To the editor:

Guo et al.1 claim desiccation tolerance in human cells producing the disaccharide trehalose. Trehalose is well known to be associated with desiccation tolerance (anhydrobiosis) in many organisms, including baker's yeast and some resurrection plants. In yeast, trehalose is thought to be necessary for anhydrobiosis (e.g., ref.2).

The demonstration that trehalose alone is sufficient to confer anhydrobiosis would be a major step forward in the understanding of this remarkable phenomenon. This might be achieved using the approach of Guo et al., whereby a desiccation-sensitive cell (in this case, a human fibroblast) is made desiccation-tolerant by enabling it to synthesize trehalose. The strategy could be generally applicable to desiccation-sensitive cell types and might be termed “anhydrobiotic engineering.“

However, the techniques used by Guo et al.—for drying cells, measuring water content of the dried cells, and demonstrating the viability of the rehydrated cells—leave their conclusions open to question.

Human fibroblasts containing trehalose were dried after removal of medium by incubating the tissue culture plate sealed with parafilm at ambient temperature for one to three days. Surprisingly, the authors claim this method results in dried cells containing no detectable residual water, comparable with cells baked at 80°C overnight. However, the Fourier transform infrared spectroscopy (FTIR) technique used was not calibrated, nor were the limits of detection determined. Also, no distinction was made between extracellular and intracellular water. It is therefore not clear how dry the fibroblasts were, or indeed whether they were dry at all.

The key attribute of anhydrobiotic organisms is the ability to continue to grow after rehydration—viability, in other words. However, Guo et al. measure viability with a “live/dead” stain, a combination of calcein AM and ethidium homodimer. Calcein AM is membrane-permeable and is converted to fluorescent calcein by active intracellular esterases, whereas ethidium is excluded by intact membranes, but taken up through the damaged plasma membrane of a dead cell, when it intercalates into DNA and exhibits enhanced fluorescence. Live cells fluoresce green, while dead cells fluoresce orange.

This was the only viability assay used by Guo et al., but we feel it is not sufficiently reliable in this context—as amply demonstrated by the accompanying paper on the cryopreservation of trehalose-containing cells3. In Figure 2 of that paper (“no poration” and “WT” data sets), a large disparity was seen by Eroglu et al. between viability as measured by the same live/dead stain and viability as measured by either plating efficiency or growth of cryopreserved mouse fibroblasts. Cells apparently alive according to the staining technique do not attach to a surface or grow and can be considered dead.

What the live/dead stain is actually measuring is the activity of cytoplasmic esterases and the integrity of the plasma membrane. Trehalose is known to protect proteins and membranes (e.g., ref. 4) from desiccation damage, and it may be that some stabilization of these cell components was provided by the trehalose in the cells described by Guo et al. What we feel has not been demonstrated is true viability, i.e., growth.

Given these doubts about the validity of the data of Guo et al., it is difficult to agree with the authors that they have engineered mammalian cells that retain viability in the absence of water.