Nalcn Is a "Leak" Sodium Channel That Regulates Excitability of Brainstem Chemosensory Neurons and Breathing

Abstract

The activity of background potassium and sodium channels determines neuronal excitability, but physiological roles for "leak" Na(+) channels in specific mammalian neurons have not been established. Here, we show that a leak Na(+) channel, Nalcn, is expressed in the CO2/H(+)-sensitive neurons of the mouse retrotrapezoid nucleus (RTN) that regulate breathing. In RTN neurons, Nalcn expression correlated with higher action potential discharge over a more alkalized range of activity; shRNA-mediated depletion of Nalcn hyperpolarized RTN neurons, and reduced leak Na(+) current and firing rate. Nalcn depletion also decreased RTN neuron activation by the neuropeptide, substance P, without affecting pH-sensitive background K(+) currents or activation by a cotransmitter, serotonin. In vivo, RTN-specific knockdown of Nalcn reduced CO2-evoked neuronal activation and breathing; hypoxic hyperventilation was unchanged. Thus, Nalcn regulates RTN neuronal excitability and stimulation by CO2, independent of direct pH sensing, potentially contributing to respiratory effects of Nalcn mutations; transmitter modulation of Nalcn may underlie state-dependent changes in breathing and respiratory chemosensitivity.

Significance statement: Breathing is an essential, enduring rhythmic motor activity orchestrated by dedicated brainstem circuits that require tonic excitatory drive for their persistent function. A major source of drive is from a group of CO2/H(+)-sensitive neurons in the retrotrapezoid nucleus (RTN), whose ongoing activity is critical for breathing. The ionic mechanisms that support spontaneous activity of RTN neurons are unknown. We show here that Nalcn, a unique channel that generates "leak" sodium currents, regulates excitability and neuromodulation of RTN neurons and CO2-stimulated breathing. Thus, this work defines a specific function for this enigmatic channel in an important physiological context.

Keywords: Phox2b; respiratory chemoreceptors; retrotrapezoid nucleus; substance P; ventilation.

Figure 1.

Nalcn is expressed at variable levels in RTN neurons. A, In situ hybridization for Nalcn transcripts and immunohistochemistry for GFP in RTN neurons from Phox2b::GFP mice. B, GFP-labeled RTN neurons acutely dissociated from acute slices (left) were harvested for multiplex, nested sc-PCR (right). Phox2b- and VGlut2-expressing RTN neurons also expressed Nalcn (boxed); fractional Nalcn expression for all cells tested (n = 164). GAD1 was detected in pooled non-GFP neurons (p); amplicons were not detected in control reactions with no template (nt). C, Firing rate histograms from exemplar GFP-labeled RTN neurons recorded in acute slices depicting effects of bath pH on discharge in cells that were characterized as either Type I (pH₅₀ = 7.23) or Type II (pH₅₀, 7.48). D, Left, Correlation between Nalcn expression levels (by sc-qPCR, normalized to GAPDH) and pH₅₀ values in individual recorded RTN neurons (R² = 0.46, p < 0.0001). Right, Nalcn expression in Type I (pH₅₀ ≤ 7.42, n = 10) and Type II (pH₅₀ > 7.42, n = 18) RTN neurons. **p < 0.0001 (unpaired t test).

Figure 2.

Nalcn knockdown in vitro decreases RTN neuronal firing. A, Schematic of AAV construct and examples of GFP-expressing, AAV2-infected (mCherry⁺) RTN neurons in cultured brainstem slice. B, Nalcn expression determined by sc-qPCR in all RTN neurons tested (relative to GAPDH). This complete dataset includes acutely dissociated GFP-labeled RTN neurons (diss); it also includes RTN neurons from cultured slices that were uninfected (uninf) or infected with scrambled (scr) or Nalcn shRNA that were harvested after dissociation or after electrophysiological recording. ****p < 0.0001 (ANOVA with pairwise post hoc comparisons by Bonferroni test). C, D, Cell-attached recording of effects of changing bath pH on firing rate in RTN neurons infected with AAV2 for scrambled control shRNA (C) or Nalcn shRNA (D). E, Resting membrane potential was determined under whole-cell current-clamp conditions in a subset of cells following cell-attached recordings. ****p < 0.0001 (ANOVA with pairwise post hoc comparisons by Bonferroni test). F, Effect of bath pH on firing rate of RTN neurons determined by cell-attached recording in cultured slices. Left, The slopes of best fit regression lines for control (uninfected and scrambled shRNA) and Nalcn shRNA-expressing RTN neurons were not different; the control line was shifted significantly toward higher firing rates (p < 0.01). Right, Cumulative probability histogram of pH₅₀ values for control RTN neurons (green represents uninfected, n = 28; white represents scrambled shRNA, n = 24) and for RTN neurons expressing Nalcn shRNA (red represents n = 34). Overlaid line is best fit Gaussian curve. **p < 0.01 (Kolmogorov–Smirnov test). G, Whole-cell voltage-clamp steps were used to characterize the pH-sensitive background K⁺ current by digital subtraction (pH 7.8- pH 7.0) in RTN neurons expressing either scrambled (n = 5) or Nalcn shRNA (n = 4); fits to the Goldman-Hodgkin-Katz (GHK) equation for a K⁺-selective current are overlaid.

Figure 3.

Nalcn knockdown in vitro reduces a TTX-resistant leak Na⁺ current. A, Whole-cell voltage-clamp recordings of membrane currents with Cs-based internal solution in RTN neurons that were uninfected, or infected with either scrambled shRNA or Nalcn shRNA under control conditions (Na-HEPES bath solution) or in a bath solution that replaced Na with NMDG. Arrows indicate zero current; capacitive transients truncated for presentation. B, Holding current (at −60 mV) and input conductance for control RTN neurons (uninfected, n = 7; or scrambled shRNA, n = 5) and for RTN neurons expressing Nalcn shRNA (n = 11). *p < 0.05 (ANOVA, pairwise post hoc comparisons by Bonferroni test). ***p < 0.001 (ANOVA, pairwise post hoc comparisons by Bonferroni test). C, Current-voltage (I–V) plot for control (uninfected, n = 8; scrambled shRNA, n = 6) and Nalcn shRNA-expressing RTN neurons (n = 15); currents normalized to cell capacitance (current density, pA/pF). D, E, G, Effect of Na⁺ substitution (with NMDG) on holding current in exemplar RTN neurons that were uninfected (D), or expressing either scrambled shRNA (E) or Nalcn shRNA (G). F, H, I–V curves illustrating effects of Na⁺ substitution for control RTN neurons (F; data pooled from uninfected and scrambled shRNA-expressing cells, n = 14) and for RTN neurons expressing Nalcn shRNA (H, n = 15); currents normalized to cell capacitance (current density, pA/pF). I, NMDG-sensitive holding current (at −60 mV; left) and input conductance (right) for control RTN neurons (uninfected, n = 7; scrambled shRNA, n = 5) and for RTN neurons expressing Nalcn shRNA (n = 11). **p < 0.01 (ANOVA, pairwise post hoc comparisons by Bonferroni test). ***p < 0.001 (ANOVA, pairwise post hoc comparisons by Bonferroni test). J, Correlation between Nalcn expression levels (by sc-qPCR, normalized to GAPDH) and NMDG-sensitive current amplitude for individual RTN neurons (n = 6, n = 5, and n = 7; R² = 0.86, p < 0.0001).

Figure 4.

SP activates Nalcn to cause excitation of RTN neurons. A, Cell-attached recording of firing rate during changes in bath pH and application of SP (1 μm, bath) in an RTN neuron expressing scrambled control shRNA (top) or Nalcn shRNA (bottom). B, Cell-attached recording of firing rate during changes in bath pH and application of 5-HT (5 μm, bath) in RTN neurons expressing scrambled control shRNA (top) or Nalcn shRNA (bottom). C, Summary data of changes in firing rate at pH 7.3 evoked by SP (left) or 5-HT (right) in control RTN neurons (uninfected or scrambled shRNA) and Nalcn shRNA-expressing RTN neurons. N values provided on plot. *p < 0.05 (ANOVA, pairwise post hoc comparisons by Bonferroni test). NS, Not significant (p > 0.49, ANOVA). D, Expression of Nalcn, VGlut2, NK1R, and NK3R determined by multiplex sc-qPCR in RTN neurons in acutely dissociated GFP-labeled RTN neurons (diss) and from RTN neurons expressing scrambled (scr) or Nalcn shRNA (sh) obtained from cultured slices; only Nalcn expression was reduced by Nalcn shRNA. ***p < 0.001 (ANOVA with pairwise post hoc comparisons by Bonferroni test). N values provided on plot. E, Whole-cell voltage-clamp recordings from exemplar RTN neurons expressing scrambled shRNA (Ea) or Nalcn shRNA (Eb) showing SP effects on holding current (top) and input conductance (bottom). Ec, Amplitude of peak SP-induced current in control RTN neurons (uninfected, n = 6; scrambled shRNA, n = 7) and Nalcn shRNA-expressing RTN neurons (n = 11). **p < 0.01 (ANOVA, pairwise post hoc comparisons by Bonferroni test).

Figure 5.

Specific and selective Nalcn knockdown in RTN neurons in vivo. A, Lentivirus was injected into the RTN of Phox2b::GFP mice based on stereotactic coordinates and antidromic field potentials elicited by facial nerve stimulation. B, Schematic of lentiviral construct for expressing Nalcn (or control) shRNA in RTN neurons in vivo. C, Left, Color merge of photomicrographs from GFP- and mCherry-immunostained sections depicting RTN neurons that were uninfected (green arrows) or infected with Nalcn shRNA-expressing lentivirus (yellow arrows); uninfected, non-RTN neurons are also apparent (white arrows). Right, Corresponding image from the same section showing Nalcn expression by in situ hybridization. D, Color merge of GFP and mCherry immunostaining (top) and Nalcn in situ hybridization (bottom) from RTN of mouse injected with control shRNA lentivirus. E, Color merge of GFP and mCherry immunostaining (top) and NK1R in situ hybridization (bottom) from RTN of mouse injected with Nalcn (left) or control (right) shRNA lentivirus (arrows in D, E as described in panel C). F, Cell counts through the RTN depicting the percentage of neurons that express the indicated gene (Nalcn, NK1R, or VGlut2, by in situ hybridization), after injection with Nalcn or control shRNA lentivirus; counts include RTN neurons that were either transduced (GFP⁺/mCherry⁺) or not transduced (only GFP). Numbers of mice analyzed after receiving Nalcn shRNA or control shRNA: for Nalcn, n = 18 and n = 10; for NK1R, n = 5 each; and for VGlut2, n = 6 and n = 4. ****p < 0.0001 (ANOVA, pairwise post hoc comparisons by Bonferroni test).

Figure 6.

Nalcn knockdown in RTN attenuates CO₂-stimulated cFos expression. A, cFos expression following exposure to CO₂ (8%, 45 min) in GFP-expressing RTN neurons infected with lentivirus for control (top) or Nalcn shRNA (bottom). Arrows indicate cFos-immunoreactive RTN neurons that were either transduced (yellow) or not transduced (white) with the indicated virus. B, Cell counts through the RTN depicting the percentage of cFos-labeled RTN neurons transduced with control or Nalcn shRNA virus compared with nontransduced RTN neurons. ****p < 0.0001 (by ANOVA, pairwise post hoc comparisons by Bonferroni test).

Figure 7.

Nalcn knockdown in RTN attenuates CO₂-stimulated ventilation. A, Example records of respiratory flow data from two individual mice during exposure to gas mixtures with different CO₂ concentrations (balance O₂), before and 4 weeks following injection of lentivirus for control and Nalcn shRNA into the RTN region. The lentiviral transduction ratio and the percentage of preinjection ΔV_E at 8% CO₂ is provided for each mouse (also see D). B, Left, Effect of increasing inspired CO₂ concentrations on V_E (the product of breathing frequency, breaths/min, and normalized tidal volume, ml/breath/g) in mice 4 weeks after injection with control or Nalcn shRNA-expressing lentivirus (n = 15 each). Shaded area represents the 95% confidence for all mice before injection. ****p < 0.0001 versus control virus at 4 weeks (two-way repeated-measures ANOVA, pairwise post hoc comparisons by Bonferroni test). Right, Effect of CO₂ on respiratory frequency (fR; breaths/min) and normalized V_T (ml/breath/g) 4 weeks after virus injection in the same control and Nalcn shRNA-expressing mice (n = 15 each). Nalcn shRNA versus scrambled shRNA: *p < 0.05; ***p < 0.001; ****p < 0.0001. Nalcn shRNA, before versus after injection: ⁺⁺p < 0.01; ⁺⁺⁺p < 0.001; ⁺⁺⁺⁺p < 0.0001; two-way repeated-measures ANOVA with pairwise post hoc comparisons by Bonferroni test. C, Increase in V_E in 8% CO₂, before and after injection with control or Nalcn shRNA-expressing lentivirus. **p < 0.01 (two-way repeated-measures ANOVA, pairwise post hoc comparisons by Bonferroni test). D, Relationship between RTN transduction ratio and the relative change in response to CO₂ for each individual viral-injected mouse. The blunting effect of Nalcn depletion on CO₂-stimulated breathing was greater in mice with more infected RTN neurons (R² = 0.76 for Nalcn shRNA, p < 0.0001; slopes significantly different, **p < 0.01). The two filled symbols are data from the exemplar mice shown in A. E, Effect of changes in inspired O₂ concentrations on V_E (normoxia = 21% O₂, hyperoxia = 100% O₂; hypoxia = 10% O₂) in mice 4 weeks after injection with control or Nalcn shRNA-expressing lentivirus (n = 15 each). NS, Not significant.