University of Southern California Dissertations and Theses

Oxygen therapy for the treatment of retinal ischemia

(USC Thesis Other)

Oxygen therapy for the treatment of retinal ischemia

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content OXYGEN THERAPY FOR THE TREATMENT OF RETINAL ISCHEMIA
by
Karthik Murali
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulllment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOMEDICAL ENGINEERING)
August 2016
Copyright 2016 Karthik Murali
Epigraph
Nobody ever gures out what life is all about, and it doesn't matter. Explore the world.
Nearly everything is really interesting if you go into it deeply enough.
- Richard Feynman
ii
Acknowledgments
Completing my PhD would not have been possible without the generous help and support
of others. First and foremost, I would like to thank my adviser Dr. Mark Humayun for
shaping the way I think and for his guidance over the years. I could not have asked for
a better mentor. I thank Dr Amir Kashani for encouraging my curiosity in the unknown
and for training me to be a better scientist. I would also like to thank Dr Yu-Chong Tai
for showing me how to merge theory with practice and for training me to be a better
engineer. Finally, my thanks goes out to Dr James Weiland for his unwavering support
and for always having his door open to his students.
Research is a collaborative eort and I was fortunate to have many wonderful col-
leagues. I am deeply grateful for all their eorts. They include Nicholas Scianmarello,
Ramiro Ribeiro, Dongyang Kang, Juan Carlos Martinez, Hossein Nazari, and Mort
Arditti.
I would also like to acknowledge the following funding sources for supporting me and
my research endeavours: National Institutes of Health, Harrington Foundation, Whittier
Foundation, Perkins Foundation, and the Research to Prevent Blindness.
Finally, I thank my family, especially my parents, for being a constant source of
support and encouragement. All that I am or ever hope to be, I owe to them.
iii
Table of Contents
Epigraph ii
Acknowledgments iii
List of Tables vi
List of Figures vii
Abstract x
Chapter 1: Introduction 1
1.1 Blindness and Retinal Ischemic Diseases . . . . . . . . . . . . . . . . . . . 1
1.2 Oxygen Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Research Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Thesis Structure and Contributions . . . . . . . . . . . . . . . . . . . . . . 8
Chapter 2: Background 10
2.1 Retina Physiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Vitreous Physiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Ischemic Retina Physiology . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Oxygenator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.5 Disease Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.6 Energy Harvesters for Implants . . . . . . . . . . . . . . . . . . . . . . . . 29
2.7 Implantable Oxygen Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Chapter 3: Spatial Variation in Vitreous Oxygen Consumption 39
3.1 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Chapter 4: Intravitreal Oxygen Level Changes after Laser Photothrombo-
sis in Rabbits 59
4.1 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
iv
4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Chapter 5: Oxygen Biosensor 72
5.1 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Chapter 6: Harvesting Solar Energy to Power Ocular Implants 79
6.1 Goal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Chapter 7: Conclusion 91
References 96
v
List of Tables
2.1 Summary of various animal models of retinal ischemia . . . . . . . . . . . 27
3.1 Ascorbate concentration dierences between mid-vitreous and posterior
vitreous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.2 Coecients and goodness-of-t indicator for curve tted to both mid-
vitreous and posterior vitreous data . . . . . . . . . . . . . . . . . . . . . 52
4.1 Fluorescein Angiography results . . . . . . . . . . . . . . . . . . . . . . . . 69
6.1 Summary of Solar Energy Harvester Characteristics . . . . . . . . . . . . 86
vi
List of Figures
1.1 Vision loss after Diabetic Retinopathy. Image courtesy of National Eye
Institute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Proposed Oxygenator design and use . . . . . . . . . . . . . . . . . . . . . 4
2.1 Human retina anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Ocular circulation anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Intraretinal oxygen prole . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4 Ultra-structure of hyaluronic acid/collagen interaction in the vitreous . . 14
2.5 Ultra-structure of vitreous . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.6 Vitreous anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.7 Rheology of human vitreous . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.8 Intravitreal oxygen map . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.9 Summary of hypothesized key events in the process of ischemic retinal
neurodegeneration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.10 Oxygenator illustration . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.11 Intraretinal oxygen measurement after laser occlusion of retinal circulation 29
2.12 Spectral irradiance of the sun compared to various lamps . . . . . . . . . 31
2.13 Photodiode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.14 Quantum eciency versus wavelength for dierent photocell materials . . 33
vii
2.15 Boost converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.16 Lithium ion battery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.17 Charge stages of a lithium ion battery . . . . . . . . . . . . . . . . . . . . 35
2.18 Jablonski diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.1 Oxygen source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.2 Ex vivo porcine eye setup . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3 Biopsied vitreous regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.4 Oxygen diusion comparison between mid and posterior vitreous . . . . . 46
3.5 Kinetics of vitreous oxygen consumption . . . . . . . . . . . . . . . . . . . 47
3.6 Representative results: Vitreous oxygen decay . . . . . . . . . . . . . . . . 49
3.7 Spatial variation in vitreous oxygen consumption. Experimental data with
tted curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.1 Fluorescein Angiography data analysis . . . . . . . . . . . . . . . . . . . . 62
4.2 Intravitreal oxygen measurement positions . . . . . . . . . . . . . . . . . . 63
4.3 Representative image of probe location . . . . . . . . . . . . . . . . . . . . 64
4.4 Representative plot of preretinal oxygen tension after photothrombosis . . 64
4.5 Average intravitreal pO
2
(in mmHg) in various positions before and after
laser photothrombosis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.6 Bar plot of intravitreal oxygen tension (P1-P8) before and after occlusion 66
4.7 Average intravitreal pO
2
(in mmHg) in various positions at time Day O
+
,
Day 7 and Day 14. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.8 Bar plot of intravitreal pO
2
(P1-P8) at times Day 0
+
, Day 7, and Day 14. 67
4.9 Preretinal pO
2
over 14 days across all positions (P1-P8). . . . . . . . . . . 68
4.10 Perfusion against preretinal oxygen tension on Day 14 . . . . . . . . . . . 70
viii
5.1 Overall experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.2 Optical sensor setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.3 How the sensor works (Ocean Optics) . . . . . . . . . . . . . . . . . . . . 76
5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
6.1 Solar energy harvester schematic diagram . . . . . . . . . . . . . . . . . . 81
6.2 Photograph of an enucleated porcine eye with implanted photovoltaic array 83
6.3 Optical transmittance through tissue and materials . . . . . . . . . . . . . 84
6.4 Characteristic I-V and corresponding power curves for photovoltaic array 85
6.5 Proposed transparent packaging for electronics . . . . . . . . . . . . . . . 88
ix
Abstract
Retinal vascular diseases, such as Diabetic Retinopathy and Retinal Vein Occlusions, can
cause a profound visual handicap due to retinal ischemia. An intraocular medical de-
vice (Oxygenator) that can provide localized supplemental delivery of oxygen to ischemic
retina could reduce the progression of retinal cell loss. For such a medical device to be
feasible, we need to understand how oxygen travels in the eye, especially in the vitreous
humor. Next, we need to develop an animal model of disease for the Oxygenator's vali-
dation studies. We also need to use a non-invasive method to monitor intravitreal oxygen
levels over time. Finally, we need to alleviate the power constraints on such implantable
ocular devices.
First, to study vitreal oxygen mass transport, oxygen diusion from an intravitreal
oxygen source was studied in enucleated porcine eyes. The vitreous oxygen consump-
tion rate was modeled after a second order rate reaction equation and the rate constant
was experimentally quantied. A spatial variation in vitreal oxygen mass transport and
vitreous oxygen consumption was observed.
To develop an animal disease model for Oxygenator preclinical studies, a rabbit model
of retinal ischemia induced by laser photothrombosis was developed and characterized.
x
Intravitreal and preretinal oxygen levels was measured across 14 days in control and
ischemic animal groups. Fluorescein angiography was also performed across 14 days to
quantify perfusion in retinal vasculature. In this animal disease model, intravitreal oxygen
levels are depressed for 7 days. Between Day 7 and Day 14, the retinal vasculature is
re-perfused and intravitreal oxygen levels start to rise.
To monitor intravitreal oxygen levels over time, a prototype of a non-invasive optical
oxygen biosensor was developed. Luminescent oxygen sensitive dye molecules (platinum
metal porphyrin) embedded in an acrylic patches were placed in the vitreous of enculeated
bovine eyes. The patches were optically excited through the cornea and the emission
light was recorded to obtain oxygen concentration measurements. This prototype oxygen
biosensor can accurately measure intravitreal oxygen levels between 0 and 21%.
To reduce the power constraints on ocular medical devices, an implantable solar en-
ergy harvesting power supply was engineered. A photovoltaic array implanted under
the conjunctiva converted incident light energy to electrical energy stored in a battery.
The subconjunctival photovoltaic array, with an area of 62 mm
2
, when irradiated at
1mWmm
-2
of sunlight, can generate 1.22 mW at the battery charging circuit's output.
xi
Chapter 1
Introduction
1.1 Blindness and Retinal Ischemic Diseases
For those of us who are reading this, it is impossible to fathom what it means to be
blind. But there are more than 285 million people around the world who suer from
various degrees of blindness[World Health Organization 2014]. Not only is the emotional
and psychological toll of this condition devastating, the economic burden of this disease
is staggering. Some of the leading causes of blindness are diseases which involve retinal
ischemia (inadequate blood
ow to the eye's retina) such as Diabetic Retinopathy and
Retinal Vein Occlusions. In the US alone, Diabetic Retinopathy is projected to aect
11M people by the year 2030[National Eye Institute]. Unfortunately, current methods
of treating retinal ischemia are inadequate and do not hold much promise for people
suering from blindness.
Currently, the standard of care for ischemic retinal disease treatment involves using
intravitreal medications (injections into the eye), laser photocoagulation (burning away
parts of the retina), and Pars Plana Vitrectomy (removing the vitreous humor of the
1
(a) (b)
Figure 1.1: Vision loss after Diabetic Retinopathy. Image courtesy of National Eye
Institute
eye)[Giuliari 2012]. Intravitreal medications have been shown to decrease vascular en-
dothelial growth factor levels and vascular permeability (side eects of retinal ischemia)
but they need to be administered monthly and do not, in fact, treat the underlying
causes of retinal ischemia. Laser photocoagulation is destructive, leading to side eects
such as constricted peripheral visual eld, delayed dark adaptation, central blind spots,
and scarring. Vitrectomy itself reduces damage by hypoxia (inadequate oxygen), but
does so inadequately. It can also divert oxygen from the anterior structures, leading to
iris and angle neovascularization, and elevated intraocular pressure. In summary, the
current treatments, especially for moderate to severe ischemic retinal ischemic diseases,
have signicant room for improvement. In addition, if the current trend of treating
these conditions via intravitreal injection continues to increase, the treatments will be
very expensive and burdensome for the healthcare system[IVT Injections: Health Policy
Implications].
2
1.2 Oxygen Therapy
In an eort to target the underlying pathophysiology of retinal ishemia, our team proposed
a oxygen therapy as a method for treating retinal ischemia. Specically, we proposed
an ocular implant (Oxygenator), a bio-microelectromechanical (bio-MEMS) system that
provides localized, supplemental oxygen to the retina for treating hypoxia (Fig 1.2).
This method is backed by series of scientic ndings from the past 20 years. First,
studies in animals showed that increasing the oxygen levels in the vitreous cavity can
markedly decrease retinal damage and improve the electroretinogram (ERG) response,
which is an indicator of retinal cell health [Ben-Nun et al. 1988; Blair et al. 1991; Wilson,
Berkowitz, and Srebro 1995]. Nguyen et al later demonstrated that systemic oxygenation
in patients with diabetic macular edema can successfully treat macular edema[Nguyen
2004]. Finally, Abdallah et al demonstrated that in animal models of retinal ischemia,
localized acute oxygenation of the retina through the vitreous can spare the retina from
ischemic damage[Abdallah et al. 2011]. Based on these ndings, the Oxygenator was
developed to regulate oxygen delivery with nanoliter accuracy and direct it exclusively to
the retina.
The Oxygenator will be composed of 3 parts: the pump body, cannula and a Arcade.
The pump body, encapsulated in a hermetically sealed chamber, will contain a rellable
saline reservoir as well as the battery and programmable electronics. Electrolysis will
be performed on saline via electrodes to generate hydrogen and oxygen gas. This gas
will diuse throughout the rest of the device. The cannula is a
uidic channel that
will connect the pump body to the Arcade. The Arcade is a semipermeable membrane.
3
Oxygen diuses through the Arcade and out of the device. The small Oxygenator pump
body will be sutured to the sclera in one quadrant of the human eye: similar to the
placement of the FDA approved Argus Retinal Prosthesis. The Arcade will be implanted
inside the posterior chamber through a small incision. The Oxygenator is designed to be
an ideal treatment for moderate to severe diabetic retinopathy and retinal vein occlusions
in which there is macular ischemia and macular edema. We believe that this will be a
paradigm shift from current treatment options.
Figure 1.2: (Top Left) Oxygenator device: the device will have an electronic capsule,
an electrolyte reservoir that is connected to the intraocular unit through a cannula, and
an electrolyte rell port. (Top Right) Oxygenator Extraocular unit: The device will be
miniaturized with a low prole to t in the subconjunctival space and can be relled
using an automated pump.(Bottom Left) Location of the implant (Bottom Center) Er-
gonomic low prole design for recharging power transmission. (Bottom Right) Flexible,
and comfortable for night time recharging.
Our team has made great strides in developing this technology. We have created two
devices (passive and active) that are capable of localized supplemental oxygen delivery.
The passive device, uses diusion along the concentration gradient between the atmo-
spheric air and the vitreous humor to transport oxygen from the subconjunctival space to
4
the ischemic retina[Kang et al. 2015]. The active device, generates oxygen via electrolysis
of saline and delivers it to the ischemic retina[Scianmarello et al. 2016].
1.3 Research Goals
For this therapy to become reality and for the Oxygenator to be used clinically, we must
rst answer a few key questions. What is the safe and eective therapeutic window
for oxygenation of the ischemic retina? Too little oxygen and the therapy is ineective.
Too much oxygen puts the retinal tissue under oxidative stress[Hiramitsu et al. 1976].
However, since our therapy works by delivering oxygen through the vitreous, we must
fully understand the oxygen mass transport dynamics in the vitreous humor. Dening the
spatial and temporal dynamics of intravitreal oxygen transport will enable us to quantify
the exact concentration of oxygen that reaches the ischemic parts of the retina from our
device. However, current empirical and computational models of oxygen transport within
the vitreous only consider the retina as an oxygen source in the posterior chamber [Shui
et al. 2006; Filas, Shui, and Beebe 2013]. Can these models be extrapolated to accurately
dene oxygen transport from an external source of intravitreal oxygen? Furthermore,
while we have evidence that the vitreous consumes oxygen in an ascorbate dependent
manner, we do not fully understand the reaction kinetics of this mechanism[Shui et al.
2009].
My rst research goal was to empirically dene the mass transport dynamics of oxygen
in the eye from an intravitreal source[Murali et al. 2016; Murali, Kashani, and Humayun
5
2015]. I did this by studying the oxygen transport from the Oxygenator prototypes in
enucleated porcine eyes (Chapter 3).
Delineating the safe and eective therapeutic window for oxygenation of the ischemic
retina also requires us to create an animal model of retinal ischemia to test our oxygen
therapy on. For this purpose, we had chosen a rabbit model with retinal ischemia induced
by photothrombosis. This decision was based on reasons detailed in Chapter 2. To
appreciate the changes induced by oxygen therapy on this animal disease model, we must
rst understand the retinal changes and intravitreal oxygen level changes that occur over
time in the animal disease model without oxygen therapy. However, only the chronic
retinal vascular characteristics, along with retinal histology and immunohistochemistry,
have been well dened[Ameri et al. 2008b]. What are the intravitreal oxygen levels in
this animal disease model and how do they change over time?
This led to my second research goal; to characterize the intravitreal oxygen distri-
butions in my animal disease model. I did this by investigating the intravitreal oxygen
levels in a rabbit model of retinal ischemia induced by laser photothrombosis over the
course of 14 days (Chapter 4).
In order to accurately provide the therapeutic levels of oxygen, the Oxygenator needs
to be able to monitor the intravitreal oxygen levels and titrate the oxygen dose ac-
cordingly. However, the current method of measuring intravitreal oxygen is through an
invasive probe that is inserted into the eye. There is a need to incorporate a wireless
implantable oxygen sensor within the Oxygenator. This will allow the Oxygenator to
close the loop on oxygen dosing.
6
My third research goal was to develop a non-invasive oxygen biosensor to integrate
with the Oxygenator. I did this by engineering an optical oxygen biosensor prototype and
evaluating its feasibility to monitor the oxygen content within the implanted Oxygenator
(Chapter 5).
Finally, the Oxygenator will be powered by batteries encased in its pump body. Due
to the physical size constraints of the Oxygenator, battery size, and thus capacity, is
limited and the batteries need to be recharged frequently. One method of doing so is via
inductive coupling with time varying magnetic elds. While this method of wireless power
transfer is common among new medical device implants, the need for constant battery
recharging, from the patient's point of view, increases the complexity of the therapy.
There is currently no method of eciently harvesting energy from environmental sources
for ocular implants. Power constraints can be relaxed by energy scavengers, exploiting
environmental sources of energy, such as solar energy[Olivo, Carrara, and De Micheli
2011]. There is potential for new breakthroughs in the eld of energy scavenging for
medical device implants.
My nal research goal was to develop a energy scavenger to power ocular implants
[Murali, Scianmarello, and Humayun 2015]. I did this by engineering a photovoltaic
system to harvest light energy to power ophthalmic implants (Chapter 6).
7
1.4 Thesis Structure and Contributions
Chapter 2 contains an overview of the relevant scientic literature pertaining to my
dissertation. This covers healthy and disease physiology as well as introductions to oxygen
biosensors and energy scavenging for implants.
Chapter 3 details my work investigating the oxygen transport dynamics in enucle-
ated porcine eyes. Through my experiments studying the intravitreal oxygen diusion,
I quantied the reaction kinetics of vitreous oxygen consumption, modeled as a second
order rate reaction, and found a spatial variation in vitreous oxygen consumption.
In Chapter 4, I sought to establish an animal model for the Oxygenator studies by
studying the intravitreal oxygen levels in a rabbit model of retinal ischemia induced by
laser photothrombosis over the course of 14 days. These studies led to a full characteri-
zation of the intravitreal oxygen distribution in this animal model of retinal ischemia as
well as the establishment of a viable animal disease model for the Oxygenator pre-clinical
studies.
Chapter 5 describes my work on developing an oxygen biosensor to close the loop
on oxygen generation for the Oxygenator. I successfully demonstrated the feasibility of
accurately measuring intravitreal oxygen (0-21% oxygen tension) non-invasively through
an luminescence based optical sensor.
Chapter 6 describes my work on developing an energy harvesting system to power
ocular implants. Through my work, I built and validated a subconjunctival photovoltaic
implant that can generate up to 1.22mW to power ocular implants.
8
Finally, I summarize my scientic work and suggest ideas for future studies in Chapter
7.
9
Chapter 2
Background
2.1 Retina Physiology
The human retina is the neurosensory tissue that is responsible for converting light energy
into neuronal responses that is essential for vision. This retinal tissue is approximately
300μm thick in humans, with variations according to race, gender, and age[]Kashani et al.
2010]. The vertebrate retina is made up for the following layers of cells: retinal pigment
epithelium, photoreceptor layer, outer nuclear layer, outer plexiform layer, inner nuclear
layer, inner plexiform layer, ganglion cell layer, and nerve ber layer[Jampol 2003]. These
layers can also be classied as inner retina (inner nuclear layer through to nerve ber layer)
and outer retina (Retinal Pigment Epithelium through to outer plexiform layer).
The retinal tissue, which is one of the most metabolically active tissues in the body, is
perfused by the retinal vasculature, which is located within the retina, and the choroidal
vasculature, which is located posterior to the retinal pigment epithelium (Figure 2.2)
[Jampol 2003]. The inner retina is perfused by the retinal vasculature, which is auto
regulated and thus is responsive to systemic changes in oxygen levels. The outer retina
10
Figure 2.1: 3-D block of a portion of human retina [Webvision 2016]
is supplied solely by the choroid, which is not auto regulated[Jampol 2003]. Both these
vasculature systems serve to provide tissue with oxygen, metabolic substrates, and remove
waste products. The retinal vasculature varies from species to species. In humans, cats,
dogs, and rats, the retinal vasculature takes on a holangiotic pattern (compact plexus
of blood vessels located in the major part of the light-sensitive portion of the retina).
Whereas, in rabbits, the retinal vasculature takes on a merangiotic pattern (having blood
vessels localized to a specic region of the retina).
The oxygen levels in the retinal vasculature [Kashani et al. 2011] and the intraretinal
oxygen levels have been well studied in a variety of animal models[Wangsa-Wirawan and
Linsenmeier 2003; Linsenmeier 1986; Yu and Cringle 2001]. In cats, the oxygen tension
is around 70mmHg in the choroid, 0-20mmHg in the outer retina, and 10-20mmHg in
the inner retina. By using Fick's laws of diusion, these studies have quantied the
retinal oxygen consumption in the various layers of the retina, under dark and light
conditions[Wangsa-Wirawan and Linsenmeier 2003].
11
Figure 2.2: Anatomy of the ocular circulation. (A) Cut away drawing of the human eye
along the superior{inferior axis through the optic nerve, showing the vascular supply to
the retina and choroid. The retinal vessels are supplied by the central retinal artery. a,
artery; v, vein; n, nerve. (B) Drawing showing the vasculature of the retina and choroid.
Retinal arterioles and venules lie on the vitreal surface of the retina while capillary plexi
lie in just beneath the surface and in the inner nuclear layer. [Newman 2013]
The carbohydrate store for aerobic and anaerobic respiration comes from the ocular
blood circulation. The vitreous also can serve as a reserve of carbohydrates.
2.2 Vitreous Physiology
The vitreous makes up approximately 80% of the volume of the eye and is the largest
single structure in the eye. The vitreous can be broken up into three parts; liquid vitreous,
12
Figure 2.3: Intraretinal oxygen proles across cat retina during light and dark adapta-
tions. The retina is shown schematically at the top. The 4 cell types shown are (left to
right) retinal pigment epithelial cells rod photoreceptors, bipolar cells, and ganglion cells
[Wangsa-Wirawan and Linsenmeier 2003]
gel vitreous and posterior vitreous cortex (also called posterior vitreous hyaloid)[Sebag
2014].
The human gel vitreous is maintained by a network of ne collagen brils (collagen
II, V/XI, and IX)[Seb 2014] suspended in a network of hyaluronic acid. Hyaluronic acid's
volume goes from 0.66 cm
3
/g to 2000-3000 cm
3
/g as it goes from its un-hydrated form to
its hydrated form. This expansion in volume is responsible for a high transmission of light
with minimal scattering [Jampol 2003]. The posterior vitreous cortex is a denser version
of the gel vitreous and is the only part of the vitreous that contains cells (hyalocytes).
It is 100-300μm thick and is attached to the basal lamina of the retina and ciliary body.
While the rest of our vitreous starts o as gel vitreous, over time, the rheology of it
changes and the gel vitreous is replaced by liquid vitreous[Sebag 2014].
The vitreous anatomy is described in the Figure 2.6 [Skeie and Mahajan 2011]. The
collagen bers are most concentrated at the vitreous base and least concentrated in the
central vitreous body. At the vitreous base, the bers are lined perpendicular to the eye
13
Figure 2.4: Ultra-structure of hyaluronic acid/collagen interaction in the vitreous [Jampol
2003]
Figure 2.5: Ultra-structure of vitreous [Jampol 2003]
wall, resulting in strong mechanical vitreo-retinal adhesion in the periphery [Seb 2014].
The posterior vitreous cortex, described above, while absent over the optic disc, has
strong vitreoretinal adhesion along the borders of the optic disc [Seb 2014].
The vitreous serves as a clear optical medium in the eye. In humans, the healthy
intraocular pressure range is between 12 and 22 mmHg. The vitreous also exhibits vis-
coelastic properties which allow it to absorb energy rapidly and release it slowly; an
ability that protects the retina in head trauma and prevents sudden changes to vitreous
volume[Sebag 2014]. The vitreous also serves to maintain the molecular homeostasis of
14
Figure 2.6: Vitreous Anatomy [Skeie and Mahajan 2011]
the eye. It contains inorganic substances (such as sodium, potassium, calcium, magne-
sium, chloride, and phosphate) and inorganic substances (such as ascorbate, glucose, and
lactate)[Jampol 2003].
Of particular interest to us is ascorbate, also known as ascorbic acid or Vitamin-
C, which acts as an antioxidant in the eye. Ascorbate accumulates in ocular tissues
of many animal (and human) species at a concentration several times higher than the
plasma level[Rose and Bode 1991]. Now, ascorbate is essential for many processes, such
as synthesis of collagen, maintaining prosthetic metal ions in their reduced forms, and
scavenging free radicals to prevent oxidative damage to tissues [Padh 1991]. We know
that it plays an important role in preventing oxidative damage caused by free radical
formation during solar radiation [Rose, Richer, and Bode 1998]. It was more recently
found that ascorbate plays a role in maintaining the low intraocular oxygen environment
by reacting with oxygen in the vitreous humor through a free radical process [Shui et al.
2009].
Superoxide dismutase is also a powerful antioxidant which reacts with a high rate con-
stant with superoxide radicals, 2.7*10
5
Ms
-1
[Behndig et al. 1998]. However, extracellular
superoxide dismutase is found in low concentrations in the vitreous humor. For example,
15
if total SOD activity of all isoenzymes is used for the calculation, the pseudo-rst-order
rate constant for SOD in human vitreous is 50s
-1
compared to that of ascorbate in bovine
vitreous, which is 230s
-1
[Behndig et al. 1998]. It is for this reason that in the vitreous,
ascorbate is the predominate antioxidant.
As the human vitreous ages, it undergoes 2 types of changes; gel liquefaction (synchisis
senilis) and vitreoretinal interface weakening[Sebag 2014]. As shown in Figure 2.7, from
the age of 40, the volume of liquid vitreous increases as the volume of gel vitreous steadily
decreases. While the reason for this is unknown, the leading theory explores the role of
the hyaluronan-collagen association[Seb 2014]. Vitreoretinal interface weakening results
in a posterior vitreous detachment. However, even in a complete vitreous detachment,
the vitreoretinal interface does not separate at the vitreous base.
Figure 2.7: Rheology of human vitreous [Jampol 2003]
Transport of molecules within the vitreous humor occurs via diusion and convection.
Diusion occurs due to the Brownian motion of molecules and convection occurs due to
the movement of the eye. Even when focused on a xed point, the eye constantly executes
a series of extremely rapid angular rotations (500°s
-1
) known as saccades[Rayner 1998].
These, along with regular eye tracking motions, create
ow patterns in the vitreous
16
humor[Siggers and Ethier 2012]. In order to computationally model these processes,
researchers are investigating ways to characterize the
uid mechanic parameters of these
processes.
For example, researchers have used a variety of methods to measure the diusion
coecient of drugs and drug surrogates in the vitreous humor [Gisladottir, Loftsson, and
Stefansson 2009; Penkova et al. 2014; Xu et al. 2000]. Gisladottir postulated, and exper-
imentally showed, that diusion through vitreous would be slower compared to diusion
through saline due to a lower diusion constant. This was explained with the Stokes-
Einstein equation which predicts that diusion is inversely correlated with viscosity of a
medium. Xu's data also suggested slower diusion of molecules in vitreous humor as com-
pared to diusion through saline, at least for molecules with molecular weight of 300g/mol
or larger. Researchers when modeling the transport of small molecules like oxygen, have
assumed that its diusion coecient in vitreous is the same as it is in saline[Filas, Shui,
and Beebe 2013].
They have also studied the
uid
ow patterns of a moving eye, using eye models and
vitreous substitutes [Repetto, Stocchino, and Caerata 2005; Rossi et al. 2012]. However,
these processes are further complicated by vitreous liquefaction. While there are several
papers describing computational models of intravitreal drug transport, we are a long way
away from fully characterizing intravitreal transport[Balachandran and Barocas 2008;
Balachandran and Barocas 2011; Smith et al. 2012].
Intravitreal oxygen levels in healthy rabbit and human eyes have been dened by sev-
eral groups[Holekamp, Shui, and Beebe 2005; Shui et al. 2006]. Intravitral oxygen levels
in rabbits and humans after vitrectomy have also been dened[Barbazetto et al. 2004;
17
Holekamp, Shui, and Beebe 2005; Lange et al. 2011; Simpson et al. 2013]. Vitrectomy
increases oxygen tensions in human eyes as opposed to in rabbit eyes. This might po-
tentially be explained by the low ascorbate concentration in rabbit vitreous (0.46mM)
compared to the high concentration of ascorbate in human vitreous (2mM)[Jampol 2003;
Shui et al. 2009]. Vitrectomy results in a lower ascorbate concentration and ascorbate is
involved in the consumption of oxygen in the vitreous[Shui et al. 2009].
Figure 2.8: Intravitreal oxygen map. Values obtained from prior studies[Holekamp, Shui,
and Beebe 2005; Simpson et al. 2013; Lange et al. 2011; Shui et al. 2006; Quiram et al.
2007; Stef ansson, Novack, and Hatchell 1990].
The relative hypoxic state of the intraocular environment has led to the hypothesis
increases in intraocular oxygen levels can lead to pathologies such as nuclear sclerotic
cataract, and open angle glaucoma[Beebe et al. 2014]. Higher than normal oxygen levels
in the eye has been associated with side eects such as nuclear sclerotic cataract and
open angle glaucoma. Palmquist et al found out that out of 15 patients who underwent
hyperbaric oxygen therapy (HBO; breathing in 100% oxygen gas at> atm), all developed
early stage nuclear opacities or had progression of existing nuclear opacities, while none of
the control subjects suered similar side eects. [Palmquist, Philipson, and Barr 1984].
18
Giblin experimentally showed that increased nuclear light scattering was produced in
lenses of older guinea pigs by treatment of the animals with HBO but not with hyperbaric
air. He also observed that increased nuclear light scattering developed more rapidly in
guinea pigs exposed to longer periods of HBO[Giblin et al. 1995].
2.3 Ischemic Retina Physiology
Retinal ischemic diseases are some of the leading causes of blindness worldwide[World
Health Organization 2014]. Examples of these diseases include diabetic retinopathy, reti-
nal vein occlusion, sickle-cell disease, ocular ischemic syndrome, and retinal detachment.
Diabetic Retinopathy (DR) and Retinal Vein Occlusions (RVO) represent two of the most
prevalent amongst these diseases. With the incidence of diabetes throughout the world
projected to rise from 150 million to approximately 300 million by the year 2025, DR
represents a major threat to our global population and will likely present ever-increasing
burdens on the health care delivery system.
Studies have shown that chronic hyperglycemia contributes to the pathogenesis of
DR[Pautler 2010]. However, the exact mechanism through which hyperglycemia changes
the structure and cellular composition of the retinal microvasculature is not fully un-
derstood. But the changes in the retinal microvasculature include the thickening of
19
the retinal capillary basement membrane, pericyte apoptosis, and endothelial cell dam-
age[Pautler 2010]. This can lead to the formation of saccular capillary aneurysms, non-
perfusion and obliteration of capillaries[Engerman 1989]. The damage to retinal vascu-
lature leads to retinal ischemia, which stimulates neovascularization mediated by angio-
genic factors[Ciulla, Amador, and Zinman 2003]. The formation of new vessels leads to
the change in classication from non-proliferative DR to proliferative DR.
Retinal venous occlusions can be broken into two types; Central retinal vein occlusion
(CRVO) and branch retinal vein occlusion (BRVO). The tract of vein passing through the
lamina cribrosa is the most common site of occlusion[Prisco and Marcucci 2002]. BRVO
typically occurs at common arterio-venous crossing point[O'Mahoney, Wong, and Ray
2008]. Retinal vasculature occlusions lead to retinal ischemia.
The basic pathophysiology of above mentioned retinal ischemic diseases is ischemia
(lack of adequate blood supply) of the inner retina. Ischemia can lead through necrosis or
apoptosis through a series of events called the `Ischemic Cascade' (Figure 2.9) [Osborne
et al. 2004]. Ischemia rst leads to hypoxia, which then leads to energy failure. This
results in changes to intracellular calcium and sodium concentrations as well as free
radical production. These nally lead in cell death.
Hypoxia is dened as an inadequate supply of oxygen to the retina. While hypoxia is
necessary for the formation of a vascular system in embryos, in disease conditions, hypoxia
represents a mismatch in oxygen supply versus demand at the cellular level within ocular
tissues. This in turn leads to an increase in hypoxia inducible factor (HIF)[Osborne et al.
2004].
20
Figure 2.9: Summary of hypothesized key events in the process of ischemic retinal neu-
rodegeneration[Osborne et al. 2004].
HIF-1 protein is expressed in mammalian cells. In normoxic conditions, is degraded
by cellular proteasomal complexes. Under hypoxic conditions, levels of HIF-1 protein
build up. This promotes dimerization with HIF-1 and translocation to the nucleus where
the HIF-heterodimer acts as a transcription factor. The result is an increase in expres-
sion of genes critical to the promotion of survival in hypoxic conditions. This includes
vascular endothelial growth factor (VEGF), stromal-derived factor-1(SDF-1), and EPO,
which promote neovascularization[Osborne et al. 2004]. VEGF is produced by many cell
21
types in the retina and is critical to maintaining normal ocular function. However, over-
production of VEGF can lead to non-physiological neovascularization, which results in
further complications.
Currently, the standard of care for ischemic retinal diseases involves using monthly
intraocular medications (i.e.: anti-vascular endothelial growth factor (anti-VEGF)), de-
structive laser to the retina (pan-retinal photo-coagulation (PRP)) and/or surgery (pars
plana vitrectomy (PPV))[Giuliari 2012]. There are problems with these procedures how-
ever as described below.
Intraocular anti-VEGF medications have been shown to decrease the VEGF levels
and vascular permeability but need to be administered monthly and may increase the
risk of retina ischemia. O-label use of intravitreal steroid injections can lead to elevated
intraocular pressure and cataract formation and/or progression.
The goal of PRP is to ablate (through thermal damage) ischemic areas of the pe-
ripheral retina and reduce induction of angiogenic growth factors[Ciulla, Amador, and
Zinman 2003]. PRP laser treatment increases preretinal and intraretinal O2 concen-
tration but is highly destructive, burning the neural surface of the retina with visible,
blanching marks. Side eects of this include constricted peripheral visual eld as well as
delayed dark adaptation. Moreover, laser therapy for macular edema can lead to central
scotomas (blind spots) and scarring[Fong, Girach, and Boney 2007].
PPV surgery itself increases preretinal O2 but those changes are insucient to supply
the ischemic retina and can divert oxygen from the anterior structures of the eye leading
to formation of neo-vessels in the iris with higher risk of bleeding inside of the eye and
elevated intraocular pressure.
22
In summary, the evidence above shows that current treatments, especially for mod-
erate to severe ischemic retinal vascular disease, can be improved and if the trend of
treating these conditions via intravitreal injection continues to increase, the treatments
will also be very expensive and burdensome for the health care system[IVT Injections:
Health Policy Implications].
2.4 Oxygenator
We have invented a metabolic prosthesis, called the Oxygenator, to ll the current gaps
in the treatment of retinal ischemic diseases (Figure 2.10). The Oxygenator generates
controlled amounts of oxygen and delivers it to the ischemic tissue. It does so by using
electricity to split water into oxygen and hydrogen inside a gas permeable membrane. This
oxygen is transported out of the membrane and delivered to the tissue. The local delivery
of oxygen allows patients to experience the benets of local oxygen therapy without the
disadvantages of systemic oxygen therapy. The Oxygenator has the potential to provide
long-term therapy for patients with moderate to severe DR and RVO, who currently are
not beneting from existing therapies, as well as potentially improve outcomes and reduce
side eects in patients undergoing existing therapies.
The nal form of the Oxygenator will have an implanted unit and a wearable wireless
powering unit. The implanted unit will consist of an extraocular body, a cannula, and
a intraocular component. The extraocular body will house a receiver coil (for wireless
power transmission), electronics and a drug reservoir. This will all be packaged inside a
hermetically sealed titanium body. The cannula, which will connect the extraocular body
23
Figure 2.10: (Top Left) Oxygenator Intraocular unit: Oxygen permeable MEMS
membrane-bag with micro electrodes. (Top Right) Position of MEMS unit on retina.
(Bottom Left) Oxygenator Extraocular unit: The device has an electronic capsule, an
electrolyte reservoir that is connected to the intraocular unit through a cable, and an
electrolyte rell port. The device is miniaturized with a low prole to t in the sub-
conjunctival space and can be relled using an automated pump (Bottom Center) Single
wearable, ergonomic,
exible, low prole design for power transmission (Bottom Right)
Flexible, and comfortable band-aid design for night time use
to the intraocular component, will serve as a
uidic channel and be manufactured with
a bio-compatible silicone material. The intraocular component, which will also be made
out of a bio-compatible silicone, will be semi-permeable to oxygen gas and will serve as
the intraocular source of localized oxygen to the ischemic tissue.
The rationale behind the Oxygenator is to prevent tissue from experiencing hypoxia,
and thus halting the ischemic cascade. The retinal cells can survive up to 90 minutes
in ischemic conditions, so the Oxygenator will serve as a preventative therapy[Osborne
et al. 2004]. A hypothesis for why the retina's tolerance for ischemia is higher than
the brain's tolerance is the ability of the retina to extract energy from other sources
24
anaerobically[Osborne et al. 2004]. While the Oxygenator only prevents hypoxia, the
tissue still experiences a shortage of glucose (necessary for energy production), and a
buildup of waste products. However, studies have shown that the vitreous can serve as a
reserve of glucose[Osborne et al. 2004].
There are studies that have validated the Oxygenator's principles. Nguyen et al
showed that supplemental systemic oxygen therapy delivered through nasal cannula helps
improve health of the retina and vision in humans suering from Diabetic Macular
Edema[Nguyen 2004]. In the pilot study, patients continuously received 4ml/min of oxy-
gen via nasal cannula, attached to portable oxygen tanks, for a period of 3 months. At
the end of the study, there was a signicant reduction in the thickness of the center of
the macula. However, systemic oxygen therapy is not practical, due to the necessity of
carrying combustible oxygen tanks and the unwanted side eects associated with systemic
therapy.
Abdallah and colleagues demonstrated in vivo, the ecacy of supplemental and local
delivery of oxygen to the retina in cases of ischemia. After 90 minutes of retinal ischemia,
induced by elevated intraocular pressure, oxygen was generated in the anterior vitreous.
In the treated group, preretinal oxygen tension remained elevated compared to sham and
control groups over 12 days. The full eld ERG b wave amplitude remained constant over
2 weeks, compared to the sham and treatment groups, which experienced a decline in b-
wave amplitudes. Histology showed that the oxygenated eyes maintained good retinal
thickness and well-preserved anatomy compared to sham/control eyes. [Abdallah et al.
2011]. The retina was then studied under a microscope to determine the eectiveness of
our oxygen therapy. Evaluation of multiple sections showed that nearly 90% of the normal
25
retinal thickness was preserved in the oxygenated group, whereas retinal thicknesses in
the sham and non-treated group were only 20% to 30% that of normal.
There are several scientic questions that need to be addressed before bringing this
therapy to the patients' bedside.
2.5 Disease Model
To perform in vivo experiments with the Oxygenator to test its ecacy, we need an
appropriate animal model with retinal ischemia. Diabetic animal models are inecient
because it takes many years to induce even mild DR in pancreatectomized animals with
similar ocular dimensions as humans. Thus, an alternative way of creating retinal is-
chemia in animals is required. A number of animal models have been developed to study
the evolution of retinal ischemia and to explore potential treatments to prevent neu-
ronal degeneration[Man, Kwong, and Caprioli 2010]. Monkeys, pigs, dogs, cats, rabbits,
guinea pigs, rats, and mice have been used as animal models of retinal ischemia. Several
techniques, such as pressure elevation, laser photothrombosis, and optic nerve ligation,
have been used to create retinal ischemia in these animals[Minhas, Morishita, and Anand
2012]. The table below summarizes the various animal models of retinal ischemia (Table
2.1).
Retinal vein occlusion induced secondary to photothrombosis (pRVO) in rabbits and
dogs is a well established procedure[Ameri et al. 2008a; Hayreh, Heuven, and Hayreh
1978; Tameesh et al. 2004; Jaime et al. 2012].
26
Disease
Model
Procedure
Degree of Is-
chemia
Advantages Disadvantages
Pressure
Elevation
Raise IOP
above
systolic
pressure
Complete
(Choroidal
and retinal)
Easy method
Global ischemia
Human DR and
RVO does not re-
sult in global is-
chemia
Vessel liga-
tion
Suture op-
tic nerve,
ophthalmic
vessels
Complete
(Choroidal
and retinal)
Global ischemia
No impact on vit-
reous cavity
Dicult method
Human DR and
RVO does not re-
sult in global is-
chemia
Endothelin
administra-
tion
Intravitreal
vasocon-
strictor
injection
Variable
Easy method
Closer to human
DR and RVO dis-
ease states
Degree of is-
chemia variable
from subject to
subject
Unable to control
degree of ischemia
Photo
thrombosis
of retinal
vessels
Rose Ben-
gal and
light illumi-
nation
Partial in-
ner retinal
ischemia
Easy method
Able to control
area of ischemia
Closer to human
DR and RVO dis-
ease states
Degree of is-
chemia variable
from subject to
subject
Table 2.1: Summary of various animal models of retinal ischemia
27
pRVO works by injecting Rose Bengal (4,5,6,7-tetrachloro-2',4',5',7'-tetraiodo
uorescein),
which is a chloride substituted xanthene type dye into the body. When it is exposed to
high intensity light at 550nm, oxygen radicals and oxygen singlets are produced. Sin-
glet oxygen destroys cellular membranes and damages the endothelial cell membranes in
the retinal vasculature. This provides the stimulus for platelet aggregation and occlu-
sion[Nanda et al. 1987].
This pRVO model has a number of unique properties and advantages over other
models. First, pRVO can be made in segmental patterns that are more representative
of ischemic diseases than total retinal ischemia. Second, pRVO can be induced with
lasers and do not require invasive surgical methods that can be intrinsically confounding.
Third, the pRVO animal models cause signicant retinal edema and hemorrhage as well
as capillary non-perfusion in rabbits, dogs, and cats and are therefore representative of
human RVOs. Fourth, histopathology demonstrates complete occlusion of the vein lumen
with platelet thrombosis and degenerating red blood cells. It is important to note that
all pRVO models, regardless of species, show eventual recannulation [Royster et al. 1988].
However, repeated laser treatments can maintain occlusion for longer periods. Fifth, the
pRVO model is well-established in both merangiotic retina (rabbit) and holangiotic retina
(dog)[Ameri et al. 2008a; Hayreh, Heuven, and Hayreh 1978; Tameesh et al. 2004].
Intravitreal oxygen levels after RVO are not fully described for animals and humans
but several groups have performed intraretinal oxygen measurements after laser retinal
vein/artery occlusions. They demonstrate that the inner retinal oxygen tension drops to
0 mmHg in the ischemic retina after occlusion[Yu et al. 2007].
28
Figure 2.11: Intraretinal oxygen measurement after laser occlusion of retinal circula-
tion[Yu et al. 2007].
2.6 Energy Harvesters for Implants
One of the rst implantable medical devices was the cardiac pacemaker[\Development
of an implantable cardiac pacemaker"]. The rst pacemakers required large amounts of
power and stimulated the heart from outside the body. These were then replaced by low
power versions which used trans-cutaneous electrodes in direct contact with the heart.
By using an on-board battery, these pacemakers became fully implantable.
Today there are several implantable medical devices such as the retinal prosthesis,
glucose sensor, and cochlear implants[Smith 2009; Weiland and Humayun 2008]. The
energy sources for these devices range from rechargeable implanted batteries to external
batteries coupled wirelessly. Battery technology has greatly evolved over the years and
29
today's lithium-ion batteries have large energy densities. The ability to recharge these
batteries while implanted in the body is important since replacement of batteries requires
costly hospital visits and surgery. Thus several methods of recharging implanted batteries
have been implemented[Olivo, Carrara, and De Micheli 2011].
The most prevalent method of wireless power transfer is through inductive cou-
pling[Wang et al. 2006]. This method is popular because the time varying magnetic
elds have little interaction with biological tissue. Inductive links can provide power in
the tens of milliwatts range[Silay et al. 2009; Wang et al. 2006]. The disadvantage of
inductive coupling is that the external telemetry device needs to be placed close to the
implant which can be uncomfortable to the patient and limit his/her movement.
Energy harvesters have the ability to reduce the power constraints placed on implanted
medical devices. Energy harvested from the environment has been converted to electrical
energy to support the power grid for several years. This concept can also be applied to
biomedical implants. Scavenging energy from the environment has the ability to reduce
the number of times the user will need to recharge the battery. This could potentially
improve biomedical implants' ease of use amongst patients. It can also potentially lead
to battery-less `passive' implants.
Kinetic harvesters aim to collect the energy related to human motion. These har-
vesters can be classied into piezoelectric, electromagnetic, and electrostatic harvesters[Cadei
et al. 2014]. Electromagnetic harvesters usually are comprised of a permanent magnet
and a coil. Based on the Faraday-Neuman-Lenz law, relative motion between the coil
and the magnet generates a voltage via a varying magnetic eld. Electrostatic harvesters
30
generate a voltage based on moving the plates on variable capacitors. Piezoelectric har-
vesters utilize the property of piezo materials which generate a potential dierence when
subject to mechanical vibrations. These kinetic harvesters are most ecient are unique
oscillating frequencies. Thermoelectric harvesters exploit the Seebeck eect. When there
is a temperature gradient between two dierent metals or semiconductors, a voltage drop
is generated between them. Bio-fuel cells electrochemically convert glucose to electrical
energy[Hannan et al. 2014].
Energy harvesters using solar energy have not been investigated in detail because
of sunlight's poor trans-cutaneous transmission[Anderson and Parrish 1981]. However,
optical power transfer with near infra-red light (NIR) has been studied as a method to
power pacemakers[Goto and Nakagawa 2001]. The sun's spectral irradiance is shown in
Figure 2.12. Xenon lamps are commonly used to mimic the sun's irradiance spectrum.
Figure 2.12: Spectral irradiance of the sun compared to various lamps[Newport 2016].
The sun's peak radiation level on the Earth's surface is 1 kWm
-2
. This light intensity
varies according a variety of factors; time of the day, cloud cover, day of the month etc.
Photovoltaic cells (solar cells) are PN junctions made out of semiconductor materials
(Figure 2.13). When photons of appropriate energy impinge upon the depletion region
created by the PN junction, electron-hole pairs are created. This is created due to the
31
energy transfer from the photon to the electron in the valence band. This energy causes
the electron to move to the conductance band, leaving behind a hole carrier. The electron
and hole are separated by the junction's built in electric eld. When a wire is placed
from the cathode (N-type) to the anode (P-type), electrons will
ow through the wire.
The open circuit voltage of a photovoltaic cell increases with band-gap.
Figure 2.13: Illustration describing the function of a photodiode
[http://www.imagesco.com/articles/photovoltaic/photovoltaic-pg4.html]
Quantum eciency is dened as the ratio of the number of carriers collected by the
solar cell to the number of photons of a given energy incident on the solar cell. Dierent
materials have dierent quantum eciencies at dierent wavelengths (Figure 2.14). The
long wavelength cuto is due to photons not having enough energy to excite an electron
from the valence to conductance band. The short wavelength cuto occurs because most
of the photons are absorbed at the surface, where recombination time is short, resulting
in very few electron-hole pairs to reach the depletion region prior to recombination.
Now, the current and voltage from a photovoltaic array will generally require to a
buck or boost converter to ensure that the voltage is appropriate for charging a battery.
In my case, I'm using a boost converter and it's useful to understand how boost converters
work (Figure 2.15). During the OFF cycles of the pulse width modulated (PWM) signal,
energy is stored in the capacitor. During the PWM ON cycles, the energy is stored in
32
Figure 2.14: Quantum eciency versus wavelength for dierent photocell materi-
als[Theilmann 2012]
the inductor. When the PWM is OFF again, the energy from the inductor is stored in
the capacitor, increasing its voltage beyond that of the power supply. By storing energy
in an inductor and capacitor, the boost converter is able to eciently increase the output
voltage relative to the input voltage.
Figure 2.15: Boost converter
Power from the photovoltaic cells cannot directly power electronics. Electronics re-
quire a constant source of energy and energy harvesters are unable to provide constant
power. For example, a photovoltaic array can only generate power when it is exposed to
33
sunlight. Thus, batteries are required and the photovoltaic cell can recharge the battery
and the battery in turn, can provide a constant source of power to the electronics.
We will focus on Lithium Ion batteries due to their high power densities (0.16 Wh/g)
[Battery University]. These batteries function due to lithium ion migration from one
electrode to another. During charging, the inter-layers of the anode are doped with
Lithium ions and the inter-layers of the cathode are de-doped. During discharging, the
opposite reactions occur.
Figure 2.16: Lithium ion battery [Battery University]
The charging of the lithium ion battery can be broken into 2 main phases, 1) constant
current, and 2) constant voltage [Battery University]. The current and voltage proles
are depicted in Figure 2.17. It is essential to regulate the voltage of the charger when
recharging lithium ion batteries. Overcharging can lead to gas production at the cathode
and high pressures within the cell which can be hazardous.
34
Figure 2.17: Charge stages of a lithium ion battery. Li-ion is fully charged when the
current drops to a predetermined level or levels out at the end of Stage 2 [Battery Uni-
versity].
2.7 Implantable Oxygen Sensor
Medical, agricultural, waste management, and food packaging industries all use dissolved
oxygen sensors[Ramamoorthy, Dutta, and Akbar 2003]. The two primary types of dis-
solved oxygen sensing technologies are the electrochemical oxygen sensor and optical
oxygen sensors. The former operates by electrochemically reducing oxygen dissolved in
solution and measuring the resulting current. The latter operates by measuring the in-
tensity or lifetime of a
uorescent molecule in the presence of oxygen molecules. Optical
sensors had signicant advantages of electrochemical ones due to their fast response times,
potentially lower energy consumption, and minimal reactivity with oxygen[Ramamoorthy,
Dutta, and Akbar 2003].
Luminescence based oxygen sensors use the quenching of the dye's luminescence by
oxygen as the transduction mechanism, which is described in the general equations 2.2
below[\Luminescence-Based Oxygen Sensors"].
35
D +hv!D

(2.1)
D

+O
2
!D +O
2
(2.2)
There are two main types of quenching: static and dynamic. Static quenching involves
a reaction between the analyte and the dye molecule. Dynamic quenching is a collision
based quenching method where the excited, high energy dye molecule collides with an
oxygen molecule and transfers some of its energy to the oxygen molecule[\Luminescence-
Based Oxygen Sensors"].
In a non ratiometric dye, the luminescent molecule electrons gain energy from photons
and as they return to their previous energy level, they emit photons. In a ratiometric
dye, the
uorophore is excited by photons and electronic energy is partially transferred
to the phosphore, resulting in light emission at two dierent wavelengths.
Figure 2.18: Jablonski diagram: E { energy, A { absorption, F {
uorescence, P {
phosphorescence, IC { internal conversion, ISC { intersystem crossing, VR { vibrational
relaxation, S { singlet state, T { triplet state[Feng et al. 2012].
36
Fluorescent oxygen probes change intensity in response to oxygen molecules. This
intensity change is due to changes in extinction coecient of the
uorescent dye when
oxygen concentration varies. The ratio of phosphorescence intensity to that of
uorescence
in the absence and presence of oxygen is linked to oxygen tension through the Stern-
Volmer equation (equation 2.4)[Yoshihara et al. 2012].
Ratiometric oxygen probes
I
0
p=I
0
f
Ip
=If
= 1 +K
SV
[O
2
] (2.3)
Non-ratiometric oxygen probes
I
0
I
= 1 +K
SV
[O
2
] (2.4)
Another method of quantifying oxygen content is lifetime-based sensing[\Luminescence-
Based Oxygen Sensors"]. The
uorescence lifetime of a sample is the average duration
of time the
uorophore remains in its excited state. Following pulsed excitation, the
intensity decay of oxygen sensitive dyes are usually a single exponential.
I(t) =I
0
e
t
=
(2.5)
While there are many molecular interactions which can in
uence
uorescence decay
times, we are most interested in collisional quenching (oxygen acts as a quencher). By
37
studying the change of lifetime of the dye, we can determine the concentration of oxy-
gen[Szmacinski and Lakowicz 1994].

0

= 1 +
0
k
q
[O
2
] (2.6)
Initially, for non-ratiometric dyes, relying on oxygen induced changes in emission in-
tensity proved to be problematic. Emission intensity can easily be aected by a number of
reasons such as, imprecise or dirty optical surfaces, changes to the optical alignment from
sample to sample, changes in concentration of dye at the detection site, photo bleaching of
the dye, and variance in sensor materials[Szmacinski and Lakowicz 1994]. Non-ratiometric
dyes suered from the problem of referencing the intensity measurements. This resulted
in the need for frequent calibrations and other corrections. This issue was overcome with
the advent of wavelength ratiometric dyes[Wang et al. 2010]. Here the ratio of signals at
two emission wavelengths is used for quantitative measurements of oxygen concentration.
The reference signal created by the emission spectra of the oxygen insensitive molecule
eliminates problems with non analyte induced intensity changes[Kostov et al. 2000].
38
Chapter 3
Spatial Variation in Vitreous Oxygen Consumption
3.1 Goals
For intraocular oxygen therapy to reach the patient's bedside, we must answer the ques-
tion, what is the safe and eective therapeutic window for oxygenation of the ischemic
retina? Too little oxygen and the therapy is ineective[Hiramitsu et al. 1976]. Too much
oxygen puts the retinal tissue under oxidative stress. Before we can answer that question
however, since our therapy works by delivering oxygen through the vitreous, we must
fully understand the oxygen mass transport dynamics in the vitreous humor. Dening
the spatial and temporal dynamics of intravitreal oxygen transport will enable us to quan-
tify the exact concentration of oxygen that reaches the ischemic parts of the retina from
our device. Current empirical and computational models of oxygen transport within the
vitreous only consider the retina as an oxygen source in the posterior chamber [Shui et al.
2006; Filas, Shui, and Beebe 2013]. Can these models be extrapolated to accurately de-
ne oxygen transport from an external source of intravitreal oxygen? Furthermore, while
39
we have evidence that the vitreous consumes oxygen in an ascorbate dependent manner,
we need to better understand the reaction kinetics of this mechanism[Shui et al. 2009].
In this chapter, we take the rst step towards that goal by investigating the spatial
dynamics of vitreous oxygen consumption in porcine cadaver eyes. The reaction kinetics
of the ascorbate-oxygen reactions are studied via measuring the decay of oxygen tension in
vitreous samples. The spatial characteristics of vitreous oxygen consumption are studied
through intravitreal oxygen measured from an oxygen source.
3.2 Methods
For the following experiments, we used fresh porcine cadaver eyes (Sierra Medical Sci-
ence, Whittier CA) that were shipped and used for experiments at USC with 6 hours of
harvesting. These specimens included only the eyeball with minimal amounts of perioc-
ular adventitial tissue and typically did not include the muscle attachments or any other
components of the orbit. Only whole globes from 6 month old pigs were used for these
studies. This ensured that the globes had fully formed vitreous with high gel content and
no liquefaction.
Oxygen Source Fabrication
A custom made device was used in order to create a constant focal source of oxygen
within the cadaver eye's vitreous humor (Fig 3.1). The oxygen source device fabrication
has two main steps: mold fabrication and silicone casting. The mold was fabricated
through negative dry lm photoresist (DuPont) laminated on a fresh silicon wafer. It was
patterned via UV light exposure and developed. For the casting of silicone (MED4-4210,
40
two-part, medical-grade, NuSil Technology LLC) are mixed at a 10:1 ratio by weight,
degassed under vacuum, and applied onto the patterned mold. A hollow stainless steel
tube is inserted through the neck of the oxygen source device to ensure that air permeates
through the base of the oxygen source device and only exits through the tip of the oxygen
source device. This semi-permeable oxygen source device acts as a conduit of air from
the base to the tip and eectively acts as point source of oxygen within the vitreous[Kang
et al. 2015].
Figure 3.1: Oxygen source device. The base is 10 mm in diameter and the tip is 4 mm
in diameter. The hollow stainless steel tube connects the base to the tip. Only the base
and tip are permeable to oxygen[Kang et al. 2015].
Intravitreal Oxygen Diusion Measurements
In order to ensure visibility into the vitreous of the porcine cadaver eyes, the native cornea
and the lens were removed and replaced with a keratoprosthesis (Ocular Instruments).
Next, the custom made oxygen source shown in Fig 3.1 was placed so that the tip of the
oxygen source device was either located in the mid-vitreous or posterior vitreous. The
41
location of the tip was approximated by the length of the cannula inside the eye. The
base of the oxygen source was left outside the globe and exposed to atmospheric air. A
schematic is shown in Fig 3.2. Next, a 23 gauge valved trocar (Alcon) was implanted
180

across the semi-permeable oxygen source. This positioning allowed for placement
of a commercially available oxygen measuring probe (Oxford Optronix, OxyLab), here
on referred to as oxygen probe, in proximity of the oxygen diuser. The 2 positions of
the trocar are illustrated in Fig 3.2. We inserted the oxygen measuring probe through
the trocar and positioned it at various distances from the oxygen source. For our control
experiments, the oxygen source was removed after probe was positioned. At the start of
the experiment, the probe was positioned 0 mm from the surface of the oxygen source tip.
The eyes were then left in a dark environment for an hour for the oxygen diusion to reach
steady state. After an hour, recordings were performed. At each distance location, oxygen
tension is recorded for 1.5 minutes. In order to verify that the probe was 0 mm from the
surface of the oxygen source tip, we measured oxygen tension at the surface of the oxygen
source tip and ensured that it was equal to that of air. We used a micromanipulator
(Edmund Optics, Linear translational stage) to retract the probe to desired distances (0,
1, 2, 3, 4 mm) away from the oxygen source tip. The oxygen probe has a resolution of
0.1 mmHg and an accuracy of10% when partial pressure of oxygen (pO
2
) is below 150
mmHg and20% when the pO
2
is above 150 mmHg.
Vitreous Oxygen Consumption Rate Measurements
We dissected vitreous humor from several porcine cadaver eyes and placed them in a
beaker. Lens material, retinal tissue, and choroid tissue were carefully removed during
42
Figure 3.2: Ex vivo porcine eye preparation and intravitreal oxygen measurement meth-
ods. Left: Oxygen source is positioned in the mid-vitreous and the oxygen probe is
retracted in the direction shown by the arrow. Right: Oxygen source is positioned in
contact with the retinal tissue and the oxygen probe is retracted in the direction shown
by the arrow. The probe is positioned such that it stays in the posterior vitreous region.
The trocar is used to facilitate oxygen probe entry and retraction without creating any
motion artefact. The dashed lines indicate regions that are designated has mid vitreous
and posterior vitreous.
the dissection procedure. We used a magnetic stirrer to homogenize the vitreous for
5 minutes. This ensured that the oxygen tension in the vitreous was equal to that of
air (160 mmHg). Vitreous samples were collected with a vitrector (Alcon Constellation
Vision System). The vitreous was then transferred to a clear glass beaker and covered with
a rubber stopper. The glass vial was tightly sealed with a crimper (IVPACKS LLC). Care
was taken to minimize the pocket of air within the beaker. We introduced the oxygen
probe through the rubber septum with an 18 gauge needle and recorded the decay of
vitreous oxygen tension until oxygen tension reached 10 mmHg. This is representative
of the physiological levels of oxygen in the vitreous. At the end of the experiment,
vitreous was again collected with a vitrector and the ascorbate concentration dierences
in the samples were analyzed. In a few samples, ascorbate oxidase was added to the
43
vitreous sample. Ascorbate oxidase reduces ascorbate present in the vitreous, slowing
down vitreous oxygen consumption.
Ascorbate Measurements
We carefully dissected out samples from the vitreous core and posterior vitreous of porcine
cadaver eyes. To ensure visibility into the vitreous, we removed the cornea, anterior lens
capsule and extracted the lens leaving the posterior lens capsule intact. We attached a
3cc syringe to a vitrector (Alcon Constellation Vision System) and inserted it into the eye
via a 23 gauge trocar. With a cut rate of 5000 cuts per minute, we rst aspirated 200μL
samples from the core (Fig 3.3). Following that, samples from the posterior vitreous were
aspirated. Care was taken to not cut the retinal tissue.
Figure 3.3: : Illustration of vitreous regions that were biopsied for ascorbate measure-
ments. (A) The site of the mid-vitreous sample collection. (B) The site of the posterior
vitreous sample collection.
We tested these samples in duplicate with an ascorbic acid assay kit (Sigma Aldrich
MAK075). In this assay, ascorbic acid concentration is determined using the Ferric Reduc-
ing/Antioxidant and Ascorbic Acid (FRASC) assay. In this assay, Fe
3+
is reduced to Fe
2+
by antioxidants present in the sample, which results in a colorimetric (593 nm) product.
The addition of ascorbate oxidase to parallel samples oxidizes any ascorbic acid present
44
allowing for the measurement of the ascorbic acid concentration. A freshly prepared stan-
dard curve was used for all measurements. The specicity of this method for ascorbate
has been previously validated using gas chromatography-mass spectroscopy[Shui et al.
2009].
3.3 Results
In 6 eyes, the oxygen source was placed in the mid-vitreous (vitreous core) and in another
6 eyes it was placed in the posterior vitreous. Each of the control groups had 4 eyes.
When the oxygen probe was 0 mm away from the oxygen source, it measured the pO
2
of air which was 160 mmHg. In the presence of an oxygen source, mid-vitreous pO
2
and posterior vitreous pO
2
measurements are higher across all distances as compared to
the control pO
2
measurements (Fig 3.4). A two-sided, two-sample t-test was conducted
to statistically compare pO
2
recordings between experiments with an oxygen source and
control experiments. When the oxygen source was placed in the mid-vitreous, there was
a statistically signicant decrease in the oxygen tension with distance from the oxygen
source versus control; (p<0.005 at all distances 0, 1, 2, 3, 4 mm). When the oxygen
source was placed in the posterior vitreous, there was a statistically signicant decrease
in the oxygen tension with distance from the oxygen source versus control as well; 0 mm
(p<0.01), 1 mm (p=0.01), 2 mm (p<0.01), and 3 mm (p=0.018) distances.
Overall, oxygen tension recordings at all distances from the oxygen source were higher
when the oxygen source was placed in the mid-vitreous as compared to when it was placed
in the posterior vitreous (Fig 3.4). A two sided, two-sample t-test was conducted to
45
Figure 3.4: Comparison in oxygen diusion kinetics between mid and posterior vitreous.
Black Squares: Oxygen tension with increasing distance from an oxygen source in the
posterior vitreous. Red Circles: Oxygen tension with increasing distance from an oxygen
source in the mid-vitreous. Oxygen tension recordings are higher in the mid-vitreous
compared to the posterior vitreous as distance from the oxygen source increases. Black
Squares: Oxygen tension recordings with increasing distance from the probe in the ab-
sence of any oxygen source in the posterior vitreous. Hollow Circles: Oxygen tension
recordings with increasing distance from the probe in the absence of any oxygen source
in the mid-vitreous. Hollow Squares: Oxygen tension recordings with increasing distance
from the probe in the absence of any oxygen source in the posterior vitreous. Please note
that the open symbols are not clearly identiable because they overlap in the graph.
compare pO
2
recordings between mid-vitreous and posterior vitreous experiments with an
oxygen source. The mid-vitreous oxygen tension was signicantly higher when compared
to the posterior vitreous oxygen tension values at distances 2 and 3 mm from the oxygen
source (p<0.001).
We modeled our experimental system as a point-source of oxygen diusing radially
in spherical coordinates. Using Fick's laws of diusion, we obtain equation 3.1, where C
is the concentration of oxygen in M (Molar), r is distance away from the source in mm,
and R is the rate of oxygen consumption by the vitreous in Ms
-1
, and D is the diusion
coecient of oxygen in water at 20

C (0.00197 mm
2
s
-1
).
46
dC
dt
=
D
r
2
d
dr
(r
2
dC
dr
) +R (3.1)
Under steady state conditions, equation 3.1 can be reduced to equation 3.2. This
assumption is valid because experimentally, the oxygen tension at various distances at
the time of measurement (after 1 hour time lapse) was constant. Also based on the
characteristic diusion time constant, the diusion process would have reached steady
state.
D
r
2
d
dr
(r
2
dC
dr
) =R (3.2)
In order to determine an appropriate reaction rate equation for R, we measured the
oxygen consumption rate by the vitreous as described in the methods section and illus-
trated in Fig 3.5.
Figure 3.5: Kinetics of oxygen consumption in vitreous samples (red circle) compared to
water (black square) in a sealed chamber and to vitreous treated with ascorbate oxidase
(blue triangles).
47
Over the course of approximately 5.5 hours, the oxygen content in vitreous humor
decayed from 160 mmHg to 0 mmHg. Water was used as a control and the oxygen
tension of water remained at baseline levels throughout the duration of the experiment.
The ascorbate concentration in the vitreous humor was measured before and after the
experiment described in Fig 3.5, using the assay described in the methods section. The
ascorbate content did not decrease by more than 5% of the initial value. A two-sided,
paired, sample t-test was conducted to determine if ascorbate content changed over the
time during which vitreous oxygen levels were found to decrease. We could not nd a
signicant dierence between the ascorbate content before and after the experiment at
= 5%.
The reaction of ascorbate with oxygen is described in equation 3.3. Equation 3.3 is
not meant to suggest a direct two electron transfer from AH

toO
2
in order to generate
H
2
O
2
. The overall reaction involves 2 one electron transfers (equation 3.4 and 3.5). The
oxidation of ascorbic acid has been described as following a general solution of the second
order rate equation as described in equation 3.6.
AH

+O
2
+H
+
!A +H
2
O
2
(3.3)
AH

+O
:
2
+H
+
!A
:
+H
2
O
2
(3.4)
A
:
+O
2
!A +O
:
2
(3.5)
R =
dO
2
dt
=k[AH

][O
2
] (3.6)
48
Since the ascorbate concentration in our experimental setup did not change beyond
5%, it was assumed to be constant. Thus, the rate law equation 3.6 can be combined with
equation 3.7 to yield a single eective rate constant as described in equation equation
3.8.
d[AH

]
dt
= 0 (3.7)
ln[O
2
] = ln[O
2
]
0
k

t (3.8)
wherek

=k(AH

) (3.9)
Fig 3.6 is a plot of ln[O
2
] against time. By tting ln[O
2
] against time, we can obtain the
eective reaction rate constant k
=
ast from the slope of the tted line. The experimental
value for k

is 1:95 10
4
s
1
4:55 10
5
s
1
(SE). By accounting for the ascorbate
concentration in each sample, we obtain the nal reaction rate constantk = 1:61M
1
s
1

0:708M
1
s
1
(SE).
Figure 3.6: Representative results. Plot of ln[O
2
] against time.
49
With the reaction term equation known, we combined equation 3.6 and equation 3.2
to obtain a dierential that can be solved analytically to produce 3.10. See Appendix for
detailed solution.
C(r) =
c
1
e
r
q
AH

k
D
r
+
c
2
e
r
q
AH

k
D
r
q
AH

k
D
(3.10)
To explain the spatial variation in vitreous oxygen consumption, we hypothesized that
the ascorbate concentrations varied spatially. So, we measured the ascorbate content in
the posterior vitreous and compared it with the ascorbate content in the mid-vitreous
(Table 3.1).
Eye Ascorbate
content in
mid-
vitreous(mM)
Ascorbate
content in
posterior
vitreous(mM)
Dierence in
ascorbate
content (mM)
1 0.271 0.453 0.182
2 0.321 0.418 0.098
3 0.172 0.384 0.212
4 0.245 0.234 -0.011
5 0.301 0.39 0.088
Table 3.1: Ascorbate concentration dierences between mid-vitreous and posterior vitre-
ous
Vitreous samples of the vitreous core and posterior vitreous from 5 eyes were obtained
and analyzed for ascorbate content. A one-sided paired sample t-test was conducted to
determine the eect of vitreous location on ascorbate content. There was a signicant
dierence between the ascorbate content in the vitreous core compared to the ascorbate
content in the posterior vitreous (p = 0:02). Ascorbate content is higher in the posterior
50
vitreous as compared to the core. The mean ascorbate content in the mid-vitreous and
posterior vitreous was 0.262 mM and 0.376 mM respectively.
These ascorbate and reaction rate constant (k) values were substituted into equation
3.10 and the resulting curve was tted against the experimental data (Fig 3.7). The
oxygen source was modeled as a 2 mm radius sphere. Distance from the oxygen source
was modeled from the surface of the 2 mm radius sphere. Oxygen diusion coecient
D was assumed to be a constant[17]. Because the vitreous is approximately 99% water,
we can assume the oxygen diusion coecient to not dier signicantly between the mid
and posterior vitreous. Coecients C
1
and C
2
were obtained for the mid and posterior
vitreous based on the best t curve (OriginLab). This was done by minimizing the Chi-
squared value, which is the square root of the sum of the squares of the distance of each
data point from the theoretical curve (Table 3.2).
Figure 3.7: Spatial variation in vitreous oxygen consumption. Experimental data with
tted curve.
51
Position in
Eye
C
1
C
2
Adjusted R
2
Mid-Vitreous 0.00148 8:43 10
6
0.703
Posterior
Vitreous
0.00175 -42 0.86
Table 3.2: Coecients and goodness-of-t indicator for curve tted to both mid-vitreous
and posterior vitreous data
3.4 Discussion
Oxygen distribution in the eye is tightly regulated. This is evident by the hypoxic en-
vironment of ocular tissues, such as the lens (3-9 mmHg) and trabecular meshwork (12
mmHg)[Siegfried et al. 2010; Shui et al. 2009; Padh 1991]. Studies have shown that this
regulation is performed by vitreous humor[Shui et al. 2006; Shui et al. 2009]. In healthy
eyes, oxygen from the retinal vasculature in the anterior surface of the retina diuses
into the vitreous humor. There is a gradient in oxygen content from the retina (22
mmHg) to the posterior lens (9 mmHg) which indicates oxygen consumption by the
vitreous[Beebe et al. 2014]. Shui and colleagues experimentally demonstrated that vitre-
ous reacts with molecular oxygen via an ascorbate dependent reaction[Shui et al. 2009].
Ascorbate (Vitamin C) is essential for many processes, such as synthesis of collagen,
maintaining prosthetic metal ions in their reduced forms, and scavenging free radicals
to prevent oxidative damage to tissues[Padh 1991]. Ascorbate is present in the vitreous
humor in relatively high concentrations and plays an important role in preventing ox-
idative damage caused by free radical formation during solar radiation[Rose, Richer, and
Bode 1998]. Shui et al's recent nding suggests that ascorbate's role in maintaining a low
intraocular oxygen environment is equally important. It is interesting to note that hu-
mans and other animal species such as guinea pigs and primates are limited to whatever
52
ascorbate they can obtain from their diet since they do not have the ability to synthesize
ascorbate from glucose[Rose and Bode 1991].
The ascorbate reaction with molecular oxygen has been studied in a variety of elds
[Eison-Perchonok and Downes 1982; Eaton 1991]. Equation 3.3 is the proposed overall
stoichiometric biochemical equation. However, it is important to note that this reaction
only occurs as 2 one electron transfer reactions as described in equations 3.4 and 3.5.
This reaction will only occur in the presence of catalysts such as light, free radicals,
and transition metals; these catalytic reactions are not fully understood[Eaton 1991; Sil-
verblatt, Robinson, and King 1943; Cabelli and Bielski 1983]. Thus the reaction kinetics
have not been fully detailed either. Shui et al initially proposed a constant rate of vit-
reous oxygen consumption. Filas et al later proposed a hyperbolic function (similar to
Michaelis-Menton Kinetics) to describe the reaction kinetics[Filas, Shui, and Beebe 2013].
However, based on our empirical evidence, we believe that the reaction kinetics can be
best described by a second-order reaction rate law.
While we found a statistical dierence in the ascorbate content between the vitreous
core and posterior vitreous, ascorbate alone might not be the sole cause of the spatial
variation in vitreous oxygen consumption. The concentration of the above mentioned
catalysts might also account for the spatial variation in vitreous oxygen consumption. It is
important to note that the spatial variation in vitreous oxygen consumption was shown in
enucleated porcine eyes, used within 6 hours of enucleation. Despite a short postmortem
interval, retinal degeneration might aect the posterior vitreous oxygen consumption.
More work needs to be done in vivo to conrm the spatial variation in vitreous oxygen
consumption.
53
It is well established that ascorbate accumulates in the eye of humans and animals
at a concentration that is several times higher than that present in the blood plasma.
Prior studies have shown that ascorbate contents in the pig's vitreous, aqueous and blood
plasma are 0.28 mM, 0.57 mM, and 0.01 mM respectively[Rose and Bode 1991; Johnson
1936; DiMattio 1989]. This is especially important because ascorbate's high concentration
makes it the dominant antioxidant over Superoxide Dismutase[Behndig et al. 1998]. We
know that ascorbate does not enter the vitreous by diusion alone and its intravitreal
concentration is maintained by sodium dependent ascorbate transporter (SLC23A2). This
transporter is present in the pigmented layer of the ciliary epithelium[Tsukaguchi et
al. 1999]. Locci et al experimentally demonstrated that ascorbate is found in higher
concentrations near the basal area of the vitreous, the part of the vitreous closest to
the ciliary body[Locci et al. 2014]. Locci's study, together with our ndings of higher
ascorbate content in the posterior vitreous suggests that ascorbate concentration in the
eye exists along a gradient. We hypothesize that ascorbate might enter the eye via the
retinal pigment epithelium[Salceda and Contreras-Cubas 2007]. We also hypothesize that
the posterior vitreous in close proximity to the vitreous cortex, with its denser network
of collagen bers, holds more ascorbate as compared to the vitreous core. Ascorbate's
molecular mass is more than 10x that of oxygen. This spatial variation in ascorbate might
account for the spatial variation on vitreous oxygen consumption.
Our ndings of a spatial variation in vitreous oxygen consumption might also give us
new insight into the results of recent intravitreal oxygen studies. Quiram et al showed
that intravitreal oxygen tension in animals with posterior vitreous detachment (PVD)
along with vitreous liquefaction is signicantly higher when compared to oxygen levels
54
after vitreous liquefaction without PVD. However, oxygen levels in animals with vitreous
liquefaction without PVD are not dierent from oxygen levels in animals with neither
vitreous liquefaction nor PVD[Quiram et al. 2007]. Our understanding that the posterior
vitreous consumes oxygen at a higher rate adds to Quiram's ndings. We believe that the
vitreous cortex does not slow down the process of diusion by having a slower diusion
constant, but rather, consumes oxygen at a greater rate than the rest of the vitreous.
Conclusion
In this study, we investigated the spatial variation of vitreous oxygen consumption in
enucleated porcine eyes. We fabricated a custom oxygen source, implanted it in the
mid vitreous and the posterior vitreous and found that the oxygen concentration proles
were statistically dierent between these two locations; suggesting a spatial variation in
vitreous oxygen consumption. In conjunction with that nding, we observed statistically
dierent concentrations of ascorbate across the two locations and quantied the reaction
rate between ascorbate and oxygen in vitreous.
55
Appendix
Point source diusion equation. Assuming steady state, we get 3.12 from 3.11.
dC
dt
=
D
r
2
d
dr
(r
2
dC
dr
) +R (3.11)
D
r
2
d
dr
(r
2
dC
dr
) =R (3.12)
Reaction equation between ascorbate and oxygen
AH

+O
2
+H
+
!A +H
2
O
2
(3.13)
R =
dO
2
dt
=k[AH

][O
2
] (3.14)
d[AH

]
dt
= 0 (3.15)
ln[O
2
] = ln[O
2
]
0
k

t (3.16)
where;k

=k(AH

) (3.17)
Plug 3.14 into 3.12
D
r
2
d
dr
(r
2
dC
dr
) =k(AH

)C (3.18)
D(2
dC
dr
+r
d
2
C
dr
2
)
r
=k(AH

)C (3.19)
56
Let
C =
v(r)
r
(3.20)
Thus,
dC
dr
=
dv
dr
r

v
r
2
(3.21)
d
2
C
dr
2
=
2v
r
3
+
d
2
C
dr
2
r

2
dv
dr
r
2
(3.22)
Plug into 3.20,3.21, and 3.22, into 3.19
d
d
2
v
dr
2
=k(AH

)v (3.23)
Let
v =e
r
(3.24)
Substitute 3.14 into 3.13
((k(AH

)) +D
2
)e
r
= 0 (3.25)
=
p
k(AH

)
p
D
(3.26)
57
General solution is the sum of the above solutions
v =v
1
+v
2
(3.27)
Where
v
1
=c
1
e
p
k(AH

)
p
D
D
(3.28)
v
2
=c
2
e
p
k(AH

)
p
D
D
(3.29)
and
v =c
1
e
p
k(AH

)
p
D
D
+c
2
e
p
k(AH

)
p
D
D
(3.30)
Substitute back for C
rC =c
1
e
p
k(AH

)
p
D
D
+c
2
e
p
k(AH

)
p
D
D
(3.31)
C(r) =
c
1
e
r
q
AH

k
D
r
+
c
2
e
r
q
AH

k
D
r
q
AH

k
D
(3.32)
58
Chapter 4
Intravitreal Oxygen Level Changes after Laser
Photothrombosis in Rabbits
4.1 Goals
For the oxygen therapy to reach the patient's bedside, we need to answer the question,
what is the safe and eective therapeutic window for retinal oxygenation for the ischemic
retina? Delineating this therapeutic window requires us to create an animal model of
retinal ischemia to test our oxygen therapy on. For this purpose, we chose a rabbit model
with retinal ischemia induced by photothrombosis. To appreciate the changes induced
by oxygen therapy on this animal disease model, we must rst understand the retinal
changes and intravitreal oxygen level changes that occur over time in the animal disease
model without oxygen therapy. However, only the chronic retinal vascular characteristics,
along with retinal histology and immunohistochemistry, have been well dened[Ameri et
al. 2008b]. What are the intravitreal oxygen levels in this animal disease model and how
do they change over time?
59
In this chapter, we aim to investigate the intravitreal oxygen levels in a rabbit model
of retinal ischemia induced by laser photothrombolysis over the course of 14 days.
4.2 Methods
Pigmented rabbits, each weighing 4-6 lbs, were used in this study. All animal experi-
ments adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision
Research and were approved by the Institutional Animal Care and Use Committee of
the University of Southern California. For all animal procedures, the rabbits were anes-
thetized with subcutaneous injection of a mixture of ketamine hydrochloride (25 mg/kg)
and xylazine hydrochloride (6 mg/kg) that was repeated as necessary. The pupils were
dilated with a topical application of phenylephrine hydrochloride 2.5% and tropicamide
0.5% eye drops. Only the right eyes were used. Gentamicin eye ointment was applied
after each procedure to reduce the risk of uveitis. Animals were euthanized via ear vein
injection of sodium pentobarbital (120 mg/kg).
We used a retinal venous occlusion model induced by laser photothrombosis that
has been well described by our group [13,14]. In all animals, RVO was created using
an Argon green laser (Iris Medical Oculight Glx; IRIDEX Corporation, Mountain View,
CA) mounted on an indirect ophthalmoscope. Rose Bengal solution (20mg/kg, Sigma-
Aldrich Inc., St. Louis, MO) was prepared by diluting 30mg/ml with Ringer's solution
and ltering it twice using 45μm and 22μm syringe lters (VWR). A few seconds after
intravenous injection of Rose Bengal, Argon laser was applied to the retinal veins inside
the optic disc, or at the disc margin. All major veins were treated with a laser. In
60
cases where the retinal artery and vein were in very close proximity it is very likely that
both the vessels were treated although only the vein was purposely targeted. To ensure
a long-term occlusion, the laser was scanned over a 2-3 mm segment of the vein at the
disc margin. The laser spot size was 125μm, and the duration of each ablation was
approximately 0.5 s and10 laser shots were applied to each vessel segment with a laser
power of 150 mW.
Using a RetCam (Clarity Medical Systems Inc), at least four color photographs were
taken from each eye: the temporal medullary wing, the nasal medullary wing, the superior
retina above the disc, and the inferior retina below the disc. The RetCam imaging
system was also used to perform Fluorescein Angiography (FA). An intravenous line was
established on the marginal ear vein and 0.2 ml of 10%
uorescein (Akorn, Abit Springs,
LA) was injected and
ushed with 1 ml of normal saline. For indocyanine green (ICG), 1.5
ml solution containing 3.75 mg ICG (Akorn, Abit Springs, LA) was injected. Sequential
images were taken immediately after
uorescein injection. A
uorescein angiogram was
used to verify retinal vein occlusion following the laser treatment and at each follow-up
time point. From each FA video, 3 pieces of data were extracted; 1) Presence of blood
perfusion at proximal and distal sections of the medullary wings, marked P5-P8 in Fig
4.1, 2) Time taken for blood to traverse one optic disc diameter away from the rim of the
optic disc, and 3) Transit time, denoted as the time taken for the
uorescein to enter the
optic disc through an artery and exit to the optic disc through a retinal vein (Fig 4.1).
For partial pressure of oxygen (pO
2
) measurements, we used a ber optic probe con-
nected to a commercially available oxygen measuring system (OxyLab; Oxford Optronix
61
Figure 4.1: Fluorescein Angiography data analysis. The time taken for the
uorescein to
travel 1 optic disc diameter away from the optic disc (location indicated by the yellow
bars and white arrow) was recorded. Retinal vasculature perfusions at positions P5-8
were noted as a binary value. Transit time, denoted as the time for the
uorescein to
enter the retinal vasculature to the time where it rst exits was also recoded.
Ltd., Oxford, UK). The probe has an outer diameter of 250μm and contains a lumines-
cent dye at the tip for which the signal half-life is inversely proportional to the pO
2
. The
oxygen sensor excites the dye with a light pulse, measures the decaying signal's half-life
and converts it to its corresponding pO
2
. The oxygen probe has a resolution of 0.1 mmHg
and an accuracy of10% when pO
2
is below 150 mmHg and20% when the pO
2
is above
150 mmHg.
The probe was introduced into the vitreous cavity of anesthetized rabbits through a
23 gauge valved trocar (Alcon) 2.5 mm from the limbus in the superior nasal quadrant
of the globe. We used a micromanipulator (Edmund Optics, Linear translational stage)
to position the probe in 7 dierent positions in the vitreous (Fig 4.2). All measurements
62
were performed in light-adapted conditions. The microscope light was turned o immedi-
ately before preretinal pO2 measurements because this probe (OxyLite; Oxford Optronix
Ltd.) is very sensitive to bright light. Each measurement was averaged over 90 seconds.
Oxygen measurements were performed before laser photocoagulation (control and Day
0
-
), immediately post laser photocoagulation (Day 0
+
), 7 days, and 14 days after laser
photocoagulation. Powerlab was used to collect the analog signals from the OxyLab in-
strument and LabChart was used to sample and analyze the data. OriginLab was used
to create plots as well as carry out the statistical analysis of the data.
Figure 4.2: Oxygen tension was recorded from the following 7 locations in the vitreous,
P1 through P8.
4.3 Results
Fig 4.3 shows the position of the oxygen probe with respect to the retina. It also shows
the probe in position after the blood vessels were occluded.
Fig 4.4 is a representative plot which shows the decay of pO
2
immediately after
laser photothrombosis. The oxygen probe was positioned to record the pre-retinal pO
2
immediately over the retinal vein. The baseline pO
2
was 29 mmHg. The blood vessels
63
Figure 4.3: (Left) Color fundus image of rabbit retina with oxygen probe at position P7.
(Right) Fluorescein Angiography image after occlusion of nasal blood vessels with oxygen
probe placed over the occluded vessel.
were occluded, as indicated by the arrow in Fig 4.4, and the pO
2
fell to a steady state
level of 3 mmHg over a period of 400 s, or 6.67 mins.
Figure 4.4: Decline of pO2 after laser photothrombosis of retinal blood vessels. Measure-
ment was made at position P7 illustrated in Fig 4.2
64
Pre and post occlusion oxygen tension measurements
The pO
2
before and after laser photothrombosis are detailed in Fig 4.5 and 4.6 (N=6).
Immediately after laser photothrombosis, pO
2
decreased across all intravitreal positions.
These changes are statistically signicant in the following positions P3, P5, P6, P7, and
P8 (p <0.01).
Figure 4.5: Average intravitreal pO
2
(in mmHg) in various positions before and after
laser photothrombosis.
Oxygen tension measurements across 14 days
The pO
2
after laser photothrombosis at time points Day 0
+
(<60mins post occlusion),
Day 7 and Day 14, are detailed in Fig 4.7, 4.8, and 4.9 (N=5). Intravitreal pO
2
decreased
after laser photothrombosis and remained low on Day 7. On Day 14, the intravitreal pO
2
was signicantly elevated.
Seven days after laser photothrombosis intravitreal pO
2
was signicantly lower at all
positions except P4 (p<0.01) when compared to control eyes. Fourteen days after laser
65
Figure 4.6: Bar plot of intravitreal oxygen tension (P1-P8) before and after occlusion
photothrombosis intravitreal pO
2
was only signicantly lower at position P3 (p <0.01)
when compared to control eyes.
Compared to Day 0
+
none of the pO
2
measurements are signicantly dierent (p<0.01)
on Day 7. Compared to Day 0
+
, pO
2
measurements at positions P5, P6 and P7 are sig-
nicantly elevated on Day 14 and are more consistent with baseline pO
2
measurements.
Figure 4.7: Average intravitreal pO
2
(in mmHg) in various positions at time Day O
+
,
Day 7 and Day 14.
66
Figure 4.8: Bar plot of intravitreal pO
2
(P1-P8) at times Day 0
+
, Day 7, and Day 14.
Fundus imaging and
uorescein angiography
Fundus images and
uorescein angiography (FA) videos were taken before laser photo-
coagulation, immediately after photocoagulation, on Day 7 and on Day 14. The FA
taken within 1 hour post occlusion showed no blood
ow in the retina. In some cases, a
segmental lling defect of the choroid was noted on FA as well extending 2-3 disc diam-
eters (superiorly) from the area of laser photothrombosis. On Day 7, the fundus images
showed optic nerve swelling and the FA showed no blood
ow in retinal vasculature. This
is demonstrated in the Transit Times and 1 OD Traversal Times described in Table 4.1.
67
Figure 4.9: Preretinal pO
2
over 14 days across all positions (P1-P8).
On Day 14, the 1 OD Traversal Time is longer compared to Controls due to higher
vascular resistance. On Day 14, the Temporal Transit time was lower but this might be
due to shunting.
On Day 14, the FA results showed high variance in retinal blood
ow, ranging from
partial occlusion to full reperfusion. When comparing the preretinal positions on Day 14,
the perfused positions (as observed in the FA video) exhibited signicantly higher oxygen
tensions (p<0.01) than the non-perfused positions' oxygen tension values (Fig 4.10).
4.4 Discussion
In this study we use a modied photothrombotic RVO model in the rabbit retina to
demonstrate changes in pre-retinal pO
2
over 14 days. This is the longest period over
which in vivo measurements of pre-retinal pO
2
have been made in any RVO model.
68
Nasal Nasal Temporal Temporal
1 OD Transit 1OD Transit
Traversal Time (s) Time (s) Traversal Time (s) Time (s)
Control 0.770.3 5.50.8 0.80.3 5.81.0
Day 0
+
NA NA NA NA
Day 7 5.1* NA NA NA
Day 14 2.611.3** NA 2.450.7*** 4.420.8**
Table 4.1: Fluorescein Angiography results. 1 OD Traversal time refers to the time taken
for the dye to travel through 1 optic disc diameter away from the rim of the optic disc.
The Transit Time refers to the time take for the dye to enter the optic disc through an
artery till it re-enters the optic disc through a retinal vein. * indicates data extracted
from N = 1. ** indicates data extracted from N = 2. *** indicates data extracted from
N = 3. In other samples, no measurements were able to be recorded due to lack of blood

ow.
Preretinal pO
2
dropped signicantly at all retinal locations within an hour after laser
photothrombosis and
uorescein angiography conrmed lack of retinal blood
ow during
this same time interval. The rapid formation of blood clots after laser photothrombosis
and the high rate of oxygen consumption by the retina can explain this rapid drop in
preretinal pO
2
[Yu and Cringle 2001]. We also observed a lack of
uorescein lling of some
segments of the choroidal vasculature at the site of laser but measurements of pre-retinal
pO
2
were equally depressed at multiple locations throughout the retina suggesting that
the lack of choroidal blood
ow did not contribute signcantly to pre-retinal pO
2
.
At Day 7, we observed that preretinal pO
2
remain depressed and were not signicantly
dierent than one hour after laser photothrombosis at all retinal locations. The FA
also showed that there was no notable revascularization of the retina at Day 7. This is
contrary to previously published studies where blood
ow through the retinal vasculature
was observed at Day 7, with revascularization increasing with time and stabilizing at Day
21[Ameri et al. 2008b; Royster et al. 1988]. In order to avoid revascularization by Day 7,
we modied the laser photothrombosis by targeting a 2-3 mm segment of the retinal vein
69
Figure 4.10: Perfusion against preretinal oxygen tension on Day 14. Perfused areas of
the retina on Day 14 correspond to higher preretinal oxygen tension recordings. The
dierence is statistically signicant (p < 0.01).
instead of just one spot (see methods). This allowed for a more robust vein occlusion
that was ophthalmoscopically visible and conrmed by FA. From our experience, this
technique ensures that no revascularization occurs by Day 7.
On Day 14, preretinal pO
2
in some rabbits started to rise, resulting in a high vari-
ance in intravitreal pO
2
levels. This corresponded to revascularization of occluded blood
vessels on
uorescein angiography. In almost all cases, the revascularization pattern on

uorescein angiography at day 14 was dierent relative to before laser photothrombosis.
In many cases the presence of collateral vessels was noted. However, no neovascularization
was present at Day 14. The correlation between perfusion and preretinal pO2 on Day 14
(Fig 4.10) suggests that revascularization of the retinal vasculature is the main contribu-
tor to the elevated pre-retinal pO2 values, as opposed to other contributing factors such
as oxygen diusion from the choroid or reduced retinal oxygen consumption.
70
Our results indicate the viability of rabbit laser photothrombosis model of retinal
ischemia for studies investigating intravitreal oxygen therapy. In this animal model, our
results show that there is a minimal intravitreal oxygen supply between Day 0 and Day 7.
Between Day 7 and Day 14, reperfusion of the retinal vasculature re-introduces a major
source of intravitreal oxygen.
Conclusion
In this study, we investigated the intravitreal oxygen levels in a rabbit model of retinal
ischemia induced by laser photothrombosis over the course of 14 days. We observed
that preretinal oxygen levels decreased within 60 minutes of laser photocoagulation, with
intravitreal oxygen levels decreasing at a slower rate compare to pre-retinal oxygen levels.
At Day 7, intravitreal oxygen levels were depressed but by Day 14, revascularization
occurred and intravitreal oxygen levels started to rise. Our results suggest that the
rabbit model of laser photothrombosis can be used to study the intravitreal eects of
vitreal oxygen therapy for the treatment of retinal ischemia.
71
Chapter 5
Oxygen Biosensor
5.1 Goals
In order to accurately provide the therapeutic levels of oxygen, the Oxygenator needs to
be able to monitor the intravitreal oxygen levels and titrate the oxygen dose accordingly.
However, the current method of measuring intravitreal oxygen is through an invasive
probe that is inserted into the eye. There is a need to incorporate a wireless implantable
oxygen sensor within the Oxygenator. This will allow the Oxygenator to close the loop
on oxygen dosing.
In this chapter, we explore an oxygen biosensor that can provide real time measure-
ment of intravitreal oxygen tension non invasively. I hypothesize that an optical sensor
system can measure the oxygen tension in the vitreous when the luminescent, oxygen sen-
sitive dye molecules are placed in the vitreous and the optical probe is positioned on the
corneal surface. I test this hypothesis using non-ratiometric oxygen sensors in enucleated
bovine eyes.
72
5.2 Methods
Freshly harvested bovine cadaver eyes (Sierra Medical) was used in each experiment. Each
eye was used immediately upon shipment (<6 hours after harvest) to reduce the optical
transmission loss caused by clouding of the cornea and opacication of the lens. The
muscle and conjunctival tissue was rst removed. A
ap of sclera was created at the back
of the eye using a scalpel blade and lifted open, exposing the vitreous humor. The sensor
was placed on the vitreous and the output gas tube from the mass
ow controller was
placed next to it so that the gas would be incident on the sensor. The mass
ow controller
(MCQ Instruments) was used to change the oxygen concentration in the oxygen sensor
patch. The mass
ow controller is able to mix oxygen and nitrogen gases and output them
at a rate of 200 ccm between a range of 0% oxygen to 100% oxygen with an accuracy of
1%. A re
ectance probe (Thorlabs), comprised of the emission probe and the excitation
probe, was used to emit excitation light and collect emission light. The emission probe
has 6 optical bers to collect emission light and the excitation probe has one optical ber
to emit excitation light. This type of probe was chosen to maximize the collection of
emission light. The probe was placed in contact with the corneal surface. A spectrometer
(Ocean Optics, NeoFox), connected to the other end of the emission probe, was used to
measure the phosphorescence from the sensor as well as trigger a high power LED, which
was used as an excitation light source. Since the original excitation light source within
the NeoFox unit was a weak LED, a photocurrent amplier circuit was built to provide
stronger emission light intensity at the same wavelength. The high power LED connected
to the amplier circuit was coupled to the excitation probe.
73
Figure 5.1: Overall experimental setup
The sensor uses a Platinum metal porphyrin (Pt(II)) luminescent dye molecules em-
bedded in a 0.08mm thick
exible acrylic patch. It is excited by light at 460nm. The
molecules emit light at a stokes shift of 180nm and with longer decay times. The Neo-
fox system (OceanOptics) uses a proprietary method to calculate the decay constant .
(Figure 5.3).
Once the decay rate () has been calculated, we can calculate the partial pressure of
oxygen through equation 5.2 which was derived from the Stern-Volmer equation 5.1.
F
0
F
=

0

= 1 +K
SV
C (5.1)
pO
2
=A +B

0

(5.2)
74
Figure 5.2: Optical sensor setup
The constants A and B are derived from the calibration procedure performed prior
to measurements. For the calibration, the sensor was exposed to known concentrations
of oxygen (0 and 21%) and the values were recorded. Using these two known val-
ues, the equation was solved for A and B. Since the calibration curve can be aected
by temperature, a constant temperature was maintained between calibration tests and
measurements.
Three dierent sensors and three bovine eyes were used to perform the preliminary
experiments and the results are shown below. The re
ectance probe's and the sensor's
position were unaltered between calibration steps and measurements.
75
Figure 5.3: How the sensor works (Ocean Optics)
5.3 Results
The oxygen sensor's measurements are highly accurate over the range 0 to 21% oxygen
concentration (Fig 5.4). It demonstrates the sensor's high sensitivity and linearity.
5.4 Discussion
While the oxygen sensor system described here displays a high accuracy, it has two
main limitations. Firstly, the temperature must remain constant between calibration
tests and measurements. This is because Platinum metal porphyrin (Pt(II)) luminescent
dye molecules, which make up the sensor, are very sensitive to temperature
uctuations
[\Luminescence-Based Oxygen Sensors"]. Temperature can aect the molecules' emis-
sion intensity and lifetime. However, given the vitreous' constant temperature in vivo,
temperature
uctuations are not a major limitation of this sensor.
Second, in this oxygen sensor system, the accuracy of the sensor's measurements is
highly dependent on the relative position of the sensor to the probe. Thus, the sensor
must be calibrated prior to every measurement. However, this is not feasible as the system
76
Figure 5.4: Results
cannot be re-calibrated after the initial surgery to implant the sensor. This problem
is similar to the problems described in intensity measurement techniques used on non-
ratiometric dyes [Kostov et al. 2000]. The lifetime based measurement system is supposed
to overcome the unreliability of light intensity changes because it uses the dye molecules
intrinsic property (lifetime signal) to measure oxygen tension. While the Ocean Optics
sensor claims to utilize a lifetime based measurement system, my experimental work
suggests the lifetime based measurement methods still result in sensor systems that are
not robust and that cannot handle changes in sensor-probe positions. One possible way
to overcome this would be to move to a ratiometric oxygen sensor. Ratiometric oxygen
77
sensors work by measuring light intensities at two dierent wavelengths, one of which
acts as an internal reference signal[Yoshihara et al. 2012].
If these limitations are overcome, the oxygen sensor has the potential to be integrated
with the Oxygenator device to monitor intravitreal oxygen levels over time. The biocom-
patible and oxygen permeable membrane of the Oxygenator's intraocular unit can act as
a substrate for the oxygen sensitivy dye molecules. A re
ectance probe can measure the
oxygen levels surrounding the sensor non invasively via the cornea.
This experimental data supports the hypothesis that an optical sensor system can
measure the oxygen tension in the vitreous when the luminescent, oxygen sensitive dye
molecules are placed in the vitreous and the optical probe is positioned on the corneal
surface (eectively creating a non-invasive intravitreal oxygen sensor). The next step will
be to make the sensor more robust by overcoming a major limitation of the current non-
ratiometric sensor, its high dependence on probe-sensor positions. A ratiometric sensor
can potentially overcome these limitations due to it's self-referencing ability (a comparison
between ratiometric and non-ratiometric sensors is described in Chapter 2.7).
78
Chapter 6
Harvesting Solar Energy to Power Ocular Implants
6.1 Goal
The implantable Oxygenator will be powered by batteries encased in a hermetically sealed
unit. Due to the physical size constraints of an ocular implant, the battery size and thus
the battery capacity, is limited and the batteries need to be recharged frequently. One
method of doing so is via inductive coupling with time varying magnetic elds. While
this method of wireless power transfer is common among new medical device implants,
the need for constant battery recharging, from the patient's point of view, increases the
complexity of the therapy. There is currently no method of eciently harvesting energy
from environmental sources for ocular implants. Power constraints can be relaxed by
energy scavengers, exploiting environmental sources of energy, such as solar energy[Olivo,
Carrara, and De Micheli 2011].
We propose harvesting solar energy to power ocular implants. In this chapter, we
investigate the characteristics of a subconjunctival and episcleral photovoltaic implant
to harvest solar energy. We build a photovoltaic cell array and connect it to an energy
79
harvesting circuit for recharging batteries. We also propose a packaging technique to
ensure a lengthy lifetime for the implant.
6.2 Methods
Transmittance
The principal instruments used for our transmittance experiments were a computer driven
Newport Tunable Light Source, with a 300W Xenon lamp, and a power meter (Newport
2936) with a thermopile sensor. The system was turned on for a minimum of 30 minutes
prior to each set of experiments to establish thermal equilibrium. Material optical prop-
erties were obtained with the following experimental protocol: (a) A baseline was initially
recorded for the wavelength range of interest (400-1100nm) with the power meter. (b) For
each material sample, the transmittance is calculated as a ratio compared to the baseline.
(c) After the transmittance test, each sample's thickness is measured with calipers, with
care taken not to deform the sample.
Samples of conjunctiva were extracted from the eyes of three enucleated cadaver pig
eyes. The dissection and measurement of the bulbar conjunctiva was performed within 6
hours postmortem with a pair of Wescott scissors. Care was taken to prevent collecting
the capsule of Tenon.
Samples of Polydimethylsiloxane (PDMS) were prepared with Sylgard 184 (Dow Corn-
ing). The two part PDMS was mixed at a ratio of 10:1 ratio by weight, degassed under
vacuum, and baked at 75 C for 60 minutes. The parylene lm was fabricated using a
80
chemical vapor deposition process. Samples of glass used clear borosilicate microscope
slides (VWR 48311-600).
The prepared samples were then placed between two glass slides (VWR 48311-600)
and pressed in the sample assembly for transmittance measurements. The transmittance
spectrum through the glass slides is the baseline.
Circuit conguration and components
Figure 6.1 shows the schematic diagram of the photovoltaic (PV) system. The PV cells
convert light energy to electrical energy. The energy harvesting integrated chip (IC)
boosts the voltage to an appropriate energy required for charging a rechargeable battery.
Figure 6.1: Schematic diagram of diode array, energy harvesting integrated circuit chip,
and relevant electrical components required to convert solar energy to electrical energy
to recharge a battery.
The PV cell array consists of eight Si diodes (OSIoptoelectronics) connected in a
combination of series parallel to 1) provide enough voltage to ensure high eciency of
81
the energy harvesting IC and 2) provide sucient current in the event that some of the
diodes are blocked from light. This diode has peak responsivity (0.62A/W) at 940nm, at
which there is a higher transmission of light through biological tissue. It is also made up
of Silicon and glass which are common materials found in implants. Its dimensions are
1.9x4.1mm with an active area of 1.7x2.8mm.
An energy harvesting IC (Texas Instruments BQ25504), which acts as a boost con-
verter with battery management capabilities and maximum power point tracking, was
chosen for this application. It comes in a small footprint 3x3mm VQFN package, not
including the external circuit components. It was also chosen for its high eciency over a
wide range of currents. The details of the circuit are shown in the schematic Figure 6.1.
Experimental conguration
The PV cells were connected with 0.04mm diameter gold wire and conductive epoxy
before being encapsulated in 1mm thick PDMS sheet.
We used a 300W Xenon lamp to simulate the solar radiation. Xenon lamp has a light
spectrum that has a similar prole as the sun's light spectrum. A power meter was used
to measure three levels of power intensity (1, 0.67, and 0.33 mW mm
-2
) and the PV array
was placed at those locations during experimentation.
The circuit in Figure 6.1 was modied by replacing the rechargeable battery with a
potentiometer and a 330μF capacitor in parallel. The resistance of the potentiometer
was varied to obtain an output I-V curve.
82
Surgical implantation
The PV array was implanted in freshly harvested (<6 hours postmortem) cadaver pig
eyes. Pig eyes were chosen because they are similar in size to human eyes. The bulbar
conjunctiva was dissected with Wescott scissors and the PV array was placed above the
capsules of Tenon and below the conjunctiva. 6-0 silk sutures were used to suture the PV
array to the sclera and the conjunctiva over the array. This implant is shown in Figure
6.2. In experiments with the PV array implanted, light intensity is determined by the
light irradiated on the cornea. The light incident on the cornea and the implanted PV
array is assumed to be the same. New PV arrays were used for each implantation. Each
array was tested prior to implantation to ensure similar open circuit voltages and the
closed circuit currents.
Figure 6.2: Photograph of an enucleated porcine eye with implanted photovoltaic array.
The array is implanted anteriorly under the conjunctiva and above the sclera.
6.3 Results
Optical Transmittance
The optical transmittance data is shown in Figure 6.3. All material have a sample size of
3 (n=3). The thicknesses of the various materials are: conjunctiva (0.155mm 0.05mm),
83
silicone (1.03mm 0.07mm), Parylene (5μm) and glass (1.01mm 0.03mm). Across
500-1100nm, the optical transmittance of glass, Parylene and silicone is around 0.9. For
conjunctiva, the optical transmittance increases from 0.19 to 0.8 as wavelength increases
from 400nm to 1100nm. At the wavelength where the PV cells' are most responsive
(940nm), the transmittance of conjunctiva, PDMS, Parylene, and glass are 0.72, 0.92,
0.92, and 0.92 respectively.
Figure 6.3: Optical transmittance of relevant biological tissue (conjunctiva) as well as
material used in the hermetic packaging.Dotted lines indicate the standard error
Characteristic IV curve
The I-V and power curves obtained for the photovoltaic array before cadaver eye implan-
tation are shown in Figure 6.4. The irradiation intensities were 1, 0.67 and 0.33 mW
mm
-2
. The peak power was obtained at 0.79V, 3.97mA and 0.77V, 3.08mA, and 0.83V,
84
1.52mA for irradiation intensities of 1, 0.67 and 0.33 mW mm
-2
respectively. Accord-
ing to I-V curve, at peak power, conversion eciencies of 8.26%, 9.32%, and 10.1% for
irradiation intensities of 1, 0.67 and 0.33 mW mm
-2
can be obtained.
Figure 6.4: Characteristic I-V and corresponding power curves for photovoltaic array at
1, 0.67 and 0.33 mW mm
-2
. The solid lines represent the current curves and the dashed
lines represent the power curves.
Output power
The results of the benchtop and cadaver eye tests are shown in Table 6.1. The table
shows the current and voltage at peak power for 2 scenarios: 1) the output from the
energy harvesting circuit before implantation, and 2) the output from the energy har-
vesting circuit after implantation in a pig cadaver eye. Before implantation the power
conversion eciencies for irradiation 1, 0.67 and 0.33 mW mm
-2
are 3.73%, 3.91%, and
85
3.90% respectively. After implantation, the power conversion eciencies for irradiation
1, 0.67 and 0.33 mW mm
-2
are 3.24%, 3.59%, and 3.59% respectively
Light In-
tensity
(mWmm
-2
)
Before Implantation in Eye After Implantation in Eye
Circuit Output
Eciency
(%)
Circuit Output
Eciency
(%)
Voltage
(V)
Current
(mA)
Voltage
(V)
Current
(mA)
1 3.22 0.44 3.73 3.22 0.38 3.24
0.67 3.22 0.31 3.91 3.22 0.28 3.59
0.33 3.21 0.15 3.90 3.21 0.14 3.59
Table 6.1: Summary of Solar Energy Harvester Characteristics
6.4 Discussion
Transmittance
Our experimental data on the transmittance of porcine conjunctiva is also similar to the
transmittance of rabbit conjunctiva of similar thickness[Nemati, Rylander Iii, and Welch
1996]. Nemati et al showed that for rabbit conjunctiva, the scattering coecients are
larger than the absorption coecient, implying that the primary mechanism of attenua-
tion in the conjunctiva, for 400-1100nm range, is due to scattering[Nemati, Rylander Iii,
and Welch 1996].
Performance of the Electronics
The voltage at which the most power is extracted from the photovoltaic array ranges
between 79% and 85% of the open circuit voltage across the various light intensities. We
also noted that after implantation in the eye, the conjunctiva was stretched, increasing
its transmittance.
86
Based on the data thus far, we can obtain the theoretical power eciency of the entire
solar cell harvesting system. Given the transmittance of the conjunctiva to be 70%, the
power conversion eciency of the PV array to be 9% and the eciency of the energy
harvesting circuit to be 60%, the overall theoretical power eciency should be 3.78%.
The experimental data shows average eciency to be 3.47%.
Proposed Packaging Scheme
Hermetic packaging is required to protect the electronic circuity of the implant from
the harsh environment of the human body. The package should be resistant to water
or ion ingress and should protect the electronics for at least 10 years. Our proposed
packaging will involve 3 main parts: a transparent package encapsulating the PV cells,
and a discrete capsule housing the rest of the electronics, and feedthroughs, which must
penetrate both transparent package and discrete capsule. The packaging schemes behind
discrete capsules have been described previously[Jiang and Zhou 2010]. The feedthroughs
are required so that the photovoltaic cells can be electrically connected to the rest of the
circuit. These connections are the most vulnerable leakage points of the system.
We propose a transparent packaging scheme for the PV cells (Figure 6.5). The goal
is to minimize water ingress while maximizing optical transmission and minimizing the
height of the implant. In the proposed scheme, the PV cell (400μm) is placed on a
titanium foil (50μm). A glass sheet (100μm) is placed on top of the PV cell and is
bonded to the titanium foil with epoxy. While glass-to-metal seal technology is highly
eective, with a typical hermeticity of 1*10
-8
atm cm
3
s
-1
, the high temperatures used
in the process cause damage to the photovoltaic cells[Jiang and Zhou 2010]. Thus, we
87
propose using epoxy to bond the glass to the titanium foil. A coating of 100μm silicone is
applied to reduce mechanical stress on the glass and PV cells. Finally, a conformal coating
of Parylene (20μm) is applied. The total thickness of this package will be 670μm. The
nature of low water vapor transmission rates (WVTR) of glass and titanium compared to
normal polymers, make this design very suitable for our application. Polymers' WVTR
can be several orders of magnitude higher than those of metals and glass. Previous
studies using similar designs have observed a change in failure mode from fast water
vapor diusion through barrier layers to slow undercut of the glass[Chang, Liu, and Tai
2014]. The feedthroughs can be run through the epoxy gap and coated with Parylene. It
should be noted that both epoxy and Parylene represent probable water vapor leakage
paths.
Figure 6.5: Proposed transparent packaging for photovoltaic array. The photovoltaic
cell (blue) is encapsulated by glass (grey) on top and by a titanium lm (black) below,
which are bound by epoxy (green). A conformal coating of silicone (yellow) and Parylene
(purple) is then applied.
The materials chosen for the packaging scheme are known biocompatible materials and
should prevent adverse biological reactions. While the possibility of conjunctival erosion
and irritation exists, we believe that an implant with<1mm thickness will mitigate these
problems.
88
Comparison with alternative energy harvesters
The solar energy harvester proposed in this paper, has a maximum power density of
49mWcm
3
at peak sunlight irradiance. This is obtained from a power output of the
battery charging circuit of 0.38mA at 3.22V, using 8 diodes, each with dimensions of
4.1mm x 1.9mm x 0.4mm, at irradiance 1mWmm
2
, after implantation in the cadaver
eye.
Alternative human body energy sources that can be scavenged by MEMS based har-
vesters include kinetic energy, thermal energy, and biofuel harvesters. As far as the
authors know, these technologies have not been implemented for ocular energy harvest-
ing applications. While piezoelectric harvesters have high power densities, going to 4.5
mWcm
-3
, they require signicant forces and are most eective in the knee. Electromag-
netic and electrostatic harvesters only need movement but have low power densities in the
range of a fewWcm
3
[Sue and Tsai 2012]. Thermoelectric harvesters' power eciencies
are greatly dependent on temperature gradients within human tissue. Power densities of
up around 300Wcm
3
have been reported in the presence of a 5 Kelvin temperature dif-
ference[Hannan et al. 2014]. Since the conjunctiva is relatively thin (145μm) and exposed
to ambient air, there is potential for thermoelectric harvesters to power ocular implants.
Biofuel harvesters convert glucose to electrical energy but these currently have low power
densities (2Wcm
2
)[Stetten et al. 2006]. While the power density of the solar energy
harvester is highly favorable compared to alternative energy scavengers, the solar energy
harvester is limited to functioning only when exposed to sunlight. It should be noted
89
that the solar energy harvester's power density is calculated at he output of the battery
charging circuit, and not at the photodiode array.
Conclusion
This paper suggests the feasibility of a solar energy harvesting mechanism to supplement
the power requirements of ocular drug delivery devices and biosensors. Future work must
be done to validate safety through temperature tests, optimize the system to maximize ef-
ciency, test the packaging technique, and test the system under ambient indoor/outdoor
lighting.
90
Chapter 7
Conclusion
Diabetic retinopathy (DR) and Branched/Central Retina Vein Occlusion (BRVO/CRVO)
are leading causes of blindness worldwide. Nearly all patients with type-I diabetes will
develop some manifestation of DR. In type 2 diabetic patients, 80% of insulin-dependent
patients and 50% of patients not requiring insulin therapy will have DR within 20 to
25 years following systemic disease onset. With the incidence of diabetes throughout
the world projected to rise from 150 million to approximately 300 million by the year
2025, DR alone represents a major threat to our global population and will likely present
ever-increasing burden on the health care delivery system.
The most popular of the currently available options to treat these retinal ischemic
diseases is intravitreal anti-VEGF (vascular endothelial growth factor) medication. How-
ever, this is expensive and involve monthly injections of drug into the eye. Patient
compliance with this therapy can be troublesome given that patients have to take time
o work to accommodate a monthly or a 6-week treatment plan. Not only that, if the
current trend of treating these conditions via intravitreal injection continues to increase,
the treatments will be burdensome for the health care system. In 2010 alone, Medicare
91
expenditure on intravitreal injections was$200M, with the procedure volume growing at a
rate of 10-20%[IVT Injections: Health Policy Implications]. On top of all this, anti-VEGF
medication does not treat the root cause of retinal ischemia. By preventing the release
of VEGF, this medication prevents neovascularization, without treating the underlying
cause of the VEGF release.
Localized, supplemental oxygen therapy for the retina has the potential to go much
further than all the current existing therapies to improve these patients' debilitating
condition. First, it targets retinal hypoxia, which is secondary to only ischemia, and which
is a precursor to cellular energy failure as well as the release of angiogenic growth factors.
Next, delivering oxygen therapy via the Oxygenator can improve patient compliance by
reducing the need for frequent clinic visits. In my research endeavors I have sought
to further this therapy in an eort to bring this technology to the bedside. My work,
alongside others, builds a foundation for oxygen therapy upon which the Oxygenator
technology can be developed.
Dening the therapeutic window of retinal oxygenation is a critical goal that must be
achieved for oxygen therapy to become a viable treatment option. Since localized sup-
plemental delivery of oxygen to the retina is achieved through delivering oxygen through
vitreous humor, I rst sought to fully understand the oxygen mass transport dynamics
in the vitreous humor. Through my research work, I quantied the biochemical process
in which vitreous consumes oxygen. I found that the process can be best modeled as a
second order rate equation and I experimentally determined the coecients. My work
allows us to theoretically predict the rate at which oxygen is consumed based on ascorbate
92
and oxygen concentrations. The mathematical underpinning of vitreous oxygen consump-
tion will allow for more accurate computational models of intraocular oxygen transport
which will in turn allow us to better predict the optimal placement of the Oxygenator
implant. My work also contributed to a more fundamental understanding of the vitreous
as a non-homogeneous medium. With respect to oxygen transport (and other molec-
ular transport), the vitreous has always been assumed to be a homogeneous medium.
However, my work questions this assumption and suggests that the vitreous might have
topological variations that can aect mass transport.
My experiments revealed an improved understanding of intravitreal oxygen mass
transport; however, some questions remain unanswered. First, we do not know what
the eect of convection is on mass transport in the vitreous. We know that liquefaction
of the vitreous changes its viscosity and
uid dynamics. However, we do not know how
this changes the rate of vitreal mass transport. A useful experiment will be to induce liq-
uefaction of the vitreous and study oxygen transport over time while the eye is undergoing
rotational movements. Another open question regarding vitreous oxygen consumption re-
lates to long term eects of this biochemical process. Is the ascorbate oxygen reaction
in the vitreous maintained over time? Does the eye replenish the ascorbate at the same
rate at which it is depleted through its interaction with oxygen? These questions can be
answered by monitoring the ascorbate levels through chronic in vivo experiments with
Oxygenator implants.
As I mentioned before, dening the therapeutic window of retinal oxygenation is a
critical goal. This can only be achieved by performing chronic in vivo studies. So I sought
to establish an animal disease model of retinal ischemia that can be used for studying
93
the eects of intravitreal oxygen therapy. Through my research work, I characterized the
chronic intravitreal oxygen distribution in an rabbit model of retinal ischemia induced by
laser photothrombosis. Our modied laser photothrombosis technique led to intravitreal
oxygen levels decreasing to negligible levels 7 days after the induction of retinal ischemia
and rising to almost pre-ischemia levels within 14 days. Since the Oxygenator delivers
oxygen to the ischemic retina by raising the intravitreal oxygen levels in the ischemic
eye to baseline levels, an animal model with low intraocular oxygen levels is critical to
quantifying the Oxygenator's ecacy in elevating vitreal oxygen levels. As a result of
my work, we now know that optimum duration of testing the Oxygenator's impact on
intravitreal oxygen in this animal disease model is within 7 days.
The animal disease model I established for the Oxygenator animal studies has its
limitations; more work must be done to improve the viability of the animal disease model.
How can we extend the duration of retinal ischemia in this animal disease model so that
the Oxygenator can be tested over several weeks? One possible way would be to re-occlude
the vessels using laser photothrombosis every 10 days but we do not yet know the retinal
or intravitreal oxygen eects of this procedure. The next step will be to determine ways
to extend retinal ischemia while maintaining depressed intravitreal oxygen levels in this
rabbit model of retinal ischemia.
After dening the therapeutic window of retinal oxygenation, it is imperative that the
Oxygenator technology can accurately deliver oxygen levels within this window. In order
to achieve this goal, there is a need monitor the intravitreal oxygen levels in patients
easily and titrate the oxygen dose accordingly. So I sought to create a non-invasive
optical oxygen biosensor that can be coupled to the Oxygenator implant. Through my
94
research work, I have engineered a prototype of the rst implantable intraocular oxygen
biosensor and demonstrated its proof of concept. We now know that we can measure the
luminescence of an oxygen sensitive dye molecule through the cornea, lens and vitreous
humor. In the future, the robustness of the oxygen biosensor must be improved. This can
be done by using a ratiometric oxygen sensitive dye molecule as the sensor. Its inherent
referencing ability can reduce measurement errors. The sensor must also be tested to
ensure that the dye molecules can be safely immobilized in the Oxygenator's material
and that they can accurately measure oxygen levels over time in a live eye. This can be
tested through chronic in vivo experiments.
One of the technological constraints that face medical implants today is the issue
of battery capacity. This is especially true for ocular implants, where physical size is
limited. One potential solution is to harvest energy from the environment to power these
ocular implants. So I sought to create a solar energy harvester to power ocular these
implants. Through my research work, I created a prototype of the rst subconjunctival
photovoltaic implant and demonstrated a proof of concept. As a result of my work, we
now know the quantities of power that an ocular implant can harvest from solar energy.
This technology has the potential to be used for low power ocular sensors and other low
power implants. For this technology to progress, certain key questions must be answered.
We must study the long term hermeticity of the optically transparent packaging. The
safety prole of the photovoltaic implant must also be investigated through temperature
tests.
THE END
95
References
Abdallah, Walid et al. (2011). \Vitreal oxygenation in retinal ischemia reperfusion."
In: Investigative ophthalmology & visual science 52.2, pp. 1035{42. issn: 1552-5783.
doi: 10.1167/iovs.09-4516. url: http://www.pubmedcentral.nih.gov/
articlerender.fcgi?artid=3053094{\&}tool=pmcentrez{\&}rendertype=
abstract.
Ameri, Hossein et al. (2008a). Intravitreal and subretinal injection of tissue plasminogen
activator (tPA) in the treatment of experimentally created retinal vein occlusion in
rabbits. doi: 10.1097/IAE.0b013e318140580c. url: http://www.ncbi.nlm.nih.
gov/pubmed/18301042.
Ameri, Hossein et al. (2008b). \Natural course of experimental retinal vein occlusion in
rabbit; arterial occlusion following venous photothrombosis." In: Graefe's archive for
clinical and experimental ophthalmology = Albrecht von Graefes Archiv f ur klinische
und experimentelle Ophthalmologie 246.10, pp. 1429{39. issn: 0721-832X. doi: 10.
1007/s00417-008-0878-4. url: http://www.ncbi.nlm.nih.gov/pubmed/
18642023.
Anderson, R. Rox and John A. Parrish (1981). \The Optics of Human Skin". In: Jour-
nal of Investigative Dermatology 77.1, pp. 13{19. issn: 0022202X. doi: 10.1111/
1523-1747.ep12479191. url: http://linkinghub.elsevier.com/retrieve/pii/
S0022202X15461251.
Balachandran, Ram K and Victor H Barocas (2008). Computer modeling of drug delivery
to the posterior eye: eect of active transport and loss to choroidal blood
ow. doi:
10.1007/s11095-008-9691-3. url: http://www.ncbi.nlm.nih.gov/pubmed/
18679772.
| (2011). \Contribution of saccadic motion to intravitreal drug transport: theoretical
analysis." In: Pharmaceutical research 28.5, pp. 1049{64. issn: 1573-904X. doi:
10.1007/s11095-010-0356-7. url: http://www.pubmedcentral.nih.gov/
articlerender.fcgi?artid=3074012{\&}tool=pmcentrez{\&}rendertype=
abstract.
Barbazetto, Irene a et al. (2004). \Oxygen tension in the rabbit lens and vitreous before
and after vitrectomy." In: Experimental eye research 78.5, pp. 917{24. issn: 0014-
4835. doi: 10.1016/j.exer.2004.01.003. url: http://www.ncbi.nlm.nih.gov/
pubmed/15051473.
Beebe, David C et al. (2014). \Preserve the (intraocular) environment: the importance
of maintaining normal oxygen gradients in the eye." In: Japanese journal of ophthal-
mology 58.3, pp. 225{31. issn: 1613-2246. doi: 10.1007/s10384-014-0318-4. url:
http://www.ncbi.nlm.nih.gov/pubmed/24687817.
96
Behndig, Anders et al. (1998). \Superoxide dismutase isoenzymes in the human eye." In:
Investigative ophthalmology & visual science 39.3, pp. 471{5. issn: 0146-0404. url:
http://www.iovs.org/content/39/3/471.shorthttp://www.ncbi.nlm.nih.gov/
pubmed/9501855.
Ben-Nun, J et al. (1988). \A new method for oxygen supply to acute ischemic retina."
In: Investigative ophthalmology & visual science 29.2, pp. 298{304. issn: 0146-0404.
url: http://www.ncbi.nlm.nih.gov/pubmed/3338887.
Blair, N P et al. (1991). \Limitation of retinal injury by vitreoperfusion initiated after
onset of ischemia." In: Archives of ophthalmology 109.1, pp. 113{8. issn: 0003-9950.
url: http://www.ncbi.nlm.nih.gov/pubmed/1987927.
Cabelli, Diane E. and Benon H. J. Bielski (1983). \Kinetics and mechanism for the ox-
idation of ascorbic acid/ascorbate by HO2/O2- (hydroperoxyl/superoxide) radicals.
A pulse radiolysis and stopped-
ow photolysis study". In: The Journal of Physical
Chemistry 87.10, pp. 1809{1812. issn: 0022-3654. doi: 10.1021/j100233a031. url:
http://pubs.acs.org/doi/abs/10.1021/j100233a031.
Cadei, Andrea et al. (2014). \Kinetic and thermal energy harvesters for implantable
medical devices and biomedical autonomous sensors". In: Measurement Science and
Technology 25, p. 012003. issn: 0957-0233. doi: 10.1088/0957-0233/25/1/012003.
url: http://stacks.iop.org/0957-0233/25/i=1/a=012003?key=crossref.
4565a4c5f7f5607751b3631c083710d0.
Chang, Jay Han-Chieh, Yang Liu, and Yu-Chong Tai (2014). \Long term glass-
encapsulated packaging for implant electronics". In: 2014 IEEE 27th International
Conference on Micro Electro Mechanical Systems (MEMS), pp. 1127{1130. doi:
10.1109/MEMSYS.2014.6765844. url: http://ieeexplore.ieee.org/lpdocs/
epic03/wrapper.htm?arnumber=6765844.
Ciulla, Thomas A., Armando G. Amador, and Bernard Zinman (2003). Diabetic retinopa-
thy and diabetic macular edema: Pathophysiology, screening, and novel therapies.doi:
10.2337/diacare.26.9.2653.
DeGra, B. A. and J N Demas. \Luminescence-Based Oxygen Sensors". In: Reviews in
Fluorescence 2005. 1. Boston, MA: Springer US, pp. 125{151. doi: 10.1007/0-387-
23690-2{\_}6. url: http://link.springer.com/10.1007/0-387-23690-2{\_}6.
DiMattio, J (1989). \A comparative study of ascorbic acid entry into aqueous and vitreous
humors of the rat and guinea pig." In: Investigative ophthalmology & visual science
30.11, pp. 2320{2331. url: http://www.iovs.org/content/30/11/2320.short.
Eaton, John W. (1991). \Is the lens canned?" In: Free Radical Biology and Medicine
11.2, pp. 207{213. issn: 08915849. doi: 10.1016/0891-5849(91)90173-Z. url:
http://linkinghub.elsevier.com/retrieve/pii/089158499190173Z.
Eison-Perchonok, M. H. and T. W. Downes (1982). \Kinetics of Ascorbic Acid Autoxida-
tion as a Function of Dissolved Oxygen Concentration and Temperature". In: Journal
of Food Science 47.3, pp. 765{767. issn: 0022-1147. doi: 10.1111/j.1365-2621.
1982.tb12710.x. url: http://onlinelibrary.wiley.com/doi/10.1111/j.1365-
2621.1982.tb12710.x/abstracthttp://doi.wiley.com/10.1111/j.1365-
2621.1982.tb12710.x.
97
Engerman, Ronald L (1989). \Pathogenesis of Diabetic Retinopathy". In: Diabetes 38.Oc-
tober, pp. 1203{1206. issn: 0012-1797. doi: 10.2337/diab.38.10.1203. url: http:
//diabetes.diabetesjournals.org/content/38/10/1203.abstract.
Feng, Yan et al. (2012). \Ratiometric optical oxygen sensing: a review in respect of mate-
rial design". In: The Analyst 137, p. 4885.issn: 0003-2654.doi:10.1039/c2an35907c.
Filas, Benjamen a, Ying-Bo Shui, and David C Beebe (2013). \Computational model for
oxygen transport and consumption in human vitreous." In: Investigative ophthalmol-
ogy & visual science 54.10, pp. 6549{59. issn: 1552-5783. doi: 10.1167/iovs.13-
12609. url: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=
3797591{\&}tool=pmcentrez{\&}rendertype=abstract.
Fong, Donald S., Aniz Girach, and April Boney (2007). \VISUAL SIDE EFFECTS
OF SUCCESSFUL SCATTER LASER PHOTOCOAGULATION SURGERY FOR
PROLIFERATIVE DIABETIC RETINOPATHY". In: Retina 27.7, pp. 816{824.
issn: 0275-004X. doi: 10.1097/IAE.0b013e318042d32c. url: http://content.
wkhealth.com/linkback/openurl?sid=WKPTLP:landingpage{\&}an=00006982-
200709000-00003.
Giblin, F J et al. (1995). Nuclear light scattering, disulde formation and membrane
damage in lenses of older guinea pigs treated with hyperbaric oxygen. doi: 10.1016/
S0014-4835(05)80105-8.
Gisladottir, Svanborg, Thorsteinn Loftsson, and Einar Stefansson (2009). \Diusion char-
acteristics of vitreous humour and saline solution follow the Stokes Einstein equation."
In: Graefe's archive for clinical and experimental ophthalmology = Albrecht von Grae-
fes Archiv f ur klinische und experimentelle Ophthalmologie 247.12, pp. 1677{84. issn:
1435-702X. doi: 10.1007/s00417-009-1141-3. url: http://www.ncbi.nlm.nih.
gov/pubmed/19639333.
Giuliari, Gian P (2012). \Diabetic retinopathy: current and new treatment options." In:
Current diabetes reviews 8.1, pp. 32{41. issn: 1875-6417. url: http://www.ncbi.
nlm.nih.gov/pubmed/22352446.
Goto, K and T Nakagawa (2001). \An implantable power supply with an optically
rechargeable lithium battery". In: . . . , IEEE Transactions on 48.7, pp. 830{833.
url: http://ieeexplore.ieee.org/xpls/abs{\_}all.jsp?arnumber=930908.
Hannan, Mahammad a et al. (2014). \Energy harvesting for the implantable biomedical
devices: issues and challenges." In: Biomedical engineering online 13.1, p. 79. issn:
1475-925X. doi: 10.1186/1475-925X-13-79. url: http://www.pubmedcentral.
nih . gov / articlerender . fcgi ? artid = 4075616{\ & }tool = pmcentrez{\ &
}rendertype=abstract.
Hayreh, Sohan Singh, Wichard van Heuven, and Manmohinder Singh Hayreh (1978). \Ex-
perimental Retinal Vascular Occlusion". In: Archives of ophthalmology 96, pp. 311{
323.
Hiramitsu, T et al. (1976). \Formation of lipoperoxide in the retina of rabbit exposed to
high concentration of oxygen". In: Experientia Idd. url: http://link.springer.
com/article/10.1007/BF01990199.
Holekamp, Nancy M, Ying-Bo Shui, and David C Beebe (2005). \Vitrectomy surgery
increases oxygen exposure to the lens: a possible mechanism for nuclear cataract
98
formation." In: American journal of ophthalmology 139.2, pp. 302{10. issn: 0002-
9394. doi: 10.1016/j.ajo.2004.09.046. url: http://www.ncbi.nlm.nih.gov/
pubmed/15733992.
IVT Injections: Health Policy Implications. url: http://www.reviewofophthalmology.
com/content/d/retinal{\_}insider/c/48732/ (visited on 02/05/2015).
Jaime, Gilberto Raul Lopez et al. (2012). \Acute Variations in Retinal Vascular Oxygen
Content in a Rabbit Model of Retinal Venous Occlusion". In: PLoS ONE 7.11. issn:
19326203. doi: 10.1371/journal.pone.0050179.
Jampol, Lee M. (2003). \Adler's Physiology of the Eye: Clinical Application, 10th Edi-
tion," in: Survey of Ophthalmology 48.4, p. 464.issn: 00396257.doi:10.1016/S0039-
6257(03)00069-9. url: http://linkinghub.elsevier.com/retrieve/pii/
S0039625703000699.
Jiang, Guangqiang and David D Zhou (2010). Technology Advances and Challenges in
Hermetic Packaging for Implantable Medical Devices. Ed. by David Zhou and Elias
Greenbaum. Biological and Medical Physics, Biomedical Engineering. New York, NY:
Springer New York, pp. 27{61. isbn: 978-0-387-98119-2. doi: 10.1007/978-0-387-
98120-8{\_}2. url: http://www.springerlink.com/index/10.1007/978-0-387-
98120-8.
Johnson, S W (1936). \Cataract and ascorbic acid in the guinea-pig eye." In: The
Biochemical journal 30.8, pp. 1430{7. issn: 0264-6021. url: http : / / www .
pubmedcentral . nih . gov / articlerender . fcgi ? artid = 1263202{\ & }tool =
pmcentrez{\&}rendertype=abstract.
Kang, Dongyang et al. (2015). \MEMS oxygen transporter to treat retinal ischemia".
In: 2015 28th IEEE International Conference on Micro Electro Mechanical Systems
(MEMS). Figure 1. IEEE, pp. 154{157. isbn: 978-1-4799-7955-4. doi: 10.1109/
MEMSYS.2015.7050909. url: http://ieeexplore.ieee.org/lpdocs/epic03/
wrapper.htm?arnumber=7050909.
Kashani, Amir H et al. (2010). \Retinal thickness analysis by race, gender, and age using
Stratus OCT." In: American journal of ophthalmology 149.3, 496{502.e1. issn: 1879-
1891. doi: 10.1016/j.ajo.2009.09.025. url: http://www.pubmedcentral.nih.
gov/articlerender.fcgi?artid=2826608{\&}tool=pmcentrez{\&}rendertype=
abstract.
Kashani, Amir H. et al. (2011). \Hyperspectral computed tomographic imaging spec-
troscopy of vascular oxygen gradients in the rabbit retina In Vivo". In: PLoS ONE
6.9, pp. 1{11. issn: 19326203. doi: 10.1371/journal.pone.0024482.
Kostov, Yordan et al. (2000). \Unique oxygen analyzer combining a dual emission probe
and a low-cost solid-state ratiometric
uorometer". In: Applied Spectroscopy 54.6,
pp. 864{868. issn: 00037028. doi: 10.1366/0003702001950238.
Lange, Clemens a K et al. (2011). \Intraocular oxygen distribution in advanced prolifera-
tive diabetic retinopathy." In: American journal of ophthalmology 152.3, 406{412.e3.
issn: 1879-1891. doi: 10.1016/j.ajo.2011.02.014. url: http://www.ncbi.nlm.
nih.gov/pubmed/21723532.
Linsenmeier, R a (1986). \Eects of light and darkness on oxygen distribution and con-
sumption in the cat retina." In: The Journal of general physiology 88.4, pp. 521{42.
99
issn: 0022-1295. url: http://www.pubmedcentral.nih.gov/articlerender.fcgi?
artid=2228847{\&}tool=pmcentrez{\&}rendertype=abstract.
Locci, Emanuela et al. (2014). \A metabolomic approach to animal vitreous humor to-
pographical composition: a pilot study." In: PloS one 9.5, e97773. issn: 1932-6203.
doi: 10.1371/journal.pone.0097773. url: http://www.pubmedcentral.nih.
gov/articlerender.fcgi?artid=4028277{\&}tool=pmcentrez{\&}rendertype=
abstract.
Man, Jacky, Kwong Kwong, and Joseph Caprioli (2010). \Animal Models for Retinal
Diseases". In: Neuromethods 46.2. Ed. by Iok-Hou Pang and Abbot F. Clark. doi:
10.1007/978-1-60761-541-5. url: http://www.springerlink.com/index/10.
1007/978-1-60761-541-5.
Minhas, Gillipsie, Ryuichi Morishita, and Akshay Anand (2012). \Preclinical models
to investigate retinal ischemia: advances and drawbacks." In: Frontiers in neurol-
ogy 3.May, p. 75. issn: 1664-2295. doi: 10.3389/fneur.2012.00075. url: http:
//www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3350026{\&}tool=
pmcentrez{\&}rendertype=abstract.
Murali, K, N Scianmarello, and M.S. Humayun (2015). \Harvesting solar energy to power
ocular implants". In: 2015 IEEE Biomedical Circuits and Systems Conference (Bio-
CAS). IEEE, pp. 1{4.isbn: 978-1-4799-7234-0.doi: 10.1109/BioCAS.2015.7348291.
url: http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=
7348291.
Murali, Karthik, Amir H Kashani, and Mark S Humayun (2015). \Whole vitreous humor
dissection for vitreodynamic analysis." In: Journal of visualized experiments : JoVE
99, e52759. issn: 1940-087X. doi: 10.3791/52759. url: http://www.ncbi.nlm.
nih.gov/pubmed/26065393http://www.pubmedcentral.nih.gov/articlerender.
fcgi?artid=PMC4532330.
Murali, Karthik et al. (2016). \Spatial Variations in Vitreous Oxygen Consumption". In:
Plos One 11.3, e0149961. issn: 1932-6203. doi: 10.1371/journal.pone.0149961.
url: http://dx.plos.org/10.1371/journal.pone.0149961.
Nanda, S K et al. (1987). \A new method for vascular occlusion. Photochemical initiation
of thrombosis." In: Archives of ophthalmology 105.8, pp. 1121{4.issn: 0003-9950.url:
http://www.ncbi.nlm.nih.gov/pubmed/3632422.
Nemati, B, H G Rylander Iii, and a J Welch (1996). \Optical properties of conjunc-
tiva, sclera, and the ciliary body and their consequences for transscleral cyclopho-
tocoagulation." In: Applied optics 35.19, pp. 3321{7. issn: 0003-6935. url: http:
//www.ncbi.nlm.nih.gov/pubmed/21102718.
Newman, Eric A (2013). \Functional hyperemia and mechanisms of neurovascular cou-
pling in the retinal vasculature." In: Journal of cerebral blood
ow and metabolism :
ocial journal of the International Society of Cerebral Blood Flow and Metabolism
33.11, pp. 1685{95. issn: 1559-7016. doi: 10.1038/jcbfm.2013.145. url: http:
//dx.doi.org/10.1038/jcbfm.2013.145.
Newport (2016). Optics Technical Notes. url: http://www.newport.com/ (visited on
03/01/2016).
Nguyen, Q. D. (2004). \Supplemental Oxygen Improves Diabetic Macular Edema: A
Pilot Study". In: Investigative Ophthalmology & Visual Science 45.2, pp. 617{624.
100
issn: 0146-0404. doi: 10.1167/iovs.03-0557. url: http://www.iovs.org/cgi/
doi/10.1167/iovs.03-0557.
Olivo, Jacopo, Sandro Carrara, and Giovanni De Micheli (2011). \Energy Harvesting
and Remote Powering for Implantable Biosensors". In: IEEE Sensors Journal 11.7,
pp. 1573{1586. issn: 1530-437X. doi: 10.1109/JSEN.2010.2085042. url: http:
//ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=5597914.
O'Mahoney, Paul R a, David T Wong, and Joel G Ray (2008). \Retinal vein occlusion
and traditional risk factors for atherosclerosis." In: Archives of ophthalmology 126.5,
pp. 692{699. issn: 1538-3601. doi: 10.1001/archopht.126.5.692.
Osborne, Neville N et al. (2004). \Retinal ischemia: mechanisms of damage and potential
therapeutic strategies." In: Progress in retinal and eye research 23.1, pp. 91{147.issn:
1350-9462. doi: 10.1016/j.preteyeres.2003.12.001. url: http://www.ncbi.
nlm.nih.gov/pubmed/14766318.
Padh, Harish (1991). \Vitamin C: newer insights into its biochemical functions". In:
Nutrition reviews. url: http://onlinelibrary.wiley.com/doi/10.1111/j.1753-
4887.1991.tb07407.x/abstract.
Palmquist, B M, B Philipson, and P O Barr (1984). \Nuclear cataract and myopia during
hyperbaric oxygen therapy." In: The British journal of ophthalmology 68.2, pp. 113{
117. issn: 0007-1161. doi: 10.1136/bjo.68.2.113.
Pautler, Scott E (2010). \Diabetic Retinopathy". In: ed. by David J. Browning, pp. 203{
225. doi: 10.1007/978-0-387-85900-2. url: http://www.springerlink.com/
index/10.1007/978-0-387-85900-2.
Penkova, Anita et al. (2014). \A technique for drug surrogate diusion coecient measure-
ment by intravitreal injection". In: International Journal of Heat and Mass Transfer
70, pp. 504{514. issn: 00179310. doi: 10.1016/j.ijheatmasstransfer.2013.11.
002. url: http://dx.doi.org/10.1016/j.ijheatmasstransfer.2013.11.002.
Prisco, Domenico and Rossella Marcucci (2002). \Retinal vein thrombosis: risk factors,
pathogenesis and therapeutic approach." In: Pathophysiology of haemostasis and
thrombosis 32, pp. 308{311.
Quiram, Polly A et al. (2007). \Microplasmin-induced posterior vitreous detachment
aects vitreous oxygen levels." In: Retina (Philadelphia, Pa.) 27.8, pp. 1090{6. issn:
0275-004X. doi: 10.1097/IAE.0b013e3180654229. url: http://www.ncbi.nlm.
nih.gov/pmc/articles/PMC2702988/http://www.pubmedcentral.nih.gov/
articlerender.fcgi?artid=2702988{\&}tool=pmcentrez{\&}rendertype=
abstract.
Raillard, H. \Development of an implantable cardiac pacemaker". In: 1962 IEEE Interna-
tional Solid-State Circuits Conference. Digest of Technical Papers. Vol. V. Institute of
Electrical and Electronics Engineers, pp. 88{89. doi: 10.1109/ISSCC.1962.1157379.
url: http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=
1157379.
Ramamoorthy, R, P K Dutta, and S A Akbar (2003). \Oxygen sensors: Materials, meth-
ods, designs and applications". In: Journal of Materials Science 38, pp. 4271{4282.
issn: 00222461. doi: 10.1023/A:1026370729205.
101
Rayner, K (1998). \Eye movements in reading and information processing: 20 years of
research." In: Psychological bulletin 124.3, pp. 372{422. issn: 0033-2909. doi: 10.
1037/0033-2909.124.3.372. arXiv: 0010225v1 [hep-th].
Repetto, Rodolfo, Alessandro Stocchino, and Chiara Caerata (2005). \Experimental
investigation of vitreous humour motion within a human eye model." In: Physics
in medicine and biology 50.19, pp. 4729{43. issn: 0031-9155. doi: 10.1088/0031-
9155/50/19/021. url: http://www.ncbi.nlm.nih.gov/pubmed/16177501.
Rose, R. C., S. P. Richer, and a. M. Bode (1998). \Ocular Oxidants and Antioxidant
Protection". In: Experimental Biology and Medicine 217.4, pp. 397{407. issn: 1535-
3702. doi: 10.3181/00379727-217-44250. url: http://ebm.sagepub.com/lookup/
doi/10.3181/00379727-217-44250.
Rose, Richard C. and Ann M. Bode (1991). \Ocular ascorbate transport and metabolism".
In: Comparative Biochemistry and Physiology Part A: Physiology 100.2, pp. 273{
285. issn: 03009629. doi: 10 . 1016 / 0300 - 9629(91 ) 90470 - W. url: http :
//www.sciencedirect.com/science/article/pii/030096299190470Whttp:
//linkinghub.elsevier.com/retrieve/pii/030096299190470W.
Rossi, Tommaso et al. (2012). \Ultrasound imaging velocimetry of the human vitreous."
In: Experimental eye research 99, pp. 98{104.issn: 1096-0007.doi: 10.1016/j.exer.
2012.03.014. url: http://www.ncbi.nlm.nih.gov/pubmed/22516112.
Royster, a J et al. (1988). \Photochemical initiation of thrombosis. Fluorescein angio-
graphic, histologic, and ultrastructural alterations in the choroid, retinal pigment
epithelium, and retina." In: Archives of ophthalmology 106.11, pp. 1608{14. issn:
0003-9950. url: http://www.ncbi.nlm.nih.gov/pubmed/3190547.
Salceda, Roc o and Cecilia Contreras-Cubas (2007). \Ascorbate uptake in normal and
diabetic rat retina and retinal pigment epithelium". In: Comparative Biochemistry and
Physiology Part C: Toxicology & Pharmacology 146.1-2, pp. 175{179. issn: 15320456.
doi: 10.1016/j.cbpc.2007.02.015.
Scianmarello, Nicholas et al. (2016). \OXYGEN GENERATION BY ELECTROLYSIS
TO TREAT RETINAL". In: IEEE MEMS January, pp. 399{402. doi: 10.1109/
MEMSYS.2016.7421645.
Seb, J (2014). Vitreous. Ed. by J. Sebag. New York, NY: Springer New York. isbn: 978-
1-4939-1085-4. doi: 10.1007/978-1-4939-1086-1. url: http://link.springer.
com/10.1007/978-1-4939-1086-1.
Sebag, J (2014). Vitreous. Ed. by J. Sebag. New York, NY: Springer New York. isbn: 978-
1-4939-1085-4. doi: 10.1007/978-1-4939-1086-1. url: http://link.springer.
com/10.1007/978-1-4939-1086-1.
Shui, YB Ying-Bo et al. (2009). \The gel state of the vitreous and ascorbate-dependent
oxygen consumption: relationship to the etiology of nuclear cataracts." In: Archives of
ophthalmology 127.4, pp. 475{82. issn: 1538-3601. doi: 10.1001/archophthalmol.
2008.621. url: http://archpsyc.jamanetwork.com/article.aspx?articleid=
422840http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=
2683478{\&}tool=pmcentrez{\&}rendertype=abstract.
Shui, Ying-Bo et al. (2006). \Oxygen distribution in the rabbit eye and oxygen consump-
tion by the lens." In: Investigative ophthalmology & visual science 47.4, pp. 1571{80.
102
issn: 0146-0404. doi: 10.1167/iovs.05-1475. url: http://www.ncbi.nlm.nih.
gov/pubmed/16565394.
Siegfried, Carla J et al. (2010). \Oxygen distribution in the human eye: relevance to the
etiology of open-angle glaucoma after vitrectomy." In: Investigative ophthalmology &
visual science 51.11, pp. 5731{8. issn: 1552-5783. doi: 10.1167/iovs.10-5666. url:
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3061509{\&
}tool=pmcentrez{\&}rendertype=abstract.
Siggers, Jennifer H. and C. Ross Ethier (2012). \Fluid Mechanics of the Eye". In: An-
nual Review of Fluid Mechanics 44.1, pp. 347{372. issn: 0066-4189. doi: 10.1146/
annurev-fluid-120710-101058. url: http://www.annualreviews.org/doi/abs/
10.1146/annurev-fluid-120710-101058.
Silay, K. M. et al. (2009). \Load optimization of an inductive power link for remote
powering of biomedical implants". In: Proceedings - IEEE International Symposium
on Circuits and Systems, pp. 533{536. issn: 02714310. doi: 10.1109/ISCAS.2009.
5117803.
Silverblatt, E, AL Robinson, and CG King (1943). \The kinetics of the reaction between
ascorbic acid and oxygen in the presence of copper ion". In: Journal of the American
. . . 455.4. url: http://pubs.acs.org/doi/abs/10.1021/ja01242a002.
Simpson, Andrew R H et al. (2013). \Measuring the eect of pars plana vitrectomy on vit-
reous oxygenation using magnetic resonance imaging." In: Investigative ophthalmology
& visual science 54, pp. 2028{34. issn: 1552-5783. doi: 10.1167/iovs.12-11258.
url: http://www.ncbi.nlm.nih.gov/pubmed/23439600.
Skeie, Jessica M and Vinit B Mahajan (2011). \Dissection of human vitreous body ele-
ments for proteomic analysis." In: Journal of visualized experiments : JoVE 47, e2455.
issn: 1940-087X.doi:10.3791/2455.url:http://www.jove.com/video/2455http:
//www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3182653{\&}tool=
pmcentrez{\&}rendertype=abstract.
Smith, Corey a et al. (2012). \A framework for modeling ocular drug transport and

ow through the eye using micro-CT." In: Physics in medicine and biology 57.19,
pp. 6295{307. issn: 1361-6560. doi: 10.1088/0031-9155/57/19/6295. url: http:
//www.ncbi.nlm.nih.gov/pubmed/22982646.
Smith, S. (2009). \An implantable drug delivery system with wireless power and commu-
nication". In: Bionic Health: Next Generation Implants, Prosthetics and Devices, 2009
IET October, pp. 1{88. doi: 10.1049/ic.2009.0194. url: http://ieeexplore.
ieee.org/xpl/freeabs{\_}all.jsp?arnumber=5348717.
Stef ansson, E, R L Novack, and D L Hatchell (1990). \Vitrectomy prevents retinal hypoxia
in branch retinal vein occlusion." In: Investigative ophthalmology & visual science 31.2,
pp. 284{9. issn: 0146-0404. url: http://www.ncbi.nlm.nih.gov/pubmed/2303330.
Stetten, F.v. Stetten F.v. et al. (2006). \A One-Compartment, Direct Glucose Fuel Cell
for Powering Long-Term Medical Implants". In: 19th IEEE International Conference
on Micro Electro Mechanical Systems January, pp. 934{937. issn: 1084-6999. doi:
10.1109/MEMSYS.2006.1627954.
Sue, Chung Yang and Nan Chyuan Tsai (2012). \Human powered MEMS-based energy
harvest devices". In: Applied Energy 93, pp. 390{403. issn: 03062619. doi: 10.1016/
103
j.apenergy.2011.12.037. url: http://dx.doi.org/10.1016/j.apenergy.2011.
12.037.
Szmacinski, Henryk and Joseph R Lakowicz (1994). \Lifetime-Based Sensing". In: Topics
in Fluorescence Spectroscopy. Vol. 4. Boston: Kluwer Academic Publishers, pp. 295{
334. doi: 10.1007/0-306-47060-8{\_}10. url: http://link.springer.com/10.
1007/0-306-47060-8{\_}10.
Tameesh, Mohamed K et al. (2004). Retinal vein cannulation with prolonged infusion
of tissue plasminogen activator (t-PA) for the treatment of experimental retinal vein
occlusion in dogs. doi: 10.1016/j.ajo.2004.06.083. url: http://www.ncbi.nlm.
nih.gov/pubmed/15531319.
Theilmann, PT (2012). \Wireless power transfer for scaled electronic biomedical im-
plants". PhD thesis. url: https://escholarship.org/uc/item/07d5v53v.pdf.
Tsukaguchi, H et al. (1999). \A family of mammalian Na+-dependent L-ascorbic acid
transporters." In: Nature 399.6731, pp. 70{5. issn: 0028-0836. doi: 10.1038/19986.
url: http://www.ncbi.nlm.nih.gov/pubmed/10331392.
Wang, Guoxing et al. (2006). \A dual band wireless power and data telemetry for retinal
prosthesis." In: Conference proceedings : ... Annual International Conference of the
IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine
and Biology Society. Conference 1, pp. 4392{5.issn: 1557-170X.doi:10.1109/IEMBS.
2006.260470. url: http://www.ncbi.nlm.nih.gov/pubmed/17946243.
Wang, Xu Dong et al. (2010). \Optical oxygen sensors move towards colorimetric determi-
nation". In: TrAC - Trends in Analytical Chemistry 29, pp. 319{338. issn: 01659936.
doi: 10.1016/j.trac.2010.01.004.
Wangsa-Wirawan, Norbert D and Robert A Linsenmeier (2003). \Retinal oxygen: funda-
mental and clinical aspects." In: Archives of ophthalmology 121.4, pp. 547{57. issn:
0003-9950. doi: 10.1001/archopht.121.4.547. url: http://www.ncbi.nlm.nih.
gov/pubmed/12695252.
Webvision (2016). The Organization of the Retina and Visual System. url: http://
webvision.med.utah.edu/ (visited on 03/01/2016).
Weiland, James D. and Mark S. Humayun (2008). \Visual prosthesis". In: Proceedings of
the IEEE 97.7, pp. 1529{1559. issn: 0018-9219. doi: 10.1109/JPROC.2008.922589.
Wilson, C a, B a Berkowitz, and R Srebro (1995). \Per
uorinated organic liquid as an
intraocular oxygen reservoir for the ischemic retina." In: Investigative ophthalmology
& visual science 36.1, pp. 131{41. issn: 0146-0404. url: http://www.ncbi.nlm.nih.
gov/pubmed/7822140.
Xu, J et al. (2000). \Permeability and diusion in vitreous humor: implications for drug
delivery." In: Pharmaceutical research 17.6, pp. 664{9. issn: 0724-8741. url: http:
//www.ncbi.nlm.nih.gov/pubmed/10955838.
Yoshihara, Toshitada et al. (2012). \Ratiometric molecular sensor for monitoring oxygen
levels in living cells". In: Angewandte Chemie - International Edition 51, pp. 4148{
4151. issn: 14337851. doi: 10.1002/anie.201107557.
Yu, D Y and S J Cringle (2001). Oxygen distribution and consumption within the retina
in vascularised and avascular retinas and in animal models of retinal disease. url:
http://www.ncbi.nlm.nih.gov/pubmed/11173251.
104
Yu, Dao-Yi et al. (2007). \Intraretinal oxygen distribution and consumption during retinal
artery occlusion and graded hyperoxic ventilation in the rat." In: Investigative ophthal-
mology & visual science 48.5, pp. 2290{6. issn: 0146-0404. doi: 10.1167/iovs.06-
1197. url: http://www.ncbi.nlm.nih.gov/pubmed/17460293.
105

Abstract (if available)

Abstract Retinal vascular diseases, such as Diabetic Retinopathy and Retinal Vein Occlusions, can cause a profound visual handicap due to retinal ischemia. An intraocular medical device (Oxygenator) that can provide localized supplemental delivery of oxygen to ischemic retina could reduce the progression of retinal cell loss. For such a medical device to be feasible, we need to understand how oxygen travels in the eye, especially in the vitreous humor. Next, we need to develop an animal model of disease for the Oxygenator's validation studies. We also need to use a non-invasive method to monitor intravitreal oxygen levels over time. Finally, we need to alleviate the power constraints on such implantable ocular devices. ❧ First, to study vitreal oxygen mass transport, oxygen diffusion from an intravitreal oxygen source was studied in enucleated porcine eyes. The vitreous oxygen consumption rate was modeled after a second order rate reaction equation and the rate constant was experimentally quantified. A spatial variation in vitreal oxygen mass transport and vitreous oxygen consumption was observed. ❧ To develop an animal disease model for Oxygenator preclinical studies, a rabbit model of retinal ischemia induced by laser photothrombosis was developed and characterized. Intravitreal and preretinal oxygen levels was measured across 14 days in control and ischemic animal groups. Fluorescein angiography was also performed across 14 days to quantify perfusion in retinal vasculature. In this animal disease model, intravitreal oxygen levels are depressed for 7 days. Between Day 7 and Day 14, the retinal vasculature is re-perfused and intravitreal oxygen levels start to rise. ❧ To monitor intravitreal oxygen levels over time, a prototype of a non-invasive optical oxygen biosensor was developed. Luminescent oxygen sensitive dye molecules (platinum metal porphyrin) embedded in an acrylic patches were placed in the vitreous of enculeated bovine eyes. The patches were optically excited through the cornea and the emission light was recorded to obtain oxygen concentration measurements. This prototype oxygen biosensor can accurately measure intravitreal oxygen levels between 0 and 21%. ❧ To reduce the power constraints on ocular medical devices, an implantable solar energy harvesting power supply was engineered. A photovoltaic array implanted under the conjunctiva converted incident light energy to electrical energy stored in a battery. The subconjunctival photovoltaic array, with an area of 62 mm², when irradiated at 1mWmm⁻² of sunlight, can generate 1.22 mW at the battery charging circuit's output.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Towards a high resolution retinal implant

PDF

Investigation of the electrode-tissue interface of retinal prostheses

PDF

Improving stimulation strategies for epiretinal prostheses

PDF

Manipulation of RGCs response using different stimulation strategies for retinal prosthesis

PDF

PET study of retinal prosthesis functionality

PDF

Intraocular and extraocular cameras for retinal prostheses: effects of foveation by means of visual prosthesis simulation

PDF

Saliency based image processing to aid retinal prosthesis recipients

PDF

Stimulation strategies to improve efficiency and temporal resolution of epiretinal prostheses

PDF

Modeling retinal prosthesis mechanics

PDF

Prosthetic vision in blind human patients: Predicting the percepts of epiretinal stimulation

PDF

Understanding the degenerate retina's response to electrical stimulation: an in vitro approach

PDF

Dependence of rabbit retinal synchrony on visual stimulation parameters

PDF

The effects of the vitreous state on the ocular pharmacokinetics following intravitreal drug delivery: an evaluation with quantitative magnetic resonance imaging

$Intraocular camera for retinal prostheses: refractive and diffractive lens systems$

PDF

Intraocular camera for retinal prostheses: refractive and diffractive lens systems

PDF

Novel imaging systems for intraocular retinal prostheses and wearable visual aids

PDF

Electronics design and in vivo evaluation of a wirelessly rechargeable fetal micropacemaker

PDF

Characterization of visual cortex function in late-blind individuals with retinitis pigmentosa and Argus II patients

PDF

An in vitro model of a retinal prosthesis

PDF

Quantitative MRI for oxygenation imaging in cerebrovascular diseases

PDF

Electrical stimulation of the orbicularis oculi to restore eye blink

Asset Metadata

Creator Murali, Karthik (author)

Core Title Oxygen therapy for the treatment of retinal ischemia

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Publication Date 06/01/2016

Defense Date 04/25/2016

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

Tag diabetic retinopathy,OAI-PMH Harvest,oxygen,retinal ischemia,vitreous

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Weiland, James (committee chair), Humayun, Mark (committee member), Kashani, Amir (committee member)

Creator Email karthikm.kmn@gmail.com,karthikm@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-249555

Unique identifier UC11280698

Identifier etd-MuraliKart-4421.pdf (filename),usctheses-c40-249555 (legacy record id)

Legacy Identifier etd-MuraliKart-4421.pdf

Dmrecord 249555

Document Type Dissertation

Format application/pdf (imt)

Rights Murali, Karthik

Type texts

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

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

Repository Name University of Southern California Digital Library

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