Relaxation of synaptic inhibitory events as a compensatory mechanism in fetal SOD spinal motor networks

Abstract

Amyotrophic lateral sclerosis (ALS) is a devastating neurodegenerative disease affecting motor neurons (MNs) during late adulthood. Here, with the aim of identifying early changes underpinning ALS neurodegeneration, we analyzed the GABAergic/glycinergic inputs to E17.5 fetal MNs from SOD1^G93A (SOD) mice in parallel with chloride homeostasis. Our results show that IPSCs are less frequent in SOD animals in accordance with a reduction of synaptic VIAAT-positive terminals. SOD MNs exhibited an E_GABAAR10 mV more depolarized than in WT MNs associated with a KCC2 reduction. Interestingly, SOD GABAergic/glycinergic IPSCs and evoked GABA_AR-currents exhibited a slower decay correlated to elevated [Cl^-]_i. Computer simulations revealed that a slower relaxation of synaptic inhibitory events acts as compensatory mechanism to strengthen GABA/glycine inhibition when E_GABAAR is more depolarized. How such mechanisms evolve during pathophysiological processes remain to be determined, but our data indicate that at least SOD1 familial ALS may be considered as a neurodevelopmental disease.

Keywords: ALS disease; GABA/glycine inhibition; SOD1G93A; modelling; mouse; neuroscience; patch-clamp; relaxation time course.

Figure 1.. Alterations in chloride homeostasis.

(A) Representative traces of isoguvacine (Iso) responses illustrating the reversal of the evoked GABA_AR-related current (E_GABAAR) in WT and SOD MNs. Holding voltage −70 mV, −60 mV, −50 mV and −40 mV. I/V relationships, plotted below the traces, revealed that E_GABAAR was −61.4 mV and −48.5 mV in the representative WT and SOD MNs, respectively. (B1) Mean E_GABAAR was significantly lower in E17.5 WT (n = 25) compared to SOD (n = 24) MNs from the same litter. (B2) Accordingly, the mean [Cl^-]_i, calculated from individual E_GABAAR values, indicated that SOD MNs had a higher [Cl^-]_i than WT. (B3) Mean resting membrane potential (Em) did not differ in WT and SOD MNs. (B4) Mean driving force (DF) for chloride ions was doubled in SOD MNs compared to WT MNs. Values illustrated were from the same set of MNs. ****p<0.0001, ***p<0.001, ns non-significant, Mann Whitney test.

Figure 1—figure supplement 1.. Recorded MNs belong to the lateral motor column (LMC).

(A) FoxP1 transcription factor expressed in the LMC (green staining of nuclei). (B) A recorded MN filled with neurobiotin (red staining). (C) Merged image showing the LMC location of the recorded MN. MMC: medial motor cortex. Midline is indicated by a dashed line on the left of each panel.

Figure 2.. Expression of KCC2 and NKCC1 in the E17.5 SOD spinal cord.

(A1) Analysis of the KCC2 protein in SOD and WT SCs. The stain-free staining of total proteins loaded (lower left panel) was used as the normalization control to quantify the ~140 kDa KCC2 band (upper left panel). Right panel: quantification of KCC2 WB. (A2) Analysis of the NKCC1 protein in SOD and WT SCs. The ~150 kDa NKCC1 band (upper left panel) was analyzed. Quantification of NKCC1 WB. Three loads (20 μg of protein) from SOD and WT lumbar spinal cord extracts were used, each load including five individual SCs from four different litters. (B1–B4) Representative immunohistochemical KCC2 (red) staining in WT (B1–B3) and SOD (B2–B4) lumbar 0.2 µm thickness optical sections (L3-L5 level) containing Hb9-eGFP MNs (green). Images correspond to x60 confocal acquisitions. White arrows point the KCC2 labeling surrounding spinal MNs. (B5–B6) Quantification of the global KCC2 staining (B5) in the MN area and membrane KCC2 staining (B6) in E17.5 SOD and WT SCs. (C1) Illustration of the protocol applied to assess the KCC2 efficacy as explained in the Materials and method section. (C2–C3) Evolution of E_Cl values with time during the isoguvacine (iso) application in control aCSF (C2) and in the presence of the highly specific KCC2 blocker VU0240551 (10 µM) (C3). ***p<0.001, **p<0.01, *p<0.05, ns non-significant, Mann Whitney test.

Figure 3.. Motor activity in the E17.5 SOD spinal cord.

(A) Fictive locomotor-like activity evoked by 5-HT/NMDA/DA (see Materials and methods) recorded from contralateral lumbar ventral roots (lL3, rL3; two first traces) displayed regular alternating activities (see circular plots above each raw data panel) in both WT (A1) and SOD (A2) SCs. Below the raw data are traces of the floating mean spike frequencies (mFr) and the corresponding troughs of activity (burst start, Bst) from which the phase relationships between the left and right SC sides were calculated (circular plots). (B) The cycle period was significantly longer in SOD SCs (n = 9) compared to WT (n = 8) SCs. **p<0.01, ns non-significant, Mann Whitney test.

Figure 4.. SOD E17.5 MNs exhibit a reduced IPSC frequency compared with WT MNs along with a persistent increase in their tau_decay.

(A1) Sample miniature IPSC (mIPSC) recordings obtained from representative SOD and WT MNs (holding membrane potential −70 mV). (A2) Cumulative frequency of inter-event interval (IEI) and mean values of mIPSCS revealed a small but significant reduction in frequency in SOD MNs (n = 19) compared with WT (n = 15) MNs. (A3) mIPSC peak amplitude was slightly smaller in SOD than in WT MNs (same layout as in A2). (A4) Mean tau_rise was unchanged but mean tau_decay was significantly increased in SOD mIPSCs compared with WT mIPSCs. Inset traces (mean of 30 events) from representative experiments. (A5) Blocking the GlyR or GABA_AR (black horizontal bars) reveals a tonic current. The distribution of pure GABA_AR-mediated mIPSCs in WT (n = 5) and SOD (n = 3) MNs and pure GlyR-mediated mIPSCs in WT (n = 3) and SOD (n = 5) MNs is shown between the two traces. (A6) Tau_decay of pure GlyR-mediated mIPSCs is significantly increased in SOD MNs (n = 3) compared to WT MNs (n = 5). Traces are mean of 53 events (WT) and 57 events (SOD) from representative experiments. (A7) Relaxation of pure GABA_AR-mediated mIPSCs is significantly longer in SOD MNs (n = 3) compared to WT MNs (n = 5). Traces are mean of 35 events (WT) and 35 events (SOD) from representative experiments. (A8) Mixed GlyR-GABA_AR-mediated events from representative experiments. (B1) Samples of spontaneous IPSC (sIPSC) recordings from representative SOD and WT MNs. (B2) Cumulative frequency of IEI and mean values of sIPSCS revealed a slight frequency reduction in SOD MNs (n = 17) compared to WT MNs (n = 14). (B3) sIPSC peak amplitudes were similar in SOD and WT MNs. (B4) Mean tau_rise was not significantly different in WT sIPSCs, but SOD sIPSCs exhibited a higher tau_decay. Inset traces (mean of 30 events) from representative experiments. (C1–C2) SOD eIPSCs (n = 6) displayed a higher tau_decay than WT eIPSCs (n = 10). (D1–D2) Puff-evoked SOD GABA_AR currents (n = 21) exhibited a higher tau_decay than WT GABA_AR currents (n = 16). Left traces in C1) and D1) are from representative SOD and WT eIPSCs and puff-evoked GABA_AR currents. ****p<0.0001, ***p<0.001, **p<0.01, *p<0.05, ns non-significant, Mann Whitney test.

Figure 5.. Inhibitory terminal staining in E17.5 SOD and WT SCs.

(A1-A2) Confocal visualization of Hb9-eGFP MNs (green) along with VIAAT (red) and synaptophysin (white) staining in lumbar E17.5 WT (A1) and SOD1 SCs (A2) (frontal sections). Arrows point the synaptophysin-VIAAT co-localized staining highlighted in the left corner insets (0.2 µm thickness optical section). (B1) The density of VIAAT puncta was similar in WT (n = 6) and SOD1 (n = 5) SCs. (B2) SOD SCs again exhibited a significant reduction in their density of synaptophysin terminals. (B3) VIAAT puncta co-localizing with synaptophysin puncta were slightly reduced in SOD SCs compared with WT SCs. ***p<0.001, **p<0.01, *p<0.05, ns non-significant, Mann Whitney test.

Figure 6.. Simulation of a putative compensatory role of tau_decay for the inhibitory strength of GABA/gly synaptic events in SOD-like MNs.

(A) Simulations made with a WT-like MN whose soma was continuously injected with depolarizing current (250 pA) to induce spiking discharge at a rate of ~12.5 Hz. After the spiking discharge was stabilized (t = 2 s), a train of GABA/Gly synaptic events was delivered to the soma at a frequency of 71 Hz, with either a tau_decay of 20 ms (A1) or 25 ms (A2) and E_GABAAR set at −50 mV. Using these two values of GABA/Gly tau_decay (20 ms: dark blue circles; 25 ms: light blue squares), various frequencies of GABA/Gly synaptic train (from 0 to 100 Hz) were tested with either E_GABAAR = -50 mV (A3) or E_GABAAR = −60 mV (A4). The horizontal dashed line represents the stabilized spiking frequency before application of the GABA/Gly synaptic train. Inhibitory effects (below dashed line) are in red, excitatory effects (above this line) are in green. (B1–B5) Simulations made with a SOD-like MN, in which the soma was continuously depolarized (200 pA injection) to induce spiking discharge at ~12.5 Hz (same layout as in A). Note that in the SOD-like MN with E_GABAAR = −50 mV, excitatory effects are observed when the GABA/Gly synaptic frequency is below 50 Hz (B3). This effect is not observed when E_GABAAR = −60 mV (B4) or is almost absent in the WT-like MN with E_GABAAR = −50 mV.

Figure 6—figure supplement 1.. Simulated E17.5 WT-like and E17.5 SOD-like MNs.

(A, B) The two simulated MNs differ only in the length of their terminal dendrites, being shorter in the SOD-like MN. (C) Schematic drawing of the simulated cell body, dendrites and axon along with the location of GABA/Gly synaptic input. AIS: axon initial segment.

Figure 6—figure supplement 2.. Firing activity and synaptic inputs recorded from E17.5 MNs during fictive locomotion.

(A) Activity recorded simultaneously in a lumbar ventral root (L5) and a spinal MN (L5 level). (B) Quantification of the ventral root activity during a locomotor-related burst, including number of events (spikes), mean spike frequency (FMean) and mean instantaneous frequency (FiMean). (C) Analysis of the instantaneous frequency of excitatory (Glu, blue) and inhibitory (GABA/Gly, pink) synaptic events recorded in voltage-clamp mode (holding membrane potential −45 mV) from an E17.5 lumbar MN during a locomotor-like sequence. Note that GABA/Gly synaptic events may recur at a frequency of >100 Hz during locomotor bursts. (D) Detection of a barrage of inhibitory (upward projecting) synaptic events (upper panel, corresponding to position one in C) and mixed inhibitory (upward) and excitatory (downward) synaptic events (lower panel, position two in C).

Figure 7.. Functional consequence of increasing tau_decay on locomotor activity.

(A) Effect of tau_decay on the locomotor rhythm when E_GABAAR (E_Cl) is set to −60 mV (A1-A2, B1) and −50 mV (B2).

Figure 8.. Summary.

(A) Schematic drawing summarizing the altered inhibitory inputs (~25% reduction, see Figure 3B3) to fetal SOD MNs. [Cl^-]_i is higher in SOD MNs (A1) than in WT MNs (A2) because of a KCC2 down-regulation, leading to an increased GABA/Gly-induced depolarizing effect (see insets). (B) Consequence of increasing tau_decay on the GABA/gly inhibitory effect in SOD-like MNs. Due to an accumulation in the intracellular compartment, E_GABAAR exerts a strong depolarizing effect. A burst of spikes generated by MNs is hardly blocked by a barrage of GABA/gly events (see blue traces) when tau_decay is set to 20 ms. Increasing tau_decay to 25 ms allows a better summation of the shunting component of the depolarizing GABA/gly post-synaptic event leading to a better clamp of E_m towards E_GABAAR and to the blockade of MN discharge (see green traces). (C) Impact of a larger tau_decay in SOD MNs on the frequency of the locomotor rhythm.