- Email: [email protected]
Frequencies of chromosome-specific mosaicisms in trophoectoderm biopsies detected by next-generation sequencing
ORIGINAL ARTICLE: GENETICS
Frequencies of chromosome-specific mosaicisms in trophoectoderm biopsies detected by next-generation sequencing Gary Nakhuda, M.D.,a,b Chen Jing, M.S.,a Rachel Butler, M.S.,a,b Colleen Guimond, M.S.,a,b Jason Hitkari, M.D.,a,b Elizabeth Taylor, M.D.,a,b Niamh Tallon, M.D.,a,b and Albert Yuzpe, M.D.a,b a
Olive Fertility Centre; and b University of British Columbia, Vancouver, British Columbia, Canada
Objective: To examine the chromosome-specific frequencies of mosaicism detected by next-generation sequencing (NGS) compared with constitutional aneuploidy. Design: Retrospective cross-sectional review of NGS results from trophectoderm biopsies analyzed by per-chromosome prevalence of mosaicism and constitutional aneuploidy. Setting: Private fertility clinic. Patient(s): A total of 378 patients who underwent preimplantation genetic screening by NGS for routine clinical indications from February 2016 to April 2017. Intervention(s): None. Main Outcome Measure(s): Aneuploidies and mosaicisms were tabulated per chromosome, and whole-chromosome and segmental mosaicisms were also analyzed. Result(s): NGS results were analyzed from 1,547 blastocysts. Mosaicism was detected as the sole abnormality in 17.5% (n ¼ 270) of samples but were also found in 196/634 aneuploid embryos, so the overall incidence of mosaicism per biopsy was 30.1%. Mosaicism did not statistically vary when stratified by maternal age. The mean rate of overall mosaicism per chromosome was 2.46%. When whole chromosome and segmental mosaicisms were compared, unequal frequencies were found in several chromosomes. Trisomy was more frequently detected as whole-chromosome mosaicism, although monosomy was more frequently seen in segmental mosaicism. Aneuploidy and mosaicism displayed different patterns of distribution in various chromosomes. Conclusion(s): Mosaicism is unequally detected in various chromosomes and appears distinct from the distribution pattern of constitutional aneuploidy. Whole chromosome and segmental mosaicisms are also differentially detected. These results contribute to the study of mosaicism, illuminating a differential pattern of detection across the genome. (Fertil SterilÒ 2018;109:857–65. Ó2018 by American Society for Reproductive Medicine.) Key Words: Mosaicism, PGS, CCS, aneuploidy, NGS Discuss: You can discuss this article with its authors and other readers at https://www.fertstertdialog.com/users/16110-fertilityand-sterility/posts/29129-25175
M
osaicism is the coexistence of two or more cell lines with differing chromosomal complements. Caused by mitotic segregation errors during somatic division, it can be generalized or isolated to specific organ systems and associated
with pathology, but it is also found in healthy individuals (1). Mosaicism was documented when G-banding was used to investigate karyotypes in 6–8day-old blastocysts (2) and by fluorescence in-situ hybridization (FISH) at the cleavage stage (3). After the signif-
Received October 19, 2017; revised January 2, 2018; accepted January 8, 2018. G.N. has nothing to disclose. C.J. has nothing to disclose. R.B. has nothing to disclose. C.G. has nothing to disclose. J.H. has nothing to disclose. E.T. has nothing to disclose. N.T. has nothing to disclose. A.Y. has nothing to disclose. Reprint requests: Gary Nakhuda, M.D., Olive Fertility Centre, 555 W 12th Ave., Suite 300 East Tower, Vancouver, BC V5Z3X7 (E-mail: [email protected]). Fertility and Sterility® Vol. 109, No. 5, May 2018 0015-0282/$36.00 Copyright ©2018 American Society for Reproductive Medicine, Published by Elsevier Inc. https://doi.org/10.1016/j.fertnstert.2018.01.011 VOL. 109 NO. 5 / MAY 2018
icant limitations of FISH for preimplantation genetic screening (PGS) were recognized, newer molecular cytogenetic techniques, such as singlenucleotide polymorphism (SNP) array, array comparative genome hybridization (aCGH), and quantitative polymerase chain reaction (PCR), were introduced for 24-chromosome copy number analysis. Most recently, nextgeneration sequencing (NGS) has been applied to PGS as a potentially more efficient and affordable technique (4). In contrast to the other molecular techniques, which are relatively less 857
ORIGINAL ARTICLE: GENETICS sensitive to mosaicism, NGS has been reported to detect mosaicism in 30% of trophectoderm (TE) specimens (5). The detection limit for the proportion of aneuploid cells appears to be dependent on the NGS protocol and has been reported to be as low as 10%–16%, although lower detection limits are associated with greater diagnostic inaccuracy (6). According to the Preimplantation Genetic Diagnosis International Society (PGDIS), embryos with <20% aneuploid cells may be considered euploid and >80% aneuploid cells suggest constitutional aneuploidy; thus, embryos with aneuploid percentages of 20%–80% could be classified as mosaic (7). Consequently, NGS has catalyzed a renewed interest in the phenomenon of embryonic mosaicism, stirring controversy about the significance of these findings, especially regarding the viability of mosaic embryos and the possible manifestations of the aneuploid cell lines. A few publications have demonstrated that a minority of mosaic blastocysts can result in live birth, albeit with lower implantation rates and higher spontaneous abortion (SAB) rates (6, 8, 9). However, given the relative novelty of NGS-detected mosaicism in PGS blastocysts, long-term follow-up has yet to be reported. Pathologic manifestations of mosaicisms are well documented, creating understandable concern regarding the outcomes of mosaic embryo transfers. On the other hand, there is no deterministic evidence to predict how embryonic mosaicism will manifest, and in fact, reasonable evidence to suggest that many embryonic mosaicisms may be clinically irrelevant owing to mechanisms such as self-correction, apoptosis, or preferential allocation (10). Furthermore, there is concern that the increased frequency of mosaicism may be an artifact of NGS technology and may lead to the discarding of potentially viable embryos (11). Contributing to the controversy is the fact that there are several presentations of mosaicisms in TE cells, including whole-chromosome, segmental (or partial), and complex. It has been proposed that some mosaic embryos may be more suitable to transfer than others, as a function of the specific chromosomes that are affected and the degree of mosaicism defined by the percentage of aneuploid cells (7, 12). Although the general prevalence of mosaicism in TE has been reported, there is a paucity of published data regarding how mosaicism affects specific chromosomes, as has been reported with constitutional aneuploidy (13–15). Also, the frequency of whole-chromosome, segmental, and complex mosaicisms as detected by NGS is not well documented. Therefore, we reviewed the NGS results of our PGS cohort to examine the patterns and prevalence of chromosomespecific mosaicisms in TE samples. In addition, because aneuploidy and mosaicism are attributed to different mechanisms, we compared the chromosome-specific aneuploidies rates with those of mosaicisms to determine if distinct distributions are exhibited. We hypothesized that documenting the chromosome-specific frequencies of mosaicism may elucidate the clinical significance of mosaicism and provide insight to the biologic and technical factors that lead to its detection.
858
METHODS We performed a retrospective cross-sectional review of all NGS results in patients presenting for PGS from February 2016 to March 2017. Ovarian stimulation was performed with a combination of highly purified urinary gonadotropin (Menopur; Ferring) and recombinant FSH (Puregon; Merck) in an antagonist protocol (Orgalutran; Merck). Final maturation was triggered with 5–10 IU hCG or GnRH agonist (Decapeptyl; Ipsen) according to physician discretion. Transvaginal oocyte retrieval was performed 35 hours after trigger. Oocytes were denuded and intracytoplasmic sperm injection was performed 4 hours after retrieval and cultured to day 3 in GlobalPlus media (Life Global). Assisted hatching of viable cleavage stage embryos was performed on day 3 with the Zilos-tk Zona Infrared Laser Optical System (Hamilton-Thorne), and culture was continued in individual droplets until day 5 or 6, at which time biopsy was performed on blastocysts with a viable grade, generally considered to be 3CC or greater according to Gardner scoring (16). In HEPES supplemented with 20% serum, micromanipulation and biopsy were performed with the use of laser pulses to release five to ten TE cells for analysis, which were transferred into microcentrifuge tubes and shipped to Reprogenetics, Los Angeles. According to Reprogenetics protocol, PGS was performed as follows. Whole-gene DNA amplification (WGA) from the TE cells was performed with the Sureplex Amplification System (Illumina). WGA was followed by tagmentation of DNA and subsequent PCR reaction of adapter sequences for amplification of insert DNA to add index sequences. Dual-indexed sequencing of pooled libraries was performed on Illumina MiSeq with the use of the Veriseq PGS system (Illumina). Copy number analysis was made with the use of Bluefuse Multi software (Bluefuse/Illumina). Mosaic calls were made when 20%–80% of the cells were aneuploid. Validation of mosaicism detection has been previously documented with a series of mixing experiments using known proportions of euploid and aneuploid cell lines, ranging from 0 to 100% (6). Statistical analysis was performed with the use of R v. 3.3.2. Comparisons of equal frequency were performed with the use of chi-square goodness-of-fit test. When stratified by age group, chi-square test of independence was performed to determine whether age was related to other factors. Oneway analysis of variance was performed when comparing mean error rate of different chromosome structures. In all cases, statistical significance was considered to be at P< .05. The study was approved by the University of British Columbia Institutional Review Board.
RESULTS A total of 1,582 blastocysts were biopsied from 378 patients who underwent 448 egg retrievals. The average age of the patient was 36.54 (range 28–47). Thirteen TE specimens could not be processed owing to degraded DNA and 22 were reported as ‘‘no diagnosis,’’ resulting in a 97.8% diagnosis rate (n ¼ 1,547). Of the embryos, 643 (41.6%) were classified as euploid, 634 (40.9%) aneuploid, and 270 (17.5%) mosaic. Furthermore, of the 634 aneuploid specimens, 196 were also affected by
VOL. 109 NO. 5 / MAY 2018
Fertility and Sterility®
FIGURE 1
Probability of aneuploidy and mosaicism in specific chromosomes. Nakhuda. Frequencies of mosaicisms by NGS. Fertil Steril 2018.
mosaicism. Therefore, mosaicisms were found in a total of 466 TE biopsies (30.1%). In the 634 aneuploid embryos, a total of 964 aneuploidies were found. The mean aneuploidy rate per chromosome was 2.49%. The chromosomes most affected by aneuploidy per biopsy were chromosomes 22 (n ¼ 109; 7.05%), 16 (n ¼ 88; 5.69%), and 21 (n ¼ 86; 5.56%); the least affected were chromosomes 17 (n ¼ 19; 1.23%) and Y (n ¼ 4; 0.26%). Only aneuploidies of chromosomes 15, 16, 18, 21, and 22 were each present in more than 3% of biopsies. In the 466 biopsies affected by mosaicism, 952 total mosaicisms were found. The mean mosaicism rate per chromosome was 2.46%. Per biopsy, mosaicisms occurred most frequently in chromosomes 21 (n ¼ 62; 4.01%), 22 (n ¼ 59; 3.81%), and 2 (n ¼ 56; 3.62%) and least frequently in chromosomes 10 (n ¼ 24; 1.55%), 12 (n ¼ 21; 1.36%), and Y (n ¼ 4; 0.26%). Only chromosomes 1, 2, 6, 21, and 22 were found to be mosaic in more than 3% of biopsies. Figure 1 illustrates the frequencies of aneuploidy and mosaicism per chromosome. When stratified by maternal age, the frequency of euploidy and aneuploidy per biopsy were significantly correlated (P<2.2e16), but the frequency of mosaicism was not (P¼ .7152; Supplemental Table 1, available online at www.fertstert.org). When stratified for paternal age group, there also appeared to be a significant trend for the frequency of aneuploidy (P¼8.76e07), but not for mosaicism (P¼ .2762). However, because maternal age and paternal age were strongly correlated (r ¼ 0.43; P<2.2e16), we stratified paternal age controlling for maternal age. As a result, for each maternal age group, the frequencies of aneuploidy per biopsy among paternal age groups were not significant (<35 y: P¼ .2656; 35–39 y: P¼ .2731; >39 y: P¼ .4618). Conversely, when maternal age was stratified controlling VOL. 109 NO. 5 / MAY 2018
for paternal age, the frequency of aneuploidy per biopsy by maternal age remained significant (<35 y: P¼ .0033; 35– 39 y: P¼1.19e08; >39 y: P¼2.85e14). Regarding frequency of mosaicism, controlling for maternal and paternal age did not yield any significant differences. To investigate the patient effect, a mixed-effects logistic regression was constructed including patient as a random effect. However, including patient effect in the model only slightly changed the P values for age effect. Age was still a highly significant effect for euploidy and aneuploidy, but not a significant effect for mosaicism. Therefore, it is reasonable to infer that the intrinsic dependence within the data does introduce meaningful bias to the results. Mosaicism was subdivided to whole chromosome and segmental errors to determine if these conditions were equally represented in all chromosomes. Whole-chromosome (n ¼ 611; 64.2%) occurred more frequently than segmental (n ¼ 341; 35.8%) mosaicisms (P<2.2e16). Whole-chromosome mosaicisms occurred most frequently in chromosome 22 (n ¼ 58; 3.75%), and least in Y (n ¼ 4; 0.26%). When only segmental mosaicisms were considered, chromosome 1 (n ¼ 28; 1.81%) was most frequently affected, and 19 and 22 (each: n ¼ 1; 0.06%) were least affected of the numeric chromosomes, and no segmental mosaicisms were found in Y. Most chromosomes were equally affected by whole and segmental errors, but comparisons per chromosome showed statistically significant greater frequencies of wholechromosome mosaicisms in chromosomes 3, 6, 8, and 19– 22, whereas chromosomes 1, 9, and X trended toward more segmental mosaicisms (Fig. 2). Whole-chromosome and segmental mosaicisms were categorized as monosomy or trisomy (Table 1). For wholechromosome mosaicisms, trisomy was significantly more 859
ORIGINAL ARTICLE: GENETICS
FIGURE 2
Probability of whole-chromosome and segmental mosaicism in specific chromosomes. Nakhuda. Frequencies of mosaicisms by NGS. Fertil Steril 2018.
frequent than monosomy (P¼ .0002), but for segmental mosaicisms trisomy was less frequent (P¼ .0005). When stratified by maternal age, there were no differences in wholechromosome or segmental monosomy or trisomy (Supplemental Table 2, available online at www.fertstert.org). When compared for the number of mosaicisms per embryo, there were no differences between age groups, with single mosaicisms being most common at all ages (Table 2). Finally, to determine if chromosome structure predicted the frequency of mosaicism, chromosomes were categorized as acrocentric (13, 14, 15, 21, 22), metacentric (1, 3, 16, 19, 20), or submetacentric (2, 4–12, 17, 18, X, Y). Per chromosome, compared with the two other categories, acrocentric chromosomes were significantly more affected by aneuploidy (P¼ .0305), but no significant differences were seen for mosaicism.
DISCUSSION Our data confirm the fact that mosaicism is frequently reported when NGS is used for PGS of TE specimens. In our cohort, 17.5% of biopsies yielded mosaicism as the sole finding, but a total of 30.1% of all biopsies resulted in at least one mosaicism finding when those affected with constitutional aneuploidy were also considered. We found that mosaicisms are not distributed evenly in every chromosome, and the pattern of distribution has certain similarities with that of aneuploidies, albeit with some interesting differences. Furthermore, when mosaicisms are stratified to whole chromosome versus segmental, distributions also appear distinct. To the best of our knowledge this is the first study to compare the chromosome-specific distribution of mosaicisms with 860
constitutional aneuploidies in TE samples screened with the use of NGS. The distribution of mosaicisms and aneuploidies appears to vary depending on the specific chromosome examined. For example, chromosomes 21 and 22 are frequently affected by both aneuploidy (5.56% and 7.05%), and mosaicism (4.01% and 3.81%), suggesting that blastocyst development is tolerant to both meiotic and mitotic errors of these chromosomes. Other chromosomes, however, appear to be differentially affected. For example, several chromosomes, including 1, 2, and 6, have a greater incidence of mosaicism (3.23%, 3.62%, and 3.04%) compared with aneuploidy (1.29%, 1.55%, and 1.62%). It could be speculated that aneuploidies in these chromosomes bias against survival to the blastocyst stage and are therefore less commonly seen in TE biopsies. Alternatively, perhaps these chromosomes are less prone to meiotic than to mitotic errors. In contrast, aneuploidies of some chromosomes, such as 15 (5.11%) and 16 (5.69%), are found with relatively high frequency, but mosaicisms of those same chromosomes are less common (2.2% and 2.07%). These chromosomes may be potentially more prone to meiotic errors, but not to postzygotic events. When mosaicisms were subdivided, whole-chromosome errors were found to be significantly more common than segmental errors (64.2 vs. 35.8%; P<2.2e16). Furthermore, whole-chromosome and segmental mosaicisms were not equally represented in all chromosomes. Most notably, whole-chromosome errors essentially comprised all of the mosaic errors of chromosomes 19 and 22, potentially suggesting that these chromosomes are more susceptible to chromatid segregation errors. On the other hand, mosaicisms of chromosomes 1, 2, 5, 9, and X were found to be more often affected by segmental rather than whole-chromosome errors, VOL. 109 NO. 5 / MAY 2018
Fertility and Sterility®
TABLE 1 Frequency of monosomy and trisomy in whole and segmental chromosome mosaicisms. Mosaicism
Monosomy, n (%)
Trisomy, n (%)
P value
Whole Segmental
259 (16.7) 203 (13.1)
352 (22.8) 138 (8.9)
.0002 .0005
Nakhuda. Frequencies of mosaicisms by NGS. Fertil Steril 2018.
potentially alluding to mitotic subcentromeric instability. In an extensive review of meiotic and mitotic chromosome errors detected in both day 3 and blastocyst stages, McCoy et al. reported segmental errors in 146/17,219 (0.8%) of TE biopsies (15), in contrast to 341/1,547 of (22.0%) biopsies in our cohort. This variation may represent differences in the detection threshold of SNP array, which was used in McCoy et al.'s cohort, compared with NGS, which we used. Fragouli et al. recently reported a live birth rate of 57.1% (8/14) with no miscarriages when blastocysts with segmental mosaicisms were transferred, an outcome similar to that for euploid blastocysts (17). They suggested that cells with segmental mosaicisms are subject to cell cycle checkpoint mechanisms that cause arrest of these abnormal cells, resulting in embryos that ultimately have a normal chromosome complement. However, segmental aneuploidies have been documented in 6% of miscarriages (18) and in association with congenital anomalies in newborns (19), so it may be premature to conclude that segmental mosaicisms are always benign. The significance of whole chromosome versus segmental mosaicisms requires further investigation. The chromosome-specific distribution of constitutional aneuploidies was examined by Franasiak et al., who reviewed results of 15,169 TE samples tested by means of quantitative PCR and SNP array in a general infertility population, of which 40.7% were aneuploid, which was similar to the frequency of aneuploidy in our cohort (40.9%) (13). Fragouli et al. also reviewed chromosome-specific frequency of aneuploidies from aCGH results from TE biopsies (7). In descending order, chromosomes 22, 16, 15, 21, and 19 were the most frequently affected by abnormalities in both of the above studies. From 17,219 TE biopsies examined with the use of SNP array, McCoy et al. reported the frequency of meiotic error from TE biopsies to be greatest in chromosomes 16, 22, 15, and 21 (15). In the present cohort, the chromosomes most frequently affected by aneuploidy, in descending order,
were 22, 16, 21, 15, and 18. Although the data from the present study were generally concordant with previous studies, the differences may be attributable to factors such as discrepancies in study populations, sample size, or treatment-specific factors (20). Alternatively, variation may be attributable to the different cytogenetic techniques used in each study. For example, whereas aCGH concurs with euploid and aneuploid results as detected with the use of NGS, most mosaic TE samples detected by NGS were labeled euploid by aCGH (9). An undetermined proportion of the 17.5% of TE samples in our population that were identified as solely mosaic may have been labeled either euploid or aneuploid by other cytogenetic techniques, potentially skewing the aneuploidy distribution in specific chromosomes. Using SNP array, McCoy et al. characterized the origin of aneuploidy per chromosome based on the chromosomal signature to distinguish between meitoic and mitotic aneuploidy (15). Aneuploidies where both parental homologues were identified were determined to be of meiotic origin, although those with a gain or loss of at least one paternal chromosome copy were identified as mitotic error. Overall, at least one chromosomal error was found in 44.3% of TE biopsies. In contrast to our data, where mosaicism was determined based on the bioinformatic parameters dictated by the NGS protocol without direct evidence of mitotic origin, McCoy et al. inferred mosaicism by deduction of mitotic error by means of the SNP array chromosomal signature of parental homologues. They found that mitotic chromosome losses exceed gains, whereas our data showed a higher incidence of whole and partial trisomies compared with monosomies. Given the significant reduction in complex mitotic errors at the blastocyst stage, the authors concluded that such errors would not likely lead to clinical miscarriage, although others have documented mosaicism as a notable cause of SAB (9). On a related note, it is also of interest to juxtapose aneuploidy distributions in SAB findings with PGS results. In a 12year review of 2,301 specimens, trisomies 16, 21, and 22 were found to be the most common aneuploidies in SAB (21), which is similar to aneuploidies detected in PGS studies, including the present one. However, some aneuploidies, such as trisomies 1 and 19, are seen with relative frequency in PGS studies but were rarely found in SAB specimens. It has been hypothesized that aneuploidies of gene-rich chromosomes may lead to implantation failure or very early miscarriages, thus failing to develop to the point where errors in these chromosomes are seen in SAB (21).
TABLE 2 Number of mosaicisms stratified by age. Age (y)
1 Mosaicism (% per biopsy)
2 Mosaicisms (% per biopsy)
3 Mosaicisms (% per biopsy)
>3 Mosaicisms (% per biopsy)
83 (5.4%) 26 (5.4%) 34 (4.8%) 23 (6.4%)
45 (2.9%) 14 (2.9%) 23 (3.3%) 8 (2.2%)
64 (4.1%) 18 (3.7%) 31 (4.4%) 15 (4.1%)
Overall 274 (17.7%) <35 93 (19.3%) 35–39 123 (17.5%) >39 58 (16.0%) c2 (age vs. no. of mosaicism)
Total biopsies 1,547 482 703 362 0.7911
Nakhuda. Frequencies of mosaicisms by NGS. Fertil Steril 2018.
VOL. 109 NO. 5 / MAY 2018
861
ORIGINAL ARTICLE: GENETICS Although maternal age effect on meiotic aneuploidy is well known, it is apparent that age does not appear to increase mitotic errors that lead to mosaicism (15, 22, 23), even though the first few mitotic divisions, which are most error prone, are driven by maternal transcripts (24, 25). The lack of a maternal age effect on mosaicism was reinforced in our data, with neither the frequency per chromosome nor the number of affected chromosomes per biopsy increasing with age. Anaphase lag and nondisjunction are the mechanisms most often associated with mitotic aneuploidy, but chromosome demolition, endoreplication, cytokinesis errors, cell fusion, and chromosome breakage also have been implicated (10). A paternal effect may also be involved with mosaicism, evidenced by data demonstrating that embryos derived after testicular sperm extraction (TESE) from males with nonobstructive azoospermia have higher rates of mosaicism compared with embryos from ejaculated sperm (26). When paternal age was disentangled from maternal age, we did not see a paternal age effect on either aneuploidy or mosaicism, although TESE specimens were not specifically examined in our cohort. Previous reports have examined the relationship between chromosome structure and the incidence of aneuploidy per chromosome. Franasiak et al. concluded that short and medium acrocentric chromosomes were the most likely to have errors in copy number, and the constitutional aneuploidy rate in those chromosomes rose disproportionately with age, although they did not report mosaicism or mitotic error rates (13). McCoy et al. reported a negative correlation between meiotic error and chromosome length (15). They found a positive correlation with mitotic errors and chromosome length in day 3 blastomeres, but not in TE samples. In our cohort, acrocentric chromosomes were found to have the highest frequency of aneuploidy and mosaicism per chromosome, but the finding was statistically significant only for aneuploidy. However, when whole-chromosome and segmental mosaicisms were considered separately, whole-chromosome mosaicisms were statistically more common in acrocentric chromosomes. Another interesting observation in our data was that some longer chromosomes, particularly 1, 2, 5, 9, and X (Fig. 2), are more frequently affected by segmental rather than whole-chromosome mosaicisms. It could be speculated that longer chromosomes are more prone to breakage during somatic division, yet without attachment to the centromere these errors may lead to cell cycle arrest rather than perpetuating segmental aneuploidy in successive divisions (17). Alternatively, higher rates of segmental errors in larger chromosomes could be the result of technical artifacts related to WGA or biased coverage of the larger chromosomes during NGS. Detection and reporting of mosaicism is dependent on the technical limitations of the cytogenetic method and the bioinformatic interpretation. Mitotic aneuploidies were initially reported with the use of karyotype analysis, but the technique was limited by the requirement for dividing metaphase cells, resulting in fewer than 25% of embryos that could be analyzed (2). FISH also has the ability to detect mosaicism, but sequential hybridization with the use of FISH has diminishing efficiency with each round, with predictive values for 862
abnormal results as low as 41% when even five chromosomes are investigated (27, 28). Furthermore, FISH can not generally detect segmental aneuploidy, because probes are specific for the particular locus being tested (29). Newer methods allow reliable 24-chromosome copy number analysis, but SNP array and aCGH rely on DNA amplification, which may introduce bias due to excessive amplification of specific regions or, in the case of aCGH, failure of amplification of one parental allele (30, 31). In contrast, mosaicism can be reliably detected only with the use of SNP array when 50% of cells are aneuploid (32, 33). NGS has a broader dynamic range for copy number detection, increasing the sensitivity for mosaicism (23). However, a randomized blinded trial comparing quantitative PCR and NGS did not show a difference in sensitivity to mosaicism with the use of default commercial settings when whole-chromosome mosaicisms were examined (34). Determination of whole-chromosome and segmental mosaicisms on the Veriseq PGS system was validated by analyzing aneuploid-euploid DNA ratios ranging from 0:5 to 5:0, based on the assumption that a typical TE biopsy contains five cells (9, 17). With little overlap between the various admixtures, the percentage of normal cells were reliably detected. When copy number falls between disomy and aneusomy, mosaicism is inferred. Ratios >80% or <20% can not be detected in a five-cell sample, but it is generally agreed that results beyond those ranges can be considered to be aneuploid or euploid, respectively (6, 12). However, it has been suggested that under certain conditions, the finding of mosaicism may be artifactual or limited by a high false positive rate (11, 34). Segmental aneuploidies and mosaicisms detected with the use of NGS have been confirmed with whole-blastocyst FISH analysis (35). Segmental mosaicisms can be detected with a high degree of sensitivity with the use of the Veriseq system owing to better coverage of the genome compared with targeted amplification methods, but that may come at the expense of lower specificity resulting in more false positive results (36). Furthermore, the validation of mosaicism with admixture studies relied on fibroblast and TE cell lines with aneuploidies of specific chromosomes, namely, chromosomes 13, 18, and 21 for whole gains, chromosome X for whole losses, chromosomes 10p and13q for partial losses, and specific deletions and duplications (6). Implicit to the conclusions is that detection of these errors would be equivalent regardless of the chromosome affected, however, detection thresholds for specific chromosomes may be variable depending on the technique used (37), thus limiting the ability to categorically define mosaicism, and subsequently affecting the interpretation of clinically reported results, such as those reported here. Assuming that mosaicism is not a technical artifact, the clinical relevance is ultimately determined by its implications for healthy live births. Several reports have documented live births from transfer of mosaic embryos, although implantation failure and miscarriage occur at significantly greater rates compared with euploid embryo transfers (8, 9, 17). Embryos with segmental mosaicisms have been found to have ongoing pregnancy rates equivalent to euploid blastocysts (17). The largest reported series to date, details VOL. 109 NO. 5 / MAY 2018
Fertility and Sterility® the transfer outcomes (but not live birth outcomes) from 143 transfers of mosaic embryos, documenting a 41% (58/143) ongoing implantation rate (6). In that study, complex mosaicisms (three or more mosaic chromosomes) had the lowest implantation rate, but segmental, single, and double mosaicisms had statistically similar outcomes. They did not demonstrate a difference in outcomes when 20%–40% mosaicism was compared with >40%, but that was attributed to small sample size. It seems clear that some mosaic embryos may lead to ongoing pregnancy. However, it is important to emphasize that longitudinal follow-up on the health of the children resulting from mosaic embryos has yet to be performed. Evidence suggests that mosaicism may be associated with numerous pathologic conditions, including autism, cancer, and organ-specific pathology (38, 39). Even when aneuploid cells are limited to the TE with a euploid ICM, as is the case with confined placental mosaicism, intrauterine growth restriction and fetal death may result (40). In pediatric diseases that warrant genetic testing, up to 2% can be attributed to mosaicism (41, 42). Mosaicism can also result in viable embryos with uniparental disomy (UPD) due to monosomy or trisomy rescue of a meiotic aneuploidy, which has significant disease implications (1). Such issues notwithstanding, further studies are required to determine which embryonic mosaicisms are harbingers of clinical sequelae and which are benign. Some guidance has been offered regarding which mosaic embryos can be considered as safest for transfer by the PDGIS (7). Mosaic monosomy was stated to be preferable to mosaic trisomy, because monosomy is not commonly observed as a clinical syndrome (other than 45 X0). However, caution with mosaic monosomy has been suggested because reciprocal trisomy may exist from mitotic nondisjunction (43). Furthermore, the infrequent documentation of live births with autosomal monosomy may represent a detection failure, because low level mosaic monosomy may be below the detection limit of microarray techniques but have nevertheless been associated with significant pathology (44). Munne et al. did not show a difference in implantation rates when mosaic monosomy and trisomy were compared (6). Our data showed a higher prevalence of mosaic trisomy when whole chromosomes were involved and of mosaic monosomy when segmental lesions were involved. Segmental mosaicisms were not addressed by the PGDIS position statement, but limited data suggest that live birth rates are equivalent to those of euploid transfers (17). It has also been documented that segmental mosaicisms can result in pathologic outcomes attributable to segmental UPD or mosaicisms for insertions, deletions, and point mutations (40). Another suggested criteria for consideration of mosaic embryo transfer is the percentage of aneuploid cells (7), which has been found to correlate with embryo viability (45). However, in patients found to have mosaicism in cultured lymphocytes, the proportion of aneuploid cells did not correlate with the severity of the phenotype, and low-level mosaicism may still be associated with pathology (44). The level of mosaicism is not addressed in our data. Needless to say, the data on outcomes of mosaic embryo transfers are still very preliminary and will not likely VOL. 109 NO. 5 / MAY 2018
be satisfactorily addressed for the many years required to determine if pathology manifests over the long term. There are several other issues that limit the interpretation of the present findings. This study was retrospective and observational, demonstrating the relative frequency of chromosome-specific mosaicisms but not offering any direct insight into the clinical relevance of the findings. Ultimately, only a large-scale, randomized, nonselection study would truly resolve the concerns. In addition, as has been noted in other studies of this nature (13), the number of embryos tested per patient is variable, thus potentially skewing the findings as a function of unequal contributions from individuals. However, a logistic regression model that included patient effect did not suggest an intrinsic dependence on results for mosaicism. Furthermore, our data can not necessarily be generalized, because it has been documented that there could be significant variability introduced by iatrogenic factors, such as the excessive force during the biopsy process, or the presence of dead or apoptotic cells could lead to artifactual loss or gain of chromosomes (23). Finally, questions persist regarding the technical limitations that may affect the diagnostic accuracy of mosaicism detection with currently available NGS-PGS platforms, challenging the interpretation of these findings (11, 34, 36). It has been suggested that the highest level of evidence of mosaicism would be demonstrated from reciprocal errors found in double biopsies of a blastocyst (37). Unfortunately, that standard may be impractical to apply clinically, so clinicians are still faced with the challenge of interpreting results as they are reported, which is based on the intermediate log-2 ratios. The data presented here are insufficient to draw any clinical conclusions or to validate or refute the technical aspects. However, such data may assist in patient counseling and spur future research by comparing the prevalence of chromosome-specific mosaicism in blastocysts with those associated with pathologic findings.
CONCLUSION Our data illustrate the differences in detection of mosaicism compared with constitutional aneuploidy with the use of NGS, suggesting that mitotic and meiotic error have characteristic distributions across the genome that vary according to the specific chromosome. In addition, whole-chromosome and segmental mosaicisms are differentially distributed, potentially suggesting chromosome-specific susceptibilities to mitotic error. Alternatively, the lower prevalence of certain mosaicisms in TE biopsies may be interpreted as a lower tolerance of errors in gene-rich regions which reduces survival to the blastocyst stage (15). Additional research is required to determine why certain chromosomes seem more predisposed to meiotic or mitotic error, and why segmental and whole-chromosome mitotic errors occur with preponderance for particular chromosomes. The contribution of technical limitations and artifact associated with NGS also can not be ruled out as a source of confusion, potentially resulting in false positive mosaic results. Larger-scale multicenter studies could confirm if our findings are robust and, most importantly, if 863
ORIGINAL ARTICLE: GENETICS specific embryonic mosaicisms are consistent with healthy live birth.
20.
REFERENCES
21.
1. 2.
3.
4.
5.
6.
7.
8. 9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
864
Biesecker LG, Spinner NB. A genomic view of mosaicism and human disease. Nat Rev Genet 2013;14:307–20. Clouston HJ, Fenwick J, Webb AL, Herbert M, Murdoch A, Wolstenholme J. Detection of mosaic and nonmosaic chromosome abnormalities in 6- to 8day-old human blastocysts. Hum Genet 1997;101:30–6. Delhanty JDA, Griffin DK, Handyside AH, Harper J, Atkinson GHG, Pieters MHEC, et al. Detection of aneuploidy and chromosomal mosaicism in human embryos during preimplantation sex determination by fluorescent in situ hybridisation, (FISH). Hum Mol Genet 1993;2:1183–5. Zheng H, Jin H, Liu L, Liu J, Wang WH. Application of next-generation sequencing for 24-chromosome aneuploidy screening of human preimplantation embryos. Mol Cytogenet 2015;8:38. Tormasi S, Capaldi R, Gouw F, Munne S, Wells D, Fragouli E. The use of next generation sequencing to determine the developmental potential of mosaic blastocysts. Fertil Steril 2016;105(2 Suppl 1):e41. Munn e S, Blazek J, Large M, Martinez-Ortiz PA, Nisson H, Liu E, et al. Detailed investigation into the cytogenetic constitution and pregnancy outcome of replacing mosaic blastocysts detected with the use of highresolution next-generation sequencing. Fertil Steril 2017;108:62–71.e8. Preimplantation Genetic Diagnosis International Society. PGDIS position statement on chromosome mosaicism and preimplantation aneuploidy testing at the blastocyst stage. July 19, 2016. Available at: http://pgdis. org/docs/newsletter_071816.html. Accessed September 4, 2017. Greco E, Minasi MG, Fiorentino F. Healthy babies after intrauterine transfer of mosaic aneuploid blastocysts. N Engl J Med 2015;373:2089–90. Maxwell SM, Colls P, Hodes-Wertz B, McCulloh DH, McCaffrey C, Wells D, et al. Why do euploid embryos miscarry? A case-control study comparing the rate of aneuploidy within presumed euploid embryos that resulted in miscarriage or live birth using next-generation sequencing. Fertil Steril 2016;106: 1414–9.e5. Mantikou E, Wong KM, Repping S, Mastenbroek S. Molecular origin of mitotic aneuploidies in preimplantation embryos. Biochim Biophys Acta 2012;1822:1921–30. Esfandiari N, Bunnell ME, Casper RF. Human embryo mosaicism: did we drop the ball on chromosomal testing? J Assist Reprod Genet 2016;33: 1439–44. Besser AG, Mounts EL. Counselling considerations for chromosomal mosaicism detected by preimplantation genetic screening. Reprod Biomed Online 2017;34:369–74. Franasiak JM, Forman EJ, Hong KH, Werner MD, Upham KM, Treff NR, et al. Aneuploidy across individual chromosomes at the embryonic level in trophectoderm biopsies: changes with patient age and chromosome structure. J Assist Reprod Genet 2014;31:1501–9. Fragouli E, Alfarawati S, Spath K, Jaroudi S, Sarasa J, Enciso M, et al. The origin and impact of embryonic aneuploidy. Hum Genet 2013;132:1001– 13. McCoy RC, Demko ZP, Ryan A, Banjevic M, Hill M, Sigurjonsson S, et al. Evidence of selection against complex mitotic-origin aneuploidy during preimplantation development. PLoS Genet 2015;11:e1005601. Gardner DK, Lane M, Stevens J, Schlenker T, Schoolcraft WB. Blastocyst score affects implantation and pregnancy outcome: toward a single blastocyst transfer. Fertil Steril 2000;73:1155–8. Fragouli E, Alfarawati S, Spath K, Babariya D, Tarozzi N, Borini A, et al. Analysis of implantation and ongoing pregnancy rates following the transfer of mosaic diploid-aneuploid blastocysts. Hum Genet 2017;2017:805–13. Martínez MC, M endez C, Ferro J, Nicolas M, Serra V, Landeras J. Cytogenetic analysis of early nonviable pregnancies after assisted reproduction treatment. Fertil Steril 2010;93:289–92. Wellesley D, Dolk H, Boyd PA, Greenlees R, Haeusler M, Nelen V, et al. Rare chromosome abnormalities, prevalence and prenatal diagnosis rates from population-based congenital anomaly registers in Europe. Eur J Hum Genet 2012;20:521–6.
22.
23. 24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
Munne S, Magli C, Adler A, Wright G, de Boer K, Mortimer D, et al. Treatment-related chromosome abnormalities in human embryos. Hum Reprod 1997;12:780–4. Menasha J, Levy B, Hirschhorn K, Kardon NB. Incidence and spectrum of chromosome abnormalities in spontaneous abortions: new insights from a 12-year study. Genet Med 2005;7:251–63. Antonarakis SE, Avramopoulos D, Blouin JL, Talbot CC, Schinzel AA. Mitotic errors in somatic cells cause trisomy 21 in about 4.5% of cases and are not associated with advanced maternal age. Nat Genet 1993;3:146–50. Munne S, Wells D. Detection of mosaicism at blastocyst stage with the use of high-resolution next-generation sequencing. Fertil Steril 2017:1085–91. Braude P, Bolton V, Moore S. Human gene expression first occurs between the four- and eight-cell stages of preimplantation development. Nature 1988;332:459–61. Katz-Jaffe MG, Trounson AO, Cram DS. Chromosome 21 mosaic human preimplantation embryos predominantly arise from diploid conceptions. Fertil Steril 2005;84:634–43. Silber S, Escudero T, Lenahan K, Abdelhadi I, Kilani Z, Munne S. Chromosomal abnormalities in embryos derived from testicular sperm extraction. Fertil Steril 2003;79:30–8. Liu J, Tsai YL, Zheng XZ, Baramki TA, Yazigi RA, Katz E. Potential use of repeated fluorescence in situ hybridization in the same human blastomeres for preimplantation genetic diagnosis. Fertil Steril 1998;70:729–33. Scriven PN, Bossuyt PMM. Diagnostic accuracy: Theoretical models for preimplantation genetic testing of a single nucleus using the fluorescence in situ hybridization technique. Hum Reprod 2010;25:2622–8. Wilton L, Voullaire L, Sargeant P, Williamson R, McBain J. Preimplantation aneuploidy screening using comparative genomic hybridization or fluorescence in situ hybridization of embryos from patients with recurrent implantation failure. Fertil Steril 2003;80:860–8. Handyside AH, Robinson MD, Simpson RJ, Omar MB, Shaw MA, Grudzinskas JG, et al. Isothermal whole genome amplification from single and small numbers of cells: a new era for preimplantation genetic diagnosis of inherited disease. Mol Hum Reprod 2004;10:767–72. Iwamoto K, Bundo M, Ueda J, Nakano Y, Ukai W, Hashimoto E, et al. Detection of chromosomal structural alterations in single cells by SNP arrays: a systematic survey of amplification bias and optimized workflow. PLoS One 2007;2:e1306. Treff NR, Levy B, Su J, Northrop LE, Tao X, Scott RT. SNP microarray–based 24 chromosome aneuploidy screening is significantly more consistent than FISH. Mol Hum Reprod 2010;16:583–9. Northrop LE, Treff NR, Levy B, Scott JT. SNP microarray–based 24 chromosome aneuploidy screening demonstrates that cleavage-stage FISH poorly predicts aneuploidy in embryos that develop to morphologically normal blastocysts. Mol Hum Reprod 2010;16:590–600. Goodrich D, Tao X, Bohrer C, Lonczak A, Xing T, Zimmerman R, et al. A randomized and blinded comparison of qPCR and NGS-based detection of aneuploidy in a cell line mixture model of blastocyst biopsy mosaicism. J Assist Reprod Genet 2016;33:1473–80. Vera-Rodríguez M, Michel CE, Mercader A, Bladon AJ, Rodrigo L, Kokocinski F, et al. Distribution patterns of segmental aneuploidies in human blastocysts identified by next-generation sequencing. Fertil Steril 2016;105:1047–55.e2. Goodrich D, Xing T, Tao X, Lonczak A, Zhan Y, Landis J, et al. Evaluation of comprehensive chromosome screening platforms for the detection of mosaic segmental aneuploidy. J Assist Reprod Genet 2017;34:975–81. Capalbo A, Ubaldi FM, Rienzi L, Scott R, Treff N. Detecting mosaicism in trophectoderm biopsies: current challenges and future possibilities. Hum Reprod 2016;32:492–8. Yurov YB, Vorsanova SG, Iourov IY, Demidova IA, Beresheva AK, Kravetz VS, et al. Unexplained autism is frequently associated with low-level mosaic aneuploidy. J Med Genet 2007;44:521–5. Jacobs KB, Yeager M, Zhou W, Wacholder S, Wang Z, Rodriguez-Santiago B, et al. Detectable clonal mosaicism and its relationship to aging and cancer. Nat Genet 2012;44:651–8. Spinner NB, Conlin LK. Mosaicism and clinical genetics. Am J Med Genet C Semin Med Genet 2014;166:397–405.
VOL. 109 NO. 5 / MAY 2018
Fertility and Sterility® 41.
42.
43.
Ballif BC, Rorem EA, Sundin K, Lincicum M, Gaskin S, Coppinger J, et al. Detection of low-level mosaicism by array CGH in routine diagnostic specimens. Am J Med Genet A 2006;140:2757–67. Cheung SW, Shaw CA, Scott DA, Patel A, Sahoo T, Bacino CA, et al. Microarray-based CGH detects chromosomal mosaicism not revealed by conventional cytogenetics. Am J Med Genet A 2007;143:1679–86. Scott RT, Galliano D. The challenge of embryonic mosaicism in preimplantation genetic screening. Fertil Steril 2016;105:1150–2.
VOL. 109 NO. 5 / MAY 2018
44.
45.
Hochstenbach R, Krijtenburg PJ, van der Veken LT, van der Smagt J, Roeleveld-Versteegh A, Visser G, et al. Monosomy 20 mosaicism revealed by extensive karyotyping in blood and skin cells: Case report and review of the literature. Cytogenet Genome Res 2014;144:155–62. Bolton H, Graham SJL, Van der Aa N, Kumar P, Theunis K, Fernandez Gallardo E, et al. Mouse model of chromosome mosaicism reveals lineage-specific depletion of aneuploid cells and normal developmental potential. Nat Commun 2016;7:11165.
865
Frequencies of chromosome-specific mosaicisms in trophoectoderm biopsies detected by next-generation sequencing Gary Nakhuda, M.D.,a,b Chen Jing, M.S.,a Rachel Butler, M.S.,a,b Colleen Guimond, M.S.,a,b Jason Hitkari, M.D.,a,b Elizabeth Taylor, M.D.,a,b Niamh Tallon, M.D.,a,b and Albert Yuzpe, M.D.a,b a
Olive Fertility Centre; and b University of British Columbia, Vancouver, British Columbia, Canada
Objective: To examine the chromosome-specific frequencies of mosaicism detected by next-generation sequencing (NGS) compared with constitutional aneuploidy. Design: Retrospective cross-sectional review of NGS results from trophectoderm biopsies analyzed by per-chromosome prevalence of mosaicism and constitutional aneuploidy. Setting: Private fertility clinic. Patient(s): A total of 378 patients who underwent preimplantation genetic screening by NGS for routine clinical indications from February 2016 to April 2017. Intervention(s): None. Main Outcome Measure(s): Aneuploidies and mosaicisms were tabulated per chromosome, and whole-chromosome and segmental mosaicisms were also analyzed. Result(s): NGS results were analyzed from 1,547 blastocysts. Mosaicism was detected as the sole abnormality in 17.5% (n ¼ 270) of samples but were also found in 196/634 aneuploid embryos, so the overall incidence of mosaicism per biopsy was 30.1%. Mosaicism did not statistically vary when stratified by maternal age. The mean rate of overall mosaicism per chromosome was 2.46%. When whole chromosome and segmental mosaicisms were compared, unequal frequencies were found in several chromosomes. Trisomy was more frequently detected as whole-chromosome mosaicism, although monosomy was more frequently seen in segmental mosaicism. Aneuploidy and mosaicism displayed different patterns of distribution in various chromosomes. Conclusion(s): Mosaicism is unequally detected in various chromosomes and appears distinct from the distribution pattern of constitutional aneuploidy. Whole chromosome and segmental mosaicisms are also differentially detected. These results contribute to the study of mosaicism, illuminating a differential pattern of detection across the genome. (Fertil SterilÒ 2018;109:857–65. Ó2018 by American Society for Reproductive Medicine.) Key Words: Mosaicism, PGS, CCS, aneuploidy, NGS Discuss: You can discuss this article with its authors and other readers at https://www.fertstertdialog.com/users/16110-fertilityand-sterility/posts/29129-25175
M
osaicism is the coexistence of two or more cell lines with differing chromosomal complements. Caused by mitotic segregation errors during somatic division, it can be generalized or isolated to specific organ systems and associated
with pathology, but it is also found in healthy individuals (1). Mosaicism was documented when G-banding was used to investigate karyotypes in 6–8day-old blastocysts (2) and by fluorescence in-situ hybridization (FISH) at the cleavage stage (3). After the signif-
Received October 19, 2017; revised January 2, 2018; accepted January 8, 2018. G.N. has nothing to disclose. C.J. has nothing to disclose. R.B. has nothing to disclose. C.G. has nothing to disclose. J.H. has nothing to disclose. E.T. has nothing to disclose. N.T. has nothing to disclose. A.Y. has nothing to disclose. Reprint requests: Gary Nakhuda, M.D., Olive Fertility Centre, 555 W 12th Ave., Suite 300 East Tower, Vancouver, BC V5Z3X7 (E-mail: [email protected]). Fertility and Sterility® Vol. 109, No. 5, May 2018 0015-0282/$36.00 Copyright ©2018 American Society for Reproductive Medicine, Published by Elsevier Inc. https://doi.org/10.1016/j.fertnstert.2018.01.011 VOL. 109 NO. 5 / MAY 2018
icant limitations of FISH for preimplantation genetic screening (PGS) were recognized, newer molecular cytogenetic techniques, such as singlenucleotide polymorphism (SNP) array, array comparative genome hybridization (aCGH), and quantitative polymerase chain reaction (PCR), were introduced for 24-chromosome copy number analysis. Most recently, nextgeneration sequencing (NGS) has been applied to PGS as a potentially more efficient and affordable technique (4). In contrast to the other molecular techniques, which are relatively less 857
ORIGINAL ARTICLE: GENETICS sensitive to mosaicism, NGS has been reported to detect mosaicism in 30% of trophectoderm (TE) specimens (5). The detection limit for the proportion of aneuploid cells appears to be dependent on the NGS protocol and has been reported to be as low as 10%–16%, although lower detection limits are associated with greater diagnostic inaccuracy (6). According to the Preimplantation Genetic Diagnosis International Society (PGDIS), embryos with <20% aneuploid cells may be considered euploid and >80% aneuploid cells suggest constitutional aneuploidy; thus, embryos with aneuploid percentages of 20%–80% could be classified as mosaic (7). Consequently, NGS has catalyzed a renewed interest in the phenomenon of embryonic mosaicism, stirring controversy about the significance of these findings, especially regarding the viability of mosaic embryos and the possible manifestations of the aneuploid cell lines. A few publications have demonstrated that a minority of mosaic blastocysts can result in live birth, albeit with lower implantation rates and higher spontaneous abortion (SAB) rates (6, 8, 9). However, given the relative novelty of NGS-detected mosaicism in PGS blastocysts, long-term follow-up has yet to be reported. Pathologic manifestations of mosaicisms are well documented, creating understandable concern regarding the outcomes of mosaic embryo transfers. On the other hand, there is no deterministic evidence to predict how embryonic mosaicism will manifest, and in fact, reasonable evidence to suggest that many embryonic mosaicisms may be clinically irrelevant owing to mechanisms such as self-correction, apoptosis, or preferential allocation (10). Furthermore, there is concern that the increased frequency of mosaicism may be an artifact of NGS technology and may lead to the discarding of potentially viable embryos (11). Contributing to the controversy is the fact that there are several presentations of mosaicisms in TE cells, including whole-chromosome, segmental (or partial), and complex. It has been proposed that some mosaic embryos may be more suitable to transfer than others, as a function of the specific chromosomes that are affected and the degree of mosaicism defined by the percentage of aneuploid cells (7, 12). Although the general prevalence of mosaicism in TE has been reported, there is a paucity of published data regarding how mosaicism affects specific chromosomes, as has been reported with constitutional aneuploidy (13–15). Also, the frequency of whole-chromosome, segmental, and complex mosaicisms as detected by NGS is not well documented. Therefore, we reviewed the NGS results of our PGS cohort to examine the patterns and prevalence of chromosomespecific mosaicisms in TE samples. In addition, because aneuploidy and mosaicism are attributed to different mechanisms, we compared the chromosome-specific aneuploidies rates with those of mosaicisms to determine if distinct distributions are exhibited. We hypothesized that documenting the chromosome-specific frequencies of mosaicism may elucidate the clinical significance of mosaicism and provide insight to the biologic and technical factors that lead to its detection.
858
METHODS We performed a retrospective cross-sectional review of all NGS results in patients presenting for PGS from February 2016 to March 2017. Ovarian stimulation was performed with a combination of highly purified urinary gonadotropin (Menopur; Ferring) and recombinant FSH (Puregon; Merck) in an antagonist protocol (Orgalutran; Merck). Final maturation was triggered with 5–10 IU hCG or GnRH agonist (Decapeptyl; Ipsen) according to physician discretion. Transvaginal oocyte retrieval was performed 35 hours after trigger. Oocytes were denuded and intracytoplasmic sperm injection was performed 4 hours after retrieval and cultured to day 3 in GlobalPlus media (Life Global). Assisted hatching of viable cleavage stage embryos was performed on day 3 with the Zilos-tk Zona Infrared Laser Optical System (Hamilton-Thorne), and culture was continued in individual droplets until day 5 or 6, at which time biopsy was performed on blastocysts with a viable grade, generally considered to be 3CC or greater according to Gardner scoring (16). In HEPES supplemented with 20% serum, micromanipulation and biopsy were performed with the use of laser pulses to release five to ten TE cells for analysis, which were transferred into microcentrifuge tubes and shipped to Reprogenetics, Los Angeles. According to Reprogenetics protocol, PGS was performed as follows. Whole-gene DNA amplification (WGA) from the TE cells was performed with the Sureplex Amplification System (Illumina). WGA was followed by tagmentation of DNA and subsequent PCR reaction of adapter sequences for amplification of insert DNA to add index sequences. Dual-indexed sequencing of pooled libraries was performed on Illumina MiSeq with the use of the Veriseq PGS system (Illumina). Copy number analysis was made with the use of Bluefuse Multi software (Bluefuse/Illumina). Mosaic calls were made when 20%–80% of the cells were aneuploid. Validation of mosaicism detection has been previously documented with a series of mixing experiments using known proportions of euploid and aneuploid cell lines, ranging from 0 to 100% (6). Statistical analysis was performed with the use of R v. 3.3.2. Comparisons of equal frequency were performed with the use of chi-square goodness-of-fit test. When stratified by age group, chi-square test of independence was performed to determine whether age was related to other factors. Oneway analysis of variance was performed when comparing mean error rate of different chromosome structures. In all cases, statistical significance was considered to be at P< .05. The study was approved by the University of British Columbia Institutional Review Board.
RESULTS A total of 1,582 blastocysts were biopsied from 378 patients who underwent 448 egg retrievals. The average age of the patient was 36.54 (range 28–47). Thirteen TE specimens could not be processed owing to degraded DNA and 22 were reported as ‘‘no diagnosis,’’ resulting in a 97.8% diagnosis rate (n ¼ 1,547). Of the embryos, 643 (41.6%) were classified as euploid, 634 (40.9%) aneuploid, and 270 (17.5%) mosaic. Furthermore, of the 634 aneuploid specimens, 196 were also affected by
VOL. 109 NO. 5 / MAY 2018
Fertility and Sterility®
FIGURE 1
Probability of aneuploidy and mosaicism in specific chromosomes. Nakhuda. Frequencies of mosaicisms by NGS. Fertil Steril 2018.
mosaicism. Therefore, mosaicisms were found in a total of 466 TE biopsies (30.1%). In the 634 aneuploid embryos, a total of 964 aneuploidies were found. The mean aneuploidy rate per chromosome was 2.49%. The chromosomes most affected by aneuploidy per biopsy were chromosomes 22 (n ¼ 109; 7.05%), 16 (n ¼ 88; 5.69%), and 21 (n ¼ 86; 5.56%); the least affected were chromosomes 17 (n ¼ 19; 1.23%) and Y (n ¼ 4; 0.26%). Only aneuploidies of chromosomes 15, 16, 18, 21, and 22 were each present in more than 3% of biopsies. In the 466 biopsies affected by mosaicism, 952 total mosaicisms were found. The mean mosaicism rate per chromosome was 2.46%. Per biopsy, mosaicisms occurred most frequently in chromosomes 21 (n ¼ 62; 4.01%), 22 (n ¼ 59; 3.81%), and 2 (n ¼ 56; 3.62%) and least frequently in chromosomes 10 (n ¼ 24; 1.55%), 12 (n ¼ 21; 1.36%), and Y (n ¼ 4; 0.26%). Only chromosomes 1, 2, 6, 21, and 22 were found to be mosaic in more than 3% of biopsies. Figure 1 illustrates the frequencies of aneuploidy and mosaicism per chromosome. When stratified by maternal age, the frequency of euploidy and aneuploidy per biopsy were significantly correlated (P<2.2e16), but the frequency of mosaicism was not (P¼ .7152; Supplemental Table 1, available online at www.fertstert.org). When stratified for paternal age group, there also appeared to be a significant trend for the frequency of aneuploidy (P¼8.76e07), but not for mosaicism (P¼ .2762). However, because maternal age and paternal age were strongly correlated (r ¼ 0.43; P<2.2e16), we stratified paternal age controlling for maternal age. As a result, for each maternal age group, the frequencies of aneuploidy per biopsy among paternal age groups were not significant (<35 y: P¼ .2656; 35–39 y: P¼ .2731; >39 y: P¼ .4618). Conversely, when maternal age was stratified controlling VOL. 109 NO. 5 / MAY 2018
for paternal age, the frequency of aneuploidy per biopsy by maternal age remained significant (<35 y: P¼ .0033; 35– 39 y: P¼1.19e08; >39 y: P¼2.85e14). Regarding frequency of mosaicism, controlling for maternal and paternal age did not yield any significant differences. To investigate the patient effect, a mixed-effects logistic regression was constructed including patient as a random effect. However, including patient effect in the model only slightly changed the P values for age effect. Age was still a highly significant effect for euploidy and aneuploidy, but not a significant effect for mosaicism. Therefore, it is reasonable to infer that the intrinsic dependence within the data does introduce meaningful bias to the results. Mosaicism was subdivided to whole chromosome and segmental errors to determine if these conditions were equally represented in all chromosomes. Whole-chromosome (n ¼ 611; 64.2%) occurred more frequently than segmental (n ¼ 341; 35.8%) mosaicisms (P<2.2e16). Whole-chromosome mosaicisms occurred most frequently in chromosome 22 (n ¼ 58; 3.75%), and least in Y (n ¼ 4; 0.26%). When only segmental mosaicisms were considered, chromosome 1 (n ¼ 28; 1.81%) was most frequently affected, and 19 and 22 (each: n ¼ 1; 0.06%) were least affected of the numeric chromosomes, and no segmental mosaicisms were found in Y. Most chromosomes were equally affected by whole and segmental errors, but comparisons per chromosome showed statistically significant greater frequencies of wholechromosome mosaicisms in chromosomes 3, 6, 8, and 19– 22, whereas chromosomes 1, 9, and X trended toward more segmental mosaicisms (Fig. 2). Whole-chromosome and segmental mosaicisms were categorized as monosomy or trisomy (Table 1). For wholechromosome mosaicisms, trisomy was significantly more 859
ORIGINAL ARTICLE: GENETICS
FIGURE 2
Probability of whole-chromosome and segmental mosaicism in specific chromosomes. Nakhuda. Frequencies of mosaicisms by NGS. Fertil Steril 2018.
frequent than monosomy (P¼ .0002), but for segmental mosaicisms trisomy was less frequent (P¼ .0005). When stratified by maternal age, there were no differences in wholechromosome or segmental monosomy or trisomy (Supplemental Table 2, available online at www.fertstert.org). When compared for the number of mosaicisms per embryo, there were no differences between age groups, with single mosaicisms being most common at all ages (Table 2). Finally, to determine if chromosome structure predicted the frequency of mosaicism, chromosomes were categorized as acrocentric (13, 14, 15, 21, 22), metacentric (1, 3, 16, 19, 20), or submetacentric (2, 4–12, 17, 18, X, Y). Per chromosome, compared with the two other categories, acrocentric chromosomes were significantly more affected by aneuploidy (P¼ .0305), but no significant differences were seen for mosaicism.
DISCUSSION Our data confirm the fact that mosaicism is frequently reported when NGS is used for PGS of TE specimens. In our cohort, 17.5% of biopsies yielded mosaicism as the sole finding, but a total of 30.1% of all biopsies resulted in at least one mosaicism finding when those affected with constitutional aneuploidy were also considered. We found that mosaicisms are not distributed evenly in every chromosome, and the pattern of distribution has certain similarities with that of aneuploidies, albeit with some interesting differences. Furthermore, when mosaicisms are stratified to whole chromosome versus segmental, distributions also appear distinct. To the best of our knowledge this is the first study to compare the chromosome-specific distribution of mosaicisms with 860
constitutional aneuploidies in TE samples screened with the use of NGS. The distribution of mosaicisms and aneuploidies appears to vary depending on the specific chromosome examined. For example, chromosomes 21 and 22 are frequently affected by both aneuploidy (5.56% and 7.05%), and mosaicism (4.01% and 3.81%), suggesting that blastocyst development is tolerant to both meiotic and mitotic errors of these chromosomes. Other chromosomes, however, appear to be differentially affected. For example, several chromosomes, including 1, 2, and 6, have a greater incidence of mosaicism (3.23%, 3.62%, and 3.04%) compared with aneuploidy (1.29%, 1.55%, and 1.62%). It could be speculated that aneuploidies in these chromosomes bias against survival to the blastocyst stage and are therefore less commonly seen in TE biopsies. Alternatively, perhaps these chromosomes are less prone to meiotic than to mitotic errors. In contrast, aneuploidies of some chromosomes, such as 15 (5.11%) and 16 (5.69%), are found with relatively high frequency, but mosaicisms of those same chromosomes are less common (2.2% and 2.07%). These chromosomes may be potentially more prone to meiotic errors, but not to postzygotic events. When mosaicisms were subdivided, whole-chromosome errors were found to be significantly more common than segmental errors (64.2 vs. 35.8%; P<2.2e16). Furthermore, whole-chromosome and segmental mosaicisms were not equally represented in all chromosomes. Most notably, whole-chromosome errors essentially comprised all of the mosaic errors of chromosomes 19 and 22, potentially suggesting that these chromosomes are more susceptible to chromatid segregation errors. On the other hand, mosaicisms of chromosomes 1, 2, 5, 9, and X were found to be more often affected by segmental rather than whole-chromosome errors, VOL. 109 NO. 5 / MAY 2018
Fertility and Sterility®
TABLE 1 Frequency of monosomy and trisomy in whole and segmental chromosome mosaicisms. Mosaicism
Monosomy, n (%)
Trisomy, n (%)
P value
Whole Segmental
259 (16.7) 203 (13.1)
352 (22.8) 138 (8.9)
.0002 .0005
Nakhuda. Frequencies of mosaicisms by NGS. Fertil Steril 2018.
potentially alluding to mitotic subcentromeric instability. In an extensive review of meiotic and mitotic chromosome errors detected in both day 3 and blastocyst stages, McCoy et al. reported segmental errors in 146/17,219 (0.8%) of TE biopsies (15), in contrast to 341/1,547 of (22.0%) biopsies in our cohort. This variation may represent differences in the detection threshold of SNP array, which was used in McCoy et al.'s cohort, compared with NGS, which we used. Fragouli et al. recently reported a live birth rate of 57.1% (8/14) with no miscarriages when blastocysts with segmental mosaicisms were transferred, an outcome similar to that for euploid blastocysts (17). They suggested that cells with segmental mosaicisms are subject to cell cycle checkpoint mechanisms that cause arrest of these abnormal cells, resulting in embryos that ultimately have a normal chromosome complement. However, segmental aneuploidies have been documented in 6% of miscarriages (18) and in association with congenital anomalies in newborns (19), so it may be premature to conclude that segmental mosaicisms are always benign. The significance of whole chromosome versus segmental mosaicisms requires further investigation. The chromosome-specific distribution of constitutional aneuploidies was examined by Franasiak et al., who reviewed results of 15,169 TE samples tested by means of quantitative PCR and SNP array in a general infertility population, of which 40.7% were aneuploid, which was similar to the frequency of aneuploidy in our cohort (40.9%) (13). Fragouli et al. also reviewed chromosome-specific frequency of aneuploidies from aCGH results from TE biopsies (7). In descending order, chromosomes 22, 16, 15, 21, and 19 were the most frequently affected by abnormalities in both of the above studies. From 17,219 TE biopsies examined with the use of SNP array, McCoy et al. reported the frequency of meiotic error from TE biopsies to be greatest in chromosomes 16, 22, 15, and 21 (15). In the present cohort, the chromosomes most frequently affected by aneuploidy, in descending order,
were 22, 16, 21, 15, and 18. Although the data from the present study were generally concordant with previous studies, the differences may be attributable to factors such as discrepancies in study populations, sample size, or treatment-specific factors (20). Alternatively, variation may be attributable to the different cytogenetic techniques used in each study. For example, whereas aCGH concurs with euploid and aneuploid results as detected with the use of NGS, most mosaic TE samples detected by NGS were labeled euploid by aCGH (9). An undetermined proportion of the 17.5% of TE samples in our population that were identified as solely mosaic may have been labeled either euploid or aneuploid by other cytogenetic techniques, potentially skewing the aneuploidy distribution in specific chromosomes. Using SNP array, McCoy et al. characterized the origin of aneuploidy per chromosome based on the chromosomal signature to distinguish between meitoic and mitotic aneuploidy (15). Aneuploidies where both parental homologues were identified were determined to be of meiotic origin, although those with a gain or loss of at least one paternal chromosome copy were identified as mitotic error. Overall, at least one chromosomal error was found in 44.3% of TE biopsies. In contrast to our data, where mosaicism was determined based on the bioinformatic parameters dictated by the NGS protocol without direct evidence of mitotic origin, McCoy et al. inferred mosaicism by deduction of mitotic error by means of the SNP array chromosomal signature of parental homologues. They found that mitotic chromosome losses exceed gains, whereas our data showed a higher incidence of whole and partial trisomies compared with monosomies. Given the significant reduction in complex mitotic errors at the blastocyst stage, the authors concluded that such errors would not likely lead to clinical miscarriage, although others have documented mosaicism as a notable cause of SAB (9). On a related note, it is also of interest to juxtapose aneuploidy distributions in SAB findings with PGS results. In a 12year review of 2,301 specimens, trisomies 16, 21, and 22 were found to be the most common aneuploidies in SAB (21), which is similar to aneuploidies detected in PGS studies, including the present one. However, some aneuploidies, such as trisomies 1 and 19, are seen with relative frequency in PGS studies but were rarely found in SAB specimens. It has been hypothesized that aneuploidies of gene-rich chromosomes may lead to implantation failure or very early miscarriages, thus failing to develop to the point where errors in these chromosomes are seen in SAB (21).
TABLE 2 Number of mosaicisms stratified by age. Age (y)
1 Mosaicism (% per biopsy)
2 Mosaicisms (% per biopsy)
3 Mosaicisms (% per biopsy)
>3 Mosaicisms (% per biopsy)
83 (5.4%) 26 (5.4%) 34 (4.8%) 23 (6.4%)
45 (2.9%) 14 (2.9%) 23 (3.3%) 8 (2.2%)
64 (4.1%) 18 (3.7%) 31 (4.4%) 15 (4.1%)
Overall 274 (17.7%) <35 93 (19.3%) 35–39 123 (17.5%) >39 58 (16.0%) c2 (age vs. no. of mosaicism)
Total biopsies 1,547 482 703 362 0.7911
Nakhuda. Frequencies of mosaicisms by NGS. Fertil Steril 2018.
VOL. 109 NO. 5 / MAY 2018
861
ORIGINAL ARTICLE: GENETICS Although maternal age effect on meiotic aneuploidy is well known, it is apparent that age does not appear to increase mitotic errors that lead to mosaicism (15, 22, 23), even though the first few mitotic divisions, which are most error prone, are driven by maternal transcripts (24, 25). The lack of a maternal age effect on mosaicism was reinforced in our data, with neither the frequency per chromosome nor the number of affected chromosomes per biopsy increasing with age. Anaphase lag and nondisjunction are the mechanisms most often associated with mitotic aneuploidy, but chromosome demolition, endoreplication, cytokinesis errors, cell fusion, and chromosome breakage also have been implicated (10). A paternal effect may also be involved with mosaicism, evidenced by data demonstrating that embryos derived after testicular sperm extraction (TESE) from males with nonobstructive azoospermia have higher rates of mosaicism compared with embryos from ejaculated sperm (26). When paternal age was disentangled from maternal age, we did not see a paternal age effect on either aneuploidy or mosaicism, although TESE specimens were not specifically examined in our cohort. Previous reports have examined the relationship between chromosome structure and the incidence of aneuploidy per chromosome. Franasiak et al. concluded that short and medium acrocentric chromosomes were the most likely to have errors in copy number, and the constitutional aneuploidy rate in those chromosomes rose disproportionately with age, although they did not report mosaicism or mitotic error rates (13). McCoy et al. reported a negative correlation between meiotic error and chromosome length (15). They found a positive correlation with mitotic errors and chromosome length in day 3 blastomeres, but not in TE samples. In our cohort, acrocentric chromosomes were found to have the highest frequency of aneuploidy and mosaicism per chromosome, but the finding was statistically significant only for aneuploidy. However, when whole-chromosome and segmental mosaicisms were considered separately, whole-chromosome mosaicisms were statistically more common in acrocentric chromosomes. Another interesting observation in our data was that some longer chromosomes, particularly 1, 2, 5, 9, and X (Fig. 2), are more frequently affected by segmental rather than whole-chromosome mosaicisms. It could be speculated that longer chromosomes are more prone to breakage during somatic division, yet without attachment to the centromere these errors may lead to cell cycle arrest rather than perpetuating segmental aneuploidy in successive divisions (17). Alternatively, higher rates of segmental errors in larger chromosomes could be the result of technical artifacts related to WGA or biased coverage of the larger chromosomes during NGS. Detection and reporting of mosaicism is dependent on the technical limitations of the cytogenetic method and the bioinformatic interpretation. Mitotic aneuploidies were initially reported with the use of karyotype analysis, but the technique was limited by the requirement for dividing metaphase cells, resulting in fewer than 25% of embryos that could be analyzed (2). FISH also has the ability to detect mosaicism, but sequential hybridization with the use of FISH has diminishing efficiency with each round, with predictive values for 862
abnormal results as low as 41% when even five chromosomes are investigated (27, 28). Furthermore, FISH can not generally detect segmental aneuploidy, because probes are specific for the particular locus being tested (29). Newer methods allow reliable 24-chromosome copy number analysis, but SNP array and aCGH rely on DNA amplification, which may introduce bias due to excessive amplification of specific regions or, in the case of aCGH, failure of amplification of one parental allele (30, 31). In contrast, mosaicism can be reliably detected only with the use of SNP array when 50% of cells are aneuploid (32, 33). NGS has a broader dynamic range for copy number detection, increasing the sensitivity for mosaicism (23). However, a randomized blinded trial comparing quantitative PCR and NGS did not show a difference in sensitivity to mosaicism with the use of default commercial settings when whole-chromosome mosaicisms were examined (34). Determination of whole-chromosome and segmental mosaicisms on the Veriseq PGS system was validated by analyzing aneuploid-euploid DNA ratios ranging from 0:5 to 5:0, based on the assumption that a typical TE biopsy contains five cells (9, 17). With little overlap between the various admixtures, the percentage of normal cells were reliably detected. When copy number falls between disomy and aneusomy, mosaicism is inferred. Ratios >80% or <20% can not be detected in a five-cell sample, but it is generally agreed that results beyond those ranges can be considered to be aneuploid or euploid, respectively (6, 12). However, it has been suggested that under certain conditions, the finding of mosaicism may be artifactual or limited by a high false positive rate (11, 34). Segmental aneuploidies and mosaicisms detected with the use of NGS have been confirmed with whole-blastocyst FISH analysis (35). Segmental mosaicisms can be detected with a high degree of sensitivity with the use of the Veriseq system owing to better coverage of the genome compared with targeted amplification methods, but that may come at the expense of lower specificity resulting in more false positive results (36). Furthermore, the validation of mosaicism with admixture studies relied on fibroblast and TE cell lines with aneuploidies of specific chromosomes, namely, chromosomes 13, 18, and 21 for whole gains, chromosome X for whole losses, chromosomes 10p and13q for partial losses, and specific deletions and duplications (6). Implicit to the conclusions is that detection of these errors would be equivalent regardless of the chromosome affected, however, detection thresholds for specific chromosomes may be variable depending on the technique used (37), thus limiting the ability to categorically define mosaicism, and subsequently affecting the interpretation of clinically reported results, such as those reported here. Assuming that mosaicism is not a technical artifact, the clinical relevance is ultimately determined by its implications for healthy live births. Several reports have documented live births from transfer of mosaic embryos, although implantation failure and miscarriage occur at significantly greater rates compared with euploid embryo transfers (8, 9, 17). Embryos with segmental mosaicisms have been found to have ongoing pregnancy rates equivalent to euploid blastocysts (17). The largest reported series to date, details VOL. 109 NO. 5 / MAY 2018
Fertility and Sterility® the transfer outcomes (but not live birth outcomes) from 143 transfers of mosaic embryos, documenting a 41% (58/143) ongoing implantation rate (6). In that study, complex mosaicisms (three or more mosaic chromosomes) had the lowest implantation rate, but segmental, single, and double mosaicisms had statistically similar outcomes. They did not demonstrate a difference in outcomes when 20%–40% mosaicism was compared with >40%, but that was attributed to small sample size. It seems clear that some mosaic embryos may lead to ongoing pregnancy. However, it is important to emphasize that longitudinal follow-up on the health of the children resulting from mosaic embryos has yet to be performed. Evidence suggests that mosaicism may be associated with numerous pathologic conditions, including autism, cancer, and organ-specific pathology (38, 39). Even when aneuploid cells are limited to the TE with a euploid ICM, as is the case with confined placental mosaicism, intrauterine growth restriction and fetal death may result (40). In pediatric diseases that warrant genetic testing, up to 2% can be attributed to mosaicism (41, 42). Mosaicism can also result in viable embryos with uniparental disomy (UPD) due to monosomy or trisomy rescue of a meiotic aneuploidy, which has significant disease implications (1). Such issues notwithstanding, further studies are required to determine which embryonic mosaicisms are harbingers of clinical sequelae and which are benign. Some guidance has been offered regarding which mosaic embryos can be considered as safest for transfer by the PDGIS (7). Mosaic monosomy was stated to be preferable to mosaic trisomy, because monosomy is not commonly observed as a clinical syndrome (other than 45 X0). However, caution with mosaic monosomy has been suggested because reciprocal trisomy may exist from mitotic nondisjunction (43). Furthermore, the infrequent documentation of live births with autosomal monosomy may represent a detection failure, because low level mosaic monosomy may be below the detection limit of microarray techniques but have nevertheless been associated with significant pathology (44). Munne et al. did not show a difference in implantation rates when mosaic monosomy and trisomy were compared (6). Our data showed a higher prevalence of mosaic trisomy when whole chromosomes were involved and of mosaic monosomy when segmental lesions were involved. Segmental mosaicisms were not addressed by the PGDIS position statement, but limited data suggest that live birth rates are equivalent to those of euploid transfers (17). It has also been documented that segmental mosaicisms can result in pathologic outcomes attributable to segmental UPD or mosaicisms for insertions, deletions, and point mutations (40). Another suggested criteria for consideration of mosaic embryo transfer is the percentage of aneuploid cells (7), which has been found to correlate with embryo viability (45). However, in patients found to have mosaicism in cultured lymphocytes, the proportion of aneuploid cells did not correlate with the severity of the phenotype, and low-level mosaicism may still be associated with pathology (44). The level of mosaicism is not addressed in our data. Needless to say, the data on outcomes of mosaic embryo transfers are still very preliminary and will not likely VOL. 109 NO. 5 / MAY 2018
be satisfactorily addressed for the many years required to determine if pathology manifests over the long term. There are several other issues that limit the interpretation of the present findings. This study was retrospective and observational, demonstrating the relative frequency of chromosome-specific mosaicisms but not offering any direct insight into the clinical relevance of the findings. Ultimately, only a large-scale, randomized, nonselection study would truly resolve the concerns. In addition, as has been noted in other studies of this nature (13), the number of embryos tested per patient is variable, thus potentially skewing the findings as a function of unequal contributions from individuals. However, a logistic regression model that included patient effect did not suggest an intrinsic dependence on results for mosaicism. Furthermore, our data can not necessarily be generalized, because it has been documented that there could be significant variability introduced by iatrogenic factors, such as the excessive force during the biopsy process, or the presence of dead or apoptotic cells could lead to artifactual loss or gain of chromosomes (23). Finally, questions persist regarding the technical limitations that may affect the diagnostic accuracy of mosaicism detection with currently available NGS-PGS platforms, challenging the interpretation of these findings (11, 34, 36). It has been suggested that the highest level of evidence of mosaicism would be demonstrated from reciprocal errors found in double biopsies of a blastocyst (37). Unfortunately, that standard may be impractical to apply clinically, so clinicians are still faced with the challenge of interpreting results as they are reported, which is based on the intermediate log-2 ratios. The data presented here are insufficient to draw any clinical conclusions or to validate or refute the technical aspects. However, such data may assist in patient counseling and spur future research by comparing the prevalence of chromosome-specific mosaicism in blastocysts with those associated with pathologic findings.
CONCLUSION Our data illustrate the differences in detection of mosaicism compared with constitutional aneuploidy with the use of NGS, suggesting that mitotic and meiotic error have characteristic distributions across the genome that vary according to the specific chromosome. In addition, whole-chromosome and segmental mosaicisms are differentially distributed, potentially suggesting chromosome-specific susceptibilities to mitotic error. Alternatively, the lower prevalence of certain mosaicisms in TE biopsies may be interpreted as a lower tolerance of errors in gene-rich regions which reduces survival to the blastocyst stage (15). Additional research is required to determine why certain chromosomes seem more predisposed to meiotic or mitotic error, and why segmental and whole-chromosome mitotic errors occur with preponderance for particular chromosomes. The contribution of technical limitations and artifact associated with NGS also can not be ruled out as a source of confusion, potentially resulting in false positive mosaic results. Larger-scale multicenter studies could confirm if our findings are robust and, most importantly, if 863
ORIGINAL ARTICLE: GENETICS specific embryonic mosaicisms are consistent with healthy live birth.
20.
REFERENCES
21.
1. 2.
3.
4.
5.
6.
7.
8. 9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
864
Biesecker LG, Spinner NB. A genomic view of mosaicism and human disease. Nat Rev Genet 2013;14:307–20. Clouston HJ, Fenwick J, Webb AL, Herbert M, Murdoch A, Wolstenholme J. Detection of mosaic and nonmosaic chromosome abnormalities in 6- to 8day-old human blastocysts. Hum Genet 1997;101:30–6. Delhanty JDA, Griffin DK, Handyside AH, Harper J, Atkinson GHG, Pieters MHEC, et al. Detection of aneuploidy and chromosomal mosaicism in human embryos during preimplantation sex determination by fluorescent in situ hybridisation, (FISH). Hum Mol Genet 1993;2:1183–5. Zheng H, Jin H, Liu L, Liu J, Wang WH. Application of next-generation sequencing for 24-chromosome aneuploidy screening of human preimplantation embryos. Mol Cytogenet 2015;8:38. Tormasi S, Capaldi R, Gouw F, Munne S, Wells D, Fragouli E. The use of next generation sequencing to determine the developmental potential of mosaic blastocysts. Fertil Steril 2016;105(2 Suppl 1):e41. Munn e S, Blazek J, Large M, Martinez-Ortiz PA, Nisson H, Liu E, et al. Detailed investigation into the cytogenetic constitution and pregnancy outcome of replacing mosaic blastocysts detected with the use of highresolution next-generation sequencing. Fertil Steril 2017;108:62–71.e8. Preimplantation Genetic Diagnosis International Society. PGDIS position statement on chromosome mosaicism and preimplantation aneuploidy testing at the blastocyst stage. July 19, 2016. Available at: http://pgdis. org/docs/newsletter_071816.html. Accessed September 4, 2017. Greco E, Minasi MG, Fiorentino F. Healthy babies after intrauterine transfer of mosaic aneuploid blastocysts. N Engl J Med 2015;373:2089–90. Maxwell SM, Colls P, Hodes-Wertz B, McCulloh DH, McCaffrey C, Wells D, et al. Why do euploid embryos miscarry? A case-control study comparing the rate of aneuploidy within presumed euploid embryos that resulted in miscarriage or live birth using next-generation sequencing. Fertil Steril 2016;106: 1414–9.e5. Mantikou E, Wong KM, Repping S, Mastenbroek S. Molecular origin of mitotic aneuploidies in preimplantation embryos. Biochim Biophys Acta 2012;1822:1921–30. Esfandiari N, Bunnell ME, Casper RF. Human embryo mosaicism: did we drop the ball on chromosomal testing? J Assist Reprod Genet 2016;33: 1439–44. Besser AG, Mounts EL. Counselling considerations for chromosomal mosaicism detected by preimplantation genetic screening. Reprod Biomed Online 2017;34:369–74. Franasiak JM, Forman EJ, Hong KH, Werner MD, Upham KM, Treff NR, et al. Aneuploidy across individual chromosomes at the embryonic level in trophectoderm biopsies: changes with patient age and chromosome structure. J Assist Reprod Genet 2014;31:1501–9. Fragouli E, Alfarawati S, Spath K, Jaroudi S, Sarasa J, Enciso M, et al. The origin and impact of embryonic aneuploidy. Hum Genet 2013;132:1001– 13. McCoy RC, Demko ZP, Ryan A, Banjevic M, Hill M, Sigurjonsson S, et al. Evidence of selection against complex mitotic-origin aneuploidy during preimplantation development. PLoS Genet 2015;11:e1005601. Gardner DK, Lane M, Stevens J, Schlenker T, Schoolcraft WB. Blastocyst score affects implantation and pregnancy outcome: toward a single blastocyst transfer. Fertil Steril 2000;73:1155–8. Fragouli E, Alfarawati S, Spath K, Babariya D, Tarozzi N, Borini A, et al. Analysis of implantation and ongoing pregnancy rates following the transfer of mosaic diploid-aneuploid blastocysts. Hum Genet 2017;2017:805–13. Martínez MC, M endez C, Ferro J, Nicolas M, Serra V, Landeras J. Cytogenetic analysis of early nonviable pregnancies after assisted reproduction treatment. Fertil Steril 2010;93:289–92. Wellesley D, Dolk H, Boyd PA, Greenlees R, Haeusler M, Nelen V, et al. Rare chromosome abnormalities, prevalence and prenatal diagnosis rates from population-based congenital anomaly registers in Europe. Eur J Hum Genet 2012;20:521–6.
22.
23. 24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
Munne S, Magli C, Adler A, Wright G, de Boer K, Mortimer D, et al. Treatment-related chromosome abnormalities in human embryos. Hum Reprod 1997;12:780–4. Menasha J, Levy B, Hirschhorn K, Kardon NB. Incidence and spectrum of chromosome abnormalities in spontaneous abortions: new insights from a 12-year study. Genet Med 2005;7:251–63. Antonarakis SE, Avramopoulos D, Blouin JL, Talbot CC, Schinzel AA. Mitotic errors in somatic cells cause trisomy 21 in about 4.5% of cases and are not associated with advanced maternal age. Nat Genet 1993;3:146–50. Munne S, Wells D. Detection of mosaicism at blastocyst stage with the use of high-resolution next-generation sequencing. Fertil Steril 2017:1085–91. Braude P, Bolton V, Moore S. Human gene expression first occurs between the four- and eight-cell stages of preimplantation development. Nature 1988;332:459–61. Katz-Jaffe MG, Trounson AO, Cram DS. Chromosome 21 mosaic human preimplantation embryos predominantly arise from diploid conceptions. Fertil Steril 2005;84:634–43. Silber S, Escudero T, Lenahan K, Abdelhadi I, Kilani Z, Munne S. Chromosomal abnormalities in embryos derived from testicular sperm extraction. Fertil Steril 2003;79:30–8. Liu J, Tsai YL, Zheng XZ, Baramki TA, Yazigi RA, Katz E. Potential use of repeated fluorescence in situ hybridization in the same human blastomeres for preimplantation genetic diagnosis. Fertil Steril 1998;70:729–33. Scriven PN, Bossuyt PMM. Diagnostic accuracy: Theoretical models for preimplantation genetic testing of a single nucleus using the fluorescence in situ hybridization technique. Hum Reprod 2010;25:2622–8. Wilton L, Voullaire L, Sargeant P, Williamson R, McBain J. Preimplantation aneuploidy screening using comparative genomic hybridization or fluorescence in situ hybridization of embryos from patients with recurrent implantation failure. Fertil Steril 2003;80:860–8. Handyside AH, Robinson MD, Simpson RJ, Omar MB, Shaw MA, Grudzinskas JG, et al. Isothermal whole genome amplification from single and small numbers of cells: a new era for preimplantation genetic diagnosis of inherited disease. Mol Hum Reprod 2004;10:767–72. Iwamoto K, Bundo M, Ueda J, Nakano Y, Ukai W, Hashimoto E, et al. Detection of chromosomal structural alterations in single cells by SNP arrays: a systematic survey of amplification bias and optimized workflow. PLoS One 2007;2:e1306. Treff NR, Levy B, Su J, Northrop LE, Tao X, Scott RT. SNP microarray–based 24 chromosome aneuploidy screening is significantly more consistent than FISH. Mol Hum Reprod 2010;16:583–9. Northrop LE, Treff NR, Levy B, Scott JT. SNP microarray–based 24 chromosome aneuploidy screening demonstrates that cleavage-stage FISH poorly predicts aneuploidy in embryos that develop to morphologically normal blastocysts. Mol Hum Reprod 2010;16:590–600. Goodrich D, Tao X, Bohrer C, Lonczak A, Xing T, Zimmerman R, et al. A randomized and blinded comparison of qPCR and NGS-based detection of aneuploidy in a cell line mixture model of blastocyst biopsy mosaicism. J Assist Reprod Genet 2016;33:1473–80. Vera-Rodríguez M, Michel CE, Mercader A, Bladon AJ, Rodrigo L, Kokocinski F, et al. Distribution patterns of segmental aneuploidies in human blastocysts identified by next-generation sequencing. Fertil Steril 2016;105:1047–55.e2. Goodrich D, Xing T, Tao X, Lonczak A, Zhan Y, Landis J, et al. Evaluation of comprehensive chromosome screening platforms for the detection of mosaic segmental aneuploidy. J Assist Reprod Genet 2017;34:975–81. Capalbo A, Ubaldi FM, Rienzi L, Scott R, Treff N. Detecting mosaicism in trophectoderm biopsies: current challenges and future possibilities. Hum Reprod 2016;32:492–8. Yurov YB, Vorsanova SG, Iourov IY, Demidova IA, Beresheva AK, Kravetz VS, et al. Unexplained autism is frequently associated with low-level mosaic aneuploidy. J Med Genet 2007;44:521–5. Jacobs KB, Yeager M, Zhou W, Wacholder S, Wang Z, Rodriguez-Santiago B, et al. Detectable clonal mosaicism and its relationship to aging and cancer. Nat Genet 2012;44:651–8. Spinner NB, Conlin LK. Mosaicism and clinical genetics. Am J Med Genet C Semin Med Genet 2014;166:397–405.
VOL. 109 NO. 5 / MAY 2018
Fertility and Sterility® 41.
42.
43.
Ballif BC, Rorem EA, Sundin K, Lincicum M, Gaskin S, Coppinger J, et al. Detection of low-level mosaicism by array CGH in routine diagnostic specimens. Am J Med Genet A 2006;140:2757–67. Cheung SW, Shaw CA, Scott DA, Patel A, Sahoo T, Bacino CA, et al. Microarray-based CGH detects chromosomal mosaicism not revealed by conventional cytogenetics. Am J Med Genet A 2007;143:1679–86. Scott RT, Galliano D. The challenge of embryonic mosaicism in preimplantation genetic screening. Fertil Steril 2016;105:1150–2.
VOL. 109 NO. 5 / MAY 2018
44.
45.
Hochstenbach R, Krijtenburg PJ, van der Veken LT, van der Smagt J, Roeleveld-Versteegh A, Visser G, et al. Monosomy 20 mosaicism revealed by extensive karyotyping in blood and skin cells: Case report and review of the literature. Cytogenet Genome Res 2014;144:155–62. Bolton H, Graham SJL, Van der Aa N, Kumar P, Theunis K, Fernandez Gallardo E, et al. Mouse model of chromosome mosaicism reveals lineage-specific depletion of aneuploid cells and normal developmental potential. Nat Commun 2016;7:11165.
865