American Journal of Medical Genetics 39321-331 (1991)
Distribution of Aneuploidy in Human Gametes:
Comparison Between Human Sperm and Oocytes
Renee H. Martin, Evelyn KO, and Alfred Rademaker
Division of Medical Genetics, Department of Pediatrics, Faculty of Medicine, University of Calgary (R.H.M.), and
Medical Genetics Clinic, Alberta Children$ Hospital (R.H.M., E.K.), Calgary, Alberta, Canada; Cancer Center,
Biometry Section, Northwestern University Medical School, Chicago, Illinois (A.R.I
The frequency and distribution of aneuploidy
was compared in 11,615 karyotyped human
sperm and 772 karyotyped human oocytes to
determine if all chromosomes are equally
likely to be involved in aneuploid events or if
some chromosomes are particularly susceptible to nondisjunction. The frequency of hypohaploidy and hyperhaploidy was compared
among different chromosome groups and individual chromosomes for human sperm and
oocytes. In general, hypohaploid chromosome complements were more frequent than
hyperhaploid complements, in sperm and oocytes. The distribution of chromosome loss in
the hypohaploid complements indicated that
significantly fewer of the large chromosomes
and significantly more of the small chromosomes were lost, suggesting that technical loss
predominantly affects small chromosomes. A
conservative estimate of aneuploidy (2 X hyperhaploidy) was approximately 3-4% in the
human sperm and 18-19% in human oocytes.Al1 chromosome groups were represented among hyperhaploid human sperm
and oocytes. For human sperm, the observed
frequency of hyperhaploidy equaled the expected frequency based on the assumption
that the frequency of nondisjunction is equal
for all chromosome groups, with two exceptions: group G and the sex chromosomes.
Among individual chromosomes in human
sperm, chromosomes 1 and 21 and the sex
chromosomes had a significant excess of hyperhaploidy. For human oocytes, there were
fewer hyperhaploid oocytes than expected for
chromosome groups C and F and more than
expected for chromosome groups D and G.
Among individual chromosomes there was a
Received for publication April 2,1990; revision received July 18,
1990.
Address reprint requests to Dr. R.H. Martin, Medical Genetics
Clinic, Alberta Children’s Hospital, 1820 Richmond Rd. S.W., Calgary, Alberta, Canada T2T 5C7.
0 1991 Wiley-Liss, Inc.
significant excess for chromosome 21. These
results indicate that all chromosomes are susceptible to nondisjunction but that chromosome 21 is particularly prone to aneuploidy in
both human sperm and oocytes. They also
demonstrate that sex chromosome aneuploidy is common in human sperm but not in
human oocytes.
KEY WORDS: human sperm chromosomes,
human oocyte chromosomes,
cytogenetics of human gametes, etiology of aneuploidy
INTRODUCTION
Aneuploidy has been observed for almost every chromosome in human spontaneous abortions [Hassold et
al., 19801 whereas in human liveborn infants aneuploidy is generally only seen for chromosomes 13, 18,
and 21 and the sex chromosomes [de Grouchy and Turleau, 19841. There is some controversy as to whether
these chromosomes are particularly susceptible to nondisjunction or whether aneuploidy for other chromosomes is not compatible with survival to term. Many
studies have focused on the former explanation and have
attempted to elucidate factors that make the D and G
group chromosomes more prone to nondisjunction. For
example, satellite association of the acrocentric chromosomes [Hansson, 1979; Henderson et al., 19731, or a
relatively smaller number of chiasmata [Henderson and
Edwards, 1968; Luthardt et al., 19731have been postulated as the cause of an increased frequency of nondisjunction for these chromosomes. Data from studies on
spontaneous abortions indicate that survival of various
specific aneuploidies is largely responsible for the
skewed representation a t birth since trisomies that are
common in spontaneous abortion (e.g.,trisomy 16)never
survive to term [Hassold et al., 19801.
It is important to determine the incidence of nondisjunction for different chromosomes because this information will give us important clues about the mechanisms of nondisjunction and survival of chromosomally
abnormal embryos. If all chromosome groups have the
322
Martinet al.
same frequency of nondisjunction, then the mechanism
that causes nondisjunction must be common to all chromosomes, e.g., errors of spindle formation or attachment. If specific chromosomes have a n increased frequency of nondisjunction, then we could focus on the
specific characteristics of the chromosomes (e.g., small
size, presence of heterochromatin, nucleolar organizing
regions) to elucidate some of the factors that influence
the rate of nondisjunction.
During the last decade it has become possible to study
the chromosome constitution of human gametes. The
chromosomes in human sperm can be visualized after
cross-species fertilization with golden hamster oocytes
[Rudak et al., 1978; Martin, 19831. Analysis of “spare
oocytes” from in vitro fertilization (IVF) programs provided our first information on the chromosome constitution of human oocytes [Wramsby and Liedholm, 1984;
Martin et al., 1986bl.. A number of laboratories from
around the world have reported on these cytogenetic
analyses of human sperm and oocytes. Many of the
studies were based on small samples. We have pooled
results from these studies to provide information on the
frequency and distribution of aneuploidy among the various chromosome groups in human sperm and oocytes.
MATERIALS AND METHODS
Studies on Human Sperm
Published reports on cytogenetic analyses of human
sperm were included if the specific karyotypes of the
aneuploid complements were detailed in the report. We
also included recent unpublished studies from our laboratory. The majority of the donors were normal healthy
men (Table I). Data from 31 men with constitutional
chromosomal rearrangements (such as translocations
and inversions) were also included if no interchromosoma1 effect had been ascertained in these studies. Aneuploid complements unrelated to the chromosome rearrangement were analyzed from the studies presented in
Table 11.
Sperm capacitation and pretreatment differed in the
studies. Most studies used BWW medium [Martin,
19831 for capacitation of sperm [Balkan et al., 1983;
Balkan and Martin, 1983a, b; Benet and Martin, 1988;
Brandriff et al., 1985a; Martin, 1984,1986,1988a;Martin et al., 1986a, 1987; Rudak et al., 19781. Test-yolk
buffer [Brandriff et al., 1985bl was also used for storage
and capacitation of sperm [Brandriff et al., 1985a, 1986;
Martin, 1988b, 1990; Martin et al., 1990a, b; Templado
et al., 19881. Some studies did not clearly identify which
type of sperm pretreatment was used [Jenderny and
Rohrborn, 1987; Pellestor et al., 1987, 1989; Sele et al.,
19851. We have demonstrated that the two methods of
sperm pretreatment have no effect on the frequency of
aneuploidy in human sperm [Martin et al., 19891.
Some aneuploid sperm complements were classified
only to the level of chromosome group whereas others
had the extra individual chromosome identified. Therefore the analysis was performed both for chromosome
groups and for individual chromosomes. Since some
studies have demonstrated a n excess of hypohaploid
complements compared to hyperhaploid complements
that might reflect technical artefact for some of the
hypohaploid complements [Martin et al., 1987; Pellestor
et al., 19871, the analysis was performed separately for
hypohaploid and hyperhaploid complements rather
than lumping them together as aneuploid.
Studies on Human Oocytes
Published reports on cytogenetic analyses of human
oocytes were included if the specific karyotypes of the
aneuploid complements were detailed in the report.
Studies on fertilized oocytes and embryos were not included since we wanted to be able to separate maternal
and paternal errors. There reports are presented in
Table 111.
Follicular stimulation protocols differed among the
studies and these are recorded in Table 111. Most studies
have not found a correlation between the frequency of
chromosome abnormalities and the use of different stimulation protocols [Plachot et al., 1988; van Blerkom and
Henry, 19881.Most studies identified chromosomes only
to the chromosome group and not to individual chromosomes. However, a small number of investigators did
identify individual chromosomes so the analysis was
performed both for chromosome groups and for individual chromosomes. Since many studies on human oocytes
demonstrated a n excess of hypohaploid complements
compared to hyperhaploid complements that might reflect technical artefact for some of the hypohaploid complements [Djalali et al., 1988; Martin et al., 1986b1,the
analysis was performed separately for hypohaploid and
hyperhaploid complements, as it was for the sperm
karyotypes.
Data Analysis
The aneuploid chromosome complements were classified a s hyperhaploid (n + 1)or hypohaploid (n - 1).The
data were analyzed for the frequency of hyperhaploidy
and hypohaploidy, respectively, within the Denver chromosome groups (A through G and the sex of chromosomes) and also for individual chromosomes. For statistical analysis, it was assumed that all chromosomes had
a n equal probability of nondisjunction and two-tailed
binomial tests using exact binomial probabilities were
performed [Rosner, 19861 to determine if each chromosome (or chromosome group) had a frequency of hyperhaploidy (or hypohaploidy) that differed significantly
from the expected frequency. The analyses were performed separately for sperm chromosome complements
from normal donors and donors with a constitutional
chromosomal rearrangement. Chi square analysis was
performed to determine if there were significant differences between the two groups.
RESULTS
Human Sperm
A total of 11,615human sperm chromosome complements was analyzed including 8,356 sperm complements from normal donors and 3,259 complements from
men with constitutional chromosomal abnormalities.
A summary of the frequency and type of chromosome
abnormalities in 8,356 sperm from normal donors is
presented in Table I. The men were 18 to 55 years old.
The mean frequency of numerical abnormalities in the
-~
.~
Brandriff et al., 1985
Jenderny and Rohrborn,
1987
Martin et al., 1987 and
unpublished data
Rudak et a1 , 1978
Sele et al., 1985
Mean
Study
~
60
70
5,629
18-55 (32.3)
Unknown
Unknown
2,468
129
No. of
karyotypes
21-49 (32.9)
Unknown
Age range
(mean)
3.3
7.1
3.1
3.5
0.9
0.8
5%
Hypoploid
~
1.7
5.7
1.9
0.6
0.7
0.8
5.0
12.9
5.0
4.2
1.7
1.6
~
Aneuploid
Hyperploid
~~
5%
r%.
~~
~~
3.3
11.4
3.8
1.4
Conservative estimate
of aneuploidy (%)
(2 x hyper)
1.5
1.6
TABLE I. Summarv of Snerm Chromosome Abnormalities in SDerm f i o m Normal Men
C/c
1.7
1.4
5.3
9.4
7.7
6.2
Structural
-
Studv
Reciprocal translocations
Brandriff et al., 1986
Spriggs et al.
(unpubl. data)
Martin et al.
(unpubl. data)
Martin et al., 1990a
Martin et al., 1990b
Martin, 1984
Martin, 1988
Balkan and Martin,
1983
Temulado et al.. 1988
PeLstor et al., '1989
Fbbertsonian translocations
Balkan and Martin,
1983
Syme and Martin
(unpubl. data)
Pellestor et al., 1987
Pellestor, 1990
Martin, 1988
Inversions
Martin, 1990
Balkan et al., 1983
Martin, 1986
Other
Hulten and Martin
(unpubl. data)
Benet and Martin, 1988
Martin et al., 1986
Mean
10.0
5.5
7.7
5.8
2.7
120
399
13
171
37
75
283
40
25-55 (38.6)
30
30
34-39 (36.5)
26
27-35 (31.2)
144
121
94
210
37
32
33
37
55
74
78
67
121
41
40
24
27
30-39 (34.5)
419
35-52 (43.5)
23
0
6.0
183
37
27
0.9
0
1.2
5.5
427
145
0
1.4
1.1
1.4
4.8
5.5
6.8
5.3
0
3.3
0
2.6
0
0
0.2
4.3
4.0
1.1
2.1
2.5
5.3
12.8
6.0
2.5
2.9
0
8.0
8.8
0
0
0
2.3
0
Hyperploid
Hypoploid
25-31 (28)
24-29 (26.5)
%
%
No. of
karyotypes
Age range
(mean)
5.5
0.1
6.5
6.2
2.1
5.8
5.3
15.4
7.5
2.5
3.1
4.3
12.0
9.9
10.8
6.0
7.7
8.2
2.7
6.0
2.1
5.5
Aneuploid
%
%
0
2.7
2.3
10.9
2.7
9.7
19.0
22.9
2.5
5.3
0
6.6
0
2.9
1.3
0
10.7
5.3
8.7
5.3
1.1
5.1
3.0
0
0.5
8.7
8.0
2.1
11.7
19.5
7.7
15.8
18.9
9.8
0
1.7
1.0
0
4.7
0
4.0
20.0
Structural
1.9
0
Conservative estimate
of aneuploidy (%)
(2 x hyper)
TABLE 11. Donors With a Constitutional Chromosome Rearrangement: Summary of Sperm Chromosome Abnormalities Unrelated to
the Chromosome Rearrangement
47
Unknown
CC/HMG, HMG,
FSH, LHRH
agonist/HMG
CC/HCG HCG
*
8
32-40 (?)
44
24-35 (29.4)
135
188
51
24-39 (33.6)
Unknown
CC
Unknown
251
27-42 (33.5)
22-40 (30.6)
34
24-39 (31)
No. of
karyotypes
14
+ HMG/HCG
CC + HMG +
FSH + LH,
2 HCG
CC + HMG
CC, HMG, CC +
HMG
FSH + HMG/HCG
Age range
(mean)
25-38 (?)
1.5
9.2
16.1
17.0
12.5
6.7
12.8
12.5
8.0
2.3
25 .O
10.6
13.7
8.0
13.1
19.6
5.9
Hyperploid
14.3
Hypoploid
21.4
23.5
%
%
25.1
8.1
29.8
25.0
18.6
25.0
33.3
21.1
29.4
Aneuploid
35.7
%
18.5
3.0
34.0
25.0
16.0
4.5
27.5
15.9
11.8
(2 x hyper)
28.6
(%I
Conservative
estimate of
aneuploidy
%
1.2
0
4.3
0
0
4.5
2.0
0.4
0
Structural
abnormals
0
a
CC, clomiphene citrate; HCG, human chorionic gonadotrophin; HMG, human menopausal gonadotrophin; FSH, follicle stimulating hormone; LH, luteinizing
hormone.
Van Blerkom
and Henry,
1988
Mean
Pellestor and
Sele, 1988
Papadopoulos
e t al., 1989
Plachot et
al., 1987
Study
Wramsby et
al., 1987
Djalali et al.,
1988
Bongso et
al., 1988
Ma et al.,
1989
Martin et
al., 198613
Type Of
stimulation"
CC/HCG
TABLE 111. Summary of Chromosome Abnormalities in Human Oocytes
326
Martinet al.
five studies was 5.0% (1.6 to 12.9%) with 1.9% (0.6 to
5.7%) hyperhaploid and 3.1% (0.8to 7.1%) hypohaploid.
Since there was a n excess of hypohaploid complements,
a conservative estimate of aneuploidy was obtained by
doubling the frequency of hyperhaploid complements.
This estimate yielded a mean of 3.8%(1.4 to 11.4%).The
mean frequency of structural abnormalities was 5.3%
(1.4 to 9.4%).
A summary of the frequency and types of chromosome
abnormalities (unrelated to the constitutional chromosome rearrangement) in 3,259 sperm from donors with
rearranged chromosomes is presented in Table 11. The
men were 24 to 55 years old. The mean frequency of
numerical abnormalities in the 21 studies was 6.5% (2.1
to 15.4%) with 1.1%(0 to 4.3%) hyperhaploid and 5.3%
(0 to 12.0%)hypohaploid. The conservative estimate of
aneuploidy was 2.3% (0 to 8.7%).The mean frequency of
structural abnormalities was 9.7% (0 to 22.9%).
The data were analyzed separately for the two groups
of men and then pooled after statistical analysis demonstrated no differences in the two groups of men. The
observed and expected frequency of hyperhaploid sperm
in each chromosome group is presented in Table IV. In
all chromosome groups, the observed frequency of hyperhaploid sperm equaled the expected frequency, with
the exception of the G group and sex chromosomes, both
of which showed a significant increase in the frequency
of hyperhaploidy. The C group and the sex chromosomes
have minimum and maximum values because in some
chromosome spreads it was not possible to distinguish
between a n X chromosome or a C group chromosome.
But even in the minimum sex chromosome category,
there was a highly significant increase in the frequency
of hyperhaploidy.
The observed and expected frequency of hyperhaploid
sperm among individual chromosomes is presented in
Table V. None of the observed numbers was significantly different from the expected number of 3.04, except for chromosomes 1 and 21 and the sex chromosomes. Table V lists the number of hyperhaploid sperm
for each sex chromosome, but the nondisjunctional event
occurred in the homologous pair of sex chromosomes.
Therefore, the minimum value to be compared with the
other chromosomes is 12 (an extra X or Y chromosome).
TABLE V. Hyperhaploid Sperm in
Individual Chromosomes
Chromosome no.
Chromosome group
D
E
F
G*
X or Y**
X or Y or C
Observed"
Expectedb
13
8
11 (min)-20 (max)
9
10
4
19
12
21
12.4
8.3
28.9
12.4
12.4
8.3
8.3
4.1
amin, minimum value; max, maximum value.
Expected = (number ofchromosomes/group)/23 x total hyperhaploid
b
(95).
* P = ,003
**P = ,002
0.02
3.04
2l
3 12 P < ,0001
7
=
[total hyperhaploid (70)1/23
The observed and expected frequency of hypohaploid
sperm in each chromosome group is presented in Table
VI. Groups A, 3, and C had a significant deficit of hypohaploidy, and groups E and G and the sex chromosomes
had a significant excess of hypohaploidy, based on the
assumption that chromosome loss was random. Hypohaploid sperm for individual chromosomes is presented
in Table VII. Chromosomes 1, 2, 3, 4, 7, and 9 had
significantly less hypohaploidy than expected and chromosomes 18, 21, and 22 and the sex chromosomes had
significantly more than expected. Thus, i t appears that
smaller chromosomes were preferentially lost.
Human Oocytes
A total of 772 karyotyped human oocyte chromosome
complements was analyzed from women ranging in age
TABLE IV. Hyperhaploid Sperm in Each
Chromosome Group
A
B
C
=
Expecteda
3
10 P = .002
3
X
Y
X or Y
'Expected
Observed
-__
8P
4
1
4
0
0
1
0
6
2
0
0
2
1
3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
TABLE VI. Hypohaploid Sperm in Each
Chromosome Group
Chromosome.~
group
__
A*
B*
C*
D
E**
F
G**
X or Y**
X or Y or C
Observed
Expecteda
18
21
103 (min)-107 (max)
61
90
52
95
30
34
61.8
41.2
144.3
61.8
61.8
41.2
41.2
20.6
aExpected = (number of chromosomes/group)/23 x total hypohaploid
(474).
* P < 0.05: fewer chromosomes lost than expected.
**P < 0.05: more chromosomes lost than expected.
Aneuploidy in Human Gametes
TABLE VII. Hypohaploid Sperm in
Individual Chromosomes
Chromosome no.
1*
2*
3*
4*
5
6
7*
8
9*
10
11
12
13
14
15
16
17
Observed
9
3
6
6
12
14
8
17
5
17
Exoected"
18.87
18
20
19
12
15
24
20
43
25
23
44
44
30
18**
19
20
21**
22**
X or Y**
quency. For groups C and F, there were significantly
fewer hyperhaploid complements than expected and for
groups D and G, there were significantly more hyperhaploid oocytes. The C group and the X chromosome
have minimum and maximum values because in some
chromosome spreads it was not possible to distinguish
between a n X chromosome or a C group chromosome.
The observed and expected frequency of hyperhaploid
oocytes among individual chromosomes is presented in
Table IX. None of the observed numbers was significantly different from the expected number of 0.3 except
for chromosome 21.
The observed and expected frequency of hypohaploid
oocytes in each chromosome group is presented in Table
X. Groups A, B, and C had a significant deficit of hypohaploidy and groups D and G had a significant excess of
TABLE IX. Hyperhaploid Oocytes i n
Individual Chromosomes
~
Chromosome no.
1
2
3
aExpected = [total hypohaploid (434)1/23.
* P < 0.05: fewer chromosomes lost than expected.
**P < 0.05: more chromosomes lost than expected.
4
from 22 to 42 years. A summary of the frequency and
types of chromosome abnormalities in the oocytes is
presented in Table 111.The mean frequency of numerical
abnormalities in the nine studies was 25.1% (8.1 to
35.7%) with 9.2% (1.5 to 17%)hyperhaploid and 16.1%
(6.7 to 25.0%) hypohaploid. Since there was a n excess of
hypohaploid complements, a conservative estimate of
aneuploidy was obtained by doubling the frequency of
hyperhaploid complements. This estimate yielded a
mean of 18.5% (3 to 34%). The mean frequency of structural abnormalities was 1.2% (0 to 4.5%).
The observed and expected frequency of hyperhaploid
oocytes in each chromosome group is presented in Table
VIII. For chromosome groups A, B, and E, the expected
frequency of hyperhaploidy equaled the observed fre-
-
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
X
UExpected
TABLE VIII. Hyperhaploid Oocytes in Each
Chromosome G r o w
~
~~~~
~
~~~~
Observed.
8
8
2 (mini-6 (max)
18
8
2
31
0 imin)
4 imax)
Expectedb
10.6
7.0
24.6
10.6
10.6
7.0
7.0
3.5
amin, minimum; max, maximum value.
bExpected = (number of chromosomes/group)23 x total hyperhaploid
(81).
* P < .05: fewer hyperhaploid oocytes.
**P< .05: more hyperhaploid oocytes.
***P < ,0001: more hyperhaploid oocytes.
=
Observed
ExDecteda
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
2
0
0
3 P = ,005
1
0
0.3
[total hyperhaploid (8,1123.
TABLE X. Hypohaploid Oocytes in Each
Chromosome Group
~~~
Chromosome group
A
B
C*
D**
E
F*
G***
X
X or C
327
Chromosome group
A*
B*
C*
D**
E
F
G**
X
X or C
Observed
6
7
31 (mink36 (max)
40
24
16
38
4
9
Expecteda
.22.3
14.9
52.0
22.3
22.3
14.9
14.9
7.4
aExpected = (number of chromosomes/group)/23 x total hypohaploid
1171).
* P < 0.05: fewer chromosomes lost than expected.
**P < 0.05: more chromosomes lost than expected.
328
Martinet al.
hypohaploidy, based on the assumption that chromosome loss was random. Hypohaploidy among individual
chromosomes is presented in Table XI. Only chromosome 18 showed a significant excess of hypohaploidy.
DISCUSSION
The summary of chromosome abnormalities in human sperm and oocytes clearly demonstrates that we
have a much larger body of knowledge on human sperm
than human oocytes since we have summarized data on
11,615 karyotyped human sperm and only 772 karyotyped human oocytes. One of the reasons for this discrepancy is the difficulty of obtaining human oocytes. Millions of human sperm are present in a single ejaculate
whereas surgery is required to obtain a few human
oocytes. Also the cross-species fertilization technique in
which hamster oocytes are fertilized by human sperm to
produce pronuclear human sperm chromosomes [Rudak
e t al., 1978; Martin, 19831 has provided a method to
analyze the chromosomes in human sperm at a mitotic
stage. These mitotic chromosomes are amenable to
banding techniques and can be analyzed with the same
precision as somatic chromosomes. Thus, most reports
on human sperm chromosome complements provide detailed information with every individual chromosome
identified and the location of chromosome breaks specified [Brandriff et al., 1985; Martin et al., 19871. In
contrast, data on human oocytes has largely come from
IVF programs in which the oocytes were not fertilized
and were then used for cytogenetic analysis. The morphology of meiotic chromosomes is poor and banding is
often unsatisfactory. Therefore, chromosomes are often
identified only to the Denver groups and not individual
chromosomes. For example, Plachot et al. [1987] reported on cytogenetic abnormalities in 151 unfertilized
TABLE XI. Hypohaploid Oocytes in
Individual Chromosomes
Chromosome no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
X
Observed
0
1
0
0
0
1.7
0
1
2
5 P
4
4
2
3
Expected”
=
,047
4
aExpected = [total hypohaploid (38)1123.
oocytes. For the majority of the oocytes, only chromosome counts were provided; 47 oocytes were analyzed to
the level of the Denver chromosome groups and none
were identified to the level of individual chromosomes.
Thus, the data on human oocytes are not as reliable as
those on human sperm.
The mean frequency of chromosome abnormalities in
human sperm is approximately 10% (Tables I and 11).
The overall frequency of abnormalities in human oocytes is approximately 20% (Table 111).This frequency of
abnormalities in human oocytes must be considered a
minimal estimate because it does not take into account
abnormalities from the second meiotic division. The
higher frequency of chromosome abnormalities in human oocytes compared to sperm agrees with conclusions
drawn from the investigation of the parental origin of
autosomal trisomies in which the maternal contribution
is invariably higher than the paternal [Juberg and
Mowrey, 1983; Kupke and Muller, 19891.
Another difference between the human sperm and
oocyte data is the proportion of gametes with numerical
and structural abnormalities. Aneuploidy accounts for
less than half of the abnormalities in human sperm,
whereas more than 90% of the abnormalities are aneuploid in human oocytes. Structural abnormalities are
observed in approximately 5%of human sperm whereas
only 1%are observed in human oocytes (Tables I, 11, and
111). These data agree with results from studies on human newborn infants demonstrating that more than
80% of de novo chromosome structural abnormalities
are paternal in origin [Ejima et al., 1988; Olson and
Magenis, 19881.
In most cytogenetic studies of human sperm and oocytes there has been a n excess of hypohaploid complements compared to hyperhaploid complements (Tables I,
11, and 111). Theoretically, the process of nondisjunction
should provide a n equal frequency of hyperhaploidy and
hypohaploidy. Most of the studies use a variation of
Tarkowski’s technique [Tarkowski, 19661 for chromosome fixation and this technique is associated with chromosome loss [Kamiguchi and Mikamo, 1986; Martin et
al., 19761. Analysis of the hypohaploid complements
demonstrated that significantly fewer large chromosomes and significantly more small chromosomes were
lost than expected for both human sperm (Tables VI and
VII) and oocytes (Tables X and XI). It could be argued
that these results support the contention that some hypohaploidy results from technical artefacts, since small
chromosomes might be lost more easily from the slide
during fixation than large chromosomes. However,
there was a significantly increased frequency of hyperhaploidy for G group chromosomes (in particular chromosome 21) for both human sperm (Tables IV and V) and
oocytes (Tables VIII and IX) demonstrating that a t least
part of the excess in this group was caused by nondisjunction. It is also possible that anaphase lag could preferentially affect small chromosomes and be responsible
for some of the excess hypohaploidy.
All chromosome groups were represented among hyperhaploid human sperm and oocytes (Tables IV and
VIII).
In human sperm, the observed frequency of hyper-
Aneuploidy in Human Gametes
haploidy equaled the expected frequency based on the
assumption that the frequency of nondisjunction is
equal in all chromosome groups, with two exceptions:
group G and the sex chromosomes. There was a highly
significant increase in the frequency of hyperhaploidy
in both of these chromosome groups. Among individual
chromosomes in human sperm the distribution of hyperhaploidy was again quite uniform with the exception of
chromosomes 1and 21 and the sex chromosomes, which
all had a significant excess of hyperhaploidy (Table V).
The increased frequency of hyperhaploidy of chromosome 1 is surprising since this is the only chromosome
that has not been observed as a trisomy in spontaneous
abortions. However trisomy 1has been observed in a n
eight-cell human pre-embryo [Watt et al., 19871and i t is
possible that trisomy 1is common in human conceptuses
but is always lost in a n early preimplantation stage. The
increased frequency of hyperhaploidy of chromosome 21
and the sex chromosomes was even more dramatic.
These results corroborate data from studies on human
liveborn offspring and spontaneous abortions. A number
of studies have demonstrated that the paternal contribution to trisomy 21 is approximately 20% [Juberg and
Mowrey, 19831whereas other autosomal trisomies have
a much lower frequency of paternal origin [Hassold et
al., 1984; Kupke and Muller, 19891. Therefore it is not
surprising that we found a n increased frequency of hyperhaploidy of chromosome 21.
The highly significant increase in the frequency of
hyperhaploidy of the sex chromosomes in human sperm
is interesting since, in many studies of meiosis I in
human males, the X and Y chromosomes are often unpaired [Laurie and Hulten, 19851. It is possible that
these univalents might predispose to nondisjunction of
the sex chromosomes in males. There are four common
gonosomal aneuploidies in humans: 4 5 3 , 47,XYY,
47,XXY, and 47,XXX. Recent evidence has demonstrated that most of these originate from a paternal
error, unlike autosomal aneuploidies, in which maternal errors predominate [Hassold et al., 19841.Hassold et
al. [19881, in a combined cytogenetic and molecular
analysis of 35 spontaneously aborted and 5 liveborn
45,X conceptions, determined that 80% had occurred
because of the loss of a paternal sex chromosome. This
value agrees well with the earlier estimate of 77% paternal errors based on the Xg blood-group data of Sanger et
al. [1977]. In the data from human sperm chromosome
complements, there was a significantly increased frequency of sex-chromosomal hyperhaploidy and a concomitant increase in the frequency of sex-chromosomal
hypohaploidy. All cases of 47,XYY must originate from
paternal nondisjunction, and Jacobs et al. [19881 have
demonstrated that the additional chromosome in
47,XXY conceptions is paternal in half of the cases. Our
data, demonstrating a n excess of gonosomal aneuploidy
in sperm, corroborate these results. Apparently there
has been only one report on the origin of the extra X
chromosome in 47,XXX conceptuses, and this study
demonstrated more maternal errors, although only nine
cases were studied [May et al., 19881.
In human oocytes, there were fewer hyperhaploid oocytes than expected for chromosome groups C and F and
329
more than expected for chromosome groups D and G
(Table VIII). These trends mirror the data obtained from
studies of human spontaneous abortions wherein trisomies for groups C and F are rare whereas trisomies for
groups D and G are common [Hassold et al., 19801. Since
very few human oocytes have been analyzed cytogenetically to the level of individual chromosomes, the
data on hyperhaploidy on individual chromosomes is
very scanty. However, there was a significant excess of
chromosome 21. This is not unexpected since trisomy 21
is the most common trisomy observed among human
liveborn infants [de Grouchy and Turleau, 19843. Although trisomy 16 is the most common autosomal trisomy in spontaneous abortions [Hassold et al., 19801,
hyperhaploidy of chromosome 16 was not observed in
human oocytes. This may simply reflect the small sample size.
Overall, the distributions of aneuploidy in human
sperm and oocytes have marked similarities. Firstly, all
chromosome groups are represented among hyperhaploid human sperm and oocytes, not just those that
are common in human liveborn infants and spontaneous
abortions. This indicates that there must be a common
mechanism of nondisjunction affecting all chromosomes, such as spindle fiber formation or kinetochore
attachment. Secondly, there is a highly significant increase in the frequency of hyperhaploidy for chromosome 21 in both human sperm and oocytes. Therefore
there must be some aspect of chromosome21 that causes a
predisposition to nondisjunction. A number of theories
have been proposed, e.g., satellite association [Hanson,
1979; Henderson et al., 19731, which could affect meiosis
in men and women. The increased incidence of nondisjunction of chromosome 21 in women might be explained
by nucleolar persistence during the long dictyotene stage
in female meiosis [Vagner-Capodanoet al., 19871. Henderson and Edwards 119681 suggested a lack of pairing of 21
because of decreased number of chiasmata. Warren et al.
[19871examined recombination frequencies between polymorphic markers on chromosome 21 and found evidence
for asynapsies rather than desynapsis. This mechanism
would also be compatible with errors in male and female
meiosis. However,none of these theories offers a n explanation of why chromosome 21 would have a higher frequency
of nondisjunction than chromosome 22.
There are also major differences in the data from human
sperm and oocytes. Firstly, there is a much higher frequency of chromosome abnormalities in human oocytes
than in human sperm, especially when one considers only
numerical abnormalities. The mean conservative estimate of aneuploidy in human oocytes was approximately
18-19% compared to 3-4% in human sperm. Secondly,
there was a highly significant increase in the frequency of
hyperhaploidy (and hypohaploidy) for the sex chromosomes in human sperm that was not observed in human
oocytes. This suggests that much of the gonosomal aneuploidy of spontaneous abortions and liveborn infants represents paternal nondisjunction. Data corroborating these
findings in human sperm are emerging from studies of the
parental origin of sex chromosome aneuploidies in human
liveborn infants and spontaneous abortions (as discussed
above).
330
Martinet al.
frequency in human males with normal karyotypes. Ann Hum Genet
49:189-201.
Luthardt FW, Palmer CG, Uy P-L (1973): Chiasma and univalent frequencies in aging female mice. Cytogenet Cell Genet 12:68-79.
Ma S, Kalousek DK, Zouves C, Yuen BH, Gomel V, Moon YS (1989):
Chromosome analysis of human oocytes failing to fertilize in vitro.
Fertil Steril 51:992-997.
Martin RH (1983): A detailed method for obtaining preparations of human sperm chromosomes. Cytogenet Cell Genet 35253-256.
Martin
RH (1984): Analysis of human sperm chromosome complements
REFERENCES
from a male heterozygous for a reciprocal translocation
t(11;2)(q23qll). Clin Genet 25:357-361.
Balkan W, Martin RH (1983a): Chromosome segregation into the spermatozoa of two men heterozygous for different reciprocal translocaMartin RH (1986): Sperm chromosome analysis in a man heterozygous
tions. Hum Genet 63:345-348.
for a paracentric inversion of chromosome 7(qllq22). Hum Genet
73:97-100.
Balkan W, Martin RH (198313): Segregation of chromosomes into the
spermatozoa of a man heterozygous for a 14;21 Robertsonian trans- Martin RH (1988a):Meiotic segregation of human sperm chromosomes in
location. Am J Med Genet 16:169-172.
translocation heterozygotes: report of a t(9;10)(q34qll)and a review of
Balkan W, Burns K, Martin RH (1983):Sperm chromosome analysis of a
the literature. Cytogenet Cell Genet 47:48-51.
man heterozygous for a pericentric inversion of chromosome number
Martin RH (1988b): Cytogenetic analysis of sperm from a male hetero3. Cytogenet Cell Genet 35:295-297.
zygous for a 13;14 Robertsonian translocation. Hum Genet 80:
Benet J, Martin RH (1988): Sperm chromosome complements in a
357-361.
47,XYY man. Hum Genet 78:313-315.
Martin RH (1990): A cytogenetic study of the frequency of recombined
Bongso A, Chye NS, Ratman S, Sathananthan H, Wong PC (1988):
chromosomes in sperm from a man heterozygous for an inversion of
Chromosome anomalies in human oocytes failing to fertilize after
chromosome 3 (inv(3)(p25;q21)).Am J Hum Genet (in press).
insemination in vitro. Hum Reprod 3:645-649.
Martin RH, Dill FJ, Miller JR (1976): Nondisjunction in aging female
Brandriff B, Gordon L, Ashworth L, Watchmaker G, Moore D, Wyrobek
mice. Cytogenet Cell Genet 17:150.
A J , Carrano AV (1985a): Chromosomes of human sperm: variability
Martin RH, Hildebrand KA, Yamamoto J , Peterson D, Rademaker AW,
among normal individuals. Hum Genet 70:18-24.
Taylor P, Lin CC (1986a): The meiotic segregation of human sperm
Brandriff B, Gordon L, Watchmaker G (1985b):Human sperm chromochromosomes in two men with accessory marker chromosomes. Am J
somes obtained from hamster eggs after sperm capacitation in TESTMed Genet 25381-388.
yolk buffer. Gamete Res 11:253-259.
Martin RH, Mahadevan MM, Taylor PJ, Hildebrand K, Long-Simpson L,
Brandriff B, Gordon L, Ashworth LK, Littman V, Watchmaker G, CarPeterson D, Yamamoto J, Fleetham J (1986b):Chromosomal analysis
ran0 AV (1986):Cytogenetics of human sperm: meiotic segregation in
of unfertilized human oocytes. J Reprod Fertil 78:673-1678.
two translocation carriers. Am J Hum Genet 38197-208.
Martin RH, Rademaker AW, Hildebrand K, Long-Simpson L, Peterson D,
de GrouchyJ,Turleau C (1984):“ClinicalAtlas ofHuman Chromosomes,”
Yamamoto J (1987): Variation in the frequency and type of sperm
2nd ed. New York: John Wiley & Sons.
chromosomal abnormalities among normal men. Hum Genet
77:108-114.
Qalali M, Rosenbusch B, Wolf M, Sterzik K (1988): Cytogenetics of
unfertilized human oocytes. J Reprod Fertil 84:647-652.
Martin RH, Templado C, KOE, Rademaker A (1989):The effect of culture
Ejima Y, Sasaki MS, Kaneko A, Tanooka H (1988): Qpes, rates, origin
conditions and media on the frequency of chromosomal abnormalities
and expressivity of chromosome mutations involving 13q14 in retiin human sperm chromosome complements. Mol Reprod Dev 25101noblastoma patients. Hum Genet 79:118-123.
104.
Hansson A (1979):Satellite association in human metaphases.A compar- Martin RH, McGillivray B, Barclay L, Hildebrand K, KO E (1990a):
ative study of normal individuals, patients with Down syndrome and
Sperm chromosome analysis in a man heterozygous for a reciprocal
their parents. Hereditas 90:59-83.
translocation 46,XYt(12;20)(q24.3;qll).
Hum Reprod 5606-609.
Hassold T, Chen N, Funkhouser J , Jooss T, Manuel B, Matsuura J, Martin RH, Barclay L, Hildebrand K, KOE, Fowlow SB (1990b): CytoMatsuyama A, Wilson C, Yamane JA, Jacobs PA (1980):A cytogenetic
genetic analysis of 400 sperm from three translocation heterozygotes.
study of 1000 spontaneous abortions. Ann Hum Genet 44:151-164.
Hum Genet (in press).
Hassold T, Chiu D, Yamane JA (1984): Parental origin of autosomal May KM, Hassold T, Jacobs P, Lee M, Robinson A (1988): The triple X
trisomies. Ann Hum Genet 48:129-144.
syndrome parental origin of non-disjunction, Am J Hum Genet
43(Suppl):A114.
Hassold T, Benham F, Leppert M (1988): Cytogenetic and molecular
analysis of sex chromosome monosomy. Am J Hum Genet 42: Olson SB, Magenis RE (1988): Preferential paternal origin of de novo
534-541.
structural chromosome rearrangements. In Daniel A (ed):“The Cytogenetics of Mammalian Autosomal Rearrangements.” New York:
Henderson SA, Edwards RG (1968): Chiasma frequency and maternal
Alan R. Liss, pp 583-599.
age in mammals. Nature 218:22-28.
Henderson SA, Warburton D, Atwood KC (1973): Ribosomal DNA con- Papadopoulos G, Randall J , Templeton AA (19891: The frequency of
nectives between human acrocentric chromosomes. Nature
chromosome anomalies in human unfertilized oocytes and uncleaved
zygotes after insemination in vitro. Hum Reprod 4568-573.
245:95-97.
Jacobs P, Hassold T, Whittington E, Butler G, Collier S, Keston M, Lee M Pellestor F, Sele B (1988):Assessment of aneuploidy in the human female
(1988): Klinefelter’s syndrome: an analysis of the origin of the addiby using cytogenetics of IVF failures. Am J Hum Genet 42:274-283.
tional sex chromosome using molecular probes. Ann Hum Genet
Pellestor
F (1990):Analysis of meiotic segregation in a man heterozygous
52:93-109.
for a 13;15 Robertsonian translocation and a review of the literature.
Jenderny J , Rohrborn G (1987):Chromosomeanalysis ofhuman sperm. 1.
Hum Genet 8549-54.
First results with a modified method. Hum Genet 76:385-388.
Pellestor F, Sele B, Jalbert H (1987): Chromosome analysis of sperJuberg RC, Mowey PN (1983): Origin of non-disjunction in trisomy 21
matozoa from a male heterozygous for a 13;14Robertsonian translocasyndrome: all studies compiled, parental age analysis and internation. Hum Genet 76:116-120.
tional comparisons. Am J Med Genet 16:111-116.
Pellestor F, Sele B, Jalbert H, Jalbert P (1989):Direct segregation analKamiguchi Y, Mikamo K (1986):An improved,efficient method for analyysis of reciprocal translocations: a study of 283 sperm karyotypes from
sing human sperm chromosomes using zona-free hamster ova. Am J
four carriers. Am J Hum Genet 44:464-473.
Hum Genet 38:724.
Plachot M, de Grouchy J,Junca A, Mandelbaum J , Turleau C, Couillin P,
Kupke KG, Muller U (1989):parental origin of the extra chromosome in
Cohen J, Salat-Baroux J (1987): From oocyte to embryo: a model,
Trisomy 18. Am J Hum Genet 45599-605.
deduced from in vitro fertilization,for natural selection against chromosome abnormalities. Ann Genet 30:22-32.
Laurie DA, H u l t h MA (1985): Further studies on bivalent chiasma
ACKNOWLEDGMENTS
Renee Martin is an Alberta Heritage Foundation for
Medical Research Scientist, whose research is funded by
the Medical Research Council of Canada and the Alberta
Children’s Hospital Foundation. Sincere thanks go to
Lynne Bell for preparing the manuscript and to Leona
Barclay and Kathy Hildebrand for technical assistance.
Aneuploidy in Human Gametes
Plachot M, Veiga A, Montagut J, de Grouchy J, Calderon G, Lepretre S,
Junca AM, et al. (1988):Are clinical and biological IVF parameters
correlated with chromosomal disorders in early life: a multicentric
study. Hum Reprod 3:627-635.
Rosner B (1986):“Fundamentalsof biostatistics.” 2nd Ed. Duxbury Press,
Boston, p 222.
Rudak E, Jacobs PA, Yanagimachi (1978):Direct analysis of the chromosome constitution of human spermatozoa. Nature 274911-913.
Sanger R, Tippett P, Gavin J, Teesdale P, Daniels GL (1977):Xg groups
and sex chromosome abnormalities in people of northern European
ancestry: an adddendum. J Med Genet 14:210-213.
Sele B, Pellestor F, Jalbert P, Estrade C (1985):Analyse cytogenetique
des pronucleus a partir du modele de fecondation interspecifique
homme hamster. Ann Genet 28:81.
Tarkowski AK (1966):An air-drying method for chromosomepreparation
from mouse eggs. Cytogenetics 5394-400.
Templado C, Navarro J, Benet J , Genesca A, Mar Penez M, Egozcue J
(1988):Human sperm chromosome studies in a reciprocal translocation t(2;5). Hum Genet 79:24-28.
331
Vagner-Capodano AM, Hartung M, Stahl A (1987):Nucleolus, nucleolar
chromosomes and nucleolus-associated chromatin from early
dyplotene to dictyotene in the human oocyte. Hum Genet
75:140-146.
van Blerkom J, Henry GI (1988):Cytogenetics analysis of living human
oocytes: cellular basis and developmental consequences of perturbations in chromosomal organization and complement. Hum Reprod
3:777-790.
Warren AC, Chakravarti A, Wong C, Slaugenkaupt SA, Halloran SL,
Watkins PC, Metaxotou C, Antonarakis SE (1987): Evidence for reduced recombination on the nondisjoined chromosome 21 in Down
syndrome. Science 237:652-654.
Watt JL, Templeton AA, Messinis I, Bell L, Cunningham P, Duncan Ro
(1987): Wsomy 1 in an eight cell human preembryo. J Med Genet
24:60-64.
Wramsby H, Liedholm P (1984):A gradual fKation method for chromosomal preparations of human oocytes. Fertil Steril 41:736-738.
Wramsby H, Fredga K, Liedholm P (1987): Chromosome analysis of
human oocytes recovered from preovulatory follicles in stimulated
cycles. N Engl J Med 316:121-124.