Sperm cryopreservation is routinely used presently in the management of human male infertility (Holt, 1997; Donnelly et al., 2001b). Despite this, the current cryo-techniques used for human spermatozoa are still imperfect. To date, nearly all cryobiological investigations on spermatozoa or routine freezing involve the use of conventional (programmable or standard vapour) freezing. The effectiveness of the cryo-technique is associated with permeable and non-permeable cryoprotectants. These are used to prevent the formation of ice crystals during freezing and, thus, avoid structural damage and motility loss after cryopreservation.
The decline in spermatozoa motility after cryopreservation is a topic of current research since it is one of the factors that are first affected (Critser et al., 1987b; Watson, 1995). However, the mechanism through which motility is decreased is still unclear. This mechanism may be mechanical or of a physical–chemical aetiology. Permeable cryoprotectants play a leading role, while the non-permeable protective agents play a supporting role and, in most cases, cannot protect the cells in the absence of a permeable cryoprotectant. The properties of permeable cryoprotectants are directly related to osmotic and toxic damage with concurrent cell saturation before cooling (Sherman, 1973; Watson, 1979; Gao et al., 1997) and/or removal after thawing (Watson, 1995; Gao et al., 1995, 1997). In conventional freezing, mechanical cell injury could occur by rapid freezing leading to intracellular or extracellular ice crystal formation and signs of osmotic damage (Watson, 1995; Gao et al., 1995, 1997). Conventional freezing causes extensive chemical—physical damage to the extracellular and intracellular membranes of the sperm that are attributable to changes in the lipid phase transition and/or increased lipid peroxidation (Alvarez and Storey, 1992, 1993; Mossad et al., 1994) during cooling or after thawing, with the consequence of a decrease both in sperm velocity and in the percentage of motile spermatozoa (Critser et al., 1987a,b; Keel et al., 1987; Mossad et al., 1994; Watson, 1995; Leffler and Walters, 1996). It has been established that the production of reactive oxygen species leads to increased lipid peroxidation after cryopreservation (Alvarez and Storey, 1992) and is significantly associated with a loss of sperm motility (Aitken et al., 1989; O’Connell et al., 2002). As previously suggested (Alvarez and Storey, 1992, 1993; Chatterjee and Gagnon, 2001), the injury to human spermatozoa induced by cryopreservation mainly occurs during thawing. This damage could, at least in part, be related to reduced antioxidant defence activity during cooling and/or structural damage to the cytoskeleton and/or antioxidant enzymes during cryopreservation (Alvarez and Storey, 1992, 1993). All these findings suggest that the slow freezing of sperm, aside from ice crystal formation, is intrinsically deleterious. Egg yolk, a natural complex mixture of cholesterol, phospholipids and antioxidants, has been used in sperm cryopreservation for many years to reduce the negative effects of osmotic shock. How this protective effect is produced is not entirely clear, since egg yolk is such a complex mixture, but it may play a role in reducing the deleterious effects of hyperosmotic salt solutions on membrane structures during rapid cooling (Watson, 1976, 1995; Ostashko, 1978, 1995; Pursel et al., 1978; Holt et al., 1988, 1992; Katkov et al., 1996). However, its most important role could be fortification of the cell membrane by the lipid components of the egg yolk (Ostashko, 1978). Some of these components (low-density lipoprotein fraction, glycolipids, cholesterol) may also become incorporated into the membranes, reducing their tendency to gel during cooling, as described for sperm and erythrocytes (Watson, 1976; Parks and Lynch, 1992; Ostashko 1978). Further, Chatterjee and Gagnon (2001) demonstrated that egg yolk–Tris–glycerol cryoprotectant medium (EYTG) is an efficient scavenger of NO–, O2– and H2O2 radicals. Thus, we may be dealing with a dual action whereby the cell membrane becomes coated and isolated from direct contact with the cryoprotective agents while preserving its fluidity and flexibility at a lower temperature, thereby reducing cytoskeletal damage. In a previous study (Isachenko and Nayudu, 1999), we showed that by combining the two factors, an increased temperature and the inclusion of egg yolk in both the vitrification and dilution media, the survival of mouse germinal vesicle oocytes was significantly improved after warming. The inclusion of egg yolk in the vitrification/dilution medium also improved the maturation rate and the proportion of normal metaphase forms, suggesting a major effect on reducing internal cell damage. However, the effectiveness (avoiding intracellular ice formation) of permeable and non-permeable cryoprotectants during conventional freezing is only revealed when we use a slow cooling rate (Gao et al., 1997). All these negative effects of freezing on cells can also lead to chromatin damage. The assessment of sperm nucleus integrity due to such causes is very important, since, as recently described (Spano et al., 1999; Sakkas and Tomlinson, 2000), chromatin abnormalities affect sperm quality and male fertility status. Fraga et al. (1991) correlated damaged sperm DNA with mutagenic effects. It has also been shown that, besides having significant effects on sperm morphology and membrane integrity, freezing/thawing the sperm of fertile and infertile men also leads to significant chromatin damage (Royere et al., 1988, 1991; Hammadeh et al., 1999; Donnelly et al., 2001a,b). Other studies have demonstrated (Balhorn et al., 1988; Manicardi et al., 1995) that any defects in sperm chromatin structure in infertile men with increased DNA instability are sensitive to denaturing stress. This denaturing stress may be induced by a treatment such as freezing. Despite all this, the oocyte is able to repair a small amount of sperm DNA damage (>8%; Ahmadi and Ng, 1999), though this repair seems to be insufficient to support subsequent embryo development (Ahmadi and Ng, 1999) and can lead to decreased conception rates or failed conception (Hunter, 1976; Royere et al., 1988). Proportions of spermatozoa with fragmented DNA have been negatively correlated with fertilization rates in IVF (Sun et al., 1997) and ICSI (Lopes et al., 1998).
In contrast to slow freezing, vitrification as a rule involves the use of very high concentrations (3.5–8 mol/l) of permeating cryoprotectants and high cooling rates (up to 1013°K/min). According to the literature, the critical cooling speed for the vitrification of pure water varies dramatically depending on the method used, from 107 to 1013°K/min (for references see Figure 9 in Karlsson and Cravalho, 1994). Given that high concentrations of cryoprotectants have a marked toxic effect (Fahy, 1986; Pegg and Diaper, 1988; Shaw et al., 2000), it is possible to decrease this toxicity using a combination of two cryoprotectants (e.g. ethylene glycol and DMSO), and/or to expose cells to pre-cooled concentrated solutions in a stepwise manner (Fahy et al., 1984; Fahy, 1986). Another strategy is to reduce the amount of cryoprotectant and simultaneously increase the cooling and warming rates (Liebermann and Tucker, 2002).
Luyet (1937) first mentioned the possibility of using the vitrification technique (a small specimen cooled very rapidly was vitrified without substantial loss of viability (for references see, for example, Fahy, 1988). The following year, Luyet and Hodapp (1938) reported the survival of frog spermatozoa vitrified in liquid nitrogen, and a few years later, Schaffner (1942) successfully vitrified fowl spermatozoa using a modification of Luyet technique. Nevertheless, all subsequent attempts to vitrify mammalian spermatozoa using this approach resulted in low or null survival (Hoagland and Pincus, 1942; Smith, 1961) mostly because of the critical speed of freezing and warming, which is very high for low concentrations of cyroprotectants. Such high speeds were unattainable by investigators at this time. Unfortunately, the high concentrations of cryoprotectants (30–50% compared with 5–7% for slow freezing) used in classic vitrification cannot be applied to spermatozoa due to their lethal osmotic effects (Holt, 1997; Katkov et al., 1998; Mazur et al., 2000). However, it has been established as dogma that the vitrification of large cells, tissues and even organs can only be achieved using high concentrations of combinations of permeable and impermeable cryoprotectants (Fahy, 1988). The total concentrations of such substances must be at least 50% (w/w) (if vitrification is conducted at atmospheric pressure) to reach the zone of stable vitrification. Concurrently, the speed of cooling and warming should be relatively low. These conditions can be very damaging for cells and lead to subsequent biochemical alterations and lethal osmotic injury (Fahy, 1984), although some of the deleterious effects of cryoprotectants on mammalian sperm can be avoided by adopting optimal regimes of addition and removal of the cryoprotectant (Sherman, 1973; Watson, 1979; Critser et al., 1988; Gao et al., 1995; Leffler and Walters, 1996; Katkov et al., 1998; Katkov, 2002). These regimes are, however, ineffective for human and animal spermatozoa treated with high concentrations of cryoprotectants. Thus, at present, the only alternative to this is the use of very rapid cooling and warming rates along with a very small specimen size. Such were the conditions used in the present study. The sample size can be minimized using different carrier systems [open-pulled straws (OPS; Vajta et al., 1997), flexipet-denuding pipette (FDP; Oberstein et al., 2001; Liebermann et al., 2002), micro drops (Papis et al., 2001), electron microscopy copper grids (Martino et al., 1996; Hong et al., 1999), hemi-straw system (Kuwayama and Kato, 2000), small nylon coils (Kurokawa et al., 1996), nylon mesh (Matsumoto et al., 2001) or cryoloop (Oberstein et al., 2001)] such that the duration of solidification of the liquid phase during freezing is reduced.
When the conditions of vitrification are: a very high speed of cooling [up to 720 000°K/min in the initial phase of cooling, as theoretically calculated (Isachenko et al., 2003)], a short cooling time (5–8 s) and a small specimen size (20 µl), not many nuclei of crystallization form and these seem to be insufficiently large to damage human spermatozoa. In these conditions, the probability of substantial devitrification (recrystallization) of the vitrified solution is also low due to the high speed and short time of warming (direct dissolution in a large volume of agitated warm water), and the small size of the specimen (extracellular recrystallization) and cells (intracellular recrystallization). The substantial compartmentalization of intracellular compounds may contribute to the successful survival of spermatozoa.
Moreover, it has been established that the amount of osmotically inactive water is higher in spermatozoa, where it is bound to several macromolecule structures such as DNA, histones, hyaluronidase, etc., than in oocytes or embryos. According to our calculations (see mathematical equation in Isachenko et al., 2003), amounts of high molecular weight components can be 6–8 times higher than in embryos and this will invariably affect the viscosity and glass transition temperature of the intracellular cytosol in sperm, but the probability of lethal ice formation during cooling/warming would be higher for embryos. There is indirect evidence to show that the intracellular components of sperm may act as natural cryoprotectants, including the fact that mice spermatozoa, among the most osmotically fragile of all species (Katkov et al., 1998), can be frozen successfully in the absence of permeable cryoprotectants, using only protein- and sugar-rich skimmed milk and raffinose as extracellular non-permeable cryoprotectants (Yokoyama et al., 1990; Nakagata and Takeshima, 1992; Koshimoto et al., 2000). Our estimations for albumin indicate that, in general, during cooling and especially during rewarming/resuscitation, the small amount of specimen and cells, high viscosity of the solution and high speed of cooling and warming (Isachenko et al., 2003) used would avoid devitrification (especially intracellular) (Karlsson and Cravalho, 1994; Karlsson, 2001). Seemingly, the presence of relatively high concentrations of albumin, which is highly efficient at inhibiting lipid peroxidation (Karow, 1997), and sugars substantially raised the viscosity of the solution, especially at the lower temperatures. In addition, the small sample size and number of cells made vitrification stable during both cooling and warming, leading to good results after warming. It would be logical to suppose that sperm DNA would be damaged using such an extreme cryo-protocol for sperm preservation. However, Evenson et al. (1991) found no difference in their SCSA results for non-cryopreserved or cryopreserved sperm, and slowly or flash-frozen sperm. Duty et al. (2002) confirmed these findings and reported that flash freezing in liquid nitrogen with no cryoprotectants most closely reproduced the results of a freshly obtained human ejaculate. These authors assumed that the particular DNA packaging of human sperm protects the DNA from intracellular fluid shifts and ice crystal formation during cryopreservation. All these data were confirmed by our results.
In conclusion, our results show that the vitrification of human spermatozoa in the absence of conventional cryoprotectants is indeed feasible. The DNA integrity of vitrified sperm is comparable with that shown by standard slow-frozen/thawed spermatozoa, yet the method is quick and simple and does not require special cryobiological equipment.
We presently are conducting studies to determine: how the different stages of vitrification compare with those of slow freezing in terms of inducing lipid cell membrane peroxidation; and the roles played by the different cryoprotective (glycerol, ethylene glycol, DMSO) and supporting (egg yolk and human serum albumin) agents in preventing damage to the cytoskeleton and/or antioxidant enzymes in the presence of reactive oxygen species during slow freezing or vitrification.
The authors thank Ms Doris Peters, Ms Ingrid Orth and Ms Martina Becker for technical assistance.