<%BANNER%>

Gene-Induced Chondrogenesis of Mesenchymal Stem Cells through Viral Gene Delivery

Permanent Link: http://ufdc.ufl.edu/UFE0041956/00001

Material Information

Title: Gene-Induced Chondrogenesis of Mesenchymal Stem Cells through Viral Gene Delivery
Physical Description: 1 online resource (126 p.)
Language: english
Creator: Bush, Marsha
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: aav, adenovirus, bmp, cartilage, cell, chondrogenesis, gene, ihh, lentivirus, mesenchymal, sox9, stem, therapy
Biochemistry and Molecular Biology (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Articular cartilage is a highly specialized tissue that allows for near frictionless motion of diarthrodial joints. When cartilage is damaged as a result of injury or disease, the body?s innate capacity for repair is often insufficient due to the low cellular density, avascular and aneural nature of cartilage. Typically the repair tissue is a fibrotic scar which lacks the structural properties of native cartilage. Adult mesenchymal stem cells (MSCs) are well-suited for cell-based therapies for cartilage repair since they are readily available from many tissue sources and capable of differentiating along multiple mesenchymal lineages. Gene transfer to MSCs is a viable method for achieving sustained local expression of specific protein factors, which has been shown to induce chondrogenesis in vitro and may enhance chondrogenic differentiation in vivo. Previous experiments have shown that delivery of the cDNAs for TGF-? and BMP-2 induces chondrogenesis of bovine MSCs in high density aggregate culture. In these studies, we expanded the analyses to include additional cDNAs whose protein products are associated with chondrogenic differentiation during development, including bone morphogenetic proteins BMP-4, BMP-7, developmental morphogen Indian hedgehog (Ihh), transcription factor Sox9, and connective tissue growth factor (CTGF). We transduced early passage bovine MSCs with adenoviral vectors carrying the complete cDNAs for the candidate transgenes at doses ranging from 10 virus particles/cell to 100,000 vp/cell. Chondrogenesis was evaluated by gross examination of aggregate morphology, toluidine blue staining for proteoglycan expression and collagen types I, II, and X immunohistochemistry. The greatest biological responses for each transgene were observed in the dose range of 100-1000 vp/cell; higher viral doses appeared to inhibit chondrogenesis, while lower viral doses were insufficient to yield a pronounced effect. We found that gene transfer of the cDNAs for BMP-4, BMP-7, Ihh, and Sox9 induced chondrogenesis of bovine MSCs while CTGF was not chondrogenic. BMP-4 supports chondrogenesis as effectively as other members of the TGF-beta superfamily such as BMP-2. In further experiments to elucidate the most suitable gene delivery system for animal models and clinical applications, we evaluated vector systems currently available: adeno-associated virus (AAV), lentivirus, and non-viral transfection relative to adenovirus. Simple transfection alone was unable to generate the levels of expression necessary to promote chondrogenesis in our system. Both adenovirus and AAV transduction of MSCs resulted in robust BMP-4 expression over the course of 21 days, complete with chondrogenic differentiation and the expression of cartilage matrix proteins. Lentivirus, which offers the potential for long-term gene expression, was ill-suited for our application because it required long-term culture selection of cells, which altered cellular morphology and negatively impacted cell survival. These data indicate that scAAV serotype 2 delivery of BMP-4 promotes chondrogenesis of MSCs.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Marsha Bush.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Ghivizzani, Steven.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041956:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041956/00001

Material Information

Title: Gene-Induced Chondrogenesis of Mesenchymal Stem Cells through Viral Gene Delivery
Physical Description: 1 online resource (126 p.)
Language: english
Creator: Bush, Marsha
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: aav, adenovirus, bmp, cartilage, cell, chondrogenesis, gene, ihh, lentivirus, mesenchymal, sox9, stem, therapy
Biochemistry and Molecular Biology (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Articular cartilage is a highly specialized tissue that allows for near frictionless motion of diarthrodial joints. When cartilage is damaged as a result of injury or disease, the body?s innate capacity for repair is often insufficient due to the low cellular density, avascular and aneural nature of cartilage. Typically the repair tissue is a fibrotic scar which lacks the structural properties of native cartilage. Adult mesenchymal stem cells (MSCs) are well-suited for cell-based therapies for cartilage repair since they are readily available from many tissue sources and capable of differentiating along multiple mesenchymal lineages. Gene transfer to MSCs is a viable method for achieving sustained local expression of specific protein factors, which has been shown to induce chondrogenesis in vitro and may enhance chondrogenic differentiation in vivo. Previous experiments have shown that delivery of the cDNAs for TGF-? and BMP-2 induces chondrogenesis of bovine MSCs in high density aggregate culture. In these studies, we expanded the analyses to include additional cDNAs whose protein products are associated with chondrogenic differentiation during development, including bone morphogenetic proteins BMP-4, BMP-7, developmental morphogen Indian hedgehog (Ihh), transcription factor Sox9, and connective tissue growth factor (CTGF). We transduced early passage bovine MSCs with adenoviral vectors carrying the complete cDNAs for the candidate transgenes at doses ranging from 10 virus particles/cell to 100,000 vp/cell. Chondrogenesis was evaluated by gross examination of aggregate morphology, toluidine blue staining for proteoglycan expression and collagen types I, II, and X immunohistochemistry. The greatest biological responses for each transgene were observed in the dose range of 100-1000 vp/cell; higher viral doses appeared to inhibit chondrogenesis, while lower viral doses were insufficient to yield a pronounced effect. We found that gene transfer of the cDNAs for BMP-4, BMP-7, Ihh, and Sox9 induced chondrogenesis of bovine MSCs while CTGF was not chondrogenic. BMP-4 supports chondrogenesis as effectively as other members of the TGF-beta superfamily such as BMP-2. In further experiments to elucidate the most suitable gene delivery system for animal models and clinical applications, we evaluated vector systems currently available: adeno-associated virus (AAV), lentivirus, and non-viral transfection relative to adenovirus. Simple transfection alone was unable to generate the levels of expression necessary to promote chondrogenesis in our system. Both adenovirus and AAV transduction of MSCs resulted in robust BMP-4 expression over the course of 21 days, complete with chondrogenic differentiation and the expression of cartilage matrix proteins. Lentivirus, which offers the potential for long-term gene expression, was ill-suited for our application because it required long-term culture selection of cells, which altered cellular morphology and negatively impacted cell survival. These data indicate that scAAV serotype 2 delivery of BMP-4 promotes chondrogenesis of MSCs.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Marsha Bush.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Ghivizzani, Steven.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041956:00001


This item has the following downloads:


Full Text





GENE-INDUCED CHONDROGENESIS OF MESENCHYMAL STEM CELLS
THROUGH VIRAL GENE DELIVERY




















By

MARSHA LYNN BUSH


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2010

































2010 Marsha Lynn Bush
































To my parents, Diana Bush and Jim Bush









ACKNOWLEDGMENTS

I thank my family for their unconditional love and support throughout my ongoing

educational pursuits; I thank Grandma for being my penpal, Geoffrey and Aaron for

being calm and injecting humor into my life, and Mama, Daddy, and Kim for being

patient listeners and providing sound advice.

I thank my mentor, Dr. Steve Ghivizzani, for his leadership and encouragement as

he taught me the importance of clear, persuasive writing, the value of effective

presentation skills, and how to think like a scientist. He provided me ample opportunities

to learn and experiment in his lab and to discuss the scientific merit of many

experimental outcomes. I greatly appreciate my advisory committee: Drs. Jorg Bungert,

Bryon Petersen, Ed Scott, and Phyllis LuValle, who have provided many hours of

guidance. My lab colleagues, who have become a surrogate family over the years, have

always been there for me: Jesse Kay, Rachael Watson, Carrie Saites, Celine Theodore,

Paddy Levings, Anthony Dacanay, and Jeetpaul Saran.

My first research experiences were in the lab of Dr. Chuck Fox at the University of

Kentucky College of Agriculture, Department of Entomology. I'm grateful for the time he

spent teaching me how to design experiments with statistically significant sample sizes

and appropriate controls. I further extend my gratitude to the many teachers and

professors who encouraged me to pursue my dreams and to never doubt myself.

I thank my dear friend Joe Menotti for his faith in me when the task felt

overwhelming. Finally, I thank my friends and fellow scientists whom I have met through

the Interdisciplinary Program (IDP) in Biomedical Sciences-especially Amanda

DuBose, Melissa Marzahn, Megan Greenlee, Dacia Kwiatkowski, Ahu Demir and









Maggie and Levi Watson-for their consolation over failed experiments and their shared

joy over triumphs, no matter how big or small.









TABLE OF CONTENTS

page

ACKNOWLEDGMENTS .................. ....................... ......... ............... 4

L IS T O F T A B L E S ........................................................................................... ........ 9

LIST OF FIGURES................................. ......... 10

LIS T O F A B B R EV IA T IO N S ............................................................ ............... 12

ABSTRACT .................................... ................................... ........... 15

CHAPTER

1 INTR O D U CTIO N ................................................................. .. ......... 17

Cartilage Biology..................................................... 17
A architecture of the K nee ............................................. .................... .... 17
A rticular C cartilage Structure ................................... ............................. ....... 18
Proteoglycans ......... ........... ......... ................ ......... ...... 19
Chondrocytes .................................. ............... 21
C o llage ns ...................... .......... ............................... ............... .. .. 2 3
Adjacent cells and tissues support articular cartilage............................. 25
C cartilage D evelopm ent ..................................................... ......................... 26
Osteogenic and Chondrogenic Differentiation and Regulation in
D eve lopm e nt ................................... .... .................... ..... .......... 26
B M P R eceptors and S ignaling ............................... ................ .... ............... 3 1
C cartilage R epair.................................... ........ ...... .......................... 32
The Body's Natural Approaches to Cartilage Defect Repair.......................... 32
Surgical Approaches to Cartilage Repair .............. ........ .......................... 33
Nonreparative restorative techniques ............ ..................................... 33
Reparative procedures....................................... ................. 34
Restorative strategies ......... .... ................ .......... ..... ........... 37
Reconstructive methods to treat cartilage defects ................... ........ 39
Gene- and Cell-Based Approaches to Cartilage Repair................................. 41
MSCs: the logical cell type for chondrogenesis............................... ..... 42
Mechanical stim ulation to promote cartilage............................................ 44
Regeneration and Repair Using MSCs: Gene Therapy and Orthopaedics ............ 44

2 MATERIALS AND METHODS ............. ......... ............................ 55

In Vitro Cell Culture.......................................... ........... 55
HEK293 and 293FT Cell Culture .............. ................. ................................ 55
Harvesting and Culturing Bovine MSCs ....... ..... ........... ....................... 55
Aggregate (Pellet) Culture ................................... .................... ... ...... 56
Chondrogenic M edia Form ulation ...................... ........... ....... ............. 56


6









Virus Preparation and Transgene Expression ............... .... ................. 57
Construction and Generation of Recombinant Adenoviruses Containing
Chondrogenic Transgenes ............... .. ...... ...................... ................. 57
Adenovirus Propagation and Amplification ...... ................ .......................... 57
Generation of Lentivirus .............. ......... .. .. ... ........... ................. 59
Construction and Generation of scAAV Vectors.............................. ............... 59
Gene Transfer to MSCs to Induce and Enhance Chondrogenesis ................ 60
Plasmid DNA transfection .......... .. .......... ......... ......................... 60
Transgene delivery using adenovirus ............ ..................................... 61
Methods to Detect Transgene Products ........ .... ..................... ............... 61
W western blot.......................... ............ ......... ......... 61
ELISA to detect secreted transgene products ................................ 62
Histology and Immunohistochemistry.................................. ...................... 62
R NA Extraction, RT, and rtPC R .................................. .................................. 63
In V ivo E x pe rim e nts ............... ........................................................ 6 4
Intra-A rticular Injections ........................................................... ..... 64
Harvesting Tissue, Decalcification, and Histology ................ ........... ........ 64

3 GENE DELIVERY STIMULATES CHONDROGENESIS OF MSCS ................... 67

Introduction ............... ....... ........................................................... ...... 67
R a tio n a le ................................................ ....... .......... ...... 7 0
R e s u lts .... ............... .. ....... ... .................................. 7 1
Cre-lox Recombination and Adenovirus Propagation............. ............... 71
Isolation of M SCs .......... ......... ........ ...... ............ ..... .......... 73
Gene-Mediated Chondrogenesis of MSCs...................................... ........... .. 74
Adenoviral-Mediated Delivery of BMP-4, BMP-7, Ihh, and Sox9 Drives
Chondrogenic Differentiation of M SCs .............................. ..................... 76
Adenoviral Delivery of CTGF Stimulates Proliferation of MSC Aggregates...... 81
Combinations of Sox9 with BMPs Induced Chondrogenesis........................... 82
Wistar Rat Responses to Adenoviral Transgene Delivery.............................. 84
D is c u s s io n ......... ............................................... .................................... 8 6

4 COMPARISON OF THE EFFECTS OF ADENOVIRAL, LENTIVIRAL, AND AAV
TRANSGENE DELIVERY TO MESENCHYMAL STEM CELLS........................... 99

Introduction ............... ....... ........................................................... ...... 99
Results ...................... ...... .......... .... ......... .. .......... ............... 100
Plasmid DNA Transfection of MSCs Results in Transient Expression ........... 100
Lentiviral Transduction of MSCs Proves Challenging................................. 101
Adenoviral-Mediated Delivery of BMP-4 to MSCs Induces Chondrogenesis. 102
Self-Complementary AAV-Mediated Delivery of BMP-4 to MSCs is
Com parable to Adenovirus.............................. .............. 102
D is c u s s io n ......................................................................... .... ........ 1 0 3

5 SUMMARY AND FUTURE DIRECTIONS ........... ....................................... 111









LIS T O F R E F E R E N C E S ............................................................................. ........ 115

BIOGRAPHICAL SKETCH ....................................................... 125









LIST OF TABLES

Table page

1-1 Collagen types. .................................. ........................ ................ 52

1-2 Morphogens and Growth Factors in Cartilage Development............................ 53

4-1 Comparison of gene delivery vectors reveals pros and cons............................. 110









LIST OF FIGURES


Figure page

1-1 The four zones of articular cartilage stretch from the superficial surface to the
deep zone, where hypertrophic chondrocytes are replaced by calcified
m atrix ..................... ....... ..... ..... ....... ............. ......... 47

1-2 The human knee contains specialized structures to withstand the forces of
m o v e m e nt........................ ..... .......... ..... .................................... 4 8

1-3 The arrangement of structures in the articular cartilage matrix is designed to
absorb forces, especially compression...... ................. ..... .............. 49

1-4 A proteoglycan aggregate is made up of many proteoglycan subunits
attached to a hyaluronic acid backbone via link protein.............................. 50

1-5 Fibrillar collagens, including collagen type II, form a characteristic triple helix... 51

2-1 Assay for in vitro chondrogenesis ............ ........... .......... .... ............. 66

3-1 ELISAs of conditioned media from MSC aggregates transduced with BMP-4
and BMP-7 indicated protein production in response to viral doses
adm inistered.................................. .......... ...... ......... ........... 89

3-2 Western blotting confirmed transgene expression following adenoviral gene
delivery to M SCs. ............ .......... .. ........ ............. ... ...... ......... 90

3-3 MSC aggregates expressing BMP-4 were evaluated for chondrogenesis
through toluidine blue staining of proteoglycans and immunohistochemistry
for collagen types I and II. ....................................... ................ 91

3-4 MSC aggregates infected with Ad.BMP-7 over a range of doses yielded
pronounced chondrogenesis at the 1000 vp/cell dose only.............................. 92

3-5 Indian hedgehog aggregates expressed proteoglycan and collagen type II
following adenoviral gene delivery....... ......... .............. ............... 93

3-6 Sox9 induced cellular proliferation and matrix synthesis with less
differentiation than other transgenes tested. ............................. ... ................ 94

3-7 Collagen type X staining of Sox9 and BMP-4 aggregates............... ............. 95

3-8 CTGF expression in MSC aggregates promotes cellular proliferation but no
chondrogenic differentiation.................... ........ ........................... 96

3-9 Sox9 in combination with BMP-4 and BMP-2 in MSC aggregate culture had
no additive effects upon chondrogenesis.. ............... ................................... 97









3-10 Wistar rat knees exhibited varying responses to adenoviral delivery of
pleiotropic transgenes. ............ ......... ..................... ........... ............... 98

4-1 ELISA of conditioned media demonstrates expression of BMP-4 as
measured 48 hours post-transfection............................ ................... 106

4-2 GFP expression in 293 cells 72 hours after infection with 15 pl of LV-GFP in
a minimal volume of serumless media...... ... .. ........... ............... .... .... 107

4-3 LV-BMP-4 expression was significantly less than Ad.BMP-4 expression in
bovine M SCs. ............. .... ..................................................... 107

4-4 AAV serotypes 1, 2, 5, and 8 were screened for transduction efficiency on
low passage bovine MSCs in monolayer..................................... ................ 108

4-5 MSCs transduced with scAAV-BMP-4 serotypes 2 and 5 demonstrate matrix
protein synthesis in aggregate culture............... ........ .............................. 109









LIST OF ABBREVIATIONS

AAV adeno-associated virus

ACT autologous chondrocyte transplantation

BM bone marrow

BMP bone morphogenetic protein

bp base pairs

C degrees Centigrade

CaPO4 calcium phosphate

CAR Coxsackie-adenovirus receptor

CDMP cartilage derived morphogenic protein

cDNA complementary DNA

CMV cytomegalovirus

CsCI cesium chloride

DMEM Dulbecco's Modified Eagle Medium

DNA deoxyribonucleic acid

ECM extracellular matrix

ELISA enzyme-linked immunosorbant assay

FBS fetal bovine serum

FGF fibroblast growth factor

GAG glycosaminoglycan

GDF growth/differentiation factor

GFP green fluorescent protein

hr hour

IGF insulin-like growth factor

Ihh Indian hedgehog









kg

MSC

pg

mg

min

mL

mM

MMPs

ng

nm

OA

PBS

PCR

pg

PG

RA

rAAV

RNA

sc

scAAV

SDS-PAGE

ssAAV

TGF-P3

trs

U


kilogram

mesenchymal stem cell

microgram

milligram

minute

milliliter

millimolar

matrix metalloproteases

nanogram

nanometer

osteoarthritis

phosphate buffered saline

polymerase chain reaction

picogram

proteoglycan

rheumatoid arthritis

recombinant adeno-associated virus

ribonucleic acid

self-complementary

self-complementary adeno-associated virus

sodium dodecyl sulfate-polyacrylamide gel electrophoresis

single-stranded adeno-associated virus

transforming growth factor beta

terminal resolution site

units









UV ultraviolet

vg vector genome

vp virus particle

VSV-G vesicular stomatitis virus G protein









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

GENE-INDUCED CHONDROGENESIS OF MESENCHYMAL STEM CELLS
THROUGH VIRAL GENE DELIVERY


By

Marsha Lynn Bush

August 2010

Chair: Steven C. Ghivizzani
Major: Medical Sciences Biochemistry and Molecular Biology


Articular cartilage is a highly specialized tissue that allows for near frictionless

motion of diarthrodial joints. When cartilage is damaged as a result of injury or disease,

the body's innate capacity for repair is often insufficient due to the low cellular density,

avascular and aneural nature of cartilage. Typically the repair tissue is a fibrotic scar

which lacks the structural properties of native cartilage. Adult mesenchymal stem cells

(MSCs) are well-suited for cell-based therapies for cartilage repair since they are readily

available from many tissue sources and capable of differentiating along multiple

mesenchymal lineages. Gene transfer to MSCs is a viable method for achieving

sustained local expression of specific protein factors, which has been shown to induce

chondrogenesis in vitro and may enhance chondrogenic differentiation in vivo.

Previous experiments have shown that delivery of the cDNAs for TGF-3 and BMP-

2 induces chondrogenesis of bovine MSCs in high density aggregate culture. In these

studies, we expanded the analyses to include additional cDNAs whose protein products

are associated with chondrogenic differentiation during development, including bone

morphogenetic proteins BMP-4, BMP-7, developmental morphogen Indian hedgehog









(Ihh), transcription factor Sox9, and connective tissue growth factor (CTGF). We

transduced early passage bovine MSCs with adenoviral vectors carrying the complete

cDNAs for the candidate transgenes at doses ranging from 10 virus particles/cell to

100,000 vp/cell. Chondrogenesis was evaluated by gross examination of aggregate

morphology, toluidine blue staining for proteoglycan expression and collagen types I, II,

and X immunohistochemistry. The greatest biological responses for each transgene

were observed in the dose range of 100-1000 vp/cell; higher viral doses appeared to

inhibit chondrogenesis, while lower viral doses were insufficient to yield a pronounced

effect. We found that gene transfer of the cDNAs for BMP-4, BMP-7, Ihh, and Sox9

induced chondrogenesis of bovine MSCs while CTGF was not chondrogenic. BMP-4

supports chondrogenesis as effectively as other members of the TGF-3 superfamily

such as BMP-2.

In further experiments to elucidate the most suitable gene delivery system for

animal models and clinical applications, we evaluated vector systems currently

available: adeno-associated virus (AAV), lentivirus, and non-viral transfection relative to

adenovirus. Simple transfection alone was unable to generate the levels of expression

necessary to promote chondrogenesis in our system. Both adenovirus and AAV

transduction of MSCs resulted in robust BMP-4 expression over the course of 21 days,

complete with chondrogenic differentiation and the expression of cartilage matrix

proteins. Lentivirus, which offers the potential for long-term gene expression, was ill-

suited for our application because it required long-term culture selection of cells, which

altered cellular morphology and negatively impacted cell survival. These data indicate

that scAAV serotype 2 delivery of BMP-4 promotes chondrogenesis of MSCs.









CHAPTER 1
INTRODUCTION

Cartilage Biology

Cartilage is a specialized connective tissue found in various locations throughout

the human body and is subdivided into three main types: 1) hyaline articularr), 2) elastic,

and 3) fibrocartilage1. Hyaline cartilage, named for its glassy appearance, covers

articulating surfaces, such as those at the ends of long bones; therefore, it is also

known as articular cartilage. Its unique architecture offers firm support with some

pliability and consists of 80-90% water, with the remainder of the tissue composed

primarily of collagen type II fibers and proteoglycans. Cells known as chondrocytes

reside in the cartilage at low density and work to slowly remodel the matrix to keep it

structurally and functionally sound2. In childhood, hyaline cartilage is a key component

of the growth plates, the actively growing regions that add length to the ends of long

bones1 (Fig.1-1A).

Architecture of the Knee

Diarthrodial joints, such as the knee, have a specialized architecture that is

capable of withstanding repeated extreme forces over the lifetime of an individual, often

without problems for 80 or more years. Mesenchymal tissues in the knee joint, including

cartilage, tendons, ligaments, synovium, and meniscus, have specific features that

enable smooth locomotion and resistance to compressive and tensile forces (Fig. 1-2).

Ligaments present within and without the joint capsule guide movement of the femur

and tibia to control extension and flexion, secure the articulating bones when standing,

and prevent hyperextension or overflexion of the knee joint. The knee can absorb

vertical forces equal to nearly seven times body weight; however, the knee joint is









susceptible to damage from horizontal forces or twisting movements such as those that

occur in football and other contact sports.

Articular Cartilage Structure

Articular cartilage covers and protects the joints at sites of articulation, such as the

knee, ankle, elbow, and knuckle. To withstand the stress of movement of the human

body, articular cartilage must be durable yet provide an effective cushion against load-

bearing and impact, and it must be smooth to provide nearly frictionless motion. The

nature and structure of articular cartilage impart these necessary properties.

The extracellular matrix (ECM) has unique properties that enable articular cartilage

to rebound after impact. Moreover, it is remarkably durable and is capable of

withstanding repeated exposure to extreme forces and pressure, which effectively

cushions the ends of the bones. Extracellular matrix is synthesized and maintained by

the resident population of chondrocytes through anabolic and catabolic mechanisms

and is composed mainly of collagen fibrils and proteoglycans3 (Fig. 1-3).

As depicted in Figure 1-1 B, articular cartilage is highly organized and consists of

four zones, or layers: 1) superficial, 2) transitional, 3) radial, and 4) calcified cartilage3'4.

The superficial zone contains chondrocytes that are flattened and lie parallel to the

surface. Collagen type II fibrils also have a parallel orientation in this layer, which allows

the cartilage to withstand the shear forces generated during normal joint loading. The

transitional zone, as its name suggests, exhibits a cell morphology and ECM

composition intermediate to that of the superficial zone and the deeper radial zone.

Transitional layer chondrocytes are rounded and produce more proteoglycan and less

collagen than in the superficial layer. Deep to this zone lies the radial zone, where the

cells are aligned in columns perpendicular to the joint surface4. The radial zone is the









largest layer, with the thickest collagen fibrils, the most concentrated proteoglycans, and

the lowest water content. The radial zone is separated from the deepest zone, calcified

cartilage, by the tidemark, an area that appears to help tether the cartilage by increasing

the contact area between the layers5. The zone of calcified cartilage forms the transition

between soft articular cartilage and the hard underlying bone4'5. Cells in this zone are

isolated and almost completely surrounded by calcified cartilage, suggesting that they

have a low metabolic rate, but it is known they are not completely inactive5.

Throughout the body, articular cartilage is thicker, with greater proteoglycan

content, in load-bearing areas and thinner in areas where loading is minimal4.

Mechanical loading affects the morphology and the metabolic activity of chondrocytes4,

and controlled loading is thought to contribute to enhanced cartilage healing after injury.

Proteoglycans

The extracellular matrix consists of a network of proteoglcyans (PGs) and collagen

fibers arranged to control the flow of water molecules to cushion the joint from forces

associated with normal motion. Proteoglycans give the tissue its ability to resist

compression and remain durable for up to 80 years, or more, in some humans.

Proteoglycans are proteins bound to long-chain polysaccharides known as

glycosaminoglycans, or GAGs. Large aggrecating proteoglycans known as aggrecans

are the most abundant proteoglycans in cartilage, consisting of a linear protein core with

numerous GAG chains of chondroitin sulfate or keratan sulfate attached6 (Fig. 1-4). In

addition to large aggregating proteoglycans like aggrecan, small non-aggregating

proteoglycans, including decorin, biglycan, and fibromodulin, bind to other matrix

molecules and help to stabilize the matrix7.









Hyaluronic acid (HA), or hyaluronate, molecules are the central organizing units of

the cartilage matrix. They are long chains with up to 100 glycoproteins attached to and

extending from them. Link protein connects each proteoglycan subunit to the

hyaluronate backbone. The large concentration of negative charges from the GAG side

chains attract and hold polar water molecules in the matrix by osmotic pressure8. This

charge repulsion keeps the chains separated in a characteristic bottle-brush formation

(Fig. 1-4). When hydrated, proteoglycans account for most of the physiological mass of

cartilage. They provide a structure to hold and control the flow of water, the integral

factor in cartilage's ability to provide its protective cushioning properties throughout

lifetime of the individual.

The architecture of articular cartilage regulates the flow of fluid in the cartilage and

cushions the joint from mechanical stresses such as compressive loads9. When

cartilage is compressed during normal joint motion, water flows out to the joint cavity,

mixing with the existing synovial fluid that lubricates the joint. The negative charges of

the GAGs repel one another when forced into close proximity, enabling them to resist

further compression. Additionally, collagen type II fibrils provide the strength needed to

retain the shape of cartilage under compression and hinder the expansion of

proteoglycans7'10. When the pressure is released, the cartilage springs back to its

original form and water molecules are drawn back into the proteoglycans. This flow of

liquid into and out of cartilage also carries nutrients to and wastes from chondrocytes,

providing nourishment, which helps explain why periods of inactivity can result in

weaker cartilage that injures easily1.









If proteoglycans are damaged by trauma, infection or enzymes, such as those

released during inflammatory disease, their structure degenerates and their water-

holding capacity changes8. The collagen meshwork also begins to break down, which

reduces constraint on the proteoglycans and allows them to imbibe additional water,

thus reducing cartilage stiffness. Over time damaged cartilage loses its resilience to

mechanical forces in the joint, and the subchondral bone can become exposed, causing

severe pain and disability due to the abundant nerve supply and bleeding from the bone

marrow and surrounding vasculature.

Due to the density of the extracellular matrix secreted and maintained by

chondrocytes and the physical requirements of the tissue, there is no room for blood

vessels or nerve fibers to permeate cartilage. Indeed, the permeation of the matrix with

vascular or neural tissues would significantly compromise its protective properties.

Instead, nutrients diffuse through the perichondrium, a well-vascularized dense irregular

connective tissue that surrounds the surface of most cartilage structures1.

The matrix, while continually undergoing remodeling by chondrocytes, also serves

to protect chondrocytes from injury resulting from normal use of the joint. Because

cartilage lacks innervation, injuries that occur within the cartilage layer often go

unnoticed until there is penetration of the subchondral bone. Additionally, the structure

of the matrix determines the types of molecules and which concentrations reach the

cells encased within.

Chondrocytes

Chondrocytes are mature cartilage cells found in deep pits called lacunae1 where

they function to maintain cartilage homeostasis through the perpetual degradation and

synthesis of matrix components". Chondrocytes diffusely populate articular cartilage,









exhibit a low metabolic rate compared to tissues such as muscle, and are restricted

from moving and dividing by the dense matrix fibers. The cells do not interact with one

another, but instead interact with the ECM (Fig. 1-1 B). In the deepest layers of cartilage,

chondrocytes exist in a hypoxic environment, so metabolism is mainly anaerobic, with

the conversion of glucose to lactic acid4. In normal articular cartilage, there is a low

turnover of the extracellular matrix, and chondrocytes are thought to rarely divide4'12

When articular cartilage sustains damage, however, the chondrocytes form clusters and

cellular activity increases4. Aging chondrocytes, on the other hand, gradually lose their

ability to divide, which supports the observation that the cartilage in older individuals

heals more slowly than that of young people13'14. Sports injuries in adolescence or

young adulthood may set the stage for pain later in life as those early events damage

cartilage, causing it to unnaturally erode. Damaged cartilage may calcify, escalating the

extent of damage as this transition to bone results a loss of nutrient supply and

subsequent chondrocyte death.

The cartilage microenvironment may have a profound influence upon chondrocyte

lifespan. The chondrocyte has a finite lifespan in the epiphyseal growth plate, whereas it

has a very long, stable phenotype evident in articular cartilage15. Chondrocytes of

articular cartilage are long-lived cells of mesenchymal origin, but their lifecycle is

unconfirmed. It is possible that they are post-mitotic and replaced by new cells

infiltrating from the subchondral bone or that a small population of progenitor cells exists

within the cartilage.

Chondrocyte dedifferentiation in culture. The stability of the phenotype of

articular chondrocytes is critically dependent on physical environment and cell density.









In monolayer culture, these cells progressively lose their chondrogenic phenotype,

transitioning from cubiodal cells with high ECM synthesis to spindle-shaped cells that

primarily produce collagen type I. Dedifferentiation can be avoided or delayed by

changing the geometry of the cell culture through the use of 3D matrices such as

agarose or alginate, high density micromass cultures, or pellet cultures15

Collagens

The dry mass of articular cartilage is composed primarily of type II collagen and

aggrecan, a proteoglycan. Collagen type II makes up 90% of the collagen present, with

minor collagens types VI, IX, and XI making up the other 10% (Table 1-1). Collagen

fibrils provide tensile strength, serving as a natural scaffold to anchor and organize cells

and hydrated proteoglycans and possibly guiding cell signaling responses during

development and repair16

Collagens differ in their ability to form fibers and to organize the fibers into

networks (Table 1-1). The typical structure of fibrillar collagens is a right-handed triple

helix, as shown in Figure 1-5, which arises from an abundance of three principal amino

acids: glycine, proline, and hydroxyproline2. These form a repeating pattern of Gly-Pro-

X where X is any amino acid. The side chain of glycine, a single H atom, is the only side

chain capable of fitting into the crowded center of the triple strand helix.

Type I collagen fibrils have enormous tensile strength; that is, this type of collagen

can be stretched without being broken. These fibrils, roughly 50 nm in diameter and

several micrometers long, are packed side-by-side into parallel bundles, termed

collagen fibers. In tendons, collagen type I fibers connect muscles with bones and must

withstand enormous forces. Gram for gram, type I collagen is stronger than steel2.









Collagen type II, the primary collagen of articular cartilage, has fibrils that are

smaller in diameter than type I and are oriented randomly in the viscous proteoglycan

matrix (Fig. 1-3). Such rigid macromolecules impart a strength and compressibility to the

matrix and allow it to resist large deformations in shape. This property allows joints to

absorb shock. Type II fibrils are cross-linked to proteoglycans in the matrix by collagen

type IX, a collagen consisting of two long triple helices connected by a flexible kink (Fig.

1-5B). The globular N-terminal domain extends from the composite fibrils, as does a

heparan sulfate molecule, a type of large, highly charged polysaccharide that is linked

to the type IX collagen chain at the flexible kink. These protruding nonhelical domains

are thought to anchor the fibril to proteoglycans and other components of the matrix.

The interrupted triple-helical structure of type IX collagen prevents it from assembling

into fibrils; instead, collagen type IX associates with fibrils formed from other collagen

types and thus is called fibril-associated collagen (Table 1-1).

Post-translational modification of procollagen is crucial for the formation of mature

collagen molecules and their assembly into fibrils; defects in this process have serious

consequences2. For example, the activity of prolyl hydroxylases requires an essential

cofactor, ascorbic acid (vitamin C). In cells deprived of ascorbate, as in the disease

scurvy, the procollagen chains are not hydroxylated sufficiently to form stable triple

helices at normal body temperature, nor can they form normal fibrils. Consequently,

nonhydroxylated procollagen chains are degraded within the cell.

Type II collagen fibers are firmly embedded in subchondral bone and rise from

there to the cartilage surface, where they bend to form arches8. The overlapping fibers









form a mesh framework that gives stability to cartilage (Fig. 1-3). Collagen fibers provide

the framework that allows proteoglycans to hold water molecules within the matrix.

Adjacent cells and tissues support articular cartilage

The adjacent cells and tissues in the joint-synoviocytes, bone marrow, MSCs,

meniscus, ligaments and tendons-provide support to articular cartilage. Most proximal

to the bones and cartilage of the joint is the synovium, a thin layer of vascularized

connective tissue that lacks a basement membrane. Two cell types are present: type A,

macrophage-like cells, and type B, fibroblast-like cells. Macrophage-like or phagocytic

cells remove microbes and the debris that results from normal wear and tear in the joint.

Type B synoviocytes produce synovial fluid, which consists of hyaluronan, or hyaluronic

acid, lubricin, proteinases, and collagenases, and serves to surround, lubricate, and

nourish the joint space. This fluid forms a thin layer (roughly 50 pm) at the surface of

cartilage and also seeps into microcavities and irregularities in the articular cartilage

surface, filling all empty space. Normal synovial fluid increases the viscosity and

elasticity of articular cartilage, and there is also some evidence that it helps regulate

synovial cell growth.

The liquid present in articular cartilage effectively serves as a synovial fluid

reserve. During movement, the synovial fluid held in the cartilage is squeezed out

mechanically to maintain a layer of fluid on the cartilage surface (so-called weeping

lubrication). Synovial fluid is chiefly responsible for reducing friction between apposing

surfaces of cartilage and absorbing shock. In addition, movement of synovial fluid

supplies oxygen and nutrients to and removes carbon dioxide and metabolic wastes

from chondrocytes within articular cartilage.









Cartilage Development

Osteogenic and Chondrogenic Differentiation and Regulation in Development

Cartilage morphogenesis and osteogenesis are influenced by developmental

signals. Current research is exploring which signals direct cartilage to remain at the

ends of long bones when the rest of the bone ossifies. Cartilage serves as the blueprint

for subsequent bone and joint morphogenesis as well as tendon and ligament

insertions15

Pre-cartilage condensation observed in early limb buds is a transient phase of

skeletogenesis, in which a cartilaginous framework serves as a scaffold for later

ossification of skeletal elements17. To form cellular condensations, cells actively move

toward a center, resulting in an increase in cells per unit area rather than an increase in

cellular proliferation7'18. Mesenchymal cell condensation is crucial for chondrogenesis

and is associated with an increase in cell-cell and cell-matrix interactions through cell-

cell adhesion molecules and gap junctions, which facilitate intercellular

communication17'19. Cell-cell interactions presumably trigger one or more signal

transduction pathways that initiate chondrogenic differentiation. N-cadherin and N-CAM

are two cell adhesion molecules expressed in condensing mesenchyme, and they play

a role in mediating mesenchymal condensation. N-cadherin is responsible for cell-cell

adhesion and its expression likely modulates the progression of chondrogenesis. The

actions of N-cadherin can be modulated by Wnts, a family of secreted glycoproteins that

influence cellular condensation and chondrogenic differentiation in early development20.

During embryonic development, MSCs give rise to cartilage of two types:

permanent and transient. Permanent hyaline cartilage arises from MSCs at the distal

ends of developing bones. Following the initial cellular condensation event, MSCs









differentiate toward stable chondrocytes that synthesize the extracellular matrix of

articular cartilage. Transient cartilage forms a skeletal framework which is later replaced

by mineralized bone during the process of endochondral ossification.

When considering the candidate biological factors for enhancing the repair

response of cartilage, members of the transforming growth factor-3 (TGF-3) superfamily

are of particular interest because of their abilities to promote chondrogenic activity in

vitro and in vivo. The TGF-3 superfamily comprises structurally related regulatory

molecules that include the five TGF-3 isoforms, bone morphogenetic proteins (BMPs),

growth/differentiation factors (GDFs), activins, inhibins, nodal, and glial-derived

neurotrophic factor (GDNF)21.

TGF-p. TGF-3 has been shown to induce mesenchymal chondrogenesis in

cultures of cell lines and primary adult MSCs22'23'24'25 and has been implicated in

embryonic cartilage formation26. TGF-3 also regulates the growth and synthetic

processes of chondrocytes, stimulating extracellular matrix synthesis and chondrocyte

proliferation in cell and organ culture27'28'29. Similar increases in matrix synthesis have

been reported in chondrocyte cultures genetically modified with TGF-31 cDNA30

Administration of recombinant TGF-31 in vivo has been reported to increase

proteoglycan synthesis and restore proteoglycan levels in the knees of arthritic mice31

TGF-31 promotes cellular proliferation and initiates and maintains chondrogenesis

of mesenchymal progenitor cells32. Other cell types, including both embryonic and adult

fibroblasts, are able to undergo differentiation into a chondrogenic phenotype in the

presence of chondrogenic inducers33, demonstrating that TGF-3 can influence

differentiation of more cell types than just chondroprogenitors. TGF-betas have been









shown to activate intracellular signaling cascades, particularly those cascades

containing MAP kinases, p38, ERK-1 and JNK to promote cartilage-specific gene

expression32

Bone morphogenetic proteins. Bone morphogenetic proteins (BMPs) make up

nearly one-third of the TGF-3 superfamily and are closely linked to the formation of

bone, cartilage and connective tissues in vivo34'35. BMPs regulate a diverse range of

developmental processes during embryogenesis and postnatal development, and they

control the differentiation of musculoskeletal tissues including bone, cartilage, tendon,

and ligaments36 (Table 1-2). The homeostasis of articular cartilage in the joint is

maintained as a balance between anabolic morphogens such as bone morphogenetic

proteins (BMPs) and cartilage-derived morphogenetic proteins (CDMPs) and catabolic

cytokines such as IL-1, IL-17, and TNF-a15

Studies both in vitro and in vivo have shown that BMP signaling is required for the

formation of cartilaginous condensations and for the differentiation of precursor cells

into chondrocytes37. Like TGF-3, BMPs -2, -4, and -7 can stimulate chondrogenesis of

mesenchymal progenitor cells in vitro38'39'40. BMP-2 is shown to accelerate the healing

of osteochondral defects in vivo when delivered via a collagen sponge41'42. BMP-7 is

used clinically for spinal fusion and fracture repair in the long bones43. Implantation of

periosteal derived progenitor cells genetically modified to express BMP-7 improved

repair of rabbit osteochondral defects, indicating the efficacy of a combined gene

therapy and tissue engineering approach44. Further clinical applications of BMP-7/OP-1

include incorporation of the recombinant human protein into collagen scaffolds for repair

of full-thickness canine osteochondral defects. Repair tissue in treated canine knee









defects had a hyaline cartilage-like appearance and integrated continuously with the

intact cartilage adjacent to the defect45. Both BMP-2 and BMP-7 improve cartilage repair

in studies of artificial cartilage plugs containing either collagen or hydroxyapatite

blended with biodegradable polymers46.

BMP-2 and BMP-4 are equally effective in promoting chondrogenesis in primary

hMSCs in aggregate culture, but BMP-4 aggregates show a lower tendency to progress

toward hypertrophy, a crucial characteristic to consider for cartilage repair47. Local

delivery of BMP-4 by retrovirally transduced MDSCs (muscle-derived stem cells) shows

enhanced chondrogenesis and significantly improved articular cartilage repair in rats48

IGF-1. Insulin-like growth factors are another class of molecules with potential for

improving endogenous repair. IGF-1 is a major regulator of matrix synthesis in articular

cartilage as it stimulates chondrocyte metabolism and promotes healing of cartilage

lesions in vivo,49,50 but its role in chondrogenesis is less clear. Ad.IGF-1 effectively

transduces chondrocytes, MSCs, and synovial cells resulting in IGF-1 production

sufficient to stimulate matrix gene expression and proteoglycan production49. When

chondrocytes are cultured in vitro and exposed to varying concentrations of Ad. IGF-1,

the cells are readily transduced; they produce significant amounts of IGF-1 to promote

increased cartilage matrix gene expression and resist de-differentiation for 28 days51.

IGF-1, when combined with TGF-3, can have synergistic effects in promoting

chondrogenesis52

Sox9. In addition to growth factors, research has identified increasing numbers of

biological molecules that are involved in regulation of chondrogenic differentiation,

expression of cartilage matrix genes, and accelerated repair of cartilage defects in









animal models. Of these, the transcription factor SRY-related HMG box gene 9 (Sox9)

is most closely associated with the expression of cartilage ECM genes and cartilage

formation53'54'55'56. Mutations in Sox9 result in campomelic dysplasia, a semilethal

skeletal malformation syndrome and XY sex reversal57.

Sox9, dubbed the chondrogenic master gene, binds to regulatory sequences in the

promoter region of several cartilage genes, thus enhancing their expression in

chondrocytes56. Sox9 regulates the expression of chondrogenic genes such as

aggrecan and collagen types II, IX and XI during chondrocyte differentiation52. In the

collagen type II al gene (Col2al), Sox9 binds within a 48 bp enhancer region located in

the first intron and acts in concert with two cofactors: long form of SRY-related HMG-

box gene 5 (L-Sox5) and SRY-related HMG-box gene 6 (Sox6)58. The two cofactors are

normally expressed in MSCs; therefore, Sox9 gene delivery is sufficient to enhance the

level of chondrocytic genes58. Because it functions intracellularly and cannot be

delivered in soluble form, gene transfer of Sox9 cDNA to mesenchymal progenitor cells

offers a means to investigate the reparative potential of this molecule.

Ihh and PTHrP. Chondrocyte proliferation and maturation are key points of

regulation that may influence a repair process based on differentiation of progenitor

cells. Indian hedgehog (Ihh), a member of the hedgehog family of cell surface-

associated ligands, is expressed in prehypertrophic chondrocytes of the growth plate,

and functions to inhibit chondrocyte hypertrophy by maintaining expression of

parathyroid hormone related peptide (PTHrP) through a negative feedback loop59

Altering the expression of these proteins during chondrogenesis may serve to delay the

onset of hypertrophy, chondrocyte apoptosis, and formation of bone, while increasing









the pool of proliferating chondrocytes. Indeed, addition of PTHrP has been shown to

inhibit chondrocyte hypertrophy during in vitro chondrogenesis of primary adult

MSCs60,61.

BMP Receptors and Signaling

Two major types of membrane-bound serine/threonine kinase receptors are

required for BMP signal transduction: BMPR-I and BMPR-II. Additionally, there are two

options for intracellular signaling: the Smad and MAP kinase pathways. When BMPs

bind to pre-formed heteromeric BMPR-I:BMPR-II receptor complexes, the Smad

pathway is activated36, whereas BMP ligand binding that induces the formation of

heteromeric receptor complexes induces the MAPK (mitogen-activated protein kinase)

pathway35. The BMP-specific Smad proteins include receptor-regulated Smads 1, 5,

and 8 (R-Smads), a common-partner Smad 4 (Co-Smad), and inhibitory Smads 6 and 7

(I-Smads).

During joint morphogenesis, BMP binding proteins play a role in defining the

boundaries between muscle, cartilage, perichondrium, and tendon/ligament36. BMP

signaling is constrained on many levels by antagonists such as Noggin, Chordin,

follistatin, ventropin, twisted gastrulation, Gremlin, Cerberus, and DAN 62. Noggin binds

to BMP-2 and -4 with high affinity and blocks their interaction with BMP receptors.

Chordin also binds to BMP-2 and -4 to govern pattern formation (as originally studied in

Drosophila). Antagonists, including Chordin, are proteolytically activated by an MMP,

BMP-1, which was misidentified upon its discovery. BMP-1 is not a BMP per se, rather,

it is a BMP inhibitor. DAN family members are newly identified BMP antagonists based

on screens of cDNA libraries; their role in articular cartilage development and

homeostasis as well as arthritis is not yet clear63.









Cartilage Repair

Arthritic conditions affect over 70 million adults in the United States and cause an

economic burden in excess of $60 billion annually. These numbers are projected to

climb substantially as the population increases in age and the incidence of obesity

continues to skyrocket.

Pathological conditions, such as osteoarthritis and rheumatoid arthritis, and

traumatic conditions, such as intra-articular fracture or cartilage tearing from ligament

injury, all yield damage to articular cartilage64. Articular cartilage defects fall into two

categories: partial and full-thickness. Partial thickness defects are limited to the cartilage

layer only, and do not penetrate through to the subchondral bone. In the absence of

blood, a reservoir of stem cells and growth factors, there is little potential for the defect

to repair spontaneously. These defects deteriorate with time and can lead to additional

problems as the synovial lining becomes irritated by loose cartilage flaps, and the knee

locks due to cartilage detachment65. Full-thickness defects, on the other hand, penetrate

through the cartilage to the subchondral bone causing rupture of the local vasculature

and access to the marrow. This provides an avenue for progenitor cells from the bone

marrow to enter the defect and promote spontaneous healing through the formation of

fibrocartilaginous repair tissue.

The Body's Natural Approaches to Cartilage Defect Repair

Unlike bone, which has great regenerative potential, cartilage has no vascularity or

innervation; therefore, it has a low innate capability for self repair and regeneration.

Injury to cartilage usually heals through formation of a fibrocartilage scar. Fibrocartilage,

consisting predominantly of type I collagen fibrils with unordered proteoglycans and a

random cell arrangement, has inferior mechanical and biological properties compared to









the highly ordered network of type II collagen fibrils, proteoglycans and chondrocytes

present in hyaline cartilage. Over time, fibrocartilage repair tissue degenerates, resulting

in permanent loss of structure and function and leading to severe pain66

Although they do not lead to cartilage healing, palliative therapies are the primary

approach to treat symptoms of knee lesions. A combination of physiotherapy to

maintain range of motion and strengthen the affected limb, weight loss to decrease

forces exerted upon the knee, and NSAIDs for pain relief are among the conservative

treatment procedures commonly used. Intra-articular injections of analgesics along with

steroids will relieve pain, and injections with hyaluronate promote increased lubrication

and decreased friction on joint surfaces, showing a modest improvement over placebo

in clinical studies11'65

Surgical Approaches to Cartilage Repair

Nonreparative restorative techniques

Nonreparative restorative techniques for damaged cartilage include debridement,

chondral shaving, and joint lavage11. These techniques can be performed

arthroscopically, and are thought to relieve pain and improve mobility, but they do not

on their own restore the structure and function of diseased cartilage67'68'69. One major

drawback of nonreparative cartilage restoration techniques is that long-term benefits are

reduced due to a loss of chondrocytes at the border between healthy and damaged

cartilage after the injured cartilage is removed7.

Debridement. The debridement procedure, developed by Magnuson in 1941,

involves removal of inflammatory cells and other fragments such as chondral flaps,

osteophytes, torn ligaments, degenerated menisci and other debris resulting from

arthritis-mediated joint damage71'72'73. Debridement is intended to eliminate the









biochemical and mechanical factors that cause arthritis symptoms, resulting in pain

relief and improvement of joint functions. The effects are short-lived and somewhat

unpredictable, with 1/3 to 2/3 of patients showing improvement of symptoms at follow-

up evaluation73'74

Chondral shaving. Chondral shaving, developed in 1908 by Budinger, excises

damaged cartilage to relieve pain, using a motorized shaver. Shaving the cartilage is

thought to convert the fibrillated damaged surface to a smooth surface; however, further

examination has shown that shaving yields a rough surface with grooves1. Adverse

effects on chondrocytes and cartilage matrix can result from chondral shaving, most

notably chondrocyte death from the heat generated by frictional resistance to the tools

used. Additionally, there is a risk of chondral tears that do not heal, leading to

progressive degeneration of the remaining cartilage4'75

Knee joint lavage. Joint lavage is the technique of rinsing the joint with a

physiological fluid to remove debris76. It is frequently coupled with a debridement

procedure77, and is usually performed when more conservative treatments, such as

debridement and chondral shaving, prove inadequate. Suction of the fluid removes

degradation products, inflammatory cells and degradative enzymes; however, the

irrigation fluid can potentially harm cartilage and does nothing to halt disease

progression11'69

Reparative procedures

Reparative strategies, also termed marrow stimulation techniques, aim to initiate

bleeding from the subchondral bone, which releases progenitor cells, among other cells

with chondrogenic potential, from the vascular system to the site of cartilage injury78'79.









This causes a blood clot to form, plugging the injury site and paving the way for

chondroprogenitor cells to generate a fibrocartilage scar 1'80. Although fibrocartilage has

inferior mechanical capabilities, it does serve a purpose in filling the chondral defect and

covering the underlying bone, thus reducing pain and swelling65. While fibrocartilage fills

the site of cartilage injury and temporarily alleviates pain, it is functionally inadequate as

a replacement for articular cartilage in the long term58. Reparative procedures include

arthroscopic abrasion arthroplasty, subchondral drilling, microfracture and

spongialization, a modified technique that combines debridement and suchondral

drilling.

Arthroscopic abrasion arthroplasty. Arthroscopic abrasion arthroplasty is a

minimally invasive procedure that involves burring the exposed bone to access the

vasculature of the subchondral plate, which promotes formation of a blood clot and

subsequent formation of fibrocartilaginous repair tissue. In short term follow-up, nearly

half of the patients showed improvement; however, some studies have shown that

breakdown of repair tissue can occur as early as one year after the procedure". Other

studies indicate that when the joint is properly protected, fibrocartilaginous repair tissue

maintains integrity for up to six years76

Microfracture. The microfracture technique was developed by Steadman to

enhance chondral resurfacing by allowing the influx of marrow elements-mesenchymal

stem cells, growth factors, and other proteins-to create a microenvironment that would

promote new tissue formation and take advantage of the body's natural healing

process79. This arthroscopic procedure involves a preparative debridement step to

remove the damaged cartilage and expose the subchondral bone followed by a V-









shaped piercing using a specialized awl which makes multiple perforations

approximately 3 mm apart76'79. Subchondral piercing causes bleeding from the bone

marrow and leads to formation of a blood clot populated with platelets, growth factors,

and progenitor cells, which adheres to the exposed bone surface, fills the defect, and

progresses toward a cartilage-like repair tissue11'79. In younger, active patients (<35

years of age), microfracture remains the procedure of choice for cartilage lesions

smaller than 2.5 cm81. Long-term follow-ups from 2 to 12 years report pain relief and

restored knee function for 75% of patients with deep subchondral defects4'60'76.

Conversely, another study shows a decline in positive clinical outcomes 2 years

postoperatively, especially in older patients82. Drawbacks to microfracture include the

poor biomechanical nature of the resulting fibrocartilage repair tissue, incomplete filling

of the defect, and the potential for abnormal bone growth into the cartilage lesion76

Physical rehabilitation, typically with continuous passive motion and protected weight

bearing, is an important step in improving cartilage repair after microfracture, possibly

due to increased movement of synovial fluid throughout the joint space, which carries

nutrients to chondrocytes and enhances synthesis of matrix proteins83'84.

Subchondral drilling. Drilling into the subchondral bone is a technique

established by Pridie in 1959 to develop bleeding channels through the subchondral

bone to promote the formation of cartilage to resurface the exposed bone. It is reported

that 85% of patients whose knees were treated with this procedure demonstrated

satisfactory long-term outcomes76; however, in a rabbit model, the newly formed

cartilage repair tissue lost its hyaline-like morphology within 1 year of treatment and

instead resembled dense collagenous tissue85. Adverse effects of drilling include









damage to the subchondral bone from heat generated during the procedure and the

potential for formation of a subchondral hematoma.

Spongialization. Spongialization is a modification of debridement and drilling,

where the entire injured cartilage and the subchondral bone beneath it are removed,

which exposes the spongy, or cancellous, bone86'87. In contrast to subchondral drilling,

spongialization removes the highly innervated subchondral plate, thus removing a

source of pain, and it may promote improved healing. Follow-ups by Ficat, the

developer of this procedure, showed improved joint function and pain relief in 70-80% of

patients, but this procedure has not gained popularity, perhaps because it is invasive

and may cause thermal necrosis of the surrounding cells that are the target for

stimulation86,88

Restorative strategies

Restorative strategies for the joint include high tibial osteotomy (HTO), knee

replacement (either total or unicompartmental), and transplantation of bone, cartilage, or

tissue. Knee replacement procedures involve removal of bone from femoral and tibial

surfaces and resurfacing with prosthetic implants.

High tibial osteotomy. High tibial osteotomy (HTO) is a technique developed by

Jackson in 1958, and is used in patients who experience cartilage degradation in the

medial compartment of the knee. This procedure aims to transfer weight bearing from

damaged regions to those with healthy cartilage to relieve pain and prevent further

osteochondral damage1. Especially useful in younger, active patients who are not

ready for knee replacement, HTO shows satisfactory clinical outcomes in 80% of

patients 5 years post-op, and 60% at 10 years after treatment". Since HTO is only









useful in a select group of patients, knee replacement is often the technique of choice

for middle-aged and elderly patients since it has reliable long-term outcomes.

Partial knee replacement. With partial knee replacement, only the damaged

surfaces of the knee are replaced with prostheses while intact surfaces are not altered.

Typically, younger patients take advantage of partial knee replacement because it is

less invasive, has a lower cost and has a shorter recovery time than total knee

replacement surgery. Even though partial knee arthroplasty has a 98% survival rate at

10 year follow-up and delays the need for invasive total knee arthroplasty by 10 years or

more, the cartilage in surrounding areas progressively deteriorates due to osteoarthritis

as shown by radiographs of patients' knees.

Total knee replacement. When conservative treatments fail, total knee

replacement, or total knee arthroplasty (TKA), is recommended as an effective and

durable procedure to restore mobility and relieve pain in people suffering from end-

stage knee lesions1. TKA is the most invasive of all the procedures and can be

performed as revision surgery for each of the aforementioned procedures. Prostheses

consist of a femoral and tibial component and some also involve patellar resurfacing1.

The anterior cruciate ligament (ACL) is always removed and in some instances the

posterior cruciate ligament (PCL) is substituted or removed. For elderly patients, total

knee arthroplasty is currently the recommended technique to restore function of the

entire knee after articular cartilage damage". In patients greater than 70 years old, the

10 year follow-up success rate remains high, but in patients less than 50 years old, the

failure rate is higher, likely due to the longevity and vigor of joint use. The average









duration of implant survival remains 10-15 years, with some reports of 20 years. When

the implant degrades, revision surgery is necessary to restore joint mobility.

Reconstructive methods to treat cartilage defects

Reconstructive methods aim to fill the cartilage defect with autologous cells or

tissues, and surgeons attempt to carry them out in the least invasive method possible8.

These methods use either pieces of cartilage tissue (mosaicplasty) or autologous

chondrocytes harvested from non-load-bearing areas autologouss chondrocyte

transplantation)58. The main drawbacks are that there is a very limited supply of non-

weight-bearing cartilage tissue available for harvest, and collecting tissue introduces

new sites of damage to the articular cartilage of the joint58.

Osteochondral transplantation or mosaicplasty. Osteochondral transplantation

(osteochondral grafting) includes autologous and allograft transplants. Autologous

osteochondral transplantation (OATS), also termed mosaicplasty, consists of removal of

cartilage from the defect site down to the subchondral bone, followed by creation of

small holes 15 mm deep and perpendicular to the cartilage surface11'80'89. Next,

osteochondral grafts 10-15 mm long are harvested from non-weight-bearing surfaces of

the patient's articular cartilage and inserted into the donor sites. As the site heals,

fibrocartilage is the most prevalent repair tissue observed1.

While mosaicplasty promotes pain relief and joint function improvement, there is

risk of donor site morbidity, the original contour of the femoral condyle may be difficult to

recreate, and chondrocyte death can result in degeneration of the graft. Additionally,

there is a chance of bone and cartilage collapse. Removal of cartilage from non-load-

bearing areas of the joint may lead to defects at those donor sites, resulting in more

pain and additional cartilage damage from the lack of chondrocytes available for repair.









Allograft osteochondral transplantation is indicated for large osteochondral defects

that exceed those which can be repaired via mosaicplasty and other less invasive

techniques. Allografts replace the injured cartilage and its underlying subchondral bone

with a histocompatible fresh or fresh frozen cartilage segment taken from an organ

donor. Chondrocyte viability and biomechanical integrity deteriorate over time in

storage; therefore, fresh osteochondral allograft (less than 2 weeks old) is

recommended to optimize the success of the procedure1. Long-term follow-ups report

85% and 74% graft survival at 10 and 15 years post-operatively11.

Autologous chondrocyte transplantation. Autologous chondrocyte

transplantation (ACT), also dubbed autologous chondrocyte implantation (ACI), is a cell-

based technique first described by Brittberg in 1994, to relieve pain and restore function

in knees affected by either chondral or osteochondral defects11'90. Chondrocytes are

harvested arthroscopically from a non-weight-bearing portion of the joint and cultured

for approximately 6 weeks. After cell expansion in vitro, the cultured cells are surgically

introduced to the cartilage lesion of the patient 1'80

To prepare the defect, the site is debrided and covered by a periosteal flap

harvested from the patient's tibia or femur. The periosteal flap is sealed or glued to the

edge of the defect, and the cultured cells are delivered and held in place underneath the

flap. The cells in suspension eventually attach and secrete extracellular matrix. ACT is

useful for medium to large sized defects with positive results for up to 10 years90. It is

reported that in the majority of patients (80-90%), hyaline-like repair tissue forms and

pain relief and joint mobility improve89,91'92. Furthermore, integration of the newly formed

repair tissue with the surrounding cartilage is reported in nearly 90% of patients at five









year follow-up92. Additional studies show, however, a high proportion of fibrocartilage

present in the repair site, further demonstrating the difficulty of regenerating articular

cartilage93

Disadvantages of ACT include the extreme expense, leakage of chondrocytes

from beneath the periosteal flap at the recipient site, the need for two separate

surgeries for harvest and implantation, the risk of chondrocyte dedifferentiation in

monolayer cell culture, uneven distribution of cells introduced to the defect, and long

recovery time following the operation94. Further complications following ACT include

fibrous overgrowth of the periosteal flap and separation and detachment of the

periosteum from the repairing cartilage95.

An improved variation of ACT applied clinically is known as characterized

chondrocyte implantation (CCI), where patients' chondrocytes are sorted and selected

based on cell surface antigens that indicate greater potential to produce hyaline

cartilage. This method is currently marketed at ChondroCelect by Tigenix of Belgium11

Gene- and Cell-Based Approaches to Cartilage Repair

What we observe is that no matter the restorative or reconstructive procedure

used, the fibrocartilage repair tissue or the prosthesis eventually degrades. Nothing can

match the native articular cartilage originally produced by the patient. The goal of tissue

engineering, gene-, and cell-based therapies is to replicate the natural structure and

durability of articular cartilage to enable long-lasting repair of damaged cartilage. Ideal

strategies would restore the structural and functional elements of articular cartilage

through transplant of a viable cartilage substitute, revised methods to enhance cartilage

repair, or methods to prevent or halt cartilage damage.









Clearly there is a need for a cartilage repair strategy that introduces autologous,

multipotent cells to the joint in a way that enables them to integrate and survive within

the existing cartilage structure. MSCs appear as viable candidates as they are found

abundantly in tissues throughout the body22,23

MSCs: the logical cell type for chondrogenesis

Cells derived from mesenchymal tissues with the ability to proliferate extensively,

self-renew, and undergo multipotent differentiation are broadly defined as mesenchymal

stem cells, or MSCs. Experiments performed by Friedenstein and others described the

presence of mesenchymal stem cells (MSCs) in the bone marrow that could be isolated

through their intrinsic property to adhere to tissue culture plastic96'97. These early

observations led to the standard accepted assay used to identify MSCs, the colony-

forming unit-fibroblast (CFU-F) assay, which identifies adherent, spindle-shaped cells

that proliferate to form colonies.

There are no definitive surface markers for MSCs; however, they are frequently

positive for STRO-1, CD73, CD90, CD105, CD106, CD146, CD166 and negative for

CD45, CD11b, CD34, CD31, CD11798'99. Rather, they are characterized by their ease of

isolation and their rapid growth in vitro while maintaining their differentiation potential,

allowing for extensive culture expansion to obtain large quantities suitable for

therapeutic use. These properties make MSCs ideal building blocks for tissue

engineering efforts to regenerate tissues and repair structures damaged by arthritic

conditions and cartilage injuries100. MSCs offer advantages for cell therapy because

they are easier to culture and manipulate ex vivo than chondrocytes, which de-

differentiate when removed from their native ECM and grown in monolayer.









MSCs can be readily harvested from multiple tissue sources including bone

marrow, adipose tissue, synovial fat pad and periosteum101. Cells with stem-like

characteristics can also be harvested from skin, liver, skeletal muscle, dental pulp and

cartilage102'103,104. MSCs have been shown to differentiate into chondroprogenitors,

osteoblasts, adipocytes, myoblasts and hepatocytes in vitro'05. Mesenchymal stem cells

contribute to the regeneration of mesenchymal tissues throughout the body such as

bone, cartilage, muscle, ligament, tendon, adipose and marrow stromal06

MSCs play an important role in human development, growth, repair, regeneration,

and homeostasis. Their multilineage potential makes them a useful model to investigate

mechanisms of cartilage tissue development and regulation, especially following

treatment with proteins from the TGF-3 superfamily32. Few reports exist of human MSC

implantation for cartilage repair, but studies are taking place to evaluate the use of

MSCs rather than chondrocytes for repair of cartilage defects in the knee.

It is thought that with increasing age there is a significant decline in the abundance

of MSCs, their lifespan, and their differentiation potential. This is a logical assumption

since it has been observed that as chondrocytes age they generate smaller, less

functional ECM proteins and are less responsive to growth factors and mechanical

stimuli, although an equally poor repair response of cartilage to damage is observed

throughout embryonic, immature, and mature cartilage107'108. However, MSCs extracted

from the synovial fat pad have been shown to maintain their osteogenic differentiation

potential throughout life and could perhaps offer more chondrogenic potential than bone

marrow-derived MSCs109









MSCs provide an autologous cell source, eliminating much of the risk of disease

transmission and rejection of donor cells. There is also controversial evidence that

MSCs may have immunosuppressive effects110. MSCs can be prompted toward

chondrogenic differentiation to provide cells for direct delivery to articular cartilage

defects or to provide cells to seed scaffolds which are then implanted into defects.

When used in a gene therapy approach, MSCs can be genetically enhanced to express

specific growth and differentiation factors which could not only influence their

differentiation in a paracrine manner, but also stimulate neighboring cells present in the

cartilage implant site.

Stem cells can serve as vehicles for gene delivery to damaged articular cartilage

by transfecting cells with recombinant cDNAs encoding chondrogenic proteins and

growth factors52. Furthermore, viral transduction of MSCs with chondrogenic cDNAs can

result in longer expression of the gene product and greater potential to influence

cartilage defect repair.

Mechanical stimulation to promote cartilage

Mechanical stimulation is a natural component of the chondrocyte environment

and is known to affect gene expression and re-differentiation of chondrocytes58.

Chondrocytes, and other cell types such as MSCs, respond to dynamic compression by

changing their gene expression profile58. In therapeutic applications, it is important to

remember that controlled movement post-implantation could be essential for adequate

healing and development of articular cartilage rather than fibrocartilage.

Regeneration and Repair Using MSCs: Gene Therapy and Orthopaedics

MSCs are ideally suited for cartilage gene therapy applications as they can be

prompted to differentiate along a chondrogenic lineage by growth and differentiation









factors. Members of the TGF-3 superfamily play an integral role in cartilage

development; thus, they are logical choices for cartilage development in vitro. Gene

delivery using viral vectors enables rapid, robust expression of chondrogenic proteins in

MSCs, prompting their differentiation along a chondrogenic lineage. Such modified

MSCs can then be implanted into cartilage defects, where they will continue to express

the transgene products, thus impacting neighboring chondrocytes embedded in the

ECM. Differentiated MSCs can take the place of the limited number of chondrocytes

that are currently harvested for transplant procedures and will serve to fill cartilage

defects with chondroprogenitor cells which secrete cartilaginous matrix and repair the

site of damage with a natural tissue.

As the field of gene therapy expands, numerous approaches are being explored to

improve treatments for cancer, AIDS, hemophilia, cystic fibrosis and an array of other

diseases. Studies of gene therapy applications for orthopaedic conditions arose from

research directed toward the treatment of rheumatoid arthritis1'. Rather than a gene

therapy method to correct a genetic abnormality, this work was based upon the concept

of using gene transfer as a protein delivery system to treat chronic joint disease. By

delivering cDNAs encoding naturally-occurring anti-inflammatory or anti-arthritic proteins

to cells in the synovial lining, these cells would then serve as factories for the local

overproduction and secretion of the therapeutic proteins into the joint space and

surrounding tissues. Studies in several laboratories have shown that exogenous cDNAs

can be efficiently delivered and expressed at levels within the joint sufficient to have

beneficial effects in a variety of animal models112.









The success of gene therapy studies in rheumatoid arthritis has led to the

exploration of the application of gene delivery for other orthopaedic applications such as

bone,113,114,115 ligament, tendon,116,117 and cartilage44 healing. Most of these

investigations use gene delivery in the same manner, as a system for localized,

sustained production of bioactive molecules to promote healthy regeneration. Gene

transfer can be used as a means to achieve sustained synthesis of specific proteins

within a cartilaginous lesion, and this can be used to augment the differentiation of

mesenchymal stem cells toward chondrogenesis in vivo. Gene transfer can also be

used to stimulate existing cells and tissues, such as muscle, to repair large segmental

defects in bone118. A similar technique can be applied to repair cartilage defects with

chondrocytes (ACT) or MSCs that are stimulated by virally-delivered chondrogenic

transgenes.


















--. ...


A).






Articlar -" <- Superficial zone



m ar .- t o* p tii. C lm of p
S- Trnsitial zone


esdte e in th rI z--. Radial zone
St t' emal
pace b c i B Dp-Caltfied cartilage
A ...-.Subcnondral bone


B).

Figure 1-1. The four zones of articular cartilage stretch from the superficial surface to
the deep zone, where hypertrophic chondrocytes are replaced by calcified
matrix. A) Growth plate of proximal tibia. Columns of prehypertrophic
chondrocytes are present in the radial zone, progressing to enlarged, rounded
chondrocytes that undergo hypertrophy, followed by apoptosis and
replacement by calcified cartilage. B) Depiction of the zones of articular
cartilage with resident chondrocyte population. The collagen fibrils, depicted
as pale gray lines, differ in their orientation throughout the zones to allow
movement, provide structural support, limit the expansion of proteoglycans,
and absorb stresses. (Part B adapted from Ahmed and Hincke, Tiss Eng Part
B, 2010).










Anterior View


PaIeINY.-rniraI
wrmen Laeal 1ibloferrioral
Impr ieam nrt idria.
Synovialri .
membrane Menscus Medal
Tibia titbiAem iBi
S| compartment
I 'b



Figure 1-2. The human knee contains specialized structures to withstand the forces of
movement. Depicted here, articular cartilage lines the ends of the femur and
tibia and the inner patella to enable near frictionless motion of the knee joint.
The specialized architecture of articular cartilage can withstand repeated
extreme forces, often for more than 80 years, without damage. Image
adapted from Ahmed & Hincke, Tiss Eng Part B, 2010.


Medial View
























Figure 1-3. The arrangement of structures in the articular cartilage matrix is designed to
absorb forces, especially compression. Collagen fibers form a supportive
framework for cartilage, and they constrict the expansion of hydrated
proteoglycans. Chondrocytes (not shown), the resident cells of cartilage,
populate the matrix at a low density and serve to remodel the extracellular
matrix. Image adapted from Mow and Ratcliffe, Biomechanics of Diarthrodial
Joints, "Structure and Function of Articular Cartilage and Meniscus," 1990.












/ -Hyaluonic Acid

L ik Proteins
S- "^ -Kcaon Sulfole
etz/ir owbim Su/fale






'-- c ~Proteogfycan SuJunif






Figure 1-4. A proteoglycan aggregate is made up of many proteoglycan subunits
attached to a hyaluronic acid backbone via link protein. Proteoglycans consist
of a protein core with glycosaminoglycan side chains of keratan sulfate or
chondroitin sulfate attached. The negative charges of the side chains repel
one another and form their characteristic bottlebrush-like structure. Image
adapted from Mow and Ratcliffe, Biomechanics of Diarthrodial Joints,
"Structure and Function of Articular Cartilage and Meniscus," 1990.










T,r. -ll fibril


I N-teritirin




( -^ FleoibIl
kink
Collagen Fibers Type IX







50 nm


Figure 1-5. Fibrillar collagens, including collagen type II, form a characteristic triple
helix. Multiple tripeptides are bundled together to form fibrils, which in turn,
are packed into parallel bundles to yield fibers with incredible strength.
Collagen type IX, with its flexible kink, crosslinks collagen type II fibrils to
proteoglycans in the cartilage extracellular matrix. Figures are adapted from
Sigma-Aldrich Life Sciences, 2009, and Lodish, et al. Molecular Cell Biology,
1999.









Table 1-1. Collagen types.
Collagen type Representative Tissues References
Fibrillar collagens
I Abundant throughout human body; skin, bone, scar tissue, 2,119
fibrocartilage, tendon, ligaments, dentin, artery walls,
interstitial tissues
II Articular cartilage and vitreous humor of the eye 2,119,120

III Produced by young fibroblasts prior to type I collagen, 2
reticular fibers, skin, muscle, blood vessels, intestines,
uterus
V Interstitial tissues, similar to type I, also associated with 2,119
placenta (fetal tissues), synovial membranes and cell culture
XI Hyaline cartilage 119

Nonfibrillar collagens
IV Basal lamina, lens of the eye, part of filtration system of
capillaries and glomeruli of the kidney nephrons
VI Most interstitial tissues, similar to type I, blood vessels, skin, 119
intervertebral disc
VII Forms anchoring fibrils in dermal epidermal junctions

VIII mainly endothelial cells 121

IX Hyaline cartilage, associated with type II and XI fibrils in 121
cartilage, vitreous humor, FACIT collagen

X Hypertrophic and mineralizing cartilage, growth plate 119

XII Embryonic tendon and skin, periodontal ligament, FACIT 119
collagen
XIII Endothelial cells 119

XIV Fetal skin and tendon, FACIT collagen 119

Collagens are an essential part of the integumentary system. Of the 29 types of collagen
identified and described in literature, the fourteen most prevalent are listed above. Collagen type
I is the most abundant collagen in the human body. Collagen type II is the main component of
articular cartilage, and is supported by types IX and XI. A subgroup of nonfibrillar collagens,
including types IV and VIII, form sheets that create structures such as basement membranes
that surround tissues. FACIT, Fibril Associated Collagens with Interrupted Triple Helices, refers
to a type of collagen which is also a proteoglycan. FACIT collagens include types IX, XII, XIV,
XIX, and XXI.









Table 1-2. Morphogens and Growth Factors in Cartilage Development


Morphogen Name
Bone Morphogenetic Protein (BMP)
BMP-2


BMP-4


BMP-3

BMP-3B

BMP-5
BMP-6


BMP-7


BMP-8


BMP-9

BMP-10


Cartilage derived morphogenic protein
(CDMP)


Alternate Descriptors

BMP-2A


BMP-2B


GDF-10


OP-1
osteogenicc protein 1)

OP-2


GDF (Growth/
differentiation factor)


Function(s)

Cartilage and bone
morphogenesis, used
clinically for bone repair36
Cartilage and bone
morphogenesis
Bone formation, inhibits
activity of BMP-246
Membranous bone
formation
Bone morphogenesis
Cartilage hypertrophy, bone
formation via alternate
mechanism to BMP-2 or
BMP-4
Osteogenic differentiation,
used clinically to augment
bone repair36
Bone formation, esp. active
in early phase of fracture
healing46
Anabolic factor in juvenile
cartilage
Not chondrogenic; regulates
cardiac growth and heart
chamber maturation


GDF-5, BMP-14


GDF-6, BMP-13



GDF-7, BMP-12


Mesenchymal
condensation,
chondrogenesis
Cartilage development and
hypertrophy, cartilage
formation in vitro similar to
BMP-2
Ligament and tendon
development, cartilage
formation in vitro similar to
BMP-2


CDMP-1


CDMP-2



CDMP-3









Table 1-2. Continued.
Morphogen
Transcription Factor


Name


Sox9


Alternate Descriptors


SRY (sex-determining
region Y) HMG box 9


Function(s)


Binds regulatory sequences
in Col2al promoter region;
necessary for mesenchymal
condensation


Other Morphogens

Indian Cartilage morphogenesis;
hedgehog esp. in prehypertrophic
(Ihh) chondrocytes
Many growth factors, morphogens, and transcription factors play integral roles in cartilage
morphogenesis and are activated during the body's natural cartilage repair responses. Many of
these same factors, especially BMPs, are involved in bone morphogenesis and are delivered in
clinical applications to enhance non-union fracture healing.









CHAPTER 2
MATERIALS AND METHODS

In Vitro Cell Culture

HEK293 and 293FT Cell Culture

Immortalized cell lines were cultured in 75 cm2 flasks containing Dulbecco's

modified Eagle's medium (DMEM), with 10% FBS (Gibco), 1% glutamine (Gibco), and

1% penicillin-streptomycin (Gibco), hereafter referred to as complete DMEM, at 37C in

a 5% CO2 environment. For adenovirus propagation, HEK293 cells were grown to

approximately 70-80% density in 175 cm2 flasks prior to viral infection. For lentivirus

preparation, 293FT cells were cultured in either 75 cm2 or 175 cm2 flasks and treated as

indicated below.

Harvesting and Culturing Bovine MSCs

We harvested bone marrow from the long bones of 3-day old Holstein-Fresian bull

calves from the University of Florida Department of Large Animal Sciences. The bones

were cut open with a table saw, marrow was removed with a spatula and diluted, to

prevent clotting, in MSC medium, which contains DMEM (Gibco) with 10% MSC-

qualified FBS (Gibco), 1% glutamine (Gibco), and 1% penicillin-streptomycin (Gibco).

Addition of antimycotic agents to the medium were not beneficial as these substances

interfered with adherence, growth, and expansion of MSCs. The cells were cultured at

37C in a 5% CO2 environment for 48-72 hours, at which point blood cells were

removed by changing the media. The plastic-adherent cells that remained were grown

and expanded for up to five passages for use in experiments.









Aggregate (Pellet) Culture

The pellet culture system allows cell-cell interactions analogous to those that occur

in precartilage condensation events during embryonic development122. We used

aggregate culture as a means to evaluate the chondrogenic potential of multiple

transgenes in vitro. We adapted our "high-throughput" aggregate culture system based

upon results that showed the efficiency of 96-well plates over the use of individual 15

mL conical vials when screening numerous treatment groups123

As depicted in Figure 2-1, we transduced early passage bovine MSCs with

adenoviral vectors carrying the complete cDNAs for the candidate transgenes, denoted

Ad.BMP-2, Ad.BMP-4, Ad.BMP-7, Ad.lhh, Ad.Sox9, Ad.CTGF, and so on, at doses

ranging from 10 virus particles/cell to 100,000 vp/cell. Similar dose ranges were used

for scAAV-BMP-4 transduction. MSCs were grown in monolayer in 75 cm2 flasks and

transduced with virus. Twenty-four hours later, cells were trypsinized, counted, and 2.0

x 105 cells in a 300 to 350 pl volume were pelleted by centrifugation within individual

wells of 96-well V-bottom polypropylene plates (Corning). The resulting cellular

aggregates were grown for 21 days in serum-free chondrogenic medium.

Chondrogenesis was evaluated by gross examination of aggregate morphology,

histological staining for proteoglycan expression, and immunohistochemical analysis for

collagen types I, II, and X.

Chondrogenic Media Formulation

Cell aggregates in 96-well plates were maintained in 350 pl chondrogenic medium,

which consists of serum-free DMEM (Gibco), 1% ITS (insulin, transferring, selenium)

(Sigma), 1% penicillin-streptomycin (Gibco), dexamethasone (10-7 M), ascorbate-2-

phosphate (50 pg/mL) (Sigma), proline (40 pg/mL) (Sigma), and 1 mM sodium pyruvate









(Gibco). Media were changed every 48 to 72 hours, except for days 3, 7, 14, and 21,

when media were changed to allow 24-hour accumulation of proteins for detection via

ELISA.

Virus Preparation and Transgene Expression

Construction and Generation of Recombinant Adenoviruses Containing
Chondrogenic Transgenes

Serotype 5, El- E3-deleted recombinant adenoviruses containing BMP-2, BMP-4,

BMP-7, TGF-31, Indian hedgehog, Sox9, GFP, and others were generated through the

Cre-lox recombination system developed by Hardy124. Each transgene was inserted

directionally into the adenoviral shuttle plasmid, pAdlox, containing the 3' inverted

terminal repeat of the virus, a x packaging signal, a cDNA expression cassette driven

by the cytomegalovirus (CMV) promoter/enhancer, and a loxP Cre recombinase

recognition sequence. Cotransfection of Cre8 293 cells, which constitutively express

high levels of Cre recombinase, with linearized Adlox plasmid and Y5 adenoviral

genomic DNA flanked by loxP sites generates recombinant adenovirus124. Specifically,

Cre-mediated recombination occurs between the loxP site in pAdlox vector and the 3'

loxP site in the adenoviral backbone. Any nonrecombined y5 adenovirus that is present

can be separated from the recombinant adenoviral particles via subsequent propagation

on Cre8 293 cells, whose Cre recombinase will delete the packaging signal of Y5 virus.

Adenovirus Propagation and Amplification

To generate the quantities of replication-deficient adenovirus needed for large-

scale infections, 293 cells were grown in complete DMEM in 175 cm2 flasks. At 70-80%

confluence, the media were removed and the cells were rinsed with PBS. A small

aliquot (2-5 pl) of the desired adenovirus was added to a minimal volume of serum-free









media (10 mL) and applied to the cells for 4-6 hours, at which point 12-15 mL complete

DMEM was added. Cells were incubated at 37C with 5% C02 until cells rounded up

and developed a granular appearance (signs of the cytopathic effects of a lytic virus)

usually occurring after 2-3 days. Based on the method developed by Palmer125, cells

were collected using a cell scraper just before the virus caused them to lyse, then

transferred, along with the media, to a 50 mL centrifuge tube. Following tabletop

centrifugation at 2000g for 10 minutes at 4C, the media were discarded in bleach, and

the pellet was resuspended in 4 mL Tris, pH 7.0, and stored at -80C. Cells were lysed

by three rounds of freeze/thaw, digested with benzonase (Sigma) and the final

supernatant was collected and stored on ice until cesium chloride (CsCI) gradient

purification. CsCI gradients were prepared by layering approximately 3 mL of 1.4 g/mL

CsCI solution on the bottom, 3 mL of 1.2 g/mL CsCI solution in the middle, and 4 mL of

viral cell supernatant on the top layer in chilled ultracentrifuge tubes (Beckman). After

centrifugation at 40,000 g for 1 hour at 40C, viral bands localized at the interface of the

two CsCI layers. The viral band was harvested by puncturing the centrifuge tube with a

needle and syringe. If two bands were visible, the lowest band containing the infectious

particles was harvested. The harvested band was diluted 2- to 4-fold in 10mM Tris-HCI,

pH 8.0, for recentrifugation. Three consecutive gradients were performed on each viral

prep. After the third CsCI gradient purification, the harvested adenovirus fraction was

transferred to dialysis tubing and placed in dialysis buffer at 40C for 6 to 8 hours.

Following three rounds of dialysis, the virus was stored at -800C in 50 pl aliquots. Viral

titers were estimated by optical density125.









Generation of Lentivirus

We produced lentiviral vectors by implementing a four plasmid transfection

procedure adapted from Invitrogen's ViraPower expression system. Transducing

vectors expressing the desired transgenes (GFP and BMP-4) were generated via the

insertion of the specific cDNAs into the pLenti4/V5-DEST vector via homologous

recombination. The resulting expression plasmid was mixed with the three necessary

packaging plasmids, denoted pLP1, pLP2, and pLP-VSVG. Plasmid DNA was

completed with lipofectamine and delivered to monolayer cultures of 293FT cells.

Twenty-four hours following transfection, the cell culture medium was replaced.

Conditioned medium was harvested at 48 and 72 hours to allow collection of virus.

Medium was concentrated through ultracentrifugation at 20,000g at 40C for 2 hours. The

resulting pellet was resuspended in Opti-MEM (Gibco) and used immediately or

aliquotted and stored at -800C until use.

Construction and Generation of scAAV Vectors

For generation of scAAV vector plasmids, the cDNA encoding GFP was

directionally inserted into the conventional AAV packaging vector pTRUF2 as a Notl-

Sall fragment. For generation of scAAV vector plasmids, the cDNAs for GFP and BMP-

4 were directionally inserted into the Sacll and Notl sites of the pHpa-trs-SK plasmid.

PCR was used to modify BMP-4 cDNA to introduce Sacll and Notl sites at the 3' and 5'

ends. The insert plasmid contains the CMV promoter/enhancer and the cDNA of interest

surrounded by ITRs from AAV2. The pDG-2 plasmid contains the rep and cap genes

from AAV2 and complementing adenoviral functions required for amplification and

packaging of the AAV genome. Similarly, to generate serotype 5 scAAV, the pxyz-2

plasmid was used.









AAV vectors were propagated using an adenovirus-free, two plasmid transfection

system. Using 10-layer cell factories (Nunc), the respective AAV vector plasmids were

co-transfected into 293 cells by CaPO4 precipitation with the pDG-2 or pxyz-5

packaging/helper plasmid. Sixty hours post-transfection, cells were harvested with PBS

containing 10mM EDTA, pelleted, resuspended in buffer containing 150 mM NaCI and

50 mM Tris, and lysed by three successive rounds of freeze-thaw. Cellular nucleic acids

were digested by incubation with Benzonase (Sigma). Purification of AAV from the

crude lysate was performed over iodixanol gradients followed by FPLC affinity

chromatography over mono-Q columns. The eluate was desalted and concentrated with

a Millipore Biomax 100K filter, aliquotted and stored at -800C. Viral titers were

determined by competitive quantitative PCR assay relative to well-characterized AAV

viral reference standards. Each viral preparation was examined for purity by resolution

of the viral proteins by SDS PAGE and silver stain.

Gene Transfer to MSCs to Induce and Enhance Chondrogenesis

Plasmid DNA transfection

Monolayer cultures of bovine MSCs were transfected with DNA-lipofectamine

complexes. Liposomes containing each BMP-4 expression vector were generated by

incubation of DNA with lipofectamine (Invitrogen) in Opti-MEM (Gibco). Following a 20

minute incubation, the complexes were added to cells in a minimal volume of serum-

free medium. One to three hours later, complete medium was added to cells. Medium

was changed 24 hours post-transfection. At 48 hours post-transfection, conditioned

media were collected and BMP-4 expression was characterized by ELISA. GFP

expression was observed visually in the transfection controls.









Transgene delivery using adenovirus

MSCs are receptive to adenoviral transduction since they possess the CAR

receptor. Adenovirus readily infects dividing and non-dividing cells so this was not a

limiting factor in carrying out viral transduction prior to forming cell aggregates.

Methods to Detect Transgene Products

Western blot

Since there is no commercially available ELISA to quantify expression levels of Ihh

or Sox9, Western blots of conditioned media or cell lysates, respectively, were used to

verify expression. MSCs were grown in monolayer culture, then transduced with Ad.lhh

or Ad.Sox9 in a minimal volume of Opti-MEM (Gibco). Twenty-four hours later, media

were removed and replaced with serumless DMEM. After an additional 24 hours,

conditioned media was removed from Ad.lhh-treated wells for detection of protein

expression. Since Sox9 is expressed intracellularly, Ad.Sox9 infected cells were

harvested with a cell scraper and lysed in chilled homogenization buffer. The resulting

cell extract was used for protein detection. As these were human transgenes expressed

in bovine MSCs, we were able to distinguish between endogenous and exogenous

protein production.

A BCA assay was completed to gauge total protein content and ensure

consistency in loading samples. A total of 10 pg total protein per lane was loaded into

10% or 15% Tris-HCI pre-cast gels (Bio-Rad) for Sox9 and Ihh detection, respectively.

Proteins were transferred to Immun-Blot PVDF membranes (Bio-Rad) in buffer

containing 25mM Tris, 192 mM glycine with 20% methanol and 0.1% SDS. Following

transfer, the membranes were blocked with 5% milk for 1 hour prior to application of

primary antibodies: 1:2000 rabbit anti-Sox9 (Santa Cruz Biotechnlogy, Inc.) and









1:10,000 rat anti-lhh (R&D Systems). Membranes were soaked in primary antibody

solution overnight on a low speed orbital shaker at 4C. Following incubation, the

membranes were rinsed with TBS-T and secondary antibodies conjugated to

horseradish peroxidase were applied for 45 minutes: 1:15,000 anti-rabbit-HRP (Bio-

Rad) and 1:12,000 anti-rat-lgG-HRP (Sigma-Aldrich), respectively. Membranes were

rinsed and proteins were detected with the Immun-Star HRP Chemiluminscent Kit (Bio-

Rad). For detection of 3-actin as a loading control, antibodies were stripped from the

membrane using 0.5M NaOH for 20 minutes at room temperature. As before,

membranes were rinsed with TBS-T, and blocked with 5% milk for 45 to 60 minutes.

Mouse anti-p-actin-HRP antibody (Sigma-Aldrich) was applied to the membranes at a

concentration of 1:50,000 overnight at 4C. 3-actin was visualized via

chemiluminescence.

ELISA to detect secreted transgene products

Concentrations of secreted protein products present in conditioned media were

quantified using commercially available Duo-Set ELISA kits (R&D Systems) for BMP-2,

BMP-4, BMP-7, and IGF-1 as directed by the manufacturer. The conditioned

chondrogenic media from three or more replicate aggregates were used for all data

points. Unless otherwise noted, media were placed onto cells 24 hours prior to

collection, then used immediately or stored at -20C until use.

Histology and Immunohistochemistry

Aggregates were removed from culture after 21 days and fixed in 4%

paraformaldehyde solution overnight. The aggregates were paraffin embedded, cut into

5 pm sections, and mounted on plus-charged slides (Fisher Scientific). Slides were

deparaffinized and rehydrated through a series of xylenes and graded alcohols, after









which they were stored in water for 2 minutes. Appropriate sections were stained with

toluidine blue, while alternate sections were examined for collagen type I and type II

content. Initially, heat mediated antigen retrieval was performed in Dako Target

Retrieval Solution (DakoCytomation) for 20 minutes at 950C. While this procedure works

well for many tissue samples, the heat proved too intense for aggregates, and they

often became detached at this step. We omitted the heat retrieval and found that

aggregate sections remained attached for the entire procedure, and its omission did not

impact detection of collagen.

Samples were treated with chondroitinase ABC (Sigma-Aldrich) at 0.2 U/mL for 15

to 30 minutes at 37C to cleave chondroitin sulfate polysaccharide chains. Nonspecific

binding was blocked in 15% normal serum matched to the secondary antibody species.

Slides were incubated overnight at 4C or 1 hour at room temperature with

commercially available antibodies: rabbit anti-collagen I (Chemicon) at 1:1000 and

rabbit anti-collagen II (Chemicon) at 1:500. Although the primary species reactivity was

mouse, the antibodies were shown to cross-react with bovine tissues. The fluorescent

secondary antibody, Alexa Fluor 488 donkey anti-rabbit, was used at a 1:200 dilution

and allowed to incubate for 1 hour in the dark. Samples were washed with buffer (TBS-

T), counterstained with Dapi, and coverslipped using Vectashield mountant.

RNA Extraction, RT, and rtPCR

For each treatment, ten additional pellets were harvested at days 3, 7, and 21 for

total RNA isolation. Total RNA was isolated from treated and control cell pellets using

the RNeasy mini kit (Qiagen) as directed. Aggregates were stored in RNALater

(Qiagen) until RNA extraction was performed, at which time each group of treated

pellets were frozen in liquid nitrogen and pulverized using a mortar and pestle. The









pulverized tissue was added to lysis Buffer RLT, homogenized using a 20-gauge

needle, and RNA was harvested using RNeasy spin columns following the

manufacturer's protocol (Qiagen).

We completed semi-quantitative real time RT-PCR analyses to verify transgene

expression in cases where there was no ELISA for detection (Ihh) or chondrogenic gene

products of interest were intracellular (Sox9). Real time PCR was carried out using the

Eppendorf Realplex machine and software. To synthesize cDNA, 1 pg of total RNA from

each group was reverse transcribed using random hexamer primers and MMLV reverse

transcriptase (Invitrogen). Specific primer sets were used to amplify type II collagen al

chain, type I collagen a2, type X collagen al, aggrecan, osteopontin, and fibronectin.

In Vivo Experiments

Intra-Articular Injections

To examine the effects of chondrogenic transgene expression on collateral tissues

in the joint, we performed intra-articular injections in both knee joints of Male Wistar rats

weighing 100-150 g. The adenoviral vectors of interest were suspended in phosphate

buffered saline (PBS) to a 50 pl volume and injected into the joint space of the knee

through the infrapatellar ligament. The knee diameter was measured with calipers daily,

and rats were weighed daily for 7 days. All animal procedures were approved by the

Institutional Animal Care and Use Committee of the University of Florida.

Harvesting Tissue, Decalcification, and Histology

Seven days after intra-articular injection, joint tissues were harvested and stored in

5% formic acid overnight at low speed on an orbital shaker to speed decalcification.

Following a PBS rinse, the knees were stored in 0.5 M EDTA solution on a low speed

orbital shaker at 40C until bones were sufficiently decalcified (approximately one week).









The decalcified knees were sliced medially and cut into 5 pm sections for hematoxylin &

eosin staining.









Inflot wlth rombinant virus
Hrrvnt bone marrow from oontainhg gene of Interest
udlr Isolate and culture adherent
ilbroblast (MSCs)






Palt 2 x O cla Huvt at dwy21
CIUb in Hltology (toluine blue)
Culture 24 hr chonekgonic rmdum IHC (Col lyp" I rd 10




Figure 2-1. Assay for in vitro chondrogenesis. We harvested bone marrow from 1 to 3-
day old bull caves, cultured the plastic adherent cells and expanded them in
vitro. To generate cell aggregates, we virally transduced cells in monolayer
and 24 hours later, pelleted them by centrifugation. Cells were grown in
chondrogenic medium for 21 days, then evaluated for hypertrophic
differentiation, cellular proliferation, and matrix protein production.









CHAPTER 3
GENE DELIVERY STIMULATES CHONDROGENESIS OF MSCS

Introduction

Articular cartilage is a highly specialized tissue that allows for near frictionless

motion of diarthrodial joints. When cartilage is damaged as a result of injury or disease,

natural repair processes are often insufficient to regenerate the tissue due to the lack of

vascularity, dense extracellular matrix (ECM) and low cellular density of cartilage.

Typically the repair response, if any, generates a fibrocartilage scar which lacks the

unique architecture and structural properties of native articular cartilage. In most cases,

though, damage or lesions of a significant size remain permanently. These injuries often

initiate a degenerate cycle that over time leads to generalized cartilage loss and

osteoarthritis.

Biology offers a number of potential approaches to enhancing the natural repair

response of bone marrow progenitor cells in vivo. As shown in Table 1-2, several

growth factors, morphogens and more recently, transcription factors, have been shown

to promote differentiation along chondrogenic lineages. While these substances have

shown promise in animal models of cartilage repair and regeneration, their clinical

application is hindered by delivery problems. The half-lives of many proteins are limited

in vivo, so they are difficult to administer to sites of cartilage damage at therapeutic

concentrations for sustained periods of time. Localized delivery of these agents without

involvement of non-target organs has also proven to be problematic. These limitations

may be overcome by developing techniques to transfer genes encoding chondrogenic

gene products to cells at the appropriate sites and to express those genes locally for the









necessary period of time. In this manner, the proteins of interest are synthesized locally

by cells and are presented to the microenvironment in a natural fashion.

Numerous experimental approaches are currently being explored to enhance

cartilage regeneration and repair. These include tissue engineering and gene and cell-

based therapies. Adult mesenchymal stem cells (MSCs) are a well-suited cellular

platform on which to base therapies for cartilage repair and tissue regeneration since

they have the capacity to self renew and can differentiate into multiple mesenchymal

tissues, including cartilage and bone105'106. They are also readily available from a variety

of tissue sources, including bone marrow, synovium, fat, skin, and muscle. MSCs

maintain their multilineage capacity over several passages in culture, making them

amenable to various applications, including ex vivo therapies126

Within articular cartilage, the extracellular matrix constantly undergoes remodeling

by chondrocytes, which exclusively populate the matrix at low density. Adult MSCs in

vivo serve as replacements for differentiated cells of mesenchymal tissues that naturally

expire or succumb to injury or disease127. This process of native stem cell-generated

cell replacement peaks between ages 20 through 29 and decreases with age14'122, but it

could be re-charged by the introduction of modified MSCs to sites of cartilage damage.

To date, no repair strategy has been shown to generate a durable repair tissue that can

withstand the functional demands required of articular cartilage in vivo128

Delivery of MSCs alone is not sufficient to generate appropriate repair tissue in

cartilaginous lesions, as the microenvironment is not adequate to drive and maintain

chondrogenic differentiation. Gene transfer to MSCs, however, can be adapted to

achieve sustained local expression or synthesis of specific protein factors, which may









be used to induce chondrogenesis in vitro and may enhance chondrogenic

differentiation in vivo. If appropriate stimulatory factors are delivered to sites of cartilage

damage, it may be possible to trigger signaling pathways in resident and introduced

cells that drive the cells to synthesize repair tissue identical to the original in structure

and form. Growth factors which are delivered as gene products offer advantages over

recombinant protein delivery in that the proteins are presented in a natural context, they

can be synthesized locally at the site of need for extended periods of time, and may be

less costly.

We used an aggregate culture system to evaluate the ability of candidate

transgenes to stimulate bovine MSCs toward chondrogenesis in vitro. High-throughput

aggregate culture of MSCs offers a useful means to evaluate chondrogenic potential of

multiple transgenes in vitro. Using a similar method, Johnstone et al. demonstrated

chondrogenesis in MSC aggregates in defined medium containing dexamethasone and

TGF-31 resulting in aggregates that synthesized extracellular matrix characteristic of

articular cartilage, containing proteoglycan and type II collagen22

Growth factors such as TGF-3, BMP-2, BMP-4, BMP-7, IGF-1 and FGF-2 have

stimulatory effects on cartilage function. Animal studies document the benefits of

exogenous growth factors in stimulating MSCs and grafted chondrocytes toward

chondrogenesis129. Chondrogenic genes, such as Sox9, Ihh, and BMPs 2, 4, and 7 play

key roles in the development of the cartilage anlagen within the embryo and subsequent

formation of permanent cartilage, as seen in the joints, or transient cartilage that

undergoes replacement by skeletal elements elsewhere in the body. The method of









introducing modified MSCs to cartilage defects has considerable potential to improve

the cartilage repair process.

This study was performed to evaluate BMP-4, BMP-7, Sox9, Ihh and other genes

for their chondrogenic potential when expressed by MSCs as transgenes. While there

are numerous other factors that could be studied, our evaluation includes gene products

considered to be among the principle contributors to the process of chondrogenesis and

those reported to be effective in cartilage repair in animal models. These experiments

were used to identify the most effective gene and gene combinations for stimulating

chondrogenic differentiation of MSCs in vitro. Thus, the selected cDNAs should offer a

reasonable assessment of the potential utility of gene delivery in this application.

Rationale

Multiple growth factors, morphogens, and transcription factors, including TGF-31,

BMP-2, BMP-4, IGF-1, Ihh and Sox9, are known to play a role in chondrocyte

differentiation and proliferation. Chondrogenic effects of TGF-31, BMP-2, and IGF-1,

have been well-characterized in an aggregate culture system when delivered to MSCs

via adenoviral vectors130. Although TGF-31 and BMP-2 showed strong chondrogenic

activity, they tended to drive the cells toward hypertrophy, a preliminary step in bone

formation. With the goal of identifying transgenes whose products may be more suitable

for cartilage repair, we expressed BMP-4, BMP-7, Sox9, Ihh, IGF-1, and CTGF as

transgenes in MSCs and characterized their relative abilities to induce chondrogenesis

within the pellet culture system singly and in combination. The results of these studies

provide new insight into the biological activity of each cDNA when administered to

MSCs as a gene product, and they form the basis for selection of candidate transgenes

to evaluate in vivo.









Results

Our group has shown previously that MSCs in culture are highly amenable to

infection and subsequent transduction with recombinant type 5 adenovirus vectors. For

the majority of transgenes tested, adenoviral-mediated delivery to MSCs yielded robust

expression of protein products that typically persisted for two to three weeks. The

efficiency of adenovirus gene transfer coupled with the relative ease of generating novel

recombinants and the ability to readily propagate the vector to high titers, led us to use

this technology to evaluate the chondrogenic activity of our candidate transgenes.

Cre-lox Recombination and Adenovirus Propagation

We obtained the cDNAs for BMP-4, BMP-7, Ihh, Sox9 and CTGF from the

American Type Culture Collection (ATCC) as IMAGE clones. Following amplification of

each associated vector, we isolated and directionally inserted each cDNA into the

pAdlox shuttle vector. This plasmid contains the 3' inverted terminal repeat of the

adenovirus, a native y packaging signal, and a cDNA expression cassette driven by the

cytomegalovirus promoter/enhancer followed by a loxP Cre-recombinase recognition

sequence. Following verification of each pAdlox construct by diagnostic restriction

digestion, we transfected each into cultures of 293 cells to assay for synthesis of the

respective transgene products. Forty-eight hours post-transfection the conditioned

media or cell lysates, as appropriate, were analyzed qualitatively for expression of the

respective transgene products: BMPs -4 and -7 by ELISA; Ihh, Sox9, and CTGF by

Western blot (data not shown).

Once it was confirmed that the vector constructs were indeed functional, they were

used to generate recombinant adenovirus. For this, the respective plasmids were

linearized by restriction digest and co-transfected with purified, y5 adenoviral DNA into









cultures of 293 cells engineered to constitutively express the Cre recombinase (Cre8

cells). (The y5 adenoviral genome packaging sequence (\) is flanked by loxP sites:

novel recombinant adenovirus is generated by Cre-mediated recombination between

the loxP site in the Adlox shuttle vector and the 3' loxP site in the Y5 viral backbone.) If

the far left hand portion of the viral genome contained in the Adlox plasmid was

successfully linked to the adenoviral backbone, plaques in the monolayer of cells were

usually detectable within 5-7 days. After this point the culture medium was no longer

changed, allowing the virus released from the lytic plaques to accumulate and infect

neighboring cells, generating widespread cytopathic effects in each culture. The cells

and medium were then collected, and following successive rounds of freeze-thaw, the

lysates were used to infect new cultures of Cre8 cells. Since the packaging signal of the

y5 adenoviral backbone is removed by the Cre recombinase, the Cre8 cell line is non-

permissive for its replication. Therefore, each new vector construct was passage at

least three times in Cre8 cells to eliminate any contaminating Y5 virus that may have

been propagated during growth and amplification of the new vector.

To determine the purity of each new adenovirus preparation, the genomic DNA

was isolated, digested with appropriate restriction enzymes and analyzed following

electrophoresis in 0.7% agarose gels. Once the vectors were found to be free of

detectable y5, subsequent amplifications were performed in 293 cells. In an effort to

eliminate as much contaminating cellular DNA and debris as reasonably possible,

adenoviral preparations were banded over three successive CsCI gradients and were

then dialyzed against multiple changes of dialysis buffer prior to aliquotting and storage.









Isolation of MSCs

We elected to use the bone marrow of bovine calves as a source of MSCs, as it

would provide several advantages for our studies. Foremost is that the long bones of

these animals, at birth, are at least 5 to 10 times greater in size and volume than the

corresponding bones in the adult rabbit, which is among the largest of the common

experimental animals. This large size enabled the isolation of an abundance of primary

MSCs (usually > 5 x 107) from a single bone (generally the head of the femur), requiring

minimal expansion of the cells in culture prior to their use in experiments. Further, since

male calves are routinely sold from the dairy farm soon after birth, we could make use

of these animals and in the process obtain cells from a newborn animal. In this regard,

the cells would be expected to have maximal proliferative and differentiation potential,

and thereby provide a robust readout of chondrogenic stimulation in our in vitro assays.

The use of these animals was facilitated by the Large Animal Sciences Department

whose facilities on the University of Florida campus are in close proximity to the Health

Sciences Center.

Following procurement immediately after the death of the animal, the femurs were

transected with a band saw. The red marrow was removed from the epiphyseal ends of

the femurs and tibiae, and both red and yellow marrow was scooped from the medullary

cavity with a spatula. The semi-solid marrow was placed in MSC medium (as described

in Chapter 2) and passed through a syringe several times to disaggregate the loose

stroma. Following digestion of the mixture with collagenases types I and II and neutral

protease (Worthington Biochemical Corporation) for 30 minutes at 370C, the cells were

passed through a 40 pm nylon cell strainer (BD Falcon) and plated. After the first

preparation, we found that the enzymatic digestion step was not necessary; the cells









would adhere to the culture flasks as long as the stroma was sufficiently homogenized

by hand. Although somewhat variable between preparations, typically enough of the

disaggregated marrow was obtained to seed 5 to 6 175 cm2 flasks, which would be ~30-

40% confluent with adherent fibroblastic cells at 24 to 48 hrs. Cells were allowed to

adhere for 72 hours prior to the first change in media, which removed all of the non-

adherent blood cells. Following expansion to confluence, the cells were either used

immediately for experimentation or were aliquotted for storage in liquid nitrogen.

Gene-Mediated Chondrogenesis of MSCs

Our studies of gene-induced chondrogenesis were performed using high density

aggregate culture systems. In early experiments we followed the methods of Palmer et

al. whereby approximately 2.0 x 105 MSCs suspended in culture medium were

aliquotted into 15 mL conical vials (Corning). During centrifugation at low speed, the

cells were forced into aggregates, forming pellets in the bottom of each tube. Each cell

pellet was cultured in 750 pl of chondrogenic medium supplemented with

dexamethasone in its individual 15 mL tube for the following 21 days. Media were

changed every 48 hours, with additional changes to allow for 24-hour conditioned media

collection at days 3, 7, 14, and 21. While this method proved useful for a small number

of samples, it proved unmanageable for the simultaneous culture of multiple treatment

groups, each with multiple replicates. The frequent handling of large numbers of tubes

and their individual screw caps was unwieldy, leaving the cultures prone to fungal

contamination. Supplementation of the culture medium with antimycotic agents was of

no benefit as these appeared to interfere with cell-cell adhesion and inhibited

chondrogenic differentiation.









To provide a method amenable to the analysis of large numbers of samples, we

worked to develop a "high-throughput" in vitro chondrogenesis system whereby

numerous MSC aggregates could be formed and cultured simultaneously in multi-well

plates. In this technique, the MSCs were genetically modified in monolayer, detached

using trypsin 24 hrs later, suspended in a minimal volume of culture medium and then

counted. Approximately 2.0 x 105 cells were then delivered in a 300 pl volume of

chondrogenic medium (described in Chapter 2) to individual wells of a V-bottom 96-well

plate. Once the plate was loaded with the samples of interest, using appropriate plate

adaptors, it was spun at 40C in a table top centrifuge at 500g for 5 minutes to form cell

aggregates in the bottom of each well. The wells were topped off with an additional 100

pl of chondrogenic media and grown at 370C. Over the course of 24 hours, the cells

formed rounded aggregates that did not adhere to the polypropylene plate. After 24

hours, aggregates were disturbed by pipetting the media vigorously. This provided a

means to observe whether all cells were incorporated into each aggregate rather than

disassociating upon movement. The medium for each sample was changed routinely,

as before, and the cells were cultured for 21 days.

Although the maintenance of the tiny pellets remained a somewhat challenging

and tedious procedure, we found this approach to be far more manageable and

provided more consistent results than the use of 15 mL tubes. We found, however, that

certain parameters were essential to its application. For example, useful aggregates

were generated only with the use of Corning brand, polypropylene, V-bottom, 96 well

plates. We tested similar plates from other manufacturers, such as Nunc, but were









unable to consistently generate usable pellets following centrifugation since the "V" of

their well was a slightly different angle.

Adenoviral-Mediated Delivery of BMP-4, BMP-7, Ihh, and Sox9 Drives
Chondrogenic Differentiation of MSCs

Having developed suitable techniques for the isolation of low passage MSCs and

their use in high-throughput in vitro chondrogenesis assays, we used these methods to

evaluate the relative chondrogenic activity of BMP-4, BMP-7, Ihh, Sox9 and CTGF

following adenoviral-mediated gene delivery.

Initially we focused our efforts on Ad.BMP-4 and Ad.BMP-7, since the

chondrogenic potential of the recombinant proteins has been demonstrated in several

reports. First or second passage bovine MSCs were seeded in monolayer and

expanded to ~90% confluence. To ensure that the full potential of each cDNA was

represented in our assays, we infected separate monolayer cultures with each

adenoviral vector over a range of doses spanning 10 to 105 vp/cell. The genetically

modified cells were harvested 24 hours later, and ~2.0 x 105 cells were seeded per well

for each vector and dose. The cells were pelleted by centrifugation and cultured in

defined medium containing dexamethasone for 21 days. The levels of the secreted

transgene products were measured in conditioned medium over the course of the

experiment. At the end of the incubation period, the cell pellets were harvested and

processed for histology and immunohistochemistry or were pooled and used for

isolation of RNA for subsequent analysis by qRT-PCR. Chondrogenesis was evaluated

by examination of aggregate morphology, histological staining for proteoglycan content

and immunohistochemical analysis for production of collagen types I, II, and X.









For both Ad.BMP-4 and Ad.BMP-7, the aggregates formed from MSCs infected at

the 105 vp/cell dose were not viable and disintegrated within 3 to 7 days of culture (data

not shown). As shown in Figure 3-1, ELISA measurements of the respective transgene

products produced by the remaining cell pellets between the two vectors reflected

considerably different levels of protein production and response ranges. In general,

peak production occurred near days 3 and 7, and gradually tapered over the 21 day

incubation period. Transgenic expression was largely dose dependent for both Ad.BMP-

4 and Ad.BMP-7; however, the cell pellets transduced with Ad.BMP-7 showed a

somewhat higher level of protein production relative to viral dose (Fig. 3-1B).

In response to Ad.BMP-4, the cells receiving the 10, 100 and 1000 vp/cell doses

produced peak expression levels of ~0.8, 2.0 and 7.0 ng/mL, respectively (Fig. 3-1A).

While the 1000 vp dose showed the greatest initial levels of expression, it also showed

the most precipitous reduction over time to less than 2 ng/mL. Interestingly, the pellets

infected at the 10,000 vp/cell dose produced somewhat less BMP-4 than the 1000 vp

dose throughout, and expression was more variable. BMP-4 was not detected in the

conditioned medium of the control pellets at any time.

Transgene expression from the Ad.BMP-7-infected pellets at day 3 ranged from

less than 1 ng/mL for the 10 vp/cell dose to over 50 ng/mL for the 10,000 vp/cell dose

(Fig. 3-1B). With the exception of the day 7 time point where the 1000 vp/cell dose

showed a jump in expression from 32 ng/mL to nearly 70 ng/mL, expression levels

showed a modest but gradual reduction in expression.

Histologic examination indicated distinct evidence of transgene-induced

chondrogenesis of the MSCs stimulated with BMP-4 and BMP-7. However, we found









the level of chondrogenesis to be reproducibly greater and more consistent when BMP-

4 was supplied as a transgene. As seen in Figure 3-3, relative to controls, MSC pellets

infected with either 10, 100 or 1000 vp/cell of Ad.BMP-4, expressing between 0.8 and

8.0 ng/mL of the transgene product, were highly cellular and showed positive staining

for toluidine blue and corresponding positive immunostaining for type II collagen,

characteristic of articular cartilage matrix. Pellets infected at 100 and 1000 vp/cell

doses, expressing 2-8 ng/mL BMP-4 over the 21 days produced a more dense, uniform

matrix populated with rounded chondrocytic cells in lacunae, morphologically similar to

the transitional zone of articular cartilage (Fig. 1-1B). Pellets infected at the 10,000

vp/cell dose were much smaller and appeared as loose cell aggregates with no

evidence of extracellular matrix production or cellular differentiation. This is consistent

with previous findings that excess viral load can have toxic effects, which would account

for the reduced expression of the BMP-4 transgene and lack of a biological response in

these pellets.

For the pellets transduced with Ad.BMP-7, pronounced chondrogenesis was

observed only in the pellets infected at the 1000 vp/cell dose, which produced between

30 and 70 ng/mL of BMP-7. As shown in Figure 3-4, these pellets synthesized a dense,

uniform extracellular matrix enriched for proteoglycans and collagen type II. Rounded

chondrocytic cells were evident; however, many appeared to have begun to advance to

hypertrophy, indicated by their increased cytoplasmic volume. MSC pellets infected at

the lower doses, producing less than 10 ng/mL of BMP-7, showed minimal

metachromatic staining with toluidine blue and little evidence of cellular differentiation or

cartilage ECM synthesis. The pellets infected at the 10,000 vp/cell dose were largely









similar to those infected at the same dose with Ad.BMP-4, and despite continued

production of BMP-7 of between 30-50 ng/mL over the 21 days, showed no evidence of

chondrogenesis.

Following these experiments, we wanted to adopt a similar strategy to evaluate the

chondrogenic effects of adenoviral mediated gene transfer of Ihh and Sox9 to MSCs.

Unfortunately, since Sox9 is an intracellular transcription factor and commercial ELISAs

were not available for Ihh, it would not be possible quantify transgene products

synthesized by the pellets over time. Therefore, to verify effective gene delivery to

MSCs and gain insight into the relationship between viral dose and the level of

transgenic expression, we performed Western blot analyses of MSC cultures infected

with Ad.Sox9 or Ad.lhh at 3-fold dose increments between 10 and 10,000 vp/cell.

Cultures of uninfected MSCs were processed in parallel and used as negative controls.

Based on the negative results obtained above with extremely high vector doses, we set

10,000 vp/cell as the upper limit for infection.

As seen in Figure 3-2, detectable human Sox9 protein was found in lysates of cells

infected with as few as 30 vp/cell of Ad.Sox9, with peak expression associated with

doses of 100 and 300 vp/cell. Analogous to the expression data for Ad.BMP-4, a

marked reduction in Sox9 production was seen at 1000 vp/cell and higher doses, such

that it was below the limit of detection in cells infected with 3000 and 10,000 vp/cell.

Western blots of media conditioned by MSCs infected with Ad.Ihh also showed

dose dependent expression, but with a slightly different profile. Protein bands were

faintly visible in the 300 vp/cell dose lane increasing in intensity with viral dose to a

maximum at 3000 vp/cell (Fig. 3-2). At the 10,000 vp/cell dose, however, there was no









visible protein band. The dramatic reductions in transgene expression seen with high

vector doses were again consistent with toxicity from excessive viral load.

Having established a working relationship between viral dose and transgene

expression for Ad.Sox9 and Ad.Ihh, we infected additional cultures of MSCs with each

vector over a similar dose range. As described above, we then seeded the genetically

modified cells into aggregate culture and assayed for chondrogenic induction 21 days

later.

Histologic analysis of the recovered cell pellets showed that both Ihh and Sox9

were capable of driving chondrogenesis when expressed as transgenes; however

significant differences were noted in their stimulatory capacities. For MSCs infected with

Ad.Ihh, pellets formed from cells at the 300 and 1000 vp/cell doses demonstrated the

most robust chondrogenesis (Fig. 3-5). The majority of the cells in each had a rounded,

mature, chondrocytic phenotype and produced large quantities of ECM enriched for

proteoglycans and collagen type II, with low collagen type I content. Although

chondrogenesis was readily discernable in pellets at the 3000 vp/cell dose, they were

noticeably smaller, with less abundant matrix and reduced cellularity. Aggregates

formed from MSCs at the 10,000 vp/cell dose never increased in size, became

fragmented within 1 to 2 weeks and disassociated by day 21. Aggregates formed from

MSCs infected at doses below 100 vp/cell gradually diminished in size and at the end

the incubation period were too small to paraffin embed.

Aggregates formed from MSCs infected with Ad.Sox9 showed the greatest

response to the transgene at the 100 vp/cell to 1000 vp/cell range (Fig. 3-6). The

volume and cellularity of the aggregates was considerably greater than those infected at









lower doses and the uninfected controls. The pellets showed robust extracellular matrix

production with increased proteoglycans (indicated by metachromatic toluidine blue

staining) and increased collagen type II content, consistent with that of articular

cartilage. Distinct from the aggregates expressing the other growth factors, the

morphology of the cells appeared considerably less hypertrophic. The cells (and their

surrounding lacunae) were noticeably smaller in diameter, and contained less

cytoplasmic volume. While aggregates from the 10,000 vp/cell dose also showed

pronounced synthesis of cartilaginous matrix, these pellets were smaller in diameter.

During endochondral bone growth, the synthesis of collagen type X by growth

plate chondrocytes, in conjunction with increased cellular volume, is indicative of their

maturation to a terminal hypertrophic state. From this, the cells undergo apoptosis, and

the residual cartilage matrix is replaced by bone. The morphology of the chondrogenic

cells in the pellets expressing BMP-4, BMP-7 and Ihh suggests that prolonged

stimulation with these factors has driven the differentiation of the MSCs toward an early

hypertrophic phenotype, more so than overexpression of Sox9. To further examine this,

we used immunohistochemistry to stain for the presence of collagen type X in

chondrogenic sections from Ad.BMP-4 and Ad.Sox9 infected pellets. As shown in

Figure 3-7, strong collagen type X staining is seen throughout the outer regions of the

Ad.BMP-4 pellet, while the Ad.Sox9 pellets show little positive staining over

background.

Adenoviral Delivery of CTGF Stimulates Proliferation of MSC Aggregates

In an effort to identify additional cDNAs with chondrogenic potential that might be

of use in cartilage repair, we also tested the effect of adenoviral mediated expression of

connective tissue growth factor (CTGF) in MSCs in aggregate culture. Similar to TGF-









11, CTGF stimulates fibroblast proliferation, differentiation and extracellular matrix

synthesis, and is important for chondrocyte differentiation and maturation131

As above, early passage of MSC cultures were infected with a range of doses of

Ad.CTGF and placed into aggregate culture for 3 weeks. As with the other transgenes,

the aggregates infected with certain vector doses visibly increased in size relative to

controls, during incubation. Histologic analysis showed that the increase in pellet

volume could be attributed to intense cellular proliferation, which was particularly

evident at the 1000 vp/cell dose (Fig. 3-8). However, despite a clear biological response

to the transgene product, there was no apparent chondrogenic differentiation or

cartilage matrix protein production. At higher levels of Ad.CTGF infection, such as those

at 10,000 vp/cell and higher, the cells did not proliferate; the aggregates remained

small, and they fragmented after 21 days of culture.

Combinations of Sox9 with BMPs Induced Chondrogenesis

The process of chondrogenesis in vivo is complex, requiring the orchestrated

expression and interplay of numerous growth and transcription factors. Using gene

delivery we have shown that overexpression of certain of these proteins individually is

sufficient to initiate chondrogenic differentiation of mesenchymal progenitors in vitro.

The resulting phenotype of the cells and their synthesized matrix demonstrate key

hallmarks of articular cartilage, but fall short of the architecture of the native tissue.

Indeed it is likely that successful regeneration of articular cartilage will require the

delivery of multiple factors and temporal regulation of their expression. Toward this

direction we initiated preliminary studies of the effects of gene combinations on in vitro

chondrogenesis. For these experiments, we co-expressed the most potent of the

chondroinductive genes we have tested thus far, BMP-4 and BMP-2 (characterized in









earlier studies in our group), with the chondrogenic transcription factor Sox9. We

hypothesized that the enhanced expression of Sox9 would help to stabilize the

chondrocyte phenotype and block progression to hypertrophy, significantly improving

the quality of the resulting cartilaginous tissue.

In exploring gene combinations, we infected MSCs in monolayer with pairs of

vectors at doses that individually produced a robust chondrogenic response: i.e.

Ad.BMP-4 at 100 and 1000 vp/cell; and Ad.BMP-2 at 1000 vp/cell were mixed with

Ad.Sox9 at 100 or 1000 vp/cell. We then seeded the genetically modified cells into

aggregate culture and analyzed the effects of co-expression of the respective

transgenes on chondrogenesis in our in vitro assay.

During the incubation period, the majority of the cell aggregates showed a steady

increase in diameter, and none of the gene combinations caused the pellets to

reproducibly disintegrate. Histological analysis showed enhanced production of cartilage

matrix components in pellets from each of the various treatment groups, relative to

controls, indicating that the transgene products were functionally expressed. However,

none of the gene combinations appeared to enhance chondrogenesis in vitro relative to

single gene delivery.

While the results of these experiments were somewhat negative, they serve to

emphasize the complexity of cell signaling pathways and cellular differentiation. The

adenoviral vector, with its CMV-promoter driven expression cassette, is designed to

provide extraordinarily high levels of transgenic expression, and likely serves to

continually saturate ligand-specific surface receptors and downstream signaling

pathways. Effectively modulating the effects of such a powerful stimulus may prove to









be challenging, requiring far more sophisticated gene delivery approaches than the

crude systems used here.

Wistar Rat Responses to Adenoviral Transgene Delivery

When designing a gene-based strategy for cartilage repair, the selection of an

appropriate transgene is fundamental to the overall success of the procedure-its

efficacy and its safety. For most applications, it is probable that gene delivery vectors,

their secreted transgene products, or genetically modified cells will emigrate from the

repair site to adjacent tissues. Because many of the agents considered to have the

greatest chondrogenic activity are pleiotropic, with broad stimulatory activities in various

tissues and cell types, it is critical to thoroughly understand the possible impact of their

constitutive, local overexpression on the joint tissues and the health of the prospective

patient.

Using our in vitro assay we have identified several proteins (BMP-4, BMP-7, Ihh,

Sox9) with the capacity to drive mesenchymal chondrogenesis when expressed in

MSCs as transgenes. In earlier related studies we found that the cDNA for BMP-2 was

also highly effective. Having established the relative chondrogenic activity of these

genes in vitro, we wanted to determine their potential to stimulate pathologic side effects

when over-expressed intra-articularly. For these in vivo studies, we selected the knee

of the rat as our animal model since the rat is the smallest common experimental animal

with a defined joint space that can be reliably targeted by intra-articular injection.

We delivered the adenoviral vectors containing BMP-2, BMP-4, BMP-7, Sox9 and

Ihh bilaterally at doses of 5 x 108 and 2 x 109 vp/pl to the knees of Wistar rats (2 rats per

vector and dose). Parallel groups of rats were similarly injected with Ad.GFP to control

for the effects of adenoviral delivery. At 7 days post injection the animals were killed,









and the knees were harvested, decalcified and processed for histologic analysis. (The

vector doses used in these experiments were based on our previous experience with

adenoviral gene transfer to joints. They were intended to provide sufficient transgene

expression to enable a robust depiction of the stimulatory properties of each transgene

product intra-articularly without endangering the overall health of the animals.)

H&E stained sections of the knees receiving Ad.GFP showed a mild leukocytic

infiltration and a slight thickening of the synovial intima, typical of an adenoviral injection

into the joint. Otherwise, the morphology of the tissues was normal. In stark contrast,

adenoviral delivery of the various chondrogenic cDNAs elicited a wide variety of

biological responses. The most dramatic effects occurred in the joints injected with

Ad.BMP-2. The representative field shown in Figure 3-10 shows massive hypertrophy of

the synovium and capsular tissues that completely displaced the adipose layer that

normally supports the synovial lining. The expanded tissue was fibrotic in many areas

and populated with spindled fibroblasts. About half of the tissue mass, (particularly

those areas proximal to articular cartilage), was strikingly chondrogenic and was heavily

populated with chondrocytes and chondroblastic cells. The joints receiving Ad.BMP-4

and Ad.BMP-7 showed a dramatic hypertrophy of the synovial lining, caused by

extensive chondrometaplasia throughout the subsynovium. There was little if any fibrotic

component to the expanded tissues; the increased volume was comprised almost

exclusively of chondroblastic cells.

Distinct from overexpression of the BMPs, adenoviral delivery of Ihh induced a

mild to moderate fibrosis broadly across the synovial lining, marked by an increase in

both collagen fibers and fibroblastic cells. Ad.Sox9 was associated with a moderate









synovitis throughout the lining, with increased numbers of synovial fibroblasts and

infiltrating leukocyes relative to the Ad.GFP control.

Discussion

Our experiments demonstrate that adenoviral-mediated delivery of chondrogenic

growth factors to MSCs serves as an effective method to induce chondrogenic

differentiation of these cells in aggregate culture. In comparing multiple growth factor

cDNAs delivered across a range of adenoviral doses, we found that delivery of ~100 to

1000 vp/cell provided the most effective levels of transgene expression. Of the factors

we evaluated, BMP-4 induced the most robust chondrogenesis, with aggregates

exhibiting a dense extracellular matrix populated with chondrocytic cells resembling

those found in deep articular cartilage. BMP-7 and Indian hedgehog exhibited less

desirable chondrogenic responses; while aggregates showed extracellular matrix

production and the presence of chondrocytic cells, these secreted factors were prone to

drive cells toward terminal differentiation and hypertrophy. Sox9, while less prone to

induce hypertrophic differentiation, induced production of cartilaginous matrix and

maintained the cells at a prehypertrophic state, as shown with collagen type II

expression and the absence of collagen type X. CTGF induced proliferation of the cells

at the 1000 vp/cell dose but failed to promote chondrogenesis in our system.

Members of our lab have shown previously that gene delivery of BMP-2 and its

expression in the 10-100 ng/mL range is required to achieve optimal chondrogenic

differentiation of MSCs. Here, using a similar system, we have shown that BMP-4

production of <10 ng/mL is sufficient to promote robust chondrogenic differentiation of

the cells. These data indicate that BMP-4 is highly potent, and it induces chondrogenic

responses equal to those observed at higher ng/mL levels of BMP-2 production.









Overall, in pellets modified to express secreted growth factors, we observed more

substantial chondrocytic differentiation than in those expressing the intracellular factor

Sox9. With secreted factors, even though a proportion of MSCs within the pellet express

the transgene, they have both autocrine and paracrine effects, which may enhance

chondrogenesis throughout neighboring cells within the aggregate. Conversely,

transgene products that function intracellularly may have advantages for use in studies

in vivo because the risk of exposing cells in neighboring tissues such as the synovium,

meniscus, tendon, and muscle to pleiotropic growth factors will be significantly reduced.

Chondrogenesis in vivo is a complex process involving the coordinate interplay of

numerous factors. As such, it is probably unrealistic to expect complete articular

cartilage regeneration via transfer and expression of a single gene. Based upon

preliminary work demonstrating enhanced chondrogenesis when IGF-1 was delivered to

MSCs in combination with TGF-31 or BMPs, we opted to similarly test our battery of

transgenes in combinations. As all potential combinations of chondrogenic cDNAs and

doses are too numerous to test feasibly, we used the information gathered about each

single transgene to deduce the most effective strategy. We hypothesized that delivery of

Sox9, which is expressed intracellularly, combined with the secreted proteins, BMP-2 or

BMP-4, would provide an effective combination to drive chondrogenic differentiation in

MSCs while maintaining cells in a pre-hypertrophic state. Although the combinations we

evaluated did not significantly alter chondrogenesis, they illustrate the complexity of

modulating cell signaling.

Although each transgene tested induced measurable chondrogenic effects on

bovine MSCs in aggregate culture, the microenvironment of the joint space in vivo is









quite different. Intra-articular delivery of Ad.BMP-2, -4, -7, Ad.Ihh, and Ad.Sox9 to the

intact knees of Wistar rats enabled us to evaluate potential side effects of ectopic

growth factor expression. BMP-2, -4 and -7 prompted cellular expansion and ectopic

cartilage formation to varying degrees while Ihh caused a weak fibrosis across the

synovial lining and Sox9 induced mild inflammation. Although these rat knees likely

demonstrate the worst case scenario of potential side effects of gene enhanced

cartilage repair, they provide a vivid representation of the potency of the various

transgene products and their capacity to impact the biology of the articular connective

tissues. Conversely, they likewise demonstrate the exquisite sensitivity of synovial

fibroblasts to proliferate and differentiate following stimulation with certain BMPs.

Methods designed to induce cellular differentiation in cartilage and bone repair in vivo

should take into account the high capacity for toxic side effects in adjacent tissues.

Although adenoviral vectors have a reputation for causing adverse immune

responses in vivo, this vector system may have a future in ex vivo approaches to tissue

repair. With their ability to readily infect MSCs and induce a high level of gene

expression, adenoviral vectors can serve as efficient protein delivery vehicles to modify

cells that are introduced to sites of tissue damage. Such a method will bypass the

potential for host immune responses while allowing for potent expression of

chondrogenic cDNAs.















- -- 10 vp/cell
---100 vp/cell
--1000 vp/cell
-A-- 10,000 vp/cell
control


Day3 Day7 Dayl4 Day21
Time (d)


- -.. 10 vp/cell
- 100 vp/cell
---1000 vp/cell
-A- 10,000 vp/cell
-X control


Day 3


Day 7


Day 14


Day 21


B). Time (d)


Figure 3-1. ELISAs of conditioned media from MSC aggregates transduced with BMP-4
and BMP-7 indicated protein production in response to viral doses
administered. Triplicate samples of conditioned media were obtained at days
3, 7, 14, and 21 following a 24 hour incubation, and protein levels were
quantified by ELISA. Data are shown as the mean SEM. A) BMP-4 ELISA
indicates a dose-dependent response following Ad.BMP-4 transduction. The
decrease in protein production following the 10,000 vp/cell dose indicates
toxicity in response to excessive adenovirus. B) BMP-7 ELISA shows protein
production in response to Ad.BMP-7 transduction at higher concentrations
than those observed with BMP-4.


80

70

_ 60
E
50

' 40

m 30

20


I\ _


S- 7













a. a >S

Indian
hedgehog,

Beta actin -- -



Beta actin l*o I .


Figure 3-2. Western blotting confirmed transgene expression following adenoviral gene
delivery to MSCs. As shown in the top panel, Ad.lhh was delivered at a range
of doses to bovine MSCs in monolayer, in a minimal volume of media. Media
were replaced 24 hours later, and harvested at 48 hours post-infection for
detection of Ihh protein. The presence of bands at 300, 1000, and 3000
vp/cell doses correlated with the matrix synthesis and chondrogenic
differentiation observed with toluidine blue staining of Ihh aggregates at these
doses (Fig. 3-5). Following detection of Ihh, antibodies were stripped from the
membrane, and 3-actin was detected as a loading control. In the bottom
panel, Ad.Sox9 was delivered at a range of doses, and 48 hours post-
transduction, cells were harvested with a cell scraper and lysed with chilled
homogenization buffer. Because the transcription factor Sox9 remains
intracellular, cell lysates were used to detect Sox9 production. As with Ihh, the
range in which Sox9 bands were visible correlated to the dose range in which
proteoglycan synthesis was most robust as shown through toluidine blue
staining (Fig. 3-6). Again, 3-actin was used as a loading control.













Toluidine
blue




Collagen
type II




Collagen
type I


Figure 3-3. MSC aggregates expressing BMP-4 were evaluated for chondrogenesis
through toluidine blue staining of proteoglycans and immunohistochemistry
for collagen types I and II. Aggregates infected with 10, 100, or 1000 vp/cell
were highly cellular, the proteoglycans of their synthesized matrix stained
deep purple with toluidine blue, and they showed positive immunostaining for
collagen type II, as shown by green fluorescence. The 10,000 vp/cell dose
resulted in fragmented pellets with little to no cellular differentiation or matrix
production. All aggregates were negative for collagen type I production.













Toluidineue '' s






Collagen
type II





Collagen
type I




Figure 3-4. MSC aggregates infected with Ad.BMP-7 over a range of doses yielded
pronounced chondrogenesis at the 1000 vp/cell dose only. Doses of 10 and
100 vp/cell showed minimal toluidine blue staining and poor collagen type II
expression. Aggregates receiving 1000 vp/cell synthesized dense, uniform
extracellular matrix enriched for proteoglycans and collagen type II. Rounded,
chondrocytic cells were present in the matrix, and many appeared to advance
toward hypertrophy, with increased cytoplasmic volume as shown above.
Pellets receiving 10,000 vp/cell showed no evidence of matrix protein
production or chondrogenic differentiation. The high dose of adenoviral vector
proved toxic. All aggregates were negative for collagen type I production.










300 vp/cell 1000 vp/cell


blue




Collagen
type II




Collagen
type I


Figure 3-5. Indian hedgehog aggregates expressed proteoglycan and collagen type II
following adenoviral gene delivery. Indian hedgehog (Ihh) expression in
pellets after 21 days in culture revealed a dose-dependent response.
Aggregates in the range of 300 to 1000 vp/cell demonstrated the most robust
chondrogenesis by forming large, rounded aggregates that secreted
proteoglycans, forming a cartilage matrix that contained chondrocytic cells.
MSCs receiving doses of 100 vp/cell or less or 10,000 vp/cell or greater
formed fragmented aggregates that did not proliferate, were smaller than the
non-transduced controls (shown in Figure 3-3), and disassociated by the 21-
day timepoint. All aggregates were negative for collagen type I.


100 vp/cell


3000 vplcell










100 vp/cell 1000 vp/cell


Toluidine
Blue





Collagen
type II





Collagen
type I



Figure 3-6. Sox9 induced cellular proliferation and matrix synthesis with less
differentiation than other transgenes tested. Sox9 expression in MSCs
resulted in aggregates of a uniform round shape consisting of proteoglycan
matrix and prehypertrophic-like cells. Similar to the other transgenes tested,
the most chondrogenic effects were observed after delivery of 100 to 1000
vp/cell. Doses above and below this range continued to promote collagen
type II expression, yet they were insufficient to promote robust cellular
proliferation. All aggregates were negative for collagen type I production.


10,000 vp/cell


10 vp/cell











BMP-4 1000 vD/cell


Collagen
type X







Figure 3-7. Collagen type X staining of Sox9 and BMP-4 aggregates. BMP-4
aggregates express collagen type X and undergo differentiation toward a
terminal hypertrophic state, shown through green fluorescence. Sox9 pellets
show no signs of collagen type X, with staining equivalent to background
levels, indicating their maintenance at a prehypertrophic state of
differentiation. 20x magnification.


Sox9 1000 volcell









Negative control


Figure 3-8. CTGF expression in MSC aggregates promotes cellular proliferation but no
chondrogenic differentiation. Delivery of 1000 vp/cell of Ad.CTGF resulted in cellular
proliferation and large aggregates; however, toluidine blue staining showed no evidence
of proteoglycan matrix production or chondrogenic differentiation. Likewise, aggregates
were negative for collagen type II production (lack of green fluorescence) when
evaluated immunohistochemically. Cells receiving viral doses of 10 and 100 vp/cell
formed aggregates that were too small to paraffin embed. Viral doses of 10,000 vp/cell
or greater caused aggregates to fragment and disassociate in culture. 10x
magnification.


Toluidine blue













BMP-4 and Sox9
controls



D E F


Sox9 + BMP-4
combinations



G H


BmP-2 control
Soxg + BMP-2
combinations





Figure 3-9. Sox9 in combination with BMP-4 and BMP-2 in MSC aggregate culture had
no additive effects upon chondrogenesis. Panels A, B, C, and G are single
doses, D-F, H-I are combinations. Toluidine blue staining shows matrix
synthesis at all doses; however, combinations resulted in smaller aggregates
that showed less matrix synthesis and less chondrogenic differentiation. A)
Ad.BMP-4 100 vp/cell, B) Ad.BMP-4 1000 vp/cell, C) Ad.Sox9 1000 vp/cell,
D) Ad.Sox9 1000 vp/cell + Ad.BMP-4 1000 vp/cell, E) Ad.Sox9 100 vp/cell +
Ad.BMP-4 1000 vp/cell, F) Ad.Sox9 100 vp/cell + Ad.BMP-4 100 vp/cell, G)
Ad.BMP-2 1000 vp/cell, H) Ad.Sox9 1000 vp/cell + Ad.BMP-2 1000 vp/cell, I)
Ad.Sox9 100 vp/cell + Ad.BMP-2 1000 vp/cell. Images are displayed at 10x
magnification.









control BMP-2 BMP-4









BMP-7 Indian hedgehog Sox9










Figure 3-10. Wistar rat knees exhibited varying responses to adenoviral delivery of
pleiotropic transgenes. Adenoviral vectors containing GFP and the
chondrogenic cDNAs BMP-2, BMP-4, BMP-7, Ihh, and Sox9 were injected
intra-articularly into both knee joints of healthy male Wistar rats at doses of 5
x 108 and 2 x109 vp/pl. Effects from BMPs ranged from the massive synovial
hypertrophy, fibrosis, and ectopic chondrogenesis resulting from Ad.BMP-2 to
moderate synovial hypertrophy and chondrometaplasia with little fibrosis from
Ad.BMP-4 and Ad.BMP-7. Indian hedgehog (Ihh) induced synovial fibrosis
and Sox9 was associated with a mild synovitis, comparable to that observed
in the Ad.GFP control animals.









CHAPTER 4
COMPARISON OF THE EFFECTS OF ADENOVIRAL, LENTIVIRAL, AND AAV
TRANSGENE DELIVERY TO MESENCHYMAL STEM CELLS

Introduction

Our studies have focused on the genetic modification of MSCs for use in the repair

and regeneration of articular cartilage. We have found that adenoviral-based vectors

efficiently transduce these cells, and readily enable the expression of transgene

products at biologically relevant levels. Indeed, adenoviral-mediated delivery and

expression of certain growth factor cDNAs in MSCs can be used to effectively direct the

differentiation of these cells along chondrogenic pathways. While highly useful tools for

studies in vitro and in experimental animals, the most widely used generations of this

vector still contain the majority of the native adenoviral genes. Low-level expression of

these genes in transduced cells is frequently associated with inflammatory reactions in

vivo and elimination of transduced cells by cytoxic T cells in immunologically competent

hosts.

The relative ease of use of the adenoviral vector system, coupled with its highly

efficient gene transfer, provided us with a valuable tool to screen a battery of candidate

cDNAs for their chondrogenic potential. From this screen we have identified several that

merit evaluation in vivo in cartilage repair models. While the adenovirus has been the

workhorse of our in vitro studies, its immunogenic profile excludes it from use in humans

and may negatively impact the process of tissue repair. In an effort to identify a gene

delivery system more suitable for use in animal models and possible clinical application,

we evaluated the utility of the well-characterized vector systems currently available,

(lentivirus, adeno-associated virus and non-viral transfection) for their capacity to effect

gene-mediated chondrogenesis of MSCs.









For these studies, we first inserted the cDNA for BMP-4 into AAV, lentivirus and

plasmid-based vectors. We then determined the capacity of each vector to deliver the

BMP-4 transgene to bovine MSCs and stimulate chondrogenesis in the aggregate

culture system. The BMP-4 cDNA was selected for these studies as it appeared to have

the greatest chondrogenic activity in adenovirally-transduced aggregates. Thus, the

ability to functionally deliver this cDNA would set the minimum standard for efficacy.

Results

Plasmid DNA Transfection of MSCs Results in Transient Expression

To generate the plasmids needed to produce recombinant viruses, we inserted the

cDNA for human BMP-4 into specific CMV-promoter driven vector constructs: AAV

(pHPA-trs-sk-BMP-4), adenovirus (pAdlox-BMP-4), lentivirus (pCDH-puro-BMP-4). We

then used these plasmids in DNA transfections to gauge expression resulting from

nonviral gene transfer. Plasmid DNA transfections of BMP-4 cDNA were carried out in

24-well plates of bovine MSCs grown to approximately 80% confluence. DNA was

diluted in Opti-MEM (Gibco) and mixed with lipofectamine (Invitrogen) at a ratio of lpg

of DNA to 2.5 pl of lipofectamine. DNA:lipofectamine complexes were added drop-wise

to cells and allowed to incubate for 24 hours. At 48 and 72 hours post-transfection,

conditioned media were collected for quantification of BMP-4 by ELISA. As shown in

Figure 4-1, transfections with each plasmid construct yielded BMP-4 in the 1.0 ng/mL

range, which was significantly higher than background levels from control cells.

However, transgene expression was transient. ELISAs of conditioned media at 72 hours

following transfection showed concentrations of BMP-4 near background levels (data

not shown). Such low levels of short-lived expression were not sufficient to drive

chondrogenesis in our aggregate culture assay.


100









Lentiviral Transduction of MSCs Proves Challenging

Preliminary data showing robust GFP expression in 293 cells following lentiviral

transduction established a basis for the use of lentivirus to deliver chondrogenic

transgenes, such as BMP-4, to bovine MSCs in monolayer (Fig. 4-2). To produce

recombinant lentivirus, we implemented a four plasmid transfection procedure adapted

from the ViraPower expression system (Invitrogen). Transducing vectors expressing the

desired transgenes (GFP and BMP-4) were generated via the insertion of the specific

cDNAs into the pLenti4/V5-DEST vector via homologous recombination. The resulting

expression plasmid was mixed with the three necessary packaging plasmids, denoted

pLP1, pLP2, and pLP-VSVG. Plasmid DNA was completed with lipofectamine and

delivered to monolayer cultures of 293FT cells. Despite numerous attempts, we were

unable to generate sufficiently high viral titers with the ViraPower system. In an effort to

improve viral production, we then evaluated the plasmid, pCDH-puro, for its potential to

generate appreciable titers of lentivirus.

Lentivirus generated using the pCDH-puro-BMP-4 construct induced BMP-4

expression in monolayer MSCs (Fig. 4-3), but again the viral titers generated were

insufficient to induce chondrogenesis within our pellet culture system. In an effort to

enhance this, we selected for cells that expressed BMP-4 using puromycin. We cultured

the cells under selection in preparation for subsequent expansion and use in aggregate

culture; however, the combination of BMP-4 expression and puromycin in the cell

culture medium appeared to inhibit MSC growth and proliferation, rendering the cells

unsuitable for aggregate culture. Despite repeated attempts with each system, we were

unable to produce reasonable titers for use directly without the need for cell selection.


101









Adenoviral-Mediated Delivery of BMP-4 to MSCs Induces Chondrogenesis

As discussed extensively in Chapter 3, bovine MSCs were transduced in

monolayer with an El, E3-deleted serotype 5 adenoviral vector carrying human BMP-4

cDNA and grown in aggregate culture in our in vitro chondrogenesis assay. Histologic

evaluation of aggregates after 21 days in culture showed that Ad.BMP-4 effectively

induced chondrogenesis in MSC aggregates as indicated by staining for matrix

proteoglycans and type II collagen. Aggregates infected with 100 to 1000 vp/cell were

large, round, and produced a dense, uniform matrix populated with rounded

chondrocytic cells in lacunae, morphologically similar to the transitional or deep zones

of articular cartilage (Fig. 3-3).

Self-Complementary AAV-Mediated Delivery of BMP-4 to MSCs is Comparable to
Adenovirus

Numerous AAV capsid serotypes are available for cross-packaging of vector

genomes, and each one changes the tropism of the virus. In an effort to identify the

most effective AAV serotype to target bovine MSCs, we packaged scAAV vector

containing GFP into serotypes 1, 2, 5, and 8. Following administration of 104 vg/cell to

MSCs in monolayer, transduction efficiency was determined visually by GFP

expression. As shown in Figure 4-4, after 24 hours, GFP expression in transduced cells

was evident for serotypes 2 and 5, and the intensity increased over the course of 7

days. By day 7, all 4 serotypes exhibited detectable GFP expression, but type 2 was by

far the most robust, followed by type 5 (Fig. 4-4). The order of efficacy for AAV

transduction of MSCs was AAV2>5>1 >8.

Based upon these data, we packaged the scAAV vector plasmids containing BMP-

4 (pHPA-trs-sk-BMP-4) to generate AAV serotypes 2 and 5 for use in our in vitro


102









chondrogenesis assay. We infected MSCs over a 3-fold range of doses from 102-104

vg/cell. After 21 days in culture, the pellets were evaluated histologically with toluidine

blue staining for proteoglycans. As shown in Figure 4-5, consistent with data from GFP

expression, infection with scAAV2-BMP-4 induced the most complete chondrogenic

response observed to date. The 3000 vg/cell dose yielded aggregates that were large,

round, highly cellular, and strikingly uniform in appearance. Aggregates showed robust

proteoglycan synthesis with chondrocytic cells uniformly distributed throughout.

Aggregates receiving the 1000 vg/cell dose exhibited similarly intense matrix

production, yet the periphery of these pellets was differentiated. Conversely, scAAV5-

BMP-4 delivered across the same range of doses produced aggregates with little

apparent chondrogenic differentiation but showed evidence of enhanced proteoglycans

synthesis.

Discussion

Using our pellet culture system, we compared the ability of recombinant lentivirus,

AAV, adenovirus and plasmid transfection to deliver and functionally express the BMP-4

cDNA in bovine MSCs, as indicated by chondrogenic differentiation in vitro. The

purpose of these studies was to identify from among the most developed gene transfer

systems, the one most suitable for use in vivo in studies of cartilage repair. As shown in

Table 4-1, each vector system offers advantages for gene therapy with certain

limitations for human applications.

In our hands, plasmid transfection and lentiviral transduction were unsatisfactory

for our in vitro chondrogenesis assay. While DNA transfection of MSCs provided BMP-4

expression near the minimal functional level (~1 ng/mL), it persisted for just over 48 hrs,

which was insufficient to induce a meaningful response. With the lentivirus, technical


103









difficulties prevented us from reproducibly generating vectors at usable titer. While we

had intended to evaluate these systems based on transgene expression profiles, in the

case of lentivirus, vector production proved too cumbersome to be practical for our

purposes.

Recombinant AAV is emerging as the vector of choice for human gene therapy

applications because of its perceived safety, as the transduced cells do not express

viral genes. In order to improve the efficiency of AAV transduction in our investigations,

we utilized self-complementary AAV vectors. In a study completed by Kay et al.,

scAAV2 transduced synovial fibroblasts far more efficiently than conventional single-

stranded AAV vectors, and expression levels of the hlL-1 Ra transgene were similar to

that noted previously for adenovirus and lentivirus vectors132

Self-complementary AAV mediated gene delivery of BMP-4 induced a strikingly

robust and uniform chondrogenic response, generating cartilaginous pellets qualitatively

superior to any other method. If these results hold, they indicate that AAV is clearly the

optimal vector system for this type of application. Our goal was to identify a system that

could provide functional transgene expression but was less immunogenic than

adenovirus. In most gene transfer applications, relative to AAV, adenoviral vectors

generate considerably higher levels of transgene expression. Based on this, we

expected that AAV would provide only borderline functional levels of expression, forcing

a trade-off between efficacy and low immunogenicity. That AAV may be capable of

mediating more effective transgenic expression and a more favorable immune profile

than adenovirus represents a significant finding.


104









At this point we do not know what may have contributed to the enhanced

chondrogenic response provided by scAAV BMP-4. It is possible that it was fortuitous,

caused by changes in culture conditions beyond our awareness. Alternatively, it is

possible that the differences in morphology between pellets genetically modified with

adenovirus and AAV can be attributed to the nature of their vector genomes. The

adenoviral vector genome is about 35 kb in length and contains nearly all of the native

viral genes. The scAAV vector is ~2.5 kb and is comprised only of an expression

cassette flanked by small DNA hairpins. In our chondrogenic assays we infected MSCs

in culture with a range of adenoviral vector particles. Pellets receiving the highest viral

doses frequently showed toxic effects reflected by reduced transgene expression or

loss of viability. At lower doses these effects may be less pronounced, but still

negatively influence the biology of the infected cells, resulting in pellets whose

morphology is asymmetric.


105














1.4 BMP-4



0.8
0.6
o 0.4
0
0 0.2
0
non-transfected pHPA-trs-sk-GFP pHPA-trs-sk- Adlox-BMP4 pCDH1-BMP4
control BMP4
24-hr conditioned media, serumless


Figure 4-1. ELISA of conditioned media demonstrates expression of BMP-4 as
measured 48 hours post-transfection. Bovine MSCs in monolayer were
transfected with lipofectamine:cDNA complexes as directed by the
manufacturer (Invitrogen). BMP-4 expression from MSCs transfected with
AAV (pHPA-trs-sk), adenovirus (Adlox), and lentivirus (pCDH1) plasmid
constructs carrying human BMP-4 cDNA was significantly higher than non-
transfected control cells; however, this expression was transient. After 72
hours, BMP-4 expression matched that of negative controls (not shown).


106



























a minimal volume of serumless media. A) brightfield B) fluorescence, 20x
magnification. This demonstrates our ability to generate lentivirus expressing
GFP. This virus was then applied to bovine MSCs to gauge expression of
lentivirus in primary cells prior to generating lentivirus carrying BMP-4 cDNA.



8 I
BMP-4
7 -

E6
5.
4
3




0 '





Figure 4-3. LV-BMP-4 expression was significantly less than Ad.BMP-4 expression in
bovine MSCs. Bovine MSCs were grown to 70% confluence in monolayer
prior to infection with virus in a minimal volume of media. 48 hours after
infection, BMP-4 production was measured from 24-hour conditioned media.
While Ad.BMP-4 equivalent to 1000 vp/cell yielded 6 ng/mL of BMP-4,
lentivirus yielded less than 0.5 ng/mL. Data are shown as mean SEM.
8:;, ,
BMPn
7,
-J .
26,b
,C)
Fiue 42 F xrsini 9 cls7 or fe netinwt 5tIo VGPi










r-,
0














lentivirus yielded less than 0.5 ng/mk. Data are shown as mean SEM.


107









Dav 3 Dav 5


Type 1




Type 2




Type 5




Type 8



Figure 4-4. AAV serotypes 1, 2, 5, and 8 were screened for transduction efficiency on
low passage bovine MSCs in monolayer. Self-complementary AAV-CMV-GFP
was delivered at a dose of 104 vg/cell, and transduction efficiency was
evaluated visually with GFP. Types 2 and 5 were most effective, as shown by
robust GFP expression beginning at day 1 and increasing in intensity over the
course of 7 days. Serotypes 1 and 8 were less effective at the same dose.
We chose to use serotypes 2 and 5 for subsequent AAV constructs carrying
BMP-4.


108


Dav 1


Dav 7










1000 vg/cell


scAAV2
BMP-4









scAAV5
BMP-4






Figure 4-5. MSCs transduced with scAAV-BMP-4 serotypes 2 and 5 demonstrate
matrix protein synthesis in aggregate culture. Toluidine blue staining shows
that scAAV2-BMP-4 promotes prehypertrophic differentiation in MSC
aggregates while scAAV5-BMP-4-aggregates show limited chondrogenic
differentiation but substantial proteoglycan production.


109


3000 vg/cell









Table 4-1. Comparison of gene delivery vectors reveals pros and cons


Vector
Adenovirus


Lentivirus





scAAV


Plasmid DNA


Advantages
Highly efficient gene transfer
Readily infects MSCs
Rapid, robust expression
Long-term expression not required for
chondrogenesis to occur
Useful as a protein delivery system

Infects MSCs
Potential for sustained, long-term expression
of growth factors
Can select for cells that express protein
products

Readily infects MSCs over a range of doses
Effectively delivers chondrogenic transgenes
Regarded as safe for clinical applications


Ease of production
Stability
Safety


Disadvantages
Expression tapers off after 21d
Hyperplasia in synovium when virus is injected
freely-need to control delivery for ex vivo
applications


Difficult to achieve high viral titers
Selection is required
Selection conditions are inhibitory to cell expansion
and subsequent use in chondrogenesis assay


Potential for less efficient gene delivery than other
viral vectors
Pre-existing humoral immunity to many serotypes


Inefficient gene transfer
Transient expression
Cannot readily apply a dose range


Comparison of gene delivery vectors reveals advantages and disadvantages of each system, with adenovirus and AAV
offering the most potential for safe, efficient delivery and expression of chondrogenic cDNAs.


110









CHAPTER 5
SUMMARY AND FUTURE DIRECTIONS

The unique architecture of articular cartilage limits its ability for self-repair, and

since cartilage is frequently damaged by traumatic injury or disease, enhancing

methods to repair damaged cartilage is of great clinical relevance. Tissue engineering

techniques along with stem cell and gene-based therapies have the potential to improve

cartilage repair and may eventually eliminate the need for invasive surgical procedures

and total knee replacement.

Tissue engineering techniques and stem cell therapies hold great promise for

improvements to articular cartilage repair, especially in the knee joint. Current cell-

based surgical treatments employ the use of healthy chondrocytes harvested from non-

weight-bearing areas of cartilage. Chondrocytes are available in very limited quantities

and require additional time for expansion in vitro prior to introduction to the cartilage

lesion. Autologous mesenchymal stem cells, on the other hand, can be harvested in

large quantities from bone marrow, fat, muscle and other tissues, and they can be virally

transduced to express chondrogenic transgenes. Modified MSCs introduced to sites of

cartilage damage may prompt the generation of functional hyaline-like cartilage rather

than a fibrocartilage scar.

In order to apply these ex vivo techniques, however, a number of challenges must

be overcome. These include further work on identifying the optimal combination of stem

cells, scaffolds and growth factors, and refining the conditions to enhance cell

expansion and chondrogenesis in vitro and integration of the cells and scaffolds with

existing cartilage in vivo. Future research efforts may focus on biodegradable scaffolds


111









laden with cells that will promote chondrogenesis and integrate at the wound margins to

yield fully functional articular cartilage rather than fibrocartilage.

The mode of delivery of cells to repair articular cartilage depends upon the size of

the defect. Small, localized defects may be repaired by direct application of modified

MSCs to the site, similar to delivery of chondrocytes in current chondrocyte transplant

procedures, whereas larger cartilage lesions would rely upon scaffolds to fill the defect

and hold the cells in place. Modified MSCs that express growth factors and other

chondrogenic proteins will provide the appropriate cues to initiate repair responses

within cartilage. Caution must be applied, though, as induction of certain signaling

pathways may trigger osteogenic differentiation.

Surgical attempts to repair damaged cartilage mentioned in Chapter 1 include

autologous chondrocyte transplantation (ACT), which is limited to treating trauma, and

is unable to repair large cartilage full-thickness defects. Gene-based therapies, on the

other hand, deliver transgene products to injured tissue to catalyze a healing response

without the need for surgery. A current phase 1 clinical trial uses a modified version of

ACT to initiate cartilage repair by infecting primary autologous chondrocytes with a virus

modified to express TGF-P1. The modified cells are cultured and expanded in vitro, then

injected into the knee joint of patients with degenerative arthritis (TissueGene,

Gaithersburg, MD). This type of approach bypasses the risk of inflammatory responses

to adenovirus, and it ensures that only the desired cell type is exposed to the potent,

chondroinductive transgene.

In Chapter 3, we set out to evaluate the ability of candidate chondrogenic

transgenes to stimulate MSCs toward chondrogenic differentiation. Based upon findings


112









that TGF-31 stimulates local proliferation and chondrogenic differentiation in

mesenchymal progenitor cells, we designed our study to include members of the TGF-3

superfamily as well as other developmental morphogens and transcription factors

(Table 1-1). These factors included BMPs 2, 4, and 7, Indian hedgehog, and Sox9 as

well as connective tissue growth factor (CTGF). Of all transgenes evaluated in our

aggregate culture system, we found that BMP-4 expression resulting from doses of

Ad.BMP-4 of 100 to 1000 vp/cell led to the most robust chondrogenic response. When

Ad.BMP-4 was injected intra-articularly in rat knees, however, this secreted factor

promoted dramatic cellular expansion and chondrometaplasia of the synovial layer.

Because such potent chondrogenic inducers can impact the highly receptive progenitor

cells of the synovial lining, it remains important to consider transgenes that are

expressed intracellularly, such as Sox9. Although we selected BMP-4 for our studies in

Chapter 4, Sox9 remains an important transgene for consideration in developing

cartilage repair models, especially in preventing ectopic cartilage formation while still

promoting defect healing.

Our evaluation of viral vectors in Chapter 4 enabled us to determine the optimal

vector for delivery of chondrogenic transgenes-in this case BMP-4-to MSCs. We

discovered that scAAV2-BMP-4 infects bovine MSCs optimally, and when administered

to MSCs at 3000 vg/cell, it drives BMP-4 expression that promotes more robust

chondrogenic differentiation and matrix synthesis than that shown in Ad.BMP-4

aggregates. Moreover, aggregates treated with scAAV2-BMP-4 were uniformly round

without the irregular edges and asymmetry often observed in Ad.BMP-4 aggregates.


113









Delivery of modified stem cells could become a method of choice for clinical

applications, in both human and veterinary medicine. Patients' own cells can be

harvested, modified with scAAV-BMP-4 (or a virus carrying other chondrogenic factors),

and reintroduced to the cartilage defect site within one surgical procedure to augment

repair of cartilage damaged by sports injury (as in the case of football players and

racehorses) or arthritic disease (such as in older adults and in dogs). A gene- or

modified cell-delivery approach is not strictly limited to cartilage applications; this

technique can be used to enhance bone fracture healing, heart tissue repair, and

numerous other conditions that impact animals and people.


114









LIST OF REFERENCES


1. Marieb, E.N. Tissue: a living connection. Human Anatomy & Physiology 119-133
(1998).

2. Lodish, H. et al. Collagen: the fibrous proteins of the matrix. Molecular Cell Biology
(1999).

3. Bastiaansen-Jenniskens, Y.M. et al. Contribution of collagen network features to
functional properties of engineered cartilage. Osteoarthr. Cartil. 16, 359-366
(2008).

4. Gikas, P.D., Aston, W.J.S. & Briggs, T.W.R. Autologous chondrocyte implantation:
where do we stand now? J. Orthop. Sci. 13, 283-292 (2008).

5. Oegema, T.R., Carpenter, R.J., Hofmeister, F. & Thompson, R.C. The interaction
of the zone of calcified cartilage and subchondral bone in osteoarthritis. Microsc.
Res. Tech 37, 324-332 (1997).

6. Ghivizzani, S.C., Oligino, T.J., Robbins, P.D. & Evans, C.H. Cartilage injury and
repair. Phys. Med. Rehabil. Clin. N. Am. 11, 289-307, vi (2000).

7. Buckwalter, J.A. & Mankin, H.J. Articular cartilage: tissue design and chondrocyte-
matrix interactions. Instr. Course Lect. 47, 477-486 (1998).

8. Bentley, G. & Minas, T. Treating joint damage in young people. BMJ 320, 1585-
1588 (2000).

9. Goldring, M.B. Update on the biology of the chondrocyte and new approaches to
treating cartilage diseases. Best Pract. Res. Clin. Rheumatol. 20, 1003-1025
(2006).

10. Buschmann, M.D. & Grodzinsky, A.J. A molecular model of proteoglycan-
associated electrostatic forces in cartilage mechanics. J. Biomech. Eng. 117, 179-
192 (1995).

11. Ahmed, T.A.E. & Hincke, M.T. Strategies for articular cartilage lesion repair and
functional restoration. Tissue Eng Part B Rev 16, 305-329 (2010).

12. Muir, H. The chondrocyte, architect of cartilage. Biomechanics, structure, function
and molecular biology of cartilage matrix macromolecules. Bioessays 17, 1039-
1048 (1995).

13. Martin, J.A. & Buckwalter, J.A. Aging, articular cartilage chondrocyte senescence
and osteoarthritis. Biogerontology 3, 257-264 (2002).

14. Martin, J.A. & Buckwalter, J.A. Roles of articular cartilage aging and chondrocyte
senescence in the pathogenesis of osteoarthritis. Iowa Orthop. J. 21, 1-7 (2001).


115









15. Reddi, A.H. Cartilage morphogenetic proteins: role in joint development,
homoeostasis, and regeneration. Ann. Rheum. Dis. 62 Suppl 2, ii73-78 (2003).

16. Ghosh, S., Laha, M., Mondal, S., Sengupta, S. & Kaplan, D.L. In vitro model of
mesenchymal condensation during chondrogenic development. Biomaterials 30,
6530-6540 (2009).

17. DeLise, A.M., Fischer, L. & Tuan, R.S. Cellular interactions and signaling in
cartilage development. Osteoarthr. Cartil 8, 309-334 (2000).

18. Navagiri, S.S. & Dubey, P.N. The role of hypertrophic cartilage in endochondral
ossification. Z Mikrosk Anat Forsch 90, 435-446 (1976).

19. Coelho, C.N. & Kosher, R.A. Gap junctional communication during limb cartilage
differentiation. Dev. Biol 144, 47-53 (1991).

20. Tuan, R.S. Cellular signaling in developmental chondrogenesis: N-cadherin, Wnts,
and BMP-2. J. Bone Joint Surg. Am. 85-A Suppl 2, 137-141 (2003).

21. Moses, H.L. & Serra, R. Regulation of differentiation by TGF-beta. Curr. Opin.
Genet. Dev 6, 581-586 (1996).

22. Johnstone, B., Hering, T.M., Caplan, A.I., Goldberg, V.M. & Yoo, J.U. In vitro
chondrogenesis of bone marrow-derived mesenchymal progenitor cells. Exp. Cell
Res. 238, 265-272 (1998).

23. Yoo, J.U. et al. The chondrogenic potential of human bone-marrow-derived
mesenchymal progenitor cells. J. Bone Joint Surg. Am. 80, 1745-1757 (1998).

24. Barry, F., Boynton, R.E., Liu, B. & Murphy, J.M. Chondrogenic differentiation of
mesenchymal stem cells from bone marrow: differentiation-dependent gene
expression of matrix components. Exp. Cell Res. 268, 189-200 (2001).

25. Denker, A.E., Nicoll, S.B. & Tuan, R.S. Formation of cartilage-like spheroids by
micromass cultures of murine C3H10T1/2 cells upon treatment with transforming
growth factor-beta 1. Differentiation 59, 25-34 (1995).

26. Hill, D.J. & Logan, A. Peptide growth factors and their interactions during
chondrogenesis. Prog. Growth Factor Res 4, 45-68 (1992).

27. Qi, W.N. & Scully, S.P. Effect of type II collagen in chondrocyte response to TGF-
beta 1 regulation. Exp. Cell Res 241, 142-150 (1998).

28. Morales, T.I., Joyce, M.E., Sobel, M.E., Danielpour, D. & Roberts, A.B.
Transforming growth factor-beta in calf articular cartilage organ cultures: synthesis
and distribution. Arch. Biochem. Biophys 288, 397-405 (1991).


116









29. Luyten, F.P., Chen, P., Paralkar, V. & Reddi, A.H. Recombinant bone
morphogenetic protein-4, transforming growth factor-beta 1, and activin A enhance
the cartilage phenotype of articular chondrocytes in vitro. Exp. Cell Res 210, 224-
229 (1994).

30. Shuler, F.D. et al. Increased matrix synthesis following adenoviral transfer of a
transforming growth factor betal gene into articular chondrocytes. J. Orthop. Res
18, 585-592 (2000).

31. Glansbeek, H.L., van Beuningen, H.M., Vitters, E.L., van der Kraan, P.M. & van
den Berg, W.B. Stimulation of articular cartilage repair in established arthritis by
local administration of transforming growth factor-beta into murine knee joints. Lab.
Invest 78, 133-142 (1998).

32. Tuli, R. et al. Transforming growth factor-beta-mediated chondrogenesis of human
mesenchymal progenitor cells involves N-cadherin and mitogen-activated protein
kinase and Wnt signaling cross-talk. J. Biol. Chem 278, 41227-41236 (2003).

33. Lengner, C.J. et al. Primary mouse embryonic fibroblasts: a model of mesenchymal
cartilage formation. J. Cell. Physiol 200, 327-333 (2004).

34. Wozney, J.M. & Rosen, V. Bone morphogenetic protein and bone morphogenetic
protein gene family in bone formation and repair. Clin. Orthop. Relat. Res 26-37
(1998).

35. Wan, M. & Cao, X. BMP signaling in skeletal development. Biochem. Biophys. Res.
Commun 328, 651-657 (2005).

36. Oshin, A.O. & Stewart, M.C. The role of bone morphogenetic proteins in articular
cartilage development, homeostasis and repair. Vet. Comp. Orthop. Traumatol. 20,
151-158 (2007).

37. Yoon, B.S. & Lyons, K.M. Multiple functions of BMPs in chondrogenesis. J. Cell.
Biochem 93, 93-103 (2004).

38. Ahrens, M. et al. Expression of human bone morphogenetic proteins-2 or -4 in
murine mesenchymal progenitor C3H10T1/2 cells induces differentiation into
distinct mesenchymal cell lineages. DNA Cell Biol 12, 871-880 (1993).

39. Denker, A.E., Haas, A.R., Nicoll, S.B. & Tuan, R.S. Chondrogenic differentiation of
murine C3H10T1/2 multipotential mesenchymal cells: I. Stimulation by bone
morphogenetic protein-2 in high-density micromass cultures. Differentiation 64, 67-
76(1999).

40. Aikawa, T. et al. Establishment of bone morphogenetic protein 2 responsive
chondrogenic cell line. J. Bone Miner. Res 11, 544-553 (1996).


117









41. Sellers, R.S., Peluso, D. & Morris, E.A. The effect of recombinant human bone
morphogenetic protein-2 (rhBMP-2) on the healing of full-thickness defects of
articular cartilage. J Bone Joint Surg Am 79, 1452-1463 (1997).

42. Sellers, R.S. et al. Repair of articular cartilage defects one year after treatment with
recombinant human bone morphogenetic protein-2 (rhBMP-2). J Bone Joint Surg
Am 82, 151-160 (2000).

43. White, A.P. et al. Clinical applications of BMP-7/OP-1 in fractures, nonunions and
spinal fusion. Int Orthop 31, 735-741 (2007).

44. Mason, J.M. et al. Cartilage and bone regeneration using gene-enhanced tissue
engineering. Clin. Orthop. Relat. Res S171-178 (2000).

45. Cook, S.D., Patron, L.P., Salkeld, S.L. & Rueger, D.C. Repair of articular cartilage
defects with osteogenic protein-1 (BMP-7) in dogs. J Bone Joint Surg Am 85-A
Suppl 3, 116-123 (2003).

46. Miljkovic, N.D., Cooper, G.M. & Marra, K.G. Chondrogenesis, bone morphogenetic
protein-4 and mesenchymal stem cells. Osteoarthr. Cartil. 16, 1121-1130 (2008).

47. Steinert, A.F. et al. Hypertrophy is induced during the in vitro chondrogenic
differentiation of human mesenchymal stem cells by bone morphogenetic protein-2
and bone morphogenetic protein-4 gene transfer. Arthritis Res. Ther 11, R148
(2009).

48. Kuroda, R. et al. Cartilage repair using bone morphogenetic protein 4 and muscle-
derived stem cells. Arthritis Rheum 54, 433-442 (2006).

49. Nixon, A.J. et al. Insulinlike growth factor-I gene therapy applications for cartilage
repair. Clin. Orthop. Relat. Res S201-213 (2000).

50. Nixon, A.J., Saxer, R.A. & Brower-Toland, B.D. Exogenous insulin-like growth
factor-I stimulates an autoinductive IGF-I autocrine/paracrine response in
chondrocytes. J. Orthop. Res 19, 26-32 (2001).

51. Brower-Toland, B.D. et al. Direct adenovirus-mediated insulin-like growth factor I
gene transfer enhances transplant chondrocyte function. Hum. Gene Ther 12, 117-
129 (2001).

52. Khan, W.S., Johnson, D.S. & Hardingham, T.E. The potential of stem cells in the
treatment of knee cartilage defects. Knee (2010).

53. de Crombrugghe, B. et al. Transcriptional mechanisms of chondrocyte
differentiation. Matrix Biol 19, 389-394 (2000).


118









54. Kolettas, E., Muir, H.I., Barrett, J.C. & Hardingham, T.E. Chondrocyte phenotype
and cell survival are regulated by culture conditions and by specific cytokines
through the expression of Sox-9 transcription factor. Rheumatology (Oxford) 40,
1146-1156 (2001).

55. Bi, W., Deng, J.M., Zhang, Z., Behringer, R.R. & de Crombrugghe, B. Sox9 is
required for cartilage formation. Nat. Genet 22, 85-89 (1999).

56. Lefebvre, V., Behringer, R.R. & de Crombrugghe, B. L-Sox5, Sox6 and Sox9
control essential steps of the chondrocyte differentiation pathway. Osteoarthr. Cartil
9 Suppl A, S69-75 (2001).

57. Wagner, T. et al. Autosomal sex reversal and campomelic dysplasia are caused by
mutations in and around the SRY-related gene SOX9. Cell 79, 1111-1120 (1994).

58. Kupcsik, L., Stoddart, M.J., Li, Z., Benneker, L.M. & Alini, M. Improving
Chondrogenesis: Potential and Limitations of SOX9 Gene Transfer and Mechanical
Stimulation for Cartilage Tissue Engineering. Tissue Eng Part A (2010).

59. Vortkamp, A. et al. Regulation of rate of cartilage differentiation by Indian
hedgehog and PTH-related protein. Science 273, 613-622 (1996).

60. Mak, K.K., Kronenberg, H.M., Chuang, P., Mackem, S. & Yang, Y. Indian
hedgehog signals independently of PTHrP to promote chondrocyte hypertrophy.
Development 135, 1947-1956 (2008).

61. Weiss, S., Hennig, T., Bock, R., Steck, E. & Richter, W. Impact of growth factors
and PTHrP on early and late chondrogenic differentiation of human mesenchymal
stem cells. J. Cell. Physiol 223, 84-93 (2010).

62. Walsh, D.W., Godson, C., Brazil, D.P. & Martin, F. Extracellular BMP-antagonist
regulation in development and disease: tied up in knots. Trends Cell Biol 20, 244-
256 (2010).

63. Reddi, A.H. Interplay between bone morphogenetic proteins and cognate binding
proteins in bone and cartilage development: noggin, chordin and DAN. Arthritis
Res. 3, 1-5 (2001).

64. Aigner, T. & Stove, J. Collagens--major component of the physiological cartilage
matrix, major target of cartilage degeneration, major tool in cartilage repair. Adv.
Drug Deliv. Rev. 55, 1569-1593 (2003).

65. Detterline, A.J., Goldberg, S., Bach, B.R. & Cole, B.J. Treatment options for
articular cartilage defects of the knee. Orthop Nurs 24, 361-366; quiz 367-368
(2005).

66. Ahmed, T.A.E. & Hincke, M.T. Strategies for articular cartilage lesion repair and
functional restoration. Tissue Eng. Part B Rev. (2010).


119









67. Siparsky, P., Ryzewicz, M., Peterson, B. & Bartz, R. Arthroscopic treatment of
osteoarthritis of the knee: are there any evidence-based indications? Clin. Orthop.
Relat. Res 455, 107-112 (2007).

68. Stuart, M.J. & Lubowitz, J.H. What, if any, are the indications for arthroscopic
debridement of the osteoarthritic knee? Arthroscopy 22, 238-239 (2006).

69. Litzner, J., Kasten, P., Ginther, K. & Kirschner, S. Surgical options for patients
with osteoarthritis of the knee. Nat Rev Rheumatol 5, 309-316 (2009).

70. Hunziker, E.B. & Quinn, T.M. Surgical removal of articular cartilage leads to loss of
chondrocytes from cartilage bordering the wound edge. J Bone Joint Surg Am 85-A
Supply 2, 85-92 (2003).

71. Magnuson, P.B. The classic: Joint debridement: surgical treatment of degenerative
arthritis. Clin. Orthop. Relat. Res 4-12 (1974).

72. Day, B. The indications for arthroscopic debridement for osteoarthritis of the knee.
Orthop. Clin. North Am 36, 413-417 (2005).

73. Clarke, H.D. & Scott, W.N. The role of debridement: through small portals. J
Arthroplasty 18, 10-13 (2003).

74. McLaren, A.C., Blokker, C.P., Fowler, P.J., Roth, J.N. & Rock, M.G. Arthroscopic
debridement of the knee for osteoarthrosis. Can J Surg 34, 595-598 (1991).

75. Kim, H.K., Moran, M.E. & Salter, R.B. The potential for regeneration of articular
cartilage in defects created by chondral shaving and subchondral abrasion. An
experimental investigation in rabbits. J Bone Joint Surg Am 73, 1301-1315 (1991).

76. Nehrer, S. & Minas, T. Treatment of articular cartilage defects. Invest Radiol 35,
639-646 (2000).

77. Jackson, R.W. & Dieterichs, C. The results of arthroscopic lavage and debridement
of osteoarthritic knees based on the severity of degeneration: a 4- to 6-year
symptomatic follow-up. Arthroscopy 19, 13-20 (2003).

78. Ficat, R.P., Ficat, C., Gedeon, P. & Toussaint, J.B. Spongialization: a new
treatment for diseased patellae. Clin. Orthop. Relat. Res 74-83 (1979).

79. Steadman, J.R., Rodkey, W.G. & Rodrigo, J.J. Microfracture: surgical technique
and rehabilitation to treat chondral defects. Clin. Orthop. Relat. Res S362-369
(2001).

80. Bekkers, J.E.J., Inklaar, M. & Saris, D.B.F. Treatment selection in articular cartilage
lesions of the knee: a systematic review. Am. J. Sports Med. 37 Suppl 1, 148S-
55S (2009).


120









81. Steinwachs, M.R., Guggi, T. & Kreuz, P.C. Marrow stimulation techniques. Injury
39 Suppl 1, S26-31 (2008).

82. Kessler, M.W., Ackerman, G., Dines, J.S. & Grande, D. Emerging technologies and
fourth generation issues in cartilage repair. Sports Med Arthrosc 16, 246-254
(2008).

83. Steadman, J.R. et al. Outcomes of microfracture for traumatic chondral defects of
the knee: average 11-year follow-up. Arthroscopy 19, 477-484 (2003).

84. Salter, R.B. et al. The biological effect of continuous passive motion on the healing
of full-thickness defects in articular cartilage. An experimental investigation in the
rabbit. J Bone Joint Surg Am 62, 1232-1251 (1980).

85. Mitchell, N. & Shepard, N. The resurfacing of adult rabbit articular cartilage by
multiple perforations through the subchondral bone. J Bone Joint Surg Am 58, 230-
233 (1976).

86. Ficat, R.P., Ficat, C., Gedeon, P. & Toussaint, J.B. Spongialization: a new
treatment for diseased patellae. Clin. Orthop. Relat. Res 74-83 (1979).

87. Bhosale, A.M. & Richardson, J.B. Articular cartilage: structure, injuries and review
of management. Br. Med. Bull 87, 77-95 (2008).

88. Hunziker, E.B. Articular cartilage repair: basic science and clinical progress. A
review of the current status and prospects. Osteoarthr. Cartil 10, 432-463 (2002).

89. Recht, M. et al. MR imaging of cartilage repair procedures. Skeletal Radiol 32, 185-
200 (2003).

90. Hettrich, C.M., Crawford, D. & Rodeo, S.A. Cartilage repair: third-generation cell-
based technologies-basic science, surgical techniques, clinical outcomes. Sports
Med Arthrosc 16, 230-235 (2008).

91. Polster, J. & Recht, M. Postoperative MR evaluation of chondral repair in the knee.
Eur J Radiol 54, 206-213 (2005).

92. Ferruzzi, A. et al. Autologous chondrocyte implantation in the knee joint: open
compared with arthroscopic technique. Comparison at a minimum follow-up of five
years. J Bone Joint Surg Am 90 Suppl 4, 90-101 (2008).

93. McNickle, A.G., L'Heureux, D.R., Yanke, A.B. & Cole, B.J. Outcomes of autologous
chondrocyte implantation in a diverse patient population. Am J Sports Med 37,
1344-1350 (2009).

94. McNickle, A.G., Provencher, M.T. & Cole, B.J. Overview of existing cartilage repair
technology. Sports Med Arthrosc 16, 196-201 (2008).


121









95. Recht, M.P., Goodwin, D.W., Winalski, C.S. & White, L.M. MRI of articular
cartilage: revisiting current status and future directions. AJR Am J Roentgenol 185,
899-914 (2005).

96. Friedenstein, A.J., Piatetzky-Shapiro, I.I. & Petrakova, K.V. Osteogenesis in
transplants of bone marrow cells. J Embryol Exp Morphol 16, 381-390 (1966).

97. Friedenstein, A.J., Chailakhjan, R.K. & Lalykina, K.S. The development of fibroblast
colonies in monolayer cultures of guinea-pig bone marrow and spleen cells. Cell
Tissue Kinet 3, 393-403 (1970).

98. Noth, U., Steinert, A.F. & Tuan, R.S. Technology insight: adult mesenchymal stem
cells for osteoarthritis therapy. Nat Clin Pract Rheumatol 4, 371-380 (2008).

99. Kolf, C.M., Cho, E. & Tuan, R.S. Mesenchymal stromal cells. Biology of adult
mesenchymal stem cells: regulation of niche, self-renewal and differentiation.
Arthritis Res. Ther 9, 204 (2007).

100. Chen, F.H. & Tuan, R.S. Mesenchymal stem cells in arthritic diseases. Arthritis
Res. Ther 10, 223 (2008).

101. Dragoo, J.L. et al. Tissue-engineered cartilage and bone using stem cells from
human infrapatellar fat pads. J Bone Joint Surg Br 85, 740-747 (2003).

102. Huang, G.T., Gronthos, S. & Shi, S. Mesenchymal stem cells derived from dental
tissues vs. those from other sources: their biology and role in regenerative
medicine. J. Dent. Res 88, 792-806 (2009).

103. Alongi, D.J. et al. Stem/progenitor cells from inflamed human dental pulp retain
tissue regeneration potential. Regen Med (2010).doi:10.2217/rme.10.30

104. Dowthwaite, G.P. et al. The surface of articular cartilage contains a progenitor cell
population. J. Cell. Sci. 117, 889-897 (2004).

105. Pittenger, M.F. et al. Multilineage potential of adult human mesenchymal stem
cells. Science 284, 143-147 (1999).

106. Prockop, D.J. Marrow stromal cells as stem cells for nonhematopoietic tissues.
Science 276, 71-74 (1997).

107. Tew, S. et al. Differences in repair responses between immature and mature
cartilage. Clin. Orthop. Relat. Res S142-152 (2001).

108. Martin, J.A. & Buckwalter, J.A. The role of chondrocyte senescence in the
pathogenesis of osteoarthritis and in limiting cartilage repair. J. Bone Joint Surg.
Am. 85-A Suppl 2, 106-110 (2003).


122









109. Khan, W.S., Adesida, A.B., Tew, S.R., Andrew, J.G. & Hardingham, T.E. The
epitope characterisation and the osteogenic differentiation potential of human fat
pad-derived stem cells is maintained with ageing in later life. Injury 40, 150-157
(2009).

110. Krampera, M. et al. Bone marrow mesenchymal stem cells inhibit the response of
naive and memory antigen-specific T cells to their cognate peptide. Blood 101,
3722-3729 (2003).

111. Evans, C.H. et al. Gene transfer to human joints: progress toward a gene therapy
of arthritis. Proc. Natl. Acad. Sci. U.S.A 102, 8698-8703 (2005).

112. Ghivizzani, S.C., Oligino, T.J., Glorioso, J.C., Robbins, P.D. & Evans, C.H. Direct
gene delivery strategies for the treatment of rheumatoid arthritis. Drug Discov.
Today 6, 259-267 (2001).

113. Baltzer, A.W. et al. Genetic enhancement of fracture repair: healing of an
experimental segmental defect by adenoviral transfer of the BMP-2 gene. Gene
Ther 7, 734-739 (2000).

114. Peterson, B. et al. Healing of critically sized femoral defects, using genetically
modified mesenchymal stem cells from human adipose tissue. Tissue Eng 11, 120-
129 (2005).

115. Lieberman, J.R. et al. The effect of regional gene therapy with bone morphogenetic
protein-2-producing bone-marrow cells on the repair of segmental femoral defects
in rats. J Bone Joint Surg Am 81, 905-917 (1999).

116. Lou, J., Tu, Y., Ludwig, F.J., Zhang, J. & Manske, P.R. Effect of bone
morphogenetic protein-12 gene transfer on mesenchymal progenitor cells. Clin.
Orthop. Relat. Res 333-339 (1999).

117. Lou, J., Tu, Y., Burns, M., Silva, M.J. & Manske, P. BMP-12 gene transfer
augmentation of lacerated tendon repair. J. Orthop. Res 19, 1199-1202 (2001).

118. Betz, O.B. et al. Healing of large segmental bone defects induced by expedited
bone morphogenetic protein-2 gene-activated, syngeneic muscle grafts. Hum.
Gene Ther 20, 1589-1596 (2009).

119. van der Rest, M. & Garrone, R. Collagen family of proteins. FASEB J 5, 2814-2823
(1991).

120. Gelse, K., Poschl, E. & Aigner, T. Collagens--structure, function, and biosynthesis.
Adv. Drug Deliv. Rev. 55, 1531-1546 (2003).

121. Gordon, M.K. & Hahn, R.A. Collagens. Cell Tissue Res. 339, 247-257 (2010).


123









122. Lee, J.W., Kim, Y.H., Kim, S., Han, S.H. & Hahn, S.B. Chondrogenic differentiation
of mesenchymal stem cells and its clinical applications. Yonsei Med. J 45 Suppl,
41-47 (2004).

123. Welter, J.F., Solchaga, L.A. & Penick, K.J. Simplification of aggregate culture of
human mesenchymal stem cells as a chondrogenic screening assay.
BioTechniques 42, 732, 734-737 (2007).

124. Hardy, S., Kitamura, M., Harris-Stansil, T., Dai, Y. & Phipps, M.L. Construction of
adenovirus vectors through Cre-lox recombination. J. Virol. 71, 1842-1849 (1997).

125. Palmer, G.D. et al. Gene transfer to articular chondrocytes with recombinant
adenovirus. Cytokines and Colony Stimulating Factors 215, 235-246 (2003).

126. Colter, D.C., Class, R., DiGirolamo, C.M. & Prockop, D.J. Rapid expansion of
recycling stem cells in cultures of plastic-adherent cells from human bone marrow.
Proc. Natl. Acad. Sci. U.S.A 97, 3213-3218 (2000).

127. Caplan, A.I. Review: mesenchymal stem cells: cell-based reconstructive therapy in
orthopedics. Tissue Eng. 11, 1198-1211 (2005).

128. Steinert, A.F. et al. Major biological obstacles for persistent cell-based regeneration
of articular cartilage. Arthritis Res. Ther 9, 213 (2007).

129. Nixon, A.J. et al. Gene therapy in musculoskeletal repair. Ann. N. Y. Acad. Sci
1117, 310-327 (2007).

130. Palmer, G.D. et al. Gene-induced chondrogenesis of primary mesenchymal stem
cells in vitro. Mol. Ther. 12, 219-228 (2005).

131. Song, J.J. et al. Connective tissue growth factor (CTGF) acts as a downstream
mediator of TGF-betal to induce mesenchymal cell condensation. J. Cell. Physiol
210, 398-410 (2007).

132. Kay, J.D. et al. Intra-articular gene delivery and expression of interleukin-1Ra
mediated by self-complementary adeno-associated virus. J Gene Med 11, 605-614
(2009).


124









BIOGRAPHICAL SKETCH

Marsha Lynn Bush was raised in the tourist town of Cave City, Kentucky, where

she grew up on a farm, in a zoo, and next door to the longest cave in the world. After

graduating valedictorian from Barren County High School in 1999, Marsha attended the

University of Kentucky. She spent the summers of 2000, 2002, and 2003 serving as a

guide at Mammoth Cave National Park. It was here that Marsha learned of the variety of

research projects that were taking place spanning the fields of geology, biology,

hydrology, microbiology, anthropology, genealogy, and paleontology. During the

summer of 2001, Marsha began her own research pursuits as a student in the Kentucky

Young Scientist Summer (KYSS) Research Program. Under the guidance of Dr.

Charles (Chuck) Fox, Marsha took part in a population genetics experiment examining

maternal effects on offspring lifespan in the seed beetle and stored grain pest,

Callosobruchus maculatus. An offshoot of this project became her primary research

project for completion of her Bachelor's degree in agricultural biotechnology. Prior to

graduating in December 2003, Marsha spent the spring semester as the biotechnology

lab intern at The Land at Epcot. As part of Epcot Science, Marsha learned about plant

tissue culture as she propagated a variety of species used within The Land

greenhouses and grew, packaged, and managed the sales of Mickey's Mini Gardens at

four locations within Walt Disney World. This internship experience opened the door for

Marsha's return to Florida in the spring of 2004 as a reproductive biology intern at

Disney's Animal Kingdom Wildlife Tracking Center. This experience cemented her

desire to take part in research efforts that would benefit both animals and people.

In August 2004, Marsha began studies in the Interdisciplinary Program (IDP) in

Biomedical Sciences at the University of Florida College of Medicine. In May 2005,


125









Marsha joined the lab of Dr. Steve Ghivizzani in the Department of Orthopaedics and

Rehabilitation where she has investigated the induction of chondrogenesis in

mesenchymal stem cells following delivery of trangenes via viral vectors. Her work may

contribute to the understanding of gene therapy treatments for repair of articular

cartilage damage resulting from injuries or diseases such as arthritis. Following

completion of her Ph.D. in 2010, Marsha will continue her education at the University of

Wisconsin-Madison School of Veterinary Medicine, as she combines her interests in

veterinary medicine and translational research.


126





PAGE 1

GENE-INDUCED CHONDR OGENESIS OF MESENCHYMAL STEM CELLS THROUGH VIRAL GENE DELIVERY By MARSHA LYNN BUSH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORID A IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010 1

PAGE 2

2010 Marsha Lynn Bush 2

PAGE 3

To my parents, Diana Bush and Jim Bush 3

PAGE 4

ACKNOWLEDGMENTS I thank my family for t heir unconditional love and s upport throughout my ongoing educational pursuits; I thank Grandma fo r bei ng my penpal, Geoffrey and Aaron for being calm and injecting humor into my life, and Mama, Daddy, and Kim for being patient listeners and prov iding sound advice. I thank my mentor, Dr. Steve Ghivizzani, for his leadership and encouragemen t as he taught me the importance of clear, persuasive writing, the value of effective presentation skills, and how to think like a sci entist. He provided me ample opportunities to learn and experiment in his lab and to discuss the scientific merit of many experimental outcomes. I great ly appreciate my advisory co mmittee: Drs. Jrg Bungert, Bryon Petersen, Ed Scott, and Phyllis LuValle, who have provided many hours of guidance. My lab colleagues, who have become a surrogate family over the years, have always been there for me: Jesse Kay, Rach ael Watson, Carrie Saites, Celine Theodore, Paddy Levings, Anthony Dacanay, and Jeetpaul Saran. My first research exper iences were in the lab of Dr. Chuck Fox at the University of Kentucky College of Agriculture, Department of Entomology. I m grateful for the time he spent teaching me how to design experiments with stati stically significant sample sizes and appropriate controls. I fu rther extend my gratitude to the many teachers and professors who encouraged me to pursue my dreams and to never doubt myself. I thank my dear friend Joe Menotti for his faith in me when the task felt overwhelming. Finally, I thank my friends and fellow scientists whom I have met through the Interdisciplinary Program (IDP) in Biomedical Sciencesespecially Amanda DuBose, Melissa Marzahn, Megan Greenlee Dacia Kwiatkowski, Ahu Demir and 4

PAGE 5

Maggie and Levi Watsonfor their consolation over failed experiments and their shared joy over triumphs, no matte r how big or small. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDG MENTS..................................................................................................4 LIST OF T ABLES............................................................................................................9 LIST OF FI GURES ........................................................................................................10 LIST OF ABBR EVIATIONS...........................................................................................12 ABSTRACT ................................................................................................................... 15 CHAPTE R 1 INTRODUC TION....................................................................................................17 Cartilage Bi ology ..................................................................................................... 17 Architecture of the Kn ee...................................................................................17 Articular Cartil age Struct ure .............................................................................18 Proteoglyc ans ............................................................................................19 Chondrocyt es .............................................................................................21 Collagens ................................................................................................... 23 Adjacent cells and tissues s upport articular cartil age................................. 25 Cartilage Deve lopment...........................................................................................26 Osteogenic and Chondrogenic Differentiation and Regulation in Developm ent.................................................................................................26 BMP Receptors and Signalin g ..........................................................................31 Cartilage R epair ...................................................................................................... 32 The Bodys Natural Approaches to Cartilage Defe ct Repai r.............................32 Surgical Approaches to Cartilage Repair ......................................................... 33 Nonreparative restorat ive techniques........................................................33 Reparative pr ocedures ...............................................................................34 Restorative st rategies................................................................................ 37 Reconstructive methods to treat cartilage defects.....................................39 Geneand Cell-Based Approac hes to Cartil age Repa ir ...................................41 MSCs: the logical cell type for ch ondr ogenesis ..........................................42 Mechanical stimulation to promo te ca rtilage .............................................. 44 Regeneration and Repair Using MSCs: Gene Therapy and Or tho paedics.............44 2 MATERIALS A ND METHOD S................................................................................55 In Vitro Cell Cult ure .................................................................................................55 HEK293 and 293FT Ce ll Cultur e ......................................................................55 Harvesting and Culturi ng Bovine MSCs ...........................................................55 Aggregate (Pellet) Culture ................................................................................56 Chondrogenic Medi a Formula tion ....................................................................56 6

PAGE 7

Virus Preparation and Transgene Expr ession........................................................57 Construction and Generation of Reco mbinant Adenoviruses Contai ning Chondrogenic Transg enes............................................................................57 Adenovirus Propagat ion and Amplif icatio n.......................................................57 Generation of Lentivir us ...................................................................................59 Construction and Generati on of s cAAV Ve ct ors...............................................59 Gene Transfer to MSCs to I nduce and Enhance Chondrogenes is...................60 Plasmid DNA tr ansfectio n..........................................................................60 Transgene delivery us ing adenovir us ........................................................61 Methods to Detect Transgene Pr oducts........................................................... 61 Western blot...............................................................................................61 ELISA to detect secr e ted transgene products ............................................62 Histology and Immuno histochemis try...............................................................62 RNA Ex traction, RT, and rt PCR.......................................................................63 In Vivo Expe riments ................................................................................................64 Intra-Ar ticular Injections....................................................................................64 Harvesting Tissue, Decalcif ication, and Histol ogy ............................................64 3 GENE DELIVERY STIMULATES CHONDROGENESIS OF MSCS.......................67 Introducti on .............................................................................................................67 Rationa le .................................................................................................................70 Result s....................................................................................................................71 Cre-lox Recombination and Adenovirus Propagatio n .......................................71 Isolation of MSCs .............................................................................................73 Gene-Mediated Chondr ogenesis of MSCs .......................................................74 Adenov iral-Mediated Delivery of BMP-4, BMP-7, Ihh, and Sox9 Driv es Chondrogenic Different iation of MSCs ..........................................................76 Adenoviral Delivery of CT GF Stimulates Proliferat ion of MSC Aggregates......81 Combinations of Sox 9 with BMPs Induced Ch ondrogenesi s............................82 Wistar Rat Responses to A denoviral Transgene Delivery................................ 84 Discussio n..............................................................................................................86 4 COMPARISON OF THE EFFECTS OF ADENOVIRAL, LENTIVIRAL, AND AAV TRANSGENE DELIVERY TO MESENCHYMAL ST EM CE LLS .............................99 Introducti on.............................................................................................................99 Results .................................................................................................................. 100 Plasmid DNA Transfection of MSCs Re sults in Transient Expression........... 100 Lentiviral Transduction of MS Cs Proves Ch allenging.....................................101 Adenov iral-Mediated Delivery of BMP4 to MSCs Induces Chondrogenes is.102 Self-Complementary AAV-Mediated Delivery of BMP-4 to MSCs is Comparable to Adenovirus ..........................................................................102 Discussio n ............................................................................................................103 5 SUMMARY AND FUTURE DIRECTIONS............................................................111 7

PAGE 8

LIST OF REFE RENCES.............................................................................................115 BIOGRAPHICAL SKETCH ..........................................................................................125 8

PAGE 9

LIST OF TABLES Table page 1-1 Co llagen types .............................................................................................52 1-2 Morphogens and Growth Fact ors in Cartilage Development........................53 4-1 Comparison of gene delivery vectors reveals pros and c ons.....................110 9

PAGE 10

LIST OF FIGURES Figure page 1-1 The four zones of articular cartilage st retch from the superfi cial surface to the deep zone, where hypertrophic chondrocyt es are replaced by calcified matrix ..................................................................................................................47 1-2 The human knee contains specialized st ructures to withstand the forces of movement ...........................................................................................................48 1-3 The arrangement of struct ures in the articular cart ilage matrix is designed to absorb forces, especia lly compre ssion............................................................... 49 1-4 A proteoglycan aggregate is made up of many proteoglycan subunits attached to a hyaluronic acid backbone via lin k protei n...................................... 50 1-5 Fibrillar collagens, including collagen type II, form a characteristic triple helix... 51 2-1 Assay for in vitro chondrogenes is.......................................................................66 3-1 ELISAs of conditioned media fr om MSC aggregates tr ansduced with BMP-4 and BMP-7 indicated protein production in response to viral doses administe red.......................................................................................................89 3-2 Western blotting confirmed trans gene expression following adenoviral gene delivery to MSCs................................................................................................ 90 3-3 MSC aggregates expressing BMP-4 were evaluated for chondrogenesis through toluidine blue staining of proteoglycans and immunohistochemistry for collagen types I and II. ..................................................................................91 3-4 MSC aggregates infected with Ad.BMP -7 over a range of doses yielded pronounced chondrogenesis at t he 1000 vp/cell dos e only................................92 3-5 Indian hedgehog aggregates expressed proteoglycan and collagen type II following adenoviral gene deliver y...................................................................... 93 3-6 Sox9 induced cellular prolifer ation and matrix synthesis with less differentiation than other transgenes te sted.......................................................94 3-7 Collagen type X staining of Sox9 and BMP-4 aggregates .................................. 95 3-8 CTGF expression in MSC aggregates promotes cellular proliferation but no chondrogenic diffe rentiati on................................................................................96 3-9 Sox9 in combinati on with BMP-4 and BMP-2 in MSC aggregate culture had no additive effects upo n chondrogenes is.. .........................................................97 10

PAGE 11

3-10 Wistar rat knees exhibited varyi ng responses to adenoviral delivery of pleiotropic transgenes ........................................................................................98 4-1 ELISA of conditio ned media demonstrates ex pression of BMP-4 as measured 48 hours post-tr ansfection. .............................................................. 106 4-2 GFP expression in 293 cells 72 hours after infection with 15 l of LV-GFP in a minimal volume of se rumless media.............................................................. 107 4-3 LV-BMP-4 expression was significant ly less than Ad.BMP-4 expression in bovine MSCs .................................................................................................... 107 4-4 AAV serotypes 1, 2, 5, and 8 were screened for transduction efficiency on low passage bovine MSCs in monolay er.......................................................... 108 4-5 MSCs transduced with scAAV-BMP-4 se rotypes 2 and 5 demonstrate matrix protein synthesis in ag gregate cultur e..............................................................109 11

PAGE 12

LIST OF ABBREVIATIONS AAV adeno-associated virus ACT autologous chondro cyte transplantation BM bone marrow BMP bone morphogenetic protein bp base pairs C degrees Centigrade CaPO4 calcium phosphate CAR Coxsackie-adenovirus receptor CDMP cartilage derived morphogenic protein cDNA complementary DNA CMV cytomegalovirus CsCl cesium chloride DMEM Dulbeccos Modified Eagle Medium DNA deoxyribonucleic acid ECM extracellular matrix ELISA enzyme-linked immunosorbant assay FBS fetal bovine serum FGF fibroblast growth factor GAG glycosaminoglycan GDF growth/differentiation factor GFP green fluorescent protein hr hour IGF insulin-like growth factor Ihh Indian hedgehog 12

PAGE 13

kg kilogram MSC mesenchymal stem cell g microgram mg milligram min minute mL milliliter mM millimolar MMPs matrix metalloproteases ng nanogram nm nanometer OA osteoarthritis PBS phosphate buffered saline PCR polymerase chain reaction pg picogram PG proteoglycan RA rheumatoid arthritis rAAV recombinant adeno-associated virus RNA ribonucleic acid sc self-complementary scAAV self-complementary adeno-associated virus SDS-PAGE sodium dodecyl sulfatepolyacrylamide gel electrophoresis ssAAV single-stranded aden o-associated virus TGFtransforming growth factor beta trs terminal resolution site U units 13

PAGE 14

UV ultraviolet vg vector genome vp virus particle VSV-G vesicular stomatitis virus G protein 14

PAGE 15

Abstract of Dissertation Pr esented to the Graduate School of the University of Florida in Partial Fulf illment of the Requirements for t he Degree of Doctor of Philosophy GENE-INDUCED CHONDR OGENESIS OF MESENCHYMAL STEM CELLS THROUGH VIRAL GENE DELIVE RY By Marsha Lynn Bush August 2010 Chair: Steven C. Ghivizzani Major: Medical Sciences Biochemistry and Molecular Biology Articular cartilage is a highly specializ ed tissue that allows for near frictionless motion of diarthrodial joints. When cartilage is damaged as a result of injury or disease, the bodys innate capacity for repair is often insufficient due to the low cellular density, avascular and aneural nature of cartilage. Typically the repai r tissue is a fibrotic scar which lacks the structural properties of native cartilage. Adult mesenchymal stem cells (MSCs) are well-suited for cellbased therapies for cartilage repair since they are readily available from many tissue sources and c apable of differentiating along multiple mesenchymal lineages. Gene tr ansfer to MSCs is a viable method for achieving sustained local expression of specific protei n factors, which has been shown to induce chondrogenesis in vitro and may enhance chondrogenic differentiation in vivo Previous experiments have shown that delivery of the cDNAs for TGFand BMP2 induces chondrogenesis of bovine MSCs in high density aggregate culture. In these studies, we expanded the analyses to include additional cDNAs whose protein products are associated with chondrogenic different iation during development, including bone morphogenetic proteins BMP-4, BMP-7, developmental mor phogen Indian hedgehog 15

PAGE 16

(Ihh), transcription factor Sox9, and connec tive tissue growth factor (CTGF). We transduced early passage bovine MSCs with adenoviral vectors carrying the complete cDNAs for the candidate transgenes at doses ranging from 10 virus particles/cell to 100,000 vp/cell. Chondrogenesis was evaluated by gross examination of aggregate morphology, toluidine blue st aining for proteoglycan expr ession and collagen types I, II, and X immunohistochemistry. The greatest biological responses for each transgene were observed in the dose range of 100-1000 vp/cell; higher viral doses appeared to inhibit chondrogenesis, while lower viral doses were insufficient to yield a pronounced effect. We found that gene transfer of the cDNAs for BMP-4, BMP-7, Ihh, and Sox9 induced chondrogenesis of bovine MSCs wh ile CTGF was not chondrogenic. BMP-4 supports chondrogenesis as effectivel y as other members of the TGFsuperfamily such as BMP-2. In further experiments to elucidate t he most suitable gene delivery system for animal models and clinical applications, we evaluated v ector systems currently available: adeno-associated virus (AAV), lentiv irus, and non-viral transfection relative to adenovirus. Simple transfection alone was unable to generate the levels of expression necessary to promote chondrogenesis in our system. Both adenovirus and AAV transduction of MSCs resulted in robust BMP4 expression over the course of 21 days, complete with chondrogenic differentiation and the expression of cartilage matrix proteins. Lentivirus, which offers the pot ential for long-term gene expression, was illsuited for our application because it required l ong-term culture selection of cells, which altered cellular morphology and negatively im pacted cell survival. These data indicate that scAAV serotype 2 delivery of BM P-4 promotes chondrogenesis of MSCs. 16

PAGE 17

CHAPTER 1 INTRODUCTION Cartilage Biology Cartilage is a specialized connective tissue found in various locations throughout the human body and is subdivided into three main types: 1) hyaline (articular), 2) elastic, and 3) fibrocartilage1. Hyaline cartilage, named for its glassy appearance, covers articulating surfaces, such as those at the ends of long bones; ther efore, it is also known as articular cartilage. Its unique architecture offers firm support with some pliability and consists of 80-90% water, with the remainder of the tissue composed primarily of collagen type II fibers and proteoglycans. Cells known as chondrocytes reside in the cartilage at low density and work to slowly remodel the matrix to keep it structurally and functionally sound2. In childhood, hyaline cartilage is a key component of the growth plates, the actively growi ng regions that add lengt h to the ends of long bones1 (Fig.1-1A). Architecture of the Knee Diarthrodial joints, such as the knee, hav e a specialized architecture that is capable of withstanding repeated extreme forces over the lifetime of an individual, often without problems for 80 or more years. Mese nchymal tissues in the knee joint, including cartilage, tendons, ligaments, synovium, and m eniscus, have specific features that enable smooth locomotion and resistance to comp ressive and tensile forces (Fig. 1-2). Ligaments present within and without the join t capsule guide movement of the femur and tibia to control extension and flexion, secure the articulating bones when standing, and prevent hyperextension or overflexion of the knee joint. The knee can absorb vertical forces equal to nearly seven time s body weight; however, the knee joint is 17

PAGE 18

susceptible to damage from horizontal forces or twisting movements such as those that occur in football and other contact sports. Articular Cartilage Structure Articular cartilage covers and protects the join ts at sites of articulation, such as the knee, ankle, elbow, and knuckle. To withstand the stress of mo vement of the human body, articular cartilage must be durable yet provide an effective cushion against loadbearing and impact, and it must be smooth to provide nearly frictionless motion. The nature and structure of articular cartil age impart these necessary properties. The extrac ellular matrix (ECM) has unique properties that enable articular cartilage to rebound after impact. Moreover, it is remarkably durable and is capable of withstanding repeated exposure to extreme forces and pressure, which effectively cushions the ends of the bones. Extracellula r matrix is synthesized and maintained by the resident population of chondrocytes th rough anabolic and catabolic mech anisms and is composed mainly of co llagen fibrils and proteoglycans3 (Fig. 1-3). As depicted in Figure 1-1B, articular cart ilage is highly organized and consists of four zones, or layers: 1) superficial, 2) tr ansitional, 3) radial, and 4) calcified cartilage3,4. The superficial zone contains chondrocytes that are flattened and lie parallel to the surface. Collagen type II fibrils also have a para llel orientation in this layer, which allows the cartilage to withstand the shear forces generated during normal joint loading. The transitional zone, as its name suggests exhibits a cell morphology and ECM composition intermediate to that of the superficial z one and the deeper radial zone. Transitional layer chondrocytes are rounded and produce more proteoglycan and less collagen than in the superficial layer. Deep to this zone lies the radial zone, where the cells are aligned in columns perpendicular to the joint surface4. The radial zone is the 18

PAGE 19

largest layer, with the thicke st collagen fibrils, the most c oncentrated proteoglycans, and the lowest water content. The radial zone is separated from the deepest zone, calcified cartilage, by the tidemark, an area that appears to help tether the cartilage by increasing the contact area between the layers5. The zone of calcified cartilage forms the transition between soft articular cartil age and the hard underlying bone4,5. Cells in this zone are isolated and almost completely surrounded by calcified cartilage, suggesting that they have a low metabolic rate, but it is k nown they are not completely inactive5. Throughout the body, articular cartilage is thicker, with greater proteoglycan content, in load-bearing areas and thinner in areas where loading is minimal4. Mechanical loading affects the morphology and the metabolic activity of chondrocytes4, and controlled loading is thought to contribut e to enhanced cartilage healing after injury. Proteoglycans The extracellular matrix consists of a network of proteoglcyans (PGs) and collagen fibers arranged to control the flow of water molecules to cushion the joint from forces associated with normal motion. Proteoglycans give the tissue its ability to resist compression and remain durable fo r up to 80 years, or more, in some humans. Proteoglycans are proteins bound to l ong-chain polysaccharides known as glycosaminoglycans, or GAGs Large aggrecating proteoglycans known as aggrecans are the most abundant proteoglycans in cartilage consisting of a linear protein core with numerous GAG chains of chondroitin su lfate or keratan sulfate attached6 (Fig. 1-4). In addition to large aggregating proteoglycans like aggrecan, small non-aggregating proteoglycans, including decorin, biglycan, and fibromodulin, bind to other matrix molecules and help to stabilize the matrix7. 19

PAGE 20

Hyaluronic acid (HA), or hyaluronate, mo lecules are the central organizing units of the cartilage matrix. They are long chains wi th up to 100 glycoproteins attached to and extending from them. Link protein connects each proteoglycan subunit to the hyaluronate backbone. The large concentrati on of negative charges from the GAG side chains attract and hold polar water molecules in the matrix by osmotic pressure8. This charge repulsion keeps the chains separated in a characteristic bottle-brush formation (Fig. 1-4). When hydrated, prot eoglycans account for most of the physiological mass of cartilage. They provide a structure to hold and control the flow of water, the integral factor in cartilages ability to provide its protective cu shioning properties throughout lifetime of the individual. The architecture of articular cartilage regulat es the flow of fluid in the cartilage and cushions the joint from mechanical stresses such as compressive loads9. When cartilage is compressed during normal joint moti on, water flows out to the joint cavity, mixing with the existing synovial fluid that lubricates the joint. The negative charges of the GAGs repel one another when forced into close proximity, enab ling them to resist further compression. Addition ally, collagen type II fibrils provide the strength needed to retain the shape of cartilage under comp ression and hinder the expansion of proteoglycans7,10. When the pressure is released, the cartilage springs back to its original form and water molecules are drawn ba ck into the proteoglycans. This flow of liquid into and out of cartilage also carries nutrients to and wastes from chondrocytes, providing nourishment, which helps explain wh y periods of inactivity can result in weaker cartilage that injures easily1. 20

PAGE 21

If proteoglycans are damaged by trauma, infection or enzymes, such as those released during inflammatory disease, their structure degenerates and their waterholding capacity changes8. The collagen meshwork also begins to break down, which reduces constraint on the pr oteoglycans and allows them to imbibe additional water, thus reducing cartilage stiffness. Over ti me damaged cartilage loses its resilience to mechanical forces in the joint, and the s ubchondral bone can become exposed, causing severe pain and disability due to the abundant nerve supply and bleeding from the bone marrow and surrounding vasculature. Due to the density of the extracellula r matrix secreted and maintained by chondrocytes and the physical r equirements of the tissue, t here is no room for blood vessels or nerve fibers to permeate cartilage. Indeed, the permeation of the matrix with vascular or neural tissues would significantly compromise its protective properties. Instead, nutrients diffuse thr ough the perichondrium, a well-v ascularized dense irregular connective tissue that surrounds the su rface of most cartilage structures1. The matrix, while continually undergoing re modeling by chondrocytes, also serves to protect chondrocytes from injury result ing from normal use of the joint. Because cartilage lacks innervation, injuries that occu r within the cartilage layer often go unnoticed until there is penetrati on of the subchondral bone. A dditionally, the structure of the matrix determines the types of mo lecules and which concentrations reach the cells encased within. Chondrocytes Chondrocytes are mature cartilage cells found in deep pits called lacunae1 where they function to maintain cartilage homeos tasis through the perpetual degradation and synthesis of matrix components11. Chondrocytes diffusely populate articular cartilage, 21

PAGE 22

exhibit a low metabolic rate compared to tissues such as muscle, and are restricted from moving and dividing by the dense matrix fibers. The cells do not interact with one another, but instead interact with the ECM (Fig. 1-1B). In the deepest layers of cartilage, chondrocytes exist in a hypoxic environmen t, so metabolism is mainly anaerobic, with the conversion of glucose to lactic acid4. In normal articular cartilage, there is a low turnover of the extracellu lar matrix, and chondrocytes are thought to rarely divide4,12. When articular cartilage sustains damage, however, the chondrocytes form clusters and cellular activity increases4. Aging chondrocytes, on the ot her hand, gradually lose their ability to divide, which supports the observa tion that the cartilage in older individuals heals more slowly than that of young people13,14. Sports injuries in adolescence or young adulthood may set the stage for pain late r in life as those early events damage cartilage, causing it to unnaturally erode. Da maged cartilage may calcify, escalating the extent of damage as this transition to bone results a loss of nutrient supply and subsequent chondrocyte death. The cartilage microenvironment may hav e a profound influence upon chondrocyte lifespan. T he chondrocyte has a finite lifespan in the epiphyseal growth plate, whereas it has a very long, stable phenotype evident in articular cartilage15. Chondrocytes of articular cartilage are long-lived cells of mesenchymal origin, but their lifecycle is unconfirmed. It is possible that they ar e post-mitotic and replaced by new cells infiltrating from the subchondral bone or that a small populatio n of progenitor cells exists within the cartilage. Chondrocyte dedifferentiation in culture. The stability of the phenotype of articular chondrocytes is critically dependent on physical environment and cell density. 22

PAGE 23

In monolayer culture, thes e cells progressively lose their chondrogenic phenotype, transitioning from cubiodal cells with high ECM synthesis to spindle-shaped cells that primarily produce collagen type I. Dedifferentiation can be avoided or delayed by changing the geometry of the ce ll culture through the use of 3D matrices such as agarose or alginate, high density micr omass cultures, or pellet cultures15. Collagens The dry mass of articular cartilage is composed primarily of type II collagen and aggrecan, a proteoglycan. Collagen type II makes up 90% of the collagen present, with minor collagens types VI, IX, and XI maki ng up the other 10% (Table 1-1). Collagen fibrils provide tensile st rength, serving as a natural scaffold to anchor and organize cells and hydrated proteoglycans and possibly gu iding cell signaling responses during development and repair16. Collagens differ in their ability to form fibers and to organize the fibers into networks (Table 1-1). The typica l structure of fibrillar coll agens is a right-handed triple helix, as shown in Figure 1-5, which aris es from an abundance of three principal amino acids: glycine, proline, and hydroxyproline2. These form a repeating pattern of Gly-ProX where X is any amino acid. The side chain of glycine, a single H atom, is the only side chain capable of fitting into the crowded center of the triple strand helix. Type I collagen fibrils have enormous tensile strength; that is, th is type of collagen can be stretched without being broken. Thes e fibrils, roughly 50 nm in diameter and several micrometers long, are packed side-by-side into parallel bundles, termed collagen fibers. In tendons, collagen type I fi bers connect muscles with bones and must withstand enormous forces. Gram for gram type I collagen is stronger than steel2. 23

PAGE 24

Collagen type II, the primary collagen of arti cular cartilage, has fibrils that are smaller in diameter than type I and are or iented randomly in the viscous proteoglycan matrix (Fig. 1-3). Such rigid macromolecules impart a strength and compressibility to the matrix and allow it to resist large deformations in shape. This property allows joints to absorb shock. Type II fibrils are cross-linked to proteoglycans in the matrix by collagen type IX, a collagen consisting of two long triple helices connected by a flexible kink (Fig. 1-5B). The globular N-terminal domain extends from the composite fibrils, as does a heparan sulfate molecule, a type of large, highly charged polysaccharide that is linked to the type IX collagen chain at the flexible kink. These protruding nonhelical domains are thought to anchor the fibril to proteogl ycans and other components of the matrix. The interrupted triple-helical structure of type IX collag en prevents it from assembling into fibrils; instead, collagen type IX associ ates with fibrils formed from other collagen types and thus is called fibril-associated colla gen (Table 1-1). Post-translational modification of procollage n is crucial for the formation of mature collagen molecules and their assembly into fi brils; defects in this process have serious consequences2. For example, the activity of prolyl hydroxylases requires an essential cofactor, ascorbic acid (vitam in C). In cells deprived of ascorbate, as in the disease scurvy, the procollage n chains are not hydroxylated su fficiently to form stable triple helices at normal body temperature, nor can they form normal fibrils. Consequently, nonhydroxylated procollagen chains are degraded within the cell. Type II collagen fibers are firmly embedded in subchondral bone and rise from there to the cartilage surface, where they bend to form arches8. The overlapping fibers 24

PAGE 25

form a mesh framework that gives stability to cartilage (Fig. 1-3). Collagen fibers provide the framework that allows proteoglycans to hold water molecules within the matrix. Adjacent cells and tissues support articular cartilage The adjacent cells and tissues in t he jointsynoviocytes, bone marrow, MSCs, meniscus, ligaments and tendonsprovide support to articular cartilage. Most proximal to the bones and cartilage of the joint is the synovium, a thin layer of vascularized connective tissue that lacks a basement me mb rane. Two cell types are present: type A, macrophage-like cells, and type B, fibroblast-like cells. Macrophage-like or phagocytic cells remove microbes and the debris that resu lts from normal wear and tear in the joint. Type B synoviocytes produce synovial fluid, whic h consists of hyaluronan, or hyaluronic acid, lubricin, proteinases, and collagenases and serves to surround, lubricate, and nourish the joint space. This flui d forms a thin layer (roughly 50 m) at the surface of cartilage and also seeps into microcavities a nd irregularities in the articular cartilage surface, filling all empty space. Normal synovial fluid increases the viscosity and elasticity of articular cartilage, and there is also some evidence that it helps regulate synovial cell growth. The liquid present in articular cartilage effectively serves as a synovial fluid reserve. During movement, the synovial fl uid held in the cartilage is squeezed out mechanically to maintain a layer of flui d on the cartilage surface (so-called weeping lubrication). Synovial fluid is chiefly res ponsible for reducing fr iction between apposing surfaces of cartilage and absorbing shock. In addition, movement of synovial fluid supplies oxygen and nutrients to and remove s carbon dioxide and metabolic wastes from chondrocytes within articular cartilage. 25

PAGE 26

Cartilage Development Osteogenic and Chondrogenic Differentia tion and Regulation in Development Cartilage morphogenesis and osteogenesis are influenced by developmental signals. Current research is exploring which signals direct cartilage to remain at the ends of long bones when the rest of the bone ossifies. Cartil age serves as the blueprint for subsequent bone and joint morphogenesi s as well as tendon and ligament insertions15. Pre-cartilage condensation observed in ear ly limb buds is a transient phase of skeletogenesis, in which a cartilaginous framework serves as a scaffold for la ter ossification of skeletal elements17. To form cellular condensations, cells actively move toward a center, resulting in an increase in cells per unit area rather than an increase in cellular proliferation17,18. Mesenchymal cell condensation is crucial for chondrogenesis and is associated with an increase in cell-cell and cell-matrix interactions through cellcell adhesion molecules and gap juncti ons, which facilitate intercellular communication17,19. Cell-cell interactions presum ably trigger one or more signal transduction pathways that initiate chondroge nic differentiation. N-cadherin and N-CAM are two cell adhesion molecules expressed in condensing mesenchyme, and they play a role in mediating mesenchymal condensation. N-cadherin is responsible for cell-cell adhesion and its expression likely modulates the progression of chondrogenesis. The actions of N-cadherin can be modulated by Wnts a family of secreted glycoproteins that influence cellular condensation and chondro genic differentiation in early development20. During embryonic development, MSCs give rise to cartilage of two types: permanent and transient. Permanent hyaline cart ilage arises from MSCs at the distal ends of developing bones. Following the init ial cellular condensation event, MSCs 26

PAGE 27

differentiate toward stable chondrocytes th at synthesize the extracellular matrix of articular cartilage. Transient cartilage forms a skeletal framework which is later replaced by mineralized bone during the process of endochondral ossification. When considering the candidate biologi cal factors for enhancing the repair response of cartilage, members of the transforming growth factor(TGF) superfamily are of parti cular interest because of their abilities to promote chondrogenic activity in vitro and in vivo The TGFsuperfamily comprises structurally related regulatory molecules that include the five TGFisoforms, bone morphogenetic proteins (BMPs), growth/differentiation factors (GDFs), ac tivins, inhibins, nodal, and glial-derived neurotrophic factor (GDNF)21. TGF. TGFhas been shown to induce mesenchymal chondrogenesis in cultures of cell lines and primary adult MSCs22,23,24,25 and has been implicated in embryonic cartilage formation26. TGFalso regulates the growth and synthetic processes of chondrocytes, stimulating extr acellular matrix synthesis and chondrocyte proliferation in ce ll and organ culture27,28,29. Similar increases in matrix synthesis have been reported in chondrocyte cultur es genetically modified with TGF1 cDNA30. Administration of recombinant TGF1 in vivo has been reported to increase proteoglycan synthesis and restore proteoglycan levels in the knees of arthritic mice31. TGF1 promotes cellular proliferation and in itiates and maintains chondrogenesis of mesenchymal progenitor cells32. Other cell types, including both embryonic and adult fibroblasts, are able to undergo differentiation into a chondrogenic phenotype in the presence of chondrogenic inducers33, demonstrating that TGFcan influence differentiation of more cell types than just chondroprogenitors. TGF-betas have been 27

PAGE 28

shown to activate intracellular signal ing cascades, particularly those cascades containing MAP kinases, p38, ERK-1 and JNK to promote cartilage-specific gene expression32. Bone morphogenetic proteins. Bone morphogenetic proteins (BMPs) make up nearly one-third of the TGFsuperfamily and are closely linked to the formation of bone, cartilage and connective tissues in vivo34,35. BMPs regulate a diverse range of developmental processes during embryogen esis and postnatal development, and they control the differentiation of musculoskeletal tissues in cluding bone, ca rtilage, tendon, and ligaments36 (Table 1-2). The homeostasis of arti cular cartilage in the joint is maintained as a balance between anabolic morphogens such as bone morphogenetic proteins (BMPs) and cartilage-derived mor phogenetic proteins (CDMPs) and catabolic cytokines such as IL-1, IL-17, and TNF15. Studies both in vitro and in vivo have shown that BMP signaling is required for the formation of cartilaginous condensations and fo r the differentiation of precursor cells into chondrocytes37. Like TGF, BMPs -2, -4, and -7 can stimulate chondrogenesis of mesenchymal progenitor cells in vitro38,39,40. BMP-2 is shown to accelerate the healing of osteochondral defects in vivo when delivered via a collagen sponge41,42. BMP-7 is used clinically for spinal fusion an d fracture repair in the long bones43. Implantation of periosteal derived progenitor cells genetically modified to express BMP-7 improved repair of rabbit osteochondral defects, i ndicating the efficacy of a combined gene therapy and tissue engineering approach44. Further clinical applications of BMP-7/OP-1 include incorporation of the recombinant human protein in to collagen scaffolds for repair of full-thickness canine osteochondral def ects. Repair tissue in treated canine knee 28

PAGE 29

defects had a hyaline cartil age-like appearance and integrat ed continuously with the intact cartilage adjacent to the defect45. Both BMP-2 and BMP-7 im prove cartilage repair in studies of artificial cartilage plugs c ontaining either collagen or hydroxyapatite blended with biodegradable polymers46. BMP-2 and BMP-4 are equally effective in promoting chondrogenesis in primary hMSCs in aggregate culture, but BMP-4 aggr egates show a lower tendency to progress toward hypertrophy, a crucial characteri stic to consider for cartilage repair47. Local delivery of BMP-4 by retrov irally transduced MDSCs (muscle-derived stem cells) shows enhanced chondrogenesis and significantly improved articular cartilage repair in rats48. IGF-1. Insulin-like growth factors are another class of molecules with potential for improving endogenous repair. IGF-1 is a major regulator of matrix synthesis in articular cartilage as it stimulates chondrocyte me tabolism and promotes healing of cartilage lesions in vivo ,49,50 but its role in chondrogenesis is less clear. Ad.IGF-1 effectively transduces chondrocytes, MSCs, and synovial cells resulting in IGF-1 production sufficient to stimulate matrix gene expression and proteoglycan production49. When chondrocytes are cultured in vitro and exposed to varying concentrations of Ad.IGF-1, the cells are readily transduced; they produc e significant amounts of IGF-1 to promote increased cartilage matrix gene expression and resist de-differentiation for 28 days51. IGF-1, when combined with TGF, can have synergistic effects in promoting chondrogenesis52. Sox9. In addition to growth factors, resear ch has identified increasing numbers of biological molecules that are involved in regulation of chondroge nic differentiation, express ion of cartilage matrix genes, and acce lerated repair of cartilage defects in 29

PAGE 30

animal models. Of these, the transcription factor SRY-related HMG box gene 9 (Sox9) is most closely associated with the expr ession of cartilage ECM genes and cartilage formation53,54,55,56. Mutations in Sox9 result in ca mpomelic dysplasia, a semilethal skeletal malformation syndrome and XY sex reversal57. Sox9, dubbed the chondrogenic mast er gene, binds to regul atory sequences in the promoter region of seve ral cartilage genes, thus enhan cing their expression in chondrocytes56. Sox9 regulates the expression of chondrogenic genes such as aggrecan and collagen types II, IX and XI during chondrocyte differentiation52. In the collagen type II 1 gene (Col2a1), Sox9 binds within a 48 bp enhancer region located in the first intron and acts in concert with tw o cofactors: long form of SRY-related HMGbox gene 5 (L-Sox5) and SRY-re lated HMG-box gene 6 (Sox6)58. The two cofactors are normally expressed in MSCs; therefore, So x9 gene delivery is sufficient to enhance the level of chondrocytic genes58. Because it functions in tracellularly and cannot be delivered in soluble form, gene transfer of So x9 cDNA to mesenchymal progenitor cells offers a means to invest igate the reparativ e potential of this molecule. Ihh and PTHrP. Chondrocyte proliferation a nd maturation are key points of regulation that may influence a repair pr ocess based on differentiation of progenitor cells. Indian hedgehog (Ihh), a member of the hedgehog family of cell surfaceassociated ligands, is expresse d in prehypertrophic chondrocyt es of the grow th plate, and functions to inhibit chondrocyte hyper trophy by maintaining expression of parathyroid hormone related peptide (PTHrP) through a negative feedback loop59. Altering the expression of these proteins dur ing chondrogenesis may serve to delay the onset of hypertrophy, chondrocyte apoptosis, and formation of bone, while increasing 30

PAGE 31

the pool of proliferating chondrocytes. Indeed, addition of PTHrP has been shown to inhibit chondrocyte hypertrophy during in vitro chondrogenesis of primary adult MSCs60,61. BMP Receptors and Signaling Two major types of membrane-bound seri ne/threonine kinase receptors are required for BMP signal transduction: BMPR-I and BMPR-II. Additionally, there are two options for intracellular signaling: the Smad and MAP ki nase pathways. When BMPs bind to pre-formed heteromeric BMPR-I:B MP R-II receptor complexes, the Smad pathway is activated36, whereas BMP ligand binding that induces the formation of heteromeric receptor complexes induces th e MAPK (mitogen-activated protein kinase) pathway35. The BMP-specific Smad proteins include receptor-regulated Smads 1, 5, and 8 (R-Smads), a common-par tner Smad 4 (Co-Smad), and inhibitory Smads 6 and 7 (I-Smads). During joint morphogenesis, BMP binding prot eins play a role in defining the boundaries between muscle, cartilage, peric hondrium, and tendon/ligament36. BMP signaling is constrained on many levels by antagonists such as Noggin, Chordin, follistatin, ventropin, twisted gastr ulation, Gremlin, Cerberus, and DAN 62. Noggin binds to BMP-2 and -4 with high affinity and blo cks their interaction with BMP receptors. Chordin also binds to BMP-2 and -4 to govern pattern formation (as originally studied in Drosophila ). Antagonists, including Chordin, are proteolytically activated by an MMP, BMP-1, which was misidentified upon it s discovery. BMP-1 is not a BMP per se rather, it is a BMP inhibitor. DAN family mem bers are newly identif ied BMP antagonists based on screens of cDNA libraries; their role in articular cartilage development and homeostasis as well as arthritis is not yet clear63. 31

PAGE 32

Cartilage Repair Arthritic conditions affect over 70 million adul ts in the United States and cause an economic burden in excess of $60 billion annually. These numbers are projected to climb substantially as the population incr eases in age and the incidence of obesity continues to skyrocket. Pathological conditions, such as ost eoarthritis and rheumatoid arthritis, and traumatic conditions, such as intra-articular fracture or cartilage tearing from ligament injury, all yield damage to articular cartilage64. Articular cartilage defects fall into two categories: partial and full-thickness. Partia l thickness defects are limited to the cartilage layer only, and do not penetrate through to t he subchondral bone. In the absence of blood, a reservoir of stem cells and growth fact ors, there is little pot ential for the defect to repair spontaneously. These defects deteriora te with time and can lead to additional problems as the synovial lining becomes irrita ted by loose cartilage flaps, and the knee locks due to cartilage detachment65. Full-thickness defects, on the other hand, penetrate through the cartilage to the subchondral bone c ausing rupture of the local vasculature and access to the marrow. This provides an avenue for progenitor cells from the bone marrow to enter the defect and promote spontaneous healing through the formation of fibrocartilaginous repair tissue. The Bodys Natural Approaches to Cartilage Defect Repair Unlike bone, which has great regenerative potential, cartilage has no vascularity or innervation; therefore, it has a low innate capability for self repair and regeneration. Injury to cartilage usually hea ls through formation of a fibroc artilage scar. Fibrocartilage, consisting predominantly of type I collagen fibrils with unordered proteoglycans and a random cell arrangement, has inferior mechani cal and biological properties compared to 32

PAGE 33

the highly ordered network of type II co llagen fibrils, proteoglycans and chondrocytes present in hyaline cartilage. Over time, fi brocartilage repair tissue degenerates, resulting in permanent loss of structure and function and leading to severe pain66. Although they do not lead to cartilage healin g, palliative therapies are the pri mary approach to treat symptoms of knee lesions. A combination of physiotherapy to maintain range of motion and strengthen the affected limb, weight loss to decrease forces exerted upon the knee, and NSAIDs for pain relief are among the conservative treatment procedures commonly used. Intra-arti cular injections of analgesics along with steroids will relieve pain, and injections with hyaluronate promote increased lubrication and decreased friction on joint surfaces, sho wing a modest improvement over placebo in clinical studies11,65. Surgical Approaches to Cartilage Repair Nonreparative restorative techniques Nonreparative restorative techniques for damaged cartilage include debridement, chondral shaving, and joint lavage11. These techniques can be performed arthroscopically, and are thought to relieve pain and improve mobility, but they do not on their own restore the structur e and function of diseased cartilage67,68,69. One major drawback of nonreparative cartil age restoration techniques is that long-term benefits are reduced due to a loss of chondrocytes at the border betw een healthy and damaged cartilage after the injured cartilage is removed70. Debridement. The debridement procedure, developed by Magnuson in 1941, involves removal of inflammatory cells and other fragments such as chondral flaps, osteophytes, torn ligaments, degenerated m enisci and other debris resulting from arthritis-mediated joint damage71,72,73. Debridement is int ended to eliminate the 33

PAGE 34

biochemical and mechanical factors that caus e arthritis symptoms, resulting in pain relief and improvement of joint functions. The effects are short-lived and somewhat unpredictable, with 1/3 to 2/3 of patients showing improvem ent of symptoms at followup evaluation73,74. Chondral shaving. Chondral shaving, developed in 1908 by Budinger, excises damaged cartilage to reliev e pain, using a motorized shaver. Shaving the cartilage is thought to convert the fibrillat ed damaged surface to a smooth surface; however, further examination has shown t hat shaving yields a rough surface with grooves11. Adverse effects on chondrocytes and cartilage matrix c an result from chondral shaving, most notably chondrocyte death from the heat generated by frictional resistance to the tools used. Additionally, there is a risk of chondr al tears that do not heal, leading to progressive degeneration of the remaining cartilage4,75. Knee joint lavage. Joint lavage is the technique of rinsing the joint with a physiologica l fluid to remove debris76. It is frequently coupled with a debridement procedure77, and is usually performed when more conservative treatments, such as debridement and chondral shaving, prove inadequate. Suction of the fluid removes degradation products, inflammatory cells and degradative enzymes; however, the irrigation fluid can potentially harm cartilage and does nothing to halt disease progression11,69. Reparative procedures Reparative strategies, also termed marrow st imulation techniques, aim to initiate bleeding from the subchondral bone, which releases progeni tor cells, among other cells with chondrogenic potential, from the vascular system to the site o f cartilage injury78,79. 34

PAGE 35

This causes a blood clot to form, pluggi ng the injury site and paving the way for chondroprogenitor cells to gener ate a fibrocartilage scar11,80. Although fibrocartilage has inferior mechanical capabilities, it does serv e a purpose in filling the chondral defect and covering the underlying bone, t hus reducing pain and swelling65. While fibrocartilage fills the site of cartilage injury and temporarily allevi ates pain, it is functionally inadequate as a replacement for articular cartilage in the long term58. Reparative procedures include arthroscopic abrasion arthroplasty, subchondral drilling, microfracture and spongialization, a modified technique that combines debridement and suchondral drilling. Arthroscopic abrasion arthroplasty. Arthroscopic abrasion arthroplasty is a minimally invasive procedure that involves burring the expos ed bone to access the vasculature of the subchondral plate, which promotes formation of a blood clot and subsequent formation of fibrocartilaginous r epair tissue. In short term follow-up, nearly half of the patients showed improvement; however, some studies have shown that breakdown of repair tissue can occur as early as one year after the procedure11. Other studies indicate that when the joint is proper ly protected, fibrocartilaginous repair tissue maintains integrity for up to six years76. Microfracture. The microfracture technique was developed by Steadman to enhance chondral resu rfacing by allowing th e influx of marrow elementsmesenchymal stem cells, growth factors, and other prot einsto create a microenvironment that would promote new tissue formation and take advantage of the bodys natural healing process79. This arthroscopic procedure involves a preparative debridement step to remove the damaged cartilage and expose the subchondral bone followed by a V35

PAGE 36

shaped piercing using a specialized awl which makes multiple perforations approximately 3 mm apart76,79. Subchondral piercing causes bleeding from the bone marrow and leads to formation of a blood clot populated with platelets, growth factors, and progenitor cells, which adheres to the ex posed bone surface, fills the defect, and progresses toward a cartilage-like repair tissue11,79. In younger, active patients (<35 years of age), microfracture remains the procedure of choice for cartilage lesions smaller than 2.5 cm81. Long-term follow-ups from 2 to 12 years report pain relief and restored knee function for 75% of patients with deep subchondral defects4,60,76. Conversely, another study shows a decline in positive clinical outcomes 2 years postoperatively, especially in older patients82. Drawbacks to microfracture include the poor biomechanical nature of the resulting fibrocartilage repair tissue, incomplete filling of the defect, and the potential for abnorma l bone growth into the cartilage lesion76. Physical rehabilitation, typically with continuous passive motion and protected weight bearing, is an important step in improving cartilage repair a fter microfracture, possibly due to increased movement of synovial fluid throughout the joint space, which carries nutrients to chondrocytes and enhances synthesis of matrix proteins83,84. Subchondral drilling. Drilling into the subchondral bo ne is a technique established by Pridie in 1959 to develop bleeding channels through the subchondral bone to promote the formation of cartilage to resurface the exposed bone. It is reported that 85% of patients whose knees were treated with this procedure demonstrated satisfactory long-term outcomes76; however, in a rabbit model, the newly formed cartilage repair tissue lost its hyaline-like morphology within 1 year of treatment and instead resembled dense collagenous tissue85. Adverse effects of drilling include 36

PAGE 37

damage to the subchondral bone from heat generated during the pr ocedure and the potential for formation of a subchondral hematoma. Spongialization. Spongialization is a modification of debridement and drilling, where the entire injured cartilage and the subchondral bone beneath it are removed, which exposes the spongy, or cancellous, bone86,87. In contrast to subchondral drilling, spongialization removes the highly innervated subchondral plate, thus removing a source of pain, and it may promote im proved healing. Follow-ups by Ficat, the developer of this procedure, showed improved joint function and pain relief in 70-80% of patients, but this procedure has not gained popularit y, perhaps because it is invasive and may cause thermal necrosis of the surr ounding cells that are the target for stimulation86,88. Restorative strategies Restorative strategies for the joint include high tibi al osteotomy (HTO), knee replacement (either total or unicompartmental), and transplant ation of bone, cartilage, or tissue. Knee replacement procedures involv e removal of bone from femoral and tibial surfaces and resurfacing wit h prosthetic implants. High tibial osteotomy. High tibial osteotomy (HTO ) is a technique developed by Jackson in 1958, and is used in patients wh o experience cartilage degradation in the medial com partment of the k nee. This procedure aims to transfer weight bearing from damaged regions to those with healthy cartil age to relieve pain and prevent further osteochondral damage11. Especially useful in younger, active patients who are not ready for knee replacement, HTO shows sati sfactory clinical outcomes in 80% of patients 5 years post-op, and 60% at 10 years after treatment11. Since HTO is only 37

PAGE 38

useful in a select group of patients, knee re placement is often the technique of choice for middle-aged and elderly patients since it has reliable long-term outcomes. Partial knee replacement. With partial knee repl acement, only the damaged surfaces of the knee are replaced with prostheses while intact surfaces are not altered. Typically, younger patients take advantage of partial knee replacement because it is less invasive, has a lower cost and has a shorter recovery time than total knee replacement surgery. Even thoug h partial knee arthroplasty has a 98% survival rate at 10 year follow-up and delays the need for invasive total knee arthroplasty by 10 years or more, the cartilage in surrounding areas progressi vely deteriorates due to osteoarthritis as shown by radiographs of patients knees. Total knee replacement. When conservative treat ments fail, total knee replacement, or total knee arthroplasty (TKA), is recommended as an effective and durable procedure to restore mobility and relieve pain in people suffering from endstage knee lesions11. TKA is the most invasive of all the procedures and can be performed as revision surgery for each of the aforementioned proc edures. Prostheses consist of a femoral and tibial component and some also involve patellar resurfacing11. The anterior cruciate ligament (ACL) is always removed and in some instances the posterior cruciate ligament (PCL) is substitu ted or removed. For el derly patients, total knee arthroplasty is currently the recomm ended technique to restore function of the entire knee after articular cartilage damage11. In patients greater than 70 years old, the 10 year follow-up success rate remains high, but in patients less than 50 years old, the failure rate is higher, likely due to the l ongevity and vigor of joint use. The average 38

PAGE 39

duration of implant survival remains 10-15 y ears, with some reports of 20 years. When the implant degrades, revision surgery is necessary to restore joint mobility. Reconstructive methods to treat cartilage defects Reconstructive methods aim to fill the cartilage defect with autologous cells or tissues, and surgeons attempt to carry them out in the least invasive method possible80. These methods use either pieces of cartilage tissue (mosaicplasty) or autologous chondrocytes harvested from non-loadbearing areas (autologous chondrocyte transplantation)58. The main drawbacks are that there is a very limited supply of nonweight-bearing cartilage tissue available for harvest, and collecting tissue introduces new sites of damage to the arti cular cartilage of the joint58. Osteochondral transplantation or mosaicplasty. Osteochondral transplantation (osteochondral grafting) includes autologous and allograft transplants. Autologous osteochondral transplantation (OATS), also te rmed mosaicplasty, consists of removal of cartilage from the defect site down to t he subchondral bone, followed by creation of small holes 15 mm deep and perpendicular to the cartilage surface11,80,89. Next, osteochondral grafts 10-15 mm long are harvested from non-we ight-bearing surfaces of the patients articular cartilage and inserted into the donor sites. As the site heals, fibrocartilage is the most pr evalent repair tissue observed11. While mosaicplasty promotes pain relief and joint function improvement, there is risk of donor site morbidity, the original cont our of the femoral condyle may be difficult to recreate, and chondrocyte death can result in degeneration of the graft. Additionally, there is a chance of bone and cartilage collap se. Removal of cartilage from non-loadbearing areas of the joint may lead to defects at those donor sites, resulting in more pain and additional cartilage dam age from the lack of chondrocyt es available for repair. 39

PAGE 40

Allograft osteochondral transplantation is indicated for large osteochondral defects that exceed those which can be repaired via mosaicplasty and other less invasive techniques. Allografts replace the injured cartilage and its underlying subchondral bone with a histocompatible fres h or fresh frozen cartilage segment taken from an organ donor. Chondrocyte viability and biomechanical integrity deteriorate over time in storage; therefore, fresh osteochondral allograft (less than 2 weeks old) is recommended to optimize the success of the procedure11. Long-term follow-ups report 85% and 74% graft survival at 10 and 15 years post-operatively11. Autologous chondrocyte transplantation. Autologous chondrocyte transplantation (ACT), also dubbed autologous chondrocyte implantatio n (ACI), is a cellbased technique first described by Brittberg in 1994, to relieve pain and restore function in knees affected by either chondral or osteochondral defects11,90. Chondrocytes are harvested arthroscopically from a non-weightbearing portion of the joint and cultured for approximately 6 weeks. After cell expansion in vitro the cultured cells are surgically introduced to the cartil age lesion of the patient11,80. To prepare the defect, the site is debr ided and covered by a periosteal flap harvested from the patients tibi a or femur. The periosteal fl ap is sealed or glued to the edge of the defect, and the cultured cells ar e delivered and held in place underneath the flap. The cells in suspension eventually atta ch and secrete extracellular matrix. ACT is useful for medium to large sized defects with positive results for up to 10 years90. It is reported that in the majority of patients (80-90%), hyaline-like repair tissue forms and pain relief and joint mobility improve89,91,92. Furthermore, integration of the newly formed repair tissue with the surrounding cartilage is reported in nearly 90% of patients at five 40

PAGE 41

year follow-up92. Additional studies show, however, a high proportion of fibrocartilage present in the repair site, further demonstrat ing the difficulty of regenerating articular cartilage93. Disadvantages of ACT include the extrem e expense, leakage of chondrocytes from beneath the periosteal flap at the re cipient site, the need for two separate surgeries for harvest and im plantation, the risk of chondrocyte dedifferentiation in monolayer cell culture, uneven dist ribution of cells introduced to the defect, and long recovery time following the operation94. Further complications following ACT include fibrous overgrowth of the periosteal flap and separation and detachment of the periosteum from the repairing cartilage95. An improved variation of ACT applied clinically is known as characterized chondrocyte implantation (CCI), where patients chondrocytes are sorted and selected based on cell surface antigens that indicate greater potential to produce hyaline cartilage. This method is currently market ed at ChondroCelect by Tigenix of Belgium11. Geneand Cell-Based Approaches to Cartilage Repair What we observe is that no matter the restorative or reconstructive procedure used, the fibrocartilage repair tissue or t he prosthesis eventually degrades. Nothing can match the native articular cartilage originally produced by the patient. The goal of tissue engineering, gene-, and cell-bas ed therapies is to replicat e the natural structure and durability of articular cartilage to enable long-lasting repair of damaged cartilage. Ideal strategies would restore the structural and functional elem ents of articular cartilage through transplant of a viable cartilage substitute, revised methods to enhance cartilage repair, or methods to prevent or halt cartilage damage. 41

PAGE 42

Clearly there is a need for a cartilage repair strategy that introduces autologous, multipotent cells to the joint in a way that enables them to integrate and survive within the existing cartilage structure. MSCs appear as viable candidates as they are found abundantly in tissues throughout the body22,23. MSCs: the logical cell type for chondrogenesis Cells derived from mesenchymal tissues with the ability to prolif erate extensively, self-renew, and undergo multipotent differentia tion are broadly def ined as mesenchymal stem cells, or MSCs. Experiments perform ed by Friedenstein and others described the presence of mesenchymal stem cells (MSCs) in the bone marrow that could be isolated through their intrinsic property to adhere to tissue culture plastic96,97. These early observations led to the standard accepted assay used to identify MSCs, the colonyforming unit-fibroblast (CFU-F) assay, which identifies a dherent, spindle-shaped cells that proliferate to form colonies. There are no definitive surface markers for MSCs; however, they are frequently positive for STRO-1, CD73, CD90, CD 105, CD106, CD146, CD166 and negative for CD45, CD11b, CD34, CD31, CD11798,99. Rather, they are characterized by their ease of isolation and their rapid growth in vitro while maintaining their differentiation potential, allowing for extensive cult ure expansion to obtain la rge quantities suitable for therapeutic use. These properties make MSCs ideal building blocks for tissue engineering efforts to regenerate tissues and repair st ructures damaged by arthritic conditions and cartilage injuries100. MSCs offer advantages for cell therapy because they are easier to culture and manipulat e ex vivo than chondrocytes, which dedifferentiate when removed from thei r native ECM and grown in monolayer. 42

PAGE 43

MSCs can be readily harvested from mult iple tissue sources including bone marrow, adipose tissue, synovial fat pad and periosteum101. Cells with stem-like characteristics can also be harvested from skin, liver, skeletal muscle, dental pulp and cartilage102,103,104. MSCs have been shown to differentiate into chondroprogenitors, osteoblasts, adipocytes, myoblasts and hepatocytes in vitro105. Mesenchymal stem cells contribute to the regeneration of mesenchymal tissues throughout the body such as bone, cartilage, muscle, ligament, tendon, adipose and marrow stroma106. MSCs play an important role in human dev elopment, growth, repair, regeneration, and homeostasis. Their multilineage potentia l make s them a useful model to investigate mechanisms of cartilage tissue development and regulation, especially following treatment with protei ns from the TGFsuperfamily32. Few reports exist of human MSC implantation for cartilage repai r, but studies are taking place to evaluate the use of MSCs rather than chondrocytes for repai r of cartilage defects in the knee. It is thought that with increas ing age there is a signific ant decline in the abundance of MSCs, their lifespan, and t heir differentiation potential. This is a logical assumption since it has been observed that as chondr ocytes age they generate smaller, less functional ECM proteins and are less respon sive to growth factors and mechanical stimuli, although an equally po or repair response of cartilage to damage is observed throughout embryonic, immature, and mature cartilage107,108. However, MSCs extracted from the synovial fat pad hav e been shown to maintain their osteogenic differentiation potential throughout life and could perhaps o ffer more chondrogeni c potential than bone marrow-derived MSCs109. 43

PAGE 44

MSCs provide an autologous cell source, elim inating much of the risk of disease transmission and rejection of donor cells. There is also controversial evidence that MSCs may have immunosuppressive effects110. MSCs can be prompted toward chondrogenic differentiation to provide cells fo r direct delivery to articular cartilage defects or to provide cells to seed scaffo lds which are then implanted into defects. When used in a gene therapy approach, MSCs can be genetically enhanced to express specific growth and differentiation factor s which could not only influence their differentiation in a paracrine manner, but also stimulate neighboring ce lls present in the cartilage implant site. Stem cells can serve as vehicles for gene delivery to damaged articular cartilage by transfecting cells with recombinant cDNAs encoding chondr ogenic proteins and growth factors52. Furthermore, viral transduction of MSCs with chondrogenic cDNAs can result in longer expression of the gene product and gr eater potential to influence cartilage defect repair. Mechanical stimulation to promote cartilage Mechanical stimulation is a natural co mponent of the c hondrocyte environment and is known to affect gene expression and re-differentia tion of chondrocytes58. Chondrocytes, and other cell types such as MSCs, respond to dynamic compression by changing their gene expression profile58. In therapeutic applications it is important to remember that controlled movement postimplantation could be essential for adequate healing and development of articular cartilage rather than fibrocartilage. Regeneration and Repair Using MSCs: Gene Therapy and Orthopaedics MSCs are ideally suited for cartilage gene therapy applications as they can be prompted to differentiate along a chondroge nic lineage by growth and differentiation 44

PAGE 45

factors. Members of the TGFsuperfamily play an integral role in cartilage development; thus, they are logical choice s for cartilage development in vitro. Gene delivery using viral vectors enables rapid, ro bust expression of chondrogenic proteins in MSCs, prompting their differentiation al ong a chondrogenic lineag e. Such modified MSCs can then be implanted into cartilage def ects, where they will continue to express the transgene products, thus impacting nei ghboring chondrocytes embedded in the ECM. Differentiated MSCs can take the pl ace of the limited num ber of chondrocytes that are currently harvested for transplant procedures and will serve to fill cartilage defects with chondroprogenitor cells which secr ete cartilaginous matrix and repair the site of damage with a natural tissue. As the field of gene therapy expands, num erous approaches are being explored to improve trea tments for cancer, AIDS, hemophilia, cystic fibrosis and an array of other diseases. Studies of gene therapy applicati ons for orthopaedic conditions arose from research directed toward the tr eatment of rheumatoid arthritis111. Rather than a gene therapy method to correct a genetic abnormality, this work was based upon the concept of using gene transfer as a protein delivery syst em to treat chronic joint disease. By delivering cDNAs encoding naturally-occurring ant i-inflammatory or anti-arthritic proteins to cells in the synovial lining, these cells would then serve as factories for the local overproduction and secretion of the therapeut ic proteins into the joint space and surrounding tissues. Studies in several labo ratories have shown that exogenous cDNAs can be efficiently delivered and expressed at levels within the joint sufficient to have beneficial effects in a va riety of animal models112. 45

PAGE 46

The success of gene therapy studies in rheumatoid arthritis has led to the exploration of the application of gene delivery for other ort hopaedic applications such as bone,113,114,115 ligament, tendon,116,117 and cartilage44 healing. Most of these investigations use gene delivery in the same manner, as a system for localized, sustained production of bioactive molecules to promote healthy re generation. Gene transfer can be used as a means to achieve su stained synthesis of specific proteins within a cartilaginous lesion, and this can be used to augment the differentiation of mesenchymal stem cells toward chondrogenesis in vivo Gene transfer can also be used to stimulate existing cells and tissues, su ch as muscle, to repair large segmental defects in bone118. A similar technique can be applied to repair cartilage defects with chondrocytes (ACT) or MSCs that are st imulated by virally-delivered chondrogenic transgenes. 46

PAGE 47

A). B). Figure 1-1. The four zones of articular cart ilage stretch from the su perficial surface to the deep zone, where hypertrophic chondroc ytes are replaced by c alcified matrix. A) Growth plate of proximal tibia. Columns of prehypertrophic chondrocytes are present in the radial zone, progressing to enlarged, rounded chondrocytes that undergo hypertr ophy, followed by apoptosis and replacement by calcified cartilage. B) Depiction of the zones of articular cartilage with resident chondrocyte popul ation. The collagen fibrils, depicted as pale gray lines, differ in their orientation throughout the zones to allow movement, provide structural support, lim it the expansion of proteoglycans, and absorb stresses. (Part B adapted from Ahmed and Hincke, Tiss Eng Part B 2010). 47

PAGE 48

Figure 1-2. The human knee cont ains specialized structures to withstand the forces of movement. Depicted here, articular cartilage lines the ends of the femur and tibia and the inner patella to enable near frictionless motion of the knee joint. The specialized architecture of arti cular cartilage can withstand repeated extreme forces, often for more t han 80 years, without damage. Image adapted from Ahmed & Hincke, Tiss Eng Part B 2010. 48

PAGE 49

Figure 1-3. The arrangement of st ructures in the articular cart ilage matrix is designed to absorb forces, especially compression. Collagen fibers form a supportive framework for cartilage, and they c onstrict the expans ion of hydrated proteoglycans. Chondrocytes (not shown) the resident cells of cartilage, populate the matrix at a low dens ity and serve to remodel the extracellular matrix. Image adapted fr om Mow and Ratcliffe, Biomechanics of Diarthrodial Joints Structure and Function of Articu lar Cartilage and Meniscus, 1990. 49

PAGE 50

Figure 1-4. A proteoglycan aggregate is made up of ma ny proteoglycan subunits attached to a hyaluronic acid backbone vi a link protein. Proteoglycans consist of a protein core with glycosaminoglyc an side chains of keratan sulfate or chondroitin sulfate attached. The negativ e charges of the side chains repel one another and form their characteristic bottl ebrush-like structure. Image adapted from Mow and Ratcliffe, Biomechanics of Diarthrodial Joints Structure and Function of Articu lar Cartilage and Meniscus, 1990. 50

PAGE 51

Figure 1-5. Fibrillar collagens, including collagen type II, form a characteristic triple helix. Multiple tripeptides are bundled t ogether to form fibrils, which in turn, are packed into parallel bundles to yi eld fiber s with incredible strength. Collagen type IX, with its flexible kink, cr osslinks collagen type II fibrils to proteoglycans in the cartilage extrace llular matrix. Figures are adapted from Sigma-Aldrich Life Scienc es, 2009, and Lodish, et al. Molecular Cell Biology, 1999. 51

PAGE 52

Table 1-1. Collagen types. Collagen type Representative Tissues References Fibrillar collagens I Abundant throughout human body; skin, bone, scar tissue, fibrocartilage, tendon, ligaments, dentin, artery walls, interstitial tissues 2,119 II Articular cartilage and vitreous humor of the eye 2,119,120 III Produced by young fibroblasts prior to type I collagen, reticular fibers, skin, muscle, blo od vessels, intestines, uterus 2 V Interstitial tissues, similar to type I, also associated with placenta (fetal tissues ), synovial membranes and cell culture 2,119 XI Hyaline cartilage 119 Nonfibrillar collagens IV Basal lamina, lens of the eye, part of filtration system of capillaries and glomeruli of the kidne y nephrons VI Most interstitial tissues, similar to type I, blood vessels, skin, intervertebral disc 119 VII Forms anchoring fibrils in dermal epidermal junctions VIII mainly endothelial cells 121 IX Hyaline cartilage, associated with type II and XI fibrils in cartilage, vitr eous humor, FACIT collagen 121 X Hypertrophic and mineralizing cartilage, growth plate 119 XII Embryonic tendon and skin, periodontal ligament, FACIT collagen 119 XIII Endothelial cells 119 XIV Fetal skin and tendon, FACIT collagen 119 Collagens are an essential part of the integumentary system. Of the 29 types of collagen identified and described in literature, the fourteen most prevalent are listed above. Collagen type I is the most abundant collagen in the human body. Collagen type II is the main component of articular cartilage, and is supported by types IX and XI. A subgroup of nonfibrillar collagens, including types IV and VIII, form sheets that create structures such as basement membranes that surround tissues. FACIT, Fibril Associated Collagens with Interrupted Triple Helices, refers to a type of collagen which is also a proteoglycan. FACIT collagens include types IX, XII, XIV, XIX, and XXI. 52

PAGE 53

Table 1-2. Morphogens and Growth Factors in Cartilage Development Morphogen Name Alternate Descriptors Function(s) Bone Morphogenetic Protein (BMP) BMP-2 BMP-2A Cartilage and bone morphogenesis, used clinically for bone repair36 BMP-4 BMP-2B Cartilage and bone morphogenesis BMP-3 Bone formation, inhibits activity of BMP-246 BMP-3B GDF-10 Membranous bone formation BMP-5 Bone morph ogenesis BMP-6 Cartilage hypertrophy, bone formation via alternate mechanism to BMP-2 or BMP-4 BMP-7 OP-1 (osteogenic protein 1) Osteogenic differentiation, used clinically to augment bone repair36 BMP-8 OP-2 Bone formation, esp. active in early phase of fracture healing46 BMP-9 Anabolic factor in juvenile cartilage BMP-10 Not chondrogenic; regulates cardiac gro w th and heart chamber ma turation Cartilage derived morphogenic protein (CDMP) GDF (Growth/ differentiation factor) CDMP-1 GDF-5, BMP-14 Mesenchymal condensatio n, chondrogen esis CDMP-2 GDF-6, BMP-13 Cartilage development a nd hypertrophy, cartilage formation in vitro similar to BMP-2 CDMP-3 GDF-7, BMP-12 Ligament and tendon development, cartilage formation in vitro similar to BMP-2 53

PAGE 54

Table 1-2. Continued. Many growth factors, morphogens, and transcription factors play integral roles in cartilage morphogenesis and are activated during the bodys natural cartilage repair responses. Many of these same factors, especially BMPs, are involved in bone morphogenesis and are delivered in clinical applications to enhance non-union fracture healing. Morphogen Name Alternate Descriptors Function(s) Transcription Factor Sox9 SRY (sex-determining region Y) HMG box 9 Binds regula tory sequences in Col2a1 promoter region; necessary for mesenchymal condensation Other Morp hogens Indian hedgehog (Ihh) Cartilage morphogenesis; esp. in preh ypertrophic chondrocytes 54

PAGE 55

CHAPTER 2 MATERIALS AND METHODS In Vitro Cell Culture HEK293 and 293FT Cell Culture Immortalized cell lines were cultured in 75 cm2 flasks containing Dulbeccos modified Eagles medium (DMEM), with 10% FBS (Gibco), 1% glutamine (Gibco), and 1% penicillin-streptomycin (Gibco ), hereafter referred to as complete DMEM, at 37C in a 5% CO2 environment. For adenovirus propagation, HEK293 cells were grown to approximately 70-80% density in 175 cm2 flasks prior to viral infection. For lentivirus preparation, 293FT cells were cultured in either 75 cm2 or 175 cm2 flasks and treated as indicated below. Harvesting and Culturing Bovine MSCs We harvested bone marrow from the long bones of 3-day old Holstein-Fresian bull calves from the University of Florida D epartment of Large Animal Sciences. The bones were cut open with a table saw, marrow was removed with a spatul a and diluted, to prevent clotting, in MSC medium, whic h contains DMEM (Gibco) with 10% MSCqualified FBS (Gibco), 1% glutamine (Gibco ), and 1% penicillin-s treptomycin (Gibco). Addition of antimycotic agents to the medium were not beneficial as these substances interfered with adherence, growth and expansion of MSCs. The cells were cultured at 37C in a 5% CO2 environment for 48-72 hours, at which point blood cells were removed by changing the media. The plasti c-adherent cells that remained were grown and expanded for up to five passages for use in experiments. 55

PAGE 56

Aggregate (Pellet) Culture The pellet culture system allows cell-cell in teractions analogous to those that occur in precartilage condensation even ts during embryonic development122. We used aggregate culture as a means to evaluate the chondrogenic potential of multiple transgenes in vitro We adapted our high-throughput aggregate culture system based upon results that showed the efficiency of 96-we ll plates over the use of individual 15 mL conical vials when screening numerous treatment groups123. As depicted in Figure 2-1, we transduced early passage bovine MSCs with adenoviral vectors carrying the complete cDNAs for the candidate transgenes, denoted Ad.BMP-2, Ad.BMP-4, Ad.BMP -7, Ad.Ihh, Ad.Sox9, Ad.C TGF, and so on, at doses ranging from 10 virus particles/cell to 100,000 vp/cell. Similar dose ranges were used for scAAV-BMP-4 transduction. MSCs were grown in monolayer in 75 cm2 flasks and transduced with virus. Twenty-f our hours later, cells were trypsinized, counted, and 2.0 x 105 cells in a 300 to 350 l volume were pe lleted by centrifugation within individual wells of 96-well V-bottom polypropylene plat es (Corning). The resulting cellular aggregates were grown for 21 days in serum-free chondrogenic medium. Chondrogenesis was evaluated by gross examination of aggregate morphology, histological staining for proteoglycan expr ession, and immunohistochemical analysis for collagen types I, II, and X. Chondrogenic Media Formulation Cell aggregates in 96-well plates were maintained in 350 l chondrogenic medium, which consists of serum-free DMEM (Gibco), 1% ITS (insulin, transferrin, selenium) (Sigma), 1% penicillin-streptomycin (Gibco), dexamethasone (10-7 M), ascorbate-2phosphate (50 g/mL) (Sigma), proline (40 g/mL) (Sigma), and 1 mM sodium pyruvate 56

PAGE 57

(Gibco). Media were changed every 48 to 72 hours, except for days 3, 7, 14, and 21, when media were changed to allow 24-hour accumulation of proteins for detection via ELISA. Virus Preparation and Transgene Expression Construction and Generation of Recombinant Adenoviruses Containing Chondrogenic Transgenes Serotype 5, E1E3-deleted recombinant ade noviruses containing BMP-2, BMP-4, BMP-7, TGF1, Indian hedgehog, Sox9, GFP, and others were generated through the Cre-lox recombination system developed by Hardy124. Each transgene was inserted directionally into the adenoviral shuttle pl asmid, pAdlox, containing the 3 inverted terminal repeat of the virus, a packaging signal, a cDNA expression cassette driven by the cytomegalovirus (C MV) promoter/enhancer, and a loxP Cre recombinase recognition sequence. Cotransfection of Cr e8 293 cells, which constitutively express high levels of Cre recombinase, with linearized Adlox plasmid and 5 adenoviral genomic DNA flanked by loxP sites generates recombinant adenovirus124. Specifically, Cre-mediated recombination occurs between the loxP site in pAdlox vector and the 3 loxP site in the adenoviral ba ckbone. Any nonrecombined 5 adenovirus that is present can be separated from the recombinant adenovir al particles via subsequent propagation on Cre8 293 cells, whose Cre recombinas e will delete the packaging signal of 5 virus. Adenovirus Propagation and Amplification To generate the quantities of replication-deficient aden ovirus needed for largescale infections, 293 cells were grown in complete DMEM in 175 cm2 flasks. At 70-80% confluence, the media were removed and the cells were rinsed with PBS. A small aliquot (2-5 l) of the desired adenovirus was added to a minimal volume of serum-free 57

PAGE 58

media (10 mL) and applied to the cells for 46 hours, at which poin t 12-15 mL complete DMEM was added. Cells were incubated at 37C with 5% CO2 until cells rounded up and developed a granular appearance (s igns of the cytopathic e ffects of a lytic virus) usually occurring after 2-3 days. Ba sed on the method developed by Palmer125, cells were collected using a cell scraper just befor e the virus caused them to lyse, then transferred, along with the m edia, to a 50 mL centrif uge tube. Following tabletop centrifugation at 2000g for 10 minutes at 4C, the medi a were discarded in bleach, and the pellet was resuspended in 4 mL Tris, pH 7. 0, and stored at -80C. Cells were lysed by three rounds of freeze/thaw, diges ted with benzonase (Sigma) and the final supernatant was collected and stored on ice until cesium chloride (CsCl) gradient purification. CsCl gradients we re prepared by layering approx imately 3 mL of 1.4 g/mL CsCl solution on the bottom, 3 mL of 1.2 g/mL CsCl solution in the middle, and 4 mL of viral cell supernatant on the top layer in ch illed ultracentrifuge tubes (Beckman). After centrifugation at 40,000 g for 1 hour at 4C, viral bands lo calized at the interface of the two CsCl layers. The viral band was harvested by puncturing the centrifuge tube with a needle and syringe. If two bands we re visible, the lowest band containing the infectious particles was harvested. The harvested band was diluted 2to 4-fold in 10mM Tris-HCl, pH 8.0, for recentrifugation. Three consecutive gradients were performed on each viral prep. After the third CsCl gr adient purification, the harvested adenovirus fraction was transferred to dialysis tubing and placed in dialysis buffer at 4C for 6 to 8 hours. Following three rounds of dialysis, t he virus was stored at -80C in 50 l aliquots. Viral titers were estimated by optical density125. 58

PAGE 59

Generation of Lentivirus We produced lentiviral vect ors by implementing a f our plasmid transfection procedure adapted from Invitrogens ViraPower expres sion system. Transducing vectors expressing the desired transgenes (G FP and BMP-4) were generated via the insertion of the specific cDNAs into the pLenti4/V5-DEST vector via homologous recombination. The resulting expression pl asmid was mixed with the three necessary packaging plasmids, denoted pLP1, pLP2, and pLP-VSVG. Plasmid DNA was complexed with lipofectamine and delivered to monolayer cultures of 293FT cells. Twenty-four hours following tr ansfection, the cell culture medium was replaced. Conditioned medium was harve sted at 48 and 72 hours to allow collection of virus. Medium was concentrated through ultracentrifugation at 20,000 g at 4oC for 2 hours. The resulting pellet was resuspended in Opti -MEM (Gibco) and used immediately or aliquotted and stored at -80oC until use. Construction and Generation of scAAV Vectors For generation of scAAV vector plasmids, the cDNA encoding GFP was directionally inserted into the conventional AAV packaging vect or pTRUF2 as a NotISalI fragment. For generation of scAAV vector plasmids, the cDNAs for GFP and BMP4 were directionally inserted into the SacII and NotI sites of the pHpa-tr s-SK plasmid. PCR was used to modify BMP-4 cDNA to in troduce SacII and NotI sites at the 3 and 5 ends. The insert plasmid contains the CMV pr omoter/enhancer and the cDNA of interest surrounded by ITRs from AAV2. The pDG-2 plasmid contains the rep and cap genes from AAV2 and complementing adenoviral f unctions required for amplification and packaging of the AAV genome. Similarly, to generate serotype 5 scAAV, the pxyz-2 plasmid was used. 59

PAGE 60

AAV vectors were propagated using an adenovir us-free, two plasmid transfection system. Using 10-layer cell factories (Nunc), the respective AAV vector plasmids were co-transfected into 293 cells by CaPO4 precipitation with the pDG-2 or pxyz-5 packaging/helper plasmid. Sixty hours post-transfection, cells were harvested with PBS containing 10mM EDTA, pelleted, resus pended in buffer containing 150 mM NaCl and 50 mM Tris, and lysed by three successive rounds of freeze-thaw. Cellular nucleic acids were digested by incubation with Benzonase (Sigma). Purification of AAV from the crude lysate was performed over iodixanol gradients followed by FPLC affinity chromatography over mono-Q columns. T he eluate was desalted and concentrated with a Millipore Biomax 100K filter, aliquotted and stored at -80C. Viral titers were determined by competitive quantitative PCR assay relative to well-characterized AAV viral reference standards. Each viral preparation was examined for purity by resolution of the viral proteins by SDS PAGE and silver stain. Gene Transfer to MSCs to I nduce and Enhance Chondrogenesis Plasmid DNA transfection Monolayer cultures of bov ine MSCs were transfected with DNA-lipofectamine complexes. Liposomes contai ning each BMP-4 expression vector were generated by incubation of DNA wit h lipofec tamine (Invitrogen) in Opti-MEM (Gibco). Following a 20 minute incubation, the complexes were added to cells in a minimal volume of serumfree medium. One to three hour s later, complete medium was added to cells. Medium was changed 24 hours post-transfection. At 48 hours post-transfection, conditioned media were collected and BMP-4 expression was characterized by ELISA. GFP expression was observed visually in the transfection controls. 60

PAGE 61

Transgene delivery using adenovirus MSCs are receptive to adenoviral transduction since they possess the CAR receptor. Adenovirus readily infect s dividi ng and non-dividing cells so this was not a limiting factor in carrying out viral tr ansduction prior to forming cell aggregates. Methods to Detect Transgene Products Western blot Since there is no commercially available ELISA to quantify expres sion levels of Ihh or Sox9, Western blots of conditioned media or cell lysates, respectively, were used to verify expression. MSCs were grown in mo nolayer culture, then transduced with Ad.Ihh or Ad.Sox9 in a minimal volu me of Opti-MEM (Gibco). Tw enty-four hours later, media were removed and replaced with seruml ess DMEM. After an additional 24 hours, conditioned media was removed from Ad.I hh-treated wells for detection of protein expression. Since Sox9 is expressed intr acellularly, Ad.Sox9 infected cells were harvested with a cell scraper and lysed in chil led homogenization buffer. The resulting cell extract was used for protein detection. As these were human transgenes expressed in bovine MSCs, we were able to di stinguish between e ndogenous and exogenous protein production. A BCA assay was completed to gauge total protein content and ensure consistency in loading samples. A total of 10 g total pr otein per lane was loaded into 10% or 15% Tris-HCl pre-cast gels (Bio-Rad ) for Sox9 and Ihh detection, respectively. Proteins were transferred to Immun-Blot PVDF membranes (Bio-Rad) in buffer containing 25mM Tris, 192 mM glycine with 20% methanol and 0.1% SDS. Following transfer, the membranes were blocked with 5% milk for 1 hour prior to application of primary antibodies: 1:2000 rabbit anti-Sox9 (Santa Cruz Biotechnlogy, Inc.) and 61

PAGE 62

1:10,000 rat anti-Ihh (R&D S ystems). Membranes were soaked in primary antibody solution overnight on a low speed orbital s haker at 4C. Following incubation, the membranes were rinsed with TBS-T and secondary antibodies conjugated to horseradish peroxidase were applied for 45 minutes: 1:15,000 anti-rabbit-HRP (BioRad) and 1:12,000 anti-rat-IgG-HRP (Sigma-Ald rich), respectively. Membranes were rinsed and proteins were detec ted with the Immun-Star HRP Chemiluminscent Kit (BioRad). For detection of -actin as a loading control, ant ibodies were stripped from the membrane using 0.5M NaOH for 20 minutes at room temperat ure. As before, membranes were rinsed with TBS-T, and blo cked with 5% milk for 45 to 60 minutes. Mouse anti-actin-HRP antibody (Sigma-Aldrich) was applied to the membranes at a concentration of 1:50,000 overnight at 4C. -actin was visualized via chemiluminescence. ELISA to detect secreted transgene products Concentrations of secreted protein products present in conditioned media were quantified using commercially av ailable Duo-Set ELISA kits (R&D Systems) for BMP-2, BMP-4, BMP-7, and IGF-1 as directed by the manufacturer. The conditioned chondrogenic media from three or more r eplicate aggregates were used for all data points. Unless otherwise noted, media wer e placed onto cells 24 hours prior to collection, then used immediately or stored at -20C until use. Histology and Immu nohistochemistry Aggregates were removed from cult ure after 21 days and fixed in 4% paraformaldehyde solution overnight. The aggr egates were paraffin embedded, cut into 5 m sections, and mounted on plus-charged slides (Fisher Scientific). Slides were deparaffinized and rehydrated through a series of xylenes and graded alcohols, after 62

PAGE 63

which they were stored in water for 2 minut es. Appropriate sections were stained with toluidine blue, while alternat e sections were examined for collagen type I and type II content. Initially, heat mediated antigen retr ieval was performed in Dako Target Retrieval Solution (DakoCytomation) for 20 mi nutes at 95C. While this procedure works well for many tissue samples, the heat proved too intense for aggregates, and they often became detached at this step. We om itted the heat retrieval and found that aggregate sections remained attached for the entire procedure, and its omission did not impact detection of collagen. Samples were treated with chondroitinase ABC (Sigma-Aldrich) at 0.2 U/mL for 15 to 30 minutes at 37C to cleave chondroitin sulfate polysaccharide chains. Nonspecific binding was blocked in 15% normal serum matched to the secondary antibody species. Slides wer e incubated overnight at 4 C or 1 hour at room temperature with commercially available antibodies: rabbi t anti-collagen I (Chem icon) at 1:1000 and rabbit anti-collagen II (Chemicon) at 1:500. Although the primary species reactivity was mouse, the antibodies were shown to cross-react with bovine tissues. The fluorescent secondary antibody, Alexa Fluor 488 donkey ant i-rabbit, was used at a 1:200 dilution and allowed to incubate for 1 hour in the dark. Samples were washed with buffer (TBST), counterstained with Dapi, and coverslipped using Vectashield mountant. RNA Extraction, RT, and rtPCR For each treatment, ten additional pellets were harvested at days 3, 7, and 21 for total RNA isolation. Total RNA was isolated from treated and contro l cell pellets using the RNeas y mini kit (Qiagen) as direct ed. Aggregates were stored in RNALater (Qiagen) until RNA extraction was performed, at which time each group of treated pellets were frozen in liquid nitrogen and pulverized using a mortar and pestle. The 63

PAGE 64

pulverized tissue was added to lysis Buffer RLT, homogenized using a 20-gauge needle, and RNA was harvested using RN easy spin columns following the manufacturers protocol (Qiagen). We completed semi-quantitative real ti me RT-PCR analyses to verify transgene expression in cases where there was no ELI SA for detection (Ihh) or chondrogenic gene products of interest were intracellular (Sox 9) Real time PCR was carried out using the Eppendorf Realplex machine a nd software. To synthesize cDNA, 1 g of total RNA from each group was reverse transcribed using random hexamer primers and MMLV reverse transcriptase (Invitrogen). Specific primer sets were used to amplify type II collagen 1 chain, type I collagen 2, type X collagen 1, aggrecan, osteopontin, and fibronectin. In Vivo Experiments Intra-Articular Injections To examine the effects of chondrogenic tr ansgene expression on collateral tis sues in the joint, we performed intra-articular injecti ons in both knee joints of Male Wistar rats weighing 100-150 g. The adenoviral vectors of interest were suspended in phosphate buffered saline (PBS) to a 50 l volume and in jected into the joint space of the knee through the infrapatellar ligament. The knee diameter was measured with calipers daily, and rats were weighed daily for 7 days. All animal procedures were approved by the Institutional Animal Care and Use Committe e of the University of Florida. Harvesting Tissue, Decalci fication, and Histology Seven day s after intra-articular injection, joint tissues were harvested and stored in 5% formic acid overnight at low speed on an orbital shaker to speed decalcification. Follo wing a PBS rinse, the knees were stor ed in 0.5 M EDTA solution on a low speed orbital shaker at 4C until bones were suffi ciently decalcified (approximately one week). 64

PAGE 65

The decalcified knees were sliced medially a nd cut into 5 m sections for hematoxylin & eosin staining. 65

PAGE 66

Figure 2-1. Assay for in vitro chondrogenesis. We harvested bone marrow from 1 to 3day old bull caves, cultured the plas tic adherent cells and expanded them in vitro To generate cell aggregates, we vira lly transduced cells in monolayer and 24 hours later, pelleted them by c entrifugation. Cells were grown in chondrogenic medium for 21 days, then evaluated for hypertrophic differentiation, cellular proliferation, and matrix protein production. 66

PAGE 67

CHAPTER 3 GENE DELIVERY STIMULATES CHONDROGENESIS OF MSCS Introduction Articular cartilage is a highly specializ ed tissue that allows for near frictionless motion of diarthrodial joints. When cartilage is damaged as a result of injury or disease, natural repair processes are often insufficien t to regenerate the ti ssue due to the lack of vascularity, dense extracellula r matrix (ECM) and low cellular density of cartilage. Typically the repair response, if any, generates a fibrocartilage scar which lacks the unique architecture and structural properties of native articular cartil age. In most cases, though, damage or lesions of a significant size remain permanently. These injuries often initiate a degenerate cycle that over time leads to generalized cartilage loss and osteoarthritis. Biology offers a number of potential approaches to enhancing the natural repair response of bone marrow progenitor cells in vivo As shown in Table 1-2, several growth factors, morphogens and more recent ly, transcription factors, have been shown to promote differentiation along c hondrogenic lineages. While these substances have shown promise in animal models of cartilage repair and regeneration, their clinical application is hindered by deliv ery problems. The half-lives of many proteins are limited in vivo so they are difficult to administer to sites of cartilage damage at therapeutic concentrations for sustained periods of time Localized delivery of these agents without involvement of non-target organs has also pr oven to be problematic. These limitations may be overcome by developing techniques to transfer genes encoding chondrogenic gene products to cells at the appropriate sites and to express those genes locally for the 67

PAGE 68

necessary period of time. In th is manner, the proteins of inte rest are synthesized locally by cells and are presented to the mi croenvironment in a natural fashion. Numerous experimental approaches are currently being explored to enhance cartilage regeneration and repair. These include tissue engi neering and gene and cellbased therapies. Adult mesenchymal stem cells (MSCs) are a well-suited cellular platform on which to base therapies for cartilage repair and tissue regeneration since they have the capacity to self renew and can differentiate into multiple mesenchymal tissues, including cartilage and bone105,106. They are also readily available from a variety of tissue sources, including bone marrow, synovium, fat, skin, and muscle. MSCs maintain their multilineage capacity over several passages in culture, making them amenable to various applications, including ex vivo therapies126. Within articular cartilage, the extracellu lar matrix constantly undergoes remodeling by chondrocytes, which exclusively populate the matrix at low density. Adult MSCs in vivo serve as replacements for differentiated ce lls of mesenchymal tissues that naturally expire or succumb to injury or disease127. This process of native stem cell-generated cell replacement peaks between ages 20 through 29 and decreases with age14,122, but it could be re-charged by the introduction of modi fied MSCs to sites of cartilage damage. To date, no repair strategy has been shown to generate a durable repair tissue that can withstand the functional demands required of articular cartilage in vivo128. Delivery of MSCs alone is not suffici ent to generate appropriate repair tissue in cartilaginous lesions, as the microenvironment is not adequate to drive and maintain chondrogenic differentiation. Gene transfe r to MSCs, however, can be adapted to achieve sustained loca l expression or synthesis of specific protein factors, which may 68

PAGE 69

be used to induce chondrogenesis in vitro and may enhance chondrogenic differentiation in vivo If appropriate stimulatory factors ar e delivered to sites of cartilage damage, it may be possible to trigger si gnaling pathways in resident and introduced cells that drive the cells to synthesize repair tissue identical to the original in structure and form. Growth factors which are delivered as gene products offer advantages over recombinant protein delivery in that the proteins are presented in a natural context, they can be synthesized locally at the site of need for extended periods of time, and may be less costly. We used an aggregate culture system to evaluate the ability of candidate transgenes to stimulate bovine MSCs toward chondrogenesis in vitro High-throughput aggregate culture of MSCs offers a useful means to evaluate cho ndrogenic potential of multiple transgenes in vitro Using a similar method, Johnstone et al. demonstrated chondrogenesis in MSC aggregates in defi ned medium containing dexamethasone and TGF1 resulting in aggregates that synthesized extracellular matrix characteristic of articular cartilage, containi ng proteoglycan and type II collagen22. Growth factors such as TGF, BMP-2, BMP4, BMP7, IGF-1 and FGF-2 have stimulatory effects on cartilage function. Animal studies document the benefits of exogenous growth factors in stimulating MSCs and grafted chondrocytes toward chondrogenesis129. Chondrogenic genes, such as Sox9, I hh, and BMPs 2, 4, and 7 play key roles in the development of the cart ilage anlagen within the embryo and subsequent formation of permanent cartilage, as seen in the joints, or transient cartilage that undergoes replacement by skeletal elements elsewhere in the body. The method of 69

PAGE 70

introducing modified MSCs to cartilage defects has considerable potential to improve the cartilage repair process. This study was performed to evaluate BM P-4, BMP-7, Sox9, Ihh and other genes for their chondrogenic potentia l when expressed by MSCs as transgenes. While there are numerous other factors t hat could be studied, our eval uation includes gene products considered to be among the principle contri butors to the process of chondrogenesis and those reported to be effective in cartilage repair in animal models. These experiments were used to identify the most effectiv e gene and gene combinations for stimulating chondrogenic differentiation of MSCs in vitro Thus, the selected cDNAs should offer a reasonable assessment of the pot ential utility of gene delivery in this application. Rationale Multiple growth factors, morphogens, and transcription factors, including TGF1, BMP-2, BMP-4, IGF-1, Ihh and Sox9, are known to play a role in chondrocyte differentiation and proliferation. Chondrogenic effects of TGF1, B MP-2, and IGF-1, have been well-characterized in an aggregate culture system when delivered to MSCs via adenoviral vectors130. Although TGF1 and BMP-2 showed strong chondrogenic activity, they tended to drive the cells towa rd hypertrophy, a preliminary step in bone formation. With the goal of identifying trans genes whose products may be more suitable for cartilage repair, we expressed BMP-4, BMP-7, Sox9, Ihh, IGF-1, and CTGF as transgenes in MSCs and characterized their relative abilities to induce chondrogenesis within the pellet culture system singly and in co mbination. The results of these studies provide new insight into the biological ac tivity of each cDNA when administered to MSCs as a gene product, and they form the basis for selection of candidate transgenes to evaluate in vivo 70

PAGE 71

Results Our group has shown previously that MS Cs in culture are highly amenable to infection and subsequent transduction with recombinant type 5 adenovirus vectors. For the majority of transgenes tested, adenovir al-m ediated delivery to MSCs yielded robust expression of protein products that typica lly persisted for two to three weeks. The efficiency of adenovirus gene transfer coupled wit h the relative ease of generating novel recombinants and the ability to readily propagate the vector to high titers, led us to use this technology to evaluate the chondrogenic activity of our candidate transgenes. Cre-lox Recombination and Adenovirus Propagation We obtained the cDNAs for BMP-4, BMP-7, Ihh, Sox9 and CTGF from the American Type Culture Collection (ATCC) as IMAGE clones. Following amplification of each assoc iated vector, we isolated and dire ctionally inserted each cDNA into the pAdlox shuttle vector. This plasmid contains the 3 inve rted terminal repeat of the adenovirus, a native packaging signal, and a cDNA ex pression cassette driven by the cytomegalovirus promoter /enhancer followed by a loxP Cre-recombinase recognition sequence. Following verification of each pAdl ox construct by diagnostic restriction digestion, we transfected each into cultures of 293 cells to assay for synthesis of the respective transgene products. Forty-eight hours post-transfection the conditioned media or cell lysates, as appropriate, were a nalyzed qualitatively for expression of the respective transgene products: BMPs -4 and -7 by ELISA; Ihh, Sox9, and CTGF by Western blot (data not shown). Once it was confirmed that the vector constructs were indeed functional, they were used to generate recombinant adenovirus. For this, the respective plasmids were linearized by restriction digest and co-transfected with purified, 5 adenov iral DNA into 71

PAGE 72

cultures of 293 cells engineer ed to constitutively express the Cre recombinase (Cre8 cells). (The 5 adenoviral genome packaging sequence ( ) is flanked by loxP sites: novel recombinant adenovirus is generated by Cre-mediated recombination between the loxP site in the Adlox shuttle vector and the 3 loxP site in the 5 viral backbone.) If the far left hand portion of the viral genome contained in the Adlox plasmid was successfully linked to the adenoviral backbone, plaques in the monolayer of cells were usually detectable within 5-7 days. After this point the culture medium was no longer changed, allowing the virus rel eased from the lytic plaques to accumulate and infect neighboring cells, generating widespread cytopathic effects in each culture. The cells and medium were then collected, and followi ng successive rounds of freeze-thaw, the lysates were used to infect new cultures of Cre8 cells. Since the packaging signal of the 5 adenoviral backbone is removed by the Cre recombinase, the Cre8 cell line is nonpermissive for its replication. Therefore, each new vector construct was passaged at least three times in Cre8 cells to eliminate any contaminating 5 virus that may have been propagated during growth and amplification of the new vector. To determine the purity of each new adenov irus preparation, the genomic DNA was isolated, digested with appropriate restriction enzymes and analyzed following electrophor esis in 0.7% agarose gels. Once the vectors were found to be free of detectable 5, subsequent amplifications were performed in 293 cells. In an effort to eliminate as much contam inating cellular DNA and debris as reasonably possible, adenoviral preparations were banded over th ree successive CsCl gradients and were then dialyzed against multiple changes of dialysi s buffer prior to aliquotting and storage. 72

PAGE 73

Isolation of MSCs We elected to use the bone marrow of bovine calves as a source of MSCs, as it would provide several advantages for our studi es. Foremost is that the long bones of these animals, at birth, are at least 5 to 10 times greater in size and volume than the corresponding bones in the adult rabbit, whic h is among the larges t of the common experimental animals. This large size enabled the isolation of an abundance of primary MSCs (usually > 5 x 107) from a single bone (generally t he head of the femur), requiring minimal expansion of the cells in culture prior to their use in experiments. Further, since male calves are routinely sold from the dairy farm soon after birth, we could make use of these animals and in the process obtain cells from a newborn animal. In this regard, the cells would be expected to have maximal proliferative and differ entiation potential, and thereby provide a robust readout of chondrogenic stimulation in our in vitro assays. The use of these animals was facilitated by the Large Animal Sciences Department whose facilities on the University of Florida campus are in close proximity to the Health Sciences Center. Following procurement immediat ely after the death of the animal, the femurs were transected with a band saw. The red marrow was removed from the epiphyseal ends of the femurs and tibiae, and both red and yello w marrow was scooped fr om the medullary cavity with a spatula. The semi-solid marro w was placed in MSC medium (as described in Chapter 2) and passed through a syringe several times to disaggregate the loose stroma. Following digestion of the mixture with collagenases types I and II and neutral protease (Worthington Biochemical Corporation) for 30 minutes at 37C, the cells were passed through a 40 m nylon cell strainer (BD Falcon) and plated. After the first preparation, we found that t he enzymatic digestion step was not necessary; the cells 73

PAGE 74

would adhere to the culture flasks as long as the stroma was sufficiently homogenized by hand. Although somewhat variable between preparations, typically enough of the disaggregated marrow was obtained to seed 5 to 6 175 cm2 flasks, which would be ~3040% confluent with adherent fibroblastic cells at 24 to 48 hrs. Cells were allowed to adhere for 72 hours prior to the first change in media, which removed all of the nonadherent blood cells. Following ex pansion to confluence, the cells were either used immediately for experimentation or were aliquotted for storage in liquid nitrogen. Gene-Mediated Chondrogenesis of MSCs Our studies of gene-induced chondrogenesis were performed using high density aggregate culture systems. In early experiment s we followed the methods of Palmer et al. whereby approximately 2.0 x 105 MSCs suspended in culture medium were aliquotted into 15 mL conica l vials (Corning). During centrifugation at low speed, the cells were forced into aggregates, forming pelle ts in the bottom of each tube. Each cell pellet was cultured in 750 l of chondrogenic medium supplemented with dexamethasone in its individual 15 mL t ube for the following 21 days. Media were changed every 48 hours, with additional changes to allow for 24-hour conditioned media collection at days 3, 7, 14, and 21. While th is method proved useful for a small number of samples, it proved unmanagea ble for the simultaneous cult ure of multiple treatment groups, each with multiple replicates. The frequent handling of la rge numbers of tubes and their individual screw caps was unwield y, leaving the cultures prone to fungal contamination. Supplementation of the culture medium with antimycotic agents was of no benefit as these appeared to interfer e with cell-cell adhesi on and inhibited chondrogenic differentiation. 74

PAGE 75

To provide a method amenable to the analysi s of large numbers of samples, we worked to develop a high-throughput in vitro chondrogenesis system whereby numerous MSC aggregates could be formed and cu ltured simultaneously in multi-well plates. In this technique, the MSCs were genetically modified in monolayer, detached using trypsin 24 hrs later, suspended in a mi nimal volume of cu lture medium and then counted. Approximately 2.0 x 105 cells were then delivered in a 300 l volume of chondrogenic medium (described in Chapter 2) to individual wells of a V-bottom 96-well plate. Once the plate was loaded with the samples of inte rest, using appropriate plate adaptors, it was spun at 4C in a table top centrifuge at 500 g for 5 minutes to form cell aggregates in the bottom of each well. The wells were topped off with an additional 100 l of chondrogenic media and grown at 37C. Ov er the course of 24 hours, the cells formed rounded aggregates that did not adhere to the polypr opylene plate. After 24 hours, aggregates were disturbed by pipetti ng the media vigorously. This provided a means to observe whether all cells were in corporated into each agg regate rather than disassociating upon movement. The medium for each sample was changed routinely, as before, and the cells were cultured for 21 days. Although the maintenance of the tiny pellets remai ned a somewhat challenging and tedious procedure, we found this approach to be far more manageable and provided more consistent results than the use of 15 mL tubes. We found, however, that certain parameters we re essential to its application. For example, useful aggregates were generated only with the use of Corni ng brand, polypropylene, V-bottom, 96 well plates. We tested similar plates from other manufacturers, such as Nunc, but were 75

PAGE 76

unable to consistently generate usable pellets following centrifugation since the V of their well was a slightly different angle. Adenoviral-Mediated Delivery of BM P-4, BMP-7, Ihh, and Sox9 Drives Chondrogenic Differentiation of MSCs Having developed suitable techniques for the isolation of low passage MSCs and their use in high-throughput in vitro chondrogenesis assays, we used these methods to evaluate the relative chondrogenic activity of BMP-4, B MP-7, Ihh, Sox9 and CTGF following adenoviral-medi ated gene delivery. Initially we focused our efforts on Ad.BMP-4 and Ad.BMP-7, since the chondrogenic potential of the recombinant proteins has been demonstrated in several reports. First or second passage bovi ne MSCs were seeded in monolayer and expanded to ~90% confluence. To ensure that the full potentia l of each cDNA was represented in our assays, we infected separate monolayer cultures with each adenoviral vector over a range of doses spanning 10 to 105 vp/cell. The genetically modified cells were harvested 24 hours later, and ~2.0 x 105 cells were seeded per well for each vector and dose. The cells were pelleted by centrifugation and cultured in defined medium contai ning dexamethasone for 21 days. The levels of the secreted transgene products were measured in conditi oned medium over the course of the experiment. At the end of t he incubation period, the cell pellets were harvested and processed for histology and immunohistoche mistry or were pooled and used for isolation of RNA for subsequent analysis by qRT-PCR. Chondrogenesis was evaluated by examination of aggregate morphology, histol ogical staining fo r proteoglycan content and immunohistochemical analysis for production of collagen types I, II, and X. 76

PAGE 77

For both Ad.BMP-4 and Ad.BMP-7, the aggregat es formed from MSCs infected at the 105 vp/cell dose were not viable and disintegr ated within 3 to 7 days of culture (data not shown). As shown in Figure 3-1, ELI SA measurements of the respective transgene products produced by the remaining cell pe llets between the two vectors reflected considerably different levels of prot ein production and response ranges. In general, peak production occurred near days 3 and 7, and gradually tapered over the 21 day incubation period. Transgenic expression was largely dose dependent for both Ad.BMP4 and Ad.BMP-7; however, the cell pellets transduced with Ad.BMP-7 showed a somewhat higher level of pr otein production relative to viral dose (Fig. 3-1B). In response to Ad.BMP-4, the cells receiving the 10, 100 and 1000 vp/cell doses produced peak expression levels of ~0.8, 2. 0 and 7.0 ng/mL, respectively (Fig. 3-1A). While the 1000 vp dose showed the greatest initial levels of expression, it also showed the most precipitous reduction over time to less than 2 ng/mL. Intere stingly, the pellets infected at the 10,000 vp/ce ll dos e produced somewhat less BMP-4 than the 1000 vp dose throughout, and expression was more vari able. BMP-4 was not detected in the conditioned medium of the control pellets at any time. Transgene expression from the Ad.BMP-7-i nfected pellets at day 3 ranged from less than 1 ng/mL for the 10 vp/cell dose to over 50 ng/mL for the 10,000 vp/cell dose (Fig. 3-1B). With the exception of the day 7 ti me point where the 1000 vp/cell dose showed a jump in expression from 32 ng/mL to nearly 70 ng/mL, expression levels showed a modest but gradual reduction in expression. Histologic examination indicated distinct ev idence of transgene-induced chondrogenesis of the MSCs stimulated wit h BMP-4 and BMP-7. Ho wever, we found 77

PAGE 78

the level of chondrogenesis to be reproducibly greater and more consistent when BMP4 was supplied as a transgene. As seen in Figur e 3-3, relative to controls, MSC pellets infected with either 10, 100 or 1000 vp/cell of Ad.BMP-4, expressing between 0.8 and 8.0 ng/mL of the trans gene product, were highly cellular and showed positive staining for toluidine blue and corresponding posit ive immunostaining for type II collagen, characteristic of articular cartilage ma trix. Pellets infected at 100 and 1000 vp/cell doses, expressing 2-8 ng/mL BMP-4 over the 21 days produced a more dense, uniform matrix populated with r ounded chondrocytic cells in lacunae, morphologically similar to the transitional zone of articular cartilage (Fig. 1-1B). Pellets infected at the 10,000 vp/cell dose were much smaller and appeared as loose cell aggregates with no evidence of extracellular matrix production or cellular differentiation. This is consistent with previous findings that ex cess viral load can have toxic effects, which would account for the reduced expression of the BMP-4 transgene and lack of a biological response in these pellets. For the pellets transduced with Ad .BMP-7, pronounced chondrogenesis was observed only in the pellets infected at t he 1000 vp/cell dose, which produced between 30 and 70 ng/mL of BM P-7. As shown in Figure 3-4, these pellets synthesized a dense, uniform extracellular matrix enriched fo r proteoglycans and collagen type II. Rounded chondrocytic cells were evident; however, many appeared to have begun to advance to hypertrophy, indicated by their increased cyt oplasmic volume. MSC pellets infected at the lower doses, producing less than 10 ng/mL of BMP-7, showed minimal metachromatic staining with toluidine blue and little evidence of cellular differentiation or cartilage ECM synthesis. The pellets infect ed at the 10,000 vp/cell dose were largely 78

PAGE 79

similar to those infected at the same dose with Ad.BMP-4, and despite continued production of BMP-7 of betw een 30-50 ng/mL over the 21 da ys, showed no evidence of chondrogenesis. Following these experiments, we wanted to adopt a similar strategy to evaluate the chondrogenic effects of adenoviral mediated gene transfer of Ihh and Sox9 to MSCs. Unfortunately, since Sox9 is an intracellular transcription factor and commercial ELISAs were not available for Ihh, it would not be possible quantify transgene products synthesized by the pellets over time. Ther efore, to verify effective gene delivery to MSCs and gain insi ght into the relations hip between viral dose and the level of transgenic expression, we performed Western blot analyses of MSC cultures infected with Ad.Sox9 or Ad.lhh at 3-fold dose increments between 10 and 10,000 vp/cell. Cultures of uninfected MSCs were proce ssed in parallel and used as negative controls. Based on the negative results obtained above wit h extremely high vector doses, we set 10,000 vp/cell as the upper limit for infection. As seen in Figure 3-2, detec table human Sox9 protein was found in lysates of cells infected with as few as 30 vp/cell of Ad .Sox9, with peak expr ession associated with doses of 100 and 300 vp/cell. Analogous to the expres sion data for Ad.BMP-4, a marked reduction in Sox9 production was seen at 1000 vp/cell and higher doses, such that it was below the limit of detection in cells infected with 3000 and 10,000 vp/cell. Western blots of media c onditioned by MSCs infected with Ad.Ihh also showed dose dependent expression, but with a slightly different profile. Protein bands were faintly visible in the 300 vp/cell dose lane incr easing in intensity with viral dose to a maximum at 3000 vp/cell (Fig. 3-2). At the 10,000 vp/cell dose, however, there was no 79

PAGE 80

visible protein band. The dramatic reductions in transgene expression seen with high vector doses were again consistent with toxicity from excessive viral load. Having established a working relationship between viral dose and transgene expression for Ad.Sox9 and Ad.Ihh, we infe cted additional cultures of MSCs with each vector over a similar dose range. As de scribed above, we then seeded the genetically modified cells into aggregate culture and assayed for chondrogenic induction 21 days later. Histologic analysis of the recovered cell pellets showed that both Ihh and Sox9 were capable of driving chondrogenesis when expressed as transgenes; however significant differences were noted in their stimulatory capacities. F or MSCs infected with Ad.Ihh, pellets formed from cells at t he 300 and 1000 vp/cell doses demonstrated the most robust chondrogenesis (Fig. 3-5). The majo rity of the cells in each had a rounded, mature, chondrocytic phenotype and produced large quantities of ECM enriched for proteoglycans and collagen type II, with low collagen type I content. Although chondrogenesis was readily disc ernable in pellets at the 3000 vp/cell dose, they were noticeably smaller, with less abundant matr ix and reduced cellularity. Aggregates formed from MSCs at the 10,000 vp/ce ll dose never increased in size, became fragmented within 1 to 2 weeks and disassoci ated by day 21. Aggregates formed from MSCs infected at doses below 100 vp/cell gradually diminished in size and at the end the incubation period were t oo small to paraffin embed. Aggregates formed from MSCs infected with Ad.Sox9 showed the greatest response to the transgene at the 100 vp/cell to 1000 vp/cell range (Fig. 3-6). The volume and cellularity of t he aggregates was considerably gr eater than those infected at 80

PAGE 81

lower doses and the uninfected controls. The pe llets showed robust extracellular matrix production with increased proteoglycans (ind icated by metachromatic toluidine blue staining) and increased collagen type II content, consistent with that of articular cartilage. Distinct from t he aggregates expressing the ot her growth factors, the morphology of the ce lls appeared considerably less hy pertrophic. The cells (and their surrounding lacunae) were noticeably sm aller in diameter, and contained less cytoplasmic volume. While aggregates from the 10,000 vp/cell dose also showed pronounced synthesis of cartilaginous matrix, these pellets were smaller in diameter. During endochondral bone growth, the synt hesis of collagen type X by growth plate chondrocytes, in conjunction with increased cellular volu me, is indicati ve of their maturation to a terminal hypertrophic state. From this, the cells undergo apoptosis, and the residual cartilage matrix is replaced by bone. The morphology of the chondrogenic cells in the pellets expressing BMP-4, BMP-7 and Ihh suggests that prolonged stimulation with these factors has driven the differentiation of the MSCs toward an early hypertrophic phenotype, more so than overexpression of Sox9. To further examine this, we used immunohistochemistry to stain fo r the presence of collagen type X in chondrogenic sections from Ad.BMP-4 and Ad.Sox9 infected pellets. As shown in Figure 3-7, strong collagen type X staining is seen throughout the outer regions of the Ad.BMP-4 pellet, while the Ad.Sox9 pellet s show little positive staining over background. Adenoviral Delivery of CTGF Stimul ates Proliferation of MSC Aggregates In an effort to identify additional cDNAs with chondrogenic potential that might be of use in cartilage repair, we also tested t he effect of adenoviral mediated expression of connective tissue growth factor (CTGF) in MSCs in aggregate cult ure. Similar to TGF81

PAGE 82

1, CTGF stimulates fibroblas t proliferation, differentia tion and extracellular matrix synthesis, and is important for chondr ocyte differentiation and maturation131. As above, early passage of MSC cultures were infected with a range of doses of Ad.CTGF and placed into aggregate culture fo r 3 weeks. As with the other transgenes, the aggregates infected with certai n vector doses visibly increased in size relative to controls, during incubation. Histologic anal ysis showed that the increase in pellet volume could be attributed to intense cellu lar proliferation, which was particularly evident at the 1000 vp/cell dose (Fig. 3-8). Ho wever, despite a clear biological response to the transgene product, there was no apparent chondrogenic differentiation or cartilage matrix protein production. At higher levels of Ad.CTGF infection, such as those at 10,000 vp/cell and higher, the cells did not proliferate; the aggregates remained small, and they fragmented a fter 21 days of culture. Combinations of Sox9 with BMPs Induced Chondrogenesis The process of chondrogenesis in vivo is complex, requiring the orchestrated expression and interplay of numerous grow th and transcription factors. Using gene delivery we have shown that over expression of certain of these proteins individually is sufficient to initiate chondrogenic diffe rentiation of mesenchymal progenitors in vitro The resulting phenotype of the cells and th eir synthesized matrix demonstrate key hallmarks of articular cartilage, but fall short of the architecture of the native tissue. Indeed it is likely that successful regenerat ion of articular cartilage will require the delivery of multiple factors and temporal regulation of thei r expression. Toward this direction we initiated preliminary studies of the effects of gene combinations on in vitro chondrogenesis. For these experiments, we co-expressed the most potent of the chondroinductive genes we have tested thus far, BMP-4 and BMP-2 (characterized in 82

PAGE 83

earlier studies in our group) with the chondrogenic transcr iption factor Sox9. We hypothesized that the enhanced expression of Sox9 would help to stabilize the chondrocyte phenotype and block progression to hypertrophy, significantly improving the quality of the resulting cartilaginous tissue. In exploring gene combinations, we infe cted MSCs in monolayer with pairs of vectors at doses that i ndividually produced a robust ch ondrogenic response: i.e. Ad.BMP-4 at 100 and 1000 vp/cell; and Ad.BMP -2 at 1000 vp/cell were mixed with Ad.Sox9 at 100 or 1000 vp/cell. We then seeded the genetically modified cells into aggregate culture and analyzed the effects of co-expression of the respective transgenes on chondrogenesis in our in vitro assay. During the incubation period, the majority of the cell aggregates showed a steady increase in diameter, and none of the gene combinations caused the pellets to reproducibly disintegrate. Histological anal ysis showed enhanced production of cartilage matrix components in pellets from each of the various treatment groups, relative to controls, indicating that the transgene produc ts were functionally expressed. However, none of the gene combinations appeared to enhance chondrogenesis in vitro relative to single gene delivery. While the results of these experiments were somewhat negative, they serve to emphasize the complexity of cell signaling pathways and ce llular differentiation. The adenoviral vector, with its CMVpromoter driven expressi on cassette, is designed to provide extraordinarily high levels of transgenic expre ssion, and likely serves to continually saturate ligand-specific su rface receptors and downstream signaling pathways. Effectively modulating the effects of such a powerful stimulus may prove to 83

PAGE 84

be challenging, requiring far more s ophisticated gene deliver y approaches than the crude systems used here. Wistar Rat Responses to Adenoviral Transgene Delivery When designing a gene-based strategy for ca rtilage repair, the selection of an appropriate transgene is fundamental to the overall success of the procedureits efficacy and its safety. For most applications, it is probable that gene delivery vectors, their secreted transgene products, or genetically modified cells will emigrate from the repair site to adjacent tissues. Because many of the agents c onsidered to have the greatest chondrogenic activity ar e pleiotropic, with broad stimul atory activities in various tissues and cell types, it is critical to thoroughly understand the possible impact of their constitutive, local overexpression on the join t tissues and the health of the prospective patient. Using our in vitro assay we have identified several proteins (BMP-4, BMP-7, Ihh, Sox9) with the capacity to drive mesenc hymal chondrogenesis when expres sed in MSCs as transgenes. In earlier related studies we found that the cDNA for BMP-2 was also highly effective. Having established the relative chondrogenic activity of these genes in vitro we wanted to determine their potential to stimulate pathologic side effects when over-expressed intra-articularly. For these in vivo studies, we selected the knee of the rat as our animal model since the ra t is the smallest co mmon experimental animal with a defined joint space that can be reliably targeted by intra-articular injection. We delivered the adenoviral vectors containing BMP-2, BMP-4, BMP-7, Sox9 and Ihh bilaterally at doses of 5 x 108 and 2 x 109 vp/l to the knees of Wi star rats (2 rats per vector and dose). Parallel groups of rats were similarly injected with Ad.GFP to control for the effects of adenoviral delivery. At 7 days post injection the animals were killed, 84

PAGE 85

and the knees were harvested, de calcified and processed for histologic analysis. (The vector doses used in these experiments we re based on our previous experience with adenoviral gene transfer to joints. They were intended to provide sufficient transgene expression to enable a robust depiction of the stimulatory properties of each transgene product intra-articularly without endangering the overall health of the animals.) H&E stained sections of the knees receiving Ad.GFP showed a mild leukocytic infiltration and a slight thickening of the synov ial intima, typical of an adenoviral injection into the joint. Otherwise, the morphology of the tissues was normal. In stark contrast, adenoviral deliv ery of the various chondrogenic cDNAs elicited a wide variety of biological responses. The most dramatic ef fects occurred in the joints injected with Ad.BMP-2. The representative field shown in Figure 3-10 shows massive hypertrophy of the synovium and capsular tissues that comp letely displaced the adipose layer that normally supports the synovial lining. The expanded tissue was fibrotic in many areas and populated with spindled fibrob lasts. About half of the tissue mass, (particularly those areas proximal to articular cartil age), was strikingly chondrogenic and was heavily populated with chondrocytes and chondroblastic ce lls. The joints receiving Ad.BMP-4 and Ad.BMP-7 showed a dramatic hypertrop hy of the synovial lining, caused by extensive chondrometaplasia throughout the subsynovium. There was little if any fibrotic component to the expanded tissues; the increased volume was comprised almost exclusively of chondroblastic cells. Distinct from overexpression of the BM Ps, adenoviral delivery of Ihh induced a mild to moderate fibrosis broadly across the synovial lining, marked by an increase in both collagen fibers and fibroblastic cells. Ad.Sox9 was associated with a moderate 85

PAGE 86

synovitis throughout the lining, with incr eased numbers of synovial fibroblasts and infiltrating leukocyes relative to the Ad.GFP control. Discussion Our experiments demonstrate that adenoviral-mediated delivery of chondrogenic growth factors to MSCs serves as an effective method to induce chondrogenic differentiation of these cells in aggregate cult ure. In comparing mult iple growth factor cDNAs delivered across a range of adenoviral doses, we found that delivery of ~100 to 1000 vp/cell provided the most effective leve ls of transgene expres sion. Of the factors we evaluated, BMP-4 induced the mo st robust chondrogenesis, with aggregates exhibiting a dense extracellu lar matrix populated with chondrocytic cells resembling those found in deep articular cartilage BMP-7 and Indian hedgehog exhibited less desirable chondrogenic responses; while aggregates showed extracellular matrix production and the presence of c hondrocytic cells, these secreted factors were prone to drive cells toward terminal differentiati on and hypertrophy. Sox9, while less prone to induce hypertrophic differentiation, induc ed production of cartilaginous matrix and maintained the cells at a prehypertrophi c state, as shown with collagen type II expression and the absence of co llagen type X. CTGF induced proliferation of the cells at the 1000 vp/cell dose but failed to promote chondrogenesis in our system. Members of our lab have shown previous ly that gene delivery of BMP-2 and its expression in the 10-100 ng/mL range is requ ired to achieve optim al chondrogenic differentiation of MSCs. Here, using a si milar system, we have shown that BMP-4 production of <10 ng/mL is sufficient to promote robust chondrogeni c differentiation of the cells. These data indicate that BMP-4 is highly potent, and it induces chondrogenic responses equal to those observed at highe r ng/mL levels of BMP-2 production. 86

PAGE 87

Overall, in pellets modified to express se creted growth factors, we observed more substantial chondrocytic differ entiation than in those expressing the intracellular factor Sox9. With secreted factors, even though a proportion of MSCs within the pellet express the transgene, they have both autocrine and paracrine effects, which may enhance chondrogenesis throughout neighboring cells within the aggregate. Conversely, transgene products that function intracellularly may have advantages for use in studies in vivo because the risk of exposing cells in nei ghboring tissues such as the synovium, meniscus, tendon, and muscle to pleiotropic growth factors will be significantly reduced. Chondrogenesis in vivo is a complex process involving the coordinate interplay of numerous factors. As such, it is probably unrealistic to expect complete articular cartilage regeneration via transfer and expr ession of a single gene. Based upon preliminary work demonstrating enhanced chondrogenesis when IGF-1 was delivered to MSCs in combination with TGF-1 or BMPs, we opted to similarly test our battery of transgenes in combinations. As all potential combinations of chondrogenic cDNAs and doses are too numerous to test feasibly we used the information gathered about each single transgene to deduce the most effective st rategy. We hypothesiz ed that delivery of Sox9, which is expressed intracellularly, co mbined with the secret ed proteins, BMP-2 or BMP-4, would provide an effective combinat ion to drive chondrogeni c differentiation in MSCs while maintaining cells in a pre-hyper trophic state. Although the combinations we evaluated did not significantly alter chondrogenesis, they illustrate the complexity of modulating cell signaling. Although each transgene tested induced measurable chondrogenic effects on bovine MSCs in aggregate culture, t he microenvironment of the joint space in vivo is 87

PAGE 88

quite different. Intra-articula r delivery of Ad.BMP-2, -4, -7 Ad.Ihh, and Ad.Sox9 to the intact knees of Wistar rats enabled us to evaluate potential side effects of ectopic growth factor expression. BMP-2, -4 and -7 prompted cellular expansion and ectopic cartilage formation to varying degrees whil e Ihh caused a weak fibrosis across the synovial lining and Sox9 induc ed mild inflammation. Although these rat knees likely demonstrate the worst case scenario of potential side effects of gene enhanced cartilage repair, they provide a vivid repr esentation of the potency of the various transgene products and their capacity to impact the biology of the articular connective tissues. Conversely, they likewise demonstrat e the exquisite sensitivity of synovial fibroblasts to proliferate and differentiate following stimulation with certain BMPs. Methods designed to induce cellular di fferentiation in cartilage and bone repair in vivo should take into account the high capacity for toxic side effects in adjacent tissues. Although adenoviral vectors have a reput ation for causing adverse immune responses in vivo this vector system may ha ve a futu re in ex vivo approaches to tissue repair. With their ability to readily infect MSCs and induce a high level of gene expression, adenoviral vectors can serve as effi cient protein delivery vehicles to modify cells that are introduced to sites of tissue damage. Such a method will bypass the potential for host immune responses whil e allowing for potent expression of chondrogenic cDNAs. 88

PAGE 89

A). 0.0 2.0 4.0 6.0 8.0 10.0 Day 3 Day 7Day 14Day 21 Time (d)BMP-4 (ng/ml) 10 vp/cell 100 vp/cell 1000 vp/cell 10,000 vp/cell control B). 0 10 20 30 40 50 60 70 80 90 Day 3 Day 7 Day 14 Day 21 Time (d)BMP-7 (ng/ml) 10 vp/cell 100 vp/cell 1000 vp/cell 10,000 vp/cell control Figure 3-1. ELISAs of conditioned media from MSC aggr egates transduced with BMP-4 and BMP-7 indicated protein production in response to viral doses administered. Triplicate samples of conditioned media were obtained at days 3, 7, 14, and 21 following a 24 hour in cubation, and protein levels were quantified by ELISA. Data are shown as the mean S EM. A) BMP-4 ELISA indicates a dose-dependent response following Ad.BMP-4 transduction. The decrease in protein production follow ing the 10,000 vp/cell dose indicates toxicity in response to excessive adenovirus. B) BMP-7 ELISA shows protein production in response to Ad.BMP-7 tr ansduction at higher concentrations than those observed with BMP-4. 89

PAGE 90

Figure 3-2. Western blotting confirm ed transgene expression fo llowing adenoviral gene delivery to MSCs. As shown in the top panel, Ad.Ihh was delivered at a range of doses to bovine MSCs in monolayer, in a minimal volume of media. Media were replaced 24 hours later, and har vested at 48 hours post-infection for detection of Ihh protein. The presence of bands at 300, 1000, and 3000 vp/cell doses correlated with the ma trix synthesis and chondrogenic differentiation observed with toluidine bl ue staining of Ihh aggregates at these doses (Fig. 3-5). Following detection of Ihh, antibodies were stripped from the membrane, and -actin was detected as a loading control. In the bottom panel, Ad.Sox9 was delivered at a range of doses, and 48 hours posttransduction, cells were harvested with a cell scraper and lysed with chilled homogenization buffer. Because the transcription factor Sox9 remains intracellular, cell lysates were used to detect Sox9 production. As with Ihh, the range in which Sox9 bands were visible correlated to the dose range in which proteoglycan synthesis was most robust as shown through toluidine blue staining (Fig. 3-6). Again, -actin was used as a loading control. 90

PAGE 91

Figure 3-3. MSC aggregates expressing BMP-4 were evaluated for chondrogenesis through toluidine blue staining of proteoglycans and immunohistochemistry for collagen types I and II. Aggregates infected with 10, 100, or 1000 vp/cell were highly cellular, the proteoglycans of their synthesized matrix stained deep purple with toluidine blue, and they showed positive immunostaining for collagen type II, as shown by green fluorescence. The 10,000 vp/cell dose resulted in fragmented pellets with little to no cellular differentiation or matrix production. All aggregates were negative for collagen type I production. 91

PAGE 92

Figure 3-4. MSC aggregates infected with Ad .BMP-7 over a range of doses yielded pronounced chondrogenesis at the 1000 vp/cell dose only. Doses of 10 and 100 vp/cell showed minimal toluidine blue staining and poor collagen type II expression. Aggregates receiving 1000 vp/cell synthesized dense, uniform extracellular matrix enriched for proteoglycans and collagen type II. Rounded, chondrocytic cells were present in t he matrix, and many appeared to advance toward hypertrophy, with increased cytoplasmic volume as shown above. Pellets receiving 10,000 vp/cell show ed no evidence of matrix protein production or chondrogenic differentiation. The high dose of adenoviral vector proved toxic. All aggregates were negat ive for collagen type I production. 92

PAGE 93

Figure 3-5. Indian hedgehog aggregates expr essed proteoglycan and collagen type II following adenoviral gene delivery. Indian hedgehog (Ihh) expression in pellets after 21 days in culture revealed a dose-dependent response. Aggregates in the range of 300 to 1000 vp/cell demonstrated the most robust chondrogenesis by forming large, rounded aggregates that secreted proteoglycans, forming a cartilage matrix that contained chondrocytic cells. MSCs receiving doses of 100 vp/cell or less or 10,000 vp/cell or greater formed fragmented aggregates that did not proliferate, were smaller than the non-transduced controls (shown in Figur e 3-3), and disassociated by the 21day timepoint. All aggregates were negative for collagen type I. 93

PAGE 94

Figure 3-6. Sox9 induc ed cellular proliferation and matrix synthesis with less differentiation than other transgenes te sted. Sox9 expression in MSCs resulted in aggregates of a uniform round shape consisting of proteoglycan matrix and prehypertrophic-like cells. Sim ilar to the other transgenes tested, the most chondrogenic effects were observed after delivery of 100 to 1000 vp/cell. Doses above and below this range continued to promote collagen type II expression, yet they were in sufficient to promote robust cellular proliferation. All aggre gates were negative for collagen type I production. 94

PAGE 95

Figure 3-7. Collagen type X staining of Sox9 and BMP4 aggregates. BMP-4 aggregates express collagen type X and undergo differentiation toward a terminal hypertrophic state, shown thr ough green fluorescence. Sox9 pellets show no signs of collagen type X, with staining equivalent to background levels, indicating their maintenance at a prehypertrophic state of differentiation. 20x magnification. 95

PAGE 96

Figure 3-8. CTGF expression in MSC aggregates promotes cellular proliferation but no chondrogenic differentiation. Delivery of 100 0 vp/cell of Ad.CTGF resulted in cellular proliferation and large aggregates; however, to luidine blue staining showed no evidenc e of proteoglycan matrix producti on or chondrogenic differentiati on. Likewise, aggregates were negative for collagen type II production (lack of green fluorescence) when evaluated immunohistochemically. Cells rece iving viral doses of 10 and 100 vp/cell formed aggregates that were too small to paraffin embed. Viral doses of 10,000 vp/cell or greater caused aggregates to fragment and disassociate in culture. 10x magnification. 96

PAGE 97

Figure 3-9. Sox9 in comb ination with BMP-4 and BMP-2 in MSC aggregate culture had no additive effects upon chondrogenesis. Panels A, B, C, and G are single doses, D-F, H-I are combinations. Tolu idine blue staini ng shows matrix synthesis at all doses; however, combinations resulted in smaller aggregates that showed less matri x synthesis and le ss chondrogenic differentiation. A) Ad.BMP-4 100 vp/cell, B) Ad.BMP-4 1000 vp/cell, C) Ad.Sox9 1000 vp/cell, D) Ad.Sox9 1000 vp/cell + Ad.BMP-4 1000 vp/cell, E) Ad.Sox9 100 vp/cell + Ad.BMP-4 1000 vp/cell, F) Ad.Sox9 100 vp /cell + Ad.BMP-4 100 vp/cell, G) Ad.BMP-2 1000 vp/cell, H) Ad.Sox9 1000 vp/cell + Ad.BMP-2 1000 vp/cell, I) Ad.Sox9 100 vp/cell + Ad.BMP-2 1000 vp /cell. Images are displayed at 10x magnification. 97

PAGE 98

Figure 3-10. Wistar rat knees exhibited varying responses to adenoviral delivery of pleiotropic transgenes. Adenoviral ve ctors containing GFP and the chondrogenic cDNAs BMP-2, BMP-4, BMP-7, Ihh, and Sox9 were injected intra-articularly into both knee joints of healthy male Wistar rats at doses of 5 x 108 and 2 x109 vp/l. Effects from BMPs ranged from the massive synovial hypertrophy, fibrosis, and ectopic chondroge nesis resulting from Ad.BMP-2 to moderate synovial hypertrophy and chondrom etaplasia with little fibrosis from Ad.BMP-4 and Ad.BMP-7. Indian hedgehog (Ihh) induced synovial fibrosis and Sox9 was associated wit h a mild synovitis, comparable to that observed in the Ad.GFP control animals. 98

PAGE 99

CHAPTER 4 COMPARISON OF THE EFFECTS OF ADENOVIRAL, LENTIVIRAL, AND AAV TRANSGENE DELIVERY TO MESENCHYMAL STEM CELLS Introduction Our studies have focused on the genetic modifi cation of MSCs for use in the repair and regeneration of articular cartilage. We have found that adenoviral-based ve ctors efficiently transduce these cells, and r eadily enable the expr ession of transgene products at biologically relevant le vels. Indeed, adenoviral-mediated delivery and expression of certain growth fa ctor cDNAs in MSCs can be used to effectively direct the differentiation of these cells along chondrogeni c pathways. While highly useful tools for studies in vitro and in experimental animals, the mo st widely used generations of this vector still contain the majority of the native adenoviral genes. Low-level expression of these genes in transduced cells is frequently associated with infla mmatory reactions in vivo and elimination of transduced cells by cytox ic T cells in immunologically competent hosts. The relative ease of use of the adenovir al vector system, coupled with its highly efficient gene transfer, provided us with a val uable tool to screen a battery of candidate cDNAs for their chondrogenic potential. From th is screen we have identified several that merit evaluation in vivo in cartilage repair models. While the adenovirus has been the workhorse of our in vitro studies, its immunogenic profile excludes it from use in humans and may negatively impact the proc ess of tissue repair. In an effort to identify a gene delivery system more suitable for use in anima l models and possible clinical application, we evaluated the utility of t he well-characterized vector systems currently available, (lentivirus, adeno-associated virus and non-viral trans fection) for their ca pacity to effect gene-mediated chondrogenes is of MSCs. 99

PAGE 100

For these studies, we first inserted the cDNA for BMP-4 into AAV, lentivirus and plasmid-based vectors. We then determined the capacity of each vector to deliver the BMP-4 transgene to bovine MSCs and stimulate chondrogenesis in the aggregate culture system. The BMP-4 cDNA was selected for these studies as it appeared to have the greatest chondrogenic activity in adenov irally-transduced aggregates. Thus, the ability to functionally deliver this cDNA would set the minimum standard for efficacy. Results Plasmid DNA Transfection of MSCs Results in Transient Expression To generate the plasmids needed to produce recombinant viruses, we inserted the cDNA for human BMP-4 into specific CMV-promoter driven vect or constructs: AAV (pHPA-trs-sk-BMP-4), adenovirus (pAdlox-BMP-4), lentivir us (pCDH-puro-BMP-4). We then used these plasm ids in DNA transfect ions to gauge expression resulting from nonviral gene transfer. Plasmid DNA transfections of BMP-4 cDNA were carried out in 24-well plates of bovine MSCs grown to approximately 80% c onfluence. DNA was diluted in Opti-MEM (Gibco) and mixed with li pofectamine (Invitrog en) at a ratio of 1g of DNA to 2.5 l of lipofectamine. DNA: lipofectamine complexes were added drop-wise to cells and allowed to incubate for 24 hours. At 48 and 72 hours post-transfection, conditioned media were colle cted for quantification of BMP4 by ELISA. As shown in Figure 4-1, transfections with each plasmid construct yielded BMP-4 in the 1.0 ng/mL range, which was significantly higher t han background levels from control cells. However, transgene expression was transient. ELISAs of condition ed media at 72 hours following transfection showed concentration s of BMP-4 near background levels (data not shown). Such low levels of short-lived expression were not sufficient to drive chondrogenesis in our aggregate culture assay. 100

PAGE 101

Lentiviral Transduction of MSCs Proves Challenging Preliminary data showing r obust GFP expression in 293 cells following lentiviral transduction established a basis for the use of lentivirus to deliver chondrogenic transgenes, such as BM P-4, to bovine MSCs in monolayer (Fig. 4-2). To produce recombinant lentivirus, we implemented a f our plasmid transfection procedure adapted from the ViraPower expressi on system (Invitrogen). Transducin g vectors expressing the desired transgenes (GFP and BMP-4) were generat ed via the insertion of the specific cDNAs into the pLenti4/V5-DEST vector via homologous recombination. The resulting expression plasmid was mixed with the th ree necessary packaging plasmids, denoted pLP1, pLP2, and pLP-VSVG. Plasmid DNA wa s complexed with lipofectamine and delivered to monolayer cultures of 293FT ce lls. Despite numerous attempts, we were unable to generate sufficiently high viral titers with the ViraPower system. In an effort to improve viral production, we then evaluated the plasmid, pCDH-puro, for its potential to generate appreciable titers of lentivirus. Lentivirus generated usi ng the pCDH-puro-BMP-4 c onstruct induced BMP-4 expression in monolay er MSCs (Fig. 4-3), but again the viral titers generated were insufficient to induce chondrogenesis within our pellet culture system. In an effort to enhance this, we selected for cells that expre ssed BMP-4 using puromycin. We cultured the cells under selection in preparation for subsequent expansion and use in aggregate culture; however, the comb ination of BMP-4 expressi on and puromycin in the cell culture medium appeared to i nhibit MSC growth and prolifer ation, rendering the cells unsuitable for aggregate culture. Despite r epeated attempts with each system, we were unable to produce reasonable titers for use dire ctly without the need for cell selection. 101

PAGE 102

Adenoviral-Mediated Delivery of BMP4 to MSCs Induces Chondrogenesis As discussed extensively in Chapter 3, bovine MSCs were transduced in monolayer with an E1, E3-deleted serotype 5 adenoviral vector ca rrying human BMP-4 cDNA and grown in aggr egate culture in our in vitro chondrogenesis assay. Histologic evaluation of aggregates after 21 days in cult ure showed that Ad.BMP-4 effectively induced chondrogenesis in MSC aggregates as indicated by stai ning for matrix proteoglycans and type II collagen. Aggregates infected with 100 to 1000 vp/cell were large, round, and produced a dense, uniform matrix populated with rounded chondrocytic cells in lacunae, morphologically similar to the trans itional or deep zones of articular cartilage (Fig. 3-3). Self-Complementary AAV-Mediated Delivery of BMP-4 to MSCs is Comparable to Adenovirus Numerous AAV capsid serotypes are available for cross-packaging of vector genomes, and each one changes the tr opism of the virus. In an effort to identify the most effective AAV serotype to target bovine MSCs, we packaged scAAV vector containing GFP into serotypes 1, 2, 5, and 8. Following administration of 104 vg/cell to MSCs in monolayer, transduction effici ency was determined visually by GFP expression. As shown in Figure 4-4, after 24 hours, GFP expression in transduced cells was evident for serotypes 2 and 5, and the in tensity increased over the course of 7 days. By day 7, all 4 serotypes exhibited det ectable GFP expression, but type 2 was by far the most robust, followed by type 5 (Fig. 4-4). The order of efficacy for AAV transduction of MSCs was AAV2>5>1>8. Based upon these data, we packaged the scAAV vector plasmids containing BMP4 (pHPA-trs-sk-BMP-4) to generate AAV serotypes 2 and 5 for use in our in vitro 102

PAGE 103

chondrogenesis assay. We infected MSCs over a 3-fold range of doses from 102-104 vg/cell. After 21 days in culture, the pellets were evaluated histolog ically with toluidine blue staining for proteoglycans. As shown in Figure 4-5, consistent with data from GFP expression, infection with scAAV2-BMP-4 induced the most complete chondrogenic response observed to date. The 3000 vg/cell dos e yielded aggregates that were large, round, highly cellular, and strikingly unifo rm in appearance. Aggregates showed robust proteoglycan synthesis with chondrocytic cells uniformly distributed throughout. Aggregates receiving the 1000 vg/cell dose exhibited similarly intense matrix production, yet the periphery of these pellets was differentiated. Conversely, scAAV5BMP-4 delivered across the same range of doses produced aggr egates with little apparent chondrogenic differentiation but showed evidence of enhanced proteoglycans synthesis. Discussion Using our pellet culture system, we compar ed the ability of recombinant lentivir us, AAV, adenovirus and plasmid transfection to deliver and functionally express the BMP-4 cDNA in bovine MSCs, as indicated by chondrogenic differentiation in vitro The purpose of these studies was to identify fr om among the most developed gene transfer systems, the one most suitable for use in vivo in studies of cartilage repair. As shown in Table 4-1, each vector system offers advantages for gene therapy with certain limitations for human applications. In our hands, plasmid transfection and lent iviral transduction were unsatisfactory for our in vitro chondrogenesis as say. While DNA transfection of MSCs provided BMP-4 expression near the minimal functional level (~1 ng/mL), it persisted fo r just over 48 hrs, which was insufficient to induce a meaningful response. With the lentivirus, technical 103

PAGE 104

difficulties prevented us from reproducibly g enerating vectors at us able titer. While we had intended to evaluate these systems bas ed on transgene expression profiles, in the case of lentivirus, vector production proved too cumbersome to be practical for our purposes. Recombinant AAV is emerging as the ve ctor of choice for human gene therapy applications because of its perceived safety, as the transduced cells do not express viral genes. In order to improve the efficien cy of AAV transduction in our investigations, we utiliz ed self-complementary AAV vectors. In a study completed by Kay et al., scAAV2 transduced synovial fibroblasts far mo re efficiently than conventional singlestranded AAV vectors, and expression levels of the hIL-1Ra transgene were similar to that noted previously for adenov irus and lentivirus vectors132. Self-complementary AAV m ediated gene delivery of BMP-4 induced a strikingly robust and uniform chondrogenic re sponse, generating cartilagino us pellets qualitatively superior to any other method. If these results hold, they indi cate that AAV is clearly the optimal vector system for this type of application. Our goal was to identify a system that could provide functional transgene expr ession but was less immunogenic than adenovirus. In most gene transfer applications relative to AAV, adenoviral vectors generate considerably higher levels of transgene expression. Based on this, we expected that AAV would provide only borderline functional levels of expression, forcing a trade-off between efficacy and low immuno genicity. That AAV may be capable of mediating more effective transgenic expre ssion and a more favorable immune profile than adenovirus represents a significant finding. 104

PAGE 105

At this point we do not know what may have contributed to the enhanced chondrogenic response provided by scAAV BMP-4. It is possibl e that it was fortuitous, caused by changes in culture conditions bey ond our awareness. Alternatively, it is possible that the differences in morphol ogy between pellets genetically modified with adenovirus and AAV can be attributed to the nat ure of their vector genomes. The adenoviral vector genome is about 35 kb in length and contains nearly all of the native viral genes. The scAAV vector is ~2.5 kb and is comprised only of an expression cassette flanked by small DNA hairpins. In our chondrogenic assays we infected MSCs in culture with a range of adenoviral vector par ticles. Pellets receiving the highest viral doses frequently showed toxic effects refl ected by reduced transgene expression or loss of viability. At lower doses thes e effects may be less pronounced, but still negatively influence the biology of the infe cted cells, resulting in pellets whose morphology is asymmetric. 105

PAGE 106

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 non-transfected control pHPA-trs-sk-GFPpHPA-trs-skBMP4 Adlox-BMP4pCDH1-BMP424-hr conditioned media, serumlessConcentration (ng/mL) BMP-4 Figure 4-1. ELISA of condi tioned media demonstrates expression of BMP-4 as measured 48 hours post-transfection. Bovine MSCs in monolayer were transfected with lipofectamine:cDNA complexes as directed by the manufacturer (Invitrogen). BMP-4 expr ession from MSCs transfected with AAV (pHPA-trs-sk), adenovirus (Adlox ), and lentivirus (pCDH1) plasmid constructs carrying human BMP-4 cDNA was significantly higher than nontransfected control cells; however, th is expression was transient. After 72 hours, BMP-4 expression matched that of negative controls (not shown). 106

PAGE 107

Figure 4-2. GFP expression in 293 ce lls 72 hours after infection with 15 l of LV-GFP in a minimal volume of serumless media. A) brightfield B) fluorescence, 20x magnification. This demonstrates our ab ility to generate lentivirus expressing GFP. This virus was then applied to bovine MSCs to gauge expression of lentivirus in primary cells prior to generating lentivirus c arrying BMP-4 cDNA. 0 1 2 3 4 5 6 7 8 Control MSCLV-GFP 48 hrLV-BMP-4 48 hrLV-BMP-4 72 hrAd.BMP-4 48 hrConditioned mediaConcentration (ng/mL) BMP-4 Figure 4-3. LV-BMP-4 expression was signif icantly less than Ad.BMP-4 expression in bovine MSCs. Bovine MSCs were grow n to 70% confluence in monolayer prior to infection with virus in a mini mal volume of media. 48 hours after infection, BMP-4 production was measur ed from 24-hour cond itioned media. While Ad.BMP-4 equivalent to 1000 vp /cell yielded 6 ng/mL of BMP-4, lentivirus yielded less than 0.5 ng/mL. Data are shown as mean S EM. 107

PAGE 108

Type 1 Type 2 Type 5 Type 8 Day 1 Day 3 Day 5 Day 7 Figure 4-4. AAV serotypes 1, 2, 5, and 8 were screened for transduction efficiency on low passage bovine MSCs in monolay er. Self-complementary AAV-CMV-GFP was delivered at a dose of 104 vg/cell, and transduction efficiency was evaluated visually with GFP. Types 2 and 5 were most effective, as shown by robust GFP expression beginning at day 1 and increasing in intensity over the course of 7 days. Serotypes 1 and 8 we re less effective at the same dose. We chose to use serotypes 2 and 5 for subsequent AAV constructs carrying BMP-4. 108

PAGE 109

109 Figure 4-5. MSCs transduced with scAAV-BMP-4 serotypes 2 and 5 demonstrate matrix protein synthesis in aggregate cu lture. Toluidine blue staining shows that scAAV2-BMP-4 promotes preh ypertrophic differentiation in MSC aggregates while scAAV5-BMP-4-aggr egates show limited chondrogenic differentiation but substant ial proteoglycan production.

PAGE 110

110 Table 4-1. Comparison of gene delivery vectors reveals pros and cons Vector Advantages Disadvantages Adenovirus Highly efficient gene transfer Readily infects MSCs Rapid, robust expression Long-term expression not required for chondrogenesis to occur Useful as a protein delivery system Expression tapers off after 21d Hyperplasia in synovium when virus is injected freelyneed to control delivery for ex vivo applications Lentivirus Infects MSCs Potential for sustained, long-term expressi on of growth factors Can select for cells that express protein products Difficult to achieve high viral titers Selection is required Selection conditions are inhibitory to cell expansion and subsequent use in chondrogenesis assa y scAAV Readily infects MSCs over a range of doses Effectively delivers chondrogenic transgenes Regarded as safe for c linical applications Potential for less efficient gene delivery than other viral vectors Pre-existing humoral imm unity to many sero types Plasmid DNA Ease of production Stability Safety Inefficient gene transfer Transient expression Cannot readily apply a dose range Comparison of gene delivery vectors reveals advantages and disadvantages of each system, with adenovirus and AAV offering the most potential for safe, efficient delivery and expression of chondrogenic cDNAs.

PAGE 111

CHAPTER 5 SUMMARY AND FUTURE DIRECTIONS The unique architecture of articular cart ilage limits its ability for self-repair, and since cartilage is frequently damaged by traum atic injury or disease, enhancing methods to repair damaged cartilage is of great clinical relevance. Tissue engineering techniques along with stem cell and gene-based therapies have the potential to improve cartilage repair and may eventually eliminate the need for invasive surgical procedures and total knee replacement. Tissue engineering techniques and stem cell therapies hold great promise for improvements to articular cartilage repair, especially in the knee joint. Current cellbased surgical treatments employ the use of healthy chondrocytes harvested from nonweight-bearing areas of cart ilage. Chondrocytes are availa ble in v ery limited quantities and require additional time for expansion in vitro prior to introduction to the cartilage lesion. Autologous mesenchymal stem ce lls, on the other hand, can be harvested in large quantities from bone marrow, fat, muscle and other tissues, and they can be virally transduced to express chondrogenic transgenes. Modified MSCs introduced to sites of cartilage damage may prompt th e generation of functional hy aline-like cartilage rather than a fibrocartilage scar. In order to apply these ex vivo techniques, however, a number of challenges must be overcome. These include further work on i dentifying the optimal combination of stem cells, scaffolds and growth factors, and refining the conditions to enhance cell expansion and chondr ogenesis in vitro and integration of the cells and scaffolds with existing cartilage in vivo Future research efforts may focus on biodegradable scaffolds 111

PAGE 112

laden with cells that will promote chondrogenesis and integrate at the wound margins to yield fully functional articular cartilage rather than fibrocartilage. The mode of delivery of cells to repair articular cartilage depends upon the size of the defect. Small, localized defects may be repaired by direct application of modified MSCs to the site, similar to delivery of c hondrocytes in current chondrocyte transplant procedures, whereas larger cartilage lesions would rely upon scaffolds to fill the defect and hold the cells in place. Modified MSCs that express growth factors and other chondrogenic proteins will provide the appropria te cues to initiate repair responses within cartilage. Caution must be applied, though, as induc tion of certain signaling pathways may trigger osteogenic differentiation. Surgical attempts to re pair damaged cartilage menti oned in Chapter 1 include autologous chondrocyte transplant ation (ACT), which is limit ed to treating trauma, and is unable to repair large cartilage full-thickn ess defects. Gene-based therapies, on the other hand, deliver transgene products to inju red tissue to catalyze a healing r esponse without the need for surgery. A current phase 1 c linical trial uses a modified version of ACT to initiate cartilage repair by infecti ng primary autologous chondrocytes with a virus modified to express TGF1. The modified cells are cultured and expanded in vitro then injected into the knee joint of patient s with degenerative arthritis (TissueGene, Gaithersburg, MD). This type of approach bypasses the risk of inflammatory responses to adenovirus, and it ensures that only the desired cell type is exposed to the potent, chondroinductive transgene. In Chapter 3, we set out to evaluate the ability of candidate chondrogenic transgenes to stimulate MSCs toward chon drogenic differentiation. Based upon finding s 112

PAGE 113

that TGF1 stimulates local proliferati on and chondrogenic differentiation in mesenchymal progenitor cells, we designed our study to include members of the TGFsuperfamily as well as other developmental morphogens and transcription factors (Table 1-1). These factors included BMPs 2, 4, and 7, Indian hedge hog, and Sox9 as well as connective tissue growth factor (CTGF). Of all transgenes evaluated in our aggregate culture system, we found that BMP-4 expression resulting from doses of Ad.BMP-4 of 100 to 1000 vp/cell led to the most robust chondrogenic response. When Ad.BMP-4 was injected intra-articularly in rat knees, however, this secreted factor promoted dramatic cellular expansion and chondrometaplasia of the synovial layer. Because such potent chondrogenic inducers can impact the highly receptive progenitor cells of the synovial lining, it remains important to consider transgenes that are expressed intracellularly, su ch as Sox9. Although we sele cted BMP-4 for our studies in Chapter 4, Sox9 remains an important transgene for cons ideration in developing cartilage repair models, especially in prevent ing ectopic cartilage formation while still promoting defect healing. Our evaluation of viral vectors in Chapt er 4 enabled us to determine the optimal vector for delivery of chondrogenic transgenesin this case BMP-4to MSCs. We discovered that scAAV2-BMP-4 infects bov ine MSCs optimally, and when administered to MSCs at 3000 vg/cell, it drives BMP4 expression that promotes more robust chondrogenic differentiation and matrix synt hesis than that shown in Ad.BMP-4 aggregates Moreover, aggregates treated with scAAV2-BMP-4 were uniformly round without the irregular edges and asymmetry of ten observed in Ad.BMP-4 aggregates. 113

PAGE 114

Delivery of modified stem cells could become a method of choice for clinical applications, in both human and veterinary medicine. Patients own cells can be harvested, modified with scAAVBMP-4 (or a virus carrying ot her chondrogenic factors), and reintroduced to the cartilage defect site within one surgical procedure to augment repair of cartilage damaged by sports injury (as in the case of football players and racehorses) or arthritic disease (such as in older adults and in dogs). A geneor modified cell-delivery approach is not strictly limited to cartilage applications; this technique can be used to enhance bone frac ture healing, heart tissue repair, and numerous other conditions t hat impact animals and people. 114

PAGE 115

LIST OF REFERENCES 1. Marieb, E.N. Tiss ue: a living connection. Human Anatomy & Physiology 119-133 (1998). 2. Lodish, H. et al. Collagen: the fibrous proteins of the matrix. Molecular Cell Biology (1999). 3. Bastiaans en-Jenniskens, Y.M. et al. Cont ribution of collagen net work features to functional properties of engineered cartilage. Osteoarthr. Cartil. 16, 359-366 (2008). 4. Gikas, P.D., Aston, W.J. S. & Briggs, T.W.R. Autolog ous chondrocyte implantation: where do we stand now? J. Orthop. Sci. 13 283-292 (2008). 5. Oegema, T.R., Carpenter, R.J., Hofmeister, F. & Thom pson, R.C. The interaction of the zone of calcified cartilage and subchondral bone in osteoarthritis. Microsc. Res. Tech 37 324-332 (1997). 6. Ghivizzani, S.C., Oligino, T.J., Robbi ns, P.D. & Evans, C.H. Cartilage injury and repair. Phys. Med. Rehabil. Clin. N. Am. 11, 289-307, vi (2000). 7. Buckwalter, J.A. & Mankin, H.J. Articu lar cartilage: tissue design and chondrocytematrix interactions. Instr. Course Lect. 47, 477-486 (1998). 8. Bentley, G. & Minas, T. Tr eating joint damage in young people. BMJ 320 15851588 (2000). 9. Goldring, M.B. Update on the biology of the chondrocyte and new approaches to treating cartilage diseases. Best Pract. Res. Clin. Rheu matol. 20, 1003-1025 (2006). 10. Buschmann, M.D. & Grodzinsky, A.J. A molecular model of proteoglycanassociated electrostatic forc es in cartilage mechanics. J. Biomech. Eng. 117 179192 (1995). 11. Ahmed, T.A.E. & Hincke, M.T. Strategies for articular cartilage lesion repair and functional restoration. Tissue Eng Part B Rev 16, 305-329 (2010). 12. Muir, H. The chondrocyte, architect of cartilage. Biomechanics, structure, function and molecular biology of cartilage matrix macromolecules. Bioessays 17, 10391048 (1995). 13. Martin, J.A. & Buckwalter, J.A. Agi ng, articular cartilage chondrocyte senescence and osteoarthritis. Biogerontology 3 257-264 (2002). 14. Martin, J.A. & Buckwalter, J.A. Roles of articular cartilage aging and chondrocyte senescence in the pathogenes is of osteoarthritis. Iowa Orthop. J. 21, 1-7 (2001). 115

PAGE 116

15. Reddi, A.H. Cartilage morphogenetic proteins: role in joint development, homoeostasis, and regeneration. Ann. Rheum. Dis. 62 Suppl 2 ii73-78 (2003). 16. Ghosh, S., Laha, M. Mondal, S., Sengupta, S. & Kapl an, D.L. In vitro model of mesenchymal condensation during chondr ogenic development. Biomaterials 30, 6530-6540 (2009). 17. DeLise, A.M., Fischer, L. & Tuan, R.S. Cellular interactions and signaling in cartilage development. Osteoarthr. Cartil 8 309-334 (2000). 18. Navagiri, S.S. & Dubey, P.N. The role of hypertrophic ca rtilage in endochondral ossification. Z Mikrosk Anat Forsch 90, 435-446 (1976). 19. Coelho, C.N. & Kosher, R.A. Gap j unctional communication during limb cartilage differentiation. Dev. Biol 144 47-53 (1991). 20. Tuan, R.S. Cellular si gnaling in developmental chondrogenesis: N-cadherin, Wnts, and BMP-2. J. Bone Joint Surg. Am. 85-A Suppl 2 137-141 (2003). 21. Moses, H.L. & Serra, R. Regulation of differentiation by TGF-beta. Curr. Opin. Genet. Dev 6 581-586 (1996). 22. Johnstone, B., Hering, T.M., Caplan, A.I. Goldberg, V.M. & Yoo, J.U. In vitro chondrogenesis of bone marrow-derived mesenchymal progenitor cells. Exp. Cell Res. 238 265-272 (1998). 23. Yoo, J.U. et al. The chondrogenic potential of human bone-marrow-derived mesenchymal progenitor cells. J. Bone Joint Surg. Am. 80 1745-1757 (1998). 24. Barry, F., Boynton, R.E., Liu, B. & Mu rphy, J.M. Chondrogenic differentiation of mesenchymal stem cells from bone marrow: differentiation-dependent gene expressi on of matrix components. Exp. Cell Res. 268 189-200 (2001). 25. Denker, A.E., Nicoll, S.B. & Tuan, R.S. Formation of cartilage-like spheroids by micromass cultures of mu rine C3H10T1/2 ce lls upon treatment with transforming growth factor-beta 1. Differentiation 59, 25-34 (1995). 26. Hill, D.J. & Logan, A. Peptide growth factors and their interactions during chondrogenesis. Prog. Growth Factor Res 4 45-68 (1992). 27. Qi, W.N. & Scully, S.P. Effect of ty pe II collagen in chondrocyte response to TGFbeta 1 regulation. Exp. Cell Res 241 142-150 (1998). 28. Morales, T.I., Joyce, M.E., Sobel, M.E., Daniel pour, D. & Roberts, A.B. Transforming growth factor-beta i n calf ar ticular cartilage organ cultures: synthesis and distribution. Arch. Biochem. Biophys 288 397-405 (1991). 116

PAGE 117

29. Luyten, F.P., Chen, P., Paralkar V. & Reddi, A.H. Recombinant bone morphogenetic protein-4, trans forming growth factor-beta 1, and activin A enhance the cartilage phenotype of articu lar chondrocytes in vitro. Exp. Cell Res 210 224229 (1994). 30. Shuler, F.D. et al. Increased matrix synthesis following adenoviral transfer of a transforming growth factor beta1 gene into articular chondrocytes. J. Orthop. Res 18, 585-592 (2000). 31. Glansbeek, H.L., van Beun ingen, H.M., Vitters, E.L., van der Kraan, P.M. & van den Berg, W.B. Stimulation of articular ca rtilage repair in established arthritis by local administration of trans forming growth factor-beta into murine knee joints. Lab. Invest 78 133-142 (1998). 32. Tuli, R. et al. Transforming growth fa ctor-beta-mediated c hondrogenesis of human mesenchymal progenitor cells involves N-cadherin and mitogen-activated protein kinase and Wnt signaling cross-talk. J. Biol. Chem 278 41227-41236 (2003). 33. Lengner, C.J. et al. Pri mary mouse embry onic fibroblasts: a model of mesenchymal cartilage formation. J. Cell. Physiol 200 327-333 (2004). 34. Wozney, J.M. & Rosen, V. Bone morphogenetic protei n and bone morphogenetic protein gene family in bone formation and repair. Clin. Orthop. Relat. Res 26-37 (1998). 35. Wan, M. & Cao, X. BMP si gnaling in skeletal development. Biochem. Biophys Res. Commun 328 651-657 (2005). 36. Oshin, A.O. & Stewart, M.C. The role of bone morphogenet ic proteins in articular cartilage development, homeostasis and repair. Vet. Com p. Orthop. Traumatol. 20, 151-158 (2007). 37. Yoon, B.S. & Lyons, K. M. Multiple functions of BMPs in chondrogenesis. J. Cell. Biochem 93 93-103 (2004). 38. Ahrens, M. et al. Expr ession of human bone morphogeneti c proteins-2 or -4 in murine mesenchymal progenitor C3H10T1/2 cells induces differentiation into distinct mesenchymal cell lineages. DNA Cell Biol 12, 871-880 (1993). 39. Denker, A.E., Haas, A.R., Nicoll, S.B. & Tuan, R.S. Chondrogenic differentiation of murine C3H10T1/2 multipotential mesenc hymal cells: I. Stimulation by bone morphogenetic protein-2 in highdensity micromass cultures. Differentiation 64 6776 (1999). 40. Aikawa, T. et al. Establishment of bone morphogenetic protein 2 responsive chondrogenic cell line. J. Bone Miner. Res 11 544-553 (1996). 117

PAGE 118

41. Sellers, R.S., Peluso, D. & Morris, E.A. The effect of recombinant human bone morphogenetic protein-2 (rhBM P-2) on the healing of full-thickness defects of articular cartilage. J Bone Joint Surg Am 79, 1452-1463 (1997). 42. Sellers, R.S. et al. Repai r of articular cartilage defects one year after treatment with recombinant human bone morphogenetic protein-2 (rhBMP-2). J Bone Joint Surg Am 82 151-160 (2000). 43. White, A.P. et al. Clinical applications of BMP-7/OP-1 in fractures, nonunions and spinal fusion. Int Orthop 31, 735-741 (2007). 44. Mason, J.M. et al. Cartilage and bone regeneration usi ng gene-enhanced tiss ue engineering. Clin. Orthop. Relat. Res S171-178 (2000). 45. Cook, S.D., Patron, L.P., Salkeld, S.L. & Rueger, D.C. Repair of articular cartilage defects with osteogenic protein-1 (BMP-7) in dogs. J Bone Joint Surg Am 85-A Suppl 3 116-123 (2003). 46. Miljkovic, N.D., Cooper, G.M. & Marra K.G. Chondrogenesis, bone morphogenetic protein-4 and mesenchymal stem cells. Osteoarthr. Car til. 16, 1121-1130 (2008). 47. Steinert, A.F. et al. Hypertrophy is induced during the in vitro chondrogenic differentiation of human mesenchymal stem cells by bone morphogenetic protein-2 and bone morphogenetic prot ein-4 gene transfer. Arthri tis Res. Ther 11, R148 (2009). 48. Kuroda, R. et al. Cartilage repair us ing bone morphogenetic protein 4 and musclederived stem cells. Arthritis Rheum 54 433-442 (2006). 49. Nixon, A.J. et al. Insulinlike growth fa ctor-I gene therapy applic ations for cartilage repair. Clin. Orthop. Relat. Res S201-213 (2000). 50. Nixon, A.J., Saxer, R. A. & Brower-Toland, B. D. Exogenous insulinlike growth factor-I stimulates an autoinductive IGF-I autocrine/paracrine response in chondrocytes. J. Orthop. Res 19, 26-32 (2001). 51. Brower-Toland, B.D. et al. Direct adenovirus-mediated insu lin-like growth factor I gene transfer enhances transplant chondrocyte function. Hum. Gene Ther 12, 117129 (2001). 52. Khan, W.S., Johnson, D.S. & Hardingham, T.E. The potent ial of stem cells in the treatment of knee cartilage defects. Knee (2010). 53. de Crombrugghe, B. et al. Transcr iptional mechanisms of chondrocyte differentiation. Matrix Biol 19 389-394 (2000). 118

PAGE 119

54. Kolettas, E., Muir, H.I., Barrett, J.C. & Hardingham, T.E. Chondrocyte phenotype and cell survival are regulated by cultur e conditions and by specific cytokines through the expression of So x-9 transcription factor. Rheumatology (Oxford) 40 1146-1156 (2001). 55. Bi, W., Deng, J.M., Zhang, Z., Behri nger, R.R. & de Crombrugghe, B. Sox9 is required for cartilage formation. Nat. Genet 22, 85-89 (1999). 56. Lefebvre, V., Behringer, R.R. & de Crombrugghe, B. L-So x5, Sox6 and Sox9 control essential steps of the chondrocyte differentiation pathway. Osteoarthr. Cartil 9 Suppl A S69-75 (2001). 57. Wagner, T. et al. Autosomal sex reversal and campomelic dysplasia are caused by mutations in and around the SRY-related gene SOX9. Cell 79, 1111-1120 (1994). 58. Kupcsik, L., Stoddart, M.J., Li, Z., Benneker, L.M. & Alini, M. Improving Chondrogenesis: Potential and Limitations of SOX9 Gene Transfer and Mechanical Stimulation for Cartil age Tissue Engineering. Tissue Eng Part A (2010). 59. Vortkamp, A. et al. Re gulation of rate of cartil age differentiation by Indian hedgehog and PTH-related protein. Science 273 613-622 (1996). 60. Mak, K.K., Kronenberg, H.M., Chuang, P., Mackem, S. & Yang, Y. Indian hedgehog signals independently of PTHrP to promote chondrocyte hypertrophy. D evelopment 135 1947-1956 (2008). 61. Weiss, S., Hennig, T., Bo ck, R., Steck, E. & Richter, W. Impact of growth factors and PTHrP on early and late chondrogenic di fferentiation of human mesenchymal stem cells. J. Cell. Physiol 223 84-93 (2010). 62. Walsh, D.W., Godson, C., Brazil, D.P. & Martin, F. Ex tracellular BMP-antagonist regulation in development and disease: tied up in knots. Trends Cell Biol 20, 244256 (2010). 63. Reddi, A.H. Interplay between bone morphogenetic proteins and cognate binding proteins in bone and cartilage deve lopment: noggin, chordin and DAN. Arthritis Res. 3 1-5 (2001). 64. Aigner, T. & Stve, J. Collagens--major component of the physiological cartilage matrix, major target of cartilage degenerat ion, major tool in cartilage repair. Adv. Drug Deliv. Rev. 55 1569-1593 (2003). 65. Detterline, A.J., Goldberg, S., Bach, B.R. & Cole, B.J. Tr eatment options for articular cartilage defects of the knee. Orthop Nurs 24 361-366; quiz 367-368 (2005). 66. Ahmed, T.A.E. & Hincke, M.T. Strategies for articular cartilage lesion repair and functional restoration. Tissue Eng Part B Rev. (2010). 119

PAGE 120

67. Siparsky, P., Ryzewicz, M., Peterson, B. & Bartz, R. Arthroscopic treatment of osteoarthritis of the knee: are t here any evidence-based indications? Clin. Orthop. Relat. Res 455 107-112 (2007). 68. Stuart, M.J. & Lubowitz, J.H. What, if any, are the indications for arthroscopic debridement of the osteoarthritic knee? Arthroscopy 22, 238-239 (2006). 69. Ltzner, J., Kasten, P., G nther, K. & Kirschner, S. Su rgical options for patients with osteoarthritis of the knee. Nat Rev Rheu matol 5 309-316 (2009). 70. Hunzik er, E.B. & Quinn, T.M. Surgical re moval of articular cartilage leads to loss of chondrocytes from cartilage bordering the wound edge. J Bone Joint Surg Am 85-A Suppl 2 85-92 (2003). 71. Magnuson, P.B. The classic: Joint debri dement: surgical treatment of degenerative arthritis. Clin. Orthop. Relat. Res 4-12 (1974). 72. Day, B. The indications for arthroscopi c debridement for osteoar thritis of the knee. Orthop. Clin. North A m 36 413-417 (2005). 73. Clarke, H.D. & Scott, W.N. The role of debri dement: through small portals. J Arthroplasty 18, 10-13 (2003). 74. McLaren, A.C., Blokker, C.P., Fowler, P.J., Roth, J.N. & Ro ck, M.G. Arthroscopic dbridement of the knee for osteoarthrosis. Can J Surg 34 595-598 (1991). 75. Kim, H.K., Moran, M.E. & Salter, R.B. The potential for regeneration of articular cartilage in defects created by chondral shaving and subchondral abrasion. An experimental investigation in rabbits. J Bone Joint Surg Am 73 1301-1315 (1991). 76. Nehrer, S. & Minas, T. Treatment of articular cartilage defects. Invest Radiol 35 639-646 (2000). 77. Jackson, R.W. & Dieterichs, C. The resu lts of arthroscopic lavage and debridement of osteoarthritic knees based on the severi ty of degeneration: a 4to 6-year symptomatic follow-up. Arthroscopy 19, 13-20 (2003). 78. Ficat, R.P., Ficat, C., Gedeon, P. & Toussaint, J.B. Spongialization: a new treatment for diseased patellae. Clin. Orthop. Relat. Res 74-83 (1979). 79. Steadman, J.R., Rodkey, W.G. & Rodrigo, J.J. Microfracture: surgical technique and rehabilitation to treat chondral defects. Clin. Orthop. Relat. Res S362-369 (2001). 80. Bekkers, J.E.J., Inklaar, M. & Saris, D.B.F. Treatment sele ction in articular cartilage lesions of the knee: a systematic review. Am. J. Sports Med. 37 Suppl 1 148S55S (2009) 120

PAGE 121

81. Steinwachs, M.R., Gugg i, T. & Kreuz, P.C. Ma rrow stimulation techniques. Injury 39 Suppl 1 S26-31 (2008). 82. Kessler, M.W., Ackerman, G., Dines, J. S. & Grande, D. Emerging technologies and fourth generation issues in cartilage repair. Sports Med Arthrosc 16 246-254 (2008). 83. Steadman, J.R. et al. Outc omes of microfracture for tr aumatic chondral defects of the knee: average 11-year followup. Arthroscopy 19, 477-484 (2003). 84. Salter, R.B. et al. The biological effect of continuous passive motion on the healing of full-thickness defects in articular cart ilage. An experimental investigation in the rabbit. J Bone Joint Surg Am 62, 1232-1251 (1980). 85. Mitchell, N. & Shepard, N. The resurf acing of adult rabbit articular cartilage by multiple perforations through the subchondral bone. J Bone Joint Surg Am 58 230233 (1976). 86. Ficat, R.P., Ficat, C., Gedeon, P. & Toussaint, J.B. Spongialization: a new treatment for diseased patellae. Clin. Orthop. Relat. Res 74-83 (1979). 87. Bhosale, A.M. & Richardson, J.B. Articular cartilage: stru cture, injuries and review of management. Br. Med. Bull 87, 77-95 (2008). 88. Hunzik er, E.B. Articular cartilage repair: basic scienc e and clinical progress. A review of the current status and prospects. Osteoarthr. Cartil 10, 432-463 (2002). 89. Recht, M. et al. MR imaging of cartilage repair procedures. Skeletal Radiol 32 185200 (2003). 90. Hettrich, C.M., Crawford, D. & Rodeo, S.A. Cartilage repair: third-generation cellbased technologies--basic science, surgic al techniques, clinical outcomes. Sports Med Arthrosc 16 230-235 (2008). 91. Polster, J. & Recht, M. Postoperative MR evaluation of chondral repair in the knee. Eur J Radiol 54, 206-213 (2005). 92. Ferruzzi, A. et al. Autologous chondro cyte implantation in the knee joint: open compared with arthroscopic technique. Com parison at a minimum follow-up of five years. J Bone Joint Su rg Am 90 Suppl 4 90-101 (2008). 93. McNickle, A.G., L'Heureux D.R., Yanke, A.B. & Cole, B.J. Outcomes of autologous chondrocyte implantation in a diverse patient population. Am J Sports Med 37 1344-1350 (2009). 94. McNickle, A.G., Provencher, M.T. & Cole, B. J. Overview of ex isting cartilage repair technology. Sports Med Arthrosc 16, 196-201 (2008). 121

PAGE 122

95. Recht, M.P., Goodwin, D.W. Winalski, C.S. & White, L.M. MRI of articular cartilage: revisiting current status and future directions. AJR Am J Roentgenol 185 899-914 (2005). 96. Friedenstein, A.J., Piatetzky-Shapiro, I. I. & Petrakova, K.V. Osteogenesis in transplants of bone marrow cells. J Embryol Exp Morphol 16, 381-390 (1966). 97. Friedenstein, A.J., Chailakhjan, R.K. & Lalykina, K.S. The devel opment of fibroblast colonies in monolayer cultures of guinea-pig bone marrow and spleen cells. Cell Tissue Kinet 3 393-403 (1970). 98. Nth, U., Steinert, A.F. & Tuan, R.S. Technology insight: adult mesenchymal stem cells for osteoarthritis therapy. Nat Clin Pract Rheuma tol 4 371-380 (2008). 99. Kolf, C.M., Cho, E. & Tuan, R.S. Mesenchymal stromal cells. Biology of adult mesenchymal stem cells: regulation of niche, self-renewal and differentiation. Arthritis Res. Ther 9 204 (2007). 100. Chen, F.H. & Tuan, R.S. Mesenchyma l stem cells in arthritic diseases. Arthritis Res. Ther 10, 223 (2008). 101. Dragoo, J.L. et al. Tissue-engineered cartilage and bone using stem cells from human infrapatellar fat pads. J Bone Joint Surg Br 85, 740-747 (2003). 102. Huang, G.T., Gronthos, S. & Shi, S. Me senchymal stem cells deriv ed from dental tissues vs. those from other sources: their biology and ro le in regenerative medicine. J. Dent. Res 88, 792-806 (2009). 103. Alongi, D.J. et al. Stem/progenitor cells from inflamed human dental pulp retain tissue regeneration potential. Regen Med (2010).doi:10.2217/rme.10.30 104. Dowthwaite, G.P. et al. The surface of articular cartilage contains a progenitor cell population. J. Cell. Sci. 117 889-897 (2004). 105. Pittenger, M.F. et al. Multilin eage pot ential of adult human mesenchymal stem cells. Scienc e 284 143-147 (1999). 106. Prockop, D.J. Marrow st romal cells as stem cells for nonhematopoietic tissues. Science 276 71-74 (1997). 107. Tew, S. et al. Differences in repai r responses between immature and mature cartilage. Clin. Orthop. Relat. Res S142-152 (2001). 108. Martin, J.A. & Buckwalter, J.A. Th e role of chondrocyte senescence in the pathogenesis of osteoarthritis and in limiting cartilage repair. J. Bone Joint Surg. Am 85-A Suppl 2 106-110 (2003). 122

PAGE 123

109. Khan, W.S., Adesida, A. B., Tew, S.R., Andrew, J.G. & Hardingham, T.E. The epitope characterisation and the osteogenic differentiation potent ial of human fat pad-derived stem cells is mainta ined with ageing in later life. Injury 40, 150-157 (2009). 110. Krampera, M. et al. Bone marrow mesenchymal stem cells inhibit the response of naive and memory antigen-specif ic T cells to their cognate peptide. Blood 101 3722-3729 (2003). 111. Evans, C.H. et al. Gene transfer to human joints: progress toward a gene therapy of arthritis. Proc. Natl. Acad. Sci. U.S.A 102 8698-8703 (2005). 112. Ghivizzani, S.C., Oligino, T.J., Glorio so, J.C., Robbins, P.D. & Evans, C.H. Direct gene delivery strategies for the trea tment of rheumatoid arthritis. Drug Discov. Today 6 259-267 (2001). 113. Baltzer, A.W. et al. Genetic enhancement of fracture repair: healing of an experimental segmental defect by adenoviral transfer of the BMP-2 gene. Gene Ther 7 734-739 (2000). 114. Peterson, B. et al. Healing of criti c ally sized femoral defects, using genetically modified mesenchymal stem cells from human adipose tissue. Tissue Eng 11, 120129 (2005). 115. Lieberman, J.R. et al. The effect of regional gene th erapy with bone morphogenetic protein-2-producing bone-marrow cells on t he repair of segmental femoral defects in rats. J Bone Joint Surg Am 81, 905-917 (1999). 116. Lou, J., Tu, Y., Ludwig, F.J., Zhang, J. & Manske, P.R. Effect of bone morphogenetic protein-12 gene transfer on mesenchymal progenitor cells. Clin. Orthop. Relat. Res 333-339 (1999). 117. Lou, J., Tu, Y., Burns, M., Silva, M. J. & Manske, P. BMP-12 gene transfer augmentation of lace rated tendon repair. J. Orthop. Res 19, 1199-1202 (2001 ). 118. Betz, O.B. et al. Heal ing of large segmental bone def ects induced by expedited bone morphogenetic protein-2 gene-acti vated, syngeneic muscle grafts. Hum Gene Ther 20, 1589-1596 (2009). 119. van der Rest, M. & Garrone, R. Collagen family of proteins. FASEB J 5 2814-2823 (1991). 120. Gelse, K., Pschl, E. & Aigner, T. Coll agens--structure, function, and biosynthesis. Adv. Drug Deliv. Rev. 55, 1531-1546 (2003). 121. Gordon, M.K. & H ahn, R.A. Collagens. Cell Tissue Res. 339 247-257 (2010). 123

PAGE 124

122. Lee, J.W., Kim, Y.H., Kim, S., Han, S.H. & Hahn, S.B. Chondrogenic differentiation of mesenchymal stem cells and its clinical applications. Yonsei Med. J 45 Suppl 41-47 (2004). 123. Welter, J.F., Solchaga, L.A. & Penick, K. J. Simplification of aggregate culture of human mesenchymal stem cells as a chondrogenic screening assay. BioTechniques 42 732, 734-737 (2007). 124. Hardy, S., Kitamura, M., Harris-Stansil, T., Dai, Y. & Phipps, M.L. Construction of adenovirus vectors through Cr e-lox recombination. J. Virol. 71, 1842-1849 (1997). 125. Palmer, G.D. et al. Ge ne transfer to articular c hondrocytes with recombinant adenovirus. Cytokines and Colony Stimulating Factors 215 235-246 (2003). 126. Colter, D.C., Class, R., DiGirolamo, C.M. & Prockop, D.J. Rapid expansion of recycling stem cells in cultures of plas tic-adherent cells from human bone marrow. Proc. Natl. Acad. Sci. U.S.A 97 3213-3218 (2000). 127. Caplan, A.I. Review: mesenchymal stem cells: cell-based reconstructive therapy in orthopedics. Tissue Eng. 11, 1198-1211 (2005). 128. Steinert, A.F. et al. Major biological obs tacles for persistent cell-based regeneration of articular cartilage. Arthritis Res. Ther 9 213 (2007). 129. Nixon, A.J. et al. Gene t herapy in musculoskeletal repair. Ann. N. Y. Acad. Sci 1117 310-327 (2007). 130. Palmer, G.D. et al. Gene-induced chondr ogenesis of primary mesenchymal stem cells in vitro. Mol. The r. 12, 219-228 (2005). 131. Song, J.J. et al. Connective tissue grow th factor (CTGF) ac ts as a downstream mediator of TGF-beta1 to induc e mesenchymal cell condensation. J. Cell. Physiol 210 398-410 (2007). 132. Kay, J.D. et al. Intraarticular gene delivery and express ion of interleukin-1Ra mediated by self-complement ary adeno-ass ociated virus. J Gene Med 11, 605-614 (2009). 124

PAGE 125

BIOGRAPHICAL SKETCH Marsha Lynn Bush was raised in the touris t town of Cave City, Kentucky, where she grew up on a farm, in a zoo, and next door to the longest cave in the world. After graduating valedictorian from Barren County High School in 1999, Marsha attended the University of Kentucky. She spent the summers of 2000, 2002, and 2003 serving as a guide at Mammoth Ca ve National Park. It was here that Mars ha learned of the variety of research projects that were taking pl ace spanning the fields of geology, biology, hydrology, microbiology, anthropology, gene alogy, and paleontology. During the summer of 2001, Marsha began her own research pursuits as a student in the Kentucky Young Scientist Summer (KYSS) Research Program. Under the guidance of Dr. Charles (Chuck) Fox, Marsha took part in a population genetics ex periment examining maternal effects on offspring lifespan in the seed beetle and stored grain pest, Callosobruchus maculatus An offshoot of this projec t became her prim ary research project for completion of her Bachelors degree in agricultural biotechnology. Prior to graduating in December 2003, Ma rsha spent the spring seme ster as the biotechnology lab intern at The Land at Epcot. As part of Epcot Science, Marsha learned about plant tissue culture as she propagated a vari ety of species used within The Land greenhouses and grew, packaged, and managed the sales of Mickeys Mini Gardens at four locations within Walt Disney World. Th is internship experience opened the door for Marshas return to Florida in the spring of 2004 as a reproductive biology intern at Disneys Animal Kingdom Wildlife Tracki ng Center. This experience cemented her desire to take part in research efforts t hat would benefit both animals and people. In August 2004, Marsha began studies in the Interdisciplinary Program (IDP) in Biomedical Sciences at the University of Florida College of Medicine. In May 2005, 125

PAGE 126

Marsha joined the lab of Dr. Steve Ghivi zzani in the Departm ent of Orthopaedics and Rehabilitation where she has investigated the induc tion of chondrogenesis in mesenchymal stem cells following delivery of trangenes via viral vectors. Her work may contribute to the understanding of gene therapy treatments for repair of articular cartilage damage resulting from injuries or diseases such as arthritis. Following completion of her Ph.D. in 2010, Marsha will c ontinue her education at the University of Wisconsin-Madison School of Veterinary Medici ne, as she combines her interests in veterinary medicine and translational research. 126