Abstract

Valuable information on the cytogenetic constitution of female gametes has been deduced from the direct, so-called conventional analysis of oocytes remaining unfertilized in programmes of assisted reproduction. Additional, indirect conclusions have become possible by PGD of the polar bodies. Both techniques provided evidence for the co-existence of two aneuploidy-causing mechanisms during first maternal meiosis; non-disjunction (ND) of bivalents results in the loss or gain of whole chromosomes in metaphase II complements, whereas a precocious division (pre-division, PD) of univalents leads to the loss or gain of single chromatids. As to the distribution of ND and PD, however, direct oocyte chromosome studies and PGD tell surprisingly different stories. Moreover, first and second polar body analyses contradict the data derived from DNA polymorphism studies concerning the distribution of first and second meiotic division errors. An increased awareness of these problems appears necessary because important decisions are made on the basis of PGD results.

Introduction

There has been a long-standing interest in analysing the chromosomal constitution of human oocytes to assess the incidence of cytogenetic abnormalities arising during maternal meiosis. Apart from a few studies on non-inseminated cells donated for research, oocytes failing to fertilize after the application of assisted reproductive technologies have represented the most important material for this purpose. During the past 20 years, extensive data from conventional karyotyping and also molecular cytogenetic studies have become available (Rosenbusch, 2004; Pellestor et al., 2005). They revealed the co-existence of two aneuploidy-causing mechanisms in meiosis I: undoubtedly what must be considered one of the pre-eminent findings in this field of research. The corresponding mechanisms are known as non-disjunction (ND) of bivalents and precocious division or pre-division (PD) of univalents. In metaphase II (MII) complements, ND will result in the loss or gain of whole chromosomes, whereas PD will lead to the loss or gain of single chromatids (Angell, 1997). More recently, PGD of the first and second polar bodies have provided the opportunity to draw indirect conclusions on the chromosomal composition of the oocyte (Kuliev and Verlinsky, 2004; Kuliev et al., 2005), thereby supporting the existence of ND and PD.

Several years ago, we published a first commentary on a seemingly important aspect: the distribution of ND and PD in aneuploid oocytes (Rosenbusch et al., 2001). We used data from direct cytogenetic analysis of the oocyte and from PGD of the first polar body (PB) and surprisingly noted an intriguing incompatibility of the results derived from the two different sources. Briefly, direct oocyte chromosome analyses indicated a slight excess of PD (51.2%) compared with ND (48.8%). Extra chromosomes or chromatids appeared with a frequency of ∼20% each, whereas loss of chromosomes or chromatids each contributed ∼30% to numerically abnormal gametes. However, a completely different picture was obtained from PB diagnoses, suggesting that only ∼13% of the corresponding oocytes would show ND, whereas ∼87% would be assigned to PD. Moreover, chromosome hyperhaploidy considerably exceeded the incidence of hypohaploidy, and extra chromatids were three times more frequent than missing chromatids—findings that also contradicted the tendencies seen in direct cytogenetic analysis.

In the meantime, additional data from conventional oocyte chromosome (Pellestor et al., 2002, 2003) and PB analyses (Kuliev et al., 2005) have been published, and it appeared of interest whether they might help to clarify the situation.

Material, methods and definitions

An unfertilized MII oocyte is characterized by the presence of the first PB. Provided that first meiotic division proceeded normally, both the secondary oocyte and the first PB will carry a haploid 23, X chromosome set. Second meiotic division will be resumed after sperm penetration, causing a separation of oocyte chromosomes into single chromatids and extrusion of 23 chromatids into the second PB. As reviewed elsewhere (Pellestor et al., 2005), oocytes have been investigated by different techniques, including the conventional karyotyping of homogeneously stained or, more rarely, G- or R-banded chromosomes and molecular cytogenetic methods such as fluorescence in situ hybridization (FISH) or spectral karyotyping (SKY). Oocyte chromosomes show a particular morphology with compact, that is, highly condensed arms that render karyotyping much more difficult than for mitotic chromosomes. The sister chromatids are loosely associated and tend to overlap or separate from each other. The structure of the complementary chromosome set of the first PB is inferior to that of the oocyte and mostly appears fuzzy, clumped or fragmented. Because this feature prevents attempts of conventional karyotyping, the method of choice for PGD of the PB is FISH using probes for some selected chromosomes.

The present contribution has been intended as a ‘food for thought’ and not as a complete review. Therefore, only two sources, which currently represent the most extensive and coherent data provided by two different groups of investigators, have been considered. As regards PGD, Tables 2 and 3 published by Kuliev et al. (2005) have been selected. They show the results of FISH analyses of the first PB using probes specific for chromosomes 13, 16, 18, 21 and 22 and include the following abnormalities arising during meiosis I:

  • 23+1: metaphase with an additional chromosome;

  • 23–1: metaphase with a missing chromosome;

  • 23+1/2: metaphase with an additional chromatid;

  • 22+1/2: metaphase in which a chromosome is replaced by a single chromatid.

Attention has to be paid to the fact that these data have been extrapolated to the oocyte (Table I), because a ‘23–1’ constitution of the first PB will equate to a ‘23+1’ complement in the oocyte etc.

Table I.

Distribution of non-disjunction and pre-division in female gametes in relation to the patients’ age

Non-disjunction [n (%)] Pre-division [n (%)] Age (graphic
 
 

 
 
 23+1 23–1 23+1/2 22+1/2  
First polar body diagnosis using FISH for chromosomes 13, 16, 18, 21 and 22 (Kuliev et al., 2005
4019 77 (1.9) 9 (0.2) 836 (20.8) 341 (8.5) ≥35 (38.4) 
Direct karyotyping of oocyte chromosomes (Pellestor et al., 2003)a 
1367 22 (1.6) 27 (2.0) 35 (2.6) 48 (3.5) 19–46 (31.6) 
321 10 (3.1) 8 (2.5) 21 (6.5) 27 (8.4) ≥35 (38.0) 
243 9 (3.7) 6 (2.5) 20 (8.2) 21 (8.6) ≥36 (39.0) 
194 8 (4.1) 6 (3.1) 17 (8.8) 18 (9.3) ≥37 (39.6) 
81 4 (4.9) 5 (6.2) 10 (12.3) 10 (12.3) ≥40 (41.4) 
Non-disjunction [n (%)] Pre-division [n (%)] Age (graphic
 
 

 
 
 23+1 23–1 23+1/2 22+1/2  
First polar body diagnosis using FISH for chromosomes 13, 16, 18, 21 and 22 (Kuliev et al., 2005
4019 77 (1.9) 9 (0.2) 836 (20.8) 341 (8.5) ≥35 (38.4) 
Direct karyotyping of oocyte chromosomes (Pellestor et al., 2003)a 
1367 22 (1.6) 27 (2.0) 35 (2.6) 48 (3.5) 19–46 (31.6) 
321 10 (3.1) 8 (2.5) 21 (6.5) 27 (8.4) ≥35 (38.0) 
243 9 (3.7) 6 (2.5) 20 (8.2) 21 (8.6) ≥36 (39.0) 
194 8 (4.1) 6 (3.1) 17 (8.8) 18 (9.3) ≥37 (39.6) 
81 4 (4.9) 5 (6.2) 10 (12.3) 10 (12.3) ≥40 (41.4) 

FISH, fluorescence in situ hybridization.

a

The total number of karyotypes was 1397. However, the authors excluded 30 cells (29 with structural chromosome aberrations and one case of tetraploidy) from statistical calculation.

To facilitate comparisons, complex cases of aneuploidy were not considered, that is, we have excluded the simultaneous occurrence of hypohaploidy and hyperhaploidy in the same metaphase and the co-existence of events caused by ND and PD. Concerning direct oocyte karyotypes, the data given in Table 2 by Pellestor et al. (2003) have been adopted, although a few complements had more than one additional or missing chromosome/chromatid. However, it was not possible to exclude these complements, because, from the available information (Pellestor et al., 2002, 2003), they could not be allocated to the corresponding female age groups (see below). Because the main intention was to reveal the distribution of ND and PD, this was considered a minor shortcoming that has been ignored.

Results

The information obtained from the first PB analyses (Kuliev et al., 2005) leads to the following conclusions on the chromosomal constitution of the corresponding oocytes (Table I); extra chromosomes (1.9%) and chromatids (20.8%) are more frequent than missing chromosomes (0.2%) or chromatids (8.5%). The overall rate of aneuploidy (excluding complex cases as mentioned above) is 31.4% with substantially fewer chromosome (2.1%, 86/4019) than chromatid errors (29.3%, 1177/4019).

Direct examination of the oocyte chromosomes reveals a decreased frequency of aneuploidy, although this approach evaluates the whole haploid chromosome set instead of five selected chromosomes. In total, Pellestor et al. (2002, 2003) observed 151 (10.8%) aneuploid cells, comprising 5.4% hypohaploidy, 4.1% hyperhaploidy, 0.8% complex aneuploidies and 0.05% extreme aneuploidies with less than 18 chromosomes. Their data also indicate a preponderance of chromatid errors compared with whole chromosome ND. In detail, however, there are intriguing differences (Table I). Considering all analysable karyotypes, the rates for additional (1.6%) or missing (2.0%) chromosomes found by Pellestor et al. (2003) are nearly equal, whereas there are more complements with missing (3.5%) than with additional (2.6%) chromatids. Thus, the information provided by the two techniques still remains contradictory with regard to the total incidence of aneuploidy and the participation of ND and PD. However, there is an important aspect that has been disregarded in our previous article (Rosenbusch et al., 2001): the fact that the data from PGD refer to patients ≥35 years of age (Kuliev et al., 2003, 2005). Consequently, the results of direct oocyte karyotyping provided by Pellestor et al. (2003) have been subdivided according to the patients’ age (Table I). Four age groups have been considered, of which one (≥35 years, mean: 38.0 years) compares favourably with the group examined by Kuliev et al. (2005). However, the discrepancies persist. First PB analyses suggest an approximately 10-fold higher incidence of chromosome hyperhaploidy versus hypohaploidy and a preponderance of additional versus missing chromatids. Independent of the patients’ age, these tendencies are not supported by conventional oocyte chromosome studies.

Finally, concomitant analysis of both PBs has offered the unique possibility to judge not only first but also second maternal meiotic division errors. Here, however, another surprise is waiting, because in contrast to reports on the predominant origin of chromosome 16 and 21 errors in meiosis I and chromosome 18 errors in meiosis II, ‘the direct data showed no significant difference in the origin of the chromosome 21 errors, and the opposite tendencies for the chromosome 16 and 18 errors’ (Kuliev and Verlinsky, 2004). In short, these extensive PB analyses contradict the conclusions made on the basis of DNA polymorphism studies (Hassold and Hunt, 2001).

Comments

In search of reasons for the above-mentioned discrepancies, we inevitably encounter phenomena summarized under the term ‘technical problems’. They have extensively been discussed for conventional oocyte chromosome studies, and it has been shown that a variety of pitfalls exist that affect the interpretation of meiotic metaphases (Michelmann and Tavmergen, 1991). It is evident that particularly the evaluation of hypohaploidy depends on the technique of cell fixation (Rosenbusch, 2004; Pellestor et al., 2005). However, the quality of chromosome preparation is not only important for the success of conventional karyotyping but also for FISH assays. The latter may have introduced a significant bias in oocyte chromosome studies, because results can be obtained from poor-quality complements that would not be analysable by the conventional approach (Pellestor et al., 2005). The particular morphology of oocyte chromosomes may facilitate the overlapping of chromosome territories and subsequently of fluorescent signals (Pellestor et al., 2005), and moreover, the results seem to depend on the probes used. Just to mention a few examples, Dailey et al. (1996) reported an error rate of 10.6%, whereas Eckel et al. (2003) demonstrated that only 10 of 17 metaphases remained abnormal as soon as the preparations analysed by dual-colour FISH were re-examined by multilocus FISH.

The quality of PB chromosomes is even worse than that of oocyte chromosomes, and their examination is further complicated by the fact that the PB undergoes rapid ageing in vitro, frequently accompanied by fragmentation. Because missing signals in the PB may arise from hybridization failures or artefactual loss of chromosomal material (Verlinsky et al., 1998), we assumed that the excess of hyperhaploidy (23+1 and 23+1/2) assigned to the oocyte by PB analyses can be explained by an incorrect evaluation of the complementary PB chromosome sets (23–1 and 22+1/2, respectively) (Rosenbusch et al., 2001). More recently, Pujol et al. (2003) examined oocyte/first PB doublets to estimate the percentage of missing chromosomes/chromatids that would be artefactual when analysing the first PB alone. This incidence amounted to 25.8%. Gutiérrez-Mateo et al. (2004) studied the first PB by comparative genomic hybridization and the oocyte chromosomes by FISH and obtained reciprocal results in 88.1%. This approach comprised 42 first PB–MII doublets donated by 33 patients with a mean age of 35.8 years. Counting only cases with reciprocal results and only complements with loss or gain of one chromosome/chromatid, the data provided by Gutiérrez-Mateo et al. (2004) include six cells with ND (two with gain and four with loss of a chromosome, respectively) and six gametes with PD (two with gain and four with loss of a chromatid, respectively). These findings agree better with the previously reported tendency deduced from direct oocyte chromosome analyses but not with indirect conclusions from first PB studies (Rosenbusch et al., 2001).

Of course, technical problems are only one aspect. Another point at issue is the origin of the material because PGD results are based on fresh oocytes, whereas most conventional karyotyping and FISH studies refer to oocytes remaining unfertilized in vitro that have been kept in culture for varying periods of time. However, a few karyotyping studies applying Tarkowski’s (1966) technique could be performed on fresh, non-inseminated oocytes. They reported aneuploidy rates of about or >20% that is in agreement with data from unfertilized, aged gametes fixed with the same technique (Pellestor et al., 2005). Sandalinas et al. (2002) were able to analyse 47 fresh, non-inseminated oocytes by SKY. In this small sample, loss of a whole chromosome (five cells) was more frequent than presence of an additional chromosome (one case), but there were slightly more complements with an additional than with a missing chromatid (four versus three oocytes). However, the incidence of ND, unbalanced and balanced PD increased with patients’ age, suggesting that even when the effects of in vitro ageing have been minimized, other parameters need careful consideration. These may also include specific features of gametogenesis. For instance, an extra chromosome in the first PB with no corresponding loss of material in the oocyte (or vice versa) could be explained by the presence of trisomic precursor cells (Cupisti et al., 2003; Pujol et al., 2003). However, a significant contribution of this factor to the observed discordance in the distribution of ND and PD seems rather improbable.

Finally, it should be noted that follow-up studies of the oocytes classified as abnormal by (first + second) PB analyses could not confirm the predicted results in 33.8% of cases (Verlinsky et al., 1998). This observation and the preceding considerations may raise the question about the dimension of the error rate that we are willing to accept. Admittedly, the problem may be less severe for countries in which PGD is followed by the biopsy of one or more blastomeres of the preimplantation embryo. However, in countries such as Germany, this technique is not allowed. Here, the results of PGD must be used to decide whether the pronuclear stage can develop into an embryo. The above-mentioned inconsistencies suggest that the reliability of PB analyses by FISH for assessing numerical anomalies in the corresponding oocyte has to be verified in further studies. Preferably, this should be done in a blind and independent manner, that is, oocyte and PB/PBs should be examined by different investigators who do not know which of these meiotic products belong together. Those future studies should, if possible, include fresh, non-inseminated and/or in vitro matured oocytes and new techniques that allow the analysis of more chromosomes. Such an approach also makes sense in view of the contradictory information on the frequency of meiosis I and II errors for chromosomes 16, 18 and 21. If the predictions made by PGD were correct, different rules must be valid for very early aneuploid stages that fail to implant than for those affected conceptuses that end as spontaneous abortions or survive to term. In conclusion, there are important reasons for continuing efforts in this area of research. The ‘lessons from PGD’ (Kuliev and Verlinsky, 2004) but also from direct oocyte chromosome analyses have to be scrutinized before we start rewriting the corresponding chapters on maternal meiosis.

References

Angell
R
First-meiotic-division nondisjunction in human oocytes
Am J Hum Genet
 , 
1997
, vol. 
61
 (pg. 
23
-
32
)
Cupisti
S
Conn
CM
Fragouli
E
Whalley
K
Mills
JA
Faed
MJW
Delhanty
JDA
Sequential FISH analysis of oocytes and polar bodies reveals aneuploidy mechanisms
Prenat Diagn
 , 
2003
, vol. 
23
 (pg. 
663
-
668
)
Dailey
T
Dale
B
Cohen
J
Munné
S
Association between nondisjunction and maternal age in meiosis-II human oocytes
Am J Hum Genet
 , 
1996
, vol. 
59
 (pg. 
176
-
184
)
Eckel
H
Kleinstein
J
Wieacker
P
Stumm
M
Multi-locus (ML)-FISH is a reliable tool for nondisjunction studies in human oocytes
Cytogenet Genome Res
 , 
2003
, vol. 
103
 (pg. 
47
-
53
)
Gutiérrez-Mateo
C
Benet
J
Wells
D
Colls
P
Bermúdez
MG
Sánchez-García
JF
Egozcue
J
Navarro
J
Munné
S
Aneuploidy study of human oocytes first polar body comparative genomic hybridization and metaphase II fluorescence in situ hybridization analysis
Hum Reprod
 , 
2004
, vol. 
19
 (pg. 
2859
-
2868
)
Hassold
T
Hunt
P
To err (meiotically) is human: the genesis of human aneuploidy
Nat Rev Genet
 , 
2001
, vol. 
2
 (pg. 
280
-
291
)
Kuliev
A
Verlinsky
Y
Meiotic and mitotic nondisjunction: lessons from preimplantation genetic diagnosis
Hum Reprod Update
 , 
2004
, vol. 
10
 (pg. 
401
-
407
)
Kuliev
A
Cieslak
J
Ilkevitch
Y
Verlinsky
Y
Chromosomal abnormalities in a series of 6733 human oocytes in preimplantation diagnosis for age-related aneuploidies
Reprod Biomed Online
 , 
2003
, vol. 
6
 (pg. 
54
-
59
)
Kuliev
A
Cieslak
J
Verlinsky
Y
Frequency and distribution of chromosome abnormalities in human oocytes
Cytogenet Genome Res
 , 
2005
, vol. 
111
 (pg. 
193
-
198
)
Michelmann
HW
Tavmergen
E
Chromosomenanalysen aus unbefruchteten menschlichen Eizellen nach In-vitro-Insemination. Wie sind die Ergebnisse zu bewerten?
Fertilität
 , 
1991
, vol. 
7
 (pg. 
129
-
139
)
Pellestor
F
Andréo
B
Arnal
F
Humeau
C
Demaille
J
Mechanisms of non-disjunction in human female meiosis: the co-existence of two modes of malsegregation evidenced by the karyotyping of 1397 in-vitro unfertilized oocytes
Hum Reprod
 , 
2002
, vol. 
17
 (pg. 
2134
-
2145
)
Pellestor
F
Andréo
B
Arnal
F
Humeau
C
Demaille
J
Maternal aging and chromosomal abnormalities: new data drawn from in vitro unfertilized human oocytes
Hum Genet
 , 
2003
, vol. 
112
 (pg. 
195
-
203
)
Pellestor
F
Anahory
T
Hamamah
S
The chromosomal analysis of human oocytes. An overview of established procedures
Hum Reprod Update
 , 
2005
, vol. 
11
 (pg. 
15
-
32
)
Pujol
A
Boiso
I
Benet
J
Veiga
A
Durban
M
Campillo
M
Egozcue
J
Navarro
J
Analysis of nine chromosome probes in first polar bodies and metaphase II oocytes for the detection of aneuploidies
Eur J Hum Genet
 , 
2003
, vol. 
11
 (pg. 
325
-
336
)
Rosenbusch
B
The incidence of aneuploidy in human oocytes assessed by conventional cytogenetic analysis
Hereditas
 , 
2004
, vol. 
141
 (pg. 
97
-
105
)
Rosenbusch
B
Glaeser
B
Schneider
M
Aneuploidy in human oocytes: nondisjunction or predivision?
Cytogenet Cell Genet
 , 
2001
, vol. 
94
 (pg. 
241
-
243
)
Sandalinas
M
Márquez
C
Munné
S
Spectral karyotyping of fresh, non-inseminated oocytes
Mol Hum Reprod
 , 
2002
, vol. 
8
 (pg. 
580
-
585
)
Tarkowski
AK
An air-drying method for chromosome preparations from mouse eggs
Cytogenetics
 , 
1966
, vol. 
5
 (pg. 
394
-
400
)
Verlinsky
Y
Cieslak
J
Ivakhnenko
V
Evsikov
S
Wolf
G
White
M
Lifchez
A
Kaplan
B
Moise
J
Valle
J
, et al.  . 
Preimplantation diagnosis of common aneuploidies by the first- and second-polar body FISH analysis
J Assist Reprod Genet
 , 
1998
, vol. 
15
 (pg. 
285
-
289
)