Sp1-regulated expression of p11 contributes to motor neuron degeneration by membrane insertion of TASK1

Disruption in membrane excitability contributes to malfunction and differential vulnerability of specific neuronal subpopulations in a number of neurological diseases. The adaptor protein p11, and background potassium channel TASK1, have overlapping distributions in the CNS. Here, we report that the transcription factor Sp1 controls p11 expression, which impacts on excitability by hampering functional expression of TASK1. In the SOD1-G93A mouse model of ALS, Sp1-p11-TASK1 dysregulation contributes to increased excitability and vulnerability of motor neurons. Interference with either Sp1 or p11 is neuroprotective, delaying neuron loss and prolonging lifespan in this model. Nitrosative stress, a potential factor in human neurodegeneration, stimulated Sp1 expression and human p11 promoter activity, at least in part, through a Sp1-binding site. Disruption of Sp1 or p11 also has neuroprotective effects in a traumatic model of motor neuron degeneration. Together our work suggests the Sp1-p11-TASK1 pathway is a potential target for treatment of degeneration of motor neurons.

siRNAp11-induced hyperpolarization in SMNs wt (Fig. 1l) could affect Ca 2+ influx via GluRs and VSCCs. To test this hypothesis, selective inhibitors for NMDA (APV) and AMPA (NBQX) receptors, as well as for L-type (nifedipine), N-type (ω-conotoxin) or P/Qtype (ω-agatoxin) VSCCs were added to the bath solution before and during the excitotoxic challenge ( Supplementary Fig. 3g). siRNAp11 strongly attenuated the drop of resting [Ca 2+ ]i induced by APV, nifedipine, and ω-conotoxin, but not by NBQX and ω-agatoxin ( Supplementary Fig. 2h,i). These results suggest that p11 knockdown reduced Ca 2+ entry through NMDA receptors, as well as L-type and N-type channels, but not via AMPA receptors or P/Q-type channels, at baseline. Interestingly, while APV reduced peak [Ca 2+ ]i (-33.4 ± 3.1%), NBQX fully blocked the rapid response to glutamate in cRNA-treated SMNs wt .
In addition, siRNAp11-induced reduction in peak [Ca 2+ ]i was accentuated by APV, but unchanged by any of the VSCCs inhibitors ( Supplementary Fig. 3j). These findings agree with the relief of Mg 2+ block of NMDA receptors induced by Vm depolarization and/or AMPA receptors-mediated fast excitation 1,2 .
Strikingly, APV fully prevented glutamate-induced Ca 2+ deregulation in cRNAtreated SMNs wt , whereas AMPA receptors and P/Q-type VSCCs inhibitors drastically decreased the percentage of them undergoing deregulation ( Supplementary Fig. 3f). siRNAp11 strengthened the effects of AMPA, L-, N-, and P/Q-type inhibitors since the fraction of SMNs wt suffering Ca 2+ deregulation was reduced in the presence of NBQX, nifedipine, ω-conotoxin, and ω-agatoxin ( Supplementary Fig. 3f). Within the analyzed time window, contribution of L-and N-type channels to glutamate-induced Ca 2+ deregulation was lesser than that of P/Q-type. The strengthening of L-/N-type dependence under siRNAp11 treatment could be the result of two mechanisms. First, p11 down-regulation induces Vm hyperpolarization, thereby reducing opening probability of non-fully inhibited L-and/or Ntype channels under the drug concentrations used in our experiments. Second, if full inhibition of L-and/or N-type channels had been reached in our experiments, siRNAp11induced hyperpolarization would be expected to reduce Ca 2+ entry via NMDARs and/or nontargeted VSCCs. Furthermore, in the few cells than experienced Ca 2+ deregulation, the Ltype, but not the N-type, inhibitor prevented siRNAp11-promoted delays in Ca 2+ deregulation ( Supplementary Fig. 3k). These results suggest that, at least in the fraction of cells that still was glutamate-sensitive in the time window we studied, siRNAp11-induced delay in Ca 2+ deregulation was L-type-dependent. None of the siRNAp11-treated SMNs wt showed this last phase under APV, NBQX, and ω-agatoxin ( Supplementary Fig. 3f). Therefore, p11 downregulation attenuates Ca 2+ entry into the motoneuron through GluRs and VSCCs. Chronic incubation (≥ 4 h) of brainstem slices with dexamethasone (Dexa., 1 µM) increased mRNA p11 /p11 levels in the HN (a), enhanced IME (b), and reduced the impact of pH changes on Vm, and R N (c) of HMNs. gapdh and α-tub were the housekeeping gene for qRT-PCR and the internal protein loading reference for western blotting, respectively. Number of independent samples is in parentheses. Summary data are shown as mean ± s.e.m. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; by Mann-Whitney U test (a) or Student t-test (b,c). Source data are provided as a Source Data file. . Box-plot of the integrated intensity (in arbitrary units, au) of p11 immunofluorescence after the indicated treatments (n ≥ 6 fields for each of 3 independent experiments). Number of independent samples is in parentheses. Scale bars, 100 µm. ***p < 0.001; by Student t-test. Source data are provided as a Source Data file.       (4) (7) (4) (4) (7) (7) (4) (4) Step length (mm)

Supplementary Figure 9. Retinoic acid prevents/reverses RNS/nerve injury-induced p11 upregulation and inhibition of TASK-like channels.
(a-c) Co-incubation with retinoic acid (20 µM. ≥ 4 h) fully avoided both, DETA/NO-induced upregulation of p11 protein, but not mRNA levels (a), and DETA/NO-evoked inhibition of TASK channels in motoneurons (b,c). For more details see figure 5h and supplementary figure 8a,b. Goldman-Hodgkin-Katz fits for control and DETA/NO (dashed lines), previously reported 10 , are shown here for comparison. Results for dexamethasone (Dexa) have been presented in supplementary figure 1 and they are additionally presented here with comparative purposes. (d,e) SMI32-immunolabelled brainstem sections obtained at the level of the HN from two injured animals (d) and mean number of HMNs (e) in the intact and injured sides at the indicated ages. XIIth nerve crushing was inflicted at P3 (see f). Mean number of HMNs in the intact HN was taken as 100% for each time point tested. Scale bar, 300 µm. n = 3 rats/group. (f) Experimental model. Crushing was inflicted to the XIIth nerve of P3 rats. Pups were untreated and brainstem slices were incubated with retinoic acid (≥ 4 h) before HMN recordings or protein extraction as indicated in the schematic (bottom). (g) Plot represents pH-induced changes (from pH 8.2 to 6.2) in Vm and R N for each listed treatment group. (h) Immunoblots for p11 in HN samples obtained from brainstem slices of lesioned animals following incubation with retinoic acid. Plot shows averaged expression of p11 in the injured versus intact HN for each experimental condition. Control and/or DETA/NO groups in c,g and h are the same as for supplementary figure 8f and figure 5g, respectively. gapdh and α-tub were internal controls for qRT-PCR and western blotting, respectively. Number of independent samples is in parentheses. Summary data are shown as mean ± s.e.m. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; Mann-Whitney U test (   (3) (7) (6) (4)      Comparison with Riluzole. Cumulative probability curves of symptom onset (a,g), survival (c,i), time course of body weight (b,h), time to fall from rotarod (d,j), step length (e,k), and time required to run along a track of 50 cm (f,l), for SOD1 G93A mice receiving the treatments indicated in a, for a-f and g, for g-l.
(a-f) Mit-A treatment (30 µg/kg/d) began at presymptomatic (P60) or early symptomatic (P90) stages, by drug dissolution in the drinking water, supplemented with 1% sucrose. A higher dose of Mit-A (150 µg/kg/d) was tested beginning at P90. All animals were treated with vehicle (drinking water supplemented with 1% sucrose) from P30 up to beginning of Mit-A treatment. (g-l) Mit-A treatment (30 µg/kg/d) or Riluzole (25 mg/kg/d) began at pre-symptomatic (P60) stage, by drugs dissolution as above. All animals were treated with vehicle (drinking water supplemented with 1% sucrose) from P30 up to beginning of Mit-A/Riluzole treatment. The vehicle and Mit-A groups are the same as for figure 7f-h and supplementary figure 14e-g, and they are shown here for comparison purposes. (m) qRT-PCR of DNA, isolated from the tail of a random sampling of SOD1 G93A mice included in these series of experiments, discarded that differences between treatments were the consequence of different expression degrees of the human transgene. gapdh was the housekeeping gene. Number of mice per group and test is indicated in parentheses. Summary data are shown as mean ± s.e.m. n.s., not significant; by one-way ANOVA with post hoc Holm-Sidak method (m). Statistics outputs determined by Log-Rank test, Kaplan-Meier analysis (a,c,g,i) or two-way ANOVA (b,d-f,h,j-l) between each Mit-A/Riluzole-treated pools and the vehicle group or between Mit-A and Riluzole are indicated on plots. Source data are provided as a Source Data file.