Adult onset amyotrophic lateral sclerosis (ALS) arises due to progressive and irreversible functional deficits to the central nervous system, specifically the loss of motor neurons. Sporadic ALS causality is not well understood, but is almost certainly of multifactorial origin involving a combination of genetic and environmental factors. The discovery of endemic ALS in the native Chamorro population of Guam during the 1950s and the co-occurrence of Parkinsonism and dementia in some patients led to searches for environmental toxins that could be responsible. In the present paper, we report that an environmental neurotoxin enhances mutant superoxide dismutase (SOD)-induced spinal motor neuron death and pathology and induces motor axon abnormalities. These results cumulatively confirm earlier findings that exposure to an environmental toxin is sufficient to produce the disease phenotype and indicate a role for gene-environment interaction in some forms of the disease.
Amyotrophic lateral sclerosis (ALS) is a progressive paralytic disorder that arises due to degeneration of motor neurons in the brain and spinal cord that invariably leads to respiratory failure and death. About 10% of all ALS cases are familial (FALS), inherited in an autosomal dominant manner, and approximately 20% of FALS cases are caused by mutations in the antioxidant enzyme, superoxide dismutase 1 (SOD1) [
While studies that focus on putative environmental toxins are relatively few, to date those that seek to find the link between toxins and genetic susceptibility are rarer [
Studies of the ALS disease cluster among the Chamorro people of Guam shortly after World War II provided evidence of environmental causation [11-13]. Similar to classical ALS, these Guamanian cases usually showed signs of upper and lower motor neuron losses developing invariably during the course of the illness while extraocular muscle weakness and objective sensory deficits were largely absent [14,15]. An atypical form of parkinsonism that presented with dementia emerged also with high incidence, sometimes overlapping with ALS in the same patient. All cases showed a wide distribution of Alzheimer’s-like neurofibrillary tangles in the absence of senile plaques [
Similar data has been reported for SOD1G93A mice without cycad exposure, where the primary difference involved decreases in spinal cord white matter volume in transgenic animals [
Although the data supporting neurotoxins in ALS, including the impact of SGs is intriguing, it is uncertain whether these or other toxins are alone responsible for the pathological outcomes or whether they may work in conjunction with genetic susceptibility factors. The idea of genetic and environmental interplay has been an established focus for numerous studies, and is not limited to neurodegenerative diseases [31,32]. Based on the above studies using cycad seed flour or the derived toxins, we tested the hypothesis that similar toxins could modify the onset and progression of disease symptoms in SOD1 transgenic mice. To explore this hypothesis, the current study distinguished between the pathological consequences of either genetic predisposition to neurodegeneration or neurotoxin acting alone on motor axons, and the site of primary pathogenesis when both stressors act in combination.
A colony of mice heterozygous for the G37R mutation (line 29) of the human gene for SOD1, were obtained from the laboratory of Dr. Neil Cashman (Vancouver, BC, Canada). The original breeder mice were produced by Dr. D.W. Cleveland by microinjecting a plasmid encoding wild type human SOD1 contained within 12 kb genomic DNA fragment into hybrid (C57BL/6J x C3H/- HeJ) F2 mouse embryos. The G37R mutation was introduced into the human SOD1 gene by PCR using a mutagenic primer. Transgenic mice from this founder line (line 29) express a (7-fold) increase in SOD1 activity in spinal cord, with pathology restricted to motor neurons in the spinal cord and brainstem [
Animals were observed daily for development of tremor, weakness, and loss of weight. At regular 3 week intervals, starting at 4 weeks of age when no signs of disease symptoms were present and up to 51 weeks of age, animals were monitored for onset of disease phenotype. Motor performance was monitored by Rotarod testing using an IITC Rotarod (IITC Life Science Inc., Woodland Hills, CA; #755). Mice were placed on a partitioned rotating rod 1.25 inch in diameter and tested at a constant speed of 30 RPM. Animals fell onto a platform that is sensed by magnetic switches and the time that the animal remains on the rod (up to a maximum of 180 seconds) is displayed and recorded. Rotarod testing was performed on the same day as grip endurance testing.
For immunocytochemistry and other cell labelling, we processed CNS tissue as we previously described [
We counted all motor neurons and measured gliosis intensities in every tenth transverse section of lumbar cord between L3 and L5. Soma diameters were measured by calculating the average length of two lines spanning the diameter of the cell and intersecting like crosshairs with the intersection point on the center of the nucleus. Data are presented as means ± standard error of the means (S.E.M.) as indicated. All results were analyzed to determine the effect of SG alone, G37R mutation alone, or the interaction between these two variables by ANOVA via Statistica 8.0 statistical software (StatSoft Inc., Tulsa, OK). Statistical significance between groups on rotarod performance was analyzed using repeated measures ANOVA via Statistica 8.0 statistical software (StatSoft Inc., Tulsa, OK) and graphed using Prism 5 (GraphPad Software Inc., San Diego, CA). Statistical significance between experimental and control groups for all other behavioral and histological assessments were calculated using a Student’s t-test. The t-test was used when only two groups and one condition were compared, as in the case of the quantitative histological data. Distributions of Nissl stained soma diameters and toluidine blue stained ventral roots were tested for normality using a one-sample Kolmogorov-Smirnov test, and differences between groups for sex and cell morphology were detected using one-way ANOVA. Following that, a post hoc test using multiple comparison Tukey HSD test was used to find out which groups were significantly different from controls. The analysis and graphs of distribution functions were computed by SPSS version 17.0 (New York, USA). Differences were considered statistically significant at p < 0.05. P values were expressed as exact values except in cases where p < 0.0001. A statistician reviewed the statistical content and methodology.
Motor coordination deficits on the rotarod began appearing in wild type mice with SG 2 weeks after beginning SG treatment in males (
645.81, p < 0.0001). Similar to the male counterparts, an interaction between SG and SOD1G37R genotype was also observed in females (F = 1699.02, p < 0.0001). Wild type mice without SG (controls) showed a similar level of rotarod performance throughout the study, as demonstrated by a minimal change in performance with increasing age.
No significant differences in ChAT-positive cell counts were found in SG-fed wild types (p = 0.12 for males, p = 0.09 for females), although the cell count was noticeably lower in this group for both sexes (26% decrease compared to controls;
(F = 1.7, DF = 40, p = 0.15). In comparison with wild type controls, significant decreases in ChAT-positive cell count were found in G37R mice (p = 0.019 for males, p = 0.0017 for females). Significant main effects were observed for genotype in both males (F = 1.2, DF = 34, p = 0.002) and females (F = 1.04, DF = 40, p = 0.0002). For SG-fed G37R mice the decrease in ChAT-positive cells was more pronounced (p = 0.0034 for males, p = 0.0008 for females). Analysis of gene × diet interaction showed no significance in males (F = 0.27, DF = 34, p = 0.61) and females (F = 0.01, DF = 1, p = 0.91). However, when all the animals from both sexes were combined, significant main effects of both diet (p = 0.007) and genotype (p < 0.0001) were found. In contrast, no significant diet × genotype interaction was observed (p = 0.66). The appearances and numbers of motor neurons were not significantly different between the left and right ventral horns (data not shown).
All groups exhibited morphologically abnormal neurons that appeared to be undergoing degeneration among a larger population of healthy cells (
SG-fed wild type (Tukey HSD post hoc test; p < 0.001), control-fed G37R (Tukey HSD post hoc test; p = 0.001), and SG-fed G37R (Tukey HSD post hoc test; p < 0.001) female mice indicates a decline in soma diameters of neurons with an apparently normal morphology (
Astrocyte gliosis was determined by quantifying the immunohistochemical labeling intensity of green fluorescent GFAP positive cells in the ventral horn of the lumbar spinal cord in conjunction with blue Hoechst staining (
throughout the entire grey matter (Figures 5(g) and (l)). Furthermore, control-fed G37R mouse cord showed astrocytes that were exclusively in the activated state with enlarged cell bodies and thickened, retracted processes (Figures 5(j) and (k)) or astrocytes that exclusively showed increased branching of long thin processes (Figures 5(h) and (i)). Treatment with SG in G37R animals did not increase astrocyte proliferation further in males (student’s t-test; p = 0.48), but significantly decreased GFAP label intensity in females (student’s t-test; p = 0.017, compare Figures 5(h)-(j) with m). All GFAPpositive cells in this group were astrocytes with an activated morphology (Figures 5(m) and (n)).
Significant main effects of genotype on increased GFAP green fluorescence intensity was observed in males (one-way ANOVA; F = 3.69, DF = 222, p << 0.0001) and females (one-way ANOVA; F = 2.49, DF = 221, p << 0.0001). Interaction between diet and genotype was significant for both males (one-way ANOVA; p = 0.007) and females (one-way ANOVA; p = 0.001). In summary, SG-fed mice showed significant increases in GFAP label intensity in the lumbar cord grey matter (males p < 0.0001; females p < 0.05) compared to non SG-fed wild type mice with astrocytes in the resting state. Mutant mice showed even greater significant GFAP infiltration (males p < 0.0001; females p < 0.0001) compared to wild type mice with astrocytes either exclusively in an altered resting morphology or an active morphology. The combination of SG with the mSOD mutation did not change astrocyte numbers, but induced active astrocyte morphology in all animals.
Microglial proliferation was described by quantifying
the immunohistochemical labeling intensity of red fluorescent Iba1 positive cells in the ventral horn of the lumbar cord in conjunction with blue Hoechst staining (
Significant main effects of genotype on increased Iba1 red fluorescence intensity was observed in males (oneway ANOVA; F = 3.4, DF = 220, p << 0.0001) and females (one-way ANOVA; F = 2.75, DF = 220, p << 0.0001). Interaction between diet and genotype was significant for both males (one-way ANOVA; p = 0.008) and females (one-way ANOVA; p = 0.002). In summary, SG-fed mice no changes in Iba1 label intensity in the lumbar cord grey matter (males p = 0.35; females p = 0.51) compared to non SG-fed wild type mice with microglia in the resting state. Mutant mice showed significant microglial infiltration (males p << 0.0001; females p << 0.0001) compared to wild type mice with microglia in a transition state between resting and active morphology, but most with an active morphology. The combination of SG with the mSOD mutation did not change microglia numbers in females (p = 0.28), but significantly increased microgliosis in males (p = 0.051) while active astrocyte morphology was observed in all animals.
Evaluations of muscle end plates with rhodamine conjugated α-bungarotoxin immunofluorescence showed significant abnormalities in all G37R groups of both sexes
(
The axon diameter distribution histogram showed a normal distribution of toluidine blue stained motor axon diameters for the controls (Kolmogorov-Smirnov test, p > 0.01) while distribution of axonal diameters for other groups did not show a normal distribution (Kolmogorov-Smirnov test, p < 0.01). The frequency distributions of axon diameters were different between wild type and G37R mice (
the controls. Motor axons from SG-fed wild types did not show any changes in diameter compared to wild type controls. Both G37R groups showed a similar wide distribution range, although skewed toward smaller axon diameters compared to wild types.
The reduced proportion of large diameter motor axons in the lumbar ventral root of both G37R groups and the consequent significant reduction in axon diameter range are further illustrated in the cumulative distribution functions (
The current results provide some of the first evidence in relation to ALS that a gene-toxin interaction may be part of interlocking etiologies for the disease. In the present case, a toxin was imposed on top of a genetic mutation that already can induce a profound ALS phenotype. In spite of this, toxin exposure exacerbated the mutant gene toxic gain of function impact by adding to and accelerating the series of events leading to motor neuron death. Given that over 90% of all ALS cases are sporadic, it may be reasonable to postulate that the opposite also occurs, namely that primarily toxin etiologies act upon genetic susceptibility factors to induce the disease state. In this view, sporadic cases might arise in some fraction f the population containing such susceptibility genes in context to a relatively ubiquitous toxic burden.
The present data suggest that a determined search for such susceptibility genes might be of great benefit for pre clinical screening of potential ALS patients.
The results of the present study demonstrate the contribution of an identified neurotoxin linked to some forms of ALS to the progression of motor deficits and neurodegenerative processes of a mouse model of ALS. Chronic dietary administration of this toxic SG also induced significant reductions in motor function during behavioral tests as reported in previous studies [21,30]. These studies showed that motor coordination and grip endurance were impaired from 8 weeks of chronic SG exposure in males and from 18 and 38 weeks, respectively, in females. In addition, hindlimb extension showed deficits by 12 weeks of SG feeding in males, and 26 weeks in females. There was a non-significant trend towards motor neuron loss and increased gliosis in the lumbar ventral horn at disease end-stage, and there was significant gene × diet interaction effect on gliosis. At disease end stage, when degenerative changes in motor neurons and cell death are well underway, marked synaptic loss in motor endplates and reduction in axon caliber occur in mutant mice. Chronic treatment with SG had no effect on furthering these degenerative changes in peripheral motor components in G37R mice. Although in vitro studies have shown that synthetic cycad steryls act directly on neural cells [18,20], SG appears to be primarily involved in inflammatory reactions and motor neuron pathology in the spinal cord. The fact that SG-mediated neuropathology was only noted in the spinal cord where neuronal cell death mainly occurs indicates that it could contribute to cell dysfunction leading to neurodegeneration in SOD1 mutant mice. Thus, our data suggest that environmental toxins initiate neural pathogenesis by triggering inflammation-mediated neurodegeneration with an additive effect on the processes triggered by mutant SOD1.
Our results are in agreement with our earlier behavioral [21,23] and histological [24,34] studies on the effects of chronic cycad exposure on motor impairments in mice. Data have also been reported for synthetic cycad toxins, where significant motor dysfunction is evident following 30 weeks of chronic treatment, and the onset of motor neuron death was observed from 15 weeks [
The results of this study showed significant leftward shifts in the cumulative histogram of soma diameter distribution in non-SG-fed and SG-fed G37R cords of both sexes. Studies in G93A mice reported preferential loss of the most forceful motor units paralleled by denervation of large type IIB (fast-fatigable) muscle fibers and indicate conversion to motor units innervated by smaller motor axons during the presymptomatic stage [
Prolonged dietary exposure to SG elicited a significant astrocytic response in the ventral lumbar horn of both sexes that was more pronounced in males than females (58% increase in males compared to 20% in females). This finding is consistent with previous studies of toxin administration to male mice [23,30], and provides novel data for glial responses in females. G37R expression elicited a prominent microglial and astroglial response, some of which showed morphological changes reminiscent of activation. Prominent hypertrophic astrocytes and irregular amoeboid shaped microglia are predominantly located in the anterior horn of symptomatic G37R mice [39,40]. Enhanced gliosis as measured by increased microglial cell numbers and GFAP staining intensity was observed in other studies of this line of symptomatic G37R mouse [
Synaptic loss at motor end plates of the gastrocnemius muscle of mutant SOD1 mice was consistent with the remarkable decline in motor function and paralysis of the hind limbs. Numerous other studies show early vulnerability of fast-type neuromuscular synapses and functional motor unit losses [
Another major finding demonstrated here is a significant reduction in motor axon caliber in G37R mice that is unaffected by chronic SG dietary administration. The numbers of large caliber axons is significantly lower in G37R ventral roots compared with wild types, while the number of small caliber axons is increased. This finding is consistent with increased numbers of small axon fibers and greater frequency of axonal degeneration in the ventral spinal root of ALS patients [
We wish to thank Drs. A. Eisen, I. Mackenzie, and C. Seow for their valuable comments and advice. We are grateful to Dr. N. Cashman for the SOD1G37R breeders and Dr. H. Stewart for advice with animal breeding and husbandry. The authors wish to thank D. Sommerfeld and B. Hilton for help with behavioural monitoring and data collecting. Special thanks to A. Kam for his support and encouragement. We acknowledge financial support from the National Institutes of Health (NIH5R01N6SO51723), ALS Association, Canadian Institutes of Health Research, and Scottish Rite Charitable Foundation of Canada (G.L.).