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

Abstract

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.

Fig. 1

p11 controls MN IME through TASK1, but not TASK3 subunits. a Co-immunoprecipitation (IP) of p11 and TASK1 in protein fractions of homogenized brain from an adult mouse. An anti-sheep IgG was a control. Elution 1 (E1) and 2 (E2) are displayed. b Left, experimental model for whole-cell patch clamp recordings from HMNs in brainstem slices. Inset, a micropipette (white arrow) close to the MN pool is showed. Right, voltage responses (top) to step hyperpolarizing currents (bottom) of two HMNs recorded from P7 rats receiving indicated oligonucleotides at P5. Dots, time points used to measure the peak voltage response to construct the I–V plot (see the “Methods” section). c, d Microinjection of siRNA_p11 (2 μg/2 μl) into the fourth ventricle of P5 rats reduced mRNA_p11 levels in the brainstem at P7. e Vm and R_N of HMNs recorded under pH 7.2. f Schematic summarizing effects of extracellular pH on MN IME and TASK-mediated K⁺ currents. Int intracellular, Ext extracellular. g Changes in Vm and R_N induced by variation of extracellular pH from 8.2 to 6.2 (TASK-dependent changes). h Time-course of Vm and R_N obtained from recorded SMNs^wt at the indicated days-in vitro (DIV) after seeding. i Effect of siRNA_p11 or cRNA (2 μM) on p11 and TASK1 in SMNs^wt by western blot. j Confocal images showing that siRNA_p11 (2 μM) leads to a drastic reduction in p11 immunolabelling (red) together with an increase in TASK1 expression (green) in neuritic processes of SMNs^wt, as compared to cRNA (2 μM). Scale bar, 5 μm. k Box-plots of the mean (top) and integrated (bottom) intensity (in arbitrary units, a.u.) of TASK1 immunolabelling in neurites after treatments. l, m As in e, g, respectively, but from SMNs of the indicated genotypes under the specified treatments. Number of independent samples in each group is in parentheses. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant; by Student t-test d, e, g, k, l, one-way analysis of variance (ANOVA) with post hoc Holm–Sidak method i, m (Vm) or ANOVA on Ranks with post hoc Dunn’s method m (R_N). Source data are provided as a Source Data file

Fig. 2

p11 controls Ca²⁺ signaling and vulnerability to an excitotoxic stimulus through TASK1, but not TASK3 subunits. a Dimensional displays (left) and time course of [Ca²⁺]_i alterations (right) obtained from SMNs treated with cRNA or siRNA_p11 (2 μM) from 1 DIV. Red horizontal lines indicate the glutamate (Glu, 150 μM) exposure interval. b Log-Rank test, Kaplan–Meier analysis reported that siRNA_p11 (dotted lines) delayed Ca²⁺ deregulation in SMNs^wt and SMNs^task3−/− (p < 0.001), but not SMNs^task1−/− (p = 0.761) relative to that in the cRNA condition. n = 36–49 SMNs. Ca²⁺ imaging studies were performed at 3–4 DIV. c Schematic summarizing effects of listed drugs on TASK-mediated currents. d Effects of bupivacaine (Bupiv; 40 μM, 60 min) and Glu (150 μM, 30 min), added alone or in combination, on SMNs^wt survival. e Effects of isoflurane (Iso; 0.8 mM each 10 min for 70 min) and Glu, added alone or in combination, on SMNs^wt, SMNs^task1−/−, and SMNs^task3−/− survival. f Influence of AEA (10 μM) on Vm of a SMN^wt at 4–5 DIV. g Changes in Vm (left) and R_N (right) of SMNs^wt, SMNs^task1−/−, and SMNs^task3−/− induced by AEA at the indicated concentrations. Number of SMNs in each group is in parentheses. h Effects of AEA (10 μM, 60 min) added alone or in combination with Glu on the survival of the indicated groups of SMNs. i Dose-dependent effects of siRNA_p11 on the survival of SMNs^wt exposed to Glu (30 min, 150 μM). cRNA (5 μM) was taken as an additional control. Mean number of SMNs^wt in the untreated condition was taken as 100%. j Mean effects of siRNA_p11 (2 μM) on survival of SMNs^task1−/− and SMNs^task3−/−. No NTFs, culture medium not supplemented with neurotrophic factors. Schematics in d–f, h, i represent experimental protocols. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant; by Student's t-test g or one-way analysis of variance (ANOVA) with post hoc Holm–Sidak method d, e, h–j. d, e, h–j n ≥ 3 independent experiments. k Schematic diagram depicting a neuroprotective strategy by interfering with p11 function (with siRNA_p11 in the example). Source data are provided as a Source Data file

Fig. 3

Neuroprotective effects of siRNA_p11 in the SOD1-G93A mouse model of ALS. a In situ hybridization (ISH, from the Allen Brain Atlas) and immunohistochemistry (ICH) showing the expression pattern of p11 in the lumbar spinal cord of adult wild-type mice. b, c Confocal images of lumbar spinal cords immunolabeled for p11 and the MN marker SMI32 b or the astroglial marker GFAP c. d mRNA (left) and protein (right) levels of p11 and TASK1 in the spinal cord of SOD1-G93A mice and their Non-Tg littermates, at the indicated stages (see schematic on top). e p11 immunofluorescence (left) and mean intensity (arbitrary units, a.u.; right) of labeling in lumbar MNs of indicated mice and age. MNs were identified as in b. f Confocal images showing astroglial reaction in the ventral horn of SOD1-G93A mice. Double immunolabeling reported p11-immunoreactive astrocytes in the boxed area. g Schematic of timing for injection system implantation and cRNA/siRNA_p11 administration. h p11 immunohistochemistry performed in lumbar spinal cord sections obtained from indicated mice which received cRNA or siRNA_p11 from P60. Insets, examples of two vacuolated MNs with different levels of p11 immunostaining. i Number of SMI32-identified MNs in the L3–L5 spinal cord segments of 4-month-old mice with the indicated genotypes that received the stated treatments beginning at P60. The box indicates the area of the ventral horn containing the MN pools that were quantified. Scale bars, b, c, e, 50 μm; a, c, 100 μm; h, i, 300 μm. j–l Cumulative probability curves of symptom onset j, survival l, and time course of mean body weight k for cRNA-treated and siRNA_p11-treated SOD1-G93A mice. Number of independent samples in each group is in parentheses. Statistic outputs are displayed in plots. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; by Student t-test d, e, i or two-way ANOVA with post hoc Holm–Sidak method k. Log-Rank test, Kaplan–Meier analysis was applied to j and l. Source data are provided as a Source Data file

Fig. 4

Cell-type specific p11 dysregulation is pivotal in a traumatic model of MN degeneration. a In situ hybridization (ISH, from the Allen Brain Atlas) and immunohistochemistry (ICH) showing the expression pattern of p11 in the HN of adult wild-type mice. b Confocal images showing p11 in SMI32-labeled HMNs (top) but not in GFAP-positive astrocytes (bottom). c Top, experimental design. Bottom, images illustrating neuronal-specificity of a lentivirus, which direct the expression of eGFP. d Effect of resection of a 2-mm segment of the XIIth nerve and immediate proximal stump ligation on the number of HMNs in adult mice. A neutral red-stained coronal section at the indicated post-injury time is shown on top. e Western blots (top) and plot (bottom, left) showing p11 upregulation in the injured (Inj) relative to the intact (Int) HN. Empty circle, sham condition. Bottom, right, qRT-PCR analysis. f p11 upregulation (top) and astroglial reaction (bottom) in the insulted HN after nerve injury. g Double immunohistochemistry showed p11 expression in SMI32-identified HMNs but not in GFAP-positive astrocytes in the damaged side. h Effect of microinjection of indicated LVVs in the midline between both HNs c on protein and mRNA levels for p11 in HNs. Neocortex was a control for systemic transduction. gapdh and α-tub were internal controls for qRT-PCR and western blotting, respectively. i LVV-shRNA_p11 microinjection, 3 days before nerve injury (see schematic in j), reduced p11 labeling in SMI32-identified HMNs relative to LVV-shRNA_lacz. Mean intensity (arbitrary units, au) of p11 labeling in injured/intact HMNs (top) or in the intact/non-injected side (bottom) after indicated LVVs injection. j LVV-shRNA_p11 reduced neuronal loss, as assessed by Nissl staining in the injured versus the intact HN. In d, e, i (top), j, data values in the injured side are displayed relative to those obtained for the intact side. Scale bars, c (right), 20 μm; a–c (left), f, g, i, 100 μm; d, j, 500 μm. Number of independent samples is in parentheses. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; by Mann–Whitney test d, e, h, Student t-test i or one-way ANOVA with post hoc Holm–Sidak method j. Source data are provided as a Source Data file

Fig. 5

RNS upregulates p11 and inhibits TASK channels in a traumatic model of MN degeneration. a Effects of nerve injury or persistent action of the NO-PKG pathway on MN IME and TASK-mediated currents. b–e Expression of phosphorylated ATF-1 (pATF-1) and RhoA proteins b, indicated mRNAs c, d and p11 e in HNs from neonatal brainstem slices incubated for 6 h in aCSF alone (control), or supplemented with either DETA/NO (1 mM), DETA/NO+PKG-i (10 μM), DETA/NO+H1152 (20 μM), or DETA/NO+Y27632 (10 μM). In c, mRNA levels of DETA/NO versus control condition are presented. In b–e, n ≥ 3 independent experiments. f Experimental model. Crushing was inflicted to the XIIth nerve of P3 rats. Pups were untreated, L-NAME-treated, or they received a single intracerebroventricular microinjection of cRNA or siRNA_p11. g Immunoblots for RhoA and p11 of HNs from P7 rats treated as shown. gapdh and α-tub were the internal controls for qRT-PCR and western blotting, respectively. h Averaged I–V relationships of the TASK-mediated current for the indicated treatments (see experimental design in f) (see Supplementary Fig. 9b). Data were well fitted (r > 0.9) with the Goldman–Hodgkin–Katz equation (solid line). Current reversed close to −90 mV, near the reversal potential predicted for K⁺ (E_K). Goldman–Hodgkin–Katz fits for control and DETA/NO (dashed lines), previously reported, are shown for comparison. i Top, current-clamp recordings of the voltage responses to a series of depolarizing and hyperpolarizing current pulses (0.5 s duration, 0.04 nA increments) from a control and a damaged HMN recorded 4 days after axonal injury. Bottom, time courses illustrating alterations in current holding (I_holding) recorded in response to extracellular pH changes for two P7 HMNs held at −65 mV. j Plots represent TASK-dependent changes in Vm and R_N for each listed treatment group. Number of independent samples is in parentheses. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; by Student's t-test c, g or one-way ANOVA with post hoc Holm–Sidak method b, d, e, j. k Schematic diagram depicting the mechanism by which RNS impacts on MN IME. Source data are provided as a Source Data file

Fig. 6

RNS stimulates the human p11 promoter through a Sp1-binding site. a, b Luciferase activity (in arbitrary units, a.u.) coupled to the human p11 promoter from transfected dNSC34 incubated for 48 h either with the indicated DETA/NO concentrations a or with drugs modulating the NO/sGC/PKG-ROCK signaling pathway b. c Luciferase activity displayed by the indicated serial 5′ promoter deletions in untreated (control) or DETA/NO-treated transfected dNSC34s. NO-induced changes, in folds relative to the mean value for untreated (control) condition, are presented on the right plot. d Top, sequence of the human p11 promoter from −96 to +7. A regulatory AP-2/Sp1-binding site is highlighted (blue). G/C to T site-directed mutations are also shown (Mut 1–Mut 5). Middle, as in c for the −96 to +89 segment of the promoter including each of mutations. The −70 to +89 segment was used as a NO-unresponsive control. As a negative control, mutation of the AP-2/Sp1-binding site downstream −70 did not alter NO-responsiveness. Bottom, NO-responsive GC-box element identified by site-directed mutagenesis. e–g EMSAs obtained by incubating ³²P-labeled −96 to +89 wild-type or Mut 5 probes with three independent nuclear protein extracts (1, 2, 3) from dNSC34s or in the absence of nuclear extracts (−) e. Competition assays by co-addition of canonical sequences for Sp1 and AP-2 at the indicated concentrations f, or super-shift analysis with antibodies (2 μg μl⁻¹) against Sp1 and/or Sp3 g. h Time course of changes in mRNA_Sp1 and mRNA_p11 expression in SMNs^wt incubated with the stated drugs beginning at 6 DIV relative to that obtained from untreated SMNs^wt at the respective time points. DETA/NO (0.1 mM) for 24 or 72-h did not affect mRNA levels. DETA (1 mM, 72 h) was a negative control. n = 3 independent experiments. i Levels of mRNA_Sp1 and mRNA_p11 in SMNs^wt that received the indicated treatments for 48 h. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; by Student's t-test c, d, one-way ANOVA with post hoc Holm–Sidak method a, b, h, i. j The schematic represents the molecular pathway subserving p11 upregulation under RNS. Source data are provided as a Source Data file

Fig. 7

Neuroprotective effects of the Sp1-interfering agent Mit-A in the SOD1-G93A mouse model of ALS. a mRNA (left) and protein (right) levels for Sp1 in the spinal cord of SOD1-G93A mice and their Non-Tg littermates at the indicated ages. b Confocal images showing Sp1-immunolabeled nuclei of SMI32-positive lumbar MNs from 3-month-old Non-Tg and SOD1-G93A mice. Boxed areas in low magnification images for the Non-Tg condition are displayed at higher magnification together with similar areas obtained from transgenic mice. Box plot represents the mean intensity of nuclear Sp1-labeling of Non-Tg or SOD1-G93A MNs. c Mechanism of action by which Mit-A might protect neurons against excitotoxic degeneration. d Daily administration of Mit-A (30 μg kg⁻¹ day⁻¹) in the drinking water, beginning at P30, dysregulated mRNA (left) and protein (right) levels of p11 and TASK1, in the lumbar spinal cord of SOD1-G93A mice at P120. e Number of SMI32-positive MNs (left), in the L3–L5 spinal cord segment of SOD1-G93A mice and their Non-Tg littermates, treated with Mit-A or vehicle. Scale bars, 50 μm. f–h Cumulative probability curves of symptoms onset f and survival h, and time course of body weight g, against the age of SOD1-G93A mice receiving the indicated treatments. Mit-A treatment as in d. i–k Representative electromyographic recordings from the gastronecmius muscle of Non-Tg and SOD1-G93A mice at weeks 13 i, j and 15 k. Transgenic mice received treatments as in f–h. In i, fibrillation potentials (dots) were recorded by needle electromyography from a 13-week-old transgenic mouse. These spontaneous muscle discharges were absent in Non-Tg mice j. l, m Maximal (l) and mean (m) frequency of fibrillations in SOD1-G93A animals at the indicated weeks treated with vehicle or Mit-A. Number of independent samples is in parentheses. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; by Student's t-test a, b, d, e, l, m or two-way ANOVA with post hoc Holm–Sidak method g. Statistic outputs determined by Log-Rank test, Kaplan–Meier analysis f, h or two-way ANOVA g between treatments are stated on plots. Source data are provided as a Source Data file

Fig. 8

TASK1-dependent neuroprotective effects of Mit-A in the traumatic model of MN degeneration. a–c Effects of resection of a 2-mm-segment of the XIIth nerve on mRNA_Sp1 and Sp1 levels a, number of Sp1-immunoreactive nuclei b, and optical density of nuclear staining c in the HN of the injured (inj) versus the intact (int) side at the stated days post-lesion. Coronal sections immunolabelled for Sp1 b, c, some of them counterstained by Nissl (c, bottom), obtained 7 days after nerve injury. At least 50 Sp1-immunoreactive nuclei per animal and side were analyzed in c. d Immunoblots for Sp1 from HNs of the intact and injured sides at 7 days post-lesion in mice treated with L-NAME (180 mg kg⁻¹ day⁻¹, i.p.) or its inactive stereoisomer D-NAME from day 5th after nerve resection. gapdh and α-tub/β-act were internal controls for qRT-PCR and western blotting, respectively. e, f Plots represent the number of neurons (in percent) in the injured versus the intact side taken as 100%. In e, administration of vehicle and drugs were intraperitoneal (i.p.). Mit-A was orally administered except when indicated as i.p. applied in f. Nissl stained coronal sections obtained from mice sacrificed 3 weeks after nerve injury receiving indicated treatments are shown to the left in f. Dotted lines delimit the HN in each section. Animals treated with vehicle in e and vehicle i.p. in f are the same. f Control, untreated animals; vehicle, drinking water supplemented with 1% sucrose; Mit-A-30 (30 μg kg⁻¹ day⁻¹); Mit-A-300 (300 μg kg⁻¹ day⁻¹). Note that Mit-A was neuroprotective for task3^−/−, but not for task1^−/− mice. Scale bars, b 350 μm; c 100 μm; f 500 μm. Number of independent samples is indicated in parentheses. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; by paired (injured vs. intact side within a treatment, a–f), unpaired (comparison between two groups, f) Student t-test or one-way ANOVA with post hoc Holm–Sidak method (comparison between more than two treatments, e, f). Source data are provided as a Source Data file

Fig. 9

Cell-type specific p11 knockdown is neuroprotective for MNs in the SOD1-G93A mouse model of ALS. a Bilateral microinjections of neuron-specific LVVs were performed in the ventral horn of 2-month-old SOD1-G93A mice (see schematic on top). Reference section (from the Allen Brain Atlas) illustrating injection protocol (left) is the same than in Fig. 3a. Bright-field (middle panels) and fluorescence (right panels) images of the same coronal sections showing injection places (boxed areas) of Chicago sky blue at the same coordinates than LVVs. Insets, details of injection places. b Representative acetyl-cholinesterase (AChE)-immunostained L1–L3 sections (from the Allen Brain Atlas) showing horizontal plane sectioning performed to analyze LVV-treated mice. c–e Confocal images showing p11 immunolabelled GFP-positive SMI32-identified lumbar MNs from 4-month-old SOD1-G93A mice after intraspinal microinjection of LVV-shRNA_lacz c, LVV-dnSp1 d or LVV-shRNA_p11 e performed in 2-month-old animals. Note that while glial-like structures (arrows) show similar p11 labeling intensity in all conditions, transduced MNs in d and e display considerably less intensity than in c. f Mean fluorescence intensity (arbitrary units, a.u.) of p11 staining in GFP−/SMI32-positive lumbar MNs. g, h Representative SMI32-immunostained ventral horn (L1–L3) horizontal sections obtained from 4-month-old mice of the stated genotypes receiving indicated treatments. High magnification photomicrographs of some vacuolated MNs (arrows) are shown. Scale bars, a 500 μm; c–e 50 μm; g, h 20 and 100 μm, high and low magnification images, respectively. i Number of MNs in the L1–L3 spinal cord segment of 4-month-old mice with the indicated genotypes that received the stated treatments. Number of independent samples is in parentheses. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; by one-way ANOVA with post hoc Holm–Sidak method. Source data are provided as a Source Data file

Fig. 10

Schematic diagram modeling the role of the Sp1–p11–TASK1 triad in neurodegeneration. a Physiological expression of the transcription factor Sp1 maintains baseline levels of p11, which regulates IME by controlling TASK1 trafficking from the endoplasmic reticulum (ER) to the plasma membrane. p11–TASK1 complex dissociation is required to overcome ER retention and normal recycling of TASK1 to the surface. Under healthy conditions, when neuronal IME is fine-tuned, moderated Ca²⁺ influx, through GluRs and/or VSCCs, leads to physiological responses, which are controlled by mechanisms of intracellular Ca²⁺ buffering and extrusion. b In neurological conditions, over-expression and/or de novo expression of nNOS could be the switch on (star) to initiate a positive feedback loop that contributes to the degeneration and death of sick neurons. Usually, NO production in the nervous system is directed by Ca²⁺ influx driven by NMDA receptors. The calmodulin (CaM)–Ca²⁺ interaction, which is now enabled to activate nNOS, allows for the coupling of pathological NO synthesis to Ca²⁺ entry. NO reacts rapidly with the superoxide anion (O₂⁻) to form the highly toxic product peroxynitrite (ONOO⁻) and/or promotes S-nitrosylation, triggering protein misfolding and aggregation, thus leading to neurotoxicity (omitted for simplicity). In addition, a persistent pathological concentration of NO, involving the sGC/PKG-RhoA/ROCK pathway (omitted for simplicity), induces Sp1 and stimulates the human p11 promoter depending on, at least, a Sp1-binding site (GC-box). Subsequent p11 upregulation sequesters TASK1 to the ER, thus disrupting the normal recycling of this subunit to the surface. This TASK1 uptake/replenishment unbalance enhances IME, which, in turn, increases the opening probability of VSCCs and, GluRs in the presence of glutamate, thereby exacerbating Ca²⁺ entry into the cell. Ca²⁺ overload might contribute to neurodegeneration and, finally, to neuron death. Whether NO of glial origin and/or other unidentified factors promote Sp1 expression/binding and/or p11 expression are not discarded and merit for further investigation