The voltage-gated calcium channel blocker lomerizine is neuroprotective in motor neurons expressing mutant SOD1, but not TDP-43

Authors

  • Luan T. Tran,

    1. Department of Neurology/Neurosurgery, Montreal Neurological Institute, McGill University, Montreal, QC, Canada
    Search for more papers by this author
  • Benoit J. Gentil,

    1. Department of Neurology/Neurosurgery, Montreal Neurological Institute, McGill University, Montreal, QC, Canada
    Search for more papers by this author
  • Kathleen E. Sullivan,

    1. Department of Neurology/Neurosurgery, Montreal Neurological Institute, McGill University, Montreal, QC, Canada
    Search for more papers by this author
  • Heather D. Durham

    Corresponding author
    1. Department of Neurology/Neurosurgery, Montreal Neurological Institute, McGill University, Montreal, QC, Canada
    • Address correspondence and reprint requests to Dr Heather D. Durham, Rm 649, Montreal Neurological Institute, 3801 University St., Montreal, QC H3A 2B3, Canada. E-mail: heather.durham@mcgill.ca

    Search for more papers by this author

Abstract

Excitotoxicity and disruption of Ca2+ homeostasis have been implicated in amyotrophic lateral sclerosis (ALS) and limiting Ca2+ entry is protective in models of ALS caused by mutation of SOD1. Lomerizine, an antagonist of L- and T-type voltage-gated calcium channels and transient receptor potential channel 5 transient receptor potential channels, is well tolerated clinically, making it a potential therapeutic candidate. Lomerizine reduced glutamate excitotoxicity in cultured motor neurons by reducing the accumulation of cytoplasmic Ca2+ and protected motor neurons against multiple measures of mutant SOD1 toxicity: Ca2+ overload, impaired mitochondrial trafficking, mitochondrial fragmentation, formation of mutant SOD1 inclusions, and loss of viability. To assess the utility of lomerizine in other forms of ALS, calcium homeostasis was evaluated in culture models of disease because of mutations in the RNA-binding proteins transactive response DNA-binding protein 43 (TDP-43) and Fused in Sarcoma (FUS). Calcium did not play the same role in the toxicity of these mutant proteins as with mutant SOD1 and lomerizine failed to prevent cytoplasmic accumulation of mutant TDP-43, a hallmark of its pathology. These experiments point to differences in the pathogenic pathways between types of ALS and show the utility of primary culture models in comparing those mechanisms and effectiveness of therapeutic strategies.

image

Calcium sensitivity is a factor in motor neuron vulnerability in ALS. The voltage-gated calcium channel blocker lomerizine normalized [Ca2+] and reduced toxicity of mutant Cu/Zn-superoxide dismutase (SOD1) causing familial ALS1. Calcium homeostasis was not disrupted in motor neurons expressing ALS-associated mutants of TAR DNA-binding protein 43 (TDP-43) or FUS, nor was lomerizine protective, affirming differences in pathogenic mechanism and therapeutic efficacy in the forms of ALS.

Abbreviations used
[Ca2+]

Ca2+ concentration

ALS10

ALS due to TARDP mutation

ALS1

ALS due to SOD1 mutation

ALS6

ALS due to FUS mutation

ALS

amyotrophic lateral sclerosis

AMPA

alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate

DMSO

dimethyl sulfoxide

DRG

dorsal root ganglia

eGFP

enhanced green fluorescent protein

FITC

fluorescein isothiocyanate

FUS

fused in sarcoma/translated in liposarcoma

GluR2

glutamate receptor subunit 2

PBS

phosphate-buffered saline

SOD1

Cu/Zn-superoxide dismutase

TDP-43

TAR DNA-binding protein 43

VGCC

voltage-gated calcium channel

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease arising sporadically or associated with one of the several genetic mutations. The first ALS gene identified was SOD1 encoding Cu/Zn-superoxide dismutase (ALS1) (Rosen et al. 1993), but several others have been identified including the genes encoding RNA-binding proteins [transactive response DNA-binding protein 43 (TDP-43) (Arai et al. 2006; Neumann et al. 2006), fused in sarcoma/translated in liposarcoma (FUS) (Kwiatkowski et al. 2009; Vance et al. 2009) and rho guanyl-nucleotide exchange factor, a rhoGEF nucleotide exchange factor (Droppelmann et al. 2013)]. Loss of respiratory and motor function ultimately results from dysfunction and loss of motor neurons. A key question is how their physiological properties render these neurons vulnerable to these genetic mutations and to other factors underlying sporadic and familial ALS.

Affected pools of motor neurons are particularly susceptible to glutamate excitotoxicity through calcium-dependent mechanisms, with their high level of Ca2+-permeable glutamate receptors (Williams et al. 1997; Carriedo et al. 1998; Vandenberghe et al. 2000; Kawahara et al. 2003; Tateno et al. 2004; Corona and Tapia 2007), and low levels of cytosolic calcium-binding proteins (Ince et al. 1993; Alexianu et al. 1994; Elliott and Snider 1995; Junttila et al. 1995; Reiner et al. 1995; Lips and Keller 1998). Excitotoxic mechanisms can be exacerbated in ALS by loss of the astrocytic excitatory amino acid transporter-2 glutamate transporter for neurotransmitter reuptake (Rothstein et al. 1995). Furthermore, calcium is increased in mitochondria of motor nerve terminals of sporadic ALS patients (Siklos et al. 1996) and Q/R editing of glutamate receptor subunit 2, the subunit responsible for preventing Ca2+ permeability of alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA) glutamate receptors, can be reduced (Takuma et al. 1999). Most evidence for loss of calcium homeostasis comes from experimental models of ALS1 (Roy et al. 1998; Siklós et al. 1998; Kruman et al. 1999; Jaiswal et al. 2009; Tradewell et al. 2011) in which neuroprotection is achieved by pharmacological manipulation of Ca2+ influx (Roy et al. 1998; Ghadge et al. 2003; Van Damme et al. 2003, 2005; Tateno et al. 2004; Tortarolo et al. 2006)] or by increasing Ca2+-buffering capacity (Roy et al. 1998; Beers et al. 2001).

Taken together, these observations suggest that maintaining calcium homeostasis might be of therapeutic benefit, particularly in ALS1. Although manipulating glutamate receptor composition or activity is protective in culture models, there are limitations to applying those strategies in patients in whom neurotransmission must be maintained. Another target is voltage-gated calcium channels (VGCC). VGCC inhibitors have a history of long-term clinical use, particularly in cardiovascular disease. Lomerizine is a dual L/T-type channel blocker used for prophylaxis of migraine, particularly in Japan. The drug has a good safety profile with long-term administration in humans, with less effect on blood pressure or heart rate (Imai et al. 2007), and distributes to nervous tissue (Awata et al. 1995). This drug also blocks the transient receptor potential channel 5 an oxidative stress-sensitive, Ca2+-permeable channel (Ishii et al. 2011). We evaluated the neuroprotective properties of lomerizine in a primary culture model of ALS1. Expression of SOD1 mutants in cultured motor neurons, but not SOD1WT, results in formation of inclusions containing the mutant protein and loss of viability over 1–2 weeks (Durham et al. 1997). More upstream abnormalities in cultured motor neurons expressing SOD1G93A are initial increase in mitochondrial [Ca2+] followed by dramatic mitochondrial fragmentation (and partial reduction in membrane potential – Δψ) and increased endoplasmic reticular [Ca2+]. Cytosolic [Ca2+] increases and is significantly higher in neurons harboring inclusions of mutant protein (Tradewell et al. 2011). In this study, impaired mitochondrial dynamics, including inhibition of mitochondrial fusion and mitochondrial axonal transport were also demonstrated, as previously reported in other models (De Vos et al. 2007; Magrane et al. 2009, 2012; Vande Velde et al. 2011; Song et al. 2012). Lomerizine treatment reduced all of these manifestations of mutant SOD1 toxicity.

Despite the evidence for excitotoxicity contributing to pathogenesis and efficacy of anti-glutamate approaches in experimental models of ALS1, neither glutamate receptor blockers nor the VGCC antagonists, nimodipine, and verapamil, have proven efficacious in sporadic ALS patients (Miller et al. 1996a,b; Kriz et al. 2003; Zinman and Cudkowicz 2011). Riluzole, a sodium channel blocker that reduces transmitter release, is minimally effective (Bensimon et al. 1994). One possibility is that these mechanisms are not major contributors to pathogenesis in all types of ALS or in all individuals. Indeed, Armstrong and Drapeau recently reported that VGCC agonists restore synaptic function in zebrafish larvae expressing mutant TDP-43 (Armstrong and Drapeau 2013).

In this study, we assessed the involvement of calcium and relevance of lomerizine, in forms of ALS resulting from mutations in the RNA-binding proteins, Fused in Sarcoma (FUS; ALS6) and TDP-43 (ALS10). Overexpression of wild type or mutant human TDP-43 or FUS did not alter Ca2+ levels in cultured mouse motor neurons, nor did lomerizine prevent the accumulation of cytoplasmic TDP-43 associated with its toxicity.

Methods

Dissociated spinal cord DRG cultures

Cultures were prepared from E13 CD1 mouse embryos and cultured in hormone- and growth factor-enriched medium, and motor neurons were identified, as previously described (Roy et al. 1998). Dams at 13-day gestation were purchased from Charles River Labs (St. Constant, QC, Canada). Cultures were used in experiments 3–6 weeks following dissociation to allow maturation of motor neurons. All studies were carried out in accordance with the Canadian Council on Animal Care and ARRIVE guidelines.

Gene transfer

Plasmids: Human SOD1WT or the ALS-causing mutant, SOD1G93A, in pCEP4 (injected at 200 μg/mL); SOD1WT-enhanced green fluorescent protein (eGFP) or SOD1G93A-eGFP in eGFPN1 (injected at 4 μg/mL) (Tradewell et al. 2011); eGFPN1 (injected at 2 μg/mL); eGFP-FUSWT, eGFP-FUSR521H, and eGFP-FUSP525L in pEGFP-C1 (Tradewell et al. 2012) and TDP43WT-eGFP and TDP43G348C-eGFP and TDP-43A315T-eGFP in pEGFP-N1, or TDP-43 WT and mutants with a C-terminal FLAG tag (Kabashi et al. 2010) (injected at 20 μg/mL); mitochondrial ratiometric pericam (mtpericam; from Dr Atsushi Miyawaki, RIKEN Brain Science Institute, Wako City, Saitama, Japan; injected at 25 μg/mL); pOCTeGFP in pcDNA3 (with the mitochondrial targeting signal sequence of ornithine carbamoyltransferase; from Dr. Heidi McBride, McGill University, Montreal, QC, Canada; injected at 2 μg/mL).

Drug treatment

A stock solution of 100 mM lomerizine (Cedarlane, Burlington, ON, Canada) was prepared in dimethyl sulfoxide (DMSO) and aliquots were stored at −80°C. Lomerizine was further diluted to 400 μM in DMSO prior to use and further diluted in culture medium to working concentrations.

Immunocytochemistry

Spinal cord–dorsal root ganglia (DRG) cultures on coverslips were fixed in 3% paraformaldehyde in phosphate-buffered saline (PBS) for 10 min; permeabilized with 0.5% Nonidet-P40 in PBS for 1 min; fixed again with paraformaldehyde for 2 min; blocked with 5% horse serum in PBS for 30 min; incubated with the primary antibody for 1 h; washed with PBS 3 times for 3 min each; incubated with secondary antibody for 30 min; washed with PBS three times for 3 min each, and mounted on glass slides using Immumount (Thermo Fisher Scientific, Mississauga, ON, Canada).

Antibodies: Human SOD1 was detected by immunocytochemistry using mouse monoclonal antibody SD-G6, which does not cross-react with murine SOD1 (Sigma Chemical Co., St. Louis, MO, USA; 1 : 300). Human and mouse TDP-43 was detected using rabbit polyclonal antibody from Proteintech (Chicago, IL, USA; 10782-2-AP; 1 : 500). Motor neuron morphology was visualized by ICC with mouse anti-microtubule-associated protein 2 (Novus Biologicals, Littleton, CO, USA; 1 : 400). Secondary antibodies for ICC (InVitrogen Life Sciences, Burlington, ON, Canada) were anti-mouse IgG-Alexa Flour 488 (1 : 200), anti-rabbit IgG-Alexa Flour 594 (1 : 200), and anti-mouse IgG-Cy3 (1 : 300).

Imaging of epifluorescence

Cultures (living or fixed and mounted on slides) were placed on the stage of an Axiovert 35 inverted microscope (Carl Zeiss Canada Ltd, Toronto, ON, Canada) equipped with epifluorescence optics and a Lambda 10-B/SmartShutter™ with Lambda 10-C filter wheel (Sutter Instruments, Co., Novato, CA, USA) containing excitation and emission filters (Chroma Technologies, Rockingham, VT, USA). Images, below fluorescence saturation level, were captured using an ORCA-ER cooled CCD digital camera (Hamamatsu Photonics K. K., Hamamatsu City, Japan) and analyzed using Universal MetaFluor® Imaging Software (Molecular Devices, Inc., Downingtown, PA, USA).

Mitochondrial length and transport

SOD1WT or SOD1G93A (untagged) was expressed with pOCT-eGFP to visualize mitochondria. For transport studies, coverslips were placed in a live cell imaging chamber (Harvard Apparatus, Montreal, QC, Canada) and mounted on the stage of the Zeiss Axiovert 35 microscope. Images were captured every 5 s for 100 frames (8 min) using a 63× 1.4 NA Apochromat objective using the ORCA-ER cooled CCD camera and Metafluor® software. Deconvolution was performed using 3-D Huygens deconvolution software (Scientific Volume Imaging, Hilversum, The Netherlands). Universal Imaging Metamorph® software (Molecular Devices, Inc.) was used to generate videos and kymographs of mitochondrial position, from which movement was assessed as moving (anterograde or retrograde) or stationary. Movement was defined as at least 4 μm change in position over four consecutive frames. Cultures were fixed after imaging and immunolabeled with antibody to human SOD1 (clone # SD-G6; Sigma-Aldrich, St. Louis, MO, USA) to confirm plasmid expression. Mitochondrial length was measured from images recorded during transport studies.

Mitochondrial [Ca2+] and Δψ

Mitochondrial [Ca2+] was measured as previously described (Tradewell et al. 2011) using the genetically encoded, mitochondrially targeted sensor, mtpericam (Shimozono et al. 2002). The mtpericam was excited alternately through 436/10 nm and 494/18 nm filters and fluorescence emission was collected at 535/30 nm. Ca2+ binding causes an increase in excitation at 494 nm and decrease in excitation at 436 nm; thus, the ratio of fluorescence at 494/436 is directly proportional to mitochondrial [Ca2+]. All experimental groups were evaluated on each day. Because of day-to-day variability in measurements, each ratio was normalized to the mean of all ratios collected on the same day to combine data collected in multiple experiments.

Mitochondrial membrane potential (Δψ) was measured after 30-min incubation with 100 nM tetramethylrhodamine methylester (TMRM; InVitrogen Life Sciences) as described previously (Tradewell et al. 2011).

Cytosolic Ca2+

Cytosolic [Ca2+] was measured in motor neurons microinjected with plasmid encoding SOD1WT-eGFP or SOD1G93A-eGFP using the ratiometric indicator, fura-2 (InVitrogen Life Sciences; used at 10 μM) as previously described (Tradewell et al. 2011). eGFP epifluorescence was used to identify neurons expressing constructs for calcium imaging. Fura-2 was excited using 340/10 nm and 380/10 nm filters alternately and emission through a long pass 510 nm filter was captured by Metafluor®. Cytosolic [Ca2+] was determined as the ratio of fluorescence at 340/380 nm excitation. Data for each experimental group were collected using the same parameters.

Quantitation of neurons with inclusions

Motor neurons were microinjected with plasmid encoding untagged SOD1G93A and cultures were treated with vehicle or lomerizine. Three days later, cultures were fixed and immunolabeled with antibody recognizing human, but not mouse, SOD1 (SD-G6). The percentage of motor neurons with aggregated (vs. diffuse) SOD1G93A was calculated in three cultures per condition and expressed as mean ± SEM. Note, SOD1WT does not form these inclusions (Durham et al. 1997) and therefore was not evaluated in this experiment.

Motor neuron viability

Viability of motor neurons was evaluated at days 1–5 inclusive, following microinjection of expression plasmids plus 70 kDa dextran-FITC. The number of neurons containing the marker was counted by epifluorescence microscopy and morphology was evaluated by phase microscopy. Counts were normalized to the number of neurons on day 1 as previously described (Roy et al. 1998) and graphed as mean ± SEM. A minimum of three cultures per condition was evaluated and the experiment was repeated using at least one additional culture batch.

Statistical analysis

For comparison of two groups, significance of difference between group means was assessed by unpaired t-test. For more than two experimental groups, one-way or two-way anova followed by Tukey's Honestly Significant Difference post hoc analysis was performed as indicated in figure legends. Statistical significance was established at p < 0.05. On figures, only significant differences important to the point of the experiment are illustrated. For viability data, significance of differences between means was calculated using SPSS (IBM, Armonk, NY, USA) Life Table, with significance established at p < 0.05.

Results

The VGCC inhibitor, lomerizine, attenuated glutamate excitotoxicity in cultured motor neurons

To demonstrate the effectiveness of lomerizine in limiting intracellular [Ca2+], its ability to inhibit glutamate-induced death of motor neurons and the associated rise in cytosolic [Ca2+] was evaluated. The concentration was extrapolated from previous studies. Lomerizine inhibited the low- and high-voltage activated Ca2+ currents in dissociated rat brain neurons at a threshold concentration of 0.01 μM and IC50 of 1.9 μM (Akaike et al. 1993) and H2O2-induced Ca2+ influx in hippocampal neurons was inhibited by 1 μM lomerizine (Ishii et al. 2011). Pre-treatment with 1 μM lomerizine significantly reduced acute death of motor neurons in spinal cord-DRG cultures exposed to 50 μM glutamate, a concentration that killed approximately 40% of motor neurons in the culture by 6 h (Fig. 1a), and inhibited the rise in cytosolic [Ca2+] that occurred with glutamate treatment (Fig. 1b).

Figure 1.

Lomerizine protected motor neurons in murine spinal cord–dorsal root ganglia (DRG) cultures from glutamate excitotoxicity, prolonging viability and attenuating glutamate-induced rise in intracellular [Ca2+]. (a) Motor neurons were marked for counting by intranuclear microinjection of dextran-FITC, and then treated with either vehicle [0.0025% dimethyl sulfoxide (DMSO)] or 1 μM lomerizine. Neurons retaining the marker were counted by epifluorescence microscopy prior to exposure of cultures to 50 μM glutamate, and at 3 and 6 h thereafter. Counts were expressed as percent of the number present before glutamate treatment. Shown are means ± SEM for data collected from three cultures per condition, 83 and 180 neurons, respectively, in control and lomerizine-treated cultures (*p < 0.05 Two-way anova, Tukey Honestly Significant Difference (HSD) post hoc analysis). (b) Following addition of vehicle or lomerizine to spinal cord-DRG cultures, cells were loaded with the Ca2+ indicator, fura-2, then exposed to 50 μM glutamate. Fura-2 fluorescence at 340/380 nm was measured in the same neurons before and after addition of glutamate and expressed as percent change induced by glutamate (10 neurons in three cultures were analyzed per group). Lomerizine reduced the glutamate-induced increase in cytosolic [Ca2+] in motor neurons from 52% to 35% over baseline; n = 10 neurons per treatment. (*p < 0.05, t-test).

Lomerizine preserved mitochondrial morphology and axonal transport in motor neurons expressing SOD1G93A

Longer term experiments were carried out to evaluate the effect of lomerizine on mutant SOD1 toxicity using previously established endpoints and time points (Tradewell et al. 2011). 0.5 μM lomerizine was sufficient to significantly prevent the mitochondrial fragmentation (quantified as decreased length) of mitochondria induced by SOD1G93A (Fig. 2a,b), but did not alter mitochondrial length in neurons expressing SOD1WT, which does not affect mitochondrial morphology under these conditions (Tradewell et al. 2011)].

Figure 2.

Lomerizine prevented mitochondrial fragmentation and inhibition of axonal transport induced by Cu/Zn-superoxide dismutase (SOD1)G93A. (a) Mitochondrial length was measured from images of axonal segments of motor neurons expressing enhanced green fluorescent protein (eGFP) targeted to mitochondria (pOCT-eGFP). Shown are micrographs of mitochondrial eGFP epifluorescence in axons of motor neurons expressing SOD1WT or SOD1G93A for 3 days, treated with 0.5 μM lomerizine, or 0.0025% dimethyl sulfoxide (DMSO) as vehicle; 127–264 mitochondria per group. Scale bar = 10 μm. (b) Graphed are means ± SEM of mitochondrial length under the various experimental conditions. (c) Mitochondrial transport was evaluated in the same neurons by time-lapse vital imaging of mitochondrial eGFP. Plotted are the mean percentages of moving versus static mitochondria during an 8-min recording period. Two independent experiments were conducted, each with three cultures per condition; 168–325 mitochondria per group (*significantly different at p < 0.05; two-way anova, Tukey Honestly Significant Difference (HSD) post hoc analysis).

Mutant SOD1 impairs axonal transport of mitochondria and alters their distribution (De Vos et al. 2007; Magrane et al. 2009, 2012; Vande Velde et al. 2011). Mitochondrial motility was monitored in cultured motor neurons expressing pOCT-eGFP by time-lapse imaging of axons. Three days following microinjection of expression vectors, the percentage of moving mitochondria was significantly reduced in neurons expressing SOD1G93A relative to neurons expressing SOD1WT. Lomerizine maintained mitochondrial motility in motor neurons expressing SOD1G93A (Fig. 2c) without affecting the normal mitochondrial motility in neurons expressing SOD1WT.

Lomerizine prevented dysregulation of calcium homeostasis by SOD1G93A

Ca2+ regulates mitochondrial transport through the Miro-Milton complex that links kinesin motors to mitochondria (Liu and Hajnoczky 2009). Since calcium dysregulation is a feature of the SOD1G93A phenotype, the effect of lomerizine on Ca2+ accumulation in mitochondria of motor neurons was evaluated using a genetically encoded, Ca2+-sensitive pericam. Treatment with lomerizine prevented the increase in mitochondrial [Ca2+] that occurred by 1 day following microinjection of plasmid encoding SOD1G93A-eGFP (Fig. 3a), as well as the decrease in Δψ (Fig. 3b).

Figure 3.

Lomerizine prevented elevation of mitochondrial and cytosolic Ca2+ in motor neurons expressing SOD1G93A. In motor neurons expressing SOD1G93A, treatment with 0.5 μM lomerizine prevented (a) the increase in mitochondrial Ca2+, measured using a ratiometric pericam (494/436 fluorescence ratio), (b) the reduction in Δψ, measured using TMRM (496/436 fluorescence ratio), and (c) elevation of cytosolic [Ca2+], measured using fura-2 (340/380 fluorescence ratio). Lomerizine had no significant effect on these parameters in neurons expressing SOD1WT. Plotted are means ± SEM of measurements taken 1 day following microinjection of plasmid vectors, from neurons in three different culture batches, 44–57 neurons per treatment group (*significantly different at p < 0.05, two-way anova, Tukey post hoc analysis).

A major function of mitochondria is to buffer intracellular Ca2+ to maintain cytosolic calcium homeostasis. In motor neurons expressing SOD1G93A-eGFP, increased cytosolic [Ca2+] followed accumulation of Ca2+ in mitochondria (Tradewell et al. 2011). As with mitochondrial [Ca2+], treatment with 0.5 μM lomerizine was highly preventative, maintaining cytosolic [Ca2+] comparable to levels in neurons expressing SOD1WT-eGFP (Fig. 3c). No significant protection was observed with 0.1 or 0.25 μM lomerizine (data not shown).

Lomerizine both prolonged viability and reduced formation of inclusions in motor neurons expressing SOD1G93A

Wild-type SOD1 is diffusely distributed in motor neurons (Fig. 4ai), whereas SOD1G93A forms aggregates in a percentage of neurons (Fig. 4aii) and causes motor neuron death (Fig. 4c) (Durham et al. 1997; Roy et al. 1998; Batulan et al. 2006; Tradewell et al. 2011). SOD1WT does not affect motor neuron viability relative to injection control, nor does it form inclusions (Durham et al. 1997). Treatment with lomerizine reduced the percentage of motor neurons containing inclusions on day 3 (Fig. 4b) and increased survival of SOD1G93A-expressing motor neurons to levels comparable to those expressing SOD1WT (Fig. 4c).

Figure 4.

Lomerizine decreased formation of inclusions in neurons expressing SOD1G93A and prolonged their viability. Human SOD1WT or SOD1G93A were expressed in motor neurons by intranuclear microinjection of plasmid. Dextran-FITC was included as a marker of injected neurons. (a) Plasmid-derived protein was detected by immunocytochemistry with antibody recognizing human, but not endogenous murine, SOD1. SOD1WT remains diffusely distributed (i), whereas a proportion of motor neurons expressing mutant protein contain SOD1-immunolabeled inclusions (ii). (b) Graphed is the mean percentage of SOD1G93A-expressing neurons containing inclusions (± SEM) in cultures treated with 0.5 μM lomerizine, 3 days following plasmid microinjection. (c) Increased survival of SOD1G93A-expressing neurons was as previously reported (see text). Treatment with lomerizine prolonged viability of motor neurons expressing SOD1G93A, but had no significant effect on viability of neurons expressing SOD1WT. Presented are means ± SEM. of results from three cultures per condition, 20–60 motor neurons per culture (*significantly different at p < 0.05, t-test for (b) and two-way anova, Tukey Honestly Significant Difference (HSD) post hoc analysis for (c)).

In summary, lomerizine was effective in preventing multiple measures of SOD1G93A in motor neurons, consistent with calcium dysregulation being a significant contributor to toxicity.

Cytosolic [Ca2+] was not elevated in motor neurons expressing mutant FUS or TDP-43

Many promising treatments identified in preclinical studies using mutant SOD1 transgenic mice have failed to translate into efficacy in human trials, leading to criticism of the ALS1 model for relevance to other forms of ALS, including sporadic disease. Thus, to investigate if lomerizine might be effective against other forms of ALS, experiments were carried out in models of ALS6 and ALS10, because of mutations in FUS (Kwiatkowski et al. 2009; Vance et al. 2009) and TARDP (Arai et al. 2006; Neumann et al. 2006), respectively. Both encode RNA-binding proteins (FUS and TDP-43) that accumulate and aggregate in the cytoplasm of cultured motor neurons [see (Tradewell et al. 2012) and (Kabashi et al. 2010)].

The first experiment was to determine if dysregulation of cytosolic Ca2+ was a feature of mutant FUS or TDP-43 toxicity. Motor neurons were microinjected with plasmids encoding eGFP-tagged wild-type human TDP-43 or the ALS-causing mutants, G348C or A315T and wild-type human FUS or the ALS-causing mutants, R521H and P525L. Cytosolic [Ca2+] was measured using fura-2 on day 5, a time when cytoplasmic accumulation and aggregation of these proteins is even more profound than at day 3, but neurons remain viable (Kabashi et al. 2010; Tradewell et al. 2012). The 340/380 nm ratios of fura-2 fluorescence, proportional to [Ca2+], did not differ in neurons expressing wild-type TDP-43 or FUS compared to neurons injected with empty plasmid (Fig. 5a), wild-type TDP-43 compared to the G348C or A315T mutants (Fig. 5b), or wild-type FUS compared to the R521H or P525L mutants (Fig. 5c). Nor did cytoplasmic [Ca2+] differ according to whether FUS or TDP-43 was distributed predominantly in the nucleus or in the cytoplasm (see Figure S1).

Figure 5.

(a–c) Neither wild type nor mutant TAR DNA-binding protein 43 (TDP-43) or fused in sarcoma/translated in liposarcoma (FUS) altered cytosolic [Ca2+] in cultured motor neurons. Cytosolic Ca2+ was measured using fura-2 as an indicator of Ca2+ homeostasis in motor neurons expressing enhanced green fluorescent protein (eGFP)-tagged human wild-type TDP-43 or FUS or with mutations causing familial amyotrophic lateral sclerosis (ALS). Cytosolic [Ca2+] was measured 5 days following microinjection of plasmid encoding (a) FUSWT or TDP-43WT compared to empty plasmid; (b) TDP-43WT, TDP-43A315T or TDP-43G348C, or (c) FUSWT, FUSR521H or FUSP525L. Plotted are mean ± SEM of 340/380 nm ratios of fura-2 fluorescence from three culture batches (26–36 neurons per condition). There was no significant difference in relative cytosolic [Ca2+] among these groups (p > 0.05, one-way anova with Tukey Honestly Significant Difference (HSD) post hoc analysis). (d) Neither WT nor mutant TDP-43 affected mitochondrial calcium levels. Mitochondrial [Ca2+] was measured using the mitochondrial pericam in motor neurons microinjected with empty plasmid or plasmid encoding FLAG-tagged TDP-43WT, TDP-43G348C, or TDP-43A315T. Shown are means ± SEM of 496/436 nm ratios of pericam fluorescence in 21–23 neurons per group. (no significant differences among groups at p < 0.05, one-way anova with Tukey HSD post hoc analysis).

Whereas TDP-43 or FUS (either WT or mutants) failed to significantly alter cytosolic [Ca2+] in motor neurons, it is possible that calcium load was increased, but effectively buffered, as in the initial stages of mutant SOD1 toxicity. We therefore measured mitochondrial [Ca2+] in motor neurons expressing WT, G348C or A315T TDP-43, but found no significant differences (Fig. 5d).

Lomerizine treatment failed to prevent redistribution and cytoplasmic accumulation of mutant TDP-43

Although Ca2+ homeostasis seemed to be intact in motor neurons expressing mutant TDP-43 or FUS, experiments were conducted to determine if lomerizine would have an impact on toxicity, using accumulation of TDP-43 in the cytoplasm as the endpoint. Cultured motor neurons were injected with TDP-43WT-eGFP, TDP-43G348C-eGFP or TDP-43A315T-eGFP encoding plasmids, treated with 0.5 μM lomerizine or vehicle, and on day 5 post-microinjection, counted and classified as having either predominantly nuclear or cytosolic TDP-43-eGFP localization (Fig. 6a). TDP-43 was predominantly cytoplasmic in a significantly greater percentage of motor neurons expressing the mutants relative to TDP-43WT, but the proportion was not significantly affected by treatment with lomerizine (Fig. 6b). Given that lomerizine did not affect localization of TDP-43, experiments were not conducted with FUS.

Figure 6.

Subcellular localization of TAR DNA-binding protein 43 (TDP-43) is not affected by lomerizine treatment. Redistribution of mutant TDP-43 to the cytoplasm is an established pathological hallmark, and was used as a measure of toxicity in this experiment. Following microinjection of plasmid encoding enhanced green fluorescent protein (eGFP)-tagged TDP-43WT, TDP-43A315T, or TDP-43G348C, expression was evaluated over 5 days by eGFP epifluorescence microscopy, following which cultures were fixed and immunolabeled with antibody to microtubule-associated protein 2 to visualize the perikarya and dendrites of motor neurons and location of the nucleus. EGFP epifluorescence was rated as mainly cytosolic or mainly nuclear (as shown in Fig. 6a). Distribution on days 3 and 5 are shown in Fig. 6b. More neurons had cytoplasmic mutant TDP-43 compared to wild type, as expected, but treatment with lomerizine had no significant effect on distribution. Plotted are mean percentages ± SEM of motor neurons with predominantly cytosolic eGFP-TDP-43 localization; data were collected from three cultures, 18–50 cells per condition (*significantly different from WT at p < 0.01, two-way anova, Tukey Honestly Significant Difference (HSD) post hoc analysis).

Discussion

Primary culture models of familial ALS were used to further investigate the relevance of calcium dysregulation in pathogenesis and the utility of VGCC inhibition to mitigate calcium dysregulation and toxicity. Plasmids encoding mutant proteins linked to familial ALS or their wild-type counterparts were expressed in motor neurons of long-term dissociated cultures of murine spinal cord-DRG to evaluate toxicity in the context of the unique physiological properties of motor neurons that contribute to their vulnerability. Relevant properties are expression of glutamate receptors (Williams et al. 1997; Carriedo et al. 1998; Vandenberghe et al. 2000; Kawahara et al. 2003; Tateno et al. 2004; Corona and Tapia 2007) and low levels of the cytosolic Ca2+ buffering proteins, calbindin, and parvalbumin (Ince et al. 1993; Alexianu et al. 1994; Elliott and Snider 1995; Junttila et al. 1995; Reiner et al. 1995; Lips and Keller 1998).

As pointed out in the Introduction, excitotoxic mechanisms are implicated in the pathogenesis of ALS, and studied most in models of in ALS1 because of SOD1 mutation. Motor neurons cultured from mutant SOD1 transgenic mouse embryos show increased excitability of action potential production (Kuo et al. 2003); increase in persistent sodium current (Kuo et al. 2005); increased conductance of AMPA receptors (Pieri et al. 2003), and enhanced sensitivity to glutamate (Kruman et al. 1999), all of which would increase Ca2+ influx. The initiating mechanism remains unclear, although a possibility is formation of pore-like structures in membranes. SOD1A4V formed tetrameric pore-like, ion-conducting structures in reconstituted in lipid membrane, as well as depolarizing the membrane potential of mouse neuroblastoma cells and increasing intracellular [Ca2+] when added to the culture medium (Allen et al. 2012).

Even low level, chronic stimulation of glutamate receptors promotes toxicity of mutant SOD1 in motor neurons through Ca2+ -mediated mechanisms. Blocking glutamate receptors, specifically Ca2+ -permeable AMPA receptors, or over-expressing the calcium-binding protein, calbindin, or calpastatin prevented formation of inclusions and cell death induced by mutant SOD1 in cultured motor neurons (Roy et al. 1998; Tradewell and Durham 2010). In this study, we report that the VGCC inhibitor, lomerizine, was highly effective in preventing not only Ca2+ dysregulation, but mitochondrial abnormalities, formation of inclusions, and motor neuron death induced by expression of SOD1G93A.

Toxicity of mutant SOD1 is complex, involving multiple intracellular pathways even though caused by a single missense mutation. Considerable evidence points to mitochondria being an important target (De Vos et al. 2007; Magrane et al. 2009, 2012; Vande Velde et al. 2011). Ca2+ regulation within mitochondria is disrupted, as is their transport. Ca2+ regulates the association of mitochondria with kinesin motors through the Miro-Milton complex for transport; increased Ca2+ inhibits this association and therefore, microtubule-based mitochondrial transport, making this pathway a likely target (Liu and Hajnoczky 2009). Lomerizine maintained mitochondrial transport in motor neurons expressing SOD1G93A, likely by limiting Ca2+ influx to levels buffering mechanisms could handle. The preservation of mitochondrial morphology by lomerizine (prevention of fragmentation) infers maintenance of the balance of fission/fusion, which on its own has been shown to be neuroprotective in experimental models of excitotoxicity (Cheung et al. 2007). Elevated Ca2+ is known to promote mitochondrial stress and reactive oxygen species generation, factors that could contribute to mitochondrial fragmentation. Interestingly isradipine, an L-type VGCC inhibitor, prevented mitochondrial reactive oxygen species generation associated with Ca2+ entry during pacemaker activity in substantia nigral pars compacta dopaminergic neurons vulnerable to Parkinson's disease (Guzman et al. 2010; Ilijic et al. 2011) and is in early phase clinical trials for Parkinson's (Simuni et al. 2010).

Mutations promote misfolding of SOD1 which is associated with aggregation and inappropriate interactions with cellular constituents including mitochondrial membrane proteins (Vande Velde et al. 2008, 2011; Israelson et al. 2010). In our previous study, cultured motor neurons containing mutant SOD1 inclusions had significantly higher cytosolic [Ca2+] than those with mutant SOD1 distributed diffusely (Tradewell et al. 2011). That lomerizine reduced formation of inclusions in motor neurons expressing SOD1G93A also points the interplay of calcium dysregulation with protein misfolding and aggregation.

Pathogenesis of ALS is complex. This is a heterogeneous disease and mechanisms are likely to differ between ALS1 and other forms. Even when caused by a missense mutation, the mutant protein can disrupt multiple functions of motor neurons and interactions with other cells. Thus, it may not be surprising that any one drug is yet to have a marked impact on the disease in patients, including riluzole, which is approved for treatment of ALS (Traynor et al. 2006; Aggarwal and Cudkowicz 2008). The VGCC inhibitors, nimodipine and verapamil, were previously tested in small pilot clinical trials with sporadic ALS patients, but failed to show significant benefit (Miller et al. 1996a,b). One possibility is that this class of drugs is not sufficient to target excitotoxicity, reduce calcium load or promote protein folding at clinical doses, but might show synergistic neuroprotective effects when used in combination with other agents. Another possibility is that Ca2+ dysregulation is less important in the pathogenesis of non-SOD1 ALS. To test the latter, we measured cytosolic Ca2+ in motor neurons expressing ALS-linked mutants of the RNA-binding proteins, FUS and TDP-43. These mutant proteins accumulated in the cytoplasm and aggregated, a hallmark of their toxicity [for literature review and establishment of the culture models see (Tradewell et al. 2012) and (Kabashi et al. 2010)]; however, no elevation of cytosolic [Ca2+] was detected in comparison to neurons expressing wild-type protein, suggesting that Ca2+ dyshomeostasis is not a major component early in the pathogenic cascade or that the normal buffering mechanisms are sufficient. The lack of elevation in mitochondrial [Ca2+] in neurons expressing WT or mutant TDP-43 argues against increased calcium load. Within that context, it is not surprising that lomerizine failed to affect mislocalization of mutant of TDP-43. Furthermore, the Ca2 + -blocking AMPA receptor antagonist, 6-cyano-7-nitroquinoxaline-2,3-dione had no significant effect on mislocalization of mutant FUS in cultured motor neurons or the formation of cytoplasmic inclusions (Tibshirani, Gupta and Durham, unpublished data), whereas this treatment was highly protective against mutant SOD1 toxicity (Roy et al. 1998). Toxicity of FUS and TDP-43 mutants also differs from SOD1 mutants in other ways. Any disruption of mitochondrial morphology or loss of viability is much less dramatic in cultured motor neurons over-expressing mutant FUS or TDP-43 compared to mutant SOD1 (Kabashi et al. 2010; Tradewell et al. 2012).

In conclusion, lomerizine preserved motor neuron viability and mitigated elevated intracellular Ca2+ associated with either glutamate excitotoxicity or expression of an ALS-linked SOD1 mutant; however, Ca2+ overload was not a major feature in models of ALS6 or ALS10, because of mutations in FUS and TDP-43, respectively, and lomerizine treatment failed to prevent cytoplasmic accumulation and aggregation of TDP-43. For this reason, lomerizine might not be an effective treatment across the entire ALS spectrum, but might be useful in ALS1. This study is an example of how primary culture models of multiple forms of familial ALS can be used to identify commonalities and differences in the pathogenic pathways of different forms of ALS and effectiveness of candidate therapies.

Acknowledgments and conflict of interest disclosure

This work was supported by the Canadian Institutes for Health Research (MOP-77743 to HDD) and the Muscular Dystrophy Association (MDA93897 to HDD). The authors thank Sandra Minotti for the spinal cord-DRG cultures, Miranda Tradewell for training in microfluorometric imaging, Laura Cooper for technical assistance, Dr. Heidi McBride for the pOCT-eGFP plasmid, and Dr. Atsushi Miyawaki for the mitochondrial pericam plasmid. This work was included in the M.Sc. thesis of Luan Tran. The authors have no conflict of interest to declare.

Ancillary