Three-dimensional quantification of the spatiotemporal co-evolution of vascular and neuronal networks within intact eyes

Material Information

Three-dimensional quantification of the spatiotemporal co-evolution of vascular and neuronal networks within intact eyes
Singh, Jasmine N ( author )
Physical Description:
1 electronic file (41 pages) : ;


Subjects / Keywords:
Eye ( lcsh )
Retrolental fibroplasia ( lcsh )
Eye ( fast )
Retrolental fibroplasia ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Retinopathy of prematurity (ROP) is a disease that affects blood vessel development and distribution in the eyes of infants born prematurely resulting in vision impairment and loss. While many techniques exist for investigating this disease, a common method used to investigate structure-function relationships in ROP is serial sectioning followed by two-dimensional image analysis. Often, serial sectioning is fraught with inconsistencies due to tissue tearing and folding which may introduce optical artifacts during imaging. Of note are the errors that occur when multiple physical sections are computationally reconstructed to quantify the spatial location of fluorescent labels within the original three-dimensional tissue. In this work we utilize passive CLARITY technique PACT, which renders tissue optically transparent through the establishment of a monomer hydrogel matrix and removal of light scattering lipids to generate optically transparent eyes. To measure fluorescent labels within these intact eyes, we have used a newly developed and unique digital scanned light sheet microscope (DSLM) specifically designed to quantify fluorescently labeled signaling molecules and structures within PACT treated samples. These two techniques combined provide a methodology to quantify the three-dimensional distribution of key signaling molecules and structures during development of the eye, bypassing issues inherent in serial sectioning, two-dimensional imaging and computational reconstruction. Here, we report quantification of the vascular and neuronal network structures in intact control and endotoxin diseased model rat eyes. We provide a comparative analysis evaluating both two-dimensional and three-dimensional imaging techniques and find that network features in developing eyes are more accurately quantified using our three-dimensional imaging approach.
Thesis (M.S.)-University of Colorado at Denver.
Includes bibliographic references
System Details:
System requirements: Adobe Reader
Statement of Responsibility:
by Jasmine N. Singh.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
958707316 ( OCLC )


This item has the following downloads:

Full Text
B.S., University of Colorado, 2013
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Masters of Integrated Science
Integrated Science Program

This thesis for the Master of Integrated Science degree by
Jasmine N. Singh
has been approved for the
Integrative Science program
Douglas Shepherd, Chair
Xiaojun Ren
Joseph A. Brzezinski
Date: April 29th, 2016

Singh, Jasmine (MIS, Masters of Integrated Science)
Three-Dimensional Quantification of the Spatiotemporal Co-evolution of Vascular and
Neuronal Networks Within Intact Eyes
Thesis directed by Assistant Professor Douglas P. Shepherd
Retinopathy of prematurity (ROP) is a disease that affects blood vessel development
and distribution in the eyes of infants born prematurely resulting in vision impairment and
loss. While many techniques exist for investigating this disease, a common method used to
investigate structure-function relationships in ROP is serial sectioning followed by two-
dimensional image analysis. Often, serial sectioning is fraught with inconsistencies due to
tissue tearing and folding which may introduce optical artifacts during imaging. Of note are
the errors that occur when multiple physical sections are computationally reconstructed to
quantify the spatial location of fluorescent labels within the original three-dimensional tissue.
In this work we utilize passive CLARITY technique PACT, which renders tissue optically
transparent through the establishment of a monomer hydrogel matrix and removal of light
scattering lipids to generate optically transparent eyes. To measure fluorescent labels within
these intact eyes, we have used a newly developed and unique digital scanned light sheet
microscope (DSLM) specifically designed to quantify fluorescently labeled signaling
molecules and structures within PACT treated samples. These two techniques combined
provide a methodology to quantify the three-dimensional distribution of key signaling
molecules and structures during development of the eye, bypassing issues inherent in serial
sectioning, two-dimensional imaging and computational reconstruction. Here, we report
quantification of the vascular and neuronal network structures in intact control and endotoxin
diseasaed model rat eyes. We provide a comparative analysis evaluating both two-

dimensional and three-dimensional imaging techniques and find that network features in
developing eyes are more accurately quantified using our three-dimensional imaging
The form and content of this abstract are approved. I recommend its publication.
Approved: Douglas P. Shepherd

I. RETINOPATHY OF PREMATURITY...................................................1
Histology and Sectioning for Two-Dimensional Analyses of ROP..................5
Fluorescence Labeling Techniques..............................................7
Current Methodologies Used to Better Understand ROP...........................8
New Methodologies Available for Investigating ROP: Digital Scanned Light Sheet
Research Plan: Specific Aims.................................................10
Aim I: What is The Two-Dimensional Distribution of Blood Vessels in ROP and Control
Eyes (0-6 months.......................................................10
Aim II: Light Sheet Microscope Acquisition of Control and Retinopathy Eyes in Three-
Dimensions (2-8 months)................................................10
Aim III: Comparison of Two-Dimensions and Three-Dimensional Techniques for
Observing and Analyzing Fluorescence in Disease Model (6-12 months)....ll
II. DIGITAL SCANNED LIGHSHEET MICROSCOPY.......................................12
Literature Review of DSLM....................................................12
Image Analysis Software......................................................15
III. PASSIVE CLARITY TECHNIQUE, PACT...........................................17
Literature Review of PACT....................................................17
IV. METHODS....................................................................20

Animal Protocol................................................................20
Embedding and Cryo-sectioning..................................................20
Fluorescent Immunohistochemistry...............................................20
Two-Dimensional Fluorescence Microscopy Image Acquisition......................22
Three-Dimensional Digital Scanned Light Sheet Microscopy Image Acquisition.....23
Image Analysis.................................................................24
V. RESULTS.......................................................................25
Determining the Correct Antibody...............................................25
Two-Dimensional Images and Analysis............................................25
FIJI Quantification of Pixel Intensity Data..................................25
Three-Dimensional Images and Analysis..........................................26
FIJI Quantification of Pixel Intensity Data..................................26
VI. DISCUSSION...................................................................29

Retinopathy of prematurity (ROP) is a disease that develops in infants born at or
prior to 31 weeks and can result in impaired vision and blindness1. A common feature of
ROP includes abnormally developed blood vessels in the deepest most posterior layer of
the eye, the retina (Fig. 1). ROP was first documented in 1944 and associated with preterm
birth, low birth weight and overgrowth of portions of the eye2, and in the seventy years
following its discovery, researchers have sought to elucidate the mechanisms behind these
characteristic features of ROP. In normal infants, the eye develops in utero in the relatively
hypoxic environment of the amniotic sack and in this hypoxic environment, growth factors
such as vascular endothelial growth factor (VEGF) function to upregulate the formation of
new blood vessel branches in the retina3. However, if the infant is born prematurely, they are
briefly given supplemental oxygen to support the function of their lungs and in doing this,
vascular development within the eye is arrested as the excess oxygen downregulates
angiogenic promoting growth factors like VEGF4. After this brief therapeutic treatment, the
infant is taken off of supplemental oxygen, resulting in upregulation of VEGF expression
once again, however, because the previous angiogenic event was abruptly arrested, new
angiogenic events occurs abnormally due to increased expression of VEGF5 and other
angiogenic promoting growth factors (Fig. 2). Research has demonstrated that the longer an
infant is exposed to therapeutic oxygen, the worse their ROP will be, and as such, clinicians
try to minimize the duration of oxygen therapy6.

Ciliary body
Canal of
schlemm \
Conjunctiva V
Vein (central retinal)
- Rectus medialis
Figure 1: In retinopathy of prematurity, posterior and deep layer of the eye, the retina,
demonstrates over vascularization and can detach from external layers of the eyes such as
the choroid and sclera. This detachment leads to blindness. Image obtained from master eyes
associates 2016.
Literature Review of ROP
In analyzing the molecular mechanism involved in ROP, novel findings have
elucidated an intricate signaling cascade involving Hypoxia Inducible Factor-1 (HIF-1) and
Vascular Endothelial Growth Factor (VEGF). In normoxic (oxygen rich) tissue in adults, HIF-
1 is hydroxylated and bound to the Von Hipple-Lindau protein (VHL)7. Once bound to the
VHL protein, HIF-1 is ubiquinated and prompted for proteasomal degradation8. In hypoxic
(oxygen poor) conditions as seen in the amniotic sack, the enzyme responsible for
hydroxylating HIF-1, prolyl hydroxylase, is down regulated and HIF-1 accumulates in the
cytoplasm of hypoxic cells 5. As HIF-1 concentrations increase in the cell, HIF-1a and p
subunits dissociate from the whole HIF-1 protein complex and translocate to the nucleus to
bind to the promoter region of the VEGF gene, thereby upregulating VEGF expression.
Because the preterm fetus is still in a growth state, exposure to concentrations of oxygen
greater than that observed in the womb are hyperoxic to the infants resulting in oxygen
toxicity1. Under these hyperoxic conditions, HIF1 is destroyed, VEGF expression is down
regulated and angiogenesis is inhibited9. This down regulation occurs during important

developmental stages for the infant, and inhibition of neovascularization leads to reduced
oxygen levels in portions of the eye. As less oxygen saturates the tissues in the infants
eyes, fewer HIF-1 molecules are ubiquinated and more HIF-1a & p subunits can translocate
to the nucleus and bind to the promoter region of the VEGF gene7 (Fig. 2). With this binding,
VEGF expression is once again upregulated, however, due to the fact that the previous
normal angiogenesis was abruptly arrested, this subsequent angiogenesis occurs
abnormally, resulting in malformation of optic vasculature1.
Figure 2: In normal conditions in a developing fetus, HIF-1 protein persists in the cell and a
and p subunits dissociate from the protein complex and translocate to the nucleus to
upregulate VEGF expression. Under the hyperoxic conditions observed in preterm birth, HIF-
1 protein is hydroxylated, ubiquinated and degraded. As a result, HIF-1 regulated
angiogenesis is arrested. Variations in oxygen concentration in infant blood leads to variations
in the 2 mechanisms involved in this pathway and these variations results in ROP
ROP exists in various stages that require different methods of treatment. Preterm
birth of 31 weeks is identified as the first stage of ROP in which the pre-term infant has
received some amount of therapeutic oxygen1. This stage is physically visible in the eye as
a demarcation line or an area in the medial portion of the eye distinguishing the anterior

from posterior areas of the eye and is associated with a decrease in growth factors HIF-1
and VEGF and results in an abrupt stop in angiogenesis. The longer this reduction is
observed, the greater the likelihood the infants will develop severe ROP6. The second and
third stages are identified as a ridge forming at the demarcation line and extra-retinal
fibrovascularization. These stages of ROP are identified as the point at which
vascularization of the retina has abnormally increased. This increase occurs only when the
infants exposure to oxygen decreases during perinatal periods, resulting in an increase in
HIF-1 and VEGF expression10 and in the fourth stage, extra retinal fibrovascularization,
blood vessels can be observed growing out of the plane of the retina into the vitreous
humor. Transcript analysis of ROP model eyes at these two stages have shown an increase
in VEGF transcripts associated with this re-initiation of aberrant angiogenesis9. At this point
therapeutic anti-VEGF11 and anti-HIF12 medications can be administered to inhibit the over
activity of angiogenesis. The fourth and fifth stages of ROP are identified as partial and full
retinal detachment, and at this point in ROP, the re-initiation of angiogenesis has occurred
such that swelling of portions of the retina develop, resulting in retinal detachment. This
stage usually results in permeant partial or complete blindness. Progression of ROP through
each of these stages can occur over a matter of hours in preterm infants and because of
this, clinician take extra measures to modulate preterm infants exposure to oxygen6
Like VEGF, other growth factors have been implicated in the onset of ROP. Insulin-
like growth factor 1 (IGF-1) has been found to function in a similar manner to VEGF
throughout ROP progression13. Previous research has demonstrated that through
knockdown of IGF-1, angiogenesis is inhibited as observed in the first stages of ROP and
when IGF-1 is over expressed, angiogenesis occurs more rapidly as seen in later stages of
ROP. It was concluded that this relationship occurs due to IGF-Ts ability to phosphorylate
Akt, a signaling molecule involved in VEGF dependent angiogenesis13. Additional evidence

has shown that IGF binding protein-3 (IGFBP-3) can increase vessel growth in mice and is
also under expressed in infants with advanced ROP1. With this, it has been postulated that
IGFBP-3 can function to inhibit oxygen induced blood vessel loss by inhibiting IGF/VEGF
dependent angiogenesis and promote regrowth after blood vessels have been destroyed in
ROP onset14. With these findings, researchers continue to explore therapeutic uses for IGF
in treating retinopathy of prematurity
Histology and Sectioning for Two-Dimensional Analyses of ROP
For investigation of tissues, researchers often conduct histological analysis that
involves thin sectioning or cutting lateral slices of tissue. Each researcher utilizes varying
techniques for histological sectioning. But in general, tissues are fixed in paraformaldehyde
and then embedded in either paraffin in heat for wax sectioning or optimal cutting
temperature (OCT) media for cryo-sectioning (Fig. 3). In each instance, sections can be
between 5 pm and 120 pm thick. Once these section have been cut, they are adhered to
slides and treated with fluorescent stains15 for imaging.
Figure 3: Eves embedded in optical cutting temperature media are cryo-frozen prior to
sectioning. Sectioned portions of the eyes are stained with antibody labels and imaged (Red-
Calretinin and Green-Lectin Groffonia simplicifolia isolectin 4B). Embedded image obtained
from Peter et al. Pathology innovations LLC, 2003

While histological sectioning has been a gold standard for researching retinopathy,
sectioning itself is fraught with error. During embedding and sectioning of the eye,
mishandling of the tissues can easily results in tears or creased that become apparent
during imaging (Fig. 3). Similarly, embedding tissues in hot paraffin wax renders many
antigens disrupted or destroyed as heating (approximate 50C) in wax can degrade
endogenous proteins folded structure16. Due to the thin layers of cut sections, tissues within
each section are prone to tearing during handling and adherence to slides. Because the eye
is composed of various types of cells with different characteristics, certain portions of the
eye such as the sclera, do not adhere to the slide easily and can be completely removed
during washes. With all these issues, the final product of the section adhered to a slide is
quite different from the eyes that are observed in vivo. Researchers have utilized these
histological sections with image registration software to reconstruct an image of the entire
organ systems. However, because of the errors inherent in histological sectioning such as
tissue tearing or inconsistent fluorescent labeling, reconstructed eyes often provide
incomplete or inaccurate information regarding structure wide protein distributions in the
To circumvent the issues observed with histological sectioning, researchers now
focus their efforts on more localized dissection techniques such as flowering of the retina. In
flowering, eyes must first be dissected, removing the internal lens and vitreous as well as
the external sclera and choroid. The remaining anatomy, the retina forming a small cup
attached to the optic nerve is then flowered. Small incisions are made along the edges of
the retina allowing it to lay flat and resemble a flower. This technique has been referred to
as flowering but is more commonly known as whole mount. Researchers utilize different
variations of labeling and dissecting, but in general, the thick piece of tissue is then whole
mounted and fixed to a slide and prompted for subsequent imaging15.

Fluorescence Labeling Techniques
To determine the total expression of certain proteins or cell types in an organ
system, researchers utilize immunohistochemistry (IHC). In IHC, tissues are treated with
immunoglobulins or antibodies for the detection of other proteins. First the tissue is fixed in
alcohols or aldehydes to strengthen its morphological structure. Then the tissue is treated
with mild buffer to restore any altered antigens in preparation for antibody labeling. After
this, the tissue is treated with animal serum to block non-specific binding or binding of the
antibody to proteins other than the protein of interest. Once the tissue has been prepared it
is then treated with antibodies that bind to the protein of interest. In this technique, a primary
antibody from one species of animal is used to bind to the protein of interested. A secondary
antibody is conjugated to a fluorescence dye and has binding specificity to the primary
antibody and so can bind to the site of the primary antibody while also emitting fluorescent
light. Because this labeling requires the use of two antibodies, the prevalence of non-
specific binding increases. However, with the use of more overall antibody, robust
fluorescence of proteins of interested can be attained as concentrations of each antibody
can be greater17 (Fig. 3 & 4).

Figure 4: Whole mount retinal cup. In flowering the retinal cup, we were able to acquire
information regarding the distribution of blood vessels in the retina in two-dimensions using
confocal fluorescence microscopy. Blood vessels are labeled with GSi4B Lectin conjugated
to 488 fluorescence molecule.

Current Methodologies Used to Better Understand ROP
Laboratory methodologies used for analyzing ROP pathogenesis in eye samples
have focused on the use of scanning electron microscopy (SEM) and confocal fluorescence
microscopy. Using SEM, a sample is bombarded with accelerating electrons that loss
energy in the form of light or heat once they strike the specimen. This emitted heat or light is
detected and then used to determine spatial information about the structure of the sample18.
While SEM can be used to image specimens, it cannot penetrate through the tissue of a
specimen and, as a result, only the topography of a specimen maybe obtained18. Similarly,
confocal fluorescence microscopy can be used to image the eye with additional dissection
methods. Many advances in understanding ROP have been made through whole mount
techniques. In flowering the retina and staining it with fluorescent antibodies, much
information can be obtained regarding the spatial distribution of proteins in endothelial cells
and neurons and as such whole mounting optic tissue has become a gold standard
technique in investigating ROP. However, flowering not only disrupts the normal
conformation of the retina, but also may distort the vasculature, resulting in abnormal folding
or bending of blood vessels that can produce imaging artifacts. Additionally, because this
technique relies on converting the three-dimensional retinal cup into two-dimensions, many
morphological features such as retinal swelling and detachment are not easily observable.
In clinics, ophthalmoscopes or high resolution cameras and light are placed in front
of the pupil of patients eyes and can observe the internal portions of the retina. This
technique is widely used in live patients to diagnose ROP19. Because it is predominantly
done in live patients, additional techniques such as fluorescent labeling and microscopy
cannot be done with this technique. While each of these technique have, in some capacity,
provided researchers and healthcare providers more information regarding ROP, each has
its own shortcomings. With whole mount techniques altering the three-dimensional

conformation of the eyes and ophthalmoscopes implementation in only living patients, a
more comprehensive technique must be utilized to understand the onset of ROP in a spatial
manner. As such a more comprehensive three-dimensional imaging technique can be
employed to evaluate the role of over expression of signaling molecules on the overall
structure of the eye and once such methods is light sheet microscopy.
New Methodologies Available for Investigating ROP: Digital Scanned Light Sheet
Recent advances in microscopy and imaging have provided new techniques for
quantifying fluorescent signaling with intact organs such as the eye in three-dimensions.
One such example is the advent of digital scanned light sheet microscopy (DSLM) in which
a thin sheet of light is created by passing a beam of laser light through a beam expanding
objective. This thin sheet is then reflected off of a vertically oscillating mirror to create a light
sheet20. This sheet is then projected into the sample resulting in selective illumination of thin
sheets within the whole sample, exciting fluorescent markers incorporated throughout the
tissue. Images obtained using digital scanning light sheet microscopy (DSLM) can then be
reconstructed to generate a high-resolution three-dimensional images of the organ.
With the feature of gross tissue penetration, DSLM offers new effective ways to
observe an organ or tissue sample in its normal structural conformation from the inside-out.
However, because most tissue contains many light diffracting molecules like lipids,
biochemical techniques may be employed to render the tissue optically transparent and one
such technique is Clear Lipid-exchanged Acrylamide-hybridized Rigid
Imaging/lmmunostaining/ln situ hybridization-compatible Tissue-hYdrogel or CLARITY. By
replacing lipids in the tissue with transparent acrylamide hydrogel, CLARITY effectively
renders tissues optically transparent21. In combining light sheet microscopy, optical tissue
clearing and fluorescent labeling, high resolution images can be generated and used to

acquire information regarding the spatial distribution of proteins throughout the eye during
disease onset and progression.
Research Plan: Specific Aims
Aim I: What is The Two-Dimensional Distribution of Blood Vessels in ROP and Control Eves
(0-6 months)
The goal of my first aim was to observe variations in blood vessel distribution for both
control and experimental eye models that were exposed to Escherichia coli endotoxin which
previously has shown to produce an insult to pulmonary vasculature3. To do this I cryo-
sectioned and stained the eyes using the IHC protocol from Biolegends titled
Immunohistochemistry protocol for frozen sections and then imaged the antibody stained
histological sections using a Nikon Eclipse Ti fluorescent microscope. This microscope
provided images of whole retina sections including blood vessel and neuron network for
further analysis in image analysis software, FIJI22. In using this technique, I was able to
identify portions of the eye such as the retina and choroid to focus on when imaging via the
light sheet microscopy in three-dimensions and acquired gross spatial information regarding
the morphology of the eyes neuronal and vascular components (Fig. 1).
Aim II: Light Sheet Microscope Acquisition of Control and Retinopathy Eves in Three-
Dimensions (2-8 months)
In parallel with histological imaging from Aim 1, I prepared another sample of eyes
for light sheet imaging. To properly image any tissue using the light sheet microscopy, the
tissue must be optically transparent and to do this I use the general passive CLARITY tissue
clearing protocol, PACT23. Images were acquired using custom LabVIEW digital imaging
software and analyzed using additional three-dimensional image analysis software, FIJI and
Vaa 3D. From these analyses, I have obtained information regarding quantity and structural
variations that exist in blood vessels and neurons in both control and endotoxin models.

Aim III: Comparison of Two-Dimensions and Three-Dimensional Techniques for Observing
and Analyzing Fluorescence in Disease Model (6-12 months)
After acquiring two and three-dimensional images, I conducted a series of
comparative analyses. I evaluated any difference that exist in vascular and neuronal
networks between the control group and the experimental eyes in both the two-dimensional
and three-dimensional images. Through the use of particle counting algorithms for analyzing
the two-dimensional images and network analysis algorithms for analyzing the three-
dimensional images I observed differences in protein expression in neurons and endothelial
cells between control and diseased eyes. With the particle counting algorithm I was able
ascertain the number of fluorescence pixels at or above a pre-determined intensity and in
doing this, I acquired information regarding the number of fluorescently labeled proteins as a
function of the number of photons reaching the detection camera to generate these high
intensity pixels in the image. To standardize the values of pixel intensities I obtained, I then
calculated the number of high intensity pixels per the imaging area in pm3. For three-
dimensional analysis, I generated network maps of both the blood vessels and neurons
distributed throughout the eyes. Using a network analysis algorithm in Vaa 3D I obtained
network characteristics like average branch diameters and average network volume.

While many microscopy techniques offer the ability to generate images from samples
in three-dimensions, very few offer a thorough and concise delivery of high resolution
images with minimal reconstruction required. Recent modifications to the traditional light
sheet fluorescence microscope have enable researchers to image optically sectioned
tissues through the use of light and optics manipulation in a technique dubbed Digital
Scanned Light-Sheet Microscopy or DSLM. In utilizing a beam of laser light, thin sheets of
fluorescent light can penetrate cleared tissue and excite fluorescently labeled proteins. In
doing this, researchers are able to conduct a comprehensive interrogation of protein
distributions at the level of a whole organ. Additionally, with these developments, images
demonstrate a high signal to noise ratio because the excitation light is striking the sample at
an angle perpendicular to the detection objective. With the ability to generate high resolution
images with rapid turn-around time, DSLM is ideal for analyzing protein distribution in whole
organ tissues systems.
Literature Review of DSLM
Novel findings in microscopy techniques have played a pivotal role in the
advancements of biomedical research, and throughout scientific history, microscopy
techniques such as fluorescence microscopy and electron microscopy have allowed
researchers to image previously elusive things such as viruses and proteins. While many
advantages exist with each of these microscopy techniques, each also demonstrates
limitations. Due to limited selective illumination and the presence of light scattering
molecules in samples that produce background fluorescence, fluorescent microscopy allows
for imaging only of flat (200 pm or less in thickness) two-dimensional samples. Because of

this, tissues must be structurally altered to accommodate the increased detection of
background out of focus fluorescence. Similarly, with electron microscopy, penetration of
thick samples is not possible and as such only topographical images of tissue are obtained.
While two-photon and confocal fluorescence microscopy offers light penetration in thick
tissue samples, each utilizes a beam of high energy light to excite fluorophores within the
tissue and as a result photo bleaching occurs rapidly with only small sections of the tissue
being imaged at a time. As such, reconstruction of three-dimensional images from image
obtained using two-photon or confocal fluorescence microscopy can be time consuming and
computationally intensive. Of these techniques none has the ability to image large sections
of tissues with high contrast and limited photo bleaching to generate a three-dimensional
One microscopy technique with capacity to image large thick tissues samples is light
sheet microscopy. In 1993 the one of the first examples of fluorescence light sheet
microscopy, dubbed orthogonal plane fluorescence optical sectioning (OPFOS) was
created24. This method of microscopy utilized a laser, beam expander and cylindrical lens to
generate a sheet of laser light. This sheet of light was then used to penetrate and illuminate
sections of an optically cleared and fluorescently labeled cochlea. Images of each section of
the cochlea were captured and reconstructed into a three-dimensional image. From this,
internal features of the cochlea, such as the cochlear duct, otherwise hidden to the naked
eye were observed for the first time in complete in vivo structures.
Advances in three-dimensional microscopy techniques have continued since OPFOS
in 1993 and one technique that has refined many features of OPFOS is DSLM. In this
technique, a plane of light is generated by passing laser light through a beam expanding
objective. This plan of light is then reflected off of an oscillating mirror and laterally
translated into the image plane creating a complete light sheet. With this rapid scanning
beam, many advantages to imaging are observed including uniformed delivery of light

throughout the specimen and enhanced in-plane volumetric imaging with micron optical
sectioning to generate images with greater signal to noise ratio and higher resolution
images25. Most DSLM are comprised of 6 pieces or subunits; a light source, a beam
shaping device, a scanning excitation system, an apparatus for holding the specimen, a
detection system, and the software used to generate and analyze three-dimensional images
20. Generally, both excitation and detection objectives are immersed in media matching the
index of refraction of the tissues or specimen mounted in the microscope. In doing this,
scattering of excitation light either through the media or through the specimen is reduced
allowing for decreased noise and increased signal detection.
With the set up described above, some disadvantages have been discovered. First,
immersion objectives specialized for certain index matching medias can be very expensive
and maybe hard to purchase for one lab or research group. In using common water
immersion objectives which may be a little less expensive, the index of refraction for tissues
seldom match the index of refraction for water and as a results, scattering of excitation light
will produce images with low signal to noise ratio and low resolution. Second, because the
sample is moved instead of the light source, image artifacts can be detected as a result of
moving the sample. To address these two issue our lab has made some modifications to the
light sheet microscope in mounting the sample to a rotatable stage and using air objectives.
To address the issues of limited focal length in both excitation and detective objectives we
have included the use of an electrically tunable lens (ETL) to our apparatus set up. ETLs are
lenses made of a polymer that changes shape when subject to electric current, and in
applying a current to the ETLs, the curvature of the lens changes as well as the focal
length20. Motivated by previous work on remote-focusing DSLM using ETLs, we present a
C-DSLM (cleared tissue digital scanned light-sheet microscopy), in a technique that
manipulates remote focusing to concisely separate axial positioning of the excitation and

detection planes within the specimen and remove movement of the sample to account for
variations in optical properties of cleared tissue while keeping the specimen stationary.
With C-DSLM, for efficient excitation of endogenously expressed or exogenously
labeled fluorescence proteins, all portions of the samples must be rendered optically
transparent and with advances in biochemistry, this is now possible. One clearing method
CLARITY replaces light diffracting lipids in a tissues sample with an optically transparent
hydrogel26. Not only does this allow for the retention of the tissues original structure, but
also preserves DNA, RNA and proteins endogenously found within the sample. As such,
fluorescent staining techniques can be conducted in samples that have been cleared. This
technique has enabled researchers to observe gene expression and protein distribution in
tissues that are otherwise difficult to image23.
Image Analysis Software
The availability of powerful software is required for analyzing the large amount of
images produced by C-DSLM and two open source software, Vaa 3D and FIJI, are widely
available to researchers currently. Both software function to analyze fluorescent images
through observing pixel intensity, and provide users a comprehensive approach to
fluorescent analysis in various types of images. Because of their use of mathematic
algorithms in identifying pixel intensity, both softwares allow researchers the ability to
distinguish image artifacts from real fluorescent spots through thresholding, thereby
enabling accurate quantification of fluorescently labeled proteins in images.
FIJI is a java based program that was developed in 1987s as NIH Image by Wayne
Rasband and in developing NIH Image, Rasband sought to provide researchers with
affordable and user friendly image analysis software compatible with Mac computers22. In
the 30 years following its development, the Java based program has been refined not just by
Rasband but also by users of the software and now includes three hundred macros and
over five hundred plugins developed by researchers around the world22. Because of its

open source availability, researchers can now use FIJI to analyze various types of biological
images ranging from images of single cell cultures to three-dimensional reconstructions of
neuronal networks. As such, FIJI has provided biological researchers powerful analysis
capabilities needed for quantifying fine features in cells and tissues alike.
Another three-dimensional image analysis tool dubbed Vaa 3D is a Dos based
program that, like FIJI, has the ability to quantify the number of fluorescence points found in
a large image sequences. Additionally, Vaa 3D utilizes exemplar points and network maps
established by user to count the number of fluorescent spots with similar characteristics or to
determine network maps of blood vessels and neurons (Fig. 5). From these network maps
information regarding frequency and degree of branching as well as blood vessel diameter
can be obtained. With this, Vaa 3D can generate clean high resolution images of not only
protein distributions but also neuronal and vascular networks in three-dimensions27.
Figure 5: Images for fluorescence distributions of endothelial cell networks in three-
dimensions. From the reconstruction of vascular networks using Vaa 3D, information
regarding degree and characteristics of branching can be determined, difference in network
characteristics can be used to evaluate differences in control and disease model eyes. A)
Volume vascular system at the retina, identified by locating the optical nerve and then
traversing directly into the eye. B) Multiscale tubeness filtering to highlight the vascular
network within the retina layer (scale bar 500 Dm)

With the many advancements in fluorescence microscopy, imaging tissues with light
detracting molecules continues to be problematic and as such, a need for optically
transparent tissues have grown. Research conducted using C-DSLM prior to tissue clearing
methods utilized already optically transparent models such as zebrafish. However, as
interests in analyzing the three-dimensional nature of mammalian tissues have grown, novel
methods for rendering these tissues optically transparent have also advanced. One such
technique, CLARITY, utilizes methods in which tissue is rendered optically transparent while
maintaining the structural integrity of endogenously expressed proteins26. These techniques
exploit the same biochemistry used in standard SDS-page by utilizing acrylamide to form a
hydrogel matrix within the tissues and then uses SDS to removed light diffracting lipid
molecules. By removing light diffracting lipid molecules, CLARITY has built on previous work
conducted to render tissues optically transparent by increasing the distance between lipids
allowing more excitation light into the detection objective of microscopes. However, in using
the milder SDS detergent as compared to previously used ethers and hydrofurans, the
longevity of fluorescent molecules in the tissues in increased23. In employing techniques
from previous work that manipulate light detracting lipid molecules in tissues to reduce
background fluorescence and generate high resolution images, CLARITY has refined tissue
Literature Review of PACT
With the advent of optical sectioning and recent advances in light sheet microscopy,
the need to remove light scattering molecules from tissue for efficient image acquisition has
grown. One technique that achieves this is passive CLARITY technique (PACT). PACT was

introduced in 2013 in a study examining neuronal distributions within intact mouse brains,
and also demonstrated translation into whole human tissue samples 26. In this study the
basic chemical mechanisms by which optical transparency is rendered in tissues is outlined
and in building on previous work in the field, researchers were able to effectively remove
light diffracting lipid molecules from this tissues (Fig 6). As outlined in the 2013 paper,
tissues are incubated in acrylamide monomer which, upon polymerization, creates a
hydrogel matrix that can subsequently crosslink with endogenous proteins, DNA and RNA.
Next, eyes are degassed using Nitrogen to remove any reactive oxidative species that may
function to quench endogenous or IHC labeled fluorescent molecules. After this, the tissue
is incubated for several days in a solution containing high concentrations of sodium dodecyl
sulfate (SDS) a strong amphipathic detergent. Due to the charge and molecular structure of
SDS, the detergent can coat lipids in negative charge, facilitating their aggregation into
larger lipid micelles. With the formation of lipids micelles, light diffracting lipid molecules can
effectively be removed from the tissues and for this research, washing the eyes in
phosphate buffered solution (PBS) removed all of the micelles present in the tissue while
also maintaining its structural integrity26. Once these light diffracting molecules are removed
from the tissue, fluorescently labeled tissues may immediately be incubated in index
matching media and prompted for imaging
Figure 6: Stage of tissue clearing in PACT CLARITY procedure. 1. The eyes is harvested from
the infant rat. 2. The lens and cornea are dissected out. 3. The eye is incubated in acrylamide
monomers. 4. Acrylamide monomer polymerizes forming the hydrogel matrix. 5. The hydrogel
matrix cross-links with endogenous biomolecules like proteins, DNA and RNA. 6. The light
diffracting lipid molecules are washed out of the eyes with SDS and PBS. 7. The eye is labeled
with antibodies.

PACT has some inherent issues that continue to be addressed and one such
obstacle is the implementation of antibody labeling in cleared tissues. Due to the
concentration of acrylamide used to establish the hydrogel matrix, diffusion of full antibodies
through the matrix can be problematic and to alleviate this issue, researchers have
employed alternative methods. First, the permeability of the hydrogel matrix has proven to
be a function of the concentration of acrylamide used in the matrix establishment and in
reducing the concentration of acrylamide, a more porous hydrogel matrix is obtained.
Additionally, because a large volume of tissue has undergone treatment with acrylamide and
SDS buffers, it has become necessary to use a much larger concentration of antibody, an
approximate ten-fold increase, than that observed in histological IHC. This ensures that
more antibody persists in the tissues and reach all of the target protein epitopes than
antibody that are lost in treatment with chemical buffers. Finally, researchers have employed
the method of FAB or fragmented antigen binding in which the antibody is fragmented into
particles approximately 50 kDa in weight and then introduced into the cleared tissues. FAB
techniques have demonstrated more precise labeling as penetration of the antibody
fragments into the hydrogel is easier with fragments of smaller sizes. In utilizing these
variations in antibody staining and PACT, methods for obtaining sample perfect for image
continues to be refined23.

Animal Protocol
Albino adult female rats (Sprague Dawley) were purchased from Charles River
Laboratories. Mothers of control rats received intra-amniotic injections of saline and mothers
of diseased rats received intra-amniotic endotoxin injections from Escherichia coli at 20
days gestation. All protocols for rearing rats and administering endotoxins are outlined in
Tang et al. 2010. Two days after receiving injections caesarian sections were preformed and
maternal rats were sacrificed. Diseased and control pups were reared at room air for 2
weeks with a surrogate mother. After 14 days pups were sacrificed and eyes were
harvested. Eyes were incubated in 4%paraformaldehyde for 30 minutes and then the lens
and cornea of the eyes were removed.
Embedding and Cryo-sectioning
Eye were incubated in 30% sucrose for 24 hours, then a 50-50 mixture containing
30% sucrose and Optimal Cutting Temperature Media (OCT) from Scigen (catalog #4583)
for 24 hours. After this eyes were incubated in 100% OCT for 24 hours, placed in cryo-molds
(Tissue-Tek Cryomolds Catalog #4557) and froze at -80C. Frozen embedded eyes were
then sectioned on a microtome to 7 urn thick sections and adhered to slides (Fisher brand
Superfrost plus microscope slides catalog #12-550-15). Slides were then stored at -20C
and stained within 1 month of storage.
Fluorescent Immunohistochemistry
Slides containing eye sections were removed from -20C freezer and incubated at
room temperature for 3 minutes. Slides were then washed in PBS for five minutes and then
incubated in blocking buffer for 2 hours. Blocking buffer contains 10% normal horse serum

(Vector labs lot#ZB0630), 11% Bovine serum albumin (BSA) (Fisher Scientific CAS 9048-
46-8) and 0.5% Triton 100 (Sigma Aldrich T9284). Slides were then incubated in 1:250
concentration of mouse Calretinin antibody (Merk Millipore catalog #1568) and 1:500
concentration of Biotinylated GSi4B Lectin (Vector Labs Catalog #B-1205) over night. Eyes
were then washed 5 times in PBS for 5 minutes and then incubated in 1:500 concertation of
goat anti-mouse secondary antibody conjugated to Alex 647 flor (Thermifisher Labs) and
1:750 concentration of 488 conjugated streptavidin (Jackson ImmunoResearch Labs catalog
#016-540-084) for 2 hours. Slides again were washed in PBS 5 times for 5 minutes and
mounted with DAPI hard mounting media (Vector Laboratories Catalog # H-1500) and a
coverslip. Slides were imaged on a Nikon Eclipse Ti fluorescent microscope and fluorescent
molecules were excited with light at 640 nm of light and 488 nm of light. Images were then
analyzed using particle counting plugins in FIJI image analysis software.
To render tissues optically transparent, eye harvested from pups were incubated in
4% acrylamide and 0.25% photoinitiator,2,2-Azobis[2-(2-imidazolin-2-yl) propane]
dihydrochloride (VA-044, Wako Chemicals USA) monomer solution overnight a 4C. Infused
samples were then degassed in nitrogen for 5 minutes and incubated at 37C for 2 hours to
initiate hydrogel polymerization and crosslinking of hydrogel matrix to endogenous proteins.
Eyes were removed from hydrogel solution and washed in PBS. Eye were then incubated in
8% SDS solution for 5 days at 37C with gentle shaking. Eyes were then washed 5-8 times
over the course of an 8-hour day in PBS to remove excess SDS in sample tissue. Eye were
then incubated in primary antibody solution containing a 1:100 dilution of antibody in .02M
phosphate buffer with 2% normal horse serum and 0.2% triton X 100 for five days. Primary
antibodies used were Calretinin mouse antibody (Merk Millipore catalog #1568) and
Biotinylated GSi4B Lectin (Vector Labs Catalog #B-1205 (Fig. 7). Eyes were then washed in

PBS to remove excess primary antibody and incubate in secondary antibody solution for 2
days. Secondary antibody solution contained the same amount of serum and triton X 100
with 1:250 dilution of goat anti-mouse secondary antibody conjugated to Alex 647
(Thermifisher Labs catalog # ab150159) and 488 conjugated streptavidin (Jackson
ImmunoResearch Labs catalog #016-540-084). After this, eyes were washed in PBS to
remove excess secondary antibody and then incubated in RIMs imaging media until cleared.
RIMs imaging media contained 40 g of Histodenz (Sigma D2158) in 30 ml of .02M
phosphate buffer in 0.1% triton X 100.
Figure 7: Eyes harvested from infant rats display light diffracting properties and opaque
tissues. Upon undergoing PACT CLARITY, the eyes are render completely optically
transparent and are not easily visible to the naked eye.
Two-Dimensional Fluorescence Microscopy Image Acquisition
Once microscope slides with histological sections are mounted with Dapi hard
mounting media and coverslip, the slides were mounted onto the Nikon Eclipse Tl
fluorescence microscope and imaged in NIS imaging software. Upon starting NIS imaging
software, the sections were observed at 4x and 10x magnifications to determine general
morphological characteristics. Then at 40x magnification, the exposure times of the camera
for both 488 nm and 640 nm lasers was set and recorded and tiling parameters were set.
Tiling image run was then initiated with check points set ever few imaging areas for adjusting

the lens focus. Once an entire tiled image was generated, files were saved as tif and exported
onto an external hard drive
Three-Dimensional Digital Scanned Light Sheet Microscopy Image Acquisition
Once the eyes were incubated in Histodenz index matching media and rendered
optically transparent, the eyes were the adhered to 20 pi pipette tips and mounted into a
quartz cuvette containing Histodenz (Fig. 8). The lasers and mirror were then turned one via
the power strip and the lasers were turned on in Obis software. Once the sample was
mounted onto the microscope and the lasers were operating, lens and laser calibration and
image acquisition was conducted in LabVIEW. Upon starting LabVIEW software, both
excitation and detection lenses were calibrated for both 488 nm and 640 nm lasers by
shining the non-oscillating or stationary laser beam through portion of Histodenz not
occupied by tissue. The cuvettes position in the z-plane was then adjusted until the laser
beam appeared as a thin, evenly distributed beam of light. After this initial lens calibration,
the mirror governing the formation of the light sheet was turned on and the tissue samples
was moved into the light sheet. To calibrate the detection objective focal length, the laser is
moved through the sample in the z-plane in 150-200 pm units several times and for each
position, the voltage applied to the detection objective lens is changed until the image
rendered in LabVIEW is clear. In doing this, the exact distance in the z-plan within the tissue
sample corresponds to the correct focal length of the detection objective lens and once
these calibration values have been recorded and saved, image acquisition was run. To
accommodate the working distance of the lens, a maximum number of sheets in the z-plane
were calculated and this values was used in LabVIEW during image acquisition. Obtained
image stacks were saved as tiff.

Figure 8: Cleared and fluorescently labeled eyes are mounted into quartz cuvette containing
index matching media. Cleared eyes are mounted to 20 pi pipette tip at the site of the optic
nerves and places in the index matching media. ETL facilitates the imaging of the whole eye
will little movement of the sample
Image Analysis
Images obtained from both the light sheet and two-dimensional fluorescence
microscopy techniques were analyzed in ImageJ. For the two-dimensional image analysis,
the fluorescent multichannel image was split into separate channels, a threshold was
applied to each image channel and the particle counting algorithm in FIJI was used to
generate pixel counts at or over the threshold intensity. Similarly, with the images obtained
from the light sheets, ten random images from the 500 image stack were obtained, a
threshold was applied and the same particle counter algorithm in FIJI was conducted on
each of these ten images. Once particle values were obtained, the numbers of intense
pixels per imaging area was calculated for each image. The average and standard
deviations were then obtained for both the two-dimensional images and the stacks of ten
images from the whole light sheet image. The averages were compared and p-values were
determined from a standard t-test.

Determining the Correct Antibody
For determining the correct antibodies for both neuronal and endothelial labeling
some challenges arose. First, many antibodies compatible with mouse or other murine
species such as anti-ERG-1 and anti-Von Willibrand Factor showed limited specificity in rat
model eyes and this was observed by limited specific fluorescence during imaging.
Additionally, one antibody similar to anti-Von Willibrand Factor and specific to rats, anti-
PECAM, also demonstrated low specific binding in rat eyes in both histological sections and
PACT cleared tissues. Due to the issues faced in using antibody proteins in labeling
endothelial cells in rat eyes, we explored the use of alternative sugar binding molecules for
endothelial labeling and found that a biotinylated Lectin produced the most robust and
specific labeling. Unlike antibody labeling in endothelial cells, antibody labeling in neuronal
cells proved to be much easier and of the two antibodies used, anti-Calretinin and anti-Pax
6, the Calretinin antibody demonstrated a much more specific binding to neuronal cells in
the retina.
Two-Dimensional Images and Analysis
FIJI Quantification of Pixel Intensity Data
For image analysis of two-dimensional images, histological sections were labeled with anti-
Calretinin in the 640 nm channel and biotinylated Lectin GSi4b in the 488 nm channel,
images were loaded into FIJI and channels were split. A flood filled threshold in FIJI was
then applied to the images and particle counting algorithm in FIJI was ran. Particles for
approximately four different control and diseased two-dimensional images were processed
in this way for both neuronal labeled and endothelial labeled images. All of the pixel intensity
counts obtained for each image were divided by the imaging volume in pm3. An average and

standard deviation in counts per volume were obtained. The variations in counts of intense
pixels observed in disease and control models demonstrated no significant difference
however, these counts may be inaccurate due to nonspecific binding of each antibody
observed in the sclera of both control and diseased models.
Three-Dimensional Images and Analysis
FIJI Quantification of Pixel Intensity Data
In keeping with the same sampling and processing methods as that in the two-
dimensional image analysis, endothelial and neuronal images for both control and diseased
models were processed separately in FIJI. Each image sequence generated by the light
sheet microscope contained approximately 500 image stacks and of these 500 stacks, 10
stacks were used to obtain pixel intensity with the particle counting algorithm as 10 image
stacks roughly equate to the thickness of one histological sections, or 7 pm. Each of these
10 stacks were processed identical to the histological images, a flood filled threshold was
set and the particle counting algorithm was implemented for determining the number of
pixels above an average intensity. When counts were obtained from each of the ten images
in one image stack, the counts per imaging area were obtained. These average counts and
standard deviations were calculated for 4 different images stacks (Fig. 9).

2D histological sections

Neuronal Cells
Endothelial Cells
PACT treated eyes
£ 160e-4'
x HOe-4-
B 60e-4a
10 e-^a
Neuronal Cells
In progress
Endothelial Cells
Figure 9: Average pixel counts per imaging area with average standard deviations for the
entire dataset are represented by the error bars. A large range in standard deviation exist in
the analysis of two-dimensional imaging methods. Despite the greater range that exists in
utilizing this methods, a statistical significance is observable between control and diseased

Vaa 3D Network Analysis Data
Because so much variation was present in the data collected from the FIJI particle
counting software, we decided to implement the use of an alternative software Vaa 3D for
determining the characteristics of blood vessel networks. Much like the particle counting
algorithm previously implemented in FIJI, network tracing algorithms in Vaa 3D utilized pixel
intensity to identify fluorescently labeled network structures in tissues. In employing this
technique for analyzing network structures, information regarding branch diameter and total
network volume were obtained and compared among control and diseased eyes, however,
due to limited time we were unable to conduct further analysis beyond the control model
(Fig. 10)
Figure 10: Network analysis conducted on PACT cleared eyes show blood vessel networks
in the highly vascularized retina of control rat eyes. Blood vessels are labeled with Lectin
GSi4B saccharide targeting endothelial cell surface proteins. Average vascular length =
26,9509 pm, network volume = 5.2e7 pm3 (Scale bar = 500 pm)

In implementing two dimensional imaging techniques such as confocal or two-photon
fluorescence microscopy, researchers have acquired greater insight into the nature of ROP
as a disease. In employing these techniques researchers have been able to demonstrate
that ROP is characterized by an increase in vascularization of the retina due to an increase
in expression of the VEGF gene under the regulation of the HIF-1 protein. This increase in
vascularization has been easily observed using two-dimensional fluorescent microscopy
techniques. With these findings however, little or no information has been obtained about
the influence of over vascularization on the three-dimensional structure of the eyes or or its
various layers beyond the possibility for retinal detachment. To evaluate the relationship that
exists between the retina and the structural integrity of the eye during angiogenesis, we
have utilized novel techniques in three-dimensional fluorescence microscopy, and have
found differences in the data collected.
In implementing fluorescent microscopy techniques for the analysis of two-
dimensional histological sections of rat eyes, we have found that the number cells as
demonstrated by intense pixel counts are lower in endothelial cells and greater in neuronal
cells in the endotoxin diseased model eye. This supports previous findings that endotoxin
induces an insults to angiogenesis in pulmonary tissues28. Additionally, the increase in
neuron counts in the diseased eye could be associated with inhibition of neuronal pruning
events that occur in rats during their first weeks of life. While these data provide some
preliminary insight into vascular and neuronal counts in the endotoxin model, the high
degree of variation that exists in pixel intensity counts may be the result of sample bias or
the product of mislabeled tissues.

The variations observed in pixel intensity counts obtained from analysis of two-
dimensional images could also result from some experimental errors. First, treatment of thin
histological sections is prone to error during handling and fluorescent labeling. As a result
images provide only nominal true information regarding cell counts and distributions within
an organ system. Additionally, in two-dimensional image analysis of either cell counts or
network structure, high resolution images must be generated. We utilized tiling options
available in fluorescence microscopy software which was not only time consuming but
generated high background noise which inhibited our ability to confidently generate cell
counts or acquire network information. Finally, due to nonspecific binding of fluorescent
antibodies in scleral region of the eyes, fluorescence intensity counts prove to be unreliable.
Due to the issues encountered in quantifying fluorescence in histological section,
researchers most commonly employ flat mounting techniques to obtained such quantifiable
information of control and diseased eyes and this provides an alternative method for
comparison of two-dimensional imaging to three-dimensional imaging techniques in the
In conducting three dimensional imaging of the eyes utilizing the same methods for
analyzing the two-dimensional images, we have generated counts with slightly more
confidence as shown by a significant difference in p-values from t-testing, however, the
large variation in pixel intensity values still remains. To circumvent this issue, we have
employed the use of an alternative network analysis in Vaa 3D software to generate more
reliable information regarding vascular and neuronal networks. Not only have we obtain
more precise information regarding the vascular and neuronal network characteristics, but
have used fluorescently labeled tissues to render complete neuronal and vascular networks
in 3D within intact eyes. Furthermore, we have observed the effects of over vascularization
in the various layers of the eye and how these effects disrupt the normal structure of the

whole eye. In conducting PACT CLARITY in rat eyes with subsequent antibody labeling we
encountered some issues many of which have been address in literature currently
One of the biggest obstacle encountered during the preparation and imaging of the
tissues was implementing antibody labeling in PACT cleared eyes. Due to the dense
distribution of the polymer matrix in the tissues, penetration of certain antibodies into deep
portions of the tissue proved to be difficult and inefficient, to address this issues we used
antibody alternative GSi4B Lectin. Current methods explored in the literature implement
FAB or fragmented antigen binding in which full antibodies are fragmented and reassemble
upon binding to target protein epitopes in the tissue. Another technique used to circumvent
issues with antibody penetration is the use of Llama antibodies which are much smaller than
polyclonal or monoclonal antibodies from other species and as such demonstrate good
penetration in dense tissue samples like those observed in PACT 21. Some speculation
remains regarding the effects of bis-acrylamide on the epitope of proteins of interest in
inhibiting effective fluorescent binding however, this has not been researched extensively.
While many techniques are available for characterizing the onset and progression of
disease, we have found that implementing PACT tissue clearing techniques with C-DSLM
provides a highly comprehensive approach for investigating disease. With the ability to
generate high resolution images of fluorescently labeled intact tissues without altering native
conformation, more reliable analysis can be conducted for investigating protein
concentrations and distributions. In implementing PACT with C-DSLM further information
can be obtained regarding the timing of key events in ROP over the course of infant
perinatal development. To enhance the efficiency of these technique in future work,
adaptions to antibody labeling techniques can utilize FAB or llama antibody proteins for
targeting epitopes of interested. Information obtained regarding important events in ROP
onset can then be used for further identifying and treating ROP during medically significant

time points in preterm infants both with respect to vascular and neuronal development and
also with respect to molecular signaling.

1. Cavallaro, G., Luca, F., Paola, B. & Fabio, M. The pathophysiology of retinopathy of
prematurity: an update of previous and recent knowledge. Acta Ophthalmol. (Copenh.)
92, 2-20 (2014).
2. Terry, T. L. Retrolental Fibroplasia in the Premature Infant: V. Further Studies on
Fibroplastic Overgrowth of the Persistent Tunica Vasculosa Lentis. Trans. Am.
Ophthalmol. Soc. 42, 383-396 (1944).
3. Smith, L. E. H. Pathogenesis of retinopathy of prematurity. Semin. Neonatol. 8, 469-473
4. Montoya, R. V., Clapp, C., Rivera, J. C. & Quiroz-Mercaso, H. intraocular and systemic
levels of vascular endothelial growth factor in advanced cases of retinopathy of
prematurity. Clin. Ophthamology 4, 947-953 (2010).
5. Chen, J. & Smith, L. E. H. Retinopathy of Prematurity. Angiogenesis 10, 133-140
6. Gole, G. et al. The international classification of retinopathy of prematurity revisited.
Arch. Ophthalmol. 123, 991-999 (2005).
7. Kurihara, T. et al. von Hippel-Lindau protein regulates transition from the fetal to the
adult circulatory system in retina. Developmental, 1563-1571 (2010).
8. Linden, T. & Wenger, R. The antimycotic ciclopirox olamine induces HIF-1 alpha
stability, VEGF expression, and angiogenesis. FASEB J. 17, 761-763 (2003).
9. Pierce, E. ., Avery, R. L., Aiello, L. P. & Smith, L. E. H. Vascular endothelial growth
factor.vascular permeability factor expression in a mouse model of retinal
neovascularization. 92, (1995).
10. Bach, L. A. Endothelial cells and the IGF system. J. Mol. Endocrinol. 54, R1-R13
11. Ferrara, N., Hillan, K. J., Gerber, H.-P. & Novotny, W. Discovery and development of
bevacizumab, an anti-VEGF antibody for treating cancer. Nat. Rev. Drug Discov. 3,
391-400 (2004).
12. Jiang, J. et al. Inhibition of retinal neovascularization by gene transfer of small interfering
RNA targeting HIF-1a and VEGF. J. Cell. Physiol. 218, 66-74 (2009).

13. Hellstrom, A. Low IGF-I suppresses VEGF-survival signaling in retinal endothelial cells:
Direct Correlation with clinical retinoapthy of prematurity. 98, 5804-5808 (2001).
14. Lofqvist, C. et al. IGFBP3 suppresses retinopathy through suppression of oxygen-
induced vessel loss and promotion of vascular regrowth. Proc. Natl. Acad. Sci. U. S. A.
104, 10589-10594 (2007).
15. Claybon, A. & Bishop, A. J. R. Dissection of a Mouse Eye for a Whole Mount of the
Retinal Pigment Epithelium. J. Vis. Exp. (2011). doi: 10.3791/2563
16. MacIntyre, N. Unmasking antigens for immunohistochemistry ProQuest. Available at:
origsite=gscholar. (Accessed: 16th March 2016)
17. Fritschy, J.-M. & Hartig, W. in Encyclopedia of Life Sciences (ed. John Wiley & Sons,
Ltd) (John Wiley & Sons, Ltd, 2001).
18. Coene, W. M. & Van Dyck, D. Maximum-likelihood methof for focus-variatkon image
reconstruction in highnresolution transmission electron microscopy. Ultra microscopy 64,
109-135 (1996).
19. Wu, C., Petersen, R. & VanderVeen, D. RetCam Imaging for Retinopathy of Prematurity
Screening. J. AAPOS 10, 107-111 (2006).
20. Fahrbach, F. Rapid 3D light-sheet microscopy with a tunable lens. Opt. Express 21,
21010-21026 (2013).
21. Yang, B. etal. Single-cell phenotyping within transparent intact tissue through whole-
body clearing. Cell 158, 945-958 (2014).
22. Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH image to ImageJ: 25 years of
image analysis: for the past 25 years NIH image and ImageJ software have been
pioneers as open tools for the analysis of scientific images. We discuss the origins,
challenges and solutions of these two programs, and how their history can serve to
advise and inform other software projects. Nat. Methods 9, 671+ (2012).
23. Treweek, J. B. etal. Whole-body tissue stabilization and selective extractions via tissue-
hydrogel hybrids for high-resolution intact circuit mapping and phenotyping. Nat. Protoc.
10, 1860-1896 (2015).

24. Voie, A. H., Burns, D. H. & Spelman, F. A. Orthogonal-plane fluorescence optical
sectioning: Three-dimensional imaging of macroscopic biological specimens. J. Microsc.
170, 229-236 (1993).
25. Keller, P. J., Schmidt, A. D., Wittbrodt, J. & Stelzer, E. H. K. Reconstructions of
Zebrafish Early Embryonic Development by Scanned Light Sheet Microscopy. Science
233, (2008).
26. Chung, K. & Deisseroth, K. Structural and molecular interrogation of intact biological
systems. Nature 0, (2013).
27. Peng, H., Ruan, Z., Long, F., Simpson, J. H. & Myers, E. W. V3D enables real-time 3D
visualization and quantitative analysis of large-scale biological image data sets. Nat.
Biotechnol. 28, 348-353 (2010).
28. Tang, J.-R. et al. Moderate postnatal hyperoxia accelerates lung growth and attenuates
pulmonary hypertension in infant rats after exposure to intra-amniotic endotoxin. Am. J.
Physiol. Lung Cell. Mol. Physiol. 299, L735-748 (2010).

Full Text