BIOLOGY OF REPRODUCTION 51, 414-424 (1994)
Effects of Temperature and Restoration of Osmotic Equilibrium during Thawing on the
Induction of Plasma Membrane Damage in Cryopreserved Ram Spermatozoa'
W.V. HOLT 2 and R.D. NORTH
Institute of Zoology, Zoological Society of London, Regent's Park, London NW1 4RY, United Kingdom
ABSTRACT
The objective of this investigation was to examine the nature of freeze/thaw-induced plasma membrane damage in an effort
to validate hypotheses about cryoinjury in ram spermatozoa. Spermatozoa were loaded with fluorescein diacetate (FDA), a marker
for plasma membrane integrity, and cooled (15°C/min) to temperatures between -10°C and -30°C on a cryomicroscope stage.
Post-thaw fluorescence intensity measurements of individual cells indicated that freezing to temperatures between -10°C and
-15°C did not induce significant membrane permeabilization. However, freezing below -15°C was followed by membrane permeabilization immediately after thawing.
0
A majority (> 60% ) of flagellar plasma membranes of cells frozen to - 10 C remained ultrastructurally intact during thawing;
principal-piece membranes were more robust than middle piece membranes (p = 0.001). Significant middle-piece membrane
0
breakage was, however, induced as the post-thaw temperature increased from +10°C to +30°C (10 C, 64 ± 12.3% intact mem0
0
branes [mean ± SEMI; 30 C, 43 ± 12.5% intact membranes [mean + SEM]; p = 0.0085). Cells frozen to -30 C did not exhibit
this thawing effect, although the distinction between middle-piece and principal-piece plasma membranes was evident (p =
0.002). All sperm head plasma membranes were damaged by freezing and thawing to any combination of temperatures. Although
acrosomes became swollen after freezing and thawing, the incidence of outer acrosomal membrane vesiculation remained at
control (unfrozen) levels with all treatments used.
Experimental exposure to the hyperosmotic conditions generated during freezing induced little flagellar membrane permeabilization, but significant damage was caused by restoration of osmotic equilibrium. It is suggested that membranes are initially
destabilized during the freezing process, both by low temperature effects and by exposure to high salt concentrations. The
resultant post-thaw degeneration of the plasma membrane is caused by a combination of temperature and osmotic effects.
INTRODUCTION
reasons for attempting to impressing
several
There are
procedures.
cryopreservation
of
semen
prove the success
immunosuppresof
human
for
transmission
The potential
sive virus (HIV) through contaminated donor semen has
resulted in the requirement that frozen semen from tested
donors be used in donor insemination programs [1]. In
conservation biology, semen cryopreservation procedures
are required for a wide range of animal spermatozoa in
connection with the objective of preserving genetic material as a resource for supporting the survival of endangered
species [2]. In agriculture, the use of frozen boar and ram
semen is insufficiently reliable and effective for routine
purposes. Although it has been routine practice to breed
dairy cattle by artificial insemination with frozen semen,
nonetheless a major proportion of bull spermatozoa are
damaged and rendered infertile by the cryopreservation
technique [3].
The development of semen cryopreservation procedures was recently reviewed by Hammerstedt et al. [4] and
Watson [5], who expressed the view that further progress
in improving sperm survival would not be achieved simply
by modifying established cryopreservation diluents. A more
Accepted April 22, 1994.
Received January 6, 1994.
'Supported by a grant from the Medical Research Council and the Agriculture
and Food Research Council. The silicon-intensified target camera used in these experiments was provided by The Royal Society, London.
2Correspondence. FAX: 71-586-2870.
fundamental understanding of the biophysical and biochemical processes that accompany sperm freezing and
thawing is therefore an essential prerequisite to the design
of successful cryopreservation protocols.
Hypotheses for cell cryoinjury and cryoprotectant activity
have been proposed (for reviews, see [6,7]) and can be
summarized as follows. When cells are frozen, the crystallization of water induces the generation of hyperosmotic,
unfrozen pockets of high solute concentration. At moderate
freezing rates, hyperosmotic regions form within the extracellular environment; these induce removal of intracellular water, consequent cell shrinkage, and exposure of the
cell surface to anisosmotic (> 600 mOs/Kg) conditions.
Rapid. freezing may cause lethal intracellular ice nucleation
if sufficient freezable water remains in the cytoplasm when
the nucleation temperature is reached. Thawing involves a
reversal of these effects, with a decrease in the solute concentration of the external milieu and restoration of the intracellular water content and cell volume.
Spermatozoa deviate markedly from the ideally spherical
and homogeneous structures treated by the theoretical
model. However, the relevance of the osmotic interaction
hypothesis has been supported by experiments in which
spermatozoa were exposed to hypertonic sodium chloride
solutions. The extent of sperm damage induced by unfrozen hypertonic solutions correlated well with the damaging effects observed at subzero temperatures, where the
effective increases in salt concentration could be calculated
from colligative solution theory [8, 9].
414
SPERM MEMBRANE FREEZING
In a recent investigation aimed at testing the applicability
to spermatozoa of the hypotheses summarized above [10],
direct observation of ram spermatozoa by cryomicroscopy
produced a number of unexpected and novel findings. The
purpose of the present series of experiments was to formulate and investigate a number of alternative, but related,
hypotheses arising from these observations.
The previous study indicated that plasma membrane integrity was maintained throughout the cooling and freezing
process even in the absence of cryoprotectants. Membrane
damage was not manifested immediately upon thawing, but
occurred during post-thaw re-warming within specific temperature ranges between 2°C and 30°C. The temperature of
post-thaw damage induction was lowered by increasingly
severe treatments such as exposure to detergent or lower
minimum temperatures during the freezing process.
Membrane integrity was inferred in two different ways.
The first was through observation of post-thaw retention of
intracellular fluorescein after pre-freeze uptake of fluorescein diacetate (FDA). Nonfluorescent FDA is able to permeate into normal cells with intact plasma membranes. Intracellular esterases convert the FDA into the fluorescent
compound, fluorescein. Fluorescein is unable to pass out
through an intact plasma membrane, but can flow out of a
cell with a sufficiently damaged membrane. Consequently,
after exposure to FDA, cells with an intact plasma membrane will fluoresce and remain fluorescent with time, except for photobleaching of the fluorophore, or will lose
fluorescence as the fluorescein diffuses out. Secondly,
membrane integrity was inferred through observation that
exclusion of extracellular ATP by intact plasma membranes
prevented reactivation of cells immobilized by cold-shock
and freezing. Reactivation occurred, however, if the cells
were re-warmed to approximately 120C. Since detergentdemembranated, and therefore immotile, ram spermatozoa
can be reactivated by exogenous ATP [11], it was argued
that reactivation could be induced after other forms of
membrane damage but would occur only when the plasma
membrane became sufficiently damaged to permit inward
diffusion of ATP.
These observations are consistent with previous ultrastructural findings on routinely frozen human spermatozoa
that showed thawing to have caused more morphological
damage than freezing [12]. In the previous experiments, the
minimum post-thaw temperature permitting motility reactivation by exogenous ATP (9-14°C) was approximately 7°C
higher than the corresponding temperature at which fluorescein efflux occurred, perhaps indicating that the two
techniques detected different facets of membrane damage.
The simplest hypothesis to explain these data is that
membrane integrity is retained during exposure to the high
solute concentrations produced by freezing, but lost when
solute dilution restores isotonicity during thawing. This
possibility was investigated in the previous study, where it
was shown that significant fluorescein loss was delayed for
415
180 sec when the frozen spermatozoa were thawed and kept
at 0°C. Moreover, fluorescein loss was induced only when
completion of the thawing process was followed by an increase in temperature from 0 to 100C. Thus dilution effects
are evidently not the sole cause of membrane damage and
the simple hypothesis is not supported.
To investigate this effect further, the extent of membrane
damage caused solely by the restoration of isosmotic conditions was examined in the present study. An experiment
in which spermatozoa were exposed to a hypertonic environment for a short time at 5C, and then re-warmed to
room temperature, was undertaken. The spermatozoa were
therefore exposed to many of the effects of freezing and
thawing, but not the low temperature itself. The hypothesis
would predict that spermatozoa remain intact while exposed to hypertonic media, but that considerable membrane damage is induced upon redilution to isotonic conditions. Recent experiments of this type with human
spermatozoa [13] demonstrated that these cells do, indeed,
remain substantially intact during severe hypertonic stress
(> 1000 mOs), suffering plasma membrane damage upon
subsequent restoration of isotonicity.
An alternative interpretation of the previous findings on
ATP-induced reactivation is that thawing does not lead to
immediate rupture and lysis of the sperm plasma membrane during restoration of the osmotic equilibrium. A hypothetical, transitionally damaged state may exist after
thawing whereby entry of exogenous ATP is temporarily
blocked at low post-thaw temperatures (i.e., 1-10 0 C) even
though fluorescein efflux occurs readily. Such a state could
occur only if the plasma membranes remained structurally
intact, retaining some of their semipermeable properties.
This hypothesis was examined in the present study by
investigation of the following testable predictions. 1) Structurally damaged plasma membranes would be infrequently
observed at low post-thaw temperatures (i.e., 0-5C) and
2) structural plasma membrane damage would be significantly increased by elevating the post-thaw temperature
above 10°C. Transmission electron microscopy (TEM) and
scanning electron microscopy (SEM) were used to examine
the ultrastructural integrity of sperm plasma membranes after freezing and thawing. Major breaches of the plasma
membrane, compatible with unimpeded solute diffusion,
would be detectable by this method.
Our previous experiments [10] also demonstrated a direct relationship between the ATP-induced reactivation
temperature of frozen/thawed cells and the minimum temperature imposed during freezing. To test the hypothesis
that freezing to lower temperatures influenced the post-thaw
membrane destabilization temperature, further experiments were aimed specifically at comparing the post-thaw
leakiness characteristics of spermatozoa frozen to different
minimum temperatures. It would be predicted from the
previous data that when spermatozoa are frozen to -10°C,
they are more fluorescein-retentive after thawing than are
416
HOLT AND NORTH
spermatozoa frozen to below -20 0C. A semi-quantitative
image analysis method of assessing fluorescein loss from
individual cells was developed for this purpose, validated,
and used to examine the kinetics of fluorescence decline
in frozen/thawed cells.
Cryoprotectants were not used within the present experiments, which were aimed at examining the effects initially observed when cryoprotectants are omitted from the
cell suspensions. However, the previous studies [10] clearly
demonstrated that the presence of 2% glycerol shifted the
post-thaw fluorescein loss into a higher temperature range
but did not prevent it. It is therefore argued that clues to
the detrimental actions of freezing and thawing can reasonably be obtained when cryoprotectant is omitted.
Different responses between major plasma membrane
domains would not be unexpected. Since the observations
on sperm reactivation could apply only to the flagellum,
separate examination of sperm head, middle-piece, and
principal-piece plasma membranes was undertaken in the
ultrastructural study. The fluorescence image assessment
technique was directed toward the middle piece. This was
largely imposed by practical considerations, since fluorescence measurements from sperm heads were orientationsensitive and subject to error whereas the middle-piece was
the most reliably measured region of the flagellum.
MATERIALS AND METHODS
Semen Collection and Handling
Ram spermatozoa were collected with the aid of an artificial vagina and washed once in Krebs'-Henseleit-Ringer
medium (KHR; 1:10 [v/v] dilution). This involved centrifugation for 5 min at 900 x g and resuspension in KHR at
a concentration of either 1000 x 106 spermatozoa/ml for
cryomicroscopy and SEM or 200 x 106 spermatozoa/ml for
use in the hypertonic exposure experiment. Alternatively,
the spermatozoa were allowed to swim up into fresh KHR
for TEM experiments. Once prepared, samples were maintained at 22°C for 2-3 h and serially subsampled as appropriate.
Effects of Hypertonic Media and Restoration of Isotonicity
Exposure to hypertonic solutions. Aliquants (0.25 ml)
of sperm suspensions (200 x 106 /ml in isotonic KHR; 282
mOs/Kg) in conical plastic 10-ml tubes were cooled slowly
(0.25°C/min) from 30 0C to 5C. The osmolarity was then
elevated, as required by the experimental design, in a single step by adding 0.25 ml of precooled KHR whose NaCl
content had been increased such that the resultant solution
was appropriately hypertonic. The target hyperosmotic
strengths were 600 mOs/Kg and 1000 mOs/Kg; for controls, 0.25 ml of normal-strength KHR was added.
The samples were allowed to remain in hyperosmotic
media for 5 min, whereupon a 50-iil aliquant of each was
removed for evaluation of membrane integrity (see below).
A further 4.5 ml (i.e., 1:10 dilution) of modified KHR was
then added to each tube in a single step; in this case, the
NaCl content was appropriately below normal to restore
isotonicity. This step was performed at room temperature
to simulate the thawing and re-warming process. A further
aliquant of sperm suspension (100 dul) was then removed
for assessment of plasma membrane integrity.
Assessment of plasma membrane integrity. The spermatozoa were assessed through use of a modification of the
carboxyfluorescein diacetate:propidium iodide technique
described by Harrison and Vickers [14]. Sperm samples were
diluted to approximately 7.5 x 106 /ml in the following buffer
solution: 20 mM HEPES buffer (pH 7.5), 1 mg/ml polyvinylpyrrolidone, 1 mg/ml polyvinyl alcohol, 220 mM sucrose, 10 mM NaCl, 2.5 mM KCl, 10 mM glucose, 0.005 mg/
ml propidium iodide, 0.02 mg/ml carboxyfluorescein diacetate (CFDA), 0.06 pIl/ml formalin (40% stock solution).
CFDA was used here instead of the FDA used elsewhere in
the present experiments because it is more stably retained
by intact cells and preparations can be examined up to 1
h after dilution into the staining medium. Similarly, intact
cells do not permit entry of propidium iodide within that
period.
Aliquants (20 ,1 l) of this sperm suspension were placed
on glass slides and mounted beneath a coverslip. They were
then examined by fluorescence microscopy using a 40x
objective, and approximately 200 cells per slide were scored
as "intact" (green fluorescence) or "permeable" (red fluorescence).
Experimental design. Semen samples from four rams
(replicates) were used in an experiment to distinguish between the damaging effects of 1) exposing spermatozoa to
hypertonic media (KHR adjusted to 600 and 1000 mOs/Kg;
control samples exposed to normal-strength KR, 282 mOs/
Kg) at 5°C and 2) subsequent restoration of isotonicity. The
same samples were examined during and after hypertonic
exposure.
Cryomicroscopy and FluorescenceAssessment
A CM-3 cryomicroscope stage (Planer Biomed, Sunburyon-Thames, Surrey, UK) mounted on a Zeiss Axioskop microscope (Carl Zeiss, Oberkochen, Germany) equipped with
a 40x objective was used for these experiments. For observation of spermatozoa during freezing and thawing, 50
Il of the stored sperm suspension was diluted 1:10 with
KHR, and 5 l of stock FDA solution was added (Sigma
Chemical Co., Poole, Dorset, UK; 0.5 mg/ml in dimethyl
sulfoxide). No cryoprotectants were added. Spermatozoa
were placed on the cryomicroscope stage at 37°C without
any further incubation period and used in the freeze/thaw
experiments. Unless otherwise specified, samples were
cooled at 15°C/min and warmed at 10°C/min. Ice nucleation was induced automatically by the centrally directed
growth of spontaneously formed ice crystals from the cooler
SPERM MEMBRANE FREEZING
edges of the preparation. Observations were confined to a
region approximately 2 mm wide, adjacent to the thermocouple used for sensing and controlling the stage temperature. To minimize photobleaching effects, spermatozoa
were not exposed to ultraviolet light during cooling and
freezing, but only when the temperature reached -2°C prior
to thawing.
A silicon-intensified target video camera (Hamamatsu
C2400; Hamamatsu City, Japan) was used to produce images that were captured directly into a digital frame store
(DT2861; Data Translation, Marlboro, MA; 16 frames at 512
x 512 resolution). The manufacturer's specifications indicated that the camera output was a linear function of the
face plate illuminance; no electronic modification of this
relationship (image enhancement) was introduced. To reduce background noise, each stored image was derived by
averaging 8 sequential frames generated at 25 frames/sec
(i.e., representing 0.3 sec). Photobleaching and cell damage
caused by the illuminating ultraviolet light were reduced
by placing a neutral density filter in the incident light path
(25% transmission); such illumination permitted untreated
spermatozoa to remain motile for approximately 3 min.
Individual spermatozoa could be recognized in successive images of a series. Single pixel (pixel dimension approximately 1 x 1.5 ,um) grey-scale measurements from
black (0) to white (255) were made by the use of a manual
cursor and simple image analysis software (developed by
Dr. D.M. Holburn, Department of Engineering, University
of Cambridge Cambridge, UK). Fluorescence intensity measurements were made on sperm middle pieces. Background fluorescence was also measured in the immediate
vicinity of each cell; this was subtracted from the cellular
values to compensate for spatial variations in image brightness. Subsequent data analysis was performed upon the absolute differences. Image capture was controlled by routines written in the Data Translation programming language,
IrisTutor.
During the cryomicroscope experiments, the need to
observe individually identifiable cells for finite periods of
time, up to 1 min, meant that typically, 5-10 usable cells
were visible throughout the period of interest. Experiments
were replicated several times with the same ejaculate and
several ejaculates were used, although on different days.
For the purpose of clarity in data presentation, comparisons
within single ejaculates or representative data from single
ejaculates are shown in the present paper.
Validation of the image assessment technique. Preliminary testing and validation of the fluorescence assessment
system were carried out with glutaraldehyde-fixed spermatozoa, labeled with fluorescein isothiocyanate (FITC) by
a standard protocol [15]. Multiple binding sites for FITC
would be responsible for fluorescence in such cells; therefore leakage of fluorescein would not occur and photobleaching effects under the standard illumination conditions used in the present experiments would be detectable.
417
Although the use of this technique involves the assumption
that the photobleaching kinetics of free (from FDA) and
bound (from FITC) fluorescein are similar, this is probably
justified.
A series of measurements were made of FITC-labeled,
glutaraldehyde-fixed spermatozoa. Grey-level intensity
measurements were made every 6 sec over the middle pieces
of spermatozoa held at O°C for a period of 1.5 min. Greylevel values varied between spermatozoa, but there was no
evidence of fluorescent fading within the measurement time
scale. To confirm that lowered temperatures did not reduce
fluorescence levels, further experiments were performed
in which the measurements were made on spermatozoa
held at -60 0C. No evidence of a systematic effect of temperature upon fluorescence intensity was obtained.
SEM
Experimental technique. Four microliters of the washed
sperm preparation was placed upon the cryomicroscope
stage assembly at an initial temperature of 30°C, directly
above the temperature-sensing thermocouple. Samples were
not mounted beneath a coverslip. The temperature was reduced to -10 0C at 15°C/min, and the samples were held
at -10°C for 5 min. The temperature was then raised to
various values above the melting point (10C, 5°C, and 10°C).
Ice nucleation occurred spontaneously but reproducibly,
probably induced by the presence of frost adjacent to the
sample. A control uncooled and unfrozen treatment was
also included with each replicate. Once the target temperatures were reached, 4-ul aliquants of temperature-equilibrated fixative (glutaraldehyde, 4% [v/v]; formaldehyde, 3%
[w/v]; sucrose 4% [w/v]; 100 mM phosphate buffer [pH 7.4];
GFPS) were placed upon the melted samples and left for
1 min.
Polylysine-coated nickel grids were drawn through the
fixative-sperm mixture, immediately placed into a large drop
of fixative on a wax surface, and then allowed to complete
fixation overnight. The grids were then prepared for SEM
by critical-point drying and gold coating and were examined with a JEOL 1200X electron microscope (Jeol Ltd., Tokyo, Japan) in SEM mode.
Evaluation of cell damage. Evaluation of the frozen
and thawed spermatozoa was performed without knowledge of the treatments concerned. The experiment was
replicated with spermatozoa from three ejaculates, and
treatments were randomized and coded. An unfrozen control treatment was included in the experiment. Approximately 50 spermatozoa were examined for each treatment;
every spermatozoon was classified as either "intact" or
"damaged." If damaged, spermatozoa were classified further as 1) sperm head vesiculated, 2) acrosome lost or
damaged, 3) head separated from tail, 4) tail abnormality.
TEM
Sperm preparation. To examine the incidence of sperm
plasma membrane damage after freezing and thawing,
418
HOLT AND NORTH
0
FIG. 1. Transverse sections through the middle pieces of spermatozoa frozen to -10°C and then thawed to 30 C. Evidence of plasma membrane
of principal piece from sperC)
Sections
000).
(x56
sections
both
in
undamaged
morphologically
are
breakage is seen in (A) but not in (B). Mitochondria
matozoa frozen to -30°C and then thawed to 30°C. In one case the plasma membrane is intact (arrow), whereas in the other it is damaged or missing
(x 57 300).
spermatozoa were frozen and thawed within 0.5 ml Cassou
straws of the type routinely used for semen cryopreservation and then processed for examination by TEM. Swim-up
populations of spermatozoa were prepared as described
above; the sperm concentration was adjusted to 100 x 106/
ml. Spermatozoa were loaded at 30°C into 0.5-ml plastic
straws, which were then sealed with polyvinyl alcohol sealant and frozen to either -10C or -30°C. For these bulk
samples, we were unable to reproduce the linear cooling
profiles of the cryomicroscope, and the freezing methods
used represent the best compromises possible.
0
Freezing technique. Straws to be frozen at -30 C were
placed upright in a controlled-rate freezer (Planer R204;
Planer Products Ltd, Sunbury-on-Thames, Surrey, UK) with
the chamber precooled to -30°C. Internal straw tempera0
ture declined rapidly (60°C/min) to approximately -2 C,
at which point freezing occurred. The temperature then increased instantly to +2°C, remained constant for about 50
sec, and then continued to decrease at an initial rate of 7°C/
min. After the straws had been in the chamber for 5 min,
the chamber was heated to the desired target temperature
(10C, 50C, 100C, or 30°C). The initial re-warming rate was
approximately 30°C/min; the thawing process retarded the
heating rate, slowing it to 3°C/min in the temperature range
0-5°C.
After 5-min equilibration at the new temperature, the
samples were transferred to a water bath at the target thawing temperature and allowed to thaw for a further 2 min.
The straw contents were emptied into 2 ml of GFPS, osmicated, and prepared for TEM. Treatments were performed sequentially but in randomized order to eliminate
possible systematic effects due to storage at room temperature. Control samples from each replicate were loaded into
straws, placed in the freezing machine for the appropriate
time, and then emptied into fixative.
Freezing straws to -10 0 C presented a practical problem.
When a controlled-rate cell freezer was used, the latent heat
of fusion released during freezing caused the sample temperature to increase by several degrees and remain static
for up to 2 min. The relatively high chamber temperature
(-10°C) provided insufficient temperature gradient to cool
the samples as required. Therefore an alternative freezing
technique was used that involved placing sealed straws into
precooled ethylene glycol at -10°C for 5 min to maximize
heat transfer.
The cooling profile was nonlinear, but interruption of
cooling by heat of fusion was minimized. The straw contents cooled at 23°C/min to -3 0C. Freezing occurred consistently between -3°C and -4 0C, although deliberate nucleation of ice was not performed. The sample temperature
then increased rapidly to approximately +2°C, where it remained constant for 20 sec before resumption of cooling
to -10 0C at an initial rate of 7°C/min. After freezing and
holding in ethylene glycol for 5 min, the straws were moved
0
to a water bath at the target thawing temperature (1 C, 5°C,
0
10 C, or 30°C), allowed to thaw for 2 min, and fixed as
described above.
Experimental design. To examine plasma membrane
breakage and acrosomal damage in detail, two separate randomized block experiments were set up, each using semen
samples from four rams (replicates). Spermatozoa were
0
loaded into straws as described above, cooled to -10 C in
0
one experiment and -30 C in the other, and then warmed
to 1°C, 5°C, 100C, or 30°C. Samples were then fixed using
temperature-equilibrated fixative. Unfrozen control treatments involved placing the straws in a water bath at 30°C
for 5 min.
Evaluation of sections. Spermatozoa were evaluated
by examining transverse sections of middle piece and principal piece (Fig. 1) for the integrity of the surrounding plasma
membrane. Sections were scored as damaged or intact as
defined in Figure 1. Plasma membranes over sagittally sectioned sperm heads were scored as intact or damaged, and
acrosomes were scored as 1) intact or swollen (Figs. 2, A
and C) or 2) vesiculated (Fig. 2B). Approximately 50 appropriately sectioned principal pieces, middle pieces, and
sperm heads were examined for each treatment. It is recognized that assessment of transverse sections probably
419
SPERM MEMBRANE FREEZING
FIG. 2. Sagittal sections of spermatozoa frozen to -10°C and then thawed to 1°C A). Only traces of plasma
membrane are evident, but one acrosome appears intact, without any indication of swelling (magnification x25 000).
B) Sections thawed to 5C. Vesiculation of the outer acrosomal membrane is apparent (x25 000). C) Thawed to 30°C.
While only fragments of plasma membrane remain, the outer acrosomal membrane is deformed but structurally intact
(x33 600).
underestimates the absolute extent of membrane damage.
However, within experiments this approach provided a
consistent method of evaluating the relative effects of different treatments.
StatisticalAnalysis
Statistical analyses were performed with CSS/Statistica
software (StatSoft UK, Letchworth, UK). Counts obtained from
the quantitative SEM and TEM experiments were expressed
as percentages and transformed to angles (arcsin) for oneor two-way analysis of variance. Paired treatment contrasts
within an analysis were made using Least Significant Difference (LSD) tests. Results from the fluorescent assessments were examined by linear regression analysis. Differences in fluorescent decay characteristics between groups
of control and frozen spermatozoa were examined by anal-
ysis of covariance to assess departure from parallelism and
differences between intercepts. The effects of hypertonic
media and restoration of isotonicity were examined using
analysis of variance for repeated measures.
RESULTS
Exposure to Hypertonic Solutions and Restoration of
Isotonicity
The results are illustrated in Figure 3. After slow cooling
to 5C, approximately 50% of control spermatozoa remained intact; no further cell damage was induced by subsequent dilution and re-warming. The incidence of cell
damage was slightly, but not significantly (p = 0.36), increased by the initial exposure to hyperosmotic solutions.
420
HOLT AND NORTH
Control
60
600 mOs
1000 mOs
100
80
40
9-
60
a)
u1
Ct
,
A
20
40
I
I
I
0
--
H I
-
-
H I
20
|
-
. **
.-
-
0
H I
FIG. 3. Proportions of intact spermatozoa after spermatozoa were cooled
to 5°C, exposed to hypertonic solutions (H) for 5 min, and then resuspended in isotonic medium (I). indicateses significant reduction of sperm
integrity (p < 0.01).
0
0
-r--)
5
I.D
a)
10
15
20
25
20
25
Time (sec)
100
*E
Sperm damage was significantly increased, however, by
restoration of isotonic conditions (p = 0.0007). Examination of individual contrasts confirmed that this effect was
significant for both 600 mOs/Kg (p = 0.003) and 1000 mOs/
Kg (p = 0.002).
80
Cryomicroscopy
Measurement of fluorescence decay after thawing. To
examine the kinetics of fluorescein leakage in frozen/thawed
cells, grey-level measurements were made over a 24-sec period immediately after thawing as the temperature in-
20
C
C
5
0
40
0
0
5
10
15
Time (sec)
100
CI)
60
FIG. 5. Intensity, scale of 0 (black) to 255 (white), of middle piece fluorescence in FDA-labeled ram spermatozoa frozen to -30°C, re-warmed at
10°C/min, and measured at 1.3-sec intervals after thawing. Mean values (+
SEM) are shown for three (group A) and four (group B) spermatozoa from
separate ejaculates. Intensity values were derived by subtraction of background measurements from cellular measurements and did not exceed 100.
80
60
C
._
CI
C)
40
creased from 0 to 4C (warming rate 10°C/min). Samples
of spermatozoa were frozen to -10 0C, -15°C, and -30°C
at 15°C/min. Image capture was initiated immediately upon
thawing. Control treatments involved maintaining spermatozoa upon the cryomicroscope stage for 24 sec at 20 0C.
Comparisons were made within four separate ejaculates; for
practical reasons the ejaculates were examined on different
M
20
0
Time (sec)
FIG. 4. Intensity, scale of 0 (black) to 255 (white), of middle piece fluorescence in FDA-labeled ram spermatozoa. Solid circles indicate control
values (mesan + SEM, n = 8) from live cells maintained at 20TC for 25 sec.
Solid squareres indicate values (mean + SEM, n = 5) for cells frozen to -10°C,
warmed at 10°C/min (0.16°C/sec), and then measured at 6-sec intervals after thawin g. Rates of fluorescence loss are not significantly different (p =
0.515). Intedensity values were derived by subtraction of background measurements from cellular measurements and did not exceed 100.
days.
After cooling to -10 0C and subsequent thawing, spermatozoa exhibited a linear decline in fluorescence (Fig. 4)
during the initial 25-sec post-thaw period. Analysis of co-
variance showed that the fluorescence loss rate was not significantly higher than that for control (Fig. 4) unfrozen
treatment (departure from parallelism;p = 0.515). A com-
bination of photobleaching and/or slow fluorescein re-
421
SPERM MEMBRANE FREEZING
TABLE 1. Membrane integrity of acrosomal and sperm head plasma membranes after freezing to -10C or -30°C
and warming to various temperatures.
Minimum freezing temperature
-10°C
(n= 4)
Rewarmed to:
1°C
5°C
10°C
30°C
Unfrozen control
Acrosome
intact or swollen
(mean + SEM)
57.5
61.0
72.0
51.5
68.0
+
+
+
+
+
10.9
12.8
6.7
8.2
6.9
-30°C
(n = 4)
Plasma membrane
intact
0
0
0
0
73.8 ± 6.8
Acrosome
intact or swollen
(mean + SEM)
87.5 +
74.8 +
80.5+
75.6 +
92.8 +
Plasma membrane
intact
4.8
13.8
11.7
13.3
1.4
0
0
2.6 ± 1.6
0
86.3 ± 4.9
aResults are expressed as percentages.
lease from the cells would account for the similar fluorescent losses in the control and frozen treatments, and it is
concluded that cooling to -10°C did not induce a significant increase in fluorescein efflux rate. However, initial fluorescence intensity of the frozen/thawed sample was significantly lower (difference between intercepts;p = 0.021)
than the intensity of the control. Because the initial intensity represents the fluorescence as close to the instant of
thawing as possible, this loss must have been incurred during the cooling and freezing processes not experienced by
the controls.
Freezing to -15 0C and re-warming caused a rapid decline in fluorescence within the first 5 sec after thawing
(data not shown); this corresponded to a temperature increase of approximately 1°C (i.e., 5 sec at +10°C/min).
Similar results were obtained with spermatozoa cooled to
-30 0C. Using images captured at 1.3-sec intervals (0.3 sec
for integration of 8 images + 1 second pause) after thawing, it was possible to resolve this fluorescent loss in more
detail. Figure 5 (A and B) shows representative plots made
from two groups of spermatozoa from separate ejaculates.
Immediately after thawing, the fluorescence intensity values
were not significantly lower than those of cells cooled to
-100 C; however, the intensities diminished rapidly below
the detection limit within 5 sec.
TEM
Sperm head plasma membranes and acrosomes. Virtually 100% of spermatozoa subjected to freezing exhibited
plasma membrane loss or damage over the head region,
regardless of the treatment (Table 1; Fig. 2, A-C).
Acrosomes were partially disrupted by freezing and
thawing. Although many acrosomes became swollen, they
retained an intact outer acrosomal membrane (Table 1; Fig.
2, A and C). Between 50% and 85% of spermatozoa maintained outer acrosomal membrane integrity after freezing
to -10°C or -30°C. Thawing temperature had no statistically significant influence upon this effect. However, in some
spermatozoa frozen to - 10°C and thawed to 1°C, the extent
of acrosomal swelling was minimal (Fig. 2A). Acrosomal
membrane vesiculation was not significantly increased by
freezing and thawing (p > 0.5).
Flagellarplasma membranes. The data for flagellar
plasma membranes are summarized in Figure 6. A surprisingly high incidence of intact flagellar plasma membranes
persisted after freezing and thawing to -10°C or -30°C.
However, the absolute values are probably overestimated
(see Materials and Methods), as longitudinal sections of
middle pieces (Fig. 7) revealed that membrane damage could
occur in localized regions overlying only one or two mitochondrial gyres (Fig. 7, arrows). In general, the principal
pieces retained more intact plasma membranes than middle pieces after freezing to either -10°C (p < 0.001) or
-30°C (p = 0.002).
Freezing to - 100C and then thawing to 1°C, 50C, or 10°C
failed to induce significant membrane disruption to middle-piece plasma membranes (p > 0.3 for comparisons with
control; Fig. 1,A and B). Further warming to 300C increased
the incidence of damaged middle-piece membranes (individual contrasts for 30 0C vs. 1C, 5°C, and 10°C;p = 0.045,
0.042, and 0.063, respectively).
Freezing to -30 0C and thawing to any of the target temperatures increased the incidence of damaged middle-piece
plasma membranes (p < 0.01) by 25-30%. However, a surprisingly high incidence (50-60%) of intact membranes was
still evident after freezing and thawing. Membrane damage
was not further increased by elevated thawing temperatures.
Principal-piece plasma membrane integrity was not significantly affected by freezing to -10 0C and then thawing
to 1C (p = 0.125). However, more cells became damaged
when the temperature was increased to 5C or above (Fig.
6) (individual contrasts for 1°C vs. 5C, 10°C, and 30°C; p
= 0.037, 0.017, and 0.014, respectively). Nevertheless, even
after thawing and re-warming to 30°C, more than 70% of
principal pieces were surrounded by an intact plasma
membrane (Figs. C and 7A).
Freezing to -30 0C and thawing caused a slight reduction
in the incidence of intact principal-piece plasma membranes (from 90% to approximately 77% intact); this effect
422
HOLT AND NORTH
100
A
PP *
80
'I
'''-.
MP
60
40
20
0
·
Control
·
·
·
1°
·
5"
·
10
·
30
·
·
Frozen to -10'C
B
100
80
PP e......
0
FIG. 7. Longitudinal section through the flagellum of a spermatozoon
frozen to -10°C. A) Thawed to 30°C; morphologically intact plasma membrane is seen along the principal piece (x39 600). B) Thawed to 1°C; although the plasma membrane along the middle piece is mainly intact, sites
of localized damage (arrows) are apparent (x28 600).
CO
60
spermatozoa, respectively (mean + SEM). These values did
not differ significantly from each other or from values for
8.6% intact
the unfrozen control, which contained 68.9
spermatozoa.
No between-treatment, or control vs. frozen, differences
were detected for any of the categories of damaged cells.
0
c
.1;
40
20
Rewarm temp. °C
Control
0 _········
1"
5"
10°
30°
DISCUSSION
Frozen to -30'C
FIG. 6. Percentages of intact plasma membranes of middle pieces (MP,
solid squares) and principal pieces (PP, solid diamonds) in spermatozoa
frozen to -10°C (A) or -30°C (B) and warmed to various post-thaw temperatures. Control treatments were fixed without prior cooling below 30°C.
was not statistically significant for any of the target re-warming temperatures (0.15 < p < 0.28). In this experiment,
significant variation between animals (p < 0.01) for the degree of membrane damage sustained obscured treatment
effects.
SEM
To see whether or not major plasma membrane damage
was induced when spermatozoa were frozen to -10°C in
a linear cooling profile, semen samples were frozen to -10°C
at 15°C/min on the cryomicroscope stage. They were rewarmed to 1°C, 5°C, or 10 0C at 10°C/min and prepared for
SEM as described in MaterialsandMethods. From the original hypothesis it was predicted that gross membrane damage would not be visible until the thawed spermatozoa
reached 10°C or above.
Samples frozen to -10C and re-warmed to 1° C, 5°C, or
0
10 C contained 53 + 8.9%, 56 + 6.4%, and 65 + 6.4% intact
The primary purpose of the present study was to investigate the related hypotheses that 1) the membrane damage
manifested after thawing is influenced by the minimum
temperature reached during freezing and 2) plasma membrane damage in frozen spermatozoa is induced not only
during thawing itself, but also during temperature elevation
after thawing.
Support for the hypothesis that conditions experienced
during freezing govern the severity of plasma membrane
damage was provided by two observations. 1) The fluorescein efflux rate for spermatozoa frozen to -10°C and then
re-warmed was not significantly greater than that for unfrozen control cells, and 2) increased efflux rate was induced by further cooling to -15C or -30°C. Since the
efflux rate must be governed by the permeability of the
diffusion barriers, it is concluded that the process of cooling from -10°C to -15°C increases plasma membrane or
mitochondrial permeability. The exact cause of the lesions
cannot be determined from the present studies.
Unless special precautions were taken to inhibit ice nucleation, the KHR solution used in these experiments began
to freeze on the cryomicroscope stage at about -2°C. As
the temperature decreased, the amount of unfrozen liquid
diminished and by -10°C the whole sample was frozen,
SPERM MEMBRANE FREEZING
although narrow unfrozen channels containing highly concentrated solute probably persist at this temperature [16,17].
Major differences induced by freezing to either -10 0C or
lower temperatures are therefore attributable either to intrinsic effects of temperature (for example, upon membrane lipid organization) or to increased osmolarity of the
extracellular environment. Crowe et al. [18] considered the
induction of lipid phase transitions by ice formation unlikely.
Treating spermatozoa with hyperosmotic conditions indicated that restoration of isotonic conditions caused considerably more plasma membrane damage than the initial
exposure. In terms of freeze/thawing processes, this result
supports the view that thawing is the more deleterious.
However, recent experiments with human spermatozoa [13]
have demonstrated that the conditions experienced during
exposure to the hyperosmotic conditions govern the severity of the subsequent cell damage. The investigators examined hyperosmotic exposure up to 5000 mOs and found
that human spermatozoa become the most significantly affected when placed in environments exceeding 3000 mOs.
The relationship between plasma membrane damage and
the minimum freezing temperature can probably be explained, therefore, as a consequence of exposure to increasingly hypertonic environments.
Although the post-thaw fluorescein efflux rate of spermatozoa cooled to -10°C was similar to that for unfrozen
control cells, the initial post-thaw fluorescence intensity of
the frozen spermatozoa was significantly lower. As this was
not attributable to photobleaching, which was prevented
before thawing, it might have been caused by slow leakage
of fluorescein during the period immediately before freezing. Slow leakage of fluorescein from intact cells has been
attributed to passage through membrane channels [19].
The ultrastructural evidence indicated that changes in
flagellar membrane permeability were not induced by gross
plasma membrane rupture during either freezing or thawing. This supports the hypothesis that permeabilization
mechanisms do not necessarily involve visible structural
damage. Middle- and principal-piece plasma membrane
damage was, however, induced when spermatozoa frozen
to -10°C were re-warmed through the temperature zone
of 10-30°C. The mechanism causing this effect is unclear,
but it may involve changes in membrane lipid organization
or protein conformation. Ample evidence exists for lipid
phase transition processes within this temperature range.
In previous studies of membrane lipid dynamics, lateral
phase transitions were detected at 17°C, 22°C [20], and 26°C
[21] in ram sperm plasma membranes; at 13.6°C and 21.1°C
in human and goat sperm plasma membranes, respectively
[22]; and centered on 18°C in boar sperm [23]. Protein tertiary structure is perturbed during freezing, partly as a result of water removal, and experiments have been performed that indicate the possible value of trehalose and
proline in protecting against this effect [24]. Spermatozoa
423
that had been cooled to -30°C failed to manifest this ultrastructural effect of re-warming but appeared on the basis
of the fluorescence data to have already undergone a permeability change during freezing. The sample cooling profiles for the bulk samples used for TEM studies differed
markedly from the linear protocols achieved with the
cryomicroscope. Although these differences did not lead to
conflicting experimental data, the possibility that spermatozoa in the two systems interact differently with crystalline
ice and hyperosmotic fluids should not be overlooked.
The ultrastructural observation that 100% of sperm head
plasma membranes were ruptured after any freezing treatment highlights the differences in membrane stability between major cellular compartments. A number of cytoskeletal proteins have been identified in spermatozoa (for
review, see [25]), but their role in maintaining the structural integrity of the sperm plasma membrane is still unclear. It seems likely, however, that the close attachment of
the middle-piece and principal-piece plasma membrane to
the underlying cytoplasmic structures provides considerable physical support against cryoinjury. These findings also
imply that caution should be exercised when water permeability coefficients for spermatozoa are calculated. Such
studies are undertaken in an effort to derive optimal cryopreservation protocols from theoretical principles, but at
present the predicted and experimentally derived ideal
cooling rates differ by 100- to 1000-fold [26]. One explanation might be that techniques used to determine cell
swelling and membrane permeabilization in response to
changes in tonicity, i.e., flow cytometry and fluorescence
microscopy, provide average values for cells that do not
truly represent any of the individual compartments. The
evidence presented here demonstrates that the plasma
membrane overlying the sperm head is considerably more
labile than that of the flagellum.
The results obtained in the present study are compatible
with the original hypothesis that increased membrane permeability induced by freezing is not necessarily accompanied by major structural damage. They also show that
restoration of isotonic equilibrium is a significant cause of
membrane damage during cryopreservation and indicate
that better post-thaw semen handling techniques are likely
to improve the viability of cryopreserved semen.
The present findings also suggest that treatments capable
of preventing molecular reorganization of the plasma
membrane during freezing would probably inhibit subsequent thawing-induced membrane permeabilization and
assist in the maintenance of sperm viability. Snchez-Partida
[27] reported beneficial effects on the post-thaw motility of
ram spermatozoa when proline and glycine betaine were
included in the cryoprotective diluents. This accords with
the view that proline functions as a cryoprotectant for lipids
and proteins in membranes, interacting directly with membrane lipids and altering their hydration state and phase
behavior [24]. Egg yolk lipoproteins may also act in this way
424
HOLT AND NORTH
and have been shown to cause significant inhibition of the
post-thaw fluorescein loss [10] from ram spermatozoa. Their
interaction with plasma membrane components has been
suggested previously [28]. The observation that detergent
treatment of egg yolk lipoproteins improves their cryoprotective efficacy [29], perhaps by modifying their tertiary
structure, also suggests that a specific interaction with
membrane lipids or proteins is required for cryoprotection.
ACKNOWLEDGMENTS
We are grateful to Dr. D.M. Holburn, Department of Engineering, University of
Cambridge, for developing the image analysis software used in this study and to Dr.
P.F. Watson (Royal Veterinary College, London) for providing the ram semen.
REFERENCES
1. Barratt CLR, Matson PL, Holt WV. British Andrology Society guidelines for screening of donors for donor insemination. Hum Reprod 1993; 8:1521-1523.
2. Holt WV. Advances in artificial insemination and semen freezing in mammals.
In: Moore HDM, Holt WV, Mace GM (eds.). Symp Zool Soc Lond 1992; 64:1935.
3. Shannon P. Factors affecting semen preservation and conception rates in cattle.
J Reprod Fertil 1978; 54:519-527.
4. Hammerstedt RH, Graham JK, Nolan JP. Cryopreservation of mammalian sperm:
what we ask them to survive. J Androl 1990; 11:73-88.
5. Watson PF. Cryopreservation of ram semen. In: Jornadas Inter. Reprod. Anim.
Al.I.1990; 5:95-112. (Asociacion Espanola de Especialistas en reproduccion animal, inseminacion artificial y nuevas tecnologias. Acta ponencias y communicaciones 14-17 June, Zaragoza).
6. Mazur P. Freezing of living cells: mechanisms and implications. Am J Physiol
1984; 247:C125-C142.
7. Mazur P. Basic concepts in freezing cells. In: Johnson LA,Larsson K (eds.), Proceedings of the First International Conference on Deep Freezing of Boar Semen.
Sweden: Uppsala Swedish University of Agricultural Sciences; 1985: 91-111.
8. Lovelock JE, Polge C. The immobilization of spermatozoa by freezing and thawing and the protective action of glycerol. Biochem J 1954; 58:618-622.
9. Watson PF, Duncan AE. Effect of salt concentration and unfrozen water fraction
on the viability of slowly frozen ram spermatozoa. Cryobiology 1988; 25:131142.
10. Holt WV, Head MF, North RD. Freeze-induced membrane damage in ram spermatozoa is manifested after thawing-observations with experimental cryomicroscopy. Biol Reprod 1992; 46:1086-1094.
11. Vishwanath R, Swan MA, White IG. Effect of triton X-100 on ultrastructure, reactivation, and motility characteristics of ram spermatozoa. Gamete Res 1986; 15361371.
12. Woolley DM, Richardson DW. Ultrastructural injury in human spermatozoa after
freezing and thawing. J Reprod Fertil 1978; 53:389-394.
13. Gao DY, Ashworth E, Watson PF, Kleinhans FW, Mazur P, Critser JK Hyperosmotic tolerance of human spermatozoa-separate effects of glycerol, sodiumchloride, and sucrose on spermolysis. Biol Reprod 1993; 49:112-123.
14. Harrison R, Vickers SE. Use of fluorescent probes to assess membrane integrity
in mammalian spermatozoa. J Reprod Fertil 1990; 88:343-352.
15. Pearse AGE. Histochemistry, Theoretical and Applied. London: Churchill; 1968.
16. Nei T. Mechanism of haemolysis of erythrocytes by freezing, with special reference to freezing at near-zero temperatures. In: Wolstenholme GEW, O'Connor
M (eds.), The Frozen Cell. London: Churchill; 1970: 131-147.
17. Mazur P, Rigopoulos M. Contribution of unfrozen fraction and of salt concentration to the survival of slowly frozen human erythrocytes: influence of warming
rate. Cryobiology 1983; 20:274-289.
18. Crowe JH, Carpenter JF, Crowe LS, Anchordoguy TJ. Are freezing and dehydration similar stress vectors? A comparison of modes of interaction of stabilizing
solutes with biomolecules. Cryobiology 1990; 27:219-231.
19. McGann LE, Yang H, Walterson M. Manifestations of cell damage after freezing
and thawing. Cryobiology 1988; 25:178-185.
20. Holt WV, North RD. Thermotropic phase transitions in the plasma membrane of
ram spermatozoa. J Reprod Fertil 1986; 78:447-457.
21. Wolf DE, Maynard VM, McKinnon CA, Melchior DL. Lipid domains in the ram
sperm plasma membrane demonstrated by differential scanning calorimetry. Proc
Natl Acad Sci USA 1990; 87:6893-6896.
22. Crowe JH, Hoekstra FA, Crowe LM, Anchordoguy TJ, Drobnis E. Lipid phase transitions measured in intact cells with Fourier transform infrared spectroscopy.
Cryobiology 1989; 26:76-84.
23. Drobnis EZ, Crowe LM, Berger T, Anchordoguy TJ, OverstreetJW, CroweJH. Cold
shock damage is due to lipid phase-transitions in cell-membranes: a demonstration using sperm as a model. J Exp Zool 1993; 265:432-437.
24. Rudolph AS, Crowe JH. Membrane stabilization during freezing. The role of two
natural cryoprotectants, trehalose and proline. Cryobiology 1985; 22:367-377.
25. Camatini M, Colombo A, Bonfanti P. Cytoskeletal proteins involved in cell polarity: mammalian spermatozoa and insect enterocytes. In: Lanzavecchia G, Valvassori R (eds.), Form and Function in Zoology, Selected Symposia and Monographs UZI; 1991; 5:15-31.
26. Noiles EE, Mazur P, Watson PF, Kleinhans FW, Critser JK Determination of water
permeability coefficient for human spermatozoa and its activation-energy. Biol
Reprod 1993; 48:99-109.
27. S/mchez-Partida LG, Maxwell WMC, Paleg LG, Setchell BP. Proline and glycine
betaine in cryoprotective diluents for ram spermatozoa. Reprod Fertil Dev 1992;
4:113-118.
28. Watson PF. The protection of ram and bull spermatozoa by the low density lipoprotein fraction of egg yolk during storage at 5°C, and deep freezing. J Therm
Biol 1976; 1:137-141.
29. Pontbriand D, Howard JG, Schiewe MC, Stuart LD, Wildt DE. Effect of cryoprotective diluent and method of freeze-thawing on survival and acrosomal integrity
of ram spermatozoa. Cryobiology 1989; 26:341-354.