**2.5 Discussion**

The greater challenge of studies related to the cryopreservation of the ovarian tissue is to define a freezing protocol adapted to the different cell types such as oocytes, follicular cells, stroma cells, etc. One of the objectives of our team was to compare the effects of different freezing parameters based on the morphology and the viability of the follicles, the evaluation of the ultrastructure of the ovarian tissue, the DNA fragmentation of the oocytes or the graft of the ovarian tissue. When the mathematical model was validated, the use of experimental fractional designs allowed us to know simultaneously the individual and the relative effects of different chemo-physical freezing parameters for each species. This statistical method firstly allows a global evaluation of cryopreservation protocols and discriminates the most valuable factors. Finally, the factors which seemed to have a discriminating effect on follicular morphological preservation were evaluated in a wider population. (Neto et al., 2008).

The results of the experimental design in the doe rabbit and in the bitch show that the postseeding freezing rate is one of the most important chemophysical factors influencing the morphology or the viability of ovarian follicles. A slow freezing rate (0.3°C/min) seems to be more appropriate for the cryopreservation of the doe rabbit and bitch ovarian tissue. Nevertheless, no influence of the freezing rate was observed in the queen and in the cow. Most of the authors use a very slow freezing rate, which is derived from embryo freezing protocols. However, few studies show the importance of this freezing parameter. In contrast to our results in the bitch and in the doe rabbit, but in accordance with our result in the

the follicular growth potential were maintained. The frozen-thawed tissues were grafted to female SCID mice as previously described. The histological assessments of the follicle population revealed a significant increase in the density and distribution of secondary follicles from eight weeks post grafting compared to the follicle population at 1 week (*P*<0.05). Consequently, the shift from primordial-primary follicles to secondary follicles occurred in a time laps of eight weeks. Moreover morphologically normal follicles were observed until 16 weeks post transplantation and intact secondary follicles with more than 3 layers of granulosa cells and a normal oocyte surrounded by a well-defined zona pellucida were present at this time. Despite a massive follicular loss touching particularly the early follicles and occurring just after grafting, the graft survived long term xenografting. Similarly, after an important loss just after grafting (one week grafts) the stromal cell number increased during the graft period, to reach a density comparable to fresh ovarian tissue at 16 weeks. Otherwise, the vascularization setting-up was assessed by immunohistochemistry as developed previously. The vessel density analysis revealed that already at one week post grafting the vessel density within the graft was comparable to the fresh control ovarian tissue. Moreover, the vessel density tended to increase at 16 weeks post grafting compared to the other groups (fresh control, one and eight weeks, *P*<0.05) even if no significant difference was registered. No antral follicles was present at the end of the graft time, however, persisting vaginal cornification was noticed on the recipients from 5-9 days post grafting, and estrus behavior was observed several times during the graft period in the recipients cages indicating an hormonal activity resulting from the graft. Taken together, these in vivo results confirm the good preservation of the bitch ovarian tissue by

The greater challenge of studies related to the cryopreservation of the ovarian tissue is to define a freezing protocol adapted to the different cell types such as oocytes, follicular cells, stroma cells, etc. One of the objectives of our team was to compare the effects of different freezing parameters based on the morphology and the viability of the follicles, the evaluation of the ultrastructure of the ovarian tissue, the DNA fragmentation of the oocytes or the graft of the ovarian tissue. When the mathematical model was validated, the use of experimental fractional designs allowed us to know simultaneously the individual and the relative effects of different chemo-physical freezing parameters for each species. This statistical method firstly allows a global evaluation of cryopreservation protocols and discriminates the most valuable factors. Finally, the factors which seemed to have a discriminating effect on follicular morphological preservation were evaluated in a wider

The results of the experimental design in the doe rabbit and in the bitch show that the postseeding freezing rate is one of the most important chemophysical factors influencing the morphology or the viability of ovarian follicles. A slow freezing rate (0.3°C/min) seems to be more appropriate for the cryopreservation of the doe rabbit and bitch ovarian tissue. Nevertheless, no influence of the freezing rate was observed in the queen and in the cow. Most of the authors use a very slow freezing rate, which is derived from embryo freezing protocols. However, few studies show the importance of this freezing parameter. In contrast to our results in the bitch and in the doe rabbit, but in accordance with our result in the

applying our cryopreservation method.

population. (Neto et al., 2008).

**2.5 Discussion** 

queen and in the cow, Demirci *et al.* observed a high (but not significant) proportion of follicles without any morphological defect after a post-seeding freezing rate of 2°C/min in the ewe (Demirci et al., 2001). Nevertheless, Gook *et al.* also observed a better follicular preservation when using a slow freezing rate (0.3°C/min) with human ovarian tissue (Gook et al., 1999). Whereas Cleary *et al.* observed no difference in terms of follicular growth after grafting, between a conventional embryo freezing protocol (0.3°C/min) and a passive cooling at 1°C/min from 0°C to -84°C on the mouse ovarian tissue (Cleary et al., 2001). Although these two cooling rates (0.3°C/min and 2°C/min) could be considered as slow, these results may be explained by a difference in cell dehydration during the post-seeding step. With rapid cooling rates, we can hypothesise that time required for the exosmose of the cell water is insufficient and consequently promotes the formation of lethal intracellular ice. While at very slow cooling rates, high level of dehydration occurs with concomitant increasing in solute concentration (salting out). Investigations on the freezing rate were extended in the queen with the evaluation of the third freezing phase. When associated with PROH, slow cooling in solid phase seems to be more appropriate. Nevertheless, several authors use cryopreservation protocols with a direct immersion into liquid nitrogen after – 40°C, such as Rodrigues et al. (2004) in the goat, Lucci et al. (2004) in the zebu cow, Santos et al. (2006) in the ewe or Lima et al. (2006) in the cat. Births had been obtained after graft of ovarian fragments frozen with such a protocol in the rabbit doe (Almodin et al., 2004b) and in the ewe (Almodin et al., 2004a), but not in the cat where in vivo follicular growth were observed when using a freezing protocol with controlled third phase (Bosch et al., 2004).

The experimental designs revealed a crucial role of the permeating CPA in the doe rabbit, in the cow and in the bitch, added to a crucial role of the non permeating CPA in the doe rabbit and in the cow. Among the various freezing protocols described in the literature, those using DMSO or PROH as permeating CPA seems to be more efficient, whatever the species. Our results suggest that PROH improve the follicular quality after freezing in the doe rabbit and in the queen. Results of experimental design obtained in the cow suggested a better morphological preservation rate when using PROH. However, in this species, standard comparison between DMSO and PROH doesn't confirm these results. Contrary to that, bitch ovarian tissue seems to be better preserved in freezing medium composed of DMSO. These results were confirmed by the follicular growth observed after xenograft. It can be hypothesized that DMSO penetrate better within the tissue than PROH. Indeed, the bitch ovarian tissue as the goat or the ewe is rich in collagenous fibbers and more fibrous than the doe rabbit ovarian tissue for example. Therefore, a good ability to penetrate within the ovarian tissue is an important characteristic for the chosen CPA Nevertheless, both CPA have sensibly the same molecular weight (PROH: 76.10 g/mol, DMSO: 78.14 g/mol) with a lower weight for PROH. Thus a better penetration of the DMSO cannot be explained by this physical parameter. The explanation may come from the toxicity of both CPAs. Our team also investigated the toxicity of the both CPAs on bitch ovarian tissue after equilibration steps at room temperature without freezing and registered a deleterious effect of PROH compared to DMSO on preantral follicle viability in this species.

Except in canine and feline models, addition of non-permeating CPA in the freezing medium seems to improve the protective effect of CPAs. The protective effect of PROH seems to be improved when it is associated with trehalose, in the doe rabbit and in the cow. This observation was confirmed by electron microscopy evaluation of the doe rabbit ovarian

New Approaches of Ovarian Tissue Cryopreservation from Domestic Animal Species 225

For the first time a complete evaluation process of important factors influencing the morphology and the viability of preantral follicles has been performed after equilibration process and freezing in different species. These results suggest that cryopreservation of ovarian tissue is a promising and suitable technique that could be used as complementary tool to embryo cryopreservation, to preserve the animals' genetic resources by the female pathway. Doe rabbit could also be used as a biomedical model to investigate the long term consequences of cryopreservation on ovarian follicles and the birth of future progenies.

In definitive, the use of factorial fractional experimental design approach allowed us to develop suitable cryopreservation protocol in different species, while reducing the number of experiments and increasing the number of parameters evaluated. However it can be noticed in our model species, but also in the literature, that results can be radically different according to the species. Moreover, among the numerous articles published on ovarian tissue cryopreservation heterogeneous results can be observed in the same species after applying roughly or widely different slow-freezing protocols. One of the candidates to explain such disparity in the obtained results is the amount of ice crystals formed during slow freezing. Indeed, the strategy of slow cooling is to decrease cell temperature enough slowly to allow removal of most of the freezable intracellular water before reaching the ice nucleation temperature. The main objective of this method is to avoid intracellular ice crystal formation which is known to be lethal. However ice crystallization still occurs extracellularly with the risk of tissue shrinkage or disorganization of the tissue

As ice formation and melting are exothermic and endothermic phenomena respectively, they can be objectified and studied by thermodynamical measures. Among the various physical methods of analysis, Differential Scanning Calorimetry (DSC) is an interesting tool. In fact, DSC gives the opportunity to measure important parameters of a cryopreservation solution under dynamic conditions. A cryopreservation solution can thus be characterized by its thermal properties such as temperatures of phase transitions and quantity of ice crystallized and melted. Two types of DSC are commonly used: power-compensation DSC and heat-flux DSC. Our team chose the first one in order to study cryopreservation solutions

The power compensation DSC is based on the "zero balance principle" as explained as follow. The sample and a reference are placed in two microfurnaces which are continuously cooled by liquid nitrogen. The temperature of each microfurnace can be, on the one hand, precisely measured by a temperature sensor and, on the other hand, precisely adjusted by a heating resistor. Each microfurnace contains one sensor and one resistor. The principle of the power compensation DSC is to maintain the two microfurnaces under the same temperature regardless of phase transitions or reactions occurring in the sample. Thus, when a phase transition occurs in the sample, the heat released or absorbed by the sample has to be compensated by the heating resistor which is below the sample. Consequently, the calorimeter measures a difference between the heating powers provided by the two resistors. This difference reveals the phase transition. When this phase transition is

crystallization, this difference allows also us to measure the quantity of ice formed.

with a more fundamental approach than with biological methods.

**3. Perspectives** 

components.

tissue subjected to cryopreservation. Ultrastructure of doe rabbit follicles after cryopreservation was well preserved, but stromal cells and fibroblasts were damaged. Such alterations have been observed in human tissues after cryopreservation (Navarro-Costa et al., 2005; Santos et al., 2010). However, fibroblasts can easily be reproduced by cell division, indicating that damage to the stroma can be repaired. Collagen fibers did not seem to be damaged by cryopreservation in this study, but they were sparse in the doe rabbit ovary. This observation may explain the fragility of doe rabbit ovarian tissue during the equilibration, freezing and thawing steps (Neto et al., 2005).

Sugar are not systematically associated with permeating CPA, but Marsella et al. (2008), showed the advantageous effect of sucrose. Trehalose has been frequently used in embryo cryopreservation, but not in ovarian tissue cryopreservation. Sucrose and trehalose share the property to stabilise cellular membranes and proteins via the formation of hydrogen bonds with polar residues of phospholipidic membrane. This property allows preserving the membrane integrity under anhydrous conditions. Moreover, it modifies the temperature at which the separation of lipid phase occurs during cooling (Crowe et al., 1984; Crowe et al., 1985; Crowe et al., 2001). As compared to other sugars, trehalose seems to have a higher capacity to preserve biomolecules, cellular membrane and cells in a drying or in a freezing state (Crowe et al., 1996; Storey et al., 1998; Sano et al., 1999; Welsh and Herbert, 1999).

Few comparable studies have been reported in the cryopreservation of different mammalian species. Despite encouraging results in the different studied species, and except in the queen, none of evaluated protocols allows preserving the same proportion of normal follicles than before freezing. Most of authors observed a reduction of normal follicles in frozen/thawed ovarian tissue compared with fresh control when using similar freezing protocols in the mouse (Candy et al., 1997), the goat (Rodrigues et al., 2004), the cow (Lucci et al., 2004), and the ewe (Demirci et al., 2002). As for the queen, no morphological difference was observed in human follicles before and after cryopreservation (Hovatta et al., 1996; Fabbri et al., 2006) Newton observed similar proportions of "viable" follicles after freezing when using DMSO, ethylene glycol or PROH and xenografting (Newton et al., 1996).

Live births in the rabbit doe and follicular growths observed in the bitch after grafting of cryopreserved ovarian tissue shows the efficacy of evaluated freezing protocols. Almodin et al*.* obtained live offspring after grafting of small fragments of cryopreserved rabbit ovarian tissue using 1.5 M DMSO and a very slow post seeding freezing rate (Almodin et al., 2004b). In the bitch, results about in vivo growth obtained after cryopreservation are relatively scarce compared to other species. Ishijima et al. (2006) tried to transplant vitrified ovarian tissue (2M DMSO, 3M PROH, 1M acetamide) into immunodeficient mice during 4 weeks and observed signs of growth in the early follicles (primordial-primary follicles). However, they noticed an important follicular loss occurring just after grafting which is in accordance with our results obtained with slow-frozen tissue. The time necessary for setting up of the neovascularization within the grafted tissue seems to be more deleterious for the cells than the cryopreservation technique itself. Except our results obtained on the bitch ovarian tissue cryopreservation no other results using slow freezing of female germ cells was obtained in this species. Furthermore, live birth has not yet been obtained after ovarian tissue cryopreservation, but the difficulties to mastered *in vitro* maturation and fertilization steps in canines do not contribute to the development of this technique.

For the first time a complete evaluation process of important factors influencing the morphology and the viability of preantral follicles has been performed after equilibration process and freezing in different species. These results suggest that cryopreservation of ovarian tissue is a promising and suitable technique that could be used as complementary tool to embryo cryopreservation, to preserve the animals' genetic resources by the female pathway. Doe rabbit could also be used as a biomedical model to investigate the long term consequences of cryopreservation on ovarian follicles and the birth of future progenies.
