Vero-B4 kidney cells showed a massive aggravation of cold-induced cell injury when stored at 4°C in RPMI 1640 cell culture medium, as compared to other cell culture media or KH buffer, and the combination of glucose, low calcium and high phosphate concentrations appeared to account for this phenomenon.
Mechanisms of cold-induced cell injury
Various cell types display iron-dependent cold-induced cell injury which is triggered by an increase in cytosolic chelatable iron ions [12, 31, 32]. Iron-dependent ROS formation leads to apoptotic and necrotic cell death via mitochondrial alterations, i.e. induction of the mitochondrial permeability transition (MPT) [13–16]. In all cell types used in this study, cold-induced cell injury could be inhibited by the addition of iron chelators, which indicates that it is mainly iron-dependent. In diverse endothelial cells, the extent of cold-induced lethal cell injury is dependent on the confluence state of the cell cultures, with late confluent cells being particularly prone to injury . On the other hand, subconfluent and early confluent cells are more susceptible to loss of cell-cell and cell-substrate interactions (S. Knoop, U. Rauen, unpublished results). The classical hypothesis of cold-induced cell injury proposes another mechanism based on sodium influx and cell swelling caused by inhibition of the Na+/K+-ATPase [19, 20]. This mechanism could, however, not be verified in adherent rat hepatocytes , and also appeared to play little role in the current study (practically no cold-induced injury in the sodium-rich KHG in the presence of deferoxamine, Figures 1, 7, Table 5).
Enhancement of cold-induced injury by RPMI
RPMI 1640 cell culture medium is the standard culture medium suggested by the German Collection of Microorganisms and Cell Cultures (DSMZ) for culture of Vero-B4 cells  and does not show any toxicity at 37°C. Therefore, it was surprising to find that this medium strongly enhanced cold-induced cell injury in these cells (Figure 1). The effect could not be attributed to sodium, as proposed in the classical mechanism, since aggravation did not occur in KH buffer or in other cell culture media with even higher sodium concentrations (see Table 3) but was specific to RPMI. The enhancement was, surprisingly, caused by the triple combination of glucose, low calcium and high phosphate concentrations and was iron dependent, as it was completely inhibited in the presence of iron chelators (Figure 1) and appeared to be mediated by MPT (inhibition by the MPT inhibitor combination tfp/fructose; Figure 6).
Role of calcium in MPT induction and cold-induced injury
Numerous factors have been discussed to trigger MPT or to sensitize mitochondria to MPT, thus leading to MPT induction either alone or in various combinations [34–38]. Amongst these dozens of factors are increased matrix calcium concentrations, high inorganic phosphate concentrations, decreased mitochondrial membrane potential, oxidation of pyridine nucleotides and of sulfhydryl groups, oxidizing agents and oxidative stress and a mitochondrial matrix pH around 7.4. While accumulation of calcium in the mitochondrial matrix has been and predominantly still is regarded as a prerequisite of MPT induction [35, 37], MPT has also been described to occur in the absence of major Ca2+ changes, especially when Pi is elevated [34, 35, 39].
It has been shown in various cell types that cytosolic and/or mitochondrial Ca2+ concentrations increase during cold ischemia/hypoxia or during subsequent reperfusion/reoxygenation [39–41]. Therefore, most organ preservation solutions contain no or very little Ca2+. Protection against anoxia-induced MPT by low extracellular calcium concentrations, associated with a decrease in mitochondrial matrix calcium content, was seen by Pastorino et al. . Also in pure hypothermic injury (without accompanying hypoxia) increases in cytosolic and mitochondrial calcium have been reported and related to cell injury [43–45].
However, Ca2+-free incubation aggravated cold-induced cell injury in rat hepatocytes [46–48] and liver endothelial cells  and Ca2+-free incubation was associated with increased ROS formation at 37°C . In line with this, addition of calcium to clinically used (phosphate-rich) preservation solutions reduced cell damage in rat livers [50, 51] and rat aorta , and decreased lipid peroxidation . Here, in the presence of glucose, low Ca2+ concentrations also increased cold-induced injury in Vero-B4 cells (Figure 3).
Role of phosphate in MPT induction
Increased concentrations of inorganic phosphate (Pi) are another well-known trigger for MPT [34, 35, 37, 53]. The deleterious effect of high matrix Pi concentrations has been explained by the buffering capacity of Pi yielding a matrix pH in favor of MPT [35, 37], by the ability of Pi to decrease the levels of ADP [37, 53], which is supposed to be a potent inhibitor of MPT, or by increasing ROS formation [34, 54, 55]. To promote MPT, phosphate apparently needs to enter the mitochondrial matrix [35, 37, 55]. In the present study, increased extracellular phosphate concentrations (5.6 mM) in KHG alone had no influence on cold-induced injury of Vero-B4 cells – only in combination with decreased Ca2+ concentrations (0.42 mM) we saw the aggravating effect (Figure 3). This finding is in contrast to the literature, where increased matrix concentrations of calcium are mostly regarded as prerequisite for MPT, although increased Pi levels appear to lower the threshold for Ca2+, even to physiological levels [34, 35, 37, 39, 55]. Potentially, in our setting, enhanced ROS formation during cold storage, likely further increased by low Ca2+ concentrations , sensitized the mitochondria for phosphate-triggered MPT.
Role of glucose in enhancement of cold-induced injury
During cold storage, iron-dependent injury in Vero-B4 cells was slightly (KH) or moderately (Ca2+- or Pi-modified KH) enhanced by addition of glucose (Figure 3). Lehnen-Beyel et al.  found that in L929 cells, addition of glucose caused an increase in intracellular levels of NADH which enhanced redox-cycling of iron ions and thus aggravated iron-dependent cell injury. Here, the addition of iron chelators inhibited cold-induced injury in all cell types, showing that the injury is also iron-dependent. Since cell lines tend to be highly glycolytic, increased reduction of nicotinamide adenine dinucleotides, i.e. increased availability of NADH, which fosters iron redox-cycling , might be the reason for the effect of extracellular glucose in Vero-B4 cells. In hepatocytes, in which no noticeable aggravation of cell injury was seen in the presence of glucose, endogenous glucose from glycogenolysis is likely to be available for metabolism also during incubation in glucose-free KH buffer. However, it should be noted that glucose did not only exhibit injurious features during cold incubation/rewarming but also appeared to be necessary as a substrate for the cells (Figure 1).
The aggravation of cold-induced cell injury by the triple combination seen here is thus likely to be caused by several interacting mechanisms: during cold storage, iron-dependent ROS formation is further increased due to both, the low Ca2+ concentration and simultaneously increased iron redox-cycling fostered by glucose via NADH availability. The increased ROS formation likely sensitizes the mitochondria for MPT. Additionally, MPT is promoted by increased concentrations of inorganic phosphate. Neither of the components alone nor in different combinations of two did approximate the level of cell injury seen for Vero-B4 cells in RPMI during cold incubation – the interaction of all three factors appeared to be necessary for the injurious effect.
Enhancement not specific for Vero-B4 cells but differences in sensitivity between cell types
The aggravation of cold-induced injury by a combination of glucose, low calcium and high phosphate, although not specific for Vero-B4 cells, seems to be particularly pronounced in this cell type. Porcine aortic endothelial cells also displayed an aggravation of cold-induced injury, but only in more extreme, but still clinically relevant conditions, i.e. the nominal absence of Ca2+ and presence of higher phosphate concentrations (0 mM/25 mM as in KHG(Ca--,P++)). In rat hepatocytes, similar phosphate concentrations also induced cell injury, but in these cells, an injurious effect of high phosphate was also seen at 37°C, in line with data described previously , and roughly doubled during cold storage (Figure 7).
Consequences for cell pausing media and organ preservation solutions
Not only RPMI 1640 cell culture medium and organ preservation solutions, but also many well-established buffer solutions (0.05 M phosphate buffer (50 mM phosphate, no calcium), phosphate-buffered saline (12 mM phosphate, no calcium)) display similar characteristics as the modified solutions used here. University of Wisconsin solution , which is used for organ preservation in the clinical setting, combines high Pi with nominal absence of Ca2+ (although in the absence of glucose) in concentrations that are identical with the concentrations we here used in KH(Ca--,PP++) and KHG(Ca--,P++). In Euro Collins solution, the Pi concentration is even higher and the solution contains glucose . Considering that hepatocytes and endothelial cells were severely damaged in this environment at 4°C, it should be considered to use phosphate-free/phosphate-poor solutions for cell, tissue and organ preservation. Also, the choice of pausing medium for cell cultures and the solutions used in processing steps performed at lower temperatures should be carefully made, considering that some of the solutions severely aggravate cold-induced cell injury. Addition of iron chelators provides significant protection in many solutions [23, 25, 32, 59–61]. However, with increasing storage time, differences between different base solutions become more distinct even in the presence of iron chelators [59, 61]. As iron chelators can cause iron depletion of cells [62, 63] and thus interfere with cell proliferation after rewarming, their concentration should be kept at the lowest effective concentration; therefore, enhancing/injurious effects of the base solutions should be minimized. Thus, we suggest using iron chelator-containing solutions based on buffers other than phosphate, such as recently described for cold storage of various cell types [59, 64, 65].
Comparison of cell culture media with cold storage solutions
The organ preservation solution UW has been adopted for short-term cold storage of cells with reasonably good results [66–68]. However, for various cell types, UW did not provide better protection than cell culture medium, and cold-induced cell injury in UW could also be greatly reduced by the addition of iron chelators [25, 32, 59]. In Vero-B4 cells, UW provided better protection than RPMI during one week of cold storage, but the protective effect was lost after longer cold storage periods (B. Akyildiz, U. Rauen, unpublished results). After two weeks of cold storage in UW solution plus three hours of rewarming, LDH release of Vero-B4 cells was about 50% whereas it was less than 10% after cold storage in ChillProtec and ChillProtec Plus, commercially available cell storage solutions. However, the focus of the current study was to understand the surprising RPMI effect, not to compare or further optimize cold storage solutions.
Limitations of the current study
Limitations of the current study are the relatively short cold storage period, the relatively short follow-up period and the lack of comparison to different cold storage solutions. The storage period of one week and the short follow-up period of 3 h rewarming were chosen to study the disastrous RPMI effect, which is already marked at these time points. However, the short rewarming period does not account for late apoptosis or for proliferative dysfunction of the surviving cells after cold storage. Optimization of cold storage solutions in addition requires longer cold storage periods and proper comparison with the different cold storage solutions available and is currently in progress.