Effects of cryoprotectants and ice-seeding temperature on intracellular freezing and survival of human oocytes

Fouad S. Trad1,4, Mehmet Toner2 and John D. Biggers3,5

1 Department of Obstetrics, Gynecology and Reproductive Biology, Brigham and Women's Hospital, Harvard Medical School, 75 Francis Street, Boston, MA 02115, 2 Center for Engineering in Medicine and Surgical Services, Massachusetts General Hospital, Harvard Medical School, 51 Blossom Street, Boston, MA 02114 and 3 Department of Cell Biology, Harvard Medical School, 240 Longwood Avenue, Boston, MA, USA


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 References
 
The accurate determination of the freezing conditions that promote intracellular ice formation (IIF) is crucial for designing cryopreservation protocols for cells. In this paper, the range of temperatures at which IIF occurs in human oocytes was determined. Fresh oocytes with a germinal vesicle, failed-to-fertilize (metaphase I and metaphase II stages) and polyspermic eggs were used for this study. The occurrence of IIF was first visualized at a cooling rate of 120°C/min using a programmable thermal microscope stage connected to a videomicroscope. Then, with a cooling rate of 0.2°C/min, the seeding temperature of the extracellular ice was modified to decrease the incidence of IIF and increase the survival rate of frozen–thawed human oocytes. After adding different cryoprotectants, the median temperature of IIF (TMED) was decreased by ~23°C in mouse and only by ~6.5°C in human oocytes. Using 1.5 M propylene glycol and seeding temperatures of -8.0, -6.0 and -4.5°C, the incidence of IIF was 22/28 (78%), 8/24 (33%) and 0/33 (0%) and the 24 h post-thaw survival rate was 10/31(32%), 19/34 (56%) and 52/56 (93%) respectively. The results show that IIF occurs more readily in human oocytes, and that ice seeding between -6°C and -8°C triggers IIF in a large number of human oocytes. Undesirable IIF can be prevented and survival rates maximized by raising the seeding temperature as close as possible to the melting point of the solution, which in our instrument was -4.5°C.

Key words: cryobiology/freezing/human/intracellular ice formation/oocyte


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 References
 
Routine human oocyte cryopreservation remains an elusive objective (Oktay et al., 1988Go; Bernard and Fuller, 1996Go). The need for such a technique has become more pressing with the increase of donor oocyte programmes and the potential risk of disease transmission. Women who are at risk of losing their ovarian function to surgery, cancer treatment or premature menopause will also benefit from oocyte banking. This technique should also circumvent many ethical and legal objections to human embryo cryopreservation.

Good fertilization and cleavage rates (Al-Hasani et al., 1987Go; Chen, 1988Go; Gook et al., 1994Go, 1995Go; Toth et al., 1994Go; Kazem, 1995Go) as well as a few pregnancies (Van Uem et al., 1987Go; Chen, 1988Go; Tucker et al., 1996Go; Porcu et al., 1997Go) have been reported using human oocytes that survived cryopreservation. Although post-thaw survival rates ranging from 20 to 80% have been achieved, these freezing protocols have not been robust enough to translate into a reproducible, clinically useful technique. Pre-freezing events, such as oocyte exposure to cryoprotectant solutions and simple cooling, have been shown to be relatively innocuous, and do not inhibit fertilization and development (Hunter et al., 1991Go; Bernard et al., 1992Go; Gook et al., 1995Go). Thus, the step that limits the efficiency of this multi-step process seems to be confined to the freeze and thaw procedure.

Hitherto, attempts to improve the freeze and thaw process in human oocytes have used protocols empirically derived from animal models. An alternative approach relies on studying the biophysical characteristics of oocytes under sub-zero conditions by observing water transport and intracellular ice formation (IIF) using a cryomicroscope. These biophysical characteristics are then used in a theoretical model to derive an optimized freezing protocol (Karlsson et al., 1996Go). The currently accepted theory states that the avoidance of large and numerous intracellular ice crystals is a necessary, but not sufficient, condition for a cell to survive freezing (Mazur, 1984Go). One strategy to prevent IIF after extracellular ice seeding is to provide time for the cell to dehydrate before reaching the lower temperatures at which ice nucleation occurs. Typically, dehydration of oocytes occurs during a slow (<2°C/min) freezing, with the diffusion of intracellular water to the hyperosmotic extracellular media generated by the incorporation of pure water in the growing crystals. However, slowly cooled cells suffer damage due to long exposure to high electrolyte concentrations, excessive cell dehydration, mechanical effects of the external ice, and a combination thereof. Thus the cooling rates should be fast enough to minimize the long exposure of oocytes to deleterious freezing conditions, but slow enough to avoid the damaging effect of IIF. To design a freezing protocol, the biophysical parameters governing not only water transport across the oolemma but also ice nucleation need to be determined. To date, the parameters for water transport are known for bovine (Myers et al., 1987Go) and mouse oocytes at supra-zero (Leibo, 1980Go) and sub-zero temperatures (Toner et al., 1990aGo). Ice nucleation parameters are known for both mouse oocyte (Toner et al., 1990bGo; Karlsson et al., 1996Go) and macaque (Younis et al., 1996Go) oocytes. Some water transport parameters have been determined in human oocytes (Hunter et al., 1992Go), but not those concerned with ice nucleation parameters.

In this paper, we show that the median temperature at which IIF occurs in human oocytes is not depressed to a significant level by the addition of several cryoprotectants. This unexpected finding indicates that IIF occurs more readily in human oocytes. We also report that the extracellular ice seeding temperature significantly affects IIF in human oocytes. The data are used to suggest modifications of a current method of freezing human oocytes to increase the post-thaw survival rate.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 References
 
Ovarian stimulation, oocyte retrieval and culture protocol
Human oocytes were obtained from patients undergoing in-vitro fertilization (IVF) at the Center for Reproductive Medicine, Department of Obstetrics, Gynecology and Reproductive Biology, Brigham and Women's Hospital. Only those oocytes not useful clinically were used for this study. The institutional Human Research Committee approved the project. Ovarian stimulation was achieved using various combinations of gonadotrophin and gonadotrophin releasing hormone agonists, depending on each patient. Follicular growth was monitored by ultrasound and oestradiol-17ß measurements. Human chorionic gonadotropin (HCG) was administered after three follicles reached a diameter of 18–20 mm. Oocyte retrieval was performed transvaginally 36 h later. Oocyte–cumulus complexes were individually cultured in Ham's F-10 medium without hypoxanthine (N3389; Sigma Chemical Company, St Louis, MO, USA) supplemented with 10% synthetic serum substitute (SSS) (99193; Irvine Scientific, Irvine, CA, USA) at 37.2°C in an atmosphere of 5.5% CO2 in air.

Human oocytes
The following categories of human oocytes were used for this study.

Fresh oocytes with a germinal vesicle
These were retrieved from patients scheduled for intracytoplasmic sperm injection (ICSI) who had their oocytes evaluated for maturity prior to injection. The cumulus mass was completely removed 4 h after the retrieval with a small bore glass pipette following a 30 s exposure to 80IU/ml hyaluronidase (H3506; Sigma) in HEPES-buffered HTF (9963; Irvine Scientific). The immature eggs with a germinal vesicle were selected for the study.

Oocytes after exposure to spermatozoa
Oocyte–cumulus complexes were inseminated 4–6 h after retrieval using 1 500 000 spermatozoa/ml. The spermatozoa were processed with a Percoll column (Punjabi et al., 1990Go). Fertilization was assessed at 14–19 h post insemination. Two categories of these oocytes were selected for study.

Failed-to-fertilize.
Unfertilized oocytes were evaluated for maturity and classified as metaphase I (MI) and metaphase II (MII) oocytes, based on the presence or absence of a polar body. Eggs were used within 24–30 h of the retrieval. Because of their scarce number, unfertilized oocytes with a germinal vesicle were not used for this study.

Polyspermic embryo.
Fertilized oocytes with an abnormal number of pronuclei were used while still in the uncleaved stage. Eggs were used 24–30 h after retrieval.

Mouse eggs
The procedure for the collection of mouse oocytes has been described previously (Jackson and Keissling, 1989Go). Four to six week old BDF mice were obtained from the Jackson Laboratories (Bar Harbor, ME, USA). Female mice were induced to superovulate by intraperitoneal (i.p.) injection of 5IU of pregnant mare serum gonadotrophin (Sigma, G-4877), followed 48 h later by an i.p. injection of 5IU of HCG (Sigma, CG-5). Twelve to 15 h later, metaphase II oocytes were collected and used for experiments.

Cryoprotectant solutions
The following saline solutions were used in this work: Dulbecco's phosphate-buffered solution (DPBS) (450–1300EC; Gibco) and PB1 (DPBS + glucose) (4501300eb; Gibco). Both solutions were supplemented with penicillin G (0.095 g/l) and streptomycin (0.05 g/l). All freezing solutions were prepared using PB1. The following cryoprotectant solutions were prepared: (i) 1.5 M propylene glycol (P1009; Sigma) in PB1 (PG), (ii) 1.5 M dimethyl sulphoxide (5879; Sigma) in PB1 (DMSO), and (iii) 1.5 M ethylene glycol (E-9129; Sigma) in PB1 (EG). For the oocyte cryopreservation experiments, PG, DMSO, and EG were supplemented with 20% SSS in order to minimize oocyte loss due to sticking in the straws. For the cryomicroscopy experiments, the freezing solutions were used without adding SSS, since oocytes were not recovered after a cryomicroscopy experiment. Only 1.5 M cryoprotectant solutions were used, except when a stepwise addition of 0.5 M and 1.0 M solutions of PG was needed. For the removal of cryoprotectant solutions from oocytes, 1.08 M sucrose (S0389; Sigma) in PB1 supplemented with 20% SSS was used for both conventional and modified cryopreservation protocols.

Oocyte cryomicroscopy
The occurrence of IIF in oocytes, in the presence and absence of a cryoprotectant, was observed using the cryomicroscopy system described by Cosman et al. (1989). Our equipment employed a programmable thermal microscope stage (Thermascope; Interface Technique, Cambridge, MA, USA) connected to a video microscopy system. Depending on availability, two to 15 human oocytes were equilibrated in the cryoprotectant solutions for 20 min and then placed in a 50 µl drop on the cryomicroscope stage under a coverslip sealed with silicone grease. The initiation of the sample freezing was achieved by slightly supercooling the solution to –6.5°C and manually triggering seeding extracellular ice by contacting the edge of the sample with a chilled forceps. Controlled cooling at a rate of 120°C/min to –60°C was initiated as soon as the growing ice front had engulfed all oocytes in the field of view. Oocytes were not held at the extracellular ice seeding temperature because the thin layer of 50 µl solution had a thickness of ~50–100 µm which allows immediate latent heat dissipation. A rapid cooling rate of 120°C/min was chosen to minimize water efflux, thus reducing the effect of water transport on ice nucleation. Under bright-field illumination, IIF was manifested by a sudden darkening of the cytoplasm believed to be caused by the light scattering due to microscopic ice crystals and/or bubbles in the oocyte. The temperature at which IIF occurred in each oocyte was determined by reviewing the video recordings of each freezing experiment. The observed IIF temperature for each individual oocyte was corrected for the thermal gradient across the stage by subtracting the corresponding temperature of the ice front during the thaw, and adding the melting point of the solution (Toner et al., 1990bGo). In subsequent experiments, the effect of varying the cooling rates and extracellular ice seeding temperatures on IIF was also determined. Cooling rates of 0.2°C/min and 10°C/min were used to investigate the role of dehydration on IIF. Also, at a cooling rate of 0.2°C/min, the extracellular ice seeding temperature was modified by supercooling the solution to –4.5, –5.5, –6.0, –6.5 and –8°C before contacting the edge of the sample with a chilled forceps. The incidence of IIF for each experimental setting was recorded.

Oocyte cryopreservation
These experiments were designed to test the modified freezing protocol associated with a minimal incidence of IIF by assessing the viability of frozen–thawed oocytes. Control experiments were conducted using the one step method for human embryo cryopreservation derived from a bovine embryo freezing protocol (Leibo 1984Go) and then modified (Nowrouzi et al., 1991Go) for human use.

The conventional cryopreservation protocol
A Biocool freezer (FTS System, Stone Bridge, NY, USA) was used, with a reagent grade alcohol at 95% (Thomas Scientific, 4916-E52) and a TPC-44.E programmable controller set for the following protocol: starting temperatures -6°C and -8°C, ramp rate -0.2°C/min, holding temperature -40°C, holding time 15-30 min. In preparation for the freezing protocol, the oocytes were transferred to a PB1 solution for 10 min followed by PG equilibrated at room temperature for a maximum of 20 min. The 0.25 ml straws (IMV, France, A101) were loaded as follows: 10 mm of PG, 10 mm of air, 10 mm of PG containing the oocytes, 10 mm of air, 65 mm of sucrose solution, 10mm of air. The air was drawn up until the first PG column moistened the PVA powder between the two plugs and no more air could be drawn in. Both ends of the straw were sealed by heat and the plugged end was fitted with a 0.5 ml straw (IMV, A102) to be used as a handle. The straws were maintained in a horizontal position until they were briefly plunged into the alcohol bath at -6°C and -8°C. The top of the sucrose column and the bottom of the PG column not containing the oocytes were then seeded using a chilled forceps. All the straws were placed perpendicularly in a freezing basket (IMV, M001) suspended in the 95% alcohol bath and held for 15 min at the extracellular ice seeding temperature. After confirming that the straws were frozen, the ramp was activated as described above. At the end of the 15 min holding time at -40°C, the straws were quickly transferred to a liquid nitrogen canister, and then moved to a storage tank for 1–7 days.

For thawing, the straw was removed from the liquid nitrogen tank and laid at room temperature for 2 min with both ends supported by a slant rack. If, after wiping with a Kim wipe, condensation formed, an additional 30 s was allowed. After taking the handle off, and holding the unplugged end, the straw was shaken a few times like a thermometer until there was only one air space at this end. The straw was then placed, plugged end down, in 15 ml centrifuge tube containing Milli-Q water in a 37°C water bath for 3 min. The straw was then turned over, plugged end up, and placed in a 15ml centrifuge tube containing Milli-Q water at room temperature for 1 min. The straw was emptied on a dry dish under direct vision using a dissecting microscope. The oocytes were washed twice in PB1 at room temperature and allowed to equilibrate for 10 min before transfer to a culture dish in the conditions described above.

The modified cryopreservation protocol
To induce seeding of extracellular ice at -4.5°C, the following modifications were introduced to the conventional protocol described above. The Biocool unit was programmed as follows: segment 1: start temperature -3.5°C, ramp -0.2°C/min, holding temperature -15°C, holding time 0 min; segment 2: ramp -1°C/min, holding temperature -40°C, holding time 15–30 min. The IMV straws were loaded as follows: 10 mm of PG, 10 mm of air, 20 mm of PG containing the oocytes, 10 mm of air, 55 mm of sucrose and 10 mm of air. The unplugged end was fitted with the 0.5 ml IMV straw. The straws were held perpendicularly, plugged end up, until all the oocytes migrated to the lower meniscus of the 20mm PG column. Their exact position was checked under a dissecting microscope. Meanwhile, the freezing programme was activated from the starting temperature of -3.5°C, then put on hold to maintain the alcohol bath temperature at -4°C. The straw was inverted, and briefly dipped in the alcohol bath for 10–15 s at a 45° angle. All the temperatures were measured at the tip of a temperature probe located at the level of the 20 mm PG column. The lower meniscus of the 20 mm PG column was seeded in a location opposite to the oocytes, and replaced promptly in the alcohol bath in the slanted position. Simultaneously, the freezing ramp was reactivated to reach -4.5°C. This temperature was held for 20 min. After checking for seeding of the sucrose column without removing the straw from the alcohol bath, we proceeded with the protocol as above.

Attention to the following details was important to ensure adequate extracellular ice seeding. Because seeding of extracellular ice in an adjacent column at higher temperatures did not ensure proper ice formation in the cryoprotectant column containing the oocytes, the latter was seeded directly. Consequently, and to prevent accidental exposure of the eggs to ice seeding, the size of the cryoprotectant column containing the oocytes was doubled to 20 mm. The oocytes were also moved as far as possible from the seeding area by placing the straw perpendicularly, and their location was verified at the opposite meniscus prior to seeding of extracellular ice. The straw was placed in a 45° slanted position in the alcohol bath to allow the oocytes to move slowly toward the frozen area. In a perpendicular position, the oocytes reached the bottom of the PG column in 2–5 min, as opposed to 10–15 min in the slanted position. This timing was critical. It allowed new ice to grow upward from the frozen region and trap the oocytes migrating downward in hyperosmolar channels located between the growing ice crystals.

Evaluation of oocyte survival
Oocyte survival was assessed at 2 and 24 h post-thaw. Signs of oocyte damage included zonal fracture, discoloured ooplasm, pyknosis and cytoplasmic disruption. A very good correlation was previously reported between delayed (24 h) morphological assessment, and vital staining when demonstrating viability of frozen–thawed human oocytes (Pensis et al., 1989Go).

Biometrical methods
Box plots

The distributions of the temperature of IIF under various experimental conditions were plotted as proportional notched box plots to show the 10th, 25th, 50th (median), 75th, 90th percentiles (Kafadar, 1985Go). Observations that fall outside the 10th and 90th percentile were considered outliers. The width of the box is proportional to the number of observations. The notches were constructed using the formula 1.57x(75th percentile – 25th percentile)/n. If the notches do not overlap, the medians are considered significantly different at a significance level of P = 0.05. The box plots were computed using the Number Cruncher Statistical System 97 (NCSS, Kaysville, UT, USA).

Contingency tables

The data in Tables I and IIGoGo are presented as doubly ordered rxc contingency tables. The homogeneities of the row (seeding temperatures) distributions were tested using the one-sided exact Jonckheere–Terpstra test of significance (Pirie, 1983Go). The value of P and its 99% confidence limits were computed by the Monte Carlo method using the StatXact 3 for Windows computer package (Cytel Software Corporation, Cambridge, MA, USA).


View this table:
[in this window]
[in a new window]
 
Table I. The numbers of human oocytes developing intracellular ice (IIF), and their proportions, after freezing at -0.2°C/min using various seeding temperatures in 1.5 M propylene glycol
 

View this table:
[in this window]
[in a new window]
 
Table II. Post-thaw morphological survival rates after 2 and 24 h in frozen–thawed human oocytes at different seeding temperatures
 
Logistic regression

Two of the sets of data in Tables I and IIGoGo were also examined by fitting linear logistical regression equations of the form:


where y = logit p = ln (p/(1 – p)), p is the proportion responding, T is the extracellular ice seeding temperature, b0 is the intercept, and b1 is the slope of the regression. All computations were done using the LogXact for Windows computer package (Cytel).

Model threshold distributions

Consider a population of n oocytes. Hypothetically each oocyte will form intracellular ice as a result of extracellular ice seeding at some threshold temperature (TII). Likewise the oocyte will suffer irreversible damage and die as the result of seeding at some threshold temperature (TID). These two temperatures are not necessarily the same, although they may be highly correlated. In both cases frequency distributions [f(TID), f(TII)] of threshold values can be used to model the population of oocytes. The corresponding cumulative distributions of survivors (F(TID)) and oocytes in which intracellular ice forms (F(TII)) are given by:


Specifically, we will assume that the two frequency distributions can be modelled by the logistic distribution, which is very close to the normal distribution and has properties that make it relatively simple to use (Malik, 1985Go). Finney (Finney 1947Go) has described a comparable use of the logistic distribution, including the computations involved, in a discussion of tolerance distributions which arise in the field of biological assay. He showed that the logistic distribution can be estimated by substituting equation (1) in the equation:


Thus the values of b0 and b1 obtained by fitting (1) to the data in Tables I and IIGoGo have been substituted in (3) to estimate the two frequency distributions of the threshold temperatures that cause irreversible damage [f(TID)] and IIF [f(TII)]. The corresponding cumulative distributions can be computed by transforming the fitted linear equations from the logit response scale to the proportion response scale using the inverse of equation (1):


Results

Preliminary observations
Preliminary experiments with human oocytes showed significant differences in the median temperature of IIF (TMED) in fresh oocytes with a germinal vesicle, unfertilized MI or MII oocytes and polyspermic oocytes (Figure 1Go). The differences were considered marginal and sufficiently small to justify pooling the data from all types of human oocytes when analysing all subsequent experiments.



View larger version (14K):
[in this window]
[in a new window]
 
Figure 1. Proportional, notched box plots depicting the 10th, 25th, 50th (median), 75th and 90th percentiles of the intracellular ice formation (IIF) temperatures of several types of human oocytes [fresh germinal vesicle (GV) stage, metaphase I (MI) stage failed-to-fertilize, metaphase II (MII) stage failed-to-fertilize, and polyspermic], frozen in Dulbecco's phosphate-buffered saline (DPBS) at a rate of 120°C/min. If the notches do not overlap, the medians are considered significantly different for a P-value of 0.05. Observations that fall outside the 10th and 90th percentiles are considered outliers and are shown as closed circles.

 
The effects of PG, DMSO and EG on the distribution of the temperatures of IIF in human oocytes are shown in Figure 2Go. The distributions of IIF vary considerably in location and scatter. The TMED in human oocytes frozen without cryoprotectant was –8.5°C. The addition of the cryoprotectants (PG, DMSO and EG) only slightly depressed the temperature of IIF in human oocytes. PG, DMSO and EG depressed the TMED by only 5.0°C, 8.2°C and 6.5°C, respectively.



View larger version (12K):
[in this window]
[in a new window]
 
Figure 2. Proportional, notched box plot of the intracellular ice formation (IIF) temperatures of human oocytes frozen in the presence of Dulbecco's phosphate-buffered saline (DPBS) and of PB1 supplemented with the cryoprotectants propylene glycol (PG), dimethyl sulphoxide (DMSO) and ethylene glycol (EG) at a rate of 120°C/min.

 
Comparison of human and mouse oocytes
The unexpected finding that all three cryoprotectants were ineffective in depressing the IIF temperature of human oocytes prompted us to test rigorously the accuracy of our experimental approach using mouse oocytes as a well-established model system. The TMED in mouse oocytes frozen without cryoprotectant was -12.5°C, which was decreased by 19.6°C, 25.9°C and 25.1°C by the addition of PG, DMSO and EG, respectively (Table IIIGo). Furthermore, the large difference between human and mouse oocytes in the depression of TMED produced by the addition of cryoprotectant has been confirmed by freezing the oocytes of the two species together. Figure 3AGo–D shows a representative set of eight human and 15 mouse oocytes during seeding at -6.5°C, and subsequent cooling at 120°C/min in PG. The micrograph depicts the behaviour of oocytes at -14.4°C, -21.1°C and -37.7°C, respectively. When the temperature reached -14.4°C, IIF had been observed in 7/8 human oocytes and in 0/15 mouse oocytes. Mouse oocytes only began to undergo IIF when the temperature reached -21.1°C (2/15). When the temperature reached -37.7°C, IIF was observed in 14/15 mouse oocytes.


View this table:
[in this window]
[in a new window]
 
Table III. Temperature (T) of intracellular ice formation (IIF) in mouse oocytes frozen at -120°C/min, in the absence and presence of cryoprotectants
 


View larger version (102K):
[in this window]
[in a new window]
 
Figure 3. Cryostage pictures combining eight large human oocytes (right) and 15 smaller mouse oocytes (left) with temperatures measured at the tip of the thermocouple. The oocytes were frozen in propylene glycol at a rate of 120°C/min. (A) Temperature -6.5°C: seeding is occurring with the ice front rapidly moving across the stage. (B) Temperature -14.4°C: IIF (sudden darkening of the cytoplasm) occurred in 7/8 human oocytes and none of the mouse oocytes. (C) Temperature -21.1°C: intracellular ice formation (IIF) occurred in all human oocytes and in 2/15 mouse oocytes. (D) Temperature -37.7°C: IIF occurred in 14/15 of the mouse oocytes. Original magnification x50.

 
Figure 4Go shows the distributions of the temperatures of IIF in all the human and mouse oocytes frozen together in PG. The location of the two distributions is highly significantly different with TMED in mouse of -32.7°C and in human of -14.4°C.



View larger version (9K):
[in this window]
[in a new window]
 
Figure 4. Proportional, notched box plots showing the 10th, 25th, 50th (median), 75th and 90th percentiles of the intracellular ice formation (IIF) temperatures of mouse and human oocytes frozen simultaneously in 1.5 M propylene glycol at a rate of 120°C/min.

 
Possible confounding factors
In an attempt to probe further the difference in IIF temperatures of mouse and human oocytes, the following series of experiments was done.

Uptake of cryoprotectant (PG)
Given the larger size of human oocytes, the equilibration of PG with human oocytes was evaluated from their volumetric behaviour in PG, as assessed by measuring the oocyte projected area and converting it to volume by assuming spherical geometry. Specifically, five human oocytes were placed in 1.5 M PG. They shrank rapidly within 30 s and then, relatively slowly, returned to their near-original volume in ~6 min (data not shown). This is the classical physiological response of a cell exposed to a permeable solute, and clearly demonstrates that human oocytes were nearly equilibrated with 1.5 M PG during our cryomicroscopy experiments.

Cytoplasmic damage due to osmotic shock
To rule out the possibility that osmotic damage to human oocytes during one-step addition of 1.5 M PG was causing elevated IIF temperatures, 33 human oocytes were first exposed sequentially to 0.5 M, 1.0 M and 1.5 M PG for 5 min each and then frozen at a rate of 120°C/min in 1.5 M PG. The TMED was -12.7°C, very similar to our previous estimate of -13.5°C. Furthermore, gross membrane damage after prolonged exposure (90 min) of five human oocytes to PG could not be demonstrated in the form of superficial blebs.

Rate of freezing
Forty-seven human oocytes in PG were frozen at a rate of -10°C/min. Although the TMED decreased to -17.7°C, potentially accounting for minimal water diffusion, it remained well above nucleation temperatures of mouse oocytes (below -30°C).

Oocyte culture
Since human oocytes were cultured for ~24 h prior to the cryomicroscopy experiments, the effect of culture on mouse oocyte IIF was also investigated. Twelve mouse oocytes cultured for 24 h were frozen in PG at 120°C/min. The observed TMED of -28.8°C was similar to the TMED determined in fresh mouse oocytes of -32.1°C.

Effect of seeding temperature on IIF
Groups of human oocytes were frozen on the cryomicroscope at a rate of -0.2°C/min, until the temperature was -40°C. The groups were seeded at -8.0, -6.5, -6.0, -5.5 and -4.5°C. The numbers of oocytes undergoing IIF were determined at the end of each run. The pooled results are shown in Table IGo. The exact Jonckheere-Terpstra test of significance shows that there are very significant differences between the effects of seeding temperature on the incidence of IIF (P < 0.001). The probability of IIF progessively falls as the seeding temperature is increased above 6.5°C. A linear logistic regression was fitted to the data (Table IGo, Figure 5Go), and its estimated slope was found to be significantly negative. The very significant deviance, however, indicates that the model is a poor fit to the data. Further analysis of the residuals showed that this lack of fit is mainly due to a small group of eight oocytes all of which developed intracellular ice after seeding at -6.5°C. When this group was omitted from the regression analysis the deviance became non-significant although the estimated slope was only slightly changed. Thus the sigmoid regression line shown in Figure 5Go was computed, retaining the group with apparently extreme responses.



View larger version (9K):
[in this window]
[in a new window]
 
Figure 5. Estimated cumulative distributions of intracellular ice formation (IIF) and survivors 24 h post-thaw on seeding temperature in human oocytes. The bar represents the `safe window' for seeding temperatures in human oocytes.

 
Effect of modified freezing protocol on survival after thawing
The importance of increasing the extracellular ice seeding temperature was verified by comparing frozen–thawed oocyte survival rates in the conventional freezing protocol with seeding temperatures of -6.0°C and -8.0°C, and also in a modified protocol with ice seeding temperature of -4.5°C. The pooled results are shown in Table IIGo. The exact Jonckheere–Terpstra test shows that there are very significant differences between the effects of seeding tempertature (P < 0.001). The probability of survival increased as the seeding temperature was raised. A linear logistic regression has been fitted to the data obtained 24 h post-thaw (Table IIGo, Figure 5Go). The deviance was not significant, indicating that the model adequately represents the data. The estimated slope was very significantly positive showing a clear increase in the incidence of post-thaw survival as the seeding temperature was raised.

Discussion
Cryomicroscopy has been used to determine the characteristics of IIF in human oocytes with and without several penetrating cryoprotectants. The results demonstrated that although the IIF temperature of human oocytes was not depressed by the addition of several cryoprotectants, a modified freeze and thaw protocol for human oocytes with high survival rates could still be designed.

Choosing an optimal seeding temperature
The ascending and descending sigmoid regression lines shown in Figure 5Go are independent estimates of the cumulative proportions of oocytes undergoing IIF and oocytes surviving after extracellular ice seeding in a temperature range of -8°C and -4.5°C. Both sets of oocytes were frozen in the presence of PG. The curves are almost inversely symmetrical, and intersect at a point whose abscissa corresponds to a seeding temperature of 6.7°C. This relationship is similar to the high correlation observed between the incidence of IIF and cell survival on cooling rates (Toner et al., 1993Go). Fifty-three per cent of the oocytes are expected to form intracellular ice and survive respectively when seeded at this temperature. The estimated frequency distribution of the threshold seeding temperatures at which individual oocytes form intracellular ice [f(TII)] is shown in Figure 6Go. Similarly the estimated frequency distribution of the threshold seeding temperatures when individual oocytes suffer irreversible damage [f(TID)] is also shown in Figure 6Go. A prominent feature of both distributions is the variability of the two parameters over a wide range of extracellular ice seeding temperatures. A few oocytes are very sensitive to the extracellular ice seeding temperatures and undergo IIF. As a result, these sensitive oocytes suffer irreversible damage at relatively high extracellular ice seeding temperatures (about -4.5°C). Alternatively, a few oocytes are very insensitive to the extracellular ice seeding temperature and do not undergo IIF. These oocytes suffer irreversible damage at only relatively low extracellular ice seeding temperatures (below -8°C). From a practical point of view, these results show that the initiation of the freezing protocol by seeding the extracellular ice at -8°C results in the loss of many human oocytes. On the other hand, seeding extracellular ice at -6°C will result in a significantly higher proportion of oocytes surviving. Nevertheless, particularly sensitive oocytes will still be lost. We thus propose that when human oocytes are frozen, the extracellular ice seeding temperature should be raised to a window between -4°C and -4.5°C to minimize the occurrence of IIF and save the more sensitive oocytes (Figure 6Go). The effects of seeding temperature on the survival of human and mouse oocytes are noticeably different. The cumulative incidence of survival of mouse embryos decreases steeply only when the seeding temperature is lowered below -7°C (Whittingham, 1977Go) or -8°C (Miyamoto, 1981Go). Thus, the acceptable window for extracellular ice seeding in the mouse is somewhat wider (3–4°C) than that in the human, allowing much more flexibility in the choice of seeding temperatures.



View larger version (9K):
[in this window]
[in a new window]
 
Figure 6. Frequency distributions of the incidence of human oocyte intracellular ice formation (IIF) and irreversible damage 24 h post-thaw on seeding temperature.

 
The cause of the high variability associated with the effects of extracellular ice seeding temperatures on the incidence of IIF and irreversible cell damage in human oocytes is unknown. In making measurements of the membrane water permeability and inactive volume of human oocytes, (Hunter et al., 1992Go), encountered very high variability relative to that seen in the mouse. They also were unable to identify the cause of the variability. Finally, these distribution curves are obtained from a mixed population of human oocytes. Further research to determine the threshold distributions is required using more specific populations of human oocytes, such as immature oocytes at the germinal vesicle stage and polyspermic oocytes. Further research is also needed on the fertilization and developmental potential of thawed oocytes produced after using our cryopreservation protocol, the incidence of polyploidy and the effects on the microfilament network which may affect the osmotic behaviour of oocytes (see Al-Hasani et al., 1987Go; Pickering et al., 1988Go; Le Gal et al., 1994Go; Van Blerkom and Davis, 1994Go; George et al., 1996Go).

Quantitative differences between species in the effects of cryoprotectants
When human oocytes were frozen at -120°C/min in the presence of cryoprotectant, the TMED was only slightly depressed (~6.5°C) (Figure 1Go). The fall in the IIF temperature was far less than that which occurs in mouse oocytes (~23°C; Table IGo) (Karlsson et al., 1996Go). The reason for these differences may be related to the 3-fold difference in volume between the small mouse oocyte and the large human oocyte. Consistent with this possibility, a comparable relatively small depression of IIF (~4°C) produced by glycerol was observed in the large cow oocyte (Myers et al., 1987Go). In contrast the published literature on the nucleation temperature of ice in water droplets of diameters comparable to those of mouse and human oocytes (80–129 µm) does not support this explanation, because of a very weak measured dependence of nucleation on droplet volume. However, it is important to note that these water droplet studies are performed under either homogeneous or relatively ineffective heterogeneous nucleation conditions at temperatures below -30°C [review: Hobbs (1974)]. On the other hand, human oocytes underwent IIF at much higher temperatures, indicating a very effective heterogeneous catalysis of IIF. In fact, a model based on the modified classical nucleation theory predicts that the likelihood of IIF in mouse oocytes is a strong function of the surface area of the cell (thus, cell volume) under consideration (Toner, 1993Go). Other cell-specific factors may also be involved since the depression of the temperature for IIF produced by DMSO in several cell types from different species was not correlated with volume (Hubel et al., 1991Go). Further detailed studies are required to explain the observed interspecies differences.

The large difference between human and mouse oocytes in the range of temperatures when IIF occurs should arouse caution in extrapolating protocols optimized in the mouse for use in the human. In a typical freezing protocol with a freezing rate of <2°C/min, most mouse oocytes will loose 90% of their water before reaching the critical ice nucleation temperature zone between -30°C and -40°C, and will thus avoid lethal IIF (Mazur, 1977Go). In the human oocyte, we have shown that lowering the rate of freezing to -0.2°C/min is not successful in preventing IIF, presumably because the cell has not dehydrated sufficiently when it enters the critical IIF temperature zone. The insufficient dehydration is probably related to the remarkably higher ice nucleation temperatures observed in the human oocytes. More specifically, extracellular ice seeding at -6°C and -8°C, customary in most freezing protocols, triggers IIF in a significant number of human oocytes during the holding period (Table IGo). Our results have shown that the undesirable IIF can be prevented by increasing the temperature of extracellular ice seeding as close as possible to the melting point of the solution, which in our case was -4.5°C. Presumably the higher extracellular ice seeding temperature allows the human oocyte to dehydrate extensively prior to reaching the critical temperature zone of IIF, thereby preventing the occurrence of IIF. In our protocol, a freezing rate of -0.2°C/min down to -15°C was empirically adopted to guarantee near equilibrium conditions between cytoplasmic and extracellular water as the oocyte approaches the IIF temperature range so as to ensure maximal dehydration. Thereafter, the freezing rate was increased to -1°C/min to reduce the exposure time to the potentially damaging extracellular electrolyte and solute concentrations which develop during freezing (Karlsson et al., 1996Go). Although this modified two-step cooling protocol changed both the extracellular seeding temperature and the cooling rate compared to the conventional protocol, the total duration of the protocol between the seeding of the extracellular ice and the plunge temperature into liquid nitrogen at -40°C is similar for all protocols (~80 min). The rationale for constant protocol duration is the strong association observed between protocol duration and mouse oocyte viability in our earlier studies (Figure 5Go of Karlsson et al., 1996Go). Since these settings were arbitrarily determined, this protocol could be further optimized by using the theoretical model described elsewhere (Toner et al., 1991Go; Karlsson et al., 1996Go).

Seeding temperatures and IIF in other cells
The effects of extracellular ice seeding temperature and cooling rate on IIF has been theoretically investigated with several mammalian cells, including red blood cells (Diller, 1975Go), granulocytes (Schwartz and Diller, 1984Go), hepatocytes (Toner et al., 1992Go) and mouse oocytes (Toner et al., 1993Go). All of these studies clearly indicate strong dependence of IIF behaviour of mammalian cells on the extracellular ice seeding temperatures. Typically, for faster cooling rates, the cumulative frequency of IIF increases for a given extracellular ice seeding temperature because less time is available for the cells to dehydrate at faster cooling rates. These studies underscore the important role of the extracellular ice seeding temperature on the fate of cells with respect to undergoing IIF. In the case of human oocytes, minimal depression of IIF in the presence of cryoprotectants requires using both a very slow cooling rate and a high extracellular ice seeding temperature to afford enough cellular dehydration to minimize (or prevent) IIF.

Are spare human oocytes useful in research?
There should always be concern about whether reliable conclusions can be reached when spare human oocytes and embryos are used for research (see Winston et al., 1993Go). Two observations suggest that the cell membranes of the failed-fertilized and polyspermic oocytes we have used retained their semipermeable properties which play a key role in the formation of intracellular ice. Firstly, the TMED of spare human oocytes frozen with no cryoprotectant -8.5°C, (Figure 1Go) is only 4°C above the TMED observed when freshly isolated mouse oocytes were frozen without cryoprotectant. Secondly, there is a close inverse symmetry of the cumulative distributions of IIF and survivors when spare human oocytes are frozen (Figure 5Go).

To conclude, we found that human oocytes display minimal or no depression of ice nucleation temperatures irrespective of the cryoprotectant used, and thus human oocytes are potentially more prone to lethal IIF during a freeze–thaw cycle. Increasing the seeding temperature was found to provide a narrow window for human oocytes to dehydrate without incurring IIF. Modification of a freezing protocol accordingly was associated with excellent survival rates of human oocytes. These conclusions might be extrapolated to the cryopreservation of fresh mature human oocytes and to human pronucleate embryos.


    Acknowledgments
 
The authors thank all the members of the Division of Reproductive Medicine at the Brigham and Women's hospital for their support and technical assistance, particularly Drs J.Hill, M.D.Hornstein, B.N.Clarke, G.Adaniya and A.Nurredine. We thank Dr J.Karlsson for teaching the use of the cryobiology equipment, and Dr A.Eroglu for editing the pictures. We are also grateful to Drs J.Baltz and S.Leibo for helpful discussions, and to Drs Catherine Racowsky and Betsey Williams for helpful criticism of the manuscript. This project was funded by the American Society of Reproductive Medicine/Mead Johnson/Bristol–Myers Squibb research grant, and by the American College of Obstetrics and Gynecology/Mead Johnson/Bristol–Myers Squibb fellowship award. The cryomicroscope system was purchased by a grant to J.D.B. from the Milton Fund, Harvard University.


    Notes
 
4 Present address: Trad Hospital, 53 Mexico Street, Beirut, Lebanon Back

5 To whom correspondence should be addressed Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 References
 
Al-Hasani, S., Diedrich, K., Van der Ven, H. et al. (1987) Cryopreservation of human oocytes. Hum. Reprod., 2, 695–700.[Abstract]

Bernard, A. and Fuller, B.J. (1996) Cryopreservation of human oocytes: a review of current problems and perspectives. Hum. Reprod. Update, 2, 193–207.[Abstract]

Bernard, A., Hunter, J.E., Fuller, B.J. et al. (1992) Fertilization and embryonic development of human oocytes after cooling. Hum. Reprod. 7, 1447–1450.[Abstract]

Chen, C. (1988) Pregnancies after human oocyte cryopreservation. Ann. NY Acad. Sci., 541, 541–549.[ISI][Medline]

Cosman, M.D., Toner, M., Kandel, J. et al. (1989) An integrated cryomicroscopy system. Cryo-lett., 10, 17–38.[ISI]

Diller, K.R. (1975) Intracellular freezing: effect of extracellular supercooling. Cryobiology, 21, 480–485.

Finney, D.J. (1947) The principles of biological assay. J. R.. Statist. Soc., 9 (Suppl.), 46–91.

George, M.A., Pickering, S.J., Braude, S.J. et al. (1996) The distribution of {alpha}- and ß-tubulin in fresh and aged human and mouse oocytes exposed to cryoprotectant. Mol. Hum. Reprod., 2, 445–456.[Abstract]

Gook, D.A., Osborn, S.M., Bourne, H. et al. (1994) Fertilization of human oocytes following cryopreservation: normal karyotypes and absence of stray chromosomes. Hum. Reprod., 9, 684–691.[Abstract]

Gook, D.A., Schiewe, M.C., Osborn, S.M. et al. (1995) Intracytoplasmic injection and embryo development of human oocytes cryopreserved using 1,2-propanediol. Hum. Reprod., 10, 2637–2641.[Abstract]

Hobbs, P.V. (1974) Ice Physics. Clarendon Press, Oxford.

Hubel, A., Toner, M., Cravalho, E.G. et al. (1991) Intracellular ice formation during the freezing of hepatocytes cultured in double collagen gel. Biotechnol. Prog., 7, 554–559.[ISI][Medline]

Hunter, J.E., Bernard, A.G., Fuller, B.J. et al. (1991) Fertilization and development of the human oocyte following exposure to cryoprotectants, low temperatures and cryopreservation. Hum. Reprod., 6, 1460–1465.[Abstract]

Hunter, J.E., Bernard, A.G., Fuller, B.J. et al. (1992) Measurement of membrane water permeability (LP) and its temperature dependence (activation energy) in human fresh and failed to fertilize oocytes and mouse oocytes. Cryobiology, 29, 240–249.[ISI][Medline]

Jackson, K.V. and Kiessling, A.A. (1989) Fertilization and cleavage of mouse oocytes exposed to the conditions of human oocyte retrieval for in vitro fertilization. Fertil. Steril., 51, 675–681.[ISI][Medline]

Kafadar, K. (1985) Notched-box-and-whisker plot. In Encyclopedia of Statistical Sciences, Vol. 6. Wiley, New York, pp. 367–370.

Karlsson, J.O., Eroglu, A., Toth, T.L. et al. (1996) Fertilization and development of mouse oocytes cryopreserved using a theoretically optimized protocol. Hum. Reprod., 11, 1296–1305.[Abstract]

Kazem, R., Thompson, L.A., Srikantharajah, A. et al. (1995) Cryopreservation of human oocytes and fertilization by two techniques: in vitro fertilization and intracytoplasmic sperm injection. Hum. Reprod., 10, 2650–2654.[Abstract]

Le Gal, F., Gasqui, P. and Renard, J.P. (1994) Differential osmotic behavior of mammalian oocytes before and after maturation; a quantitative analysis using goat oocytes as a model. Cryobiology, 31, 154–170.[ISI][Medline]

Leibo, S.P. (1980) Water permeability and its activation energy of fertilized and unfertilized mouse ova. J. Memb. Biol., 53, 179–188.[ISI][Medline]

Leibo, S.P. (1984) A one step method for direct non-surgical transfer of frozen–thawed bovine embryos. Theriogenology, 21, 767–790.[ISI]

Malik, H.J. (1985) Logistic distribution. In Encyclopedia of Statistical Sciences, Vol. 5. Wiley, New York, pp. 123–128.

Mazur, P. (1977) The role of intracellular freezing in the death of cells cooled at supraoptimal rates. Cryobiology, 14, 251–272.[ISI][Medline]

Mazur, P. (1984) Freezing of living cells: mechanisms and implications. Am. J. Physiol., 143, C125–C142.

Miyamoto, H. and Ishibashi, T. (1981) Effects of the temperature ice-seeding on survival of frozen-and-thawed mouse morulae. Experientia, 37, 187–18.[ISI][Medline]

Myers, S.P., Lin, T., Pitt, R.E. and Steponkus, P.L. (1987) Cryobehaviour of immature bovine oocytes. Cryo-lett., 8, 260–275.[ISI]

Nowrouzi, K., Baker, A.F., Check, J.H. et al. (1991) Cryopreservation using a shorter cooling program and a single step removal of cryoprotectant during thawing. Abstract of the 7th annual meeting of the ESHRE and the 7th World Congress on IVF and Assisted Procreation, pp. 221.

Oktay, K., Newton, H., Aubard, Y., Salha, O. et al. (1988) Cryopreservation of immature human oocytes and ovarian tissue: an emerging technology? Fertil. Steril., 69, 1–7.

Pensis, M., Loumaye, E. and Psalti, I. (1989) Screening of conditions for rapid freezing of human oocytes: preliminary study toward their cryopreservation. Fertil.Steril., 52, 787–794.[ISI][Medline]

Pickering, S., Johnson, M., Braude, P. et al. (1988) Cytoskeletal organization in fresh, aged and spontaneously activated human oocytes. Hum. Reprod., 3, 978–989.[Abstract]

Pirie, W. (1983) Jonckheere tests for alternative alternatives. In Encyclopedia of Statistical Sciences, Vol. 4. John Wiley, New York, pp. 315–318.

Porcu, E., Fabbri, R., Seracchioli, R. et al. (1997) Birth of a healthy female after intracytoplasmic injection of cryopreserved human oocytes. Fertil. Steril., 68, 724–726.[ISI][Medline]

Punjabi, U., Gerris, J., van Bijlen, J. et al. (1990) Comparison between different pre-treatment techniques for sperm recovery prior to IVF and GIFT. Hum. Reprod., 5, 75–83.[Abstract]

Schwartz, G.J. and Diller, K.R. (1984) Intracellular freezing of human granulocytes. Cryobiology, 21, 654–660.[ISI][Medline]

Toner, M. (1993) Nucleaction of ice crystals inside biological cells. In Steponkus, P. (ed.), Advances in Low-Temperature Biology. JAI Press, London. Vol. 2, pp. 1–51.

Toner, M., Cravalho, E.G. and Armant, D.R. (1990a) Water transport and estimated transmembrane potential during the freezing of mouse oocytes. J. Memb. Biol., 115, 261–272.[ISI][Medline]

Toner, M., Cravalho, E.G. and Karel, M. (1990b) Thermodynamics and kinetics of intracellular ice formation during freezing biological cells J. Appl. Phys., 67, 1582–1593. [Erratum: J. Appl. Phys., 70, 4653.][ISI]

Toner, M., Cravalho, E.G., Karel, M. et al. (1991) Cryomicroscopic analysis of intracellular ice formation during freezing of mouse oocytes without cryoadditives. Cryobiology, 28, 55–71.[ISI][Medline]

Toner, M., Tompkins, R.G., Cravallo, E.G. et al. (1992) Transport phenomena during freezing of isolated hepatocytes. Amer. Inst. Chem. Engng J., 38, 1512–1522.

Toner, M., Cravallo, E.G. and Karel, M. (1993) Cellular response of mouse oocytes to freezing stress: prediction of intracellular ice formation. J. Biomech. Engng, 115, 169–174.

Toth, T.L., Baka, S.G., Veeck, L.L. et al. (1994) Fertilization and in vitro development of cryopreserved human prophase I oocytes. Fertil. Steril., 61, 891–894.[ISI][Medline]

Tucker, M., Wright, G., Morton, P. et al. (1996) Preliminary experience with human oocyte cryopreservation using 1,2-propanediol and sucrose. Hum. Reprod., 11, 1513–1515.[Abstract/Free Full Text]

Van Blerkom, J. and Davis, P.W. (1994) Cytogenetic, cellular, and developmental consequences of cryopreservation of immature and mature mouse and human oocytes. Microsc. Res. Technique, 27, 165–193.[ISI][Medline]

Van Uem, J.F.H.M., Siebzehnrubl, E.R., Schuh B. et al. (1987) Birth after cryopreservation of unfertilized oocytes. Lancet, i, 752–753.

Whittingham, D.G., (1977) Some factors affecting embryo storage in laboratory animals. In The Freezing of Mammalian Embryos, Ciba Symposium 52 (new series). Elsevier, New York, pp. 97–127.

Winston, N.J., Braude, P.R. and Johnson, M.H. (1993) Are failed-fertilized human oocyte useful? Hum. Reprod., 8, 503–507.[ISI][Medline]

Younis, A.I., Toner, M., Albertini, D.F. et al. (1996) Cryobiology of non-human primate oocytes Hum. Reprod., 11, 156–165.

Submitted on July 17, 1998; accepted on January 29, 1999.