After intracerebral hemorrhage, oligodendrocyte precursors

proliferate and differentiate inside white-matter tracts in the

rat striatum

by

Michael Jason Emmanuel Joseph

A thesis submitted in conformity with the requirements

for the degree of Master of Science
 Graduate Department of Physiology

University of Toronto

© Copyright by Michael Jason Emmanuel Joseph 2016

ii

After intracerebral hemorrhage, oligodendrocyte precursors

proliferate and differentiate inside white-matter tracts in the rat

striatum

Michael Jason Emmanuel Joseph

Master of Science

Department of Physiology

University of Toronto

2016

Abstract

Using a rat model of ICH, I quantified survival, proliferation and differentiation of

oligodendrocyte-lineage cells (Olig2+) in the peri-hematoma region, surrounding striatum

and contralateral striatum from 1–28 days. In the peri-hematoma, their density increased

dramatically and peaked at 7 days, coinciding with disorganization of myelinated axon

bundles. By 3 days, many were proliferating locally. By 14 days, their density declined,

and by 28 days, it was the same as the contralateral striatum. At 14 and 28 days, many

surviving axons were aligned into white-matter bundles, which appeared less swollen or

fragmented. Oligodendrocyte maturation was prevalent over the 28-day period. Densities

of immature OPCs (NG2+Olig2

+) and mature (CC-1

+Olig2

+) oligodendrocytes increased

dramatically over the first week. Regardless of the maturation state, they increased

preferentially inside the white-matter bundles. Thus, endogenous oligodendrocyte

precursors proliferate and differentiate in the peri-hematoma region and have the

potential to re-myelinate axons tracts after hemorrhagic stroke.

iii

Acknowledgements

The work presented in this thesis would not have been possible without the

support and guidance of several individuals. First and foremost, I would like to thank my

supervisor, Dr. Lyanne Schlichter, for providing me with this opportunity and offering

her guidance throughout this project. Dr. Schlichter’s passion for science and excitement

when discussing project ideas and data is truly inspiring. Dr. Schlichter is very precise in

her writing and data analysis, which has driven me to approach the work with the same

level of detail. She has been very supportive throughout my research and especially

during the process of publishing my first paper. I feel very lucky to have been trained by

a scientist such as herself.

Secondly, I would like to thank my committee members, Dr. James Eubanks and

Dr. Isabelle Aubert. Both Dr. Eubanks and Dr. Aubert went above and beyond as

committee members. They were incredibly supportive during the beginning phase of my

project when I was trouble shooting different experiments and constantly forced me to

think about the big picture. Their comments and suggestions during my committee

meetings helped drive the aims of the project.

I would also like to thank all the members of the Schlichter lab: Dr. Jayalakshmi

Caliaperumal, Dr. Starlee Lively, Sarah Hutchings, Tamjeed Siddiqui, Doris Lam,

Raymond Wong, Roger Ferreira, and Xiaoping Zhu. They have allowed me to bounce

ideas, listened to my rationales and have provided me with very special friendships both

inside and outside the lab. I would especially like to thank Jayalakshmi, Starlee and Sarah

for training me in many of the lab techniques and providing helpful advice throughout my

degree. Jayalakshmi also conducted the experiments on the sagittal tissue sections, which

iv

have been an important element of our story.

Finally, I would like to thank my friends and family. Many of them don’t have the

same appreciation for science as I do; yet, they have provided me with incredible support

and advice throughout my degree and have always had faith in me.

v

Table of Contents

Abstract .............................................................................................................................. ii

Acknowledgements .......................................................................................................... iii

Table of Contents .............................................................................................................. v

List of tables..................................................................................................................... vii

List of figures .................................................................................................................. viii

List of abbreviations ........................................................................................................ ix

Publications ....................................................................................................................... x

Chapter 1. Introduction.................................................................................................... 1

1.1 Intracerebral hemorrhage (ICH) in humans ....................................................... 1

1.1.1 Incidence ............................................................................................................ 1

1.1.2 Prognosis ............................................................................................................ 1

1.1.3 Regions affected................................................................................................. 2

1.1.4 Pathophysiology ................................................................................................. 3

1.2 Animal models of intracerebral hemorrhage ....................................................... 4

1.2.1. Autologous blood model ................................................................................... 5

1.2.2 Collagenase model ............................................................................................. 6

1.3 White matter versus gray matter ........................................................................... 8

1.4 White matter damage after ICH ........................................................................... 8

1.4.1 White matter damage in human patients ............................................................ 8

1.4.2 White matter damage in ICH animal models ..................................................... 9

1.4 Oligodendrocytes................................................................................................... 10

1.5 Objectives/Hypotheses .......................................................................................... 14

Chapter 2. Materials and Methods................................................................................ 16

2.1 Animals and surgery ............................................................................................. 16

2.2 Tissue preparation ................................................................................................ 17

2.3 Cresyl violet staining ............................................................................................. 17

2.4 Immunohistochemistry: Staining and analysis .................................................. 18

2.5 Data analysis .......................................................................................................... 22

Chapter 3. Results ........................................................................................................... 25

3.1 Myelin is damaged but some myelinated axons survive in the peri-hematoma

after ICH ...................................................................................................................... 25

3.2 The density of oligodendrocyte-lineage cells increases as a result of

proliferation in the peri-hematoma region ............................................................... 29

3.3 Oligodendrocyte precursor cells (OPCs) and mature oligodendrocytes

increase transiently in the peri-hematoma, preferentially inside white-matter

tracts ............................................................................................................................. 36

vi

Chapter 4. Discussion ..................................................................................................... 43

4.1 Background/overview ........................................................................................... 43

4.2 Salient findings ...................................................................................................... 44

4.3 Ischemic stroke versus ICH .................................................................................. 45

4.3.1 Extent of myelin and axon damage and possible white matter recovery ......... 45

4.3.2 Oligodendrocyte-lineage cells ......................................................................... 47

4.3.3 OPC proliferation ............................................................................................. 49

4.3.4 OPC differentiation .......................................................................................... 52

4.3.5 Oligodendrocyte-lineage cell death ................................................................. 54

4.4 Oligodendrogenesis in rodents versus humans ................................................... 54

4.5 What this study adds ............................................................................................ 55

Chapter 5. Future Directions ......................................................................................... 56

Reference list ................................................................................................................... 59

Copyright Acknowledgements ....................................................................................... 80

vii

List of tables

Table 2.1. Antibodies used to identify white matter, oligodendrocytes and their

maturation state…………………………………………………………………………19

viii

List of figures

Figure 1.1. Development of oligodendrocytes…………………………………………..12

Figure 2.1. Secondary antibody absorption and emission spectra………………………21

Figure 2.2. Hematoma development in the first month after ICH………………………23

Figure 3.1. Myelin damage and axon survival in the peri-hematoma after ICH………..27

Figure 3.2. Numbers of oligodendrocyte-lineage cells increase after ICH, especially in

the peri-hematoma……………………………………………………………………….31

Figure 3.3. Oligodendrocyte-lineage cells preferentially increase inside white-matter

bundles…………………………………………………………………………………...33

Figure 3.4. Oligodendrocyte-lineage cells proliferate in the damaged striatum,

particularly in the peri-hematoma region………………………………………………...35

Figure 3.5. Oligodendrocyte precursor cells (OPCs) increase preferentially in the peri-

hematoma and inside white-matter bundles……………………………………………...37

Figure 3.6. Mature oligodendrocytes increase after ICH, especially inside white-matter

bundles in the peri-hematoma……………………………………………………………40

ix

List of abbreviations

ICH Intracerebral hemorrhage

tPA Tissue plasminogen activator

BBB Blood brain barrier

CNS Central nervous system

MRI Magnetic resonance imaging

DTI Diffusion tensor imaging

TEM Transmission electron microscopy

MBP Myelin basic protein

dMBP Degraded myelin basic protein

APP Amyloid precursor protein

SVZ Subventricular zone

OPC Oligodendrocyte precursor cell

NG2 Neural/glial antigen 2

PDGFRα Platelet-derived growth factor receptor alpha

NSC Neural stem cell

GST-π Glutathione S-transferase pi

APC Adenomatous polyposis coli

CNPase 2’,3’-cyclic nucleotide 3’-phosphodiesterase (CNPase)

MOG Myelin oligodendrocyte glycoprotein

PLP Proteolipid protein

PBS Phosphate buffered saline

PFA Paraformaldehyde

PBT Buffer: 1 M PBS, 0.1% bovine serum albumin and 0.2% Triton X-100

pan-NF Pan-axonal neurofilament

Olig2 Oligodendrocyte lineage transcription factor 2

DAPI 4’-6-Diamidino-2-phenylindole

ANOVA Analysis of variance

CSF Cerebrospinal fluid

MCAO Middle cerebral artery occlusion

IGF-1 Insulin-like growth factor-1

PDGF Platelet-derived growth factor

FGF-2 Fibroblast growth factor-2

BDNF Brain-derived neurotrophic factor

TGF-β Transforming growth factor beta

EGF Epidermal growth factor

Ascl1 Achaete-scute homolog 1

LPS Lipopolysaccharide

x

Publications

Joseph MJ, Caliaperumal J, Schlichter LC. After intracerebral hemorrhage,

oligodendrocyte precursors proliferate and differentiate inside white matter tracts

in the rat striatum. Transl Stroke Res. 2016. doi:10.1007/s12975-015-0445-3.

1

Chapter 1. Introduction

1.1 Intracerebral hemorrhage (ICH) in humans

1.1.1 Incidence

Stroke is a leading cause of death worldwide (Thrift et al., 2014). It consists of

two main types, ischemic stroke and intracerebral hemorrhage (ICH). Ischemic stroke is

caused by the transient or permanent occlusion of a cerebral artery (Donnan et al., 2008).

ICH is caused by bleeding in the brain parenchyma due to: (i) the spontaneous rupture of

small penetrating arteries or arterioles (primary ICH; ~80% of cases); or (ii) pre-existing

or acquired conditions (secondary ICH) such as vascular malformations, brain tumours,

or use of anticoagulants (Sahni and Weinberger, 2007). The incidence of ICH is

increasing, likely due to the aging population demographic (Krishnamurthi et al., 2014).

ICH comprises only 10-20% of all stroke cases (Feigin et al., 2009); however,

approximately 30-40% of all ischemic strokes undergo some degree of hemorrhagic

transformation (Wang and Lo, 2003). Treatment with tissue plasminogen activator (tPA)

in ischemic stroke patients to promote re-vascularization might also lead to hemorrhagic

transformation (Sussman and Connolly, 2013). Unfortunately, despite the prevalence of

ICH, there is less research into its pathophysiology and the underlying cellular processes.

Given that primary ICH is more prevalent and develops in the absence of any underlying

conditions or treatments, it is the focus of this study.

1.1.2 Prognosis

The overall mortality from ICH is ~40% by 1 month (van Asch et al., 2010) and

>70% at 5 years (Poon et al., 2014). ICH has a significantly higher mortality rate

2

compared to ischemic stroke within the first 3 weeks (Andersen et al., 2009). The causes

of death vary according to the time course of the disease. Patients who survive the initial

bleed often die due to brain cell death secondary to the bleeding, increased intracranial

pressure, hematoma expansion, or cardiac causes (Naidech et al., 2009; Godoy et al.,

2015). Few survivors regain functional independence (Poon et al., 2014; Hemphill et al.,

2015). No medical or surgical intervention has been demonstrated to reduce mortality or

improve outcomes in patients with ICH. Thus, current treatment focuses on supportive

care and on reducing the impact of the damage (Anderson, 2009; Sangha and Gonzales,

2011).

1.1.3 Regions affected

Clinically, primary ICH is often categorized according to the location of the

bleed, lobar or deep. This distinction is based on the presumed underlying etiology. In

lobar ICH, cerebral amyloid angiopathy, which predominately affects arteries, arterioles

and capillaries in the cerebral cortex (Vinters, 1987), plays a dominant role in its

pathogenesis (Ritter et al., 2005). The deposition of amyloid protein in the vessel walls

causes the vessels to become brittle and prone to breakage. ICH caused by hypertension,

the most significant modifiable risk factor of ICH (Ariesen et al., 2003), typically occurs

in deep brain areas (Jackson and Sudlow, 2006). In a population-based study, common

locations included the basal ganglia, internal capsule, periventricular white matter,

thalamus (50%), cerebellum (10%) and pons (6%) (Flaherty et al., 2006). Hypertension-

related hemorrhages typically occur in these deep areas because they are located close to

the high-pressure area of the circle of Willis (Sutherland and Auer, 2006).

3

For this study, I induced ICH in the rat striatum, the largest component of the basal

ganglia, where a considerable proportion of ICHs occur in humans. Because the striatum

is fairly large; it can contain the bleed and reduce the risk of blood leaking into the

ventricle. Unless otherwise specified, the majority of the background research is obtained

from animal models of striatal ICH.

1.1.4 Pathophysiology

The burst of a single vessel gives rise to an initial hematoma. The primary injury

phase after ICH occurs within minutes to hours following onset. The resulting hematoma

causes local tissue destruction (Qureshi et al., 2001; Keep et al., 2012). As the

hemorrhage develops, tissue is physically destroyed by blood entering the brain, and the

mass effect causes shrinkage of surrounding structures, including the ventricles (Qureshi

et al., 2001). Hematoma expansion is common following ICH in humans, is present in

~73% of cases (Davis et al., 2006) and correlates with outcome (Brott et al., 1997).

Edema and changes in cerebral blood flow also occur (Gebel et al., 2002).

The introduction of blood components and generation of post-hemorrhage cellular

debris exacerbates tissue damage in the region immediately adjacent to the hematoma,

the peri-hematoma. This comprises the secondary injury phase. Thrombin and iron

released from ruptured erythrocytes diffuse out of the hematoma and act as

chemoattractants for the subsequent infiltration and activation of inflammatory cells (Xue

and Del Bigio, 2001; Xi et al., 2004; Wang and Dore, 2007). There is a prominent

inflammatory response in which microglia, the brain’s innate immune cells, become

activated; and circulating leukocytes (neutrophils and macrophages) are recruited to the

4

site of injury (Taylor and Sansing, 2013). The inflammatory response during the

secondary phase can last for days to weeks (Xue and Del Bigio, 2001).

Damage to the peri-hematoma region is delayed. Thus, it is considered potentially

salvageable tissue, and is therefore a target for some therapeutic interventions. Thus, I

have emphasized the peri-hematoma region throughout this thesis, as opposed to the

likely unsalvageable hematoma.

1.2 Animal models of intracerebral hemorrhage

Various animal models of ICH have been developed in order to study the

pathophysiology at a preclinical level (James et al., 2008; Ma et al., 2011). Animal

models also allow a defined onset of injury, the investigation of novel therapies and gene

knockouts, and homogeneous experimental groups.

The earliest ICH animal model is from 1963, using dogs (Whisnant et al., 1963),

and this was followed by monkeys (Segal et al., 1982), rats and rabbits (Ropper and

Zervas, 1982; Kaufman et al., 1985), pigs (Wagner et al., 1996), and most recently, mice

(Nakamura et al., 2004). As important as it is to find an animal subject that will mimic

processes in the human brain, there are other challenges in creating an animal model.

Rhesus monkeys, one of the earliest model species that was used, are closely related to

humans genetically and have several physiological and neurobiological similarities

(Rhesus Macaque Genome et al., 2007). However, monkeys and dogs are costly and there

have been various levels of restrictions and regulations in using them (Ma et al., 2011).

Thus, these ICH models were quickly discontinued.

Mice and rats are predominately used (MacLellan et al., 2012). These species are

inexpensive and easily housed, have a vast array of reagents for immunohistochemistry

5

and molecular biology, well-developed paradigms for testing behaviour and functional

outcomes, and the potential for using transgenic and knockout animals (Kirkman et al.,

2011). Rats offer several advantages over mice. The cerebrovascular anatomy and

physiology is reasonably similar between rats and humans (Yamori et al., 1976). There is

also a positive correlation between brain size and the amount of myelinated axons (Wang

et al., 2008), and the larger brain of rats is especially useful for quantifying differences

across brain regions. Two common methods of inducing ICH in rodents are to inject

either bacterial collagenase or autologous blood into the striatum. These models have

contributed significantly to our knowledge of ICH-induced injury mechanisms.

For this thesis, I have used the collagenase model of ICH in rats. I will briefly discuss

both the autologous blood and collagenase models and the strength of the collagenase

model.

1.2.1. Autologous blood model

The first blood model was developed in 1982 and was one of the earliest animal

models of ICH. Fresh blood from the ventricle of a donor rat was rapidly infused into the

caudate nucleus of the experimental ICH rat (Ropper and Zervas, 1982). In 1984, Bullock

and colleagues introduced a variation of this model using autologous blood from the

femoral artery rather than donor blood. A needle was used to bridge the femoral artery

and the caudate nucleus (Bullock et al., 1984). However, there was a lack of

reproducibility with this model, likely because of variations in blood pressure between

animals, affecting the rate of blood infusion (Ma et al., 2011). Later, researchers

employed a microinfusion pump to allow autologous blood (again extracted from the

femoral artery) to be infused at a constant rate into the caudate nucleus, creating a

6

controllable and reproducible hematoma (Yang et al., 1994). Two major issues remained.

(i) There was often reflux of blood up the needle-penetration tract into the ventricular

system, or extension of blood into the subdural space, causing inconsistent hematoma

volumes. (ii) Damage to the femoral artery affected the assessment of neurobehavioural

deficits (Ma et al., 2011). The autologous blood model was later refined to include two

injections: after an initial injection, the blood was allowed to clot and the rest was

injected (Deinsberger et al., 1996). This minimized blood reflux through the needle-

penetration tract. Other improvements include extracting blood from the tail artery and

co-injecting heparin to prevent clotting (Belayev et al., 2003). However, heparin also

antagonizes thrombin, a known mediator of secondary injury (Xi et al., 1998), which

might affect the damage produced by the model.

The autologous blood model mimics a single large bleed that can occur in human

patients, and is especially useful for studying the mechanisms of mechanical injury

immediately after ICH (MacLellan et al., 2008). The disadvantages are that it does not

create spontaneous bleeding, is not suitable for studying the re-bleeding that is often seen

in humans (James et al., 2008), and it is technically difficult to perform.

1.2.2 Collagenase model

Injecting bacterial collagenase into the brain to simulate ICH was first introduced

in 1990 (Rosenberg et al., 1990). Collagenase is an enzyme that catalyzes the hydrolysis

of collagen, which is an integral component of the extracellular matrix of the blood brain

barrier (BBB). Bleeding occurs within 10 minutes of collagenase injection (Rosenberg et

al., 1990) and lasts for several hours (Adeoye et al., 2011), mimicking the recurrence of

bleeding and hematoma expansion in human patients (Passero et al., 1995;Qureshi et al.,

7

2001). An advantage of this model is that it causes spontaneous bleeding in the

parenchyma, which closely recapitulates spontaneous human ICH. In order to limit the

area of blood vessels affected and the size of the bleed, most experimenters inject

collagenase in a very small volume of saline (~0.5 μl). This model has been used to

create hematomas of a consistent size and in a precise location (MacLellan et al., 2012).

It also avoids the blood backflow up the needle-penetration tract that occurs in the

autologous blood model.

Despite its advantages, the collagenase model does not accurately mimic the

rupture of a single artery or arteriole normally observed in human patients. Instead,

collagenase dissolves the extracellular matrix around capillaries to produce a hemorrhage

(Del Bigio et al., 1996). It was initially thought that collagenase was directly toxic or

resulted in an exaggerated inflammatory response (Del Bigio et al., 1996) compared to

human ICH; however, in vitro studies argue against this (Matsushita et al., 2000; Chu et

al., 2004).

Several studies have directly compared the autologous blood and collagenase

models (James et al., 2008; MacLellan et al., 2008; Manaenko et al., 2011). Neither

model completely mimics the spontaneous rupture of a single artery; however, the

collagenase model is often considered superior for several reasons. After collagenase-

induced ICH, more neurons are lost outside of the hematoma and BBB damage is evident

(MacLellan et al., 2008), making this model useful for examining the progress of brain

injury outside the hematoma. The autologous blood model minimally affects tissue

surrounding the hematoma. Secondly, the collagenase model results in long-lasting

functional deficits, while functional recovery is rapid after autologous blood injection

8

(MacLellan et al., 2008). Thus, the collagenase model is much more similar to human

ICH where patients often have severe, persistent functional deficits.

Given that the collagenase model more closely recapitulates human ICH, it was chosen

as the model for this study.

1.3 White matter versus gray matter

White matter and gray matter comprise two distinct tissues in the central nervous

system (CNS). The two regions are visually distinguishable by colour, hence the

nomenclature. The regions are also topographically different. Gray matter is composed of

neuronal cell bodies and unmyelinated axons, which communicate in local circuits. White

matter is comprised of axons, myelin sheaths that enwrap them, and oligodendrocytes,

the cells that produce myelin (Petty and Wettstein, 1999). Myelin is necessary for

saltatory conduction of action potentials along axons, via sodium ion fluxes at the nodes

of Ranvier. Thus, myelinated axons are essential for communicating between distant

areas in the CNS, with millions of afferent and efferent axons arranged in white-matter

tracts.

The striatum contains a large number of myelinated axon bundles that are surrounded by

neuronal cell bodies. Thus, the induction of ICH in the striatum allows a unique

opportunity to examine both gray matter and white matter injury in a single structure.

1.4 White matter damage after ICH

1.4.1 White matter damage in human patients

White-matter damage often occurs after acute CNS injury (Matute and Ransom,

2012) and is considered an important predictor of the outcome, including after stroke

9

(Medana and Esiri, 2003; Lee et al., 2010). Because of its crucial role, even small white-

matter lesions can produce devastating neurological deficits (Stys and Lipton, 2007).

After ICH, primary damage to white matter by the bleed or secondary injury events can

lead to Wallerian degeneration (Fukui et al., 1994; Kazui et al., 1994). Significant white-

matter damage was seen in CT scans of 77% of patients (Smith et al., 2004) and severe

white-matter lesions were associated with large ICH volumes (Lou et al., 2010), severe

Glasgow Coma Scale scores (Lee et al., 2010), and cognitive decline (Benedictus et al.,

2015). Magnetic resonance imaging (MRI), which shows a strong contrast between

signals from gray and white matter, showed that white matter hyper-intensity volumes

were higher after ICH than ischemic strokes (Rost et al., 2010). Diffusion tensor imaging

(DTI) has been used to reveal the integrity of white-matter tracts after ICH (Goh et al.,

2015) and predict future motor function (Yoshioka et al., 2008; Kuzu et al., 2012;

Koyama et al., 2013). Transmission electron microscopy (TEM) has been used on

resected tissue from the hematoma of ICH patients and shows myelin sheath thickening

and loosening around the axons, abnormal axonal neurofilaments and axonal atrophy

(Wang et al., 2015).

1.4.2 White matter damage in ICH animal models

Most preclinical, animal studies of ICH have focused on gray matter (Kellner and

Connolly, 2010; MacLellan et al., 2012) with very little research on white-matter damage

and its progression. A review by MacLellan et al., 2012 indicates that only about 6% of

experimental ICH studies have quantified white matter injury. Our lab has been

exploiting the well-established collagenase model of ICH in the rat striatum to quantify

spatial and temporal changes in white-matter damage in the peri-hematoma and

10

surrounding striatum (Wasserman and Schlichter, 2008; Wasserman et al., 2008; Moxon-

Emre and Schlichter, 2011; Lively and Schlichter, 2012a). One of the main components

of myelin is myelin basic protein (MBP). Our lab has used antibodies against MBP, and

against degraded myelin basic protein (dMBP). dMBP staining is inversely related to

MBP staining so that as the level of healthy myelin (MBP) decreases, the level of

unhealthy myelin (dMBP) increases. Our lab has shown significant loss of healthy myelin

in the hematoma at 3 days and progressive, extensive damage to myelin in the peri-

hematoma during the first week after ICH. At 7 days, axons, labeled with NF200, in the

peri-hematoma were diffuse and no longer recognizable as bundles. The accumulation of

amyloid precursor protein (APP), which indicates impaired axonal transport, was shown

at the hematoma edge and extending into the surrounding striatum. The damage to myelin

bundles and axons correlated temporally and spatially with inflammation and microglial

activation (Wasserman and Schlichter, 2008; Wasserman et al., 2008; Moxon-Emre and

Schlichter, 2011; Lively and Schlichter, 2012a). Others have shown widespread edema in

the white matter for at least 1 week after ICH (Del Bigio et al., 1996) and a reduction in

the volume of the corpus callosum at 6 weeks (MacLellan et al., 2008). At 100 days, the

number of white-matter bundles in the hematoma was reduced; however, quantifying the

entire striatum revealed an increased number of bundles, though each appeared thinner

(Felberg et al., 2002). Functional recovery and efficacy of potential treatments will

require both axon survival and re-myelination.

1.4 Oligodendrocytes

Oligodendrocytes are the myelinating cells of the CNS. Oligodendrocytes wrap

layers of specialized cell membrane around axons to form myelin, which decreases the

11

capacitance and increases the resistance across the axonal membrane, speeding neural

communication (Tomassy et al. 2016). It has also been proposed that oligodendrocytes

provide trophic support to neurons and to long axons that might not receive adequate

support from intra-axonal trafficking alone (Nave, 2010).

The process of oligodendrogenesis is depicted in Fig. 1.1. In the healthy brain, the

SVZ is a major site of oligodendrogenesis during development (Menn et al., 2006) and

throughout life (Young et al., 2013; Kang et al., 2014). Oligodendrocyte precursor cells

(OPCs) express neural/glial antigen 2 (NG2) and platelet-derived growth factor receptor

alpha (PDGFRα) (Sim et al., 2006). They are generated from neural stem cells (NSCs) in

the SVZ. Some OPCs migrate from the SVZ to other brain regions, where they remain as

precursors and proliferate at a low rate (Young et al., 2013; Kang et al., 2014), or

undergo maturation (Dawson et al., 2003). During maturation, the NG2 and PDGFRα

markers are replaced with glutathione S-transferase pi (GST-π), adenomatous polyposis

coli (APC; also called CC-1) and myelin-specific proteins such as 2’,3’-cyclic-nucleotide

3’-phosphodiesterase (CNPase), myelin oligodendrocyte glycoprotein (MOG),

proteolipid protein (PLP) and MBP (Baumann and Pham-Dinh, 2001).

Once an OPC differentiates into a mature oligodendrocyte, it extends multiple

processes that individually ensheath axon segments and generate the spiral layers of

modified cell membrane that compose myelin (Geren and Raskind, 1953). Live imaging

of myelination by rodent oligodendrocytes in vitro, or in vivo during zebrafish

development have demonstrated that oligodendrocytes establish all of their myelin

wrappings within a few hours after differentiation; after which, there appears to be little

plasticity (Watkins et al., 2008; Czopka et al., 2013). A single oligodendrocyte can form

12

Figure 1.1 Development of oligodendrocytes

Neural

stem cell OPC

NG2-

PDGFRa-

Olig2+/-

Sox10-

NG2+

PDGFRa+

Olig2+

Sox10+

Mature

oligodendrocyte

CC-1+

CNPase

MBP

PLP

MOG

Olig2+

Sox10+

13

wrappings within a few hours after differentiation; after which, there appears to be little

plasticity (Watkins et al., 2008; Czopka et al., 2013). A single oligodendrocyte can form

myelin segments on up to 50 neighbouring axons (Baumann and Pham-Dinh, 2001).

In the healthy rodent brain, myelinating oligodendrocytes continue to be

generated from OPCs in adulthood (Dimou et al., 2008; Rivers et al., 2008; Emery, 2010;

Young et al., 2013). Oligodendrocyte turnover is thought to sustain basal myelin turnover

and brain plasticity (Young et al., 2013; Gibson et al., 2014). However, adult OPCs differ

from their perinatal counterparts in cell cycle duration, rates of migration, differentiation

and responses to growth factors (Simon et al., 2011). After CNS injury and myelin

damage, new oligodendrocytes must be generated, recruited to the damaged region and

undergo maturation for successful re-myelination to occur. In some models of acute CNS

injury, OPC proliferation increases. For instance, OPCs rapidly re-entered the cell cycle

after a stab wound (Simon et al., 2011), and they proliferated from 1 to 7 days after

collagenase-induced intraspinal hemorrhage (Sahinkaya et al., 2014). Increased OPC

proliferation was seen in the SVZ in animal models of multiple sclerosis (Nait-Oumesmar

et al., 1999; Nait-Oumesmar et al., 2007).

Re-myelination can be a highly efficient repair process (Miron et al., 2013). It has

been studied extensively in multiple sclerosis (Harlow et al., 2015), spinal cord injury

(Papastefanaki and Matsas, 2015) and ischemic stroke (Itoh et al., 2015). Cell

replacement therapy to increase the population of oligodendrocytes (Lopez Juarez et al.,

2015) or manipulating signaling pathways to enhance myelination (Flores et al., 2008;

Goebbels et al., 2010; Snaidero et al., 2014) are some examples of treatments being

attempted. However, the success or failure of re-myelination has not yet been

14

investigated after ICH. The paucity of information raises several important questions. Do

oligodendrocyte-lineage cells remain in the damaged region after ICH? If so, do they

proliferate and remain as non-myelinating OPCs, or do they differentiate into mature

cells? Does oligodendrogenesis occur within the peri-hematoma? Here, we addressed

these questions from 1 to 28 days after collagenase-induced ICH in the rat striatum.

1.5 Objectives/Hypotheses

The objective of this study was to characterize the spatial and temporal profile of

oligodendrocytes after ICH. I induced an experimental ICH in the rat striatum, and used

immunohistochemical techniques to label oligodendrocyte-lineage cells and distinguish

OPCs from mature oligodendrocytes. I quantified the cells from 1 to 28 days after ICH

and investigated regional differences in the peri-hematoma, surrounding striatum and

contralateral striatum. Sagittal brain sections were used to investigate the localization of

oligodendrocytes inside the longitudinal, corticostriatal white matter tracts of the striatum

compared to outside the tracts. Finally, I investigated the time course of oligodendrocyte

proliferation after ICH.

Hypothesis 1: Given that CNS injury, including ICH, damages myelin, there will be an

initial loss of oligodendrocytes in the striatum after ICH. There will also be an increase in

OPC proliferation to expand the population of oligodendrocytes.

Hypothesis 2: Given that mature oligodendrocytes are necessary to produce myelin, after

ICH, there will be an early increase in OPCs in the damaged striatum, and over time, they

will differentiate into mature oligodendrocytes.

15

Hypothesis 3: Given that oligodendrocytes are predominately located within white

matter tracts and that oligodendrocytes are necessary to myelinate axons,

oligodendrocytes will be observed preferentially localized within white matter tracts

rather than outside.

16

Chapter 2. Materials and Methods

2.1 Animals and surgery

All procedures were performed in accordance with guidelines established by the

Canadian Council on Animal Care and approved by the University Health Network

Animal Care Committee. Male Sprague-Dawley rats weighing 300–350 g (Charles River,

Saint-Constant, Quebec, Canada) were randomly assigned to either the saline control

(n=3) or ICH group (n=3 or 4 per time point). They were housed at two rats per cage in a

temperature- and humidity-controlled room under a diurnal light cycle (lights on from 6

am to 6 pm), and provided with food and water ad libitum.

Surgical procedures were performed aseptically. Rats were anesthetized with

isoflurane (4% induction, 2% maintenance) and placed in a stereotaxic frame (David

Kopf Instruments, Tujungam, CA). Their heart rate and breathing rate were monitored

before and during the surgery using the MouseOX® system (Starr Life Sciences, St.

Laurent, Quebec, Canada). The surgical site was cleansed with povidone-iodine. A 2 cm

long incision was made to expose the skull. Subsequently, a 1 mm diameter burr hole was

drilled into the skull 3.0 mm lateral and 0.1 mm anterior to bregma. A 26-gauge needle

attached to a micro-pump (model UltraMicroPump II; World Precision Instruments,

Sarasota, FL) was inserted 6.0 mm deep from the surface of the skull. Rats were injected

with 0.5 μL of sterile 0.9% saline (controls) or saline containing 0.03 U of type IV

collagenase (Sigma-Aldrich, Oakville, Ontario, Canada) for experimental ICH. The

injection rate was 100 nL/min. The dose of collagenase was chosen after a titration pilot

study so that the hematoma was restricted to the striatum, with sufficient tissue remaining

to have regions of both peri-hematoma and surrounding striatum. Upon completion of

17

injection, the needle was left in place for 10 min to prevent backflow. The wound was

closed with suture clips, and Xylocaine was delivered subcutaneously near the injection

site to reduce post-operative pain. Body temperature was maintained throughout the

surgery and recovery period using a thermostatic heating pad. All animals regained

consciousness within 10 min, and only rats with an ICH demonstrated an ipsilateral

turning bias. Their ability to eat, drink, and groom themselves was not impaired, and no

rats died because of surgery or ICH induction.

2.2 Tissue preparation

At 1, 3, 7, 14 or 28 days after surgery, rats were euthanized by an overdose of

isoflurane. They were transcardially perfused through the heart with 150 mL of 1 M

phosphate buffered saline (PBS; Wisent Bioproducts, St-Bruno, Quebec, Canada),

followed by 150 mL of 4% paraformaldehyde (PFA; pH 7.4). Brains were removed, post-

fixed in 4% PFA overnight, and cryoprotected by sequential incubation in 10% sucrose

for 24 h followed by 30% sucrose for 48 h. The brains were cut coronally or sagittally

and embedded in tissue-freezing medium (Ted Pella, Redding, CA), flash frozen in a

mixture of dry ice and 95% ethanol, and stored at –40°C until sectioning. Then, 16 μm

thick sections were cut using a cryostat (model CM350S; Leica, Richmond Hill, Ontario,

Canada), placed onto gelatin-coated glass slides (1% gelatin, 0.5% chromium potassium

sulfate), and stored at –40°C until used.

2.3 Cresyl violet staining

Hematoma development at each time point was examined in coronal and sagittal

sections using cresyl violet. Frozen brain sections were initially dehydrated in ethanol to

remove phospholipids and fixation chemicals (Crapo et al., 2011), and then rehydrated in

18

distilled water for 1 min before adding 0.25% cresyl violet (Sigma-Aldrich) in a 200 mM

acetate buffer (pH 4.0) until well stained (Sirkin, 1983). The sections were immersed for

1 min in 0.25% glacial acetic acid in ethanol to differentiate the white matter, incubated

for 4 min in CitriSolv (Fisher Scientific, Pittsburgh, PA) to clear the unstained tissue, and

then mounted using DPX (Sigma-Aldrich). Images were made using a flatbed scanner

(model HP Scanjet 4070; Hewlett-Packard, Palo Alto, CA).

2.4 Immunohistochemistry: Staining and analysis

Immunohistochemistry was used to examine the extent of white matter injury, and

the density, proliferation, and differentiation of oligodendrocyte-lineage cells. Frozen

brain sections were thawed at room temperature and rehydrated for 20 min in a buffer

containing 1 M PBS at pH 7.5, 0.1% bovine serum albumin (Sigma-Aldrich), and 0.2%

Triton X-100 (Sigma-Aldrich) (PBT). To reduce nonspecific antibody binding, sections

were blocked for 2 h in PBT containing 10% donkey serum (Jackson Immunoresearch,

Baltimore, PA). For staining, sections were incubated overnight at room temperature in

PBT buffer containing 5% donkey serum and two or three of the following primary

antibodies. Antibodies, their sources and dilutions used are listed in Table 2.1.

To assess myelin and white-matter tracts, we used polyclonal antibodies directed

against MBP; i.e., rabbit anti-MBP (1:1000; ab40390, Abcam, Toronto, Ontario, Canada)

or chicken anti-MBP (1:400; ab106583, Abcam). Axons were labeled with a mouse

monoclonal antibody against pan-axonal neurofilaments (pan-NF; 1:1000; 837901,

Covance, Montreal, Quebec, Canada). To label all oligodendrocyte-lineage cells, we used

antibodies against ‘oligodendrocyte lineage transcription factor 2’ (Olig2); either rabbit

polyclonal anti-Olig2 (1:500; AB9610, Millipore, Etobicoke, Ontario, Canada) or mouse

19

Table 2.1. Antibodies used to identify white matter, oligodendrocytes and their

maturation state

Antibody Used to Label Host species Concentration Source

MBP Myelin Rabbit/chicken 1:1000/1:400 Abcam, Toronto,

ON

Pan-NF Neurofilaments Mouse 1:1000 Abcam

Olig2 Oligodendrocyte

lineage cells

Rabbit/mouse 1:500 Millipore,

Etobicoke, ON

NG2 OPCs Mouse 1:500 Millipore

CC-1 Mature

oligodendrocytes

Mouse 1:500 Millipore

Ki67 Proliferating cells Mouse 1:100 Abcam

DAPI Nuclei of all cells 1:5000 Sigma Aldrich,

Oakville, ON

MBP, myelin basic protein; pan-NF, pan-axonal neurofilament; Olig2, oligodendrocyte

lineage transcription factor 2; NG2, neural/glial antigen 2; CC1, adenomatous polyposis

coli; DAPI, 4’-6-diamidino-2-phenylindole

20

monoclonal anti-Olig2 (1:500; MABN50, Millipore). OPCs were identified by double

staining with anti-Olig2 and a mouse monoclonal antibody against neural/glial antigen 2

(NG2) chondroitin sulfate proteoglycan (1:500; MAB5384, Millipore). In the adjacent

serial brain section, mature oligodendrocytes were identified by double staining with anti-

Olig2 and a mouse monoclonal antibody against adenomatous polyposis coli (APC; also

called CC-1; 1:500; OP80, Millipore). Cell proliferation was monitored with a mouse

antibody against Ki67 (1:100; ab15580, Abcam). After labeling with primary antibodies,

brain sections were washed in PBT (3×10 min) and incubated with appropriate secondary

antibodies for 2 h at room temperature in the dark. Secondary antibodies (all from

Jackson Immunoresearch) were applied at 1:200 dilution (Alexa Fluor 488-conjugated

donkey anti-mouse, Alexa Fluor 488-conjugated donkey anti-rabbit, Alexa Fluor 594-

conjugated donkey anti-mouse, Alexa Fluor 594-conjugated donkey anti-rabbit, Dylight

405-conjugated donkey anti-chicken), and were diluted in PBT and 5% donkey serum.

For double and triple staining, secondary antibodies were chosen from different host

species to avoid cross-reactivity and with different absorption and emission spectra (Fig.

2.1) to minimize spectral overlap and ensure that a given antigen is only stained in one

colour. To label cell nuclei, 4’-6-diamidino-2-phenylindole (DAPI; 1:5000; Sigma-

Aldrich) was applied for 10 min. After washing in PBT (3×10 min), slides were cover-

slipped using Dako Fluorescent Mounting Medium (Dako North America, Inc.,

Carpinteria, CA) and stored in the dark at 4°C. To minimize variability,

immunohistochemistry for a given antibody combination was conducted on the same day

for all rats and time points using the same solutions.

Antibody-stained sections were examined using a confocal microscope (model

21

Figure 2.1 Secondary antibody excitation and emission spectra. The excitation

(dotted line) and emission (solid line) spectra are shown, expressed as the percentage of

relative intensity. a Double staining. DAPI (blue), Alexa Fluor 488 (green), Alexa Fluor

594 (orange) and b Triple staining. Alexa Fluor 405 (blue), Alexa Fluor 488 (green),

Alexa Fluor 594 (orange). Alexa Fluor 405 is shown rather than the Dylight 405

fluorophore that was used in the experiment; however, the two have very similar

absorption and emission spectra. Spectra were obtained using the Fluorescence

SpectraViewer tool by ThermoFisher Scientific:

https://www.thermofisher.com/ca/en/home/life-science/cell-analysis/labeling-

chemistry/fluorescence-spectraviewer.html

https://www.thermofisher.com/ca/en/home/life-science/cell-analysis/labeling-chemistry/fluorescence-spectraviewer.htmlhttps://www.thermofisher.com/ca/en/home/life-science/cell-analysis/labeling-chemistry/fluorescence-spectraviewer.html

22

LSM700; Zeiss, Oberkochen, Germany) and quantified using Image J software,

version 1.48v (National Institutes of Health, Bethesda, MD). Images used within each

figure were captured with the same pinhole size and detector sensitivity settings, and with

short exposure times to avoid fading. For sections with DAPI staining, the blue channel

was always imaged last because of DAPI’s broad emission spectrum; thus minimizing

signal bleed into the green channel. Staining areas and cell counts were determined in one

coronal or sagittal section per animal, with four sample boxes for each region examined

(see schematic in Fig. 2.2). The hematoma boundary was readily identified by the

presence of blood. All peri-hematoma sample boxes were placed immediately adjacent to

the hematoma, with one side of each box aligned with the hematoma boundary. When

sampling the surrounding ipsilateral striatum, each box was located beyond the peri-

hematoma boxes, ventral to the corpus callosum. For the subventricular zone (SVZ),

boxes were selected immediately adjacent to the lateral ventricle starting in the dorsal-

most region and moving ventrally. For the contralateral region, boxes were randomly

placed in the center of the striatum. When cells were counted, all sample boxes were

320×320 μm for coronal sections. The boxes were slightly larger (400×400 μm) for

sagittal sections because of variations in white-matter tract diameter and density, and to

have more tracts per image.

2.5 Data analysis

Oligodendrocyte-lineage cells were quantified at 1, 3, 7, 14, and 28 days after

ICH. Cells were counted manually using ImageJ software by an observer who was

blinded to the time point and location (i.e., peri-hematoma, surrounding striatum,

contralateral striatum). For coronal sections, counts from four sampling boxes at each

23

Figure 2.2 Hematoma development in the first month after ICH and sampling

regions used. Representative cresyl violet-stained sections from rats on Day 1 after

saline injection (control), and on Days 1, 3, 7, 14 and 28 after injection of type IV

collagenase into the right striatum. For Day 7 after ICH, sample boxes are shown for the

peri-hematoma (red), surrounding ipsilateral striatum (blue), and contralateral striatum

(black). a Coronal sections. The hematoma appears pale, which is especially clear on the

right side of the sections on Days 1 and 3. On Days 7 and 14, the hematoma is smaller

and darkly stained. b Sagittal sections. Scale bars = 1 mm.

24

location and time were converted to densities (number of cells/area) and averaged, and

the mean ± SEM was calculated from three rats at each time point. To assess time-

dependent changes within each region, a one-way analysis of variance (ANOVA) was

followed by Tukey’s multiple-comparisons test. For sagittal sections, we first quantified

the number of cells inside versus outside white-matter tracts in the peri-hematoma region.

Then, the area represented by white-matter tracts was calculated by summing the areas of

each tract, outlined using the Polygon selection tool on ImageJ. The area outside the

bundles was calculated by subtracting the area of the white-matter tracts from the total

image area. Counts for oligodendrocyte-lineage cells were converted to densities (number

of cells/area) and averaged. Values are reported as mean ± SEM, from three or four rats

at each time point, and comparisons between inside versus outside white-matter tracts

were made with a two-way ANOVA followed by Tukey’s multiple-comparisons test.

Throughout, differences were considered statistically significant if p

25

Chapter 3. Results

Figures 3.3, 3.5c-d and 3.6c-d have been contributed by Jayalakshmi Caliaperumal, a

postdoctoral fellow in Dr. Schlichter’s lab.

3.1 Myelin is damaged but some myelinated axons survive in the peri-hematoma

after ICH

As expected, there was no hemorrhage in control, saline-injected rats; while a

hemorrhage developed when collagenase was injected into the striatum (Fig. 2.2). At 1

day, collagenase injection produced a large hematoma that was confined to the striatum,

and in both coronal (Fig. 2.2a) and sagittal (Fig. 2.2b) sections, cell damage was

apparent, as shown by reduced cresyl violet staining of Nissl bodies. While a hematoma

was evident throughout the 28-day period, the area of diminished cresyl violet staining

decreased with time and the blood clot was gradually resorbed. In addition, the lateral

ventricles expanded (ventriculomegaly), which is a common outcome after ICH and is

due to impaired drainage of cerebrospinal fluid (CSF) (Del Bigio, 1993) or tissue loss

(MacLellan et al., 2008). Others, using the same ICH model, have observed similar

changes to the size of the hematoma and ventricles for 28 days and longer (MacLellan et

al., 2008; Nguyen et al., 2008). The darker staining at 7 and 14 days is due to

chromogenic iron that is released from ruptured erythrocytes (Auriat et al., 2012;

Caliaperumal et al., 2012). In order to quantify changes in the spatial distribution of

oligodendrocyte-lineage cells, I used coronal sections and examined three regions: the

peri-hematoma, surrounding striatum, and contralateral striatum. Approximate sample

locations are illustrated by the boxes in Fig. 2.2. Then, sagittal sections were used to

quantify cells inside versus outside white-matter bundles in the peri-hematoma region.

26

The sample boxes were slightly larger to visualize more bundles and a larger proportion

of each bundle.

To examine the progression of damage to axons and myelin in a qualitative

manner, coronal (Fig. 3.1a) and sagittal (Fig. 3.1b) sections were used. Myelinated axon

tracts in the striatum arise from excitatory cortical neurons (Calabresi et al., 1996). They

were visualized using an antibody against MBP, and axons were labeled with a pan-NF

antibody that recognizes the three major neurofilament subunits: light, medium and heavy

chain. The coronal sections show approximately circular myelinated axon tracts

extending in the rostral-caudal axis (Fig. 3.1a). In the saline-injected (control) striatum,

bundles are filled with densely packed myelinated axons, as indicated by the abundant

overlap of the MBP and pan-NF stains (Fig. 3.1a). At high magnification, the two stains

do not precisely co-localize [also see (Lively and Schlichter, 2012b)] because myelin

wraps around each neurofilament-containing axon. In sagittal sections from saline-

injected animals, there were dense longitudinal tracts of myelinated axons throughout the

striatum (Fig. 3.1b). The sagittal orientation was especially useful for assessing damage

to white-matter bundles close to the hematoma.

Our lab previously showed a progressive, dramatic decrease in staining for normal

MBP and a concomitant increase in damaged MBP in the peri-hematoma region from 1

to 7 days after ICH (Wasserman and Schlichter, 2008; Moxon-Emre and Schlichter,

2011; Lively and Schlichter, 2012a). In coronal (Fig. 3.1a) and sagittal sections (Fig.

3.1b), despite the normal variation in diameter of striatal white-matter tracts, a similar

increase in myelin damage is clearly evident over the first week. There was progressive

bundle swelling and fragmentation. Interestingly, at 7 days, myelin and axon staining

27

28

29

Figure 3.1 Myelin damage and axon survival in the peri-hematoma after ICH.

Representative images of sections double-labeled with an antibody against

neurofilaments (pan-NF; green), and an antibody against myelin basic protein (MBP;

red). a Coronal sections from the region adjacent to the hematoma, which is to the right

but beyond the image (as indicated by →H). The peri-hematoma is visible in the right-

most portion of each image, and the surrounding striatum is visible toward the left.

Higher-magnification images of the boxed regions are shown to the right. b Sagittal

sections. All images were taken in the peri-hematoma, the region immediately adjacent to

the hematoma, which is below each portion shown (as indicated by H). Scale bars = 50

μm and apply to all panels.

were not well-organized into dense bundles, and the high-magnification coronal image

shows diffuse neurofilament staining inside fragmented bundles. Beyond the first week,

the bundles of myelinated axons appeared denser, less swollen, and less fragmented (i.e.,

day 14 was similar to day 3). The high-magnification coronal image at 28 days clearly

shows that the axons were more aligned into dense myelinated tracts. From the sagittal

images, it appears that some bundles were thinner at 14 and 28 days than in the saline

control. The improved appearance of the myelinated axon tracts after the first week, and

presence of both axons and myelin in the peri-hematoma region throughout the first

month after ICH, raises the prospect for re-myelination. Thus, I next determined if

oligodendrocyte-lineage cells survived and differentiated into mature cells that are

capable of myelin production.

3.2 The density of oligodendrocyte-lineage cells increases as a result of proliferation

in the peri-hematoma region

I quantified temporal changes in the density of oligodendrocyte-lineage cells in

the peri-hematoma region, surrounding striatum, and undamaged contralateral striatum

30

from 1 to 28 days. All oligodendrocyte-lineage cells, regardless of maturation state,

express the transcription factor, Olig2 (Zhou et al., 2000) and, as expected, I found that

Olig2 co-localized with the nuclear marker, DAPI (Fig. 3.2a). It was previously shown

that saline injection does not cause significant damage (DeBow et al., 2003; Wasserman

and Schlichter, 2007a; Wasserman et al., 2008; Caliaperumal et al., 2012). As expected,

saline-injected controls showed the same density of oligodendrocyte-lineage cells at 1, 3

and 7 days; hence, ICH samples were compared with the 1-day saline control. After ICH,

the density of Olig2+ cells increased from 1 to 7 days in the peri-hematoma region, at

which time they were ~11-fold higher than controls (Fig. 3.2b) and formed a dense band

around the hematoma (Fig. 3.3a; inset to right). Their numbers were markedly decreased

at 14 and 28 days. In the surrounding striatum further from the hematoma, and in the

contralateral striatum, the density of Olig2+ cells was elevated at 1, 3, 7 and 14 days, and

then decreased to the control level at 28 days.

Sagittal sections from the peri-hematoma region were examined to ask whether the

oligodendrocyte-lineage cells increased inside myelinated white-matter tracts, and in the

neuron-rich neuropil, which contains many cell bodies of medium spiny neurons that are

damaged after ICH (Wasserman and Schlichter, 2007b; 2008; Moxon-Emre and

Schlichter, 2011; Lively and Schlichter, 2012a). The bundles of myelinated axons were

swollen and MBP was degraded at 7 days, as previously shown. As shown in Figure 3.1,

many myelinated bundles had re-appeared by 14 and 28 days (Fig. 3.3a). Although

Olig2+ cells were present both inside and outside the bundles, they preferentially

accumulated inside the white-matter tracts (Fig. 3.2b). Again, their density peaked at 7

days, and then decreased.

31

Figure 3.2 Numbers of oligodendrocyte-lineage cells increase after ICH, especially

in the peri-hematoma. a Oligodendrocyte-lineage cells of any maturation state were

identified using an antibody against the nuclear antigen, Olig2 (red). Cell nuclei were

stained with DAPI (blue), and the enlarged inset shows an example of staining co-

localization. Representative images from 1 to 28 days after ICH show the contralateral

(undamaged) striatum, the peri-hematoma region and the striatum surrounding the peri-

hematoma. Scale bar = 50 μm (main panels), 10 μm (inset). Inset to right: Low

magnification image of oligodendrocyte-lineage cells (Olig2; red) surrounding the

hematoma at 7 days. Scale bar = 50 μm b Time-dependent changes in density of

oligodendrocyte-lineage cells were determined by counting double-labeled Olig2+DAPI

+

nuclei. Values represent mean ± SEM from 3 rats at each time point, and the dashed line

represents the mean for 1-day saline controls (n=3). Statistical comparisons were based

on a one-way ANOVA followed by Tukey’s post-hoc test. Comparisons show

differences: † from the preceding time point, # between ICH and 1 day saline controls; 1

symbol p

32

contralateral striatum, 28 vs 1, 3 and 7 days.

33

Figure 3.3 Oligodendrocyte-lineage cells preferentially increase inside white-matter

bundles. a Representative images from sagittal sections taken from the peri-hematoma

region, with myelin bundles labeled with MBP (green) and oligodendrocyte-lineage cells

of any maturation state labeled with Olig2 (red). MBP has been false coloured from blue

to green using ImageJ to better visualize the bundles. Scale bar = 50 m and applies to all

panels. b Density of Olig2+ cells inside or outside white-matter bundles in the peri-

hematoma region. Values are mean ± SEM from 4 rats at each time point after ICH. The

dashed lines indicate mean values for 1 day saline controls (n=3) for oligodendrocyte-

lineage cells inside (red) and outside bundles (black). Statistical comparisons were based

on a two-way ANOVA followed by Tukey’s post-hoc test. Comparisons show

differences: † from the preceding time point, # from the saline control, * between those

34

inside and outside the bundles; 1 symbol p

35

Figure 3.4 Oligodendrocyte-lineage cells proliferate in the damaged striatum,

particularly in the peri-hematoma region. a Representative images show

oligodendrocyte-lineage cells labeled for Olig2 (green), proliferating cells stained with an

antibody against Ki67 (red), and nuclei labeled with DAPI (blue). Images were taken

from the ipsilateral side in the peri-hematoma region (left panels), surrounding striatum

(middle) and SVZ (right). Scale bar = 50 μm and applies to all panels. b Higher-

magnification images of typical single, double- and triple-stained nuclei in each region at

3 days after ICH. c Density of proliferating oligodendrocyte-lineage cells (i.e., cells with

Olig2+Ki67

+ nuclei) in the peri-hematoma and surrounding striatum from 1 to 28 days

after ICH. d Percentage of oligodendrocyte-lineage cells that were proliferating. Values

are mean ± SEM from 3 rats at each time. Statistical comparisons were based on a one-

way ANOVA followed by Tukey’s post-hoc test and show differences † from the

36

preceding time point; 1 symbol p

37

38

Figure 3.5 Oligodendrocyte precursor cells (OPCs) increase preferentially in the

peri-hematoma and inside white-matter bundles. a OPCs were identified as cells with

Olig2+ nuclei (red) that were double-labeled with an antibody against NG2 chondroitin

sulfate proteoglycan (green). Representative images were taken from the contralateral

striatum, and the peri-hematoma region and the surrounding striatum on the ipsilateral

side. The higher-magnification inset shows a typical OPC, with an Olig2+ nucleus

surrounded by NG2. Scale bar = 50 μm (main panels), 10 μm (inset). b Time-dependent

changes in density of OPCs (Olig2+NG2

+ cells). Values represent mean ± SEM from 3

rats at each time point, and the dashed line represents the mean for 1-day saline controls

(n=3). Statistical comparisons were based on a one-way ANOVA followed by Tukey’s

post-hoc test. Comparisons show differences: † from the preceding time point, # from the

1 day saline control; 1 symbol p

39

striatum, 3 vs 14 and 28 days. c Representative images from sagittal sections taken from

the peri-hematoma region were triple-stained for Olig2 (red), NG2 (green), and myelin

basic protein (MBP; blue). Scale bar = 50 m and applies to all panels. d Density of

OPCs (Olig2+NG2

+ cells) inside or outside white-matter bundles in the peri-hematoma

region. Values are mean ± SEM from 4 rats at each time point after ICH. The dashed

lines indicate mean levels for 1day saline controls (n=3) for OPCs inside (red) and

outside the bundles (black). Statistical comparisons were based on a two-way ANOVA,

followed by Tukey’s post-hoc test. Comparisons show differences: † from the preceding

time point, # from the saline control, * between those inside and outside the bundles; 1

symbol p

40

41

Figure 3.6 Mature oligodendrocytes increase after ICH, especially inside white-

matter bundles in the peri-hematoma. a Mature oligodendrocytes were identified as

cells with Olig2+ nuclei (red) that were double-labeled with an antibody against

adenomatous polyposis coli (APC, also called CC-1; green). Representative images were

taken from the contralateral striatum, and the peri-hematoma and surrounding striatum on

the ipsilateral side. The higher-magnification inset shows a typical mature

oligodendrocyte with an Olig2+ nucleus surrounded by CC-1. Scale bar = 50 μm (main

panels), 10 μm (inset). b Time-dependent changes by density of mature oligodendrocytes

(Olig2+CC-1

+ cells). Values represent mean ± SEM from 3 rats at each time point, and

the dashed line represents the mean for 1-day saline controls (n=3). Statistical

comparisons were based on a one-way ANOVA followed by Tukey’s post-hoc test.

Comparisons show differences: † from the preceding time point, # from the 1 day saline

control; 1 symbol p

42

differences that are not shown with symbols are: (i) in the peri-hematoma, 7 vs 1 and 28

days; and (ii) in the surrounding striatum, 7 vs 1 and 28 days and 14 vs 1 and 3 days. c

Representative images from sagittal sections from the peri-hematoma region that were

triple-stained for Olig2 (red), CC-1 (green), and myelin basic protein (MBP; blue). Scale

bar = 50 m and applies to all panels. d Density of mature oligodendrocytes (Olig2+/CC-

1+) inside or outside white-matter bundles in the peri-hematoma region. Values are mean

± SEM from 4 rats at each time point. The dashed lines indicate mean values for 1 day

saline controls (n=3) for oligodendrocytes inside (red) and outside the bundles (black).

Statistical comparisons were based on a two-way ANOVA, followed by Tukey’s post-

hoc test. Comparisons show differences: † from the preceding time point, # from the

saline control, * between those inside and outside the bundles; 1 symbol p

43

Chapter 4. Discussion

4.1 Background/overview

White-matter damage often occurs after acute CNS injury (Matute and Ransom,

2012) and is considered an important predictor of the outcome, including after stroke

(Medana and Esiri, 2003; Lee et al., 2010). After ICH, significant white-matter damage

was seen in CT scans of 77% of patients (Smith et al., 2004) and severe white-matter

lesions were associated with large ICH volumes (Lou et al., 2010) and severe Glasgow

Coma Scale scores (Lee et al., 2010). MRI showed that white-matter hyper-intensity

volumes were higher after ICH than ischemic strokes (Rost et al., 2010). In most

experimental stroke studies that consider white matter, the focus has been on ischemia,

and on treatment targets ranging from reducing immediate damage to promoting

endogenous repair/regeneration to replacing cells using exogenous sources. In evaluating

approaches for repairing white matter, it is important to know the fate of axons and

oligodendrocytes but almost nothing is known about these aspects after ICH.

When I began this research, I came across very few studies that examined

oligodendrocytes after ICH using immunohistochemistry. The markers used in these

studies were CNPase (Jeong et al., 2003; Masuda et al., 2010; Lee et al., 2015) and MBP,

which are both expressed by mature oligodendrocytes and myelinated axons (Zhao et al.,

2009; Zhou et al., 2014). With these stains it is difficult to count cells in vivo, especially

within white-matter tracts, because the myelin is also stained. Thus, none of these studies

quantified the temporal or spatial distributions of oligodendrocyte-lineage cells after ICH,

and instead focused on a single time point or region. The most similar study to ours

showed that after experimental ICH in the mouse cortex, staining of mature

44

oligodendrocytes with RIP antibody (which recognizes CNPase) decreased at 7 days

(Masuda et al., 2010). The location with respect to the hematoma was not stated; and

temporal changes, proliferation and differentiation of oligodendrocyte-lineage cells were

not addressed. Two studies attempted cell replacement therapy using neural stem cells

(NSCs) (Jeong et al., 2003) and umbilical cord-derived mesenchymal stem cells (Liu et

al., 2010) from embryonic human tissue. The first study did not find any CNPase-stained

transplanted cells in the peri-hematoma (Jeong et al., 2003), and the second study simply

stated that mesenchymal stem cells have the potential to differentiate into

oligodendrocytes (Liu et al., 2010). Before treatments are pursued to replace

oligodendrocytes or enhance oligodendrogenesis after ICH, it is imperative to identify

their endogenous levels after injury. The present study appears to be the first to address

the fate of oligodendrocyte-lineage cells after ICH in the rat. I focused on the peri-

hematoma region and surrounding striatum because tissue is immediately damaged inside

the hematoma and is unlikely to be salvageable (Ziai, 2013). To that end, this work

provides the most detailed assessment of the time course and location of

oligodendrocytes outside the hematoma. I will discuss the findings in light of the limited

literature.

4.2 Salient findings

There are several salient findings. (i) Axons and oligodendrocytes survived for at

least a month in the peri-hematoma region and surrounding striatum and by 2 weeks,

there were signs that the myelinated corticostriatal axon tracts were recovering. (ii) The

density of both immature OPCs and mature oligodendrocytes increased dramatically in

the peri-hematoma region for the first week and then declined to about control levels.

45

Their source was likely local because considerable proliferation of oligodendrocyte-

lineage cells was seen in the peri-hematoma and surrounding striatum (peak at 3 days)

but not in the SVZ during the first month. (iii) Both immature OPCs and mature

oligodendrocytes were present throughout the first month. When their density peaked at

about 1 week, they were preferentially located inside white-matter bundles. Thus, the

endogenous cells that are required for re-myelination are present in the right place for the

first month after ICH.

4.3 Ischemic stroke versus ICH

4.3.1 Extent of myelin and axon damage and possible white matter recovery

We used coronal and sagittal sections to assess white-matter tracts that are

comprised of myelinated corticostriatal axons. In this model, our lab previously

quantified a progressive loss of normal myelin and accumulation of damaged myelin in

the first week (Wasserman and Schlichter, 2008; Moxon-Emre and Schlichter, 2011;

Lively and Schlichter, 2012b). Here, similar progressive myelin damage was seen.

However, a new finding was that both axons and oligodendrocytes survived in the peri-

hematoma region and surrounding striatum for at least 28 days; the longest time tested.

Initially, as white-matter damage progressed during the first week, the myelinated tracts

in the peri-hematoma region became swollen, which is consistent with previously

identified white-matter edema (Del Bigio et al., 1996). The bundles became increasingly

fragmented during the first week but still recognizable, and extensive, diffuse

neurofilament staining remained at 7 days. Our lab previously found that there is a

similar spatial distribution of myelin damage after ICH and transient ischemia in the

striatum; however, the damage occurred more rapidly after ICH. Complete loss of MBP

46

and a dramatic increase in dMBP took 1–3 days in the hematoma after ICH, versus 3–7

days in the infarct core after ischemia (Wasserman and Schlichter, 2008; Moxon-Emre

and Schlichter, 2010; 2011; Lively and Schlichter, 2012b).

In earlier ICH studies, our lab observed transient accumulation of APP and

SC1/hevin within axons in the peri-hematoma at 1 and 3 days (Wasserman and

Schlichter, 2008; Moxon-Emre and Schlichter, 2011; Lively and Schlichter, 2012b). At

the time, our lab noted that accumulation of APP indicates impaired axoplasmic transport

but does not necessarily mean the axons are irreversibly damaged. This axon impairment

was first seen at the hematoma edge and it progressed over time into the surrounding

striatum, with entire myelinated bundles staining for APP (Moxon-Emre and Schlichter,

2011). This is a different finding from transient ischemia in the striatum, where this axon

impairment was limited to the infarct core (Moxon-Emre and Schlichter, 2010). We had

not previously examined times beyond 7 days in either stroke model. We were surprised

to find that at 14 and 28 days after ICH, many discrete myelinated bundles of axons were

again present and appeared relatively normal. While it is possible that severely

fragmented tracts had been removed, there was no apparent decrease in the number of

bundles compared with the undamaged striatum. Instead, in sagittal sections, some

bundles appeared thinner (a characteristic of regenerated myelin (Gledhill and

McDonald, 1977)). This is consistent with an observation at 100 days after autologous

blood injection into the rat striatum (Felberg et al., 2002). An earlier striatal ICH study in

mice reported fewer corticostriatal axons at 7 days, but an increase at 9 weeks (Barratt et

al., 2014). Perhaps because myelin damage in the striatum progresses more rapidly after

ICH than ischemia, and axon damage is far more extensive, oligodendrogenesis might be

47

initiated faster after ICH to repair the damage.

The present finding that axon bundles were poorly organized at 7 days but well-

organized at 14 and 28 days, suggests that some recovery had occurred. While a large

battery of tests is needed to assess neurological outcomes after damage to the striatum

(MacLellan et al., 2009), several studies of collagenase-induced ICH in the rat show

varying degrees of spontaneous recovery. Over several weeks, ICH animals exhibited

recovery in several motor and cognitive tasks (MacLellan et al., 2006; Hartman et al.,

2009; MacLellan et al., 2009; Beray-Berthat et al., 2010) but still had some sensorimotor

deficits up to 8 weeks (Hartman et al., 2009; Beray-Berthat et al., 2010). Not surprisingly,

greater recovery is generally seen for small to medium-sized lesions (MacLellan et al.,

2006): the size range in the present study. In the most similar study of ICH in the

striatum, motor deficits had resolved at 30 days, as judged by the composite neurological

deficit scale, which assesses spontaneous ipsilateral circling, hind limb retraction,

bilateral forepaw grasp, beam walking ability, and forelimb flexion (MacLellan et al.,

2009).

4.3.2 Oligodendrocyte-lineage cells

The main emphasis of this study was on the fate of oligodendrocyte-lineage cells.

In the peri-hematoma region, regardless of their maturation state, their numbers increased

substantially from 1 to 7 days, and they were preferentially increased inside white-matter

bundles. They were also increased in the surrounding striatum, though the number inside

or outside white-matter bundles wasn’t quantified in this region. In both the peri-

hematoma and surrounding striatum, the density of oligodendrocyte-lineage cells then

declined at 14 and 28 days; however, it is important to note that their numbers did not

48

decline below control levels at any time examined. In contrast, few oligodendrocyte-

lineage cells were found inside the hematoma.

Several early studies of transient ischemia that highlighted the importance of

studying oligodendrocytes have shown double-labeling with Tau-1, a microtubule

associated protein whose accumulation indicates cell injury (Dewar and Dawson, 1995;

Pantoni et al., 1996; Irving et al., 1997; Gresle et al., 2006). More recently, studies have

examined the fate of oligodendrocyte-lineage cells after transient ischemia. In one such

study, a 2 h rat middle cerebral artery occlusion (MCAO) model was used (Jiang et al.,

2011), which created a large lesion in the striatum and overlying cortex.

Oligodendrocyte-lineage cells in the striatum peri-infarct region decreased slightly at 1

day and increased at both 7 and 14 days. This contrasts with striatal ICH, where I did not

observe a reduction in oligodendrocyte-lineage cells in the peri-hematoma, and found that

the increase peaked at 7 days and then declined at 14 days.

By 7 days after ICH, the oligodendrocyte-lineage cells formed a ring around the

hematoma. This ring was similar to the previously identified peri-hematoma ring of

activated microglia/macrophages (Moxon-Emre and Schlichter, 2011); in which case, the

immune cells were preferentially inside white-matter bundles that labeled for dMBP. I

have not come across any studies describing a similar finding after ischemia. This raises

important questions for future study. Are activated microglia/macrophages phagocytosing

myelin debris to promote re-myelination? Do they initially support the survival and

proliferation of oligodendrocyte-lineage cells, and then at later times, kill and remove

excess cells? Another possibility is that the decline in oligodendrocyte-lineage cells at

later times results from toxic effects of blood components. In vitro, blood proteins can

49

suppress migration of OPCs and thrombin can suppress OPC differentiation (Juliet et al.,

2009). Iron liberated from ruptured erythrocytes causes oxidative stress (Wu et al., 2011;

Auriat et al., 2012), and iron remains elevated for at least one month in the collagenase

ICH model (Auriat et al., 2012; Caliaperumal and Colbourne, 2014).

4.3.3 OPC proliferation

In principle, the increase in density of oligodendrocyte-lineage cells in the peri-

hematoma and surrounding striatum after ICH could be due to proliferation, migration

from other areas or both. In the present study, the striatum contained many proliferating

oligodendrocyte-lineage cells but very few were seen in the SVZ. In rodent models of

transient ischemia, NG2+ OPCs were also prevalent in the peri-infarct region, increasing

at 3 and 7 days (Ohta et al., 2003; Tanaka et al., 2003; Jiang et al., 2011). (Sozmen et al.,

2009) showed increased OPC proliferation at 7 days by double staining NG2 and BrdU,

followed by a decline at 28 days.

Ki67 and BrdU have both been widely used as markers of cell proliferation (Kee

et al., 2002) and can be used together, as in a recent article using the intraspinal

hemorrhage model (Sahinkaya et al., 2014). In that study, rats received intraperitoneal

injections of BrdU at varying intervals after hemorrhage. This allowed the investigators

to examine the number of OPCs that had proliferated during those time intervals (using

BrdU) and track their maturation (by double labeling with CC-1), as well as those cells

actively proliferating at the time of sacrifice (using Ki67). For the present study, only

Ki67 was used and the results indicate the number of actively proliferating OPCs present

at the time the animal was sacrificed. In future, a BrdU staining paradigm should be used

with the ICH model to estimate how soon newly proliferating oligodendrocytes mature

50

after ICH. However, BrdU can also be incorporated during DNA repair (Zheng et al.,

2011) and might not be a true reflection of oligodendrogenesis.

The SVZ is considered a source of neural progenitor cells that migrate out and can

differentiate into oligodendrocytes. However, previous ICH studies have investigated the

SVZ only from a neurogenesis perspective (Masuda et al., 2007; Yang et al., 2008; Yan

et al., 2009; Otero et al., 2012; Xu et al., 2012; Kang et al., 2014). The site of

oligodendrogenesis might also depend on the disease model and where the damage is

located. For instance, in Huntington’s disease, oligodendrocyte-lineage cells proliferated

within the striatum (McCollum et al., 2013) but not in the SVZ (Lazic et al., 2006; Kohl

et al., 2010; Simpson et al., 2011; McCollum et al., 2013). After de-myelination of the

mouse corpus callosum, OPC proliferation was seen within the SVZ (Menn et al., 2006;

Xing et al., 2014). After transient ischemia, OPC proliferation was not examined in the

SVZ, but fate-mapping showed some SVZ-derived oligodendrocytes in the striatum (Li et

al., 2010; Zhang et al., 2011). The striatum might also represent a key area for

oligodendrogenesis, as there was a larger increase in Olig2+ cells in the striatal peri-

infarct compared to the cortical peri-infarct after MCAo (Jiang et al., 2011). Furthermore,

oligodendrogenesis might depend on the extent of damage. In the present study, a small

to medium-sized hematoma was induced in the striatum, with a significant amount of

surrounding striatal tissue remaining. In an intraspinal hemorrhage model, it was

suggested that the greater the amount of spared tissue, the more OPCs might proliferate

early after injury (Sahinkaya et al., 2014).

The oligodendrocyte density also increased in the contralateral striatum from 1 to

14 days but I did not detect proliferating OPCs at any time in the undamaged site. In

51

future, with optical tagging or genetic markers, it might be possible to determine whether

their source in the contralateral striatum is from within the striatum or from the corpus

callosum, SVZ or another location. I could not find previous publications addressing

oligodendrocyte responses in the contralateral side after ICH. However, there are

numerous indications that the contralateral hemisphere does respond after ICH; i.e.,

changes in gene expression (Wasserman et al., 2007; Lively and Schlichter, 2012a),

enhanced dendritic arborization (Nguyen et al., 2008), loss of corticostriatal axons

(Barratt et al., 2014), and increased axial diffusivity (Fan et al., 2013).

There is increasing evidence that growth factors are important signalling

molecules for CNS re-myelination, much as they are for myelination during development

(reviewed in (Hinks and Franklin, 2000)). Inhibiting growth factors pharmacologically,

reduced re-myelination (McKay et al., 1997; 1998) while re-myelination was enhanced

by delivery of exogenous growth factors (Yao et al., 1995; Cannella et al., 1998;

McTigue et al., 1998). The substantial increase in OPC proliferation suggests that ICH

up-regulates some growth factors in the peri-hematoma and surrounding striatum. There

are several good candidates. Insulin-like growth factor-1 (IGF-1) stimulates OPC

proliferation in vitro (McMorris and Dubois-Dalcq, 1988). Platelet-derived growth factor

(PDGF) activates PDGFRα on OPCs and increases their proliferation and survival in

vitro (Noble et al., 1988; Barres et al., 1992; Grinspan and Franceschini, 1995) and in

organotypic slices (Hill et al., 2013). Fibroblast growth factor-2 (FGF-2) stimulates OPC

proliferation (Grinspan et al., 1993) and up-regulates their expression of PDGFRα, which

prevents them from maturing, and thus, prolongs their responses to PDGF (McKinnon et

al., 1990). In future, it would be interesting to examine the spatial and temporal

52

expression of these potential mitogens after ICH. There appears to be very limited

information about their expression after ICH, with data only from two microarray studies.

Within 24 h after ICH, FGF-2 was ~1.5-fold lower in the peri-hematoma region of 4 ICH

patients (Rosell et al., 2011), and the FGF-2 receptor, FGF2rb, was ~2 fold lower in the

striatum and cortex in the rat autologous blood-injection model (Lu et al., 2006).

Interestingly, FGF-2 supplementation reduced edema and improved some motor

functions in the collagenase- and autologous-blood injection ICH models in mice (Huang

et al., 2012). In a spinal cord demyelination