that chromosomal mosaicism, the presence of two or more distinct cell populations, and aneuploidies are prevalent throughout human pre-and post-implantation development (19,20). Many factors conduce to the phenomenon, such as controlled ovarian hyperstimulation, embryo culture condition and so on (19). Investigated by Babariya and colleagues, they considered that during the first few mitotic divisions following fertilization, control of the cell cycle was more relaxed, permitting DNA double strand breaks to occur and persist through cell division (21). Those aberrations could be rescued after further embryo development by different pathways: anaphase-lag, nondisjunction, or chromosome demolition correction or clonal depletion correction (19,22). Anaphase-lag correction describes that the process of anaphase lagging reverts the reciprocal chromosomes back to disomy. Nondisjunction correction describes that the three chromosomes will be divided into 2 daughter cells unevenly, one with 2 chromosomes and the other with 4 chromosomes. Then the tetrasomic cell will be depleted by apoptosis. Chromosome demolition correction indicates that aneuploid cells can give rise to two diploid cells by deliberate fragmentation of one of the three chromosomes during metaphase or anaphase. Clonal depletion correction depicts that aneuploid cells will be depleted by apoptosis or reduced proliferation. However, the potential of embryo to develop to a baby largely depends on the cell lineage and sufficient euploid cells. In Bolton et al.'s mouse model, they revealed that aneuploid cells in the fetal linage were eliminated by apoptosis, whereas those in the placental lineage showed severe proliferative defects (23). And they also showed that mosaic embryos have full developmental potential when they contained sufficient normal cells (23).
Laboratory error may also contribute to the inaccuracy. Assuming a normal euploid cell might have been biopsied, the laboratory assessment might be inaccurate and provide a misdiagnosis of aneuploidy (24). So, then some embryos with potential development are inadvertently being classified as unviable and lost chance to produce a healthy baby. There is no doubt that it is a big hit to poorprognosis patients who have only one embryo to transfer.
Despite the small patient numbers, the results we yielded here were consistent with another 3 aforementioned studies (2)(3)(4) in the regard that mosaic or aneuploid embryo could have potential to give birth to a healthy baby (No. 5,6,7 and 8). To the best of our knowledge, there is no report on chromosomal results of PGT babies after transferring euploid embryos hitherto. But if the accuracy of TE biopsy is doubtful, the chromosomal analysis of those babies produced by euploid PGT embryos is worthwhile to concern. Therefore, it is the first report that presumed euploid embryos could possibly produce babies with chromosomal abnormalities.
Although implanting mosaic or aneuploid embryos reduces the pregnancy rate (25), it could be a try for those patients with only an aneuploid embryo according to our investigation. Some researchers also commended that IVF clinicians could transfer abnormal embryos according the type of involved chromosomes and the degree of mosaicism of human embryo (3,18,19). Like, trisomies 2,3, 7 and 8 are the most frequently detected mosaic autosomal trisomies on chorionic villus sampling (CVS) and normally lead to a chromosomally normal fetus (26). Our data found segmental aneuploid chromosomes 2, 7, 16, 19, 20, X could produce healthy babies, especially for chromosomes 16  Additionally, we also calculated the sensitivity, specificity, PPV, NPV and accuracy of 10Mb and 4Mb resolutions, based on the chromosomal results of born babies as "gold standard". Although the sensitivity is low (25%), 10Mb had higher specificity (100%), PPV (100%) and diagnostic accuracy (82.4%) compared to the 4Mb. As we all know, 4Mb have higher resolution and sensitivity than 10Mb.
However, such a high resolution screens out some abnormal embryos with subsequent developmental potential, making them discarded and lose chance to be implanted. Nonetheless, a relatively rough resolution (10Mb) assured the detection capability of confirming newborns with normal chromosomal profiles, that is, providing an opportunity to those slightly abnormal embryos for self-correction. In 2018, a study pointed out the ability to detect a segmental aneuploidy of <10 Mb is not guaranteed and multiple other factors must be taken into account when using the technology for an inherited segmental change (28). The low Kappa values that reflected the predictive value for chromosomal profiles between TE biopsies and born baby confirmed the possible self-correction mechanisms during the development of embryos. But it is widely acknowleged that TE biopsy is less invasive and much safer for PGT. We should not deny its superiority in PGT and need to find a way to improve its application.
In our experience, we suggest that we should analyze TE biopsy in both 10Mb and 4Mb resolutions to uncover severely adverse chromosomal aberrations but use 10Mb resolution to guide transfer.
Although our study firstly pointed out the inconsistency among TE biopsy of different NGS resolutions and peripheral blood samples of born babies, some inevitable limitations need to be acknowledged and discussed here. Firstly, it is a retrospective and observational study with a quite small sample size.
Secondly, we just compared TE biopsy in 10Mb and 4Mb resolutions. More precise or rougher resolution is supposed to be investigated in the further studies. Finally, it is worth noting that NGS and SNP array in our investigation are two distinct techniques using different platforms, which might result in analysis bias. For the data we and other teams provided(2-4), a research with large sample number is required in the future.

Conclusions
Our data demonstrated that the chromosomal results between peripheral blood samples and TE biopsies of born babies were not completely congruent. Aneuploid and mosaic embryos had potential to produce healthy babies, whereas normal embryos also had chance to produce babies with chromosomal abnormality. In spite of low sensitivity of both resolutions, 10Mb had higher specificity, positive predictive value and diagnostic accuracy than 4Mb. We also suggested that TE biopsy be analyzed in both 10Mb and 4Mb resolutions to uncover severely adverse chromosomal aberrations but use 10Mb resolution to guide transfer.   Abbreviations: PPV, positive predictive value; NPV, negative predictive value; TE, trophectoderm; NGS, next generation sequencing.