Supplementary Figure 14: Representative trace and average data for the astrocyte membrane potential in different concentrations of bath K⁺.

The slope of the line shows a 55 mV depolarization for a 10-fold change in K⁺ concentration. Additional Note: A salient feature across mouse models of HD is that MSNs display depolarized membrane potentials by as much as ∼12 mV. A vaguely formulated idea suggests that reducing astrocyte Kir4.1 function with Ba²⁺ in WT slices may reveal secondary consequences for MSNs, i.e., this approach will elevate K⁺ near neurons, depolarize MSNs and thus reveal astrocyte Kir4.1 contributions to altered MSN properties frequently observed in mouse models of HD. However, this view has problems. First, although Ba²⁺ is a useful tool to isolate astrocyte Kir4.1 currents, it also blocks MSN K⁺ channels, meaning that it will be impossible to tell if any measured effects are due to Ba²⁺ actions on astrocytes or MSNs directly. Second, the suggested experiment can only work if brain slices can buffer K⁺ locally, i.e. that local K⁺ is not set by the bath K⁺ concentration. We tested for this by measuring the astrocyte membrane potential in bath solutions of different K⁺ concentration, exploiting the fact that astrocytes function as K⁺ microelectrodes because of their high K⁺ conductance. The Nernst equation predicts that the membrane potential of an astrocyte with a high K⁺ conductance should change by ∼56 mV for a ten-fold change in bath K⁺ concentration. If brain slices can in fact buffer K⁺ in the vicinity of a cell, then we would predict a deviation from this, because the local K⁺ concentration would be set by buffering and not by the bath K⁺ concentration. We found that striatal astrocytes showed a near perfect Nernst-like ∼55 mV depolarization for a 10-fold change in bath K⁺ concentration. These data show that isolated preparations such as brain slices cannot buffer K⁺. Thus, the idea of blocking astrocyte Kir4.1 channels with Ba²⁺ and looking for downstream effects on MSNs because of disrupted local K⁺ buffering is flawed.