UNDERSTANDING IN EARLY STAGES OF EMBRYOS CHROMOSOMAL ABNORMALITIES MUCH BETTER
By Ilya Volodyaev, PhD., et al., (Anna Ivanova, Elena Korchivaia, Ilya Mazunin), Senior Researcher, Laboratory of Developmental Biophysics, Department of Embryology, Faculty of Biology, Moscow State University, Russia; and Russian national representative (basic science) to ESHRE. He can be contacted though The Reproductive Times or directly at ivolodyaev@gmail.com or ivolodyaev@emcmos.ru.
Revised: September 23rd, 2024
Especially in human embryos, chromosomal abnormalities in early embryonic development are a major concern, albeit a routine finding. They, however, can also be found in some other mammals. They can involve all cells of an embryo or be mosaic, and they can affect one or more complete chromosomes or be segmental, affecting only small pieces of chromosomes (Figure 1). In testing embryos for such chromosomal abnormalities, the technology matters. Some reported abnormalities are definitely artifacts of current testing technologies, but their exact share and effects on the accuracy of testing – in the fertility field called preimplantation genetic testing for aneuploidy or PGT-A – are still largely unknown. In this article, we address the distinction between true biological aneuploidies and false-positive diagnoses based on technical errors, and also discuss the prevalence and dynamics of chromosomal abnormalities and the role of embryo self-correction.
Figure 1. Types of chromosomal aberrations and their representations in an NGS profile.
WHOLE-CHROMOSOMAL ABNORMALITIES, A KNOWN ENTITY
Whole-chromosomal abnormalities (WCAs) in human oocytes and embryos are well-documented. A mother’s age plays a critical role, with women under age 35 experiencing WCAs in approximately 20–30% of eggs, with an additional ca. 6% increase in aneuploidy for every year after that (Figure 2). These abnormalities are observed both in natural and stimulated cycles (1) and remain consistent across different studies. Interestingly, sperm WCAs show age independence, and affect only about 3% of sperm cells (2).
Figure 2. Proportion of euploid human embryos as a function of female’s age (autologous oocytes). Note the linear trend in the advanced maternal age. Data from (5).
Whole-chromosomal abnormalities, particularly in human embryos, are largely attributed to the instability of the oocyte’s meiotic spindle, which lacks microtubule-organizing centers and certain molecular motors, and thus appears much more error-prone than in most other species (3,4). Interestingly, besides humans, such intriguing properties of meiotic spindle have been documented across several other species, namely pigs, horses, and primates (5-7).
SEGMENTAL ANEUPLOIDIES: VARIABILITY AND CHALLENGES
Unlike WCAs, segmental aneuploidies (SAs) display significant variability in reported prevalence, ranging from less than 3% in some studies (1) to as high as 30% or more (8,9). Although these estimates vary, one consistent finding is that SAs are independent of age and WCAs. These abnormalities are often platform-dependent, meaning their detection rates fluctuate with the technology used (10). For instance, studies reporting fewer SAs tend to have more concordant results in repeated biopsies, suggesting that many SAs reported in other studies may be artifacts.
The reported incidence of SAs is around 3–10% in oocytes, 5–20% in sperm, 15–25% in cleavage-stage embryos, and 7–15% in blastocysts (8,11,12). SAs are associated with relatively poorer development of embryos, which, moreover, more frequently arrest than euploid embryos (13). However, the true significance and nature of an SA diagnoses in an embryo is still debated and may often be associated with technical artifacts. The frequency of misdiagnoses of SAs is at least one potential reason why among all chromosomal abnormalities, embryos with SAs, if nevertheless transferred - demonstrated the best pregnancy and live birth rates among transfers of embryos with any chromosomal abnormality.
MOSAICISM: A CONTROVERSIAL DIAGNOSIS
Whole-chromosomal mosaicism (WCM), like SAs, exhibits a wide range of reported rates — from less than 3% (14,15) to over 30% (16,17) and is age- and WCA-independent. It apparently is the consequence of a range of phenomena, from real biological issues to technical artifacts and misinterpretations. Their true distribution still remains to be determined and is among the most hotly debated issues in reproductive medicine.
Some authors believe that most reported cases of mosaicism are artifacts due to variability in testing platforms and yet unresolved issues of single-cell DNA testing (18-20). Several among them, moreover, then argue that true mosaicism likely affects only about 5–15% of blastocysts. Based on single-cell studies (21), others consider the real mosaicism rate to stand at ca. 80% of all blastocysts. Whatever one wishes to believe, to minimize false diagnoses, PGT-A platforms must be optimized as much as possible and the seemingly amazing plasticity of embryos and, therefore, their ability to self-correct must be further explored,
One must also remember that, while the correct definition of mosaicism is the presence of two or more cell lines with different karyotypes in a whole organism (in this case, a blastocyst), typical PGT-A reports refer to mosaicism only based on a 5—10-cell biopsy of trophectoderm. This incorrect definition of mosaicism by PGT-A laboratories will, therefore, only reflect a small fraction of the actual mosaic aneuploidy in the tested human embryo. Consequently, the diagnosis of a full aneuploidy will be incorrect if only a small island of aneuploid cells was at random biopsied (while the embryo contains both aneuploid cells and euploid cells). Similarly, an embryo may be incorrectly identified as euploid if by chance only euploid cells have been biopsied, while surrounding tissues contain aneuploid cell islands. Underdiagnosis of mosaicism is also supported by ca. 80% of blastocyst-stage embryos in single cell analyses demonstrating aneuploid cells (21).
With ca. 20% of all embryos affected by meiotic aneuploidy (affecting all cells in an embryo) and all other aneuploidies being mitotic (and, therefore, present in islands of cells) - if ca. 80% of all embryos demonstrate at least some aneuploid cells by single cell analysis - ca. 60% of all embryos must be to different degrees mosaic.
The diagnosis of mosaicism is also affected by PGT-a laboratories using arbitrary thresholds in defining mosaicism: An embryos in many PGT-A laboratories in the U.S and Europe can be defined as euploid with <20%, <40%, or even <50% aneuploid DNA in a trophectoderm biopsy of 5- 6 cells. Which of these definitions a laboratory uses to define euploidy will, of course determine the percentage of mosaic embryos because they will represent either the range of 20-80%, 40-80%, or even 50-80%. The cut-off for aneuploidy, in contrast, has remained consistent at 80%. It is, however, important to remember that under the correct biological definition of “mosaicism,” any presence of a second cell lineage in the organism, from 0.1% to 99.9% in theory represents “mosaicism.”
Consequently, most embryos currently signed out by PGT-A laboratories are biologically mosaic and many embryos signed out as aneuploid also are in reality mosaic. This, however, does not negate that percentages of aneuploid cells in a mosaic embryo matter because we know from animal models that the ability to self-correct declines with increasing percentages of aneuploid cells in an embryo (22).
That mosaic embryos more likely arrest than euploid embryos (13), may also be a reason for underestimation of mosaicism at the blastocyst stage. Mitotic errors occur mainly during the first and second divisions of the zygote. The proportion of embryos at the cleavage stage affected by mosaicism- including possible artefacts - is, therefore, higher than at the blastocyst stage.
All of this, paradoxically, however, also does not exclude overdiagnosis of mosaicism when “noisy NGS profiles” are overinterpreted as mosaic instead of being marked as “unreadable.” This, of course, can also contribute to laboratory-specific differences in prevalence of mosaicism diagnoses in the same patient populations (19). We below offer a short introduction to the potential mathematical estimation of chromosomal artifacts in PGT-A technologies.
UNDERSTANDING THE MATHEMATICS BEHIND CHROMOSOMAL ARTIFACTS
When it comes to assessing chromosomal abnormalities, whether whole-chromosomal or segmental, a critical challenge is differentiating true biological events from technical artifacts introduced during testing. Technical artifacts have attracted too little attention in the medical literature. Understanding how these artifacts arise requires not only biological insight but also a mathematical perspective. Below, we provide a brief overview of the formulas used to estimate the probability of artifacts in whole-genome amplification (WGA) technologies, such as Next-Generation Sequencing (NGS), and Fluorescence In Situ Hybridization (FISH).
Artifact Probability in Whole-Genome Amplification Technologies
In WGA technologies, artifacts arise from such factors as DNA degradation or “loss” and non-uniform amplification. The probability of encountering such artifacts can be mathematically estimated using the following formula:
where: a is probability of degradation of 1 DNA copy; pk, probability of k DNA copies degraded; N, number of DNA copies in the sample; n, number of DNA copies required for a stable signal; G, number of “independent genome segments” (which is a conditional division of the genome into fragments to account for their independent degradation processes).
This equation illustrates how the probability of artifacts decreases as the number of DNA copies in the sample increases. Simply put, the more DNA is present for analysis, the more reliable will the results be because of fewer artifacts being introduced during the testing process.
Artifact Probability in FISH
For FISH, which is used to detect specific chromosomes, artifacts arise from hybridization failures, chromosome overlap, and “false signals”. The probability of encountering an artifact in FISH can be approximated using the following formula:
where: p is probability of false signals; B, probability of hybridization failure for a single chromosomal type; s, average relative area occupied by one chromosome at FISH on which the next chromosome must not overlap; m, number of chromosomes labeled in one cell; n, number of chromosomal types labeled in one cell; N, number of cells in the biopsy sample. Note that this formula is approximately true for small n and becomes false for large n.
Unlike WGA, the probability of artifacts in FISH increases with the number of cells in the sample. This fundamental difference makes FISH and WGA assessments of chromosomal abnormalities inherently incomparable. Even when the two methods yield similar results, variations in the number of cells tested (unavoidable in real practice) will cause significant discrepancies.
WHAT THESE FORMULAS TELL US
The differences in the mathematical estimations of artifacts between WGA and FISH underscore the complexities of detecting chromosomal abnormalities accurately. While WGA improves reliability with more DNA copies, FISH becomes less reliable as the sample size increases. Although this may seem like a currently unimportant issue considering FISH is basically no longer used in PGT-A, one faces these discrepancies every time when PGT-A data from different stages of embryo development are compared. The PGT 1.0 to PGT 2.0 shift included changes of technology and of stage of embryo biopsy. A large majority of PGT-tested cleavage-stage embryos underwent FISH testing, while blastocysts have been mostly subjected to aCGH or NGS.
CAN EMBRYOS SELF-CORRECT?
One of the most intriguing and still controversial topics in embryonic development is the promising evidence that mammalian embryos, as well as human embryonic stem cell models and cultures can self-correct (21-23). Thousands of healthy births have followed the transfer of mosaic embryos, challenging our understanding of embryo biology (21-27). The significantly lower rates of mosaicism among fetuses and children suggest that either current testing methods are misidentifying euploid embryos as mosaics, or that some embryos have the ability to correct chromosomal errors during development (20-22,25,28-30). The first question is more technical and seemingly uninteresting, while the second option involves deep insights into embryonic development and the true beauty of regulatory processes and self-organization. However, it is the first “uninteresting” technical issue that must be scrupulously optimized not to be trapped by above noted technical artifacts.
The self-correction issue has recently gotten strong support from a study by Danish investigators (31) who reported ESC-derived mouse primitive endoderm lineage (producing the extraembryonic endoderm supporting embryo development but not part of the fetus) to have so much plasticity and potency that it can regenerate a complete blastocyst which continues post-implantation development. Though this, of course, does not imply genetic self-correction and does not explain how aneuploid cells are removed, it appears reasonable to assume that an epiblast that can regenerate the extraembryonic cell lineage, also can replace its own chromosomal abnormal cells (interestingly, the extraembryonic cell lineage in mouse (20) and human embryos (21) does not appear to have the same plasticity and ability to self-correct). A better understanding of the embryo’s ability to self-correct has, however, still to be pursued.
CONCLUSION
While whole-chromosomal aneuploidies are relatively well-understood, segmental aneuploidies and mosaicism remain more open to discussion, largely due to the limitations and artifacts of the current testing technologies. The role of mosaicism in embryo development, and the possibility of embryo self-correction represent critical questions in the field of reproductive medicine that warrant further examination and changes in current reporting schemes in PGT-A.
After the very recent statement of the Practice Committees of the American Society for Reproductive Medicine (ASRM) and the Society for Assisted Reproductive Technology (SART), noting absence of any outcome utility for the PGT-A procedure for in vitro fertilization (IVF) (32), the future of PGT-A appears at a crossroad. If PGT-A is to remain in continued use in association with IVF, a clear outcome utility for the procedure must be found and current laboratory methodologies must be further improved. In practical terms this means a better understanding of technical artifacts, a significant decrease in overdiagnoses, potentially the addition of new types of diagnosable anomalies due to more in-depth diagnostic techniques and technologies (33), and improvements in the ability to separate embryo plasticity and self-organization from technical artifacts.
References
1. Labarta et al.,. J Clin Endocrinol Metab 2012;97:E1987-E1994
2. Bell et al.,. Nat 2020;583:259-264.
3. Holubcova, et al., Science 2015;348:1143-11477.
4. So et al., Science 2022;375.
5. Hornak et al., Plos One 2012;7:e30335.
6. Hornak et al., CytogenGenome Res 2016;150:60-70.
7. Dupont, et al., Fertil Steril 2010;93:2545-520.
8. Babariya et al., Hum Reprod 2017;32:2549-2560.
9. Navratil et al.,Molec Hum Reprod 2020;26:269-76.
10. Treff NR, Franasiak JM. Fertil Steril 2017;107:27-31.
11. Treff NR, Marin DT. Fertil Steril 2021;116:1205-11.
12. Magli eet al.,Hum Reprod 2020;35:759-69.
13 McCoy et al., Genome Med 2023;15(1):77
14. Katz-Jaffe et al., FertilSteril 2017;108:e87-e8.
15. CCRM 2017. . CCRM Reports the Lowest Chromosomal Mosaicism Rates in the U.S. At Less Than Three Percent Across the CCRM Network. In: Peel J, ed. CCRM Fertility, 2017.
16. Escudero et al., Fertil Steril 2016;105:e20-e21.
17. Munne et al., Am J Hum Genet 1994;55:150-159.
18. Treff et al., Hum Reprod 2010;16:583-589.
19. Kullyev et al., In: Preimplantation Genetic Diagnosis International Society PGDIS 2023. Paris, France, 2023
20. Swain EJ. Reprod BioMed Online 2019;39:599-607
21. Yang et al. Nat Cell Biology 2021;23:314-321.
22. Bolton et al., Nat Commun 2016;29(7):11165
23. Viotti et al. Fertil Steril 2023;120(5):957-966
24. Barad et al., Hum Reprod 2022;37(6):1194-1206
25. Coticchio et al., Hum Reprod Update 2021;27:848-865.
26. Munne et al., Eur J Med Genet 2020;63:103741.
27. Yakovlev et al., Reprod Biomed Online 2022;45:88-100.
28. Capalbo et al., Hum Reprod 2017;32:492-8.
29. Esfandiari et al., J Assist Reprod Genet 2016;33:1439-1444.
30. Popovic et al., , Hum Reprod Update 2020;26:313-334.
31. Linneberg-Agerholm et al.,Cell 2024;S0092-8674(24)00595-
32. Practice Committee of the ASRM and SART. Fertil Steril 2024; 122(3);421-434
33. Essers et al., Nat Med 2023;29:3233-3242.