OUP user menu

Moderate Environmental Enrichment Mitigates Tauopathy in a Neurofibrillary Tangle Mouse Model

Inbal Lahiani-Cohen MSc, Athanasios Lourbopoulos MD, PhD, Ester Haber PhD, Lea Rozenstein-Tsalkovich, Oded Abramsky MD, PhD, Nikolaos Grigoriadis MD, PhD, Hanna Rosenmann PhD
DOI: http://dx.doi.org/10.1097/NEN.0b013e318221bfab 610-621 First published online: 1 July 2011


Epidemiological studies show that stimulating activities reduce therisk of dementia. In animal models of Alzheimer disease, there have been conflicting results of the effects of environmental enrichment (EE) on disease-related amyloid pathology. Here, we tested the direct effect of EE, independently of amyloid pathology, on brain neurofibrillary tangles (NFTs), which best correlate with dementia. We exposed transgenic mice (E257K/P301S-Tau-Tg driven by the natural tau promoter) to moderate nonstrained EE or regular environment. Concomitant with neurogenesis, we detected a decrease in NFT burden and a decrease in the activation of microglia in EE versus regular-environment mice. There was also a trend toward improvementin cognitive tasks in the EE mice. Increased immunoreactivity of brain-derived neurotrophic factor, which is involved in the regulation of tau phosphorylation, was detected in the EE mice, suggesting its possible involvement in the beneficial effects on NFTs and other parameters in the EE mice. These results suggest that NFTs may be directly responsive to environmental stimulating activities and that even nonstrained activities may mitigate tauopathies independent of theinvolvement of amyloid.

Key Words
  • Alzheimer disease
  • Brain-derived neurotrophic factor
  • Environmental enrichment
  • Neurofibrillary tangles
  • Neurogenesis
  • Tau
  • Tauopathy


Retrospective epidemiological studies suggest that stimulating activities, specifically a high level of education, a mentally demanding occupation, and challenging/active leisure and physical activity over a lifetime, are associated with a reduced incidence of dementia (1-3). Prospective studies have also shown that complex mental activity is associated with a reduced risk of developing dementia (4-6), and there is further support for the benefit of physical activity (7). These effects are generally attributed to brain cognitive reserve, that is, increased neuronal connectivity in brain regions involved in learning and memory that may compensate for and withstand detrimental symptoms of clinical dementia before they manifest.

Stimulating activities in animals using “environmental enrichment” (EE), which involves enhanced levels of sensory, cognitive, and motor stimulation, elicit various plastic responses in adult brains, including increased dendritic arborization and neurogenesis and improved learning (8). These effects are likely modulated by microglia (9-12). The effects of EE have been intensively investigated in transgenic (tg) models of Alzheimer disease (AD), particularly in mice with amyloid plaque neuropathology. These studies demonstrate cognitive improvement, but effects on amyloid pathology are variable (most studies report a significant decrease [13-16]), but involvement of both amyloid-related and -unrelated mechanisms (17) and no decrease in amyloid deposition (18) have been reported. Two studies have shown an increase in amyloid load in enriched animals, although cognitive deficits were mitigated (19, 20). These discrepancies suggest that amyloid pathology may not be the primary or only target responding to stimulating activities. The complexity of the responsiveness to EE has been further supported by the finding that ApoE4, the most prevalent AD genetic risk factor, had a detrimental effect of enhanced apoptosis, as opposed to the beneficial effect of EE in ApoE3 mice (21, 22).

Much less is known about the effect of EE on neurofibrillary tangles (NFTs), the aggregates of the phosphorylated microtubule-associated protein tau. Neurofibrillary tangles have been considered the best correlate of dementia in AD (23-25). Neurofibrillary tangles are also characteristic of other tauopathies that lack amyloid plaques, such as frontotemporal dementia, Pick disease, and others. In tg mice that express anti-nerve growth factor (NGF) antibodies, EE induced a decrease in amyloid burden without any affect on tau pathology (13); this result may be related to the NGF deficits (26-28). A recent study in APPswe/PS1{Delta}E9 mice (which have early tau pathology in the form of phosphorylated tau rather than mature aggregated NFTs in addition to amyloid pathology) showed a decrease in phosphorylated tau (29). This might be secondary to the decrease in amyloid oligomers because tau phosphorylation is regulated by pathways downstream of amyloid pathology (30).

Here, we studied the effects of EE on NFTs independent of other brain pathologies using the tauopathy mouse model that we recently generated. It expresses pathogenic human tau (E257T/P301S) under the regulation of the natural tau promoter (tolerated level of expression) and is characterized by pure NFT pathology with cognitive deficits (31). We found a decrease in NFT burden in the tauopathy mice exposed to EE accompanied by some cognitive improvement and an increase in the brain-derived neurotrophic factor (BDNF), the neuroprotective factor that has been reported to reduce tau phosphorylation via the tyrosine receptor kinase B (32). Leem et al (33) recently showed a decrease in phosphorylated tau (but not in aggregated NFTs or in cognitive deficits) after a long-term high-intensity exercise (possibly representing activities suitable for professional athletes) in mice that overexpress human tau. Our results suggest that NFTs developed under authentic tau regulation are reduced directly, and not in a manner depending on amyloid pathology, after a moderate degree of stimulating activities (which may reflect activity levels feasible in the general population).

Materials and Methods


E257T/P301S human tau protein (DM-Tau-tg) mice (31) were further crossed with C57BL mice for more than 6 generations to obtain tg offspring that could be identified by polymerase chain reaction analysis of tail genomic DNA. Experiments were approved by the animal ethics committee.

Environmental Enrichment and Study Design

Four-month-old DM-Tau-tg mice were housed for 9 months in EE or regular-environment (RE) cages. For moderate EE, the cages were large (610 × 435 × 215 mm) and equipped with a running wheel for tolerated voluntary exercise and differently shaped objects (tunnels, boxes, cubes, balls, shelters, ladder, labyrinth) that were substituted with others once a week in such a way that the same objects were used every other week. Regular environment consisted of standard laboratory cages (420 × 260 × 180 mm) without objects. Control groups consisted of RE DM-Tau-tg mice and RE wild-type (WT)-non-tg mice.

Two separate, independent experiments were performed. Experiment 1 included 13 DM-Tau-tg mice (5 males, 8 females) exposed to EE (EE-DM-Tau-tg mice), 11 DM-Tau-tg mice (5 males, 6 females) exposed to RE (RE-DM-Tau-tg mice), and 14 (6 males, 8 females) non-tg nonenriched mice (RE-WT). Experiment 2 included 9 EE-DM-Tau-tg (3 males, 6 females), 9 RE-DM-Tau-tg (3 males, 6 females), and 11 (5 males, 6 females) RE-WT mice.

Tissue Collection

Animals were killed at the end of each experiment under deep anesthesia and quickly transcardially perfused with phosphate-buffered saline. Brains were quickly removed; 1 hemisphere was frozen at −80°C, and the other was postfixed for 20 hours in 4% paraformaldehyde in phosphate-buffered saline (pH 7.2, ice-cold) and processed for sagittal paraffin sectioning at 6 μm.

Histology and Immunohistochemistry

Paraffin-embedded sections were silver-impregnated by Gallyas's (34) silver method. Neurofibrillary tangles were also detected by immunohistochemistry (IHC) using the AT8 mouse monoclonal Ab (Innogenetics, Ghent, Belgium) that recognizes tau phosphorylated at 202/205, as well as tau pathology in DM-Tau-tg mice (31). Microglial cells were stained with biotinylated tomato lectin (3 mg/mL Lycopersicon esulentum tomato; Sigma, St Louis, MO) and the Iba-1 antibody (WAKO, Osaka, Japan). Neuroblasts were detected using the doublecortin marker (DCX; Santa Cruz Biotechnology, Santa Cruz, CA). Brain-derived neurotrophic factor immunoreactivity was detected using a rabbit antibody (N-20; Santa Cruz Biotechnology).

Immunostaining was performed using the Mouse-on-Mouse system (Vector Laboratories, Burlingame, CA) for AT8, the EnVision System HRP Kit (DAKO Cytomation, Glostrup, Denmark) for BDNF and Iba-1, and the LSAB technique (LSAB2 System HRP; DAKO Cytomation) for lectin. Paraffin sections were deparaffinized and rehydrated in graded alcohols, and antigen retrieval was performed with citrate buffer pH 6 in a food steamer device (Braun, Kronberg, Germany) for 60 minutes. Endogenous peroxidase was blocked with 0.3% H2O2 in methanol followed by incubation in the appropriate blocking buffer (10% fetal bovine serum in Tris-buffered saline for all primary antibodies used except lectin; 0.3% Triton X in Tris-buffered saline for lectin) for 30 minutes. Sections were incubated overnight at 4°C with the primary antibodies AT8 (1:50), Iba-1 (1:2000), DCX (1:200), and BDNF (1:800) for 2 hours for lectin; 3,3′-diaminobenzidine tetrahydrochloride (DAB; DAKO) was used as a chromogen, and sections were counterstained with hematoxylin. Slides were then dehydrated in graded ethanols and covered with entelan.

Double immunofluorescence staining for combinations of Iba-1 and M1 or M2 microglia/macrophage phenotypes were performed using similar protocols. Sections from experimental autoimmune encephalomyelitis experiments were used as positive controls in these studies. Chemicals were purchased from Sigma unless specified otherwise.

Neuropathologic Evaluation

Neuropathologic evaluation was performed on 4 brain sections, spaced approximately 60 to 100 μm apart, under light microscope (Zeiss Axioplan 2; Carl Zeiss MicroImaging GmbH, Gottingen, Germany), with the aid of a CCD camera (Nikon, Tokyo, Japan) (35-37). Evaluation of neuropathology was blindly performed by 2 independent observers under 20× optical fields using a rectangular grid applied into the prefrontal lenses (prefrontal grid), with dimensions 610 × 610 μm for the 20× magnification. Higher magnification (40×) was used where necessary. The surface of each section was scanned, and the mean corresponding positive counted cells per surface area of the grid were used to calculate their mean densities per squared millimeter. Lectin- or Iba-1-positive microglia/macrophages, Gallyas-positive dark neurons, and AT8-positive cells per squared millimeter were evaluated in the cortex, hippocampus, thalamus, striatum, and brainstem. Cells positive for DCX were counted under 40× optical fields in the subgranular zone of the hippocampus and subventricular zone/rostral migratory stream. Four sagittal sections containing these structures and spaced approximately 60 to 80 μm apart were used. Data from DCX-positive cells were expressed as “cells per field."

Microglial activation states were quantified by determining the ramification index (RI) ranging from 0 for ramified “resting” cells to 1 for “active” amoeboid cells (38, 39), using the ImageJ software. Briefly, digital images from lectin- and Iba-1-positive microglia in the cortex, hippocampus, striatum, and brainstem were captured under 40× optical fields; 20 cells per section in 3 sections were evaluated per animal. These digital images were then separately processed with the binary mode of the ImageJ software (ImageJ ver.1.38; National Institutes of Health) to calculate the RI, using the cell perimeter and area in the following formula: RI = (4π × area)/perimeter2 (38).

To evaluate BDNF immunoreactivity, digital images were captured under 20× optical fields (3 sections per animal, 20 images per section evenly and repetitively distributed among the cortex, hippocampus, and brainstem) and processed in the ImageJ software using an adopted method previously described for semiautomated quantification of immunostained specimens (40). Briefly, digital images of DAB-stained sections were analyzed into their red-green-blue components; the blue channel was selected because DAB-stained areas demonstrate the greatest intensity levels for this channel (41, 42). The blue channel was then thresholded to segment the DAB-stained regions, binary transformed into black-white images by the ImageJ software, and compared with the original 3-colored image to validate its transformation. Automatic calculation of the total “black” pixels extracted the BDNF-immunostained area in each image. These data were used to express the mean BDNF relative signal in pixels.

Enzyme-Linked Immunosorbent Assay

The mouse BDNF enzyme-linked immunosorbent assay kit (Acris Antibodies, Herford, Germany) was used according to the manufacturer's instructions for analyzing BDNF levels in hippocampal homogenates of the EE-DM-Tau-tg mice and RE-DM-Tau-tg mice obtained as described (31).

Behavioral Examinations


The T-maze test was used for assessing the spatial short-term memory and alternation behavior, that is, determining the mouse's ability to recognize and differentiate between a new unknown and a familiar compartment (31, 43). The T-shaped maze was made of plastic with 2 arms 45 cm in length that extended at a right angle from a 57-cm-long alley. The arms had a width of 10 cm and were surrounded by 10-cm-high walls. The test consists of 2 trials with an intertrial interval of 1 hour, during which time the animals were put back to their home cages. During an 8-minute acquisition trial, one of the short arms was closed. In a 3-minute retention trial, mice had access to both arms and to the alley. Numbers of entries into the unfamiliar arm and the time spent in the unfamiliar arm were recorded. Mice normally tend to enter more times and spend more time in the new unknown arm than in the familiar one or in the alley.

Eight-Arm Maze

For the evaluation of spatial memory related to hippocampal-cortical function, we used the 8-arm radial maze scaled for mice (44, 45). The animals were introduced to the radial maze and were observed until they made entries to all 8 arms or until they completed 25 entries, whichever came first. The number of entries needed to complete a full round of 8 arms (once to each of the arms) within the 25 trials was recorded. The lower the number of entries needed, the better the cognitive score (best = 8, worse = 25). Maze performance was calculated each day for 5 consecutive days. Results were presented as area under the curve using the formula as follows: (Day 2 + Day 3 + Day 4 + Day 5) − 4 × (Day 1) with the entries indicated by day number (31, 46, 47).

Data Analysis

Results are presented as mean ± SEM. For comparisons of the quantity or intensity of stained cells and behavioral performance in the mazes between the study groups, the unpaired parametric Student t-test or nonparametric Mann-Whitney U test were used when appropriate.


Decreased NFT Burden in EE-DM-Tau-tg Mice

We exposed the DM-Tau-tg mice to an EE paradigm that involved sensory, cognitive, and motor stimulation in the form of a moderate/nonstrained paradigm (substituting the objects with others only once a week, in such a way that the same objects were used every other week, as opposed to other reported protocols where objects were changed or completely substituted more often [48]), together with a tolerated voluntary nonforced physical exercise. The EE started at 4 months of age (ie, 2 months before the onset of NFT pathology in this model) and continued for approximately 9 months. This long-term exposure to EE was used to allow high effectiveness, in a manner similar to that of O'Callaghan et al (49) who showed long-term EE protection against age-related hippocampal changes.

In the first experiment, the EE-DM-Tau-tg mice showed a significant decrease of tau pathology/ NFT burden, as indicated by the significant lower burden of cells stained for the Gallyas staining for NFTs/tau pathology in the brain relative to the RE-DM-Tau-tg mice (5.46 ± 0.34 vs 9.6 ± 0.6, a 43.3% lower burden in enriched-mice [p < 0.0001]): the decrease was 53.6% (p < 0.0001), 56.6% (p = 0.006), and 18.9% in the cortex, hippocampus, and brainstem, respectively (Figs. 1A, B). This decrease in NFTs in the brain was further confirmed by IHC with AT8 (0.34 ± 0.05 vs 0.54 ± 0.06, a decrease of 37.1% from RE- to EE-DM-Tau-tg mice [p = 0.01]): the decrease was 34.4% and 45.7% (p = 0.05) in the cortex and brainstem, respectively (Figs. 2A, B). In the second experiment, there was a significantly lower burden of Gallyas-positive cells in the brain of EE- versus RE-DM-Tau-tg mice (7.44 ± 1.1 vs 15.3 ± 1.7, a 51.4% lower burden [p < 0.0001]): the decrease was 50.2% (p < 0.0001), 35.7%, and 35.2% (p = 0.08) in the cortex, hippocampus, and brainstem, respectively. This decrease in NFTs was confirmed by IHC with AT8 (0.8 ± 0.04 vs 1.04 ± 0.06, a decrease of 23.1% from RE- to EE-DM-Tau-tg mice [p = 0.003]): the decrease was 14% and 31.9% (p = 0.002) in the cortex and brainstem, respectively.


Reduced neurofibrillary tangle (NFT) burden assessed by the Gallyas method in environmentally enriched (EE)-DM-Tau-tg mice. (A) There are fewer Gallyas-positive cells (arrows) in the cortex, hippocampus, and brainstem of EE-DM-Tau-tg versus regular-environment (RE)-DM-Tau-tg mice. Scale bars = 100 μm. Insets are higher-magnification images from the same fields. (B) Semiquantitative assessment of Gallyas staining (p < 0.0001, p < 0.0001, and p = 0.006 for total brain, cortex, and hippocampus, respectively).


Reduced neurofibrillary tangle (NFT) burden assessed by AT8 immunohistochemistry in environmentally enriched (EE)-DM-Tau-tg mice. (A) A decrease in AT8-positive cells is evident in the cortex, hippocampus, and brainstem of EE-DM-Tau-tg versus regular-environment (RE)-DM-Tau-tg mice. Scale bars = 100 μm. Insets are higher-magnification images from the same fields. (B) Semiquantitative assessment of AT8 staining. p = 0.01 and p = 0.05 for total brain and brainstem, respectively.

Evidence for Neurogenesis in the Enriched-DM-Tau-tg

Neurogenesis occurs in adult mammalian brains, mainly in the hippocampal dentate gyrus and in the subventricular zone (50, 51), We assessed neurogenesis using the microtubule-associated protein DCX, which is expressed in newly formed neurons (52). More DCX-positive cells were detected in the EE-DM-Tau-tg mice versus RE-DM-Tau-tg mice, particularly in the subgranular zone in hippocampus (1.53 ± 0.15 vs 0.47 ± 0.1 [p < 0.0001]). Differences were less evident in the subventricular zone and rostral migratory stream (42.69 ± 10.7 vs 37.59 ± 10.24 [p = 0.76]) (Fig. 3). This result suggests that our EE paradigm was effective with respect to inducing neurogenesis.


Neurogenesis in environmentally enriched (EE)-DM-Tau-tg mice. (A) Greater numbers of doublecortin marker (DCX)-positive cells are seen mainly in the hippocampal subgranular zone (SGZ) and, to a lesser extent, in the subventricular zone (SVZ) and the rostral migratory stream (RMS) of EE-DM-Tau-tg versus regular-environment (RE)-DM-Tau-tg mice. (B, C) Semi-quantitative assessment of DCX in the SGZ in hippocampus (p < 0.0001) (B), and in SVZ and RMS (p = 0.76) (C). LV indicates lateral ventricle.

A Slight Decrease in the Activated State of Microglia in the EE-DM-Tau-tg Mice

By lectin staining, no significant difference in microglial burden was detected between EE- and RE-DM-Tau-tg mice (5.24 ± 0.35 vs 5.02 ± 0.32). Interestingly, there was lower RI (thinner processes) in EE- versus RE-DM-Tau-tg mice (0.0635% ± 0.03% vs 0.071% ± 0.03%, a decrease of 10.6% [p = 0.01]) (Fig. 4). Similar results were obtained with Iba-1 Ab, that is, there was no difference in microglial numbers between EE- and RE-DM-Tau-tg mice (11.40 ± 0.31 vs 10.98 ± 0.25), whereas there was a significant decrease in RI of EE- versus RE-DM-Tau-tg mice (0.0795% ± 0.003% vs 0.092% ± 0.004%, a decrease of 13.6% [p = 0.02]) (Fig. 5). These results suggest that, although microglial cell counts were not affected, they were less activated in the EE- DM-Tau-tg mice. Specific staining for discrimination of either an M1 or M2 phenotype of these microglia did not reveal any positivity of these cells, possibly indicating an M0 phenotype (data not shown). As expected, no perivascular macrophages were observed in any mice.


There is a slight decrease in activated microglia (assessed by lectin staining) (arrows) in environmentally enriched (EE)-DM-Tau-tg mice. (A) Microglial burden and activation in the cortex, hippocampus, and brainstem of EE-DM-Tau-tg and regular-environment (RE)-DM-Tau-tg mice. Although there are similar numbers of microglia (arrows) in the groups, they have a lower ramification index (RI) (less activated, thinner processes) in EE-DM-Tau-tg versus RE-DM-Tau-tg mice. Scale bars = 100 μm. Insets are higher-magnification images from the same fields. (B, C) Quantitative assessment of total burden of microglia (B) and of activated microglia (C) (p = 0.01).


There is a slight decrease in activated microglia (assessed by Iba-1 staining) in environmentally enriched (EE)-DM-Tau-tg mice. (A) Microglial burden and activation in the cortex, hippocampus, and brainstem of EE-DM-Tau-tg and regular-environment (RE)-DM-Tau-tg mice. Although there are similar numbers of microglia (arrows) in both groups, they have a lower ramification index (RI) (thinner processes) in EE-DM-Tau-tg versus RE-DM-Tau-tg mice. Scale bars = 100 μm. Insets are higher-magnification images from the same fields. (B, C) Quantitative assessment of total burden of microglia (B) and of activated microglia (C) (p = 0.02).

Increased BDNF in EE-DM-Tau-tg Mice

Immunohistochemistry with anti-BDNF Ab revealed that EE-DM-Tau-tg mice had increased BDNF staining compared with that in RE-DM-Tau-tg mice (increase of 19.2% [p = 0.01]). This was observed in cell processes and axons (Fig. 6). The increase was 35.7% (p = 0.001), 29.5%, and 9.7% (p = 0.003) in the cortex, hippocampus, and brainstem, respectively. Analysis of the BDNF levels in hippocampal homogenates by enzyme-linked immunosorbent assay revealed a trend of increased level in the EE-DM-Tau-tg mice relative to RE-DM-Tau-tg mice (29.52 ± 4.47 vs 24.08 ± 4.59 pg/μg protein, a nonsignificant increase of 18.4%). RE-DM-Tau-tg mice expressed lower BDNF levels than RE-WT mice (15.5% [p = 0.047], 27.6% [p = 0.017], 41.6% [p = 0.037], and 7.2% [p = 0.027], respectively) in total brain, cortex, hippocampus, and brainstem, respectively.


Increased brain-derived neurotrophic factor (BDNF) in environmentally enriched (EE)-DM-Tau-tg mice. (A) There is greater BDNF immunoreactivity in the cortex, hippocampus, and brainstem of EE-DM-Tau-tg versus regular-environment (RE)-DM-Tau-tg mice. Boxes in the photos indicate areas of magnified insets displaying the BDNF-positive processes and axons. Differences are indicated as “more and denser brown signals.” (B) Semiquantitative assessment of BDNF immunoreactivity. p = 0.01, p = 0.0001, and p = 0.003 for total brain, cortex, and brainstem, respectively.

Behavioral Tests

In the T-maze, EE-DM-Tau-tg mice showed a better performance than the RE-DM-Tau-tg mice, with a higher number of entries into the unfamiliar arm, reaching close to the number of the RE-WT mice (3.62 ± 0.28, 2.89 ± 0.3, and 3.71 ± 0.2 entries in EE-DM-Tau-tg, RE-DM-Tau-tg, and RE-WT mice, respectively). The difference between RE-DM-Tau-tg and RE-WT was significant (p = 0.04), but the EE-DM-Tau-tg mice (albeit showing a similar performance to that of RE-WT mice) were not significantly different from RE-DM-Tau-tg mice (p = 0.1). However, the similar performances of EE- DM-Tau-tg and RE-WT mice (3.62 ± 0.28 and 3.71 ± 0.2, respectively, p = 0.8) support the idea that enrichment restored cognitive function to the RE-DM-Tau-tg (Fig. 7A). A similar improvement in EE-DM-Tau-tg mice was found when calculating the time in seconds spent in the unfamiliar arm of the T-maze (not shown).


Cognitive tasks in environmentally enriched (EE)-DM-Tau-tg mice. (A) In the T-maze, EE-DM-Tau-tg mice show a nonsignificant better performance versus regular-environment (RE)-DM-Tau-tg mice, that is, there are higher numbers of entries into the unfamiliar arm (p = 0.1). The similar performance of the EE-DM-Tau-tg and the RE wild-type (WT) mice (nonsignificant difference [NS], p = 0.8) suggests that the low performance of the RE-DM-Tau-tg (vs RE WT mice, p = 0.04) was restored in EE-DM-Tau-tg mice. (B) In the 8-arm maze, EE-DM-Tau-tg mice needed fewer entries into the arms (calculated as area under the curve [AUC]) versus RE-DM-Tau-tg mice; their performance was similar to that of RE-WT mice, although there were no significant differences among the groups.

In the 8-arm maze, there was a similar trend of better performance in EE- versus RE-DM-Tau-tg mice, approaching the area under the curve of the RE-WT mice (6.55 ± 3.8, 10.67 ± 2.82, and 6.75 ± 2.49 entries in EE-DM-Tau-tg, RE-DM-Tau-tg, and RE-WT mice, respectively) but without reaching statistical significance (Fig. 7B).


Although firm epidemiological evidence points to the beneficial effect of stimulating activities on symptoms of dementia and AD, mechanistic support from experimental animal studies is limited. Despite intensive investigation, there is still controversy regarding the possibility that the protective effect of EE is mediated via the amyloid pathology, including the possibility that amyloid plaques may be exacerbated after EE (13, 14, 16-20, 29, 53, 54). Less focus has been placed on the responsiveness of NFT pathology to EE. We used an EE paradigm that may have a therapeutic and applicable advantage because it involves moderate sensory and cognitive stimulation and includes tolerated voluntary exercise. Moreover, the DM-Tau-tg mice model is driven by the natural tau promoter, leading to a tolerated level of expression of the tg mutant tau, which contrasts with some other tauopathy tg models in which unrelated promoters may lead to very high levels of tau transgene overexpression and are accompanied by motor deficits that are not necessarily related to tauopathy (55-58). Our finding that NFTs are reduced in various brain regions by EE indicates that the NFT pathology developed under natural physiological regulation responds directly to EE, that is, without involvement of amyloid pathology; moreover, this seemed to be accompanied by some cognitive benefit. The data also suggest that EE may protect not only against AD but also against other tauopathies in which only NFTs are detected. Neurofibrillary tangle reduction in response to a tolerated EE paradigm may have a broader applicability because exposure to moderate stimulating activities would be applicable to most of the population. In contrast, exposing tau pathology mouse model (under the regulation of the enolase-promoter) (59) to a long-term high-intensity exercise reduced the degree of tau phosphorylation (33). However, it is not clear whether this intensive physical activity (which may correspond more to professional athletes) also reduced mature NFTs or whether intensive physical activity had any beneficial effects on cognitive deficits.

The EE DM-Tau-tg mice showed evidence of neurogenesis, indicating that EE likely elicited plastic responses in the adult brain, specifically by neurogenesis (8, 51). Our finding that neurogenesis in EE DM-Tau-tg mice was particularly evident in the hippocampus is similar to the report that neurogenesis specifically occurred in the hippocampus of enriched WT mice (52). It will be interesting to investigate whether the extent of neurogenesis taking place in the DM-Tau-tg mice is similar to that in WT- mice exposed to EE. This will indicate whether the pathologic tau (double-mutant) alters/impairs the neurogenesis and plastic responses.

No differences in numbers of microglia were detected between EE-DM-Tau-tg and RE- DM-Tau-tg mice, possibly indicating that the increase in microglia reported in EE WT mice may be impaired in the EE-DM-Tau-tg mice because of some alteration/deficit in the tau function. This has been reported in enriched PS1ΔE9- and PS1M146L mice (9, 11), and there is a report on the lack of increase in microglial burden in enriched APPswe/PS1ΔE9 mice (29). Yet, although no change in microglial count was detected, a decrease in the microglial activation state was detected with both lectin staining and Iba-1 IHC. Morphology provides some indication about the activation state of microglia, although it cannot predict function--either proinflammatory, M1 type, or anti-inflammatory, M2 type (60). The slight, although significant, decrease in the activation state of microglia by EE treatment may suggest a reduction of noxious stimuli for microglia (61, 62). However, the parenchymal microglia did not acquire a typical M1 or M2 phenotype (data not shown), a finding that may not support a direct role of the microglial cells in reducing NFTs by the EE. On the basis of the reports of a positive correlation between NFT burden and microglial-burden in brains of both tauopathy patients (63, 64) and NFT Tg mice (65), a lower microglial activation that accompanied the lower NFT burden in our EE-DM-Tau-tg mice may not be surprising. Although activation of microglia has been reported to precede the formation of NFTs in naive NFT Tg mice (51), we speculate that NFTs may also affect the microglial activation, that is, once a decrease in tangles occurs, there may be a lower activation signal for microglia. In other words, the decrease in microglial activation might be an epiphenomenon reflecting the decrease in the neurodegenerative process in the EE-DM-Tau-tg mice than a causative effect.

We observed an increase in BDNF in neuronal and glial processes in the EE-DM-Tau-tg versus RE-DM-Tau-tg mice. This indicates that the pathogenic DM-Tau did not block this BDNF elevation reported in WT mice exposed to EE (66-68). This increase in BDNF after EE may shed some light on the mechanisms involved in the reduction in NFT burden in the enriched tauopathy mice because BDNF has been reported to reduce tau phosphorylation via the tyrosine receptor kinase B, involving the GSK-3β and PI-3 kinase, which participate in the signaling transduction cascade regulating tau phosphorylation (32). Other studies also showed that the regulation of GSK-3b/PI-3 is associated with BDNF (69-73). Moreover, Berardi et al showed that EE of mice that cannot express NGF, another neurotrophic factor that induces tau dephosphorylation (27, 28), reduced amyloid pathology but did not reduce tau phosphorylation, lending further support to the notion that expression of neuroprotective neurotrophic factors under EE may be involved in tau dephosphorylation. Some additional indirect support to the relevance of the increase in BDNF to the decrease in NFTs in the enriched mice may originate from our finding that RE DM-Tau-tg mice also showed a decrease in BDNF level relative to RE WT mice; EE seemed to rescue both deficits.

Our DM-Tau-tg mice also showed a trend of improvement in cognitive tasks after exposure to EE. This improvement in performance of the EE-DM-Tau-Tg, which approached the performance of the RE-WT mice, may be related to a reversal/prevention, at least partially, of the detrimental effects characteristic to the DM-Tau-tg mice. It is possible that not only anti-NFT processes but also other general EE-induced processes (e.g. neurogenesis and synaptic remodeling) are involved in the improvement in cognitive tasks in our NFT model. Further studies are needed to dissect the contribution of such processes, but our present results indicate that mice with NFTs are responsive to EE and that their cognitive status can be improved.

Environmental enrichment was beneficial when it was started at an early age, thereby raising the possibility that EE may be beneficial for the prevention of NFT pathology and cognitive deficits. While being exposed to stimulating activities from a young age (specifically of nonextreme regimes, yet long-term) seems feasible to the general population, it would be important to investigate whether exposure to stimulating activities at older ages (when the tauopathy disease is already evident) would also be beneficial. Further experiments are also needed to address how long is the beneficial effect preserved after cessation of EE. In addition, the effect of EE on biochemical properties of tau, such as the response of soluble oligomers (toxic forms 74), as well as on the detergent solubility of tau, are important issues to be investigated.

In conclusion, our results showed a decrease in NFT burden accompanied by an improvement in memory in an authentic tauopathy model after a tolerated paradigm of EE, pointing to a direct effect of EE on NFTs rather than a secondary anti-NFT effect of EE on amyloid pathology. This anti-NFT effect correlated with increased BDNF. The responsiveness of NFTs even under moderate/nonextreme stimulating activities may have promising implications to the general population.


The authors thank Fanny Baitscher from the Department of Neurology, Hadassah, and Evangelia Nousiopoulou from the Department of Neurology, AHEPA, for technical assistance.


  • This study was supported by Grant No. 5721 from the Chief Scientist Office of the Ministry of Health, Israel.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
View Abstract