Spontaneous and Induced Nontransgenic Animal Models of AD

Abstract

Alzheimer’s disease (AD), the most common neurodegenerative and dementing disorder, is characterized by extracellular amyloid deposition, intracellular neurofibrillary tangle formation, and neuronal loss. We are still behind in AD research in terms of knowledge regarding understanding its pathophysiology and designing therapeutics because of the lack of an accurate animal model for AD. A complete animal model of AD should imitate all the cognitive, behavioral, and neuropathological features of the disease. Partial models are currently in use, which only mimic specific and not all of the components of AD pathology. Currently the transgenic animals are the popular models for AD research, but different genetic backgrounds of these transgenic animals remain a major confounding factor. This review attempts to summarize the current literature on nontransgenic animal models of AD and to highlight the potential of exploiting spontaneous and induced animal models for neuropathological, neurochemical, neurobehavioral, and neuroprotective studies of AD.

Introduction

Around the world, about 24 million people are having dementia. ¹ Alzheimer’s disease (AD), the most common form of dementia, is an age-related progressive neurodegenerative disorder associated with loss of cognition and memory. Pathophysiologically, AD is characterized by degeneration of neurons and synapses, deposition of amyloid beta (Aβ) plaques, and accumulation of neurofibrillary tangles (NFTs) in brain. ² Amyloid plaques consist of accumulated or aggregated Aβ peptides that are excised from the transmembrane protein, amyloid precursor protein (APP). These plaques thus formed are known to have toxic properties causing synaptic dysfunction, mitochondrial damage, oxidative stress (OXS), excitotoxicity, and microglial activation leading to neurodegeneration. The NFTs are the result of hyperphosphorylation of tau protein, which lead to microtubule instability, consequently affecting the axonal morphology. ³

Etiology of AD is multifactorial, with both genetic and environmental factors implicated in its pathogenesis. ⁴ The disease is basically classified into 2 types, sporadic AD (SAD), the common form accounting for 90% to 95% of the cases for which no defined cause is known, and familial AD (FAD) that shows autosomal dominant inheritance, and the genes involved are mutations in APP, presenilin 1 (PS1), and PS2. ⁵ Advancing age appears to be the greatest risk factor for AD. Its incidence increases in people of 60 years and older. Other risk factors may include positive family history of dementia, female gender, and head trauma. ⁶

The pathophysiology of AD is complex involving many abnormal mechanisms that ultimately lead to the disease process. Our understanding of the pathomechanism of AD is constantly changing. The accumulation of stable amyloid plaques is earlier thought to be and widely accepted as a central pathogenetic event of the AD. All mutations known to cause AD increase the production of Aβ peptide. Association of neuronal loss and deficiency of neurotransmitter system leads to the appearance of this dementia syndrome. Besides this, the neuronal loss is associated with neurochemical deficits that disrupt cell-to-cell communications, abnormal synthesis, and accumulation of cytoskeletal proteins (eg, tau that leads to the formation of NFT), loss of synapses, damage through oxidative metabolism, and cell death. ⁷ Recently, this classical amyloid cascade hypothesis (ACH) has been challenged; and Aβ plaques, leading to the formation of NFTs the well-known cardinal features of AD, are now thought to be an adaptive response instead of causative agents of AD. ⁸ So the notion that Aβ and phosphorylated tau (P-tau) are pathologic molecules is slowly changing, and their physiological roles that they represent a cellular adaptive strategy to OXS are just being realized. Plaques and tangles act as compensatory response to OXS and mitochondrial dysfunction, followed by involvement of other disturbed mechanisms such as astrocytic and microglial activation, inflammation, cholesterol imbalance, excitotoxicity, Ca²⁺ homeostasis, autophagy, and mitotic dysfunction. ^{9 –11} Although the exact pathophysiological cascade remains to be understood, these diverse mechanisms ultimately result in neuronal death. The role of these individual processes in AD has already been comprehensively discussed. ^{12 –15}

The poor prognosis and hence delayed treatment of AD is due to our inadequate understanding about the disease with respect to its origin and pathogenesis. The diversity of mechanisms also remains an obstacle in creating a model that depicts the natural history of AD in humans. Most commonly used models in neuroscience are partial models that focus only on a particular aspect of disease, instead of modeling full-blown pathology. A valid model should have similar features of symptomatology, etiology, background, and treatment basis as seen in the target species, the species to be modeled. ¹⁶

Animal Models and AD

Although studies on patients with AD would be the best possible experimental approach, we are put to face a lot of setbacks because of the puzzling nature of the disease, such as, the long disease process in humans, variable asymptomatic period, interindividual variation, and lack of a noninvasive confirmatory diagnostic technique. Animal models possess some advantageous features that make them essential for research like shorter life spans, shorter gestation period, and high progeny.

Animal models of AD can be classified into spontaneous and induced models. As the name suggests, spontaneous models are supposed to develop AD condition without artificial manipulation, for example, aging animals. Induced models display the condition as a result of artificial manipulation such as surgical procedures, administration of some drug, or application of some environmental stimuli and genetic manipulations (ie, transgenic animals). On the basis of the manipulation done, induced models can be further classified into transgenic and nontransgenic models. Both the groups have their own merits and demerits, which need careful analysis.

Transgenic mice are the most commonly used animal models for studying AD and are an essential tool in studying in vivo pathophysiology. With the identification of genetic factors involved in AD, the development of several transgenic mouse models have taken place, constructed with either bearing an exogenous gene or knockout of a particular gene. The common rationale employed in creating transgenic model of AD is the overexpression of the transgene carrying FAD mutations under different promoters. Genes, namely, mutant APP, PS-1 and PS-2, apolipoprotein E (ApoE), and tau are used to construct transgenic mouse models of AD. ^17,18

Although transgenic models have served the purpose of providing relevant information on the individual role and particular genes in the pathophysiology of AD, there are certain lacunae in using transgenic animals, which makes the use of nontransgenic models more preferable. The altered genetic component of these animals cannot be considered as general for all. Above all, transgenic animal models only account for early-onset FAD, a form representing less than 5% of the patients with AD. But to study SAD, we require a model that would represent the natural history of AD in humans. Various nontransgenic models are being adopted, which show one or more hallmarks of AD-like senile plaques, NFTs, OXS, and cognitive impairment. However, no single animal model has been produced, which replicates all aspects analogous to AD in humans. The transgenic models have already been discussed elsewhere. ^{19 –22} In this article, we overview the various nontransgenic animal models and their utility in understanding pathophysiology and designing therapeutics against AD.

Spontaneous Models

To date, a variety of animals ranging from worms to polar bears have been tried as AD models. Few species such as dogs, cats, sheep, and nonhuman primates spontaneously develop plaque or tau pathology along with neurobehavioral impairment.

Dogs have a long history of research in aging, and among its various breeds, beagles are the most commonly studied. Beagles above 10 years of age are considered equivalent to 66 to 96 years of human. ²³ Aging in beagle represents features of normal aging, mild cognitive impairment as well as early or mild AD. ^24,25 Behavioral studies exhibited learning and memory deficits in aged dogs, ²⁶ with learning assessed by different learning tasks such as size concept, size discrimination and spatial learning, ^27,28 and memory as tested by object recognition, spatial memory, and visuospatial working memory tasks. ^29,30 In dogs, this cognitive impairment began around 6 to 7 years of age. ³¹ Also, the animals exhibited diffuse amyloid deposits in the prefrontal, temporal, and occipital cortices, reduced levels of antioxidant enzymes, superoxide dismutase (SOD), and glutamine synthetase, increased OXS, and DNA and RNA damage. Dogs share an identical amino acid sequence of Aβ 42 peptide as that of human, which may explain the presence of accumulated Aβ in their brains. However, these animals lacked tau accumulation that probably may be attributed to the interspecies differences in tau protein sequences. ³² Thus, aging dogs are probably a reliable model depicting early AD, showing cognitive decline along with pathological features as Aβ accumulation, oxidative damage, and neuronal loss.

The other example of natural animal model of AD would be aged Octodon degu, a rodent that has been shown to develop both plaque formation and tangle accumulation, after 3 to 4 years of age. Its Aβ 42 peptide is highly homologous to humans, with a single amino acid substitution, whereas all other murine forms differ in 3 amino acids from humans. With advancing age, it displays both intra- and extracellular deposition of Aβ, intracellular accumulation of tau, and acetyl cholinesterase positive neurons. These neuropathological features were found to correlate with memory impairment in an age-dependent manner. ³³ Also, ubiquitination and astrocytic changes are seen, thus making it the first rodent model mimicking complete pathology of AD. ³⁴

Aging sheep represents an animal model of neurofibrillary degeneration, depicting structures that resemble Alzheimer-like NFTs and neuritic plaques. Immunohistochemistry and staining techniques such as thioflavin S and silver staining have shown the presence of neurofibrillary-like changes naturally in the cerebral cortex of aging sheep, Ovis aries, with features starting as early as 5 years, which increases in number and size with advancing age. Although there is no cognitive decline in these sheep, they can be employed to reveal information regarding AD pathoetiology. ³⁵

The cholesterol-fed aged New Zealand white rabbit, best small animal model for coronary artery disease (CAD), is also considered for studying AD. These rabbits manifest senile plaques that can be explained by the identical sequence of Aβ peptide in rabbits and humans. Although the aged rabbits do not spontaneously develop plaque pathology, Aβ accumulation in them is observed with increased exposure to cholesterol and copper. ³⁶ Intake of copper in these cholesterol-fed rabbits led to increased Aβ accumulation that is due to the ability of Cu to reduce Aβ clearance from rabbit brain. ¹⁷ Other AD features observed in cholesterol-fed aged rabbits were ApoE immunoreactivity, inflammation in terms of microglial activation, presence of OXS markers, that is increased SOD and learning impairment. ³⁷

Chick embryos are also an interesting natural animal model to study APP pathology. Although aged chickens , that is, upto 10 years of age do not develop any of the AD pathology, APP expression in chick embryo parallels the human system. Chick embryos express APP 695 and APP 751, both of which are major APP isoforms in humans, ³⁸ along with the gene-encoding proteins involved in the production of Aβ, including β secretase APP-cleaving enzyme (BACE)-1, BACE-2, PS-1, PS-2, and nicastrin. Also, neuropil thread formation has been observed in chick embryo on mitochondrial inhibition. ³⁹ The cellular and developmental biology of APP thus can be investigated in chick embryo, and it can also serve as an assay system for testing drugs targeting APP and its derivatives.

Primates always have advantage over nonprimates in their use in examining AD pathology owing to their phylogenetic proximity and highly conserved APP sequence compared to humans. ^40,41 Primates display age-dependent AD-like pathologies, predominantly tau abnormalities. ⁴² Tauopathy , that is, the presence of paired helical filaments in neocortex has been seen in aged chimpanzees (Pan troglodytes). ⁴³ Tamarins also represent a model of early AD pathology, as plaque formation occurs in them naturally with advancing age. In these non-human primates, Aβ plaque deposition starts at around 13 years of age in frontal and temporal cortex followed by occipital cortex, in a similar pattern as observed in humans, and deposition of Aβ42 was more than that of Aβ40. Along with amyloid pathology, activated microglia, reactive astrocytes, and ubiquitin-positive neuritis have also been observed in primates. Mouse lemurs with advancing age develop many symptoms of accumulated Aβ plaques, NFTs, and loss of cholinergic neurons, as seen in humans with AD. ⁴⁴

Aged bear and polar bear brains also show the features of diffused plaques as well as NFTs. Aged bears show cytoskeletal abnormalities similar to human AD. ⁴⁵ Also, wolverine with increasing age demonstrated senile plaques and NFTs along with other cerebral lesions as seen in AD. ^46,47 Showing both the cardinal features of AD, wolverine and polar bear imitate AD features more closely than most of the other species.

Cognitive impairment may be present in dolphins, which is usually thought to be shown as coastal beaching behavior of aged dolphins. With advancing age, dolphins develop Aβ deposition in cerebellum and medulla oblongata, as shown by congo red staining. This may be because of homologous amino acid sequence of APP, PS1, PS2, and BACE in humans and dolphin. ⁴⁸

Aging rodent offers knowledge about age-associated cholinergic hypofunction in AD, although they do not develop plaques and NFTs. ⁴⁹ Senescence-accelerated mouse (SAM) is an “accelerated senescence” animal model established through phenotypic selection of AKR/J mice. The SAMP8 substrain shows age-related learning and memory deficits and is therefore a model for dementia. ^50,51

Appropriateness of Aging Animals

Spontaneous models exhibit AD-like characteristics with advancing age without any external induction. However, the use of spontaneous animal models in AD research is limited by availability, economical, and ethical reasons. Their inaccessibility to laboratory research along with long life span and cost restricts their utility in AD research. For studying primates, ethical and economical factors are also a major concern. Thus, aging animals are not usually considered appropriate for studying AD in a laboratory setup.

Induced Animal Models

Nontransgenic models have been created by inducing drugs in the animals to exhibit AD characteristics. Induced animal models for AD can be classified depending on the drug/toxic agent used to induce the lesion, its route of administration, the region of the brain affected, or the extent of morphological, behavioral, and pathological changes observed.

Models Featuring Amyloid Pathology

Aspects of AD can be modeled in rodents by direct intracerebral injection of Aβ. This causes learning and memory deficits in treated animals, with the severity of deficits observed dependent on the species of Aβ infused. Various models have been designed to induce amyloid pathology in animals, keeping in mind the amyloid hypothesis of AD. ⁵² Injecting Aβ peptide in rat brain has been shown to produce amyloid plaque formation and disruption of long-term potentiation and behavior. But different models created with variable Aβ peptides have presented contrary results in terms of biochemical, histological, and behavioral patterns. The Aβ1-40, the soluble form, is more readily cleared from brain by phagocytosis in comparison to longer, fibril-forming peptides Aβ1-42. ^53,54 However, the same Aβ1-40 protein can be made to deposit when injected with growth factors or in a preaggregated confirmation. Similarly, intrahippocampal injection of Aβ1-40 protein in rats displays working memory impairment when assessed through radial arm maze task. The memory deficits produced were acute but transient. ⁵⁵

Variant soluble forms of Aβ protein (Aβ1-28, Aβ12-28, Aβ18-28, Aβ12-20) also produce short-term amnesia in rats as evidenced by footshock active avoidance test. ⁵⁶ The aggregated form of Aβ25-35 injection is more capable of producing impaired cognition than soluble Aβ25-35. ^57,58 The more aggregated form of Aβ (Aβ1-43) and Aβ in combination (Aβ1-40 and Aβ1-43) produced memory impairment, neurodegeneration, and microglial activation after 7 weeks of intrahippocampal injection. Single injection of Aβ 1-42 intracerebroventricularly (icv) induced neuroinflammation and OXS, leading to learning and memory impairment in rats. ⁵⁹ Also, intraparenchymal administration is seen to be more effective in contributing to memory impairment than intracerebroventricular injection. Thus, exogenous administration of Aβ in rat brain brings about changed outcomes according to the type and site of Aβ infused. ^60,61

Attempts have been made to inject an additional compound to enhance the effectiveness of Aβ. The AD model has also been constructed by injecting ibotenic acid in combination with Aβ in rats. ⁶² Coinjection of Aβ and ibotenic acid produced neurodegeneration in distant areas such as CA1, CA4, and dentate gyrus. The animals also demonstrated impaired memory, neurochemical changes such as reduced SOD activity, elevated malondialdehyde (MDA) levels, and neuronal loss. ^63,64 Intracerebroventricular injection of Aβ1-40 with continuous intraperitoneal injection of aluminum chloride (AlCl₃) everyday for 3 weeks resulted in amyloid pathology and impaired cognition in rats as seen in AD. ⁶⁵

Another AD model, the Samaritan’s Alzheimer’s Rat Model, is constructed by slowly infusing Aβ peptide with ferrous sulfate heptahydrate and l-buthionine-(S, R)-sulfoximine in Long Evans rat hippocampus via osmotic pump. This model is advantageous in the aspect that rats developed the characteristics of AD within 4 weeks and displayed deposition of Aβ, increased tau levels in cerebrospinal fluid, enhanced OXS, and impaired memory. ^66,67

Different viral vectors, parvovirus, adeno-associated virus, have also been tried as vehicles for delivering Aβ complementary DNA-encoded gene in adult Wistar rats, the outcome of which was enhanced expression of Aβ in the hippocampus, leading to cognitive deficits. ⁶⁸

Several models have been designed for enhancing endogenous Aβ levels to induce plaque formation. Thiorphan is one such compound in application for increasing Aβ accumulation in brain. It reduces degradation of Aβ in brain by inhibiting peptidase neprilysin used for Aβ degradation. Continuous infusion of thiorphan via miniosmotic pump in the hippocampus of rat brain for 4 weeks resulted in Aβ accumulation and plaque formation with impaired learning and memory. ⁶⁹ Similar results were obtained in rabbits treated with thiorphan, which exhibit increased levels of Aβ in cerebrospinal fluid and cortex after 5 days of administration. ⁷⁰ Injection of Aβ42 and thiorphan in middle-aged (16-17 years) Rhesus monkeys also produced memory impairment, Aβ accumulation as observed in the neurons of the basal ganglia, cortex, and hippocampus, followed by neuronal atrophy and loss along with degeneration of choline acetyltransferase–positive cholinergic neurons. ⁷¹

Models Featuring Tau Pathology

The NFTs are the second pathological hallmark of AD. Tau hyperphosphorylation is an initial requisite for formation of NFTs. Tau function is regulated by a balance between several protein kinases and phosphatases, with hyperphosphorylation preceding formation of NFTs. So an AD animal model can be created by producing a hyperphosphorylation dynamics by either activation of kinases or inhibition of phosphatases.

Okadaic acid (OKA) is a known inhibitor of serine/threonine protein phosphatases, which acts by inhibiting protein phosphatase 2A (PP-2A). This results in hyperphosphorylation of neurofilament—H/M subunits and the disruption of microtubules. Infusion of OKA in rat induces progressive cognitive deficiency, NFTs-like conformational changes in the brain, and oxidative damage in both the cortex and the hippocampus. ⁷² The OKA when injected in cerebral cortex of sheep produces tau alterations along with dystrophic neurites. ⁷³ The OKA is also a strong inducer of oxidative stress in cultured cells. ^74,75 Injection of OKA into the nucleus basalis of Meynert in rats lead to cholinergic deficit by decreasing acetylcholine levels.

Calyculin is another inhibitor of PP-2A and PP-1; and in human neuroblastoma cells treated with calyculin (10 nm), there was increased phosphorylation of tau at tau-1 and PHF-1 leading to neurofilament accumulation, and there was also a dose-dependent decrease in cell viability. ⁷⁶

Wortmannin is another compound that turns down protein kinase B (PKB) suppressing PI3 K and thus activates glycogen synthase kinase 3 (GSK-3) by decreasing the phosphorylated state of GSK-3. Stereotaxic administration of wortmannin in lateral ventricle of rat resulted in tau hyperphosphorylation. The level of tau hyperphosphorylation was high for upto 12 hours of injection after which the level started decreasing. The site of phosphorylation was paired helical filament epitope 1 (Ser396/404) in hippocampus and surrounding areas of injection. ^77,78 Wortmannin also produces oxidative stress and alters cell morphology, but amyloid production is less.

Kinase activation has been tried with isoproterenol (0.02 µmol/L), an activator of protein kinase A; when injected in the rat hippocampus bilaterally, it leads to tau hyperphosphorylation at similar PHF-1 and tau-1 epitopes as seen in AD. Isoproterenol also activates Ca/calmodulin-dependent kinase II and cyclin-dependent kinase 5 and causes inhibition of PP-2A expression. Memory impairment was seen in these rats after 48 hours of isoproterenol administration ⁷⁹ along with elevated oxidative stress. ⁸⁰

Haloperidol, an old psychotic drug, has also been reported to cause neuronal damage and oxidative stress on treatment. It was noted that haloperidol administration is associated with tau hyperphosphorylation and hence microtubule disassembly. In a study done on neuroblastoma N1E-115 cells treated with haloperidol (100 mmol/L), it was seen to induce oxidative stress, tau hyperphosphorylation, and hence cytoskeleton disassembly. ⁸¹

Colchicine can induce tauopathy by directly affecting the microtubule polymerization. Colchicine, a known cytotoxicant, binds irreversibly to tubulin dimers thus preventing their polymerization and leading to neurofibrillary degeneration. Intracerebroventricular administration of colchicine produces cognitive impairment and oxidative stress in rats, leading to neurotoxicity. ⁸² Apart from NFT, colchicine models also exhibit loss of cholinergic neurons with reduction in acetylcholinesterase and choline acetyltransferase activity, elevated lipid peroxidation, and nitrite concentration along with lowered levels of glutathione S-transferase (GST), reduced glutathione (GSH), catalase, and SOD. ^83,84 It also leads to morphological changes and destruction of hippocampal cells, impaired axonal transport followed by neuronal loss. The effect of colchicine appears to be time and dose dependent, producing biochemical, histological, and neurobehavioral changes after 2 to 3 weeks of administration. ⁸⁵

With the finding that SAD is recognized as an insulin-resistant brain state (IRBS), another nontransgenic animal model has been proposed that is developed by intracerebroventricular injection of subdiabetogenic doses of streptozotocin, which serves as an experimental model for early pathophysiological changes of SAD. The AD-like changes seen in IRBS rats were impaired learning and spatial memory, neuroinflammation, altered synaptic proteins and insulin/insulin-like growth factor 1 signaling, and increased hyperphosphorylated tau in the brain. ⁸⁶ The STZ-icv-treated animals also were found to develop progressive cholinergic deficits, glucose hypometabolism, oxidative stress, and neurodegeneration that share many features in common with SAD in humans. ⁸⁷

Various studies have linked homocysteine (hcy) levels to AD. Serum hcy levels were observed to be significantly higher in patients with AD when compared to control individuals. ⁸⁸ The exact mechanism behind is not yet known; however, tau hyperphosphorylation was found to be linking hcy and AD. Chronic administration of hcy to rats resulted in memory impairment. The hcy when injected into lateral ventricle of rat brain produces hyperphosphorylation of tau at PHF-1 (Ser396/404) and tau-1 (Ser198/199/202) epitopes. The levels of PP-2A in these animals were seen to be reduced than normal, suggesting hcy could lead to AD pathology through PP-2A downregulation. ⁸⁹ The hcy also seems to play a role in iron dysregulation or oxidative stress cycle, vascular, and neurotoxic pathophysiologic mechanisms contributing to AD pathogenesis. However, hyperhomocysteinemia may also underlie the secondary cause of deficiencies of vitamin B12, B6, or folate. ^90,91

Models Showing Other Features of AD

Apart from amyloid plaques and NFTs, AD pathophysiology also includes other mechanisms like oxidative stress, neuroinflammation, mitochondrial dysfunction, microglial activation, and cell cycle aberration. These mechanisms have also been targeted to create animal models of AD, thus depicting one or more features.

In accordance with cholinergic hypothesis, studies have shown that scopolamine, a muscarinic cholinergic antagonist, was capable of inducing transient memory impairment in rats and dogs. ⁹² Scopolamine was seen to impair encoding and memory consolidation without affecting memory retention. ⁹³ Scopolamine administration was associated with increased acetylcholinesterase activity and no effect on choline acetyltransferase activity. There was also increased oxidative damage to brain as measured by elevated levels of MDA, catalase, and SOD. ⁹⁴ Although there is cholinergic impairment, weakened learning, and oxidative damage, no Aβ or tau accumulation is observed.

The strong association between aluminum (Al) in water/atmosphere and AD as suggested by various studies has prompted the use of Al compounds to induce AD. Rat model created by administering 1.6 mg Al/kg body weight per day corresponding to the maximum Al intake in humans resulted in impaired cognition in rats along with oxidative stress, inflammation, and tau hyperphosphorylation along with the loss of PP-2A. ⁹⁵ But Aβ accumulation was not prominent, and although Al exposure leads to the formation of neuropil threads, the NFTs were not detected. Other salts of Al like Al sulfate/AlCl₃/Al maltolate have also been tried in various animals. The Al sulfate chronically given to mice in drinking water for 12 months resulted in oxidative damage and Aβ deposition as viewed in congophilic amyloid angiopathy. ⁹⁶ Injection of AlCl₃ in the rabbit brain induced the formation of NFTs in brain after 60 days of administration. Also administration of AlCl₃ at a dose of 50 mg/kg per d for 3 months resulted in neurofibrillary degeneration in hippocampus and cortex of mice brain. Aluminum maltolate when given intracisternaly in aged New Zealand white rabbits produced neurofibrillary tangles, amyloid plaque deposition, oxidative stress, cognitive impairment, neuronal cell loss, and apoptosis. In comparison to other Al salts, Al maltolate provides free aqueous Al at physiological pH and hence is preferred to induce AD pathology. All other Al salts—Al lactate, AlF, and AlSiO₄—failed to show AD neuropathology in aged rabbits. Thus, Al-maltolate-treated aged rabbits demonstrated AD pathology by covering histopathological, neurochemical, and behavioral aspects of AD. ⁹⁷

Among the environmental factors causing SAD, the damaging effects of hypoxia on neurodegeneration of AD have also been considered, because of the fact that cellular hypoxia is believed to be one of the major critical initiating event in AD pathogenesis. Hypoxia has been linked to increased Aβ production through many ways. ⁹⁸ It is reported that hypoxia leads to enhanced expression of APP which is a substrate for Aβ, upregulates cleavage of APP by increasing BACE1 gene expression, reduces Aβ degradation by decreasing production of neprilysin, and also brings Ca²⁺ dysfunction. Hypoxia also disturbs Aβ homeostasis in the central nervous system by opening tight junctions, thus causing dysfunction of blood–brain barrier. Hypoxic injury to neurons has also been demonstrated to cause DNA damage, mitochondrial dysfunction, and activation of cyclin-dependent kinase 5 and GSK-3β which are substrates of tau hyperphosphorylation. These cellular mechanisms via hypoxia contribute to AD pathology. ⁹⁹ Chronic intermittent hypoxia exposure to rats for 3 days resulted in significant increase in the production of Aβ peptides. ¹⁰⁰ Thus, hypoxic exposure to the animal can also be considered to create an induced nontransgenic animal model of AD.

Summary and Conclusion

The best model for AD would be the one that reproduces maximum features as seen in human AD; however, it is not always possible due to various interspecies considerations and differences. Some animal species, like dogs and primates, spontaneously develop AD pathology with plaques or tauopathies. But their use is largely restricted due to economical and practical reasons. Thus, AD research is mainly dependent on induced models, the majority of which in use are the transgenic forms. The altered genetic background of the transgenic mice has a great influence on the behavioral and molecular experiments, so it should be carefully considered. Lesion-induced models that involve administration of amyloid peptide or toxins via different routes of administration would differ in picture, as they produce features depending on the type and compound used, mode, route, duration, and administration of drug. The different mechanisms that are affected creates another obstacle to reproduce AD. An ideal model for AD would be the one that would be created by a mechanism that precedes the actual disease process.

Alternatively, the induction process can be combined to study more than one pathogenic mechanisms of AD simultaneously, for instance a model that would produce manifestations of both arms of disease process, namely, Aβ and tauopathy. After careful consideration, it is suggested that AD should be modeled using a combination model to mimic all the pathological and behavioral features of the disease. A combination model can be created to study different combinations of pathogenic mechanisms, for example, lesion-induced models by double injection of toxins, such as amyloid peptide to increase brain Aβ deposition and another toxin to induce tau pathology. However, such complex systems have rarely been used, because of the greater variability in the results obtained.

In this article, we suggest that the use of a combination model with stereotaxic injection of OKA followed by hypoxic exposure would be near ideal. The OKA mimics and produces tau hyperphosphorylated state by downregulating PP-2A, allowing it to aggregate. Hypoxia causes the production of Aβ by various mechanisms. The effect of OKA induction in rats can be potentiated by exposing the same rats to hypoxia for a brief period.

Among the induced drugs, the OKA in addition to tau pathology and cognitive impairment also produces oxidative stress. ^65,72 The OKA also induces NFT formation, a cardinal feature of AD. The OKA treatment to cultured neurons has been shown to induce accumulation of APP-β-C terminal fragment that might be involved in AD pathogenesis. Exposing OKA-treated rats to hypoxic environment may individually as well as synergistically lead to Aβ production, mitochondrial dysfunction, oxidative stress, and tau hyperphosphorylation. ⁹⁷ By unfurling all the features of AD, this approach may provide better understanding of the biological mechanisms that underlie the symptoms of the disease.

No animal spontaneously develops AD, nor can it be experimentally induced fully; however, there is a huge research need for good models of AD. Induced models of this disease have been, and indeed remain, important in the study of AD pathogenesis and possible therapeutic interventions. To recapitulate the major pathological features of AD, the proposed combination model may be considered. Rodent model based on hypoxia and OKA treatment may mimic the main neuropathological, neurochemical, and neurobehavioral aspects of AD and would also provide opportunities to test various therapeutic agents effectively.