influence of selected diets on neural insulin activity and ......modulate cognitive function, at...
TRANSCRIPT
Influence of Selected Diets on Neural Insulin Activity and
Cognitive Function
by
Matthew David Parrott
A thesis submitted in conformity with the requirements
for the degree of Doctor of Philosophy
Department of Nutritional Sciences
University of Toronto
© Copyright by Matthew David Parrott, 2014
Influence of Selected Diets on Neural Insulin Activity and Cognitive
Function
Matthew David Parrott
Doctor of Philosophy
Department of Nutritional Sciences
University of Toronto
2014
Abstract
Diets high in saturated fat are related to worse cognitive function while those high in fish,
vegetables and fruit have been associated with better cognitive function and lower dementia risk.
While the precise physiological mechanisms underlying these dietary influences are not
completely understood, modulation of brain-insulin activity and neuroinflammation were
speculated to contribute. The results of this thesis suggest that the associations of fish, fruit,
vegetables and saturated fat with cognitive function reflect adherence to a broader set of dietary
patterns in older adults whose own relationship with cognition may be dependent on individual
differences in environmental conditions, innate metabolism, and/or genetics. Although markers
of brain insulin signaling were related to optimal cognitive function in rodent studies, diet-
induced modulation of the hippocampal insulin signaling pathway was, at best, indirect. Diet-
induced, peripheral insulin resistance and metabolic dysfunction were closely related to cognitive
deficits.
ii
Table of Contents
Table of Contents ........................................................................................................................... iii
List of Tables ................................................................................................................................ vii
List of Figures .............................................................................................................................. viii
List of Abbreviations ..................................................................................................................... ix
1 Chapter 1: General Introduction ................................................................................................ 1
2 Chapter 2: Background Literature & Theoretical Overview ..................................................... 4
2.1 Neuropathologic events associated with disruptions to brain insulin, insulin-mediated cell signaling, inflammation and oxidative stress ............................................................... 5
2.1.1 Role of insulin signaling in memory processing ..................................................... 6
2.1.2 Loss of insulin signaling and associated neuropathologic events ........................... 7
2.1.3 Insulin resistance, inflammation & brain function ................................................. 8
2.1.4 Insulin resistance and neurotrophins ....................................................................... 9
2.2 Association of foods and nutrients with cognition: Potential links to brain insulin signaling and neuroinflammation ..................................................................................... 10
2.2.1 Saturated fat .......................................................................................................... 11
2.2.2 Omega-3 fatty acids and fish oils .......................................................................... 11
2.2.3 Dietary antioxidants of plant origin ...................................................................... 13
2.3 Summary and Rationale .................................................................................................... 14
2.4 Objectives & Hypotheses .................................................................................................. 15
2.4.1 Objectives ............................................................................................................. 16
2.4.2 Hypotheses ............................................................................................................ 16
3 Chapter 3: Relationship between diet quality and cognition depends on socioeconomic position in healthy older adults ................................................................................................ 18
3.1 Abstract ............................................................................................................................. 19
3.2 Introduction ....................................................................................................................... 20
3.3 Methods ............................................................................................................................. 21 iii
3.3.1 Participants ............................................................................................................ 21
3.3.2 Statistical analysis ................................................................................................. 22
3.4 Results ............................................................................................................................... 24
3.4.1 Participant characteristics ..................................................................................... 24
3.4.2 Dietary patterns ..................................................................................................... 26
3.4.3 Final models .......................................................................................................... 26
3.5 Discussion ......................................................................................................................... 43
3.6 Acknowledgements ........................................................................................................... 46
4 Chapter 4: Whole-food diet worsened cognitive dysfunction in an Alzheimer’s disease mouse model ............................................................................................................................ 47
4.1 Abstract ............................................................................................................................. 48
4.2 Introduction ....................................................................................................................... 49
4.3 Methods ............................................................................................................................. 50
4.3.1 Mice and diets ....................................................................................................... 50
4.3.2 Cognitive testing ................................................................................................... 53
4.3.3 Genotyping by polymerase chain reaction (PCR) ................................................ 56
4.3.4 Hippocampal gene expression analysis (Quantitative reverse real-time PCR) .... 57
4.3.5 Cortical Aβ burden ................................................................................................ 59
4.3.6 Statistical analysis ................................................................................................. 59
4.4 Results ............................................................................................................................... 60
4.4.1 Cognitive function ................................................................................................ 60
4.4.2 Gene expression .................................................................................................... 61
4.4.3 Cortical Aβ burden ................................................................................................ 61
4.5 Discussion ......................................................................................................................... 71
4.6 Acknowledgements ........................................................................................................... 77
4.6.1 Disclosure Statements ........................................................................................... 77
iv
5 Chapter 5: Rosiglitazone prevents hippocampal-dependent memory deficits associated with peripheral metabolic dysfunction in a rat model of diet-induced obesity ........................ 78
5.1 Abstract ............................................................................................................................. 79
5.2 Introduction ....................................................................................................................... 80
5.3 Methods ............................................................................................................................. 81
5.3.1 Subjects and Diets ................................................................................................. 81
5.3.2 Variable-interval delayed alternation (VIDA) task ............................................... 84
5.3.3 Intracerebroventricular insulin infusion and tissue collection .............................. 84
5.3.4 Plasma biochemistry ............................................................................................. 85
5.3.5 Hippocampal gene expression analysis (Quantitative reverse transcription real-time PCR) ...................................................................................................... 85
5.3.6 Immunoblot analysis of hippocampal protein abundance .................................... 88
5.3.7 Statistical Analyses ............................................................................................... 88
5.4 Results ............................................................................................................................... 90
5.4.1 Variable-Interval Delayed Alternation (VIDA) task ............................................ 90
5.4.2 Fasting plasma biochemistry and body measurements ......................................... 91
5.4.3 Insulin-stimulated differences in hippocampal gene expression and protein abundance ............................................................................................................. 92
5.4.4 Correlations between plasma biomarkers and hippocampal-dependent memory . 92
5.4.5 Correlations between hippocampal insulin-signaling proteins and hippocampal-dependent memory .......................................................................... 93
5.5 Discussion ....................................................................................................................... 103
6 Chapter 6: General Discussion of Thesis Results ................................................................. 110
6.1 Overview of Objectives & Summary of Results ............................................................. 111
6.2 Separate Dietary Components & Associations with Global Diet Quality ...................... 114
6.3 Effects of Diet Quality on Cognitive Function ............................................................... 115
6.4 Functional Insulin & Cognitive Functions ...................................................................... 118
v
6.5 Cognitive Impairment: Combined Effects of Inflammation & Reduced Functional Insulin Activity ............................................................................................................... 119
6.6 Implications & Directions for Future Research .............................................................. 120
6.7 Conclusions ..................................................................................................................... 124
References ................................................................................................................................... 126
vi
List of Tables Table 3.1. Characteristics of the analyzed and unanalyzed participants ...................................... 25
Table 3.2. Factor loadings for dietary patterns1 ........................................................................... 28
Table 3.3. Baseline characteristics of the NuAge study participants across quintiles of diet
pattern score .................................................................................................................................. 32
Table 3.4. Associations between dietary patterns and cognitive function, and interactions with
socioeconomic indicators, in participants of the NuAge study1 ................................................... 35
Table 3.5. Associations between dietary patterns and cognitive function within socioeconomic
subgroups1 ..................................................................................................................................... 38
Table 4.1. Composition of experimental diets ............................................................................. 52
Table 4.2. Genes targeted for reverse transcription quantitative real-time PCR ......................... 58
Table 4.3. Body weights of experimental animals ....................................................................... 62
Table 4.4. Average number of trials required to reach criterion on the brightness discrimination
test ................................................................................................................................................. 67
Table 4.5. Cortical Aβ content of transgenic animals by dieta .................................................... 70
Table 5.1. Composition of experimental diets .............................................................................. 83
Table 5.2. Genes targeted for reverse transcription quantitative real-time PCR ......................... 87
Table 5.3. Fasting plasma biochemistry and body measurements* .............................................. 95
vii
List of Figures
Figure 3.1. Association between 3MS score and test year stratified by selected socioeconomic
indicators and prudent pattern score ............................................................................................. 40
Figure 3.2. Association between 3MS score and test year stratified by education and Western
pattern score. ................................................................................................................................. 42
Figure 4.1. Latencies for the spatial memory test acquisition and probe trial performance ........ 63
Figure 4.2. Errors for the spatial memory test acquisition ........................................................... 64
Figure 4.3. Latencies for the non-matching-to-sample test ......................................................... 65
Figure 4.4. Errors for the non-matching-to-sample test. .............................................................. 66
Figure 4.5. Statistically significant differences in hippocampal gene expression ....................... 68
Figure 4.6. Statistically non-significant differences in hippocampal gene expression ........ 69
Figure 5.1. Performance on the variable-interval delayed alternation (VIDA) test ............ 94
Figure 5.2. Hippocampal gene expression ................................................................................... 97
Figure 5.3. Insulin-stimulated, hippocampal abundance of insulin signaling proteins ............... 99
Figure 5.4. Correlations between plasma biomarkers and hippocampal-dependent memory ... 100
Figure 5.5. Correlations between hippocampal p-Akt and hippocampal-dependent memory . 102
viii
List of Abbreviations
3MS, Modified Mini-Mental State Examination
Aβ, amyloid-beta peptide
AβO, amyloid-beta peptide oligomer
ANOVA, Analysis of variance
AD, Alzheimer’s disease
APP, Alzheimer precursor protein
B, parameter estimate
BBB, blood-brain barrier
BDNF, brain-derived neurotrophic factor
CI, confidence interval
CON, control diet
CNS, central nervous system
CPS, composite plasma score
DHA, docosahexaenoic acid
DIO, diet-induced obesity
FOXO, Forkhead box
GFAP, glial fibrillary acidic protein
GSK3A, glycogen synthase kinase-3 alpha
HFD, high fat diet
ICV, intracerebroventricular
ix
IGF-1, insulin-like growth factor-1
IL-1β, interleukin-1 beta
ITI, inter-trial interval
IR, insulin resistance
IRc, insulin receptor
IRI, insulin resistance index
IRS1, insulin receptor substrate-1
LTP, long-term potentiation
MAPK1, mitogen activated protein kinsase-1
MCP-1, monocyte chemoattractant protein-1
n-3, omega-3
NMTS, non-matching-to-sample
MetSyn, metabolic syndrome
p-Akt, phosphorylated Akt
PP, plant polyphenols
PI3k, phosphotidylinositol-3 kinase
PIK3R1, phosphatidylinositol 3-kinase, p85α regulatory subunit
PPARγ, Peroxisome proliferator-activated receptor gamma
PTP1B, protein-tyrosine phosphatase 1B
ROS, reactive oxygen species
ROSI, rosiglitazone
SEM, standard error of the mean x
SEP, socioeconomic position
SNP, single nucleotide polymorphism
SOCS3, suppressor of cytokine signaling-3
T2DM, type-2 diabetes mellitus
Tg, transgene
TNFA, tumor necrosis factor-alpha
TZD, thiazolidinedione
VIDA, variable-interval delayed alternation
WFD, whole-food diet
Wt, wildtype
xi
1
1 Chapter 1: General Introduction
A number of prospective observational studies suggest that the dietary profile benefiting
retention of cognitive function with aging contains weekly servings of fish and multiple daily
servings of darkly or brightly colored fruits, and vegetables. Alternatively, these and other
studies find that decreased cognitive function associates with diets relatively high in saturated
fat.
How each dietary profile is contributing to metabolism and cognitive function is debatable. For
instance, darkly or brightly coloured fruits and vegetables are sources of polyphenolic
compounds which may have direct effects on cell signaling, growth, and differentiation in
addition to being the most abundant and powerful dietary antioxidants. Although not well
understood, plant polyphenols as a group may play a large role in findings of better cognitive
outcomes with consumption of plant matter (fruits/vegetables), which is a view supported by
findings that food sources of antioxidant vitamins are associated with cognitive benefits while
pill-forms are not. Certainly, studies in animal models of aging and Alzheimer’s disease (AD)
have found that dietary polyphenols reduce markers of neuroinflammation and stimulate the
activity of phosphotidylinositol-3 kinase (PI3k)—a ubiquitous enzyme involved in many cellular
responses with particular importance to insulin signaling—resulting in improved neuronal
survival and memory. Similarly, omega-3 (n-3) fatty acids, in addition to their acknowledged
role in supporting membrane-bound protein functions and neurotransmisson, may have anti-
inflammatory and pro-survival capabilities by modulating cytokine activity, neurotrophin
expression, and anti-apoptotic pathways, including those influenced by PI3k. Animal studies
find that diets low in n-3 fat compromise cognitive function, promote oxidative stress, and
increase neuropathology. These roles are complemented by epidemiologic data showing that
intake of n-3 fatty acids, particularly those from fish and fish oil, reduces risk for dementia and
cognitive decline.
Alternatively, feeding animals diets high in saturated fat leads to cognitive impairment, reduced
neurotrophin levels, and increased AD-related pathology—mirroring epidemiologic findings.
2
Human food consumption patterns indicate that diets high in saturated fat usually contain lower
amounts of plant matter, and place individuals at increased risk for obesity. Observational
studies indicating positive or negative effects of dietary fat on cognitive decline also find that
such influences are modified by the relative abundance of different fatty acids so that negative
effects of saturated fat are most evident in the context of low n-3 intake and vice versa. These
results could indicate that foods contained in “bad” and “good” dietary profiles are somewhat
exclusive. Interestingly, those dietary patterns related to the prevention and promotion of type-2
diabetes are similar to those patterns modulating risk of dementia and cognitive decline. This,
and other more direct evidence, suggests that changes in insulin sensitivity are related to loss of
brain insulin activity and impaired function.
The strength and challenge of my research relates to the fact that it addresses the less studied role
that global dietary quality, rather than single nutrients or nutrient classes, may play on cognition.
I hypothesized that intake of certain dietary components (fish, fruits and vegetables; saturated
fat) may represent adherence to a broader set of somewhat exclusive dietary patterns that
modulate cognitive function, at least in part, through the interrelated mechanisms of brain
insulin-signaling and inflammation. This thesis focuses on examining the possible mechanistic
relationship between diet-induced changes in molecular components of the brain insulin
signaling pathway, neuroinflammation, and cognitive performance in rodents. It also examines
whether the foods or nutrients hypothesized to influence these mechanisms belong to a broader
set of dietary patterns in older adults. This novel and broad approach is desirable because studies
examining effects of individual nutrients do not model human consumption patterns and
potentially lead to the erroneous expectation that supplementation with single source nutrients
are sufficient to offset the negative attributes of an overall poor diet. Additionally, by focusing
on only one potential biologic mediator, the integrated effects of biologic mechanisms are easily
overlooked. Importantly, diet-associated diseases like cardiovascular disease and type-2
diabetes, which increase risk of cognitive decline and dementia, are neither isolated in their
etiology nor confined to single biologic systems, while the benefits of more ‘healthy’ eating
patterns can target an equally broad spectrum of biologic supports.
3
This thesis will feature an overview of the literature and theoretical framework that led to the
development of the specific objectives and hypotheses (Chapter 2). This chapter was adapted
from a published book chapter and review article representing the state of the literature in 2006-
2007 when the thesis was envisioned. Further updates from more recent studies are included in
the subsequent three chapters (Chapters 3-5) presenting the observational and experimental data,
as well as, in the final chapter which provides a general discussion of the thesis results and
specific conclusions (Chapter 6).
4
2 Chapter 2: Background Literature & Theoretical Overview
The following chapter was adapted with permission from:
Parrott MD, Greenwood CE. Is there a role for nutrition in cognitive
rehabilitation? In Cognitive Neurorehabilitation, 2nd Edition; eds Winocur G,
Robertson I, Stuss D. New York: Cambridge University Press, 2008. Reprinted
with permission
Material on these pages is copyright Cambridge University Press or reproduced with
permission from other copyright owners. It may be downloaded and printed for personal
reference, but not otherwise copied, altered in any way or transmitted to others (unless
explicitly stated otherwise) without the written permission of Cambridge University
Press. Hypertext links to other Web locations are for the convenience of users and do not
constitute any endorsement or authorisation by Cambridge University Press
Parrott MD, Greenwood CE. Dietary influences on cognitive function with aging:
from high-fat diets to healthful eating. Ann N Y Acad Sci 2007;1114:389-97.
Copyright © 2007 John Wiley Sons, Inc. Reprinted by permission John Wiley
Sons, Inc.
5
The maintenance of cognitive function with aging is a current and growing concern. While
pharmaceutical and therapeutic approaches continue to be actively researched, there is increasing
evidence indicating a role for lifestyle factors in successful brain aging. One such factor is diet,
or a person’s usual pattern of eating and drinking, which is proposed to exert both beneficial and
detrimental influences. Our early studies in rats were amongst the first to demonstrate adverse
cognitive effects of chronic consumption of high fat diets in young adult animals, especially diets
high in saturated fat [1]. In the intervening years, animal and human epidemiological studies
have both confirmed the detrimental role of high saturated fat consumption and identified
positive attributes of more healthful eating patterns. Interestingly, dietary profiles thought to
benefit or impair cognitive function correspond to those profiles exerting similar effects on risk
for many chronic diseases which are not thought to have primary cognitive involvements (i.e.
diabetes, coronary heart disease). As will be discussed, this relationship may result from shared
dietary influences on the overlapping physiologic mechanisms responsible for disease risk and
altered cognition with aging with a specific focus on brain insulin signaling and
neuroinflammation.
2.1 Neuropathologic events associated with disruptions to brain
insulin, insulin-mediated cell signaling, inflammation and
oxidative stress
It is increasingly recognised that disruptions in brain insulin mediated cell signaling, apparent in
those with insulin resistance (IR) or type-2 diabetes mellitus (T2DM), result in brain insults
leading to cognitive decline and neuropathologic progression of Alzheimer’s disease (AD) [2-4].
This section will review data demonstrating that:
• Brain insulin signaling is involved in processes underlying memory
• Disruption of brain insulin signaling is associated with development of neuropathologic
hallmarks of Alzheimer’s disease, including accumulation of amyloid-beta peptide (Aβ)
6
involved in plaque formation and production of hyperphosphorylated tau proteins involved in
neurofibrillary tangle formation.
• Loss of insulin signaling can interfere with the action of neurotrophins, and facilitate
neuroinflammatory processes which collectively impede synaptic plasticity and contribute to
neuronal death.
2.1.1 Role of insulin signaling in memory processing
From a mechanistic perspective, there is compelling evidence that brain insulin signalling is
essential for memory processing, including the:
• localization of the insulin receptor (IRc) to key brain regions including the frontal and
cerebral cortices, hippocampus and medial temporal lobe [3,5,6];
• evidence for in situ insulin synthesis [7-9];
• identification of the major signalling pathways including phosphotidylinositol-3 kinase
(PI3k) and its downstream effector Akt and the cytoplasmic intermediate protein Shc and
their convergence on mitogen activated protein kinase activation [5,10-13];
• their ability to stimulate synthesis of proteins necessary and sufficient for long-term memory
formation [4,14,15];
• their enhancement when animals are exposed to learning paradigms [5,16,17];
• the ability of insulin signalling to modulate long-term potentiation (LTP)—a molecular
model of memory—by regulation of the pre- and postsynaptic synthesis and activity of
neurotransmitters including acetylcholine, gamma-aminobutyric acid, serotonin, dopamine,
and glutamate [4,7,18].
These effects are in addition to the more traditional role of insulin in stimulating cerebral glucose
metabolism in specific brain areas versus total brain glucose uptake—in this sense brain remains
an insulin insensitive organ. For instance, overlapping distributions of insulin, IRc, and insulin
sensitive glucose transporters in the hippocampus provide a platform for insulin-stimulated
7
glucose uptake which is known to improve a wide range of memory functions [3,19]. Consistent
with this molecular role are studies indicating that acute insulin elevations facilitate memory
function when given at optimal doses to rodents [20] and humans provided there is adequate
glucose availability [21,22].
Brain insulin signaling may become increasingly impaired with IR and T2DM since chronic
elevations in plasma insulin downregulate brain insulin transport and IRc expression, such that
the brain experiences an insulin-deficient state [23,24], placing individuals at increased risk for
cognitive decline [25]. Indeed, cognitive deficits are already apparent in those with IR. For
example, we report decrements in delayed verbal memory that associate with measures of IR
[26] and glycemic control [27,28], as well as impaired long-term memory and indication of
disrupted brain insulin signaling in genetically obese and insulin resistant Zucker rats [29].
Many patients with AD show evidence of deficient brain insulin signaling, including lower CSF
insulin levels, and resistance to insulin-mediated memory facilitation, compared with healthy
controls [30,31], and progressive loss of insulin and IRc expression accompanied by loss of
downstream signaling through PI3k/Akt [32,33]. Thus, progressive loss of brain insulin
signaling occurs in tandem with cognitive decline throughout the spectrum of age-associated
memory loss to AD, with many arguing a cause and effect relationship.
2.1.2 Loss of insulin signaling and associated neuropathologic events
While loss of insulin signaling in and of itself could interfere with cognitive function due to
impaired memory processing, other, pathologic events associated with brain insulin deficiency
could have more serious consequences as they relate to neuronal plasticity and survival.
Specifically, brain insulin deficiency leads to increased brain accumulation of Aβ—accumulation
of which is thought to be an initiating event in AD pathogenesis [34]—by decreasing local
degradation [3,35-38] and impairing its brain export [39-42]. Additionally, loss of insulin
mediated PI3k/Akt signalling, increases caspase-3 activity contributing to apoptotic cell death
[43] and production of short cytoplasmic amyloid precursor protein fragments [44-47] which in
turn may increase Aβ generation [45,48,49]. Aβ species may then exert positive feedback on
caspase activity allowing for escalation of Aβ accumulation [50,51] whose cytotoxic properties
8
are linked to widespread disruption of LTP and neurotransmission [34,52]. Loss of insulin-
mediated cell signaling may also contribute to the formation of neurofibrillary tangles, another
hallmark of AD highly correlated with cognitive deterioration [53], since insulin signaling
suppresses enzymes involved in tau hyperphosphorylation [54-56].
2.1.3 Insulin resistance, inflammation & brain function
Recent observational studies have implicated serum levels of pro-inflammatory cytokines with
lower cognitive status, lower nerve conduction velocity and greater cognitive decline among
senior citizens [57-59], even after adjustments for a wide variety of socio-demographic variables.
Low-grade, systemic inflammation, as often found in insulin-resistant individuals [60], is
associated with decline and early cognitive deterioration [61]. Our own studies in adults with
T2DM suggest that those individuals carrying a single nucleotide polymorphism (SNP) which
reduces the expression of tumor necrosis factor-alpha have better delayed verbal memories and
show less loss of its function over 48 weeks compared with those not carrying the SNP [62].
Indeed not only do individuals diagnosed with metabolic syndrome, a condition characterised by
IR, exhibit greater cognitive impairment than healthy controls, but within the metabolic
syndrome group, those participants with the greatest inflammation exhibited the greatest
impairment [63]. The role of peripheral hyperinsulinemia to produce systemic inflammation is
well known [64], and can translate into increased markers of central nervous system (CNS)
inflammation [65]. Interestingly, increased CNS inflammation was positively correlated with
changes in Aβ suggesting that synchronous hyperinsulinemia-induced increases in Aβ and
inflammation may represent an important pathway through which IR promotes both cognitive
deterioration and AD pathology. This is compounded by the facts that Aβ cytotoxicity, through
its pro-oxidant properties [66], may feedback to promote further production [67] and that
inflammatory cytokines downregulate brain expression of Aβ scavengers, which could lead to
increased deposition. Increased pro-inflammatory cytokine levels can downregulate PI3K/Akt in
aged rat brains leading to subsequent caspase activation [68] suggesting that insulin-related
signaling is negatively affected by inflammation ultimately leading to increased neuronal death.
9
2.1.4 Insulin resistance and neurotrophins
The interplay between insulin levels and neurotrophins represents another way in which IR can
negatively impact on brain function. Traditional neurotrophins, including brain-derived
neurotrophic factor (BDNF), like insulin, use tyrosine kinase signal transduction to activate
downstream targets including PI3k and Akt/protein kinase-B [69]. Obese and typically insulin-
resistant mice exhibit reduced BDNF levels [70] which can adversely affect LTP, neuronal
survival and brain plasticity, and consequently memory and learning [69]. Conversely, treatment
of neurons with BDNF promotes PI3K/Akt signaling, and reduces caspase activity, highlighting
the possible importance of neurotrophins in the regulation of insulin signaling and vice-versa
[71]. Plasma hyperinsulinemia also downregulates the transport of insulin-like growth factor-1
(IGF-1) across the blood–brain barrier from the periphery where it is synthesised [72]. Although
not a traditional neurotrophin, IGF-1 rapidly and significantly stimulates the process of
membrane assembly at the axonal growth cone through direct stimulation of the PI3K/Akt
pathway – an effect not shared with BDNF [73,74] and increases neurite sprouting and
outgrowth [75,76]. The high degree of amino acid homology between IGF-1 and insulin allows
for cross-reactivity between their respective membrane receptors, making their signaling
cascades almost indistinguishable [77,78]. Furthermore, like insulin, IGF-1 activity can be
reduced by pro-inflammatory cytokines [79,80]. Thus, IR, and/or loss of functional brain
insulin, may negatively impact on the processes of neuronal growth, plasticity and survival
through its ability to reduce brain levels of neurotrophins like IGF-1 and BDNF. Given recent
findings that learning may be accompanied by hippocampal neurogenesis in adult rodent brains
[81], and the reliance of neurogenesis on similar biologic signals, like IGF-1 and BDNF [69],
this negative pathway has the potential to impair a number of processes recruited during learning
which could ultimately lead to worse cognitive outcomes.
10
2.2 Association of foods and nutrients with cognition: Potential links
to brain insulin signaling and neuroinflammation
A number of large, prospective and cross-sectional observational studies find that the dietary
profile benefiting cognitive function with aging contains weekly servings of fish [82-86] and
multiple daily servings of darkly or brightly coloured fruits and leafy vegetables [87-89].
Alternatively, these and other studies find that decreased cognitive function associates with diets
relatively high in total, trans and saturated fat [90]. Both brain-specific mechanisms and the fact
that inappropriate intake of these nutrients may concomitantly elevate risk for chronic diseases,
such as cardiovascular disease, hypertension [91] and T2DM [92-95], which are themselves
independent risk factors for cognitive decline likely explain their influence. Rather than
concentrate on diet’s recognised role in maintaining vascular health, we will argue that the
presence of diet-induced, peripheral insulin resistance may be accompanied by defective
neuronal insulin signaling with downstream consequences that could be a major modifiable
factor contributing to cognitive deficits. In contrast, a diet including fish, fruits, and vegetables
is linked to preservation and/or protection against many adverse processes which need to
minimized to maintain neuronal health—a prerequisite for retention of cognitive function with
aging. Specifically, this section will outline how:
• Dietary components, especially saturated fat, can lead to loss of insulin signaling and
promotion of inflammation.
• Reduced omega-3 (n-3) fatty acid availability may decrease levels in the brain resulting in
reduced insulin signaling and compromised cognitive function in aging animals; reductions
in dietary n-3 fatty acids promote AD neuropathology; n-3 fatty acids modulate production
of neurotrophins, molecular components of the insulin signaling pathway, and
neuroinflammatory markers.
• Plant polyphenols improve markers of oxidative stress and inflammation; support brain
insulin signaling and other cell survival pathways; improve cognitive function in aged and
AD- engineered animals.
11
2.2.1 Saturated fat
Excess dietary fat, particularly saturated fat, worsens IR in humans while animal studies indicate
that dietary saturated fat intake can actually induce IR [96,97]. Corroborating this assertion are
our studies of high saturated fat feeding that consistently link both the level and type of fat to
cognitive deficits in young adult rats. These deficits are widespread, influencing a number of
cognitive domains, with hippocampally mediated memory functions being the most adversely
affected [98]. High fat feeding also contributes to AD pathology by inducing even higher levels
of Aβ in Tg2576 mice [99], and increasing brain inflammation and decreasing BDNF levels
[70,100,101]. Thus, studies indicate that consumption of high-fat diets adversely influence many
biological parameters including neuronal signaling cascades involved in memory,
neurotransmission, neuronal growth and survival, AD pathology, and neurotrophin activity –
either directly or through promotion of IR – ultimately leading to impairments in behavioral
function. Importantly, in most instances, animal diets in these studies were modeled to provide
dietary fat levels consistent with upper limits of typical North American diets and epidemiologic
data demonstrating relationships between high-fat diets and poorer cognitive function [90] that
reflect human consumption patterns.
The preceding discussion focused on negative effects of diet on brain integrity and function.
However, there is especially compelling evidence that two nutrients – omega-3 fatty acids and
plant-based antioxidants – are especially beneficial to cognition and may act to influence brain
insulin signaling and neuroinflammation.
2.2.2 Omega-3 fatty acids and fish oils
Although dietary n-3 fatty acids appear protective against AD and cognitive decline in large,
prospective studies, this effect is somewhat specific to docosahexaenoic acid (DHA) [82,85,86].
Docosahexaenoic acid is a long-chain (22 carbons, 6 double bonds) n-3 fatty acid found in
especially high amounts in fish oil. Intake changes even in late-life alter brain fatty acid profiles
and behavioral outcomes [98,102]. It has been speculated that humans evolved consuming a diet
containing up to 15 times the proportion of n-3 fatty acids found in the present North American
12
diet [103,104], making many argue that our current diet is relatively deficient in them and that
this deficiency may contribute to a number of brain-related disorders [105]. Since mammals
cannot directly synthesise DHA, it must be obtained pre-formed from the diet or from other
dietary n-3 fatty acids that are inefficiently converted to DHA (1–6%) through a series of
enzymatic steps, setting the stage for inadequate brain supply of DHA just as local turnover
increases and conversion decreases with aging [102,105].
Docosahexaenoic acid is comparatively enriched in the brain, where it can be synthesised from
precursors in astrocytes and concentrated in neurons [106,107]. The structural predominance of
DHA in the brain may be linked to its functional importance since changes in availability and
content influence neural membrane-bound enzyme and ion channel activities, membrane fluidity,
LTP, and neurotransmitter release [105,108]. While there is considerable evidence indicating a
developmental role for n-3 fats, dietary deficiency, or reduction, even in old and adult animals
impairs learning and readily depletes neuronal membrane content [109,110]. Furthermore,
dietary supplementation with DHA readily improves membrane content and neurotransmitter
receptors that were adversely affected by dietary depletion [102,111]. In the Tg2576 mouse
model of AD, reductions in dietary DHA led to decreased brain levels and adversely impacted on
Aβ deposition, plaque load, dendritic spine formation and synaptic loss, brain protein oxidation
and neurotransmitter receptors [112-114]. In the one study addressing functional outcomes,
impaired performance in the Morris Water Maze following DHA depletion was prevented by a
DHA replete diet [113]. These, and other studies [115], consistently link loss of brain DHA with
reduced PI3k/ Akt activity resulting in downstream caspase activation and neuronal apoptosis in
both wildtype and AD-engineered animals. A link between neuronal survival and DHA is further
supported by its ability to promote neurite growth in culture [116]. Furthermore, DHA induces
anti-apoptotic and neuroprotective, anti-inflammatory gene-expression programs in the brain
through conversion to neuroprotectin D1. This less well-known role for DHA was subsequently
shown to protect neurons from Aβ- induced neurotoxicity which is highly dependent on
promotion of oxidative stress/inflammation [117,118]. Dietary n-3 fatty acid enrichment also
attenuates inflammatory responses by shifting production of local inflammatory mediators in the
13
brain such as prostaglandins from pro- to anti-inflammatory forms [119], and suppresses adverse
age-related changes in cortical interleukin and PI3k activity [68].
Taken together, studies indicate that DHA deficiency and IR can work through shared
mechanisms to impair neuronal health and survival by interfering with PI3k/Akt signaling and
promoting oxidative stress and neuroinflammation. Importantly dietary DHA repletion can
improve IRc signaling in peripheral tissues of insulin resistant individuals [120], and prevent the
development of IR by high fat feeding in animals [121]. These mechanisms work in concert
with DHA effects on general membrane function, neurotransmitters and neuronal survival and
growth.
2.2.3 Dietary antioxidants of plant origin
Epidemiological evidence indicates that consumption of dietary antioxidants, particularly those
derived from food versus supplemental sources, decrease risk of AD and cognitive decline
[88,122-125]. Dietary antioxidants scavenge the reactive oxygen species (ROS) responsible for
oxidative damage which can induce production of pro-inflammatory cytokines, and regulate Aβ
cytotoxicity [126,127]. The most abundant dietary antioxidants are the polyphenols. Their total
dietary intake could be as high as 1 g/day which is substantively higher than that of all other
classes of known antioxidants. Although ubiquitous in most plant foods, their main sources are
darkly or brightly coloured fruits, vegetables, and plant-derived beverages such as tea [128].
Polyphenol intake is inversely related to the incidence of chronic diseases including coronary
heart disease, diabetes, and cancer [129,130]. There is also growing evidence that these potent
antioxidants and cell-signaling effectors also play a protective role in the brain. Polyphenols not
only improve the status of different oxidative stress biomarkers [131], but may also directly
modulate enzymes involved in signal transduction resulting in modification of redox status of the
cell, and activation of survival pathways [128]. For example, green tea polyphenols directly
influence many signaling pathways including PI3k/Akt [132,133] independent of their
antioxidant roles, resulting in reduced Aβ fibril formation, soluble Aβ release, and potent radical-
scavenging/anti-inflammatory properties. Similarly, blueberry polyphenols exert high
antioxidant capacity, and blueberry-polyphenol enriched rodent diets consistently prevent age-
14
related deficits in learning and memory by decreasing brain ROS levels and altering neuronal
signalling [134,135]. In these studies greater cognitive benefits have been associated with
consumption of foods with higher antioxidant capacity, such as blueberries, compared to those
with lower levels of antioxidants, including spinach and strawberry. Blueberry extract prevented
cognitive decline in animal models of AD and increased extracellular-signal regulated kinases
and protein kinase-C activity—both of which are also regulated by insulin and downstream of
PI3k [136]. In normal, aged rats blueberry polyphenols increased neurotrophin levels which
associated with reduced memory errors [137]. Comparable studies in other AD mouse models
supplemented with spice-polyphenol curcumin demonstrated potent reductions in Aβ, plaque
formation, Aβ fibril formation, oxidative stress, and many pro-inflammatory markers including
IL-1β even when administered in aged mice that already possess significant AD pathology
[138,139]. Thus, plant based dietary antioxidants have an important role to play in controlling
brain inflammation, influencing beneficial neuronal signaling and behaviour, as well as, having
positive impacts on limiting pathological neurodegeneration. The mechanisms employed by
polyphenols, once again, indicate the degree of convergence between diet, inflammation, and
insulin signaling as they relate to cognitive function and brain health.
2.3 Summary and Rationale
In summary, diet influences biologic systems intimately involved in supporting cognition.
These systems converge at two critical points: (1) sustainability of insulin and insulin-related cell
signaling and (2) limiting inflammation. The dietary components discussed in this review all
share the ability to influence these processes, and are prominent in the literature as being related
to age-related cognitive changes. The data equally suggest that the global characteristics of a
‘healthful’ diet [140] cannot be attributed to individual nutrients since a ‘healthful’ diet targets
multiple systems which may work synergistically as it relates to cognitive function and that
focusing on only one biologic system, or nutrient, may be insufficient to explain beneficial
effects of the overall diet. For example, plant matter contributes a complex array of antioxidant
compounds which seem important to controlling inflammatory processes that interfere with
proper brain function, as well as, acting to stimulate molecular components of the insulin
15
signaling pathway like PI3k. In addition to long-chain omega-3 fatty acids, fish is a source of
essential micronutrients including selenium, iron, and iodine which may also support cognitive
function. It has been shown that natural mixtures of plant polyphenols seem to provide greater
antioxidant capacity than isolated sources of polyphenols as found in many synthetic
preparations—a property attributed to the recycling of oxidized compounds in natural mixtures
[141-146]. Importantly, the nature of the dietary components shown to be most prominently
related to cognition and possible modulation of insulin signaling and inflammatory pathways
may also represent alternative dietary profiles or patterns. For instance, human food
consumption patterns indicate that diets high in saturated fat usually contain lower amounts of
plant matter, and place individuals at increased obesity risk [147,148] while fish consumption is
associated with higher intake of fruits and vegetables [149]. These results suggest that foods
contained in potentially “good” and “bad” dietary profiles are somewhat exclusive, and that their
combined nutrient exposure may produce synergies that are not apparent when provided as
separately. Thus, observational studies focussing on the cognitive influences of a single food or
nutrient may actually be capturing the influence of a broader pattern, or be confounded by an
overall poor diet. Futhermore, a “mixed” approach in dietary interventions may allow for
simultaneous recruitment of multiple systems associated with neuroprotection while minimizing
the consequences of an overall poor diet. Although not well researched, this view is supported
by findings that a low-fat diet enriched with both DHA and polyphenol extract improved
learning in old rats compared with a diet high in saturated fat. Importantly, benefits were only
seen with the combined use of DHA and polyphenols, but not when either component was
separately provided [150].
2.4 Objectives & Hypotheses
Based on the theoretical framework outlined above, the overall goal of this thesis was to
determine whether dietary components, identified as being related to cognitive function by
observational studies (fish, fruits & vegetables, saturated fat), exert their neurocognitive effects
as a part of a broader set of dietary patterns in older adults, and by influencing markers of
neuroinflammation, peripheral insulin sensitivity and/or brain insulin signaling in rodents.
16
2.4.1 Objectives
The following specific objectives were addressed by a series of observational and rodent studies:
1. To determine whether empirically-derived dietary pattern(s) in older adults are associated
with intake of the dietary components tested in rodents (See Objective 2), and to explore their
possible association(s) with cognitive function (Chapter 3).
2. To determine the effects of diets associated with consumption of saturated fat or fish, fruits,
and vegetables on cognitive function in rodents (Chapters 4 and 5). These dietary
components were targeted based on their prominence in the epidemiological literature, and
potential to influence insulin signaling and neuroinflammation.
3. To determine if diet-induced behavioural changes in rodents were accompanied by changes
in markers of brain-insulin activity and neuroinflammation (Chapters 4 and 5).
4. To directly investigate the role of diet-induced, peripheral insulin resistance in promoting
cognitive deficits and defects in hippocampal insulin sensitivity (Chapter 5).
2.4.2 Hypotheses
In older adults: (A) dietary patterns associated with consumption of the dietary components
targeted in the animal studies would be identified in older adults; and (B) that a dietary pattern
associated with consumption of fish, fruits and vegetables would exhibit a beneficial relationship
with cognitive function whereas another pattern associated with saturated fat intake would
exhibit an adverse relationship (Chapter 3).
In animals, it was expected that a combined whole-food diet consisting of fish, vegetables and
fruit would improve cognitive function in a transgenic mouse model of Alzheimer’s disease
(Chapter 4) whereas a high saturated fat intake would produce memory deficits in rats (Chapter
5).
17
In animals, diet-induced changes in cognitive function would be accompanied by corresponding
patterns of enhanced (Chapter 4) or diminished (Chapter 5) hippocampal insulin-signaling and
peripheral insulin resistance where measured (Chapter 5).
18
3 Chapter 3: Relationship between diet quality and cognition
depends on socioeconomic position in healthy older adults
This chapter is adapted with permission from an article published in The Journal of
Nutrition © 2013 (copyright the American Society for Nutrition). The original
article was published as the following:
Parrott MD, Shatenstein B, Ferland G, Payette H, Morais JA, Belleville S, Kergoat
M-J, Gaudreau P, Greenwood CE. Relationship between diet quality and cognition
depends on socioeconomic position in healthy older adults. J Nutr 2013;143:1767-
1773.
Student’s Contribution: MDP conceived of the analytical plan, conducted the statistical
analyses, and wrote the manuscript. Co-authors were responsible for collecting the data or
supervising the student.
19
3.1 Abstract
Both diet quality and socioeconomic position (SEP) have been linked to age-related cognitive
changes, but there is little understanding of how the socioeconomic context of dietary intake may
shape its cognitive impact. We examined whether equal adherence to ‘prudent’ and ‘Western’
dietary patterns, identified by principle components analysis, was associated with global
cognitive function (Modified Mini-Mental State Examination; 3MS) in independently living
older adults with different SEP (68-84 y; n = 1099). The interaction of dietary pattern adherence
with household income, educational attainment, occupational prestige, and a composite indicator
of SEP combining all three was examined in multiple-adjusted mixed models over 3 years of
follow-up in participants of the NuAge study. Adherence to the prudent pattern (vegetables,
fruits, fish, poultry, and lower-fat dairy) was related to higher 3MS scores at recruitment only in
upper categories of income [B = 0.56 (95% CI = 0.11, 1.01)], education [B = 0.44 (95% CI =
0.080, 0.80)], or composite SEP [B = 0.37 (95% CI = 0.045, 0.70)]. High prudent pattern
adherence was associated with less cognitive decline only in those with low composite SEP [B =
0.25 (95% CI = 0.0094, 0.50)]. Conversely, adherence to the Western pattern (meats, potatoes,
processed foods, and higher-fat dairy) was associated with more cognitive decline [B = -0.23
(95% CI = -0.43, -0.032)] only in those with low educational attainment. In summary, among
individuals with equivalent diet quality, the magnitude and characteristics of the diet-cognition
relationship depended on their socioeconomic circumstances. These results suggest that
interventions promoting retention of cognitive function through improved diet quality would
provide maximum benefit to those with relatively low SEP.
20
3.2 Introduction
The impact of global dietary quality on age-related cognitive change is of growing interest to
investigators attempting to examine dietary patterns rather than individual foods or nutrients
[151]. To date several prospective studies have related dietary patterns, reflecting high diet
quality, to lower rates of cognitive decline and incidence of dementia [152-157]. Socioeconomic
position (SEP)—represented by indicators like income, education and occupation—is an
established determinant of dietary intake such that higher SEP is generally related to better diet
quality [158]. Interestingly, SEP is also associated with differences in cognitive function across
the lifecourse [159-162]. Accumulated animal evidence suggests that both diet [163,164] and
conditions simulating SEP [165,166] modulate neurobiological mechanisms that mediate
changes in brain structure and function as a response to these life experiences—a capacity
broadly referred to as brain plasticity. Accordingly, socioeconomic indicators and nutrient
biomarkers have also been linked to differences in brain morphology and, in some cases, patterns
of activation [167-173]. The interrelated nature of SEP, diet quality and cognition increases the
risk for significant confounding of the diet-cognition relationship in which purported dietary
impacts may be proxy for the ‘true’ influence of SEP [174]. However, there are suggestions of a
more complicated relationship between diet and other aspects of lifestyle as it relates to
cognition. Lifestyle behaviours like diet, physical activity, and social engagement not only
cluster together, but also exert additive effects on cognitive function such that their combined
impact is greater than either separately [175,176]. Animal studies have found that cognitively
stimulating environments, a factor linked to SEP in humans, can augment or mask the
behavioural impacts of dietary interventions [177,178]. In prospective studies the deleterious
association of high sodium intake with accelerated cognitive decline [179] and the protective
influence of a Mediterranean diet on incidence of Alzheimer disease, at least qualitatively,
appear dependent on an individual’s level of physical activity [180]. Collectively these studies
suggest that dietary effects on cognition are dependent on other aspects of an individual’s
lifestyle. Given its potential to influence neurobiological and behavioural outcomes, we
hypothesized that SEP may modulate the impact of diet quality on cognitive function. Therefore
the objective of this study was to examine whether equal adherence to ‘prudent’ and ‘Western’
21
dietary patterns, identified by principle components analysis, was associated with global
cognitive function in older adults with different socioeconomic position.
3.3 Methods
3.3.1 Participants
The Québec Longitudinal Study on Nutrition and Successful Aging (NuAge) is a prospective
cohort study of independent older adults (68-84 years; n=1793) randomly selected from the
Québec health insurance registry; the methodology has been described elsewhere [181].
Eligible, community-dwelling seniors were fluent in French or English, able to walk unassisted
for 100 meters or climb ten stairs without rest, cognitively unimpaired, and free of disabilities in
activities of daily living. Recruitment occurred between December 2003 and April 2005. This
analysis is based on the first three years of follow-up. The NuAge protocol was approved by
research ethics boards at the Institut universitaire de gériatrie de Montréal and the Geriatric
University Institute of Sherbrooke. All participants provided written informed consent.
Non-dietary variables were collected at recruitment during a structured, computer-assisted
interview for use as covariates. Self-reported variables included questionnaires measuring
current physical activity [182], social engagement [183], perceived health status [184], and
depressive symptomatology [185]. Standing height and weight, waist circumference, and seated
blood pressure were directly measured. Presence of hypertension (self-report/medication/blood
pressure > 140/90) and type-2 diabetes (self-report/medication/fasting plasma glucose
concentration ≥ 7.0 mmol/L) were determined. Global cognitive function was evaluated
annually with the Modified Mini-Mental State Examination (3MS) [186]. Validated French and
English versions of all questionnaires were available and administered based on the participant’s
preference.
Individuals were classified as belonging to upper or lower categories of certain socioeconomic
indicators. A binary variable was created based on Goyder and Franks’ scale of occupational
prestige [187] using self-reported National Occupational Classification skill type categories and
22
descriptions of longest serving occupation. Individuals were assigned into upper or lower
occupational prestige categories based on a midpoint split of the rank-order described by
McLaren and Godley [188]. Upper and lower categories of income and educational attainment
were created by splitting the sample at the second tertile of income ($44800) and the median
years-of-education (12 y) respectively. Finally, a composite indicator of SEP was created by
constructing an additive scale combining participants’ income, education and occupational
prestige indicators. This scale ranged from 0-3, where 3 denoted better SEP, and was collapsed
into two categories based on the median score (M=1).
Diet was assessed at entry by a validated semi-quantitative food frequency questionnaire (FFQ)
estimating usual intake of 78 foods or food groups over the previous 12 months [189]. Exposure
to each item was calculated by converting frequency categories to daily servings of a standard
portion. For example, the category “3-5 times per week” was converted to 0.57 servings/day.
Individuals with implausible or incomplete FFQs [190], as well as, those missing information on
income, education, or occupation were screened out of the study population. Individuals with
history of Parkinson’s disease, muscular dystrophy, or stroke were also excluded resulting in a
final sample of 1099 participants of whom 179 were lost to follow-up. The potential impact of
attrition was addressed using the last observation carried-forward (LOCF) approach [191].
3.3.2 Statistical analysis
Analyses were conducted using SAS 9.1 (SAS institute, Cary, N.C.). Dietary patterns were
identified by principle components analysis of FFQ-exposures using PROC FACTOR, and were
orthogonally rotated using the “varimax” option. The number of factors retained for subsequent
analyses was determined by considering Eigenvalues (>1), the Scree plot, and interpretability of
the resulting patterns. FFQ-exposures with factor loadings greater than 0.155, representing the
critical value for a correlation based on the sample size (n = 1099), were considered as
significant contributors. Dietary pattern scores indicating adherence of a participant to each
retained dietary pattern were obtained for subsequent use in multi-adjusted models. Linear trends
in the association between quintile categories of dietary pattern scores and selected covariates
23
and nutrient intakes were identified using general linear models for continuous variables, or the
Maentel-Haentszel chi square statistic for categorical variables.
The association between dietary patterns and cognitive function was assessed using multiple-
adjusted mixed models with a random intercept and an unstructured covariance structure in
PROC MIXED. Time was coded as a continuous variable expressed as years since study entry.
Covariates included in the adjusted model were identified as being associated with both diet
pattern and 3MS scores during pre-screening, or were considered important based on the
literature. Like others, covariates were included with their interactions with time in order to
account for potential effects on overall cognitive function and the rate of decline [88,153].
Energy intake was included in the adjusted model so that the results could be interpreted as being
independent of the absolute amount of food consumed. The parameter estimate (B) for dietary
pattern score (‘diet’) represented the association of dietary patterns with mean, or overall,
cognitive performance throughout the follow-up period, and the estimate for a ‘diet x time’
interaction represented the association with rate of cognitive decline over the follow-up period
(i.e. slope). Positive estimates indicated that greater adherence to a dietary pattern was
associated with greater mean cognitive performance or less decline during follow-up. To
examine whether the impact of diet quality on cognition was dependent on SEP, each indicator
of SEP was tested for an interaction with the dietary main effects (‘diet x indicator’ and ‘diet x
indicator x time’). In order to understand the underlying relationships, statistically significant
interactions were decomposed by testing for the dietary main effects in the upper and lower
categories of the implicated socioeconomic indicators. For instance, significant ‘diet x indicator’
interactions led to examination of the association of dietary pattern adherence with overall
performance in the implicated socioeconomic sub-group. Significant ‘diet x indicator x time’
interactions triggered similar tests for dietary impacts on overall performance, cognitive decline,
and performance at recruitment in the implicated socioeconomic sub-group. Values in the text
are (mean ± SD). Statistical significance was set at the P < 0.05 level.
24
3.4 Results
3.4.1 Participant characteristics
There were 1099 participants at recruitment of whom 74, 31, and 74 were lost to follow-up at the
second, third, and fourth annual visits respectively. Mean, unadjusted 3MS scores at baseline
and each subsequent annual visit were as follows: 94.0 ± 4.3 (n = 1096), 93.0 ± 5.5 (n = 997),
93.0 ± 6.0 (n = 983), 92.6 ± 6.2 (n = 920). The unanalyzed group of participants excluded from
the parent study (n = 694) differed from the analytic sample (n = 1099) at recruitment in that it
contained more females, exhibited lower global cognitive function, had lower household income,
and attained fewer years-of-education (Table 3.1).
25
Table 3.1. Characteristics of the analyzed and unanalyzed participants
Variable Analyzed (n = 1099) Unanalyzed (n = 694) P
3MS score 94.0 ± 4.3 (99.7)1 93.4 ± 4.5 (99.6) <0.01
Female, % 49.4 (100) 53.3 (99.6) <0.01
Age, y 74.1 ± 4.1 (100) 74.9 ± 4.2 (99.7) <0.01
Household income, $ 39707 ± 22502 (100) 34832 ± 20335 (65.0) <.001
Education, y 11.9 ± 4.6 (100) 11.2 ± 4.3 (99.7) <0.01
Smoking, pack-years 15.1 ± 25.6 (98.1) 12.1 ± 23.9 (98.4) 0.015
PASE score 104.1 ± 52.2 (98.6) 94.1 ± 50.4 (97.8) <0.01
BMI, kg/m2 27.9 ± 4.7 (99.6) 27.9 ± 4.6 (98.7) 0.78
Hypertension, % 60.1 (100) 61.3 (99.7) 0.63
Type-2 diabetes, % 13.4 (100) 13.6 (99.7) 0.90
1 mean ± SD (% reported). P-values are for differences from general linear models for
continuous variables, and the Maentel-Haentszel chi square statistic for categorical variables.
3MS, Modified Mini-Mental State Examination; PASE, Physical Activity Scale for the
Elderly.
26
3.4.2 Dietary patterns
Three patterns with Eigenvalues greater than one remained after examination of the Scree plot.
The first two patterns with the highest Eigenvalues, accounting for 5.5% and 4.9% of the total
variance, were retained for rotation because they were the most interpretable and distinct (Table
3.2). The first, termed the “prudent pattern”, was associated with intakes of vegetables, fruits,
fatty fish, lower-fat dairy products, poultry, and legumes. The second, termed the “Western
pattern”, was associated with intakes of beef, potatoes, white bread, baked goods, processed
meats, higher-fat dairy products, and salty snacks.
Comparison of dietary intakes across quintiles of dietary pattern score revealed linear trends that
generally confirmed expectations based on the profile of foods comprising each pattern (Table
3.3). For instance, higher quintiles of prudent pattern score were associated with greater energy-
adjusted intakes of dietary fibre, vitamin C, and a higher polyunsaturated:saturated fatty acid
ratio whereas the opposite was true of the Western pattern. Higher quintiles of prudent pattern
score were generally associated with better indications of health and socioeconomic position
whereas the opposite was true of the Western pattern (Table 3.3).
3.4.3 Final models
When tested as main effects higher adherence to the prudent pattern was related to better overall
cognitive performance, but was not associated with cognitive decline (Table 3.4). Education,
income, and composite SEP were significant effect-modifiers (Table 3.4). After decomposing
these interactions it was revealed that adherence to the prudent pattern was related to higher 3MS
scores at recruitment in the upper category of each indicator [Education: B = 0.44 (95% CI =
0.080, 0.80); Income: B = 0.56 (95% CI = 0.11, 1.01); Composite SEP: B = 0.37 (95% CI =
0.045, 0.70)] (Table 3.5). Furthermore, high prudent pattern adherence was associated with less
cognitive decline in those with low composite SEP [B = 0.25 (95% CI = 0.0094, 0.50)] (Table
3.5). These interactions are presented graphically in Figure 3.1.
Greater adherence to the Western pattern was related to worse overall cognitive performance, but
27
was not related to cognitive decline when tested as a main effect (Table 3.4). Education was a
significant effect-modifier in the adjusted models (Table 3.4). After decomposing these
interactions it was revealed that adherence to the Western pattern was related to worse overall
performance [B = -1.06 (95% CI = -1.65, -0.48)] and more cognitive decline [B = -0.23 (95% CI
= -0.43, -0.032)] only in the lower education group (Table 3.5). This interaction is presented
graphically in Figure 3.2.
After running the mixed models using the last observation carried forward method, the “diet x
composite SEP” interaction for the prudent pattern was no longer statistically significant.
Therefore, this interaction was considered an artifact resulting from loss to follow-up, and was
not explored further. There were no other material changes to the results displayed in Table 3.4.
The significant interactions seen in Table 3.4 were confirmed after converting diet quality into a
categorical variable by classifying individuals into upper or lower categories of adherence to
each dietary pattern based on the median dietary pattern score (data not shown). The mean
dietary pattern scores in upper and lower categories, and their associated variances, were not
significantly different between socioeconomic groups (data not shown). In the case of the
Western pattern, the diet-education interaction was confirmed after these conditions were met by
excluding the top 5% of diet scores in the lower education group. Therefore, it was found that
individuals with equivalent adherence to each dietary pattern were present within each high or
low socioeconomic subgroup, and that the interactions in Table 2 did not merely reflect
socioeconomic gradients in food selection.
28
Table 3.2. Factor loadings for dietary patterns1
Food Frequency Questionnaire-item Prudent Western
Green, leafy vegetables 0.616 --
Cruciferous vegetables 0.585 --
Green, red, yellow sweet peppers 0.535 --
Carrots 0.473 --
Other vegetables 0.459 --
Other fruits 0.441 --
Salad dressings, mayonnaise dips 0.433 --
Tomatoes 0.428 --
Green/yellow beans, green peas, corn 0.385 --
Apples, pears 0.382 --
Salmon, trout, sardines, herring, tuna 0.357 --
Berries 0.338 --
Yogurt 0.369 -0.244
Citrus fruit 0.324 --
Nuts, peanuts, other seeds 0.323 --
Bananas 0.316 --
29
Melons 0.282 -0.204
Beans, peas, lentils, hummus, beans with pork 0.273 --
Tofu, foods with soya or vegetable protein 0.269 -0.174
Rice, rice noodles, couscous 0.268 --
Poultry 0.263 --
Skim milk 0.254 --
Tomato or vegetable soups 0.225 0.185
Other fish 0.211 --
Cheeses 0.207 --
Seafood 0.203 --
High fibre breakfast cereals 0.202 -0.160
Sunflower seeds 0.157 -0.157
Commercial sliced white bread -0.248 0.489
Beef -- 0.522
Boiled, mashed, or baked potatoes -- 0.492
Sauces (brown, white, BBQ, gravy) -- 0.466
Baked goods (cakes, pies, donuts, pastries) -- 0.453
Sausages, hot dogs -- 0.437
30
French-fries or pan fried potatoes -- 0.416
Ham, cold cuts, smoked meat, bacon -- 0.386
Pork -- 0.369
Milk or cream in coffee/tea -- 0.336
Sugar in coffee/tea -- 0.345
Butter on bread or cooked vegetables -- 0.311
Coffee/tea -- 0.311
Ice cream ice milk, frozen yogurt -- 0.299
Eggs, omelettes, quiches -- 0.290
Regular soft drinks -- 0.273
Pizza -- 0.257
Cookies -- 0.255
Salty snacks (chips, salted crackers, popcorn, pretzels) -- 0.237
Margarine on bread or cooked vegetables -- 0.233
Jam, honey, sweet spreads, maple products -- 0.226
Milk-based desserts, puddings -- 0.223
Candies, chocolate -- 0.213
Sugar added to cereal -- 0.192
31
Liver, other organ meats 0.189 0.190
Fruit drinks with added sugar -- 0.188
Pasta with tomato sauce -- 0.185
Other soups -- 0.179
Pasta with cream sauce -- 0.178
Soya drinks -- -0.164
Beer -- 0.163
1Loadings (>0.155) represent correlation of a Food Frequency Questionnaire-item
with diet pattern score.
32
Table 3.3. Baseline characteristics of the NuAge study participants across quintiles of diet
pattern score
Variable Q1 Q2 Q3 Q4 Q5 P-trend1
Prudent pattern
N 218 217 211 235 218 --
3MS score 93.2 ± 4.42 93.7 ± 4.4 93.5 ± 4.3 94.6 ± 4.3 94.9 ± 4.1 <0.01
Female, % 34.4 50.7 50.7 51.9 59.2 <0.01
Age, y 74.2 ± 4.0 74.4 ± 4.2 74.5 ± 4.1 73.2 ± 4.1 74.2 ± 4.3 <0.01
Income, %3 30.3 28.1 28.4 41.7 41.7 <0.01
Education, % 40.8 46.1 47.9 56.2 59.2 <0.01
Occupation, % 38.1 41.5 45.5 53.2 55.0 <0.01
Composite SEP, % 58.3 62.2 65.4 72.3 75.2 <0.01
Smoking, pack-y 19.8 ± 30.6 15.4 ± 24.3 12.5 ± 21.7 13.5 ± 24.2 14.5 ± 26.2 0.037
PASE score 106 ± 50 104 ± 55 101 ± 52 103 ± 53 107 ± 50 0.78
BMI, kg/m2 29.1 ± 4.5 27.5 ± 4.9 27.5 ± 4.4 27.8 ± 4.7 27.5 ± 4.7 <0.01
WC, cm 100.2 ± 12.5 95.0 ± 13.8 94.4 ± 12.8 94.7 ± 13.0 93.8 ± 12.5 <0.01
Hypertension, % 69.3 61.7 57.3 57.4 55.0 <0.01
Fiber, mg/kcal/d 7.8 ± 2.3 9.1 ± 2.4 9.8 ± 2.4 9.8 ± 2.2 11.2 ± 2.5 <0.01
Vitamin C, μg/kcal/d 54.5 ± 29 70.8 ± 34 80.2 ± 32 82.6 ± 36 101.5 ± 43 <0.01
Folate, μg/kcal/d 0.15 ± 0.03 0.17 ± 0.04 0.19 ± 0.04 0.19 ± 0.04 0.22 ± 0.05 <0.01
33
Vitamin K, ng/kcal/d 49 ± 25 62 ± 27 70 ± 33 73 ± 33 93 ± 46 <0.01
Potassium, mg/kcal/d 1.49 ± 0.30 1.68 ± 0.32 1.78 ± 0.30 1.79 ± 0.33 1.97 ± 0.34 <0.01
SFA, mg/kcal/d 12.9 ± 3.2 12.5 ± 2.9 12.2 ± 2.4 12.4 ± 2.7 11.4 ± 2.2 <0.01
PUFA:SFA 0.61 ± 0.25 0.64 ± 0.26 0.66 ± 0.25 0.65 ± 0.25 0.74 ± 0.31 <0.01
Fatty fish, serving/wk 0.45 ± 0.47 0.66 ± 0.61 0.84 ± 0.77 1.05 ± 1.05 1.54 ± 1.47 <0.01
Leafy greens, serving/d 0.38 ± 0.36 0.74 ± 0.47 1.0 ± 0.64 1.2 ± 0.73 1.9 ± 1.2 <0.01
Berries, serving/d 0.092 ± 0.13 0.12 ± 0.18 0.18 ± 0.21 0.22 ± 0.28 0.37 ± 0.43 <0.01
Western pattern
Sample size 201 225 221 229 223 --
3MS score 94.4 ± 4.5 94.6 ± 4.0 94.1 ± 4.3 93.9 ± 4.4 93.1 ± 4.5 <0.01
Female, % 65.7 60.0 51.6 44.1 27.3 <0.01
Age, y 74.5 ± 4.2 74.0 ± 4.3 74.2 ± 4.2 74.1 ± 4.0 73.9 ± 4.0 0.51
Income, %3 36.3 36.4 34.8 33.2 30.5 0.14
Education, % 59.7 56.0 57.0 43.7 35.4 <0.01
Occupation, % 57.2 52.9 51.6 42.4 30.9 <0.01
Composite SEP, % 73.1 72.0 70.6 63.8 55.2 <0.01
Smoking, pack-y 8.7 ± 18.6 11.3 ± 22.4 17.5 ± 28.4 15.0 ± 23.6 22.5 ± 30.8 <0.01
PASE score 96 ± 47 97 ± 47 104 ± 54 111 ± 53 112 ± 57 <0.01
BMI, kg/m2 27.7 ± 5.4 27.1 ± 4.0 27.8 ± 4.7 28.2 ± 4.7 28.6 ± 4.6 <0.01
34
WC, cm 92.9 ± 13 92.7 ± 12 94.8 ± 13 96.5 ± 13 100.7 ± 13 <0.01
Hypertension, % 53.7 60.0 61.5 66.8 57.8 0.17
Fiber, mg/kcal/d 11.5 ± 2.9 10.1 ± 2.4 9.6 ± 2.2 8.8 ± 2.0 7.9 ± 2.0 <0.01
Vitamin C, μg/kcal/d 96.9 ± 52.7 84.7 ± 35.6 79.6 ± 33.0 69.0 ± 28.9 61.8 ± 29.9 <0.01
Folate, μg/kcal/d 0.21 ± 0.06 0.19 ± 0.05 0.18 ± 0.04 0.18 ± 0.04 0.17 ± 0.04 <0.01
Vitamin K, ng/kcal/d 85 ± 46 73 ± 37 69 ± 30 64 ± 31 56 ± 27 <0.01
Potassium, mg/kcal/d 1.96 ± 0.38 1.80 ± 0.34 1.74 ± 0.33 1.66 ± 0.31 1.55 ± 0.28 <0.01
SFA, mg/kcal/d 10.8 ± 2.7 12.1 ± 2.9 12.4 ± 2.4 12.8 ± 2.6 13.4 ± 2.5 <0.01
PUFA:SFA 0.79 ± 0.4 0.68 ± 0.3 0.62 ± 0.2 0.62 ± 0.2 0.60 ± 0.2 <0.01
Fatty fish, serving/wk 1.12 ± 1.33 0.91 ± 0.84 0.84 ± 0.84 0.91 ± 0.98 0.77 ± 0.98 <0.01
Leafy greens, serving/d 1.26 ± 1.10 1.06 ± 0.90 1.01 ± 0.81 0.97 ± 0.78 0.96 ± 0.87 <0.01
Berries, serving/d 0.23 ± 0.30 0.21 ± 0.29 0.16 ± 0.23 0.19 ± 0.27 0.19 ± 0.32 0.055
1P-value for linear trend. 3MS, Modified Mini-Mental State Examination; PASE, Physical
Activity Scale for the Elderly; PUFA, polyunsaturated fatty acid; SEP, socioeconomic position;
SFA, saturated fatty acid; WC, waist circumference. 2 Values are mean ± SD. 3Percentage in
upper category of each socioeconomic indicator.
35
Table 3.4. Associations between dietary patterns and cognitive function, and interactions with
socioeconomic indicators, in participants of the NuAge study1
Unadjusted
(n = 1099)
Adjusted2
(n = 1001)
B 95% CI P B 95% CI P
Prudent pattern
Diet 0.68 0.39, 0.97 <0.01 0.35 0.031, 0.67 0.032
Diet x time 0.025 -0.067, 0.11 0.63 0.036 -0.077, 0.15 0.53
Diet x education 0.25 -0.31, 0.80 0.38 0.42 -0.11, 0.95 0.12
Diet x education x time 0.18 0.0029, 0.36 0.046 0.21 0.024, 0.40 0.027
Diet x income 0.26 -0.32, 0.85 0.38 0.25 -0.29, 0.79 0.37
Diet x income x time 0.26 0.078, 0.44 <0.01 0.30 0.11, 0.49 <0.01
Diet x occupation 0.41 -0.15, 0.97 0.15 0.33 -0.20, 0.85 0.33
Diet x occupation x time 0.16 -0.022, 0.33 0.085 0.18 -0.0025, 0.37 0.053
Diet x composite SEP 0.34 -0.28, 0.96 0.28 0.61 0.018, 1.19 0.043
Diet x composite SEP x
time
0.34 0.14, 0.54 <0.01 0.37 0.16, 0.58 <0.01
Western pattern
Diet -0.82 -1.10, -0.55 <0.01 -0.55 -0.92, -0.17 <0.01
36
Diet x time -
0.083
-0.17, 0.0021 0.056 -
0.081
-0.21, 0.050 0.23
Diet x education -0.48 -1.02, 0.055 0.078 -0.55 -1.06, -0.033 0.037
Diet x education x time -0.18 -0.36, -
0.0072
0.041 -0.18 -0.36, -
0.0036
0.046
Diet x income -0.15 -0.72, 0.42 0.60 -0.17 -0.69, 0.35 0.52
Diet x income x time -
0.042
-0.22, 0.14 0.64 -
0.045
-0.23, 0.14 0.63
Diet x occupation -0.21 -0.77, 0.34 0.45 -0.29 -0.81, 0.23 0.27
Diet x occupation x time -
0.045
-0.22, 0.13 0.62 -
0.014
-0.19, 0.17 0.88
Diet x composite SEP -0.29 -0.85, 0.27 0.32 -0.45 -0.98, 0.082 0.097
Diet x composite SEP x
time
-0.14 -0.32, 0.042 0.13 -0.15 -0.33, 0.042 0.13
1 B, parameter estimate; CI, confidence interval. Estimates (B) for ‘Diet’ represent change in
mean 3MS score over the entire follow-up period per unit increase of dietary pattern score.
Estimates for ‘Diet x time’ represents the annual change in 3MS score per unit increase in dietary
pattern score. Positive estimates indicated that greater adherence to a dietary pattern was
associated with better overall cognitive function or less decline during follow-up. The estimates
for models testing for interactions with socioeconomic indicators are (Diet x indicator and ‘Diet
x indicator x time’) cannot be interpreted in this manner, but were included to establish which
indicators should be carried forward for socioeconomic sub-group analysis of dietary
relationships with cognitive function.
37
2 Energy intake, age, education, sex, physical activity, medication usage, vitamin supplement
usage, natural health product usage, social engagement, depression, perceived health status,
smoking, waist circumference, body mass index, hypertension, type-2 diabetes, systolic blood
pressure, income, education, occupation and their interactions with time. Terms were not
included twice in the same model.
38
Table 3.5. Associations between dietary patterns and cognitive function within socioeconomic
subgroups1
Stratified group B 95% CI P n
Prudent pattern
Diet at recruitment
Low education -0.019 -0.49, 0.45 0.94 505
High education 0.44 0.080, 0.80 0.017 493
Low income 0.058 -0.32, 0.44 0.19 657
High income 0.56 0.11, 1.01 0.015 341
Low composite SEP -0.010 -0.037, 0.017 0.47 336
High composite SEP 0.37 0.045, 0.70 0.026 662
Diet x time Low composite SEP 0.25 0.0094, 0.50 0.042 337
High composite SEP -0.045 -0.17, 0.079 0.48 664
Western pattern
Diet Low education -1.06 -1.65, -0.48 <0.01 506
High education -0.15 -0.61, 0.31 0.52 495
Diet x time Low education -0.23 -0.43, -0.032 0.023 506
39
High education 0.058 -0.12, 0.23 0.52 495
1 B, parameter estimate; CI, confidence interval; SEP, socioeconomic position. Estimates (B) for
‘Diet’ represent change in mean 3MS score over the entire follow-up period per unit increase of
dietary pattern score. Estimates for ‘Diet x time’ represents the annual change in 3MS score per
unit increase in dietary pattern score. Positive estimates indicated that greater adherence to a
dietary pattern was associated with better overall cognitive function or less decline during
follow-up. Models were adjusted for the same variables as in Table 3.4
40
Figure 3.1. Association between 3MS score and test year stratified by selected socioeconomic
indicators and prudent pattern score
High prudent pattern adherence Low prudent pattern adherence
41
In the upper categories of income, education, and composite SEP (panels B, D, F) performance
at entry (Y1) was significantly better in the highest (solid) versus lowest (dashed) tertile of
prudent pattern score. In the lower categories of income and education (panels A and C), there
was no significant association between prudent score and performance. However, in the lower
category of composite SEP (panel E) the rate of cognitive decline was significantly slower in the
highest (solid) versus lowest (dashed) tertile of prudent pattern score. 3MS, Modified Mini-
Mental State Examination.
42
Figure 3.2. Association between 3MS score and test year stratified by education and Western
pattern score.
In the lower category of education (Panel A) overall performance was significantly lower, and
the rate of cognitive decline was significantly higher, in those with above-median (solid) versus
below-median (dashed) Western pattern scores. There was no significant association between
Western pattern score and cognitive performance in those in the higher category of education
(Panel B). 3MS, Modified Mini-Mental State Examination.
High Western pattern adherence Low Western pattern adherence
43
3.5 Discussion
In this study the magnitude and characteristics of the diet-cognition relationship depended on an
individual’s socioeconomic position (SEP). For instance, cognitive benefits of adherence to a
prudent dietary pattern were seen irrespective of SEP, but differed in their form such that higher
adherence at recruitment was associated with less decline in those with low SEP whereas it was
associated with better performance at entry among those with high SEP. Alternatively, worse
overall performance and more cognitive decline were associated with higher adherence to a
Western dietary pattern at recruitment only in those with relatively low educational attainment.
These interactions were not merely the product of socioeconomic gradients in diet quality as they
reflected cognitive performance of individuals with dissimilar SEP but equivalent diet quality.
Several longitudinal studies have shown that indices of diet quality similar to the prudent pattern
(containing vegetables, fruits, fish) are associated with better cognitive function in older adults
[152,153,155-157,192]. To our knowledge no longitudinal studies of cognition have examined
analogues to the Western pattern (containing meats, processed foods, high fat dairy), but its
association with worse cognition is consistent with similar dietary patterns examined in cross-
sectional studies [174,193]. Like the current analysis, these studies have linked diet quality to
late-life cognition even after adjusting for health behaviours, chronic disease, and
sociodemographics. This study is unique in the number of indicators examined and its finding of
effect-modification. One cross-sectional study determined that education attenuated the
relationship between dietary patterns and cognition by acting as a strong confounder, but did not
find evidence for a diet-education interaction or examine additional socioeconomic indicators
[174].
It is unclear why the prudent pattern interacted with a broader set of indicators than the Western
pattern, or why the interactions were restricted to specific dimensions of cognitive performance
(decline vs. overall vs. entry). A number of factors, including the timing and/or duration of
observation, as well as, the specific nature of dietary and socioeconomic influences on cognitive
function may have contributed to these differences. For instance, the impact of SEP on cognition
44
has been linked to its influence on attaining peak performance in midlife which may impact on
the timing and trajectory of subsequent declines in later life [160]. Further complicating the
picture, some have found that slower cognitive decline in later life has been associated with
upward mobility in SEP [159] while such an association with diet has been shown to require
consistently high diet quality starting at least from midlife [194].
It is useful to study multiple socioeconomic indicators as they may work on cognition through
distinct mechanisms. Income and education have been linked to increased access to health-
enhancing goods, services and knowledge while occupational prestige may reflect psychosocial
dimensions including the stress associated with being part of a discriminated class and levels of
social support [195]. In this study, the composite indicator was used to examine the combined
impact of simultaneously belonging to the most disadvantaged category of each indicator
studied. Our results are consistent with studies finding additive impacts of multiple lifestyle
factors including diet, smoking, and physical activity on cognitive function [175,176]. In these
studies, the proposed impact of lifestyle factors on cognition were unidirectional (i.e. high
physical activity + non-smoking) whereas the current study examined combined impact of
factors where the independent effects on cognition may be in opposition (i.e. high education +
low diet quality). We propose that these behavioural associations are consistent with the concept
of cognitive reserve which postulates that individual differences in lifestyle may allow for more
successful accommodation of age-related brain changes by protecting the amount of neural
substrate or the efficiency of brain networks mediating performance [196]. These differences in
cognitive reserve may arise from the previously suggested impacts of diet quality and SEP on
brain plasticity. For instance, high adherence to the prudent pattern was broadly beneficial
because it promotes neurobiological processes necessary for the development of reserve
capacity. Conversely, the negative impact of the Western pattern could be offset by increased
educational attainment which is considered to be an important proxy of cognitive reserve.
This study may provide insight into situations where SEP and its general influence on diet
quality are mismatched. Such a mismatch is not inconceivable as a reasonably large proportion
of individuals with low diet quality also have relatively high SEP just as relatively high diet
45
quality can be maintained by some individuals with low SEP [158,197]. We hypothesize that
interventions aimed at shifting dietary intake to better resemble the prudent pattern, and less
resemble the Western pattern, would be most beneficial to those with relatively low SEP as they
collectively appear most sensitive to both the detrimental and beneficial impacts of diet quality.
Interpretation of these results is subject to some limitations. For instance, independent older
adults without cognitive impairment at baseline were purposefully selected for participation, and
the impacts of the dietary patterns were observed in a range of 3MS scores greater than those
indicating substantive cognitive impairment. Consequently, these results may not be applicable
to populations with lower levels of cognitive function. Since the 3MS provides a single measure
of global cognition, we could not examine whether the observed associations were restricted to
specific cognitive domains. Although we employed multiple indicators reflecting both
psychosocial and material dimensions of SEP, it could be useful to study additional indicators
across the lifecourse (i.e. wealth, household conditions, parental SEP) as upward mobility in SEP
has been shown to impact cognition [159]. Differences between the analyzed and unanalyzed
participants in income and education may have resulted in overestimated and underestimated
associations of the prudent and Western patterns respectively. The length of follow-up in the
present study was relatively short compared to other studies of the diet-cognition relationship,
but this limitation would be expected to bias towards a null finding as it related to detecting a
relationship with cognitive decline.
In summary, socioeconomic position altered the characteristics and magnitude of the relationship
between diet quality and cognition. Since individuals within the same category of diet quality
performed differently depending on their socioeconomic circumstances, it may be unrealistic to
expect diet to act on cognition in isolation from SEP. These results also suggest that
interventions promoting retention of cognitive function through improved diet quality would
provide maximum benefit to those with relatively low socioeconomic position.
46
3.6 Acknowledgements
All authors read and approved the final version of the paper, and were involved in
conception of the research plan. BS, GF, HP, JAM, SB, MK, and PG collected the data. MDP
analyzed the data and wrote the paper. CEG and MDP had responsibility for the final content.
47
4 Chapter 4: Whole-food diet worsened cognitive
dysfunction in an Alzheimer’s disease mouse model
This chapter is adapted with permission from an article published in Neurobiology
of Aging © 2014 (copyright Elsevier). The original article was published as the
following:
Parrott MD, Winocur G, Bazinet RP, Ma DWL, Greenwood CE. Whole-food diet worsened
cognitive dysfunction in an Alzheimer’s disease mouse model. Neurobiol Aging 2014; e-pub
ahead of print 15 August 2014; doi:10.1016/j.neurobiolaging.2014.08.013
Student’s Contribution: MDP conceived of the research plan; raised the animals and maintained
the breeding colony; designed and administered the experimental diets; collected the tissue;
conducted the gene expression analysis; conducted the statistical analyses; wrote the manuscript.
GW was responsible for cognitive assessments. Amyloid-beta peptide abundance was
determined with the assistance of Dr. Joanne McLaurin and Ms. Mary Hill.
48
4.1 Abstract
Food combinations have been associated with lower incidence of Alzheimer’s disease (AD). We
hypothesized that a combination whole-food diet (WFD) containing freeze-dried fish, vegetables
and fruits would improve cognitive function in TgCRND8 mice by modulating brain insulin-
signaling and neuroinflammation. Cognitive function was assessed by a comprehensive battery
of tasks adapted to the Morris water maze. Unexpectedly, a ‘Diet x Transgene’ interaction was
observed in which transgenic animals fed the WFD exhibited even worse cognitive function than
their transgenic counterparts fed the control diet on tests of spatial memory (P<0.01) and
strategic rule learning (P=0.034). These behavioural deficits coincided with higher hippocampal
gene expression of tumor necrosis factor-α (P=0.013). There were no differences in cortical
amyloid-β peptide species according to diet. These results indicate that a dietary profile
identified from epidemiological studies exacerbated cognitive dysfunction and
neuroinflammation in a mouse model of familial AD. We suggest that normally adaptive
cellular responses to dietary phytochemicals were impaired by amyloid-beta deposition leading
to increased oxidative stress, neuroinflammation, and behavioural deficits.
49
4.2 Introduction
The availability of transgenic animal models of Alzheimer’s disease (AD) has permitted
investigations into the impact of nutrients and other compounds isolated from foods on disease-
related neuropathology, particularly amyloid-beta peptide (Aβ), and behavioural deficits.
Among the most studied compounds are docosahexaenoic acid (DHA), antioxidant vitamins, and
certain dietary phytochemicals which have been shown to be beneficial in many [113,198-202],
but not all cases [203]. These beneficial findings have been attributed to control of oxidative
stress and, despite concerns over lack of direct in vivo evidence [204], neuroinflammation.
However, it is becoming increasingly evident that dietary compounds, especially DHA and
phytochemicals, exert a pleiotropic effect by modulating cell signaling pathways that impact on
additional mechanisms like synaptic plasticity and the enzymatic processing of Aβ [205-208].
Many of these cell signaling changes involve regulators of cellular energy homeostasis and
growth that are also commonly modulated by insulin. Interestingly, impaired neuronal insulin
signaling is a prominent feature of AD that has been linked to disease severity, Aβ deposition,
and degree of cognitive dysfunction [209-212].
Despite the diversity of potential mechanisms and promising results in animal models, clinical
trials have found limited or no benefits of single nutrients like antioxidant vitamins [213,214]
and DHA [215,216] in AD. This mismatch between basic and clinical studies has led some to
speculate that background diet may be an important determinant of intervention success such that
single nutrient supplementation is unlikely to benefit those with replete diets, or conversely, that
supplementation with a single nutrient is unlikely to overcome the negative impacts of overall
low diet quality [164]. The recognition that whole foods provide a wide array of compounds that
may interact to produce synergistic effects has led to interest in the role that food combinations
or diet quality may play in preventing AD with some positive results in the epidemiological
literature [140,192].
The present study is the first in a systematic investigation of the impact of a combined whole-
food diet (WFD) on cognitive function and Aβ deposition in a transgenic mouse model of AD.
50
Given the evidence surrounding their efficacy from epidemiological and basic studies, food
sources of DHA, antioxidant vitamins, and phytochemicals—namely freeze-dried fish, fruits, and
vegetables—were targeted. Since these dietary compounds have been shown to influence brain
insulin signaling and neuroinflammation, cerebral gene expression of these pathways was also
determined. The following tests of learning and memory were administered to measure
cognitive functions known to be affected in AD: (1) spatial memory [217] in which distal
environmental cues are used to find a submerged platform. This test is sensitive to impairment
within the hippocampus, a subcortical structure that is widely implicated in the memory loss
reliably seen early in AD. (2) Non-matching-to-sample (NMTS) which requires that animals
differentiate between sample and test stimuli and select one according to a learned rule. NMTS
and similar rule-learning tasks incorporate conditional and working memory components that are
critical for many types of problem solving under the control of the pre-frontal cortex [218].
These abilities, along with the integrity of the frontal lobes, are increasingly compromised in AD
as the disease progresses. (3) Brightness discrimination learning in which mice must
discriminate between black and white stimuli to find the platform. This task is believed to
depend on the caudate nucleus and related striatal structures [219], a brain system that is affected
in the later stages of AD. Our working hypothesis was that the WFD containing fish, vegetables
and fruits would beneficially influence cognitive performance and Aβ deposition through
modulation of brain insulin signaling and neuroinflammation.
4.3 Methods
4.3.1 Mice and diets
TgCRND8 mice [220] overexpressing mutations in the human APP gene (KM670/671NL,
V717F) and maintained on a mixed C3H/C57 outbred background were obtained courtesy of the
Tanz Centre for Research in Neurodegenerative Diseases. Animals were housed 3-4 per cage (L:
29 cm/W: 18 cm/H: 12 cm) in a facility with controlled temperature (21°C), humidity (40%), and
light cycle (12 h light/dark). Mice had ad libitium access to either the WFD or control diets from
weaning (aged 3 weeks) until sacrifice at 7 months of age (Harlan Teklad, Madison, WI; See
51
Table 4.1 for detailed composition). The WFD contained skinless, freeze-dried Atlantic salmon
(prepared by Guelph Food Technology Centre, Guelph, ON) and a proprietary mixture of
powdered, freeze-dried vegetables and fruits (BerryGreen®, New Chapter, Brattleboro, VT).
The three most abundant ingredients of this mixture were spinach, blueberries, and cruciferous
vegetables (kale, cabbage, broccoli, brussel sprouts). The total fat content and fatty acid profile
of the freeze-dried salmon was determined by flame-ionized gas chromatography as described
previously [221]. The total phenolic content and oxygen radical absorbance capacity expressed
in millimoles of Trolox equivalents (mmolTE) of the fruit and vegetable mixture were
determined by an independent lab (Brunswick Laboratories, Southborough, MA). Based on
these analyses, the WFD provided 2.46 mg docosahexaenoic acid (0.246% wt/wt), 1.10 mg total
phenolics, and 0.018 mmolTE per gram of diet. The control diet was formulated to have the
same energy density (3.8 kcal/g), macronutrient composition (17% fat/64% carbohydrate/19%
protein per kcal of diet), and fibre content as the WFD. Corn oil acted as the main source of
dietary fat.
At four months of age, mice were transferred to Trent University for cognitive testing. Testing
commenced after a two week acclimatization period, and lasted for an additional 2.5 months.
Animals were sacrificed at 7 months of age by pentobarbital overdose. Following rapid excision,
brains were dissected on a cold surface in PBS, flash frozen in liquid nitrogen, and stored at -
80°C.
52
Table 4.1. Composition of experimental diets
Ingredient (g/kg) Whole-food diet Control diet
Freeze-dried Atlantic salmon 200 --
Corn oil -- 70
Casein 35 200
L-cystine 3 3
Corn starch 386.486 397.486
Maltodextrin 132 132
Sucrose 100 100
Fruit & vegetable powder 60 --
Soybean oil 6 --
Cellulose 30 50
Mineral mix, AIN-93G-MX 35 35
Vitamin mix, AIN-93-VX 10 10
Choline bitartrate 2.5 2.5
tert-Butylhydroquinone 0.014 0.014
53
4.3.2 Cognitive testing
The spatial memory, NMTS, and brightness discrimination tasks were administered in a circular
pool (130 cm diameter and approximately 30 cm high), located in the centre of a room (360 cm x
360 cm). The pool was filled with water rendered opaque by diluted, non-toxic white tempera
paint, to a depth of 18cm, and maintained at room temperature (21o C). An inverted flower pot
(15 cm high by 10 cm in diameter) with a white surface, situated a few cm below the surface of
the water, served as a platform on which the mice could climb to escape the water. A heat lamp
near the pool provided a warm area where mice waited between trials. Throughout testing, the
water was cleaned after each trial and changed every 2 to 3 days.
For the spatial memory and the NMTS tasks, the pool was divided into six zones of
approximately equal size. Swimming patterns of mice were monitored by an overhead video
camera connected to a recorder and data processing system. The system enabled computation of
the time required to find and climb on the platform and the time spent in the platform zone.
Records were kept of the animals’ swimming routes that were used to count errors and are
available on request. For the brightness discrimination task, the pool was fitted with a T-maze
whose walls extended 10 cm above the water surface. The stem of the “T” was 27 cm long. The
horizontal arm was 65 cm long with slats along the walls into which black or white panels were
inserted. The submerged platform was located at the end of the panel designated as the positive
arm.
These tasks are commonly used in our lab to assess the effects of various types of brain
dysfunction on cognitive performance in mice [222] and rats [223]. All testing was conducted by
a single experimenter who was blind to the treatment history.
4.3.2.1 Spatial memory
Initially, mice received two days of orientation training, consisting of 5 trials/day in which mice
were placed individually in the pool and allowed to swim to and climb upon the platform, which
was visible a few cm above the surface of the liquid. The location in which the mice were placed
in the pool and the location of the platform were varied from trial to trial. A trial continued until
54
the mouse mounted the platform with all four paws, or until 120 sec. elapsed. The mouse was
allowed to remain on the platform for 10 sec.; if it failed to find the platform in the allotted time,
it was manually guided to the platform where it was allowed to remain for 10 sec. The mouse
was then removed and placed in a clean cage under the heat lamp to await the next trial. The
mice were run in squads of 4-5, allowing for an interval of 2-3 min between trials.
Spatial memory testing began on Day 3. The platform was now below the surface of the water
and always located in the centre of the north-east zone of the pool. For each trial, the mouse was
placed in the water at the edge of the pool, facing the wall, at a different location. The starting
locations were determined by a semi-random sequence, such that, except for the north-east zone,
each location was used at least once each day. The starting location was never in the north-east
zone. Trial administration was identical to that followed in orientation training, with each trial
continuing until the mouse mounted the platform with all four paws, or until 120 sec. elapsed.
As before, the mouse was allowed to remain on the platform for 10 sec.; if it failed to find the
platform in the allotted time, it was manually guided to the platform for 10 sec. The mouse was
then removed and placed in a clean cage under the heat lamp to await the next trial. Each mouse
received 5 trials/day for 7 consecutive days, following this procedure. On Day 8, the first two
trials were conducted in the usual manner. On the third trial, which served as a probe trial, the
platform was removed and the mice were allowed to swim for 60 sec. The interval between
Trials 2 and the probe test (Trial 3) was the same as for all other trials. Trials 4 and 5 followed
the usual procedure with the submerged platform returned to its location.
Two response measures were recorded for each trial of Days 1 to 7 -- latency and errors. The
latency was the time required to reach and climb onto the platform, measured from when the
mouse was placed in the water. An error was counted each time the mouse entered a zone not
containing the platform, or when the mouse left the zone that contained the platform without
successfully mounting it. If the mouse failed to find the platform within 120 sec., it was given an
error score of 30 for that trial. On the probe trial of Day 8, the time spent in the zone that
normally contained the platform was the measure of interest.
55
4.3.2.2 Non-matching-to-sample (NMTS)
The stimuli for the sample and test trials were black and white cylinders (30 cm long x 3 cm in
diameter), suspended 5 cm above the surface of the water. The position of the cylinders was
controlled manually by the experimenter through a system of pulleys, weights, and wires that ran
inconspicuously outside the perimeter of the pool and along the ceiling.
For each sample trial, the black or white cylinder was suspended above the submerged platform.
During the subsequent test trial, both cylinders were present, but the cylinder that was not
present during the preceding sample trial was suspended over the platform and cued its location.
Thus, if on a given sample trial, the black cylinder signaled the location of the platform then, on
the succeeding test trial, the white cylinder signaled its location. The locations of the cylinder
and platform varied between sample trials. The black or white cylinder was selected as the
sample stimulus for each pair of trials according to a semi-random schedule that ensured that
each cylinder was the sample for 50% of the trials. For each test trial, the platform was moved to
another zone with the non-sample cylinder located directly above it. The sample stimulus was
also moved to a different zone. The zone that contained the submerged platform was changed
after each sample and test trial, according to a random schedule, in order to eliminate the use of
spatial cues. All zones were used equally for locating cues in the sample and test trials and,
within the zones, the platform was positioned randomly.
NMTS testing began 10 days after the completion of the spatial memory test. At the beginning
of each sample trial, the mouse was placed in the pool at the same location (south-east zone),
facing the wall of the pool, and allowed to swim to the submerged platform under the sample
cylinder. The mouse remained on the platform for 20 sec. The mouse was then removed and
placed in a clean cage under the heat lamp while the platform was moved and the cylinders put in
position for the test trial. The organization of the cylinders and platform took about 10 sec. The
mouse was then placed in the pool at the usual location and allowed to swim to the submerged
platform or until 60 sec. had elapsed.
56
The time required to reach and climb onto the platform (latency) was recorded. An error was
counted each time the mouse entered a zone not containing the platform, or when the mouse left
the zone that contained the platform without successfully mounting it. If the mouse failed to find
the platform within 60 sec., it was given an error score of 15 for that trial. In either case, the
mouse was allowed 10 sec. on the platform before being returned to a holding cage under the
heat lamp, to await the next pair of trials. The mice were tested in squads of 4 or 5, which
allowed for an interval of 4 to 5 min. between each pair of trials. Ten daily sessions, each
consisting of 5 pairs of sample and test trials, were administered. Latency and error scores for
each sample and test trial were recorded.
4.3.2.3 Brightness discrimination learning
The discrimination learning task was administered two weeks after the NMTS task. In this test,
mice learned to discriminate between the black and white arms of the T-maze. For half the mice
the black arm was positive with the submerged platform located at the end of that arm; for the
other half the white arm was positive. The position of the panels was determined by a random
schedule. At the beginning of each trial, the mouse was placed in the stem at the edge of the
pool and allowed to find and mount the submerged platform. Each mouse received 20 daily
sessions of 5 trials/day until a criterion of 8 of 10 errorless trials over two consecutive days was
achieved. An error was scored each time a mouse’s entire body entered the incorrect arm and
when a mouse left the correct arm after having entered it. Mice were scored on the number of
trials required to reach criterion with individuals that did not reach criterion by day 20 being
assigned a score of 100.
4.3.3 Genotyping by polymerase chain reaction (PCR)
Tail clippings were digested overnight at 55°C by proteinase K and cell lysis buffer (Cell
Signaling Technology, Danvers, MA). DNA was extracted using phenol followed by
precipitation and washing with ethanol. DNA was resuspended in 25 μL of Tris-EDTA buffer,
and aliquots were amplified by PCR using primers for the APP transgene (5´-
AGAAATGAAGAAACGCCAAGCGCCGTGACT-3´ and 5´-
57
TGTCCAAGATGCAGCAGAACGGCTACGAAAA-3´). Each 25 μL PCR contained 0.3 mM
deoxyribonucleotide mix (Fermentas, R0192), 1.5 mM of magnesium chloride, and 1.25 units of
Platinum Taq DNA polymerase with its buffer (Invitrogen, 10966-018). The thermocycler
program consisted of a period of 3 minutes at 94°C followed by 30 repeated cycles of 20 seconds
at 94°C, 20 seconds at 68°C and 90 seconds at 72°C. The program completed with 7 minutes at
72°C. The amplified PCR products were run through a 2% agarose gel containing SYBR safe
DNA Gel Stain (Invitrogen, S33102), and visualized using the Fluorochem image system (Model
8000, Alpha Innotech Corp., San Leandro, CA).
4.3.4 Hippocampal gene expression analysis (Quantitative reverse real-time
PCR)
Total RNA was isolated and purified from hippocampi using the Trizol method (Life
Technologies, Carlsbad, CA) and RNeasy mini kit (Qiagen, Venlo, Netherlands) according to the
manufacturers’ instructions. RNA purity and quantity were assessed by using a NanoDrop 1000
(NanoDrop Technologies, Wilmington, USA) to measure the 260 nm to 280 nm UV absorbance
ratio and 260 nm absorbance respectively. One microgram of total RNA was reverse transcribed
into cDNA using a High-Capacity cDNA Reverse Transcription kit (Life Technologies,
Carlsbad, CA). Quantitative real-time PCR was performed with TaqMan Gene Expression
Master Mix and TaqMan Gene Expression Assays on an ABI Prism 7000 SDS system (Life
Technologies, Carlsbad, CA). Assays were selected to target markers of insulin-signaling and
inflammation (Table 4.2). Relative differences in gene expression were quantified by using the
∆∆CT method to calculate fold-change values between treatments [224]. Wildtype mice on the
control diet acted as the reference group, and expression was normalized by the average CT of
two endogenous control genes (18S, GAPDH).
58
Table 4.2. Genes targeted for reverse transcription quantitative real-time PCR
Target gene (Gene symbol) Assay ID
Eukaryotic 18S rRNA (18S) Hs99999901_g1
Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) Mm99999915_g1
Insulin degrading enzyme (IDE) Mm00473077_m1
Insulin receptor (INSR) Mm00439693_m1
Phosphatidylinositol 3-kinase, p85 alpha regulatory subunit (PIK3R1) Mm00803160_m1
Insulin receptor substrate-1 (IRS1) Mm01278327_m1
Glycogen synthase kinase-3 alpha (GSK3A) Mm01719732_m1
Protein kinase C, alpha (PRKCA) Mm00440858_m1
Mitogen activated protein kinase-1 (MAPK1) Mm00442479_m1
Tumor necrosis factor alpha (TNFA) Mm00443258_m1
Glial fibrillary acidic protein (GFAP) Mm01253033_m1
59
4.3.5 Cortical Aβ burden
Cortices from transgenic mice were homogenized in buffered sucrose solution followed by either
a mixture of 0.4% diethylamine and 100 mM NaCl for soluble Aβ or cold formic acid for the
isolation of total Aβ. After neutralization, diluted samples were analyzed for Aβ40 and 42 using
commercially available sandwich ELISA kits (BioSource, Burlington, ON) according to the
manufacturer’s instructions, and as performed previously [225].
4.3.6 Statistical analysis
The variables analyzed for the spatial memory acquisition, NMTS, and discrimination learning
tests were the latency (in seconds) and the number of errors exhibited on each trial across all
testing days. Only latencies are presented in the main text as the latency and error scores yielded
the same pattern of results for all tests; error scores are provided in supplementary figures. The
variable of interest for the spatial and cued memory probe trials was the amount of time spent in
the platform zone.
Linear mixed models (PROC MIXED; SAS v9.3) with an autoregressive covariance structure
were used to test differences between groups on behavioural measures that involved repeated
measures. The fixed effect of trials was assessed to indicate whether mice exhibited significant
improvements on the respective tasks with practice. The fixed effects of diet, transgene, and
their interaction were tested as the primary outcomes. Statistically significant (P < 0.05) ‘diet x
transgene’ interactions were decomposed by testing for differences between the four treatment
groups using Tukey’s post-hoc test to adjust for multiple comparisons. A similar approach was
taken when analyzing the probe trial, gene expression, and cortical Aβ data except that the fixed
effects were tested in a general linear model (PROC GLM).
Of the 39 mice transferred for cognitive testing, seven mice prematurely died (6 transgenic/1
wildtype; 5 control diet/2 whole-food diet). Mice were excluded from the statistical analysis of
the cognitive test during which they died. Accelerated mortality is a characteristic of the
TgCRND8 mouse line that is shared by humans with AD [220].
60
4.4 Results
Mice on the WFD were heavier than those on the control diet at death (P = 0.021) (Table 4.3).
Total per capita consumption by mice on the WFD appeared somewhat higher than mice on the
control diet during the same two-week period (31 g/mouse/week vs. 28 g/mouse/week). Group
housing prevented measurement of food intake at an individual level, and therefore, the use of
inferential statistics.
4.4.1 Cognitive function
4.4.1.1 Spatial memory test
During the acquisition stage which reflects spatial learning and memory, there was a significant
main effect of trials (P < 0.01) which showed no interaction with either diet (P = 0.66) or
transgene (P = 0.47), indicating comparable improvement in all groups with practice.
Nevertheless, there was a ‘diet x transgene’ interaction (P < 0.01; Figure 4.1, Panel A) such that
transgenic animals exhibited longer latencies over the testing period than the wildtype mice, but
the Tg-WFD group took even longer than the Tg-Con group to find the platform. There was no
difference in performance according to diet in wildtype animals. A similar pattern of results was
obtained when errors were analyzed (Figure 4.2). During the probe trial test of spatial memory
this interaction was not observed (P = 0.99), and there was no main effect of diet (P = 0.29).
However, transgenic animals spent less time in the platform zone than the wildtype mice (P =
0.024; Figure 4.1, Panel B).
4.4.1.2 Non-matching-to-sample test
There was a significant main effect of trials (P = 0.036) which showed no interaction with either
diet (P = 0.35) or transgene (P = 0.59), indicating comparable improvement in all groups with
practice. A ‘diet x transgene’ interaction (P < 0.01; Figure 4.3) revealed that transgenic animals
exhibited longer latencies over the testing period than the wildtype mice, but the Tg-WFD group
took even longer than the Tg-Con group to find the platform. There was no significant
difference in performance according to diet in wildtype animals. A similar pattern of results was
61
observed when errors were analyzed (Figure 4.4).
4.4.1.3 Brightness discrimination learning
Transgenic mice required more trials to reach criterion than wildtype mice (P = 0.025; Table
4.4). There was no main effect of diet (P = 0.19) or a ‘diet x transgene’ interaction (P = 0.53).
4.4.2 Gene expression
Transgenic animals exhibited higher expression of the astrocyte marker glial fibrillary acidic
protein (GFAP; P < 0.01), as well as lower expression of the insulin-signaling associated genes
mitogen activated protein kinase-1 (MAPK1; P = 0.018) and glycogen synthase kinase-3 alpha
(GSK3A; P = 0.049) (Figure 4.5). A significant ‘diet x transgene’ interaction was observed for
the pro-inflammatory cytokine tumor necrosis factor-alpha (P = 0.013; TNFA) such that the Tg-
WFD group exhibited the higher expression than all other groups which did not significantly
differ from each other (Figure 4.5). No significant differences in expression were observed for
the remaining target genes (Figure 4.6).
4.4.3 Cortical Aβ burden
Among the transgenic animals, there were no significant differences between diet groups on any
of the measures of Aβ burden (Table 4.5).
62
Table 4.3. Body weights of experimental animals
Group Body weight (g)a
Tg-WFD (n = 7) 32.1 ± 2.0
Tg-Con (n = 7) 28.0 ± 2.0
Wt-WFD (n = 9) 35.4 ± 1.8
Wt-Con (n = 9) 30.3 ± 1.8
Abbreviations: Tg, transgenic; WFD, whole-food diet; Wt, wildtype; Con, control diet. a Mice
on WFD were heavier (P = 0.021). There was no significant effect of transgene (P = 0.18) or a
‘diet x transgene’ interaction (P = 0.80). (mean ± SEM)
63
Figure 4.1. Latencies for the spatial memory test acquisition and probe trial performance
During spatial memory acquisition (Panel A), transgenic mice receiving the WFD performed
even worse than transgenic mice on the control diet. There were no differences according to diet
in the wildtype mice. Performance by groups not sharing a letter are statistically different over
the entire testing period. The same pattern of performance is seen when errors are analyzed
(Figure 4.2). In the probe trial (Panel B) transgenic mice performed worse than wildtype. n = 8-
10/group. All data mean ± SEM. (Tg, transgenic; WFD, whole-food diet; Con, control diet; Wt,
wildtype).
B
a
bc
c
A
Days1 2 3 4 5 6 7
Mea
n la
tenc
y (s
)
0
20
40
60
80
100
Tg-WFD Tg-Con Wt-WFD Wt-Con
Tg-WFD Tg-Con Wt-WFD Wt-Con
Tim
e in
pla
tform
zon
e (s
)
0
5
10
15
20 Diet, P = 0.29Transgene, P = 0.024
64
Figure 4.2. Errors for the spatial memory test acquisition
Task performance improved over time (main effect of trials P < 0.01), irrespective of diet and
transgene. There was a significant ‘diet x transgene’ interaction (P < 0.01) such that transgenic
animals made more errors over the testing period than the wildtype mice, but the Tg-WFD group
made even more errors that the Tg-Con group. Performance by groups not sharing a letter are
statistically different over the entire testing period. n = 8-10/group. All data mean ± SEM. (Tg,
transgenic; WFD, whole-food diet; Con, control diet; Wt, wildtype).
Days1 2 3 4 5 6 7
Erro
rs
0
5
10
15
20
25
ab
cc
Tg-WFD Tg-Con Wt-WFD Wt-Con
65
Figure 4.3. Latencies for the non-matching-to-sample test
Transgenic mice receiving the WFD performed even worse than transgenic mice on the control
diet. There were no differences according to diet in the wildtype mice. Performance by groups
not sharing a letter are statistically different over the entire testing period. The same pattern of
performance is seen when errors are analyzed (Figure 4.4). n = 8-10/group. All data mean ±
SEM. (Tg, transgenic; WFD, whole-food diet; Con, control diet; Wt, wildtype).
Days1 2 3 4 5 6 7 8 9 10
Mea
n La
tenc
y (s
)
0
10
20
30
40
50
a
b
cc
Tg-WFD Tg-Con Wt-WFD Wt-Con
66
Figure 4.4. Errors for the non-matching-to-sample test.
Task performance improved over time (main effect of trials P < 0.034), irrespective of diet and
transgene. There was a significant ‘diet x transgene’ interaction (P < 0.01) such that transgenic
animals made more errors over the testing period than the wildtype mice, but the Tg-WFD group
made even more errors than the Tg-Con group. Performance by groups not sharing a letter are
statistically different over the entire testing period. n = 8-10/group. All data mean ± SEM. (Tg,
transgenic; WFD, whole-food diet; Con, control diet; Wt, wildtype).
Days1 2 3 4 5 6 7 8 9 10
Erro
rs
0
2
4
6
8
10
12
14
a
b
cc
Tg-WFD Tg-Con Wt-WFD Wt-Con
67
Table 4.4. Average number of trials required to reach criterion on the brightness discrimination
test
Experimental Group Trials-to-criteriona
Tg-WFD (n = 7) 43.6 ± 11.9
Tg-Con (n = 7) 64.3 ± 10.3
Wt-WFD (n = 9) 26.7 ± 9.3
Wt-Con (n = 9) 35.0 ± 7.8
Abbreviations: Tg, transgenic; WFD, whole-food diet; Wt, wildtype; Con, control diet. a
Transgenic mice required more trials to reach criterion (P = 0.025). There was no significant
effect of diet (P = 0.19) or a ‘diet x transgene’ interaction (P = 0.54). All data mean ± SEM.
68
Figure 4.5. Statistically significant differences in hippocampal gene expression
Hippocampal gene expression of mitogen activated protein kinase-1 (MAPK1; Panel A) and
glycogen synthase kinase-3α (GSK3A; Panel B) were significantly lower in transgenic mice
compared to the wildtype mice. Transgenic mice exhibited higher expression of glial fibrillary
acidic protein (GFAP; Panel C) compared to wildtype. Transgenic mice receiving the WFD
exhibited higher tumor necrosis factor-α expression compared to all other treatment groups
which did not differ from each other (TNFA; Panel D; Bars not sharing a letter are statistically
different). n = 5 independent samples/group. All data mean ± SEM. (Tg, transgenic; WFD,
whole-food diet; Con, control diet; Wt, wildtype).
Tg-WFD Tg-Con Wt-WFD Wt-Con
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Tg-WFD Tg-Con Wt-WFD Wt-Con
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
Tg-WFD Tg-Con Wt-WFD Wt-Con
Fold
cha
nge
in m
RN
A
0.0
1.0
2.0
3.0
4.0
5.0
6.0
Tg-WFD Tg-Con Wt-WFD Wt-Con
Fold
cha
nge
in m
RN
A
0.0
0.5
1.0
1.5
2.0
2.5
3.0
MAPK1
Diet, P = 0.91Transgene, P = 0.018
GSK3A
Diet, P = 0.67Transgene, P = 0.049
Diet, P = 0.24Transgene, P < 0.01
GFAP TNFA
Diet x Transgene, P = 0.013
a
bb
b
A B
C D
69
Figure 4.6. Statistically non-significant differences in hippocampal gene expression
There were no differences in gene expression of insulin receptor (INSR; Panel A), insulin
receptor substrate-1 (IRS1; Panel B), phosphatidylinositol 3-kinase, p85 alpha regulatory subunit
(PIK3R1; Panel C), protein kinase C, alpha (PRKCA; Panel D), and insulin degrading enzyme
(IDE; Panel E). n=5 independent samples/group. All data mean ± SEM. (Tg, transgenic; WFD,
whole-food diet; Con, control diet; Wt, wildtype).
Tg-WFD Tg-Con Wt-WFD Wt-Con
Fold
cha
nge
in m
RN
A
0.0
0.5
1.0
1.5
2.0
2.5
Tg-WFD Tg-Con Wt-WFD Wt-Con
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
Tg-WFD Tg-Con Wt-WFD Wt-Con
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Tg-WFD Tg-Con Wt-WFD Wt-Con
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
Tg-WFD Tg-Con Wt-WFD Wt-Con
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
INSR
Diet, P = 0.53Transgene, P = 0.13
A IRS1
Diet, P = 0.57Transgene, P = 0.12
B
PIK3R1
Diet, P = 0.73Transgene, P = 0.084
C PRKCA
Diet, P = 0.73Transgene, P = 0.18
D
IDE
Diet, P = 0.63Transgene, P = 0.17
E
70
Table 4.5. Cortical Aβ content of transgenic animals by dieta
Aβ species (ng/mg) Whole-food diet Control diet P
Aβ42 soluble 31.30 ± 3.29 38.62 ± 3.47 0.15
Aβ42 insoluble 1319.02 ± 109.52 1236.76 ± 61.16 0.52
Aβ40 soluble 28.61 ± 4.44 37.31 ± 4.79 0.21
Aβ40 insoluble 679.80 ± 98.95 678.77 ± 63.10 0.99
Aβ42:Aβ40 soluble 1.16 ± 0.088 1.07 ± 0.062 0.42
Aβ42:Aβ40 insoluble 2.08 ± 0.21 1.86 ± 0.081 0.35
a P for differences between diet groups. n = 7/grp. mean ± SEM
of wet weight.
71
4.5 Discussion
The impaired performance of the Tg-CON group on the behavioural tasks is consistent with
similar reports of cognitive deficits in this transgenic model of AD [226]. The unexpected
finding was that transgenic mice fed the WFD were even more impaired on the spatial memory
and NMTS tasks. We initially predicted that that the WFD would ameliorate behavioural
deficits by enhancing brain insulin signaling and reducing neuroinflammation. In fact, the
results indicate that an interaction between transgene driven Aβ deposition and the WFD
produced a heightened neuroinflammatory response that coincided with exacerbation of
behavioural deficits. The behavioural tasks selected for this study assessed various aspects of
learning and memory which can be dissociated and linked to different brain regions. The spatial
memory test, as measured in the Morris water maze, is a form of context-dependent reference
memory that depends on the functional integrity of the hippocampus [217]. The NMTS
conditional rule-learning task, because of the inherent strategic and working memory
components, is identified with frontal lobe function [218]. The brightness discrimination task
assesses non-conditional learning that is believed to depend on striatal structures [219]. In AD,
hippocampal, frontal-lobe, and striatal functions are compromised to varying degrees. The
deficits of transgenic mice on all tasks are consistent with this pattern and support the use of this
transgenic strain as a model of AD. Further, the finding that WFD impaired performance on the
spatial memory and NMTS tasks, but not the brightness discrimination task, provides insight into
brain mechanisms that were susceptible to dietary effects.
The question arises as to whether impairment in the ability to detect crucial environmental
stimuli or perform the appropriate swimming behaviour could account for observed deficits on
the various tasks. This interpretation receives some support from the relatively poor early
performance of the transgenic groups on the spatial memory and NMTS tasks. However, this
effect has been observed in previous work with other impaired mouse models [222] and
attributed to initial disorientation related to the animals’ cognitive impairment. Several lines of
evidence in the present study also argue against a performance deficit interpretation. First, the
Tg-CON and the Tg-WFD groups improved similarly to wildtype animals over trials during the
72
acquisition phase, and they subsequently performed similarly on the probe test of the spatial
memory task. This indicates that, although they were impaired in finding the platform’s location
during training, once learned, the Tg-WFD and Tg-CON groups remembered it and had no
trouble using relevant cues to swim in the appropriate zone of the maze. Second, similar
performance of the Tg-WFD and Tg-Con groups on the brightness discrimination task argues
against a deleterious effect of the WFD on sensori-motor function. The present study did not
directly assess performance-related variables and such an investigation may be warranted.
However, on balance, the behavioural and biological evidence points more strongly to a
disruption of cognitive processes in transgenic animals and an exacerbation of this effect by the
WFD on at least some of the tasks.
Group differences in cognitive function did not match differences in body weight, and the WFD
did not seem to adversely influence food intake or body weight. Great care was taken to ensure
that the WFD met established guidelines for nutritional adequacy [227]. It could be argued that
the WFD represents a more ‘natural’ mouse diet compared to most laboratory diets as feral mice
are opportunistic omnivores that, depending on availability, frequently consume vegetation and
large amounts of animal protein [228-230]. Therefore, it seems unlikely that WFD-induced
impairment of cognitive function was related to generalized toxicity or pervasive physical
illness.
Cerebral insulin-signaling and neuroinflammation are considered to play an interconnected role
in AD pathogenesis [209-212], and were important mechanisms of interest in this study.
Although changes in whole-body insulin sensitivity were not assessed, similarities in
macronutrient distribution and energy density make large diet-induced differences in insulin
sensitivity unlikely. We found that transgenic animals exhibited reduced hippocampal gene
expression of MAPK1 which is part of an insulin receptor substrate-1 (IRS-1) independent
insulin signaling pathway that is dysregulated in AD [212,231-233] , and appears critically
important for learning, memory, and synaptic plasticity [234,235]. This finding is consistent
with another study involving TgCRND8 mice that found less hippocampal activation of MAPK1
both in basal conditions and under cholinergic stimulation [236]. We also observed that
73
transgenic animals exhibited lower expression of glycogen synthase kinase-3 alpha (GSK3A)
which has been associated with cognitive impairments in animals [237]. Therefore, aberrant
signaling through the MAPK1 and GSK3A pathways may have contributed to impaired
performance by the transgenic animals on every behavioural test in this study. However, they
did not appear to coincide with exacerbation of these behavioural deficits by the WFD on certain
tasks.
A more robust neuroinflammatory response in animals on the WFD, as determined by greater
expression of an overlapping set of neuroinflammatory genes, better matched the pattern of
exacerbated behavioural deficits. Compared to their wildtype littermates, transgenic animals
exhibited higher expression of glial fibrillary acidic protein (GFAP), a marker of astrocyte
activation, which agrees with studies identifying reactive astrocytes as an early event in the
TgCRND8 mouse [238] that is shared with human AD [239]. High GFAP expression has also
been shown to be inversely related to cognitive function [240], and to coincide with impaired
brain insulin-signaling [33] which agrees with our findings of reduced MAPK1 and GSK3A
expression by the transgenic animals. More importantly, the transgenic animals on the WFD
exhibited the highest TNFA expression in conjunction with the poorest performance on tests of
spatial learning and strategic rule learning. Increased expression of TNFA has been linked to
AD pathogenesis [241] and cognitive dysfunction in AD animal models [242-244]. Thus,
combined elevation of GFAP and TNFA seems to distinguish the Tg-WFD from the Tg-Con
group which only exhibited increased expression of GFAP.
Exacerbated behavioural deficits and elevated TNFA expression did not coincide with
differences in the deposition of soluble or insoluble Aβ species in the transgenic animals. A
number of factors may be responsible for this finding. Firstly, glial activation has been linked to
phagocytic Aβ clearance in AD mouse models overexpressing pro-inflammatory cytokines
including TNFA [245-247]. Interestingly, in AD mouse models improvements in behaviour
associated with TNFA inhibition have occurred both with [244] and without [242] any impacts
on Aβ pathology. These studies suggest that activated glia may limit Aβ deposition even as
production of potentially disruptive substances for behaviour, like TNFA, are elevated.
74
Secondly, we did not measure the abundance of AβOs which may be more proximally related to
producing neuroinflammatory and insulin resistant brain states. Interestingly, behavioural
benefits of dietary antioxidants [248,249], blueberries [250], and an insulin-sensitizing drug
[251] have occurred without any effects on Aβ deposition in animal models.
The combined administration of multiple whole foods precludes conclusions as to which specific
dietary component is responsible for promoting neuroinflammation and behavioural dysfunction
in the TgCRND8 mouse. It is clear that this synergistic effect reflected the adverse interaction of
the WFD with transgene-driven Aβ deposition. The WFD differed from the control diet in its
inclusion of freeze-dried salmon, vegetables, fruits, and smaller amounts of other botanicals. We
propose that Aβ impaired the normally adaptive cellular response to the phytochemicals
contained in this complex mixture. This hypothesis is based on the capacity of many foodborne
phytochemicals to activate cellular stress response pathways, partly due to their direct pro-
oxidant effects, that serve to upregulate endogenous antioxidant defense systems [252,253]. The
benefit of this indirect antioxidant action, or hormetic effect, derives from the ability of low-level
oxidative stress to precondition cells so that they are better prepared when larger insults strike
[254]. However, such benefits assume unimpaired activation of stress response pathways and
downstream antioxidant defense systems. Interestingly, activation of the phytochemical-
sensitive nuclear factor erythroid-2 p45-related factor 2/antioxidant response element
(Nrf2/ARE) stress response pathway has been shown to involve several insulin signaling
molecules [253,255-257] that have been shown to be downregulated in human AD and AD
animal models including in this study [33,209,212,258-260]. Furthermore, damage to small
molecule antioxidants like glutathione and reduced activity of enzymes that participate in certain
cellular antioxidant defenses are seen in human AD and animal models of Aβ deposition [261-
269]. The unintended consequence of combining sustained exposure to phytochemicals with
Aβ-related dysregulation in cellular responses may be the promotion of oxidative stress. Since
oxidative stress can regulate Aβ-cytotoxicity and transcription of pro-inflammatory cytokines, it
may explain the adverse behavioural impact of the WFD [126,127]. This framework agrees with
findings of elevated TNFA expression in transgenic animals receiving the WFD without such an
effect in wildtype mice, or transgenic mice receiving the control diet, that presumably had intact
75
cellular response systems or were not exposed to similarly high amounts of phytochemicals
respectively. Epidemiological studies have generally indicated that dietary phytochemical
consumption is related to better cognitive function and reduced incidence of dementia [270-272].
However, one study found that certain phytochemical subclasses exerted positive or null effects
on episodic memory, but were negatively associated with executive function [272]. This
domain-specific effect resembled our results in which frontal lobe-dependent strategic rule
learning was further impaired in transgenic mice on the WFD, but there was no such effect on
the highly hippocampus-dependent spatial memory probe.
Our results highlight the importance of better understanding the role that global diet quality may
play in moderating single nutrient effects. Some of the main components of the whole-food diet
included blueberries [250,273,274], spinach [273], and DHA [113,275] which have been shown
to beneficially influence cognitive function in animal models of aging or AD when administered
as extracts. However, when combined together in this study they adversely affected some
behavioural outcomes. These findings contrast with studies in which combining DHA with
polyphenolic compounds [259,276] seemed to elicit greater benefits than either alone on
behaviour and Aβ deposition in the Tg2576 mouse. However, not all dietary combinations have
proven beneficial. Co-administration of vitamins E and C to APP/PS1 mice resulted in spatial
memory impairments that were not seen when vitamin C was administered alone [248]. In a
similar mouse model combining DHA with phospholipid precursors exacerbated Aβ deposition
compared to either separately whereas co-administering both these compounds with additional
micronutrients produced an anti-amyloidogenic effect [277]. Combining DHA with a high
saturated fat mixture appeared to promote amyloidogenic processing, and abolished the anti-
amyloidogenic effect of DHA in the TgCRND8 mouse [278]. Collectively, these studies indicate
that dietary effects may reflect the interaction between single dietary components, and that these
combinations can sometimes produce unexpected results.
It is unclear whether the impact of the WFD on behaviour and gene expression is due to the
composite effect of the entire diet, or related to the relative abundance of a particular dietary
component. For instance, past studies have found behavioural benefits to the administration of
76
berry extracts in AD mouse models [250,274]. However, the WFD in this study also contained
relatively large amounts of cruciferous vegetables (kale, cabbage, broccoli, brussel sprouts,
radishes) which are a rich source of glucosinolate phytochemicals. Glucosinolate exposure has
been shown to generate reactive oxygen species in animal models [279] which may underlie their
chemopreventative activity in cancer cells [280]. If such pro-oxidant effects apply to brain tissue
undergoing active Aβ deposition, that would support our hypothesis that chronic exposure to
large amounts of phytochemicals may underlie the cognitive deficits and neuroinflammatory
gene expression associated with WFD consumption by the transgenic animals. In terms of
comparative dosages, the dietary concentration of DHA was lower than studies which have
observed positive [113,259] and null [203] effects on behaviour in AD mouse models, but
comparable to another in which DHA exerted beneficial influences on electrophysiological and
behavioural outcomes [275]. The dietary concentration of phytochemicals, as assessed by total
polyphenol content [274] , and dietary antioxidant capacity [250] appear higher than other
studies in AD mouse models. However, this was by design as other studies used purified
extracts or compounds which may have had higher bioavailability and would not have been
subject to food-food interactions [281]. The dosage of dietary bioactives in this study were most
likely higher than could be reasonably attained from non-supplemental sources in the human
diet, and there is always the possibility that the WFD would exert different effects in mice than
in human subjects. However, such limitations apply to most dietary studies in AD mouse models
with implications that are beyond the scope of this paper. It was not our intent to use this study
to develop specific recommendations, but to test a novel hypothesis based on supporting
observational evidence with relevance to human health. Future attempts to determine the relative
importance of specific dietary components versus their dosage may be informative.
In conclusion, a whole-food diet based on the epidemiological literature exacerbated cognitive
dysfunction in a mouse model of familial AD possibly by enhancing neuroinflammation. These
unexpected results highlight the potential complexity of food-food interactions, and the
potentially unexpected ways in which diet may influence AD progression. We feel this study
supports those who caution against high dose supplemental consumption of phytochemicals, and
promote the need to better assess the safety profile of food-borne compounds [282]. Such
77
caution seems warranted, as individuals with cognitive impairments have been shown to be
interested consumers of herbal remedies and supplements which are widely perceived to be
harmless by elderly people [283,284].
4.6 Acknowledgements
This work was supported by a grant from the Canadian Institutes of Health Research. We thank
Ms. Rosemary Ahrens, Mr. Jeremy Audia, Ms. Mary Hill, and Dr. Joanne McLaurin for their
technical assistance and advice.
4.6.1 Disclosure Statements
None of the authors report any potential or actual conflicts of interest. Procedures were carried
out in accordance to the policies set out by the Canadian Council on Animal Care, and were
approved by animal care committees at Trent University and the University of Toronto.
78
5 Chapter 5: Rosiglitazone prevents hippocampal-dependent
memory deficits associated with peripheral metabolic
dysfunction in a rat model of diet-induced obesity
Student’s Contribution: MDP conceived of the research plan; raised the animals; designed,
produced and administered the experimental diets; collected the food intake data; collected the
tissue and body measurements; conducted intracerebroventricular insulin infusions; conducted
the gene expression analysis; isolated the hippocampal proteins and assisted in analyses of
protein abundance; assisted in biochemical assessments of plasma biomarkers; conducted the
statistical analyses; wrote the manuscript. Dr. Gordon Winocur was responsible for cognitive
assessments.
79
5.1 Abstract
Diet-induced obesity (DIO) has consistently been shown to impair cognitive function in rodents.
Our main objective was to determine whether hippocampal and/or peripheral insulin resistance
(IR) was involved in mediating this impairment. Weanling rats fed either control (CON; 12%
fat) or high fat diets (HFD; 41% fat) supplemented with or without the insulin-sensitizing drug
rosiglitazone (5mg/kg body weight) for three months were tested on an operant bar pressing task
of learning and memory. Relative to CON, HFD caused deficits in hippocampal-dependent
memory which were prevented by dietary co-administration of rosiglitazone. Rosiglitazone
corrected an indicator of peripheral IR, but there were no group differences in hippocampal IR as
determined by relative abundance of phosphorylated Akt (p-Akt) in relation to the total Akt pool
(p-Akt/Akt) following intracerebroventricular insulin infusion. Memory deficits were most
strongly associated with a composite indicator of peripheral metabolic dysfunction that reflected
simultaneous, additive impacts of IR involving adipose tissue inflammation (monocyte
chemoattractant protein-1) and hyperleptinemia. Interestingly, insulin-stimulated p-Akt
abundance was associated with better memory only in those animals with relatively low levels of
peripheral metabolic dysfunction. These results parallel clinical evidence highlighting the
susceptibility of the hippocampus to obesity-related metabolic disorders, and strongly implicate
peripheral IR involving adipose tissue inflammation as a major mediator of memory impairments
associated with DIO. Hippocampal IR per se did not appear to be involved in this process.
Further investigation of alternative, brain-related mechanisms such as changes in insulin and/or
glucose availability and regulation of downstream insulin-stimulated transcriptional activation
seem warranted.
80
5.2 Introduction
Midlife obesity has consistently been linked with accelerated cognitive aging and increased
incidence of dementia in later-life [285,286], and is associated with worse cognitive performance
even in adolescents and young adults [287,288]. The core metabolic defect associated with
obesity-related disorders like type-2 diabetes and the metabolic syndrome (MetSyn) is insulin
resistance which has been associated with poor cognitive outcomes [289-293]. Molecular
components of the insulin signaling pathway have been shown to be involved in supporting
memory consolidation and functional plasticity [6,234,294]. Administering insulin directly into
the central nervous system (CNS) has been shown to enhance cognitive performance [295-298]
while chronically high plasma insulin concentrations, as seen in insulin resistance, and obesity
have been shown to reduce insulin delivery into the CNS [299-301]. Therefore, the dynamic
between peripheral and central insulin resistance would appear to be important mechanisms of
interest when examining the adverse effects of obesity on cognitive function.
We have consistently observed hippocampal-dependent memory deficits in a model of mild to
moderate, high fat diet-induced obesity [1]. Others have attempted to distinguish the cognitive
impairing effects of diet-induced obesity from peripheral insulin resistance per se through the
administration of insulin-sensitizing drugs with mixed results [302-304]. Drawing conclusions
from these studies, or making comparisons to our own, is made difficult by differences in the
onset, duration, and type of drug and/or dietary treatments which may influence the profile of
metabolic and cognitive results. Therefore, our main objective was to determine whether
hippocampal and/or peripheral insulin resistance was involved in mediating memory deficits
found in our established model of diet-induced obesity (DIO). Given reports of its cognitive
enhancing effects [251,305,306], the peroxisome proliferator-activated receptor-gamma (PPARγ)
agonist rosiglitazone—a member of the thiazolidinedione (TZD) class of insulin-sensitizing
drugs—was co-administered with a high fat diet to prevent the development of peripheral insulin
resistance. We tested learning and memory using the same operant-based bar pressing task
previously shown to be sensitive to cognitive impairments in this model of DIO. Correlations
between cognitive function and blood-borne indicators of peripheral metabolic dysfunction or
81
markers of in vivo hippocampal insulin signaling following intracerebroventricular insulin
infusion were examined. We also set out to test for possible dependency between peripheral
metabolic dysfunction and hippocampal indicators of insulin resistance in mediating any
potential cognitive deficits.
5.3 Methods
5.3.1 Subjects and Diets
Weanling, male Long-Evans rats (Charles River Laboratories Inc., St-Constant, QC) were
randomly assigned to one of four experimental groups at 4 weeks of age when body weight
ranged from 56 to 81 grams. Rats were individually housed in ventilated plastic boxes with
bedding in facilities with controlled temperature (22 ± 1°C) and light cycle (12 h light/dark)
consistent with guidelines established by the Canadian Council on Animal Care. All procedures
were approved by animal care committees at Trent University and the University of Toronto.
Each experimental group received one of four powdered diets consisting of a high fat diet (HFD)
or control diet (CON) and the same diets supplemented with the insulin-sensitizing drug
rosiglitazone (HFD-ROSI or CON-ROSI). Consistent with our past studies [178,307,308], the
high fat diet provided 41% of energy from fat versus 12% for the control diet (Table 5.1). The
main source of dietary fat for the high fat and control diets was beef tallow (Dyets Inc.,
Bethlehem, PA) and soybean oil (Persall Naturals Ltd., Waterford, ON) respectively. The
amount of vitamin and mineral mix included in the HFD was adjusted upwards in proportion to
the higher energy density (4.5 kcal/g) compared to the CON (3.8 kcal/g). The powdered diets
were packed into glass feeding jars held in place by a stainless steel bracket attached to a large
tray that collected spillage. Fresh diet was prepared twice weekly, and stored in airtight
containers at 4°C until feeding. Rosiglitazone (Cayman Chemical Company Inc., Ann Arbor,
MI) was mixed into freshly prepared diet to provide a constant average dose (5 mg/kg of body
weight) within each experimental group. The amount of drug mixed into the diet for each
experimental group, in order to maintain a constant mean dosage, was updated biweekly based
on weekly measurements of body weight and food intake. Food intake was determined for each
82
rat by subtracting the weight of diet left in the jars from the amount of initial diet provided taking
into account any spillage. Spillage was determined by weighing the amount of diet powder that
could be separated from bedding and other materials covering the collection tray using a wire
mesh sieve. Rats had ad libitum access to their experimental diets for a 93-day passive feeding
phase at the University of Toronto, and for 1 week after being transferred to Trent University for
behavioural testing. Rats were then restricted to 80% of mean ad libitum consumption of their
particular experimental group starting at the beginning of behavioural testing until sacrifice at
approximately 5 months of age. This restriction was not meant to cause great weight loss, but
only to ensure motivation for tests involving food rewards. If animals lost more than 10% of
pre-restriction body weight, the food supply was adjusted to restore any losses. Water was freely
available throughout the study.
83
Table 5.1. Composition of experimental diets
Ingredient (g/kg) High fat diet Control diet
Casein 238 200
Corn starch 449.42 649.49
Beef tallow 180 --
Soybean oil 12 50
Safflower oil 8 --
Cellulose 50 50
Mineral mix, AIN-93G-MX 42 35
Vitamin mix, AIN-93-VX 12 10
L-cystine 3.57 3
Choline bitartrate 5 2.5
tert-Butylhydroquinone 0.01 0.01
84
5.3.2 Variable-interval delayed alternation (VIDA) task
The VIDA task has been described in detail elsewhere [308,309]. All testing was conducted in
computer-controlled Skinner boxes, outfitted with a single retractable lever to the right of a
central feeder, housed in a sound-proof chamber and illuminated by a 3-Watt light centrally
installed in the roof chamber.
Food-restricted rats were shaped to press the lever for food according to a continuous
reinforcement schedule (CRF). CRF training, which consisted of one 30-min session per day,
continued until a response rate of 80 responses per session was achieved over two consecutive
days. During CRF training, each leverpress was rewarded with a single 45-mg Noyes food
pellet. After each session, rats were returned to their cages and received their daily allotment of
diet.
VIDA testing was initiated the day after criterion was reached in CRF training. Each test session
consisted of 14 reinforced (go) trials alternating with 14 non-reinforced (no-go) trials, which
were 20 seconds long. During the go trials, each leverpress produced a 45-mg Noyes food pellet,
whereas leverpresses during the no-go trials were not rewarded. The go and no-go trials were
separated by a variable inter-trial interval (ITI), during whch the lever was retracted. ITI’s were
0, 2.5, 5, 10, 20, 40, or 80 seconds long with each interval occurring twice after go trials and
twice after no-go trials, so that each ITI occurred four times per session. The ITI sequence
varied for each session, which always began with a go trial. Testing sessions continued daily for
15 consecutive days.
5.3.3 Intracerebroventricular insulin infusion and tissue collection
Two weeks after testing sessions were completed, animals received intracerebroventricular
(ICV) infusions of insulin into the third ventricle immediately preceding sacrifice and tissue
collection. The dose and timing of the infusion was based on previous studies that examined
insulin-stimulated enhancement of cognition and activation of hippocampal insulin signaling
[298,310]. After an overnight fast, rats were anesthetized with isoflurance gas (3% induction, 1-
85
2% maintenance) and their head secured into a stereotaxic frame (Stoelting). Before the incision
was made, 50 uL of 0.1% Sensorcaine was injected subcutaneously at the incision site. The skull
was exposed and a small hole was drilled (anterior/posterior, − 4.3 mm and medial/lateral
0.0 mm and dorsal/ventral, − 4.2 mm from bregma). A 6 uL injection of human recombinant
insulin (Sigma-Aldrich, St. Louis, MO) dissolved in sterile saline (6 mU) was infused at a
constant rate (1 uL/min) by a motorized stereotaxic injector (Stoelting) using a 10 uL syringe
with a 26s gauge, bevelled injection needle (Hamilton). The needle was left in a place for 30
minutes following completion of the infusion after which it was slowly removed. Rats were then
immediately decapitated by guillotine, and trunk blood was collected into EDTA coated tubes
and centrifuged. Following rapid excision, hippocampi were dissected from the brain on a cold
surface in PBS, flash frozen in liquid nitrogen, and stored at -80°C. Epidydimal fat pads were
removed and weighed. Animals were weighed immediately prior to anesthetization. Rats
underwent ICV infusions and were sacrificed based on a random schedule over the same two
week period.
5.3.4 Plasma biochemistry
Plasma glucose concentration was determined by commercial glucometer (OneTouch Ultra,
LifeScan Canada Ltd., Burnaby, BC) directly from trunk blood at sacrifice. Colormetric assays
were used to determine plasma concentrations of triacylglycerols and free fatty acids (Cayman
Chemical Company Inc., Ann Arbor, MI) from centrifuged trunk blood according to the
manufacturer’s directions. Plasma concentrations of leptin, insulin, and monocyte
chemoattractant protein-1 were determined using the Rat Adipokine panel kit (Millipore,
Billerica, MA) with Luminex multiplex reagents and Luminex 100 detection system (Luminex
Corp., Austin, TX) according to the manufacturer’s instructions.
5.3.5 Hippocampal gene expression analysis (Quantitative reverse transcription
real-time PCR)
Total RNA was isolated and purified from hippocampi using the Trizol method (Life
Technologies, Carlsbad, CA) and RNeasy mini kit (Qiagen, Venlo, Netherlands) according to the
86
manufacturers’ instructions. RNA purity and quantity were assessed by using a NanoDrop 1000
(NanoDrop Technologies, Wilmington, USA) to measure the 260 nm to 280 nm UV absorbance
ratio and 260 nm absorbance respectively. One microgram of total RNA was reverse transcribed
into cDNA using a High-Capacity cDNA Reverse Transcription kit (Life Technologies,
Carlsbad, CA). Quantitative real-time PCR was performed with TaqMan Gene Expression
Master Mix and TaqMan Gene Expression Assays on an ABI Prism 7000 SDS system (Life
Technologies, Carlsbad, CA). Assays were selected to target markers of insulin-signaling,
inflammation, and neurotransmission (Table 5.2). Relative differences in gene expression were
quantified by using the ∆∆CT method to calculate fold-change values between treatments [224].
Rats on the control diet without rosiglitazone (CON) acted as the reference group, and
expression was normalized by the CT of an endogenous control gene (GAPDH).
87
Table 5.2. Genes targeted for reverse transcription quantitative real-time PCR
Target gene (Gene symbol) Assay ID
Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) Rn99999916_s1
Insulin receptor (INSR) Rn00567070_m1
Insulin receptor substrate-1 (IRS1) Rn02132493_s1
Phosphatidylinositol 3-kinase, p85 alpha regulatory subunit (PIK3R1) Rn00564547_m1
Protein kinase C, alpha (PRKCA) Rn01496145_m1
Leptin receptor (LEPR) Rn01433205_m1
Glutamate receptor, ionotropic, N-methyl D-aspartate 2A (GRIN2A) Rn00561341_m1
Glutamate receptor, ionotropic, N-methyl D-aspartate 2B (GRIN2B) Rn00680474_m1
Tumor necrosis factor alpha (TNFA) Rn99999017_m1
Glial fibrillary acidic protein (GFAP) Rn00566603_m1
Interleukin 1-beta (IL1B) Rn00580432_m1
88
5.3.6 Immunoblot analysis of hippocampal protein abundance
The total membrane fraction of hippocampal proteins were isolated and analyzed as described
previously [311]. Briefly, one hippocampal hemisphere from each individual rat was
homogenized in ice-cold homogenization buffer and centrifuged for 10 min at 500 g at 4 °C.
The supernatant was saved, and protein concentration was determined by the Bradford method
using bovine serum albumin as a standard. Proteins were separated by SDS/PAGE (10%),
transferred to nitrocellulose (NC) membranes and blocked in Tris-buffered saline (TBS) plus
10% nonfat dry milk (NFDM) for 60 min. NC membranes were incubated with primary antisera
to the insulin receptor-beta (Santa Cruz Laoratories, Santa Cruz, CA; 1:1000), or Akt (1:1000) or
phosphorylated-Akt (1:1,000; serine 473) prepared in TBS/5% NFDM overnight at 4 °C with
gentle shaking. NC membranes were then washed with TBS plus 0.05% Tween 20 (TBST) and
incubated with peroxidase-labeled species-specific secondary antibodies (1:5,000) at room
temperature for 60 min. NC membranes were then washed with TBST and developed using
enhanced chemiluminescence reagents (ECL, Amersham) as described by the manufacturer.
Computer-assisted microdensitometry of autoradiographic images was determined on the MCID
image analysis system (Imaging Research Inc., St. Catherines, ON).
5.3.7 Statistical Analyses
For the VIDA task, data are expressed as the go/no-go latency ratio which was calculated at each
ITI by dividing the mean latency to the first leverpress in the go trials by the mean latency to first
leverpress in the no-go trials, with lower latency ratios indicating better performance. Data were
collapsed across days prior to statistical analyses consistent with our previous studies. To assess
rats’ ability to learn the basic alternation rule, data from ITI-0 were averaged across five blocks
of 3 days for each animal. To examine performance at increasing delays between the go and no-
go trials, analyses were limited to the last 3 days of testing (Block 5). Data for ITI-5 and ITI-10
were averaged and referred to as the ‘short delay’, whereas data for ITI-40 and ITI-80 were
collapsed and referred to as the ‘long delay’. In previous work with brain damaged rats [312],
rule learning on the VIDA task at ITI-0 and performance at short delays were linked to a neural
89
network in which the prefrontal cortex appears to be the principal region. The hippocampus is
not part of this network but damage to this structure selectively impairs VIDA task performance
at long delays. Therefore, performance at the long delay is referred to as hippocampal-dependent
memory. Thirty-six animals in total were analyzed for their VIDA task performance (HFD, n =
10; HFD-ROSI, n = 9; CON, n = 7; CON-ROSI, n = 10).
Repeated-measures factorial analysis of variance (ANOVA) using a linear mixed model (PROC
MIXED; SAS v9.3) with an autoregressive covariance structure tested differences between
groups in VIDA task performance at ITI-0 over the 15-day testing period. The fixed effect of
‘Block’ was assessed to indicate whether rats exhibited significant improvements in learning the
basic alternation rule with practice. Differences in the rate of change in learning the basic
alternation rule over the testing period were tested by the interaction between block and a
variable representing membership in one of the four experimental groups (Block x Group). The
fixed effects of high fat diet (HFD), rosiglitazone (ROSI), and their interaction (HFD x ROSI)
indicated whether mean performance over the entire testing period differed based on HFD and
ROSI assignment. One-way ANOVA was used to test differences by experimental group,
particularly when additive or interactive effects of HFD and ROSI were indicated, using the
Tukey post-hoc test to adjust for multiple comparisons. A similar approach was taken when
analyzing VIDA task performance at short and long delays (during Block 5), gene expression,
protein abundance, plasma biochemistry, body measurements, and mean energy intake except
that the fixed effects were tested in a general linear model (PROC GLM) using the Student-
Newman-Keuls test to adjust for multiple comparisons when differences by ‘Group’ were of
interest.
Correlation of plasma biomarkers and hippocampal protein abundance with VIDA task
performance, where there were statistically significant differences between groups at Block 5,
were tested using a general linear model (PROC GLM). Since we did not administer a glucose
tolerance test, an Insulin Resistance Index (IRI) was calculated based on the homeostasis model
assessment-insulin resistance score by multiplying fasting plasma glucose and insulin
concentrations. To test whether the possible association between insulin resistance and VIDA
90
task performance was dependent on specific alterations in peripheral metabolic profile, we tested
for interactions between IRI and the other plasma biomarkers (MCP-1, triacylglycerols, free fatty
acids). Observational studies have indicated that a combined indicator reflecting metabolic
dysfunction is more consistently related to cognition than its individual components [293].
Therefore, we combined statistically significant associations into a single multiple regression
model after converting fasting plasma concentrations to z-scores. A composite plasma score
(CPS) was based on the multiple regression equation from a model in which plasma biomarkers
exhibited statistically significant partial correlations. Possible dependency between blood borne
and brain-based biomarkers in mediating VIDA task performance was tested by statistical
interactions between CPS and each hippocampal protein. Interactions approaching statistical
significance were decomposed by their main effects to determine the nature of any possible
relationships with VIDA task performance at the long delay.
5.4 Results
5.4.1 Variable-Interval Delayed Alternation (VIDA) task
The ability to learn the basic alternation rule was determined by examining performance at ITI-0,
where there was no delay between the go and no-go trials (Figure 5.1, Panel A). Performance
improved with practice as evidenced by decreases in the go/no-go latency ratio over time (blocks
of 3 days) for all animals collectively (P < 0.01), and within each experimental group (all P’s ≤
0.01). Two-way ANOVA by mixed modelling determined that, collectively, animals consuming
the high fat diet exhibited worse average performance over the entire testing period compared to
animals consuming the control diet (HFD, P < 0.01; ROSI, P = 0.95; HFD x ROSI, P = 0.76).
Subsequent analysis examining differences between experimental groups at each 3-day block of
time indicated that this statistically significant HFD effect was largely driven by transiently
superior performance of the Con-ROSI group compared to the HFD-only group at blocks 3 (P <
0.01) and 4 (P < 0.01). Experimental groups did not differ at any other time including at Block 5
when all groups behaved similarly during this last stage of testing (P = 0.58). All groups showed
91
similar rates of learning as determined by testing for a ‘Block x Group’ interaction in the mixed
model (P = 0.32).
In addition, latency to the first leverpress for go and no-go trials during the first day of testing at
ITI-0 was compared between experimental groups. It was reasoned that overall longer latencies
when animals were naive to the task could reflect reduced physical ability secondary to obesity
and/or motivation to perform the task. There were no statistically significant differences on
either the go (P = 0.44) or no-go (P = 0.43) trials. Taken together, the data suggest similar
motivational-performance properties and comparable rates of alternation rule learning in all
experimental groups.
The impact of imposing a memory demand, by increasing ITI, was examined at Block 5 when
animals were most practiced at the VIDA task. Memory for the alternation rule at long delays
between lever presentations (ITI-40 and ITI-80), which is highly dependent on the hippocampus,
exhibited a ‘HFD x ROSI’ interaction such that the HFD-only group performed worse than all
other experimental groups which did not differ in their performance (HFD, P < 0.01; ROSI, P =
0.43; HFD x ROSI, P=0.021) (Figure 5.1, Panel B). There were no differences between
experimental groups for memory at the short delay (ITI-5 and ITI-10) where performance is less
dependent on the hippocampus (HFD, P = 0.15; ROSI, P = 0.41; HFD x ROSI, P = 0.36) (Figure
5.1, Panel B).
5.4.2 Fasting plasma biochemistry and body measurements
Fasting plasma glucose and the Insulin Resistance Index (IRI) were the only variables to exhibit
the same pattern of differences between treatment groups as VIDA task performance at the long
delay (i.e. ‘HFD x ROSI’ interaction) (Table 5.3). Rosiglitazone didn’t spare rats from the
increases in plasma triacyclglycerides and leptin resulting from HFD consumption. Statistically
significant main effects for HFD and ROSI indicated that they both contributed to increased
body weight, body weight gain, epidydimal fat pad mass, and mean weekly energy intake. Even
in the case of relative obesity, however, rats receiving rosiglitazone had the lowest fasting
plasma insulin concentrations. Based on differences between experimental groups the HFD-only
92
group had greater plasma insulin compared to groups receiving rosiglitazone (HFD-ROSI and
CON-ROSI) with the CON-only group possessing an intermediate value.
5.4.3 Insulin-stimulated differences in hippocampal gene expression and
protein abundance
For the most part, hippocampal expression of genes related to insulin and leptin signaling,
neuroinflammation, and glutaminergic neurotransmission did not differ between experimental
groups (Figure 5.2). The only statistically significant result involved less expression of the
phosphatidylinositol 3-kinase, p85α regulatory subunit (PIK3R1) in animals receiving the HFD
compared to animals receiving the control diet irrespective of rosiglitazone administration (HFD,
P=0.030; Rosi, P=0.44; HFD x Rosi, P=0.14) (Figure 5.2, Panel A).
Similarly, there were few differences in insulin-stimulated protein abundance of the insulin
receptor, total Akt, or in the p-Akt/Akt expression ratio which was expected to best reflect the
degree of hippocampal insulin sensitivity (Figure 5.3). However, hippocampal p-Akt abundance
was slightly lower in rats receiving rosiglitazone irrespective of high fat diet consumption
(Figure 5.3, Panel C).
5.4.4 Correlations between plasma biomarkers and hippocampal-dependent
memory
Elevated plasma leptin or Insulin Resistance Index (IRI ) were both correlated with worse
performance on the VIDA task at the long delay (Figure 5.4, Panels A and D). Fasting plasma
glucose was also positively correlated with latency ratio at the long delay (r = 0.50, P < 0.01).
There were no other statistically significant associations between plasma biomarkers or body
measurements and VIDA task performance. After pre-planned tests of statistical interactions
between IRI and other plasma biomarkers (triacylglycerols, free fatty acids, leptin, MCP-1), it
was determined that IRI was related to worse hippocampal-dependent memory only in those
animals with above average plasma concentrations of the inflammatory biomarker MCP-1
(Figure 5.4, Panel C). There was no statistically significant relationship between insulin
93
resistance and behaviour in animals with less than average concentrations of circulating MCP-1
(Figure 5.4, Panel B).
In a multiple regression model (R2 = 0.51, P < 0.01), the partial correlations for leptin and the
‘IRI x MCP-1’ interaction remained statistically significant (data not shown) suggesting they
were independently associated with VIDA task performance. A Composite Plasma Score (CPS)
was calculated, using z-scores for each variable based on the multiple regression equation (CPS
= leptin + IRI + MCP-1 + (IRI x MCP-1)), to unify independent plasma biomarkers into a single
variable. Similar to VIDA task performance (Figure 5.1, Panel B), differences in CPS score
between treatment groups exhibited a ‘HFD x ROSI’ interaction (P = 0.023) (Figure 5.4, Panel
E). CPS was more highly correlated with VIDA task performance than any other single variable
(Figure 5.4, Panel F). Furthermore, all individuals from the HFD-only group belonged to the
upper category (above average) of CPS such that they were overrepresented in the upper
category compared with individuals from all other groups (Fisher’s Exact Test, right-sided, two-
sided, and whole-table P’s < 0.01) (data not shown).
5.4.5 Correlations between hippocampal insulin-signaling proteins and
hippocampal-dependent memory
There were no statistically significant associations between VIDA task performance and any of
the measured hippocampal proteins. After pre-planned tests of statistical interactions between
protein abundance and CPS, only the ‘CPS x p-Akt’ interaction approached statistical
significance (P = 0.051). After decomposition, it was determined that this interaction reflected an
association of higher hippocampal p-Akt to better VIDA task performance only in animals with
relatively low CPS (Figure 5.5, Panels A and B; high and low categories represent a median
split). Since worse VIDA task performance in the upper versus lower category of CPS (P =
0.023) was not mirrored by a significant difference in hippocampal p-Akt content (P = 0.50), this
result suggests that metabolic dysregulation in the periphery changed the relationship between
VIDA task performance and p-Akt as opposed to modulating its abundance.
94
Figure 5.1. Performance on the variable-interval delayed alternation (VIDA) test
There were no differences between groups in the rate of learning the basic alternation rule over
15 days, displayed as five blocks of three days, when there was no delay between presentations
of the retractable lever (ITI-0) (Panel A). Furthermore, there were no group differences in mean
performance at ‘Block 5’ during the last stage of testing (Panel A). VIDA task performance at
ITI-0 is dependent on the prefrontal cortex. Memory (Panel B) for the alternation rule at the long
delay between lever presentations (ITI-40 and ITI-80) was impaired in the HFD group compared
to every other group who did not differ in their performance (HFD x ROSI, P=0.021) (Bars not
sharing the same superscript letter differ). VIDA task performance at the long delay is
dependent on the hippocampus. There were no differences in performance between groups at the
short delay (ITI-5 and ITI-10 seconds) where performance is less dependent on the hippocampus
(Panel B). All data mean ± SEM. n = 7-10/group. (HFD, high fat diet; CON, control diet;
ROSI, rosiglitazone).
Block
1 2 3 4 5
Late
ncy
Rat
io
0.0
0.2
0.4
0.6
0.8
1.0
1.2
HFD HFD-ROSI CON-ROSICON
A
a
b
bb
B
Long Delay Short Delay
Late
ncy
Rat
io
0.0
0.2
0.4
0.6
0.8
1.0 HFDHFD-ROSICONCON-ROSI
HFD x ROSI, P = 0.021
N.S.D.
95
Table 5.3. Fasting plasma biochemistry and body measurements*
HFD HFD-ROSI CON CON-ROSI Two-way ANOVA
Glucose (mmol/L) 14.6 ± 0.8a 9.9 ± 0.9b 8.1 ± 1.0b 7.6 ± 0.8b HFD P < 0.01, ROSI P < 0.01,
HFD x ROSI P = 0.025
Insulin (ng/mL) 4.2 ± 0.5a 2.2 ± 0.5b 2.8 ± 0.5ab 2.1 ± 0.5b HFD P = 0.15, ROSI P < 0.01,
HFD x ROSI P = 0.19
Insulin Resistance Index 65.4 ± 7.5a 21.9 ± 7.5b 23.7 ± 8.1b 16.9 ± 6.6b HFD P < 0.01, ROSI P < 0.01,
HFD x ROSI P = 0.021
Leptin (ng/mL) 8.9 ± 0.9a 8.9 ± 1.0a 4.4 ± 1.1b 4.9 ± 0.9b HFD P = 0.15, ROSI P < 0.01,
HFD x ROSI P = 0.19
MCP-1 (pg/mL) 325 ± 36 296 ± 39 260 ± 42 305 ± 34 HFD P = 0.46, ROSI P = 0.85,
HFD x ROSI P = 0.33
Triacylglycerol (mg/dL) 82.3 ± 9.0a 81.5 ± 9.6a 56.1 ± 10.4b 64.1 ± 8.5b HFD P = 0.028, ROSI P = 0.70,
HFD x ROSI P = 0.64
Free fatty acids (µmol/L) 905 ± 96 975 ± 103 1167 ± 111 1053 ± 91 HFD P = 0.10, ROSI P = 0.83,
HFD x ROSI P = 0.37
96
Body weight (g) 667 ± 15b 722 ± 16a 587 ± 18c 631 ± 15bc HFD P < 0.01, ROSI P = 0.015
HFD x ROSI P = 0.92
Body weight gain (g) 599 ± 14b 654 ± 15a 523 ± 17c 563 ± 14bc HFD P < 0.01, ROSI P < 0.01,
HFD x ROSI P = 0.62
Epidydimal fat pad mass
(g)
21.1 ± 1.6ab 25.7 ± 1.7a 12.6 ± 1.9c 16.9 ± 1.6bc HFD P < 0.01, ROSI P = 0.015,
HFD x ROSI P = 0.92
Weekly energy intake
(kcal)
811 ± 14a 861 ± 15b 720 ± 17c 761 ± 14c HFD P < 0.01, ROSI P < 0.01,
HFD x ROSI P = 0.81
*Differences based on the Student-Newman-Keuls post hoc adjustment are indicated by superscript letters. Values with the
same letter are not significantly different (mean ± SEM; n = 6-10/group). Abbreviations: HFD, high fat diet; ROSI,
rosiglitazone; MCP-1, monocyte chemoattractant protein-1.
97
Figure 5.2. Hippocampal gene expression
HFD HFD-ROSI CON CON-ROSI
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2PIK3R1
HFD, P = 0.030ROSI, P = 0.44
A
HFD HFD-ROSI CON CON-ROSIFo
ld c
hang
e in
mR
NA
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6INSR
HFD, P = 0.17ROSI, P = 0.37
B
HFD HFD-ROSI CON CON-ROSI
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8IRS1
HFD, P = 0.38ROSI, P = 0.81
C
HFD HFD-ROSI CON CON-ROSI
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
PRKCAHFD, P = 0.97ROSI, P = 0.87
D
HFD HFD-ROSI CON CON-ROSI
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4LEPR
HFD, P = 0.17ROSI, P = 0.64
E
HFD HFD-ROSI CON CON-ROSI
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6TNFA
HFD, P = 0.68ROSI, P = 0.71
F
HFD HFD-ROSI CON CON-ROSI
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
GFAPHFD, P = 0.71ROSI, P = 0.92
G
HFD HFD-ROSI CON CON-ROSI
Fold
cha
nge
in m
RN
A
0.0
0.5
1.0
1.5
2.0
IL1BHFD, P = 0.75ROSI, P = 0.22
H
HFD HFD-ROSI CON CON-ROSI
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6GRIN2A
HFD, P = 0.63ROSI, P = 0.58
I
HFD HFD-ROSI CON CON-ROSI
Fold
cha
nge
in m
RN
A
0.0
0.2
0.4
0.6
0.8
1.0
1.2
GRIN2BHFD, P = 0.29ROSI, P = 0.93
J
98
Collectively, animals consuming the high fat diet exhibited lower expression of the phosphatidylinositol-3 kinase, p85α regulatory subunit
(PIK3R1) irrespective of rosiglitazone co-administration (Panel A). Expression of the remaining genes involved with insulin signaling,
neuroinflammation, and glutaminergic neurotransimission did not differ between experimental groups (Panels B, C, D, E, F, G, H, I, J). n
= 5 independent samples/group. All data mean ± SEM. Definitions for gene abbreviations are found in Table 2. (HFD, high fat diet;
CON, control diet; ROSI, rosiglitazone)
99
Figure 5.3. Insulin-stimulated, hippocampal abundance of insulin signaling proteins
There were no differences in the abundance of the insulin receptor (IR) or Akt between
experimental groups (Panels A and B). Collectively, animals receiving rosiglitazone exhibited
slightly lower abundance of phosphorylated-Akt (p-Akt) irrespective of high fat diet
consumption (Panel C). However, this effect did not translate into differences in the ratio of p-
Akt to Akt which best reflects the degree of hippocampal insulin resistance (Panel D). n = 6-
10/group. All data mean ± SEM. (HFD, high fat diet; CON, control diet; ROSI, rosiglitazone)
HFD HFD-ROSI CON CON-ROSI
p-A
kt /
Akt
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
HFD HFD-ROSI CON CON-ROSI
IR le
vels
(% o
f con
trol
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
HFD HFD-ROSI CON CON-ROSI
Akt
leve
ls (%
of c
ontr
ol)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
HFD HFD-ROSI CON CON-ROSI
p-A
kt le
vels
(% o
f con
trol
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2 HFD, P = 0.89ROSI, P = 0.038
C
HFD, P = 0.33ROSI, P = 0.86
AHFD, P = 0.90ROSI, P = 0.92
B
HFD, P = 0.94ROSI, P = 0.20
D
100
Figure 5.4. Correlations between plasma biomarkers and hippocampal-dependent memory
Insulin Resistance Index(High MCP-1)
0 10 20 30 40 50 60 70 80
Late
ncy
ratio
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Insulin Resistance Index(Low MCP-1)
0 20 40 60 80 100 120 140
Late
ncy
ratio
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Insulin Resistance Index
0 20 40 60 80 100 120 140
Late
ncy
ratio
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Plasma Leptin (ng/mL)
0 2 4 6 8 10 12 14 16
Late
ncy
ratio
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
HFD HFD-ROSI CON CON-ROSI
Com
posi
te P
lasm
a Sc
ore
0
1
2
3
4
5
6
7
Composite Plasma Score
0 2 4 6 8 10 12
Late
ncy
ratio
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
r = 0.42, P = 0.022
A
P = 0.19 r = 0.76 P < 0.01
r = 0.54, P < 0.01
HFD x ROSI, P = 0.023a
b
b b
r = 0.64, P < 0.01
B C
D E F
101
Insulin Resistance Index (IRI) and plasma leptin were associated with worse VIDA task performance at the long delay (Panels A and D).
Further analysis revealed that the adverse relationship between IRI and performance was restricted to animals with relatively high plasma
concentrations of MCP-1 representing a statistically significant ‘IRI x MCP-1’ interaction (Panels B and C). A Composite Plasma Score
(CPS) reflecting plasma leptin concentration and the ‘IRI x MCP-1’ interaction exhibited the same pattern of differences between
experimental groups as VIDA task performance at the long delay (HFD x ROSI, P = 0.023; n = 6-9/grp; mean ± SEM) (Panel E). Bars
not sharing the same letter differ. The CPS was more highly correlated with hippocampal-dependent memory than any other plasma
biomarker (Panel F). n = 29 for all scatterplots. (HFD, high fat diet; CON, control diet; ROSI, rosiglitazone)
102
Figure 5.5. Correlations between hippocampal p-Akt and hippocampal-dependent memory
Hippocampal abundance of phosphorylated-Akt (p-Akt) was associated with better VIDA task
performance at the long delay in animals with relatively low peripheral metabolic dysfunction as
indicated by the Composite Plasma Score (CPS) (n = 11) (Panel A). There was no relationship
between hippocampal-dependent memory and p-Akt in rats with relatively high CPS (n = 13)
(Panel B). High and low categories of CPS reflect a median split.
Hippocampal p-Akt(Low Composite Plasma Score)
0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60
Late
ncy
ratio
0.0
0.2
0.4
0.6
0.8
1.0
Hippocampal p-Akt(High Composite Plasma Score)
0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55
Late
ncy
ratio
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
r = -0.66, P = 0.027
A B
P = 0.33
103
5.5 Discussion
Our primary objective was to determine whether insulin resistance played a role in mediating
cognitive deficits associated with a high fat diet model of diet-induced obesity (DIO). We found
that rosiglitazone prevented development of diet-induced deficits in hippocampal-dependent
memory such that group mean differences in behaviour matched differences in peripheral insulin
resistance as indicated by the insulin resistance index (IRI). Since there were no statistically
significant differences in the hippocampal p-Akt/Akt ratio between experimental groups
following intracerebroventricular (ICV) insulin infusion, it does not appear that differences in
hippocampal insulin sensitivity were involved in mediating the behavioural deficits.
Correlational analyses further defined the peripheral and cerebral factors associated with
differences in memory. For instance, peripheral insulin resistance was related to worse VIDA
task performance at long delays only in individuals with relatively high plasma MCP-1
concentrations reflecting an ‘IRI x MCP-1’ interaction. In a multiple regression model, this
interaction explained variance in VIDA task performance independent of plasma leptin which
exhibited its own association with worse hippocampal-dependent memory. A composite plasma
score (CPS), encompassing all of these blood-borne biomarkers, exhibited the single largest
correlation with VIDA task performance and the same pattern of differences between
experimental groups as behaviour. Furthermore, animals in the high fat diet group were
overrepresented in the highest category of CPS compared to all other diet groups. These results
suggest that the observed memory deficits were better explained by the combined impact of
indicators related to peripheral metabolic dysfunction versus their separate effects. Similarly,
observational studies have found that risk for age-related cognitive impairments were more
strongly related to a composite measure of the metabolic syndrome (MetSyn) versus its
individual components, and were restricted to those with higher expression of inflammatory
markers [293,313].
The profile of metabolic dysfunction represented by the composite plasma score implicates a
combination of insulin resistance involving adipose tissue inflammation and hyperleptinemia in
the development of memory deficits. A high fat diet favours relatively rapid expansion of
104
adipose tissue involving adipocyte hypertrophy which is associated with increased secretion of
pro-inflammatory cytokines compared to hyperplastic expansion by increasing the number of
adipocytes [314,315]. MCP-1 is secreted by adipocytes in response to localized expression of
pro-inflammatory cytokines [316] and is critically involved in the infiltration of adipose tissue by
activated macrophages which have been linked to progression of obesity-related metabolic
dysfunction and insulin resistance [314-316]. In this study, the adverse effect of insulin
resistance on memory was dependent on elevated plasma MCP-1, and animals on the high fat
diet were overrepresented in the upper category of a group exhibiting simultaneous elevations in
both insulin resistance and MCP-1. Therefore, our results support others who directly linked
adipose tissue inflammation to the development of hippocampal-dependent memory deficits
caused by diet-induced obesity [317]. Our results indicate the importance of co-existent insulin
resistance in this process as there were no differences in plasma MCP-1 between experimental
groups, and MCP-1 was not independently associated with VIDA task performance.
Interestingly, PPARγ-dependent transcriptional activity promotes pre-adipocyte differentiation
and has been shown to be repressed by a high fat diet in adipose tissue [318,319]. Thus, PPARγ
agonism by rosiglitazone may have allowed for hyperplastic adipose tissue expansion, and
prevented the adipose tissue inflammation associated with expansion by adipocyte hypertrophy.
Although direct examination of adipose tissue histology would be required to confirm this
hypothesis, our findings suggest that such potentially beneficial effects were independent from
obesity per se as rosiglitazone independently increased both body weight and epidydimal fat pad
mass.
Elevations in plasma leptin were independently associated with worse VIDA task performance
across all experimental groups. This result agrees with previous work linking hyperleptinemia to
adverse effects on hippocampal structural integrity and functional plasticity in a rodent model of
genetic obesity [320,321]. These adverse effects contrast with growing literature supporting a
role for leptin in the facilitation of hippocampal function [322,323]. This mismatch may reflect
decreased access of leptin to hippocampal targets in the obese state. Accordingly, obese animals
exhibit less efficient transport of leptin across the BBB and less activation of hippocampal leptin
signaling in response to peripheral leptin administration [324,325]. While multiple mechanisms
105
may exist, plasma triacylglycerols appear to be important mediators of this transport defect
[326]. In this study, animals consuming the high fat diet exhibited similar plasma leptin and
triacylglycerol concentrations irrespective of whether or not they received rosiglitazone. The
fact that plasma leptin was similar between groups with markedly different VIDA task
performance, but was associated with hippocampal-dependent memory function across all
groups, appears to highlight the importance of considering the additive impact of insulin
resistance and hyperleptinemia as represented by the composite plasma score. Under this
conceptual framework, the adverse impacts of less hippocampal leptin availability could be
accommodated as long as insulin resistance involving adipose tissue inflammation was
controlled.
Analyses of gene expression and protein abundance did not strongly support a role for
hippocampal insulin resistance in mediating the observed memory deficits, and suggest that
alternative mechanisms including reduced insulin or glucose availability may be responsible.
We found that activation of the insulin-signaling pathway in the hippocampus, as indicated by
the insulin-stimulated p-Akt/Akt ratio, did not differ between groups. There were also no
differences in hippocampal abundance of the insulin receptor or total Akt. These results agree
with studies finding no differences in insulin-stimulated activation of signaling proteins in
animals consuming a high fat diet [327,328], but disagree with another study reporting decreased
insulin-stimulated p-Akt levels [329]. Similar variability is seen for the basal activation state
with studies indicating increased [330] or no differences in p-Akt hippocampal abundance in
models of diet-induced obesity [302,329]. We found that p-Akt abundance was lower in animals
receiving rosiglitazone irrespective of high fat diet consumption, and gene expression of PIK3R1
was lower in animals receiving the HFD irrepsepctive of rosiglitazone administration.
Although one might expect these two components to move in tandem, the PIK3R1 gene is
spliced into variants producing proteins with differing efficiencies in insulin signal transduction
that have been shown to be functionally dissociated from PI3k activity and/or Akt
phosphorylation in some cases [331-334]. In an effort to maximize statistical power we did not
include a saline control in the ICV protocol, however, increased abundance of p-Akt relative to
the total Akt pool (p-Akt/Akt ratio) is a better indicator of insulin-stimulated cell signaling than
106
total p-Akt abundance which may reflect both basal and insulin-stimulated activation states. In
this study, lower abundance of hippocampal p-Akt in experimental groups receiving
rosiglitazone corresponded with lower levels of circulating insulin. Interestingly, a positive
correlation between fasting plasma and CSF insulin concentrations has been reported in rodents
which suggests that group differences in fasting p-Akt abundance may reflect differences in CNS
insulin availability [335]. If true, our results suggest that small differences in fasting CNS
insulin may not be as important to cognitive function as insulin flux across the blood-brain
barrier (BBB) during periods of transient change in plasma insulin—a process that is impaired by
diet-induced obesity and insulin resistance [299,300,335]. Transient increases in circulating
[336] and CNS insulin concentrations [295] have been shown to enhance cognitive function even
in obese and diabetic individuals [296,297]. We have previously shown that peripheral glucose
injection greatly attenuated hippocampal-dependent memory deficits in the current model of diet-
induced obesity [308]. This attenuating effect may reflect increased availability of CNS insulin
and glucose, or some combination of both factors. Research demonstrating transient decreases in
rat hippocampal glucose levels during memory-demanding cognitive tasks, and its replenishment
with glucose administration [337,338], suggests that under conditions of increased neuronal
activity local neuronal glucose supply may become limiting. Since glucose is the principle
substrate for brain energy metabolism and is used for neurotransmitter synthesis [339], local
glucose depletion may be detrimental to actively firing neurons. Recent studies have indicated
that individuals with insulin resistance exhibit cerebral glucose hypometabolism which has been
associated with worse cognitive function [340-343]. Interestingly, enhancement of synaptic
plasticity and cognitive function by rosiglitazone has been associated with upregulated
abundance of hippocampal glucose transporters [306]. It should be noted that our findings do
not preclude contributions by other mechanisms linked to the adverse effect of DIO on
hippocampal memory including changes in BBB integrity [344], cerebrovascular reactivity
[345], neurotransmission [346,347], antioxidant defense systems[347,348], and
neuroinflammation [317]. Although we did not find differences in the expression of genes
associated with neuroinflammation or glutaminergic neurotransmission, this does not preclude
their possible involvement at the protein level or in different models of diet-induced obesity.
Thus, a number of mechanisms may act to cause the hippocampal-dependent memory deficits
107
associated with diet-induced obesity, but results from the model used in this study strongly
suggest that hippocampal insulin-sensitivity per se is not one of them.
Although we found no differences in hippocampal insulin sensitivity, peripheral metabolic
dysfunction had a profound impact on the relationship between central insulin-signaling and
memory. Greater abundance of hippocampal p-Akt was associated with better memory only in
animals with relatively low levels of peripheral metabolic dysfunction as indicated by the CPS.
There was no difference in hippocampal p-Akt abundance between upper and lower categories of
CPS. This finding bears some similarity to an examination of skeletal muscle in which obese
subjects exhibited normal insulin-stimulated phosphorylation of Akt compared to lean
individuals, but Akt activity was correlated with glucose disposal rate only in the lean subjects
[349]. Interestingly, the basal activation state of hippocampal insulin signaling molecules in
post-mortem tissue from non-diabetic individuals was closely related to cognitive ability even
after adjusting for the presence of Alzheimer’s disease neuropathology [212]. Our results
suggest that downstream mediators of the relationship between p-Akt and memory were
influenced by peripheral metabolic dysfunction rather than differences in the abundance of p-Akt
which did not differ between upper and lower categories of CPS. One such set of regulators
belong to the Forkhead box (FOXO) family of transcription activators which are downstream
substrates of Akt in hippocampal neurons [350] and have been shown to negatively regulate
peripheral insulin action in diet-induced obesity partly by suppressing activation of PPARγ
transcriptional activation [318]. This suppression has been attributed to less endogenous
phosphorylation of FOXO proteins resulting in their localization to the nucleus [351] leading to
increased FOXO transcriptional activity and direct interference in the ability of PPARγ to bind
promoter regions on its target genes [352]. Interestingly, hippocampal PPARγ agonism by
rosiglitazone has been linked to cognitive enhancements [251] in a transgenic mouse model of
AD which also exhibits perturbed insulin signaling [209]—much as seen in DIO either through
intrinsic signaling defects or lack of functional insulin. Thus, metabolic dysfunction in the
periphery may be linked with less hippocampal phosphorylation of FOXO by Akt leading to
suppression of PPARγ transcriptional activity. This theoretical model relies on assumptions
regarding patterns of Akt and FOXO activation that would definitely need to be confirmed.
108
However, it has the benefit of linking rosiglitazone to improvements in cognitive function by
acting to reduce peripheral metabolic dysfunction and/or direct hippocampal agonism of PPARγ.
Investigation of SOCS3 (suppressor of cytokine signaling-3) and PTP1B (protein-tyrosine
phosphatase 1B), which suppress leptin receptor signaling and regulate insulin action, may also
provide insight into the mechanisms at play in this study as their overlapping functions have the
potential to produce significant crosstalk between signaling pathways in DIO where insulin and
leptin resistance develop together [353,354].
Our results cannot distinguish between the peripherally mediated and potential direct effects of
rosiglitazone in ameliorating hippocampal-dependent memory deficits associated with a high fat
diet. However, there are strong suggestions that rosiglitazone crosses the blood-brain barrier to
activate CNS PPARγ including a recent study in which oral rosiglitazone increased hippocampal
PPARγ binding activity, and ICV administration of a PPARγ antagonist reversed the cognitive
enhancements associated with peripheral administration [251]. Furthermore, CNS PPARγ
activity has been linked with the hyperphagia and weight gain frequently seen with rosiglitazone
treatment, and is required for its insulin-sensitizing effects on peripheral tissues [355,356]—all
of which were observed in this study. Such results imply that TZD’s influence complex and
bidirectional dependency between brain and peripheral tissues during the development of DIO.
Differences in pharmacological mechanisms may explain why cognitive deficits associated with
a high fat diet are more consistently prevented, including our own study, by rosiglitazone [303]
versus other insulin-sensitizing drugs like metformin [302,304]. However, these studies also
differed in the onset and duration of the dietary and pharmacological treatments, as well as, in
the metabolic profiles elicited by these factors. Future comparative studies would benefit from
the use of more similar experimental models. Understanding the activity of key enzymes
involved in metabolic signaling versus only their abundance may be informative as these two
parameters have been shown to be discordant in type-2 diabetics [357,358]. Based on our
results, assessment of both basal and activated insulin signaling states would also appear to be
important.
In conclusion, a high fat diet caused deficits in hippocampal-dependent memory which were
109
prevented by oral co-administration of the insulin-sensitizing drug rosiglitazone. Memory
function was most strongly associated with a composite indicator of peripheral metabolic
dysfunction that reflected simultaneous, combined impacts of insulin resistance involving
adipose tissue inflammation and hyperleptinemia. This adverse metabolic milieu appeared to
modify the relationship between a marker of hippocampal insulin-signaling and memory function
such that insulin-stimulated p-Akt abundance was associated with better memory only in those
with relatively low levels of peripheral metabolic dysfunction
110
6 Chapter 6: General Discussion of Thesis Results
111
6.1 Overview of Objectives & Summary of Results
The overall goal of this thesis was to determine whether selected dietary components exerted
their neurocognitive effects as a part of a broader set of dietary patterns, and by influencing brain
insulin signaling and/or peripheral insulin sensitivity.
A major objective of this thesis was to investigate the impact of dietary components, identified
from observational studies of age-related cognitive decline and dementia, on cognition in animal
models. These components included both foods (fish, fruits, vegetables) and a macronutrient
class (saturated fat). Since foods are not consumed in isolation from each other and have been
shown to be auto-correlated within an individual’s diet, another key objective was to investigate
the effect of food combinations on cognition in both animal models and older adults. In animals,
it was expected that a combined whole-food diet consisting of fish, vegetables and fruit would
improve cognitive function in a transgenic mouse model of Alzheimer’s disease (AD) whereas a
high saturated fat intake would produce memory deficits in rats. In older adults, the underlying
hypotheses were that: (A) dietary patterns associated with consumption of the dietary
components targeted in the animal studies would be identified in older adults; and (B) that a
dietary pattern associated with consumption of fish, fruits and vegetables would exhibit a
beneficial relationship with cognitive function whereas another pattern associated with saturated
fat intake would exhibit an adverse relationship. Finally, molecular components of the insulin-
signaling pathway are involved in neuronal processes that support cognitive function. Clinical
and observational studies have found that whole body/peripheral insulin sensitivity is related to
cognition. Peripheral insulin sensitivity has been shown to have an impact on central, or brain,
insulin availability which may impact on central insulin-signaling activity and, thus, cognition.
Therefore, the final objectives were to investigate whether: (A) diet-induced differences in
cognitive function were accompanied by molecular differences in brain insulin-signaling
molecules; and (B) whether diet-induced differences in peripheral insulin sensitivity impacted
brain insulin-signaling and cognition. Since changes in insulin sensitivity and insulin signaling
are often associated with inflammatory status expression of neuroinflammatory genes were also
monitored.
112
The dietary components tested in animal models were identified from the observational literature
as it stood at the start of this PhD project. This literature suggested that less age-related cognitive
decline and/or incidence of AD was related to increased consumption of fish, vegetables, fruit,
and polyunsaturated fatty acids. Saturated fat stood out as the main dietary exposure related to
adverse cognitive outcomes. The literature surrounding the diet-cognition relationship has
expanded rapidly in the ensuing years to include great interest in the potential role for food
combinations, or dietary patterns, which are thought to stand as an indicator of global diet quality
[151]. The benefits of studying the effects of global diet quality versus single foods or nutrients
have been suggested to include more realistic modelling of human consumption patterns and
possible synergism between dietary components. Foods, rather than nutrients, acted as the
favoured targets for the animal (Chapter 4) and observational studies (Chapter 3) in this thesis as
they provided many of the nutrients and phytochemicals that had been proposed to be beneficial
to cognitive function. For instance, fish, vegetables, and fruit collectively provide PUFA, plant
polyphenols, and other dietary antioxidants. The literature at the time had not really identified
foods that were adversely related to cognition so saturated fat was targeted for animal
investigation of an adverse diet (Chapter 5), and the relationship between dietary patterns in the
observational study (Chapter 3) with saturated fat intake was of interest.
Chapter 3 addressed the objective of investigating whether empirically-derived dietary pattern(s)
in older adults are associated with intake of the dietary components tested in rodents, selected
due to their prominence in the epidemiological literature, and exploring their possible
association(s) with cognitive function. It was found that intake of the selected dietary
components (fish, vegetables and fruit; saturated fat) were associated with adherence to a broader
set of dietary patterns which were themselves related to cognitive function. In this study the
magnitude and characteristics of the diet-cognition relationship depended on an individual’s
socioeconomic position (SEP). For instance, cognitive benefits of adherence to a prudent dietary
pattern were seen irrespective of SEP, but differed in their form such that higher adherence at
recruitment was associated with less decline in those with low SEP whereas it was associated
with better performance at entry among those with high SEP. Alternatively, worse overall
performance and more cognitive decline were associated with higher adherence to a Western
113
dietary pattern at recruitment only in those with relatively low educational attainment. These
interactions were not merely the product of socioeconomic gradients in diet quality as they
reflected cognitive performance of individuals with dissimilar SEP but equivalent diet quality.
In Chapter 4, a whole-food diet (WFD) containing freeze-dried fish, vegetables, and fruit,
exacerbated cognitive dysfunction in a mouse model of familial AD possibly by enhancing
neuroinflammation. Lower hippocampal expression of the MAPK1 and GSK3A genes,
indicating downregulated brain insulin signaling, may have contributed to impaired performance
by the transgenic animals on every behavioural test in this study. However, they did not appear
to coincide with exacerbation of these behavioural deficits by the WFD on certain tasks. In
contrast, higher hippocampal expression of TNFA in transgenic animals receiving the WFD
suggested that unexpected promotion of neuroinflammation mediated the observed behavioural
deficits. The WFD was composed based on observational studies which identified its
components as being related to less incidence of AD. Thus, Chapter 4 addressed objectives
related to investigating the effects of a diet associated with consumption of fish, vegetables, and
fruit on cognition, brain insulin-signaling, and neuroinflammation.
Chapter 5 found that a high saturated fat diet caused deficits in hippocampal-dependent memory
which were prevented by co-administration of the insulin-sensitizing drug rosiglitazone.
Memory function was most strongly associated with a composite indicator of peripheral
metabolic dysfunction that reflected simultaneous, combined impacts of insulin resistance
involving adipose tissue inflammation and hyperleptinemia. This adverse metabolic milieu
appeared to modify the relationship between a marker of hippocampal insulin-signaling and
memory function such that insulin-stimulated p-Akt abundance was associated with better
memory only in those animals with relatively low levels of peripheral metabolic dysfunction.
This chapter addressed objectives related to investigating the impact of saturated fat on cognition
and brain insulin signaling, as well as, the impact of peripheral insulin sensitivity on central
insulin signaling.
The following sections will discuss how the studies contributing to this thesis advanced scientific
knowledge in relation to each thesis objective. Following this discussion, potential limitations
114
and opportunities for future research will be highlighted. Although the studies comprising this
thesis did not always produce results that agreed with initial hypotheses, they collectively
suggest that global diet quality, brain insulin-signaling, peripheral insulin sensitivity, and diet-
induced, albeit indirect, changes in insulin-signaling all play a role in determining cognitive
function.
6.2 Separate Dietary Components & Associations with Global Diet Quality
The results of Chapter 3 indicate that dietary patterns associated with better and worse cognitive
performance were also related to intakes of fish, fruit, vegetables, and saturated fat in the
expected directions (Table 3.2 and 3.3). In agreement with our expectations, we detected a
dietary pattern in older adults which was most highly associated with vegetables, fruits, and fish
as the most highly loaded source of animal protein (Table 3.2). Higher adherence to this
‘prudent’ pattern was associated with lower intake of saturated fat, and higher intake of PUFA in
relation to saturated fat, more antioxidant vitamins, more B-vitamins, more vitamin K, and more
potassium (Table 3.3). We were the first to empirically identify a dietary pattern which was
associated with worse cognitive function in a longitudinal study. Higher adherence to this
‘Western’ pattern was associated with higher consumption of meats, potatoes, processed foods,
and higher-fat dairy (Table 3.2). Interestingly, higher adherence to the Western pattern was
associated with linear trends for reduced consumption of leafy greens and fatty fish which were
two top-loaded foods for the prudent pattern (Table 3.3). This profile of food consumption was
associated with higher intake of saturated fat, and less intake of PUFA in relation to saturated fat,
fewer antioxidant vitamins, fewer B-vitamins, less vitamin K, and less potassium (Table 3.3).
Thus, in agreement with our hypothesis, the pre-selected dietary components identified from the
observational literature for intervention in rodents appeared to be associated with adherence to
larger dietary patterns in older adults.
The dietary patterns identified in this study were similar to those related to coronary heart disease
identified in North American samples using similar statistical techniques [359]. They also
resembled dietary patterns related to cognitive function and/or risk of dementia in recent
115
prospective studies in which dietary quality was assessed by indices measuring adherence to a
Mediterranean diet [153,156], a recommended food score [157], and identification of patterned
nutrient intakes related to risk of Alzheimer’s disease using reduced rank regression [154].
These studies relied on a priori assumptions of dietary quality in that they identified foods or
nutrients of interest based on theoretical knowledge. Intakes of many single nutrients (vitamin E,
PUFA, B-vitamins, saturated fat) and foods (fish, vegetables, fruit, nuts, legumes) associated
with dietary patterns in these studies were also related to adherence to the empirically derived
patterns in this study.
6.3 Effects of Diet Quality on Cognitive Function
The results reported in Chapters 3 and 4 generally established that food combinations or dietary
patterns associated with fish, fruit and vegetable consumption can influence cognition. Since
adherence to the Western dietary pattern was associated with higher intake of saturated fat, there
are also links between the results reported in Chapter 3 and Chapter 5. In accordance with our
objectives and hypotheses, we found that diets associated with saturated fat intake adversely
influenced measures of global cognitive function in older adults (Chapter 3) and hippocampal-
dependent memory in rodents (Chapter 5). Since higher adherence to the Western pattern was
also related to less consumption fruits, vegetables, and other nutrients which may influence
cognitive function, a specific role for saturated fat cannot be directly established in the human
study. However, the results of Chapter 5 support the existence of a specific role for saturated fat
as great care was taken to ensure that the high fat diet contained adequate, and similar, amounts
of micronutrients and essential fatty acids relative to the control diet. In contrast to saturated fat,
our hypothesis that fish, fruit, and vegetable consumption would be associated with cognitive
benefits received mixed support. In Chapter 3, adherence to the prudent dietary pattern was
related to better overall cognitive performance over three years of follow-up in older adults.
However, the combination fish, vegetable, and fruit WFD in Chapter 4 exacerbated cognitive
dysfunction in a transgenic mouse model of AD. This unexpected, diet-induced cognitive
impairment was not seen in wildtype mice whose performance was unaffected by consumption
of the WFD, and that did not experience rapid, genetically determined amyloid-β peptide (Aβ)
deposition. These results indicate that factors inherent to the individual, in this case a transgene,
116
largely determined the impact of diet on cognitive function. Interestingly, we find a similar
interaction between dietary and individual-level factors when examining the neurocognitive
effects of diets associated with saturated fat. In Chapter 3 a difference in lived experience, rather
than a genetically determined predisposition, modified the dietary effect such that adherence to
the Western pattern in Chapter 3 was negatively associated with cognition only in individuals
with relatively low education. This result was interpreted to mean that education provided
redundancy within neural networks that maintained performance in the face of the potential
adverse effects of the Western diet. Interestingly, a socially and physically stimulating
environment has been shown to greatly attenuate the adverse impact of a high saturated fat diet
on hippocampal-dependent memory in the same animal model used in Chapter 5 [178]. The
results from older adults in Chapter 3 indicate that adherence to the prudent dietary pattern was
associated with better cognitive function irrespective of education or other socioeconomic
indicators which would seem to indicate that cognitive benefits of diet may be seen despite
individual differences in lived experience that were sufficient to modify the adverse influence of
diet quality. It is interesting to speculate that the WFD, in Chapter 4, did not exert any
neurocognitive effects in the wildtype mice due to their relatively young age, compared to the
older adults in Chapter 3, and because the control diet provided mostly adequate amounts of
essential nutrients. However, diet-induced, hippocampal-dependent memory deficits were seen
in the young adult rats fed a high fat diet from weaning in Chapter 5. These finding may indicate
that the onset and duration of an individual’s exposure to a given diet not only determines its
neurocognitive effect, but also suggests that the potential benefits of global diet quality are more
likely to be seen in older age compared to its detrimental effects which can be seen at much
younger ages. This apparent disconnect may reflect the importance of adequate diet quality
during brain development.
Based on visual inspection of Figure 3.1, it appears that higher adherence to the prudent pattern
may actually be associated with accelerated cognitive decline in upper categories of income,
education, and composite SEP. However, statistically significant relationships between prudent
pattern adherence and cognitive decline (‘diet x time’ interaction) were not detected within any
of the upper categories of SEP upon stratification. Despite appearances, therefore, differences in
117
cognitive function were restricted to study entry and could not be extended to include differences
in rates of decline. Perhaps with longer follow-up or less variable estimates of cognitive function
such a relationship could be reliably detected, and would provide an interesting parallel to the
results of Chapter 4 in which a diet resembling the prudent pattern exacerbated cognitive
dysfunction in a transgenic mouse model of AD. It would also be interesting if adherence to the
prudent pattern exhibited both adverse and beneficial relationships with cognitive function
depending on the timescale of observation. As alluded to earlier, the fact that benefits of the
prudent pattern appear restricted to performance at entry in upper categories of SEP may relate to
a number of factors including the timing and/or duration of observation, as well as, the specific
nature of dietary and income-related influences on cognitive function. For instance, both diet
and SEP may exert their effect on cognitive function over the entire lifetime as indicated by
studies implicating SEP in mediating peak cognitive performance earlier in the lifecourse [160]
and the potential importance of mid-life health behaviours on late-life health [360]. It is
expected that individuals with more indication of reserve will exhibit delayed onset of faster
rates of decline [361]. Thus, higher scores at entry in more adherent individuals may reflect
higher peak performance and delayed onset of decline due to greater cognitive reserve resulting
from greater lifetime exposure to the prudent diet compared to less adherent individuals.
Without information on changes in diet and SEP throughout the lifecourse, however, definitive
conclusions cannot be made.
In summary, the results reported in this thesis largely confirm that diets associated with the
consumption of fish, fruits, vegetable, and saturated fat influence cognitive function when
studied using observational and experimental approaches. These dietary components are shown
to be associated with a larger set of dietary patterns in independently living, healthy older adults.
Both the experimental and observational evidence suggest a certain dependency of diet on other
lifestyle factors, or inherent individual differences, in mediating its neurocognitive effects.
These individual-level differences appear to largely determine whether a given diet has a
beneficial, adverse, or null impact on cognition.
118
6.4 Functional Insulin & Cognitive Functions
The major thesis objectives addressed by the rodent studies (Chapters 4 and 5) related to whether
diet-induced differences in cognitive function were accompanied by corresponding patterns of
enhanced (Chapter 4) or diminished hippocampal insulin-signaling and/or whole-body insulin
sensitivity (Chapter 5). Generally, these studies indicated that markers of brain insulin signaling
were related to optimal cognitive function, but diet-induced modulation of the hippocampal
insulin signaling pathway was, at best, indirect. However, diet-induced, peripheral insulin
resistance and metabolic dysfunction were closely related to cognitive deficits (Chapter 5).
In Chapter 5 indicators of functional insulin activity in both the brain and peripheral tissues were
assessed as changes in both compartments were postulated to be involved in processes leading to
memory impairment. Assessments were confined to brain-based indicators of insulin signaling
in Chapter 4 as the experimental diets were isocaloric, matched for macronutrient distribution,
and the nutrients supplied by the diet were postulated to act directly on the brain (plant
polyphenols, omega-3 fats). Contrary to the initial hypotheses, both rodent studies found that
dietary interventions did not directly alter molecular components of the insulin signaling
pathway in the brain. For instance, expression of most insulin-related genes and proteins did not
appear to be affected by consumption of the whole-food or high fat diets. These results stand in
contrast to the adverse relationship between peripheral insulin resistance and hippocampal-
dependent memory in Chapter 5. In that study a high fat diet induced peripheral insulin
resistance, but did not have a corresponding effect on hippocampal insulin resistance per se as
indicated by insulin-stimulated phosphorylation of Akt relative to its total cellular abundance.
However, a molecular indicator of hippocampal insulin signaling (p-Akt) was related to better
memory in a subset of individuals exhibiting a relatively low degree of peripheral metabolic
dysfunction which included insulin resistance. This relationship was not observed in individuals
with a relatively high degree of diet-induced peripheral metabolic dysfunction. Therefore, the
observed memory deficits involved changes in both peripheral and central insulin signaling to the
extent that diet-induced changes in peripheral insulin resistance could be said to disrupt the
beneficial relationship between an important component of the insulin signaling pathway and
behaviour. As mentioned in Chapter 5, lower abundance of hippocampal p-Akt in experimental
119
groups receiving rosiglitazone corresponded with lower levels of circulating insulin. A positive
correlation between fasting plasma and CSF insulin concentrations has been reported in rodents
which suggests that group differences in fasting p-Akt abundance may reflect differences in CNS
insulin availability [335]. Interestingly, this correlation does not apply to obese animals that are
likely to be insulin resistant. These studies highlight the importance of considering insulin
signaling in the brain as a product of functional insulin activity—a property that reflects both the
amount of CNS insulin present and neuronal capacity to transduce external hormone into a
cellular signal. Impairments in either factor could lead to adverse effects on cognition. For
instance, there is now fairly conclusive evidence that Alzheimer’s disease is characterized by an
insulin-resistant brain state resulting from activation of cellular stress kinases by oligomeric Aβ
and subsequent inhibitory phosphorylation of insulin receptor substrate-1 [209,212]. This
pathological downregulation of neuronal insulin signaling could impair optimal activation of
insulin signaling molecules like MAPK1 which was shown to be downregulated in Chapter 4 and
implicated in memory consolidation. Conversely, there is controversy as to if, or when, the
brains’ of obese individuals become insulin resistant. A few lines of evidence strongly suggest
that impaired access of circulating insulin into the brain is at least as important to obesity-related
cognitive impairments as the loss of neuronal insulin sensitivity per se. For instance, insulin
transport across the blood-brain barrier (BBB) is impaired by diet-induced obesity and peripheral
insulin resistance [299,300,335] whereas direct administration of insulin into the brains of obese
and diabetic individuals has been shown to enhance cognitive performance [296,297]. This
evidence agrees with our finding of neuronal insulin responsiveness by obese animals in Chapter
5, and supports our interpretation that alternative mechanisms including decreased insulin access
may be responsible for the observed memory deficits.
6.5 Cognitive Impairment: Combined Effects of Inflammation & Reduced Functional Insulin Activity
The results of Chapters 4 and 5 suggest that reductions in peripheral and hippocampal functional
insulin activity work in concert with inflammation to produce cognitive impairments. In Chapter
5, the adverse relationship between peripheral insulin resistance and hippocampal-dependent
memory was apparent only in individuals with relatively high plasma monocyte chemoattractant
120
protein-1 (MCP-1) which is an indicator of elevated adipose tissue inflammation. This
dependent relationship appeared to require both factors to be elevated as MCP-1, or insulin
resistance in the subset of individuals with relatively low MCP-1, was not independently
correlated with hippocampal-dependent memory. In contrast to this dependent relationship, the
results of Chapter 4 suggest that neuroinflammation and impaired hippocampal insulin signaling
combined to produce progressively greater impairments in cognitive function. This potential,
additive effect is reflected in the way that transgenic animals consuming the WFD exhibited
lower expression of insulin related genes, more robust neuroinflammatory gene expression, and
worse cognitive function compared to their transgenic counterparts on the control diet who
exhibited less robust neuroinflammation. Accordingly, wildtype animals with apparently
unperturbed expression of hippocampal insulin signaling and neuroinflammatory genes
performed better than either transgenic group.
6.6 Implications & Directions for Future Research
The results of this thesis have implications for topics of future research aimed at understanding
the effects of diet quality and insulin activity on cognitive function. Foods (fish, fruits and
vegetables) and nutrients (saturated fat) shown to be separately associated with age-related
cognitive changes were found to be related to a broader set of dietary patterns which were
themselves related to cognition (Chapter 3). The nature of this finding implies that observational
studies examining the intake of single foods or food groups on cognition may actually be
capturing the influence of these broader patterns. The dietary pattern scores assigned to each
individual in Chapter 3 reflected total intake of the foods associated with each pattern weighted
by the correlation of each food with the overall pattern. For instance, it would be possible to
attain similarly high prudent pattern scores such that lesser consumption of fish could be offset
by greater consumption of leafy greens, and vice versa. Therefore, the diet pattern score reflects
the collective intake of the foods comprising the overall pattern. Although it would be unlikely
that a high dietary pattern score could be obtained by someone who avoided the top loaded
foods, the current studies cannot reveal whether any health related effects are due to the
collective influence of all the foods within the diet or extremely sensitive to a few or even one
food. Future attempts to unravel the collective effect of all foods loading onto a specific pattern
121
from the specific effect of certain foods may be informative not only in designing strategies for
maximizing retention of cognition with aging, but also in the fundamental understanding of how
diet influences health. The results of this thesis suggest that specific foods or food groups within
the diet may indeed moderate any overall effects on cognition. As previously mentioned,
adherence to the prudent pattern in Chapter 3 was associated with consumption of vegetables,
fruits and fish and related to better cognition whereas consumption of the whole-food diet
containing the same types of food by rodents in Chapter 4 resulted in cognitive dysfunction. It
was speculated that the high proportion of cruciferous vegetables in the WFD may have
contributed to placing extreme demands on a damaged cellular stress response system in the
transgenic mice. Although cruciferous vegetable intake was also related to adherence to the
prudent pattern in Chapter 3, it is interesting to speculate that the high proportion of cruciferous
vegetables may have contributed to discordant findings between Chapter 4 and Chapter 3 where
the top-loaded plants in the prudent dietary pattern were leafy greens.
Individual differences in environmental conditions, innate metabolism and/or genetics moderated
the effects of diet on cognitive performance in each study to varying degrees. Thus, further
research into the specific dimensions and underlying mechanisms of these effects may aid in the
development of dietary recommendations to maximize retention of cognitive function with
aging. In Chapter 4, only mice expressing human mutations for familial AD were adversely
influenced by the WFD. This transgene created an adverse brain environment which included
heightened neuroinflammation. In Chapter 5, diet-induced insulin resistance was most adversely
related to memory in those individuals with higher indications of adipose tissue inflammation.
Since living conditions and diet for these animals were strictly controlled, differences in adipose
tissue inflammation, and cognition, were also likely to be genetically driven. It would be
interesting to assess whether similar diets would be associated with cognitive deficits in human
participants carrying mutations related to familial AD, inflammatory status, and adipose tissue
remodeling. Analyses could also be conducted based on differences in plasma biomarkers
related to these same processes.
122
In contrast to the rodent studies, there were large differences in living conditions between
participants in Chapter 3. The analytical approach in that study was to statistically control for
these differences, and to assess the extent to which dietary effects on cognition depended on
some of these differences. As mentioned in Chapter 3, the dependency between diet quality and
socioeconomic indicators did not merely result from gradients in dietary intake. However,
socioeconomic gradients in dietary intake did exist in Chapter 3 such that adherence to the
prudent and Western patterns were positively and inversely related to SEP respectively. Thus,
individuals who were likely to be most affected by poor diet quality, were also those who may be
most resistant to interventions aimed at improving diet quality. Future research into strategies
and policies to improve diet quality in individuals with relatively low SEP would seem
warranted.
It was logistically not possible to test the effects of the whole-food and high fat diets in both rats
and transgenic mice, or to use pair-feeding to separate the effects of energy intake from diet per
se. Future studies could certainly take these approaches. It would be interesting to determine if
the observed effects on cognitive function were influenced by both the choice of animal model
and ‘control’ diet. The choice of control diet may have relevance to cognition as past studies in
transgenic mouse models of AD have found that DHA benefitted behaviour and markers of
synaptic function in comparison to control diets depleted in n-3 fatty acids [112,113] whereas a
DHA-enriched diets were shown to reduce Aβ deposition in comparison to normal chow and n-3
depleted diets [114]. The relationships between diet quality and cognition in Chapter 3 were
adjusted for differences in energy intake, but not macronutrient distribution. Conversely, the
diets compared in Chapter 4 were designed to have similar macronutrient distributions and
energy densities whereas the diets in Chapter 5 were neither macronutrient balanced or
isocaloric. It is unknown what effect these differences may have in the interpretation of the
results. In Chapter 5 the HFD-ROSI group exhibited greater body weight and energy intake than
the memory impaired HFD group, but performed no differently on the VIDA task than animals
consuming the hypocaloric control diet which acted as the cognitively intact reference group.
Thus, dietary composition seemed to matter more than level of consumption in this study.
However, it should be noted that rosiglitazone disconnected overconsumption of the high fat diet
123
from many of its adverse metabolic consequences—a situation that cannot be realistically
translated into recommendations for human populations. Since energy intake was statistically
adjusted for in Chapter 3, those results would suggest that dietary composition is at least as
important as differences in energy intake in determining cognitive performance.
The results of this thesis suggest that future studies should focus on the totality of metabolic
dysfunction accompanying insulin resistance and non-classical markers of insulin signaling when
assessing the potential role that functional insulin plays in mediating cognition. Although the
experiments in Chapters 4 and 5 partially support a role for brain insulin signaling in mediating
cognition, they did not directly observe diet-induced differences in molecular components of the
brain insulin signaling pathway. These results may have been limited by choices made in the
selection of specific targets and the methods used to assess them. For instance, Chapter 4 was
confined to gene expression analyses of widely distributed molecular components of the insulin
signaling pathway whereas Chapter 5 analyzed both gene expression and abundance of a smaller
group of protein targets. As mentioned in Chapter 5, studies in diabetic subjects have shown that
assessments of enzymatic activity can reveal differences in insulin-responsive cellular targets in
the absence of differences in absolute protein abundance. As discussed in Chapter 5,
downstream perturbations in the insulin-signaling pathway at the level of transcriptional
activation may have influenced cognition, and would be overlooked by studies that focus on
prototypical insulin signaling targets such as PI3k/Akt. Recent animal and pathological studies
have indicated that phosphorylation of specific residues on IRS1 are involved in mediating
neuronal insulin resistance and cognitive dysfunction in Alzheimer’s disease [209,212]. Thus,
assessing dietary effects on these specific targets may yield informative results. Chapter 5
implicated diet-induced metabolic dysfunction involving inflammation, insulin resistance, and
leptin resistance as a major determinant of hippocampal-dependent memory. We did not have
the opportunity to analyze additional plasma biomarkers such as adipokines and other pro-
inflammatory cytokines, but measurement of these additional targets may produce a composite
score of metabolic dysfunction with an even closer relationship with cognitive function. In
Chapter 3, we controlled for a number of metabolic indicators including waist circumference,
type-2 diabetes, and hypertension so that the relationships with diet could be interpreted as being
124
independent of these conditions. This interpretation may not reflect the true dimensions or
magnitude of the diet-cognition relationship, however, as diet could conceivably act on the brain
via disturbances associated with metabolic diseases. Formal assessments of how strong dietary
effects per se are from the effects of diet-related metabolic diseases would seem to be warranted.
6.7 Conclusions
In summary, the results this thesis suggest that the associations of fish, fruit, vegetables and
saturated fat with cognitive function reflect adherence to a broader set of dietary patterns whose
own relationship with cognition may be dependent on individual differences in environmental
conditions, innate metabolism, and/or genetics. Although markers of brain insulin signaling
were related to optimal cognitive function in rodent studies, diet-induced modulation of the
hippocampal insulin signaling pathway was, at best, indirect. Diet-induced, peripheral insulin
resistance and metabolic dysfunction were closely related to cognitive deficits. The following
specific conclusions were made:
1. Consumption of fish, fruits, vegetables, and saturated fat was associated with adherence to a
broader set of dietary patterns in older adults (Chapter 3). Furthermore, the relationships
between these dietary patterns and cognition were in the expected directions based on their
association with those separate dietary components. However, the relationships between
dietary patterns and cognition were dependent on indicators of socioeconomic position which
are known to exert their own effects on cognition and brain development.
2. The studies in rodents (Chapter 5) and older adults (Chapter 3) are similar in their findings
that diets associated with relatively high saturated fat intake were related to worse cognitive
function.
3. A combination fish, vegetable, and fruit diet unexpectedly worsened cognitive dysfunction in
a transgenic mouse model of familial AD (Chapter 4). Cognitive function in wildtype mice,
lacking genetically determined neuropathology, was unaffected by the dietary combination.
125
In contrast, a dietary pattern associated with consumption of vegetables, fruit, and fish was
related to better cognitive function in older adults (Chapter 3) for whom we do not currently
have any genetic information.
4. Although markers of hippocampal insulin signaling were related to optimal cognitive
function in animal studies, diet-induced modulation of the hippocampal insulin signaling
pathway was, at best, indirect (Chapters 4 and 5).
5. Diet-induced changes in peripheral insulin resistance and metabolic dysfunction was related
to worse hippocampal-dependent memory in rats (Chapter 5).
6. Dietary modulation of inflammatory pathways play an important role in modifying the
influence of peripheral insulin sensitivity (Chapter 5), and amplifying the effect of reduced
hippocampal insulin signaling (Chapter 4), on cognitive function.
7. Collectively, these studies suggest that the relationship between diet quality and cognitive
function is dependent on individual differences in environmental conditions, innate
metabolism, and/or genetics.
126
References
1. Winocur G, Greenwood CE (2005) Studies of the effects of high fat diets on cognitive function in a rat model. Neurobiol Aging 26 Suppl 1: 46-49.
2. Craft S (2005) Insulin resistance syndrome and Alzheimer's disease: age- and obesity-related effects on memory, amyloid, and inflammation. Neurobiol Aging 26: 65-69. Epub 2005 Nov 2002.
3. Craft S, Watson GS (2004) Insulin and neurodegenerative disease: shared and specific mechanisms. Lancet Neurol 3: 169-178.
4. Zhao WQ, Alkon DL (2001) Role of insulin and insulin receptor in learning and memory. Mol Cell Endocrinol 177: 125-134.
5. Zhao W, Chen H, Xu H, Moore E, Meiri N, et al. (1999) Brain insulin receptors and spatial memory. Correlated changes in gene expression, tyrosine phosphorylation, and signaling molecules in the hippocampus of water maze trained rats. J Biol Chem 274: 34893-34902.
6. Zhao WQ, Alkon DL (2001) Role of insulin and insulin receptor in learning and memory. Mol Cell Endocrinol 177: 125-134.
7. Devaskar SU, Giddings SJ, Rajakumar PA, Carnaghi LR, Menon RK, et al. (1994) Insulin gene expression and insulin synthesis in mammalian neuronal cells. J Biol Chem 269: 8445-8454.
8. Rulifson EJ, Kim SK, Nusse R (2002) Ablation of insulin-producing neurons in flies: growth and diabetic phenotypes. Science 296: 1118-1120.
9. Schechter R, Holtzclaw L, Sadiq F, Kahn A, Devaskar S (1988) Insulin synthesis by isolated rabbit neurons. Endocrinology 123: 505-513.
10. Foulstone EJ, Tavare JM, Gunn-Moore FJ (1999) Sustained phosphorylation and activation of protein kinase B correlates with brain-derived neurotrophic factor and insulin stimulated survival of cerebellar granule cells. Neurosci Lett 264: 125-128.
11. Orban PC, Chapman PF, Brambilla R (1999) Is the Ras-MAPK signalling pathway necessary for long-term memory formation? Trends Neurosci 22: 38-44.
12. Ryu BR, Ko HW, Jou I, Noh JS, Gwag BJ (1999) Phosphatidylinositol 3-kinase-mediated regulation of neuronal apoptosis and necrosis by insulin and IGF-I. J Neurobiol 39: 536-546.
127
13. Yang H, Raizada MK (1999) Role of phosphatidylinositol 3-kinase in angiotensin II regulation of norepinephrine neuromodulation in brain neurons of the spontaneously hypertensive rat. J Neurosci 19: 2413-2423.
14. Alkon DL, Epstein H, Kuzirian A, Bennett MC, Nelson TJ (2005) Protein synthesis required for long-term memory is induced by PKC activation on days before associative learning. Proc Natl Acad Sci U S A 102: 16432-16437. Epub 12005 Oct 16428.
15. Kahn CR, White MF, Shoelson SE, Backer JM, Araki E, et al. (1993) The insulin receptor and its substrate: molecular determinants of early events in insulin action. Recent Prog Horm Res 48: 291-339.
16. Bank B, DeWeer A, Kuzirian AM, Rasmussen H, Alkon DL (1988) Classical conditioning induces long-term translocation of protein kinase C in rabbit hippocampal CA1 cells. Proc Natl Acad Sci U S A 85: 1988-1992.
17. Olds JL, Anderson ML, McPhie DL, Staten LD, Alkon DL (1989) Imaging of memory-specific changes in the distribution of protein kinase C in the hippocampus. Science 245: 866-869.
18. Havrankova J, Schmechel D, Roth J, Brownstein M (1978) Identification of insulin in rat brain. Proc Natl Acad Sci U S A 75: 5737-5741.
19. Korol DL, Gold PE (1998) Glucose, memory, and aging. Am J Clin Nutr 67: 764S-771S.
20. Park CR, Seeley RJ, Craft S, Woods SC (2000) Intracerebroventricular insulin enhances memory in a passive-avoidance task. Physiol Behav 68: 509-514.
21. Craft S, Asthana S, Newcomer JW, Wilkinson CW, Matos IT, et al. (1999) Enhancement of memory in Alzheimer disease with insulin and somatostatin, but not glucose. Arch Gen Psychiatry 56: 1135-1140.
22. Reger MA, Watson GS, Frey WH, 2nd, Baker LD, Cholerton B, et al. (2006) Effects of intranasal insulin on cognition in memory-impaired older adults: Modulation by APOE genotype. Neurobiol Aging 27: 451-458. Epub 2005 Jun 2016.
23. Baura GD, Foster DM, Kaiyala K, Porte D, Jr., Kahn SE, et al. (1996) Insulin transport from plasma into the central nervous system is inhibited by dexamethasone in dogs. Diabetes 45: 86-90.
24. Wallum BJ, Taborsky GJ, Jr., Porte D, Jr., Figlewicz DP, Jacobson L, et al. (1987) Cerebrospinal fluid insulin levels increase during intravenous insulin infusions in man. J Clin Endocrinol Metab 64: 190-194.
25. Awad N, Gagnon M, Messier C (2004) The relationship between impaired glucose tolerance, type 2 diabetes, and cognitive function. J Clin Exp Neuropsychol 26: 1044-1080.
128
26. Kaplan RJ, Greenwood CE, Winocur G, Wolever TM (2000) Cognitive performance is associated with glucose regulation in healthy elderly persons and can be enhanced with glucose and dietary carbohydrates. American Journal of Clinical Nutrition 72: 825-836.
27. Greenwood CE, Kaplan RJ, Hebblethwaite S, Jenkins DJ (2003) Carbohydrate-induced memory impairment in adults with type 2 diabetes. Diabetes Care 26: 1961-1966.
28. Papanikoloau Y, Palmer H, Binns MA, Jenkins DJA, Greenwood CE (2006) Better cognitive performance following a low-glycaemic-index compared with a high-glycaemic-index carbohydrate meal
in adults with type 2 diabetes. Diabetologia 49: 855–862.
29. Winocur G, Greenwood CE, Piroli GG, Grillo CA, Reznikov LR, et al. (2005) Memory impairment in obese Zucker rats: an investigation of cognitive function in an animal model of insulin resistance and obesity. Behav Neurosci 119: 1389-1395.
30. Craft S, Asthana S, Cook DG, Baker LD, Cherrier M, et al. (2003) Insulin dose-response effects on memory and plasma amyloid precursor protein in Alzheimer's disease: interactions with apolipoprotein E genotype. Psychoneuroendocrinology 28: 809-822.
31. Craft S, Peskind E, Schwartz MW, Schellenberg GD, Raskind M, et al. (1998) Cerebrospinal fluid and plasma insulin levels in Alzheimer's disease: relationship to severity of dementia and apolipoprotein E genotype. Neurology 50: 164-168.
32. Frolich L, Blum-Degen D, Bernstein HG, Engelsberger S, Humrich J, et al. (1998) Brain insulin and insulin receptors in aging and sporadic Alzheimer's disease. J Neural Transm 105: 423-438.
33. Rivera EJ, Goldin A, Fulmer N, Tavares R, Wands JR, et al. (2005) Insulin and insulin-like growth factor expression and function deteriorate with progression of Alzheimer's disease: link to brain reductions in acetylcholine. J Alzheimers Dis 8: 247-268.
34. Oddo S, Caccamo A, Shepherd JD, Murphy MP, Golde TE, et al. (2003) Triple-transgenic model of Alzheimer's disease with plaques and tangles: intracellular Abeta and synaptic dysfunction. Neuron 39: 409-421.
35. Authier F, Posner BI, Bergeron JJ (1996) Insulin-degrading enzyme. Clin Invest Med 19: 149-160.
36. Farris W, Mansourian S, Chang Y, Lindsley L, Eckman EA, et al. (2003) Insulin-degrading enzyme regulates the levels of insulin, amyloid beta-protein, and the beta-amyloid precursor protein intracellular domain in vivo. Proc Natl Acad Sci U S A 100: 4162-4167. Epub 2003 Mar 4112.
129
37. Farris W, Mansourian S, Leissring MA, Eckman EA, Bertram L, et al. (2004) Partial loss-of-function mutations in insulin-degrading enzyme that induce diabetes also impair degradation of amyloid beta-protein. Am J Pathol 164: 1425-1434.
38. Vekrellis K, Ye Z, Qiu WQ, Walsh D, Hartley D, et al. (2000) Neurons regulate extracellular levels of amyloid beta-protein via proteolysis by insulin-degrading enzyme. J Neurosci 20: 1657-1665.
39. Carro E, Trejo JL, Gomez-Isla T, LeRoith D, Torres-Aleman I (2002) Serum insulin-like growth factor I regulates brain amyloid-beta levels. Nat Med 8: 1390-1397. Epub 2002 Nov 1394.
40. Gasparini L, Xu H (2003) Potential roles of insulin and IGF-1 in Alzheimer's disease. Trends Neurosci 26: 404-406.
41. Nestler JE, Barlascini CO, Tetrault GA, Fratkin MJ, Clore JN, et al. (1990) Increased transcapillary escape rate of albumin in nondiabetic men in response to hyperinsulinemia. Diabetes 39: 1212-1217.
42. Stein TD, Johnson JA (2002) Lack of neurodegeneration in transgenic mice overexpressing mutant amyloid precursor protein is associated with increased levels of transthyretin and the activation of cell survival pathways. J Neurosci 22: 7380-7388.
43. Yuan J, Yankner BA (2000) Apoptosis in the nervous system. Nature 407: 802-809.
44. Galvan V, Chen S, Lu D, Logvinova A, Goldsmith P, et al. (2002) Caspase cleavage of members of the amyloid precursor family of proteins. J Neurochem 82: 283-294.
45. Gervais FG, Xu D, Robertson GS, Vaillancourt JP, Zhu Y, et al. (1999) Involvement of caspases in proteolytic cleavage of Alzheimer's amyloid-beta precursor protein and amyloidogenic A beta peptide formation. Cell 97: 395-406.
46. Lu DC, Rabizadeh S, Chandra S, Shayya RF, Ellerby LM, et al. (2000) A second cytotoxic proteolytic peptide derived from amyloid beta-protein precursor. Nat Med 6: 397-404.
47. Su JH, Kesslak JP, Head E, Cotman CW (2002) Caspase-cleaved amyloid precursor protein and activated caspase-3 are co-localized in the granules of granulovacuolar degeneration in Alzheimer's disease and Down's syndrome brain. Acta Neuropathol (Berl) 104: 1-6. Epub 2002 May 2009.
48. Dumanchin-Njock C, Alves da Costa CA, Mercken L, Pradier L, Checler F (2001) The caspase-derived C-terminal fragment of betaAPP induces caspase-independent toxicity and triggers selective increase of Abeta42 in mammalian cells. J Neurochem 78: 1153-1161.
130
49. Tesco G, Koh YH, Tanzi RE (2003) Caspase activation increases beta-amyloid generation independently of caspase cleavage of the beta-amyloid precursor protein (APP). J Biol Chem 278: 46074-46080. Epub 42003 Sep 46075.
50. Marin N, Romero B, Bosch-Morell F, Llansola M, Felipo V, et al. (2000) Beta-amyloid-induced activation of caspase-3 in primary cultures of rat neurons. Mech Ageing Dev 119: 63-67.
51. Uetsuki T, Takemoto K, Nishimura I, Okamoto M, Niinobe M, et al. (1999) Activation of neuronal caspase-3 by intracellular accumulation of wild-type Alzheimer amyloid precursor protein. J Neurosci 19: 6955-6964.
52. Kar S, Slowikowski SP, Westaway D, Mount HT (2004) Interactions between beta-amyloid and central cholinergic neurons: implications for Alzheimer's disease. Journal of Psychiatry & Neuroscience 29: 427-441.
53. Arriagada PV, Growdon JH, Hedley-Whyte ET, Hyman BT (1992) Neurofibrillary tangles but not senile plaques parallel duration and severity of Alzheimer's disease. Neurology 42: 631-639.
54. Hong M, Lee VM (1997) Insulin and insulin-like growth factor-1 regulate tau phosphorylation in cultured human neurons. J Biol Chem 272: 19547-19553.
55. Schechter R, Beju D, Miller KE (2005) The effect of insulin deficiency on tau and neurofilament in the insulin knockout mouse. Biochem Biophys Res Commun 334: 979-986.
56. Schechter R, Beju D, Miller KE (2005) Insulin deficiency induces neuron cell death in knock-out mice by hyperphosphorylation of JNK. Program No 2522 Abstract Viewer/Itinerary Planner. Washington, DC: Society for Neuroscience.
57. Di Iorio A, Cherubini A, Volpato S, Sparvieri E, Lauretani F, et al. (2006) Markers of inflammation, vitamin E and peripheral nervous system function: the InCHIANTI study. Neurobiol Aging 27: 1280-1288.
58. Weaver JD, Huang MH, Albert M, Harris T, Rowe JW, et al. (2002) Interleukin-6 and risk of cognitive decline: MacArthur studies of successful aging. Neurology 59: 371-378.
59. Wright CB, Sacco RL, Rundek TR, Delman JB, Rabbani LE, et al. (2006) Interleukin-6 is associated with cognitive function: the Northern Manhattan Study. J Stroke Cerebrovasc Dis 15: 34-38.
60. Odegaard JI, Chawla A (2013) Pleiotropic actions of insulin resistance and inflammation in metabolic homeostasis. Science 339: 172-177.
131
61. Dimopoulos N, Piperi C, Salonicioti A, Mitropoulos P, Kallai E, et al. (2006) Indices of low-grade chronic inflammation correlate with early cognitive deterioration in an elderly Greek population. Neurosci Lett 398: 118-123.
62. Chui MH, Papanikolaou Y, Fontaine-Bisson B, Turcotte J, Wolever TM, et al. (2007) The TNF-alpha-238G > a single-nucleotide polymorphism protects against memory decline in older adults with type 2 diabetes. Behav Neurosci 121: 619-624.
63. Yaffe K, Kanaya A, Lindquist K, Simonsick EM, Harris T, et al. (2004) The metabolic syndrome, inflammation, and risk of cognitive decline. JAMA 292: 2237-2242.
64. Caballero AE (2003) Endothelial dysfunction in obesity and insulin resistance: a road to diabetes and heart disease. Obes Res 11: 1278-1289.
65. Fishel MA, Watson GS, Montine TJ, Wang Q, Green PS, et al. (2005) Hyperinsulinemia provokes synchronous increases in central inflammation and beta-amyloid in normal adults. Arch Neurol 62: 1539-1544.
66. Behl C, Davis JB, Lesley R, Schubert D (1994) Hydrogen peroxide mediates amyloid beta protein toxicity. Cell 77: 817-827.
67. Misonou H, Morishima-Kawashima M, Ihara Y (2000) Oxidative stress induces intracellular accumulation of amyloid beta-protein (Abeta) in human neuroblastoma cells. Biochemistry 39: 6951-6959.
68. Maher FO, Martin DS, Lynch MA (2004) Increased IL-1beta in cortex of aged rats is accompanied by downregulation of ERK and PI-3 kinase. Neurobiol Aging 25: 795-806.
69. Mattson MP, Maudsley S, Martin B (2004) BDNF and 5-HT: a dynamic duo in age-related neuronal plasticity and neurodegenerative disorders. Trends Neurosci 27: 589-594.
70. Duan W, Guo Z, Jiang H, Ware M, Mattson MP (2003) Reversal of behavioral and metabolic abnormalities, and insulin resistance syndrome, by dietary restriction in mice deficient in brain-derived neurotrophic factor. Endocrinology 144: 2446-2453.
71. Lesne S, Gabriel C, Nelson DA, White E, Mackenzie ET, et al. (2005) Akt-dependent expression of NAIP-1 protects neurons against amyloid-{beta} toxicity. J Biol Chem 280: 24941-24947. Epub 22005 Mar 24929.
72. Yu Y, Kastin AJ, Pan W (2006) Reciprocal interactions of insulin and insulin-like growth factor I in receptor-mediated transport across the blood-brain barrier. Endocrinology 147: 2611-2615.
73. Laurino L, Wang XX, de la Houssaye BA, Sosa L, Dupraz S, et al. (2005) PI3K activation by IGF-1 is essential for the regulation of membrane expansion at the nerve growth cone. J Cell Sci 118: 3653-3662.
132
74. Pfenninger KH, Laurino L, Peretti D, Wang X, Rosso S, et al. (2003) Regulation of membrane expansion at the nerve growth cone. J Cell Sci 116: 1209-1217.
75. Aizenman Y, de Vellis J (1987) Brain neurons develop in a serum and glial free environment: effects of transferrin, insulin, insulin-like growth factor-I and thyroid hormone on neuronal survival, growth and differentiation. Brain Res 406: 32-42.
76. Caroni P, Grandes P (1990) Nerve sprouting in innervated adult skeletal muscle induced by exposure to elevated levels of insulin-like growth factors. J Cell Biol 110: 1307-1317.
77. Dupont J, LeRoith D (2001) Insulin and insulin-like growth factor I receptors: similarities and differences in signal transduction. Horm Res 55 Suppl 2: 22-26.
78. Pandini G, Frasca F, Mineo R, Sciacca L, Vigneri R, et al. (2002) Insulin/insulin-like growth factor I hybrid receptors have different biological characteristics depending on the insulin receptor isoform involved. J Biol Chem 277: 39684-39695.
79. Hansen LL, Ikeda Y, Olsen GS, Busch AK, Mosthaf L (1999) Insulin signaling is inhibited by micromolar concentrations of H(2)O(2). Evidence for a role of H(2)O(2) in tumor necrosis factor alpha-mediated insulin resistance. J Biol Chem 274: 25078-25084.
80. Maher FO, Clarke RM, Kelly A, Nally RE, Lynch MA (2006) Interaction between interferon gamma and insulin-like growth factor-1 in hippocampus impacts on the ability of rats to sustain long-term potentiation. J Neurochem 96: 1560-1571.
81. Winocur G, Wojtowicz JM, Sekeres M, Snyder JS, Wang S (2006) Inhibition of neurogenesis interferes with hippocampus-dependent memory function. Hippocampus 16: 296-304.
82. Barberger-Gateau P, Letenneur L, Deschamps V, Peres K, Dartigues JF, et al. (2002) Fish, meat, and risk of dementia: cohort study. BMJ 325: 932-933.
83. Kalmijn S, Launer LJ, Ott A, Witteman JC, Hofman A, et al. (1997) Dietary fat intake and the risk of incident dementia in the Rotterdam Study. Ann Neurol 42: 776-782.
84. Kalmijn S, van Boxtel MP, Ocke M, Verschuren WM, Kromhout D, et al. (2004) Dietary intake of fatty acids and fish in relation to cognitive performance at middle age. Neurology 62: 275-280.
85. Morris MC, Evans DA, Bienias JL, Tangney CC, Bennett DA, et al. (2003) Consumption of fish and n-3 fatty acids and risk of incident Alzheimer disease. Arch Neurol 60: 940-946.
86. Morris MC, Evans DA, Tangney CC, Bienias JL, Wilson RS (2005) Fish consumption and cognitive decline with age in a large community study. Arch Neurol 62: 1849-1853. Epub 2005 Oct 1810.
133
87. Kang JH, Ascherio A, Grodstein F (2005) Fruit and vegetable consumption and cognitive decline in aging women. Ann Neurol 57: 713-720.
88. Morris MC, Evans DA, Tangney CC, Bienias JL, Wilson RS (2006) Associations of vegetable and fruit consumption with age-related cognitive change. Neurology 67: 1370-1376.
89. Ortega RM, Requejo AM, Andres P, Lopez-Sobaler AM, Quintas ME, et al. (1997) Dietary intake and cognitive function in a group of elderly people. American Journal of Clinical Nutrition 66: 803-809.
90. Morris MC, Evans DA, Bienias JL, Tangney CC, Bennett DA, et al. (2003) Dietary fats and the risk of incident Alzheimer disease.[erratum appears in Arch Neurol. 2003 Aug;60(8):1072]. Archives of Neurology 60: 194-200.
91. Newman AB, Fitzpatrick AL, Lopez O, Jackson S, Lyketsos C, et al. (2005) Dementia and Alzheimer's disease incidence in relationship to cardiovascular disease in the Cardiovascular Health Study cohort. J Am Geriatr Soc 53: 1101-1107.
92. Arvanitakis Z, Wilson RS, Bienias JL, Evans DA, Bennett DA (2004) Diabetes mellitus and risk of Alzheimer disease and decline in cognitive function. Arch Neurol 61: 661-666.
93. Luchsinger JA, Tang MX, Shea S, Mayeux R (2004) Hyperinsulinemia and risk of Alzheimer disease. Neurology 63: 1187-1192.
94. Ott A, Stolk RP, van Harskamp F, Pols HA, Hofman A, et al. (1999) Diabetes mellitus and the risk of dementia: The Rotterdam Study. Neurology 53: 1937-1942.
95. Peila R, Rodriguez BL, Launer LJ (2002) Type 2 diabetes, APOE gene, and the risk for dementia and related pathologies: The Honolulu-Asia Aging Study. Diabetes 51: 1256-1262.
96. Riccardi G, Giacco R, Rivellese AA (2004) Dietary fat, insulin sensitivity and the metabolic syndrome. Clin Nutr 23: 447-456.
97. Storlien LH, Jenkins AB, Chisholm DJ, Pascoe WS, Khouri S, et al. (1991) Influence of dietary fat composition on development of insulin resistance in rats. Relationship to muscle triglyceride and omega-3 fatty acids in muscle phospholipid. Diabetes 40: 280-289.
98. Winocur G, Greenwood CE (2005) Studies of the effects of high fat diets on cognitive function in a rat model. Neurobiol Aging 26: 46-49. Epub 2005 Oct 2010.
99. Ho L, Qin W, Pompl PN, Xiang Z, Wang J, et al. (2004) Diet-induced insulin resistance promotes amyloidosis in a transgenic mouse model of Alzheimer's disease. FASEB J 18: 902-904. Epub 2004 Mar 2019.
134
100. Molteni R, Wu A, Vaynman S, Ying Z, Barnard RJ, et al. (2004) Exercise reverses the harmful effects of consumption of a high-fat diet on synaptic and behavioral plasticity associated to the action of brain-derived neurotrophic factor. Neuroscience 123: 429-440.
101. Zhang X, Dong F, Ren J, Driscoll MJ, Culver B (2005) High dietary fat induces NADPH oxidase-associated oxidative stress and inflammation in rat cerebral cortex. Exp Neurol 191: 318-325.
102. Bourre JM (2004) Roles of unsaturated fatty acids (especially omega-3 fatty acids) in the brain at various ages and during ageing. J Nutr Health Aging 8: 163-174.
103. Simopoulos AP (1991) Omega-3 fatty acids in health and disease and in growth and development. Am J Clin Nutr 54: 438-463.
104. Simopoulos AP (1999) Essential fatty acids in health and chronic disease. Am J Clin Nutr 70: 560S-569S.
105. Young G, Conquer J (2005) Omega-3 fatty acids and neuropsychiatric disorders. Reprod Nutr Dev 45: 1-28.
106. Moore SA (1993) Cerebral endothelium and astrocytes cooperate in supplying docosahexaenoic acid to neurons. Adv Exp Med Biol 331: 229-233.
107. Moore SA, Yoder E, Murphy S, Dutton GR, Spector AA (1991) Astrocytes, not neurons, produce docosahexaenoic acid (22:6 omega-3) and arachidonic acid (20:4 omega-6). J Neurochem 56: 518-524.
108. Horrocks LA, Farooqui AA (2004) Docosahexaenoic acid in the diet: its importance in maintenance and restoration of neural membrane function. Prostaglandins Leukot Essent Fatty Acids 70: 361-372.
109. Barcelo-Coblijn G, Hogyes E, Kitajka K, Puskas LG, Zvara A, et al. (2003) Modification by docosahexaenoic acid of age-induced alterations in gene expression and molecular composition of rat brain phospholipids. Proc Natl Acad Sci U S A 100: 11321-11326.
110. Catalan J, Moriguchi T, Slotnick B, Murthy M, Greiner RS, et al. (2002) Cognitive deficits in docosahexaenoic acid-deficient rats. Behav Neurosci 116: 1022-1031.
111. Dyall SC, Michael GJ, Whelpton R, Scott AG, Michael-Titus AT (2006) Dietary enrichment with omega-3 polyunsaturated fatty acids reverses age-related decreases in the GluR2 and NR2B glutamate receptor subunits in rat forebrain. Neurobiol Aging.
112. Calon F, Lim GP, Morihara T, Yang F, Ubeda O, et al. (2005) Dietary n-3 polyunsaturated fatty acid depletion activates caspases and decreases NMDA receptors in the brain of a transgenic mouse model of Alzheimer's disease. Eur J Neurosci 22: 617-626.
135
113. Calon F, Lim GP, Yang F, Morihara T, Teter B, et al. (2004) Docosahexaenoic acid protects from dendritic pathology in an Alzheimer's disease mouse model. Neuron 43: 633-645.
114. Lim GP, Calon F, Morihara T, Yang F, Teter B, et al. (2005) A diet enriched with the omega-3 fatty acid docosahexaenoic acid reduces amyloid burden in an aged Alzheimer mouse model. J Neurosci 25: 3032-3040.
115. Akbar M, Calderon F, Wen Z, Kim HY (2005) Docosahexaenoic acid: a positive modulator of Akt signaling in neuronal survival. Proc Natl Acad Sci U S A 102: 10858-10863. Epub 12005 Jul 10822.
116. Calderon F, Kim HY (2004) Docosahexaenoic acid promotes neurite growth in hippocampal neurons. J Neurochem 90: 979-988.
117. Lukiw WJ, Cui JG, Marcheselli VL, Bodker M, Botkjaer A, et al. (2005) A role for docosahexaenoic acid-derived neuroprotectin D1 in neural cell survival and Alzheimer disease. J Clin Invest 115: 2774-2783.
118. Marcheselli VL, Hong S, Lukiw WJ, Tian XH, Gronert K, et al. (2003) Novel docosanoids inhibit brain ischemia-reperfusion-mediated leukocyte infiltration and pro-inflammatory gene expression. J Biol Chem 278: 43807-43817. Epub 42003 Aug 43815.
119. Cao D, Zhou C, Sun L, Xue R, Xu J, et al. (2005) Chronic administration of ethyl docosahexaenoate reduces gerbil brain eicosanoid productions following ischemia and reperfusion. J Nutr Biochem 9: 9.
120. Taouis M, Dagou C, Ster C, Durand G, Pinault M, et al. (2002) N-3 polyunsaturated fatty acids prevent the defect of insulin receptor signaling in muscle. Am J Physiol Endocrinol Metab 282: E664-671.
121. Storlien LH, Kraegen EW, Chisholm DJ, Ford GL, Bruce DG, et al. (1987) Fish oil prevents insulin resistance induced by high-fat feeding in rats. Science 237: 885-888.
122. Engelhart MJ, Geerlings MI, Ruitenberg A, van Swieten JC, Hofman A, et al. (2002) Dietary intake of antioxidants and risk of Alzheimer disease. JAMA 287: 3223-3229.
123. Hu P, Bretsky P, Crimmins EM, Guralnik JM, Reuben DB, et al. (2006) Association between serum beta-carotene levels and decline of cognitive function in high-functioning older persons with or without apolipoprotein E 4 alleles: MacArthur studies of successful aging. J Gerontol A Biol Sci Med Sci 61: 616-620.
124. Morris MC, Evans DA, Bienias JL, Tangney CC, Bennett DA, et al. (2002) Dietary intake of antioxidant nutrients and the risk of incident Alzheimer disease in a biracial community study. Jama 287: 3230-3237.
136
125. Zandi PP, Anthony JC, Khachaturian AS, Stone SV, Gustafson D, et al. (2004) Reduced risk of Alzheimer disease in users of antioxidant vitamin supplements: the Cache County Study. Arch Neurol 61: 82-88.
126. Kaltschmidt B, Uherek M, Volk B, Baeuerle PA, Kaltschmidt C (1997) Transcription factor NF-kappaB is activated in primary neurons by amyloid beta peptides and in neurons surrounding early plaques from patients with Alzheimer disease. Proc Natl Acad Sci U S A 94: 2642-2647.
127. Shi Q, Gibson GE (2007) Oxidative stress and transcriptional regulation in Alzheimer disease. Alzheimer Dis Assoc Disord 21: 276-291.
128. Scalbert A, Johnson IT, Saltmarsh M (2005) Polyphenols: antioxidants and beyond. Am J Clin Nutr 81: 215S-217S.
129. Huxley RR, Neil HA (2003) The relation between dietary flavonol intake and coronary heart disease mortality: a meta-analysis of prospective cohort studies. Eur J Clin Nutr 57: 904-908.
130. Knekt P, Kumpulainen J, Jarvinen R, Rissanen H, Heliovaara M, et al. (2002) Flavonoid intake and risk of chronic diseases. Am J Clin Nutr 76: 560-568.
131. Williamson G, Manach C (2005) Bioavailability and bioefficacy of polyphenols in humans. II. Review of 93 intervention studies. Am J Clin Nutr 81: 243S-255S.
132. Mandel SA, Avramovich-Tirosh Y, Reznichenko L, Zheng H, Weinreb O, et al. (2005) Multifunctional activities of green tea catechins in neuroprotection. Modulation of cell survival genes, iron-dependent oxidative stress and PKC signaling pathway. Neurosignals 14: 46-60.
133. Weinreb O, Mandel S, Amit T, Youdim MB (2004) Neurological mechanisms of green tea polyphenols in Alzheimer's and Parkinson's diseases. J Nutr Biochem 15: 506-516.
134. Joseph JA, Shukitt-Hale B, Casadesus G (2005) Reversing the deleterious effects of aging on neuronal communication and behavior: beneficial properties of fruit polyphenolic compounds. Am J Clin Nutr 81: 313S-316S.
135. Lau FC, Shukitt-Hale B, Joseph JA (2005) The beneficial effects of fruit polyphenols on brain aging. Neurobiol Aging 26: 128-132. Epub 2005 Sep 2027.
136. Joseph JA, Denisova NA, Arendash G, Gordon M, Diamond D, et al. (2003) Blueberry supplementation enhances signaling and prevents behavioral deficits in an Alzheimer disease model. Nutr Neurosci 6: 153-162.
137
137. Casadesus G, Shukitt-Hale B, Stellwagen HM, Zhu X, Lee HG, et al. (2004) Modulation of hippocampal plasticity and cognitive behavior by short-term blueberry supplementation in aged rats. Nutr Neurosci 7: 309-316.
138. Lim GP, Chu T, Yang F, Beech W, Frautschy SA, et al. (2001) The curry spice curcumin reduces oxidative damage and amyloid pathology in an Alzheimer transgenic mouse. J Neurosci 21: 8370-8377.
139. Yang F, Lim GP, Begum AN, Ubeda OJ, Simmons MR, et al. (2005) Curcumin inhibits formation of amyloid beta oligomers and fibrils, binds plaques, and reduces amyloid in vivo. J Biol Chem 280: 5892-5901. Epub 2004 Dec 5897.
140. Scarmeas N, Stern Y, Tang MX, Mayeux R, Luchsinger JA (2006) Mediterranean diet and risk for Alzheimer's disease. Ann Neurol 59: 912-921.
141. Chen CY, Milbury PE, Kwak HK, Collins FW, Samuel P, et al. (2004) Avenanthramides and phenolic acids from oats are bioavailable and act synergistically with vitamin C to enhance hamster and human LDL resistance to oxidation. J Nutr 134: 1459-1466.
142. Chen CY, Milbury PE, Lapsley K, Blumberg JB (2005) Flavonoids from almond skins are bioavailable and act synergistically with vitamins C and E to enhance hamster and human LDL resistance to oxidation. J Nutr 135: 1366-1373.
143. Filipe P, Lanca V, Silva JN, Morliere P, Santus R, et al. (2001) Flavonoids and urate antioxidant interplay in plasma oxidative stress. Mol Cell Biochem 221: 79-87.
144. Hwang J, Sevanian A, Hodis HN, Ursini F (2000) Synergistic inhibition of LDL oxidation by phytoestrogens and ascorbic acid. Free Radic Biol Med 29: 79-89.
145. Laranjinha J, Vieira O, Madeira V, Almeida L (1995) Two related phenolic antioxidants with opposite effects on vitamin E content in low density lipoproteins oxidized by ferrylmyoglobin: consumption vs regeneration. Arch Biochem Biophys 323: 373-381.
146. Rossetto M, Vanzani P, Mattivi F, Lunelli M, Scarpa M, et al. (2002) Synergistic antioxidant effect of catechin and malvidin 3-glucoside on free radical-initiated peroxidation of linoleic acid in micelles. Arch Biochem Biophys 408: 239-245.
147. Davis JN, Hodges VA, Gillham MB (2006) Normal-weight adults consume more fiber and fruit than their age- and height-matched overweight/obese counterparts. J Am Diet Assoc 106: 833-840.
148. Nelson LH, Tucker LA (1996) Diet composition related to body fat in a multivariate study of 203 men. J Am Diet Assoc 96: 771-777.
138
149. Barberger-Gateau P, Jutand MA, Letenneur L, Larrieu S, Tavernier B, et al. (2005) Correlates of regular fish consumption in French elderly community dwellers: data from the Three-City study. Eur J Clin Nutr 59: 817-825.
150. Shirai N, Suzuki H (2004) Effect of dietary docosahexaenoic acid and catechins on maze behavior in mice. Ann Nutr Metab 48: 51-58.
151. Alles B, Samieri C, Feart C, Jutand MA, Laurin D, et al. (2012) Dietary patterns: a novel approach to examine the link between nutrition and cognitive function in older individuals. Nutr Res Rev 25: 207-222.
152. Barberger-Gateau P, Raffaitin C, Letenneur L, Berr C, Tzourio C, et al. (2007) Dietary patterns and risk of dementia: the Three-City cohort study. Neurology 69: 1921-1930.
153. Feart C, Samieri C, Rondeau V, Amieva H, Portet F, et al. (2009) Adherence to a Mediterranean diet, cognitive decline, and risk of dementia. JAMA 302: 638-648.
154. Gu Y, Nieves JW, Stern Y, Luchsinger JA, Scarmeas N (2010) Food combination and Alzheimer disease risk: a protective diet. Arch Neurol 67: 699-706.
155. Kesse-Guyot E, Andreeva VA, Jeandel C, Ferry M, Hercberg S, et al. (2012) A healthy dietary pattern at midlife is associated with subsequent cognitive performance. J Nutr 142: 909-915.
156. Tangney CC, Kwasny MJ, Li H, Wilson RS, Evans DA, et al. (2011) Adherence to a Mediterranean-type dietary pattern and cognitive decline in a community population. Am J Clin Nutr 93: 601-607.
157. Wengreen HJ, Neilson C, Munger R, Corcoran C (2009) Diet quality is associated with better cognitive test performance among aging men and women. J Nutr 139: 1944-1949. Epub 2009 Aug 1912.
158. Aggarwal A, Monsivais P, Cook AJ, Drewnowski A (2011) Does diet cost mediate the relation between socioeconomic position and diet quality? Eur J Clin Nutr 65: 1059-1066.
159. Haan MN, Zeki Al-Hazzouri A, Aiello AE (2011) Life-span socioeconomic trajectory, nativity, and cognitive aging in Mexican Americans: the Sacramento Area Latino Study on Aging. J Gerontol B Psychol Sci Soc Sci 66: i102-i110.
160. Karlamangla AS, Miller-Martinez D, Aneshensel CS, Seeman TE, Wight RG, et al. (2009) Trajectories of cognitive function in late life in the United States: demographic and socioeconomic predictors. Am J Epidemiol 170: 331-342. Epub 2009 Jul 2015.
161. Turrell G, Lynch JW, Kaplan GA, Everson SA, Helkala EL, et al. (2002) Socioeconomic position across the lifecourse and cognitive function in late middle age. J Gerontol B Psychol Sci Soc Sci 57: S43-51.
139
162. Zeki Al Hazzouri A, Haan MN, Kalbfleisch JD, Galea S, Lisabeth LD, et al. (2011) Life-course socioeconomic position and incidence of dementia and cognitive impairment without dementia in older Mexican Americans: results from the Sacramento area Latino study on aging. Am J Epidemiol 173: 1148-1158.
163. Gomez-Pinilla F (2008) Brain foods: the effects of nutrients on brain function. Nat Rev Neurosci 9: 568-578.
164. Morris MC (2012) Nutritional determinants of cognitive aging and dementia. Proc Nutr Soc 71: 1-13.
165. Davidson RJ, McEwen BS (2012) Social influences on neuroplasticity: stress and interventions to promote well-being. Nat Neurosci 15: 689-695.
166. Hackman DA, Farah MJ, Meaney MJ (2010) Socioeconomic status and the brain: mechanistic insights from human and animal research. Nat Rev Neurosci 11: 651-659.
167. Bowman GL, Silbert LC, Howieson D, Dodge HH, Traber MG, et al. (2012) Nutrient biomarker patterns, cognitive function, and MRI measures of brain aging. Neurology 78: 241-249.
168. Butterworth P, Cherbuin N, Sachdev P, Anstey KJ (2012) The association between financial hardship and amygdala and hippocampal volumes: results from the PATH through life project. Soc Cogn Affect Neurosci 7: 548-556.
169. Gianaros PJ, Marsland AL, Sheu LK, Erickson KI, Verstynen TD (2012) Inflammatory Pathways Link Socioeconomic Inequalities to White Matter Architecture. Cereb Cortex.
170. Hanson JL, Chandra A, Wolfe BL, Pollak SD (2011) Association between income and the hippocampus. PLoS One 6: 0018712.
171. McEwen BS, Gianaros PJ (2010) Central role of the brain in stress and adaptation: links to socioeconomic status, health, and disease. Ann N Y Acad Sci 1186: 190-222.
172. Noble KG, Grieve SM, Korgaonkar MS, Engelhardt LE, Griffith EY, et al. (2012) Hippocampal volume varies with educational attainment across the life-span. Front Hum Neurosci 6:307.: 10.3389/fnhum.2012.00307. Epub 02012 Nov 00309.
173. Tangney CC, Aggarwal NT, Li H, Wilson RS, Decarli C, et al. (2011) Vitamin B12, cognition, and brain MRI measures: a cross-sectional examination. Neurology 77: 1276-1282.
174. Akbaraly TN, Singh-Manoux A, Marmot MG, Brunner EJ (2009) Education attenuates the association between dietary patterns and cognition. Dement Geriatr Cogn Disord 27: 147-154. Epub 2009 Feb 2002.
140
175. Lee Y, Kim J, Back JH (2009) The influence of multiple lifestyle behaviors on cognitive function in older persons living in the community. Prev Med 48: 86-90.
176. Sabia S, Nabi H, Kivimaki M, Shipley MJ, Marmot MG, et al. (2009) Health behaviors from early to late midlife as predictors of cognitive function: The Whitehall II study. Am J Epidemiol 170: 428-437.
177. Nippak PM, Mendelson J, Muggenburg B, Milgram NW (2007) Enhanced spatial ability in aged dogs following dietary and behavioural enrichment. Neurobiol Learn Mem 87: 610-623.
178. Winocur G, Greenwood CE (1999) The effects of high fat diets and environmental influences on cognitive performance in rats. Behav Brain Res 101: 153-161.
179. Fiocco AJ, Shatenstein B, Ferland G, Payette H, Belleville S, et al. (2011) Sodium intake and physical activity impact cognitive maintenance in older adults: the NuAge Study. Neurobiol Aging 33: e21-28.
180. Scarmeas N, Luchsinger JA, Schupf N, Brickman AM, Cosentino S, et al. (2009) Physical activity, diet, and risk of Alzheimer disease. JAMA 302: 627-637.
181. Gaudreau P, Morais JA, Shatenstein B, Gray-Donald K, Khalil A, et al. (2007) Nutrition as a determinant of successful aging: description of the Quebec longitudinal study Nuage and results from cross-sectional pilot studies. Rejuvenation Res 10: 377-386.
182. Washburn RA, McAuley E, Katula J, Mihalko SL, Boileau RA (1999) The physical activity scale for the elderly (PASE): evidence for validity. J Clin Epidemiol 52: 643-651.
183. Lefrancois R, Leclerc G, Dube M, Hamel S, Gaulin P (2001) Valued activities of everyday life among the very old: a one-year trend. Activities Adaptation Aging 25: 19-35.
184. Fillenbaum GG, Smyer MA (1981) The development, validity, and reliability of the OARS multidimensional functional assessment questionnaire. J Gerontol 36: 428-434.
185. Yesavage JA, Brink TL, Rose TL, Lum O, Huang V, et al. (1982) Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res 17: 37-49.
186. Teng EL, Chui HC (1987) The Modified Mini-Mental State (3MS) examination. J Clin Psychiatry 48: 314-318.
187. Goyder J, Frank K (2007) A scale of occupational prestige in Canada, based on NOC major groups. Can J Sociol 32: 63-83.
188. McLaren L, Godley J (2009) Social class and BMI among Canadian adults: a focus on occupational prestige. Obesity (Silver Spring) 17: 290-299.
141
189. Shatenstein B, Nadon S, Godin C, Ferland G (2005) Development and validation of a food frequency questionnaire. Can J Diet Pract Res 66: 67-75.
190. Shatenstein B, Ferland G, Belleville S, Gray-Donald K, Kergoat MJ, et al. (2012) Diet quality and cognition among older adults from the NuAge study. Exp Gerontol 47: 353-360.
191. Streiner DL (2002) The case of the missing data: methods of dealing with dropouts and other research vagaries. Can J Psychiatry 47: 68-75.
192. Gu Y, Nieves JW, Stern Y, Luchsinger JA, Scarmeas N (2010) Food combination and Alzheimer disease risk: a protective diet. Arch Neurol 67: 699-706.
193. Torres SJ, Lautenschlager NT, Wattanapenpaiboon N, Greenop KR, Beer C, et al. (2012) Dietary patterns are associated with cognition among older people with mild cognitive impairment. Nutrients 4: 1542-1551.
194. Cadar D, Pikhart H, Mishra G, Stephen A, Kuh D, et al. (2012) The role of lifestyle behaviors on 20-year cognitive decline. J Aging Res.
195. Braveman PA, Cubbin C, Egerter S, Chideya S, Marchi KS, et al. (2005) Socioeconomic status in health research: one size does not fit all. JAMA 294: 2879-2888.
196. Steffener J, Stern Y (2012) Exploring the neural basis of cognitive reserve in aging. Biochim Biophys Acta 1822: 467-473.
197. Williams L, Ball K, Crawford D (2010) Why do some socioeconomically disadvantaged women eat better than others? An investigation of the personal, social and environmental correlates of fruit and vegetable consumption. Appetite 55: 441-446.
198. Frautschy SA, Hu W, Kim P, Miller SA, Chu T, et al. (2001) Phenolic anti-inflammatory antioxidant reversal of Abeta-induced cognitive deficits and neuropathology. Neurobiol Aging 22: 993-1005.
199. Lee YJ, Choi DY, Lee YK, Lee YM, Han SB, et al. (2012) 4-O-methylhonokiol prevents memory impairment in the Tg2576 transgenic mice model of Alzheimer's disease via regulation of beta-secretase activity. J Alzheimers Dis 29: 677-690.
200. Murakami K, Murata N, Ozawa Y, Kinoshita N, Irie K, et al. (2011) Vitamin C restores behavioral deficits and amyloid-beta oligomerization without affecting plaque formation in a mouse model of Alzheimer's disease. J Alzheimers Dis 26: 7-18.
201. Rezai-Zadeh K, Arendash GW, Hou H, Fernandez F, Jensen M, et al. (2008) Green tea epigallocatechin-3-gallate (EGCG) reduces beta-amyloid mediated cognitive impairment and modulates tau pathology in Alzheimer transgenic mice. Brain Res 1214: 177-187.
142
202. Wang J, Ho L, Zhao W, Ono K, Rosensweig C, et al. (2008) Grape-derived polyphenolics prevent Abeta oligomerization and attenuate cognitive deterioration in a mouse model of Alzheimer's disease. J Neurosci 28: 6388-6392.
203. Arendash GW, Jensen MT, Salem N, Jr., Hussein N, Cracchiolo J, et al. (2007) A diet high in omega-3 fatty acids does not improve or protect cognitive performance in Alzheimer's transgenic mice. Neuroscience 149: 286-302.
204. Orr SK, Trepanier MO, Bazinet RP (2013) n-3 Polyunsaturated fatty acids in animal models with neuroinflammation. Prostaglandins Leukot Essent Fatty Acids 88: 97-103.
205. Davinelli S, Sapere N, Zella D, Bracale R, Intrieri M, et al. (2012) Pleiotropic protective effects of phytochemicals in Alzheimer's disease. Oxid Med Cell Longev 2012: 386527.
206. Frautschy SA, Cole GM (2010) Why pleiotropic interventions are needed for Alzheimer's disease. Mol Neurobiol 41: 392-409.
207. Grimm MO, Kuchenbecker J, Grosgen S, Burg VK, Hundsdorfer B, et al. (2011) Docosahexaenoic acid reduces amyloid beta production via multiple pleiotropic mechanisms. J Biol Chem 286: 14028-14039.
208. Zhao Y, Calon F, Julien C, Winkler JW, Petasis NA, et al. (2011) Docosahexaenoic acid-derived neuroprotectin D1 induces neuronal survival via secretase- and PPARgamma-mediated mechanisms in Alzheimer's disease models. PLoS One 6: e15816.
209. Bomfim TR, Forny-Germano L, Sathler LB, Brito-Moreira J, Houzel JC, et al. (2012) An anti-diabetes agent protects the mouse brain from defective insulin signaling caused by Alzheimer's disease- associated Abeta oligomers. J Clin Invest 122: 1339-1353.
210. Craft S, Cholerton B, Baker LD (2013) Insulin and Alzheimer's disease: untangling the web. J Alzheimers Dis 33 Suppl 1: S263-275.
211. de la Monte SM (2012) Contributions of brain insulin resistance and deficiency in amyloid-related neurodegeneration in Alzheimer's disease. Drugs 72: 49-66.
212. Talbot K, Wang HY, Kazi H, Han LY, Bakshi KP, et al. (2012) Demonstrated brain insulin resistance in Alzheimer's disease patients is associated with IGF-1 resistance, IRS-1 dysregulation, and cognitive decline. J Clin Invest 122: 1316-1338.
213. Galasko DR, Peskind E, Clark CM, Quinn JF, Ringman JM, et al. (2012) Antioxidants for Alzheimer disease: a randomized clinical trial with cerebrospinal fluid biomarker measures. Arch Neurol 69: 836-841.
214. Sano M, Ernesto C, Thomas RG, Klauber MR, Schafer K, et al. (1997) A controlled trial of selegiline, alpha-tocopherol, or both as treatment for Alzheimer's disease. The Alzheimer's Disease Cooperative Study. N Engl J Med 336: 1216-1222.
143
215. Freund-Levi Y, Eriksdotter-Jonhagen M, Cederholm T, Basun H, Faxen-Irving G, et al. (2006) Omega-3 fatty acid treatment in 174 patients with mild to moderate Alzheimer disease: OmegAD study: a randomized double-blind trial. Arch Neurol 63: 1402-1408.
216. Quinn JF, Raman R, Thomas RG, Yurko-Mauro K, Nelson EB, et al. (2010) Docosahexaenoic acid supplementation and cognitive decline in Alzheimer disease: a randomized trial. JAMA 304: 1903-1911.
217. Morris R (1984) Developments of a water-maze procedure for studying spatial learning in the rat. J Neurosci Methods 11: 47-60.
218. Moscovitch M, Winocur G (1995) Frontal lobes, memory, & aging. In: Grafman J, Holyak K, Boller F, editors. Structure and Function of Human Prefrontal Cortex. New York: New York Academy of Sciences. pp. 119-150.
219. McDonald R, Ergis A-M, Winocur G (1999) Functional dissociation of brain regions in learning an memory. In: Foster J, Jelcic M, editors. Memory: Sysems, Process, or Function. Oxford: Oxford University Press. pp. 66-104.
220. Chishti MA, Yang DS, Janus C, Phinney AL, Horne P, et al. (2001) Early-onset amyloid deposition and cognitive deficits in transgenic mice expressing a double mutant form of amyloid precursor protein 695. J Biol Chem 276: 21562-21570.
221. Reza-Lopez SA, Anderson GH, Szeto IM, Taha AY, Ma DW (2009) High vitamin intake by Wistar rats during pregnancy alters tissue fatty acid concentration in the offspring fed an obesogenic diet. Metabolism 58: 722-730.
222. Winocur G, Vardy J, Binns MA, Kerr L, Tannock I (2006) The effects of the anti-cancer drugs, methotrexate and 5-fluorouracil, on cognitive function in mice. Pharmacol Biochem Behav 85: 66-75.
223. Winocur G, Thompson C, Hakim A, Greenwood C (2013) The effects of high- and low-risk environments on cognitive function in rats following 2-vessel occlusion of the carotid arteries: a behavioral study. Behav Brain Res 252: 144-156.
224. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25: 402-408.
225. McLaurin J, Kierstead ME, Brown ME, Hawkes CA, Lambermon MH, et al. (2006) Cyclohexanehexol inhibitors of Abeta aggregation prevent and reverse Alzheimer phenotype in a mouse model. Nat Med 12: 801-808.
226. Romberg C, Horner AE, Bussey TJ, Saksida LM (2013) A touch screen-automated cognitive test battery reveals impaired attention, memory abnormalities, and increased response inhibition in the TgCRND8 mouse model of Alzheimer's disease. Neurobiol Aging 34: 731-744.
144
227. Reeves PG, Nielsen FH, Fahey GC, Jr. (1993) AIN-93 purified diets for laboratory rodents: final report of the American Institute of Nutrition ad hoc writing committee on the reformulation of the AIN-76A rodent diet. J Nutr 123: 1939-1951.
228. Houtcooper WC (1978) Food habits of rodents in a cultivated ecosystem. J Mammal 59: 427-430.
229. Landry SO (1970) The Rodentia as omnivores. Q Rev Biol 45: 351-372.
230. Tann CR, Singleton GR, Coman BJ (1993) Diet of the house mouse, Mus domesticus, in the Mallee wheatlands of North-western Victoria. Wildl Res 18: 1-12.
231. Khan TK, Alkon DL (2006) An internally controlled peripheral biomarker for Alzheimer's disease: Erk1 and Erk2 responses to the inflammatory signal bradykinin. Proc Natl Acad Sci U S A 103: 13203-13207.
232. Veeranna, Kaji T, Boland B, Odrljin T, Mohan P, et al. (2004) Calpain mediates calcium-induced activation of the erk1,2 MAPK pathway and cytoskeletal phosphorylation in neurons: relevance to Alzheimer's disease. Am J Pathol 165: 795-805.
233. Hallock P, Thomas MA (2012) Integrating the Alzheimer's disease proteome and transcriptome: a comprehensive network model of a complex disease. OMICS 16: 37-49.
234. Nelson TJ, Sun MK, Hongpaisan J, Alkon DL (2008) Insulin, PKC signaling pathways and synaptic remodeling during memory storage and neuronal repair. Eur J Pharmacol 585: 76-87.
235. Sweatt JD (2004) Mitogen-activated protein kinases in synaptic plasticity and memory. Curr Opin Neurobiol 14: 311-317.
236. Giovannini MG, Cerbai F, Bellucci A, Melani C, Grossi C, et al. (2008) Differential activation of mitogen-activated protein kinase signalling pathways in the hippocampus of CRND8 transgenic mouse, a model of Alzheimer's disease. Neuroscience 153: 618-633.
237. Maurin H, Lechat B, Dewachter I, Ris L, Louis JV, et al. (2013) Neurological characterization of mice deficient in GSK3alpha highlight pleiotropic physiological functions in cognition and pathological activity as Tau kinase. Mol Brain 6: 27.
238. Dudal S, Krzywkowski P, Paquette J, Morissette C, Lacombe D, et al. (2004) Inflammation occurs early during the Abeta deposition process in TgCRND8 mice. Neurobiol Aging 25: 861-871.
239. Robinson CA, Clark AW, Parhad IM, Fung TS, Bou SS (1994) Gene expression in Alzheimer neocortex as a function of age and pathologic severity. Neurobiol Aging 15: 681-690.
145
240. Kashon ML, Ross GW, O'Callaghan JP, Miller DB, Petrovitch H, et al. (2004) Associations of cortical astrogliosis with cognitive performance and dementia status. J Alzheimers Dis 6: 595-604; discussion 673-581.
241. Tarkowski E, Andreasen N, Tarkowski A, Blennow K (2003) Intrathecal inflammation precedes development of Alzheimer's disease. J Neurol Neurosurg Psychiatry 74: 1200-1205.
242. Gabbita SP, Srivastava MK, Eslami P, Johnson MF, Kobritz NK, et al. (2012) Early intervention with a small molecule inhibitor for tumor necrosis factor-alpha prevents cognitive deficits in a triple transgenic mouse model of Alzheimer's disease. J Neuroinflammation 9: 99.
243. Medeiros R, Prediger RD, Passos GF, Pandolfo P, Duarte FS, et al. (2007) Connecting TNF-alpha signaling pathways to iNOS expression in a mouse model of Alzheimer's disease: relevance for the behavioral and synaptic deficits induced by amyloid beta protein. J Neurosci 27: 5394-5404.
244. Tweedie D, Ferguson RA, Fishman K, Frankola KA, Van Praag H, et al. (2012) Tumor necrosis factor-alpha synthesis inhibitor 3,6'-dithiothalidomide attenuates markers of inflammation, Alzheimer pathology and behavioral deficits in animal models of neuroinflammation and Alzheimer's disease. J Neuroinflammation 9: 106.
245. Chakrabarty P, Herring A, Ceballos-Diaz C, Das P, Golde TE (2011) Hippocampal expression of murine TNFalpha results in attenuation of amyloid deposition in vivo. Mol Neurodegener 6: 16.
246. Chakrabarty P, Jansen-West K, Beccard A, Ceballos-Diaz C, Levites Y, et al. (2010) Massive gliosis induced by interleukin-6 suppresses Abeta deposition in vivo: evidence against inflammation as a driving force for amyloid deposition. FASEB J 24: 548-559.
247. Shaftel SS, Kyrkanides S, Olschowka JA, Miller JN, Johnson RE, et al. (2007) Sustained hippocampal IL-1 beta overexpression mediates chronic neuroinflammation and ameliorates Alzheimer plaque pathology. J Clin Invest 117: 1595-1604.
248. Harrison FE, Allard J, Bixler R, Usoh C, Li L, et al. (2009) Antioxidants and cognitive training interact to affect oxidative stress and memory in APP/PSEN1 mice. Nutr Neurosci 12: 203-218.
249. Quinn JF, Bussiere JR, Hammond RS, Montine TJ, Henson E, et al. (2007) Chronic dietary alpha-lipoic acid reduces deficits in hippocampal memory of aged Tg2576 mice. Neurobiol Aging 28: 213-225.
146
250. Joseph JA, Denisova NA, Arendash G, Gordon M, Diamond D, et al. (2003) Blueberry supplementation enhances signaling and prevents behavioral deficits in an Alzheimer disease model. Nutr Neurosci 6: 153-162.
251. Denner LA, Rodriguez-Rivera J, Haidacher SJ, Jahrling JB, Carmical JR, et al. (2012) Cognitive enhancement with rosiglitazone links the hippocampal PPARgamma and ERK MAPK signaling pathways. J Neurosci 32: 16725-16735a.
252. Calabrese V, Cornelius C, Dinkova-Kostova AT, Iavicoli I, Di Paola R, et al. (2012) Cellular stress responses, hormetic phytochemicals and vitagenes in aging and longevity. Biochim Biophys Acta 1822: 753-783.
253. Son TG, Camandola S, Mattson MP (2008) Hormetic dietary phytochemicals. Neuromolecular Med 10: 236-246.
254. Mattson MP (2008) Hormesis defined. Ageing Res Rev 7: 1-7.
255. Wang X, Wu H, Chen H, Liu R, Liu J, et al. (2012) Does insulin bolster antioxidant defenses via the extracellular signal-regulated kinases-protein kinase B-nuclear factor erythroid 2 p45-related factor 2 pathway? Antioxid Redox Signal 16: 1061-1070.
256. Wu CC, Hsu MC, Hsieh CW, Lin JB, Lai PH, et al. (2006) Upregulation of heme oxygenase-1 by Epigallocatechin-3-gallate via the phosphatidylinositol 3-kinase/Akt and ERK pathways. Life Sci 78: 2889-2897.
257. Yu R, Lei W, Mandlekar S, Weber MJ, Der CJ, et al. (1999) Role of a mitogen-activated protein kinase pathway in the induction of phase II detoxifying enzymes by chemicals. J Biol Chem 274: 27545-27552.
258. Chua LM, Lim ML, Chong PR, Hu ZP, Cheung NS, et al. (2012) Impaired neuronal insulin signaling precedes Abeta42 accumulation in female AbetaPPsw/PS1DeltaE9 mice. J Alzheimers Dis 29: 783-791.
259. Ma QL, Yang F, Rosario ER, Ubeda OJ, Beech W, et al. (2009) Beta-amyloid oligomers induce phosphorylation of tau and inactivation of insulin receptor substrate via c-Jun N-terminal kinase signaling: suppression by omega-3 fatty acids and curcumin. J Neurosci 29: 9078-9089.
260. Maesako M, Uemura K, Kuzuya A, Sasaki K, Asada M, et al. (2012) Gain of function by phosphorylation in Presenilin 1-mediated regulation of insulin signaling. J Neurochem 121: 964-973.
261. Anantharaman M, Tangpong J, Keller JN, Murphy MP, Markesbery WR, et al. (2006) Beta-amyloid mediated nitration of manganese superoxide dismutase: implication for oxidative stress in a APPNLH/NLH X PS-1P264L/P264L double knock-in mouse model of Alzheimer's disease. Am J Pathol 168: 1608-1618.
147
262. Cristalli DO, Arnal N, Marra FA, de Alaniz MJ, Marra CA (2012) Peripheral markers in neurodegenerative patients and their first-degree relatives. J Neurol Sci 314: 48-56.
263. Garg SK, Vitvitsky V, Albin R, Banerjee R (2011) Astrocytic redox remodeling by amyloid beta peptide. Antioxid Redox Signal 14: 2385-2397.
264. Ghosh D, LeVault KR, Barnett AJ, Brewer GJ (2012) A reversible early oxidized redox state that precedes macromolecular ROS damage in aging nontransgenic and 3xTg-AD mouse neurons. J Neurosci 32: 5821-5832.
265. Kharrazi H, Vaisi-Raygani A, Rahimi Z, Tavilani H, Aminian M, et al. (2008) Association between enzymatic and non-enzymatic antioxidant defense mechanism with apolipoprotein E genotypes in Alzheimer disease. Clin Biochem 41: 932-936.
266. Lee HP, Pancholi N, Esposito L, Previll LA, Wang X, et al. (2012) Early induction of oxidative stress in mouse model of Alzheimer disease with reduced mitochondrial superoxide dismutase activity. PLoS One 7: e28033.
267. Puertas MC, Martinez-Martos JM, Cobo MP, Carrera MP, Mayas MD, et al. (2012) Plasma oxidative stress parameters in men and women with early stage Alzheimer type dementia. Exp Gerontol 47: 625-630.
268. Schuessel K, Schafer S, Bayer TA, Czech C, Pradier L, et al. (2005) Impaired Cu/Zn-SOD activity contributes to increased oxidative damage in APP transgenic mice. Neurobiol Dis 18: 89-99.
269. Sofic E, Sapcanin A, Tahirovic I, Gavrankapetanovic I, Jellinger K, et al. (2006) Antioxidant capacity in postmortem brain tissues of Parkinson's and Alzheimer's diseases. J Neural Transm Suppl: 39-43.
270. Devore EE, Kang JH, Breteler MM, Grodstein F (2012) Dietary intakes of berries and flavonoids in relation to cognitive decline. Ann Neurol 72: 135-143.
271. Commenges D, Scotet V, Renaud S, Jacqmin-Gadda H, Barberger-Gateau P, et al. (2000) Intake of flavonoids and risk of dementia. Eur J Epidemiol 16: 357-363.
272. Kesse-Guyot E, Fezeu L, Andreeva VA, Touvier M, Scalbert A, et al. (2012) Total and specific polyphenol intakes in midlife are associated with cognitive function measured 13 years later. J Nutr 142: 76-83.
273. Joseph JA, Shukitt-Hale B, Denisova NA, Bielinski D, Martin A, et al. (1999) Reversals of age-related declines in neuronal signal transduction, cognitive, and motor behavioral deficits with blueberry, spinach, or strawberry dietary supplementation. J Neurosci 19: 8114-8121.
148
274. Shih PH, Chan YC, Liao JW, Wang MF, Yen GC (2010) Antioxidant and cognitive promotion effects of anthocyanin-rich mulberry (Morus atropurpurea L.) on senescence-accelerated mice and prevention of Alzheimer's disease. J Nutr Biochem 21: 598-605.
275. Arsenault D, Julien C, Tremblay C, Calon F (2011) DHA improves cognition and prevents dysfunction of entorhinal cortex neurons in 3xTg-AD mice. PLoS One 6: e17397.
276. Giunta B, Hou H, Zhu Y, Salemi J, Ruscin A, et al. (2010) Fish oil enhances anti-amyloidogenic properties of green tea EGCG in Tg2576 mice. Neurosci Lett 471: 134-138.
277. Broersen LM, Kuipers AA, Balvers M, van Wijk N, Savelkoul PJ, et al. (2013) A specific multi-nutrient diet reduces Alzheimer-like pathology in young adult AbetaPPswe/PS1dE9 mice. J Alzheimers Dis 33: 177-190.
278. Amtul Z, Keet M, Wang L, Merrifield P, Westaway D, et al. (2011) DHA supplemented in peptamen diet offers no advantage in pathways to amyloidosis: is it time to evaluate composite lipid diet? PLoS One 6: e24094.
279. Valgimigli L, Iori R (2009) Antioxidant and pro-oxidant capacities of ITCs. Environ Mol Mutagen 50: 222-237.
280. Singh SV, Singh K (2012) Cancer chemoprevention with dietary isothiocyanates mature for clinical translational research. Carcinogenesis 33: 1833-1842.
281. Spencer JP, Abd El Mohsen MM, Minihane AM, Mathers JC (2008) Biomarkers of the intake of dietary polyphenols: strengths, limitations and application in nutrition research. Br J Nutr 99: 12-22.
282. Egert S, Rimbach G (2011) Which sources of flavonoids: complex diets or dietary supplements? Adv Nutr 2: 8-14.
283. Dergal JM, Gold JL, Laxer DA, Lee MS, Binns MA, et al. (2002) Potential interactions between herbal medicines and conventional drug therapies used by older adults attending a memory clinic. Drugs Aging 19: 879-886.
284. Shahrokh LE, Lukaszuk JM, Prawitz AD (2005) Elderly herbal supplement users less satisfied with medical care than nonusers. J Am Diet Assoc 105: 1138-1140.
285. Anstey KJ, Cherbuin N, Budge M, Young J (2011) Body mass index in midlife and late-life as a risk factor for dementia: a meta-analysis of prospective studies. Obes Rev 12: e426-437.
286. Dahl AK, Hassing LB (2012) Obesity and Cognitive Aging. Epidemiol Rev.
149
287. Bove RM, Brick DJ, Healy BC, Mancuso SM, Gerweck AV, et al. (2013) Metabolic and endocrine correlates of cognitive function in healthy young women. Obesity (Silver Spring) 21: 1343-1349.
288. Yau PL, Castro MG, Tagani A, Tsui WH, Convit A (2012) Obesity and metabolic syndrome and functional and structural brain impairments in adolescence. Pediatrics 130: e856-864.
289. Gudala K, Bansal D, Schifano F, Bhansali A (2013) Diabetes mellitus and risk of dementia: A meta-analysis of prospective observational studies. J Diabetes Investig 4: 640-650.
290. Rasgon NL, Kenna HA, Wroolie TE, Kelley R, Silverman D, et al. (2011) Insulin resistance and hippocampal volume in women at risk for Alzheimer's disease. Neurobiol Aging 32: 1942-1948.
291. Ronnemaa E, Zethelius B, Sundelof J, Sundstrom J, Degerman-Gunnarsson M, et al. (2008) Impaired insulin secretion increases the risk of Alzheimer disease. Neurology 71: 1065-1071.
292. Xu W, Caracciolo B, Wang HX, Winblad B, Backman L, et al. (2010) Accelerated progression from mild cognitive impairment to dementia in people with diabetes. Diabetes 59: 2928-2935.
293. Yaffe K (2007) Metabolic syndrome and cognitive disorders: is the sum greater than its parts? Alzheimer Dis Assoc Disord 21: 167-171.
294. Fadel JR, Jolivalt CG, Reagan LP (2013) Food for thought: the role of appetitive peptides in age-related cognitive decline. Ageing Res Rev 12: 764-776.
295. Dhamoon MS, Noble JM, Craft S (2009) Intranasal insulin improves cognition and modulates beta-amyloid in early AD. Neurology 72: 292-293; author reply 293-294.
296. Hallschmid M, Benedict C, Schultes B, Born J, Kern W (2008) Obese men respond to cognitive but not to catabolic brain insulin signaling. Int J Obes (Lond) 32: 275-282.
297. Novak V, Milberg W, Hao Y, Munshi M, Novak P, et al. (2014) Enhancement of vasoreactivity and cognition by intranasal insulin in type 2 diabetes. Diabetes Care 37: 751-759.
298. Park CR, Seeley RJ, Craft S, Woods SC (2000) Intracerebroventricular insulin enhances memory in a passive-avoidance task. Physiol Behav 68: 509-514.
299. Baura GD, Foster DM, Porte D, Jr., Kahn SE, Bergman RN, et al. (1993) Saturable transport of insulin from plasma into the central nervous system of dogs in vivo. A mechanism for regulated insulin delivery to the brain. J Clin Invest 92: 1824-1830.
150
300. Kaiyala KJ, Prigeon RL, Kahn SE, Woods SC, Schwartz MW (2000) Obesity induced by a high-fat diet is associated with reduced brain insulin transport in dogs. Diabetes 49: 1525-1533.
301. Kern W, Benedict C, Schultes B, Plohr F, Moser A, et al. (2006) Low cerebrospinal fluid insulin levels in obese humans. Diabetologia 49: 2790-2792.
302. McNeilly AD, Williamson R, Balfour DJ, Stewart CA, Sutherland C (2012) A high-fat-diet-induced cognitive deficit in rats that is not prevented by improving insulin sensitivity with metformin. Diabetologia 55: 3061-3070.
303. Pathan AR, Gaikwad AB, Viswanad B, Ramarao P (2008) Rosiglitazone attenuates the cognitive deficits induced by high fat diet feeding in rats. Eur J Pharmacol 589: 176-179.
304. Pintana H, Apaijai N, Pratchayasakul W, Chattipakorn N, Chattipakorn SC (2012) Effects of metformin on learning and memory behaviors and brain mitochondrial functions in high fat diet induced insulin resistant rats. Life Sci 91: 409-414.
305. Abbatecola AM, Lattanzio F, Molinari AM, Cioffi M, Mansi L, et al. (2010) Rosiglitazone and cognitive stability in older individuals with type 2 diabetes and mild cognitive impairment. Diabetes Care 33: 1706-1711.
306. Wang BW, Hok V, Della-Chiesa A, Callaghan C, Barlow S, et al. (2012) Rosiglitazone enhances learning, place cell activity, and synaptic plasticity in middle-aged rats. Neurobiol Aging 33: 835 e813-830.
307. Greenwood CE, Winocur G (1996) Cognitive impairment in rats fed high-fat diets: a specific effect of saturated fatty-acid intake. Behav Neurosci 110: 451-459.
308. Greenwood CE, Winocur G (2001) Glucose treatment reduces memory deficits in young adult rats fed high-fat diets. Neurobiol Learn Mem 75: 179-189.
309. Winocur G (1986) Memory decline in aged rats: a neuropsychological interpretation. J Gerontol 41: 758-763.
310. Grillo CA, Piroli GG, Hendry RM, Reagan LP (2009) Insulin-stimulated translocation of GLUT4 to the plasma membrane in rat hippocampus is PI3-kinase dependent. Brain Res 1296: 35-45.
311. Piroli GG, Grillo CA, Hoskin EK, Znamensky V, Katz EB, et al. (2002) Peripheral glucose administration stimulates the translocation of GLUT8 glucose transporter to the endoplasmic reticulum in the rat hippocampus. J Comp Neurol 452: 103-114.
312. Winocur G (1991) Functional dissociation of the hippocampus and prefrontal cortex in learning and memory. Psychobiology 19: 11-20.
151
313. Roberts RO, Geda YE, Knopman DS, Cha RH, Boeve BF, et al. (2010) Metabolic syndrome, inflammation, and nonamnestic mild cognitive impairment in older persons: a population-based study. Alzheimer Dis Assoc Disord 24: 11-18.
314. Denis GV, Obin MS (2013) 'Metabolically healthy obesity': origins and implications. Mol Aspects Med 34: 59-70.
315. McArdle MA, Finucane OM, Connaughton RM, McMorrow AM, Roche HM (2013) Mechanisms of obesity-induced inflammation and insulin resistance: insights into the emerging role of nutritional strategies. Front Endocrinol (Lausanne) 4: 52.
316. Deshmane SL, Kremlev S, Amini S, Sawaya BE (2009) Monocyte chemoattractant protein-1 (MCP-1): an overview. J Interferon Cytokine Res 29: 313-326.
317. Erion JR, Wosiski-Kuhn M, Dey A, Hao S, Davis CL, et al. (2014) Obesity elicits interleukin 1-mediated deficits in hippocampal synaptic plasticity. J Neurosci 34: 2618-2631.
318. Kim JJ, Li P, Huntley J, Chang JP, Arden KC, et al. (2009) FoxO1 haploinsufficiency protects against high-fat diet-induced insulin resistance with enhanced peroxisome proliferator-activated receptor gamma activation in adipose tissue. Diabetes 58: 1275-1282.
319. Rosen ED, Sarraf P, Troy AE, Bradwin G, Moore K, et al. (1999) PPAR gamma is required for the differentiation of adipose tissue in vivo and in vitro. Mol Cell 4: 611-617.
320. Grillo CA, Piroli GG, Evans AN, Macht VA, Wilson SP, et al. (2011) Obesity/hyperleptinemic phenotype adversely affects hippocampal plasticity: effects of dietary restriction. Physiol Behav 104: 235-241.
321. Grillo CA, Piroli GG, Junor L, Wilson SP, Mott DD, et al. (2011) Obesity/hyperleptinemic phenotype impairs structural and functional plasticity in the rat hippocampus. Physiol Behav 105: 138-144.
322. Harvey J (2007) Leptin regulation of neuronal excitability and cognitive function. Curr Opin Pharmacol 7: 643-647.
323. Morrison CD (2009) Leptin signaling in brain: A link between nutrition and cognition? Biochim Biophys Acta 1792: 401-408.
324. Banks WA, DiPalma CR, Farrell CL (1999) Impaired transport of leptin across the blood-brain barrier in obesity. Peptides 20: 1341-1345.
325. Burguera B, Couce ME, Curran GL, Jensen MD, Lloyd RV, et al. (2000) Obesity is associated with a decreased leptin transport across the blood-brain barrier in rats. Diabetes 49: 1219-1223.
152
326. Banks WA, Coon AB, Robinson SM, Moinuddin A, Shultz JM, et al. (2004) Triglycerides induce leptin resistance at the blood-brain barrier. Diabetes 53: 1253-1260.
327. Mielke JG, Nicolitch K, Avellaneda V, Earlam K, Ahuja T, et al. (2006) Longitudinal study of the effects of a high-fat diet on glucose regulation, hippocampal function, and cerebral insulin sensitivity in C57BL/6 mice. Behav Brain Res 175: 374-382.
328. Ross AP, Bruggeman EC, Kasumu AW, Mielke JG, Parent MB (2012) Non-alcoholic fatty liver disease impairs hippocampal-dependent memory in male rats. Physiol Behav 106: 133-141.
329. Pipatpiboon N, Pratchayasakul W, Chattipakorn N, Chattipakorn SC (2012) PPARgamma agonist improves neuronal insulin receptor function in hippocampus and brain mitochondria function in rats with insulin resistance induced by long term high-fat diets. Endocrinology 153: 329-338.
330. Muller AP, Cammarota M, Dietrich MO, Rotta LN, Portela LV, et al. (2008) Different effect of high fat diet and physical exercise in the hippocampal signaling. Neurochem Res 33: 880-885.
331. Mauvais-Jarvis F, Ueki K, Fruman DA, Hirshman MF, Sakamoto K, et al. (2002) Reduced expression of the murine p85alpha subunit of phosphoinositide 3-kinase improves insulin signaling and ameliorates diabetes. J Clin Invest 109: 141-149.
332. Terauchi Y, Tsuji Y, Satoh S, Minoura H, Murakami K, et al. (1999) Increased insulin sensitivity and hypoglycaemia in mice lacking the p85 alpha subunit of phosphoinositide 3-kinase. Nat Genet 21: 230-235.
333. Ueki K, Algenstaedt P, Mauvais-Jarvis F, Kahn CR (2000) Positive and negative regulation of phosphoinositide 3-kinase-dependent signaling pathways by three different gene products of the p85alpha regulatory subunit. Mol Cell Biol 20: 8035-8046.
334. Ueki K, Fruman DA, Yballe CM, Fasshauer M, Klein J, et al. (2003) Positive and negative roles of p85 alpha and p85 beta regulatory subunits of phosphoinositide 3-kinase in insulin signaling. J Biol Chem 278: 48453-48466.
335. Begg DP, Mul JD, Liu M, Reedy BM, D'Alessio DA, et al. (2013) Reversal of diet-induced obesity increases insulin transport into cerebrospinal fluid and restores sensitivity to the anorexic action of central insulin in male rats. Endocrinology 154: 1047-1054.
336. Watson GS, Craft S (2004) Modulation of memory by insulin and glucose: neuropsychological observations in Alzheimer's disease. Eur J Pharmacol 490: 97-113.
337. McNay EC, Fries TM, Gold PE (2000) Decreases in rat extracellular hippocampal glucose concentration associated with cognitive demand during a spatial task. Proc Natl Acad Sci U S A 97: 2881-2885.
153
338. McNay EC, Gold PE (2001) Age-related differences in hippocampal extracellular fluid glucose concentration during behavioral testing and following systemic glucose administration. J Gerontol A Biol Sci Med Sci 56: B66-71.
339. Messier C (2004) Glucose improvement of memory: a review. Eur J Pharmacol 490: 33-57.
340. Baker LD, Cross DJ, Minoshima S, Belongia D, Watson GS, et al. (2011) Insulin resistance and Alzheimer-like reductions in regional cerebral glucose metabolism for cognitively normal adults with prediabetes or early type 2 diabetes. Arch Neurol 68: 51-57.
341. Burns CM, Chen K, Kaszniak AW, Lee W, Alexander GE, et al. (2013) Higher serum glucose levels are associated with cerebral hypometabolism in Alzheimer regions. Neurology 80: 1557-1564.
342. Emmanuel Y, Cochlin LE, Tyler DJ, de Jager CA, Smith AD, et al. (2013) Human hippocampal energy metabolism is impaired during cognitive activity in a lipid infusion model of insulin resistance. Brain Behav 3: 134-144.
343. Nugent S, Croteau E, Pifferi F, Fortier M, Tremblay S, et al. (2011) Brain and systemic glucose metabolism in the healthy elderly following fish oil supplementation. Prostaglandins Leukot Essent Fatty Acids 85: 287-291.
344. Kanoski SE, Zhang Y, Zheng W, Davidson TL (2010) The effects of a high-energy diet on hippocampal function and blood-brain barrier integrity in the rat. J Alzheimers Dis 21: 207-219.
345. Haast RA, Kiliaan AJ (2014) Impact of fatty acids on brain circulation, structure and function. Prostaglandins Leukot Essent Fatty Acids.
346. Valladolid-Acebes I, Merino B, Principato A, Fole A, Barbas C, et al. (2012) High-fat diets induce changes in hippocampal glutamate metabolism and neurotransmission. Am J Physiol Endocrinol Metab 302: E396-402.
347. Yin QQ, Pei JJ, Xu S, Luo DZ, Dong SQ, et al. (2013) Pioglitazone improves cognitive function via increasing insulin sensitivity and strengthening antioxidant defense system in fructose-drinking insulin resistance rats. PLoS One 8: e59313.
348. Morrison CD, Pistell PJ, Ingram DK, Johnson WD, Liu Y, et al. (2010) High fat diet increases hippocampal oxidative stress and cognitive impairment in aged mice: implications for decreased Nrf2 signaling. J Neurochem 114: 1581-1589.
349. Kim YB, Nikoulina SE, Ciaraldi TP, Henry RR, Kahn BB (1999) Normal insulin-dependent activation of Akt/protein kinase B, with diminished activation of phosphoinositide 3-kinase, in muscle in type 2 diabetes. J Clin Invest 104: 733-741.
154
350. Zheng WH, Kar S, Quirion R (2002) Insulin-like growth factor-1-induced phosphorylation of transcription factor FKHRL1 is mediated by phosphatidylinositol 3-kinase/Akt kinase and role of this pathway in insulin-like growth factor-1-induced survival of cultured hippocampal neurons. Mol Pharmacol 62: 225-233.
351. Nakae J, Cao Y, Oki M, Orba Y, Sawa H, et al. (2008) Forkhead transcription factor FoxO1 in adipose tissue regulates energy storage and expenditure. Diabetes 57: 563-576.
352. Fan W, Imamura T, Sonoda N, Sears DD, Patsouris D, et al. (2009) FOXO1 transrepresses peroxisome proliferator-activated receptor gamma transactivation, coordinating an insulin-induced feed-forward response in adipocytes. J Biol Chem 284: 12188-12197.
353. Morrison CD, Huypens P, Stewart LK, Gettys TW (2009) Implications of crosstalk between leptin and insulin signaling during the development of diet-induced obesity. Biochim Biophys Acta 1792: 409-416.
354. Niswender KD, Schwartz MW (2003) Insulin and leptin revisited: adiposity signals with overlapping physiological and intracellular signaling capabilities. Front Neuroendocrinol 24: 1-10.
355. Lu M, Sarruf DA, Talukdar S, Sharma S, Li P, et al. (2011) Brain PPAR-gamma promotes obesity and is required for the insulin-sensitizing effect of thiazolidinediones. Nat Med 17: 618-622.
356. Ryan KK, Li B, Grayson BE, Matter EK, Woods SC, et al. (2011) A role for central nervous system PPAR-gamma in the regulation of energy balance. Nat Med 17: 623-626.
357. Kim YB, Ciaraldi TP, Kong A, Kim D, Chu N, et al. (2002) Troglitazone but not metformin restores insulin-stimulated phosphoinositide 3-kinase activity and increases p110beta protein levels in skeletal muscle of type 2 diabetic subjects. Diabetes 51: 443-448.
358. Kim YB, Peroni OD, Franke TF, Kahn BB (2000) Divergent regulation of Akt1 and Akt2 isoforms in insulin target tissues of obese Zucker rats. Diabetes 49: 847-856.
359. Hu FB, Rimm EB, Stampfer MJ, Ascherio A, Spiegelman D, et al. (2000) Prospective study of major dietary patterns and risk of coronary heart disease in men. Am J Clin Nutr 72: 912-921.
360. Willcox BJ, He Q, Chen R, Yano K, Masaki KH, et al. (2006) Midlife risk factors and healthy survival in men. JAMA 296: 2343-2350.
361. Stern Y (2009) Cognitive reserve. Neuropsychologia 47: 2015-2028. Epub 2009 Mar 2013.