Functional Neuroimaging of Recovery from Focal …...ii Functional Neuroimaging of Recovery from...

Functional Neuroimaging of Recovery from Focal

Ischemic Stroke

by

Evelyn MR Lake

A thesis submitted in conformity with the requirements

for the degree of Doctor of Philosophy

Medical Biophysics

University of Toronto

© Copyright by Evelyn MR Lake 2016

ii

Functional Neuroimaging of Recovery from Focal Ischemic

Stroke

Evelyn MR Lake

Doctor of Philosophy

Medical Biophysics University of Toronto

2016

Abstract

Ischemic stroke is the leading global cause of healthy life-years-lost, yet, physiotherapy

remains the only means to improve long-term outcome in the vast majority of patients.

The absence of more effective interventions in the subacute stage reflects uncertainty

surrounding the mechanisms that govern recovery. The present work investigated

endogenous neurovascular adaptation, as well as the delayed neurogliovascular unit

modulation via GABAA antagonism and cyclooxygenase-1 (COX-1) inhibition.

Longitudinal functional and structural magnetic resonance imaging (MRI), intracranial

array electrophysiology, Montoya Staircase testing, and immunofluorescence were

employed to examine subacute functional and structural changes in the peri-infarct zone

in a rodent model of focal ischemia. In the absence of treatment, early subacute stage

was characterized by a persistent skilled reaching deficit and stable lesion volume. Peri-

infarct resting perfusion and vascular reactivity to hypercapnia were elevated a week

post stroke, while the peri-lesional neuronal network was silenced and somatotopy

abolished. By 21 days post-stroke, peri-lesional blood flow resolved to the contra-

iii

lesional level, but peri-lesional vascular reactivity remained elevated. Concomitantly,

neuronal response amplitudes increased with distance from the necrotic core,

suggesting functional remodelling of the lesion periphery, buttressed by increased

spontaneous activity. The peri-infarct showed increased vascular density, neuronal loss,

and astrocytic activation, while microglia and macrophage recruitment was widespread.

GABAA antagonism resulted in progressive improvement in skilled reaching

performance and a decrease in stroke volume, while COX-1 inhibition preserved peri-

lesional hyperperfusion, increased neuronal survival and decreased microglia and

macrophage recruitment. Combined, these studies provide evidence of highly dynamic

functional changes in the peri-infarct zone weeks following ischemic injury, suggesting

an extended temporal window for therapeutic interventions. Delayed pharmacological

modulation of GABAA or COX-1 activity may exert multiple beneficial effects on

neurogliovascular function and hence may be promising treatment targets in the

subacute stage of stroke.

iv

Acknowledgements

For encouragement, and tolerance, I would like to thank my family. In particular,

my husband Taylor who has withstood numerous disappointments when I chose to work

in place of coming home, or forgave me when I failed (again) to predict when a project

could possibly be completed. I would like to thank him for his patience, and for adding

richness to my life outside of the lab. I would also like to give a special 'thank you' to my

father and mother. My Dad for making the pursuit of my studies feel natural and

valuable, and my Mom for letting me know that no-matter-the-outcome she will still be

proud of me.

For guidance, enthusiasm and patience, I would like to thank my supervisor.

Without Bojana, I would have experienced none of the events or opportunities which

have filled my time as a graduate student, and shaped my outlook as a budding

scientist. I would also like to thank my co-supervisor for his uncompromising character

and generosity. I feel fortunate to have had the opportunity to learn from Greg how to

think and express myself as a young researcher, and the value of enjoying the company

of my colleagues. With some luck, a lot of hard work, and careful attention to detail, I

hope I can live up to their expectations (one day).

Lastly, I would like to thank the members of the lab (anyone and everyone) who

has contributed to the overall very positive experience I have had working towards the

completion of this thesis. I have been incredibly fortunate to have worked with a group

of people whom I've enjoyed spending time with on a daily basis.

v

Table of Contents

List of Tables

List of Figures

1 Introduction

1.1 Global stroke burden

1.2 Ischemic stroke burden and treatment in North America

1.3 Neuroprotection and STAIRs

1.4 Mediators of ischemic stroke progression

1.5 Vascular and neuronal network remodelling

1.6 Inflammation

1.7 Endogenous recovery from ischemic stroke

1.8 Functional MRI in chronic ischemic stroke patients

1.9 Preclinical fMRI

1.10 Modelling ischemic stroke in rats

1.11 Behavioural tests

1.12 The present work

1.13 Summary of thesis chapters

List of contributions

2 The effects of delayed reduction of tonic inhibition on ischemic lesion and sensorimotor function

2.1 Introduction

2.2 Methods

2.2.1 Subjects

2.2.2 Stroke induction

2.2.3 Drug treatment

2.2.4 Magnetic Resonance Imaging

2.2.5 Behavioural assessment

2.2.6 Histology

2.2.7 Statistical analysis

2.3 Results

2.3.1 Exclusion criteria

2.3.2 Structural assessment


vi

2.3.4 Histology

2.4 Discussion

2.5 Conclusion

2.6 Supplementary material

3 Neurovascular unit remodelling in the subacute stage of stroke recovery

3.1 Introduction

3.2 Methods

3.2.1 Inclusion/exclusion


3.2.3 Electrophysiological recordings

3.2.4 Immunohistochemistry

3.2.5 Statistical Analysis

3.3 Results

3.3.1 Effect of focal ischemia on forelimb skilled reaching ability

3.3.2 Spontaneous neuronal activity in the ipsi-lesional cortex

3.3.3 Somatotopy in the ipsi-lesional cortex

3.3.4 Stroke volumes in the ipsi-lesional cortex

3.3.5 Resting perfusion and perfusion responses to hypercapnia

3.3.6 Inflammatory, neuronal and vascular changes ipsi-lesionally

3.4 Discussion


3.4.2 Injury induced angiogenesis

3.4.3 Remapping of cortical somatotopy following focal ischemia

3.4.4 Changes in neuronal excitability following focal ischemia

3.4.5 Widespread neuro-inflammation

3.5 Conclusion

4 The effects of delayed COX-1 inhibition on recovery from focal ischemic injury

4.1 Introduction

4.2 Methods

4.2.1 Inclusion/exclusion criteria

4.2.2 FR122047 administration




4.3 Results

4.3.1 Resting perfusion and vascular reactivity to hypercapnia


vii

4.4 Discussion

4.4.1 Peri-lesional hemodynamics post ischemia

4.4.2 Ischemia-induced angiogenesis

4.4.3 FR122047 treatment effects

5 Discussion

5.1 Modulating excitability

5.2 Endogenous neuro/angio-genesis

5.3 Inflammation

5.4 Normalization

5.5 Conclusion References (alphabetical order) Appendix 1 – Continuous Arterial Spin Labelling (CASL) Appendix 2 – Functional MRI in chronic ischemic stroke (Review)

viii

List of Tables

2 The effects of delayed reduction of tonic inhibition on ischemic lesion and sensorimotor function 2.1 Physiological parameters during stroke induction and imaging 2.2 Injury characterization 2.3 Skilled reaching ability 2.4 Animal body weight

3 Neurovascular unit remodelling in the subacute stage of stroke recovery 3.1 Physiological monitoring data 3.2 Inclusion criteria for behaviour and imaging protocols 3.3 Channel-wise ipsi-/contra-lateral spontaneous activity power ratio 3.4 Spontaneous activity power ratios 3.5 Stroke volumes on T2-weighted MRI 3.6 Sections used for each ROI in the immunofluorescence analysis 3.7 Optical density ratios on immunofluorescence

4 The effects of delayed COX-1 inhibition on recovery from focal ischemic injury 4.1 Physiological monitoring 4.2 Number of animals 4.3 ROI volumes 4.4 Cortical hemodynamics

ix

List of Figures

2 The effects of delayed reduction of tonic inhibition on ischemic lesion and sensorimotor function 2.1 MRI data segmentation 2.2 Decrease in the volume of injury with treatment 2.3 Amelioration of skilled reaching deficit with treatment 2.4 Relationship between skilled reaching ability and volume of injury 2.5 H&E staining 2.6 Animal body weight 2.7 T2-weighted MRI in sham stroke with L-655,708 2.8 GFAP and NeuN immunohistochemistry

3 Neurovascular unit remodelling in the subacute stage of stroke recovery 3.1 MINCTools classify results 3.2 Somatosensory evoked potentials 3.3 Montoya staircase performance 3.4 Electrophysiological recordings of spontaneous activity 3.5 Evoked LFP responses in the contralateral hemisphere 3.6 Electrophysiological recordings of evoked responses 3.7 Stroke volumes on T2-weighted MRI 3.8 Resting blood flow and cerebrovascular reactivity to 10% CO2 3.9 GFAP and Iba-1 immunofluorescence 3.10 NeuN and RECA-1 immunofluorescence 3.11 Immunofluorescence and corresponding T2-weighted MRI

4 The effects of delayed COX-1 inhibition on recovery from focal ischemic injury 4.1 Schematic of ALZET-pump implantation 4.2 Resting perfusion 4.3 Blood flow response to hypercapnia 4.4 Iba-1 and GFAP immunohistochemistry 4.5 NeuN and RECA-1 immunohistochemistry

1

Chapter 1

1.1 Global stroke burden

Stroke is the leading cause of adult neurological disability worldwide, with the

majority of patients suffering moderate to severe disabilities, and up to one third

requiring institutionalization [Evenson et al. 2001]. With a lifetime cost between $59,800

and $230,000 USD per patient (estimated in 13 countries including Canada and the

United States), the socio-economic burden on patients, family members, and health

care services is already daunting [Feigin et al. 2003, Caro et al. 2000]. Moreover, it is

estimated that by 2020, demographic and epidemiological changes will result in

stroke/coronary-artery disease emerging as the leading cause of lost healthy life-years

worldwide [Feigin et al. 2003, WHO 2000]. If the hitherto trends continue, in the year

2030, there will be almost 12 million stroke deaths, 70 million new stroke survivors, and

more than 200 million disability-adjusted life-years lost globally [Feigin et al. 2014].

There is a clear need to identify ways through which stroke can be prevented, and

means to improve recovery following injury [Krueger et al. 2015].

1.2 Ischemic stroke burden and treatment in North America

Over 85% of all strokes that occur in Canada are ischemic, arising from an

occlusion in a cerebral vessel by a blood clot [Statistics Canada CANSIM 2012]. The

majority of ischemic stroke patients (approximately 90%) arrive at a care facility beyond

the window opportunity for acute stage recanalization treatment [Harsany et al. 2014].

Furthermore, approximately one third do not benefit [Bhatia et al. 2010, Mazighi et al.

2009, Riedel et al. 2011, Smith et al. 2008]. This leaves most patients with

physiotherapy as the only available means to improve long-term outcome [Mikulik et al.

2015, Hill et al. 2005]. That there are not more effective means to rehabilitate this

patient population largely reflects our uncertainty surrounding the underlying

mechanisms that govern recovery following ischemic injury [Teasell et al. 2014].

2

1.3 Neuroprotection and STAIRs

During the last three decades, the search for ischemic stroke therapeutic agents

has focused on neuroprotection [Moretti et al. 2015]. Some examples include: free

radical scavengers, gamma-aminobutyric acid (GABA) agonists, competitive or non-

competitive N-methyl-D-aspartate (NMDA) antagonists, and growth factors [Wahlgren et

al. 2004]. Unfortunately, despite dozens of clinical trials, no neuroprotective agent has

shown success in clinical trials, which has spurred much discussion among basic

researchers and clinicians as to how translation might be improved [Wahlgren et al.

2004, AIadecola et al. 2011]. Limitations of preclinical stroke modelling has been

proposed as a possible reason for the failure of so many therapeutics in clinical trials

that were successful in preclinical experiments [Wahlgren et al. 2004, AIadecola et al.

2011].

Notably, stroke is closely associated with numerous comorbidities (e.g. diabetes,

hypertension, or obesity), whereas the animals in preclinical trials are typically young,

lean and healthy [Wahlgren et al. 2004, Sutherland et al. 2012]. Additionally, it is

common for preclinical studies to examine only male animals [Wahlgren et al. 2004,

Sutherland et al. 2012]. Furthermore, despite the high frequency of late arrivals in

hospital, treatment is often administered within 1-2 hours following the onset of ischemia

in preclinical models [Wahlgren et al. 2004, Sutherland et al. 2012]. Upon re-

examination, many candidate neuroprotective treatments were administered to patients

much longer after the onset of ischemia than they had been given to animals in

successful preclinical experiments [Wahlgren et al. 2004]. Thus, neuroprotection is most

effective within a short time window, much like recanalization treatments, which greatly

restricts the number of patients who stand to benefit from these strategies. In addition,

many preclinical studies were confounded by drugs which caused hypothermia (which

itself is neuroprotective through several pathways: e.g. reducing free radical formation,

attenuating protein kinase C activity, and slowing cellular metabolism [Sutherland et al.

2012, Yenari et al. 2008, Berger et al. 2002, Globus et al. 1995]). Moreover, it is likely

that hypothermia is effective because so many pathways are affected which promote

cell survival; whereas many neuroprotective treatment strategies affect only one of

many targets in the cell death cascade [AIadecola et al. 2011]. Finally, outcome is most

3

commonly evaluated using tests of gross motor function and quality of life assessments

in clinical practice; whereas stroke volume is the most common outcome metric in

preclinical research [Wahlgren et al. 2004, Sutherland et al. 2012]. In addition,

evaluation of outcome in preclinical models is often early; therefore, delayed cell death

has been mistaken for cell survival [Xu et al. 2013]. Upon re-evaluation, when body

temperature is controlled and therapeutic intervention and evaluation of outcome are

delayed, the majority of neuroprotectants show no preclinical benefit. Therefore, it has

been suggested that to overcome the translational roadblock, preclinical trials that test

novel therapeutics need to adhere to a more strict set of guidelines to better model the

patient population.

To address these shortcomings, The Stroke Therapy Academic Industry

Roundtable (STAIR) working group has provided an initial (1999) and updated (2009)

set of recommendations to improve translation. Some of the recommendations from

STAIRs include: increased sample sizes, randomization, blinding, exclusion reporting,

as well as the noted points discussed above [Howells et al. 2013, Moretti et al. 2015].

Unfortunately, O’Collins et al. (2006) found that less than 1% of drugs reported to be

effective in animal models published between 1990 and 2003 met the STAIRs criteria

[O’Collins et al. 2006]. In addition, the failure to translate neuroprotective agents from

“bench to bedside” may also be impeded by an overly neurocentric approach to stroke

recovery [Turner et al. 2013, Moretti et al. 2015].

1.4 Mediators of ischemic injury progression

While much research has focused on preserving the neuronal cell population

following ischemia, less attention has been paid to re-establishing normal signalling

between neurons, glia, and vessels, which is critical for healthy brain function [Weber et

al. 2006]. Together, these cells are referred to as the neurogliovascular unit, which

serves as a conceptual framework for brain function [del Zoppo et al. 2009. Iadecola et

al. 2004, Lambrechts et al. 2008]. Recovery from ischemic injury depends not only on

the health of each cellular component, but also on the interplay between cell types.

Notably, these interactions change during the stages of injury progression: broadly,

transitioning (in the acute stage) from a predominantly cytotoxic environment, to an

4

environment that promotes repair and remodelling (in later stages) [BIadecola et al.

2011]. The design of effective therapeutic interventions hence depends critically on

understanding inter-cellular interactions within the neurogliovascular unit and their

evolution following injury.

Within minutes of occlusion, hypoxia triggers events which further reduce blood

flow [Atochin et al. 2007, Ishikawa et al. 2004], initiate pro-inflammatory signalling

[BIadecola et al. 2011], and increase the permeability of the blood brain barrier (BBB,

facilitating the infiltration of circulating immune cells and subsequent activation of

resident immune cells [Konsman et al. 2007, Engelhardt et al. 2009]) [BIadecola et al.

2011]. Specifically, increased intra-vascular shear stress generates reactive oxygen

species (ROS) and activates platelets and endothelial cells [Carden et al. 2000,

Peerschke et al. 2010, Pinsky et al. 1996, Eltzschig et al. 2011]. Further, oxidative

stress on endothelial cells reduces the bio-availability of nitric oxide (NO, a vasodilator,

and inhibitor of platelet aggregation and leukocyte adhesion) [Yilmaz et al. 2010]. In

addition, intra-vascular fibrin formation, and constriction of the micro-vascular bed (by

pericytes) increases the likelihood of trapped platelets/leukocytes and causes micro-

occlusions [Yemisci et al. 2009, del Zoppo et al. 1991, Hyman et al. 1991]. In addition,

pro-inflammatory signalling increases the release of cytokines, vaso-active mediators,

and proteases [Konsman et al. 2007, Lindsberg et al. 2010, Strbian et al. 2006 Sairanen

et al. 1997].

When cells begin to die, control of vessel tone is reduced [Burnstock et al. 2013],

and pro-inflammatory signalling increases [Mogensen 2009]. Specifically, neuronal

death disrupts the cell-to-cell interaction between neurons and microglia causing

microglia activation (and loss of immuno-suppression) [Matsumoto et al. 2007, Cardona

et al. 2006]. Once activated, microglia behave like macrophages (adopting ameboid

morphology, assuming migratory capacity, engaging in phagocytosis, and presenting

antigens), and begin to express neurotransmitter receptors [Pocock et al. 2007]. If

neurotransmitters are present (if there are surviving neurons), they down-regulate the

production of cytokines [Pocock et al. 2007], and suppress mast cells [Peachell et al.

2007, Samson et al. 2003]. Furthermore, cell death releases intra-cellular components

[BIadecola et al. 2011]: notably, adenosine/uridine-triphosphate (ATP/UTP), which

activates microglia [Kono et al. 2008, Bours et al. 2006], and promotes platelet

5

aggregation [Bune et al. 2010], and danger associated molecular pattern molecules

(DAMPs) which activate dendritic cells (ie. accessory cells, which process antigen

material and present it to T cells [den Haan et al. 2014]), microglia, peri-vascular

macrophages, and endothelial cells [Marsh et al. 2009]. In addition, lytic enzymes,

which degrade matrix proteins [BIadecola et al. 2011], and promote microglia activation

[Chapman et al. 2000] are released.

With time, these cytotoxic processes give way to pathways that re-establish

tissue homeostasis through the removal of dead cells, resolution of inflammation

[BIadecola et al. 2011], and release of growth factors [Spite et al. 2010, Nathan et al.

2010] which promote neuronal sprouting, neurogenesis, angiogenesis, and gliogenesis.

However, these late-stage processes are not well understood [BIadecola et al. 2011,

Greenberg et al. 2006, Carmichael et al. 2010, Zhang et al. 2002]. Notably, the

resolution of inflammation was previously thought to be a passive process which

resulted from the exhaustion of pro-inflammatory signalling; however, it has since been

shown that there are a large number of mediators involved in actively suppressing

inflammation following ischemic injury [Spite et al. 2010]. Furthermore, some of these

mediators are neuroprotective [Lalancette-Hebert et al. 2007, AIadecola et al. 2011], and

promote angiogenesis [Zhang et al. 2002]; highlighting the interplay between cellular

components within the neurogliovascular unit during recovery.

In summary, acute cytotoxic processes post-stroke eventually give way to, as of

yet still poorly understood mechanisms of neurogliovascular repair and remodelling

(notably, endogenous neuro- and angio-genesis and inflammation). In patients, injury-

induced changes within the neurovascular unit, are likely responsible for some

spontaneous recovery [Yu et al., 2016]. However, for the majority of patients,

endogenous mechanisms of healing are insufficient for a full recovery [Teasell et al.

2014]. Through a better understanding of mechanisms which govern remodelling, novel

treatments accessible to a much broader patient population stand to be realized

[Wiltrout et al. 2007, Liu et al. 2014, Sawada et al. 2014, Ruan et al. 2015, Marlier et al.

2015, Yu et al., 2016].

6

1.5 Vascular and neuronal network remodelling

In the subacute stage after stroke, there is dramatic vascular outgrowth mediated

by the release of pro-angiogenic growth factors (e.g. endothelial nitric oxide synthase,

cystathionine-γ-lyase, vascular endothelial growth factor (VEGF) / VEGF-Receptor-2

(VEGF-R-2) and angiopoietin-1 (Ang-1) / Ang-1 receptor Tie2) [Ruan et al. 2015].

Concomitantly, there is evidence of injury-induced neurogenesis (proliferation,

migration, and maturation) and the release of associated growth factors (e.g. brain

derived neurotrophic factor (BDNF), erythropoietin, insulin-like growth factor-1 (IGF-1),

epidermal growth factor (EGF), fibroblast growth factor-2 (FGF-2), VEGF and

chemokines including stromal cell-derived factor-1 (SDF-1), chemokine receptor 4

(CXCR4), monocyte chemo-attractant protein-1 (MCP-1) and matrix

metalloproteinases), which critically depends on angiogenesis and the micro-

environment created by the neo-vasculature [Sawada et al. 2014].

Specifically, vessel density and length increases during the subacute stage, and

neuroblasts migrate toward the site of injury along new and pre-existing vessels [Thored

et al. 2007, Kojima et al. 2010, Grade et al. 2013]. The release of BDNF, by endothelial

cells, plays a critical role in guiding neuroblast migration [Grade et al. 2013]. In addition,

neuronal progenitor cells (NPCs) proliferate and migrate from subcortical white matter to

peri-lesional tissue where they closely associate with new endothelial cells [Ohab et al.

2006]. NPC migration and residency within peri-lesional tissue depends on Ang1/Tie2

as well as SDF1/CXCR4 endothelial cell production/association [Ohab et al. 2006].

Furthermore, evidence of altered gene expression within the post-stroke brain promotes

NPC proliferation, migration, differentiation and cell survival [cf. Review by Zhang et al.

2015]. For example, the regulation of genes that transcribe: transforming growth factor-

β, signaling pathway genes (e.g. notch 2 and 3), transcription factors (e.g. sox 2,3,9,

and 18, FKHR1 and Hfhbf1), neurotransmitters (e.g. glutamate/neutral amino-acid

transporter protein family members), and angiogenesis promoters (e.g. VEGF, Ang-

2/Tei2 and Ephrin family genes) [Liu et al. 2007] are altered following ischemia.

Manipulating endogenous mechanisms of vascular/neuronal repair/outgrowth has

shown promise in preclinical studies, but treatment optimization and a deeper

understanding of the underlying mechanisms are needed. For example, SDF1β/Ang1

7

treatment improves functional outcome following stroke [Gerlai et al. 2000], however,

without an effect on the number of surviving resident NPCs within peri-lesional tissue

[Ohab et al. 2006]. Thus, Ang1/Tie2 and SDF1/CXCR4 signaling likely influence NPC

migration, but have no effect on cell survival or differentiation [Ohab et al. 2006].

Alternatively, enhancing injury-induced angiogenesis through exercise both before [Ding

et al. 2006] and after [Hu et al. 2010] stroke increases VEGF [Tang et al. 2010], and

tumor necrosis factor (TNF)-α [Ding et al. 2006] and improves outcome [cf. Review by

Wang et al. 2015]. Similarly, intracerebroventricular administration of VEGF reduces

infarct volume, enhances neuronal survival, and increases angiogenesis [Sun et al.

2003]. Furthermore, select pharmacological agents including: metformin [Mao et al.

2013, Zhang et al. 2002, Lee et al. 2004], fluoxetine/simvastatin/ascorbic acid [Corbett

et al. 2015] and rolipram [Li et al. 2011, Hu et al. 2016] enhance angio-/neuro-genesis

and improve functional outcome when administered 24 hours following stroke.

1.6 Inflammation

Concomitant with neuro-/angio-genesis in the subacute stage, inflammation has

a significant effect on recovery following stroke. The inflammatory process begins within

minutes to hours of the onset of ischemia, when endothelial cells are triggered to

express adhesion molecules (e.g. selectins [Goussev et al. 1998, Huang et al. 2000,

Mocco et al. 2002], and intracellular cell adhesion molecule-1 (ICAM-1) [Kanemoto et al.

2002, Vemuganti et al. 2004, Liesz et al. 2011, Li et al. 2016]) which facilitate circulating

immune cell infiltration. Attenuating immune cell infiltration and resident inflammatory

cell (e.g. microglia and astrocytes) activation in the acute phase, reduces lesion volume

and improves functional outcome at later stages [Kim et al. 2014, Becker et al. 2001,

Relton et al. 2001]. Furthermore, attenuating disruption of the BBB by inhibiting

proteases (e.g. matrix matalloproteinase [Lapchak et al. 2000, Yang et al. 2007] and

endogenous tissue plasminogen activator) released by activated microglia [Lapchak et

al. 2000, Yepes et al. 2000, Cinelli et al. 2001]) decreases edema and the likelihood of

hemorrhagic transformation [cf. Review by AJin et al. 2010]. In the subacute stage, a

redox-mediated inflammatory response continues to evolve [Kawabori et al. 2015].

However, the effect of inflammation on outcome in later stages has yet to be fully

8

characterized [Lai et al. 2006, Kim et al. 2014, Kawabori et al. 2015] and includes both

beneficial (e.g. neuroprotection [Lai et al. 2006], necrotic cell debris scavenging, and

removal of neutrophils - which release cytotoxic mediators [Weston et al. 2007, Kim et

al. 2014]) as well as detrimental effects (e.g. oxidative and nitrosative stress,

perpetuation of inflammation and neurodegeneration [cf. Review by Frank-Cannon et al.

2009]).

Treatment strategies which modulate microglial activation (e.g. edaravone - a

free radical scavenger [Zhang et al. 2005], hyperbaric oxygen treatments [Gunther et al.

2005], and broad-spectrum antibiotics (doxycycline and minocycline) [Yrjanheikki et al.

1998 and 1999]) are neuroprotective, reduce infarct volume, improve behavioural

outcome, and reduce astrogliosis [Yrjanheikki et al. 1998 and 1999, Zhang et al. 2005,

Gunther et al. 2005]). However, directly inhibiting the proliferation of microglia increases

infarct volume and neuronal apoptosis [Lalancette-Hebert et al. 2007]. Similarly, in the

subacute stage, astrocytes exhibit both positive and negative effects on outcome (e.g.

express pro-inflammatory mediators, contribute to reactive gliosis, and mediate glial

scar formation - which both prevents axonal in-growth and re-innervation and forms a

barrier between viable and damaged tissue [Nowicka et al. 2008]) [Kim et al. 2014]. In

lieu of directly targeting inflammatory cells (which have biphasic effects), some

subacute treatment strategies have targeted downstream metabolites (e.g. interleukins

(IL), transforming growth factor (TGF)-β, tumor necrosis factor (TNF)-α, and

prostaglandins, PGs), with varied effects [cf. Review by Allan et al. 2001].

For example, targeting some ILs (e.g. IL-4, -6 and 10) [Xiong et al. 2011, Clark et

al. 2000, Herrmann et al. 2003, Jung et al. 2011, Yamashita et al. 2005, Gertz et al.

2012, Strle et al. 2001] has a beneficial effect on outcome whereas targeting others

(e.g. IL-1α, and β) [Pinteaux et al. 2006, Pradillo et al. 2012, Boutin et al. 2001, Basu et

al. 2005, Mulcahy et al. 2003] has a detrimental effect on outcome. Targeting TGF-β or

TNF-α has both positive and negative effects at different stages during injury

progression [Lu et al. 2005, Pang et al. 2001, Lu et al. 2005, Hallenbeck et al. 2002,

Pan et al. 2007]. Similarly, targeting PGs exerts a host of both positive and negative

effects following stroke [cf. Review Yagami et al. 2015]. In addition to chemokines,

cytokines, and PGs, nitric oxide (NO), and reactive oxygen species (ROS) play critical

roles in ischemic injury progression by affecting neuro-inflammation [Kim et al. 2014].

9

Specifically, NO (generated by endothelial, neuronal or inducible synthases – eNOS,

nNOS and iNOS respectively) regulated vascular tone, inhibits platelet aggregation, and

neuronal signalling, and promotes leukocyte adhesion [AIadecola et al. 1995]. Inhibition

of iNOS is neuroprotective [BIadecola et al. 1995, Zhao et al. 2000]. Similarly, ROS

produced by inflammatory cells promote neuronal death and exacerbate injury [Kim et

al. 2014].

In conclusion, the immune response following stroke is long-lasting and involves

a wide range of interacting mediators and metabolites. It is at present not clear how to

modulate the innate immune response so as to promote recovery and repair in the

subacute stage following focal ischemia and no strategies have yet demonstrated

clinical benefit. However, preclinical studies have identified a host of various factors that

appear to be promising treatment targets warranting further investigation.

1.7 Endogenous recovery from ischemic stroke

Most ischemic stroke patients report some degree of improvement with time after

injury indicating that rehabilitation supported endogenous repair during the weeks to

months following a stroke can amount to some recovery [Krishnamurthi et al. 2013,

Hallet et al. 2001, Seil et al. 1997, Steinberg et al. 1997]. Constraint-induced movement

therapy (CIMT) is the most effective in helping patients improve performance [Kwakkel

et al. 2015, Langhorne et al. 2011, Veerbeek et al. 2014]. As many as 25% of patients,

when engaged in CIMT programs, are able to obtain a level of proficiency similar to

healthy subjects [Taub et al. 1993, et al. Miltner et al. 1999, Liepert et al. 2000, Lum et

al. 2002, Fasoli et al. 2003, Fasoli et al. 2004, Taub et al. 2006, Lai et al. 2002].

However, CIMT is not practical for large patient cohorts, and physical gains typically

plateau after 3-4 months [Kwakkel et al. 2016]. Notwithstanding, evidence of recovery in

later stages suggests that the therapeutic window for enhancement of endogenous

healing may be much longer than previously assumed [Hallett et al. 2001, Seil et al.

1997, Steinberg et al. 1997, Lai et al. 2002, Duncan et al. 2000, Krishnamurthi et al.

2013].

In clinical practice, stroke recovery is evaluated using one or more structured

tests which measure impairment using simple physical tasks and quality of life

10

assessments [Gladstone et al. 2002, Salter et al. 2013]. Notably, the most commonly

employed test of impairment is the Fugl-Meyer scale, which can be used alone or in

combination with other tests (e.g. the Chedoke-McMaster Disability Inventory [Gowland

et al. 1995], or the Chedoke Arm and Hand Activity Inventory [Barreca et al. 1999]).

Although tests are carefully designed to evaluate whether a patient is improving, they

are confounded by self-reporting bias, a high incidence of very gradual gains (recovery

over a long period often results in an underestimation of progress), and limited

sensitivity (inability of some tests such as Fugl-Meyer scale to characterize mild

impairments) [Gladstone et al. 2002]. Further, they offer little insight into the underlying

processes driving or impeding recovery [Donovan et al. 2008]. To address this gap,

magnetic resonance imaging (MRI) and functional MRI (fMRI) have been proposed as

sensitive metrics of neuro-anatomical and neuro-physiological changes and thus

correlates of functional recovery.

MRI is a particularly well-suited non-invasive tool for the study of ischemic stroke

recovery as it measures both structure and function. Blood oxygenation level dependent

(BOLD) imaging and arterial spin labelling (ASL) can be used to assess metabolism and

blood flow. For example, ischemia triggers peri-lesional angiogenesis thus changing the

micro-vascular architecture and affecting cerebral blood flow (CBF) and cerebral blood

volume (CBV). Furthermore, the cerebral metabolic rate of oxygen consumption

(CMRO2) is modulated by: neuronal loss [De Girolami et al. 1984], glial recruitment and

activation [AJin et al. 2010], generation of new vessels [Hayward et al. 2011, Lin et al.

2008, Dijkhuizen et al. 2003, Lin et al. 2002] axonal sprouting [Carmichael et al. 2001],

and synaptogenesis [Stroemer et al. 1992, Stroemer et al. 1993, Stroemer et al. 1998].

Other advantages of MRI/fMRI include: the ability to monitor changes throughout the

recovery process with whole brain coverage, and the ready application in both patients

as well as in preclinical experiments, thereby helping to bridge the translational gap

[Hallet et al. 2001, Seil et al. 1997, Steinberg et al. 1997, Lee et al. 1995].

1.8 Functional MRI in chronic ischemic stroke patients

The most frequently applied fMRI technique is BOLD imaging. However, because

the parameters dictating the BOLD signal are differentially affected by ischemic injury

11

progression, the interpretation of BOLD signal changes following ischemic injury is

complex and BOLD neuroimaging in isolation has been challenged to impact clinical

research; thus motivating the implementation of complementary techniques.

Furthermore, pronounced alterations in BOLD signal amplitude and kinetics in both ipsi-

and contra-lesional hemispheres have been found altered whether or not patients

showed a good recovery [Newton et al. 2002, Blicher et al. 2012, Pineiro et al. 2002,

Rossini et al. 2004, Altamura et al. 2007, Roc et al. 2006].

Notwithstanding, delayed, low amplitude, absent, or negative BOLD responses

are most commonly purported to indicate dysfunction. BOLD response attenuation has

thus been proposed to result from compromised local perfusion due to structural

damage to the vasculature [Pineiro et al. 2002], exhausted vasomotor reactivity [Rossini

et al. 2004, Altamura et al. 2007], or absent neuronal activity [Blicher et al. 2012, Roc et

al. 2006, Krainik et al. 2005]. To test these hypotheses, some BOLD fMRI studies have

probed neuronal activity while others have assessed cerebrovascular reactivity [Rossini

et al. 2004, Altamura et al. 2007, Binkofski et al. 2004]. The results are, however,

conflicting, so that more work is needed to probe the mechanisms underlying BOLD

changes following focal ischemia.

For example, employing BOLD fMRI and motor evoked potentials (assessed with

trans-cranial magnetic stimulation, TMS), Binkofski et al. (2004) observed a transient

lack of BOLD activity despite clinical improvement and preserved motor evoked

potentials (TMS). Similarly, Rossini et al. (2004) and Altamura et al. (2007) observed

absent or reduced BOLD activity bilaterally, yet measured stereotypical somatosensory

evoked fields bilaterally on magnetoencephalography (MEG) [Rossini et al. 2004,

Altamura et al. 2007]. Combined, these data provide evidence that at least in well-

recovered patients BOLD fMRI attenuation may not reflect an absence of neuronal

activity.

Impaired cerebrovascular reactivity has been associated with BOLD attenuation

in some, but not all studies. For example, Rossini et al. (2004) measured

cerebrovascular reactivity to hypercapnia with trans-cranial Doppler, but found no

correlation between BOLD response amplitude and the degree of vascular reactivity

[Rossini et al. 2004]. Similarly, Blicher et al. (2012) observed no BOLD signal change,

on average, yet ASL fMRI and VASO-FLAIR (vascular-space-occupancy fluid

12

attenuated inversion recovery) showed task elicited CBF and CBV increases [Blicher et

al. 2012]. On the other hand, Krainik et al. (2005) found that in fully recovered patients

hypocapnia caused a smaller BOLD signal change ipsi- relative to contra-lesionally and

relative to control subjects (bilaterally), while activity-evoked BOLD signal change was

attenuated ipsi-lesionally [Kainik et al. 2005]. Therefore, unlike Rossini et al. (2004) and

Blicher et al. (2012), Krainik and colleagues found impaired cerebrovascular reactivity to

predict BOLD response attenuation [Krainik et al. 2005]. Reports of persistent neuronal

activity, vascular reactivity impairment, and alterations in BOLD responses have

prompted further mechanistic hypotheses of modified neurovascular coupling due to

abnormal vascular anatomy and/or changes in tissue metabolic needs [Blicher et al.

2012, Krainik et al. 2005].

In addition to BOLD response attenuation at or near the lesion site, more

widespread alterations have been observed ipsi- and/or contra-lesionally [Newton et al.

2002, Cramer et al. 1997, Marshall et al. 2000, Ward et al. 2003, Feydy et al. 2002, Cao

et al. 1998]. However, whether abnormal bilateral activation or increased activation

volume signifies recovery [Cramer et al. 1997, Marshall et al. 2000], or is a

manifestation of dysfunction [Ward et al. 2003, Feydy et al. 2002] in the weeks-months

following focal ischemia remains controversial [Newton et al. 2002, Cao et al. 1998,

Buma et al. 2010].

In summary, clinical research has provided evidence of modulation post ischemic

injury that can be measured with fMRI, but findings on the correlation between fMRI

response features and clinical outcome have been inconsistent [Buma et al. 2010].

Mechanistically, there has been some, though still varied, support for fMRI modulation

resulting from alterations in cerebrovascular reactivity and neurovascular coupling,

though details of these mechanisms remain uncertain. A detailed mechanistic

understanding of the determinants of fMRI contrast post ischemic injury necessitates

measurements of changes in the morphology and function of specific cellular

populations and hence calls for multi-modal studies in preclinical models.

13

1.9 Preclinical fMRI

Preclinical studies offer well-controlled experimental conditions including tightly

regulated ischemic injury parameters as well as carefully controlled comorbidities, and

thus provide a powerful means to examine the neurobiological determinants of stroke

recovery and establish a quantitative model of fMRI contrast post ischemic injury.

However, the translational potential of preclinical findings critically depends on careful

experimental design and thoughtful consideration of salient inter-species differences. Of

note, there are structural, functional and metabolic differences between humans and

animals (the discussion here is limited to rats) in addition to differences in preclinical vs.

clinical fMRI implementation.

The higher nominal spatial resolution used in preclinical fMRI (e.g. 0.32 mm3 as

in Weber et al. (2008)) vs. clinical fMRI (e.g. 27 mm3) research is to be evaluated in the

light of a three orders of magnitude smaller rat vs. human total brain volume [Weber et

al. 2008, Kirilina et al. 2015]. A preclinical fMRI voxel thus represents an approximately

100x larger relative brain volume than that comprised in a clinical fMRI voxel.

Additionally, cortical neuronal density in rats is approximately twice that of humans,

whereas the ratio of glia to neurons is lower in rats than in humans [De Felipe et al.

2002, Nedergaard et al. 2003]. Rodents have much less white matter than humans:

white to grey matter ratio is 60:40 in humans vs. 14:86 in rats [Kraft et al. 2012].

However, fundamental cortical communication processes (e.g., action potentials, pre-

and post-synaptic potentials, glial neurotransmitter and K+ clearance, etc.) have been

estimated to have approximately the same relative energy costs in rats and humans

[Hyder et al. 2013]. Finally, the sensitivity of fMRI to magnetic susceptibility gradients

limits the invasiveness of manipulations that can be performed on the animals and

hence the spectrum of complementary techniques that can be done in combination with

fMRI.

1.10 Modelling ischemic stroke in rats

Much work has gone into the development of models of focal ischemia in the rat

brain that result in focal lesions and robust, persistent sensorimotor deficits. These

14

factors, combined with the ease of maintenance, short breeding cycles, diminished

ethical controversy relative to higher order mammals, and a large body of data on

functional hyperemia in the absence of injury, make the rat a particularly attractive

model in which to conduct studies of the pathophysiological mechanisms over the

course of recovery from ischemic injury.

The most commonly used model of focal ischemia employed in neuroimaging

studies is middle cerebral artery (MCA) occlusion. To occlude the MCA necessitates a

midline incision of the neck to expose the carotid bifurcation and cutting the external

carotid artery. The resulting damage (from the surgical preparation alone) affects the

temporal, lingual, and pharyngeal musculature, which impairs mastication and

swallowing causing post-surgical weight loss, poor motor performance, and dehydration

in more than 40% of animals [Dittmar et al. 2003]. In addition, the infarct that results

from MCA occlusion occupies much of the ipsilesional hemisphere (35-65%) and often

causes hypothalamic injury (which rarely occurs in human stroke), and complicates the

interpretation of outcome due to effects damage to this region has on thermo-regulation

and motivation [Lynch et al. 1999, AGerriets et al. 2004]. Furthermore, only ~10% of all

stroke injuries seen in patients affect in excess of 40% of the ipsilesional hemisphere

[Carmichael et al. 2005]: these strokes are characterized by progressive edema, arterial

compression, and infarct expansion [Hacke et al. 1996, Carmichael et al. 2005]. In

addition, treatment of patients with malignant infarcts is largely ineffective: more than

80% die, and those that live have severe brain damage and frequently require

craniectomy to avoid herniation [Hacke et al. 1996, Berrouschot et al. 1998, Schwab et

al. 1998, Carmichael et al. 2005]. On the other hand, typical infarct volumes in patients

that are able to participate in population studies and clinical trials are between 4.5 and

14% of the affected hemisphere [Carmichael et al. 2005].

Alternative models to MCAO include distal MCA occlusion (e.g. the 'three-vessel'

model), embolic models (e.g. micro-/macro-spheres, thrombotic clot, or photo-

thrombosis), and endothelin-1 (ET-1) injection [Carmichael et al. 2005]. The three-

vessel occlusion model is one of the more common methods [Carmichael et al. 2005],

whose implementations either permanently or transiently occlude the MCA, left common

carotid artery (CCA), or right CCA [Carmichael et al. 2005]. This model avoids

hypothalamic, hippocampal, and mid-brain damage and therefore does not result in the

15

disruption of thermo-regulation [Yamashita et al. 1997]. Resulting infarct volumes are

smaller (affecting 14-26% of the ipsi-lesional hemisphere) than those in standard MCAO

[Kanemitsu et al. 2002, Lambertsen et al. 2002, Guegan et al. 2000]. Unfortunately,

distal MCA occlusion requires a small craniotomy, separation of parotid gland and

temporalis muscle to access the CCAs and the MCA on the surface of the brain, which

requires extensive surgical expertise [Carmichael et al. 2005].

Examples of less invasive models use emboli to occlude cerebral vessels: either

micro-/macro- spheres, thrombi, or photo-thrombosis [Carmichael et al. 2005]. Micro-

(~50μm in diameter [Miyake et al. 1993, Mayzel-Oreg et al. 2004]) or macro- (300-

400μm in diameter [Gerriets et al. 2003A,B]) spheres injected intra-arterially lodge in

either distal small vessels (micro) or the MCA (macro) causing permanent occlusion

[Carmichael et al. 2005]. The macro-sphere MCA occlusion results in an injury similar to

the standard MCA methodology without the same degree of surgical intervention and

without causing hypothalamic damage [Carmichael et al. 2005]. Micro-spheres are less

predictable and result in multi-focal infarcts distributed throughout the brain causing a

diffuse injury, which can be desirable in some specific areas of ischemic stroke research

(e.g. study of lacunar or covert stroke) [Carmichael et al. 2005]. Thromboembolic

models of ischemia use spontaneously formed clots, an autologous blood clot placed

into the blood stream, or thrombin-induced clots within the MCA [Zhang et al. 1997,

Beech et al. 2001]. In general, these methodologies have poor survival, result in

extremely variable infarct volumes and locations (and thus variable behavioural

outcomes), but are still useful for studying endogenous or drug mediated thrombolysis

[Carmichael et al. 2005].

As a final example of embolism models, photo-thrombosis involves local intra-

vascular photo-oxidation to generate highly circumscribed ischemic cortical lesions

following the injection of a photosensitive dye (e.g. Rose-bengal) [Carmichael et al.

2005]. Irradiation (through intact skull, or a cranial window) of an animal with circulating

photosensitizing dye generates intra-vascular singlet oxygen, focal endothelial damage,

platelet activation and micro-vascular occlusion [Watson et al. 1985, Dietrich et al. 1986,

Dietrich et al. 1987, Que et al. 1999]. Although, photothrombosis can be relatively non-

invasive, and offers the ability to target specific areas (down to individual micro-

vessels), this method causes rapid cell death [Kim et al. 2000, Schroeter et al. 1997,

16

Jander et al. 2007], and generates an ischemic core with little to no penumbra. The

resulting lesion exhibits substantial vasogenic edema [Nagayama et al. 2000, Kim et al.

2000, Hayashi et al. 1999, Katsman et al. 2003], and simultaneous cytotoxic

intracellular edema [Watson et al. 1985, van Bruggen et al. 1992, Lee et al. 1996] which

are not typically seen in stroke patients [Provenzale et al. 2003], but are frequently

observed in traumatic brain injury [Hu et al. 2001].

Endothelin-1 (ET-1), a potent vaso-constrictor, can be applied either directly to

the exposed MCA, injected proximal to the MCA, applied topically to brain tissue, or

injected into brain tissue to target different brain regions. When directly applied to the

MCA, ET-1 causes a dose-dependent CBF reduction in to downstream regions [Reid et

al. 1995]. The persistence of the CBF reduction and resulting lesion volume are dose

dependent [Reid et al. 1995]. When injected intra-cortically, ET-1 progressively reduces

CBF, reaching a maximum within one hour. A slight reperfusion of peri-lesional tissue

begins seven hours later and only after 48 hours does reperfusion reach the ischemic

core [Windle et al. 2006]. In regions with severely compromised perfusion, CBF drops

by ~80% [Windle et al. 2006].

1.11 Behavioural tests

Of key importance for the assessment of clinically relevant functional outcome

post-stroke [Endres et al. 2008], sensitive behavioural tests have been developed to

measure functional deficits resulting from focal ischemia in the rat. The most common

method to evaluate behavioural recovery following ischemia involves gross neurological

test batteries. Common tests include: body positioning, spontaneous activity,

respiration, tremor, activity, startle response, gait, pelvic and tail elevation, and touch

escape [Hunter et al. 2000]. Performing neurological test batteries is generally

straightforward and does not require specialized equipment or facilities; but the scores

are subjective, and deficits often resolve spontaneously within a few days to a week

after stroke [Murphy & Corbett 2009].

Alternative types of behavioural testing include forelimb asymmetry, cylinder

[Jones et al. 1994], beam, and ladder walking tasks [Metz et al. 2002], which measure

animals' foot faults or placement errors [Murphy & Corbett 2009]. Applied following an

17

ischemic injury that predominantly affects one hemisphere, these tests measure the

relative degree of dependence on the unaffected limbs, with recovered or healthy

animals using all limbs equally [Murphy & Corbett 2009]. During the cylinder task, the

animal is placed within a small vertical cylinder that permits video recording from below.

The number of contacts made with either the left or right forelimb (or both) with the

cylinder wall is recorded [Jones et al. 1994]. During the beam and ladder walking tasks,

animals are placed on a narrow beam or ladder that must be traversed (>1 meter) to

reach a darkened goal box [Metz et al. 2002]. Foot faults (slips) are counted as the

animal traverses the beam or ladder [Metz et al. 2002]. In general, deficits measured

with the cylinder task or the beam/ladder walking tasks resolve within a few weeks

following focal ischemia [Murphy & Corbett 2009] depending on injury location and

volume. Although these tests typically show more persistent deficits than the

neurological test batteries, they require specific facilities and equipment, and the scores

are still somewhat subject to experimenter bias.

Skilled reaching tests, such as the Montoya staircase test [Montoya et al. 1991],

evaluate forepaw dexterity, which tends to recover less well than other motor functions

[Murphy & Corbett 2009]. During these tasks, rodents reach for individual pellets (treats)

placed either on a shelf within a test chamber [Whishaw et al. 1993], or on a series of

steps within a test box [Montoya et al. 1991]. Mild food-deprivation motivates pellet

retrieval. Reaching success is scored by recording the number of pellets eaten/dropped

during multiple repeated tests (typically 15 minutes in the test chamber or box). These

tests are quantitative, not subject to experimenter bias, and sensitive to forelimb

impairments that persist for a very long time after stroke (in excess of 5 weeks)

[Biernaskie et al. 2004]. Although skilled reaching tasks require animal training and

specific facilities, it is generally thought that these behavioural tests are the most

rigorous for evaluating forelimb deficits following focal ischemia [Murphy & Corbett

2009].

1.12 The present work

The aim of the present work was to investigate both endogenous neurovascular

remodelling and the amenability of neurovascular modulation by delayed

18

pharmacological interventions. We evaluated changes in neuro-anatomy,

hemodynamics, neuronal excitability, and behaviour in the subacute phase of ischemic

injury progression in an ET-1 model of focal ischemia in rats. To provide a

comprehensive characterization of recovery, we used 7T multi-modal (structural and

functional) MRI, intra-cortical electrophysiology, the Montoya staircase skilled reaching

task, and immunohistochemistry. These assays were acquired in the subacute period

following focal ischemia (7-21 days following stroke): the present findings thus apply to

a therapeutic window that is accessible to a broad population of patients. The findings

highlight the complex relationship between neurophysiological state and behaviour and

provide evidence of highly dynamic endogenous functional changes during the weeks

following injury. Furthermore, we observed beneficial effects of both GABA-antagonism

and COX-1 inhibition, indicating that these pathways may provide promising treatment

targets in the subacute stage of stroke recovery.

1.13 Summary of thesis chapters

Our first study was motivated by recent evidence that, contrary to the long-

standing dogma [Schiene et al. 1996], acute stage hyper-excitability is followed by

abnormally high levels of tonic inhibition, mediated by decreased extra-synaptic GABA-

uptake, which results in neuronal hypo-excitability [Clarkson et al. 2010]. Clarkson et al.

(2010) show that delayed L-655,708 (a GABAA-receptor inverse agonist) treatment

(administered three days post-stroke) improves sensorimotor function in a mouse

photothrombotic model. We evaluated L-655,708 treatment in ET-1 injected rats and

utilized the more sensitive Montoya skilled reaching task and T2-weighted MRI. In the

present work, treatment-induced reduction in volumes of necrotic core and peri-lesional

tissue and improved reaching ability were observed. Thereby providing further evidence

that treatment with L-655,708 in the subacute stage may ameliorate some of the effects

of focal ischemia induced injury. Modulating neuronal excitability to increase plasticity

and enhance learning and memory is emerging as a promising strategy for improving

outcome following stroke at later time points [cf. Review by Carmichael 2011]. One of

the ground breaking works in this endeavour is the aforementioned study by Clarkson et

al (2010). As discussed above, a drawback of the photothrombotic model (employed by

19

Clarkson and colleagues) is that it produces a necrotic core with limited peri-lesional

tissue (where much remodelling occurs following injury). In the present work, we elected

the ET-1 model, which produces an injury composed of a relatively large peri-infarct

zone and small necrotic core to increase the potential for remodelling and recovery. In

addition, ET-1 injection has been shown to induce flow reduction and reperfusion

kinetics similar to those observed in human stroke [Olsen & Lassen 1984, Mohr et al.

1986, Heiss et al. 2000, Biernaskie et al. 2001, Carmichael 2005]. The translational

importance of protracted reperfusion vs. rapid flow changes in transient MCAO has

been emphasized in a recent review (cf. Review by Hossmann 2012) and is an

important advantage of the ET-1 model. Furthermore, we employed the Montoya

staircase task, which provides a clinically relevant measurement of performance (upper

limb impairment being the most common motor deficit in patients) [Kleim et al. 2007],

and is purported to be particularly resistant to spontaneous recovery [Murphy & Corbett

2009] which underscores the relevance of the present findings. In summary, we were

careful to choose a model of injury, therapeutic window, neuroimaging technique and

behavioural assay with which to test L-655,708 treatment that are amenable to

translation.

To further our understanding of changes in the neurovascular unit in the

subacute stage following stroke, we examined injury progression in naïve animals.

Increased resting perfusion and endogenous angiogenesis in this period have been

shown to correlate with behavioural recovery [Hayward et al. 2011, Dijkhuizen et al.

2003, Zhang et al. 2013]. In addition to Montoya Staircase testing and T2-weighted MRI,

resting perfusion, and vascular reactivity to hypercapnia were estimated with ASL.

Further, cross-sectional intra-cortical EEG recordings of spontaneous activity and

responses to bilateral forepaw stimulation were collected. Histological analysis was

performed following imaging. Stroke produced a stable ischemic lesion volume, and a

persistent impairment of forelimb skilled reaching ability. At 7 days, resting perfusion as

well as vascular reactivity were approximately twice that of sham-operated animals peri-

lesionally. By 21 days, hyperperfusion pseudo-normalized to contra-lesional levels;

while vascular reactivity remained elevated. EEG recordings showed that the

topological pattern of LFP response amplitudes to bilateral forepaw stimulation were

abolished peri-lesionally at 7 days. By three weeks, LFP response amplitudes

20

progressively increased with distance from the site of ET-1 injection. In parallel, the

average spontaneous neuronal activity in ischemic animals progressed from being

depressed ipsi-lesionally (by approximately 50%) to being twice that observed contra-

lesionally. On immunohistochemical analysis, blood vessel endothelial cell density was

increased peri-lesionally; while neuronal loss and inflammation were widespread. On

the whole, this study provided evidence of highly dynamic endogenous processes of

structural and functional remodelling within both the vascular and neuronal networks

during the weeks following focal ischemia. In general, the evolution of endogenous

changes within the neurovascular unit during the subacute stage has been

understudied. Further, the overwhelming majority of neuroimaging studies have been

conducted in permanent or transient MCAO models. This is significant because these

models cause extensive damage (affecting >40% of the ipsi-lesional hemisphere [Hacke

et al. 1996, Carmichael 2005]) which is typically fatal in patients (>80% die shortly

following stroke [Hacke et al. 1996, Carmichael 2005, Berrouschot et al. 1998, Schwab

et al. 1998]). Furthermore, in the case of transient occlusion models, the prompt

recirculation of the ischemic territory has been shown to cause a markedly different

pathophysiology than observed in clinical stroke and as such has been called to be

eliminated from preclinical stroke research [Hossmann 2012]. The present findings (in

the ET-1 model) are particularly significant as they present a unique multi-modal

neuroimaging characterization of ischemic injury progression in a clinically relevant

model of injury.

Having characterized endogenous neurovascular remodelling, we next attempted

to enhance it through selective COX-1 inhibition, so as to attenuate the subacute peri-

lesional neuro-inflammation observed in the second study. While most work on post-

stroke anti-inflammatories to date has focused on COX-2 inhibition, its detrimental

clinical side effects [cf. Reviews Spite et al. 2010, Yagami et al. 2015], have prompted a

recent shift toward downstream arachidonic acid metabolites and COX-1 [Yagami et al.

2015, Rainsford et al. 2007, Aid et al. 2011, Liedtke et al. 2012, Perrone et al. 2010], but

the latter had not been hitherto examined in the subacute stage of focal ischemia in

vivo. Following 12 days of COX-1 inhibition begun a week post stroke, we observed

sustained peri-lesional resting perfusion elevation and a trend toward normalization of

hypercapnia elicited peri-lesional perfusion increases. Immunohistochemical analysis

21

revealed that treated rats exhibited increased neuronal survival, and attenuated peri-

lesional inflammation. These findings suggest that COX-1 inhibition in the subacute

stage may ameliorate some of the effects of focal ischemia induced injury. Inflammation

plays an as-of-yet incompletely understood role in subacute ischemic injury progression.

To-date, no anti-inflammatory therapeutics have been successfully translated to the

clinic [Yagami et al. 2015]. Again, many previous studies have been confounded by

testing candidate therapeutics at times that are inaccessible in the majority of patients

and by choosing models which do not capture critical characteristics of stroke pathology

[Hossmann 2012, Yagami et al. 2015]. As discussed above, in the present work we

employed a model which better recapitulates key features of stroke pathology, and

tested COX-1 inhibition in the subacute stage which increases the translational potential

of the present findings. Further, FR122047 has to-date not been tested as a candidate

treatment strategy for selective COX-1 inhibition in the subacute stage of ischemic injury

progression in vivo.

In summary, these studies yielded an unprecedented level of detailed in vivo

characterization of neurogliovascular changes in relation to behavioural performance

during endogenous recovery and in the presence of pharmacological interventions

during the subacute stage of ischemic injury progression. Through further study of the

underlying mechanisms and therapeutic strategies investigated in the present work,

greatly needed therapeutic strategies for ischemic stroke patients effective during an

extended time-window post stroke stand to be realized.

22

List of Contributions

Chapter 2: The effects of delayed reduction of tonic inhibition on ischemic lesion and sensorimotor function. Evelyn MR Lake, Joydeep Chaudhuri, Lynsie AM Thomason, Rafal Janik, Milan Ganguly, Dale Corbett, Greg J Stanisz, and Bojana Stefanovic

Published in Journal of Cerebral Blood Flow and Metabolism

Chapter 3: Neurovascular function in the subacute stage of focal cerebral ischemia. Evelyn MR Lake, Paolo Bazzigaluppi, James Mester, Lynsie AM Thomason, Rafal Janik, Mary Brown, JoAnne McLaurin, Peter L Carlen, Dale Corbett, Greg J Stanisz and Bojana Stefanovic

Under review with NeuroImage

Chapter 4: The effects of delayed COX-1 inhibition on recovery from focal ischemic injury. Evelyn MR Lake, James Mester, Lynsie AM Thomason, Conner Adams, Paolo Bazzigaluppi, Margaret Koletar, Rafal Janik, JoAnne McLaurin, Greg J Stanisz and Bojana Stefanovic

Under review with Journal of Magnetic Resonance Imaging Chapter 6: Functional MRI in chronic ischemic stroke (Review). Evelyn MR Lake,

Paolo Bazzigaluppi and Bojana Stefanovic

In press in Philosophical Transactions B

23

Text reproduced from publication in JCBFM

Chapter 2

The Effect of Delayed Reduction of Tonic Inhibition on Ischemic

Lesion and Sensorimotor Function

Abstract

To aid in development of chronic stage treatments for sensorimotor deficits induced by

ischemic stroke, we investigated the effects of GABA antagonism on brain structure and

fine skilled reaching in a rat model of focal ischemia induced via cortical micro-injections

of endothelin-1 (ET-1). Beginning 7 days after stroke, animals were administered a

novel GABAA inverse-agonist, L-655,708, at a dose low enough to afford α5-GABAA

receptor specificity. A week following stroke, the ischemic lesion comprised a small

hypointense necrotic core (6 ± 1 mm3) surrounded by a large (62 ± 11 mm3)

hyperintense peri-lesional region; the skilled reaching ability on the Montoya staircase

test was decreased to 34 ± 2% of the animals’ pre-stroke performance level. Upon L-

655,708 treatment, animals showed a progressive decrease in total stroke volume (13 ±

4 mm3/week, p=0.002), with no change in animals receiving placebo (p>0.3).

Concomitantly, treated animals’ skilled reaching progressively improved, by 9 ± 1% per

week, so that after two weeks of treatment, these animals performed at 65 ± 6% of their

baseline ability, which was 25 ± 11% better than animals given placebo (p=0.04). These

data indicate beneficial effects of delayed, sustained low-dose GABAA antagonism on

neuroanatomical injury and skilled reaching in the chronic stage of stroke recovery in an

ET-1 rat model of focal ischemia.

24

2.1 Introduction

The majority of ischemic stroke patients (as much as 92% in some centres)

arrive at a care facility beyond the window of opportunity (i.e., 3-4.5 hrs after the onset

of symptoms) for acute treatment with tissue plasminogen activator (t-PA), leaving

rehabilitation as the only means to improve outcome [Statistics Canada CANSIM 2012,

Hill et al. 2005]. In the long term, most ischemic stroke patients suffer moderate to

severe disabilities, and up to one third require institutionalization [Evenson et al. 2001].

The need for restorative treatments that may be administered during the chronic stage

(days to weeks following an ischemic event) is thus pressing.

It is now widely believed that the chronic stage recovery involves tissue

remodelling - through the generation of new neurons and glia, axonal sprouting, and/or

synaptogenesis [Witte et al. 2000] - in the peri-infarct zone, a surviving meta-stable

region of “at-risk” tissue that exhibits heterogeneity in both acute stage perfusion deficit

and long-term tissue fate [Wieloch et al. 2006]. Indeed animal studies suggest that the

period to support and/or enhance remapping of sensorimotor function within the peri-

infarct zone may be sufficiently long to enable effective treatment in the sub-acute and

chronic stages [Clarkson et al. 2010, Brown et al. 2009]. Although neurons in the peri-

infarct zone had long been considered hyper-excitable [Schiene et al. 1996], a recent

study in a mouse photothrombic model of focal ischemia by Clarkson et al. (2010)

reports that the initial acute phase of hyper-excitability is followed (on the third day after

photo-thrombosis) by abnormally high levels of tonic inhibition (proposed to be mediated

by decreased extra-synaptic GABA-uptake) resulting in neuronal hypo-excitability

[Clarkson et al. 2010]. Clarkson et al. (2010) show that infusing L-655,708, a GABAA-

receptor inverse agonist, using ALZET-1002 mini-pumps 3 days post-stroke improves

sensorimotor function on the grid-walking and cylinder tasks. By 7 days following stroke,

animals treated with 400 μg/kg/day of L-655,708 have fewer foot-faults on grid walking

task (1.8 times the forelimb foot faults before stroke in the treated group vs. 2.5 times in

the control group) and lower affected-to-unaffected forelimb use difference on cylinder

task (19% discrepancy in the treated group vs. 35% discrepancy in the control group).

Spurred by this report and in keeping with the recommendations of the Stroke

Treatment Academic Industry Roundtable (STAIR) [Khale et al. 2012], we investigated

25

the effects of L-655,708 on brain anatomy and forelimb skilled reaching over three

weeks following focal ischemia induced by cortical micro-injection of endothelin-1 (ET-

1).

We used rats to evaluate the effects of L-655,708 in higher-order species and to

enable the utilization of the Montoya skilled reaching task which has been shown highly

sensitive to sensorimotor injury while exhibiting much less spontaneous recovery than

that seen with measures of spontaneous activity and neurological test batteries [Murphy

& Corbett 2009]. Cortical micro-injection of ET-1 was performed to more faithfully

recapitulate the kinetics of flow impairment observed in human stroke, and to produce a

lesion comprised of a substantial peri-infarct zone surrounding the necrotic core, as

frequently observed in patients, where peri-infarct region comprises up to 35% of the

total volume of injury [Heiss et al. 2000, Windle et al. 2006, Carmichael et al. 2005,

Olsen et al. 1984, Mohr et al. 1986, ABiernaskie et al. 2001]. In contrast to Clarkson et

al. (2010), we implanted the L,655-708 tablets subcutaneously and used a dosing

regimen, following Atack et al. (2006), that results in a steady-state plasma drug

concentration that is low enough to produce selective α5-subunit-containing GABAA-

receptor occupancy during the course of treatment [Atack et al. 2006]. We thereby

assessed drug effects following a clinically relevant delivery method at a dose that

minimizes side-effects by preserving receptor subtype specificity [Atack et al. 2006].

2.2 Methods

2.2.1 Subjects

All experimental procedures in this study were approved by the Animal Care

Committee at the Sunnybrook Research Institute. Thirty-seven adult male Sprague-

Dawley rats (Charles River, Montreal, Canada) weighing 340±50g (mean ± standard

deviation, SD) at the time of stroke induction were included in this study. Animals were

housed in pairs on a 12 hour light/dark cycle. Administration of drug, MR imaging and

behaviour trials were performed during the light phase. Food and water were freely

available except during behavioural test periods (14 consecutive days prior to stroke,

and on days 4-6, 11-13, and 18-20) when food was restricted to 12-15g per day.

26

Throughout the study, the body weight of each animal was maintained at above 90% of

the free-feeding weight.

2.2.2 Stroke induction

All animals underwent the same stroke induction procedure under isoflurane

anesthesia (5% induction and 2-2.5% maintenance). Continuous physiological

monitoring was conducted throughout the surgical procedure (see Table 2.1) to ensure

physiological stability. An ischemic injury was induced via two intra-cortical micro-

injections of ET-1, into the forelimb area of the right primary sensorimotor cortex, as

described in detail below [Windle et al. 2006].

Animals were secured in a small animal stereotaxic apparatus (David KOPF

Instruments). Under aseptic condition, a midline incision was made, and two burr holes

(2mm in diameter) were drilled (relative to Bregma) at 0.0mm AP, -2.5mm ML; and at

2.3mm AP, -2.5mm ML over the right sensorimotor cortex using a high-speed micro-drill

(Foredom Electric Co., Bethel Connecticut) [Windle et al. 2006]. A 10-μl Hamilton

Syringe was used to deliver 800 picomoles of ET-1 suspended in 4μl of phosphate

buffered saline (PBS) at -2.3mm DV: 400 picomoles were delivered in 2μl aliquots

through each burr hole. After lowering the needle to -2.5mm and retracting it to -2.3mm

DV, a one-minute delay was allowed before injection began. A further one-minute delay

was kept between the delivery of each μl, and a two-minute delay preceded needle

retraction. ET-1 was injected at a rate of 1μl per minute, for a total delivery time

(including 4, one-minute delays, and 2, two-minute) of 12 minutes. Burr holes were

closed with bone wax and the scalp was sutured over the skull. For analgesia, animals

were given a subcutaneous dose of Marcaine (0.2mg/kg) suspended in PBS at the

beginning and at the end of surgery.

Table 2.1

Temperature [˚C] Heart Rate [bpm] Breath Rate [bpm] O2 Saturation [%]

Stroke Induction 37.5 ± 0.2 319 ± 20 45 ± 5 98 ± 1

MRI Sessions 37.2 ± 0.5 331 ± 43 68 ± 15 99 ± 1

27

Table 2.1 Physiological Parameters During Stroke Induction and Imaging.

Physiological parameters during stroke induction, with animals on 2-3% isoflurane; and

during MRI, with animals receiving a continuous intravenous infusion of propofol

(45mg/kg/hr).

2.2.3 Drug treatment

All animals underwent daily subcutaneous tablet implantation for two weeks

beginning one week after stroke induction. Animals received tablets containing either

1.5mg of L-655,708 (Tocris Bioscience, Cat. No. 1327) in 58.5mg of high-viscosity

hydroxy-propyl-methyl-cellulose (HV-HPMC) (Sigma-Aldrich, CAS No. 9004-65-3), for

the treated group; or 60mg of HV-HPMC, for the control group [Atack et al. 2006]. We

thereby followed the dosing formulation that affords α5-subunit-containing GABAA-

receptor specificity and clinically relevant dosing kinetics, as described by Atack et al.

(2006), resulting in low, sustained drug concentration in the blood plasma (125-

150ng/ml) and brain tissue (50-60ng/ml) [Atack et al. 2006]. Treatment group allocation

was randomized.

Solid tablets were formulated in-house using a hand press. Animals were

anesthetized with isoflurane (5% induction and 2-2.5% maintenance) for each tablet

implantation procedure, which lasted less than 10 minutes. Under aseptic conditions, an

incision was made in the skin overlying the fat pad on the neck of each animal and a

tablet placed between the fat pad and the skin. The skin was then sutured and a

subcutaneous dose of Marcaine (0.1mg/kg) suspended in PBS administered. During the

course of treatment, three independent incisions were made for pill implantation. Each

incision was used repeatedly by removing the stitches from the previous day and

placing a new tablet within the pocket. To avoid excessive trauma to the tissue, each

pocket was used for only five consecutive pill implantations. Animals were monitored

daily; we observed no instances of infection or evidence of significant irritation.

28


All animals were imaged prior to stroke and at weekly intervals thereafter (on

days 7, 14 and 21) on a 7T Bruker BioSpec system. Animals were briefly anesthetized

with isoflurane (5% induction and 2-2.5% maintenance) for the placement of an

intravenous catheter; isoflurane was discontinued as soon as the catheter was secured.

Once responsive, a bolus (7.5mg/kg) of propofol was administered intravenously and its

continuous infusion (45mg/kg/hr) maintained while T2-weighted structural images were

acquired. Continuous physiological monitoring was conducted throughout all imaging

sessions (see Table 2.1) to ensure physiological stability.

A birdcage body coil was used for signal excitation and a quadrature receive-only

coil for signal reception. T2-weighted structural images were obtained with a rapid

acquisition with relaxation enhancement (RARE) sequence using a RARE factor of 8,

repetition time, TR of 5500ms, echo time, TE of 47ms, and a matrix size of 128x256.

Forty-five 0.5-mm thick coronal slices with a nominal in-plane spatial resolution of

0.1x0.1mm2 were obtained in 12 minutes.

Images were imported into the ImageMagick Display Studio LLC Command-Line

tool for semi-automated segmentation [ImageMagick 2013]. Using predetermined signal

thresholds for two volumes of interest (VOIs) – necrotic core and peri-lesional zone -

serial coronal images were segmented [Kidwell et al. 2003]. Following previous work,

voxels with signal more than two standard deviations (two SDs) above the mean signal

in the corresponding contra-lesional region of interest (ROI) were classified as peri-

lesional tissue; and voxels with signal 2 SDs below the mean signal in the

corresponding contra-lesional ROI were classified as necrotic tissue [Kidwell et al. 2003,

van der Zijden et al. 2008, Neumann-Haefelin et al. 2000, Wegener et al. 2006]. Single

voxel dilation and small hole (<6 voxels in 3D diameter) filling were applied within slices

and between neighbouring slices. The volumes of necrotic core, peri-lesional tissue and

the total stroke volume (comprising necrotic and peri-lesional tissue) were computed at

each time point for each animal.

29


For behavioural assessment, each animal was placed in a staircase test box

which isolates the left from the right forelimb with a central platform. During each trial,

three Noyes precision pellets (45mg, Research Diets Inc, New Brunswick, NJ) were

placed on each of fourteen steps, seven on the right side and seven on the left side of

the box. Animals were allowed 15 minutes within the staircase box to retrieve and eat

pellets. Whenever a pellet was dropped, it fell to the base of the box where it could not

be retrieved. At the conclusion of each trial, the number of pellets successfully eaten

with either right or left forepaw was recorded.

To become proficient at the skilled reaching task, animals were trained for two

trials per day (with a minimum of three hours between trials) for a period of two weeks

prior to stroke induction. Baseline reaching ability was estimated as the average number

of pellets successfully eaten during the final six consecutive trials (i.e. trials performed

on the final three days of the two week training period) [Windle et al. 2006, Montoya et

al. 1991]. Following earlier work [Jeffers et al. 2014], to reach baseline training criteria

(before stroke), animals had to have eaten, on average across the final six training

trials, at least 12 pellets with each forepaw. Three days following stroke, all animals

were tested for a deficit in reaching ability twice daily for three consecutive days (on

days 4, 5 and 6 after stroke) to estimate initial behavioural impairment (before

treatment). Testing was repeated over the course of treatment, on days 11, 12, and 13,

and on days 18, 19 and 20. There was thus ~24 hours between the final behavioural

testing day and the corresponding weekly MRI session. Reaching ability was reported

as the number of pellets successfully eaten on each side during all trials. All behavioural

training, testing and analysis were conducted without knowledge of treatment group

allocation.

2.2.6 Histology

Animals were euthanized by isoflurane overdose. Brains were removed

immediately and placed in 10% buffered formalin. Following full fixation (i.e., after at

least 4 days in formalin), brains were processed using a Leica Biosystems

(Heidelberger, Germany) automated tissue processor and mounted in paraffin wax

30

using a Leica Biosystems embedding station. To sample the volume of injury for

hematoxylin and eosin (H&E) staining, we collected 50 coronal sections, 4-μm thick at

40μm intervals (between 0.0 and 2.3 AP). Stained sections were imaged with an

SCN400 Leica Biosystems scanner and viewed using the Leica Image Viewer.


Linear mixed effect (lme) modelling and ANOVA were used to evaluate the effect

of L-655,708 treatment on MRI regional volumes and forelimb reaching performance

over the course of recovery using the lme function in the nlme package within the R

software [Baaven et al. 2008]. As described by Laird and Ware [Laird et al. 1982], the

lme modelling produces sensible restricted maximum likelihood estimates in the

presence of unbalanced allocation of subjects by factor, as was the case presently. We

modelled each MR volumetric measure (of necrotic core, peri-lesional zone and total

stroke) and skilled reaching performance as linear functions of two fixed effects:

treatment (L-655,708 or placebo) and time, expressed as days after injury. Contrast

within each group across time and between the two groups at two and three weeks after

stroke were considered. Absolute, as well as relative (to volume of injury before

treatment or pre-stroke ability) changes were evaluated. Subjects were treated as

random effects, thus accounting for across-subject variation.

2.3 Results

2.3.1 Exclusion criteria

Of the 37 animals which underwent stroke induction, 9 either died during ET-1

injection or had to be sacrificed because of excessive weight loss, evidence of

persistent dehydration, or loss of the righting reflex within 48 hours following ET-1

injection. Two animals were excluded from all analysis after being identified as outliers

in terms of the stroke volume at day 7 post ET-1 injection: one animal had a stroke

volume ~4 times the mean stroke volume of the cohort, and the other had a stroke

volume ~5 times less than the mean stroke volume of the cohort. Transient RF coil

issues precluded acquisition of MRI data in 5 animals. Finally, 6 animals did not reach

31

the minimum criterion for baseline reaching ability (average of 12 pellets eaten during

the final 6 training sessions) and were thus excluded from the behavioural analysis. In

total, 20 animals (N=10 treated with L-655,708 and N=10 given a placebo) were

included in the structural MRI analysis. Twenty animals (N=11 treated with L-655,708,

with 7 that had MRI; and N=9 given a placebo, with 8 that had MRI) were included in the

behavioural assessment.

2.3.2 Structural assessment

Representative images resulting from the semi-automated segmentation of

structural MRI data are overlaid on the corresponding T2 weighted RARE coronal

images in Figure 2.1. The mean volumes of each region of injury, in the two treatment

groups, are listed before and after treatment in Table 2.2. Seven days after stroke

(before treatment), the average stroke volume, across all animals, of 68 ± 11 mm3 was

comprised of 6 ± 1 mm3 of necrotic core and 62 ± 11 mm3 of peri-lesional tissue, with no

difference in necrotic core (p=0.5), peri-lesional (p=0.4), or total stroke (p=0.4) volume

between L-655,708 and placebo treated animals in either the structural or behavioural

analysis cohorts.

32

Table 2.2

Before Treatment [7 days] Post-Treatment [21 days]

Volume [mm3] All Subjects [N = 20]

L-655,708 [N = 10]

Placebo [N = 10]

L-655,708 [N = 10]

Placebo [N = 10]

Total Stroke 68 ± 11 77 ± 19 58 ± 12 52 ± 18 60 ± 18

Necrotic core 6 ± 1 6 ± 1 5 ± 1 2.1 ± 0.3 4 ± 1

Peri-infarct 62 ± 11 71 ± 19 53 ± 11 50 ± 18 56 ± 18

Skilled Reaching All Subjects [N = 20]

L-655,708 [N = 11]

Placebo [N = 9]

L-655,708 [N = 11]

Placebo [N = 9]

% Pre-stroke ability 34 ± 2 37 ± 3 31 ± 3 62 ± 4 37 ± 3

Table 2.2 Injury characterization. The regional volumes and skilled reaching

performance prior to treatment (i.e., one week following stroke) and after two weeks of

treatment (i.e., three weeks following stroke). The VOIs are quoted for the entire cohort

(N=21, column two); L-655,708 treated animals (N=11, columns three and five); and

animals given placebo (N=10, columns four and six). The total stroke volume comprises

the necrotic core and the peri-lesional tissue. Skilled reaching is expressed in % of pre-

stroke performance. Errors are stated as the SEM.

33

Figure 2.1

Figure 2.1 MRI data segmentation. Coronal T2-weighted MRI slices collected 7 and 14

days (i.e. before and after treatment) for an animal given a placebo (a. and b.) and an

animal treated with L-655,708 (c. and d.) with semi-automated segmentation based

ROIs overlaid on the right. Following previous work, voxels with signal >2 SDs of the

mean signal in the corresponding contra-lesional ROI were classified as peri-lesional

tissue (yellow), and voxels with signal <2 SDs of the mean signal in the corresponding

contra-lesional ROI were classified as necrotic tissue (red) [Jeffers et al. 2014, Baaven

et al. 2008, Laird et al. 19892, Lively et al. 2011, Grotta et al. 2008].

34

Figure 2.2

Figure 2.2 Decrease in the volume of injury with treatment. Mean volumes within

each treatment group and the regression lines corresponding to peri-lesional tissue (a.),

necrotic core (b.), and total stroke volume (c.) vs. time for animals treated with L-

655,708 (teal) and animals given a placebo (red). In animals treated with L-655,708,

each VOI decreased with treatment: peri-lesional (11 ± 3 mm3/week, p=0.007), necrotic

(2.0 ± 0.4 mm3/week, p=1e-5), and total stroke (13 ± 4 mm3/week , p=0.002). Animals

given a placebo exhibited no change in any VOI over time (p>0.3). The slope for peri-

lesional, necrotic, and total stroke volume was different between the two treatment

groups (p = 0.03, 0.01, and 0.01, respectively). The shaded grey area is the 95%

confidence level interval for the prediction of the linear model. The difference in each

VOI between 7 and 21 days following stroke (before and after treatment) are plotted in

d. (peri-lesional), e. (necrotic), and f. (total stroke). The VOI change with treatment was

significantly different between the two groups: this difference amounted to 24 ± 11mm3

in peri-lesional zone (p=0.04); 3.3 ± 1.5mm3 in necrotic core (p=0.04); and 27 ± 13mm3

total stroke (p=0.03) from day 7 to day 21 after stroke.

35

Unlike placebo treated animals, L-655,708 treated animals exhibited a

progressive decrease in the volumes of injury, i.e., the peri-lesional zone, necrotic core,

and total stroke. The cohort-wise volumes of injury at each imaging session are shown

in Figure 2.2 along with the linear regression fits for each of the VOI across time (7-21

days after stroke). The slope of the VOI vs. time regression was different between

treatment groups, with p=0.03 for peri-lesional zone; p=0.01 for necrotic core; and

p=0.01 for total stroke volume. In particular, each injury volume decreased with time in

the L-655,708 treated but not in the placebo animals. The peri-lesional volume changed

at a rate of -10.5 ± 0.5 mm3/week (p=0.007) in the L-655,708 animals vs. 1 ± 4

mm3/week (p=0.7) in the placebo group; the necrotic core changed at a rate of -2.0 ±

0.4 mm3/week (p=1e-5) vs. -0.4 ± 0.4 mm3/week (p=0.3) in the placebo group; and total

stroke volume changed at a rate of -13 ± 4 mm3/week (p=0.002) vs. 1 ± 4 mm3/week

(p=0.8) in the placebo group.

We also contrasted, across the two treatment groups, the changes in each

volume of injury before vs. after treatment. Animals treated with L-655,708 had a

greater decrease in the volume of injury than animals given a placebo. Specifically,

compared to the placebo treated animals, the L-655,708 animals exhibited a 24 ± 11

mm3 greater reduction in peri-lesional volume (p=0.04); a 3.3 ± 1.5 mm3 greater

reduction in necrotic core volume (p=0.04), for a 27 ± 12 mm3 greater decrease in total

stroke volume (p=0.03). These treatment-dependent drops in the volumes of injury are

shown in Figure 2.2 d. (peri-lesional zone), e. (necrotic core), and f. (total stroke).


Before stroke, there was no difference in mean reaching ability between animals

in the L-655,708 treated (N=11) vs. placebo groups (N=9). On average, animals

consumed 15.8 ± 0.3 pellets on the left side during the final 6 training sessions (animals

in the L-655,708 treatment group consumed 15.1 ± 0.4 pellets, while animals in the

placebo group consumed 16.7 ± 0.3 pellets; i.e., they consumed 72 ± 2% and 80 ± 2%

of pellets available on the left side, respectively). Seven days after stroke (prior to

treatment), the mean reaching ability decreased similarly (p=0.8) in the L-655,708 and

placebo groups, to 6 ± 1 pellets (26 ± 2% of available pellets or 37 ± 4% of baseline

ability) and 5 ± 1 pellets (25 ± 3% of available pellets or 31 ± 4% of baseline ability),

36

respectively. The skilled reaching performance levels before and after the treatment are

listed in Table 2.2.

Following earlier work, all behavioural scores were normalized, within each

subject, to the subject’s reaching success at baseline, i.e. prior to stroke. The affected

(left) forepaw performance on the skilled reaching task is plotted as the percent of pre-

stroke ability (±SEM) at 7 days after stroke (prior to treatment) and at 14 and 21 days

after stroke (during and after treatment) for each treatment group in Figure 2.3 a.

Superimposed is the regression of the reaching ability vs. time in each cohort. The rate

of reaching ability improvement was higher in the L-655,708 than in the placebo treated

animals (p=0.03). In particular, the reaching performance improved with time, by 9 ± 1%

per week in the L-655,708 group (p=1e-5) vs. 5 ± 1% per week in the placebo group

(p=0.0006). By 14 days after stroke (after 7 days of treatment), L-655,708 treated

animals consumed 23 ± 10% more pellets than the animals given the placebo (p=0.04),

improving 17 ± 8% more than controls (p=0.05). This difference increased to 25 ± 11%

by 21 days after stroke (following 14 days of treatment) (p=0.04), an improvement 19 ±

7% greater in the treated vs. in the placebo administered animals (p=0.02). Figure 2.3 b.

and 2.3 c. display the skilled reaching improvement in the two groups at 14 and 21 days

following stroke relative to their respective performance at 7 days after stroke.

We next investigated whether behavioural recovery assessed with the Montoya

staircase test related to the cortical changes observed on MRI. Figure 2.4 shows the

regression of skilled reaching performance against peri-lesional (2.4 a.), necrotic (2.4

b.), and total stroke (2.4 c.) volumes for all animals for which both behavioural and MRI

data were available (N=15). Skilled reaching performance decreased with increasing

peri-lesional, necrotic, and total stroke volume. Behavioural performance, normalized to

pre-stroke performance, decreased (p=0.00001) at a rate of 0.24 ± 0.06 per mm3 of

peri-lesional volume, 3.2 ± 0.6 per mm3 of necrotic volume, and 0.24 ± 0.06 per mm3 of

total stroke volume.

37

Figure 2.3

Figure 2.3 Amelioration of skilled reaching deficit with treatment. (a.) The mean

affected (left) forepaw reaching ability normalized subject-wise to the reaching ability

prior to stroke (± SEM) for both groups. For animals treated with L-655,708 (teal), the

slope is 9 ± 1% per week (p=1e-5); and for animals given a placebo (red), the slope is 5

± 1% per week (p=0.0006). The difference in slopes between the two treatment groups

was significant (p=0.03). The shaded grey area is the 95% confidence level interval for

the prediction of the linear model. The improvement within groups was evaluated by

taking the difference between performance at 14 (b.) or 21 (c.) days after stroke and 7

days after stroke. Animals treated with L-655,708 showed greater improvement than

animals given a placebo at 14 days (by 17 ± 8%, p=0.05) and at 21 days (by 19 ± 7%,

p=0.02) after stroke.

38

Figure 2.4

Figure 2.4 Relationship between skilled reaching ability and volume of injury. The

number of pellets eaten during each trial (normalized to performance prior to stroke) are

plotted against the peri-lesional, necrotic, and total stroke volumes for each animal.

From the lme modelling, across all animals, skilled reaching performance decreased

with increasing volume of each region of injury. Reaching performance decreased by

0.24 ± 0.06 % per mm3 of peri-lesional tissue (a.) (p=0.00001), 3.2 ± 0.6 % per mm3 of

necrotic core (b.) (p=0.00001), and 0.24 ± 0.06 % per mm3 of total stroke volume (c.)

(p=1e-5). The shaded grey area is the 95% confidence level interval for the prediction

of the linear model.

39

2.3.4 Histology

In agreement with the literature reports on ischemic stroke modelling via direct

cortical injections of ET-1, H&E staining showed a narrow region devoid of any cells at

the site of the ET-1 injection, surrounded by a region of tissue rarefaction, as illustrated

in Figure 2.5. Glial scar formation in the peri-lesional region was apparent on

immunohistochemical staining with GFAP, Iba-1, and NeuN (cf. Supplementary Figure

2.1) [Lively et al. 2011]. Close visual inspection of H&E sections and T2-weighted MRI

data suggested that hypo-intense ROI (necrotic core) on MRI corresponded to cell-free

region on H&E; whereas the surrounding, hyper-intense ROI exhibited tissue rarefaction

on H&E, increased T2-weighted signal intensity likely resulting from accompanying

edema and inflammation [Lively et al. 2011].

Figure 2.5

Figure 2.5 H&E staining. (a.) One of 45 coronal T2-weighted MR-images collected in

vivo 21 days following stroke induction in an animal given placebo, with the

corresponding inset of the region of injury. The three regions are labelled as healthy

tissue (1), peri-lesional zone (2), and necrotic core (3). (b.) The corresponding H&E

section and inset show differences in morphology of the three regions. Hypo-intense

ROI (necrotic core) on MRI corresponds to cell-free region on H&E; it is surrounded by

a hyper-intense ROI that exhibits tissue rarefaction and putative inflammatory cells

(confirmed by immunohistochemical analysis, cf. Supplementary Figure 2.2.3). The

hypointense region is dense fibrous tissue, which was not preserved in the thin

histological section.

40

2.4 Discussion

In the present study, we evaluated the effects of L-655,708 treatment

administered during the chronic phase of ischemic injury progression in an ET-1 model

of focal ischemia in rats. Sustained, low-dose treatment with L-655,708 progressively

reduced volumes of necrotic core and peri-lesional tissue and improved fine motor

reaching ability. We observed behavioural improvement to be inversely related to the

necrotic, peri-lesional, and total stroke volume across all animals. We have thus shown

that in a higher-order species, using a model of injury that produces more clinically

relevant perfusion kinetics, and a dosing regimen that preserves receptor subunit

specific effects, sustained GABA antagonism starting a week after the ischemic insult

shows potential as a late time point treatment strategy for ischemic stroke.

Notably, the low-dose administered in the present study likely afforded GABAA

alpha-5 receptor subtype specificity within the brain and kept the concentration within

the plasma low enough to avoid off-target effects [Atack et al. 2006]. However, GABAA

receptors are expressed in tissues outside of the central nervous system including

peripheral neurons and non-neuronal cells (e.g., smooth muscle cells on the uterus and

bladder, and by endocrine cells) [Hedblom & Kirkness 1997] and GABAA receptors have

been broadly implicated in cell proliferation [Young & Bordey 2009]. However, the roles

GABAA receptors play within peripheral tissues are presently not well understood

[Young & Bordey 2009].

The translational obstacles encountered with numerous acutely administered

neuro-protective treatment strategies have spurred great research interest in delayed

neuro-restorative treatments [Grotta et al. 2008, National Institute of Neurological

Disease and Stroke 2012]. Of the recent preclinical late time point treatment strategies

(cf. 2014 review by Chen et al. (2014)) shown to promote neurological recovery when

administered in the sub-acute and chronic phases (≥4 hours after injury) in rodent

models [Chen et al. 2014], only a few commence ≥24 hours following ischemia and

show an effect on lesion volume. Hitherto studies in rat models employ erythropoietin

(EPO) 24 hrs following embolic MCAO [Ding et. al 2010]; granulocyte colony-stimulating

factor 24 hrs following ligation of the right MCA and bilateral common carotid arteries

41

[Shyu et. al 2004]; heparin-binding epidermal growth factor-like growth factor 24hrs

following 90 mins of MCAO [Jin et. al 2004]; and vascular endothelial growth factor 24

hrs after 90 minutes of MCAO [Sun et. al 2003]. These treatments reduce lesion volume

by ~20-45% and/or reduce shrinkage of the ipsi-lesional hemisphere as evaluated via

longitudinal MRI [Ding et. al 2010, Shyu et. al 2004] or histological analysis [Jin et. al

2004, Sun et. al 2003]. Mechanistically, the treatments have been implicated to induce

angiogenesis [Shyu et al. 2004, Taguchi et al. 2004], neurogenesis [Shyu et al. 2004,

Taguchi et al. 2004], white matter re-organization [Shyu et al. 2004], recruitment of

autologous hematopoietic stem cells [Jin et al. 2004], and the enhanced survival of

ischemic tissue [Sun et al. 2003, Taguchi et al. 2004]. However, given their early

initiation time, these treatment strategies probably affect downstream steps in cell-death

pathways comprising the acute injury process: the mechanistic changes in the tissue

effected through present treatment, commenced a full week following ischemic insult

are likely distinct [Sun et al. 2003, Taguchi et al. 2004].

In turn, two investigations in murine models that report decreased lesion volume

with treatment begun more than 24 hrs following the onset of ischemia employ human

cord blood-derived CD34+ cells 48 hrs following permanent MCAO [Taguchi et. al

2004]; and erythropoeitin (EPO) 4 or 11 days following ET-1 induced ischemia [Wang

et. al 2012]. These treatments reduce lesion volume by ~30-80% as evaluated using a

cortical thickness index [Taguchi et. al 2004] or histological analysis [Xiong et. al 2014].

Associated neovascularization [Wang et al. 2012], endogenous neurogenesis [Wang et

al. 2012] and attenuation of inflammatory responses [Xiong et al. 2014] have been

suggested therein to elicit neuro-protection and neuro-regeneration: these are the likely

mechanisms also invoked presently upon L-655,708 treatment, though further work on

the elucidation of underlying cellular processes is clearly warranted.

The classification of the volumes of injury into peri-lesional zone and necrotic

core based on T2-weighted MRI signal intensity was buttressed qualitatively via

examination of H&E slices, and immunohistochemical analysis (see Supplementary

Figure 2.1) with T2-weighted hypo-intensity arising from the region devoid of cells on

H&E; and T2-weighted hyper-intensity resulting from edema and inflammation in the

surrounding region [Grotta et al. 2008]. The ~30% reduction in the volume of peri-

lesional T2-weighted hyperintense tissue following 14 days of treatment thus likely

42

results from resolving edema and inflammation and hence the normalization of peri-

lesional T2-values [Pomeroy et al. 2011]. Immunohistochemical analysis will be

incorporated into future work to elucidate the changes in cellular populations underlying

the structural remodelling observed in L-655,708 treated animals.

In light of the specificity and difficulty of the Montoya staircase test, spontaneous

recovery of skilled reaching ability as assessed by this test is limited, underscoring the

significance of the present observation of recovery to ~62% of baseline reaching ability

after two weeks of L-655,708 treatment [Windle et al. 2006]. To the best of our

knowledge, enriched rehabilitation (ER, which combines intensive daily reaching with

the impaired forelimb and an enriched cage environment) has hitherto been the only

late time point therapeutic intervention (used alone or in combination) to result in

sensorimotor improvement on the Montoya staircase test following ischemic injury

[Biernaskie et al. 2001, Biernaskie et al. 2004]. In two studies by Biernaskie et al.

(2001/2004), following stereotaxic micro-injection of ET-1 to the distal portion of the

MCA, recovery to ~50% of pre-stroke ability on the Montoya staircase test was shown

by 40 days after ischemic insult with ER therapy initiated within the first week after

stroke [ABiernaskie et al. 2001, Biernaskie et al. 2004]. Jeffers et al. (2014) in turn

reported that following 2 weeks of epidermal growth factor and EPO infusion into the

ipsi-lateral ventricle and an additional 2 weeks of ER therapy, animals recovered to

~55% of their pre-stroke performance; and further on, at 10 weeks after stroke, reached

~60% of their pre-stroke performance level [Jeffers et al. 2014]. Of note, however,

Biernaskie et al. and Jeffers et al. (2014) observed more severe deficits in skilled

reaching ability one week following ischemic insult (~20 and ~25% of pre-stroke

performance, respectively) than that seen in the present study (with decrease to ~34%

of pre-stroke performance at 7 days post-stroke) [Jeffers et al. 2014]. Future studies will

investigate if further or faster gains in recovery can be attained when L-655,708

treatment is combined with mild ER and how a combined intervention affects structural

remodelling and functional MRI responses.

43

2.5 Conclusion

To the best of our knowledge, the present work is the first to investigate L-

655,708 efficacy in rats, employ a systemic low dose regimen a week following the

ischemic insult, observe progressive decrease in volumes of injury on MRI and an

improvement in skilled reaching ability, and relate volume of injury to behavioural

performance. The results thus provide evidence that treatment with L-655,708 in the

chronic stage of stroke recovery may ameliorate some of the focal ischemia induced

injury in terms of both structural brain damage and skilled reaching deficit. Future work

will investigate if further gains in recovery can be achieved by combining L-655,708 with

enriched rehabilitation and on elucidating the mechanisms underlying structural

remodelling and functional recovery so as to allow treatment optimization.

44

2.6 Supplementary material

Table 2.3 Skilled reaching ability (mean ± SD)

Injection Treatment N Baseline 7 days 14 days 21 days

PBS L-655,708 7 100 ± 10 101 ± 13 105 ± 10 107 ± 11

ET-1 L-655,708 11 100 ± 15 37 ± 27 51 ± 32 62 ± 31

ET-1 Placebo 9 100 ± 13 31 ± 24 28 ± 25 37 ± 25

Table 2.4

Group Starting weight (on day -6)

Growth after injection (day 0-21)

Injection Treatment N Mean SEM SD [g /week] p-value

PBS L-655,708 7 325 7 18 9.8 ± 0.2 < 0.0001

ET-1 L-655,708 5 336 21 47 11.8 ± 0.2 < 0.0001

ET-1 Placebo 6 357 12 29 2.8 ± 0.2 0.1

Table 2.4 Animal body weight. The mean animal weights at the start of baseline

skilled reaching ability evaluation (6 days before ET-1 or PBS injection) are listed in

columns 4-7. Post-surgery rate of change (mean ± SEM) in animal weight from day 0 to

day 21 is shown in column 7, with the p-value on this slope quoted in column 8.

45

Figure 2.6

Figure 2.6 Animal body weight. Body weights (N=18) of PBS injected and treated with

L-655,708 (red); ET-1 injected and treated with L-655,708 (green); ET-1 injected and

given a placebo (blue). Weight was measured daily during the behavioural training and

testing periods, namely: prior to stroke induction (training) on days -6 through -1; prior to

treatment on days 4-6 after surgery; after one week of L-655,708 treatment or placebo

administration on days 11-13; and after two weeks of L-655,708 treatment or placebo

administration on days 18-20. Body weight in grams is shown as mean ± SEM. There

was no dependence of starting weight (Supplementary Table 2.6.2.) on group (p=0.2).

Comparing starting weight between groups: there was neither a difference between L-

655,708 treated groups nor between ET-1 injected groups (p=0.6 and 0.4 respectively),

however, there was a difference between the placebo and PBS injected groups

(p=0.03), with the mean weight of the former being greater. The slope of weight after

ET-1 or PBS injection, was positive in L-655,708 treated groups (ET-1: 11.8 ± 0.2

g/week, and PBS: 9.8 ± 0.2 g/week, p-values <0.00001), while the placebo group

showed a trend (2.8 ± 0.2 g/week, p=0.1). The rate of increase in weight of the placebo

group was different from both L-655,708 treated groups after injection (p-values <

0.00001), while there was no difference between L-655,708 treated groups (p=0.2).

After ET-1 or PBS injection (but before treatment) there was a difference in weight

between the placebo group and both L-655,708 treated groups (p-values=0.02), but no

46

difference between L-655,708 treated groups (p=0.6). After the first week of treatment,

the weight of the placebo group was different from the ET-1 injected L-655,708 treated

group (p=0.05), and showed a trend towards being different from the placebo group

(p=0.09), with still no difference between L-655,708 treated groups (p=0.5). At the final

time point (after two weeks of treatment), there were no differences in weight between

groups (p-values > 0.2). The shaded grey area is the 95% confidence level interval for

the prediction of the linear model.

47

Figure 2.7

Figure 2.7 T2-weighted MRI in sham stroke with L-655,708 treatment. Six images

collected from a representative animal prior to (A and D), 7 days (B and E), and 21 days

(C and F) after sham stroke induction (intra-cortical injection of PBS). Seven days after

sham surgery this animal was treated with L-655,708 for 14 days (i.e. until the study end

point 21 days after sham stroke induction). The first row of images (A, B, and C) were

collected at the level of the anterior injection site (-2.3mm Bregma), and the second row

of images (D, E, and F) were collected at the level of the posterior injection site (0.0mm

Bregma). The black arrows (in B, C, E, and F) indicate the site of PBS injection. Minimal

signal contrast along the needle tract was observed following PBS injection (in B, C, E,

and F). No change in image contrast was observed between 7 and 21 days (before and

after L-655,708 treatment).

48

Figure 2.8

Figure 2.8 GFAP and NeuN Immunohistochemistry. Immunohistochemical results for

the regions of interest in two animals with ET-1 induced injury and corresponding T2-

weighted MR images (a.i. and b.i.). (a.) was administered a placebo and (b.) was

treated with L-655,708; both were sacrificed three weeks after stroke (i.e. after two

weeks of treatment). In a./b.i. the T2-weighted cortical regions of interest (~0.0 Bregma)

are shown for both animals (scale bar 1mm). The immunohistological sections

corresponding to the same ROI as a./b.i. are shown in a./b.ii. (NeuN - red, DAPI - blue)

and a./b.iii. (GFAP - green, DAPI - blue) [scale same as a./b.i.]. Surrounding the

49

necrotic core, devoid of NeuN or GFAP positive cells, is the peri-lesional zone

characterized by low neuronal density (NeuN, red), and astrogliosis (GFAP, green).

Within a./b.i. and a./b.iii. six smaller regions of interest are indicated corresponding to

images a./b.iv. (NeuN - left hemisphere), a./b.v. (NeuN - right hemisphere), a./b.vi.

(GFAP - left hemisphere), and a./b.vii. (GFAP - right hemisphere)

[scale shown in a.iv. 300μm, equivalent for a./b.iv./vii.]

50

Although the work presented in Chapter 2 employed a pharmacological agent to

spur recovery, the overall aim of the treatment was to support the endogenous repair

processes. To expand the set of endogenous targets, we conducted our next set of

studies on treatment-naïve animals. Prior work had demonstrated increased resting

perfusion and angiogenesis in the peri-infarct tissue: these vascular alterations have

been further correlated to spontaneous behavioural recovery on neurological test

batteries and beam walking and cylinder tasks [Hayward et al. 2011, Dijkhuizen et al.

2003, Zhang et al. 2013]. The experiments of Chapter 3 build upon this work and have

furthered our understanding of neurogliovascular changes in the subacute stage of

stroke by examining neuronal activity, hemodynamics, neuro-inflammation, and motor

recovery. In addition to Montoya Staircase testing and T2-weighted MRI, CASL fMRI

data was collected to evaluate resting perfusion, and vascular reactivity to hypercapnia,

a common clinical and preclinical test of cerebral vascular function. Here we present a

multi-modal neuroimaging study in the ET-1 model in place of MCAO which has

dominated previous work. The failure of dozens of treatment strategies which showed

promise in preclinical studies to translate to patients, motivated a closer examination of

stroke modelling. It has been concluded that permanent and transient MCAO models

are inadequate as neither creates an injury akin to what is observed in the vast majority

of patients. Specifically, permanent MCAO results in an injury which is substantially

larger than patients can sustain and survive [Carmichael 2011], and transient MCAO

has been shown to cause an injury that does share the kinetics or cell death pathology

of naturally occurring stroke [Hossmann 2012]. The ET-1 model was chosen for the

present study because it addresses the shortcomings of MCAO models [Olsen &

Lassen 1984, Mohr et al. 1986, Heiss et al. 2000, ABiernaskie et al. 2001, Carmichael

2005] and has not been employed before in a multi-modal neuroimaging study which

examines subacute recovery using the present techniques to interrogate multiple

components of the neurovascular unit.

51

Abbreviated text under review with NeuroImage

Chapter 3

Neurovascular Unit Remodelling in the Subacute Stage of Stroke

Recovery

Abstract

Brain plasticity following focal cerebral ischemia has been observed in both stroke

survivors and in preclinical models of stroke. Endogenous neurovascular adaptation is

at present incompletely understood yet its potentiation may improve long-term functional

outcome. We employed longitudinal MRI, intracranial array electrophysiology, Montoya

Staircase testing, and immunofluorescence to examine function of brain vessels,

neurons, and glia in addition to forelimb skilled reaching during the subacute stage of

ischemic injury progression. Focal ischemic stroke (~100mm3 or ~20% of the total brain

volume) was induced in adult Sprague-Dawley rats via direct injection of endothelin-1

(ET-1) into the right sensori-motor cortex, producing sustained impairment in left

forelimb reaching ability. Resting perfusion and vascular reactivity to hypercapnia in the

peri-lesional cortex were elevated by approximately 60% and 80% respectively seven

days following stroke. At the same time, the normal topological pattern of local field

potential (LFP) responses to peripheral somatosensory stimulation was abolished and

the average power of spontaneous LFP activity attenuated by approximately 50%

relative to the contra-lesional cortex, suggesting initial response attenuation within the

peri-infarct zone. By 21 days after stroke, peri-lesional blood flow resolved, but peri-

lesional vascular reactivity remained elevated. Concomitantly, the LFP response

52

amplitudes increased with distance from the site of ET-1 injection, suggesting functional

remodelling from the core of the lesion to its periphery. This notion was further

buttressed by the lateralization of spontaneous neuronal activity: by day 21, the average

ipsi-lesional power of spontaneous LFP activity was almost twice that of the contra-

lesional cortex. Over the observation period, the peri-lesional cortex exhibited increased

vascular density, along with neuronal loss, astrocytic activation, and recruitment and

activation of microglia and macrophages, with neuronal loss and inflammation extending

beyond the peri-lesional cortex. These findings highlight the complex relationship

between neurophysiological state and behaviour and provide evidence of highly

dynamic functional changes in the peri-infarct zone weeks following the ischemic insult,

suggesting an extended temporal window for therapeutic interventions.

53

3.1 Introduction

Recent work has investigated the determinants of prognosis post stroke and

there is widespread recognition of the importance of understanding the mechanisms of

endogenous recovery as a means of guiding the development of new treatments [Lee &

van Donkelaar 1995, Seil 1997, Steinberg & Augustine 1997, Hallett 2001]. A prominent

topic of current enquiry is injury-induced vascular remodelling. An increase in vascular

density on post mortem neuropathological examination in stroke patients correlates with

functional improvement and longer survival after ischemic injury [Krupiński et al. 1992,

Krupiński et al. 1994, Szpak et al. 1999]. Furthermore, brain perfusion patterns (on

single-photon emission computed tomography imaging) in patients with hemispheric

stroke (in the subacute period, 1-12 days after injury) correlate with outcome, extent and

injury severity [Alexandrov et al. 1996]. The poorest outcomes are seen in patients with

decreased perilesional perfusion, whereas patients with normal or elevated perfusion

show better recovery [Alexandrov et al. 1996].

A number of preclinical studies have reported that histopathological evidence of

angiogenesis correlates with transient in vivo increases in cerebral blood flow (CBF)

and/or cerebral blood volume (CBV) in the peri-lesional tissue during the subacute

phase (one to three weeks following stroke) [Lin et al. 2002, Dijkhuizen et al. 2003, Lin

et al. 2008, Hayward et al. 2011, Martin et al. 2012, Chang et al. 2013]. In preclinical

models, a transition from hypo- (in the initial 24-48 hours) to hyperperfusion is

associated with angiogenesis one to two weeks following injury [Dijkhuizen et al. 2003,

Hayward et al. 2011, Zhang et al. 2013] and improved performance on neurological test

batteries, beam walking and cylinder tests [Lin et al. 2002, Dijkhuizen et al. 2003, Lin et

al. 2008, Haywayrd et al. 2011, Martin et al. 2012, Zhang et al. 2013]. The details of

spatio-temporal changes in neurovascular morphology and function post ischemia

remain uncertain yet are critical for the design of more effective therapeutic approaches

in the subacute stage of stroke.

To address this gap, we injected endothelin-1 (ET-1), a potent vaso-constrictor,

into the forelimb region of the right sensorimotor cortex to produce a robust, focal region

of ischemia. The ET-1 model was chosen as it produces a spatially targeted, focal

http://www.ncbi.nlm.nih.gov/pubmed/?term=Dijkhuizen%20RM%5BAuthor%5D&cauthor=true&cauthor_uid=12533611

http://www.ncbi.nlm.nih.gov/pubmed/?term=Dijkhuizen%20RM%5BAuthor%5D&cauthor=true&cauthor_uid=12533611

54

lesion, similar in volume to what is observed in stroke survivors [Carmichael et al.

2005]); and exhibits flow impairment and resolution kinetics similar to those observed in

human stroke [Olsen & Lassen 1984, Mohr et al. 1986, Heiss et al. 2000, ABiernaskie et

al. 2001, Carmichael 2005]. Brain morphology and vascular function were assessed

using in situ T2-weighted magnetic resonance imaging (MRI) and continuous arterial

spin labelling (CASL) functional MRI; neuronal activity was evaluated using intracortical

array electrophysiology, and neuronal, glial, and vascular morphology were assessed

using pathological analysis. These biological assays were related to skilled reaching

ability on the Montoya Staircase test. Assays were conducted over 21 days following

ischemic insult to examine the evolution of functional and structural changes in the

neurovascular unit.

3.2 Methods

All data processing was performed blinded to surgery group allocation (stroke or

sham) and injury status (seven, or 21 days after stroke or sham-surgery). Please refer

above section 2.2.2 for stroke induction methodology and 2.2.5 for a description of

Montoya staircase testing (please note testing was repeated post-surgery at one and

three weeks in the present study).

3.2.1 Inclusion/exclusion

Ninety-six (N=65 with stroke and N=31 sham-operated; average weight ±

standard error: 383 ± 8g) adult male Sprague-Dawley rats were included in this study.

Animals were housed in pairs on a reversed 12-hour light/dark cycle. MRI, behavioural

trials, and electrophysiological recordings were performed during the dark phase.

Continuous monitoring of breath rate, heart rate, blood O2 saturation, transcutaneous

CO2, end-tidal CO2 and body temperature, was conducted throughout all experiments

necessitating anesthesia (See Table 3.1 for the average physiological levels recorded

under different anesthesia conditions). Inclusion criteria are summarized in Table 3.2 for

each assay.

55

Table 3.1

Anesthesia Breath Rate [breath/min.]

Heart Rate [beat/min.]

Arterial O2 Sat. [%]

Transcutaneous CO2 [mmHg]

Temperature [ºC]

Isoflurane 48 ± 7 330 ± 20 98.5 ± 0.6 NA 37.3 ± 0.5

Propofol 58 ± 24 368 ± 34 98 ± 1 70 ± 11 37 ± 1

α-Chloralose 73 ± 16 384 ± 27 98.8 ± 0.5 58 ± 18 36 ± 3

Table 3.1 Physiological monitoring data. During each experimental protocol where

anesthesia (isoflurance, propofol, and α-chloralose) was given (stroke induction, MRI,

tracheostomization, cranial window preparation, electrophysiological recording) breath

rate (breaths/minute), heart rate (beats/minute), arterial O2 saturation (%),

transcutaneous CO2 (mmHg) and temperature (ºC) were monitored and recorded. The

mean ± SEM are reported.

56

Table 3.2

Control Ischemic

Experiment Criterion Day Total Excluded Total Excluded

Skilled Reaching

Baseline Ability <12 pellets eaten

-4 → -1 9 0 10 1

Handed >20% discrepancy

7 9 0 9 0

Stroke Induction Survival >48 hours 7 31 0 65 8

T2-weighted MRI

Small lesion <15mm³ 7 NA NA 53 14

Cortical Damage 7 31 1 NA NA

Evoked LFP Low SNR LFP < 3 SDs of baseline

7 10 3 5 0

21 10 2 8 0

Spontaneous LFP

Low SNR LFP < 3 SDs of baseline

7 10 4 5 0

21 10 4 8 4

Resting Perfusion

Motion >0.25mm 7 26 3 34 3

21 18 5 25 8

Responses to CO2

Motion >0.25mm 7 4 0 11 3

21 6 0 10 2

SDs: standard deviations

Table 3.2 Exclusion criteria were applied for each metric: skilled reaching ability,

ischemic damage (successful stroke induction or sham-operation), evoked and

spontaneous LFP recordings, resting perfusion and vascular reactivity to hypercapnic

challenges. For both groups, the total number of rats as well as the number of animals

excluded are summarized for each criterion at seven and 21 days. Animals that did not

reach minimum proficiency on the skilled reaching task (retrieving at least 12 pellets,

N=1) or showed evidence of handedness (a discrepancy of pellets eaten between

forelimbs’ >20%, N=0) during the final 4 days of training were excluded. Animals in the

ischemic group whose lesion volume on T2-weighted MRI at day seven was <15mm3

(N=14) were excluded, as well as sham-operated animals with evidence of damage

(N=1). From previous work employing the Montoya Staircase test, a reaching deficit

rarely manifests if lesion volume is <15mm3. Low signal-to-noise-ratio (SNR) in both

57

electrophysiological and MRI acquisition protocols resulted in exclusion. During intra-

cranial electrophysiological recordings low SNR resulted from head-stage amplifier

failure, persistent hypercapnia, or damage during craniotomy. None of the evoked LFP

recordings from stroke animals were excluded. At seven (N=3) and 21 (N=2) days after

sham-surgery, a total of five evoked LFP recordings made in sham-operated animals

were thus excluded. Four spontaneous electrophysiological recordings were excluded

at each time point from the sham group, and (N=4) recordings from stroke animals were

excluded at 21 days. During MRI acquisitions, low SNR resulted from motion exceeding

the nominal in-plane EPI resolution. This resulted in exclusion of resting perfusion

measurements: three from both the sham and stroke groups at seven days, and N=5

from the sham group, and N=8 from the stroke group at 21 days. In addition, N=5 data

sets were excluded from the perfusion responses to hypercapnia data-set: all from the

stroke group, N=3 at seven days, and N=2 at 21 days. In light of varying attrition rates,

the sample sizes yielded sensitivity to a change in reaching ability of 12% (equivalent to

the number placed on an individual step); a 17% difference between evoked response

amplitude in neighbouring electrodes; a 6% difference in ipsi-/contra-lateral power; and

a 10% lateralization in either resting perfusion or perfusion response to hypercapnia.

58


All animals were imaged seven and 21 days following stroke induction (or sham)

surgery on a 7T animal system (Bruker, BioSpec, Etlingen, Germany). Animals were

immobilized with ear bars and an incisor bar, and a stable plane of anesthesia was

maintained with an intravenous infusion of 45mg/kg/hr of propofol (Pharmascience Inc.,

Montreal Quebec, Canada). Please refer to above section 2.2.4 for structural T2-

weighted imaging protocol and lesion segmentation methodology. Propofol anesthesia

was chosen in light of its being well tolerated and allowing rapid recovery, thus being

well suited to longitudinal experiments; in addition to having been successfully used for

cerebral blood flow quantification via ASL in rats [Griffin et al. 2010].

CASL: A custom built labelling coil was positioned at the level of the carotid

arteries and a quadrature receive-only coil was used for signal detection. Using a 1.5-

second adiabatic labelling pulse, and a 0.4-second post-labelling delay, single-average,

single-shot echo-planar images (EPI) were obtained from five 1.5mm thick coronal

slices positioned over the sensorimotor cortex, with a 0.25x0.25mm2 in-plane resolution,

TR/TE of 2000/8.3ms, and inter-slice gap of 0.5mm. Sixty EPI frames were collected to

estimate resting perfusion. For vessel reactivity measurements, animals were

tracheostomized, mechanically ventilated, and challenged by six presentations of a

hypercapnic mixture in ON:OFF periods of 1:4 minutes (ON: 10% CO2, 31% O2 and

59% N2, OFF: 0% CO2, 31% O2 and 69% N2). Inspired mixture composition and delivery

were controlled by a programmable GasMixer (GSM-3, CWE Inc., Boston MA).

CASL data processing: All CASL data were motion corrected (AFNI, Analysis of

Functional NeuroImages [Cox 1996], 2dImReg), masked (to isolate grey matter), and

spatially blurred within the grey matter mask (AFNI 3dBlurToFWHM, full-width-half-

maximum 0.55mm) prior to fitting the data using a Generalized Linear Model (AFNI

3dDeconvolve). Subject-specific hemodynamic response functions were produced by

averaging the signal in the left (contra-lateral) cortical grey matter [Kang et al. 2003]. A

threshold was applied to resulting maps of perfusion signal changes elicited by

hypercapnia and resting perfusion to correct for multiple comparisons (false discovery

rate q<0.01). In each animal, we manually identified a training set of approximately 40-

60 voxels in both perfusion response and resting perfusion maps residing in the contra-

59

and ipsi-lateral cortices. The pial surface and boundary with corpus callosum were

excluded, along with major vessels. Using these training data, classification was

performed with a probabilistic classifier (MINCTools, [Neelin & Fonov], Classify) and

thresholds applied to the resulting probabilistic maps: at >80% a posteriori probability of

belonging to the left cortex (contralateral region of interest, ROI) and at >75% a

posteriori probability of belonging to the affected right cortex (ipsilateral ROI), cf. Figure

3.1. The CASL measurements were then reported in terms of the right-to-left ratio

following earlier work [Heiss et al. 1997, Shen et al. 2011, Hayward et al. 2011,

Wegener et al. 2013], given the presence of nuisance signal variability sources that

affect both hemispheres equally, including inter-subject variability in T1, inversion

efficiency, and global responses to anesthesia. Notwithstanding, to explore the effect of

ischemia on the left (contralateral) sensorimotor cortex hemodynamics, the contralateral

CASL data variation in relation to the average non-lesioned whole brain grey matter

signal were also examined.

3.2.3 Electrophysiological recordings

Recordings were performed either seven or 21 days following stroke or sham-

surgery. Animals were anesthetized with isofluorane (5% induction and 2-2.5%

maintenance) and positioned in a stereotaxic frame (David KOPF Instruments, Tujunga

California, USA). Using a high-speed micro-drill (Foredom Electric Co., Bethel

Connecticut, USA), skull was removed to expose dura, which was carefully resected, to

create two cranial windows over the left and right somatosensory areas (~9.0 x 4.4mm

in size, extending from 2.0mm to -7.0mm in the A-P direction, and from -0.1mm to -

4.5mm in the M-L direction). Two multi-electrode arrays (MEA - MicroProbes,

Gaithersburg Maryland, USA) composed of 16 (2x8) Pl/Ir electrodes (tip diameter:

125μm, impedance: 0.5MΩ, inter-electrode spacing: 250μm, for a total MEA width of

0.5mm and a total length of 2mm) were lowered into the exposed cortices (to 250μm D-

V) for intracortical recordings. The MEA centre was placed at -1.0mm A-P and -2.5mm

M-L, so as to align the most anterior pair of electrodes with the posterior injection site.

After surgery and electrode placement, isoflurane was discontinued and a continuous

infusion of α-chloralose (27mg/kg/hr) begun. Acquisition bandwidth was set to 0.3Hz-

5kHz. Signal was amplified 20x at the head-stage and 50x by the amplifier (Model 3600

60

- AM-systems, Carlsborg Washington, USA). Data were acquired using a 32-channel

SciWorks DataWave Acquisition System (AM-systems, Carlsborg Washington, USA),

with a sampling rate of 20kHz. LFP responses were extracted off-line by filtering the raw

traces with a low-pass filter with a cutoff frequency of 300Hz. Only events identified as a

negative deviation of the field potential larger than three SDs from mean baseline signal

amplitude were used in the analysis (cf. Figure 3.2).

Figure 3.1

Figure 3.1 MINCTools classify results. An example of the ROIs resulting from the

probabalistic map thresholded in the contralatertal (a.) and ipsilateral (b.) hemisphere,

superimposed on the perfusion map (c.). Probabilistic maps were thresholded at >80%

a posteriori probability of belonging to the contralateral ROI and at >75% a posteriori

probability of belonging to the ipsilateral ROI.

61

Figure 3.2

Figure 3.2 Representative somatosensory evoked potentials. Evoked LFP

responses recorded at four sites (a.) at seven (i.) and 21 days (ii.) after stroke.

Contralesional traces are shown in grey, and ipsilesional in orange. Seven days after

stroke, ipsilesional responses were strongly attenuated (a. i.). Twenty-one days after

stroke, at the same distance from Bregma, ipsilesional responses were no longer

obviously attenuated. (b.) Spontaneous activity recordings from two recording sites:

closest to Bregma (top trace) and distal from Bregma (bottom trace), contralesionally

(left column) and ipsilesionally (right column). Seven days post-stroke (b.i.) spontaneous

activity is reduced ipsi- versus contra-lesionally, but there is no obvious ipsilesional

attenuation 21 days post-stroke (b.ii.).

62

Spontaneous Activity: One minute of spontaneous activity was recorded before

beginning bilateral forepaw stimulation. The Fast Fourier Transform was computed for

each two-second interval using a non-over-lapping running window within each channel.

Power was normalized between hemispheres by taking the ratio between corresponding

channels in each hemisphere (e.g., ipsi-lateral channel #1 divided by contra-lateral

channel #1). Following earlier work [Womelsdorf et al. 2014], the channel-wise ipsi- to

contra-lateral power ratios were analyzed for each frequency band of interest: Theta (3-

10Hz), Alpha (8-14Hz), Beta-1 (15-20Hz), Beta-2 (20- 35Hz) and Gamma (35-90Hz).

These bands have been hypothesized to arise from dynamic motifs that are based on

structural circuits and dictate behavioural function, after input-output transformation

[Womelsdorf et al. 2014].

Evoked Activity: LFP recordings were acquired during bilateral forepaw

stimulation (three pulses at 10Hz, 0.5-0.8mA) delivered every 10 seconds and repeated

30 times. The first 10 responses following the presentation of a stimulus and not

preceded (within 100 milliseconds) by a spontaneous event were averaged and

analyzed. The difference was taken between the average responses of neighbouring

channels in a rostro-caudal direction to eliminate signal common to neighbours and

isolate signal from the immediate vicinity of each channel. Finally, signal amplitude was

normalized for each MEA: the amplitude of the average LFP response in each channel

of an MEA was divided by the maximum response amplitude recorded by that MEA.

The normalized signal was plotted as a function of distance from Bregma.


Following the final MRI session, brains were extracted for immunohistochemical

analysis. Animals were euthanized with an over-dose of isoflurane; the brains were

removed from the skull, placed in paraformaldehyde (PFA, Sigma-Aldrich, St. Louis, MI)

and stored at 4ºC. After 24 hours in PFA, brains were rinsed in PBS and transferred to a

15% sucrose solution and subsequently to a 30% sucrose solution (Sigma-Aldrich, St.

Louis, MI). For each primary antibody, six or seven evenly spaced free floating 40μm

coronal sections were collected within the boundaries of the ischemic injury (defined by

hyperintensity on in vivo T2-weighted images) between 3.8mm and -3.5mm from

Bregma in the A-P direction.

63

Sections were singly labelled with primary antibodies, followed with secondary

antibodies conjugated with Alexa 488 fluorescent dye (1/200, Goat Anti-Rabbit A-11008,

Donkey Anti-Mouse A-21202, ThermoFisher Scientific, Calgary, AB). The primary

antibodies used were GFAP (1/500, Z0334, Dako, Burlington, ON), Iba-1 (1/500, 019-

19741, Wako Chemicals, Cape Charles, VA), NeuN (1/200, MAB377, EMD Millipore,

Billerica, MA), and RECA-1 (1/100, MCA970R, AbD Serotec Bio-Rad Co, Raleigh, NC).

Briefly, frozen cryo-protected sections were transferred to 24 well dishes and washed

3x10 minutes with PBS. Sections were blocked for one hour at room temperature in

PBS containing 2-10% Donkey or Goat Serum and 0.1-0.5% Triton X-100. Sections

selected for incubation with Iba-1 antibody were exposed to boiling in 10mM citrate

buffer containing 0.05% Tween20 for antigen retrieval prior to blocking. All primary

antibodies were diluted in the same blocking buffer and incubated overnight at 4ºC, with

the exception of RECA-1 sections, which were incubated at room temperature.

On the following morning, sections were washed 3x10 minutes with PBS and

then incubated for two hours at room temperature with fluorescent conjugated

secondary antibodies diluted in the same blocking buffer. Nuclear counter-stain

NucBlue Reagent (DAPI two drops/ml, R37606, Molecular Probes, ThermoFisher

Scientific, Calgary, AB) was added at the same time as the secondary antibody.

Sections were washed thoroughly in PBS, and mounted on VWR Brand frosted slides

with mounting medium (Polyvinyl alcohol mounting medium with DABCO #10981

Sigma-Aldrich, St. Louis, MI) and glass cover-slips.

All sections were digitized on a Zeiss ApoTome.2 Microscope using

StereoInvestigator (MBF, Biosciences, Williston, VT) and an air 40x objective (with

numerical aperture of 0.95, and working distance of 0.25mm; Zeiss, Oberkochen,

Germany). Each section was registered to the Paxinos & Watson [Paxinos & Watson

2005] digital rat brain atlas and ROIs identified based on anatomical locations where

increases in GFAP/Iba-1 signal were visually apparent. The ROIs thus identified were:

cerebral cortex, caudate putamen, piriform cortex, and olfactory tubercle. The average

signal intensity per unit area for each ROI was computed for both hemispheres in all

sections stained with GFAP or Iba-1. The ipsi- to contra-lateral inter-hemispheric signal

ratio for each ROI was then computed and compared between groups.

64

For NeuN and RECA-1 stained sections, binary images were generated in

ImageJ. A Sobel edge detection filter was first applied to calculate local derivatives

along each imaging axis. The edge-enhanced image was then computed by quadrature

summation of the two derivatives [Hast 2014]. Next, images were blurred with a

Gaussian filter (kernel size of 3 pixels), and the background subtracted (using a sliding

paraboloid of 50 pixels) [Sternberg 1983]. Finally, images were binarized using the

Make Binary algorithm in ImageJ, at a set threshold so that the fractional area occupied

by positively stained cells could be computed. The percent of each ROI area occupied

by positively stained NeuN or RECA-1 cells was computed and the ipsi- to contra-lateral

inter-hemispheric ratio of these values compared between groups.


Please refer to above section 2.2.7 for statistical analysis methodology. Here,

MRI (lesion volume, resting perfusion, reactivity to 10% CO2), intra-cortical

electrophysiological (evoked and spontaneous activity), pathological assays (GFAP,

Iba-1, NeuN, and RECA-1), and skilled reaching performance were modelled as linear

functions of group (stroke or sham) and time after surgery. Contrast within and between

groups at both one and three weeks after injury were considered.

3.3 Results

3.3.1 Effect of focal ischemia on forelimb skilled reaching ability

To evaluate fine motor function before ischemia and during ischemic injury

progression, animals were trained and tested on the Montoya Staircase skilled reaching

task, which provides a sensitive measure of forelimb motor performance [Montoya et al.

1991]. During the 15-minute task, animals retrieved and ate pellets with either the left or

right forelimb. At the conclusion of the task, the number of pellets successfully eaten

with each limb was recorded. Figure 3.3 shows the number of pellets eaten with the left

(contra-lateral) forelimb plotted against time in both stroke and sham groups. All values

are quoted as the mean ± SEM.

On average, across eight baseline trials (prior to stroke induction or sham-

65

surgery), animals retrieved 14.9 ± 0.2 pellets (or 71 ± 1% of pellets available) on the left

side (N=18). Prior to surgery, there was no difference in performance between cohorts

(stroke animals: 14.6 ± 0.3 pellets, 70 ± 1%; sham-operated animals: 15.4 ± 0.3 pellets,

73 ± 1%). One week following surgery, the left forelimb reaching performance of

animals with stroke deteriorated to 2.8 ± 0.3 pellets or 13 ± 1% (P<0.0001), while

animals that underwent sham surgeries showed no impairment (consuming 14.9 ± 0.3

pellets, or 71 ± 1%). Two weeks later (21 days after stroke or sham-surgery), animals

with stroke remained impaired (3.2 ± 0.4 pellets or 15 ± 2%, P<0.0001) while sham-

operated animals still performed at baseline level (15.5 ± 0.4, 74 ± 2%). These data

demonstrate a substantial decrease in the skilled reaching ability of the contralateral

limb in animals with ischemic lesions seven days after injury with no evidence of

spontaneous recovery two weeks thereafter; while sham-operated animals showed no

effect of surgery or time on reaching performance. Of note, given the extensive

preclinical and clinical literature on ipsilateral limb impairment [Wetter et al. 2005,

Yarosh et al. 2004, and Kwon et al. 2007], the ipsilateral (i.e. right) forelimb also showed

a decrease in reaching success following stroke (data not shown), although the

ipsilateral deficit was smaller than the contralateral deficit at both seven and 21 days

(P<0.00001).

66

Figure 3.3

Figure 3.3 Montoya staircase performance. Pellets eaten during each trial with the

contralateral (left) forepaw are shown for each subject for each trial on days five-seven

(N=9 stroke, N=9 sham-operated), and 19-21 (N=7 stroke, N=9 sham-operated) after

injection. Baseline reaching ability, 14.9 ± 0.2 pellets or 71 ± 1% of pellets available on

the left side, was determined based on eight trials prior to surgery (days -5 until -1). At

baseline, there was no difference in performance of stroke (14.6 ± 0.3 pellets, 70 ± 1%)

and sham animals (15.4 ± 0.3 pellets, 73 ± 1%). There was no time-dependent change

in the reaching ability of sham-operated animals: they consumed an average 14.9 ± 0.3

(71 ± 1%) pellets on day seven and 15.5 ± 0.4 (74 ± 2%) pellets on day 21. Stroke

caused a decrease in the reaching performance: injured rats consumed 2.8 ± 0.3 (13 ±

1%) (P<0.0001) pellets on day seven and 3.2 ± 0.4 (15 ± 2%) (P<0.0001) pellets on day

21, with no difference in performance between seven and 21 days post stroke. On day

seven and 21, ischemic animals consumed fewer pellets with the contralateral forepaw

than sham animals (P<0.0001).

67

3.3.2 Spontaneous neuronal activity in the ipsi-lesional cortex

At either seven or 21 days after stroke-induction (or sham-surgery), both

somatosensory evoked and spontaneous LFP responses were recorded simultaneously

from the right (ipsilateral) and left (contralateral) hemispheres, using a pair of MEAs [in

a 2x8 configuration]. Each MEA was placed with the most anterior pair of electrodes in

line with the posterior ET-1 or PBS injection site (0.0 Bregma) and the long axis of the

array (eight electrodes in length) extending in the A-P direction to +3.0mm Bregma

(Figure 3.4 a. shows the surgical preparation for intracortical electrophysiological

recordings of both evoked and spontaneous activity). For a direct measure of neuronal

activity, spontaneous LFP events were recorded for one minute prior to recording

evoked activity. Following earlier work [Womelsdorf et al. 2014], lateralization (a

discrepancy between hemispheres) of spontaneous neuronal activity is reported as the

(ipsi- to contra-lateral) ratio of the spectral power in five frequency bands: Theta (3-

10Hz), Alpha (8-14Hz), Beta-1 (15-20Hz), Beta-2 (20- 35Hz) and Gamma (35-90Hz).

Figure 3.4 depicts MEA positioning (a.); power recorded by each array (ipsi- and contra-

lateral) for a representative animal from each group (sham and stroke) (b.i.); the

average across each MEA for these animals (b.ii.); and the group-wise ipsi- to contra-

lateral ratio average for both time points (c). Table 3.3 lists the ipsi- to contra-lateral

ratios for the five aforementioned frequency bands for both groups.

Table 3.3

Channel-wise ipsi- to contra-lateral power ratio [a.u.]

Sham animals Stroke animals

Band Day 7 (N=5) Day 21 (N=5) Day 7 (N=5) Day 21 (N=4)

Theta [3-10Hz] 1.1 ± 0.2 1.2 ± 0.1 0.49 ± 0.05 3.7 ± 0.4 ** †

Alpha [8-14Hz] 0.92 ± 0.09 1.06 ± 0.09 0.53 ± 0.05 2.8 ± 0.2 ** †

Beta-1 [15-20Hz] 0.81 ± 0.06 0.91 ± 0.08 0.58 ± 0.04 2.5 ± 0.2 ** †

Beta-2 [20-35Hz] 1.04 ± 0.07 1.1 ± 0.1 0.57 ± 0.04 * 2.3 ± 0.1 ** †

Gamma [35-90Hz] 1.01 ± 0.05 1.1 ± 0.1 0.74 ± 0.04 1.61 ± 0.06 **

Table 3.3 Channel-wise ipsi-/contra-lateral spontaneous activity power ratio. The

channel-wise ipsi- to contra-lateral power ratios averaged across channels and across

68

animals in the sham and stroke groups for each band of interest (± SEM) are reported

at seven and 21 days after surgery. At seven days, stroke animals had a lower Beta-2

band inter-hemispheric ratio (* P=0.03) relative to sham animals. Inter-hemispheric

ratios increased in every band in stroke animals between seven and 21 days (**

P<0.008). Further, between groups, the inter-hemispheric ratios in stroke animals was

greater than in sham animals in all bands († P<0.03) except Gamma [35-90Hz]

(P=0.08).

Figure 3.4 (next page) Intra-cranial electrophysiological recordings of evoked

responses. (a.) A schematic of the surgical preparation for intracortical

electrophysiological recordings, with the extent of cranial windows outlined in (a.i.) by

blue rectangles. The injection sites made on day 0 for ET-1 or PBS (sham) injections

are indicated by black outlined circles with red centres (a.i.). A photograph of a double

cranial window preparation from a stroke rat seven days after stroke is shown in (a.ii.).

The approximate placement of the MEAs is indicated by grey rectangles, with

electrodes shown as black circles. (b.i.) Recordings of spontaneous activity in Theta

band from contra- and ipsi-lateral hemispheres of a representative sham and

representative stroke rat on day seven. The corresponding average Theta band power

across all channels from these two animals is shown in (b.ii.) results are plotted as the

mean ± SEM. (c.) Inter-hemispheric channel-wise power ratios (sham in green; stroke in

orange) averaged across channels, in each band, on days seven and 21 following

stroke (results are plotted as the mean ± SEM). Compared to sham animals, stroke rats

exhibited decreased ipsilesional Beta-2 power on day seven (P=0.03); but increased

ipsilesional power in all bands (with the exception of Gamma) on day 21 (P<0.03). The

average inter-hemispheric power ratio in stroke rats increased between day seven and

21 in all frequency bands (P<0.008).

69

Figure 3.4

70

There was no lateralization of the spectral power in spontaneous activity

recordings in sham-operated animals: their average ipsi- to contra-lateral power ratios

(across all frequency bands) were 0.98 ± 0.04 (N=5) on day seven and 1.09 ± 0.04

(N=5) on day 21 (refer to Table 3.3 for results from separate bands). In contrast, stroke

animals showed decreased Beta-2 band (20-35Hz) power, by 51 ± 5%, relative to that

of sham animals, on day seven (P=0.03). By 21 days after stroke, the ipsi- to contra-

lesional power ratio in stroke animals increased, by an average of 67 ± 9%, indicating a

dramatic rise in the ipsi-lesional spontaneous activity. Moreover, the power of ipsi-

lesional (relative to contra-lesional) cortical activity on day 21 was higher, by 20 ± 10%

than that observed in the sham group. In summary, these data show that an initial

silencing in the ipsi-lesional hemisphere progresses to hyperactivity by 21 days after

injury.

3.3.3 Somatotopy in the ipsi-lesional cortex

As the ischemic injury targeted the forelimb representation, forepaw stimulation

was employed to assess the effects of ischemia. The physiological pattern of average

normalized response amplitudes as a function of distance (0.0→3.0mm Bregma)

recorded from contralateral (injection naïve) hemispheres seven and 21 days after

surgery are shown in Figure 3.5. Response amplitude (estimated by the magnitude of

the first negative peak – N1) was normalized within each MEA by dividing the average

amplitude (averaged across the first 10 responses) recorded by each electrode by the

amplitude of the largest mean response in that array. The evoked LFP response

amplitudes from the contralateral hemispheres followed the expected somatotopic

organization, thus getting progressively smaller from 0.0 to +3.0mm from Bregma in the

A-P direction at both time points in the stroke (day seven: N=5 and day 21: N=8) and

sham rats (day seven: N=7 and day 21: N=8) (P<0.0001). This decrease in LFP

amplitude with distance in the contralateral hemisphere did not depend on either time or

group.

71

Figure 3.5

Figure 3.5 Evoked LFP responses in the contralateral hemisphere. Normalized LFP

responses to bilateral forepaw stimulation as a function of distance from Bregma on day

seven (a.) and 21 (b.) in the contralateral hemisphere of sham (circles, light grey) and of

stroke rats (triangles, dark grey). The N1 amplitude was normalized within each MEA by

dividing the average amplitude recorded by each electrode by the amplitude of the

largest average response in that array. The amplitude of the LFP responses decreased

moving from 0.0 Bregma towards +3.0mm of Bregma in the A-P direction on day seven

and 21 in both groups (P<0.0001). This topological organization of forelimb responses

did not depend on either time point or group (stroke vs. sham). Results are plotted as

the mean ± SEM.

72

Evoked LFP responses in the ipsilateral (injected) hemispheres are shown in

Figure 3.6. In sham-operated animals, the LFP responses were not distinguishable

between hemispheres. In contrast, lesioned hemispheres of stroke animals exhibited

evoked LFP response patterns that were distinct from the LFP responses in either the

contra-lesional hemispheres of ischemic animals or those in either hemisphere of sham

animals (P<0.0001). Seven days post-stroke, the somatotopic organization in ipsi-

lesional hemispheres of stroke rats was completely abolished, with no dependence of

evoked LFP response amplitude on distance from Bregma. On day 21 post-stroke,

intracranial recordings provided evidence of somatotopic remodelling: LFP response

amplitude to forelimb stimulation increased with distance from the injection site

(P<0.0001). In summary, the results showed that the somatotopic representation in the

forelimb cortex was disrupted seven days after ischemia yet remodelled two weeks

later.

73

Figure 3.6

Figure 3.6 Intra-cranial electrophysiological recordings of spontaneous activity.

Left-to-right normalized LFP responses to bilateral forepaw stimulation, averaged

across all animals, as a function of distance from Bregma: (a.) day seven: sham group

(i.) and stroke group (ii.). (b.); day 21: sham group (i.) and stroke group (ii.). In sham

animals, LFP amplitude decreased with distance from Bregma seven and 21 days after

surgery, with no dependence of LFP amplitude on time after injection. Ipsilateral LFP

responses in stroke animals were different than those in control animals at both time

points (P<0.0001). Ipsilesionally there was no dependence of LFP amplitude on

distance from Bregma on day seven following stroke. On day 21 after stroke,

ipsilesional LFP response amplitude increased from Bregma to +3.0mm in the A-P

direction (P<0.0001). Results are plotted as the mean ± SEM.

74

3.3.4 Stroke volumes in the ipsi-lesional cortex

To assess the morphological consequences of stroke induction, in vivo T2-

weighted MR-images were acquired in all animals seven days after surgery. Table 3.4

lists the stroke volumes in both cohorts, as estimated by hyperintensity on T2-weighted

MRI. Representative in vivo T2-weighted MR-images are shown for one animal from the

ischemic group acquired at seven and 21 days after stroke in Figure 3.7. The volume of

stroke remained stable between seven and 21 days across all stroke animals.

Table 3.4

Experimental Group Day 7 [mm³] N Day 21 [mm³] N

Behaviour

i. Montoya Staircase Test 94 ± 14 9 100 ± 17 9

Intra-cranial Electrophysiology

i. Evoked Activity Day 7 81 ± 16 5 NA NA

ii. Evoked Activity Day 21 104 ± 8 8 97 ± 13 8

iii. Spontaneous Activity Day 7 81 ± 16 5 NA NA

iv. Spontaneous Activity Day 21 97 ± 6 6 113 ± 6 6

Functional MRI

i. Resting Perfusion Day 7 75 ± 8 29 NA NA

ii. Resting Perfusion Day 21 85 ± 9 17 105 ± 5 17

iii. Reactivity to 10% CO2 Day 7 72 ± 18 10 NA NA

iv. Reactivity to 10% CO2 Day 21 95 ± 14 9 102 ± 8 9

Table 3.4 Stroke volumes on T2-weighted MRI. For each experimental group

(behaviour, intracranial electrophysiology, and functional MRI), there was no difference

in the mean lesion volume (on T2-weighted MRI) at either time point, nor was there a

change in lesion volume between time points. The mean ± SEM are reported for the

stroke group at seven and 21 days after stroke from animals meeting the inclusion

criteria (listed in Table 3.2).

75

Figure 3.7

Figure 3.7 Stroke volumes on T2-weighted MRI. ET-1 injection (indicated by white

arrows) collected at day seven (a.) and 21 (b.) showing ischemia-induced hyperintensity

(due to edema/inflammation) (cf. Methods: Magnetic Resonance Imaging for lesion

segmentation steps). The stroke volume of all animals included in this work are plotted

in (c.); the mean ± SEM for day seven (N=29) was 92 ± 6 mm³ and for day 21 (N=20)

101 ± 6 mm³. There was no difference in stroke volume between time points.


CASL MRI experiments were performed to estimate resting perfusion and

reactivity to hypercapnic challenges, at seven and 21 days after stroke or sham-surgery.

The ipsilateral ROI signal was normalized to the contralateral ROI signal to facilitate

comparisons across groups and time. CASL maps overlaid on structural T2-weighted

MR-images from representative stroke and sham animals at seven and 21 days after

surgery are shown in Figure 3.8, with resting perfusion shown in Figure 3.8 (a.). Sham

animals showed no lateralization in resting perfusion either seven (0.96 ± 0.01) or 21

days (0.92 ± 0.02) post-surgery. In contrast, stroke animals exhibited resting perfusion

lateralization at day seven (1.63 ± 0.08), but not at day 21 (0.96 ± 0.07). In particular,

ipsilateral resting perfusion was elevated (relative to contralateral levels) in stroke

animals compared to sham-operated animals at seven days (P=0.00001), but not at 21

days (P=0.7). Further, ipsi- relative to contra-lateral perfusion decreased with time in

stroke animals (P=0.00001), but did not change in sham-operated animals (P=0.1).

76

Contralateral resting perfusion of either group did not differ from that of healthy grey

matter either seven days (with contralateral to whole brain non-lesioned grey matter

perfusion ratio of 0.92 ± 0.03 in stroke and 0.98 ± 0.02 in sham-operated rats) or 21

days post-surgery (0.91 ± 0.03 in stroke and 0.98 ± 0.02 in sham-operated rats).

Furthermore, normalized contralateral resting perfusion did not change with time in

either group (P=0.6).

Reactivity to hypercapnia at seven and 21 days after surgery is shown in Figure

3.8 (b.). As with resting perfusion, sham animals exhibited no lateralization in perfusion

responses on day seven: 1.1 ± 0.2; or on day 21: 1.2 ± 0.1. In contrast, stroke rats

showed elevated ipsi- relative to contra-lateral vascular reactivity at day seven: 1.8 ± 0.2

and 21: 2.1 ± 0.3 post-stroke. Perfusion response lateralization did not change with time

in either group (sham: P=0.7; stroke: P=0.4). Ipsilesional perfusion responses were

elevated (relative to contralaterally) in stroke animals compared to sham-operated

animals at both seven (P=0.04) and 21 days (P=0.03). Contralateral perfusion

responses were not different from those in healthy grey matter in either group either

seven days (with contralateral to whole brain non-lesioned grey matter perfusion

response ratio of 0.9 ± 0.1 in stroke and 0.9 ± 0.1 in sham-operated rats), or 21 days

(0.89 ± 0.06 in stroke and 0.95 ± 0.08 in sham-operated rats) post-surgery.

Furthermore, contralateral perfusion responses did not change with time in either group

(P=0.6).

77

Figure 3.8

78

Figure 3.8 Functional MRI of resting blood flow, and cerebrovascular reactivity to

10% CO2. Overlaid on T2-weighted MRI is the resting perfusion (a.) and blood flow

response to 10% CO2 (b.) in representative animals (i., ii., iv., v.). CASL data across all

animals within each group are shown in (iii.) mean ± SEM. Unlike sham rats, stroke rats

exhibited lateralization in baseline perfusion and vascular reactivity. In particular, stroke

rats showed ipsilesional hyperperfusion (at 1.63 ± 0.08 fold contralesional levels) and

hyper-reactivity (at 1.8 ± 0.2 fold contralesional levels) on day seven. By 21 days

following ischemic insult, baseline perfusion of stroke rats was no longer lateralized

(P<0.0001), yet ipsilesional vascular reactivity remained elevated (at 2.1 ± 0.3 fold

contralesional levels). There was no lateralization of either resting perfusion or perfusion

response to hypercapnia in sham animals at either time point.

3.3.6 Inflammatory cell, neuronal and vascular changes ipsi-lesionally

Serial coronal sections were prepared with GFAP (astrocytes), Iba-1

(microglia/macrophage), NeuN (neurons) and RECA-1 (endothelial cells)

immunofluorescent stains. The number of sections included in these analyses are

summarized in Table 3.5. GFAP showed widespread astrocytic activation within the

ipsilesional hemisphere, with a pronounced increase in the astrocytic activation in the

peri-lesional cortex. Dense clusters of active astrocytes were also present in the

caudate putamen, piriform cortex, and olfactory tubercle. These ROIs were chosen by

visual inspection of GFAP and Iba-1 stained sections. The ipsi- to contra-lateral signal

ratio for each of the ROIs for the aforementioned immunofluorescent stains in both

groups (stroke and sham) at day seven and 21 are shown in Figure 3.9 a. (GFAP) and

b. (Iba-1) and Figure 3.9a. (NeuN) and b. (RECA-1).

At both time points, ipsi- (relative to contra-lesional) GFAP and Iba-1 signal

intensity was elevated in all ROIs in the stroke vs. sham group. Table 3.6 summarizes

the ipsi- to contra-lateral signal intensity ratios for both time points for each ROI. There

was a time-dependent increase in the inter-hemispheric ratio of GFAP signal in stroke

animals in all ROIs except for the piriform cortex between day seven and 21, indicating

progressive astrocytic activation (see Figure 3.10 a.). Conversely, there was no change

in the ipsi- to contra-lesional signal intensity ratio of Iba-1 from stroke animals between

79

day seven and 21, indicating no change in microglia/macrophage recruitment and

activation over this period (see Figure 3.10 b.). Sham animals showed no lateralization

of GFAP or Iba-1 signals at either time point.

Table 3.5

Day 7 Day 21

Stain ROI Sham Stroke Sham Stroke

GFAP i. Cortex 24 21 24 12

ii. Caudate Putamen 17 13 16 8

iii. Piriform Cortex 23 20 23 12

iv. Olfactory Tubercle 12 8 14 7

Iba-1 i. Cortex 24 24 24 12



iv. Olfactory Tubercle 12 5 9 5

NeuN i. Cortex 24 23 23 11



RECA-1 i. Cortex 24 24 24 12



Table 3.5 Number of sections used for each ROI in the immunofluorescence

analysis. For each immunofluorescent stain (GFAP, Iba-1, NeuN, and RECA-1), for

each ROI (cortex, caudate putamen, piriform cortex, and olfactory tubercle), the number

of coronal sections included to estimate the mean signal density ipsi- to contra-lateral

ratio are listed for the sham, and stroke groups at seven and 21 days after stroke or

sham surgery.

80

Table 3.6

Ipsi- to contra-lateral Fluorescence [a.u.]

Sham Stroke

Stain ROI Day 7 Day 21 Day 7 Day 21

GFAP i. Cortex 1.15 ± 0.04 0.98 ± 0.03 2.5 ± 0.1 ‡ 3.0 ± 0.3 ‡

ii. Caudate Putamen 0.98 ± 0.07 0.87 ± 0.1 1.7 ± 0.2 † 4.6 ± 0.7 †

iii. Piriform Cortex 0.96 ± 0.05 0.91 ± 0.08 1.6 ± 0.2 * 1.9 ± 0.4 **

iv. Olfactory Tubercle 1.08 ± 0.05 0.99 ± 0.08 1.2 ± 0.2 ** 2.6 ± 0.3 ‡

Iba-1 i. Cortex 1.08 ± 0.04 1.07 ± 0.04 3.8 ± 0.5 ‡ 4.6 ± 0.8 ‡

ii. Caudate Putamen 0.99 ± 0.06 1.05 ± 0.06 3.6 ± 0.6 † 2.4 ± 0.5 **

iii. Piriform Cortex 0.89 ± 0.04 0.77 ± 0.04 1.71 ± 0.5 ** 2.0 ± 0.2 ‡

iv. Olfactory Tubercle 0.97 ± 0.1 0.99 ± 0.07 1.80 ± 0.4 * 1.7 ± 0.4 *

RECA-1 i. Cortex 1.00 ± 0.04 1.02 ± 0.04 1.6 ± 0.1 ‡ 2.7 ± 0.4 ‡

ii. Caudate Putamen 0.91 ± 0.06 1.08 ± 0.05 0.9 ± 0.1 0.96 ± 0.08

iii. Piriform Cortex 1.01 ± 0.05 1.01 ± 0.08 1.0 ± 0.3 0.94 ± 0.06

NeuN i. Cortex 1.03 ± 0.04 0.97 ± 0.05 0.35 ± 0.05 ‡ 0.36 ± 0.06 ‡

ii. Caudate Putamen 0.96 ± 0.08 0.9 ± 0.1 0.32 ± 0.07 † 0.10 ± 0.04 **

iii. Piriform Cortex 0.96 ± 0.04 0.9 ± 0.1 0.5 ± 0.2 ** 0.5 ± 0.2

Table 3.6 Ipsi- to contra-lateral ratios of optical density from immunofluorescence

analysis. The average ipsi- to contra-lateral signal intensity (GFAP and Iba-1) or

relative area (RECA-1 and NeuN) across all animals at seven and 21 days after

surgery. In comparison to sham rats, stroke animals showed increased ipsilesional

astrogliosis (GFAP) and macrophage/microglia recruitment in all ROIs. Angiogenesis

was observed only in the peri-lesional cortical tissue, while neuronal loss was evident in

all ROIs except for the olfactory tubercle. Results are plotted as the mean ± SEM. *

P<0.05, ** P<0.01, † P<0.001, ‡ P<0.0001.

81

Figure 3.9

Figure 3.9 GFAP and Iba-1 immunofluorescence. Interhemispheric ratio of GFAP (a.)

and Iba-1 (b.) signal intensity per unit area averaged across all animals on days seven

and 21 in the cortex (i.), caudate putamen (ii.), piriform cortex (iii.), and olfactory

tubercle (iv.). On day seven and 21, stroke rats showed increased ipsilesional GFAP

and Iba-1 signal in all ROIs, with the exception of GFAP in the olfactory tubercle on day

seven. GFAP interhemispheric asymmetry increased between day seven and 21 in

stroke rats in all ROIs expect for the piriform cortex. Results are plotted as the mean ±

SEM.

82

Figure 3.10

Figure 3.10 NeuN and RECA-1 immunofluorescence. Interhemispheric ratio of

percent area occupied by positively stained NeuN (a.) and RECA-1 (b.) cells on day

seven and 21 in the cortex (i.), caudate putamen (ii.), and piriform cortex (iii.). Stroke

rats exhibited neuronal loss on both day seven and 21 in all ROIs, with the exception of

the piriform cortex on day 21. There were no time-dependent changes in NeuN staining

in either cohort. Stroke animals showed increased endothelial cell density (on RECA-1)

ipsilesionally on day seven that increased further over the next two weeks. Results are

plotted as the mean ± SEM. (* P≤0.01)

83

On day seven post surgery, the ipsi- to contra-lesional ratio of NeuN signal was

lower in stroke animals compared to shams: 0.35 ± 0.05 (stroke rats) vs. 1.03 ± 0.04

(sham animals) (P<0.0001). Neuronal loss was also observed in the caudate putamen:

0.32 ± 0.07 (stroke) vs. 0.96 ± 0.08 (sham) (P=0.0004), and piriform cortex: 0.5 ± 0.2

(stroke) vs. 0.96 ± 0.04 (sham) (P=0.004). Neuronal loss thus extended to regions distal

from the site of injection. The decrease in neuronal density persisted on day 21 in the

cortex 0.36 ± 0.06 (stroke) vs. 0.97 ± 0.05 (sham) (P<0.0001) and caudate putamen

0.10 ± 9 (stroke) vs. 0.9 ± 0.1 (sham) (P=0.007).

The RECA-1 ipsi- to contra-lateral cortical signal ratio was elevated in stroke

animals compared to that of sham animals at both time points, indicating peri-lesional

angiogenesis on day seven: 1.6 ± 0.1 (stroke) vs. 1.00 ± 0.04 (sham), (P=0.0001); and

21: 2.7 ± 0.4 (stroke) vs. 1.02 ± 0.04 (sham) (P<0.0001). Further, this ratio rose with

time in stroke animals (P=0.002) indicating progressive morphological remodelling of

the micro-vasculature in the peri-lesional tissue (see Figure 3.10 b.). No changes were

observed in the caudate putamen, piriform cortex or olfactory tubercle. Figure 3.11

summarizes the pathological findings for a representative stroke animal at seven days.

84

Figure 3.11

Figure 3.11 Immunofluorescence in a representative subject and corresponding

T2 -weighted MRI. (a.) Atlas drawing [Paxinos & Watson 2005] at -0.36mm A-P from

Bregma, hence proximal to the posterior site of ET-1 injection, with an overlay

summarizing the immunohistochemical changes observed in stroke animals (scale bar:

1mm). Coloured regions indicate areas of astrocytic activation (yellow),

microglia/macrophage recruitment (green), decreased neuronal density (blue), and

increased vascular density (peach). b. T2-weighted slice roughly corresponding to the

atlas section in (a.) acquired in a representative rat seven days post-stroke. (c.)

Example coronal sections from rat in (b.) at -0.36mm from Bregma (scale bar: 1mm)

and insets showing the region surrounding the ET-1 injection site (i. GFAP ii. Iba-1 iii.

NeuN iv. RECA-1). Scale bar: 500μm. Pir: piriform cortex; Tu: olfactory tubercle.

85

3.4 Discussion

The present work provided a multi-modal characterization of the evolution of

neurogliovascular function in the subacute stage of the focal ischemic injury. Over three

weeks following ischemic insult, ET-1 micro-injection produced a stable lesion volume

(~100 mm3 or ~20% of the total brain volume), and a persistent impairment of left

forelimb skilled reaching ability. Blood vessel density was increased ipsi- vs. contra-

laterally and neuronal loss and inflammation were widespread. A week following stroke,

resting perfusion and vascular reactivity to 10% CO2 in ipsilesional cortices were

elevated by 60% and 80% relative to corresponding levels contralesionally. Further, the

topological pattern of LFP response amplitudes to bilateral forepaw stimulation were

abolished and the average spontaneous neuronal activity attenuated by ~50%. Three

weeks after stroke, ipsilesional hyperperfusion pseudo-normalized; while vascular

reactivity remained at double the contralesional level. LFP evoked response amplitudes

progressively increased with distance from the site of injury and the average ipsilesional

spontaneous neuronal activity was twice that observed contralesionally.


The transient hyperperfusion and persistent hyper-reactivity observed in the

ipsilesional cortex in this work are consistent with a number of prior studies in rat MCAO

models [Lin et al. 2008, Wang et al. 2002, Martin et al. 2012, Wegener et al. 2013]. Two

or four days following transient MCAO, rats are hyper-perfused [Wang et al. 2002] and

hyperperfusion persists between four and seven days following two-hour intra-luminal

MCAO, with peri-lesional perfusion being double that of sham-operated rats at seven

days [Martin et al. 2012], in excellent agreement with present findings. In a 60-minute

MCAO model, Wegener et al. (2013) identified three outcome patterns dependent on

lesion size and location, including limited hyperperfusion on day one, maximal

hyperperfusion on day four, and decreased vaso-reactivity on day four but elevated on

day 14, which is consistent with the present observations. The doubling of resting

perfusion at seven days in the current study also falls within the reported range of

86

ischemia-induced ipsi- vs. contra-lesional perfusion elevation: 120-200% reported

previously [van Lookeren et al. 1999, Wang et al. 2002, Martin et al. 2012].

In contrast, Shen et al. (2011) report hyperperfusion at 48 hours to resolve by

seven days, but also show perfusion changes strongly dependent on MCAO duration. In

addition, Shen and colleagues observe decreased peri-lesional perfusion responses to

hypercapnia concomitant to increased resting perfusion at 48 hours post stroke. Due to

the highly dynamic changes in the peri-lesional vascular function during the first week

following stroke [Wegener et al. 2013], it is hard to relate the vaso-reactivity at seven

days (our first post-stroke sampling point) to a 48-hour measurement. As with prior

hemodynamic studies on ischemic injury progression [Heiss et al. 1997, Shen et al.

2011, Hayward et al. 2011, Wegener et al. 2013], the current CASL protocol was limited

in that only relative measures of cortical hemodynamics were made. Within-subject

measurements of T1 relaxation time, inversion efficiency, and sampling at a series of

post-labelling intervals should be undertaken in future studies to allow absolute

perfusion quantification, following optimization of these sequences to minimize total

scan time and thus curb attrition. Moreover, propofol anesthesia, employed in current

study during all MRI experiments, induces 20-60% regional vaso-constriction [Cenic et

al. 2000, Werner et al. 1993], likely due to reduced metabolic demand [Werner et al.

1993, Dam et al., 1990]. Although logistically complex, modifications to the present

protocol to enable imaging in awake animals could be used to evaluate possible

interactions between the effects of ischemia and propofol anesthesia.

3.4.2 Injury induced angiogenesis

Ischemia-induced angiogenesis has been previously reported [Hayward et al.

2011, Wegener et al. 2013] and likely underlies the elevated peri-lesional resting

perfusion observed in the present work. Between two and seven days following MCAO,

Martin et al. (2012) show an increase in endothelial cell (CD31+) number peri-lesionally

between two and seven days following MCAO. Also in an MCAO model, Wegener et al.

(2013) and Hayward et al. (2011), show peri-lesional RECA-1 staining at 140% the

contralesional level two [Wegener et al. 2013] and 12 weeks [Hayward et al. 2011] post-

stroke. Furthermore, tripling of micro-vascular density following stroke was seen by Lin

87

et al. (2008) 14-21 days post-MCAO, who suggested it resulted from a delayed surge of

angiogenesis. The presently observed increase in endothelial cell density in the

ipsilesional cortex is thus in excellent agreement with prior studies.

However, although peri-lesional vascular density remained elevated at 21 days,

resting perfusion normalized. Pruning, recruitment of mural cells, the generation of an

extracellular matrix, specialization of the vessel wall, and the functional integration of

nascent vasculature are highly dependent on the spatio-temporal interaction of new

micro-vessels and the surrounding neurons and glia [cf. Review by Korn & Augustin

2015]. peri-lesional neuronal loss may have led to extensive pruning of new vessels and

hence unperfused vessels by day 21 [Korn & Augustin 2015]. Indeed, by 13 weeks

following transient MCAO, peri-lesional micro-vessel density (doubled at four weeks)

regresses to contralesional levels [Yu et al. 2007]. Furthermore, peri-lesional hyper-

reactivity to hypercapnia may have resulted from lower resistance of newly added,

immature vessels [White & Bloor 1992].

3.4.3 Remapping of cortical somatotopy following focal ischemia

Electrophysiological recordings show a somatotopic organization of the primary

somatosensory cortex in the healthy rat brain [Hosp et al. 2008]. Accordingly, we

observed monotonically decreasing forepaw stimulation elicited LFP response

amplitudes contralaterally in both groups and ipsilaterally in the sham group. One day

following transient MCAO, disruption of motor-evoked potentials ipsilesionally has been

observed and is thought to result from persistent transmission failure of cortical

synapses, which likely contributes to motor dysfunction [Bolay & Dalkara 1998]. In the

present work, no evidence of a spatial pattern in ipsilesional LFP response amplitudes

was observed a week following stroke.

Somatosensory LFP responses to forepaw stimulation improve one week

following ischemic damage to the striatum and fully recover by four weeks [Shih et al.

2014]. Thus, distal to the site of injury, LFP responses are compromised transiently

following stroke, but can recover to be indistinguishable from those in controls [Shih et

al. 2014]. In the present work, we observed a new topological pattern to emerge by

three weeks: evoked potential amplitudes increased monotonically with distance from

88

the site of injury. The present findings hence likely represent recovery of function in the

lesion periphery [Shih et al. 2014], and possibly remapping.

Following an ischemic injury, sensory and motor function performed by injured

tissue prior to infarction has the potential to remap to contra- and peri-lesional areas. In

general, remapping manifests as gross physiological changes in the responsiveness of

neuronal networks demonstrating widespread functional plasticity during the weeks-

months following stroke [Murphy & Corbett 2009, Carmichael 2012]. Although studies

have provided evidence of remapping during the chronic phase of ischemic injury

progression, what the features of the remapped response (magnitude, kinetics, or

spatial pattern) imply for functional outcome is still controversial [Buma et al. 2010,

Buetefisch 2015]. In the present work, neurophysiological changes were not

accompanied by improvement of skilled reaching ability on the Montoya Staircase test,

possibly as a result of the incomplete recovery process after three weeks. Furthermore,

the Montoya staircase task is particularly resistant to spontaneous recovery (associated

with cortical reorganization) [Murphy & Corbett 2009]. However, significant time-

dependent performance improvement on skilled reaching tasks has been observed with

frequent testing in animals with small stroke volumes (10-15mm3), and evidence of

cortical remapping concomitant with behavioural performance improvement with

intensive reach training therapy [Nishibe et al. 2015]. Specifically, before intervention, at

~2 weeks post-stroke, Nishibe et al. (2015) found forelimb motor maps to be

significantly smaller in stroked animals employing intracortical micro-stimulation (where

stimulation evoked movements were mapped using a single electrode) relative to sham-

operated animals (controls). By 5 weeks (after 20 days of rehabilitation therapy), the

motor map area of the rehabilitation group extended to 3.5mm from Bregma and was

14% larger than that in the controls; the motor map area of the group receiving no

rehabilitation therapy was 37% smaller than that of controls [Nishide et al. 2015]. By 38

days, the motor maps in rehabilitated animals extended to 4.5mm from Bregma [Nishide

et al. 2015]. These extents of spatial remodeling are consistent with those seen in the

present work.

Montoya staircase testing was employed in the present work because it is

quantitative, not subject to experimenter bias [Murphy & Corbett 2009], and reports on

clinically relevant performance (upper limb impairment being the most common motor

89

deficit in patients) [Kleim et al. 2007]. Notwithstanding, a battery of behavioural tests

should be used in future studies to assess overall functional gains over a longer

observation period.

3.4.4 Changes in neuronal excitability following focal ischemia

As further evidence of neuronal functional adaptation, a pronounced shift in the

inter-hemispheric asymmetry of spontaneous neuronal activity was observed in

ischemic animals between seven and 21 days after stroke. The ipsilesional cortex was

hypo-active (by ~50% relative to contralesional levels) at day seven, yet became hyper-

active by day 21 (at two fold contralesional levels). Magnetoencephalography in patients

with partial motor recovery similarly show an increase in band power ipsi- relative to

contra-lesionally [Tecchio et al. 2006]. Relative hyperactivity is thought to result from

increased intra-regional neuronal firing synchrony and, at a cellular level, an increase in

intrinsic neuronal excitability [Tecchio et al. 2006].

Preclinically, Clarkson et al. (2010) reports elevated tonic inhibition

(hypoexcitability) at one and two weeks following ischemic insult in a mouse

photothrombotic model. Unlike the present study, Clarkson et al. (2010) do not observe

evidence of delayed endogenous hyper-excitability. However, the latter assessment in

Clarkson et al. (2010) was made at two weeks (vs. three weeks in the current work),

which may have been too early for endogenous hyper-excitability to manifest. In

addition, the photothrombotic model employed by Clarkson et al. (2010), produces a

small necrotic core with no re-perfusion and very little peri-lesional tissue, which may

have created a different ischemic micro-environment and subacute stage injury

progression than the one produced by ET-1 injection in the present study.

Notwithstanding, the success of treatment with an inverse agonist of GABA-A receptor

by Clarkson et al. (2010) and in our earlier study [Lake et al., 2015] suggests that

accelerating and/or potentiating delayed endogenous hyper-excitability (observed in the

present work) may prove beneficial.

3.4.5 Widespread neuro-inflammation

90

The delayed astrocytic activation accompanied by decreased neuronal density

(>50% decrease in NeuN staining) in the peri-lesional cortex observed in the present

work are in accordance with earlier reports on the time course of inflammatory cell

infiltration and neuronal cell death in the subacute phase of ischemic injury progression

[Rossini et al. 2001, Nguemeni et al. 2015]. It is noteworthy that the area of reduced

neuronal density tightly circumscribed the site of injection and comprised a tissue

volume of ~0.065mm3, thus being much smaller than the area to which electrode arrays

were sensitive (~0.2mm3). Neuronal death was thus limited to the ischemic core, while

neurons in the peri-lesional tissue exhibited time-dependent changes in activity, as

evidenced on intracortical electrophysiology recordings.

Removed from the site of injury, a marked increase in astrocytic and

microglia/macrophage activation was found in the ipsilesional caudate putamen, piriform

cortex, and olfactory tubercle at both time points. Although each of these structures has

been implicated in focal ischemic injury progression (namely astrogliosis in the olfactory

tubercle, and neuro-inflammation in the piriform cortex and caudate putamen);

secondary ischemic injury effects in these brain regions are not commonly reported and

their implications are unknown [Speliotes et al. 1995, Gualtieri et al. 2012]. A key

integrating structure, the olfactory tubercle is connected to both the piriform cortex (from

which it receives monosynaptic olfactory input) and the caudate putamen (to which

there are medium-sized dense-spine cell projections), which may explain why damage

in these structures was observed [Fallon 1983, Wesson & Wilson 2011].

Although concomitant decreases in neuronal density in the caudate putamen and

piriform cortex were found, there was no effect of stroke on neuronal density in the

olfactory tubercle. Further, no regions distal to the site of injection showed changes in

vascular density on immunohistochemistry; and no morphological or vascular functional

changes were seen in any of these distal structures on MRI. The former suggests that

angiogenesis may only accompany a supra-threshold degree of injury; whereas the

latter is unsurprising given the limited sensitivity of structural MRI to subtle changes in

tissue morphology. Finally, it is also possible that back-flow of ET-1 and its subsequent

distribution through the subarachnoid space could have caused some distal injury.

91

3.5 Conclusion

On the whole, the present data demonstrated the complexity of the relationship

between structure and function of brain neuronal and vascular networks, and

emphasized the need for concomitant characterization of both over prolonged

observation periods following ischemic insult. It is likely that the effects of cortical

reorganization, via spatially-specific modulation of neurovascular function, on

behavioural outcome critically depend on the timing and extent of such changes. The

current findings warrant further exploration of neuronal activity patterns, vascular

remodelling and neuro-inflammation during the subacute stage of ischemic injury

progression and the effects their modulation may have on long-term neurophysiological

measures and behavioural outcome. As the vast majority of patients arrive at a care

facility well outside of the hyper-acute window, identifying ways through which outcome

can be improved during the subacute and chronic stages is of utmost importance.

92

On the whole, the data presented in Chapter 3 demonstrate the complexity of the

relationship between structure and function of brain neuronal and vascular networks,

emphasizing the need for their concomitant characterization. The data show highly

dynamic endogenous processes of structural and functional remodelling within both

vascular and neuronal networks over the weeks following ischemic injury. Although

skilled reaching remained impaired despite the neurovascular remodelling observed, it

is likely that effective rehabilitation treatments will be achieved once a mechanistic

understanding of the relationship between timing and spatial extent of cortical neuronal

network reorganization and behavioural performance has been gathered. Chapter 4

describes a study in which we attempted to enhance the endogenous healing through

attenuation of the peri-lesional neuro-inflammation via selective COX-1 inhibition in the

subacute stage of ischemic injury progression, a treatment strategy which has been

called for [Perrone 2010], but to-date not tested in a clinically relevant model of focal

ischemia in the subacute period.

93

Chapter 4

Delayed COX-1 promotes recovery of the neurovascular unit

following focal ischemic injury

Abstract

Stroke is the leading cause of adult disability worldwide. The absence of more effective

interventions in the chronic stage of stroke reflects uncertainty surrounding the

mechanisms that govern recovery and affect neurogliovascular function in the peri-

infarct region. To address this gap, we assessed the effects of selective

cyclooxygenase-1 (COX-1) inhibition via chronic FR122047 treatment beginning 7 days

following focal ischemia employing the endothelin-1 model in Sprague-Dawley rats.

Stroke caused an increase in peri-lesional resting perfusion and an elevation in peri-

lesional perfusion responses to hypercapnia at 7 days after injury. At 21 days post-

injury, placebo administered rats showed normalization of peri-lesional perfusion but

persistent hyper-reactivity to hypercapnia, whereas treated animals exhibited sustained

peri-lesional hyperperfusion while peri-lesional responses to hypercapnia were no

longer different from contra-lesional hypercapnic reactivity. These hemodynamic

changes were accompanied by peri-lesional neuronal loss, increased endothelial

density, and widespread microglial and astrocytic activation in all animals. When

compared to placebo administered rats, treated animals showed increased peri-lesional

neuronal survival and decreased microglia/macrophage recruitment. These findings

shed new light on the role of COX-1 in chronic ischemic injury progression and suggest

94

that delayed selective COX-1 inhibition has multiple beneficial effects on the

neurogliolvascular unit in the peri-infarct zone.

95

4.1 Introduction

A prominent pathophysiological consequence of stroke is the inflammatory

response, comprising leukocyte infiltration, activation and accumulation of microglia,

macrophages and astrocytes, and an increased production of inflammatory mediators

[cf. Reviews Spite et al. 2010, BIadecola et al. 2011, and Kim et al. 2014]. Hitherto,

candidate anti-inflammatory therapeutics have been preponderantly applied within the

acute period. Chronically, inflammatory processes evolve and exert both deleterious

and beneficial effects on remodelling and repair [Spite et al. 2010, BIadecola et al.

2011]. Following stroke, a hypoxia-induced increase in intracellular Ca2+ raises the

production of arachidonic acid (AA) [Stanimirovic et al. 2000]. Cyclooxygenases (COX-1

and COX-2) metabolize AA to different bioactive prostaglandins (PGs) [Yagami et al.

2015]. Some PGs ameliorate excito-toxicity [Kawano et al. 2006, Saleem et al. 2009]

while others activate poly adenosine diphosphate (ADP) ribose polymerase (PARP) and

caspase-3, causing an accumulation of ubiquitinated proteins and contributing to

neuronal apoptosis [Liu et al. 2013]. Furthermore, select PGs regulate blood vessel tone

[cf. Review Félétou et al. 2011] acting as either vasodilators or constrictors [Sanchez-

Moreno et al. 2004], control platelet aggregation [Yagami et al. 2015], and affect

leukocyte activation and leukocyte-endothelial interactions [Sanchez-Moreno et al.

2004].

Although promising effects have been observed with COX-2 inhibition in the

acute stage of stroke in preclinical research [cf. Reviews by Spite et al. 2010 and

Yagami et al. 2015], long-term placebo-controlled clinical studies have revealed serious

cardiovascular side-effects of COX-2 inhibitors [Ott et al. 2003, Nussmeier et al. 2005,

Cannon et al. 2006]. In the view of these challenges [cf. Review Katz et al. 2013],

attention has begun to turn to downstream PGs as well as to selective COX-1 inhibition

[Yagami et al. 2015, Rainsford et al. 2007, Aid et al. 2011, Liedtke et al. 2012, Perrone

et al. 2010]. Indeed, several PGs that ameliorate excitotoxicity [Yagami et al. 2015,

Kawano et al. 2006, Ahmed et al. 2011, Liang et al. 2011] in humans, but not

necessarily rodents, derive from COX-2 activity [Song et al. 2008]. In addition, PGs that

act as vasodilators and anti-thrombotic agents, are predominantly produced by COX-2

96

(in humans and rodents) [Sanchez-Moreno et al. 2004]. On the other hand, a number of

PGs that promote neuronal apoptosis [Liu et al. 2013], constrict vessels, and promote

platelet aggregation [Yagami et al. 2015], are preponderantly produced by COX-1 in

humans and rodents [Kobayashi et al. 2004, McAdam et al. 1999]. Notwithstanding, few

studies have investigated the role of COX-1 in the pathogenesis of stroke, especially in

the chronic stage, and the available data on the (sub)acute period are conflicting

[Yagami et al. 2015, Aid et al. 2011, Liedtke et al. 2012, Perrone et al. 2010]. COX-1

null mice have been shown to exhibit a greater reduction in peri-lesional blood flow four

days post-stroke and increased ischemic injury [BIadecola et al. 2001]. In contrast, a

single intraperitoneal dose of Valeryl Salicylate, a COX-1 inhibitor, six hours following

ischemia attenuated inflammation, neuronal loss, and oxidative stress on post mortem

histology in a gerbil model of transient global ischemia [Candelario-Jalil et al. 2003].

Moreover, COX-1 inhibitors are particularly attractive given their long-standing safety

record and their widespread use in managing inflammation. In addition, COX-1 inhibition

has shown benefit in preclinical models of other central nervous system pathologies

characterized by chronic inflammation, including rodent models of Alzheimer’s and

Parkinson’s disease [Choi et al. 2009, McKee et al. 2008, Nomura et al. 2011].

To investigate the role of COX-1 in the chronic stage of stroke, we assessed

cerebrovascular function, inflammation, and neuronal survival, following

intracerebroventricular administration of FR122047, a highly selective COX-1

antagonist. To probe these changes we employed immunohistochemistry and high field

MRI. Stroke was produced by intracortical micro-injection of endothelin-1 into

sensorimotor cortex of adult rats. Longitudinal structural and functional in vivo imaging,

prior to intervention (7 days following stroke) as well as 48 hours following cessation of

treatment (21 days following stroke), revealed spatio-temporal modulation of

cerebrovascular function, which was associated with changes in neuronal survival and

neuro-inflammation on histological analysis of brain tissue. Delaying treatment until 7

days following ischemia provides the opportunity to probe pre-intervention compromises

in behavioural performance and significantly elongates the therapeutic window, while

evaluation at 21 days post-stroke allows assessment of long-term outcome.

97

4.2 Methods

All data processing was performed blinded to group allocation (placebo or

FR122047) and time post ischemia (7 vs. 21 days). During all surgical procedures

(which were performed under aseptic conditions) and imaging sessions, body

temperature, breath rate, heart rate, and blood oxygen saturation were monitored

continuously and kept within physiological range (cf. Table 4.1). Refer to section 2.2.2

for stroke induction methodology.

3.2.1 Inclusion/exclusion criteria

Twenty-three adult male Sprague-Dawley rats weighing 297 ± 23g (mean ± SD)

were included in this study. Three rats went into respiratory arrest and died during

endothelin-1 injection: this attrition is consistent with our previous work in this model

[Lake et al. 2015, Chapter 2]. Hardware failure resulted in the death of one rat during

the hypercapnic challenge in MRI 7 days after stroke induction; one rat died during

ALZET mini-pump implantation from intracranial bleeding following catheter

misplacement, and one rat was euthanized for excessive weight loss (>10% decrease

from pre-stroke body weight) between imaging sessions. Finally, one FR122047 treated

rat was excluded at the 21 day imaging session because the catheter (for FR122047

delivery) had detached from the skull. Consequently, there were N=19 (N=9 placebo,

and N=10 FR122047 treated) MRI data sets collected at 7 days (prior to intervention);

and N=16 (N=7 placebo, and N=9 FR122047 treated) at 21 days (post-intervention), (cf.

Table 4.2 for animal numbers summary).

98

Table 4.1

Temperature [°C]

Breath Rate [breaths/min.]

Heart Rate [beats/min.]

Arterial O2 Saturation [%]

Isoflurane during stroke induction (N=13)

Pre-ET-1 37 ± 0.3 53 ± 6 346 ± 21 99 ± 0.6

Post-ET-1 37 ± 0.4 57 ± 7 342 ± 24 98 ± 0.6

Breath Rate [breaths/min.]

Heart Rate [beats/min.]

Expired CO2 [%]

7 Days post-stroke MRI under propofol (N=17)

CO2 Off 54 ± 13 640 ± 141 10 ± 3

CO2 On 57 ± 9 625 ± 59 16 ± 5

21 Days post-stroke MRI under propofol (N=6)

Placebo

CO2 Off 59 ± 11 605 ± 25 14 ± 12

CO2 On 56 ± 14 601 ± 24 15 ± 6

FR122047 treated (N=6)

CO2 Off 42 ± 15 668 ± 11 9 ± 4

CO2 On 41 ± 18 623 ± 4 13 ± 5

Table 4.1 Physiological monitoring. Data was collected during surgical procedures

(under isoflurane) during stroke induction; as well as during MRI (under propofol). Rats

were free breathing during surgery, but intubated and mechanically ventilated for MRI.

Results are reported as the mean ± the standard deviation.

99

Table 4.2

Placebo FR122047

Experimental Group 7 Days 21 Days 7 Days 21 Days

T2-weighted MRI and resting perfusion 9 7 10 9

Reactivity to hyper-capnia 9 6 8 9

Immunohistochemistry 0 3 0 4

Exclusion Criteria 7 Days 21 Days 7 Days 21 Days

Death during ET-1 micro-injection 2 Na 1 Na

Hardware failure during MRI 0 0 1 0

Poor SNR during MRI acquisition 0 1 2 0

Catheter failure Na 1 Na 1

Weight loss 0 1 0 0

Table 4.2 Number of animals. Number of animals included within each experimental

group and number of animals excluded.

100

4.2.2 FR122047 administration

FR122047 was delivered intracerebroventricularly to reduce off-target effects due

to the expression of COX-1 throughout the body [Perrone et al. 2010]. All rats were

anesthetized with isoflurane (5% induction and 2-2.5% maintenance) and implanted

with ALZET-osmotic pumps (ALZET Osmotic Pumps, Cupertino, CA, USA) 7 days after

stroke. In preparation, 24-48 hours prior to implantation, ALZET-osmotic pumps (2ML2,

delivery rate 5µl/hour) were loaded with 1.6mg FR122047 (Cedarlane Laboratories LTD,

Burlington ON, Canada) suspended in 2ml of sterile saline (dosage of ~400µg/kg/day)

or sterile saline alone (placebo). Super glue (Loctite, Westlake Ohio, USA) was used to

secure ~8 inches of PV-56 tubing (BTPE-50 polyethylene tubing, Instech Laboratories,

Inc., Polymouth Meeting PA USA) to each ALZET-osmotic pump flow moderator to

serve as a delivery catheter. Rats were randomly assigned to either FR122047 treated

or placebo groups.

A filled pump was inserted into the peritoneal cavity and the musculoperitoneal

layer and peritoneal wall were closed leaving the PV-56 tubing catheter penetrating the

abdominal cavity at the anterior end of the incision. Using blunt dissection the

cutaneous layer was separated from underlying tissue to allow the catheter to be fed

subcutaneously from the abdomen to the base of the skull. At this point, rats were

transferred to a stereotaxic apparatus (David KOPF Instruments, Tujunga California,

USA). Using a high-speed micro-drill (Foredom Electric Co., Bethel Connecticut, USA),

a burr hole was made through the interparietal bone 2-mm posterior of lambda and 2-

mm right of the midline. The catheter was placed through the burr hole within the lateral

ventricle of the right hemisphere and secured to the skull with super glue. For analgesia,

rats were given a subcutaneous dose of Lidocaine (0.2mg/kg) at each incision site at

the beginning and at the end of surgery. All incisions were closed with 4.0 absorbable

sutures (UNIFY PGA sutures, AD Surgical, Sunnyvale CA, USA).

Twelve days following pump implantation, all rats underwent the ALZET-osmotic

pump extraction procedure (prior to the second MR imaging session). With the rat in the

prone position, a 1-cm midline incision was made at the base of the skull in the M-L

direction to access the catheter. The catheter was cut and sealed leaving 1-cm of tubing

still within and attached to the back of the skull. The incision overlying the base of the

101

skull was sutured. The ALZET-osmotic pump and attached catheter were next removed

from the abdominal cavity. Lidocaine (0.2mg/kg) was administered subcutaneously at

each incision site (cf. Fig. 4.1). Animals were monitored closely (twice daily) for 72

hours post-surgery (stroke induction and pump implantation/removal). Thereafter, the

surgical area was inspected daily, to look for signs of inflammation/infection. No

instances of infection were observed. Frequent post-operation checks of weight,

grooming, and behaviour were performed.


All animals were imaged 7 and 21 days following stroke, and a stable plane of

anesthesia was maintained with an intravenous infusion of 45mg/kg/hr of propofol

(Pharmascience Inc., Montreal Quebec, Canada). Notably, intravenous propofol is

commonly used in longitudinal functional neuroimaging studies in rats [Haesel et al.

2015]. In the present work, titration experiments have shown that a low dose of propofol

(45mg/kg/hr) is well tolerated and yields robust responses to hypercapnia as well as

rapid recovery. Animals were intubated and mechanically ventilated with 31% O2 and

69% N2. Following the first imaging session, animals were extubated and recovered.

Structural imaging: Please refer to section 2.2.4 for T2-weighted imaging

protocol (cf. Table 4.3 for lesion volumes). CASL acquisition: Refer above to section

3.2.2 for imaging protocol (note: fewer hypercapnic challenges were presented in the

present work – limited to 4). CASL data processing: Refer to section 3.2.2. (note:

subject-specific hemodynamic response function were produced by averaging the signal

in the left (notionally unaffected) cortical grey matter across hypercapnic challenges

time-locked to the onset of the increase in the end-tidal CO2 signal).

102

Figure 4.1

Figure 4.1 Schemiatic of ALZET-pump implantation. Chronic placebo or FR122047

delivery was accomplished via intraperitoneally (IP) implanted ALZET-osmotic pumps

with catheters fed subcutaneously to the right lateral ventricle. The injection sites where

endothelin-1 was delivered are indicated with red dots in (a.). Seven days after

endothelin-1 injection, a third burr hole was drilled for catheter insertion (indicated by

the green dot in (a.)). The opposite end of the catheter was attached to the ALZET-

pump implanted IP. The turquoise line indicates the coronal plane ~4.7mm posterior of

Bregma and correspond to the T2-weighted image (acquired from a representative rat

21 days following stroke) shown in (b.). The black arrow in (b.) indicates the catheter

placed within the right lateral ventricle.

103

Table 4.3

Placebo FR122047

ROI 7 Days N 21 Days N 7 Days N 21 Days N

T2 hyper-intensity 56 ± 12 9 73 ± 25 7 49 ± 9 10 55 ± 14 9

Resting hyperperfusion 40 ± 8 9 39 ± 8 7 27 ± 6 8 24 ± 7 9

Vascular hyper-reactivity to CO2 29 ± 6 9 41 ± 12 6 34 ± 5 8 41 ± 7 9

Table 4.3 ROI volumes. The mean volume ± SEM [mm3] found to belong to structural

or functional ROIs for placebo and FR122047 treated rats at 7 and 21 days after stroke

(before and after intervention). There were no differences in any of the ROI volumes

between groups (placebo versus FR122047) at either 7 or 21 days following stroke or

between time points in either group.


Refer to section 3.2.4 above for staining methodology and digitization of stained

sections. Note: the fractional area occupied by positively stained GFAP and Iba-1 cells

was compared between groups across all brain sections (with ventricles excluded by

manual segmentation). The NeuN and RECA-1 staining was examined in the peri-

lesional ROI, which was defined by directly projecting manually segmented areas of

intense cortical staining on GFAP and Iba-1 in adjacent brain sections.


Refer above to section 2.2.7. Note: Lesion volume, average resting perfusion,

and average perfusion response to hypercapnia were modelled as linear functions of

hemisphere (peri- versus contra-lesional), group (placebo versus FR122047) and time

after stroke (7 versus 21 days). The fractional area of peri-lesional tissue occupied by

cells stained for NeuN, and RECA-1; and the fractional area occupied by Iba-1 and

GFAP positive cells were modelled as linear functions of group (placebo versus

FR122047).

104

4.3 Results

4.3.1 Resting perfusion and vascular reactivity to hypercapnia

Maps of representative resting perfusion and perfusion responses to hypercapnia

are shown in Fig. 4.2 and 4.3 respectively, along with plots of the mean group-wise

signal within contra- and peri-lesional ROIs and the inter-cortical signal ratios. Resting

perfusion and perfusion response to hypercapnia (mean ± SEM, in percent) in placebo

and FR122047 groups across time are listed in Table 4.4. To facilitate inter-subject

comparisons given variability in labelling coil placement in relation to the carotids and

subjects’ responses to anesthesia, the peri-to-contra lesional ratios were computed in

each rat and averaged within groups (also listed in Table 4.4).

Table 4.4

Placebo FR122047

Day Left Right Ratio [R/L] Left Right Ratio [R/L]

Resting perfusion

7 2.2 ± 0.4% 3.4 ± 0.5% 1.7 ± 0.2 1.8 ± 0.3% 3.0 ± 0.4% 1.8 ± 0.1

21 1.2 ± 0.3% 1.2 ± 0.3% 1.0 ± 0.2 1.7 ± 0.4% 2.7 ± 0.6% 1.8 ± 0.1

Perfusion response to hypercapnia

7 93 ± 25% 154 ± 28% 1.9 ± 0.2 98 ± 20% 168 ± 31% 1.8 ± 0.2

21 116 ± 18% 217 ± 35% 1.9 ± 0.2 119 ± 27% 181 ± 47% 1.7 ± 0.2

Table 4.4 Cortical hemodynamics. Mean signal ± SEM [%] for resting perfusion, and

perfusion response to hypercapnia in placebo administered and FR122047 treated rats

before (7 days) and after (21 days) intervention in the peri-lesional ROI (right) and in the

contra-lesional ROI (left) averaged across animals. The inter-hemispheric ratio ([R/L]

right normalized to left) is reported in columns 4 and 7. For the inter-hemispheric ratio,

the average signal within each animal’s right ROI was divided by the average signal

within the animals left ROI, and the average ratio across animals is reported.

105

Figure 4.2

106

Figure 4.2 Resting Perfusion. Representative maps of resting perfusion from a

placebo (a.) and FR122047 treated rat (b.) at 7 (i.) and 21 (ii.) days after stroke.

(Reactivity maps from the same subjects are shown in Fig. 4.2.) The resting perfusion

(mean ± SEM) in contra- and peri-lesional ROIs are plotted in (c.) for 7 days (i.) (N=9

placebo, and N=10 FR122047 treated rats) and 21 days (ii.) (N=7 placebo, and N=9

FR122047 treated rats). Signal in (c.) is reported as the ASL signal [%] (i.e. estimated

difference between ctl and tag frames was divided by the dc signal estimate) averaged

within each ROI for each animal, and then averaged across animals. The inter-

hemispheric resting perfusion ratio ([R/L] right normalized to left) is plotted in (d.). For

the inter-hemispheric ratio, the average signal within each animal’s right ROI was

divided by the average signal within the animals left ROI, and the average ratio across

animals is reported. Placebo rats’ data are plotted in grey and FR122047 treated rats’

data in green. Resting perfusion was elevated 7 days post-stroke peri- vs. contra-

lesionally in both groups (P<0.01). At 21 days, after intervention, there was no longer a

difference in resting perfusion between hemispheres in placebo rats (P=0.8). In

contrast, the peri-lesional cortex of FR122047 treated rats remained hyper-perfused

(P=0.006). Furthermore, resting perfusion was greater in the peri-lesional ROI in

FR122047 relative to placebo rats (P=0.05). Across time, placebo rats showed a

decrease in peri-lesional resting perfusion (P=0.01); whereas FR122047 treated rats

showed no change (P=0.7). By 21 days, the inter-hemispheric ratio within placebo rats

had decreased (P=0.03), whereas there was no time-dependent change in the

interhemispheric ratio of FR122047 treated rats (P=0.8). No changes were observed in

contra-lesional perfusion in either group.

107

Figure 4.3

108

Figure 4.3 Blood flow response to hypercapnia. Representative maps of reactivity to

hypercapnia from a placebo (a.) and FR122047 treated rat (b.) at 7 (i.) and 21 (ii.) days

after stroke. The perfusion change elicited by hypercapnia in contra- and peri-lesional

ROIs are plotted in (c.) for 7 days (i.) (N=9 placebo, and N=8 FR122047 treated rats)

and 21 days (ii.) (N=6 placebo, and N=9 FR122047 treated rats). Signal in (c.) is

reported as the fASL signal [%] (thus dividing the estimated fASL regressor’s weighting

factor by the estimated the ASL regressor’s weighting factor) averaged within each ROI

for each animal, and averaged across animals. The inter-hemispheric ratio of perfusion

responses is plotted in (d.). For the inter-hemispheric ratio, the average signal within the

right ROI for each animal was divided by the average signal within the left ROI, and we

plot the average ratio across animals. Placebo rats’ data are plotted in grey and

FR122047 treated rats’ data in green. At 7 days post-stroke, peri-lesional vascular

reactivity to hypercapnia was elevated, in comparison to the contra-lesional levels, in

both groups (P<0.002). At 21 days after stroke, peri-lesional perfusion responses to

hypercapnia were elevated (P=0.005), while FR122047 treated rats no longer showed

lateralization (P=0.2). However, there were no significant differences between groups,

nor were there time-dependent differences in either hemisphere in either group.

109

Prior to intervention, there were no differences between groups in lesion volume,

resting perfusion, or perfusion responses to hypercapnia. In both groups, resting

perfusion was elevated peri- vs. contra-lesionally: 3.4 ± 0.5% vs. 2.2 ± 0.4% in placebo

(N=9, P=0.01) and 3.0 ± 0.4% vs. 1.8 ± 0.03% in FR122047 treated rats (N=10,

P=0.0001). Across the two groups, peri-lesional perfusion was thus 170 ± 10% of the

contra-lesional perfusion and perfusion increase to hypercapnia was 180 ± 10% of the

contra-lesional level. Following intervention, resting perfusion was no longer lateralized

in placebo rats: 1.2 ± 0.3% in both hemispheres (N=7, P=0.8). In contrast, resting

perfusion in FR122047 treated rats remained elevated peri- vs. contra-lesionally: 2.7 ±

0.6% vs. 1.7 ± 0.4% (N=9, P=0.006). Furthermore, resting perfusion was greater in the

peri-lesional ROI of FR122047 treated rats relative to placebo rats (P=0.05). Across

time, the placebo group exhibited a decrease in peri-lesional resting perfusion: from 3.4

± 0.5% at 7 vs. 1.2 ± 0.3% by 21 days (P=0.01); whereas FR122047 treated rats

showed no time-dependent changes (P=0.6). No changes in resting perfusion were

observed in the contra-lesional ROI in either group.

In response to hypercapnia, peri-lesional vascular reactivity was almost doubled

7 days post-stroke peri- vs. contra-lesionally: 154 ± 28% vs. 93 ± 25% in placebo

animals (N=9, P=0.0009); and 168 ± 31% vs. 98 ± 20% in FR122047 treated rats (N=8,

P=0.002). At 21 days post-stroke, peri-lesional perfusion responses to hypercapnia

were still strongly elevated in placebo rats: 217 ± 35% peri- vs. 116 ± 18% contra-

lesionally (N=6, P=0.005). In contrast, FR122047 treated rats showed no lateralization

of perfusion responses to hypercapnia: 181 ± 47% vs. 119 ± 27% (N=9, P=0.2).

However, there were no significant differences between groups in peri- or contra-

lesional hemispheres at either time point, nor were there time-dependent changes in

perfusion responses to hypercapnia in either hemisphere in either group.


Representative Iba-1 and GFAP images from a placebo and a FR122047 treated

rat are shown in Fig. 4.4, along with the mean ± SEM Iba-1 and GFAP fractional areas.

The corresponding NeuN and RECA-1 data are shown in Fig. 4.5. The corresponding

NeuN and RECA-1 data are shown in Fig. 4. The fractional area occupied by

microglia/macrophages was smaller in FR122047 treated than in placebo rats: 17 ± 1%

110

in FR122047 vs. 20 ± 1% in placebo rats (P=0.05) 21 days post-stroke. The fractional

area occupied by astrocytes was not different between the two groups: 38 ± 2% in

FR122047 vs. 43 ± 4% in placebo rats (P=0.4). NeuN+ cells occupied a smaller cortical

area peri- vs. contra-lesionally: 14.9 ± 0.8% peri- vs. 32 ± 3% contra-lesionally in

placebo (P=0.03); and 22 ± 1% peri- vs. 36 ± 2% contra-lesionally in FR122047 treated

rats (P=0.009). However, FR122047 treated rats showed increased peri-lesional

neuronal survival: the fractional area occupied by neurons was greater in FR122047

treated rats relative to that of placebo rats (P=0.02). Peri-lesional angiogenesis was

observed in both groups: endothelial cells occupied a greater fractional area peri- vs.

contra-lesionally, with 51 ± 2% peri- vs. 22 ± 1% contra-lesionally in placebo (P=0.006),

and 49 ± 3% peri- vs. 22 ± 2% contra-lesionally in FR122047 treated rats (P=0.003).

There was no effect of group on the fractional area occupied by endothelial cells

(P=0.3). (In addition, the average stroke volume on T2-weighted MRI at 7 days post-

stroke in rats analyzed with immunohistochemistry was 50 ± 10 mm³, N=7, with no

differences between groups: 56 ± 20 mm³ in placebo, N=3 vs. 44 ± 11 mm³ in

FR122047 treated rats, N=4.)

111

Figure 4.3

Figure 4.4 Iba-1 and GFAP immunohistochemistry. Representative Iba-1 images

from a placebo rat (i.) and from a FR122047 treated rat (ii.) are shown in (a.).

Representative GFAP images from neighbouring coronal sections are shown in (c.).

The average Iba-1 and GFAP positive fractional areas across rats for both groups

(within the whole brain ROI) are plotted in (b.). The fractional area occupied by

microglia/macrophages (Iba-1) was smaller in FR122047 treated relative to that in

placebo administered rats: 17.1 ± 0.9% in FR122047 treated vs. 20 ± 1% in placebo

rats (P=0.05). Scale bars: 1mm.

112

Figure 4.5

Figure 4.5 NeuN and RECA-1 immunohistochemistry. Representative NeuN images

from a placebo rat (i.) and from a FR122047 treated rat (ii.) are shown in (a.).

Representative RECA-1 images from neighbouring coronal sections are shown in (c.).

The average fractional areas occupied by neurons and endothelial cells are plotted in

(b.). FR122047 treated rats showed greater peri-lesional neuronal survival than did the

113

placebo rats: 14.7 ± 0.7% of peri-lesional cortex was occupied by NeuN+ cells in

FR122047 treated vs. 10 ± 1% in placebo rats (P=0.02). Scale bars: 1mm.

4.4 Discussion

The present study characterized neurogliovascular changes elicited by selective

COX-1 inhibition via intracerebroventricular FR122047 treatment initiated 7 days

following stroke in an endothelin-1 model of focal ischemia. Stroke induced a transient

increase in peri-lesional resting perfusion and a sustained elevation in peri-lesional

vascular reactivity to hypercapnia in placebo rats. In FR122047 treated rats, peri-

lesional hyperperfusion persisted at 21 days post-stroke while perfusion responses to

hypercapnia showed a tendency to normalize. The peri-lesional vascular reactivity was,

however, highly variable across the FR122047 group. Immunohistochemistry revealed

peri-lesional neuronal loss, increased peri-lesional endothelial density, and widespread

increases in microglial/macrophage recruitment and astrocytic reactivity in all rats.

However, relative to placebo rats, FR122047 treated rats showed increased peri-

lesional neuronal survival and decreased peri-lesional microglial/macrophage

recruitment.

4.4.1 Peri-lesional hemodynamics post ischemia

At one week post-stroke, across all rats, peri-lesional perfusion was 170 ± 10%

and perfusion responses to hypercapnia were 180 ± 10% of the respective contra-

lesional levels. These elevations are consistent with previous reports in rat models of

focal ischemic injury [Wang et al. 2002, Martin et al. 2012, Wegener et al. 2013], but in

contrast with other studies [Wegener et al. 2013, Shen et al. 2011]. In a rat transient

occlusion model of stroke, peri-lesional hyperperfusion is observed 2-4 days following

ischemia [Wang et al. 2002]. In a 2-hour intraluminal middle cerebral artery occlusion

(MCAO) model of focal ischemia, the 7-day peri-lesional perfusion is double that of

sham-operated rats [Martin et al. 2012], in excellent agreement with the present finding.

Further, in a rat 60-minute transient MCAO model, Wegener et al. (2013) identifies three

outcome patterns. (1) Large hemispheric cortico-subcortical strokes, with

hyperperfusion on day 4 post-stroke and a further increase in resting perfusion on day

114

14, with decreased vaso-reactivity at both time points. (2) Moderate cortico-subcortical

strokes, with maximal hyperperfusion on day 4, decreased vaso-reactivity on day 4, and

elevated vaso-reactivity on day 14. (3) Predominantly subcortical infarctions, with

maximal hyperperfusion on day one; persistent hyperperfusion on day 14, and

unaffected or increased vaso-reactivity in punctate areas within the ischemic lesion on

day 4, and more pronounced hyper-reactivity by day 14. Although somewhat variable,

prior work in preclinical models of focal ischemia provides evidence of peri-lesional

hyperperfusion and hyper-reactivity to hypercapnia during the sub-acute period, in

agreement with the present work.

On the other hand, Shen et al. (2011) report that hyperperfusion peaks 48 hours

after MCAO and resolves by day 7 in most rats. The detection of hyperperfusion in this

study is, however, strongly dependent on occlusion time (30-/60-/90-minutes) [Shen et

al. 2011]: hyperperfusion is detected (at least once) in all of the short, in half of the mid-

length, and in none of the long occlusion time groups [Shen et al. 2011] (shorter MCAO

times have been suggested less likely to damage blood vessels). Furthermore, they

observe that hypercapnia decreases perfusion response to hypercapnia at 48 hours

within the hyper-perfused peri-lesional tissue relative to the contra-lesional perfusion

response [Shen et al. 2011]. Presently observed hyperperfusion at 7 days was, in

contrast, accompanied by increased peri-lesional perfusion responses to hypercapnia:

due to the highly dynamic changes in the peri-lesional vascular function during the first

week following stroke, the interpolation of present 7-day findings to 2-day time point is

exceedingly difficult and likely accounts for the discrepancy.

4.4.2 Ischemia-induced angiogenesis

In this work, elevated resting perfusion likely resulted from increased vascular

density due to injury-induced angiogenesis, which has been widely observed

histologically following focal ischemia in rats [Marntin et al. 2012, Wegener et al. 2013,

Lin et al. 2008, Hayward et al. 2011]. Martin et al. (2012) show increased numbers of

endothelial cells (CD31+) between 2 and 7 days following focal ischemia. Wegener et al.

(2013), report peri-lesional RECA-1 staining at 140% the contra-lesional level in rats

with evidence of increased vaso-reactivity, whereas rats with normal perfusion

responses to CO2 exhibit no lateralization in RECA-1 staining. Likewise, Hayward et al.

115

(2011) estimate blood vessel branch number (using RECA-1 staining) at 140% the

contra-lesional level concomitant to hyperperfusion up to 12 weeks following MCAO. Lin

et al. (2008) observe the density of peri-lesional small caliber vessels (<30µm) 3 days

after stroke to be 3-times the density in sham-operated controls [Lin et al. 2008]. Of

note, vascular networks formed by vessel sprouting undergo extensive pruning to form

a mature network [cf. Review Korn et al. 2015] and prior work in a rat model of MCAO

suggests that angiogenesis following focal ischemia begins with the hyper-dilation of

peri-lesional vessels and subsequent generation of microvessels during the weeks

following ischemic injury [Morris et al. 2003]. Across all animals presently examined with

RECA-1 immunohistochemical analysis (N=7), there was a doubling of peri-lesional

vascular density relative to contra-lesional levels. The increase in peri-lesional perfusion

observed in all rats prior to intervention was thus consistent with ischemia-induced

angiogenesis.

4.4.3 FR122047 treatment effects

Peri-lesional resting perfusion normalized to contra-lesional levels in placebo

rats, but remained at double the contra-lesional level in FR122047 treated rats (2.7 ±

0.6% peri- vs. 1.7 ± 0.4% contra-lesionally, P=0.006). The placebo rats showed a

decrease in peri-lesional resting perfusion over time (3.4 ± 0.5% at 7 vs. 1.2 ± 0.3% at

21 days, P=0.01); whereas FR122047 treated rats showed no time-dependent changes

in peri-lesional perfusion (3.0 ± 0.4% at vs. 2.7 ± 0.6% at 21 days, P=0.6). Inter-

hemispheric differences in perfusion responses to hypercapnia were significant in both

groups at 7 days, and persisted in placebo rats (217 ± 35% peri- vs. 116 ± 18% contra-

lesionally, P=0.005), but were no longer observed in FR122047 treated rats (181 ± 47%

peri- vs. 119 ± 27% contra-lesionally, P=0.2) given the high variability in peri-lesional

perfusion responses in treated animals. Accordingly, peri-lesional perfusion responses

were not distinguishable between groups, nor did peri-lesional perfusion responses in

FR122047 treated rats change with time.

Increased peri-lesional neuronal survival in FR122047 treated rats likely resulted

from attenuation of secondary cell death, which has been reported previously 1-3 weeks

following focal ischemia [Wegener et al. 2006]. In support of this assertion, COX-1

inhibition (via Valeryl Salicylate (20 mg/kg) administered intraperitoneally 6 hours post-

116

reperfusion) decreases neuronal loss in a gerbil model of global ischemia [Candelario-

Jalil et al. 2003]. This effect has been suggested to result from a reduction in PG-D2

production – largely mediated by COX-1 post ischemia [Song et al. 2008] – that reduces

the PG-D2-induced increase in neuronal apoptosis by activation of poly ADP ribose

polymerase and caspase-3 [Liu et al. 2013].

Pruning, recruitment of mural cells, the generation of an extracellular matrix,

specialization of the vessel wall, and the functional integration of nascent vasculature

are highly dependent on the spatio-temporal interaction of nascent vessels with

surrounding neurons and glia [cf. Review Korn et al. 2015]. Increased peri-lesional

neuronal survival may thus have promoted the integration and maturation of the nascent

peri-lesional vessels, hence contributing to persistent peri-lesional hyperperfusion in

FR122047 treated rats. In contrast, greater peri-lesional neuronal loss in the placebo

rats may have led to more extensive pruning during the maturation of the nascent

vessels, thus rendering some of them non-perfused [Korn et al. 2015], accounting for

the reduction in peri-lesional perfusion with time in the placebo animals. Peri-lesional

hyper-reactivity to hypercapnia similarly may have been due to the lower resistance of

nascent vessels [White et al. 1992]. Persistent peri-lesional hyperperfusion in treated

rats may have also been influenced by an FR122047-mediated reduction in PGs that

promote vaso-constriction and are predominately produced via COX-1 in activated

microglia [Yagami et al. 2015, Kobayashi et al. 2004, McAdam et al. 1999, Giulian et al.

1996], whose peri-lesional density was decreased in FR122047 treated relative to

placebo rats. Reduced microglial/macrophage density in FR122047 treated rats is in

general agreement with previous reports on COX-1 inhibition effects in other models of

chronic inflammatory responses [Choi et al. 2009, McKee et al. 2008, Nomura et al.

2011].

The present study provided evidence of beneficial effects of delayed, selective,

cerebral COX-1 inhibition on the neurogliovascular unit in the peri-infarct zone.

Treatment resulted in increased peri-lesional neuronal survival, decreased recruitment

of microglia/macrophages, and sustained elevation in peri-lesional perfusion. These

histopathological and imaging findings warrant further investigation of the effect of

selective COX-1 inhibition on functional outcome in the chronic stage of stroke recovery.

Furthermore, future measurements of PGs levels post COX-1 inhibition, along with in

117

vivo measurements of changes in vascular architecture and function as well as neuronal

excitability will further the mechanistic understanding of the observed effects. In

addition, future studies which include behavioural testing to evaluate the effects of

FR122047 treatment on functional outcome are needed as metrics of behavioural

changes are the gold standard outcome measure of recovery from ischemic injury

[Murphy & Corbett 2009].

118

Chapter 5

There is a pressing need to identify strategies to improve functional recovery

over weeks and months- following ischemic injury [Teasell et al. 2014]. That there are

not more effective means to rehabilitate this growing patient population largely reflects

the uncertainty surrounding the underlying mechanisms that govern long-term recovery

following ischemic injury. The present work investigated the complex relationship

between structure and function of brain neuronal and vascular networks during the

weeks following unilateral focal ischemic stroke. Observations of concomitant structural,

functional, and behavioural changes over several weeks following ischemic insult were

reported during both endogenous recovery as well as with the application of L-655,708,

a GABA-receptor inverse agonist to increase neuronal excitability and FR122047, a

selective COX-1 inhibitor.

5.1 Modulating excitability

Following an ischemic injury, some sensory and motor function performed by

injured tissue prior to infarction is remapped to contra-lesional as well as peri-lesional

areas. Remapping is manifested as gross physiological changes in the responsiveness

of neuronal networks demonstrating widespread functional plasticity during the weeks-

months following injury [Murphy & Corbett 2009, Carmichael et al. 2012]. There is

growing interest in the potential that modulating excitability may have in facilitating

remapping; particularly given the evidence of improved recovery in patients receiving

trans-cranial direct current stimulation (tDCS) treatment [cf. Review by Khedr et al.

2010, Floel 2014 and Jones et al. 2015]. tDCS is thought to enhance motor learning and

increase synaptic plasticity [Fritsch et al. 2010, Monte-Silva et al. 2013, Stagg et al.

2009] by inducing subthreshold membrane depolarization (anodal tDCS), and/or

hyperpolarization (cathodal tDCS) [cf. Review by Savic et al. 2016]. These effects are

governed by altered NMDA and GABA receptor activity [Fritsch et al. 2010, Stagg et al.

2009]. Although, tDCS treatment remains to be optimized [Khedr et al. 2010], the

119

prevailing opinion is that higher session frequency results in greater functional benefit

[Jones et al. 2015, Monte-Silva et al. 2013, Floel 2014].

Pharmacological manipulations designed to enhance neuronal plasticity by

modulating excitability following stroke have also been proposed as a means to improve

outcome (e.g. blockers of tonic GABA activity and positive allosteric modulators of

AMPA (α-amino-3-hydroxyl-5-methyl-4-isoxazolepropionate) receptors) [cf. Review by

Carmichael et al. 2012]. Agents that target α-5 containing GABAA receptors (which

spare phasic GABA signalling) regulate cognitive processing, and play key roles in

memory acquisition, consolidation, and retrieval [Glykys et al. 2007, Gabriella et al.

2010, Carmichael et al. 2012]. The work presented in Chapter 2 investigated the effects

of L-655,708 (an inverse α-5 containing GABAA receptor agonist) and found that in the

subacute stage, L-655,708 elicits a decrease in stroke volume and improves skilled

reaching ability. This work provides further evidence that modulating excitability in the

subacute stage may be a promising therapeutic strategy for ischemic stroke.

A potential avenue of future investigations would be to combine L-655,708

treatment with tDCS and physical therapy (e.g. enriched environment and/or daily reach

training [BBiernaskie et al. 2001, Clarke et al. 2014]). It is conceivable that the combined

therapeutic effect would exceed the sum of the functional gains from either strategy

does alone. Moreover, fMRI measurements (e.g. BOLD and/or ASL) and

electrophysiological recordings could reveal more widespread evidence of remodelling

facilitated by L-655,708 treatment, or evidence of contra- or ipsi-lesional

remapping/recruitment, furthering our understanding of the underlying processes of

remodelling. A review of the strengths and challenges in applying fMRI in preclinical

modeling of stroke is presented in Appendix 2. Finally, it would be of great utility for the

development of GABA inhibition based intervention to examine changes in local

concentration of GABA with magnetic resonance spectroscopy (MRS), which has been

previously employed to demonstrate tDCS-mediated changes in GABA [Stagg et al.

2009]. These data would offer insight into downstream effects of GABA antagonism and

help to identify the therapeutic window for treatments which modulate excitability.

120

5.2 Endogenous neuro/angio-genesis

Brain exhibits endogenous angio- and neuro-genesis post-stroke [cf. Reviews by

Wiltrout et al. 2007, Liu et al. 2014, Sawada et al. 2014, Ruan et al. 2015, Marlier et al.

2015, Yu et al., 2016]. Promoting endogenous repair following neuronal loss caused by

injury or disease is thought to have great potential to improve outcome for many

patients [cf. Review by Jessberger et al. 2016]. Specifically, stroke induces significant

vascular outgrowth [Ruan et al. 2015], which creates a unique micro-environment

necessary for neurogenesis in the adult brain [Sawada et al. 2014, Thored et al. 2007,

Kojima et al. 2010, Grade et al. 2013, Ohab et al. 2006, Gerlai et al. 2000, Marlier et al.

2015, Nishijima et al. 2010, Greenberg et al. 2013, Jin et al. 2002, Shen et al. 2004,

Zhang et al. 2015]. Thus, enhancing injury-induced angiogenesis has been studied as a

potential therapeutic strategy in preclinical models of ischemia [Ding et al. 2006, Hu et

al. 2010, Wang et al. 2015, Tang et al. 2010, Sun et al. 2003]. For example, exercise

(either before or after stroke) increases vessel density in the peri-infarct zone and

improves functional outcome. In addition, a handful of pharmacological agents have

been identified to enhance neuro-/angio-genesis when administered 24 hours following

ischemia in animal models [Zhou et al. 2001, Manwani et al. 2013, Selvin et al. 2008,

Jin et al. 2014, Mantovani et al. 2013, Kiegerl et al. 2009, Corbett et al. 2015, Hu et al.

2016]. However, insight into endogenous neuro-/angio-genesis in preclinical models has

predominantly been gleaned from invasive or terminal experiments (e.g. microscopy,

and histochemistry) [cf. Review by Jiang et al. 2016]. Neuroimaging assays that may be

conducted both preclinically and in patients may likely help translate what has been

learnt in animal models to the clinic. Further, non-invasive methodologies allow the

assessment of tissue remodelling longitudinally and can provide valuable in vivo

measurements of function, which is currently not well characterized [Jiang et al. 2016].

Chapter 3 built upon a handful of preclinical studies that have reported transient

increases in CBF and/or CBV in peri-lesional tissue during the weeks following focal

ischemia and linked these findings to injury-induced angiogenesis on

immunohistochemistry. Concomitant to marked hemodynamic changes on in vivo ASL,

we observed evidence of both evoked and spontaneous neuronal activity that are likely

manifestations of remapping. However, the presently observed neurophysiological and

121

hemodynamic changes were not accompanied by improved skilled reaching ability,

which may be due to a highly non-linear relationship between neurophysiological state

and behaviour. Moreover, functional gains may had been apparent on other behavioural

tests.

Future investigations examining the time course of neuronal activity changes

following injury would profit from being conducted in awake and behaving animals with

permanently implanted intra-cranial electrodes [cf. Review by Oliveira et al. 2008]. Such

an experiment would reveal more precisely when changes in evoked and spontaneous

neuronal activity arise and thus more clearly define when interventions (to spur

angiogenesis or raise excitability) might be most beneficial. Alternatively, investigating

the effects of pro-angiogenic treatment (e.g. VEGF or Ang-1), exercise (endogenously

pro-angiogenic) or enhancers of neuronal excitability (e.g. L-655,708 as in Chapter 2)

on CBF and CBF elicited changes, neuronal-excitability and behaviour may reveal

neuro-imaging biomarkers of functional improvement. Finally, as a plateau was not

observed, the observation period should be extended to assess longer term changes.

5.3 Inflammation

Hitherto, the majority of candidate anti-inflammatory therapeutics have been

applied in the acute period [cf. Review by Spite et al. 2010, BIadecola et al. 2011 and

Kim et al. 2014]. Chronically, continued inflammatory processes have both deleterious

and beneficial effects on remodelling and repair [BIadecola et al. 2011, Spite et al.

2010]. Among the anti-inflammatories preclinical studies, a large proportoin find COX-2

inhibition to improve outcome [cf. Reviews by Spite et al. 2010, Yagami et al. 2015].

However, long-term placebo-controlled clinical studies reveal unfortunate cardiovascular

side-effects of COX-2 inhibition [Ott et al 2003, Nussmeier et al. 2005, Cannon et al.

2006, Blobaum et al. 2006], curbing translation of these agents. In the study presented

in Chapter 4, we reproduced the fMRI results presented in Chapter 3 in the vehicle-only

control group and demonstrated that 12 days of intracerebroventricular administration of

FR122047 (a selective COX-1 inhibitor) preserved peri-lesional hyperperfusion,

increased neuronal survival, and decreased microglia and macrophage recruitment and

activation. To the best of our knowledge, the present work is the first to examine

122

selective pharmacological COX-1 inhibition as a focal ischemic stroke treatment

strategy in the subacute phase in vivo. These findings warrant further investigation of

the effect of selective COX-1 inhibition on functional outcome, treatment optimization,

and the underlying mechanisms of FR122047-mediated changes in the

neurogliovascular unit, especially the underlying effects on PG-D2 and TX-A2

production. Furthermore, we presently administered FR122047 intracerebroventricularly

to avoid off-target effects (as COX-1 is expressed throughout the body) [Perrone et al.

2010]. However, in an effort to increase the translational potential of this treatment

strategy, future work should endeavour to test a less invasive means of delivering

FR122047 treatment. In addition, combining L-655,708 and FR122047 treatment to

simultaneously combat the deleterious effects of hyper-excitability and neuro-

inflammation might prove to be a powerful combinatorial subacute intervention and

should also be investigated.

The present work was conducted exclusively in male animals. Before

menopause, women suffer fewer strokes than do men [Stagmeyr et al. 1997, Sudlow

and Warlow et al. 1997]. Following menopause the incidence of stroke equalizes

between women and men [Anderson et al. 1991, Hurn and Macrea 2000] possibly as a

result of the lost neuroprotective effects of estrogen [Hurn and Macrea 2000] and/or the

associated higher cerebral blood flow of pre-menopausal women [Esposito et al. 1996].

However, sex differences are present in the pediatric population and persist following

menopause suggesting that the prescence of reproductive steroids do not wholly

account for the sexual dimorphism in stroke [cf. Review by Liu et al. 2009]. Notably,

outcome after an ischemic injury in post-menopausal women is worse than men

[Bushnell 2008]. Specifically, women suffer from more severe disabilities [Bushnell

2008], are more likely to be institutionalized [Bushnell 2008], and less likely to reach

independence in instrumental activities [Lai et al. 2005]. Furthermore, there are sex

differences in neuroprotection and cell death [Liu et al. 2009], auto-regulation by

immune cells [Tipton & Sullivan 2014] and response to rehabilitation [Langdon et al.

2014]. Future investigations of the present findings should thus be conducted in both

male and female animals. These investigations may address behavioural phenotyping

differences both between sexes and across straing [Meziane et al. 2007].

123

5.4 Limitations of Normalization

In the present work, functional MRI measures (perfusion and reactivity)

presented in Chapters 3 and 4 were obtained by dividing the average parameter

estimate within the ipsilateral ROI by the average parameter estimate within the

contralateral ROI in each animal. These ratios were then averaged within groups. This

division was necessary in Chapter 3 due to variation in the signal-to-noise ratio between

recordings on different imaging days. In Chapter 4, we report both the absolute signal

change in right and left ROIs as well as the inter-hemispheric ratio. Although commonly

employed, use of the contralateral hemisphere signal as an 'internal reference' is

confounding as the contralateral hemisphere has been shown to be involved in the

progression of ischemic injury [c.f. Review in Appendix 2]. Notwithstanding, comparing

the ipsilateral measurements to the whole brain parameter estimates (as was done in

Chapter 3), or contrasting the non-normalized signal changes (Chapter 4) yielded the

same contrasts. While preferable, absolute quantification of CBF was not undertaken in

the present work. It would have entailed T1 relaxometry and inversion efficiency

measurements in each subject, in addition to the quantification of steady-state arterial

magnetization, thus significantly prolonging the MRI protocol, which would have likely

increased the attrition given the fragile state of these animals.

In Chapter 3 we report normalized electrophysiological data on evoked and

spontaneous activity. In these experiments, the signal from neighbouring electrodes

was first subtracted to increase sensitivity to changes in the neighbourhood of each

electrode. In the evoked activity experiment, each pairwise difference signal traces was

then averaged across 10 stimulus presentations and the response amplitude of that

average trace estimated; and divided by the maximum response amplitude recorded by

that array. These values were then averaged across spatially corresponding positions in

different animals and resulting across-subject averages plotted as a function of distance

from Bregma. In the spontaneous activity experiment, the FFT of the recorded signal

was computed for each electrode and the power within each frequency band of interest

averaged across all electrodes of the array. In each animal, the average power

recorded by the ipsilateral array was then divided by the average power recorded by the

contra-lateral array. The sources of noise in these experiments were differences

124

between the right and left craniotomies, the hardware differences between the two

MEAs, as well as environmental noise sources, necessitating normalization. To reduce

systematic error, we randomized the placement of each MEA between right vs. left

hemisphere.

In Chapter 3, the area occupied by positively stained cells in

immunohistochemical analysis was also normalized. For the ipsilateral and contralateral

ROIs in each animal, the fractional area occupied by positively stained cells was

computed and the ipsilateral ROI value was then divided by the contralateral ROI value.

The inter-hemispheric ratios were then averaged across animals within each group. For

each stain, all sections were prepared simultaneously. Furthermore, all imaging and

data processing parameters were held constant for all sections. As with the fMRI

analysis, normalization of the immunohistochemical data thus used the contralateral

hemisphere value as an internal control. In Chapter 4, we reported the fractional area

occupied by the cell types of interest without normalization, yet observed the same

contrasts in the placebo animals as those observed in untreated animals of Chapter 3.

Future work should further characterize the immunohistochemical results reported in the

present work using a more rigorous methodology (e.g., stereology).

5.5 Conclusion

As the vast majority of patients arrive at a care facility well outside of the window

of opportunity for treatment using current therapies, identifying ways through which

outcome can be improved during the subacute and chronic stages is of utmost

importance. The present work investigated endogenous neurovascular processes of

subacute recovery as well as two pharmacological interventions: sustained low-dose

GABAA antagonism (L-655,708) and selective cyclooxygenase-1 (COX-1) inhibition

(FR122047). On the whole, the present work shows the complexity of the relationship

between neurovascular structure, function, and behaviour, and emphasized the need for

multi-modal characterization of neurophysiological state over prolonged observation

periods following ischemic insult. Furthermore, the beneficial effects of L-655,708 and

FR122047 when administered in the subacute stage of injury progression warrant

further investigation of the underlying mechanisms of action. All together, these data

125

show new evidence of highly dynamic structural and functional remodelling within the

neurogliolvascular network over the weeks following ischemic injury and the successful

pharmacological modulation of these processes to improve outcome.

126

References

Ahmad M, Ahmad AS, Zhuang H, et al. Stimulation of prostaglandin E2-EP3 receptors exacerbates stroke and excitotoxic injury. J Neuroimmunol. 2007;184:172–179

Aid S, Bosetti F. Targeting cyclooxygenases-1 and -2 in neuro-inflammation: Therapeutic implications. Biochimie. 2011;93:46–51

Alexandrov AV, Clack SE, Ehrlich LE, Bladin CF, Smurawska LT, Pirisi A, Caldwell CB. Simple visual analysis of brain perfusion on HMPAO SPECT predicts early outcome in acute stroke. Stroke. 1996;26:1537-42

Allan SM, Rothwell NJ. Cytokines and acute neurodegeneration. Nat Rev Neurosci. 2001: 734–744

Altamura C, Torquati K, Zappasodi F, Ferretti A, Pizzella V, Tibuzzi F, Vernieri F, Pasqualetti P, Landi D, Del Gratta C et al. fMRI-vs-MEG evaluation of post-stroke interhemispheric asymmetries in primary sensorimotor hand areas. Exp Neurol. 2007;204(2):631-9

Anderson KM, Wilson PW, Odell PM, Kannel WB. An updated coronary risk profile. I statement for health professionals. Circulation. 1991;83:356-62

Atack JR, Pike A, Clarke A, Cook SM, Sohal B, McKernan RM et al. Rat pharmacokinetics and pharmacodynamics of a sustained release formulation of the GABA(A) Alpha-5-selective compound L-655,708. Drug Metab Dispos. 2006;34:887-93

Atochin DN, Wang A, Liu VW, Critchlow JD, Dantas AP, Looft-Wilson R, Murata T, Salomone S, Shin HK, Ayata C, Moskowitz MA, Michel T, Sessa WC, Huang PL. The phosphorylation state of eNOS modulates vascular reactivity and outcome of cerebral ischemia in vivo. J Clin Invest. 2007;117(7):1961-7

Astrakas LG, Naqvi SHA, Kateb B, Tzika AA. Functional MRI using robotic MRI compatible devices for monitoring rehabilitation from chronic stroke in the molecular medicine era (review). Int J Mol Med. 2012; 29(6):963-73

Baaven RH, Davidson DJ, Bates DM. Mixed-effects modelling with crossed random effects for subjects and items. J Mem Lang. 2008:390-412

Barreca S, Gowland C, Stratford P, et al. Development of the Chedoke arm and hand activity inventory: item selection. Physiother Can. 1999;209-11

Basu A, Lazovic J, Krady JK, Mauger DT, Rothstein RP, Smith MB, Levison SW. Interleukin-1 and the interleukin-1 type 1 receptor are essential for the progressive neurodegeneration that ensues subsequent to a mild hypoxic/ischemic injury. 2005

Becker K, Kindrick D, Relton J, Harlan J, Winn R. Antibody to the alpha4 integrin decreases infarct size in transient focal cerebral ischemia in rats. Stroke. 2001;32(1):206–211

Beech JS, Williams SC, Campbell CA, Bath PM, Parsons AA, Hunter AJ, et al. Further characterisation of a thromboembolic model of stroke in the rat. Brain Res. 2001;895:18 –24

Berger C, Schabitz WR, Georgiadis D, Steiner T, Aschoff A, Schwab S. Effects of hypothermia on excitatory amino acids and metabolism in stroke patients: a microdialysis study. Stroke 2002;33:519–24

127

Berrouschot J, Sterker M, Bettin S, Koster J, Schneider D. Mortal of space-occupying (“malignant”) middle cerebral artery infarction under conservative intensive care. Intensive Care Med. 1998;24:620–3

Bhatia R, Hill MD, Shobha N, Menon B, Bal S, Kochar P, Watson T, Goyal M, Demchuk AM. Low rates of acute recanalization with intravenous recombinant tissue plasminogen activator in ischemic stroke: real-world experience and a call for action. Stroke 2010;41:2254–8

ABiernaskie J, Corbett D, Peeling J, Wells J, Lei H. A serial MR study of cerebral blood flow changes and lesion development following endothelin-1-induced ischemia in rats. Magn Reson Med. 2001;46:827-30

BBiernaskie J, Corbett D. Enriched rehabilitative training promotes improved forelimb motor function and enhanced dendritic growth following focal ischemic injury. J Neurosci. 2001;21: 5272–80

Biernaskie J, Chernenko G, Corbett D. Efficacy of rehabilitative experience declines with time after focal ischemic brain injury. J Neurosci. 2004;24:1245–4

Binkofski F, Seitz RJ. Modulation of the BOLD- response in early recovery from sensorimotor stroke. Neurology. 2004;63(7):1223-9

Blicher JU, Stagg CJ, O’Shea J, Østergaard L, MacIntosh BJ, Johansen-Berg H, Jezzard P, Donahue MJ. Visualization of altered neurovascular coupling in chronic stroke patients using multi-modal functional MRI. J Cereb Blood Flow Metab. 2012(11):2044-54

Blobaum AL, Marnett LJ. NSAID action and the foundations for cardiovascular toxicity. Cardiotoxicity Non-Cardiovascular Drugs. 2010;257–285

Bolay H, Dalkara T. Mechanisms of motor dysfunction after transient MCA occlusion: persistent transmission failure in cortical synapses is a major determinant. Stroke. 1998;29(9):1988-93

Bours MJ, Swennen EL, Di Virgilio F, Cronstein BN, Dagnelie PC. Adenosine 5'-triphosphate and adenosine as endogenous signaling molecules in immunity and inflammation. Pharmacol Ther. 2006;112:358–404

Boutin H, LeFeuvre RA, Horai R, Asano M, Iwakura Y, Rothwell NJ. Role of IL-1alpha and IL-1beta in ischemic brain damage. J Neurosci. 2001;21(15):5528–5534

Brown CE, Aminoltejari K, Erb H, Winship IR, Murphy TH. In vivo voltage-sensitive dye imaging in adult mice reveals that somatosensory maps lost to stroke are replaced over weeks by new structural and functional circuits with prolonged modes of activation within both the peri-infarct zone and distant sites. J Neurosci. 2009;29:1719-34

Buetefisch CM. Role of the Contralesional Hemisphere in Post-Stroke Recovery of Upper Extremity Motor Function. Front Neurol. 2015;6:214

Buga AM, Di Napoli M, Popa-Wagner A. Preclinical models of stroke in aged animals with or without comorbidities: role of neuro-inflammation. Biogerontology. 2013;14(6):651-62

Buma FE, Lindeman E, Ramsey NF, Kwakkel G. Functional neuroimaging studies of early upper limb recovery after stroke: a systematic review of the literature. Neurorehabil Neural Repair. 2010;24(7):589-608

Bune LT, Thaning P, Johansson PI, Bochsen L, Rosenmeier JB. Effects of nucleotides and nucleosides on coagulation. Blood Coagul Fibrinolysis. 2010;21:436–41

Burnstock G, Ralevic V. Purinergic signaling and blood vessels in health and disease. Pharmacol Rev. 2013;66(1):102-92

128

Bushnell CD. Stroke and the female brain. Nature Clinical Practice Neurology. 2008;4:22-33

Buxton RB,Wong EC, Frank LR. Dynamics of blood flow and oxygenation changes during brain activation: the balloon model. Magn Reson Med. 1998;39(6):855–64

Candelario-Jalil E, González-Falcón A, García-Cabrera M, Álvarez D, Al-Dalain S, Martínez-Sánchez G, León OS, Sprinter JE. Assessment of the relative contribution of COX-1 and COX-2 isoforms to ischemia-induced oxidative damage and eurodegeneration following transient global cerebral ischemia. J Neurochem. 2003;86(3):545-55

Candelario-Jalil E, Pinheiro de Oliveira AC, Gräf S, Bhatia HS, Hüll M, Muñoz E, Fiebich BL. Resveratrol potently reduces prostaglandin E2 production and free radical formation in lipopolysaccharide-activated primary rat microglia. J neuro-inflammation. 2007;4:25

Cannon CP, Curtis SP, FitzGerald GA, Krum H, Kaur A, Bolognese JA, Reicin AS, Bombardier C, Weinblatt ME, van der Heijde D, Erdmann E, Laine L. Cardiovascular outcomes with etoricoxib and diclofenac in patients with osteoarthritis and rheumatoid arthritis in the multinational etoricoxib and diclofenac arthritis long-term (MEDAL) programme: A randomized comparison. Lancet 2006;368:1771–1781

Cao Y, D’Olhaberriague L, Vikingstad EM, Levine SR, Welch KM. Pilot study of functional MRI to assess cerebral activation of motor function after poststroke hemiparesis. Stroke. 1998;29(1):112-22

Carden DL, Granger DN. Pathophysiology of ischaemia-reperfusion injury. J Pathol. 2000; 190:255–66

Cardona AE, et al. Control of microglial neurotoxicity by the fractalkine receptor. Nat Neurosci. 2006;9:917–24

Carey JR, Kimberley TJ, Lewis SM, Auerbach E, Dorsey L, Rundquist P, Ugurbil K. Analysis of fMRI and finger tracking training in subjects with chronic stroke. Brain. 2002;125:773–788

Carmichael ST, Wei L, Rovainen CM, Woolsey TA. New patterns of intracortical projections after focal cortical stroke. Neurobiol Dis. 2001;8(5):910-22

Carmichael ST. Rodent models of focal stroke: size, mechanism and purpose. NeuroRx. 2005;2:396-409

Carmichael ST. Translating the frontiers of brain repair to treatments: starting not to break the rules. Neurobiol Dis. 2010;37:237–42

Carmichael TS, Brain excitability in stroke: the yin and yang of stroke progression. Arch. Neurol. 2012;69:161-7

Caro JJ, Huybrechts KF, Duchesne I. Management patterns and costs of acute ischemic stroke : an international study. Stroke. 2000;31:582–90

Center for Biologics Evaluation and Research. US License 1048: Activase ® . Rockville, Food and Drug Administration, 2003

Chapman GA, et al. Fractalkine cleavage from neuronal membranes represents an acute event in the inflammatory response to excitotoxic brain damage. J Neurosci. 2000;20:RC87

Chang CH, et al. High resolution structural and functional assessments of cerebral microvasculature using 3D Gas delta-R2*-mMRA. PLoS One. 2013;8:1-10

Chen JJ, Pike GB. Global cerebral oxidative metabolism during hypercapnia and

129

hypocapnia in humans: implications for BOLD fMRI. J Cereb Blood Flow Metab, 2010;6:1094-9

Chen J, Venkat P, Zacharek A, Chopp M. Neurorestorative therapy for stroke. Front Hum Neurosci. 2014:8(318);1-12

Choi SH, Bosetti F. Cyclooxygenase-1 null mice show reduced neuro-inflammation in response to β-amyloid. Aging 2009;1:234–244

Cinelli P, Madani R, Tsuzuki N, Vallet P, Arras M, Zhao CN, Osterwalder T, Rulicke T, Sonderegger P. Neuroserpin, a neuroprotective factor in focal ischemic stroke. Mol Cell Neurosci. 2001;18(5):443–457

Clark WM, Rinker LG, Lessov NS, Hazel K, Hill JK, Stenzel-Poore M, Eckenstein F. Lack of interleukin-6 expression is not protective against focal central nervous system ischemia. Stroke. 2000;31(7):1715–1720

Clarke J, Langdon KD, Corbett D. Early poststroke experience differentially alters periinfarct layer II and III cortex. J Cereb Blood Flow Metab. 2014;34(4):630-7

Clarkson AN, Huang BS, Macisaac SE, Mody I, Carmichael T. Reducing excessive GABA-mediated tonic inhibition promotes functional recovery after stroke. Nature. 2010;468:305-9

Committee for Proprietary Medicinal Products (CPMP): Actilyse. Annex I. Amended Summary of Product Characteristics of the Reference Member State. The European Agency for the Evaluation of Medical Products (EMEA). 2002

Corbett AM, Sieber S, Wyatt N, Lizzi J, Flannery T, Sibbit B, Sanghvi S. Increasing neurogenesis with fluoxetine, simvastatin and ascorbic Acid leads to functional recovery in ischemic stroke. Recent Pat Drug Deliv Formul. 2015;9(2):158-66

Cox RW. AFNI: software for analysis and visualization of functional magnetic resonance neuroimages. Comput Biomed Res. 1996;29:162-73

Cramer SC, Nelles G, Benson RR, Kaplan JD, Parker RA, Kwong KK, Kennedy DN, Finklestein SP, Rosen BR. A functional MRI study of subjects recovered from hemiparetic stroke. Stroke 1997; 28:2518–27

Chung S, You J, Tack G, Yi J, Lee B. The mean intensity of activation than the number of activated voxels is a better index to reflect the levels of visuospatial performance. Conf Proc IEEE Eng Med Biol Soc. 2005;5:5298–5301

Chung SC, Sohn JH, Lee B, Tack GR, Yi JH, You JH, Kwon JH, Kim HJ, Lee SY. A comparison of the mean signal change method and the voxel count method to evaluate the sensitivity of individual variability in visuospatial performance. Neurosci Lett. 2007;418:138–142

Da Fonseca ACC, Matias D, Garcia C, et al. The impact of microglial activation on blood-brain barrier in brain diseases. Frontiers in Cellular Neuroscience. 2014;8:362

Davis TL, Kwong KK,Weisskoff RM, Rosen BR. Calibrated functional MRI: mapping the dynamics of oxidative metabolism. Proc Natl Acad Sci USA. 1998;95(4):1834–9

De Felipe L, Alonso-Nanclares L, Arellano JL. Microstructure of the neocortex: Comparative aspects. J Neurocytol. 2002;31(3-5):299-316

De Girolami U, Crowell RM, Marcoux FW. Selective necrosis and total necrosis in focal cerebral ischemia: neuropathologic observations on experimental middle cerebral artery occlusion in the macaque monkey. J Neuropathol Exp Neurol.

130

1984;43(1):57-71 del Zoppo GJ, Schmid-Schonbein GW, Mori E, Copeland BR, Chang CM.

Polymorphonuclear leukocytes occlude capillaries following middle cerebral artery occlusion and reperfusion in baboons. Stroke. 1991;22:1276–83

del Zoppo GJ. Inflammation and the neurovascular unit in the setting of focal cerebral ischemia. Neuroscience. 2009;158(3):972-82

den Haan JMM, Arens R, van Zelm MC. The activation of the adaptive immune system: Cross-talk between antigen-presenting cells, T cells and B cells. Immunology Letters. 2014;162(2):103–12

Dietrich WD, Ginsberg MD, Busto R, Watson BD. Photochemically induced cortical infarction in the rat. 2. Acute and subacute alterations in local glucose utilization. J Cereb Blood Flow Metab . 1986;6:195–202

Dietrich WD, Watson BD, Busto R, Ginsberg MD. Metabolic plasticity following cortical infarction: a 2-deoxyglucose study. In: Cerebrovascular disorders (Raichel ME, Powers WJ, eds). New York: Raven Press. 1987;285–95

Dijkhuizen RM, Singhal AB, Mandeville JB, Wu O, Halpern EF, Finklestein SP, Rosen BR, Lo EH. Correlation between brain reorganization, ischemic damage, and neurologic status after transient focal cerebral ischemia in rats: a functional magnetic resonance imaging study. J Neurosci. 2003;23(2):510-7

Ding YH, Li J, Yao WX, Rafols JA, Clark JC, Ding Y. Exercise preconditioning upregulates cerebral integrins and enhances cerebrovascular integrity in ischemic rats. Acta Neuropathol. 2006;112:74–84

Ding G, Jiang Q, Li L, Zhang L, Wang Y, Zhang ZG et al. Cerebral tissue repair and atrophy after embolic stroke in rat: a magnetic resonance imaging study of erythropoietin therapy. J Neurosci Res. 2010;88(14): 3206-14

Dittmar M, Spruss T, Schuierer G, Horn M. External carotid artery territory ischemia impairs outcome in the endovascular filament model of middle cerebral artery occlusion in rats. Stroke. 2003;34:2252-7

Donovan NJ, Kendall DL, Heaton SC, Kwon S, Velozo CA, Duncan PW. Conceptualizing functional cognition in stroke. Neurorehabil Neural Repair. 2008;22(2):122–35

Duncan PW, Lai SM, Keighley J. Defining post-stroke recovery: implications for design and interpretation of drug trials. Neuropharmacology. 2000;39(5):835–41

Eltzschig HK, Carmeliet P. Hypoxia and inflammation. N Engl J Med. 2011;364:656–65 Endres M, Engelhardt, B, Koistinaho, J, Lindvall, O, Meairs, S, Mohr, JP, Planas, A,

Rothwell, N, Schwaninger, M, Schwab, ME, Vivien, D, Wieloch, T, Dirnagl U. Improving outcome after stroke: overcoming the translational roadblock. Cerebrovasc Dis. 2008;25(3):268-78

Engelhardt B, Sorokin L. The blood-brain and the blood-cerebrospinal fluid barriers: function and dysfunction. Semin Immunopathol. 2009;31:497–511

Esposito G, Van Horn JD, Weinberger DR, Berman KF. Gender differences in cerebral blood flow as a function of cognitive state with PET. Journal of nuclear medicine: official publication Society of Nuclear Medicine 1996;37:559-64

Evenson KR, Rosamond WD, Morris DL. Prehospital and in-hospital delays in acute stroke care. Neuroepidemiology. 2001;20:65-76

Fallon JH. The islands of Calleja complex of rat basal forebrain II: connections of medium and largesized cells. Brain Resb. Bull. 1983;10:775–93

131

Fasoli SE, Krebs HI, Stein J, Frontera WR and Hogan N: Effects of robotic therapy on motor impairment and recovery in chronic stroke. Arch Phys Med Rehabil. 2003;84(4):477-82

Fasoli SE, Krebs HI, Stein J, Frontera WR, Hughes R and Hogan N. Robotic therapy for chronic motor impairments after stroke: follow-up results. Arch Phys Med

Rehabil. 2004;85(7): 1106‐11 Feigin VL, Carlene MM Lawes, Derrick A Bennett, and Craig S Anderson. Stroke

epidemiology: a review of population- based studies of incidence, prevalence, and case-fatality in the late 20th century . Lancet Neurology 2003;2:43-553

Feigin VL, Forouzanfar MH, Krishnamurthi R, et al. Global and regional burden of stroke during 1990–2010: findings from the Global Burden of Disease Study 2010. Lancet. 2014;383(9913):245-54

Félétou M, Huang Y, Vanhoutte PM. Endothelium-mediated control of vascular tone: COX-1 and COX-2 products. Br J Pharmacol. 2011;164(3):894-912

Feydy A, Carlier R, Roby-Brami A, Bussel B, Cazalis F, Pierot L, Burnod Y, Maier MA. Longitudinal study of motor recovery after stroke: recruitment and focusing of brain activation. Stroke. 2002;33(6):1610-7

Floel A. tDCS-enhanced motor and cognitive function in neurological diseases. Neuroimage 2014;85: 934-47

Frank-Cannon TC, Alto LT, McAlpine FE, Tansey MG. Does neuro-inflammation fan the flame in neurodegenerative diseases? Mol Neurodegener. 2009;4:47

Fritsch B, Reis J, Martinowich K, Schambra HM, Ji Y, Cohen LG, Lu B. Direct current stimulation promotes BDNF-dependent synaptic plasticity: potential implications for motor learning. Neuron . 2010;66:198-204

Gabriella G, Giovanna C. γ-Aminobutyric acid type A (GABA(A)) receptor subtype inverse agonists as therapeutic agents in cognition. Methods Enzymol. 2010;485:197–211

Gerlai R, Thibodeaux H, Palmer JT, van Lookeren Campagne M, Van Bruggen N. Transient focal cerebral ischemia induces sensorimotor deficits in mice. Behav Brain Res. 2000;108:63–71

AGerriets T, Stolz E, Walberer M, Kaps M, Bachmann G, Fisher M. Neuroprotective effects of MK-801 in different rat stroke models for permanent middle cerebral artery occlusion: adverse effects of hypothalamic damage and strategies for its avoidance. Stroke. 2003;34:2234 –9

BGerriets T, Li F, Silva MD, Meng X, Brevard M, Sotak CH, Fisher M. The macrosphere model: evaluation of a new stroke model for permanent middle cerebral artery occlusion in rats. J Neurosci Methods. 2003;122:201–11

Gerriets T, Stolz E, Walberer M, Mȕller C, Rottger C, Kluge A, Manfred K , Fisher M, Bachmann G. Complications and pitfalls in rat stroke models for middle cerebral artery occlusion. Stroke. 2004;35:2372-7

Gertz K, Kronenberg G, Kalin RE, Baldinger T, Werner C, Balkaya M, Eom GD, Hellmann- Regen J, Krober J, Miller KR, Lindauer U, Laufs U, Dirnagl U, Heppner FL, Endres M. Essential role of interleukin-6 in post-stroke angiogenesis. Brain. 2012;135(Pt 6):1964–1980

Giulian D, Corpuz M, Richmond B, Wendt E, Hall ER. Activated microglia are the principal glial source of thromboxane in the central nervous system. Neurochem Int. 1996; 29(1):65–76

132

Gladstone DJ, Danells CJ, Black SE. The fugl-meyer assessment of motor recovery after stroke: a critical review of its measurement properties. Neurorehabil Neural Repair. 2002;16(3):232-40

Globus MY, Alonso O, Dietrich WD, Busto R, Ginsberg MD. Glutamate release and free radical production following brain injury: effects of posttraumatic hypothermia. J Neurochem 1995;65:1704–11

Glykys J, Mody I. Activation of GABAA receptors: views from outside the synaptic cleft. Neron. 2007;56(5):763–70

Goussev AV, Zhang Z, Anderson DC, Chopp M. P-selectin antibody reduces hemorrhage and infarct volume resulting from MCA occlusion in the rat. J Neurol Sci. 1998;161(1):16–22

Gowland C, VanHullenaar S, Torresin W, et al. Chedoke-McMaster stroke assessment: development, validation and administration manual. Hamilton (ON), Canada: Chedoke-McMaster Hospitals and McMaster University; 1995

Grade S, Weng YC, Snapyan M, Kriz J, Malva JO, Saghatelyan A. Brain-derived neurotrophic factor promotes vasculature-associated migration of neuronal precursors toward the ischemic striatum. PLoS One. 2013;8(1):e55039

Greenberg DA, Jin K. Growth factors and stroke. NeuroRx. 2006;3:458–65 Greenberg DA, Jin K. Vascular endothelial growth factors (VEGFs) and stroke. Cell Mol

Life Sci. 2013;70(10):1753-61 Griffin KM, Blau CW, Kelly ME, O'Herlihy C, O'Connell PR, Jones JF, Kerskens CM.

Propofol allows precise quantitative arterial spin labelling functional magnetic resonance imaging in the rat. Neuroimage. 2010;51(4):1395-404.

Grotta JC, Jacobs TP, Koroshetz WJ, Moskowitz MA. Stroke Program Reviews Group: An interim report. Stroke; 2008(39):1364-70

Gualtieri F, Curia G, Marinelli C, Biagini G. Increased perivascular laminin redicts damage to astrocytes in CA3 and piriform cortex following chemoconvulsive treatments. Neurosci. 2012;218:278-94

Guegan C, Sola B. Early and sequential recruitment of apoptotic effectors after focal permanent ischemia in mice. Brain Res. 2000;856: 93–100

Gunther A, Kuppers-Tiedt L, Schneider PM, Kunert I, Berrouschot J, Schneider D, Rossner S. Reduced infarct volume and differential effects on glial cell activation after hyperbaric oxygen treatment in rat permanent focal cerebral ischaemia. Eur J Neurosci. 2005;21(11):3189–3194

Hacke W, Schwab S, Horn M, Spranger M, De Georgia M, von Kummer R. “Malignant” middle cerebral artery territory infarction: clinical course and prognostic signs. Arch Neurol. 1996;53:309 –315

Haensel JX, Spain A, Martin C. A systematic review of physiological methods in rodent pharmacological MRI studies. Psychopharmacology (Berl). 2015;232(3):489-99

Hallenbeck JM. The many faces of tumor necrosis factor in stroke. Nat Med. 2002;8(12):1363– 1368

Hallett M. Plasticity of the human motor cortex and recovery from stroke. Brain Res Brain Res Rev. 2001;36(2-3):169-74

Harsany M, Tsivgoulis G, Alexandrov AV. Intravenous thrombolysis in acute ischemic stroke: standard and potential future applications. Expert Rev Neurother. 2014;14(8):879-92

Harston G, Tee Y, Blockley N, Okell TW, Thandeswaran S, Shaya G, Sheerin F,

133

Cellerini M, Payne S, Jezzard P, Chappell, Kennedy J. Identifying the ischaemic penumbra using pH-weighted magnetic resonance imaging. Brain. 2015;138:36-42

Hast A. Simple filter design for first and second order derivatives by a double filtering approach. Pattern Recognition Letters. 2014;42;65-71

Hayashi T, Sakurai M, Itoyama Y, Abe K. Oxidative damage and breakage of DNA in rat brain after transient MCA occlusion. Brain Res. 1999;832:159 –63

Hayward NM, Yanev P, Haapasalo A, Miettinen R, Hiltunen M, Gröhn O, Jolkkonen J. Chronic hyperperfusion and angiogenesis follow subacute hypo-perfusion in the thalamus of rats with focal cerebral ischemia. J Cereb Blood Flow Metab. 2011;31(4):1119-32

Heiss WD, Graf R, Löttgen J, Ohta K, Fujita T, Wagner R, Grond M, Weinhard K. Repeat positron emission tomographic studies in transient middle cerebral artery occlusion in cats: residual perfusion and efficacy of postischemic reperfusion. J Cereb Blood Flow Metab 1997;17:388-400

Heiss WD. Ischemic Penumbra: Evidence from functional imaging in man. J Cereb Blood F Met. 2000;20:1276-93

Herrmann O, Tarabin V, Suzuki S, Attigah N, Coserea I, Schneider A, Vogel J, Prinz S, Schwab S, Monyer H, Brombacher F, Schwaninger M. Regulation of body temperature and neuroprotection by endogenous interleukin-6 in cerebral ischemia. J Cereb Blood Flow Metab. 2003; 23(4):406–415

Hill MD, Buchan AM. Thrombolysis for acute ischemic stroke: results of the Canadian Alteplase for Stroke Effectiveness Study. Can Med Assoc J. 2005;172:1307-12

Hoge RD. Calibrated fMRI. Neuroimage. 2012;62(2):930-7 Hosp JA, et al. Thin-film epidural microelectrode arrays for somatosensory and motor

cortex mapping in rat. J. Neurosci. Methods. 2008;172:255-62 Hossmann KA. The two pathophysiologies of focal brain ischemia: implications for

translational stroke research. J. Cereb. Blood Flow Metab. 2012;32:1310-16 Howells DW, Macleod MR. Evidence-based translational medicine. Stroke .

2013;44:1466–71 Hu X, Wester P, Brannstrom T, Watson BD, Gu W. Progressive and reproducible focal

cortical ischemia with or without late spontaneous reperfusion generated by a ring-shaped, laser-driven photothrombotic lesion in rats. Brain Res Brain Res Protoc. 2001;7:76- 85

Hu X, Zheng H, Yan T, Pan S, Fang J, Jiang R, Ma S. Physical exercise induces expression of CD31 and facilitates neural function recovery in rats with focal cerebral infarction. Neurol Res. 2010;32:397–402

Hu S, Cao Q, Xu P, Ji W, Wang G, Zhang Y. Rolipram stimulates angiogenesis and attenuates neuronal apoptosis through the cAMP/cAMP-responsive element binding protein pathway following ischemic stroke in rats. Exp Ther Med. 2016;11(3):1005-10

Huang J, Choudhri TF, Winfree CJ, McTaggart RA, Kiss S, Mocco J, Kim LJ, Protopsaltis TS, Zhang Y, Pinsky DJ, Connolly ES Jr. Postischemic cerebrovascular E-selectin expression mediates tissue injury in murine stroke. Stroke. 2000;31(12):3047–3053

Hunter AJ, Hatcher J, Virley D, Nelson P, Irving E, Hadingham SJ, Parsons AA. Functional assessments in mice and rats after focal stroke.

134

Neuropharmacology. 2000;39(5):806-16 Hurn PD, Macrae IM. Estrogen as a neuroprotectant in stroke. I Cereb Blood Flow

Metab. 2000;20:631-52 Hyder F, Rothman DL, Bennett MR. Cortical energy demands of signaling and

nonsignaling components in brain are conserved across mammalian species and activity levels. Proc Natl Acad Sci USA. 2013;110(9):3549-54

Hyman MC, et al. Self-regulation of inflammatory cell trafficking in mice by the leukocyte surface apyrase CD39. J Clin Invest. 2009;119:1136–49

AIadecola C, Zhang F, Xu S, Casey R, Ross ME. Inducible nitric oxide synthase gene expression in brain following cerebral ischemia. J Cereb Blood Flow Metab. 1995;15(3):378–384

BIadecola C, Zhang F, Xu X. Inhibition of inducible nitric oxide synthase ameliorates cerebral ischemic damage. Am J Physiol. 1995;268(1 Pt 2):R286–292

Iadecola C, Sugimoto K, Niwa K, et al. Increased susceptibility to ischemic brain injury in cyclooxygenase-1-deficient mice. J Cereb Blood Flow Metab. 2001;21:1436-41

Iadecola C. Neurovascular regulation in the normal brain and in Alzheimer's disease. Nat Rev Neurosci. 2004;5(5):347-60

AIadecola C, Anrather J. Stroke research at a crossroad: asking the brain for directions. Nat Neurosci. 2011;14(11):1363-8

BIadecola C, Anrather J. The immunology of stroke: from mechanisms to translation. Nat Med. 2011;17:796–808

ImageMagick Studio LLC. http://www.imagemagick.org/script/index.php. Copyright 1999-2013 Non-profit organization dedicated to making software imaging solutions freely available. 2013

Ishikawa M, Zhang JH, Nanda A, Granger DN. Inflammatory responses to ischemia and reperfusion in the cerebral microcirculation. Front Biosci. 2004;9:1339–47

Jander S, Schroeter M, Stoll G. Role of NMDA receptor signaling in the regulation of inflammatory gene expression after focal brain ischemia. J Neuroimmunol. 2007;109:181–7

Jeffers MS, Hoyles A, Morshead C, Corbett D. Epidermal growth factor and erythropoietin infusion accelerate functional recovery in combination with rehabilitation. Stroke. 2014;45(6):1856-8

Jessberger S. Neural repair in the adult brain. F1000Res. 2016;5 Jiang Q. Magnetic resonance imaging and cell-based neurorestorative therapy after

brain injury. Neural Regen Res. 2016;11(1):7-14 Jin K, Zhu Y, Sun Y, Mao XO, Xie L, Greenberg DA. Vascular endothelial growth factor

(VEGF) stimulates neurogenesis in vitro and in vivo. Proc Natl Acad Sci USA. 2002;99(18):11946-50

Jin K, Sun Y, Xie L, Childs J, Mao XO, Greenberg DA. Post-ischemic administration of heparin-binding epidermal growth factor-like growth factor (HB-EGF reduces infarct size and modifies neurogenesis after focal cerebral ischemia in the rat. J Cereb Blood F Met. 2004;24:399-408

Jin R, Yang G, Li G. Molecular insights and therapeutic targets for blood-brain barrier disruption in ischemic stroke: critical role of matrix metalloproteinases and tissue-type plasminogen activator. Neurobiol Dis. 2010;38(3):376–385

AJin R, Yang G, Li G. Inflammatory mechanisms in ischemic stroke: role of inflammatory cells. J Leukoc Biol. 2010;87(5):779-89

135

BJin R, Yang G, Li G. Molecular insights and therapeutic targets for blood-brain barrier disruption in ischemic stroke: critical role of matrix metalloproteinases and tissue-type plasminogen activator. Neurobiol Dis. 2010;38(3):376–385

Jin Q, Cheng J, Liu Y, Wu J, Wang X, Wei S, Zhou X, Qin Z, Jia J, Zhen X. Improvement of functional recovery by chronic metformin treatment is associated with enhanced alternative activation of microglia/macrophages and increased angiogenesis and neurogenesis following experimental stroke. Brain Behav Immun. 2014;40:131-42

Jones TA, Schallert T. Use-dependent growth of pyramidal neurons after neocortical damage. J Neurosci. 1994;14:2140–52

Jones TA, Adkins DL. Motor System Reorganization After Stroke: Stimulating and Training Toward Perfection. Physiology (Bethesda). 2015;30(5):358-70

Jung JE, Kim GS, Chan PH. Neuroprotection by interleukin-6 is mediated by signal transducer and activator of transcription 3 and antioxidative signaling in ischemic stroke. Stroke. 2011;42(12):3574–3579

Kakuta H, Zheng X, Oda H, Harada S, Sugimoto Y, Sasaki K, Tai A. Cyclooxygenase-1-Selective Inhibitors Are Attractive Candidates for Analgesics That Do Not Cause Gastric Damage. Design and in Vitro/in Vivo Evaluation of a Benzamide-Type Cyclooxygenase-1 Selective Inhibitor. J Med Chem. 2008;51:2400-11

Kanemitsu H, Nakagomi T, Tamura A, Tsuchiya T, Kono G, Sano K. Differences in the extent of primary ischemic damage between middle cerebral artery coagulation and intraluminal occlusion models. J Cereb Blood Flow Metab. 2002 ;22:1196 –204

Kanemoto Y, Nakase H, Akita N, Sakaki T. Effects of anti-intercellular adhesion molecule-1antibody on reperfusion injury induced by late reperfusion in the rat middle cerebral artery occlusion model. Neurosurgery. 2002;51(4):1034–1041

Kang JK, Bénar C, Al-Asmi A, et al. Using patient-specific hemodynamic response functions in combined EEG-fMRI studies in epilepsy. Neuroimage. 2003;20:1162-70

Katsman D, Zheng J, Spinelli K, Carmichael ST. Tissue micro-environments within functional cortical subdivisions adjacent to focal stroke. J Cereb Blood Flow Metab. 2003;23:997–1009

Katz JA. COX-2 inhibition: what we learned--a controversial update on safety data. Pain Med 2013;14:29-34

Kawabori M, Yenari MA. The role of the microglia in acute CNS injury.Metab Brain Dis. 2015;30(2):381-92

Kawano T, Anrather J, Zhou P, Park L, Wang G, Frys KA, Kunz A, Cho S Orio M, Iadecola C. Prostaglandin E2 EP1 receptors: downstream effectors of COX-2 neurotoxicity. Nat Med 2006;12(2):225– 229

Khale MP, Bix GJ. Successfully climbing the “STAIRs”: Surmounting failed translation of experimental ischemic stroke treatments. Stroke Research and Treatment. 2012;2013:374098

Khedr EM, Fetoh NA. Short- and long-term effect of rTMS on motor function recovery after ischemic stroke. Restor Neurol Neurosci. 2010;28(4):545-59

Kidwell CS, Alger JR, Saver JL. Beyond mismatch: evolving paradigms in imaging the ischemic penumbra with multi-modal magnetic resonance imaging. Stroke.

136

2003;34:2729- 35 Kigerl KA, Gensel JC, Ankeny DP, Alexander JK, Donnelly DJ, Popovich PG.

Identification of two distinct macrophage subsets with divergent effects causing either neurotoxicity or regeneration in the injured mouse spinal cord. J. Neurosci. 2009;29(43):13435–44

Kim GW, Sugawara T, Chan PH. Involvement of oxidative stress and caspase-3 in cortical infarction after photothrombotic ischemia in mice. J Cereb Blood Flow Metab. 2000;20:1690 –1701

Kim JY, Kawabori M, Yenari MA. Innate inflammatory responses in stroke: mechanisms and potential therapeutic targets. Curr Med Chem. 2014;21(18):2076-97

Kimberley TJ, Lewis SM. Understanding Neuroimaging. Phys Ther. 2007;87:670–683 Kimberley TJ, Khandekar G, Borig M. fMRI reliability in subjects with stroke. Exp Brain

res. 2008;186:183–190 Kirilina E, Lutti A, Poser BA, Blankenburg F, Weiskopf N. The quest for the best: The

impact of different EPI sequences on the sensitivity of random effect fMRI group analyses. NeuroImage. 2015;126:49-59

Kleim JA, Boychuk JA, Adkins DL. Rat models of upper extremity impairment in stroke. ILAR J. 2007;48(4):374-84

Kobayashi T, Tahara Y, Matsumoto M, Iguchi M, Sano H, Murayama T, Arai H, Oida H, Yurugi-Kobayashi T, Yamashita JK, Katagiri H, Majima M, Yokode M, Kita T, Narumiya S. Roles of thromboxane A(2) and prostacyclin in the development of atherosclerosis in apoE-deficient mice. J Clin Invest. 2004;114(6):784-94

Kojima T, Hirota Y, Ema M, Takahashi S, Miyoshi I, Okano H, Sawamoto K. SVZ-derived neural progenitor cells migrate along a blood vessel scaffold toward the post-stroke striatum. Stem Cells. 2010;28(3):545-54

Kono H, Rock KL. How dying cells alert the immune system to danger. Nat Rev Immunol. 2008;8:279–89

Konsman JP, Drukarch B, Van Dam AM. (Peri)vascular production and action of proinflammatory cytokines in brain pathology. Clin Sci (Lond). 2007;112:1–25

Korn C, Augustin HG. Mechanisms of Vessel Pruning and Regression. Dev Cell. 2015;34(1):5-17

Krafft PR, Bailey EL, Lekic T, Rolland WB, Altay O, Tang J, Wardlaw JM, Zhang JH, Sudlow CL. Etiology of stroke and choice of models. Int J Stroke. 2012;7(5):398-406

Krainik A, Hund-Georgiadis M, Zysset S, von Cramon DY. Regional impairment of cerebrovascular reactivity and BOLD signal in adults after stroke. Stroke. 2005;36(6):1146-52

Krishnamurthi RV, Feigin VL, Forouzanfas MH, Mensah GA, Connor M, Bennett DA, Moran AE, Sacco RL, Anderson LM, Truelsen T, et al. Global and regional burden of first-ever ischaemic and haemorrhagic stroke during 1990-2010: findings from the Global Burden of Disease Study 2010. Lancet Glob Health. 2013;1(5):259-81

Krueger H, Koot J, Hall RE, O'Callaghan C, Bayley M, Corbett D. Prevalence of Individuals Experiencing the Effects of Stroke in Canada: Trends and Projections. Stroke. 2015;46(8):2226-31

Krupiński J, Kaluza J, Kumar P, Kumar S, Wang JM. Some remarks on the growth-rate and angiogenesis of microvessels in ischemic stroke. Morphometric and immunocytochemical studies. Patol Pol. 1993;44:203–9

137

Krupiński J, et al. Role of angiogenesis in patients with cerebral ischemic stroke. Stroke. 1994;25:1794-8

Kwakkel G, Veerbeek JM, van Wegen EE, Wolf SL. Constraint-induced movement therapy after stroke. Lancet Neurol. 2015;14(2):224-34

Kwakkel G, Winters C, van Wegen EE, Nijland RH, van Kuijk AA, Visser-Meily A, de Groot J, de Vlugt E, Arendzen JH, Geurts AC, Meskers CG. Effects of Unilateral Upper Limb Training in Two Distinct Prognostic Groups Early After Stroke: The EXPLICIT-Stroke Randomized Clinical Trial. Neurorehabil Neural Repair. 2016.

Lai SM, Studenski S, Duncan PW, Perera S. Persisting consequences of stroke measured by the Stroke Impact Scale. Stroke. 2002;33(7):1840-4

Lai AY, Todd KG. Microglia in cerebral ischemia: molecular actions and interactions. Can J Physiol Pharmacol. 2006;84(1):49–59

Laird NM and JH Ware. Random-effects models for longitudinal data. Biometrics. 1982;38(4):963-74

Lake EM, Chaudhuri J, Thomason L, et al. The effects of delayed reduction of tonic inhibition on ischemic lesion and sensorimotor function. J Cereb Blood Flow Metab. 2015;35:1601-9

Lalancette-Hebert M, Gowing G, Simard A, Weng YC, Kriz J. Selective ablation of proliferating microglial cells exacerbates ischemic injury in the brain. J Neurosci. 2007;27:2596–605

Lambertsen KL, Gregersen R, Finsen B. Microglial-macrophage synthesis of tumor necrosis factor after focal cerebral ischemia in mice is strain dependent. J Cereb Blood Flow Metab. 2002 ;22:785–97

Lambrechts D Zacchigna S, Carmeliet P. Neurovascular signalling defects in neurodegeneration. Nature reviews. 2008;9:169–81

Langdon KD, Granter-Button S, Harley CW, Moody-Corbett F, Peeling J, Corbett D. A cognitive rehabilitation paradigm effective in male rats lacks efficacy in female rats. J Cereb Flow Metab. 2014;34:1673-80

Langhorne P, Bernhardt J, Kwakkel G. Stroke rehabilitation. Lancet. 2011; 377(9778):1693–702

Lapchak PA, Chapman DF, Zivin JA. Metalloproteinase inhibition reduces thrombolytic (tissue plasminogen activator)-induced hemorrhage after thromboembolic stroke. Stroke. 2000;31(12): 3034–3040

Lee RG, van Donkelaar P. Mechanisms underlying functional recovery following stroke. Can J Neurol Sci. 1995;22(4):257-63

Lee VM, Burdett NG, Carpenter A, Hall LD, Pambakian PS, Patel S, Wood NI, James MF. Evolution of photochemically induced focal cerebral ischemia in the rat. Magnetic resonance imaging and histology. Stroke. 1996;27:2110–8

Lee JK, Kim JE, Sivula M, Strittmatter SM. Nogo receptor antagonism promotes stroke recovery by enhancing axonal plasticity. J Neurosci. 2004;24(27):6209–17

Li YF, Cheng YF, Huang Y, Conti M, Wilson SP, O'Donnell JM, Zhang HT. Phosphodiesterase-4D knock-out and RNA interference-mediated knock-down enhance memory and increase hippocampal neurogenesis via increased cAMP signaling. J Neurosci. 2011;31(1):172-83

Li C, Wang J, Fang Y, Liu Y, Chen T, Sun H, Zhou XF, Liao H. Nafamostat Mesilate Improves Function Recovery after Stroke by Inhibiting neuro-inflammation in Rats. Brain Behav Immun. 2016

138

Liang X, Lin L, Woodling NS, et al. Signaling via the prostaglandin E(2) receptor EP4 exerts neuronal and vascular protection in a mouse model of cerebral ischemia. J Clin Invest. 2011;121:4362– 4371

Liedtke AJ, Crews BC, Daniel CM, Blobaum AL, Kingsley PJ, Ghebreselasie K, Marnett LJ. Cyclooxygenase-1-selective inhibitors based on the (E)-2'-des-methyl-sulindac sulfide scaffold. J Med Chem. 2012;55(5):2287-300

Liepert J, Bauder H, Wolfgang HR, Miltner WH, Taub E, Weiller C: Treatment-induced cortical reorganization after stroke in humans. Stroke. 2000;31(6):1210-6

Liesz A, Zhou W, Mracsko E, Karcher S, Bauer H, Schwarting S, Sun L, Bruder D, Stegemann S, Cerwenka A, Sommer C, Dalpke AH, Veltkamp R. Inhibition of lymphocyte trafficking shields the brain against deleterious neuro-inflammation after stroke. Brain. 2011;134(Pt 3):704–720

Lin TN, Sun SW, Cheng WM, Li F, Chang C. Dynamic changes in cerebral blood flow and angiogenesis after transient focal cerebral ischemia in rats: evaluation with serial magnetic resonance imaging. Stroke. 2002;33(12):2985-91

Lin CY, Chang C, Cheung WM, Lin MH, Chen JJ, Hsu CY, Chen JH, Lin TN. Dynamic changes in vascular permeability, cerebral blood volume, vascular density, and size after transient focal cerebral ischemia in rats: evaluation with contrast-enhanced magnetic resonance imaging. J Cereb Blood Flow Metab. 2008;28(8):1491-501

Lindsberg PJ, Strbian D, Karjalainen-Lindsberg ML. Mast cells as early responders in the regulation of acute blood-brain barrier changes after cerebral ischemia and hemorrhage. J Cereb Blood Flow Metab. 2010;30:689–702

Liu XS, Zhang ZG, Zhang RL, Gregg S, Morris DC, Wang Y, Chopp M. Stroke induces gene profile changes associated with neurogenesis and angiogenesis in adult SVZ progenitor cells. J Cereb Blood Flow Metab. 2007;27(3):564-74

Liu M, Dziennis S, Hurn PD, Alkayed NJ. Mechanisms of gender-linked ischemic brain injury. Restor Neurol Neurosci. 2009;27:163-79

Liu H, Li W, Rose ME, Pascoe JL, Miller TM, Ahmad M, Poloyac SM, Hickey RW, Graham SH. Prostaglandin D2 toxicity in primary neurons is mediated through its bioactive cyclopentenone metabolites. Neurotoxicology. 2013;39:35–44

Liu J, Wang Y, Akamatsu Y, Lee CC, Stetler RA, Lawton MT, Yang GY. Vascular remodelling after ischemic stroke: mechanisms and therapeutic potentials. Prog Neurobiol. 2014;115:138-56

Lively S, Moxon-Emre I, Schlichter CL. SC1/Hevin and Reactive Gliosis after Transient Ischemic stroke in young and aged rats. J Neuropathol Exp Neurol. 2011;70:913-29

Logothetis NK. What we can do and what we cannot do with fMRI. Nature. 2008; 453(7197):869-78

Lu YZ, Lin CH, Cheng FC, Hsueh CM. Molecular mechanisms responsible for microglia-derived protection of Sprague-Dawley rat brain cells during in vitro ischemia. Neurosci Lett. 2005; 373(2):159–164

Lum PS, Burgar CG, Shor PC, Majmundar M and van der Loos M. Robot-assisted movement training compared with conventional therapy techniques for the rehabilitation of upper-limb motor function after stroke. Arch Phys Med Rehabil. 2002;83(7):952-9

Lynch CD, Cooney PT, Bennett SA, Thornton PL, Khan AS, Ingram RL, Sonntag WE.

139

Effects of moderate caloric restriction on cortical microvascular density and local cerebral blood flow in aged rats. Neurobiol Aging. 1999;20(2):191-200S

Mantovani A, Biswas SK, Galdiero MR, Sica A, Locati M. Macrophage plasticity and polarization in tissue repair and remodelling. J Pathol. 2013;229(2):176–85

Manwani B, McCullough LD. Function of the master energy regulator adenosine monophosphate-activated protein kinase in stroke. J. Neurosci. Res. 2013;91(8):1018–29

Mao L, Jia J, Zhou X, Xiao Y, Wang Y, Mao X, Zhen X, Guan Y, Alkayed NJ, Cheng J. Delayed administration of a PTEN inhibitor BPV improves functional recovery after experimental stroke. Neuroscience. 2013;231:272–81

Marlier Q, Verteneuil S, Vandenbosch R, Malgrange B. Mechanisms and Functional Significance of Stroke-Induced Neurogenesis. Front Neurosci. 2015;9:458

Martin A, et al. Imaging of perfusion, angiogenesis, and tissue elasticity after stroke. J. Cereb. Blood Flow Metab. 2012;32:1496-1507

Martirosian P, Boss A, Schraml C, Schwenzer NF, Graf H, Claussen CD, Schick F. Magnetic resonance perfusion imaging without contrast media. Eur J Nucl Med Mol Imaging. 2010;37(1):52-64

Matsumoto H, Kumon Y, Watanabe H, Ohnishi T, Takahashi H, Imai Y, Tanaka J. Expression of CD200 by macrophage-like cells in ischemic core of rat brain after transient middle cerebral artery occlusion. Neurosci Lett. 2007;418:44–8

Metz GA, Whishaw IQ. Cortical and subcortical lesions impair skilled walking in the ladder rung walking test: a new task to evaluate fore- and hindlimb stepping, placing, and co-ordination. J Neurosci Methods. 2002;115:169–179

Marsh BJ, Williams-Karnesky RL, Stenzel-Poore MP. Toll-like receptor signaling in endogenous neuroprotection and stroke. Neuroscience. 2009;158:1007–20

Marshall RS, Perera GM, Lazar RM, Krakauer JW, Constantine RC, DeLaPaz RL. Evolution of cortical activation during recovery from corticospinal tract infarction. Stroke. 2000;31(3):656-61

Mayzel-Oreg O, Omae T, Kazemi M, Li F, Fisher M, Cohen Y, et al. Microsphere-induced embolic stroke: an MRI study. Magn Reson Med. 2004;51:1232–8

Mazighi M, Serfaty JM, Labreuche J, Laissy JP, Meseguer E, Lavallee PC, et al. Comparison of intravenous alteplase with a combined intravenous-endovascular approach in patients with stroke and confirmed arterial occlusion (RECANALISE study): a prospective cohort study. Lancet Neurol. 2009;8:802-9

McAdam BF, Catella-Lawson F, Mardini IA, Kapoor S, Lawson JA, FitzGerald GA. Systemic biosynthesis of prostacyclin by cyclooxygenase (COX)-2: the human pharmacology of a selective inhibitor of COX-2. Proc Natl Acad Sci U S A. 1999;96(1):272-7

McKee AC, Carreras I, Hossain L, Ryu H, Klein W. L, Oddo S, LaFerla F. M, Jenkins B. G, Kowall N. W, Dedeoglu A. Ibuprofen reduces Aβ, hyperphosphorylated tau and memory deficits in Alzheimer mice. Brain Res. 2008;1207:225–236

Meziane H, Ouagazzal AM, Aubert L, Wietrzych M, Krezel W. Estrous cycle effects on behavior of C57BL/6J and BALB/cB-gamma-J female mice: implications for phenotyping strategies. Genes, Brain and Behavior. 2007;6:192-200

Mikulik R, Wahlgren N. Treatment of acute stroke: an update. Journal of Internal Medicine. 2015;145:165

Milidonis X, Marshall I, Macleod MR, Sena ES. Magnetic resonance imaging in

140

experimental stroke and comparison with histology: systematic review and meta-analysis. Stroke. 2015;46(3):843-51

Miltner WH, Bauder H, Sommer M, Dettmers C and Taub E. Effects of constraint-induced movement therapy on patients with chronic motor deficits after stroke: a replication. Stroke. 1999;30(3):586-92

Miyake M, Takeo S, Kaijihara H. Sustained decrease in regional blood flow after microsphere injection in rats. Stroke. 1993 ;24:415– 20

Mocco J, Choudhri T, Huang J, Harfeldt E, Efros L, Klingbeil C, Vexler V, Hall W, Zhang Y, Mack W, Popilskis S, Pinsky DJ, Connolly ES Jr. HuEP5C7 as a humanized monoclonal anti-E/P- selectin neurovascular protective strategy in a blinded placebo-controlled trial of nonhuman primate stroke. Circ Res. 2002;91(10):907–914

Mogensen TH. Pathogen Recognition and Inflammatory Signaling in Innate Immune Defenses. Clin Microbiol Rev. 2009;22(2):240-73

Mohr JP, Gautier JC, Hier D, Stein RW. Middle cerebral artery disease. Stroke: Pathophysiology, Diagnosis and Management. 1986;1:377-450

Monte-Silva K, Kuo MF, Hessenthaler S, Fresnoza S, Liebetanz D, Paulus W, Nitsche MA. Induction of late LTP-like plasticity in the human motor cortex by repeated non-invasive brain stimulation. Brain Stimul. 2013;6:424-32

Montoya CP, Campbell-Hope LJ, Pemberton KD, Dunnett SB. The “staircase test”: a measure of independent forelimb reaching and grasping abilities in rats. J Neurosci Methods. 1991;36(2-3):219-28

Mohr JP, Grotta JC, Wolf PA, Moskowitz MA, Mayberg MR, Von Kummer R. Stroke: pathophysiology, diagnosis, and management. Elsevier Health Sciences. 2011

Moretti A, Ferrari F, Villa RF. Neuroprotection for ischaemic stroke: Current status and challenges . Pharmacol Ther. 2015;146:23-34

Morris DC, Yeich T, Khalighi MM, Soltanian-Zadeh H, Zhang ZG, Chopp M. Microvascular structure after embolic focal cerebral ischemia in the rat. Brain Res. 2003;972(1-2):31-7

Murphy TH, Corbett D. Plasticity during stroke recovery: from synapse to behaviour. Nat Rev Neurosci. 2009;10(12):861-72

Mulcahy NJ, Ross J, Rothwell NJ, Loddick SA. Delayed administration of interleukin-1 receptor antagonist protects against transient cerebral ischaemia in the rat. Br J Pharmacol. 2003;140(3): 471–476

Nagayama T, Lan J, Henshall DC, Chen D, O’Horo C, Simon RP, Chen J Induction of oxidative DNA damage in the peri-infarct region after permanent focal cerebral ischemia. J Neurochem. 2000;75:1716 –28

Nathan C, Ding A. Nonresolving inflammation. Cell. 2010;140:871–82 Nedergaard M, Ransom B, Goldman SA. New roles for astrocytes: redefining the

functional architecture of the brain. Trends Neurosci. 2003;26(10):523-30 Neelin & Fonov et al. Copyright © 1997-2016 McConnell Brain Imaging Centre.

MINCTools. http://www.bic.mni.mcgill.ca/ServicesSoftware/MINC Neumann-Haefelin T, Kastrup A, de Crespigny A, Yenari MA, Ringer T, Sun GH et al.

Serial MRI after transient focal cerebral ischemia in rats: dynamics of tissue injury, blood-brain barrier damage, and edema formation. Stroke. 2000;31:1965-72

Newton J, Sunderland A, Butterworth SE, Peters AM, Peck KK, Gowland PA. A pilot

141

study of event- related functional magnetic resonance imaging of monitored wrist movements in patients with partial recovery. Stroke. 2002;33(12):2881-7

Nguemeni C, Gomez-Smith M, Jeffers MS, Schuch CP, Corbett D. Time course of neuronal death following endothelin-1 induced focal ischemia in rats. J. Neurosci. Methods. 2015;242:72-6

Nishibe M, Urban ET 3rd, Barbay S, Nudo RJ. Rehabilitative training promotes rapid motor recovery but delayed motor map reorganization in a rat cortical ischemic infarct model. Neurorehabil Neural Repair. 2015;29(5):472-82

Nishijima T, Piriz J, Duflot S, Fernandez AM, Gaitan G, Gomez-Pinedo U, Verdugo JM, Leroy F, Soya H, Nuñez A, Torres-Aleman I. Neuronal activity drives localized blood-brain-barrier transport of serum insulin-like growth factor-I into the CNS. Neuron. 2010;67(5):834-46

Niwa K, Haensel C, Ross ME, Iadecola C. Cyclooxygenase-1 participates in selected vasodilator responses of the cerebral circulation. Circ Res. 2001;88(6):600-8

Nomura DK, Morrison BE, Blankman JL, Long JZ, Kinsey SG, Marcondes MCG, Ward AM, Hahn YK, Lichtman AH, Conti B, Cravatt BF. Endocannabinoid Hydrolysis Generates Brain Prostaglandins That Promote neuro-inflammation. Science (Washington, DC, U.S.) 2011;334:809–813

Nowicka D, Rogozinska K, Aleksy M, Witte OW, Skangiel-Kramska J. spatio-temporal dynamics of astroglial and microglial responses after photothrombotic stroke in the rat brain. Acta Neurobiol Exp (Wars). 2008;68(2):155–168

Nussmeier NA, Whelton AA, Brown MT, Langford RM, Hoeft A, Parlow JL, Boyce SW, Verburg KM. Complications of the COX-2 inhibitors parecoxib and valdecoxib after cardiac surgery. N Engl J Med. 2005;352:1081–1091

O’Collins VE, Macleod MR, Donnan GA, Horky LL, van der Worp BH, Howells DW. 1,026 experimental treatments in acute stroke. Ann Neurol. 2006;59:467–77

Ohab JJ, Fleming S, Blesch A, Carmichael ST. A neurovascular niche for neurogenesis after stroke. J Neurosci. 2006;26(50):13007-16

Oliveira LMO, Dimitrov D. Surgical Techniques for Chronic Implantation of Microwire Arrays in Rodents and Primates. In: Nicolelis MAL, editor. Methods for Neural Ensemble Recordings. 2nd edition. Boca Raton (FL): CRC Press/Taylor & Francis; 2008. Chapter 2.

Olsen TS, Lassen NA. A dynamic concept of middle cerebral artery occlusion and cerebral infarction in the acute state based on interpreting severe hyperemia as a sign of embolic migration. Stroke. 1984;15:458-68

Ott E, Nussmeier NA, Duke PC, Feneck RO, Alston RP, Snabes MC, Hubbard RC, Hsu PH, Saidman LJ, Mangano DT. Efficacy and safety of the cyclooxygenase 2 inhibitors parecoxib and valdecoxib in patients undergoing coronary artery bypass surgery. J Thorac Cardiovasc Surg. 2003;125:1481–1492

Pan W, Kastin AJ. Tumor necrosis factor and stroke: role of the blood-brain barrier. Prog Neurobiol. 2007;83(6):363–374

Pang L, Ye W, Che XM, Roessler BJ, Betz AL, Yang GY. Reduction of inflammatory response in the mouse brain with adenoviral-mediated transforming growth factor-ss1 expression. Stroke. 2001;32(2):544–552

Parkes LM, Tofts PS. Improved accuracy of human cerebral blood perfusion measurements using arterial spin labeling: accounting for capillary water permeability. Magn Reson Med. 2002;48(1):27-41

142

Paxinos G, Watson C. The Rat Brain in Stereotaxic Coordinates. Fifth Edition. 2005 Peachell P. Regulation of mast cells by β-agonists. Clin Rev Allergy Immunol.

2006;31:131–42 Peerschke EI, Yin W, Ghebrehiwet B. Complement activation on platelets: implications

for vascular inflammation and thrombosis. Mol Immunol. 2010;47:2170–5 Peng SL, Ravi H, Sheng M, Thomas BP, Lu H. Searching for a truly “iso-metabolic” gas

challenge in physiological fMRI. J Cereb Blood Flow Metabol. 2016 Perrone MG, Scilimati A, Simone L, Vitale P. Selective COX-1 Inhibition: A Therapeutic

Target to be Reconsidered. Curr Med Chem. 2010;17(32):3769-805 Pineiro R, Pendlebury S, Johansen-Berg H, Matthews PM. Altered hemodynamic

responses in patients after subcortical stroke measured by functional MRI. Stroke. 2002;33:103-9

Pinsky DJ, et al. Hypoxia-induced exocytosis of endothelial cell Weibel-Palade bodies. A mechanism for rapid neutrophil recruitment after cardiac preservation. J Clin Invest. 1996;97:493–500

Pinteaux E, Rothwell NJ, Boutin H. Neuroprotective actions of endogenous interleukin-1 receptor antagonist (IL-1ra) are mediated by glia. Glia. 2006;53(5):551–556

Pocock JM, Kettenmann H. Neurotransmitter receptors on microglia. Trends Neurosci. 2007; 30:527–35

Pomeroy V, Aglioti SM, Mark VW, McFarland D, Stinear C, Wolf SL et al. Neurological principles and rehabilitation of action disorders: rehabilitation interventions. Neurorehab Neural Re. 2011;5:35-43

Pradillo JM, Denes A, Greenhalgh AD, Boutin H, Drake C, McColl BW, Barton E, Proctor SD, Russell JC, Rothwell NJ, Allan SM. Delayed administration of interleukin-1 receptor antagonist reduces ischemic brain damage and inflammation in comorbid rats. J Cereb Blood Flow Metab. 2012;32(9):1810–1819

Provenzale JM, Jahan R, Naidich TP, Fox AJ. Assessment of the patient with hyperacute stroke: imaging and therapy. Radiology . 2003;229:347–59

Que M, Schiene K, Witte OW, Zilles K. Widespread up-regulation of N-methyl-D-aspartate receptors after focal photothrombotic lesion in rat brain. Neurosci Lett. 1999;273:77– 80

Rainsford KD. Anti-inflammatory drugs in the 21st century. Subcell Biochem. 2007;42:3–27

Reid JL, Dawson D, Macrae IM. Endothelin, cerebral ischaemia and infarction. Clin Exp Hypertens. 1995;(1-2):399-407

Relton JK, Sloan KE, Frew EM, Whalley ET, Adams SP, Lobb RR. Inhibition of alpha4 integrin protects against transient focal cerebral ischemia in normotensive and hypertensive rats. Stroke. 2001;32(1):199–205

Riedel CH, Zimmermann P, Jensen-Kondering U, Stingele R, Deuschl G, Jansen O. The importance of size: successful recanalization by intravenous thrombolysis in acute anterior stroke depends on thrombus length. Stroke. 2011;42:1775–7

Roc AC, Wang J, Ances BM, Liebeskind DS, Kasner SE, Detre JA. Altered hemodynamics and regional cerebral blood flow in patients with hemodynamically significant stenoses. Stroke. 2006;37(2):382-7

Rossini PM, et al. Interhemispheric difference of sensory hand area after monohemispheric stroke: MEG/MRI integrative study. NeuroImage.

143

2001;14:474-85 Rossini C, Altamura A, Ferretti F, Vernieri F, Zappasodi M, Caulo, Pizzella V. Does

cerebrovascular disease affect the coupling between neuronal activity and local haemodynamics? Brain. 2004;127(1):99-110

Ruan L, Wang B, ZhuGe Q, Jin K. Coupling of neurogenesis and angiogenesis after ischemic stroke. Brain Res. 2015;1623:166-73

Sahin B, et al. Brain volumes of the lamb, rat and bird do not show hemispheric asymmetry: a stereological study. Image Anal. Sterol. 2001;20:9-13

Sairanen TR, Lindsberg PJ, Brenner M, Siren AL. Global forebrain ischemia results in differential cellular expression of interleukin-1beta (IL-1beta) and its receptor at mRNA and protein level. J Cereb Blood Flow Metab. 1997;17:1107–20

Saleem S, Ahmad AS, Maruyama T, Narumiya S, Dore S. PGF(2alpha) FP receptor contributes to brain damage following transient focal brain ischemia. Neurotox Res. 2009;15(1):62–70

Salter K, Campbell N, Richardson M, Mehta S, Jutai J, Zettler L, Moses M, McClure A, Mays R, Foley N, Teasell R. Evidence-Based Review of Stroke Rehabilitation Chapter 21: Outcome measures in stroke rehabilitation 2013. Heart and stroke foundation Canadian partnership for stroke recovery. Evidence Based Review of Stroke Rehabilitation, London Ontario Canada. 2013

Samson MT, et al. Differential roles of CB1 and CB2 cannabinoid receptors in mast cells. J Immunol. 2003;170:4953–62

Sanchez-Moreno C, Dashe JF, Scott T, Thaler D, Folstein MF, Martin A. Decreased levels of plasma vitamin C and increased concentrations of inflammatory and oxidative stress markers after stroke. Stroke J Cereb Circ. 2004;35(1):163–168

Savic B, Meier B. How Transcranial Direct Current Stimulation Can Modulate Implicit Motor Sequence Learning and Consolidation: A Brief Review. Front Hum Neurosci. 2016;10:26

Sawada M, Matsumoto M, Sawamoto K. Vascular regulation of adult neurogenesis under physiological and pathological conditions. Front Neurosci. 2014;8:53

Schiene K, Bruehl C, Zilles K, Qu M, Hagemann G, Kraemer M et al. Neuronal hyper-excitability and reduction of GABAA-receptor expression in the surround of cerebral photo-thrombosis. J Cereb Blood F Met. 1996;16:906-14

Schroeter M, Jander S, Huitinga I, Witte OW, Stoll G. Phagocytic response in photochemically induced infarction of rat cerebral cortex. The role of resident microglia. Stroke. 1997;28:382–6

Schwab S, Steiner T, Aschoff A, Schwarz S, Steiner HH, Jansen O, et al. Early hemicraniectomy in patients with complete middle cerebral artery infarction. Stroke. 1998;29:1888–93

Schwab JM, Nguyen TD, Postler E, Meyermann R, Schluesener HJ. Selectiveaccumulation of cyclooxygenase-1-expressing microglial cells/macrophages in lesions of human focal cerebral ischemia. Acta Neuropathol. 2000;99(6):609-14

Seil FJ. Recovery and repair issues after stroke from the scientific perspective. Curr Opin Neurol. 1997;10(1):49-51

Selvin E, Hirsch A. Contemporary risk factor control and walking dysfunction in individuals with peripheral arterial disease: NHANES 1999– 2004. Atherosclerosis. 2008;201(2)425–33

144

Shen Q, Goderie SK, Jin L, Karanth N, Sun Y, Abramova N, Vincent P, Pumiglia K, Temple S. Endothelial cells stimulate self-renewal and expand neurogenesis of neural stem cells. Science. 2004;304(5675):1338-40

Shen Q, Du F, Huang S, Duong TQ. spatio-temporal characteristics of postischemic hyperperfusion with respect to changes in T1, T2, diffusion, angiography, and blood-brain barrier permeability. J Cereb Blood Flow Metab. 2011;31(10):2076-85

Shen Q, Huang S, Duong TQ. Ultra-high spatial resolution basal and evoked cerebral blood flow MRI of the rat brain. Brain Res. 2015;1599:126-36

Shih YY, Huang S, Chen YY, Lai HY, Kao YC, Du F, Hui ES, Duong TQ. Imaging neurovascular function and functional recovery after stroke in the rat striatum using forepaw stimulation. J Cereb Blood Flow Metab. 2014(9):1483-92

Shu CY, Herman P , Coman D, Sanganahalli BG , Wang H, Juchem W, Rothman DL, de Graaf RA, Hyder F. Brain region and activity-dependent properties of M for calibrated fMRI. Neuroimage. 2016;125:846:856

Shyu WC, Lin SZ, Yang HI, Tzeng YS, Pang CY,Yen PS et al. Functional recovery of stroke rat induced by granulocyte colony-neurogenesis and angiogenesis and improves neurological function in rats. Circulation. 2004;110:1847-54

Simon AB, Buxton RB. Understanding the dynamic relationship between cerebral blood flow and the BOLD signal: implications for quantitative functional fMRI. Neuroimage 2015;116:158:161

Smith WS, Sung G, Saver J, Budzik R, Duckwiler G, Liebeskind DS, Lutsep HL, Rymer MM, Higashida RT, Starkman S, Gobin YP. Multi MERCI Investigators; Frei D, Grobelny T, Hellinger F, Huddle D, Kidwell C, Koroshetz W, Marks M, Nesbit G, Silverman IE: Mechanical thrombectomy for acute ischemic stroke: final results of the Multi MERCI trial. Stroke. 2008;39:1205–12

Song WL, Wang M, Ricciotti E, Fries S, Yu Y, Grosser T, Reilly M, Lawson JA, FitzGerald GA. Tetranor PGDM, an abundant urinary metabolite reflects biosynthesis of prostaglandin D2 in mice and humans. J Biol Chem. 2008; 283(2):1179-88

Speliotes EK, et al. Increased expression of basic fibroblast growth factor (bFGF) following focal cerebral infarction in the rat. Brain Res. Mol. Brain Res. 1995;29:31-42

Spite M, Serhan CN. Novel lipid mediators promote resolution of acute inflammation: impact of aspirin and statins. Circ Res. 2010;107:1170–84

Stagg CJ, Best JG, Stephenson MC, O'Shea J, Wylezinska M, Kincses ZT, Morris PG, Matthews PM, Johansen-Berg H. Polarity-sensitive modulation of cortical neurotransmitters by transcranial stimulation. J Neurosci. 2009; 29(16):5202-6

Stanimirovic D, Satoh K. Inflammatory mediators of cerebral endothelium: a role in ischemic brain inflammation. Brain Pathol. 2000;10(1):113-26

Statistics Canada CANSIM Table 102-0529: deaths, by cause: diseases of the circulatory system. 2012, 2000–2006

Stegmayr B, Asplund K, Kuulasmaa K, Rajakangas AM, Thorvaldsen P, Tuomilehto J. Stroke incidence and mortality correlated to stroke risk factors in the WHO MONICA Project. An ecological study of 18 populations. Stroke. 1997;28:1367-74

Steinberg BA, Augustine JR. behavioural, anatomical, and physiological aspects of

145

recovery of motor function following stroke. Brain Res Brain Res Rev. 1997;25(1):125-32

Sternberg SR. Biomedical image processing. Computer. 1983;16:2-34 Strbian D, Karjalainen-Lindsberg ML, Tatlisumak T, Lindsberg PJ. Cerebral mast cells

regulate early ischemic brain swelling and neutrophil accumulation. J Cereb Blood Flow Metab. 2006;26:605–12

Strle K, Zhou JH, Shen WH, Broussard SR, Johnson RW, Freund GG, Dantzer R, Kelley KW. Interleukin-10 in the brain. Crit Rev Immunol. 2001;21(5):427–449

Stroemer RP, Kent TA, Hulsebosch CE. Increase in synaptophysin immunoreactivity following cortical infarction. Neurosci Lett. 1992;147(1):21-4

Stroemer RP, Kent TA, Hulsebosch CE. Acute increase in expression of growth associated protein GAP-43 following cortical ischemia in rat. Neurosci Lett. 1993;162(1-2):51-4

Stroemer RP, Kent TA, Hulsebosch CE. Neocortical neural sprouting, synaptogenesis and behavioural recovery following neocortical infarction in rats. Stroke. 1998;29(11):2381-93

Sudlow CL, Warlow CP. Comparable studies of the incidence of stroke and its pathological types: results from an international collaboration. International Stroke Inschidence Collaboration. Stroke. 1997;28:491-9

Sun Y, Jin K, Xie L, Childs J, Mao XO, Logvinova A et al. VEGF-induced neuro-protection, neurogenesis and angiogenesis after focal cerebral ischemia. J Clin Invest. 2003;111:1843-51

Sutherland BA, Minnerup J, Balami JS, Arba F, Buchan AM, Kleinschnitz C. Neuroprotection for ischaemic stroke: translation from the bench to the bedside. Int J Stroke. 2012;7:407–18

Swanson LW. Mapping the human brain: past, present, and future. Trends Neurosci. 1995;18:471-4

Szpak GM, et al. Border zone neovascularization in cerebral ischemic infarct. Folia Neuropathol. 1999;37:264–8

Taguchi A, Soma T, Tanaka H, Kanda T, Nishimura H, Yoshikawa H et al. Administration of CD34+ cells after stroke enhances neurogenesis via angiogenesis in a mouse model. J Clin Invest. 2004;114:330–8

Tang K, Xia FC, Wagner PD, Breen EC. Exercise induced VEGF transcriptional activation in brain, lung and skeletal muscle. Respir Physiol Neurobiol. 2010;170:16–22

Taub E, Miller NE, Novack TA, Cook EW 3rd, Fleming WC, Nepomuceno CS, Connell JS, Crago JE. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehabil. 1993;74(4):347-54

Taub E, Uswatte G, King DK, Morris D, Crago JE, Chatterjee A. A placebo-controlled trial of constraint-induced movement therapy for upper extremity after stroke. Stroke. 2006;37(4):1045-9

Teasell R, Rice D, Richardson M, Campbell N, Madady M, Hussein N, Murie-Fernandez M, Page S. The next revolution in stroke care. Expert Rev Neurother. 2014:14(11):1307-14

Tecchio F, et al. Brain plasticity in recovery from stroke: An MEG assessment. NeuroImage. 2006;32:1326-34

Thored P, Wood J, Arvidsson A, Cammenga J, Kokaia Z, Lindvall O. Long-term

146

neuroblast migration along blood vessels in an area with transient angiogenesis and increased vascularization after stroke. Stroke. 2007;38(11):3032-9

Tipton AJ, Sullivan JP. Sex differences in T cells in hypertension. Clin Ther. 2014;36:1882-900

Turner RC, Lucke-Wold B, Lucke-Wold N, Elliott AS, Logsdon AF, Rosen CL, et al. Neuroprotection for ischemic stroke: moving past shortcomings and identifying promising directions. Int J Mol Sci. 2013;14:1890–917

Ueki M, Linn F, Hossmann KA. Functional activation of cerebral blood flow and metabolism before and after global ischemia of rat brain. J Crebe Blood Flow and metab. 1988;8:486:494

van Bruggen N, Cullen BM, King MD, Doran M, Williams SR, Gadian DG, et al. T2- and diffusion-weighted magnetic resonance imaging of a focal ischemic lesion in rat brain. Stroke. 1992;23:576–82

van der Zijden JP, van der Toorn A, van der Marel K, Dijhuizen RM. Longitudinal in vivo MRI of alterations in peri-lesional tissue after transient ischemic stroke in rats. Exp Neurol. 2008;212:207-12

van Lookeren CM, et al. Secondary reduction in the apparent diffusion coefficient of water, increase in cerebral blood volume, and delayed neuronal death after middle cerebral artery occlusion and early repercussion in the rat. J. Cereb. Blood Flow Metab. 1999;19:1354-64

Veerbeek JM, van Wegen E, van Peppen R, van der Wees PJ, Hendriks E, Rietberg M, Kwakkel G. What is the evidence for physical therapy poststroke? A systematic review and meta-analysis. PLoS One. 2014;9(2)

Vemuganti R, Dempsey RJ, Bowen KK. Inhibition of intercellular adhesion molecule-1 protein expression by antisense oligonucleotides is neuroprotective after transient middle cerebral artery occlusion in rat. Stroke. 2004;35(1):179–184

Virley D, Beech JS, Smart SC, Williams SC, Hodges H, Hunter AJ. A temporal MRI assessment of neuropathology after transient middle cerebral artery occlusion in the rat: correlations with behaviour. J Cereb Blood Flow Metab. 2000;20:563-82

Wahlgren NG, Ahmed N. Neuroprotection in cerebral ischemia: facts and fancies – the need for new approaches. Cerebrovasc Dis. 2004;17(1):153-66

Wang L, Yushmanov VE, Liachenko SM, Tang P, Hamilton RL, Xu Y. Late reversal of cerebral and water diffusion after transient focal ischemia in rats. J. Cereb. Blood Flow Metab. 2002;22:253-61

Wang Y, Cooke MJ, Morshead CM, Shoichet MS. Hydrogel delivery of erythropoietin to the brain for endogenous stem cell stimulation after stroke injury. Biomaterials. 2012(33):2681-62

Wang Y, Li M, Dong F, Zhang J, Zhang F. Physical exercise-induced protection on ischemic cardiovascular and cerebrovascular diseases. Int J Clin Exp Med. 2015:15;8(11):19859-66

Ward NS, Brown MM, Thompson AJ, Frackowiak RS: Neural correlates of outcome after stroke: a cross-sectional fMRI study. Brain. 2003;126(6):1430-48

Watson BD, Dietrich WD, Busto R, Wachtel MS, Ginsberg MD. Induction of reproducible brain infarction by photochemically initiated thrombosis. Ann Neurol. 1985;17:497–504

Weber R, Ramos-Cabrer P, Wiedermann D, van Camp N, Hoehn M. A fully non-

147

invasive and robust experimental protocol for longitudinal fMRI studies in the rat. Neuroimage. 2006;29(4):1303-10

Weber R, Ramos-Cabrer P, Justicia C, Wiedermann D, Strecker C, Sprenger C, Hoehn M. Early prediction of functional recovery after experimental stroke: functional magnetic resonance imaging, electrophysiology, and behavioural testing in rats. J Neurosci. 2008;28(5):1022-9

Wegener S, Weber R, Ramos-Cabrer P, Uhlenkueken U, Sprenger C, Wiedermann D et al. Temporal profile of T2-weighted MRI distinguishes between pannecrosis and selective neuronal death after transient focal cerebral ischemia in the rat. J Cereb Blood F Met. 2006;26:38-47

Wegener S, Artmann J, Luft AR, Buxton RB, Weller M, Wong EC. The time of maximum post-ischemic hyperperfusion indicates infarct growth following transient experimental ischemia. PLoS One. 2013;8(5)

Wesson DW, Wilson DA. Sniffing out the contributions of the olfactory tubercle to the sense of smell: hedonics, sensory integration, and more? Neurosci. Biobehav. Rev. 2011;35:655-68

Weston RM, Jones NM, Jarrott B, Callaway JK. Inflammatory cell infiltration after endothelin-1- induced cerebral ischemia: histochemical and myeloperoxidase correlation with temporal changes in brain injury. J Cereb Blood Flow Metab. 2007;27(1):100–114

Whishaw IQ, Pellis SM, Gorny B, Kolb B, Tetzlaff W. Proximal and distal impairments in rat forelimb use in reaching follow unilateral pyramidal tract lesions. Behav Brain Res. 1993;56:59–76

White FC, Bloor CM. Coronary vascular remodelling and coronary resistance during chronic ischemia. Am J Cardiovasc Pathol. 1992; 4:193-202

WHO. The world health report 2000: Health systems - improving performance. Geneva: WHO, 2000

Wieloch T, Nikolich K. Mechanisms of neural plasticity following brain injury. Curr Opin Neurobiol. 2006;16:258-64

Wiltrout C, Lang B, Yan Y, Dempsey RJ, Vemuganti R. Repairing brain after stroke: a review on post-ischemic neurogenesis. Neurochem Int. 2007;50(7-8):1028-41

Windle V, Szymanska A, Granter-Button S, White C, Buist R, Peeling J et al. An analysis of four different methods of producing focal cerebral ischemia with endothelin-1 in the rat. Exp Neurol. 2006;201:324-34

Witte OW, Bidmon HJ, Schiene K, Redecker C, Hagemann G. Functional differentiation of multiple peri-lesional zones after focal cerebral ischemia. J Cereb Blood F Met. 2000;20:1149-65

Womelsdorf T, Valiante TA, Sahin NT, Miller KJ, Tiesinga P. Dynamic circuit motifs underlying rhythmic gain control, gating and integration. Nat Neurosci. 2014;17(8):1031-9

Wong EC. An introduction to ASL labeling techniques. J Magn Reson Imaging. 2014;40(1):1-10

Woodruff TM, Thundyil J, Tang SC, Sobey CG, Taylor SM, Arumugam TV. Pathophysiology, treatment, and animal and cellular models of human ischemic stroke. Molecular degeneration. 2011;6-11

Xiong X, Barreto GE, Xu L, Ouyang YB, Xie X, Giffard RG. Increased brain injury and worsened neurological outcome in interleukin-4 knockout mice after transient

148

focal cerebral ischemia. Stroke. 2011;42(7):2026–2032 Xiong Y, Zhu WZ, Zhang Q, Wang W. Observation of post-MCAO cortical inflammatory

edema in rats by 7.0Tesla MRI. J Huazhong Univ Sci Technol Med Sci. 2014;34(1):120-4

Xu SY, Pan SY. The failure of animal models of neuroprotection in acute ischemic stroke to translate to clinical efficacy. Med Sci Monit Basic Res. 2013;19:37-45

Yagami T, Koma H, Yamamoto Y. Pathophysiological Roles of Cyclooxygenases and PG in the Central Nervous System. Mol Neurobiol. 2015

Yamashita K, Busch E, Wiessner C, Hossmann KA. Thread occlusion but not electrocoagulation of the middle cerebral artery causes hypothalamic damage with subsequent hyperthermia. Neurol Med Chir. 1997;37:723–7

Yamashita T, Sawamoto K, Suzuki S, Suzuki N, Adachi K, Kawase T, Mihara M, Ohsugi Y, Abe K, Okano H. Blockade of interleukin-6 signaling aggravates ischemic cerebral damage in mice: possible involvement of Stat3 activation in the protection of neurons. J Neurochem. 2005;94(2): 459–468

Yang Y, Yang T, Li Q, Wang CX, Shuaib A. A new reproducible focal cerebral ischemia model by introduction of polyvinylsiloxane into the middle cerebral artery: a comparison study. Neurosci Methods. 2002;118(2):199-206

Yang Y, Estrada EY, Thompson JF, Liu W, Rosenberg GA. Matrix metalloproteinase-mediated disruption of tight junction proteins in cerebral vessels is reversed by synthetic matrix metalloproteinase inhibitor in focal ischemia in rat. J Cereb Blood Flow Metab. 2007;27(4):697– 709

Yemisci M, et al. Pericyte contraction induced by oxidative-nitrative stress impairs capillary reflow despite successful opening of an occluded cerebral artery. Nat Med. 2009;15:1031–7

Yenari M, Kitagawa K, Lyden P, Perez-Pinzon M. Metabolic down-regulation: a key to successful neuroprotection? Stroke. 2008;39:2910–7

Yepes M, Sandkvist M, Wong MK, Coleman TA, Smith E, Cohan SL, Lawrence DA. Neuroserpin reduces cerebral infarct volume and protects neurons from ischemia-induced apoptosis. Blood. 2000;96(2):569–576

Yilmaz G, Granger DN. Leukocyte recruitment and ischemic brain injury. Neuromolecular Med. 2010;12:193–204

Yrjanheikki J, Keinanen R, Pellikka M, Hokfelt T, Koistinaho J. Tetracyclines inhibit microglial activation and are neuroprotective in global brain ischemia. Proc Natl Acad Sci U S A. 1998;95(26):15769–15774

Yrjanheikki J, Tikka T, Keinanen R, Goldsteins G, Chan PH, Koistinaho J. A tetracycline derivative, minocycline, reduces inflammation and protects against focal cerebral ischemia with a wide therapeutic window. Proc Natl Acad Sci U S A. 1999;96(23):13496–13500

Yu SW, Friedman B, Cheng Q, Lyden PD. Stroke-evoked angiogenesis results in a transient population of microvessels. Cereb Blood Flow Metab. 2007;27(4):755-63

Yu CW, Liang X, Lipsky S, Karaaslan C, Kozakewich H, Hotamisligil GS, Bischoff J, Cataltepe S. Dual role of fatty acid-binding protein 5 on endothelial cell fate: a potential link between lipid metabolism and angiogenic responses. Angiogenesis. 2016;19(1):95-106

Zhang Z, Zhang RL, Jiang Q, Raman SB, Cantwell L, Chopp M. A new rat model of

149

thrombotic focal cerebral ischemia. J Cereb Blood Flow Metab. 1997;17:123–35

Zhang Z, Chopp M. Vascular endothelial growth factor and angiopoietins in focal cerebral ischemia. Trends Cardiovasc Med. 2002;12(2):62-6

Zhang L, Schallert T, Zhang ZG, Jiang Q, Arniego P, Li Q, Lu M, Chopp M. A test for detecting long-term sensorimotor dysfunction in the mouse after focal cerebral ischemia. J. Neurosci. Methods. 2002;117(2):207–14

Zhang N, Komine-Kobayashi M, Tanaka R, Liu M, Mizuno Y, Urabe T. Edaravone reduces early accumulation of oxidative products and sequential inflammatory responses after transient focal ischemia in mice brain. Stroke. 2005;36(10):2220–2225

Zhang P, et al. Early exercise improves cerebral blood flow through increased angiogenesis in experimental stroke rat model. J Neuroeng. Rehabil. 2013;10;1-10

Zhang ZG, Chopp M. Promoting brain remodelling to aid in stroke recovery. Trends Mol Med. 2015;21(9):543-8

Zhao X, Haensel C, Araki E, Ross ME, Iadecola C. Gene-dosing effect and persistence of reduction in ischemic brain injury in mice lacking inducible nitric oxide synthase. Brain Res. 2000; 872(1–2):215–218

Zhe-Sun P, Zhou J, Sun W, J Huang, van Zijl P. Detection of the ischemic penumbra using pH weighted fMRI. J Cereb Blood Flow Metabol. 2007;27:1129-1136

Zhe-Sun P, Benner T, Copen WA, Sorensen G. Early experience of translating pH-weighted MRI to image human subject at 3 Tesla. Stroke, 2010:147-155

Zhou G, Myers R, Li Y, Chen Y, Shen X, Fenyk-Melody J, Wu M, Ventre J, Doebber T, Fujii N, Musi N, Hirshman MF, Goodyear LJ, Moller DE. Role of AMP-activated protein kinase in mechanism of metformin action. J Clin Invest. 2001:108(8):1167–74

150

Appendix 1 – Continuous ASL

An MR image can be sensitized to the effect of in-flowing blood spins if those

spins are in a different magnetic state (“labelled”) than the static tissue [Martirosian et

al. 2010]. When magnetically labelled spins diffuse into the extracellular space and

exchange with the unperturbed tissue spins, the result is a net reduction of tissue

magnetization [Parkes & Tofts 2002]. Repeating the experiment in a “control” condition

(in which arterial spins are not labelled [Martirosian et al. 2010]) and comparing the

signal between conditions results in a direct measurement of perfusion [Parkes & Tofts

2002].

In the continuous ASL experiment (CASL, employed in the present work) the

“label” is created by inverting the magnetization of arterial protons using a long radio

frequency (RF) pulse in combination with a slice-selective gradient for a flow-driven

adiabatic inversion of arterial magnetization [Parkes & Tofts 2002]. During adiabatic

labeling, the magnetization of spins crossing the labeling plane are inverted. The net

result being the creation of a continuous stream of inverted or 'labelled' protons which

flow towards the imaging plane placed downstream of the labelling plane [Martirosian et

al. 2010]. The ASL labeling pulse consists of a lengthy (~1-2 second long) constant

gradient ('Gz') (applied parallel to the direction of flow) and an RF pulse ('B1') applied

perpendicular to the gradient (along the transverse plane). In the present work the

direction of flow is parallel to the main magnetic field where labelling occurs at the level

of carotid arteries [Wong 2014]. From the perspective of the moving spin, the effective

field ('Beff' – from B1 and Gz) rotates at the frequency of the applied RF pulse: B1 is fixed

(in the transverse plane), while the longitudinal component is proportional to the position

of the spin [Wong 2014]. As the spin travels across the labelling plane Beff (beginning at

a magnitude of '+Z') reduces to zero then grows to reach '-Z' causing an inversion of the

spin if rotation of Beff is slow compared to the precession frequency (this is the adiabatic

condition) [1] – reproduced from Wong 2014. In Equation [1], 'Ɣ' denotes the

gyromagnetic ratio (the constant nucleus-specific MR frequency at a given field

strength).

151

[1]

A cartoon of the labelling (difference in signal achieved between a labeled image

and a control image) and labelling/imaging planes overlaid on an image of a rat head

are shown in the figure below. The plot below shows the signal difference with the

passage of the label during the CASL experiment. In the first part, (1), there is no signal

change as the label has yet to reach the imaging plane. In time (denoted tA), the label

begins to reach the imaging plane resulting in a non-zero difference in signal. The signal

difference increases as the label continues to arrive at the imaging plane during the

second part (2) of the experiment. Relaxation and outflow after labelling stops (at a time

denoted: tA + tL), result in a reduction of the signal difference during the final part of the

experiment (3). Adapted from Parkes & Tofts (2002).

For ASL experiments, the in-flow/out-flow arterial/venous magnetization (denoted

by 'ma' and 'mv' in equation [2]) are included in the Bloch equation (which expresses net

nuclear magnetization (M) as a function of relaxation times) [Parkes & Tofts 2002]. The

152

simplest description of ASL theory uses a single compartment model where within an

imaging voxel spins instantaneously cross from intra-vascular to extra vascular space:

ie there's only one compartment effectively. There are two key assumptions in this

model: (1) water enters and exits through exchanging arteries/veins, (2) the

compartment is well mixed. Therefore, within a voxel inside the imaging plane, spins

cross between the intra- and extra-vascular compartments at a perfusion rate (f: blood

perfusion, ml blood/min/100ml tissue) [Parkes & Tofts 2002]. Spins relax at the tissue

relaxation rate 'T1' (or the longitudinal relaxation time of water in tissue). Equation [2]

(reproduced from Parkes & Tofts 2002), M(t) denotes tissue magnetization, M0 denotes

the longitudinal equilibrium magnetization of tissue.

[2]

In a CASL experiment, signal (difference in magnetization between the control

and labelled images) depends on the inversion efficiency (labelling efficiency, denoted

as 'α'), the longitudinal relaxation times of blood (T1b) and tissue (T1app), tA – the time

taken for the label to travel between the labelling and imaging planes (arrival time) and

the labelling time (tL) – duration of the labelling pulse [Parkes & Tofts 2002]. The

solution to the Bloch equation [2] during the three parts of the CASL experiment (refer to

figure above) are: [3] before the arrival of the label (part one of the experiment: t < tA),

[4] during the in-flow of the label (part two of the experiment: tA < t < tA + tL), and [5]

during the label out-flow (part three of the experiment: t > tA + tL). Solution is reproduced

from Parkes & Tofts (2002).

[3]

[4]

[5]

Down-stream of the labelling plane, images are acquired rapidly (in the present

153

work) using echo-planar imaging (EPI). We employed CASL imaging to measure

relative perfusion without quantifying absolute perfusion. For these measurements we

acquired EPI images following labelling and control (when no label was introduced)

preparations as described above in an alternating control/label sequence (repeated a

minimum of 30 times). To estimate relative perfusion the data series was fit voxel-wise

as described in Methods Chapter 3.

154

In press for publication in Philosophical Transaction B

Appendix 2

Functional MRI in Chronic Ischemic Stroke (Review)

Abstract

Ischemic stroke is the leading cause of adult disability worldwide. Effective rehabilitation

is hindered by uncertainty surrounding the underlying mechanisms that govern long-

term ischemic injury progression. Despite its potential as a sensitive non-invasive in vivo

marker of brain function that may aid in development of new treatments, BOLD fMRI

has found limited application in the clinical research on chronic stage stroke

progression. Stroke affects each of the physiological parameters underlying the BOLD

contrast, markedly complicating the interpretation of BOLD fMRI data. This review

summarizes current progress on application of BOLD fMRI in chronic stage of ischemic

injury progression and discusses means by which more information may be gained from

such BOLD fMRI measurements. Concomitant measurements of vascular reactivity,

neuronal activity, and metabolism in preclinical models of stroke are reviewed along

with illustrative examples of post-ischemic evolution in neuronal, glial, and vascular

function. The realization of the BOLD fMRI potential to propel stroke research is

predicated on the carefully designed preclinical research establishing an ischemia-

specific quantitative model of BOLD signal contrast so as to provide the framework for

interpretation of fMRI findings in clinical populations.

155

2.1 Introduction

Although widely used in neuroscience for the study of healthy human brain

function, blood oxygenation level dependent (BOLD) functional magnetic resonance

imaging (fMRI) has found limited application in clinical research on brain disorders and

diseases. Indeed, the perceived clinical potential of BOLD fMRI remains unrealized. In

the present review, we focus on the application of fMRI to studies of brain function in the

chronic stage of stroke progression. The large societal and personal burden of stroke,

the mechanistic basis of stroke in vascular injury, and the strong dependence of BOLD

signal on blood flow, make fMRI application to chronic stage of stroke a particularly

significant research direction. Indeed, BOLD neuroimaging has been proposed a robust

in vivo marker of ischemia-induced brain dysfunction/stroke recovery that could aid in

identification of the determinants of prognosis, and guide the development of new

ischemic stroke treatments [Hallet et al. 2001, Seil et al. 1997, Steinberg et al. 1997,

Lee et al. 1995].

The positive BOLD response is robustly observed in various brain regions under

a variety of input tasks/stimuli across many physiological conditions and in many

species. It reflects a focal decrease in deoxyhemoglobin (dHb) content, which due to the

paramagnetism of deoxyhemoglobin, its sequestration to red blood cells, and the

geometry of the focal vascular architecture, results in the decrease of local microscopic

magnetic field heterogeneities and thus lengthening of the apparent tissue spin-spin

relaxation time during periods of increased neuronal activity. The BOLD fMRI signal is

hence a function of deoxyhemoglobin (dHb) concentration, which is in turn driven by

adjustments in cerebral blood flow (CBF), cerebral blood volume (CBV), and cerebral

metabolic rate of O2 consumption (CMRO2); as well as the distribution of dHb in the

tissue, as determined by the microvascular architecture [Buxton et al. 1998, Davis et al.

1998]. Stroke affects each of the physiological parameters underlying BOLD contrast:

ischemia triggers peri-lesional angiogenesis (preceded acutely by vessel loss) thus

changing microvascular architecture and affecting CBF and CBV; and modulates

CMRO2 due to neuronal loss [De Girolami et al. 1984], glial recruitment and activation

[BJin et al. 2010], generation of new vessels [Hayward et al. 2011, Lin et al. 2008,

156

Dijhuizen et al. 2003, Lin et al. 2002, Carmichael et al. 2001] axonal sprouting

[Carmicharl et al. 2001], and synaptogenesis [Stroemer et al. 1993, Stroemer et al.

1994 Stroemer et al. 1998]. Consequently, BOLD may be a sensitive and widely

accessible marker of ischemic sequelae. On the other hand, because all of the

parameters dictating the BOLD signal are differentially affected by ischemic injury

progression, the interpretation of BOLD signal change following ischemic injury is

complex [Logothetis et al. 2008] and BOLD neuroimaging in isolation has been

challenged to impact clinical research on stroke recovery.

This review summarizes progress on clinical and preclinical application of BOLD

fMRI in chronic stage of ischemic injury progression and discusses means by which

more information may be gained from such BOLD fMRI measurements. The focus on

the chronic stage is motivated by an increasingly recognized need for interventions in

the weeks to months following ischemic injury as the majority of ischemic stroke

patients (as much as 97% in some centres) arrive at a care facility beyond the

therapeutic window for safe application of currently available treatments [Harsany et al.

2014]. This research direction is further supported by the incidence and temporal

evolution of spontaneous recovery: as many as 25% of ischemic stroke patients show

improvement with physical therapy and continue to recover function more than 12

months following stroke [Lee et al. 1995, Taub et al. 1993, Miltner et al. 1999, Liepert et

al. 2000, Lum et al. 2002, Fasoli et al. 2003, Fasoli et al. 2004, Taub et al. 2006, Lai et

al. 2002]. The therapeutic window for novel treatments, aimed at enhancement of

endogenous healing, may thus be much longer than is often assumed [Hallet et al.

2001, Seil et al. 1997, Steinberg et al. 1997, Lai et al. 2002, Duncan et al. 2000,

Krishnamurthi et al. 2013]. BOLD fMRI may thus also be valuable for elucidation of the

mechanisms which govern recovery in the long-term and for identification of new,

chronic stage treatment targets so as to improve recovery in a much wider population of

patients.

2.2 Clinical fMRI research on ischemic stroke recovery

In clinical practice, stroke recovery is evaluated using one or more structured

neuropsychological tests which incorporate simple physical tasks and quality of life

157

assessments [Salter et al. 2013]. Although neuropsychological tests are carefully

designed to evaluate whether a patient is improving, they are confounded by self-

reporting bias and high incidence of very gradual gains (recovery over a long period

often results in a flawed perception of progress). Further, they offer little insight into the

underlying processes driving or impeding recovery [Donovan et al. 2008]. To address

this gap, several studies have tested whether fMRI can provide a more objective metric

of behavioural impairment and/or functional recovery while still correlating with

performance on neuropsychological assessments.

2.3 Behavioural correlates of BOLD fMRI response

On the whole, evoked BOLD fMRI response measurements months following

ischemic stroke have shown pronounced alterations in amplitude and kinetics of the

BOLD signal both in the ipsi- and in the contralesional cortex months after the ischemic

insult. In majority of patients studied to date, however, BOLD response was altered

whether or not the patient had made good recovery, suggesting BOLD fMRI may not be

a sensitive marker of functional recovery. For example, in patients with residual

paralysis 9-57 months after stroke, Blicher et al. (2012) used unilateral isolated wrist

extension–flexion and observed positive BOLD response in 55%, negative BOLD

response in 9%, and no BOLD response in 36% of patients [Blicher et al. 2012]. In

patients with poor outcome, imaged >270 days after stroke, Newton et al. employed

near-isometric wrist extension movements and found the peak BOLD response in

ipsilesional M1 to be delayed relative to contralesional M1 responses during movement

of either wrist [Newton et al. 2002]. The patients exhibited abnormally delayed M1

responses in the ipsilesional hemisphere when moving the paretic wrist. In mildly

impaired patients performing a finger- or hand-tapping task (4-660 days after stroke),

Pineriro et al. (2002) observed the rate of BOLD signal increase and maximum BOLD

signal change in the sensorimotor cortex of both hemispheres to be decreased by ~30%

in patients relative to controls [Pineiro et al. 2002]. Rossini et al. (2004) and Altamura et

al. (2007) studied individuals who experienced agood recovery: 12 months following

injury, direct median nerve stimulation elicited no BOLD activation in either hemisphere

in 25% of these patients; while in 45% of the patients, activation was present only in the

158

contralesional hemisphere [Rossini et al. 2004, Altamura et al. 2007]. In another study

of patients with good recovery 1-2 years after stroke, a visually guided bilateral handball

squeeze task produced a biphasic BOLD response, with patients showing earlier onset

and extended duration of the initial, negative BOLD response followed by positive BOLD

response with a delayed time-to-peak (Roc et al. (2006)) but normal peak magnitude

[Roc et al. 2006].

2.4 Mechanistic basis of BOLD fMRI response in stroke recovery

Notwithstanding the lack of correlation between BOLD fMRI alterations and

behavioural performance, researchers have also attempted to use BOLD fMRI

measurements to gain insight into the neurobiological changes in the chronic stage of

ischemic injury progression. Delayed, low amplitude, absent, or negative BOLD

responses have been purported to indicate dysfunction. BOLD response attenuation

was thus proposed to result from compromised local perfusion due to structural damage

to the vasculature (caused by either ischemia itself or pre-existing comorbidities)

[Pineiro et al. 2002], exhausted vasomotor reactivity due to chronically maximally dilated

vasculature [Rossini et al. 2004, Altamura et al. 2007], or absent neuronal activity

[Blicher et al. 2012, Roc et al. 2006, Krainik et al. 2005]. To test these hypotheses,

some fMRI studies probed neuronal activity while others assessed cerebrovascular

reactivity [Rossini et al. 2004, Altamura et al. 2007, Binkofski et al. 2004].

Employing tactile exploration of objects to elicit BOLD signal changes and motor

evoked potentials (assessed with TMS), Binkofski et al. (2004) conducted a longitudinal

multi-modal neuroimaging study in patients imaged at one, 2-4 weeks and one month

following injury [Binkofski et al. 2004]. One week after stroke, BOLD signal change and

TMS revealed the representation of the paretic hand fingers to be remapped to the

adjacent cortical areas. At 2-4 weeks, there was a transient lack of BOLD responses

despite clinical improvement of hand function and preserved motor evoked potentials on

TMS. At one month after ischemia, BOLD responses were again detected; motor

evoked potentials were still present, and further clinical improvement was made.

Similarly, Rossini et al. (2004) and Altamura et al. (2007) observed absent or reduced

BOLD responses in both hemispheres, yet measured stereotypical somatosensory

159

evoked fields from ipsi- and contralesional hemispheres via MEG [Rossini et al. 2004,

Altamura et al. 2007]. Combined, these data provide evidence that at least in the well-

recovering patients BOLD fMRI attenuation was not due to an absence of neuronal

activity.

With respect to impairment in cerebrovascular reactivity, some studies did not

while other did find it correlated with BOLD attenuation. Rossini et al. (2004) measured

cerebrovascular reactivity to hypercapnia (inhalation of 7% CO2) with transcranial

doppler colour-coded duplex sonography [Rossini et al. 2004]. In 25% of the patients,

missing BOLD fMRI activation was accompanied by impaired vascular reactivity to

hypercapnia. In the remaining 75% of patients, who exhibited some BOLD response, no

correlation was found between the amplitude of the BOLD response and the degree of

hypercapnic reactivity of the major arterial vessels; though impaired microvascular

reactivity was thus not excluded. Moreover, Blicher et al. (2012) observed no BOLD

signal change, on average, in M1 in response to unilateral isolated wrist extension–

flexion in patients with residual paralysis, yet ASL and VASO-FLAIR (vascular-space-

occupancy fluid attenuated inversion recovery) showed this task elicited CBF and CBV

increases that were indistinguishable from those observed in control subjects [Blicher et

al. 2012].

On the other hand, Krainik et al. (2005) used fMRI to estimate cerebrovascular

reactivity by comparing BOLD signal between normal breathing and hyperventilation-

elicited hypocapnia (which increases oxygen extraction, induces vaso-constriction and

reduces CBF) in fully recovered patients with frontal lobe stroke 11-85 days after injury

[Krainik et al. 2005]. Hyperventilation resulted in a smaller BOLD signal decrease in the

ipsilesional primary sensorimotor cortex and supplementary motor areas relative to

contralesional hemispheres and relative to control subjects (bilaterally).

Hyperventilation-elicited BOLD signal decrease in the contralesional supplementary

motor areas was larger than that in the control subjects. In a separate experiment in the

same study, BOLD response during a repeated hand grip task and was decreased in

the ipsilesional primary sensorimotor cortex and ipsilesional supplementary motor areas

despite a full clinical recovery and no evidence of lesions in the primary sensorimotor

cortex, supplementary motor areas, cerebellum, basal ganglia, or thalamus on 3D T₁-

weighted MRI [Krainik et al. 2005]. The authors suggested that impaired

160

cerebrovascular reactivity was a predictor of impaired motor-related BOLD response in

the sensorimotor cortex during contralateral movements [Krainik et al. 2005]. The

discrepancy between this study and those by Rossini et al. (2004) and Blicher et al.

(2012) may be attributed to methodological differences.

Preceding reports of persistent neuronal activity, vascular reactivity impairment,

and alterations in BOLD response amplitude and temporal profile have prompted further

mechanistic hypotheses of modified neurovascular coupling due to abnormal vascular

anatomy and/or changes in tissue metabolic needs. Ultrastructural changes in cerebral

vessels caused by atherosclerosis, peri-lesional gliosis, and/or disruption of aminergic

and cholinergic fibers that innervate the vasculature in the peri-lesional tissue could

modulate blood flow and volume and alter vascular reserve capacity and/or

neurovascular coupling [Blicher et al. 2012, Krainik et al. 2005]. Further, BOLD

attenuation may be secondary to increases in basal CBF of the peri-infarct zone due to

persistent collateral perfusion [Krainik et al 2005].

2.5 BOLD fMRI response remapping

In addition to BOLD response attenuation, BOLD fMRI studies of ischemic injury

progression have also shown evidence of altered spatial pattern of BOLD fMRI

responses. In particular, focal signal enhancement within the primary motor cortex,

increased activity in non-primary sensorimotor areas, and/or increased activity in the

contralateral hemisphere have all been reported [Newton et al. 2002, Cramer et al.

1997, Marshall et al. 2000, Ward et al. 2003, Feydy et al. 2002, Cao et al. 1998].

However, whether such remapping of the BOLD response provides salient insight and if

so, whether it signifies recovery or is a manifestation of dysfunction remains

controversial [Newton et al. 2002, Cao et al. 1998, Buma et al. 2010].

Some studies have provided evidence of BOLD response remapping, but such

remapping was not found to correlate with outcome. For instance, employing the finger

opposition task in patients 5-43 months after injury, with varied NIHSS and Rankin

scores, Cao et al. (1998) examined event related BOLD responses [Cao et al. 1998].

Seventy five percent of patients showed extended activation of the contralesional

sensorimotor cortex; 50% showed bilateral activation of the primary sensorimotor

161

cortex; 50% showed contralesional activation alone during paretic hand movement;

25% showed activation in contralesional premotor and dorsolateral prefrontal cortex,

and 25% of patients exhibited activation in the contralesional supramarginal gyrus and

premotor cortex during paretic hand movement. Despite such extensive BOLD

response reorganization in the patients, no correlation between degree of remapping

and neuropsychological test scores was observed.

On the other hand, some authors have postulated that fMRI response

remapping, namely activation in the contralesional motor cortex, increased sensorimotor

activation volume, and activation within the peri-lesional tissue indicate remapping of

neuronal activity patterns and/or recruitment of additional neurons and are

manifestations of recovery. In particular, Cramer et al. (1997) employed a finger- or

hand-tapping task to measure BOLD responses in recovered stroke patients between

11 days and 15 months following ischemia [Cramer et al. 1997]. Seventy percent of the

patients were found to have a greater activation volume in the contralesional

sensorimotor cortex than that observed in control subjects. Most patients with greater

activation volumes in the contralesional sensorimotor cortex also showed increased

activation in the contralesional cerebellum, premotor cortex and both the ipsilesional

and contralesional supplementary motor areas [Cramer et al. 1997]. In addition, degree

of peri-infarct activation correlated with clinical improvement. Similarly, Marshall et al.

(2000) used the finger-thumb opposition task to measure BOLD responses in patients

within the first few days of stroke onset and again 3-6 months thereafter [Marshall et al.

2000]. In the first few days of injury and in the chronic phase after stroke, patients

showed greater activation in the ipsilesional sensorimotor cortex, ipsilesional posterior

parietal, and bilateral prefrontal regions than did the control subjects. Moreover,

ipsilesional (relative to contralesional) activity in the sensorimotor cortex increased

progressively as the paretic hand regained function [Marshall et al. 2000]. Combined,

these studies suggest that a dynamic bihemispheric reorganization of motor networks

likely contributes to recovery following focal ischemia [Cramer et al. 1997, Marshall et

al. 2000].

On the other hand, some authors suggested that remapping of the BOLD

response correlates with a poor prognosis. For instance, Ward et al. (2003) employed a

dynamic visually paced handgrip task at least 3 months after stroke in patients with

162

documented deficits lasting a minimum of 48 hours following stroke, and with a varied

degree of recovery at the time of imaging [Ward et al. 2002]. Patients with poor recovery

were more likely to recruit a number of motor-related brain regions over and above

those activated in the control group, whereas patients showing recovery were more

likely to show an fMRI activation pattern similar to control subjects [Ward et al. 2002].

Across the entire patient group, Ward et al. (2003) was able to demonstrate a negative

correlation between outcome and the degree of task-related activation in the

supplementary motor areas, cingulate motor areas, premotor cortex, posterior parietal

cortex, and cerebellum [Ward et al. 2002]. In another example, Feydy et al. (2002) used

repetitive first clench (N=10), or a modified movement for those unable to fist clench

(N=4), to measure evoked BOLD responses at three imaging sessions over a period of

1-6 months following ischemic stroke [Feydy et al. 2002]. At earlier time points, patients

showed contralateral and ipsilateral recruitment of sensorimotor cortex, supplementary

motor areas, frontal premotor cortex, frontal premotor areas, and, less commonly,

posterior parietal areas. By latter imaging time points, this recruitment subsided in 66%

of patients; the remaining 34% of the patients showed a higher incidence of M1 lesions

[Feydy et al. 2002]. Authors suggested that ipsilateral recruitment after stroke

corresponds to a compensatory corticocortical process which may be due to a decrease

in inhibition caused by an M1 lesion and damage to the underlying reciprocal

corticocortical connections from M1 to the secondary motor areas, leading to more

extensive ipsilateral and contralateral activation in worse-off patients.

In summary, clinical research has provided ample evidence of BOLD modulation

post ischemic injury, but findings on the correlation between BOLD fMRI response

features (magnitude, kinetics, or spatial pattern) and clinical outcome have been

inconsistent [Buma et al. 2010]. Mechanistically, there has been some, though still

varied, support for BOLD modulation resulting from alterations in impairment in

cerebrovascular reactivity and alterations in neurovascular coupling, though details of

these mechanisms remain uncertain. Concomitant measurements of metabolism,

hemodynamics and BOLD in ischemic stroke patients during the weeks to months

following injury need to be more widely employed [Astrakas et al. 2012] to improve our

understanding of the natural progression of the disease in patients and facilitate

interpretation of BOLD response signals in studies of novel therapeutic agents. At the

163

same time, a detailed mechanistic understanding of the determinants of BOLD contrast

post ischemic injury will necessitate measurements of changes in morphology and

function of specific cellular populations and hence calls for studies in preclinical models

of ischemic stroke.

2.6 Preclinical fMRI research on ischemic stroke recovery

Preclinical studies offer well-controlled experimental conditions including tightly

regulated ischemic injury parameters as well as careful control of comorbidities, and

thus provide a powerful means for examining the neurobiological determinants of stroke

recovery and establishing a quantitative model of BOLD signal contrast post ischemic

injury. However, translation potential of preclinical findings critically depends on careful

experimental design and thoughtful consideration of salient interspecies differences. Of

note, there are structural, functional and hence metabolic differences between humans

and animals in addition to differences in preclinical vs. clinical BOLD fMRI

implementation, and rats in particular. The higher nominal spatial resolution used in

preclinical fMRI (e.g. 0.32 mm3 as in Weber et al. (2008)) vs. clinical fMRI (e.g. 27 mm3)

research is to be evaluated in the light of three orders of magnitude smaller rodent vs.

human total brain volume [Weber et al. 2008, Kirilina et al. 2015, Swanson 1995].

Preclinical BOLD fMRI voxel thus represents 100 times larger relative brain volume than

that comprised in a clinical BOLD fMRI voxel. Additionally, cortical neuronal density in

rats is approximately twice that of humans, whereas the ratio of glia to neurons is lower

in rats (~1:1) than in humans [De Felipe et al. 2002, Nedegaard et al. 2003].

Interestingly, fundamental cortical communication processes (e.g., action potentials,

pre- and postsynaptic potentials, glial neurotransmitter and K+ clearance, etc.) were

estimated to have the same relative energy costs in rats and in humans [Hyder et al.

2013]. On a final methodological note, the sensitivity of BOLD fMRI to magnetic

susceptibility gradients limits the invasiveness of manipulations that can be performed

on the animals and hence the spectrum of complementary techniques that can be done

in combination with fMRI.

Since combining fMRI with other methodologies represents a major technical

challenge, there are only a few multi-modal preclinical fMRI reports which monitor

164

ischemia induced neurovascular remodelling during the chronic period (more than 7

days after the injury): further their conclusions with respect to neurovascular coupling

post ischemia have not been consistent. In particular, Shih et al. (2014) measured intra-

cortical LFPs and CBF and CBV fMRI signal changes with forepaw stimulation up to 4

weeks following a 20-min middle carotid artery (MCA) occlusion [Shih et al. 2014].

Noxious forepaw stimulation-elicited negative fMRI responses were associated with LFP

recovery 28 days following ischemic insult. Negative fMRI responses were interpreted

as evidence of uncoupling between neuronal activity and local vascular reactivity. On

the other hand, Weber et al. (2008) performed the adhesive tape removal test, BOLD

fMRI and EEG (through skull) up to 7 weeks following MCA occlusion [Weber et al.

2008]. At 2-7 days after ischemia, rats were impaired and showed attenuated BOLD

responses to forepaw stimulation; within 1-4 weeks, they exhibited recovery on the

adhesive tape removal test and showed re-establishment of normal BOLD responses,

suggesting that BOLD fMRI could be used as an indicator of functional loss and

restoration. Moreover, the authors observed a tight coupling of electrical brain activity

and hemodynamic responses in early and late phases after stroke suggesting

preservation of neurovascular coupling in the somatosensory cortex in this model.

Complexity of the peri-lesional neurovascular changes induced by focal ischemia is

further exemplified by two multi-modal preclinical fMRI examinations from our own work.

Using intra-cortical micro-injection of endothelin-1 into the right sensorimotor cortex of

adult Sprague Dawley rats, we induced ischemic lesions of 50 ± 10 mm³ in volume on

T₂-weighted MRI. Three weeks following stroke, BOLD and continuous ASL were

acquired to measure BOLD and CBF signal changes elicited by hypercapnia (10%

CO₂), while skilled reaching ability was evaluated using the Montoya Staircase skilled

reaching task [Montoya et al. 1991]. In addition, we recorded intra-cortical spontaneous

local field potential (LFP) responses simultaneously from both hemispheres using a pair

of multi-electrode arrays composed of 16 (2x8) Pl/Ir electrodes (placed within the cortex

such that the anterior pair of electrodes were within the core of the lesion) or performed

immunohistochemical analysis post mortem. These data allowed the investigation of

BOLD signal changes, cerebrovascular reactivity, motor recovery, and neuronal

functioning, as well as post mortem examination of changes in select cellular

populations.

165

In the first example, we considered the subacute period (7 days following

ischemic injury). Representative fMRI and immunohistochemical results are shown in

Figure 2.1. At this time point, hypercapnia elicited an attenuated BOLD response in peri-

lesional M1/M2 ROI relative to the contralesional M1/M2 ROI. The peri-lesional BOLD

response was attenuated by ~60%: 1.09 ± 0.07% in the peri-lesional cortex vs. 2.5 ±

0.1% in the contralesional cortex. However, perfusion increase to hypercapnia in the

peri-lesional M1/M2 cortex was triple that in the contralesional cortex (39 ± 5% peri-

lesionally vs. 12 ± 2% contralesionally). peri-lesional resting perfusion was twice the

contralesional perfusion: 1.66 ± 0.09% vs. 0.85 ± 0.04%. peri-lesional BOLD fMRI

response attenuation thus accompanied vascular hyper-reactivity.

Immunohistochemical analysis of these brains demonstrated peri-lesional neuro-

inflammation, with dense microglial invasion (Figure 2.1.b.i.) and pronounced astrocytic

reactivity (Figure 2.1.b.ii.). Proximal to the site of injury, RECA-1 staining (Figure

2.1.b.iii.) showed an increase in vascular endothelial density indicating endogenous

angiogenesis, as previously reported [Hayward et al. 2011, Lin et al. 2008, Dijhuizen et

al. 2003, Lin et al. 2002]. In parallel experiments, Montoya staircase testing

demonstrated ~70-80% impairment in skilled reaching ability of the left forepaw. Intra-

cortical LFP recordings collected 7 days after injury showed a 10-30% lower ipsi- to

contra-lesional power ratio (across all frequency bands), suggesting silencing of the

neuronal network in the peri-infarct zone.

BOLD response attenuation and motor impairment were thus accompanied by a

hyper-reactive and proliferating vasculature on one hand; and neuronal hypo-activity on

the other hand. We hypothesized that the pronounced increase in peri-lesional vascular

reactivity and local baseline perfusion were triggered by increased metabolic needs (O₂

consumption and glutamate/glutamine cycling) of the densely packed

neuroinflammatory cells. Direct measurements of tissue metabolism are needed to

examine this hypothesis.

166

Figure 2.1

Figure 2.1 Functional MRI and immunohistochemistry in a rat model of focal

sensorimotor ischemia. A week following right sensorimotor cortex injection of

endothelin-1, BOLD fMRI responses (a.i.) and blood flow changes (a.ii.) elicited by

hypercapnia (10% CO₂), as well as resting perfusion (a.iii.) are shown overlaid on the

corresponding structural T₂-weighted images. peri-lesional M1/M2 BOLD response

(1.09 ± 0.07%) was attenuated by ~60% relative to its level in the contralesional cortex

167

(2.5 ± 0.1%), with an ipsi- to contra-lesional ratio of cortical ΔBOLD of 0.44 ± 0.02. peri-

lesionally, blood flow increase in M1/M2 ROI in response to hypercapnia, of 39 ± 5%,

was strongly potentiated (ratio: 3.25 ± 0.04) relative to that in the contralateral cortical

ROI (12 ± 2%). Similarly, peri-lesional resting perfusion (1.66 ± 0.09%) was doubled

(ratio: 1.95 ± 0.01) relative to that contralesional cortex. post mortem

immunohistochemistry was performed in the same animals immediately following

imaging, with Iba-1 staining for microglia (b.i.), GFAP for astrocytic activation (b.ii.), and

RECA-1 for blood vessel endothelium (b.iii.). Microglial recruitment (b.i.) as well as a

pronounced astrocytic activation (b.ii.) were evident in the peri-infarct zone.

Endogenous angiogenesis was also observed in the peri-lesional tissue (b.iii.). Scale

bar in (a.) 2mm, scale bar in (b.) 500μm.

168

In the second example, we considered the neurovascular state three weeks

following the endothelin-1 elicited ischemic insult. Representative fMRI and

electrophysiological results are shown in Figures 2.2 and 2.3, respectively. In these

experiments we investigated the effects of selective cyclooxygenase-1 (COX-1)

inhibition through the chronic administration of FR122047 (Cedarlane Laboratories LTD,

Burlington ON, Canada) from 7 to 19 days following stroke, followed by fMRI or

electrophysiological recordings 20 or 21 days after stroke. In clinical and preclinical

ischemic stroke studies, COX-1 metabolites have been shown to contribute to injury

during the chronic phase of stroke [Perrone et al. 2010, Schwab et al. 2000]. In

preclinical studies, COX-1 inhibition has been shown to prevent neuronal loss [Choi et

al. 2009], reduce oxidative stress (by disrupting free radical production) [Candelario-Jalil

et al. 2003, Candelario-Jalil et al. 2007], attenuate neuro-inflammation [Perrone et al.

2010], and increase resting vascular tone [Niwa et al. 2001]. Through present sustained

inhibition of COX-1 in the chronic stage of stroke, we expected to ameliorate

neurovascular function through reduction in microglial invasion. We therefore measured

BOLD and flow responses to hypercapnia (Figure 2.2) in vehicle-administered and

FR122047-treated animals and subsequently recorded neuronal activity via intracranial

electrophysiology (Figure 2.3).

Twenty-one days after stroke, the BOLD response to hypercapnia in the vehicle-

administered animal was not lateralized (2.6 ± 0.1% vs. 2.8 ± 0.1%). However, peri-

lesional blood flow increase to hypercapnia was double that of the contralesional cortex

(ratio of 2.3 ± 0.1). In the FR122047 treated animal, BOLD response to hypercapnia

was attenuated by 40% peri-lesionally (3.7 ± 0.3% in the peri- vs. 6.1 ± 0.3% in the

contralesional cortex). Nonetheless, its peri-lesional perfusion response to hypercapnia

was elevated by 20% over the contralesional level (ratio of 1.2 ± 0.1). Although peri-

lesional vascular reactivity was elevated in both animals, the control animal exhibited no

BOLD response lateralization whereas the FR122047 rat showed attenuated BOLD

response in the peri-lesional tissue.

Moreover, the non-lateralized BOLD response of the control animal is in contrast

to attenuated BOLD response in the subacute study, though both are accompanied by

elevated peri-lesional vascular reactivity. It is possible that with time, ongoing peri-

lesional vascular remodelling catches up with the metabolic demands of the tissue.

169

Figure 2.2

170

Figure 2.2 Functional MRI following chronic COX-1 inhibition in a rat model of

focal sensorimotor ischemia. Maps of BOLD signal changes elicited by hypercapnia

(10% CO₂) in two representative rats after ALZET-mini pump intracerebroventricular

administration of vehicle (a.i.) or FR122047 (a.ii.) are shown overlaid on the

corresponding T₂-weighted structural images (SI, signal intensity). The treatment was

initiated 7 days following injection of endothelin-1 into the right sensorimotor cortex and

terminated after 10 days, with imaging performed 24 hours following treatment

cessation. The average normalized signal intensity (SI) in the M1/M2 ROIs (top, black

trace: contralateral ROI; bottom, blue trace: peri-lesional ROI) for the vehicle-

administered control subject are plotted in (a.iii.) and for the FR122047 treated animal

are displayed in (a.iv.). These traces were normalized, animal-wise, to the maximum

signal in that animal’s contralateral M1/M2 ROI. Periods during which 10% CO₂ was

delivered are indicated by grey rectangles. Concomitant blood flow changes are shown

in (b.i.) for a vehicle administered rat; and in (b.ii) for a FR122047 treated animal. The

control animal showed no lateralization in ΔBOLD (2.6 ± 0.1% vs. 2.8 ± 0.1%), with a

peri- to contralesional ratio of 0.9 ± 0.1; and doubled peri- vs. contralesional ΔCBF, with

a peri- to contralesional ratio of 2.3 ± 0.1. In the FR122047 treated animal, BOLD

response was attenuated peri- (3.7 ± 0.3%) vs. contralesionally (6.1 ± 0.3%); whereas

peri-lesional Δperfusion was increased by 20% relative to that of the contralesional

cortex (ratio of 1.2 ± 0.1). Scale bar (a.i. and b.i.) 2mm. Although the across-brain

average BOLD responses were higher in the treated rat than in the vehicle-administered

rat, this inter-subject variability was not significant cohort-wise; but it did motivate the

interhemispheric normalization within each subject.

171

Figure 2.3

172

Figure 2.3 Neuronal activity following chronic COX-1 inhibition in a rat model of

focal sensorimotor ischemia. Intracranial array electrophysiological recordings were

done in the same animals shown in Figure 2.2. Coherence and power estimates were

evaluated in the control animal (a.) and in the FR122047 treated animal (b.). Normalized

heat-plots of the multi-electrode array recordings from contralesional (left, i.), and

ipsilesional (right, ii.) hemispheres. We recorded from 0 to 0.5mm from Bregma and

plotted distance along the x-axis, and from 0.3Hz to 100Hz and plotted frequency along

the y-axis. The contralesional hemispheres (a.i. and b.i.) magnitude squared coherence

(msc) estimates between neighbouring electrodes showed a high degree of coherence

across the cortex throughout the frequency spectrum in the right (contralesional)

hemispheres. In the placebo administered animal, the ipsilesional hemisphere (a.ii.)

msc was reduced, especially immediately proximal to the site of injury (within 0-0.25mm

of Bregma). In the FR122047 treated animal, the msc reduction in the peri-lesional

tissue was less pronounced. The cross-electrode averaged msc is plotted for the

contralesional hemisphere (in green) and for the ipsilesional hemisphere (in black) for

the control animal (a.iii.) and for the FR122047 treated animal (b.iii.). The contralesional

hemispheres (green in a.iii. and b.iii.) showed a higher coherence than the ipsilesional

hemispheres (black in a.iii. and b.iii.) across all frequency bands. Contra- vs. ipsilesional

msc ratios in each band are quoted below and above the corresponding traces in a.iii

and b.iii. The ipsi- to contralesional ratio of the average neuronal power within each

frequency band is plotted in (a.iv. and b.iv.). In both animals, the average neuronal

power was attenuated in the ipsilesional hemisphere relative to contralesional

hemisphere in each frequency band. However, in all frequency bands, this attenuation

was greater in the vehicle only administered animal relative to the FR122047 treated

animal, theta: 73% vs. 90%, alpha: 37% vs. 58%, beta: 21% vs. 31%, low gamma: 12%

vs. 13%, and gamma: 24% vs. 28%.

173

To examine the attenuation of BOLD response in the peri-lesional zone in the

FR122047 treated animal, despite the increased vascular reactivity, we performed

intracranial array electrode recordings in the same animals (Figure 2.3). During rest, the

brain generates oscillations in different frequency bands, which are believed to be the

neuronal correlate of different behaviours and cognitive states (for review see

Womelsdorf et al. 2014). We found high coherence in the contralesional cortex of both

animals (green traces in Figure 2.3.a.iii. and b.iii.). In the peri-lesional cortices (black

traces in Figure 2.3.a.iii. and b.iii.), the coherence was lower, indicating stroke-induced

impairment in neuronal functioning; the treated animal, however, exhibited a smaller

degree of attenuation across all frequency bands (Figure 2.3.b.iii.) than did the placebo

administered animal (Figure 2.3.a.iii.). Notwithstanding the attenuated BOLD response,

these data suggest COX-1 inhibition in the chronic phase after ischemia may exert a

beneficial effect on neuronal communication and putative cortical reorganization. To

examine underlying neuronal metabolic activity, we estimated neuronal power as a

correlate of neuronal metabolic intake. In both animals the average neuronal power was

attenuated in the peri-lesional relative to that in the contralesional cortex in each

frequency band. However, in all frequency bands this attenuation was smaller in the

FR122047 treated animal than in the placebo administered animal: (control vs. treated):

theta: 73% vs. 90%, alpha: 37% vs. 58%, beta: 21% vs. 31%, low gamma: 12% vs.

13%, and gamma: 24% vs. 28% (cf. Figure 2.3. a.iv. and b.iv.). In summary, the treated

animal exhibited higher relative (peri-lesional vs. contralesional) coherence and higher

relative neuronal power. The peri-lesional BOLD response attenuation in FR122047

treated animal was thus associated with improved neuronal function and may be

attributable to high oxygen extraction accompanying elevated neuronal activity, spurred

by inhibition of microglial invasion by COX-1. These examples are in line with the acute

phase observations of depressed CBF and metabolic responses despite the recovery of

somatosensory evoked potentials [Ueki et al. 1988]. Combined, these findings

emphasize the need for multi-modal examination of neuronal, vascular, and glial

function as well as metabolism over the course of ischemic injury progression so as to

establish a framework for interpreting BOLD signal changes in the highly dynamic peri-

infarct region.

174

2.7 Conclusions and future work

Notwithstanding the complexities, BOLD fMRI still has significant potential to

advance the research of neurophysiological changes in the chronic stage of stroke

recovery, due to its easy implementation in patient populations and its robustness.

Longitudinal BOLD fMRI measurements based on signal intensity changes have been

consistently proven more reliable than activation volumes in ischemia patients

[Kimberley et al. 2008] and in healthy subjects [Chung et al. 2005, Chung et al. 2007,

Carey et al. 2002, Kimberley et al. 2007]. For the potential of fMRI to be fully realized,

however, a much more comprehensive understanding of the neurovascular coupling

over this period is needed. Such understanding necessitates that BOLD fMRI be widely

acquired and combined with other, more specific cellular measures as infarct affects all

cell types, neurons, glia, and vascular endothelial cells, both acutely and chronically. As

was the case with BOLD fMRI applications to neuroscientific research in physiological

conditions, multi-modal preclinical studies can be used to quantitatively characterize

neurovascular state in ischemic progression and yield an ischemia-specific model of

BOLD signal contrast to support the interpretation of fMRI findings in clinical research.

In particular, combining BOLD fMRI with in vivo measurements of glucose and oxygen

metabolism in addition to direct measures of vascular state will provide important insight

into the determinants of BOLD contrast in ischemic injury progression. In addition,

combining BOLD with CBF measurements in calibrated fMRI experiments may provide

quantitative estimates of the relative oxygen consumption during the execution of a task

[Hoge et al. 2012, Shu et al. 2016, Simon et al. 2015, Chen et al. 2010]. This approach

relies on CMRO2 measurements during iso-metabolic perturbation and the conservation

of the assumed blood flow and volume coupling between iso-metabolic stimulation and

the functional stimulus/task under investigation. The identification of clinically

translatable iso-metabolic perturbers remains an active area of research [Peng et al.

2016], as does the relationship between flow and volume changes under various

(patho)physiological conditions. Further insight into the metabolic changes in the peri-

infarct tissue may be attained by combining BOLD fMRI with complementary MRI

approaches to assess tissue pH, such as, amide proton transfer imaging. It’s preclinical

applications have shown that MCAO-induced pH deficits extend beyond the region of

175

diffusion changes, but are more circumscribed than hypo-perfusion [Zhe-Sun et al.

2007], and promising preliminary results have also been obtained in stroke patients

[Zhe-Sun et al. 2010, Harston et al. 2015]. Due to the spatial heterogeneity of the meta-

stable peri-infarct zone, these studies will profit from improved spatial resolution [Shen

et al. 2015]. Moreover, such assessments will need to be undertaken longitudinally, at

multiple time-points following injury due to the highly dynamic neurovascular changes in

the chronic stage of ischemic injury progression. Finally, the generalizability of the

preclinical findings will be greatly extended by the careful consideration of the

appropriateness of the preclinical model of ischemia [Murphy & Corbett 2009],

modelling of relevant comorbidities that are known to exert a profound impact on

outcomes in patients [Buga et al. 2013], as well as standardization of these protocols to

enable the integration of data across different studies and centres [Milidonis et al. 2015].

The mechanistic understanding of ischemic injury and recovery ensuing from the careful

application of BOLD fMRI in combination with other modalities in preclinical models will

be invaluable in the development of new therapeutic treatments for stroke.

In particular, fMRI may prove a valuable tool in the preclinical

development of novel pleiotropic therapeutic approaches, called for by the evidence of

peri-lesional plasticity and thus potential for functional recovery and the hitherto poor

efficacy of exclusively neuroprotective approaches [Woodruff et al. 2011]. In addition to

affecting neuronal populations, these treatments will target inflammatory cells and

cerebral vasculature in the chronic stage of stroke [Niwa et al. 2001], making multi-

modal fMRI (combining BOLD fMRI with MRI-based measurements of blood flow,

glucose metabolism, pH, and inflammation) a useful tool for the comprehensive

characterization of treatment effects. Another important aspect in treatment

development for stroke is elucidating the contributors to subacute reperfusion

damage/edema. Recent research suggests that reperfusion-induced injury in the

subacute stage may be prevented by inhibiting microglial and astrocytic production of

matrix metalloproteinases responsible for delayed BBB breakdown (reviewed in by Da

Fonseca et al. 2014). fMRI is particularly beneficial for these investigations due to its

robustness, sensitivity and easy repeatability, all the while providing whole brain

coverage and thus affording the interpretation of signal changes in affected areas in

relation to those in distal (notionally intact) brain regions. fMRI use in the development

176

of such interventions may yield indispensable information on spatio-temporal changes in

neurovascular functioning (e.g. oxygen extraction with calibrated BOLD) by at risk

tissue, furthering our understanding of the mechanisms of subacute reperfusion injury,

which may allow the extension of the time-window for safe and effective application of

thrombolytics.

Functional Neuroimaging of Recovery from Focal …...ii Functional Neuroimaging of Recovery from...

Documents

Transcript of Functional Neuroimaging of Recovery from Focal …...ii Functional Neuroimaging of Recovery from...