PEPTIDE-BASED SYSTEMS FOR THE TARGETED DISRUPTION AND
TREATMENT OF STAPHYLOCOCCUS EPIDERMIDIS BIOFILMS
by
CHRISTOPHER MICHAEL HOFMANN
Submitted in partial fulfillment of the requirements
For the degree of Doctor of Philosophy
Thesis Advisor: Dr. Roger E. Marchant
Department of Biomedical Engineering
CASE WESTERN RESERVE UNIVERSITY
May, 2012
CASE WESTERN RESERVE UNIVERSITY SCHOOL OF GRADUATE STUDIES
We hereby approve the thesis/dissertation of
Christopher Michael Hofmann
Candidate for the Doctor of Philosophy degree *.
(signed) Roger E. Marchant (chair of the committee)
James M. Anderson
Anirban Sen Gupta
Brian Cobb
(date) March 29, 2012
*We also certify that written approval has been obtained for any proprietary material contained therein.
To my wife and parents, who taught me that anything is possible.
TABLE OF CONTENTS
TABLE OF CONTENTS 1
LIST OF TABLES 4
LIST OF FIGURES 5
ACKNOWLEDGEMENTS 7
ABSTRACT 8
CHAPTER 1: INTRODUCTION 10 1.1 Clinical Incidence of Infection 10 1.2 Pathogenicity of Implanted Medical Device Infections 11 1.3 Staphylococcus epidermidis Biofilms 13 1.3.1 The Bacteria 13 1.3.2 Primary Adhesion to Surfaces 18 1.3.2.1 Nonspecific Binding 18 1.3.2.2 Bacterial-Mediated Nonspecific Adhesion 23 1.3.2.3 Bacterial-Mediated Specific Binding 24 1.3.2.3.1 Role of Bacterial Polysaccharides as Ligands 25 1.3.2.3.2 Role of Bacterial Proteins as Ligands 26 1.3.3 Biofilm Formation 29 1.3.3.1 Polysaccharide Intercellular Adhesin (PIA) 29 1.3.3.2 Accumulation Associated Protein (AAP) 33 1.3.3.3 Other Matrix Components (eDNA, Bhp, EC TA) 35 1.3.3.4 Quorum Sensing System (Agr, LuxS) 37 1.4 The Host Response 40 1.5 References 45
1
CHAPTER 2: BACTERIAL RESISTANCE TO ANTIBIOTICS 71 2.1 Antibiotics and Resistance to Treatment 71 2.2 Overview of Antibiotics 72 2.3 Antibiotic Resistance 80 2.4 The Biofilm as a Source of Protection 87 2.5 References 98
CHAPTER 3: STRATEGIES FOR PREVENTING AND TREATING 109 DEVICE INFECTIONS 3.1 Modifications of the Biomaterial 109 3.1.1 Surface Modifications of Materials to Prevent Adhesion 109 3.1.2 Incorporation of Antimicrobial Agents 116 3.2 Specific Targeting Strategies 127 3.1.1 Bacteriophages 127 3.1.2 Antibodies and Opsonization 130 3.3 Specific Aims and Hypothesis 133 3.4 References 135
CHAPTER 4: FIBRINOGEN-BASED LIGAND FOR SPECIFIC 143 TARGETING AND DELIVERY TO SURFACE-ADHERENT S. EPIDERMIDIS 4.1 Introduction 143 4.2 Materials and Methods 146 4.3 Results 151 4.4 Discussion 158 4.5 Conclusions 163 4.6 Future Directions 164 4.7 Acknowledgements 167 4.8 References 168
2
CHAPTER 5: DISRUPTION OF S. EPIDERMIDIS BIOFILM 173 FORMATION USING A TARGETED CATIONIC PEPTIDE 5.1 Introduction 173 5.2 Materials and Methods 175 5.3 Results 178 5.4 Discussion 183 5.5 Conclusions 187 5.6 Acknowledgements 188 5.7 References 189
CHAPTER 6: TARGETED DELIVERY OF VANCOMYCIN TO 197 S. EPIDERMIDIS BIOFILMS USING A FIBRINOGEN- DERIVED PEPTIDE 6.1 Introduction 197 6.2 Materials and Methods 200 6.3 Results 207 6.4 Discussion 212 6.5 Conclusions 218 6.6 Acknowledgements 218 6.7 References 219
CHAPTER 7: CONCLUSIONS AND PERSPECTIVES 227 7.1 Summary of Completed Work 227 7.2 Targeted Vancomycin: Synthesis Considerations 229 7.3 Future Directions - Flow System 246 7.4 Acknowledgements 246 7.4 References 250
CHAPTER 8: BIBLIOGRAPHY 253
3
LIST OF TABLES
TABLE 2.1 Clinically Administered Antibiotics 73
TABLE 2.2 Modes of Antibiotic Resistance 82
TABLE 6.1 MIC and MBC Values for vancomycin, 208 6-20-PEG3400-VAN, and 6-20-PEG5000-Van
4
LIST OF FIGURES
FIGURE 1.1 Schematic of biofilm formation process on 12 cardiovascular biomaterials
FIGURE 1.2 Structure of Polysaccharide Intercellular Adhesin (PIA) 32
FIGURE 2.1 Binding interaction between vancomycin and 86 Peptidoglycan Acyl-D-Ala-D-Ala
FIGURE 4.1 6-20-NG synthesis scheme 147
FIGURE 4.2 Analysis of 6-20 purity by MALDI-TOF mass 152 spectrometry and RP-HPLC
FIGURE 4.3 Blocking non-specific 6-20-NG adhesion to substrate 153
FIGURE 4.4 Normalized 6-20-NG adhesion to PET and S. epidermidis 155
FIGURE 4.5 Scanning electron microscope images of peptide 156 blocking studies
FIGURE 4.6 Peptide blocking of 6-20-NG to S. epidermidis 157
FIGURE 5.1 Optical density growth curves of surface-adherent 179 S. epidermidis
FIGURE 5.2 Effects of peptide on the composition of surface-adherent 181 biofilms as determined by quantitative fluorescence microplate readings
FIGURE 5.3 Biofilm structure after 21 hours as observed by 182 SEM and PIA staining with wheat germ agglutinin
FIGURE 5.4 Biofilm composition after 21 hours as determined by the 184 XTT metabolic assay
FIGURE 6.1 Chemical structure of vancomycin 202
FIGURE 6.2 Two-step reaction scheme for synthesis of 6-20-PEGX-VAN 203
5
FIGURE 6.3 Retention of 6-20-PEG3400-VAN, 6-20-PEG5000-VAN, 210 and vancomycin by S. epidermidis
FIGURE 6.4 Retention of targeted antibiotics by 24 hour 216 S. epidermidis biofilms as determined by indirect ELISA
FIGURE 7.1 MALDI-TOF of vancomycin 231
FIGURE 7.2 MALDI-TOF of SM-PEG(12)-Vancomycin synthesized 232 in aqueous conditions (PBS, pH 7.4)
FIGURE 7.3 MALDI-TOF of SM-PEG(2)-Vancomycin synthesized 234 in organic conditions (DMF w/ TFA)
FIGURE 7.4 MALDI-TOF of SM-PEG(12)-Vancomycin synthesized 235 in organic conditions (DMF w/ TFA)
FIGURE 7.5 MALDI-TOF of 6-20-PEG(2)-Vancomycin synthesized 237 in organic conditions (DMF w/ TFA)
FIGURE 7.6 pKa values of vancomycin 239
FIGURE 7.7 Charge vs. pH plot for vancomycin and PEGx-Vancomycin 240
FIGURE 7.8 Proposed synthesis scheme for soluble 242 6-20-PEGx-Vancomycin
FIGURE 7.9 Proposed synthesis scheme for on-resin modification of 244 6-20 peptide and attachment of MAL-PEGx-NHS crosslinker
FIGURE 7.10 MALDI-TOF of MAL-PEG(2)-6-20 synthesized 245 according to scheme in Figure 7.8
6
ACKNOWLEDGEMENTS
The Marchant lab has been a wonderful place to learn, and without its many members, past and present, I could never have finished this project. I greatly appreciate the freedom Dr. Marchant allowed me in developing my project, as it allowed me to truly take ownership of the work and its outcome. Former graduate students Sharon Sagnella, Eric Anderson, Arya Kumar, Coby Larsen, and Jeff Beamish were great sources of knowledge and insight as I started out on this journey, while current graduate students Lynn Dudash, Lin Lin, Derek
Jones, and Jenny Bastijanic have been great sources of support as I worked to wrap things up. I would especially like to thank Kyle Bednar, who worked with me during his undergraduate time at Case. Without his contributions this work would have taken far longer, and I wish him the best of luck as he now pursues his own Ph.D. at the University of Cincinnati.
Most importantly, I want to thank my wife Chrissy for her constant support and encouragement over the years. Her enthusiasm and motivation served as a constant source of inspiration, giving me the strength to keep moving forward even when things weren’t looking so bright.
Such an accomplishment would not have been possible without the help and support of countless other people over the years. To those that I wasn’t able to thank by name, please know that I will be forever grateful for everything you have done on my behalf.
7
Peptide-Based Systems for the Targeted Disruption and Treatment of Staphylococcus epidermidis Biofilms
Abstract
by
CHRISTOPHER MICHAEL HOFMANN
Complications due to nosocomial infections of implanted medical devices pose a significant health risk to patients, with Staphylococcus epidermidis often implicated in the case of blood-contacting biomaterials. One method by which S. epidermidis initially adheres to biomaterials uses the plasma protein fibrinogen as an intermediary, where the S. epidermidis surface protein SdrG binds to a short amino acid sequence near the amino terminus of the B chain of fibrinogen. This study reports on the use of this fibrinogen-derived 6-20 peptide for the targeted disruption of S. epidermidis biofilm formation using a cationic peptide, as well as the specific delivery of vancomycin to S. epidermidis biofilms.
S. epidermidis virulence relies mainly upon its ability to form a biofilm, the main component of which is polysaccharide intercellular adhesin (PIA). The synthetic 6-20 peptide was utilized to deliver a cationic polylysine peptide (G3K6) to the bacterial surface and disrupt the charge-charge interactions needed for PIA retention and biofilm stability. The effects of the 6-20-G3K6 peptide on biofilm formation were assessed using optical density, fluorescently labeled wheat germ
8
agglutinin, nucleic acid stain (SYTO 9), and a metabolic assay (XTT, 2,3-bis(2- methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide inner salt).
Biofilms formed in the presence of 6-20-G3K6 peptide (100 M) resulted in a
37.9% reduction in PIA content and a 17.5% reduction of adherent bacteria relative to biofilms grown in the absence of peptide. These studies demonstrate the targeting ability of the 6-20 peptide towards biomaterial-adherent S. epidermidis, and highlight the potential for disrupting the early stages of biofilm formation.
Targeted 6-20-PEGx-VAN vancomycin derivatives were then synthesized using a flexible, variable length poly(ethylene glycol) linker between the peptide and antibiotic. Initial binding to surface adherent S. epidermidis was increased in a concentration-dependent manner relative to vancomycin for all equivalent concentrations ≥4 g/ml, with targeted vancomycin content up to 22.9 times that of vancomycin alone. Retention of the targeted antibiotics was measured after an additional 24 hour incubation period, revealing levels 1.3 times that of vancomycin. The results demonstrate the improved targeting and retention of vancomycin within a biofilm due to the incorporation of a specific targeting motif.
9
CHAPTER 1
INTRODUCTION
1.1 CLINICAL INCIDENCE OF INFECTION
There are close to 40 million surgical procedures involving the insertion of
artificial devices performed each year in the United States[1]. Of these implanted
devices, close to 1 million will become infected, accounting for nearly 45% of the total nosocomial infections reported annually[2, 3]. Approximately 835,700 cardiovascular devices are implanted each year, including mechanical heart valves, vascular grafts, pacemaker-defibrillators, and ventricular assist devices[1,
2], of which 31,680 devices are projected to become infected[2]. While the overall incidence of infection is relatively low at just under 4%, the associated morbidity/mortality rate can exceed >25%. In addition to the health risks, infected medical devices have a significant financial impact on patients. The average cost of medical and surgical treatments for infected cardiovascular devices ranges from $35,000 to $50,000, with an estimated annual cost of $1.2 billion[2]. The total annual cost of treatment for all nosocomial infections (both device and non-device related) is estimated to be $11 billion[3].
10
1.2 PATHOGENICITY OF IMPLANTED MEDICAL DEVICE INFECTIONS
Staphylococci, enterococci, enterobacteriaceae, and Candida spp. are common
pathogens associated with infections of indwelling medical devices[4], with the likelihood of infection, as well as the organism implicated in the infection, greatly
dependent upon the type and location of the implant itself[4]. In the case of
intravascular implants, coagulase-negative staphylococci (CoNS), particularly
Staphylococcus epidermidis, are the most common cause of infection[1, 5].
Staphylococcus epidermidis is a naturally occurring commensal organism found
on the surface of the human body, making up the majority of the bacterial
microflora[6]. The only infection known to be caused by the typically
noninvasive S. epidermidis in the immunocompetent adult is native valve
endocarditis; all other infections require the presence of a foreign body[7]. It is
likely that introduction of these microorganisms to an implanted device occurs at
the time of surgery, as it has been found that 5,000 to 50,000 skin particles are
transferred daily from physicians skin flora in intensive care units, and even
under aseptic conditions 90% of clean wounds contain pathogenic bacteria.[3]
S. epidermidis infections rely upon the ability of the bacteria to adhere to a
surface and successfully form a biofilm, as opposed to S. aureus infections which
are often associated with toxins and other virulence factors[5, 8]. A two-stage
11
•Mature biofilm •Leukocytes release oxidative/nitrogen and intermediates contentsgranular •Bacterial •Bacterial proliferation, aggregation, andslime formation iovascular biomaterials. Bacteria first adhere e and form an extracellular polysaccharide matrix ng which time leukocytes are able to phagocytose the Time •Specific bacterial to thrombus adhesion formations via protein intermediaries Biomaterial
•Platelet activation and adhesion •Leukocytes phagocytosis of bacteria S. epidermidis Leukocyte Activated Platelet Plasma Protein Plasma : Schematic of biofilm formation process on card •Plasma proteins proteins adsorb•Plasma onto biomaterial •Non-specific bacterial adhesion •Leukocyte chemotaxis and migration Figure 1.1 to surface-adsorbed proteins and/or platelets, duri bacteria. Following adhesion, the bacteria proliferat that protects them from the host defenses.
12
process can be used to describe the course of such foreign-body infections
(Figure 1.1), wherein initial adhesion of the bacteria to the material occurs first, followed by proliferation, matrix secretion, and cell-cell adhesion leading to the formation of a mature, multi-layered biofilm[5]. While planktonic S. epidermidis is known to be susceptible to a large number of antibiotics[9], the biofilm environment offers the encapsulated bacteria increased resistance, often times able to survive antibiotic concentrations several orders of magnitude higher than the minimum inhibitory concentration (MIC) measured in planktonic suspensions[10-12].
1.3 STAPHYLOCOCCUS EPIDERMIDIS BIOFILMS
1.3.1 The Bacteria
Staphylococcus epidermidis is a gram positive bacterium characterized by a low
G-C content and typically aggregating in grape-like clusters[8, 13]. As one of several gram positive, low G-C bacteria that share the cocci shape, further description is necessary in order to uniquely characterize S. epidermidis.
Staphylococcus test positive for the enzyme catalase, and as such can be easily distinguished from the similarly shaped and catalase-negative Streptococcus[13].
In order to distinguish Staphylococcus form Micrococcus, the oxidation- fermentation test is used. Micrococcus (an obligate aerobe) can only produce acid from glucose aerobically, while Staphylococcus (a falcultative aerobe), can produce acid from glucose both aerobically and anaerobically[13]. Finally, the
13
genus Staphylococcus is divided into two main groups based on expression of the
clotting enzyme coagulase. Staphylococcus epidermidis is non-pigmented and
coagulase-negative, whereas Staphylococcus aureus is yellow-pigmented and
coagulase-positive[8, 14].
As a gram positive bacterium, S. epidermidis is characterized by a thick cell
wall surrounding the cell’s cytoplasmic membrane. The cytoplasmic membrane
is an 8 nm thick phospholipid bilayer stabilized by hopanoids (similar to sterols
found in eukaryotic cells)[13]. The membrane acts as a diffusion barrier, and
along with active transport systems located within the membrane, allows the
bacteria to concentrate large amounts of dissolved solutes within the cell. The
resulting turgor pressure is substantial, equal to about 2 atmospheres in
Escherichia coli[13]. In order to counter this pressure, a thick peptidoglycan cell wall is constructed that provides the strength necessary to maintain membrane integrity. Inhibition of peptidoglycan synthesis or degradation of an existing cell wall will result in cell lysis[15].
Peptidoglycan is a linear polymer made up of two alternating sugar residues,
N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc), linked
through β-1→4 bonds[15]. The D-Lactoyl group of each MurNAc residue is
substituted with a tetrapeptide sequence of L-Alanine−γ-D-Glutamic acid− L-
Lysine (or in some cases 2,6 Diaminopimelic Acid)−D-Alanine, although the
exact sequence is known to have some variations from strain to strain[15, 16].
14
The D-L-D-L-D sequence prevents the formation of α-helical secondary structure and gives the tetrapeptide more flexibility[17]. The peptidoglycan strands, with an average length of 18 disaccharide units in S. aureus, form sheets around the
cytoplasmic membrane which are subsequently crosslinked by an interpeptide
bridge[15]. The interpeptide bridge forms a crosslink between the amino group
of the L-lysine (or 2,6 diaminopimelic acid) and the carboxyl group of the
terminal D-alanine residue on a neighboring tetrapeptide[15]. While most gram negative bacteria form crosslinks using a direct amide bond between the two residues, gram positive bacteria utilize a peptide bridge consisting of one to
seven amino acids of varying composition[13, 15]. This peptide interbridge is the
greatest source of diversity in the bacterial peptidoglycan structure. S. aureus, for
example, typically utilizes a five-reside glycine linker[13]. S. epidermidis, on the
other hand, shows more variation. While 20% of S. epidermidis strain Texas 26
possessed penta-glycine bridges, 55% were found to have Gly-Gly-Ser-Gly-Gly,
15% were Ser-Gly-Ser-Gly-Gly, and 10% were Gly-Ser-Gly-Gly-Gly [18].
Meanwhile, S. epidermidis strain 66 contains an alanine residue in place of one of
the glycines of the penta-glycine bridge[17].
The peptidoglycan structure serves as an anchor for both polysaccharides and cell wall proteins. The principal polysaccharide attached to the bacterial surface of gram positive bacteria is teichoic acid (TA), which can either be anchored to the cell wall (cell wall teichoic acid, CW TA) or anchored to the cytoplasmic
15
membrane by way of a lipid-spanning anchor (lipoteichoic acid, LTA)[19].
Teichoic acid is generally described as a (1,3)-phosphodiester-linked polyglycerol
phosphate (S. epidermidis) or polyribitol phosphate (S. aureus) with an overall
negative charge[20, 21]. The 2-position of the glycerol residues in S. epidermidis
can be substituted with α-glucose (α-Glc), α-N-acetylglucosamine (α-GlcNAc), D- alanine, or α-6-alanyl glucose (α-Glc6Ala)[22]. D-alanine is of particular importance, as each D-ala substitution imparts a positive charge into the TA molecule, countering some (but not all) of the negative charge provided by the phosphate backbone. An S. aureus mutant lacking the ability to incorporate D-ala onto the TA molecule lost its ability to colonize glass and plastic surfaces, presumably due to an increased negative surface charge of the bacteria that resulted in unfavorable electrostatic interactions with the surfaces[23]. It is also speculated that the net negative charge of TA is utilized to anchor non-covalently bound extracellular molecules such as autolysin E (AtlE) and polysaccharide intercellular adhesin (PIA).
Bacterial surface proteins are also anchored to the cell wall, with eleven genes encoding putative surface-anchored proteins already identified[24] These cell- wall-anchored (CWA) proteins often times play key roles in bacterial adherence and evasion from the host immune system[25]. S. aureus Protein A is a prototypical surface protein, and most CWA proteins of gram positive bacteria can be described by the same general structure[26]. At the N-terminus, there is
16
an ~40 amino acid secretion signal responsible for mediating translocation across the cell’s cytoplasmic membrane[24]. Bordering the secretion signal there is an A- region containing the ligand-binding site, a B-repeat region containing tandemly repeated sequences, and a proline-rich wall-spanning region.[27] Finally, there is a sorting signal region required for properly locating the protein on the bacterial surface. This sorting signal consists of a wall-anchoring LPXTG motif, a hydrophobic transmembrane segment, and a positively charged cytoplasmic tail[26]. While the LPXTG motif is highly conserved amongst gram positive bacteria, the hydrophobic domain and the positively charged tail are variable in sequence and length[26].
Attachment of CWA proteins to the cell wall is mediated by transpeptidases known as sortases. The S. aureus transpeptidase Sortase A (SrtA) is considered the archetypical sortase for gram positive bacteria, although five distinct subfamilies of sortases have been identified, with most bacteria encoding two or more distinct sortases[28]. S. epidermidis RP62A encodes just one sortase, SrtA[24,
28]. After a protein is exported into the extracellular space, it is retained within the cell membrane at its C-terminal by the sorting signal’s positively charged tail.
Following retention, SrtA proteolytically cleaves the protein between the threonine and glycine of the LPXTG motif, and the carboxyl group of the threonine is amide-linked to a free amino group of the peptidoglycan crossbridge[25]. Since a mature cell wall is highly crosslinked and therefore
17
contains very few free crossbridge amines, it is likely that protein anchoring
takes place prior to completion of the peptidoglycan crosslinking process[25].
1.3.2 Primary Adhesion to Surfaces
Prior to colonization, bacteria present in the bulk fluid environment are
transported towards the surface by Brownian motion, sedimentation, and
convective mass transport[29]. For static systems, sedimentation is the most important factor, and Brownian motion and convection can often times be neglected. Meanwhile, for flow systems, convective mass transport and diffusion
become much more significant[30]. Once S. epidermidis is brought in to close
proximity with the indwelling biomaterial, both nonspecific interactions and
specific binding events work to anchor the bacteria to the surface.
1.3.2.1 Nonspecific Binding
Bacterial adhesion to abiotic surfaces was first proposed by Marshall et al. to involve an initial reversible sorption step, followed by a surface-dependent irreversible sorption process[31]. As such, it was suggested that the adhesion of a bacterium to a surface could be explained by the DLVO colloidal stability theory
(after Derjaguin, Landau, Verwey, and Overbeek)[31]. The DLVO theory has since become a commonly accepted approach to modeling bacterial adhesion[32,
33].
18
The classical DLVO theory describes the total interaction energy (WDLVO) between a sphere (bacteria) and planar surface (biomaterial surface) in aqueous media as a balance between the attractive van der Waals energy (WvdW) and repulsive electrostatic double layer energy (WEl)[32, 33]. The van der Waals and
electrostatic double layer energies are additive and can be represented as a
function of separation distance (d) as follows: