Immune system alterations in amyotrophic lateral sclerosis


S. W. Pedersen, Department of Neurology, Glostrup Hospital, DK-2600 Glostrup, Denmark

Tel.: +45 38 633468

Fax: +45 38 633926



Amyotrophic lateral sclerosis is a disease of which the underlying cause and pathogenesis are unknown. Cumulatative data clearly indicates an active participation by the immune system in the disease. An increasingly recognized theory suggests a non-cell autonomous mechanism, meaning that multiple cells working together are necessary for the pathogenesis of the disease. Observed immune system alterations could indicate an active participation in this mechanism. Damaged motor neurons are able to activate microglia, astrocytes and the complement system, which further can influence each other and contribute to neurodegeneration. Infiltrating peripheral immune cells appears to correlate with disease progression, but their significance and composition is unclear. The deleterious effects of this collaborating system of cells appear to outweigh the protective aspects, and revealing this interplay might give more insight into the disease. Markers from the classical complement pathway are elevated where its initiator C1q appears to derive primarily from motor neurons. Activated microglia and astrocytes are found in close proximity to dying motor neurons. Their activation status and proliferation seemingly increases with disease progression. Infiltrating monocytes, macrophages and T cells are associated with these areas, although with mixed reports regarding T cell composition. This literature review will provide evidence supporting the immune system as an important part of ALS disease mechanism and present a hypothesis to direct the way for further studies.


Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease where the underlying cause and exact mechanisms have yet to be elucidated. The French neurologist Jean Martin Charcot described the most common clinical features of ALS over 150 years ago. The selective loss of motor neurons in the spinal cord, brain stem, and motor cortex leads to progressive muscle weakness, paralysis, respiratory muscle failure, and death usually within 2–5 years. Sporadic ALS (sALS) accounts for 90% of cases, and the remaining 10% arises as familial ALS (fALS), where 20% is caused by mutations in the superoxide dismutase-1 (SOD1) gene [1, 2].

Neuroinflammation is a term describing a process, which includes activation of glial cells and infiltration of peripheral immune cells. This is known to occur in ALS and other neurodegenerative diseases. Neuroinflammation was originally regarded as a protective response trying to limit damage and initiate repair, but emerging evidence suggests that it could execute deleterious actions. These deleterious actions performed mainly by the immune system could possibly impair neuronal survival and stimulate the degenerative process. It is even believed that microglia and astrocytes are able to initiate the neurodegenerative process under certain circumstances [3].

Several studies have explored the possible involvement of immune cells, using primarily animal models where mice overexpress the human mutant SOD1 (mSOD1) protein and, as a consequence, develop disease characteristics comparable to human cases [2]. This literature study will attempt to explore whether the immune system has a connection to ALS disease mechanisms. The study will focus on the presumed deleterious actions of especially microglia, astrocytes, the complement system, and peripheral T lymphocytes.


Amyotrophic lateral sclerosis – genetics and pathogenesis

Genetic studies have discovered several genes causing and associated with ALS. A mutation in the SOD1 gene causes ALS and accounts for approximately 1–2% of all cases. Less common are mutations in transactive response DNA-binding protein, causing mutant TAR DNA–binding protein 43 (TDP-43), which has been identified as a major component of the ubiquitinated inclusions found in patients with sALS. A dominant mutation in the fused in sarcoma or translocation in liposarcoma (FUS/TLS) gene has been observed in several fALS cases, where the corresponding protein is a co-activator of nuclear factor κB (NF-κB). Nuclear factor κB might be involved in an inflammatory response based upon recent data [1, 2, 4].

Mice and rats overexpressing the human mSOD1 protein develop disease characteristics similar to those found in humans and are therefore used extensively to study ALS. There is, however, no animal model available for sALS, but spinal cord samples of wild-type SOD1 from patients with sALS have shown conformational changes resembling those found in mSOD1. This suggests that modified wild-type SOD1 could contribute to sALS pathogenic mechanisms. Therefore, findings regarding fALS could be relevant in transferring them to sALS studies [2].

Regarding the pathogenesis of ALS, several hypotheses have been proposed, including glutamate excitotoxicity, oxidative injury, autoimmunity, and neuroinflammation [5]. It is well known that motor neurons in patients with ALS are dying. Why on the other hand, it is still under extensive research. Selective expression of mSOD1 in motor neurons or microglia did not result in motor neuron loss. This observation supports the theory of a non-cell autonomous mechanism, meaning that a deleterious interplay between different cell types is necessary for the pathogenesis of the disease [2]. Given that ALS seems to be a non-cell autonomous disease and that immune system alterations have been found in the spinal cord and motor cortex of patient with sALS and fALS, the immune system could play a pivotal role in this context [5].

Local immune system

Activation of astrocytes and microglia is accepted as a part of the pathological process in ALS [6]. The non-cell autonomous mechanisms thought to occur in ALS combined with the evidence of neuroinflammation are just partly responsible for the increased acknowledgement of the role of these cells.


Microglia are considered the resident immune cells of the central nervous system (CNS). Under normal conditions, microglia exhibit a deactivated phenotype, monitoring the extracellular environment, and may play an important role in maintenance of homeostasis through communications with astrocytes and neurons [1, 2]. Classical activation of microglia leads to a neurotoxic M1 phenotype, whereas alternative activation induces a protective M2 phenotype. See Table 1 for properties [1, 7, 8].Activated microglia are found in the motor cortex and spinal cord of patients with ALS, where the intensity of activation correlates with severity of upper neuron damage and is associated with infiltration of T cells [2, 8]. In vivo models have revealed that microglial proliferation and activation possibly occur already at earlier stages of disease, which subsequently increase with disease progression [9].

Table 1. Properties of M1 and M2 microglia

Triggered: GM-CSF

Primed by: IFN-γ

Triggered: IL-4 and IL-13

Primed by: M-CSF


Inflammatory cytokines:

TNF-α, IFN-γ, IL-1β, IL-6, and IL-12


Anti-inflammatory factors:

IGF-1, IL-4, and IL-10


Lipopolysaccharide (LPS) is known to be an activator of microglia through Toll-like receptor (TLR), facilitated by the protein CD14 [1, 4]. Chronic infusion with LPS in a presymptomatic ALS mouse exacerbated the disease progression and ameliorated the innate immune response. A similar study found increased levels of the helper T cell 1 (Th1) promoting cytokine interleukin (IL)-12 in the brains of mSOD1 mice. In a cell culture system stimulated with LPS, mSOD1 microglia released higher levels of pro-inflammatory cytokines such as tumor necrosis factor-α (TNF-α) and IL-6 compared to wild-type microglia [1, 8].


Further evidence confirming a role for microglial activation in motor neuron pathology was observed in transgenic (Tg) SOD1G93A mice, which showed a connection between activated microglia and elevated cyclooxygenase-2 (Cox-2) and prostaglandin E2 (PGE2) levels surrounding affected motor neurons at early symptomatic stages. In a similar mouse model, PGE2 signaling displayed evidence of a possible neurotoxic pathway. Decreased inflammatory effectors such as Cox-2 resulted from the deletion of the PGE2 receptor, which led to improved motor strength and survival time in these mice [2, 3].


Microglia and astrocytes expressing mSOD1 increase their activation of Nox-2. This subunit of the NADPH oxidase complex is thereby locked in a superoxide-producing form resulting in increased production of reactive oxygen species (ROS) [2, 3, 10].


Astrocytes constitute the majority of glial cells in the CNS. They are not professional immune cells, but can upon activation influence ongoing immune responses [2]. Astrocytes execute functions that are beneficial for motor neurons under normal circumstances. These are apparently lost in astrocytes expressing mSOD1. Protection against glutamate excitotoxicity is compromised by loss of the excitatory amino acid transporter 2 (EAAT2) glutamate transporter, which has shown severely reduced levels in sALS as well as in patients with fALS [6, 8]. Astrocytes can upon activation secrete a variety of substances including monocyte chemoattractant protein-1 (MCP-1), Cox-2, prostaglandins, glutamate, IL-6, TNF-α, ROS, and Fas ligand [9, 11].

The complement system

The complement system is a collection of proteins and proteolytic enzymes. These factors can upon activation initiate a self-amplifying cascade. Of the three activation pathways, the classical pathway was found to be active in human ALS tissue. Some factors are locally produced in the CNS, including the initiator of the classical pathway C1q, C3, and C4, whereas others derive from the blood [12]. Several independent studies have reported elevations of C1q in spinal cord tissue and isolated motor neurons in Tg mice models of ALS [13].

Quantitative PCR in a human study found that C1q mRNA did not show significant difference compared to controls, unlike C4 mRNA, which was up-regulated. In situ hybridization revealed distribution of the mRNAs in the spinal cord and motor cortex, where expression by both motor neurons and glial cells was observed. Immunohistochemistry revealed strong staining for the C1q protein antibody in motor neurons, which colocalized with a stronger C1q mRNA presence. This suggests that most of the locally produced C1q is of neuronal origin [12].

The downstream factors C3 and C5b-9 were also found to have elevated mRNA and protein levels. C3 protein deposition was found on activated microglia and astrocytes, whereas C5b-9 colocalized with activated microglia and neurons [12]. In the CNS of patients with ALS, complement receptors and MHC molecules are highly expressed by activated microglia [1].

Peripheral immune cells

T cells

There is strong evidence of T-cell infiltration in the CNS of patients with ALS. In close proximity to dying motor neurons, observations revealed increased levels of CD4+ and CD8+ T cells in the spinal cords of mSOD1 mice and in the brain parenchyma of patients with ALS [1]. The different subclasses of CD4+ and CD8+ cells could each contribute to inflammation in various ways. See Table 2 [9, 14].

Table 2. Properties of different T-cell subsets
Thought statusPro-inflammatoryAnti-inflammatoryAnti-inflammatoryPro-inflammatory
SecreteIFN-γ, IL-1, and TNF-βIL-4 and IL-13IL-4, IL-10, and TGF-βIFN-γ, Fas-L, perforin, and granzymes
Possible functionsStimulate/induce M1 microglia through secreted cytokinesStimulate/induce M2 microglia through secreted cytokines. Up-regulate the release of IGF-1Maintain immunological tolerance and homeostasis. Down-regulate production of inflammatory cytokines. Suppress peripheral immune cell activationStimulate/induce M1 microglia through secreted cytokines

A study of brains and spinal cords from diseased patients with ALS found significant elevated levels of CD8+ T cells compared to controls, but found no evidence of CD4+ cells in the tissues [12]. Another ALS study of patients' blood by flow cytometry detected elevated numbers of CD8+ cells and decreased numbers of CD4+ Tregs, but found no difference in CD4+ T cells [15]. Decreased numbers of Tregs in the blood of patients with ALS have also been reported at early stages, which was thought to be caused by recruitment to the CNS, while another study found that this correlated with faster disease progression [5, 14, 15].

Monocytes and macrophages

Monocytes and macrophages displayed elevated levels in all patients from a sALS study. Their degree of activation was also directly related to the progression rate. MCP-1 is essential for the migration of these cells to relevant areas. Elevated levels of this chemoattractant have been detected in serum and cerebrospinal fluid from patients with ALS. Astrocytes in an ALS spinal cord showed the highest immunoreactivity for MCP-1, indicating that astrocytes might have a role in mediating this response. Activated microglia, monocytes, and dendritic cells are able to present antigens to Th1 cells, which in turn enhances the inflammatory response by production of cytokines [9].

B cells

B cells are not considered to be involved in ALS disease mechanisms, based on findings in both human and animal studies [7, 12].


The objective of this review was to explore the connection between the immune system and ALS. This connection is confirmed by accumulating amounts of evidence revealing a relationship between immune system alterations and ALS disease mechanisms. Observations from gene expression profiling indicate that inflammatory cascades are activated prior to the initiation of the neurodegenerative process. This suggests that the immune system could already be involved in the presymptomatic phase of ALS. The nature of this involvement is, however, still unclear [1].

One of many theories is that onset and early stages are determined by protein aggregates in motor neurons, whereas progression and duration are determined by surrounding cells [2]. Several contradicting theories regarding ALS and the immune system exist, making it difficult to assess the time of contribution belonging to the different components. In fact, alterations in the following suggested pathways may occur simultaneously, rather than sequentially [6]. The remaining discussion will therefore focus on plausible mechanisms involved, and less focus will be placed on the time aspect.

Damaged motor neurons

Damaged motor neurons could be able to induce proliferation and activation of microglia and astrocytes through several routes. Mutated protein aggregates, like mSOD1, are thought to be secreted by motor neurons through a chromogranin chaperone-like process. An observation of extracellular mSOD1 in the spinal cords of patients with ALS possibly supports this theory [1, 2]. Similar to LPS, extracellular mSOD1 recognized by TLR and associated protein CD14 could induce classical activation of microglia [1, 8].

Damaged motor neurons may secrete ATP, which subsequently could activate microglia and astrocytes through the purinergic receptor P2X7. Observations of extracellular ATP in the spinal cords of patients with ALS support this as a plausible pathway [1, 8].

The complement factor C1q produced by MNs can directly activate microglia toward an inflammatory phenotype and initiate the classical complement pathway [12].


Activation of microglia, astrocytes, and the complement system may result in the disruption of the blood–brain barrier, which combined could attract and activate adaptive immune cells [12].

Activated complement factors are further able to opsonize membranes and form the membrane attack complex (MAC/C5b-9). By self-amplification, the downstream activation of C3, C5, and C5b-9 will result in even more activated complement components and cellular damage. Inhibitors normally regulate these activation cascades, but this inhibition seemingly fails to cope with the inflammatory response in ALS. Defective complement inhibition is associated with, among others, induction of autoimmune responses [12].

Astrocytes and microglia both appear to suppress their respective trophic actions and acquire a more cytotoxic phenotype as the disease progresses. Both cells are able to modulate activation of each other, but the nature of this communication is largely unknown. Possible mediators include IL-6 and excess glutamate [9].

The pro-inflammatory cytokines secreted by M1 microglia are able to induce activation of the transcription factor NF-κB, which modulates expression of various proteins such as Cox-2. Cyclooxygenase-2 released by microglia and astrocytes is able to increase the production of inflammatory cytokines and prostaglandins. Prostaglandins could in turn contribute to excitotoxicity and oxidative damage by triggering glutamate release from astrocytes and inducing formation of free radicals [4, 11].

Infiltrating peripheral immune cells could be locally activated and acquire various properties depending on the nature of the surrounding milieu. Observations in several human studies have revealed mixed results, which might be explained by the nature of these studies and their vulnerability to bias [15]. Overall, the evidence clearly suggests that T cells are somehow involved in the disease, at least at end stages where CD8+ T cells are likely to dominate. CD4+ Tregs are considered important for immunological homeostasis and tolerance and are known to be able to suppress activation of peripheral immune cells. Decreased numbers of these cells could indicate an early attempt to recruit these cells to the CNS to fight the local immunological responses, or it could be a sign of failing tolerance [15].

Death of a motor neuron

The non-cell autonomous mechanism thought to be responsible for motor neuron death has been extensively studied. The notion that motor neuron death is a result of multiple cells in collaboration is one interesting aspect. Another interesting aspect is the actual effect the mSOD1 has on these cells compared to their normal counterparts. Of particular interest is the fact that mSOD1 astrocytes have shown to be more toxic than normal reactive astrocytes, and they release factors especially toxic to motor neurons. The theory that motor neuron death is caused by a non-cell autonomous mechanism is strongly supported by observations of mSOD1 astrocytes. Wild-type motor neurons die not only in the presence of mSOD1 astrocytes, but also when cultured in a milieu defined by mSOD1 astrocytes [16].

Activated microglia and astrocytes could be able to induce motor neuron death by releasing different substances capable of activating TNF-α-mediated apoptotic mechanisms, Fas- or NO-induced apoptotic pathways [9]. Isolated motor neurons from Tg mSOD1 mice had increased sensitivity to Fas- or NO-triggered death compared to wild-type motor neurons, suggesting that motor neurons containing mSOD1 are especially vulnerable to these substances [1, 16].

Elevated glutamate levels in the synaptic cleft resulting from excess glutamate release or loss of the EAAT2 transporter could stimulate receptors on motor neurons, leading to increased calcium influx and motor neuron death [2]. Increased levels of glutamate have been found in the cerebrospinal fluid of patients with ALS [6]. The only clinically approved drug for ALS to date, riluzole, reduces glutamate release from motor neuron terminals. This confirms glutamate excitotoxicity as part of ALS pathogenesis [8].

Excess levels of ROS produced by microglia and astrocytes could contribute to motor neuron death by several mechanisms. ROS are able to damage plasma membrane lipids and proteins, thereby compromising the integrity of neighboring cells. They can oxidize proteins of importance to motor neuron survival, such as the insulin-like growth factor 1 (IGF-1) receptor. They can promote excitotoxicity by damaging the EAAT2 glutamate transporter on astrocytes [3]. In this manner, microglia and astrocytes are able to cause oxidative stress, but whether this is a primary or secondary event is unknown. Oxidative stress is known to potentiate the effects of other pathological processes such as excitotoxicity and protein aggregation. In fact, these processes are increasingly recognized as integrated, but the mechanism behind this is unknown [6, 9].

Disease progression

In ALS, there seems to be a disruption of balance between clearance and repair in the CNS. The different pathways that have been proposed may induce positive feedback loops which exacerbate the neurotoxic effects of the immune system, making protective functions more and more insufficient [12].

One of the main challenges is to determine which cells are present at different times. As mentioned previously, it has been postulated that microglial activation has occurred at early stages of disease, which increases with disease progression and is associated with T-cell infiltration. The decreased levels of Tregs in the blood of patients in early stages are thought to be caused by recruitment to the CNS. Given the deleterious effects of M1 microglia, it is tempting to focus on these cells, but new evidence indicates that the neuroprotective M2 microglia play a much greater role than previously assumed, at least in early stages [16, 17].

If the hypothesis that protein aggregation in motor neurons is the trigger is right, then it is likely that local and peripheral immune cells would respond. The first line of defense should have been microglia, astrocytes, and the innate immune system including the complement system, monocytes, macrophages, and natural killer cells. Given that the first responders are not all protective in nature could partially be caused by the direct effect of extracellular and intracellular mSOD1. Then, as a consequence of protein aggregation itself and the innate immune response, the adaptive immune system is included. Trying to repair the damages is the initial goal. See Fig. 1A Early stages.

Figure 1.

[Correction added on 25th April 2013, after first online publication: In Figure 1, part A should have been labeled ‘Early stages’, and part B should have been labeled ‘Late stages’.] (A) Early stages: the theory is that T-cell subsets and M2 microglia are activated as an attempt to halter the neurodegenerative process and protect damaged motor neurons. Motor neurons on the other hand act self-destructive by secreting protein aggregates and C1q, which activates M1 microglia and the classical complement pathway, respectively. Astrocytes are solely harmful by secretion of neurotoxic substances and mediators which influences, recruits, and communicates with other immune cells. (B) Late stages: a mouse model of ALS revealed that decreased numbers of regulatory T cells correlated with increased M1 microglia and CD8+ activity. This shift in microglial and T-cell subsets aberrates the protective aspect of the immune system, resulting in accelerated motor neuron degeneration. Communication between cells further exacerbate the vicious circle. COX-2: cyclooxygenase-2; Fas-L: Fas ligand; IFN-γ: interferon-γ; IGF-1: insulin-like growth factor-1; IL-4: interleukin-4; MAC: membrane attack complex; MCP-1: monocyte chemoattractant protein-1; PGE2: prostaglandin E2; ROS: reactive oxygen species; TNF-α: tumor necrosis factor-α. Red arrows indicate a harmful pathway, whereas green arrows illustrate protective pathways. The numbers refer to Table 3 and illustrate drug targets. Protein aggregation is the trigger of disease, in this case mutated superoxide dismutase-1. Of importance, the mutation is present in every cell, making them more vulnerable to certain stimulus and obtain various properties which may differ from their normal state.

As the disease progresses, a shift occurs, where the previously dominating M2 microglia and Tregs are outnumbered by CD8+ and M1 microglia. This happens to coincide with decreased numbers of Tregs in the spinal cords of mSOD1 mice. The reduction in Tregs and possibly other neuroprotective cells may alter the surroundings in a way favorable to deleterious cells, causing the shift and the rising numbers of CD8+ and M1 microglia.

Astrocytes appear to be more consistent in the way they contribute. The neurotoxicity of astrocytes could be caused by several factors including failure to clear debris in the CNS, failure to maintain the blood–brain barrier, failure to maintain normal glutamate transport, release of pro-inflammatory cytokines and other neurotoxic factors like ROS, and release of MCP-1 and hence recruitment of peripheral immune cells.

It seems to be a losing battle – more and more motor neurons are dying, which means more and more antigen-presenting cells are returning to the region's lymph nodes where they could stimulate T cells. The blood–brain barrier may suffer from lack of protection by local cells, causing it to leak and allowing more and more peripheral immune cells to enter. At some point, the neurotoxic stimulus outweighs the neuroprotective, causing the observed shift in cells [16, 17]. See Fig. 1B Late stages.

Challenges in research and drug development

Research and drug development in ALS are currently facing many challenges. How reliable is it to transfer results from animal studies to human theories? On the other hand, how reliable are human studies? ALS is a relatively rare disease with short survival time, making studies vulnerable to bias. A frequent problem is patient groups which are too small and often comprised of patients at different stages of the disease. There are surprisingly few studies researching the protective functions of the immune system. Drugs that show beneficial effects in animals have failed to provide the same effects in humans, which could have several explanations. The drugs being studied are often administrated presymptomatically or before onset, which to date is not possible in humans, because diagnosis is not confirmed before symptoms or motor neuron loss is present. Crucial for further drug testing and human research is discovering biological markers for the disease, which could result in earlier diagnosis and improved drug research.

Despite great efforts, there is still no cure for ALS. Multiple theories regarding the pathogenesis of disease have resulted in numerous drug therapy strategies, none of which has proven to be the miracle cure. The many setbacks during clinical trials have led to the notion that the answer lies in a combination therapy, rather than a miracle pill. To date, no results from major studies on combination therapies are available. The development of an effective ALS treatment is complicated, not simply because of the many mechanisms at play or lack of biomarkers making early diagnosis possible, but also because there is still too much we do not know about the disease. The task is further complicated because drugs targeting the immune system may be beneficial at one stage of disease, while it may be harmful at another. The future of ALS treatment appears to lie in a combination cocktail, and finding the right combination is key. In a combination cocktail, combining drugs that target protein aggregation, glutamate excitotoxicity, and immune modulation could possibly be a perfect match. See Table 3 for a selection of promising drugs [18-26].

Table 3. Selection of promising and relevant drugs for ALS treatment
AgentTargets and applicationActionsAnimal studiesHuman studiesAdvantages and identifier
  1. ALS, amyotrophic lateral sclerosis; EAAT2, excitatory amino acid transporter 2; CNS, central nervous system; SOD1, superoxide dismutase-1.

  2. Numbered agents have their targets equivalently numbered in Fig 1A,B.

(1) RiluzoleGlutamate excitotoxicityDecreases the presynaptic release of glutamateSignificantly delayed onset and slowed down loss of motor function in preclinical trialsTreatment is safe and well tolerated, with moderate increased survival time of 2–9 monthsDue to loss of efficacy, research of effects in older patients, in bulbar ALS, and in patients with advanced ALS is needed 

Glutamate excitotoxicity

Beta-lactam antibiotic

Increases activity and gene expression of EAAT2 glutamate transporterIncreased levels of EAAT2, delayed motor neuron loss, and increased survival time in mSOD1 micePhase III RCT completed 2012. Study results were negativeRequires intravenous injection. Limited safety profile in humans. Ceftriaxone is known to have excellent CNS penetration.NCT00349622
ArimoclomolProtein aggregationInhibition of protein aggregation by inducing heat shock proteins during cellular stress.Significantly increased survival time in mSOD1 mice when given before or at symptom onsetSafe and well tolerated in phase II clinical trial. Currently recruiting patients for a phase II/III RCTGiven that presymptomatic administration is not possible in humans, the drug could be less efficient. Good CNS penetrationNCT00706147

Oxidative stress and mitochondrial dysfunction.

Chirally pure R-enantiomer of pramipexole which is used in parkinson's disease

The precise mechanism of action is unclearIn mSOD1 mice, the drug revealed increased survival, antioxidant, anti-apoptotic, and neuroprotective effectsClinical trials observed: slowed disease progression, slowed functional decline, and decreased mortality. A large-scale phase III study started in 2011Safe and well tolerated in a two-part phase II studyNCT01281189
(2) AEOL-10150Oxidative stressAntioxidant analogous to the catalytic site of superoxide dismutase. Neutralizes reactive oxygen speciesThree different mice models showed beneficial effects when given at symptom onset. Significant increased survival timePhase 1 study was safe and well tolerated. Further studies were suspended due to financial reasonsGiven that presymptomatic administration is not possible in humans, the drug could be less efficient 
CreatineOxidative stress and mitochondrial dysfunctionStabilizes the mitochondrial transition pore and facilitates mitochondrial ATP synthesis. Antioxidant propertiesIncreased survival time in transgenic mSOD1 mice when given before disease onset

Several phase II clinical trials gave negative results.

Clinical trials with higher doses and combination with tamoxifen are ongoing

Oral administration, elevated brain penetration, and good safety profile.

Given that presymptomatic administration is not possible in humans, the drug could be less efficient

(3) CelecoxibCox-2 inhibitorInhibition of cyclooxygenase-2 and hence inhibition of prostaglandin triggered glutamate excitotoxicity and oxidative stressPositive preclinical studies on mSOD1 transgenic mice. Increased survival and motor neuron protectionA phase III RCT showed no beneficial effects on patients with ALSCox-2 inhibition cannot be rejected as a viable CNS target because the lack of detectable action of celecoxib in the CNS limited the phase III clinical trial 
(4) Minocycline

Apoptosis and microglial modulation.

Tetracycline antibiotic

Modulate microglial activation via TLR4 receptor inhibition. Inhibition of cytochrome c and caspase-3 activationBeneficial effects in preclinical trialsSafe and well tolerated in phase II clinical trials. A phase III RCT revealed greater and faster declinein patients with ALSDue to alterations in microglial phenotypes, timing of administration might be crucial to the outcome! 
(5) Glatiramer acetate

T-cell modulation.

Treatment for multiple sclerosis

Enhances regulatory T-cell immunity and stimulates TH2 cells. May have anti-glutaminergic and growth factor stimulating effectsConflicting results, some mice models showed increased survival time, while others did notA phase III RCT showed no beneficial effects. Extrapolation from preclinical data showed the doses used were too low, making the results inconclusiveAdministered as subcutaneous injectionsNCT00326625



A peroxisome proliferator-activated receptor-γ agonist with potent anti-inflammatory effectsImproved muscle strength, delayed disease onset, and prolonged survival in SOD1 transgenic micePhase I/II RCT combining pioglitazone and tretinoin in patients on riluzole. Study results were negative 



Cyclosporine AImmunosuppressive agentBinds small intracellular regulatory proteins, leading to inhibition of the phosphatase calcineurin and T-cell activationIncreased survival time in transgenic mice with permissive BBBEarly studies showed no beneficial effects in patients with ALSCannot pass the blood–brain barrier! Neuroprotective effects from this drug are therefore limited, unless the blood–brain barrier is compromised 


The evidence of immune system alterations in ALS is of such magnitude and correlation to disease mechanisms that it cannot be ignored. The trigger of disease appears to be mutated protein aggregates, like mSOD1, TDP-43, or FUS/TLS. Regardless of mutation, they all trigger mechanisms which culminate in the same disease. Mutations associated with ALS may not trigger exactly the same reactions or influence cells identically, but there are common denominators. The immune system is always involved, the transcription factor NF-κB is thought to be a key player, and the fact that wild-type SOD1 from patients with sALS resembles mSOD1 found in patients with fALS suggests that other mutations could induce many of the same responses.

Present in each cell, mSOD1 is able to alter the natural order of CNS. Motor neurons expressing mSOD1 are more sensitive to certain neurotoxic substances. Motor neurons are able to activate M1 microglia and the classical complement pathway, which exerts harmful actions toward motor neurons. Astrocytes expressing mSOD1 have lost their neuroprotective properties and are extremely toxic to motor neurons by secreting glutamate, Fas-L, ROS, and other cytokines. In addition, by secreting MCP-1, astrocytes are able to recruit peripheral immune cells which also contribute to neurodegeneration. Cytokines and other messengers enable communication, stimulation, and inhibition between cells.

Amyotrophic lateral sclerosis is a disease with many unanswered questions. A deeper understanding of disease mechanisms is desperately needed to develop a treatment. Furthermore, biomarkers making early diagnosis possible are crucial for effective therapeutic research. When moving forward, more research on possible common denominators between different mutations is needed, which could reveal common pathways of destruction in the CNS of patients with ALS. Finding common pathways would represent attractive drug targets. When the trigger and mechanisms leading to motor neuron degeneration are ascertained, additional research examining whether and how protective immune cell subsets are involved could reveal drug targets that enforce neuroprotection.

Mutated protein aggregates may very well be the trigger of this devastating disease, but by itself it is not sufficient enough to cause neurodegeneration. The involvement of the immune system in ALS fits the profile of a catalyst, which enables and dictates the degenerative process.


Neuroimmunologist Matilda Degn Vinther is acknowledged for providing critical comments to the manuscript, and Gilda Leslie Kischinovsky is acknowledged for proofreading and language editing in the manuscript.

Conflicts of interest

We have nothing to disclose in relation to this project.