Dimethyl sulphoxide addition or withdrawal causes biphasic volume changes and its withdrawal causes t-system vacuolation in skeletal muscle

This letter addresses a methodological problem regarding the use of dimethyl sulphoxide (DMSO) in physiological solutions that is likely to influence the interpretation of some of the experimental results in a recent article (Lindinger et al. 2011). Indeed, examination of the literature reveals that at least one previous article in this journal may have been affected by the same issue (Yeung et al. 2002), suggesting the need to draw attention to this matter.

DMSO is commonly used as a solvent for poorly water-soluble drugs in physiological experimentation. Lindinger et al. (2011) employ it as a solvent for bumetanide, diluting a 20 mm stock solution of bumetanide in DMSO to produce an experimental solution containing 0.14 mm bumetanide and 0.7% DMSO; they also use it as a solvent for calcein-AM at a stated final concentration of 0.256%. The protocols employed do not control for any influence of addition or withdrawal of DMSO. Unfortunately, this is a significant omission for two reasons. First, DMSO is known to be semi-membrane permeant in other tissues and therefore causes biphasic volume changes on both addition and withdrawal (Pfaff et al. 1998). Second, 0.7% DMSO is in fact 98.6 mm, given that pure DMSO has a molecular mass of 78.13 g mol⁻¹ and a density of 1.10 g ml⁻¹ (Sigma-Aldrich catalogue, 2011).

In previously unpublished control experiments recording fibre volume changes using xz-plane confocal microscopy of whole Rana temporaria cutaneous pectoris muscles (as described by Ferenczi et al. 2004), I noticed that exposure of muscle fibres to Ringer solution containing 100 mm (0.71%) DMSO produces shrinkage and subsequent volume recovery similar to that seen on addition of 100 mm glycerol, though with a faster time-course of volume recovery (Fig. 1). Subsequent washout of DMSO with normal Ringer solution then produces swelling followed by volume recovery (not shown) and t-system vacuolation even with as little as 0.2% (28 mm) DMSO (Fig. 2), again similar to the influence of glycerol withdrawal. It can be noted that simultaneous withdrawal of ouabain and 1% (140 mm) DMSO has been shown to produce vacuolation of mouse muscle fibres, although this was ascribed to withdrawal of ouabain, rather than DMSO, by the authors (Yeung et al. 2002).

Figure 1.

A shows a confocal xz-plane image of the full thickness of a whole cutaneous pectoris muscle, bathed in Ringer solution containing lissamine rhodamine (62.5 μg ml⁻¹), obtained from Rana temporaria killed by concussion followed by pithing (Schedule 1: Animals (Scientific Procedures) Act 1986). Muscle length was held constant; thus, cross-sectional areas were measured from such images over time to provide (B) an accurate measurement of relative fibre volume during the addition of 100 mm DMSO (filled triangles) or 100 mm glycerol (open squares); (mean ± standard deviation of n = 4 fibres in each case). Both DMSO and glycerol produced rapid fibre shrinkage followed by slower but complete volume recovery. The recovery was slower in glycerol, suggesting that DMSO is more membrane permeable than glycerol.

Figure 2. Withdrawal of 0.2% (28 mm) DMSO produces pronounced t-system vacuolation, revealed by confocal microscopy of a fibre within a whole cutaneous pectoris muscle bathed in lissamine rhodamine (62.5 μg ml⁻¹).

The t-system origin of the vacuoles is clearly visible by inspection of the smaller vacuoles.

Unfortunately, these observations are of considerable relevance to any analysis of the experiments described by Lindinger et al. (2011). Thus, for example, their Fig. 4B demonstrates the volume changes resulting from the simultaneous addition of 35% hypertonic Tyrode solution and washout of bumetanide, following a 30 min exposure to isotonic bumetanide-containing Tyrode solution. This implies the simultaneous addition of ∼62 mm NaCl and withdrawal of ∼100 mm DMSO. It may also be noteworthy that Lindinger et al. describe an approximately 50 s delay between solution change and re-initiation of imaging, in which time both DMSO and glycerol addition had produced >10% fibre shrinkage in Fig. 1 here.

It is also possible that the calcein-loading procedure, which involved addition and withdrawal of DMSO, produced vacuolation of the muscle fibres. Exposure to hypertonic solutions might then produce devacuolation (Gallagher & Huang, 1997). The volume of the t-system in a vacuolated fibre can increase from its normal value of <0.5% to up to 10–15% (Krolenko & Lucy, 2001), and therefore it is possible that t-system volume changes could significantly influence the density of the intracellular calcein dye used in these experiments.

The probable influence of DMSO addition and withdrawal on cell volume and possibly on dye concentration during the experiments of Lindinger et al. (2011) makes analysis of some parts of their results very difficult. To my knowledge, the influence of DMSO addition and withdrawal has not been studied in mammalian muscle. Thus, appropriate control experiments are required to reveal which components, if any, of the observed volume changes were artefacts of changing DMSO concentrations. In addition, experiments involving the addition or withdrawal of pharmacological agents dissolved in a vehicle should maintain a constant level of that vehicle throughout.

The addition of such controls to the experiments of Lindinger et al. would be welcome, as the suggestion that the sodium–potassium–chloride cotransporter (NKCC) activity can produce significant regulatory volume increase in skeletal muscle fibres raises interesting questions about membrane potential control. For example, full recovery from exposure to a 125 mm hypertonic solution (Figs 3, 4 and 5 of Lindinger et al. 2011) would, if resulting entirely from NKCC activity, require intracellular Cl⁻ to increase by ∼60 mm. Assuming extracellular [Cl⁻] is 140 mm in the isotonic solutions and 200 mm in the hypertonic solutions, then a value of 6 mm for intracellular [Cl⁻] would give a reasonable value of −80 mV for E_Cl. A 60 mm increase in intracellular [Cl⁻] would then reduce E_Cl to just −28 mV. Given the normal high Cl⁻ permeability of skeletal muscle, one might therefore expect such a regulatory volume increase to be accompanied by very significant depolarization unless background ionic permeabilities were greatly changed.

In conclusion, experimental design must take into account that addition or withdrawal of DMSO from physiological solutions can influence cell volume and produce vacuolation of skeletal muscle.