Early Intervention with a Multi-Ingredient Dietary Supplement Improves Mood and Spatial Memory in a Triple Transgenic Mouse Model of Alzheimer’s Disease

Abstract

The increasing global burden of Alzheimer’s disease (AD) and failure of conventional treatments to stop neurodegeneration necessitates an alternative approach. Evidence of inflammation, mitochondrial dysfunction, and oxidative stress prior to the accumulation of amyloid-β in the prodromal stage of AD (mild cognitive impairment; MCI) suggests that early interventions which counteract these features, such as dietary supplements, may ameliorate the onset of MCI-like behavioral symptoms. We administered a polyphenol-containing multiple ingredient dietary supplement (MDS), or vehicle, to both sexes of triple transgenic (3xTg-AD) mice and wildtype mice for 2 months from 2–4 months of age. We hypothesized that the MDS would preserve spatial learning, which is known to be impaired in untreated 3xTg-AD mice by 4 months of age. Behavioral phenotyping of animals was done at 1-2 and 3-4 months of age using a comprehensive battery of tests. As previously reported in males, both sexes of 3xTg-AD mice exhibited increased anxiety-like behavior at 1-2 months of age, prior to deficits in learning and memory, which did not appear until 3-4 months of age. The MDS did not reduce this anxiety or prevent impairments in novel object recognition (both sexes) or on the water maze probe trial (females only). Strikingly, the MDS specifically prevented 3xTg-AD mice (both sexes) from developing impairments (exhibited by untreated 3xTg-AD controls) in working memory and spatial learning. The MDS also increased sucrose preference, an indicator of hedonic tone. These data show that the MDS can prevent some, but not all, psychopathology in an AD model.

Keywords

Alzheimer’s disease anhedonia anxiety dietary supplements learning memory mice mild cognitive impairment reversal learning transgenic working memory

INTRODUCTION

Alzheimer’s disease (AD) afflicts over 33 million people worldwide (http://www.who.int) at a cost of US$ 604 billion and is projected to reach 135 million cases by 2050. Conventional treatments, such as cholinergic drugs, fail to stop the progression of the disease, highlighting the need for new treatments [1]. In AD, the rapid decline in cognitive ability reflects damage to brain regions important in learning, memory and mood regulation, notably the hippocampus and prefrontal cortex [2 –4]. Thus, AD is also associated with high levels of depression (43.6%) and anxiety (25.4%) far exceeding those in age-matched controls [5].

Most research into AD pathogenesis has focused on counteracting the accumulation of misfolded proteins amyloid-β (Aβ) and hyperphosphorylated tau that contribute to plaques and tangles throughout the brains of AD patients. This work has led to the prominent “amyloid cascade hypothesis” of AD [6, 7]. Briefly, the cascade hypothesis proposes that excessive production or impaired clearance of Aβ leads to its subsequent accumulation, which triggers a neurotoxic cascade resulting in tau hyperphosphorylation and aggregation, synaptic atrophy, inflammation and oxidative damage, Ca^{2 +} imbalance and ultimately neuronal death.

The amyloid cascade hypothesis is well supported by evidence from in vitro and animal studies [7 –9]. However, it appears to have some serious limitations, most notably the fact that all anti-amyloidogenic drugs have thus far failed in human clinical trials [10, 11]. Furthermore, downstream effectors such as synapse loss [12] and reduced levels of brain-derived neurotrophic factor (BDNF) [13, 14], rather than plaque or tangle burden, are better predictors of cognitive symptoms in AD patients. Moreover, an emerging literature implicates exacerbation of age-related physiological changes in cognitive decline and AD. Specifically, the preclinical stage of AD, mild cognitive impairment (MCI) [15], has been linked to elevated levels of inflammation [16], insulin resistance [17, 18] and oxidative stress [19], and reduced levels of BDNF [13] and brain glucose metabolism [20]. Therefore, in contrast to most other studies on AD model animals [21, 22], we chose to focus on prevention rather than treatment.

Lifestyle based approaches to MCI and AD prevention, such as nutritional supplements, are emerging as alternatives to pharmaceutical compounds [1, 11]. For example, 6 months of supplementation with the omega-3 fatty acid docosahexaenoic acid (1.16 g/d DHA) in healthy adults was found to improve episodic memory, working memory, and attention [23]. Further, in a 6–8 year longitudinal study, the risk of AD was 60% lower in those who drank polyphenol-containing fruit or vegetable juices at least 3x per week [24].

Such findings suggest that a combination of polyphenols and unsaturated fatty acids may protect against MCI and consequently AD. We propose that a particularly promising approach is to use a broad-based multiple ingredient dietary supplement (MDS; Table 1) developed to target age-related alterations in inflammation, oxidative stress, mitochondrial dysfunction, insulin resistance and membrane integrity [25 –29], all of which are also implicated in AD disease progression as discussed above. This MDS has been shown to reduce age-related declines in spatial learning, brain volume, neuronal atrophy, neuronal death and DNA damage in aged mice [25, 26].

Table 1

Ingredients included in the multiple ingredient dietary supplement

Ingredient	Daily Dose (mg)	Ingredient	Daily Dose (mg)
Acetyl L-Carnitine	14.4	L-Glutathione	0.36
Acetylsalicylic Acid	2.5	Magnesium	0.72
Ascorbic Acid (Vit. C)	3.6	Melatonin	0.01
Bioflavinoids	4.32	N-Acetyl Cystein	7.2
Cholecalciferol (Vit. D)	0.0000625	Niacin (Vit. B3)	0.72
Chromium Picolinate	0.00144	Potassium	0.36
		Pyridoxine
Cobalamin (Vit. B12)	0.00072	Hydrochloride (Vit. B6)	0.72
Curcumin	1.8	Quercitin	0.9
D-a-tocopherol (Vit. E)	0.965	Rutin	0.72
Folic Acid	0.01	Selenium	0.00108
Garlic	0.0216	Thiamine (Vit. B1)	0.72
Ginger	7.2	Ubiquinone (CoQ10)	0.44
Gingko Biloba	1.44	Wild Fish Oil	21.6
Ginseng	8.64	a-Lipoic Acid	0.72
Green Tea	7.2	ß-Carotene	0.03

The current study administered the MDS to both sexes of 3xTg-AD mice [30] and wildtype (WT) mice and compared their behavior across a wide range of measures (Table 2) to vehicle-treated 3xTg-AD and WT controls. This battery of tests was designed to clarify which behavioral alterations appear in this mouse model at the earliest stages; most previous studies examined behavioral outcomes at 6 months of age or later (e.g., [31 –33]), reported conflicting sex differences (e.g., [34, 35]), used only a single behavioral measure starting at 2–4 months of age [34, 36], or evaluated only one sex (e.g., [37, 38]).

Table 2

Mouse behavioural test battery

Day(s)	Test	Measures
– 4 to 0	Animal Handling	n/a
1	visual placing response, righting reflex, grasping reflex, hind limb placing response, postural reflex, negative geotaxis	basic motor and visual reflexes
	Rotarod	sensorimotor coordination
2	Hanging Basket	muscle strength
3–8	Olfactory Acuity and Discrimination	olfactory acuity and discrimination
9	Beam Walking	balance/motor coordination
10 - 11	Open Field	anxiety (agoraphobia/photophobia)
12	Elevated Plus Maze	anxiety (acrophobia/photophobia)
13–16	Novel Object Recognition	novelty detection
17–21	INBEST Monitoring	spontaneous activity, food and water intake, sucrose preference
22–27	Spontaneous Alternation	visuospatial working memory and navigation strategy
28–36	Morris Water Maze	visual acuity, Spatial learning/memory

The importance of including both sexes is becoming increasingly appreciated, as male and female 3xTg-AD mice exhibit different trajectories of Aβ accumulation [39] and behavioral abnormalities [33, 35]. Behavioral alterations, progressing from anxiety to learning and memory impairments, emerge in males at 2–4 months of age [30 , 41]), while females develop similar symptoms at 4–6 months of age [33 , 43]. After initial presentation, it is unclear whether memory deteriorates more rapidly in males or females, as different behavioral tests have yielded conflicting results (e.g., [34, 35]). Behavioral deficits are followed by the emergence of amyloid and tau accumulation starting at 3–6 months of age, with more rapid accumulation in females [44 –47]. Plaques appear later, at 8–14 months of age in females and 16–18 months of age in males [38 , 48]. Still later, tangles emerge at 16–18 months of age in females [44, 48] and 21–26 months of age in males [46, 49].

Given the rate of disease progression described above, we anticipated that by 3-4 months of age 3xTg-AD mice would exhibit anxiety-like behavior (males only; [38]) and selective deficits on the Morris water maze (MWM; both sexes [36]). Moreover, we predicted especially pronounced deficits in the MWM for the high interference reversal trials, during which the mouse is required to neglect the original location of the submerged platform and learn a new location. High interference learning is markedly impaired in MCI patients [50] and has been shown previously to be dependent upon hippocampal neurogenesis in animal models [51, 52]. Moreover, adult neurogenesis and BDNF are among the earliest biomarkers depleted in rodent models of AD (by 2–4 months of age [53 –56]). Given that the MDS (and similar supplements) upregulate neurogenesis [57, 58] and BDNF [57 , 59–61], we hypothesized that the MDS would ameliorate deficits in the 3xTg-AD mice on water maze reversal trials.

MATERIALS AND METHODS

Animals

The subjects were 26 male and 20 female B6;129-Psen1^tm1Mpm Tg(APPSwe, tauP301L) 1Lfa/Mmjax (3xTg-AD) mice, plus 25 male and 26 female mice of the wildtype genetic background B6129SF2/J strain (WT). Fewer 3xTg-AD females and WT males were used due to a lack of availability from the breeding colony (see below). The 3xTg-AD mouse has been described in detail elsewhere [30]. It is a homozygous carrier of 3 human mutations for genes associated with early onset AD and frontotemporal dementia affecting amyloid and tau: APPswe, Psen1, and tauP301L. This model was chosen as one of the few to exhibit both amyloid and tau pathology concomitant with impaired memory and altered mood-related behaviors [62, 63]. Mice used in the study were obtained from an in-house breeding colony in the Psychology Department Animal Facility at McMaster University, which was established using breeders ordered from Jackson Laboratories (Bar Harbor, USA).

All animals, including breeders, were housed with a reversed 12:12 h light/dark cycle (lights off from 7:00–19:00) and stable temperature (∼22°C±0.5°C) and humidity (∼62%). Upon weaning, experimental mice were housed individually and provided with Harlan^TM Teklad 22/5 Rodent Diet chow and water ad libitum. All mice were caged with woodchip bedding and provided with nestlets as a source of environmental enrichment. The mice were then tattooed on the ears under isoflurane anesthesia for identification purposes and subsequently handled for 5 days (∼1 min each per day) prior to the start of baseline behavioral testing (at ∼4.5–5 weeks of age) using the behavioral battery (Table 2). Animals were also given ad libitum access to sucrose-water (1%) for 10 h during INBEST testing. After baseline behavioral testing, mice of each genotype were randomly assigned to the treatment (MDS) or control (vehicle) condition (see below), yielding 8 groups: MDS 3xTg-AD females (n = 10), vehicle 3xTg-AD females (n = 10), MDS 3xTg-AD males (n = 13), vehicle 3xTg-AD males (n = 13), MDS WT females (n = 13), vehicle WT females (n = 13), MDS WT males (n = 12), vehicle WT males (n = 13). All procedures were approved by the McMaster University Animal Research Ethics Board.

Multiple ingredient dietary supplement

The multi-ingredient dietary supplement (MDS) was first developed to determine whether targeting multiple cellular processes with nutritional compounds might attenuate the processes associated with the premature aging phenotype of transgenic growth hormone mice. Specifically, increased oxidative stress, inflammation, impaired glucose metabolism, membrane deterioration, and mitochondrial dysregulation are processes common to normal brain aging and exacerbated in neuropathologies [64 –69]. Ingredients were selected based on established efficacy in one or more of the above processes, leading to the combination of 30 ingredients (Table 1) that comprise the MDS. See Lemon et al. [28], for the processes targeted by each ingredient. The MDS has demonstrated significant protective effects on cognition, longevity and motor function in normal and TGM mice [25–29 , 70].

The combination of ingredients, preparation method and doses of the MDS used were identical to those described previously [27]. These parameters were originally defined according to recommendations for human consumption adapted to mice, accounting for differences in body size and metabolic rate. Briefly, the 30 MDS ingredients were mixed into aqueous solution, pipetted onto a small piece of bagel and left to dry. These bagel chips were then provided to mice between 17:00–19:00, i.e., leading up to the sleep cycle (lights on), because of the presence of melatonin in the MDS. Mice assigned to the vehicle control condition were provided with plain bagel chips only (no MDS). All bagel chips were consumed within several minutes. Bagel chips were fed to the mice daily from 2 months of age for 2 months, until the end of behavioral testing at 4 months of age.

Behavioral test battery

The battery of tests (described below; Table 2) was used to assess a wide variety of motor, sensory, mood-related, learning and memory functions in mice that are known (except for reflexes) to be altered in 3xTg-AD mice at different ages. Animals were tested in the battery (Table 2) at 1-2 months of age (baseline), randomly assigned to receive the MDS or vehicle daily for 2 months and subsequently retested at 3-4 months of age. The 1-2 and 3-4 month testing ages were therefore chosen because they represent distinct stages of early disease progression in this specific strain (see Introduction). This combination of tests represents an expansion from a similar battery used previously [38], and is, to our knowledge, the most comprehensive battery of tests that has been administered to 3xTg-AD mice. This broad range of measures was designed to capture the spectrum of cognitive, behavioral, and psychiatric symptoms commonly exhibited by patients with MCI or AD. While it is not possible to evaluate some symptoms (such as hallucinations [71]) in mice, we were able to test many of the behavioral characteristics that are affected in MCI or AD such as motor co-ordination, anxiety, working memory, recognition memory, visuospatial learning and memory, olfactory function, and anhedonia [15 , 71–73]. Subjects were tested individually in random order (generated using R software) on each day of the battery (tests administered in fixed order) to ensure that no animals were tested earlier or later in the day on average relative to others. Note that an abridged description of each test is included below. Please see the Supplementary Materials for additional details.

Motor and visual reflexes

On the first day of testing, several reflexes were assessed to determine if any mouse was affected by gross neurological deficits that would impair visual or motor function to a degree that would limit their ability to complete the other tests. Reflexes tested were: visual placing response, righting reflex, grasping reflex, hind-limb clasping reflex, postural reflex, and negative geotaxis. Testing procedures have been described elsewhere [74, 75]. Performance was scored only on a pass/fail basis because more detailed assessments were done in subsequent tests (e.g., beam walking and rotarod). Since no mice exhibited abnormal responses, this portion of the battery is not covered further in this report.

Motor coordination and muscle strength

Although motor coordination, strength and mobility are considered to be affected only later in the progression of AD (http://www.alz.org; [76], there is some evidence that fine/complex coordination might be impacted in MCI [72]. The rotarod [77] and hanging basket test [78] assess gross coordination and muscle strength, respectively. Since performance on these tests can be affected by body weight [33], we also recorded weight as a potential covariate.

The rotarod

The procedure and apparatus (MedAssociates, Inc., St. Albans, VT) used in the present experiment were described previously [38]. Mice were tested on 3 trials in a single session, and the average fall time was used as a measure of performance.

The hanging basket test

The mouse was placed on top of a wire mesh that made up the bottom surface of a transparent Plexiglas basket (29 cm×19 cm×16 cm). The basket was then turned upside down and rested on a padded table top. The time until falling (up to 10 min) was recorded by stopwatch, and then the mouse was returned to its home cage. Each mouse was tested on 2 trials.

Beam walking test

On this test the animals crossed a narrow beam from a moderately aversive location (brightly illuminated open space) to a less aversive location (enclosed shelter). The apparatus consisted of a translucent Plexiglas beam (70 cm×1 cm²) with a small rectangular (6 cm×8 cm) Plexiglas platform on one end and a large circular platform (diameter = 28.5 cm) at the other. The protocol was comprised of 3 shaping trials and a single testing trial. In each shaping trial the mouse was placed at a different starting location on the beam, initially adjacent to the large platform and on successive trials, starting progressively farther along the beam toward the small rectangular platform. After the 3rd shaping trial, the mouse was placed on the small rectangular platform and observed as it crossed the beam. Mice were scored for time and number of foot slips.

Olfactory acuity and discrimination

Olfactory function is becoming increasingly recognized as a potential diagnostic tool in helping to discriminate between healthy older adults, those with MCI, and those with AD [79, 80].

Acuity testing

The olfactory acuity testing method used has been described previously [38]. 5 trials were run over 5 days (1/d) using a different concentration (wt/vol, 1%, 0.1%, 0.01%, 0.001% or 0% /control) of peanut butter each day. In each trial, 60 uL of the peanut butter (in mineral oil) solution was pipetted onto 3 cm² square pieces of filter paper that were left in the cage with the mouse for 3 min. Mice were scored for sniffing time.

Discrimination testing

The olfactory discrimination testing was done the day after acuity testing, using the same equipment. 0.001% cinnamon and 0.001% paprika (President’s Choice, Loblaws Inc., Brampton, ON) dissolved in mineral oil were used as odorants, as they are considered to be appetitively neutral [81]. In a single session, mice were exposed to 4 consecutive trials lasting 2 min each using cinnamon as the odorant. Between trials, the testing cages were cleaned using 1.5% acetic acid in aqueous solution. Paprika was used as the odorant on the 5th trial. The difference in sniffing time between trial 4 (last cinnamon trial) and trial 5 (the paprika trial) was used as the measure of performance, where animals which exhibited an increase in sniffing time were considered able to discriminate between the two odorants.

Open field

Anxiety-like behavior was of great interest in the present study, as 83% of MCI patients who also suffer from anxiety disorders later convert to AD [73]. The open field is a common test of anxiety and exploratory behavior in rodents [82]. The apparatus used was an empty white polyethylene water maze pool (height = 60 cm, diameter = 112 cm). Each animal was tested in a single 20-min trial initiated by releasing the subject facing the wall of the apparatus. Trials were recorded using an overhead camera. For tracking purposes, the open field was divided into 3 equal area zones (3284 cm²): outside, intermediate and center, using the arena settings in Ethovision XT (Noldus, Toronto, ON). Live tracking data were subsequently processed to obtain measures including latency to approach, time spent in the center zone, total distance and average speed of ambulation, freezing time and frequency, and the incidence of defecation. Generally, more anxious mice are less willing to explore exposed areas, tend to freeze more often and leave more fecal boli than calmer mice [82, 83].

Elevated plus maze

The plus maze is another standard test of anxiety in rodents [82]. The apparatus consisted of a black Plexiglas plus shaped maze (4 arms each 56 cm×12.5 cm) elevated 60 cm off the floor, with 2 closed arms (15.5 cm high walls) and 2 open arms (no walls). Mice were placed in the center of the maze facing an open arm and left to explore the maze for a single 10-min trial before being returned to their home cages. Mice were scored for time spent in the open arms.

Integrated behavioral station (INBEST)

The INBEST is a Plexiglas monitoring apparatus (39 cm L×53 cm W×50 cm H) designed by Sakic et al. [84] that enables simultaneous automated tracking of multiple behaviors over extended periods in a minimally-stressful environment. Mice were placed in an INBEST box for a single 10-h trial before being returned to their home cages. INBEST measures included in the present study were sucrose preference, locomotor activity and food consumption. Locomotor activity was recorded as total ambulatory distance and average speed calculated from Ethovision XT tracking data. Sucrose preference is a measure of anhedonia [85] and was calculated by comparing consumption of a 1% sucrose solution (w/v) to total consumption of sucrose-water plus water alone. This concentration of sucrose was chosen based on previous literature [86]. Sucrose preference was of interest in this study because symptoms of depression, including anhedonia/apathy, reportedly increase the risk of conversion from MCI to AD by a factor of 1.9 [73].

Novel object recognition test

Recognition memory, rather than episodic memory [37], was measured in this study, as both are impaired in MCI patients [15]. The novel object test is a widely used means of evaluating object recognition memory in rodents [87], which takes advantage of the natural inclination of rodents to interact preferentially with novel objects over familiar ones. The apparatus consisted of a white high-density polyethylene box (area = 60 cm²×wall height = 40 cm). An assortment of glass, plastic, polished rock and painted metal statues were used as objects.

We employed a single trial, non-matching-to-sample procedure similar to that reported by others [87]. On the first 3 days of testing, mice were placed in the apparatus for a 10-min habituation trial (per day) with no objects present. On the 4th day, mice were tested in 2 trials in which objects were also present in opposite corners of the box. The mouse was left to explore and interact with the objects for 5 min. It was then returned to its home cage for 5 min while the apparatus and objects were cleaned and one of the objects was switched with a different one. The mouse was then returned to the box for 3 min. Mice were scored for the percentage of time spent interacting with the novel object out of the total time spent interacting with both objects.

Spontaneous alternation

This test takes advantage of rodents’ natural tendency to explore new environments and is generally considered a metric of spatial working memory [88], which is impaired in AD patients [15] and 3xTg-AD mice [34]. The apparatus and protocol used in the present study were described in detail previously [38]. Briefly, the apparatus was a plus-shaped maze with arms and sliding doors. Large objects in the testing room served as available visual cues for navigation. There were 2 stages per trial. Each stage was started by placing the mouse in a blocked section of a predetermined starting arm. During the first stage, 2 arms were blocked, and the mouse was restricted to the start arm and one additional arm. The mouse was released from the starting location, and when it entered the other arm it was locked inside the arm for 60 s. It was then returned to its home cage for 30 s while the maze was cleaned and an additional arm was made to be accessible.

In the second stage of trials on days 1–5, the mouse was returned to the same starting location as for stage 1. It was then released after 5 s and permitted to enter one of the two other open arms. The mouse’s choice (alternation or perseveration) was recorded. Two trials were administered each day. On stage 2 of the trials on day 6, instead of returning the mice to the same starting location, they were started on the opposite side of the maze, enabling evaluation of which navigation strategy had been used during days 1–5.

Morris water maze

Impairments in visuospatial learning and memory are well-known symptoms of AD [15]. In addition, learning under conditions of high interference, of which water maze reversal trials are an example, is markedly impaired in MCI patients [50]. The Morris water maze is a standard test of spatial learning and memory in rodents [89]. The apparatus and procedure used have been described elsewhere [38]. Briefly, the apparatus was a white pool (height = 60 cm, diameter = 112 cm), partially filled with tap water (24–26°C) and surrounded by a variety of visual cues. Testing was conducted over 8 days at baseline. On the first day of 9 days of follow-up testing, a single 2 min probe trial, to assess residual memory for the former platform location, was administered. Four trials per day were administered on all subsequent days, consisting of cued acquisition trials in which a transparent platform (diameter = 15 cm) was placed in the center of a quadrant and made visible with a black rubber cork.

Sixteen acquisition trials were administered over the next 4 days, in which the platform was submerged. To assess memory for the platform location, the following day consisted of 4 probe trials in which the platform was removed from the pool. The next day consisted of cued reversal trials (i.e., platform visible again) with the platform relocated to the opposite quadrant. On the final day, 4 reversal trials were administered in which the platform was submerged again but in the 2nd location.

All trials were filmed using an overhead video camera, and Ethovision XT (Noldus, Toronto, ON) tracking data were scored for swim path length, average swimming speed, and latency to locate the platform. For probe trials only, the time spent in the target quadrant was used as a measure of memory. For other trials, there were differences in swimming speed between groups, so swim path length was used instead of latency and as a measure of learning. For the cued trials, swim path length was considered to be a measure of visual acuity, whereas on acquisition and reversal trials when the platform was hidden, it was considered to be a measure of spatial learning and memory.

Statistical analyses and graphics

Olfactory acuity and water maze data were analyzed using repeated-measures genotype×sex×diet×trial/session (or concentration) mixed effects ANOVAs. As they measure different cognitive processes, the cued acquisition/reversal, probe, acquisition and reversal trials of the water maze were analyzed separately. Performance on these measures was predicted a priori to improve linearly across trials for olfactory acuity, water maze acquisition, and water maze reversal (e.g., [38]), therefore, the linear change across trials (i.e., the first order polynomial contrast) was also evaluated for these ANOVAs. In cases where the sphericity assumption was violated (Mauchly’s test), the corrected Greenhouse-Geisser p-values are reported instead of the uncorrected p-values. If significant interactions were found between two between-subjects factors, the simple main effects were similarly evaluated. If a significant interaction was detected, post-hoc t-tests were used, with Tukey’s HSD correction for multiple comparisons applied to p-values. Since we hypothesized that the MDS would selectively preserve learning on the hidden platform reversal trials of the water maze in 3xTg-AD mice, an a priori contrast [90] [(3xTg-AD×MDS) > (3xTg-AD×vehicle = WT×MDS = WT×vehicle)] was also evaluated for the data from these water maze trials only.

All other measures were evaluated with univariate 2×2×2 (genotype×sex×diet) factorial ANOVAs unless either normality (Shapiro-Wilk test) or homoscedasticity (Levene’s test) was violated, in which case an equivalent factorial permutation test (10,000,000 iterations) was used [91, 92]. Instead of comparing data from baseline to 3-4 month testing using repeated measures ANOVAs, difference scores from baseline were examined in separate factorial ANOVAs or permutation tests following the same logic described above. This was done to improve the power of the tests to detect effects of the MDS based on the experimental design.

T-tests were also used to assess whether the change in spontaneous alternation rate between days 1–5 and day 6 was significantly non-zero for each group at baseline or follow-up testing, to determine the navigation strategy that had likely been used by the mice in that group during the previous 5 days. Statistical tests were conducted using the car [93], afex [94], lsmeans [95] and lmPerm [91] packages in R [96]. An α-level of 0.05 was used to determine statistical significance. Figures were generated using the ggplot2 package in R [97], and tables were created using Microsoft Excel 2010 (Microsoft Canada Inc., Mississauga, ON). All subjects were included in all analyses, therefore, as specified in the Animals section above, group sample sizes for statistical comparisons were: MDS 3xTg-AD females (n = 10), vehicle 3xTg-AD females (n = 10), MDS 3xTg-AD males (n = 13), vehicle 3xTg-AD males (n = 13), MDS WT females (n = 13), vehicle WT females (n = 13), MDS WT males (n = 12), vehicle WT males (n = 13). With respect to the results of the ANOVAs and factorial permutation tests, this is equivalent to: MDS (n = 48) versus vehicle (n = 49) for main effects of the MDS; males (n = 51) versus females (n = 46) for main effects of sex; and 3xTg-AD (n = 46) versus WT (n = 51) for main effects of genotype.

Partial eta-squared ( $η_{p}^{2}$ ) was used as a measure of effect size for all effects and interactions reported for ANOVAs and permutation tests that modelled non-repeated-measures data [98]. Generalized eta-squared ( $η_{g}^{2}$ ) was used as an effect size measure for repeated- measures data [99]. r_contrast-squared ( $r_{c}^{2}$ [100, 101]) was used as an effect size measure for the post-hoc t-tests, first order (linear) polynomial contrasts (olfactory acuity and water maze data), as well as the a priori contrast [(3xTg-AD×MDS)< (3xTg-AD×vehicle = WT×MDS = WT×vehicle)] used in analyzing the water maze reversal trial data. For reference, Cohen’s benchmarks for small (0.01), medium (0.06), and large (0.14) effects, are recommended for these measures [98 , 102]. Cohen’s d [102] was also used as an effect size measure for the post-hoc one-sample t-test comparing the performance of 3xTg-AD females to chance performance (30 s in the target quadrant) on the first probe trial of the Morris water maze (for reference: small effect d = 0.2, medium effect = 0.5, large effect d = 0.8).

RESULTS

Motor tests

The rotarod, beam walking and hanging basket tests were used to evaluate motor co-ordination and muscle strength. Data from these tests (and many of the others in the battery) were non-normally distributed or heteroscedastic, so the equivalent factorial permutation tests were used for analyses instead of ANOVA in such cases (see Methods). A genotype×sex×diet factorial permutation test on the rotarod fall time (Supplementary Figure 1a) at baseline did not uncover any effects or interactions when body weight was included as a co-variate in the model (Supplementary Table 1). At 3-4 months of age (Supplementary Table 1), after 2 months of MDS supplementation, there were still no differences between groups on the rotarod using raw scores (Supplementary Figure 1b) or difference scores (Supplementary Figure 1c) from baseline ([3-4 month score]-[baseline score]). Equivalent permutation tests for beam walking crossing time and the number of foot slips exhibited a similar pattern (Supplementary Table 2), with no significant effects or interactions for genotype, sex, or diet at baseline or at 3-4 months of age (Supplementary Figure 2). Thus, motor co-ordination does not differ between genotypes or sexes and is not affected by 2 months of MDS supplementation from 2–4 months of age. At baseline, but not 3-4 months of age, female mice did, however, take longer to fall in the hanging basket test (Supplementary Figure 3; Supplementary Table 3) than males (main effect of sex, permutation p = 0.002, $η_{p}^{2}$ = 0.104), even when body weight was included as a covariate (main effect of sex, permutation p = 0.002).

Table 3

Effects of the MDS on behavioral alterations in 3xTg-AD vs. WT mice up to 4 months of age

Behavioral Domain	Test	Alterations in 3xTg-AD mice	Effect of MDS
	rotarod	none	null
motor (co-ordination, strength)	hanging basket	none	null
	beam walking	none	null
sensory (olfaction and vision)	olfactory acuity	decreased in females at 1-2 months of age	decreased in females of both genotypes
	olfactory discrimination	none	null
	water maze cued trials	none	null
anxiety-like behavior	open field	increased in both sexes at 1-2 months of age	null
(photophobia/agoraphobia)	elevated plus maze	increased in both sexes at 1-2 months and	null
		3-4 months of age
depressive-like behavior (anhedonia)	sucrose preference	none	increased in both genotypes and sexes
recognition memory	novel object	impaired in both sexes at 3-4 months of age	null
working memory	spontaneous alternation	impaired in both sexes at 3-4 months of age	prevented deficits
long term memory	water maze probe trials	impaired in females only at 3-4 months of age	null
spatial learning and memory	water maze acquisition trials (low interference)	none	null
	water maze reversal trials (high interference)	impaired in both sexes at 3-4 months of age	prevented deficits

Fig.1

Box plots of the time (s) spent in the open arms of the elevated plus maze at 1-2 months of age (a), 3-4 months of age (b), and the difference scores (c) for each group between 1-2 and 3-4 months of age. 3xTg-AD mice spent less time than WT mice in the open arms at both 1-2 and 3-4 months of age (2×2×2 ANOVA, ^***p < 0.001). (AFD = MDS 3xTg-AD females [n = 10]; AFV = vehicle 3xTg-AD females [n = 10]; AMD = MDS 3xTg-AD males [n = 13]; AMV = vehicle 3xTg-AD males [n = 13]; WFD = MDS WT females [n = 13]; WFV = vehicle WT females [n = 13]; WMD = MDS WT males [n = 12]; WMV = vehicle WT males [n = 13]).

Fig.2

Preference for 1% sucrose water over unsweetened water ([weight 1% sucrose water consumed] / [total liquid consumed]; mean±SE) during INBEST testing at 1-2 months of age (a), 3-4 months of age (b), and the difference scores (c) between 1-2 and 3-4 months of age. At baseline (a), but not 3-4 months of age (b), sucrose preference was higher for males than females (2×2×2 ANOVA, main effect of sex, ^*p = 0.028). c) Examination of the difference scores determined that there was a significant increase with MDS supplementation (2×2×2 ANOVA, main effect of diet, ^*p = 0.013). (AFD = MDS 3xTg-AD females [n = 10]; AFV = vehicle 3xTg-AD females [n = 10]; AMD = MDS 3xTg-AD males [n = 13]; AMV = vehicle 3xTg-AD males [n = 13]; WFD = MDS WT females [n = 13]; WFV = vehicle WT females [n = 13]; WMD = MDS WT males [n = 12]; WMV = vehicle WT males [n = 13]).

Fig.3

Preference for the novel object ([time spent interacting with the novel object] / [total time spent interacting with both objects]; mean±SE) during novel object testing at 1-2 months of age (a), 3-4 months of age (b), and the difference scores (c) for each group between 1-2 and 3-4 months of age. A) At baseline, there were no differences between groups. b) By 3-4 months of age WT mice spent proportionally more time interacting with the novel object compared to 3xTg-AD mice (2×2×2 ANOVA, main effect of genotype, ^***p < 0.001). c) This represented a greater increase from baseline testing for WT mice compared to 3xTg-AD mice (2×2×2 ANOVA, main effect of genotype, ^***p < 0.001). (AFD = MDS 3xTg-AD females [n = 10]; AFV = vehicle 3xTg-AD females [n = 10]; AMD = MDS 3xTg-AD males [n = 13]; AMV = vehicle 3xTg-AD males [n = 13]; WFD = MDS WT females [n = 13]; WFV = vehicle WT females [n = 13]; WMD = MDS WT males [n = 12]; WMV = vehicle WT males [n = 13]).

Olfactory function

Olfactory acuity was evaluated with increasing concentrations of peanut butter dissolved in mineral oil. Total sniffing time was used as a measure of acuity. A genotype×sex×diet×concentration mixed effects ANOVA did not reveal any significant effects or interactions between groups at baseline (Supplementary Table 4). Since we anticipated that sniffing time would increase with concentration, we also examined post hoc linear contrasts across concentrations for each sex×genotype group (repeated measures one-way ANOVA for each group). Tukey HSD corrected comparisons indicated that all groups showed a linear increase in sniffing time except for the 3xTg-AD females (t_(1,374) = 1.567, p = 0.118, $r_{c}^{2}$ = 0.007; 3xTg-AD males, t_(1,374) = 6.955, p < 0.001, $r_{c}^{2}$ = 0.115; WT females, t_(1,374) = 3.201, p = 0.002, $r_{c}^{2}$ = 0.027; WT males, t_(1,374) = 4.537, p < 0.001, $r_{c}^{2}$ = 0.052; Supplementary Figure 4a-c), suggesting that olfactory function may be impaired in 3xTg-AD females from 1-2 months of age (Supplementary Table 4).

Fig.4

Alternation rate (mean±SE) during days 1–5 of spontaneous alternation testing at 1-2 months of age (a), 3-4 months of age (b), and the difference scores (c) for each group between 1-2 and 3-4 months of age. A) There were no differences between groups at baseline. b) At 3-4 months of age WT females alternated more frequently than WT males (2×2×2 ANOVA, genotype×sex interaction, ^*p = 0.012). c) Comparison of the difference scores between baseline and follow-up testing (2×2×2 ANOVA) revealed a relative increase in alternation rates in WT females (genotype×sex interaction, ^*p = 0.025) compared to WT males but also that alternation rates were decreased in vehicle-supplemented 3xTg-AD mice but not MDS-supplemented 3xTg-AD mice (genotype x diet interaction, ^#p = 0.023). (AFD = MDS 3xTg-AD females [n = 10]; AFV = vehicle 3xTg-AD females [n = 10]; AMD = MDS 3xTg-AD males [n = 13]; AMV = vehicle 3xTg-AD males [n = 13]; WFD = MDS WT females [n = 13]; WFV = vehicle WT females [n = 13]; WMD = MDS WT males [n = 12]; WMV = vehicle WT males [n = 13]).

After exposure to the MDS for 2 months and at 3-4 months of age, there was no longer a genotype × sex × concentration difference (Supplementary Table 4). Instead, based on a genotype × sex× diet × concentration mixed effects ANOVA, there was a main effect of sex (F_(1,89) = 4.051, p = 0.047, $η_{g}^{2}$ = 0.021), a sex x diet interaction (F_(1,89) = 5.740, p = 0.019, $η_{g}^{2}$ = 0.029), a sex × concentration interaction (F_(4,356) = 2.605, p = 0.047, $η_{g}^{2}$ = 0.015), and a sex x diet x concentration interaction (F_(4,356) = 3.159, p = 0.022, $η_{g}^{2}$ = 0.019; Supplementary Figure 4d-f). Follow-up Tukey HSD corrected comparisons for sex×diet groups (averaging across genotype: MDS supplemented females, MDS supplemented males, vehicle-control females and vehicle-control males) were also examined for each concentration. These revealed lower sniffing times in MDS-supplemented females than MDS-supplemented males for the 2nd lowest (0.01% PB; t_(1,308) = – 2.899, p = 0.021, $r_{c}^{2}$ = 0.027) and highest (1% PB, t_(1,308) = – 2.875, p = 0.022, $r_{c}^{2}$ = 0.026) peanut butter concentrations. Sniffing time for MDS supplemented females was also lower than the vehicle control females at the highest concentration (1% PB, t_(1,308) = – 3.546, p = 0.003, $r_{c}^{2}$ = 0.039). Olfactory discrimination was assessed by comparing the difference in time spent sniffing the odorant between the last exposure of 0.001% cinnamon (trial 4/5) and 1st exposure of 0.001% paprika (trial 5/5), where animals that exhibited a dishabituation effect or an increase in sniffing time were considered able to discriminate between the two odorants. At baseline, the genotype×sex×diet permutation test did not yield any significant effects or interactions, and there were also no effects or interactions in the equivalent analysis of the data from follow-up testing (Supplementary Figure 5, Supplementary Table 5). There were therefore no effects of 2 months of MDS consumption on olfactory discrimination ability or any differences between males and females of either 3xTg-AD or WT mouse strains at 3-4 months of age (Supplementary Table 5).

Fig.5

Swim path length (cm, mean±SE) on the reversal trials of the water maze (2nd hidden platform location) at 1-2 (a-c) and 3-4 months of age (d-f). There were no differences between groups at 1-2 months of age, and all groups improved linearly across trials (a-c; 2×2×2×4 ANOVA, linear contrast for trial, p < 0.001). f) At 3-4 months of age, only vehicle-supplemented 3xTg-AD mice failed to improve across trials (2×2×2×4 ANOVA, interaction between a linear contrast for trial and the hypothesis contrast [vehicle-supplemented 3xTg-AD>MDS-supplemented 3xTg-AD = MDS-supplemented WT = vehicle-supplemented], ^*p = 0.011).

Anxiety-like behavior

A genotype×sex×diet factorial permutation test determined that at baseline, 3xTg-AD mice spent less time in the open arms of the plus maze than wild type mice, for both males and females (main effect of genotype, p < 0.001, $η_{p}^{2}$ = 0.293; Fig. 1a, Supplementary Table 6). By 3-4 months of age, 3xTg-AD mice still spent less time in the open arms than WT mice (ANOVA; main effect of genotype F_(1,86) = 67.585, p < 0.001, $η_{p}^{2}$ = 0.440; Fig. 1b; Supplementary Table 6). No additional effects or interactions were revealed using the difference scores from baseline (Fig. 1c).

Several potential anxiety-related behaviors were measured via video tracking in the open field test, specifically: total time spent in the center zone, latency to approach the center, total freezing time, and the number of fecal boli secreted. In contrast to the plus maze results, 3xTg-AD females (not males) spent more time in the center of the open field than WT females at 1-2 months of age (permutation test; genotype×sex interaction, p = 0.013, $η_{p}^{2}$ = 0.067; Supplementary Figure 6a, Supplementary Table 7). Follow-up tests for the simple main effect of genotype for males and females separately showed that only 3xTg-AD females exhibited this behavior (p < 0.001, $η_{p}^{2}$ = 0.214, for effect of genotype in females; p = 0.823, $η_{p}^{2}$ < 0.001, for males).

The permutation test for latency to approach the center at baseline also yielded a significant genotype×sex interaction (p = 0.039, $η_{p}^{2}$ = 0.048; Supplementary Figure 7a, Supplementary Table 7). Follow-up tests for the simple main effect of genotype in each sex determined that only 3xTg-AD males took longer to approach the center (for females, p = 0.367, $η_{p}^{2}$ = 0.029; for males, p = 0.013, $η_{p}^{2}$ = 0.090; Supplementary Figure 8a). In addition, both sexes of 3xTg-AD mice spent more time frozen than WT mice (main effect of genotype, permutation p = 0.004, $η_{p}^{2}$ = 0.135; Supplementary Table 7). The ANOVA for the number of fecal boli left in the open field (Supplementary Figure 9a) yielded significant main effects of genotype (F_(1,89) = 5.437, p = 0.026, $η_{p}^{2}$ = 0.055) and sex (F_(1,89) = 4.644, p = 0.041, $η_{p}^{2}$ = 0.048), where more fecal boli were secreted respectively by 3xTg-AD mice than WT mice and by males more than females. By 3-4 months of age, there was no longer a genotype x sex interaction for time spent in the center of the open field, only a main effect of sex (females > males; permutation p = 0.015, $η_{p}^{2}$ = 0.064; Supplementary Figure 6b, Supplementary Table 7). There were no other significant effects or interactions detected, and a factorial permutation test on the center time difference scores also did not yield any significant effects or interactions (Supplementary Figure 6c, Supplementary Table 7).

For the open field center approach latency at follow-up, males took longer to approach the center than females (2×2×2 factorial permutation test; main effect of sex, p = 0.022, $η_{p}^{2}$ = 0.054; Supplementary Figure 7b, Supplementary Table 7), and there were no other significant main effects or interactions using the raw 3-4 month testing scores or difference scores (Supplementary Figure 7c). There were also no main effects or interactions from the equivalent permutation tests for freezing times at 3-4 months of age (or the change from baseline; Supplementary Figure 8, Supplementary Table 7). There was a main effect of sex according to the ANOVA for the number of fecal boli left in the open field at 3-4 months of age (males > females, F_(1,89) = 11.301, p = 0.001, $η_{p}^{2}$ = 0.113; Supplementary Figure 9b), and there were no main effects or interactions detected using the difference scores (Supplementary Figure 9c, Supplementary Table 7). Taken together, although 3xTg-AD mice exhibit anxiety-like behavior on both the open field and elevated plus maze at 1-2 months of age, the difference is only observed on the plus maze by 3-4 months of age. Lastly, the MDS does not seem to affect anxiety-like behavior that is already present prior to the start of treatment (at 1-2 months of age), given that it did not alter performance on the plus maze (Fig. 1), or any of the open field measures (Supplementary Figures 6–9).

Anhedonia, activity, and appetitive behavior

The INBEST apparatus enabled us to monitor habitual motor and appetitive behavior while also affording an opportunity to evaluate sucrose preference over a 10-h session. Sucrose preference was found to be higher in males than females at baseline (ANOVA; main effect of sex, F_(1,89) = 5.556, p = 0.028, $η_{p}^{2}$ = 0.058; Fig. 2a) but not by 3-4 months of age (Fig. 2b, Supplementary Table 8). The effect of MDS supplementation, but not genotype, was significant when the difference scores were examined in an ANOVA (main effect of diet, F_(1,89) = 6.370, p = 0.013, $η_{p}^{2}$ = 0.068), indicating that MDS supplementation increases sucrose preference (Fig. 2c, Supplementary Table 8). The effect of the MDS was attenuated slightly but remained significant (F_(1,39) = 5.378, p = 0.026, $η_{p}^{2}$ = 0.073) when the difference scores for food eaten during INBEST testing were included in the analysis as a covariate. This suggests that there is a genuine increase in hedonic motivation with MDS supplementation that is independent of appetite.

Both the total ambulatory distance and average movement speed were used as measures of motor activity. At baseline, genotype×sex×diet factorial permutation tests did not uncover any differences in ambulatory distance (Supplementary Figure 10a, Supplementary Table 8), although 3xTg-AD mice moved faster than WT mice (main effect of genotype, p = 0.004, $η_{p}^{2}$ = 0.088; Supplementary Figure 11a). After supplementation with the MDS at follow-up testing, there were no differences between groups in terms of either ambulatory distance or speed (factorial permutation tests; Supplementary Figures 10b and 11b), indicating that the baseline genotype difference in speed did not persist with age. Mirroring this change between baseline and follow-up testing was the finding of more negative difference scores in speed for 3xTg-AD mice than for WT mice (factorial permutation test; main effect of genotype, p = 0.011, $η_{p}^{2}$ = 0.069; Supplementary Figure 11c, Supplementary Table 8).

Object recognition memory

Recognition memory was assessed based on the preference of mice to explore a novel object over a familiar one. Examination of the fraction of time spent interacting with the new object out of the total interaction time for both objects in the arena did not reveal any effects or interactions between genotypes, sexes, or pre-treatment differences between supplementation groups (MDS versus vehicle) at baseline (2×2×2 ANOVA; Fig. 3a, Supplementary Table 9). However, by 3-4 months of age 3xTg-AD mice spent less time interacting with the novel object than WT mice (main effect of genotype, F_(1,89) = 18.107, p < 0.001, $η_{p}^{2}$ = 0.169; Fig. 3b, Supplementary Table 9). This main effect was also the only significant one when the difference scores were analyzed (F_(1,89) = 54.528, p < 0.001, $η_{p}^{2}$ = 0.336; Fig. 3c, Supplementary Table 9). Thus, compared to WT mice, both sexes of 3xTg-AD mice exhibit an attenuation of the developmental improvement in object recognition memory between 1-2 and 3-4 months of age that is not affected by the MDS.

Working memory

Working memory was assessed here using the spontaneous alternation test. For the day 1–5 alternation rate at baseline, the 2×2×2 ANOVA did not reveal any main effects or interactions (Fig. 4a, Supplementary Table 10). At 3-4 months of age, analysis of the raw scores revealed only a significant interaction between genotype and sex (F_(1,89) = 6.555, p = 0.012, $η_{p}^{2}$ = 0.069; Fig. 4b, Supplementary Table 10). When the difference scores from baseline to 3-4 month testing were compared, this interaction was still present (F_(1,89) = 5.231, p = 0.025, $η_{p}^{2}$ = 0.056), but there was also a genotype x diet interaction (F_(1,89) = 5.359, p = 0.023, $η_{p}^{2}$ = 0.057; Fig. 4c).

Follow-up ANOVAs for the simple main effects of these interactions determined that vehicle-control (F_(1,46) = 5.64, p = 0.022, $η_{p}^{2}$ = 0.112), but not MDS-supplemented (F_(1,47) = 0.314, p = 0.578, $η_{p}^{2}$ = 0.007), 3xTg-AD mice exhibited a decrease in spontaneous alternation rate from baseline relative to their WT counterparts (Fig. 4c). There was also a relative increase in alternation rates for WT females compared to 3xTg-AD females (F_(1,44) = 7.039, p = 0.011, $η_{p}^{2}$ = 0.138), but not for male WT mice compared to 3xTg-AD mice (F_(1,49) = 0.226, p = 0.636, $η_{p}^{2}$ = 0.005). Thus, MDS supplementation maintains spontaneous alternation rates in 3xTg-AD mice, similar to the level of WT mice, at 3-4 months of age. Furthermore, WT female mice show an increase in alternation rate between 1-2 and 3-4 months of age (Supplementary Table 10). In terms of navigation strategy, separate t-tests for the difference scores between the day 1–5 average alternation rate and day 6 alternation rate of each group at both 1-2 and 3-4 months of age determined that no group exhibited a change in alternation rate at day 6 from their day 1–5 average (data not shown).

Visuospatial learning and memory

Visual acuity and visuospatial learning and memory under conditions of low and high interference were evaluated using different stages of the water maze. On the cued acquisition trials, mixed effects ANOVAs (genotype×sex×diet×trial) determined that there were differences between 3xTg-AD mice and WT mice in terms of swimming speed at baseline (main effect of genotype, F_(1,89) = 18.12, p < 0.001, $η_{g}^{2}$ = 0.081; Supplementary Figure 12a-c) and at 3-4 months of age (main effect of genotype, F_(1,89) = 33.49, p < 0.001, $η_{g}^{2}$ = 0.087; Supplementary Figure 12d-f, Supplementary Table 11). Therefore, swim path length, rather than the latency to locate the platform, was used as a measure of performance on trials in which there was either a visible or submerged platform. On probe trials, during which the platform was removed from the pool, the time spent in the target quadrant was used as an indicator of memory for the location of the first platform.

Analysis of the swim path length data on the cued acquisition and cued reversal trials (i.e., platform was visible) using mixed effects ANOVAs did not indicate a disparity in visual acuity between any groups at either testing age (Supplementary Figures 13 and 14; Supplementary Table 11). On the acquisition trials, during which the platform was submerged in the first testing location (low interference), the 2×2×2×16 (genotype×sex×diet×trial) mixed effects ANOVA did not detect any effects or interactions at baseline or at 3-4 months of age, aside from the expected improvement across trials for all groups (main effect of trial; F_(15,1335) = 7.154, p < 0.001, $η_{g}^{2}$ = 0.063 at baseline; F_(10,855) = 3.399, p < 0.001, $η_{g}^{2}$ = 0.032, at 3-4 months of age; Supplementary Figure 15, Supplementary Table 12). These results indicate that all groups were able to learn the

first platform location equally well.

Examination of the time spent in the target quadrant in probe trials using a mixed effects ANOVA detected only a significant main effect of sex (F_(1,89) = 6.5, p = 0.013, $η_{g}^{2}$ = 0.042; males higher than females) and trial (F_(3,267) = 11.427, p < 0.001, $η_{g}^{2}$ = 0.049) at 1-2 months of age (Supplementary Figure 16a-c, Supplementary Table 13). Addition of a linear contrast to the ANOVA for the trial factor confirmed that, as expected, mice spent progressively less time in the target quadrant across trials when the platform was not in the pool (i.e., an extinction effect, t_(1,288) = – 4.940, p < 0.001, $r_{c}^{2}$ = 0.078). At 3-4 months of age, there was a genotype x sex x trial interaction (F_(3,267) = 3.530, p = 0.020, $η_{g}^{2}$ = 0.015; Supplementary Figure 16b, Supplementary Table 13). Follow-up Tukey HSD-corrected comparisons of each genotype×sex group on each trial demonstrated that 3xTg-AD females spent less time in the target quadrant of the maze than WT females on trial 1 only (t_(1,207) = – 3.506, p = 0.003, $r_{c}^{2}$ = 0.056). Although this memory impairment in 3xTg-AD females was not ameliorated by the MDS (the genotype x sex x diet interaction in the ANOVA was not significant, Supplementary Table 13), a further t-test demonstrated that the time they spent in the target quadrant was significantly greater than chance (t₍₁₉₎ = 2.832, p = 0.011, d = 0.633; Supplementary Figure 16b), suggesting that they could remember the trained platform location. On the probe trial used prior to the start of the water maze cued trials at 3-4 months of age (i.e., to test for residual memory for the former platform location after testing at 1-2 months of age), one tailed t-tests for each group confirmed that no group spent more than 30 s, or 25%, of the trial in the previously rewarded quadrant (all p-values >0.85, data not shown). This indicated that there was no residual memory of the platform

location from the previous round of testing.

On the reversal trials, in which the platform was submerged in the second location, at 1-2 months of age, the genotype×sex×diet×trial ANOVA on the swim path length data confirmed only the expected linear improvement of all groups across trials (Fig. 5a-c; main effect of trial, F_(3,255) = 4.009, p = 0.011, $η_{g}^{2}$ = 0.063; linear contrast for trial, t_(1,265) = – 2.967, p < 0.001, $r_{c}^{2}$ = 0.023; Supplementary Table 12). For the data from the follow-up testing at 3-4 months of age, we examined the a priori hypothesis that only vehicle-control 3xTg-AD mice would fail to exhibit an improvement across reversal trials using: 1) a linear contrast for the trial factor and, 2) a planned contrast of 1, – 1/3, – 1/3, and – 1/3 for the genotype×diet groups [(vehicle-control 3xTg-AD)> (MDS-supplemented 3xTg-AD = MDS-supplemented WT = vehicle-control)] within a 2×4×4 (sex×[genotype×diet]×trial) mixed effects ANOVA. As predicted, this analysis revealed only a significant interaction for the two contrasts (F_(1,368) = 6.468, p = 0.011, $r_{c}^{2}$ = 0.018; Fig. 5f, Supplementary Table 12), indicating that only vehicle-control 3xTg-AD mice had difficulty learning the second platform location. A summary of this and the other effects of the MDS reported above are presented in Table 3.

DISCUSSION

While we have previously observed benefits of the MDS on the age-related decline in spatial learning and memory among wild-type mice following an extended period of supplementation (∼1 year [25]), the present study demonstrates that similar benefits can be obtained in an AD mouse model with only 2 months of supplementation. That strong benefits can be observed after such a short intervention period is remarkable considering that most similar studies (e.g., [21 , 103]) report positive results after longer treatment periods (at least 3–6 months). Specifically, we show here that the MDS prevents the emergence of deficits in working memory and reversal learning that occur in 3xTg-AD by 4 months of age. The reversal learning deficit is expected as part of a human-like MCI phenotype (see Methods for a description), whereas the working memory impairment seems to emerge earlier in 3xTg-AD mice [34] than in humans [15]. Thus, the MDS may impart a degree of resilience against the deterioration of learning and memory in AD.

On the spontaneous alternation test, we also did not observe a decrease in performance when the starting viewpoint was shifted (i.e., alternation rate on days 1–5 versus day 6). This suggests the use of an allocentric rather than egocentric navigation strategy [104]. The particular strategy used was of interest, since (in humans) allocentric navigation preferentially engages the hippocampus, while egocentric navigation depends more on parietal association and striatal areas. Furthermore, there is a shift toward egocentric navigation with age [104] that may reflect changes in hippocampal volume that also occur with age and in AD [105]. Thus, the decreased alternation rate among 3-4-month-old control 3xTg-AD mice is not due to a change in navigation strategy (i.e., from allocentric to egocentric).

Evidence from the literature suggests that performance on the reversal trials of the water maze depends upon adult hippocampal neurogenesis [106 –109] (although see [110]). More generally, adult neurogenesis seems to be important for hippocampal-dependent learning under conditions of high interference [51 , 111]. Similarly, in AD patents, deficits on a high-interference, picture-matching memory test are also correlated with Aβ levels [112]. Promoting neurogenesis may therefore impart resilience against Aβ toxicity. In addition, the benefit of the MDS on reversal learning shown here may be due to preserved levels of neurogenesis in MDS-supplemented 3xTg-AD mice, which merits further investigation.

Our wide-ranging battery of tests across different ages has also yielded a more detailed and comprehensive picture of the early behavioral trajectory of the 3xTg-AD mouse than has been available thus far. To our knowledge, this is the first study aimed at preventing the development of symptoms in an AD mouse model using a complex supplement that was administered starting at 2 months of age. The fact that not all our results were positive suggests that the use of supplements alone may not be sufficient to prevent AD onset. Similarly, at least two other recent studies used complex supplements (at least 5 ingredients) in AD model mice. In one [21], a combination of 11 ingredients, 10 of which are also in the MDS used here, was administered to Tg2576 mice for 6 months starting at 6 months of age (behavioral deficits begin at 3 months of age in this strain). Benefits of varying magnitude were observed for all 3 measures studied: the novel object test and water maze, and levels of Aβ oligomers in whole brain homogenates. However, without the use of a more comprehensive behavioral battery of tests, it is unknown if the treatment could impact other aspects of an AD-like phenotype such as sensory/motor function or emotionality (see below for more information). Another study [22] using 3xTg-AD mice reported sex-dependent benefits of 4 months of treatment starting at 7–10 months of age on mitochondrial function (cytochrome C oxidase activity; males only) and a delayed-match-to-position, short-term memory task (only a 30 s delay; males only), using a 29-component supplement containing 5 similar ingredients to the MDS (turmeric/curcumin, green tea, ginger, fish oil, and vitamin D). However, the treatment did not improve memory deficits exhibited by the 3xTg-AD mice over longer periods (30 min or 24 h). Additionally, mitochondrial function was reduced in supplemented females, highlighting the importance of studying both sexes during pre-clinical experiments.

Our results suggest that, prior to 4 months of age, the 3xTg-AD mouse models the human behavioral phenotype of an MCI-like syndrome reasonably well, although not perfectly. For instance, the 3xTg-AD mouse did exhibit anxiety-like behavior (both sexes) and deficits in olfactory acuity (females only), but not motor co-ordination impairments, prior to 4 months of age. These changes in olfaction and anxiety level were the earliest behavioral symptoms to appear in these animals, developing prior to the spatial learning, working memory, or object recognition memory impairments at 3-4 months of age. Although symptoms of depression and anxiety are both common in patients with dementia [5], our results suggest the possibility that, among behavioral symptoms, anxiety or olfactory deficits may appear earlier than anhedonia or memory impairments in some older adults who later develop MCI. These findings are in agreement with those of Marchese et al. [38] who reported similar findings in male 3xTg-AD mice only. In that study, compared to WT males, 3xTg-AD males exhibited elevated anxiety-like behavior on the step-down test at 1.5 months of age and increased olfactory acuity starting at 6 months of age. Here, we confirm that the anxiety-like behavior is also present in 3xTg-AD females at 1-2 months of age, and we show that olfactory alterations (increased sensitivity in males, decreased sensitivity in females) are present in 3xTg-AD mice at 1-2, but not 3-4, months of age. Our finding of a beneficial effect of the MDS on sucrose preference also suggests that supplementation with the MDS may help protect older adults at risk for MCI from developing anhedonia or other depressive symptoms.

Regarding the observed anxiety, there is evidence of increased HPA axis activation in 3-4-month-old 3xTg-AD mice due to upregulation of glucocorticoid and mineralocorticoid receptors, but not corticosteroid levels, in CA3 and the DG of the hippocampus [113]. If this upregulation is also present at 1-2 months of age in this strain, then it might explain the increased anxiety in transgenic mice evident in our plus maze and open field data and would merit further research. Additionally, the stress response (increased corticosterone levels) of 3xTg-AD females, but not males, following water maze testing is increased over WT females between 6–15 months of age [35], emphasizing the relevance of comparing performance on tests of spatial memory to anxiety measures in both males and females when interpreting treatment effects.

The MDS did not reverse this pre-existing anxiety or prevent the appearance of deficits in novel object recognition (both males and females) or long-term recall on the water maze probe trials (females only) in 3xTg-AD mice by 3-4 months of age. Previous work with the MDS in 12-month-old C57BL6 mice demonstrated superior object recognition memory following 9+ months of supplementation, suggesting that longer treatment periods may have a positive impact on recognition memory in 3xTg-AD mice as well [114]. In fact, given the strength of some results with our relatively short treatment period suggests that some of the AD features that proved refractory to supplementation might still show some benefit by a longer treatment period. It also appears that the MDS negatively impacts olfactory acuity in 3-4-month-old females of both genotypes. This was surprising, since previous work on the MDS has found beneficial effects on olfactory function in older animals (9 months and older [26, 27]), suggesting that this side effect is age-dependent and is unlikely to be an issue if treatment is started at a later age in wild type animals. The sources of these deficits require further investigation.

Our recent work suggests that the MDS in combination with aerobic exercise [57] may yield greater benefits than either treatment alone, particularly when elevated levels of stress are involved in a disease phenotype, such as is the case for the 3xTg-AD mouse [35, 113]. Again, the benefit of supplementation (and perhaps the interaction with exercise) may have been improved with a longer supplementation period. With respect to the current water maze results, if 3xTg-AD females experience a greater stress response to testing than males at 3-4 months of age, it could explain why they, but not 3xTg-AD males, exhibited a deficit in long term memory based on the water maze probe trials that was not responsive to the MDS alone. Given that impairment in 3xTg-AD females on the probe trials was relatively small (i.e., they could still remember the platform location), we recommend that future studies of young 3xTg-AD mice use converging evidence from additional performance measures, some of which (e.g., platform crossings or average proximity to the platform) are more sensitive to group differences than the time spent in the target quadrant [115].

Our previous finding that both exercise and the MDS, but neither alone, were sufficient to improve anhedonia, hippocampal neurogenesis, BDNF, and hippocampal size in chronically stressed mice [57] suggests that the MDS and exercise together may prevent the deficit in probe trial performance observed here in 3xTg-AD females. This is consistent with other literature showing increased effects of combined diet and exercise [59 , 117].

Taken together, the results presented here suggest that as little as 2 months of supplementation with the MDS protects cognitive functions such as working spatial memory in juvenile 3xTg-AD mice. These functions are supported by highly plastic structures such as the prefrontal cortex and hippocampus that are affected most severely in AD [1]. However, such a short period of supplementation was insufficient to preserve recognition memory or reverse pre-existing anxiety, functions which depend more on other brain areas such as the perirhinal cortex (supports object recognition [118]) or amygdala (affects anxiety-like behavior [119]). Future studies are indicated to determine if longer periods of supplementation or combination with exercise can reverse these other impairments or impact AD-relevant biomarkers (e.g., amyloid-β or hyperphosphorylated tau or BDNF). Regardless, the present data suggest that dietary supplements can obtain remarkable results, at least in mice. Ours and other studies suggest that supplements may serve as a key part of a multi-pronged intervention program to protect against AD-related behavioral changes.

Footnotes

ACKNOWLEDGMENTS

We would like to thank Dr. M. Kapadia for advice on the olfactory testing protocol. We also recognize the substantial efforts of undergraduate students J. Tran, M. Lysenko-Martin, Z. Patel, N. Ramseyer, M. Kryzewska, G. Kaur, N. Elsaadawy, D. Chok, H. Shahid, M. Campbell, L. Ribeiro, L. Babatinca, K. Thomas, B. Terpou, A. Mizzi, M. Radenovic and I. Perera in assisting with the behavioral testing, the scoring of behavioral video recordings, and administration of the dietary supplement. This work was supported by grants from the Natural Sciences and Engineering Research Council of Canada (NSERC) to S.B. (RGPIN-2014-03580), D.R.B. (RGPIN 238495-10), and J.M.W. (RGPIN 194616-11). It was also supported by a grant from the Canadian Institutes of Health Research to J.M.W. (MOP 11927), grants from the Alzheimer’s Society of Canada (grant #17-04) and the Canadian Consortium on Neurodegeneration in Aging Team 2 to M.F., and an NSERC Canada Graduate Scholarship to C.H.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Mattson

(2004) Pathways towards and away from Alzheimer’s disease., Nature 430, 631–639.

Burgess

, Maguire

, O’Keefe

(2002) The human hippocampus and spatial and episodic memory. Neuron 35, 625–641.

Sekeres

, Bonasia

, St-Laurent

, Pishdadian

, Winocur

, Grady

, Moscovitch

(2016) Recovering and preventing loss of detailed memory: Differential rates of forgetting for detail types inepisodic memory. Learn Mem 23, 72–82.

Snyder

, Soumier

, Brewer

, Pickel

, Cameron

(2011) Adult hippocampal neurogenesis buffers stress responses and depressive beavihor. Nature 476, 458–461.

Lyketsos

, Lopez

, Jones

, Fitzpatrick

, Breitner

, DeKosky

(2013) Prevalence of neuropsychiatric symptoms results from thecardiovascular health study. JAMA 288, 1475–1483.

Hardy

, Higgins

(1992) Alzheimer’s disease: The amyloid cascade hypothesis. Science 256, 184–185.

Selkoe

, Hardy

(2016) The amyloid hypothesis of Alzheimer’sdisease at 25 years. EMBO Mol Med 8, 1–14.

Q-LQ-L

, Yang

, Rosario

, Ubeda

, Beech

, Gant

, Chen

, Hudspeth

, Chen

, Zhao

, Vinters

, Frautschy

, Cole

(2009) Beta-amyloid oligomers induce phosphorylation of tau andinactivation of insulin receptor substrate via c-Jun N-terminalkinase signaling: Suppression by omega-3 fatty acids and curcumin. J Neurosci 29, 9078–9089.

Jin

, Shepardson

, Yang

, Chen

, Walsh

, Selkoe

(2011) Soluble amyloid beta-protein dimers isolated from Alzheimer cortexdirectly induce Tau hyperphosphorylation and neuritic degeneration. Proc Natl Acad Sci U S A 108, 5819–5824.

10.

Herrup

(2015) The case for rejecting the amyloid cascadehypothesis. Nat Neurosci 18, 794–799.

11.

Mattson

(2015) Late-onset dementia: A mosaic of prototypicalpathologies modifiable by diet and lifestyle. NPJ Aging Mech Dis 1, 15003.

12.

Terry

, Masliah

, Salmon

, Butters

, DeTeresa

, Hill

, Hansen

, Katzman

(1991) Physical basis of cognitive alterationsin Alzheimer’s disease: Synapse loss is the major correlate of cognitive impairment. Ann Neurol 30, 572–580.

13.

Peng

, Wuu

, Mufson

, Fahnestock

(2005) Precursor form of brain-derived neurotrophic factor and mature brain-derivedneurotrophic factor are decreased in the pre-clinical stages ofAlzheimer’s disease. J Neurochem 93, 1412–1421.

14.

Michalski

, Corrada

, Kawas

, Fahnestock

(2015) Brain-derived neurotrophic factor and TrkB expression in the “oldest-old,” the90+ Study: Correlation with cognitive status and levels of solubleamyloid-beta. Neurobiol Aging 36, 3130–3139.

15.

Morris

, Storandt

, Miller

, McKeel

, Price

, Rubin

, Berg

(2001) Mild cognitive impairment represents early-stage Alzheimer disease. Arch Neurol 58, 397–405.

16.

Tarkowski

, Andreasen

, Tarkowski

, Blennow

(2003) Intrathecal inflammation precedes development of Alzheimer’s disease. JNeurol Neurosurg Psychiatry 74, 1200–1205.

17.

Talbot

, Wang

, Kazi

, Han

, Bakshi

, Stucky

, Fuino

, Kawaguchi

, Samoyedny

, Wilson

, Arvanitakis

, Schneider

, Wolf

, Bennett

, Trojanowski

, Arnold

(2012) Demonstrated brain insulin resistance in Alzheimer’s disease patients is assocaited with IGF-1 resisitance, IRS-1 dysregulation, and cogntivedecline. J Clin Invest 122, 1316–1338.

18.

Willette

, Bendlin

, Starks

, Birdsill

, Johnson

, Christian

, Okonkwo

, La Rue

, Hermann

, Koscik

, Jonaitis

, Sager

, Asthana

(2015) Association of insulin resistance with cerebral glucose uptake in late middle– aged adults atrisk for Alzheimer disease. JAMA Neurol 72, 1013.

19.

Lovell

, Soman

, Bradley

(2011) Oxidatively modified nucleic acids in preclinical Alzheimer’s disease (PCAD) brain. Mech Ageing Dev 132, 443–448.

20.

Mosconi

, Pupi

, De Leon

(2008) Brain glucose hypometabolismand oxidative stress in preclinical Alzheimer’s disease. Ann NY Acad Sci 1147, 180–195.

21.

Parachikova

, Green

, Hendrix

, Laferla

(2010) Formulation of a medical food cocktail for Alzheimer’s disease: Beneficial effects on cognition and neuropathology in a mouse model of the disease. PLoS One 5, e14015.

22.

Wolf

, Braden

, Bimonte-Nelson

, Kusne

, Young

, Engler-Chiurazzi

, Garcia

, Walker

, Moses

GSD

, Tran

, LaFerla

, Lue

, Lombardo

, Valla

(2012) Broad-based nutritionalsupplementation in 3xTg mice corrects mitochondrial function and indicates sex-specificity in response to Alzheimer’s disease intervention. J Alzheimers Dis 32, 217–232.

23.

Stonehouse

, Conlon

, Podd

, Hill

, Minihane

, Haskell

, Kennedy

(2013) DHA supplementation improved both memory andreaction time in healthy young adults: A randomized controlled trial1–3. Am J Clin Nutr 94, 1134–1143.

24.

Dai

, Borenstein

, Wu

, Jackson

, Larson

(2006) Fruit andvegetable juices and Alzheimer’s disease: The Kame Project. AmJ Med 119, 751–759.

25.

Aksenov

, Long

, Liu

, Szechtman

, Khanna

, Matravadia

, Rollo

(2013) A complex dietary supplement augments spatial learning, brain mass, and mitochondrial electron transport chain activity inaging mice. Age (Dordr) 35, 23–33.

26.

Lemon

, Aksenov

, Samigullina

, Aksenov

, Rodgers

, Rollo

, Boreham

(2016) A multi-ingredient dietary supplement abolishes large-scale brain cell loss, improves sensory function, and prevents neuronal atrophy in aging mice. Environ Mol Mutagen 57, 382–404.

27.

Lemon

, Boreham

, Rollo

(2003) A dietary supplement abolishes age-related cognitive decline in transgenic mice expressing elevated free radical processes. Exp Biol Med 228, 800–810.

28.

Lemon

, Rollo

, McFarlane

, Boreham

(2008) Radiation-induced apoptosis in mouse lymphocytes is modified by acomplex dietary supplement: The effect of genotype and gender. Mutagenesis 23, 465–472.

29.

Aksenov

, Long

, Lokuge

, Foster

, Liu

, Rollo

(2010) Dietary amelioration of locomotor, neurotransmitter andmitochondrial aging. Exp Biol Med 235, 66–76.

30.

Oddo

, Caccamo

, Shepherd

, Murphy

, Golde

, Kayed

, Metherate

, Mattson

, Akbari

, LaFerla

(2003) Triple-transgenic model of Alzheimer’s disease with plaques andtangles: Intracellular Aβ and synaptic dysfunction. Neuron 39, 409–421.

31.

Sterniczuk

, Antle

, Laferla

, Dyck

(2010) Characterization of the 3xTg-AD mouse model of Alzheimer’s disease: Part 2. Beavihoral and cognitive changes. Brain Res 1348, 149–155.

32.

Filali

, Lalonde

, Theriault

, Julien

, Calon

, Planel

(2012) Cognitive and non-cognitive beavihors in the triple transgenic mousemodel of Alzheimer’s disease expressing mutated APP, PS1, and Mapt(3xTg-AD). Behav Brain Res 234, 334–342.

33.

Stover

, Campbell

, Van Winssen

, Brown

(2015) Analysis of motor function in 6-month-old male and female 3xTg-AD mice. Behav Brain Res 281, 16–23.

34.

Stevens

, Brown

(2015) Reference and working memory deficits inthe 3xTg-AD mouse between 2 and 15-months of age: A cross-sectionalstudy. Behav Brain Res 278, 496–505.

35.

Clinton

, Billings

, Green

, Caccamo

, Ngo

, Oddo

, McGaugh

, LaFerla

(2007) Age-dependent sexual dimorphism in cognitionand stress response in the 3xTg-AD mice. Neurobiol Dis 28, 76–82.

36.

Billings

, Oddo

, Green

, McGaugh

, LaFerla

(2005) Intraneuronal Aß causes the onset of early Alzheimer’s disease-related cognitive deficits in transgenic mice. Neuron 45, 675–688.

37.

Davis

, Easton

, Eacott

, Gigg

(2013) Episodic-like memory for what-where-which occasion is selectively impaired in the 3xTgADmouse model of Alzheimer’s disease. J Alzheimers Dis 33, 681–698.

38.

Marchese

, Cowan

, Head

, Ma

, Karimi

, Ashthorpe

, Kapadia

, Zhao

, Davis

, Sakic

(2014) Autoimmune manifestations in the3xTg-AD model of Alzheimer’s disease. J Alzheimers Dis 39, 191–210.

39.

Carroll

, Rosario

, Kreimer

, Villamagna

, Gentzschein

, Stanczyk

, Pike

(2010) Sex differences in β-amyloid accumulation in 3xTg-AD mice: Role of neonatal sex steroid hormone exposure. Brain Res 1366, 233–245.

40.

Davis

, Burnett

, Gigg

(2017) Water and T-maze protocols areequally efficient methods to assess spatial memory in 3xTg Alzheimer’s disease mice. Behav Brain Res 331, 54–66.

41.

Attar

, Liu

, Chan

W-TC

, Hayes

, Nejad

, Lei

, Bitan

(2013) Ashortened Barnes maze protocol reveals memory deficits at 4-monthsof age in the triple-transgenic mouse model of Alzheimer’s disease. PLoS One 8, e80355.

42.

Pietropaolo

, Feldon

, Yee

(2014) Environmental enrichmenteliminates the anxiety phenotypes in a triple transgenic mouse model of Alzheimer’s disease. Cogn Affect Behav Neurosci 14, 996–1008.

43.

García-Mesa

, López-Ramos

, Giménez-Llort

, Revilla

, Guerra

, Gruart

, Laferla

, Cristòfol

, Delgado-García

, Sanfeliu

(2011) Physical exercise protectsagainst Alzheimer’s disease in 3xTg-AD mice. J Alzheimers Dis 24, 421–454.

44.

Hirata-Fukae

, Li

H-F

, Hoe

H-S

, Gray

, Minami

, Hamada

, Niikura

, Hua

, Tsukagoshi-Nagai

, Horikoshi-Sakuraba

, Mughal

, Rebeck

, LaFerla

, Mattson

, Iwata

, Saido

, Klein

, Duff

, Aisen

, Matsuoka

(2008) Females exhibit more extensive amyloid, but not tau, pathology in an Alzheimer transgenic model. Brain Res 1216, 92–103.

45.

Mastrangelo

, Bowers

(2008) Detailed immunohistochemical characterization of temporal and spatial progression of Alzheimer’sdisease-related pathologies in male triple-transgenic mice. BMC Neurosci 9, 81.

46.

España

, Giménez-Llort

, Valero

, Miñano

, Rábano

, Rodriguez-Alvarez

, LaFerla

, Saura

(2010) Intraneuronal β-amyloid accumulation in the amygdala enhances fear and anxiety in Alzheimer’s disease transgenic mice. Biol Psychiatry 67, 513–521.

47.

Sawmiller

, Habib

, Li

, Darlington

, Hou

, Tian

, Shytle

, Smith

, Giunta

, Mori

, Tan

(2016) Diosmin reduces cerebral Aβ levels, tau hyperphosphorylation, neuroinflammation, and cognitive impairment in the 3xTg-AD mice. J Neuroimmunol 299, 98–106.

48.

Walls

, Ager

, Vasilevko

, Cheng

, Medeiros

, LaFerla

(2014) P-Tau immunotherapy reduces soluble and insoluble tau in aged3xTg-AD mice. Neurosci Lett 575, 96–100.

49.

Liu

, Breitbart

, Sun

, Mehta

, Boutajangout

, Scholtzova

, Wisniewski

(2014) Blocking the apolipoprotein E/amyloid βinteraction in triple transgenic mice ameliorates Alzheimer’s disease related amyloid β and tau pathology. J Neurochem 128, 577–591.

50.

Ally

, Hussey

, Ko

, Molitor

(2013) Pattern separation andpattern completion in Alzheimer’s disease: Evidence of rapid forgetting in amnestic mild cognitive impairment. Hippocampus 23, 1246–1258.

51.

Luu

, Sill

, Gao

, Becker

, Wojtowicz

, Smith

(2012) Therole of adult hippocampal neurogenesis in reducing interference. Behav Neurosci 126, 381–391.

52.

Winocur

, Becker

, Luu

, Rosenzweig

, Wojtowicz

(2012) Adult hippocampal neurogenesis and memory interference. Behav Brain Res 227, 464–469.

53.

Francis

, Kim

, Barakat

, Fraenkl

, Yücel

, Peng

, Michalski

, Fahnestock

, McLaurin

, Mount

(2012) Object recognition memory and BDNF expression are reduced in young TgCRND8 mice. Neurobiol Aging 33, 555–563.

54.

Rodríguez

, Jones

, Tabuchi

, Allan

, Knight

, LaFerla

, Oddo

, Verkhratsky

(2008) Impaired adult neurogenesis in thtate gyrus of a triple transgenic mouse model of Alzheimer’s disease. PLoS One 3, e2935.

55.

Demars

, Hu

Y-S

, Gadadhar

, Lazarov

(2010) Impaired neurogenesis is an early event in the etiology of familial Alzheimer’s disease intransgenic mice. J Neurosci Res 88, 2103–2117.

56.

Hamilton

, Aumont

, Julien

, Vadnais

, Calon

, Fernandes

KJL

(2010) Widespread deficits in adult neurogenesis precede plaque and tangle formation in the 3xTg mouse model of Alzheimer’s disease. Eur J Neurosci 32, 905–920.

57.

Hutton

, Déry

, Rosa

, Lemon

, Rollo

, Boreham

, Fahnestock

, deCatanzaro

, Wojtowicz

, Becker

(2015) Synergistic effects of diet and exercise on hippocampal function inchronically stressed mice. Neuroscience 308, 180–193.

58.

Valente

, Hidalgo

, Bolea

, Ramirez

, Anglés

, Reguant

, Morelló

, Gutiérrez

, Boada

, Unzeta

(2009) A diet enriched in polyphenols and polyunsaturated fatty acids, LMN diet, induces neurogenesis in the subventricular zone and hippocampus of adult mouse brain. J Alzheimers Dis 18, 849–865.

59.

Fahnestock

, Marchese

, Head

, Pop

, Michalski

, Milgram

, Cotman

(2012) BDNF increases with beavihoral enrichment and an anti oxidant diet in the aged dog. Neurobiol Aging 33, 546–554.

60.

Rendeiro

, Vauzour

, Kean

, Butler

, Rattray

, Spencer

JPE

, Williams

(2012) Blueberry supplementation induces spatial memory improvements and region-specific regulation of hippocampal BDNF mRNA expression in young rats. Psychopharmacology (Berl) 223, 319–330.

61.

Williams

, El Mohsen

, Vauzour

, Rendeiro

, Butler

, Ellis

, Whiteman

, Spencer

JPE

(2008) Blueberry-induced changes inspatial working memory correlate with changes in hippocampal CREB phosphorylation and brain-derived neurotrophic factor (BDNF) levels. Free Radic Biol Med 45, 295–305.

62.

Hall

, Roberson

(2012) Mouse models of Alzheimer’s disease. Brain Res Bull 88, 3–12.

63.

Webster

, Bachstetter

, Nelson

, Schmitt

, Van Eldik

(2014) Using mice to model Alzheimer’s dementia: An overview of theclinical disease and the preclinical beavihoral changes in 10 mouse models. Front Genet 5, 1–23.

64.

Lyras

, Cairns

, Jenner

, Halliwell

(1997) Anassessment of oxidative damage to proteins, lipids, and DNA in brainfrom patients with Alzheimer’s disease. J Neurochem 68, 2061–2069.

65.

Mattson

, Pedersen

, Duan

, Culmsee

, Camandola

(1999) Cellular and molecular mechanisms underlying perturbed energymetabolism and neuronal degeneration in Alzheimer’s and Parkinson’s diseases. Ann N Y Acad Sci 893, 154–175.

66.

Butterfield

, Drake

, Pocernich

, Castegna

(2001) Evidence of oxidative damage in Alzheimer’s disease brain: Central role for amyloid beta-peptide.. Trends Mol Med 12, 548–554.

67.

Mutlu-Türkoğlu

, Ilhan

, Oztezcan

, Kuru

, Aykaç-Toker

, Uysal

(2003) Age-related increases in plasmamalondialdehyde and protein carbonyl levels and lymphocyte DNAdamage in elderly subjects. Clin Biochem 36, 397–400.

68.

Sastre

, Pallardó

, Viña

(2003) The role of mitochondrial oxidative stress in aging. Free Radic Biol Med 35, 1–8.

69.

Antier

, Carswell

, Brosnan

, Hamilton

, Macrae

, Groves

, Jardine

, Reid

, Dominiczak

(2004) Increased levels of superoxide in brains from old female rats. Free Radic Res 38, 177–183.

70.

Long

, Aksenov

, Rollo

, Liu

(2012) A complex dietary supplement modulates nitrative stress in normal mice and in a newmouse model of nitrative stress and cognitive aging. Mech Ageing Dev 133, 523–529.

71.

Mega

, Cummings

, Fiorello

, Gornbein

(1996) The spectrum of beavihoral changes in Alzheimer’s. Neurology 46, 130–135.

72.

Kluger

, Gianutsos

, Golomb

, Ferris

, George

, Franssen

, Reisberg

(1997) Patterns of motor impairment in normal aging, mild cognitive decline, and early Alzheimer’ disease. J Gerontol Ser B Psychol Sci Soc Sci 52B, P28–P39.

73.

Palmer

, Berger

, Monastero

, Winblad

, Bäckman

, Fratiglioni

(2007) Predictors of progression from mild cognitive impairment to Alzheimer disease. Neurology 68, 1596–1602.

74.

Heyser

(2004) Assessment of developmental milestones in rodents. Curr Protoc Neurosci Chapter 8, Unit 8.18.

75.

Guyenet

, Furrer

, Damian

, Baughan

, La Spada

, Garden

(2010) A simple composite phenotype scoring system for evaluating mouse models of cerebellar ataxia, J Vis Exp, doi: 10.3791/1787

76.

Pettersson

, Olsson

, Wahlund

(2005) Motor function insubjects with mild cognitive impairment and early Alzheimer’s disease. Dement Geriatr Cogn Disord 19, 299–304.

77.

Jones

, Roberts

(1968) The quantitative measurement of motorinco-ordination in naive mice using an accelerating rotarod. J Pharm Pharmacol 20, 302–304.

78.

Klein

, Vykoukal

, Lechler

, Zeitler

, Gehmert

, Schreml

, Alt

, Bogdahn

, Prantl

(2012) Noninvasive in vivo assessment of muscle impairment in the mdx mouse model - A comparison of two common wire hanging methods with two different results. J Neurosci Methods 203, 292–297.

79.

Devanand

, Michaels-Marston

, Liu

, Pelton

, Padilla

, Marder

, Bell

, Stern

, Mayeux

(2000) Olfactory deficits inpatients with mild cognitive impairment predict Alzheimer’s diseaseat follow-up. Am J Psychiatry 157, 1399–1405.

80.

Djordjevic

, Jones-Gotman

, De Sousa

, Chertkow

(2008) Olfaction in patients with mild cognitive impairment and Alzheimer’s disease. Neurobiol Aging 29, 693–706.

81.

Gheusi

, Cremer

, McLean

, Chazal

, Vincent

, Lledo

(2000) Importance of newly generated neurons in the adult olfactory bulbfor odor discrimination. Proc Natl Acad Sci U S A 97, 1823–1828.

82.

Carola

, D’Olimpio

, Brunamonti

, Mangia

, Renzi

(2002) Evaluation of the elevated plus-maze and open-field tests for the assessment of anxiety-related beavihours in inbred mice. Behav Brain Res 134, 49–57.

83.

Lister

(1990) Ethologically-based animal models of anxiety disorders. Pharmacol Ther 46, 321–340.

84.

Sakic

, Cooper

MPA

, Taylor

, Stojanovic

, Zagorac

, Kapadia

(2015) Beavihoral phenotyping of murine disease models with the integrated behoral station (INBEST), J Vis Exp, doi: 10.3791/51524

85.

Willner

, Muscat

, Papp

(1992) Chronic mild stress-induced anhedonia: A realistic animal model of depression. Neurosci Biobehav Rev 16, 525–534.

86.

Strekalova

, Spanagel

, Bartsch

, Henn

, Gass

(2004) Stress-induced anhedonia in mice is associated with deficits inforced swimming and exploration. Neuropsychopharmacology 29, 2007–2017.

87.

Bevins

, Besheer

(2006) Object recognition in rats and mice: Aone-trial non-matching-to-sample learning task to study “recognition memory.”, Nat Protoc 1, 1306–1311.

88.

Lalonde

(2002) The neurobiological basis of spontaneous alternation. Neurosci Biobehav Rev 26, 91–104.

89.

Morris

(1984) Developments of a water-maze procedure for studying spatial learning in the rat. J Neurosci Methods 11, 47–60.

90.

Fox

(2008) Applied regression analysis and generalized linear models, 2nd ed., Sage Publications, Thousand Oaks, CA.

91.

Anderson

, Braak

(2003) Permutation tests for multi-factorialanalysis of variance. J Stat Comp Sim 73, 85–113.

92.

Wheeler

, Torchiano

(2016) : The R project for statistical computing.URL:, Permutation testsfor linear models in R URL: https://github.com/mtorchiano/lmPerm.

93.

Fox

, Weisberg

, (2011) An {R} companion to applied regression, 2nd ed., Sage Publications, Thousand Oaks, CA.

94.

Singmann

, Bolker

, Westfall

, Aust

(2016) Analysis of factorial experiments. The R project for statistical computing. URL: https://CRAN.R-project.org/package=afex.

95.

Lenth

(2016) Least-squares means: The R package lsmeans. J Stat Softw 69, 1–33.

96.

R Core Team (2017) R: A language and environment for statistical computing. R Foundation for Statistical Computing. URL: https://www.R-project.org/.

97.

Wickham

(2009), ggplot2: Elegant graphics for data analysis Springer-Verlag, New York.

98.

Richardson

(2011) Eta squared and partial eta squared as measuresof effect size in educational research. Educ Res Rev 6, 135–147.

99.

Bakeman

(2005) Recommended effect size statistics for repeatedmeasures designs. Behav Res Methods 37, 379–384.

100.

Rosnow

, Rosenthal

(2003) Effect sizes for experimenting psychologists. Can J Exp Psychol 57, 221.

101.

Furr

(2004) Interpreting effect sizes in contrast analysis., 1, Understanding Stat 3, 125.

102.

Cohen

(1969), Statistical power analysis for the behaviouralsciences Academic Press, New York.

103.

, Chung

, An

, Lim S

jin

, Choi

, Yu

, Kim

, Kang

, Cho

, Shim

, Bae

(2016) Standardized herbal formula PM012 decreases cognitive impairment and promotes neurogenesis in the 3xTg-AD mouodel of Alzheimer’s disease. Mol Neurobiol 58, 5401–5412.

104.

Bohbot

, McKenzie

, Konishi

, Fouquet

, Kurdi

, Schachar

, Boivin

, Robaey

(2012) Virtual navigation strategies from childhood to senescence: Evidence for changes across the life span. Front Aging Neurosci 4, 28.

105.

Jack

Jr. , Petersen

, Xu

, O’Brien

, Smith

, Ivnik

, Boeve

, Tangalos

, Kokmen

(2000) Rates of hippocampal atrophy correlate with change in clinical status in aging and AD. Neurology 55, 484–489.

106.

Garthe

, Behr

, Kempermann

(2009) Adult-generated hippocampal neurons allow the flexible use of spatially precise learning strategies. PLoS One 4, e5464.

107.

Garthe

, Huang

, Kaczmarek

, Filipkowski

, Kempermann

(2014) Not all water mazes are created equal: Cyclin D2 knockout mice withconstitutively suppressed adult hippocampal neurogenesis do showspecific spatial learning deficits. Genes, Brain Behav 13, 357–364.

108.

Hörster

, Garthe

, Walker

, Ichwan

, Steiner

, Khan

, Lie

, Nicola

, Ramirez-Rodriguez

, Kempermann

(2017) Kiplis required for functionally relevant adult hippocampal neurogenesisin mice. Stem Cells 35, 787–799.

109.

Zhang

C-L

, Zou

, He

, Gage

, Evans

(2008) A role for adult TLX-positive neural stem cells in learning and beavihour. Nature 451, 1004–1007.

110.

Brozka

, Pistikova

, Radostova

, Vales

, Svoboda

, Grzyb

, Stuchlik

(2017) Adult neurogenesis reduction by a cytostatic treatment improves spatial reversal learning in rats. Neurobiol Learn Mem 141, 93–100.

111.

Becker

(2005) A computational principle for hippocampal learningand neurogenesis. Hippocampus 15, 722–738.

112.

Wesnes

, Annas

, Basun

, Edgar

, Blennow

(2014) Performance on a pattern separation task by Alzheimer’s patients shows possible links between disrupted dentate gyrus activity and apolipoprotein E∈4 status and cerebrospinal fluid amyloid-β42 levels. Alzheimers Res Ther 6, 20.

113.

Hebda-Bauer

, Simmons

, Sugg

, Ural

, Stewart

, Beals

, Wei

, Watson

, Akil

(2013) 3xTg-AD mice exhibit an activated central stress axis during early-stage pathology. J Alzheimers Dis 33, 407–422.

114.

Maei

, Zaslavsky

, Teixeira

, Frankland

(2009) What is themost sensitive measure of water maze probe test performance? Front Integr Neurosci 3, 4.

115.

Aksenov

(2014) Complex dietary Interventions to Slow Rates of Aging. (Doctoral dissertation) Retrieved from macsphere.mcmaster.ca. (http://hdl.handle.net/11375/15347.; Identifier #5485838).

116.

Pop

, Head

, Hill

M-A

, Gillen

, Berchtold

, Muggenburg

, Milgram

, Murphy

, Cotman

(2010) Synergistic effects oflong-term antioxidant diet and beavihoral enrichment on beta-amyloidload and non-amyloidogenic processing in aged canines. J Neurosci 30, 9831–9839.

117.

, Ying

, Gomez-Pinilla

(2008) Docosahexaenoic acid dietary supplementation enhances the effects of exercise on synaptic plasticity and cognition. Neuroscience 155, 751–759.

118.

Wan

, Aggleton

, Brown

(1999) Different contributions of the hippocampus and perirhinal cortex to recognition memory. J Neurosci 19, 1142–1148.

119.

McHugh

, Deacon

RMJ

, Rawlins

JNP

, Bannerman

(2004) Amygdala andventral hippocampus contribute differentially to mechanisms of fearand anxiety. Behav Neurosci 118, 63–78.