quantitative analysis of capillary networks in the human ... · quantitative analysis of capillary...

QUANTITATIVE ANALYSIS OF

CAPILLARY NETWORKS

IN THE HUMAN RETINA

Dr Geoffrey Zhi Peng Chan MBBS(Hons)

This Thesis is presented for the degree of Master of Medical Science

of The University of Western Australia

Centre for Ophthalmology and Visual Science,

Lions Eye Institute, Perth, Australia

2015

Acknowledgements

I would like to express my sincerest gratitude to my Supervisors, Professor Dao-Yi

Yu and Professor William H Morgan for their support and guidance in all steps of this

Thesis. They have taught me just how little we know of the human body and,

importantly, opened my eyes to another aspect of medicine through their selfless

pursuit of knowledge. Thank you to Dr Paula Yu who performed histologic staining

on many of the specimens used for this research.

I am grateful to all the lovely people in my research group including Priscilla, Hong

Fang and Min for sharing in the laughter and dedication during our countless

weekends and hours spent at the lab. A special thank you to Chandrakumar

Balaratnasingam, my dear friend and mentor who showed me through action that a

tall tree can always grow taller, and that nothing can take the place of hard work and

perseverance no matter what the pursuit in life.

To my mother Oi Yin, father Yat Ming, sister Sharon, Singh family and my beloved

Trish: thank you for supporting me through thick and thin. It is only because of you

that I have been so privileged to be where I am today through your hours of support

and counsel.

In addition, thank you to Professor Stephen Cringle, Kathryn Morgan, Dean Darcey

and Marinko and team who have offered invaluable assistance. The majority of this

work has been peer reviewed and published in Investigative Ophthalmology and

Visual Science. Grant support was provided by the National Health and Medical

Research Council of Australia and the Australian Research Council Centre of

Excellence in Vision Science.

Table of Contents

Abstract ......................................................................................................................... 10

Chapter 1.0: Introduction ........................................................................................... 13

1.1 Background ..................................................................................................... 13

1.2 The Retina.......................................................................................................... 17

1.2.1 Retinal Histology............................................................................................... 17

1.2.2 Retinal Coupling Physiology: The Microenvironment ..................................... 19

1.2.3 Retinal Capillary Morphometry ........................................................................ 20

1.2.4 Previous Studies of the Retinal Microvasculature ............................................ 21

1.2.5 Current Understanding of Retinal Microvasculature ........................................ 22

1.2.6 Current Understanding of Oxygen Demand in the Retina ............................... 23

1.2.7 The Retina as a Surrogate Measure of Vascular Health .................................... 26

1.2.8 The Perifoveal Region ....................................................................................... 28

1.3 Optical Vascular Detection Methods ................................................................... 29

1.3.1 Fundus Fluorescein Angiography ..................................................................... 30

1.3.2 Microperfusion Fixation Staining and Confocal Microscopy .......................... 32

1.3.3 Scanning Laser Ophthalmoscopy ...................................................................... 34

1.3.4 Adaptive Optics ................................................................................................. 34

1.3.5 Optical Coherence Tomography ....................................................................... 35

Chapter 2.0: Quantitative Morphometry of Perifoveal Capillary Networks in the Human Retina .............................................................................................................. 37

2.1 Introduction ..................................................................................................... 37

2.2 Methods ........................................................................................................... 37

2.2.1 Human Donor Eyes ....................................................................................... 37

2.2.2 Tissue Preparation ......................................................................................... 39

2.2.3 Immunolabelling ............................................................................................ 39

2.2.4 Microscopy .................................................................................................... 41

2.2.5 Image Preparation .......................................................................................... 43

2.2.6 Qualitative Differentiation of Capillary Networks ........................................ 43

2.2.7 Morphometric Quantification of Capillary Networks ................................... 44

2.2.8 Statistical Analysis ........................................................................................ 44

2.2.9 Inter-observer Correlation and Measurement Reproducibility ...................... 46

2.3 Results ............................................................................................................. 46

2.3.1 Eye donors ..................................................................................................... 46

2.3.2 General ........................................................................................................... 46

2.3.3 Qualitative study of retinal capillary networks .............................................. 48

2.3.4 Quantitative Characteristics of the Nerve Fibre Layer Network ................... 48

2.3.5 Quantitative Characteristics of the Retinal Ganglion Cell and Superficial Inner Plexiform Layer network ........................................................................................ 49

2.3.6 Quantitative Characteristics of the Deep Inner Plexiform and Superficial Inner Nuclear Layer network ........................................................................................... 49

2.3.7 Quantitative Characteristics of the Deep Inner Nuclear Layer Network ....... 49

2.3.8 Morphometric Comparisons Between Networks .......................................... 55

2.3.9 Inter-observer Correlation and Measurement Reproducibility ...................... 55

2.4 Discussion ....................................................................................................... 56

Chapter 3.0: Quantitative Changes in Perifoveal capillary networks in patients with vascular comorbidities ................................................................................................. 60

3.1 Introduction ..................................................................................................... 60

3.2 Methods ........................................................................................................... 60

3.2.1 Human Donor Eyes ....................................................................................... 60

3.2.2 Tissue Preparation ......................................................................................... 61

3.2.3 Microscopy .................................................................................................... 63




3.2.7 Statistical Analysis ........................................................................................ 65

3.3 Results ............................................................................................................. 66

3.3.1 Eye Donors .................................................................................................... 66

3.3.2 General ........................................................................................................... 66

3.3.3 Morphometric Characteristics of Capillary Networks in Disease Eyes ........ 67

3.3.4 Comparisons between Capillary Networks in Disease Eyes ......................... 71

3.3.5 Comparisons between Disease and Control Eyes .......................................... 72

3.4 Discussion ....................................................................................................... 77

Chapter 4.0: Validating the utility of speckle variance optical coherence tomography for quantitatively analysing human perifoveal capillary networks............................... 81

4.1 Introduction ..................................................................................................... 81

4.2 Methods ........................................................................................................... 82

4.2.1 Live Human Subjects ..................................................................................... 82

4.2.2 Human Donor Eyes ...................................................................................... 82

4.2.3 svOCT Imaging technique ............................................................................ 82

4.2.4 Tissue Preparation for Confocal Microscopy ................................................ 87

4.2.5 Confocal Microscopy techniques .................................................................. 87




4.2.9 Statistical Analysis ....................................................................................... 89

4.3 Results ............................................................................................................. 91

4.3.1 Demographics ................................................................................................ 91

4.3.2 Morphometric Comparisons Between svOCT and Confocal Eyes .............. 91

4.3.3 Quantitative Analysis of Capillary Diameter ................................................ 94

4.3.4 Quantitative Analysis of Capillary Density ................................................... 96

4.4 Discussion ....................................................................................................... 97

Chapter 5.0: General Discussion .............................................................................. 100

REFERENCES ........................................................................................................... 105

APPENDIX ................................................................................................................. 122

Table of Figures

Figure 1: The Human Retina.................................................................................................... 18

Figure 2: Oxygen Demand in Retinal Subcompartments ........................................................ 25

Figure 3: Fluorescein Angiogram Image ................................................................................. 31

Figure 4: Confocal microscope ................................................................................................ 33

Figure 5: Microcannulation of the Central Retinal Artery....................................................... 40

Figure 6: Human perifovea ...................................................................................................... 42

Figure 7: Methodology for Quantification of Capillary Networks .......................................... 45

Figure 8: Co-localization of Capillary Networks within Retinal Layers ................................. 47

Figure 9: Nerve Fibre layer Capillary Network ....................................................................... 51

Figure 10: Ganglion Cell layer/superficial Inner Nuclear layer Network ............................... 52

Figure 11: Deep Inner Plexiform layer/superficial Inner Nuclear layer Network ................... 53

Figure 12: Deep Inner Nuclear layer Network ........................................................................ 54

Figure 13: Perifoveal Capillary Networks in an eye with Vascular Comorbidities ................ 69

Figure 14: Quantitative Comparisons of Capillary Network morphometry. ........................... 74

Figure 15: Morphometric Comparisons between control and disease eyes ............................. 75

Figure 16: Comparison of Three-dimensional Capillary Network Morphometry between

control and disease eyes ........................................................................................................... 76

Figure 17: Speckle Variance Optical Coherence Tomography derived Cross Section of the

Human Perifovea ..................................................................................................................... 85

Figure 18: Generation of Speckle Variance Optical Coherence Tomography Images ............ 86

Figure 19: Methodology for Quantifying Capillary Network Morphometry .......................... 90

Figure 20: Morphometric comparisons between svOCT and Histology derived Images ........ 92

Figure 21: Quantitative Comparisons of Capillary Network Morphometry ............................ 93

Abbreviations

AO Adaptive Optics

bfGF Basic fibroblast growth factor

FA Fluorescein Angiography

HGF Hepatocyte growth factor

IGF-1 Insulin-like growth factor 1

INL Inner Nuclear Layer

IPL Inner Plexiform Layer

NFL Nerve Fibre Layer

OCT

ONL

OPL

Optical Coherence Tomography

Outer Nuclear Layer

Outer Plexiform Layer

PDGF Platelet derived growth factor

PGF-1 Placental growth factor

RGCL

RPE

Retinal Ganglion Cell Layer

Retinal Pigment Epithelium

SLO Scanning Laser Ophthalmoscopy

svOCT Speckle Variance Optical Coherence Tomography

TGFβ Transforming growth factor beta

VEGF Vascular endothelial growth factor

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan

Abstract

Purpose: To improve our understanding of the human capillary microvasculature by

determining correlations between the location of neuronal subcellular compartments

and the morphometric features of regional capillary networks in the layered retina. To

detect if patients with cardiovascular comorbidities but no clinically detectable retinal

disease demonstrate quantitative alterations to capillary networks. To determine

whether speckle variance optical coherence tomography can be used as a clinical

diagnostic tool by accurately providing qualitative and quantitative information

regarding capillary networks in the human retina.

Methods: The first experiment involved studying the retinal capillary networks in 17

healthy human donor eyes. A specific location, the perifoveal region located 2 mm

nasal to the fovea was studied. Micropipette technology was used to cannulate the

central retinal artery and label the retinal microcirculation using a phalloidin perfusate.

Confocal scanning laser microscopy was used for capillary imaging. Antibodies

including gamma-synuclein, goalpha and parvalbumin antibodies were also used to co-

localize the nerve fibre layer (NFL), retinal ganglion cell layer (RGCL), inner

plexiform layer (IPL) and inner nuclear layer (INL) to capillary networks. Capillary

diameter, capillary loop area and capillary density were measured between networks.

For the second experiment, comparisons were made between 10 eyes from patients

with systemic vascular comorbidities and the 17 control eyes utilized in the first

experiment. All eyes were absent of clinically evident ocular disease. Identical micro-

cannulation techniques were used to label the retinal microvasculature and the same

perifoveal region was studied. Two- and three-dimensional image reconstructions were

used to perform quantitative measurements of individual capillary networks within the

perifovea. Parameters measured included capillary diameter, capillary loop area,

capillary loop length, capillary density and capillary surface area.

For experiment three, imaging comparisons were made between 9 live human subjects

and 8 human donor eyes. A prototype speckle variance optical coherence tomography

(svOCT) system was used to generate en face retinal capillary networks. Capillary

networks corresponding to the NFL, RGCL, IPL and INL were identified based on

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan morphometric features. Confocal scanning laser microscopy was used for capillary

imaging of histology specimens. Capillary diameter and density measurements were

compared between imaging techniques and within svOCT derived networks in the

perifovea.

Results: In the first experiment studying perifoveal capillary networks of healthy

human donors, it was found that capillary networks in the human perifovea are

morphometrically heterogeneous. Four capillary networks were identified in the retinal

layers corresponding to: (1) NFL (2) RGCL and superficial portion of IPL (3) Deep

portion of IPL and superficial portion of INL (4) Deep portion of INL. Laminar

configurations were present in NFL and deep INL networks. Remaining networks

demonstrated three-dimensional configurations. Capillary density was greatest in the

networks serving the IPL (P < 0.015). Capillary loop area was smallest in the two

innermost networks (P < 0.003). There was no difference in capillary diameter between

networks (P = 0.715).

In the second experiment comparing capillary networks in patients with vascular

comorbidities to healthy eyes, capillary diameter was found to be increased in both the

RGC and sIPL capillary networks (P = 0.005). Capillary loop area and capillary loop

length was reduced in all capillary networks in patients with vascular comorbidities (all

P < 0.020) except the deep capillary network of the INL (P = 0.863). Capillary density

was reduced in the NFL capillary network in patients with vascular comorbidities (P =

0.033). There was no difference in the relative occupied capillary surface area between

control and diseased eyes (all P > 0.061).

In the third experiment comparing imaging techniques, svOCT imaging of the retina

was demonstrated to accurately delineate all orders of retinal microvasculature.

Capillary networks were noted to be morphometrically similar in images derived from

svOCT and histological techniques. Capillary diameter was uniformly increased in all

svOCT derived capillary network measurements (all P < 0.001). There was no

difference in capillary density between svOCT and histology techniques (P > 0.094).

Conclusions: The experiments outlined in this thesis provides insight into the capillary

networks of the human perifovea and changes which may occur in the context of

disease. It highlights how morphometric features of regional capillary networks likely

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan serve a critical role in supporting neuronal homeostasis and that characteristics of

perifoveal capillary networks are non-uniformly altered in patients with vascular

comorbidities. The results of this study also suggest that svOCT imaging allows

accurate morphometric assessment of individual networks in the perifovea and that this

technique may allow the early detection and understanding of vascular-mechanisms in

retinal ischaemic diseases.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Chapter 1.0: Introduction

1.1 Background

Retinal vascular diseases are a major cause of visual morbidity worldwide yet there

remains only limited knowledge regarding quantitative aspects of human retinal

capillary micro-anatomy and structure.1-3 Such quantitative information would be

pertinent in predicting capillary perfusion as a function of the human cardiovascular

system. Histological studies involving non-human primates have revealed that retinal

capillaries demonstrate a multi-laminar morphology with a sheet like superficial

capillary layer and separate, elaborate three-dimensional deep capillary networks.4, 5

However, the human and primate retinal capillary structure are fundamentally distinct

and it is for this reason that the sequence of pathogenic changes in retinal vascular

diseases, such as diabetic retinopathy, are vastly different between the two species.6, 7

In order to improve our understanding of cellular mechanisms underlying human retinal

vascular diseases it is imperative that we have a thorough documentation of normal

human capillary micro-architecture.

The capacity of the retina to partition light stimuli into a series of complex visual

signals, prior to cortical transmission, is largely attributed to the layered organization of

its neuronal populations.8 Each neuronal population involved in visual processing has

distinct metabolic demands with significant disparity in the rate of oxygen consumption

between retinal layers.9, 10 It is expected that the distribution of capillary networks

within the retina correlate with the metabolic demands of soma, dendrites and

synapses– neuronal layers with the greatest energy demand may have the greatest

capillary supply.11 Few studies however have aimed to validate this hypothesis.

Unlike other regions in the central nervous system, the retinal circulation must achieve

neuronal nutrition without compromising the optical properties of the pathway

transmitting light to the photoreceptor layer in the outer retina.12 Increases in neuronal

energy demand cannot therefore be compensated by a simple increase in capillary

number. It is likely that retinal capillary networks are morphometrically adapted in

order that the balance between cellular nutrition and optical clarity can be achieved. In

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan other organs, microcirculatory adaptive mechanisms that serve to satisfy the unique

metabolic demands of distinct cell populations include modifications in capillary

diameter,13-21 inter-capillary distance,13 capillary density13, 18, 20, 21 and the area

bounded by capillary loops.22 Similar capillary network adaptations may be present in

the human retina.

Capillary endothelia facilitate the exchange of nutrients and toxic wastes between

neurons and supporting glia.23-25 By doing so they perform a critical role in

modulating retinal homeostasis. Cardiovascular comorbidities have the potential to

alter the blood-retina barrier and are therefore considered important risk factors for

retinal disease.26-28 However, it is unknown if the quantitative characteristics of retinal

capillary networks are altered before disease-induced alterations to the blood-retina

barrier become clinically manifest. Such knowledge is important as it may permit the

application of emerging state-of-the-art technology to image capillary networks and

thus detect early retinal disease.29 Furthermore, because break down of the blood-

retinal barrier results in vasogenic oedema and neural tissue damage with resultant loss

of vision,30 the clinical application of such histo-pathological knowledge could be of

major benefit in the prevention of vision loss. Understanding capillary network

changes in patients with cardiovascular comorbidities may provide valuable insights

into pathogenic mechanisms that underlie retinal vascular diseases.

The experiments outlined in this thesis attempt to document qualitative and quantitative

aspects of human capillary micro-architecture in the perifoveal region. This work will

significantly improve our understanding of normal human retinal anatomy and provide

the basis for understanding important sight-threatening diseases such as diabetic

retinopathy and age related macular degeneration. In particular, changes in the

presence of cardiovascular comorbidities, which may precede the development of

clinically manifest ocular disease, are studied. Comparisons between our histological

techniques and speckle variance optical coherence tomography (svOCT) as a non

invasive imaging technique are also examined closely.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Chapter 1 provides background information about the perifoveal region of the retina

and current understandings of capillary microvascular anatomy. Previous investigations

that have demonstrated important relationships between retinal vascular characteristics

and cardiovascular morbidity are discussed. The principles of modern retinal vascular

detection methods are explained with an emphasis on microperfusion fixation

techniques and optical coherence tomography.

Chapter 2 outlines an experiment in which perfusion- based labeling techniques,

utilizing micropipette cannulation of the central retinal artery and targeted antibodies,

are used to accurately label the human retinal circulation. Immunolabelling and

microscopy techniques are coupled with this perfusion methodology to perform

detailed analyses of capillary network relationships in the human perifovea.

Chapter 3 outlines an experiment in which the same perfusion-based labelling

techniques are employed with two- and three-dimensional image analysis techniques to

document peri-foveal capillary network changes in patients with a history of

cardiovascular morbidity but no clinically detectable retinal disease. We also

determine if capillary networks are selectively or uniformly altered by cardiovascular

morbidity.

Chapter 4 outlines an in-vivo experiment in which a modern svOCT technique is used

to image the perifoveal capillaries. Qualitative and quantitative aspects of human

capillary microvasculature derived from the svOCT technique are compared to

histological measurements.

Chapter 5 is a general discussion of the conclusions of this research and how findings

may allow clinico-histological correlation with modern imaging devices to aid in the

early detection of retinal vascular disease.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 1.2 The Retina

1.2.1 Retinal Histology

The vertebrate neuro-retina is a light sensitive layer of tissue, lining the inner surface of

the eye with an average thickness of only 300 µm.31 It consists of 10 distinct layers

running from inner (closest to vitreous body) to outer retina (Figure 1) which allows

transmission of photoreceptive information towards the optic nerve32 :

1. Inner limiting membrane- boundary between retina and vitreous formed by basement membrane of Muller cells and supporting astrocytes

2. Nerve Fibre layer - formed by retinal ganglion cell axons to form optic nerve

3. Ganglion Cell layer- cell body or soma of retinal ganglion cells and displaced amacrine cells

4. Inner plexiform layer- synapse between bipolar cell axons and retinal ganglion cell dendrites. Laterally connecting amacrine cell input is present.

5. Inner nuclear layer- cell body of bipolar cells, amacrine, horizontal and displaced ganglion cells

6. Outer plexiform layer- synapse between rod spherules, cone pedicules with bipolar and horizontal cell dendrites

7. Outer nuclear layer- cell bodies of rod and cone cells

8. External limiting membrane- junctional complexes (zonula adherens) between rods, cones and Muller cells

9. Photoreceptor layer inner segments

10. Photoreceptor layer outer segments- with (9), collectively contain rod and cones which represent the light-sensitive elements of the retina

In addition, retinal health is supported by glial cells and retinal pigment epithelium- a nonsensory layer of cuboidal cells which extend between the optic disk and ora serrata.


Figure 1: The Human Retina

Hematoxylin & Eosin stain cross section of the human retina demonstrating the

histological layers. ILM (Inner limiting membrane). RNFL (Retinal nerve fiber layer).

IPL (Inner plexiform layer). INL (Inner nuclear layer). OPL (Outer plexiform layer).

ONL (Outer nuclear layer). RPE (Retinal pigment epithelium). The choroid and sclera

are also visible. (Adapted from Deltagen Histology Atlas, 2006. Available from:

http://www.deltagen.com/target/histologyatlas/HistologyAtlas.html)

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 1.2.2 Retinal Coupling Physiology: The Microenvironment

Capillary- neuronal relationships have been the subject of extensive investigation

within the human central nervous system.24, 25, 33, 34 A particular pattern that is

recognized is that blood supply and degree of capillarization is adjusted to the

metabolic demands of regional tissue.25, 35-37 Half the intercapillary distance represents

the maximal distance which oxygen must travel by diffusion to supply tissues. In

selected regions of rabbit brains, pattern of distribution of blood flow were closely

correlated with that of capillary density along with a positive correlation to cytochrome

oxidase activity.21, 38 Close correlations have been shown among blood flow, glucose

utilization and capillary density in the brain in adult primates.39 In squirrel monkey

retina, the nerve fiber layer was noted to be contributing more to total vascularity than

would be predicted based on tissue thickness or volume measurements.4

Similarly, the primate retina is expected to exhibit variation in vascularity across

neuronal layers that reflect local variations in metabolic demand. The study of these

patterns in humans could therefore be of interest in establishing the metabolic

requirements of certain neuronal layers and modifications in both physiological and

pathological states. The study of local variations in blood supply and quantitative

changes in the microarchitecture may also help to explain differential vulnerability of

certain neurons in different disorders and have significant implication for physiological

studies of neural function.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 1.2.3 Retinal Capillary Morphometry

Retinal processing of the spatial and temporal properties of the visual stimulus results

in a remarkably intricate axonal signal that demands a disproportionately high volume

of cortical representation, relative to organ size, for sensory interpretation.40 It is

estimated that 90% of all sensory information integrated by the brain are of visual

origin and approximately one third of the brain’s volume is devoted to visual

processing. 41, 42 A complex system of microvascular networks are required to satisfy

the energy demands necessary for pre-cortical retinal processing.43 To date, there

remains a paucity of information concerning the structure and location of capillary

networks that support human retinal function. Such information may be critical for

understanding vascular mechanisms involved in retinal homeostasis and disease.

The capacity of the retina to partition the visual stimulus into a series of complex

psychophysical signals, prior to cortical transmission, is largely attributed to the layered

organisation of neuronal populations.44 Each neuron involved in visual processing has

distinct metabolic demands which translates to a significant disparity in the rate of

oxygen consumption between retinal layers. There is overwhelming evidence to

suggest that the organization of the microvasculature plays a critical role in satisfying

the energy demands of neuronal elements, however it remains unclear if retinal

capillary morphometry is regionally adapted as a consequence of intra-retinal variation

in neuronal function.4, 5

The morphometric organization of the retinal circulation is constrained by the optical

properties of the eye. Unlike other regions in the central nervous system, the retinal

circulation must achieve neuronal nutrition without compromising the optical

properties of the pathway transmitting light to the photoreceptor layer in the outer

retina - an increase in neuronal energy demand cannot therefore be compensated by a

simple increase in capillary number. Several studies, involving non-ocular

microvascular systems, have demonstrated important correlations between regional

homeostatic activity and capillary network morphometry. Previously reported

microcirculatory adaptive mechanisms that serve to satisfy the unique metabolic

demands of varying cell populations include modifications in capillary diameter13-20, 45 ,

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan inter-capillary distance13 , capillary density14, 18, 20, 21, 45, 46 and the area bounded by

capillary loops22 . It is possible that similar vascular adaptations maintain the balance

between cellular nutrition and optical clarity in the human retina.

1.2.4 Previous Studies of the Retinal Microvasculature

The microcirculation of the retina has been a subject of extensive investigation owing

to its importance in retinal health and disease.47 However, due to limitations in

imaging techniques, the precise connections between microvascular layers of the

human retina remains unknown.48 Given that the choroid, under normal atmospheric

conditions is only able to supply a small proportion of the oxygen needs of the inner

retina, it was postulated that an intricate capillary network provided by the retinal

circulation would be needed in order to support this region.49, 50

Using a modified benzidine-staining method for delineating vessels, Michaelson's

classic 1954 anatomical dissertation titled 'Retinal circulation in man and animals'

marked the beginning of a period of rapid expansion in our knowledge about the retinal

circulation.51 In this work, an emphasis was placed on a laminar distribution of

capillaries in the retina. Pioneering studies subsequently performed by Snodderly in

squirrel monkey retinas involved the study of vascular patterns in whole mount

preparations which identified regional variations in the vascular network with

topography of the fovea.4, 5 However, structural and functional differences between

experimental animals and humans were noted to exist and comprehensive studies of

human retinal microvasculature were hindered by the difficulty of obtaining and

preparing human tissue.52 Human retinas undergo rapid postmortem deterioration with

changes including swelling, distortion and fragmentation.53 Additionally, lab methods

were restricted to studying the architecture through techniques such as trypsin digest

which are noted to alter the three dimensional arrangement of tissue.54 Colloidal iron

and vascular corrosion casts were also employed but are subject to pressure artefact and

do not allow for optical sectioning.55, 56

More recent studies have utilized confocal microscopy to delineate the three-

dimensional angioarchitecture of the mouse retina, describing the three dimensional

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan arrangement as a series of 'capillary hammocks' between major retinal arteries and

veins.57 Immunofluorescent staining techniques combined with confocal imaging of

the capillary networks have allowed a qualitative analysis of separate retinal capillary

plexuses.21 However, the complex three dimensional microvascular arrangement of the

retinal capillaries and their quantitative properties have not been fully determined due

to limitations in resolution capabilities to date.

1.2.5 Current Understanding of Retinal Microvasculature

Unlike a three dimensional network of capillaries that is found in most other organs, the

markedly lamellar structure of the retina renders its capillary networks largely two-

dimensional. The major branches of the retinal artery runs peripherally from the optic

nerve head with successive divisions of the vessel remaining at a superficial level

within the nerve fibre layer.58 At this point, studies of the retinal circulation have

identified two distinct capillary beds subserving the inner retina. One of these capillary

beds, the superficial group, continues on the vitreal layers to serve the nerve

fibre/ganglion cell layers with a deeper group running at a steeper angle to a deeper

level where it is related to the inner nuclear and plexiform layers. There is a distinctive

difference in vascular pattern, mesh density, and presence or absence of larger vessels

noted between the two layers.59 Specifically, the superficial plexus closely

approximates the larger vessels with a more complex capillary distribution and higher

density compared to a mostly planar distribution of the deeper network. Anastomotic

capillaries run from superficial to deep networks, and capillaries, which may run for

part of its course in one layer, may change to the other.

This pattern is modified in thicker parts of the retina with the superficial layer becomes

increasingly more three dimensional, rendering up to four capillary layers noted around

the macular region and optic disc.5 Iwasaki and Inomata were able to show striking

regional variations in capillary lamination within the human perifovea according to

thickness of the ganglion cell layer.60 This is consistent with subsequent studies

showing that the density of capillary volume is known to change in different regions of

the retina with a rise in percentage of retinal area occupied by capillaries noted as a

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan function of increasing eccentricity from the foveal avascular zone.5 The notion that

some capillary planes may have a denser network than others and vary depending on

retinal region implies that retinal blood flow might be preferentially distributed to

certain retinal layers. Such information may confer a deeper understanding of neuronal

activity in both physiological and dysfunctional states.

1.2.6 Current Understanding of Oxygen Demand in the Retina

Given that there is a delicate balance between oxygen supply and consumption in the

retina which places it at particular risk for ischaemic damage, the ability to understand

the different metabolic requirements of different cell types may offer insights into

differential vulnerability in these pathological processes. Vulnerability to ischaemic

damage in the retina stems from a need to balance the density of capillaries and

corresponding oxygen diffusion distances with interference of light pathways. This

results in an average pO2 that is low (20mmHg) relative to other organs of the body.61-

66 Furthermore, the energy demands of the retina on a per gram basis have been

described as higher than that of the brain with the majority supplied from oxidative

metabolism coupled to ATP synthesis.

Several attempts have been made to understand the oxygen requirements of different

retinal subcompartments. Krogh's oxygen diffusion models and conceptualization that

rate of utilization of oxygen is a vital determinant to tissue distribution is particularly

applicable to the structurally complex networks of the inner retina. Through these

calculations, the inner retina is noted to have a high oxygen extraction fraction

indicating a relatively sparse retinal vasculature given its high oxygen consumption

rate.65

The retina has a highly layered structure with spatially separated vascular components.

Thus, the retina is particularly amenable to the measurement of intraretinal oxygen

distribution to predict the environment in particular retinal layers. Experimental models

produced by Yu et al have been generated to calculate these different oxygen

consumption rates in different layers of the inner retina and has highlighted the

intraretinal oxygen environment in a range of species (Figure 2).67-71 In rat retina, such

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan models have provided insights into oxygen metabolism within living retina.

Specifically, there are three zones of dominant oxygen consumption identified in this

species, namely the inner plexiform layer, outer plexiform layer and inner segments of

the photoreceptors.69 Relatively consistent minimum levels of oxygen delivered to the

inner plexiform layer reflects the relatively high oxygen demand requirements of

synaptic activity. This is supported by different studies which have analyzed the

cerebral oxidative capacity in monkey retinas via the surrogate measure of cytochrome

oxidase histochemistry.71 Specifically, this enzyme which exists in the inner membrane

of mitochondria indicates distribution of mitochondria in the human retina

corresponding to denser staining patterns approximating inner plexiform and ganglion

cell layers.72-76

Retinal blood vessels are capable of adaptive structural change in response to varying

conditions and functional demands. In response to systemic hyperoxia, oxygen

regulatory mechanisms which act to dampen the high oxygen environment within the

inner retina have been established. Specifically, measurement of intraretinal oxygen

environments have noted muted oxygen responses stemming from an increase in

oxygen consumption from inner and outer plexiform layers.10 The importance of such

regulatory mechanisms controlling oxygen levels in the inner retina is difficult to

determine but has implications on processes such as laser photocoagulation which

involve alteration of the intraretinal oxygen environment.77 From an architectural

standpoint, this would also suggest the need for a well adapted deep capillary plexus

which would be projected to supply oxygen to the metabolically active inner plexiform

layer, and to a lesser extent, the outer plexiform layer and photoreceptors.

These findings illustrate that oxygen metabolism of inner retinal subcompartments is

complex.49 The notion that powerful oxygen regulating mechanisms are present

indicates that a controlled oxygen environment is critical in the maintenance of retinal

health. Inferences drawn from these studies may also suggest which layers are most

vulnerable to hypoxic injury.


Figure 2: Oxygen Demand in Retinal Subcompartments

A diagram showing oxygen tension in the inferior rat retina correlated with a perfusion-fixed histological cross section. Magnification x 590. This allows correlation of the oxygen profile with anatomical subcompartments. (Adapted from Yu et al, 1994)

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 1.2.7 The Retina as a Surrogate Measure of Vascular Health

Capillary microvasculature is fundamental to the maintenance of neuronal health as the

primary site of nutrient supply and removal of metabolic waste products.78 Changes in

capillary density and alterations in vessel diameter may result in increased vascular

resistance.78 Capillary changes in the aging human cerebral cortex include an increased

mean capillary diameter and volume, increased length of capillaries per unit volume, a

decreased specific surface area and a reduced mean intercapillary distance. Burns et al

also noted changes in cerebral capillary morphology with age including a larger mean

capillary diameter due to loss of endothelial cells and subsequent elongation and

thinning of remaining cells.13 This was noted to correlate with a thinning of the

capillary wall suggesting an increased vulnerability to mechanical damage with

implications for distributions of pressure and wall shear stress in remaining

microvessels. Transport and exchange capability of oxygen would be expected to

decline through such changes in the microvasculature architecture.

Due to the delicate balance between oxygen supply and demand which places the retina

at particular risk of ischaemic damage, numerous studies have explored the relationship

between abnormalities in the retinal microvasculature and systemic vascular diseases.79-

85 Retinal vascular network structure may be adversely altered by diabetes mellitus,

atherosclerosis and high blood pressure. The presence of chronic hypertension is the

most important systemic risk factor for development of both retinal artery and vein

occlusions.86-88 Chapman et al showed narrowing of bifurcation angles and an increase

of the length/diameter ratio of individual vessels within a hypertensive model of a

retinal arterial tree, resulting in potentially detrimental flow patterns.89, 90 Conversely,

various epidemiological studies have shown that retinal vascular changes are predictive

of cardiovascular morbidity with a reduced retinal arteriolar caliber associated with an

increased risk of coronary artery disease. This suggests a complex interplay between

micro- and macro-vascular disease states.82

Chronic hypertension leads to retinal arteriolar narrowing and arteriovenous nicking as

a reflection of impaired vasomotor tone, whilst retinal haemorrhages, microaneurysms

and cotton wool spots represent early small-vessel arteriosclerosis and retinal

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan ischaemia.91, 92 Early changes in diabetic retinopathy include thickening of the

endothelial basement membrane, disruptions in blood flow, loss of mural cells and

genesis of acellular capillaries. With disease progression, endothelial apoptosis and

capillary rarefaction lead to a hypoxic inner retina with attempted compensation via

pathologic angiogenesis.28, 93-96 Precursor anatomic alterations in the capillary

microvasculature before such clinical ocular disease is evident remain unknown. Such

changes, which may manifest as morphological differences in vascular caliber and

tortuosity measurements may reflect deviations from an optimal design principle and

represent the earliest change of impaired microcirculatory transport. The added notion

that different vascular disorders seem to preferentially affect superficial or deeper

capillary plexuses with hypertensive alterations mainly arising in superficial nerve fibre

layer capillaries is of interest.97

Systemically, abnormalities of the retinal circulation not only indicate retinal

dysfunction and disease but may also reflect impaired peripheral circulation, useful as a

biological model to explore the relationship of microcirculatory disease to the

development of cardiovascular disorders. Ophthalmic signs of retinopathy such as

microaneurym, retinal haemorrhage and soft exudates, and biomarkers of

microvascular injury stemming from hypertension and other vascular comorbidities

have also been associated with increased coronary vascular disease mortality.98-100

Histopathologic studies have showed that in patients with retinal vascular alterations,

64% would have concomitant renal vascular changes and 90% would demonstrate

cerebral vascular changes.101, 102 Thus, surrogate assessment of systemic vascular

health through imaging of the retinal vasculature is possible and may indicate the

vascular status of various organs beyond the eye.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 1.2.8 The Perifoveal Region

An important anatomical landmark of the retina is the macula. Within this, the foveala

is an avascular region of the retina containing exclusively cone cells, upon which the

fovea is centered spanning 2500 microns in diameter. The parafovea incorporates the

surrounding intermediate region where ganglion cell layers are composed of more than

five rows of retinal ganglion cells, with the perifovea, in turn encapsulating a region

with the outermost edge located 2750 microns from the centre of the fovea.103

Collectively, these regions consist of multiple retinal layers that are encapsulated by the

macula lutea that appears clinically as a yellow rim of tissue in the central retina

representing the area where vision is sharpest.32, 104

The optic disc is located nasal to the fovea. Thus, ganglion cell fibers arising from the

temporal retina must avoid the macula by anatomically separating to enter the optic

disc at either the superior or inferior pole. Conversely, those arising from the nasal

aspect of the fovea can travel uninterrupted directly to the disc as the papillomacular

bundle.105, 106

The perifoveal region is histologically characterised by approximately four rows of

retinal ganglion cell nuclei and five to seven rows of nuclei in the inner and outer

nuclear layers.32 Nerve fibre layer thickness in the perifoveal macula-papillary bundle

is also greater than most other parts of the macula correlating to an increased capillary

volume in this region. The perifoveal region is an area of interest due to its high density

of axons carrying information from the central retina, the difficultly in access and

labelling by traditional staining methods and previous studies showing that this region

is the most densely vascularized region of the retina.47, 107 For this thesis, the perifovea

and the inner retinal layers corresponding to this region were studied.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 1.3 Optical Vascular Detection Methods

The retinal circulation demonstrates an absolute end-arterial system without

anastomotic supply. The delicate balance of oxygen supply and consumption in the

retina makes the retina particularly vulnerable to ischaemic injury.108 As such, the risk

of irreversible damage follows acute ischaemia, with the presence of capillary drop out

underlying the pathological process of various retinal vascular diseases including

central retinal vein occlusion and diabetic retinopathy.109, 110 Furthermore, there is

mounting evidence that quantitative characteristics of retinal capillary networks are

selectively altered and may occur early in such ischaemic disease processes. Thus, the

ability for the early detection of capillary changes is paramount so that pathogenic

mechanisms leading to ischaemia may be identified and addressed.

The human retina is routinely imaged using a variety of different techniques. Each

modality offers distinct advantages and are divided into techniques requiring tissue

label- fluorescein angiography and microperfusion fixation staining as opposed to label

free techniques- fundus photography, scanning laser ophthalmoscopy, adaptive optics

and optical coherence tomography.111-113 Imaging of the retinal capillaries is difficult

because of their small size and arrangement in multiple planes. Thus techniques must

have high spatial resolution to directly image microscopic vasculature structures such

as capillaries and arterioles.114 Understanding pathogenesis of diseases at a subclinical

stage paves way for early detection and treatment.115 Such retinal imaging modalities

have been developed to allow a more objective and precise assessment of retinal

vascular changes over time, translating to better diagnosis and monitoring of retinal

disease outcomes in clinical practice. Expanding interest in imaging has been

associated with an increase in publications utilizing this technology for clinical

application.112

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 1.3.1 Fundus Fluorescein Angiography

Fundus fluorescein angiography (FA) is the current gold standard technique used in

clinical practice for the examination of the retinal and choroidal circulation.116, 117 It

has a broad clinical application in the management of retinal vascular diseases and

involves intravenous administration of a fluorescein dye with an angiogram obtained by

photographing the fluorescence emitted after light excitation of the retina.118 Normal

circulatory filling is evaluated over time with pathological changes characterized by the

presence of either hyper- or hypo-fluorescence.

Previous studies have showed that meaningful and reliable analysis of capillary

morphology and density in multi-layered capillary network regions via FA is

suboptimal (Figure 3). Detail of the retina is noted to decline proportionally with

distance away from the fovea avascular zone with major limitations including a high

background noise to signal ratio. FA provides a wide-field view of the

microvasculature with limitations in transverse and axial resolution by ocular

monochromatic aberrations.119 Additionally, administration of fluorescein carries the

risk of severe adverse effects including a small but significant risk of anaphylaxis and

death.120 As such, although it accurately delineates large and medium-caliber retinal

vessels, this techniques' ability to allow reproducible delineation of architecture at the

capillary level is limited.


Figure 3: Fluorescein Angiogram Image

Fluorescein angiogram image demonstrates variability in ability to distinguish capillary detail against background fluorescence in a healthy young subject.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 1.3.2 Microperfusion Fixation Staining and Confocal Microscopy

Confocal microscopy utilizes point illumination and the addition of a spatial pinhole in

an optically conjugate plane to eliminate out-of-focus signals (Figure 4). In doing so,

this optical technique provides for exquisite axial and lateral resolution.121 However,

this technique is mainly limited to the ex vivo examination of tissue which has been

excised, preserved and histologically stained with or without the use of fluorescent

markers.122 Drawbacks limiting the use of this technology in the clinical setting

include its relatively slow speed of data acquisition and safety issues regarding the

necessity of fluorescent dye. Given that much of the light from sample fluorescence is

blocked by the pinhole, decreased signal intensity, particularly at deeper planes means

restriction of depth penetration to less than 400µm.123

Confocal microscopy has been successfully applied in microperfused and flat-mounted

retinal preparations which have given valuable information regarding cellular layers

and spatial relationships between microvasculature, neurons and glia in the human

retina.47 Three dimensional information at the cellular level and quantitative

information regarding capillary meshwork density at different layers have been

possible. High resolution depth resolved topographic images have been achieved with

in vivo systems. Technology that permits in vivo confocal scanning of the rat retina has

been developed, but few studies have validated this technique for the reliable

identification of capillary detail in human subjects.124, 125


Figure 4: Confocal microscope

A schematic of a confocal microscope set up demonstrating how addition of a confocal

detector pinhole reduces out of focus light, thereby resulting in better image contrast

and resolution. (Adapted from Department of Bioimaging, John Innes Centre, 2007.

Available from: jic.ac.uk/microscopy/more/T5_8.htm)

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 1.3.3 Scanning Laser Ophthalmoscopy

Scanning laser ophthalmoscopy (SLO) utilizes a small spot of low energy collimated

laser light directed through the pupil, capturing reflections from its surrounds within the

pupil boundary to create an image.126-128 Point by point successive illumination of the

retina is made in combination with a scanning system that allows a large area of fundus

to be imaged. Images are collated via appropriate software and a raster scan is

generated. Adaptation of a confocal technique with addition of a pinhole to this system

eliminates light beyond the plane of focus to minimize scatter and increase axial

resolution of the imaging plane. The output of the photo detecting technique is recorded

and displayed as a mean image of several sampling frames. In combination with

adaptive optics (see below), high imaging speed and adaptations to accommodate for

minor lateral physical changes such as eye movement have been integrated.129 This

technique has been utilized with digital imaging analysis techniques to visualize and

quantify macular networks of capillaries as well as measure flow velocities within

them.127 The Heidelberg Retina Tomograph (HRT, Heidelberg Engineering,

Dossenheim, Germany) is a commercially available confocal SLO system applied to

receive three-dimensional images of the retinal surface in vivo. It has successful

applications in the structural analysis of both the optic disc and peripapillary retinal

nerve fiber layers in glaucomatous eye disease.130, 131

1.3.4 Adaptive Optics

Adaptive optics (AO) is a technique applied in retinal imaging to decrease in vivo

optical aberrations and improve microscopic resolution.129 Distinct from being a

standalone vascular detection method, it refers to technology adapted to improve the

performance of optical systems. It was first envisioned in 1953 and was originally

developed for astronomical telescopes to decrease atmospheric aberrations on the

creation of wavefront distortions.132, 133

AO systems are composed of three principle components including a deformable mirror

that compensates for measured aberrations, a wave front sensor which detects

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan monochromatic wavefront aberrations and specialized software to facilitate interaction

and dynamic adjustment. It is combined with other imaging modalities such as fundus

cameras, SLO and OCT to improve lateral resolution capabilities to the order of 2

µm.129 The availability of commercial systems has dramatically increased the clinical

utility of these other imaging modalities with AO-SLO detection of retinal

microvascular damages reflecting presence of microaneurysms, increased FAZ size and

capillary dropout in patients with diabetes mellitus.134, 135 AO-OCT could be the

desirable imaging modality in the future offering better axial resolution and sensitivity

to weak reflections compared to AO-SLO systems.112

1.3.5 Optical Coherence Tomography

First reported in 1991, optical coherence tomography (OCT) has widespread use in the

modern clinical setting with the ability to provide cross-sectional information of the

vitreoretinal interface and optic nerve head.114 The principles behind this technique are

analogous to how sound waves are utilized in ultrasonography, instead using light

waves to obtain a reflectivity profile of tissue and visualization of structures at all

levels of the retina. Low coherence interferometric detection of backscattered photons

is used. Specifically, shorter light wavelengths permit imaging at higher resolution and

are measured indirectly via an optical technique in which signal carrying light returning

from the eye interferes with emitted light that has travelled a known path length. The

original standard OCT is a time-domain technique as depth information is acquired as a

sequence of samples, over time.

Modern improvements to the OCT technique include the development of spectral

domain-OCT which allows for a faster acquisition of data, higher resolution and

sensitivity.136 Increased signal to noise performance is secondary to a reduction in

detection bandwidth required in this modification. Additional functional extensions

including speckle variance techniques which provide further biological information that

is uniquely suited to visualize tissue at the microcirculation level. Portability, relative

affordability and increasing spatial resolution have allowed recent groups to visualize

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan microvasculature as well as quantify physiological blood flow through detection of

speckle spatial frequencies in OCT intensity.137

Speckle variance optical coherence tomography (svOCT) is an imaging modality that is

gaining increasing popularity for the study of retinal vascular anatomy and disease138 ,

along with other label-free OCT based flow contrast techniques.139-141 svOCT is a non-

invasive technique based on red blood cell movement that is capable of providing real-

time angiographic information about retinal capillary networks without the

administration of intravenous dye.142, 143 For these reasons, svOCT is potentially an

attractive tool with a myriad of clinical applications.


Chapter 2.0: Quantitative Morphometry of Perifoveal Capillary Networks in the Human Retina

2.1 Introduction

To our knowledge, there have not been any studies that have quantified or delineated

the morphometric characteristics of capillary networks in the human perifovea relative

to retinal layers. Our laboratory has developed novel perfusion-based labelling

techniques, utilising micropipette cannulation of the central retinal artery and targeted

antibodies, to accurately label the human retinal circulation.47, 144, 145 We have also

coupled immunolabelling and microscopy techniques with this perfusion methodology

to perform detailed analyses of intercellular relationships in the human macula.146 The

present experiment employs previously validated state-of-the-art technology to quantify

the morphometric features of capillary networks in a specialised region of the retina.

The perifoveal retina, located between the optic disk and fovea, is examined.32 The

region corresponding to the maculo-papillary bundle is analysed due to a relative

increase in nerve fibre layer (NFL) and retinal ganglion cell (RGC) thickness within

this region.147 The aim of this study is to improve our understanding of vascular

mechanisms that support retinal homeostasis.

2.2 Methods

This study was approved by the human research ethics committee at The University of

Western Australia. All human tissue was handled according to the tenets of the

Declaration of Helsinki.

2.2.1 Human Donor Eyes

A total of 17 human eyes from 14 donors were used for this study. All eyes were

obtained from the Lions Eye Bank of Western Australia (Lions Eye Institute, Western

Australia) after removal of corneal buttons for transplantation. Donor eyes used for

this research had no documented history of eye disease. The demographic data and

medical co-morbidities of each donor are presented in Table 1.


Patient ID Sex Age Eye

Co-morbid conditions

Time to perfusion (hrs)

A M 32 L MVA 20 B M 23 L Suicide 22 C M 53 L MVA 14 D M 66 R+L Colon Cancer 15 E M 22 L Suicide 15 F M 27 L MVA 8 G M 59 R+L Melanoma 12 H M 39 L Bacterial Endocarditis 20 I M 68 R COPD 11 J# M 60 L Prostate Cancer 18 K# M 65 R Cardiomyopathy 6.5 L# F 64 L Short illness 5 M# M 72 R Drowning 15 N▲ M 60 R+L Alzheimers Disease 2.5

Table 1 – Donor demographic details. Age (years), sex (M = male or F = female),

cause of death and time to perfusion for each eye donor is provided. Donor eyes that

were flat mounted for co-localisation studies are designated (#) and those that were

sectioned transversely for co-localisation studies are designated (▲). MVA = motor

vehicle accident, COPD = chronic obstructive pulmonary disease


2.2.2 Tissue Preparation

Our previously reported technique of central retinal artery cannulation, microvascular

fixation and targeted endothelial cell labelling was used for this work (Figure 5).146

Briefly, the central retinal artery was cannulated using a glass micropipette and the

retinal circulation perfused with a mixture of oxygenated Ringer’s solution and 1%

bovine serum albumin. After the 20-minute Ringer’s wash the retinal circulation was

perfusion-fixed using a solution of 4% paraformaldehyde in 0.1 M phosphate buffer.

0.1% Triton-X-100 in 0.1 M phosphate-buffered solution was then used to aid in the

permeabilisation of endothelial cell membranes. Detergent was removed from the

retinal circulation by perfusion with 0.1M phosphate buffer solution. Endothelial

microfilaments and nuclei were labelled over 2 hours by perfusion with a solution

comprising of phalloidin conjugated to Alexa Fluor 546 (30 U; A22283; Invitrogen,

Carlsbad, CA) and bisbenzimide (H 33258; 1.2 µg/mL; Sigma-Aldrich, St. Louis, MO).

Residual label was cleared from the vasculature by further perfusion with 0.1 M

phosphate buffer. Eye cups were immersion fixed in 4% paraformaldehyde overnight

prior to dissection. The posterior globe was dissected at the equator to allow viewing

of the posterior retina. Relaxing radial incisions were used to permit retinal flat

mounting.

2.2.3 Immunolabelling

To aid co-localization between capillary networks and retinal layers further

immunolabelling was performed in 4 post-perfused flat mount sections and 2 transverse

retinal sections. In 2 flat mount sections, the relationship between capillary networks,

the NFL and RGCL was explored by incubating the retina with rabbit monoclonal γ-

synuclein antibody148 (Abcam ab55424, 1:200) followed by incubation with goat anti-

rabbit antibody conjugated with Alexa Fluor 488 (Invitrogen A-11008, 1:200). In 2

flat mount sections, the relationship between capillary networks, bipolar cells and

horizontal cells were explored using mouse-Goα antibody149 (Millipore MAB3073;

1:200) and rabbit Parvalbumin antibody150 (Swant PV 25;1:500) respectively. Goat

anti-mouse conjugated with Alexa Fluor 488 and anti-rabbit antibodies conjugated with

Alexa Fluor 633 were used for secondary labelling. All primary antibodies were

incubated for 3 days followed by overnight incubation of secondary antibodies.


Figure 5: Microcannulation of the Central Retinal Artery

Micropipette technology was used to cannulate the central retinal artery for perfusion, fixation and immunolabelling of various retinal structures.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Transverse retinal sections used for co-localization studies were of 12 µm thickness and

were prepared from regions that were used for flat mount confocal microscopy studies.

Transverse retinal sections were immunolabelled with γ-synuclein antibody148 (Abcam

ab55424, 1:200), Goα antibody149 (Millipore MAB3073; 1:200) and parvalbumin

antibody150 (Swant PV 25;1:500). Lectin-TRITC151 (Sigma L5266, 1:40) was also used

to label blood vessels in transverse sections.

2.2.4 Microscopy

In all flat mount retinal specimens confocal microscopy images were acquired from the

peri-foveal region - located 2 mm nasal to the center of the fovea (Figure 6).32 Wide-

field images were captured prior to confocal microscopy with the aid of a x4 dry lens

(Plan NA 0.2; Nikon, Tokyo, Japan) and a fluorescent microscope (Eclipse E800;

Nikon). Wide-field images were used to accurately measure distances from the centre

of the fovea prior to confocal scanning. Confocal microscope images were captured

using a Nikon C1 Confocal with EZ-C1 (v. 3.20) image acquisition software. A x20

dry objective lens (NA 0.4) was used for all scans. Using a motorised stage, a series of

Z stacks were captured for each specimen beginning from the vitreal surface, at the

level of the inner limiting membrane, to the outer retina. Each z-stack consisted of a

depth of optical sections collected at 0.35 µm increments along the z-plane.

Images of different wavelengths were acquired sequentially. Visualisation of sections

labelled with Alexa Fluor® 408 secondary antibody was achieved by laser excitation at

a 405 nm line from an argon laser with emissions detected through a 450/35 nm band

pass filter. Visualisation of sections labelled with Alexa Fluor® 488 secondary

antibody was achieved by laser excitation at a 488 nm line from an argon laser with

emissions detected through a 525/50 nm band pass filter. Visualisation of sections

labelled with Alexa Fluor® 564 secondary antibody was achieved by laser excitation

at a 545 nm line from an argon laser with emissions detected through a 595/40 nm band

pass filter. Visualisation of sections labelled with Alexa Fluor® 637 secondary

antibody was achieved by laser excitation at a 637 nm line from an argon laser with

emissions detected through a 450/35 nm band pass filter.


Figure 6: Human perifovea

Colour fundus (A) and corresponding Zeiss Cirrus high-definition ocular coherence tomography image (B) from a healthy subject illustrate the perifoveal region that was studied (fenestrated box). Nerve fibre layer thickness in the perifoveal region (indicated by arrows) is greater than most other parts of the macula. Toluidine blue stained section from a separate subject demonstrates normal perifoveal histology (C). This region has approximately four rows of retinal ganglion cell nuclei and five to seven rows of nuclei in the inner and outer nuclear layers. OD = Optic Disk, NFL = Nerve Fibre Layer, RGCL = Retinal Ganglion Cell Layer, IPL = Inner Plexiform Layer and INL = Inner Nuclear Layer. Scale bar in colour fundus and ocular coherence tomography image = 1000 µm. Scale bar in toluidine blue histology = 50 µm.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 2.2.5 Image Preparation

ImagePro Plus (Media Cybernetics, Version 7.1) and Image J (version 1.43, National

Institute of Health, USA, http:/rsb.info.nih.gov/ij) were used to quantify confocal

microscope images. All images for the manuscript were prepared using Adobe

Photoshop (version 12.1, Adobe Systems Inc.) and Adobe Illustrator CS5 (version

12.1.0, Adobe Systems Inc.). Confocal images in this manuscript were pseudo-

coloured using Look Up Tables available on ImageJ.

2.2.6 Qualitative Differentiation of Capillary Networks

Previous central nervous system studies have demonstrated that capillary networks

within the brain, subserving distinct neuronal and glial populations, are characterised

by unique cyto-architectural features.19, 20, 45, 152 Morphometric criteria previously

defined by these structure-function histological studies were used to partition the retinal

circulation into different capillary networks.

A z stack of confocal images that only contained information derived from the vascular

channel of each eye was viewed using an animation sequence on Image J and a

capillary network was defined as being different when one of the following

morphometric criteria was satisfied:

• Alteration in projected direction and orientation of capillaries5, 19, 153

• Alteration in capillary branching pattern5, 19

• Reduced presence of capillaries within retinal tissue

Qualitative division of the retinal circulation into different capillary networks was

performed by 4 separate observers using the above criteria. Each observer recorded the

location of the first and final image slice, within the z stack, for each capillary network.

Inter-observer correlation between the number and location of capillary networks were

subsequently performed (described below).

Following the division of z stacks into separate capillary networks, information

acquired from remaining laser channels were merged with vascular channels and used

to co-localize individual capillary networks with neuronal structures in the layered

retina. Specifically, capillary networks were localized in relation to nuclei, the nerve

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan fiber layer, ganglion cell layer, bipolar cells and horizontal cells. Quadruple-labelled

transverse retinal sections were also used to aid in co-localization studies.

2.2.7 Morphometric Quantification of Capillary Networks

Quantitative data from capillary networks were attained following z projection of all

images between the first and final image slice for each network as determined above.

Capillary morphometric measurements were performed using previously defined

histologic parameters.15 ,22, 45 Manual tracing methods were used to attain the

following quantitative measurements from each capillary network (Figure 7):

• Capillary diameter - Defined as the perpendicular distance across the maximum

chord axis of each vessel. Each confocal image was partitioned into 9 equal

regions (Fig. 7B) and measurements were obtained from each region to ensure

representative sampling. An average of 45 measurements was obtained from

each image. Capillaries were defined as vessels with an absence of smooth

muscle cells with diameter less than 10 µm.19

• Capillary loop area - Defined as the area circumscribed within visually enclosed

capillary loops.22

• Capillary density - Defined as percentage of the sample area occupied by

capillary lumens.

2.2.8 Statistical Analysis

All data is expressed in terms of mean and standard error which were calculated using

Sigmastat (Sigmastat, ver. 3.1; SPSS, Chicago, IL). Multiple measurements from eyes

with data taken from right and left eyes of the same individual were analyzed using R

(R Foundation for Statistical Computing, Vienna, Austria).154 One-way analysis of

variance (ANOVA) testing was performed to compare measurements between layers.

The model used included “Right” or “Left” nested within “eye donor” as random

effects using linear mixed modeling to test measurement differences between retinal

layers154 . The assignment of donor as a random effect was used to account for the

effects of intra-“eye” correlation and similarly “Right” and “Left” to account for right

and left eye correlation.


Figure 7: Methodology for Quantification of Capillary Networks

Representative Z-projected confocal microscope image of the deep inner nuclear layer network (A) and a corresponding manually traced image (B) illustrates the vascular parameters that were measured. Vessel density (highlighted in red) was expressed as a percentage of total area. Capillary loop area (highlighted in green) expressed as µm2 and capillary diameter (blue marks) expressed as µm were also measured. Each image was divided into 9 equal portions (fenestrated lines) and capillary diameter measurements were obtained from each region. Scale bar = 100 µm.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 2.2.9 Inter-observer Correlation and Measurement Reproducibility

To assess inter-observer variation in distinguishing capillary networks, four masked

observers were asked to identify stack numbers for the beginning and end of each

morphometrically distinct capillary network in five eye specimens. Two-way ANOVA

was performed modeling stack number with observer and layer to see if observer

accounted for any variation in first or last stack number identified.

Twelve capillary microvasculature images derived from three different specimens were

quantified on three separate occasions, each at least 1 week apart, by the same masked

observer who performed all the data analyses. Capillary diameter, capillary loop area

and capillary density were quantified for each image. Two-way ANOVA was used to

assess the effect of day of measurement and capillary network layer on each of the

capillary morphometric parameters with eye donor as a random effect. “Right” or

“Left” was not considered because only one eye from each eye donor was used.

2.3 Results

2.3.1 Eye donors

The mean age of donors was 50.71 ± 4.84 years (age range, 22 – 72 years). We

examined 6 right eyes and 11 left eyes from a total of 13 male and 1 female donors.

The average post mortem time before eyes were perfused was 13.14 ± 3.51 hours.

2.3.2 General

All orders of retinal microvasculature were clearly labelled after perfusion labelling via

the central retinal artery. Endothelial cells and nuclei were clearly identified following

excitation with separate laser channels. Immersion immunolabelling of flat-mount and

transverse retinal sections, post-perfusion, also allowed identification of capillaries

relative to the NFL, RGCL, inner plexiform layer (IPL) and inner nuclear layer (INL).

Transverse retinal sections (Figure 8), from the region of interest, demonstrated a well

delineated NFL, RGCL, IPL and INL. In the perifoveal region the RGCL

demonstrated several strata of cell bodies (Fig. 6C).


Figure 8: Co-localization of Capillary Networks within Retinal Layers

Triple- (A) and quadruple-labelled (B) transverse retinal sections demonstrate the relationship between capillary networks and nerve fibre layer (NFL), retinal ganglion cell layer (RGCL), inner plexiform layer (IPL) and inner nuclear layer (INL). Endothelial cells are labelled with Phalloidin, nuclei are labelled with Hoescht, RGCL and NFL are labelled with γ-synuclein. Goα and Parvalbumin were used to label Bipolar and Horizontal cells, respectively. Four capillary networks (fenestrated red brackets) were identified in the inner retina. Scale bar = 50 µm.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 2.3.3 Qualitative study of retinal capillary networks

Four morphologically varied retinal capillary networks were consistently reported by

the four observers (summary Fig. 8). The morphometric features of each capillary

network were as follows:

• Nerve Fibre Layer (NFL) network (Fig. 9) – Characterised by long capillary

segments that were predominantly oriented parallel to the direction of retinal

ganglion cell axons. A small number of shorter capillary segments, that

interconnected long radial capillaries, were also seen in this network.

Interconnecting capillaries were oriented either diagonal or orthogonal to long

segments.

• Retinal Ganglion Cell and Superficial Inner Plexiform Layer (RGCL/sIPL)

network (Fig. 10) – Characterised by a dense meshwork of three dimensional

vessels that were arranged in a lattice pattern with reduced inter-capillary

spaces. Capillaries in this network demonstrated looping hairpin turns that

projected vertically. There was also close approximation between RGCL

capillaries and the larger retinal vessels.

• Deep Inner Plexiform and Superficial Inner Nuclear Layer (dIPL/sINL) network

(Fig. 11) – Capillaries in this network comprised of vertical and oblique

segments that resulted in an irregularly shaped loop configuration. Capillaries

in this network demonstrated great tortuosity.

• Deep Inner Nuclear Layer (dINL) Network (Fig. 12) – Capillaries in this

network were arranged in a one-dimensional laminar configuration. Capillaries

ran a linear trajectory with little tortuosity.

2.3.4 Quantitative Characteristics of the Nerve Fibre Layer Network (NFL)

Mean capillary diameter in the NFL network was 8.30 ± 0.07 µm (n = 347). A total of

84 capillary loops were measured with a mean area of 4940.31 ± 529.38 µm2.

Capillary density in the NFL network was 17.37 ± 0.99%. NFL capillary

measurements from donors are presented in Table 2.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 2.3.5 Quantitative Characteristics of the Retinal Ganglion Cell and Superficial Inner Plexiform Layer network (RGC/sIPL)

Mean capillary diameter in the RGCL/sIPL network was 8.29 ± 0.07 µm (n = 304). A

total of 102 capillary loops were measured with a mean area of 3954.04 ± 505.30 µm2.

Capillary density in the RGCL/sIPL network was 22.32 ± 0.99%. RGCL/sIPL

capillary measurements from donors are presented in Table 2.

2.3.6 Quantitative Characteristics of the Deep Inner Plexiform and Superficial Inner Nuclear Layer network (dIPL/sINL)

Mean capillary diameter in the dIPL/sINL network was 8.25 ± 0.07 µm (n = 325). A

total of 63 capillary loops were measured with a mean area of 5424.03 ± 602.72 µm2.

Capillary density in the dIPL/sINL network was 19.56 ± 0.99%. dIPL/sINL network

capillary measurements from donors are presented in Table 2.

2.3.7 Quantitative Characteristics of the Deep Inner Nuclear Layer Network (dINL)

Mean capillary diameter in the dINL network was 8.26 ± 0.07 µm (n = 372). A total of

66 capillary loops were measured with a mean area of 6866.52 ± 584.58 µm2.

Capillary density in the dINL was 17.95 ± 0.99%. dINL capillary measurements from

donors are presented in Table 2.


Table 2 – Quantitative capillary network data for donors. Mean capillary diameter, capillary loop area and capillary density measurements

for each individual donor eye are provided. SE = Standard error. NFL = Nerve Fibre Layer, RGCL = Retinal Ganglion Cell Layer, sIPL =

superficial Inner Plexiform Layer, dIPL = deep Inner Plexiform Layer, sINL = superficial Inner Nuclear Layer and dINL = deep Inner

Nuclear Layer.

Specimen Capillary diameter (µm) Capillary loop area (µm2) Capillary density (%) NFL RGCL/sIPL dIPL/sINL dINL NFL RGCL/sIPL dIPL/sINL dINL NFL RGCL/sIPL dIPL/sINL dINL A 8.6 8.1 8.5 7.6 2462.5 4823.8 9397.2 15593.1 15.0 23.4 16.5 14.7 B 8.1 7.8 6.5 7.4 7494.5 1814.7 0.0 4841.8 14.8 25.3 14.4 18.9 C 8.3 8.8 8.8 9.0 4684.0 8754.2 3436.5 4169.8 18.6 16.8 20.4 17.6 D 8.4 8.3 8.7 9.2 5394.9 2826.6 7461.2 6307.3 22.4 26.0 24.6 24.2 D 8.8 8.0 8.3 7.5 6987.6 2782.1 2354.3 4562.3 20.2 34.6 25.0 21.7 E 8.6 8.8 8.1 8.1 5959.9 6000.8 8412.4 7158.17 22.0 16.2 14.1 15.2 F 7.9 8.3 8.4 8.9 3543.7 5318.5 5401.9 7326.5 10.5 21.1 20.7 18.9 G 7.9 8.3 8.2 8.5 5688.8 0.0 9063.6 9813.8 19.2 19.6 19.7 15.2 H 8.7 8.6 8.1 8.4 3789.6 6234.2 4472.2 0.0 17.4 16.8 15.0 12.6 I 7.6 8.3 8.8 8.4 1645.9 5622.6 3219.5 4371.5 14.8 21.2 23.1 21.3 J 8.2 7.3 8.4 8.0 4788.9 3500.3 3979.8 8799.2 16.3 24.6 21.7 17.2

Mean 8.30 8.29 8.25 8.26 4940.31 3954.04 5424.03 6866.52 17.37 22.32 19.56 17.95

SE 0.07 0.07 0.07 0.07 529.38 505.30 602.72 584.58 0.99 0.99 0.99 0.99


Figure 9: Nerve Fibre layer Capillary Network

Endothelial cells (labelled with phalloidin) stain red, nuclei (labelled with Hoescht) stain blue and NFL axons (labelled with Anti-γ-synuclein) stain green. Confocal capillary images captured from a single laser channel (A) demonstrate that the trajectory of a large number of capillaries parallel axons in the nerve fibre layer. Merged images (B) demonstrate a paucity of nuclei within this network. Capillary images from a separate donor (C) merged with a stain specific for retinal ganglion cell axons (D) demonstrate co-localization of this network with the nerve fibre layer of the retina. Scale bar = 100 µm.


Figure 10: Ganglion Cell layer/superficial Inner Nuclear layer Network

Endothelial cells (labelled with phalloidin) stain red, nuclei (labelled with Hoescht) stain blue and retinal ganglion cells (labelled with Anti-γ-synuclein) stain green. Confocal capillary images captured from a single laser channel (A) demonstrate the complex capillary configuration in this network. The proximity of capillaries to larger order arterioles and veins in this network is also demonstrated. Merged images (B) demonstrate a high concentration of nuclei within this network. Capillary images from a separate donor (C) merged with a stain specific for retinal ganglion cell nuclei (D) demonstrate co-localization of this network with the ganglion cell layer of the retina. Scale bar = 100 µm.


Figure 11: Deep Inner Plexiform layer/superficial Inner Nuclear layer Network Endothelial cells (labelled with phalloidin) stain red, nuclei (labelled with Hoescht) stain blue, bipolar cells and their cell processes (labelled with Goα) stain green. Confocal capillary images captured from a single laser channel (A) demonstrate the irregularly shaped loop configuration of capillaries in this network. Merged images (B) demonstrate a high concentration of nuclei within this network. Capillary images from a separate donor (C) merged with bipolar cell markers (D) demonstrate the intimate relationship of capillaries to processes of bipolar cells which co-localize this network with the inner plexiform layer of the retina. Scale bar = 100 µm.


Figure 12: Deep Inner Nuclear layer Network Endothelial cells (labelled with phalloidin) stain red, nuclei (labelled with Hoescht) stain blue, bipolar cells (labelled with Goα) stain green. Confocal capillary images captured from a single laser channel (A) demonstrate the planar configuration of capillaries in this network. Merged images (B) demonstrate a high concentration of nuclei in the region of this network. Capillary images from a separate donor (C) merged with Bipolar cell markers (D) demonstrate a lower concentration of nuclei in comparison to the superficial INL network. Scale bar = 100 µm.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 2.3.8 Morphometric Comparisons Between Networks

There was no difference in capillary diameter between networks (P = 0.715). Capillary

loop area was smallest in the RGCL/sIPL network and was significantly smaller than

the dIPL/sINL network (P = 0.028) and dINL network (P < 0.003) but not the NFL

network (P > 0.050). There was no difference in capillary loop area between

dIPL/sINL and dINL networks (P = 0.088) and also between NFL and dIPL/sINL (P =

0.632) and dINL networks (P = 0.107).

Capillary density was greatest in the RGCL/sIPL network and was significantly greater

than the NFL network (P = 0.015) and dINL (P = 0.004). There was no difference in

capillary density between RGCL/sIPL network and dIPL/sINL networks (P = 0.074),

dIPL/sINL and dINL (P = 0.069), NFL and dIPL/sINL (P = 0.179) and also NFL and

dINL networks (P = 0.689).

2.3.9 Inter-observer Correlation and Measurement Reproducibility

The assignment of first and last stack numbers, to define capillary networks, was not

different between the four observers (P = 0.999). Analysis of measurement

reproducibility did not reveal a significant difference in capillary diameter

measurement between the 3 measurement days (P = 0.836). There was also no

significant difference between the 3 measurement days for capillary loop area (P =

0.976) and capillary density measurements (P = 0.549).

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 2.4 Discussion

The major findings from this study are: (1) Capillary networks in the human perifovea

demonstrate morphometric variation according to retinal layer. (2) Capillary density

varies between retinal layers and is greatest in the RGCL/sIPL and dIPL/sINL

networks. (3) Capillary loop area is smallest in the two innermost networks. (4) There

is no difference in capillary diameter between the layers of the retina.

The perifovea is a specialised region of the human eye that is histologically and

functionally distinct in comparison to other portions of the retina.32 Unlike the

peripheral retina, which is comprised of only a single row of ganglion cells, the nasal

perifovea is characterised by 4-5 rows of retinal ganglion cells.32 Additionally, the

nerve fibre layer in the perifovea is thicker than in other parts of the macula.155 The

macula-papillary bundle traverses the perifovea and is the conduit through which a

large quantity of pre-cortically processed retinal information is transmitted to the brain.

Consequently, diseases that preferentially affect the macula-papillary bundle result in

devastating visual morbidity. Understanding the structure of capillary networks serving

the perifovea may provide insights into vascular-mediated mechanisms that satisfy the

metabolic demands of this region.

The present study identified four morphometrically different capillary networks within

the human perifovea. This finding was verified by four masked, independent

observers. The innermost and outermost networks, situated in the NFL and deep

portion of the INL respectively, demonstrated a laminar, one-dimensional

configuration. Capillary networks in the NFL were also observed to project parallel to

the trajectory of retinal ganglion cell axons and resembled the microcirculation

described in skeletal muscle where capillaries are oriented parallel to the direction of

muscle fibres.156 Interconnecting, orthogonally-oriented anastamoses are also seen in

the NFL, similar to skeletal muscle capillary systems.156 In contrast, the capillary

networks located in the RGCL/sIPL and dIPL/sINL demonstrated a tortuous, three-

dimensional architecture that resembled the voroni tessellation described in cortical

capillary beds.157 The variation in retinal capillary network morphology identified in

the present study demonstrates important parallels to the human cerebral cortex where

the microcirculation is also altered according to neuronal layer.158-160

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Unlike the brain, the retina is readily accessible for investigating the physiological

behaviour of subcellular components within distinct neuronal layers. Using oxygen

sensitive microelectrode techniques, intra-retinal oxygen distribution and oxygen

uptake in different cellular layers has been quantified during physiological and non-

physiological states.49, 50, 63, 71, 77, 161-165 These previous studies have identified three

distinct regions of high oxygen uptake: (1) The inner segment of photoreceptors. (2)

The inner plexiform layer. (3) The outer plexiform layer.10, 49, 69 The relationship, and

proximity, between each region of high oxygen uptake and the local microcirculation

however is vastly different. Detailed studies have shown that photoreceptors, including

their nuclei, the high energy-consuming inner segments and the photosensitive outer

segments, lie within the avascular layers of the retina.166 Oxygen and nutrient supplies

to photoreceptors are completely dependent upon diffusion mechanisms from choroidal

and deep retinal vascular beds. Although oxygen tension in the choroid is high, oxygen

tension at the level of the inner segments is paradoxically low.10 In contrast, the inner

half of the retina is supported by a sparse distribution of retinal vessels with significant

disparities in oxygen levels between inner retinal layers.10 The present study identifies

important relationships between neuronal sub-compartments and regional capillary

network morphometry and may be important for understanding vascular-mediated

mechanisms that account for the heterogeneous oxygen profile across the inner retina.

It may also identify vascular-mediated mechanisms that permit momentary variations

in neuronal metabolic demands to be satisfied.49, 71 Retinal glia are likely to play a

critical role in modulating changes in regional blood supply consequent to variations in

neuronal demands.167

Investigating how retinal capillary network topography is coupled with regional

neuronal demands is important for understanding physiological mechanisms that

support retinal homeostasis. The present study demonstrates that the inner plexiform

layer is supported by two capillary networks - situated in the inner and outer boundaries

of the IPL. Capillary density is greatest in these networks, relative to other retinal

layers, suggesting that the energy demands of neuronal arborisations are high. The

three-dimensional organisations of these networks most likely serve to increase oxygen

delivery and waste removal within the IPL.152, 168 It was surprising that the central

portion of the IPL was relatively devoid of vasculature, however this region is known

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan to have considerable Muller cell support.169 We speculate that Muller cells play a vital

role in supporting neuronal homeostasis in the mid-portion of the IPL by transporting

nutrients from adjacent capillary networks to the dense arborisations in this region.

The importance of the glia-neuron-vascular unit in retinal health and disease has been

extensively documented.170 In addition to Muller cells, the pattern of mitochondrial

distribution is also expected to sustain metabolically active retinal layers that have a

scant capillary circulation.171

Inter-capillary areas bear important relationships to oxygen diffusion properties and it

is postulated that decreasing inter-capillary areas result in decreased oxygen diffusion

times22 – the net effect being increased ATP production. Capillary loop area was

lowest in NFL and RGCL/sIPL networks suggesting that oxygen diffusion may be an

important mechanism by which neuronal function is supported in these layers. The

significant differences in capillary loop area between the two networks that serve the

IPL also suggests that the process of oxygen diffusion plays a disparate role in

supporting the inner and outer portions of the IPL. Mean capillary diameter is also

known to influence the rate of capillary oxygen exchange.172 Unlike studies in intra-

cortical capillary networks173 we did not detect significant differences between

perifoveal capillary network diameters.

The purpose of this study was to identify major differences in capillary network

morphometry between perifoveal layers. Our experimental model of central artery

cannulation and perfusion is best suited for such an investigation as it ensures reliable

and complete labelling of the retinal microcirculation. Confocal microscopy and

immunohistochemical techniques developed in our laboratory also ensured accurate

correlation of capillary-neuronal relationships. Trypsin digestion174 and vascular

casting175 techniques were previously used to study the retinal microcirculation

however inadvertent tissue destruction consequent of these methodologies limited

accurate delineation of such inter-relationships.

The results of this study suggest that capillary network morphometry is coupled with

neuronal demands in the human perifovea. It also demonstrates that the correlation

between neuronal metabolic activity and capillary network location is not exact. The

IPL is supported by capillary networks that are situated on the boundaries of this layer

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan and not within it. Such organisation may have relevance for maintaining the optical

properties of the eye. The results of this study also indicate that complex cellular and

subcellular mechanisms, other than capillary networks, including glial cells and

mitochondria are important for supporting retinal homeostasis. It is important to

further investigate how these complex mechanisms are perturbed in ocular diseases as

it may provide insights into pathophysiological processes involved in retinal vascular

diseases. Capillary network morphometry is altered with disease15 and it will therefore

be important to perform similar morphometric studies using abnormal human eyes.


Chapter 3.0: Quantitative Changes in Perifoveal capillary networks in patients with vascular comorbidities

3.1 Introduction

Utilising novel micro-cannulation techniques developed in our laboratory we have

previously quantified somecharacteristics of capillary networks that serve the retina.59,

144, 145, 176, 177 This study had identified four different capillary networks in the normal

human retina, with the quantitative and morphometric characteristics of each network

varying with retinal depth and eccentricity.176, 177

The present study employs two- and three-dimensional image analysis techniques to

document perifoveal capillary network changes in patients with a history of

cardiovascular morbidity but no clinically detectable retinal disease. We also

determine if capillary networks are selectively or uniformly altered by cardiovascular

morbidity. The purpose of this study is to investigate the consequence of

cardiovascular morbidity on retinal capillary network morphometry, the results of

which may allow clinico-histological correlation with modern imaging devices to aid in

the early detection of retinal vascular disease. As the retina is considered a window to

the brain, findings in this study may also provide a valuable contribution to previous

investigations that have demonstrated important relationships between retinal vascular

characteristics and cardiovascular morbidity.80-85

3.2 Methods


Western Australia. All human tissue was handled according to the tenets of the

Declaration of Helsinki.


Human donor eyes used for this research were absent of clinically evident ocular

disease. Ocular disease in donor eyes was excluded using patient records, post mortem

examination by an ophthalmologist and also examination of retinal flat mounts with a

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan fluorescent microscope after perfusion. Specifically, examination of retinal flat mounts

did not demonstrate the presence of microaneurysms, neovascular vessels, leakage,

arterial/vein occlusions, vascular anomalies or areas of retina non-perfusion in donor

eyes. A total of 10 human eyes from 6 donors with vascular comorbidities and 17

control eyes from 12 donors were used. Most of the control eyes used in this report

were also used in our previous publication.177 Additionally, six control eyes from 4

new donors were also used – 2 of these donors died from hanging (ages 15 and 19

years) and the other 2 died from metastatic cancer (ages 67 and 73). In this study, the

group with cardiovascular comorbidities is referred to as the 'disease' group and eyes

from control donors as the 'control' group. Patients in the disease group suffered from

one or more of the following vascular comorbid conditions: hypertension, congestive

cardiac failure, dyslipidemia, atherosclerosis, atrial fibrillation or

hypercholesterolaemia (Table 3). All eyes were obtained from the Lions Eye Bank of

Western Australia (Lions Eye Institute, Western Australia). The demographic data and

vascular co-morbidities of each donor are presented in Table 3.

3.2.2 Tissue Preparation

Our previously reported method of micro-cannulation and targeted perfusion based

labelling techniques was utilized to label the retinal microvasculature. 59, 144, 145, 176, 177

The central retinal artery was cannulated using a glass micropipette and the retinal

circulation perfused for 20 minutes with a mixture of oxygenated Ringer’s solution and

1% bovine serum albumin. A solution of 4% paraformaldehyde in 0.1 M phosphate

buffer was then used for perfusion-fixation with permeabilisation of endothelial cell

membranes aided by a 0.1%Triton-X-100 in 0.1 M phosphate-buffered solution.

Detergent was removed from the retinal circulation by perfusion with 0.1M phosphate

buffer solution. Endothelial microfilaments were labelled over 2 hours by perfusion

with a solution comprising of either phalloidin conjugated to Alexa Fluor 546 (30 U;

A22283; Invitrogen, Carlsbad, CA) or Lectin-TRITC1616 (Sigma L5266, 1:40).

Nucleus labelling was achieved with bisbenzimide (1.2 µg/mL; Sigma-Aldrich, St.

Louis, MO). Residual label was cleared from the vasculature by further perfusion with

0.1 M phosphate buffer. Labelled specimens were immersion fixed in 4%

paraformaldehyde overnight prior to dissection and flat mounting.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Patient

ID Sex Age Eye Vascular

Comorbidities Cause of death Time to

perfusion (hrs)

A F 56 R+L

Hypercholesteraemia Renal transplant failure 8.5

B M 65 R+L

Hypertension Atherosclerosis

ICH 14

C M 64 L

Hypertension MI 14

D M 60 R+L

Atrial Fibrillation Hypertension Heart Failure

MI 8.5

E M 64 L

Hypertension Dyslipidaemia

MI 22

F M 79 R+L

Hypertension Atherosclerosis

ICH 15

G M 32 L - MVA 20

H M 23 L -

Suicide 22

I M 53 L -

MVA 14

J M 66 R+L - Metastatic Colon Cancer 15

K M 22 L - Suicide 15

L M 27 L - MVA 8

M M 59 R+L - Melanoma 12

N M 39 R+L -

Bacterial Endocarditis 20

O M 15 R+L - Hanging 22.5

P M 19 R - Hanging 17

Q M 67 L - Metastatic Lung Cancer 12

R F 73 R - Metastatic Breast Cancer 15

Table 3 – Donor demographic details. Age (years), sex (M = male or F = female), eye (R= right or L = left), vascular comorbidities, cause of death and time to perfusion for each eye donor is provided. ICH = Intracranial haemorrhage. MI = Myocardial infarction.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 3.2.3 Microscopy

The perifoveal region, located 2 mm nasal to the center of the fovea, was imaged using

confocal laser scanning microscopy.32 The region that was studied occupied an area of

4.1 x 105 µm2. The perifovea was examined because we had previously quantified the

morphometric characteristics of capillary networks in this region in normal human eyes

and thus were able to make direct and reliable comparisons with eyes from patients

with cardiovascular comorbidities.177 The other reason the perifovea was analysed is

because it is traversed by the macula-papillary bundle and thus plays a critical role in

visual processing. Furthermore, unlike the peripheral retina which is comprised of only

a single row of ganglion cells, the nasal perifovea is characterised by 4-5 rows of

retinal ganglion cells.32 Wide-field images utilizing a x4 dry lens (Plan NA 0.2;

Nikon, Tokyo, Japan) and a fluorescence microscope (Eclipse E800; Nikon) were used

to accurately measure distances from the centre of the fovea prior to confocal scanning.

Confocal microscope images were captured using a Nikon C1 Confocal with EZ-C1 (v.

3.20) image acquisition software. A x20 dry objective lens (NA 0.4) was used for all

scans. Using a motorised stage, a series of Z stacks were captured for each specimen

beginning from the vitreal surface at the level of the inner limiting membrane to the

outer retina. Each z-stack consisted of a depth of optical sections collected at 0.35 µm

increments along the z-plane. Visualisation of sections was achieved by sequential

laser excitation at a 488-nm and 561-nm line from an argon laser with emissions

detected through a 525/50-nm 595/40-nm band pass filter respectively.

3.2.4 Image Preparation

ImagePro Plus (Media Cybernetics, Version 7.1) and Image J (version 1.43, National

Institute of Health, USA, http:/rsb.info.nih.gov/ij) software were used to quantify

confocal microscope images. All images for the manuscript were prepared using

Adobe Photoshop (version 12.1, Adobe Systems Inc.) and Adobe Illustrator CS5

(version 12.1.0, Adobe Systems Inc.). Confocal images in this manuscript were

pseudo-coloured using Look Up Tables available on ImageJ.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Three-dimensional (3-D) reconstruction of capillary networks was performed using

Imaris software (Verision 7.4.2; Bitplane, Zurich, Switzerland). To minimize artefacts

caused by minor fluctuations in signal intensity, the dataset was refined by utilizing the

Gaussian filter tool and background subtraction prior to 3-D reconstruction (both, Filter

Width determined by the minimum capillary diameter within each confocal stack).


Our previously defined morphometric criteria, that has been validated by inter-observer

correlation studies, was used to partition the retinal circulation into separate capillary

networks. 176, 177 Capillary branching patterns, trajectory andposition with respect to

the nuclear layers were used to stratify the retinal circulation into separate networks.

Qualitative studies were only performed on eyes from the disease group as we had

previously reported the characteristics of different networks in normal eyes. 177

3.2.6 Morphometric Quantification of Capillary Networks

Our previously published methods were used to quantify the morphometric features of

capillary networks in the disease group.177 Manual tracing methods using Image J

software (version 1.43, National Institute of Health, USA, http:/rsb.info.nih.gov/ij) was

employed to obtain capillary diameter, capillary loop area, capillary loop length and

capillary density measurements from each network.19, 20, 45, 95, 152 Capillary loop area

was defined as the area enclosed by regional capillaries that intersected to form a

closed loop within a network. The boundary of individual capillary loops in each

network was determined by carefully scrolling through image slices within the stack.

Three-dimensional image reconstructions (described below) were also used to

accurately delineate the boundaries of capillary loops. Once the boundaries of

individual capillary loops were determined, the z-projected image of the stack was used

to calculate the loop area. Capillary loop length was measured as the maximal linear

distance between inner vessel walls formed by enclosed capillary loops. This typically

represented the distance between two opposing capillary loop apices on a z-projection.

Capillary density was determined by calculating the percentage of the image that was

occupied by capillary lumens. Manual tracing techniques were used to determine the

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan area of the image occupied by capillary lumens. Capillary density was determined by

dividing this measurement by the total area of the image. In some eyes, data was

unable to be attained from all capillary networks due to limitations in image quality.

Imaris statistics sum function was applied to three-dimensional image reconstructions

to express the surface area of each capillary network as a proportion of the total surface

area occupied by capillaries in the peri-foveal region. This calculation was performed

using eyes in the control and disease group and is referred to as the relative occupied

capillary surface area.


All data are expressed in terms of mean and standard error, which were calculated

using R (R Foundation for Statistical Computing, Vienna, Austria). The disease eye

cohort data was initially analyzed followed by comparisons with control groups. In the

disease group analysis, capillary network layer, age and post-mortem time were

assigned as covariates and morphometric measurements as the response variable.

Multiple measurements from eyes, with data taken from right and left eyes of the same

individual were analyzed with ‘‘Right’’ or ‘‘Left’’ eye nested within ‘‘eye donor’’,

assigned as random effects to account for factor correlations.23 Analysis of covariance

(ANCOVA) testing between four capillary networks generated three comparisons, thus

a P value of <0.017 was considered significant.

In the analysis comparing disease and control eye cohorts, we grouped measurements

from a single capillary network. Similarly, group (control or disease group), age and

post-mortem time were assigned as covariates with morphometric measurements as the

response variable. ‘Right’’ or ‘‘Left’’ eye nested within eye donor were assigned as

random effects to account for factor correlations.154 For these comparisons,

ANCOVA testing was performed with a P value of <0.050 considered significant.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 3.3 Results

3.3.1 Eye Donors

The mean age of donors in the control group was 43.6 + 5.1 years (age range 15 to 73

years). We examined 5 right eyes and 12 left eyes from a total of 12 donors (10 male

and 2 female). The average postmortem time before eyes were perfused was 15.7 +1.3

hours.

The mean age of donors in the disease group was 64.8 +2.5 (age range 56 to 79 years).

We examined 4 right eyes and 6 left eyes from a total of 6 donors (5 male and 1

female). The average postmortem time before eyes were perfused was 12.2 + 1.4

hours. There was a significant difference between the mean age of donor eyes (P =

0.005).

3.3.2 General

All orders of retinal microvasculature were clearly identifiable by our perfusion

labelling technique. The position and morphology of separate capillary networks in

disease eyes were similar to what we observed in control eyes.5, 6 We identified 4

separate capillary networks in disease eyes corresponding to the following regions

within the retina:

1. The nerve fibre layer (NFL)

2. The retinal ganglion cell/superficial inner plexiform layer (RGC/sIPL)

3. The deep inner plexiform layer/superficial inner nuclear layer (dIPL/sINL)

4. The deep inner nuclear layer (dINL)

A total of 1348 capillary diameter and 315 capillary loop area measurements were

made in control eyes. A total of 1547 capillary diameter and 280 capillary loop area

measurements were performed in disease eyes.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 3.3.3 Morphometric Characteristics of Capillary Networks in Disease Eyes

Morphometric characteristics of the 4 different capillary networks in diseased eyes are

shown in Figure 13.

1. NFL capillary network

Capillaries of the NFL network were characterized by a predominance of capillary

segments that were oriented parallel to the direction of retinal ganglion cell axons (Fig.

13B). Short interconnecting segments bridged these capillaries in a diagonal or

perpendicular fashion. Mean capillary diameter in the NFL network was 8.64 ±

0.06µm. Mean capillary loop area was 2496.36 ± 713.48µm2. Mean capillary loop

length in the NFL network was 83.25 ± 17.14µm. Capillary density in the NFL

network was 12.78% ± 1.66%. Mean relative occupied capillary surface area was

13.62% ± 2.31%. NFL capillary measurements from donors are presented in Table 4.

2. RGC/sIPL capillary network

Capillaries of the RGC/sIPL network were characterized by a dense meshwork of 3D

vessels (Figure 13C). Reduced inter-capillary distance was a predominant feature of

this network with capillary segments demonstrating sharp diagonal and orthogonal

branching patterns. Mean capillary diameter in the RGCL/sIPL network was 8.93 ±


length in the RGCL/sIPL network was 62.41 ± 3.10µm. Capillary density in the

RGCL/sIPL network was 25.46% ± 1.76%. Mean relative occupied capillary surface

area was 40.55% ± 4.44%. RGCL/sIPL capillary measurements from donors are

presented in Table 4.

3. dIPL/sINL capillary network

Capillaries in the dIPL/sINL network were highly tortuous and demonstrated irregular

shaped loop configurations (Figure 13D). Mean capillary diameter in the dIPL/sINL

network was 8.96 ± 0.06µm. Mean capillary loop areas was 2813.66 ± 427.61µm2.

Mean capillary loop length in the dIPL/sINL network was 73.86 ± 6.49µm. Capillary

density in the dIPL/sINL network was 17.77% ± 1.55%. Mean relative occupied

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan capillary surface area was 19.35% ± 2.79%. dIPL/sINL capillary measurements from

donors are presented in Table 4.

4. dINL capillary network

Capillaries of the dINL network were characterized by a predominantly one

dimensional laminar configuration where capillaries ran a trajectory with little

tortuosity (Figure 13E). Mean capillary diameter in the dINL network was 9.10 ±


length in the dINL network was 147.86 ± 9.46µm. Capillary density in the dINL

network was 21.55% ± 1.38%. Mean relative occupied capillary surface area was

26.48% ± 1.83%. dINL capillary measurements from donors are presented in Table 4


Figure 13: Perifoveal Capillary Networks in an eye with Vascular Comorbidities

Three dimensional reconstruction of human perifoveal capillary networks in an eye with vascular comorbidities (A). Confocal capillary images captured from a single laser channel demonstrates the NFL network (B), RGC/sIPL network (C), dIPL/sINL network (D) and dINL network (E). Arrow indicates direction from vitread to sclerad surface. Scale bar = 100 µm.


Table 4 – Quantitative capillary network data for individual donors. Mean capillary diameter, capillary loop area, capillary density and

relative capillary surface area measurements for donor eyes is provided. SE = Standard error. NFL = Nerve Fibre Layer, RGCL = Retinal

Ganglion Cell Layer, sIPL = superficial Inner Plexiform Layer, dIPL = deep Inner Plexiform Layer, sINL = superficial Inner Nuclear

Layer and dINL = deep Inner Nuclear Layer. Data unable to be obtained is designated (-)

Specimen Capillary diameter (µm) Capillary loop area (µm2) Capillary density (%) Relative capillary surface area (%)

NFL RGCL/sIPL dIPL/sINL dINL NFL RGCL/sIPL dIPL/sINL dINL NFL RGCL/sIPL dIPL/sINL dINL NFL RGCL/sIPL dIPL/sINL dINL A 8.9 9.0 9.0 - 10164.8 1473.4 5453.3 - 17.7 29.5 22.8 - - - - -

A 8.7 9.1 9.6 - 0.0 2513.5 4244.1 - 14.0 27.4 25.4 - - - - -

B 8.6 9.0 8.2 8.6 1802.0 1308.7 3878.1 6287.1 21.2 23.1 14.8 20.4 22.8 30.5 21.5 25.1 B 9.6 8.7 9.2 8.5 3041.8 1084.0 3794.5 5661.9 14.1 34.3 18.9 18.6 13.9 47.8 15.4 22.8 C 6.8 8.1 8.8 - 2434.9 1551.8 3186.9 - 4.3 18.3 13.8 - - - - -

D 8.6 9.4 9.4 9.7 623.9 1834.1 3420.0 5123.7 15.4 27.6 17.2 29.4 17.3 31.7 20.6 30.4 D 9.1 8.9 9.0 9.9 0.0 1521.1 824.4 5793.9 14.6 28.2 18.1 29.8 10.5 55.8 13.4 20.3 E 9.2 9.3 - - 0.0 5429.6 - - 12.6 14.9 - - - - - -

F 8.3 8.6 8.6 8.8 3180.5 1967.6 0.0 11025.0 7.2 26.3 9.9 17.8 7.8 46.5 13.8 31.9 F 9.1 8.8 8.8 9.0 1732.6 2606.0 1416.1 4038.5 6.7 24.8 18.9 21.2 9.4 30.9 31.4 28.2

Mean 8.64 8.93 8.96 9.10 2496.36 1734.01 2813.66 6004.32 12.78 25.46 17.77 21.55 13.62 40.55 19.35 26.48

SE 0.06 0.07 0.06 0.07 713.48 141.66 427.61 526.20 1.66 1.76 1.55 1.38 2.31 4.44 2.79 1.83


3.3.4 Comparisons between Capillary Networks in Disease Eyes

Age was not associated with capillary diameter, loop area, loop length, density or

occupied capillary surface area measurements in any of the control (P > 0.107) or

disease networks (P > 0.172). Thus, age was removed from our multivariate model for

all subsequent comparisons. Capillary diameter was smallest in the NFL network and

was significantly smaller than the RGC/sIPL network (P = 0.001) and dIPL/sINL

network (P = 0.002) but not the dINL network (P = 0.062). There was no difference in

capillary diameter between the RGC/sIPL network and dIPL/sINL network (P = 0.489).

There was also no difference in capillary diameter between RGC/sIPLand dINL


Capillary loop area was smallest in the RGC/sIPL network and was significantly

smaller than in the dIPL/sINL and dINL networks (all; P< 0.003), but not the NFL

network (P= 0.086). Capillary loop area was largest in the dINL network and was

significantly larger than the NFL, RGCL/sIPL and dIPL/sINL networks (all P< 0.003).

Capillary loop area was not significantly different between the NFL and dIPL/sINL


Capillary loop length was largest in the dINL network and was significantly larger than

the NFL, RGCL/sIPL and dIPL/sINL networks (all P< 0.003). Capillary loop length

was smallest in the RGC/sIPL network and was significantly smaller than the dINL (P<

0.001), but not the NFL network (P= 0.062) or dIPL/sINL network (P= 0.035).

Capillary loop length was not significantly different between the NFL and dIPL/sINL


Capillary density was smallest in the NFL network and was significantly smaller than

the RGC/sIPL network (P = 0.001) and dINL network (P = 0.011), but not the

dIPL/sINL network (P = 0.043). Capillary density was significantly smaller in the

dIPL/sINL network than in the RGC/sIPL network (P = 0.001). There was no

difference in capillary density between the dINL network and the RGC/sIPL network

(P = 0.075) and dIPL/sINL network (P = 0.017).

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Relative occupied capillary surface area was greatest in the RGCL/sIPL network and

was significantly greater than in the NFL (P = 0.004) and dIPL/sINL(P = 0.012)

network, but not in the dINL network(P = 0.038). Relative occupied capillary surface

area was significantly larger in the dINL than in the NFL (P = 0.010). There was no

difference in capillary surface area between the dIPL/sINL and the NFL (P = 0.152) or

the dINL network (P = 0.055).

3.3.5 Comparisons between Disease and Control Eyes

Morphometric comparisons between control and disease eyes are displayed as bar-

charts in Figure 14. Morphometric comparisons between control and disease eyes for

NFL, RGC/sIPL, dIPL/sINL and dINL networks are provided in Figure 15.

Comparisons of the three-dimensional characteristics of these networks are provided in

figure 16.

1. NFL capillary network

There was no difference in NFL capillary diameter between disease and control groups

(P = 0.469). Capillary loop area (P = 0.020), capillary loop length (P = 0.009) and

density (P = 0.033) was reduced in disease eyes (Figs. 15A, 15B and 16A) . There was

no difference in relative occupied capillary surface area between disease and control

groups (P = 0.279).

2. RGC/sIPL capillary network

Capillary diameter in the RGC/sIPL network in the disease group was greater than

control eyes (P = 0.005). Disease eyes had smaller capillary loop area (P = 0.001) and

capillary loop lengths (P = 0.003) however there was no significant difference in

capillary density (P = 0.271) between the two groups (Figs. 15C, 15D and 16B). There

was no difference in relative occupied capillary surface area between disease and

control groups (P = 0.218).

3. dIPL/sINL capillary network

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan There was no difference in dIPL/sINL capillary diameter between disease and control

groups (P = 0.065). Disease eyes had smaller capillary loop areas (P = 0.008) and

capillary loop lengths (P = 0.005) but there was no significant difference in capillary

density (P = 0.173) between the two groups (Figs. 15E, 15F and 16C). There was no

difference in relative occupied capillary surface area between disease and control

groups (P = 0.061).

4. dINL capillary network

There was no difference in dINL capillary diameter between disease and control groups

(P = 0.073). There was no difference in capillary loop area (P = 0.863), capillary loop

length (P = 0.444), capillary density (P = 0.075) and relative occupied capillary surface

area (P = 0.440) between the two groups (Figs. 15G, 15H and 16D).


Figure 14: Quantitative Comparisons of Capillary Network Morphometry

Comparisons between control and disease eyes for capillary diameter (A), capillary loop area (B), capillary density (C) and relative occupied capillary surface area (D) are presented. Mean ± standard error for each parameter is provided. * denotes a significant difference between control and disease eyes (P < 0.05).


Figure 15: Morphometric Comparisons between control and disease eyes

NFL (A and B), RGC/sIPL (B and C), dIPL/sINL (E and F) and dINL (G and H) networks. Some capillary loops in each image are shaded in blue to allow comparison of capillary loop area between the two groups. Scale bar = 100 µm.


Figure 16: Comparison of Three-dimensional Capillary Network Morphometry between control and disease eyes

There was no difference in relative occupied surface area between the two groups in the NFL network (A and B), RGC/sIPL network (C and D), dIPL/sINL network (E and F) and dINL network (G and H). Scale bar = 200 µm.


The major findings from this study are as follows: (1) Cardiovascular comorbidities

induce alterations to capillary diameter, capillary loop area and capillary density

measurements within the human perifovea. (2) Capillary networks are altered in a non-

uniform manner by cardiovascular comorbidities. Specifically, capillary diameter

increased in the RGC/sIPL network. Capillary loop length and capillary loop area

decreased in all networks other than the deep INL network in patients with

cardiovascular comorbidities. Capillary density was reduced in the NFL network only.

Capillary endothelia maintain intimate relationships with neurons and glia cells and

thereby modulate retinal homeostasis.49, 146, 167, 178, 179 The morphometric organization

and position of capillary networks, within the layered retina, is believed to correlate

with the metabolic demands of regional neuronal subtypes.10, 49, 168, 173, 179-182

Quantifiable parameters such as capillary loop area, capillary loop length, capillary

density and capillary diameter convey vital information about the vasculogenic

mechanisms that are involved in retinal nutrition.20, 45, 95, 152 Through the complex

process of neurovascular coupling, capillary diameter is linked to mean erythrocyte

transit time through a microcirculation and contributes to the tissue oxygen extraction

fraction.181, 182 Similarly, capillary loop area and density measurements allow inference

about diffusion mechanisms that are involved in satisfying regional energy demands.22,

95, 168, 173, 183

Cardiovascular comorbidities, including ischemic heart disease, hypertension, smoking

and and hypercholesterolemia have been implicated in the pathogenesis of retinal

vascular disease.80, 82, 84, 85, 98-100, 184, 185 Detailed histo-pathological studies have

demonstrated alterations to endothelial tight junctions, basement membranes and mural

cells in patients with cardiovascular disease.82, 83, 98 The net effect of these histological

changes is the breakdown of the BRB. Clinical manifestations of BRB perturbations

include microaneurysms, exudates, haemorrhages, oedema and nerve fibre layer

ischemia.85, 95, 186-188 When present in the macula or perifovea, these changes can result

in rapid and irreversible vision loss.189 Although transverse histological investigations

have been used to study circulatory changes in the eye,119, 190, 191 to our knowledge,

two- and three-dimensional quantitative techniques have not been employed to quantify

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan disease-induced capillary network alteration in the human retina. The chronological

relationship between retinal capillary network alteration and endothelial dysfunction,

whether it is antecedent or subsequent, has also not been explored. This information

may be important for understanding patho-physiological mechanisms involved in

retinal vascular diseases.

In this study, capillary loop area measurements were used to speculate upon the role of

diffusion mechanisms in regional cellular nutrition. Mathematical studies have shown

that steady state oxygen diffusion in vascular networks with an arbitrary geometry can

be modelled to a three-dimensional space.192-194 Therefore, although the two- and

three-dimensional configuration of capillary loops were different between networks, it

is likely that the functional properties of each loop with regards to oxygen diffusion are

similar. The present study demonstrated a decrease in capillary loop area in all retinal

networks, other than the dINL, in patients with cardiovascular comorbidities.

Decreasing intercapillary areas may result in decreased oxygen diffusion times and may

therefore be an important compensatory mechanism by which the microcirculation

ensures adequate cellular nutrition in early retinal disease.22, 95 The proximity of the

deep inner nuclear layer to the outer retina, where the choroid is the predominant

supplier of metabolic substrates,191 may have been the reason why capillary loop areas

were not altered in this network.

In the capillary network subserving the retinal ganglion cell and superficial inner

plexiform layer, the mean diameter of capillaries in patients with cardiovascular

comorbidities is increased. These findings are similar to what has been described in

cerebro-cortical models of hypoxia.193 In the parietal cortex, experimental hypoxia

induces heterogeneous changes to capillary network flow and also an increase in red

blood cell velocity.195 Disease-induced increases in capillary diameter are speculated

to underlie these changes.196 Capillary diameter in the brain has also been shown to

increase in hypercapnic hyperemia. 196 197, 198 The net effect of increasing capillary

diameter is an increase in blood flow and thus improved tissue oxygenation.199 Such an

increase in capillary diameter may explain, at least to some degree, the 'non-difference'

in corresponding capillary density and occupied surface area measurements between

disease and control eyes. Patients with cardiovascular comorbidities may experience

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan some degree of retinal hypoxia relative to normal eyes. We speculate that in the

earliest stages of hypertension and other cardiovascular comorbid disorders, a

compensatory increase in capillary diameter may be one means of satisfying neuronal

demands in a relatively hypoxic environment before later clinical disease such as

decreased arteriolar diameter becomes manifest.85

The molecular basis for capillary network alteration in patients with cardiovascular

comorbidities may be the upregulation of cytokines and growth factor gradients within

the retina.83, 166, 200 Vascular endothelial growth factor is a major stimulator of

endothelial movement and an important determinant of capillary network morphology

in embryogenesis.201, 202 Other growth factors involved in vasculogenesis include

fibroblast growth factor and transforming growth factor beta.203, 204 We speculate that

upregulation of growth factors in the retina in patients with cardiovascular

comorbidities, possibly due to hypoxia, may underlie the process of capillary network

alteration. Disease-induced VEGF expression in the retina is non-uniform, being

predominantly expressed by Muller cells, endothelia, astrocytes, retinal pigment

epithelium and retinal ganglion cells.205 The non-uniform pattern of capillary network

alteration in patients with cardiovascular comorbidities may bear important correlations

to the predominant cell type that is activated to produce VEGF in the presence of

cardiovascular disease. Further work, however, is required to explore these

correlations.

Micro-cannulation techniques employed in this study allowed the complete labelling of

the retinal circulation and permitted us to reliably quantify the morphometric features

of capillary networks in the perifovea. The results of this study suggest that capillary

network morphometry is altered before clinical signs of BRB breakdown are evident in

patients with cardiovascular comorbidities. Our previous study has compared the

morphometric features of human retinal capillary networks between images derived

from fluorescein angiography and post mortem histopathological slides.119 This

previous study has shown that capillary details derived from fluorescein angiography

and fundus photography is limited to the inner capillary networks with very little

information concerning the deep networks being present in images derived by these

modalities. The density of capillary structures in fluorescein images is also

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan significantly lower than images derived from the same retinal eccentricity by confocal

scanning laser microscopy techniques using post-mortem tissue. Therefore, current

imaging modalities that are employed in clinical practice do not allow a layered

analysis of human capillary networks.119 However, the emergence of high penetration

optical coherence tomography technology29 may permit non-invasive means of

studying distinct capillary networks in the human eye and may be useful for the early

detection of capillary alteration in patients with cardiovascular comorbidities.

Application of such technologies to monitor the temporal sequence of capillary network

alteration in at-risk patients may also allow the timely intervention of therapy.

Although this experiment provides valuable information concerning the pathogenesis of

retinal vascular disease there were several limitations of this study. The sample size in

this study was too small to separate the effects of individual vascular comorbid disease

on capillary network morphometry. The spectrum of systemic disease severity for each

donor was also unknown. Furthermore, if the sample size was larger, it would have

been useful to compare the morphometric features of capillary networks between

treated and untreated patients in the disease group.

It is also possible that some of the control donors were suffering from undiagnosed

vascular comorbidities at the time of death and this remains a limitation with all post

mortem histopathological studies. The findings of this study are also limited to one

retinal eccentricity. It is also possible that capillary network alteration is specific to the

form of cardiovascular insult and retinal eccentricity. The purpose of this study was to

investigate the effects of cardiovascular risk factors on the retinal microcirculation and

although there was no documented evidence of cardiovascular comorbidity in the

control group, some of the patients in this group may not have been healthy at the time

of death. This in turn may have had some influence on the results of this study. Other

limitations of this study include the potential of post mortem alterations to vascular

morphology and the long post mortem time, of greater than 12 hours, in 4 patients in

the disease group.


Chapter 4.0: Validating the utility of speckle variance optical coherence tomography for quantitatively analysing human perifoveal capillary networks

4.1 Introduction

There remains a need for development of modern non invasive techniques that may

provide meaningful and reliable imaging of capillary morphology and structure.111

Speckle variance ocular coherence tomography (svOCT) represents a modern non-

invasive technique with increased imaging speed and performance that may provide

retinal capillary information in real time.142, 143, 206 The amount of capillary detail that

can be derived from individual retinal networks, however, remains unknown. This

experiment utilizes a custom-built svOCT device to qualitatively and quantitatively

study the various capillary networks in the human perifovea. A prototype speckle

variance optical coherence tomography system based on a wavelength-swept source

engine was used to generate en face images of retinal capillary networks. This study

aims to document qualitative and quantitative aspects of human capillary

microvasculature derived from this svOCT technique compared to our previous

histological data. Understanding the capillary detail that can be reliably extracted will

form the basis for expansion of such techniques for clinical diagnostic and investigative

purposes.

In this report, we used a prototype svOCT device to quantitatively and qualitatively

study perifoveal capillary networks in healthy human eyes. We first illustrate how this

device has the capacity to image and isolate distinct capillary networks in the perifovea.

Quantitative measurements derived from this device are then validated by making

comparisons between svOCT and age-matched histology data. These results are used

to illustrate the advantages and limitations of this device for studying retinal vascular

diseases.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 4.2 Methods


Western Australia and The University of British Columbia. All live-patient imaging

was performed at the Eye Care Centre in Vancouver. All human tissue was handled

according to the tenets of the Declaration of Helsinki.

4.2.1 Live Human Subjects

17 eyes from 9 healthy human subjects, aged between 28 to 60 were imaged using

svOCT. In this study, the group from which speckle variance OCT measurements were

made are referred to as the 'svOCT' group and eyes from the confocal microscopy

images as the 'histology' group.


Eleven human eyes from 8 donors, aged between 22 and 66 were used for this study.

Donor eyes used for this research had no documented history of eye disease. All eyes

were obtained from the Lions Eye Bank of Western Australia (Lions Eye Institute,

Western Australia) after removal of corneal buttons for transplantation. The donor eyes

used in this report were also used in our previous publication.177 The demographic

data of human subjects used for svOCT imaging and human eye donors for confocal

microscopy are presented in Table 5.

4.2.3 svOCT Imaging technique

Speckle variance OCT images of human subjects were acquired from a graphics

processing unit-accelerated svOCT clinical prototype.143 The details of the acquisition

system have previously been published.207 In short, the OCT system was based on a

1060nm swept-source with 100 kHz A-scan rate. The axial resolution was ~ 6µm in

tissue (based on the coherence length of the Axsun swept-source) and the lateral

resolution was estimated to be ~8.5µm based on the Gaussian waist of the focused spot

on the retina. Real time processing of the OCT intensity image data was performed

using our open source GPU algorithm.208 The retinal region that was imaged was ~

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 3mm nasal to the fovea, evaluated based on real-time volumetric OCT images (Figure

17A). For the speckle variance calculation, three repeat acquisitions at each B-scan

location were acquired. The scan area was sampled in a 300x300(x3) grid with a ~ 1x1

mm field of view in 3.15 seconds (Figure 18). En face visualization of the retinal

microvasculature was processed and displayed in real-time using our open source GPU

algorithm for svOCT.143 Scan dimensions were calibrated based on the eye length of

each participant, measured using the Zeiss IOLMaster 500 (IOL master 500, Carl Zeiss,

Jena, Germany). Images were acquired from both eyes of 9 patients.


Patient

ID Sex Age Eye

A M 36 R+L

B F 36 R+L

C F 28 R+L

D F 30 R+L

E M 31 R+L

F M 59 R+L

G F 60 R+L

H M 56 R+L

I F 52 R

Table 5 – Donor demographic details. Human subjects used for this research had no documented medical comorbidities. Age (years), sex (M = male or F = female), eye laterality (R = right or L = left)..


Figure 17: Speckle Variance Optical Coherence Tomography derived Cross Section of the Human Perifovea

Speckle variance optical coherence tomography (A) and a corresponding transverse confocal microscopy retinal section (B) from a healthy subject illustrate the various retinal layers in the perifoveal region. On the confocal image, endothelial cells are labelled with Phalloidin and nuclei are labelled with Hoescht antibody. Retinal layers corresponding to the nerve fibre layer (yellow), ganglion cell layer (green), inner nuclear layer (light blue) and outer nuclear layer (red) are clearly demarcated in each image. Capillary networks were co-localised within these retinal layers. Scale bar = 50 µm.


Figure 18: Generation of Speckle Variance Optical Coherence Tomography Images

Schematic demonstrating how speckle variance optical coherence tomography images were generated. Three repeat acquisitions at each B-scan location were acquired. The scan area was sampled in a 300x300(x3) grid with a ~ 1x1 mm field of view to generate volumetric data. En face visualization of the retinal microvasculature was processed and displayed in real-time using an open source GPU algorithm. (Adapted from Xu et al, 2014)


4.2.4 Tissue Preparation for Confocal Microscopy

Our method of micro-cannulation and targeted perfusion based labelling techniques

was utilized to label the retinal microvasculature in human donor eyes.47, 146 The

central retinal artery was cannulated and the retinal circulation perfused for 20 minutes

with a mixture of oxygenated Ringer’s solution and 1% bovine serum albumin.

Fixation with a solution of 4% paraformaldehyde in 0.1 M phosphate buffer was

followed by permeabilisation of endothelial cell membranes by a 0.1%Triton-X-100 in

0.1 M phosphate-buffered solution. Detergent was removed from the retinal circulation

by perfusion with 0.1M phosphate buffer solution. Endothelial microfilaments were

labelled over 2 hours by perfusion with a solution comprising of either phalloidin

conjugated to Alexa Fluor 546 (30 U; A22283; Invitrogen, Carlsbad, CA) or Lectin-

TRITC (Sigma L5266, 1:40). Nuclei were labelled with bisbenzimide (1.2 µg/mL;

Sigma-Aldrich, St. Louis, MO). Residual label was cleared from the vasculature by

further perfusion with 0.1 M phosphate buffer. Labelled specimens were immersion

fixed in 4% paraformaldehyde overnight prior to dissection and flat mounting.

4.2.5 Confocal Microscopy techniques

Images were captured from the retinal eccentricity that was located 2mm nasal to the

center of the fovea (Figure 17B). Confocal images were captured using a Nikon C1

Confocal equipped with three lasers (wavelengths 408 nm, 488 nm and 561nm) and

EZ-C1 (v. 3.20) image acquisition software. Images of different wavelengths were

acquired by sequential laser excitation. A 20x dry objective lens (NA 0.4) was used for

all scans capturing a 636.5 by 636.5µm region of interest. Using a motorised stage, a

series of z-stacks were captured for each specimen beginning from the vitreal surface,

at the level of the inner limiting membrane, to the outer retina. Each z-stack consisted

of a depth of optical sections collected at 0.35 µm increments along the z-plane.

4.2.6 Image Preparation

ImagePro Plus (Media Cybernetics, Version 7.1) software was used to quantify

confocal microscope images. All images for the manuscript were prepared using

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Adobe Photoshop (version 12.1, Adobe Systems Inc.) and Adobe Illustrator CS5

(version 15.1.0, Adobe Systems Inc.).


Our previously defined morphometric criteria, that has been validated by inter-observer

correlation studies, was used to partition the retinal circulation into separate capillary

networks.5, 19, 153, 177 Capillary branching patterns, trajectory and position respective to

nuclear layers were used to stratify the retinal circulation into 4 separate networks

corresponding to the following regions within the retina:

1. The nerve fibre layer (NFL)

2. The retinal ganglion cell/superficial inner plexiform layer (RGC/sIPL)

3. The deep inner plexiform layer/superficial inner nuclear layer (dIPL/sINL)

4. The deep inner nuclear layer (dINL)

The svOCT images were processed using the GPU-accelerated program,143 and the

retinal layers were automatically segmented using a 3D Graph Cut based segmentation

tool implemented in Matlab.209 En face retinal network images were generated from

the same 4 capillary networks using the morphometric criteria described above. The en

face retinal network images were generated by summing the svOCT data along the

depth direction within each segmented retinal network. A region of interest

corresponding to 636.5µm by 636.5µm was cropped from en face images permitting

direct comparison between svOCT and histology images.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 4.2.8 Morphometric Quantification of Capillary Networks

Our previously published methods were used to quantify the morphometric features of

capillary networks.177, 210 Quantitative data from capillary networks were attained

following z projection of all images between the first and final image slice for each

network. Manual tracing methods using ImagePro Plus (Media Cybernetics, Version

7.1) was employed to attain the following quantitative measurements from each

capillary network (Fig. 19):

• Capillary diameter- Defined as the perpendicular distance across the maximum

chord axis of each vessel. Each confocal image was partitioned into 9 equal

regions (Fig. 19A) and measurements were obtained from each region to ensure

representative sampling. An average of 45 measurements was obtained from

each image.

• Capillary density- Reflected by measurements including number of vessel

intersections per 100µm and intercapillary distance. A 3x3 grid consisting of

two equally spaced horizontal and vertical perpendicular line segments was

superimposed over images. Manual counts of capillary intersections across the

4 line segments covering a distance of 636.5µm were recorded (Fig. 19B) and

expressed as an average number of capillary intersections per 100µm.

Intercapillary distance was based on an average of 4 inter capillary distance

measurements calculated by dividing line segment distance by number of vessel

intersections.


All data is expressed in terms of mean and standard error which were calculated using

Sigmastat (Sigmastat, ver. 3.5; SPSS, Chicago, IL). The svOCT cohort data was

initially analyzed followed by comparisons with histology groups. For these

comparisons, T-testing was performed with a P value of < 0.050 considered significant.

Similarly, Mann-Whitney rank sum test was utilized with a significant value of P <

0.050 where normality failed. SigmaPlot (Sigmaplot, ver. 12.0; Systat Software, Inc.,

San Jose, CA) was used to create the bar graphs that are presented in this manuscript.


Figure 19: Methodology for Quantifying Capillary Network Morphometry

Representative speckle variance optical coherence tomography en face images of the deep inner nuclear layer network are used to illustrate how capillary diameter (A) and capillary density measurements (B) were acquired. Each image was first partitioned into a 3x3 grid (solid yellow lines). Capillary diameter (red lines) was measured by determining the perpendicular distance across the maximum chord axis for each vessel. Capillary density was calculated by determining the number of vessel intersections (red squares) per 100µm of line segment. Intercapillary distance was also used as an indice of capillary density. Scale bar = 100 µm.


4.3 Results

4.3.1 Demographics

The mean age of subjects in the svOCT group was 43.11 ± 4.46 years. We examined 9

right eyes and 8 left eyes from a total of 9 subjects (4 male and 5 female). One eye was

excluded due to significant motion artefact.

The mean age of donors in the histology group was 40.12 ± 6.05 years. We examined

3 right and 8 left eyes from a total of 8 male donors. There was no statistically

significant difference between the mean age of svOCT and histology eyes (P = 0.630).

4.3.2 Morphometric Comparisons Between svOCT and Confocal Eyes

All retinal microvascular networks were visualized with the svOCT imaging

technique. Morphometric characteristics of the 4 different capillary networks between

svOCT and histology techniques are shown in Figure 20. Quantitative comparisons

between svOCT and histology images are displayed as bar-charts in Figure 21.

Although retinal ganglion cell axons could not be distinguished by this technique, NFL

capillaries in the svOCT images were seen to be predominantly oriented parallel to one

another as in our previous descriptions of this network. svOCT-imaged capillaries of

the RGC/sIPL network were closely approximated to the larger vessels and comprised

a dense meshwork. svOCT-imaged capillaries of the dIPL/sINL network were

associated with a prominent homogeneous back-scattered signal and ran a tortuous

trajectory. In contrast, the dINL was the most easily discernible network, comprised of

a flat network of capillaries with large sloping capillaries in loop configurations.

In general, the svOCT retinal layer images closely approximated those derived from

histological techniques. Across all svOCT-imaged retinal networks, background was

noisier and capillary margins were less well defined.


Figure 20: Morphometric Comparisons between svOCT and Histology derived Images

Representative images of NFL network (A and B), RGC/sIPL network (C and D), dIPL/sINL network (E and F) and dINL network (G and H) for eyes are provided. Scale bar = 100 µm.


Figure 21: Quantitative Comparisons of Capillary Network Morphometry

Comparisons between speckle variance optical coherence tomography and histology images for capillary diameter (A), capillary intersections per 100µm (B) and intercapillary distance (C) are presented. Mean ± standard error for each parameter is provided. *denotes a significant difference between svOCT and histology eyes (P < 0.050).

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 4.3.3 Quantitative Analysis of Capillary Diameter

A total of 1760 capillary diameter measurements were made in histology eyes and 3060

capillary diameter measurements were performed in svOCT eyes.

The average capillary diameter for all networks in the svOCT and histology techniques

was 10.05 ± 0.05 µm and 8.31 ± 0.06 µm respectively. Mean capillary diameters for

each retinal network for both svOCT and histology techniques are provided in Table 6.

There was no difference in retinal network capillary diameter within the histology

cohort (P > 0.216). Within svOCT images, there was a significant difference between

capillary network diameters between all networks (all P < 0.033) except the RGC/sIPL

and dIPL/sINL network (P = 0.565). Comparisons between svOCT and histology

showed a significantly increased capillary diameter for all retinal networks (all P <

0.001) as demonstrated in Figure 21A.


Table 6 - Quantitative retinal capillary network data acquired using svOCT and histology imaging techniques.

Mean capillary diameter, number of vessel intersections per grid, number of capillary intersections per 100µm and intercapillary distance

(ICD) with corresponding standard errors for each network is provided. NFL = Nerve Fibre Layer, RGCL = Retinal Ganglion Cell Layer,

sIPL = superficial Inner Plexiform Layer, dIPL = deep Inner Plexiform Layer, sINL = superficial Inner Nuclear Layer and dINL = deep

Inner Nuclear Layer.

Capillary Diameter (microns)

No. Vessel Intersections per grid

No. Capillary Intersections per 100µm ICD (microns)

svOCT NFL 9.86 + 0.06 37.5 + 1.32 1.47 + 0.04 73.46 + 3.61

RGCL/sIPL 10.01 + 0.05 38.47 + 1.49 1.51 + 0.04 70.94 + 3.43

dIPL/sINL 9.98 + 0.05 41.12 + 1.04 1.62 + 0.03 64.31 + 1.71

dINL 10.34 + 0.05 40.06 + 1.22 1.57 + 0.04 66.05 + 1.96

Histology NFL 8.32 + 0.06 35.80 + 2.97 1.41 + 0.12 79.21 + 8.73

RGCL/sIPL 8.28 + 0.07 40.80 + 1.93 1.60 + 0.08 65.60 + 2.89

dIPL/sINL 8.33 + 0.06 38.4 + 2.46 1.51 + 0.10 71.56 + 5.09

dINL 8.30 + 0.05 39.2 + 2.24 1.57 + 0.09 68.84 + 3.97


4.3.4 Quantitative Analysis of Capillary Density

The average number of capillary intersections per 100µm for all networks in the

svOCT and histology techniques was 1.54 ± 0.05µm and 1.51 ± 0.09µm respectively.

Mean capillary intersections per 100µm for each retinal network for both svOCT and

histology techniques are provided in Table 6. Capillary intersections per 100µm was

largest in the dIPL/sINL network on svOCT and was significantly larger than the NFL

network (P = 0.038). There was no difference in capillary intersections per 100µm

between the NFL, RGC/sIPL and dINL networks (all P >0.155). There were no

differences in capillary intersections per 100µm between capillary networks in the

histology specimens (all P > 0.196). The number of capillary intersections per 100µm

comparisons between svOCT and histology did not show a significant difference for all

retinal networks (all P > 0.157).

The intercapillary distances for all networks in the svOCT and histology techniques

was 68.69 ± 2.67µm and 71.30 ± 5.17µm respectively. Mean intercapillary distances

for each retinal network for both svOCT and histology techniques are provided in Table

6. Intercapillary distance was largest in the NFL network on svOCT and was

significantly larger than the dIPL/sINL network (P = 0.029). There was no difference

in intercapillary distance between the NFL, RGC/sIPL and dINL networks (all P

>0.080). There were no differences in intercapillary distance between capillary

networks in the histology specimens (all P > 0.177). Intercapillary distance

comparisons between svOCT and histology did not show a significant difference for all

retinal networks (all P > 0.094).


The major findings from this study are as follows: (1) Speckle variance optical

coherence tomography has the capacity to delineate the morphometric characteristics of

different capillary networks in the human perifovea. (2) The morphological

characteristics of capillary networks presented in svOCT images are highly comparable

to age-matched histology images. (3) Capillary diameter measurements in svOCT

images are significantly greater than histological data for all networks. (4) Capillary

density measurements are not significantly different between svOCT and histological

data for all networks.

Retinal neurons in the human perifovea are nourished by four morphologically distinct

capillary networks.177 We have previously shown that the anatomical configuration of

each capillary network is altered in accordance with the unique metabolic demands of

the regional neuronal milieu. The vulnerability of these networks to systemic

cardiovascular disease is not uniform. Our prior work has provided evidence that

hypertension and other cardiovascular co-morbidities induce selective alteration to

capillary networks in the perifovea.210 Therefore, in addition to providing vital

information about neurovascular coupling mechanisms that regulate retinal

homeostasis, the study of capillary networks is important for our understanding of

vascular patho-biology.

Speckle variance OCT is a non-invasive imaging device that is gaining greater

application in clinical medicine.138, 142 In ophthalmology, svOCT has been used to

acquire real-time, flow contrast images of the macula and optic nerve head.143 A major

advantage of svOCT is the ability to perform angiography without contrast

administration. This avoids many potential adverse effects that are commonly seen

with fluorescein angiography. The capability of svOCT to identify microvasculature

by calculating interframe intensity variation of structural images overcomes Doppler

angle limitation- the larger range of Doppler frequency shifts and increased spectral

broadening as beam-flow measurement angle increases. This also makes it less user-

dependent than Doppler OCT techniques.138, 211, 212 The manner in which the svOCT

data is collected and processed, repeat B-scans at the same location, also makes it more

sensitive to detecting flow contrast in regions of low flow. Finally, the relatively fast

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan image acquisition and processing time and the capacity to provide real-time images that

are immediately available to the clinician are other major advantages of svOCT.

The application of svOCT technology to image the clinical manifestations of age

related macular degeneration have previously been described, however to our

knowledge, this is the first report to validate and quantitatively compare svOCT to

histological data.213 Using our prototype svOCT device we were able to reliably image

the various capillary networks in the perifovea. Specifically, we were able to image the

NFL network, the RGC/sIPL network, the dIPL/sINL network and the dINL network

with great accuracy. Capillaries in the NFL network were seen to run parallel to the

direction of axons while capillaries in the dINL network demonstrated a planar

configuration with multiple closed loops. Capillaries in RGC/sIPL and dIPL/sINL

networks were observed to have a tortuous trajectory and displayed a complex three-

dimensional configuration. We found that the morphological characteristics of

networks in svOCT images were highly comparable to our histological images.

The mean capillary diameter was significantly greater in svOCT images compared to

age-matched histological images for all networks. This can be attributed to the

difference in lateral resolution of the two imaging systems; the lateral point spread

function (PSF) of the svOCT imaging beam can be modelled as a Gaussian with 1/e2

radius of 17 µm, artificially broadening the measured width of the capillaries. This

difference may also have been due to the effects of fixation artefact and shrinkage in

histology specimens.214 Other plausible explanations for increased capillary diameter

in svOCT images may be due to the dynamic effects of physiological variables such as

blood pressure, tissue interstitial pressure and vessel wall elasticity acting upon

capillary beds.215 These variables have been shown to influence the anatomical

dimensions of live tissue. Pericyte mediated reduction of capillary diameter has also

been shown in the context of neuronal death to influence capillary diameter and this

effect may also have resulted in smaller diameters in histological images.216

Investigation of these potential physiological confounding factors with a high

resolution svOCT system is left to future work.

In this experiment we used the number of capillary intersections per 100μm and the

inter-capillary distance as indices of capillary density. Inter-capillary distance

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan measurements in the retina using svOCT closely approximated what has previously

been reported in the human gray matter.217 The mean value of 68.69μm was

significantly less than 123μm which is the reported value of inter-capillary distance in

human white matter.218 Similar to our histology data, we found that capillary density

in svOCT images was greatest in the network bordering the plexiform layer and least in

the nerve fibre layer network. When comparisons were made between svOCT and

histology data we did not find a difference in capillary intersections per 100μm and

inter-capillary distance for any network.

This work validates the utility of our prototype svOCT for studying human perifoveal

capillary networks but we acknowledge several limitations of this device. Firstly,

patient co-operation and steady fixation are essential for acquiring high quality images

of capillary networks. Although the image acquisition time was only 2.7 seconds it

may not be possible for patients with significant visual disease to fixate for this period

of time. In these patients, motion artefacts may obscure important capillary detail.

Modifications to this device with shorter image acquisition times136, 219 or the

incorporation of motion tracking technology220, 221 may potentially overcome this

limitation. Secondly, the axial resolution of this device is 6 µm and the lateral spatial

resolution is 17µm. The resolving capability of this device is therefore still not as

precise as what is currently attainable histologically. Furthermore, the duration of an

imaging session, speed of the OCT system, and lateral resolution combine to limit the

field of view that can be acquired with this technique in a clinical setting.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Chapter 5.0: General Discussion

The human retinal capillary circulation is a complex vascular network that must be

architecturally designed to support neuronal layers without compromising the optical

properties of the light transmitting pathway.12 Our histological studies of human retina

have revealed variable morphology of capillary networks sub-serving distinct neuronal

layers with disparate metabolic demands. Multi-layered networks have been

demonstrated with a laminar, one-dimensional configuration noted for the innermost

and outermost networks in contrast to the tortuous, three-dimensional configurations

supporting the high synaptic activity of the inner plexiform layer.176, 177 Such an

arrangement is important considering that the high energy demands of the inner retina

is supplied from a single central retinal artery, with a retinal circulation that must

expand to supply an area greater than 1000 mm2.222 Maintaining capillary perfusion

and regulatory capabilities through regional variations in vascular pattern and

distribution ensures that the high density and functional requirements of the retinal

neuronal populations are satisfied.

The findings from our histological studies of healthy human perifoveal capillary

microvasculature have established that four morphologically distinct capillary networks

exist in this region with unique alterations suited to meet the metabolic demands of

neuronal sub-compartments.177 Denser and more complex vascular networks sub-serve

regions of higher oxygen consumption required for the high ATP synaptic activity of

the plexiform layer.

In our comparison between healthy eyes and those with cardiovascular co-morbidities,

quantifiable alterations in retinal capillary networks were found to occur. Specifically,

reductions in capillary diameter and capillary density corresponded to increased oxygen

diffusion distances, possibly stemming from capillary dropout and photoreceptor

loss.210 Although in vivo studies have qualified such disease induced capillary

network alterations in the human retina, to our knowledge, previously employed

techniques have not been used to quantitatively analyse changes to human capillaries in

their separate networks.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Our findings from the study of eyes with cardiovascular comorbidities also

demonstrated that capillary networks of the perifovea are altered in a non-uniform

manner with the presence of vascular comorbidities. With the exception of the

innermost retinal network, situated in the NFL, the diameter of capillaries increased.

Partial occlusion of small capillaries and capillary rarefaction are involved in the

pathophysiology of diabetes and the relatively smaller diameter of the NFL in patients

with vascular comorbidities may highlight a particular susceptibility to such disease

processes. This study had identified differential vulnerability of networks in vascular-

mediated disease and it for this reason that the ability to resolve individual capillary

networks is important.

Previous studies note that retinal hypertensive alterations mainly arise from the

superficial capillaries of the NFL.97 A smaller vessel diameter and findings of a

decreased NFL network density in relation to other capillary networks may also

contribute to a local increase in vascular resistance. This suggests that alterations to

perifoveal capillary blood flow may preferentially affect the NFL and may render this

capillary network particularly susceptible to vascular insults. We speculate that such

changes in vascular topography may stem from dysregulation in cytokine and local

growth factors which modulate capillary behaviour in retinal disease.83, 166, 200 In

response to hyperglycaemia and ischaemia, paracrine expression of neurotrophic and

angiogenic factors such as VEGF are upregulated.223 Other hypoxia induced molecular

transcription factors such as platelet derived growth factor (PDGF), insulin-like growth

factor 1(IGF-1), hepatocyte growth factor (HGF), basic fibroblast growth factor (bFGF)

have also been implicated in abnormal vascularization.224 In rats, the overexpression of

placental growth factor (rPGF-1) induced the dilation and increased tortuosity of retinal

vessels, the formation of microaneuryms and vascular sprouts.203 Microvascular cell

loss and progressive occlusion of capillaries results have been theorized to stem from

the induction of growth factors such as transforming growth factor beta (TGFβ).

Intercapillary areas, a reflection of the distance which oxygen and metabolites must

travel by diffusion were reduced in all eyes with cardiovascular comorbidities except in

the deepest capillary network of the inner nuclear layer. Decreasing intercapillary areas

result in decreased oxygen diffusion times, and we postulate that these morphological

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan alterations may represent compensatory mechanisms to factors limiting blood supply in

systemic vascular conditions such as findings of reduced capillary blood flow

velocities. This remains in keeping with other studies where perifoveal intercapillary

areas were significantly increased at baseline in patients with essential hypertension

compared to healthy subjects.225 The deep inner nuclear layer capillary network may

remain unchanged due to a relatively low oxygen demand in a region devoid of high

metabolic synaptic activity. Furthermore, the proximity of the dINL to the outer retina

where the choroid is the predominant supplier of metabolic substrates may also mean

that morphological alterations to compensate oxygen supply are not required until later

disease stages. Further work, however, is required to validate this hypothesis.

Limitations of the studies involving confocal microscopy of perifoveal networks

include the potential for postmortem artifact in influencing morphometric

measurements, in particular capillary diameter findings which may be subject to

postmortem induced vessel dilatation. However, the effects of these mechanisms are

expected to be equal in both health and disease. In our study of disease, capillary

networks are accounted for through standardization with a control group that had

identical times to perfusion. Our sample size encompassed a wide variety of vascular

comorbidities, constrained by the difficulty to obtain enough postmortem eyes to gather

sufficient power to explore relationships with single vascular diseases. These findings

are limited to one retinal eccentricity and it is likely that the capillary networks of the

peripheral retina may behave differently to the perifovea, reflecting a differing density

of the retinal neuropile and corresponding metabolic requirements.

The svOCT technique employed in the final study provided accurate in vivo

information about distinct capillary networks in the human eye and permitted us to

reliably quantify morphometric features of networks in the perifovea. Similar

appreciable morphometry and density measurements validate this technique with

remarkable concordance despite a complex three dimensional microvascular structure

reflecting a thick retinal ganglion cell and nerve fibre in the perifoveal region studied.

High sensitive vascular detection down to the capillary level is possible and

incorporation with relative depth information makes this an attractive tool to study the

dynamic relationship between neuronal activity and surrounding blood vessels. Also,

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan the ability of this technique to overcome doppler angle makes it less user dependent

than doppler optical coherence tomography techniques. Significant movement artefact

that can distort signal are improved by short data acquisition times. However, limited

resolution and depth of penetration limits its use in imaging of extensive microvascular

systems. Commercial availability and cost also restrict the use of this technique at

present.114, 226-229

Comparison of the svOCT derived images to histology identified uniformly larger

capillary diameters in svOCT derived images. Possible reasons for this include

differences between fixed versus dynamic live specimens with an active process of

dilation of capillaries proven in vivo in mouse cerebral cortex.216, 230 Other plausible

explanations for increased capillary diameter may be the result of other physiological

forces exerted in live specimens including fluctuations in blood pressure, tissue

interstitial pressure and vessel wall tension or elasticity.215 Pericyte mediated reduction

of capillary diameter has been shown in the context of neuronal death and may have

resulted in smaller diameters derived from histology.216 Post hoc analysis found that

the dIPL/sINL had the densest capillary network in contrast to significantly larger

intercapillary distance measurements established in the NFL. These results were

consistent with our previous findings that density was greatest for networks bordering

the plexiform layer and least in the nerve fibre layer. Such capillary density

information as a surrogate measure of capillary perfusion distance allows inferences to

made regarding the metabolic activity of tissue and identification of alterations in tissue

metabolic requirements in health and disease. The knowledge that capillary changes

may occur prior to development of ocular microvascular disease may aid in the

application of such novel in vivo technology to image the retinal microcirculation231

and allow earlier disease detection.228

A difficult distinction between homogeneous back-scattered signal was particularly

prominent in the RGC/sIPL and dIPL/sINL layers from the svOCT derived images

which is likely shadow artefact from the complex three dimensional arrangements in

these networks. Techniques to enhance separation of structures from tissue background

signal are being developed and it is likely this will improve resolution capabilities

toward the future.114, 137, 141, 206 It is not clear whether regional variability in the

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan intensity of the blood vessel signal as well as subtle differences in capillary diameter

along the trajectory of capillaries convey true information regarding haemodynamic

blood flow. Work remains to elucidate whether this technique is able to provide such

information on blood flow physiology and functionality. The foveal avascular zone

may represent an attractive area for study of this potential correlation given the variable

flow and relative decrease in retinal tissue thickness and complexity of the retinal

network in this region.103, 232

In summary, this thesis has furthered our understanding of the detailed qualitative and

quantitative characteristics of the human perifoveal microvasculature. We have

detected and speculated on why changes to specific capillary networks may occur in

isolation with certain disease states. The results of this thesis also suggest that svOCT

is a promising technique for the qualitative and quantitative assessment of capillary

networks in the human retina by allowing for noninvasive, real-time high resolution

volumetric imaging.

In this thesis, we have proven that capillary networks may be distinguished within their

distinct capillary networks in vivo and it will therefore be important to perform similar

in vivo morphometric studies using abnormal human eyes. These studies would be

aimed at distinguishing whether quantitative measurements are sensitive in detecting

early capillary changes in patients with retinal vascular disease such as capillary drop

out associated with diabetic retinopathy. Such early detection of changes in individual

capillary networks in vivo would also allow for a temporal understanding of disease

processes- whether capillary network changes precipitate or provide compensatory

responses in retinal vascular disease may be elucidated. Clinical utility could be

extended beyond risk stratification to the quantification and response monitoring of

capillary networks to various therapeutic modalities e.g anti-VEGF therapy. Such in

vivo information may validate the utility of this device for managing patients and prove

revolutionary for stratifying patients into high-risk groups for retinal disease

progression.


References

1. Tsaloumas MD, Kirwan J, Vinall H, et al. Nine year follow-up study of morbidity and mortality in retinal vein occlusion. Eye (Lond). 2000;14:821-827.

2. Laouri M, Chen E, Looman M, Gallagher M. The burden of disease of retinal vein occlusion: review of the literature. Eye (Lond). 2011.

3. Murray CJL, Lopez AD. The Global Burden of Disease: a comprehensive assessment of mortality and dissability from diseases, injuries and risk factors in 1990 and projected to 2020. In: Cambridge, MA: Harvard School of Public Health, Harvard University Press; 1996.

4. Snodderly DM, Weinhaus RS. Retinal vasculature of the fovea of the squirrel monkey, Saimiri sciureus: Three-dimensional architecture, visual screening, and relationships to the neuronal layers. Journal of Comparative Neurology. 1990;297:145-163.

5. Snodderly DM, Weinhaus RS, Choi JC. Neural-vascular relationships in central retina of macaque monkeys (Macaca fascicularis). J Neurosci. 1992;12:1169-1193.

6. Johnson MA, Lutty GA, McLeod DS, et al. Ocular structure and function in an aged monkey with spontaneous diabetes mellitus. Exp Eye Res. 2005;80:37-42.

7. Haverkamp S, Haeseleer F, Hendrickson A. A comparison of immunocytochemical markers to identify bipolar cell types in human and monkey retina. Vis Neurosci. 2003;20:589-600.

8. MacNichol EF, Jr. Retinal processing of visual data. Proc Natl Acad Sci U S A. 1966;55:1331-1344.

9. Cringle SJ, Yu DY, Yu PK, Su EN. Intraretinal oxygen consumption in the rat in vivo. Invest Ophthalmol Vis Sci. 2002;43:1922-1927.

10. Yu D-Y, Cringle SJ. Oxygen distribution and consumption within the retina in vascularised and avascular retinas and in animal models of retinal disease. Prog Retina Eye Res. 2001;20:175-208.

11. Kiryu J, Asrani S, Shahidi M, Mori M, Zeimer R. Local response of the primate retinal microcirculation to increased metabolic demand induced by flicker. Invest Ophthalmol Vis Sci. 1995;36:1240-1246.

12. Yu DY, Yu PK, Balaratnasingam C, Cringle SJ, Su EN. Microscopic structure of the retina and vasculature in the human eye. In: Méndez-Vilas A, Díaz J (eds), Microscopy: Science, Technology, Applications and Education: Formatex Research Center; 2010:867-875.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 13. Burns EM, Kruckeberg TW, Gaetano PK. Changes with age in cerebral

capillary morphology. Neurobiol aging. 1981;2:283-291.

14. HALE AR, REED AF. Studies in cerebral circulation. Methods for the qualitative and quantitative study of human cerebral blood vessels. Am Heart J. 1963;66:226-242.

15. Bell MA, Ball MJ. Morphometric comparison of hippocampal microvasculature in ageing and demented people: diameters and densities. Acta Neuropathol. 1981;53:299-318.

16. Cassot F, Lauwers F, Fouard C, Prohaska S, Lauwers-Cances V. A novel three-dimensional computer-assisted method for a quantitative study of microvascular networks of the human cerebral cortex. Microcirculation. 2006;13:1-18.

17. Sun XF, Zhang H, Wu XC, Han YM, Hou GQ, Xian MS. Microvascular corrosion casting of normal tissue, transitional mucosa and adenocarcinoma in the human colon. Acta Oncol. 1992;31:37-40.

18. Lambova SN, Hermann W, Muller-Ladner U. Comparison of qualitative and quantitative analysis of capillaroscopic findings in patients with rheumatic diseases. Rheumatol Int. 2011.

19. Richard E, van Gool WA, Hoozemans JJ, et al. Morphometric changes in the cortical microvascular network in Alzheimer's disease. J Alzheimers Dis. 2010;22:811-818.

20. Sposito NM, Gross PM. Morphometry of individual capillary beds in the hypothalamo-neurohypophysial system of rats. Brain Res. 1987;403:375-379.

21. Foreman DM, Bagley S, Moore J, Ireland GW, McCleod D, Boulton ME. Three dimensional analysis of the retinal vasculature using immunofluorescent staining and confocal laser scanning microscopy. Br J Ophthalmol. 1996;80:246-251.

22. Arend O, Wolf S, Jung F, et al. Retinal microcirculation in patients with diabetes mellitus: dynamic and morphological analysis of perifoveal capillary network. Br J Ophthalmol. 1991;75:514-518.

23. Aird WC, Rosenberg RD. Vascular Endothelium: Physiology, Pathology, and Therapeutic Opportunities New York: Schattauer; 1997.

24. Iadecola C. Neurovascular regulation in the normal brain and in Alzheimer's disease. Nat Rev Neurosci. 2004;5:347-360.

25. Paemeleire K. The cellular basis of neurovascular metabolic coupling. Acta Neurol Belg. 2002;102:153-157.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 26. Alder VA, Su EN, Yu D-Y, Cringle S, Yu PK. Overview of studies on

metabolic and vascular regulatory changes in early diabetic retinopathy. [Review] [26 refs]. Aust N Z J Ophthalmol. 1998;26:141-148.

27. Amin RH, Frank RN, Kennedy A, Eliott D, Puklin JE, Abrams GW. Vascular endothelial growth factor is present in glial cells of the retina and optic nerve of human subjects with nonproliferative diabetic retinopathy. Invest Ophthalmol Vis Sci. 1997;38:36-47.

28. Antonetti DA, Lieth E, Barber AJ, Gardner TW. Molecular mechanisms of vascular permeability in diabetic retinopathy. Semin Ophthalmol. 1999;14:240-248.

29. Makita S, Jaillon F, Yamanari M, Miura M, Yasuno Y. Comprehensive in vivo micro-vascular imaging of the human eye by dual-beam-scan Doppler optical coherence angiography. Opt Express. 2011;19:1271-1283.

30. Klaassen I, Van Noorden CJ, Schlingemann RO. Molecular basis of the inner blood-retinal barrier and its breakdown in diabetic macular edema and other pathological conditions. Prog Retin Eye Res. 2013;34:19-48.

31. Grover S, Murthy RK, Brar VS, Chalam KV. Comparison of retinal thickness in normal eyes using Stratus and Spectralis optical coherence tomography. Invest Ophthalmol Vis Sci. 2010;51:2644-2647.

32. Hogan MJ, Alvarado JA, Weddell JE. Histology of the Human Eye: An Atlas and Textbook Philadelphia: W.B. Saunders Company; 1971.

33. Nordenstrom BEW. Neurovascular activation requires conduction through vessels. Physiological Chemistry and Physics and Medical NMR. 1989;21:249-256.

34. Ward N, LaManna J. The neurovascular unit and its growth factors: coordinated response in the vascular and nervous systems. Neurological Research. 2004;26:870-883.

35. Rosenthal G, Hemphill JC, III, Sorani M, et al. Brain tissue oxygen tension is more indicative of oxygen diffusion than oxygen delivery and metabolism in patients with traumatic brain injury. Crit Care Med. 2008;36:1917-1924.

36. Hughson RL. Regulation of blood flow at the onset of exercise by feed forward and feedback mechanisms. Can J Appl Physiol. 2003;28:774-787.

37. Lecrux C, Hamel E. The neurovascular unit in brain function and disease. Acta Physiol (Oxf). 2011;203:47-59.

38. Ridge JW. The distribution of cytochrome-oxidase activity in rabbit brain. Biochem J. 1967;102:612-617.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 39. Zeller K, Rahner-Welsch S, Kuschinsky W. Distribution of Glut1 glucose

transporters in different brain structures compared to glucose utilization and capillary density of adult rat brains. J Cereb Blood Flow Metab. 1997;17:204-209.

40. Azzopardi P, ", Cowey A. The overrepresentation of the fovea and adjacent retina in the striate cortex and dorsal lateral geniculate nucleus of the macaque monkey. Neuroscience. 1996;72:627-639.

41. Dhande OS, ", Huberman AD, ". Retinal ganglion cell maps in the brain: implications for visual processing. Curr Opin Neurobiol. 2014;24:133-142.

42. Barton RA, ". Visual specialization and brain evolution in primates. Proc Biol Sci. 1998;265:1933-1937.

43. van Essen DC, ", Anderson CH, ", Felleman DJ. Information processing in the primate visual system: an integrated systems perspective. Science ". 1992;255:419-423.

44. Masland RH. The neuronal organization of the retina. Neuron. 2012;76:266-280.

45. Shaver SW, Sposito NM, Gross PM. Quantitative fine structure of capillaries in subregions of the rat subfornical organ. J Comp Neurol. 1990;294:145-152.

46. Tuor UI, Kurpita G, Simone C. Correlation of local changes in cerebral blood flow, capillary density, and cytochrome oxidase during development. J Comp Neurol. 1994;342:439-448.

47. Yu PK, Balaratnasingam C, Cringle SJ, McAllister IL, Provis J, Yu DY. Microstructure and network organization of the microvasculature in the human macula. Invest Ophthalmol Vis Sci. 2010;51:6735-6743.

48. Kong X, Wang K, Sun X, Witt RE. Comparative study of the retinal vessel anatomy of rhesus monkeys and humans. Clin Experiment Ophthalmol. 2010;38:629-634.

49. Yu DY, Cringle SJ, Yu PK, Su EN. Intraretinal oxygen distribution and consumption during retinal artery occlusion and graded hyperoxic ventilation in the rat. Invest Ophthalmol Vis Sci. 2007;48:2290-2296.

50. Yu D-Y, Cringle SJ, Alder VA, Su E-N, Yu PK. Intraretinal oxygen distribution and choroidal regulation in the avascular retina of guinea pigs. Am J Physiol. 1996;270:H965-H973.

51. Michaelson IC. Review of adult conditions of inner ocular vessels in the mammalian eye. In: Thomas CC (ed), Retinal Circulation in Man and Animal Illinois: Springfield; 1954:101-112.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 52. Schroder S, Brab M, Schmid-Schonbein GW, Reim M, Schmid-Schonbein H.

Microvascular network topology of the human retinal vessels. Fortschr Ophthalmol. 1990;87:52-58.

53. Huang JC, Voaden MJ, Zarbin MA, Marshall J. Morphologic preservation and variability of human donor retina. Curr Eye Res. 2000;20:231-241.

54. Chou JC, Rollins SD, Fawzi AA. Trypsin digest protocol to analyze the retinal vasculature of a mouse model. J Vis Exp. 2013:e50489.

55. Bhutto IA, ", Amemiya T. Retinal vascular architecture is maintained in retinal degeneration: corrosion cast and electron microscope study. Eye (Lond) ". 2001;15:531-538.

56. Amemiya T, Bhutto IA. Retinal vascular changes and systemic diseases: corrosion cast demonstration. Ital J Anat Embryol. 2001;106:237-244.

57. Paques M, Tadayoni R, Sercombe R, et al. Structural and hemodynamic analysis of the mouse retinal microcirculation. Invest Ophthalmol Vis Sci. 2003;44:4960-4967.

58. Varma DD, Cugati S, Lee AW, Chen CS. A review of central retinal artery occlusion: clinical presentation and management. Eye (Lond). 2013;27:688-697.

59. Yu PK, Balaratnasingam C, Cringle SJ, McAllister IL, Provis J, Yu DY. Microstructure and network organization of the microvasculature in the human macula. Invest Ophthalmol Vis Sci. 2010;51:6735-6743.

60. Iwasaki M, Inomata H. Relation between superficial capillaries and foveal structures in the human retina. Invest Ophthalmol Vis Sci. 1986;27:1698-1705.

61. Ahmed J, Braun RD, Dunn R, Linsenmeier RA. Oxygen Distribution in the Macaque Retina. Invest Ophthalmol Vis Sci. 1993;34:516-521.

62. Birol G, Wang S, Budzynski E, Wangsa-Wirawan ND, Linsenmeier RA. Oxygen distribution and consumption in the macaque retina. Am J Physiol (Heart & Circ Physiol). 2007;293:H1696-H1704.

63. Cringle SJ, Yu PK, Su EN, Yu DY. Oxygen distribution and consumption in the developing rat retina. Invest Ophthalmol Vis Sci. 2006;47:4072-4076.

64. Linsenmeier RA, Braun RD. Oxygen distribution and consumption in the cat retina during normoxia and hypoxemia. J Gen Physiol. 1992;99:177-197.

65. Schumacker PT, Samsel RW. Analysis of oxygen delivery and uptake relationships in the Krogh tissue model. Am J Physiol. 1989;67:1234-1244.

66. Severns ML, Adams JM. The relation between krogh and compartmental transport models. Journal of Theoretical Biology. 1982;97:239-249.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 67. Cringle SJ, Yu D-Y, Alder VA. Intraretinal oxygen tension in the rat eye.

Graefes Arch Clin Exp Ophthalmol. 1991;229:574-577.

68. Cringle SJ, Yu D-Y, Alder VA. Intravitreal and intraretinal oxygen tension in the rat eye. Adv Exp Med Biol. 1992;316:113-117.

69. Cringle SJ, Yu D-Y, Yu PK, Su E-N. Intraretinal oxygen consumption in the rat in vivo [published erratum appears in Invest Ophthalmol Vis Sci 2003 Jan;44(1)]. Invest Ophthalmol Vis Sci. 2002;43:1922-1927.

70. Cringle SJ, Yu D-Y. Intraretinal oxygenation and oxygen consumption in the rabbit during systemic hyperoxia. Invest Ophthalmol Vis Sci. 2004;45:3223-3228.

71. Yu D-Y, Cringle SJ, Su E-N. Intraretinal oxygen distrubution in the monkey retina and the response to systemic hyperoxia. Invest Ophthalmol Vis Sci. 2005;46:4728-4733.

72. Barka T, Anderson PJ. Oxidases and peroxidases- Cytochrome oxidase. In: Histochemistry theory, pratice and bibliography. New York: Hoeber Medical Division, Harper and Row; 1963:318-323.

73. Brown GC, Crompton M, Wray S. Cytochrome oxidase content of rat brain during development. Biochim Biophys Acta. 1991;1057:273-275.

74. Buttery RG, Weller WL, Haight JR. Retinal cytochrome oxidase, A missing link in the electron transport chain? Neuroscience Letters. 1997;30:s52.

75. Chen E, Soderberg PG, Lindstrom B. Activity distribution of cytochrome oxidase in the rat retina. A quantitative histochemical study. Acta Ophthalmol. 1989;67:645-651.

76. Wong-Riley MT. Cytochrome oxidase: an endogenous metabolic marker for neuronal activity. Trends Neurosci. 1989;12:94-101.

77. Yu D-Y, Cringle SJ, Su E-N, Yu PK, Humayun MS, Dorin G. Laser induced changes in intraretinal oxygen distribution in pigmented rabbits. Invest Ophthalmol Vis Sci. 2005;46:988-999.

78. Liu D, Wood NB, Witt N, Hughes AD, Thom SA, Xu XY. Computational analysis of oxygen transport in the retinal arterial network. Curr Eye Res. 2009;34:945-956.

79. Wolf S, Arend O, Schulte K, Ittel TH, Reim M. Quantification of retinal capillary density and flow velocity in patients with essential hypertension. Hypertension. 1994;23:464-467.

80. London A, Benhar I, Schwartz M. The retina as a window to the brain-from eye research to CNS disorders. Nat Rev Neurol. 2013;9:44-53.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 81. Cooper ME, Bonnet F, Oldfield M, Jandeleit-Dahm K. Mechanisms of diabetic

vasculopathy: an overview. Am J Hypertens. 2001;14:475-486.

82. Wong TY, Hubbard LD, Klein R, et al. Retinal microvascular abnormalities and blood pressure in older people: the Cardiovascular Health Study. Br J Ophthalmol. 2002;86:1007-1013.

83. Durham JT, Herman IM. Microvascular modifications in diabetic retinopathy. Curr Diab Rep. 2011;11:253-264.

84. Kawasaki R, Xie J, Cheung N, et al. Retinal microvascular signs and risk of stroke: the Multi-Ethnic Study of Atherosclerosis (MESA). Stroke. 2012;43:3245-3251.

85. Cheung CY, Ikram MK, Sabanayagam C, Wong TY. Retinal microvasculature as a model to study the manifestations of hypertension. Hypertension. 2012;60:1094-1103.

86. Recchia FM, ", Brown GC. Systemic disorders associated with retinal vascular occlusion. Curr Opin Ophthalmol. 2000;11:462-467.

87. . Natural history and clinical management of central retinal vein occlusion. The Central Vein Occlusion Study Group. Arch Ophthalmol. 1997;115:486-491.

88. Hayreh SS. Prevalent misconceptions about acute retinal vascular occlusive disorders. Prog Retin Eye Res. 2005;24:493-519.

89. Chapman N, ", Haimes G, et al. Acute effects of oxygen and carbon dioxide on retinal vascular network geometry in hypertensive and normotensive subjects. Clin Sci (Lond) ". 2000;99:483-488.

90. King LA, ", Stanton AV, et al. Arteriolar length-diameter (L:D) ratio: a geometric parameter of the retinal vasculature diagnostic of hypertension. J Hum Hypertens. 1996;10:417-418.

91. Hayreh SS, Zimmerman B, McCarthy MJ, Podhajsky P. Systemic diseases associated with various types of retinal vein occlusion. Am J Ophthalmol. 2001;131:61-77.

92. Rathaus M, Bernheim J. Oxygen species in the microvascular environment: regulation of vascular tone and the development of hypertension. Nephrol Dial Transplant. 2002;17:216-221.

93. Alder VA, Su EN, Yu D-Y, Cringle SJ, Yu PK. Diabetic retinopathy: early functional changes. Clinical & Experimental Pharmacology & Physiology. 1997;24:785-788.

94. Alder VA, Yu D-Y, Cringle SJ, Su EN. Experimental approaches to diabetic retinopathy. Asia Pac J Ophthalmol. 1992;4:20-25.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 95. Arend O, Wolf S, Remky A, et al. Perifoveal microcirculation with non-insulin-

dependent diabetes mellitus. Graefes Arch Clin Exp Ophthalmol. 1994;232:225-231.

96. Ashton N. Studies of the retinal capillaries in relation to diabetic and other retinopathies. Br J Ophthalmol. 1963;47:521-538.

97. Chatterjee S, Chattopadhyay S, Hope-Ross M, Lip PL. Hypertension and the eye: changing perspectives. J Hum Hypertens. 2002;16:667-675.

98. Wong TY, Klein R, Klein BE, Tielsch JM, Hubbard L, Nieto FJ. Retinal microvascular abnormalities and their relationship with hypertension, cardiovascular disease, and mortality. Surv Ophthalmol. 2001;46:59-80.

99. De Silva DA, Woon FP, Manzano JJ, et al. The relationship between aortic stiffness and changes in retinal microvessels among Asian ischemic stroke patients. J Hum Hypertens. 2012;26:716-722.

100. Cheung CY, Thomas GN, Tay W, et al. Retinal vascular fractal dimension and its relationship with cardiovascular and ocular risk factors. Am J Ophthalmol. 2012;154:663-674.

101. Yau JW, ", Lee P, et al. Retinal vein occlusion: an approach to diagnosis, systemic risk factors and" management. Intern Med J. 2008;38:904-910.

102. Deva R, ", Alias MA, et al. Vision-threatening retinal abnormalities in chronic kidney disease stages 3 to 5. Clin J Am Soc Nephrol. 2011;6:1866-1871.

103. Dubis AM, Hansen BR, Cooper RF, Beringer J, Dubra A, Carroll J. Relationship between the foveal avascular zone and foveal pit morphology. Invest Ophthalmol Vis Sci. 2012;53:1628-1636.

104. Hendrickson A. Organization of the Adult Primate Fovea.

105. Fitzgibbon T, Taylor SF. Retinotopy of the human retinal nerve fibre layer and optic nerve head. J Comp Neurol. 1996;375:238-251.

106. Mikelberg FS, Drance SM, Schulzer M, Yidegiligne HM, Weis MM. The normal human optic nerve. Axon count and axon diameter distribution. Ophthalmology. 1989;96:1325-1328.

107. Provis JM, Hendrickson AE. The foveal avascular region of developing human retina. Arch Ophthalmol. 2008;126:507-511.

108. Hyman C. The concept of end arteries and diversion of blood flow. Invest Ophthalmol. 1965;4:1000-1003.

109. Li W, ", Yanoff M, et al. Retinal capillary pericyte apoptosis in early human diabetic retinopathy. Chin Med J (Engl ). 1997;110:659-663.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 110. Arend O, ", Remky A, et al. Capillary density and retinal diameter

measurements and their impact on altered retinal circulation in glaucoma: a digital fluorescein angiographic study. Br J Ophthalmol. 2002;86:429-433.

111. Abramoff MD, ", Garvin MK, ", Sonka M. Retinal Imaging and Image Analysis. IEEE Trans Med Imaging. 2010;3:169-208.

112. Patton N, ", Aslam TM, et al. Retinal image analysis: concepts, applications and potential.". Prog Retin Eye Res. 2006;25:99-127.

113. Kozak I, ", Williams L, ". Seeing is believing: Updates in retinal imaging. Saudi J Ophthalmol. 2014;28:79-80.

114. Standish B, Mariampillai A, Leung M, Vitkin IA. Optical Coherence Tomography: Principles and Applications of Microvascular Imaging.

115. Patton N, ", Aslam T, et al. Retinal vascular image analysis as a potential screening tool for cerebrovascular disease: a rationale based on homology between cerebral and retinal microvasculatures. J Anat. 2005;206:319-348.

116. Bursell SE, Clermont AC, Shiba T, King GL. Evaluating retinal circulation using video fluorescein angiography in control and diabetic rats. Curr Eye Res. 1992;11:287-295.

117. Cavallerano AA. Ophthalmic fluorescein angiography. Optometry Clinics. 1996:1-23.

118. NOVOTNY HR, ALVIS DL. A method of photographing fluorescence in circulating blood in the human retina. Circulation. 1961;24:82-86.

119. Mendis KR, Balaratnasingam C, Yu P, et al. Correlation of histologic and clinical images to determine the diagnostic value of fluorescein angiography for studying retinal capillary detail. Invest Ophthalmol Vis Sci. 2010;51:5864-5869.

120. Bearelly S, Rao S, Fekrat S. Anaphylaxis following intravenous fluorescein angiography in a vitreoretinal clinic: report of 4 cases. Can J Ophthalmol. 2009;44:444-445.

121. Fine A, Amos WB, Durbin RM, McNaughton PA. Confocal microscopy: applications in neurobiology. Trends in Neurosciences. 1988;11:346-350.

122. Cavanagh HD, Petroll WM, Jester JV. Confocal microscopy: Uses in measurement of cellular structure and function. Prog Retina Eye Res. 1995;14:527-565.

123. Wang TD. Confocal microscopy from the bench to the bedside. Gastrointest Endosc. 2005;62:696-697.

124. Geng Y, ", Greenberg KP, et al. In vivo imaging of microscopic structures in the rat retina. Invest Ophthalmol Vis Sci. 2009;50:5872-5879.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 125. Masters BR, Bohnke M. Three-dimensional confocal microscopy of the living

human eye. Annu Rev Biomed Eng. 2002;4:69-91.

126. Elsner AE, Burns SA, Hughes GW, Webb RH. Reflectometry with a scanning laser ophthalmoscope. Applied Optics. 1992;31:3697-3711.

127. Wolf S, Arend O, Toonen H, Bertram B, Jung F, Reim M. Retinal capillary blood flow measurement with a scanning laser ophthalmoscope. Ophthalmology. 1991;98:996-1000.

128. Zhong Z, Song H, Chui TY, Petrig BL, Burns SA. Noninvasive measurements and analysis of blood velocity profiles in human retinal vessels. Invest Ophthalmol Vis Sci. 2011;52:4151-4157.

129. Godara P, ", Dubis AM, et al. Adaptive optics retinal imaging: emerging clinical applications. Optom Vis Sci. 2010;87:930-941.

130. Alexandrescu C, Dascalu AM, Panca A, et al. Confocal scanning laser ophthalmoscopy in glaucoma diagnosis and management. J Med Life. 2010;3:229-234.

131. Zangwill LM, Chan K, Bowd C, et al. Heidelberg retina tomograph measurements of the optic disc and parapapillary retina for detecting glaucoma analyzed by machine learning classifiers. Invest Ophthalmol Vis Sci. 2004;45:3144-3151.

132. Roorda A. Applications of adaptive optics scanning laser ophthalmoscopy. Optom Vis Sci. 2010;87:260-268.

133. Liang J, Williams DR, Miller DT. Supernormal vision and high-resolution retinal imaging through adaptive optics. J Opt Soc Am A Opt Image Sci Vis. 1997;14:2884-2892.

134. Burns SA, Elsner AE, Chui TY, et al. In vivo adaptive optics microvascular imaging in diabetic patients without clinically severe diabetic retinopathy. Biomed Opt Express. 2014;5:961-974.

135. Massamba N, Querques G, Lamory B, Querques L, Souied E, Soubrane G. In vivo evaluation of photoreceptor mosaic in type 2 idiopathic macular telangiectasia using adaptive optics. Acta Ophthalmol. 2010.

136. van Velthoven ME, Faber DJ, Verbraak FD, van Leeuwen TG, de Smet MD. Recent developments in optical coherence tomography for imaging the retina. Prog Retin Eye Res. 2007;26:57-77.

137. Kim DY, ", Fingler J, et al. Noninvasive imaging of the foveal avascular zone with high-speed, phase-variance " optical coherence tomography. Invest Ophthalmol Vis Sci. 2012;53:85-92.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 138. Mariampillai A, Standish BA, Moriyama EH, et al. Speckle variance detection

of microvasculature using swept-source optical coherence tomography. Opt Lett. 2008;33:1530-1532.

139. Wang RK, An L, Francis P, Wilson DJ. Depth-resolved imaging of capillary networks in retina and choroid using ultrahigh sensitive optical microangiography. Opt Lett. 2010;35:1467-1469.

140. Fingler J, Zawadzki RJ, Werner JS, Schwartz D, Fraser SE. Volumetric microvascular imaging of human retina using optical coherence tomography with a novel motion contrast technique. Opt Express. 2009;17:22190-22200.

141. Jia Y, Tan O, Tokayer J, et al. Split-spectrum amplitude-decorrelation angiography with optical coherence tomography. Opt Express. 2012;20:4710-4725.

142. Mahmud MS, Cadotte DW, Vuong B, et al. Review of speckle and phase variance optical coherence tomography to visualize microvascular networks. J Biomed Opt. 2013;18:50901.

143. Xu J, Wong K, Jian Y, Sarunic MV. Real-time acquisition and display of flow contrast using speckle variance optical coherence tomography in a graphics processing unit. J Biomed Opt. 2014;19:026001.

144. Yu PK, Tan PE, Morgan WH, Cringle SJ, McAllister IL, Yu DY. Age-related changes in venous endothelial phenotype at human retinal artery-vein crossing points. Invest Ophthalmol Vis Sci. 2012;53:1108-1116.

145. Kang MH, Balaratnasingam C, Yu P, et al. Morphometric characteristics of central retinal artery and vein endothelium in the normal human optic nerve head. Invest Ophthalmol Vis Sci. 2011;52:1359-1367.

146. Yu PK, Balaratnasingam C, Morgan WH, Cringle SJ, McAllister IL, Yu DY. The structural relationship between the microvasculature, neurons, and glia in the human retina. Invest Ophthalmol Vis Sci. 2010;51:447-458.

147. Radius RL. Thickness of the retinal nerve fiber layer in primate eyes. Arch Ophthalmol. 1980;98:1625-1629.

148. Surgucheva I, Weisman AD, Goldberg JL, Shnyra A, Surguchov A. Gamma-synuclein as a marker of retinal ganglion cells. Mol Vis. 2008;14:1540-1548.

149. Haverkamp S, Haeseleer F, Hendrickson A. A comparison of immunocytochemical markers to identify bipolar cell types in human and monkey retina. Visual Neuroscience. 2003;20:589-600.

150. Kivela T. Parvalbumin, a horizontal cell-associated calcium-binding protein in retinoblastoma eyes. Invest Ophthalmol Vis Sci. 1998;39:1044-1048.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 151. Gariano RF, Iruela-Arispe ML, Sage EH, Hendrickson AE.

Immunohistochemical characterization of developing and mature primate retinal blood vessels. Invest Ophthalmol Vis Sci. 1996;37:93-103.

152. Shaver SW, Pang JJ, Wall KM, Sposito NM, Gross PM. Subregional topography of capillaries in the dorsal vagal complex of rats: I. Morphometric properties. J Comp Neurol. 1991;306:73-82.

153. Less JR, Skalak TC, Sevick EM, Jain RK. Microvascular architecture in a mammary carcinoma: branching patterns and vessel dimensions. Cancer Res. 1991;51:265-273.

154. R Development Core Team. R: A Language and Environment for Statistical Computing. In: Vienna, Austria: R Foundation for Statistical Computing; 2011.

155. Alamouti B, Funk J. Retinal thickness decreases with age: an OCT study. Br J Ophthalmol. 2003;87:899-901.

156. Fujino H, Kondo H, Murakami S, et al. Differences in capillary architecture, hemodynamics, and angiogenic factors in rat slow and fast plantarflexor muscles. Muscle Nerve. 2012;45:242-249.

157. Safaeian N, Sellier M, David T. A computational model of hemodynamic parameters in cortical capillary networks. J Theor Biol. 2010;271:145-156.

158. Cobb S. Cerebral Circulation. A critical discussion of the symposium. In: Cobb S, Frantz AM, Penfield W, Riley HA, Association for Research in Nervous and Mental Disease (eds), The Circulation of the Brain and Spinal Cord Baltimore: The Williams and Wilkins; 1938:710-752.

159. Craigie EH. The vascularity of the cerebral cortex of the albino rat. J Comp Neurol. 1921;33:193-212.

160. Motti ED, Imhof HG, Yasargil MG. The terminal vascular bed in the superficial cortex of the rat. An SEM study of corrosion casts. J Neurosurg. 1986;65:834-846.

161. Yu D-Y, Cringle SJ, Alder VA, Su EN. Intraretinal oxygen distribution in rats as a function of systemic blood pressure. Am J Physiol. 1994;36:H2498-H2507.

162. Yu D-Y, Cringle SJ. Oxygen distribution in the mouse retina. Invest Ophthalmol Vis Sci. 2006;47:1109-1112.

163. Yu D-Y, Cringle SJ. Low oxygen consumption in the inner retina of the visual streak of the rabbit. Am J Physiol (Heart C). 2004;286:H419-H423.

164. Yu D-Y, Cringle SJ, Su E-N, Yu PK. Retinal degeneration and local oxygen metabolism. Exp Eye Res. 2005;80:745-751.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 165. Yu D-Y, Cringle SJ, Su EN, Yu PK. Retinal oxygen distribution in a rat model

of retinal ischemia. Invest Ophthalmol Vis Sci. 2000;41:S19.

166. Gariano RF, Gardner TW. Retinal angiogenesis in development and disease. Nature. 2005;438:960-966.

167. Metea MR, Newman EA. Glial cells dilate and constrict blood vessels: a mechanism of neurovascular coupling. J Neurosci. 2006;26:2862-2870.

168. Rackl A, Pawlik G, Bing RJ. Cerebral capillary topography and red cell flow in vivo. In: Ceros-Navaroo J, Fritschka E (eds), Cerebral Microcirculation and Metabolism New York: Raven Press; 1981:17-21.

169. Bringmann A, Pannicke T, Grosche J, et al. Muller cells in the healthy and diseased retina. Prog Retin Eye Res. 2006;25:397-424.

170. Bringmann A, Uckermann O, Pannicke T, Iandiev I, Reichenbach A, Wiedemann P. Neuronal versus glial cell swelling in the ischaemic retina. Acta Ophthalmol Scand. 2005;83:528-538.

171. Germer A, Biedermann B, Wolburg H, et al. Distribution of mitochondria within Muller cells--I. Correlation with retinal vascularization in different mammalian species. J Neurocytol. 1998;27:329-345.

172. Ellsworth ML, Popel AS, Pittman RN. Assessment and impact of heterogeneities of convective oxygen transport parameters in capillaries of striated muscle: experimental and theoretical. Microvasc Res. 1988;35:341-362.

173. Pawlik G, Rackl A, Bing RJ. Quantitative capillary topography and blood flow in the cerebral cortex of cats: an in vivo microscopic study. Brain Res. 1981;208:35-58.

174. Danis RP, Wallow IHL. HRP/Trypsin technique for studies of the retinal vasculature. Invest Ophthalmol Vis Sci. 1986;27:434-437.

175. Bek T, Jensen PK. Three-dimensional structure of human retinal vessels studied by vascular casting. Acta Ophthalmol (Copenh). 1993;71:506-513.

176. Tan PE, Yu PK, Balaratnasingam C, et al. Quantitative confocal imaging of the retinal microvasculature in the human retina. Invest Ophthalmol Vis Sci. 2012;53:5728-5736.

177. Chan G, Balaratnasingam C, Yu PK, et al. Quantitative morphometry of perifoveal capillary networks in the human retina. Invest Ophthalmol Vis Sci. 2012;53:5502-5514.

178. Attwell D, Buchan AM, Charpak S, Lauritzen M, Macvicar BA, Newman EA. Glial and neuronal control of brain blood flow. Nature. 2010;468:232-243.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 179. Drake CT, Iadecola C. The role of neuronal signaling in controlling cerebral

blood flow. Brain and Language. 2007;102:141-152.

180. Pournaras CJ, Rungger-Brandle E, Riva CE, Hardarson SH, Stefansson E. Regulation of retinal blood flow in health and disease. Prog Retin Eye Res. 2008;27:284-330.

181. Jespersen SN, Ostergaard L. The roles of cerebral blood flow, capillary transit time heterogeneity, and oxygen tension in brain oxygenation and metabolism. J Cereb Blood Flow Metab. 2012;32:264-277.

182. Bedggood P, Metha A. Direct visualization and characterization of erythrocyte flow in human retinal capillaries. Biomed Opt Express. 2012;3:3264-3277.

183. Remky A, Wolf S, Knabben H, Arend O, Reim M. Perifoveal capillary network in patients with acute central retinal vein occlusion. Ophthalmology. 1997;104:33-37.

184. Crosby-Nwaobi R, Heng LZ, Sivaprasad S. Retinal vascular calibre, geometry and progression of diabetic retinopathy in type 2 diabetes mellitus. Ophthalmologica. 2012;228:84-92.

185. Leung H, Wang JJ, Rochtchina E, Wong TY, Klein R, Mitchell P. Impact of current and past blood pressure on retinal arteriolar diameter in an older population. J Hypertens. 2004;22:1543-1549.

186. Schmidt D. The mystery of cotton-wool spots - a review of recent and historical descriptions. Eur J Med Res. 2008;13:231-266.

187. Tang J, Mohr S, Du YD, Kern TS. Non-uniform distribution of lesions and biochemical abnormalities within the retina of diabetic humans. Curr Eye Res. 2003;27:7-13.

188. Klein R, Meuer SM, Moss SE, Klein BEK. The relationship of retinal microaneurysm counts to the 4-year progression of diabetic retinopathy. Arch Ophthalmol. 1989;107:1780-1785.

189. Tarr JM, Kaul K, Wolanska K, Kohner EM, Chibber R. Retinopathy in diabetes. Adv Exp Med Biol. 2012;771:88-106.

190. Chapman N, Dell'omo G, Sartini MS, et al. Peripheral vascular disease is associated with abnormal arteriolar diameter relationships at bifurcations in the human retina. Clin Sci (Lond). 2002;103:111-116.

191. Grizzi F, Russo C, Colombo P, et al. Quantitative evaluation and modeling of two-dimensional neovascular network complexity: the surface fractal dimension. BMC Cancer. 2005;5:14.

192. Beard DA, Bassingthwaighte JB. Modeling advection and diffusion of oxygen in complex vascular networks. Ann Biomed Eng. 2001;29:298-310.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 193. Levy MN, SAUCEDA G. Diffusion of oxygen from arterial to venous segments

of renal capillaires. Am J Physiol. 1959;196:1336-1339.

194. Pannabecker TL, Dantzler WH. Three-dimensional architecture of collecting ducts, loops of Henle, and blood vessels in the renal papilla. Am J Physiol Renal Physiol. 2007;293:F696-F704.

195. Hudetz AG, Biswal BB, Feher G, Kampine JP. Effects of hypoxia and hypercapnia on capillary flow velocity in the rat cerebral cortex. Microvasc Res. 1997;54:35-42.

196. Atkinson JL, Anderson RE, Sundt TM, Jr. The effect of carbon dioxide on the diameter of brain capillaries. Brain Res. 1990;517:333-340.

197. Stefanovic B, Hutchinson E, Yakovleva V, et al. Functional reactivity of cerebral capillaries. J Cereb Blood Flow Metab. 2008;28:961-972.

198. Duelli R, Kuschinsky W. Changes in brain capillary diameter during hypocapnia and hypercapnia. J Cereb Blood Flow Metab. 1993;13:1025-1028.

199. van der Linden P, De HS, Belisle S, et al. Comparative effects of red blood cell transfusion and increasing blood flow on tissue oxygenation in oxygen supply-dependent conditions. Am J Respir Crit Care Med. 2001;163:1605-1608.

200. Mowat FM, Luhmann UF, Smith AJ, et al. HIF-1alpha and HIF-2alpha are differentially activated in distinct cell populations in retinal ischaemia. PLoS ONE. 2010;5:e11103.

201. le NF, Fleury V, Pries A, Corvol P, Eichmann A, Reneman RS. Control of arterial branching morphogenesis in embryogenesis: go with the flow. Cardiovasc Res. 2005;65:619-628.

202. Gerhardt H, Golding M, Fruttiger M, et al. VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia. J Cell Biol. 2003;161:1163-1177.

203. Kowalczuk L, Touchard E, Omri S, et al. Placental growth factor contributes to micro-vascular abnormalization and blood-retinal barrier breakdown in diabetic retinopathy. PLoS ONE. 2011;6:e17462.

204. Rousseau B, Larrieu-Lahargue F, Bikfalvi A, Javerzat S. Involvement of fibroblast growth factors in choroidal angiogenesis and retinal vascularization. Exp Eye Res. 2003;77:147-156.

205. Wang J, Xu X, Elliott MH, Zhu M, Le YZ. Muller cell-derived VEGF is essential for diabetes-induced retinal inflammation and vascular leakage. Diabetes. 2010;59:2297-2305.

206. Mariampillai A, Leung MK, Jarvi M, et al. Optimized speckle variance OCT imaging of microvasculature. Opt Lett. 2010;35:1257-1259.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 207. Young M, Lebed E, Jian Y, Mackenzie PJ, Beg MF, Sarunic MV. Real-time

high-speed volumetric imaging using compressive sampling optical coherence tomography. Biomed Opt Express. 2011;2:2690-2697.

208. Jian Y, Wong K, Sarunic MV. Graphics processing unit accelerated optical coherence tomography processing at megahertz axial scan rate and high resolution video rate volumetric rendering. J Biomed Opt. 2013;18:26002.

209. Lee KK, Mariampillai A, Yu JX, et al. Real-time speckle variance swept-source optical coherence tomography using a graphics processing unit. Biomed Opt Express. 2012;3:1557-1564.

210. Chan G, Balaratnasingam C, Yu PK, et al. Quantitative changes in perifoveal capillary networks in patients with vascular comorbidities. Invest Ophthalmol Vis Sci. 2013;54:5175-5185.

211. Park MY, Jung SE, Byun JY, Kim JH, Joo GE. Effect of beam-flow angle on velocity measurements in modern Doppler ultrasound systems. AJR Am J Roentgenol. 2012;198:1139-1143.

212. Larina IV, Larin KV, Justice MJ, Dickinson ME. Optical Coherence Tomography for live imaging of mammalian development. Curr Opin Genet Dev. 2011;21:579-584.

213. Ellabban AA, Tsujikawa A, Matsumoto A, et al. Macular choroidal thickness and volume in eyes with angioid streaks measured by swept source optical coherence tomography. Am J Ophthalmol. 2012;153:1133-1143.

214. Rastogi V, Puri N, Arora S, Kaur G, Yadav L, Sharma R. Artefacts: a diagnostic dilemma - a review. J Clin Diagn Res. 2013;7:2408-2413.

215. Tawia SA, Rogers PA. In vivo microscopy of the subepithelial capillary plexus of the endometrium of rats during embryo implantation. J Reprod Fertil. 1992;96:673-680.

216. Hamilton NB, Attwell D, Hall CN. Pericyte-mediated regulation of capillary diameter: a component of neurovascular coupling in health and disease. Front Neuroenergetics. 2010;2.

217. Mintun MA, Lundstrom BN, Snyder AZ, Vlassenko AG, Shulman GL, Raichle ME. Blood flow and oxygen delivery to human brain during functional activity: theoretical modeling and experimental data. Proc Natl Acad Sci U S A. 2001;98:6859-6864.

218. Yoshii Y, Sugiyama K. Intercapillary distance in the proliferating area of human glioma. Cancer Res. 1988;48:2938-2941.

219. Motaghiannezam R, Fraser S. Logarithmic intensity and speckle-based motion contrast methods for human retinal vasculature visualization using swept source optical coherence tomography. Biomed Opt Express. 2012;3:503-521.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan 220. Braaf B, Vienola KV, Sheehy CK, et al. Real-time eye motion correction in

phase-resolved OCT angiography with tracking SLO. Biomed Opt Express. 2013;4:51-65.

221. Vienola KV, Braaf B, Sheehy CK, et al. Real-time eye motion compensation for OCT imaging with tracking SLO. Biomed Opt Express. 2012;3:2950-2963.

222. Gariano RF, Iruela-Arispe ML, Hendrickson AE. Vascular development in primate retina: comparison of laminar plexus formation in monkey and human. Invest Ophthalmol Vis Sci. 1994;35:3442-3455.

223. Thangarajah H, Yao D, Chang EI, et al. The molecular basis for impaired hypoxia-induced VEGF expression in diabetic tissues. Proc Natl Acad Sci U S A. 2009;106:13505-13510.

224. Simo R, Carrasco E, Garcia-Ramirez M, Hernandez C. Angiogenic and antiangiogenic factors in proliferative diabetic retinopathy. Curr Diabetes Rev. 2006;2:71-98.

225. Kutschbach P, Wolf S, Sieveking M, Ittel TH, Schulte K, Reim M. Retinal capillary density in patients with arterial hypertension: 2-year follow-up. Graefes Arch Clin Exp Ophthalmol. 1998;236:410-414.

226. Chauhan DS, Marshall J. The interpretation of optical coherence tomography images of the retina. Invest Ophthalmol Vis Sci. 1999;40:2332-2342.

227. Gabriele ML, Wollstein G, Ishikawa H, et al. Three dimensional optical coherence tomography imaging: advantages and advances. Prog Retin Eye Res. 2010;29:556-579.

228. Mrejen S, Spaide RF. Optical coherence tomography: imaging of the choroid and beyond. Surv Ophthalmol. 2013;58:387-429.

229. van Velthoven ME, ", Faber DJ, et al. Recent developments in optical coherence tomography for imaging the retina. Prog Retin Eye Res. 2007;26:57-77.

230. Bergers G, Song S. The role of pericytes in blood-vessel formation and maintenance. Neuro Oncol. 2005;7:452-464.

231. Schmoll T, Singh AS, Blatter C, et al. Imaging of the parafoveal capillary network and its integrity analysis using fractal dimension. Biomed Opt Express. 2011;2:1159-1168.

232. Hayreh SS. Physiological Anatomy of the Retinal Vasculature. In: Dartt DA (ed), Encyclopedia of the Eye Oxford: Academic Press; 2010:431-438.

Thesis- Quantitative Analysis of Perifoveal Capillary Networks in the Human Retina By Dr Geoffrey Zhi Peng Chan Appendix Nil supplementary material attached.

quantitative analysis of capillary networks in the human ... · quantitative analysis of capillary...

Documents