In 1907, Alois Alzheimer described a 51-year-old female patient who had developed a rapid loss of memory and had become disoriented in time and space. As the illness progressed, she became bedridden and incontinent and died four and a half years after the onset of illness. An examination post-mortem revealed an evenly atrophic brain with striking neurofibrillary pathology. Alzheimer also described the presence of unusual deposits in the cortex that were refractory to staining1.
Definitive diagnosis of AD occurs during post-mortem examination upon detection of two hallmark pathologies. The first is amyloid plaques, which consist of β-amyloid (Aβ). The length of Aβ can vary, but a 42-amino acid variant (Aβ42) is considered neurotoxic due to its propensity to readily aggregate into oligomers and fibrils9. The folding of Aβ into neurotoxic oligomeric, protofibrillar, and fibrillar assemblies is hypothesized to be the key pathologic event in AD. Aβ is formed through cleavage of the Aβ precursor protein by two endoproteinases, β-secretase and γ-secretase, that cleave the Aβ N-terminus and C-terminus, respectively5. The second pathological hallmark is the appearance of intraneuronal aggregates composed of the microtubule-associated protein tau9.
Presently, the only approved therapies for AD are the cholinesterase inhibitors (ChEIs) and an N-methyl-D-aspartate (NMDA) receptor antagonist, which increase cognitive function, but do not slow the rate of decline31.
Developing Novel Treatments
Discovering new pharmacological therapies is not a linear process. The development can cycle around several different steps before a medication comes onto the market. The first step is identifying the area of disease where there is a medical need that is not met, in this case to identify an Alzheimer's disease medication that can slow the rate of decline of the disease and maintain the cognitive function of sufferers. The next step is to identify a target for new medication. There are several targets under study in Alzheimer's therapy, for example the Aβ protein itself5, or the γ-secretase enzyme that modulates Aβ formation29. Chemical leads can be built upon compounds that already exist8 or can be an endogenous ligand5. Compounds are then screened to find those that have the best efficacy or selectivity for a receptor8. Once a compound has been identified, pre-clinical studies are used to test other facets, such as safety and absorption13. If pre-clinical studies are successful, the compound moves forward to clinical trials. Phase I is the first time a new compound is tested in human subjects and is designed to determine safety, tolerance, pharmacokinetic and pharmacodynamic properties, and early indications of efficacy14. Phase II investigates the effects of the drug in subjects with Alzheimer's disease. The aim is to demonstrate it's clinical effectiveness. This depends upon outcome measures. Many Alzheimer's disease studies use analyses of cognitive function as outcome measures in the absence of a viable biomarker4. Positive outcome in Phase II trials leads to Phase III, which assess the safety and efficacy in the target group in thousands of patients38.
Oral administration is a popular route owing to its convenience37. Absorption is slower and less complete than other forms of administration and some drugs are subject to extensive first pass metabolism in the liver or gut wall prior to reaching the systemic circulation, limiting their bioavailability37. Intravenous administration offers immediate and complete absorption, however the risk of serious adverse events and acute allergic reactions (caused by high peak concentrations) is increased37.
Transgenic mouse models are essential in studying the pathology of Alzheimer's disease9. β-amyloid precursor protein (APP)-overproducing mice develop amyloid deposits similar to those found in the human brain, the amount of which increases with age9. Despite chronic APP production, plaques typically accumulate in mid-to-late adulthood in the majority of these animals, although this is largely dependent on gene expression levels or the number of mutations introduced9. Aβ plaques found in the brains of Alzheimer's disease transgenic mice appear structurally similar to those found in the human brain; they initiate as diffuse plaques consisting mainly of Aβ42, develop a dense Aβ42 core, and then incorporate Aβ40 as well as numerous other non-Aβ components such as ubiquitin and α-synuclein10.
The Aβ cascade hypothesis predicts that tau hyperphosphorylation occurs as a downstream consequence of Aβ accumulation. APP-overexpressing transgenic mice have provided evidence both for and against this. APP-overexpressing models do not develop neurofibrillary tangles, yet many do show detectable tau hyperphosphorylation11.
The development of prevention therapies for Alzheimer's disease would greatly benefit from biomarkers that are sensitive to the subtle brain changes that occur in the preclinical stage of the disease. Reductions in the cerebral metabolic rate of glucose (CMRglc), a measure of neuronal function, have been studied for use as a tool in the early diagnosis of AD. In vivo brain 2-[18F]fluoro-2-Deoxy-D-glucose-positron emission tomography (FDG-PET) imaging demonstrates consistent and progressive CMRglc reductions in Alzheimer's disease patients, the extent and topography of which correlate with symptom severity30. There is also increasing evidence that hypometabolism appears during the preclinical stages of Alzheimer's disease and can predict decline years before the onset of symptoms30. Currently, there are no available tests for the definitive diagnosis of Alzheimer's disease in vivo, the clinical diagnosis remains a behavioral diagnosis after the exclusion of other causes, limiting the potential for early intervention and prevention research30.
Oral Small Molecules
Drugs that modulate the γ-secretase enzyme cleaving of β-amyloid precursor protein (APP) to release the different forms of Aβ are candidates for treatment of Alzheimer's disease29. An example is tarenflurbil (formerly R-flurbiprofen), a selective Aβ42- lowering agent that has been shown both in vitro and in vivo to modulate γ-secretase activity and reduce Aβ42 production in favor of shorter less toxic forms of Aβ (eg, Aβ38 and Aβ37)26,27. In mouse models of Alzheimer's disease, tarenflurbil was able to prevent learning and memory deficits and reduce Aβ42 brain concentrations26,28.
Based upon these promising results, a multicenter, randomized, double-blinded, placebo controlled trial of patients with mild Alzheimer's disease was conducted at 133 trial sites in the United States between February 21, 2005, and April 30, 200829. Concomitant treatment with cholinesterase inhibitors or memantine was permitted in this study, whilst Tarenflurbil, 800 mg, or placebo, were administered twice a day. The co-primary efficacy end points were the change, from baseline to the 18th month of the trial, in the total score on the subscale of the Alzheimer Disease Assessment Scale−Cognitive Subscale (ADAS-Cog, 80-point version) and Alzheimer Disease Cooperative Studies-activities of daily living (ADCS-ADL) scale. Additional prespecified slope analyses explored the possibility of disease modification. Tarenflurbil had no beneficial effect on the co-primary outcomes using an intent-to-treat analysis, whilst no significant differences occurred in the secondary outcomes. The tarenflurbil group had a small increase in frequency of dizziness, anemia, and infections. Tarenflurbil did not slow cognitive decline or the loss of activities of daily living in patients with mild AD29.
It is possible in drug development to manipulate an existing molecule so that it has greater efficacy. Russo et al8 studied Serotonin 5-HT4 receptor (5-HT4R) agonists, which are of particular interest for the treatment of Alzheimer's disease because of the positive effects of 5-HT4R on learning and memory performances35, whilst also regulating the production of Aβ36. The study looked at synthesising a 5-HT4R agonist that was more potent and selective than those agonists already existing. In the study, two libraries of molecules based upon the scaffold of ML10302, a highly specific and partial 5-HT4R agonist, were synthesised and then their binding affinities and agonist properties evaluated. In vivo, the two best performing compounds exhibited neuroprotective activity by increasing the level of the soluble form of the amyloid precursor protein (sAPPR) in the cortex and hippocampus of mice. One of these compounds could also inhibit Aβ fibril formation in vitro.
Immunotherapy
Immunotherapy involves the activation of cell-mediated or humoral (antibody) immune responses to eliminate noxious agents from the body5. Two classes of immunotherapeutic intervention have been explored: active and passive immunization5. In the former class, various types of Aβ immunogens are used to elicit endogenous Aβ-specific antibody production5. In the latter class, antibodies are produced exogenously, e.g., through monoclonal antibody methods, and then administered passively to the affected host5.
Schenk et al.6 published the first report of successful removal of parenchymal amyloid plaques in the PDAPP transgenic mouse model of AD actively vaccinated against fibrillar Aβ42.
Based upon the effectiveness shown by active anti-Aβ vaccination experiments in mouse models, and positive preclinical safety data in guinea pigs, rabbits and monkeys, Elan Pharmaceuticals began a small single-dose phase I study in 200013. In this study, 24 patients received an intramuscular injection of aggregated human Aβ1-42 plus QS-21 adjuvant (AN1792+QS-21), which was apparently well tolerated14. Consequently, vaccine safety, tolerability and immunogenicity were assessed in a multidose phase II trial that included 80 patients who were immunized with variable doses of AN1792+QS-21 or QS-21 alone, followed by up to four more injections of a modified vaccine containing a polysorbate-80 preservative (AN1792+QS-21/PS-80)14. Antibody titer responses (anti-AN1792 > 1:1000) were reported in approximately 59% of patients by the end of the vaccination schedule14. Although 24% of treated individuals reported some treatment-related adverse effects, such as accidental injury, infection and confusion, none of these were found to be related to dose and incidence of AN1792 administration14. Additionally, vaccinated patients showed less decline than placebo controls on the Disability Assessment for Dementia test after 84 weeks of treatment14.
A multicenter, randomized, double-blinded, placebo-controlled phase IIa trial of AN1792 began in October 2001, involving 372 patients (50-85 years old) diagnosed with probable Alzheimer's disease, with vaccine dosing scheduled every three months for one year13. The patient cohort was divided into two groups, 300 individuals received AN1792+QS-21/PS-80, whilst the remaining 72 patients were given placebo and evaluated for biomarker concentration, brain volume, cognitive performance and day-to-day functioning as measures of treatment efficacy15. Study dosing was however halted in January 2002, following the report of aseptic meningoencephalitis in 18 patients (6%) who had received one to three vaccinations15-17. All of the affected individuals belonged to the active treatment group, although no correlation could be made between antibody titers or the total number of injections and symptom severity and rate of relapse17.
Following the premature end of the AN1792 phase IIa trial, postmortem examinations of the brains of 12 immunized patients (2 of which developed meningoencephalitis) have described a focal absence of diffuse and neuritic Ab-positive plaques throughout the neocortex, the lack of dystrophic neurites, as well as the presence of activated microglia or macrophages associated with Aβ deposits in nearly all cases18-22. Very extensive to nearly complete Aβ removal was reported in four individuals21-22. Although cerebrospinal fluid (CSF) tau concentrations were significantly decreased in antibody responders, no changes were observed in the number of tau-positive neurofibrillary tangles, while the severity of cerebral amyloid angiopathy and associated microhemorrhages were either unaltered or increased18-21,23. In the two patients who developed meningoencephalitis, inflammatory T-cell infiltrates were observed throughout the brain, as well as white matter infiltration of CD68-immunoreactive macrophages18,19.
Rozkalne et al3 studied the long term effect of a single dose of passive immunotherapy upon a mouse model. To determine whether Aβ plaques are present 30 days after a single dose of anti-Aβ antibody treatment, plaques were labelled with R1282 immunostaining and the plaque burden calculated in treated and untreated areas of the cortex. Treatment with anti-Aβ antibody 3D6 did not lead to a statistically significant reduction in plaques observed 30 days post-treatment, however there was a reduction in the number of axonal dystrophies per plaque and an increase in synapse density. The authors believed that this indicated that despite early clearance of plaques with passive immunotherapy, there is a probable re-deposition within 30 days after a single dose of monoclonal antibody in that particular mouse model.
Despite the variability in functional outcome between vaccinated cohorts, the apparent effectiveness of anti-Aβ antibodies in reducing plaque load in distinct brain areas has encouraged the continued use of immunotherapy for the treatment of AD13.
Bapineuzumab, a humanized anti-Aβ monoclonal antibody, has been evaluated in a multiple ascending dose, safety, and efficacy study in mild to moderate Alzheimer's disease4. This study of 234 patients, randomly assigned to IV bapineuzumab or placebo in 4 dose cohorts (0.15, 0.5, 1.0, or 2.0 mg/kg). The trial took place over 78 weeks with patients receiving 6 infusions, 13 weeks apart. The study looked for significant differences within dose cohorts using the Alzheimer's Disease Assessment Scale-Cognitive and Disability Assessment for Dementia scales. No significant differences were found in their primary efficacy analysis, however exploratory analyses showed potential treatment differences on the cognitive and functional endpoints in those who had completed the study and those that do not carry a risk allele for the condition known as APOE ε4. Reversible vasogenic oedema, detected on brain MRI in 12/124 (9.7%) bapineuzumab-treated patients, was more frequent in higher dose groups and APOE ε4 carriers. Whilst their primary efficacy outcomes in this phase II trial were not significant, bapineuzumab progressed to a phase III clinical with special attention to APOE ε4 noncarrier status38.
Asuni et al.32 published the first report on tau immunotherapy. They used the transgenic P301L mouse, which expresses an FTDP-17- associated tau mutation and develops pathology in the motor cortex, brainstem, and spinal cord33. Using a peptide corresponding to residues 379-408 of tau, with the phosphorylation of Ser396 and Ser404, two phosphoserine residues commonly associated with neurofibrillary tangles, they were able to show that animals immunized with the immunogen for two to five months demonstrated reductions in insoluble tau and increases in soluble tau in immunohistochemical and biochemical analyses. Behavioral analyses using the rotarod and traverse beam showed improved performance after immunization as compared to controls treated with adjuvant alone.
Conclusion
In order for an AD drug or therapy to be approved, there is a current requirement that the therapy slow cognitive decline in a therapeutic setting and be safe12. Given the slow and variable rate of decline especially among patients with mild AD or mild cognitive impairment of the AD type, a definitive double blind placebo controlled, phase 3 trial requires thousands of patients and at least 12-18 months of testing12. Such trials typically cost upwards of $50 million dollars12. The increasing complexity and cost of these trials reduces the number of different therapies that can definitively be tested in humans with Alzheimer's disease. It is becoming increasingly important to define what counts as efficacy in a clinical trial, be a it clear slowing of cognitive decline, a modulation of a defined biomarker, or some combination of the two. It is also important to consider that therapeutic studies of agents targeting tau and Aβ may have limited benefit, or in the case of bapineuzumab, may only benefit a small subsection of patients with the condition.
The outcomes of clinical trials are also influenced by the stage of the disease targeted. The use of anti-Aβ immunotherapies may prove to be most useful for those patients with mild cognitive impairment (MCI) and other presymptomatic individuals who are under environmental or genetic risk of developing Alzheimer's disease13. There is some evidence suggesting that patients with MCI steadily progress to greater stages of dementia and have 30% less neurons in the entorhinal cortex, compared to nondemented controls24, 25. Therefore, targeting anti-Aβ therapeutics toward at-risk individuals may prove effective at slowing and/or reversing disease progression13.
Alternatively, it may be that 'cocktail' or multimodal therapy may be needed to significantly impact the clinical course of Alzheimer's disease. For example, targeting tau, Aβ, inflammation and cognitive symptoms, via a cocktail of drugs and therapies, may be more effective then monotherapy pertaining to just one of those areas12.
In order to fully treat the condition, it is important to fully determine exactly why neurons die in Alzheimer's disease. Every step in determining the pathology of the disease leads to new therapeutic targets or strategies. A major problem on the road to understanding the downstream pathological cascades initiated by Aβ is that we do not have animal models of Aβ deposition that fully replicate one of the primary features of Alzheimer's disease pathology, namely neuronal loss34. Whilst current models are clearly useful for target and therapeutic validation, they are unable to substitute for clinical trials in humans, which, as seen throughout this report, tends to be the stage in development where many of these therapies have fallen by the wayside.
Identifying individuals at risk for Alzheimer's disease and determining pre-morbid and disease-associated biomarkers may turn out to be just as important as a discovering a novel therapeutic strategy with respect to ultimately developing an effective treatment for the condition12.
