SESSION 10

CRYOBIOLOGY - IV

Chairman: J. Karlsson

INTRODUCTION

To understand the mechanisms of both the freezing injuries of cells and the protection of cells due to cryoprotectants in cryopreservation of cells, the microscopic behavior of the ice crystal and the cells during the freezing of cell suspension has been actively studied using optical and electron microscopes. However, the freezing of cell suspensions generally proceeds transiently and spatially in three-dimensions(3D). Therefore, 3D observation in real time is required to understand the details of microstructure in the freezing process of cell suspensions. For such observation, optical and electron microscopes have limitations because these microscopes produce only 2D images of the sample. Furthermore, the required fixation of the sample does not allow observation of the same sample. This lack of information means that presumptions must be made concerning the phenomena that occur during the freezing.

A confocal laser scanning microscope (CLSM) with a high-performance computer was recently developed. This is a noninvasive method that produces optical tomograms of biological materials without fixation and slicing of a sample. In this study, the behavior of ice crystals and human red blood cells during extracellular-freezing was visualized in 3D in real time using a CLSM and a fluorescent dye. The influence of the addition of cryoprotectant and the cooling rate were investigated on the morphology of ice crystals and the interaction between ice crystals and cells near the freezing interface.

MATERIALS AND METHODS

Suspensions of human red blood cells stained with a fluorescent dye were frozen. For the solutions, physiological saline (PS) alone and physiological saline with 2.4M glycerol (PS+Gly) were used. For fluorescence observation, a fluorescent dye, acridine orange (AO, CH₂₀N₃Cl, molecular wight=301.82), was added to the solutions.

A sample was set between a glass microslide and a glass coverslip, and then was frozen on the surface of the directional solidification stage. This method allows us to independently control the temperature gradient G and advancing velocity V of the gradient at the freezing interface, resulting in the cooling rate H=GV. At a constant G=13.5^oC/mm, H was varied in the range up to 20.0^oC/min.

A Leica erect-type CLSM (TCS4D) was used. This has an Ar-Kr (argon-krypton) laser (488nm, 568nm and 643nm in wavelength) and two photomultipliers. The scanning of laser beam on a horizontal section combined with the vertical movement of the stage with a sample on it produced continuous, noninvasive, 3D optical tomograms.

The AO has a maximum-excitation wavelength of 492nm(blue), and two maximum- emission wavelengths of 530nm(green) and 640nm(red). It is a monomer in solution and becomes a dimer when taken into a cell membrane. The monomer and dimer generate green and red fluorescences, respectively. The ice does not generate fluorescence because during the freezing the dye is not trapped in ice. Therefore, the ice crystal, cells, and unfrozen solution can be distinguished with different colors.

RESULTS AND DISCUSSION

The morphology of the ice crystals depends on the cooling rate and the concentration of chemical additives. For the range of the parameters studied, optical microscopy shows that the ice crystals in PS typically have two morphologies in 2D: a flat interface and a cellular interface, and that the ice crystals in PS+Gly typically have a dendritic appearance. For these typical morphologies, the 3D microstructure was measured by the CLSM system.

Freezing of Red Blood Cells in Physiological Saline (PS)

The ice crystals have a flat interface at lower cooling rates and a cellular interface at higher cooling rates than the critical cooling rate for transition from a flat one to a cellular one. The interface which appeared flat in 2D at H=3.0^oC/min, had a wedge-shaped interface in 3D, whose upper part protruded into the higher-temperature region. The unfrozen solution remained in a thin layer on and under the ice. The solid-liquid interface advanced into the unfrozen solution with red blood cells dispersed in the bottom region of the sample while maintaining nearly the morphology. The red blood cells were pushed by the interface, causing them to accumulate in the unfrozen solution in front of the interface. These accumulated red blood cells were trapped in the unfrozen solution between the ice crystals and the upper or lower glass substrate, and then became flatter in the colder region.

An increase in cooling rate at a constant temperature gradient promotes constitutional supercooling in the unfrozen solution adjacent to the freezing interface, resulting in morphological instability of the flat interface in 2D. The instability changes the interface from flat into cellular. As the cellular ice crystals grew in the unfrozen solution with the red blood cells in the bottom region, most of the cells were pushed aside by the cellular ice crystals and thus accumulated and compacted in the unfrozen solution between the finger-like ice crystals. Each finger-like ice crystal in the horizontal section appeared as a cluster of ice in the vertical section of uniform temperature. The clusters of ice were arranged at roughly periodic intervals and were inclined to the vertical direction. The change of the shape of the cross section in the direction of temperature-decrease clearly indicates that the growth of ice crystal in the vertical section was anisotropic.

Freezing of Red Blood Cells in Physiological Saline with Glycerol (PS+Gly)

Secondary and tertiary arms of the dendritic structure are clearly seen in the horizontal section. The red blood cells remained dispersed during the freezing. This independence of the dispersion of red blood cells results from the "flexible" interaction between ice crystals and cells. In this "flexible" interaction, as the ice crystal grows, the morphology of the ice-solution interface changes in response to the shape of cells, and the cells are almost not moved by the interface. This contrasts with the interaction between ice crystals and cells in PS.

The microstructure in the vertical section of uniform temperature in PS+Gly has the following similarities with the cellular ice crystals in PS: 1)arrangement of clusters of ice crystals at nearly periodic intervals and with an inclination, and 2) anisotropic growth of ice. However, a few cracks and red blood cells were also in the clusters of ice crystals in the vertical section. The cracks were filled with unfrozen solution and corresponded to the narrow space between the primary arm and the branch of the dendrites. At a higher cooling rate, the branching of ice happens more frequently, and correspondingly the number of cracks in the vertical section of ice increased. The outline of the ice was not only wavy but also conformed locally to the shape of the red blood cells whereas the outline of ice in PS was a monotonic curve.

Ice Fraction in the Freezing Region near the Tip of Ice Crystal

Ice fraction, defined as the ratio of the area occupied by the ice in each vertical section, was estimated from a series of vertical tomograms for the cellular ice crystals in PS and the dendritic ice crystals in PS+Gly. The ice fraction increased with distance and temperature-decrease from the tip of ice crystal, reaching near saturation in each cell suspension. The ice fraction in PS has a larger value than that in PS+Gly due to the lower initial-concentration of solute in the PS than in the PS+Gly. For reference, the ice fraction was also calculated from the liquidus of solution at thermal equilibrium. For PS, the experimental value tends to be slightly smaller than the calculated value, while for PS+Gly, the values are similar. The cause may be mainly that the mass transfer in the direction of temperature-decrease in PS proceeded due to the diffusion during the freezing more rapidly than in PS+Gly and consequently the solute-concentration at the tip of ice crystal became higher than the initial value.

SUMMARY

The 3D behavior of ice crystals and red blood cells during the extracellular-freezing was made clear for all flat, cellular, and dendritic solid-liquid interfaces, and the ice fraction in the freezing region was obtained. The results also indicates that using a CLSM with a fluorescent dye is very effective for the purpose of this study.

EXTENDED ABSTRACT

In this paper we report on our attempt to recover unprotected erythrocytes frozen at 10,000 ^oC/min. To this end we vitrified the intracellular solution of frozen erythrocytes using a laser pulse before thawing the sample. Electron microscopy studies have already shown that unprotected erythrocytes frozen at the rate of 10^{4 o}C/min have substantial IIF;¹ and numerous studies, including ours, indicate that rapid warming of erythrocytes frozen at this rate results in 100% hemolysis. The hypothesis driving this work, however, is that the formation of ice crystals during the rapid freezing of cells is completely innocuous, but the kinetics of crystal growth and reformation during warming increase dramatically as the ice crystal fraction within the cell increases. The apparent existence of a level of lethal IIF is really a kinetic barrier beyond which so called "rapid warming" is still too slow to prevent lethal damage during warming. By vitrifying the intracellular solution prior to thawing the cells, we can partially decouple the damaging effects of crystal formation from the damage incurred during warming.

The technique used to vitrify the intracellular solution is to use a pulse from a Q- switched laser at a wavelength that selectively targets the intracellular solution. This allows for the intracellular solution to be "instantaneously" melted while the surrounding solution remains at cryogenic temperatures. Conductive cooling of the intracellular solution results in cooling rates of more than 1 million degrees per second which is fast enough to vitrify pure water.² The efficacy of this technique in creating amorphous ice and a numerical analysis of the resulting cooling rates has been reported elsewhere.³ Erythrocytes were chosen for this study because of their naturally occurring chromophore (hemoglobin) which allows them to be selectively targeted by 532 nm. radiation when they are frozen in a PBS solution. The pulse width was 7 ns.

Human blood was diluted with PBS and loaded in cryostraws. The cells were frozen by immersion in a methanol slurry and subsequent transfer to liquid nitrogen. After the administration of the laser pulse the straws were thawed by immersion in warm water in a manner identical to control experiments. Different energy fluxes were administered to the straws to study the effect of energy density on hemolysis and to find the optimal energy flux. The diameter of the incident beam was held constant at 3mm.

All of the cells that were frozen using the described protocol and were then thawed without laser treatment lysed. In samples that were laser irradiated up to 80% of the cells remained intact after thawing. Studies were performed to determine the percentage of lysed cells as a function of laser energy flux. Below a certain threshold energy (2.5 J/cm²) all of the cells lysed. For energies between 3 and 6 J/cm² the percentage of cells that did not lyse increased roughly monotonically. As the incident energy was increased beyond 6 J/cm² the percentage of non lysed cells decreased rapidly, but appeared to stabilize at about 25%.

The cells that did not lyse after laser vitrification and thawing exhibited intact membranes that were stable over time. The cells were osmotically active; and in particular erythrocyte ghosts could be formed by exposing them to a hypotonic solution just as is true for control cells. The cellular retention of hemoglobin, and proper cell morphology were verified by using Wright's stain. The eosin in Wright's stain binds to hemoglobin. Dark staining of the cells indicated clearly that they retained their hemoglobin. This was also confirmed by the observation of red spots in centrifuge tubes after the cells were spun down. The morphology of the cells was excellent. The cells maintained their discoid shape, and showed no signs of crenelation.

A number of controls were run to prove that the beneficial effect of the laser irradiation was the vitrification of the intracellular solution. Identical procedures were used for the freezing and laser irradiation of fibroblasts, which do not selectively absorb the radiation; and the laser pulses had no effect on their survival. In addition erythrocytes were exposed to laser pulses with much longer pulse widths, thereby resulting in much slower cooling rates; and they also experienced no beneficial effect from the laser pulses. Only 7 ns. pulses of radiation that selectively targeted intracellular solution provided protection form cell lysis during warming.

These findings do not prove that a new and improved cryopreservation technique for unprotected erythrocytes can be developed based on laser recovery. They do indicate, however, that the primary damage to cells that experience large amount of IIF may be solely a result of damage incurred during the warming process. They also indicate that this damage can be avoided if the cells are warmed fast enough.

1. Nei, T. (1976), "Freezing injury to erythrocytes I. Freezing patterns and post-thaw hemolysis," Cryobiology, v. 13, pp. 278-286.
2. Hallbrucker, A., Mayer, E. and Johari, G.P. (1989), "Glass-liquid transition and the enthalpy of devitrification of annealed vapor-deposited amorphous solid water. A comparison with hyperquenched glassy water," Journal of Physical Chemistry, v. 93, pp. 4986-4990.
3. Fowler, A.J. and Toner M. (1997), "Cryopreservation of Cells using Ultra-Rapid Freezing," Advances in Heat and Mass Transfer in Biotechnology, BED-Vol. 37, pp. 179-183.

ABSTRACT IS NOT RECEIVED.

14 small metal spheres of diameter 0.3~3.0mm are plunged into subcooled LN₂ at speeds of 0~6m/s,and the moving distances inside the LN₂ are 5cm. As the results of the increasing of the subcooled degree, or the decreasing of the diameter of sphere, the curves of boiling heat flux or cooling rate would be shifted upward and rightward dramatically. As the subcooling is 13K, the heat flux is up to 1.4 10⁶W/m², the cooling rate is also up to 8200 K/s at the CHF point of boiling curves for sphere of diameter 0.287mm. It is resulted that if the time of moving inside the liquid nitrogen is not longer than the time required for forming stable vapor in the liquid, the quenching boiling heat transfer is not influenced by the quenching speed. And when the subcooling is large enough, or the sphere diameter is small enough, the curve shapes of the boiling heat flux would be changed from "M" style to "L" style. As the subcooling beyond 13k, "L" shape boiling curves are recorded for sphere of diameter 0.287mm.