Human Reproduction vol.13 no.1 pp.154–160, 1998
The meiotic competence of in-vitro matured human
oocytes is influenced by donor age: evidence that
folliculogenesis is compromised in the reproductively
aged ovary
Kimberly Volarcik1, Leon Sheean2, James
Goldfarb2, Linda Woods1, Fadi W.Abdul-Karim3
and Patricia Hunt1,4
1Department
2Department
of Genetics and Centre for Human Genetics,
of Reproductive Biology and 3Department of
Pathology, Case Western Reserve University School of Medicine
and University Hospitals of Cleveland, Cleveland, Ohio 441064955, USA
4To
whom correspondence should be addressed
The human oocyte appears to be particularly prone to
meiotic errors, and the incidence of these errors is strongly
influenced by maternal age. We have initiated studies of
human oocytes from unstimulated ovaries and have
observed age-related effects on the meiotic process in
oocytes from unselected antral follicles. Specifically, in
oocytes obtained from donors over the age of 35 years, the
majority of oocytes that extruded a first polar body in
culture and arrested at second meiotic metaphase had
aberrations in spindle formation and chromosome alignment. Similarly, observations of a limited number of oocytes
at first meiotic metaphase suggest disturbances at this stage
of meiosis as well. Finally, preliminary results of nondisjunction studies suggest that the frequency of errors in
chromosome segregation at the first meiotic division is
influenced by donor age in in-vitro matured oocytes as it
is in oocytes undergoing meiotic maturation in vivo. These
data provide direct evidence that the meiotic competence
of oocytes from unstimulated ovaries declines with donor
age. Similarly, studies of in-vitro fertilization (IVF) pregnancies in older women indicate that the developmental
competence of the human oocyte declines with age. Since
both meiotic and developmental competence are acquired
during the late stages of oocyte growth, we postulate that
an age-related decline in the process of folliculogenesis
results in reduced oocyte quality and that the well characterized age-related increase in meiotic non-disjunction is
one symptom of compromised oocyte growth.
Key words: age-related non-disjunction/folliculogenesis/meiotic competence
Introduction
One of the earliest and most profound ageing effects in the
human is the decline in reproductive function that becomes
evident during the fourth and fifth decades of life in the human
female. It remains unclear whether this decline is the result of
age-related changes in the oocyte, the uterus, the neuroendo154
crine system or a combination of factors. However, one of
the best characterized aspects of reproductive decline is the
decreased ability of the oocyte to segregate chromosomes
successfully during the completion of meiosis. The frequency
of chromosomally abnormal offspring resulting from errors
during female meiosis is directly correlated with maternal age,
and it has been estimated that by the fifth decade of life as
many as 50% of all ovulated oocytes may be chromosomally
abnormal (Hassold, 1986).
Little is known about the basis of age-related chromosome
mal-segregation. It is commonly believed to originate in
maternal meiosis I because, in the human female, oocytes
enter meiosis during the fetal period and remain suspended in
prophase of the first meiotic division at the diplotene (dictyate)
stage until ovulation. As a result, completion of the first
meiotic division may take 40 years or longer. The duration of
the division has led to speculation that the basis of the age
effect on trisomy is due to events occurring prenatally at the
time of entry into meiosis I (Henderson and Edwards, 1968);
or during the interval between meiotic arrest and re-entry
(Penrose, 1965); or during the periovulatory period at the time
of re-entry into meiosis I (Crowley et al., 1979; Sugawara and
Mikamo, 1983; Eichenlaub-Ritter et al., 1988; Warburton,
1989; Gaulden, 1992; Tarin, 1995). Alternatively, changes in
the uterine environment rather than the oocyte have been
postulated to influence the survival of the chromosomally
abnormal conceptus (Ayme and Lippman-Hand, 1982). Evidence from genetic studies (reviewed in Hassold et al., 1993)
argues against changes in the uterine environment (the socalled ‘relaxed selection’ model) as a causal factor. The data
accumulated to date, however, do not allow us to distinguish
between factors acting prenatally (at the time of meiotic entry)
or postnatally (during the prolonged resting phase or at the
time of resumption and completion of the first meiotic division).
The mammalian female meiotic process is complicated by
the fact that the ability to resume and complete the first meiotic
division is linked to the process of oocyte growth. In the
neonate, oocytes are arrested in prophase of the first meiotic
division and are surrounded by a single layer of somatic
cells. During the complex process of folliculogenesis, these
primordial follicles undergo significant growth and development to produce a periovulatory follicle containing a mature
oocyte. Recent studies in the mouse have demonstrated that
both meiotic competence (the ability to resume and complete
the first meiotic division and to arrest at the second meiotic
metaphase) and developmental competence (the ability to
undergo fertilization, to cleave, and to support embryonic
development) are acquired by the oocyte in a stepwise fashion
during the late stages of folliculogenesis (Eppig et al., 1994).
© European Society for Human Reproduction and Embryology
Folliculogenesis and the reproductively aged ovary
The importance of the final stages of follicular maturation on
the meiotic process suggests that somatic factors may play a
role in subsequent meiotic events.
Recently, we have suggested that the control of the female
meiotic process differs in a fundamental respect from the male
meiotic and the mitotic process, lacking a sensitive cell cycle
checkpoint control mechanism that monitors the alignment of
chromosomes at metaphase (Hunt et al., 1995; LeMaire-Adkins
et al., 1997). We have postulated that this lack of a stringent
checkpoint control provides an explanation for the higher error
rate in female meiosis. This hypothesis, however, does not
explain the influence of maternal age on the incidence of
meiotic errors. To elucidate the basis of the age-related
increase in chromosome non-disjunction, we have combined
immunofluorescence staining and molecular cytogenetic techniques to study in-vitro matured human oocytes from donors
ranging in age from 18 to 55 years. Our results demonstrate
an age-related increase in several different types of meiotic
aberrations, including defects in meiotic spindle formation and
chromosome alignment, and chromosome segregation errors
at the first meiotic division. We interpret these observations
as evidence of an age-related decline in the process of
folliculogenesis and suggest that meiotic non-disjunction may
be but one consequence of a more generalized decline in
oocyte growth in the ageing human ovary.
Materials and methods
Sources of human oocytes
Human oocytes were obtained from women between the ages of 18
and 55 years. Cancer patients were excluded to avoid possible
confounding effects of chemotherapeutic agents. Oocytes were recovered from unstimulated ovaries by two different methods. First,
oocytes were retrieved from antral follicles dissected from specimens
of ovarian tissue removed in the course of routine gynaecological
surgery. Surgical specimens of ovarian tissue were obtained through
the NCI-funded Human Cooperative Tissue Network at Case Western
Reserve University. Ovarian tissue specimens were placed in sterile
containers at room temperature containing Dulbecco’s modified
Eagle’s medium (DMEM) medium (Gibco-Life Technologies, Grand
Island, NY, USA) and transported to the laboratory within several
hours of surgery. Second, with prior patient consent, oocytes were
aspirated from follicles visible on the surface of the ovary of women
undergoing laparoscopic tubal ligation surgery at University Hospitals
of Cleveland. Follicular fluid aspirates were placed in sterile tubes,
the aspiration needle was rinsed with sterile, prewarmed HEPES
buffered human tubal fluid (HTF) media (Irvine Scientific, Santa
Ana, CA, USA), and both follicular fluid and washes were placed in
a 37°C heating block and immediately transported to the laboratory.
The two sources of oocytes, those obtained from antral follicles
dissected from surgical specimens of ovarian tissue and those aspirated
from visible follicles on the surface of the ovary of patients undergoing
laparoscopic tubal ligation, provided access to slightly different
populations: older donors predominated in the ovarian tissue group,
a proportion of these specimens represented diseased ovaries, and
oocytes were obtained from antral follicles ranging in size from 5 to
15 mm. In contrast, follicular aspiration on women undergoing
laparoscopic tubal ligation provided access to a younger group of
oocyte donors, predominately disease-free ovaries and large antral
follicles, since only follicles visible on the surface of the ovary were
aspirated. Despite these differences, the behaviour and characteristics
of the oocytes obtained from the two sources were not significantly
different (with the exception of a slightly accelerated rate of polar
body extrusion among oocytes obtained from follicular aspirates) and
oocytes from the two sources were pooled for the purposes of analysis.
Characterization and culture of human oocytes
Ovarian tissue specimens were carefully dissected to expose intact
follicles, and follicles were removed from the surrounding stromal
tissue. The diameter of individual follicles was measured and oocytes
were recovered from follicles .4 mm in diameter. The oocytes were
aspirated from these follicles using a finely drawn Pasteur pipette.
For samples from follicular aspirations, the follicular fluid and washes
were examined with a stereo microscope to identify the oocyte and
adherent cumulus cells.
The quality of the oocytes obtained from either source was assessed
using the following diagnostic indicators: presence or absence of
adherent cumulus cells, oocyte diameter and shape, and cytoplasmic
characteristics. Oocytes exhibiting clear signs of degeneration (e.g.
highly condensed, dark cytoplasm) were classified as ‘dead on arrival’
and discarded. All remaining oocytes were cultured individually in
10 µl drops of one of three randomly assigned culture media (Ham’s
F10, minimal essential medium (MEM) or Waymouth’s medium;
Gibco-Life Technologies) supplemented with sodium pyruvate (0.23
mM), 2.5 µg human follicle stimulating hormone (FSH) (National
Hormone and Pituitary Program, the National Institute of Diabetes
and Digestive and Kidney Diseases (NIDDKD), the National Institute
of Child Health and Human Development (NICHD), and the US
Dept. of Agriculture (USDA)), and 10% follicular phase serum
(obtained from a recently gravid female donor). Oocytes were cultured
at 37°C in an atmosphere of 5% CO2, 5% O2 and 90% N2. After 24 h
in culture, the oocytes were examined for germinal vesicle breakdown
(GVBD) and the presence of a polar body. In most cases, the cumulus
cells were tightly associated with and obscuring the oocyte and it
was necessary to dissociate them mechanically by gently pipetting
the oocyte with a finely drawn glass pipette. Oocytes exhibiting a
polar body after 24 h in culture were fixed and processed as described
below. In addition, a subset of oocytes scored as GVBD after 24 h
were fixed to obtain oocytes at first meiotic metaphase. However, the
majority of oocytes showing no evidence of polar body extrusion
after 24 h in culture were maintained in culture and monitored for
polar body extrusion for an additional 24 h. At the end of the 48-h
culture period, oocytes that had undergone GVBD but showed no
sign of polar body extrusion were fixed and immunoreacted to
determine the stage of meiotic arrest, whereas oocytes that failed to
undergo GVBD were considered incapable of meiotic maturation and
were discarded.
Oocyte analysis
Oocytes were analysed in one of three different ways: first, for 21 of
the 73 oocytes that extruded a polar body, air dried cytogenetic
preparations were made according to the technique described by
Tarkowski (1966). Subsequent fluorescence in-situ hybridization
(FISH) analysis was performed as described below to obtain information on the segregation of homologous chromosomes at the first
meiotic division. Second, the remaining 52 oocytes exhibiting a polar
body were fixed intact in fibrin clots attached to a microscope slide
and subjected to sequential immunofluorescence staining and FISH
analysis as previously described (Hunt et al., 1995). This approach
allowed us to obtain information on both the morphology of the
second meiotic spindle and the segregation of specific homologous
chromosomes at the first meiotic division. Finally, all oocytes that
underwent GVBD but did not extrude a polar body were fixed intact
155
K.Volarcik et al.
and subjected to immunofluorescence staining to determine the meiotic
stage of the cell and to obtain information about meiotic spindle
formation and chromosome behaviour at the first division.
To visualize the meiotic spindle, immunofluorescence staining with
a monoclonal antibody to β-tubulin (Sigma) and chromatin staining
with propidium iodide (Sigma) were performed as previously
described (Hunt et al., 1995). The microtubule and chromatin characteristics were utilized to determine the meiotic stage of all oocytes
undergoing GVBD during the culture period.
FISH with chromosome-specific probes was performed on all
oocytes that extruded a first polar body and on the few oocytes
obtained at anaphase of the first meiotic division. Pericentromeric
probes for alpha satellite sequences of chromosomes 16, 18 and the
X and a distal 21q probe were utilized for these analyses. Hybridizations were performed using the following probes: a direct labelled
Spectrum Orange probe for chromosome 18 (Vysis), a direct labelled
Spectrum Green probe for the X chromosome (Vysis), a biotin
labelled probe for chromosome 16 (Oncor), and a digoxigenin labelled
probe for chromosome 21 (Oncor). Slides were denatured at 72°C for
10 min in 70% formamide and the chromosome 16 and chromosome 21
probes were prepared in 50% formamide solution, denatured at 72°C
for 5 min, and applied to the slide. A coverslip was added and sealed
with rubber cement, and the slides were hybridized overnight at 37°C
in a humid chamber. Following hybridization, the slides were washed
in 50% formamide/23sodium chloride/sodium citrate (SSC) at 37°C
for 10 min followed by a 5 min wash in 23SSC at 37°C. The slides
were then blocked in 43SSC/3% BSA for 20 min, detected with
avidin and antidigoxigenin conjugated fluorochromes for 1 h at 37°C,
counterstained with propidium iodide and imaged on a confocal
microscope. The chromosome 21 and 16 signals were subsequently
quenched by exposure to light and the slides were rehybridized with
X- and chromosome 18-specific probes. The second hybridization
was the same as the first except that the slides were not denatured
and, since the probes were directly labelled, the blocking and detection
steps were omitted.
Following immunofluorescence staining and after each of the
hybridizations the oocytes were visualized on a BioRad MRC600
confocal system equipped with a krypton/argon laser and attached to
a Zeiss Axioplan fluorescence microscope. The meiotic spindle and
chromosomes were visualized by three-dimensional optical sectioning.
Statistical analysis
Standard goodness of fit analyses were used to determine if the
differences observed between oocytes obtained from the two different
sources or from different aged donors were significant. Significance
levels were at P 5 0.05.
Results
Oocytes were obtained from two different sources, specimens
of ovarian tissue and direct aspiration of follicles, providing
access to oocytes from women of a range of ages. A total of
298 ovarian tissue specimens and follicular aspirates from 291
patients were obtained. In total 403 oocytes were retrieved
from these specimens; 160 from donors under 35 years of age
and 243 from donors 35 years of age and older. Because
oocytes were obtained from unselected antral follicles, 262 of
the 403 (65%) oocytes were dead upon retrieval or died in
culture and were excluded from the study. An additional 24
oocytes failed to undergo GVBD during the 48 h culture
period and were similarly excluded. The number of such
oocytes was independent of the source of the oocyte (i.e.
156
Table I. Meiotic maturation by donor age
Oocytes from
Donors
,35 years
Oocytes from
Donors
ù35 years
Total oocytes analysed
GVBD oocytes analysed at 24 h
44
9 (20.5)
73
18 (24.7)
Total oocytes for analysis of
meiotic maturation
Oocytes extruding a polar body
a) ,24 h
b) .24 h
MI arrested oocytes after 48 h
35
55
30
17
13
5
(86)
(57)
(43)
(14)
43
16
27
12
(78)
(37)
(63)
(22)
Values in parentheses are percentages.
GVBD 5 germinal vesicle breakdown.
ovarian tissue specimens or follicular aspirates) and the age
of the donor.
Effect of maternal age on meiotic maturation
In total, 117 oocytes were studied; 44 oocytes from donors
,35 years and 73 oocytes from donors ù35 years of age
(Table I). A small proportion (27/117) of the oocytes were
fixed at the GVBD stage after 24 h in culture to obtain oocytes
at the first meiotic division. These oocytes have been omitted
from the analysis of meiotic maturation which follow, but the
results of immunofluorescence studies are discussed later.
Information on the meiotic maturation of the remaining 90
oocytes is detailed in Table I. Thirty of the 35 (86%) oocytes
obtained from donors under the age of 35 and 43 of the 55
(78%) oocytes from donors 35 years of age and older extruded
a polar body in culture. Although this difference was not
significant, the data suggested that the time of polar body
extrusion may be influenced by donor age. However, careful
examination of the data revealed an association with the source
of the oocytes, not with donor age; a comparison of the rate
of polar body extrusion for oocytes obtained from the two
different sources demonstrates that 13 out of 15 (87%) of the
oocytes obtained from aspirates extruded a polar body by 24 h,
as compared to 20 out of 58 (35%) of the oocytes obtained
from ovarian tissue. This difference is highly significant
(χ22df 5 13.02, P , 0.005) and, moreover, is independent of
donor age. Thus the apparent age effect was due to the fact
that a greater proportion of oocytes from younger donors were
obtained from follicular aspirates. This difference in the rate
of meiotic maturation between oocytes obtained from the
two different sources undoubtedly reflected the fact that the
oocytes obtained via aspiration were from larger, more mature
antral follicles. The data on oocytes that underwent GVBD
but did not extrude a polar body confirmed this; five of the
35 (14%) oocytes from donors ,35 years and 12 of the 55
(22%) oocytes from donors ù35 years underwent GVBD but
did not extrude a polar body after 48 h in culture. Although
the difference between donor age groups in the number of
meiotically arrested oocytes is not significant, all 17 MI
arrested oocytes were obtained from ovarian tissue.
Folliculogenesis and the reproductively aged ovary
Table II. Aberrations in MII spindle formation and chromosome alignment
Donor age
(years)
Total oocytes
analysed
Oocytes with
aberrations
,35
.35
18
34
2 (11)a
24 (71)a
8
12
1 (13)
6 (50)
10
22
1 (10)
18 (82)
Oocytes with polar body at ,24 h
,35
.35
Oocytes with polar body after .24 h
,35
.35
Values in parentheses are percentages.
aSignificantly different χ2
1df 5 16.64, P , 0.001.
Effect of maternal age on spindle formation and chromosome
alignment
Table II summarizes the results of the combined immunofluorescence studies conducted to examine spindle organization
and chromosome alignment in oocytes arrested at second
meiotic metaphase. As can be seen in Table II, the meiotic
aberrations observed at this stage were strongly related to
donor age. Specifically, aberrations in the second meiotic
spindle apparatus were observed in only two of the 18 (11%)
MII arrested oocytes obtained from donors ,35 years, but
were seen in 24 of the 34 (71%) MII oocytes from donors
ù35 years. This difference is highly significant (χ21df 5 16.64;
P , 0.001) and provides evidence that the meiotic competence
of the human oocyte declines with age. In most cases a second
meiotic spindle was evident but the metaphase alignment of
the chromosomes was disrupted; however, in a few cases
bipolar spindle formation was completely disrupted. Both the
frequency and extent of disruption was highest among those
oocytes that extruded a polar body after .24 h in culture
(Table II). Examples of the types of aberrations observed are
shown in Figure 1.
The age-related difference in the frequency of MII arrested
oocytes with aberrations in spindle formation and chromosome
alignment was not a reflection of differences in the source of
the oocytes. If oocytes obtained from ovarian tissue were
analysed separately, one out of 14 (7%) of the oocytes
obtained from donors ,35 years of age exhibited aberrations
as compared with 22 out of 31 (71%) of oocytes obtained
from donors over 35 years of age (χ22df 5 5.1, P , 0.025).
Although the same trend is observed for oocytes obtained
from follicular aspirates, the small number of such oocytes
precludes a similar comparison by donor age.
To determine if aberrations in meiotic spindle formation
and chromosome alignment were also a feature of the first
meiotic division, 27 GVBD oocytes were fixed at 24 h.
Immunofluorescence staining revealed that 11 of the 27 oocytes
were, in fact, at metaphase I, the remainder being either at
prometaphase or anaphase of the first division. In two of the
11 metaphase I oocytes (from donors aged 30 and 35 years)
the alignment of the chromosomes was disrupted. As was the
case for the majority of MII oocytes, both aberrant MI oocytes
had a bipolar spindle with the majority of the chromosomes
aligned at the spindle equator but with multiple unaligned
chromosomes located between the spindle equator and the
spindle poles. Although based on a limited number of observations, these data suggest that disturbances in the meiotic
process are also evident at the first meiotic division.
Effect of maternal age on chromosome segregation at the
first meiotic division
Data on the segregation of homologous chromosomes at the
first meiotic division were obtained for one or more of the
four chromosomes studied on a total of 43 oocytes; 20 oocytes
obtained from donors under 35 years of age and 23 oocytes
obtained from donors ù35 years of age. The data are summarized by donor age and chromosome in Table III. Chromosomespecific rates of mal-segregation ranged from 6% (2/34) for
the X chromosome to 23% (5/22) for chromosome 16; the
combined total for all four chromosomes was 11% (10/90).
The total rate of mal-segregation for oocytes obtained from
women under the age of 35 years was approximately half that
for oocytes obtained from women 35 years and older (8 and
14% respectively). Although not statistically significant, these
results were consistent with an effect of increasing maternal
age on chromosome segregation at the first meiotic division.
Discussion
The association between meiotic chromosome non-disjunction
and advancing maternal age in natural pregnancies has been
recognized for years, although the mechanism by which
age influences chromosome segregation remains unknown.
Similarly, aneuploidy is a common occurrence in IVF programmes. Cytogenetic studies of oocytes that failed to fertilize
have suggested that a significant proportion of the oocytes
obtained via ovarian stimulation for IVF are aneuploid
(Wramsby et al., 1978; Martin et al., 1986; Plachot et al.,
1988; Pellestor, 1991; and reviewed in Hassold et al., 1993).
Furthermore, the majority of morphologically abnormal
oocytes and embryos have been suggested to be aneuploid.
For example, 45% of oocytes with cytoplasmic abnormalities
were reported to be aneuploid (Van Blerkom and Henry, 1992)
and 60% of embryos arresting in the preimplantation stages
showed chromosome mosaicism on FISH analysis (Munné
et al., 1994), suggesting that mitotic as well as meiotic nondisjunction is an impediment to successful assisted reproduction.
To investigate the basis of the age-related increase in meiotic
chromosome non-disjunction in the human female, we initiated
studies of in-vitro matured human oocytes from unstimulated
ovaries. Oocytes from unstimulated ovaries were chosen for
two reasons. Firstly, this allowed us more closely to simulate
the normal in-vivo situation, an important consideration since
hormonal stimulation of the ovary itself might influence the
rate of meiotic non-disjunction. Secondly, more complete data
are available on non-disjunction rate and the influence of
maternal age in pregnancies resulting from natural cycles than
those resulting from the various ovarian stimulation regimes
used in assisted reproduction. Two sources of oocytes, those
obtained from antral follicles dissected from surgical specimens
of ovarian tissue and those aspirated from visible follicles on
157
K.Volarcik et al.
Figure 1. Composite photomicrograph of second meiotic spindles from human oocytes. Each oocyte was stained with an antibody to βtubulin (green) and with a chromatin stain (red) and visualized using a confocal fluorescence microscope. The oocyte in (a) was obtained
from a young donor, whereas the oocytes in (b–f) were obtained from donors over 35 years of age. (a) Normal, bipolar MII spindle
showing tight alignment of all chromosomes at the spindle equator. Note that chromosomes and some microtubule staining are also evident
in the polar body (lower left). (b) A comparable stage oocyte obtained from an older donor. Note that the normal bipolar spindle appears
slightly disrupted with astral microtubules evident. Also notice that the metaphase alignment of the chromosomes is disrupted. Chromosome
and microtubule staining are also evident in the polar body (lower right). (c, d) Additional examples of MII spindles showing gross
disturbances in the metaphase alignment of the chromosomes. All oocytes were obtained from donors ù35 years of age.
Table III. Chromosome segregation at MI
X
18
16
Age (years)
Norm
Abn
Norm
Abn
,35
14
1
11
0
ù35
18
1
14
2
21
Norm
Total
Abn
Norm
Abn
Norm
Abn
7
2
4
0
36
10
3
2
1
44
3
(8%)
7
(14%)
Norm 5 normal, Abn 5 abnormal.
the surface of the ovary of patients undergoing laparoscopic
tubal ligation, provided access to patients of a range of ages.
The primary intent of our study was to analyse the mechanism of age-related non-disjunction in the human oocyte.
Nevertheless, the technique of combined immunofluorescence
staining and FISH allowed us simultaneously to evaluate the
meiotic stage of the oocyte, structural aspects of the meiotic
spindle apparatus and the alignment of the chromosomes, and
the segregation behaviour of specific chromosomes at the first
meiotic division. Since the maternal age curve for human
158
trisomy rises precipitously after the age of 35 years (Hassold
and Chiu, 1985; Hassold, 1986), we subdivided the oocytes
into two groups; those obtained from donors ,35 years and
those obtained from donors 35 years of age and older.
Previous reports of the rate of spontaneous maturation of
human oocytes from unstimulated ovaries range from 30 to
70% (Edwards, 1965; Tsuji et al., 1985; Cha et al., 1991;
Gomez et al., 1993). These rates are low by comparison with
the maturation rate for human immature oocytes obtained from
stimulated ovaries (Gomez et al., 1993) and immature oocytes
Folliculogenesis and the reproductively aged ovary
from non-stimulated ovaries of other mammalian species
(Thibault, 1977). Our meiotic maturation results confirm these
observations. This points out an important limitation in the
use of oocytes from unselected antral follicles of unstimulated
ovaries—that is, a significant proportion of the oocytes are
atretic and will not be suitable for meiotic studies. Additionally,
the in-vitro maturation of oocytes raises the possibility that
the observed abnormalities may be an artefact of in-vitro
culture. However, as detailed below, our data from the subset
of oocytes which underwent spontaneous meiotic maturation
in vitro mimic the in-vivo situation in several respects, and
hence may provide insight to age-related changes in the
human oocyte.
The overall rate of meiotic maturation (as evidenced by
polar body extrusion) was only slightly lower among oocytes
obtained from donors over the age of 35 years. Nevertheless,
analysis of MII arrested oocytes from these donors provides
evidence of an age-related decline in the meiotic competence
of human oocytes. Specifically, the majority of oocytes from
donors 35 years of age and older that resumed and successfully
completed the first meiotic division in vitro had defects in
second meiotic metaphase spindle formation and/or chromosome alignment. These observations confirm and extend the
recent findings of Battaglia et al. (1996), who studied a
somewhat different population of oocytes. That is, we studied
oocytes from antral follicles of unstimulated ovaries, while
they examined periovulatory follicles exposed to exogenous
human chorionic gonadotrophin (HCG) 32 h prior to collection,
and donors were restricted to women aged 20–25 years and
40–45 years. Despite these differences, the incidence of second
meiotic metaphase defects was virtually identical: Battaglia
et al. reported aberrations in 17% of the MII oocytes obtained
from 20–25 year old donors and 79% of oocytes obtained
from donors aged 40–45 years, whereas in our studies the
frequency of aberrations was 21% for oocytes from donors
,35 years and 71% for oocytes from donors 35 years of age
and older. The striking similarity between the studies suggests
that in-vitro matured oocytes obtained from unselected preantral follicles can serve as a useful model for the more
difficult-to-obtain human periovulatory oocyte.
In addition to defects in the second meiotic metaphase of
MII arrested oocytes, we observed errors in chromosome
alignment in a small group of oocytes obtained at first meiotic
metaphase. In contrast to second meiotic metaphase which is
a cell cycle arrest point, first meiotic metaphase is a transient
phase, making detailed analysis of meiotic spindle formation
and chromosome alignment extremely difficult. Although the
number of oocytes we analysed was limited, our results suggest
that age-related defects in chromosome alignment are also a
feature of the first meiotic division. This is not surprising,
since the vast majority of errors in chromosome segregation
have their genesis at MI (Hassold et al., 1993). Indeed the
results of our segregation analysis of chromosomes 16, 18, 21
and the X chromosome in MII arrested oocytes suggest a high
frequency of MI errors in in-vitro matured, unstimulated
human oocytes.
Taken together, the meiotic aberrations we observed in
oocytes from reproductively aged donors suggest that the
meiotic competence of the human oocyte becomes severely
compromised with age. Similarly, the success rate for IVF
involving women .40 years indicates an age-related reduction
in the developmental competence of the human oocyte. Not
surprisingly, fewer oocytes are obtained following hyperstimulation of reproductively aged women and, although the morphology, fertilization rate and rate of embryo transfer have not
been significantly different, the pregnancy rate is dramatically
reduced by comparison with younger patients (Wood et al.,
1992). Furthermore, the advent of ovum donation programmes
has clearly demonstrated that the problem is related to oocyte
quality; for older women the rate of implantation is higher
with donor than with self oocytes and, additionally, the
pregnancy rate is not different between young donors and
reproductively aged recipients (Navot et al., 1991).
The study of oocytes obtained at successive stages of growth
during the first wave of folliculogenesis in the murine ovary
has demonstrated that both the ability to resume and complete
the first meiotic division and the ability to support embryonic
development are acquired during the late stages of oocyte
growth (Eppig et al., 1994). Accordingly, we hypothesize that
the age-associated meiotic defects observed in our studies and
those of Battaglia et al. (1996) and the reduced developmental
potential of human oocytes from older IVF patients are
the result of a defect in the oocyte growth process in the
reproductively aged ovary. According to this hypothesis, the
well characterized age-related increase in meiotic chromosome
non-disjunction represents one symptom of a more generalized
decline in oocyte quality.
If this hypothesis is correct, it suggests that the primary
defect may be at the somatic level. In the mouse, defects in
oocyte maturation have been recognized in several strains.
The two best characterized are the I/LnJ and LT/Sv strains,
which show atypical progression of both nuclear and cytoplasmic maturation. Although the nature of the defect(s)
remains unknown, it has been postulated to be at least partly
due to somatic factors (reviewed in Eppig and Wigglesworth,
1994). A less well characterized mouse, the XYPos sex reversed
mouse, shows defects in the meiotic maturation and complete
developmental incompetence of all oocytes (Hunt, in preparation and Merchant-Larios et al., 1994). Interestingly, meiotic
studies reveal an extraordinarily high incidence of disturbances
in meiotic chromosome behaviour (including a high incidence
of meiotic non-disjunction) which are strikingly similar to the
defects we have observed in human oocytes obtained from
older donors (Hunt and LeMaire, 1992 and Hunt, unpublished).
The characterization of the somatic defects in these mouse
mutants will serve to increase our understanding of somatic
influences on oocyte growth and development and may provide
insight to human age-related aneuploidy.
Whether the age-related increase in meiotic non-disjunction
is unique to the human female or is a characteristic of other
long-lived mammals remains unclear. Information on other
mammalian species is limited and the best studied non-human
mammal, the mouse, shows only a modest age-related increase
in meiotic non-disjunction by comparison with the human
female. However, the fact that the first meiotic division in the
female is the longest known cell division, being initiated
159
K.Volarcik et al.
during the fetal period but not completed until ovulation—
over 50 years for women at the end of their reproductive
lifespan—has formed the basis for many of the hypotheses
proposed to explain human meiotic non-disjunction. According
to our hypothesis, it is not the length of the division itself that
is the important factor in age-related non-disjunction, but the
fact that the oocyte acquires the ability to resume and complete
the first meiotic division during the final stages of this growth
(Eppig et al., 1994).
If our conclusions are correct, i.e. if the entire process of
folliculogenesis becomes compromised in the reproductively
aged ovary, then the development of a culture system to
support the in-vitro growth of human follicles, as has recently
been achieved in the mouse, may provide a means of circumventing these problems and hence of obtaining high quality
oocytes from women in the latter years of their reproductive
lifespan. Conversely, defining the culture conditions necessary
to support normal growth and development of the immature
human oocyte provides a powerful new approach to understanding the factors that control and influence the meiotic process
and, importantly, to the age-related changes that result in the
extraordinarily high frequency of errors in meiotic chromosome
segregation in our species.
Acknowledgements
We gratefully acknowledge the support of the Gynecology attending
physicians and surgical staff at University Hospitals of Cleveland for
their participation in the ovarian aspirations. We thank Nina Desai
and Terry Hassold for helpful comments on the manuscript and Paula
Embury and Renée LeMaire-Adkins for technical support. This research
was supported by NIH grant R01 HD31866 and a grant from the Emory
University Research Committee to PAH. FSH was generously provided
by the National Hormone and Pituitary Program, the National Institute
of Diabetes and Digestive and Kidney Diseases, the NICHD, and the
US Department of Agriculture. Ovarian tissue specimens were obtained
through the NCI-funded Human Cooperative Tissue Network at Case
Western Reserve University.
References
Ayme, S. and Lippman-Hand, A. (1982) Maternal-age effect on aneuploidy:
does altered embryonic selection play a role? Am. J. Hum. Genet., 34,
558–565.
Battaglia, D.E., Goodwin, P. and Klein, N.A. (1996) Influence of maternal
age on meiotic spindle assembly in oocytes from naturally cycling women.
Hum. Reprod., 11, 2217–2222.
Cha, K.Y., Koo, J.J., Ko, J.J. et al. (1991) Pregnancy after in vitro fertilization
of human follicular oocytes collected from nonstimulated cycles, their
culture in vitro and transfer in a donor oocyte program. Fertil. Steril., 55,
109–113.
Crowley, P., Gulah, D., Hayden, T. et al. (1979) A chiasma–hormonal
hypothesis relating Down’s syndrome and maternal age. Nature, 280,
417–418.
Edwards, R.G. (1965) Maturation in vitro of human oocytes. Lancet, 2, 926.
Eichenlaub-Ritter, U., Chandley, A.C. and Gosden, R. G. (1988) The CBA
mouse as a model for age-related aneuploidy in man: studies of oocyte
mutation, spindle formation and chromosome alignment during meiosis.
Chromosoma, 96, 220–226.
Eppig, J.J. and Wigglesworth, K. (1994) Atypical maturation of oocytes of
strain I/LnJ mice. Hum. Reprod., 9, 1136–1142.
Eppig, J.J., Shultz, R.M., O’Brien, M. and Chesnel, F. (1994) Relationship
between the developmental programs controlling nuclear and cytoplasmic
maturation of mouse oocytes. Dev. Biol., 164, 1–9.
Gaulden, M.E. (1992) Maternal age effect: the enigma of Down syndrome
and other trisomic conditions. Mutat. Res., 296, 69–88.
160
Gomez, E., Tarin, J.J. and Pellicer, A. (1993) Oocyte maturation in humans:
the role of gonadotropins and growth factors. Fertil. Steril., 60, 40–45.
Hassold, T. (1986) Chromosome abnormalities in human reproduction. Trends
Genet., 2, 105–110.
Hassold, T. and Chiu, D. (1985) Maternal age specific rates of numerical
chromosome abnormalities with special reference to trisomy. Hum. Genet.,
70, 11–17.
Hassold, T., Hunt, P.A. and Sherman, S. (1993) Trisomy in humans: incidence,
origin and etiology. Curr. Opin. Genet. Dev., 3, 398–4903.
Henderson, S.A. and Edwards, R.G. (1968) Chiasma frequency and maternal
age in mammals. Nature, 218, 22–28.
Hunt, P.A. and LeMaire, R. (1992) Sex-chromosome pairing: evidence that
the behavior of the pseudoautosomal region differs during male female
meiosis. Am. J. Hum. Genet., 50, 1162–1170.
Hunt, P.A., LeMaire, R. Embury, P. et al. (1995) Analysis of chromosome
behavior in intact mammalian oocytes: monitoring the segregation of a
univalent chromosome during mammalian female meiosis. Hum. Mol.
Genet., 4, 2007–2012.
LeMaire-Adkins, R., Radke, K. and Hunt, P.A. (1997) Lack of checkpoint
control at the metaphase-anaphase transition: a mechanism of meiotic nondisjunction in mammalian females. J. Cell Biol., in press.
Martin, R.H., Mahadevan, M.M. and Taylor, P.J. (1986) Chromosomal analysis
of unfertilized human oocytes. J. Reprod. Fert., 78, 673–678.
Merchant-Larios, H., Clarke, H.J. and Taketo, T. (1994) Developmental arrest
of fertilized eggs from the B6.YDOM sex-reversed female mouse. Dev.
Genet., 15, 435–442.
Munné, S., Grifo, J., Cohen, J. and Weier, H.-U.G. (1994) Chromosome
abnormalities in human arrested preimplantation embryos: a multiple-probe
FISH study. Am. J. Hum. Genet., 55, 150–159.
Navot, D., Bergh, P.A., Williams, M.A et al. (1991) Poor oocyte quality rather
than implantation failure as a cause of age-related decline in female fertility.
Lancet, 337, 1375–1377.
Pellestor, F. (1991) Frequency and distribution of aneuploidy in human female
gametes. Hum. Genet., 86, 283–288.
Penrose, L.S. (1965) Mongolism as a Problem in Human Biology; The Early
Conceptus, Normal and Abnormal. Thesis, University of St. Andrews,
Dundee, pp. 94–97.
Plachot, M., Veiga, A., Montagut, J. et al. (1988) Are clinical and biological
IVF parameters correlated with chromosomal disorders in early life: a
multicentre study. Hum. Reprod., 3, 627–635.
Sugawara, S. and Mikamo, K. (1983) Absence of correlation between univalent
formation and meiotic nondisjunction in the aged female Chinese hamster.
Cytogenet. Cell Genet., 35, 34–40.
Tarin, J.J. (1995) Aetiology of age-associated aneuploidy: a mechanism based
on the ‘free radical theory of ageing’. Mol. Hum. Reprod., 1, 1563–1565.
Tarkowski, A.K. (1966) An air drying method for chromosome preparations
from mouse eggs. Cytogenetics, 5, 394–400.
Thibault, C. (1977) Are follicular maturation and oocyte maturation
independent processes? J. Reprod. Fertil., 51, 1–15.
Tsuji, K., Sowa, M. and Nakano, R. (1985) Relationship between human
oocyte maturation and different follicular sizes. Biol. Reprod., 32, 413–417.
Van Blerkom, J. and Henry, G. (1992) Oocyte dysmorphology and aneuploidy
in meiotically mature human oocytes after ovarian stimulation. Hum.
Reprod., 7, 379–390.
Warburton, D. (1989) The effect of maternal age on the frequency of trisomy:
change in meiosis or in utero selection? In Hassold, T.G. and Epstein, C.J.
(eds), Molecular and Cytogenetic Studies of Non-disjunction. Alan R. Liss,
New York, pp. 165–181.
Wood, C., Calderon, I. and Crombie, A. (1992) Age and fertility: results of
assisted reproductive technology in women over 40 years. J. Assist. Reprod.
Genet., 9, 482–484.
Wramsby, H., Fredga, K. and Liedholm, P. (1978) Chromosome analysis of
human embryos recovered from preovulatory follicles in stimulated cycles.
N. Engl. J. Med., 316, 121–124.
Received on May 8, 1997; accepted on September 24, 1997