Evidence-Based Drug Treatment in Amyotrophic Lateral Sclerosis and Upcoming Clinical Trials

Abstract

Amyotrophic Lateral sclerosis/motor neuron disease is a severe neurodegenerative disease characterized by upper and Lower motor neuron degeneration for which there is no truly effective treatment. Several therapies have shown promise in preclinical models of motor neuron disease; however, most of them failed in human studies, so that the noticeable progress in understanding the cellular mechanisms of motor neuron degeneration has not been matched with the development of therapeutic strategies to prevent disease progression or to extend survival longer than achieved by riluzole. We review treatment development in motor neuron disease and discuss the strengths and limitations of past as well as upcoming clinical trials.

Introduction

It is the objective of pharmacological therapeutic approaches to amyotrophic lateral sclerosis (ALS) to prevent the deterioration of motor neuron function and thereby, slow down or even arrest the currently inexorable course of the condition ('disease modification', ‘neuroprotection'). Rational pharmacological intervention must be directed towards modification of the pathogenesis that is thought to underlie ALS and that leads to neuronal cell loss. In 2009, it is known that the etiology of the disease is influenced by a number of factors; among them genetic factors which are the best explored. Presently, at least ten genes are known to influence the pathogenesis of the disease [Beleza-Meireles and Al-Chalabi, 2009]. Therefore, it is apparent that the etiology of motor neuron diseases (MNDs) and specifically ALS seems to be diverse, and pathogenetically orientated approaches are believed to be more promising to generate disease-modifying therapies. However, the biochemical and pharmacological influences on the pathogenetic processes are not understood in detail and are likely to be quite heterogeneous and dependent on the stage of disease; only in late stages they may end up in a final common pathway - reflecting the physiology and biochemistry of target cells. This target may also change during development and aging.

Therefore, the development of effective pharmacological therapies seems to be a major challenge and cannot be compared with many other fields in medicine.

Since ALS is such a severe clinical disease, there is also much pressure from the patients, their relatives and the public for fast development of therapeutics (in particular pharmaceuticals) and it is understandable that the history of ALS is filled with non evidence-based treatments which create unjustified promises and hopes. These approaches - not based on rationality - should have limits when the therapeutic claim does harm (including financial harm) to the patient and his relatives. A recent example for such a development is the introduction of ‘stem cell therapy’ into clinical therapeutical strategies of ALS which has never been shown to be successful in a scientifically based way, neither in a clinical trial nor on the basis of credible case reports. On the other hand, the recent past has demonstrated that this pressure on patients, physicians and scientists leads to the danger of jumping uncritically into costly clinical trials on the basis of in vitro and experimental animal studies; however, during the past decade we have seen a number of large clinical trials which were not successful. This partially even led to the reaction ‘that translational research is not valid in ALS/MND and in vitro as well as in vivo experimental research is not justified'.

Current pharmacological approaches to ALS/MND

In the early 1990s, riluzole, a putative blocker of glutamate release, was established as the first disease-modifying treatment of ALS by improving survival of the patients [Lacomblez et al. 1996; Bensimon et al. 1994]. After initial controversies on the size of its effect, riluzole soon became the gold standard for each new candidate drug. Since an effect of riluzole on functional outcome measures could not be demonstrated in both studies, survival was widely accepted as primary endpoint of interventional drug studies in ALS. It was also a rational and ethical consequence that pharmaceutical trials were designed as add-on trials to riluzole and powered in a way that the effect of riluzole on survival was reached at minimum. The worldwide acceptance of riluzole was facilitated by retrospective studies which showed that early treatment seemed to have a better effect [Riviere et al. 1998] and that retrospectively the survival rates in large established cohorts in Europe and North America have increased after the introduction of this drug [Traynor et al. 2003; Turner et al. 2002; Brooks, 1999]. These positive effects on survival exceeded 1 year, contrasting with the size effect of riluzole in the large study which was estimated to be 3–4 months. However, it is unknown if proportion of this ‘increase’ is biased by patient selection including earlier diagnosis, alterations of interactions between patients and their physicians, by symptomatic therapy, by center effects or all of these aspects. Recent results of negative intervention studies (e.g. Corcia and Meininger (2008)) clearly demonstrate that the death rates, as observed 1 year after inclusion into the trials, have declined to less than a third observed in the riluzole studies. This shows that the cohorts in which drugs are tested today differ greatly from those of the early studies like the riluzole trial. In conclusion, since the introduction of riluzole, a number of attempts were made to improve its effect, but until today each of the therapeutic trials in humans remained nonsuccessful.

The challenges of drug discovery in ALS/MND

Since the mid 1990s, the field of translational animal research in ALS/MND was largely built on the use of a commercially available transgenic animal model for anterior horn cell loss carrying a large number of toxic human mutations in the Cu/Zn SOD gene [Rothstein, 2003; Gurney et al. 1994]. It seemed to be a major advantage of this model that it did not only reproduce the clinical and neuropathological hallmarks of ALS/MND, but also had a rapid course. The latter made it possible to screen compounds rapidly. Until now, more than 100 (published and unpublished) compounds were tested with apparent success in this animal model; however, many attempts to translate these experimental studies directly into therapeutic strategies in humans failed [Ludolph, 2006; Shefner et al. 2004]. To improve the success rate and to decrease unjustified investments into human and financial resources, in The Netherlands, March 2006, clinical and preclinical researchers met at a workshop in order to define guidelines for the use of standardized methods for drug testing in mouse models.

The workshop members agreed to the following three principles on which the more detailed guidelines were built [Ludolph et al. 2007]:

1. It was necessary to develop guidelines for preclinical research, since a number of experimental drug-finding studies fall out to be unreliable for translational research, representing the step from mice to men. The guidelines are to be developed to improve the quality of these studies and consequently save resources for both, preclinical and clinical studies.

2. It was decided to consider ‘proof of concept’ and ‘preclinical studies’ separately. A proof of concept study has the goal to elucidate the mechanism of the disease, either on a biochemical or physiological background. Such a study may use a drug as an investigational tool. A preclinical study has the primary goal to develop a drug for use in humans and must be subjected to more rigid methodological considerations, including toxicity and dose-response investigations.

3. It was expressed that the presence of one model representing only a single etiological factor of ALS/MND is not satisfying. The development of additional models based on etiological factors other than SOD1 mutations is warranted, in particular to have the opportunity to validate more accurately the targets of disease-modifying drugs independently of the etiology of the disease in the model organism.

During this meeting, guidelines for experimental animal research were developed and published [Ludolph et al. 2007] but doubtlessly will require further development. These thoughts and their consequences are likely to be employed in a similar fashion to translational research in other neurodegenerative diseases.

Prevention and treatment of ALS/MND

In analogy to other neurological diseases such as stroke and multiple sclerosis, recent experimental and clinical results suggest that prevention of the disease ('treatment in the preclinical phase') may be different from clinical strategies to modify the natural history of ALS. Experimentally, the results of the Cleveland group currently have a major influence on the development of therapeutic hypotheses in the field of ALS/MND. They could demonstrate, on the basis of an experimental design, that the effects of a single drug on disease pathology are depending on the time the drug is administered and are distinct in various stages of the disease. This hypothesis refers to the assumption that onset of disease is traced back to the vulnerability of the motor neuron itself whereas progression of the clinical disease may be the consequence of immune responses as immunoinflammation in the scope of inadequate neuroprotective or neurodegenerative processes, especially activation of microglia [Boillee et al. 2006; Clement et al. 2003]. In brief, this group showed that cell death in models of ALS/MND is non cell-autonomous. More specifically, the experimental onset of the disease is determined by toxic influences inherent to motor neurons themselves whereas progression of the clinical disease may be more dependent of neuroinflammation, in particular activation of microglia. It has also been shown by this group and others [Keller et al. 2008; Marchetto et al. 2008; Yamanaka et al. 2008] that the influence of astroglia is of major relevance for the clinical disease and therefore also a target for therapeutic interventions. Certainly, these insights must be dissected in the future; to find specific targets for therapeutic interventions, it is necessary to delineate specifically which aspects of the inflammatory reaction and which protective properties of astrocytes might be efficient and useful drug targets.

Do we have evidence that the hypothesis of a non-cell autonomous pathogenesis might not only be relevant in experimental animal models but also for human ALS? The answer is, at first sight, positive. Recent studies on the therapeutic and preventive capacity of vitamin E are complementary to the results of these studies. Whereas prospective interventions with high doses of vitamin E during the clinical course of the disease had no effect on survival of ALS patients in two independent studies [Graf et al. 2005; Desnuelle et al. 2001], retrospective analyses of two ‘preclinical’ healthy cohorts [Veldink et al. 2007; Ascherio et al. 2005] showed that intake of moderate vitamin E dosages decreased the risk of ALS/MND up to the astonishing amount of 63%. Moreover, it could be shown in experimental animals that an early intervention with vitamin E slowed down the preclinical course of the disease, indicating that alpha-tocopherol intake cannot prevent disease progression but seems to modify individual susceptibility of disease onset [Bucher et al. 2007]. Accordingly, the majority of the current clinical trials attempt to modify neuroinflammation or try to enhance the protective properties of astrocytes to influence the natural history of the disease.

Recent clinical trials

The recent negative minocycline trial [Gordon et al. 2007] deserves mentioning because of a number of reasons. First, the trial was based on the rational of a number of published positive animal studies [Kriz et al. 2003, 2002; Zhang et al. 2003; Van Den Bosch et al. 2002], but in sharp contrast did not show a positive effect in humans and even showed negative effects in minocycline-treated patients in major outcome measures [Gamez, 2008]. It is possible that the treatment strategy to use the highest tolerated dose in the individual patient led to this disappointing result by allowing the drug to reach toxic levels; dose-response curves are often u-shaped because of the toxicity of a drug. The u-shaped dose-responses are related with the hormesis theory: hormesis characterizes the dose-response continuum as stimulatory at low doses and inhibitory at high doses, leading to the biphasic, hormetic dose-response curve. Since the late 1890s it has been recognized that u-shaped dose-response curves might occur as a response to a disruption in homeostasis [Zhang et al. 2008]. Cell-based u-or inverted u-shaped responses may derive from common processes involved in activation of adaptive responses required to protect cells from stressful environments. These adaptive pathways extend the region of cellular homeostasis and are protective against ultimate cell, organ and system toxicity.

However, the activation of stress responses, such as an increase in glutamate concentration, carries a significant energetic cost to the cell, leading to alterations of a variety of basal cellular functions in adapted or stressed cells. At low levels of disruption or toxicity many biological systems display an overcompensation or rebound-phenomenon response, which results in the apparent low-dose stimulation component of the response curve. At higher doses with greater initial toxicity, the system often displays a more limited capacity for a compensatory response, usually insufficient to return to control values [Calabrese, 2008; Mattson, 2008].

In this context, it is probable that metabolic homeostasis of ALS patients has been changed during disease progress with the consequence of an increased sensitivity, including paradoxic reactions on drug doses that are compatible to healthy controls. Also, the trial shows that not only strategies of experimental drug treatment in mice require guidelines but also the design of clinical trials may need refinement: the seemingly firmly established belief that one of the major functional outcome measures of ALS trials, the ALS-FRS score may be linear was shown to be incorrect in this trial. Also, the fashionable ‘lead in design’ was shown to be misleading. In principle, the concept of the ‘lead in design’ is of major interest for ALS trials since this way of showing individual functional decline before the trial increases the power and potentially decreases the costs. However, the minocycline trial clearly showed that this ‘lead in phase’ is not necessarily predictive.

A recent observational study by Luc Dupuis and collaborators also deserves conceptional attention. Based on their own findings in experimental animals with an MND which showed a survival benefit of a high energy diet [Dupuis et al. 2004], the group retrospectively analysed lipid metabolism in 369 ALS patients and compared them with controls [Dupuis et al. 2008]. Hyperlipidemia (plasma levels of cholesterol or LDL) and steatosis of the liver were more frequent in ALS patients when compared with 286 healthy controls or patients suffering from Parkinson's disease. The results of this cohort study showed also that an abnormally elevated LDL/HDL ratio increased survival by more than 12 months - counteracting current medical ‘intuition'. These results indicate that - like in other neurodegenerative diseases - metabolic factors may be important in ALS/MND and are therapeutically underestimated. To confirm these pilot results and to transform this into clinically relevant therapeutic strategies, a prospective double-blind, placebo-controlled intervention trial is warranted.

Arimoclomol, a pharmacologically improved analogue of biomoclomol has been shown to amplify heat shock protein gene expression with an elevation of the respective HSP levels [Kieran et al. 2004]. Assuming protein misfolding with consecutively oligomerization as a relevant pathological event in ALS, Cudkowicz and colleagues performed a double-blinded, placebo-controlled as well as open-label study to investigate the safety and tolerability of this drug in patients with ALS [Cudkowicz et al. 2008]. Apart from well compatibility, the small number of participants and short duration of treatment eliminate the possibility of valid, definitive efficacy analyses, so that further phase IIb and III studies are necessary to confirm or disprove the promising effects seen in the preceding animal model of SOD^G93A mice [Kalmar et al. 2008].

Trophic factors, including recombinant human insulin-like growth factor I (rhIGF-I), are possible disease modifying therapies for ALS. In a North American study [Lange et al. 1996], participants on placebo were compared with those on recombinant human insulin-like growth factor. The analysis of the randomized clinical trial showed a weighted mean difference after 9 months in favour of the treated group. However, data from a European trial were not as conclusive and did not clearly demonstrate the efficacy of rhIGF-I in reducing the symptoms of the enrolled ALS patients [Borasio etal. 1998]. Recent data from a randomized, double-blind, placebo-controlled study could not reproduce the promising results of the North American study [Sorenson et al. 2008], implying that the available randomized placebo controlled trials do not show clinical efficacy of rhIGF-I on ALS. Such controversies of ongoing studies raise not only the actual problem of inter-study reproducibility, but raise also the question of how to solve this issue. From our point of view, the lack of biomarkers providing information on efficacy at each point in the progression of the disease is one of the greatest drawbacks for clinical trials. Those markers may provide information of the pharmacodynamics and bioavailability of the drug being tested. A recent phase II clinical trial of the immunomodulator copaxone also permits major insights into the problems of translational research [Meininger et al. unpublished]. In transgenic animal studies nonconsistent results were obtained; in a first attempt, a model which expresses a lower copy number of the Cu/Zn SOD mutations was used to avoid the strong neurotoxic effect of conventional higher copy numbers [Angelov et al. 2003]; the riluzole effect and two dosages of the MAO B inhibitor rasagiline served as positive controls [Waibel et al. 2004]. This study showed a convincing positive effect. However, in the animals expressing a high number of copies of the G93A Cu/Zn SOD mutation the disease-modifying was negligible [Habisch et al. 2007]. This makes clear that (1) the genetic background of the experimental animals is important; (2) the strength of the pathogenetic ('toxic') principle may influence the result of the experimental treatment. It is interesting to note here that similar observations were made during the preclinical development of copaxone for multiple sclerosis [Haenggeli et al. 2007], an indication in which the drug is extremely successful in humans.

The analysis of the clinical cohort used in the human copaxone trial showed that survival rates were much higher than previously observed, whereas survival 1 year after inclusion in the large riluzole trials was only close to 50%, this number slightly exceeded 10% in the verum (copaxone) and placebo stratum alike. It is evident that the major reason ‘for the change of the prognosis’ of ALS was due to the fact that patients were diagnosed earlier and included into the trial at an earlier disease stage; it is unclear at the moment to which extent improvements of symptomatic treatment measures contributed to these results. However, whatever the explanation for this striking change may be, the improved survival for patients in clinical trials makes the outcome measure ‘survival’ more difficult to use and will increase the costs of clinical trials. It should also be considered that these different patient groups may theoretically react differently to interventional treatments (for summary of all studies mentioned in the text, see Table 1).

Table 1.

Human pharmacological trials in amyotrophic lateral sclerosis. Summary of all clinical trials mentioned in the text with indication of the exact doses applied, the outcome measure, duration of intake as well as results.

Drug and doses	Outcome measure	Duration of intake	Results	Remark	Reference
Alpha-tocopherol 500 mg/day	Deterioration of function, measured by the modified Norris limb scale; survival	12 months	No effect of survival or motor function	Patients with riluzole plus alpha-tocopherol remained longer in milder states of the ALS Health State scale	Desnuelle et al. (2001)
Alpha-tocopherol 5000 mg/day	Survival, calculating time to death, tracheostomy or permanent assisted ventilation, according to the WFN-Criteria of clinical trials	18 month	No influence by alpha-tocopherol intake	Functional assessments showed a marginal trend in favour of vitamin E, without reaching significance.	Graf et al. (2005)
Alpha-tocopherol 15 mg/day	Risk of developing ALS	10 years	Less than half the risk of death from ALS than non-users.	Vitamin E supplementation could have a role in ALS prevention	Ascherio et al. (2005)
Alpha-tocopherol 22 mg/day	Risk of developing ALS	12 month	High intake of vitamin E is associated with a 50-60% decreased risk of developing ALS	Intake of moderate vitamin E dosages decreased the risk of ALS/MND as a preventive measure	Veldink et al. (2007)
Minocycline 400 mg/day	Difference in rate of change in the revised ALS functional rating scale (ALSFRS-R)	9 month	ALSFRS-R score deterioration was faster in the minocycline group than in the placebo group	Minocycline has a harmful effect on patients with ALS	Gordon et al. (2007)
High energy diet	Investigation whether lipid contents may have an impact on disease progression and survival	12 month	Frequency of hyper lipidemia as revealed by increased plasma levels of total cholesterol or LDL was twofold higher in patients with ALS than in control subjects.	Hyperlipidemia is a significant prognostic factor for survival of patients with amyotrophic lateral sclerosis.	Dupuis et al. (2008)
Arimoclomol 300 mg/day	Safety and tolerability	9 month	Well tolerated and safe	Arimoclomol was shown to be safe. It crosses the blood-brain barrier.	Cudkowicz et al. (2008)
IGF-1 0.1 mg/kg/day	Safety and efficacy	9 month	No significant difference between treated groups	IGF-1 appeared to be safe and well tolerated.	Borasio et al. 1998)
IFG-1 0.1 mg/kg/day	Change in the manual muscle testing score	2 years	No significant difference between treated groups	More adverse reactions with thromboembolic events in the treated group	Sorenson et al. (2008)
Lithium 300 mg/day, plasma levels ranging from 0.4 to 0.8 mEq/liter	Progression of the ALSFRS and Norris scale	15 month	Lithium delays ALS progression in human patients.	Decreases in ALS-FRS-R and Norris scales were not statistically significant in the group treated with lithium.	Fornai et al. 2008)

Upcoming clinical trials

A recent pilot study has shown a striking disease-modifying effect of lithium therapy in mice and men [Fornai et al. 2008]. However, the mouse study has some unusual methodological aspects (including an extremely short survival time of the nontreated controls) and not surprisingly, the results of the experimental research could not be reproduced [Bendotti et al. pers. comm.]. Also, the number of patients in the nonblinded clinical trial was comparatively small and the controls were historical ones. Therefore, there is a clear need to validate the results of this pilot study in a prospective, randomized, doubleblind drug trial. This publication shows the danger of methodological questionable research for human health; the dangers and side effects of lithium therapy in elderly patients that do not have regular fluid intake are well-known and may be life-threatening for individual patients.

Also, currently attempts are made to translate the positive experimental effects of pioglitazone as shown in two solid experimental animal studies from mice to men in a phase IIa study [Kiaei et al. 2005; Schutz et al. 2005]. In a phase II trial the modifier of astrocytic function, Ono-2506, seemingly showed a positive therapeutic effect on the survival of a subgroup of patients dependent on the timing of drug application. Currently, an attempt is made to reproduce these findings in the European SIRONA trial. This trial is an interesting example for a time-adapted therapeutic approach to the human disease which might reflect modern views of the pathogenesis. In this specific trial, patients need to be recruited in early phases of the disease, resulting in hitherto unknown challenges for the diagnosis of the disease and patient recruitment. AMPA (a-amino-3-hydroxyl-5-methyl-4-isoxazolepropionate) antagonists are long-standing candidates for interventional trials in ALS/MND [Kuner et al. 2005; Spencer et al. 1986]. Consequently, they have been successfully tested by several groups for efficacy in experimental animal studies and should be analyzed for their therapeutic effects in humans suffering from ALS. Currently, the AMPA antagonist talampanel is being tested in a large study in Europe and North America. Also, the surprising and innovative observations of the effects of nutritional interventions in mice and the positive effect of hyperlipidemia in humans require confirmation in large interventional trials.

In summary, although major problems were seen in translational research and drug development in ALS/MND, the field can build its future on an extensive experience with experimental in vitro and in vivo research and also on experience with the design of large clinical international drug trials and an excellent recruitment record. Therefore in our view, the chances to develop disease-modifying drugs in ALS are excellent and these developments may serve as role models for other neurodegenerative diseases.

Conflict of interest statement

None declared.