HM Medical Clinic

 

Animal models of alzheimer's disease and drug development

Editors-in-ChiefKelvin Lam – Simplex Pharma Advisors, Inc., Arlington, MA, USA Henk Timmerman – Vrije Universiteit, The Netherlands Animal models of Alzheimer's disease and drug development Bart LaurijssensFabienne , Anisur 1BEL Pharm Consulting, Moulin d'Ozil, 07140 Chambonas, France 2CNRS UMR 7179, MNHN, 1 Av du Petit Chaˆteau, 91800 Brunoy, France Animal disease models are considered important in the development of drugs for Alzheimer's disease. This Oscar Della Pasqua – Leiden/Amsterdam Center for Drug brief review will discuss possible reasons why their Research, Leiden, The Netherlands.
success in identifying efficacious treatments has been limited, and will provide some thoughts on the role of loss in brain regions involved in learning and memory pro- animal experimentation in drug development. Specifi- cally, none of the current models of Alzheimer's dis- The interest in finding a cure or prevention for AD is ease have either construct or predictive validity, and no understandably great. Proper animal models of human AD model probably ever will. Clearly, specific animal are considered desirable if not essential in this process and much research effort has been put into that effect. As no experiments contribute to our understanding of the perfect model exists, the question becomes whether ‘the best disease and generate hypotheses. Ultimately, however, models available' are good enough. What exactly can be the hypothesis can only be tested in human patients inferred from the results and what not? Or, differently put, and only with the proper tools. These tools are a how do they contribute to our understanding and decision- pharmacologically active intervention (in humans) The objective, therefore, of this brief review is to discuss the and a clinical trial suited to evaluate the mechanism potential role of the current animal disease models for AD in of action. Integration of knowledge in quantitative drug development. Specifically, the aim is to discuss why the (sub) models is considered important if not essential animal models of AD should have such a limited success in predicting successful treatments in the clinic (or rather, in this process.
clinical trials), and to provide some thoughts on how to use animal experiments in drug development.
IntroductionAlzheimer's disease (AD) is a neurodegenerative disease Animal models of Alzheimer's disease affecting an estimated 5.4 million people globally, mainly The aetiology of AD is unknown, but there is still a general the elderly. There is currently no cure for AD, but some consensus in favour of the ‘amyloid hypothesis' even if symptomatic treatments are available. The disease is charac- it has been questioned ]. A wide range of animal models terized by its hallmark histopathological findings of extra- have been developed to mimic the human context of the disease for the purpose of developing therapeutics or disease neurofibrillary tangles of tau and by neuronal and synaptic modifying agents. In fact, in most of the animal models the *Corresponding author.: B. Laurijssens ( first goal is to simulate the neuropathological findings of AD 1740-6749/$ ß 2012 Elsevier Ltd. All rights reserved.
Drug Discovery Today: Technologies Translational pharmacology Vol. 10, No. 3 2013 followed by the correlation of cognitive function without animal models reviewed in this section either show major knowing whether the neuropathological agents have similar AD-related neuropathology or are in close phylogenetic biological consequences in humans and in animal models. It proximity to humans. In recent years dogs have been con- is beyond the scope of this review to discuss all the different sidered as a useful animal model for AD due to the close models. Here we will briefly summarize some potential ani- proximity of canine and human brain aging. Dogs develop mal models and their translation towards the clinical settings extensive Ab and diffuse plaque deposition and the extent of (). Animal models used in AD can be broadly divided Ab deposition correlates with the decline of some measures of into three categories: Natural models, Genetic models and cognitive function in the absence of neurofibrillary tangles (NFTs). Moreover, the amino acid sequence of Ab is fully conserved between dogs and humans []. Because of sponta- neous age-related Ab deposition and relative ease of cognitive Several animals including polar bears, dogs, cats, goats and function assessment, aged dogs were used to evaluate inter- sheep and some non-human primates spontaneously develop ventions with the aim of reduction of Ab load and subsequent some AD-related neuropathological features []. The few cognitive improvement ( Table 1. Some animal models of AD and clinical trials Treatment active in model Active immunization with Ab Trial stopped due to CNS side effects ] Acetylcholinesterase inhibitor, Donepezil [] Nicotinic receptor agonist ABT-418 ] Discontinued for side effects Glutamate receptor blocker Memantine ] Mouse lemur, Octodon, No pharmacological trial has been reported yet.
Polar bear, Cat, Sheep, Active immunization with aggregated Ab Trial stopped due to CNS side effects ] Passive immunization with IgG [ Vasogenic oedema and microhaemorrhage in some patients, no significant cognitive benefit, Phase-III trials underway ] b-Secretase inhibitor [] Reduce plasma Ab significantly, Under phase-III clinical trial [ g-Secretase inhibitors ] Trial halted because the drug failed to stop disease progression, associated with worsening cognitive function and increased risk of skin cancer [] g-Secretase modulators ] No benefit, trial discontinued [ Acetylcholinesterase inhibitor [ Cognitive decline cannot be prevented, cognitive benefits with treatment associated adverse effects reported [ Non-steroidal anti-inflammatory drugs ] Negative results [ No cognitive benefit [ Ttpa / APPsw mice Together with Rifampin. Some positive effect on cognitive performance [ Interventional models Intrahippocampal amyloid infusion model Intracerebroventricular Ab infusion model Marginal benefits, larger trial ongoing ] a PDAPP mice: Mutation of human APP with valine at residue 717 substituted by phenylalanine under the control of human platelet derived growth factor promoter. APP mice: Mutation of three isoforms [695, 751 and 770] of human APP under the control of various promoters. Tg2576 mice: Swedish double mutation [K670N/M671L] on the 695 human APP under the control of hamster prion protein promoter. PSAPP mice: Swedish double mutation [K670N/M671L] plus Presenilin 1 exon 9 deleted. Tg-tau mice: Human normal shortest 3R tau isoform inserted into the mouse genome under the promoter. Ttpa / APPsw mice: Tg2576 mice crossed with a-tocopherol transfer protein knockout Vol. 10, No. 3 2013 Drug Discovery Today: Technologies Translational pharmacology Among the non-human primate models, mouse lemurs Another interesting finding is the presence of extensive astro- seem to be a potential animal model that exhibit amyloid gliosis in the aged octodon brain, which is a characteristic plaque, NFTs and some other AD related neuropathology ].
feature of human AD brain. It has been hypothesized that This small prosimian primate lives 8–10 years in captivity and high homology (97.5%) of octodon Ab and human Ab might shows age-related changes similar to those of aging humans be an important factor in the appearance of AD markers in []. About 20% of mouse lemurs aged five years or older show this rodent. Breeding difficulty and comparatively longer life- significant brain atrophy ], extensive accumulation of amy- span are limitations of this rodent model ].
loid plaques, neurodegeneration [] and/or loss of cholinergic neurons []. The genes responsible for the formation of senile plaques are highly similar between mouse lemurs and Transgenic technology provides unique opportunity to repro- humans The presence of amyloid deposition has been duce the cause of familial AD by transfecting a mutant human recently linked to cerebral atrophy in aged individuals, giving amyloid precursor protein (APP). Mice have extensively been new insight into the understanding of pathological aging in used as transgenic models and facilitated our understanding of this non-human primate species Moreover, a compar- the molecular mechanisms associated with Ab-production, able decline in declarative memory and executive function deposition and clearance and the effects of Ab on neuronal has been reported in both aged human and aged mouse lemur network and synapses that play important role in cognitive whereas procedural memory appears to be conserved in both function. The APP mouse model successfully produced a wide species []. In a recent study it has been shown that cogni- range of parenchymal and vascular amyloid deposits similar to tively impaired aged mouse lemurs have cerebral atrophy those of human AD []. Although morphological similarity especially those brain regions that are responsible for cogni- does exist, there is a difference in biochemical composition of tive functions []. It should be noted that there is a differ- deposited Ab between mouse models and AD brain More- ence in distribution of Ab deposits and plaques between over, these transgenic mouse models failed to develop neuro- human and mouse lemur. In humans the Ab depositions fibrillary tangles (NFTs), an important histopathological usually start in the hippocampus but in mouse lemur they hallmark of AD result from intraneuronal aggregation of appear first in cortical regions []. As with human AD, hyperphosphorylated tau protein ]. Oddo and colleagues currently no diagnostic tools are available that can predict presented for the first time a triple transgenic mouse model which adult mouse lemur will develop AD-like symptoms.
where both plaques and NFTs were found in AD-relevant brain However, this model provides an opportunity to search for regions. These mice also developed extracellular amyloid beta these predictors.
deposits before the formation of NFTs and exhibited impaired Rhesus monkeys and humans share many diseases of aging synaptic plasticity including long-term potentiation, the key and probably are the most successful models to identify basis for cognitive function. Furthermore, transgenic mouse diagnostic markers and the development of safe and effective models also provided valuable information regarding the role treatments for human brain disorders. Monkeys of over 19 of inflammation, oxidative stress and mitochondrial dysfunc- years old showed significant amyloid-plaque-like lesions in tion in the pathogenesis of AD However, progressive areas of brain responsible for cognitive function. However, neuronal loss in hippocampus and specific neocortical regions most of the aged monkeys do not show an Alzheimer's-like of the human AD brain [is not evident in most of the syndrome and age-related rapid cognitive decline commonly transgenic mouse models, and this is a major limitation of found in AD patients is not usually found in them either [].
these murine models. In addition, these transgenic mice repre- Moreover, their long developmental period, low reproductive sent only those who are suffering from familial AD, which is output, long captive life and risk of serious zoonotic disease <1% of all AD patients. There is no mouse model that can fully transmission are major disadvantages for using these animals reproduce the features of disease progression of vast majority of in a wider range of AD research AD cases that is sporadic/late-onset AD.
Octodon degu, a rodent of South American origin has The fruit-fly Drosophila melanogaster is a widely used and recently been found to have spontaneous development of well-appreciated animal model of neurodegeneration includ- AD-related neuropathology at older age. Localisation of both ing AD []. The fruit-fly is small in size, has a simple well- intracellular and extracellular amyloid deposits and NFTs-like studied anatomy and possesses a well-organized brain.
intracellular deposition has been detected in different layers Although the fruit-fly brain has only a fraction of the cells of cortex and hippocampus of old animals. Adult octodon of the human brain and a different neuroanatomical organi- cortex shows the presence of cholinergic neurons as in zation, it is similar in the fundamental aspects of cell biology, humans but with a different distribution. They also demon- in terms of regulation of gene expression, membrane traffick- strate age-associated cognitive impairment but whether these ing, neuronal connectivity, cell signalling, synaptogenesis cognitive deficits are also associated with cholinergic neuro- and cell death []. Moreover, its short life cycle and degeneration similar to AD patients is not known yet.
completely sequenced genomes are added experimental Drug Discovery Today: Technologies Translational pharmacology Vol. 10, No. 3 2013 advantages of this model. Importantly, the transparent cuti- Animal models of disease cle of the larvae of fruit-fly allows the study of the disease It is obvious that many of the animal models of AD have progression in living intact animals which is rather difficult contributed, and continue to contribute, to our understand- in vertebrates []. The task for associative learning and ing of the processes that may or may not underlie human AD.
memory such as Pavlovian olfactory conditioning can be But can they really be called animal models of AD, and can measured in this model, which is homologous to the classical they therefore be used as such? conditioning of the eyeblink response found to be impaired The perfect animal model of a disease would be a scaled in patients diagnosed with AD []. The fact that hippocam- down replica of the human disease, representing all impor- pal-dependant cognitive functions, which are impaired early tant components from cause via structural damage to symp- in human AD, cannot be tested in invertebrates due to the toms (). It would be structurally and quantitatively lack of these brain structures is a major limitation of this identical to the human. For this, both the animal model animal model. In spite of the lack in homology between fruit- and the human disease need to be very well understood.
fly and human brain, this simple invertebrate animal pro- Such a model would have face-, predictive- and construct vided insight in the disease mechanism ranging from genetics validity [] (see ). None of the models above satisfy this to some cognitive functions that could be followed in living criterion and no animal model probably ever will, especially intact animals.
for a disease such as AD, given its complexity ]. It could be argued, therefore, that such a model should not even be an Interventional models Introduction of pharmacological or chemical substances into Even predictive validity, although very useful, is not trivial the brain or the induction of lesions in specific brain regions to establish. It requires effective drugs in the clinic, ideally may replicate some of the characteristic features of AD. Many with different mechanisms, and a good understanding of the models involve the introduction of Ab peptide into the brain false positives and false negatives. The latter is particularly of, for example, the rat [] or rhesus monkey []. Although hard to evaluate because compounds that do not work in the these models induce some of the clinical signs, they do not animal model are often not tested in the clinic (or reported).
directly resemble AD pathology Other chemical inter- In a way, one could say that the availability of an animal ventional models include scopolamine-induced amnesia, model with established good predictive validity is unfortu- introduction of inflammation with endotoxins or interfer- nately inversely correlated with the unmet medical need.
ence with brain metabolism [].
All the disease models for AD listed have at best face The lesion models involve the chemical or physical validity, where the animal model merely shares phenomen- destruction of specific brain areas, which are generally either ological similarities with the disorder under study. Hence the cholinergic (i.e. the nucleus basalis magnocellularis in problem of ‘translation' and recent interest in ‘translational rodents, e.g. ]) or involved in cognition (i.e. hippocampus, striatal and cortical brain regions). Major disadvantages of the The schematic in illustrates the complexity and the lesion models include non-specificity of the lesion, and their many possible pitfalls in trying to create an animal analogue failure to capture the disease progression and the more global for human disease. A single relevant difference, whether aspects of the disease (too specific) structural or quantitative, can ‘invalidate' the model. This As a model of disease, interventional models would gen- is, when it is not understood and considered in the transla- erally be better at identifying symptomatic or corrective tion to the human condition.
treatments, rather than disease modifying therapies that halt By contrast, many of the single components of the sche- or slow down progression, unless the ‘intervention' repre- matic in will be preserved across species (normal sents a damage early in the disease progression.
physiology). Thus, whereas it is probably impossible to repli- Furthermore, a wide range of variability including species, cate the full sequence from cause to symptoms in animals, the animal husbandry, site of injection or induction, model study of components of the hypothesized sequence could protocol and concentration and volume of substances may influence the experimental outcome. However, these models Box 1. Validation criteria for animal models ] can provide important insights such as scopolamine-induced amnesia model contributed to the role of cholinergic system Face validity: The animal model resembles the human disease condition on a superficial level, for example, biochemistry or symptomatology.
in cognition [], specific brain lesions induced memory Predictive validity: The animal model can successfully discriminate deficit models and the neuronal mechanism underlying between successful and unsuccessful treatments for the human disease memory dysfunction and the Ab/pharmaco-chemical Construct validity: The animal model is based on a sound theoretical substance induced model and the understanding of inflam- rationale, requiring good understanding of the human disease condition.
mation, neurotoxicity, neurodegeneration and synaptic Notably, construct validity does not require superficial similarity.
Vol. 10, No. 3 2013 Drug Discovery Today: Technologies Translational pharmacology Drug Discovery Today: Technologies Figure 1. A theoretical framework for a disease model. The disease cause, which can be an event or a continuous process, would result in some primary damage, followed possibly by a downstream damage cascade, the result of which may be the disruption of function, observed as a symptom. For AD this could be a genetic defect generating an excess of Ab, downstream resulting in death of cholinergic neurons causing memory impairment, observed as a poor ADAScog score. The diagnosis of the disease may involve some or all of these components.There may be parallel paths, interacting at different stages. A different cause may either generate the same secondary damage, or independently show the same symptoms. Depending on how the disease is defined this may represent heterogeneity. Also, the same cause and primary damage can have multiple secondary damages (in multiple anatomical locations) and subsequently, multiple dysfunctions and symptoms. Importantly, these cascades are part of a system, which defines such elements as the turnover rates of components, repair (i.e.
reversibility) and response (i.e. immune system), and robustness (i.e. feedback loops, collateral pathways, compensatory mechanisms). All the elements of such a model have anatomical locations, and can be described quantitatively in terms of capacity, sensitivity and rate constants. In addition, the cause has an intensity and time course associated with it. Furthermore, most if not all relationships between elements are non-linear.
identify potential rate-limiting steps, suggest possible inter- are an unavoidable part of drug development. A properly ventions and provide a rationale for the translation to conducted negative clinical trial would suggest that a revision of the hypothesis is necessary. Appropriate bio- markers can help to understand the nature of the required Reasons animal models of disease can fail revision. Some possible reasons are: So far the animal models for AD have not been very successful a. Incorrect basic structure of the model. Key elements of in identifying effective treatments (see and []).
the animal model are not major players on the causal Why? It is possible to highlight specific weaknesses (and path in humans. There is some similarity with reasons strengths) for each animal model (e.g. ]). Some of the why biomarkers fail In addition, a key element issues have been raised in the previous sections. More gen- could be regulated differently in the model species.
erally, one can identify three categories of reasons why b. The model is quantitatively wrong. Critical species clinical trial results do not match the animal model results, differences in the capacity, sensitivity or rate constants and not only because of the animal model per se. In our governing the relationship between elements are not discussion it is assumed that the animal experimentation is technically sound, potential issues with that being discussed 2. Incorrect (quantitative) translation of the intervention to the patient. Thus the target is relevant but not appropri- ately engaged. Without the appropriate pharmacological 1. The animal disease model does not truly reflect the disease biomarkers this is difficult to identify, and very little useful or represents an incorrect hypothesis. Hence the drug inferences can be made after a negative clinical trial. One target is not relevant to human disease. This can ulti- cannot distinguish between wrong target and wrong dose.
mately only be addressed in patients, and failures like this Possible explanations are: Drug Discovery Today: Technologies Translational pharmacology Vol. 10, No. 3 2013 a. Incorrect dosing regimen because of species differences engage multiple targets to treat complex disorders ]. Multi- in pharmacokinetics, distribution to target site and/or ple targets may be required to significantly affect one aspect target affinity, duration of treatment required.
of the disease, or to address the multidimensionality of the b. Physiology of the primary target is different between disorder. In addition, animal models can be used to screen species, or altered in human disease (not captured in series of compounds, and identify potential targets by their the animal model), for example, interventions that ability to modulate an aspect of the hypothesized disease significantly lower brain Ab in mice can fail to do so in AD patients because of species differences in Ab The main objective for the pharmacological intervention is to ensure that the dose regimen that will be tested in the 3. Clinical trial design not appropriate to test the interven- human patient does what it pharmacologically is meant to tion under investigation. A negative clinical trial risks do: sufficiently, long enough and at the right place. Animal abandoning a correct hypothesis and a beneficial inter- experimentation can contribute to our understanding of the vention. Although not the objective of this review, some pharmacokinetics of the compound and the target pharma- cology, as long as properly scaled between species, and pro- a. Wrong dosing regimen. For example, the translation vide guidance on what may be the required target may be adequate, but target engagement is limited by engagement. One objective is to verify this in humans early safety and tolerability.
on in clinical development, as is done for pharmacokinetics b. Study too short to be able to notice effects and, if possible, with biomarkers of pharmacology.
c. Wrong patient population. The patients are too Animal experimentation does not seem to contribute advanced to allow significant benefit of the interven- directly to the clinical trial design. However, as mentioned tion or only a subpopulation would benefit.
above, the right experiments can provide guidance regarding d. Insensitive or not relevant endpoints. The modality the timing of treatment and regarding which modalities affected by the treatment may not, or poorly so, be would be consistent with the target pharmacology.
captured by the study endpoints.
Thus, animal experimentation remains part of the scien- e. Other factors such as placebo response and data ana- tific method that iterates between hypothesis generation, predictions and experimentation. The hypothesis is in fact conceptual and should be a human disease model, attempt- The first two categories are failures to translate animal ing to address the components mentioned in Given the results to human patients. One cannot translate if the crucial complexity, describing and connecting parts of the model components leading up to a particular endpoint are not quantitatively is an important part of this process identified. This is where a conceptual framework such as Among other benefits, modelling and simulation provide a shown in would help by capturing the current state way of integrating currently available facts and beliefs, iden- of knowledge and assumptions explicitly. One could argue tifying gaps and raising questions [] and, for drug devel- that this understanding could also contribute to avoid some opment, ultimately some sense of the uncertainties of a failures of the third category, for example, timing of the particular development program.
intervention, identification of surrogate or biomarkers or For drug development some specific quantitative aspects of the modality monitored for treatment effect.
a theoretical human model to consider are suggested below, all of which are a potential source of critical species differ- Animal experimentation and Alzheimer's disease How can animal experimentation contribute to some of the challenges facing drug development in general and our Target selection: search for a cure for AD in particular? Three major challenges mirror the reasons for failure listed above and they are: first,  Relationship between target modulation and subsequent target selection, second, the (pharmacological) intervention events. How much ‘change' of the target is required to and third, the clinical patient trials.
cause changes downstream (sensitivity) and when can Target selection is probably the most difficult challenge in these changes be observed? What is the relative contribu- AD and is directly related to our, limited, understanding of tion of the target pathway to the endpoint of interest? Is the disease. An obvious role of animal experiments is to there a rationale for multiple targets? increase this understanding, as they already greatly have,  The homeostasis of the target and/or critical downstream and to generate hypotheses. However, the correctness of elements, for example, Ab, or memory function. How is it the hypothesis can only be addressed in human patients with maintained and can it be modulated? the right tool. Some have suggested to focus more on research  Related to the above, robustness and redundancies in the in human patients [and the need to simultaneously Vol. 10, No. 3 2013 Drug Discovery Today: Technologies Translational pharmacology Evaluation of the intervention: example is some modelling and simulation of brain atro- phy as biomarker []. Multiple competing factors contri-  Pharmacokinetics buting to brain volume loss measurements were identified  Can the intervention modulate the component(s) of inter- and incorporated in a model. These included, for example, est (target engagement), and if so, what is the required normal aging, disease, volume of amyloid deposits and timing of the intervention? And, again, how long does it inflammation. The model could explain why treatment take before these changes can be observed, which also with an anti-amyloid vaccine would result in a temporary depends on the endpoint? In other words, pharmacoki- acceleration of brain volume loss, as was observed in clin- netic–pharmacodynamic relationships for the target phar- ical trials.
 Are there bio- or surrogate markers that can inform on the Another example is the complexity underlying interven- target pharmacology, and, if so, what is the relationship tions attempting to lower soluble Ab in the CNS of patients, between target function and biomarker level? Again, con- suggested as potentially beneficial by the Ab hypothesis (e.g.
sidering relative contributions and time courses. A recent ]). The relationships between the different forms of Ab Effect of Q PB Variability Fraction of Plasma or Brain Basal A Blood Brain Barrier
Koff · Cx
Koff · Cx + UP0
K01 + Kon · P
KP0 + Kon · D
Total Plasma Aβ = P + Cx
Kon· D · P
Drug Discovery Today: Technologies Figure 2. A model of the Ab system in the presence of an antibody restricted to the peripheral circulation. The model describes the relationships between free soluble Ab in the brain (BS), in the CSF (CSF) and in the plasma (P), separated by the blood brain and blood CSF barrier. An anti-Ab antibody (D) can form a complex with the plasma Ab (Cx). The relationships were parameterised by rate constants (K and U) and between compartmental clearances (Q). The Kon and Koff represent the association and dissociation rate constants for the antibody with Ab.The inset shows the simulation results for a normal rat after once weekly dosing with the antibody with the fraction of basal Ab levels in the brain (dashed) and plasma (solid). The distribution of lines represents the impact of varying the value of the (unknown) parameter Q which is the clearance from plasma to brain. Although the impact of this parameter value is significant (between 10% and 40% reduction of brain Ab), it suggests that nearly 90% reduction in plasma is matched by only about 20% reduction in the brain.
Reproduced with permission from ].
Drug Discovery Today: Technologies Translational pharmacology Vol. 10, No. 3 2013 across the different compartments of interest, for example, hypotheses, it is crucial to identify a pharmacologically active brain, CSF and plasma, are not immediately obvious. This dosing regimen to test in patients. Then a negative clinical issue was highlighted for the rational selection of biomarkers patient study, although disappointing, will at least increase by Thompson and Lockhart for example Ab concentra- our knowledge regarding the disease.
tions in the CSF as marker for CNS concentrations, but is equally important for the rationalization of interventions.
Preliminary work attempting to address this was presented by Simeoni et al. and an example of the model, with an FA and AR are funded as part of the Pharma-Cog consortium intervention included, is shown in Interestingly, the by the European Community's Seventh Framework Pro- preliminary simulations with the model suggest that an anti- gramme for the Innovative Medicine Initiative under Grant body against Ab, restricted to the systemic circulation, would Agreement no. 115009. For further information please refer only modestly lower CNS levels of Ab, even if plasma levels were greatly reduced ). This is clearly a red flag when in the process of developing such an antibody, and merits further investigation of the system and assumptions. The Mattson, M.P. (2004) Pathways towards and away from Alzheimer's model made specific assumptions about the structure of disease. Nature 430, 631–639 the model and some of the parameters and ignored some Hardy, J.A. and Higgins, G.A. (1992) Alzheimer's disease: the amyloid cascade hypothesis. Science 256, 184–185 non-linearities that must be present in the system. Never- Schnabel, J. (2011) Little proteins, big clues. Nature 475, S12–S14 theless, this work could guide specific future experiments and Van Dam, D. and De Deyn, P.P. (2011) Animal models in the drug the impact of parameter assumptions can be explored by discovery pipeline for Alzheimer's disease. Br. J. Pharmacol. 164, sensitivity analysis (as was done for this presentation).
Johnstone, E. et al. (1991) Conservation of the sequence of the Alzheimer's Thus, even if animal models of AD cannot be considered disease amyloid peptide in dog, polar bear and five other mammals by models of disease with construct (or even predictive) validity, cross-species polymerase chain reaction analysis. Mol. Brain Res. 10, it is not to so that animal experiments have not been or will Languille, S. et al. (2012) The grey mouse lemur: a non human primate not be useful in understanding of the disease or that the model for ageing studies. Ageing Res. Rev. 11, 150–162 hypotheses they aim to reflect are incorrect. The challenge is Dhenain, M. et al. (2003) Regional atrophy in the brain of lissencephalic to identify which part of the overall human disease model a mouse lemur primates: measurement by automatic histogram-based segmentation of MR images. Magn. Reson. Med. 50, 984–992 particular animal experiment addresses.
Bons, N. et al. (2006) Microcebus murinus: a useful primate model for Lastly, although not specifically addressed here, anticipat- human cerebral aging and Alzheimer's disease? Genes Brain Behav. 5, ing safety issues of the intervention under study is obviously important. In particular, adverse pharmacology associated Mestre, N. and Bons, N. (1993) Age-related cytological changes and neuronal loss in basal forebrain cholinergic neurons in Microcebus murinus with the primary and/or secondary targets of the compound (Lemurian, Primate). Neurodegeneration 2, 25–32 can, should, be addressed applying similar principles as for Kraska, A. et al. (2011) Age-associated cerebral atrophy in mouse lemur primates. Neurobiol. Aging 32, 894–906 Picq, J.L. (2007) Aging affects executive functions and memory in mouse lemur primates. Exp. Gerontol. 42, 223–232 Picq, J-L. et al. (2012) Age-related cerebral atrophy in nonhuman primates The value of many of the so-called animal models for AD lies predicts cognitive impairments. Neurobiol. Aging 33, 1096–1109 Giannakopoulos, P. et al. (1997) Quantitative analysis of tau protein- not in the fact that they attempt to replicate the full spectrum immunoreactive accumulations and b amyloid protein deposits in the of the human disease, but in that they could help us, together cerebral cortex of the mouse lemur, Microcebus murinus. Acta Neuropathol.
with clinical data, to generate hypotheses and understand the complex physiology relevant to the human disease. In that Buccafusco, J.J. (2008) Estimation of working memory in macaques for studying drugs for the treatment of cognitive disorders. J. Alzheimer's Dis.
sense, they are not really animal models of disease but ‘just' animal experiments. These experiments, then, need to De Magalhaes, J. and Costa, J. (2009) A database of vertebrate longevity address specific questions that allow the construction of records and their relation to other life-history traits. J. Evol. Biol. 22, the hypotheses for the whole or parts of a quantitative human Braidy, N. et al. (2012) Recent rodent models for Alzheimer's disease: disease model. The conceptual human model is where inter- clinical implications and basic research. J. Neural Transm. 119, 173–195 ventions should be evaluated: for example, target selection, Inestrosa, N.C. et al. (2005) Human-like rodent amyloid-b-peptide determines Alzheimer pathology in aged wild-type Octodon degu.
dosing regimen and timing, biomarker and clinical endpoint Neurobiol. Aging 26, 1023–1028 selection and safety. Integration of information and ideas Games, D. et al. (1995) Alzheimer-type neuropathology in transgenic mice into a quantitative conceptual (sub) model of human disease overexpressing V717F b-amyloid precursor protein. Nature 373, 523–527 and pharmacology is therefore considered essential.
Kalback, W. et al. (2002) APP transgenic mice Tg2576 accumulate Ab peptides that are distinct from the chemically modified and insoluble For drug development, whereas there is likely to remain peptides deposited in Alzheimer's disease senile plaques. Biochemistry 41, great uncertainty around target selection and the associated Vol. 10, No. 3 2013 Drug Discovery Today: Technologies Translational pharmacology Goedert, M. et al. (2006) Tau protein, the paired helical filament and Cedarbaum, J.M. and Wagg, J.K. (2011) Predicting initial changes in brain Alzheimer's disease. J. Alzheimer's Dis. 9, 195–207 volume with effective anti-amyloid agents: just ask alice. Alzheimer's Oddo, S. et al. (2003) Triple-transgenic model of Alzheimer's disease with plaques and tangles: intracellular A [beta] and synaptic dysfunction.
Thompson, P.W. and Lockhart, A. (2009) Monitoring the amyloid beta- Neuron 39, 409–421 peptide in vivo-caveat emptor. Drug Discov. Today 14, 241–251 Du, H. et al. (2008) Cyclophilin D deficiency attenuates mitochondrial and Simeoni M, et al. (2008) Modelling beta amyloid system: sensitivity neuronal perturbation and ameliorates learning and memory in analysis at steady state and in dynamic conditions. Population Analysis Alzheimer's disease. Nat. Med. 14, 1097–1105 Group Europe (PAGE) Meeting, Marseille, France (Abstract 1396) Lee, Y.J. et al. (2010) Inflammation and Alzheimers disease. Arch. Pharm.
Cotman, C.W. and Head, E. (2008) The canine (dog) model of human Res. 33, 1539–1556 aging and disease: dietary, environmental and immunotherapy Gomez-Isla, T. et al. (1996) Profound loss of layer II entorhinal cortex approaches. J. Alzheimers Dis. 15, 685–707 neurons distinguishes very mild Alzheimers disease from nondemented Check, E. (2002) Nerve inflammation halts trial for Alzheimer's drug.
aging. J. Neurosci. 16, 4450–4491 Driscoll, M. and Gerstbrein, B. (2003) Dying for a cause: invertebrate Fillenbaum, G.G. et al. (2005) Dementia and Alzheimer's disease in genetics takes on human neurodegeneration. Nat. Rev. Genet. 4, community-dwelling elders taking vitamin C and/or vitamin E. Ann.
Pharmacother. 39, 2009–2014 Iijima, K. and Iijima-Ando, K. (2008) Drosophila models of Alzheimer's Schenk, D. (2002) Amyloid-b immunotherapy for Alzheimer's disease: the amyloidosis: the challenge of dissecting the complex mechanisms of end of the beginning. Nat. Rev. Neurosci. 3, 824–828 toxicity of amyloid-b 42. J. Alzheimers Dis. 15, 523–540 Morgan, D. (2011) Immunotherapy for Alzheimers disease. J. Intern. Med.
Sang, T.K. and Jackson, G.R. (2005) Drosophila models of neurodegenerative disease. NeuroRx 2, 438–446 Hussain, I. et al. (2007) Oral administration of a potent and selective non- Sinadinos, C. et al. (2009) Live axonal transport disruption by mutant peptidic BACE-1 inhibitor decreases b-cleavage of amyloid precursor huntingtin fragments in Drosophila motor neuron axons. Neurobiol. Dis.
protein and amyloid-b production in vivo. J. Neurochem. 100, 802–809 Townsend, M. et al. (2010) Oral treatment with a g-secretase inhibitor Nakamura, S. et al. (2001) Progressive brain dysfunction following improves long-term potentiation in a mouse model of Alzheimer's disease.
intracerebroventricular infusion of beta1-42-amyloid peptide. Brain Res.
J. Pharmacol. Exp. Ther. 333, 110–119 Kounnas, M.Z. et al. (2010) Modulation of [gamma]-secretase reduces Li, W. et al. (2010) A nonhuman primate model of Alzheimer's disease [beta]-amyloid deposition in a transgenic mouse model of Alzheimer's generated by intracranial injection of amyloid-beta42 and thiorphan.
disease. Neuron 67, 769–780 Metab. Brain Dis. 25, 277–284 Dong, H. et al. (2009) Effects of donepezil on amyloid-[beta] and synapse Duyckaerts, C. et al. (2008) Alzheimer disease models and human density in the Tg2576 mouse model of Alzheimer's disease. Brain Res.
neuropathology: similarities and differences. Acta Neuropathol. (Berl.) 115, Choi, J.K. et al. (2010) Anti-inflammatory treatment in AD mice protects Fine, A. et al. (1985) Cholinergic ventral forebrain grafts into the neocortex against neuronal pathology. Exp. Neurol. 223, 377–384 improve passive avoidance memory in a rat model of Alzheimer disease.
Boimel, M. et al. (2009) Statins reduce the neurofibrillary tangle burden in Proc. Natl. Acad. Sci. U.S.A. 82, 5227–5230 a mouse model of tauopathy. J. Neuropathol. Exp. Neurol. 68, 314–325 Ebert, U. and Kirch, W. (1998) Scopolamine model of dementia: Nishida, Y. et al. (2009) Depletion of vitamin E increases Amyloid b electroencephalogram findings and cognitive performance. Eur. J. Clin.
accumulation by decreasing its clearances from brain and blood in a Invest. 28, 944–949 mouse model of Alzheimer disease. J. Biol. Chem. 284, 33400–33408 Castane, A. et al. (2010) Selective lesions of the dorsomedial striatum Costa, R. et al. (2011) Testing the therapeutic potential of doxycycline in a impair serial spatial reversal learning in rats. Behav. Brain Res. 210, Drosophila melanogaster model of Alzheimer's disease. J. Biol. Chem. 286, Frautschy, S. et al. (2001) Phenolic anti-inflammatory antioxidant reversal Golde, T.E. et al. (2010) Targeting A [beta] and tau in Alzheimer's disease, of A [beta]-induced cognitive deficits and neuropathology. Neurobiol.
an early interim report. Exp. Neurol. 223, 252–266 Aging 22, 993–1005 Extance, A. (2010) Alzheimer's failure raises questions about disease- Willner, P. (1986) Validation criteria for animal models of human mental modifying strategies. Nat. Rev. Drug Discov. 9, 749–751 disorders: learned helplessness as a paradigm case. Prog.
Sabbagh, M. and Cummings, J. (2011) Progressive cholinergic decline in Neuropsychopharmacol. Biol. Psychiatry 10, 677–690 Alzheimer's disease: consideration for treatment with donepezil 23 mg in Horrobin, D.F. (2003) Modern biomedical research: an internally self- patients with moderate to severe symptomatology. BMC Neurol. 11, 21 consistent universe with little contact with medical reality? Nat. Rev. Drug Imbimbo, B.P. (2009) An update on the efficacy of non-steroidal anti- Discov. 2, 151–154 inflammatory drugs in Alzheimer's disease. Expert Opin. Invest. Drugs 18, Geerts, H. (2009) Of mice and men: bridging the translational disconnect in CNS drug discovery. CNS Drugs 23, 915–926 Feldman, H. et al. (2010) Randomized controlled trial of atorvastatin in van der Worp, H.B. et al. (2010) Can animal models of disease reliably mild to moderate Alzheimer disease. Neurology 74, 956–964 inform human studies? PLoS Med. 7, e1000245 Isaac, M. et al. (2008) Vitamin E for Alzheimer's disease and mild cognitive Frank, R. and Hargreaves, R. (2003) Clinical biomarkers in drug discovery impairment. Cochrane Database Syst Rev (Online) CD002854 and development. Nat. Rev. Drug Discov. 2, 566–580 Loeb, M.B. et al. (2004) A randomized, controlled trial of doxycycline and Dollery, C.T. (2010) The challenge of complexity. Clin. Pharmacol. Ther.
rifampin for patients with Alzheimer's disease. J. Am. Geriatr. Soc. 52, Aderem, A. (2005) Systems biology: its practice and challenges. Cell 121, Ahmed, T. et al. (2010) Curcuminoids enhance memory in an amyloid- infused rat model of Alzheimer's disease. Neuroscience 169, 1296–1306 Butcher, E.C. et al. (2004) Systems biology in drug discovery. Nat.
Hu, S. et al. (2009) GSK3 inhibitors show benefits in an Alzheimer's disease Biotechnol. 22, 1253–1259 (AD) model of neurodegeneration but adverse effects in control animals.
Boxenbaum, H. (1992) Pharmacokinetics: philosophy of modeling. Drug Neurobiol. Dis. 33, 193–206 Metab. Rev. 24, 89–120 Martinez, A. et al. (2011) Glycogen synthase kinase 3 inhibitors in the next Massoud, T.F. et al. (1998) Principles and philosophy of modeling in horizon for Alzheimer's disease treatment. Int. J. Alzheimers Dis. 2011, biomedical research. FASEB J. 12, 275–285

Source: http://www.oltrelasperimentazioneanimale.eu/archivio_pubmed/critica_scientifica/Laurijssens,%20Aujard,%20Rahman%20-%202013%20-%20Animal%20models%20of%20Alzheimer's%20disease%20and%20drug%20development.pdf