Revisiting rodent models: Octodon degus as Alzheimer’s disease model?

Alzheimer’s disease primarily occurs as sporadic disease and is accompanied with vast socio-economic problems. The mandatory basic research relies on robust and reliable disease models to overcome increasing incidence and emerging social challenges. Rodent models are most efficient, versatile, and predominantly used in research. However, only highly artificial and mostly genetically modified models are available. As these ‘engineered’ models reproduce only isolated features, researchers demand more suitable models of sporadic neurodegenerative diseases. One very promising animal model was the South American rodent Octodon degus, which was repeatedly described as natural ‘sporadic Alzheimer’s disease model’ with ‘Alzheimer’s disease-like neuropathology’. To unveil advantages over the ‘artificial’ mouse models, we re-evaluated the age-dependent, neurohistological changes in young and aged Octodon degus (1 to 5-years-old) bred in a wild-type colony in Germany. In our hands, extensive neuropathological analyses of young and aged animals revealed normal age-related cortical changes without obvious signs for extensive degeneration as seen in patients with dementia. Neither significant neuronal loss nor enhanced microglial activation were observed in aged animals. Silver impregnation methods, conventional, and immunohistological stains as well as biochemical fractionations revealed neither amyloid accumulation nor tangle formation. Phosphoepitope-specific antibodies against tau species displayed similar intraneuronal reactivity in both, young and aged Octodon degus. In contrast to previous results, our study suggests that Octodon degus born and bred in captivity do not inevitably develop cortical amyloidosis, tangle formation or neuronal loss as seen in Alzheimer’s disease patients or transgenic disease models.


Introduction
Senile plaques, a hallmark of Alzheimer's disease (AD), were long suggested to initiate the destructive cascade to progressive neuronal dysfunction and death. Nowadays small, soluble oligomers of β-amyloid (Aβ) are deemed the primary toxic species [1]. These oligomers disrupt a variety of receptors [2], increase membrane permeability [2] and are suspected to induce hyperphosphorylation and aggregation of tau [3]. Physiologically, Aβ is largely eliminated from the brain by LRP1 [4] and several ABC transporters (reviewed in [5,6]). The vast majority of cases occur sporadically and a large series of risk factors have been identified, including age, type 2 diabetes, high blood pressure, and various genetic factors like specific alleles of apolipoprotein E (APOE) [7,8].
A small proportion of AD cases involve genetic variations which entail alterations in amount, ratio or amino acid sequence of Aβ [9]. However, these rare inherited forms are the fundament of both, disease models and our current understanding of AD. Due to a lack of alternatives, the main focus lies on the usage of these genetically manipulated research animals (thus non-sporadic AD models), which restricts the progress of research and limits the scope of detailed analyses. To successfully combat the sporadic form of AD, models that develop the disease on a more 'natural' basis would certainly help to understand the underlying mechanisms which are essential for developing advanced and efficient therapy options.
The South American rodent Octodon degus (degu) may be a promising candidate for physiologically modelling sporadic AD, as it was reported to develop the 'full range of AD-like pathologies' [10] without any genetic manipulation. While wild degus have only a limited life expectancy (mean: <1 year; common max: 3-4 years), captivity lowers mortality and increases mean life span to 5-8 years [11] (Ebensperger L.A. & Hayes L.D., unpublished data). This captivity-dependent, aged phenotype combined with the highly homologous Aβ sequence [12], differing only in one amino acid from human Aβ (see Fig. 1b), might be the main reasons for their vulnerability. Thus, during the past years wild-caught and captive born degus were used in AD research [12][13][14][15] and have been referenced in numerous review publications [10,[16][17][18][19][20][21][22]. In this course, prominent intra-and extracellular Aβ deposits were reported in cortical and hippocampal areas of aged animals (>3 years) [12,19]. Furthermore, APP and Aβ positive axonal bulbs were observed in hippocampal white matter tracts of old animals (6 years) which preceded cerebral amyloid angiopathy [13]. Biochemical analysis of aged degus (5-years-old) indicated a correlation between Aβ*56 oligomers and tau phosphorylation on the one hand and decreased synaptic plasticity and impaired memory performance on the other hand [14]. However, the most recent study, analysing young (1 to 3 years old) and old animals (4 to 6 years old), found AD associated autophagy markers LC3 and p62 unchanged [15]. Furthermore, GFAP (glial fibrillary acidic protein), CD11b expression, oxidative stress and apoptosis markers were unaffected in cortices but elevated in hippocampi of old animals. Cortical and hippocampal levels of AD-linked IL-6 seemed increased in old degus [15].
The aim of the present study was to critically reevaluate the suitability of degus as 'natural' AD model. To characterize amyloidosis and tau deposition, different cortical and hippocampal regions of young (1-year-old) and aged (5-years-old) wild-type, colony-bred degus were screened for neurodegenerative changes.

Materials and methods
Drugs and chemicals used in the study were purchased from Carl Roth, Karlsruhe, Germany; except for those specifically mentioned.

Animals
Wild-type degus used in the present study were bred at the Institute of Biology, Otto-von-Guericke-University Magdeburg. Degus were housed in a climate-controlled environment (22°C, 55 ± 5 % humidity) on a 12 h light/dark cycle with an enriched environment. Social enrichment was provided by housing in same sex groups of up to 4 animals. Sensory, motor and cognitive enrichment was provided by housing the animals in large cages (510x420x680 mm 3 , EBECO, Castrop Rauxel, Germany), equipped with a drinking bottle, burrows for hiding, a running wheel (Europet-Bernina International, Iserlohn, Germany) for physical exercise and material for nest building. Nutritional Fig. 1 Comparison of human and rodent Aβ sequences. Part a shows a simplified phylogenetic tree of rodents. In b, human and rodent sequences of Aβ 42 are compared. The Aβ sequence of guinea pigs (green) equals the human sequence. Chinchillas, degus and naked mole rats (NMRs) share the same sequence (yellow) with one variation at position 13 as compared to humans. The lesser Egyptian jerboa has an additional difference at position 10 (orange). A larger group of rodents, including mice and rats, show 3 sequential differences at 5, 10 and 13 (red). This sequence is often erroneously referred to as "rodent Aβ", as it is the most frequent sequence in rodents. The uniformity of human and degu tau at the analysed phosphorylation sites is shown in c enrichment was provided by providing dry bread, fresh vegetables and fresh tree branches in addition to the regular pellet food (rat diet pellets and cereals; ssniff Spezialdiäten, Soest, Germany) ad libitum. A first cohort of degus was immunohistochemically analysed (J.P.) in two groups: young (12 months) and aged (60 months), both sexes, at least two animals per sex and age group. A second cohort from the same colony was independently analysed in another neuropathological laboratory (C.K.) and used for quantitative and western blot analyses: seven animals (age in months: 3, 24, 25, 56, 56, 65, and 65) Tg-mice, harbouring mutant human amyloid precursor protein (KM670/671NL) and presenilin 1 (L166P) both driven by murine Thy1.2-promotor [23], were kindly provided by the University of Tübingen, Germany. All mice were housed in a climate-controlled (22°C) environment on a 12 h light/dark cycle in same sex groups of up to 4 animals and free access to food and water. All experiments were conducted in accordance to the EU and state law of Saxony-Anhalt and approved by the local animal ethics committee.

Statistics
For all quantifications, animals of either sex were used (n ≥ 2 per sex; n ≥ 4 per group), statistical significance (p ≤ 0.05) was determined using unpaired t-tests with Welch's correction (Prism 6, GraphPad Software La Jolla, CA, USA).

Results
Routine histological Haematoxylin and Eosin (H&E) stains of young (1 year) and old (5 years) wild-type degus revealed normal age-related changes in the examined brain regions, but no obvious signs for specific lesions, neurodegeneration, or neuronal displacement (Fig. 2a, b). Neuronal marker NeuN unveiled no generalized loss of cortical neurons between both time points (Fig. 2c, d, m). Microglia-specific stains of ionized calcium-binding adapter molecule 1 (IBA1), which can unravel early signs of pathology shown by localised microglia accumulation, displayed homogenously distributed populations of resting microglia in young and aged animals (Fig. 2e, f). Individual differences in young and aged degus were apparent, but the cortical coverage was not significantly different between the two groups (Fig. 2n). Cortical clustering of microglia as seen in pre-depositing APP-transgenic mice, pinpointing towards starting Aβ deposition, was not detected. Cortical astroglia (GFAP-positive) were nearly exclusively located in cortical layer 1 and around blood vessels (Fig. 2g, h, o), without any age-dependent changes in spatial distribution or intensity.
To evaluate the extent of any cortical amyloidosis and neuronal destruction, a modified Campbell-Switzer stain was applied, but no extracellular deposits were exposed (Fig. 2i, j). This result was supported by thioflavin T stains which revealed no specific cortical fluorescence as well (Fig. 2k, l).
To examine potentially undiscovered amyloid deposits, we performed more sensitive immunohistochemical stains by employing commonly used antibodies against different Aβ epitopes (clones 6F3D, 4G8, 6E10). The epitope of clone 6E10 is located Nterminal of the H13R substitution (within amino acids 3-8). Besides high unspecific background staining, limited intracellular immunoreactivity was detected in all cortices of young and aged degus. However, extracellular deposits (e.g. plaques) could not be traced in any of the examined brain regions (Fig. 3a, b). The epitope of anti-Aβ antibody clone 6F3D (amino acids 8-17) includes the H13R substitution and showed neither intra-nor extracellular immunoreactivity in young or aged animals (Fig. 3c, d). Likewise, no aggregates could be detected using the 4G8 antibody (Fig. 3e, f ) with an epitope C-terminally of the H13R substitution (amino acids: [18][19][20][21][22]. The lack of agedependent, immunohistological changes in degus was independently confirmed by a second neuropathological laboratory (C.K.) in an additional study (1 to 5 year old degus from the same colony, anti-amyloid stains using clones 6E10 and IC16; data not shown).
Levels of cortical and hippocampal Aβ 40 and Aβ 42 in young and aged degus were quantitatively measured using immunoassays (Fig. 4), revealing very low levels of soluble and membrane-bound Aβ and low levels of protein-bound and insoluble Aβ. The levels of insoluble Aβ did not age-dependently change. However, the concentration of insoluble Aβ was generally several magnitudes below those of established AD mouse models, and even lower than those of wild-type naked mole rats and guinea pigs ( Table 2). The comparison of wild-type degus, wildtype and transgenic mice demonstrates that aged wild-type degus very closely resemble histological parameters of wild-type mice in terms of Aβ deposition and unspecific activation of microglial and astroglial cells (Fig. 5). Fig. 2 Immunohistochemical analysis of young and aged degus. H&E stain revealed normal age-related changes but no signs for lesions, neurodegeneration, or displacement in young (1-year-old, a) and aged (5-years-old, b) animals. Density of cortical neurons (NeuN-satin) remained virtually unchanged in aged degus (d), compared to young (c). IBA1-stain (e, f) revealed homologues populations of resting microglia cells (young, e; aged, f)). GFAP Immunoreactivity was slightly decreased in aged animals (h), but spatial distribution (layer 1, surrounding vessels) was similar (young: g; aged: h). Campbell-Switzer stain unveiled neither extracellular plaques nor tangles (i, j). Thioflavin T likewise indicated no amyloid plaques (young, k; aged, l). Semi-automatically determination of neurons (m) as well as microglial cells (n) and astrocytes (o) in cortices revealed no significant changes in aged animals. Scale bars = 100 μm. Data is presented as mean ± SEM (n ≥ 4) Phosphorylated tau is the second protein accumulating during disease progression and another histopathological hallmark of AD [31]. We utilized antibodies against different epitopes of phosphorylated tau to screen for neurofibrillary tangles. Although sequences of human and degu tau vary, the analysed phosphorylation sites (Ser202/Thr205, Thr212/Ser214 and Thr231) are identical (Fig. 1c). AT8 (Ser202/Thr205) labelled cortical neurons in young and aged animals (Fig. 6a, b) and AT100 (Thr212/Ser214) showed nuclear-localized reactivity (Fig. 6c, d). AT100 epitope is known for nuclear co-localization and considered not tau-specific, as it appears likewise in tau knockout mice [32]. AT180 (Thr231) equally stained cortical neurons of young and aged degus (Fig. 6e, f). In sum, the detected tau did morphologically not correspond to neurofibrillary tangles and showed no age-dependent intensification.  Moreover, the independent quantification of tau revealed neither elevated levels of total tau, nor an increase in phosphorylated or insoluble fraction (Fig. 7).

Discussion
In the near future, aging societies will be particularly challenged by age-related diseases demanding intensive care. Exceptional research endeavours are necessary to face these approaching challenges. Thus, reliable and accurate animal models are one of the major prerequisites in basic research. Although transgenic rodent models of AD are constantly refined, no model yet mimics all pathological features of this complex disease. More accurate models maintaining  Compared to wild-type degus (d, g) and mice (e, h), transgenic mice manifest with pronounced micro-(f) and astrogliosis (i). Scale bars = 200 μm the beneficial characteristics of rodents would lead to a better understanding and more expedient therapeutic approaches. Degus were described as a promising natural model of Alzheimer's disease during the last years by a Chilean group. However, in our studies we were unable to detect any systematic occurrence of the typical histopathological hallmarks of AD in relation to age. Haematoxylin and Eosin as well as NeuN stains showed not more than slight differences between young and aged degus, rather indicative for the natural aging process than pathological neurodegeneration. In contrast to previous results from Inestrosa et al. [15], we could not detect elevated GFAP expression in old degus compared to young animals. The lack of signs for microglial and astrocytic inflammation further reiterates the absence of a pathological degeneration in the examined brains of the old degus.

Aβ pathology
The sparse intracellular reactivity of anti-Aβ clone 6E10, which was lacking in clone 4G8, clone 6F3D, Campbell-Switzer and thioflavin T stains, most likely indicates an unspecific reaction [33]. In contrast, no sign of any extracellular deposition of Aβ was detected in aged animals by any of the applied staining methods (Fig. 3). Quantitative measurements underpinned the absence of considerable amounts of insoluble Aβ (Fig. 4) and revealed Aβ-levels that are in the same range as those in wild-type mice [29] and below those of wild-type naked mole rats [34]. Consistent with results of van Groen et al. no significant neuronal loss was found in the brain of 5-years-old degus [13]. These findings are in sharp contrast to observations in brains of degus obtained from their natural habitat, in which prominent intra-and extracellular Aβ deposits in cortices and hippocampi of aged animals (>3 years) were reported [12,19]. These differences may, at least in part, be caused by different rearing conditions (laboratory housing versus natural wildlife conditions) and it has to be considered that in their wildlife habitat the animals are exposed to stress, may suffer from hypertension, viral infections and diabetes, i.e. known risk factors contributing to the aetiology of AD and the early development of AD-type neuropathology.
Furthermore, the single amino-acid-difference between degus and humans at position 13 (histidine to arginine) affects a histidine residue (His 13 ) which is crucial for aggregation and toxicity of Aβ. His 13 is involved in early N-terminal β-sheet formation [35] and a substitution lowers aggregation propensity [36], neuronal binding [37], and cytotoxicity [36]. Moreover, His 13 is involved in the coordination of metal ions [38] and methylation or  substitution by arginine, as seen in degus, lowers the affinity for metal ions and thus depletes aggregation [38][39][40] and attenuates toxicity [41,42] of Aβ.
Two other species which are related to degus share a similar Aβ sequence, but, despite higher life expectancies, lack the neuropathological features as reported for degus. Naked mole rats (Heterocephalus glaber) have the identical Aβ sequence (see Fig. 1) and an exceptional lifespan of more than 30 years [34]. Although young naked mole rats naturally exhibit pronounced oxidative stress [43] and Aβ levels similar to 3xTg-AD mice [34,43], they do not develop amyloid plaques with age [34]. Furthermore, naked mole rats even present with high levels of phosphorylated tau without any tangle formation [44]. In Guinea pigs (Cavia porcellus), with a human identical Aβ sequence (see Fig. 1) and a lifespan similar to degus (average 5-7 [45]), dense amyloid deposits do not occur [45], despite similar APP processing [46,47] and high β-secretase activity [47].

Tau pathology
The additional screening for tau deposition, the second aggregating protein in AD, revealed similar intracellular reactivity in young and aged degus using phosphoepitopespecific antibodies AT8 and AT180. AT100 staining showed the previously described, unspecific nuclear localization [32]. Biochemical analysis did not reveal an age-dependent increase of total, insoluble or phosphorylated tau (Fig. 7). Some variability observed in the levels of total tau or insoluble tau could hint subsets with different aggregation propensities but the very same animals did not exhibit tau pathology in IHC, and larger number of animals would be needed to identify the existence of such subsets. Hence, no evidence for a pathological deposition of tau could be detected in the examined animals.

Methodological considerations
The animals used in a variety of studies were collected from different sources [13,48,49], including animals caught in the wild [12,50], the latter does usually not allow a precise age determination and thus hinders precisely controlled analyses. However, standardised housing conditions as used for the degus examined in the present study seem to prevent the development of ' ADlike' pathology described for wild-caught animals. Nonetheless, it would be fascinating to decipher the factors inducing the histopathological and biochemical changes in degus previously described [12,13,15]. Moreover, not only the particular species or the specific amino acid sequence seems influence the extent of amyloid deposition, but the genetic background likewise enfolds a strong effect [26]. As degus are not yet an established research model, they lack a defined, stable and characterized inbred genetic background.
An interesting approach of separating old degus with severely aberrant behaviour as disease model takes high individual variability into account and indicates that ' AD-like' pathology might not necessarily develop in old degus. However, even in this selected subgroup with increased levels of inflammatory and oxidative stress markers, no correlation between altered behaviour and specific neuropathological symptoms could be established [50]. The stated impairments of spatial memory and cognition in aged degus [14] may therefore be just a part of the normal aging process, since physiologic aging is linked to significant impairments in memory [51], cognition [52] and hippocampal long-term potentiation [53] in mice as well. Thus, symptoms of normal aging may not be misinterpreted to model AD.

Conclusion
Octodon degus was re-evaluated in the context of existing rodent AD models and human AD pathology. Performing immunohistological and molecular analyses of young and aged animals, we were able to show exclusively normal age-related cortical changes without indications for extensive degeneration as seen in patients with dementia and transgenic AD mice. Neither significant neuronal loss nor enhanced microglial activation were observed in aged animals of our degu population. Furthermore, no amyloid accumulation or tangle formation as seen in sporadic Alzheimer's disease patients could be determined. Phosphoepitope-specific antibodies against tau species displayed similar intracellular neuronal reactivity in both, young and aged degus.
Moreover, we highlighted some previous results, which stand in contrast to the assumption of degus as natural AD model and seem to be thus far neglected. Bearing that in mind, assessment of degus as AD models should be meticulously done and receive particular attention. Currently, it is not clear if unnoticed environmental or rearing factors might play a role in triggering AD-like neuropathology. Therefore, we conclude that presently, the rodent Octodon degus is neither a superior model which is more suitable than other frequently used rodent models nor a 'natural' model of Alzheimer's disease in general.