Ice formation and its elimination in cryopreservation of oocytes

Abstract

Damage from ice and potential toxicity of ice-inhibiting cryoprotective agents (CPAs) are key issues in assisted reproduction of humans, domestic and research animals, and endangered species using cryopreserved oocytes and embryos. The nature of ice formed in bovine oocytes (similar in size to oocytes of humans and most other mammals) after rapid cooling and during rapid warming were examined using synchrotron-based time-resolved x-ray diffraction. Using cooling rates, warming rates and CPA concentrations of current practice, oocytes show no ice after cooling but always develop large ice fractions - consistent with crystallization of most free water - during warming, so most ice-related damage must occur during warming. The detailed behavior of ice at warming depended on the nature of ice formed during cooling. Increasing cooling rates allows oocytes soaked as in current practice to remain essentially ice free during both cooling and warming. Much larger convective warming rates are demonstrated and will allow routine ice-free cryopreservation with smaller CPA concentrations. These results clarify the roles of cooling, warming, and CPA concentration in generating ice in oocytes and establish the structure and grain size of ice formed. Ice formation can be eliminated as a factor affecting post-thaw oocyte viability and development in many species, improving outcomes and allowing other deleterious effects of the cryopreservation cycle to be independently studied.

Keywords: assisted reproduction; cryopreservation; ice formation; oocyte; vitrification; x-ray diffraction.

Figure 1.. Correlation between optical images and diffraction patterns of cryocooled bovine oocytes at T=−173°C.

(a) Oocyte soaked in 100% strength vitrification solution (15% DMSO, 15% EG., 0.5 M sucrose) and fast cooled in LN₂, as described in the text, showing two diffuse rings characteristic of low-density amorphous ice I_LDA. (b) Oocyte soaked in 50% vitrification solution and fast cooled. Diffuse scattering characteristic of I_LDA. (c) Oocyte soaked in 90% vitrification solution and slow cooled in cold N₂ gas before plunging in LN₂. Diffuse but somewhat sharper scattering, largely consistent with I_LDA. (d) Oocyte soaked in 80% vitrification solution and slow cooled, showing well defined but somewhat broad diffraction rings at resolutions expected for cubic ice I_c. (e) Oocyte soaked in 40% vitrification solution and fast cooled. Sharp diffraction rings at cubic ice resolutions and beginning of hexagonal ice ring. (f) Oocyte soaked in 40% vitrification solution and slow cooled. Strong and azimuthally inhomogeneous rings at all expected resolutions of hexagonal ice I_h. White circle in each oocyte image has a diameter of 100 μm.

Figure 2.. Variation of ice diffraction intensity versus resolution with amount and type of ice.

Azimuthally integrated and background subtracted diffraction intensity corresponding to the samples and x-ray detector images shown in Figure 1 (a)–(f), arranged from bottom to top. The dotted black lines for the top three are fits to a model of stacking disordered ice with hexagonal plane stacking fractions Φ_h of (d) 0.35, (e) 0.40, and (f) 0.72.

Figure 3.. Analysis of ice diffraction during warming.

(a) Azimuthally integrated and background subtracted diffraction intensity versus resolution and (b) equivalent hexagonal ice unit cell volume, hexagonal stacking fraction Φ_h, and integrated diffraction intensity in ice versus time during warming of bovine oocytes. Samples from left to right: oocyte on a crystallography support, soaked in 40% vitrification solution and fast cooled; oocyte on a thick Cryotop, soaked in 60% vitrification solution and fast cooled; and oocyte on a crystallography support, soaked in 40% vitrification solution and slow cooled. Parameters were extracted from DIFFaX fits to azimuthally integrated and background subtracted diffraction patterns. Time t=0 corresponds to the opening of the valve controlling room temperature N2 gas flow. Sample warming, as indicated by a change in unit cell parameter, began ~20 ms after valve opening, due to the time required to establish warm gas flow at the sample position. Unit cell volumes of ~128.1 ų and ~130.5 ų correspond to temperatures of ~100 K and ~273 K (Fig. S7). Dotted red lines in the upper left and right panels of (b) have a slope corresponding to a warming rate of ~130,000 °C/min. The dotted turquoise line in the upper middle panel of (b) corresponds to a warming rate of 25,000 °C/min.

Figure 4.. Effect of CPA concentration and cooling rate on amount of ice formed in bovine oocytes during warming.

Maximum integrated intensity in ice diffraction rings observed during warming versus CPA concentration in the vitrification solution used. Oocytes were slow cooled (at ~30,000 °C/min, orange circles) or fast cooled (at ~600,000 °C/min, blue circles), and then warmed at ~150,000 °C/min in a N₂ gas stream. Closed and open symbols indicate oocytes that did and did not show crystalline ice diffraction after cooling and before warming. The dotted line is a fit to the fast cooled sample data. Experimental scatter may reflect variations in the initial state/quality of the oocytes, in CPA concentration within oocytes introduced in the soaks, or in the (small) amount of surface liquid (in which ice is more likely to nucleate and grow).

Figure 5.. Evolution of oocyte diffraction during warming.

Time series of diffraction images acquired during warming of (a) the oocyte of Fig. 1(b), soaked in 50% vitrification solution, moved through oil to remove surface solvent, and “fast” cooled; and (b) the oocyte of Fig. 1(a), soaked in 100% vitrification solution and “fast” cooled. In (a), ice diffraction becomes visible at 50 ms, evolves from diffuse rings at cubic ice resolutions to sharp rings at cubic ice locations at 60 ms to rings at all hexagonal ice locations at 80 ms, and has disappeared at 110 ms. In (b), only one partial, extremely weak ice diffraction ring is observed, most likely arising from surrounding solvent, and for less than 20 ms.. Fig. S11 shows corresponding intensity vs resolution plots with DIFFaX fits.