Motoneuron Death Triggered by a Specific Pathway Downstream of Fas

Neuron, Vol. 35, 1067–1083, September 12, 2002, Copyright 2002 by Cell Press

Motoneuron Death Triggered by a Specific Pathway Downstream of Fas: Potentiation by ALS-Linked SOD1 Mutations Ce´dric Raoul,1 Alvaro G. Este´vez,2 Hiroshi Nishimune,1,6 Don W. Cleveland,3 Odile deLapeyrie`re,1 Christopher E. Henderson,1,4 Georg Haase,1,5 and Brigitte Pettmann1,5 1 INSERM U. 382 Developmental Biology Institute of Marseille CNRS - INSERM - Univ. Mediterranee Campus de Luminy Case 907 13288 MARSEILLE Cedex 09 France 2 Department of Physiology and Biophysics Center for Free Radical Biology University of Alabama at Birmingham Birmingham, Alabama 35294 3 CMM-East 3080 Ludwig Institute for Cancer Research University of California San Diego La Jolla, California 92093

Summary Death pathways restricted to specific neuronal classes could potentially allow for precise control of developmental neuronal death and also underlie the selectivity of neuronal loss in neurodegenerative disease. We show that Fas-triggered death of normal embryonic motoneurons requires transcriptional upregulation of neuronal NOS and involves Daxx, ASK1, and p38 together with the classical FADD/caspase-8 cascade. No evidence for involvement of this pathway was found in cells other than motoneurons. Motoneurons from transgenic mice overexpressing ALS-linked SOD1 mutants (G37R, G85R, or G93A) displayed increased susceptibility to activation of this pathway: they were more sensitive to Fas- or NO-triggered cell death but not to trophic deprivation or excitotoxic stimulation. Thus, triggering of a motoneuron-restricted cell death pathway by neighboring cells might contribute to motoneuron loss in ALS. Introduction The remarkable degree of conservation of programmed cell death (PCD) pathways between different organisms and cell types has led to the paradigmatic concept of a “central cell death pathway” in which upstream events trigger activation of caspases, often through mechanisms involving control of cytochrome c release from mitochondria (Green and Kroemer, 1998; Pettmann and Henderson, 1998). However, it is clear that different cell 4

Correspondence: [email protected] These authors contributed equally to this work. 6 Present address: Department of Anatomy and Neurobiology, Washington University Medical School, Campus Box 8108, 660 South Euclid Avenue, St. Louis, Missouri 63110. 5

death genes are involved in death of different cell types and that a given cell may activate distinct death pathways when triggered to die by different stimuli (Pettmann and Henderson, 1998). One striking example of the former concerns Bax, a Bcl-2 family member which is potently proapoptotic in many models. Null mutant mice for bax were initially characterized for their phenotype in the immune system (Knudson et al., 1995). Although this was significant, it was far from the complete inhibition of PCD that might have been expected. In contrast, subsequent analysis of the nervous system of these mice revealed profound reductions in neuronal PCD, both during development and following experimental lesions (Deckwerth et al., 1996). These results strongly suggest that there are neuron-specific modes of PCD. If this were confirmed, it would have clear consequences for understanding not only the control of cell death during development but also for analyzing and potentially preventing the neuron-specific cell degeneration and death that characterizes neurodegenerative diseases, such as amyotrophic lateral sclerosis (ALS), Parkinson’s disease, and Huntington’s disease. We have addressed this general question in the specific context of the death of motoneurons triggered by activation of Fas, which has been recently implicated in the developmental and lesion-induced death of neurons. Fas (APO-1/CD95) is a member of the death receptor family that includes the tumor necrosis factor receptor (TNFR) and the low-affinity neurotrophin receptor p75NTR. Upon engagement by its ligand FasL, the intracellular death domain of Fas recruits the adaptor molecule FADD, leading to the activation of caspase 8 and subsequent activation of a caspase cascade resulting in cell death (Kischkel et al., 1995; Nagata, 1997). The function of Fas has been most studied in the immune system, where activation of Fas leads to deletion of activated mature T cells at the end of an immune response and to killing of virus-infected cells or cancer cells by cytotoxic T cells and natural killer cells (Krammer, 2000). In the brain, a subset of cortical neurons express Fas in close proximity to cells expressing FasL, and death can be triggered in a proportion of these neurons in vitro by Fas activation (Cheema et al., 1999). In models of brain ischemia, expression of Fas and increased levels of active caspase-8 have been reported (Herdegen et al., 1998; Matsushita et al., 2000; Matsuyama et al., 1995; Northington et al., 2001), and there is a reduction of stroke-induced brain damage in the absence of FasL (Martin-Villalba et al., 1999). We recently showed that cultured motoneurons deprived of trophic support express increased levels of FasL and can be kept alive by blocking extracellular Fas/FasL interactions, showing that Fas activation is a driving force for cell death in these conditions (Raoul et al., 1999). Moreover, even in the presence of trophic factors, exogenous activation of the Fas system induces death of ⵑ50% of motoneurons through a pathway involving caspase-8 (Raoul et al., 1999). This system provided a means for studying signaling mechanisms that are downstream of Fas, indepen-

Neuron 1068

dently of other mechanisms involved in death induced by trophic deprivation. Amyotrophic lateral sclerosis (ALS) is a late-onset neurodegenerative disease characterized by the selective degeneration and death of motoneurons in spinal cord, brainstem, and cerebral cortex. While most ALS cases appear spontaneously, about 10% are familial (FALS). Of these, 10%–20% (1%–2% of total) are linked to various dominantly inherited mutations in the gene encoding superoxide dismutase 1 (SOD1). FALS pathology is mimicked to a striking degree in several lines of transgenic mice for human ALS-linked SOD1 gene mutations (reviewed in Cleveland and Rothstein, 2001). However, the mechanism(s) by which SOD1 mutations cause ALS are still unsolved. The nature of mutant SOD1 toxicity remains highly controversial (Beckman et al., 2001; Cleveland and Rothstein, 2001; Julien, 2001; Shaw and Eggett, 2000). Moreover, the molecular pathways leading to motoneuron degeneration have not been elucidated: it has been shown that mutant SOD1 can trigger cell death in neuronal cell cultures (Ciriolo et al., 2000; Ghadge et al., 1997; Rabizadeh et al., 1995) and in motoneurons (Durham et al., 1997), but genetic, pharmacologic, and biochemical studies in ALS patients and mutant SOD1 mice have given controversial results concerning the engagement of the classical apoptotic cascade (Kostic et al., 1997; Li et al., 2000; Migheli et al., 1999; Migheli et al., 1994; Pasinelli et al., 2000). Last, the cellular site of action of the mutations remains enigmatic, since motoneuron degeneration has only been observed in transgenic mice when SOD1 mutants were ubiquitously expressed under the control of their own promoter but not after specific overexpression in neurons (Lino et al., 2002; Pramatarova et al., 2001) or in astrocytes (Gong et al., 2000). Results presented here provide a potential answer to some of these questions. We first show that death of normal motoneurons triggered by Fas activation involves transcriptional upregulation of neuronal nitric oxide synthase (NOS). We could find no evidence for this pathway in other Fas-responsive cell types. We next isolated embryonic motoneurons from transgenic mutant SOD1 mice. These motoneurons displayed greatly increased sensitivity to activation of Fas and its intracellular relay nitric oxide (NO) but not to trophic factor deprivation or excitotoxicity. Thus, activation of a motoneuron-restricted signaling pathway may underlie selective loss of motoneurons in these models of ALS. Results Neuronal NOS Is Specifically Upregulated in Motoneurons Following Fas Activation and Its Activity Is Essential for Fas-Induced Death The observation that motoneuron death triggered by exogenous activation of Fas takes at least 48 hr (Raoul et al., 1999) raised the possibility that transcriptional events might be involved. As neuronal NOS (nNOS) has been implicated in motoneuron death induced by trophic deprivation (Estevez et al., 1998b), we looked for potential involvement of nNOS and two other NOS family members, inducible NOS (iNOS) and endothelial NOS (eNOS), in Fas signaling. To avoid effects due to trophic

deprivation, embryonic mouse E12.5 motoneurons were cultured in the continuous presence of optimal concentrations of neurotrophic factors (NTFs; see Experimental Procedures) for 16 hr before addition of agonistic antiFas antibody or soluble Fas ligand (sFasL) (Raoul et al., 1999). By RT-PCR analysis 24 hr later, none of the NOS isoforms was detected in motoneurons cultured in the presence of trophic factors alone. However, agonistic anti-Fas antibody strikingly increased the level of mRNA for nNOS, without affecting the other isoforms (Figure 1A). Levels of nNOS protein detected by immunofluorescence also increased following Fas activation (Figure 1B): the fraction of intensely stained motoneurons increased from 8.7% ⫾ 1.7% to 24.7% ⫾ 3.0% (mean ⫾ SD, n ⫽ 3) at 24 hr after sFasL application (data not shown, see also Figure 3D). Thus, Fas activation in motoneurons leads to the transcription and translation of nNOS. We next asked whether functional nNOS was required for Fas-induced death. Survival of motoneurons was assessed by direct cell counting 48 hr after treatment with agonistic anti-Fas antibody in the presence of NTFs. Fas-triggered death was completely blocked by the wide-spectrum NOS inhibitor L-NAME (Table 1) and by inhibitors specific for nNOS: NPLA (Zhang et al., 1997), LVNIO (Babu and Griffith, 1998), or TRIM (Handy et al., 1995) at 10 ␮M (Figure 1C). Thus, production of NO by nNOS is necessary for Fas to exert its killing effect in motoneurons. Potential Involvement of Superoxide and Peroxynitrite in Fas-Triggered Cell Death Nitric oxide reacts spontaneously with superoxide anion to form peroxynitrite, which has been implicated in several degenerative processes. We therefore tested for the potential involvement of superoxide and peroxynitrite in Fas-mediated death. The superoxide and peroxynitrite scavenger MnTBAP (Faulkner et al., 1994) rescued about 75% of the motoneurons that normally die following Fas activation (Figure 1D). To test whether peroxynitrite was indeed produced after Fas activation, we assayed for tyrosine nitration, a footprint left by the oxidant (Reiter et al., 2000). Following labeling using an antinitrotyrosine antibody, immunoreactivity of randomly selected cells was quantified in a blinded manner using confocal microscopy (Figure 1E). Fluorescence values were grouped into three categories, corresponding to low, medium, and high intensity. In the absence of Fas activators, motoneurons showed only low fluorescence intensity, whereas following Fas activation, nearly all motoneurons were classified in the medium or high categories. L-NAME and MnTBAP each completely blocked the effects of Fas activation. Thus, peroxynitrite is produced downstream of nNOS during Fas-triggered motoneuron death. p38 Kinase Controls nNOS Activation during Fas-Induced Death How is nNOS expression induced following Fas activation? Stress-activated protein kinases such as Jun N-terminal kinases (JNK) and p38 kinase (p38) play major roles in linking membrane receptors to the transcriptional machinery (Paul et al., 1997) and in apoptosis.

Motoneuron-Restricted Death Pathway 1069

Figure 1. Requirement for Nitric Oxide in Fas-Triggered Death of Motoneurons (A) Only the neuronal isoform of NOS is upregulated following Fas activation. Motoneurons were cultured for 16 hr in the presence of an optimal cocktail of neurotrophic factors (referred to as “NTFs”: 1 ng/ml BDNF, 100 pg/ ml GDNF, 10 ng/ml CNTF) and were treated or not with agonistic anti-Fas antibody (JO2, 100 ng/ml) for 24 hr. The expression of different isoforms of NOS (n, neuronal; i, inducible; e, endothelial) was assessed by RT-PCR (35 cycles) followed by Southern blotting with specific internal probes. Concentrations of cDNA were normalized using ␤-actin (24 cycles) as standard. (B) Immunocytochemistry for nNOS on mouse motoneurons following Fas activation. Culture and treatment conditions as in (A). Typical immunofluorescence for each condition is shown; cell counts are given in the text. (C) Specific inhibitors of nNOS save motoneurons from Fas-induced death. After 16 hr of culture, increasing concentrations of the selective nNOS inhibitors NPLA, LVNIO, or TRIM were added at the same time as agonistic anti-Fas antibody (100 ng/ml). Cell survival was measured 48 hr later. (D) The free radical scavenger MnTBAP protects motoneurons from Fas-triggered death. Survival assays were performed as in (C). When added alone, MnTBAP had no effect on cell survival but was toxic above 50 ␮M. Values in (C) and (D) are means ⫾ SD of triplicates, representative of three independent experiments. (E) Fas activation leads to nitrotyrosine formation, a footprint of peroxynitrite production. Motoneurons were cultured on coverslips, fixed 24 hr after the indicated treatments, and immunostained for nitrotyrosine. Immunofluorescence of randomly selected cells was analyzed in a blinded fashion (see Experimental Procedures). Values (arbitrary units) were classified into three categories: low (0–1000), medium (1000–2000), and high (above 2000) intensity.

Furthermore, MAP kinases have been shown to be involved in the regulation of NOS expression (Da Silva et al., 1997). We therefore investigated whether MAP kinase activation played a role in Fas-induced death. We prepared extracts of motoneurons treated or not with sFasL and performed Western blot analysis using specific antibodies to phosphorylated forms of p38 and JNKs (46 and 54 kDa) (Figure 2A). By densitometric scanning, we found a modest but reproducible increase (2.9 ⫾ 0.8-fold, mean ⫾ range, n ⫽ 2) in relative levels of phospho-p38 following Fas activation but no change in the levels of phospho-JNK (Figure 2B). To confirm the functional significance of p38 activation, we tested the ability of different doses of the inhibitor SB203580 to inhibit Fas-triggered motoneuron death. Low concentrations (5 ␮M) of SB203580 block p38 selectively, whereas higher concentrations (30 ␮M) block both p38 and JNK (Le-Niculescu et al., 1999). Fas-induced death was blocked completely by 5 ␮M SB203580, and as many as 65% of Fas-sensitive motoneurons were saved even by 10-fold lower concentra-

tions of inhibitor (Figure 2C), implicating p38 kinase in Fas-induced death. In contrast, the JNK inhibitor L-JNKI1 (Bonny et al., 2001) used at 1 ␮M provided no protection against Fas-induced death, although it did save ⵑ50% of motoneurons from death induced by trophic deprivation (data not shown). We next investigated whether p38 was acting in the novel pathway involving nNOS. Inhibiting p38 activity by 5 ␮M SB203580 strongly inhibited nNOS upregulation following Fas activation (Figure 2D). Moreover, exogenous NO (20 ␮M) restored the Fas killing effect in the presence of 5 ␮M SB203580 (data not shown). Thus, one principal role of p38 in these conditions is to enhance NO production. Fas-Mediated Death Requires Signaling through Daxx-ASK1-p38 We next investigated Fas signaling upstream of p38. One of the MAPK kinase kinases (MAP3K) that can control p38 activation is the apoptosis signal-regulating kinase 1 (ASK1) (Tobiume et al., 2001). Since no specific

Neuron 1070

Table 1. Pharmacologic Analysis of the Requirement for nNOS in Fas-Induced Death Treatment at 1 DIV

Percent Surviving MNs at 3 DIV

SD

NTFs anti-Fas

100 55

4.7 5.6

anti-Fas ⫹ L-NAME L-NAME anti-Fas ⫹ D-NAME D-NAME

96.3 100.7 51 104

5.5 6.5 3 4.1

anti-Fas ⫹ L-NAME ⫹ DETANONOate L-NAME ⫹ DETANONOate anti-Fas ⫹ DETANONOate anti-Fas ⫹ DETA anti-Fas ⫹ DETA ⫹ L-NAME L-NAME ⫹ DETA DETANONOate DETA

54.2 95 54.2 54.9 98 93 93 96.3

3.7 5 5.7 5 3.8 3.5 7.5 9.2

Motoneurons were treated with the indicated reagents (100 ng/ml antiFas antibody, 100 ␮M L-NAME, 100 ␮M D-NAME, 20 ␮M DETANONOate, 20 ␮M of the DETA moiety, see Experimental Procedures). The percentage of surviving motoneurons was estimated after 48 hr. Values are means of three independent experiments ⫾ SD.

inhibitors are available for ASK1, we needed to use a dominant-negative approach. To overcome the relative refractoriness of motoneurons to standard transfection methods, we developed a novel electroporation method for transducing purified motoneurons. As described in Experimental Procedures, purified motoneurons in sus-

pension were briefly preincubated with the DNA to be transfected and then electroporated using a squarewave generator before seeding in culture wells. Although a large percentage of motoneurons were killed outright by the procedure, the remaining cells survived and developed well in culture, showing normal morphology, neurite outgrowth (Figure 3A) and responses to a range of known growth and survival factors as well as to Fas agonists (data not shown). When an expression plasmid encoding EGFP under the control of a strong promoter (CMV) was electroporated, between 50% and 70% of the surviving neurons became strongly fluorescent after less than 24 hr in culture. When two expression plasmids were transfected, the efficacy of coelectroporation was 95% ⫾ 3% (mean ⫾ range, n ⫽ 2; Figure 3A). Thus, this technique should be of general interest for studying signaling mechanisms in neurons. We transfected motoneurons with plasmids coding for EGFP in combination with plasmids coding for either wild-type ASK1 or a dominant-negative form of ASK1 in which a point mutation leads to a catalytically inactive enzyme (ASK1-K709R; [Saitoh et al., 1998]). Whereas overexpression of wild-type ASK1 had no effect on Fasinduced death, overexpression of ASK1-K709R provided complete protection (Figure 3B). These results make ASK1 an excellent candidate for the control of p38 activation in these conditions. Daxx is a Fas-associated protein that has been implicated in Fas signaling in some cell types and can act upstream of ASK1 (Chang et al., 1998). In order to deterFigure 2. Role of p38 in Fas-Induced Death

(A and B) Increase in phosphorylation of p38 kinase but not JNKs following Fas activation. Motoneurons cultured with NTFs were treated or not with sFasL for 2 hr. Phosphorylation of p38 kinase and the JNKs was analyzed on Western blots (A) using antibodies against phosphorylated p38 kinase (p-p38, 42 kDa) and phosphorylated JNKs (p-JNKs, 46 and 54 kDa). Neurofilament (NF-M) immunoreactivity was used as a loading control. Densitometric scanning (B) of Western blots confirmed the specific increase in phosphorylated p38 following Fas activation. Intensity values were corrected for loading differences (monitored through NF-M) and then expressed as the ratio of treated to untreated values. Values are means ⫾ range, n ⫽ 2. (C) The p38 kinase inhibitor SB203580 saves motoneurons from Fas-induced death. Motoneurons were incubated with increasing concentrations of SB203580 added together with agonistic anti-Fas antibody. Motoneuron survival was counted 48 hr later and expressed relative to survival in the presence of NTFs alone. Results represent means ⫾ SD of triplicates from one typical experiment out of three. (D) p38 kinase controls nNOS transcription following Fas activation. Motoneurons in the presence of NTFs were treated or not with agonistic anti-Fas antibody (100 ng/ml) in the presence or absence of SB203580 (5 ␮M). Levels of nNOS were assessed 24 hr later by RT-PCR (increasing number of cycles, 25-30-35 cycles) followed by Southern blotting with specific internal primers. Concentrations of cDNA were normalized using ␤-actin (increasing number of cycles, 15-20-25 cycles) as standard. As controls, total RNA samples incubated without reverse transcriptase (⫺) were submitted to 35 cycles of amplification for nNOS and 25 cycles for ␤-actin. The result presented is representative of three independent experiments.

Motoneuron-Restricted Death Pathway 1071

Figure 3. ASK-1 and DAXX Are Involved in Fas-Mediated Motoneuron Death (A) Freshly purified motoneurons can be efficiently cotransfected. Purified motoneurons were coelectroporated with expression plasmids coding for EGFP and for a HA-tagged form of ASK1, both under the control of a CMV promoter. After 24 hr in culture, motoneurons were immunostained for HA. The fraction of EGFP-expressing motoneurons that were also HA positive was 95% ⫾ 3% (mean ⫾ range, n ⫽ 2). (B) ASK1 is required for Fas-induced motoneuron death. Motoneurons were transfected with an expression vector coding for the EGFP (pEGFP) in combination with vectors coding for either wild-type HA-tagged ASK1 (pHA-ASK1), the dominant-negative form of ASK1 (ASK1-K709R), or an empty vector. Motoneurons from each coelectroporation were seeded in NTFs, treated (or not) 16 hr later with sFasL, and the number of EGFP-positive motoneurons determined 48 hr later. Results are expressed as the percentage of EGFPpositive motoneurons surviving in the presence versus absence of Fas activation. (C) A dominant-negative form of Daxx saves motoneurons from the Fas killing effect. Motoneurons were coelectroporated with a plasmid expressing EGFP and with plasmids coding for full-length Daxx (pDaxx-wt) or a dominant-negative form of Daxx comprising its 110 amino acid C-terminal portion (pDaxx-DN). (D) Daxx is involved in the pathway leading to nNOS upregulation. Same electroporation procedure as in (C). Cells were fixed 24 hr after Fas activation and immunostained for nNOS. The percentage of nNOS/GFP-positive motoneurons was calculated as the number of motoneurons strongly stained for nNOS and positive for EGFP divided by the total number of EGFP-positive motoneurons. (B–D) The results shown are means of three independent experiments. Errors bars indicate SD.

mine whether Daxx was implicated in Fas-induced death, motoneurons were electroporated with plasmids encoding either wild-type Daxx or a dominant-negative form containing only the C-terminal Fas binding domain (Daxx-DN) (Yang et al., 1997). Overexpression of wildtype Daxx had no effect, whereas overexpression of Daxx-DN saved motoneurons from Fas-triggered death (Figure 3C). To determine if Daxx was acting upstream of the new NO pathway, we counted the fraction of motoneurons that showed strong nNOS immunoreactivity. Daxx-DN completely prevented the Fas-induced increase in the number of nNOS-expressing cells (Figure 3D). Moreover, addition of exogenous NO totally abrogated the protective effects of ASK1-K709R and DaxxDN (data not shown), further confirming that NO production is downstream of Daxx, ASK1, and p38 kinase. Interactions between Pathways Involving Daxx/p38 and FADD/Caspase-8 The existence of the novel Fas/Daxx/Ask/p38/nNOS pathway led us to examine its functional relationship with the classical Fas pathway involving FADD-mediated activation of procaspase-8 and cytochrome c release. The classical FADD/caspase-8 pathway is indeed involved in Fas-triggered motoneuron death, since cell death can be prevented by the caspase-8 inhibitor peptide IETD (Raoul et al., 1999) and considerably reduced by overexpression of a dominant-negative form of FADD (G. Ugolini, A.O. Hueber et al., personal communication). To begin to understand the relationship between the two pathways, we therefore blocked elements in one

pathway and examined the downstream effects on the other pathway. We first focused on activation of procaspase-8, which in motoneurons involves cleavage of the inactive intermediate form (p43) into active caspase-8 (p18). When Fas-induced death was blocked using nNOS inhibitors, caspase-8 was activated normally (Figure 4A), suggesting that any effect of the Daxx/p38 pathway on caspase-8 activation must lie upstream of nNOS. We therefore inhibited p38 activity using SB203580: caspase-8 activation was completely prevented (Figure 4B). These data indicate that p38, one element of the Fas/ NO pathway, can regulate the caspase-8 pathway. We next examined the release of cytochrome c from mitochondria. To investigate the occurrence of this phenomenon in motoneurons, we followed the subcellular distribution of cytochrome c by immunolabeling (Figures 4C and 4D). In the absence of Fas activation, more than 85% of motoneurons showed a granular pattern of cytochrome c immunoreactivity, consistent with a predominantly mitochondrial localization in these conditions. Following Fas activation, however, 40% of the neurons displayed a more diffuse pattern already after 30 hr (Figure 4E), consistent with cytochrome c release into the cytoplasm. After inhibition of nNOS or inhibition of caspase-8, the fraction of motoneurons with diffuse cytochrome c was reduced by only 63% ⫾ 6% and 60% ⫾ 10% (means ⫾ SD, n ⫽ 5), respectively. In contrast, when using the p38 inhibitor SB203580 the Fas-induced cytochrome c release was completely blocked (Figure 4E). These observations suggest that p38 not only con-

Neuron 1072

Figure 4. Synergistic Interactions between the Nitric Oxide and Caspase-8 Pathways (A and B) Fas-triggered activation of caspase-8 in motoneurons is controlled by p38. Motoneurons were treated or not with the indicated reagents (100 ng/ml agonistic antiFas antibody, 100 ␮M L-NAME [A] or 5 ␮M SB203580 [B]). Inactive and active caspase8 were analyzed by immunoblotting with anticaspase-8 (p43) or anti-active-caspase-8 (p18) antibodies, respectively. ␣-tubulin immunoreactivity was used as a loading control. Western blots shown are representative of two independent experiments. (C and D) Fas activation triggers cytochrome c release from mitochondria. Motoneurons were treated with agonistic anti-Fas antibody (100 ng/ml) and, 30 hr later, immunostained with anti-cytochrome c antibody. Healthy MNs show punctate labeling of the mitochondria (C), whereas following Fas activation, a proportion of the motoneurons show more diffuse labeling but no pyknotic nucleus (D). (E) The p38 kinase controls mitochondrial cytochrome c release. Motoneurons were treated as in (C) and (D), with or without 5 ␮M SB203580, 10 ␮M L-VNIO, or 1 ␮M IETD-fmk. Cytochrome c subcellular localization was classified as either punctate mitochondrial or diffuse cytoplasmic. The percentage of motoneurons showing diffuse labeling was determined by direct counting under the fluorescence microscope. In the absence of Fas activation, the inhibitors did not modify the percentage of motoneurons with diffuse cytochrome c labeling (data not shown). (F) Synergistic activation of the FADD-caspase-8 and the NO pathway contributes to Fas-triggered death of motoneurons. Motoneurons were treated with the same reagents as in (C)–(E). At 3 DIV, cell survival was estimated by direct counting of living motoneurons. After counting, motoneurons were treated again with the same doses of reagents, and surviving motoneurons were counted at 5 DIV. Results show mean values of three independent experiments, each done in triplicate. Error bars indicate SD.

trols caspase-8 but also other downstream elements leading to cytochrome c release. Further evidence for interactions between signaling pathways is provided by data in Figure 1E and Table 1. NO alone did not trigger peroxynitrite production or cell death. However, when Fas was activated but prevented from triggering endogenous NO production by addition of L-NAME, exogenous NO led both to peroxynitrite production and to cell death. Thus, parallel Fas signaling mechanisms are required for NO to be toxic in normal motoneurons. These must be either upstream of or independent from caspase-8, since IETD did not reduce Fasinduced peroxynitrite production. Given these interactions, we looked for survival differences in conditions in which only one or both pathways were activated. We analyzed motoneuron survival 4 days after Fas-agonist addition (at 5 DIV), i.e., 2 days later than in all previous experiments. Under these conditions, single inhibitors had no or only modest effects on Fas-induced death (Figure 4F). However, when the

caspase-8 inhibitor was used in combination with an nNOS inhibitor or with the p38 inhibitor, Fas-triggered motoneuron death was completely prevented, suggesting that the FADD-caspase-8 and Daxx/p38 pathways act synergistically. Taken together, these findings demonstrate that, in order to trigger motoneuron cell death, Fas needs to activate, in addition to the classical FADD-caspase-8 cascade, a novel pathway leading through activation of Daxx, ASK1, and p38 to transcriptional upregulation of the nNOS gene (Figure 9). Cellular Specificity of the Involvement of Nitric Oxide in Fas-Triggered Cell Death Since the Daxx/Ask1/p38/nNOS pathway (Figure 9) had not been previously described, we looked for upregulation of nNOS in other cell types sensitive to Fas activation in the absence of transcriptional inhibitors. When 3T3 Swiss fibroblasts were treated with agonistic antiFas antibody without actinomycin D, about 20% of the cells died after 6 hr of treatment (Figure 5A), and as

Motoneuron-Restricted Death Pathway 1073

Figure 5. Requirement for the Nitric Oxide Pathway in Fas-Induced Death Is Restricted to Motoneurons (A) Fas-induced death of fibroblasts is independent of the NO pathway. Swiss 3T3 cells grown to near confluence were preincubated with or without 100 ␮M L-NAME, 100 ␮M D-NAME, 50 ␮M MnTBAP, or 30 ␮M DEVDfmk and then treated with agonistic anti-Fas antibody (100 ng/ml) for 16 hr. Cell survival was estimated using the MTT assay. (B) Expression of different NOS isoforms after Fas activation (agonistic anti-Fas antibody, 100 ng/ml) was assessed using RT-PCR (35 cycles for nNOS, iNOS, and eNOS) followed by Southern blotting. cDNA levels were normalized using ␤-actin (24 cycles). (C) Fas-induced death of thymocytes is independent of the NO pathway. Freshly-isolated thymocytes from P5 mice were treated or not (basal) for 24 hr with agonistic anti-Fas antibody (100 ng/ml) in the presence or absence of 100 ␮M of L-NAME, 100 ␮M of D-NAME, 50 ␮M of MnTBAP, or 30 ␮M of IETD-fmk. Cell survival was estimated by direct counting, following incubation with the vital dye calcein AM and ethidium homodimer-1 (EthD-1). Results are expressed as the ratio of living cells (calcein positive) to the sum of living cells (calcein positive) and of dying cells (EthD-1 positive). (D) Fas activation in thymocytes does not lead to upregulation of NOS. Primary thymocytes were incubated or not with agonistic anti-Fas antibody (100 ng/ml) for 16 hr. Expression of the three different isoforms of NOS was analyzed by immunoblotting. (E) Fas-induced death of mouse cortical neurons is independent of the NO pathway. Primary cortical neurons from E15.5 cerebral cortex were maintained in culture for 4 days before being incubated or not with the indicated reagents (100 ng/ml agonistic anti-Fas antibody, 100 ␮M L-NAME, 100 ␮M D-NAME, 50 ␮M MnTBAP, or 30 ␮M IETD-fmk). After 24 hr of treatment, cells were fixed and stained with DAPI; the percentage of cell death was calculated as the ratio of fragmented/condensed nuclei to total nuclei. The asterisks indicate statistically significant differences (p ⬍ 0.001, n ⫽ 12) by Student’s t test. (F) Fas activation does not modify NOS expression in cortical neurons. Cortical neurons cultured for 4 days were treated or not (basal) with agonistic anti-Fas antibody (100 ng/ml) for 16 hr. Proteins were extracted, separated by SDS-PAGE, and immunoblotted with antiiNOS, anti- eNOS, or anti-nNOS antibodies as in (D). (A, C, and E) Results show typical values of at least three independent experiments. Error bars indicate SD.

many as 80% died after 18 hr; as expected, they were saved by the caspase inhibitor DEVD (Figure 5A). However, no expression of nNOS was detected by RT-PCR after 16 hr, and no change in the constitutive levels of expression of iNOS occurred (Figure 5B). Moreover, the death of 3T3 fibroblasts was not prevented by the NOS inhibitor L-NAME or the scavenger MnTBAP. Similar re-

sults were obtained using primary cultures of thymocytes (Figures 5C and 5D). This suggested that the NO pathway might be characteristic of neurons. We therefore tested embryonic cortical neurons, the only other neuronal culture system reported to be sensitive to Fas activation (Cheema et al., 1999; Ciesielski-Treska et al., 2001; Morishima et al.,

Neuron 1074

2001). Given the relatively small reduction in overall survival in these cultures, we counted the percentage of neurons showing condensed chromatin (as detected by DAPI staining). In agreement with published data, there was a significant increase in pyknotic nuclei 24 hr after Fas activation (Figure 5E). In surprising contrast with the situation in motoneuron cultures, however, we could detect no increase in levels of nNOS by Western blotting (Figure 5F) or RT-PCR (data not shown). This could potentially reflect the low percentage of dying neurons compared to healthy cells. We therefore asked whether inhibitors that completely protect motoneurons affected Fas-triggered death of cortical neurons. Strikingly, nNOS or p38 inhibitors had no protective effect (Figure 5E and data not shown), although Fas-triggered death was completely prevented by IETD (Figure 5E). Thus, motoneurons are currently the only cell type in which the NO pathway downstream of Fas is clearly involved in triggering death. Motoneurons from SOD1 Mutant Mice Show Increased Sensitivity to Activation of the Nitric Oxide Pathway Downstream of Fas Given the existence of a signaling pathway leading specifically to death of developing motoneurons, we next explored the possibility that this pathway might be involved in pathological motoneuron death. As a model for familial amyotrophic lateral sclerosis (ALS), we used transgenic mice overexpressing the ALS-linked G93A (Gurney et al., 1994), G37R (Wong et al., 1995), or G85R (Bruijn et al., 1997) SOD1 mutants. We asked whether mutant motoneurons might show altered responses to Fas activation. The following controls were made for these experiments. (1) Transgenic mice overexpressing G93A, G37R, G85R, or wild-type (wt) SOD1 were all bred on the same genetic background (C57Bl/6). (2) Mutant and control motoneurons, when prepared in parallel from individual E12.5 embryos, displayed identical yield and viability in the presence of neurotrophic factors (see Experimental Procedures). (3) Expression of the SOD1 transgenes in motoneurons cultured for 1 day was verified using an antibody that recognizes human but not mouse SOD1. Strong hSOD1 immunoreactivity was observed in G93A, G37R, and wt SOD1 motoneurons, whereas only low signal was detected in G85R motoneurons (Figures 6A–6C and data not shown). This is consistent with reports showing that levels of the unstable G85R SOD1 protein in spinal cord (Bruijn et al., 1997) are about 10to 20-fold lower than for the G93A (Gurney et al., 1994) and G37R (Wong et al., 1995) proteins. To study the sensitivity of mutant motoneurons to Fas activation, anti-Fas antibody was added to cultures after 16 hr at final concentrations between 0.01 and 100 ng/ ml, and survival was determined 2 days later. Treatment of G93A mutant motoneurons with maximal concentrations of anti-Fas antibody led to death of the same fraction of motoneurons (40%–50%) as in control cultures. However, the dose-response curve was different. At intermediate levels of Fas activation (0.1–1 ng/ml anti-Fas antibody), the number of surviving G93A motoneurons was significantly lower than that of control motoneurons (Figure 7A). Similar observations were made using G85R

and G37R mutant motoneurons (Figures 7B and 7C). Concentrations of anti-Fas antibody required to induce the half-maximal effect (EC50), calculated from eight different experiments, were 11.3 ⫾ 3.4-fold (mean ⫾ SEM) lower for mutant than for control motoneurons (Figure 7F). Thus, mutant SOD1 significantly exacerbates the sensitivity of motoneurons to Fas-mediated death. To further confirm the increased Fas sensitivity of mutant SOD1 motoneurons, we also assessed the number of motoneurons undergoing apoptosis by combined staining with DAPI and an antibody (CM1) that recognizes activated caspase-3 (Srinivasan et al., 1998b), see Figure 7G. The percentage of CM1-positive motoneurons with pyknotic nuclei was quantified 25 hr after treatment (or not) with an intermediate dose of agonistic anti-Fas antibody (0.5 ng/ml) when the total number of motoneurons did not yet differ between mutant SOD1 and control cultures. Upon Fas activation, the percentage of motoneurons with activated caspase-3 and nuclear condensation increased in G85R motoneuron cultures to a significantly higher extent than in control cultures (Figure 7H). To exclude the possibility that this simply reflected overexpression of human SOD1, we also challenged motoneurons from mice overexpressing wild-type SOD1 with anti-Fas antibody. In striking contrast to the results with mutant SOD1, overexpression of wt SOD1 conferred nearly complete protection against Fas-induced death (Figure 7D). Concentrations of anti-Fas antibody required to kill motoneurons overexpressing wt SOD1 were at least 1000-fold higher than those required to trigger death of motoneurons expressing comparable levels of mutant SOD1 (Figure 7E). Since G37R (Borchelt et al., 1994) and G93A (Rabizadeh et al., 1995) mutants are fully catalytically active, the difference between Fas susceptibilities of mutant and wt SOD1 motoneurons must reflect the gain of toxic function previously deduced from genetic analysis of these mice (Bruijn et al., 1997; Gurney et al., 1994; Wong et al., 1995). We then used pharmacological inhibitors to confirm that mutant motoneurons were triggered to die by the pathway we had defined using normal motoneurons. Inhibiting NOS, p38, or caspase-8 completely protected mutant SOD1 motoneurons against death triggered by 0.5 ng/ml or 100 ng/ml anti-Fas (Figure 8A and data not shown). The free radical scavenger MnTBAP provided partial (75%) protection but to an indistinguishable extent in mutant and wild-type cultures (Figure 8A). These results indicate that the same pathway was involved in the Fas-triggered death of mutant motoneurons. A characteristic feature of the motoneuron-specific pathway is the induction of nNOS. However, addition of NO is not itself sufficient to trigger death of normal motoneurons (Table 1 and Estevez et al., 1999). We asked whether this might be different in mutant motoneurons. Mutant and control motoneurons were cultured for 16 hr in the presence of neurotrophic factors, exposed to increasing concentrations (1–20 ␮M) of the nitric oxide donor DETANONOate, and counted 1 day later. In striking contrast with control motoneurons, about 30% of mutant G85R and G93A motoneurons were triggered to die at 5 ␮M DETANONOate and about 50% at 20 ␮M DETANONOate (Figures 8B and 8C). Thus, activation of the NO pathway at two different levels

Motoneuron-Restricted Death Pathway 1075

Figure 6. Embryonic Motoneurons from Transgenic SOD1 Mice Express Human SOD1 in Culture (A–F) Double immunolabeling for human SOD1 (A–C) and neurofilament-145 (NF-M) (D–F) in cultured motoneurons from transgenic mice for SOD1 mutants G93A (A and D), G85R (B and E), or G37R (C and F). Intense hSOD1 labeling was observed in the cytoplasm, axons, and dendrites of G93A and G37R motoneurons and also of transgenic wt SOD1 motoneurons (data not shown), whereas hSOD1 immunoreactivity was low in G85R motoneurons and undetectable in control motoneurons (data not shown). Scale bar, 25 ␮m.

results in remarkably exacerbated killing of mutant SOD1 motoneurons. Specificity of Mutant SOD1-Related Toxicity In order to exclude the possibility that cultured SOD1 mutant motoneurons show a generalized increase in susceptibility to death-inducing agents, we explored their responses to trophic deprivation and excitotoxicity. When seeded in the absence of neurotrophic factors, about 75% of control motoneurons died by 3 DIV. Under these conditions, cell death occurred to an identical extent in parallel cultures of G37R, G85R, and wt SOD1 motoneurons (Figure 8D). Next, since excessive stimulation of glutamate receptors is potentially involved in ALS pathogenesis (Rothstein, 1995), we triggered excitotoxic death of motoneurons by adding the AMPA receptor agonist domoic acid to cultures maintained for 5 DIV with neurotrophic factors (Cisterni et al., 2001). Domoic acid induced a dose-dependent death of motoneurons that was indistinguishable between control, G85R, G93A, and wt SOD1 cultures (Figure 8E, data not shown for G93A). In vivo pathology in mutant SOD1 mice is relatively selective for motoneurons. In order to confirm that increased Fas and NO sensitivity in vitro was similarly cell type specific, we prepared cultures from different tissues of control, mutant, and wt SOD1 mice (Table 2). Thymocytes and cortical neurons, both of which are Fas sensitive (Cheema et al., 1999; Ogasawara et al., 1995), did not show increased Fas sensitivity when prepared from control, G85R, G93A, or wt SOD1 mice (Table 2 and data not shown). DRG sensory neurons showed very low Fas sensitivity with no statistically significant difference between mutants and controls. Last, cerebellar granule neurons and astrocytes, which are normally not Fas sensitive, did not become Fas sensitive in the mutants. The NO susceptibility of the mutant cells was similarly cell type specific: no difference in the response to DETANONOate was observed between cortical neu-

rons and thymocytes from mice overexpressing mutant and wild-type SOD1 mice at any concentration (data not shown). Thus, the exacerbated sensitivity of mutant motoneurons to Fas-activation and NO strikingly mirrors the cell type specificity of the disease process in vivo. Discussion Our main hypothesis in this work was that death pathways restricted to specific classes of neurons could potentially provide a basis for better understanding the precise control of developmental neuronal death and the selectivity of neuronal loss in patients with neurodegenerative diseases. By studying the signaling mechanisms used by the Fas receptor in primary motoneurons, we have defined a novel pathway whose key step is the transcriptional regulation of nNOS downstream of p38 kinase. Strikingly, we could not find evidence for this pathway in other cell types. This led us to ask whether such a mechanism might be involved in the selective motoneuron degeneration observed in mice that overexpress ALS-linked mutant forms of SOD1. Motoneurons purified from these mice show a specifically exacerbated response to exogenous Fas agonists and NO. Thus, although involvement of the NO pathway in vivo remains to be confirmed, our findings provide a rationale for the selective vulnerability of motoneurons that is characteristic of this model of ALS. Cell type-dependent differences in Fas signaling pathways have previously been reported, depending on the intensity of caspase-8 activation and the degree of involvement of the mitochondrion (Scaffidi et al., 1998, 1999). In Type I cells, activation of caspase-8 at the DISC is directly followed by caspase-3 activation independently of mitochondrial function. In Type II cells, caspase-8-dependent death can be blocked by overexpression of antiapoptotic members of the Bcl-2 family, reflecting in some cases a requirement for cleavage of the BH3-only family member Bid (Li et al., 1998; Luo et

Neuron 1076

Figure 7. Mutant SOD1 Expression in Embryonic Motoneurons Causes Increased Sensitivity to Fas Activation (A–F) Motoneurons from transgenic mice for mutant SOD G93A (A), G85R (B), G37R (C) display increased sensitivity to Fas-triggered death, whereas wt SOD1 motoneurons (D) are almost completely protected. Varying concentrations of an agonistic anti-Fas antibody were added to motoneurons from transgenic or control mice and survival assessed at 3 DIV and expressed as the percentage of the number of motoneurons surviving at the same time without anti-Fas antibody addition. Each figure is representative of one out of two to three experiments performed in quadruplicate for each condition. Error bars represent SEM. Asterisks indicate the level of statistical significance for the difference between mutant and control motoneurons at a given concentration of anti-Fas as tested by Student’s t test, two-tailed, unpaired: *p ⬍ 0.05, **p ⬍ 0.01, ***p ⬍ 0.001. No asterisk, no statistical difference. Note that all cultures were performed in the presence of optimal concentrations of neurotrophic factors. (E) Cumulative data for the Fas sensitivity of mutant SOD1 motoneurons in comparison to motoneurons overexpressing wild-type SOD1. (F) Mean concentrations of anti-Fas antibody required to trigger death of 50% of Fas-sensitive motoneurons (EC50, in ng/ml) were significantly lower (p ⬍ 0.025 by Student’s t test, two-tailed, paired) in G93A, G85R, and G37R cultures than in parallel control cultures. Histograms and error bars represent means ⫾ SEM. (G and H) Increased apoptosis in G85R motoneuron cultures following Fas activation. Cultures were treated (or not) with agonistic antiFas antibody (0.5 ng/ml) for 25 hr and stained with DAPI and an antibody against activated caspase-3 (CM-1, in red). (G) Apoptotic motoneurons with strong diffuse CM-1 labeling, neurite breakdown, and nuclear condensation (red arrow) could be easily distinguished from healthy motoneurons displaying weak CM1 labeling and normal chromatin structure (white arrows). Scale bar, 25 ␮m. (H) The fraction of CM1-positive pyknotic motoneurons increased in G85R cultures from 7.7% ⫾ 1.3% (mean ⫾ SEM) to 21.9% ⫾ 0.7% upon anti Fas treatment, i.e., to a significantly higher extent than in control cultures (from 8.9% ⫾ 1.8% to 16.4% ⫾ 0.4%; n ⫽ 3 per condition, **p ⬍ 0.006 by Student’s t test, two-tailed, unpaired).

al., 1998). In these cells, caspase-3 is activated as a result of cytochrome c release from mitochondria. Primary motoneurons correspond to neither Type I nor Type II cells, and we propose to classify them as Type III cells. In these, caspase-8 activation is not sufficiently intense to alone cause rapid cell death and requires coactivation of p38 and transcription of nNOS. It is

tempting to speculate that during evolution, the transcriptionally regulated NO pathway has been grafted onto the classical cascade as a means of providing finer control of the death of these essentially irreplaceable neurons. Other members of the same receptor family as Fas, such as TNFR1 and p75NTR, can also trigger neuronal death (Barker et al., 2001; Frade and Barde,

Motoneuron-Restricted Death Pathway 1077

Figure 8. Specific Killing of Mutant SOD1 Motoneurons by Nitric Oxide (A) Fas-induced death of both control and mutant G85R and G93A motoneurons was significantly inhibited by the reactive oxygen radical scavenger MnTBAP (50 ␮M) and completely blocked by the general NOS inhibitor L-NAME (100 ␮M), the neuronal NOS inhibitor L-VNIO (10 ␮M), and the caspase-8 inhibitor IETD-fmk (1 ␮M). (B and C) Mutant motoneurons are killed by NO. Mutant G93A (B) or G85R (C) motoneurons were exposed to DETANONOate (1–20 ␮M; plotted linearly) 16 hr after seeding and survival quantified at 2 DIV. Asterisks indicate level of statistical significance as assessed by Student’s t test, two-tailed, unpaired: *p ⬍ 0.05, **p ⬍ 0.01, ***p ⬍ 0.001. No asterisk, no statistical difference. (D) Trophic factor deprivation-induced death is similar for control, G37R, G85R, and wt SOD1 motoneurons. Cell survival in the presence of trophic factors at 3 DIV did not differ between the indicated genotypes and was set to 100%; only the histogram for control motoneurons is shown. (E) Excitotoxic death of motoneurons is similar for control, G85R, and wt SOD1 motoneurons. Cells were treated at 5 DIV with domoic acid (10⫺10 to 10⫺6 M), and survival was quantified at 7 DIV. In these experiments, domoic acid-induced cell death was completely prevented by CNQX or NBQX, two different competitive receptor antagonists (data not shown). (A–E) Data are each representative of at least two independent experiments performed in triplicate or quadruplicate. Error bars represent SD (A) or SEM (B–E).

1998; Raoul et al., 2000; Robertson et al., 2001; Terrado et al., 2000). It will be interesting to see whether these receptors, too, have developed neuron-specific signaling mechanisms. Other elements of the NO pathway defined here had been reported in the literature, but their involvement in physiologically relevant Fas signaling was in some cases controversial. Daxx was first identified as a Fas interactor that could activate JNK and apoptosis in HeLa and embryonic kidney cell lines (Yang et al., 1997). Although Daxx is expressed in the brain (Yang et al., 1997), Daxx knockout mice die before E9.5, precluding analysis of a role in neuronal death (Michaelson et al., 1999). The primary target of Daxx is the MAP3K ASK1 (Chang et al., 1998). ASK1 is essential for Fas-induced apoptosis in several cell lines (Chang et al., 1998) and for death of sympathetic neurons deprived of NGF (Kanamoto et al., 2000). In motoneurons, the key member of the SAP ki-

nase family activated downstream of ASK1 is p38 kinase, which has been implicated in many neuronal cell death mechanisms (Harper and LoGrasso, 2001; Mielke and Herdegen, 2000). Although JNK is active in motoneurons (Borasio et al., 1998; Maroney et al., 1998), levels of phospho-JNK were not modulated by Fas, and Fas-triggered death was not prevented by JNK1 inhibitor. In contrast, there was a reproducible though modest increase in relative levels of phospho-p38 following Fas activation, and nNOS upregulation, cytochrome c release, and Fas-triggered death were all blocked by low concentrations of p38 inhibitor SB203580. Activation of ASK1 often leads to activation of both JNK and p38 (Tobiume et al., 2001), but some reports indicate that they can be activated independently, depending on the specific MAP2K activated by ASK-1 (for reviews, see Mielke and Herdegen, 2000; Ono and Han, 2000). p38 can activate numerous substrates (for review, Harper

Neuron 1078

Table 2. Fas Sensitivity of Different Cell Types from Transgenic SOD1 and Control Mice Cell Survival (% ⫾ SEM) Cell Type

anti-Fas (ng/ml)

Control

Thymocytes

0.1 1 10 100 1000

87 83 66 63 53

DRG neurons

1 10 100

108.4 ⫾ 4.5 107.9 ⫾ 1.1 91.5 ⫾ 2.0

100 ⫾ 4.3 93.3 ⫾ 6.1 87.6 ⫾ 2.1

Cerebellar granule neurons

100 1000

96.5 ⫾ 1.1 97.2 ⫾ 2.6

101.7 ⫾ 1.7 96.4 ⫾ 1.7

Astrocytes

100 1000

104.8 ⫾ 2.6 103.6 ⫾ 1.2

98.7 ⫾ 1.2 95.2 ⫾ 1.5

⫾ ⫾ ⫾ ⫾ ⫾

1.4 1.4 1.2 1.4 1.2

G85R SOD1 90 90 83 77 76

⫾ ⫾ ⫾ ⫾ ⫾

0.6 0.6 1.4 1.5 2.3

wt SOD1 91 89 83 73 69

⫾ ⫾ ⫾ ⫾ ⫾

1 1 2.3 2.6 3.5

Cell Death (% ⫾ SEM) Cell Type

anti-Fas (ng/ml)

Control

G85R SOD1

wt SOD1

Cortical neurons

0.5 100

9.8 ⫾ 1.2 17.6 ⫾ 0.6

9.8 ⫾ 1.2 16.3 ⫾ 1.2

9.6 ⫾ 0.9 14.9 ⫾ 0.6

After treatment with anti-Fas antibody, cultures of different cell types from transgenic SOD1 mice did not show greater cell death than control cultures. For thymocytes, cell viability was estimated using a live/dead assay; for DRG neurons, cerebellar granule neurons, and astrocytes, survival was assessed using the MTT assay; for cortical neurons, percentage of cell death was determined after DAPI staining.

and LoGrasso, 2001), including transcription factors such as ATF-2 or CREB (Tan et al., 1996). Interestingly, transcription of the nNOS gene is CREB dependent in other systems (Sasaki et al., 2000). Our demonstration of a motoneuron-specific death pathway raised the question of its possible involvement in a condition of pathological motoneuron death. Using motoneurons isolated from three different transgenic mouse lines harboring ALS-linked SOD1 mutations, we indeed demonstrated enhanced susceptibility to activation of the Fas pathway leading to activation of caspase-8, caspase-3, and ultimately death. These abnormalities preceded by about 2 weeks the earliest reported motoneuron loss in mutant SOD1 spinal cord (Lowry et al., 2001) and by about 2 months the occurrence of other histopathological and electrophysiological changes (Frey et al., 2000; Kennel et al., 1996). Our findings do not necessarily mean that mutant SOD1 motoneurons are already affected during embryonic development, since the concentrations of death inducers used here may be higher than those encountered in vivo, and levels of intrinsic signaling intermediates may be upregulated as a result of culturing. Our results suggest that mutant SOD1 toxicity affects Fas-triggered death at a level downstream of nNOS but upstream of the late-stage apoptotic cascade: mutant motoneurons displayed increased susceptibility to exogenous NO but normal cell death responses to trophic deprivation and excitotoxicity. This suggests a key role for NO in triggering mutant motoneuron death. The nNOS inhibitor AR-R 17,477, when orally administered from day 30, has been shown to significantly prolong the life span of G93A mutant mice (Facchinetti et al., 1999). Other inhibitors however were ineffective (Facchinetti et al., 1999; Upton-Rice et al., 1999), maybe due to low bioavailability in vivo. Moreover, crossing of mutant SOD1 mice with mice carrying a loss-of-function allele

of the nNOS gene did not prolong their survival (Facchinetti et al., 1999). Nevertheless, nNOS mutant mice are not complete nulls and still produce ␤ and ␥ isoforms of nNOS, leading to reduced but still significant overall levels of nNOS catalytic activity (Facchinetti et al., 1999). Strikingly, specific increases in the same nNOS isoforms have been reported in spinal cords from sporadic ALS patients (Catania et al., 2001). Thus, published results do not exclude the involvement of NO, which is suggested by our in vitro results and the pharmacological effects of AR-R 17,477 in vivo. Neurodegeneration in transgenic SOD1 mice in vivo has only been observed when mutant SOD1 was expressed ubiquitously under the control of its own promoter (Bruijn et al., 1997; Gurney et al., 1994; Wong et al., 1995); specific overexpression in neurons (Lino et al., 2002; Pramatarova et al., 2001) or in astrocytes (Gong et al., 2000) did not lead to pathology. Our in vitro results indicate that mutant SOD1-related motoneuron death proceeds cell-autonomously but can be triggered by diffusible factors such as Fas-agonists and NO. These factors might have different sources (Figure 9). Activated microglia in culture have been shown to produce NO and FasL (Ciesielski-Treska et al., 2001, and references therein). Astrocytes have been shown to contain increased levels of NOS in mutant SOD1 mice and also in ALS patients (Almer et al., 1999; Anneser et al., 2001; Catania et al., 2001). Sera have been reported to contain high levels of agonistic Fas antibodies in some sporadic ALS patients (Yi et al., 2000). This opens the possibility that NO, Fas agonists, or other diffusible factors, produced outside motoneurons, might trigger or amplify neurodegeneration. Finally, it will be of obvious interest to examine the specific role of the Fas pathway in different experimental models of in vivo motoneuron degeneration. Our preliminary results show that traumatic neurodegeneration is

Motoneuron-Restricted Death Pathway 1079

mice were obtained from Iffacredo (L’Arbresle, France). Transgenic mice for mutant G85R, G37R, or wild-type human SOD1 were from Ludwig Institute (San Diego); transgenic mice for G93A human SOD1 were obtained from Transgenic Alliance (L’Arbresle, France). Mutant G93A (line G1H [Gurney et al., 1994]), G37R SOD1 mice (line 42 [Wong et al., 1995]), and transgenic mice for wild-type human SOD1 (line 76 [Wong et al., 1995]) were maintained as hemizygotes by crossing transgenic males with nontransgenic females of the same genetic C57BL/6 background. Offspring were genotyped by separate PCR reactions for human SOD1 exon 3 (PCR primers 5⬘-TTC TGTTCCCTTCTCACTGT and 5⬘-TCCCCTTTGGCACTTGTATT) and exon 5 (primers 5⬘-TGTTGGGAGGAGGTAGTGATTA and 5⬘-AGCA GAGTTGTGTTAGTTTTAG). In each PCR reaction, the mouse globin gene was amplified with primers 5⬘-GATCATGACCGCCGTAGG and 5⬘-CATGAACTTGTCCCAGGCTT. For G93A, G37R, and wt SOD1 embryos, transgenic and control embryos were from the same litters. During PCR genotyping of embryonic tail DNAs (3–4 hr), embryos were kept at 4⬚C in Hibernate E medium (Gibco). Mutant G85R mice (line 148 [Bruijn et al., 1997]) were maintained as homozygotes. Hemizygous G85R embryos were obtained from crosses between male G85R mice and female C57BL/6 mice, and age-matched C57BL/6 embryos were used as controls.

Figure 9. A Cell-Autonomous Death Pathway that Can Be Triggered by Environmental Stimuli The presence of mutant SOD1 sensitizes motoneurons downstream of NO. Astrocytes and microglia represent potential sources of cell death triggers.

partly prevented in mice mutant for Fas and its signaling components (G. Ugolini et al., personal communication). The next step will be to cross mutant SOD1 mice to lines in which Fas activity is attenuated; similar experiments could be performed using mouse models for spinal muscular atrophy (SMA). The standard Fas knockout mice may not be appropriate for these crosses, as they still express high levels of a truncated Fas protein (Adachi et al., 1995), which may have kept some of its signaling capacity for this novel pathway, in particular the interaction with Daxx (Yang et al., 1997). It will also be of interest to test the therapeutic efficacy of p38 inhibitors in ALS and SMA models. Precise control of neuronal cell death is a prerequisite for normal brain development, and selective loss of specific neuronal populations is one of the key common features of neurodegenerative diseases such as Alzheimer’s, Parkinson’s, or Huntington’s disease. In most familial forms of these diseases, mutant proteins are expressed in unaffected brain regions and thus are unlikely to themselves restrict the site of pathology. The model we have illustrated here for the specific case of motoneurons and mouse models of ALS may be more generally applicable: neuronal class-specific pathways for degeneration or death, potentially sensitized by genetic or environmental factors. If such pathways are indeed implicated in human pathologies, they may provide important targets for selective therapeutic intervention. Experimental Procedures Animals Mice were kept in an inverted dark/light cycle (dark 11 am to 9 pm) and positive vaginal plugs at 9 hr am recorded as E0.5. Normal CD1

Reagents Hamster anti-mouse Fas antibody (JO2 clone) and mouse monoclonal anti-cytochrome c (6H2.B4 clone) were purchased from Pharmingen International, San Diego, CA. z-DEVD-fmk, z-IETD-fmk, LEHD-CHO, manganese (III) 4,4⬘,4⬘⬘,4⬘⬘⬘ (21H,2H-porphine-5,10,15,20tetrayl) tetrakis (benzoic acid), referred to as MnTBAP, and the pyridinyl imidazole p38 inhibitor SB203580 were purchased from Calbiochem (La Jolla, CA). Rabbit polyclonal antibodies against nNOS (lot# 14552a) and DETANONOate were purchased from Cayman Chemical, Ann Arbor, MI. Production and characterization of affinitypurified rabbit polyclonal antibodies to nitrotyrosine were described elsewhere (Beckman et al., 1994). Anti-active p38 and anti-active JNK rabbit polyclonal antibodies were purchased from Promega; anti-p38 and anti-JNK rabbit polyclonal antibodies were from New England Biolabs Inc. Rabbit polyclonal antibodies against inducible NOS (lot# L03202), endothelial NOS (lot# 04766), soluble Fas-Ligand (sFasL), enhancer antibodies for tagged sFasL, rhAPO-1/Fas:FcIgG (Fas-Fc), LG-Nitro-L-arginine-methyl ester.HCl (L-NAME), LG-NitroD-arginine-methyl ester.HCl (D-NAME), [N5-[Imino(propylamino) methyl]-L-ornithine; N⍀-Propyl-L-arginine] (NPLA), [N5-(1-Imino-3butenyl)-L-ornithine; Vinyl-L-NIO] (L-VNIO), [1-(2-Trifluoromethylphenyl)imidazole] (TRIM) were from Alexis Corp., Switzerland. Polyclonal antibodies against hemagglutin (HA)-tag were from Clontech Laboratories, Inc. Rabbit polyclonal antibodies against Fas (M-20) and Fas-Ligand (N-20) were from Santa Cruz Biotechnology; antiNF-M (Ab1987) from Chemicon; mouse monoclonal anti-SOD1 (clone SD-G6) anti-NF 160 (clone NN18) were from Sigma. Expression Constructs The cDNA coding for full-length mouse Daxx, referred to as Daxxwt, or for the C-Terminal 110 amino acids of Daxx, referred to as Daxx-DN (Yang et al., 1997), were cloned into pcDNA3 (Invitrogen). HA-tagged human ASK1 (HA-ASK1-wt) and the catalytically inactive mutant ASK1-K709R (Kanamoto et al., 2000) were also cloned in pcDNA3. Motoneuron Purification and Culture Motoneuron cultures were prepared from E12.5 mice spinal cords essentially as described previously (Arce et al., 1999; Henderson et al., 1995). Motoneurons were plated in the presence of a cocktail of neurotrophic factors (referred as “NTFs”: 1 ng/ml BDNF, 100 pg/ ml GDNF, 10 ng/ml CNTF) added at the time of cell seeding. After 16 hr in culture, they were treated by addition of the different reagents diluted in Neurobasal medium. Unless otherwise indicated, motoneurons were then incubated at 37⬚C for 48 hr, and living cells were counted under phase-contrast. Since the fraction of motoneurons triggered to die was never greater than 50%, all survival graphs were plotted with 40% of optimal survival as the lowest value shown. To compare the yield of motoneurons from transgenic SOD1 and control mice, motoneurons were purified and the number of cells

Neuron 1080

in suspension counted. In both mutants and controls, the yield of purified motoneurons per spinal cord was equivalent. After seeding, motoneuron survival in the absence of Fas activation remained constant between 1 and 3 DIV. In these conditions, no significant survival differences at 3 DIV were found in a retrospective analysis of 36 parallel motoneuron cultures from controls (set to 100%) and mutants: G93A (107% ⫾ 7%, mean ⫾ SEM, n ⫽ 16), G37R (93% ⫾ 11%, n ⫽ 4), G85R (97% ⫾ 17%, n ⫽ 9), and wt SOD1 (109% ⫾ 7%, n ⫽ 7). Thus, the purification and culture procedures employed did not generate bias by selecting quantitatively different subpopulations of mutant and control motoneurons. For excitotoxicity experiments, motoneurons were cultured for 5 DIV in the continued presence of NTFs, medium was changed, and cells were treated with domoic acid and/or other indicated reagents (Cisterni et al., 2001). Cell viability was quantified at 7 DIV. For NOinduced death, motoneurons were cultured for 16 hr in the presence of NTFs alone and then treated with different concentrations of DETANONOate.

Statistical Methods The number of surviving or immunoreactive motoneurons in each condition was expressed as a percentage of the number of motoneurons surviving in the presence of trophic factors alone; triplicate or quadruplicate dishes were used for each condition. If not otherwise indicated, differences between treatments or genotypes were analyzed for their statistical significance by student’s t test, twotailed, unpaired.

Electroporation of Motoneurons Cells dissociated from E12.5 mice ventral spinal cord were centrifuged over a 5% (v/v) Optiprep cushion at 2000 rpm for 15 min. Cells at the interface were collected and washed on a BSA cushion at 1500 rpm for 5 min. Pellets were resuspended in electroporation buffer (125 mM NaCl, 5 mM KCl, 1.5 mM MgCl2, 10 mM glucose, 20 mM HEPES [pH 7.4]) at a density of 50,000 cells per 100 ␮l. Five micrograms of pEGFP-N1 (Clontech Laboratories, Inc.) and the same molar amount of the vector of interest were added to the 100 ␮l cell suspension. Motoneurons were incubated at room temperature for 15 min, transferred to a 4 mm gap cuvette (Eppendorf), and electroporated using Electro square porator娃 BTX-ECM830, Genetronics Inc. Electroporation conditions were three pulses of 5 ms at 200V with intervals of 1 s. Culture medium was added rapidly to dilute the cells after electroporation, and cells were plated in fourwell plates.

Immunocytochemistry Motoneurons seeded on glass coverslips were treated or not as indicated and fixed in 3.7% (v/v) formaldehyde, washed in PBS containing 50 mM L-Lysine, and blocked for 1 hr with 5% (v/v) donkey serum, 4% BSA, 0.1% (v/v) Triton X-100 in PBS. Cells were incubated with the primary antibodies (rabbit polyclonal anti-nNOS diluted 200-fold, polyclonal anti-HA 500-fold, anti-SOD1 (SD-G6) 600-fold, anti-NF-M (Ab1987) 500-fold, anti-activated Caspase-3 (CM1) 5000-fold) in the blocking solution, followed by fluorochromeconjugated secondary antibodies. Immunofluorescence staining of nitrotyrosine and analysis of fluorescence intensity was done as described (Estevez et al., 1998a) on 150–300 randomly chosen cells per culture condition.

Monitoring of Cytochrome c Release Motoneurons grown for 16 hr on glass coverslips were treated with the indicated reagents. After 30 hr of treatment, cells were stained using the mouse monoclonal anti-cytochrome c (6H2.B4) at 1 ␮g/ ml as the primary antibody and streptavidin-conjugated Cy3 antimouse as a secondary antibody. The diffuse labeling of cytochrome c was estimated under a fluorescence microscope and confirmed by confocal analysis. The percentage of cells presenting a diffuse labeling over the total number of labeled cells was estimated on two diameters of the well, representing about 50 different fields per culture condition.

Other Cell Cultures and Apoptosis Assays Swiss 3T3 cells were maintained in DMEM medium (Gibco BRL) supplemented with 10% fetal-calf serum and 1% penicillin-streptomycin. Cells were plated on 24-well plates and cultured to near confluency prior to treatment with IETD-fmk, L-NAME, or MnTBAP, which were added to the culture medium 2 hr prior to addition of agonistic anti-Fas antibody (JO2) at 100 ng/ml. Cell survival was measured using the MTT assay (Manthorpe et al., 1986). Thymocytes were prepared from P5 mouse thymuses as described (Fisher et al., 1996), with minor modifications. Cells were cultured for 3 hr and treated with the indicated reagents. Cell viability was estimated 24 hr later using Live/Dead Viability/Cytotoxicity Kit (Molecular Probes). The percentage of viable thymocytes was calculated as follows: [number calcein AM positive cells/(number EthD-1 ⫹ calcein AM positive cells)] ⫻ 100. Primary cortical neurons were cultured from E15 mouse embryos. Cerebral cortices were dissociated in HBSS, 0.4% BSA, 100 ␮g of DNase after digestion in 0.125% Trypsin/HBSS. Cells were then centrifuged on a BSA cushion at 1500 rpm for 5 min and plated on polyornithine/laminine-treated wells at the density of 150,000 cells/ cm2. Cortical neuron cultures were maintained for 4 days in DMEM with 2% B27 supplement (Life Technologies) and 1 mM sodium pyruvate at 37⬚C, 5% CO2. Medium was changed and cells were treated with indicated reagents for 24 hr. Neurons were fixed and nuclei were stained with DAPI. Cell death was determined by counting cells with condensed or fragmented nuclei in four to seven fields (400–600 cells) per well in a blinded manner. Data are expressed as the percentage of apoptotic cells among the total cells counted. For NO-induced death of cortical neurons, cells isolated from indicated strain were cultured for 24 hr and treated for 48 hr with increasing concentrations of DETANONOate. Cell survival was measured using MTT assay. Dorsal root ganglion neuronal cultures (DRG) were obtained from E16-17 mouse embryos following the procedure described (Varon et al., 1981). Agonistic anti-Fas antibodies were added at their final concentrations immediately after seeding. Cell survival was estimated after 48 hr, by counting MTT-positive neurons. Astrocytes were obtained from E16 mouse embryos, as described (Pettmann et al., 1991) with slight modifications. Cells were treated 24 hr after subculture with agonistic anti-Fas antibody and their survival estimated after 48 hr using the MTT assay. RT-PCR Analysis of NOS Expression For each condition, about 60–80,000 purified motoneurons were seeded in two 35 mm diameter dishes. For Swiss 3T3 fibroblasts, cells were grown to near confluency in 6 cm diameter dishes. At the indicated time and treatment, total RNA was extracted using Trizol reagent (Life Technologies) according to the manufacturer’s instructions. The retrotranscription step was performed using Superscript II (Gibco BRL) reverse transcriptase and random hexamer primers. Normalization of cDNA levels was done by PCR on a PerkinElmer Thermal Cycler using as reference primers for ␤-actin (nucleotides 304 to 327, 5⬘-TTGTAACCAACTGGGACGATATGG and nucleotides 1044 to 1067, 5⬘-GATCTTGATCTTCATGGTGCTAGG (GenBank accession number GI49865). PCR primers for mouse nNOS were nucleotides 1766 to 1786, 5⬘-CCTGGGGCTCAAATGGTATGG, and nucleotides 2633 to 2653, 5⬘-GAATAGGAGGAGACGCTGTTG (GenBank accession number D14552). Primers for mouse iNOS were nucleotides 2201 to 2221, 5⬘-CAACAGGAGAAGGGGACGAAC, and nucleotides 2791 to 2811, 5⬘-TCTCTGCCTATCCGTCTCGTC (GenBank accession number M87039). Primers for eNOS were nucleotides 1549 to 1569, 5⬘-GAAGCGTGTGAAGGCAACCAT, and nucleotides 2319 to 2335, 5⬘-TTGTGGCTCGGGTGGATTTGC (GenBank accession number U53142). Internal oligonucleotide probes used for Southern blots were 5⬘-TGACCATCGTTGACCACCACT for nNOS, 5⬘-TGAAAAGTCCAGCCGCACCAC for iNOS, 5⬘-AGAAGTGG GGGTATGCTCGGG for eNOS. Membranes were washed in decreasing concentrations of SSC and then exposed for autoradiography. Western Blot Analyses Purified motoneurons were seeded at a density of 150,000 cells per 6 cm dish. Motoneurons were cultured in the presence of NTFs for 16 hr. They were treated or not with either 100 ng/ml of sFasL ⫹

Motoneuron-Restricted Death Pathway 1081

Enhancer (1 ␮g/ml) for 120 min (p38/JNKs) or 100 ng/ml of agonistic anti-Fas antibody for 18 hr in the presence or in the absence of 5 ␮M SB203580 or 100 ␮M L-NAME. Cell extract proteins were separated on 15% SDS-PAGE and Western blotting performed using antibodies against active p38, phospho-JNKs, p38, JNKs, caspase8 p18 (SK440) (Matsushita et al., 2000), or caspase-8 (Srinivasan et al., 1998a) diluted 1000-fold. For subsequent immunodetection of NF-M or ␣-tubulin, membranes were first stripped of bound antibodies. The relative intensity of phosphorylated p38 and JNKs was determined using an AGFA densitometer, and data were analyzed using the NIH image 1.62 software. For nNOS protein detection, thymocytes and cortical neurons were seeded at a density of 150,000 cells per cm2 in 6 cm Petri dishes, treated or not with agonistic anti-Fas antibody, and lysed at the indicated time. Sample proteins were resolved by 6% SDS-PAGE and submitted to Western blotting using anti-NOS following manufacturer’s instructions. Acknowledgments We are indebted to T. Williamson (Trophos, Marseille), F. Maina, and S. Alonso (INSERM, Marseille) for critical comments on the manuscript; and to members of INSERM U.382 for helpful comments on the manuscript and throughout this work. We are particularly grateful to S. Corby for help in animal care and genotyping; AnneOdile Hueber and Chantal Bazenet for plasmids coding for Daxx and Ask1; K. Tomaselli/A. Srinivasan of Idun Pharmaceuticals, Inc. (La Jolla, CA) for the generous gift of antibodies to activated caspase-3 (CM1) and caspase-8; F. Barone for anti-active-caspase-8 antibodies. We thank G. Ugolini for allowing us to discuss unpublished results. This work was funded by Institut National de la Sante´ et de la Recherche Me´dicale (INSERM), Centre National de la Recherche Scientifique (CNRS), Association Franc¸aise contre les Myopathies (AFM), European Commission contract QLG3-CT-199900602, Amyotrophic Lateral Sclerosis Association (ALSA), and GIPAventis. C.R. was supported by fellowships from the Ministe`re de l’Education Nationale, la Recherche et la Technologie (MENRT), and from the Association pour la Recherche contre le Cancer (ARC). Received: January 24, 2002 Revised: August 21, 2002 References Adachi, M., Suematsu, S., Kondo, T., Ogasawara, J., Tanaka, T., Yoshida, N., and Nagata, S. (1995). Targeted mutation in the Fas gene causes hyperplasia in peripheral lymphoid organs and liver. Nat. Genet. 11, 294–300.

Bonny, C., Oberson, A., Negri, S., Sauser, C., and Schorderet, D.F. (2001). Cell-permeable peptide inhibitors of JNK: novel blockers of beta-cell death. Diabetes 50, 77–82. Borasio, G.D., Horstmann, S., Anneser, J.M., Neff, N.T., and Glicksman, M.A. (1998). CEP-1347/KT7515, a JNK pathway inhibitor, supports the in vitro survival of chick embryonic neurons. Neuroreport 9, 1435–1439. Borchelt, D.R., Lee, M.K., Slunt, H.S., Guarnieri, M., Xu, Z.S., Wong, P.C., Brown, R.H., Jr., Price, D.L., Sisodia, S.S., and Cleveland, D.W. (1994). Superoxide dismutase 1 with mutations linked to familial amyotrophic lateral sclerosis possesses significant activity. Proc. Natl. Acad. Sci. USA 91, 8292–8296. Bruijn, L.I., Becher, M.W., Lee, M.K., Anderson, K.L., Jenkins, N.A., Copeland, N.G., Sisodia, S.S., Rothstein, J.D., Borchelt, D.R., Price, D.L., and Cleveland, D.W. (1997). ALS-linked SOD1 mutant G85R mediates damage to astrocytes and promotes rapidly progressive disease with SOD1-containing inclusions. Neuron 18, 327–338. Catania, M.V., Aronica, E., Yankaya, B., and Troost, D. (2001). Increased expression of neuronal nitric oxide synthase spliced variants in reactive astrocytes of amyotrophic lateral sclerosis human spinal cord. J. Neurosci. 21, RC148. Chang, H.Y., Nishitoh, H., Yang, X., Ichijo, H., and Baltimore, D. (1998). Activation of apoptosis signal-regulating kinase 1 (ASK1) by the adapter protein Daxx. Science 281, 1860–1863. Cheema, Z.F., Wade, S.B., Sata, M., Walsh, K., Sohrabji, F., and Miranda, R.C. (1999). Fas/Apo [apoptosis]-1 and associated proteins in the differentiating cerebral cortex: induction of caspase-dependent cell death and activation of NF-kappaB. J. Neurosci. 19, 1754– 1770. Ciesielski-Treska, J., Ulrich, G., Chasserot-Golaz, S., Zwiller, J., Revel, M.O., Aunis, D., and Bader, M.F. (2001). Mechanisms underlying neuronal death induced by chromogranin A-activated microglia. J. Biol. Chem. 276, 13113–13120. Ciriolo, M.R., De Martino, A., Lafavia, E., Rossi, L., Carri, M.T., and Rotilio, G. (2000). Cu,Zn-superoxide dismutase-dependent apoptosis induced by nitric oxide in neuronal cells. J. Biol. Chem. 275, 5065–5072. Cisterni, C., Kallenbach, S., Jordier, F., Bagnis, C., and Pettmann, B. (2001). Death of motoneurons induced by trophic deprivation or by excitotoxicity is not prevented by overexpression of SMN. Neurobiol. Dis. 8, 240–251. Cleveland, D.W., and Rothstein, J.D. (2001). From Charcot to Lou Gehrig: deciphering selective motor neuron death in ALS. Nat. Rev. Neurosci. 2, 806–819.

Almer, G., Vukosavic, S., Romero, N., and Przedborski, S. (1999). Inducible nitric oxide synthase up-regulation in a transgenic mouse model of familial amyotrophic lateral sclerosis. J. Neurochem. 72, 2415–2425.

Da Silva, J., Pierrat, B., Mary, J.L., and Lesslauer, W. (1997). Blockade of p38 mitogen-activated protein kinase pathway inhibits inducible nitric-oxide synthase expression in mouse astrocytes. J. Biol. Chem. 272, 28373–28380.

Anneser, J.M., Cookson, M.R., Ince, P.G., Shaw, P.J., and Borasio, G.D. (2001). Glial cells of the spinal cord and subcortical white matter up-regulate neuronal nitric oxide synthase in sporadic amyotrophic lateral sclerosis. Exp. Neurol. 171, 418–421.

Deckwerth, T.L., Elliott, J.L., Knudson, C.M., Johnson, E.M., Jr., Snider, W.D., and Korsmeyer, S.J. (1996). BAX is required for neuronal death after trophic factor deprivation and during development. Neuron 17, 401–411.

Arce, V., Garces, A., de Bovis, B., Filippi, P., Henderson, C., Pettmann, B., and deLapeyriere, O. (1999). Cardiotrophin-1 requires LIFRbeta to promote survival of mouse motoneurons purified by a novel technique. J. Neurosci. Res. 55, 119–126.

Durham, H.D., Roy, J., Dong, L., and Figlewicz, D.A. (1997). Aggregation of mutant Cu/Zn superoxide dismutase proteins in a culture model of ALS. J. Neuropathol. Exp. Neurol. 56, 523–530.

Babu, B.R., and Griffith, O.W. (1998). Design of isoform-selective inhibitors of nitric oxide synthase. Curr. Opin. Chem. Biol. 2, 491–500. Barker, V., Middleton, G., Davey, F., and Davies, A.M. (2001). TNFalpha contributes to the death of NGF-dependent neurons during development. Nat. Neurosci. 4, 1194–1198. Beckman, J.S., Ye, Y.Z., Anderson, P., Chen, J., Accavetti, M.A., Tarpey, M.M., and White, C.R. (1994). Extensive nitration of protein tyrosines in human atherosclerosis detected by immunohistochemistry. Biol. Chem. Hoppe Seyler 375, 81–88. Beckman, J.S., Estevez, A.G., Crow, J.P., and Barbeito, L. (2001). Superoxide dismutases and the death of motoneurons in ALS. Trends Neurosci. 24, S15–20.

Estevez, A.G., Spear, N., Manuel, S.M., Barbeito, L., Radi, R., and Beckman, J.S. (1998a). Role of endogenous nitric oxide and peroxynitrite formation in the survival and death of motor neurons in culture. Prog. Brain Res. 118, 269–280. Estevez, A.G., Spear, N., Manuel, S.M., Radi, R., Henderson, C.E., Barbeito, L., and Beckman, J.S. (1998b). Nitric oxide and superoxide contribute to motor neuron apoptosis induced by trophic factor deprivation. J. Neurosci. 18, 923–931. Estevez, A.G., Crow, J.P., Sampson, J.B., Reiter, C., Zhuang, Y., Richardson, G.J., Tarpey, M.M., Barbeito, L., and Beckman, J.S. (1999). Induction of nitric oxide-dependent apoptosis in motor neurons by zinc-deficient superoxide dismutase. Science 286, 2498– 2500. Facchinetti, F., Sasaki, M., Cutting, F.B., Zhai, P., MacDonald, J.E.,

Neuron 1082

Reif, D., Beal, M.F., Huang, P.L., Dawson, T.M., Gurney, M.E., and Dawson, V.L. (1999). Lack of involvement of neuronal nitric oxide synthase in the pathogenesis of a transgenic mouse model of familial amyotrophic lateral sclerosis. Neuroscience 90, 1483–1492. Faulkner, K.M., Liochev, S.I., and Fridovich, I. (1994). Stable Mn(III) porphyrins mimic superoxide dismutase in vitro and substitute for it in vivo. J. Biol. Chem. 269, 23471–23476. Fisher, G.H., Lenardo, M.J., and Zuniga-Pflucker, J.C. (1996). Synergy between T cell receptor and Fas (CD95/APO-1) signaling in mouse thymocyte death. Cell. Immunol. 169, 99–106. Frade, J.M., and Barde, Y.A. (1998). Microglia-derived nerve growth factor causes cell death in the developing retina. Neuron 20, 35–41. Frey, D., Schneider, C., Xu, L., Borg, J., Spooren, W., and Caroni, P. (2000). Early and selective loss of neuromuscular synapse subtypes with low sprouting competence in motoneuron diseases. J. Neurosci. 20, 2534–2542. Ghadge, G.D., Lee, J.P., Bindokas, V.P., Jordan, J., Ma, L., Miller, R.J., and Roos, R.P. (1997). Mutant superoxide dismutase-1-linked familial amyotrophic lateral sclerosis: Molecular mechanisms of neuronal death and protection. J. Neurosci. 17, 8756–8766. Gong, Y.H., Parsadanian, A.S., Andreeva, A., Snider, W.D., and Elliott, J.L. (2000). Restricted expression of G86R Cu/Zn superoxide dismutase in astrocytes results in astrocytosis but does not cause motoneuron degeneration. J. Neurosci. 20, 660–665. Green, D., and Kroemer, G. (1998). The central executioners of apoptosis: caspases or mitochondria? Trends Cell Biol. 8, 267–271. Gurney, M.E., Pu, H., Chiu, A.Y., Dal Canto, M.C., Polchow, C.Y., Alexander, D.D., Caliendo, J., Hentati, A., Kwon, Y.W., Deng, H.X., et al. (1994). Motor neuron degeneration in mice that express a human Cu,Zn superoxide dismutase mutation. Science 264, 1772– 1775. Handy, R.L., Wallace, P., Gaffen, Z.A., Whitehead, K.J., and Moore, P.K. (1995). The antinociceptive effect of 1-(2-trifluoromethylphenyl) imidazole (TRIM), a potent inhibitor of neuronal nitric oxide synthase in vitro, in the mouse. Br. J. Pharmacol. 116, 2349–2350. Harper, S.J., and LoGrasso, P. (2001). Signalling for survival and death in neurones: the role of stress-activated kinases, JNK and p38. Cell. Signal. 13, 299–310. Henderson, C.E., Bloch-Gallego, E., and Camu, W. (1995). Purified embryonic motoneurons. In Nerve Cell Culture: A Practical Approach, J. Cohen and G. Wilkin, eds. (London: Oxford University Press), pp. 69–81. Herdegen, T., Claret, F.X., Kallunki, T., Martin-Villalba, A., Winter, C., Hunter, T., and Karin, M. (1998). Lasting N-terminal phosphorylation of c-Jun and activation of c-Jun N- terminal kinases after neuronal injury. J. Neurosci. 18, 5124–5135. Julien, J.P. (2001). Amyotrophic lateral sclerosis: unfolding the toxicity of the misfolded. Cell 104, 581–591. Kanamoto, T., Mota, M., Takeda, K., Rubin, L.L., Miyazono, K., Ichijo, H., and Bazenet, C.E. (2000). Role of apoptosis signal-regulating kinase in regulation of the c-Jun N-terminal kinase pathway and apoptosis in sympathetic neurons. Mol. Cell. Biol. 20, 196–204. Kennel, P.F., Finiels, F., Revah, F., and Mallet, J. (1996). Neuromuscular function impairment is not caused by motor neurone loss in FALS mice: an electromyographic study. Neuroreport 7, 1427–1431. Kischkel, F.C., Hellbardt, S., Behrmann, I., Germer, M., Pawlita, M., Krammer, P.H., and Peter, M.E. (1995). Cytotoxicity-dependent APO-1 (Fas/CD95)-associated proteins form a death-inducing signaling complex (DISC) with the receptor. EMBO J. 14, 5579–5588. Knudson, C.M., Tung, K.S., Tourtellotte, W.G., Brown, G.A., and Korsmeyer, S.J. (1995). Bax-deficient mice with lymphoid hyperplasia and male germ cell death. Science 270, 96–99. Kostic, V., Jackson-Lewis, V., de Bilbao, F., Dubois-Dauphin, M., and Przedborski, S. (1997). Bcl-2: prolonging life in a transgenic mouse model of familial amyotrophic lateral sclerosis. Science 277, 559–562. Krammer, P.H. (2000). CD95⬘s deadly mission in the immune system. Nature 407, 789–795. Le-Niculescu, H., Bonfoco, E., Kasuya, Y., Claret, F.X., Green, D.R.,

and Karin, M. (1999). Withdrawal of survival factors results in activation of the JNK pathway in neuronal cells leading to Fas ligand induction and cell death. Mol. Cell. Biol. 19, 751–763. Li, H., Zhu, H., Xu, C.J., and Yuan, J. (1998). Cleavage of BID by caspase 8 mediates the mitochondrial damage in the Fas pathway of apoptosis. Cell 94, 491–501. Li, M., Ona, V.O., Guegan, C., Chen, M., Jackson-Lewis, V., Andrews, L.J., Olszewski, A.J., Stieg, P.E., Lee, J.P., Przedborski, S., and Friedlander, R.M. (2000). Functional role of caspase-1 and caspase-3 in an ALS transgenic mouse model. Science 288, 335–339. Lino, M.M., Schneider, C., and Caroni, P. (2002). Accumulation of SOD1 mutants in postnatal motoneurons does not cause motoneuron pathology or motoneuron disease. J. Neurosci. 22, 4825–4832. Lowry, K.S., Murray, S.S., McLean, C.A., Talman, P., Mathers, S., Lopes, E.C., and Cheema, S.S. (2001). A potential role for the p75 low-affinity neurotrophin receptor in spinal motor neuron degeneration in murine and human amyotrophic lateral sclerosis. Amyotroph. Lateral. Scler. Other Motor Neuron Disord. 2, 127–134. Luo, X., Budihardjo, I., Zou, H., Slaughter, C., and Wang, X. (1998). Bid, a Bcl2 interacting protein, mediates cytochrome c release from mitochondria in response to activation of cell surface death receptors. Cell 94, 481–490. Manthorpe, M., Fagnani, R., Skaper, S.D., and Varon, S. (1986). An automated colorimetric microassay for neuronotrophic factors. Brain Res. 390, 191–198. Maroney, A.C., Glicksman, M.A., Basma, A.N., Walton, K.M., Knight, E., Jr., Murphy, C.A., Bartlett, B.A., Finn, J.P., Angeles, T., Matsuda, Y., et al. (1998). Motoneuron apoptosis is blocked by CEP-1347 (KT 7515), a novel inhibitor of the JNK signaling pathway. J. Neurosci. 18, 104–111. Martin-Villalba, A., Herr, I., Jeremias, I., Hahne, M., Brandt, R., Vogel, J., Schenkel, J., Herdegen, T., and Debatin, K.M. (1999). CD95 ligand (Fas-L/APO-1L) and tumor necrosis factor-related apoptosis- inducing ligand mediate ischemia-induced apoptosis in neurons. J. Neurosci. 19, 3809–3817. Matsushita, K., Wu, Y., Qiu, J., Lang-Lazdunski, L., Hirt, L., Waeber, C., Hyman, B.T., Yuan, J., and Moskowitz, M.A. (2000). Fas receptor and neuronal cell death after spinal cord ischemia. J. Neurosci. 20, 6879–6887. Matsuyama, T., Hata, R., Yamamoto, Y., Tagaya, M., Akita, H., Uno, H., Wanaka, A., Furuyama, J., and Sugita, M. (1995). Localization of Fas antigen mRNA induced in postischemic murine forebrain by in situ hybridization. Brain Res. Mol. Brain Res. 34, 166–172. Michaelson, J.S., Bader, D., Kuo, F., Kozak, C., and Leder, P. (1999). Loss of Daxx, a promiscuously interacting protein, results in extensive apoptosis in early mouse development. Genes Dev. 13, 1918– 1923. Mielke, K., and Herdegen, T. (2000). JNK and p38 stress kinases– degenerative effectors of signal-transduction-cascades in the nervous system. Prog. Neurobiol. 61, 45–60. Migheli, A., Cavalla, P., Marino, S., and Schiffer, D. (1994). A study of apoptosis in normal and pathologic nervous tissue after in situ end-labeling of DNA strand breaks. J. Neuropathol. Exp. Neurol. 53, 606–616. Migheli, A., Atzori, C., Piva, R., Tortarolo, M., Girelli, M., Schiffer, D., and Bendotti, C. (1999). Lack of apoptosis in mice with ALS. Nat. Med. 5, 966–967. Morishima, Y., Gotoh, Y., Zieg, J., Barrett, T., Takano, H., Flavell, R., Davis, R.J., Shirasaki, Y., and Greenberg, M.E. (2001). Betaamyloid induces neuronal apoptosis via a mechanism that involves the c-Jun N-terminal kinase pathway and the induction of Fas ligand. J. Neurosci. 21, 7551–7560. Nagata, S. (1997). Apoptosis by death factor. Cell 88, 355–365. Northington, F.J., Ferriero, D.M., Flock, D.L., and Martin, L.J. (2001). Delayed neurodegeneration in neonatal rat thalamus after hypoxiaischemia is apoptosis. J. Neurosci. 21, 1931–1938. Ogasawara, J., Suda, T., and Nagata, S. (1995). Selective apoptosis of CD4⫹CD8⫹ thymocytes by the anti-Fas antibody. J. Exp. Med. 181, 485–491.

Motoneuron-Restricted Death Pathway 1083

Ono, K., and Han, J. (2000). The p38 signal transduction pathway: activation and function. Cell. Signal. 12, 1–13. Pasinelli, P., Houseweart, M.K., Brown, R.H., Jr., and Cleveland, D.W. (2000). Caspase-1 and -3 are sequentially activated in motor neuron death in Cu,Zn superoxide dismutase-mediated familial amyotrophic lateral sclerosis. Proc. Natl. Acad. Sci. USA 97, 13901– 13906. Paul, A., Wilson, S., Belham, C.M., Robinson, C.J., Scott, P.H., Gould, G.W., and Plevin, R. (1997). Stress-activated protein kinases: activation, regulation and function. Cell. Signal. 9, 403–410. Pettmann, B., and Henderson, C.E. (1998). Neuronal cell death. Neuron 20, 633–647. Pettmann, B., Janet, T., Labourdette, G., Sensenbrenner, M., Manthorpe, M., and Varon, S. (1991). Biologically active basic fibroblast growth factor migrates at 27 kD in “non-denaturing” SDS-polyacrylamide gel electrophoresis. Growth Factors 5, 209–220. Pramatarova, A., Laganiere, J., Roussel, J., Brisebois, K., and Rouleau, G.A. (2001). Neuron-specific expression of mutant superoxide dismutase 1 in transgenic mice does not lead to motor impairment. J. Neurosci. 21, 3369–3374. Rabizadeh, S., Gralla, E.B., Borchelt, D.R., Gwinn, R., Valentine, J.S., Sisodia, S., Wong, P., Lee, M., Hahn, H., and Bredesen, D.E. (1995). Mutations associated with amyotrophic lateral sclerosis convert superoxide dismutase from an antiapoptotic gene to a proapoptotic gene: Studies in yeast and neural cells. Proc. Natl. Acad. Sci. USA 92, 3024–3028. Raoul, C., Henderson, C.E., and Pettmann, B. (1999). Programmed cell death of embryonic motoneurons triggered through the Fas death receptor. J. Cell Biol. 147, 1049–1062. Raoul, C., Pettmann, B., and Henderson, C.E. (2000). Active killing of neurons during development and following stress: a role for p75(NTR) and Fas? Curr. Opin. Neurobiol. 10, 111–117. Reiter, C.D., Teng, R.J., and Beckman, J.S. (2000). Superoxide reacts with nitric oxide to nitrate tyrosine at physiological pH via peroxynitrite. J. Biol. Chem. 275, 32460–32466. Robertson, J., Beaulieu, J.M., Doroudchi, M.M., Durham, H.D., Julien, J.P., and Mushynski, W.E. (2001). Apoptotic death of neurons exhibiting peripherin aggregates is mediated by the proinflammatory cytokine tumor necrosis factor-alpha. J. Cell Biol. 155, 217–226. Rothstein, J.D. (1995). Excitotoxic mechanisms in the pathogenesis of amyotrophic lateral sclerosis. Adv. Neurol. 68, 7–20. Saitoh, M., Nishitoh, H., Fujii, M., Takeda, K., Tobiume, K., Sawada, Y., Kawabata, M., Miyazono, K., and Ichijo, H. (1998). Mammalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK) 1. EMBO J. 17, 2596–2606. Sasaki, M., Gonzalez-Zulueta, M., Huang, H., Herring, W.J., Ahn, S., Ginty, D.D., Dawson, V.L., and Dawson, T.M. (2000). Dynamic regulation of neuronal NO synthase transcription by calcium influx through a CREB family transcription factor-dependent mechanism. Proc. Natl. Acad. Sci. USA 97, 8617–8622. Scaffidi, C., Fulda, S., Srinivasan, A., Friesen, C., Li, F., Tomaselli, K.J., Debatin, K.M., Krammer, P.H., and Peter, M.E. (1998). Two CD95 (APO-1/Fas) signaling pathways. EMBO J. 17, 1675–1687. Scaffidi, C., Schmitz, I., Zha, J., Korsmeyer, S.J., Krammer, P.H., and Peter, M.E. (1999). Differential modulation of apoptosis sensitivity in CD95 type I and type II cells. J. Biol. Chem. 274, 22532–22538. Shaw, P.J., and Eggett, C.J. (2000). Molecular factors underlying selective vulnerability of motor neurons to neurodegeneration in amyotrophic lateral sclerosis. J. Neurol. Suppl. 247, I17–27. Srinivasan, A., Li, F., Wong, A., Kodandapani, L., Smidt, R., Jr., Krebs, J.F., Fritz, L.C., Wu, J.C., and Tomaselli, K.J. (1998a). BclxL functions downstream of caspase-8 to inhibit Fas- and tumor necrosis factor receptor 1-induced apoptosis of MCF7 breast carcinoma cells. J. Biol. Chem. 273, 4523–4529. Srinivasan, A., Roth, K.A., Sayers, R.O., Shindler, K.S., Wong, A.M., Fritz, L.C., and Tomaselli, K.J. (1998b). In situ immunodetection of activated caspase-3 in apoptotic neurons in the developing nervous system. Cell Death Differ. 5, 1004–1016. Tan, Y., Rouse, J., Zhang, A., Cariati, S., Cohen, P., and Comb, M.J.

(1996). FGF and stress regulate CREB and ATF-1 via a pathway involving p38 MAP kinase and MAPKAP kinase-2. EMBO J. 15, 4629–4642. Terrado, J., Monnier, D., Perrelet, D., Vesin, D., Jemelin, S., Buurman, W.A., Mattenberger, L., King, B., Kato, A.C., and Garcia, I. (2000). Soluble TNF receptors partially protect injured motoneurons in the postnatal CNS. Eur. J. Neurosci. 12, 3443–3447. Tobiume, K., Matsuzawa, A., Takahashi, T., Nishitoh, H., Morita, K., Takeda, K., Minowa, O., Miyazono, K., Noda, T., and Ichijo, H. (2001). ASK1 is required for sustained activations of JNK/p38 MAP kinases and apoptosis. EMBO Rep. 2, 222–228. Upton-Rice, M.N., Cudkowicz, M.E., Mathew, R.K., Reif, D., and Brown, R.H., Jr. (1999). Administration of nitric oxide synthase inhibitors does not alter disease course of amyotrophic lateral sclerosis SOD1 mutant transgenic mice. Ann. Neurol. 45, 413–414. Varon, S., Skaper, S.D., and Manthorpe, M. (1981). Trophic activities for dorsal root and sympathetic ganglionic neurons in media conditioned by Schwann and other peripheral cells. Brain Res. 227, 73–87. Wong, P.C., Pardo, C.A., Borchelt, D.R., Lee, M.K., Copeland, N.G., Jenkins, N.A., Sisodia, S.S., Cleveland, D.W., and Price, D.L. (1995). An adverse property of a familial ALS-linked SOD1 mutation causes motor neuron disease characterized by vacuolar degeneration of mitochondria. Neuron 14, 1105–1116. Yang, X., Khosravi-Far, R., Chang, H.Y., and Baltimore, D. (1997). Daxx, a novel Fas-binding protein that activates JNK and apoptosis. Cell 89, 1067–1076. Yi, F.H., Lautrette, C., Vermot-Desroches, C., Bordessoule, D., Couratier, P., Wijdenes, J., Preud’homme, J.L., and Jauberteau, M.O. (2000). In vitro induction of neuronal apoptosis by anti-Fas antibodycontaining sera from amyotrophic lateral sclerosis patients. J. Neuroimmunol. 109, 211–220. Zhang, H.Q., Fast, W., Marletta, M.A., Martasek, P., and Silverman, R.B. (1997). Potent and selective inhibition of neuronal nitric oxide synthase by N omega-propyl-L-arginine. J. Med. Chem. 40, 3869– 3870.