University of Southern California Dissertations and Theses

Physical activity optimizes circuit-specific cellular metabolism in neuroplasticity: a role for hypoxia-inducible factor-1 and its downstream targets

(USC Thesis Other)

Physical activity optimizes circuit-specific cellular metabolism in neuroplasticity: a role for hypoxia-inducible factor-1 and its downstream targets

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
PHYSICAL ACTIVITY OPTIMIZES CIRCUIT-SPECIFIC CELLULAR METABOLISM IN
NEUROPLASTICITY: A ROLE FOR HYPOXIA-INDUCIBLE FACTOR-1 AND ITS DOWNSTREAM
TARGETS

by

Matthew Ryan Halliday

____________________________________________________________________________

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree

DOCTOR OF PHILOSOPHY
(NEUROSCIENCE)

May 2019

Advisory Committee:
Michael W. Jakowec, PhD
Giselle M. Petzinger, MD
Daniel P. Holschneider, MD
John P. Walsh, PhD

Copyright 2019 Matthew R. Halliday

TABLE OF CONTENTS

CHAPTER 1: Introduction 1

CHAPTER 2: The effects of exercise on dopamine neurotransmission in Parkinson’s
disease: targeting neuroplasticity to modulate basal ganglia circuitry 8

CHAPTER 3: Exercise enhances neuroplasticity by targeting synaptic energy metabolism
and mitochondrial function in animal models of Parkinson’s disease 22

CHAPTER 4: Exercise-induced activation of a hypoxia-inducible factor 1-controlled gene
program: implications for experience-dependent neuroplasticity 36

CHAPTER 5: Physical activity optimizes circuit-specific cellular metabolism in
neuroplasticity: a role for hypoxia-inducible factor 1 and its downstream targets 47

REFERENCES 56

APPENDIX A: Relationship between cyclophilin A levels and matrix metalloproteinase-9
activity in cerebrospinal fluid of cognitively normal apolipoprotein E4 carriers and blood-
brain barrier breakdown 62

APPENDIX B: Blood-brain barrier breakdown in the aging human hippocampus 66

APPENDIX C: Central role for PICALM in amyloid-b blood-brain barrier transcytosis and
clearance 78

APPENDIX D: Accelerated pericyte degeneration and blood-brain barrier breakdown in
apolipoprotein E4 carriers with Alzheimer’s disease 112
1

CHAPTER 1: Introduction

Accumulating evidence suggest that physical activity (e.g. treadmill running) reduces the risk of a
wide range of neurological disease, including those that are associated with compromised cognition and
brain function. On one hand, studies in human neuroscience (i.e. epidemiology and clinical trials) have
been particularly useful in assessing the ability of different forms of exercise to impact the risk of developing
disease and delaying its progression. On the other hand, research using laboratory animals, such as
rodents, allow for a reductionist approach to understand the underlying mechanisms by which exercise
impacts cellular and molecular systems. In our research group, we are interested in understanding the
underlying mechanism by which experience in the form of intensive treadmill exercise can facilitate
neuroplasticity within motor and cognitive circuits affected in Parkinson’s disease (PD). The goal of the
studies presented in this dissertation was to explore the role of hypoxia-inducible factor 1 (HIF-1) as a
potential regulatory element in cellular adaptation to increased metabolic demand imposed by intensive
treadmill running and as a facilitator of experience-dependent neuroplasticity. This chapter will provide an
introduction to Parkinson’s disease (PD) and other relevant background information underlying the research
presented in this dissertation. Chapter 2 will discuss the impact of exercise in neuroplasticity in animal
models of PD. Chapter 3 will begin to focus on the underlying molecular mechanisms by which exercise
mediates neuroplasticity with an emphasis on metabolic systems. Chapter 4 will examine the role of the
transcription factor HIF-1 and its downstream target genes in facilitating exercise-induced neuroplasticity.
Chapter 5 will discuss the spatiotemporal profile by which exercise activates the HIF-1 gene program.
Finally, Chapter 6 will summarize all of the findings in this dissertation as well as provide insight into the
significance of these studies and possible future directions.

Parkinson’s disease (PD) is a common neurodegenerative disorder characterized by progressive
loss of nigrostriatal dopaminergic neurons and the presence of a-synuclein-containing Lew body
depositions. Symptomatically, PD is characterized by motor impairments (bradykinesia, tremor, rigidity, and
postural instability), cognitive impairments (executive domain dysfunction), and mood disorders (Petzinger
et al., 2013). Dopamine replacement therapy remains the gold standard treatment option for individuals
with PD, but only alleviates motor symptoms with limited efficacy (Schapira et al., 2009). Epidemiological
2

studies examining the effects of exercise have suggested that regular physical activity can improve motor
performance in patients with PD (Chen et al., 2005; Farley and Koshland, 2005; Fox et al., 2006; Fisher et
al., 2008; Ridgel et al., 2009; Alberts et al., 2011; Schenkman et al., 2012). Furthermore, there is a growing
body of literature on human beings and animal models of PD that support the notion that the beneficial
effects of exercise occur through neuroplasticity of basal ganglia structures (Fisher et al., 2004; Petzinger
et al., 2007; Kintz et al., 2013; Toy et al., 2014). Neuroplasticity is the ability of the brain to change and
adapt (structurally and functionally) in an experience-dependent manner (Kleim and Jones, 2008).

Over the past decade, studies in animals of PD have helped to understand the mechanism related
to exercise-induced neuroplasticity in dopamine-depleted brain. These include the 1-methyl-4-phenyl-
1,2,3,6-tertrahydropyridine (MPTP) and 6-hydroxydopamine (6-OHDA) neurotoxin models. These models
have proven to be particularly useful as they both replicate many pathological and behavioral characteristics
of PD, including: (1) degeneration of nigrostriatal dopaminergic neurons, (2) depletion of striatal dopamine,
and (3) impaired motor and cognitive performance. Importantly, exercise has been shown to improve motor
performance in both models through neuroplastic mechanisms targeting the corticostriatal motor circuit
affected in PD. Exercise-induced neuroplasticity has been associated with modulation of glutamate and
dopamine neurotransmission, increasing cerebral blood flow and angiogenesis, facilitating synaptogenesis
and neurogenesis, and the release of neurotrophic factors, such as BDNF. While much progress has been
made in recent years, the underlying cellular and molecular mechanisms driving exercise-induced
physiological adaptations in the brain remain unclear.

As the principle consumer of systemically available energy, normal brain function is highly
dependent upon a continuous supply of oxygen and blood-derived nutrients. When local metabolic demand
exceeds that of available energy resources, highly conserved gene programs are activated to coordinate
an integrated physiological response to increase oxygen delivery, optimize cellular metabolism, and
establish homeostasis (Sharp and Bernaudin, 2004). Of the factors that have been shown to relay metabolic
signals to adaptive changes in homeostatic gene expression, hypoxia-inducible factor 1 (HIF-1) has been
studied most extensively. Furthermore, the role of HIF-1 in neural tissues is not restricted to the regulation
of energy metabolism, and has been shown to play a critical role in regulating neurogenesis (Zhu et al.,
3

2005; Cunningham et al., 2012; Ross et al., 2012), angiogenesis (Shweiki et al., 1992; Marti et al., 2000;
Adams and Alitalo, 2007), and mitochondrial biogenesis (Correia et al., 2011), all of which are thought to
be important for exercise-induced neuroplasticity (Petzinger et al., 2015). Thus, one potential mechanism
may be that exercise, through its activation via increased metabolic demand, activates the HIF-1 gene
program, which in turn facilitates and supports synaptic strength and connectivity during motor learning in
the brain (Figure 1).

Figure 1. Induction of hypoxia-inducible factor 1 (HIF-1) target genes by hypoxia. Cellular hypoxia inhibits
the activity of prolyl hydroxylases. Stabilization of HIF-1a leads to its phosphorylation and subsequent
dimerization with ARNT (HIF-1b). The HIF-1a/HIF-1b dimer (HIF-1) then binds to p300/CBP and this
complex binds HRE elements found on all HIF-1 target genes. This molecular sequence is required for the
activation of HIF-1-controlled transcription and leads to the activation of several HIF-1 targets including
iNOS, VEGF, EPO, GLUT1, several glycolytic enzymes, and others. Adapted from Sharp and Bernaudin,
2004.
4

Structurally, HIF-1 is a heterodimeric transcription factor composed of two subunits – HIF-1a and
HIF-1b (Chowdhury et al., 2008). Both subunits are basic helix-loop-helix (bHLH) proteins of the PER-
ARNT-SIM (PAS) family. HIF-1b is constitutively expressed in the nucleus of all cells and is required for
HIF-1-mediated transcription (Wang et al., 1995; Wood et al., 1996). Under physiological oxygen tension,
HIF-1a is continuously synthesized, but rapidly degraded. The underlying molecular basis for this regulation
is oxygen-dependent hydroxylation of two proline residues (Pro402 and Pro564) within the oxygen-
dependent degradation domain (ODD) by HIF-1a prolyl hydroxylases (Bruick and McKnight, 2001).
Hydroxylation of HIF-1a allows for the binding of the von Hippel Lindau (VHL) protein which is the
recognition component of an E3 ubiquitin ligase system that targets HIF-1a for proteasomal degradation
(Ivan et al., 2001). However, under hypoxia conditions, prolyl hydroxylation is inhibited and HIF-1a stabilizes
and accumulates within the cytoplasm exponentially (Jiang et al., 1996). HIF-1a then translocates to the
nucleus where it dimerizes with HIF-1b and the transcriptional coactivator p300/CBP (CREB (cyclic-AMP-
response element-binding protein) binding protein), forming the active transcription complex. Once
stabilized and activated, HIF-1 binds to the hypoxia-response element (HRE) consensus site – 5’RCGTG-
3’ – (where R is A or G), and elicits transcriptional activation of nearly 100 different oxygen-sensitive genes
(Wenger et al., 2005) (Figure 2).

5

Figure 2. Hypoxia-inducible factors control the expression of a wide range of genes. The complete number
of genes transcriptionally activated by HIF-1 may exceed 200 as demonstrated by recent microarray
analyses reported in the literature. This figure demonstrates that HIF-1 regulates the expression of a wide
range of genes controlling cellular and molecular processes such as energy metabolism, angiogenesis,
molecular transport, cell growth and apoptosis, vasomotor control, and others. Adapted from Chowdhury et
al., 2008.

Over the past couple of decades, HIF-1 has been highlighted as the master transcriptional regulator
of the adaptive response to hypoxia, and has been implicated in the pathophysiology of several
neurodegenerative disorders (Correia and Moreira, 2010; Zhang et al., 2011), cardiovascular disease and
6

stroke (Kasivisvanathan et al., 2011), and cancer biology (Quaegebeur and Carmeliet, 2010; Masoud and
Li, 2015). Perhaps its best characterized role is in the regulation of energy metabolism. Under hypoxic
conditions, cells undergo a metabolic shift from oxidative phosphorylation toward anaerobic glycolysis
(Ebert et al., 1995; Okino et al., 1998; Greer et al., 2012). This altered metabolic profile is controlled by HIF-
1 and is characterized by (1) upregulation of glucose transporter 1 expression (GLUT1; increased glucose
delivery), and (2) increased expression of lactate dehydrogenase (LDHA), which catalyzes the bidirectional
conversion of pyruvate to lactate (Semenza, 2002; Greer et al., 2012). Interestingly, both GLUT1 and LDHA
expression have been shown to be increased in the brain with exercise (Kinni et al., 2011). While increased
brain glucose delivery and lactate consumption are thought to be important for maintaining homeostasis
during intense physical activity, it remains to be determined whether or not HIF-1 mediates this metabolic
adaptation under physiological exercise conditions.

In addition to its role in regulating energy metabolism, HIF-1 has also been shown to have an
important role in facilitating angiogenesis (Sharp and Bernaudin, 2004; Adams and Alitalo, 2007).
Angiogenesis is a tightly controlled physiological adaptation in which new blood vessels are formed from
preexisting vascular networks. While few studies support the effects of exercise on the cerebrovasculature
in animal models of PD, studies have reported that exercise can improve cerebral blood flow (CBF) and
increase levels of vascular endothelial growth factor (VEGF) in the intact brain (Black et al., 1990; Rhyu et
al., 2010). Importantly VEGF is also and HIF-1 target gene (Shweiki et al., 1992). Thus, exercise may
promote neuroplasticity by altering blood flow and angiogenesis through a HIF-1-mediated signaling
pathway.

Interestingly, studies from Holschneider and colleagues have investigated changes that occur in
regional cerebral blood flow (CBF) with exercise in both normal rodents and models of neurodegenerative
disorders (Wang et al., 2013; Wang et al., 2016). These studies have demonstrated that exercise leads to
selective activation of rCBF in rodents who participate in exercise on a motorized running wheel. For
example, significant activation is seen within a number of distinct regions including the prefrontal cortex,
dorsal striatum, motor cortex, and the vermis of the cerebellum. Importantly, other regions of the brain either
show no change or a reduction in rCBF. Furthermore, these blood flow studies are corroborated by
7

molecular studies that indicate dorsostriatal and prefrontal activation specific subunits of AMPA receptors,
as well as electrophysiological measures of plasticity and morphological changes such as restoration of
dendritic spine density (Kintz et al., 2013; Toy et al., 2014). This described association between rCBF
changes, and molecular and morphological changes supports the notion that exercise mediates circuit-
specific change in neuroplasticity.

8

CHAPTER 2: The effects of exercise on dopamine neurotransmission in Parkinson’s disease:
targeting neuroplasticity to modulate basal ganglia circuitry

Authors: Giselle M. Petzinger, Daniel P. Holschneider, Beth E. Fisher, Sarah McEwen, Natalie Kintz,
Matthew R. Halliday, William Toy, John W. Walsh, Jeff Beeler, and Michael W. Jakowec

INTRODUCTION

This chapter presents an overview of the impact of exercise on neuroplasticity in animal models of
Parkinson’s disease (PD). Neuroplasticity is the ability of the brain to encode and learn new behaviors and
can be defined as changes in molecular and cellular processes in response to environmental experiences
such as exercise. We briefly explore the effects of exercise in the basal ganglia (called the striatum in
rodents), pertinent neurotransmitter systems and associated cortical circuitry. While this brain area and
related circuitry are known to be impaired in individuals with PD, exercise may help to restore the normal
motor and cognitive function observed in healthy individuals. Exercise has been shown to affect a number
of different neurotransmitters including dopamine (Fisher et al., 2004; Petzinger et al., 2007), glutamate
(Fisher et al., 2004; Real et al., 2010; VanLeeuwen et al., 2010), serotonin (Brown et al., 1979; Blomstrand
et al., 1989), norepinephrine (Brown et al., 1979; Semenova et al., 1981; Pagliari and Peyrin, 1995), and
acetylcholine (Nakajima et al., 2003; Uchida et al., 2006) potentially contributing to the exercise-related
benefits observed in PD. This review will focus on two neurotransmitter systems that are essential for
normal corticostriatal connectivity and function. Namely, exercise effects in dopamine (DA) and glutamate
neurotransmission as well as neuronal connectivity (dendritic morphology) in basal ganglia circuits will be
addressed. Additionally, while a wide variety of exercises have been reported to be beneficial in PD, this
review will also highlight recent studies that compare the type of exercise. By way of differential effects on
blood flow and neurogenesis, skilled vs. aerobic exercise may each have a distinct impact on
neuroplasticity. These differential effects which are brain region and circuit specific suggest a potential
interaction between the type of exercise and its impact on induced neuronal activation and regional blood
flow that may be important for facilitating repair or disease modification. Understanding the impact of
exercise in the basal ganglia and its related circuitry may represent a new frontier in understanding
mechanisms of neuroplasticity and repair and thus lead to novel therapeutics for PD.

9

PARKINSON’S DISEASE AND EXERCISE AS A MODEL FOR NEUROPLASTICITY

Parkinson’s disease is a progressive neurodegenerative disorder that is characterized by the
depletion of DA due to the degeneration of neurons in the substantia nigra pars compacta (SNpc), and to
a lesser degree the ventral tegmental area (VTA). Characteristic features of PD include motor
(bradykinesia, rigidity, tremor, gait dysfunction, and postural instability) and cognitive impairment (frontal
lobe, executive dysfunction), as well as mood disorders. In PD, studies in exercise and neuroplasticity have
focused on the basal ganglia and its cortical connections, since they comprise important motor and
cognitive circuits, respectively, that are altered in disease. The basal ganglia consist of the putamen and
caudate nucleus, collectively termed the striatum in rodents. The striatum is composed of DA-D1R and DA-
D2R-containing medium spiny neurons (MSNs) of the direct and indirect projection pathways, respectively.
Synaptic connections between DA-D1R and DA-D2R-containing MSNs and cortical glutamatergic neurons,
make up cortical-striatal circuits (Kreitzer and Malenka, 2008). In the healthy brain, these circuits are
responsible for automatic (unconscious) and volitional (goal-directed) movements as well as cognitive
processes, including executive function (EF) (Leh et al., 2010). Executive function consists of working
memory, task flexibility, and problem solving, as well as planning and execution of tasks (Dirnberger and
Jahanshahi, 2013). The key circuits affected in PD are (i) the cortico-striatal motor circuit, including the
dorsal lateral striatum (analogous to the putamen in primates), the primary motor and somatosensory cortex
and the thalamus, and (ii) the frontal-striatal circuit, including the prefrontal cortex and the dorsal medial
striatum (analogous to the caudate nucleus in primates). In Figure 3 we depict the two major cortico-striatal
circuits discussed in this review, their convergence in the striatum and modulation through DA. In PD the
early and more profound DA-depletion occurs in the dorsal lateral striatum, thus leading to early deficits in
automatic execution of routine movements (Dirnberger and Jahanshahi, 2013; Engeln, 2013). Imaging
studies suggest that as individuals with PD lose control of automatic movements that there is a shift towards
frontal-striatal volitional control of motor performance (Filoteo et al., 2014). It has been posited that deficits
in EF that are common even in early stages of PD may be due I part to an over recruitment and saturation
of the frontal-striatal circuit (Floresco, 2013). Alternatively, lesion studies have supported that direct
impairment of the dorsal striatum may lead to disruption of the frontal-striatal circuit and thus EF deficits
directly (Courtière et al., 2011). In addition to their role in PD, the two circuits described above as well as
10

DA receptors are also important in motor learning (Yawata et al., 2012). Specifically, the volitional and
automatic circuits and the DA-D1R and DA-D2R are involved in the acquisition phase of motor learning while
the automatic circuits and the DA-D2R are involved in the retention phase of motor skill learning. Exercise
that incorporates aspects of motor learning, such as skill (e.g., yoga, tai chi, treadmill running) may be
useful for examining exercise-induced mechanisms of neuroplasticity in PD.

Figure 3. Dopamine (DA) projections play a critical role in modulating both motor and cognitive circuits.
Dopamine (DA) from neurons within the substantia nigra pars compacta and ventral tegmental area of the
midbrain project to the dorsal lateral striatum of the basal ganglia and the prefrontal cortex, respectively.
The earlier and more profound depletion of DA in the dorsal lateral striatum results in impairment in
corticostriatal thalamic circuitry, which is important for automatic movements, and consequently greater
reliance on frontal striatal circuitry, important for goal-directed motor control in Parkinson’s disease (PD).
Although affected to a lesser degree, DA loss in the frontal-striatal circuit contributes to cognitive
11

impairments in PD. Animal studies are beginning to reveal evidence for exercise-induced neuroplasticity in
motor and cognitive related circuitry in PD and how the two circuits are inter-related.

EPIDEMIOLOGICAL STUDIES OF EXERCISE EFFECTS OF PD

Physical activity has been demonstrated to lead to tremendous health benefits in individuals of all
ages and in both healthy and disease states. It is only in the last two decades that epidemiological studies
have suggested that a lifetime of physical activity may provide protection from a wide range of neurological
disorders, including PD (Thacker et al., 2008), Alzheimer’s disease (AD) (Pitkälä et al., 2013), and cognitive
impairment associated with aging (Voelcker-Rehage et al., 2011). For example, a study by Chen and
colleagues demonstrated that maintaining strenuous levels of physical activity in young adulthood was
associated with a reduced risk of acquiring PD in later life (Chen et al., 2005). One potential mechanism by
which exercise may reduce an individual’s risk for common neurodegenerative disease, or age-related
cognitive decline is through enhanced brain connectivity, with concomitant increased reserve and resilience
to age-related synaptic deterioration. These exercise-induced changes in brain connectivity may occur at a
molecular and circuit level and include essential components that drive neuroplasticity: neurotransmission,
synaptogenesis, and neurogenesis. While, evidence suggests that impaired function in PD can be improved
through rehabilitation and exercise, there remains a significant gap in understanding exercise-induced
neuroplasticity in the context of a neurodegenerative disorder, such as PD. To elucidate the underlying
mechanisms of exercise-related functional improvement in people with PD, researchers have primarily
utilized rodent models, including the neurotoxin-induced 1-methyl-4-phenyl-1, 2, 3, 6-tetrahydropyridine
(MPTP) and 6-hydroxydopamine (6-OHDA) models (Schwarting and Huston, 1996; Jakowec and
Petzinger, 2004). These models are helpful as a means to investigate exercise-induced mechanisms of
neuroplasticity and brain repair in PD since they exhibit analogous pathophysiological and behavioral
characteristics of PD: (i) loss of midbrain dopaminergic neurons, (ii) depletion of striatal DA, (iii) aberrant
corticostriatal connectivity and (iv) impaired cognitive and motor performance.

12

EXERCISE EFFECTS OF DA NEUROTRANSMISSION IN ANIMAL MODELS

In PD, the classical pathophysiological model is that the loss of DA in the dorsal lateral striatum
leads to imbalance of the DA-D1R direct and DA-D2R indirect pathways, such that there is increased and
aberrant corticostriatal glutamatergic synaptic drive and hyperexcitability in the DA-D2R-containing indirect
pathway. It has been posited that restoration of DA neurotransmission along this DA-D2R pathway may
serve to normalize this aberrant form of corticostriatal synaptic plasticity. Animal studies support that
exercise benefits in PD may be due in part to facilitated DA neurotransmission. Specifically, using the MPTP
mouse model of PD, intensive daily treadmill exercise leads to improved motor function and increased DA
neurotransmission compared to the non-exercise MPTP mice. While both MPTP mice groups showed equal
levels of cell loss and DA-depletion, only exercise mice showed: (i) increased evoked DA release, (ii)
increased extracellular DA through downregulation of the dopamine transporter (DAT) expression, and (iii)
decreased clearance using fast-scan cyclic voltammetry within the dorsal striatum (Fisher et al., 2004;
Petzinger et al., 2007). In the context of motor learning and its potential role in exercise and rehabilitation-
related benefits in PD, studies in mice report that DA availability can influence motor learning (rotarod
training). Specifically, PitX3 (paired-like homeodomain transcription factor 3) transgenic mice that lack
striatal DA due to developmental loss of nigrostriatal dopaminergic neurons show deficiencies in motor
learning (Beeler et al., 2010). Conversely, restoration of DA through levodopa treatment in these mice
restores motor learning (Beeler et al., 2010). Thus, exercise effects on DA availability through altered
neurotransmission may act in part to promote mechanisms critical for motor learning and important for
restoring motor behavior in PD.

Another mechanism by which exercise can influence DA neurotransmission is through DA receptor
expression (Beaulieu and Gainetdinov, 2011). For example, exercise studies in rodents have demonstrated
increased DA neurotransmission through an increase in DA-D2R protein expression and binding within the
dorsal lateral striatum (Fisher et al., 2004). Specifically, after 28-days of intensive treadmill training in MPTP
mice, DA-D2R protein expression was increased, with no reported change in the DA-D1R. Treadmill
exercise also resulted in an increase in DA-D2R transcript within MSNs of the dorsal striatum supporting
the regulatory role of exercise at the level of gene expression (Fisher et al., 2004; Petzinger et al., 2007).
Using positron emission tomography imaging with [
18
F]-fallypride, a ligand with high specificity for DA-D2R,
13

this effect of exercise in MPTP mice was also observed through increased DA-D2R binding (Vučković et
al., 2010). These reports are consistent with studies that demonstrate an exercise-induced increase of DA-
D2R mRNA, protein, and binding in the striatum of healthy non-dopamine depleted rodents (Gilliam et al.,
1984; MacRae et al., 1987; Foley and Fleshner, 2008). Translating these MPTP animal findings to clinical
studies, an exercise-induced increase in DA-D2R expression was also observed in individuals newly
diagnosed with PD (Fisher et al., 2013). After an 8-week regimen of intensive treadmill training, subjects
who underwent PET-imaging demonstrated an 80% increase in binding of [
18
F]-fallypride within the dorsal
caudate nucleus compared to pre-exercise baseline values (Fisher et al., 2013). While the relationship
between exercise-induced motor benefits in PD and increased DA-D2R function and its role in the
establishment and maintenance of motor skill learning, may underlie this benefit (Yim et al., 2009; Beeler
et al., 2012). For example, electrophysiological studies within the striatum of animals, in conjunction with a
pharmacologically specific blockade of DA-D2Rs, have shown that antagonism of DA-D2R in either early or
late phases of motor skill learning leads to impairment in glutamatergic-dependent synaptic potentiation
and motor learning (Yim et al., 2009). These studies also demonstrate that DA-D2R related synaptic
plasticity that is responsible for motor learning is localized to the dorsal striatum. Further support for the
role of DA-D2R and motor learning comes from studies in rodents by Beeler and colleagues (Beeler et al.,
2010; Beeler et al., 2012). These researchers demonstrate that the DA-D2R is also important in the
maintenance of learned motor behaviors since pharmacological blockade of the DA-D2R, and not the DA-
D1R, in rodents leads to loss of a learned motor skill.

In addition to its role in motor performance, preliminary studies in animals suggest that an exercise-
induced increase in dorsal striatal DA-D2R expression may also contribute to the reported exercise related
improvements in executive function, including behavioral flexibility (Eddy et al., 2014). Specifically, studies
have shown that 18 days of exercise can improve discrimination testing in a set-shifting, cross-maze task
in healthy rodents. This exercise benefit was reversed through selective pharmacological blockade of the
DA-D2R. Taken together, animal studies support that exercise-induced increase in DA availability along
with increased DA-D2R expression in the dorsal striatum and its related cortical circuitry may contribute to
exercise-related effects in neuroplasticity and behavioral benefits in PD. Future studies in humans are
clearly needed to confirm this relationship.
14

EXERCISE EFFECTS ON GLUTAMATE NEUROTRANSMISSION IN ANIMAL MODELS

Glutamate neurotransmission is also important in synaptic function and especially in learning and
memory as demonstrated by its role in mediating both long-term potentiation (LTP) and long-term
depression (LTD) (Bliss and Cooke, 2011; Huganir and Nicoll, 2013). These electrophysiological properties
of synaptic connectivity are dictated by specific receptor subtypes, especially the NMDA (N-methyl-D-
aspartate) and AMPA (a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) receptor, not only through
long-term synapse-specific expression, but also by fast trafficking from intracellular stores to sites within
the postsynaptic density (Huganir and Nicoll, 2013). Changes in glutamate receptor subtype expression
(i.e. neuroplasticity) and localization on neuronal electrophysiological properties are the direct result of
experience-dependent events, including exercise. For example, using the MPTP-lesioned mouse model,
we have reported exercise-induced changes in synaptic expression of specific receptor subunits of the
AMPA receptor (Kintz et al., 2013). As previously mentioned above, DA-depletion leads to structural and
functional changes in striatal MSNs, including loss of predominantly cortico-striatal MSNs, including the
loss of predominantly cortico-striatal synaptic connections (in both direct and indirect projection pathways)
and increased glutamatergic drive in remaining cortico-striatal synaptic connections (Calabresi et al., 1993).
In the MPTP mouse, exercise is able to reverse this aberrant hyperactive glutamatergic state by two would-
be processes. First, exercise alters glutamatergic receptor subunit expression, especially the AMPA
receptor subunit GluA2, particularly localized to indirect DA-D2R-containing MSNs (Kintz et al., 2013). On
striatal MSNs, exercise increases the relative expression of GluA2, a calcium impermeable AMPA receptor
subunit type, from the calcium permeable AMPA receptor GluA1. Electrophysiological correlates
demonstrate that exercise reduces synaptic excitability and excitatory post synaptic potentials (Kintz et al.,
2013). Second, exercise reduces the presynaptic storage of glutamate, as measured the electron
microscopy (Fisher et al., 2004). Taken together, exercise reduces aberrant glutamatergic drive, thus,
restoring cortico-striatal circuit function.

15

EFFECTS OF EXERCISE ON DENDRITIC SPINE DENSITY IN THE BASAL GANGLIA

In addition to function changes, the loss of DA leads to morphological changes in glutamatergic
synapses, including a decrease in dendritic spine density and disruption of connectivity in the motor circuit
(Elsworth et al., 2013). Dendritic spine loss of MSNs has been reported in post-mortem tissues of patients
with PD, as well as in the 6-OHDA and MPTP rodent models of DA depletion (Pickel et al., 1992; Smith et
al., 2009; Villalba et al., 2009). In addition, studies have suggested that dendritic spine loss occurs
predominantly on the DA-D2R-containing MSNs early after DA depletion (Cazorla et al., 2012; Fasano et
al., 2013), but others have shown that spine loss occurs on both DA-D2R and DA-D1R following prolonged
DA depletion (Villalba et al., 2009). One possible effect of exercise toward restoration of the circuitry of the
basal ganglia may be through changes in spine density. Studies in healthy rodents subjected to
environmental enrichment, voluntary wheel running, and forced treadmill running paradigms have
demonstrated an increase in dendritic spine density in cerebellar Purkinje, CA3 hippocampal pyramidal,
and layer III cortical neurons (Eadie et al., 2005; Kozorovitskiy et al., 2005; Leggio et al., 2005; Nelson and
Iwamoto, 2006; Hu et al., 2010). Studies in MPTP mice have shown that intensive treadmill running can
reverse the loss of dendritic spines on striatal MSNs (Toy et al., 2014). Besides effects on dendritic spine
density, intensive exercise also leads to the restoration of synapses as indicated by the elevated expression
of both presynaptic (synaptophysin) and postsynaptic (PSD-95) proteins.

EXERCISE TYPE, BLOOD FLOW, AND NEUROGENESIS

Exercise type can be loosely categorized into predominantly skilled or aerobic exercise. Aerobic
exercise is a system of conditioning aimed at enhancing circulatory and respiratory efficiency that improves
the body’s use of oxygen through vigorous, sustained exercise such as running, swimming, or cycling. This
is in contrast to skilled exercise, which is a form of goal-oriented movement in which temporal and/or spatial
accuracy is important for achieving pre-determined objectives. The important relationship between the type
of exercise and nature of neuroplasticity-related changes is underscored by prior work suggesting that rats
that have undergone unskilled and repetitive exercise (aerobic exercise) have an increase in the density of
capillaries in the brain’s motor regions, without an increase in synaptic numbers (as measured by dendritic
spine density (Black et al., 1990; Kleim et al., 2002; Garcia et al., 2012). This is in contrast to rats that have
16

learned new motor skills (skilled exercise) and have a greater number of synapses per neuron, without an
increase in the density of capillaries. Recent studies in animal models, including PD, have begun to further
elucidate the differential effects of skilled versus aerobic exercise on neuroplasticity associated with
alterations in blood flow. These differential effects of skills versus aerobic exercise are observed at the level
of anatomic specificity (circuit and brain region). Specifically, recent work by our laboratory suggests that
skilled compared to non-skilled aerobic training differentially affects functional activation of the medial
prefrontal cortex in parkinsonian rats during walking (Wang et al., 2015b). Rats with bilateral, striatal 6-
OHDA lesions were exposed to forced exercise for 4 weeks, either on a simple running wheel, considered
a form of non-skilled aerobic exercise (AE), or on a complex wheel with irregularly spaced rungs, a form of
skilled aerobic exercise (SAE). Cerebral perfusion was mapped during horizontal treadmill walking or at
rest using [
14
C]-iodoantipyrine autoradiography, one week after the completion of exercise. SAE compared
to AE resulted in greater increases in regional cerebral blood flow (rCBF) during walking and at rest in the
prefrontal cortex (prelimbic area). Seed correlation analysis during locomotor walking revealed that SAE
compared to AE resulted in a much broader functional connectivity of prefrontal cortex with the striatum
providing evidence of frontal-striatal neuroplasticity in these circuits through exercise. In addition, there was
also evidence for changes in functional connectivity involving the primary and secondary motor cortices,
and primary somatosensory cortex. Lastly, prelimbic cortical activation correlated with restoration of motor
function in lesioned rats undergoing skilled aerobic exercise more than with non-skilled aerobic exercise.
These results show for the first time that SAE compared to AE results in enhancement of prefrontal cortical
mediated control of motor function. We propose that the SAE paradigm likely required greater effort in motor
preparatory processing, motor control and set shifting than that required for the AE, all key roles ascribed
to prelimbic cortex (Risterucci et al., 2003; Rosano et al., 2012). This suggests that the prefrontal cortex
and its associated pathways are a central target for experience-dependent neuroplasticity as a result of
SAE. It remains to be proven whether such recruitment of prefrontal cortex by SAE will improve performance
of the 6-OHDA rat during set-shifting tasks. If proven, this would confirm the notion that motor rehabilitation
programs for PD patients should include a relatively high cognitive demand, such that by forcing patients
to practice task-switching over a sufficient number of practice trials, they might be able to overcome their
inability to generalize learned actions to different environmental contexts (Onla-or and Winstein, 2008;
17

Petzinger et al., 2013). In addition, future research will need to examine whether any recruitment of
prefrontal cortex by SAE is due to changes in dopaminergic pathways. While dopaminergic dysfunction in
prefrontal cortex is an early feature of PD and has been linked to dopaminergic loss in the caudate and
substantia nigra (Kaasinen et al., 2001; Carbon and Eidelberg, 2006; Sawamoto et al., 2008; Berti et al.,
2010; Polito et al., 2012), its role in shaping cognitive deficits and responding to an exercise intervention
remains to be determined.

In the healthy rodent brain, studies are beginning to demonstrate that the types of exercise and
activity can differentially influence the various stages of neurogenesis including cell migration,
differentiation, maturation, and circuit integration (Klaus et al., 2009; Clark et al., 2012). Neurogenesis is
the birth of new neurons. Within the adult mammalian brain there are several unique regions that display
the birth of new cells throughout life, including the granular cell layer of the hippocampus, the subventricular
zone, and the prefrontal cortex (Lie et al., 2004). It is well established that exercise (and environmental
enrichment, especially those designs that incorporate running wheels) enhances neurogenesis in the
healthy rodent brain (van Praag et al., 1999; Opendak and Gould, 2015). For example, in both young and
age mice, voluntary wheel running has been positively correlated with increased hippocampal neurogenesis
and improved memory as demonstrated by enhanced water maze performance (van Praag et al., 1999;
Kempermann et al., 2002). Interestingly, rodents exposed to an enriched environment that incorporates
aspects of skilled activity and cognitive engagement compared to rodents exposed to a voluntary running
wheel show greater cognitive flexibility in the Morris Water Maze. This improvement may be due to both
neurogenesis and enhanced neuronal incorporation into hippocampal circuitry (Garthe et al., 2016). Such
studies suggest that while many different types of physical activity promote the survival of these newborn
cells, migration, and integration may be dependent on the degree of cognitive (skilled) engagement
(Bednarczyk et al., 2011; Curlik and Shors, 2013). The effect of skilled exercise on stages of neurogenesis
may be due to its influence on the proliferation of astrocytes, activation of microglia, and expression of
factors, such as neurotrophic factors and its receptors, which are known to be important in regulating
neuroplasticity and synaptogenesis in brain regions where the birth of new cells are promoted (Ehninger
and Kempermann, 2003; Kohman et al., 2012; Bernardi et al., 2013).

18

In rodent models of neurological disorders where reduced neurogenesis is evident, including
mouse models of Alzheimer’s disease, exercise has been shown to elevate hippocampal neurogenesis and
delay deficits in learning and memory (Kim et al., 2015). However, some studies have shown elevated or
no effects of exercise in neurogenesis in the context of a disease, such as Huntington’s disease (Kohl et
al., 2007; Catlow et al., 2009; Potter et al., 2010). In animal models of PD, exercise may facilitate
neurogenesis in the hippocampus and subventricular zone, similar to reports with wild-type animals,
however, there a few reports showing enhanced neurogenesis within the damaged striatum or midbrain
regions with physical activity (Steiner et al., 2006; Klaissle et al., 2012). Clearly a major gap in our
knowledge is whether exercise or the type of exercise influences different stages in neurogenesis in brain
regions affected by disease.

EXERCISE AND THE COUPLING OF NEURONAL ACTIVATION AND CEREBRAL BLOOD FLOW

As described in previous sections, skilled exercise may lead to the recruitment and activation of
neurons in specific circuits within the brain. On the other hand, aerobic exercise may have more global
effects on the entire brain including lowering the threshold for neuroplasticity to occur through the
expression of neurotrophic factors or other modulators of synaptic plasticity as well as increasing rCBF
(Cotman et al., 2007). However, the activation of neurons through engagement in skilled exercise and the
modulation of blood flow may not be mutually exclusive processes. Rather they may promote and regulate
neuroplasticity through overlapping and integrated mechanisms. One potential scenario may be that
intensive skilled exercise with a resultant increase in neuronal activity to a specific circuit (a motor circuit
for example) may result in elevated demand for regional oxygen consumption (resulting in oxygen depletion
within this region). Elevated oxygen consumption in turn, can activate a number of regulatory signals that
respond to changes in metabolic expenditure. For example, hypoxia-inducible factor-1 (HIF-1), and oxygen
sensitive transcription factor, is activated under conditions of low tissue oxygenation, a consequence of
increased metabolic demand (Correia and Moreira, 2010). Acute, or moderate to intensive aerobic exercise
has been shown to induce transient cerebral hypoxia, which is largely sensed by HIF-1 (Kinni et al., 2011).
Important to neuroplasticity, HIF-1 regulates the expression of a wide array of downstream target genes
implicated in promoting neurogenesis, synaptogenesis, and angiogenesis (Sharp et al., 2004). Thus,
19

exercise, through orchestrating the recruitment of circuitry, high neuronal activity along with increasing
cellular metabolic energy demand, leads to the activation of a cascade of genes important for
neuroplasticity, repair, and the establishment of homeostasis. These mechanisms linking exercise and
neuroplasticity may also involve increased rCBF to activated brain regions leading to a number of important
consequences including: (i) increasing the availability of biomolecules responding to increase energy
demand, (ii) removal of waste materials and maintenance of cellular homeostasis, (iii) increased delivery of
neurotrophic factors such as BDNF, (iv) altering the blood-brain barrier to allow the targeted passage of
biomolecules and circulating cells such as macrophages to activated sites, and (v) delivery of biomolecules
involved in the formation of synaptic connections. Thus, exercise may incorporate either of both
mechanisms to facilitate neuroplasticity. A major gap in our knowledge is the precise cause-effect
relationship between elevated metabolic demand and altered rCBF. Metabolically demanding neuronal
circuits can release nitric oxide synthase (NOS) and angiogenic factors to increase blood flow to sites where
there is demand (Fabel et al., 2003; Toda et al., 2009; Viboolvorakul and Patumraj, 2014). On the other
hand, metabolically active neuronal circuits may reinforce region increases in CBF.

CONCLUSIONS

Animal studies have been instrumental in providing evidence for exercise’s role in neuroplasticity
of corticostriatal circuits that are profoundly affected in PD. This evidence includes exercise’s role in
modulating DA and glutamate neurotransmission, synaptogenesis, and increased rCBF. In addition, recent
evidence supports that the type of exercise may have regional effects on brain circuitry, with skilled exercise
differentially affecting frontal-related circuits more so than pure aerobic exercise. Although clearly more
research is needed to address major gaps in our knowledge, we hypothesize that skilled compared to
aerobic exercise has different effects on neuroplasticity, but that these effects may not be mutually
exclusive. For example, the potential effects of different types of exercise on inducing neuroplasticity in a
circuit specific manner may occur through synergistic mechanisms that include the coupling of an increasing
neuronal metabolic demand with a corresponding increase in regional blood flow. Thus, both types of
exercise may be important for facilitating neuroplasticity. In Figure 4 we illustrate that most exercises lie
within a spectrum between aerobic and skilled exercise. For example, peddling on a recumbent bicycle
20

may be considered predominantly aerobic with minimal skill or cognitive engagement. On the other end of
the spectrum, juggling may represent a highly skilled task with minimal aerobic involvement. However,
many exercises such as swimming and running involve a combination of both skilled and aerobic exercise.
Elucidation of the relative contribution of different types of exercise on neuroplasticity and motor and
cognitive improvement in PD may provide mechanistic insights important to facilitate brain repair and modify
disease progression.

21

Figure 4. Physical activity spans the spectrum from aerobic to skilled exercise. Recent exercise studies in
animal models of PD are beginning to support the differential effects of aerobic versus skilled exercise on
the establishment and maintenance of brain circuitry. In this Figure we illustrate these concepts. One
potential hypothesis highlights aerobic exercise that may lead to a broad increase in cerebral blood flow,
including within those brain circuits in the basal ganglia and cerebellum involved in motor control. Other
global factors may also be activated including reduced oxidative stress, reduced neuro-inflammation, and
increased expression of neurotrophic factors. This is in contrast to skilled exercise that entails perceptual
and a higher-level cognitive processing that may specifically target prefrontal and associated cortical circuits
important for executive function.

22

CHAPTER 3: Exercise enhances neuroplasticity by targeting synaptic energy metabolism and
mitochondrial function in animal models of Parkinson’s disease

Authors: Matthew R. Halliday, Adam J. Lundquist, Nicolaus A. Jakowec, Giselle M. Petzinger, and Michael
W. Jakowec

INTRODUCTION

Energy metabolism, primarily the conversion of glucose to adenosine triphosphate (ATP), is
fundamental to the function of all cells including those within the nervous system. Despite its high metabolic
demand, the human brain has little reserve when substrates are limited, and the consequences of failing to
supply neurons with energy can be severe, resulting in dysfunction in neurotransmission and cell death.
The analysis of brain tissues from a wide spectrum of neurodegenerative disorders including Parkinson’s
disease (PD), Alzheimer’s disease (AD), Huntington’s disease (HD), and amyotrophic lateral sclerosis
(ALS) all implicate deficits in the machinery of energy metabolism, especially in the mitochondria. Studies
in animal models used for research in these human brain disorders support the key role of energy
metabolism and mitochondrial function. Consequently, several pharmacological therapeutics have been
tested in these disorders targeting mitochondrial function with the hope of restoring brain function. To date,
no successful drug treatments targeting neuroenergetics have translated to the clinical setting. Despite
these challenges, energy metabolism remains a viable therapeutic target with a major gap in our knowledge
as to the precise role it plays in neurodegenerative disorders. A major goal of current research is to better
understand the underlying molecular and biochemical changes that take place in glucose handling within
the brain under conditions of metabolic challenges like transient hypoxia, acute and chronic injury, and
degenerative disease.

Over the past several years, exercise has become attractive as a non-pharmacological and
inexpensive treatment modality for many brain disorders. The vast majority of patients with PD incorporate
aspects of physical activity and exercise as part of their regular standard of care. Small studies and patient
reports support the benefits of exercise in improving quality of life, optimizing drug treatment, and possibly
modifying disease progression. Research in animal models (some points highlight below) has provided
support for the beneficial effects of exercise and is beginning to identify the underlying molecular
mechanisms by which exercise alters motor behavior in PD. Even in light of these major advancements in
23

elucidating the impact of exercise in neurodegenerative disorders, major gaps in our knowledge still exist –
especially in understanding the differences linking neuronal activation and energy utilization in conditions
of dopamine depletion (the major neurotransmitter affected in PD), aging, and how they may occur in a
circuit-specific manner in different regions of the brain.

The purpose of this short review is to highlight the hypothesis that physical activity in the form of
different types of exercise (aerobic versus skill-based) is in part beneficial due to its circuit-specific effects
on synaptic connectivity, a major component of neuroplasticity. Specifically, in this review we touch upon
the effects of exercise on energy metabolism at the synapse and its ability to modulate synaptic
mitochondrial function, glucose uptake, astrocyte-neuron coupling, and the mechanism that may link central
nervous system (CNS) and peripheral (muscle) activity. Together, a better understanding of these
processes will provide a foundation for rational pharmacological drug development, guide clinicians in
prescribing the most effect forms of exercise to treat deficits in disorders like PD and give us insight into
how long-term application of exercise may be necessary as a critical component of lifestyle to better
promote optimal brain function throughout life including the later stages of aging.

EFFECTS OF EXERCISE IN AN ANIMAL MODEL OF PARKINSON’S DISEASE

Epidemiological studies have shown that exercise and other forms of physical activity reduces the
incidence of PD, AD, and mild cognitive impairment (MCI) (Chen et al., 2005; Öhman et al., 2016). To better
understand the underlying molecular mechanisms by which exercise provides benefits in PD, animal
models, especially those utilizing the dopamine depleting agents MPTP (1-methyl-4-phenyl-1, 2, 3, 6-
tetrahydropyridine) and 6-OHDA (6-hydroxydopamine), have been instrumental. Exercise in animal models
has typically been applied in the form of running on a horizontal motorized treadmill or in a voluntary running
when in a home cage. Currently there is no consensus on the parameters of exercise in terms of duration,
velocity, intensity, and level of skill. Exercise introduced prior to or during the active phase of toxin-induced
cell death can provide protection from both the death of midbrain dopaminergic neurons and the depletion
of striatal dopamine. A number of mechanisms may mediate neuroprotection including (i) elevation of
neurotrophic factors such as BDNF (brain-derived neurotrophic factor), (ii) increased resilience of
mitochondria including ATP production or protection from oxidative stress, (iii) reduced bioavailability of
24

toxin due to alterations in peripheral processing (liver metabolism), transport across the blood-brain barrier,
uptake into dopaminergic neurons (through reduced expression of the dopamine transporter, DAT), or
storage/sequestration via vesicular uptake through VMAT2 (vesicular monoamine transporter) into vesicles.

In contrast to the utilization of dopaminergic toxins to investigate aspects of neuroprotection, these
toxin-induced models can also be used to study neurorestoration, specifically the effects of exercise on
neuronal connectivity when toxin-induced cell death is complete. In the case of MPTP, cell death is
complete 3 days following the final injection, while 6-OHDA-mediated cell death is typically complete within
3 to 4 weeks. In our studies, we are interested in understanding the underlying mechanisms by which
exercise, in the form of intensive treadmill running, leads to the restoration of motor behaviors in the MPTP-
lesioned mouse model of dopamine depletion, through processes collectively termed neuroplasticity. With
MPTP-induced cell death finalized within 3 days, we initiate treadmill running 5 days following administration
of the toxin; mice run 5 days per week, 1 hour per day, with speeds reaching 24 m/min (Fisher et al., 2004).
Using this regimen we have observed that exercise leads to (i) an elevation of striatal dopamine D2 receptor
expression, (ii) reduced expression of striatal DAT, (iii) elevated evoked release of dorsolateral striatal
dopamine despite no increase in overall dopamine levels, (iv) alterations in the relative expression of the
ionotropic glutamate receptor subunits GluA1 and GluA2, and (v) restoration of dendritic spine density on
striatal medium spiny neurons (MSNs) (Fisher et al., 2004; Petzinger et al., 2007; VanLeeuwen et al., 2010;
Kintz et al., 2013; Toy et al., 2014). Together these results are consistent with the reduction in
hyperexcitability of corticostriatal inputs to striatal MSNs indicating circuit-specific changes leading to
improved motor behavior. The concept of circuit-specific changes is further supported by recent studies
examining alteration in region cerebral blood flow (rCBF) with exercise. Using the tracer [
14
C]-iodoantipyrine
in the 6-OHDA rat, Holschneider and colleagues have shown an elevation in cortical rCBF specifically within
the PFC, and that this distribution is enhanced with skilled exercise compared to non-skilled aerobic
exercise (Wang et al., 2015a).

While exercise provides general brain health in a non-specific manner, we are interested in
understanding how exercise can be used to target specific brain circuits, especially those affected in
neurological disorders like PD. Physical activity provides benefits by reducing cognitive decline with normal
25

health aging (Voelcker-Rehage et al., 2010) and different types of activity have variable effects. For
example, forms of exercise with greater skills and learning components that engage cognitive circuits show
greater benefit than those that do not engage cognitive circuits but are predominantly aerobic in nature
(Voelcker-Rehage et al., 2011). This is important in patients with PD where cognitive deficits are evident in
the early stages of disease and may contribute to the degenerative process and lead to the progression of
severe motor deficits. Interestingly, we found that some of the early synaptic changes with exercise in the
MPTP-lesioned mouse model originate from cortical afferents, especially the corticostriatal circuit
connecting the cerebral cortex with the striatum (caudate-putamen). These observations, along with the
fact that deficits in cognition can interfere with gains in the benefits of exercise, highlight the importance of
circuit-specific targets that must be engaged to enhance improvements in motor behavior. In other words,
mechanisms of neuroplasticity enhanced by activating cognitive circuits are critical for targeting
improvement in motor circuits.

Therefore, major gap in our knowledge is the identity of the molecular mechanisms by which
exercise, especially forms of exercise targeting corticostriatal afferents through skill and cognitive
engagement, lead to circuit-specific changes that result in improved motor behaviors and may modify
disease progression. What are the mechanisms that link neuronal activation, neuroplasticity, and improved
motor behaviors? One attractive hypothesis remains mechanisms involved in metabolism. Metabolism
plays a critical role in linking neuronal activity-induced changes in blood flow, breakdown of energetic
substrates (e.g., glucose and lactate) with increase expression of genes and proteins involved in
angiogenesis and synaptogenesis. Central to this hypothesis are the biosynthetic pathways involved in
energy metabolism, mitochondrial function, and synaptic connectivity. In the next section, we explore some
of the data that supports the link between metabolism and neuroplasticity, which may provide valuable
insight to better develop testable experiments to explore such connections.

BRAIN ENERGY METABOLISM IN PARKINSON’S DISEASE

The brain is the most metabolically demanding organ in the human body. Although the brain only
represents approximately 2% of total body mass, it is estimated to consume 20% and 25% of the body’s
oxygen and glucose supply, respectively (Attwell and Laughlin, 2001). Unlike most peripheral tissues, the
26

brain has limited metabolic flexibility such that is relies almost exclusively on glucose (and under specific
conditions, lactate) as its substrate for energy metabolism. Furthermore, exiguous astrocytic glycogen
stores only have the capacity to sustain normal brain function for a few minutes (Obel et al., 2012). This
highlights the brains dependency on a continuously available exogenous fuel supply. The glucose
transporter GLUT1, expressed in microvascular endothelial cells, mediates glucose transport across the
blood-brain barrier (BBB) (Simpson et al., 2007). In the brain, glucose is metabolized through four principle
biochemical pathways: (i) glycolysis, (ii) oxidative phosphorylation (OXPHOS), (iii) pentose phosphate
pathway (PPP), and (iv) glycogenesis. Importantly, each unique cell type within the brain has its own
distinctive metabolic profile, which is characterized by differential expression of these canonical pathways.
For example, neurons predominantly shuttle glucose or lactate through the OXPHOS pathway and utilize
oxygen as the final electron acceptor in the mitochondrial electron transport chain (ETC). Neurons demand
a high level of energy production. A major demand is due to the establishment and maintenance of the
membrane resting potential, where the sodium-potassium ATPase pump acts as a critical consumer of ATP
stores. In addition, it is now recognized that neuronal computation is also extremely energetically
demanding with most energy consumption occurring specifically at the synapse (Harris et al., 2012). This
demand is reflected by the observation that brain mitochondria are preferentially localized to the pre- and
post-synaptic terminals of the synapse in high numbers implicating the close relationship between energy
production and its demand (Wong-Riley, 1989). In response to neuronal activation, mitochondria provide
more than 90% of cellular energy in the form of ATP and are thus critical organelles for normal neuronal
function and information processing. Energy is used to drive vesicular packaging of neurotransmitters,
vesicle cycling and docking, and release of neurotransmitters such as glutamate from the presynaptic
terminal. At the postsynaptic terminal, ATP is necessary for receptor trafficking and signal transduction, and
when necessary, initiating a new action potential. It is also noteworthy that mitochondria not only play a role
in ATP production, but also play a critical role in other molecular functions. These include the buffering of
cellular calcium influx, as demonstrated by susceptibility to calcium-mediated excitotoxic injury through
NMDA receptor channels in many degenerative disorders when mitochondria dysfunction is evident. Also,
mitochondria generate low levels of reactive oxygen species (ROS) necessary for normal metabolic
homeostasis and when uncontrolled can activate cell death programs.
27

The importance of mitochondria at the synapse implies that their dysfunction can lead to synaptic
regulation and loss of brain connectivity. Historically, the study of neurodegenerative disorders has
correlated alterations of mitochondrial function with synaptic changes. With respect to PD, for example, the
discovery of MPTP-induced parkinsonism in the early 1980’s shed light on mitochondria as a contributor to
disease process; though inhibition of complex I of the mitochondrial ETC and accumulation of ROS, MPTP
metabolites produce elevated oxidative stress and reduce ATP production leading to cellular injury and
death (Langston et al., 1983; Langston, 2017). Since then, the involvement and mechanisms by which
mitochondria underlie neuronal dysfunction in PD pathogenesis has been shown to be much more complex.
It includes defects in mitochondrial functions including (i) trafficking, (ii) biogenesis through fusion and
fission, (iii) regulation of organelle breakdown through mitophagy, (iv) activation of apoptosis, (v)
dysregulation of mechanisms controlling oxidative stress, and (vi) calcium buffering capacity (Mounsey and
Teismann, 2010).

In addition to the valuable insights generated by toxins such as MPTP, 6-OHDA, 3-nitropropionic
acid (3-NP) and pesticides like rotenone on the relationship between mitochondria and neuronal function
in degenerative brain disorders, an important source of insight has emerged from the identification of
mechanisms by which familial forms of PD lead to disease. These familial forms of PD include, but are not
limited to, genetic mutations in a-synuclein (SNCA), DJ1, PINK1, PRKN, LRRK2, and UCHL1. Recent
research has led to the identification of the role of several of these genes as regulating mitochondrial
function and synaptic health, especially those localized to the nigrostriatal and corticostriatal pathways. The
PINK1/Parkin pathway, for instance, has been deemed the “quality control” pathway of the mitochondria;
for its role in Drp1-mediated mitochondrial fission, with mutations in this genetic pathway implicated in
dopaminergic degeneration characteristic of PD (McWilliams and Muqit, 2017). Loss of DJ-1 has been
demonstrated to cause accumulation of dysfunctional mitochondria at the synapse by disrupting lysosomal
homeostasis (Krebiehl et al., 2010). LRRK2 mutations have been implicated as an underlying cause of
mtDNA damage in neurons, the downstream effects impacting calcium imbalance, oxidative stress, and
dendritic shortening. a-synuclein has been observed to impair normal mitochondrial movement and
mitochondrial fission (biogenesis), contributing to the pathophysiology of PD (Xie and Chung, 2012).
28

Supportive of the potential role of many proteins involved in the mitochondrial pathophysiology of
degenerative brain disorders it has been suggested that in AD and HD, evidence supports that b-amyloid
peptide 1-42, the pathological forms of hyperphosphorylated tau protein, and mutant huntingtin protein
(mHTT) can impair mitochondria function leading the synaptic dysfunction and cell death (Quintanilla et al.,
2014; Quintanilla et al., 2017).

EFFECTS OF EXERCISE ON BRAIN ENERGY METABOLISM

Exercise is inherently associated with metabolism. As mentioned, energy demands rise with
increased firing of action potentials and increased synaptic neurotransmission. These metabolic challenges
are addressed by enhanced delivery of glucose and lactate to meet the demands for greater ATP
production. In addition, neuroprotective strategies are also activated through metabolic responses that help
protect neurons from acute bouts of oxidative stress; mitochondria can enter a state of senescence while
alternative sources of energy production are activated – including aerobic glycolysis for the production of
ATP (see below). In fact, the activation of aerobic glycolysis can match the ATP production of oxidative
phosphorylation for short periods of time (Barros, 2013).

In addition to increased energy requirements, the brain also engages in anabolic processes by
which structural proteins and other macromolecules necessary for neuroplasticity including the synthesis
of nascent receptors, axonal and dendritic scaffolding proteins, and other morphological changes. The
metabolic pathways, either directly or indirectly, contribute in shuttling small molecular substrates such as
glucose, lactate, and other molecules like lipids and amino acids in supplying the needed structural
components for neuroplasticity. Evidence is now emerging that exercise, specifically forms of exercise that
target corticostriatal circuits, can promote metabolic changes that can harness neuroplasticity.

Effects of exercise on glucose transport and metabolism

Exercise activates specific circuits in the brain, especially those involved in motor control linking
the cerebral cortex, basal ganglia, thalamus, and cerebellum. Enhanced neuronal activity is associated with
increased metabolic demand and elevated levels of local cerebral glucose utilization. Consistent with this
physiology, several groups have reported increased expression of the glucose transporters GLUT1
29

(endothelial- and astrocyte-specific) and GLUT3 (neuron-specific) in motor control regions following
treadmill exercise (Kinni et al., 2011; Allen and Messier, 2013; Takimoto and Hamada, 2014). Although
there have been no reported changes in GLUT1 expression in the PD brain, it has been shown that
mitochondrial dysfunction and ROS accumulation can impair glucose uptake and an inadequate glucose
can supply can have deleterious effects on dendritic morphology and impair synaptic plasticity, which
underlie aspects of PD pathology (Gandhi et al., 2009). Therefore, by driving activation of circuits affected
in PD, exercise imposes a metabolic challenge, which is met by an adaptive upregulation of glucose
transporters to improve glucose availability and to support synaptic function. In addition to changes in
GLUT1 expression, there is also an activation of enzymes involved ATP production including
dehydrogenases for lactate and glucose, and others within the TCA cycle.

Effects of exercise on lactate transport and metabolism

In recent decades, lactate has emerged from largely being considered as a metabolic waste
product to an important supplemental brain energy substrate and signaling molecule. At rest, blood-derived
lactate (~1mM) provides 8-10% of total brain energy requirements (Smith et al., 2003; Boumezbeur et al.,
2010). During moderate-to-vigorous aerobic exercise, however, blood lactate levels can increase by up to
10-fold and can supplement brain energy requirements to 25% of overall demand (van Hall et al., 2009).
Complementing elevated blood lactate levels following bouts of aerobic exercise, specific regional
upregulation of lactate transporters have been found in the cortex, hippocampus, and hypothalamus,
indicating a preferential utilization of peripheral lactate during supraphysiologic conditions (Takimoto and
Hamada, 2014). Furthermore, it has been suggested that the immediate energy requirements of sustained
synaptic activity are at least, in part, fueled by the complex oxidation of astrocyte-derived lactate (Hall et
al., 2012). The potential role of astrocytes in linking metabolism and neuroplasticity is highlighted in the
next sections.

Astrocytes, exercise, and neuroplasticity

Astrocytes fulfill numerous crucially important tasks, including mediating brain vasculature,
providing energetic substrates (such as lactate) for neighboring neurons, and recycling synaptic glutamate
30

following neuronal activity. The latter activity, the glutamate-glutamine cycle, is a cornerstone of astrocyte-
neuron coupling and serves as a key mediator of neuronal activation. With astrocytic processes enveloping
neuronal synapses, the astrocytic excitatory amino acid transporter 1 and 2 (EAAT1 and 2) serve as
sensors of neuronal activation (and subsequent glutamate release) and facilitate uptake of glutamate from
the synaptic cleft (Rothstein et al., 1994). This glutamate is then converted into glutamine, which is
subsequently recycled back to neurons for conversion into glutamate and repackaged into vesicles for
future synaptic events. The uptake of glutamate is an ATP-dependent process, due to increased sodium-
potassium-ATPase activity to meet ionic displacement during glutamate uptake (Khatri and Man, 2013),
thus inducing shifts in astrocytic metabolism to meet the neuroenergetics demands of glutamate-glutamine
cycling.

This example of astrocyte-neuron coupling forms the basis for the astrocyte-neuron lactate shuttle
(ANLS) hypothesis which postulates that (i) excitatory neurotransmission at the glutamatergic synapse
(glutamate/glutamine cycle) drives glucose uptake from the circulation via GLUT1 expressed in
microvascular endothelial cells and astrocytes, (ii) glucose is preferentially metabolized via aerobic
glycolysis in astrocytes to yield lactate, and (iii) lactate is shuttled from astrocytes to neurons via
monocarboxylate transporters (MCTs) where it is converted back to pyruvate and enters the TCA cycle in
mitochondria (Pellerin and Magistretti, 1994). This metabolic shift from glycolysis feeding the OXPHOS
pathways, to glycolysis predominantly producing lactate even in the context of available oxygen is termed
the Warburg effect; this metabolic mechanisms is a hallmark of tumor growth (Magistretti and Allaman,
2015). Astrocyte-specific metabolic changes during high levels of neuronal activity serve not only as an
opportunity to provide energy to excitatory neurons through the ANLS, but also a necessary metabolic pivot
in order to continue cycling glutamate; the production of lactate, normally an aerobic process, allows
astrocytes to oxidize NADH to NAD
+
, allowing for further ATP production through glycolysis demanded by
increased local neuronal activity.

Important for the facilitation of exercise-induced neuroplasticity, the ANLS represents a model,
which couples neuronal activity and the rapid delivery of energy substrates to the synapse. Preliminary data
from our lab suggests that exercise may target synaptic metabolism by upregulating the expression of ANLS
31

components, such as lactate dehydrogenase and MCT2 (neuron-specific lactate transporter), in the
striatum, following a single bout of treadmill exercise (unpublished data, Halliday et al.) Interestingly, cell-
specific overexpression of GLUT1 (in astrocytes) and MCT2 (in neurons) has been shown to be protective
against glutamate excitotoxicity. Taken together, these findings suggest that exercise may improve
astrocyte-neuron cooperation to facilitate lactate delivery required by active synapses in order to
supplement the increased metabolic demand, and thus support synaptic function and plasticity. In fact,
disruption of astrocytic lactate impairs long-term memory formation and is rescued by administration of
exogenous lactate but not glucose (Suzuki et al., 2011). Intra-hippocampal infusions of lactate have been
shown to enhance working memory and blocking neuronal uptake of lactate impairs memory formation in
rats (Newman et al., 2011). Further, neuronal uptake of lactate induces activation of immediate early genes
(IEGs), including Zif-268, c-Fos, and Arc, implicating lactate in neuroplasticity-related processes (Yang et
al., 2014). These findings, and other, demonstrate the critical role of astrocyte-derived lactate serving as
an important energy substrate in synaptic plasticity.

The close relationship between astrocytes and neurons and their health is paramount to adaptive
neuroplasticity and synaptogenesis. In a number of neurodegenerative disorders, astrocytic activation has
been well documents in terms of astrocyte morphology and elevated glia fibrillary acidic protein (GFAP)
expression, a marker of astrocyte inflammation. Indeed, neuronal expression of lactate dehydrogenase
subunit 1 (LDH-1) demonstrates a cell-specific ability to oxidize lactate (Bittar et al., 1996), one that may
prove to be beneficial in supplying high energetic demands during oxidative stress when glucose is
selectively shunted towards the PPP to maintain neuronal antioxidant status. Additionally, overexpression
of lactate dehydrogenase in vitro provided protection against b-amyloid peptides by shifting neuronal
energetics from OXPHOS to aerobic glycolysis, ensuring cellular resistance (Newington et al., 2012).
Dysfunctional astrocytes can have extensive molecular consequences including lactate production to
supply neurons, dysfunction of the BBB at astrocyte end feet, and diminished support through factors such
as BDNF. Thus, astrocyte integrity may serve as an early indicator of brain disease and may emerge as an
important therapeutic target.

32

Effects of exercise on brain mitochondrial function

It has been well established that aerobic exercise training induces an increase in density and
efficiency of muscle mitochondria, and these effects are associated with improved respiratory chain function
and increased ATP production, mitochondrial biogenesis, and enhanced resistance to elevated oxidative
stress. Surprisingly, much less is known about the effects of exercise on mitochondria in the brain, either
in neurons or glia. As mentioned above, neuronal mitochondria are preferentially localized to the synapse
and are responsible for suppling the energy demand of synaptic function in active neuronal circuits. In
normal aging and in neurodegenerative disorders such as PD, there are a number of alterations that can
occur in mitochondria resulting in their dysfunction, including (i) decreased activity of the ETC and other
enzymatic proteins, including those involved in controlling oxidative stress and calcium buffering, (ii)
accumulation of mtDNA mutations affecting mitochondrial proteins and RNA, (iii) altered biogenesis
including fission and fusion resulting in mitochondrial fragmentation or over-sized syncytia, and (iv) loss of
regulation of mitophagy (Schapira et al., 1990). Exercise represent a potentially important non-
pharmacological approach to target mitochondrial function at the synapse with the primary goal being to
facilitate synaptic integrity and to promote synaptic plasticity.

One means by which exercise can modulate brain mitochondria is through improved respiratory
chain function. For example, regular physical activity has been reported to increase enzymatic activity of
ETC complex I, III, and IV (Navarro et al., 2004). Furthermore, improved motor behavior following exercise
in the MPTP mouse model of PD is associated with enhanced mitochondrial function as measured by
respiratory efficiency and ATP production (Lau et al., 2011). Exercise-induced increases in ATP production
may be the result of altered expression of TCA cycle enzymes – some of which are increased in expression
(Kirchner et al., 2008). In addition to affecting mitochondrial ATP production efficiency, exercise can also
regulate mitochondrial density at the synapse. Chronic physical exercise training in the form of treadmill
running has been shown to upregulate the expression of two key proteins important for a wide spectrum of
mitochondrial functions and include the genes for silent information regulator 1 (SIRT1) and peroxisome
proliferator-activated receptor-g coactivator 1-a (PGC-1a). These proteins regulate mtDNA copy number in
several brain regions including the PFC, striatum, and hippocampus as well as cascades of enzymes
33

involved in energy metabolism, biogenesis, regulation of oxidative stress, mitophagy, and apoptosis
(Gerhart-Hines et al., 2007; Steiner et al., 2011). It is of interest to note that SIRT1 and PGC-1a are the
focus of a number of pharmacological agents to enhance mitochondrial health and one can speculate that
efficacy of any drug intervention may be enhanced with the synergistic application of exercise.

A potential role for hypoxia-inducible factor-1

When local metabolic demand through neuronal activity exceeds that of available energy
resources, highly conserved gene programs are activated which coordinate an integrated physiological
response to increase oxygen delivery, optimize cellular metabolism, increase metabolic capacity, and
establish homeostasis. The wide spectrum of molecular targets necessary for successful response indicate
the importance and complexity of changes that are necessary to survive challenges like metabolic demand.
Of the factors that have been shown to relay metabolic signals to adaptive changes in homeostatic gene
expression, hypoxia-inducible factor-1a (HIF-1a), a member of the HIF family of proteins, is one potential
candidate. HIF-1a is a cytoplasmic multimeric protein that is constitutively expressed in all cells. Its function
is regulated through targeted degradation by the proteasome where under normoxic conditions HIF-1a is
hydroxylated by the enzyme prolyl hydroxylase (PHD) and subsequently ubiquitinated by an E3 ligase for
targeted degradation. Under conditions of low oxygen, such as the transient hypoxia state during intensive
exercise, PHD is not able to promote the degradation of HIF-1a resulting in its increase stability,
translocation to the nucleus, and in complex with the protein HIF-1b, forms as an active transcription factor
that binds to the hypoxia response element (HRE) found in the promoter region of a large number of genes
including many responsible for glucose and lactate transport, energy metabolism, angiogenesis,
mitochondria function, and synaptogenesis. It has been suggested that one of the potential benefits of
exercise in the MPTP-lesioned mouse model is through the activation of HIF-1a (Smeyne et al., 2015).
Thus, the role of HIF-1a in neural tissues is not restricted to the regulation of energy metabolism through
oxygen sensing but has a potentially wider role in regulating complex molecular events including
neurogenesis, angiogenesis, and mitochondrial function (biogenesis), which are important for exercise-
induced neuroplasticity and synaptogenesis (Shweiki et al., 1992; Zhu et al., 2005; Correia et al., 2011;
Petzinger et al., 2015). Therefore, one potential mechanism may be that exercise, through increased
34

cellular metabolic energy demand, activates the HIF-1a transcriptional program, which in turn facilitates
and supports synaptic strength and connectivity during motor learning in the PD brain.

It is interesting to note that transcription factors such as HIF-1a, which is tightly regulated by oxygen
demand, are also regulated by other factors that can “sense” metabolic demand in neuronal circuits. These
other factors include reactive oxygen species (ROS) including superoxide, the glutamate-glutamine cycle,
nitric oxide, and potassium levels all highlighting the potentially complex and important role played by such
central regulators.

CONCLUSION

It is well established that exercise is beneficial in both treating and preventing many disorders in
the diseased and aging brain. It is necessary to identify the underlying molecular mechanism such that
novel treatments can target specific disorders. In terms of PD, evidence indicates that circuit-specific
pathways linking cognitive and motor circuits are critical in enhancing neuroplasticity to generate benefits
in complex behavior like movement. The purpose of this review was to highlight the potential link between
energy metabolism and neuroplasticity. Circuit-specific activation of neuronal pathways leads to changes
in glucose and lactate metabolism, altered mitochondrial function, and activation of cascades including
those through factors like HIF-1a that can induce genes and proteins necessary for angiogenesis,
synaptogenesis, and other components of neuroplasticity. While this may not represent a cure for disorders
like PD, it does implicate an important therapeutic target to complement pharmacological strategies and if
implemented early may in fact modify disease progression.

35

Figure 5. Changes in metabolism at corticostriatal pathway leads to improved circuit specific changes.
36

CHAPTER 4: Exercise-induced activation of a hypoxia-inducible factor 1-controlled gene program:
implications for experience-dependent neuroplasticity

Authors: Matthew R. Halliday, Giselle M. Petzinger, and Michael W. Jakowec

Physical activity is emerging as an effective and economical approach to treat and prevent a wide
range of neurological disorders. Studies by our research group and others have shown that exercise, in the
form of treadmill running, mediates a type of experience-dependent neuroplasticity through the modulation
of dopaminergic and glutamatergic neurotransmission, synaptogenesis, and regulation of cerebral blood
flow (Fisher et al., 2004; Petzinger et al., 2007; Kintz et al., 2013; Toy et al., 2014). However, a clear
understanding of the underlying molecular mechanisms driving exercise-induced physiological adaptation
in the brain represents a major gap in our knowledge.

Recently, hypoxia-inducible factor 1 (HIF-1), a member of the HIF family of proteins and master
transcription factor, has attracted interest as a potentially critical regulatory element and facilitator of
exercise-induced neuroplasticity due to its ability to act as a metabolic “sensor” and relay changes in
metabolic state to changes in homeostatic gene expression. Importantly, HIF-1a, the oxygen-sensitive
subunit of the heterodimeric HIF-1 transcription factor has been shown to be upregulated in the brain
following an intensive treadmill exercise training program (Kinni et al., 2011). Important to neuroplasticity,
HIF-1 regulates the expression of a constellation of downstream target genes involved with cellular and
molecular processes such as energy substrate transport (glucose and lactate), metabolism, mitochondrial
function, redox balance, and angiogenesis. Thus, one potential mechanism may be that exercise, through
increased cellular metabolic demand, activates the HIF-1 gene program to increase fuel delivery, optimize
cellular metabolism, and establish homeostasis, which in turn facilitates and supports synaptic strength and
connectivity during motor learning.

For the purpose of this study we investigated the effects of a treadmill exercise program on the
activation of a HIF-1-controlled gene program in the striatum of healthy young mice. We focused our
analysis on the striatum based on previous studies which have shown this to be a highly targeted brain
region in exercise-induced neuroplasticity (Fisher et al., 2004; Petzinger et al., 2007; Kintz et al., 2013;
37

Petzinger et al., 2013; Toy et al., 2014; Petzinger et al., 2015). Using a quantitative polymerase chain
reaction (qPCR) array we show a significant change in expression of over 60 HIF-1 target genes. We then
extend these findings using validated primer sets for a select set of genes included on the array that have
been established as being relevant to experience-dependent neuroplasticity. Taken together, these findings
suggest a potential role for HIF-1 as a regulatory element and facilitator of exercise-induced brain changes.
Further elucidation of molecular mechanisms involving metabolic pathways will allow us to better apply
exercise as a therapy for brain disorders and will aid in the search for novel therapeutic targets with the
potential to enhance the benefits of exercise on brain health.
MATERIALS AND METHODS
Animals

For all experiments conducted as part of this study, a total of 24 young adult (8-10 weeks old) male
C57BL/6J mice (Jackson Laboratory, Bar Harbor, Maine, USA) were used. For the initial qPCR array
screening, animals were randomly divided into 5 groups (n=3 / group): (i) no exercise (NE), (ii) 1-day
exercise (1DE), (iii) 3-day exercise (3DE), (iv) 5-day exercise (5DE), and (v) 10-day exercise (10DE). For
subsequent qPCR validations, only NE, 1DE, and 5DE groups were analyzed (n=3 / group). Mice were
housed five to a cage and acclimated to a 12-hour light/dark cycle. All exercise training occurred during the
animal’s normal wake period. All experiments were conducted in accordance with the NIH Guide for the
Care and Use of Laboratory Animals (NIH Publication No. 80-23, revised 1996 and were approved by the
University of Southern California Institutional Animal Care and Use Committee.

Treadmill exercise and tissue collection

The exercise protocol used in the present study was adapted from a protocol previously described
(Fisher et al., 2004). Briefly, prior to group randomization, all mice were challenged to maintain a forward
position on the 45 cm treadmill belt (Columbus Instruments, Columbus, Ohio, USA) for 5 min at 5.0 m/min
to ensure exercise competency. A metal bead curtain was used as a tactile incentive to encourage mice to
maintain a forward position on the treadmill throughout the duration of the exercise session. Mice were
exercised one hour per day, five days per week. To control for any non-exercise-related effects of treadmill
running (handling, novel environment, noise, and vibration), sedentary animals (NE) were exposed to the
38

treadmill apparatus for a time period equivalent to that of the exercised groups. Exercise mice started at a
velocity of 10.0 ± 2.5 m/min and was incrementally increased to a maximum velocity of 18.5 ± 1.5 m/min
by the second week. All exercise samples were collected immediately after the animals’ last training
session. Bilateral striatal dissections were performed on ice and striata were stored in RNAlater RNA
Stabilization Reagent (Qiagen, Hilden, Germany) at 4 °C for later use.

Quantitative polymerase chain reaction (qPCR)

RNA was isolated and purified using a RNeasy Mini Kit (Qiagen, Hilden, Germany) strictly adhering
to the manufacturer’s suggested protocol. RNA concentration and purity were analyzed at a 1:10 dilution
using a BioPhotometer (Eppendorf, Hamburg, Germany). The ratio between the absorbance at 260 nm and
280 nm was used to evaluate purity; we assumed ratios between 1.8 and 2.0 to be pure. cDNA synthesis
was performed using a QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany) and the reaction
product was stored at -20 °C for downstream qPCR applications.

Gene expression data shown in Table 1 were generated using the RT
2
Profiler PCR Array (Mouse
Hypoxia Signaling; Qiagen, Hilden, Germany) strictly following the manufacturer’s suggested protocol. 1 µg
of purified striatal RNA was placed on the 96-well array plate in total. Samples were processed on an
Eppendorf realplex
2
Mastercycler and analyzed using realplex software (Eppendorf, Hamburg, Germany).
Following our initial qPCR array screening analysis, gene expression of Hif1a (5’-
ACCTTCATCGGAAACTCCAAAG, 3’- CTGTTAGGCTGGGAAAAGTTAGG); Car9 (5’-
TGCTCCAAGTGTCTGCTCAG, 3’- CAGGTGCATCCTCTTCACTGG); Ldha (5’-
TGTCTCCAGCAAAGACTACTGT, 3’- GACTGTACTTGACAATGTTGGGA); Slc2a1 (5’-
CAGTTCGGCTATAACACTGGTG, 3’-GCCCCCGACAGAGAAGATG); and Vegfa (5’-
GCACATAGAGAGAATGAGCTTCC, 3’- CTCCGCTCTGAACAAGGCT) were analyzed with the SYBR
Green PCR Kit (Qiagen, Hilden, Germany). All qPCR data was analyzed using the 2
-DDCt
method.

Statistical Analysis

All data are reported as mean ± s.e.m. The F-test was conducted to verify that the samples were
normally distributed and express homogeneity of variance. The variances of the respective samples
39

compared between the groups were statistically similar. For multiple comparisons, a one-way analysis of
variance followed by Tukey’s post hoc test was used. A p value <0.05 was considered statistically
significant. Simple linear regression and Pearson coefficient were used to determine correlations between
variables. All data were analyzed using GraphPad Prism 6.0 software (GraphPad Software, La Jolla, CA).

Figure 1. Transcript analysis using the Qiagen Hypoxia Signaling RT
2
Profiler Array. Shown are the
functional domains of the 84 transcripts screened during the initial microarray study. These functional
domains include genes controlling angiogenesis, transcription, cell growth, DNA repair, apoptosis,
molecular transport, metabolism, cell signaling, and circadian rhythm. This Figure demonstrates the range
of transcripts explored in this study. This microarray is derived from transcripts whose expression is
elevated under conditions of reduced oxygen tension and is not designed to be tissue-specific.

40

RESULTS

In total, of the 80 oxygen-sensitive genes included on the RT
2
Profiler Array, 64 unique genes were
either up- or down-regulated by at least 20% after 1, 3, 5, or 10 days of treadmill exercise compared to
sedentary controls (Table 1). For reference, a comprehensive overview of genes included on the RT
2

Profiler PCR Array, as well as their respective functional domains are shown in Figure 1. To validate gene
expression data collected from the array screening we selected five genes (Hif1a, Car9, Ldha, Slc2a1, and
Vegfa) and analyzed their expression levels in the striatum in sedentary controls and after 1 and 5 days of
exercise in a separate group of animals. Taken together, findings presented in this chapter indicate that
intensive treadmill exercise has the potential to transcriptionally activate complex cellular and molecular
systems that control energy substrate transport, metabolism, redox homeostasis, and angiogenesis, all of
which are thought to be important for experience-dependent neuroplasticity.

Table 1. Treadmill exercise activates a wide range of HIF 1-related oxygen-sensitive genes.
Fold-change regulation
Gene symbol Protein NE 1DE 3DE 5DE 10DE
Adm Adrenomedullin 1.00 1.64 3.78 1.42 1.69
Adora2b Adenosine A2b receptor 1.00 0.99 1.06 0.63 0.63
Aldoa Fructose-bisphosphate aldolase A 1.00 0.85 0.72 0.70 0.59
Angpt4l Angiopoietin-related protein 4 1.00 3.01 2.45 2.35 1.18
Ankrd37 Ankyrin repeat domain 37 1.00 0.88 1.06 0.75 0.70
Anxa2 Annexin 2 1.00 1.57 1.29 1.54 1.07
Apex1 DNA-(apurinic or apyrimidinic site) lyase 1.00 1.29 1.36 1.03 0.85
Arnt Aryl hydrocarbon receptor nuclear translocator 1.00 1.08 1.01 0.82 0.80
Atr Serine/threonine-protein kinase ATR 1.00 2.30 0.59 0.44 0.92
Bhlhe40 Basic helix-loop-helix family member e40 1.00 1.36 1.39 0.65 0.69
Blm Bloom syndrome protein 1.00 1.02 0.95 1.04 0.80
Bnip3 BLC2/adenovirus E1B protein-interacting protein 3 1.00 1.14 0.97 0.75 1.00
Bnip3l
BLC2/adenovirus E1B protein-interacting protein 3-
like
1.00 1.05 0.76 0.81 0.75
Btg1 BTG anti-proliferation factor 1 1.00 1.60 1.23 1.29 1.04
Car9* Carbonic anhydrase 9 1.00 1.21 1.34 1.56 1.14
Ccng2 Cyclin-G2 1.00 0.98 1.00 0.88 0.76
Cops5 COP9 signalosome complex subunit 5 1.00 0.97 1.29 1.18 0.90
Ctsa Cathepsin A 1.00 1.14 1.21 1.01 1.09
Ddit4 DNA damage-inducible transcript 4 1.00 2.19 2.36 1.89 1.40
Dnajc5 DnaJ homolog subfamily C member 5 1.00 1.25 0.90 0.73 0.81
Edn1 Endothelin 1 1.00 0.73 0.42 0.69 0.49
Egln1 Egl nine homolog 1 1.00 1.09 0.96 1.04 0.80
Egln2 Egl nine homolog 2 1.00 1.20 1.43 1.12 0.85
Egr1 Early growth response protein 1 1.00 1.01 0.88 0.82 0.82
41

Eif4ebp1
Eukaryotic translation initiation factor 4E-binding
protein 1
1.00 1.12 1.69 1.17 1.04
Eno1 Enolase 1 1.00 0.94 1.19 0.94 1.04
Ero1 ERO1-like protein alpha 1.00 1.39 1.13 1.19 0.90
Fos Proto-oncogene c-Fos 1.00 3.76 1.79 1.78 1.42
Gbe1 1,4-alpha-glucan-branching enzyme 1.00 0.61 0.59 0.62 0.45
Gpi1 Glucose phosphate isomerase 1 1.00 1.18 0.95 1.01 0.91
Gys1 Glycogen synthase 1 1.00 1.14 1.32 1.39 1.29
Hif1a* Hypoxia-inducible factor 1-alpha 1.00 1.38 1.41 1.74 0.82
Hif1an Hypoxia-inducible factor 1-alpha inhibitor 1.00 0.92 0.94 0.80 0.94
Hif3a Hypoxia-inducible factor 3-alpha 1.00 1.11 1.01 1.07 1.10
Hk2 Hexokinase 2 1.00 1.29 1.06 0.86 0.98
Hmox1 Heme oxygenase 1 1.00 1.06 1.11 0.88 0.90
Ier3 Radiation-inducible immediate-early gene IEX1 1.00 1.06 0.91 0.90 1.00
Igfbp3 Insulin-like growth factor-binding protein 3 1.00 1.38 2.00 1.61 1.17
Jmjd6
Bifunctional arginine demethylase and lysyl-
hydroxylase JMJD6
1.00 1.27 1.25 1.34 1.01
Ldha* L-lactate dehydrogenase A chain 1.00 1.17 1.18 1.33 1.01
Lgals3 Galectin-3 1.00 2.33 3.18 3.32 1.45
Lox Protein-lysine 6-oxidase 1.00 2.23 1.77 2.60 1.07
Map3k1 Mitogen-activated protein kinase 1 1.00 1.13 0.66 0.76 0.54
Met Hepatocyte growth factor receptor 1.00 1.88 1.16 0.95 0.91
Mif Macrophage migration inhibitory factor 1.00 1.08 1.17 1.18 0.95
Mmp9 Matrix metalloproteinase-9 1.00 1.75 1.16 1.93 1.00
Mxi1 Max-interacting protein 1 1.00 1.07 0.99 0.81 0.85
Nampt Nicotinamide phosphoribosyltransferase 1.00 1.44 1.01 1.15 0.82
Ncoa1 Nuclear receptor coactivator 1 1.00 0.86 0.94 1.02 0.90
Ndrg1 N-Myc downstream-regulated protein 1 1.00 1.29 1.03 1.18 0.95
Nfkb1 Nuclear factor NF-kappa-B p105 subunit 1.00 1.09 0.90 1.05 0.78
Nos3 Nitric oxide synthase, endothelial 1.00 1.29 1.47 0.89 1.40
Odc1 Ornithine decarboxylase 1.00 1.24 1.09 0.97 0.92
P4ha1 Prolyl 4-hydroxylase subunit 4-alpha 1.00 1.16 1.05 1.10 0.93
P4hb Protein disulfide-isomerase 1.00 1.16 1.28 0.81 0.75
Pdk1 Pyruvate dehydrogenase kinase 1 1.00 0.93 1.01 0.78 0.54
Per1 Period circadian protein homolog 1 1.00 1.58 1.62 1.11 0.97
Pfkfb3
6-phosphofructo-2-kinase/fructose-2,6-
bisphosphate 3
1.00 1.60 1.14 0.97 1.06
Pfkfb4
6-phosphofructo-2-kinase/fructose-2,6-
bisphosphate 4
1.00 0.92 0.81 0.95 0.76
Pfkl Phosphofructokinase, liver type 1.00 0.92 0.95 0.95 0.90
Pfkp Phosphofructokinase, platelet type 1.00 0.98 0.90 0.88 0.82
Pgam1 Phosphoglycerate mutase 1 1.00 0.90 1.03 1.13 0.95
Pgf Placental growth factor 1.00 1.06 1.40 0.91 0.90
Pgk1 Phosphoglycerate kinase 1 1.00 1.31 1.21 1.14 0.95
Pim1 Serine/threonine-protein kinase pim-1 1.00 1.23 1.56 1.06 1.16
Pkm Pyruvate kinase 1.00 1.38 1.57 1.06 0.91
Plau Urokinase-type plasminogen activator 1.00 0.47 0.37 0.43 0.30
Rbpj
Recombination binding protein suppressor of
hairless
1.00 1.45 1.21 1.01 1.04
Ruvbl2 RuvB-like 2 1.00 0.99 0.82 0.89 0.95
Serpine1 Plasminogen activator inhibitor 1 1.00 1.66 1.43 1.38 0.86
42

Slc16a3 Monocarboxylate transporter 4 1.00 0.92 1.11 1.21 0.96
Slc2a1* Glucose transporter 1 1.00 1.16 1.47 1.69 1.05
Slc2a3 Glucose transporter 3 1.00 1.14 1.13 1.07 0.93
Tfrc Transferrin receptor 1.00 1.26 0.97 0.85 0.88
Tpi1 Triosephosphate isomerase 1.00 1.17 0.92 0.90 0.96
Trp53 Cellular tumor antigen p53 1.00 1.03 0.91 0.82 0.80
Txnip Thioredoxin-interacting protein 1.00 2.64 1.88 2.53 1.44
Usf2 Upstream stimulatory factor 2 1.00 1.30 1.09 0.88 0.99
Vdac1
Voltage-dependent anion-selective channel protein
1
1.00 1.26 1.77 1.45 0.93
Vegfa* Vascular endothelial growth factor A 1.00 1.23 1.30 1.53 1.06
Table 1. Microarray results of Mouse Hypoxia Signaling RT
2
Profiler Array. Transcripts from the Hypoxia
Signaling RT
2
Profiler Array are tabulated by alphabetical order of gene name. Light red indicates an
increase in expression from 20-49%, while dark red indicates an increase greater than or equal to 50%.
Conversely, light blue indicates a decrease in expression by 20-49%, while dark blue indicates a decrease
greater than or equal to 50%. No color indicates expression was not increased or decreased by 20%
compared to sedentary controls. Results from this microarray analysis were used to select a subset of
transcripts for validation in a separate set of experiments. These transcripts included Hif1a, Car9, Ldha,
Slc2a1, and Vegfa.

Treadmill exercise upregulates mRNA levels of Hif1a in mouse striatum

Previous studies have shown that HIF-1a brain expression (transcript and protein) levels are
elevated immediately following treadmill exercise (Kinni et al., 2011). However, a major gap in knowledge
is the degree to which different brain regions are affected by the observed exercise-dependent increase in
Hif1a expression following aerobic exercise. Data presented in this chapter help to fill this gap by
demonstrating that, in particular, exercise upregulates Hif1a expression in the striatum. It remains to be
determined if other brain regions are affected in the same manner. As shown in Table 1, 1 hour (1DE) of
intensive treadmill exercise increased Hif1a mRNA levels by 38% in the striatum of exercise naïve mice
compared to sedentary controls. After 3 (3DE) and 5 (5DE) days of exercise, striatal Hif1a mRNA levels
were increased by 41 and 74%, respectively (Table 1). By 10 days of exercise, Hif1a expression levels had
returned to near baseline values, suggesting the potential for a type of homeostatic adaptation had occurred
in trained animals. These data were subsequently confirmed by a separate qPCR validation and are shown
in Figure 2 (Fig. 2A).
43

HIF-1 target gene transcript levels are increased following treadmill exercise

HIF-1 controls the expression of a wide array of downstream target genes that are thought to be
important for neuroplasticity. Figure 1 illustrates that many of these genes control cellular processes such
as energy substrate transport, metabolism, redox homeostasis, cell proliferation and apoptosis, and
angiogenesis. As mentioned above, during the initial PCR array screening 64 transcripts were found to be
up- or down-regulated by at least 20% after 1, 3, 5, or 10 days of treadmill exercise suggesting a dynamic
regulation of HIF-1 targets during physical activity (Table 1). From this set of HIF-1 target genes we chose
to validate the change in expression levels of 3 unique genes thought to be important for experience-
dependent neuroplasticity: Ldha (glycolytic lactate production), Slc2a1 (transendothelial and astrocytic
glucose transport), and Vegfa (angiogenesis and neurogenesis). Results from the PCR array showed a
similar temporal pattern of change in gene expression observed with Hif1a; Ldha, Slc2a1, and Vegfa all
demonstrated an exercise dose-dependent increase in transcript expression after 1, 3, and 5 days of
treadmill exercise with a return to baseline expression levels after 10 days of exercise (Table 1). However,
while there was a trend for increased expression in the other transcripts at 1DE, they did not reach statistical
significance until 5DE (Car9, 1.56-fold increase; Ldha, 1.33-fold increase; Slc2a1, 1.69-fold increase, and
Vegfa, 1.53-fold increase). These data were subsequently confirmed by a separate qPCR validation and
are shown in Figure 2 (Fig. 2B, 2D, 2F, and 2H). Simple linear regression on Pearson’s correlation analysis
against fold change in Hif1a expression are shown in Figure 2 (2C, 2G, 2E, and 2I). All correlations were
found to be positive.
44

Figure 2. Exercise leads to the activation of Hif1a and several of its downstream targets. qPCR results in
Panel A shows increased expression of Hif1a after 1 (1DE) and 5 days (5DE) of exercise compared to no
exercise (NE) groups in the striatum. Other transcriptional targets examined by qPCR also show elevated
expression as well as a positive correlation compared to Hif1a including Car9 (Panels B and C), Ldha
(Panels D and E), Slc2a1 (Panels F and G), and Vegfa (Panels H and I). * denotes significance at p < 0.05.

DISCUSSION

This study showed that exercise, in the form of motorized treadmill exercise, activates a HIF-1-
controlled gene program in the striatum. Overall there was a high degree of consistency with respect to the
pattern of gene expression between the initial microarray screen and our subsequent qPCR analysis,
providing strong validation of the initial screen approach. While previous studies have shown an elevation
in global HIF-1a expression following treadmill exercise, a major gap in knowledge has been to what degree
particular brain regions are affected. While this study restricted its analysis to the striatum, Chapter 5 will
begin to explore the effect of treadmill exercise on other brain regions involved in exercise-induced
neuroplasticity, such as the prefrontal cortex.
HIF-1 controls the expression levels of a wide array of downstream target genes, many of which
we have reported data for here (Table 1). Our array data indicate a complex and dynamic regulation of the
HIF-1-controlled gene program in response to treadmill exercise, which involves change in expression
levels of gene involved in cellular processes such as modulation of cellular metabolism, vasomotor control
45

and regulation of cerebral blood flow, DNA repair, redox homeostasis, and cell cycle control, all of which
are thought to be important for exercise-induced neuroplasticity. For the purpose of this study, we restricted
our post hoc validation of the RT
2
Profiler Array data to 3 target genes: Ldha, Slc2a1, and Vegfa. We
selected these genes based on accumulating evidence that suggests modulation of brain metabolism and
regulation of cerebral blood flow as potential mechanisms mediating the beneficial effects of physical
exercise on brain health through neuroplasticity (Kinni et al., 2011; Wang et al., 2015a; Wang et al., 2015b)
The Ldha gene encodes for lactate dehydrogenase A, an enzyme that catalyzes the interconversion of
lactate and pyruvate with one potential role in converting astrocytic sources of lactate to pyruvate in neurons
for inclusion into the tricarboxylic (TCA) cycle (Laughton et al., 2007). Recently, lactate has gone from being
considered as a metabolic waste product to an important supplemental brain energy substrate and signaling
molecule. At rest, blood-derived lactate (~1 mM) provides approximately 10% of total brain energy
requirements (Smith et al., 2003; Boumezbeur et al., 2010). During moderate-to-vigorous aerobic exercise,
however, blood lactate levels can increase by up to 10-fold and can supplement up to 25% of total brain
energy requirements (van Hall et al., 2009). Furthermore, elevated blood lactate levels are complemented
by an upregulation of monocarboxylate lactate transporters (see Chapter 5) (Takimoto and Hamada, 2014).
In addition to exercise-induced increases in Ldha expression, treadmill exercise also upregulates the
expression of genes essential for central glucose delivery, such as Slc2a1. The Slc2a1 gene encodes for
glucose transporter type 1, a carbohydrate transporter that is distributed throughout most body tissue
including the endothelial cells that make up the blood-brain barrier and whose expression is elevated with
increased cellular glucose metabolism (Dienel, 2019). Therefore, exercise, through recruitment of relevant
circuitry, and an increased cellular metabolic demand, appears to lead to the activation of a gene program
that increases energy substrate availability in brain regions of high activity, and the optimization of cellular
metbolisms to support synaptic strength and connectivity. Lastly, the Vegfa gene encodes the signaling
protein vascular endothelial growth factor A (VEGFA) whose induction under conditions of hypoxia leads
to the promotion of angiogenesis and increased blood flow (Sharp and Bernaudin, 2004).
In addition to these 3 target genes we also chose to validate the expression levels of Car9. The
Car9 gene encodes for carbonic anhydrase 9, an enzyme that catalyzes the conversion of carbon dioxide
(CO2) and water (H2O) to carbonic acid (HCO3
-
) and a proton (H
+
). This enzyme is particularly sensitive to
46

hypoxia and its expression is dependent on HIF-1a (Olive et al., 2001). For the purpose of these studies it
served as a reliable marker to validate the potential link between exercise and hypoxia as an activator of
transcript expression. Overall these results indicate that several genes involved in the elevated metabolic
response with exercise are consistent with increased expression of HIF-1a suggesting a regulatory role for
HIF-1a in exercise-induced neuroplasticity.

47

CHAPTER 5: Physical activity optimizes circuit-specific cellular metabolism in neuroplasticity: a
role for hypoxia-inducible factor 1 and its downstream targets

Authors: Matthew R. Halliday, Dishan Abeydeera, Giselle M. Petzinger, Michael W. Jakowec

Exercise and physical activity are critical in maintaining healthy motor and cognitive circuitry. Not
only is exercise important for healthy aging, but studies from our research group have shown that exercise
can serve as a means to enhance neuroplasticity and restore motor and cognitive behavioral deficits in
both patients and in animal models of neurodegenerative disorders including Parkinson’s disease (PD) and
Huntington’s disease (HD) (Fisher et al., 2004; Fisher et al., 2013; Stefanko et al., 2017). A major gap in
knowledge is a clear understanding of the underlying molecular mechanisms by which exercise leads to
the enhancement of neuroplasticity and the strengthening of brain circuitry. For example, using animal
models of PD under conditions of high intensity exercise, we have documented changes in corticostriatal
plasticity based on a wide range of techniques including molecular, histological, morphological,
electrophysiological, and behavioral outcome measures (Fisher et al., 2004; Petzinger et al., 2007;
VanLeeuwen et al., 2010; Kintz et al., 2013; Toy et al., 2014).
Under conditions of high neuronal activity, neurons and glia undergo significant metabolic changes
with respect to utilization of glucose and lactate and the generation of adenosine-triphosphate (ATP)
(Dienel, 2019). It is hypothesized that prolonged metabolic demands in these neuronal circuits leads to
downstream activation and expression of genes and proteins involved in synaptogenesis and angiogenesis
and the consequential consolidation of these circuits and permanency of the underlying behaviors.
Metabolic demands impact the availability of oxygen and elevates the expression of factors in response to
transient hypoxia. One such responsive mediator is hypoxia-inducible factor 1a (HIF-1a), a transcription
factor induced by a number of molecular environments including hypoxia, high levels of neuronal activity,
and reactive oxygen species (ROS) (Sharp and Bernaudin, 2004). Activation of HIF-1a activates a gene
program that includes a wide range of transcripts including those for metabolism, angiogenesis,
synaptogenesis, structural proteins and others.
48

Studies from Holschneider and colleagues have investigated the changes that occur in regional
cerebral blood flow (rCBF) with exercise in both normal rodents and in models of neurodegenerative
disorders (Wang et al., 2013; Wang et al., 2016). Findings have shown that exercise leads to selective
activation of rCBF in rodents engaged in exercise on a motorized treadmill or motorized running wheel. For
example, significant activation is seen within a number of distinctive regions including the prefrontal cortex,
dorsal striatum, motor cortex, and vermis of the cerebellum. Other regions of the brain either show no
change or a reduction in rCBF. Using rCBF changes as a guide, we hypothesize that these same regions
with increased rCBF will also show increased neuroplasticity. This is supported by molecular studies where
the dorsal striatum and prefrontal cortex (PFC) show activation of specific subunits of glutamate receptors
of the AMPA subtype, as well as electrophysiological parameters of plasticity (spontaneous excitatory post-
synaptic current, sEPSCs) and morphological changes with the restoration of dendritic spine density (Kintz
et al., 2013; Toy et al., 2014). The association between changes in rCBF, and between molecular and
morphological supports the notion that exercise mediates circuit-specific changes in neuroplasticity.
The purpose of this study was the begin to identify the molecular mechanisms by which exercise, in the
form of running on a motorized treadmill, leads to the activation of genes involved in metabolic and synaptic
functions that mediate neuroplasticity. The work in this chapter seeks to build on the data presented in
Chapter 4 by validating the expression change in HIF-1a and in a number of HIF-1 target genes in a region-
specific fashion based on anatomical regions of the brain where we have previously documented changes
in cerebral blood flow and synaptogenesis. Our findings support the important role of metabolic changes
within the brain that occur during exercise that are consistent with circuit-specific activation of brain
connectivity.

MATERIALS AND METHODS
Animals
For all experiments conducted in this study, a total of 25 young adult (8-10 weeks) male C57BL/6J
mice (Jackson Laboratory, Bar Harbor, Maine, USA) were used. Animals were randomly divided into 5
groups: (i) no exercise (NE), (ii) 1-day exercise (1DE), (iii) 5-day exercise (5DE), (iv) 10-day exercise
(10DE), and 5-day exercise/5-day rest (5DE+5DR). Mice were house 5 to a cage and acclimated to a 12-
49

hour shift in light/dark cycle. All exercise training occurred during the animal’s normal wake period. All
experiments were conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals
(NIH Publication No. 80-23, revised 1996) and were approved by the University of Southern California
Institutional Animal Care and Use Committee (IACUC).
Exercise Regimen
The treadmill exercise protocol was conducted as previously described (Fisher et al., 2004). Briefly,
prior to group randomization, all mice were challenged to maintain a forward position on the 45 cm treadmill
belt (Columbus Instruments, Columbus, Ohio, USA) for 5 min at 5.0 m/min to ensure exercise competency.
A metal bead curtain was used as a tactile incentive to encourage mice to maintain a forward position on
the treadmill throughout the duration of the exercise session. Mice were exercised one hour per day, five
days per week. To control for any non-exercise effects of treadmill running (handling, novel environment,
noise, and vibration), non-exercised animals (NE) were exposed to the treadmill apparatus for a time period
equivalent to that of the exercised groups. Exercise mice started at an initial velocity of 10.0 ± 2.5 m/min
that was incrementally increased to a maximum velocity of 18.5 ± 1.5 m/min by the second week. All animals
were able to complete the exercise regimen successfully.
Tissue Preparation
Tissue was collected from all treatment groups at exercise days 1, 5, and 10. Mice were sacrificed
by cervical dislocation for fresh tissue. Striatal tissue was dissected in a block from a coronal slice: (i)
Bregma +1.20 to +0.60mm, (ii) 2.5mm lateral from midline, (iii) dorsal-ventral, inferior to the corpus callosum
and superior to the anterior commissure. All exercise samples were collected immediately after the animals’
last training session. Brain tissue was dissected on ice and stored in RNAlater RNA Stabilization Reagent
(Qiagen) for later use. For qRT-PCR analysis of specific brain regions, fresh tissue was rapidly
microdissected in blocks from i) PFC (Bregma +2.0 to +1.4 mm A.P., rostral to corpus callosum; ± 1 mm
M.L. from midline to the corpus callosum, and -1.5 to -3.0 mm D.V.); ii) STR (Bregma +1.2 to -0.2mm A.P.,
including tissue bordered ventrally by the anterior commissure, dorsally by the corpus callosum, medially
by the lateral ventricle, and ±2.5 mm laterally from the midline); and iii) ERC (Bregma -2.0 to -3.0 mm A.P.,
laterally bounded by the edge of the cortex and extending 1mm medially to the edge of the striatum, and -
3.0 to -3.5 mm D.V.) Tissues were placed in RNAlater Stabilization Solution (Qiagen, Hilden, Germany)
50

and stored at 4 °C overnight.
Quantitative Polymerase Chain Reaction (qPCR)
RNA was isolated and purified using a RNeasy Mini Kit (Qiagen, Hilden, Germany) strictly adhering
to the manufacturer’s suggested protocol. RNA concentration and purity were analyzed at a 1:10 dilution
using a BioPhotometer (Eppendorf, Hamburg, Germany). The ratio between the absorbance at 260 nm and
280 nm was used to evaluate purity; we assumed ratios between 1.8 and 2.0 to be pure. cDNA synthesis
was performed using a QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany) and the reaction
product was stored at -20 °C for downstream qPCR applications. Samples were processed on an Eppendorf
realplex
2
Mastercycler and analyzed using realplex software (Eppendorf, Hamburg, Germany). Gene
expression of Hif1a (5’- ACCTTCATCGGAAACTCCAAAG, 3’- CTGTTAGGCTGGGAAAAGTTAGG), Ldha
(5’- TGTCTCCAGCAAAGACTACTGT, 3’- GACTGTACTTGACAATGTTGGGA), Slc2a1 (5’-
CAGTTCGGCTATAACACTGGTG, 3’- GCCCCCGACAGAGAAGATG), Slc16a1 (5’-
TGTTAGTCGGAGCCTTCATTTC, 3’- CACTGGTCGTTGCACTGAATA), Slc16a7 (5’-
GCTGGGTCGTAGTCTGTGC, 3’- ATCCAAGCGATCTGACTGGAG), and Vegfa (5’-
GCACATAGAGAGAATGAGCTTCC, 3’- CTCCGCTCTGAACAAGGCT) were analyzed with the SYBR
Green PCR Kit (Qiagen, Hilden, Germany). All qPCR data was analyzed using the 2
-DDCt
method.
Statistical Methods
All data are reported as mean ± S.E.M. The F-test was conducted to verify that the samples were
normally distributed and have homogenous variances. The variances of the respective samples compared
between the groups were statistically similar. For multiple comparisons, a one-way analysis of variance
followed by Tukey’s post hoc test was used. A value of p < 0.05 was considered statistically significant. All
data were analyzed using GraphPad Prism 6.0 software (GraphPad Software, La Jolla, CA).
RESULTS
To further explore the temporal and regional patterns of expression of HIF-1a-dependent genes of
interest, we also analyzed their pattern of expression not only in the STR but also in tissues from the
Prefrontal Cortex (PFC) and Ectorhinal Cortex (ERC). These tissue selections were based on regions
where we have previously observed exercise-dependent elevation in regional cerebral blood flow (rCBF) in
the STR and PFC or no change in rCBF (ERC)(Yang et al., 2007). Tissues were collected from groups with
51

no exercise (NE), 1-day exercise (1DE), 5-days of exercise (5DE), and 10-days of exercise (10DE). In
addition, tissues were collected from a group of mice subjected to 5-days of exercise followed by 5 days of
rest (5DE+5DR) in order to examine ant retention of expression changes. These data are shown in Figure
3.
We observed a significant increase in striatal expression of Hif1a transcript in the STR and PFC at
5DE with elevated expression at 1DE in the PFC (Fig. 3AB). Expression returned to baseline by 10DE as
well as in the 5DE+5DR groups. The ERC did not show a significant change in Hif1a expression (Fig. 3C).
Ldha (lactate dehydrogenase A gene) expression was similar to previously observed with increased
expression at 5DE in both the STR and PFC with elevated expression at 10DE only the STR and not the
PFC (Fig. 3DE). There was no change in the level of Ldha transcript expression in the ERC (Fig. 3F).
Slc2a1 (glucose transporter gene) was similar to previously observed with increased expression at 5DE in
both the STR and PFC with elevated expression maintained at 10DE in both regions. At the 5DE+5DR
groups transcript levels returned to baseline (Fig. 3GH). In the ERC only the 10DE group showed elevated
expression of Slc2a1 (Fig. 3I). Slc16a1 (monocarboxylate transporter 1 gene) showed no significant
increase in transcript at the 5DE or 10DE groups in the STR but only the 5DE group in the PFC (Fig. 3 JK).
Transcript expression levels remained elevated in the 5DE+5DR groups for both the STR and PFC. In the
ERC transcript expression levels were elevated in the 10DE group and the 5DE+5DR group (Fig. 3L).
Slc16a7 (monocarboxylate transporter 2) transcript showed an elevation in expression at only the 5DE
group in both the STR and PFC (Fig. 3MN). There was no change in the level of Slc16a7 transcript
expression in the ERC Fig. 3O). Vegfa transcript showed elevated expression in the STR at 5DE and 10DE
but only at 5DE in the PFC (Fig. 3PQ). There was no significant change in Vegfa transcript expression in
the ERC (Fig. 3R).

52

0
0.5
1.0
1.5
2.0
Hif1a (fold-change regulation)
NE 1DE 5DE 10DE 5DE
5DR
+
0
0.5
1.0
1.5
2.0
Hif1a (fold-change regulation)
NE 1DE 5DE 10DE 5DE
5DR
+
0
0.5
1.0
1.5
Hif1a (fold-change regulation)
NE 1DE 5DE 10DE 5DE
5DR
+
Ldha (fold-change regulation)
0
0.5
1.0
1.5
NE 1DE 5DE 10DE 5DE
5DR
+
Ldha (fold-change regulation)
0
0.5
1.0
1.5
NE 1DE 5DE 10DE 5DE
5DR
+
NE 1DE 5DE 10DE 5DE
5DR
+
Ldha (fold-change regulation)
0
0.5
1.0
1.5
Slc2a1 (fold-change regulation)
0
0.5
1.0
1.5
2.0
NE 1DE 5DE 10DE 5DE
5DR
+
Slc2a1 (fold-change regulation)
0
0.5
1.0
1.5
2.0
NE 1DE 5DE 10DE 5DE
5DR
+
Slc2a1 (fold-change regulation)
0
0.5
1.0
1.5
2.0
NE 1DE 5DE 10DE 5DE
5DR
+
Slc16a1 (fold-change regulation)
0
0.5
1.0
1.5
2.0
NE 1DE 5DE 10DE 5DE
5DR
+
Slc16a1 (fold-change regulation)
0
0.5
1.0
1.5
2.0
NE 1DE 5DE 10DE 5DE
5DR
+
Slc16a1 (fold-change regulation)
0
0.5
1.0
1.5
2.0
2.5
NE 1DE 5DE 10DE 5DE
5DR
+
Slc16a7 (fold-change regulation)
0
0.5
1.0
1.5
NE 1DE 5DE 10DE 5DE
5DR
+
NE 1DE 5DE 10DE 5DE
5DR
+
Slc16a7 (fold-change regulation)
0
0.5
1.0
1.5
2.0
NE 1DE 5DE 10DE 5DE
5DR
+
Slc16a7 (fold-change regulation)
0
0.5
1.0
1.5
NE 1DE 5DE 10DE 5DE
5DR
+
Vegfa (fold-change regulation)
0
0.5
1.0
1.5
NE 1DE 5DE 10DE 5DE
5DR
+
Vegfa (fold-change regulation)
0
0.5
1.0
1.5
2.0
NE 1DE 5DE 10DE 5DE
5DR
+
Vegfa (fold-change regulation)
0
0.5
1.0
1.5
A B C
D E F
G H I
J K L
M N O
P Q R
53

Figure 3. Temporal and regional differences in the pattern of expression of Hif1a and selected downstream
targets. Tissues from the prefrontal cortex (PFC), striatum (STR), and ectorhinal cortex (ERC) were
collected and analyzed for gene expression by qPCR in groups with no exercise (NE, white bars), 1 (1DE),
5 (5DE), or 10 (10DE) days of exercise (black bars), and 5 days of exercise followed by 5 days of rest
(5DE+5DR), gray bars). Transcript analysis including Hif1a (Hypoxia-inducible factor 1, alpha), Ldha
(Lactate dehydrogenase A), Slc2a1 (glucose transporter 1), Slc16a1 (monocarboxylate transporter 1),
Slc16a7 (monocarboxylate transporter 2), and Vegfa (vascular endothelial growth factor A). Statistical
analysis of significant changes in expression was made versus the NE group. * denotes significance at p <
0.05. N = 5 mice per group.

DISCUSSION
As we continue to explore the effects of exercise on neuronal circuitry and their underlying
behaviors it is beginning to emerge that exercise has specific effects on the brain that are regional and
anatomically specific. This has been demonstrated in several recent reports by Holschneider and
colleagues where they utilized the tracer [
14
C]-iodoantipyrine with functional brain mapping and showed
that in both normal rodents and in models of disease different forms of exercise can alter rCBF with
increased blood flow observed in regions of the brain whose circuits are engaged in specific motor
behaviors (Yang et al., 2007; Wang et al., 2013; Wang et al., 2015a; Wang et al., 2015b; Wang et al., 2016).
For example, comparing skilled versus aerobic forms of exercise on a motorized running wheel implicated
the activation of the prefrontal cortex, dorsal striatum, motor cortex, and cerebellar vermis with increased
blood flow.
Hypoxia-inducible factor-1a plays a critical role in response to metabolic demand in the brain. In its
inactive form, HIF-1a resides within the cytoplasm as part of a multimeric protein complex. Initially identified
as a global regulatory component in cancer cells to better understand metabolic homeostasis, HIF-1a has
emerged as a potentially important component in neurons in maintaining homeostasis and responding to
cellular stress and injury. Activation occurs in response to a number of molecular perturbations including
reduced oxygen tension (hypoxia), increased reactive oxygen species (ROS), elevated levels of
metabolites and small molecules including glutamate and potassium (Sharp and Bernaudin, 2004). While
54

the precise trigger for HIF-1a activation during exercise remains unknown, it is likely due to a combination
of factors, all of which are elevated with activity.
The major findings from these studies are that within the normal rodent brain, exercise in the form
of running on a motorized treadmill leads to the early activation of the transcription factor HIF-1a and several
of its downstream targets. In addition, we found that this exercise-dependent activation of HIF-1a-
dependent targets occurs in a fashion that supports a temporal and region-specific mode of action rather
than simply global activation throughout the brain. Specifically, within regions of the brain that display
elevation of rCBF with exercise, including the STR and PFC, we observed an association in the expression
of several HIF-1a target genes. Within regions of the brain where exercise-induced rCBF elevation is not
observed such as the ERC, we do not observe such changes in gene activation. This is further supporter
by previous studies showing that in rodent models of dopamine-depletion with exercise we observe
changes in the dorsal striatum including increased synaptic density on medium spiny neurons (Toy et al.,
2014), altered glutamatergic neurotransmission including subunit expression and evoked excitatory
postsynaptic currents (EPSCs) and potentials (EPSPs) (VanLeeuwen et al., 2010; Kintz et al., 2013), as
well as restoration of the depletion of the dopamine D2 receptor (Fisher et al., 2008; Vučković et al., 2010).
Overall there was a high degree of consistency in terms of patterns of gene expression between the initial
microarray screen and subsequent qPCR analysis, providing strong validation of the initial screening
approach. The number of HIF-1 targets examined in this study was narrow in scope and primarily served
as a starting point to begin to understand the mechanisms by which exercise influences metabolism and
subsequently neuroplasticity. The Car9 gene encodes carbonic anhydrase 9, an enzyme that catalyzes the
conversion of carbon dioxide and water to carbonic acid and a proton. This enzyme is sensitive to hypoxia
and its expression is dependent on HIF-1a (Olive et al., 2001). For these studies it served as a reliable
marker to validate the potential link between exercise and hypoxia as an activator of transcript expression.
The Ldha gene encodes the enzyme lactate dehydrogenase (LDHA) that catalyzes the interconversion of
lactate and pyruvate with one potential role in converting astrocytic sources of lactate to pyruvate in neurons
for inclusion in the tricarboxylic (TCA) cycle (Laughton et al., 2007). Slc2a1 encodes the glucose transporter
type 1 that is distributed in a wide spectrum of tissues including the endothelial cells responsible for the
blood-brain barrier (BBB) and whose expression is elevated with increased cellular glucose metabolism
55

(Torp et al., 1997). The Vegfa gene encodes the signaling protein vascular endothelial growth factor A
(VEGFA) whose induction under conditions of hypoxia leads to the promotion of vascularization and
increased blood flow (Fabel et al., 2003). Elevated VEGF protein has been shown to occur with exercise
and is region-specific, occurring within activated sites within the brain including the hippocampus. Together,
the pattern of expression of these four transcripts support the notion that specific forms of behavioral
activity, such as treadmill running, leads to the activation of specific neuronal circuits in restricted regions
of the brain. These findings support the potentially important role of targeted exercise and physical activity
to engage those circuits that are affected or are at risk in both normal aging and in neurodegenerative
disorders.

56

REFERENCES

Allen A, Messier C (2013) Plastic changes in the astrocyte GLUT1 glucose transporter and beta-tubulin microtubule
protein following voluntary exercise in mice. Behav Brain Res 240:95-102.
Attwell D, Laughlin SB (2001) An energy budget for signaling in the grey matter of the brain. J Cereb Blood Flow
Metab 21:1133-1145.
Barros LF (2013) Metabolic signaling by lactate in the brain. Trends Neurosci 36:396-404.
Beaulieu JM, Gainetdinov RR (2011) The physiology, signaling, and pharmacology of dopamine receptors. Pharmacol
Rev 63:182-217.
Bednarczyk MR, Hacker LC, Fortin-Nunez S, Aumont A, Bergeron R, Fernandes KJ (2011) Distinct stages of adult
hippocampal neurogenesis are regulated by running and the running environment. Hippocampus 21:1334-
1347.
Beeler JA, Cao ZF, Kheirbek MA, Ding Y, Koranda J, Murakami M, Kang UJ, Zhuang X (2010) Dopamine-dependent
motor learning: insight into levodopa's long-duration response. Ann Neurol 67:639-647.
Beeler JA, Frank MJ, McDaid J, Alexander E, Turkson S, Bernardez Sarria MS, Bernandez MS, McGehee DS, Zhuang X
(2012) A role for dopamine-mediated learning in the pathophysiology and treatment of Parkinson's disease.
Cell Rep 2:1747-1761.
Bernardi C, Tramontina AC, Nardin P, Biasibetti R, Costa AP, Vizueti AF, Batassini C, Tortorelli LS, Wartchow KM,
Dutra MF, Bobermin L, Sesterheim P, Quincozes-Santos A, de Souza J, Gonçalves CA (2013) Treadmill
exercise induces hippocampal astroglial alterations in rats. Neural Plast 2013:709732.
Berti V, Polito C, Ramat S, Vanzi E, De Cristofaro MT, Pellicanò G, Mungai F, Marini P, Formiconi AR, Sorbi S, Pupi A
(2010) Brain metabolic correlates of dopaminergic degeneration in de novo idiopathic Parkinson's disease.
Eur J Nucl Med Mol Imaging 37:537-544.
Bittar PG, Charnay Y, Pellerin L, Bouras C, Magistretti PJ (1996) Selective distribution of lactate dehydrogenase
isoenzymes in neurons and astrocytes of human brain. J Cereb Blood Flow Metab 16:1079-1089.
Black JE, Isaacs KR, Anderson BJ, Alcantara AA, Greenough WT (1990) Learning causes synaptogenesis, whereas
motor activity causes angiogenesis, in cerebellar cortex of adult rats. Proc Natl Acad Sci U S A 87:5568-5572.
Bliss TV, Cooke SF (2011) Long-term potentiation and long-term depression: a clinical perspective. Clinics (Sao Paulo)
66 Suppl 1:3-17.
Blomstrand E, Perrett D, Parry-Billings M, Newsholme EA (1989) Effect of sustained exercise on plasma amino acid
concentrations and on 5-hydroxytryptamine metabolism in six different brain regions in the rat. Acta Physiol
Scand 136:473-481.
Boumezbeur F, Petersen KF, Cline GW, Mason GF, Behar KL, Shulman GI, Rothman DL (2010) The contribution of
blood lactate to brain energy metabolism in humans measured by dynamic 13C nuclear magnetic resonance
spectroscopy. J Neurosci 30:13983-13991.
Brown BS, Payne T, Kim C, Moore G, Krebs P, Martin W (1979) Chronic response of rat brain norepinephrine and
serotonin levels to endurance training. J Appl Physiol Respir Environ Exerc Physiol 46:19-23.
Calabresi P, Mercuri NB, Sancesario G, Bernardi G (1993) Electrophysiology of dopamine-denervated striatal
neurons. Implications for Parkinson's disease. Brain 116 ( Pt 2):433-452.
Carbon M, Eidelberg D (2006) Functional imaging of sequence learning in Parkinson's disease. J Neurol Sci 248:72-
77.
Catlow BJ, Rowe AR, Clearwater CR, Mamcarz M, Arendash GW, Sanchez-Ramos J (2009) Effects of environmental
enrichment and physical activity on neurogenesis in transgenic PS1/APP mice. Brain Res 1256:173-179.
Cazorla M, Shegda M, Ramesh B, Harrison NL, Kellendonk C (2012) Striatal D2 receptors regulate dendritic
morphology of medium spiny neurons via Kir2 channels. J Neurosci 32:2398-2409.
Chen H, Zhang SM, Schwarzschild MA, Hernán MA, Ascherio A (2005) Physical activity and the risk of Parkinson
disease. Neurology 64:664-669.
Clark PJ, Bhattacharya TK, Miller DS, Kohman RA, DeYoung EK, Rhodes JS (2012) New neurons generated from
running are broadly recruited into neuronal activation associated with three different hippocampus-
involved tasks. Hippocampus 22:1860-1867.
Correia SC, Moreira PI (2010) Hypoxia-inducible factor 1: a new hope to counteract neurodegeneration? J
Neurochem 112:1-12.
57

Correia SC, Cardoso S, Santos RX, Carvalho C, Santos MS, Perry G, Smith MA, Moreira PI (2011) New insights into the
mechanisms of mitochondrial preconditioning-triggered neuroprotection. Curr Pharm Des 17:3381-3389.
Cotman CW, Berchtold NC, Christie LA (2007) Exercise builds brain health: key roles of growth factor cascades and
inflammation. Trends Neurosci 30:464-472.
Courtière A, Hardouin J, Burle B, Vidal F, Turle-Lorenzo N, Amalric M, Hasbroucq T (2011) Dynamics of executive
control and motor deficits in parkinsonian rats. J Neurosci 31:11929-11933.
Curlik DM, Shors TJ (2013) Training your brain: Do mental and physical (MAP) training enhance cognition through
the process of neurogenesis in the hippocampus? Neuropharmacology 64:506-514.
Dirnberger G, Jahanshahi M (2013) Executive dysfunction in Parkinson's disease: a review. J Neuropsychol 7:193-
224.
Eadie BD, Redila VA, Christie BR (2005) Voluntary exercise alters the cytoarchitecture of the adult dentate gyrus by
increasing cellular proliferation, dendritic complexity, and spine density. J Comp Neurol 486:39-47.
Eddy MC, Stansfield KJ, Green JT (2014) Voluntary exercise improves performance of a discrimination task through
effects on the striatal dopamine system. Learn Mem 21:334-337.
Ehninger D, Kempermann G (2003) Regional effects of wheel running and environmental enrichment on cell genesis
and microglia proliferation in the adult murine neocortex. Cereb Cortex 13:845-851.
Elsworth JD, Leranth C, Redmond DE, Roth RH (2013) Loss of asymmetric spine synapses in prefrontal cortex of
motor-asymptomatic, dopamine-depleted, cognitively impaired MPTP-treated monkeys. Int J
Neuropsychopharmacol 16:905-912.
Engeln M (2013) Throwing some light on executive function in Parkinson's disease. Mov Disord 28:1052.
Fabel K, Tam B, Kaufer D, Baiker A, Simmons N, Kuo CJ, Palmer TD (2003) VEGF is necessary for exercise-induced
adult hippocampal neurogenesis. Eur J Neurosci 18:2803-2812.
Fasano C, Bourque MJ, Lapointe G, Leo D, Thibault D, Haber M, Kortleven C, Desgroseillers L, Murai KK, Trudeau L
(2013) Dopamine facilitates dendritic spine formation by cultured striatal medium spiny neurons through
both D1 and D2 dopamine receptors. Neuropharmacology 67:432-443.
Filoteo JV, Reed JD, Litvan I, Harrington DL (2014) Volumetric correlates of cognitive functioning in nondemented
patients with Parkinson's disease. Mov Disord 29:360-367.
Fisher BE, Petzinger GM, Nixon K, Hogg E, Bremmer S, Meshul CK, Jakowec MW (2004) Exercise-induced behavioral
recovery and neuroplasticity in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned mouse basal
ganglia. J Neurosci Res 77:378-390.
Fisher BE, Li Q, Nacca A, Salem GJ, Song J, Yip J, Hui JS, Jakowec MW, Petzinger GM (2013) Treadmill exercise elevates
striatal dopamine D2 receptor binding potential in patients with early Parkinson's disease. Neuroreport
24:509-514.
Floresco SB (2013) Prefrontal dopamine and behavioral flexibility: shifting from an "inverted-U" toward a family of
functions. Front Neurosci 7:62.
Foley TE, Fleshner M (2008) Neuroplasticity of dopamine circuits after exercise: implications for central fatigue.
Neuromolecular Med 10:67-80.
Gandhi S, Wood-Kaczmar A, Yao Z, Plun-Favreau H, Deas E, Klupsch K, Downward J, Latchman DS, Tabrizi SJ, Wood
NW, Duchen MR, Abramov AY (2009) PINK1-associated Parkinson's disease is caused by neuronal
vulnerability to calcium-induced cell death. Mol Cell 33:627-638.
Garcia PC, Real CC, Ferreira AF, Alouche SR, Britto LR, Pires RS (2012) Different protocols of physical exercise produce
different effects on synaptic and structural proteins in motor areas of the rat brain. Brain Res 1456:36-48.
Garthe A, Roeder I, Kempermann G (2016) Mice in an enriched environment learn more flexibly because of adult
hippocampal neurogenesis. Hippocampus 26:261-271.
Gerhart-Hines Z, Rodgers JT, Bare O, Lerin C, Kim SH, Mostoslavsky R, Alt FW, Wu Z, Puigserver P (2007) Metabolic
control of muscle mitochondrial function and fatty acid oxidation through SIRT1/PGC-1alpha. EMBO J
26:1913-1923.
Gilliam PE, Spirduso WW, Martin TP, Walters TJ, Wilcox RE, Farrar RP (1984) The effects of exercise training on [3H]-
spiperone binding in rat striatum. Pharmacol Biochem Behav 20:863-867.
Hall CN, Klein-Flügge MC, Howarth C, Attwell D (2012) Oxidative phosphorylation, not glycolysis, powers presynaptic
and postsynaptic mechanisms underlying brain information processing. J Neurosci 32:8940-8951.
Harris JJ, Jolivet R, Attwell D (2012) Synaptic energy use and supply. Neuron 75:762-777.
58

Hu XL, Bergström SA, Brink M, Rönnbäck A, Dahlqvist P (2010) Enriched environment increases spinophilin mRNA
expression and spinophilin immunoreactive dendritic spines in hippocampus and cortex. Neurosci Lett
476:79-83.
Huganir RL, Nicoll RA (2013) AMPARs and synaptic plasticity: the last 25 years. Neuron 80:704-717.
Jakowec MW, Petzinger GM (2004) 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned model of parkinson's
disease, with emphasis on mice and nonhuman primates. Comp Med 54:497-513.
Kaasinen V, Nurmi E, Brück A, Eskola O, Bergman J, Solin O, Rinne JO (2001) Increased frontal [(18)F]fluorodopa
uptake in early Parkinson's disease: sex differences in the prefrontal cortex. Brain 124:1125-1130.
Kempermann G, Gast D, Gage FH (2002) Neuroplasticity in old age: sustained fivefold induction of hippocampal
neurogenesis by long-term environmental enrichment. Ann Neurol 52:135-143.
Khatri N, Man HY (2013) Synaptic activity and bioenergy homeostasis: implications in brain trauma and
neurodegenerative diseases. Front Neurol 4:199.
Kim YM, Ji ES, Kim SH, Kim TW, Ko IG, Jin JJ, Kim CJ, Kim DH (2015) Treadmill exercise improves short-term memory
by enhancing hippocampal cell proliferation in quinolinic acid-induced Huntington's disease rats. J Exerc
Rehabil 11:5-11.
Kinni H, Guo M, Ding JY, Konakondla S, Dornbos D, Tran R, Guthikonda M, Ding Y (2011) Cerebral metabolism after
forced or voluntary physical exercise. Brain Res 1388:48-55.
Kintz N, Petzinger GM, Akopian G, Ptasnik S, Williams C, Jakowec MW, Walsh JP (2013) Exercise modifies α-amino-
3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor expression in striatopallidal neurons in the 1-
methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned mouse. J Neurosci Res 91:1492-1507.
Kirchner L, Chen WQ, Afjehi-Sadat L, Viidik A, Skalicky M, Höger H, Lubec G (2008) Hippocampal metabolic proteins
are modulated in voluntary and treadmill exercise rats. Exp Neurol 212:145-151.
Klaissle P, Lesemann A, Huehnchen P, Hermann A, Storch A, Steiner B (2012) Physical activity and environmental
enrichment regulate the generation of neural precursors in the adult mouse substantia nigra in a dopamine-
dependent manner. BMC Neurosci 13:132.
Klaus F, Hauser T, Slomianka L, Lipp HP, Amrein I (2009) A reward increases running-wheel performance without
changing cell proliferation, neuronal differentiation or cell death in the dentate gyrus of C57BL/6 mice.
Behav Brain Res 204:175-181.
Kleim JA, Barbay S, Cooper NR, Hogg TM, Reidel CN, Remple MS, Nudo RJ (2002) Motor learning-dependent
synaptogenesis is localized to functionally reorganized motor cortex. Neurobiol Learn Mem 77:63-77.
Kohl Z, Kandasamy M, Winner B, Aigner R, Gross C, Couillard-Despres S, Bogdahn U, Aigner L, Winkler J (2007)
Physical activity fails to rescue hippocampal neurogenesis deficits in the R6/2 mouse model of Huntington's
disease. Brain Res 1155:24-33.
Kohman RA, DeYoung EK, Bhattacharya TK, Peterson LN, Rhodes JS (2012) Wheel running attenuates microglia
proliferation and increases expression of a proneurogenic phenotype in the hippocampus of aged mice.
Brain Behav Immun 26:803-810.
Kozorovitskiy Y, Gross CG, Kopil C, Battaglia L, McBreen M, Stranahan AM, Gould E (2005) Experience induces
structural and biochemical changes in the adult primate brain. Proc Natl Acad Sci U S A 102:17478-17482.
Krebiehl G, Ruckerbauer S, Burbulla LF, Kieper N, Maurer B, Waak J, Wolburg H, Gizatullina Z, Gellerich FN, Woitalla
D, Riess O, Kahle PJ, Proikas-Cezanne T, Krüger R (2010) Reduced basal autophagy and impaired
mitochondrial dynamics due to loss of Parkinson's disease-associated protein DJ-1. PLoS One 5:e9367.
Kreitzer AC, Malenka RC (2008) Striatal plasticity and basal ganglia circuit function. Neuron 60:543-554.
Langston JW (2017) The MPTP Story. J Parkinsons Dis 7:S11-S22.
Langston JW, Ballard P, Tetrud JW, Irwin I (1983) Chronic Parkinsonism in humans due to a product of meperidine-
analog synthesis. Science 219:979-980.
Lau YS, Patki G, Das-Panja K, Le WD, Ahmad SO (2011) Neuroprotective effects and mechanisms of exercise in a
chronic mouse model of Parkinson's disease with moderate neurodegeneration. Eur J Neurosci 33:1264-
1274.
Leggio MG, Mandolesi L, Federico F, Spirito F, Ricci B, Gelfo F, Petrosini L (2005) Environmental enrichment promotes
improved spatial abilities and enhanced dendritic growth in the rat. Behav Brain Res 163:78-90.
Leh SE, Petrides M, Strafella AP (2010) The neural circuitry of executive functions in healthy subjects and Parkinson's
disease. Neuropsychopharmacology 35:70-85.
59

Lie DC, Song H, Colamarino SA, Ming GL, Gage FH (2004) Neurogenesis in the adult brain: new strategies for central
nervous system diseases. Annu Rev Pharmacol Toxicol 44:399-421.
MacRae PG, Spirduso WW, Cartee GD, Farrar RP, Wilcox RE (1987) Endurance training effects on striatal D2
dopamine receptor binding and striatal dopamine metabolite levels. Neurosci Lett 79:138-144.
Magistretti PJ, Allaman I (2015) A cellular perspective on brain energy metabolism and functional imaging. Neuron
86:883-901.
McWilliams TG, Muqit MM (2017) PINK1 and Parkin: emerging themes in mitochondrial homeostasis. Curr Opin Cell
Biol 45:83-91.
Mounsey RB, Teismann P (2010) Mitochondrial dysfunction in Parkinson's disease: pathogenesis and
neuroprotection. Parkinsons Dis 2011:617472.
Nakajima K, Uchida S, Suzuki A, Hotta H, Aikawa Y (2003) The effect of walking on regional blood flow and
acetylcholine in the hippocampus in conscious rats. Auton Neurosci 103:83-92.
Navarro A, Gomez C, López-Cepero JM, Boveris A (2004) Beneficial effects of moderate exercise on mice aging:
survival, behavior, oxidative stress, and mitochondrial electron transfer. Am J Physiol Regul Integr Comp
Physiol 286:R505-511.
Nelson AJ, Iwamoto GA (2006) Reversibility of exercise-induced dendritic attenuation in brain cardiorespiratory and
locomotor areas following exercise detraining. J Appl Physiol (1985) 101:1243-1251.
Newington JT, Rappon T, Albers S, Wong DY, Rylett RJ, Cumming RC (2012) Overexpression of pyruvate
dehydrogenase kinase 1 and lactate dehydrogenase A in nerve cells confers resistance to amyloid β and
other toxins by decreasing mitochondrial respiration and reactive oxygen species production. J Biol Chem
287:37245-37258.
Newman LA, Korol DL, Gold PE (2011) Lactate produced by glycogenolysis in astrocytes regulates memory
processing. PLoS One 6:e28427.
Obel LF, Müller MS, Walls AB, Sickmann HM, Bak LK, Waagepetersen HS, Schousboe A (2012) Brain glycogen-new
perspectives on its metabolic function and regulation at the subcellular level. Front Neuroenergetics 4:3.
Onla-or S, Winstein CJ (2008) Determining the optimal challenge point for motor skill learning in adults with
moderately severe Parkinson's disease. Neurorehabil Neural Repair 22:385-395.
Opendak M, Gould E (2015) Adult neurogenesis: a substrate for experience-dependent change. Trends Cogn Sci
19:151-161.
Pagliari R, Peyrin L (1995) Norepinephrine release in the rat frontal cortex under treadmill exercise: a study with
microdialysis. J Appl Physiol (1985) 78:2121-2130.
Pellerin L, Magistretti PJ (1994) Glutamate uptake into astrocytes stimulates aerobic glycolysis: a mechanism
coupling neuronal activity to glucose utilization. Proc Natl Acad Sci U S A 91:10625-10629.
Petzinger GM, Fisher BE, McEwen S, Beeler JA, Walsh JP, Jakowec MW (2013) Exercise-enhanced neuroplasticity
targeting motor and cognitive circuitry in Parkinson's disease. Lancet Neurol 12:716-726.
Petzinger GM, Holschneider DP, Fisher BE, McEwen S, Kintz N, Halliday M, Toy W, Walsh JW, Beeler J, Jakowec MW
(2015) The Effects of Exercise on Dopamine Neurotransmission in Parkinson's Disease: Targeting
Neuroplasticity to Modulate Basal Ganglia Circuitry. Brain Plast 1:29-39.
Petzinger GM, Walsh JP, Akopian G, Hogg E, Abernathy A, Arevalo P, Turnquist P, Vucković M, Fisher BE, Togasaki
DM, Jakowec MW (2007) Effects of treadmill exercise on dopaminergic transmission in the 1-methyl-4-
phenyl-1,2,3,6-tetrahydropyridine-lesioned mouse model of basal ganglia injury. J Neurosci 27:5291-5300.
Pickel VM, Johnson E, Carson M, Chan J (1992) Ultrastructure of spared dopamine terminals in caudate-putamen
nuclei of adult rats neonatally treated with intranigral 6-hydroxydopamine. Brain Res Dev Brain Res 70:75-
86.
Pitkälä KH, Pöysti MM, Laakkonen ML, Tilvis RS, Savikko N, Kautiainen H, Strandberg TE (2013) Effects of the Finnish
Alzheimer disease exercise trial (FINALEX): a randomized controlled trial. JAMA Intern Med 173:894-901.
Polito C, Berti V, Ramat S, Vanzi E, De Cristofaro MT, Pellicanò G, Mungai F, Marini P, Formiconi AR, Sorbi S, Pupi A
(2012) Interaction of caudate dopamine depletion and brain metabolic changes with cognitive dysfunction
in early Parkinson's disease. Neurobiol Aging 33:206.e229-239.
Potter MC, Yuan C, Ottenritter C, Mughal M, van Praag H (2010) Exercise is not beneficial and may accelerate
symptom onset in a mouse model of Huntington's disease. PLoS Curr 2:RRN1201.
Quintanilla RA, Tapia C, Pérez MJ (2017) Possible role of mitochondrial permeability transition pore in the
pathogenesis of Huntington disease. Biochem Biophys Res Commun 483:1078-1083.
60

Quintanilla RA, von Bernhardi R, Godoy JA, Inestrosa NC, Johnson GV (2014) Phosphorylated tau potentiates Aβ-
induced mitochondrial damage in mature neurons. Neurobiol Dis 71:260-269.
Real CC, Ferreira AF, Hernandes MS, Britto LR, Pires RS (2010) Exercise-induced plasticity of AMPA-type glutamate
receptor subunits in the rat brain. Brain Res 1363:63-71.
Risterucci C, Terramorsi D, Nieoullon A, Amalric M (2003) Excitotoxic lesions of the prelimbic-infralimbic areas of the
rodent prefrontal cortex disrupt motor preparatory processes. Eur J Neurosci 17:1498-1508.
Rosano C, Studenski SA, Aizenstein HJ, Boudreau RM, Longstreth WT, Newman AB (2012) Slower gait, slower
information processing and smaller prefrontal area in older adults. Age Ageing 41:58-64.
Rothstein JD, Martin L, Levey AI, Dykes-Hoberg M, Jin L, Wu D, Nash N, Kuncl RW (1994) Localization of neuronal and
glial glutamate transporters. Neuron 13:713-725.
Sawamoto N, Piccini P, Hotton G, Pavese N, Thielemans K, Brooks DJ (2008) Cognitive deficits and striato-frontal
dopamine release in Parkinson's disease. Brain 131:1294-1302.
Schapira AH, Cooper JM, Dexter D, Clark JB, Jenner P, Marsden CD (1990) Mitochondrial complex I deficiency in
Parkinson's disease. J Neurochem 54:823-827.
Schwarting RK, Huston JP (1996) The unilateral 6-hydroxydopamine lesion model in behavioral brain research.
Analysis of functional deficits, recovery and treatments. Prog Neurobiol 50:275-331.
Semenova TP, Ivanov VA, Tret'yak TM (1981) Brain levels of noradrenalin, dopamine, and serotonin in rats with
different levels of motor activity. Neurosci Behav Physiol 11:153-155.
Sharp FR, Ran R, Lu A, Tang Y, Strauss KI, Glass T, Ardizzone T, Bernaudin M (2004) Hypoxic preconditioning protects
against ischemic brain injury. NeuroRx 1:26-35.
Shweiki D, Itin A, Soffer D, Keshet E (1992) Vascular endothelial growth factor induced by hypoxia may mediate
hypoxia-initiated angiogenesis. Nature 359:843-845.
Simpson IA, Carruthers A, Vannucci SJ (2007) Supply and demand in cerebral energy metabolism: the role of nutrient
transporters. J Cereb Blood Flow Metab 27:1766-1791.
Smeyne M, Sladen P, Jiao Y, Dragatsis I, Smeyne RJ (2015) HIF1α is necessary for exercise-induced neuroprotection
while HIF2α is needed for dopaminergic neuron survival in the substantia nigra pars compacta.
Neuroscience 295:23-38.
Smith D, Pernet A, Hallett WA, Bingham E, Marsden PK, Amiel SA (2003) Lactate: a preferred fuel for human brain
metabolism in vivo. J Cereb Blood Flow Metab 23:658-664.
Smith Y, Villalba RM, Raju DV (2009) Striatal spine plasticity in Parkinson's disease: pathological or not? Parkinsonism
Relat Disord 15 Suppl 3:S156-161.
Steiner B, Winter C, Hosman K, Siebert E, Kempermann G, Petrus DS, Kupsch A (2006) Enriched environment induces
cellular plasticity in the adult substantia nigra and improves motor behavior function in the 6-OHDA rat
model of Parkinson's disease. Exp Neurol 199:291-300.
Steiner JL, Murphy EA, McClellan JL, Carmichael MD, Davis JM (2011) Exercise training increases mitochondrial
biogenesis in the brain. J Appl Physiol (1985) 111:1066-1071.
Suzuki A, Stern SA, Bozdagi O, Huntley GW, Walker RH, Magistretti PJ, Alberini CM (2011) Astrocyte-neuron lactate
transport is required for long-term memory formation. Cell 144:810-823.
Takimoto M, Hamada T (2014) Acute exercise increases brain region-specific expression of MCT1, MCT2, MCT4,
GLUT1, and COX IV proteins. J Appl Physiol (1985) 116:1238-1250.
Thacker EL, Chen H, Patel AV, McCullough ML, Calle EE, Thun MJ, Schwarzschild MA, Ascherio A (2008) Recreational
physical activity and risk of Parkinson's disease. Mov Disord 23:69-74.
Toda N, Ayajiki K, Okamura T (2009) Cerebral blood flow regulation by nitric oxide: recent advances. Pharmacol Rev
61:62-97.
Toy WA, Petzinger GM, Leyshon BJ, Akopian GK, Walsh JP, Hoffman MV, Vučković MG, Jakowec MW (2014) Treadmill
exercise reverses dendritic spine loss in direct and indirect striatal medium spiny neurons in the 1-methyl-
4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mouse model of Parkinson's disease. Neurobiol Dis 63:201-
209.
Uchida S, Suzuki A, Kagitani F, Hotta H (2006) Responses of acetylcholine release and regional blood flow in the
hippocampus during walking in aged rats. J Physiol Sci 56:253-257.
van Hall G, Strømstad M, Rasmussen P, Jans O, Zaar M, Gam C, Quistorff B, Secher NH, Nielsen HB (2009) Blood
lactate is an important energy source for the human brain. J Cereb Blood Flow Metab 29:1121-1129.
61

van Praag H, Kempermann G, Gage FH (1999) Running increases cell proliferation and neurogenesis in the adult
mouse dentate gyrus. Nat Neurosci 2:266-270.
VanLeeuwen JE, Petzinger GM, Walsh JP, Akopian GK, Vuckovic M, Jakowec MW (2010) Altered AMPA receptor
expression with treadmill exercise in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned mouse
model of basal ganglia injury. J Neurosci Res 88:650-668.
Viboolvorakul S, Patumraj S (2014) Exercise training could improve age-related changes in cerebral blood flow and
capillary vascularity through the upregulation of VEGF and eNOS. Biomed Res Int 2014:230791.
Villalba RM, Lee H, Smith Y (2009) Dopaminergic denervation and spine loss in the striatum of MPTP-treated
monkeys. Exp Neurol 215:220-227.
Voelcker-Rehage C, Godde B, Staudinger UM (2010) Physical and motor fitness are both related to cognition in old
age. Eur J Neurosci 31:167-176.
Voelcker-Rehage C, Godde B, Staudinger UM (2011) Cardiovascular and coordination training differentially improve
cognitive performance and neural processing in older adults. Front Hum Neurosci 5:26.
Vučković MG, Li Q, Fisher B, Nacca A, Leahy RM, Walsh JP, Mukherjee J, Williams C, Jakowec MW, Petzinger GM
(2010) Exercise elevates dopamine D2 receptor in a mouse model of Parkinson's disease: in vivo imaging
with [¹⁸F]fallypride. Mov Disord 25:2777-2784.
Wang Z, Guo Y, Myers KG, Heintz R, Holschneider DP (2015a) Recruitment of the prefrontal cortex and cerebellum
in Parkinsonian rats following skilled aerobic exercise. Neurobiol Dis 77:71-87.
Wang Z, Guo Y, Myers KG, Heintz R, Peng YH, Maarek JM, Holschneider DP (2015b) Exercise alters resting-state
functional connectivity of motor circuits in parkinsonian rats. Neurobiol Aging 36:536-544.
Wong-Riley MT (1989) Cytochrome oxidase: an endogenous metabolic marker for neuronal activity. Trends Neurosci
12:94-101.
Xie W, Chung KK (2012) Alpha-synuclein impairs normal dynamics of mitochondria in cell and animal models of
Parkinson's disease. J Neurochem 122:404-414.
Yang J, Ruchti E, Petit JM, Jourdain P, Grenningloh G, Allaman I, Magistretti PJ (2014) Lactate promotes plasticity
gene expression by potentiating NMDA signaling in neurons. Proc Natl Acad Sci U S A 111:12228-12233.
Yawata S, Yamaguchi T, Danjo T, Hikida T, Nakanishi S (2012) Pathway-specific control of reward learning and its
flexibility via selective dopamine receptors in the nucleus accumbens. Proc Natl Acad Sci U S A 109:12764-
12769.
Yim AJ, Andersen ML, Soeiro AC, Tufik S, Oliveira MG (2009) Acute systemic blockade of D2 receptors does not
accelerate the extinction of cocaine-associated place preference. Brain Res 1304:122-128.
Zhu LL, Zhao T, Li HS, Zhao H, Wu LY, Ding AS, Fan WH, Fan M (2005) Neurogenesis in the adult rat brain after
intermittent hypoxia. Brain Res 1055:1-6.
Öhman H, Savikko N, Strandberg TE, Kautiainen H, Raivio MM, Laakkonen ML, Tilvis R, Pitkälä KH (2016) Effects of
Exercise on Cognition: The Finnish Alzheimer Disease Exercise Trial: A Randomized, Controlled Trial. J Am
Geriatr Soc 64:731-738.

62

APPENDIX A: Relationship between cyclophilin A levels and matrix metalloproteinase-9 activity in
cerebrospinal fluid of cognitively normal apolipoprotein E4 carriers and blood-brain barrier
breakdown

Authors: Matthew R. Halliday, Nunzio Pomara, Abhay P. Sagare, Wendy J. Mack, Blas Frangione, and
Berislav V. Zlokovic

INTRODUCTION

In humans, apolipoprotein E (apoE) has 3 isoforms: apoE2, apoE3, and apoE4. APOE4 is a major
genetic risk factor for Alzheimer’s disease (AD) (Verghese et al., 2011). Apolipoprotein E4 has direct effects
on the cerebrovascular system, resulting in microvascular lesions and blood-brain barrier (BBB) damage,
as recently reviewed (Zlokovic, 2013). Neurovascular dysfunction is also present in cognitively normal
APOE4 carriers and individuals with APOE4-associated disorders, including AD (Verghese et al., 2011;
Zlokovic, 2011, 2013). Moreover, postmortem brain tissue analysis has indicated that BBB breakdown in
patients with AD is more pronounced in APOE4 carries compared with APOE3 or APOE2 (Zipser et al.,
2007; Cortes-Canteli et al., 2012; Hultman et al., 2013). Our recent studies in transgenic mice have
demonstrated that apoE4 leads to BBB breakdown by activating the proinflammatory cyclophilin A (CypA)-
matrix metalloproteinase-9 (MMP-9) pathway in brain pericytes, which in turn results in degradation of the
BBB tight junctions and basement membrane proteins (Bell et al., 2012). It has also been shown that apoE4-
mediated BBB breakdown leads to secondary neuronal injury and cognitive decline in transgenic mice (Bell
et al., 2012). Apolipoprotein E2 and apoE3 maintained normal BBB integrity in transgenic mice by
suppressing the CypA-MMP-9 pathway (Bell et al., 2012). Here, we studied the cerebrospinal fluid
(CSF)/plasma albumin quotient (Qalb), an established marker of BBB breakdown (Blennow et al., 1990),
and CypA and active MMP-9 levels in the CSF of cognitively normal individuals with different APOE
genotypes to determine whether apoE4-dependent changes in BBB permeability and CypA-MMP-9
pathway as shown in APOE4, but not APOE3 and APOE2 transgenic mice, also occur in humans.

METHODS

Participants were volunteers who were recruited through advertisements or from the Memory
Education and Research Initiative Program at the Nathan S. Kline Institute for Psychiatric Research
(Pomara et al., 2012). Participants gave their informed consent to participate in studies approved by the
63

institutional review board of the Nathan S. Kline Institute for Psychiatric Research and New York University
School of Medicine. We studied a total of 49 cognitively normal individuals as indicated by the Clinical
Dementia Rating score of 0 and Mini-Mental State Examination score of approximately 30. This study did
not exclude participants meeting criteria for major depressive disorder because there was no difference in
the studied markers of BBB damage in this group compared with controls. The studied individuals
represented 3 different APOE genotypes: APOE2/E3 (n = 11), APOE3/E3 (n = 28), and APOE3/E4 (n =
10). Within each genotype, individuals were stratified into 2 age groups – 40 through 65 years old and 66
through 85 years old – to control for age-dependent effects (Table). Cerebrospinal fluid and plasma
collection and APOE genotyping were performed as described (Pomara et al., 2012). Enzyme-linked
immunosorbent assays were used to determine levels of CypA (catalog no. sE90979Hu; USCN Life
Science), active MMP-9 (catalog no. 72017; AnaSpec), and albumin (catalog no. E-80AL; Immunology
Consultant Laboratories). Data were analyzed by multifactorial analysis of variance with 2 factors (age and
APOE genotype, with Bonferroni post hoc tests to adjust for multiple comparisons, and Pearson correlation
analysis using GraphPad Prism version 5.0. Analyses were performed by an investigator blinded to the
experimental conditions. A P value of less than 0.05 was considered statistically significant.

RESULTS

Older cognitively normal individuals carrying 1 APOE4 allele compared with younger cognitively
normal APOE4 carriers or age-matched APOE4 noncarriers had increased Qalb by approximately 77% and
67%, respectively (P < .01; Figure, A). No age-dependent increase in Qalb was associated with APOE2 or
64

APOE3 alleles. Compared with cognitively normal younger APOE4 carriers or age-matched APOE4
noncarriers, older cognitively normal APOE4 carriers had increased CSF levels of CypA by approximately
190% and 95%, respectively (P < .01; Figure, B) and active MMP-9 by 167% and 110%, respectively (P <
.05; Figure, C). No age-dependent changes in CypA or MMP-9 CSF levels were associated with APOE2 or
APOE3 alleles. Importantly, increases in Qalb values correlated positively with both CypA and active MMP-
9 CSF levels in all studied individuals (r = 0.37, P < .01; and r = 0.45, P < .01, respectively) (Figure, D and
E), indicating the greater the increase in CypA and active MMP-9 levels, the greater the magnitude of BBB
breakdown assayed by Qalb.

DISCUSSION

This study showed that APOE4 carriers may be susceptible to an age-dependent BBB breakdown
prior to onset of clinical decline as determined by Clinical Dementia Rating and Mini-Mental State
65

Examination scores. Furthermore, these findings are consistent with experimental studies suggesting that
apoE4 leads to BBB damage in transgenic mice via activation of the CypA-MMP-9 pathway (Bell et al.,
2012). These findings warrant future longitudinal studies to investigate Qalb and CSF levels of CypA and
active MMP-9 in cognitively normal APOE4 carriers as they progress to mild cognitive impairment and
eventually AD. With current diagnostic markers, by the time the earliest detectable clinical signs of disease
appear, significant brain injury has likely already occurred. Therefore, studying markers of BBB damage
along with commonly used b-amyloid 42 and tau CSF levels may contribute to early detection of vascular
dysfunction of those at risk for cognitive decline and AD.

REFERENCES:
Bell RD, Winkler EA, Singh I, Sagare AP, Deane R, Wu Z, Holtzman DM, Betsholtz C, Armulik A, Sallstrom J, Berk BC,
Zlokovic BV (2012) Apolipoprotein E controls cerebrovascular integrity via cyclophilin A. Nature 485:512-516.
Blennow K, Wallin A, Fredman P, Karlsson I, Gottfries CG, Svennerholm L (1990) Blood-brain barrier disturbance in
patients with Alzheimer's disease is related to vascular factors. Acta Neurol Scand 81:323-326.
Cortes-Canteli M, Zamolodchikov D, Ahn HJ, Strickland S, Norris EH (2012) Fibrinogen and altered hemostasis in
Alzheimer's disease. J Alzheimers Dis 32:599-608.
Hultman K, Strickland S, Norris EH (2013) The APOE ɛ4/ɛ4 genotype potentiates vascular fibrin(ogen) deposition in
amyloid-laden vessels in the brains of Alzheimer's disease patients. J Cereb Blood Flow Metab 33:1251-
1258.
Pomara N, Bruno D, Sarreal AS, Hernando RT, Nierenberg J, Petkova E, Sidtis JJ, Wisniewski TM, Mehta PD, Pratico
D, Zetterberg H, Blennow K (2012) Lower CSF amyloid beta peptides and higher F2-isoprostanes in
cognitively intact elderly individuals with major depressive disorder. Am J Psychiatry 169:523-530.
Verghese PB, Castellano JM, Holtzman DM (2011) Apolipoprotein E in Alzheimer's disease and other neurological
disorders. Lancet Neurol 10:241-252.
Zipser BD, Johanson CE, Gonzalez L, Berzin TM, Tavares R, Hulette CM, Vitek MP, Hovanesian V, Stopa EG (2007)
Microvascular injury and blood-brain barrier leakage in Alzheimer's disease. Neurobiol Aging 28:977-986.
Zlokovic BV (2011) Neurovascular pathways to neurodegeneration in Alzheimer's disease and other disorders. Nat
Rev Neurosci 12:723-738.
Zlokovic BV (2013) Cerebrovascular effects of apolipoprotein E: implications for Alzheimer disease. JAMA Neurol
70:440-444.

66

APPENDIX B: Blood-Brain Barrier Breakdown in the Aging Human Hippocampus

Authors: Axel Montagne, Samuel R. Barnes, Melanie D. Sweeney, Matthew R. Halliday, Abhay P. Sagare,
Zhen Zhao, Arthur W. Toga, Russel E. Jacobs, Charles Y. Liu, Lilyana Amezcua, Michael G. Harrington,
Helena C. Chui, Meng Law, and Berislav V. Zlokovic

ABSTRACT

The blood-brain barrier (BBB) limits entry of blood-derived products, pathogens, and cells into the
brain that is essential for normal neuronal functioning and information processing. Post-mortem tissue
analysis indicates BBB damage in Alzheimer’s disease (AD). The timing of BBB breakdown remains,
however, elusive. Using an advanced dynamic contrast-enhanced MRI protocol with high spatial and
temporal resolutions to quantify regional BBB permeability in the living human brain, we show an age-
dependent BBB breakdown in the hippocampus, a region critical for learning and memory that is affected
early in AD. The BBB breakdown in the hippocampus and its CA1 and dentate gyrus subdivisions worsened
with mild cognitive impairment that correlated with injury to BBB-associated pericytes, as shown by the
cerebrospinal fluid analysis. Our data suggest that BBB breakdown is an early event in the aging human
brain that begins in the hippocampus and may contribute to cognitive impairment.

INTRODUCTION
Neuronal computation and normal functioning of the CNS requires tight control of the chemical
composition of the neuronal ‘‘milieu’’ that is maintained by the blood-brain barrier (BBB) (Iadecola, 2004;
Zlokovic, 2008, 2011). Brain endothelial cells and perivascular mural cells, pericytes, form the BBB, which
limits entry of neurotoxic plasma-derived proteins, circulating metals, pathogens, red blood cells, and
leucocytes into the brain. Studies in murine transgenic models have shown that a chronic BBB breakdown
leads to accumulation of blood-derived neurotoxic proteins in the CNS including fibrin, thrombin,
hemoglobin, iron-containing hemosiderin, free iron, and/or plasmin (an extracellular matrix-degrading
enzyme) causing progressive neurodegeneration with loss of neurons mediated by direct neuronal toxicity,
oxidant stress, and/or detachment of neurons from their supporting extracellular matrix (Armulik et al., 2010;
Bell et al., 2010; Daneman et al., 2010; Winkler et al., 2011; Bell et al., 2012; Winkler et al., 2014).

67

68

Alzheimer’s disease (AD) is characterized by selective neuronal vulnerability resulting in
progressive loss of memory (LaFerla, 2012). Post-mortem studies have shown BBB damage in AD
including accumulation in the hippocampus and cortex of blood-derived proteins (e.g., immunoglobulins,
albumin, fibrinogen, and thrombin) (Fiala et al., 2002; Salloway et al., 2002; Zipser et al., 2007; Ryu and
McLarnon, 2009; Hultman et al., 2013; Sengillo et al., 2013) and degeneration of BBB-associated pericytes
(Farkas and Luiten, 2001; Baloyannis and Baloyannis, 2012; Sengillo et al., 2013). Brain imaging studies
have shown microbleeds and accumulation of iron in AD (Cullen et al., 2005; Goos et al., 2009; Zonneveld
et al., 2014), particularly in the hippocampus (Raven et al., 2013). Some studies using the cerebrospinal
fluid (CSF) to plasma ratio of blood-derived albumin, reported BB damage in AD particularly associated
with vascular risk factors (Blennow et al., 1990; Bowman et al., 2012) or in individuals at a genetic risk for
AD (Halliday et al., 2013). At what stage BBB breakdown occurs in the living human brain and whether it
contributes to cognitive impairment remains, however, controversial.
Here, we used an advanced dynamic contrast-enhanced MRI (DCE-MRI) and post-processing
analysis with improved spatial and temporal resolutions to quantify the BBB regional permeability Ktrans
constant in the living human brain in individuals with no cognitive impairment (NCI) and mild cognitive
impairment (MCI). Compared to a previous approach limited to measurements of BBB permeability in the
white matter (WM) (Taheri et al., 2011b; Taheri et al., 2011a), the current method allows simultaneous
measurements of BBB Ktrans permeability in different gray and WM regions. Additionally, we analyzed CSF
biomarkers of BBB breakdown, injury to brain vascular cells including pericytes and endothelial cells,
inflammatory response, neuronal injury (i.e., tau and pTau), and amyloid b-peptides (Ab). We found an age-
dependent BBB breakdown in the hippocampus, a region involved in learning and memory (Squire, 1992)
that is damaged early in AD (Braak et al., 1993; Mu and Gage, 2011; Whitwell et al., 2012). The BBB
breakdown in the hippocampus worsened with MCI that correlated with measures of injury to BBB-
associated pericytes.

69

RESULTS
NCI and MCI participants were recruited through the University of Southern California Alzheimer’s
Disease Research Center (USC ADRC) and Huntington Medical Research Institute (HMRI), Pasadena.
NCI and MCI participants were evaluated using the Uniform Data Set (Morris et al., 2006; Weintraub et al.,
2009) and additional neuropsychological tests as described in the Supplemental Information. For all
participants in this study, the DCE-MRI procedure was approved by the USC Institutional Review Board
(IRB) and HMRI IRB. Lumbar puncture was approved for older NCI and MCI participants by the USC IRB
and HMRI IRB.
BBB Breakdown in the Hippocampus during Normal Aging

We studied 12 CNS regions including hippocampus and its subdivisions CA1, CA3, and dentate
gyrus (DG), different cortical regions, subcortical regions (e.g., thalamus, striatum, and caudate nucleus),
and the WM regions including corpus callosum and internal capsule (Figures 1A and 1B) and generated
70

regional Ktrans BBB permeability maps for each individual using a modified Patlak linearized regression
mathematical analysis (Patlak and Blasberg, 1985) (Figure 1C). We determined in each individual the
arterial input function (AIF) from the common carotid artery instead of using an average value from the
superior sagittal venous sinus to determine tracer concentration in blood (Larsson et al., 2009; Taheri et
al., 2011b; Taheri et al., 2011a). Individual AIF measurements are important particularly if the studied
population diverges by age as changes in blood volume and flow may affect AIF and the Ktrans
measurements.

Unexpectedly, we found that NCI individuals (Table 1) have an age-dependent progressive loss of
BBB integrity in the hippocampus, as shown by an age-dependent increase in the Ktrans values in the
entire hippocampus, its CA1 region, and DG, but not the CA3 region (Figures 1D–1G). No significant BBB
changes during aging were found in cortical (e.g., frontal cortex, temporal cortex) or subcortical (e.g.,
thalamus, striatum) regions except for the caudate nucleus (Figures S1A–S1E). Surprisingly, we did not
find significant age-dependent changes in the BBB in subcortical WM fibers, corpus callosum, and internal
capsule (Figures S1F–S1H), even though WM is believed to be affected by vascular changes early in AD
(Iadecola, 2004; Yoshita et al., 2006). Collectively, our data suggest that early vascular leakage in the aging
human brain begins in the hippocampus, which normally shows the highest barrier properties (i.e., the
lowest Ktrans values) compared to other brain regions.

71

Accelerated BBB Breakdown in the Hippocampus in Individuals with MCI
Next, we compared the BBB permeability in young NCI and older NCI (both CDR = 0) and MCI
(CDR = 0.5) groups (Figure 2; Table 1). We found a significant increase in the BBB permeability Ktrans
values in the hippocampus (Figure 2A) and its CA1 and DG regions, but not CA3, by 41%, 107%, and 48%
in the older compared to the young NCI group and by 24%, 53%, and 27% in MCI compared to age-matched
older NCI controls, respectively (Figures 2B–2E). In contrast, there were no significant differences in the
BBB permeability in cortical, subcortical, and WM regions between young and older NCI and/or MCI and
age-matched older NCI participants (Table S1), except for an increase in the caudate nucleus in older
compared to young NCI group. We did not find significant changes in hippocampal volumes between the
studied groups determined on coronal T2-weighted MRIs (Figure 3).

72

To validate our method, we studied multiple sclerosis (MS) cases with established BBB breakdown
in the WM, as reported (Taheri et al., 2011a), as an additional neurological control. MS cases had a
diagnosis of the relapsing remitting MS and met McDonald criteria (Polman et al., 2011). We selected
younger MS cases without cognitive complaints that were age-matched to younger NCI controls (Table 1).
The MS patients did not show changes in the Ktrans values in the hippocampus or hippocampal subregions
(Figures 2B-2E), or in other CNS gray matter regions (Table S1). They had, however, increased BBB
permeability in the total WM, corpus callosum, and internal capsule compared to age-matched younger NCI
group by 32%, 26%, and 23% (p < 0.001), respectively (Table S1), consistent with the reported BBB
alterations in the WM in MS (Taheri et al., 2011a). Our data extend importantly previous findings by showing
now changes in the BBB integrity in the hippocampus and other studied gray matter regions in MS.
Molecular Biomarker CSF Analysis
A significant 30% increase in the CSF/plasma albumin ratio (Blennow et al., 1990; Bowman et al.,
2012; Halliday et al., 2013) additionally confirmed BBB breakdown in MCI individuals compared to age-
matched NCI controls (Figure S2A). The increase in the CSF plasma albumin ratio correlated with an
increase in the Ktrans values in the hippocampus and its CA1 and DG subregions (Figures S2B-S2D) that
showed an increase in the BBB permeability in MCI compared to age-matched NCI controls (Figures 2B,
2C, and 2E).
Next, we studied correlations between Ktrans values and CSF levels of soluble platelet-derived
growth factor receptor b (sPDGFRb). PDGFRb is an established marker of the BBB-associated pericytes
(Armulik et al., 2010; Bell et al., 2010; Winkler et al., 2011) that play a key role in maintaining the BBB
integrity (Zlokovic, 2008; Armulik et al., 2010; Bell et al., 2010; Daneman et al., 2010; Winkler et al., 2011;
Zlokovic, 2011; Bell et al., 2012). Pericytes degenerate in AD (Farkas and Luiten, 2001; Baloyannis and
Baloyannis, 2012; Sengillo et al., 2013) and have a key role in BBB clearance of Alzheimer’s toxin amyloid
b-peptide (Ab) (Sagare et al., 2013). Pericytes die when Ab intracellular accumulation overrides their Ab
clearance capability (Sagare et al., 2013) and when exposed to hypoxia. Here, we show that both severe
hypoxia (Bell et al., 2009) and Ab are associated with shedding of the soluble form of the receptor
(sPDGFRb) from primary human cultured pericytes (Zhu et al., 2010) (Figures S3A-S3D). Furthermore,
73

sPDGFRb levels were increased by 115% in MCI compared to age-matched NCI controls (Figure 4A).
There was a positive correlation between sPDGFRb CSF levels and the Ktrans values in the hippocampus
including its CA1 and DG subdivisions (Figures 4B-4D) that showed increased BBB permeability in MCI
compared to NCI individuals (Figures 2B, 2C, and 2E).
To validate CSF sPDGFRb as a marker of pericyte injury in vivo, we studied sPDGFRb CSF levels
in 16-month-old pericyte-deficient Pdgfrb
+/-
mice, which develop ~45%-50% loss of brain pericytes (Bell et
al., 2010), and 16-month-old Alzheimer’s Tg2576 mice, which develop an age-dependent pericyte loss from
17% at 9 months of age (Sagare et al., 2013) to 35% at 18 months of age (Park et al., 2013). There was a
significant 289% and 58% increase in sPDGFRb CSF levels in Pdgfrb
+/-
mice and Tg2576 mice,
respectively, compared to their corresponding littermate controls (Figures 4E and 4F), indicating that
sPDGFRb is a reliable CSF marker of pericyte injury in mice.
The CSF analysis revealed no injury to other cell types in the neurovascular unit (Iadecola, 2004;
Zlokovic, 2008, 2011) in NCI or MCI including endothelial cells as shown by unaltered CSF levels of
biomarkers of endothelial cell injury such as soluble intercellular adhesion molecule-1 (sVCAM-1) (Iadecola,
2004; Zlokovic, 2008); no change in the inflammatory response as shown by unaltered CSF levels of
several studied cytokines (e.g., interleukins IL-2, IL-6, and IL-8, tumor necrosis factor-a, and interferon-g);
no change in neuronal injury (e.g., tau and pTau) and Ab (e.g., Ab38, Ab40, and Ab42); and no change in
matrix metalloproteinase-9 that is involved in degradation of the BBB tight junction and the basement
membrane proteins of the vessel wall (Bell et al., 2012; Halliday et al., 2013) (Figure S4).
DISCUSSION
We developed an advanced DCE-MRI approach and post-processing analysis resulting in
improved spatial resolution and signal-to-noise ratio (SNR) of the Ktrans BBB maps with the analysis of the
arterial input function in each individual allowing for accurate measurements of the regional BBB
permeability in the living human brain in different gray and WM regions. For example, our high-resolution
hippocampal imaging allows for characterization of the Ktrans BBB values not only in the hippocampus, but
also in the hippocampal subfields. In comparison, studies on the blood-brain tumor barrier permeability
74

(Larsson et al., 2009) or BBB in stroke (Aksoy et al., 2013) do not generally require spatial resolution or
SNR as high as the present study, as changes in the barrier permeability in brain tumors or after stroke are
typically one order of magnitude or more higher than the presently measured BBB changes during normal
aging, aging associated with MCI, and/or possibly other neurodegenerative conditions. The BBB
permeability Ktrans values in the hippocampus and cortex and other brain regions in young NCI individuals
were within a range of previously reported BBB Ktrans values to small inert polar molecules in mammals
including rodents (Deane et al., 2003; Bell et al., 2010; Zlokovic, 2011).
We show that the BBB breakdown during normal aging occurs initially in the hippocampus, a region
critical for learning and memory. The BBB breakdown was more pronounced in MCI compared to age-
matched neurologically intact controls, raising a possibility that it might contribute to early cognitive
impairment. Interestingly, our data show that the BBB integrity in other brain regions including cortical and
subcortical regions or the WM remains relatively unaffected during normal aging or aging associated with
MCI. Although we did not find significant changes in hippocampal volumes between the young and older
NCI and MCI individuals, it is possible that an early and progressive increase in the BBB permeability, as
we show in the hippocampus in older NCI and MCI individuals, might precede hippocampal atrophy seen
later in AD (Apostolova et al., 2010; Whitwell et al., 2012), particularly in MCI progressing to AD. This would
be similar to findings in animal models with a chronic BBB disruption showing that vascular leakages over
time lead to hippocampal and cortical atrophy, loss of neurons, and progressive behavioral changes (Bell
et al., 2010; Winkler et al., 2011; Bell et al., 2012; Winkler et al., 2014).
Findings in murine models of a small vessel brain disease (Armulik et al., 2010; Bell et al., 2010;
Daneman et al., 2010; Bell et al., 2012) and human post-mortem AD studies (Fiala et al., 2002; Salloway
et al., 2002; Zipser et al., 2007; Ryu and McLarnon, 2009; Hultman et al., 2013; Sengillo et al., 2013) have
shown that BBB breakdown leads to tissue accumulation of potentially neurotoxic blood-derived products
that normally do not enter the brain but can damage neurons when the vessels become leaky. We show
that pericyte injury and possibly early degeneration correlates with increased BBB permeability within the
hippocampus, a region known to be affected by pericyte loss and BBB breakdown on post-mortem tissue
analysis in AD (Sengillo et al., 2013). Although, our CSF biomarkers analysis did not show endothelial cell
75

injury, involvement of inflammatory cytokines, and/or direct vasculotoxic effects of Ab in MCI, it is possible
that some of these factors could play a role in magnifying BBB damage at later disease stages during
progression to dementia due to AD, as they all were shown to alter BBB permeability in experimental models
(Zlokovic, 2011).
In summary, our data suggest loss of cerebrovascular integrity during normal aging and aging
associated with MCI that begins in the hippocampus which may contribute to early stages of dementia
associated with AD.
REFERENCES:
Aksoy D, Bammer R, Mlynash M, Venkatasubramanian C, Eyngorn I, Snider RW, Gupta SN, Narayana R,
Fischbein N, Wijman CA (2013) Magnetic resonance imaging profile of blood-brain barrier injury in patients
with acute intracerebral hemorrhage. J Am Heart Assoc 2:e000161.
Apostolova LG, Hwang KS, Andrawis JP, Green AE, Babakchanian S, Morra JH, Cummings JL, Toga AW,
Trojanowski JQ, Shaw LM, Jack CR, Petersen RC, Aisen PS, Jagust WJ, Koeppe RA, Mathis CA,
Weiner MW, Thompson PM, Initiative AsDN (2010) 3D PIB and CSF biomarker associations with
hippocampal atrophy in ADNI subjects. Neurobiol Aging 31:1284-1303.
Armulik A, Genové G, Mäe M, Nisancioglu MH, Wallgard E, Niaudet C, He L, Norlin J, Lindblom P,
Strittmatter K, Johansson BR, Betsholtz C (2010) Pericytes regulate the blood-brain barrier. Nature
468:557-561.
Baloyannis SJ, Baloyannis IS (2012) The vascular factor in Alzheimer's disease: a study in Golgi technique
and electron microscopy. J Neurol Sci 322:117-121.
Bell RD, Winkler EA, Sagare AP, Singh I, LaRue B, Deane R, Zlokovic BV (2010) Pericytes control key
neurovascular functions and neuronal phenotype in the adult brain and during brain aging. Neuron
68:409-427.
Bell RD, Deane R, Chow N, Long X, Sagare A, Singh I, Streb JW, Guo H, Rubio A, Van Nostrand W, Miano
JM, Zlokovic BV (2009) SRF and myocardin regulate LRP-mediated amyloid-beta clearance in
brain vascular cells. Nat Cell Biol 11:143-153.
Bell RD, Winkler EA, Singh I, Sagare AP, Deane R, Wu Z, Holtzman DM, Betsholtz C, Armulik A, Sallstrom
J, Berk BC, Zlokovic BV (2012) Apolipoprotein E controls cerebrovascular integrity via cyclophilin
A. Nature 485:512-516.
Blennow K, Wallin A, Fredman P, Karlsson I, Gottfries CG, Svennerholm L (1990) Blood-brain barrier
disturbance in patients with Alzheimer's disease is related to vascular factors. Acta Neurol Scand
81:323-326.
Bowman GL, Kaye JA, Quinn JF (2012) Dyslipidemia and blood-brain barrier integrity in Alzheimer's
disease. Curr Gerontol Geriatr Res 2012:184042.
Braak H, Braak E, Bohl J (1993) Staging of Alzheimer-related cortical destruction. Eur Neurol 33:403-408.
Cullen KM, Kócsi Z, Stone J (2005) Pericapillary haem-rich deposits: evidence for microhaemorrhages in
aging human cerebral cortex. J Cereb Blood Flow Metab 25:1656-1667.
Daneman R, Zhou L, Kebede AA, Barres BA (2010) Pericytes are required for blood-brain barrier integrity
during embryogenesis. Nature 468:562-566.
Deane R et al. (2003) RAGE mediates amyloid-beta peptide transport across the blood-brain barrier and
accumulation in brain. Nat Med 9:907-913.
Farkas E, Luiten PG (2001) Cerebral microvascular pathology in aging and Alzheimer's disease. Prog
Neurobiol 64:575-611.
76

Fiala M, Liu QN, Sayre J, Pop V, Brahmandam V, Graves MC, Vinters HV (2002) Cyclooxygenase-2-
positive macrophages infiltrate the Alzheimer's disease brain and damage the blood-brain barrier.
Eur J Clin Invest 32:360-371.
Goos JD, Kester MI, Barkhof F, Klein M, Blankenstein MA, Scheltens P, van der Flier WM (2009) Patients
with Alzheimer disease with multiple microbleeds: relation with cerebrospinal fluid biomarkers and
cognition. Stroke 40:3455-3460.
Halliday MR, Pomara N, Sagare AP, Mack WJ, Frangione B, Zlokovic BV (2013) Relationship between
cyclophilin a levels and matrix metalloproteinase 9 activity in cerebrospinal fluid of cognitively
normal apolipoprotein e4 carriers and blood-brain barrier breakdown. JAMA Neurol 70:1198-1200.
Hultman K, Strickland S, Norris EH (2013) The APOE ɛ4/ɛ4 genotype potentiates vascular fibrin(ogen)
deposition in amyloid-laden vessels in the brains of Alzheimer's disease patients. J Cereb Blood
Flow Metab 33:1251-1258.
Iadecola C (2004) Neurovascular regulation in the normal brain and in Alzheimer's disease. Nat Rev
Neurosci 5:347-360.
LaFerla FM (2012) Preclinical success against Alzheimer's disease with an old drug. N Engl J Med 367:570-
572.
Larsson HB, Courivaud F, Rostrup E, Hansen AE (2009) Measurement of brain perfusion, blood volume,
and blood-brain barrier permeability, using dynamic contrast-enhanced T(1)-weighted MRI at 3
tesla. Magn Reson Med 62:1270-1281.
Morris JC, Weintraub S, Chui HC, Cummings J, Decarli C, Ferris S, Foster NL, Galasko D, Graff-Radford
N, Peskind ER, Beekly D, Ramos EM, Kukull WA (2006) The Uniform Data Set (UDS): clinical and
cognitive variables and descriptive data from Alzheimer Disease Centers. Alzheimer Dis Assoc
Disord 20:210-216.
Mu Y, Gage FH (2011) Adult hippocampal neurogenesis and its role in Alzheimer's disease. Mol
Neurodegener 6:85.
Park L, Zhou J, Zhou P, Pistick R, El Jamal S, Younkin L, Pierce J, Arreguin A, Anrather J, Younkin SG,
Carlson GA, McEwen BS, Iadecola C (2013) Innate immunity receptor CD36 promotes cerebral
amyloid angiopathy. Proc Natl Acad Sci U S A 110:3089-3094.
Patlak CS, Blasberg RG (1985) Graphical evaluation of blood-to-brain transfer constants from multiple-time
uptake data. Generalizations. J Cereb Blood Flow Metab 5:584-590.
Polman CH, Reingold SC, Banwell B, Clanet M, Cohen JA, Filippi M, Fujihara K, Havrdova E, Hutchinson
M, Kappos L, Lublin FD, Montalban X, O'Connor P, Sandberg-Wollheim M, Thompson AJ, Waubant
E, Weinshenker B, Wolinsky JS (2011) Diagnostic criteria for multiple sclerosis: 2010 revisions to
the McDonald criteria. Ann Neurol 69:292-302.
Raven EP, Lu PH, Tishler TA, Heydari P, Bartzokis G (2013) Increased iron levels and decreased tissue
integrity in hippocampus of Alzheimer's disease detected in vivo with magnetic resonance imaging.
J Alzheimers Dis 37:127-136.
Ryu JK, McLarnon JG (2009) A leaky blood-brain barrier, fibrinogen infiltration and microglial reactivity in
inflamed Alzheimer's disease brain. J Cell Mol Med 13:2911-2925.
Sagare AP, Bell RD, Zhao Z, Ma Q, Winkler EA, Ramanathan A, Zlokovic BV (2013) Pericyte loss influences
Alzheimer-like neurodegeneration in mice. Nat Commun 4:2932.
Salloway S, Gur T, Berzin T, Tavares R, Zipser B, Correia S, Hovanesian V, Fallon J, Kuo-Leblanc V, Glass
D, Hulette C, Rosenberg C, Vitek M, Stopa E (2002) Effect of APOE genotype on microvascular
basement membrane in Alzheimer's disease. J Neurol Sci 203-204:183-187.
Sengillo JD, Winkler EA, Walker CT, Sullivan JS, Johnson M, Zlokovic BV (2013) Deficiency in mural
vascular cells coincides with blood-brain barrier disruption in Alzheimer's disease. Brain Pathol
23:303-310.
Squire LR (1992) Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans.
Psychol Rev 99:195-231.
Taheri S, Gasparovic C, Shah NJ, Rosenberg GA (2011a) Quantitative measurement of blood-brain barrier
permeability in human using dynamic contrast-enhanced MRI with fast T1 mapping. Magn Reson
Med 65:1036-1042.
Taheri S, Gasparovic C, Huisa BN, Adair JC, Edmonds E, Prestopnik J, Grossetete M, Shah NJ, Wills J,
Qualls C, Rosenberg GA (2011b) Blood-brain barrier permeability abnormalities in vascular
cognitive impairment. Stroke 42:2158-2163.
77

Weintraub S, Salmon D, Mercaldo N, Ferris S, Graff-Radford NR, Chui H, Cummings J, DeCarli C, Foster
NL, Galasko D, Peskind E, Dietrich W, Beekly DL, Kukull WA, Morris JC (2009) The Alzheimer's
Disease Centers' Uniform Data Set (UDS): the neuropsychologic test battery. Alzheimer Dis Assoc
Disord 23:91-101.
Whitwell JL, Wiste HJ, Weigand SD, Rocca WA, Knopman DS, Roberts RO, Boeve BF, Petersen RC, Jack
CR, Initiative ADN (2012) Comparison of imaging biomarkers in the Alzheimer Disease
Neuroimaging Initiative and the Mayo Clinic Study of Aging. Arch Neurol 69:614-622.
Winkler EA, Bell RD, Zlokovic BV (2011) Central nervous system pericytes in health and disease. Nat
Neurosci 14:1398-1405.
Winkler EA, Sengillo JD, Sagare AP, Zhao Z, Ma Q, Zuniga E, Wang Y, Zhong Z, Sullivan JS, Griffin JH,
Cleveland DW, Zlokovic BV (2014) Blood-spinal cord barrier disruption contributes to early motor-
neuron degeneration in ALS-model mice. Proc Natl Acad Sci U S A 111:E1035-1042.
Yoshita M, Fletcher E, Harvey D, Ortega M, Martinez O, Mungas DM, Reed BR, DeCarli CS (2006) Extent
and distribution of white matter hyperintensities in normal aging, MCI, and AD. Neurology 67:2192-
2198.
Zhu D, Wang Y, Singh I, Bell RD, Deane R, Zhong Z, Sagare A, Winkler EA, Zlokovic BV (2010) Protein S
controls hypoxic/ischemic blood-brain barrier disruption through the TAM receptor Tyro3 and
sphingosine 1-phosphate receptor. Blood 115:4963-4972.
Zipser BD, Johanson CE, Gonzalez L, Berzin TM, Tavares R, Hulette CM, Vitek MP, Hovanesian V, Stopa
EG (2007) Microvascular injury and blood-brain barrier leakage in Alzheimer's disease. Neurobiol
Aging 28:977-986.
Zlokovic BV (2008) The blood-brain barrier in health and chronic neurodegenerative disorders. Neuron
57:178-201.
Zlokovic BV (2011) Neurovascular pathways to neurodegeneration in Alzheimer's disease and other
disorders. Nat Rev Neurosci 12:723-738.
Zonneveld HI, Goos JD, Wattjes MP, Prins ND, Scheltens P, van der Flier WM, Kuijer JP, Muller M, Barkhof
F (2014) Prevalence of cortical superficial siderosis in a memory clinic population. Neurology
82:698-704.

78

APPENDIX C: Central role for PICALM in amyloid-b blood brain-barrier transcytosis and clearance

Authors: Zhen Zhao, Abhay P. Sagare, Qingyi Ma, Matthew R. Halliday, Pan Kong, Kassandra Kisler, Ethan
A. Winkler, Anita Ramanathan, Takahisa Kanekiyo, Guojun Bu, Nelly Chuqui Owens, Sanket V. Rege,
Gabriel Si, Ashim Ahuja, Donghui Zhu, Carol A. Miller, Julie A. Schneider, Manami Maeda, Takahiro Maeda,
Tohru Sugawara, Justin K. Ichida, and Berislav V. Zlokovic

ABSTRACT

PICALM is a highly validated genetic risk factor for Alzheimer’s disease (AD). We found that
reduced expression of PICALM in AD and murine brain endothelium correlated with amyloid-b (Ab)
pathology and cognitive impairment. Moreover, Picalm deficiency diminished Ab clearance across the
murine blood-brain barrier (BBB) and accelerated Ab pathology in a manner that was reversible by
endothelial PICALM re-expression. Using human brain endothelial monolayers, we found that PICALM
regulated PICALM/clathrin-dependent internalization of Ab bound to the low-density lipoprotein receptor
related protein-1, a key Ab clearance receptor, and guided Ab trafficking to Rab5 and Rab11, leading to Ab
endothelial transcytosis and clearance. PICALM levels and Ab clearance was reduced in AD-derived
endothelial monolayers, which was reversible by adenoviral-mediated PICALM transfer. Inducible
pluripotent stem cell–derived human endothelial cells carrying the rs3851179 protective allele exhibited
higher PICALM levels and enhanced Ab clearance. Thus, PICALM regulates Ab BBB transcytosis and
clearance, which has implications for Ab brain homeostasis and clearance therapy.
INTRODUCTION
PICALM, the gene encoding the phosphatidylinositol-binding clathrin assembly (PICALM) protein
(Dreyling et al., 1996; Tebar et al., 1999), is important for endocytosis and internalization of cell receptors
(Marsh and McMahon, 1999; Ford et al., 2001; Sorkin and von Zastrow, 2009; Treusch et al., 2011).
PICALM also mediates intracellular trafficking of endocytic proteins (Vecchi et al., 2001; Miller et al., 2011).
Several genome-wide association studies have shown the association of PICALM with AD (Harold et al.,
2009; Lambert et al., 2009; Tanzi, 2012; Lambert et al., 2013; Carrasquillo et al., 2014), a neurological
disorder characterized by neurovascular dysfunction, elevated Ab, tau pathology, and neuronal loss (Hardy
79

and Selkoe, 2002; Querfurth and LaFerla, 2010; Zlokovic, 2011). However, the role of PICALM in disease
pathogenesis remains elusive.
PICALM was originally postulated to affect disease by modifying trafficking of Ab precursor protein
(APP) (Harold et al., 2009). Recent studies suggested that PICALM protects neurons from Ab toxicity by
reversing Ab effects on clathrin-mediated endocytosis (Treusch et al., 2011) and/or by directing APP
transport to the terminal degradation pathway by autophagosomes, which reduces Ab production (Tian et
al., 2013). In contrast, viral-mediated silencing of PICALM in the hippocampal neurons in APP-
overexpressing mice has been shown to diminish Ab production, resulting in a moderate reduction in Ab
load (Xiao et al., 2012). Furthermore, PICALM influences the ratio of Ab42 to total Ab in neurons through
clathrin-mediated endocytosis of g-secretase (Kanatsu et al., 2014).
In addition to its neuronal functionality, PICALM is abundantly expressed in brain capillary
endothelium (Baig et al., 2010; Parikh et al., 2014), a site of the BBB in vivo (Zlokovic, 2011), which provides
a major pathway for Ab clearance from the brain into circulation (Shibata et al., 2000; Deane et al., 2004;
Zlokovic, 2011). Thus, PICALM is ideally situated to regulate the function of brain capillary endothelial
receptors, including receptors that mediate Ab clearance such as the low-density lipoprotein receptor-
related protein 1 (LRP1), which binds Ab and is a key Ab clearance receptor at the BBB and vascular cells
(Deane et al., 2004; Wu et al., 2005; Bell et al., 2009; Zlokovic, 2011; Kanekiyo et al., 2012; Sagare et al.,
2013). Thus, we hypothesized that PICALM influences Ab clearance across the BBB and, at the molecular
level, regulates the function of LRP1 in brain endothelial cells (Deane et al., 2004; Wu et al., 2005; Bell et
al., 2009; Zlokovic, 2011). Our data suggests that endothelial PICALM has a central role in Ab clearance
and transcytosis across the BBB, which is critical for regulation of Ab levels and homeostasis in the brain.
RESULTS
PICALM reductions in brain endothelium in AD
We found robust expression of PICALM in microvessels in aged control human brains without
dementia by immunocytochemistry (Figure 1A), immunoblotting (Figure 1B, C) and double fluorescence
80

immunostaining for PICALM and endothelial-specific Lycopersicon esculentum lectin (Figure 1D and
Supplementary Figure 1A), indicating that ~65% of the endothelial cell surface area labeled with lectin was
positive for PICALM in the hippocampus and cortex (Figure 1E). PICALM levels in isolated cortical
microvessels from control human brains were more than 1.7-fold higher than in capillary-depleted brain
homogenates containing neurons and glia (Figure 1B, C). In advanced AD (Braak stage V-VI), compared
with controls (Braak stage I), PICALM levels were reduced in cerebral microvessels (Figure 1A, B, D and
Supplementary Figure 1A) by 55-65%, as shown by immunoblotting (Figure 1E). In contrast,
immunostaining for PICALM and the neuronal marker microtubule-associated protein 2 (MAP2) revealed a
moderate increase in PICALM levels in neurons in advanced AD (Braak V-VI) compared with controls
(Figure 1D), consistent with a previous report (Ando et al., 2013). When comparing 30 AD cases (Braak
stage III-IV and V-VI) with 20 controls (Braak stage 0-1 to III) (Supplementary Table 1A), we found that
PICALM endothelial levels inversely correlated with Ab load (Figure 1F), Braak stage (Figure 1G) and
clinical dementia rating (Figure 1H), and positively correlated with mini-mental state exam scores (Figure
1I). As in humans, PICALM levels in murine brain microvessels were more than two-fold higher than in
capillary-depleted brain homogenates (Supplementary Figure 1B). Notably, elevated Ab levels in APP-
overexpressing APP
Sw/0
mice (Hsiao et al., 1996) or in endothelial cultures did not affect PICALM levels
(Supplementary Figure 1C-E), ruling out Ab as a PICALM suppressor. Thus, endothelial loss of PICALM in
AD is associated with greater Ab and AD pathology and worse cognitive impairment.
81

Diminished Ab clearance in Picalm
+/-
mice
To address whether PICALM is involved in Ab clearance across the BBB in vivo and whether it
contributes to worsening of AD pathology, we next studied clearance of human Ab40 and Ab42 in Picalm
+/-

mice (generated as shown in Supplementary Figure 2A). Complete knockout of Picalm was embryonic
lethal, but Picalm
+/-
mice developed normally and had normal blood glucose, hepatic and renal analyses
(Supplementary Figure 2B-M) and did not show behavioral changes in the first 9 months (Supplementary
Figure 2N-Q). Compared with their littermate controls, Picalm
+/-
mice had ~70% lower levels of PICALM in
brain microvessels and ~50% reduction in PICALM in capillary-depleted brain (Figure 2A, B). Using an Ab
clearance assay (Deane et al., 2004; Zlokovic, 2011), and ELISA (Sagare et al., 2013), we found that 3-
month-old Picalm
+/-
mice have 38% and 36% greater brain retention of Ab40 and Ab42, respectively, 30
min after intracerebral administration of human Ab40 or Ab42 (1 ng) than controls (Figure 2C). There was
82

no difference in retention of inulin, an inert extracellular space marker frequently used to estimate brain
interstitial fluid (ISF)-to-cerebrospinal fluid (CSF) bulk flow (Deane et al., 2004; Zlokovic, 2011) (Figure 2C).
Consistent with these results, Picalm deficiency diminished Ab40 and Ab42 efflux across the BBB by 41%
and 61%, respectively (Figure 2D and Online Methods), but did not affect Ab efflux via ISF compared with
controls (Figure 2E). We found 48% and 65% lower plasma levels of Ab40 and Ab42, respectively, in
Picalm
+/-
mice compared with control Picalm
+/+
mice (Figure 2F), confirming impaired Ab clearance from
brain to blood.

Effects of Picalm deficiency and endothelial-specific rescue in APP
sw/0
mice
To address whether Picalm deficiency can influence Ab pathology, we crossed transgenic APP
sw/0

mice, which develop Ab elevation and correlative memory deficits (Hsiao et al., 1996), with Picalm
+/-
mice.
PICALM expression in APP
sw/0
; Picalm
+/-
mice was reduced by ~70% in microvessels and ~50% in capillary-
depleted brain compared with their littermate controls (Supplementary Figure 3A, B). Using hippocampal in
vivo microdialysis (Sagare et al., 2013) (Online Methods), we found a substantial 2.4- and 2.5-fold increase
83

in the steady-state levels of soluble Ab40 and Ab42 in brain ISF of 3-month-old APP
sw/0
; Picalm
+/-
mice
compared with age-matched littermate controls, respectively (Figure 3A, B). After intraperitoneal injection
of the g-secretase inhibitor Compound E, the half-life of Ab40 and Ab42 in brain ISF was increased in
APP
sw/0
; Picalm
+/-
mice compared with APP
sw/0
; Picalm
+/-
controls from 1.2 to 1.9 h and 1.4 to 2.5 h,
respectively (Figure 3C), suggesting that the increase of ISF Ab levels was a result of diminished Ab
clearance. The increase in Ab ISF levels preceded Ab and amyloid deposition. Deposits were absent in 3-
month-old APP
sw/0
; Picalm
+/-
mice (Figure 3D), but began to accumulate at 6 months of age, as shown by
the increase in Ab load in the hippocampus and cortex and accelerated development of cerebral amyloid
angiopathy (Supplementary Figure 3C, D). Notably, 9-month-old APP
sw/0
; Picalm
+/-
mice had a substantial
3.5-4-fold increase in Ab load in the cortex and hippocampus (Figure 3E, F), which was associated with
worse performance in behavioral tests, including nest construction, burrowing, novel object location and
novel object recognition (Figure 3G-I), compared with age-matched littermate controls.

To specifically address the role of endothelial PICALM in transvascular Ab clearance, we performed
a rescue experiment. We generated APP
sw/0
; Picalm
+/-
; Tie2-Cre mice and an adeno-associated viral (AAV)-
Flex-Picalm construct for delivery of a Cre-dependent expression cassette (Flex-Picalm) specifically to brain
endothelium in the hippocampus (Figure 4A). After administration of AAV-Flex-Tdtomato into the
84

hippocampus of 5-month-old APP
sw/0
; Picalm
+/-
; Tie2-Cre mice, more than 50% of lectin-positive endothelial
vascular profiles expressed Tie2-Cre-dependent Tdtomato (Supplementary Figure 4A). Co-injection of
AAV-Flex-Tdtomato and AAV-Syn1-GFP revealed that less than 3% of hippocampal neurons expressed
Tdtomato. Together, these data confirm Tie2-Cre-dependent endothelial-specific expression of transgene
and minimal leakage in neurons. Similar results were obtained after administration of AAV-Flex-Picalm into
the hippocampus of 5-month-old APP
sw/0
; Picalm
+/-
; Tie2-Cre mice with Tie2-Cre-dependent Flag-PICALM
expression in more than 50% lectin-positive endothelial vascular profiles, and we observed negligible
expression in neurons after co-injection of AAV-Flex-Picalm and AAV-Syn1-GFP (Figure 4B). PICALM re-
expression in endothelium of APP
sw/0
; Picalm
+/-
; Tie2-Cre mice after AAV-Flex-Picalm administration in the
ipsilateral hippocampus diminished Ab load, Ab40 and Ab42 levels by 64%, 46%, and 37%, respectively,
compared with the contralateral hippocampus injected with AAV-Flex (control) virus (Figure 4C-E and
Supplementary Figure 4B). Moreover, bilateral administration of AAV-Flex-Picalm in the left and right
hippocampus improve behavior in APP
sw/0
; Picalm
+/-
; Tie2-Cre mice compared with AAV-Flex (control), as
shown in a separate group of mice (Figure 4F, G). These data strongly support our hypothesis that brain
endothelial PICALM has a central role in regulating Ab clearance from the brain by controlling its efflux at
the BBB.
Neither Picalm
+/-
nor APP
sw/0
; Picalm
+/-
mice showed changes in Ab production and processing,
brain microvascular expression of major Ab transporters, including P-glycoprotein, LRP1 and RAGE, or Ab-
degrading enzymes, including neprilysin and insulin-degrading enzyme, compared with their respective
controls (Supplementary Figure 5), which make these mechanisms unlikely contributors to decreased Ab
clearance and/or increased Ab accumulation.
PICALM/clathrin-dependent endocytosis of Ab-LRP1 complex by endothelial cells
To elucidate the molecular mechanism(s) underlying PICALM regulation of Ab clearance across
the BBB, we studied Ab internalization and trafficking in primary human brain endothelial cells (BECs) and
in an in vitro model of the BBB (Zhu et al., 2010). We found that Ab40 bound to the cell surface LRP1 in
BECs at 4 °C, as previously reported (Deane et al., 2004; Zlokovic, 2011) (Supplementary Figure 6A,
85

Supplementary Table 1 and Online Methods), and that 6-carboxyfluorescein-labeled (FAM)- Ab40-LRP1
complex colocalized rapidly (<30 s) with PICALM (Figure 5A) and proteins responsible for PICALM/clathrin-
dependent internalization of ligands (Tebar et al., 1999; Sorkin and von Zastrow, 2009) at 37 °C, including
the clathrin heavy chain (CHC; Figure 5B-D) and clathrin adaptor protein a-adaptin (AP-2; Supplementary
Figure 6B). Similar to Ab40, FAM-Ab42, but not scrambled FAM-Ab42 (Supplementary Figure 6C, D), led
to a rapid increase in colocalized Ab-LRP1/PICALM puncta. Rapid recruitment of PICALM and the clathrin
endocytic apparatus to LRP1 was confirmed by co-immunoprecipitation analysis 30 s after Ab40 treatment
at 1 nM, a concentration corresponding to Ab40 levels in the CSF (Querfurth and LaFerla, 2010; Zlokovic,
2011) (Figure 5E). In contrast with PICALM, which remains associated with LRP1 over longer periods of
time, CHC and AP-2 dissociated early from LRP1 (Figure 5E), consistent with rapid uncoating of clathrin
from internalized vesicles (Sorkin and von Zastrow, 2009). siRNA inhibition of PICALM or CHC inhibited
Ab40-induced LRP1 internalization (Figure 5F), suggesting that endocytosis of Ab-LRP1 complex requires
both PICALM and clathrin.

To further understand PICALM interaction with LRP1, we studied binding of human recombinant
PICALM to a human recombinant glutathione S-transferase (GST)-tagged LRP1 C terminus fusion protein,
86

which indicated direct binding of PICALM to LRP1 C terminus (Figure 5G), but not to GST (Supplementary
Figure 6E). Although this experiment was carried out in the absence of cells, we cannot rule out the
possibility that some intermediary proteins in the cellular milieu can facilitate and/or influence PICALM
binding to LRP1. Using co-immunoprecipitation analysis after transfection of cells with PICALM and various
mutants of the C-terminus of LRP1 cytoplasmic tail (Li et al., 2001), we found that PICALM binding to LRP1
required the YXXL motif (Figure 5H, I).

87

To determine the specificity of Ab as a ligand that enhances the binding of PICALM to the
cytoplasmic tail of LRP1, we studied other LRP1 ligands, including apolipoprotein E (apoE) and activated
a2-macroglobulin (a2M*). Incubation of BECs with astrocyte-derived lapidated apoE3, apoE4 (Bell et al.,
2012) and a2M* did not result in a rapid binding of PICALM to LRP1, whereas Ab40 binding to apoE3,
apoE4, and a2M* inhibited PICALM binding to LRP1 (Supplementary Figure 7A). Consistent with these
data, binding of Ab40 to apoE4, apoE3, and a2M* inhibited Ab40-induced internalization of LRP1
(Supplementary Figure 7B, C). Collectively, these data suggest that binding of Ab to the ectodomain of
LRP1 has a unique conformational effect on its cytoplasmic C-terminus tail, enhancing the binding of
PICALM, which initiates PICALM/clathrin-dependent endocytosis of Ab-LRP1 complex.

PICALM associates with LRP1 during Ab transcytosis across an endothelial monolayer
We next used an in vitro model of the BBB to study the role of PICALM in Ab transport across fully
confluent (>97%) human brain endothelial monolayer co-cultured with pericyte-conditioned medium, as
pericytes critically influence the BBB properties (Armulik et al., 2010; Bell et al., 2010; Daneman et al.,
2010). The endothelial monolayer had a typical cobblestone pattern on the zonula occludens (ZO-1) tight
junction protein and a cortical distribution of the F-actin cytoskeleton and expressed PICALM (Figure 6A).
A transmonolayer electrical resistance (TEER) of ~280 W cm
2
and a low paracellular permeability constant
88

(P) of ~1.21 x 10
-6
cm s
-1
determined for dextran (molecular weight, 40 kDa), a metabolically inert polar
molecule, confirmed formation of the barrier.

To ascertain whether our endothelial monolayer model is suitable for Ab transport studies, we
determined cell polarity for two key Ab transporters, LRP1, and the receptor for advanced glycation end
products (RAGE), which are expressed in brain endothelium in vivo mainly at the abluminal (basolateral)
and luminal (apical) sides, respectively, as confirmed by signal intensity profile analysis across capillary
lumens in human brain (Supplementary Figure 8A, B). Consistent with in vivo findings, LRP1 and RAGE
chiefly localized toward the basolateral and apical membrane of the endothelial monolayer, respectively
89

(Supplementary Figure 8A, B). PICALM associated with LRP1-Ab complex in the endothelial monolayer
rapidly after Ab application to the basolateral membrane (Figure 6B), as it did in BEC cultures (Figure 5A,
D, E).
The proximity ligation assay (PLA) (Winkler et al., 2010; Bell et al., 2012) confirmed that PICALM
associates rapidly with LRP1 in the endothelial monolayer after Ab40 (1 nM) addition to the basolateral
membrane with a peak at 30 s and a plateau over 4 min (Figure 6C, D). In contrast, LRP1-clathrin
association peaked in 30 s and was followed by a sharp decline in 1 min (Figure 6D), consistent with co-
immunoprecipitation data showing rapid dissociation of clathrin from internalized LRP1 (Figure 5E).

To further study the roles of LRP1, PICALM, and CHC in Ab internalization at the basolateral
membrane and the basolateral-to-apical transendothelial transport (transcytosis), we used siRNA silencing.
Ab40 (1 nM) internalization was rapid, with a t1/2 of ~17 s (Figure 6E), as reported for LRP1-mediated
endocytosis (Li et al., 2001). siRNA inhibition of LRP1, PICALM, or CHC substantially diminished Ab
internalization. Basolateral-to-apical transcytosis of Ab40 (1 nM, determined by Ab40 ELISA measurements
90

in the basolateral and apical chambers corrected for paracellular diffusion by subtracting diffusion values
of inulin, a metabolically inert polar molecule; Online Methods) was inhibited by the receptor-associated
protein (RAP) and LRP1-specific antibodies, but not by other low-density lipoprotein receptor-specific
antibodies (Supplementary Figure 8D). LRP1 siRNA, but not scrambled siRNA, also inhibited Ab40
basolateral-to-apical transendothelial transport (Supplementary Figure 8D). These data confirm that Ab
binding to LRP1 initiates Ab clearance at the basolateral membrane of the monolayer, as previously
suggested (Shibata et al., 2000; Deane et al., 2004; Zlokovic, 2011), and that LRP1 is required for Ab
transcytosis across the monolayer. As expected, silencing PICALM and CHC inhibited Ab40 transcytosis
across the monolayer by approximately 85%, as shown within 30 min of the unidirectional basolateral-to-
apical transendothelial Ab transport (Figure 6F, G). Within 5 min of Ab application to the basolateral
membrane of the endothelial membrane, Ab-LRP1 internalized complex was already found near the apical
membrane of the monolayer (Supplementary Figure 9A), confirming basolateral-to-apical transendothelial
transport.
PICALM guides Ab trafficking to Rab5 and Rab11 leading to Ab endothelial transcytosis
Given that PICALM remains associated with LRP1 over longer periods of time in contrast to CHC
and AP-2 that dissociate rapidly from internalized Ab-LRP1 vesicles (Figures 5E and 6D), we next traced
intracellular trafficking of PICALM and Ab-LRP1 using molecular markers for different steps of the
endosomal pathway. Because Rab5 and Rab7 regulate Ab endosomal trafficking in neurons (Kanekiyo et
al., 2013), we asked whether PICALM interacts with Rab GTPases (Stenmark, 2009) following Ab-LRP1
internalization. Using subconfluent endothelial cultures, we found that PICALM and Ab-LRP1 complex
colocalize with Rab5- and EEA1-positive early endosomes (Sorkin and von Zastrow, 2009) within 2 min of
Ab40 (1 nM) treatment (Figure 7A and Supplementary Figure 10A, B), but not with Rab7, a GTPase that
directs fusion of late endosomes with lysosomes (Stenmark, 2009) leading to degradation of ligands (Li et
al., 2012) (Figure 7B) or the lysosomal-specific marker LAMP1 (Supplementary Figure 10C) (Stenmark,
2009). Rather, PICALM colocalized with Rab11 (Figure 7C), a GTPase that regulates recycling of vesicles
controlling transcytosis (Xu et al., 2011; Lapierre et al., 2012; Yui et al., 2013) and exocytosis (Takahashi
et al., 2012) of ligands. Ab42 also colocalized with Rab5 and Rab11 (Supplementary Figure 11A, B). In the
91

absence of Ab, PICALM minimally colocalized with Rab5, Rab7, and/or Rab11 (Supplementary Figure 11C-
E). These data suggest that PICALM likely regulates transendothelial Ab trafficking by shunting Ab away
from a degradation pathway toward a transcytotic pathway, consistent with previously demonstrated
clathrin-independent function of PICALM (Vecchi et al., 2001; Miller et al., 2011; Tian et al., 2013). Notably,
the peak colocalization between PICALM and Rab5 (Figure 7A, D) was somewhat higher than the peak
colocalization between PICALM and Rab11 (Figure 7C, D), suggesting that transfer of PICALM-containing
endocytic vesicles from Rab5 to Rab11 could potentially be a rate-limiting step in Ab transcytosis.
Using an in vitro endothelial monolayer of the BBB, we next found that FAM-Ab associated with
Rab5 and Rab11 in the endothelium within 2 and 4 min of its application to the basolateral membrane,
respectively (Figure 7E, F). PICALM-Rab11 association peaked at 4 min of incubation with Ab40 at the
basolateral membrane and remained at a plateau at 5 min (Figure 7G, H), whereas PICALM-Rab5
association peaked at 2 min and declined in 3-4 min (Figure 7H), as shown by PLA analysis. PICALM-Rab7
association was undetectable (Figure 7H). PICALM interaction with LRP1, Rab5, and Rab11 at different
time point after Ab40 treatment has been corroborated by co-immunoprecipitation analysis (Figure 7I).
Consistent with findings showing that Rab5 is necessary for the biogenesis of early and late
endosomes (Zeigerer et al., 2012), dominant-negative Rab5-S34N mutant (Online Methods) inhibited Ab
transcytosis by more than 85% as compared with EGFP control (Figure 7J, K). Dominant-negative Rab11-
S25N mutant, but not Rab7-T22N mutant, also inhibited unidirectional Ab40 basolateral-to-apical transport
by 75% (Figure 7J, K). These data suggest that Rab11 likely controls later stages of Ab transcytosis across
the monolayer, consistent with findings showing that Rab11 regulates trafficking of vesicles controlling
transcytosis (Xu et al., 2011; Lapierre et al., 2012; Yui et al., 2013) and exocytosis (Takahashi et al., 2012)
of ligands. siRNA inhibition of Rab11b, but not Rab11a, blocked Ab transcytosis (Supplementary Figure
9B), consistent with findings that Rab11b is the major Rab11 isoform in brain endothelium, whereas Rab11a
is a major isoform in epithelial cells (Xu et al., 2011; Lapierre et al., 2012; Yui et al., 2013), as we confirmed
(Supplementary Figure 9C).
92

Notably, siRNA knockdown of PICALM inhibited Rab5 and Rab11 GTPase activity in the
endothelial monolayers treated with Ab40 by 80% and 95%, respectively (Figure 7L, M), indicating that
PICALM binding to Rab5 and Rab11 is critical for maintaining Rab5 and Rab11 GTPase activity during
endosomal trafficking of Ab. Consistent with findings showing that PICALM does not direct Ab to lysosomes
for degradation, Ab was more than 95% intact within 30 min of unidirectional transcytosis and clearance
across the monolayer (Supplementary Figure 9D), as reported previously (Deane et al., 2004). Collectively,
our data indicate that PICALM controls the transcytotic pathway mediating Ab clearance across the BBB.
PICALM levels and Ab clearance by endothelial monolayers from AD patients
To determine the role of PICALM in Ab clearance in AD brain endothelium (Supplementary Table
1B), we studied Ab transcytosis across AD-derived endothelial monolayers co-cultured with pericyte-
conditioned media. The AD endothelial monolayers had a normal cobblestone pattern, TEER values and
paracellular permeability (Supplementary Figure 12A-C). Consistent with diminished PICALM levels in AD
brain endothelium in situ and in brain capillaries (Figure 1), PICALM mRNA and protein levels were reduced
by 34-35% in cultured AD BEC (Figure 8A), resulting in ~50% diminished basolateral-to-apical transcytosis
of Ab across AD-derived endothelial monolayers compared with age-matched control (Figure 8B). LRP1
levels were also reduced in AD brain endothelium (Shibata et al., 2000; Deane et al., 2004), whereas AP-
2, CHC, Rab5 and Rab11b expression was comparable with age-matched controls (data not shown).
Adenoviral-mediated transfer of PICALM compared GFP control (Supplementary Figure 12D) substantially
improved Ab transcytosis by 63%, whereas co-transfer of PICALM and LRP1 mini-gene (Bell et al., 2012)
improved Ab transport by 89% (Figure 8B), suggesting that PICALM can be therapeutically targeted.
Ab clearance by iPSC-derived endothelium carrying the rs3851179 PICALM variants
All AD-associated single-nucleotide polymorphisms (SNPs) in PICALM are located upstream of the
coding region of the gene, and no mutation in the PICALM protein has been identified to influence AD risk
(Harold et al., 2009; Lambert et al., 2009; Tanzi, 2012; Lambert et al., 2013; Carrasquillo et al., 2014).
Some AD-associated PICALM SNPs have been suggested, however, to influence PICALM expression, for
example, rs659023 variants in peripheral blood mononuclear cells (Raj et al., 2012). To address whether
93

certain PICALM variants can influence PICALM expression and Ab clearance in endothelial cells, we
focused on highly validated and replicated rs3851179 PICALM variants who rs3851179
A
allele is associated
with a lower risk of AD than the rs3851179
G
allele (Harold et al., 2009; Lambert et al., 2009; Lambert et al.,
2013). To test the phenotypic differences between the minor protective rs3851179
A
allele and the major
non-protective rs3851179
G
allele in human endothelial cells, we used inducible pluripotent stem cells
(iPSCs). The iPSCs were generated from lymphoblasts from Coriell using the episomal plasmids and
encoding factors as we described previously (Okita et al., 2011). We next used CRISPR/Cas9 genome
editing (Mali et al., 2013) to generate the isogenic homozygous iPSC lines for the two homozygous allelic
variants (Figure 8C-E) followed by direct differentiation to generate bona fide endothelial cells (Adams et
al., 2013) and endothelial monolayers (Figure 8F, G and Online Methods). iPSC-derived endothelial cells
carrying the protective rs3851179
A
allele had 72-78% higher expression levels of PICALM mRNA and
protein (Figure 8H) and 120% higher Ab clearance (Figure 8I) than the non-protective rs3851179
G
allele.
DISCUSSION
We found that PICALM reductions in brain endothelium in AD correlated with Ab and AD
neuropathology and cognitive impairment whereas reduction and re-expression of PICALM in endothelial
cells influenced Ab clearance at the BBB, Ab deposition in the mouse brain and phenotypic manifestations
of behavior in mice. At a molecular level, using an in vitro endothelial monolayer that approximates BBB,
we found that Ab binding to the ectodomain of LRP1 enhanced the binding of PICALM, which initiated
PICALM/clathrin-dependent endocytosis of Ab-LRP1 complex, and PICALM remained associated with
LRP1 after Ab internalization and directed Ab trafficking to Rab5 and Rab11, leading to Ab transcytosis
(Supplemental Figure 13). PICALM levels and Ab clearance was greatly reduced in AD-derived endothelial
monolayers, which was reversible by adenoviral-mediated PICALM transfer. Using iPSC-derived human
endothelial cells carrying rs3851179 PICALM variants, we found that the protective rs3851179 allele led to
a higher PICALM expression and enhanced Ab clearance by endothelial cells.
AAV8 viral-mediated silencing of neuronal PICALM in the hippocampus moderately reduced Ab
production in APP/PS1 mice (Xiao et al., 2012), but whether a moderate reduction in Ab load, as we
94

observed, is beneficial for neurons after inactivation of PICALM is unclear, as PICALM also protects
neurons against Ab toxicity (Treusch et al., 2011). Mice with PICALM loss-of-function allele developed
accelerated Ab accumulation and had lower behavioral test performance than controls, raising the
possibility that, in models of global Picalm deficiency, where PICALM is deleted in multiple cell populations
(for example, endothelial cells, neurons), the effect of Picalm loss from endothelium causing faulty Ab
vascular clearance may override the discrepant reduction in Ab secretion from neurons (Xiao et al., 2012).
Future studies in transgenic mice with inducible PICALM deletion in neurons should be carried out to
address its Ab-independent effects, particularly as PICALM regulates axonal growth (Bushlin et al., 2008)
and turnover of synaptic vesicles and receptors (Harel et al., 2011).
Our data revealed that PICALM guides intracellular trafficking of Ab-LRP1 to Rab5 and Rab11,
leading to Ab transcytosis across endothelial monolayer, which does not require clathrin beyond the early
internalization steps. These findings are consistent with previously demonstrated clathrin-independent
functions of PICALM in trafficking of endocytic proteins (Vecchi et al., 2001; Miller et al., 2011; Tian et al.,
2013).
At present, little is known about upstream regulators of PICALM expression. Using APP-
overexpressing mice and brain endothelial cultures, we found that elevated Ab does not suppress PICALM
in endothelium. Multiple factors in AD including turbulent capillary flow, oxidant stress, hypoxia, or
inflammation can affect gene expression in endothelial cells (Zlokovic, 2011), potentially leading to PICALM
reductions in endothelium. Identifying inhibitors of PICALM expression in AD endothelium will require future
studies interrogating multiple genetic, environmental, and local brain factors.
Consistent with the idea that some AD-associated PICALM SNPs may influence PICALM
expression (Raj et al., 2012; Parikh et al., 2014), using iPSCs, we found that the protective rs3851179 allele
had a major effect on PICALM expression and Ab clearance by endothelial cells. Future studies correlating
PICALM levels in endothelium in human brain tissue with the genotype of patients in large cohorts of control
and AD patients could provide additional information as to how different PICALM SNPs relate to each other
and/or to other genes that influence Ab clearance, as for example apoE or clusterin (Querfurth and LaFerla,
95

2010), and interrogate the roles of vascular risk factors, environment, and lifestyle. Collectively, our findings
suggest that PICALM controls Ab transport across the BBB and clearance from brain, and is therefore an
important therapeutic target for Ab clearance therapy.
REFERENCES:
Adams RH, Alitalo K (2007) Molecular regulation of angiogenesis and lymphangiogenesis. Nat Rev
Mol Cell Biol 8:464-478.
Adams WJ, Zhang Y, Cloutier J, Kuchimanchi P, Newton G, Sehrawat S, Aird WC, Mayadas TN,
Luscinskas FW, Garcia-Cardena G (2013) Functional Vascular Endothelium Derived from
Human Induced Pluripotent Stem Cells. Stem Cell Reports 1:105-113.
Aksoy D, Bammer R, Mlynash M, Venkatasubramanian C, Eyngorn I, Snider RW, Gupta SN,
Narayana R, Fischbein N, Wijman CA (2013) Magnetic resonance imaging profile of blood-
brain barrier injury in patients with acute intracerebral hemorrhage. J Am Heart Assoc
2:e000161.
Alberts JL, Linder SM, Penko AL, Lowe MJ, Phillips M (2011) It is not about the bike, it is about the
pedaling: forced exercise and Parkinson's disease. Exerc Sport Sci Rev 39:177-186.
Allen A, Messier C (2013) Plastic changes in the astrocyte GLUT1 glucose transporter and beta-
tubulin microtubule protein following voluntary exercise in mice. Behav Brain Res 240:95-
102.
Ando K, Brion J-P, Stygelbout V, Suain V, Authelet M, Dedecker R, Chanut A, Lacor P, Lavaur J,
Sazdovitch V, Rogaeva E, Potier M-C, Duyckaerts C (2013) Clathrin adaptor CALM/PICALM
is associated with neurofibrillary tangles and is cleaved in Alzheimer's brains. Acta
Neuropathologica 125:861-878.
Apostolova LG, Hwang KS, Andrawis JP, Green AE, Babakchanian S, Morra JH, Cummings JL, Toga
AW, Trojanowski JQ, Shaw LM, Jack CR, Petersen RC, Aisen PS, Jagust WJ, Koeppe RA,
Mathis CA, Weiner MW, Thompson PM, Initiative AsDN (2010) 3D PIB and CSF biomarker
associations with hippocampal atrophy in ADNI subjects. Neurobiol Aging 31:1284-1303.
Armulik A, Genové G, Mäe M, Nisancioglu MH, Wallgard E, Niaudet C, He L, Norlin J, Lindblom P,
Strittmatter K, Johansson BR, Betsholtz C (2010) Pericytes regulate the blood-brain
barrier. Nature 468:557-561.
Attwell D, Laughlin SB (2001) An energy budget for signaling in the grey matter of the brain. J
Cereb Blood Flow Metab 21:1133-1145.
Baig S, Joseph SA, Tayler H, Abraham R, Owen MJ, Williams J, Kehoe PG, Love S (2010) Distribution
and Expression of Picalm in Alzheimer Disease. Journal of Neuropathology and
Experimental Neurology 69:1071-1077.
Baloyannis SJ, Baloyannis IS (2012) The vascular factor in Alzheimer's disease: a study in Golgi
technique and electron microscopy. J Neurol Sci 322:117-121.
Barros LF (2013) Metabolic signaling by lactate in the brain. Trends Neurosci 36:396-404.
Beaulieu JM, Gainetdinov RR (2011) The physiology, signaling, and pharmacology of dopamine
receptors. Pharmacol Rev 63:182-217.
96

Bednarczyk MR, Hacker LC, Fortin-Nunez S, Aumont A, Bergeron R, Fernandes KJ (2011) Distinct
stages of adult hippocampal neurogenesis are regulated by running and the running
environment. Hippocampus 21:1334-1347.
Beeler JA, Cao ZF, Kheirbek MA, Ding Y, Koranda J, Murakami M, Kang UJ, Zhuang X (2010)
Dopamine-dependent motor learning: insight into levodopa's long-duration response.
Ann Neurol 67:639-647.
Beeler JA, Frank MJ, McDaid J, Alexander E, Turkson S, Bernardez Sarria MS, Bernandez MS,
McGehee DS, Zhuang X (2012) A role for dopamine-mediated learning in the
pathophysiology and treatment of Parkinson's disease. Cell Rep 2:1747-1761.
Bell RD, Winkler EA, Sagare AP, Singh I, LaRue B, Deane R, Zlokovic BV (2010) Pericytes control
key neurovascular functions and neuronal phenotype in the adult brain and during brain
aging. Neuron 68:409-427.
Bell RD, Deane R, Chow N, Long X, Sagare A, Singh I, Streb JW, Guo H, Rubio A, Van Nostrand W,
Miano JM, Zlokovic BV (2009) SRF and myocardin regulate LRP-mediated amyloid-beta
clearance in brain vascular cells. Nat Cell Biol 11:143-153.
Bell RD, Winkler EA, Singh I, Sagare AP, Deane R, Wu Z, Holtzman DM, Betsholtz C, Armulik A,
Sallstrom J, Berk BC, Zlokovic BV (2012) Apolipoprotein E controls cerebrovascular
integrity via cyclophilin A. Nature 485:512-516.
Bernardi C, Tramontina AC, Nardin P, Biasibetti R, Costa AP, Vizueti AF, Batassini C, Tortorelli LS,
Wartchow KM, Dutra MF, Bobermin L, Sesterheim P, Quincozes-Santos A, de Souza J,
Gonçalves CA (2013) Treadmill exercise induces hippocampal astroglial alterations in rats.
Neural Plast 2013:709732.
Berti V, Polito C, Ramat S, Vanzi E, De Cristofaro MT, Pellicanò G, Mungai F, Marini P, Formiconi
AR, Sorbi S, Pupi A (2010) Brain metabolic correlates of dopaminergic degeneration in de
novo idiopathic Parkinson's disease. Eur J Nucl Med Mol Imaging 37:537-544.
Bittar PG, Charnay Y, Pellerin L, Bouras C, Magistretti PJ (1996) Selective distribution of lactate
dehydrogenase isoenzymes in neurons and astrocytes of human brain. J Cereb Blood Flow
Metab 16:1079-1089.
Black JE, Isaacs KR, Anderson BJ, Alcantara AA, Greenough WT (1990) Learning causes
synaptogenesis, whereas motor activity causes angiogenesis, in cerebellar cortex of adult
rats. Proc Natl Acad Sci U S A 87:5568-5572.
Blennow K, Wallin A, Fredman P, Karlsson I, Gottfries CG, Svennerholm L (1990) Blood-brain
barrier disturbance in patients with Alzheimer's disease is related to vascular factors. Acta
Neurol Scand 81:323-326.
Bliss TV, Cooke SF (2011) Long-term potentiation and long-term depression: a clinical
perspective. Clinics (Sao Paulo) 66 Suppl 1:3-17.
Blomstrand E, Perrett D, Parry-Billings M, Newsholme EA (1989) Effect of sustained exercise on
plasma amino acid concentrations and on 5-hydroxytryptamine metabolism in six
different brain regions in the rat. Acta Physiol Scand 136:473-481.
Boumezbeur F, Petersen KF, Cline GW, Mason GF, Behar KL, Shulman GI, Rothman DL (2010) The
contribution of blood lactate to brain energy metabolism in humans measured by
dynamic 13C nuclear magnetic resonance spectroscopy. J Neurosci 30:13983-13991.
Bowman GL, Kaye JA, Quinn JF (2012) Dyslipidemia and blood-brain barrier integrity in
Alzheimer's disease. Curr Gerontol Geriatr Res 2012:184042.
97

Braak H, Braak E, Bohl J (1993) Staging of Alzheimer-related cortical destruction. Eur Neurol
33:403-408.
Brown BS, Payne T, Kim C, Moore G, Krebs P, Martin W (1979) Chronic response of rat brain
norepinephrine and serotonin levels to endurance training. J Appl Physiol Respir Environ
Exerc Physiol 46:19-23.
Bruick RK, McKnight SL (2001) A conserved family of prolyl-4-hydroxylases that modify HIF.
Science 294:1337-1340.
Bushlin I, Petralia RS, Wu F, Harel A, Mughal MR, Mattson MP, Yao PJ (2008) Clathrin Assembly
Protein AP180 and CALM Differentially Control Axogenesis and Dendrite Outgrowth in
Embryonic Hippocampal Neurons. Journal of Neuroscience 28:10257-10271.
Calabresi P, Mercuri NB, Sancesario G, Bernardi G (1993) Electrophysiology of dopamine-
denervated striatal neurons. Implications for Parkinson's disease. Brain 116 ( Pt 2):433-
452.
Carbon M, Eidelberg D (2006) Functional imaging of sequence learning in Parkinson's disease. J
Neurol Sci 248:72-77.
Carrasquillo MM, Khan QuA, Murray ME, Krishnan S, Aakre J, Pankratz VS, Thuy N, Ma L, Bisceglio
G, Petersen RC, Younkin SG, Dickson DW, Boeve BF, Graff-Radford NR, Ertekin-Taner N
(2014) Late-onset Alzheimer disease genetic variants in posterior cortical atrophy and
posterior AD. Neurology 82:1455-1462.
Catlow BJ, Rowe AR, Clearwater CR, Mamcarz M, Arendash GW, Sanchez-Ramos J (2009) Effects
of environmental enrichment and physical activity on neurogenesis in transgenic PS1/APP
mice. Brain Res 1256:173-179.
Cazorla M, Shegda M, Ramesh B, Harrison NL, Kellendonk C (2012) Striatal D2 receptors regulate
dendritic morphology of medium spiny neurons via Kir2 channels. J Neurosci 32:2398-
2409.
Chen H, Zhang SM, Schwarzschild MA, Hernán MA, Ascherio A (2005) Physical activity and the
risk of Parkinson disease. Neurology 64:664-669.
Chowdhury R, Hardy A, Schofield CJ (2008) The human oxygen sensing machinery and its
manipulation. Chem Soc Rev 37:1308-1319.
Clark PJ, Bhattacharya TK, Miller DS, Kohman RA, DeYoung EK, Rhodes JS (2012) New neurons
generated from running are broadly recruited into neuronal activation associated with
three different hippocampus-involved tasks. Hippocampus 22:1860-1867.
Correia SC, Moreira PI (2010) Hypoxia-inducible factor 1: a new hope to counteract
neurodegeneration? J Neurochem 112:1-12.
Correia SC, Cardoso S, Santos RX, Carvalho C, Santos MS, Perry G, Smith MA, Moreira PI (2011)
New insights into the mechanisms of mitochondrial preconditioning-triggered
neuroprotection. Curr Pharm Des 17:3381-3389.
Cortes-Canteli M, Zamolodchikov D, Ahn HJ, Strickland S, Norris EH (2012) Fibrinogen and altered
hemostasis in Alzheimer's disease. J Alzheimers Dis 32:599-608.
Cotman CW, Berchtold NC, Christie LA (2007) Exercise builds brain health: key roles of growth
factor cascades and inflammation. Trends Neurosci 30:464-472.
Courtière A, Hardouin J, Burle B, Vidal F, Turle-Lorenzo N, Amalric M, Hasbroucq T (2011)
Dynamics of executive control and motor deficits in parkinsonian rats. J Neurosci
31:11929-11933.
98

Cullen KM, Kócsi Z, Stone J (2005) Pericapillary haem-rich deposits: evidence for
microhaemorrhages in aging human cerebral cortex. J Cereb Blood Flow Metab 25:1656-
1667.
Cunningham LA, Candelario K, Li L (2012) Roles for HIF-1α in neural stem cell function and the
regenerative response to stroke. Behav Brain Res 227:410-417.
Curlik DM, Shors TJ (2013) Training your brain: Do mental and physical (MAP) training enhance
cognition through the process of neurogenesis in the hippocampus? Neuropharmacology
64:506-514.
Daneman R, Zhou L, Kebede AA, Barres BA (2010) Pericytes are required for blood-brain barrier
integrity during embryogenesis. Nature 468:562-566.
Deane R, Wu ZH, Sagare A, Davis J, Yan SD, Hamm K, Xu F, Parisi M, LaRue B, Hu HW, Spijkers P,
Guo H, Song XM, Lenting PJ, Van Nostrand WE, Zlokovic BV (2004) LRP/amyloid beta-
peptide interaction mediates differential brain efflux of A beta isoforms. Neuron 43:333-
344.
Deane R et al. (2003) RAGE mediates amyloid-beta peptide transport across the blood-brain
barrier and accumulation in brain. Nat Med 9:907-913.
Dienel GA (2019) Brain Glucose Metabolism: Integration of Energetics with Function. Physiol Rev
99:949-1045.
Dirnberger G, Jahanshahi M (2013) Executive dysfunction in Parkinson's disease: a review. J
Neuropsychol 7:193-224.
Dreyling MH, MartinezCliment JA, Zheng M, Mao J, Rowley JD, Bohlander SK (1996) The
t(10;11)(p13;q14) in the U937 cell line results in the fusion of the AF10 gene and CALM,
encoding a new member of the AP-3 clathrin assembly protein family. Proceedings of the
National Academy of Sciences of the United States of America 93:4804-4809.
Eadie BD, Redila VA, Christie BR (2005) Voluntary exercise alters the cytoarchitecture of the adult
dentate gyrus by increasing cellular proliferation, dendritic complexity, and spine density.
J Comp Neurol 486:39-47.
Ebert BL, Firth JD, Ratcliffe PJ (1995) Hypoxia and mitochondrial inhibitors regulate expression of
glucose transporter-1 via distinct Cis-acting sequences. J Biol Chem 270:29083-29089.
Eddy MC, Stansfield KJ, Green JT (2014) Voluntary exercise improves performance of a
discrimination task through effects on the striatal dopamine system. Learn Mem 21:334-
337.
Ehninger D, Kempermann G (2003) Regional effects of wheel running and environmental
enrichment on cell genesis and microglia proliferation in the adult murine neocortex.
Cereb Cortex 13:845-851.
Elsworth JD, Leranth C, Redmond DE, Roth RH (2013) Loss of asymmetric spine synapses in
prefrontal cortex of motor-asymptomatic, dopamine-depleted, cognitively impaired
MPTP-treated monkeys. Int J Neuropsychopharmacol 16:905-912.
Engeln M (2013) Throwing some light on executive function in Parkinson's disease. Mov Disord
28:1052.
Fabel K, Tam B, Kaufer D, Baiker A, Simmons N, Kuo CJ, Palmer TD (2003) VEGF is necessary for
exercise-induced adult hippocampal neurogenesis. Eur J Neurosci 18:2803-2812.
Farkas E, Luiten PG (2001) Cerebral microvascular pathology in aging and Alzheimer's disease.
Prog Neurobiol 64:575-611.
99

Farley BG, Koshland GF (2005) Training BIG to move faster: the application of the speed-
amplitude relation as a rehabilitation strategy for people with Parkinson's disease. Exp
Brain Res 167:462-467.
Fasano C, Bourque MJ, Lapointe G, Leo D, Thibault D, Haber M, Kortleven C, Desgroseillers L,
Murai KK, Trudeau L (2013) Dopamine facilitates dendritic spine formation by cultured
striatal medium spiny neurons through both D1 and D2 dopamine receptors.
Neuropharmacology 67:432-443.
Fiala M, Liu QN, Sayre J, Pop V, Brahmandam V, Graves MC, Vinters HV (2002) Cyclooxygenase-
2-positive macrophages infiltrate the Alzheimer's disease brain and damage the blood-
brain barrier. Eur J Clin Invest 32:360-371.
Filoteo JV, Reed JD, Litvan I, Harrington DL (2014) Volumetric correlates of cognitive functioning
in nondemented patients with Parkinson's disease. Mov Disord 29:360-367.
Fisher BE, Petzinger GM, Nixon K, Hogg E, Bremmer S, Meshul CK, Jakowec MW (2004) Exercise-
induced behavioral recovery and neuroplasticity in the 1-methyl-4-phenyl-1,2,3,6-
tetrahydropyridine-lesioned mouse basal ganglia. J Neurosci Res 77:378-390.
Fisher BE, Li Q, Nacca A, Salem GJ, Song J, Yip J, Hui JS, Jakowec MW, Petzinger GM (2013)
Treadmill exercise elevates striatal dopamine D2 receptor binding potential in patients
with early Parkinson's disease. Neuroreport 24:509-514.
Fisher BE, Wu AD, Salem GJ, Song J, Lin CH, Yip J, Cen S, Gordon J, Jakowec M, Petzinger G (2008)
The effect of exercise training in improving motor performance and corticomotor
excitability in people with early Parkinson's disease. Arch Phys Med Rehabil 89:1221-
1229.
Floresco SB (2013) Prefrontal dopamine and behavioral flexibility: shifting from an "inverted-U"
toward a family of functions. Front Neurosci 7:62.
Foley TE, Fleshner M (2008) Neuroplasticity of dopamine circuits after exercise: implications for
central fatigue. Neuromolecular Med 10:67-80.
Ford MGJ, Pearse BMF, Higgins MK, Vallis Y, Owen DJ, Gibson A, Hopkins CR, Evans PR, McMahon
HT (2001) Simultaneous binding of PtdIns(4,5)P-2 and clathrin by AP180 in the nucleation
of clathrin lattices on membranes. Science 291:1051-1055.
Fox CM, Ramig LO, Ciucci MR, Sapir S, McFarland DH, Farley BG (2006) The science and practice
of LSVT/LOUD: neural plasticity-principled approach to treating individuals with Parkinson
disease and other neurological disorders. Semin Speech Lang 27:283-299.
Gandhi S, Wood-Kaczmar A, Yao Z, Plun-Favreau H, Deas E, Klupsch K, Downward J, Latchman DS,
Tabrizi SJ, Wood NW, Duchen MR, Abramov AY (2009) PINK1-associated Parkinson's
disease is caused by neuronal vulnerability to calcium-induced cell death. Mol Cell 33:627-
638.
Garcia PC, Real CC, Ferreira AF, Alouche SR, Britto LR, Pires RS (2012) Different protocols of
physical exercise produce different effects on synaptic and structural proteins in motor
areas of the rat brain. Brain Res 1456:36-48.
Garthe A, Roeder I, Kempermann G (2016) Mice in an enriched environment learn more flexibly
because of adult hippocampal neurogenesis. Hippocampus 26:261-271.
Gerhart-Hines Z, Rodgers JT, Bare O, Lerin C, Kim SH, Mostoslavsky R, Alt FW, Wu Z, Puigserver P
(2007) Metabolic control of muscle mitochondrial function and fatty acid oxidation
through SIRT1/PGC-1alpha. EMBO J 26:1913-1923.
100

Gilliam PE, Spirduso WW, Martin TP, Walters TJ, Wilcox RE, Farrar RP (1984) The effects of
exercise training on [3H]-spiperone binding in rat striatum. Pharmacol Biochem Behav
20:863-867.
Goos JD, Kester MI, Barkhof F, Klein M, Blankenstein MA, Scheltens P, van der Flier WM (2009)
Patients with Alzheimer disease with multiple microbleeds: relation with cerebrospinal
fluid biomarkers and cognition. Stroke 40:3455-3460.
Greer SN, Metcalf JL, Wang Y, Ohh M (2012) The updated biology of hypoxia-inducible factor.
EMBO J 31:2448-2460.
Hall CN, Klein-Flügge MC, Howarth C, Attwell D (2012) Oxidative phosphorylation, not glycolysis,
powers presynaptic and postsynaptic mechanisms underlying brain information
processing. J Neurosci 32:8940-8951.
Halliday MR, Pomara N, Sagare AP, Mack WJ, Frangione B, Zlokovic BV (2013) Relationship
between cyclophilin a levels and matrix metalloproteinase 9 activity in cerebrospinal fluid
of cognitively normal apolipoprotein e4 carriers and blood-brain barrier breakdown.
JAMA Neurol 70:1198-1200.
Hardy J, Selkoe DJ (2002) Medicine - The amyloid hypothesis of Alzheimer's disease: Progress and
problems on the road to therapeutics. Science 297:353-356.
Harel A, Mattson MP, Yao PJ (2011) CALM, A Clathrin Assembly Protein, Influences Cell Surface
GluR2 Abundance. Neuromolecular Medicine 13:88-90.
Harold D et al. (2009) Genome-wide association study identifies variants at CLU and PICALM
associated with Alzheimer's disease. Nature Genetics 41:1088-U1061.
Harris JJ, Jolivet R, Attwell D (2012) Synaptic energy use and supply. Neuron 75:762-777.
Hsiao K, Chapman P, Nilsen S, Eckman C, Harigaya Y, Younkin S, Yang FS, Cole G (1996) Correlative
memory deficits, A beta elevation, and amyloid plaques in transgenic mice. Science
274:99-102.
Hu XL, Bergström SA, Brink M, Rönnbäck A, Dahlqvist P (2010) Enriched environment increases
spinophilin mRNA expression and spinophilin immunoreactive dendritic spines in
hippocampus and cortex. Neurosci Lett 476:79-83.
Huganir RL, Nicoll RA (2013) AMPARs and synaptic plasticity: the last 25 years. Neuron 80:704-
717.
Hultman K, Strickland S, Norris EH (2013) The APOE ɛ4/ɛ4 genotype potentiates vascular
fibrin(ogen) deposition in amyloid-laden vessels in the brains of Alzheimer's disease
patients. J Cereb Blood Flow Metab 33:1251-1258.
Iadecola C (2004) Neurovascular regulation in the normal brain and in Alzheimer's disease. Nat
Rev Neurosci 5:347-360.
Ivan M, Kondo K, Yang H, Kim W, Valiando J, Ohh M, Salic A, Asara JM, Lane WS, Kaelin WG (2001)
HIFalpha targeted for VHL-mediated destruction by proline hydroxylation: implications
for O2 sensing. Science 292:464-468.
Jakowec MW, Petzinger GM (2004) 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned
model of parkinson's disease, with emphasis on mice and nonhuman primates. Comp
Med 54:497-513.
Jiang BH, Semenza GL, Bauer C, Marti HH (1996) Hypoxia-inducible factor 1 levels vary
exponentially over a physiologically relevant range of O2 tension. Am J Physiol 271:C1172-
1180.
101

Kaasinen V, Nurmi E, Brück A, Eskola O, Bergman J, Solin O, Rinne JO (2001) Increased frontal
[(18)F]fluorodopa uptake in early Parkinson's disease: sex differences in the prefrontal
cortex. Brain 124:1125-1130.
Kanatsu K, Morohashi Y, Suzuki M, Kuroda H, Watanabe T, Tomita T, Iwatsubo T (2014) Decreased
CALM expression reduces A beta 42 to total A beta ratio through clathrin-mediated
endocytosis of gamma-secretase. Nature Communications 5.
Kanekiyo T, Liu C-C, Shinohara M, Li J, Bu G (2012) LRP1 in Brain Vascular Smooth Muscle Cells
Mediates Local Clearance of Alzheimer's Amyloid-beta. Journal of Neuroscience
32:16458-16465.
Kanekiyo T, Cirrito JR, Liu C-C, Shinohara M, Li J, Schuler DR, Shinohara M, Holtzman DM, Bu G
(2013) Neuronal Clearance of Amyloid-beta by Endocytic Receptor LRP1. Journal of
Neuroscience 33:19276-19283.
Kasivisvanathan V, Shalhoub J, Lim CS, Shepherd AC, Thapar A, Davies AH (2011) Hypoxia-
inducible factor-1 in arterial disease: a putative therapeutic target. Curr Vasc Pharmacol
9:333-349.
Kempermann G, Gast D, Gage FH (2002) Neuroplasticity in old age: sustained fivefold induction
of hippocampal neurogenesis by long-term environmental enrichment. Ann Neurol
52:135-143.
Khatri N, Man HY (2013) Synaptic activity and bioenergy homeostasis: implications in brain
trauma and neurodegenerative diseases. Front Neurol 4:199.
Kim YM, Ji ES, Kim SH, Kim TW, Ko IG, Jin JJ, Kim CJ, Kim DH (2015) Treadmill exercise improves
short-term memory by enhancing hippocampal cell proliferation in quinolinic acid-
induced Huntington's disease rats. J Exerc Rehabil 11:5-11.
Kinni H, Guo M, Ding JY, Konakondla S, Dornbos D, Tran R, Guthikonda M, Ding Y (2011) Cerebral
metabolism after forced or voluntary physical exercise. Brain Res 1388:48-55.
Kintz N, Petzinger GM, Akopian G, Ptasnik S, Williams C, Jakowec MW, Walsh JP (2013) Exercise
modifies α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor expression in
striatopallidal neurons in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned
mouse. J Neurosci Res 91:1492-1507.
Kirchner L, Chen WQ, Afjehi-Sadat L, Viidik A, Skalicky M, Höger H, Lubec G (2008) Hippocampal
metabolic proteins are modulated in voluntary and treadmill exercise rats. Exp Neurol
212:145-151.
Klaissle P, Lesemann A, Huehnchen P, Hermann A, Storch A, Steiner B (2012) Physical activity and
environmental enrichment regulate the generation of neural precursors in the adult
mouse substantia nigra in a dopamine-dependent manner. BMC Neurosci 13:132.
Klaus F, Hauser T, Slomianka L, Lipp HP, Amrein I (2009) A reward increases running-wheel
performance without changing cell proliferation, neuronal differentiation or cell death in
the dentate gyrus of C57BL/6 mice. Behav Brain Res 204:175-181.
Kleim JA, Jones TA (2008) Principles of experience-dependent neural plasticity: implications for
rehabilitation after brain damage. J Speech Lang Hear Res 51:S225-239.
Kleim JA, Barbay S, Cooper NR, Hogg TM, Reidel CN, Remple MS, Nudo RJ (2002) Motor learning-
dependent synaptogenesis is localized to functionally reorganized motor cortex.
Neurobiol Learn Mem 77:63-77.
102

Kohl Z, Kandasamy M, Winner B, Aigner R, Gross C, Couillard-Despres S, Bogdahn U, Aigner L,
Winkler J (2007) Physical activity fails to rescue hippocampal neurogenesis deficits in the
R6/2 mouse model of Huntington's disease. Brain Res 1155:24-33.
Kohman RA, DeYoung EK, Bhattacharya TK, Peterson LN, Rhodes JS (2012) Wheel running
attenuates microglia proliferation and increases expression of a proneurogenic
phenotype in the hippocampus of aged mice. Brain Behav Immun 26:803-810.
Kozorovitskiy Y, Gross CG, Kopil C, Battaglia L, McBreen M, Stranahan AM, Gould E (2005)
Experience induces structural and biochemical changes in the adult primate brain. Proc
Natl Acad Sci U S A 102:17478-17482.
Krebiehl G, Ruckerbauer S, Burbulla LF, Kieper N, Maurer B, Waak J, Wolburg H, Gizatullina Z,
Gellerich FN, Woitalla D, Riess O, Kahle PJ, Proikas-Cezanne T, Krüger R (2010) Reduced
basal autophagy and impaired mitochondrial dynamics due to loss of Parkinson's disease-
associated protein DJ-1. PLoS One 5:e9367.
Kreitzer AC, Malenka RC (2008) Striatal plasticity and basal ganglia circuit function. Neuron
60:543-554.
LaFerla FM (2012) Preclinical success against Alzheimer's disease with an old drug. N Engl J Med
367:570-572.
Lambert J-C et al. (2009) Genome-wide association study identifies variants at CLU and CR1
associated with Alzheimer's disease. Nature Genetics 41:1094-U1068.
Lambert J-C et al. (2013) Meta-analysis of 74,046 individuals identifies 11 new susceptibility loci
for Alzheimer's disease. Nature Genetics 45:1452-U1206.
Langston JW (2017) The MPTP Story. J Parkinsons Dis 7:S11-S22.
Langston JW, Ballard P, Tetrud JW, Irwin I (1983) Chronic Parkinsonism in humans due to a
product of meperidine-analog synthesis. Science 219:979-980.
Lapierre LA, Avant KM, Caldwell CM, Oztan A, Apodaca G, Knowles BC, Roland JT, Ducharme NA,
Goldenring JR (2012) Phosphorylation of Rab11-FIP2 regulates polarity in MDCK cells.
Molecular Biology of the Cell 23:2302-2318.
Larsson HB, Courivaud F, Rostrup E, Hansen AE (2009) Measurement of brain perfusion, blood
volume, and blood-brain barrier permeability, using dynamic contrast-enhanced T(1)-
weighted MRI at 3 tesla. Magn Reson Med 62:1270-1281.
Lau YS, Patki G, Das-Panja K, Le WD, Ahmad SO (2011) Neuroprotective effects and mechanisms
of exercise in a chronic mouse model of Parkinson's disease with moderate
neurodegeneration. Eur J Neurosci 33:1264-1274.
Laughton JD, Bittar P, Charnay Y, Pellerin L, Kovari E, Magistretti PJ, Bouras C (2007) Metabolic
compartmentalization in the human cortex and hippocampus: evidence for a cell- and
region-specific localization of lactate dehydrogenase 5 and pyruvate dehydrogenase.
BMC Neurosci 8:35.
Leggio MG, Mandolesi L, Federico F, Spirito F, Ricci B, Gelfo F, Petrosini L (2005) Environmental
enrichment promotes improved spatial abilities and enhanced dendritic growth in the rat.
Behav Brain Res 163:78-90.
Leh SE, Petrides M, Strafella AP (2010) The neural circuitry of executive functions in healthy
subjects and Parkinson's disease. Neuropsychopharmacology 35:70-85.
103

Li J, Kanekiyo T, Shinohara M, Zhang Y, LaDu MJ, Xu H, Bu G (2012) Differential Regulation of
Amyloid-beta Endocytic Trafficking and Lysosomal Degradation by Apolipoprotein E
Isoforms. Journal of Biological Chemistry 287:44593-44601.
Li YH, Lu WY, Marzolo MP, Bu GJ (2001) Differential functions of members of the low density
lipoprotein receptor family suggested by their distinct endocytosis rates. Journal of
Biological Chemistry 276:18000-18006.
Lie DC, Song H, Colamarino SA, Ming GL, Gage FH (2004) Neurogenesis in the adult brain: new
strategies for central nervous system diseases. Annu Rev Pharmacol Toxicol 44:399-421.
MacRae PG, Spirduso WW, Cartee GD, Farrar RP, Wilcox RE (1987) Endurance training effects on
striatal D2 dopamine receptor binding and striatal dopamine metabolite levels. Neurosci
Lett 79:138-144.
Magistretti PJ, Allaman I (2015) A cellular perspective on brain energy metabolism and functional
imaging. Neuron 86:883-901.
Mali P, Yang L, Esvelt KM, Aach J, Guell M, DiCarlo JE, Norville JE, Church GM (2013) RNA-Guided
Human Genome Engineering via Cas9. Science 339:823-826.
Marsh M, McMahon HT (1999) Cell biology - The structural era of endocytosis. Science 285:215-
220.
Marti HJ, Bernaudin M, Bellail A, Schoch H, Euler M, Petit E, Risau W (2000) Hypoxia-induced
vascular endothelial growth factor expression precedes neovascularization after cerebral
ischemia. Am J Pathol 156:965-976.
Masoud GN, Li W (2015) HIF-1α pathway: role, regulation and intervention for cancer therapy.
Acta Pharm Sin B 5:378-389.
McWilliams TG, Muqit MM (2017) PINK1 and Parkin: emerging themes in mitochondrial
homeostasis. Curr Opin Cell Biol 45:83-91.
Miller SE, Sahlender DA, Graham SC, Hoening S, Robinson MS, Peden AA, Owen DJ (2011) The
Molecular Basis for the Endocytosis of Small R-SNAREs by the Clathrin Adaptor CALM. Cell
147:1118-1131.
Morris JC, Weintraub S, Chui HC, Cummings J, Decarli C, Ferris S, Foster NL, Galasko D, Graff-
Radford N, Peskind ER, Beekly D, Ramos EM, Kukull WA (2006) The Uniform Data Set
(UDS): clinical and cognitive variables and descriptive data from Alzheimer Disease
Centers. Alzheimer Dis Assoc Disord 20:210-216.
Mounsey RB, Teismann P (2010) Mitochondrial dysfunction in Parkinson's disease: pathogenesis
and neuroprotection. Parkinsons Dis 2011:617472.
Mu Y, Gage FH (2011) Adult hippocampal neurogenesis and its role in Alzheimer's disease. Mol
Neurodegener 6:85.
Nakajima K, Uchida S, Suzuki A, Hotta H, Aikawa Y (2003) The effect of walking on regional blood
flow and acetylcholine in the hippocampus in conscious rats. Auton Neurosci 103:83-92.
Navarro A, Gomez C, López-Cepero JM, Boveris A (2004) Beneficial effects of moderate exercise
on mice aging: survival, behavior, oxidative stress, and mitochondrial electron transfer.
Am J Physiol Regul Integr Comp Physiol 286:R505-511.
Nelson AJ, Iwamoto GA (2006) Reversibility of exercise-induced dendritic attenuation in brain
cardiorespiratory and locomotor areas following exercise detraining. J Appl Physiol (1985)
101:1243-1251.
104

Newington JT, Rappon T, Albers S, Wong DY, Rylett RJ, Cumming RC (2012) Overexpression of
pyruvate dehydrogenase kinase 1 and lactate dehydrogenase A in nerve cells confers
resistance to amyloid β and other toxins by decreasing mitochondrial respiration and
reactive oxygen species production. J Biol Chem 287:37245-37258.
Newman LA, Korol DL, Gold PE (2011) Lactate produced by glycogenolysis in astrocytes regulates
memory processing. PLoS One 6:e28427.
Obel LF, Müller MS, Walls AB, Sickmann HM, Bak LK, Waagepetersen HS, Schousboe A (2012)
Brain glycogen-new perspectives on its metabolic function and regulation at the
subcellular level. Front Neuroenergetics 4:3.
Okino ST, Chichester CH, Whitlock JP (1998) Hypoxia-inducible mammalian gene expression
analyzed in vivo at a TATA-driven promoter and at an initiator-driven promoter. J Biol
Chem 273:23837-23843.
Okita K, Matsumura Y, Sato Y, Okada A, Morizane A, Okamoto S, Hong H, Nakagawa M, Tanabe
K, Tezuka K-i, Shibata T, Kunisada T, Takahashi M, Takahashi J, Saji H, Yamanaka S (2011)
A more efficient method to generate integration-free human iPS cells. Nature Methods
8:409-U452.
Olive PL, Aquino-Parsons C, MacPhail SH, Liao SY, Raleigh JA, Lerman MI, Stanbridge EJ (2001)
Carbonic anhydrase 9 as an endogenous marker for hypoxic cells in cervical cancer.
Cancer Res 61:8924-8929.
Onla-or S, Winstein CJ (2008) Determining the optimal challenge point for motor skill learning in
adults with moderately severe Parkinson's disease. Neurorehabil Neural Repair 22:385-
395.
Opendak M, Gould E (2015) Adult neurogenesis: a substrate for experience-dependent change.
Trends Cogn Sci 19:151-161.
Pagliari R, Peyrin L (1995) Norepinephrine release in the rat frontal cortex under treadmill
exercise: a study with microdialysis. J Appl Physiol (1985) 78:2121-2130.
Parikh I, Fardo DW, Estus S (2014) Genetics of PICALM Expression and Alzheimer's Disease. Plos
One 9.
Park L, Zhou J, Zhou P, Pistick R, El Jamal S, Younkin L, Pierce J, Arreguin A, Anrather J, Younkin
SG, Carlson GA, McEwen BS, Iadecola C (2013) Innate immunity receptor CD36 promotes
cerebral amyloid angiopathy. Proc Natl Acad Sci U S A 110:3089-3094.
Patlak CS, Blasberg RG (1985) Graphical evaluation of blood-to-brain transfer constants from
multiple-time uptake data. Generalizations. J Cereb Blood Flow Metab 5:584-590.
Pellerin L, Magistretti PJ (1994) Glutamate uptake into astrocytes stimulates aerobic glycolysis: a
mechanism coupling neuronal activity to glucose utilization. Proc Natl Acad Sci U S A
91:10625-10629.
Petzinger GM, Fisher BE, McEwen S, Beeler JA, Walsh JP, Jakowec MW (2013) Exercise-enhanced
neuroplasticity targeting motor and cognitive circuitry in Parkinson's disease. Lancet
Neurol 12:716-726.
Petzinger GM, Holschneider DP, Fisher BE, McEwen S, Kintz N, Halliday M, Toy W, Walsh JW,
Beeler J, Jakowec MW (2015) The Effects of Exercise on Dopamine Neurotransmission in
Parkinson's Disease: Targeting Neuroplasticity to Modulate Basal Ganglia Circuitry. Brain
Plast 1:29-39.
105

Petzinger GM, Walsh JP, Akopian G, Hogg E, Abernathy A, Arevalo P, Turnquist P, Vucković M,
Fisher BE, Togasaki DM, Jakowec MW (2007) Effects of treadmill exercise on
dopaminergic transmission in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned
mouse model of basal ganglia injury. J Neurosci 27:5291-5300.
Pickel VM, Johnson E, Carson M, Chan J (1992) Ultrastructure of spared dopamine terminals in
caudate-putamen nuclei of adult rats neonatally treated with intranigral 6-
hydroxydopamine. Brain Res Dev Brain Res 70:75-86.
Pitkälä KH, Pöysti MM, Laakkonen ML, Tilvis RS, Savikko N, Kautiainen H, Strandberg TE (2013)
Effects of the Finnish Alzheimer disease exercise trial (FINALEX): a randomized controlled
trial. JAMA Intern Med 173:894-901.
Polito C, Berti V, Ramat S, Vanzi E, De Cristofaro MT, Pellicanò G, Mungai F, Marini P, Formiconi
AR, Sorbi S, Pupi A (2012) Interaction of caudate dopamine depletion and brain metabolic
changes with cognitive dysfunction in early Parkinson's disease. Neurobiol Aging
33:206.e229-239.
Polman CH, Reingold SC, Banwell B, Clanet M, Cohen JA, Filippi M, Fujihara K, Havrdova E,
Hutchinson M, Kappos L, Lublin FD, Montalban X, O'Connor P, Sandberg-Wollheim M,
Thompson AJ, Waubant E, Weinshenker B, Wolinsky JS (2011) Diagnostic criteria for
multiple sclerosis: 2010 revisions to the McDonald criteria. Ann Neurol 69:292-302.
Pomara N, Bruno D, Sarreal AS, Hernando RT, Nierenberg J, Petkova E, Sidtis JJ, Wisniewski TM,
Mehta PD, Pratico D, Zetterberg H, Blennow K (2012) Lower CSF amyloid beta peptides
and higher F2-isoprostanes in cognitively intact elderly individuals with major depressive
disorder. Am J Psychiatry 169:523-530.
Potter MC, Yuan C, Ottenritter C, Mughal M, van Praag H (2010) Exercise is not beneficial and
may accelerate symptom onset in a mouse model of Huntington's disease. PLoS Curr
2:RRN1201.
Quaegebeur A, Carmeliet P (2010) Oxygen sensing: a common crossroad in cancer and
neurodegeneration. Curr Top Microbiol Immunol 345:71-103.
Querfurth HW, LaFerla FM (2010) MECHANISMS OF DISEASE Alzheimer's Disease. New England
Journal of Medicine 362:329-344.
Quintanilla RA, Tapia C, Pérez MJ (2017) Possible role of mitochondrial permeability transition
pore in the pathogenesis of Huntington disease. Biochem Biophys Res Commun 483:1078-
1083.
Quintanilla RA, von Bernhardi R, Godoy JA, Inestrosa NC, Johnson GV (2014) Phosphorylated tau
potentiates Aβ-induced mitochondrial damage in mature neurons. Neurobiol Dis 71:260-
269.
Raj T, Shulman JM, Keenan BT, Chibnik LB, Evans DA, Bennett DA, Stranger BE, De Jager PL (2012)
Alzheimer Disease Susceptibility Loci: Evidence for a Protein Network under Natural
Selection. American Journal of Human Genetics 90:720-726.
Raven EP, Lu PH, Tishler TA, Heydari P, Bartzokis G (2013) Increased iron levels and decreased
tissue integrity in hippocampus of Alzheimer's disease detected in vivo with magnetic
resonance imaging. J Alzheimers Dis 37:127-136.
Real CC, Ferreira AF, Hernandes MS, Britto LR, Pires RS (2010) Exercise-induced plasticity of
AMPA-type glutamate receptor subunits in the rat brain. Brain Res 1363:63-71.
106

Rhyu IJ, Bytheway JA, Kohler SJ, Lange H, Lee KJ, Boklewski J, McCormick K, Williams NI, Stanton
GB, Greenough WT, Cameron JL (2010) Effects of aerobic exercise training on cognitive
function and cortical vascularity in monkeys. Neuroscience 167:1239-1248.
Ridgel AL, Vitek JL, Alberts JL (2009) Forced, not voluntary, exercise improves motor function in
Parkinson's disease patients. Neurorehabil Neural Repair 23:600-608.
Risterucci C, Terramorsi D, Nieoullon A, Amalric M (2003) Excitotoxic lesions of the prelimbic-
infralimbic areas of the rodent prefrontal cortex disrupt motor preparatory processes. Eur
J Neurosci 17:1498-1508.
Rosano C, Studenski SA, Aizenstein HJ, Boudreau RM, Longstreth WT, Newman AB (2012) Slower
gait, slower information processing and smaller prefrontal area in older adults. Age
Ageing 41:58-64.
Ross HH, Sandhu MS, Cheung TF, Fitzpatrick GM, Sher WJ, Tiemeier AJ, Laywell ED, Fuller DD
(2012) In vivo intermittent hypoxia elicits enhanced expansion and neuronal
differentiation in cultured neural progenitors. Exp Neurol 235:238-245.
Rothstein JD, Martin L, Levey AI, Dykes-Hoberg M, Jin L, Wu D, Nash N, Kuncl RW (1994)
Localization of neuronal and glial glutamate transporters. Neuron 13:713-725.
Ryu JK, McLarnon JG (2009) A leaky blood-brain barrier, fibrinogen infiltration and microglial
reactivity in inflamed Alzheimer's disease brain. J Cell Mol Med 13:2911-2925.
Sagare AP, Bell RD, Zhao Z, Ma Q, Winkler EA, Ramanathan A, Zlokovic BV (2013) Pericyte loss
influences Alzheimer-like neurodegeneration in mice. Nat Commun 4:2932.
Salloway S, Gur T, Berzin T, Tavares R, Zipser B, Correia S, Hovanesian V, Fallon J, Kuo-Leblanc V,
Glass D, Hulette C, Rosenberg C, Vitek M, Stopa E (2002) Effect of APOE genotype on
microvascular basement membrane in Alzheimer's disease. J Neurol Sci 203-204:183-187.
Sawamoto N, Piccini P, Hotton G, Pavese N, Thielemans K, Brooks DJ (2008) Cognitive deficits and
striato-frontal dopamine release in Parkinson's disease. Brain 131:1294-1302.
Schapira AH, Cooper JM, Dexter D, Clark JB, Jenner P, Marsden CD (1990) Mitochondrial complex
I deficiency in Parkinson's disease. J Neurochem 54:823-827.
Schapira AH, Agid Y, Barone P, Jenner P, Lemke MR, Poewe W, Rascol O, Reichmann H, Tolosa E
(2009) Perspectives on recent advances in the understanding and treatment of
Parkinson's disease. Eur J Neurol 16:1090-1099.
Schenkman M, Hall DA, Barón AE, Schwartz RS, Mettler P, Kohrt WM (2012) Exercise for people
in early- or mid-stage Parkinson disease: a 16-month randomized controlled trial. Phys
Ther 92:1395-1410.
Schwarting RK, Huston JP (1996) The unilateral 6-hydroxydopamine lesion model in behavioral
brain research. Analysis of functional deficits, recovery and treatments. Prog Neurobiol
50:275-331.
Semenova TP, Ivanov VA, Tret'yak TM (1981) Brain levels of noradrenalin, dopamine, and
serotonin in rats with different levels of motor activity. Neurosci Behav Physiol 11:153-
155.
Semenza G (2002) Signal transduction to hypoxia-inducible factor 1. Biochem Pharmacol 64:993-
998.
Sengillo JD, Winkler EA, Walker CT, Sullivan JS, Johnson M, Zlokovic BV (2013) Deficiency in mural
vascular cells coincides with blood-brain barrier disruption in Alzheimer's disease. Brain
Pathol 23:303-310.
107

Sharp FR, Bernaudin M (2004) HIF1 and oxygen sensing in the brain. Nat Rev Neurosci 5:437-448.
Sharp FR, Ran R, Lu A, Tang Y, Strauss KI, Glass T, Ardizzone T, Bernaudin M (2004) Hypoxic
preconditioning protects against ischemic brain injury. NeuroRx 1:26-35.
Shibata M, Yamada S, Kumar SR, Calero M, Bading J, Frangione B, Holtzman DM, Miller CA,
Strickland DK, Ghiso J, Zlokovic BV (2000) Clearance of Alzheimer's amyloid-beta(1-40)
peptide from brain by LDL receptor-related protein-1 at the blood-brain barrier. Journal
of Clinical Investigation 106:1489-1499.
Shweiki D, Itin A, Soffer D, Keshet E (1992) Vascular endothelial growth factor induced by hypoxia
may mediate hypoxia-initiated angiogenesis. Nature 359:843-845.
Simpson IA, Carruthers A, Vannucci SJ (2007) Supply and demand in cerebral energy metabolism:
the role of nutrient transporters. J Cereb Blood Flow Metab 27:1766-1791.
Smeyne M, Sladen P, Jiao Y, Dragatsis I, Smeyne RJ (2015) HIF1α is necessary for exercise-induced
neuroprotection while HIF2α is needed for dopaminergic neuron survival in the
substantia nigra pars compacta. Neuroscience 295:23-38.
Smith D, Pernet A, Hallett WA, Bingham E, Marsden PK, Amiel SA (2003) Lactate: a preferred fuel
for human brain metabolism in vivo. J Cereb Blood Flow Metab 23:658-664.
Smith Y, Villalba RM, Raju DV (2009) Striatal spine plasticity in Parkinson's disease: pathological
or not? Parkinsonism Relat Disord 15 Suppl 3:S156-161.
Sorkin A, von Zastrow M (2009) Endocytosis and signalling: intertwining molecular networks.
Nature Reviews Molecular Cell Biology 10:609-622.
Squire LR (1992) Memory and the hippocampus: a synthesis from findings with rats, monkeys,
and humans. Psychol Rev 99:195-231.
Stefanko DP, Shah VD, Yamasaki WK, Petzinger GM, Jakowec MW (2017) Treadmill exercise
delays the onset of non-motor behaviors and striatal pathology in the CAG. Neurobiol Dis
105:15-32.
Steiner B, Winter C, Hosman K, Siebert E, Kempermann G, Petrus DS, Kupsch A (2006) Enriched
environment induces cellular plasticity in the adult substantia nigra and improves motor
behavior function in the 6-OHDA rat model of Parkinson's disease. Exp Neurol 199:291-
300.
Steiner JL, Murphy EA, McClellan JL, Carmichael MD, Davis JM (2011) Exercise training increases
mitochondrial biogenesis in the brain. J Appl Physiol (1985) 111:1066-1071.
Stenmark H (2009) Rab GTPases as coordinators of vesicle traffic. Nature Reviews Molecular Cell
Biology 10:513-525.
Suzuki A, Stern SA, Bozdagi O, Huntley GW, Walker RH, Magistretti PJ, Alberini CM (2011)
Astrocyte-neuron lactate transport is required for long-term memory formation. Cell
144:810-823.
Taheri S, Gasparovic C, Shah NJ, Rosenberg GA (2011a) Quantitative measurement of blood-brain
barrier permeability in human using dynamic contrast-enhanced MRI with fast T1
mapping. Magn Reson Med 65:1036-1042.
Taheri S, Gasparovic C, Huisa BN, Adair JC, Edmonds E, Prestopnik J, Grossetete M, Shah NJ, Wills
J, Qualls C, Rosenberg GA (2011b) Blood-brain barrier permeability abnormalities in
vascular cognitive impairment. Stroke 42:2158-2163.
108

Takahashi S, Kubo K, Waguri S, Yabashi A, Shin H-W, Katoh Y, Nakayama K (2012) Rab11 regulates
exocytosis of recycling vesicles at the plasma membrane. Journal of Cell Science
125:4049-4057.
Takimoto M, Hamada T (2014) Acute exercise increases brain region-specific expression of MCT1,
MCT2, MCT4, GLUT1, and COX IV proteins. J Appl Physiol (1985) 116:1238-1250.
Tanzi RE (2012) The genetics of Alzheimer disease. Cold Spring Harb Perspect Med 2.
Tebar F, Bohlander SK, Sorkin A (1999) Clathrin assembly lymphoid myeloid leukemia (CALM)
protein: Localization in endocytic-coated pits, interactions with clathrin, and the impact
of overexpression on clathrin-mediated traffic. Molecular Biology of the Cell 10:2687-
2702.
Thacker EL, Chen H, Patel AV, McCullough ML, Calle EE, Thun MJ, Schwarzschild MA, Ascherio A
(2008) Recreational physical activity and risk of Parkinson's disease. Mov Disord 23:69-
74.
Tian Y, Chang JC, Fan EY, Flajolet M, Greengard P (2013) Adaptor complex AP2/PICALM, through
interaction with LC3, targets Alzheimer's APP-CTF for terminal degradation via autophagy.
Proceedings of the National Academy of Sciences of the United States of America
110:17071-17076.
Toda N, Ayajiki K, Okamura T (2009) Cerebral blood flow regulation by nitric oxide: recent
advances. Pharmacol Rev 61:62-97.
Torp R, Hoover F, Danbolt NC, Storm-Mathisen J, Ottersen OP (1997) Differential distribution of
the glutamate transporters GLT1 and rEAAC1 in rat cerebral cortex and thalamus: an in
situ hybridization analysis. Anat Embryol (Berl) 195:317-326.
Toy WA, Petzinger GM, Leyshon BJ, Akopian GK, Walsh JP, Hoffman MV, Vučković MG, Jakowec
MW (2014) Treadmill exercise reverses dendritic spine loss in direct and indirect striatal
medium spiny neurons in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)
mouse model of Parkinson's disease. Neurobiol Dis 63:201-209.
Treusch S et al. (2011) Functional Links Between A beta Toxicity, Endocytic Trafficking, and
Alzheimer's Disease Risk Factors in Yeast. Science 334:1241-1245.
Uchida S, Suzuki A, Kagitani F, Hotta H (2006) Responses of acetylcholine release and regional
blood flow in the hippocampus during walking in aged rats. J Physiol Sci 56:253-257.
van Hall G, Strømstad M, Rasmussen P, Jans O, Zaar M, Gam C, Quistorff B, Secher NH, Nielsen
HB (2009) Blood lactate is an important energy source for the human brain. J Cereb Blood
Flow Metab 29:1121-1129.
van Praag H, Kempermann G, Gage FH (1999) Running increases cell proliferation and
neurogenesis in the adult mouse dentate gyrus. Nat Neurosci 2:266-270.
VanLeeuwen JE, Petzinger GM, Walsh JP, Akopian GK, Vuckovic M, Jakowec MW (2010) Altered
AMPA receptor expression with treadmill exercise in the 1-methyl-4-phenyl-1,2,3,6-
tetrahydropyridine-lesioned mouse model of basal ganglia injury. J Neurosci Res 88:650-
668.
Vecchi M, Polo S, Poupon V, van de Loo JW, Benmerah A, Di Fiore PP (2001) Nucleocytoplasmic
shuttling of endocytic proteins. Journal of Cell Biology 153:1511-1517.
Verghese PB, Castellano JM, Holtzman DM (2011) Apolipoprotein E in Alzheimer's disease and
other neurological disorders. Lancet Neurol 10:241-252.
109

Viboolvorakul S, Patumraj S (2014) Exercise training could improve age-related changes in
cerebral blood flow and capillary vascularity through the upregulation of VEGF and eNOS.
Biomed Res Int 2014:230791.
Villalba RM, Lee H, Smith Y (2009) Dopaminergic denervation and spine loss in the striatum of
MPTP-treated monkeys. Exp Neurol 215:220-227.
Voelcker-Rehage C, Godde B, Staudinger UM (2010) Physical and motor fitness are both related
to cognition in old age. Eur J Neurosci 31:167-176.
Voelcker-Rehage C, Godde B, Staudinger UM (2011) Cardiovascular and coordination training
differentially improve cognitive performance and neural processing in older adults. Front
Hum Neurosci 5:26.
Vučković MG, Li Q, Fisher B, Nacca A, Leahy RM, Walsh JP, Mukherjee J, Williams C, Jakowec MW,
Petzinger GM (2010) Exercise elevates dopamine D2 receptor in a mouse model of
Parkinson's disease: in vivo imaging with [¹⁸F]fallypride. Mov Disord 25:2777-2784.
Wang GL, Jiang BH, Rue EA, Semenza GL (1995) Hypoxia-inducible factor 1 is a basic-helix-loop-
helix-PAS heterodimer regulated by cellular O2 tension. Proc Natl Acad Sci U S A 92:5510-
5514.
Wang Z, Guo Y, Myers KG, Heintz R, Holschneider DP (2015a) Recruitment of the prefrontal cortex
and cerebellum in Parkinsonian rats following skilled aerobic exercise. Neurobiol Dis
77:71-87.
Wang Z, Myers KG, Guo Y, Ocampo MA, Pang RD, Jakowec MW, Holschneider DP (2013)
Functional reorganization of motor and limbic circuits after exercise training in a rat
model of bilateral parkinsonism. PLoS One 8:e80058.
Wang Z, Guo Y, Myers KG, Heintz R, Peng YH, Maarek JM, Holschneider DP (2015b) Exercise alters
resting-state functional connectivity of motor circuits in parkinsonian rats. Neurobiol
Aging 36:536-544.
Wang Z, Stefanko DP, Guo Y, Toy WA, Petzinger GM, Jakowec MW, Holschneider DP (2016)
Evidence of functional brain reorganization on the basis of blood flow changes in the
CAG140 knock-in mouse model of Huntington's disease. Neuroreport 27:632-639.
Weintraub S, Salmon D, Mercaldo N, Ferris S, Graff-Radford NR, Chui H, Cummings J, DeCarli C,
Foster NL, Galasko D, Peskind E, Dietrich W, Beekly DL, Kukull WA, Morris JC (2009) The
Alzheimer's Disease Centers' Uniform Data Set (UDS): the neuropsychologic test battery.
Alzheimer Dis Assoc Disord 23:91-101.
Wenger RH, Stiehl DP, Camenisch G (2005) Integration of oxygen signaling at the consensus HRE.
Sci STKE 2005:re12.
Whitwell JL, Wiste HJ, Weigand SD, Rocca WA, Knopman DS, Roberts RO, Boeve BF, Petersen RC,
Jack CR, Initiative ADN (2012) Comparison of imaging biomarkers in the Alzheimer Disease
Neuroimaging Initiative and the Mayo Clinic Study of Aging. Arch Neurol 69:614-622.
Winkler EA, Bell RD, Zlokovic BV (2010) Pericyte-specific expression of PDGF beta receptor in
mouse models with normal and deficient PDGF beta receptor signaling. Molecular
Neurodegeneration 5.
Winkler EA, Bell RD, Zlokovic BV (2011) Central nervous system pericytes in health and disease.
Nat Neurosci 14:1398-1405.
Winkler EA, Sengillo JD, Sagare AP, Zhao Z, Ma Q, Zuniga E, Wang Y, Zhong Z, Sullivan JS, Griffin
JH, Cleveland DW, Zlokovic BV (2014) Blood-spinal cord barrier disruption contributes to
110

early motor-neuron degeneration in ALS-model mice. Proc Natl Acad Sci U S A 111:E1035-
1042.
Wong-Riley MT (1989) Cytochrome oxidase: an endogenous metabolic marker for neuronal
activity. Trends Neurosci 12:94-101.
Wood SM, Gleadle JM, Pugh CW, Hankinson O, Ratcliffe PJ (1996) The role of the aryl hydrocarbon
receptor nuclear translocator (ARNT) in hypoxic induction of gene expression. Studies in
ARNT-deficient cells. J Biol Chem 271:15117-15123.
Wu ZH, Guo H, Chow NW, Sallstrom J, Bell RD, Deane R, Brooks AI, Kanagala S, Rubio A, Sagare A,
Liu D, Li F, Armstrong D, Gasiewicz T, Zidovetzki R, Song XM, Hofman F, Zlokovic BV (2005)
Role of the MEOX2 homeobox gene in neurovascular dysfunction in Alzheimer disease.
Nature Medicine 11:959-965.
Xiao Q, Gil S-C, Yan P, Wang Y, Han S, Gonzales E, Perez R, Cirrito JR, Lee J-M (2012) Role of
Phosphatidylinositol Clathrin Assembly Lymphoid-Myeloid Leukemia (PICALM) in
Intracellular Amyloid Precursor Protein (APP) Processing and Amyloid Plaque
Pathogenesis. Journal of Biological Chemistry 287:21279-21289.
Xie W, Chung KK (2012) Alpha-synuclein impairs normal dynamics of mitochondria in cell and
animal models of Parkinson's disease. J Neurochem 122:404-414.
Xu S, Edman M, Kothawala MS, Sun G, Chiang L, Mircheff A, Zhu L, Okamoto C, Hamm-Alvarez S
(2011) A Rab11a-enriched subapical membrane compartment regulates a cytoskeleton-
dependent transcytotic pathway in secretory epithelial cells of the lacrimal gland. Journal
of Cell Science 124:3503-3514.
Yang J, Sadler TR, Givrad TK, Maarek JM, Holschneider DP (2007) Changes in brain functional
activation during resting and locomotor states after unilateral nigrostriatal damage in
rats. Neuroimage 36:755-773.
Yang J, Ruchti E, Petit JM, Jourdain P, Grenningloh G, Allaman I, Magistretti PJ (2014) Lactate
promotes plasticity gene expression by potentiating NMDA signaling in neurons. Proc Natl
Acad Sci U S A 111:12228-12233.
Yawata S, Yamaguchi T, Danjo T, Hikida T, Nakanishi S (2012) Pathway-specific control of reward
learning and its flexibility via selective dopamine receptors in the nucleus accumbens.
Proc Natl Acad Sci U S A 109:12764-12769.
Yim AJ, Andersen ML, Soeiro AC, Tufik S, Oliveira MG (2009) Acute systemic blockade of D2
receptors does not accelerate the extinction of cocaine-associated place preference.
Brain Res 1304:122-128.
Yoshita M, Fletcher E, Harvey D, Ortega M, Martinez O, Mungas DM, Reed BR, DeCarli CS (2006)
Extent and distribution of white matter hyperintensities in normal aging, MCI, and AD.
Neurology 67:2192-2198.
Yui N, Lu HAJ, Chen Y, Nomura N, Bouley R, Brown D (2013) Basolateral targeting and
microtubule-dependent transcytosis of the aquaporin-2 water channel. American Journal
of Physiology-Cell Physiology 304:C38-C48.
Zeigerer A, Gilleron J, Bogorad RL, Marsico G, Nonaka H, Seifert S, Epstein-Barash H, Kuchimanchi
S, Peng CG, Ruda VM, Del Conte-Zerial P, Hengstler JG, Kalaidzidis Y, Koteliansky V, Zerial
M (2012) Rab5 is necessary for the biogenesis of the endolysosomal system in vivo.
Nature 485:465-470.
111

Zhang Z, Yan J, Chang Y, ShiDu Yan S, Shi H (2011) Hypoxia inducible factor-1 as a target for
neurodegenerative diseases. Curr Med Chem 18:4335-4343.
Zhu D, Wang Y, Singh I, Bell RD, Deane R, Zhong Z, Sagare A, Winkler EA, Zlokovic BV (2010)
Protein S controls hypoxic/ischemic blood-brain barrier disruption through the TAM
receptor Tyro3 and sphingosine 1-phosphate receptor. Blood 115:4963-4972.
Zhu LL, Zhao T, Li HS, Zhao H, Wu LY, Ding AS, Fan WH, Fan M (2005) Neurogenesis in the adult
rat brain after intermittent hypoxia. Brain Res 1055:1-6.
Zipser BD, Johanson CE, Gonzalez L, Berzin TM, Tavares R, Hulette CM, Vitek MP, Hovanesian V,
Stopa EG (2007) Microvascular injury and blood-brain barrier leakage in Alzheimer's
disease. Neurobiol Aging 28:977-986.
Zlokovic BV (2008) The blood-brain barrier in health and chronic neurodegenerative disorders.
Neuron 57:178-201.
Zlokovic BV (2011) Neurovascular pathways to neurodegeneration in Alzheimer's disease and
other disorders. Nat Rev Neurosci 12:723-738.
Zlokovic BV (2013) Cerebrovascular effects of apolipoprotein E: implications for Alzheimer
disease. JAMA Neurol 70:440-444.
Zonneveld HI, Goos JD, Wattjes MP, Prins ND, Scheltens P, van der Flier WM, Kuijer JP, Muller M,
Barkhof F (2014) Prevalence of cortical superficial siderosis in a memory clinic population.
Neurology 82:698-704.
Öhman H, Savikko N, Strandberg TE, Kautiainen H, Raivio MM, Laakkonen ML, Tilvis R, Pitkälä KH
(2016) Effects of Exercise on Cognition: The Finnish Alzheimer Disease Exercise Trial: A
Randomized, Controlled Trial. J Am Geriatr Soc 64:731-738.

112

APPENDIX D: Accelerated pericyte degeneration and blood-brain barrier breakdown in
apolipoprotein E4 carriers with Alzheimer’s disease

Authors: Matthew R. Halliday, Sanket V. Rege, Qingyi Ma, Zhen Zhao, Carol A. Miller, Ethan A. Winkler,
and Berislav V. Zlokovic

ABSTRACT

The blood-brain barrier (BBB) limits the entry of neurotoxic blood-derived products and cells into
the brain that is required for normal neuronal functioning and information processing. Pericytes maintain
the integrity of the BBB and degenerate in Alzheimer’s disease (AD). The BBB is damaged in AD,
particularly in individuals carrying apolipoprotein E4 (APOE4) gene, which is a major genetic risk factor for
late-onset AD. The mechanisms underlying the BBB breakdown in AD remain, however, elusive. Here, we
show accelerated pericyte degeneration in AD APOE4 carriers > AD APOE3 carriers > non-AD controls,
which correlates with the magnitude of BBB breakdown to immunoglobulin G and fibrin. We also show
accumulation of the proinflammatory cytokine cyclophilin A (CypA) and matrix metalloproteinase-9 (MMP-
9) in pericytes and endothelial cells in AD (APOE4 > APOE3), previously shown to lead to BBB breakdown
in transgenic APOE4 mice. The levels of the apoE lipoprotein receptor, low-density lipoprotein receptor-
related protein-1 (LRP-1), were similarly reduced in AD APOE4 and APOE3 carriers. Our data suggest that
APOE4 leads to accelerated pericyte loss and enhanced activation of LRP-1-dependent CypA-MMP-9
BBB-degrading pathway in pericytes and endothelial cells, which can mediate a greater BBB damage in
AD APOE4 compared with AD APOE3 carriers.

INTRODUCTION

A neurovascular unit is formed by vascular endothelial cells, pericytes, form the blood-brain barrier
(BBB) in the dense microvasculature networks responsible for molecular exchange. The BBB limits entry
into the brain of potentially toxic blood-derived products and larger molecules such as peptides and proteins
but allows controlled carrier-mediated bidirectional trans-endothelial transport of nutrients, such as glucose
and amino acids from blood to brain and from brain to blood. The cross-talk between different neurovascular
unit cell types controls multiple central nervous system functions including regulation of cerebrovascular
and BBB integrity and cerebral blood flow. Recent studies in murine transgenic models have shown that
113

pericytes have a critical role in maintaining the BBB integrity, and that loss of pericytes can lead to a long-
term BBB breakdown and small vessel disease contributing to neurodegenerative changes.

Pericytes degenerate in Alzheimer’s disease (AD). Multiple studies have demonstrated loss of
cerebrovascular integrity and/or BBB damage in AD that is accelerated by the apolipoprotein E4 (APOE4)
genotype. Human apoE has three major apoE isoforms: E2, E3, and E4. Genetic evidence including recent
genome-wide association studies shows that APOE4 is the strongest and most highly replicated genetic
risk factor for late-onset AD. Individuals with one copy of APOE4 have a 3.7-fold increase in AD risk and
individuals with two copies of APOE4 a 12-fold increase in AD risk relative to APOE3/E3 individuals. Studies
using murine transgenic models have shown that APOE4 increases BBB susceptibility to injury and leads
to BBB breakdown and microvascular reductions in humanized transgenic APOE4 mice compared to
APOE3 mice. The mechanisms underlying the BBB breakdown in AD, particularly in APOE4 carriers
remain, however, elusive.
Recent studies in transgenic APOE2 and APOE3 mice have shown that astrocyte-secreted apoE2
and apoE3 maintain the BBB integrity by suppressing the proinflammatory cyclophilin A (CypA)-
metalloproteinase-9 (MMP-9) pathway in pericytes via low-density lipoprotein receptor-related protein-1
(LRP1), that is a major apoE receptor. In contrast, astrocyte-secreted apoE4 fails to effectively suppress
CypA-MMP-9 pathway in APOE4 transgenic mice leading to MMP-9-mediated degradation of the BBB tight
junction and basement membrane proteins, which cause BBB breakdown. Moreover, a recent study in
cognitively normal human APOE4 carriers compared with APOE3 carriers has shown an age-dependent
increase in CypA and active MMP-9 levels in the cerebrospinal fluid (CSF) suggestive of activation of the
CypA-MMP-9 pathway that correlated with increased CSF/plasma albumin ration indicating BBB
breakdown. Using postmortem human brain tissue analysis of biomarkers of BBB breakdown, here we
show that APOE4 compared with APOE3 accelerates pericyte loss and microvascular reductions in AD,
which correlates with the magnitude of BBB breakdown to plasma proteins immunoglobulin G and fibrin.
We then show that APOE4 compared with APOE3 leads to a greater accumulation of CypA and MMP-9 in
pericytes and endothelial cells in AD suggestive of an enhanced activation of LRP1-dependent CypA-MMP-
9 BBB-degrading pathway, which in addition to pericyte loss may contribute to accelerated BBB breakdown
in AD APOE4 compared with AD APOE3 carriers.
114

Materials and Methods

Human Postmortem Tissue Samples

Postmortem paraffin-embedded human frontal cortex tissue samples (Brodmann area 9/10) were
obtained from University of Southern California (USC) Alzheimer’s disease Research Center (ADRC) in
accordance with institutional guidelines governed by approved protocols. Informed consent was obtained
from all the participants before death, and the tissue collection was approved by the Institutional Review
Board (IRB) through the USC ADRC. All the autopsy cases underwent neuropathologic evaluation of AD
according to the Consortium to Establish a Registry for Alzheimer’s disease and by the National Institute
on Aging and Reagan Institute criteria including assignment of the Braak stage (measure of number and
distribution of neurofibrillary tangles). Aged subjects that did not carry the diagnosis of AD or another
neurodegenerative disease and showed neuropathologic findings within the normal range for age were
used as age-matched controls. Clinical dementia rating (CDR) and Mini-Mental State Examination were
available for most but not all individuals. A total of nine controls and 18 AD individuals were used for
histopathologic analyses. See Table 1 for detailed demographic data. The postmortem interval (PMI)
typically ranged between 3 and 10 hours, but individuals with shorter PMI of 1 and 2 hours (two cases), and
somewhat longer PMI of 12 to 13 hours (two cases) were also included in our cohort of n= 27 cases total.
The incidence of vascular risk factors (e.g., hypertension, atherosclerosis, etc.), the gender ration, age, and
the PMI were comparable between age-matched controls and AD patients. The cause of death in both the
groups was either cardiac or respiratory arrest.

Exclusion Criteria

In this study we excluded any patients with the following conditions: frontotemporal dementia, Lewy
body dementia, vascular dementia, Parkinson’s disease dementia, stroke, brain cancer, mental retardation,
severe depression, and/or severe weakness leading to invalid mental health assessment.

115

Tissues Immunofluorescent Staining

For all analyses, formalin-fixed, parrafin-embedded brain tissues samples were used. All tissue
was cut to a thickness of 10 um. Sections were deparrafinized with xylene and rehydrated to distilled water
after serial ethanol washes. Subsequently, sections were incubated in 1:10 diluted target antigen retrieval
solution, pH 9 (Dako, Carpinteria, CA, USA) for 20 minutes at 95⁰C. The tissue sections were blocked in
116

5% donkey serum (Jackson ImmunoResearch, West Grove, PA, USA) containing 0.05% Triton X-100
(Sigma-Aldrich, St Louis, MO, USA) and then incubated with the following primary antibodies overnight at
4⁰C: goat anti-human PDGFRβ (1:100, R&D Systems, Minneapolis, MN, USA) to detect pericyte marker
PDGFRβ, rabbit anti-human fibrinogen (1:500, Dako), goat anti-human immunoglobulin G (IgG) (1:100,
R&D Systems), monoclonal rabbit anti-human LRP-1 (1:100, Abcam, EPR3724, Cambridge, MA, USA),
rabbit anti-human CypA (1:500, Abcam, ab42408), rabbit anti-human MMP-9 (1:500, Abcam, ab38898). To
visualize brain endothelial vascular profiles, sections were incubated with Dylight 488-conjugated
Lycopersicon esculentum lectin (Vector Laboratories, Burlingame, CA, USA; DL-1174; 1:200) for 1 hour at
room temperature during the secondary antibody incubation step. Alexa Fluor 568- or 633-conjugated
donkey anti-goat secondary antibodies (1:200, Life Technologies, Grand Island, NY, USA) were used to
detect goat anti-human PDGFRβ on pericytes. To visualize fibrin and IgG accumulations, sections were
incubated with 568-conjugated donkey anti-rabbit (1:200, Life Technologies) and 568-conjugated donkey
anti-goat (1:200, Life Technologies) secondary antibodies, respectively. To visualize LRP-1, CypA, and
MMP-9, tissue sections were incubated with Alexa Fluor 568-conjugated donkey anti-rabbit antibody
(1:200, Life Technologies). All secondary antibodies were incubated for 1 hour at room temperature. Sudan
black-based Autofluorescence Eliminator Reagent (EMD Millipore, Billerica, MA, USA Cat. #2160) was
applied to the samples following the manufacturer’s protocol to reduce the tissue autofluoresence before
imaging. Tissue sections were cover slipped using fluorescent mounting media (Dako).

Confocal Microscopic Analysis

All coverslipped fluorescently mounted tissue sections were scanned using a custom-built Zeiss
510 meta-confocal laster scanning microscope with a Zeiss Apochromat 259/0.8 NA water immersion
objective (Carl Zeiss MicroImaging, Thornwood, NY, USA). A 488-nm argon laser was used to excite Alexa
Fluor 488 and DyLight 488 and the emission was collected through a 500- to 550-nm band pass filter. A
543-nm HeNe laser was used to excite Alexa Fluor 568 and the emission was collected through a 560- to
615-nm band pass filter. A 633-nm HeNe laser was used to excite DyLight 649 and the emission was
collected through a 650- to 700-nm band pass filter.

117

Quantitative Image Analysis

All the images were analyzed using NIH-developed ImageJ software (Bethesda, MD, USA). A field
size of 420 um x 420 um and a maximum z-projection of 8um were utilized for all the images. For each
analysis described below, five randomly selected fields per section from three non-adjacent tissue sections
100 um apart per specimen were analyzed.

Pericyte coverage and numbers.

The quantification analysis of pericyte coverage and numbers was restricted to PDGFRβ-positive
mural cells that were associated with brain capillaries defined as microvessels <6 um in diameter, as
previously described. The endothelial-specific L. esculentum lectin was used to visualize endothelial
vascular profiles, as previously reported. The PDGFRβ coverage of brain capillaries was determined as
PDGFRβ-positive area (percentage) occupying lectin-positive endothelial capillary profiles, as we and
others have previously described. Pericyte numbers were determined by counting the number of PDGFRβ-
positive cell bodies on the abluminal side of the endothelial membrane (lectin) that colocalized with DAPI
(4’,6-diami-dino-2-phenylindole)-positive nuclei using the ImageJ Cell Counter plug-in, as we and others
previously described. The number of pericytes was expressed per mm
2
of tissue.

Microvascular capillary length.

Microvascular capillary length was quantified as the length of lectin-positive endothelial profiles (<6
um in diameter) determined using the ImageJ plug-in length analysis tool from 15 images per specimen
derived from five randomly selected fields in the cortex (420 um x 420 um) per section from three non-
adjacent (~100 um apart) sections per specimen, as we described. The length was expressed in millimeter
of lectin-positive vascular profiles per mm
3
of brain tissue.

Extravascular leakages.

The levels of extravascular blood-derived fibrin and IgG indicating BBB breakdown were measured
as previously described. Briefly, fibrin-positive and/or IgG-positive extravascular signal was obtained by
subtracting fibrin and/or IgG signal colocalized with the lectin-positive signal from total fibrin and/or IgG-
positive area, yielding a value that reflects only extravascular accumulation of each plasma-derived protein.

118

LRP1, CypA, and MMP-9 colocalization with PDGFRβ-positive pericytes and lectin-positive endothelial
profiles.

To quantify the area positive for LRP1, CypA, and MMP-9 in pericytes and brain endothelial cells,
the LRP1-positive, CypA-positive, and MMP-9-positive signal (percent-age) occupying PDGFRβ-positive
pericyte area and/or lectin-positive brain endothelial profiles were determined. The presence of LRP1,
CypA, and MMP-9 immunoreactivity in PDGFRβ-positive pericytes and lectin-positive endothelium was
determined in single confocal planes. Only single-positive areas were used for the quantification. We
utilized a plug-in in NIH ImageJ software, which quantifies and summates colocalized immunofluorescent
signal in each individual image—instead of the projection—which is included in a z-stack. This maximizes
the specificity allotted to us by confocal microscopy and only includes colocalized signal in a single plane.
This approach limits the potential confounding effect of colocalization analysis on maximum projections of
confocal stacks that would fail to unambiguously determine the cellular origin as immunofluorescent staining
from above and/or below the cell of interest would artifactually appear to colocalize.

Post-image thresholding.

To account for background fluorescence and nonspecific staining, we have meticulously optimized
blocking and staining conditions, including but not limited to tissue preparation, standard operating
procedures for use of primary and secondary antibodies, and antigen retrieval, to maximize specific tissue
staining while decreasing background signal. All slides are stained and processed in batch with identical
primary and secondary antibodies dilutions, lot number, incubation times, and laser settings as explained
above. To minimize the background autofluoresence, we used an additional incubation step with Sudan
black reagent that was applied to all control and AD tissue specimens as described above. To account for
nonspecific background staining, tissue sections from control and AD specimens were incubated with
different sets of the respective secondary antibodies for 1 hour at room temperature as described above,
but no primary antibodies. The nonspecific staining is then measured using NIH ImageJ area analysis tool,
and then subtracted from the raw images. Importantly, the thresholded images for all studied molecules or
cell-biomarkers did not differ in any appreciable manner from the raw images as the background staining
for various secondary antibodies ranged between 3 and 5% of the total signal on the raw images. All of the
images were analyzed by a masked investigator.
119

Statistical Analysis

Sample sizes were calculated using nQUERY assuming a two-sided alpha level of 0.05%, 80%
power, and homogenous variances for the two samples to be compared. Using the means and common
standard deviation for different parameters predicted from our published data and pilot studies, the sample
sizes to detect the differences between AD and non-AD groups 25% varied between 5 and 8. Our actual
sample sizes for different cellular and molecular biomarkers varied between 6 and 12, thus satisfying a
reliable measurement of the predicted effect as defined above. The F-test was conducted to verify that the
samples are normally distributed and have homogenous variances. The variances of the respective
samples compared between the groups were statistically similar. All data represent the average±s.e.m. For
multiple comparisons, a one-way analysis of variance followed by Tukey’s post hoc test was used. A P
value of <0.05 was considered statistically significant. Simple linear regression and Pearson correlation
coefficient were used to determine the correlation between variables using GraphPad Prism 3.0 software.

Results

In all studies, we compared biomarkers of BBB breakdown in the cortical tissue (Brodmann area
9/10) derived from nine age-matched non-AD controls, six AD APOE3/APOE3 carriers (APOE3 group), six
AD APOE3/APOE4 carriers and six AD APOE4/APOE4 carriers (see Table 1). For the statistical analysis,
the AD cases with one and two APOE4 alleles were pooled together in a single APOE4 group (n=12), as
we did not find a significant effect between the two versus one APOE4 allele for the vast majority of the
studied BBB biomarkers using the present relatively small cohort of APOE4 carriers.

Accelerated Pericyte Degeneration and Microvascular Reductions in Alzheimer’s Disease Apolipoprotein
E4 Carriers

Figures 1A to 1C illustrate that the numbers of PDGFRβ-positive pericytes and pericyte coverage
of the capillary wall are reduced in AD APOE3 carriers compared with non-AD controls by 30% and 23%,
respectively, and in AD APOE4 carriers compared with AD APOE3 carriers by 31% and 15%, respectively.
Supplementary Figure 1 illustrates PDGFRβ-positive pericyte cell bodies that were used for calculations of
pericyte numbers shown in Figure 1B. Confocal microscopic analysis of lectin-positive brain endothelial
profiles detected significant reductions in the total capillary length in AD APOE3 carriers compared with
120

non-AD controls by 14%, and in AD APOE4 carriers compared with AD APOE3 carriers by an additional
20% (Figure 1D). As reported previously, reductions in the pericyte coverage correlated positively with the
reductions in total capillary length in the studied AD patients and controls as shown by individual single
point data regression analysis (Figure 1E).

Accelerated Blood-Brain Barrier Breakdown in Alzheimer’s Disease Apolipoprotein E4 Carriers Correlates
with Pericyte Degeneration

To determine the degree of BBB breakdown, we next determined the levels of fibrin and IgG
extravascular accumulations depicting vascular capillary leakages. Figures 2A and 2B, show a 6.9-fold
increase in fibrin perivascular deposits in AD APOE3 carriers compared with non-AD controls, and a 3.1-
fold increase in AD APOE4 carriers compared with AD APOE3 carriers. There was a significant (P < 0.001)
121

correlation between the loss of pericyte coverage and the magnitude of fibrin deposits in AD APOE3 and
AD APOE4 carriers and non-AD controls (Figure 2C). We also found increases in the IgG extravascular
accumulation by 5.3-fold and 2.6-fold in AD APOE3 carriers compared with non-AD controls, and in AD
APOE4 compared with AD APOE3 carriers, respectively (Figures 2D and 2E). As for fibrin, there was a
significant (P < 0.001) correlation between the loss of pericyte coverage and the magnitude of IgG
extravascular deposits (Figure 2F). In addition, we found a negative correlation between loss of pericyte
coverage and Braak stage or CDR, i.e., the greater the loss of pericyte coverage, the greater the Braak
stage and the CDR, and a positive correlation between IgG extravascular deposits and Braak stage and
CDR, i.e., the greater the IgG leakage, the greater the Braak stage and the CDR, as shown in
Supplementary Figures 2A and 2D.

Accumulation of Cyclophilin A and Matrix Metalloproteinase-9 in Pericytes and Endothelium in Alzheimer’s
Disease Apolipoprotein E4 Carriers

To better understand possible molecular mechanisms contributing to accelerated BBB breakdown
in APOE4 carriers, we next studied the relative levels of proteins involved in the LRP1-CypA-MMP-9
pathway previously shown to control the BBB integrity in transgenic APOE3 and APOE4 mice. Our data
show that LRP1 levels in pericytes and brain endothelial cells of the BBB are similarly reduced irrespectively
122

of the APOE genotype in AD APOE3 and AD APOE4 patients compared with controls by approximately
55% in pericytes and 74% in endothelial cells, as shown by triple immunostaining analysis for LRP1,
pericyte marker PDGFRβ, and endothelial-specific lectin, respectively (Figures 3A to 3C). A substantial
reduction in LRP1 endothelial levels found in this study is consistent with a previous report of 75% reduction
in LRP1 brain endothelial level in AD patients compared with non-AD controls determined with Von
Willebrand Factor as an endothelial-specific marker.

Figures 4A to 4C illustrates accumulation of CypA in pericytes and endothelial cells in AD APOE3
carriers compared with non-AD controls by 3.4-fold and 1.4-fold, respectively, and in AD APOE4 carriers
compared with AD APOE3 carriers by approximately 46% and 70%, respectively. We also found an
increase in MMP-9 levels in pericytes and endothelial cells in AD APOE3 carriers compared with non-AD
controls by 32% and 79%, respectively, and in AD APOE4 carriers compared with AD APOE3 carriers by
approximately 82% and 132%, respectively (Figures 5A to 5C). In addition, we found that increased levels
of CypA in pericytes correlate positively with Braak stage and/or CDR, i.e., the greater the CypA levels, the
greater the Braak stage and the CDR, as shown in Supplementary Figures 3A to 3D. However, there was
no statistically significant correlation between MMP-9 levels in vascular cells and Braak stage or CDR (not
shown).
Interestingly, we found positive correlation between loss of pericyte coverage and loss of LRP1, as
well as negative correlations between loss of pericyte coverage and levels of CypA and MMP-9 in pericytes,
i.e., the greater the loss of pericyte coverage, the greater the levels of CypA and MMP-9 (Supplementary
Figure 4). These findings suggest that loss of pericytes could be related to activation of CypA-MMP-9
123

pathway, although the nature of this relationship cannot be revealed by the postmortem tissue analysis and
would require additional experimental studies in transgenic APOE4 mice.

Discussion

In this study we show that AD APOE4 carriers develop accelerated pericyte degeneration and BBB
breakdown compared with AD APOE3 carriers, whereas AD APOE3 carriers show a significantly greater
loss of pericytes compared with non-AD controls, which correlates with the magnitude of BBB breakdown
to the studied plasma proteins fibrin and IgG. These findings support previous studies demonstrating BBB
damage in AD and suggest that APOE genotype influences the degree of BBB damage. The present study
extends, however, previous findings by showing that a greater BBB disruption in AD APOE4 compared with
AD APOE3 carriers could be related at least, in part, to a greater loss of pericyte population that has been
shown to lead to a small vessel disease and a chronic BBB breakdown in murine transgenic models and is
associated with BBB breakdown in human neurodegenerative disorders such as AD and amyotrophic
lateral sclerosis.
At the molecular level, we show increased accumulation of CypA and MMP-9 in pericytes and
endothelial cells in AD APOE4 carriers compared with AD APOE3 carriers, which, in turn, develop greater
levels of CypA and MMP-9 in the BBB vascular cells compared with non-AD controls. In APOE4 transgenic
mice, activation of CypA-MMP-9 pathway in pericytes has been shown to mediate BBB breakdown by
124

degrading the BBB tight junction proteins ZO-1, occluding, and claudin-5 and the basement membrane
protein of the capillary wall, collagen IV, that are all substrates for MMP-9. ApoE3, but not spoE4, effectively
inhibits CypA-MMP-9 pathway in pericytes in vitro and in transgenic APOE3 mice in vivo acting through
LRP1, which leads to LRP1-dependent transcriptional suppression of CypA and a subsequent
transcriptional MMP-9 inhibition, as reported. Compared with apoE3, apoE4 interacts weakly with LRP1 on
vascular cells, which leads to a loss of inhibition of CypA-MMP-9 pathway and its activation over time
causing a progressive age-dependent BBB breakdown as shown in transgenic APOE4 mice.

LRP1 levels are reduced in brain capillaries in AD that is not influenced by APOE genotype as we
show in the present study by demonstrating a comparable loss of LRP1 from brain capillaries in AD APOE3
and AD APOE4 carriers. On the basis of previous findings in transgenic APOE3 and APOE4 mice and
culture vascular cells, one would expect to see that diminished brain capillary LRP1 levels in AD APOE3
patients compared with non-AD APOE3 controls lessens the ability of apoE3 to inhibit CypA-MMP-9
pathway, which, in turn, results in elevated CypA and MMP-9 levels in vascular cells, as shown in this study.
This likely contributes to BBB breakdown in AD APOE3 patients similar to that reported in transgenic
APOE3 mice after LRP1 silencing. The degree of BBB breakdown in AD APOE3 carriers was, however,
less pronounced than in AD APOE4 carriers in spite of comparable LRP1 reductions. As a much weaker
ligand for LRP1 in vascular cells, apoE4 compared with apoE3 is likely much less effective in inhibiting
125

LRP1-dependent CypA-MMP-9 pathway in human AD vasculature with diminished LRP1 further elevating
CypA and MMP-9 causing a greater BBB breakdown, as found in the present study, and reported in
transgenic APOE4 mice. Thus, diminished LRP1 levels at the BBB probably generates a double hit in AD
APOE4 patients with weak ligand-receptor affinity and receptor downregulation, whereas AD APOE3
patients have a single hit and only the receptor downregulation.
Interestingly, we also found that loss of pericyte coverage, extravascular accumulations of plasma-
derived proteins (i.e., IgG) and increased CypA levels, but not MMP-9, in pericytes correlate with Braak
stage and cognitive impairment (i.e., CDR). These data may suggest a possible role of BBB breakdown in
the development of neuropathology and dementia in AD. However, human postmortem studies reflect end-
stage disease and therefore cannot determine the exact time course of the BBB breakdown, CypA and
MMP-9 changes in vascular cells, and/or APOE genotype, and whether these changes precede and/or
contribute to neurodegeneration. Future longitudinal imaging studies in the living human brain in APOE4
carriers and non-carriers are needed to address these important questions. Ideally, such studies should
combine measurements of regional BBB integrity, particularly in the hippocampus, as recently reported,
with serial CSF biomarkers analyses of vascular/BBB and/or other cell-specific injury, and brain connectivity
and structural changes, and neuropsychological testing.
Elevated CSF CypA and active MMP-9 levels, and an increase in albumin CSF/plasma ratio
suggestive of BBB breakdown have been shown in living older cognitively normal APOE4/APOE3 carriers
compared with the corresponding APOE3/APOE3 and APOE2/APOE3 carriers with no cognitive
impairment, suggesting that APOE4 allele compared with APOE3 and/or APOE2 leads to increased BBB
permeability during cognitively normal aging. In addition, a recent study on the basis of postmortem brain
tissue analysis reported that CypA mRNA levels are reduced in APOE2 carriers further supporting the role
of CypA in regulating neurovascular function.
On a technical note, an earlier study reported decreased absolute values for the capillary length
density in the cortex in control subjects. A number of factors might have contributed to discrepant results
with the present study such as to name a few, use of a different molecular marker to determine capillary
length, i.e., endothelial-specific lectin versus the basement membrane protein collagen IV, different
thickness of tissue sections, i.e., 10um versus 50um that can potentially affect differential penetration of
126

primary antibodies into the respective tissue specimens, differences in the Brodman areas studied, different
PMI, i.e., 3 to 10 hours versus 14 to 24 hours, and a different quantification method, i.e., 15 images per
specimen derived from five fields per section in three adjacent sections 100um apart per specimen versus
stereological measurements, in the present versus earlier study, respectively. The source of this
discrepancy and/or the relative contributions of the factors listed above remain unclear. Importantly, the
use of lectin-positive endothelial profiles to determine capillary length as in the present study has been
validated with another endothelial marker CD31 in murine models and has been shown to accurately reflect
the length of perfused cortical capillaries in vivo determined by a fluorescein-conjugated dextran
angiography and multiphoton microscopy in the living murine brain.
Nonetheless, the present study suggests that BBB breakdown in AD APOE4 carriers could be
related to accelerated pericyte degeneration and/or activation of CypA-MMMP-9 pathway. Although, the
relative contributions of pericyte degeneration and activation of CypA-MMP-9 pathway to BBB breakdown
remain elusive at the present time it is possible that activation of MMP-9 accelerated pericyte loss by at
least two ways: (1) by allowing a greater ingress of blood-derived products across the BBB that are taken
up and degraded by pericytes, which, over time, has been shown to lead to pericyte loss in animal models;
and (2) by degrading the extracellular matrix and important cell adhesion molecules around brain capillaries
that can lead to separation of pericytes from the capillary wall and their cell death.
Future longitudinal CSF and BBB imaging bio-markers studies in living human APOE4 and APOE3
carriers with no cognitive impairment and/or with mild cognitive impairment should continue interrogating
the role of BBB breakdown in the pathogenesis of dementia because of AD and other causes, as well as
the role of pericyte loss and the studied molecular biomarkers of BBB breakdown in the living human brain.

REFERENCES:

1. Zlokovic BV. The blood-brain barrier in health and chronic neurodegenerative disorders. Neuron 2008;
57:178–201.

2. Zlokovic BV. Neurovascular pathways to neurodegeneration in Alzheimer’s disease and other
disorders. Nat Rev Neurosci 2011; 12: 723–738.

127

3. Zlokovic B. Cerebrovascular permeability to peptides:manipulations of transport systems at the blood-
brain barrier. Pharm Res 1995; 12: 1395–1406.

4. Zlokovic BV, Lipovac MN, Begley DJ, Davson H and Rakic L. Transport of leucine-enkephalin across
the blood-brain barrier in the perfused guinea pig brain.J Neurochem 1987; 49: 310–315.

5. Zlokovic BV, Begley DJ and Chain-Eliash DG. Blood-brain barrier permeability to leucine-enkephalin,d-
Alanine2-d-leucine5-enkephalin and their N-terminal amino acid (tyrosine). Brain Res 1985; 336: 125–
132.

6. Armulik A, Genove ́ G, Ma ̈e M, Nisancioglu MH, Wallgard E, Niaudet C, et al. Pericytes regulate the
blood-brain barrier. Nature 2010; 468: 557–561.

7. Daneman R, Zhou L, Kebede AA and Barres BA. Pericytes are required for blood-brain barrier integrity
during embryogenesis. Nature 2010; 468: 562–566.

8. Bell RD, Winkler EA, Sagare AP, Singh I, LaRue B, Deane R, et al. Pericytes control key neurovascular
functions and neuronal phenotype in the adult brain and during brain aging. Neuron 2010; 68: 409–427.

9. Winkler EA, Bell RD and Zlokovic BV. Central nervous system pericytes in health and disease. Nat
Neurosci 2011;14: 1398–1405.

10. Farkas E and Luiten PG. Cerebral microvascular pathology in aging and Alzheimer’s disease. Prog
Neurobiol 2001; 64: 575–611.

11. Baloyannis SJ and Baloyannis IS. The vascular factor in Alzheimer’s disease: a study in Golgi
technique and electron microscopy. J Neurol Sci 2012; 322: 117–121.

12. Sengillo JD, Winkler EA, Walker CT, Sullivan JS, Johnson M and Zlokovic BV. Deficiency in mural
vascular cells coincides with blood-brain barrier disruption in Alzheimer’s disease. Brain Pathol 2013; 23:
303–310.

13. Blennow K, Wallin A, Fredman P, Karlsson I, Gottfries CG and Svennerholm L. Blood-brain barrier
disturbance in patients with Alzheimer’s disease is related to vascular factors. Acta Neurol Scand 1990;
81: 323–326.

14. Fiala M, Liu QN, Sayre J, Pop V, Brahmandam V, Graves MC, et al. Cyclooxygenase-2-positive
macrophages infiltrate the Alzheimer’s disease brain and damage the blood-brain barrier. Eur J Clin
Invest 2002;32: 360–371.

15. Salloway S, Gur T, Berzin T, Tavares R, Zipser B, Correia S, et al. Effect of APOE genotype on
microvascular basement membrane in Alzheimer’s disease. J Neurol Sci 2002; 203-204: 1837.

16. Cullen KM, Ko ́csi Z and Stone J. Pericapillary haem-rich deposits: evidence for microhaemorrhages
in aging human cerebral cortex. J Cereb Blood Flow Metab 2005; 25: 1656–1667.

17. Zipser BD, Johanson CE, Gonzalez L, Berzin TM,Tavares R, Hulette CM, et al. Microvascular injury
and blood-brain barrier leakage in Alzheimer’s disease. Neurobiol Aging 2007; 28: 977–986.
18. Goos JD, Kester MI, Barkhof F, Klein M, Blankenstein MA, Scheltens P, et al. Patients with Alzheimer
disease with multiple microbleeds: relation with cerebrospinal fluid biomarkers and cognition fluid
biomarkers and cognition. Stroke 2009; 40: 3455–3460.

19. Ryu JK and McLarnon JG. A leaky blood-brain barrier, fibrinogen infiltration and microglial reactivity in
inflamed Alzheimer’s disease brain. J Cell Mol Med 2009; 13: 2911–2925.

128

20. Farrall AJ and Wardlaw JM. Blood-brain barrier: ageing and microvascular disease—systematic
review and meta-analysis. Neurobiol Aging 2009; 30: 337–352.

21. Bowman GL, Kaye JA and Quinn JF. Dyslipidemia and blood-brain barrier integrity in Alzheimer’s
disease. Curr Gerontol Geriatr Res 2012; 2012: 184042.

22. Hultman K, Strickland S and Norris EH. The APOE e4/e4 genotype potentiates vascular fibrin(ogen)
deposition in amyloid-laden vessels in the brains of Alzheimer’s disease patients. J Cereb Blood Flow
Metab 2013; 33: 1251–1258.

23. Halliday MR, Pomara N, Sagare AP, Mack WJ, Frangione B and Zlokovic BV. Relationship between
cyclophilin a levels and matrix metalloproteinase 9 activity in cerebrospinal fluid of cognitively normal
apolipoprotein e4 carriers and blood-brain barrier breakdown.JAMA Neurol 2013; 70: 1198–1200.

24. Zonneveld HI, Goos JD, Wattjes MP, Prins ND, Scheltens P, van der Flier WM, et al. Prevalence of
cortical superficial siderosis in a memory clinic population. Neurology 2014; 82: 698–704.

25. Winkler EA, Sagare AP and Zlokovic BV. The pericyte: a forgotten cell type with important implica-
tions for Alzheimer’s disease? Brain Pathol 2014; 24:371–386.

26. Kim J, Basak JM and Holtzman DM. The role of apolipoprotein E in Alzheimer’s disease. Neuron
2009; 63: 287–303.

27. Verghese PB, Castellano JM and Holtzman DM. Apolipoprotein E in Alzheimer’s disease and other
neurological disorders. Lancet Neurol 2011; 10: 241–252.

28. Nishitsuji K, Hosono T, Nakamura T, Bu G and Michikawa M. Apolipoprotein E regulates the integrity
of tight junctions in an isoform-dependent manner in an in vitro blood-brain barrier model. J Biol Chem
2011; 286:17536–17542.

29. Bell RD, Winkler EA, Singh I, Sagare AP, Deane R,Wu Z, et al. Apolipoprotein E controls
cerebrovascular integrity via cyclophilin A. Nature 2012; 485: 512–516.

30. Alata W, Ye Y, St-Amour I, Vandal M and Calon F. Human apolipoprotein E e4 expression impairs
cerebral vascularization and blood-brain barrier function in mice. J Cereb Blood Flow Metab 2015; 35:
86–94.

31. Winkler EA, Sengillo JD, Sullivan JS, Henkel JS, Appel SH and Zlokovic BV. Blood-spinal cord barrier
breakdown and pericyte reductions in amyotrophic lateral sclerosis. Acta Neuropathol 2013; 125: 111–
120.

32. Sagare AP, Bell RD, Zhao Z, Ma Q, Winkler EA, Ramanathan A, et al. Pericyte loss influences
Alzheimer-like neurodegeneration in mice. Nat Commun 2013; 4: 2932.

33. Park L, Zhou J, Zhou P, Pistick R, El Jamal S, Younkin L, et al. Innate immunity receptor CD36
promotes cerebral amyloid angiopathy. Proc Natl Acad Sci USA 2013; 110: 3089–3094.

34. Deane R, Wu Z, Sagare A, Davis J, Du Yan S, Hamm K, et al. LRP/amyloid beta-peptide interaction
mediates differential brain efflux of Abeta isoforms. Neuron 2004; 43:333–344.
35. Deane R, Sagare A, Hamm K, Parisi M, Lane S, Finn MB, et al. apoE isoform-specific disruption of
amyloid beta peptide clearance from mouse brain. J Clin Invest 2008; 118: 4002–4013.

36. Shibata M, Yamada S, Kumar SR, Calero M, Bading J, Frangione B, et al. Clearance of Alzheimer’s
amyloid-ss(1-40) peptide from brain by LDL receptor-related protein-1 at the blood-brain barrier. J Clin
Invest 2000; 106:1489–1499.

129

37. Montagne A, Barnes SR, Sweeney MD, Halliday MR, Sagare AP and Zhao Z. Blood-brain barrier
breakdown in the aging human hippocampus. Neuron 2015; 85: 296–302.

38. Montagne A, Pa J and Zlokovic BV. Vascular plasticity and cognition during normal aging and
dementia. JAMA Neurol 2015; (in press).

39. Conejero-Goldberg C, Gomar JJ, Bobes-Bascaran T, Hyde TM, Kleinman JE, Herman MM, et al.
APOE2 enhances neuroprotection against Alzheimer’s disease through multiple molecular mechanisms.
Mol Psychiatry 2014; 19: 1243–1250.

40. Kreczmanski P, Schmidt-Kastner R, Heinsen H, Steinbusch HW, Hof PR and Schmitz C.
Stereological studies of capillary length density in the frontal cortex of schizophrenics. Acta Neuropathol
2005; 109: 510–518.

Abstract (if available)

Abstract Exercise and other forms of physical activity lead to the activation of specific motor and cognitive circuits within the mammalian brain. These activated neuronal circuits are subjected to increased metabolic demand and must respond to transient but significant reduction in available oxygen. The transcription factor hypoxia-inducible factor 1-α is a regulatory mediator of a wide spectrum of genes involved in metabolism, synaptogenesis, and blood flow. The purpose of this thesis was to explore the potential relationship between exercise in the form of running on a motorized treadmill and the activation of genes involved in exercise-dependent neuroplasticity. to begin to elucidate the underlying molecular mechanisms involved. Mice were subjected to treadmill exercise and striatal tissues were analyzed using a commercial microarray designed to identify transcripts whose expression is altered by exposure to hypoxia, a condition occurring in cells under a high metabolic demand. Several candidate genes were identified, and a subset involved in metabolism and angiogenesis were selected to elucidate their temporal and regional patterns of expression with exercise. Transcript analysis included Hif1a (hypoxia-inducible factor 1-alpha), Ldha (lactate dehydrogenase A), Slc2a1 (glucose transporter 1), Slc16a1 (monocarboxylate transporter 1), Slc16a7 (monocarboxylate transporter 2), and Vegfa (vascular endothelial growth factor A). Overall these results indicate that several genes involved in the elevated metabolic response with exercise are consistent with increased expression of Hif1a suggesting a regulatory role for Hif1a in exercise-enhanced neuroplasticity. Furthermore, these increases in gene expression appear regionally specific, occurring in brain regions previously shown to be sites for increased cerebral blood flow activity. Such findings are beginning to lay down a working hypothesis that specific forms of exercise lead to circuit-specific neuronal activation and can identify a potentially novel therapeutic approach to target dysfunctional behaviors sub-served by such circuitry.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Experience-dependent neuroplasticity of the dorsal striatum and prefrontal cortex in the MPTP-lesioned mouse model of Parkinson’s disease

PDF

A novel role for hypoxia-inducible factor-1alpha (HIF-1alpha) in the regulation of inflammatory chemokines and leukotriene expression in sickle cell disease

PDF

Genetic manipulation of receptor interacting protein (RIP140) uncovers its critical role in the regulation of metabolism, gene expression and insulin signaling in skeletal muscle cells

Asset Metadata

Creator Halliday, Matthew Ryan (author)

Core Title Physical activity optimizes circuit-specific cellular metabolism in neuroplasticity: a role for hypoxia-inducible factor-1 and its downstream targets

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Neuroscience

Publication Date 02/20/2019

Defense Date 01/17/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag angiogenesis,Exercise,gene expression,hypoxia,metabolism,microarray,neuroplasticity,OAI-PMH Harvest,Parkinson's disease,PCR,physical activity,striatum,synaptogenesis

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Holschneider, Daniel (committee chair), Jakowec, Michael (committee member), Petzinger, Giselle (committee member), Walsh, John (committee member)

Creator Email matthalliday2@gmail.com,mhallida@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-125117

Unique identifier UC11675343

Identifier etd-HallidayMa-7098.pdf (filename),usctheses-c89-125117 (legacy record id)

Legacy Identifier etd-HallidayMa-7098.pdf

Dmrecord 125117

Document Type Dissertation

Format application/pdf (imt)

Rights Halliday, Matthew Ryan

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA