NKCC1 cotransporter inactivation underlies embryonic development of chloride-mediated inhibition in mouse spinal motoneuron

Abstract

Early in development, GABA and glycine exert excitatory action that turns to inhibition due to modification of the chloride equilibrium potential (E(Cl)) controlled by the KCC2 and NKCC1 transporters. This switch is thought to be due to a late expression of KCC2 associated with a NKCC1 down-regulation. Here, we show in mouse embryonic spinal cord that both KCC2 and NKCC1 are expressed and functional early in development (E11.5-E13.5) when GABA(A) receptor activation induces strong excitatory action. After E15.5, a switch occurs rendering GABA unable to provide excitation. At these subsequent stages, NKCC1 becomes both inactive and less abundant in motoneurons while KCC2 remains functional and hyperpolarizes E(Cl). In conclusion, in contrast to other systems, the cotransporters are concomitantly expressed early in the development of the mouse spinal cord. Moreover, whereas NKCC1 follows a classical functional extinction, KCC2 is highly expressed throughout both early and late embryonic life.

Figure 1. Schematic representation of the in vitro embryonic mouse spinal cord preparation used for electrophysiology

A, dissection procedure. After removing the meninges, the spinal cord is dorsally opened (A1) and positioned in the recording chamber ventral side up allowing direct access to motoneurons. A2, schematic view of coronal section of the opened spinal cord. B, patch-clamp recording of visually identified motoneurons. Partial upper view of the laid down spinal cord (B1). The dashed line indicates the midline (m) with the two symmetric columns of motoneurons, laterally positioned. B2, example of a visually identified motoneuron showing the relative position of puff pipette used to eject pharmacological agents and the patch-clamp pipette. B3, perforated patch clamp recording of a motoneuron. The depolarization is due to pressure ejection of the GABA_AR agonist isoguvacine. d, dorsal zone; m, midline; Mns, motoneuronal column; v, ventral zone.

Figure 2. Embryonic maturation of KCC2 and NKCC1 at lumbar level

A1, KCC2-ir on a frontal section of lumbar spinal cord at E11.5. Note that the staining is restricted to the ventro-lateral grey matter in the ventral area (v). A2, at E12.5, KCC2-ir has expanded to the whole ventral grey matter. A3, at E13.5, the KCC2-ir is detected in the dorsal zone and in dorso-medial areas. A4, at E14.5, the whole grey matter shows KCC2-ir with a stronger staining in ventral areas compared to dorsal. A5 and 6, from E16.5 to P0, KCC2-ir spreads across the entire grey matter. B1, at E11.5, NKCC1-ir is detected in the ventro-lateral and dorso-lateral grey matter and in the marginal zone. The inset, in the bottom right-hand corner shows, as for all panels, a higher magnification of the NKCC1 staining within the motoneuronal area corresponding to boxed areas drawn on the global view. B2 and 3, from E12.5 to E13.5, the NKCC1 staining remains stable in the ventral area whereas immunoreactivity progressively invades the entire dorsal zone. B4–6, from E14.5 to P0, NKCC1-ir remains detectable in the ventral area but appears stronger in the dorsal area. Staining was performed with the monoclonal T4 antibody. Scale bar, 100 μm. dh, dorsal horn; mz, marginal zone; v, ventricle. vh, ventral horn; vz, ventricular zone. C and D, quantitative analyses of mean KCC2- (C) and NKCC1-ir (D) in the ventral horn (see circled areas in A and B). Note that the mean intensity of both KCC2 and NKCC1 staining remains stable in the lumbar ventral area during ontogeny (no statistical differences between stages, one-way ANOVA). Each histogram corresponds to 4–12 preparations. Values are mean ±s.e.m.

Figure 4. GABA_AR-induced effect on motoneuron membrane potential at E13.5, E16.5 and P0

A brief pulse of depolarizing current (i) is injected into recorded motoneurons in order to bring the membrane potential above spike threshold. Then, a brief pressure application of isoguvacine (100 μm, 20 ms, 4 p.s.i.) is used to activate GABA_AR. A, at E13.5, at resting membrane potential (V_rest, A1), either depolarizing current or application of isoguvacine elicits a depolarization, on top of which a spike is generated. The amplitude of this depolarization is increased when the membrane potential is held (V_h) at −70 mV and an action potential is still triggered. B, at E16.5, a depolarizing pulse of current still elicits an action potential whatever the membrane potential (B1, B2) whereas the stimulation of GABA_AR evokes a hyperpolarization at V_h−50 mV (B1) and a depolarization at V_rest (B2). C, at birth the injected current triggers, at threshold level, a train of action potentials (C1, C2) whereas the application of isoguvacine still hyperpolarizes the membrane potential at V_h−50 mV (C1) but does not alter the membrane potential at V_h−70 mV (C2).

Figure 3. Ontogenic evolution of motoneuronal membrane properties from stages E13.5 to P0

A, comparison of motoneuronal response to depolarizing pulses, at E13.5, E16.5 and P0, showing that spike thresholds are similar at all stages of development whereas the resting membrane potential (V_rest) is 20 mV lower at E16.5 and P0 compared to E13.5. B, corrected spike threshold evolution throughout embryonic development. C, evolution of V_rest (corrected values) from E13.5 to P0. At each stage of development, circles represent individual measurement whereas black squares indicate mean values (±s.e.m.). D, evolution of motoneuronal R_in. The horizontal grey bar indicates the limits of the mature input resistance (between maximal and minimal s.e.m. values), between E16.5 and P0. E, maturation of motoneuronal membrane capacitance between E13.5 and P0. Values are given in Table 1.

Figure 5. Embryonic maturation of GABA_AR-related effects on membrane potential

A1 and B1, perforated patch clamp recordings showing currents evoked by brief application of isoguvacine, from different holding potentials, at E13.5 (A1) and E16.5 (B1). Current versus voltage plots (same neurons shown in A1 and B1) allowing calculation of GABA_AR equilibrium potential: −33 mV at E13.5 (A2) and −50 mV at E16.5 (B2) (measured by linear regression analysis, r² > 0.96 in both cases). C, evolution of corrected V_rest (white circles), spike threshold (grey triangles) and E_Cl (black circles) in motoneurons during embryonic development. From E13.5 to E15.5, spike threshold and E_Cl are not significantly different (non-parametric unpaired t test), whereas from E16.5 to birth, E_Cl is significantly lower than spike threshold (**P < 0.01; ***P < 0.0001). Values are means ±s.e.m. (n values indicated in Table 2).

Figure 6. Functional role of KCC2 and NKCC1 at different stages of development

A, representative experiment at E13.5 illustrating the reversal of isoguvacine-evoked currents in control conditions, during the blockade of NKCC1 cotransporter (bumetanide, 20 μm, 40 min) or when the KCC2 cotransporter is blocked (furosemide, 200 μm, 5 min). All current traces are from the same E13.5 motoneuron. The duration of washout is around 30 min. B, quantitative analysis of the effects of the cotransporter blockers at three key developmental stages (E13.5, E16.5–E17.5 and P0). At E13.5, NKCC1 and KCC2 blockers induce opposite variations of E_Cl. At E16.5–E17.5, although KCC2 blocker remains efficient, blocking the NKCC1 cotransporter reveals a decreased effect on E_Cl. At P0, only the KCC2 blocker is effective. n, number of motoneurons tested; n.s., non-significant difference; *P < 0.05 (one-way ANOVA followed by a post hoc Tukey/s test).

Figure 7. Distribution of NKCC1 and KCC2 proteins on motoneurons

A1–9, confocal images of an E13.5 Neurobiotin-injected motoneuron (A1) show the presence of KCC2-ir (A2–5) and NKCC1-ir (A6–9) clusters in the cytoplasm and in the periphery of the cell membrane (arrowheads). A3 and A7 are image displays in X–Y direction. A4 and A8 are line images displayed in X–Z direction (at the level of the purple vertical line in A2–3, A6–7). A5 and A9 are line images displayed in Y–Z direction (at the level of the yellow horizontal line in A2–3, A6–7). B1–9, confocal views of a P0 Neurobiotin-injected motoneuron (B1). KCC2-ir clusters (B2–5) appear dense in the cytoplasm and in the periphery of the cell membrane (arrowheads) while NKCC1-ir clusters, although still detected close to the cell membrane (arrowheads), appear less numerous in the cytoplasm (B6–9). Same multi-plane views as in A. Images are single optical sections (0.2 μm thick). C and D, quantitative analysis of the density (number per volume in μm³) of KCC2- and NKCC1-ir clusters in the cytoplasm and in the vicinity of the plasma membrane. KCC2-ir cluster density appears stable in the cytoplasm during the entire embryonic life (green circles) and decreases in the vicinity of cell membrane (black squares). NKCC1 cluster density diminishes during development in the cytoplasm (blue circles) and around the plasma membrane (black squares). Linear regression analysis for KCC2-ir clusters: slope significantly different from zero in the cytoplasm (P < 0.0001) but not around the cell membrane (P = 0.8250). Linear regression analysis for NKCC1 clusters: slope significantly non-zero in the cytoplasm (P < 0.01) and around the cell membrane (P < 0.01). Scale bars, 10 μm. C, caudal; L, lateral; M, medial; R, rostral. Data, from 5–6 Neurobiotin-injected motoneurons at each stage, are expressed as means ±s.e.m.

Figure 8. Mechanisms underlying the developmental shift in GABA_AR/GlyR responses in the mouse spinal cord

A, before E15.5, both KCC2 and NKCC1 are active and [Cl⁻]_i is high. Therefore, the activation of the Cl⁻-gated receptors induces a large outward chloride current that depolarizes the membrane and triggers action potentials (see inset, right upper corner). B, between E16.5 and E18.5, [Cl⁻]_i is reduced due to a decrease of the NKCC1 efficacy. At these stages, although the activation of GABA_AR and GlyR leads to depolarization, no action potential is elicited (inset). C, at birth, NKCC1 has lost its efficiency and only the chloride extruder KCC2 remains functional, lowering [Cl⁻]_i. Inset: around resting potential, the activation of GABA_AR/GlyR does not modify the membrane potential.