SCM DNA-based PGT-A has emerged as a valuable option that does not have invasive risks and has the capability to detect aneuploidies with excellent sensitivity and reasonable specificity. Although a plethora of validation studies in clinical IVF underline SCM DNA-based PGT-A excellent performance, several reports have resulted in discordancy to known existence of mosaicism in TE, corresponding whole blastocysts, and maternal DNA contamination [21]. It was reported that SCM contributes to maternal DNA contamination [18]. This maternal DNA can cause poor concordance between DNA from SCM and TE biopsies. There are thin cytoplasmic projections called transzonal projections (TZPs) that are located inside zona [22]. TZPs from cumulus cells connect to the oocyte and are crucial for normal oocyte formation. Existing TZPs within zona pellucida could either be a potential source of cell-free DNA, or could interfere with cell-free DNA released from the blastocyst. In our study, culture medium, which was collected from the drop where empty zona pellucida were individually cultured for 24 hours, served as a testing sample to explore if zona pellucida is associated which cell-free DNA; for this reason, zona pellucida were also removed immediately following TE biopsies. The corresponding zona-free blastocysts served as the gold standard of reference in our research cohort. After test results, we eliminated zona pellucida as the source of cell-free DNA in the spent media. Therefore, any further testing of zona pellucida would be an unnecessary testing of background noise added onto NGS and would minimize the possibility of maternal contamination to the testing sample.
As diagnostic tests require both sensitivity and specificity to be as close as possible to 100%, our SCM DNA-based PGT-A results indicated high sensitivity with limited specificity. Therefore SCM DNA-based PGT-A can only be classified as a screening method, rather than a diagnostic test. In comparison to our negative control, the empty culture drop, successful DNA amplification was observed in 100% TE and WB group samples, but only 78.67% in SCM samples. The DNA amplification QC metrics indicate the DNA sample quality from SCM is suboptimal and is likely linked to degraded DNA in nature. According to current diagnostic accuracy studies, any one of the elements in Table 3 may directly or indirectly affect the overall diagnostic accuracy through prevalence of the aneuploidy. All our samples, SCM, TE and WB, were from the same corresponding blastocysts. By using samples from the corresponding blastocysts, we minimized the role of the prevalence of the aneuploidy on overall diagnostic accuracy.
It is noteworthy that overall concordance characterized by dichotomy can only distinguish between euploidy and aneuploidy blastocysts. Our TE-to-WB overall concordance rate is 94.67% which is acceptable and comparable to other studies [10, 11, 23, 24]. Thus, PGT-A with TE biopsied samples for euploidy/aneuploidy detection are accurate. At the same time, specificity rates of TE-to-WB is 71.43%. Overall concordance rates that are high may be attributed to the fact that the blastocysts were donated and had been previously PGT-A tested chromosome aberrations blastocysts. These characteristics increase the risk of positive diagnosis bias, which in turn increases the likelihood of a misled, biased conclusion. Regarding technical limitations, the poor quality of degraded DNA in SCM also increases levels of chromosomal aberrations in SCM. This bias leads to artefactual aneuploidies unless appropriate protocol is applied. NGS data analysis would have to be altered specifically for degraded DNA amplification. The interpretation of results would also have to be based on DNA with intrinsically degraded low quality in SCM. So, both positive-diagnosis biases and DNA sample qualities in SCM associated artefactual aneuploidies with inflated aneuploidy levels in SCM samples. Unlike the overall concordance characterized by dichotomy, full chromosome concordances not only enhanced assessment of ploidy status, but also offered more detailed measurements with more precise quantitative and qualitative comparisons of chromosomal aberration including mosaicism, segmental aneuploidies, and a considerable number of chaotic NGS results. Our study shows the full concordance rate of TE-to-WB is 69.33%. This rate is similar to 75% found in other studies [11]. Our SCM overall concordance rate (89.83%) is comparable to the rates have been reported [11, 12] and is higher than Ho, et al [10]. However, the full concordance rate of SCM-WB is 32.2%. These full concordance rates are remarkably lower than other study [11], but are similar to this one [12]. These lower rates are likely due to the DNA sample quality in SCM. As stated in other reports, [25–27], this discrepancy is unsurprising given the likely degraded nature of the DNA in SCM.
In theory, it is possible to achieve 100% accurate assay of full chromosome concordance between DNA in SCM/TE and the paired blastocyst. In practice, however, it is unlikely to obtain ‘true data’ for an embryo if there exists mosaicism and reciprocal aneuploidy. SCM DNA is thought to be released as a consequence of cell apoptosis or necrosis death in the blastocyst.
As we have frequently observed in our daily clinical laboratory, apoptotic cells may be either phagocytosed by neighboring cells, or expelled into the perivitelline space (Fig. 4a) or blastocoel cavity (Fig. 4b)[28, 29]. Evidently, DNA quality in SCM stands out as the most limiting factor for the diagnostic efficiency of genetic data analysis using cell free DNA. Generally, commercial WGA buffers are designed for cell/cell-free samples in small volumes. Yet, the SCM sample from the culture drop has a relatively larger volume, which requires reaction components be scaled up. This requirement not only increases costs, but also necessitates additional validation and optimization. A reduced volume (12 µl) of blastocyst culture media[16] could certainly concentrate DNA in SCM, but it also may deviate from the manufacturer’s recommendations. Reducing the volume would therefore require further validation to ensure that the blastocyst’s developmental potential is not compromised. In imperfect concordant cases, only a small proportion of reciprocal chromosomal/sub-chromosomal gain-loss complementary pairs were observed in the current study. Biologically, alternative rigorous criteria to classify embryos as mosaic would require the presence of reciprocal aneuploidy in two different cells/cell lines or two different biopsied samples from the same embryo [30].
The frequency of whole chromosome or sub-chromosomal arm gains and losses should be similar. For example, take one biopsy that displayed a monosomy (2n-1) of a specific chromosome and another biopsy from the same embryo displaying trisomy (2n + 1) for the same chromosome [25, 31]. It is expected that at least one embryo should display reciprocal errors for at least one chromosome. Indeed, in our study, we observed 11 reciprocal chromosome arm loss-gain complementary pairs located in four different chromosomes in six blastocysts, which is similar to other observations [8, 18] exhibiting shared complementary aneuploidies. Coincidentally, in current study, all four chromosomes associated with reciprocal gain-loss complementary pairs were low GC-content chromosomes. Studies on human-inherited diseases and cancers also revealed that DNA breakpoints tend to occur in DNA sequences with low GC content [20]. Due to small sample size, we cannot robustly ascertain whether a low GC-content sequence would be more vulnerable to DNA breakpoints than a high GC-content sequence. However, utilizing GC content and reciprocal sub-chromosomal arm gain-loss complementary as a reference may prove a more efficient tool in distinguishing real DNA segments from NGS data noise generated from a sample of DNA in SCM [19, 32] .
Contrarily, our data also shows that the chromosomal aneuploidies were mostly consistent rather than reciprocal. In human and nonhuman primate preimplantation embryos, mis-segregated chromosome was sequestrated into a small nucleus-like structure adjacent to but outside of the primary nucleus [33, 34]. Consequently, chromothripsis may trigger the chromosome inside micronuclei to shatter and the sequential reassembly of fragments through breakage-fusion-bridge cycles, aberrant epigenetic regulation, abortive apoptosis, and other yet unknown mechanisms. Thus, loss and gain reciprocal chromosomal/ chromosome arms may exist as individual chromosomes or segments instead of complementary pairs. This discovery may denote a mechanism that explains why we can only detect small numbers of reciprocal chromosome arm gains and losses. Micronuclei must be regarded as a unique source of unstable genomes, damaged DNA of genetic variations, and possibly, the source of cell-free DNA in the SCM as well.