Human Ovarian Tissue

The structure of the human ovary is a crucial consideration in the potential success of cryopreservation. The human ovarian cortex is predominantly (>80%) populated with quiescent primordial stage follicles [14–16], each consisting of an oocyte (approximately one-third the diameter of a mature oocyte) surrounded by a single layer of flattened pre-granulosa cells. A much lower proportion of follicles will have initiated development (primary follicles) and in approximately 3% of follicles division of granulosa cells will have occurred (proliferating follicles). In contrast to the rodent ovary, in which much of the initial ovarian tissue cryopreservation was performed, more advanced follicles characterized by at least two layers of granulosa cells (secondary follicles) and follicles with small antral cavities are rarely observed in the human ovary.

Primordial follicles in the human ovary are situated approximately one millimeter below the cortical epithelium [17] (Figure 25.2) embedded in a dense cortex of stromal cells and collagen bundles. At birth, it is estimated that 1 million primordial follicles [18] are present within the human ovary decreasing to approximately 25,000 by the age of 37 [19, 20]. Obviously, these are the candidate follicles to preserve with any cryopreservation regimen. The aim of cryopreservation is not only to preserve the structure and function of both the pre-granulosa cells and oocyte within individual primordial follicles but also to preserve the majority of the primordial follicles. This is complicated by the enormous variation in the density of the ovarian cortex and the distribution of follicles between patients. Histological examination of ovarian tissue destined for cryopreservation from over 150 patients, ranging in age from the early teens to over 40 years of age (Gook unpublished), shows that the cortex of the adult human ovary is extremely fibrous with highly variable distribution of follicles [17, 21]. In the young teenage ovary, abundant follicles appear to be evenly distributed around the cortex. However, in ovaries from women of more advanced reproductive age, follicles are depleted from specific regions, altering the appearance to that of follicle clusters with fewer individual follicles which are, in turn, becoming more sparsely distributed [22]. The lack of uniformity across the cortex will impact not only on the evaluation of methodology but also on the potential clinical success in older women requesting ovarian tissue cryopreservation. This has led to the conclusion that, for some older patients referred for ovarian tissue cryopreservation (such as breast cancer patients), cryopreservation of ovarian tissue may not provide a realistic option for fertility preservation and that an age limit of 35 should be applied [23].

Figure 25.2 A slice of human ovarian cortex stained with a live/dead stain, showing a cluster of primordial follicles just below the surface epithelium

Laboratory Procedures

Many of the clinics undertaking ovarian tissue cryopreservation are based in Europe and are required to adhere to the EU Tissues and Cells Directive EC/2004/23. Although not mandatory outside of the EU, the basic principles of the directive are sensible recommendations for treatment of biological material destined for transplantation after storage i.e. an accredited facility, traceability of identification, quality management, standard operating procedures, minimizing microbial contamination, trained staff and monitoring of critical equipment. The implications of some of these factors with respect to ovarian tissue warrant further discussion. Leibowitz L15 is the culture medium used to process ovarian tissue by many groups. Although cell culture quality, this only establishes sterility and that included reagents are not cytotoxic. This commercially available medium is not approved for clinical use and therefore should not theoretically be used in a clinical procedure. Commercially available media and cryoprotectant solutions used in assisted reproductive technology and approved for such use should be used in procedures for the handling and preparation of ovarian tissue whenever possible. When cryoprotectant solutions used for ovarian tissue freezing are not available commercially, it may be necessary that they are formulated in-house, where quality of reagents is paramount and a system of quality assurance should be established [24].

Included in the EU tissue directive is a requirement to minimize microbial contamination, and reduce risk of exposure to staff, indicating that all processing should be in a biological class II hazard cabinet. Although this is particularly important in the context of samples known to be positive for HIV or hepatitis, it is also relevant to urgent ovarian tissue storage cases where this status is unlikely to be known. This consideration is also relevant to the final storage conditions, where the possibility of cross contamination should be eliminated by storage in vapor phase or in a secondary sealed system. It is important to note that all cryogenic vials stored under liquid nitrogen allow seepage of liquid nitrogen and therefore require a heat sealed plastic sleeve covering.

Births from ovarian tissue stored for ≥9 years and subsequently grafted in patients without ovaries [11, 25] have established that, when stored at the appropriate temperature, tissue viability has been maintained. It is also clear from this data, and from the nature of patients storing ovarian tissue, that the duration of storage is often likely to exceed a decade or longer before thawing. Therefore, it is paramount to ensure maintenance of storage temperature whether through automatic filling or manual measurement and filling, together with rigorous temperature monitoring and remote alert alarms. Recent catastrophic failures of storage tanks emphasize the importance of monitoring, alarm testing and the documentation of a disaster plan. The highest risk of failure appears to be associated with the smaller tanks and will be likely to reoccur, causing distress for many patients and potential legal ramifications, unless a quick response backup system is routinely available. In some countries, in order to facilitate compliance with many of these requirements, a centralized processing and storage facility has been established where expertise, standard operating procedures and appropriate backup systems are available [26, 27]. Centralized facilities can also be advantageous when seeking charity or government funding [28]. However, this model requires transport of tissue, during which appropriate conditions must be maintained for long periods of time.

Transport

The most significant loss of follicles results not from cryopreservation but from ischemia; this occurs following grafting [29], but also if there is a delay before freezing as is the case for transport. Baird et al. [30] reported that 65% of sheep follicles are lost following exposure to 0°C for 2–3 hours in a cell culture medium (Leibovitz L15) and only a further 7% as a result of cryopreservation. Under the same conditions, a slightly higher loss due to ischemia (78%) was observed in mouse ovaries with no further reduction as a consequence of cryopreservation [31]. It is routine practice for organs destined for transplantation to be perfused with, and transported for a number of hours in, a basic salt solution at 4°C in order to reduce ischemia. However, it is questionable whether this rationale would apply to ovarian tissue in which the follicles are located just under the surface epithelium where there is minimal circulation.

Ovarian tissue is frequently transported for 4–5 hours at 4°C [32] and has been transported in some cases for up to 28 hours [33]. The consequences of these conditions are largely unknown for human tissue although there have been pregnancies and births after orthotopic grafting of tissue that had been transported for 4–5 hours [32, 34] and 20 hours [35]. In the latter case healthy follicles were observed on the peritoneal grafted tissue at the caesarean delivery and sampling of this tissue revealed numerous healthy follicles at all stages of development [36]. Similarly, healthy follicles at all stages were observed in human cryopreserved ovarian tissue which had been transported for 4 hours at 4°C and subsequently xenografted into an immunodeficient mouse [21]. However, a measure of lipid peroxidation of human ovarian tissue following exposure to 4°C for 24 and 48 hours indicates that degradation of membranes is increased with cold transport [37]. Animal studies confirm similar outcomes i.e. that short exposure (3 hours), to either ~20°C or 0°C, had no impact on ischemia as measured by oxygen consumption whereas extended exposure (24 and 48 hours) did increase ischemia in the tissue [38]. Although ischemic damage to the tissue has been reported, there is a general consensus from animal studies that there is no impact on primordial follicle morphology following exposure to 4°C for 18–24 hours. However, there is an impact on the morphology of growing follicles and the subsequent implantation of embryos derived from the oocytes in grafted ovaries exposed for >4 hours to 4°C (reviewed [39]).

It is possible that cold ischemia may be reduced by transport in a more appropriate medium. A comparison of exposure of human ovarian tissue to 4°C for 24 hours in a Leibowitz based medium or a histidine–tryptophan–ketoglutarade solution (HTK, an organ transport medium) showed enhanced follicle survival and a lower level of lipid peroxidation in the HTK [40]. HTK medium is currently used for extended duration transport of ovarian tissue [40]. However, a similar comparison of a culture medium designed for gametes and embryos (Quinns Hepes modified HTF) and HTK for a shorter duration (2 hours) followed by xenografting reported a dramatic loss of stromal tissue structure and follicles regardless of the medium used [22]. Others have also examined the potential of antioxidants to reduce ischemic damage. Although ascorbic acid did improve nuclear integrity (DNA fragmentation), there was no improvement in primordial follicle health [38].

Oocyte Cumulus Complexes Recovered From Ovarian Tissue

Handling temperature can also have an impact on cumulus oocyte complexes (COCs) that may be isolated from small antral follicles during tissue preparation. No viable COCs were recovered from tissue transported for 22 hours at 4°C [41], but similar maturation rates (26%) were observed for COCs with and without transport for 4 hours at 4°C [42]. However, these rates are lower than reported (50%) when temperature was maintained at 37°C and isolation was performed in an IVF medium [43]. Whether the oocytes in the non-transport group in the Yin et al. study are compromised due to processing of the tissue below 37°C or the use of a simple salt solution is unknown.

Generally, two types of COC are recovered during tissue preparation; those with very few layers of corona cells, frequently with corona cells in the perivitelline space (an indicator of atresia, Figure 25.3A), and those with a dense compact mass of corona cells in which only a vague outline of the oocyte can be seen (Figure 25.3B). Following in vitro maturation (IVM), approximately 50% will mature [43–47]. Subsequent fertilization and embryo development have also been achieved and live births reported [48–50]. COCs may also provide potential additional material in the prepubertal patient but, although higher numbers of oocytes were recovered, a higher proportion degenerated and fewer matured in culture compared to those obtained from adults [45]. Advances in knowledge of the interaction between the oocyte and associated follicular cumulus cells in the maintenance of meiotic arrest and the subsequent re-initiation of meiosis gained from animal studies [51, 52] has also improved outcomes with IVM of animal [53, 54] and human oocytes [55]. Some of these developments have been applied to COCs from ovarian tissue with promising results [43], but the normality of these mature oocytes requires further assessment.

Figure 25.3 Oocyte cumulus complexes isolated during preparation of human ovarian tissue for cryopreservation; (A) a poor quality oocyte with corona cells within the perivitelline space, (B) an oocyte embedded in large cumulus mass

Follicles can also be collected from human ovarian tissue following enzyme and mechanical isolation [56–58]. Multistep culture systems have achieved growth of such follicles to over 10 times their original size [59] and development of oocytes within these follicles to the metaphase II stage [60, 61]. Although these follicles can be cryopreserved, the efficiency of this has yet to be accurately assessed in the human. However, isolated murine follicles have been successfully cryopreserved using both slow freezing [62–65] and, more recently, using vitrification [66–68].

Cryoprotectants

Cellular density and tissue geometry will affect infiltration of cryoprotectants. Diffusion of cryoprotectants is relatively rapid in murine ovaries, which consist almost completely of abutting developing follicles with almost no fibrotic material and are, therefore, relatively porous. However, the densely fibrotic human ovarian cortex is very different. Murine ovarian tissue is, therefore, an unsuitable model for the human in this respect.

In the case of permeating cryoprotectants, the aim is to gradually displace cellular water without inducing excessive shrinkage. To overcome the difficulties associated with the density of human ovarian tissue, the use of higher concentrations of cryoprotectants has been suggested. However, under these circumstances, the reduced aqueous phase and hyperosmotic conditions result in excessive shrinkage of cells and loss of cell–cell communication [72]. This in turn may result in subsequent impairment of tissue function even though morphology may appear normal. The reduction in filamentous actin traversing the zona of oocyte cumulus complexes isolated from cryopreserved murine ovaries indicates that this may occur in ovarian tissue [73].

Four permeating cryoprotectants: glycerol (GLY), dimethyl sulfoxide (DMSO), ethylene glycol (EG) and propanediol (propylene glycol; PROH) have been used in human and animal ovarian tissue cryopreservation. For all permeating cryoprotectants a concentration of 1.4 or 1.5 mol/l has generally been used for slices of human and animal ovarian tissue [74]. However, higher concentrations of PROH (2 and 4 mol/l at ambient temperature) do not appear to cause follicular toxicity when compared to control tissue exposed to no cryoprotectant [75]. In contrast, concentrations of DMSO above 2 mol/l were toxic. Follicular toxicity has also been shown for concentrations of EG above 2.0 mol/l [76].

Although GLY was used in initial rodent studies [1, 2, 77], it has not been used clinically for human tissue. This is probably due to the reported low proportion of follicles (10%) surviving after cryopreservation of human tissue with glycerol, the subsequent lack of follicles observed in the majority of grafts using this tissue [78] and the failure of follicles to survive in vitro culture [79]. The poor outcome observed with glycerol is likely to be a consequence of the slow rate at which glycerol permeates tissue relative to other cryoprotectants [80], although this can be compensated for by increasing the dehydration time [81] or applying a slower cooling rate.

The rate of cryoprotectant penetration through tissue is also a function of temperature. There is a suggestion that the membrane permeability coefficient (Lp) of ovarian cells is lower at 4°C than at 27°C for all four cryoprotectants [82]. Newton 1998 [80] showed that both DMSO and EG had penetrated through 76% of human ovarian tissue in 20 minutes at 4°C but that this occurred twice as rapidly at 37°C. Of interest is the additional observation that neither of these cryoprotectants had permeated all of the tissue even after 90 minutes at 37°C. In contrast, PROH penetration was slower at 4°C than DMSO and EG, requiring 30 minutes to achieve 76% penetration but at 37°C PROH had a significantly higher rate of diffusion, achieving 100% penetration by 15 minutes. It is no surprise then that reduced follicle survival was observed following dehydration of larger pieces of human ovarian tissue at 4°C for 30 minutes in PROH (44%) relative to DMSO (84%) and EG (74%)[78]. High follicle survival following cryopreservation under the same conditions using EG has also been observed by others [83]. Similar levels of survival (>80%) have been achieved using appropriate conditions for individual cryoprotectants, that is, DMSO at 4°C and PROH at room temperature [84]. However, to achieve a similar survival rate using a short (10 minutes) exposure time with larger (2 × 10 × 10 mm) pieces of sheep ovarian cortex, which is also relatively fibrous, a higher concentration of either PROH or DMSO (2 mol/l) was necessary [75]. Effective dehydration of tissue is therefore dependent on the tissue geometry together with the rate of penetration of the cryoprotectant which is in turn a function of temperature. Modeling the cryoprotectant loading and displacement of water [82] suggests that, with a decrease in surface area to volume ratio at a set Lp (which relates to the temperature and cryoprotectant), a longer duration of exposure is required. Therefore, to conclude from some of the earlier studies that a cryoprotectant is unsuitable for ovarian tissue cryopreservation is not justified. It would appear that insufficient dehydration may have resulted from sub-optimal conditions for a specific cryoprotectant and that, with optimization, a variety of protocols may be appropriate.

Non-permeating cryoprotectants, generally sucrose at a concentration of 0.1 mol/l, have also been used to facilitate dehydration of ovarian tissue [15, 84–88]. Studies have shown that increasing the sucrose concentration significantly improves outcomes for embryo [89] and oocyte [90] cryopreservation, but addition of various concentrations of sucrose in combination with DMSO did not appear to result in improved cryopreservation of ovarian tissue [80] and no other studies have specifically compared protocols with and without sucrose.

The use of sucrose with the permeating cryoprotectant PROH [15] has been investigated in an attempt to optimize dehydration of both pre-granulosa cells and oocytes within primordial follicles [91]. Equivalent proportions of morphologically intact oocytes could be achieved with shorter time exposure when the sucrose concentration was increased from 0.1 to 0.2 mol/l. However, extended exposure to higher sucrose resulted in decreased oocyte survival [91] and the observation, within the surviving oocytes, that the majority of cytoplasm consisted of vacuolation and lysed mitochondria. This phenomenon was not observed with the lower sucrose concentration (0.1 mol/l) which, in contrast, resulted in an improvement in the proportion of morphologically intact oocytes and pre-granulosa cells with time. The proportion of surviving oocytes with normal cytoplasm also increased with time in 0.1 mol/l sucrose [15]. Extended exposure to the elevated sucrose concentration may result in inappropriate osmotic gradients, which is also the likely explanation for the low survival of pre-granulosa cells (Figure 25.4B) and poor cytoplasmic morphology in oocytes when applying two-step compared to single-step dehydration with equivalent final cryoprotectant concentrations [15]. By manipulation of dehydration using PROH and sucrose at ambient temperature, a high proportion of intact pre-granulosa cells (74%), intact oocytes (91%) and oocytes with normal cytoplasmic appearance (95%) could be achieved (Figure 25.4 C). However, morphology of the stromal tissue was consistently poor irrespective of the regimen (Figure 25.4B & C) compared to non-cryopreserved (Figure 25.4A), highlighting the relative complexity associated with dehydration of tissue containing multiple cell types.

Figure 25.4 Primordial follicles present in ovarian tissue following cryopreservation. (A) non-cryopreserved, (B) dehydrated using 1.5 M PROH and 0.2 M sucrose (C) dehydrated using 1.5 M PROH and 0.1 M sucrose

Rate of Cooling and Warming

To achieve the highest survival of any cell, the rate of cooling should be sufficiently slow to allow the majority of water (>90%) to leave the cell while also limiting the exposure of the cell to high solute effects which are lethal. This has a greater consequence in tissue due to ice and solute effects occurring in extracellular spaces [92]. Dependent on the rate of cooling, as a cell decreases in volume due to further dehydration there is an increase in the extracellular volume. As a consequence of the increase in extracellular volume and the deposit of salts in these extracellular spaces, there is a change in the cell’s membrane permeability parameters (Lp and E). It has been predicted that at rates below 6°C/min, cells within tissue will be completely dehydrated [92].

In the majority of studies using a controlled rate of cooling, the rate used has been similar to the rates used for controlled rate embryo cryopreservation (2°C/min to ice seeding temperature followed by 0.3°C/min). As with embryo and oocyte cryopreservation, the rate of cooling for ovarian tissue has never been systematically optimized. Mazur [69] estimated that the theoretical rate of cooling for organs or tissues should be <1°C/min, but a comparison of the standard slow rate (0.3°C/min) and a faster cooling rate (2°C/min) with large pieces (200 mm³) of ovine ovarian tissue demonstrated a significant improvement in follicular survival at the faster rate [75]. In an attempt to investigate this for human ovarian tissue, the slow rate (0.3°C/min) was compared to two faster cooling rates following the same pre-freeze dehydration and the morphology of each cell type determined [15]. Poor morphology of the whole tissue was observed with a rapid rate (to be discussed in more detail later in the context of vitrification). At an intermediate rate (~36°C/min), a high proportion of the stromal cells and collagen bundles appeared normal but only half the oocytes were intact and almost all pre-granulosa cells were abnormal (swollen and with swollen nuclei), again emphasizing the importance of variation in cell size. With the slow rate, approximately half of both the stromal and pre-granulosa cells and over 80% of the oocytes were normal. Normal appearance in the pre-granulosa cells and oocytes could be further enhanced by applying the slow rate after more extensive dehydration prior to cryopreservation but this was at the expense of stromal cell survival [91, 93].

Recent evaluation of equine and macaque ovarian tissue, using differential scanning calorimetry and low temperature microscopy, has demonstrated the impact of the cryoprotectant structure, concentration of cryoprotectant and rate of cooling on the membrane permeability parameters. Although the membrane potential for individual cells is independent of these factors, the extracellular composition in tissue alters the membrane permeability of the surrounding cells. For example, when using EG, the Lp is three times faster when equine ovarian tissue is cooled at 0.5°C/min compared to at 40°C/min [92]. Interestingly, when a combination of two permeating cryoprotectants is used the membrane permeability is the same regardless of the cooling rate [92]. Although such a combination has not been used for human ovarian tissue with a slow cooling rate, this warrants further investigation.

During controlled rate cryopreservation, cryoprotectant crystallization will occur, and the temperature at which this occurs is specific for the cryoprotectant. For PROH or DMSO this occurs at −6 to −8°C. Without manual seeding, crystallization will be initiated at any solid surface i.e. throughout the tissue. Demirci et al. [75] reported a dramatic deviation from the normal cooling curve resulting in reduced follicular survival in the absence of manual seeding. Manual seeding at a slightly higher temperature (−5°C compared to −7°C for DMSO) appeared to improve follicle survival [79].

Similarly, damage can occur during thawing as a result of crystallization of water. During warming, ice crystallization growth is again nucleated at surfaces and is exacerbated by a slow warming rate and larger fluid volume where thermal exchange is slower. Irrespective of the cryopreservation method used (controlled rate or vitrification), thawing should be as rapid as possible. Although critical, the thawing temperature is rarely reported and, again, almost no studies have systematically investigated this aspect of cryopreservation with animal or human ovarian tissue. Generally, thawing rates reported are slow, involving initial exposure to room temperature followed by 37 °C or just exposure to 37°C (Table 25.1).

Table 25.1 Cryopreservation methodology used in clinical cases where oocytes and embryos have been derived from heterotopic grafts

Size of tissue slices (mm)	Cryoprotectants	Dehydration	Start Temp	Seed temp	Thaw temp	First thaw solution	Outcome	Reference
5 × 5 × 1 and 15 × 5 × 2	1.5 M DMSO +0.1 M sucrose	4°C 30 min	0°C	−7°C	Air 30 sec then 37°C 2 min	1.5 M DMSO + 0.1 M sucrose	20 oocytes, 8× MII, 2 × fertilized, 1 × 3 cell, 1 × 4 cell ET	[5]
5 × 5 × 1	1.5 M EG +0.1 M sucrose	4°C 30 min	*	−9°C	37°C	0.75 EG + 0.25 M sucrose	3 oocytes, 2 × MII, 1 × 4 cell ET 1 × 5 cell ET	[7, 154]
5 × 5 × 1 and 10 × 10 × 1	1.5 M DMSO +0.1 M sucrose	4°C 30 min	*	−7°C	35°C 2–3 min	1.0 M DMSO + 0.1 M sucrose	6 oocytes, 4 × fertilized, 1 × 6 cell, 1 × 3 cell, 1 × 2 cell (no ET)	[9]
5 × 5 × 2	1.5 M DMSO +0.1 M sucrose	4°C 30 min	4°C	−7°C	Air 2 min then 25°C 2 min	1.5 M DMSO	2 oocytes, 1 × 3 cell ET	[8]
10 × 10 × 1	1.5 M DMSO +0.2 M sucrose	4°C 30 min	4°C	−9°C	Air 30 sec then 37°C 2 min	*	20 oocytes, 6 × MII, 1 × 7 cell (no ET)	[6, 88]
*	1.5 M DMSO +0.1 M sucrose	4°C 30 min	4°C	−7°C	Air 30 sec then 37°C 2 min	1.4 M DMSO + 0.2 M sucrose	10 oocytes, 9 × embryos, 4 × blastocysts (2 ET)	[10]
5 × 1 × 1	1.5 M PROH +0.1 M sucrose	22°C 90 min	22°C	−7°C	~95°C 50 sec	1.0 MPROH +0.2 M sucrose	62 MII, 38 × fertilized, 34 × embryos, 3 births	Gook unpublished

All cryopreserved using controlled rate slow freezing at 2°C/min to seed temperature followed by 0.3°C/min to −40 °C or −50 °C

* = not provided, DMSO = dimethyl sulfoxide, EG = ethylene glycol, PROH = propanediol, MII = metaphase II oocyte, ET = embryo transfer

Exposure to some cryoprotectants, such as DMSO, at higher temperatures may increase their toxicity. A comparison of thawing of murine ovarian tissue, cryopreserved in DMSO, at 27, 37 or 47°C showed a significant reduction in follicle survival at the higher temperatures [79]. In our experience, thawing tissue cryopreserved in PROH in a 1 ml volume in a cryovial requires over 3 minutes incubation at 37°C to achieve liquefaction (a warming rate of 61°C/min Figure 25.5A). By reducing the fluid volume (0.5 ml) and warming at 37°C, the rate increases to 91°C/min (Figure 25.5B) and when warming in a ~95°C water bath the rate of warming to liquefaction is increased to 223.6°C/min (Figure 25.5B). Once liquefaction is observed, the vial is removed from the water bath and the temperature does not exceed 22°C. Tissue warmed rapidly using this latter procedure resulted in pregnancies following grafting at heterotopic sites [11] suggesting no effect on developmental potential.

Figure 25.5 Graph of the rate of warming of vials containing; A. 1 ml or B. 0.5 ml cryoprotectant solution and warmed slowly (room temperature for 30 sec followed by 37°C, 37°C) or rapidly (~95°C)

Rate of Cooling

Vitrification offers the potential benefit of overcoming many of the issues associated with ice crystallization discussed earlier. The problem associated with vitrification of ovarian tissue, however, is how to achieve the ultrarapid cooling and warming rates required. Vitrification of murine ovaries has been successful with subsequent births of pups reported [94, 95] following vitrification in cryo-straws and plunging in liquid nitrogen. However, when a faster cooling rate was achieved by direct contact with liquid nitrogen, it resulted in better preservation as evidenced by significantly more morphologically normal, viable follicles following grafting [29] and pups [94]. This improvement, however, may also be partly due to more appropriate dehydration prior to vitrification. Vitrification in cryo-straws resulted in poorer outcomes relative to controlled rate cryopreservation for all parameters measured. Wang et al. [96] have also shown the importance of the cooling rate for vitrification with both mouse and human ovarian tissue. The impact of the cooling rate is further emphasized with higher viability in follicles following ovine ovarian tissue vitrification by direct contact with liquid nitrogen relative to solid surface contact vitrification [97] and human ovarian tissue vitrified on metal needles compared to direct plunging [98, 99] or in a metal container [100]. Using the same concept as the metal needle to conduct temperature, human ovarian tissue has been successfully vitrified using a row of needles where tissue is placed across needles with direct contact to liquid nitrogen and, more recently, on a metal carrier which is subsequently sealed providing a closed vitrification system [101].

Notwithstanding the earlier discussion, slower cooling rates and warming have been successfully applied for vitrification of macaque ovarian tissue in straws following assessment of the cryoprotectant concentrations required to achieve glass transition [102]. However, this requires very high concentrations of cryoprotectants (4.8M EG + 3.6 M glycerol + polymers).

Cryoprotectant

It is of concern that, in order to achieve dehydration of ovarian tissue and facilitate vitrification with ultrarapid rates of cooling, exposure to very high concentrations of cryoprotectants is necessary. In contrast to embryo and oocyte vitrification, in which exposure to these very high concentrations of cryoprotectants is limited to less than 30 s and many normal births have been reported, dehydration of ovarian tissue will require significantly longer times in high concentrations of cryoprotectant. Ovine hemi-ovaries have been successfully vitrified resulting in live births, although one of the four births was a large lamb that died shortly after delivery as a result of malformations [103]. The cocktail of cryoprotectants (2.6 M DMSO, 2.6 M acetamide, 1.3 M PROH, and 7.5 mM polyethylene glycol) used in this study has also been used for human ovarian tissue with evidence of preservation of normal follicular morphology [104]. In contrast, poor developmental capacity was observed following vitrification of murine embryos with this protocol [105]. Also relevant to this, is the linear relationship between concentration of the cryoprotectants DMSO and PROH and the formation of formaldehyde [106].The ovine study highlights the need for these protocols to not only be evaluated at the level of follicle health, which is the main outcome in most of the studies discussed below, but also in relation to the live birth outcomes.

In one human ovarian tissue vitrification procedure, tissue was exposed to 1.0 M DMSO followed by 2.8 M DMSO at ambient temperature for 15 minutes at each concentration [107]. To facilitate dehydration others have exposed human ovarian tissue to the lower concentration, but for 15 minutes at 38°C, followed by a reduced time (2 minutes) in a higher concentration [37]. Whether high apoptosis in this study is due to toxicity from DMSO exposure at 38°C or insufficient dehydration is difficult to ascertain. Dehydration prior to controlled rate cryopreservation in DMSO (1.5 M) has routinely been performed at 4°C to limit toxicity [108]. Using this approach, successful vitrification of human ovarian tissue has been achieved with dehydration in 1.5 M DMSO at 4°C followed by a combination of PROH and EG (total molarity 4.4 M) [109, 110]. The concept of a combination of cryoprotectants to achieve high molar concentration but with reduced toxicity was initially suggested by Ali and Shelton [111] for embryo vitrification but has also been applied to ovarian tissue. The reduced toxicity was confirmed in a study of monkey ovarian tissue vitrified in a single cryoprotectant at high concentration (EG 5.6 M) or an equivalent molar concentration but with a combination of EG and DMSO [112]. Tissue subsequently grafted following vitrification with the combination produced mature oocytes which fertilized [113]. This combination of cryoprotectants has now been used for vitrification of human ovarian tissue and two births reported [114]. These results are promising but there are concerns associated with high levels of residual cryoprotectants used for vitrification still present in human ovarian tissue after the completion of warming [115], which need to be addressed.

Care must be exercised, when comparing outcomes from slow freezing and vitrification, to ensure that appropriate methodology has been used for both and that the outcomes are meaningful (see subsequently). A comparison of the Keros vitrification method and slow freezing using DMSO concluded that slow freezing was superior [116], but in both cases tissue was warmed using the same slow rate. This is likely to have severe consequences on the vitrified tissue, still containing water which will result in ice crystal growth during slow warming. Studies on cooling and warming rates with varying concentrations of cryoprotectants have indicated the warming rate to be the most critical parameter [117].

Histological Evaluation

In contrast to evaluation of outcomes from embryo or gamete cryopreservation, assessing the survival and viability of cryopreserved ovarian tissue poses specific challenges. In many of the studies discussed previously, success has been measured in terms of follicle survival after isolation from thawed tissue or apoptosis of follicles. The validity of this approach depends on three fundamental assumptions: (a) that lysed or destroyed follicles will be detected after isolation, (b) that loss of follicles as a result of enzyme digestion will be equivalent for cryopreserved and fresh tissue, and (c) specifically in the case of human ovarian tissue, that there is an even distribution of follicles throughout the tissue. However, once an oocyte within a primordial follicle has lysed, it is essentially impossible to identify the follicle. Therefore, only follicles with an intact oocyte will be included in these studies, and also in those studies assessing apoptosis, resulting in overestimation of survival. In addition, much of the stromal tissue is damaged with controlled rate cryopreservation rendering follicles within it more vulnerable to enzymes and this is also likely to vary between protocols. Finally, due to the extent of variation in follicle distribution in human ovarian tissue, it is invalid to express the number of follicles present after cryopreservation as a proportion of the number in a non-cryopreserved sample. Although these criticisms weaken some of the conclusions which have been drawn earlier, there are no other available studies of this type on cryopreservation of ovarian tissue.

In contrast, it is potentially possible to overcome these problems by histological evaluation of the entire tissue, but there are very few studies which have attempted this [31] and expressed normal morphology as a proportion of the total number of follicles within a piece of tissue [15, 91, 118]. Histological examination, at both the light and electron microscope level, has generally assessed only a small sample to estimate overall follicle integrity [109, 119, 120], although this has allowed detection of abnormalities such as oocyte shrinkage [121], vacuolated areas within the oocyte cytoplasm [122], loss of mitochondrial cristae [123] and lysis of pre-granulosa cells [124].

There are only two morphometric studies of human ovarian tissue which assess cryopreservation, one of which is an evaluation of the most commonly used procedure [3] using DMSO as a cryoprotectant and controlled rate cooling on tissue from six patients[118]. In this study, almost half of the follicles and the vast majority of oocytes (80%) were damaged. Parallel assessment of apoptosis confirmed these observations. The other study assessed the proportion of intact pre-granulosa cells and oocytes together with the relative normality of the oocyte cytoplasm as estimated by vacoulation and normal mitochondria within every oocyte for a range of cryopreservation procedures using PROH and sucrose as cryoprotectants [15, 91]. Observations were verified by electron microscopic evaluation of a small number of follicles (Figure 25.6). The highest proportion of both oocytes (85%) and pre-granulosa cells (74%) with normal morphology was observed following dehydration for 90 minutes in 1.5 M PROH + 0.1 M sucrose at ambient temperature followed by a slow controlled rate of freezing. Although this type of morphometric assessment has provided evidence of morphological normality after cryopreservation, it gives no indication of viability or function.

Figure 25.6 Antral follicle development in cryopreserved ovarian tissue from four women xenografted for 23 weeks under the kidney capsule in SCID mice

Viability staining has also been used to assess follicles within a piece of tissue [125, 126] following cryopreservation. Generally, this staining will identify live cells, on the basis of an intact membrane and cytoplasm which is functionally capable of cleaving a chromagen, and dead cells on the basis of their inability to exclude a nuclear stain [76, 83]. Again, this form of assessment has limitations [121]. Tissue pieces must be very small to facilitate diffusion of dye and only follicles with a live oocyte are detected. Although the nuclear stain will detect the germinal vesicles (GV) of lysed oocytes, these are indistinguishable from the nuclei of stromal or pre-granulosa cells, resulting in the potential for overestimation of follicle viability. Finally, it is important to remember that we cannot infer, on the basis of viability staining, that the developmental potential of these primordial follicles has been retained following cryopreservation.

In Vitro Developmental Potential

Assessment of expression of developmental potential in vitro is an attractive possibility, but requires an understanding of the requirements for initiation of growth in primordial follicles. Additionally, many months of culture may be required. More advanced follicles isolated at the secondary stage, in which initiation of granulosa cell proliferation has occurred, can be successfully grown in culture [59, 60, 127–130] and this has also been achieved following controlled rate cryopreservation [131, 132] and vitrification [133]. However, the predominant follicle present in the human ovary is the primordial. There is evidence that murine primordial follicles, grown in a two phase culture system, can produce live pups [134, 135] but this has not been established using cryopreserved ovarian tissue. Primordial follicles isolated from human cryopreserved tissue have been xenografted in an attempt to bypass the potential inadequacy of culture conditions but only 20% were present after 7 days [136]. The culture of isolated primordial follicles, whether from fresh or cryopreserved tissue, has proved problematic in animal models [137–139] and in human [85, 140, 141]. However, follicular development to the secondary stage has been established by culturing primordial follicles within stromal tissue [142, 143]. Development from primordial to early antral stage in vitro has been demonstrated using a two-stage culture system with fresh tissue [57, 144] and subsequent encapsulation of the secondary follicles resulted in a mature oocyte [61]. Results to date on demonstration of developmental potential using in vitro systems, while promising, are limited to relatively few follicles.

Assessment of Developmental Potential In Vivo

As a result of the difficulties associated with culture of primordial follicles in vitro, various in vivo approaches involving grafting of cryopreserved ovarian tissue, either alone or in combination with final maturation in vitro, have been applied. Clear evidence of preservation of developmental potential has been established by the birth of live offspring from primordial follicles following grafting of ovine and murine cryopreserved ovarian tissue [3, 108, 145]. Heterotopic grafting of murine cryopreserved tissue with subsequent IVM has also resulted in live births [146].

In the human situation, xenografting of cryopreserved tissue into immunodeficient mice has been used to assess preservation of developmental potential in a number of studies and has established that follicles are viable and capable of development [31, 86, 147–149]. Developmental capacity to the antral stage has been shown to be preserved with the commonly used cryopreservation regimen using DMSO [148, 149] and also the PROH/sucrose procedure [147]. Although primary and secondary follicles within human ovarian tissue may have survived cryopreservation, the time required for these antral follicles to develop post grafting (5–6 months) suggests that they have developed from primordial follicles. Reproducibility of preservation has been established by development of numerous antral follicles using tissue cryopreserved from multiple patients with both procedures [150–152] (Figure 25.6 PROH procedure). Full developmental competence has been shown to be preserved with both procedures with evidence of ovulation (corpora lutea) [149, 153] and mature oocytes [150, 151, 153]. Unfortunately, although mature oocytes were aspirated from follicles in our laboratory, experiments to determine subsequent fertilization were prohibited by law in Australia.