While many aspects of these developments were used in previously documented protocols [2
]; however, the reported results differ widely among labs and cell lines. Our clear distinction and innovation was to use the cumulated data in concert as a whole. Another problem is that a wide variety of the methods of assessment of recovery used in the abovementioned publications makes direct quantitative comparison of different results very difficult if not sometimes impossible. We have decided not to use any colony assay due to large variability of the size and amount of the colonies that can be plated into a particular well, thus, making the “number of colonies
” as an unreliable parameter for quantification. We instead, used the number (% to the total amount of harvested cells) of GFP-positive cells that excluded a negative viability marker 7AAD as the characteristic of the attached cells. We also precisely (using QUANTA Coulter counter option) measured the amount of nonattached and attached cells because it was also impossible to plate the same amount of cells per a well even if the cells were dissociated prior to freezing and plating, and especially if they were frozen in clumps. Thus, combination of the assessment of % the cells that reattached to the surface of the well versus floating non-attached cells and the percentage of Oct4-positive viable (7AAD-) reattached cells would give us a fair assessment of the overall recovery, that is, the yield of viable pluripotent SCs after CP (VY
). If that parameter is normalized to the yield of nonfrozen cells that are split according to the standard protocol (in clumps), that would constitute the assessment of cryosurvival (SR
, which also includes survival to CP-related protocols such as addition and removal of the CPAs and centrifugation). Those two parameters (VY
) would give an objective assessment of a particular CP protocol as the whole.
We previously hypothesized that DMSO, as a potent differentiating agent, can be intrinsically toxic for hPSCs and should be substituted with a non-toxic CPA [1
]. And indeed, DMSO has been found to be toxic to ESCs even in miniscule concentrations if applied for long time ([19
], our unpublished data
). In a pilot study, we found that EG provided equal to DMSO protection in cryopreservation of 293 cells [25
]. We have also reported that EG was less toxic to 2-cell mouse embryos [26
] while it manifested similar permeability characteristics (see [27
] for references). It made EG as an attractive alternative to DMSO so we first checked the potential toxicity by exposing iPSCs to 10% v/v concentrations of DMSO, EG, and two other widely used in cryopreservation CAPs propylene glycol and glycerol. We found that the toxicity DMSO was substantially higher than that of the 3 other cryoprotectants: 30-min exposure in 10%- DMS caused loss of 30% of initially plated pluripotent cells in comparison to 5% in control untreated with the agent cells. Ethylene glycol was less toxic (76% of viable Nanog+ cell were recovered), which was comparable with PG. Glycerol was the least toxic with recovery comparable with control (P
> .1) so the relative toxicity could be expressed as DMSO > >EG ~ PG > GLY (). DMSO has been well known as a powerful inducer of differentiation and apoptosis [28
We also found that DMSO is substantially more toxic to hESC than EG, PG, and GLY and comparable to ethanol and acetaldehyde (a paper will be published elsewhere). In contrast, in general EG is considered less toxic in cryobiological literature, especially at high concentration used for vitrification and somewhat similar to PG, even though the data are controversial. In the majority type of cells, glycerol is considered the least toxic while at the same time, it is the least permeable. In general, a permeable CPA with smaller molecular weight such as DMSO, methanol, or ethanol is the most permeable and more most toxic than CPAs with higher MW such as glycerol. We have found similar effects for 2-cell mouse embryos, and for hESCs, as well as in this study. However, application of the standard freezing protocol (cells frozen in clusters in cryovials) using the same 10% v/v concentration (though slightly different osmolality) revealed that the cell recovery rate was as follows: EG ~ DMSO > PG > >GLY. (Figures and , group “STD”).
Cryopreservation of dissociated hPSCs
has recently drawn attention after publications on the use of Accutase as a less harmful detaching and dissociating agent [31
] and especially after Watanabe et al. reported the protective role of a Roh-associated kinas (ROCK) inhibitor Y-23672 after hES cell detachment and dissociation [32
]. In all recently available reports [8
] the authors found beneficial role of cryopreservation of hES and iPS cells after dissociation and application of Y-23672 even though direct comparison of the efficiency is not possible mostly due to different methods of estimation of viability and recovery of pluripotent cells and PS cell colonies. The mechanism of the protective role of RI can be as inhibitor of apoptosis that was caused by the activation of the ROCK pathways during dissociation and detachment, pointed out in the original Watanabe's work; at the same time, freezing (and cooling) per se
can cause specific cold and osmotic related damages such as rearrangement of F-actin, activation of p53, and ROS production [36
]. Overall, as we can see in , CP of dissociated cells had clear advantage over the standard CP in clumps, especially if EG was used (51% versus 8% resp., P
< .001), in a range reported by other investigators as well as in good concordance with our data on Oct4/eGFP-infected human ESCs (results will be published elsewhere). We did not explore addition of Y-23672 one hr prior freezing for CP of iPSCs as others did [10
] but we found it somewhat more beneficial for CP of Oct4/eGFP-infected human ESCs in comparison to addition of Y-23672 after freezing only.
Cryopreservation of adherent stem cells in monolayers in plates is a relatively novel method. Beside some anecdotic reports on freezing mouse ESC plates, and recent work by Nagy and colleagues [37
] and Miyamato et al. [39
], very few reports are published on CP of adherent human pluripotent SCs [7
]. We believe that CP in adherent state is not only convenient for several applications such as high cell-based throughput and high-content screening but it also causes less damage so we (CELLTRONIX) developed in 2006–2008 a “universal” technological approach that we call ComfortFreeze®
, that can be applied not only to the pluripotent stem cells but hESC/iPSC-derived neural and cardiac progeny as well (results to be published elsewhere).
Efficient, robust, and customer-friendly CP technologies for primary hPSCs would substantially shorten the time to restore cryobanked colonies, improve shipment safety, eliminate possible biased selective pressure within a PSC-line after CP and distribution of cells frozen in 96-well plates that could be used immediately for embryotoxicity and drug screening in PSC-based toxicology in vitro kits.