The Effect of Dimethylsulfoxide on the Water Transport Response of Rat Hepatocytes During Freezing

Abstract

Successful improvement of cryopreservation protocols for cells in suspension requires knowledge of how such cells respond to the biophysical stresses of freezing (intracellular ice formation, water transport) while in the presence of a cryoprotective agent (CPA). This work investigates the biophysical water transport response in a clinically important cell type—isolated hepatocytes—during freezing in the presence of dimethylsulfoxide (DMSO). Sprague-Dawley rat liver hepatocytes were frozen in Williams E media supplemented with 0, 1, and 2 M DMSO, at rates of 5, 10, and 50°C/min. The water transport was measured by cell volumetric changes as assessed by cryomicroscopy and image analysis. Assuming that water is the only species transported under these conditions, a water transport model of the form dV/dT = f(Lpg ([CPA]), ELp ([CPA]), T(t)) was curve-fit to the experimental data to obtain the biophysical parameters of water transport—the reference hydraulic permeability (Lpg ) and activation energy of water transport (ELp )—for each DMSO concentration. These parameters were estimated two ways: (1) by curve-fitting the model to the average volume of the pooled cell data, and (2) by curve-fitting individual cell volume data and averaging the resulting parameters. The experimental data showed that less dehydration occurs during freezing at a given rate in the presence of DMSO at temperatures between 0 and −10°C. However, dehydration was able to continue at lower temperatures (<−10°C) in the presence of DMSO. The values of Lpg and ELp obtained using the individual cell volume data both decreased from their non-CPA values—4.33 × 10−13 m3 /N-s (2.69 μm/min-atm) and 317 kJ/mol (75.9 kcal/mol), respectively—to 0.873 × 10−13 m3 /N-s (0.542 μm/min-atm) and 137 kJ/mol (32.8 kcal/mol), respectively, in 1 M DMSO and 0.715 × 10−13 m3 /N-s (0.444 μm/min-atm) and 107 kJ/mol (25.7 kcal/mol), respectively, in 2 M DMSO. The trends in the pooled volume values for Lpg and ELp were very similar, but the overall fit was considered worse than for the individual volume parameters. A unique way of presenting the curve-fitting results supports a clear trend of reduction of both biophysical parameters in the presence of DMSO, and no clear trend in cooling rate dependence of the biophysical parameters. In addition, these results suggest that close proximity of the experimental cell volume data to the equilibrium volume curve may significantly reduce the efficiency of the curve-fitting process.