Technical Comments

Comment on “Local impermeant anions establish the neuronal chloride concentration”

See allHide authors and affiliations

Science  05 Sep 2014:
Vol. 345, Issue 6201, pp. 1130
DOI: 10.1126/science.1255337


Glykys et al. (Reports, 7 February 2014, p. 670) proposed that cytoplasmic impermeant anions and polyanionic extracellular matrix glycoproteins establish the local neuronal intracellular chloride concentration, [Cl]i, and thereby the polarity of γ-aminobutyric acid type A (GABAA) receptor signaling. The experimental procedures and results in this study are insufficient to support these conclusions. Contradictory results previously published by these authors and other laboratories are not referred to.

Glykys et al. (1) performed experiments with the Cl indicator Clomeleon, which has a Kd of ~90 mM (2, 3). Therefore, most of the data points (the ones with low [Cl]i) were obtained in a range, in which it is difficult to reliably calculate [Cl]i from the yellow fluorescent protein/cyan fluorescent protein (YFP/CFP) ratio. The low Cl sensitivity of Clomeleon at low [Cl]i and the potential temporal instability of fluorescent signals (CFP, YFP, background, and the like) require thorough control measurements, which are not provided in this publication. Moreover, despite the broad distributions of [Cl]i in the present study (from almost 0 mM to >50 mM), in many cases the distributions presented for individual experiments are much narrower [figures 3, B and I, and S6B in (1)].

The authors report nearly identical [Cl]i values for immature and adult hippocampal neurons, which is in contradiction to numerous previous reports using electrophysiological measurements of γ-aminobutyric acid (GABA) reversal potentials [e.g., (4, 5)] or fluorimetric [Cl]i measurements with Clomeleon [e.g., (2, 3)]. The authors also report that inhibition of the cation-Cl transporter NKCC1 with bumetanide does not significantly influence [Cl]i, which is in disagreement with previous studies in comparable age groups demonstrating that inhibition of NKCC1 reduces [Cl]i [e.g., (3, 6)] and attenuates the depolarizing effect of GABA (710). Even in the current study, bumetanide caused a substantial reduction in basal [Cl]i (from 16.6 ± 14.6 to 7.1 ± 10 mM) [figure 3, H and I, in (1)], but this effect is not discussed appropriately. Furthermore, a variety of studies demonstrated that overexpression, knockdown, or knockout of the Cl transporters KCC2 and NKCC1 affects [Cl]i homeostasis [e.g., (1113)]. We postulate that the discrepancies between the data obtained in this present study and previously published results must be discussed in this paper.

We also wonder why the ouabain effect on the cellular volume was determined with the Na+-sensitive dye sodium-binding benzofuran isophthalate (SBFI) and not by Clomeleon fluorescence, although the latter was used by the authors to measure volume changes induced by hyperosmotic solution and epileptiform activity [figures 3, F and G, and 4 in (1)], as well as in previous papers by these authors (3, 14). Measurements of [Cl]i during ouabain application would be an ideal test for the main hypothesis of the authors. In addition, to verify “that extracellular Na+ does not exert a significant Donnan effect,” it must be demonstrated that ouabain affects the intracellular Na+ concentration, which is not shown in the manuscript. Finally, blocking Na+/K+ adenosine triphosphatase (ATPase) massively interferes with the driving forces for many secondary active transporters, which is not adequately discussed in the current Report.

To test prediction 2, the authors use weak organic acids to increase the intracellular concentration of anions. However, d-lactate and pyruvate permeate the membrane in the uncharged form and dissociate inside the cell, with the vast majority of the protons being buffered intracellularly (15). The resulting decrease in the concentration of intracellular buffer anions will thus roughly balance the additional amount of organic anions introduced by the weak acids, leaving the total intracellular charges mostly unchanged. In contrast, and unlike what was stated by Glykys et al., gluconate is a membrane-impermeable weak acid. Thus, the observation that this weak acid induces the same effect as the membrane-permeable weak acids d-lactate and pyruvate further disproves that this effect can be attributed to an accumulation of intracellular anionic charges.

The [Cl]i increase induced by epileptiform activity in organotypic slice cultures [figure 4 in (1)] has already previously been reported by the authors and has been linked to seizure-induced elevations of [K+]e, shifting the driving force on NKCC1 (14). Thus, this finding is not new and nicely supports the conventional hypothesis of [Cl]i homeostasis.

In our opinion, the experiments shown by Glykys et al. (1) do not provide convincing data to support the hypothesis that fixed anionic charges establish the transmembrane Cl distribution. Astonishingly, Glykys et al. fully ignore the majority of previous observations on the role of KCC2 and NKCC1 in neuronal Cl regulation (including their own previous papers), which are in contradiction with their current hypothesis. Therefore, we are afraid that the authors’ major hypothesis is in disagreement with the newly provided experimental evidence, as well as with previously published data.


Stay Connected to Science

Navigate This Article