Background

Amyotrophic lateral sclerosis (ALS) is a severe degenerative disorder, mainly affecting the motor neurons in the spinal cord, brainstem, and cortex. Most of the cases (about 90 %) are sporadic. Familial cases have been linked to mutations in a number of genes, including Cu/Zn superoxide dismutase (SOD1), TAR DNA binding protein (TDP-43), and chromosome 9 open reading frame 72 (C9ORF72) repeat expansions [1]. Mutant SOD1 is the best characterized form of familial ALS, accounting for 15–20 % of familial cases [2]. The etiology of sporadic ALS is unknown, although it is generally believed that sporadic and familial ALS may share pathological mechanisms. The pathophysiology of the multifactorial-multisystemic ALS disease includes various mechanisms. Among them are oxidative damage, high levels of glutamate, a reduced secretion of neurotrophic factors, protein aggregations, malfunctioning of the mitochondria, rupture in the axonal passage, destruction in the calcium metabolism, and changes in the skeletal proteins [3]. Although ALS is not primarily considered an inflammatory or immune-mediated disease, immune mechanisms appear to play a role in the pathogenesis of the disease. In both ALS patients and animal models, inflammatory responses have been observed [3–5]. Microglia [26].

End-stage analysis

To determine the mortality in a reliable and human fashion, end stage was defined as the inability of the mice to right themselves 30 s after being placed on one of their sides.

Statistical analysis

Data were expressed as mean ± standard error of the mean (SEM). Significance was determined by either one- or two-way analysis of variance (ANOVA) followed by a posteriori Bonferroni’s test for multiple comparisons provided by GraphPad Prism version 5.00 for Windows (GraphPad Software, San Diego, CA, USA). Survival was evaluated by Kaplan-Meier analysis.

Results

In order to study the role of cPLA2α in the development of ALS, cPLA2α protein expression was analyzed in the spinal cord of hmSOD1 mice (Fig. 1a) at different stages during their life span (Fig. 1b). Immunofluorescence staining and quantitation showed a significant (p < 0.001) elevation of cPLA2α protein expression in the spinal cord (Fig. 1a) of 6-week-old hmSOD1 mice, before the loss of motor neuronal function that appeared around week 17 (Fig. 1c). cPLA2α upregulation was also detected in the mice brainstem from week 6 onward and showed similar results (not shown). Neuronal or glial cells were stained by specific markers in the spinal cord of hmSOD1 mice at the same time points (Fig. 1d). Immunofluorescence staining and quantitation of neurons in the spinal cord using anti-NeuN showed a significant (p < 0.001) reduction in the neuron cell area from week 19 that was further reduced at the end stage. The neurons were counted according to their size and we found a similar reduction of around 40 % at 19 weeks and around 60 % at end stage in each group, suggesting that the reduction in cell area is due to both the number and the size of the neurons. A significant (p < 0.001) appearance of activated microglia (M1 phenotype) could be detected in the spinal cord of 15-week-old hmSOD1 mice by Iba-1 or CD40 staining that was gradually increased with age. The activation of microglia at week 15 preceded neuronal damage (Fig. 1d) and the appearance of motor neuron dysfunction (Fig. 1c). Activated astrocytes determined by GFAP were significantly (p < 0.001) detected in the spinal cord of 19-week-old mSOD1 mice and were dramatically increased at the end stage. Double staining with Iba-1 and anti-CD169 for the detection of monocytes revealed a significant (p < 0.001) appearance of monocytes in the spinal cord only at the end stage (Fig. 1e). As shown by confocal microscopy in Fig. 2a and in the inset Fig. 2b, double staining analysis of cPLA2α and the different cell types suggests that the main cells expressing cPLA2α protein are the neurons. High levels of cPLA2α protein were detected in the large neurons present in the spinal cord of mSOD1 mice at 6 and 11 weeks and in the small and damaged neurons at 19 weeks and end stage. The slight reduction of cPLA2α protein expression at the end stage (Fig. 1a) is probably due to the reduced number and size of the neurons (Fig. 1d) which are the main cells expressing high levels of cPLA2α. Confocal microscopy of higher magnification of the double staining analysis of cPLA2α and Iba-1 or GFAP showed that elevated cPLA2α could be detected in activated microglia and astrocytes (Fig. 2b). Confocal microscopy of double staining of the activated form of cPLA2α, phospho-cPLA2α on serine 505 (pcPLA2α), and the markers of the different cell types (Fig. 2c) showed that pcPLA2α was detected in the nuclei of the cells.

Fig. 1
figure 1

cPLA2α protein levels in the spinal cord of hmSOD1 mice during the development of the disease in hmSOD1 mice. a Representative cPLA2α protein expression in the spinal cord tissue sections along the hmSOD1 mouse life span. Scale bars = 250 μm for upper section and 50 μm for lower sections. The mean ± SEM of the cPLA2α fluorescence intensity is presented in the bar graph (n = 8 mice for each time point, five fields were analyzed for each mouse). *p < 0.001—significance from WT mice. b, c Development of the disease monitored by Kaplan-Meier survival curve and motor performance on accelerating rotarod test (n = 12 mice in each group). d Representative spinal cord sections of neurons (NeuN), microglia (Iba-1 and CD40), and astrocytes (GFAP) during the hmSOD1 mouse life span. The mean ± SEM of percentage of the cell area is presented in the bar graphs (n = 8 for each time point, five fields were analyzed for each mouse). Scale bars = 50 μm and in the inset 20 μm. *p < 0.001—significance from WT mice. e Representative staining of peripheral monocytes detected by double staining of Iba-1 and CD169, shown by arrows. Scale bars = 250 μm for left section and 50 μm for right section in each pair. Scale bars in the insets = 20 μm. The mean ± SEM of number of monocytes and of percentage of the cell area is presented in the bar graphs (n = 8 for each time point, five fields were analyzed for each mouse). *p < 0.001—significance in comparison to week 19

Full size image
Fig. 2
figure 2

Elevated cPLA2α protein expression in the different cell types of the spinal cord. a Representative pictures of immunofluorescence double staining of cPLA2α (green) with motor neurons (NeuN), microglia (Iba-1), or astrocytes (GFAP) in the spinal cord during hmSOD1 mouse life span (out of 8 mice in each time point, presented in Fig. 1a). DAPI staining for cell nuclei. Scale bars = 20 μm. b Insets of the immunofluorescence double staining of cPLA2α with motor neurons, microglia, or astrocytes (presented in a), in a higher magnification. Scale bars = 20 μm. c Immunofluorescence double staining of phospho-cPLA2α (p-cPLA2) and motor neurons (NeuN), microglia (Iba-1) and astrocytes (GFAP). Shown representative immunofluorescence confocal microscopy pictures (out of 8 mice in each group). Lower panel presents the cell staining and DAPI to show the cell nuclei. Scale bars = 20 μm

Full size image

To determine whether cPLA2α has a role in the induction of the disease, its expression was blunted in the brain and spinal cord by means of specific oligonucleotide antisense against cPLA2α at 6-week-old hmSOD1 mice, when the elevation of cPLA2α was first detected. Ten micrograms per day of AS or the corresponding sense or saline was continuously pumped into the right lateral ventricle as done in our earlier study in a mouse model of amyloid beta brain infusion [19]. Using this methodology, it was shown [27] that significant oligonucleotide concentrations were achieved in the brain and brainstem and in all levels of the spinal cord. AS brain infusion to 6-week-old mice over a period of 6 weeks significantly prevented upregulation of cPLA2α in the brainstem as detected by immunofluorescence (Fig. 3a) and in the spinal cord as detected by Western blot analysis (Fig. 3b). As expected, at 12 weeks, there was no neuronal damage as well as no activation of microglia or astrocytes and the AS brain infusion had no effect (Fig. 3a). This treatment that did prevent the initial elevation of cPLA2α had no effect on the development of the disease assayed by motor function on rotarod (Fig. 3c). The initial value of 80 % of pre-symptomatic is due to the pump implantation surgery.

Fig. 3
figure 3

Reduction of cPLA2α upregulation at the early stage did not affect the development of the disease. Mice were brain infused with 10 μg/day AS (n = 12) or the corresponding sense (n = 6) or saline (n = 6) for 6 weeks starting at 6-week-old hmSOD1 mice. a Representative immunofluorescence staining of cPLA2α, neurons (NeuN), microglia (Iba-1), and astrocytes (GFAP) in the brain stem tissue section of WT mice, hmSOD1 mice (ALS), hmSOD1 mice brain infused with AS (ALS + AS) or sense (ALS + SE) at 12 weeks. Mice receiving saline or sense showed similar results and were combined under the sense-treated group. A double staining of cPLA2α and NeuN is presented. Scale bars = 50 μm, and in the insets = 20 μm. Shown are representative images of 12 mice in each group. The mean ± SEM of the fluorescence intensity (for cPLA2α) and percentage of the cell area (for cell markers) are presented in the bar graph. *p < 0.001—significance from WT mice and from AS-treated hmSOD1 mice. b A representative immunoblot analysis of cPLA2α and the corresponding calreticulin protein expression in the spinal cord lysates of the mice. cPLA2α protein expression was determined by dividing the intensity of each cPLA2α by the intensity of the corresponding calreticulin band after quantitation by densitometry and expressed in the bar graph as arbitrary units. The bar graph is the mean ± SE of 8 mice in each group. *p < 0.001—significance from WT mice and from AS-treated hmSOD1 mice. c Motor function by rotarod test of AS or sense brain infused-hmSOD1 mice (n = 12 in each group)

Full size image

To determine the role of cPLA2α in the progression of the disease, 10 μg/day AS or sense was brain infused at week 15 (shortly before the onset of motor neuronal dysfunction) for 6 weeks (due to the pump’s limit). As shown in Fig. 4a, AS brain infusion prolonged the survival in mice by 12 days (<0.05). AS brain infusion significantly delayed the onset of motor neuron dysfunction by 3 weeks and there was a significant (p < 0.001) difference between the two groups until 21 weeks of age (Fig. 4b). Immediately after the pump implantation, there was a reduction in motor function, which was recovered only in the AS-treated hmSOD1 mice. The evaluation of the overall motor symptoms change in mice done by Ladder test (Fig. 4c) showed also a delay of 3 weeks in the loss of motor neuron function in the AS-treated hmSOD1 mice and a significant (p < 0.001) difference from the sense-treated hmSOD1 mice until 23 weeks of age.

Fig. 4
figure 4

AS brain infusion to hmSOD1 mice shortly before the onset of motor neuron dysfunction delayed development of the disease. Fifteen-week-old hmSOD1 mice were brain infused with 10 μg/day AS or the corresponding sense during 6 weeks (n = 12 in each group). AS brain infusion prolonged survival (a), delayed loss of motor function analyzed by Rotarod (b), and delayed neurological score analyzed by a ladder (c). *p < 0.001—significance from sense-treated hmSOD1 mice

Full size image

To determine the processes that are affected by cPLA2α during the development of the disease symptoms and to characterize the effect of prevention of cPLA2α upregulation on the different processes taking place at the first phase of development of motor neuron dysfunction, hmSOD1 mice were brain infused with AS or sense at week 15 for 3–4 weeks and the experiments were terminated for analysis at 18–19 weeks of age, at a time when the two mouse groups—AS and sense control—exhibited a significant difference in the disease symptoms. Non-treated hmSOD1 mice and WT mice at the same age were also studied. As shown in the immunofluorescence analysis and quantitation in Fig. 5a, cPLA2α protein expression was much lower in the spinal cord of 18–19-week-old AS-treated hmSOD1 mice in comparison with hmSOD1 mice or sense-treated hmSOD1 mice. AS treatment prevented the reduction in the number of the neurons (detected by NeuN), prevented astrocyte activation (detected by GFAP), and inhibited microglia activation (detected by Iba-1 and by CD40) compared with hmSOD1 mice or sense-treated hmSOD1 mice. In AS-treated hmSOD1 mice, reduction of cPLA2α upregulation led to diminution of the key inflammatory enzymes; iNOS and COX-2 as detected by Western blot analysis in the spinal cord lysates (Fig. 5b).

Fig. 5
figure 5

Reduction of cPLA2α upregulation in the spinal cord of hmSOD1 mice shortly before the onset of motor dysfunction reduced neuronal death and gliosis. Fifteen-week-old hmSOD1 mice were brain infused with 10 μg/day AS or the corresponding sense for 3–4 weeks and analyzed at week 18–19. a Representative immunofluorescence staining of cPLA2α, neurons (NeuN), microglia (Iba-1), and astrocytes (GFAP) in the spinal cord tissue section of WT mice, hmSOD1 mice (ALS), hmSOD1 mice brain infused with AS (ALS + AS) or sense (ALS + SE) at 18–19 weeks. Scale bars = 50 μm. Shown are representative images of 12 mice in each group. The mean ± SEM of the fluorescence intensity (for cPLA2α) and percentage of the cell area (for cell markers) are presented in the bar graph. *p < 0.001—significance from WT mice and from AS-treated hmSOD1 mice. b Representative immunoblot analysis of cPLA2α, COX-2, and iNOS and the corresponding calreticulin protein expression in the spinal cord lysates of the mice. Densitometry analysis for the three proteins was done as described for cPLA2α in Fig. 3b. The bar graph is the mean ± SE of 8 mice in each group. *p < 0.001—significance from WT mice and from AS-treated hmSOD1 mice

Full size image

Discussion

We show herein that cPLA2α is upregulated in the spinal cord of 6-week-old mice long before the appearance of the disease symptoms, neuronal death, or gliosis and remained elevated during the whole life span of the hmSOD1 mice. Prevention of cPLA2α upregulation at the pre-symptomatic period (between 6 and 12 weeks) did not affect the course of the disease. However, prevention of cPLA2α upregulation shortly before the onset of the disease symptoms significantly (p < 0.001) delayed the loss of motor neuronal function, suggesting that cPLA2α upregulation in the spinal cord plays a role in the disease pathology. In line with our results, the role of elevated spinal cord cPLA2α in neuronal damage was reported also in spinal cord injury [28] and in spinal inflammatory hyperalgesia [29].

In the present study we used, specific antisense brain infusion in order to determine the specific role of cPLA2α upregulation in the spinal cord in the pathogenesis of ALS. The advantage of the AS technique over knockout mice is its ability to inhibit the induction of an accelerated protein at the site of inflammation, and enables its expression at basal levels, whereas in knockout mouse models total elimination of a protein may result in dramatic changes of homeostasis and/or compensation by upregulation of other proteins (with similar biological features). The use of AS strategy is reinforced by two recent studies in cPLA2α knockout mice reporting that total elimination of cPLA2α resulted with abnormalities in architecture of synapses and cortical neurons [30] as well as in alteration of brain phospholipid composition [31]. AS brain infusion to 15-week-old hmSOD1 mice (for 6 weeks) shortly before the development of motor dysfunction resulted in reduction of cPLA2α elevated protein expression (as detected at 18–19 weeks in the spinal cord) and significantly delayed the development of the disease symptoms. This blunting of cPLA2α upregulation resulted in inhibition of activation of microglia and astrocytes as well as inhibition of motor neuronal death, as detected by immunofluorescence analysis in the spinal cord of 18–19-week-old hmSOD1 mice.

A generation of mice expressing a conditional deletion of the mutant SOD1 gene showed that each individual cell types in the spinal cord contributes to the development of ALS [8, 32]. Likewise, replacing the myeloid lineage of mutant SOD1 mice with wild-type microglia slowed disease progression [6]. Since cPLA2α upregulation was detected in each cell type in the spinal cord, it probably contributes by different mechanisms to motor neuron dysfunction and cell death: (i) elevated neuronal cPLA2α can act directly in inducing neuronal death (ii) while elevated glial cPLA2α can act indirectly by activating microglia and/or astrocytes. In our recent study in primary neuronal cultures, we reported that cPLA2α activation and upregulation induced apoptotic neuronal death [18]. Concomitant with our results, cDNA microarray analysis to monitor gene expression in the spinal cord of hmSOD1 mice during neurodegeneration revealed changes in apoptosis-related gene expression [4]. In addition, the role of cPLA2α in glutamate excitotoxicity and oxidative stress-induced cell death in primary spinal cord neuron cultures was recently reported [33]. Taken together the studies in neuronal cultures and the abundant levels of cPLA2α in motor neurons in the spinal cord, it is suggested that neuronal cPLA2α plays a direct role in motor neuron cell death. Similar to other neurodegenerative diseases and acute CNS insults, one of the most striking hallmarks of ALS shared by familial and sporadic patients, as well as by rodent models, is neuro-inflammation, characterized by extensive microglial activation and astrogliosis that contribute to disease progression, rather than to its resolution [2, 34–36]. Reduction of cPLA2α upregulation in the spinal cord of the hmSOD1 mice prevented astrocyte activation detected by GFAP immunostaining that is in line with in vitro studies [37] reporting that activation of cPLA2α led to an increase in oxidative stress in astrocytes. We show here a massive activation of microglia in the spinal cord (detected by immunostaining of Iba-1 and CD40) that preceded the changes in the motor neurons, in accordance with other studies [6, 38, 39]. This microglia activation was shown to be cPLA2α dependent coincided with ours and others’ studies in cell cultures demonstrating the specific role of microglial cPLA2α in the activation and transformation of microglia to M1 phenotype. We have previously reported that cPLA2α activity regulated the production of superoxides by NOX-2 NADPH oxidase and the induction of COX-2 and iNOS via nuclear factor kB (NF-kB) in microglia cultures [17]. Interestingly, microglial NF-kB specifically has been shown to play a major role in the development of the ALS in hmSOD1 mice [40]. We show here that reduction of cPLA2α in the spinal cord also decreased iNOS and COX-2 upregulation that produce two major proinflammatory mediators; nitric oxide and PGE2, respectively. As shown in the present study for cPLA2α, it was also reported that in both early symptomatic and end-stage transgenic hmSOD1 mice, neurons and to a lesser extent glial cells in the spinal cord exhibit robust COX-2 [41] and iNOS immunoreactivity [42]. Likewise, similar to cPLA2α, COX-2 was dramatically increased in postmortem spinal cord samples from sporadic ALS patients [ 41]. Nitric oxide and superoxides both under cPLA2α regulation [17] can form the toxic reagent peroxynitrite [43, 44]. In this context, a recent study reported that inhibition of cPLA2α activity ameliorated experimental autoimmune encephalomyelitis via blocking peroxynitrite formation in mouse spinal cord white matter [45]. These results suggest that glial cPLA2α is involved by different inflammatory processes in the development of ALS. In accordance with this suggestion, injections of PLA2 into the spinal cord of mice caused inflammation and oxidative stress [46].

The cause of cPLA2α elevation in the brain and spinal cord as early as 6-week-old mice is not yet clear. SOD1 insoluble protein complexes (IPCs) were detected in motor neurons of 30-day-old mSOD1 mice [47], before the manifestation of ALS pathology, and several months before the appearance of inclusion bodies. It may be possible that the formations of IPCs in the neurons triggered the elevation of cPLA2α. The role of this early elevation of cPLA2α is also not known and prevention of this elevation for 6 weeks had no effect on the course of the disease.

The role of monocyte recruitment to the spinal cord in ALS is under debate. On one hand, it was reported that inhibition of monocyte recruitment to the spinal cord delayed motor dysfunction and increased the survival of ALS mice [48]. In contrast, another study [49] reported that monocytes could not be detected in the spinal cord of SOD1G93A mice and that microglia are not derived from infiltrating monocytes. Likewise, a recent study [50] reported that in the mutant SOD1G93A mouse model of ALS, the choroid plexus of the brain did not support leukocyte trafficking during disease progression, due to a local reduction in IFN-γ levels. The results of the present study, in agreement with the two later reports, show that the recruitment of monocytes detected by co-immunostaining with CD169 and Iba-1 was evident only at the end stage.

Conclusions

We show here that cPLA2α is elevated at 6 weeks, long before the appearance of the disease symptoms. cPLA2α elevation at the onset of symptoms plays a role in the induction of motor dysfunction. Reduction of cPLA2α upregulation shortly before the loss of motor neurons function significantly delayed the appearance of the disease symptoms. Elevated neuronal and glial cPLA2α may contribute by different aspects involved in the pathogenesis of ALS. Thus, cPLA2α may offer an efficient target for treatment of ALS and potentially other multifactorial neurodegenerative diseases.

Abbreviations

ALS, amyotrophic lateral sclerosis; AS, antisense oligonucleotide against cPLA2α; C9ORF72, chromosome 9 open reading frame 72; COX-2, cyclooxygenase-2; cPLA2α, cytosolic phospholipase A2 alpha; hmSOD1, human mutant SOD1G93A; iNOS, inducible nitric oxide synthase; NF-kB, nuclear factor kB; SOD1, superoxide dismutase; TDP-43, TAR DNA binding protein