The Effects of Endothelial Progenitor Cell Therapy on Fracture ...

The Effects of Endothelial Progenitor Cell Therapy on Fracture Healing and Infection

Status Outcomes in a Low-Grade Infected Rat Critical-Size Defect Model

by

Richard Magony

A thesis submitted in conformity with the requirements for the degree of Master of Science

Institute of Medical Science

University of Toronto

© Copyright by Richard Magony, 2020

ii

The Effects of Endothelial Progenitor Cell Therapy on Fracture Healing and Infection

Status Outcomes in a Low-Grade Infected Rat Critical-Size Defect Model

Richard Magony Master’s of Science

Institute of Medical Science University of Toronto

2020

Abstract

Low-grade infection represents a common cause of nonunion and complicates its

management. Endothelial progenitor cells (EPCs) are capable of healing bone defects,

but their potential antimicrobial properties remain unknown. We aimed to evaluate

EPCs’ bone healing and infection treatment potential in an infected nonunion setting,

with and without local antibiotics. In our first three experiments, we established a

consistent low-grade infected nonunion model, selected an effective local antibiotic

therapy, encountered EPC functionality issues while investigating bone healing

outcomes, and troubleshooted to reproduce high bone healing rates. Our final

experiment analyzed infection outcomes and did not demonstrate any significant effect

of EPCs on low-grade infection eradication in the presence or absence of local

antibiotics. However, the final study was small-scale and lacked sufficient statistical

power to demonstrate an effect. Therefore, further experimentation with larger-scale

study designs are necessary to more adequately investigate the combination of EPCs

and local antibiotics as a potential single-stage therapy for low-grade infected

nonunions.

iii

Acknowledgements

I wish to extend my utmost gratitude towards each person who had a significant role in

my graduate training experience. The transition from undergraduate- to graduate-level

studies was a difficult adjustment that I could not have done without all the help I

received from my supervisors and colleagues. I would like to start by thanking Dr. Aaron

Nauth and Dr. Emil Schemitsch for offering their guidance and support at every step of

the way and teaching me how to critically analyze my work. I admire the passion that

you both have for advancing the field of orthopaedic surgery and am truly proud to have

had a role in your research towards new therapies for your patients. I must also thank

my Program Advisory Committee members, Dr. John Davies and Dr. Matthew Muller,

for providing me with their insight and constructive criticism along the way. Their unique

perspectives on my work were helpful for designing my experiments and coming up with

creative solutions for any experimental issues I encountered. Thus, I cannot overstate

how thankful I am for all the opportunities I had to connect with such bright and highly

accomplished individuals and draw from their areas of expertise.

I would also like to thank the lab personnel who taught me the various wet and dry

techniques necessary for my project and heavily contributed to the progress I made. Dr.

Charles Godbout, our lab’s former research associate, provided me with significant

mentorship in the early stages of my degree and helped me lay the groundwork for my

project during my first year. Stephane Gagnon, our lab’s technician, performed a vital

role in essentially all of my project’s wet lab procedures and demonstrated an utmost

commitment to my work, having to learn complex techniques for animal surgeries within

a very short timeframe. I recognize that this involved countless hours of frustration and

perseverance, and so I cannot thank you enough for the sacrifices you made in order

for us to accomplish the project goals. Finally, Hening Sun and Ikran Ali served as great

role models for me as highly motivated fellow graduate students who always performed

their best work and were willing to lend a helping hand with my experiments. It was truly

a pleasure to be a part of this talented and cooperative research team and I wish the

best of luck for all of its members in their respective future endeavours.

iv

Contributions

The research presented in this thesis was designed and performed by the author

(Richard Magony) under the guidance of supervisors and Program Advisory Committee

members and with the help of the following technical personnel:

Surgical Personnel Stephane Gagnon – performed cell preparations and the majority of animal surgeries

Charles Godbout – performed animal surgeries in the pilot study

St. Michael’s Research Vivarium Personnel Danielle Gifford and fellow staff members – helped with preparing the operating rooms,

provided the necessary supplies for surgical procedures, and provided the equipment

for performing radiographic imaging

I am also grateful for the financial support provided by the 2019-2020 Queen Elizabeth

II Graduate Scholarships in Science and Technology. Without such grants, our work

would not be possible.

v

Table of Contents

Abstract ........................................................................................................................... ii Acknowledgements ....................................................................................................... iii Contributions ................................................................................................................. iv

List of Figures ................................................................................................................ ix

List of Tables .................................................................................................................. x

Chapter 1: Literature Review ......................................................................................... 1

1.1 Fracture Nonunion and Segmental Bone Defects ................................................. 1

1.1.1 Definitions .............................................................................................................. 1

1.1.2 Nonunion Clinical Relevance and Diagnosis ..................................................... 3

1.1.3 Bone Healing Biology ........................................................................................... 4

1.1.3.1 Primary Healing .................................................................................................. 4

1.1.3.2 Secondary Healing ............................................................................................. 7

1.1.4 Nonunion Risk Factors ....................................................................................... 10

1.1.5 Nonunion Treatment Strategies ......................................................................... 10

1.2 Surgical Site Infections, Antibiotic Therapies and Biofilms .......................... 12

1.2.1 Surgical Site Infections in Orthopaedic Surgery .............................................. 12

1.2.1.1 Definition and Epidemiology ........................................................................... 12

1.2.1.2 Clinical Significance ........................................................................................ 13

1.2.1.3 Preventative Measures in the OR ................................................................... 13

1.2.1.4 Microbiology and Diagnosis ............................................................................ 14

1.2.1.5 Low-Grade Infections and Nonunions ...................................................... 18

1.2.1.6 Animal Model of Low-Grade Infected Nonunion ....................................... 19

1.2.2 Antibiotic Therapies and Biofilm Infections ................................................ 20

1.2.2.1 Antibiotic Prophylaxis in Orthopaedic Surgery ........................................ 20

1.2.2.2 Local Antibiotics in Perioperative Prophylaxis ........................................ 22

1.2.2.3 Biofilm Infections and Antibiotics ............................................................. 23

1.2.2.4 Local Antibiotic Therapies and Low-Grade Infected Nonunions ............ 26

1.3 Endothelial Progenitor Cells (EPCs) .................................................................. 27

1.3.1 Clinical Relevance ........................................................................................... 27

1.3.2 Classification ................................................................................................... 28

1.3.3 Recent Discoveries of EPC-based Therapies in Fracture Healing ............. 29

vi

1.3.4 EPCs and Infection .......................................................................................... 31

1.3.5 EPCs and Antibiotics ...................................................................................... 31

Chapter 2: Rationale, Aims, and Hypothesis ............................................................. 33

Chapter 3: Experiment #1 – Pilot Study ..................................................................... 36

3.1 Rationale and Aims ................................................................................................ 36

3.2 Methods ................................................................................................................... 36

3.2.1 Experimental Design ........................................................................................... 36

3.2.2 Bacteria and Gelfoam Scaffold Preparation ..................................................... 37

3.2.3 Surgical Procedures ........................................................................................... 38

3.2.4 Euthanasia and Tissue Sample Harvest ........................................................... 39

3.2.5 Tissue Culture, Sonication and Microbiological Analysis............................... 39

3.3 Results .................................................................................................................... 41

Chapter 4: Experiment #2 - EPCs, Antibiotics, and Bone Healing in a Non-contaminated Critical-Size Defect Model ................................................................... 43


4.2 Methods ................................................................................................................... 43


4.2.2 Cell Isolation, Culture, and Characterization .................................................... 44

4.2.3 Gelfoam Scaffold Preparation and Cell Seeding .............................................. 45


4.2.5 Euthanasia and Harvest ...................................................................................... 47

4.2.6 Radiography......................................................................................................... 48

4.2.7 Tissue Homogenization and Microbiological Culture ...................................... 48

4.2.8 Study Power and Statistical Analyses .............................................................. 49

4.3 Results .................................................................................................................... 49

4.3.1 Radiography......................................................................................................... 49

4.3.2 Microbiology ........................................................................................................ 54

4.3.3 Summary of Experimental Errors and Contingency Plans ............................. 54

Chapter 5: Experiment #3 - Acute and Delayed EPC Treatment and Bone Healing ........................................................................................................................................ 55


5.2 Methods ................................................................................................................... 55


vii

5.2.2 Radiography......................................................................................................... 56

5.2.3 Cell Isolation and Culture ................................................................................... 56



5.2.6 Euthanasia and Harvest ...................................................................................... 57

5.3 Results .................................................................................................................... 57

Chapter 6: Experiment #4 - EPCs, Local Antibiotics, and Infection Outcomes in a Contaminated Critical-size Defect Model ................................................................... 58


6.2 Methods ................................................................................................................... 58


6.2.2 Bacteria and Gelfoam Scaffold Preparation ..................................................... 59

6.2.3 Radiography......................................................................................................... 59

6.2.4 Cell Isolation and Culture ................................................................................... 60



6.2.7 Tissue Culture, Sonication, and Microbiological Analysis.............................. 61

6.2.8 Statistical Analyses ............................................................................................. 61

6.3 Results .................................................................................................................... 61

6.3.1 Radiography......................................................................................................... 62

6.3.2 Microbiology ........................................................................................................ 64

Chapter 7: Discussion .................................................................................................. 66

7.1 Low-Grade Infection Animal Model ...................................................................... 67

7.2 Combination Antibiotic Therapy ........................................................................... 68

7.3 Technical Errors and EPC Functionality .............................................................. 69

7.4 EPCs, Antibiotics, and Bone Healing ............................................................ 71

7.5 EPCs, Antibiotics, and Infection ........................................................................... 72

7.6 Limitations .............................................................................................................. 73

Chapter 8: Conclusions ............................................................................................... 77

Chapter 9: Future Directions ....................................................................................... 78

References .................................................................................................................... 82

viii

List of Abbreviations AAOS – American Academy of Orthopaedic Surgeons

Ac-LDL – acetylated low density lipoprotein

AICBG – autologous iliac crest bone graft

ANOVA – analysis of variance

ASHP – American Society of Health System Pharmacists

BMP – bone morphogenetic protein

CAC – circulating angiogenic cell

CD14 – cluster of differentiation molecule 14



CI – confidence interval

CFU – colony-forming units

CT – computed tomography

ECFC – endothelial colony-forming cell

EDTA – ethylenediaminetetraacetic acid

EGF – epidermal growth factor

EGM-2MV – microvascular endothelial cell growth medium-2

EPC – endothelial progenitor cell

E-EPC – early endothelial progenitor cell

L-EPC – late endothelial progenitor cell

FBS – fetal bovine serum

FDA – Food and Drug Administration

FGF – fibroblast growth factor

IGF – insulin-like growth factor

IL – interleukin

MAC – myeloid angiogenic cell

MSC – mesenchymal stem cell

MRSA – methicillin-resistant staphylococcus aureus

N/A – not available

ix

NSAID – non-steroidal anti-inflammatory drug

OR – operating room

PBS – phosphate buffer solution

PJI - prosthetic joint infection

RANK-L – receptor activation of nuclear factor κ B

RCT – randomized controlled trial

RIA – reamer-irrigator-aspirator

SOP – standard operating procedure

SSI – surgical site infection

TGF-β – transforming growth factor beta

TNF-a – tumor necrosis factor alpha

TNS – trypsin neutralizing solution

TNTC – too numerous to count

TSA – tryptic soy agar

TSB – tryptic soy broth

UEA-1 – Ulex europaeus agglutinin 1 lectin

V+R – vancomycin and rifampin

VEGF – vascular endothelial growth factor

List of Figures

Figure 1-1: Illustration of the ‘diamond concept’ of bone healing.

Figure 1-2: Illustration of the biological events occurring at different phases of

secondary fracture healing.

Figure 1-3: Diagnostic algorithm for periprosthetic joint infection developed by the

Infectious Disease Society of America.

Figure 1-4: Results of scanning electron microscopy analysis for biofilm-producing

staphylococcal strains and S. epidermidis ATCC 35984.

Figure 1-5: Radiographic union was improved when EPCs were applied acutely and in

delayed fashion.

Figure 4-1: Mean 10-week radiographic scores across treatment groups.

x

Figure 4-2: Serial radiographs (2 wks, 6 wks, and 10 wks from left to right) of one rat

from each of the five treatments.

Figure 5-1: Representative radiographs (2 wks, 6wks, and 10wks from left to right)

radiographs from rats in the acute and delayed EPC treatment groups.

Figure 6-1: Graphical representation of radiographic scores for bone healing and

infection status across treatment groups

Figure 6-2: Representative radiographs (0 wks and 2 wks from left to right) from each

treatment group.

Figure 6-3: Graphical representation of infection outcomes at 2 weeks post-treatment.

List of Tables

Table 3-1: Tissue culture results for different S. epidermidis doses and antibiotic

regimens.

Table 4-1: Number of rats treated per group.

Table 4-2: Radiographic scoring system based on defect filling and callus density.

Table 4-3: Bone healing outcomes at 10 weeks after treatment and contamination.

outcomes based on cultures of tissue biopsies taken during the second stage surgery.

Table 4-4: Contamination outcomes at 0 weeks based on overnight cultures of biopsies

taken during second stage surgery.

Table 4-5: Results from Tukey’s multiple comparison test comparing mean radiographic

bone healing scores for treatment groups in experiment #2.

Table 5-1: Summary of healing status outcomes and mean radiographic scores

between the acute and delayed EPC treatment groups.

Table 6-1: Radiographic scoring system based on signs of infection.

Table 6-2: Summary of radiographic scores for bone healing and infection status at 2

weeks post-treatment for all treatment groups.


infection scores for treatment groups in experiment #4.

1

Chapter 1: Literature Review

1.1 Fracture Nonunion and Segmental Bone Defects

1.1.1 Definitions

A segmental bone defect refers to a circumferential absence of bone tissue at a site

where it normally exists, often resulting from trauma, disease or tumor resection. The

most common cause is high-energy trauma, which can lead to devastating patient

outcomes depending on numerous patient-dependent and -independent factors. In

some cases, the bone defect(s) created is large enough that the gap will not heal

without an intervention. Surgeons refer to these as “critical-size” defects, which typically

have a >50% circumferential loss or a length of >2 cm in adult patients (Keating et al.

2005). In animal models, critical-size bone defects are defined as the smallest size of

defects that would either not heal independent of treatment or heal less than 10% of the

time over the course of the animal’s lifetime (Hollinger and Kleinschmidt 1990, Gugala

and Gogolewski 1999).

Without any treatment, critical-size defects fail to heal and the resulting scenario is

known as fracture nonunion. More precise definitions of nonunion vary due to its

complex aetiology, but the commonly used US Food and Drugs Administration (FDA)

criteria define it as the lack of bone healing over 9 months after injury with absent

radiological signs of healing progress over 3 consecutive months (Somford et al. 2013).

A variety of nonunion risk factors have been identified, which will be described in more

detail later. Since bone healing is such a complex process, these factors can prevent

proper fracture healing via different physiological mechanisms, ultimately producing

different types of nonunions.

2

In order to categorize nonunions, Weber and Cech designed a system that classifies

them into three types according to radiographic criteria: hypertrophic, atrophic, or

oligotrophic (Weber and Cech 1976). Hypertrophic nonunions are identified

radiographically as having a large quantity of callus formation at the fracture site and

occur when biomechanical stability is insufficient. Such nonunions can be treated by the

addition of mechanical stability with internal fixation either with or without bone graft

(Uzun et al. 2015). Atrophic nonunions are characterized by their lack of callus

formation resulting from an impaired biological environment and disrupted

vascularization near the bone ends. Surgical interventions are focused on both

establishing biomechanical stability and enhancing bone biology to promote bone

healing (Said et al. 2013). The standard treatment includes autogenous bone grafting

and internal fixation, although bone morphogenetic proteins (BMPs) may also be used

as a biological stimulus. Finally, oligotrophic nonunions have minimal callus formation

and some vascularity, but either biological or mechanical factors may contribute to the

lack of healing and appropriate treatment strategies depend on which factors are

present (Schmal et al. 2020).

It is important to note the key differences between critical-size defects and nonunions.

While nonunions lack either sufficient biological support or biomechanical stability for

proper bone healing, they can occur without a fracture gap or bone defect. As for

critical-size defects, they may have optimal biological environments but bone loss is

usually the main issue. Thus, bone forming treatments are not always necessary for

nonunion cases but is required for healing critical-size defects (Keating et al. 2005).

Nonetheless, the concept of the critical-size defect has been applied in animal studies

by our lab, and others, in order to model a nonunion scenario and investigate new

fracture healing therapies (Bates et al. 2017, Li et al. 2014, Li et al. 2009, Atesok et al.

2010, Seebach et al. 2012).

3

1.1.2 Nonunion Clinical Relevance and Diagnosis

Despite advances in orthopaedic surgical interventions, the management of segmental

bone defects and nonunions remains a significant clinical problem faced by orthopaedic

surgeons. Reported incidences of nonunion vary between 1.9-4.9% of fractures

depending on the type of bone and many other factors, affecting up to 100,000 fracture

patients each year in the United States (Stewart 2019, Hak et al. 2014). Nonunions

impose a substantial socioeconomic burden since they require a high cost of treatment

and cause a tremendous loss of work productivity (Kanakaris and Giannoudis 2007).

Chronic pain at the fracture site is often severe enough to prevent the patient from

participating fully in social activities in addition to their work. Thus, the physical

debilitation caused by nonunion severely impacts the patients’ mental and physical

health-related quality of life, as reflected by studies’ survey results. In fact, these results

indicate that health effects due to femoral and tibial nonunion are worse than many

other serious chronic conditions, including end-stage hip arthrosis and congestive heart

failure (Brinker et al. 2013, Brinker et al. 2016). Therefore, the discovery of more

effective and efficient therapies could both help alleviate the tremendous suffering of

fracture nonunion patients and reduce the significant financial burden associated with

nonunion management.

The diagnosis of fracture nonunion is based on a combination of clinical and

radiographic evidence. Clinically, symptoms including the inability to bear weight, pain

at the fracture site and tenderness on palpation are positive signs (Bhandari et al.

2012). Radiologically, absence of callus formation is an important indicator in addition to

the persistence of fracture lines. Ambiguous radiographic results can be supplemented

by computed tomography (CT) scanning when determining whether or not to proceed

with surgery. However, CT has insufficient diagnostic accuracy on its own (Hak et al.

2014, Schmal et al. 2020). As described earlier, physicians use radiographs to

distinguish between hypertrophic, atrophic and oligotrophic nonunions and guide their

decision-making.

4

1.1.3 Bone Healing Biology

Successful fracture healing resulting in bone repair is a complex biological process that

depends on multiple necessary components, summarized by the diamond concept. This

framework emphasizes the importance of the following four key components:

1. Osteogenic cells

2. Osteoconductive scaffolds

3. Growth factors

4. Biomechanical stability

Two additional factors have since been added to this list: vascularization and host

factors. Sufficient vascularization allows the delivery of osteogenic cells and

osteoinductive mediators to the fracture site, which facilitates processes including

osteogenesis and bone remodeling (Wang et al. 2017). Moreover, events such as

traumas and surgery can damage the local vasculature, disrupting these processes and

contributing to fracture nonunion. In addition, the presence of certain host factors such

as systemic comorbidities also tends to negatively impact bone healing outcomes,

including chronic conditions like diabetes, peripheral neuropathy, obesity, rheumatoid

arthritis and malnutrition (Corupoglu et al. 2013). Therefore, a wide variety of patient-

dependent and independent factors are responsible for influencing fracture healing

outcomes, contributing to the wide variety of patients at risk.

With regards to fracture healing, there are two broad categories: primary (direct) and

secondary (indirect) healing.

1.1.3.1 Primary Healing

Primary healing occurs when the fracture ends are in contact with one another and

there is limited interfragmentary strain. This requires anatomic reduction of the fracture

fragments, compression across the fracture and stability provided by internal fixation.

Specifically, contact healing can occur if the gap between fragments is less than 0.01

5

mm and the interfragmentary strain is less than 2% (Shapiro 1988). During this process,

cutting cones create longitudinal cavities at the fracture site through osteoclastic

tunneling, which is later replaced by bone and the Harversian system via osteoblastic

activity. These Haversian blood vessels allow the delivery of osteoblastic progenitor

cells to the fracture site. Eventually, the bone remodels into a lamellar structure with

little to no periosteal callus formation.

The other type of primary healing is known as gap healing, which occurs under similar

conditions of anatomic reduction and low interfragmentary strain. However, the fracture

gap must be between 800 µm and 1 mm wide (Kaderly 1991). This process takes

longer, since the lamellar bone is initially oriented perpendicularly to the bone axis and

must be subsequently reconstructed and replaced by longitudinal osteons, taking

between 3 and 8 weeks. After this, the bone undergoes more remodelling in a similar

fashion to the contact healing process in order to finally establish lamellar bone (Shapiro

1988).

6

Figure 1-1: Illustration of the ‘diamond concept’ of bone healing. Reproduced with

permission from Andrzejowski, P. & Giannoudis, P. V. 2019. The ‘diamond concept’ for

long-bone nonunion management. J Orthop Traumatol. 20: 21. Terms of distribution:

http://creativecommons.org/licenses/by/4.0/

7

1.1.3.2 Secondary Healing

The majority of fracture healing occurs via secondary healing, which does not require

anatomical reduction and rigid stability but instead benefits from micromotion. Both

endochondral and intramembranous ossification are integral to this type of healing

(Gerstenfeld et al. 2006). This process is much more complex than primary healing,

involving the following detailed steps.

The fracture injury triggers the formation of a haematoma and initiates the fracture

healing cascade of events via the secretion of inflammatory cytokines. The fracture

haematoma consists of endothelial cells, mesenchymal stem cells (MSCs), and immune

cells, which ultimately help create the extracellular matrix prior to soft callus formation.

Pro-angiogenic and pro-inflammatory factors are secreted, including tumor necrosis

factor-a (TNF-a), bone morphogenic proteins (BMPs), interleukin-1 (IL-1), IL-6, IL-11,

and IL-18 (Gerstenfeld et al. 2003). Aside from inflammatory signalling, TNF-a plays a

chemotactic role and also facilitates the osteogenic differentiation of MSCs. IL-1 and IL-

6 have significant roles in fracture healing, contributing to cartilaginous callus

production, angiogenesis and the differentiation of osteoblasts and osteoclasts (Kon et

al. 2001, Lee and Lorenzo 2006, Sfeir et al. 2007, Yang et al. 2007). Vascular

endothelial growth factor (VEGF) is largely responsible for promoting angiogenesis at

the fracture site during this early stage. These events take place during the first 1 to 5

days following the fracture (Sheen and Garla 2020).

In days 5 to 11, fibrin-rich granulation tissue begins to form within the haematoma. This

tissue serves as a matrix that delivers signalling factors crucial for migration,

proliferation and differentiation of cells at the fracture site. BMPs help recruit MSCs from

many local areas including the cortex, periosteum, bone marrow and nearby soft

tissues, as well as systemically from other hematopoietic regions. These MSCs

differentiate into fibroblasts, chondroblasts and osteoblasts, which are all involved in

soft callus formation via chondrogenesis (the formation of cartilage). Transforming

growth factor-beta (TGF-β) proteins are important in this signalling cascade (Yang et al.

2007, Sheen and Garla 2020, Marsell and Einhorn 2011).

8

Days 11 to 28 are characterized by endochondral ossification, or the transformation of

cartilaginous callus into immature bone. The increased expression of receptor activator

of nuclear factor κ B (RANK-L) facilitates the differentiation of cells including

chondroblasts, chondroclasts, osteoblasts and osteoclasts, which together cause the

resorption and calcification of soft callus. Subperiosteal woven bone is formed, and new

blood vessels help deliver migrating MSCs to the fracture site. These events allow the

development of hard, bony callus (Sheen and Garla 2020).

From day 18 onwards, the long process of bone and vascular remodelling takes place

over many months, ultimately resulting in the formation of mature bone. A balance

between bone resorption via osteoclasts and bone formation via osteoblasts, known as

“coupled remodelling”, is essential to this process. Compact bone replaces the centre of

the hard callus while lamellar bone replaces the peripheral areas, gradually

transforming the hard callus into normal bone structure (Sheen and Garla 2020).

9

Figure 1-2: Illustration of the biological events occurring at different phases of

secondary fracture healing. The primary metabolic phases (blue bars) of fracture

healing overlap with biological phases (brown bars). The time scale of healing is

equivalent to a mouse closed femur fracture fixed with an intramedullary rod.

Abbreviations: PMN, polymorphonuclear leukocyte. Reproduced with permission from

Einhorn, T. A. & Gerstenfeld, L. C. 2015. Fracture healing: Mechanisms and

interventions. Nat Rev Rheumatol. 11(1): 45-54.

10

1.1.4 Nonunion Risk Factors

Since fracture healing is a complex process influenced by biological and mechanical

variables, there is a large number of factors that can impair healing and cause the

progression of a fracture to nonunion. Firstly, the main patient-related factors identified

are comorbidities and medications. Diseases like diabetes, obesity, anaemia,

malnutrition, peripheral vascular disease and usage of medications such as non-

steroidal anti-inflammatory drugs (NSAIDs) have negative biological impacts on bone

repair. Similarly, certain unhealthy lifestyle habits such as smoking and excessive

alcohol consumption can also impair healing. Aside from host characteristics, many

factors related to the fracture itself and the surgical treatments performed on them also

significantly impact healing outcomes. For instance, since biomechanical stability and

anatomical alignment are prerequisites for successful bone repair, fracture

characteristics including interfragmentary gap, bone displacement, fixation methods and

surgical techniques are critically important (Hernandez et al. 2012, Bishop et al. 2012,

Gaston and Simpson 2007, Zura et al. 2016). Infection and osteonecrosis are also

common causes. One study with prospective and retrospective data on 100 patients

who received surgical treatment for their long bone nonunion reflects the prevalence of

four major causative agents of nonunion. From greatest to least common, the factors

were mechanical (58%), dead bone with gap (46%), host (43%) and infection (38%)

(Mills et al. 2016). This study also showed that the cause of nonunion was also often

multifactorial with 69% of patients having more than one factor present. These risk

factors are widely known by surgeons, but evidence of their contributions to nonunion

progression originates from observational and preclinical studies. Hence, individual risk

factors are not relied upon for definitive nonunion prognoses. They can, however, offer

insight into nonunion aetiology and be considered together to estimate a patient’s

overall risk of nonunion.

1.1.5 Nonunion Treatment Strategies

The current gold standard treatment for bone defects is autologous iliac crest bone

grafting (AICBG). Autologous bone graft provides osteoinductive mediators, osteogenic

11

cells and an osteoconductive scaffold simultaneously, allowing it to successfully heal

atrophic nonunions that require enhanced biological factors. In addition, the

histocompatibility of autologous bone graft reduces the risk of immune rejection.

Unfortunately, harvest site pain and lengthened hospital stays are common

complications. Intramedullary femoral reaming and debris aspiration using the Reamer-

Irrigator-Aspirator (RIA) is a more recently developed bone grafting method that also

offers a promising source of the three components provided by bone graft. However,

similar to AICBG, bone graft supply is limited and RIA requires another surgical

procedure that presents an additional risk of complications. Thus, there has been an

increasing interest in researching other bone regeneration therapies that are less

invasive, less painful and more cost-effective (Nauth et al. 2018).

Current methods of treating nonunions mainly fall under the three categories of bone

grafting, biologic therapies and cell-based therapies. Bone marrow aspirate concentrate

is an example of a promising future grafting option that provides osteogenic cells and

osteoinductive mediators which can be combined with unlimited, commercially available

osteoconductive scaffolds while largely avoiding an invasive harvest procedure. Current

biologic therapies include demineralized bone matrix, BMPs, and systemic parathyroid

hormone therapy, but they lack conclusive evidence of clinical efficacy and also carry

their own share of complications. Finally, research on cell-based therapies is gaining

more attention since the derivation of osteogenic cells is the most challenging

component of bone regeneration faced by surgeons. These therapies often include

bone marrow derivatives, such as bone marrow-derived mesenchymal stromal cells and

bone marrow concentrate, which are often seeded on structures known as scaffolds that

facilitate cellular interactions promoting tissue regeneration (Palombella et al. 2019).

Current research is still in its early stages, but further investigations may lead to new

strategies of harvesting osteogenic cell populations and concentrating them on carriers

to promote bone regeneration in fractures.

12

1.2 Surgical Site Infections, Antibiotic Therapies and Biofilms

1.2.1 Surgical Site Infections in Orthopaedic Surgery

1.2.1.1 Definition and Epidemiology

According to the Centers for Disease Control and Prevention, surgical site infection

(SSI) is defined as an infection associated with a surgical procedure that occurs at the

surgical site within 30 days after the operation or up to one year afterwards if the

procedure included implantation of a prosthetic device(s) (Mangram et al. 1999). SSI

remains a common complication in surgery, affecting up to 300,000 patients per year in

the United States (Ban et al. 2017). Recent studies report SSI rates of 0.3-2.6% in

orthopaedic surgery, but rates vary significantly between different types of operations

and levels of risk (Al-Mulhim et al. 2014, Bhat et al. 2018, Brophy et al. 2019). For

instance, SSIs occur in as low as 1% of low-risk patients in clean procedures, such as

joint replacement surgeries, and as high as 16% of high-risk patients in contaminated

procedures, including many emergency trauma cases (Debarge et al. 2007, Kapadia et

al. 2013, Nichols 2004). SSI risk is influenced by a multitude of factors that can be

categorized as patient-related or procedure-related. Patient-related factors include host

characteristics and lifestyle choices like age, diabetes, smoking, previous infections,

liver or kidney diseases, excessive alcohol consumption and drug usage (Mangram et

al. 1999, Florschutz et al. 2015). Examples of procedure-related factors include

operation duration, operating room (OR) ventilation, surgical techniques and the quality

of skin and instrument sterilization. Some scoring systems have been developed to help

calculate the level of SSI risk for a given patient based on the risk factors present

(Mangram et al. 1999).

13

1.2.1.2 Clinical Significance

A case-control study from 2002 identified several major impacts that orthopaedic SSIs

have on patients and the healthcare system. Regarding hospital experiences, patients

with SSIs typically underwent longer hospitalization periods and required more total

hospitalizations and surgical procedures over the course of their treatment. Moreover,

SSIs placed a significant extra financial burden on the treatment of orthopaedic patients,

costing an additional $27,969 per patient over a 1-year study period. Finally, quality of

life measures were significantly lower for infected patients, especially those measures

related to physical capabilities (Whitehouse et al. 2002). Therefore, strategies that

effectively treat SSIs and prevent their occurrence are critical to both patient outcomes

and healthcare system efficiency.

1.2.1.3 Preventative Measures in the OR

Through understanding the SSIs’ typical modes of transmission, standard practices

have been developed for OR staff to follow as preventative measures. Studies found

that airborne particles serve as vectors for contaminants in up to 98% of orthopaedic

SSIs, whereas only 2% of infections are caused by pathogens directly from the patients’

skin (Talon et al. 2006). Amongst these airborne cases, 30% involve direct transmission

to the surgical site and 70% occur indirectly via contamination of surgical instruments or

the surgeon’s hands, followed by transmission to the site (Whyte et al. 1982). Thus, lots

of preventative measures taken by OR personnel are focused on controlling air quality

in addition to maintaining sterile hands and surfaces. For instance, airflow parameters

such as direction and rate are optimized to guide airborne contaminants away from the

surgical site. Air changes and filtration are also relied upon for maintaining clean air in

the OR (Chauveaux 2015). Foot traffic is considered a risk factor, since more personnel

causes more general movement in the OR and more door openings, which can allow

external contaminants in. For this reason, OR personnel numbers are often limited to 5

to 6 (Sadrizadeh 2014).

14

Aside from maintaining air quality, many steps are taken to sterilize all surfaces at risk

of contact with the wound area or surgeon. In some institutions, patients are required to

shower using an antiseptic detergent solution both the day before and the morning of

their operation. Prior to the operation, the patient’s skin is cleaned near the surgical site

and antiseptic, often an alcohol-based solution, is applied to the area. The surgical staff

also follow a mandatory hand hygiene protocol that includes handwashing with soap

followed by antiseptic application. Staff don special surgical attire such as gowns,

masks, caps and gloves for protection. Moreover, surgeons practice double-gloving to

lower the risk of perforation; however, outer glove perforation still consistently occurs

after approximately 90 minutes, making frequent glove changing a necessary

precaution (Beldame et al. 2012). During the operation, irrigation is often performed to

cleanse the open wound area. Finally, surgical staff use draping to maintain a sterile

field while operating. Through these thorough precautionary measures, the risk of

contamination from the direct contact of non-sterile surfaces is minimized (Chauveaux

2015).

An additional critical anti-infection strategy that has become a routine part of

orthopaedic surgical practice is antibiotic prophylaxis. This is discussed in more detail

below.

1.2.1.4 Microbiology and Diagnosis

Conclusions regarding the most prevalent pathogenic species responsible for SSIs are

fairly consistent across studies. Most infections in orthopaedic surgeries are caused by

endogenous bacteria commonly found in the skin flora, including Staphylococcus

aureus, coagulase-negative staphylococci and gram-negative bacteria (Debarge et al.

2007, Edmiston et al. 2005). However, as mentioned earlier, surgical staff and materials

may also serve as vectors for exogenous pathogens during operations, most of which

are staphylococcal and streptococcal species (Debarge et al. 2007).

15

Normally, a combination of clinical, radiological and laboratory evidence is relied upon

when diagnosing SSIs. Sinus tracts, purulent drainage, multiple positive intraoperative

cultures and histopathological signs of microorganisms are all listed as definitive

evidence of an infection. On the other hand, signs including acute or chronic pain,

inflammation, hardware loosening, osteolysis, impaired bone healing, a single positive

intraoperative culture and elevated serum inflammatory markers only serve as

suggestive evidence of an infection (Osmon et al. 2013). Moreover, the above criteria

are used as general indicators of infection presence. Classification systems that offer

more specific diagnoses of infection types have been developed, but they are not used

consistently across different hospitals.

One system developed by Trampuz and Zimmerli (2005) categorizes prosthetic joint

infections (PJIs) associated with fracture-fixation devices in a manner that is convenient

for clinical applications. Infections are classified according to associations between the

causing species’ virulence, clinical signs and the timing of infection onset, establishing

three PJI groups:

1. Early infections have an onset of less than 2 weeks and are characterized

by acute local pain, fever, erythema and edema. They are typically caused by

highly virulent pathogens such as Staphylococcus aureus or gram-negative

bacilli.

2. Delayed or low-grade infections appear after 2-10 weeks and are

characterized by chronic or worsening pain, nonunion, implant loosening and

sometimes sinus tract development. They are typically caused by less virulent

pathogens such as coagulase-negative staphylococci.

3. Late infections appear after 10 weeks and are characterized by similar

symptoms as delayed infections. They are also typically caused by less

virulent pathogens (Trampuz and Zimmerli 2005).

The orthopaedic literature consistently reports S. aureus as the most common infecting

pathogen, causing up to 20% of all SSIs (Anderson et al. 2010, Pull ter Gunne and

16

Cohen 2009, Korol et al. 2013). For this reason, a significant amount of research in

orthopaedic surgery has been devoted to investigating the aetiology of S. aureus-

induced SSIs. Epidemiological data on coagulase-negative staphylococcal infections is

less clear. However, more and more studies over the past few decades are

demonstrating that such chronic, low-grade infections have a higher prevalence in

orthopaedic patients than previously thought, are difficult to manage in terms of bone

reconstruction and can directly inhibit bone repair. Therefore, there is an emerging need

for new strategies of managing these recalcitrant infections concurrently with

addressing the need for fracture healing augmentation. To simplify categorization, “low-

grade” will be used to refer to both delayed and late infections moving forward.

17

Figure 1-3: Diagnostic algorithm for periprosthetic joint infection developed by the

Infectious Disease Society of America. ESR, erythrocyte sedimentation rate; CRP, C-

reactive protein. Reproduced with permission from Osmon, D. R., Berbari, E. F.,

Berendt, A. R., Lew, D., Zimmerli, W., Steckelberg, J. M., Rao, N., Hanssen, A., Wilson,

W. R., Infectious Diseases Society of America. 2013. Diagnosis and management of

prosthetic joint infection: Clinical practice guidelines by the Infectious Diseases Society

of America. Clin Infect Dis. 56(1):e1–e25.

18

1.2.1.5 Low-Grade Infections and Nonunions

Low-grade infections present a unique diagnostic concern for orthopaedic surgeons in

that clinical evidence is often insufficient for an accurate infection diagnosis. The

subclinical nature of low-grade infections necessitates the use of more sensitive

detection methods such as laboratory and imaging tests in order to avoid missing the

diagnosis based on absent clinical signs. Intraoperative culturing of tissue biopsies

taken from areas near the surgical site is routinely performed to check for multiple

positive cultures with phenotypically consistent microorganisms, which is deemed

reliable evidence of infection (Osmon et al. 2013). Research into better detection

methods is ongoing, and recent studies of implant-associated infections demonstrated

that culturing sonication fluid represents a new approach to identifying the presence of

low-grade infections. Sonication is highly sensitive in these cases since it can effectively

detach bacterial colonies that adhere to metal hardware in biofilms, which is a common

property of low-virulence infections (Evangelopoulos et al. 2013, Maniar et al. 2016).

However, there is still some controversy over whether tissue culturing or sonication

procedures are more reliable than the other.

By including more sensitive detection protocols in addition to clinical evaluation, recent

studies are revealing a higher prevalence of low-grade infections in nonunion cases

than previously thought. In the aforementioned study on 100 long bone nonunion

patients, 5% of cases with nonunion presumed to be aseptic were identified as infected

with multiple positive culture tests or the presence of an abscess near the implanted

metal hardware. In addition, eight patients received treatment for their infections and

each one had multiple positive culture tests despite lacking clinical evidence of ongoing

infections and having normal levels of serum markers (Mills et al. 2016). According to

one report, such “surprise” positive intraoperative cultures were found in 20% of

nonunion patients undergoing revision surgeries and are linked to lower rates of union,

more recalcitrant infections, and a greater number of subsequent surgeries in

comparison to cases with negative cultures (Olszewski et al. 2016). Low-grade

infections evidently contribute to the pathogenesis of nonunions and add great

19

complexity to their management since multiple stages of surgery are often necessary for

dealing with infection eradication and bone regeneration sequentially.

The first revision surgery usually involves debridement of the wound area, removal of

loose or contaminated hardware, re-establishment of fixation and local and systemic

antibiotic administration specific to the infection present. A period is then required for

adequate antibiotic treatment and to allow infection clearance. If there is sufficient

clinical and serological evidence that the infection has been cured, a subsequent

reconstructive surgery is performed to augment bone healing using a variety of

techniques depending on the type of nonunion present (Nauth et al. 2018). This staged

surgical approach places a significant burden on both the patient and the healthcare

system in terms of financial cost and resources. The separated surgical stages may

also cause an extensive period of physical incapacitation, resulting in the

aforementioned loss of work productivity for a long time. Furthermore, each individual

surgery involves a risk of additional complications, such as more infections.

Experimentation with novel therapies using animal models is invaluable in discovering

more efficient treatment strategies for infected nonunion cases.

1.2.1.6 Animal Model of Low-Grade Infected Nonunion

In the orthopaedic infection literature, there is only a single in vivo study performed by

Lovati and colleagues (2016) that modelled nonunion characterized by subclinical S.

epidermidis infection. Sub-critical defects were created and stabilized with internal plate

fixation in rats subjected to three different doses of S. epidermidis inoculum: 103, 105,

and 108 colony-forming units (CFUs). At 56 days post-surgery, animals were sacrificed

and evaluated for bone healing and infection status using micro-CT, microbiological and

histological analyses. They discovered a dose-dependent effect of inoculum dose on

bone healing outcomes with a higher nonunion rate at each increasing dose. Moreover,

rats inoculated with 103 CFU lacked clinical signs of infection and exhibited variable

responses in terms of both bone healing and infection clearance. In three out of five

rats, bacteria were found in cultured samples and micro-CT and histology revealed mild

20

osteolysis and impeded bone healing. The remaining two rats showed no bacteria in

their samples and had imaging results comparable to the control group, suggesting that

their infections may have been naturally eradicated by their own immune system

responses. On the other hand, the 105 and 108 CFU doses led to more severe signs of

osteomyelitis, greater neutrophil counts in post-operative blood samples and higher

nonunion rates compared to the low dose and control groups. Therefore, this study’s

results demonstrated that 103 CFU is within the range of S. epidermidis doses that

establishes a low-grade infection in rats, whereas 105 CFU and above leads to more

obvious clinical infections. In addition, the spontaneous infection clearance observed at

the low dose is an important consideration when designing future experiments that

investigate low-grade infected nonunions in animal models (Lovati et al. 2016).

1.2.2 Antibiotic Therapies and Biofilm Infections

1.2.2.1 Antibiotic Prophylaxis in Orthopaedic Surgery

Since the first clinical trials investigating prophylactic antibiotics in the late 20th century,

there has been a plethora of research in support of antibiotic prophylaxis as a

preventative tactic against SSIs (Lidwell et al. 1984, Boxma et al. 1996). For instance, a

systematic review conducted a meta-analysis on 7 studies with patients receiving total

hip and knee replacements and demonstrated an overall 81% reduction in relative risk

and 8% reduction in absolute risk with the use of prophylactic antibiotics compared to

no prophylaxis (Albulhairan et al. 2008). The current high level of evidence supporting

perioperative antibiotic prophylaxis has led to its widespread adoption in guidelines for

standard surgical procedures (Bratzler and Hunt 2006, Sewell et al. 2011, Evan et al.

2011). However, studies indicate that not all surgeons consistently follow the guidelines

and that some controversy still exists over certain aspects of prophylaxis including the

best timing of antibiotic administration, appropriate antibiotic selection, number of doses

and duration of prophylaxis.

21

Despite the heterogeneity in the literature, the general consensus on the best timing of

administration is between 30 and 60 minutes prior to skin incision, while a preoperative

window greater than 60 minutes leads to higher rates of SSIs (Hansen et al. 2014,

Thonse et al. 2004). This period is appropriate for the short half-lives of commonly used

antibiotics and allows them sufficient time to achieve the minimum inhibitory

concentration for the infection present at the surgical site (Thonse et al. 2004,

Andersson et al. 2012). There are some exceptions where antibiotics may be

administered up to 2 hours before incision, such as for vancomycin, but administration

longer than 2 hours beforehand leads to a substantially higher risk of SSI (Hansen et al.

2014, Burke 2001).

Antibiotics are selected for recommendation based on their efficacy against common

infecting pathogens, ease of administration, low cost, and lack of toxic effects.

Cephalosporins like cefazolin and cefuroxime are strongly advised for prophylactic use

because of their strength against most S. aureus strains and gram-negative bacilli.18

Accordingly, cefazolin is routinely administered as a first-line systemic antibiotic in

multiple doses during orthopaedic procedures, as illustrated by the high

correspondence rate of surgeons (96%) in a study on surgical fixation of closed long

bone fractures (Hansen et al. 2014, Gans et al. 2017). Cefuroxime is another

cephalosporin that is popular in orthopaedic surgery, particularly in patients undergoing

hip and knee arthroplasty (Hansen et al. 2014, Salkind and Rao 2011, Tornero et al.

2015). With the increasing prevalence of methicillin-resistant S. aureus (MRSA)

infections, vancomycin has also gained popularity. Vancomycin demonstrates a high

antibacterial efficacy against MRSA and coagulase-negative staphylococcal infections,

making it a good alternative drug in these cases in addition to cases of allergies to beta-

lactam antibiotics (Salkind ad Rao 2011). However, there is insufficient evidence to

suggest that vancomycin is safe to use as a routine prophylactic therapy given its

relative lack of efficacy against methicillin-sensitive staphylococci and the growing

number of vancomycin-resistant bacterial strains, in addition to concerns over potential

harmful side effects (Evan et al. 2011, Hansen et al. 2014).

22

Some guidelines exist for the duration of prophylaxis and antibiotic dosing regimens, but

these practices are still highly variable amongst surgeons. The American Academy of

Orthopaedic Surgeons (AAOS) recommends that prophylaxis duration is limited to 24

hours post-surgery due to its lack of proven health benefits, high costs, risk of

generating antibiotic-resistant strains and potential for toxic effects (Bryson et al. 2016).

This tends to be followed in practice; however, aside from the initial preoperative dose

given within 1 hour of incision, the timing of subsequent doses throughout the 24-hour

period varies from surgeon to surgeon (Dhammi et al. 2015, Bryson et al. 2016).

Moreover, evidence from a systematic review suggests that there is no difference in the

risk of SSI development between patients receiving single versus multiple doses

(McDonald et al. 1998). The American Society of Health System Pharmacists (ASHP)

also recommends administration of the minimum dose that covers the duration from

incision to wound closure, which can typically be accomplished with a single dose. In

spite of these evidence-based guidelines, the majority of surgeons practice multi-dose

prophylactic regimens, reflecting the lack of standardization of prophylactic protocols

(Dhammi et al. 2015, Bryson et al. 2016).

1.2.2.2 Local Antibiotics in Perioperative Prophylaxis

In recent decades, there has been a growing interest in the local application of

antibiotics across surgical specialties. In comparison with systemic antibiotics, local

antibiotics can achieve high doses targeted to the surgical site while keeping systemic

levels minimized. Not only could this theoretically enhance SSI prevention, but it also

could keep local concentrations above the minimum inhibitory concentrations of

infective pathogens for longer periods, effectively reducing the need for additional post-

operative doses (Fleischman and Austin 2017). Thus, local antibiotic delivery may have

applications in different fields and has been investigated in orthopaedic, dermatologic,

cardiothoracic, colorectal and abdominal surgery. A variety of delivery methods have

also been explored, including irrigation solutions, bone graft, bone cement, powders,

beads, ointments, pastes, sponges and fleeces. However, high quality clinical trials are

23

generally lacking, limiting the strength of conclusions that can be drawn regarding local

antibiotic applications (Huiras et al. 2012, Fernicola et al. 2020).

Despite the dearth of strong evidence, local intra-wound administration of powdered

vancomycin has become routine amongst orthopaedic spine surgeons. This trending

practice is likely based on the promising retrospective data seen largely in the spine

literature but also in individual studies scattered across other orthopaedic subspecialties

including total joint arthroplasty, trauma, foot and ankle, and elbow (Fleischman and

Austin 2017). In most of these studies, patients who received topical vancomycin

treatment had significantly lower rates of SSI development compared to those who did

not. However, the single randomized controlled trial (RCT) that investigated local

vancomycin therapy found that it did not exhibit any significant additional antibacterial

effect, contradicting the retrospective data. Experimental designs were also inconsistent

across studies and poor in some cases, limiting the quality of their comparative

analysis.20 Similar to other fields, further investigation with better experimental designs

(i.e. RCTs) must be done before drawing definitive conclusions regarding the safety and

efficacy of local antibiotic prophylaxis in orthopaedic surgery (Fernicola et al. 2020).

1.2.2.3 Biofilm Infections and Antibiotics

Less virulent pathogens in chronic and implant-associated orthopaedic infections often

have the capacity to form biofilms. These structures act as defensive barriers against

antibacterial agents and the host’s immune system, making the treatment of biofilm

infections more complicated than those involving only planktonic bacterial forms

(Stoodley et al. 2011, Zimmerli and Moser 2012, Arciola et al. 2018). Studies suggest

that biofilms often form on implant surfaces because they provide opportune

environments for adhesive bacterial colonization, resulting in a 100-fold increased

likelihood of infection development (Lidwell et al. 1983, Gristina 1994). This poses a

major obstacle for orthopaedic surgeons since metal hardware like locking plates and

intramedullary nails are commonly applied to provide internal fixation. Thus, research

investigating biofilm pathogenesis is important for offering insight into the mechanisms

24

of their development and the types of methods that can be used for prevention and

eradication.

Biofilms are large and organized microbial structures consisting of extracellular

polymeric matrices encasing an abundance of bacterial colonies, often from multiple

species (Costerton et al. 1995). Physiologically distinct colonies arise from variable

regional growth conditions due to differences in nutrient availability throughout the

biofilm. Oxygen and nutrient availability is particularly low in the biofilm’s interior, forcing

microorganisms to survive by adopting metabolically inactive states (Costerton et al.

1995, Werner et al. 2004). These dormant cells are referred to as “persisters” and

exhibit an exceptionally high antibiotic tolerance (Lewis 2010) Furthermore, the biofilm’s

matrix structure acts as a penetration barrier for antibacterial agents such as antibiotics

and immune cells (Zimmerli and Moser 2012, Arciola et al. 2018, Winkler 2017). These

properties play a significant role in biofilm infections’ resistance to immune responses

and common antibiotic therapies. As such, large efforts in research have been

dedicated to discovering new therapeutic options that effectively manage biofilm-

producing bacterial strains.

25

Figure 1-4: Results of scanning electron microscopy analysis for biofilm-producing

staphylococcal strains and S. epidermidis ATCC 35984. After a 24-h incubation period,

a three-dimensional biofilm structure was formed, and bacterial clusters appeared to be

coated on all sides with a gelatinous material as indicated by the red arrows (C). After a

72-h incubation period, mature three-dimensional biofilm structures were formed. Water

channels (green arrows) and thread-like appendages (blue arrows) between subunits of

bacterial colonies were distinctly observed (D). Reproduced with permission from Hou,

W., Sun, X., Wang, Z. and Zhang, Y. 2012. Biofilm-forming capacity of Staphylococcus

epidermidis, Staphylococcus aureus, and Pseudomonas aeruginosa from ocular

infections. Microbiol Immunol. 53(9): 5624-5631.

26

1.2.2.4 Local Antibiotic Therapies and Low-Grade Infected Nonunions

As discussed earlier, there is a need for more efficient and effective strategies for

dealing with low-grade infected nonunions. In revision surgeries, local antibiotic therapy

represents a potentially new method of targeting high doses of antibiotics at the surgical

site for infection treatment. However, the high prevalence of biofilm-producing

staphylococcal infections in orthopaedic patients must be considered when selecting

appropriate local antibiotics. Such chronic infections often involve mature biofilms that

are much more difficult to eradicate than those that are early on in their development.

Certain common antibiotics may prevent biofilm production but are much less effective

at disrupting them once established. For instance, both in vitro and in vivo preclinical

studies consistently show that vancomycin reliably prevents common SSIs if

administered early on after contamination, including those caused by biofilm-producing

staphylococcal strains (Chilukuri and Shah 2005, Ribiero et al. 2012, Pagano et al.

2004, Howlin et al. 2015). This coincides with the promising retrospective evidence in

the spine surgery literature. However, vancomycin loses its efficacy when applied either

against established staphylococcal biofilms or in a delayed fashion after the

contamination of implants (Chilukuri and Shah 2005, Ribiero et al. 2012, Pagano et al.

2004, Howlin et al. 2015, Darouiche and Hamill 1994, Park et al. 2017). This was further

demonstrated in an in vivo animal study where local vancomycin was administered as a

powder or in impregnated beads either acutely at 6-h or after biofilm formation at 24-h.

The study reported that although both forms of treatment were effective acutely, their

delayed treatment led to a significantly reduced antimicrobial therapeutic effect

(Tennent et al. 2016). Thus, despite its frequent application as an adjunct therapy for

prophylaxis, local vancomycin may not the best option for managing chronic infections

that likely have an established biofilm in the setting of nonunion.

Nonetheless, vancomycin still has potential applications in combination antibiotic

therapies, especially with rifampin. A collection of in vitro, in vivo and clinical evidence

has demonstrated the high efficacy of rifampin against staphylococcal biofilm

eradication. However, the Infectious Disease Society of America recommends against

rifampin usage as a monotherapy due to a high risk of causing spontaneous

27

antimicrobial resistance (Liu et al. 2011). Fortunately, rifampin-based combined

antibiotic therapies have also been successful and even more effective in some cases,

particularly when including vancomycin (Claessens et al. 2015, Saginur et al. 2006,

Thompson et al. 2017, Brinkman et al. 2017, Niska et al. 2013). Most studies

investigating this regimen applied rifampin as a systemic antibiotic. The one animal

experiment testing topical rifampin powder showed that it had a high antibacterial effect

against S. aureus biofilm infection, and that its combination with local vancomycin

therapy led to similar results. On the other hand, local vancomycin monotherapy

showed a substantially lower antibacterial efficacy than both of these treatments (Shiels

et al. 2018). This suggests a potential role for local rifampin powder with vancomycin in

eradicating staphylococcal biofilm infections, but more preclinical studies are necessary.

Furthermore, this has not yet been tested in a S. epidermidis infected animal model,

which would offer insight into its therapeutic potential for low-grade infected nonunion

patients.

1.3 Endothelial Progenitor Cells (EPCs)

1.3.1 Clinical Relevance

The term EPC is typically used to describe cells capable of differentiating into mature

endothelial cells and forming new blood vessels. They were first successfully isolated by

Asahara and colleagues in 1997, generating great excitement over their angiogenic

applications in regenerative medicine and leading to a heightened global interest in

EPC-related research over the past two decades (Asahara et al. 1997). Studies have

clearly demonstrated their ability to enter into the circulation from the bone marrow,

home to tissue sites and generate a new vascular network in response to ischemic

signals (Lee et al. 2019, Arici et al. 2015). So far, clinical trials have reported successful

outcomes for EPC-based therapies in patients with coronary artery disease, peripheral

artery disease and ischemic stroke (Lara-Hernandez et al. 2010, Zhu et al. 2016, Fang

28

et al. 2018). Such studies have demonstrated that EPCs’ angiogenic properties

displayed in both in vitro and in vivo animal studies can be translated to human patients

for a variety of pathologies. Despite these positive results and the apparent safety of

EPC-based cell therapies, several factors continue to prevent their widespread adoption

into clinical practice. Isolation and characterization techniques vary from study to study,

creating ambiguity over the EPC’s identity and making it difficult to generate valid

conclusions. This confusion is partly due to the disagreement that still exists over

functional definitions of EPCs, causing inconsistent usage of the term amongst authors.

In addition, current isolation techniques have low yields and cell passaging can change

cell identities and abolish their multipotency (Medina et al. 2017, Chopra et al. 2018).

These issues must be addressed through further preclinical work before EPCs can be

used in widespread clinical testing and applications.

1.3.2 Classification

Despite the ongoing debate over EPC definitions, significant progress has been made in

characterizing EPC subpopulations through examining cell surface markers, functions,

sources and protein expression profiles. EPCs have been broadly categorized into two

groups: early EPCs (E-EPCs) and late EPCs (l-EPCs). E-EPCs are derived from a

shorter culture period of seven to ten days, display a spindle-shaped appearance and

demonstrate monocytic features. In contrast, l-EPCs appear in culture after two to four

weeks with a cobblestone-like morphology and have an endothelial phenotype (Banno

and Yoder 2019, Medina et al. 2017, Chopra et al. 2018).139,141,142 Despite this historical

use of terminology, there has been a recent shift towards more specific functional

definitions of EPCs. E-EPCs are now more commonly called myeloid angiogenic cells

(MACs) to describe their monocytic lineage and paracrine-mediated angiogenic effects.

They have also been called circulating angiogenic cells (CACs), but recent literature

recommends against this term due to a lack of evidence of their “circulating” status.

Furthermore, l-EPCs are called endothelial colony-forming cells (ECFCs) to reflect their

committed endothelial lineage and direct angiogenic effects. ECFCs are coined “true

EPCs” because of their bona fide EPC characteristics including their endothelial

29

phenotype, in vitro clonogenic potential and ability to induce neovascularization in vivo

(Medina et al. 2017, Chopra et al. 2018, Chong et al. 2016). Unlike MACs, they have

not been investigated as therapeutic agents in clinical trials but ongoing research is

focused on improving ECFC identification and preparation approaches for clinical

translation (Medina et al. 2017).

1.3.3 Recent Discoveries of EPC-based Therapies in Fracture Healing

A series of preclinical studies in the past decade has provided early evidence of EPC-

based therapy’s potential fracture healing applications. For clarification, the cells used

most closely resembled MACs due to their shorter culture periods, but they will be

referred to simply as EPCs from here onwards to remain consistent with the authors’

terminology. The first several studies demonstrated that EPC-based therapy applied

acutely can facilitate the healing of segmental bone defects that would otherwise

progress to nonunion (Li et al. 2014, Atesok et al. 2010, Seebach et al. 2012).

Heightened expressions BMP-2 and VEGF were also observed, which may play a role

in the angiogenic and osteogenic effects of EPCs (Li et al. 2009, Li et al. 2014). This

served as early evidence of EPC-based therapy’s promise for clinical translation.

However, the acute timing of EPC administration was thought to be an inaccurate

representation of real clinical scenarios, since bone graft or substitutes are usually given

after the initial inflammatory period in order to avoid complications. Levels of

inflammatory cytokines like IL-1 and IL-6 are also elevated early on and may positively

affect EPC-induced bone healing outcomes. Thus, in vivo experimentation with delayed

EPC administration was considered a necessary next step towards clinical translation.

Bates and colleagues (2017) inserted EPCs into the defect of rats at 3 weeks post-

creation and found that the healing response to delayed treatment did not differ from

that of acute treatment, further supporting the notion that EPCs could be a novel

effective treatment option for nonunions and large segmental bone defects in various

scenarios. Further investigation of EPC-based therapy in the presence of nonunion-

associated risk factors may offer insight into the overall efficiency of EPCs for nonunion

treatment.

30

Figure 1-5: Radiographic union was improved when EPCs were applied acutely and in

delayed fashion. Reproduced with permission from Bates, B. D., Godbout, C.,

Ramnaraign, D. J., Schemitsch, E. H. and Nauth, A. Delayed Endothelial Progenitor

Cell Therapy Promotes Bone Defect Repair in a Clinically Relevant Rat Model. Stem

Cells International. 2017, Article ID 7923826, 10 pages, 2017.

31

1.3.4 EPCs and Infection

Given that infection is a frequent contributing factor to nonunion, there is a theoretical

rationale behind EPCs’ possible dual efficacy as both an antimicrobial and fracture

healing therapy. As noted above, EPCs have well-documented pro-angiogenic

capabilities that facilitate neovascularization and allow their long-term successful

integration of newly engineered tissues. These vascularized networks enable oxygen

and nutrient delivery to the tissue and provide local immunosurveillance. This potentially

creates a more immunoprotected environment at the fracture site, helping prevent

infection establishment and development. Moreover, the types of EPCs investigated in

cell therapies exhibit an immune phenotype and share similar expression profiles with

monocyte-derived-macrophages, further supporting the notion that they may mediate

immune responses (Cheng et al. 2013, Medina et al. 2017). For instance, one study

reported that early EPCs secreted much higher levels of CCL5 in comparison to adult

blood late EPCs and cord blood late EPCs (Zhang et al. 2009). Since CCL5 facilitates

leukocyte migration to inflammatory sites, this is one possible mechanism by which

EPCs may contribute to fighting against infection by foreign bodies. Thus, it is worth

investigating EPCs’ potential antimicrobial effects in the context of low-grade infected

nonunion, where the presence of a recalcitrant infection acts as an additional obstacle

to successful bone repair. If EPCs could be demonstrated to be effective as both

antimicrobial and bone reconstructive agents, they could represent promising

candidates for a novel, single-stage surgical strategy aimed towards treating infected

nonunions. This could ultimately decrease the amount of time and number of resources

required to manage these cases, benefitting both the patient and the healthcare system.

1.3.5 EPCs and Antibiotics

If EPCs have any clinical applicability in infected nonunion scenarios, it is important to

consider how the presence of antibiotics might influence their effects. Revision

surgeries for infected nonunions typically involve local antibiotic treatment in various

forms, and the heavy concentration of antibiotics at the surgical site may have

detrimental effects on EPC-mediated outcomes. To date, there is a lack of in vitro

32

studies assessing the effects of antibiotics on EPCs, but similar experimentation has

been done on related cell types including osteoblasts, endothelial cells and

mesenchymal stem cells (Edin et al. 1996, Rathbone et al. 2011, Zhang et al. 2014).

This literature reports no detrimental effects of vancomycin on osteoblast proliferation or

viability below 1000 µg/mL, which is far beyond the equivalent clinical dosage (Edin et

al. 1996, Rathbone et al. 2011). Similar outcomes were observed in an in vivo rat study,

where there was no difference in spinal fusion rates or levels of bone formation between

vancomycin-treated and control group rats (Mendoza et al. 2016). On the other hand, in

vitro work revealed rifampin’s more potent cytotoxic effects, with decreased osteogenic

activity at concentrations within the realm of clinical dosages (Rathbone et al. 2011,

Zhang et al. 2014, Isefuku et al. 2001). This was in contrast to outcomes in two animal

studies, where rifampin positively impacted bone regeneration in sterile and

contaminated scenarios. The first study reported greater bone volumes in the calvarial

defects of rats administered autogenous graft with rifampin compared to those that

received the graft without antibiotics (Durmuşlar et al. 2016). The second study found

greater bone volumes in the defects of contaminated rats treated with topical rifampin

relative to controls, as determined by radiology, histology and micro-computed

tomography (Shiels et al. 2018). Thus, both antibiotics appear to be safe for use at least

in the context of bone healing, but in vitro evidence indicates that rifampin may have a

negative impact on the viability and function of EPCs at the defect site. Investigation

using an in vivo model is needed to provide further insight into the feasibility of this dual

treatment in clinical practice.

33

Chapter 2: Rationale, Aims, and Hypothesis

Fracture healing is a complex process that depends on the timely sequence of multiple

biological and mechanical factors. A disruption to any of these processes can disrupt

normal bone healing and may cause catastrophic consequences, such as fracture

nonunion. Infection has been identified as a common risk factor for nonunion, and the

management of infected nonunions is a complex process typically involving multiple

surgical stages aimed at sequentially eradicating infection followed by promoting bone

healing. This process is resource-intensive and time-consuming, and creates a

substantial burden to both the healthcare system and the patients. Amongst the various

types of infection seen in orthopaedic surgery, there is a rising awareness of the

association of low-grade infections with nonunions. These infections are difficult to

detect and are particularly resistant to eradication. For such cases, there is an

increasing need for new strategies that tackle both issues of infection eradication and

fracture healing simultaneously, in order to treat infected nonunion cases more

efficiently.

Colleagues recently demonstrated that EPCs can effectively heal chronic nonhealing

bone defects in rats when administered locally in a delayed fashion (Bates et al. 2017).

These results suggest that EPCs are a promising therapeutic option for patients

suffering from fracture nonunion, although refinement of EPC identification and handling

methods is a necessary step before clinical translation becomes more realistic.

Furthermore, there is reason to believe that EPC-based therapy could have

antimicrobial effects as well given EPCs’ immune phenotype and capacity to form new

vascular networks that potentially enhance the tissue’s local immunosurveillance.

Should EPC-based therapy prove to be antibacterial, it may serve as a promising

component of a single-stage surgical strategy for managing infected fracture nonunions.

In addition, since local antibiotic use is gaining popularity as a surgical technique for

both infection treatment and prevention, it is worth evaluating how it may modify EPC-

mediated bone healing and infection outcomes.

34

Thus, our study aimed to investigate the efficacy of EPC-based therapy for both fracture

healing and infection treatment in the context of local antibiotics in a low-grade infected,

critical-size bone defect model. We hypothesized that:

1. EPC-based therapy would promote both bone healing and infection eradication.

2. Local antibiotics administered in addition to EPCs would further facilitate these

outcomes.

The original project was designed to test these research questions within a single,

comprehensive experiment. Due to technical shortcomings with EPC handling and

sterility techniques, the project had to be modified and carried out as a series of

experiments instead. They are outlined below:

1. Experiment #1: Pilot Study The aim of this study was to identify the optimal inoculation dose of S.

epidermidis to establish a reliable model of low-grade infection, as

demonstrated by consistently positive microbiological cultures and absent

signs or symptoms of severe infection (significant pain, purulence, hardware

failure, or sepsis). In addition, we sought to investigate the effects of different

local antibiotics to identify appropriate candidates for local antibiotic therapy in

subsequent experiments.

2. Experiment #2: EPCs, Antibiotics and Bone Healing in a Non-contaminated Critical-Size Defect Model In this study, we aimed to assess the impact of local antibiotic therapy on

EPC-mediated bone healing outcomes using radiography. Animals were not

inoculated with bacteria in order to isolate the effects of the local antibiotics

on the cell therapy in a sterile model. Unfortunately, accidental

contaminations occurred and introduced the presence of infection in some

animals.

35

3. Experiment #3: Acute and Delayed EPC Treatment and Bone Healing This experiment’s purpose was to troubleshoot the EPC functionality issues

we encountered in experiment #2 by refining EPC isolation and handling

techniques and assessing bone healing efficacy in both acute and delayed

EPC treatment contexts.

4. Experiment #4: EPCs, Local Antibiotics and Infection Outcomes in a Contaminated Critical-Size Defect Model We conducted this study to examine the potential antimicrobial properties of

EPCs both in the presence and absence of local antibiotic therapy. Due to

timeline constraints, we chose to analyze infection status as the primary

outcome measure instead of bone healing (which would require a much

longer timeframe and significantly more testing to assess).

36

Chapter 3: Experiment #1 – Pilot Study

3.1 Rationale and Aims Our pilot study served two important purposes. First, we aimed to establish a consistent

animal model of S. epidermidis-induced low-grade infection in a segmental bone defect

at 2 weeks post-inoculation. This period would potentially allow a sufficient timeframe

for biofilm development, which is an important characteristic in modeling infected

nonunions receiving delayed treatment. Moreover, we selected our dosing range based

on the results from a previous study that investigated a similar model (Lovati et al.

2016). Second, we aimed to discover an appropriate local antibiotic therapy for future

investigations using this infection model. Based on the in vitro and in vivo evidence of

rifampin’s efficacy, we decided to trial a combined local antibiotic therapy with

vancomycin and rifampin that abides by the Infectious Disease Society of America’s

guidelines and includes the current most commonly used local antibiotic in orthopaedic

surgery.

3.2 Methods

3.2.1 Experimental Design

Rats were randomly allocated to receive inoculation with one of four doses of S.

epidermidis (ATCC#35984/RP62A): 102 (n = 2), 103 (n = 2), 104 (n = 3) or 105 (n = 3)

CFUs. This dosing range was chosen to target a S. epidermidis-induced low-grade

infection based on results from a previous study that investigated a similar animal model

(Lovati et al. 2016). Each rat underwent an initial fracture surgery to create a 5-mm

critical-size defect in the right femur and establish internal fixation using a mini-plate and

37

screws. Gelfoam scaffolds were prepared with 100 uL of inoculum and placed into the

defect site. Two weeks later, a second surgery was performed to collect muscle tissue

biopsies, debride the defect site, remove newly-grown bone, clear the medullary canal

and add one of four randomized antibiotic treatments to the defect area on a new

scaffold: no treatment (n = 1), vancomycin (n = 3), rifampin (n = 2) or vancomycin +

rifampin (n = 4). Animals were sacrificed 2 weeks after the second surgery. Muscle

tissue samples and the metal hardware were collected upon sacrifice. Cultures of

homogenized tissue samples and sonication fluid post-second surgery and post-

sacrifice were used for microbiological analysis, and radiographs were taken at the

same timepoints to monitor bone healing progress.

3.2.2 Bacteria and Gelfoam Scaffold Preparation

The following protocol was established based on trials that consistently produced the

targeted bacterial dose. Stocks of S. epidermidis were stored in a -80°C freezer on

porous beads. Three mL of tryptic soy broth (TSB) was inoculated with a single bead in

a 10 mL snap cap falcon tube and placed in a shaker at 37°C and 250 rpm for a 17-hour

incubation period. The bacterial growth suspension was then transferred to a 15 mL

falcon tube and centrifuged at 8000 rpm for 10 minutes at 4°C. After discarding the

supernatant, the pellet was resuspended in 1 mL phosphate buffer solution (PBS).

Serial dilutions were then prepared in PBS and TSB from this new suspension to make

solutions with approximately 106, 105, 104, 103 and 102 CFU concentrations. Gelfoam

scaffolds (5 x 5 x 5 mm) prepared under sterile conditions in 1.5 mL microcentrifuge

tubes were inoculated with 50 uL of one of the serial dilutions (chosen based on the

targeted dose) and incubated at 37°C for 30 minutes. These scaffolds were then

transferred into new sterile 1.5 mL tubes with 50 uL TSB and stored on ice until

application during surgery on the same day. For the control rat, the scaffold was not

inoculated and left sterile. Bacterial concentrations were confirmed by spreading 100 uL

of each serially diluted suspension on a tryptic soy agar (TSA) plate, incubating

overnight at 37°C and counting colonies on the following day.

38

3.2.3 Surgical Procedures

All animal protocols in this study were approved by the St. Michael’s Hospital Animal

Care Committee. Our critical-size femoral defect rat model has been previously

described (Bates et al. 2017). Briefly, 10 male Fischer 344 inbred rats (250-300g)

underwent an initial femur fracture surgery to create a critical-size defect followed by a

debridement and treatment surgery two weeks (13-15 days) later. They were

anesthetized with 2.5-3.0% isofluorane via inhalation and 0.05 mg/kg buprenorphine

analgesic via subcutaneous injection. Their right legs were shaved and alcohol and

betadine were applied topically to disinfect the skin around the area of incision. The

entire right femur was exposed via a lateral incision to the right leg, and an oscillating

saw was used to create two osteotomies under irrigation with 0.9% sterile saline

solution. This was done to create a 5-mm segmental bone defect in the middle third of

the femur. Internal fixation was established using a mini-plate secured to the bone ends

with two distal and two proximal 1.5-mm self-tapping cortex screws. A contaminated or

non-contaminated gelfoam scaffold was placed in the defect site, based on group

allocation. Afterwards, the wound area was carefully closed via both intramuscular and

intracutaneous suturing (vicryl 5-0), and it was cleaned by topical application of

hydrogen peroxide. Rats were allowed unrestricted weight bearing activity in their

cages. Second stage surgeries were performed on each rat two weeks (13-15 days)

later for biopsy sampling, wound debridement and antibiotic treatment. The same

protocols used in the fracture surgery were followed to anesthetize the rats and prepare

their right legs for incision. After exposing the right femur, three muscle tissue biopsies

(15-25 mg) were excised from near the defect area (#1), near the proximal end of the

plate (#2) and near the distal end of the plate (#3). They were weighed using a precision

scale and placed in sterile 1.5 mL microcentrifuge tubes on ice for tissue

homogenization post-surgery. Newly grown bone was removed and the medullary canal

was cleaned using a 25G needle. The defect area was then rinsed with sterile PBS

before adding a sterile gelfoam scaffold followed by topical application of vancomycin

(25 mg/kg), rifampin (25 mg/kg), vancomycin and rifampin (both 25 mg/kg) or no

antibiotic at the defect site and along the metal plate. Before spreading, 20 mL sterile

saline was added to the vancomycin to facilitate dissolution, whereas the rifampin was

39

applied as a dry powder. The wound area was then closed and cleaned in a similar

fashion to the fracture surgery.

3.2.4 Euthanasia and Tissue Sample Harvest

Two weeks after the second surgery, each rat was anaesthetized with 5.0% isofluorane

via inhalation and sacrificed via cervical dislocation. The right leg was shaved and

cleaned with alcohol and betadine before incision. Muscle tissue samples (15-25 mg)

were taken from the following four sites around the right femur and placed into sterile

1.5 mL centrifuge tubes: above the metal plate near the defect site (#1), within the

defect site (#2), near the proximal end of the plate (#3) and near the distal end of the

plate (#4). In addition, the plate and screws were removed from each right femur and

placed in a sterile 1.5 mL centrifuge tube with sterile PBS for sonication later on. The

sonication fluid was cultured as an additional sample (#5). The muscle tissue samples

were kept on ice for subsequent microbiological analysis. The right femur was removed

from each rat, wrapped in gauze soaked with PBS and placed in a 50 mL falcon tube to

be stored in a -80°C freezer.

3.2.5 Tissue Culture, Sonication and Microbiological Analysis

Muscle biopsies and tissue samples collected during the second stage surgery and

during harvest, respectively, were transported on ice to a biosafety cabinet that was

cleaned with 70% ethanol before use. Sterile PBS was added to the tubes at a ratio of 5

µL PBS/mg tissue, and tubes with plates and screws were filled with sterile PBS.

Tissues were then ground within their respective tubes using a pestle to achieve as

much homogenization in solution as possible. The tubes containing the metal hardware

in PBS underwent sonication at 35 kHz for 3 minutes. Ten µL of each solution

(homogenized tissue solution and sonication fluid) plus two serial dilutions were

dispensed on TSA plates in linear tracks and incubated overnight at 37°C. The total

number of colonies within each track was counted on the next day. Cultures were

40

considered as positive for infection if 1 or more colonies were present in the undiluted

10 uL track. Rats were considered as positive for infection at the second stage surgery

if at least two of the three biopsies had positive culture results, and the same two-

sample threshold was applied to the samples taken during the harvest.

41

3.3 Results

Rat ID number

S.

epidermidis

dose (CFU)

Antibiotic

therapy

Second surgery

culture outcomes

(CFU for biopsies in

order: 1, 2, 3)

Harvest culture

outcomes

(CFU for

samples in

order: 1, 2, 3,

4, 5)

154848 102 Vancomycin

+ Rifampin 5, 10, 54 0**

154851 102 Vancomycin 0, 0, 7 10, 39, 19, 5,

TNTC

154852 103 Vancomycin 7, 8, 17 N/A, 47, 30,

10, TNTC


+ Rifampin 4, 1, 1 0**


+ Rifampin 1, 2, 2 0**

155320 104 Rifampin TNTC** 0**

155441 104 None TNTC, 12, TNTC N/A*

155440 105 Rifampin TNTC** 0, 2, 0, 3, 0

155442 105 Vancomycin 21, 3, 20 14, 27, 2, 12, 1


+ Rifampin 18, 3, 44 4, 3, 1, 2, 2

Table 3-1: Tissue culture results for different S. epidermidis doses and antibiotic

regimens. Abbreviations: TNTC, too numerous to count. N/A, not available.

*Rat died during second surgery.

**All samples shared the same result.

Green: positive, red: negative.

42

The lowest dose (102 CFU) was the least reliable in establishing an infection, with one

of the two rats testing negative in the second surgery. All three cultures were positive in

each rat inoculated with the doses 103, 104 and 105 CFU, indicating that they were able

to consistently establish an infection at two weeks post-inoculation. However, clear

differences in infection severity were observed between these groups. Cultures from

two 104 CFU-inoculated rats and one 105 CFU-inoculated rat had colonies that were

TNTC. In contrast, the two 103 CFU-treated rats each had three positive cultures, but

neither had cultures with TNTC colonies. This represented a lower grade of infection

relative to those established with the higher doses. For this reason, 103 CFU was

selected as the optimal S. epidermidis inoculation dose to model a low-grade infection

in our future experiments.

Regarding the harvest culture outcomes, rats treated with rifampin either as a

monotherapy or dual therapy with vancomycin had far lower colony counts than

vancomycin alone. This is clearly illustrated by the rifampin or vancomycin and rifampin

groups’ counts, where most cultures have zero colonies and others have fewer than

five. On the other hand, rats that received vancomycin alone typically had positive

cultures with more than ten colonies and, in some cases, TNTC colonies. Therefore, the

combined vancomycin and rifampin treatment demonstrated a high bacterial eradication

efficacy comparable to that of rifampin monotherapy across different inoculation doses.

Based on this, the vancomycin and rifampin combination therapy was deemed suitable

for experimental use against low-grade infections in order to effectively eradicate S.

epidermidis while abiding by the Infectious Disease Society of America’s guidelines (Liu

et al. 2011).

Statistical analyses were not performed on the pilot study’s results since its raw data

offered sufficient information to guide the subsequent experimental designs. It should

also be noted that the single control animal receiving no antibiotic treatment died during

its second stage surgery. However, its results were considered unnecessary for drawing

conclusions about rifampin’s antibacterial efficacy relative to vancomycin.

43

Chapter 4: Experiment #2 - EPCs, Antibiotics, and Bone Healing in a Non-contaminated Critical-Size Defect Model

4.1 Rationale and Aims

The purpose of this experiment was to examine the potential impacts of local antibiotic

therapy on EPC-mediated bone healing in a segmental bone defect. Given rifampin’s

reported detrimental effects on osteogenic cell viability and function, we were concerned

that its administration in conjunction with EPC-based therapy would impede the positive

bone healing effects of EPCs demonstrated in previous experiments (Lovati et al. 2016,

Li et al. 2011). Thus, we designed our treatment groups such that we could observe any

additional effects that local rifampin contributed to EPC-mediated bone healing when

included in a combined antibiotic therapy with local vancomycin. Finally, we decided

that a non-contaminated model was appropriate for isolating the impact of local

antibiotics on EPC function in the absence of infection. Our goal was to eventually

perform a similar experiment in an infected model to examine how the presence of

infection may additionally modify bone healing outcomes, but subsequent experimental

plans were revised due to unforeseen technical issues.

4.2 Methods


Rats were randomly allocated to receive one of the following five treatments: 1. no

treatment (n=9), 2. local vancomycin and rifampin (n=7), 3. EPCs (n=7), 4. EPCs and

local vancomycin (n=8), or 5. EPCs and local vancomycin and rifampin (n=10). Table 4-1 summarizes the group numbers. Similar to the pilot study, each rat underwent an

initial fracture surgery to establish a 5-mm critical-size defect in the right femur and

44

internal fixation using a mini-plate and screws. An empty gelfoam scaffold was placed in

the defect site. Two weeks later, a second surgery was done to collect muscle biopsies,

debride the defect site, remove newly-grown bone, clear the medullary canal and

administer one of the five randomized treatments. Tissue biopsies were homogenized

and cultured overnight to confirm asepsis. Animals were sacrificed 10 weeks after the

second surgery to allow sufficient time for full bone healing, which was monitored with

biweekly radiographs. Muscle tissue samples and the metal plate and screws were

collected upon sacrifice for microbiological analysis.

Rat Group Number of Rats Treated

Control 9

Vancomyin + Rifampin 7

EPC 7

EPC + Vancomycin 8

EPC + Vancomycin + Rifampin 10

Table 4-1: Number of rats treated per group.

4.2.2 Cell Isolation, Culture, and Characterization

Cell isolation was performed on syngeneic donor rats eight days before second

surgeries for rats receiving EPC treatment. One donor rat was used per treated rat, and

they were sacrificed at a weight of 250-300g via cervical dislocation. Tibiae and femora

were harvested under sterile conditions and transported in sterile PBS to the biosafety

cabinet. A rongeur was used to remove the metaphysis in each bone, and the bone

marrow was flushed with room temperature PBS into a 50 mL falcon tube until the bone

looked clear. The flushed solution was centrifuged at 360g for 10 minutes to pellet the

cells. During centrifugation, fibronectin was aspirated from pre-coated flasks and two

subsequent gentle washes were performed on each flask with 4 mL sterile PBS. After

centrifugation, the supernatant was aspirated and the pellet was resuspended in 25 mL

45

cell culture media (EGM-2MV) and mixed via pipetting. Cell media consisted of EGM-

2MV SingleQuotsTM (Lonza) added to endothelial basal medium. After allowing larger

debris to settle at the bottom of the falcon tube over 15-20 seconds, 12.5 mL of media

was transferred into a prepared flask. The media was gently mixed again in the flask.

The remaining cells and media in the tube were mixed and large debris was allowed to

settle before transferring the media into a second flask and mixing. Two flasks were

used per donor rat. The flasks were then placed inside the incubator and left for two

days until the first media change. At that point, the media was gently aspirated and the

flasks were subsequently washed twice with 4 mL pre-warmed PBS. Finally, 10 mL

fresh pre-warmed media was added. Media changes were performed every second day

until cell lifting on the day of the second surgery, with the exception of changing media

on Monday morning if the last change was performed on the previous Friday afternoon.

Although we did not perform any strict characterization procedures besides visualization

of each flask’s cultured cells under a microscope, EPCs have been characterized in

previous work that applied similar isolation and culturing methods (Bates et al. 2017).

Briefly, cells were able to uptake acetylated low density lipoprotein (Ac-LDL) and bind

Ulex europaeus agglutinin 1 lectin (UEA-1), formed tube-like structures when seeded on

Matrigel for 24 hours, and displayed spindle-like morphological characteristics after 7-8

days in culture.

4.2.3 Gelfoam Scaffold Preparation and Cell Seeding

In preparation for cell seeding, PBS was warmed and 3 mL frozen aliquots of trypsin-

EDTA (0.05% trypsin, 0.53 mM EDTA with sodium bicarbonate, Wisent # 325-542-EL)

were thawed in a 37°C water bath. For cell detachment from the flasks, cell media was

kept on ice; alternatively, 5-6 mL frozen aliquots of Trypsin Neutralizing Solution (TNS,

Lonza # CC-5002) were thawed on ice. Gelfoam scaffolds (5 x 5 x 5 mm) were cut and

stored in 1.5 mL sterile microtubes.

46

The following steps describe the trypsinization process performed to detach the cells

from the flasks, count the cells and create the targeted cell dose (2 x 106). Media was

aspirated and T75 flasks were rinsed twice with 6 mL PBS. Each flask received 3 mL of

cold Trypsin-EDTA and was placed in a 37°C incubator for 5 min. Flasks were observed

under a microscope to check for cell detachment and they were tapped 3-4 times to

facilitate this process. Five mL of cold EGM-2MV media or 5-6 mL of TNS were added

and mixed in each flask to inactivate the trypsin. Cell resuspension was facilitated via

pipetting the trypsin-EDTA-media mixture up and down, and the suspension was

transferred to a 50 mL tube (a second tube was required for more than 4 flasks). Flasks

were then rinsed with 1.5-2 mL PBS, which was also transferred to the 50 mL tube(s)

before they were centrifuged at 220 x g and at room temperature for 5 min. The

supernatant was aspirated and the cell pellets were resuspended in 100 uL PBS per

EPC donor rat plus an additional 20 uL The resuspension was transferred to a sterile

1.5 mL microtube and kept on ice. Ten uL cell suspension was combined with 190 uL

PBS and 200 uL trypan blue to make a 1/40 dilution; furthermore, 1 volume of the 1/40

dilution was mixed with 1 volume of trypan blue to create a 1/80 dilution. Both dilutions

were loaded in the hemocytometer for cell counting. Live and dead cells were counted

in each quadrant, and live and total cell concentration were calculated. Based on the

calculated result, appropriate dilutions were made to the suspension in order to produce

a concentration of 20 x 106 cells/mL. This would allow the target dose of 2 x 106 EPCs

in the 100 uL cell suspension absorbed in the gelfoam scaffold.

The following steps describe how the gelfoam scaffolds were seeded with cells. Using

an 18G needle, a hole was created aseptically in the middle of each scaffold that did not

extend all the way through. After thorough re-mixing, 100 uL of suspension was

dispensed into the hole using a pipet. Cell suspension that did not absorb was re-

aspirated and dispensed into the scaffold’s hole again until less than 10 uL remained

unabsorbed. Seeded scaffolds were kept on ice until insertion into the bone defect

during surgery. This period lasted 1.5 to 7 hours, as all scaffolds were seeded with cells

in the morning and up to five surgeries were performed sequentially afterwards.

47


The same surgical procedures were followed as in section 3.1.3 for 41 male Fischer 344

rats (250-300g), with the exception of the gelfoam preparations during both surgeries

and a slight modification to the tissue biopsy extraction protocol. During the fracture

surgery, all rats received a sterile empty gelfoam scaffold to establish our non-

contaminated groups. In the second stage surgery, the first tissue biopsy was taken

from near the defect site and the second included a combination of tissue sampled from

near the proximal and distal ends of the plate. Moreover, the new gelfoam scaffold

inserted was either empty or contained 2 x 106 EPCs. The rats then received no

antibiotics, locally applied vancomycin or locally applied vancomycin and rifampin near

the defect site and along the plate before wound closure. The same doses were used

as in section 3.1.3. This established five treatment groups: control (n=9), vancomycin +

rifampin (n=7), EPC (n=7), EPC + vancomycin (n=8), and EPC + vancomycin + rifampin

(n=10). The original plan was to ultimately compare the results of these non-

contaminated groups with contaminated groups receiving the same treatments,

although the latter part of the study was not carried out due to negative bone healing

outcomes. This will be explained in further detail below.

4.2.5 Euthanasia and Harvest

Similar methods were followed as in section 3.1.4, except euthanasia and harvest were

performed on the rats at 10 weeks post-treatment. In addition, the left (contralateral)

femora were harvested along with the right (operative) femora with the intention of

comparing levels of bone formation and strength. Both legs were shaved and the right

leg was cleaned with alcohol and betadine before incision to prevent contamination of

the muscle tissue samples. Tissue sampling, removal and sonication of metal hardware,

and storage of bones were all performed in an identical fashion to the methods

described in experiment #1.

48

4.2.6 Radiography

Beginning on the week of the second surgery, radiographs were taken biweekly until

sacrifice at 10 weeks to assess bone healing based on an antero-posterior view of the

right femur. Tube to leg distance, kilovolts and milliamps were standardized. Final (10-

wk) radiographs were scored based on defect filling, callus density and healing status

by two blinded orthopaedic surgeons. Disagreements in healing status were resolved by

consensus. Table 4-2 shows the radiographic scoring system. The healing status of

each radiograph was scored as nonunion, partial union or full union

Defect filling Callus density Score

0% N/A 0

1-25% Low 1

High 2

26-50% Low 3

High 4

51-75% Low 5

High 6

76-100% Low 7

High 8

Table 4-2: Radiographic scoring system based on defect filling and callus density.

Abbreviations: N/A, not applicable.

4.2.7 Tissue Homogenization and Microbiological Culture

The methods described in section 3.1.5 were used for tissue homogenization, hardware

sonication and microbiological culturing.

49

4.2.8 Study Power and Statistical Analyses

An a priori power calculation was performed with the G*Power software (Düsseldorf,

Germany) using biomechanics data from our colleagues’ previous work (Bates et al.

2017). We considered a p-value of <0.05 as statistically significant and selected our

sample size based on a statistical power of 0.8. In terms of the results, mean

radiographic bone healing scores were compared using one-way analysis of variance

(ANOVA) and Tukey’s post-hoc multiple comparison tests.

4.3 Results

4.3.1 Radiography

Overall, radiographic results revealed healing rates that were much lower than

expected, particularly in the treatment groups that included EPC administration. For the

non-EPC-treated groups, 0/9 (0%) of the control rats and 0/7 (0%) of the rats receiving

vancomycin and rifampin achieved union by 10 weeks. This was expected since the

bone defects were critical-size and no bone augmentation treatment was given.

However, rats in the EPC group did not demonstrate much better healing outcomes, as

only 1/7 (14%) achieved union. This was in stark contrast to the results from colleagues’

previous work, reporting a 100% healing rate for the same treatment in the same model

(Bates et al. 2017). The addition of antibiotics led to marginally better healing outcomes,

with 2/8 (25%) rats in the EPC + vancomycin group and 3/10 (30%) rats in the EPC +

vancomycin + rifampin group achieving union.

Comparison of mean radiographic scores for bone healing reaffirmed the healing status

observations. All mean scores were in the two lowest quartiles of defect filling. Figure 4-1 illustrates a trending increase in radiographic scores with EPCs and antibiotics

administration, with the EPC + vancomycin + rifampin group having the highest score

50

followed by the EPC + vancomycin group and then the EPC group. However, an

ANOVA test revealed no significant differences between groups (p = 0.54), and Tukey’s

multiple comparison test (Table 4-5) showed no significant differences for all pairwise

comparisons.

Group

Number of

animals that

achieved union at

10 wks

Number of animals

that achieved

partial union at

10 wks

Number of animals

that progressed to

nonunion at 10 wks

Control (n=9) 0/9 (0%) 0/9 (0%) 9/9 (100%)

Vancomycin +

Rifampin (n=7) 0/7 (0%) 1/7 (14%) 6/7 (86%)

EPC (n=7) 1/7 (14%) 0/7 (0%) 6/7 (86%)

EPC+V (n=8) 2/8 (25%) 2/8 (25%) 4/8 (50%)

EPC+V+R (n=10) 3/10 (30%) 3/10 (30%) 4/10 (40%)

Table 4-3: Bone healing outcomes at 10 weeks after treatment.

Abbreviations: V, vancomycin. R, rifampin. Wks, weeks.

Group

Number of animals

contaminated at 0 wks

Control (n=9) 2/9 (22%)

Vancomycin + Rifampin (n=7) 1/7 (14%)

EPC (n=7) 4/7(57%)

EPC+V (n=8) 3/8 (38%)

EPC+V+R (n=10) 1/10 (10%)

Table 4-4: Contamination outcomes at 0 weeks based on overnight cultures of biopsies

taken during second stage surgery.

51

Figure 4-1: Mean 10-week radiographic scores across treatment groups. ANOVA and

Tukey’s multiple comparison tests showed no significant differences between scores.

Values are reported as Mean ± SD. Abbreviations: Ctrl, control.

2.28 ± 1.302.93± 1.37

3.07 ± 2.13

3.50 ± 2.69

4.00 ± 2.82

0

1

2

3

4

5

6

7

8

Ctrl V+R EPC EPC+V EPC+V+R

Mea

n Ra

diog

raph

ic B

one

Heal

ing

Scor

e

Treatment Groups

52

Control

Vancomycin +

Rifampin

EPC

EPC +

Vancomycin

EPC + Vancomycin +

Rifampin

Figure 4-2: Serial radiographs (2 wks, 6 wks and 10 wks from left to right) of one rat from each of the five treatments. Progressive union is indicated with yellow arrows in the EPC + vancomycin + rifampin radiographs while the other series’ show nonunion.

53

Tukey’s multiple comparison test 95.00% CI of diff. Adjusted p

value

Control (n=9) vs. V+R (n=9) -3.831 to 2.530 0.9991

Control (n=9) vs. EPC (n=7) -3.974 to 2.387 0.9514

Control (n=9) vs. EPC+V (n=8) -4.289 to 1.844 0.7821

Control (n=9) vs. EPC+V+R (n=10) -4.622 to 1.178 0.4437

V+R (n=9) vs. EPC (n=7) -3.516 to 3.231 >0.9999

V+R (n=9) vs. EPC+V (n=8) -3.838 to 2.695 0.9866

V+R (n=9) vs. EPC+V+R (n=10) -4.182 to 2.039 0.8585

EPC (n=7) vs. EPC+V (n=8) -3.695 to 2.838 0.9955

EPC (n=7) vs. EPC+V+R (n=10) -4.039 to 2.182 0.9105

EPC+V (n=8) vs. EPC+V+R (n=10) -3.494 to 2.494 0.9888


bone healing scores for treatment groups in experiment #2. Adjusted p values and 95%

confidence intervals (CIs) are shown for each pairwise group comparison.

Abbreviations: V+R, vancomycin + rifampin.

54

4.3.2 Microbiology

According to the tissue culture results, 11/41 (27%) rats were contaminated overall.

Table 4-3 displays the contamination rates for each of the five groups, with the highest

rate observed in the EPC group (57%) and the lowest in the EPC + vancomycin +

rifampin group (10%). Since these groups were not inoculated with bacteria at any

point, these rates demonstrated that practices during surgeries and tissue culturing

were not sufficiently sterile and needed refinement for future experiments.

4.3.3 Summary of Experimental Errors and Contingency Plans

The major discrepancy in healing rates between these results and those from previous

studies was likely due to both technical differences in EPC isolation, culture and

gelfoam scaffold seeding and the unintended contamination of several animals as

shown by the microbiological culture results. Experiment #3 was designed to

troubleshoot issues with the EPC handling protocols in an effort to reproduce the

previously observed high EPC-induced healing rates. Extra precautionary measures

were taken in the operating room to reduce the risk of introducing contamination,

including using sterile gowns for all members of the surgical team, performing more

frequent glove changes and autoclaving surgical instruments halfway through the

operating day. Healing outcomes were evaluated in both acute and delayed treatment

scenarios to investigate if EPC administration timing had any additional effect on bone

healing efficacy.

Furthermore, the purpose of experiment #4 was to investigate EPCs’ potential

antimicrobial effects and their interactive effects with local antibiotics in the context of an

infected critical-size defect model. It was designed as a shorter 4-week study to be

finished within the required timeline for degree completion, using infection status as a

primary outcome (rather than infection status and bone healing, which would require a

substantially longer timeline).

55

Chapter 5: Experiment #3 - Acute and Delayed EPC Treatment and Bone Healing


The purpose of this experiment was to troubleshoot the technical difficulties we

encountered in experiment #2 regarding EPC functionality. After thorough examination

of our EPC isolation and handling protocols in experiment #2, we implemented two

distinct changes in our techniques that were thought to have likely impacted the quality

and dose of EPCs seeded in the gelfoam scaffolds. In addition to testing these

alterations, we chose to compare bone healing outcomes between animals

administered EPC treatment in a delayed versus an acute manner. This would allow us

to determine if the delayed treatment setting played a role in the diminished EPC

functionality we observed, as the 2-week period in between the first and second stage

surgeries differed from the 3-week timeline followed in our colleagues’ previous study

that showed successful bone healing (Bates et al. 2017).

5.2 Methods


Rats were randomly selected to receive one of the following two treatments: 1. acute

EPC treatment (n=5) or 2. delayed EPC treatment (n=5). Each animal underwent a

femur fracture surgery to create a 5-mm critical-size defect in their right femur and

establish internal fixation using a mini-plate and screws. The rats in the acute treatment

group received a gelfoam scaffold with EPCs (2 x 106) within the same surgery,

whereas rats in the delayed treatment group underwent a second surgery two weeks

56

(13-15 days) later to clean the defect area and receive EPC treatment. Radiographs

were taken post-operatively every two weeks until animal sacrifice at ten weeks to

monitor bone healing progression.

5.2.2 Radiography

The same protocol for taking radiographs was followed as in section 4.1.6. Final (10-

week) radiographs were scored based on defect filling, callus density and healing status

by two blinded orthopaedic surgeons. The same scoring system was used.

5.2.3 Cell Isolation and Culture

The same methods were followed as in section 4.1.2.


Similar procedures were followed as in section 4.1.3, except that gelfoam dimensions

were increased (7 x 7 x 7 mm) and cells were resuspended in media instead of PBS

after centrifugation for 5 minutes.


Similar procedures were followed as in section 3.1.3 for 10 male Fischer 344 rats (250-

300g), except for the following modifications. All gelfoam scaffolds were kept sterile and

muscle tissue biopsies were not taken from any rats since infection was not a part of the

study’s design. The five rats treated with EPCs in a delayed fashion underwent two

stages of surgery separated by two weeks, receiving an empty gelfoam scaffold during

the initial fracture surgery and a scaffold with EPCs (2 x 106) in the second stage

surgery. In the other group, the five rats treated with EPCs in an acute fashion received

an EPC-seeded gelfoam scaffold within the initial fracture surgery, so only one stage of

surgery was required.

57

5.2.6 Euthanasia and Harvest

Rats were euthanized at 10 weeks post-treatment. After shaving both legs, both femora

were extracted and stored in the same way as in section 3.1.4. Again, muscle tissue

samples were not taken as in the other studies since infection outcomes were not

assessed.

5.3 Results Bone healing outcomes were extremely consistent across the acute and delayed EPC

treatment groups, with all rats achieving union by 10 weeks. This was further reflected

by the consistently high radiographic scores, as each score was in the highest quartile

of defect filling. Mean radiographic scores are reported along with the healing status

outcomes in Table 5-1. Representative radiographic images of union progression are

shown for both treatment groups in Figure 5-1.

EPC treatment timing Number of rats that

achieved union Mean radiographic score

Acute 5/5 (100%) 7.3 ± 0.12

Delayed 5/5 (100%) 7.2 ± 0.20

Table 5-1: Summary of healing status outcomes and mean radiographic scores

between the acute and delayed EPC treatment groups.

Figure 5-1: Representative radiographs (2 wks, 6wks, and 10wks from left to right)

radiographs from rats in the acute and delayed EPC treatment groups.

Acute

Delayed

58

Chapter 6: Experiment #4 - EPCs, Local Antibiotics, and Infection Outcomes in a Contaminated Critical-size Defect

Model


In this final experiment, we aimed to investigate EPC-based therapy’s potential

antimicrobial properties in the setting of a low-grade infected segmental bone defect

and analyze how the presence of local antibiotics might modify EPC-mediated infection

outcomes. Our hypothesis that EPCs would demonstrate antimicrobial effects is based

on EPCs’ immune-like phenotypic characteristics and their capacity to induce

neovascularization that may enhance the tissue site’s local immunosurveillance (Cheng

et al. 2013, Medina et al. 2017). Due to degree timeline constraints, we had to forego

bone healing evaluation (which would require a 12-week period for each animal and

more extensive testing) and focus on infection status as our primary outcome measure

over a 4-week timeframe. This period was sufficient to allow infection development over

the first 2 weeks and treatments to exhibit any antimicrobial effects over the following 2

weeks.

6.2 Methods


Rats were randomly selected to receive one of the following four treatments: 1. no

treatment (n=6), 2. local vancomycin and rifampin (n=5), 3. EPCs (n=6), or 4. EPCs and

local vancomycin and rifampin (n=4). Each animal underwent a femur fracture surgery

59

to create a 5-mm critical-size defect in the right femur and establish internal fixation

using a mini-plate and screws. In this same surgery, a gelfoam scaffold contaminated

with approximately 103 CFU S. epidermidis was inserted into the defect area before

wound closure. Two weeks (13-15 days) later, the second stage surgery was performed

to clean the defect area, remove newly grown bone, extract muscle tissue biopsies and

apply the appropriate treatment. Biopsies were homogenized and cultured to confirm

infection status. EPC isolation and culture began eight days prior to the second surgery.

Radiographs were taken on the same week as the second surgery and two weeks later

prior to sacrifice in order to assess radiographic evidence of bone healing and infection.

Muscle tissue samples and the metal plate and screws were collected upon sacrifice for

microbiological analysis.

6.2.2 Bacteria and Gelfoam Scaffold Preparation

The same procedures were followed as in section 3.1.2, except the gelfoam scaffolds

were cut into dimensions of 7 x 7 x 7 mm for better absorption of the EPC resuspension

and, for the first surgery, they were all inoculated with the 104 serial dilution (50 uL) in

order to target the 103 CFU dose.

6.2.3 Radiography

Since sacrifice was performed at two weeks after the second surgery, two radiographs

were collected per rat: one prior to the second surgery and one prior to sacrifice. The

same procedures were followed for these radiographs as in section 4.1.6. The final (2-

week) radiographs were scored for bone healing and infection status based on criteria

outlined in Table 4-2 and Table 6-1. Each radiograph’s ultimate infection score was a

sum of the points allocated to each individual infection sign present.

60

Signs of infection Points

No change 0

Loosening of hardware 1

Presence of osteolysis 2

Periosteal reaction distant from fracture

site 3

Fracture deformity 4

Table 6-1: Radiographic scoring system based on signs of infection.

6.2.4 Cell Isolation and Culture

The same methods were followed as in section 4.1.2.


Similar procedures were followed as in section 4.1.3, except with the same

modifications as listed in section 5.1.4.


Similar procedures were followed as in section 3.1.3 for 21 male Fischer 344 rats (250-

300g), except all the gelfoam scaffolds inserted into the defect site during the fracture

surgery were contaminated with 103 CFU S. epidermidis since all animal groups were

meant to have an established low-grade infection by the second surgery. In addition, the

same modification as in section 4.1.4 was made to the tissue biopsy extraction protocol

regarding which sites the biopsies were taken from.

61

6.2.7 Tissue Culture, Sonication, and Microbiological Analysis

Similar procedures were followed as in section 3.1.5 except with a few additional steps

and a modified infection status threshold. Rats were considered positive for infection if

at least one out of the two biopsies taken during the second surgery had a positive

culture result. In addition to overnight cultures on TSA plates, 10 uL of each

homogenized biopsy or sample was inoculated in 3 mL TSB. These broth cultures were

kept in incubation at 37°C for up to 14 days to confirm the presence or absence of

bacterial growth. Animals with negative biopsy and corresponding broth cultures were

considered not infected and excluded from further analysis. Hemolysis, coagulase and

catalase tests were also performed on sample colonies from several randomly selected

rats to supplement morphological evidence of S. epidermidis presence.

6.2.8 Statistical Analyses

Mean radiographic scores for bone healing and infection status were compared using

one-way ANOVA and Tukey’s post-hoc multiple comparisons tests.

6.3 Results

The following results describe the radiographic and microbiological culture outcomes for

the 21 animals that were ultimately included in the analysis. A total of 25 animals were

operated on, but 4 were excluded from the final analysis. Three naturally resolved their

infections by the second stage surgery and 1 showed radiographic evidence of

premature mechanical failure due to technical errors with hardware installation during

the first stage of surgery.

62

6.3.1 Radiography

Images showed nonunion in all rats at 2 weeks, which was expected since the

timeframe for the given treatments to have an effect was too short to allow full bone

healing. This corresponded with extremely low radiographic bone healing scores across

all treatments, as shown in Table 6-2. Regarding infection scores, the EPC +

vancomycin + rifampin group had the lowest score (0.88 ± 0.59), followed by the

vancomycin + rifampin group (1.30 ± 0.44). The control and EPC groups were equal

with the highest scores (2.75 ± 0.94 and 2.75 ± 0.51, respectively). However, ANOVA

and Tukey’s multiple comparison tests showed no significant differences between

groups, even when comparing the highest and lowest scores (p = 0.29).

Treatment group Radiographic score for

bone healing @ 2 wks

Radiographic score for

infection status @ 2 wks

Control (n = 6) 0.50 ± 0.18 2.75 ± 0.94

EPC (n = 6) 0.92 ± 0.20 2.75 ± 0.51

V+R (n = 5) 0.70 ± 0.20 1.30 ± 0.44

EPC+V+R (n = 4) 0.38 ± 0.24 0.88 ± 0.59

Table 6-2: Summary of radiographic scores for bone healing and infection status at 2

weeks post-treatment for all treatment groups. ANOVA and Tukey’s multiple

comparison tests showed no significant differences between groups. Abbreviations:

V+R, vancomycin + rifampin. Wks, weeks.

63

Figure 6-1: Graphical representation of radiographic scores for bone healing and

infection status across treatment groups. Values are represented as Mean ± SD.

Abbreviations: Ctrl, control. V+R, vancomycin + rifampin.

Figure 6-2: Representative radiographs (0 wks and 2 wks from left to right) from each

treatment group. Red arrows point to signs of osteolysis and blue arrows point to signs

0

1

2

3

4

5

6

Ctrl EPC V+R EPC+V+R

Mea

n Ra

diog

rpah

ic S

core

Treatment Group

Bone Healing Infection Status

Control

EPC

V+R

EPC+V+R

64

of periosteal reaction distant from the defect site. Abbreviations: V+R, vancomycin +

rifampin.

6.3.2 Microbiology

Regarding infection status, cultures of tissue samples and sonicated hardware showed

a downward trend. The EPC + vancomycin + rifampin group had the lowest proportion

(25%) of rats infected at 2 weeks, followed by the vancomycin + rifampin group (40%),

the EPC group (67%) and finally the control group (83%). The data suggested that

EPCs did not cause any notable additional bacterial eradication in the presence or

absence of local antibiotics.

Figure 6-3: Graphical representation of infection outcomes at 2 weeks post-treatment.

Rats were considered infected if at least 2 out of 5 samples collected at harvest were

positive. Abbreviations: Ctrl, control. V+R, vancomycin + rifampin.

5/6 (83%)

4/6 (67%)

2/5 (40%)

1/4 (25%)

0

10

20

30

40

50

60

70

80

90

100

Ctrl EPC V+R EPC+V+R

Perc

enta

ge o

f Rat

s Inf

ecte

d @

2 W

ks

Treatment Groups

65

Tukey’s multiple comparison test 95.00% CI of diff. Adjusted p

value

Control (n=6) vs. EPC (n=6) -2.587 to 2.587 >0.9999

Control (n=6) vs. V+R (n=5) -1.263 to 4.163 0.4484

Control (n=6) vs. EPC+V+R (n=4) -1.017 to 4.767 0.2885

EPC (n=6) vs. V+R (n=5) -1.263 to 4.163 0.4484

EPC (n=6) vs. EPC+V+R (n=4) -1.017 to 4.767 0.2885

V+R (n=5) vs. EPC+V+R (n=4) -2.581 to 3.431 0.9773


infection scores for treatment groups in experiment #4. Adjusted p values and 95%

confidence intervals (CIs) are shown for each pairwise group comparison.

Abbreviations: V+R, vancomycin + rifampin.

66

Chapter 7: Discussion

In our studies, we sought to investigate the effects of EPC-based therapy and local

antibiotic therapy on bone healing and infection status outcomes in the context of a low-

grade infected nonunion. Infection is one of the most common risk factors associated

with fracture nonunion and often requires a complex, multi-staged surgical management

strategy. Typically, surgeons focus the first stage of treatment on treating the infection

and the second stage on facilitating bone repair. This process has an immense cost and

renders the patient physically incapacitated for a long period, placing a heavy financial

burden on the patient and healthcare system and having detrimental impacts on the

patient’s quality of life. The development of new therapies that can manage infection

and fracture healing outcomes simultaneously within a single stage would provide

invaluable benefits for those with infected nonunions. Our colleagues have previously

demonstrated the successful healing of critical-size segmental bone defects with the

local administration of EPCs (Bates et al. 2017). Based on their phenotypic similarities

to macrophages and their capacity to induce neovascularization, we were interested in

assessing the potential antimicrobial effects that EPCs may have in infected nonunion

scenarios (Cheng et al. 2013, Medina et al. 2017). If they exhibited dual antimicrobial

and osteogenic properties, they would make promising cell candidates for a single-

staged therapy. Moreover, since local administration of antibiotics is gaining popularity

as a surgical strategy for infection treatment and prevention amongst orthopaedic

surgeons, we also sought to evaluate how it may impact EPC-mediated healing

outcomes.

We chose to focus on low-grade infections given the rising awareness of their role in

nonunion aetiology. Referring to a previous in vivo study with a similar model, we

successfully established a rat model of S. epidermidis-induced infected nonunion in our

pilot study (experiment #1). Unfortunately, we encountered setbacks in our main

experiment (experiment #2). We began by performing surgeries for our non-

contaminated treatment groups, but flaws in our EPC isolation and handling techniques

led to poor bone healing results in the EPC control group. As a result, we could not

67

generate definitive conclusions about the effects of local antibiotics on EPC-induced

bone healing in a sterile wound. Given our remaining time, we designed an experiment

to troubleshoot our technical errors (experiment #3) and a subsequent study to

investigate EPCs’ potential antimicrobial properties with and without local antibiotics

(experiment #4). In experiment #3, our improved techniques led to high bone healing

rates in rats treated with EPCs, both in an acute and delayed fashion. Regarding

experiment #4, our radiographic and microbiological culture results demonstrated that

our model resulted in a high rate of persistent infection in the absence of any treatments

and that EPCs did not contribute any additional eradication of low-grade infection either

in the presence or absence of local antibiotics.

7.1 Low-Grade Infection Animal Model

Regarding the different inoculation doses trialled, our pilot study’s results were

consistent with previous work that investigated a low-grade infected segmental bone

defect rat model (Lovati et al. 2016). The 104 and 105 CFU groups had positive biopsy

cultures with bacterial densities greater than 5000 CFU/g tissue, which were considered

too high to be indicative of low-grade infections. In comparison, the 103 CFU group had

bacterial densities ranging from 50-850 CFU/g tissue, illustrating a much less severe

type of infection. This aligned with observations in Lovati et al.’s study (2016), where

rats in the 105 CFU group showed more severe signs of osteomyelitis and abscess

formation compared to the 103 CFU group. Given the low densities and consistently

positive cultures found post-inoculation, 103 CFU appeared to be the most reliable dose

for establishing a S. epidermidis-induced low-grade infection over 2 weeks.

Similar to Lovati et al.’s study, we experienced natural infection clearance in some rats

treated with the 103 CFU dose. In total, 25 rats were operated on for experiment #4, but

3 rats had to be excluded from the data analysis due to both biopsy cultures testing

negative. This was a lower rate (12%) than in Lovati et al.’s study (2016), which

reported 2/5 (40%) rats with negative culture outcomes at the same dose. Nonetheless,

the observed infection eradication resulting from the host’s own immune response

68

indicated that 103 CFU was on the border of doses able to establish a persistent

infection over 2 weeks. This was also reflected by the pilot study’s outcomes for the

lowest dose (102 CFU) group, as one out of these two rats had a negative biopsy test.

This along with absent clinical signs of infection suggested that the 103 CFU dose was

appropriate for modelling low-grade S. epidermidis infections in rats and should be used

in similar future in vivo experiments. The main caveat for using this dose is that a

certain rate of spontaneous infection clearance should be anticipated and prepared for

with additional animals available to make up for these exclusions. This exact rate is

unclear, but the proportions in our experience (12%) and Lovati et al.’s experiment

(40%) may be considered as initial references (Lovati et al. 2016).

Furthermore, results from cultures of sonication fluid provided some evidence of

successful biofilm development on the implanted plate and screws. In 12 out of the 21

rats analyzed in experiment #4, the cultured sample of sonication fluid had higher

colony counts than any of the other tissue samples taken during the harvest (well over

4000 CFUs on average). Moreover, it was the most common positive sample in cases

that only had 1 out of 5 positive cultures, demonstrating the bacteria’s tendency to grow

on the hardware. Given this strain’s well known biofilm-forming capability, we can infer

that the large quantities of colonies detected on the hardware were likely present as an

established biofilm, although we lack certain evidence (Chusri et al. 2012). This is a

critical aspect of the low-grade infection animal model since biofilms heavily contribute

to the infection’s recalcitrance (Stoodley et al. 2011, Zimmerli and Moser 2012, Arciola

et al. 2018). Without the necessary equipment and training, imaging techniques were

not used in this study but should be included in future investigations if possible in order

to include visual evidence of biofilm formation.

7.2 Combination Antibiotic Therapy

Both radiographic and bacterial culture results further supported the literature’s reports

of vancomycin and rifampin’s combined therapeutic efficacy against established

staphylococcal biofilm infections (Claessens et al. 2015, Saginur et al. 2006, Thompson

69

et al. 2017, Brinkman et al. 2017, Niska et al. 2013). In the pilot study, the combined

vancomycin and rifampin treatment was able to completely eradicate the S. epidermidis

infections generated by the 102, 103 and 104 CFU doses when administered 2 weeks

post-inoculation, and it paralleled rifampin monotherapy’s effects at the 104 and 105

CFU doses. This provided sound evidence for its selection as an effective dual antibiotic

regimen in experiment #4, which showed some degree of infection eradication in the

group receiving the combined therapy compared to the controls. The vancomycin and

rifampin-treated group had markedly lower final radiographic scores (1.30 ± 0.44) and

positive culture outcomes at the harvest (40%) compared to the control group (2.75 ±

0.94 and 83%, respectively). However, these differences may have been due to chance

since statistical significance was not proven for the radiographic scores (p = 0.448), but

this could have been due to the study lacking statistical power with relatively small

sample sizes (n = 4-6). An experimental design with larger group sizes would have to be

performed to confirm if the observed trend was true and to offer more solid evidence of

local vancomycin and rifampin’s therapeutic efficacy in low-grade infected segmental

bone defects. Nonetheless, this study potentially serves as an early steppingstone

towards applications of local rifampin within a combined therapy for common

staphylococcal biofilm infections.

7.3 Technical Errors and EPC Functionality

The significant discrepancy between bone healing outcomes in experiment #2 and

previous studies done by our colleagues reflected technical shortcomings in our EPC

isolation and handling methods. While Bates et al. (2017) found that 100% of critical-

size bone defects achieved union with acute or delayed EPC treatment, only 14% of

rats with EPC treatment achieved union in our study. Marginally higher healing

outcomes were seen in the antibiotic-treated EPC groups, with 25% and 30% of EPC-

treated rats achieving union when administered with vancomycin alone and vancomycin

+ rifampin, respectively. Nonetheless, the unexpectedly low healing efficacy indicated

that our EPCs had reduced functionality compared to previous work. Although

unintentional contaminations may have played a confounding role in the experiment,

70

these alone did not explain the low healing outcomes observed, since union rates

remained extremely low even after excluding contaminated rats from our analysis.

Upon closer examination of our EPC seeding protocol, we discovered several technical

faults that likely hindered EPC-mediated effects. Firstly, our gelfoam scaffold

dimensions of 5 x 5 x 5 mm were too small to allow absorbance of the entire EPC

suspension volume, resulting in some cell loss and an administered EPC dose below

our target (2 x 106). This was visually apparent during surgeries with residual EPC

suspension often remaining in the scaffold’s carrier tube after it was transferred to the

defect site. Scaffold dimensions were increased to 7 x 7 x 7 mm in experiment #3 to

account for this loss, enabling a better absorption of the cell suspension for every

animal. Moreover, we used PBS as our EPC resuspension solvent as opposed to cell

culture media in experiment #2, which was devoid of factors necessary for cellular

growth. We propose that this negatively impacted the EPCs’ health in the period

between cell lifting and scaffold placement in the defect and suppressed their

osteogenic properties. For experiment #3, we switched to resuspending in cell culture

media (EGM-2 MV) that contained various supplements (FBS, hydrocortisone, hFGF-B,

R3-IGF-1, ascorbic acid, hEGF and GA-1000), hoping that enhanced cell growth

conditions would lead to superior EPC-mediated healing outcomes. Furthermore, we

supposed that some other gradual changes in EPC isolation and handling that were

difficult to track may have also played a role in our technical learning curve, as we

observed improving healing outcomes in EPC-treated rats and increasing cell yields

over time during experiment #2.

The results that followed demonstrated how critical the aforementioned technical details

were. Sequential radiographs in experiment #3 displayed a gradual progression to union

in 100% of rats in both the acute (n = 5) and delayed (n = 5) EPC treatment groups, with

consistent strong evidence of union achievement by 6 weeks post-treatment. These

high healing rates matched those previously demonstrated by our colleagues and

reassured us that our initial isolation and expansion techniques were sufficient. The

same protocol changes were implemented in experiment #4 so we also concluded that

71

EPCs administered in that study had the same health quality as those in experiment #3.

Thus, our experimental errors and successful troubleshooting steps emphasized the

importance of technical experience and attention to fine details in EPC preparation

standard operating procedures (SOPs). The changes we implemented seemed subtle,

but they led to marked differences in cell-mediated effects with an 84% higher union

rate. These details were clarified in our laboratory’s SOPs for future EPC-related

investigations. In addition to these technical modifications, we believe that increased

experience in the operating room and the biosafety cabinet led to improved sterility

during surgeries and cell culturing practices, reducing the presence of contaminations

and their impact on our results.

7.4 EPCs, Antibiotics, and Bone Healing

The effects of local antibiotics on EPCs in the context of fracture healing has not yet

been investigated, but extrapolation from related studies can offer some insight into

what can be expected. So far, both in vitro and in vivo studies generally report no

detrimental effects of vancomycin on cell-mediated osteogenesis within the clinically

relevant dosing range (Edin et al. 1996, Mendoza et al. 2016, Rathbone et al. 2011). As

for rifampin, in vitro evidence demonstrates the potent toxicity of rifampin to osteogenic

cells at much lower concentrations that overlap with clinical doses, generating concern

over its application in close proximity to EPCs within the defect site (Isefuku et al. 2001,

Rathbone et al. 2011, Zhang et al. 2014). If rifampin reduces EPC viability within the

clinical dosing range (i.e. 25 mg/kg), this may translate to impeded bone regeneration in

in vivo models. Fortunately, two animal studies showed positive bone regeneration

effects of local rifampin administration with bone graft in a sterile wound and without any

bone augmentation treatment in a contaminated wound (Shiels et al. 2018, Durmuşlar

et al. 2016). This suggests that rifampin’s cytotoxicity may be modified by other factors

in an in vivo environment and offers early promise for its therapeutic potential in fracture

healing. Altogether, our understanding of local rifampin’s effects in this context is less

72

clear than local vancomycin’s, which is reflected by orthopaedic surgeons’ preference

towards using the latter for routine antibiotic prophylaxis.

Our experiment #2 was designed to explore this knowledge gap, but unfortunately the

difficulties encountered with EPC-mediated bone healing prevented us from drawing

any definitive conclusions. Despite the overall low healing rates in the EPC-treated

groups, we observed higher union rates when EPCs were combined with local antibiotic

administration. However, the differences failed to achieve statistical significance, so

strong conclusions could not be made based on comparisons of the groups’ mean

radiographic scores and radiographic healing outcomes. Moreover, the uneven

distribution of contaminations across treatment groups confounds this relationship.

Exclusion of the contaminated rats from the analysis led to similar results, but the

sample sizes for some groups became too low to consider the results very meaningful.

Contaminations aside, the positive trend with local vancomycin and rifampin

administration is a promising sign for their application with EPC-based fracture healing

therapies, but future investigations with better EPC isolation, handling and sterility

techniques are necessary. For now, we can surmise that the addition of local

vancomycin and rifampin may not strongly inhibit EPC-induced bone healing based on

our study’s weak evidence in accordance with the current in vivo literature.

7.5 EPCs, Antibiotics, and Infection

Thus far, there is a dearth of literature investigating the role of EPCs in infection.

Our postulation of EPCs’ potential antimicrobial properties is limited to theory based on

their phenotypic similarity to certain immune cells and their enhancement of local

immunosurveillance via neovascularization (Cheng et al. 2013, Medina et al. 2017). In

contrast, the antimicrobial activity of MSCs is well-documented and has demonstrated

efficacy in infection scenarios associated with sepsis, acute lung injury, acute

respiratory distress syndrome, orthopaedic injury, and tuberculosis. Recently, some

studies have even reported MSC-based cell therapy’s efficacy against chronic and drug-

resistant infections through potentiating antibiotics’ bactericidal activity and directly

73

contributing to biofilm eradication themselves (Sutton et al. 2016). So far, a few

mechanisms that govern their antimicrobial properties have been discovered. First,

MSCs produce a variety of antimicrobial peptides that both directly and indirectly kill

bacteria via damaging their cell membranes and recruiting immune cells (Mahlapuu et

al. 2016). They also secrete factors that enhance the phagocytic activity of

macrophages and neutrophils, facilitating the host’s innate immune responses (Rabani

et al. 2018, Chow et al. 2020). Finally, MSC-secreted factors also induce neutrophil

extracellular trap formation, which plays an important role in neutrophils’ bactericidal

activity (Chow et al. 2020). Given their similar origin and comparative tissue

regenerative properties, EPCs may also have antimicrobial properties yet to be

discovered. Research should start with assessing the bactericidal activity of EPCs’ in

vitro and infection outcomes for infected animals administered EPC-based therapies.

To our knowledge, experiment #4 was the first in vivo study that ever assessed infection

outcomes after EPC-based therapy. Altogether, our study provides insufficient data to

suggest that EPCs have any antibacterial properties on their own. Although differences

were noted in microbiological culture outcomes, they were not large enough to draw

confident conclusions regarding EPCs’ effects on infection status in the presence or

absence of local antibiotics. Moreover, no significant differences were detected across

all groups regarding mean radiographic scores of infection status. We believe that our

study was underpowered and that increased sample numbers would be necessary to

better evaluate EPCs’ potential antimicrobial properties. Thus, future investigations

should include much larger groups and a longer post-operative timeline to see if bone

healing and infection status outcomes are more pronounced at later timepoints.

7.6 Limitations

Several limitations in our experimental design restricted the scope of our conclusions.

Firstly, given the troubleshooting steps we had to take after experiment #2, we had

insufficient time to complete another in vivo experiment within the degree’s required

timeline. This prevented us from achieving one of our stated goals of assessing the

74

impact of local antibiotics on EPC-mediated bone healing. As noted above, some

positive trends were observed, but in the context of the overall low rates of healing we

are unable to make any firm conclusions with regard to the impact of local antibiotics on

EPC-mediated bone healing. We planned experiment #4 in order to address the aim of

assessing the effect of EPCs and local antibiotics on infection eradication in a model of

established low-grade infection within the timeframe for degree completion. A smaller

sample size was necessary to stay within the required timelines, which in turn reduced

our statistical power. Spontaneous infection clearance was an additional variable that

was difficult to account for when designing our experimental timeline, especially in the

midst of the COVID-19 global pandemic. Increasing restrictions on lab activities

prevented us from achieving completely balanced groups (n=4-6) with makeup

surgeries. This may have critically impacted the study’s statistical power at such a small

sample size and led to the lack of significant results. Therefore, future investigations

should be designed with larger treatment groups to see if the trends are consistent with

more animals and reflect true effects.

Protocols used to identify and define infection were somewhat arbitrary and lacked

specific bacterial characterization. We considered two or more positive cultures as

definitive evidence of infection at the study endpoint based on the Infectious Disease

Society of America’s guidelines. However, this criteria excluded several animals that

presented a single positive culture, and we have reason to believe that they were

indeed infected despite not meeting the threshold for culture positivity. Specifically,

three of them had a high density of colonies in their sonication fluid sample suggesting

that a biofilm likely developed on the plate and screws, but more sensitive biofilm

detection methods would be necessary for confirmation. For instance, if we

supplemented the culture results with scanning electron microscopic visualization of the

biofilm similarly to Lovati et al. (2016), we could perhaps identify positive infection

scenarios where the culture threshold was not met. In the future, collaborating with

other labs that have the required imaging technology could help with biofilm

identification.

75

More thorough bacterial identification testing would also improve our certainty regarding

the infecting species present. We applied catalase, coagulase and blood agar tests on

colonies in multiple rats’ samples in experiment #4, and the results were always

consistent with S. epidermidis characteristics. Aside from these tests, morphological

characteristics of all colonies were consistent with S. epidermidis criteria. We also

purchased novobiocin disks but laboratory restrictions had already begun before we

were able to conduct tests. Regardless, our tests were sufficient to eliminate S. aureus

as a common potential contaminant given the different beta-hemolytic result S.

epidermidis produces after incubation on blood agar. A gram-staining test would also be

ideal to distinguish staphylococcal colonies from gram-negative bacilli, another common

pathogen found in orthopaedic SSIs (Debarge et al. 2007, Edmiston Jr et al. 2005).

Furthermore, a differently designed control treatment could better isolate the effects of

EPCs on bone healing and infection outcomes. We applied a sterile and dry gelfoam

scaffold to the defect sites of control rats, but this does not account for the possible

benefits of media administration. As described earlier, our cell culture media contained

several supplements meant to support endothelial cell growth that may have also

enhanced activity in other cell types, including osteoblasts. Without including media in

our control gelfoam scaffolds, we are unsure about whether or not these growth factors

were partially responsible for any bone regeneration observed in experiments #2 and

#3. A similar logic extends to infection status in experiment #4, although there is less

reason to believe that the media would affect these outcomes.

With an increasing interest in the clinical translation of EPC-based therapies, cell

characterization is becoming a more and more important step in EPC preparation. The

highly ambiguous usage of the term “EPC” is still an obstacle to having more defined

cell subpopulations for tissue engineering applications. So far, the subcategorization of

EPCs into MACs and ECFCs has provided a useful dichotomy that should be integrated

into standard nomenclature until more appropriate definitions are agreed upon. We are

still limited to calling our cells “EPCs” since our studies did not include stringent

identification schemes. The short culture period and spindle-like morphology of our

76

EPCs suggest that they are most likely MACs, but supplementing their characterization

with other methods like immunophenotyping would allow a more definitive classification.

For example, staining for protein uniquely expressed by MACs such as CD45, CD14

and CD31 would offer more resounding evidence of our cells’ identity (Medina et al.

2017). Precise identification protocols would also likely be used in clinical applications of

EPC-based therapies, so similar procedures should be done in our preclinical

investigations to establish consistency. The sooner we adopt these practices, the

sooner we can phase out the name “EPC” for more accurate terms.

77

Chapter 8: Conclusions

Over our series of experiments, we were able to draw multiple conclusions. In our pilot

study, we successfully established a consistent animal model of S. epidermidis-induced

low-grade infection and demonstrated local rifampin’s antimicrobial efficacy either as a

monotherapy or as part of a combined therapy with local vancomycin. Moreover,

although our second experiment showed unexpectedly low bone healing rates in EPC-

treated groups irrespective of antibiotic application, we managed to troubleshoot our

technical difficulties and produce high healing rates of critical-size bone defects with

both acute and delayed manners of EPC treatment consistent with previous

experimental results (Atesok et al. 2010, Li et al. 2011, Bates et al. 2017). Finally, the

radiographic and microbiological outcomes in our final experiment did not offer any

evidence of EPCs contributing antimicrobial effects in a low-grade infection nonunion

scenario, either in the presence or absence of local antibiotics. However, this study was

limited by its small sample size and lacked sufficient statistical power to draw definitive

conclusions. Future studies should be designed with larger groups and longer

experimental timelines to better assess both bone healing and infection status

outcomes after the administration of EPCs and local antibiotics. These should offer

stronger implications regarding the usefulness of this combined therapy for patients

suffering from low-grade infected nonunions.

78

Chapter 9: Future Directions

Our experiments’ results show promise for the application of EPCs with local antibiotics

in infected nonunion scenarios. However, some remaining unanswered questions

should guide future investigations in this field.

1. Do local antibiotics impact EPC-mediated bone healing outcomes in

segmental bone defects?

Antibiotic prophylaxis is a routine part of nonunion revision surgeries, so surgeons must

assess the feasibility of integrating antibiotics with bone augmentation therapies,

especially when investigating potential single-stage surgical strategies for infected

nonunion scenarios. Antibiotics must be carefully selected such that they can exert their

antibacterial effects while not interfering with bone healing outcomes. Since local

antibiotic administration represents a potential novel prophylactic strategy, it should be

tested in animal models for its therapeutic efficacy in combination with bone

reconstructive treatments. Regarding EPCs, local antibiotics have not yet been

investigated with EPC-based therapy in terms of bone healing. So far, in vitro studies

report that vancomycin does not negatively impact osteogenic cell viability, whereas

rifampin does (Edin et al. 1996, Rathbone et al. 2011, Zhang et al. 2014, Isefuku et al.

2011). Moreover, local vancomycin therapy did not affect bone formation in vivo, and

studies actually reported positive bone healing effects from local rifampin therapy

(Shiels et al. 2018, Durmuşlar et al. 2016, Mendoza et al. 2016). Although this literature

is promising, it does not offer a clear indication of how local antibiotics would affect

EPC-mediated bone repair in a critical-size defect scenario. This investigation would

offer insight into their potential combination as a single-stage therapy for infected

nonunions. Our original main experiment (#2) was designed to address this question,

but the technical difficulties that we encountered with our initial EPC experiments

resulted in overall low rates of bone healing and precluded us from drawing definitive

conclusions on this. Further animal experiments with better EPC isolation and handling

79

techniques are necessary to deepen our understanding in this area. Moreover, this

should be investigated in both sterile and contaminated defects. The sterile scenario

would allow isolation of local antibiotics’ effects, whereas the contaminated scenario

would demonstrate how bacterial presence may modify these outcomes.

2. Do EPCs exhibit antimicrobial effects with and without local antibiotics? Vascularization is an essential component in successful fracture healing biology. As

demonstrated by previous studies, EPC-based therapy has shown strong success in

fracture healing largely due to the cells’ unique ability to induce local neovascularization

in the regenerated bone tissue (Li et al. 2012, Giles et al. 2017, Seebach et al. 2012).

We believe that these new blood vessel networks may also facilitate greater local

immunosurveillance in the new bone, allowing a more robust immune response to

foreign pathogens. Moreover, some EPCs have cell surface markers resembling

immune cell phenotypes, suggesting that they may be able to directly respond to

antigens as well (Cheng et al. 2013, Medina et al. 2017). We aimed to evaluate EPCs’

potential antimicrobial effects in experiment #4, but our study was small-scale and had

insufficient statistical power to show any effects. Both alone and in combination with

local antibiotic therapy, EPCs did not show any significant effect on infection

eradication. A follow-up study should be designed with larger groups that

simultaneously investigates bone healing and infection status over the course of ten

weeks post-treatment. A larger sample size would increase the study’s power, the

longer timeframe would allow sufficient time to monitor complete bone healing, and any

effects on infection status by each respective treatment could potentially become more

pronounced over time. Such a study may provide more insight into whether or not EPCs

can exert their own antimicrobial effects or have synergistic effects with local antibiotics.

80

3. Is local antibiotic prophylaxis effective and safe for clinical translation? Currently, systemic administration of cephalosporins is the standard mode of

perioperative antibiotic prophylaxis (Hansen et al. 2014, Gans et al. 2017). Local

antibiotic administration has demonstrated successful SSI prevention in multiple

surgical fields and has the potential to be a more effective mode of perioperative

prophylaxis, but its possible adverse effects have yet to be fully understood (Chen et al.

2018). For this reason, local antibiotic prophylaxis has not been adopted into surgical

guidelines yet, although many orthopaedic spine surgeons already believe strongly in

the efficacy of local intra-wound vancomycin powder and apply it routinely to prevent

SSIs (Fleischman et al. 2017). While there are ongoing clinical trials testing intra-wound

vancomycin powder’s safety and efficacy, there has been little to no investigation of

other powdered antibiotics in orthopaedic trauma. This will be crucial for expanding the

scope of local antibiotic prophylaxis since other antibiotic species may provide greater

therapeutic efficacy than vancomycin in certain infection scenarios. For instance, our

pilot study (experiment #1) demonstrated that the administration of local rifampin

powder with local vancomycin powder led to a far superior eradication of S. epidermidis-

induced biofilm infections than the vancomycin alone. Along with two other studies done

by Shiels et al. (2018), this was one of very few investigations of intra-wound rifampin

powder administration. Further in vivo animal experiments should be performed to

evaluate the efficacies of different local antibiotics regimens in a variety of infection

scenarios while also monitoring signs of toxic side effects. Through such investigations,

we could develop safety profiles for local antibiotics and better understand their

spectrum of applications in orthopaedic surgery.

Infected nonunions represent a problematic clinical scenario often involving a resource-

intensive treatment approach. The research presented here along with the work of

others provides a steppingstone towards higher level investigations of EPCs and local

antibiotics as a single-stage treatment strategy for low-grade infected nonunions. With

further support from ongoing preclinical studies and clinical trials, new cell-based and

local antibiotic therapies may ultimately reach clinical translation. Such treatment

81

strategies could potentially revolutionize the management of infected nonunions,

making it a truly exciting time for research in this field!

82

References Al-Mulhim, F. A., Baragbah, M. A., Sadat-Alit, M., Alomran, A. S., Azam, M. Q. 2014.

Prevalence of surgical site infection in orthopedic surgery: A 5-year analysis. Int Surg. 99(3): 264-268.

AlBulhairan, B., Hind, D. & Hutchinson, A. 2008. Antibiotic prophylaxis for wound infections in total joint arthroplasty: A systematic review. J Bone Joint Surg Br. 90(7): 915-919.

Anderson, D. J., Arduino, J. M., Reed, S. D., Sexton, D. J., Kaye, K. S., Grussemeyer, C. A., Peter, S. A., Hardy, C., Choi, Y. I., Friedman, J. Y., Fowler Jr, V. G. 2010. Variation in the type and frequency of postoperative invasive Staphylococcus aureus infections according to type of surgical procedure. Infect Control Hosp Epidemiol. 31(7):701–709.

Andersson, A. E., Bergh, I., Karlsson, J., Eriksson, B. I., Nilsson, K. 2012. The application of evidence-based measures to reduce surgical site infections during orthopedic surgery – Report of a single-center experience in Sweden. Patient Saf Surg. 6(1): 11

Andrzejowski, P. & Giannoudis, P.V. 2019. The ‘diamond concept’ for long bone non-union management. J Orthop Traumatol. (2019) 20: 21.

Arici, V., Perotti, C., Fabrizio, C., Del Fante, C., Ragni, F., Alessandrino, F., Viarengo, G., Pagani, M., Moia, A., Tinelli, C., Bozzani, A. 2015. Autologous immuno magnetically selected CD133+ stem cells in the treatment of no-option critical limb ischemia: clinical and contrast enhanced ultrasound assessed results in eight patients. J Transl Med. 13: 342.

Asahara, T., Murohara, T., Sullivan, A., Silver, M., van der Zee, R., Lit, T., Witzenbichler, B., Schatteman, G., Isner, J. M. 1997. Isolation of putative progenitor endothelial cells for angiogenesis. Science. 275(5302): 964-967.

Atesok, K., Li, R., Stewart, D. J., Schemitsch, E. H. 2010. Endothelial progenitor cells

promote fracture healing in a segmental bone defect model. J Orthop Res. 28(8): 1007-1014.

Ban, K. A., Minei, J. P., Laronga, C., Harbrecht, B. G., Jensen, E. H., Fry, D. E., Itani, K. M. F., Dellinger, E. P., Ko, C. Y., Duane, T. M. 2016. American College of Surgeons and Surgical Infection Society: surgical site infection guidelines, 2016 update. J Am Coll Surgeons. 224(1): 59-74.

Banno, K. & Yoder, M. C. 2019. Endothelial stem and progenitor cells for regenerative medicine. Curr Stem Cell Rep. 5: 101-108.

83

Bates, B. D., Godbout, C., Ramnaraign, D. J., Schemitsch, E. H., Nauth, A. 2017. Delayed endothelial progenitor cell therapy promotes bone defect repair in a clinically relevant rat model. Stem Cells Int. 2017: 7923826.

Beldame, J., Lagrave, B., Lievain, L., Lefebvre, B., Frebourg, N., Dujardin, F. 2012. Surgical glove bacterial contamination and perforation during total hip arthroplasty implantation: When gloves should be changed. Orthop Traumatol Surg Res. 98(4): 432–440.

Bhandari, M., Fong, K., Sprague, S., Williams, D., Petrisor, B. 2012. Variability in the definition and perceived causes of delayed unions and nonunions: A cross-sectional, multinational survey of orthopaedic surgeons. J Bone Joint Surg Am. 94(15): e1091-1096.

Bhat, A. K., Parikh, N. K. & Acharya, A. 2018. Orthopaedic surgical site infections: A

prospective cohort study. Can J Infect Control. 33(4): 227-229.

Bishop, J. A., Palanca, A. A., Bellino, M. J., Lowenberg, D. W. 2012. Assessment of compromised fracture healing. J Am Acad Orthop Surg. 20(5): 273-282.

Boxma, H., Broekhuizen, T., Patka, P., Oosting, H. 1996. Randomised controlled trial of

single-dose antibiotic prophylaxis in surgical treatment of closed fractures: the Dutch Trauma Trial. Lancet. 347(9009): 1133–1137.

Bratzler, D. W., Dellinger, E. P., Olsen, K. M., Perl, T. M., Auwaerter, P. G., Bolon, M. K., Fish, D. N., Napolitano, L. M., Sawyer, R. G., Slain, D., Steinberg, J. P., Weinstein, R. A. 2013. Clinical practice guidelines for antibiotic prophylaxis in surgery. Am J Health Syst Pharm. 70(3): 195-283.

Bratzler, D. W. & Hunt, D. R. 2006. The Surgical Infection Prevention and Surgical Care Improvement Projects: National initiatives to improve outcomes for patients having surgery. Clin Infect Dis. 43(3): 322-330.

Brinker, M. R., Hanus, B. D., Sen, M., O’Connor, D. P. 2013. The devastating effects of tibial nonunion on health-related quality of life. J Bone Joint Surg Am. 95(24): 2170-2176.

Brinker, M. R., O’Connor, D. P. 2016. Management of aseptic tibial and femoral

diaphyseal nonunions without bony defects. Orthop Clin North Am. 47(1): 67-75.

Brinkman, C. L., Schmidt-Malan, S. M., Mandrekar, J. N., Patel, R. 2017. Rifampin-based combination therapy is active in foreign-body osteomyelitis after prior rifampin monotherapy. Antimicrob Agents Chemother. 61(2): e01822-16.

Brophy, R. H., Bansal, A., Rogalski, B. L., Rizzo, M. G., Weiner, E. J., Wolff, B. D., Goldfarb, C. A. 2019. Risk factors for surgical site infection after orthopaedic

84

surgery in the ambulatory surgical center setting. J Am Acad Orthop Surg. 27(20): e928-e934.

Bryson, D. J., Morris, D. L. J., Shivji, F. S., Rollins, K. R., Snape, S., Ollivere, B. J. 2016. Antibiotic prophylaxis in orthopaedic surgery: Difficult decisions in an era of evolving antibiotic resistance. Bone Joint J. 98-B(8): 1014-1019.

Brophy, R. H., Bansal, A., Rogalski, B. L., Rizzo, M. G., Weiner, E. J., Wolff, B. D., Goldfarb, C. A. 2019. Risk factors for surgical site infection after orthopaedic surgery in the ambulatory surgical center setting. J Am Acad Orthop Surg. 27(20): e928-e934.

Burke, J. P. 2001. Maximizing appropriate antibiotic prophylaxis for surgical patients: An

update from LDS Hospital, Salt Lake City. Clin Infect Dis. 33(Suppl 2): S78–S83. Chauveaux, D. 2015. Preventing surgical-site infections : Measures other than

antibiotics. Orthop Traumatol Surg Res. 101 (1 Suppl): S77-S83.

Chen, A. F., Fleischman, A., Austin, M. S. 2018. Use of intrawound antibiotics in orthopaedic surgery. J Am Acad Orthop Surg. 26(17): e371-378.

Cheng, S., Chang, S., Tsai, T., Wu, C., Lin, W., Lin, W., Cheng, C. 2013. Differential expression of distinct surface markers in early endothelial progenitor cells and monocyte-derived macrophages. Gene Expr. 16(1): 15-24.

Chilukuri, D. M., Shah, J. C. 2005. Local delivery of vancomycin for the prophylaxis of prosthetic device-related infections. Pharm Res. 22(4): 563-572.

Chong, M. S. K., Ng, W. K. & Chan, J. K. Y. 2016. Concise review: Endothelial

progenitor cells in regenerative medicine: Applications and challenges. Stem Cells Transl Med. 5(4): 530-538.

Chopra, H., Hung, M. K., Kwong, D. L., Zhang, C. F., Pow, E. H. N. 2018. Insights into

endothelial progenitor cells: Origin, classification, potentials, and prospects. Stem Cells Int. 2018: 9847015.

Chow, L., Johnson, V., Impastato, R., Coy, J., Strumpf, A., Dow, S. 2020. Antibacterial

activity of human mesenchymal stem cells mediated directly by constitutively secreted factors and indirectly by activation of innate immune effector cells. Stem Cells Transl Med. 9(2): 235-249.

Chusri, S., Sompetch, K., Mukdee, S., Jansrisewangwong, S., Srichai, T., Maneenoon,

K., Limsuwan, S., Voravuthikunchai, P. 2012. Inhibition of Staphylococcus epidermidis biofilm formation by traditional Thai herbal recipes used for wound treatment. Evid Based Complement Alternat Med. 2012: 159797.

85

Claessens, J., Roriz, M., Merckx, R., Baatsen, P., Van Mellaert, L., Van Eldere, J. 2015. Inefficacy of vancomycin and teicoplanin in eradicating and killing Staphylococcus epidermidis biofilms in vitro. Int J Antimicrob Agents. 45(4): 368-375.

Copuroglu, C., Calori, G. M. & Giannoudis, P. V. 2013. Fracture non-union: Who is at risk? Injury. 44(11): 1379-1382.

Costerton, J. W., Lewandowski, Z., Caldwell, D. E., Korber, D. R. & Lappin-Scott, H. M. 1995. Microbial biofilms. Annu Rev Microbiol. 49: 711–745.

Dapunt, U., Spranger, O., Gantz, S., Burckhardt, I., Zimmermann, S., Schmidmaier, G., Moghaddam, A. 2015. Are atrophic long-bone nonunions associated with low-grade infections? Ther Clin Risk Manag. 11: 1843-1852.

Darouiche, R. O. & Hamill, R. J. 1994. Antibiotic penetration of and bactericidal activity within endothelial cells. Antimicrob Agents Chemother. 38(5): 1059-1064.

De Beer, J., Petruccelli, D., Rotstein, C., Weening, B., Royston, K., Winemaker, M. 2009. Antibiotic prophylaxis for total joint replacement surgery: Results of a survey of Canadian orthopedic surgeons. Can J Surg. 52(6): E229-234.

Debarge, R., Nicolle, M. C., Pinaroli, A., Ait Si Selmi, T., Neyret, P. 2007. Surgical site infection after total knee arthroplasty: A monocenter analysis of 923 first-intention implantations. Rev Chir Orthop Reparatrice Appar Moteur. 93(6): 582–587.

Dhammi, I. K., Haq, R. U. & Kumar, S. 2015. Prophylactic antibiotics in orthopaedic

surgery: Controversial issues in its use. Indian J Orthop. 49(4): 373-376.

Dudareva, M., Barrett, L., Figtree, M., Scarborough, M., Watanabe, M., Newnham, R., Wallis, R., Oakley, S., Kendrick, B., Stubbs, D., McNally, M. A., Bejon, P., Atkins, B. A., Taylor, A., Brent, A. J. 2018. Sonication versus tissue sampling for diagnosis of prosthetic joint and other orthopedic device-related infections. J Clin Microbiol. 56(12): e00688-18.

Durmuşlar, M. C., Balli, U., Türer, A., Önger, M. E., Çelik, H. H. 2016. Radiological and stereological evaluation of the effect of rifampin in bone healing in critical-size defects. J Craniofac Surg. 27(6): 1481-1485.

Edin, M. L., Miclau, T., Lester, G. E., Lindsey, R. W., Dahners, L. E. 1996. Effect of

cefazolin and vancomycin on osteoblasts in vitro. Clin Orthop. 333: 245-251. Edmiston Jr, C. E., Seabrook, G. R., Cambria, R. A., Brown, K. R., Lewis, B. D.,

Sommers, J. R., Krepel, C. J., Wilson, P. J., Sinski, S., Towne, J. B. 2005. Molecular epidemiology of microbial contamination in the operating room environment: Is there a risk for infection? Surgery. 138(4): 573–579.

86

Einhorn, T. A. & Gerstenfeld, L. C. 2015. Fracture healing: Mechanisms and interventions. Nat Rev Rheumatol. 11(1): 45-54.

Evan, R. P., Clyburn, T. A., Moucha, C. S., Prokuski, L. 2011. Surgical site infection prevention and control: An emerging paradigm. Instr Course Lect. 60: 539-543.

Evangelopoulos, D. S., Stathopoulos, I. P., Morassi, G. P., Koufos, S., Albami, A., Karampinas, P. K., Stylianakis, A., Kohl, S., Pneumaticos, S., Vlamis, J. 2013. Sonication: A valuable technique for diagnosis and treatment of periprosthetic joint infections. Sci World J. 2013: 375140.

Fang, J., Guo, Y., Tan, S., Li, Z., Xie, H., Chen, P., Wang, K., He, Z., He, P., Ke, Y.,

Jiang, X., Chen, Z. 2018. Autologous endothelial progenitor cells transplantation for acute ischemic stroke: A 4-year follow-up study. Stem Cells Transl Med. 8(1): 14-21.

Fedorovich, N. E., Haverslag, R. T., Dhert, W. J. A., Alblas, J. 2010. The role of

endothelial progenitor cells in prevascularized bone tissue engineering: Development of heterogenous constructs. Tissue Eng Part A. 16(7): 2355-2367.

Fernicola, S. D., Elsenbeck, M. J., Grimm, P. D., Pisano, A. J., Wagner, S. C. 2020. Intrasite antibiotic powder for the prevention of surgical site infection in extremity surgery: A systematic review. J Am Acad Orthop Surg. 28(1): 37-43.

Fey, P. D. & Olson, M. E. 2010. Current concepts in biofilm formation of Staphylococcus epidermidis. Future Microbiology. 5(6): 917-933.

Fleischman, A. N. & Austin, M. S. 2017. Local intra-wound administration of powdered antibiotics in orthopaedic surgery. J Bone Jt Infect. 2(1): 23-28.

Florschutz, A. V., Fagan, R. P., Matar, W. Y., Sawyer, R. G., Berrios-Torres, S. I. 2015. Surgical site infection risk factors and risk stratification. J Am Acad Orthop Surg. 23 Suppl: S8-S11.

Galandiuk, S., Polk Jr, H. C., Jagelman, D. G., Fazio, V. W. 1989. Re-emphasis of

priorities in surgical antibiotic prophylaxis. Surg Gynecol Obstet. 169(3): 219–222.

Gans, I., Jain, A., Sirisreetreerux, N., Haut, E. R., Hasenboehler, E. A. 2017. Current practice of antibiotic prophylaxis for surgical fixation of closed long bone fractures: A survey of 297 members of the Orthopaedic Trauma Association. Patient Saf Surg. 11: 2.

Gaston, M. S. & Simpson, A. H. R. W. 2007. Inhibition of fracture healing. J Bone Joint Surg Br. 89(12): 1553-1560.

87

Gerstenfeld, L. C., Alkhiary, Y. M., Krall, E. A., Nicholls, F. H., Stapleton, S. N., Fitch, J. L., Bauer, M., Kayal, R., Graves, D. T., Jepsen, K. J., Einhorn, T. A. 2006. Three-dimensional reconstruction of fracture callus morphogenesis. J Histochem Cytochem. 54(11): 1215-1228.

Gerstenfeld, L. C., Cullinane, D. M., Barnes G. L., Graves, D. T., Einhorn, T. A. 2003. Fracture healing as a post-natal developmental process: Molecular, spatial, and temporal aspects of its regulation. J Cell Biochem. 88(5): 873–884.

Giannoudis, P. V., Einhorn, T.A. & Marsh, D. 2007. Fracture healing: the diamond concept. Injury. 38 Suppl 4:S3-S6.

Giles, E. M., Godbout, C., Chi, W., Glick, M. A., Lin, T., Li, R., Schemitsch, E. H., Nauth, A. 2017. Subtypes of endothelial progenitor cells affect healing of segmental bone defects differently. Int Orthop. 41(11): 2337-2343.

Gristina AG. Implant failure and the immuno-incompetent fibro- inflammatory zone. Clin Orthop Relat Res 1994(298):106–18.

Gugala, Z. & Gogolewski, S. 1999. Regeneration of segmental diaphyseal defects in sheep tibiae using resorbable polymeric membranes: A preliminary study. J Orthop Trauma. 13(3):187–195.

Hak, D. J., Fitzpatrick, D., Bishop, J. A., Marsh, J. L., Tilp, S., Schnettler, R., Simpson, H., Volker, A. 2014. Delayed union and nonunions: epidemiology, clinical issues, and financial aspects. Injury. 45 Suppl 2: S3-S7.

Hansen, E., Belden, K., Silibovsky, R., Vogt, M., Arnold, W. V., Bicanic, G., Bini, S. A., Catani, F., Chen, J., Ghazavi, M. T., Godefroy, K. M., Holham, P., Hosseinzadeh, H., Kim, K. I. I., Kirketerp-Møller, Lidgren, L., Lin, J. H., Lonner, J. H., Moore, C. C., Papagelopoulos, P., Poultsides, L., Randall, R. L., Roslund, B., Saleh, K., Salmon, J. V., Schwarz, E. M., Stuyck, J., Dahl, A. W., Yamada, K. 2014. Perioperative antibiotics. J Arthroplasty. 29(2 Suppl): 29-48.

Hanssen, A. D., Osmon, D. R. & Nelson, C. L. 1997. Prevention of deep periprosthetic

joint infection. Instr Course Lect. 46: 555-567. Hernandez, R. K., Do, T. P., Critchlow, C. W., Dent, R. E., Jick, S. S. 2012. Patient-

related risk factors for fracture-healing complications in the United Kingdom General Practice Research Database. Acta Orthop. 83(6): 653-660.

Howlin, R. P., Brayford, M. J., Webb, J. S., Cooper, J. J., Aiken, S. S., Stoodley, P.

2015. Antibiotic-loaded synthetic calcium sulfate beads for prevention of bacterial colonization and biofilm formation in periprosthetic infections. Antimicrob Agents Chemother. 59(1): 111-120.

88

Huiras, P., Logan, J. K., Papadopoulos, S., Whitney, D. 2012. Local antimicrobial administration for prophylaxis of surgical site infections. Pharmacotherapy. 32(11): 1006-1019.

Hollinger, J. O. & Kleinschmidt, J. C. 1990. The critical size defect as an experimental model to test bone repair materials. J Craniofac Surg. 1(1): 60-68.

Isefuku, S., Joyner, C. J. & Simpson, A. H. 2001. Toxic effect of rifampicin on human osteoblast-like cells. J Orthop Res. 19(5): 950-954.

Kaderly, R. E. 1991. Primary bone healing. Semin Vet Med Surg (Small Anim). 6(1): 21–25.

Kanakaris, N. K., Giannoudis, P. V. 2007. The health economics of the treatment of long-bone non-unions. Injury. 38 Suppl 2: S77-S84.

Kannan, R. Y., Salacinski, H. J., Sales, K., Butler, P., Seifalian, A. M. 2005. The roles of

tissue engineering and vascularization in the development of micro-vascular networks: a review. Biomaterials. 26(14): 1857-75.

Kapadia, B. H., Johnson, A. J., Daley, J. A., Issa, K., Mont, M. A. 2013. Pre-admission cutaneous chlorhexidine preparation reduces surgical site infections in total hip arthroplasty. J Arthroplasty. 28(3): 490–493.

Keating, J. F., Simpson, A. H. R. W., Robinson, C. M. 2005. The management of fractures with bone loss. J Bone Joint Surg Br. 87(2): 142-150.

Kluin, O. S., Busscher, H. J., Neut, D., Van der Mei, H. C. 2016. Poly(trimethylene carbonate) as a carrier for rifampin and vancomycin to target therapy-recalcitrant staphylococcal biofilms. J Orthop Res. 34(10): 1828-1837.

Kon, T., Cho, T. J., Aizawa, T., Yamakazi, M., Nooh, N., Graves, D., Gerstenfeld, L. C., Einhorn, T. A. 2001. Expression of osteoprotegerin, receptor activator of NF-kappaB ligand (osteoprotegerin ligand) and related proinflammatory cytokines during fracture healing. J Bone Miner Res. 16(6): 1004–1014.

Korol, E., Johnston, K., Waser, N., Sifakis, F., Jafri, H. S., Lo, M., Kyaw, M. H. 2013. A systematic review of risk factors associated with surgical site infections among surgical patients. PLoS One. 8(12): e83743.

Lara-Hernandez, R., Lozano-Vilardell, P., Blanes, P., Torreguitart-Mirada, N., Galmés, A., Besalduch, J. 2010. Safety and efficacy of therapeutic angiogenesis as a novel treatment in patients with critical limb ischemia. Ann Vasc Surg. 24(2): 287-294.

89

Lee, N. G., Jeung, I. C., Heo, S. C., Song, J., Kim, W., Hwang, B., Kwon, M., Kim, Y., Lee, J., Park, J., Shin, M., Cho, Y., Son, M., Bae, K., Lee, S., Kim, J., Min, J. 2019. Ischemia-induced Netrin-4 promotes neovascularization through endothelial progenitor cell activation via Unc-5 Netrin receptor B. The FASEB Journal. 34: 1231-1246.

Lee, S. K. & Lorenzo, J. 2006. Cytokines regulating osteoclast formation and function. Curr Opin Rheumatol. 18(4): 411-418.

Lewis, K. 2010. Persister cells. Annu Rev Microbiol. 64: 357–372.

Li, R., Atesok, K., Nauth, A., Wright, D., Qamirani, E., Whyne, C. M., Schemitsch, E. H. 2011. Endothelial progenitor cells for fracture healing: A microcomputed tomography and biomechanical analysis. J Orthop Trauma. 25(8): 467-471.

Li, R., Nauth, A., Gandhi, R., Syed, K., Schemitsch, E. H. 2014. BMP-2 mRNA expression after endothelial progenitor cell therapy for fracture healing. J Orthop Trauma. 28 Suppl 1: S24-S27.

Li, R., Nauth, A., Li, C. H., Qamirani, E., Atesok, K., Schemitsch, E. H. 2012. Expression of VEGF gene isoforms in a rat segmental bone defect model treated with EPCs. J Orthop Trauma. 26(12): 689-692.

Li, R., Stewart, D. J., von Schroeder, H. P., Mackinnon, E. S., Schemitsch, E. H. 2009. Effect of cell-based VEGF gene therapy on healing of a segmental bone defect. J Orthop Res. 27(1): 8-14.

Lidwell, O. M., Lowbury, E.J., Whyte W., Blowers, R., Stanley, S. J., Lowe, D. 1983. Airborne contamination of wounds in joint replacement operations: the relationship to sepsis rates. J Hosp Infect. 4(2): 111-131.

Lidwell, O. M., Lowbury, E. J., Whyte, W., Blowers, R., Stanley, S. J., Lowe, D. 1984. Infection and sepsis after operations for total hip or knee-joint replacement: Influence of ultraclean air, prophylactic antibiotics and other factors. J Hyg (Lond). 93(3): 505–529.

Liu, C., Bayer, A., Cosgrove, S. E., Daum, R. S., Fridkin, S. K., Gorwitz, R. J., Kaplan, S. L., Karchmer, A. W., Levine, D. P., Murray, B. E., Rybak, M. J. Talan, D. A., Chambers, H. F. 2011. Clinical practice guidelines by the Infectious Diseases Society of America for the treatment of methicillin-resistant Staphylococcus aureus infections in adults and children. Clinical Infectious Diseases. 52(3): e18-e55.

90

Lovati, A. B., Romanò, C. L., Bottagisio, M., Monti, L., De Vecchi, E., Previdi, S., Accetta, R., Drago, L. 2016. Modelling Staphylococcus epidermidis-induced non-unions: Subclinical and clinical evidence in rats. PLoS One. 11(1): e0147447.

Mahlapuu, M., Håkansson, J., Ringstad, L., Björn, C. 2016. Antimicrobial peptides: an

emerging category of therapeutic agents. Front Cell Infect Microbiol. 6: 194. Mangram, A. J., Horan, T. C., Pearson, M. L., Silver, L. C., Jarvis, W. R. 1999.

Guidelines for prevention of surgical site infection, 1999. Centers for Disease Control and Prevention (CDC) Hospital Infection Control Practices Advisory Committee. Am J Infect Control. 27(2): 97-132.

Maniar, H. H., Wingert, N., McPhillips, K., Foltzer, M., Graham, J., Bowen, T. R., Horwitz, D. S. 2016 Role of sonication for detection of infection in explanted orthopaedic trauma implants. J Orthop Trauma. 30(5): e175-e180.

Marsell, R., Einhorn, T. A. 2011. The biology of fracture healing. Injury. 42(6): 551-555.

Mathur, P., Trikha, V., Farooque, K., Sharma, V., Jain, N., Bhardwaj, N., Sharma, S., Misra, M. C. 2013. Implementation of a short course of prophylactic antibiotic treatment for prevention of postoperative infections in clean orthopaedic surgeries. Indian J Med Res. 137(1): 111–116.

McDonald, M., Grabsch, E., Marshall, C., Forbes, A. 1998. Single- versus multiple-dose antimicrobial prophylaxis for major surgery: A systematic review. Aust N Z J Surg. 68(6): 388–396.

Medina, R. J., Barber, C. L., Sabatier, F., Dignat-George, F., Melero-Martin, J. M., Khosrotehrani, K., Ohneda, O., Randi, A. M., Chan, J. K. Y., Yamaguchi, T., Van Hinsbergh, V. W. M., Yoder, M. C., Stitt, A. W. 2017. Endothelial progenitors: A consensus statement on nomenclature. Stem Cells Transl Med. 6(5): 1316-1320.

Mendoza, M. C., Sonn, K. A., Kannan, A. S., Bellary, S. S., Mitchell, S. M., Singh, G.,

Park, C., Yun, C., Stock, S. R., Hsu, E. L., Hsu, W. K. 2016. The effect of vancomycin powder on bone healing in a rat spinal rhBMP-2 model. J Neurosurg Spine. 25(2): 147-153.

Metsemakers, W. J., Morgenstern, M., McNally, M. A., Moriarty, T. F., McFadyen, I., Scarborough, M., Athanasou, N. A., Ochsner, P. E., Kuehl, R., Raschke, M., Borens, O., Xie, Z., Velkes, S., Hungerer, S., Kates, S. L., Zalavras, C., Giannoudis, P. V., Richards, R. G., Verhofstad, M. H. J. 2018. Fracture-related infection: A consensus on definition from an international expert group. Injury. 49(3): 505-510.

91

Mills, L., Tsang J., Hopper, G., Keenan, G., Simpson, A. H. R. W. 2016. The multifactorial aetiology of fracture nonunion and the importance of searching for latent infection. Bone Joint Res. 5(10): 512-519.

Nauth, A., Lee, M., Gardner, M. J., Brinker, M. R., Warner, S. J., Tornetta 3rd, P.,

Leucht, P. 2018. Principles of nonunion management: state of the art. J Orthop Trauma. 32 Suppl 1:S52-S57.

Nelson, C. L., Jones, R. B., Wingert, N. C., Foltzer, M., Bowen, T. R. 2014. Sonication of antibiotic spacers predicts failure during two-stage revision for prosthetic knee and hip infections. Clin Orthop Relat Res. 472(7): 2208-2214.

Nichols, R. L. 2004. Preventing surgical site infections. Clin Med Res. 2(2): 115-118.

Niska, J. A., Shahbazian, J. H., Ramos, R. I., Francis, K. P., Bernthal, N. M., Miller, L. S. 2013. Vancomycin-rifampin combination therapy has enhanced efficacy against an experimental Staphylococcus aureus prosthetic joint infection.

Olszewski D., Streubel P.N., Stucken C., Ricci, W. M., Hoffman, M. F., Jones, C. B.,

Sietsema, D. L., Tornetta 3rd, P. 2016. Fate of patients with a “surprise” positive culture after nonunion surgery. J Orthop Trauma. 30(1):e19–e23.

Osmon, D. R., Berbari, E. F., Berendt, A. R., Lew, D., Zimmerli, W., Steckelberg, J. M.,

Rao, N., Hanssen, A., Wilson, W. R., Infectious Diseases Society of America. 2013. Diagnosis and management of prosthetic joint infection: Clinical practice guidelines by the Infectious Diseases Society of America. Clin Infect Dis. 56(1):e1–e25.

Palombella, S., Lopa, S., Gianola, S., Zagra, L., Moretti, M., Lovati, A. B. 2019. Bone

marrow-derived cell therapies to heal long-bone nonunions: A systematic review and meta-analysis – Which is the best available treatment? Stem Cells Int. 2019: 3715964, 12 pgs.

Pagano, P. J., Buchanan, L. V., Dailey, C. F., Haas, J. V., Van Enk, R. A., Gibson, J. K.

2004. Effects of linezolid on staphylococcal adherence versus time of treatment. Int J Antimicrob Agents. 23(3): 226-234.

Park, K., Greenwood-Quaintance, K. E., Schuetz, A. N., Mandrekar, J. N., Patel, R.

2017. Activity of tedizolid in methicillin-resistant Staphylococcus epidermidis experimental foreign body-associated osteomyelitis. Antimicrob Agents Chemother. 61(2): e01644-16.

Pull ter Gunne, A. F. & Cohen, D. B. 2009. Incidence, prevalence, and analysis of risk

factors for surgical site infection following adult spinal surgery. Spine. 34(13):1422–1428.

92

Rabani, R., Volchuk, A., Jerkic, M., Ormesher, L., Garces-Ramirez, L., Canton, J., Masterson, C., Gagnon, S., Tatham, K. C., Marshall, J., Grinstein, S., Laffey, J. G., Szaszi, K., Curley, G. F. 2018. Mesenchymal stem cells enhance NOX2-dependent reactive oxygen species production and bacterial killing in macrophages during sepsis. Eur Respir J. 51(4): 1702021.

Rathbone, C. R., Cross, J. D., Brown, K. V., Murray, C. K., Wenke, J. C. 2011. Effect of

various concentrations of antibiotics on osteogenic cell viability and activity. J Orthop Res. 29(7): 1070-1074.

Ribiero, M., Monteiro, F. J. & Ferraz, M. P. 2012. Infection of orthopedic implants with

emphasis on bacterial adhesion process and techniques used in studying bacterial-material interactions. Biomatter. 2(4): 176-194.

Rupp, M., Biehl, C., Budak, M., Thormann, U., Heiss, C., Alt, V. 2018. Diaphyseal long bone nonunions – types, aetiology, economics, and treatment recommendations. Int Orthop. 42(2): 247-258.

Sadrizadeh, S., Tammelin, A., Ekolind, P., Holmberg, S. 2014. Influence of staff number and internal constellation on surgical site infection in an operating room. Particuology. 4(13): 42–51.

Saginur, R., Stdenis, M., Ferris, W., Aaron, S. D., Chan, F., Lee, C., Ramotar, K. 2006. Multiple combination bactericidal testing of staphylococcal biofilms from implant-associated infections. Antimicrob Agents Chemother. 50(1): 55-61.

Said, G. Z., Farouk, O. A., Said, H. G., El-Sharkawi, M. M. M. 2013. Non-anatomical surgical solutions for difficult non-unions: case series. Trauma Mon. 17(4): 404-408.

Salkind, A. R.& Rao, K. C. 2011. Antibiotic prophylaxis to prevent surgical site infections. Am Fam Physician. 83(5): 585-590.

Schemitsch, E. H. 2017. Size matters: Defining critical in bone defect size! J Orthop Trauma. 31 Suppl 5:S20-S22.

Schmal, H., Brix, M., Bue, M., Ekman, A., Ferreira, N., Gottlieb, H., Kold, S., Taylor, A.,

Tengberg, P. T., Ban, I. 2020. Nonunion – consensus from the 4th annual meeting of the Danish Orthopaedic Trauma Society. EFORT Open Rev. 5(1): 46-57.

Seebach, C., Henrich, D., Wilhelm, K., Barker, J. H, Marzi, I. 2012. Endothelial progenitor cells improve directly and indirectly early vascularization of mesenchymal stem cell-driven bone regeneration in a critical bone defect in rats. Cell Transplant. 21(8): 1667-1677.

93

Sewell, M., Adebibe, M., Jayakumar, P., Jowett, C., Kong, K., Vemulapalli, K., Levack, B. 2011. Use of the WHO surgical safety checklist in trauma and orthopaedic patients. Int Orthop. 35(6): 897-901.

Sfeir, C., Ho, L., Doll, B. A., Azari, K., Hollinger, J. O. 2007. Fracture repair. In Lieberman, J. R., Friedlaender, G. E. (Eds.). Bone Regeneration and Repair: Biology and Clinical Applications. Totowa, NJ: Humana Press; p. 21–44.

Shapiro, F. Cortical bone repair. The relationship of the lacunar-canalicular system and intercellular gap junctions to the repair process. J Bone Joint Surg Am. 70(7):1067–1081.

Sheen, J. R. & Garla, V. V. 2020. Fracture healing overview. StatPearls. Treasure

Island, FL: StatPearls Publishing. Accessed at: https://www.ncbi.nlm.nih.gov/books/NBK551678/.

Shiels, S. M. Tennent, D. J., Lofgren, A. L., Wenke, J. C. 2018. Topical rifampin powder

for orthopaedic trauma part II: topical rifampin powder allows for spontaneous bone healing in sterile and contaminated wounds. J Orthop Res. 36(12): 3142-3150.

Shiels, S. M., Tennent, D. J. & Wenke, J. C. 2018. Topical rifampin powder for orthopaedic trauma part I: rifampin powder reduces recalcitrant infection in a delayed treatment musculoskeletal trauma model. J Orthop Res. 36(12): 3136-3141

Somford, M. P., Van den Berkerom, M. P. J., Kloen, P. 2013. Operative treatment for femoral shaft nonunions, a systematic review of the literature. Strategies Trauma Limb Reconstruction. 8(2): 77-88.

Stewart, S. K. 2019. Fracture non-union: a review of clinical challenges and future research needs. Malays Orthop J. 13(2): 1-10.

Stoodley, P., Ehrlich, G. D., Sedghizadeh, P. P., Hall-Stoodley, L., Baratz, M. E.,

Altman, D. T., Sotereanos, N. G., Costerton, J. W., Demeo, P. 2011. Orthopaedic biofilm infections. Curr Orthop Pract. 22(6): 558-563.

Sutton, M. T., Fletcher, D., Episalla, N., Auster, L., Kaur, S., Gwin, M. C., Folz, M.,

Velasquez, D., Roy, V., Van Heeckeren, R., Lennon, D. P., Caplan, A. I., Bonfield, T. L. 2017. Mesenchymal stem cell soluble mediators and cystic fibrosis. J Stem Cell Res Ther. 7(9): 400.

Talon, D., Schoenleber, T., Bertrand, X., Vichard, P. 2006. Performances of different types of airflow system in operating theatre. Ann Chir. 131(5): 316–321.

94

Tennent, D. J., Shiels, S. M., Sanchez Jr, C. J., Niece, K. L., Akers, K. S., Stinner, D. J., Wenke, J. C. 2016. Time-dependent effectiveness of locally applied vancomycin powder in a contaminated traumatic orthopaedic wound model. J Orthop Trauma. 30(10): 531-537.

Thompson, J. M., Saini, V., Ashbaugh, A. G., Miller, R. J., Ordonez, A. A., Ortines, R.

V., Wang, Y., Sterling, R. S., Jain, S. K., Miller, L. S. 2017. Oral-only linezolid-rifampin is highly effective compared with other antibiotics for periprosthetic joint infection: Study of a mouse model. J Bone Joine Surg Am. 99(8): 656-665.

Thonse, R., Sreenivas, M., Sherman, K. P. 2004. Timing of antibiotic prophylaxis in surgery for adult hip fracture. Ann R Coll Surg Engl. 86(4): 263–266.

Tornero, E., García-Ramiro, S., Martínez-Pastor, J. C., Bori, G., Bosch, J., Morata, L., Sala, M., Basora, M., Mensa, J., Soriano, A. 2015. Prophylaxis with teicoplanin and cefuroxime reduces the rate of prosthetic joint infection after primary arthroplasty. Antimicrob Agents Chemother. 59(2): 831-837.

Trampuz, A. & Zimmerli W. 2006. Diagnosis and treatment of infections associated with fracture-fixation devices. Injury. 37 Suppl 2: S59-S66.

Uçkay, I., Harbarth, S., Peter, R., Lew, D., Hoffmeyer, P., Pittet, D. 2010. Preventing surgical site infections. Expert Rev Anti Infect Ther. 8(6): 657-670.

Uzun, M., Çakar, M., Bülbül, A. M., Kara, A. 2015. Treatment of aseptic hypertrophic

nonunion of the lower extremity with less invasive stabilization system (new approach to hypertrophic nonunion treatment). Adv Orthop. 2015: 631254.

Wang, X., Friis, T., Glatt, V., Crawford, R., Xiao, Y. 2017. Structural properties of fracture haematoma: Current status and future clinical implications. J Tissue Eng Regen Med. 11(10): 2864-2875.

Wang, W. & Yeung, K. W. K. 2017. Bone grafts and biomaterials substitutes for bone

defect repair: A review. Bioact Mater. 2(4): 224-247.

Weber, B. G. & Cech, O. Pseudarthrosis. New York: Grune and Stratton. 1976. Werner, E., Roe, F., Bugnicourt, A., Franklin, M. J., Heydorn, A., Molin, S., Pitts, B. &

Stewart, P. S. 2004. Stratified growth in Pseudomonas aeruginosa biofilms. Appl Environ Microbiol. 70(10): 6188–6196.

Whitehouse, J. D., Friedman, N. D., Kirkland, K. B., Richardson, W. J., Sexton, D. J. 2002. The impact of surgical-site infections following orthopaedic surgery at a community hospital and a university hospital: adverse quality of life, excess length of stay, and extra cost. Infect Control Hosp Epidemiol. 23(4): 183-189.

95

Whyte, W., Hodgson, R. & Tinkler, J. 1982. The importance of airborne bacterial contamination of wounds. J Hosp Infect. 3(2): 123-135.

Winkler, H. 2017. Treatment of chronic orthopaedic infection. EFORT Open Rev. 2(5): 110-116.

Yang, X., Ricciardi, B.F., Hernandez-Soria, A., Shi, Y., Camacho, N. P., Bostrom, M. P. G. 2007. Callus mineralization and maturation are delayed during fracture healing in interleukin-6 knockout mice. Bone. 41(6): 928–936.

Zhang, Y., Ingram, D. A., Murphy, M. P., Saadatzadeh, M. R., Mead, L. E., Prater, D. N., Rehman, J. 2009. Release of proinflammatory mediators and expression of proinflammatory adhesion molecules by endothelial progenitor cells. Am J Physiol Heart Circ. 296: H1675-H1682.

Zhang, Z., Wang, X., Luo, F., Yang, H., Hou, T., Zhou, Q., Dai, F., He, Q., Xu, J. 2014. Effects of rifampicin on osteogenic differentiation and proliferation of human mesenchymal stem cells in the bone marrow. Genet Mol Res. 13(3): 6398-6410.

Zhu, J., Song, J., Yu, L., Zheng, H., Zhou, B., Weng, S., Fu, G. 2016. Safety and efficacy of autologous Thymosin B4 pre-treated endothelial progenitor cell transplantation in patients with acute ST segment elevation myocardial infarction: a pilot study. Cytotherapy. 18(8): 1037-1042.

Zimmerli, W. & Moser, C. 2012. Pathogenesis and treatment concepts of orthopaedic biofilm infections. FEMS Immunol Med Microbiol. 65(2): 158-168.

Zura, R., Mehta, S., Della Rocca, G. J., Grant Steen, R. 2016. Biological risk factors for nonunion of bone fracture. JBJS Rev. 4(1): 01874474-201601000-00005.

The Effects of Endothelial Progenitor Cell Therapy on Fracture ...

Documents