Alzheimer's Disease (AD) is a complex condition attributable to several etiological factors, impacting a multitude of biological pathways. Developing a research model that truly encapsulates this complexity of AD has, thus, been a scientific challenge. Even so, many models have been developed to gain insights into various aspects of AD. Animal models of AD include both transgenic animals as well as natural, non-transgenic models of AD. Together, these animal models can be used to simulate AD pathology to understand disease progression. Several in vitro models have also been developed to understand the more complex tissue-specific pathologies associated with AD. These models, especially those currently utilized, are described in the following sections, and Table 1. The Alzforum website has more extensive listing of research models [1], with 177 models as of August 5, 2019 (increasing from 159 models on September 3, 2018, and 124 models on April 7, 2017). Another important consideration using, especially, rodent models, is the difference in their brain masses: a mouse brain is about 0.4 g while a human brain is about 1300 - 1400 g, a very significant difference. The failure of therapeutic development for Alzheimer's disease for the past four decades has led researchers to question not only the value of the current models, but also on using animal models in the first place at all [2]. One recent example is the failed clinical trial of the anti-inflammatory tetracycline minocycline [3].

Transgenic mice have served as a genetic tool to study the effects of several genes implicated in etiology of AD. The most widely-used and notable mouse models express the human APP. The levels and spatial expression of the hAPP protein are driven by varying the promoters used in mice. Commonly used promoters are platelet-derived growth factor B-chain PDGF-B (e.g., J20), thymocyte differentiation antigen 1 Thy-1 (e.g., 5xFAD), and prion protein (PrP) genes.

Other mice models include Tau transgenic mice that express the human tau protein using various promoters, APP-tau double transgenic mice, triple transgenic mice that express APP, PSEN11, and Tau, and five transgenic mice that express 5 familial AD mutant genes (Table 1). These mice exhibit the expected, hallmark neurological symptoms associated with AD such as cognitive deficits, motor deficits, and memory loss, albeit at varying levels.

Researchers select the models based on their research topics. For example, JH Lee et al utilized Tg2576, 5xFAD, TgCRND8, and PS/APP mice to study autophagic build-up of Aβ in neurons [4]. C Adaikkan et al used Tg(APPSwFlLon, PSEN1∗M146L∗L286V), Tg(Prnp-MAPT∗P301S)PS19, and 5XFAD mice to evaluated the effect of gamma entrainment on amyloid plaques and mouse cognitive functions [5]. Faraco G et al studied the effect of high-salt diets on tau hyperphosphorylation and dementia with Tau knockout and APPSwedish mice [6]. Ciccone R et al investigated neuronal hyperactivity in Tg2576 mice, which carry the human APP Swedish 670/671 mutation (K670N e M671L) [7]. Zott B et al studied the effects of beta-amyloid peptide dimers on the suppression of glutamate reuptake and neuronal hyperactivation with APP23xPS45 mice [8]. Nortley R et al evaluated the capillary diameters in the brain cortices of wild-type and App^NL-G-F knock-in AD mice [9]. APP_swePSEN1_dE9 transgenic mice express a chimeric mouse/human amyloid precursor protein (Mo/HuAPP695swe) and a mutant human presenilin 1 (PS1-dE9) in CNS neurons and is a commonly used model. Both mutations lead to early-onset Alzheimer's disease. Animal models may display gender-based phenotype, for example, 5XFAD males develop the amyloid phenotype later than females.

APP/PS1 mice were also used to investigate the mechanisms of the pathological changes in the blood–brain barrier in AD, which are still not completely understood [10]. The study reported suppressed Wnt/β-catenin signalling induced by an increase of GSK3β activation in the cerebral microvessels of the experimental mice. Furthermore, targeting LRP6 activation of the Wnt/β-catenin pathway by the upstream regulator LRP6 in the brain endothelium restored blood–brain barrier malfunction [10].

As APP/PS1 mouse model develops symptoms resembling the clinical features of slow plaque development, it was chosen to develop male APP/PS1mTmG;Flt3-Cre mice with Flt3-Cre–driven expression of a dual-fluorescence mTmG reporter [11]. The study has revealed that circulating monocytes contributed to a large population of macrophages in the choroid plexus, meninges and perivascular spaces. Overall, the obtained data demonstrated that peripheral monocytes infiltrated the brain parenchyma to decrease plaque load [11].

Transgenic mice for other genes related to Alzheimer's disease have also been commonly used. Tg-Smpdstop mice were crossed with Tie2-Cre mice to obtain acid sphingomyelinase (ASM) overexpressing mice [12]. According to the earlier reports, ASM has been found to be involved in the pathogenesis of AD. Using the ASM overexpressing animal model, BJ Choi et al demonstrated that the plasma ASM correlated with Aβ accumulation, increased inflammation and decreased microglia phagocytic function in the central nervous system, suggesting the involvement of plasma ASM in the development of AD [12]. Increased astrocytic Ca2+ signaling is another potential mechanism of the AD development. Tg-ArcSwe mouse model has been used to show an impaired astrocytic Ca2+ response to locomotion and an uncoupling of astrocytic Ca2+ signaling in plaque-bearing mice [13]. Moreover, the authors have found a reduced norepinephrine signaling during spontaneous running suggesting a possible mechanistic explanation of the observed reduced astrocytic Ca2+ responses [13]. Since loss of function of triggering receptor expressed on myeloid cells 2 (TREM2) correlates with elevated risk of AD, ATV:TREM2, a TREM2-activating antibody containing a monovalent transferrin receptor (TfR) binding site in the Fc domain has been developed to facilitate blood–brain barrier transcytosis [14]. Using transgenic mouse models of AD, the study has shown that ATV:TREM2 stimulated brain microglial activity and mitochondrial metabolism [14].

Transgenic rats have also been developed to express human APP, human PS1-APP together, or the human tau protein [15].

Non-transgenic models are used to study AD, given that memory and cognition loss are common traits of all ageing animals, regardless of species. These animals include rodents such as rats and guinea pigs, as well as dogs and non-human primates. Rodents can also be induced to develop AD by administering a high-fat diet or via administration of chemicals in the brain, or via the imposition of radiofrequency lesions to the brain to induce cognitive deficits [15]. The advantages and disadvantages of these models are summarized in Table 1.

In vitro models facilitate a more direct and in-depth examination of AD-related pathologies on a cellular and molecular level. These include tissue models such as brain slices and cultured brain tissue, as well as cell models such as AD-derived induced pluripotent stem cells and neuroblastoma cells, and molecular simulation models such as antibubble biomachinery developed to study the impact of inflammation on AD (Table 1).

The aged canine has been used as model of beta-amyloid (Aβ) pathogenesis since it exhibits a similar decline in many different cognitive domains as patients with AD. The Canis familiaris amyloid precursor protein (APP) sequence shares 98% homology and the same amino acid sequence with the human APP [22]. Similar to humans, aged canines demonstrate significantly increased cortical and hippocampal atrophy as compared to non-aged control groups [23]. In addition, consistent with those abnormalities reported in humans, the molecular cascades underlying neuronal dysfunction [24] and decreases in specific neurotransmitter systems [25] in the dog brain include progressive accumulation of Aβ in diffuse plaques [26], cerebrovascular dysfunction [27] and cumulative oxidative damage of lipids, proteins, or DNA/RNA [23, 26, 28].

Aged beagle canines exhibit Aβ42-positive diffuse and compacted plaques, cerebral amyloid angiopathy and deposit of pyroglutamate (pyroGlu-3 Aβ) in blood vessels [29, 30]. Furthermore, the levels of Aβ1-42 and the Aβ42/40 ratio are significantly elevated in dogs manifesting mild cognitive deterioration than in either cognitively unimpaired or severely affected dogs [31]. Biliverdin reductase-A (BVR-A) is a pleiotropic enzyme associated with AD and statins have shown to reduce the risk of AD [32]. BVR-A protein levels correlated with lower oxidative stress and β-secretase 1 levels in the parietal cortex and enhancement in size discrimination learning of aged beagles treated with atorvastatin [33].

With regard to veterinary clinical trials in dogs, a new cell therapeutic approach, which was based on direct microinjection of autologous skin-derived neuroprecursors into the bilateral hippocampus, was initiated in aged dogs diagnosed with Canine Cognitive Dysfunction [34]. The efficacy was evaluated using the validated Canine Cognitive Dysfunction Rating Scale (CCDR). According to the trial results, 80% of the dogs showed improvement on the primary clinical CCDR endpoint, including two dogs, which demonstrated full syndromal reversal [34].

Taken together, aged dogs manifest several structural and molecular key elements of human AD and can be used as a model to elucidate the underlying mechanism of action of the disease further.

This article originally appeared as a section in an earlier Alzheimer's disease article. Dr. Zacharoula Konsoula contributed to the section on the aged dog in October 2018.