University of Southern California Dissertations and Theses

Cell crosstalk in the healthy knee influences remodeling and inflammatory gene expression

(USC Thesis Other)

Cell crosstalk in the healthy knee influences remodeling and inflammatory gene expression

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content RUNNING HEAD: Cell Crosstalk in the Healthy Knee 1

CELL CROSSTALK IN THE HEALTHY KNEE INFLUENCES REMODELING AND
INFLAMMATORY GENE EXPRESSION

By

Bianca N. Lennarz

A Thesis Presented to the
FACULTY OF THE KECK SCHOOL OF MEDICINE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
MASTER OF SCIENCE
MEDICAL PHYSIOLOGY
In the Department of Physiology and Biophysics
December 2017

CELL CROSSTALK IN THE HEALTHY KNEE 2
ACKNOWLEDGMENTS
First, I would like to express my sincere gratitude to my advisor, Dr. Austin Mircheff, not only
for the opportunity to participate in this exciting research, but also for his constant support and
for sharing his knowledge with me. This thesis would not have been possible without his efforts
to guide me. I would also like to thank Dr. Yanru Wang for mentoring me in the laboratory and
for contributing to this project.

My gratitude also extends to my other committee members, Dr. Thomas Vangsness Jr. and Dr.
Harvey Kaslow, for the opportunity, for keeping me on track, and for motivating me.

I would like to thank my mother, my father, and my sister for all of their love and encouragement
throughout my education. I would also like to give special thanks to my boyfriend, for all of his
love, support, and patience throughout this process. I would not be where I am today without
them.

CELL CROSSTALK IN THE HEALTHY KNEE 3
TABLE OF CONTENTS
Acknowledgements ..........................................................................................................................2
Table of Contents .............................................................................................................................3
Abstract ............................................................................................................................................5
Chapter 1: Introduction ....................................................................................................................7
Chapter 2: Methods ........................................................................................................................16
Animals ..............................................................................................................................16
Media and Reagents ...........................................................................................................17
Procedures ..........................................................................................................................18
Ex Vivo Model .......................................................................................................18
Cell Culture and Co-Culture ......................................................................18
Synoviocytes ..................................................................................18
Chondrocytes .................................................................................19
Co-Cultures ....................................................................................20
Imaging ......................................................................................................20
RNA Isolation from Cultured Cells ...........................................................21
In Vivo Model ........................................................................................................23
RNA Isolation from Animal Tissue ...........................................................23
RNA Sequencing Analysis ....................................................................................24
Chapter 3: Results ..........................................................................................................................26
Cell Culture Media – Minimizing Fetal Bovine Serum Content .......................................26
Crosstalk Between Chondrocytes and Synoviocytes in an Ex Vivo Model ......................31
Morphology............................................................................................................31
RNA expression .....................................................................................................31
CELL CROSSTALK IN THE HEALTHY KNEE 4
Influence of Age on Gene Expression in Cartilage, Synovium, and Fat ...........................33
Chapter 4: Discussion ....................................................................................................................35
Effect of Various Media on Growth and Morphology .......................................................35
Crosstalk Between Chondrocytes and Synoviocytes in an Ex Vivo Model ......................36
Influence of Age on Gene Expression in Cartilage, Synovium, and Fat ...........................38
Future Directions ...............................................................................................................39
References ......................................................................................................................................43
Figures............................................................................................................................................46

CELL CROSSTALK IN THE HEALTHY KNEE 5
ABSTRACT
Osteoarthritis is a painful, irreversible inflammatory disease that affects every component of the
joint. The inflammatory cycle remodels the synovia, which is responsible for maintaining the
homeostasis and nutrients of the joint, and degrades the cartilage. It has been shown that
osteoarthritis has a unique phenotype in each individual, even at the same stage of disease, which
is a reason that current treatments may not work to alleviate symptoms or halt progression. This
ex vivo and in vivo study examined the gene expression and relationship between tissue of the
healthy knee joint. We conducted an ex vivo experiment to mimic a healthy knee environment,
in which we studied the influence between synoviocytes and chondrocytes. The in vivo study
examined the cartilage, synovia, and infrapatellar fat of healthy knee joints between young and
old rabbits, to study the gene expression between the tissue and age, to examine if osteoarthritis
had naturally occurred in the tissue through aging in healthy rabbits. Female New Zealand white
rabbits with healthy joints were euthanized prior to collection of tissue during necropsy. For the
ex vivo experiments, chondrocytes and synoviocytes were isolated from cartilage and synovial
membrane, respectively. These cells were grown in culture, then later plated into mono-culture
and co-culture. For the in vivo model, the synovia and the fat pad were extracted and the
cartilage was excised from both knees of 16-week old female virgin rabbits and ≥8-month old
female retired breeder rabbits. The RNA was isolated directly from the mono-culture, co-
culture, and the tissue, and the gene expression of each group of cells was determined by RNA-
sequencing analysis. The data thus far of the ex vivo model indicated that synoviocyte mediators
up-regulated 211 and down-regulated 74 chondrocyte genes, and that chondrocyte mediators up-
regulated 131 and down-regulated 61 synoviocyte genes (P < 0.05, fold change ≥ +/−2). The
data from in vivo model resulted in an up-regulation of 117 genes and a down-regulation of 766
genes between young and old cartilage (FDR-step-up < 0.05, fold change ≥ +/−2); an up-
CELL CROSSTALK IN THE HEALTHY KNEE 6
regulation of 66 genes and a down-regulation of 76 genes between young and old synovia; and
an up-regulation of 80 genes and a down-regulation of 171 genes between young and old fat (P <
0.05, fold change ≥ +/−2). These data suggested that synoviocytes have a significant influence
on the gene expression of chondrocytes, and vice versa, affecting many genes associated with
inflammation, cartilage degradation, and cellular structure. Synoviocyte-derived and
chondrocyte-derived mediators seemed to act independently of immune cells that cause
inflammatory processes. Within the healthy knee in the in vivo model, synovia and fat possessed
a similar gene expression to each other, but both largely differed from cartilage. Distinctive
groups between age were observed in the PCA analysis, which suggested that aging can cause
the change of gene expression associated with inflammation, cartilage phenotype, and cellular
structure. Mapping the gene expression of the tissues in the diseased knee can provide a basis
for designing new, individualized treatments for patients who suffer from osteoarthritis.

CELL CROSSTALK IN THE HEALTHY KNEE 7
CHAPTER 1: INTRODUCTION
The leading causes of disability in developed countries are chronic inflammatory diseases
(Woolf et al., 2003). Osteoarthritis, a chronic inflammatory disease, occurs as the cartilage
between the joints, particularly of the knees, degrades over time from natural usage, persistent
stress, or trauma. Muscle strength and ligament tone change as an individual becomes older,
leading to greater chances of injury and mechanical stress, and therefore potentially causing
osteoarthritis. This degradation leads to limited mobility, stiffness, pain, and inflammation of the
joints (Woolf et al., 2003). Not only does osteoarthritis limit individuals from participating in
physical activity and employment, but it can also lead to other chronic illnesses and conditions
such as diabetes and obesity, which in turn, can result in worsening osteoarthritis (Barbour et al.,
2017).
There are many types of arthritis; osteoarthritis is distinctly different than the others, and
is one of the most common subcategories of the disease. Unlike in rheumatoid arthritis, for
example, osteoarthritis is not due to any autoimmune disease, it usually occurs in older
individuals, and the degradation process is slow. Rheumatoid arthritis is an autoimmune disease,
caused by genetic predisposition to autoantibody formation (McInnes et al., 2011). It can be
triggered by environmental stimuli, can occur at any age, can have a slow or a fast onset, can
affect multiple joints simultaneously, and occurs symmetrically in anatomy (Arthritis
Foundation, n.d.). Osteoarthritis is a degenerative joint disease, and affects every component of
the joint, including the bones, which are composed of the femur and tibia in the case of the knee;
the articular cartilage surrounding the femoral condyles and tibial plateau of the joints; the
meniscal cartilage between the joints, which act as the shock absorbers of the joint; the fat pad
located underneath the patella; and the synovial membrane, or synovium, which is tissue that
encloses the joint, produces lubricating synovial fluid to eliminate friction to aid in movement,
CELL CROSSTALK IN THE HEALTHY KNEE 8
and provides nutrients and a homeostatic milieu for the other tissues in the knee that are
otherwise poorly vascularized.
In the conventional view of osteoarthritis, the “wear and tear” through movement and
pressure on the joint causes an inflammatory response. Cartilage, that have been stressed or
damaged, signal and attract immune cells to the synovium where the inflammation process
initiates. The immune cells cause inflammation of the tissues and release more chemicals that
recruit even more immune cells that proliferate. This, in turn, causes more degradation to the
cartilage and meniscus and causes remodeling of the synovium, which causes a release of signals
to activate pain receptors, thickening of the joint capsule, and a loss of joint space. However, the
reverse can occur too, where inflammation of the synovium can begin independently from initial
cartilage degradation. The synovitis initiates the cascade that leads to degradation of cartilage
and meniscus, which leads to more synovitis. The infrapatellar fat pad has also been considered
to contribute in the inflammation process as well (Clockaerts et al., 2009). Either initiation
process becomes a continuous and accelerated cycle of inflammation, pain, and joint remodeling.
In a healthy knee, chondrocytes and synoviocytes release microparticles (M
P
), that
contain exosomes, which allow the cells to communicate with each other. Exosome-mediated
signaling also occurs in inflamed tissues, and contributes to the immune response cascade.
Immune cells, such as T lymphocytes (T cells), monocytes, monocyte-derived macrophages, and
dendritic cells are found in healthy synovia. These immune cells are also found in osteoarthritic
synovia, but at increased levels. The innate immune response of the inflamed joint is comprised
of anabolic factors, which include anti-inflammatory cytokines and growth factors, and of
catabolic factors, which include pro-inflammatory cytokines and proteolytic enzymes (Orlowsky
et al., 2015). IL-1 and TNFα are pro-inflammatory cytokines that are produced during the
inflammatory process are noted as major contributors to cartilage catabolism in osteoarthritis
CELL CROSSTALK IN THE HEALTHY KNEE 9
(Sokolove et al., 2013). Leukemia inhibitory factor (LIF) is also noted to have a possible pro-
inflammatory link to osteoarthritis (Orlowsky et al., 2015). Adaptive immune responses also
play a role in osteoarthritis. Macrophages use phagocytosis to engulf and digest debris,
pathogens, protein or antigen fragments, and M
P
. From the products, the macrophages separate
epitopes to display to T cells, which carry specific epitope receptors. Generally, T cells
recognize epitopes from the presentation of a professional antigen presenting cell.
Although permanent cures have not been discovered, medical treatment does exist, but is
limited to pain management, physical activity, systemic or local administration of non-steroidal
and steroidal anti-inflammatory drugs, and even joint replacements when treatment alone does
not alleviate pain (Arthritis Foundation, n.d.). The overall disease process does not appear to
consist of only one phenotype and mechanism, but instead of many different disease phenotypes.
Recent research has revealed that the disease process and immune response differ from one
patient to the next, which indicates that treatments should be unique and individualized for each
patient. Our project’s concept of multiple osteoarthritic phenotypes was based on the study by
Vangsness et al (2011), where synovial fluid cytokine concentrations and degree of cartilage
degradation (graded according to the ICRS scale) were determined in patients who underwent
meniscus repair. The data showed a correlation that a higher level of cytokines produced a
greater severity of disease, but also that the variability in signaling molecules was large between
patients (Vangsness et al., 2011). Data from principle component analysis (PCA) indicated that
osteoarthritis has multiple phenotypes. These data showed that knees from different patients
with the same degree of cartilage damage differed in projections between the various principle
components (PCs). This suggested that the molecular composition of not only the patients at
different stages of disease progression, but also within patients in the same disease stage, had
significantly diverse phenotypes (Fig. 1). Even patients with the same degree of cartilage
CELL CROSSTALK IN THE HEALTHY KNEE 10
damage, synovial fluid contained varied signaling molecule concentrations. There was no
control in this study, although the purpose was to clinically study if there was a relationship or
pattern in current osteoarthritic patients. There are a variety of influences that may account for
the diverse phenotypes, including genetic differences, differences in physiology, differences
between gender and age, and differences in how and in what degree the disease was initiated
(Vangsness et al., 2011). The uniqueness of the disease per individual poses significant
obstacles to discovering new medical therapies that may be suited for each disease phenotype.
Current therapies and clinical trials neglect the aspect of diversity of the disease, which therefore
may fail to detect beneficial responses between participant subgroups.
There are many models of disease to conduct osteoarthritic research, such as rats, mice,
guinea pigs, rabbits, and swine, among others. There is no single animal model available that
completely replicates all of the features of human osteoarthritis because of the heterogeneous
nature of the disease (McCoy, 2015). Swine would be ideal for this study, since they resemble
humans more closely not only in terms of size, physiology, and anatomy, but also in respect to
the immune mechanism involved specifically in osteoarthritis (Cruz et al., 2016). To balance
biological relevance and costs, rabbits, specifically New Zealand white rabbits, were chosen to
be used for this project. Natural degeneration occurs very slowly in the joints of rabbits and
other model animals, which closely simulates the joint disease progression in humans (Kuyinu et
al., 2016). Rabbits, though, can experience acute and rapid induced cartilage degradation,
therefore, making them suitable models of osteoarthritis (Bendele, 2001). In addition, the rabbit
knee is large enough to provide material for extensive analyses, in contrast to the knees of
smaller animal models, such as rats or mice.
Osteoarthritis occurs in either sex, although the disease phenotype and disease severity
can vary between male and female due to hormonal and anatomic structural differences (Hame et
CELL CROSSTALK IN THE HEALTHY KNEE 11
al., 2013). Osteoarthritis is also more prevalent in the female sex. For these reasons, we used
female rabbits in this project. Additionally, resources limited us from using both sexes in this
project.
Our project focuses on studying the immune cell network of the chronic inflammatory
disease of osteoarthritis and mimicking the events following immune cell activation within the
joint tissues in order to develop more specific and individualized treatments. This project was
divided into multiple studies, focusing on a study to optimize cell growth in various medium, an
ex vivo model of a healthy knee, and an in vivo model of a healthy knee, and the plan of future
studies of an ex vivo diseased model and an in vivo diseased model. The project focuses on the
difference between a closed system, referring to the ex vivo studies, and an open system,
referring to the in vivo studies. For example, the natural immune response in vivo will differ
from the reaction ex vivo in a diseased environment. Since the disease progresses in vivo, the
body’s natural immune response will have also infiltrated the joint space, whereas this cannot
occur ex vivo. Because of this important difference, both ex vivo and in vivo studies will be
performed, gene expression in all conditions will be examined, and the results will be compared
not only between the ex vivo and in vivo diseased studies to their respective controls, but the
control and diseased state of the ex vivo and in vivo studies will also be compared to each other.
It was hypothesized that synoviocytes and chondrocytes influence each other’s gene
expression under both controlled and diseased environments. Since the in vivo study
investigated the age-related progression of osteoarthritis, the gene expression at various ages
may show changes that suggest slowly-developing inflammatory processes, that may occur
before obvious histopathology has occurred. In the ex vivo study, synoviocytes were predicted
to release mediators that influenced gene expression in chondrocytes, and chondrocytes were
predicted to also release mediators that influence gene expression in synoviocytes. Before being
CELL CROSSTALK IN THE HEALTHY KNEE 12
able to experimentally assess these hypotheses, a culture medium that included the smallest
possible amount of serum but that would also optimize cell growth, had to be found. The co-
culture ex vivo experiment followed, and investigated if synoviocytes influenced chondrocyte
gene expression differently than synoviocytes grown alone, and vice versa with chondrocytes.
The age related in vivo comparisons then investigated if cartilage, synovia, and fat were
influenced depending on the age of the animal. With these two types of control studies on
healthy knees, the combined data will be compared to investigate if the gene expression in mono-
cultured or co-cultured synoviocytes accurately mimicked the gene expression in intact
synovium, as well as if the gene expression in mono-cultured or co-cultured chondrocytes
accurately mimicked the gene expression of intact cartilage.
In order to optimize cell growth, synoviocytes, which are the cells composing synovium,
and chondrocytes, which are the cells composing cartilage, were grown in various media.
Previous experiments, such as the study conducted by Cheng et al (2012), isolated and
maintained synoviocytes effectively in Dulbecco’s modification of Eagle’s medium (DMEM)
medium supplemented with as low as 10% fetal bovine serum (FBS) (Cheng et al., 2012).
Haerdi-Landerer et al (2011) isolated synoviocytes with as low as 10% fetal calf serum (Haerdi-
Landerer et al., 2011). Dulbecco's modified Eagle medium/Ham's Nutrient Mixture F-12
(DMEM/F-12) with 10% FBS, DMEM/F-12 with 10% Nu-serum (containing only 2.5% FBS),
and serum-free human fibroblast medium were used by themselves or combined in various
proportions and at various stages of the culture period to study their ability to support cellular
growth of both chondrocytes and synoviocytes. The goal was to minimize the concentration of
serum in the media and to optimize cell growth, since the presence of serum increases the
variability in the cell culture, and to use the optimized condition for the remainder of the projects
that involved cell cultures of theses cell types.
CELL CROSSTALK IN THE HEALTHY KNEE 13
The ex vivo study was designed to mimic the in vivo growth of cells and crosstalk of cell
communication within a healthy knee joint. The question in study asked, how do healthy
synoviocytes influence healthy chondrocytes, and vice versa? To study this question, the
experiment was designed with co-cultures. In this study, synoviocytes were grown in
Transwell
®
inserts and chondrocytes were grown in the wells of the Transwell
®
plates (Corning,
Inc.). Via these co-cultures, the synoviocytes and chondrocytes were physically separated from
each other by a membrane, but could still communicate with each other through this membrane,
and could theoretically alter each other’s gene expression. Mono-cultured control cells and co-
cultured cells were stained with calcein and imaged under fluorescence microscope to observe
any phenotypic differences. The RNA was isolated directly from the wells of the mono-cultured
and co-cultured cells, and the gene expression of each group of cells was determined by RNA-
sequencing analysis.
The in vivo model not only served as the control of the natural joint environment, but
also examined the possibility that the knee joint environment naturally changed through age.
This in vivo model examined the cartilage, synovia, and patellar fat of healthy knee joints
between young and old rabbits, then studied gene expression differences between the tissues and
between the age groups. The patellar fat pad and the synovium were extracted and the cartilage
was excised from both the right and left knee of a total of twenty rabbits, divided by age of either
young or old. The RNA was isolated directly from the tissue from necropsy, and was sent for
RNA sequencing analysis.
In both studies, RNA was isolated from each tissue sample or cell sample, and was sent
for RNA sequencing analysis. RNA sequencing is a high-throughput sequencing method of
quantifying and mapping transcriptomes and analyzing gene expression, by allowing the
researcher to quantify and profile the RNA that is present in a given sample (Wang et al., 2009).
CELL CROSSTALK IN THE HEALTHY KNEE 14
RT-qPCR is normally used when targeting specific genes; RNA-sequencing analysis is used for
detecting known and unknown genes, and was better suited for our study. Sample RNA is
fragmented and goes through a process, termed library preparation, in which the RNA fragments
are converted to cDNA, adapter sequences are added to the ends, and the cDNA are clonally
amplified (Cresko Lab of the University of Oregon, n.d.). The cDNA is then sequenced using a
sequencing platform, in this case, Illumina TruSeq technology, and the resulting short read
sequences are analyzed. For this project, RNA sequencing analysis allows comparisons between
and within genes of various conditions and cell types. RNA sequencing has become the model
method for analyzing and studying biological questions and disease processes. Using total RNA
measures all of the RNA present, not only including mRNA, but also including RNA types
which are not translated into genes but can influence the genes that are translated from mRNA.
For this reason, total RNA sequencing analysis can become more expensive than mRNA
sequencing analysis, which serves as a limitation. Because this project is focused on the coding
regions of the genome, it was better to use mRNA analysis. Advantages of mRNA sequencing
analysis, versus other methods of gene expression analysis, include high resolution, low
background interference, and ability to detect both novel genes and known genes (Wang et al.,
2009). Genomic sequences that have yet to be determined as well as a wide selection of species
can be studied, and no prior knowledge of the transcriptome is required (Wang et al., 2009).
These studies provide the insight to perform the long-term strategy of disease induction
in vivo, in which the idea will be briefly introduced, since the culmination of the other studies
will ultimately be compared to the in vivo diseased state. Surgical and non-surgical models have
not been shown to accurately mimic human age-related osteoarthritis, since the inflammatory
processes are quickly initiated and developed, can surpass the severity of natural disease
progression, and the inflammatory processes may have resolved spontaneously on their own.
CELL CROSSTALK IN THE HEALTHY KNEE 15
Surgical models to induce disease have included those such as meniscectomies (Moskowitz et
al., 1973); joint immobilizations, destabilizations, and muscle weakness (Egloff et al., 2014); or
anterior cruciate ligament transections (Bluteau et al., 2002). Non-surgical methods have
included those such as inflammatory induction via injections of cartilage degrading enzymes
(Kikuchi et al., 1998); fragments of fibronectin which cause an increased release of proteases
that cause cartilage damage (Homandberg et al., 1993); inflammatory cytokines and foreign
molecules (Ross et al., 2012); high doses of estradiol (Tsai et al., 1993); or metabolic poisons
(Williams et al., 1984).
Because such experimental interventions do not strongly imitate the disease response of
interest, the condition produced by each method may be unresponsive to treatments that would
have otherwise been effective for osteoarthritis. Instead, the experimental method that will be
utilized in the future disease induction of this project will be novel adoptive transfer surgeries,
adapted from a study on lacrimal gland inflammation by Mircheff et al (2016). In this study, an
inflammatory response was induced in lacrimal glands via the activation of ex vivo peripheral
blood lymphocytes that adoptively transferred pathogenic T cells to the autologous rabbit
lacrimal glands (Mircheff et al., 2016). As a similar model, rabbits will be used for the isolation
of bone marrow monocytes, chondrocytes, and synoviocytes. Bone marrow monocytes will be
taken from the rabbit femur to be matured into dendritic cells (mDC), which will later be used to
activate the peripheral blood lymphocytes (PBL) immune cells taken from a second rabbit.
Cartilage will be excised from the knee joints and synovial membrane will be extracted from the
joint space of the knee of the first rabbit. The microparticles (M
P
) produced by these
chondrocytes and synoviocytes contain the auto-antigens that mDC process and present to PBL
in ex vivo mixed cell reactions. Activated T cells will then be used in an adaptive transfer
surgery, where they will be injected into the synovial fluid space of the autologous rabbits’ knee
CELL CROSSTALK IN THE HEALTHY KNEE 16
(the rabbit of which the PBL was taken from) to induce disease. In this final study, we predict
that osteoarthritis will develop rapidly once the chondrocyte M
P
and the synoviocyte M
P
are
introduced. Just as the other studies, the RNA will be isolated from the tissue and will be sent
for RNA sequencing analysis to analyze the gene expression. The significant genes discovered
from the RNA analysis could serve as future therapeutic targets. Once it is used for the diseased
models and in human clinical trials, we can analyze the disease phenotype of each individual,
since different phenotypes will require different therapeutic targets.

CHAPTER 2: METHODS
Animals
Female New Zealand white rabbits were obtained from Irish Farms in Norco, CA, and
from Western Oregon Rabbit Co. in Philomath, OR. All experiments adhered to the Guidelines
for Use of Animals in Research and were performed according to a protocol that was approved
by the University of Southern California (USC) Institutional Animal Care and Use Committee.
All rabbits were housed in the USC Health Sciences Campus Vivarium. The ex vivo study
consisted of 8 rabbits, divided into four 16-week old rabbits (2.0-2.5 kg) used for extracting gene
expression data, and four (3.0-4.0 kg) rabbits used in fluorescence imaging. The in vivo study
consisted of twenty rabbits, divided into two groups: “young” (16-week old rabbits) and “old”
(≥8-month old rabbits). Five 16-week old virgins and three ≥8-month old retired breeders were
from Irish Farms, and four 16-week old virgins and eight ≥8-month old retired breeders were
from Western Oregon Rabbit Co.
Rabbits were euthanized prior to the collection of tissues. Based on weight, animals were
anesthetized with a 1:1 mixture of Ketamine and Xylazine injected intramuscularly with a 27-
CELL CROSSTALK IN THE HEALTHY KNEE 17
gauge needle. Once under anesthesia, the animals were euthanized intravenously with Euthasol,
using a 27-gauge needle.
Media and Reagents
The isolation medium (IM) was composed of DMEM/F12 (1:1) (Cellgro Mediatech,
Inc.), Hepes Buffer Solution (Omega Scientific, Inc.), antibiotics
(Penicillin/Streptomycin/Glutamate) (Cellgro Mediatech, Inc.), L-Proline (40 µg/mL) (Sigma
Aldrich, Inc.), L-Ascorbic acid (25 µg/mL) (Sigma Aldrich, Inc.), and Fungizone (Amphotericin
0.25 µg/mL) (Gibco by Life Technologies Corp.). Digestion medium (DM) for both cell types
contained IM with 0.2% collagenase II (Gibco by Life Technologies Corp.).
Preliminary experiments (described in RESULTS) were performed to identify the
minimum amount of fetal bovine serum (FBS) necessary for preservation of cell viability and
morphology of synoviocytes and chondrocytes, with a partially defined supplement (Nu-Serum,
which contained 10% FBS) being used to replace 100% FBS and comparing those growth results
to those of using serum-free human fibroblast cell culture media (Axol Bioscience, Cambridge,
United Kingdom). On the basis of these experiments, medium containing 90% Dulbecco's
modified Eagle medium/Ham's Nutrient Mixture F-12 (DMEM/F12) (1:1) combined with 10%
Nu-Serum, which contained 2.5% FBS, was used for the remainder of the experiments. As a
result, the synoviocyte complete culture medium (SCCM) was composed of DMEM/F12 (1:1),
Nu serum containing media, antibiotics (Penicillin/Streptomycin/Glutamate), and MEM non-
essential amino acids (Corning Cellgro Mediatech, Inc.). The chondrocyte CCM (CCCM) was
composed of DMEM/F12 (1:1), Nu serum containing media, antibiotics
(Penicillin/Streptomycin/Glutamate), MEM non-essential amino acids, L-Proline (40 µg/mL),
and L-Ascorbic acid (25 µg/mL).
CELL CROSSTALK IN THE HEALTHY KNEE 18
Procedures
Ex Vivo Model
Cell culture and co-culture. Using aseptic techniques, cartilage was excised from both the tibia
and femur of each knee joint with a #11 sterile scalpel blade, and synovial membrane was
extracted from the joint space of each knee of each rabbit using a #10 sterile scalpel blade and
sterile forceps. Dissected synovium and cartilage were placed into separate polypropylene tubes
containing a small volume of IM.
The tissues were cultured according to the protocol and according to the results of the cell
media experimentation. The synovium and cartilage were then placed in separate sterile petri
dishes with a small amount of IM to prevent the tissue from drying out. The tissue was cut into
smaller pieces with separate pairs of scalpel blades. Both tissues types were transferred to
respective 50 mL Erlenmeyer flasks, and the petri dishes were rinsed with small amounts of IM
to ensure that all of the tissue was collected into each flask. Tissue fragments were allowed to
settle for 5-10 minutes, and the supernatant was aspirated and discarded.
SYNOVIOCYTES. A volume of 2-4 mL of trypsin-EDTA 0.25% (Hyclone Laboratories, GE
Healthcare Life Sciences) was pipetted into the flask with the synovium. The flask was capped
and placed into a water bath and shaker for 20 minutes at 37°C. The flask was removed from the
water bath, 5 mL of IM was added, and the tissue was dispersed by trituration using a 5 or 10 mL
plastic pipette and automatic pipettor. The tissue was allowed to settle, and the supernatant was
aspirated. Another 5 mL of IM was added into the flask to repeat this step, then 5 mL of DM
was then added to the flask. The flask was capped and placed back into the water bath and
shaker until the tissue exhibited signs of digestion, a time between 60-100 minutes. At this
point, the flask was removed from the water bath, 15 mL of IM was added, and the cell sediment
CELL CROSSTALK IN THE HEALTHY KNEE 19
was dispersed with a pipette and automatic pipettor. A 70 µm cell strainer was set on a 50 mL
conical tube and was dampened by a small amount of IM. All of the contents of the flask were
pipetted onto the cell strainer. The tube with the strained cells was centrifuged at 600g for 5
minutes and the supernatant was discarded. A small volume of SCCM was added to the cell
pellet. A small volume of 10 µL of synoviocytes were stained with 10 µL trypan blue, pipetted
into a chamber of a slide, and counted via a TC20™ Automated Cell Counter (Bio-Rad
Laboratories, Inc., Hercules, CA). A volume of SCCM was added to the tube to adjust the cell
concentration to 8x10
3
–1x10
4
cells/mL, before distributing the cells into a sterile six well plate.
CHONDROCYTES. Cartilage was placed in a 50 mL flask, allowed to settle, the supernatant
was discarded, and 5-10 mL of DM was added. The flask was capped and placed in a 37°C
humidified incubator with 5% CO
2
for 4 hours, with periodic manual swirling of the flask. After
digestion, 15 mL of IM was added and the cell aggregates were broken by using a pipette and
automatic pipettor. A 100 µm cell strainer was placed on a 50 mL conical tube, dampened with
a small amount of IM, and all of the cell contents of the flask were strained through. These
filtered cells were centrifuged at 500g for 5 minutes and the supernatant was discarded. A small
volume of CCCM was added to the remaining cell pellet, 10 µL of these chondrocytes were
stained with 10 µL trypan blue, pipetted into a chamber of a slide, and the cells were counted via
a TC20™ Automated Cell Counter. The volume of CCCM was increased and adjusted to a cell
concentration of 8x10
3
–1x10
4
cells/mL before being distributed into a sterile six well plate.
Both cell types were incubated in a 37°C humidified incubator with 5% CO
2
while
monitored and observed daily for growth. On day 4, old CCM was pipetted off and replaced
with fresh CCM. After cells grew to confluence, they were split, whether it was to co-culture or
to the next passage. The old medium was pipetted off and discarded, and the wells (or flasks)
CELL CROSSTALK IN THE HEALTHY KNEE 20
were washed with Dulbecco's phosphate-buffered saline (DPBS). The cells were detached from
the bottom of the wells by adding trypsin-EDTA 0.25% and tilting the wells until visible
detachment was observed, then directly after, adding trypsin inhibitor (Sigma Aldrich, Inc.) (at a
1:1 ratio with volume of trypsin-EDTA 0.25%). Equal volumes of F12 were added to dilute each
well and to harvest the cells into separate 15 mL or 50 mL conical tubes per cell type depending
on the total volume from all of the wells or flasks, then centrifuged at 300g for 5 minutes. The
supernatant was discarded, and the pelleted cells were re-suspended in their appropriate culture
media volume to distribute among the wells or flasks of which the cells were being transferred.
It was determined that both chondrocytes and synoviocytes needed to be confluent on
approximately the same day before being re-suspended and placed in co-culture. If the cells
were not simultaneously confluent, the confluent cell type was passaged while waiting on the
other cell type to proliferate to confluence.
CO-CULTURES. Synoviocytes and chondrocytes were separated into control wells and co-
culture wells (Fig. 2). For co-cultures, synoviocytes were grown in the Transwell
®
inserts and
chondrocytes were grown in the wells of the Transwell
®
plates (Corning, Inc.). This served to
represent the interaction between healthy synoviocytes and chondrocytes in the normal, or
control, environment of the healthy knee. For control mono-cultures, both cell types were placed
alone in separate inserts and in separate wells. The plates were then placed in a 37°C humidified
incubator with 5% CO
2
and monitored each day for 5-6 days, while each cell type grew to
confluence. The cells were then processed either for imaging or for isolation of RNA for further
analysis.
Imaging. The LIVE-or-DIE™ Cell Viability/Cytotoxicity Kit (GeneCopoeia, Inc.) was used to
highlight live cells versus dead cells, and provided a clearer view of the shape and concentration
CELL CROSSTALK IN THE HEALTHY KNEE 21
of cells, which was useful in comparing cell types or comparing various time points of cell
growth. Calcein was retained within intact cells, producing a green fluorescence. Propidium
iodide (PI) entered cells with damaged membranes and bound to nucleic acids, producing a red
fluorescence. The wells containing cells were washed with DPBS, and a sufficient amount of
DPBS was added to cover just the bottom of the wells after the last wash. Buffer was distributed
to each well. The two dyes were added together in a solution with DPBS, vortexed, and added to
each well of the controls and co-cultures. The stained constructs were then covered and placed
in a 37°C humidified incubator with 5% CO
2
for 30-45 minutes, then placed under an EVOS
fluorescence microscope (Life Technologies, NY). Transmittance images were captured, as well
as fluorescence images using the fluorescein optical filter (485 ± 10 nm) and rhodamine optical
filter (530 ± 12.5 nm) at various magnifications of 4x, 10x, and 20x.
RNA isolation from cultured cells. RNA was purified from the control cells and co-cultured cells
using the “Isolation of Total RNA from Animal Cells” protocol of the QIAGEN
®
RNeasy
®
Mini
Kit, in combination with DNase digestion using the QIAGEN
®
RNase-Free DNase Set
(QIAGEN, Hilden, Germany). Before beginning the isolation, 10 µL of beta-mercaptoethanol
was added to every 1 mL of Buffer RLT, 70% ethanol was added to Buffer RPE, and the DNase
I stock solution was prepared by mixing the solid DNase I (1500 Kunitz units) in 550 µL of
RNase free water. All surfaces and tools were sprayed with RNaseZap
®
to remove RNase
contamination (AMBION, Inc., Austin, Texas).
In the first step of isolation, the cell culture medium was aspirated, and the cells were
disrupted by the addition of 350 µL Buffer RLT directly into the wells. The cells were scraped
off and transferred into microcentrifuge tubes by micropipette, and then vortexed. Each sample
was homogenized by pipetting the lysate into QIAshredder spin columns placed in 2 mL
CELL CROSSTALK IN THE HEALTHY KNEE 22
collection tubes and centrifuging for 2 minutes at maximum speed. A volume of about 350 µL
of 70% ethanol was added to the homogenized lysate from the collection tube, mixed well via
pipetting, and a total of 700 µL of the sample was transferred to an RNeasy
®
mini column placed
in a 2 mL collection tube. Each sample was centrifuged for 15 seconds at 10,000 rpm, and the
flow-through was discarded. Buffer RW1 (350 µL) was added to each RNeasy
®
mini column,
and the columns were centrifuged for 15 seconds at 10,000 rpm to wash; the flow-through was
discarded. DNase digestion was included into the protocol at this point; for every sample, 10 µL
DNase I stock solution was added to 70 µL of Buffer RDD for each sample, and this stock
solution was mixed gently by inverting. For each sample, 80 µL of the DNase I incubation
mixture was pipetted directly onto the RNeasy
®
silica gel membrane and left to sit on the bench
top for 15 minutes. After this period, Buffer RW1 (350 µL) was added into the RNeasy
®
mini
column, and the column was centrifuged for 15 seconds at 10,000 rpm to wash; the flow-through
was discarded. Each RNeasy
®
column was transferred to a new 2 mL collection tube, and Buffer
RPE (500 µL) was added into the RNeasy
®
columns, and the columns were centrifuged for 15
seconds at 10,000 rpm to wash the column; the flow-through was discarded. Buffer RPE (500
µL) was again added to the RNeasy
®
column and centrifuged for 2 minutes at 10,000 rpm to dry
the silica gel membrane. To elute, the RNeasy
®
column was transferred to a new 1.5 mL
collection tube; 40 µL of RNase-free water was directly added onto the RNeasy
®
silica gel
membrane and centrifuged for 1 minute at 10,000 rpm. The samples were then taken to the
Research Center for Liver Diseases Core Laboratory at USC to perform quality control (QC)
analysis with a Thermo NanoDrop 8000 spectrophotometer with PC control and data acquisition.
After the measurements were completed, the samples were labeled and placed in a -20°C freezer
for storage.
CELL CROSSTALK IN THE HEALTHY KNEE 23
In Vivo Model
RNA isolation from animal tissue. Rabbits were euthanized prior to the collection of
infrapatellar fat pads, synovium, and cartilage using aseptic techniques. The infrapatellar fat
pads were removed by using a #10 sterile scalpel blade and sterile forceps. The synovial
membrane was excised from the joint space of the right and left knee of each rabbit using a #10
sterile scalpel blade and sterile forceps. Cartilage was excised from the plateau of each tibia and
the lateral and medial condyles of each femur using a #11 sterile scalpel blade. Each sample of
tissue was placed in RNAlater
®
RNA Stabilization Reagent in nucleotide-free microcentrifuge
tubes (AMBION, Inc., Austin, Texas).
Before beginning the isolation, 10 µL of beta-mercaptoethanol was added to every 1 mL
of Buffer RLT, 70% ethanol was added to Buffer RPE, and the DNase I stock solution was
prepared by mixing the solid DNase I (1500 Kunitz units) in 550 µL of RNase free water. Each
tissue sample was also weighed; 50-60 mg of each fat sample, 30-40 mg of each synovium
sample, and 30-40 mg of each cartilage sample were separated for RNA isolation.
The RNAlater
®
was removed from each tube, and the tissue was placed into a sterile
closed system micro tissue homogenizer polypropylene tube for homogenization
(Kimble™ BioMasherII™). Buffer RLT (600 µL) was added into each new tube, and the tissue
was homogenized by grinding with a polyacetal pestle. Both the interior of the tube and the
pestle tip were lined with an abrasive surface for easier disruption. The lysates were centrifuged
for 3 minutes at full speed, and the supernatants were removed and transferred to new
microcentrifuge tubes. The following steps, including the addition of 70% ethanol, the protocol
for DNase digestion, the subsequent washes, and the RNA collection, are as described above in
RNA isolation from cultured cells. The samples were then taken to the Research Center for Liver
CELL CROSSTALK IN THE HEALTHY KNEE 24
Diseases Core Laboratory at USC to perform quality control (QC) analysis with a Thermo
NanoDrop 8000 spectrophotometer with PC control and data acquisition. Once the
measurements were completed, the samples were labeled and placed in a -20°C freezer for
storage.
RNA Sequencing Analysis
The isolated RNA extracts from each sample were sent for mRNA sequencing analysis at the
USC Molecular Genomics Core. Libraries were prepared using the Illumina TruSeq RNA
Library Prep Kit, and sequenced using the Illumina NextSeq sequencer at 25 million reads and
read length of 75 base-pair single ends (Illumina, Inc., San Diego, CA).
With guidance from the USC Norris Medical Library Bioinformatics Service, the raw
reads were input onto Partek Flow
®
software (Partek Inc., St. Louis, MO). Here, the data went
through pre-alignment quality control, which separated higher quality data with a score (QC) ≥
20. The data was then trimmed, which removed low quality bases that could have a negative
impact on the alignment results. Next, the data was aligned using STAR 2.4, which is the aligner
that works best for RNA sequencing analysis and is the fastest, most sensitive, and most precise
aligner (Dobin et al., 2013), where the reads were then aligned to reference sequences. The
aligned reads then went through post-alignment quality control analysis (QA/QC) and quantified
based on expression abundance of gene counts. This filtered data was then normalized to ensure
that the data was comparable.
At this step, it was possible to generate and view the principle component analysis (PCA)
graph of each project. PCA was utilized to visualize high dimension data in one graph; as a
visual, rather than a numerical representation. Each principle component (PC) generated a
percentage to show the contribution each PC made to the variability of the data. Loadings that
were generated by PCA represented the contribution that each variable made to each PC;
CELL CROSSTALK IN THE HEALTHY KNEE 25
projections generated by PCA quantitatively represented the variation of phenotype. Statistical
analysis to detect differential gene expression (GSA) was performed for the ex vivo study, and
comparisons were selected by the user to compare between conditions (control versus co-culture
of each cell type, and between cell type controls). Statistical analysis to detect differential
expression (GSA) and ANOVA were performed for the in vivo study, and comparisons were
selected by the user to compare tissue types of different ages, and between tissue types,
respectively. The differentially-expressed genes from each condition were separated and could
be visualized in volcano plots. The list of differentially-expressed genes taken from Partek
Flow
®
were divided into up-regulated and down-regulated genes. The differentially-expressed
genes were determined with a significance criterion of FDR-step-up < 0.05 and fold change ≥
+/−2, or if little to no genes matched this level of significance, P < 0.05 and fold change ≥ +/−2
was used as the significance criteria. The FDR (false discovery rate) is more stringent than the P
value, has greater power, and eliminates multiple testing bias, which is why the FDR was
considered as the first choice before considering the P value when determining which genes were
significant in these studies. The P value was chosen instead, only when using FDR did not result
in any differentially expressed genes.
The differentially-expressed genes were submitted to Ingenuity Pathway Analysis (IPA)
(QIAGEN, Redwood City, CA), to determine the biological and molecular functions and disease
relevance; the gene descriptions and functions were also found from IPA, Ensembl, or NIH
DAVID (Huang et al., 2009). The normalized data were also inputted into Partek Genomics
Suite
®
to further deconstruct the PCA data to generate the loadings, projections, and eigenvalues
of each PC. SPSS (v. 24) was used to run ANOVA tests across groups, with a priori α = 0.05,
and Fisher’s least significant difference was used for post hoc tests.

CELL CROSSTALK IN THE HEALTHY KNEE 26
CHAPTER 3: RESULTS
Cell Culture Media – Minimizing Fetal Bovine Serum Content
The purpose of this series of experiments was to minimize the concentration of serum in
the media for both synoviocytes and chondrocytes, since the presence of serum could contribute
to the variability of the results in cell culture. In the first trial introducing serum-free (SF) media
to chondrocytes, the cells were split into their third passage and distributed into a six well plate.
Each well contained a different proportion of media: 100% DMEM/F-12 Nu-Serum
®
containing
media (composed of 90% DMEM/F-12 and 10% Nu-Serum
®
(containing 2.5% FBS)); 100% SF
media; 50% DMEM/F-12 Nu-Serum
®
containing media (composed of 45% DMEM/F-12 and
5% Nu-Serum
®
(containing 1.25% FBS)) with 50% SF media; 25% DMEM/F-12 Nu-Serum
®

containing media (composed of 22.5% DMEM/F-12 and 2.5% Nu-Serum
®
(containing 0.625%
FBS)) with 75% SF media; and 75% DMEM/F-12 Nu-Serum
®
containing media (composed of
67.5% DMEM/F-12 and 7.5% Nu-Serum
®
(containing 1.875% FBS)) with 25% SF media (Fig.
3a). Because the wells that contained Nu-Serum
®
were confluent by day 6, all of the cells except
those in 100% SF media were split to passage 4, where another six well plate was used
containing the same proportions as its previous passage. The cells from each well of passage 4
and the 100% SF well of passage 3 were monitored and observed.
In this first trial using chondrocytes during passage 3, ascending order of Nu-Serum
®

containing media was associated with increased cell growth and confluence – the well with
100% DMEM/F-12 Nu-Serum
®
containing media contained the cells with most rapid
confluence, followed by the proportion of 75% DMEM/F-12 Nu-Serum
®
containing media, 50%
DMEM/F-12 Nu-Serum
®
containing media, 25% DMEM/F-12 Nu-Serum
®
containing media,
and finally 100% SF media. The cells in complete 100% SF media did not grow to confluence,
and the cells in 25% DMEM/F-12 Nu-Serum
®
containing media with 75% SF media grew
CELL CROSSTALK IN THE HEALTHY KNEE 27
almost confluent but at an extremely slow pace compared to those in 100% DMEM/F-12 Nu-
Serum
®
containing media. Then, chondrocytes were passaged, and the cells in Nu-Serum
®

containing media were able to survive and proliferate; the cells in the higher concentration of
Nu-Serum
®
grew at a faster rate than cells in lower concentration, while the cells in 100% SF of
both passages 3 and 4 were monitored and never grew to confluence since the cell growth halted
and the cells died.
SF media was introduced to synoviocytes when splitting cells to passage 2. The medium
was distributed among the cells of a six well plate: 100% DMEM/F-12 Nu-Serum
®
containing
media, 100% SF media, 50% DMEM/F-12 Nu-Serum
®
containing media with 50% SF media,
25% DMEM/F-12 Nu-Serum
®
containing media with 75% SF media, and 75% DMEM/F-12 Nu-
Serum
®
containing media with 25% SF media (Fig. 3b). Each well except for the cells in 100%
SF media were passaged on day 4 of their progress into the same proportions and monitored in
this later stage of cell growth. On day 3 of the passage, each well was split to passage 4, but into
different proportions to test the lowest functional ratio of SF media to Nu-Serum
®
containing
media: 100% DMEM/F-12 Nu-Serum
®
containing media and 100% SF media as controls; 25%
DMEM/F-12 Nu-Serum
®
containing media with 75% SF media; 20% DMEM/F-12 Nu-Serum
®

containing media (composed of 18% DMEM/F-12 and 2% Nu-Serum
®
(containing 0.5% FBS))
with 80% SF media; 12.5% DMEM/F-12 Nu-Serum
®
containing media (composed of 11.25%
DMEM/F-12 and 1.25% Nu-Serum
®
(containing 0.313% FBS)) with 87.5% SF media; and 10%
DMEM/F-12 Nu-Serum
®
containing media (composed of 9% DMEM/F-12 and 1% Nu-Serum
®

(containing 0.25% FBS)) with 90% SF media. Cells were monitored for their growth.
In the first trial using synoviocytes during passage 2, the cells cultured in a higher
concentration of SF media grew at a much slower rate than cells cultured in any concentration of
Nu-Serum
®
containing media. Similar to chondrocytes, synoviocytes grown in 100% DMEM/F-
CELL CROSSTALK IN THE HEALTHY KNEE 28
12 Nu-Serum
®
containing media grew to confluence the fastest, and those grown in 100% SF
media were unable to grow to confluence nor be passaged. In the third passage, the same results
were observed; synoviocytes in 100% SF media did not grow to confluence, while those in 100%
DMEM/F-12 Nu-Serum
®
containing media grew to confluence the fastest, followed by cells
equally in 75% DMEM/F-12 Nu-Serum
®
containing media, 50% DMEM/F-12 Nu-Serum
®

containing media, and 25% DMEM/F-12 Nu-Serum
®
containing media, and lastly, the newest
plated 100% SF media. In the fourth passage, the cells were visually monitored, and 20%
DMEM/F-12 Nu-Serum
®
containing media was observed to be the lowest proportion that could
be used, since, below that, at a concentration of 0% DMEM/F-12 Nu-Serum
®
containing media,
cells did not grow to confluence and grew at an extremely slow rate.
Next, primary cells were cultured in 100% DMEM/F-12 Nu-Serum
®
containing media,
then passaged once confluent to separate six well plates, using media containing reduced
proportions of Nu-Serum
®
: 100% DMEM/F-12 Nu-Serum
®
containing media; 100% SF media;
50% DMEM/F-12 Nu-Serum
®
containing media with 50% SF media; 75% DMEM/F-12 Nu-
Serum
®
containing media with 25% SF media; and 25% DMEM/F-12 Nu-Serum
®
containing
media with 75% SF media. Both cell types were split into the second passage one week later
when all of the cells were confluent except for those in 100% SF media. Cells were separated
into six well plates in 100% DMEM/F-12 Nu-Serum
®
containing media, 100% SF media, 50%
DMEM/F-12 Nu-Serum
®
containing media with 50% SF media, 20% DMEM/F-12 Nu-Serum
®

containing media with 80% SF media, 12.5% DMEM/F-12 Nu-Serum
®
containing media with
87.5% SF media, and 10% DMEM/F-12 Nu-Serum
®
containing media with 90% SF media, and
cells in 20% DMEM/F-12 Nu-Serum
®
containing media with 80% SF media were passaged into
two T25 flasks per cell type for observation. Synoviocytes and chondrocytes were split into the
CELL CROSSTALK IN THE HEALTHY KNEE 29
third passage in T75 flasks and six well plates and were distributed into the same ratios as they
were in the second passage (Fig. 3c).
It was observed that the primary cells that were split to their first passage in a six well
plate as well as in their second passage in a six well plate, grew best in highly concentrated Nu
serum containing media. When the cells were split to their third passage, containing both cell
types in six well plates and T25 flasks, the synoviocytes grew confluent much more rapidly in
Nu-Serum
®
containing media than the chondrocytes also in Nu-Serum
®
containing media. Both
cell types were then split into two T75 flasks per cell type in passage 4. Results were consistent
in that both cell types did not grow well in flasks with a higher proportion of SF media versus in
wells with a higher proportion of Nu-Serum
®
containing media.
Next, instead of using cells that had already established growth, primary cells of
chondrocytes and synoviocytes were grown in various proportions to find the optimized media
for unestablished cells. This time, SF media was introduced directly into the primary culture.
The chondrocytes and synoviocytes were cultured into separate six well plates, containing
various media proportions in each occupied well: 20% DMEM/F-12 Nu-Serum
®
containing
media with 80% SF media; 25% DMEM/F-12 Nu-Serum
®
containing media with 75% SF
media; and 30% DMEM/F-12 Nu-Serum
®
containing media (composed of 27% DMEM/F-12
and 3% Nu-Serum
®
(containing 0.75% FBS)) with 70% SF media (Fig. 3d). Both cell types
were observed for growth and morphological changes.
When the primary cells were directly cultured using the experimental proportions, 20%
DMEM/F-12 Nu-Serum
®
containing media with 80% SF media was observed to be the lowest
limit at that point of Nu-Serum
®
utilization for both cell types in terms of cell growth, but not for
phenotype. Synoviocytes were observed to grow to confluence slowly, and the chondrocytes
grew very little without reaching to confluence and detached from the plate completely. Lastly,
CELL CROSSTALK IN THE HEALTHY KNEE 30
when 10% DMEM/F-12 Nu-Serum
®
containing media was used in culture, it was then deemed
as the lowest concentration that successfully grew both cell types. The cells did not grow as
quickly as in 100% DMEM/F-12 Nu-Serum
®
containing media, although both cell types grew at
the rate equivalent to 20% DMEM/F-12 Nu-Serum
®
containing media. Any concentration lower
than 10% was not deemed fit for cell growth.
Lastly, since both cell types were successfully grown in the lowest Nu-Serum
®

containing media in the experimental proportions, both cell types were then cultured and
observed beginning from primary growth directly from the animal, in a lower dosage of 10%
DMEM/F-12 Nu-Serum
®
containing media. Both cell types were observed for growth and
morphological changes.
In 100% DMEM/F-12 Nu-Serum
®
containing media, the phenotype of synoviocytes
appeared to be elongated and organized in one-layer side by side (Fig. 4); the phenotype of
chondrocytes appeared small, circular, and grown side by side (Fig. 5). The phenotypes of these
two cell types were similar to those observed in previous ex vivo studies (Fig. 6a-b). Any
amount of SF media appeared to have changed the morphology of the cells; synoviocytes grew
in layers on top of each other and had a fibroblastic appearance stretching to attach to each other
(Fig. 7); the morphology of the chondrocytes transformed from being small, circular cells into
elongated, fibroblastic cells (Fig. 8).
As a result of this experiment, the proportion of 100% DMEM/F-12 Nu-Serum
®

containing media – media composed of 90% DMEM/F-12 and 10% Nu-Serum
®
(2.5% final FBS
content) – was utilized in all subsequent experiments that included cell culture of synoviocytes
and chondrocytes. Although the cells successfully grew with smaller Nu-Serum
®
proportions,
both cells drastically changed morphologically because of the proportion of SF media that was
CELL CROSSTALK IN THE HEALTHY KNEE 31
also included. The resulting media that we subsequently used represents a 75% decrease in
serum from the lowest FBS content found in the literature.
Crosstalk Between Chondrocytes and Synoviocytes in an Ex Vivo Model
The purpose of these experiments was to study the crosstalk of synoviocytes and
chondrocytes in a simulated construction of a healthy joint, by investigating how synoviocyte-
derived mediators influenced phenotype and gene expression of chondrocytes, and vice versa.
Morphology
The florescence imaging presented a change in phenotype between mono-cultured and
co-cultured chondrocytes, but no significant observable phenotypic change between mono-
cultured and co-cultured synoviocytes (Figs. 9a-b and 10a-b). Chondrocytes gained a slightly
elongated, larger shape grown in co-culture, compared to the smaller, rounded cells grown in
mono-culture (Figs. 11a-d and 12a-b). The mono-cultured chondrocytes were consistent in
phenotype to normal chondrocytes grown ex vivo in previous studies (Fig. 6). However, healthy
chondrocytes exhibit different phenotypes depending on whether the cells are in vivo or grown
ex vivo. In vivo, small groups of chondrocytes are grouped into separated lacunae, i.e., cavities
in the matrix; this separation did not occur ex vivo in mono-culture. The chondrocytes in vivo
then undergo hypertrophy and eventually calcify (Fig. 13). However, the phenotype of the
chondrocytes in co-culture, appears similar to the phenotype of the chondrocytes in vivo. Both
co-cultured and in vivo chondrocytes resemble each other, in that both tend toward a longer and
larger cell shape, as shown when chondrocytes develop in vivo. Mono-cultured chondrocytes
did not adopt this appearance, even after being plated for the same duration of time as the co-
cultured chondrocytes.
RNA Expression
CELL CROSSTALK IN THE HEALTHY KNEE 32
Synoviocyte mediators significantly (P < 0.05 and fold change ≥ +/−2) affected the gene
expression of 285 chondrocyte genes (Fig. 14a). These mediators influenced the down-
regulation of 74 chondrocyte genes, including the expression of a collagen degrading enzyme,
MMP12. Synoviocyte mediators influenced the up-regulation of 211 chondrocyte genes, of
which at least 10 genes contributed to cellular morphology, including genes associated with cell-
extracellular matrix adhesion, cell-cell adhesion, cytoskeletal function, and collagen assembly.
They also up-regulated the expression of ADAMTS4, a more abundant collagen degrading
enzyme, and up-regulated at least 9 chondrocyte genes responsible for inflammatory responses,
such as CCL5, CXCL13, and VCAM-1 (Fig. 14b).
Chondrocyte-derived mediators significantly (P < 0.05 and fold change ≥ +/−2) caused
the up-regulation of 131 synoviocyte genes and the down-regulation of 61 synoviocyte genes, for
a total of 192 synoviocyte genes (Fig. 15a). Many genes involved in extracellular matrix and
collagen assembly were influenced, as well as genes involved in morphology and the formation
of the cytoskeleton. The chondrocyte mediators altered the expression of collagen degradation
genes, such as the up-regulation of ADAMTS8 and ADAMTS15, but the down-regulation of
ADAMTS3 and ADAMTSL1, as well as the up-regulation of the collagen degradation inhibitor,
TIMP1. Many genes involved in inflammatory responses were also up-regulated; these included
the chemokines, CXCL10 and CXCL13, and the cytokines, LIF and BAFF (Fig. 15b). LIF was
previously noted as a gene associated with osteoarthritic inflammation (Orlowsky et al., 2015).
Principle component analysis (PCA) was performed on both synoviocyte and
chondrocyte data, as a means of visual representation to interpret the impact of change that one
cell type had on the other, in both controlled and conditioned settings. PCA data showed that
both cell types, as well as their respective conditioned and control groups, are separated (Fig.
16). PC1 and PC2 accounted for most of the variability in the abundance data. In PC1, the
CELL CROSSTALK IN THE HEALTHY KNEE 33
synoviocytes, regardless of being conditioned or control samples, are grouped together in terms
of projection value, and are separated from the chondrocytes. In PC2, the conditioned cells are
distinctly separated from their control counterparts. Loadings (L) of the PCA data showed that a
total of 7568 genes (|0.5| ≤ L) were strongly affected by PC1, and 2271 genes (|0.3| ≤ L < |0.5|)
were moderately affected by PC1, which accounted for 39.4% of the total variability. 2659
genes (|0.5| ≤ L) were strongly affected by PC2, and 2999 genes (|0.3| ≤ L < |0.5|) were
moderately affected by PC2, which accounted for 14.1% of the total variability. 2023
genes (|0.5| ≤ L) were strongly affected by PC3, and 3031 genes (|0.3| ≤ L < 0.5) were
moderately affected by PC3, which accounted for 11.6% of the total variability.
Influence of Age on Gene Expression in Cartilage, Synovium, and Fat
The goal of this experiment was to study influences that age had on gene expression in
cartilage, synovium, and infrapatellar fat. The data from knees of five “young” rabbits and three
“old” rabbits have been received and analyzed; samples from other rabbit studies are currently in
RNA-sequencing processing. Of the available results, between tissue types, volcano plots
showed that cartilage had more significant genes that differed from synovia and fat, though
synovia and fat had very few significant genes that differed from each other (P < 0.05 and fold
change ≥ +/−2). Cartilage differed from synovia with respect to 6500 genes, 2362 of which were
expressed more highly in cartilage, and 4138 of which were expressed more highly in synovia.
Cartilage differed from fat with respect to 6332 genes, 4168 of which were more highly
expressed in cartilage, and 2164 of which were more expressed in fat (FDR-step-up < 0.05 and
fold change ≥ +/−2). Synovia differed in only 272 genes compared to fat, 131 of which were
more highly expressed in synovia, and 141 of which were more highly expressed in fat (P < 0.05
and fold change ≥ +/−2).
CELL CROSSTALK IN THE HEALTHY KNEE 34
Between age, volcano plots showed that aging resulted in the change of expression of
more genes in cartilage, which resulted in a higher difference in up and down regulated genes in
distribution along with more significant genes, than synovia or fat. Between young and old
cartilage, aging was associated with an up-regulation of 117 genes and a down-regulation of 766
genes (FDR-step-up < 0.05 and fold change ≥ +/−2) (Fig. 17a). Of the significantly changed
cartilage genes, notable genes included those involved with collagen synthesis and those
involved in the formation and maintenance of cartilage and bone were down-regulated; BMP1,
which codes for a protein that induces cartilage and bone formation; CRTAP, which codes for a
scaffolding protein involved in bone; and GDF6, which codes for a protein that is involved in
bone differentiation. Inflammatory genes such as CCDC3, PROC, and RABGEF1, which
negatively regulate the pro-inflammatory response, were down-regulated, and CCRL2, CD40,
PIK3CG, and PYCARD, which are involved in activation of the adaptive or innate immune
response, were up-regulated (Fig. 17b). Between young and old synovia tissue, aging was
associated with 66 genes that were up-regulated, and 76 genes that were down-regulated (P <
0.05 and fold change ≥ +/−2) (Fig. 18a). Of the significantly changed synovium genes, many
were involved in inflammation, such as GABRB2, IRF4, and TMIGD3, and involved in pain
sensation, such as KCNK18, which were up-regulated (Fig. 18b). Between young and old fat
tissue, aging was associated with an up-regulation of 80 genes, and a down-regulation of 171
genes (P < 0.05 and fold change ≥ +/−2) (Fig. 19a). Of the significantly changed fat genes,
notable down-regulated genes included NOV, which is responsible for enhancing articular
chondrocyte phenotype and is important for skeletal structure, and HTRA1, which is involved in
bone mineralization and cytoskeletal framework; notable up-regulated genes included those
involved in inflammatory processes, such as CD69 and ICAM5 (Fig. 19b). Synovia and fat
shared many genes, including novel protein coding genes, novel small nuclear RNA (snRNA)
CELL CROSSTALK IN THE HEALTHY KNEE 35
genes, down-regulation of collagen genes such as COL1A2 and COL3A1, down-regulation of
many genes associated with regulation of the cell cycle, and up-regulation of inflammatory genes
such as CCR3 and IRF4, and bone demineralization genes such as SMPD3.
Loadings from the PCA data showed that a total of 8184 genes (|0.5| ≤ L) were strongly
affected by PC1, and 2991 genes (|0.3| ≤ L < |0.5|) were moderately affected by PC1, which
accounted for 34.1% of the total variability; 2864 genes (|0.5| ≤ L) were strongly affected
by PC2, and 3905 genes (|0.3| ≤ L < |0.5|) were moderately affected by PC2, which accounted for
13.6% of the total variability; and 1215 genes (|0.5| ≤ L) were strongly affected by PC3, and
3692 genes (|0.3| ≤ L < |0.5|) were moderately affected by PC3, which accounted for 8.54% of
the total variability. Projections between tissue types from the PCA data showed that cartilage is
clearly distinct from synovia and fat with respect to PC1-2, PC4-5, and PC14-23. Fat and
synovia were clearly distinct from each other with respect to PC2, PC4-6, PC10-11, PC14, and
PC23. Overall, cartilage was very distinct in gene expression than from synovium and fat, and
synovium and fat were both very similar in gene expression, as shown in the overall PCA plot
(Fig. 20). Age-related differences appeared in fat with respect to PC1, PC2, PC 3, and PC 4;
age-related differences appeared in synovia with respect to PC1 and PC3; age-related differences
also appeared in cartilage with respect to PC1 and PC3 (Fig. 21a-c).

CHAPTER 4: DISCUSSION
Effect of Various Media on Growth and Morphology in Cell Culture
The rate at which cells grew to confluence depended on the amount of Nu-Serum
®

containing media they were cultured in, regardless of cell type. Nu-Serum
®
contained the lowest
concentration of serum that would both successfully sustain cells and did not alter their
phenotype. Serum-free media caused both cell types to grow more slowly, differentiate, and
CELL CROSSTALK IN THE HEALTHY KNEE 36
change phenotypes. Not only did the synoviocytes differentiate, but when splitting the cells,
they also had a thick appearance and adhered into one mass instead of separating, which made it
difficult to distribute equally amongst the wells and flasks. The morphology of the chondrocytes
developed a similar fibroblastic appearance as the synoviocytes, which gave the prediction that
since SF media is tailored to culture fibroblasts, then cells grown at any concentration of this
media would differentiate and have a phenotype that resembled fibroblasts. It was determined
that it was best to use 10% Nu-Serum
®
containing media, that contained 2.5% FBS, to culture
both cell types.
Crosstalk Between Chondrocytes and Synoviocytes in an Ex Vivo Model
The observable difference in chondrocyte morphology ex vivo shows that synoviocyte
mediators influenced chondrocytes to assume a larger and more elongated phenotype, but there
was no observable difference in synoviocyte morphology ex vivo. The similarity between the
morphology of chondrocytes in co-culture compared to the morphology of chondrocytes in vivo
also supports the hypothesis that the co-culture closely simulates the in vivo environment where
synovia interacts with cartilage. These observations support the hypothesis that synovia can
initiate remodeling in cartilage. The two cell types release mediators to the synovial fluid, and
they are, therefore, presumed to interact and communicate with each other. It may not only be
the mediators that influence each cell type in vivo, but possibly also neurotransmitters and
factors from immune cells and from the blood. However, once they are placed into separate
cultures, both cell types will not conserve the interactions that influenced their phenotypes and
their gene expression programs. As the cells continue to proliferate apart from each other in
culture, they very well may present with different phenotype and gene expression than when they
are interacting. Once these cells are in contact again, as in co-culture, both cell types will be able
to influence each other independently of other factors that were present in the intact knee.
CELL CROSSTALK IN THE HEALTHY KNEE 37
The RNA expression data thus far suggest that the interaction between synoviocytes and
chondrocytes ex vivo act independently of the immune cell mediators that cause inflammation in
the tissue in vivo. The pro-inflammatory state can be initiated and accelerated due to the
crosstalk between cell types in the ex vivo system. Immune cells in vivo are in contact with
synoviocytes, which also communicate via the synovial fluid to the chondrocytes, which
suggests that synovia can initiate the inflammatory response as the source of osteoarthritic
symptoms. The RNA expression data indicate genes of interest that potentially contribute to the
development of chronic inflammatory disease in the joint. The crosstalk between synoviocytes
and chondrocytes up-regulates many pro-inflammatory genes. The PCA data indicates that the
cell types and whether they are conditioned or control are all different from each other,
determined by their separation on the PCA plot, which suggests that the co-cultured cells were
influenced by mediators of the opposite cell type, which conclusively develop a different gene
expression from their mono-cultured counterpart. The conventional view is that synoviocytes
control the milieu of the joint by releasing soluble mediators that influence chondrocytes; but
based on the findings, chondrocytes also affect synoviocyte gene expression. The data showed
that both chondrocytes or synoviocytes develop a gene expression that could induce the
pathogenesis of osteoarthritis when in contact with the other, which suggests that both tissues
induce disease due to the crosstalk that they share. Chondrocyte mediators influenced many
synoviocyte genes involved in collagen and extracellular matrix synthesis and degradation, as
well as up-regulation of genes associated with the immune response. This suggests that
synoviocyte collagen and extracellular matrix synthesis genes influence synovia remodeling,
while degradation genes influence both synovia extracellular matrix and cartilage. Synoviocyte
mediators influenced many chondrocyte genes involved in cell morphology and the up-
regulation of many genes associated with the immune response, which suggests that
CELL CROSSTALK IN THE HEALTHY KNEE 38
chondrocytes are also responsible in the remodeling of synovia and for the inflammatory
response in the tissue.
Influence of Age on Gene Expression in Cartilage, Synovium, and Fat
Within the healthy knee, there were greater differences in the gene expression between
cartilage and synovium and between cartilage and fat. Synovium and fat which possessed a
similar gene expression to each other, and both are different from cartilage. These data showed
that synovium and fat of the joint are remarkably similar in gene expression. Many differentially
expressed genes that were up-regulated are related to promoting inflammatory responses, and
genes that were down-regulated were related to differentiation in chondrocytes or adipocytes,
adipogenesis, and chondrogenesis, which may cause proliferation arrest or prevent repair in the
case of damage. Other down-regulated genes were associated with morphological changes in
connective tissue and cartilage, as well as extracellular matrix and collagen assembly and
degradation. These results suggest that, even in the healthy knee, remodeling takes place during
aging, and this remodeling affects all of the tissues.
The relationship between PC1 and PC2 of fat, PC1 and PC3 of synovium, and PC1 and
PC3 of cartilage suggest age-related differences since clear groups are formed in the data
between younger and older animals. Even within the natural separation of age groups in the
tissue-specific PCA plots, the tissues are phenotypically diverse. Similar to the results from
Vangsness et al (2011) where molecular phenotype was diverse between patients, our results
confirm that tissue from different individuals can also have various tissue phenotype in terms of
gene expression, regardless of age or inflammatory status. These data also suggest that gene
expression changes occur with age, even in healthy animals, leading toward a natural
inflammatory phenotype as aging occurs.
CELL CROSSTALK IN THE HEALTHY KNEE 39
Future Directions
There are many paths that can be taken from this point, providing a very abundant and
promising future of research. Currently, 6-month old rabbits are also being used as an in vivo
control as a middle-aged model, to compare with the “young” and the “old” rabbits. This group
of 6-month old rabbits will help to demonstrate the timeline and progression of natural
osteoarthritis as rabbits age.
A next step would be in pursuing an ex vivo osteoarthritic model. The goal of this study
will be to mimic the environment of tissue inflammation, such as what occurs during
osteoarthritis, by constructing a mixed cell reaction, but in a co-culture. A mixed cell reaction
would be in a Transwell
®
insert, while synoviocytes or chondrocytes would be in the well. This
mixed cell reaction would be created with dendritic cells that have been primed with M
p
, a
preparation that contains exosomes, and PBL. The RNA of both chondrocytes and synoviocytes
could then be isolated and analyzed using RNA-sequencing analysis. This will examine the
crosstalk that occurs between synoviocytes and chondrocytes, including their gene expression
and any morphological changes that could give move insight on what changes occur in the
inflammatory process. This would mimic the inflammatory process of the knee as a closed
system model, to be later compared to the in vivo, open system, model.
As briefly described in the INTRODUCTION, after collecting and analyzing all of the
control data, we will proceed with the in vivo diseased model to compare the control versus the
diseased phenotypes. The diversity of osteoarthritic phenotypes and the complexity of each
phenotype prevent current treatments from being effective; therefore, various phenotypes would
be generated and characterized in an in vivo disease induced model, which would be compared
to the non-diseased induced rabbits, in order to separate genes that have changed between them
as therapeutic gene targets, of which could then be translated into clinical practice. Essentially,
CELL CROSSTALK IN THE HEALTHY KNEE 40
two rabbits will be required for one disease induction. Each “M
p
/DC Source Rabbit” will be
euthanized prior to the collection of tissues. As before, cartilage will be collected from both the
lateral and medial condyles of the femur as well as from the tibia of the rabbit, and synovial
membrane will be extracted from the joint space. These chondrocytes and synoviocytes will be
maintained in optimized cell culture medium (10% Nu serum containing 2.5% fetal bovine
serum) per cell type until grown to confluence through each passage. The supernatant culture
media will be centrifuged to concentrate the M
P
, which the cells will have released to their
supernatant culture media. Bone marrow cells have already been isolated from the femur bone
marrow of ten “M
p
/DC Source Rabbits” and have been cryopreserved, will later be matured into
dendritic cells. The dendritic cells will naturally engulf the M
P
and will be activated by adding
lipopolysaccharide (LPS). The activated dendritic cells will then be X-irradiated in order to stop
further proliferation and then placed into ex vivo mixed cell reactions with peripheral blood
lymphocytes (PBL) taken from the arterial blood of each “Study Rabbit”. [
3
H]-thymidine will be
added to the medium to measure PBL proliferation. These activated T cells will then be injected
into the knees of each “Study Rabbit” to induce disease by adoptive transfer. In the adoptive
transfer surgery, a total volume of 100 µL of saline mixed with activated PBL (10
6
cells) will be
injected into the joint space. Once the disease is introduced to the animal, the rabbit will be
monitored for osteoarthritic processes as symptoms develop over time, and later euthanized for
endpoint necropsy. In order to eliminate variability, the M
P
and monocytes from each “M
p
/DC
Source Rabbit” will be pooled together, so that the pooled M
P
are presented to the pooled mDC,
which creates a constant of M
P
and mDC that will be distributed amongst all trials of T cells
from each “Study Rabbit” in order to keep a consistent repertoire of epitopes presented to every
PBL preparation; in this way, only the “Study Rabbit” will differ for every trial.
CELL CROSSTALK IN THE HEALTHY KNEE 41
After isolating the RNA and sending the samples for RNA sequencing analysis of the
tissues from the necropsies, we can then compare and contrast the gene expression between
healthy rabbits and osteoarthritic induced rabbits, both ex vivo and in vivo. The gene expression
in healthy and conditioned young rabbit cartilage and synovium will be compared with the gene
expression in mono-cultured chondrocytes and synoviocytes and co-cultured chondrocytes and
synoviocytes. These comparisons would determine if the in vivo profiles of both cell types are
similar to the co-culture studies. After gathering all over- and under-expressed genes, it would
be appropriate to investigate why the level of expression of notable significant genes have
changed and if these changes were due to molecular signaling factors. The diseased models,
which will study the mixed cell reaction ex vivo as well as study the inflamed tissues in vivo,
will help to investigate if the soluble mediators or the microparticles during cellular crosstalk are
responsible for the changes in gene expression. The mRNA expression profile of the mixed cell
reaction ex vivo and of the induced inflammation of tissues in vivo will identify candidate
mediators. Proteomic methods will then confirm which mediators are actually expressed. The
influence of each expressed mediator will then individually be removed either with neutralizing
antibodies or with receptor antagonists, to study whether each mediator carries signals to the
synovia, whether each carries signals to the cartilage, or whether each drives the mixed cell
reaction or the inflammatory process.
Future studies may include conducting the same ex vivo and in vivo experiments, but
also using meniscal cartilage, to understand its normal and immune activated role. Meniscal
cartilage acts as a shock absorber and load bearer, and has also been shown to degrade due to
cartilage damage of the femoral condyle and tibial plateau, as well as increase the severity of
disease if a meniscectomy is performed (Hunter et al., 2006). In addition, not only could all of
the ex vivo and in vivo studies be performed on differences in age, but the variable of hormone
CELL CROSSTALK IN THE HEALTHY KNEE 42
levels in the female rabbits can also be considered in the in vivo studies. Women in menopause
have higher chances of developing not only osteoarthritis, but also higher chances in developing
osteoporosis, because of estrogen deficiency (Bay-Jensen et al., 2012). Hormone levels also may
also have an effect on virgin females versus females which have given birth in the past (such as
the “young”, virgin, adult rabbits versus the “old”, retired breeders used in the in vivo study).
Differences in sex, by also using male rabbits, can also be examined in the future.
A possibility in the future is to consider exploring supplements that indicate improvement
in cartilage health, and to determine not only if they function as actual treatments, but also
investigate which genes these drugs up- or down-regulate. Overall, the combination of studies
under this project contribute to the larger aim of studying human biopsies, comparing and
contrasting the phenotypes between the animal model and human model, and, ultimately,
creating new diagnostics and individualized treatments for patients suffering from unique disease
phenotypes of osteoarthritis.

CELL CROSSTALK IN THE HEALTHY KNEE 43
REFERENCES
Arthritis Foundation. (n.d.). Rheumatoid Arthritis. Retrieved from
http://www.arthritis.org/about-arthritis/types/rheumatoid-arthritis

Barbour, K. E., Helmick, C. G., Boring, M., and Brady, T. J. (2017). “Vital signs: Prevalence of
doctor-diagnosed arthritis and arthritis-attributable activity limitation — United States, 2013–
2015.” MMWR Morb Mortal Wkly Rep, 66, 246–253.

Bay-Jensen, A. C., Slagboom, E., Chen-An, P., Alexandersen, P., Qvist, P., Christiansen, C.,
Meulenbelt, I., and Karsdal, M. A. (2012). “Role of hormones in cartilage and joint metabolism:
understanding an unhealthy metabolic phenotype in osteoarthritis.” Menopause: The Journal of
The North American Menopause Society, 20(5), 578–586.

Bendele, A. M. (2001). “Animal models of osteoarthritis.” J Musculoskelet Neuronal Interact,
1(4), 363–376.

Bluteau, G., Gouttenoire, J., Conrozier, T., Mathieu, P., Vignon, E., Richard, M., Herbage, D.,
and Mallein-Gerin, F. (2002). “Differential gene expression analysis in a rabbit model of
osteoarthritis induced by anterior cruciate ligament (ACL) section.” Biorheology, 39, 247–258.

Cheng, M., Johnson, V. M., and Murray, M. M. (2012). “Effects of age and platelet-rich plasma
on ACL cell viability and collagen gene expression.” Journal of orthopaedic research: official
publication of the Orthopaedic Research Society, 30(1), 79–85.

Clockaerts, S., Bastiaansen-Jenniskens, Y. M., Runhaar, J., Van Osch, G. J., Van Offel, J. F.,
Verhaar, J. A., De Clerck, L. S., and Somville, J. (2010). “The infrapatellar fat pad should be
considered as an active osteoarthritic joint tissue: a narrative review.” Osteoarthritis Cartilage,
18(7), 876–882.

Cresko Lab of the University of Oregon. (n.d.). RNA-seqlopedia. Retrieved from
https://rnaseq.uoregon.edu

Cruz, R., Ramírez, C., Rojas, O. I., Casas-Mejía, O., Kouri, J. B., and Vega-López, M. A.
(2016). “The pig as an osteoarthritis translational research model.” J Ost Arth, 1(103).

Dobin, A., Davis, C. A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut, P., Chaisson,
M., and Gingeras, T. R. (2013). “STAR: ultrafast universal RNA-seq aligner.” Bioinformatics,
29, 15–21.

Egloff, C., Sawatsky, A., Leonard, T., Hart, D. A., Valderrabano, V., and Herzog, W. (2014).
“Effect of muscle weakness and joint inflammation on the onset and progression of osteoarthritis
in the rabbit knee.” Osteoarthritis Cartilage, 22, 1886–1893.

Firestein, G. S. (1996). “Invasive fibroblast-like synoviocytes in rheumatoid arthritis.” Arthritis
& Rheumatism, 39(11), 1781–1790.

CELL CROSSTALK IN THE HEALTHY KNEE 44
Haerdi-Landerer, C. M., Steiner, A., and Suter, M. M. (2011). “Primary bovine synoviocyte
cultures: A useful tool for in vitro drug testing?” The Veterinary Journal, 188(1), 58–63.
Hame, S. L., and Alexander, R. A. (2013). “Knee Osteoarthritis in Women.” Current Reviews in
Musculoskeletal Medicine, 6(2), 182–187.

Homandberg, G. A., Meyers, R., and Williams, J. M. (1993). “Intraarticular injection of
fibronectin fragments causes severe depletion of cartilage proteoglycans in vivo.” J Rheumatol,
20, 1378–1382.

Huang, W., Sherman, B. T., and Lempicki, R. A. (2009). “Systematic and integrative analysis of
large gene lists using DAVID bioinformatics resources.” Nat Protoc, 4, 44–57.

Hunter, D. J., Zhang, Y. Q., Niu, J. B., Tu, X., Amin, S., Clancy, M., Guermazi, A., Grigorian,
M., Gale, D., and Felson, D. T. (2006). “The association of meniscal pathologic changes with
cartilage loss in symptomatic knee osteoarthritis.” Arthritis & Rheumatism, 54(3), 795–801.

Kikuchi, T., Sakuta, T., and Yamaguchi, T. (1998). “Intra-articular injection of collagenase
induces experimental osteoarthritis in mature rabbits.” Osteoarthritis Cartilage, 6, 177–186.

Kuyinu, E. L., Narayanan, G., Nair, L. S., and Laurencin, C. T. (2016). “Animal models of
osteoarthritis: classification, update, and measurement of outcomes.” Journal of Orthopaedic
Surgery and Research, 11, 19.

McCoy, A. M. (2015). “Animal models of osteoarthritis: Comparisons and key considerations.”
Vet Pathol, 52, 803–818.

McInnes, I. B., and Schett, G. (2011). “The pathogenesis of rheumatoid arthritis.” N Eng J Med,
365(23), 2205–2219.

Mircheff, A. K., Wang, Y., Schechter, J. E., Meng, L., Tong, W., Attar, M., Chengalvala, M.,
Harmuth, J., and Prusakiewicz, J. J. (2016). “Multiple natural and experimental inflammatory
rabbit lacrimal gland phenotypes.” Ocul Surf, 14, 460–483.

Moskowitz, R.W., Davis, W., Sammarco, J., Martens, M., Baker, J., Mayor, M., Burstein, A. H.,
and Frankel, V. H. (1973). “Experimentally induced degenerative joint lesions following partial
meniscectomy in the rabbit.” Arthritis Rheum, 16, 397–405.

Orlowsky, E. W., and Kraus, V. B. (2015). “The role of innate immunity in osteoarthritis: When
our first line of defense goes on the offensive.” The Journal of Rheumatology, 42(3), 363–371.

Ross, T. N., Kisiday, J. D., Hess, T., and McIlwraith, C. W. (2012). “Evaluation of the
inflammatory response in experimentally induced synovitis in the horse: a comparison of
recombinant equine interleukin 1 beta and lipopolysaccharide.” Osteoarthritis Cartilage, 20,
1583–1590.

Sokolove, J., and Lepus, C. M. (2013). “Role of inflammation in the pathogenesis of
osteoarthritis: Latest findings and interpretations.” Therapeutic Advances in Musculoskeletal
Disease, 5(2), 77–94.
CELL CROSSTALK IN THE HEALTHY KNEE 45

Takizawa, P. (n.d.). Histology at Yale. Retrieved from
http://medcell.med.yale.edu/histology/histology.php

Thirion, S., and Berenbaum, F. (2004). “Culture and phenotyping of chondrocytes in primary
culture.” Methods in Molecular Medicine: Cartilage and Osteoarthritis Volume 1, 100, 1–15.

Tsai, C.L., and Liu, T. K. (1993). “Estradiol-induced knee osteoarthrosis in ovariectomized
rabbits.” Clin Orthop Relat Res, 291, 295–302.

Vangsness, C.T. Jr., Burke, W. S., Narvy, S. J., MacPhee, R. D., and Fedenko, A. N. (2011).
“Human knee synovial fluid cytokines correlated with grade of knee osteoarthritis – a pilot
study.” Bull NYU Hosp Jt Dis, 69(2), 122–127.

Wang, Z., Gerstein, M., and Snyder, M. (2009). “RNA-Seq: A revolutionary tool for
transcriptomics.” Nature reviews: Genetics, 10(1), 57–63.

Williams, J. M., and Brandt, K. D. (1984). “Immobilization ameliorates chemically-induced
articular cartilage damage.” Arthritis Rheum, 27, 208–216.

Woolf, A. D., and Pfleger, B. (2003). “Burden of major musculoskeletal conditions.” Bulletin of
the World Health Organization, 81(9), 646–656.

CELL CROSSTALK IN THE HEALTHY KNEE 46
FIGURES

Figure 1. Principle component analysis (PCA) generated from data from the Vangsness et al
(2011) study. The relationship between level of cartilage damage (ICRS scale) and molecular
phenotype is very diverse, even within patients of the same ICRS grade.

CELL CROSSTALK IN THE HEALTHY KNEE 47

Figure 2. Co-culture set-up for the ex vivo model. Cell types were plated individually in mono-
culture as controls, as well as together in co-culture. The co-culture was designed to mimic the
in vivo growth of cells and crosstalk of cell communication within a healthy knee joint. Via co-
cultures, synoviocytes and chondrocytes are physically separated from each other by a
membrane, but can still communicate with each other through the membrane.

CELL CROSSTALK IN THE HEALTHY KNEE 48

Figure 3a-d. Visualizations of the cell culture design. The media type in each well has
been abbreviated. Each well contained a different proportion of media: 100% DMEM/F-12 Nu-
Serum
®
containing media (90% DMEM/F-12, 10% Nu-Serum
®
(2.5% FBS)); 100% SF media;
50% DMEM/F-12 Nu-Serum
®
containing media (45% DMEM/F-12, 5% Nu-Serum
®
(1.25%
FBS)) with 50% SF media; 25% DMEM/F-12 Nu-Serum
®
containing media (22.5% DMEM/F-
12, 2.5% Nu-Serum
®
(0.625% FBS)) with 75% SF media; and 75% DMEM/F-12 Nu-Serum
®

containing media (67.5% DMEM/F-12, 7.5% Nu-Serum
®
(1.875% FBS)) with 25% SF media.
3a. In the first trial introducing serum-free (SF) media, chondrocytes were distributed into a six
well plate and split. 3b. SF media was introduced to synoviocytes when splitting cells to passage
2 into a six well plate. 3c. Both cell types were both split into T25 then T75 flasks, concurrently
with six well plates. 3d. Both primary cell types were cultured into separate six well plates: 20%
DMEM/F-12 Nu-Serum
®
containing media (18% DMEM/F-12, 2% Nu-Serum
®
(0.5% FBS))
with 80% SF media; 25% DMEM/F-12 Nu-Serum
®
containing media (22.5% DMEM/F-12,
2.5% Nu-Serum
®
(0.625% FBS)) with 75% SF media; and 30% DMEM/F-12 Nu-Serum
®

containing media (27% DMEM/F-12, 3% Nu-Serum
®
(0.75% FBS)) with 70% SF media.
A C
D
B
CELL CROSSTALK IN THE HEALTHY KNEE 49

Figure 4. Normal phenotype of synoviocytes grown in 100% DMEM/F-12 Nu-Serum
®

containing media (containing 10% Nu-Serum
®
and 2.5% FBS) shown with 4x, 10x, and 20x
magnification. The cells appeared to grow side by side in one layer, had an elongated
phenotype, and grew to confluence, covering the entire plate.

20x

4x

10x

CELL CROSSTALK IN THE HEALTHY KNEE 50

Figure 5. Primary chondrocytes in 100% DMEM/F-12 Nu-Serum
®
containing media (composed
of 10% Nu-Serum
®
, which contained 2.5% FBS) shown with 4x, 10x, and 20x magnification.
Even straight from the body into cell culture, the chondrocytes grew healthy in this media
composition. The cells grew side by side, were small in size, and displayed a phenotype that was
mostly circular in shape.

10x

20x

4x

CELL CROSSTALK IN THE HEALTHY KNEE 51

Figure 6a-b. Normal phenotypes of synoviocytes and chondrocytes. 6a. Normal phenotype of
cultured synoviocytes (Firestein, 1996). 6b. Normal phenotype of cultured chondrocytes
(Thirion et al., 2004).

A
CELL CROSSTALK IN THE HEALTHY KNEE 52

Figure 7. Primary synoviocytes in 20% DMEM/F-12 Nu-Serum
®
containing media with 80%
SF-human fibroblast media (composed of 2% Nu-Serum
®
, which contained 0.5% FBS) shown
with 4x, 10x, and 20x magnification. Observable differences are evident between synoviocytes
grown without SF media (Fig. 4) versus synoviocytes grown with SF media. The cells grew
layered upon one another, and gained a fibroblastic phenotype.

20x

4x

10x

CELL CROSSTALK IN THE HEALTHY KNEE 53

Figure 8. Primary chondrocytes grown in 20% DMEM/F-12 Nu-Serum
®
containing media with
80% SF-human fibroblast media (composed of 2% Nu-Serum
®
, which contained 0.5% FBS)
shown with 4x, 10x, and 20x magnification. Chondrocytes grew at a slow rate, halted growth
before reaching confluence, and gained a fibroblastic phenotype.

20x

4x

10x

CELL CROSSTALK IN THE HEALTHY KNEE 54

Figure 9. Transmittance image and overlaid fluorescence image (10x magnification) of normal
mono-cultured synovoicytes on Day 5 of a trial in the ex vivo experiment. Note: Transmittance
images have contrast altered for purposes of better viewing the morphology of the cells.

Figure 10. Transmittance image and overlaid fluorescence image (10x magnification) of co-
cultured synovoicytes on Day 5 of the same rabbit as the control, in the ex vivo experiment. In
this co-culture, synoviocytes were grown in the insert, and chondrocytes were grown in the well.
Phenotypically, synoviocytes did not seem to differ from its control; although, gene expression
of the cells had changed, it did not appear visually. Note: Transmittance images have contrast
altered for purposes of better viewing the morphology of the cells.
CELL CROSSTALK IN THE HEALTHY KNEE 55

Figure 11a-b. Transmittance image and overlaid fluorescence image of mono-cultured
chondrocytes in the ex vivo experiment. 11a. Day 1 of the mono-culture (4x magnification).
11b. Cells retained a simliar phenotype by Day 5 in mono-culture (10x magnification). Note:
Transmittance images have contrast altered for purposes of better viewing the morphology of the
cells.

A
B
CELL CROSSTALK IN THE HEALTHY KNEE 56

Figure 12. Transmittance image and fluorescence image (10x magnification) of co-cultured
chondrocytes on Day 5 of the same rabbit as the control, in ex vivo experiment. The phenotype
of the chondrocytes had elongated and grown considerably larger compared to the
simultaneously grown control counterpart. Note: Transmittance images have contrast altered for
purposes of better viewing the morphology of the cells.

CELL CROSSTALK IN THE HEALTHY KNEE 57

Figure 13. In vivo phenotypes of chondrocytes. This histology image shows the stages of
chondrocyte development beginning from resting cells (top) into the ossification into bone
(bottom) in vivo. Chondrocytes in vivo are separated into lacunae, i.e., cavities in the matrix.
They eventually undergo hypertrophy and then calcify (Takizawa, n.d.).

CELL CROSSTALK IN THE HEALTHY KNEE 58

A
B
CELL CROSSTALK IN THE HEALTHY KNEE 59
Figure 14a-b. The volcano plot and table display the differnetially expressed genes that have
been up-regulated and down-regulated. 14a. This volcano plot shows chondrocyte genes
influenced by synoviocyte-derived mediators that changed between mono-culture and co-culture
(P < 0.05 and fold change ≥ +/−2). This image was generated using Partek Flow
®
software.
14b. This table provides the notable differentially-expressed chondrocyte genes and their
relevant function.

CELL CROSSTALK IN THE HEALTHY KNEE 60

A
B
CELL CROSSTALK IN THE HEALTHY KNEE 61
Figure 15a-b. The volcano plot and table display the differnetially expressed genes that have
been up-regulated and down-regulated. 15a. This volcano plot shows synoviocyte genes
influenced by chondrocyte-derived mediators that changed between mono-culture and co-culture.
This image was generated using Partek Flow
®
software. 15b. This table provides the notable
differentially-expressed synoviocyte genes and their relevant function.

CELL CROSSTALK IN THE HEALTHY KNEE 62

Figure 16. Principle component analysis (PCA) plot of mono-cultured and co-cultured
synoviocytes and chondrocytes. Projections were plotted with respect to PC1, PC2, and PC3, of
which contained the greatest variablity in the data. Each data point on the PCA plot represents
all of the data per each sample. The mono-cultured (control) synoviocytes were grouped
separately from the co-cultured (conditioned) synoviocytes, which both were clearly grouped
separately from the both the mono-cultured and co-cultured chondrocytes. If the three-
dimensional plot was rotated, there would also be a distinct separation between mono-cultured
(control) chondrocytes and co-cultured (conditioned) chondrocytes. This image was generated
using Partek Flow
®
software.

CELL CROSSTALK IN THE HEALTHY KNEE 63

A
CELL CROSSTALK IN THE HEALTHY KNEE 64

Figure 17a-b. The volcano plot and table display the differentially expressed cartilage genes that
have been up-regulated and down-regulated by age. 17a. This volcano plot shows the number of
cartilage genes influenced by age (FDR-step-up < 0.05 and fold change ≥ +/−2). This image was
generated using Partek Flow
®
software. 17b. This table provides the notable differentially-
expressed cartilage genes and their relevant function.
B
CELL CROSSTALK IN THE HEALTHY KNEE 65

A
B
CELL CROSSTALK IN THE HEALTHY KNEE 66
Figure 18a-b. The volcano plot and table display the differnetially expressed synovia genes that
have been up-regulated and down-regulated by age. 18a. This volcano plot shows the number of
synovia genes influenced by age (P < 0.05 and fold change ≥ +/−2). This image was generated
using Partek Flow
®
software. 18b. This table provides the notable differentially-expressed
synovia genes and their relevant function.

CELL CROSSTALK IN THE HEALTHY KNEE 67

A
B
CELL CROSSTALK IN THE HEALTHY KNEE 68
Figure 19a-b. The volcano plot and table display the differnetially expressed fat genes that have
been up-regulated and down-regulated by age. 19a. This volcano plot shows the number of fat
genes influenced by age (P < 0.05 and fold change ≥ +/−2). This image was generated using
Partek Flow
®
software. 19b. This table provides the notable differentially-expressed fat genes
and their relevant function.

CELL CROSSTALK IN THE HEALTHY KNEE 69

Figure 20. Principle component analysis (PCA) plot of young and old cartilage, synovium, and
fat. Projections were plotted with respect to PC1, PC2, and PC3, of which contained the greatest
variablity in the data. Each data point on the PCA plot represents all the data per each sample in
a three-dimensional space. The plot suggests that synovium and fat are very similar in gene
expression, since the data points overlap one another and are intermingled without a definitive
separation in grouping. Cartilage data is drasitcally different from the other two tissues.
Because separation in age is difficult to view in this plot, PCA was run on each tissue
individually, as seen in Figure 21a-c. This image was generated using Partek Flow
®
software.

CELL CROSSTALK IN THE HEALTHY KNEE 70

Figure 21a-c. Principle component analysis (PCA) plot of cartilage, synovium, and fat from five
young, virgin rabbits and three old, retired breeders. The principle components (PCs) that
contained the greatest loading proportion of differentially-expressed genes per each tissue type,
were the chosen PCs plotted for comparison. Each data point on the PCA plot represents all of
the data per each sample in a two-dimensional space. Each plot contained natural separations
that grouped the data by age; even within those groups, the tissues were phenotypically diverse.
21a. Projections from cartilage were plotted with respect to PC1 and PC3. 21b. Projections from
synovium were plotted with respect to PC1 and PC3. 21c. Projections from fat were plotted with
respect to PC1 and PC2.
R8
R9
R10
R11
R12
R13
R14
R15
-60
-40
-20
0
20
40
60
80
-200 -100 0 100 200
PC3
PC1
Cartilage
Young
Old
R8
R9
R10
R11
R12
R13
R14
R15
-80
-60
-40
-20
0
20
40
60
-200 -100 0 100
PC3
PC1
Synovium
Young
Old
R8
R9
R10
R11
R12
R13
R14
R15
-80
-60
-40
-20
0
20
40
60
80
-300 -200 -100 0 100
PC2
PC1
Fat
Young
Old
A B
C

Abstract (if available)

Abstract Osteoarthritis is a painful, irreversible inflammatory disease that affects every component of the joint. The inflammatory cycle remodels the synovia, which is responsible for maintaining the homeostasis and nutrients of the joint, and degrades the cartilage. It has been shown that osteoarthritis has a unique phenotype in each individual, even at the same stage of disease, which is a reason that current treatments may not work to alleviate symptoms or halt progression. This ex vivo and in vivo study examined the gene expression and relationship between tissue of the healthy knee joint. We conducted an ex vivo experiment to mimic a healthy knee environment, in which we studied the influence between synoviocytes and chondrocytes. The in vivo study examined the cartilage, synovia, and infrapatellar fat of healthy knee joints between young and old rabbits, to study the gene expression between the tissue and age, to examine if osteoarthritis had naturally occurred in the tissue through aging in healthy rabbits. Female New Zealand white rabbits with healthy joints were euthanized prior to collection of tissue during necropsy. For the ex vivo experiments, chondrocytes and synoviocytes were isolated from cartilage and synovial membrane, respectively. These cells were grown in culture, then later plated into mono-culture and co-culture. For the in vivo model, the synovia and the fat pad were extracted and the cartilage was excised from both knees of 16-week old female virgin rabbits and ≥8-month old female retired breeder rabbits. The RNA was isolated directly from the mono-culture, co-culture, and the tissue, and the gene expression of each group of cells was determined by RNA-sequencing analysis. The data thus far of the ex vivo model indicated that synoviocyte mediators up-regulated 211 and down-regulated 74 chondrocyte genes, and that chondrocyte mediators up-regulated 131 and down-regulated 61 synoviocyte genes (P < 0.05, fold change ≥ +/−2). The data from in vivo model resulted in an up-regulation of 117 genes and a down-regulation of 766 genes between young and old cartilage (FDR-step-up < 0.05, fold change ≥ +/−2)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

SLIT3 gene expression by 1,25(OH)₂D₃ in an endometriosis stromal cell line

PDF

Fibroblastic connective tissue cells: the blastema stem cells and source of large-scale chondrogenesis in the regenerating lizard tail

PDF

A novel role for hypoxia-inducible factor-1alpha (HIF-1alpha) in the regulation of inflammatory chemokines and leukotriene expression in sickle cell disease

PDF

Cell and gene therapy in the murine model of adenosine deaminase deficiency

PDF

Silencing of expression of the DRIP-80 gene correlates with aberrations in calcium ion distribution in nickel compound- and 3-methylcholanthrene-transformed C3H/10T1/2 mouse embryo fibroblast cel...

PDF

The calcium-sensing receptor in the specification of normal and malignant hematopoietic cell localization in the bone marrow microenvironment

PDF

Linking air pollution to integrative gene and metabolites networks in young adult with asthma

Asset Metadata

Creator Lennarz, Bianca N. (author)

Core Title Cell crosstalk in the healthy knee influences remodeling and inflammatory gene expression

School Keck School of Medicine

Degree Master of Science

Degree Program Medical Physiology

Publication Date 08/28/2017

Defense Date 08/17/2017

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

Tag cartilage,cell crosstalk,cellular crosstalk,chondrocytes,gene expression,Inflammation,inflammatory diseases,knee,knee joint,OAI-PMH Harvest,osteoarthritis,synoviocytes,synovium

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Mircheff, Austin (committee chair), Kaslow, Harvey (committee member), Vangsness, Thomas Jr. (committee member)

Creator Email biancalennarz@gmail.com,blennarz@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-424802

Unique identifier UC11264376

Identifier etd-LennarzBia-5693.pdf (filename),usctheses-c40-424802 (legacy record id)

Legacy Identifier etd-LennarzBia-5693.pdf

Dmrecord 424802

Document Type Thesis

Rights Lennarz, Bianca N.

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