Human Reproduction vol.14 no.10 pp.2506–2512, 1999
DNA flow cytometry of human semen
U.B.Hacker-Klom1,5, W.Göhde2, E.Nieschlag3 and
H.M.Behre3,4
1Clinic
and Policlinic of Radiotherapy–Radiooncology, 2Institute of
Radiobiology, 3Institute of Reproductive Medicine and 4Department
of Obstetrics and Gynaecology, University of Münster, Germany
5To
whom correspondence should be addressed at: Klinik und
Poliklinik für Strahlentherapie-Radioonkologie, Albert-SchweitzerStr. 33, D-48149 Münster, Germany
The aim of this study was the evaluation of DNA flow
cytometry for the analysis of male infertility. 171 ejaculates
from 155 patients with fertility problems were analysed by
flow cytometry and by conventional microscopical procedures. Using flow cytometry, it was possible to determine
the relative proportions of the various cell populations:
mature haploid and abnormal diploid mature spermatozoa,
cellular fragments, immature germ cells (haploid round
spermatids, diploid cells, S phase and 4C cells), and of
leukocytes as indicators of infection. A linear association
was observed between sperm concentration in semen as
quantified by light microscopy and by flow cytometry, even
with fewer than 20H106 spermatozoa/ml. Eight classes of
histograms, each with differing fractions of spermatozoa
and other particles, were obtained and correlated with the
results of the spermiograms. During the 10 year follow-up,
the two patient groups with a low sperm concentration or a
high concentration of cellular debris exhibited significantly
impaired fertility. The two patient groups with ù5% diploid
spermatozoa and with malcondensed sperm chromatin
were also subfertile. No ovulatory disorders were revealed
in the 155 female partners. DNA flow cytometry thus
provides an additional dimension to semen analysis not
easily gained by other methods and has the advantage of
being rapidly performed and interpreted. We therefore
recommend application of this technique in the diagnosis
of male infertility.
Key words: DNA flow cytometry/light microscopy/male infertility/semen
Introduction
Public concern has been generated recently following claims
of a reduction of human semen quality during the past 50
years: in a meta-analysis of 61 studies carried out at various
centres worldwide (Carlsen et al., 1992), a significant decrease
in mean sperm concentration occurred between 1940 and
1990 from 11.33106/ml to 6.63106/ml (P , 0.0001). Other
subsequent studies (Auger et al., 1995; Irvine et al., 1996)
2506
reported similar results. It has been proposed that lifestyle,
environmental toxins or prenatal exposure to oestrogens might
contribute to this phenomenon (Sharpe and Skakkebæk, 1993).
However, the lowering of reference values for ‘normal’ sperm
concentration from 603106/ml in the 1940s (MacLeod and
Heim, 1945) to the present value of 203106/ml (WHO, 1987)
alone could be responsible for the observation (Bromwich et al.,
1994). The standard determination of sperm concentration is
by use of a haemocytometer. It has been shown, however, that
there is a high variability in results obtained by this method
(Neuwinger et al., 1990). Therefore the application of other
methods might be useful.
Since the possibility of a long-term decline in human sperm
concentration is highly controversial (Lerchl and Nieschlag,
1996), prospective studies on large population groups will be
necessary to test the hypothesis. These could be facilitated by
application of an automated method for sperm counting such
as flow cytometry (FCM). Since this allows the analysis of
thousands of spermatozoa, statistical errors are minimized.
Among the numerous studies on sperm populations in
mammals (e.g. Harrison, 1997), only some of those reporting
FCM analysis of semen samples relevant to the present study
need be mentioned here. Measurement problems resulting from
the asymmetric shape and the highly condensed chromatin
(CC) of the spermatozoa were finally overcome: Zante introduced papain for the decondensation of sperm chromatin and
used a flow cytometer less susceptible to optical-geometric
artefacts (Zante et al., 1977), while Dean improved the
measuring accuracy by hydrodynamic orientation of spermatozoa (Dean et al., 1978). Further studies followed (Otto et al.,
1979a; Johnson et al., 1993; Ashwood-Smith, 1994; Fugger
et al., 1995).
The reduced fluorescence intensity of stained sperm DNA
when using epi-illumination systems allows reliable differentiation between haploid spermatozoa and haploid round
spermatids and between diploid spermatozoa and other diploid
cells (early germ cells or somatic cells such as leukocytes or
epithelial cells) (Van Dilla et al., 1977; Otto et al., 1979b;
Hartmann et al., 1982). Cell sorting of testicular cells and
subsequent DNA FCM using experimental animals has been
used to confirm the identity of the cells in each DNA histogram
peak (Hacker-Klom et al., 1989). Since the histograms obtained
from both testicular and ejaculated cells are qualitatively
similar in various different mammals including the human, the
conclusions drawn from the sorting data may also be applicable
to man. Clinical application of FCM as a suitable method for
supplementing the information obtained from the spermiogram
of patients with infertility was proposed by Otto et al. (1979b).
In the present study, we demonstrate that FCM provides
© European Society of Human Reproduction and Embryology
Flow cytometry of semen
Figure 1. DNA histograms of normal human semen and
corresponding cumulative frequency distribution. The relative DNA
content is indicated on the x axis, which is divided into 256
channels, and the number of cells per channel on the y axis. The
frequencies of the five different cell populations are: (i) sub-haploid
region ,1CC (debris): 21%; (ii) 1CC peak (mature haploid
spermatozoa): 71%; (iii) 1C peak (round spermatids): 1.2%; (iv)
2CC peak (diploid spermatozoa): 4.7%; and (v) .2CC level
including cells at 2C (leukocytes; G1-spermatogonia, primary
spermatocytes etc.), in the DNA synthesis phase and at 4C: 2.3%.
Figure 3. Difference of sperm concentration (106/ml) (log)
measured by light microscopy (LM) and flow cytometry (FCM)
against sperm concentration (106/ml) measured by FCM as a
double log plot (cf. Bland and Altmann, 1986).
information about the condensation state of chromatin and the
ploidy status of spermatozoa additional to the data obtained
by conventional light microscopy. Sperm counts are obtained
by FCM. A sample of 155 patients with infertility was analysed
and classified into eight groups with differing sperm DNA
histograms. To the best of our knowledge this is the first 10
year follow-up study revealing subfertile patient groups by
DNA content of spermatozoa.
Materials and methods
Patients and sample collection
A total of 171 ejaculates of 155 consecutive patients attending the
Institute of Reproductive Medicine for suspected infertility was
analysed. The patients were Caucasian and 20–53 years old (median:
32 years). The samples were obtained by masturbation after 2–7 days
of abstinence. If two or three ejaculates per patient were available, they
were all analysed to assess the reproducibility of the results obtained.
All female partners were seen for infertility work-up in the
Department of Obstetrics and Gynaecology of the University. No
ovulatory disorders were revealed in the 155 women. In two women
bilateral occlusion was detected by laparoscopy. In-vitro fertilization
was therefore performed.
Figure 2. Sperm concentration (106/ml) of semen samples as
measured by flow cytometry (FCM) and conventionally by light
microscopy (LM): (a) all 63 concentrations; (b) 31 samples with
sperm concentration ,20 3106/ml.
FCM
Aliquots of the 171 ejaculates were fixed with 10% acetic acid for
24 h to 2 months at 5°C. The semen samples were prepared slightly
modified from the previous protocol (Otto et al., 1979a). Thus 0.2 ml
of fixed cells were diluted in 3 ml DAPI solution (5 µg 49,6diamidino-2-phenylindole 2·HCl/ml 0.1 mol/l Tris-HCl buffer,
pH 8.0, with 40 mmol/l MgCl2). In special cases, the fixed cells were
centrifuged for 10 min at 200 g, resuspended in 1 ml 0.1 N sodium
citrate-sodium hydroxide buffer, pH 6.4, containing 300 Anson units
papain/ml (Anson unit: 1 unit will produce a ∆a280 of 0.001 per
min at pH 2.0 at 37°C, measured as TCA-soluble products using
haemoglobin as substrate), 20 mM dithioerythritol (DTE), 1 mmol/l
ethylenediaminetetraacetic acid (EDTA) disodium salt and 1%
dimethylsulphoxide (DMSO) and incubated for 15 min at 20°C. The
2507
U.B.Hacker-Klom et al.
Table I. Summary of the data from 171 semen samples within the eight classes as determined by DNA flow
cytometry (FCM).
Class no.
No. of
samples
% 1CC
(FCM)
1
2
3
4
5
6
7
8
Total:
12
42
10
47
16
18
16
10
171
92
82
67
79
63
62
64
8.5
68
6
6
6
6
6
6
6
6
6
3.2
6.0
4.7
10
14
21
19
13
23
Sperm cell
concentration (3106/ml)
106
25
67
67
6.8
37
2.2
0.041
33
6
6
6
6
6
6
6
6
6
62
23
61
10
6.7
23
2.2
0.058
44
% normal
spermatozoa (LM)
58
46
38
50
33
30
40
10
41
6
6
6
6
6
6
6
6
6
12
14
16
15
13
13
11
12
18
% pregnancies
(actual no.)
25
19
30
4
0
17
0
0
13
(3)
(8)
(3)
(2)
(0)
(3)
(0)
(0)
(19)
LM 5 light microscopy.
Class 1 is characterized by ù90% 1CC, class 2 by ,90 and ù70% 1CC, class 3 by ,70% 1CC, class 4 by
ù5% 2CC, class 5 by .10% overcondensed spermatozoa, class 6 by .10% .1CC, class 7 by the lack of a
histogram of normal height at 1CC, and class 8 by .50% , 1CC (debris). The arithmetic means 6 SD of
% haploid spermatozoa at 1CC, sperm cell concentration (3106/ml), percent normal spermatozoa and
percentage of spontaneous paternities per group within the 10 year follow-up are indicated.
samples were centrifuged again, resuspended and stained with DAPI
as above.
A PAS II flow cytometer (Partec GmbH, Münster, Germany)
equipped with a mercury arc lamp (HBO 100; Osram, Munich,
Germany), a UG1 excitation filter, KG1 and BG38 heat filters, a
TK420 beam splitter, and a GG435 barrier filter (Schott Glas, Mainz,
Germany) in front of the photomultiplier was used. Each sample was
measured at least twice.
Analysis of frequency histograms
Approximately 23104 cells were measured for each frequency histogram. DNA frequency histograms were evaluated using cumulative
frequency distributions (Figure 1). We discriminated one category
more than Fosså’s group (Fosså et al., 1989), i.e. five categories: (i)
cells with sub-haploid DNA content ,1CC (debris that may be of
apoptotic origin, Darzynkiewicz et al., 1992). The term ‘CC’ means
‘condensed chromatin’ and is used to indicate haploid spermatozoa
which have a normal DNA content of IC but which is too condensed
to be stained accordingly (Zante et al., 1977); (ii) mature haploid
spermatozoa in the 1CC peak; (iii) haploid round spermatids in the
1C peak; (iv) diploid spermatozoa in the 2CC peak; and (v) cells
registered to the right of the 2CC level including 2C cells (leukocytes,
G1-spermatogonia, primary spermatocytes at preleptotene etc.), cells
in the DNA synthesis phase (S) and 4C cells (primary spermatocytes
etc.; cf. Zante et al., 1977, Fosså et al., 1989).
Diploid spermatozoa of mammals are represented by a peak of
twice the modal channel number of haploid spermatozoa as confirmed
by cell sorting (Hacker-Klom et al., 1989). Reproducibility was
examined by six consecutive measurements of one sample and yielded
a coefficient of variation (CV) of 0.067 for the percentage of 1CC
cells (mature haploid spermatozoa). The coefficient of variation
indicating measuring accuracy was determined by calculating the
ratio of the standard deviation to the mean. Sperm counts were
measured in a volume of 0.2 ml by means of a special electronic
device within the FCM (Partec GmbH).
Spermiograms
Ejaculates were analysed by light microscopy according to WHO
(1987).
Statistics
Correlation coefficients between different features of spermatozoa as
determined by FCM and light microscopy were determined for
2508
identical samples with the method of least square deviations, using a
programmable calculator assuming a linear regression. Since neither
the correlation coefficient nor regression analysis was considered
appropriate in the analysis of measurement method comparison, a
graphical technique was applied (Bland and Altman, 1986). The χ2
test with Yates’ correction for continuity for contingency tables and
the unpaired t-test were applied wherever appropriate. In general,
data are given as mean 6 SD.
Results
There was a strong linear association between sperm concentration of 63 semen samples as determined by LM and FCM
(Figures 2 and 3). All 171 semen samples can be classified
into eight classes as suggested by the FCM data (Table I).
The peak of 1CC mature haploid spermatozoa in the
DNA histograms appeared at 66% (range: 60–71%) of the
fluorescence intensity of haploid round spermatids at 1C
(Figure 4). The samples of the first three classes are normozoospermic or oligozoospermic, showing a normally distributed
1CC peak with decreasing proportion of 1CC, and were
considered normal or quasi-normal. Typical after increasing
the high voltage was the splitting of the peak at 1CC, indicating
discrimination of the two populations of X and Y chromosomebearing spermatozoa with the Y spermatozoa on the left as
demonstrated here by one sample from class 1 (Figure 4a).
The material to the left of the 1CC peak represents cellular
debris. In contrast to this histogram, the other histograms in
Figure 4 were measured at a lower input of only 314 V to the
photomultiplier. This decrease in high voltage led to a shift of
the main peak at 1CC to the left of the histogram, enabling
us to visualize the range up to 4C. Class 1 is considered
normal with ù90% 1CC: a representative histogram (Figure
4b) shows only one major peak representing normal haploid
spermatozoa at 1CC comprising 91% of the measured particles.
The rest is cellular debris to the left of the main peak (,1CC),
haploid round spermatids at 1C and diploid spermatozoa at
2CC. No signals were registered in the region to the right of
2CC (2C, S and 4C). This ejaculate is normozoospermic
(723106 spermatozoa/ml).
Flow cytometry of semen
Table II. Correlation coefficients (r) obtained between flow cytometry
(FCM) data and characteristics observed by light microscopy (LM) for
identical samples: % haploid spermatozoa (FCM) is most strongly correlated
with % normal spermatozoa (LM), % debris ,1CC with % defect sperm
tails, % diploid spermatozoa with % duplicate sperm heads, % 2C cells with
% leukocytes, and a skewing to the left of the 1CC peak with % amorphous
sperm heads
Figure 4. Ten DNA histograms of human semen: (a) the peak
representing haploid spermatozoa at 1CC is split into two peaks
representing Y and X chromosome bearing spermatozoa; (b–j) after
decreasing the high voltage to the photomultiplier, the spectrum up
to 4C is visible. (b) Class 1: DNA distribution of a
normozoospermic man representing ù90% mature haploid
spermatozoa at 1CC. (c) Class 2: nearly normal DNA histogram
with 71% mature haploid spermatozoa. (d) Class 3: more severe
disturbance of spermatogenesis representing only 67% at 1CC. (e)
Class 4: DNA histogram reflecting ù5% of diploid spermatozoa at
2CC. (f) Class 5: the chromatin condensation of 55% of the haploid
spermatozoa is disturbed as reflected by the skewing of the DNA
histogram to the left. (g) and (h) Class 6: the presence of .10% of
different immature spermatogenic cell types at 1C, 2C, S and 4C
levels besides mature haploid spermatozoa is reflected in these
histograms with the more extreme case in (h). (h) The 1C peak
(round spermatids) appears in the place of 1CC (haploid
spermatozoa) in this histogram because 1CC was shifted to the left
by lowering the high voltage applied to the photomultiplier in order
to make the whole spectrum up to 4C visible in this particular case.
(i) Class 7: the DNA histogram reflects a total count of only 3.5
3106 spermatozoa in the ejaculate. (j) Class 8: only cellular debris
is present in the ejaculate, as shown by the DNA histogram. This
patient is azoospermic.
Class 2 was characterized by a 1CC peak that contained
70–90% of the particles counted. In one representative sample
(Figure 4c), there were 71% mature haploid spermatozoa and
FCM
LM
r
% 1CC
% ,1CC
% diploid spermatozoa
% 2C
Skewing to the left at 1CC
%
%
%
%
%
0.67
0.74
0.96
0.97
0.76
normal spermatozoa
defective sperm tails
duplicates
leukocytes
amorphous spermatozoa
a great deal of cellular debris (26%). The other particles were
haploid round spermatids and diploid spermatozoa, and only
0.52% were .2CC. The ejaculate had a volume of 2.4 ml
with a pH of 7.8 and 343106 spermatozoa/ml. According to
the spermiogram, only 29% of the spermatozoa were normal.
The 42 samples that fell into this class had a mean of 46 6
14% normal spermatozoa (Table I).
Class 3 contained 10 histograms with ,70% mature haploid
spermatozoa at 1CC. The typical DNA histogram (Figure
4d) shows a peak representing haploid mature spermatozoa
comprising 67% of the particles counted, a high amount
of debris (31%), a few haploid round spermatids, diploid
spermatozoa and no cells with a DNA content higher than
2CC. The ejaculate is oligozoospermic (3.33106 spermatozoa/
ml) and teratozoospermic with only 17% normal spermatozoa.
A semen analysis of the same patient ~4 months earlier yielded
a similar spermiogram. FCM analysis assigned this earlier
sample to class 5.
Histograms of classes 4–8 characterized different types of
spermatogeneic perturbations. Class 4 has diploid mature
spermatozoa at 2CC at a high level of at least 5% and .70%
1CC. The typical histogram showed 6.3% diploid spermatozoa
(Figure 4e). This ejaculate was normozoospermic with 723106
spermatozoa/ml. In all, 47 samples fell into this class (Table
I). Within class 4, there were 8.6 6 3.5% diploid spermatozoa
compared with only 2.5 6 0.72% within group 1, and 2.1 6
0.54% double-headed spermatozoa were identified by LM in
group 4 in contrast with only 0.6 6 0.35 within group 1. The
difference is significant at the P , 0.05 level by t-test.
Class 5 was characterized by a reproducible skewing of the
1CC peak to the left (Figure 4f). Although the skewing could
not be changed by treating the samples with papain for 15–60
min (thus inducing sperm chromatin decondensation), no
explanation other than a disturbance of chromatin condensation
of the haploid spermatozoa that could make the spermatozoa
resistant to papain could be found. In this ejaculate (Figure
4f) 55% of the haploid spermatozoa were not properly condensed. The patient had a parvisaemia (pathologically small
ejaculate volume ,2 ml), with an ejaculate volume of only
1.2 ml, was normozoospermic with 403106/ml, and had 39%
normal spermatozoa. Sixteen histograms fell into this class
(Table I). Only 45% of the samples within group 5 were
normozoospermic.
Typical of class 6 is that .10% cells were .1CC, indicating
2509
U.B.Hacker-Klom et al.
Table III. DNA flow cytometry (FCM) data and sperm concentration as arithmetic means 6 standard
deviation: the 171 ejaculates are assigned to three groups (azoospermic, oligozoospermic and
normozoospermic samples)
Sperm density
(3106/ml) (n)
,%1CC: debris %1CC: haploid
spermatozoa
I: 0 (3)
77 6 19
II: .0,20 (117) 9.6 6 10.9
III: ù20 (51)
13 6 13
0
65 6 40
80 6 20
the presence of immature germ cells and/or inflammatory cells.
The representative histogram of class 6 (Figure 4g) shows
68% haploid spermatozoa, 8% debris and other cell types:
spermatogenic cells, including haploid round spermatids (9%),
diploid spermatozoa (5%), and cells with a regular 2C DNA
content (4%), e.g. leukocytes and macrophages indicating
infection. The spermiogram revealed that the ejaculate is
oligoteratozoospermic (7.63106 spermatozoa/ml with only
12% normal spermatozoa). In one extreme case, the histogram
resembles a typical DNA frequency distribution of mammalian
testicular tissue (including that of man) with four peaks (Figure
4h). In the ejaculate, there were only 0.43106 spermatozoa/
ml and 23% round cells as revealed by LM. Sixteen samples
fell into this category (Table I). Within this group, there were
only 29% normozoospermic samples and, on average, more
amorphous spermatozoa than normal forms. The frequency of
round cells was highest in this group: 9.4 6 1.03106/ml in
contrast with only 3.9 6 0.83106/ml in group 1.
Class 7 contained no or only a few spermatozoa/ml. No
complete histogram could be obtained even if the whole sample
was measured (Figure 4i). As revealed by light microscopy,
the sample came from a cryptozoospermic teratozoospermic
ejaculate (0.73106 spermatozoa/ml, total volume 5 ml containing 23106/ml round forms, and 37% normal spermatozoa
in the sediment). Sixteen samples fell into this category. The
spermiograms revealed a mean of only 2.23106 spermatozoa/
ml among samples within this group (Table I).
Class 8 was characterized by a skewing of the particle
distribution from channel 0 onwards, indicating a preponderance of cellular debris in the representative DNA histogram
(Figure 4j). This ejaculate, with a volume of 7.0 ml, pH 7.5,
was totally azoospermic. Ten samples fell into this category
(Table I). None of these samples was normozoospermic. The
spermiograms revealed that any spermatozoa present were
amorphous rather than normal, and some defect tails were
found.
Table I summarizes the results of the 171 ejaculates assigned
to classes 1–8: only 7% of the patients with fertility problems
are in the class with the highest sperm quality (class 1).
Overall, the patients presented with a mean of 68 6 23%
haploid spermatozoa, a mean sperm density of 33 6 443106/
ml and a mean of 41 6 18% normal spermatozoa. In all, 13%
of the patients naturally fathered a child during the 10 years
of observation. Only 4% of the men with ù5% diploid
spermatozoa (class 4), 17% of class 6 and none of the class 5
patients with malcondensed spermatozoa or class 7 and 8
patients with low sperm concentrations or high amounts of
2510
%1C: haploid
spermatids
%2CC: diploid
spermatozoa
% ù2C: 2C, S,
4C
0
4.8 6 6.9
1.2 6 1.6
0
3.8 6 3.0
4.3 6 3.8
4.0 6 3.7
5.9 6 7.1
0.18 6 0.38
cellular debris in the semen respectively fathered children.
The two cycles of in-vitro fertilization resulted in a pregnancy
in one woman whose husband belonged to class 3. The husband
of the woman not achieving a pregnancy by in-vitro fertilization
belonged to class 4. The χ2 test reveals a statistically significant
difference in paternity rates between the classes (P 5 0.02).
When classes 1–3 and 4–8 are compared, the difference is
highly significant (P , 0.001).
Assuming a linear correlation, the percentage of 1CC cells
in the DNA histogram was correlated best with the percentage
of normal spermatozoa in the spermiogram (r 5 0.67) (Table
II). Debris in the histograms were correlated most favourably
with spermatozoa with defective tails (r 5 0.74), diploid
spermatozoa with double-headed spermatozoa (r 5 0.96), 2C
cells with leukocytes (r 5 0.97), and skewing of peak I with
amorphous sperm heads (r 5 0.76).
The percentage of haploid spermatozoa as measured by
FCM within the group of three azoospermic patients was zero
(Table III). Within this group, there were only debris and
.2CC cells indicating inflammation. Within the group of 117
oligozoospermic patients, there were 65 6 40% haploid and
3.8 6 3.0% diploid spermatozoa (Table III). Within the 51
normozoospermic patients, 80 6 20% haploid and 4.3 6 3.8%
diploid spermatozoa were found.
Of 155 patients, 141 gave one ejaculate, 12 gave two and
two patients gave three samples. 16/24 (67%) of the samples
from the 12 patients providing two sperm samples for analysis
at an interval of several weeks were assigned twice to the
same classes; the remaining 8/24 (33%) were not. Of the six
ejaculates provided by two patients, two samples from each
patient were assigned to one class, and the other sample to
another class. Differences in classification occurred over time
for quantitative rather than qualitative reasons, for example, a
finding of 4.0% diploid sperm in the first sample and 5.1%
diploid sperm in the second.
Discussion
To overcome the influence of subjective factors in semen
analysis by conventional methods, intensive efforts are being
made to establish objective laboratory methods, e.g. by computer-aided sperm analysis (CASA). However, ‘CASA is highly
dependent on the system settings used’ (Knuth et al., 1987)
which is especially critical in cases of oligozoospermia (Chan
et al., 1989).
As has been shown in the present paper, FCM allows precise
determination of sperm concentration. Linear associations
Flow cytometry of semen
between the sperm concentration obtained by conventional
LM and by FCM for normozoospermic men and for oligozoospermic patients with counts down to 0.13106 spermatozoa/ml
respectively indicate that counting by FCM may be performed
reliably, rapidly and objectively.
The 155 patients enrolled in our study had a mean sperm
concentration of 32.6 6 43.73106/ml. The average sperm
concentration 6SD of 68 6 23% within the group of 155
patients as analysed by FCM was lower than that indicated by
Fosså et al. (1989) for patients with testicular cancer (91%)
and the corresponding control group (97%), reflecting
decreased semen quality in patients with impaired fertility.
However, sperm counts are not the only parameter determining fertility (cf. Bostofte et al., 1982). Since the evaluation of
motility and morphology by LM tends to be subjective (WHO,
1987), other parameters of fertility need to be found.
In the present paper, eight classes with different semen
quality were determined by FCM and correlated with the
results of the spermiograms. About one-third of the 155
patients were normozoospermic, with other factors having
apparently affected the fertility. Increasing fractions of up to
35% of cellular debris of possibly apoptotic origin did not
affect fertility in normo- and oligozoospermic men (classes 1–
3). One unexpected feature, however, was that only two of the
47 men with .5% diploid spermatozoa fathered children
within 10 years (class 4). Although these patients had high
mean sperm concentrations and high percentages of normal
haploid spermatozoa, the occurrence of a high frequency of
diploid spermatozoa probably reflects not only faulty meiotic
segregation involving the synaptonemal complex, but also
major disturbances of spermatogenesis (Weissenberg et al.,
1998). In certain patients, an abnormally high frequency of
diploid spermatozoa seems to be correlated with repeated
miscarriages (de Geyter et al., 1997), possibly caused by sperm
polyploidy. Polyploidy, mainly triploidy, is found in 20% of
human spontaneous abortions (Tolksdorf, 1979). In cases
with repeated abortions, sorting of diploid spermatozoa may
therefore be useful.
It was also striking that there were no paternities among the
18 class 5 patients with malcondensed sperm during the
10 year follow-up: malcondensation of sperm chromatin is
considered to be correlated with infertility not necessarily
reflected by poor sperm morphology (Evenson and Melamed,
1983; Evenson et al., 1991; Hofmann and Hilscher, 1991;
Golan et al., 1997; Spanò et al, 1998). The detection of
disturbances in murine sperm chromatin condensation after
very low-dose X-rays (Sailer et al., 1995) suggests that
screening for genetic damage could be useful in the diagnosis
and characterization of infertility. Another reason for the
skewing of the sperm histogram, besides malcondensation of
sperm chromatin, might be apoptosis (as suggested by Dr
Z.Darzynkiewicz, personal communication).
Within class 6, the presence of immature germ cells or
inflammatory cells did not affect fertility within the 10 year
observation period. It was no surprise that the 16 patients with
very low sperm concentration of zero or close to zero (class
7) and the 10 men with cellular fragments prevailing in the
ejaculates (class 8) were infertile.
As suggested by these data, DNA FCM might be helpful in
future studies of human infertility, allowing the clinician to
assess rapidly and objectively the sperm count, the state of
chromatin condensation and the presence of immature germ
cells, of inflammatory cells and of diploid spermatozoa or of
aneuploid tumour cells in the ejaculate (cf. Otto et al., 1979b;
Clausen and Åabyholm, 1980; Evenson et al., 1980; Åabyholm
and Clausen, 1981; Steen and Hanson, 1981; Evenson and
Melamed, 1983; Fosså et al., 1989; Evenson et al., 1991;
Seligmann et al., 1994; Golan et al., 1997; Weissenberg et al.,
1998). Changes induced intentionally (e.g. male contraception)
or accidentally during treatments (e.g. therapies employing
radiation, hormones or cytostatic agents and affecting spermatogenesis) can be readily monitored and documented. Environmental or occupational toxin exposure might be suggested by
identifying suspected decreases in the quality of semen in
epidemiological analyses (cf. Spanò et al., 1983, 1998).
Acknowledgements
The authors cordially thank Ms G.Bellmann, Institute of Radiation
Biology, University of Münster, for excellent technical help, Dr
D.Neugebauer, Working Group Cellular Excitation Physiology,
University of Bochum, for helpful suggestions, and Dr D.S.Joshi,
Molecular Biology and Agriculture Division, Bhabda Research
Center, Bombay, India, and Dr M.H.Brinkworth, Institute of Reproductive Medicine, University of Münster, for language editing.
References
Åabyholm, T. and Clausen, O.P.F. (1981) Clinical evaluation of DNA flow
cytometry of fine needle aspirates from the testes of infertile men. Int. J.
Androl., 4, 505–514.
Ashwood-Smith, M.J. (1994) Human sperm sex selection. Safety of human
sperm selection by flow cytometry. Hum. Reprod., 9, 757.
Auger, J., Kunstmann, J.M., Czyglik, F. et al. (1995) Decline in semen quality
among fertile men in Paris during the past 20 years. N. Engl. J. Med., 332,
281–285.
Behre, H.M., Yeung, C.H. and Nieschlag, E. (1997) Diagnosis of male
infertility and hypogonadism. In Nieschlag, E. and Behre, H.M. (eds),
Andrology: Male Reproductive Health and Dysfunction. Springer, Berlin,
p. 104.
Bland, J.M. and Altman, D.G. (1986) Statistical methods for assessing
agreement between two methods of clinical measurement. Lancet, i,
307–310.
Bostofte, E., Serup, J. and Rebbe, H. (1982) Relation between sperm count
and semen volume, and pregnancies obtained during a twenty-year followup period. Int. J. Androl., 5, 267–275.
Bromwich, P., Cohen, J., Stewart, I. et al. (1994) Decline in sperm counts:
an artefact of changed reference range of ‘normal’? Br. Med. J., 309, 19–22.
Carlsen, E., Gilwercman, A., Keiding, N. et al. (1992) Evidence for decreasing
quality of semen during past 50 years. Br. Med. J., 305, 609–613.
Chan, S.Y.W. Wang, C., Song, B.L. et al. (1989) Computer-assisted image
analysis of semen concentration in human semen before and after swim-up
separation: comparison with assessment by haemocytometer. Int. J. Androl.,
12, 339–345.
Clausen, O.P.F. and Åabyholm, T. (1980) Deoxyribonucleic acid flow
cytometry of germ cells in the investigation of male infertility. Fertil.
Steril., 34, 369–374.
Darzynkiewicz, Z., Bruno, S., Del Bino, G. et al. (1992) Features of apoptotic
cells measured by flow cytometry. Cytometry, 13, 795–808.
Dean, P.N., Pinkel, D. and Mendelsohn, M.L. (1978) Hydrodynamic orientation
of sperm heads for flow cytometry. Biophys J., 23, 7–13.
de Geyter, Ch., De Geyter, M., Castel, A.M. et al. (1997) Assisted fertility.
In Nieschlag, E. and Behre, H. M. (eds), Andrology – Male Reproductive
Health and Dysfunction. Springer, Berlin, pp. 323–346.
2511
U.B.Hacker-Klom et al.
Evenson, D.P. and Melamed, M.R. (1983) Rapid analysis of normal and
abnormal cell types in human semen and testis biopsies by flow cytometry.
J. Histochem. Cytochem., 31, 248–253.
Evenson, D.P., Darzynkiewicz, Z. and Melamed, M.R. (1980) Relation
of mammalian sperm chromatin heterogeneity to fertility. Science, 210,
1131–1133.
Evenson, D.P., Jost, L.K., Baer et al. (1991) Individuality of DNA denaturation
patterns in human sperm as measured by the sperm chromatin structure
assay. Reprod. Toxicol., 5, 115–125.
Fosså, S.D., Melvik, J.E., Juul, N.O. et al. (1989) DNA flow cytometry in
sperm cells from unilaterally orchiectomized patients with testicular cancer
before further treatment. Cytometry, 10, 345–350.
Golan, R., Shochat, L., Weissenberg, R. et al. (1997) Evaluation of chromatin
condensation in human spermatozoa: a flow cytometric assay using acridine
orange staining. Mol. Hum. Reprod., 3, 47–54.
Hacker-Klom, U., Heiden, T., Otto, F.J. et al. (1989) Radiation-induced diploid
spermatids in mice. Int. J. Radiat. Biol., 55, 797–806.
Harrison, R.A. (1997) Sperm plasma membrane characteristics and boar semen
fertility. J. Reprod. Fertil., 52, 195–211.
Hartmann, W., Hettwer, H., Hofmann, N. et al. (1982) Die
Pepsinvorbehandlung menschlichen Spermas als Standardmethode in der
impulscytophotometrischen Analyse. Andrologia, 14, 135–142.
Hofmann, N. and Hilscher, B. (1991) Use of aniline blue to assess chromatin
condensation in morphologically normal spermatozoa in normal and infertile
men. Hum. Reprod., 6, 979–982.
Irvine, S., Cawood, E., Richardson, D. et al. (1996) Evidence of deteriorating
semen quality in the United Kingdom: birth cohort study in 577 men in
Scotland over 11 years. Br. Med. J., 312, 467–471.
Johnson, L.A., Welch, G.R., Keyvanfar, K. et al. (1993) Gender preselection
in humans? Flow cytometric separation for the prevention of X-linked
diseases. Hum. Reprod., 8, 1333–1339.
Knuth, U.A., Yeung, C.H. and Nieschlag, E. (1997) Computerized semen
analysis: objective measurement of semen characteristics is biased by
subjective parameter setting. Fertil. Steril., 48, 118–124.
Lerchl, A. and Nieschlag, E. (1996) Decreasing sperm counts? A critical
review. Exp. Clin. Endocrinol. Diab., 104, 301–307.
Levinson, G., Keyvanfar, K., Wu, J.C. et al. (1995) DNA-based X-enriched
sperm separation as an adjunct to preimplantation genetic testing for the
prevention of X-linked disease. Hum. Reprod., 54, 474–482.
MacLeod, J. and Heim, L.M. (1945) Characteristics and variations in semen
specimens in 100 normal men. J. Urol., 54, 474–482.
Neuwinger, J., Behre, H.M. and Nieschlag, E. (1990) External quality control
in the andrology laboratory: an experimental multicenter trial. Fertil. Steril.,
54, 308–314.
Otto, F.J., Hacker, U., Zante, J. et al. (1979a) Flow cytometry of human
spermatozoa. Histochemistry, 61, 249–254.
Otto, F.J., Hofmann, N., Hettwer, H. et al. (1979b) Die Analyse von
Spermaproben mit Hilfe impulszytophotometrischer DNS-Bestimmungen.
Andrologia, 11, 279–286.
Sailer, B. L., Jost, L. K., Erickson, M. et al. (1995) Effects of X-irradiation
on mouse testicular cells and sperm chromatin structure. Environ. Mol.
Mutagen, 25, 23–30.
Seligman, J., Kopsower, N.S., Weissenberg, R. et al. (1994) Thiol-disulfide
status of human sperm proteins. J. Reprod. Fertil., 101, 435–443.
Sharpe, R.M. and Skakkebæk, N.E. (1993) Are oestrogens involved in falling
sperm counts and disorders of the male reproductive tract? Lancet, 341,
1392–1395.
Spanò, M., Calugi, A., Capuano, V. et al. (1983) Flow cytometry and sizing
for routine andrological analysis. Andrologia, 16, 367–375.
Spanò, M., Kolstad, A.H., Larsen, S.B. et al. (1998) The applicability of the
flow cytometric sperm chromatin structure assay in epidemiological studies.
Hum. Reprod., 9, 2495–2505.
Steen, H.B. and Hanson, U. (1981) Sperm maturation determination by flow
cytometry. Cytometry, 2, 129.
Tolksdorf, M. (1979) Chromosomale Ursachen des Spontanabortes. In Schirren,
C., Mettler, L. and Semm, K. (eds), Fortschritte der Fertilitätsforschung
FDF 8. Grosse-Verlag, Berlin, pp. 41–46.
Van Dilla, M.A., Gledhill, B.L., Lake, S. et al. (1977) Measurement of
mammalian sperm deoxyribonucleic acid by flow cytometry. Problems and
approaches. J. Histochem. Cytochem., 25, 763–773.
Weissenberg, R., Aviram, A., Lewin, L.M. et al. (1998) Concurrent use of
flow cytometry and fluorescence in-situ hybridization techniques for
detecting faulty meiosis in a human sperm sample. Mol. Hum. Reprod., 4,
61–66.
2512
WHO (1987) Laboratory Manual for the Examination of Human Semen and
Semen–Cervical Mucus Interaction, 2nd edn. Cambridge University Press,
Cambridge.
Zante, J., Schumann, J., Göhde, W. et al. (1977) DNA-fluorometry of
mammalian sperm. Histochemistry, 54, 1–7.
Received on January 13, 1999; accepted on June 9, 1999