Complement activation at the motor end-plates in amyotrophic lateral sclerosis

Background Amyotrophic lateral sclerosis (ALS) is a fatal progressive neurodegenerative disease with no available therapy. Components of the innate immune system are activated in the spinal cord and central nervous system of ALS patients. Studies in the SOD1G93A mouse show deposition of C1q and C3/C3b at the motor end-plate before neurological symptoms are apparent, suggesting that complement activation precedes neurodegeneration in this model. To obtain a better understanding of the role of complement at the motor end-plates in human ALS pathology, we analyzed post-mortem tissue of ALS donors for complement activation and its regulators. Methods Post-mortem intercostal muscle biopsies were collected at autopsy from ALS (n = 11) and control (n = 6) donors. The samples were analyzed for C1q, membrane attack complex (MAC), CD55, and CD59 on the motor end-plates, using immunofluorescence or immunohistochemistry. Results Here, we show that complement activation products and regulators are deposited on the motor end-plates of ALS patients. C1q co-localized with neurofilament in the intercostal muscle of ALS donors and was absent in controls (P = 0.001). In addition, C1q was found deposited on the motor end-plates in the intercostal muscle. MAC was also found deposited on motor end-plates that were innervated by nerves in the intercostal muscle of ALS donors but not in controls (P = 0.001). High levels of the regulators CD55 and CD59 were detected at the motor end-plates of ALS donors but not in controls, suggesting an attempt to counteract complement activation and prevent MAC deposition on the end-plates before they are lost. Conclusions This study provides evidence that complement activation products are deposited on innervated motor end-plates in the intercostal muscle of ALS donors, indicating that complement activation may precede end-plate denervation in human ALS. This study adds to the understanding of ALS pathology in man and identifies complement as a potential modifier of the disease process. Electronic supplementary material The online version of this article (doi:10.1186/s12974-016-0538-2) contains supplementary material, which is available to authorized users.


Background
Amyotrophic lateral sclerosis (ALS) is the most common adult-onset motor neuron disease [1]. It is characterized by progressive loss of both upper and lower motor neurons, leading to muscle atrophy and eventually death [2]. Most ALS cases (90 %) are sporadic, while 10 % are familial. Many genes have been identified for familial ALS;C9orf72, FUS, TARDBP are the most frequently affected genes. Mutations in the gene encoding for the copper-zinc superoxide dismutase-1 (SOD-1) are found in about 10 % of familial cases of the disease. The transgenic SOD1 G93A rodent model recapitulates onset and progression of ALS.
The mechanisms leading to ALS are still unclear; both cell autonomous and non-cell autonomous mechanisms are involved [3][4][5]. A role of neuroinflammation [6][7][8] and early involvement of the neuromuscular junction in the SOD1 G93A rodent model has been suggested [9,10]. The complement system has been also associated with neuroinflammation in the ALS rodent model [8,11]. Complement is a key component of the innate immunity, but it can cause harm to tissue. Regulators of the complement system permit elimination of pathogens or dead cells without injuring the host. When this balance is disrupted, complement activation causes injury to the host and contributes to pathology in various diseases [12][13][14].
A role for complement in the pathogenesis of ALS in man is suggested by different researchers. Elevated concentrations of complement activation products in serum and cerebrospinal fluid were detected in ALS patients. In the spinal cord and motor cortex of patients with sporadic ALS, mRNA for C1q, C4 and protein levels of complement proteins C1q, C3, and membrane attack complex (MAC) were elevated [15]. In addition, C1q and C4 were upregulated in motor neurons in murine ALS models [16,17], whereas C3 was upregulated in the anterior horn areas containing motor neuron degeneration [11]. Other studies have also shown upregulation of the major proinflammatory C5a receptor, during disease progression in mouse motor neurons [18]. SOD1 G93A rat treated with C5aR antagonist displayed a significant extension of survival time and a reduction in end-stage motor scores, suggesting an important role for complement in the disease progression [11]. Increased expression of complement components C1qB, C4, factors B, C3, and C5 and a decrease in the expression of regulators CD55 (regulator of C3) and CD59a (regulator of MAC) was detected in the lumbar spinal cord of SOD1 G93A mice [19].
We have previously shown that complement activation products C3/C3b and C1q were present at the motor end-plates of SOD1 G93A mice before the appearance of symptoms and remained detectable at the symptomatic stage, suggesting that complement activation precedes neurodegeneration and plays an early role in this model [20]. Early damage at the end-plates is in line with the "dying-back" mechanism. Retrograde degeneration is detected in ALS patients [21,22], and in transgenic SOD1 G93A mice, retraction of motor axons from their muscle synapse has been shown to occur before any symptoms of the disease appear in the muscle [23], suggesting the disease starts at the motor end-plates.
Here, we analyzed whether key complement components and regulators are also deposited at the motor end-plates in post-mortem intercostal muscles of human ALS cases. We tested for the presence of complement components C1q and MAC and for the regulators CD55 and CD59 in this tissue.

Ethics statements
Tissue was obtained and used in accordance with the Declaration of Helsinki and the Academic Medical Center Research Code provided by the Medical Ethics Committee. Informed consent was obtained from all the patients.

Tissue processing human intercostal muscle
Post-mortem intercostal muscle biopsies were collected at autopsy from sex-matched ALS (n = 11) and control (n = 6) donors at the Department of Neuropathology of the Academic Medical Center (University of Amsterdam). All cases were reviewed by a neuropathologist and diagnosed according to the standard histopathological criteria. None of our patients were on respiratory support. Muscle samples were snap-frozen in liquid nitrogen and stored at −80°C until processed. Detailed information about sex, age, and clinical features of ALS and control donors are given in Tables 1 and 2. In this study, we included material from ALS donors with familial and sporadic ALS. We used age-matched controls, which did not suffer from neuromuscular or neurological disease. The tissue was subsequently embedded in Tissue-Tek, optimal cutting temperature compound (OCT) (Sakura, Zoeterwoude, NL) and cut using cryostat (Reichert Jung; Leica, Nussloch, Germany); cryosections of 6 and 40-μm were cut and stored at −80°C until immuno-and fluorescence stainings were performed.

Nonspecific esterase reaction followed by immunostaining
Fresh frozen sections of 6 μm were tested for the nonspecific esterase (NE) reaction according to the technique of Lehrer and Ornstein [24]. After the NE staining, the sections were fixed for 10 min in 4 % paraformaldehyde (PFA) and after washed in PBS. The sections were permeabalized in PBS/0.2 % TritonX and blocked for 1 h at room temperature (RT) using PBS/5 % fetal calf serum (FCS)/0.2 % TritonX (blockmix). The primary antibodies anti-C5b-9 for MAC recognizes a neo-epitope in C9 (aE11 clone, Dako, Carpinteria, CA), anti-C1q (Dako, F 0254, Denmark), or anti-decay-accelerating factor (anti-DAF) (Abcam, Ab20145, Cambridge, MA, USA) were diluted in blockmix according to Table 3 and incubated for 1 h at room temperature. The sections were washed with PBS three times and incubated with the secondary antibody PowerVision poly-AP anti-mouse IgG or poly-AP anti-rabbit IgG (Immunologic, DPVM55A, Netherlands) for 45 min. After washing, the sections were developed with VECTOR Blue Alkaline Phosphatase (AP) Substrate Kit (SK-5300). The sections were air-dried and mounted using VectaMount (Vector laboratories, H-5000-60, USA).
Animals SOD1 G93A transgenic ALS mice [high copy number; B6SJLTg (SOD1-G93A)1Gur/J] (Gurney, 1994b) and wild-type (B6SJL) littermates were housed in groups at 20°C on 12:12 h light-dark cycle, with free access to food and water. Experimental protocols complied with national animal care guidelines, licensed by the responsible authority. All animals were free of microbiological infection (FELASA screened).
Tissue processing SOD G93A and wild-type mice SOD1 G93A and wild-type mice were sacrificed at postnatal day 47 (presymptomatic stage SOD1 G93A n = 4; wild-type n = 4) by CO 2 inhalation. The gastrocnemius muscle was dissected, post-fixed overnight in 4 % paraformaldehyde/PBS at 4°C, cryoprotected in 30 % sucrose/PBS for 72 h at 4°C, and embedded in Tissue-TEK OTC; cryosections of 40-μm were cut and stored at −80°C until used for histology.

Immunofluorescence staining
For immunofluorescence staining of SOD1 G93A gastrocnemius muscle and human intercostal muscle tissue, sections were air-dried and fixed for 10 min in 4 % PFA at −20°C. Slides were washed in PBS and permeabalized in PBS/0.2 % TritonX. Subsequently, sections were blocked for 1 h at room temperature (RT) using PBS/ 5 % FCS/0.2 % TritonX (blockmix). Primary antibodies were diluted in blockmix according to Table 3 and incubated overnight at 4°C. The following primary antibodies were used (see Table 3 for specifications): antineurofilament heavy-chain (NF-H, Abcam, Cambridge UK) nerves, anti-synaptophysin detecting the motor   Table 4 in blockmix, and incubated for 2 h at RT. Used secondary antibodies are anti-mouse FITC (Jackson Immunoresearch, West Grove, PA), anti-rabbit and anti-mouse Cy3 (Jackson Immuno Research, West Grove, PA), and anti-mouse Cy5 (Invitrogen, Germany). After washing off the secondary antibody, α-bungarotoxin (α-BTX)-Alexa 488 conjugate (Molecular Probes) (1:500), which binds to post-synaptic acetylcholine receptors on the muscle fibers, was applied for 20 min at room temperature to the sections to visualize the end-plates.
The sections then were washed in PBS and air-dried. Vectashield medium (Vector Laboratories Inc, Burlingame, USA) was used for mounting.

Quantification
For each view, Z-stacks of 40-μm thick muscle were examined and scored on the total number of immunoreactivity for Alexa 488-α-BTX/MAC, Alexa 488-α-BTX/ CD59-positive end-plates as well as the total number of NF-H staining co-localizing with specific complement antibodies C1q and CD55 (Table 3). For each muscle sample of an individual donor, 20 non-overlapping microscopic views of 40-μm thick sections were examined (total volume 6.73 × 10 7 μm 3 ).
Each motor end-plate identified with Alexa 488-α-BTX on the surface of a muscle fiber was counted, and the length of the end-plates were measured in both ALS and control muscle biopsies. The size of end-plates was measured using Leica application suite X software (LASX software, Microsystems B.V., Rijswijk, The Netherlands) 3D visualizer, excluding the end-plates that were not completely in the 3D image. The number of immunoreactive area per section was scored and expressed as standard deviation of the mean (SD).

Statistical analysis
Data analysis was performed using GraphPad Prism version 5.0 (GraphPad Software Inc, San Diego, CA, USA) statistical package. Student's t test was performed for statistical analyses comparing two groups. For comparison of more than two groups, one-way ANOVA with Bonferroni multiple comparison post hoc test was used when the data was normally distributed. For nonnormally distributed data, the Kruskal-Wallis test was used. Differences were considered statistically significant when P ≤ 0.05.

Motor end-plates in ALS intercostal muscle
The size and number of motor end-plates were analyzed in the intercostal muscle of ALS donors and age-and sex-matched controls. For analysis of both nerves and motor end-plates, confocal microscopy was performed on 40-μm thick intercostal muscle sections of ALS and control donors that were stained for Alexa 448 α-BTX detecting the motor end-plates and neurofilament heavy-chain antibody (NF-H). Alexa 448 α-BTX positive and negative end-plates were expected in the intercostal muscle of ALS post-mortem tissue given that the average age of the ALS donors was 64 years and that failure of the respiratory muscle occurs in the end-stage of the disease in these patients. All control muscles showed co-localization of α-BTX and NF-H. The BTX-positive end-plates were divided in two groups: (1) end-plates co-localizing with NF-H (innervated) (Fig. 1a, b) and (2) end-plates that showed no co-localization with NF-H (denervated) (Fig. 1c).
The average number of α-BTX-positive end-plates in the intercostal muscles were 87 in controls and 17 in ALS donors per 20 non-overlapping microscopic views. Thus, the intercostal muscle of ALS donors showed a significantly lower number of α-BTX-positive motor end-plates (P = 0.0003) (Fig. 1d). The percentage of innervated and denervated end-plates were 30 and 70 %, respectively, whereas the control donors show 100 % innervation. We also analyzed the size of the α-BTXpositive end-plates in the intercostal muscle of two ALS patients and two age-matched controls. The controls showed a mean of 16.9 μm [SD 4,34) (n = 2), and the ALS cases showed a mean of 11.10 μm [SD 6,44) (P = <0.0001) (n = 2) (Fig. 1e). All end-plates in the 20 non-overlapping views were counted.

C1q deposition on the motor end-plates in the intercostal muscle of ALS donors
C1q deposits were detected before the appearance of clinical symptoms at the muscle end-plate of the SOD1 G93A mouse model [20]. This suggests that complement activation is an early event. Here, we tested in an overview experiment whether C1q deposits are also present in the muscle of ALS donors and if C1q is specifically deposited on the motor end-plate. Immunofluorescence for NF-H and C1q was performed on the intercostal muscle of control (Fig. 2a, b, c) and ALS (Fig. 2d, e, f ) donors. For each individual, we analyzed 20 non-overlapping Z-stacks in 40-μm thick sections using confocal microscopy. C1q immunoreactivity was present in the majority of the intercostal muscle tissue of ALS donors (Fig. 2d, e, f ). An average of 14 [control vs ALS P = 0.001) of the C1q immunoreactive regions in the intercostal muscle of ALS donors were co-localizing with NF-H staining (Fig. 2g, black bar). An additional 20 areas [control versus ALS P = 0.001) of the C1q immunoreactivity were found in the vicinity of NF-H staining (Fig. 2g, gray bar). No C1q immunoreactivity was detected in the intercostal muscle of age-matched controls (Fig. 2b). The NF-H immunoreactivity was generally stronger in the intercostal muscles of control compared to ALS donors (data not shown).
To determine whether C1q is deposited on the endplates, we performed a NE staining on frozen intercostal muscles of control and ALS donors to visualize the endplates followed by an immunostaining for C1q. The immunostaining showed an extensive amount of C1q deposited on and around the end-plates of ALS donors (Fig. 2i). No C1q deposition was detected in on the end-plates of control donors (Fig. 2h). We also detected C1q on the cellular are reduced compared to controls (n = 2) (P = 0.0003 and P = <0.0001, respectively). Error bars represent standard deviation of the mean elements synaptophysin and S100b indicating C1q is also deposited at the motor nerve terminal and terminal Schwann cell in the intercostal muscle of ALS donors (Additional file 1: Figure S1B, D, arrows) but not in controls (Additional file 1: Figure S1A, C).

MAC deposition on the motor end-plates in the intercostal muscle of ALS donors
To determine whether the terminal pathway of the complement system is also activated in ALS, we tested for MAC deposition at the motor end-plates. We analyzed the intercostal muscle of ALS donors. The presence of MAC on innervated or denervated motor end-plates was measured using immunofluorescence and confocal microscopy on 40-μm thick sections. We analyzed 20 nonoverlapping Z-stacks. Human intercostal muscles of control (Fig. 3a, b, c, d) and ALS donors (Fig. 3e, f, g, h) were stained for NF-H, α-BTX detecting end-plates, and C9neo epitope, a component of the terminal complement complex MAC (C5b9). MAC immunoreactivity was detected on and around nerves and on motor end-plates in ALS patients (Fig. 3e, f, g, h). A strong MAC immunoreactiviy was detected (Fig. 3h, asterisks within insert) on the end-plates with a weak α-BTX immunoreactivity (Fig. 3h, arrow within insert). By contrast, a weak MAC immunoreactivity (Fig. 3h, asterisks) was detected on end-plates with strong α-BTX immunoreactivity (Fig. 3h, arrow) and nerves innervating the motor end-plate (Fig. 3h, arrow head). We suggest there might be a relevant anti-correlation between MAC and α-BTX immunoreactivity in the ALS samples. However, the high variability between the biological specimens and the low number of end-plates detected in these samples make it difficult to draw firm conclusions based on the measurement of fluorescence intensities.
No MAC immunoreactivity was detected on or around the end-plates of control donors (Fig. 3c, d). Quantification showed a mean of six innervated (controls vs ALS donors P = 0.01) and 11 denervated (control versus ALS  (Fig. 3i).
To determine whether MAC is deposited on the motor end-plates, we performed immunostainings for MAC followed by NE staining on the intercostal muscle of control (Fig. 3j) and ALS donors (Fig. 3k). We found MAC deposition on the motor end-plates in the intercostal muscle of ALS donors but not in controls, suggesting that the terminal pathway of the complement system is activated on the motor end-plates. We also detected MAC on the cellular elements synaptophysin and S100b indicating that MAC is also deposited at the motor nerve terminal and terminal Schwann cell in the intercostal muscle of ALS donors (Additional file 2: Figure S2B, D, arrows) but not in controls (Additional file 2: Figure S2A, C).

CD55 on the motor end-plates in the intercostal muscle of ALS donors
Regulators such as CD55 and CD59 protect tissues against an attack by the complement system. The role of these regulators in the pathogenesis in ALS is of interest. CD55 acts on the membranes of self-cells to circumvent the deposition of C3b on their surfaces [25]. We found C3/C3b deposition in the intercostal muscle of ALS donors deposited at the motor nerve terminal and terminal Schwann cells (Additional file 3: Figure S3B, D, arrows) but not in controls (Additional file 3: Figure S3A, C). Therefore, we analyzed whether CD55 is also deposited in the intercostal muscle ALS donors. We analyzed 20 non-overlapping Z-stacks in 40-μm thick sections using confocal microscopy. Human intercostal muscles of control (Fig. 4a, b, c) and ALS donors (Fig. 4d, e, f ) were  control (a, b, c, d) and ALS intercostal muscle (e, f, g, h), shows presence of MAC (white asterisks in h and enlarged in the insert) on end-plates (white arrows in h) and around nerves in ALS muscle (white arrowhead in h) but not in controls (c, d). Quantification showed a significantly higher percentage of MAC-positive innervated end-plates (P = 0.001) and denervated end-plates (P = 0.001) in ALS intercostal muscle compared to controls. Numbers of MAC-positive end-plates in 20 non-overlapping Z-stacks in 40-μm thick intercostal muscle sections is given on the y-axis. Error bars represent standard deviation of the mean (i). NE staining (dark brown) followed by an immune staining for MAC (blue) showed (k) MAC deposition deposited on the end-plates of ALS donors (white arrow in k enlarged in the insert), j but not on end-plates of control donors. n.d. not detected stained for NF-H and CD55. We identified strong staining for CD55 on and around nerves in the intercostal muscle of ALS donors (Fig. 4f ) but not in controls (Fig. 4c). Quantification showed a significantly higher percentage of CD55-positive staining. Not all staining co-localized with NF-H in the intercostal muscle of ALS donors (Fig. 4g, gray bar).
To determine whether CD55 is deposited on the endplates, a NE staining on frozen intercostal muscle of control and ALS donors was performed to visualize the end-plates followed by immunostaining for CD55. No CD55 deposition was detected on the end-plates of control donors (Fig. 4c, h), by contrast an extensive amount of CD55 was found deposited on and around the endplates of ALS donors (Fig. 4i), suggesting an increased regulation of the common complement pathway on the end-plates. We also detected CD55 on the cellular elements synaptophysin and S100b indicating that CD55 is also deposited at the motor nerve terminal and terminal Schwann cell in the intercostal muscle of ALS donors (Additional file 4: Figure S4B, D, arrows) but not in controls (Additional file 4: Figure S4A, C).

CD59 on the motor end-plates in the intercostal muscle of ALS donor
The glycolipid anchored protein CD59 has a binding site for both C8 and C9 and as such can prevent formation of MAC [26,27]. Immunofluorescence staining for NF-H, α-BTX detecting end-plates, and the regulator CD59 was performed on intercostal muscle of control (Fig. 5a-d) and ALS (Fig. 5e-h) donors. We analyzed 20 nonoverlapping Z-stacks in 40-μm thick sections using confocal microscopy. CD59 was found abundantly present on and around the motor end-plates in the intercostal muscle of ALS donors (Fig. 5g, h, asterisks) but was negative in the intercostal muscle of control donors (Fig. 5c, d). Quantification showed that this difference is significant for both innervated and denervated motor end-plates of ALS donors (Fig. 5i) (P = 0.05, P = 0.05, respectively). In addition, we show that CD59 is also deposited on the  Figure  S5B, D, arrows) but not in controls (Additional file 5: Figure S5A, C).

Discussion
Although a role for complement has been found in many neurodegenerative diseases [28][29][30][31][32][33][34], its contribution to disease progression in animal models for ALS is controversial [35,36]. We previously provided evidence for an early role of the complement system in the SOD1 G93A mouse model of familial ALS [20]. Fischer and colleagues suggest that ALS pathology starts at the muscle end-plates proceeding to the spinal cord and subsequently the brain [23]. In addition, several physiological and morphological alterations have been reported on the muscle end-plates from in vivo and ex vivo mouse and rat preparations [37][38][39][40][41][42][43].
To obtain a better understanding of the role of complement in human ALS pathology, we analyzed post-mortem tissue of ALS donors for complement activation products and its regulators. We found a lower number and a decreased size of the α-BTX-positive end-plates in the tissue of ALS donors compared to controls, suggesting that the end-plates in the intercostal muscle of ALS patients are affected.
In the ALS muscle, we found deposition of complement activation products C1q and C3 but not in controls. C1q and C3 were detected not only on and around the end-plates but also on the nerve terminal and terminal Schwann cells. C1q and C3 mRNA and protein levels were found elevated in spinal cord and motor cortex of patients with sporadic ALS [15]. In murine ALS models, C1q was also upregulated in motor neurons [16], whereas C3 is upregulated in the anterior horn areas containing motor neuron degeneration. Expression profiling in the mutant SOD1 motor neurons showed that C1q genes were upregulated early in the disease. C1q can bind antibody aggregates and activate the classic complement pathway [11,17]. This data suggests a in ALS intercostal muscle tissue on denervated end-plates (white arrow pointing to α-BTX and arrowhead pointing to NF-H in h) but not in controls. Quantification shows CD59-positive innervated and denervated motor end-plates in the intercostal muscle of ALS donors but not in controls (P = 0.05 and P = 0.05, respectively) (i). Data represents standard deviation of the mean. n.d. not detected role for C1q and C3 in ALS. However, a study by Lobsiger et al. demonstrates no significant pathogenic role for C1q and C3 proteins in the survival of SOD1 G93A ALS mice, contradicting a possible role for complement in this model [44]. This study, however, did not analyze downstream pathways, like the extrinsic pathway of complement which can lead to C5 cleavage and does not need C1q and C3 proteins for activation and MAC formation, which may be the key point at which complement-mediated neurotoxicity occurs in these ALS models [35,45].
A role for MAC in the pathology of neurological disorders is suggested, including ALS [31]. In serum of ALS patients, the terminal complement activation products C5a and MAC are elevated [46]. MAC can damage tissue and target nerves in different neurodegenerative models [34,47], suggesting a role for MAC in degeneration. Furthermore, we show that MAC is deposited on the motor end-plates on day 47 in the SOD1 G93A mouse model, suggesting that MAC deposition is an early event in this model (Additional file 6: Figure S6). This result is in contrast to a previous analysis of the SOD G93A mice [20]. In that study, no MAC was detected on the end-plates of the SOD G93A mice. We attribute this difference to the use of another antibody for the detection of C5b9. We used a monoclonal mouse anti-human C9neo, this antibody gives a specific signal on both frozen human and mouse sections. It detects C9neo and not C9; therefore, it is also more specific to recognize the C9 within the MAC, whereas the polyclonal mouse anti-rat C9 that was used previously either gives no staining or a lot of background on frozen sections. Since we find MAC deposition consistently on endplates in both human and mouse muscle, an artifact is excluded.
The present study shows deposition of MAC at the muscle end-plates of ALS donors. We show strong MAC immunoreactiviy on the end-plates with a weak α-BTX immunoreactivity in the intercostal muscle of ALS donors. By contrast, a weak MAC immunoreactivity was detected on end-plates with strong α-BTX immunoreactivity. This is compatible with a model in which MAC deposition occurs before loss of the end-plates, in fact MAC could be a contributor to disease progression and end-plate pathology. In addition, MAC was also found co-localizing with the motor nerve terminal and terminal Schwann cells.
The propensity of the MAC to "drift" from the site of activation and deposit on other sites may even result in more damage to the muscle. In general, cells are protected from complement attack by multiple complement regulators, preventing damage. This protection can be overwhelmed resulting in damage to tissue and drives inflammation [48,49].
CD55 and CD59 restrict complement activation by inhibiting C3/C5 convertase activities and membrane attack complex formation, respectively. In the actively immunized experimental autoimmune myasthenia gravis mice deficient in either CD55 or CD59, a significant increase in complement deposition at the end-plates was observed and worsened disease outcome associated with increased levels of serum cytokines was observed [50].
We demonstrate that also the regulators of the common pathway CD55 and the terminal pathway CD59 are deposited on the motor end-plates of ALS donors, but not in controls. In addition, the motor nerve terminal and terminal Schwann cells were also co-localizing with CD55 and CD59. Upregulation of the complement regulators CD55 and CD59 on the motor end-plates of ALS patients probably is an attempt to dampen the high level of complement activation and protect the tissue.
Since we also detected MAC deposition at the motor end-plates of the ALS donors, the upregulation of CD55 and CD59 is not sufficient to protect the end-plates from MAC attack.

Conclusions
In summary, we demonstrated that complement activation products C1q and MAC are deposited on motor end-plates in post-mortem tissue of ALS donors. MAC was found deposited on motor end-plates that were innervated by nerves, indicating that complement activation may precede motor end-plate denervation.
Here, we showed that the regulators CD55 and CD59 are also expressed on the motor end-plates, indicating an attempt to control the activation. This process is probably not efficient enough because MAC can still be detected on the α-BTX-positive motor end-plates. Since a role for MAC in the pathology of neurological disorders is suggested [31], detecting complement deposited at the end-plates of ALS donors, before the end-plates are lost, suggests that complement is an early event in ALS and might play an important role in the motor endplate pathology in ALS. This observation is in line with earlier studies suggesting a dying-back mechanism in ALS, meaning the disease probably starts at the motor end-plates [23]. Although this study was performed using post-mortem intercostal muscle tissue of ALS patients and there may be some limitations to our conclusions about complement being involved in motor endplate degeneration, this study adds to the understanding of ALS pathology in man.

Additional files
Additional file 1: Figure S1. Representative confocal immunofluorescence for synaptophysin (SYN-Cy3) detecting the motor nerve terminal (A, B) or S100b (Cy3) detecting the terminal Schwann cells (C, D) double stained with anti-C1q (FITC) in control (A, C) and ALS (B, D) intercostal muscle shows C1q co-localizing with both synaptophysin and