SEARCH

SEARCH BY CITATION

Keywords:

  • endoplasmic reticulum stress;
  • motor neuron;
  • superoxide dismutase;
  • unfolded protein response

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Immuncytochemistry
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Recent studies indicate that endoplasmic reticulum (ER) stress is involved in the pathogenesis of familial and sporadic amyotrophic lateral sclerosis (ALS). ER stress occurs when the ER–mitochondria calcium cycle (ERMCC) is disturbed and misfolded proteins accumulate in the ER. To cope with ER stress, the cell engages the unfolded protein response (UPR). While activation of the UPR has been shown in some ALS models and tissues, ER stress elements have not been studied directly in motor neurons. Here we investigated the expression of XBP1 and ATF6α and phosphorylation of eIF2α, and their modulation, in mutated SOD1G93A NSC34 and animal model of ALS. Expression of XBP1 and ATF6α mRNA and protein was enhanced in SOD1G93A NSC34 cells. Activation of ATF6α and XBP1 and phosphorylation of eIF2α were detectable in mutated SOD1G93A motor but not in wild-type motor neurons. Treatment with the ER stressor thapsigargin enhanced phosphorylation of eIF2α and activated proteolysis of ATF6α and splicing of XBP1 in NSC34 and motor neurons in a time-dependent manner. The present study thus provides direct evidence of activated UPR in motor neurons which overexpress human pathogenic mutant SOD1G93A, providing evidence that ER stress plays a major role in ALS.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Immuncytochemistry
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Amyotrophic lateral sclerosis (ALS) is the most common adult-onset motor neuron disease and is characterized by signs of upper (spasticity, dysphagia, dysarthria) and lower (atrophy, fasciculations) motor neuron degeneration. Approximately 90% of ALS cases are sporadic ALS; nearly 10% have a positive family history (familial ALS). Mutant Cu/Zn superoxide dismutase type I (SOD1) is the cause of 20% of familial ALS cases. Compelling evidence suggests that mutant SOD1 causes ALS through a toxic gain of function (Rothstein, 2009). Mice and rats that overexpress the human mutant SOD1 are used extensively to study ALS, because they develop adult-onset progressive muscle weakness and atrophy, caused by prominent motor neuron degeneration (Gurney et al., 1994).

Although the etiology of ALS remains unclear, recent studies indicate that endoplasmic reticulum (ER) stress is involved in the pathogenesis of familial and sporadic ALS (Ilieva et al., 2007; Atkin et al., 2008; Walker, 2010). ER stress occurs when ER calcium content is depleted (Verkhratsky, 2005) and misfolded proteins accumulate in the ER. This process is tightly linked to mitochondrial function in the ER–mitochondria calcium cycle (ERMCC), which is thought to be compromised in ALS (Berridge, 2002; Grosskreutz et al., 2010). To cope with ER stress, cells activate the unfolded protein response (UPR), which mediates an upregulation of genes encoding ER-resident chaperones such as glucose-regulated protein 78 (Grp78) and a down-regulation of general protein synthesis (Kozutsumi et al., 1988; Yoshida et al., 2001; Dudek et al., 2009). Prolonged UPR leads to activation of ER-resident caspase-12, triggering apoptosis (Nakagawa et al., 2000). The three main ER stress sensors are the double-stranded RNA-activated protein kinase-like ER kinase (PERK), the basic leucine-zipper transcription factor 6 (ATF6), and the inositol-requiring enzyme 1 (IRE1).

Under ER stress conditions several pathways are activated: first, the ER-transmembrane protein ATF6α translocates to the Golgi apparatus, where it is cleaved by proteases, and the resulting N-terminal fragment of ATF6α (p50-ATF6α) translocates to the nucleus. Second, IRE1 induces splicing of the X-box binding protein 1 (XBP1) mRNA, resulting in translation of the transcription factor spliced XBP1 protein [pXBP1(s)] (Yoshida et al., 2001). Third, the eukaryotic initiation factor-2α (eIF2α) gets phosphorylated via PERK. The resulting pXBP(s), p50-ATF6α and p-eIF2α control the genes related to protein quality control, ER translocation, folding, components of the ER-associated protein degradation pathway, and genes required for lipid synthesis (Wang et al., 2000; Parker et al., 2001; Lee et al., 2003; Sriburi et al., 2004; Thuerauf et al., 2004, 2007).

While the activation of UPR in ALS has been shown in several preparations of spinal cord from ALS patients and animal models, there is no analysis of all three UPR markers in pure motor neurons. Here we investigated the expression of XBP1, ATF6α and p-eIF2α and their induction in motor neuron cultures of transgenic mutant SOD1G93A mice. We found that ER stress is a feature of the SOD1G93A-positive motor neurons, supporting the thesis of selective vulnerability of motor neurons.

Material and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Immuncytochemistry
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Animals

All animal experiments were conducted in accordance with the requirements of the National Act on the Use of Experimental Animals in Germany, and approved by the ‘Tierschutzkommission des Landes Thüringen’. The transgenic mice overexpressing the G93A mutant of human SOD1 [B6SJL-Tg(SOD1*G93A)1Gur/J, JAX Mice stock number 004435] were obtained by the Jackson Laboratory (Bar Harbor, ME, USA) and mated with C57Bl/6 females. As controls, nontransgenic littermates were used (referred to below as WT). Animals of either sex were housed under controlled laboratory conditions (temperature 22 °C, relative air humidity 55–60%, L : D 12 : 12 h, lights on at 06 : 00 h), with free access to standard diet and tap water.

Motor neuron culture

Mouse spinal ventral cords from 13-day-old mouse embryos were dissected and motor neurons were cultured on a glial feeder layer as previously described (Van Den Bosch et al., 2000; Van Damme et al., 2003). The genotype of embryos was defined according to the genotyping protocol of the Jackson Laboratory (Bar Harbor, ME, USA). Briefly, ventral spinal cords were stored in Hanks’ balanced salt solution (HBSS; Gibco, UK) then digested for 15 min in 0.1% trypsin (Gibco) in HBSS at 37 °C, then tissue was triturated with fire-polished pipettes. By centrifugation on a 6.2% OptiPrep (Axis-shield Poc AS, Oslo, Norway) cushion the motor neurons and glial cells were purified. Glial feeder layers were prepared by plating the glial cells on poly-l-ornithin (1 mg/mL)-coated 12-mm dishes at a density of 50 000 cells per dish in DMEM/Ham′s F12-Medium (PAA, Austria) supplemented with fetal calf serum (10%) or horse serum (10%), penicillin (10 U/mL) and streptomycin (10 μg/mL). In these cultures, glial cells rapidly proliferate and reach confluency after 2–3 weeks in vitro. Before seeding the motor neurons the glial cell division was halted by exposure to 5 μm cytosine arabinoside (Calbiochem, Germany) for 24 h followed by washout. Motor neurons were seeded on a pre-established glial feeder layer at a density of 15 000 cells per dish. The motor neuron culture medium consisted of neurobasal medium, B27 neuromix (2%), N2 supplement (0.2%), L-Glutamin (1 mm), horse serum (2%), penicillin (10 U/mL), streptomycin (10 μg/mL) and BDNF (2 ng/mL), all from Gibco. The cultures were kept in a 5% CO2 humidified incubator at 37 °C and used for measurements after 2 weeks in vitro.

NSC34 cell culture

Mouse motor neuronal cell line NSC-34 is a hybrid NSC cell line (neuroblastoma × spinal cord) that resembles motor neurons, displaying a multipolar neuron-like phenotype. The cells express choline acetyltransferase and neurofilament triplet proteins, and generate action potentials (Cashman et al., 1992). Therefore it is considered the best stable motor neuronal cell line model system available. This line was stably transfected with the pTet-ON plasmid (Clontech, Palo Alto, CA, USA), coding for the reverse tetracycline-controlled transactivator, to obtain the line designed NSC-34-ON7, which displays a very low level of basal expression and high inducibility. NSC-34-ON7 was used to construct inducible cell lines expressing the cDNAs encoding human WT SOD1 or the human SOD1G93A inserted in the pTRE2 plasmids as previously described (Ferri et al., 2006). The cell lines used in this study were grown in DMEM (Gibco) supplemented with 10% tetracycline-free FCS (PAA), in an atmosphere of 5% CO2 humidified incubator at 37 °C. Before each experiment (RT-PCR, Western blot) the induction of human WT SOD1 or SOD1G93A expression was obtained by adding 2 μg/mL doxycycline (Clontech) to the culture medium for the last 48 h. For each experiment (RT-PCR, Western blot) cells were plated at a density of 2 × 105 cells/mL of culture medium.

Stress treatments

Motor neuron cultures on day 14 in culture and NSC34 cultures induced for expression of human WT SOD1 and SOD1G93A were exposed to thapsigargin (0.5 μm) for 6 and 12 h. Chemicals were first dissolved in DMSO and further diluted in neurobasal medium (Gibco) for motor neurons and in DMEM (Gibco) supplemented with 10% tetracycline-free FCS (PAA) for NSC34.

RT-PCR

Total RNA was isolated from the NSC34 line using the RNA isolating kit RNeasy Mico kit 50 (Qiagen) according to the manufacturer’s protocol. Each experiment was repeated at least three times. To remove contaminating DNA, total cell RNA was treated with the RNase-free DNase Set (Qiagen). RNA was reverse-transcribed using the cDNA kit (Biorad). The cDNA was then PCR-amplified with the Master Mix-100Rxns (5Prime) using the following primer sequences: N-ATF6α (221 bp, annealing temperature 60 °C) F 5′ GGCGCCATGGAGTCGCCTTT 3′ and R 3′ GAGCAGAAGTGGCTGCCGGG 5′; XBP1 (299 bp, 58 °C) F 5′ AGCAGCAAGTGGTGGATTTGGAAG 3′ and R 3′ AAGAGGCAACAGTGTCAGAGTCCA 5′ (Cho et al., 2009); spliced XBP1 (164 bp, 58 °C) F 5′ GGTCTGCTGAGTCCGCAGCAGG 3′ and R 3′ TGACAGGGTCCAACTTGTCCAGAA 5′ (Cho et al., 2009); GAPDH (382 bp, 58 °C) F 5′ AACTTTGGCATTGTGGAAGGGCTC 3′ and R 3′ TGGAAGAGTGGGAGTTGCTGTTGA 5′. One-tenth of each reaction product was electrophoresed on a 1% agarose gel. The PCR product bands were quantified by densitometric analysis using a Gel Analyzer from INTAS® and the ratio of their expression to the expression of the housekeeping gene (GAPDH) was calculated using ImageJ software, and statistical differences between human SOD1G93A and human WT SOD1 were calculated using Student′s t-test (nondirectional), respectively anova and Bonferroni correction for multiple comparisons (PASW® Statistics 18). Differences were considered significant at P < 0.05.

Immuncytochemistry

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Immuncytochemistry
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Cultures were immersion-fixed in 4% paraformaldehyde diluted in PBS (pH 7.4) for 20 min; after washing with PBS they were incubated at room temperature with the primary antibodies ATF6α rabbit polyclonal antibody 1 : 250 (Santa Cruz, CA, USA; sc-22799), XBP1 rabbit polyclonal antibody 1 : 100 (Santa Cruz), pXBP1(u) rabbit polyclonal 1 : 250 (Abcam; ab37151), p-eIF2α goat polyclonal 1 : 100 (Santa Cruz, CA, USA; sc-12412), SMI32 mouse monoclonal antibody 1 : 1500 (Covance, CA, USA), β3-tubulin mouse monoclonal 1 : 250 (Milipore; Cat. no. MAB1637), human SOD1 rabbit polyclonal 1 : 100 (Abcam; 52950), human and mouse SOD1 mouse monoclonal 1 : 100 (Abcam; ab20926), calnexin mouse monoclonal 1 : 250 (Abcam; 31290), IBA1 rabbit polyclonal (Wako; No. 019-19741), NG2 rabbit polyclonal (Millipore; AB5320), CNPase mouse monoclonal 1 : 500 (Abcam; ab6319), GSTpi mouse monoclonal 1 : 500 (Bioscience; 610718) and GFAP goat polyclonal 1 : 500 (Santa Cruz; sc-6171) in PBS with 0.3% Triton X-100 and 2% normal goat or donkey serum for 2 h. Specifity of all used antibodies was already demonstrated by other research groups (for details see datasheets). Following incubation with primary antibodies, the cultures were washed in PBS and incubated for 1 h with the respective secondary antibodies Alexa 488 goat anti rabbit and Alexa 594 goat anti mouse (both from Invitrogen, Paisley, UK), in 10% goat or donkey serum. DAPI (Sigma, Germany) was administered for 5 minutes. The specimens were examined using a laser confocal scanning microscope (Axiophot, Zeiss).

For quantification of nuclear staining, five sections at regular intervals (Z-stacks) through motor neuron nucleui (n = 30) were selected. All the images were captured using 40× magnifications at a constant PMT voltage. Other settings (image resolution, optical zoom, scan speed, phinhole aperture, etc.) were kept uniform (Vijayalakshmi et al., 2011). Quantitative analysis of nuclear staining was performed using ImageJ software and statistical differences between SOD1G93A and WT were calculated using Student′s t-test (nondirectional), respectively anova and Bonferroni correction for multiple comparisons (PASW® Statistics 18). Differences were considered significant at P < 0.05. Each determination was carried out three times with a minimum of three dishes of three different culture preparations.

Western blot

NSC34 cultured in cell culture flasks were harvested, pooled, washed thoroughly with PBS and centrifuged. Pellets were solubilized in SDS-PAGE loading buffer. Proteins were separated by SDS-PAGE on 5–20% gradient gels (10 μg/lane) under reducing conditions and transferred onto nitrocellulose. Western blots were incubated overnight with the primary antibody ATF6α 1 : 100, XBP1 1 : 50, p-eIF2α 1 : 100 (details see above) and β-actin rabbit polyclonal antibody 1 : 1000 (abcam, UK) to show equal sample loading. Secondary antisera used for immunostaing were horseradish peroxidase-conjugated goat anti rabbit 1 : 2000 IgG (Santa Cruz). Immunoreactivity was visualized using the ECL detection system (BioRad). Quantitative analysis of Western blot bands was performed using ImageJ software and statistical differences between SOD1G93A and WT SOD1 were calculated using Student′s t-test (nondirectional), respectively anova and Bonferroni correction for multiple comparisons (PASW® Statistics 18). Each experiment was repeated at least three times. Differences were considered significant at P < 0.05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Immuncytochemistry
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Motor neuron culture system and SOD1 expression

Motor neurons were cultured on a pre-established glial feeder layer. Glial feeder layers were stained with markers for oligodendrial glia (NG2, CNPase and GSTpi), microglia (IBA1) and astrocytes (GFAP). Up to 99% of glial feeder layer cells were stained as GFAP-positive astrocytes (Fig. 1A). Less than 1% of the cells were microglial (Fig. 1B) or oligodendroglial (Fig. 1C) origin. Motor neurons were visualized (Haastert et al., 2005) via SMI32 antibodies against nonphosphorylated medium/heavy chain neurofilaments. As expected human SOD1 was detectable in SOD1G93A motor neurons, but not in nontransgenic (WT) motor neurons (data not shown), showing the antibody′s specifity for human SOD1. Staining with a SOD1 antibody against human and mouse SOD1 showed different SOD1 patterns in transgenic and WT motor neurons. In WT, the motor neurons were weakly SOD1-positive in the nucleus and cytoplasm (Fig. 1D). The SOD1G93A motor neurons were frequently more strongly SOD1-positive and showed an enhanced SOD1 immunoreactivity in the cytoplasm and the neurites (Fig. 1E). This is in line with previous findings showing an increased SOD1 staining in the spinal cord of presymptomatic transgenic SOD1G93A mice in comparison to nontransgenic littermates (Guo et al., 2010) and regarding the interaction between WT and mutated SOD1 (Prudencio et al., 2010) and the formation of SOD1 inclusions in ALS (Watanabe et al., 2001; Jonsson et al., 2004; Bergemalm et al., 2010). Costaining of human WT SOD1 with calnexin demonstrated SOD1 localisation at the ER of NSC34 (Fig. 1G and H).

image

Figure 1.  (A–E) The motor neuron culture used consisted of SMI32-positive motor neurons (blue in D and E), GFAP-positive astrocytes (green in A and C) and, to a lesser extent, of IBA1-positive microglial (orange in B) and GSTpi-positive oligodendroglial (red in C) cells. (D) The WT motor neurons showed a weak nuclear (grey, DAPI) and cytoplasmic SOD1 staining, while (E) the more strongly stained SOD1G93A-positive motor neurones showed an enhanced SOD1 immunoreactivity in the cytoplasm and the neurites. (F) Expression of the overexpressed human SOD1 in NSC34 was not influenced by ER stress treatment (ns, nonsignificant). (G and H) Coexpression of human WT SOD1 (green) with calnexin (red) demonstrates SOD1 localisation at the ER of NSC34.

Download figure to PowerPoint

Induction of ER stress

Motor neuron cultures on day 14 in culture and NSC34 cultures, induced for expression of human WT SOD1 and human SOD1G93A, were exposed to thapsigargin (0.5 μm) for 6 or 12 h. Thapsigargin is a sarco/endoplasmic reticulum calcium ATPase-inhibitor, which induces rapid depletion of ER calcium. In NSC34 the expression of human WT and human SOD1G93A protein (Fig. 1F) was not influenced by the treatment with this ER stressor.

XBP1 was spliced in the presence of mutant SOD1

We performed immunostaining in motor neurons using a XBP1 antibody which detects both proteins, pXBP1(s) and pXBP1(u), which are produced from spliced and unspliced XBP1 mRNAs, respectively. A perinuclear XBP1 expression was found in every SOD1G93A and WT motor neurons in equal amounts. The amount of nuclear immunostaining, which is due to nuclear translocation of pXBP1(s), was 2.8-fold higher in SOD1G93A motor neurons (Fig. 2A) than in WT (= 5.9, < 0.001; Fig. 2B). To evaluate whether the response to ER stress is alterated in SOD1G93A motor neurons, the cultures were treated with thapsigargin and stained for XBP1 after 6 or 12 h. Treatment with thapsigargin led to a significantly enhanced XBP1 staining after 12 h in the nucleus of both SOD1G93A (= 83, < 0.001; Fig. 2C) and WT (= 500, < 0.001; Fig. 2D) motor neurons. After 6 h treatment with this ER stressor a significant increase in nuclear staining was only detectable in WT motor neurons (< 0.001), not SOD1G93A (= 0.06), probably due to the enhanced native nuclear XBP1 staining in the presence of SOD1G93A (data not shown). After 6 and 12 h the amount of activated XBP1 (nuclear staining) was significantly higher in SOD1G93A than in WT motor neurons (t12 h = 2.1, P12 h = 0.042, t6 h = 4.2, P6 h < 0.001; Fig. 2) but the proportionate increase or slew rate of XBP1 activation after adding thapsigargin was higher in WT motor neurons, probably because activation of XBP1 is already under way in SOD1G93A motor neurons and SOD1G93A motor neurons are adapted to ER stress in the native state.

image

Figure 2.  (A) In the untreated (native) state, immunostaining patterns of XBP1 (green) showed a perinuclear and slight nuclear localization of XBP1 in SOD1G93A motor neurones. (B) The WT motor neurones showed only perinuclear staining. (C and D) Following rapid depletion of ER calcium due to treatment with thapsigargin, XBP1 translocated to the nucleus in both (C) SOD1G93A and (D) WT motor neurones. (E and F) There was no difference between (E) SOD1G93A and (F) WT regarding the expression and localization of pXBP1(u). Motor neurones were detected with SMI32 (blue). DAPI is in grey. (G–J) Accordingly the mRNA (I,J) and protein (G,H) expression of sXBP1 in untreated (native) NSC34 was higher in the presence of SOD1G93A (G,I) than WT SOD1 (H,J), and changed in a time-dependent manner after treatment with thapsigargin (thap) in both cell types. Densitometric readings of the immunoblot bands and immunfluorescece from independent experiments are shown on the right side. The band intensities are expressed as mean ± SD (*< 0.05, WT vs. G93A or native vs. thaps).

Download figure to PowerPoint

Staining with an antibody against the pXBP1(u) showed the perinuclear localisation of pXBP(u) with no difference between WT (Fig. 2F) and SOD1G93A (Fig. 2E). After treatment with thapsigargin there was no change in the amount and localisation of pXBP1(u), especially no nuclear staining, in either type of motor neuron, demonstrating the specifity of the XBP antibodies used (data not shown). Therefore pXBP1(u), which is rapidly degraded by the proteasome and probably not induced before the late recovery phase of UPR, was not influenced by the presence of SOD1G93A and short duration (6 or 12 h) treatment with an ER stressor, which is in line with previous findings (Yoshida et al., 2006).

The effects observed in motor neurons may theoretically be due to SOD1 overexpression alone. Therefore the expression patterns of UPR markers were analysed in NSC34 to confirm the results of UPR, because mice overexpressing human WT SOD1 were not available for comparison.

There was a significantly higher expression of uXBP1 (threefold) and sXBP1 (twofold) mRNA (tsXBP1 = 39.9, PsXBP1 < 0.001; tuXBP1 = 7, PuXBP1 < 0.001) in human SOD1G93A NSC34 (Fig. 2I) than in human WT SOD1 NSC34 (Fig. 2J). Treatment with thapsigargin led to a significant and time-dependent (6 and 12 h) enhancement of XBP1 and sXBP1 mRNA expression in both human SOD1G93A (F = 1944, < 0.001; Fig. 2I) and human WT SOD1 (= 1822, < 0.001; Fig. 2J) NSC34. GAPDH was used to standardize the densitometric measurements.

Also, the protein expression of pXBP1(s) was higher in human SOD1G93A (Fig. 2G) than in human WT SOD1 NSC34 (= 6.5, = 0.003; Fig. 2H). Western blot analysis of NSC34 showed increased expression of pXBP1(s) after treatment with thapsigargin in both human SOD1G93A (= 5.8, P6 h = 0.248, P12 h = 0.03; Fig. 2G) and human WT SOD1 (F = 12.9, P6 h = 0.653, P12 h = 0.01; Fig. 2H). The expression of pXBP1(u) did not differ significantly between native human SOD1G93A and human WT SOD1 (= 0.35, > 0.05). Blots were stripped and reprobed for β-actin, demonstrating equal loading in all lanes.

ATF6α was activated in mutant SOD1 motor neurons

The ATF6α antibody used is raised against to the N-terminal fragment of ATF6α and detects the uncleaved protein and the cleaved fragment. Nuclear staining, indicating the activation of ATF6α, was seen clearly in native SOD1G93A motor neurons (Fig. 3A) but only weakly in WT motor neurons (Fig. 3B), which means that ATF6α was activated and cleaved in the presence of SOD1G93A. Comparison of mean nuclear intensity revealed a significant (= 2.7, = 0.008) higher amount of nuclear ATF6α in the presence of SOD1G93A (Fig. 3A) than in WT motor neurons (Fig. 3B). Activation of ATF6α after treatment with thapsigargin was demonstrated as a significantly enhanced ATF6α staining in the nucleus of both SOD1G93A (= 63, P6 h < 0.001, P12 h < 0.001) and WT (= 27, P6 h < 0.001, P12 h < 0.001) motor neurons (Fig. 3C and D). In contrast to XBP1, the activation of ATF6α was already demonstrable 6 h after treatment with thapsigargin in both WT and SOD1G93A motor neurons (Fig. 3; no images shown). While the amount of nuclear ATF6α reached a maximum in WT motor neurons after 6 h, the ATF6α activation continued in the SOD1G93A motor neurons up to 12 h (= 3.8, = 0.001).

image

Figure 3.  (A and B) In the untreated (native) state, immunostaining patterns of ATF6α showed (B) a slight perinuclear localization of ATFα in WT motor neurones but (A) a remarkable nuclear staining in SOD1G93A motor neurones, which indicates that ATF6α is activated. After induction of ER stress ATF6α was cleaved and the p50-ATF6α fragment translocated to nucleus in (D) WT and (C) SOD1G93A motor neurones. Motor neurones were detected with SMI32 (blue). DAPI is in grey. The mRNA of ATF6α in untreated (native) NSC34 was higher in the presence of SOD1G93A (G) than WT SOD1(H) and changed in a time-dependent manner after ER stress induction with thapsigargin (thap) in both cell types. Protein expression of ATF6α did not significantly differ between WT (F) and SOD1G93A (E), while the amount of p50-ATF6α was higher in SOD1G93A (E) compared to WT (F). Densitometric readings of the immunoblot bands and immunfluorescece from independent experiments are shown on the right side. The band intensities are expressed as mean ± SD (*< 0.05, WT vs. G93A or native vs. thaps).

Download figure to PowerPoint

In NSC34 the ATF6α mRNA was 3.5-fold enhanced in the presence of human SOD1G93A (Fig. 3G) compared to human WT SOD1 (= 20, P < 0.001; Fig. 3H). Treatment with the ER stressor significantly induced expression of ATF6α mRNA in both human SOD1G93A (= 179, < 0.001; Fig. 3G) and human WT SOD1 NSC34 (= 476, < 0.001; Fig. 3H).

Activation of ATF6α was assayed in Western blots through the decrease in the full-length protein ATF6α due to cleavage into its fragments and the appearance of its processed soluble cytosolic fragment p50-ATF6α (Cox et al., 2011). The protein level of ATF6α did not differ significantly (= 0.4, = 0.689) between human SOD1G93A (Fig. 3E) and human WT SOD1 (Fig. 3F), but the cytosolic fragment was significant (= 10.4, = 0.003) higher in the presence of SOD1G93A, indicating the activation of UPR. The equal amount of full-length ATF6 protein could be a consequence of de novo synthesis because of the ER stress-induced cleavage in the presence of SOD1G93A; this is supported by the higher native mRNA expression of ATF6α. ER stress through thapsigargin led to an increased cleavage of the ATF6α protein in both human SOD1G93A and human WT SOD1 NSC34. Therefore a significant decrease in ATF6α, due to its cleavage, was detectable after thapsigargin treatment in human SOD1G93A cells (= 44.6, P6 h = 0.015, P12 h < 0.001; Fig. 3E) and human WT SOD1 cells (= 27, P6 h = 0.013, P12 h = 0.004; Fig. 3F). Accordingly there was an increase of the cytosolic fragment after thapsigargin treatment in SOD1G93A cells (= 18.8, P6 h = 0.013, P12 h = 0.002; Fig. 3E) and human WT SOD1 cells (= 34.5, P6 h = 0.752, P12 h < 0.001; Fig. 3F).

Phosphorylation of eIF2α

Activation of the PERK pathway can be demonstrated by an enhanced phosphorylation of eIF2α via PERK. Immuncytochemistry revealed a significantly (= 3.3, = 0.002) stronger staining for p-eIF2α in the presence of SOD1G93A (Fig. 4A) than in WT motor neurons (Fig. 4B). ER stress induction with thapsigargin significantly enhanced phosphorylation of eIF2α in both SOD1G93A (= 91, = 0.006; Fig. 4C) and WT (= 178, = 0.001; Fig. 4D) motor neurons. ER stress induction led to a significant higher increase in p-eIF2α in the SOD1G93A cells than in WT (= 3.8, = 0.002).

image

Figure 4.  (A and B) In the untreated state the SOD1G93A motor neurones (A) showed a stronger immunostaining for p-eIF2α than did (B) the WT motor neurones. (C and D) Upon induction of ER stress, eIF2α became much more phosphorylated in (C) the SOD1G93A than (D) the WT motor neurons. (E and F) A greater amount of p-eIF2α could also be demonstrated in (E) the untreated (native) SOD1G93A NSC34 than in (F) the WT SOD1. (H) WT motor neurons growing on SOD1G93A astrocytes did not show a remarkable nuclear ATF6α staining. (G) Vice versa, activation of ATF6α was observed in SOD1G93A motor neurones growing on WT astrocytes. Therefore activation of UPR occurred in SOD1G93A motor neurones independently of the mutational status of surrounding astrocytes. Densitometric readings of the immunoblot bands and immunfluorescece from independent experiments are shown on the right side. The band intensities are expressed as mean ± SD (*< 0.05, WT vs. SOD1G93A or native vs. thaps).

Download figure to PowerPoint

In NSC34 the p-eIF2α protein was increased in the SOD1G93A cells (= 3.8, = 0.019; Fig. 4E). Treatment with thapsigargin increased the amount of p-eIF2α in SOD1G93A cells nonsignificant (= 0.7, P6 h = 0.212, P12 h = 0.366; Fig. 4E) and significant in WT neurons (F = 19, P6 h = 0.007; P12 h = 0.002; Fig. 4F). It is possible that the nonsignificance of the changes in p-eIF2α in SOD1G93A cells were due to its higher baseline expression.

ER stress in SOD1G93A motor neurons occurred independently of surrounding astrocytes

To confirm the findings that activation of UPR is a feature of motor neurons, the SOD1G93A motor neurons were seeded on a pre-established WT glial feeder (from nontransgenic embryonic spinal cord). Vice versa, WT motor neurons were seeded on a SOD1G93A glial feeder. Cells were stained for ATF6α because it showed the strongest difference in immunofluorescence between WT and SOD1G93A (see above). The presence of SOD1G93A astrocytes did not induce cleavage of ATF6α in cocultured WT motor neurons (Fig. 4H). However, activation of ATF6α was observed in SOD1G93A motor neurons growing on WT astrocytes (= 18, < 0.001; Fig. 4G). Therefore activation of UPR occurred in SOD1G93A motor neurons independently of the mutational status of surrounding astrocytes. Healthy astrocytes without the SOD1 mutation were not able to influence the activation of UPR in SOD1G93A motor neurons in this coculture system, providing evidence that ER stress is a genuine feature of motor neurons.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Immuncytochemistry
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

In this study all three UPR pathway markers were analyzed in cultured motor neurons. We present further evidence that ER stress and activation of UPR are features of ALS and occur in the motor neurons. Moreover we were able to show that cultured motor neurons are a useful tool for investigating the cellular reactions to ER stress. Motor neuron culture may be used to develop and evaluate new therapeutic approaches modulating ER stress responses in the context of the ERMCC in ALS. We observed activation of ATF6α, splicing of XBP1 and phosphorylation of eIF2α in motor neuronal-like cells expressing SOD1G93A and in cultured motor neurons from transgenic SOD1G93A mice. Following treatment with an ER stressor both SOD1G93A and WT SOD1 motor neuronal-like cells and motor neurons showed activation of ATF6α and XBP1 and phosphorylation of eIF2α. The activation of ATF6α occured earlier than activation of XBP1, because the former is activated through proteolysis and the latter gets upregulated by de novo synthesis. The XBP1 mRNA must be induced to a significant level to produce the sXBP1 mRNA, which is preceded by the production of active and nuclear p50-ATF6α fragments (Yoshida et al., 2001). Besides the fact that UPR is already underway in the SOD1G93A cells, it seems that regulation of UPR after additional ER stress induction is disturbed in ALS, with a prolonged and stronger activation of the UPR. Further analyses regarding the long-lasting effects of ER stress induction in mutant SOD1 are necessary to describe this deregulation in more detail.

Recent studies indicate that ER stress is involved in the pathogenesis of familial and sporadic ALS. Western blot analysis of spinal cords from mutant SOD1 transgenic mice model and from ALS patients showed upregulation of all three UPR sensors IRE1, ATF6 and PERK (Atkin et al., 2006, 2008). Translocation of ATF6α to the nucleus after ER stress treatment has been shown in motor neuronal-like cells in the presence of mutated SOD1 (Oh et al., 2008). Kikuchi et al. (2006) found an age-dependent activation of ATF6α specifically in degenerating areas such as spinal cord, but not in unaffected areas such as the cerebellum, of transgenic SOD1G93A mice. Previous studies of XBP1 have shown that spliced XBP1 mRNA was more abundant in symptomatic transgenic SOD1G93A than in age-matched nontransgenic spinal cords (Kikuchi et al., 2006), and splicing of XBP1 mRNA occurs in the presence of SOD1G93A and SOD1G85R in motor neuronal-like cells (Oh et al., 2008). However, there was no difference in XBP1 mRNA levels between symptomatic transgenic SOD1G93A and age-matched nontransgenic spinal cords (Kikuchi et al., 2006). Activation of PERK and eIF2α/p-eIF2α was observed in SOD1G93A mice (Nagata et al., 2007; Saxena et al., 2009) and ALS patients (Ilieva et al., 2007; Atkin et al., 2008). Here we were able to demonstrate that ER stress is a genuine feature of SOD1G93A motor neurons and occurs independently of the neighboring glia, supporting the thesis of selective vulnerability of motor neurons in ALS. Nevertheless, astrocytes seem to play an essential role in pathogenesis of ALS (Ilieva et al., 2009; Philips & Robberecht, 2011). Downregulation of mutant SOD1 in astrocytes delayed late disease progression (Yamanaka et al., 2008). Mutant astrocytes seem to affect motor neurons by unknown neurotoxic effectors (Nagai et al., 2007; Papadeas et al., 2011), but motor neuron degeneration is not consistently demonstrated (Gong et al., 2000). Therefore one has to take into account that the missing induction of ER stress in WT motor neurons by SOD1G93A astrocytes can also be a result of the short-lasting coculture in this cellular model. It is not yet clear whether ER stress is also generated by mutant SOD1 within astrocytes.

The functional link between the SOD1 mutation and the ER stress in ALS is still not clear. It seems that ALS-associated SOD1 mutations exert toxicity by a gain of function rather than by diminishing enzymatic activity. ER stress could be a consequence of oligomerization of the SOD1 peptide with itself or with other proteins forming aggregates (aggregation hypothesis; Shaw & Valentine, 2007). Despite this, other studies suggest that aggregation of SOD1 is not the reason for ALS disease progression (Proescher et al., 2008; Karch et al., 2009). Accumulating mutant SOD1 aggregates in the ER membranes and can bind to the calcium-dependent luminal chaperone Grp78 (Kikuchi et al., 2006), which regulates the activation of the UPR. Moreover, mutant SOD1 interacts with derlin-1, an ER protein that is necessary for dislocation of misfolded proteins from the ER to the cytosol, resulting in inhibition of the ER-associated degradation pathway (Nishitoh et al., 2008).

On the other hand, protein misfolding can occur when the ER calcium homeostasis is disturbed, because protein processing and folding are calcium-dependent and the ER is functionally and physically (Friedman et al., 2011) linked to mitochondria and its cytosolic calcium transients through the ERMCC (Grosskreutz et al., 2010). In this context a disturbance of ER calcium refilling can be followed by protein misfolding and activation of UPR. To confirm this hypothesis, further studies are necessary to evaluate the effect of drugs which interfere with ERMCC on ER stress in ALS models.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Immuncytochemistry
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

This research is supported by the Bundesministerium für Bildung und Forschung (BMBF) grant to J.G. (ERMCC-NDEG) within the framework of the ERANET–NEURON program of the European Union. T.P. is supported by an IZKF grant from University Hospital Jena. For technical assistance we thank Svetlana Tausch, Julia Oberland and Diana Freitag. For allocation of antibodies we thank Silke Keiner and Christian Schmeer.

Abbreviations
ALS

amyotrophic lateral sclerosis

ATF6

basic leucine-zipper transcription factor 6

eIF2α

eukaryotic initiation factor-2

ER

endoplasmic recticulum

ERMCC

endoplasmic recticulum mitochondria calcium cycle

Grp78

glucose-regulated protein 78

IRE1

inositol-requiring enzyme 1

PERK

double-stranded RNA-activated protein kinase-like ER kinase

pXBP1(s)

spliced XBP1 protein

pXBP1(u)

unspliced XBP1 protein

p50-ATF6α

N-terminal fragment of ATF6α

SOD1

Cu/Zn superoxide dismutase type I

UPR

unfolded protein response

WT

nontransgenic littermates (used as controls)

sXBP1

spliced XBP1

uXBP1

unspliced XBP1

XBP1

X-box binding protein 1

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Immuncytochemistry
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References
  • Atkin, J.D., Farg, M.A., Turner, B.J., Tomas, D., Lysaght, J.A., Nunan, J., Rembach, A., Nagley, P., Beart, P.M., Cheema, S.S. & Horne, M.K. (2006) Induction of the unfolded protein response in familial amyotrophic lateral sclerosis and association of protein disulfide isomerase with superoxide dismutase 1. J. Biol. Chem., 281, 3015230165.
  • Atkin, J.D., Farg, M.A., Walker, A.K., McLean, C., Tomas, D. & Horne, M.K. (2008) Endoplasmic reticulum stress and induction of the unfolded protein response in human sporadic amyotrophic lateral sclerosis. Neurobiol. Dis., 30, 400407.
  • Bergemalm, D., Forsberg, K., Srivastava, V., Graffmo, K.S., Andersen, P.M., Brannstrom, T., Wingsle, G. & Marklund, S.L. (2010) Superoxide dismutase-1 and other proteins in inclusions from transgenic amyotrophic lateral sclerosis model mice. J. Neurochem., 114, 408418.
  • Berridge, M.J. (2002) The endoplasmic reticulum: a multifunctional signaling organelle. Cell Calcium, 32, 235249.
  • Cashman, N.R., Durham, H.D., Blusztajn, J.K., Oda, K., Tabira, T., Shaw, I.T., Dahrouge, S. & Antel, J.P. (1992) Neuroblastoma × spinal cord (NSC) hybrid cell lines resemble developing motor neurons. Dev. Dyn., 194, 209221.
  • Cho, Y.M., Jang, Y.S., Jang, Y.M., Chung, S.M., Kim, H.S., Lee, J.H., Jeong, S.W., Kim, I.K., Kim, J.J., Kim, K.S. & Kwon, O.J. (2009) Induction of unfolded protein response during neuronal induction of rat bone marrow stromal cells and mouse embryonic stem cells. Exp. Mol. Med., 41, 440452.
  • Cox, D.J., Strudwick, N., Ali, A.A., Paton, A.W., Paton, J.C. & Schroder, M. (2011) Measuring signaling by the unfolded protein response. Methods Enzymol., 491, 261292.
  • Dudek, J., Benedix, J., Cappel, S., Greiner, M., Jalal, C., Muller, L. & Zimmermann, R. (2009) Functions and pathologies of BiP and its interaction partners. Cell. Mol. Life Sci., 66, 15561569.
  • Ferri, A., Cozzolino, M., Crosio, C., Nencini, M., Casciati, A., Gralla, E.B., Rotilio, G., Valentine, J.S. & Carri, M.T. (2006) Familial ALS-superoxide dismutases associate with mitochondria and shift their redox potentials. Proc. Natl Acad. Sci. USA, 103, 1386013865.
  • Friedman, J.R., Lackner, L.L., West, M., DiBenedetto, J.R., Nunnari, J. & Voeltz, G.K. (2011) ER tubules mark sites of mitochondrial division. Science, 334, 358362.
  • Gong, Y.H., Parsadanian, A.S., Andreeva, A., Snider, W.D. & 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, 660665.
  • Grosskreutz, J., Van Den, B.L. & Keller, B.U. (2010) Calcium dysregulation in amyotrophic lateral sclerosis. Cell Calcium, 47, 165174.
  • Guo, Y., Li, C., Wu, D., Wu, S., Yang, C., Liu, Y., Wu, H. & Li, Z. (2010) Ultrastructural diversity of inclusions and aggregations in the lumbar spinal cord of SOD1-G93A transgenic mice. Brain Res., 1353, 234244.
  • 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. (1994) Motor neuron degeneration in mice that express a human Cu,Zn superoxide dismutase mutation. Science, 264, 17721775.
  • Haastert, K., Grosskreutz, J., Jaeckel, M., Laderer, C., Bufler, J., Grothe, C. & Claus, P. (2005) Rat embryonic motoneurons in long-term co-culture with Schwann cells-a system to investigate motoneuron diseases on a cellular level in vitro. J. Neurosci. Methods, 142, 275284.
  • Ilieva, E.V., Ayala, V., Jove, M., Dalfo, E., Cacabelos, D., Povedano, M., Bellmunt, M.J., Ferrer, I., Pamplona, R. & Portero-Otin, M. (2007) Oxidative and endoplasmic reticulum stress interplay in sporadic amyotrophic lateral sclerosis. Brain, 130, 31113123.
  • Ilieva, H., Polymenidou, M. & Cleveland, D.W. (2009) Non-cell autonomous toxicity in neurodegenerative disorders: ALS and beyond. J. Cell Biol., 187, 761772.
  • Jonsson, P.A., Ernhill, K., Andersen, P.M., Bergemalm, D., Brannstrom, T., Gredal, O., Nilsson, P. & Marklund, S.L. (2004) Minute quantities of misfolded mutant superoxide dismutase-1 cause amyotrophic lateral sclerosis. Brain, 127, 7388.
  • Karch, C.M., Prudencio, M., Winkler, D.D., Hart, P.J. & Borchelt, D.R. (2009) Role of mutant SOD1 disulfide oxidation and aggregation in the pathogenesis of familial ALS. Proc. Natl Acad. Sci. USA, 106, 77747779.
  • Kikuchi, H., Almer, G., Yamashita, S., Guegan, C., Nagai, M., Xu, Z., Sosunov, A.A., McKhann, G.M. & Przedborski, S. (2006) Spinal cord endoplasmic reticulum stress associated with a microsomal accumulation of mutant superoxide dismutase-1 in an ALS model. Proc. Natl Acad. Sci. USA, 103, 60256030.
  • Kozutsumi, Y., Segal, M., Normington, K., Gething, M.J. & Sambrook, J. (1988) The presence of malfolded proteins in the endoplasmic reticulum signals the induction of glucose-regulated proteins. Nature, 332, 462464.
  • Lee, A.H., Iwakoshi, N.N. & Glimcher, L.H. (2003) XBP-1 regulates a subset of endoplasmic reticulum resident chaperone genes in the unfolded protein response. Mol. Cell. Biol., 23, 74487459.
  • Nagai, M., Re, D.B., Nagata, T., Chalazonitis, A., Jessell, T.M., Wichterle, H. & Przedborski, S. (2007) Astrocytes expressing ALS-linked mutated SOD1 release factors selectively toxic to motor neurons. Nat. Neurosci., 10, 615622.
  • Nagata, T., Ilieva, H., Murakami, T., Shiote, M., Narai, H., Ohta, Y., Hayashi, T., Shoji, M. & Abe, K. (2007) Increased ER stress during motor neuron degeneration in a transgenic mouse model of amyotrophic lateral sclerosis. Neurol. Res., 29, 767771.
  • Nakagawa, T., Zhu, H., Morishima, N., Li, E., Xu, J., Yankner, B.A. & Yuan, J. (2000) Caspase-12 mediates endoplasmic-reticulum-specific apoptosis and cytotoxicity by amyloid-beta. Nature, 403, 98103.
  • Nishitoh, H., Kadowaki, H., Nagai, A., Maruyama, T., Yokota, T., Fukutomi, H., Noguchi, T., Matsuzawa, A., Takeda, K. & Ichijo, H. (2008) ALS-linked mutant SOD1 induces ER stress- and ASK1-dependent motor neuron death by targeting Derlin-1. Genes Dev., 22, 14511464.
  • Oh, Y.K., Shin, K.S., Yuan, J. & Kang, S.J. (2008) Superoxide dismutase 1 mutants related to amyotrophic lateral sclerosis induce endoplasmic stress in neuro2a cells. J. Neurochem., 104, 9931005.
  • Papadeas, S.T., Kraig, S.E., O’Banion, C., Lepore, A.C. & Maragakis, N.J. (2011) Astrocytes carrying the superoxide dismutase 1 (SOD1G93A) mutation induce wild-type motor neuron degeneration in vivo. Proc. Natl Acad. Sci. USA, 108, 1780317808.
  • Parker, R., Phan, T., Baumeister, P., Roy, B., Cheriyath, V., Roy, A.L. & Lee, A.S. (2001) Identification of TFII-I as the endoplasmic reticulum stress response element binding factor ERSF: its autoregulation by stress and interaction with ATF6. Mol. Cell. Biol., 21, 32203233.
  • Philips, T. & Robberecht, W. (2011) Neuroinflammation in amyotrophic lateral sclerosis: role of glial activation in motor neuron disease. Lancet Neurol., 10, 253263.
  • Proescher, J.B., Son, M., Elliott, J.L. & Culotta, V.C. (2008) Biological effects of CCS in the absence of SOD1 enzyme activation: implications for disease in a mouse model for ALS. Hum. Mol. Genet., 17, 17281737.
  • Prudencio, M., Durazo, A., Whitelegge, J.P. & Borchelt, D.R. (2010) An examination of wild-type SOD1 in modulating the toxicity and aggregation of ALS-associated mutant SOD1. Hum. Mol. Genet., 19, 47744789.
  • Rothstein, J.D. (2009) Current hypotheses for the underlying biology of amyotrophic lateral sclerosis. Ann. Neurol., 65(Suppl 1), S3S9.
  • Saxena, S., Cabuy, E. & Caroni, P. (2009) A role for motoneuron subtype-selective ER stress in disease manifestations of FALS mice. Nat. Neurosci., 12, 627636.
  • Shaw, B.F. & Valentine, J.S. (2007) How do ALS-associated mutations in superoxide dismutase 1 promote aggregation of the protein? Trends Biochem. Sci., 32, 7885.
  • Sriburi, R., Jackowski, S., Mori, K. & Brewer, J.W. (2004) XBP1: a link between the unfolded protein response, lipid biosynthesis, and biogenesis of the endoplasmic reticulum. J. Cell Biol., 167, 3541.
  • Thuerauf, D.J., Morrison, L. & Glembotski, C.C. (2004) Opposing roles for ATF6alpha and ATF6beta in endoplasmic reticulum stress response gene induction. J. Biol. Chem., 279, 2107821084.
  • Thuerauf, D.J., Marcinko, M., Belmont, P.J. & Glembotski, C.C. (2007) Effects of the isoform-specific characteristics of ATF6 alpha and ATF6 beta on endoplasmic reticulum stress response gene expression and cell viability. J. Biol. Chem., 282, 2286522878.
  • Van Damme, P., Callewaert, G., Eggermont, J., Robberecht, W. & Van Den Bosch, L. (2003) Chloride influx aggravates Ca2+-dependent AMPA receptor-mediated motoneuron death. J. Neurosci., 23, 49424950.
  • Van Den Bosch, L., Vandenberghe, W., Klaassen, H., Van Houtte, E. & Robberecht, W. (2000) Ca(2+ )-permeable AMPA receptors and selective vulnerability of motor neurons. J. Neurol. Sci., 180, 2934.
  • Verkhratsky, A. (2005) Physiology and pathophysiology of the calcium store in the endoplasmic reticulum of neurons. Physiol. Rev., 85, 201279.
  • Vijayalakshmi, K., Alladi, P.A., Ghosh, S., Prasanna, V.K., Sagar, B.C., Nalini, A., Sathyaprabha, T.N. & Raju, T.R. (2011) Evidence of endoplasmic reticular stress in the spinal motor neurons exposed to CSF from sporadic amyotrophic lateral sclerosis patients. Neurobiol. Dis., 41, 695705.
  • Walker, A.K. (2010) Protein disulfide isomerase and the endoplasmic reticulum in amyotrophic lateral sclerosis. J. Neurosci., 30, 38653867.
  • Wang, Y., Shen, J., Arenzana, N., Tirasophon, W., Kaufman, R.J. & Prywes, R. (2000) Activation of ATF6 and an ATF6 DNA binding site by the endoplasmic reticulum stress response. J. Biol. Chem., 275, 2701327020.
  • Watanabe, M., Dykes-Hoberg, M., Culotta, V.C., Price, D.L., Wong, P.C. & Rothstein, J.D. (2001) Histological evidence of protein aggregation in mutant SOD1 transgenic mice and in amyotrophic lateral sclerosis neural tissues. Neurobiol. Dis., 8, 933941.
  • Yamanaka, K., Chun, S.J., Boillee, S., Fujimori-Tonou, N., Yamashita, H., Gutmann, D.H., Takahashi, R., Misawa, H. & Cleveland, D.W. (2008) Astrocytes as determinants of disease progression in inherited amyotrophic lateral sclerosis. Nat. Neurosci., 11, 251253.
  • Yoshida, H., Matsui, T., Yamamoto, A., Okada, T. & Mori, K. (2001) XBP1 mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor. Cell, 107, 881891.
  • Yoshida, H., Oku, M., Suzuki, M. & Mori, K. (2006) pXBP1(U) encoded in XBP1 pre-mRNA negatively regulates unfolded protein response activator pXBP1(S) in mammalian ER stress response. J. Cell Biol., 172, 565575.