|Home | About | Journals | Submit | Contact Us | Français|
Successful cryopreservation demands there be little or no intracellular ice. One procedure is classical slow equilibrium freezing, and it has been successful in many cases. However, for some important cell types, including some mammalian oocytes, it has not. For the latter, there are increasing attempts to cryopreserve them by vitrification. However, even if intracellular ice formation (IIF) is prevented during cooling, it can still occur during the warming of a vitrified sample. Here, we examine two aspects of this occurrence in mouse oocytes. One took place in oocytes that were partly dehydrated by an initial hold for 12 min at −25°C. They were then cooled rapidly to −70°C and warmed slowly, or they were warmed rapidly to intermediate temperatures and held. These oocytes underwent no IIF during cooling but blackened from IIF during warming. The blackening rate increased about 5-fold for each five-degree rise in temperature. Upon thawing, they were dead. The second aspect involved oocytes that had been vitrified by cooling to −196°C while suspended in a concentrated solution of cryoprotectants and warmed at rates ranging from 140°C/min to 3300°C/min. Survivals after warming at 140°C/min and 250°C/min were low (<30%). Survivals after warming at ≥2200°C/min were high (80%). When warmed slowly, they were killed, apparently by the recrystallization of previously formed small internal ice crystals. The similarities and differences in the consequences of the two types of freezing are discussed.
Many cell types can be successfully cryopreserved by classical slow equilibrium freezing. Examples include erythrocytes and most leucocytes, adult and embryonic stem cells, umbilical cord cells, most tissue culture lines, and (importantly) preimplantation embryos from more than 20 species of mammals. The first successful instance in the last group was with mouse embryos by Whittingham et al. in 1972 . By slow equilibrium freezing, we mean the cooling of cells at a low enough rate to ensure that the cells dehydrate osmotically during the cooling, thereby avoiding intracellular ice formation (IIF) . The required cooling rate is commonly around 1°C/min, but it can be lower or higher depending chiefly on the permeability of the cell to water, the temperature coefficient of that permeability, and the surface:volume ratio of the cell. Equations to model the interaction of these several factors were first published by Mazur .
But some cell types cannot be cryopreserved by slow equilibrium methods, or yield poor survivals. The common reasons for failure are high chill sensitivity, the inability to maintain an intracellular supercooled state, and the disruption of complex cell-cell interactions by extracellular ice (EIF). One important cell type that has been difficult to cryopreserve by slow freezing is the mammalian oocyte. Generalized high chill sensitvity is one cause. Another important one is that mature oocytes are locked in the metaphase of meiosis II; that is, the chromosomes are arrayed on the meiotic spindle. The spindle is composed of microtubulin material and, like microtubules in general, it becomes disaggregated by being cooled to temperatures near 0°C [4–6].
In the case of those cells and tissue types where equilibrium slow freezing yields inferior results, there are increasing efforts to achieve cryopreservation by vitrification procedures. The rationale is that these procedures avoid damage from EIF or IIF by avoiding or preventing them. They avoid damage from chilling injury by cooling the cells/tissues so rapidly that chilling injury does not have time to develop. In some instances, this vitrification approach has been successful where slow freezing has not. One example is the Drosophila embryo [7, 8]. In the case of mouse and human oocytes there seems to be a developing consensus that vitrification procedures yield better survivals in terms of percentages of fertilization and development to blastocyst than do slow freezing procedures [6, 9]. But still, with humans the results are not good. In the last 20 years, only 300 children have been born from frozen or vitrified oocytes .
The aim of vitrification procedures is to cool cells under conditions that convert their water and that of the surrounding medium into a noncrystalline glass or amorphous state. That requires high cooling rates and high concentrations of glass-inducing solutes in and around the cells. The two are reciprocally related in that the higher the cooling rate, the lower the required concentration of solute .
Equally important as the induction of the vitreous state during cooling is to maintain a noncrystalline state during the subsequent warming; that is, to prevent ice crystallization during warming. The same two reciprocal factors operate in the latter as in the former; namely, warming rate and solute concentration. However, the conditions for warming are even more stringent than those required during cooling.
When the temperature of the sample drops below that for homogeneous ice nucleation during cooling, stable ice nuclei form. However, the viscosity is rising to very high values, and that combined with high cooling rates essentially paralyzes the growth of these nuclei. The result is a highly viscous, mainly supercooled solution. Solutions have characteristic glass transformation temperatures (Tg) below which the viscous supercooled liquid converts to the glassy state. That transition is accompanied by a change in heat capacity. The viscosity is on the order of 1013 poises.
Three phenomena operate during warming. According to Mishima and Stanley , the first of these is the conversion of amorphous ice to ultraviscous water above Tg (−140°C for pure water). The second is devitrification. For the ultraviscous water it occurs at −120°C (TD). Devitrification is the conversion of this water to crystalline ice, a process that releases the latent heat of fusion, which can be detected and measured by calorimetry. At still higher temperatures, ice commonly undergoes a third phenomenon; that is, recrystallization. Recrystallization involves the transfer of the surface molecules of small ice crystals to large crystals, a transition driven by the greater surface free energy of the former. It is accompanied by the release of very little heat. Operationally, recrystallization is usually defined as the appearance of opacity in a previous clear sample.
The terminology in cryobiological papers on vitrification tends to be imprecise. Commonly, cells subjected to putative vitrification procedures (high solute concentrations/rapid cooling) are said to be vitrified, especially if the procedure has yielded high survivals. In fact, definitive evidence that a certain procedure has vitrified a sample are relatively rare, and substantive evidence that the cells themselves have vitrified is essentially nonexistent. Definitive evidence for vitrification is provided by a lack of an hexagonal ice x-ray diffraction pattern or by the absence of an exotherm during cooling in a differential scanning calorimeter (DSC). A number of such determinations have been made for cryobiological media ; however, little evidence is available for cells, in part because of the difficulty in separating events in the medium from events within the cell.
A major purpose of the studies reported in this paper was the elucidation of recrystallization events in mouse oocytes. The experiments were made possible by the discovery of what we refer to as class 4 mouse oocytes. These are oocytes that do not exhibit intracellular ice formation during cooling, but turn black with temperature-related kinetics from ice formation during warming [13, 14]. Class 4 oocytes are produced when oocytes in 1.5 M ethylene glycol (EG) solutions are held at −25°C for 10–12 min and then cooled rapidly to −70°C. Temperature −25°C is above the intracellular ice nucleation temperature of the oocytes in this medium ; consequently, the cytoplasm is supercooled, and a 12-min hold causes them to shrink to a water content of 36% of isotonic. We determined that when such oocytes were warmed from −70°C at 10°C/min, they began to turn black at −56°C and completed blackening at −46°C [13, 14]. When they were warmed to and held at temperatures between −65°C and −50°C, they turned black with time, and the rate of blackening rose rapidly with increased temperature . That permitted a calculation of the activation energy of the blackening process; namely, 27 kcal/mole.
In the course of studies over the past 4 yr, we have made use of oocytes vitrified in an EG-acetamide-Ficoll-sucrose (EAFS) medium in the laboratories of Drs. Magosaburo Kasai and Keisuke Edashige at Kochi University (Kochi, Japan). A second major purpose of this paper was to compare the behavior of these “vitrified” oocytes with the behavior expected on the basis of our experiments on class 4 oocytes suspended in EG.
Many of the methods were described in detail in Mazur et al. [13, 15] and Seki and Mazur ; consequently, here we give details only for those aspects that differed. The procedures for obtaining and manipulating the mouse oocytes were carried out under the University of Tennessee IACUC protocol 911-0607, approved 28 June 2007.
The MII oocytes used in this study were collected from ICR female mice and vitrified by the lead author and others in Dr. Edashige's and Dr. Kasai's laboratories. They were express shipped to us in Tennessee at −196°C in liquid nitrogen (LN2).
The vitrification solution (EAFS10/10) consists of 10% (v/v) ethylene glycol and 10% (w/v) acetamide dissolved in a stock consisting of 30% (w/v) Ficoll 70 and 0.5 M sucrose in PB1 medium. The final concentrations of sucrose and Ficoll are 0.4 M and 24% (w/v), respectively. Table 1 gives the mass composition of EAFS10/10 in units more appropriate for use with phase diagrams and physical chemical analysis. The oocytes are highly permeable to EG and acetamide . They are impermeable to the other solutes. The calculation of the mass concentrations required a determination of the densities of the stock solution and the full EAFS solution. The values were 1.1698 and 1.1521 g/ml, respectively.
The following solutions were successively aspirated into 0.25-ml straws (IMV Technologies, L'Aigle, France): a 60-mm column of 0.5 M sucrose, a 20-mm air bubble, a 5-mm column of EAFS10/10, a 5-mm air bubble, and a 12-mm column of EAFS10/10 to contain the oocytes. The oocytes were first placed for 5 min in PB1 containing 10% (v/v) EG and then washed by transfer to two successive drops of EAFS10/10 at 25°C. Three (generally) were then transferred by micropipette into the 12-mm column of EAFS10/10 in each straw, and the end of the straw opposite the polyvinyl alcohol plug was heat sealed. Two minutes after introducing the oocytes into the EAFS10/10 in the straws, the straw was placed for 3 min or more in liquid nitrogen (LN2) vapor at −120°C to −150°C and then into LN2.
For our studies on EAFS-vitrified oocytes, the straws from Japan were used directly. For our studies on oocytes in ethylene glycol solutions, the contents of the straws from Japan were thawed by immersion in room temperature water after a 10-sec exposure to room temperature air. The “thawed” oocytes were then incubated for 2–3 h in M16 medium. Those with normal morphology (about 85%) were used in subsequent experiments.
The composition of the three EG/PBS solutions used (1.5, 2.0, and 2.5 M EG) are given in Table 2. The computations were based on NaCl with the same molality as Dulbecco PBS. For reference purposes, the table also includes the composition of 3.0 M EG/NaCl, although it was not used.
A major purpose of this study was to determine the effect of warming rate on the survival of EAFS-vitrified mouse oocytes. This was achieved by subjecting them to the 10 different warming protocols listed in Table 3. Protocol 9 is the standard procedure used in Kasai's and Edashige's laboratory and in our laboratory. Although the literature well describes the procedures for warming vitrified mouse oocytes, it does not give the actual numerical rates produced by these procedures. We wished to determine them for three reasons. First was to fill in the gap. Second, we needed to know quantitative rates to carry out some of our analyses. Third, we thought it important to determine how small variations in procedure would affect those rates, and in turn affect the viability of the vitrified oocytes.
To determine the warming rates with each procedure, a 36-gauge copper-constantan thermocouple was inserted into a standard IMV 0.25-ml straw containing a 50-mm column of EAFS solution. The thermocouple was attached to a Fluke (2190A) Digital Thermometer (Seattle, WA). It outputs about three temperature readings per second (the exact output rate was determined by stopwatch for each run). The outputs were fed into a PC computer with Windows XP (Microsoft) using the Hyperterminal program provided, and were transferred to the computer disk. Changes in temperature vs. time were recorded from −196°C to about + 5°C, but the warming rates listed in the right-hand column of Table 3 were based on the time taken to traverse the range of −70°C to −35°C. The reasons for this choice will become apparent later. Each warming procedure was repeated four to five times.
Warming procedures 5–10 all involved removing the straw from LN2, holding it in room temperature air for 10 sec, and then abruptly immersing it in a water bath. The differences among the procedures were the temperature of the water bath and whether or not the straw was stirred. In warming procedures 1 and 2, the 0.25-ml straws were removed from LN2 and steadily held in 23°C air (P-1) or fanned in the air (P-2) until the contents warmed above 0°C. Procedure 3 was like procedure 8, except that during the 10-sec period in air, the 0.25-ml straw was inserted into an IMV 0.5-ml straw that had been precooled in LN2, and the assembly was then immersed in 25°C water with stirring. In procedure 4, the 0.25-ml straw was removed from LN2, immersed for ~1 sec in mineral oil at 25°C, and then immersed in a 25°C water bath.
The protocols used to generate the 10 warming rates in Table 3 were then used to warm the sealed straws containing the oocytes shipped from Japan.
As soon as the crystallized vitrification solution melted, the straw was wiped dry, and its contents were expelled into a watch glass by flushing the straws with 0.8 ml of 0.5 M sucrose/PB1. The oocytes were then pipetted into fresh 0.5 M sucrose/PB1 in a watch glass. Approximately 10 min later, they were transferred to fresh PB1 medium lacking sucrose, and then transferred to and cultured in modified M16 medium for 2 h.
Viability was assessed at three time points based on osmotic responsiveness and morphological normality. First, the oocytes were examined during the 10 min in PB1/sucrose. Membrane-intact oocytes are expected to shrink with time because the sucrose is hypertonic. Second and third, they were examined after being placed in M16 and after 2 h of incubation. They fall into two binary groups. Degenerate oocytes are clearly nonviable. The others are indistinguishable from fresh oocytes, and we know from past experiments that the plasma membranes in this latter group are intact and function normally with respect to their osmotic response to hypertonic and hypotonic media, and with respect to their ability to remain supercooled in the presence of external ice. Still other criteria of normality are given in Mazur et al. [15, pp. 48–49]. Also, M. Kasai (personal communication) has found that 84% of ICR oocytes vitrified by the EAFS10/10 procedure and dezonated developed to the two-cell stage after in vitro fertilization, nearly equal to 89% for untreated controls.
EAFS-vitrified oocytes were warmed by protocol 9 and incubated in M16 medium for 2 h as described above. They were then placed in solutions of 1.5, 2.0, or 2.5 M EG/PBS plus the ice nucleating compound Snomax (York Snow Co., Victor, NY) for 15 min, after which they were transferred to a small droplet of that medium centered in the quartz sample holder of a Linkam BCS 196 cryostage (Linkam Scientific Instruments, Wakefield, UK). The oocytes were then cooled at 50°C/min to −25°C and held for 12 min. This was followed by cooling at 50°C/min to −70°C or −80°C. Finally, they were subjected to one of two warming procedures. In one procedure they were warmed at 10°C/min to −35°C, and then at 20°C/min until thawed. In the other procedure, they were warmed at 20°C/min to temperatures between −70°C and −55°C and held for various time periods before continuing the warming and thawing. In one case, they were held at −80°C in the Linkam cryostage for 36 h. Photographs were taken throughout the process. The degree of blackening of an oocyte at a specific temperature and time is assigned a score from 1 (barely detectable blackening) to 5 (maximal blackening). The rate of blackening at a given temperature is simply the degree of blackening divided by the elapsed time in minutes. Details and photographic illustrations of these procedures are given in Mazur et al.  and Seki and Mazur . The 12-min initial hold at −25°C produces class 4 oocytes predominantly; that is, oocytes that exhibit no IIF (blackening) during cooling but do so during warming. One experimental series involved holding EAFS-vitrified oocytes at −80°C for up to 168 h. They were held in a REVCO low-temperature freezer maintained at that temperature.
Error figures in tables and error bars in graphs are standard errors (standard deviations of the mean). Statistics were calculated by Microsoft Excel and by Instat v. 3.1 (GraphPad Software, San Diego, CA).
Figure 1 shows plots of the temperature of the EAFS10/10 solution in straws as a function of time for each of the 10 warming procedures. The two horizontal lines indicate the temperature range over which the warming rates were calculated (i.e., −70°C to −35°C). The six highest rates in that range were all produced by immersing the straws in a water bath and correspond to protocols 5–10. The highest rate was 3280°C/min; the lowest was 1900°C/min, a relatively small difference. The differences depended on the temperature of the water bath (0° to 37°C) and on whether or not the straws were stirred. The rate of 2950°C/min is the standard. The differences in the warming rates became more exaggerated at temperatures above −35°C, but we believe that rates above −35°C have relatively little biological effect.
These six procedures all included a 10-sec hold in air between removing them from LN2 and immersing them in water. Its purpose is to minimize cracking of the zona pellucida from large thermal stresses. During that hold, the temperature rises from −196°C to −110°C. The four protocols producing the lowest warming rates between −70°C and −35°C either involved holding the straws in air throughout or jacketing them with a coating of mineral oil or by inserting them in a larger straw. The warming rates with these four procedures (139°C to 1220°C/min) are distinctly lower than with the fastest six.
The mostly solid curve, the circle data points, and the left ordinate in Figure 2 show the effect of the warming rate between −70°C and −35°C on the survival of EAFS-vitrified oocytes. (The other two dashed curves will be discussed later.) Survival is low and variable (14% ± 11%) at the lowest warming rate (139°C/min) and rises with increasing rate to reach a plateau of 80% with warming rates of 2200°C/min and higher. The survivals at a warming rate of 1900°C/min were somewhat anomalous. The solid circle for that rate represents the mean survival of 30 oocytes from 10 straws (57% ± 13%). However, the survival in 3 of the 10 straws containing nine oocytes was 0%. If these three straws are excluded, the resulting survival represented by the open circle is 81.0% ± 5.6%.
Our hypothesis is that the low survivals at low warming rates result from more extensive recrystallization of intracellular ice at the low warming rates. It would be highly desirable to observe the degree of blackening of the EAFS-vitrified oocytes as a function of temperature, but unfortunately this cannot be done because the frozen EAFS solution is highly opaque. The alternative is to compute the degree of blackening of the oocytes as a function of temperature for each of the 10 warming rates in this study. The results of that computation are given in Figure 3. The curves were calculated using a QuickBASIC program based on the algorithm given by Seki and Mazur . The program makes use of the activation energy for blackening measured by these authors for oocytes in 1.5 M EG; namely, 27.5 kcal/mole. Take, for example, the curve for a warming rate of 251°C/min. The computed degree of blackening slowly rises above 0 as the temperature rises above −60°C. It then begins to accelerate, reaching a Black scale value of 1 at about −45°C and a value of 3 at about −41°C. The curve for the highest warming rate (3280°C/min) has the same shape, but a Black scale of 1 is not reached until −35°C, and a Black scale of 3 until −31°C.
From these computations, we can estimate the degree of blackening of oocytes in 1.5 M EG/PBS after warming from −70°C to −45°C (Fig. 2, bottom dashed curve) or to −35°C (Fig. 2, top dashed curve) as a function of the warming rate. The two curves are dramatically different. If we consider Black stage 1 as the maximum value consistent with viability, then after warming to −45°C, none of the warming rates except the slowest should be damaging. On the other hand, by −35°C, all the warming rates, with the possible exception of the highest, become lethal.
Because of its high measured activation energy, blackening rapidly accelerates with increasing temperature. But it must cease at some temperature well below 0°C. If its rate continued to increase to near 0°C, it would become impossible to warm a sample rapidly enough to prevent lethal recrystallization. For the moment, consider −45°C and −35°C as two candidates for the cessation temperature. We shall shortly discuss a possible physical basis for that choice.
As mentioned, the mostly solid curve in Figure 2 is the plot of the measured survival of EAFS-vitrified oocytes vs. the rate at which they were warmed. To what extent is this survival curve consistent with one or the other of the two (dashed) computed blackening curves for oocytes frozen in 1.5 M EG? The upper, −35°C, blackening curve predicts that the three lowest warming rates should be highly lethal; the survival curve shows low to moderate survival for these rates. This −35°C blackening curve predicts that even the four highest warming rates could be damaging, for they yield black values slightly greater than 1. The survival curve shows high survival at those rates. If we choose a cutoff temperature of −45°C, the bottom blackening curve predicts that damage will only occur at the lowest warming rate. However, the survival curve shows moderate to severe damage at the four lowest warming rates.
There is a 2-fold problem with this comparison. The blackening curves are computed from experimental data on oocytes in 1.5 M EG, and are based on activation energies derived from those data. There are no corresponding survival data. The data in the survival curve are from observations on EAFS-vitrified oocytes, but the opacity of the frozen solution makes it impossible to observe blackening and the rate at which it occurs. Consequently, it is not possible to obtain activation energies for the process in this medium.
As shown in Tables 1 and and2,2, the intracellular concentrations of CPA differ widely in EAFS and in 1.5 M EG. In EAFS, the intracellular molalities of EG + acetamide are 6.5 molal. In 1.5 M EG, the intracellular molality of CPA (EG) is 1.6 molal. We decided, therefore, to compare the recrystallization properties of oocytes in 1.5 M EG with their properties in 2.0 M EG and, to a lesser extent, 2.5 M EG.
Seki and Mazur  reported that when class 4 oocytes in 1.5 M EG are warmed at 10°C/min from −70°C, their blackening first becomes evident at −56°C and is fully developed by −46°C. (This also proved to be the case in a more recent experiment in which oocytes were cooled to and warmed from −150°C.) We have now repeated the previous 1.5 M EG experiments for oocytes in 2.0 and 2.5 M EG. In both concentrations, the onset temperature for blackening was −65°C, and the temperature of completion was −55°C. In other words, the initial and terminating blackening points occur 10°C lower in 2.0 and 2.5 M EG than in 1.5 M EG.
This large difference raised the question of whether the activation energies would differ in 1.5 M and 2 M EG. They do not. Figure 4 compares the Arrhenius plot for the rate of blackening vs. temperature for oocytes in 2 M EG with that for 1.5 M EG. The mean rates of darkening between Black 1 and 3 in 2 M EG were 3.24 ± 0.25 min−1, 0.83 ± 0.11 min−1, 0.71 ± 0.10 min−1, 0.36 ± 0.06 min−1, and 0.10 ± 0.02 min−1 at −62.5°C, −65°C, −67.5°C, −70°C, and −72.5°C, respectively. The mean rates of darkening in 1.5 M EG were 5.45 ± 1.35 min−1, 2.92 ± 0.79 min−1, 1.08 ± 0.13 min−1, 0.51 ± 0.15 min−1, 0.21 ± 0.03 min−1, 0.06 ± 0.02 min−1, and 0.0034 ± 0.0017 min−1 at −55°C, −57.5°C, −60°C, −62.5°C, −65°C, −70°C, and −80°C, respectively. This is a 1600-fold difference in rate over a 25°C span. The two lines are nearly parallel; that is, their slopes do not differ significantly. Consequently, the two activation energies are nearly identical (26.7 kcal/mole in 2 M EG vs. 25.0 kcal/mole in 1.5 M EG). What is surprising is that although parallel, the two lines are offset by nearly 2 natural log units, with the line for 2 M being above that for 1.5 M. Specifically, at a given temperature, the rate of oocyte recrystallization in 2 M EG occurs 5.5-fold more rapidly than in 1.5 M EG. This sizeable difference in rate accounts for the finding that in the experiments involving warming at a constant rate, the temperatures for the onset and completion of blackening were 10 degrees lower in 2.0 and 2.5 M EG than in 1.5 M EG. Note that the highest temperature plotted for 2 M EG in Figure 4 is −62.5°C, whereas for 1.5 M EG it is −55°C. That is because when oocytes in 2 M EG were warmed to above −62.5°C, considerable blackening had already occurred before they arrived at the hold temperature.
We performed a similar experiment, but with the oocytes equilibrated in a mixture of 1.5 M EG and 0.5 M glycerol. They are not permeable to glycerol, consequently, in this medium; the internal concentration of EG was 1.5 M, the external concentration of EG + glycerol was 2 M, and the cells were substantially shrunken because the external nonpermeating 0.5 M glycerol plus the external PBS makes the medium hypertonic. When class 4 oocytes that had been frozen in this medium were warmed at 10°C/min, blackening began at −55°C and reached a maximum at −45°C, the same as with the oocytes frozen and warmed after suspension in 1.5 M EG alone. It appears, then, that the lower onset temperature of blackening of cells in 2 and 2.5 M EG is associated with the higher concentration of EG in the cells and not with the higher concentration of CPA in the medium. Since oocytes in EG/PBS alone are at full isotonic volume (because of their high permeability to EG), whereas those in EG/glycerol/PBS are shrunken, the experiment also indicates that initial cell volume is not a factor in the recrystallization temperature.
In connection with Figure 3, we hypothesized that −45°C and −35°C might represent temperatures above which recrystallization ceases. The choice of −35°C was based on photographic observations like those in Figure 5A. They depict the change in darkness when the three class 4 oocytes in Figure 2 of Seki and Mazur , which had been held at −62.5°C for 30 min, were then warmed at 10°C/min to −35°C and at 20°C/min to higher temperatures. The photographs show that as warming continues above −50°C, the blackened oocytes begin to become detectably paler at about −35°C, and dramatically paler by −25°C. We presume that the lightening is a consequence of some melting of intracellular ice. From the ternary phase diagram for EG/NaCl/water , we can compute how much ice melts at given temperatures. As shown in Table 4, the figure is 12% for 1.5 M EG/PBS at −35°C. As elaborated on in the Discussion, it would seem incompatible for recrystallization and melting to occur simultaneously.
Table 4 shows that in 2.0 M EG, 12% of the ice will have melted by −45°C. Consequently, we expected that in 2 M EG, oocytes would undergo the first indication of lightening at −45°C, rather than at −35°C. The column of micrographs in Figure 5B does not confirm that expectation. Those oocytes also begin to lighten at about −35°C.
Figure 6 compares the measured degree of blackening of class 4 oocytes frozen in 1.5 M EG/PBS vs. time at −80°C with the survival of oocytes vitrified in EAFS10/10 after holding them at −80°C for 0–60 h. The survival of the EAFS-vitrified oocytes remains nearly constant at about 80% over that period; that is, the slope is not significantly different from zero. However, the blackening of those frozen in 1.5 M EG rises progressively. It exceeds Black 1 in about 4 h and exceeds Black 3 in 20 h. Our presumption is that blackening beyond Black 1 is lethal. An additional two straws with five oocytes that had been vitrified in EAFS were held at −80°C for 168 h. All five survived.
Whether or not ice forms in oocytes during cryopreservation is critical to whether or not the cells survive. Our thesis is that the better the understanding of the factors that lead to or prevent IIF, the greater the likelihood of devising improved protocols. For those implementing modified cryopreservation procedures, it is important to know which steps must be held to stringent tolerances and why. Our emphasis here has been to elucidate the factors that lead to lethal IIF with rapid cooling procedures, including vitrification. Most studies have emphasized the role of cooling rate, but we have found that warming rate also has a powerful influence on whether oocytes undergo IIF or not—an influence that may transcend that of the cooling rate. Rapid warming is essential to minimize both the formation of intracellular ice crystals by devitrification and their growth to lethal size by recrystallization. We believe these results and those of succeeding experiments will eventually lead to new or revised protocols for oocyte cryopreservation that yield oocytes with increased survival and high developmental potential.
Although our emphasis is on the fundamental cryobiological aspects of the problem, we would like to begin this discussion with two rather practical aspects.
Vitrification procedures subject the “vitrified” samples to warming at high rates. Figure 2 shows our determination of how the viability of EAFS-vitrified mouse oocytes is affected by the warming rate used; namely, survival progressively increases as the warming rate rises from 140°C/min to 2200°C/min.
Our standard method of rapid warming involves the abrupt transfer of 0.25-ml straws from LN2 to a water bath at room temperature (with a brief intermediate hold in room temperature air). As shown in Figure 1 and Table 1, it produces a warming rate of 2900°C/min between −70°C and −35°C. That warming rate falls well into the plateau region of Figure 2.
As just mentioned, the standard method subjects the straws to a brief intermediate hold in air (10 sec in our case) before immersing them in the water bath. An important purpose of this hold is to minimize cracking of the zona pellucida. As seen in Figure 1, that 10 sec in air results in the straws warming at ~500°C/min to −110°C before the plunge into water. The brief exposure to −110°C causes no problems, because recrystallization, if any, would take months at that temperature. However, if for any reason the samples were to be exposed to room temperature air for ≥15 sec, their temperature will rise to −70°C or higher, and recrystallization would occur at significant rates, with likely lethality.
When straws are rapidly warmed by immersion in water, the temperature of the water bath (0°C to 35°C) and whether or not the straws are stirred or not has little effect on the warming rate between −70°C and −35°C, although it does affect the warming rate above −35°C. Only the former region appears critical.
The viability of oocytes vitrified in EAFS remains high and stable for 60 h at −80°C (Fig. 6), and even after 168 h. Since the temperature of dry ice is −79°C, our data suggest that the shipment of vitrified oocytes in dry ice may be marginally feasible. However, that appears not to be the case with morulae from the same strain of mice. Jin et al.  have just published a paper showing that morulae vitrified in very similar media and in the same laboratory as the oocytes used here underwent a substantial drop in survival after 24 h at −80°C.
Our presumption is that the protection derived from rapid warming is due to its ability to suppress recrystallization; that is, blackening. Conversely, the low survivals associated with slower warming at 140°C/min reflect the occurrence of recrystallization. Unfortunately, we cannot test this hypothesis by direct experiment. First, the vitrification procedure subjects samples to a cooling rate of ≥1000°C/min (see Note Added in Proof). We cannot obtain close to any such rates with the Linkam—the maximum is 100°C/min. Second, EAFS10/10 solutions frozen in the Linkam are completely opaque, and one cannot see whether or when the oocytes blacken. Third, according to Figures 2 and and3,3, the prevention of recrystallization would require warming the vitrified oocytes at ≥1000°C/min. That is 10× or more higher than the maximum rate attainable in the Linkam. Because of these limitations, we have devoted substantial study to the recrystallization behavior of oocytes suspended in lower solute concentrations; namely, 1.5 M and 2 M EG, concentrations at which the recrytallization events can be visualized.
In Figure 2, we see that some 15% and 30% of the EAFS-vitrified oocytes survive when warmed at 140°C/min and 250°C/min. From Figure 3, we would have expected that higher warming rates would be required to minimize recrystallization. The answer to the discrepency could lie in the water contents of the oocytes prior to the initiation of vitrification. Their approximate water content is given by the ratio of the isotonic osmolality to the total osmolality of nonpermeating solutes in the external medium. From Table 1 we see that that ratio is 0.300:1.064, or 28.2% of isotonic. In other words, the water content of the oocytes in EAFS10/10 is reduced to 28.2% of normal. It is probably lower than that because the 20.7 wt.% Ficoll will exert a degree of noncolligative osmotic dehydration in spite of its high average molecular weight. In a previous publication , we showed that when the water content of oocytes is reduced to 19.7% of normal by a 30-min hold at −25°C, they do not undergo IIF when then cooled to −70°C or during subsequent warming. We refer to such oocytes as class 5. The water content of the oocytes in EAFS is reduced fairly close to that value prior to initiating vitrification.
In Figure 3, we plotted the computed degree of blackening as a function of subzero temperature in oocytes frozen in 1.5 M EG/PBS and warmed from −70°C at rates between 140°C/min and 3280°C/min; that is, the rates to which the EAFS-vitrified oocytes were subjected. The computations are based on the Arrhenius plot published by Seki and Mazur  for oocytes in 1.5 M EG. In all cases, the oocytes exceed Black 1 by the time they have warmed to −35°C. However, in the case of the highest warming rates (1900°C/min to 3280°C/min), the excess is small (Black 1.1 to 1.5). The significance of −35°C is that we consider it to be the temperature above which no further recrystallization occurs. This premise is based on the observation that the oocytes become noticeably paler by −35°C (Fig. 5A), presumably as a consequence of partial melting of the intracellular ice. Additional recrystallization ought to be blocked by that melting. Furthermore, the greatest driving force for recrystallization would come from dendritic tips with the smallest radii of curvature, because the smallest tips will have the highest surface free energies. But the very fact that they have the highest free energies will also cause their melting points to be suppressed below that of large crystals. Consequently, the fine-tipped dendrites would be the first to melt. This decrease in melting point becomes substantial (around 10°C) when the tip radius of curvature is in the low nanometer range [19, 20].
On the other hand, if the cessation of recrystallization were to occur 10 degrees lower at −45°C, the left-hand vertical dashed line in Figure 3 indicates that blackening will remain well below 1 at all warming rates of 476°C/min and higher. In other words, if Black 1 is the threshold for lethality and if the cutoff temperature for recrystallization is −35°C, then all the warming rates used for the vitrified samples should have been lethal. But if the cutoff temperature is −45°C, then only the warming rate of 139°C/min should have been lethal. If these curves apply to oocytes in EAFS, their response would lie between the two vertical lines but closer to the left-hand line for −45°C.
A priori, we would predict that the threshold temperature for the cessation of recrystallization would decrease with increasing EG concentration, because the higher the concentration, the lower the temperature at which a given degree of melting would occur. As oocytes and their surrounding medium warm, the ice in the oocytes (and in the outside medium) begins to melt when the temperature rises above the eutectic point (about −60°C for EG, but not determined with precision). The fraction of ice (U) that is melted at successively higher temperatures is computed as U = [(100 – WT)(W°T) / (WT)] / (100 – W°T). W°T is the total wt.% solutes prior to freezing as calculated for EG from the values in Table 2. WT is the total weight percent of EG + salt in the unfrozen fraction of the cell or medium at a given subzero temperature. It is obtained from the synthesized ternary-phase diagrams for EG/NaCl/water of Kleinhans and Mazur . Each concentration of EG has a unique weight percent ratio (R) of EG to isotonic salt (Table 2, column 11), and a unique phase diagram isopleth applies to that R value. Table 4 shows the computed temperatures at which 12% of the ice is melted when 1.5, 2.0, and 2.5 M EG/isotonic salt are warmed; namely, −35°C, −45°C, and −55°C, respectively. Consequently, we would expect that oocytes that have blackened by recrystallization in 2 M would become noticeably paler by −45°C (as opposed to −35°C in 1.5 M EG). However, as shown in Figure 5, we see no substantive differences between the degrees of blackness of the oocytes in column A (1.5 M) and those in column B (2.0 M).
Part of the difficulty in interpreting these sorts of data is that we have no way to quantitatively relate what we observe visually (degree of blackness) to the physical state of the liquid and ice within the oocyte. For example, the ternary-phase diagram data on EG/NaCl/water solutions say that at −35°C, 12% of the 1.5 M EG will have melted, whereas the corresponding value for 2 M EG is 16% melted. Yet, Figure 5 shows no significant difference in the appearance of the oocytes in the two media at these temperatures. This indicates that we cannot differentiate visually between the melting of 12% and 16% of the intracellular ice.
We need to emphasize that the application of phase data to the interior of the cell requires that the cells be in equilibrium with the surrounding solution and ice. In strict thermodynamic terms, that means that the chemical potentials of water and solutes inside the cell be equal to that of the ice, liquid water, and solutes surrounding the cell. In more operational terms, it means that the molal concentrations of EG and salt be the same inside and outside the cell, and the weight percent of liquid water be the same. As shown in Table 5, that is not the case for Class 4 oocytes at the end of the 12-min hold at −25°C; their water content at that point is 36% of the initial isotonic value, a percentage that is nearly double the equilibrium water content for that temperature, and they contain an internal concentration of EG that is less than half of the equilibrium concentration. If they do not undergo IIF during the subsequent cooling to −70°C (which class 4 oocytes do not, by definition), their fractional volume of water and concentration of EG will remain unchanged, because the 50°C/min cooling rate from −25°C to −70°C is too rapid to allow significant water loss. (If they do undergo full IIF during that cooling, the liquid water content drops to zero, and the EG concentration becomes infinite.)
In the case of the class 4 oocytes studied here, once cell blackening (recrystallization) begins and is completed during the subsequent warming (or warming + hold), their internal water contents and concentrations of EG move to the equilibrium values called for by the phase diagram. This occurs because the intracellular ice that forms during warming converts pure cell water into ice, and this in turn causes the concentration of internal EG to rise to very high values and the fraction of unfrozen water to fall to very low values.
In the case of oocytes vitrified in EAFS, if they truly remain unfrozen during the entire procedure, the cell water volumes and the intracellular concentrations of EG + acetamide will remain unchanged from those calculated from Table 1 for the initial suspension; namely, a cell water volume of 0.282 (0.3/1.064), and an intracellular concentration of EG + acetamide of 23.0 molal (6.496/0.282). That will not be true, of course, if IIF occurs at any stage.
Recrystallization cannot occur below Tg, the glass transformation temperature. At the end of the 12-min hold at −25°C, the osmotic shrinkage of the oocytes has increased the internal concentration of EG to a calculated 22 wt.%. That concentration remains essentially unchanged with further cooling because of the high cooling rate used. From the data of Luyet and Rasmussen  and Hayes and Pegg , we estimate the Tg for 22 wt.% EG/water to be −134°C. Actually, it is unlikely that recrystallization could occur below the devitrification temperature (TD), which we estimate from the same sources to be −125°C. The determination of the limiting lower temperature for recrystallization would be of fundamental interest, and it is experimentally testable. The only problem is that the experiment will require a modicum of patience. Assuming that recrystallization continues at temperatures below −80°C according to the kinetics of the lower Arrhenius plot in Figure 4, we calculate that class 4 oocytes in 1.5 M EG will require 358000 yr at TD and 86 million years at Tg to reach Black 1.
To our surprise, the blackening of class 4 oocytes frozen in 2.0 and 2.5 M EG/PBS and warmed at 10°C/min from −70°C was first detected at −65°C, a full 10°C below the onset temperature in 1.5 M EG. That difference is reflected in the Arrhenius plots for the 1.5 and 2 M EG in Figure 4. The two lines are parallel, indicating the same Ea for the blackening rate, but the loge of the rate for 2 M EG is nearly twice that of the rate for 1.5 M EG at any given temperature. Our presumption is that that would also be true for 2.5 M EG had we determined the Arrhenius plot for it. As of now, we have no explanation for the offset.
With one exception, the minimum temperatures in our EG experiments were −70°C to −80°C. The exception was two runs in which five oocytes were cooled to −150°C. The class 4 oocytes that we used exhibit no IIF (blackening) at those temperatures initially, even though they are 20°C to 30°C below the homogeneous ice nucleation temperature (Th ) in 22 wt.% EG/water. (Th is the temperature at which supercooled water spontaneously nucleates. Fahy et al.  estimate it to be −52°C for a 22 wt.% generic aqueous solution.) The minimum temperatures of −70°C to −80°C are also 10°C to 20°C below the approximate eutectic point for EG/water. In contrast, they are 55°C to 65°C above the estimated Tg for 22 wt.% EG in water, and are 45°C to 55°C above the estimated devitrification temperature for that concentration [21, 22]. Considering all these facts, our presumption is that at −70°C to −80°C, 1) ice nuclei have formed, 2) some ice growth may have occurred, but in amounts too small to be optically detectable as darkening, and 3) the remainder of the cytosol is a highly viscous supercooled liquid. It is not a glass, since −70°C to −80°C is well above Tg, but its viscosity is so high that ice crystal growth from the ice nuclei is prevented in the short term. The subsequent darkening during warming could represent recrystallization in the strict sense of a solid-state conversion of small crystals into large, or it could represent the growth of ice nuclei from surrounding supercooled liquid. The two possibilities would have different thermal signatures and in theory could be distinguished calorimetrically; that is, the second possibility would be accompanied by the release of the heat of fusion. The first would have very small thermal consequences.
One final point is that when the oocytes were cooled to and warmed from −150°C, cell blackening occurred over the same temperature range as when they were cooled to and warmed from −70°C; namely, −55°C to −46°C. A temperature of −150°C is well below Tg, so true vitrification should have occurred during cooling, and devitrification should have occurred during warming. Yet, the temperature range for blackening was unchanged, and no changes in or outside the cells were evident between −150°C and −55°C.
With respect to point number 2, Stott and Karlsson  have observed using high-speed cinemicrography at about 5000 Hz that the first optical indication of IIF in endothelial cells is the very rapid propagation (5.6 μm/msec) of a transparent ice front. This is followed a few milliseconds later by darkening of the cell, a slower process that appears equivalent to the blackening that we observe at a 100-fold lower temporal resolution. The fraction of the water in the cell that is converted to ice in the first step is unknown. Such a rapid wave of ice formation could possibly occur in our oocytes, but we would not be able to observe it at our much coarser temporal resolution.
Our experiments combined with those of Kingery  show that the recrystallization of ice occurs over a broad range of temperatures; namely, from at least −80°C to −2°C. It may even occur at lower temperatures. Its rate quintuples with every 5°C rise in temperature, reflecting an activation energy of some 27 kcal/mole. The same activation energy applies to macro samples of pure ice, and it applies to the ice in 1.5 and 2.0 M EG within oocytes. The upper temperature of −2°C might appear incompatible with our previous conclusion that recrystallization in oocytes ceases at −35°C or −45°C, but it is not. Kingery studied spheres of pure ice, which of course would undergo no melting at −2°C. We worked with ice in contact with EG solutions, which begin melting at much lower temperatures; that is, above the eutectic point.
Investigators such as Luyet and Rasmussen  and Rall et al.  and Rall and Polge  have reported that bulk solutions of EG/water and of glycerol/cytoplasm in eight-cell mouse embryos recrystallize at close to a fixed temperature of about −60°C. This apparently fixed temperature, we submit, is an artefactual consequence of their using slow warming rates, which confound the effects of temperature and time.
Recrystallization, thus, is an entirely kinetic process that is distinct from devitrification, which is essentially a phase change-like process occurring when the sample warms several degrees above Tg.
We know two facts. 1) Oocytes that show no evidence of blackening during either cooling or warming (class 5) are viable after thawing. 2) However, those oocytes that blacken during either cooling or warming or during isothemal holds are invariably lethally damaged after thawing (classes 1 through 4). These conclusions are based on observations of more than 1500 oocytes. Unfortunately, what we do not know, and know of no way to determine, is how much blackening is consistent with viability. Oocytes that blacken always proceed to a terminal Black 5 either during cooling or warming. We cannot stop them at, say, Black 1 and then thaw them with no further blackening. To do so would require an estimated warming rate of at least 2000°C/min, 20-fold higher than that attainable with the Linkam. Our guess, and it is just that, is that the oocytes will tolerate blackening up to Black 1 but not beyond. That is why we have used Black 1 as a limit in some of our figures and discussion.
As we have discussed, the water content of oocytes in EAFS10/10 is reduced to about 28% of the isotonic value prior to vitrification, primarily in osmotic response to the external hypertonic sucrose, and partly in response to the concentrated Ficoll. That means that the intracellular molalities of EG and acetamide are nearly four times the values shown in Table 1 for the EAFS medium; namely, 23 molal for EG + acetamide combined. (The other solutes in EAFS do not permeate.) And if we assume that the 23 molal were to apply to EG alone, then its wt.% within the cell (neglecting intracellular salts and other endogenous solutes) would be 59 wt.%. Luyet and Rasmussen  and Hayes and Pegg  report that Tg for that concentration of EG in water is −131°C. The two papers also indicate that the devitrification of a 60 wt.% EG solution occurs at approximately −78°C (an extrapolated value in Hayes and Pegg). Hayes and Pegg estimate that the cooling rate required to vitrify such a solution is <7°C/min. They calculate that the critical warming rates required to avoid the devitrification of 45 wt.% and 47 wt.% EG/water are 30000°C/min and 9000°C/min, respectively. Since the critical rate drops by a factor of three over that 2% change in concentration, one cannot really estimate the critical warming rate for a 59 wt.% intracellular solution of EG. Some 8 yr earlier, Fahy  quoted a much higher estimate of the critical warming rate to avoid the devitrification of 45 wt.% EG in water; namely, 140 million °C/min. But he indicated that was undoubtedly an overestimate.
To what extent can we interpret our experimental EAFS data in terms of the above findings and computations? First, it appears likely that the oocytes vitrify upon cooling, for the cooling rate used is higher than that computed to be critical. Second, it appears almost certain that the oocytes do not devitrify during the 10-sec warming in air to −110°C, since that is some 30°C below the estimated devitrification temperatures (TD) reported by Luyet and Rasmussen  and Hayes and Pegg . As for those oocytes held for many hours at −80°C, that temperature is right at the border of where devitrification is expected. However, if it occurs, it is not damaging (Fig. 6). As to whether recrystallization of intracellular ice occurs at higher temperatures, we cannot say for certain. Luyet and Rasmussen  report that it occurs at about −63°C in a 40 wt.% vitrified solutions of EG in water warmed at 5°C/min. (Shaw et al.  also reported a transition for vitrified EG/water at −60°C, but based on DSC exotherms, they labeled it devitrification and not recrystallization. (Why their value of TD is 30° higher than that reported by Luyet and Rasmussen  and Hayes and Pegg  is unclear.) Our experimental data on survival vs. warming rate suggest that recrystallization is occurring in EAFS-vitrifed oocytes warmed at <1000°C/min, oocytes that we calculate contain 59 wt.% EG + acetamide. But we cannot obtain microscope confirmation.
The general view is that a proper combination of high concentrations of CPA and high warming rates are required to prevent the devitrication and recrytallization of an intracellular glass formed during cooling. The thought is that their prevention is a matter of the interaction of viscosity and time. The higher the intracellular concentration of CPA and other solutes, the higher the viscosity and the slower the growth rate of any ice embryos or small ice crystals present.
We suggest a different picture; namely, 1) recrystallization ceases when the intracellular ice begins to melt; 2) the higher the intracellular concentration of CPA and other solutes, the lower will be the temperature at which that melting begins; 3) the progressive reduction in the upper temperature boundary permitting recrystallization means a progressive reduction in the maximum rate at which recrytallization can occur; and 4) the activation energy for recrystallization is so large that the maximum growth rate of ice will be greatly restricted with the lowering of the melting point. These points are illustrated in Figure 3, which shows that the lower the temperature at which a given degree of recrystallization is attained, the slower the warming can be. Hayes and Pegg  and Kleinhans and Mazur  report that the measured and calculated temperatures, respectively, for the complete melting of a 55 wt.% solution of EG in water and in PBS is −45°C and −48°C, respectively. On the basis of Figure 3, these figures mean that a warming rate of <140°C/min would suffice to prevent more than a trace amount of recrystallization in a vitrified solution of that composition.
Favoring the view that the reduction of recrystallzation with increased CPA is not a matter of increased viscosity is the observation that the activation energy for recrystallization is the same for pure water ice and for ice formed from 1.5 and 2.0 M EG over a broad range of temperatures (−72.5°C to −2°C), during which viscosities would undergo orders of magnitude change. Of course, as exemplified in Figure 4, the Ea for recrystallization can be the same in two different solutions, and yet the blackening rates at a given temperature can differ considerably.
Our suggestion (that a reduction in recrystallization associated with an increase in CPA concentration is due to a lowering of the temperature at which a given degree of melting occurs) gets no support from our inability to see any evidence for that in Figure 5. Yet, phase-diagram considerations say it must be so; that is, the higher the concentration of CPA, the lower the temperature at which a given fraction of the ice melts (Table 5). Of course, for phase-diagram data to apply, intracellular water, ice, and solutes must be in thermodynamic equilibrium. Presumably, those equilibria are approximated with a warming rate of 10°C/min.
A final point to keep in mind is that the interior of a cell is not just a solution of EG/salt or, in the case of EAFS, a solution of EG/acetamide/salt. Many other solutes are present, including high concentrations of proteins and nucleic acids. How their presence perturbs Tg and TD or, for that matter, the behavior of water itself is scarcely known.
We have recently measured the actual cooling rate achieved by Kasai's and Edashige's procedure. It is 5-fold lower than 1000°C/min; namely, 189°C/min.
We thank Prof. Keisuke Edashige, B. Jin, Y. Kawai, Y. Kobayashi, and M. Yoshimura of Kochi University (Kochi, Japan) for providing us with the oocytes used in this study, and we thank Prof. M. Kasai for providing us with unpublished data on the percentages of EAFS-vitrified oocytes that will fertilize and develop to the two-cell stage. The computations of the water contents and solute concentrations in the cells made use of Excel spreadsheets developed by Prof. F.W. Kleinhans, Department of Physics, Indiana University-Purdue University at Indianapolis for related research.
1Supported by National Institutes of Health grant R01-RR18470.