Date Published: October 10, 2018
Publisher: Public Library of Science
Author(s): Jan Huebinger, Mária A. Deli.
Rapid cooling and re-warming has been shown promising to cryopreserve living cells, which cannot be preserved by conventional slow freezing methods. However, success is limited by the cytotoxicity of highly concentrated cryoprotective agents. Recent results have shown that cryoprotective agents do not need to suppress intracellular ice crystals completely to allow for survival after cryopreservation. Cryoprotective agents like DMSO or ethylene glycol can also lead to a tolerance of cells towards intracellular ice. It is however unclear by which mechanism this tolerance is achieved. These substances are also known to modulate properties of cellular membranes. It is shown here that cryoprotective DMSO and ethylene glycol have a clear influence on the mobility of lipids in the plasma membrane of HeLa cells. To isolate changes of the properties of plasma membranes from effects on ice formation, the membrane properties were modulated in absence of cryoprotective agents. This was achieved by changing their sterol content. In cells with elevated sterol content, an immobile lipid fraction was present, similar to cells treated with DMSO and ethylene glycol. These cells showed also significantly increased plasma membrane integrity after rapid freezing and thawing in the absence of classical cryoprotective agents. However, their intracellular lysosomes, which cannot be enriched with sterols, still got ruptured. These results clearly indicate that a modulation of membrane properties can convey cryoprotection. Upon slow cooling, elevated sterol content had actually an adverse effect on the plasma membranes, which shows that this effect is specific for rapid, non-equilibrium cooling processes. Unraveling this alternative mode of action of cryoprotection should help in the directed design of novel cryoprotective agents, which might be less cytotoxic than classical, empirically-found cryoprotective agents.
Cryopreservation, i.e. the potentially infinite storage under very cold temperatures, of living cells is of fundamental interest for biomedical research, clinical application and the preservation of endangered species. Classical slow cooling cryopreservation works by extracting water from the cells and thereby constraining ice crystallization to the extracellular medium . This is accompanied by a massive shrinkage of the cells and success of reversibility depends on energy demanding adaptation by the cells . Immortalized laboratory cell lines are usually well adapted to this, but many other cell types do not tolerate this. Therefore, rapid cooling and re-warming (often termed vitrification) is a very promising approach for the cryopreservation of cells that cannot be efficiently preserved by slow cooling approaches (e.g. [3,4]). However, this approach suffers from toxicity of the relatively high concentrated cryoprotective agents that need to be applied to the cells at temperatures above 0°C [1,5]. These cryoprotecants were thought to be necessary to avoid ice-crystallization in cells, since ice-crystals were–in analogy to slow freezing approaches–considered to be absolutely lethal [1,5]. However, in a recent study we showed that ice-crystals actually form during some of these applications, which nevertheless allowed for very high survival rates . Based on this, the term vitrification is not strictly correct for such applications, because it would imply the complete suppression of ice crystallization. These approaches are therefore called rapid-cooling and rewarming approaches here. Using such approaches, the total amount of ice or the number of ice crystals did not correlate with an increase of cell death, demonstrating that intracellular ice crystallization is not lethal upon fast cooling and warming. However, cell death occurred when samples were slowly warmed and ice could re-crystallize to fewer but bigger ice-crystals . This correlation does not prove causality between re-crystallization and cell death. Yet, it reopens the question of the cause of cell death and with that also the mode of action of cryoprotective agents. The amount of tolerable re-crystallization is dependent on the type of cryoprotective agents used . This clearly indicates that the cryoprotective effect is not solely prevention of ice nucleation or re-crystallization. The cryoprotective agents apparently provide protection against the harmful effects, which at least coincide with re-crystallization. The two most frequently considered types of cryodamage are direct damage by ice crystals to cellular membranes and high solute concentration in the unfrozen fraction around the ice crystals, which could lead to the denaturing of cellular proteins or damage to lipid bilayer membranes [1,7]. However, lipid bilayer membranes themselves also undergo phase transitions and structural changes upon cooling [8,9], which have been associated with cold shock damage in sperm cells . In all of these cases, membranes are a target for cryodamage. Small polar molecules like DMSO, glycerol or ethylene glycol can modulate the hydration layer of membranes , which changes their properties at subzero temperatures [8,12]. These substances generate also a high tolerance against re-crystallization after rapid cooling . It is therefore conceivable that they convey cryoprotection to the membranes under cryo conditions. Here, I tested therefore, if these cryoprotective agents change membrane properties and if membranes can be cryoprotected by modulating only their properties, i.e. without inhibition of ice crystallization. Modulation of plasma membranes by increasing sterol content, resulted in a clear increase in resistance to cryodamage. Lysosomes, which cannot be enriched with sterols , were still found ruptured. Additionally, fluorescence recovery after photobleaching (FRAP) analysis in living HeLa cells, showed a similar partitioning effect on the plasma membrane by cryoprotective agents and sterol enrichment. On the other hand, a marked denaturing of cytosolic proteins–as it has been observed upon relatively slow and lethal cooling –was not detectable by circular dichroism spectroscopy after lethal rapid cooling and rewarming. The observed improved resistance to cryodamage by increasing sterol content was specific for rapid cooling processes.
To gain understanding about the effect that a medium of 15% DMSO and 15% ethylene glycol (DE-medium) has on membranes of living cells, living HeLa cells were labeled with the fluorescent lipid DiOC18 and treated with DE-medium. Then a small section of the plasma membrane of these cells was bleached to perform fluorescence recovery after photobleaching (FRAP) experiments. In cells treated with DE-medium the fluorescence did not fully recover, indicating the presence of an (on the timescale of the experiment) immobile phase within the plasma membrane, which was not present in untreated cells (Fig 1A). The diffusion speed in the mobile fraction was however unaltered (Fig 1B).
It has been assumed for a long time that the success of cryopreservation depends on the avoidance of intracellular ice crystals [1,5]. However, we found previously that rapidly-cooled cells can actually tolerate intracellular ice crystals . The results presented here show that under these non-equilibrium cooling conditions cellular membranes are a major target for cryodamage. A massive denaturing of proteins, as it was observed upon relatively slow lethal cooling , was not observed under lethal rapid cooling conditions. Importantly, the damage to plasma membranes can be suppressed by solely changing their properties, without interfering with ice formation. This shows that there is a cryoprotective effect that is independent of the suppression of ice formation. In this context, it is of note that the mechanism of cryoprotection is specific for rapid cooling approaches, which prevent the samples from reaching thermodynamic equilibrium during cooling. Upon slow cooling, membrane integrity was actually adversely affected by sterol-loaded MβCD (Fig 2D) and led to strong and irreversible denaturing of proteins . This indicates that there is considerable difference in the mechanism of cryodamage between equilibrium (i.e. slow-cooling) and non-equilibrium cooling processes.