The Oxidative Stress Hypothesis Of Neurodegeneration Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

The only certain predisposition to neurodegenerative disease is age, and this has lead to proposals that there may be some common pathways between ageing and neurodegeneration. There have been many hypotheses describing potential causal mechanisms of ageing, but one of the more enduring has been Harman's 'free radical theory of ageing';1 this posits that endogenous production of reactive oxygen species (ROS) results in the accumulation of cellular damage, and subsequently, cellular senescence. ROS production is a constant fixture of living: oxidative stress is a function of metabolic rate, with normal oxidative phosphorylation producing superoxide anions with every cycle. The putative involvement of ROS in ageing lead to the oxidative stress hypothesis of neurodegeneration; this proposes that an imbalance in normal antioxidant defence and the generation of ROS results in cumulative destruction of vital neuronal structures, and eventually in neuron death and loss of functional abilities. The oxidative stress hypothesis is attractive as it links several other prominent theories of neurodegeneration: notably, the failure of mitochondrial bioenergetics, transition metal toxicity, aggregated protein toxicity, environmental toxins, neuroinflammation and excitotoxicity.

An enormous number and variety of in vitro studies have shown that it is possible to find evidence for ROS causing, or being caused by, almost any factor normally associated with the pathogenesis of a neurodegenerative disease, such as protein aggregation,2-8 mitochondrial dysfunction,9,10 or programmed cell death.11,12 It has become clear that oxidative stress is one small part of a hugely complex network of interconnecting cascades, all of which culminate in neuronal degeneration. Attempting to pin 'cause' or 'consequence' onto observations is a near impossible task: ROS have effects on almost all aspects of neurodegenerative pathology, and likewise many of these factors may reciprocally increase ROS levels. Therefore, in order to release the research field from its pursuit of demarcating primary from secondary events, it is imperative that we establish the quality of the basis of evidence upon which these studies are founded. Having done so, we may determine the relevance of oxidative stress to neurodegenerative disease by discussing its role in early pathology and evaluating the success of antioxidant therapies.

The oxidative stress hypothesis of neurodegeneration: how good is the foundation of evidence?

In the early days of the oxidative stress hypothesis, scores of studies were published using post-mortem CNS tissue to demonstrate raised markers of oxidative damage in neurodegenerative diseases such as Parkinson's disease (PD),13,14 Alzheimer's disease (AD),15-17 and amyotrophic lateral sclerosis (ALS).18,19 However, we still have no direct indicators of the level of ROS in tissue; this inescapable issue alone cautions the interpretation of all of these, and subsequent, studies. ROS are highly reactive species with very short half lives, and consequently are impossible to measure in a human in vivo. Although spin traps may be used to measure ROS in vitro, these molecules are unstable and vulnerable to modifications that may render them silent to measurement techniques.20 Consequently, the only markers that are practical and available are those which reflect the damage to cellular components caused by ROS. Many of these markers are not sufficiently reliable to produce robust conclusions, and three decades after these first measurements of oxidative damage in human neurodegeneration, there is still no single biomarker that fulfils all ideal criteria. For example, these studies often use 8-hydroxy-2-guanosine (8-OHdG) as a marker of oxidative DNA or RNA damage; however, simply isolating DNA from tissue exposes it to oxidative stress, which could artifactually increase levels of 8-OHdG.21 A similar problem arises in studies using 3-nitrotyrosine as a measure of nitrosative protein damage, as freezing tissue samples may induce further protein nitration.22 Furthermore, 8-OHdG itself is not a gold standard measurement; its generation depends upon the conditions surrounding the oxidative reaction.23 Other markers are non-specific: protein carbonyls, which are used in many of these studies as an indicator of oxidative protein damage, represent protein modifications which may be oxidative, but may also be due to aldehyde binding or protein glycation.24 More fundamentally, all of these studies have been carried out using post-mortem tissue. Long intervals between death and preservation or between death and the investigators receiving the tissue - features of many of these studies - may expose tissue to post-mortem oxidative stress and consequently introduce further artefacts; therefore, it is vital that these factors are accounted for in case-control matching. Finally, it is important to consider that observations derived from post-mortem tissue may not be directly transferrable to an in vivo disease state.

These issues problematise research into the role of oxidative stress in neurodegenerative disease. Approaches focusing on whether oxidative stress is causal, or simply a secondary consequence of other pathogenic processes, may be undermined by this lack of reliable biomarkers. An awareness of the limitations of these studies may counsel a shift in the focus of current research towards more clinically relevant considerations. It is clear that clinical improvement for patients is improbable whilst relying upon approaches which target the late stages of neurodegenerative disease; by this time, neuron death is likely to have reached an irretrievable level. Targeting the early stages of disease holds more promise for effective future therapy. In order to ascertain the relevance of the oxidative stress hypothesis, it is therefore crucial to evaluate whether oxidative damage is involved in the early pathology of neurodegenerative disease.

Is oxidative stress an early feature in models of neurodegeneration?

Whilst animal models have been invaluable in exploring the early pathogenic processes of many neurodegenerative diseases, they are particularly important in ALS. ALS is not normally associated with a prodromal phase, and consequently investigation into presymptomatic pathology is limited to animal and cellular models.

One such model of ALS is the SOD1 mouse, a model of familial ALS produced by transgenic overexpression of a mutated human superoxide dismutase 1 (SOD1) gene. Some studies using this mouse model of ALS have demonstrated an increase in oxidative damage markers before the onset of clinical symptoms. Hall and colleagues used a relatively reliable but non-specific marker, malondialdehyde (MDA), to detect lipid peroxidation in SOD1 mice. They found that MDA was higher in the transgenic mice than their non-transgenic littermates at 30 days, before any signs of motor neuron pathology or the emergence of symptoms.25 Similarly, Warita et al. found raised levels of 8-OHdG in SOD1 mice at a presymptomatic stage.26 Other studies have suggested that mitochondrial abnormalities are evident before the onset of symptoms in these mice,27,28 indicating that ROS generation may be increased before clinical disease develops.

These studies have provided support for the hypothesis that oxidative stress is a feature of the early stages of ALS. However, information extracted from models is not directly transferrable to humans. The SOD1 mouse - the mouse most commonly used to model ALS - does have some face and construct validity for investigating human familial SOD1 ALS, but these patients only represent 1-2% of all ALS sufferers, the majority of which have the sporadic disease with no evidence of changes in SOD1.

Is oxidative stress an early feature of human neurodegenerative disease?

Mild cognitive impairment (MCI) is widely considered to be a prodromal phase of AD.29 Patients with Down's syndrome (DS) have offered further insight into AD as these individuals commonly develop an early onset dementia that is histopathologically indistinguishable from AD.30 Early studies in MCI and DS were mostly carried out using post-mortem CNS tissue, or aborted DS foetuses, but later works have used urine, plasma, and CSF samples to record oxidative damage markers in vivo.

Brooksbank et al. performed an early study demonstrating that lipid peroxidation was significantly higher in the brains of DS foetuses than controls;31 however, in this study lipid peroxidation was measured not by markers, but by adding ascorbate and Fe2+ - two pro-oxidant compounds - to homogenised tissue samples. More relevant studies have been carried out in vivo in DS patients: Jovanovic et al. used 166 DS and non-DS twin pairs to demonstrate that levels of 8-OHdG and MDA were significantly raised in the urine of the DS patients.32 8-OHdG is currently the most reliable urinary marker of oxidative damage to DNA, though it is only semi-quantitative.33,34 However, MDA derived from the diet may be excreted directly from the gut into the urine,35 and consequently, with no provision for the control of participants' diet in this study, these measurements should be interpreted with caution. Nevertheless, studies such as these, when combined with the results of others,36,37 do indicate that increased oxidative damage may appear before the onset of neurodegenerative symptoms in DS patients, and add support to the theory that ROS are involved in early AD pathology.

Keller et al. used thiobarbituric acid reactive substances (TBARS) to show that lipid peroxidation was higher in the post-mortem brain tissue of MCI and early AD patients than in control brains.38 However, this result is uninformative, as TBARS are a non-specific measurement - indeed, most TBARS present in tissue fluid are derived from sources other than lipid peroxidation.39 Other studies have used more reliable markers of oxidative lipid damage, such as HNE or isoprostanes, and have concluded that lipid peroxidation is higher in MCI post-mortem brain tissue40,41 or in the CSF, urine or plasma of living MCI patients.42,43

Recently, more clinically relevant findings have emerged. Ansari and Scheff reported a negative correlation between the mini mental state examination (MMSE) score, an indicator of cognitive ability, and oxidative damage to proteins and lipids in MCI patients.44 Furthermore, Butterfield and colleagues used redox proteomics to show that oxidative damage to proteins in MCI may be specific.45 The damaged proteins they identified were involved in multiple processes, and included GLUL (implicated in synaptic plasticity and glutamate regulation), PIN1 (chaperone functions), and ENO1 and PKM2 (involved in metabolism). The authors proposed that oxidative damage to proteins such as these, and their subsequent inactivation, may lead to the progression from MCI to AD.

Incidental Lewy body disease (ILBD) is often described as a preclinical phase of PD,46 and consequently exploring biochemical indices of oxidative stress in ILBD could illuminate early pathological processes in PD. However, as ILBD is asymptomatic, it is usually discovered post-mortem, and therefore does not offer the opportunity - as with DS and MCI - of in vivo human experiments. Studies investigating the presence of markers of oxidative damage in the ILBD brain have reported mixed results.47,48

These studies in human prodromal states appear to corroborate the results of animal studies, indicating that oxidative damage is a feature of early neurodegenerative disease. However, these human studies rely on the assumption that all DS, MCI or ILBD cases included in the study would have developed AD or PD, respectively. This does not take into account the significant proportion of individuals with DS, MCI or ILBD who never progress from these prodromal states to a major neurodegenerative disease. Conclusions drawn from a cohort that includes those patients that would not have subsequently developed AD or PD cannot be interpreted as offering direct support for an early role of oxidative stress in neurodegeneration.

How effective have antioxidants been in human trials?

If oxidative stress was indeed an early and, therefore, causal feature of neurodegenerative pathology, antioxidants would offer the opportunity to target primary mechanisms of neurodegeneration and consequently delay the progression, or even prevent the onset, of disease. Hope for their efficacy was high after experiments in cellular and animal models indicated neuroprotective effects.49,50 However, once again, these models are imperfect; studies must be undertaken within a cost-effective time frame, and this may be achieved by the overexpression of transgenes to far greater levels than are found in the disease in vivo. For example, although Gurney et al. reported that vitamin E delayed disease onset in the SOD1 transgenic mouse model of ALS,51 the mutant protein was highly overexpressed and therefore the beneficial effect of vitamin E may have been falsely magnified.

Many trials have been conducted exploring the use of vitamin E as a treatment for PD, AD and ALS. Almost without exception, they have shown no beneficial effect of vitamin E, even when using varying doses.52,53 A single controlled trial in AD patients demonstrated a positive effect of vitamin E treatment on disease progression,54 but this result was only observed when the baseline MMSE score was included as a covariate; otherwise, there was no difference between the placebo and treatment arms. Even when administered in MCI patients, there was no effect of vitamin E on the progression to AD.55

The only studies that have shown any positive effects of vitamin E are prospective or cross-sectional studies, evaluating individuals' dietary intake. These have shown an association between a higher vitamin E intake and potential protection against cognitive decline,56 development of ALS,57 or AD.58,59 However, prospective and cross-sectional studies have limited relevance, as they are liable to the incorporation of many confounding factors. For example, it may be that patients that have a high vitamin E intake also ensure that their diet contains high levels of other vitamins, such as vitamin B, which may be linked to a lower risk of AD.60 Dietary intake of participants is usually assessed using a self-administered food frequency questionnaire, which is subjective and therefore not ideal for comparison. Furthermore, a prospective study with a longer duration found no effect of vitamin E on the development of AD in a cohort of healthy individuals.61

Selegiline, a monoamine oxidase B (MAO-B) inhibitor, may also have antioxidant effects and has been trialled in human ALS, AD and PD. In vitro experiments have shown that selegiline can increase the activity of the antioxidant enzyme SOD, and may also inhibit the production of ROS during DA metabolism.62 However, results from clinical trials have been mixed, and many of these trials are poorly designed. Mazzini et al. found no effect of selegiline on disease progression in ALS patients in an open study,63 and Mitchell et al. reported the same finding but had enrolled only 56 participants, of which approximately half completed the trial.64 Moreover, another more methodologically sound study also found no effect of selegiline on disease progression or survival in ALS patients.65 Some studies have shown positive results for selegiline in AD,66 but most have not.67 Many investigators did not undertake sufficient measures to minimise bias, with some studies being open or single-blinded,68-70 lacking appropriate control groups,70 or incorporating other variables such as the concurrent administration of other treatments.71,72

Selegiline has been used with some success in PD.52 However, its action as a MAO-B inhibitor suggests that results may be, at least in some part, due to effects on dopamine metabolism; positive outcomes in motor or cognitive assessments may be attributable to symptomatic treatment.

Thus, the quality of the evidence demonstrating a beneficial effect of antioxidant therapy in neurodegeneration may be called into question. Study design continues to be an issue in the interpretation of results: aside from inadequate randomisation and blinding, many studies are underpowered and fail to include sufficient follow-up periods. A lack of reliable biomarkers of oxidative stress, or even disease progression, prevents monitoring of the real effects of any administered drugs. Appropriate dosing and duration of treatment have not been resolved: for example, it may take many months of vitamin therapy in order to establish a potentially therapeutic concentration in the brain, yet treatment is often only administered for several weeks, with no attempt to establish a system for monitoring the concentration achieved. Furthermore, tests used to assess outcome measures such as cognitive function do not take into account that other factors related to the disease - such as depression, or motor and functional abilities - may affect the patient's motivation and ability during testing. Even when these methodological problems are taken into account, the sheer number of trials that have produced negative results suggest that current antioxidants do not hold much promise for the future treatment of neurodegenerative disease.

The oxidative stress hypothesis of neurodegeneration has been hailed by its advocates as offering a unifying theory for the pathogenesis of neurodegenerative disease. Although there is a wealth of literature supporting this theory, the interpretation of these positive results is limited by the absence of gold-standard markers of oxidative stress. Unfortunately, as the number of studies supporting a central role of oxidative stress in neurodegeneration has grown, these issues have tended to be overlooked, and the validity of these markers has been taken as established. Having become increasingly focused on assigning a primary or secondary role to oxidative stress in neurodegeneration, the research field risks overlooking wider questions of the relevance of oxidative stress to both the early pathology of neurodegenerative disease, and the implications for its treatment. The evidence for a role for ROS in the early pathology of PD, AD and ALS is mixed: despite studies demonstrating the presence of increased oxidative markers in prodromal conditions and presymptomatic animal models, methodological issues remain, and the impulse to infer direct support for the hypothesis from these results should be questioned. Furthermore, the results of trials of antioxidant therapies have demonstrated that the available methods of directly targeting oxidative stress in neurodegenerative disease are not effective. The expectation that antioxidants would be a successful therapeutic strategy was founded on evidence from early studies demonstrating the presence of oxidative stress in neurodegenerative disease, and later, studies showing that this was an early feature of the pathology. However, the quality of this basis of evidence, coupled with the failure of antioxidants in clinical trials, questions this expectation. It is now time to reappraise the rationale of continuing research into strategies employing antioxidants as therapies. Perhaps genetic and small molecular therapies focused on upstream events such as RNA processing or protein deposition would be a more fruitful avenue, and may offer a broader approach to the prevention of the pathological processes of neurodegeneration.