Reduced calreticulin levels link endoplasmic reticulum stress and Fas-triggered cell death in motoneurons vulnerable to ALS

Abstract

Cellular responses to protein misfolding are thought to play key roles in triggering neurodegeneration. In the mutant superoxide dismutase (mSOD1) model of amyotrophic lateral sclerosis (ALS), subsets of motoneurons are selectively vulnerable to degeneration. Fast fatigable motoneurons selectively activate an endoplasmic reticulum (ER) stress response that drives their early degeneration while a subset of mSOD1 motoneurons show exacerbated sensitivity to activation of the motoneuron-specific Fas/NO pathway. However, the links between the two mechanisms and the molecular basis of their cellular specificity remained unclear. We show that Fas activation leads, specifically in mSOD1 motoneurons, to reductions in levels of calreticulin (CRT), a calcium-binding ER chaperone. Decreased expression of CRT is both necessary and sufficient to trigger SOD1(G93A) motoneuron death through the Fas/NO pathway. In SOD1(G93A) mice in vivo, reductions in CRT precede muscle denervation and are restricted to vulnerable motor pools. In vitro, both reduced CRT and Fas activation trigger an ER stress response that is restricted to, and required for death of, vulnerable SOD1(G93A) motoneurons. Our data reveal CRT as a critical link between a motoneuron-specific death pathway and the ER stress response and point to a role of CRT levels in modulating motoneuron vulnerability to ALS.

Figure 1.

Expression of CRT, a luminal ER protein, is decreased twofold only in SOD1^G93A motoneurons and only after sFasL or NO treatment. Decrease in CRT expression is Fas-dependent and happens only in SOD1^G93A motoneurons and only after sFasL treatment. CRT staining was visualized (A) and quantified (B) in cells cultured for 24 h, then treated for 24 h with sFasL (1 or 100 ng/ml) or sLIGHT (100 ng/ml). Western blots of cultured motoneuron extracts (C) were quantified (D), confirming the immunostaining results. CRT decreased expression occurs in motoneurons expressing different human mutated SOD1, but not in motoneurons expressing the human WT SOD1 (E). CRT fluorescence was quantified in motoneurons cultured from SOD1^G93A, SOD1^G85R, and SOD1^WT embryos for 24 h and treated with DETANONOate, an NO donor (10 μm). CRT decreased expression is specific to motoneurons (F). CRT fluorescence was quantified in motoneurons and cortical and hippocampal neurons, cultured for 24 h, then treated for 24 h with NO (20 μm). Results are expressed as ratio to the CRT level measured in untreated cells. Scale bar, 20 μm. PDI, an ER stress marker shows increased expression in SOD1^G93A motoneurons after sFasL treatment (G). PDI fluorescence was quantified after sFasL (100 ng/ml) treatment as in B. The mean intensity of CRT or PDI fluorescence was assessed using NHI ImageJ software. Data are means ± SD of three independent experiments: *p < 0.05.

Figure 2.

CRT downregulation in WT motoneurons in vitro leads to death and implicates the Fas signaling pathway. Efficiency and specificity of CRT silencing with sh-CRTs in NSC34 cells (A, B) and in motoneurons (C, D). NSC34 cells were cultured for 24 h, then transfected with the EGFP vector or with shcrt1/2 or sh-control. Forty-eight hours later, cells were lysed and Western blot performed on the extracts (A). CRT levels were quantified relative to actin levels (B). Motoneurons were electroporated with the EGFP vector in combination with empty vector (e.v.), shcrt1/2, or sh-control and immunostained for CRT 48 h later (C). Immunostaining was quantified (D). Downregulation of CRT leads to motoneuron death (E, F). Motoneuron survival was assessed 1 and 2 d after electroporation with the EGFP vector in combination with e.v., shcrt1/2, or sh-control. All conditions were expressed relative to EGFP + e.v. condition (E). WT and CRT heterozygous (crt^+/−) Motoneurons were cultured and cell counting performed 1 and 2 d after seeding (F). The killing effect of CRT downregulation is dependent on Fas activation (G, H). WT motoneurons were electroporated as in (E). Motoneurons were treated 24 h later with Fas-Fc (1 μg/ml) and survival assessed 48 h later (G). WT motoneurons were electroporated as in (E), treated or not 12 h later with DETANONOate, and immunostained for Fas ligand 24 h later. FasL expression was quantified in EGFP-positive cells and expressed relative to EGFP alone condition (H). Data are means ± SD of three independent experiments: *p < 0.05, **p < 0.01, and ***p < 0.001.

Figure 3.

CRT overexpression prevents FasL/NO-induced death of SOD1^G93A motoneurons (A, B). WT or SOD1^G93A motoneurons were electroporated with EGFP/e.v. or EGFP/ CRT-HA vectors and were immunostained 24 h later for anti-HA to detect CRT and anti-VAPA to stain the ER (A). Scale bar, 20 μm. Twenty four hours after electroporation with the same vectors as in (A), motoneurons were treated or not with sFasL (100 ng/ml). Cell survival (EGFP-positive cells) was assessed 48 h later, and expressed relative to WT (B). Data are from three independent experiments ± SD: ***p <0.001.

Figure 4.

CRT expression is decreased in vivo in motoneurons in the SOD1^G93A mice lumbar spinal cord starting at the asymptomatic stage (A–D). Immunostaining of CRT and NeuN was performed on spinal cord sections at the L3–L5 levels in 30-, 38-, and 45-d-old mice (asymptomatic); 60-d-old mice (presymptomatic); 90-d-old mice (onset of the disease); and 110-d-old mice (symptomatic). A, Enlargement of the ventral horn of a 38-d-old mSOD1 spinal cord. Scale bar, 50 μm. Long arrow = low CRT expressor motoneuron, short arrow = high CRT expressor motoneuron. Quantification of CRT and NeuN immunostaining was performed using NHI ImageJ software. Data are expressed as the CRT mean fluorescence in SOD1^G93A relative to CRT fluorescence in age-matched WT mice (B). Western blot for CRT and calnexin (CNX) from lumbar ventral spinal cord extracts was performed with mice of the same age as in B. Quantification of the amounts of CRT and CNX using NHI ImageJ software, expressed as a ratio to the age-matched WT (C, D). Data in B–D are from three independent experiments ± SD for each age compared with the WT, *p < 0.05, **p < 0.01.

Figure 5.

Decrease in CRT expression occurs only in vulnerable motoneurons. A, Immunostaining of CRT was performed on lumbar spinal cord sections 7 d after injection of tetramethyl-rhodamine-dextran in the tibialis (vulnerable motoneurons) or in the soleus (resistant motoneurons) muscle. Scale bar, 50 μm. B, Immunofluorescence of CRT was quantified using NHI ImageJ software in tetramethyl-rhodamine-dextran-labeled motoneurons in WT or SOD1^G93A mice at 38 d, and data expressed relative to the WT. Data in B were from three independent experiments ± SD for each age compared with the WT. **p < 0.01.

Figure 6.

Calcium signaling is involved in SOD1^G93A motoneuron death triggered through Fas. Ratiometric [Ca²⁺]_i measurement in motoneurons. Representative motoneurons loaded with Fura-Red dye before, during, and after 25 mm KCl stimulation. Top row, Represents the calcium-insensitive, 490 nm excitation wavelength. Bottom row, Represents the calcium-sensitive, 440 nm excitation wavelength (A). Fura-Red fluorescence change when excited using 440 and 490 nm filters in response to KCl application is illustrated as a ratio of fluorescence intensities (B). Basal level of [Ca²⁺]_i (C), amplitudes of neuronal response to application of 25 mm KCl (D), and 50% recovery time after KCl application (E) were measured in different motoneurons, WT versus SOD1^G93A, treated or not with sFasL. Data are from four experiments with 20–30 neurons analyzed per experiment. Values ± SEM; **p < 0.01. Chelating intracellular calcium with BAPTA-AM prevents sFasL-induced death and CRT decrease of SOD1^G93A motoneurons (F, G). WT or SOD1^G93A motoneurons were treated with sFasL (100 ng/ml) in combination or not with BAPTA-AM (5 μg/ml) and motoneuron survival (F), and CRT levels (G) were assessed 48 and 24 h after, respectively. Increasing cytoplasmic calcium with caffeine treatment potentiates death triggered through Fas activation in WT motoneurons (H). WT motoneurons were cultured for 24 h, then treated with a sublethal dose of sFasL (1 ng/ml) and/or caffeine (10 μg/ml) and cell survival was evaluated 48 h after treatment. Data are means ± SD of three independent experiments: *p < 0.05, **p < 0.01. n.s., not significant.

Figure 7.

ER stress is involved in SOD1^G93A motoneuron death triggered through Fas. ER stress markers CHOP and P-eIF2α are upregulated in SOD1^G93A motoneurons treated with sFasL (A–C). WT or SOD1^G93A motoneurons were cultured for 24 h, then treated with sFasL (100 ng/ml) or thapsigargin (20 nm, 2 h) and immunostained 24 h after for CHOP (A, B) or P-eIF2α (C) in combination with a motoneuronal marker, SMI32. Scale bar (in A), 20 μm. The mean intensity of fluorescence was quantified using MetaMorph software. Inhibiting ER stress prevents sFasL-induced death of SOD1^G93A motoneurons (D). WT or SOD1^G93A motoneurons were cultured for 24 h then treated for 24 h with sFasL (100 ng/ml) and/or salubrinal (ER stress inhibitor, 5 μm) or Z-ATAD-fmk (caspase-12 inhibitor, 10 μm). Combined sublethal doses of sFas ligand (sFasL) and ER stress inducers result in death of WT motoneurons (E). WT motoneurons were cultured for 24 h, then treated for 48 h with sublethal doses of sFasL (1 ng/ml) and/or of ER stress inducers tunicamycin (0.05 μg/ml) or thapsigargin (0.1 nm). Cell survival was measured and results for each condition expressed relative to the nontreatment condition. Data are means ± SD of three independent experiments, expressed relative to untreated condition: *p < p 0.05, **p < 0.01, and ***p < 0.001.

Figure 8.

CRT decreased expression is upstream and independent of ER stress activation. Death of motoneurons induced by downregulation of CRT is dependent on ER stress (A, B). Motoneurons were electroporated with EGFP in combination with empty or sh-crt1/2 or sh-control vectors and treated or not with salubrinal (5 μm). Cell survival was assessed 48 h after (A). WT and crt^+/− motoneurons were treated with salubrinal 8 h after seeding and assayed for survival at 48 h (B). Blocking ER stress does not prevent decrease in CRT in SOD1^G93A (C). Motoneurons were treated 24 h after seeding with sFasL (100 ng/ml), in combination or not with salubrinal (5 μm). CRT immunostaining was performed 24 h later. Induction of ER stress in WT motoneurons does not lead to decreased expression of CRT (D). WT motoneurons were treated at 24 h after seeding with sublethal doses of sFasL (1 ng/ml), tunicamycin (0.05 μg/ml), or thapsigargin (0.1 nm). CRT/SMI32 double-immunostaining was performed 24 h later and CRT expression was quantified relative to SMI32. Data are means ± SD of three independent experiments. ER stress induced by nerve crush does not lead to CRT decrease (E, F). CRT and BiP expression were visualized (E) and quantified (F) relative to NeuN in L3–L5 sections of spinal cords from WT mice at 30 d, 2 d after sciatic nerve crush was performed on one side. C,D,E, Immunostainings were quantified with NHI ImageJ software and expressed relative to untreated and/or WT. Scale bar, 50 μm. Data in (F) are means ± SD of three independent experiments: *p < 0.05, ***p < 0.001.

Figure 9.

Model of CRT reduction and ER stress contribution to the death of SOD1^G93A vulnerable motoneurons. While Fas activation leads in all motoneurons to activation of nNOS and NO production, only in mSOD1-vulnerable motoneurons does this activation induce a decrease in CRT leading to an ER stress response and death.