Human Reproduction, Vol.31, No.2 pp. 312–323, 2016
Advanced Access publication on November 29, 2015 doi:10.1093/humrep/dev281
ORIGINAL ARTICLE Embryology
Human embryos commonly form
abnormal nuclei during development: a
mechanism of DNA damage, embryonic
aneuploidy, and developmental arrest
Daniel H. Kort 1,6,†, Gloryn Chia2,†, Nathan R. Treff3,4, Akemi J. Tanaka 2,
Tongji Xing 4, Lauren Bauer Vensand 5, Stephanie Micucci 2,
Robert Prosser 1, Roger A. Lobo 1, Mark V. Sauer 1, and Dieter Egli 2,5,*
1
Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology and Infertility, Columbia University College of Physicians
and Surgeons, 622 W. 168th Street, PH-16, New York, NY 10032, USA 2Department of Pediatrics, Columbia University of Physicians and
Surgeons, 1150 St. Nicholas Avenue, New York, NY 10032, USA 3Reproductive Medicine Associates, New Jersey, 140 Allen Road, Basking Ridge,
NJ 07920, USA 4Department of Microbiology and Molecular Genetics, Rutgers-Robert Wood Johnson Medical School, Piscataway,
NJ 08854-5635, USA 5The New York Stem Cell Foundation Research Institute, New York, NY 10032, USA 6Present address: Monmouth Medical
Center, Damien Fertility Partners, Shrewsbury, NJ 07702, USA
*Correspondence address. E-mail: [email protected] (D.E.)
Submitted on December 27, 2014; resubmitted on September 11, 2015; accepted on October 19, 2015
study question: What is the prevalence and developmental significance of morphologic nuclear abnormalities in human preimplantation
embryos?
summary answer: Nuclear abnormalities are commonly found in human IVF embryos and are associated with DNA damage, aneuploidy,
and decreased developmental potential.
what is known already: Early human embryonic development is complicated by genomic errors that occur after fertilization. The
appearance of extra-nuclear DNA, which has been observed in IVF, may be a result of such errors. However, the mechanism by which abnormal
nuclei form and the impact on DNA integrity and embryonic development is not understood.
study design, size, duration: Cryopreserved human cleavage-stage embryos (n ¼ 150) and cryopreserved blastocysts (n ¼ 105)
from clinical IVF cycles performed between 1997 and 2008 were donated for research. Fresh embryos (n ¼ 60) of poor quality that were slated for
discard were also used. Immunohistochemical, microscopic and cytogenetic analyses at different developmental stages and morphologic grades
were performed.
participants/materials, setting, methods: Embryos were fixed and stained for DNA, centromeres, mitotic activity and
DNA damage and imaged using confocal microscopy. Rates of abnormal nuclear formation were compared between morphologically normal
cleavage-stage embryos, morphologically normal blastocysts, and poor quality embryos. To control for clinical and IVF history of oocytes
donors, and quality of frozen embryos within our sample, cleavage-stage embryos (n ¼ 52) were thawed and fixed at different stages of development and then analyzed microscopically. Cleavage-stage embryos (n ¼ 9) were thawed and all blastomeres (n ¼ 62) were disaggregated,
imaged and analyzed for karyotype. Correlations were made between microscopic and cytogenetic findings of individual blastomeres and
whole embryos.
main results and the role of chance: The frequency of microscopic nuclear abnormalities was lower in blastocysts (5%; 177/
3737 cells) than in cleavage-stage embryos (16%, 103/640 blastomeres, P , 0.05) and highest in arrested embryos (65%; 44/68 blastomeres,
P , 0.05). DNA damage was significantly higher in cells with microscopic nuclear abnormalities (gH2AX (phosphorylated (Ser139) histone
H2A.X): 87.1%, 74/85; replication protein A: 72.9%, 62/85) relative to cells with normal nuclear morphology (gH2AX: 9.3%, 60/642; RPA:
5.6%, 36/642) (P , 0.05). Blastomeres containing nuclear abnormalities were strongly associated with aneuploidy (Fisher exact test,
two-tailed, P , 0.01).
†
The authors consider that the first two authors should be regarded as joint First Authors.
& The Author 2015. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please email: [email protected]
313
Abnormal nuclear formation in early human development
limitations, reasons for caution: The embryos used were de-identified, and the clinical and IVF history was unknown.
wider implications of the findings: This study explores a mechanism of abnormal embryonic development post-fertilization.
While most of the current data have explored abnormal meiotic chromosome segregation in oocytes as a primary mechanism of reproductive
failure, abnormal nuclear formation during early mitotic cell division in IVF embryos also plays a significant role. The detection of abnormal nuclear
formation may have clinical application in noninvasive embryo selection during IVF.
study funding/competing interest(s): The study was supported by Columbia University and the New York Stem Cell Foundation. Authors declare no competing interest.
Key words: mitosis / micronucleation / preimplantation development / DNA damage / mosaicism
Introduction
Despite over 30 years of scientific and clinical advances, assisted reproduction remains highly inefficient. The majority of embryos created via
assisted reproductive technologies are ultimately discarded, and most
of the embryos transferred fail to result in term delivery (Murray,
2004, 2008). While the etiology of IVF failure is multi-factorial, embryonic aneuploidy is suspected to be the dominant factor (Scott et al., 2012;
Franasiak et al., 2014). In fact, selection of euploid embryos using cytogenetic techniques has been shown to significantly improve implantation
and live born delivery rates per embryo transferred (Scott et al., 2013).
Although estimates of embryonic aneuploidy vary by patient population, laboratory and cytogenetic techniques, it is likely that over 50% of
preimplantation embryos are in fact aneuploid and incompatible with
successful pregnancy and live born delivery (Franasiak et al., 2014).
Such embryonic aneuploidy results from genomic errors(s) at one or
more stage(s) of development—oocyte meiosis I, oocyte meiosis II, fertilization, and mitosis (Fragouli et al., 2013).
Understanding when and how genomic errors occur is essential to the
field of reproductive medicine. Genomic errors that occur during oocyte
meiosis are a well-established cause of embryonic aneuploidy. Such
errors occur as a result of whole chromosome or chromatid segregation
failure in meiosis I or II (Fragouli et al., 2006, 2013) with error rates directly related to maternal age.
In addition to errors in meiosis, genomic errors may occur during the
post-fertilization mitotic divisions, resulting in embryonic mosaicism
(Capalbo et al., 2013; Mertzanidou et al., 2013). For example, when multiple blastomeres within 14 cleavage-stage embryos from young patients
with successful IVF cycles were analyzed, 71% (10/14) of the embryos
were interpreted as mosaic, although variation due to technical artifacts
was not excluded (Mertzanidou et al., 2013). In a different study, developing embryos were analyzed by sequential polar body, blastomere, and
trophoectoderm biopsy. Nearly 50% (10/21) of aneuploid blastocysts
resulted from chromosomal errors that occurred post-fertilization
(Capalbo et al., 2013).
Although embryonic mosaicism is evident, the developmental potential of mosaic embryos, or the ability of an embryo to ‘tolerate’ certain
levels of mosaicism, is unknown. Furthermore, it is unclear whether segregation errors during mitosis occur via the same mechanism as the segregation errors during meiosis.
Mitotic errors are difficult to observe in the clinical IVF setting as such
studies require destruction of the embryo and genotyping of individual
blastomeres. However, mitotic errors can manifest in the appearance
of extra-nuclear DNA, which can be observed microscopically (Alikani
et al., 2000; Wong et al., 2010; Chavez et al., 2012). Depending on the
size, shape and time of observation, these nuclear abnormalities are
described as micronuclei (small nucleus, separate from main nucleus),
multinuclei (.1 nucleus of similar size within a cell) and nucleoplasmic
bridges (oblong formation of extra-nuclear DNA between two separate
nuclei). These abnormalities are associated with an increased rate of
mitotic chromosomal abnormalities (Kligman et al., 1996) and correlate
negatively with blastocyst development, cell number and implantation
rate (Hardy et al., 1989; Pelinck et al., 1998; Alikani et al., 2000; Royen
et al., 2003; Meriano et al., 2004). Therefore, embryos with such abnormalities are generally not transferred.
Although abnormal nuclei have been observed within embryos, the
mechanism by which such abnormal nuclei form is unknown, and it is
unclear whether this phenomenon results in mitotic aneuploidy and mosaicism. Proposed mechanisms of abnormal nucleation include cleavage
failure (Hardy et al., 1993) or abnormal packaging of DNA after completion of mitosis. (Pickering et al., 1995) However, the majority of studies to
date have used bright-field microscopy, which can only distinguish large
nuclei from the surrounding cytoplasm and cannot provide accurate
quantification and mechanistic data.
Using immunohistochemical techniques, high-resolution confocal microscopy and cytogenetic analyses, we quantified and characterized
nuclear abnormalities in human preimplantation embryos of different developmental stages and morphologic grades. We show that such nuclear
abnormalities are commonly found in early human embryos and are
strongly associated with poor developmental potential, DNA damage
and the presence of centromere-less chromosomal fragments. Cytogenetic analysis following blastomere staining for DNA revealed that blastomeres with micronuclei contained (segmental) mosaicism. Therefore,
the formation of micronuclei represents a mechanism of mitotic aneuploidy, affects DNA integrity, and contributes to developmental arrest
in early human embryos.
Methods
Research samples
All protocols and procedures were approved by the Columbia University
Institutional Review Board and by the Columbia University Stem Cell
Committee.
Cryopreserved cleavage-stage embryos (n ¼ 150) and cryopreserved
blastocysts (n ¼ 105) from IVF cycles performed at the Center for
Women’s Reproductive Care at Columbia University between 1997 and
2008 were analyzed (stored in liquid nitrogen at less than 21968C). Cryopreserved embryos were given the grade of ‘good’ quality as assessed by standard grading criteria based on (cleavage stage) cell number, fragmentation and
symmetry or (blastocyst) expansion, inner cell mass and trophoectoderm
314
(Racowsky et al., 2010). Criteria for cleavage-stage embryos included 6 – 10
cells by late morning/early afternoon of day 3, ,5% fragmentation and symmetric blastomeres. Freezing criteria remained consistent between 1997 and
2008. Vitrification (fast-cooling) was used for cryopreservation. All embryos
were donated for research during a formal informed consent process, including signed documentation. Embryos were thawed using the Quinn’s Advantagew thaw kit (Sage IVF, Trumbull, CT, USA) and fixed immediately
post-thaw and prepared for imaging. No embryo culture was performed
to exclude any potential effect of cryopreservation on physiologic events
downstream of the thaw.
Fresh embryos (n ¼ 60) deemed to be ‘poor’ (Racowsky et al., 2010)
quality during in vitro development were used. These embryos did not
meet the criteria for transfer or cryopreservation at any time point and are
the complementary group of the above cryopreserved embryos. Embryos
were diagnosed as either arrested (Day 3) or poor quality (Say 6), and they
were not suitable for clinical use.
An additional cohort of cryopreserved ‘good’ quality cleavage-stage
embryos (n ¼ 52) were thawed. A total of 25 embryos were fixed immediately, and 27 allowed to develop in culture to the blastocyst stage. Embryos of
different stage and developmental potential from 11 donors were compared
through assessment of microscopic morphologic nuclear abnormalities and
markers of DNA damage.
Kort et al.
Cytogenetic analysis
Nine embryos were selected for cytogenetic analysis. Embryos were thawed,
plasma membranes were allowed to stabilize for 1 – 2 h in GlobalTotal (LifeGlobal, Guilford, CT, USA) and dissociated in PGD medium (LifeGlobal,
Guilford, CT, USA). Dissociated blastomeres were stained with Hoechst
33342 (0.1 mg/ml) for 5 – 10 min in KOH solution containing 0.01% Triton
X-100, washed three times in KOH solution and imaged using an Olympus
IX73 (Olympus, Center Valley, PA, USA). Individual blastomeres were
placed into tubes containing 1 ml KOH solution. Whole-genome amplification and single-nucleotide polymorphism (SNP) microarray analysis was performed as previously described (Treff et al., 2010). Correlations were made
between microscopic and cytogenetic findings of individual blastomeres and
whole embryos.
The GEO accession number is GSE72150.
Statistics
Statistics tests were performed on Stata/MP 13 (Stata Corp LP, College
Station, TX, USA) and GraphPad Prism 6 (GraphPad Software, Inc., La
Jolla, CA, USA) statistical and data analysis software. A value of P , 0.05
was considered significant.
Results
Immunohistochemistry of cryopreserved
human embryos
After thaw (cryopreserved samples) or collection (fresh samples), embryos
were placed in 0.5% Triton/phosphate-buffered solution (0.5% PBS-T) for
10 s, fixed in 4% paraformaldehyde for 20 min and serially washed (×2) in
PBS, all at room temperature. Embryos were blocked with 10% fetal
bovine serum in PBS for 2 h and then placed in 10 ml of primary antibody
solution (anti-centromere IgG dilution 1:100 (Antibodies, Inc. #15-235-0001),
anti-beta-tubulin (Millipore, #05-661) at dilution 1:1000 and anti-phosphoHistone 3 (Ser10) (Millipore #06-570) at dilution 1:2000) in a plastic tissue
culture dish for 2 h at room temperature or for 18 h at 48C. Embryos were
serially washed (×2) and transferred to a solution of secondary antibody
(Life Technologies, dilution 1:500) and Hoechst (dilution 1:100) for 2 h.
Embryos were serially washed (×2) and placed in 10 ml droplets of 1%
bovine serum albumin on an uncoated 30-mm glass dish (P50G-1.5-30-F
MatTek Corp., Ashland, MA, USA) covered with embryo culture oil (Irving
Scientific, Santa Ana, CA, USA).
Imaging and analysis
Confocal microscopy was performed using a Zeiss LSM (LSM 510) at magnification of 63× and analyzed using Zeiss Zen software. Images were
obtained using 355-, 488-, 555- and 633-nm lasers in addition to bright-field
microscopy. The number of confocal sections and their thickness was
adjusted to image all of the embryo. For a typical cleavage-stage embryo,
40 sections were taken of 1.8 – 2.1 micrometer thickness with overlap
between confocal sections. Cell numbers were counted manually. Nuclei
and extra-nuclear DNA were identified by the presence of Hoechst staining
over background (minimal). Nuclear size was quantified by measurement
of the horizontal diameter of the Hoechst positive mass. As nuclei are circular, diameter is one possible measure of the size. The presence or
absence of a centromere was assessed by the presence of human
anti-centromere antibody staining above nuclear background. The presence of mitosis was assessed by both the presence of phospho-histone
H3 (p-H3) staining and cellular morphologic appearance. The presence
of the mitotic spindle was assessed by the presence of beta-tubulin
above cytoplasmic background.
Microscopic nuclear abnormalities are
prevalent in good-quality human
cleavage-stage embryos
Eighty nine ‘good’ quality cleavage-stage embryos containing 640 blastomeres (average 7.2 blastomeres/embryo) were analyzed (Supplementary Video S1). Forty-eight percent (43/89) of embryos contained
blastomeres with two or more nuclei (Fig. 1a –f), while 52% (46/89)
of embryos had a single normal appearing nucleus in all blastomeres.
When analyzed in aggregate, 16% (103/640) of blastomeres from
cleavage-stage embryos were found to contain two or more nuclei
(Fig. 1e).
To quantify the degree of nuclear abnormalities within each embryo,
the frequency of nuclear abnormalities per embryo (number abnormal
blastomeres of embryo/total number of blastomeres of embryo)
was calculated. Of the embryos that contained abnormal blastomeres
(n ¼ 43), 49% (21/43) contained ,25% abnormal blastomeres, 28%
(12/43) contained 25 –50% abnormal blastomeres, 16% (7/43) contained 50 –75% abnormal blastomeres and 7% (3/43) contained
.75% abnormal blastomeres (Fig. 1f).
The prevalence of microscopic nuclear
abnormalities decreases with advanced
stage and developmental potential
To compare nuclear abnormality rates in early, non-selected, cleavagestage embryos with embryos that had successfully developed to the
blastocyst stage, 45 cryopreserved ‘good’ quality blastocysts containing
3737 cells (average 83 cells/blastocyst) were analyzed (Supplementary
Video S2, Fig. 1g and h). While the majority (38/45, 84.4%) of blastocysts
contained abnormal cells within the embryo, the frequency of abnormal
cells per embryo was low (Supplementary Fig. S1a). Fifty-three percent
(24/45) of blastocysts had an abnormality frequency (number of abnormal cells/total number of cells in embryo); ,5 and 91% (41/45) of blastocysts had an abnormality frequency of ,10% (Supplementary Fig. S1a).
Abnormal nuclear formation in early human development
315
Figure 1 Nuclear abnormalities and error rates in unselected human cleavage-stage embryos (a – f), blastocysts (g– i) and discarded arrested cleavagestage embryos (j – l). (a and b) Examples of abnormal nuclear formations in an unselected cleavage-stage embryo. Confocal sections. (c) Projection of the
nuclei of a single blastomere. (d) Number of unselected cleavage-stage embryos without (blue bar) or with one or more abnormal nuclei (red bar). (e)
Number of blastomeres with or without an abnormal nucleus. (f) Abnormality frequency per embryo. (g and h) Example of a nuclear abnormality in a blastocyst. (i) Number of normal and abnormally nucleated cells in high-quality blastocysts. (j and k) Example of an abnormal blastomere in an arrested cleavagestage embryo. (l) Quantification of normal (blue) and abnormal (red) blastomeres in arrested cleavage-stage embryos.
316
When analyzed in aggregate, 4.7% (177/3737) of cells within blastocysts
contained microscopic nuclear abnormalities (Fig. 1i).
To compare our findings in the two groups of ‘good’ quality embryos,
embryos slated for discard during in vitro development were obtained
from our IVF center for analysis. These embryos had either arrested at
the cleavage stage or progressed to the morula stage but failed to
become a normal blastocyst by Day 6. Embryos were diagnosed as
either arrested as a cleavage-stage embryo (≤8 cells) or poor quality
morulas/blastocyst-like (.8 cells) and not suitable for cryopreservation.
Thirteen arrested cleavage-stage embryos containing 71 blastomeres
(average 5.2 blastomeres/embryo) and 10 poor quality embryos containing 268 cells (average 27 cells/embryo) were analyzed using identical
protocols (Supplementary Video S3). Ninety-two percent (12/13) of
the arrested cleavage-stage embryos had nuclear abnormalities, and
the frequency of abnormalities per embryo was high (Fig. 1j – l, Supplementary Fig. S1b). When analyzed in aggregate, 65% (44/68) of blastomeres in arrested embryos contained microscopic nuclear abnormalities
(Fig. 1l). In the group of 10 failed embryos with disorganized growth past
the eight-cell stage but without normal cavitation, 10% (27/268) of
the cells contained microscopic abnormalities. Thus, the frequency
of abnormal nucleation was negatively associated (65 versus 10%,
P , 0.01) with the development even in poor-quality embryos
(Supplementary Fig. S1c).
Rates of microscopic nuclear abnormalities were compared between
groups of different developmental potential: embryos arrested at the ≤8
cell stage, embryos arrested with .8 cells, ‘good’ quality embryos
cryopreserved at the cleavage stage, and ‘good’ quality embryos cryopreserved at the blastocyst stage (Supplementary Fig. S1c). The frequency of microscopic nuclear abnormalities was lower in blastocysts
than in cleavage stage or arrested embryos and highest in embryos
arrested at the cleavage stage (P , 0.01).
To control for effects of clinical and IVF history of oocytes donors, and
quality of frozen embryos, 52 cleavage-stage embryos from 11 donors
were thawed and fixed at different stages of development. Embryos
from each donor were divided into two groups; 25 embryos (159 blastomeres) were fixed immediately after thaw (at cleavage stage), and 27
embryos were placed in culture media and were allowed to develop.
Of the 27 embryos allowed to develop, 14 embryos (78 blastomeres)
arrested at the cleavage stage and were thus fixed, 5 embryos (87
cells) developed to the morula stage and were also fixed, and 8
embryos (403 cells) developed to the blastocyst stage and were fixed.
These four groups with different developmental potential from the
same clinical source were analyzed microscopically.
Embryos that developed to the blastocyst stage had the lowest frequency of nuclear abnormalities (6.0%, 24/403, P , 0.05) (Supplementary Fig. S2). Embryos that developed to morula stage had a higher
frequency of nuclear abnormalities compared with blastocyst stage
embryos (13.8%, 12/87, P , 0.05) but a lower frequency of nuclear
abnormalities compared with arrested cleavage-stage embryos
(24.4%, 19/78, P , 0.05). Embryos that arrested at the cleavage stage
had the highest prevalence of nuclear abnormalities, significantly higher
than cleavage-stage embryos at thaw (18.9%, 30/159, P , 0.05).
These results are consistent with the above experiments using
embryos without controlling for clinical history. In both experiments,
the presence of microscopic nuclear abnormalities within blastomeres
of preimplantation embryos was negatively associated with developmental potential.
Kort et al.
Extra-nuclear DNA forms in mitosis
To directly determine whether extra-nuclear DNA originated at mitosis,
we examined mitotic cells in blastomeres and blastocysts. Though
mitotic cells were infrequent (,1%) in blastomeres, we found nuclear
DNA segregating abnormally into multiple nuclei in a mitotic blastomere
(Fig. 2a). In a morula-stage embryo, we found chromosomes exterior of
the microtubule spindle (Fig. 2b and c). Furthermore, we also observed
two nucleocytoplasmic bridges at interphase, evidence of incomplete
dissociation of sister chromatids at mitosis (Fig. 2d and e). In blastocysts,
we found a lost chromosome between the segregating chromosomes at
anaphase of mitosis (Fig. 2f) as well as misaligned metaphase (Fig. 2g –i).
Therefore, the errors found at interphase were also present and
likely originated at mitosis rather than from fragmentation of nuclei
at interphase.
Extra-nuclear DNA forms from
chromosomal fragments
In order to better characterize the mechanism by which microscopic
abnormalities develop, embryos from all three groups were stained
with anti-centromere protein CENP-A antibody (see Methods). All
nuclei of normal morphology contained centromeres, providing a positive control for the staining. The presence or absence of a centromere
within a microscopic abnormality distinguishes whole chromosomes
from chromosome fragments, and the mechanisms associated with
their formation.
Within the 43 ‘good’ (Racowsky et al., 2010) quality cleavage-stage
embryos containing nuclear abnormalities, 16.3% (7/43) contained
exclusively centromere-negative abnormalities (Fig. 3a), supporting a
mechanism of segmental chromosomal error resulting from DNA
breaks. A total of 34.9% (15/43) of the embryos contained exclusively
centromere-positive abnormalities (Fig. 3b), which may be a whole
chromosome error resulting from spindle attachment or a cohesion
error. However, segmental chromosomal errors may too have a centromere depending on the size of the segment and proximity of the break to
the centromere. A total of 48.8% (21/43) of the embryos contained both
centromere-negative and centromere-positive abnormalities (Fig. 3c),
demonstrating that different mechanisms of mitotic errors may occur
within the same embryo. When the 103 abnormal blastomeres were
analyzed in aggregate, 60% (62/103) contained centromere-positive abnormalities, 19.5% (20/103) contained centromere-negative abnormalities and 20.5% (21/103) contained both centromere-positive and
centromere-negative abnormalities. Therefore, data from centromere
staining suggest the presence of segmental errors as well as whole
chromosomal errors within cleavage-stage blastomeres.
Identical staining was performed on blastocysts to observe the
mechanisms of abnormal nucleation at the more advanced stage.
Within abnormal cells from ‘good’ (Racowsky et al., 2010) quality blastocysts, 55% (66/119) had centromere-positive abnormalities and
45% (53/119) had centromere-negative abnormalities (Fig. 3d and e).
While the overall frequency of nuclear abnormalities was low in blastocysts, the abnormalities observed were found to be both centromerepositive and centromere-negative. Therefore, similar mechanisms of abnormal nucleation, albeit at a lower frequency than in cleavage-stage
embryos, also occur in blastocysts.
In order to determine whether these same mechanisms exist within
highly abnormal embryos with little or no developmental potential,
Abnormal nuclear formation in early human development
317
Figure 2 Abnormal mitosis results in abnormal nucleation. Examples of errors in mitosis. (a) Mitotic blastomere and nuclear DNA segregating into multiple nuclei (arrowheads). (b and c) Chromosomes without centromere (arrowhead) excluded from microtubule spindle in morula-stage embryo (d and e)
Nucleocytoplasmic bridge at interphase in a morula-stage embryo. Note the thread of chromatin between nuclei and the presence of centromeres at both
junction points to the nucleus (arrowheads). The circled DNA is outside of the main nucleus, forming a mini-nucleus. (f) Anaphase at the blastocyst stage
with a lagging chromosome (arrowhead). (g– i) Confocal sections of a metaphase in a blastocyst cell. Note the bipolar arrangement of centromeres in one
section (g) and the misalignment in others (h and i). p-H3: phospho-histone H3.
identical staining was performed on embryos that had arrested at the
≤8-cell stage. Within this group, of the 44 blastomeres containing microscopic nuclear abnormalities, 55% (24/44) contained exclusively
centromere-positive abnormalities, 32% (14/44) contained exclusively
centromere-negative abnormalities and 14% (6/44) contained both
centromere-positive and centromere-negative abnormalities (Fig. 3f,
Supplementary Fig. S3a and b). All abnormalities, independent of the
type, decreased in incidence with developmental progression (Fig. 3f).
In addition, both centromere-negative (Fig. 2e –h) and centromerepositive (Fig. 2b) chromosomal errors were found at mitosis. Thus,
chromosomal segmental errors arise during mitosis and contribute significantly to the formation of microscopic nuclear abnormalities in
arrested embryos.
Depending on the type of error during mitosis, varying amounts of
extra-nuclear DNA may be present. In order to better estimate the
amount of DNA that is found in these abnormalities, the relative size
of each abnormality was calculated using the formula for relative size:
(maximum diameter of extra-nuclear DNA)/(maximum diameter of
nucleus). The relative size of the extra-nuclear DNA was plotted in
deciles (Supplementary Fig. S4a). This distribution was found to be
bimodal, concentrated at the high and low (micronucleation) frequencies. While large nuclei always contained a centromere (Fig. 3b), micronucleation was associated with a lack of centromere staining in
approximately half of the nuclei (Fig. 3f).
Depending on the size of the nucleus, abnormalities were categorized
into micronucleation (cell with one or two small additional nuclei, e.g.
Fig. 3a), binucleation (cell with two equally sized nuclei, e.g. Fig. 3b)
and multinucleation (cells with many small nuclei without a discernible
main nucleus, e.g. Fig. 1c). The frequency of these abnormalities was calculated in the 52 embryos thawed and fixed at different stages of development. While all these abnormalities could be observed at all stages of
preimplantation development, the frequency of micronucleation was
318
Figure 3 Abnormal nucleation is frequently associated with centromere loss. Examples of abnormal nucleation with or without centromere abnormalities in cleavage-stage embryos (a – c) and blastocysts (d and e). (f) Quantification of centromere-positive and -negative nuclear abnormalities at different developmental stages. Scale bars represent 10 mm.
Kort et al.
319
Abnormal nuclear formation in early human development
highest in newly thawed cleavage-stage embryos, while multinucleation
was highest in arrested cleavage-stage embryos, with a corresponding
decrease in micronucleation (Supplementary Fig. S4b). This suggests
that cells with micronuclei are unable to undergo normal mitosis, resulting in progressive fragmentation of nuclear DNA and developmental
arrest.
Microscopic nuclear abnormalities are
associated with DNA damage
Given the association of extra-nuclear DNA fragments and poor developmental potential, we predicted that the DNA found outside the
nucleus would be damaged. To test this hypothesis, 52 cleavage-stage
embryos that were thawed and fixed at different stages of development
were stained with a marker of DNA damage, phosphorylated (Ser139)
histone H2A.X (gH2AX) and replication protein A (RPA).
Of the 25 embryos (159 blastomeres) that were fixed immediately,
61/159 (38.4%) were positive for gH2AX while 46/159 (28.9%) were
positive for RPA (Fig. 4a). While embryos that arrested in culture
showed no decrease in the frequency of cells positive for gH2AX
(29/78, 37%) and RPA (23/78, 29.5%), developmental progression to
the morula and blastocyst stage led to a decrease in damaged cells.
Embryos that developed to the morula stage contained 15% (13/87)
gH2AX and 7% (6/87) RPA positive cells, while blastocysts contained
the lowest frequency of cells with damaged DNA: merely 7.7%
(31/403) of cells were positive for gH2AX, and 5.7% (23/403) for
RPA (Fig. 4a). Therefore, developmental progression correlated with a
lower frequency of DNA damage, in parallel to a reduction of abnormal
nucleation (P , 0.05).
To determine the relation between abnormal nucleation and DNA
damage, we examined blastomeres with normal and abnormal nuclear
morphology for markers of DNA damage at different developmental
stages. Independent of the developmental stage, most cells with abnormal nuclei showed markers of DNA damage, RPA and gH2AX (Fig. 4b).
DNA damage was significantly higher in cells with microscopic nuclear
abnormalities (gH2AX:87.1%, 74/85; RPA: 72.9%, 62/85) as compared
with cells with normal nuclear morphology (gH2AX: 9.3%, 60/642;
RPA: 5.6%, 36/642) (P , 0.05).
To determine whether DNA damage depended on the type of abnormal nucleation, we examined cells with two nuclei (binucleation), micronuclei, or multiple small nuclei (multinucleation) for markers of DNA
damage (Fig. 4c). Notably, all types of nuclear abnormalities were associated with DNA damage (Fig. 4d). As previously mentioned, cells with
normal nucleation exhibit lower frequency of gH2AX and RPA staining
than cells with abnormal nucleation (P , 0.05).
Microscopic nuclear abnormalities are
associated with mosaicism and segmental
aneuploidies
To determine whether microscopic nuclear abnormalities correlated
with aneuploidy, we performed DNA staining and microscopy, followed
by SNP array analysis of individual blastomeres. A total of 13 cleavagestage embryos underwent thaw, disaggregation, DNA staining and
imaging, followed by SNP array analysis of all blastomeres within the
embryo. Seventy-two of 93 (77%) blastomeres from 12 embryos were
successfully karyotyped (Supplementary Fig. S5). Only autosomes
were considered in the analysis of chromosomal aneuploidies. Four
blastomeres were called non-concurrent, which may be due to an artifact
of the method, or due to multiple segmental aneuploidies, and were not
further considered in the analysis. Reciprocal loss and gain were considered mitotic aneuploidies, while uniform gain or loss of chromosomes
was considered aneuploidies of meiotic origin.
Four embryos were found to have microscopically normal nuclei and
no evidence of micronuclei or multinuclei in all blastomeres. Three out of
four of morphologically normal embryos were 100% euploid, while one
was potentially mosaic with an indeterminate gain of chromosome 19 in
one blastomere. Overall, 1/24 (4%) karyotypically abnormal blastomere
due to mosaicism was found, while none contained meiotic aneuploidy.
Three embryos were found to have binucleation. All (3/3) binucleated
embryos contained aneuploidies of meiotic origin (chromosome gain or
loss in all blastomeres), in addition to a single-mitotic chromosome gain
in one blastomere (embryo 7, gain of chromosome 21). When calculated
per blastomere, 1/16 (0.6%) blastomeres contained mosaicism, while all
16/16 (100%) contained meiotic aneuploidy.
Four embryos contained micronucleation, of which three were successfully karyotyped in the relevant blastomeres. All (3/3) showed evidence of mitotic aneuploidy due to a reciprocal chromosomal gain/
loss between blastomeres (Fig. 5, Supplementary Fig. S5). Embryo #3
showed a reciprocal gain/loss of chromosome 15 between two sets of
blastomeres; two blastomeres with chromosome 15 gain contained
microscopically visible extra-nuclear DNA, while blastomeres with
chromosome loss did not (Fig. 5a and b). Embryo # 8 showed a similar
reciprocal gain/loss of chromosome 8 and 10 between blastomeres,
and the abnormality on chromosome 10 appeared to be segmental
with reciprocal loss and gain of 10q. Furthermore, embryo #9 showed
reciprocal segmental aneuploidies of chromosome 8 and chromosome
3, in addition to aneuploidy of meiotic origin on chromosomes 21 and
22. In total, all nine blastomeres (100%) containing micronuclei were aneuploid, eight contained a chromosome gain, and one a loss. Of the 13
blastomeres with normal nuclear morphology, 6 contained either a
chromosome gain or loss (46%). Overall, in embryos with micronucleation, 15/22 blastomeres (68%) contained chromosomal mosaicism. An
additional single embryo contained both micronucleation and binucleation with complex aneuploidy. Thus, the presence of micronuclei was
highly predictive for mitotic aneuploidy in our sample (Fisher exact
test, two-tailed, P , 0.01).
Discussion
Here, we demonstrate the presence of extra-nuclear DNA in a large
group of IVF human embryos at different stages of development. The
presence of extra-nuclear DNA was strongly associated with decreasing
developmental potential: abnormalities were infrequent in normally
developing blastocysts, intermediate in good quality cleavage-stage
embryos and highest in arrested embryos.
In further support of this observation, the frequency of microscopic
nuclear abnormalities is in line with current estimates of mosaicism in
both cleavage-stage embryos and blastocysts based on karyotypes
(Kligman et al., 1996; Wells and Delhanty, 2000; Fragouli and Wells,
2011; Mertzanidou et al., 2013). Remarkably, the frequency of nuclear
abnormalities described here and the frequency of karyotypically abnormal cells both decrease from cleavage stage to the blastocyst stage
(Barbash-Hazan et al., 2009). This can be explained by either a
genomic ‘correction’ that occurs with successful development or a
320
Kort et al.
Figure 4 Abnormal nucleation is associated with DNA damage. (a) Quantification of blastomeres with DNA damage (gH2AX: pink bars and RPA: red
bars) in IVF embryos from 11 donors thawed at cleavage stage and allowed to develop in culture. (b) Frequency of DNA damage (gH2AX and RPA) in
blastomeres with (red bars) and without (blue bars) abnormal nucleation at different developmental stages. *P-value , 0.05. (c) Representative images
of DNA damage, gH2AX (pink) and RPA (red), in normal blastomeres and blastomeres with micronucleation, binucleation and multinucleation. Scale
bars represent 10 mm. (d) Frequency of DNA damage in blastomeres with different subtypes of abnormal nucleation. gH2AX: phosphorylated
(Ser139) histone H2A.X, RPA: replication protein A, DAPI: 4′ ,6-Diamidino-2-phenylindole.
selection process in which karyotypically abnormal blastomeres, or the
embryos in which they reside, fail to develop. Given the known attrition
of developing cleavage-stage embryos within in-vitro culture systems and
the frequent finding of excluded blastomeres within growing embryos,
the latter interpretation is worth further discussion.
The presence of microscopic nuclear abnormalities was highly
predictive of cellular aneuploidy. Using DNA staining and microscopy, followed by DNA array analysis, we found that embryos containing blastomeres with microscopically abnormal nuclei were
aneuploid due to mitotic mosaicism, while embryos without abnormal nuclei were largely euploid or had only a low degree of
mosaicism. Perhaps most exciting, blastomeres in embryos with
micronucleation showed reciprocal chromosomal gains and losses
of both whole chromosomes and of chromosome segments. Therefore, formation of extra-nuclear DNA may be a primary mechanism
by which mitotic aneuploidies occur.
Though it is in principle possible that extra-nuclear DNA can be
re-integrated into the bulk DNA at the next mitosis, and though it has
Abnormal nuclear formation in early human development
321
Figure 5 Micronuclei are associated with chromosome gain. Blastomeres from the same cleavage-stage embryo are shown. Two blastomeres show
additional micronuclei and gain of chromosome 15 (a and b), while two nuclei show reciprocal chromosome loss (c and d). Scale bars represent 10 mm.
been suggested that abnormal karyotypes can self-correct, we consider
this an unlikely scenario. First, sequential mis-segregation of chromosomes resulting in a normal karyotype has only been documented in unfertilized oocytes at meiosis I and II and can only rescue ,2% of the
abnormal oocytes (Fragouli et al., 2013). Second, we found that
mitotic errors are detrimental to the integrity of the DNA, contrasting
with the mechanism of segregation errors in meiosis. While abnormal
meiotic chromosome segregation results predominantly in whole
chromosome aneuploidies (Fragouli et al., 2013) due to a decrease in
sister chromatid cohesion (Duncan et al., 2012), we found that abnormal
mitosis is often associated with damage to the DNA. Using centromere
localization probes, we have demonstrated the absence of centromeres
within extra-nuclear DNA in a large proportion of samples. In addition
using array genotyping, reciprocal segmental aneuploidies were found.
These chromosome fragments would be unable to normally segregate
at the following mitosis. The observation of cells with a large number
of micronuclei in blastomeres of arrested cleavage-stage embryos (Supplementary Fig. S4B) might reflect a progressive deterioration in the
ability to normally segregate a compromised genome.
We also found that microscopic nuclear abnormalities show evidence
of DNA damage, with extensive phosphorylation of gH2Ax. Our
cytological data are consistent with molecular analysis of cleavage-stage
embryos, showing breakage of chromosomes and frequent chromosomal rearrangements during cleavage-stage development (Wells and Delhanty, 2000; Voet et al., 2011a,b). In cultured cells, such chromosome
damage can result from abnormal DNA replication (Burrell et al.,
2013). In fact it is known that the formation of extra-nuclear DNA is associated with abnormal DNA replication, followed by extensive DNA
damage in mitosis (Crasta et al., 2012). Interestingly, break points map
to fragile sites (Wells and Delhanty, 2000), regions of the chromosomes
that are difficult to replicate. Therefore, mitotic chromosome segregation errors differ from meiotic segregation errors, both in mechanisms
of formation and in their consequence on cell-cycle progression. While
whole chromosome segregation errors occurring in meiosis can be tolerated during preimplantation development, aneuploidies arising
during mitosis are greatly decreased at the blastocyst stage and are
usually lost before establishing a clinical pregnancy (Fragouli et al.,
2013). In parallel, we find that the incidence of cells with micronuclei
and of cells with damaged DNA also decreases. The high frequency
of the nuclear abnormalities in arrested blastomeres, as well as the extensive damage to the DNA reported here, strongly suggests that these
defects are not compatible with cell-cycle progression and that these
322
cells are lost from the developing embryo by exclusion. Live-cell imaging
of early cleavage-stage embryos indeed shows decreased cell-cycle
kinetics and reduced cell division potential of aneuploid embryos
(Wong et al., 2010; Chavez et al., 2012). Cell-cycle arrest may be a key
mechanism to eliminate karyotypically abnormal cells during preimplantation development that could otherwise cause developmental defects
post implantation.
In summary, multiple lines of evidence presented here show microscopic nuclear abnormalities to be a reliable marker for chromosome
mosaicism, chromosome abnormalities and DNA damage. Previous
studies have shown an association between aneuploidy using fluorescence in situ hybridization (FISH) analysis and abnormal nucleation
detected by bright-field microscopy (Kligman et al., 1996), though the association was weaker than reported here. Differences between the two
studies might be explained by the use of improved imaging and karyotyping techniques. Micronuclei are challenging to detect in cleavage-stage
blastomeres using standard light microscopy, and FISH can only
analyze a small number of chromosomes. The use of Hoechst staining
to comprehensively detect abnormal nucleation combined with array
karyotyping of all chromosomes revealed that all embryos with micronuclei were mosaic, and all cells with micronuclei contained a chromosomal
abnormality due to a mitotic error. If further validated, and with
improved live embryo imaging, our observations may provide a convenient noninvasive method of embryo selection in IVF.
We propose that abnormal mitosis induces loss of cells in developing
embryos and significantly undermines normal growth and development.
As early cleavage stages rely on maternal products in the egg (Braude
et al., 1988), a portion of the defects may reflect the quality of the
oocyte. There may be a ‘threshold’ of normal mitosis in the first few
cell divisions necessary for successful development of the embryo. If
the fidelity of early cell divisions is above this threshold, normal development can occur. If the fidelity of early cell divisions is below this threshold,
the embryos arrest. In support of this hypothesis, studies in mice have
revealed that specific blastomeres are required for development to
term (Torres-Padilla et al., 2007). Studies using continuous, real-time
imaging combined with analysis of the cell types formed at the blastocyst
stage may shed light on this hypothesis in human development.
Supplementary data
Supplementary data are available at http://humrep.oxfordjournals.org/.
Acknowledgements
We thank Joao dePinho for assistance with sample collection and Nissim
Benvenisty of the Hebrew University of Jerusalem for critical reading of
the manuscript.
Authors’ roles
D.K. and D.E. designed the studies, D.K., D.E. and G.C. interpreted the
data and wrote the manuscript with input from all authors, D.K. and G.C.
performed confocal microscopy, N.T. performed and interpreted SNP
array analysis, J.T. performed statistics and assisted with figure
preparation and provided technical support, S.M. assisted with data acquisition, R.P. and M.V.S. prepared specimen for research use, L.B.V.
Kort et al.
and D.E. thawed blastomeres and blastocysts, and R.A.L. and M.V.S. provided supervision.
Funding
This study was supported by the Russell Berrie Foundation Program in
Cellular Therapies of Diabetes. D.E. is a NYSCF-Robertson Investigator.
D.H.K. was supported by a clinical fellowship in reproductive endocrinology and infertility. G.C. is supported by a postdoctoral fellowship from
the Agency for Science, Technology and Research, Singapore.
Conflict of interest
None declared.
References
Alikani M, Calderon G, Tomkin G, Garrisi J, Kokot M, Cohen J. Cleavage
anomalies in early human embryos and survival after prolonged culture
in-vitro. Hum Reprod 2000;15:2634 – 2643.
Barbash-Hazan S, Frumkin T, Malcov M, Yaron Y, Cohen T, Azem F, Amit A,
Ben-Yosef D. Preimplantation aneuploid embryos undergo self-correction
in correlation with their developmental potential. Fertil Steril 2009;
92:890 – 896.
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 – 461.
Burrell RA, McClelland SE, Endesfelder D, Groth P, Weller MC, Shaikh N,
Domingo E, Kanu N, Dewhurst SM, Gronroos E et al. Replication stress
links structural and numerical cancer chromosomal instability. Nature
2013;494:492 – 496.
Capalbo A, Bono S, Spizzichino L, Biricik A, Baldi M, Colamaria S, Ubaldi FM,
Rienzi L, Fiorentino F. Sequential comprehensive chromosome analysis on
polar bodies, blastomeres and trophoblast: insights into female meiotic
errors and chromosomal segregation in the preimplantation window of
embryo development. Hum Reprod 2013;28:509 – 518.
Chavez SL, Loewke KE, Han J, Moussavi F, Colls P, Munne S, Behr B, Reijo
Pera RA. Dynamic blastomere behaviour reflects human embryo ploidy
by the four-cell stage. Nat Commun 2012;3:1251.
Crasta K, Ganem NJ, Dagher R, Lantermann AB, Ivanova EV, Pan Y, Nezi L,
Protopopov A, Chowdhury D, Pellman D. DNA breaks and chromosome
pulverization from errors in mitosis. Nature 2012;482:53– 58.
Duncan FE, Hornick JE, Lampson MA, Schultz RM, Shea LD, Woodruff TK.
Chromosome cohesion decreases in human eggs with advanced
maternal age. Aging Cell 2012;11:1121 – 1124.
Fragouli E, Wells D. Aneuploidy in the human blastocyst. Cytogenet Genome
Res 2011;133:149 – 159.
Fragouli E, Wells D, Thornhill A, Serhal P, Faed MJ, Harper JC, Delhanty JD.
Comparative genomic hybridization analysis of human oocytes and polar
bodies. Hum Reprod 2006;21:2319 – 2328.
Fragouli E, Alfarawati S, Spath K, Jaroudi S, Sarasa J, Enciso M, Wells D. The
origin and impact of embryonic aneuploidy. Hum Genet 2013;
132:1001– 1013.
Franasiak JM, Forman EJ, Hong KH, Werner MD, Upham KM, Treff NR,
Scott RT Jr. The nature of aneuploidy with increasing age of the female
partner: a review of 15,169 consecutive trophectoderm biopsies
evaluated with comprehensive chromosomal screening. Fertil Steril 2014;
101:656– 663.e651.
Abnormal nuclear formation in early human development
Hardy K, Handyside AH, Winston RML. The human blastocyst: cell number,
death and allocation during late preimplantation development in vitro.
Development 1989;107:597– 604.
Hardy K, Winston RML, Handyside AH. Binucleate blastomeres in
preimplantation human embryos in vitro: failure of cytokinesis during
early cleavage. J Reprod Fertil 1993;98:549 – 558.
Kligman I, Benadiva C, Alikani M, Munne S. The presence of multinucleated
blastomere’s in human embryos is correlated with chromosomal
abnormalities. Hum Reprod 1996;11:1492 – 1498.
Meriano J, Clark C, Cedesky K, Laskin CA. Binucleated and micronucleated
blastomeres in embryos derived from human assisted reproduction cycles.
Reprod Biomed Online 2004;9:511– 520.
Mertzanidou A, Wilton L, Cheng J, Spits C, Vanneste E, Moreau Y,
Vermeesch JR, Sermon K. Microarray analysis reveals abnormal
chromosomal complements in over 70% of 14 normally developing
human embryos. Hum Reprod 2013;28:256 – 264.
Murray AW. Recycling the cell cycle: cyclins revisited. Cell 2004;
116:221– 234.
Pelinck MJ, De Vos M, Dekens M, Van der Elst J, De Sutter P, Dhont M.
Embryos cultured in vitro with multinucleated blastomeres have poor
implantation potential in human in-vitro fertilization and intracytoplasmic
sperm injection. Hum Reprod 1998;13:960 – 963.
Pickering SJ, Taylor A, Johnson MH, Braude P. An analysis of multinucleated
blastomere formation in human embryos. Mol Hum Reprod 1995;1:254–264.
Racowsky C, Vernon M, Mayer J, Ball GD, Behr B, Pomeroy KO, Wininger D,
Gibbons W, Conaghan J, Stern JE. Standardization of grading embryo
morphology. Fertil Steril 2010;94:1152– 1153.
Royen EV, Manelschots K, Vercruyssen M, De Neubourg D, Valkenburg M,
Ryckaert G, Gerris J. Multinucleation in cleavage stage embryos. Hum
Reprod 2003;18:1062 – 1069.
323
Scott RT Jr, Ferry K, Su J, Tao X, Scott K, Treff NR. Comprehensive
chromosome screening is highly predictive of the reproductive potential
of human embryos: a prospective, blinded, nonselection study. Fertil
Steril 2012;97:870 – 875.
Scott RT Jr, Upham KM, Forman EJ, Hong KH, Scott KL, Taylor D, Tao X,
Treff NR. Blastocyst biopsy with comprehensive chromosome screening
and fresh embryo transfer significantly increases in vitro fertilization
implantation and delivery rates: a randomized controlled trial. Fertil Steril
2013;100:697 – 703.
Torres-Padilla ME, Parfitt DE, Kouzarides T, Zernicka-Goetz M. Histone
arginine methylation regulates pluripotency in the early mouse embryo.
Nature 2007;445:214 – 218.
Treff NR, Su J, Tao X, Levy B, Scott RT Jr. Accurate single cell 24 chromosome
aneuploidy screening using whole genome amplification and single
nucleotide polymorphism microarrays. Fertil Steril 2010;94:2017– 2021.
Voet T, Vanneste E, Van der Aa N, Melotte C, Jackmaert S, Vandendael T,
Declercq M, Debrock S, Fryns JP, Moreau Y et al. Breakagefusion-bridge cycles leading to inv dup del occur in human cleavage stage
embryos. Hum Mutat 2011a;32:783 – 793.
Voet T, Vanneste E, Vermeesch JR. The human cleavage stage embryo is a
cradle of chromosomal rearrangements. Cytogenet Genome Res 2011b;
133:160– 168.
Wells D, Delhanty JDA. Comprehensive chromosomal analysis of human
preimplantation embryos using whole genome amplification and single
cell comparative genomic hybridization. Mol Hum Reprod 2000;6:
1055 – 1062.
Wong CC, Loewke KE, Bossert NL, Behr B, De Jonge CJ, Baer TM, Reijo
Pera RA. Non-invasive imaging of human embryos before embryonic
genome activation predicts development to the blastocyst stage. Nat
Biotechnol 2010;28:1115– 1121.