Antimicrobial peptides Experimental prevention of osteomyelitis · Osteomyelitis Osteomyelitis is a...

$: Antimicrobial peptides Experimental prevention of osteomyelitis · Osteomyelitis Osteomyelitis is a refractory condition, potentially leading to amputation or even death. Treatment$
Antimicrobial peptidesExperimental prevention of

osteomyelitis

We performed the experiments at the Oral Cell Biology Department(OCB), Academic Centre for Dentistry (ACTA), VU. The work wassupported by:

AM-Pharma, Annafonds, Biomet, BV Cyclotron, Hereaus / Cavex, Johnson& Johnson, Link, Livit Orthopedie, Nederlandse vereniging voorBiomaterialen en Tissue Engineering, Nederlandse OrthopaedischeVereniging, Ortho Biotech, Smithkline & Nephew, STEGA, Stichting AnnaFonds, Stichting BIS, Stryker, Synthes Nederland, Wright Medical,Zimmer

© Hein Stallmann, 2007. All rights reserved. No part of this book may bereproduced, stored in retrievable form, or transmitted in any form or byany means, without written permission from the author.

Cover: The Gross Clinic, (Thomas Eakins, 1875), surgical treatment ofosteomyelitis. With kind permission of Jefferson Medical College ofThomas Jefferson University, Philadelphia.

Printed by Giethoorn ten Brink, Meppel. Fonts: Dutch823 and Verdana.

VRIJE UNIVERSITEIT

Antimicrobial peptidesExperimental prevention of osteomyelitis

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad Doctor aande Vrije Universiteit Amsterdam,

op gezag van de rector magnificusprof.dr. L.M. Bouter,

in het openbaar te verdedigenten overstaan van de promotiecommissie

van de faculteit der Geneeskundeop vrijdag 9 maart 2007 om 13.45 uurin het auditorium van de universiteit,

de Boelelaan 1105

door

Hein Pieter Stallmann

geboren te Amsterdam

promotoren: prof.dr. P.I.J.M. Wuismanprof.dr. A.C. van Nieuw Amerongen

copromotor: prof.dr. E.C.I. Veerman

Say it’s only a paper moon,

Contents

Chapter Page

Abbreviations 8

Outline and aims 9

1. Introduction. AMPs: application in musculoskeletal infections 11

2. Continuous-release or burst-release of hLF1-11from bone substitutes 23

3. In vitro gentamicin release from bone substitutes Influence of carrier type on duration of release 31

4. Histatin and lactoferrin derived AMPs: activity and toxicity 41

5. In vivo release of hLF1-11 from calcium phosphate cement 51

6. Osteomyelitis prevention in rabbits using hLF1-11or gentamicin containing cement 65

7. Summary and Discussion 77

8. Samenvatting NL 89

Dankwoord 101

8

Abbreviations

AMP antimicrobial peptide

AO Arbeitsgemeinschaft für Osteosynthesefragen

ATCC American Type Culture Collection

BHI brain heart infusion

CaP calcium phosphate

dH2O demineralized water

DHVAR-5 synthetic histatin analogue

ELISA Enzyme Linked Immunosorbent Assay

ESR erythrocyte sedimentation rate

Fmoc 9-fluorenylmethoxycarbonyl

GS gentamicin sulphate

hLF1-11 human lactoferrin N-terminal amino acids 1-11

HPLC high performance liquid chromatography

KDL department of experimental animal surgery

LC-MS/MS liquid chromatography mass-spectrometry/mass-spectrometry

LF lactoferrin

LPS lipopolysaccharide

MRSA methicillin resistant Staphylococcus aureus

PBS phosphate buffered saline

PMMA polymethylmethacrylate

Q-TOF MS quadrupole-time-of-flight mass-spectrometry

SCV small colony variant

STEGA Skeletal Tissue Engineering Group Amsterdam

VRSA vancomycin-resistant Staphylococcus aureus

VRSE vancomycin-resistant Staphylococcus epidermidis

WBC white blood cell count

9

Outline and aims

he first chapter introduces the main concepts of this manuscript:osteomyelitis (bone infection), bacterial resistance andantimicrobial peptides (AMPs). As part of a solution to the firsttwo problems, AMPs were investigated for the experimentalprevention of infection. The first chapter introduces several

AMPs which have been investigated in pre-clinical studies. In ourexperiments hLF1-11 was used, it consists of amino acids one to eleven ofhuman lactoferrin. We investigated a biodegradable carrier-system forlocal treatment or prevention of bone infection. The properties of severalbiodegradable carriers were investigated in chapters two and three. Thechapters show the release profiles of hLF1-11 and gentamicin, aconventional antibiotic, from a dozen different carrier materials. Theactivity on bacterial strains was confirmed after the release process.Chapter four describes the toxic properties and antimicrobial spectrum ofhLF1-11 and control peptides. No specific toxicity was observed on bloodand bone cells. The peptide rapidly killed all clinically problematicbacterial strains. These results encouraged us to set up several animalexperiments. For these experiments a combination of biodegradablecement (Bonesource) and hLF1-11 was selected. In the rabbitexperiments, the release of hLF1-11 from calcium phosphate cement wasused as a prophylactic agent against infection with Staphylococcus aureus,the principal cause of osteomyelitis. Chapter five describes the release ofhLF1-11 from cement into bone, the AMP rapidly dissipated from thecarrier. This was comparable to the earlier in vitro experiments (chapterstwo and three). Chapter six combines all results from earlier chapters in aprevention study. The rabbits were surgically contaminated with S. aureusand injected with cement and hLF1-11. After removal and culture of thebones, a significant reduction of infection was measured by both hLF1-11and (standard) gentamicin. This confirms the clinical potential in localtreatment of infection.

Study aims

• To develop a biodegradable carrier for release of hLF1-11.

• To evaluate the cytotoxicity and antimicrobial efficacy of hLF1-11.

• To compare the results to gentamicin and other antimicrobial peptidesfor reference.

• To implant the carrier and the peptide in an animal model and studyits release end efficacy.

10

11

Introduction

Antimicrobial peptides: review of theirapplication in musculoskeletal infections

Hein P Stallmann•, Chris Faber•, Arie V Nieuw Amerongen,Paul IJM Wuisman(• Equal contribution)

Injury (2006), 37(Supplement 2): 34-40

Antimicrobial resistance is expected to increase the burden ofosteomyelitis drastically. The rise in resistant bacterial strains is drivingresearchers to find new treatment options. As a potential new antibioticclass, antimicrobial peptides (AMPs) combine several attractive intrinsicproperties. Their minimal propensity for inducing antimicrobialresistance could be of particular clinical significance.AMPs act as an essential part of the innate immune system and havebeen identified in virtually all forms of life. These short, positivelycharged peptides have a combined pore-forming and intracellular killingeffect on a broad range of microorganisms. Their reported spectrum ofaction includes resistant bacterial strains, viruses, and fungi. Moreover,immunomodulating, antitumoric, and angiogenic mechanisms have beenreported.We have designed degradable and non-degradable drug-release systemsfor local treatment with AMPs. In animal models of osteomyelitis, thesesystems reduced bone infection caused by both resistant and non-resistant strains. The systemic application of several peptides forexperimental detection and treatment of bone and soft-tissue infection isalso discussed in this chapter. Radioactive-labelled peptides haveaccurately discriminated sterile inflammation from active infection inimaging studies.Successful pre-clinical studies of AMPs indicate that clinical evaluationof these powerful antibiotic agents is in order.

1

Chapter 1

12

esistant bacterial strains on gentamicin beads characterisethe dilemma that we face in treating osteomyelitis.1 In spiteof its initial bacterial reduction, the gradual release of anantibiotic may induce resistance, thus frustrating effectivetreatment. Furthermore, adhesion of bacteria to implant

surfaces causes reduced antibiotic sensitivity and increased virulence.The acquisition of the implant surface by host cells or bacteria becomes acrucial ”race for the surface” that determines the clinical outcome.2 Theheavy impact of osteomyelitis translates into increased morbidity,mortality, and cost of healthcare resources.3-5

‘Superbugs’ such as vancomycin-resistant Staphylococcus aureus(VRSA) and vancomycin-resistant Staphylococcus epidermidis (VRSE)have become resistant to last-resort drugs. Until recently, the global risein antibiotic-resistant pathogens coincided with reduced commercialefforts to develop new antibiotic agents.6 Over the last 40 years, newclasses of antibiotics have been scarce, although there has been a positivetrend since the turn of the millennium.7

The intriguing properties of AMPs have stimulated interest in thesenatural effectors of immunity.8

Since their identification in frogsand insects, they have been found in all life forms studied for theirpresence. They constitute a mechanism of resistance which is very old inevolutionary terms, and their strength lies in a combination of poreformation, intracellular killing, and minimal induction of resistance.9 Asnew characteristics of AMPs are described, research focus is shifting toadditional properties such as immunomodulation.10,11

Several papers have reported successful in vivo infection preventionand treatment with AMPs. The controlled release of AMPs from carriermaterials to prevent and treat osteomyelitis is a focus of our AMP-relatedresearch.12-14

OsteomyelitisOsteomyelitis is a refractory condition, potentially leading to

amputation or even death. Treatment often requires multiple surgicalinterventions and local or systemic antibiotic therapy.15 It predominantly

affects both extremes of age; acutehematogenous infection occurs mainly inchildren and chronic osteomyelitis in theelderly.16 Implant surfaces constitute a

safe haven for pathogens, frustrating fracture treatment and arthroplasty.The prophylaxis and treatment of implant-associated infections wasaddressed in depth by several authors in a 2006 AO Supplement toInjury.

The old adage “once osteomyelitis always osteomyelitis” has not lostits relevance. The multifactorial nature of these infections demands anapproach that addresses all aspects related to the patient, the wound, theimplant, and the pathogen. This implies that no single solution will work

identification in frogs and insects

it predominantly affects

both extremes of age

Introduction

13

in all cases. Despite controversy on the use of prophylactic localantibiotic treatment, long term studies using Scandinavian implantregistration data seem to justify using antibiotic containing cement.17

Releasemg/g

10

20

0 1 2

(time)

1/2 [days

1/2]

GS from CaP

hLF1-11 from CaP

DHVAR-5 from PMMA

GS from PMMA

Figure 1 The early in vitro release kinetics of AMPs and gentamicin (GS)showed high release of hLF1-11 and DHVAR-5. Gentamicin was swiftlyreleased from biodegradable CaP and gradually from PMMA. Thedrug/carrier ratios were: 5.0% hLF1-11, 5.0% DHVAR-5, and 3.0%gentamicin. Data from several experiments were combined in this figure.Values and SD represent serial experiments in triplicate. The x-axisrepresents the square root of time.

For established infections, surgical debridement of all infected tissueis combined with local and systemic antibiotic therapy. Additionalmeasures to improve the patient’s immune status, reduce tissue traumaand optimise implant surfaces will further decrease the occurrence ofosteomyelitis.

Local treatment of infection has favourable pharmacokinetics, itproduces high drug concentrations at the focus of infection withoutsystemic toxicity.18 In the 1970s, Buchholz introduced the concept ofantibiotic-containing polymethyl-methacrylate (PMMA) cement forprevention of infection.19 Klemm followed suit with antibiotic beadchains for treatment of chronic infection.20 After this, biodegradablecollagen fleeces appeared, which produce high local antibioticconcentrations.21 Unlike biodegradable materials, PMMA-beads allowstaged treatment and require surgical removal after each treatmentepisode.

Chapter 1

14

Disadvantages of currently used non-biodegradable PMMA carriersinclude low antibiotic release by cements and the need to surgicallyremove PMMA beads.18 Moreover, resistant bacteria may appear on thecarrier surface during later stages of low-level antibiotic release.22 Incontrast, biodegradable antibiotic carriers may result in high levels ofrelease and obviate the need for removal as they are gradually replacedby ingrowing tissue.23,24 Furthermore, secondary release of the antibioticmay occur during the degradation phase of the carrier, which couldincrease the antimicrobial efficacy compared to non-biodegradablecarriers.25

Animal studiesIn a well-established rabbit model, we studied two AMPs for

preventing and treating femoral infection. The rabbit model is often usedto analyse the efficacy of drug-release systems.25,26 The studies comparedtwo peptides: human lactoferrin N-terminal amino acids 1-11 (hLF1-11)and a synthetic histatin analogue (DHVAR-5). Both AMPs kill a broadrange of resistant and non-resistant pathogens, including fungi.27-29

Figure 2 The three-dimensional structure of human lactoferricin, ofwhich hLF1-11 comprises the N-terminal eleven amino-acids (bottompart). The positively charged parts have been highlighted black,negatively charged parts red. The image (structure 1Z6W deposited byHunter and coworkers), was rendered with Cn3D version 4.1, producedby the National Center for Biotechnology Information,http://ncbi.nlm.nih.gov, last accessed on 27 March 2006

Biodegradable calcium phosphate (CaP) cement or PMMA was usedas the carrier matrix. Figure 1 shows the release of these peptides fromtheir carriers, compared to gentamicin release from the same carriermaterials. Figure 2 shows the three-dimensional structure of humanlactoferricin, of which amino-acids one to eleven form hLF1-11.

In a prevention study using a CaP carrier, hLF1-11 significantlyreduced S. aureus infection. Histology showed some signs of earlyremodelling but no signs of toxicity.14 After operative induction of

Introduction

15

osteomyelitis with a virulent methicillin-resistant Staphylococcus aureus(MRSA) strain, both hLF1-11 and DHVAR-5 succeeded in reducing thenumber of pathogens. However, in these one-stage treatment studies,AMPs did not kill all bacteria.12,13 This is consistent with operativetreatment of chronic osteomyelitis. Staged therapy - several operativeinsertions of antibiotic carriers - is often required.

Others have used hLF1-11 in a mouse model of thigh muscleinfection. After intravenous injection, hLF1-11 killed a number ofpathogens, including multi-drug resistant Acinetobacter baumannii.11,30,31

Steinstraesser et al reported that the designer peptide Novispirin G10adequately treated skin burns infected with resistant pathogens.32

Protegrin-1, a defensin-like molecule isolated from pig leukocytes,showed its activity in a similar burn model. Protegrin-1 was also mosteffective on several bacterial strains in a model of lung infection.33,34 Asummary of in vivo antimicrobial activity of several AMPs is given inTable 1.

AMP Microorganisms Animals and models

hLF1-11A. baumannii, C. albicans,E. coli, K. pneumoniae,MRSA, S. aureus

Mouse, rabbit 11,13,14,30,31

Bone infection, muscleinfection

DHVAR-5 MRSARabbit 12

Bone infection

Novispirin G10P. aeruginosa,K. pneumoniae

Mouse, rat 32,35

Burn infection, Lunginfection

Protegrin-1C. albicans, MRSA,P. aeruginosa, S. aureus

Mouse, pig, rat 33,34,36

Burn infection, woundinfection

Table 1 In vivo results of several broad spectrum AMPs. ParticularlyMRSA and A. baumannii notoriously cause problematic nosocomialinfections.

Antimicrobial resistanceIn developing countries, the problem is characterised by accelerating

rates of resistance driven by antibiotic misuse and shortfalls in infectioncontrol and public health. Concerted action is needed in order to preventthe emergence of new resistant strains and the spread of existing ones.The World Health Organisation targets factors such as regulated drugavailability, drug quality control, and surveillance in its containmentstrategies. Implementation of combination therapy, drug cycling

Chapter 1

16

regimes, and directly observed short-course therapy may furtherdecrease resistance.37-39

However, the problem of resistance is certainly not limited todeveloping countries. Other countries are also plagued by resistantpathogens, and are experiencing an increasing risk of community-acquired resistant strains.40 Our last-resort drugs are continuously undersiege. A recent report details the mechanism of reduced sensitivity tovancomycin in S. aureus: mutation and thickening of the cell wall due toaccumulation of excess amounts of peptidoglycan.41 Consequentially, weneed to re-invigorate our marginalised antimicrobial developmentefforts.42

Infection by resistant pathogens is a growing problem in implantsurgery.43 Infected implants are rapidly covered by pathogens that form abiofilm. The surface of such biofilms repels or inactivates antibiotics and,in deeper layers, less susceptible slow-growing bacteria survive. Small colonyvariants (SCVs) may occur, which are slow-growing bacteria with an altered cellmembrane potential that results in reduced susceptibility to, for example,aminoglycosides. The concepts of biofilm formation, SCVs, andresistance mechanisms were addressed in depth by Harris, Richards andvon Eiff.44-46 Biofilms of resistant bacteria have been found even ongentamicin-releasing PMMA cement, which stresses the importance oftimely removal of PMMA beads to prevent resistance.1,47

AMPs are poised to take on resistant pathogens. Even after repeatedsub-therapeutic exposure to AMPs in vitro, occurrence of resistantbacterial strains remained rare.48 The inhibition of resistancedevelopment has been explained by the large number of natural AMPvariants: thousands of which have been identified.9 Furthermore, thechange in cell membrane potential required to repel AMPs could bedifficult for bacteria to produce.10 However, a small number of reports onmechanisms of resistance to AMPs invites cautious introduction andprudent use of AMP-based drugs.49,50

Antimicrobial peptidesAMPs, also called cationic peptides because of their positive charge,

interact with the negatively charged cell membrane of bacteria and fungi.This interaction results in pore formation and membrane perturbation,leakage of cellular components, and cell death. AMPs can enter throughthe pores and killing occurs by electrolyte disturbance and sabotage ofthe mitochondria.51-54 Mammalian cell membranes, which are positivelycharged and rich in cholesterol, do not interact with AMPs.9 SeveralAMPs mentioned in this section are listed in Table 2. Although AMPsrarely induce resistance and readily kill a range of multi-drug-resistantpathogens, some authors have reported mechanisms of primary oracquired resistance.49,50

AMPs are poised to take

on resistant pathogens

Introduction

17

We combined two AMPs, hLF1-11 and DHVAR-5, with several CaPand PMMA carriers and analysed the release pattern in vitro. Figure 1compares the release profiles of DHVAR-5, hLF1-11, and gentamicinfrom the same carrier materials that were later used in vivo. Bothpeptides retained their activity after the release process and killed a widerange of bacteria, including resistant strains.29,55,56 By altering the porousstructure of PMMA, the addition of DHVAR-5 increased the release ofgentamicin while the cement retained adequate mechanical properties.57

The toxicity of the peptides was determined on human erythrocytesand cultured mouse bone cells. No toxicity was observed at therapeuticconcentrations. Only DHVAR-5 induced haemolysis and bone cell deathat 16 times the concentration required to kill the pathogens tested.29

Protegrin-1 and Cystatin-S were used as the positive and negativecontrols in the in vitro studies.29,56 These results were confirmed in thepreviously mentioned rabbit studies of osteomyelitis caused by MRSA.12,13

Table 2 The amino acid sequences of several peptides mentioned in thisreview. Positively charged amino acids (R, K and H) are indicated bybold typeface. Notice the asymmetrical distribution of positively chargedpeptides in the active peptides; three-dimensional folding in vivo mayenhance this asymmetry. On contact with bacteria, the positively chargedpeptides perforate the negatively charged bacterial cell membrane.Cystatin-S is an inactive control peptide used in several in vitro studies.

Biofilms, a successful defence strategy employed by bacteria on implantsurfaces, appear to be accessible to AMPs. In vitro, DHVAR-5 causedbacterial reduction in oral biofilms (plaque) on implant surfaces.27 Songand associates studied intra-tracheal administration of Novispirin G-10in a cystic fibrosis model. It effectively countered resistant Pseudomonasaeruginosa-induced lung infection in rats.5 In a similar model, Bartlettand coworkers showed IL-6 up-regulation during Novispirin G-10treatment of Klebsiella pneumoniae infection.35 The beneficial effect ofimmunomodulation by AMPs has been reported in several studies: itresulted in higher in vivo than in vitro activity on pathogenicmicroorganisms.11

The in vivo kinetics and biodistribution of injected AMPs may bemonitored scintigraphically by labelling them with radioactive

Peptide Structure

Cystatin-S (inactive peptide) SSSKEENRIIPGGIYDA

DHVAR5 LLLFLLKKRKKRKY

hLF1-11 GRRRRSVQWCA

Novispirin G10 KNLRRIIRKGIHIIKKYG

Protegrin-1 RGGRLCYCRRRFCVCVVGR

Chapter 1

18

technetium (Tc99m )58. Such peptides show a preferential accumulation atsites of bacterial or fungal infection, which makes them a potentialdiagnostic option. This process could be used to detect occult infections,monitor antibiotic therapy, and discriminate between infection andsterile inflammation.59

Additional clinical applications may be developed from existing invitro results: Lipopolysaccharide (LPS) -binding capacity of AMPs couldimprove sepsis treatment.60 The reports of anti-retroviral activity on HIVstrains and anti-tumoric activity of several synthetic AMPs enable pre-clinical studies to improve our understanding of these features.61-63

Furthermore, pore formation by AMPs can promote drug entry intobacteria and fungi, allowing a synergistic effect.64,65 Finally, the claims ofdecreased bone resorption and angiogenic properties may need furtherevaluation to show their clinical efficacy.66,67 Several clinical trialsunderline the potential of future AMP-based drugs for treating superficialinfections, oral mucositis, and paediatric sepsis.68-70

ConclusionsThe multifactorial problem of implant-related osteomyelitis requires a

combined approach. Aspects related to both patient (immune system),operation (implant type) andmicroorganism (virulence,sensitivity) dictate theoutcome. By optimising allfactors, the risk of infection may be drastically reduced. A combination oflocal and systemic antibiotic prophylaxis has shown good results inminimising infection.

Until recently, the rise in antimicrobial resistance was not matched bycommercial research efforts into new antibiotics. The occurrence ofVRSA and VRSE, resistant to ‘last-resort’ antibiotics, may herald an ageof challenging clinical problems, reminiscent of the pre-antibiotic eraearly last century.7,41 New antibiotics may be part of the solution toprevent this scenario. Prudent use of antibiotics, in healthcare and foodproduction, may further oppose the development of resistance.4

AMPs, a diverse class of natural antibiotic peptides, kill commonproblematic bacterial strains and do not easily induce resistance. In vivoexperiments indicate strong activity without signs of toxicity attherapeutic levels. In prophylactic and therapeutic models of bone andsoft-tissue infection, AMPs significantly reduced the number of offendingbacteria. Additional potential therapeutic options include: antiviral,antifungal, angiogenic, LPS-binding, and tumoricidal applications.

The retrieval of resistant bacterial strains on explanted gentamicin-containing PMMA beads (in 18 of 20 patients) highlights the clinicalimpact of antibiotic-resistance in orthopaedic infections.1 The successfulin vivo results of several AMPs in prevention and treatment studiesvalidate the need for clinical studies of bone and implant infection.

the rise in antimicrobial resistance was not

matched by commercial research efforts

Introduction

19

References

1. Neut D, van de Belt H, Stokroos I, van Horn JR, van der Mei HC, Busscher HJ. (2001).Biomaterial-associated infection of gentamicin-loaded PMMA beads in orthopaedicrevision surgery. J Antimicrob Chemother 47: 885-891.

2. Gristina AG, Oga M, Webb LX, Hobgood CD. (1985). Adherent bacterial colonization inthe pathogenesis of osteomyelitis. Science 228: 990-993.

3. Coast J, Smith RD, Millar MR. (1996). Superbugs: should antimicrobial resistance beincluded as a cost in economic evaluation? Health Econ 5: 217-226.

4. Coast J, Smith RD, Millar MR. (1998). An economic perspective on policy to reduceantimicrobial resistance. Soc Sci Med 46: 29-38.

5. Song Z, Wu H, Mygind P, Raventos D, Sonksen C, Kristensen HH, Hoiby N. (2005).Effects of intratracheal administration of novispirin G10 on a rat model of mucoidPseudomonas aeruginosa lung infection. Antimicrob Agents Chemother 49: 3868-3874.

6. Norrby SR, Nord CE, Finch R. (2005). Lack of development of new antimicrobialdrugs: a potential serious threat to public health. Lancet Infect Dis 5: 115-119.

7. Hancock RE. (2005). Mechanisms of action of newer antibiotics for Gram-positivepathogens. Lancet Infect Dis 5: 209-218.

8. Hancock RE. (1999). Host defence (cationic) peptides: what is their future clinicalpotential? Drugs 57: 469-473.

9. Zasloff M. (2002). Antimicrobial peptides of multicellular organisms. Nature 415: 389-395.

10. Hancock RE. (2001). Cationic peptides: effectors in innate immunity and novelantimicrobials. Lancet Infect Dis 1: 156-164.

11. Nibbering PH, Ravensbergen E, Welling MM, van Berkel LA, van Berkel PH, PauwelsEK, Nuijens JH. (2001). Human lactoferrin and peptides derived from its N terminusare highly effective against infections with antibiotic-resistant bacteria. Infect Immun69: 1469-1476.

12. Faber C, Hoogendoorn RJ, Stallmann HP, Lyaruu DM, Nieuw Amerongen AV,Wuisman PIJM. (2004). In vivo comparison of Dhvar-5 and gentamicin in an MRSAosteomyelitis prevention model. J Antimicrob Chemother 54: 1078-1084.

13. Faber C, Stallmann HP, Lyaruu DM, Joosten U, von Eiff C, Nieuw Amerongen AV,Wuisman PIJM. (2005). Comparable efficacies of the antimicrobial peptide humanlactoferrin 1-11 and gentamicin in a chronic methicillin-resistant Staphylococcusaureus osteomyelitis model. Antimicrob Agents Chemother 49: 2438-2444.

14. Stallmann HP, Faber C, Bronckers AL, Nieuw Amerongen AV, Wuisman PIJM. (2004).Osteomyelitis prevention in rabbits using antimicrobial peptide hLF1-11- orgentamicin-containing calcium phosphate cement. J Antimicrob Chemother 54: 472-476.

15. Cierny G, III, Mader JT, Penninck JJ. (2003). A clinical staging system for adultosteomyelitis. Clin Orthop Relat Res 7-24.

16. Le Saux N, Howard A, Barrowman NJ, Gaboury I, Sampson M, Moher D. (2002).Shorter courses of parenteral antibiotic therapy do not appear to influence responserates for children with acute hematogenous osteomyelitis: a systematic review. BMCInfect Dis 2: 16.

17. Engesaeter LB, Lie SA, Espehaug B, Furnes O, Vollset SE, Havelin LI. (2003).Antibiotic prophylaxis in total hip arthroplasty: effects of antibiotic prophylaxissystemically and in bone cement on the revision rate of 22,170 primary hipreplacements followed 0-14 years in the Norwegian Arthroplasty Register. Acta Orthop74: 644-651.

18. Henry SL, Galloway KP. (1995). Local antibacterial therapy for the management oforthopaedic infections. Pharmacokinetic considerations. Clin Pharmacokinet 29: 36-45.

19. Buchholz HW, Engelbrecht H. (1970). [Depot effects of various antibiotics mixed withPalacos resins]. Chirurg 41: 511-515.

20. Klemm KW. (1993). Antibiotic bead chains. Clin Orthop 63-76.21. Ipsen T, Jorgensen PS, Damholt V, Torholm C. (1991). Gentamicin-collagen sponge for

local applications. 10 cases of chronic osteomyelitis followed for 1 year. Acta Orthop 62:592-594.

Chapter 1

20


23. Walenkamp GH. (2001). Gentamicin PMMA beads and other local antibiotic carriers intwo-stage revision of total knee infection: a review. J Control Release 13 Spec No 1: 66-72.

24. Wichelhaus TA, Dingeldein E, Rauschmann M, Kluge S, Dieterich R, Schafer V, BradeV. (2001). Elution characteristics of vancomycin, teicoplanin, gentamicin andclindamycin from calcium sulphate beads. J Antimicrob Chemother 48: 117-119.

25. Humphrey JS, Mehta S, Seaber AV, Vail TP. (1998). Pharmacokinetics of a degradabledrug delivery system in bone. Clin Orthop Relat Res 218-224.

26. Calhoun JH, Mader JT. (1997). Treatment of osteomyelitis with a biodegradableantibiotic implant. Clin Orthop 206-214.

27. Helmerhorst EJ, Hodgson R, van 't HW, Veerman EC, Allison C, Nieuw Amerongen AV.(1999). The effects of histatin-derived basic antimicrobial peptides on oral biofilms. JDent Res 78: 1245-1250.

28. Lupetti A, Paulusma-Annema A, Welling MM, Senesi S, Van Dissel JT, Nibbering PH.(2000). Candidacidal activities of human lactoferrin peptides derived from the Nterminus. Antimicrob Agents Chemother 44: 3257-3263.

29. Stallmann HP, Faber C, Bronckers AL, Blieck-Hogervorst JM, Brouwer CP, NieuwAmerongen AV, Wuisman PIJM. (2005). Histatin and lactoferrin derived peptides:Antimicrobial properties and effects on mammalian cells. Peptides 26: 2355-2359.

30. Dijkshoorn L, Brouwer CP, Bogaards SJ, Nemec A, van den Broek PJ, Nibbering PH.(2004). The synthetic N-terminal peptide of human lactoferrin, hLF(1-11), is highlyeffective against experimental infection caused by multidrug-resistant Acinetobacterbaumannii. Antimicrob Agents Chemother 48: 4919-4921.

31. Lupetti A, Danesi R, 't Wout JW, Van Dissel JT, Senesi S, Nibbering PH. (2002).Antimicrobial peptides: therapeutic potential for the treatment of Candida infections.Expert Opin Investig Drugs 11: 309-318.

32. Steinstraesser L, Tack BF, Waring AJ, Hong T, Boo LM, Fan MH and others. (2002).Activity of novispirin G10 against Pseudomonas aeruginosa in vitro and in infectedburns. Antimicrob Agents Chemother 46: 1837-1844.

33. Steinberg DA, Hurst MA, Fujii CA, Kung AH, Ho JF, Cheng FC, Loury DJ, Fiddes JC.(1997). Protegrin-1: a broad-spectrum, rapidly microbicidal peptide with in vivoactivity. Antimicrob Agents Chemother 41: 1738-1742.

34. Steinstraesser L, Klein RD, Aminlari A, Fan MH, Khilanani V, Remick DG, Su GL,Wang SC. (2001). Protegrin-1 enhances bacterial killing in thermally injured skin. CritCare Med 29: 1431-1437.

35. Bartlett KH, McCray PB, Jr., Thorne PS. (2003). Novispirin G10-induced lung toxicityin a Klebsiella pneumoniae infection model. Antimicrob Agents Chemother 47: 3901-3906.

36. Ceccarelli AV, Cole AM, Park AK, Tahk S, Yoshioka D, Ganz T. (2001). Therapeuticeffect of a pig-derived peptide antibiotic on porcine wound infections. Comp Med 51:75-79.

37. Okeke IN, Klugman KP, Bhutta ZA, Duse AG, Jenkins P, O'Brien TF, Pablos-Mendez A,Laxminarayan R. (2005). Antimicrobial resistance in developing countries. Part II:strategies for containment. Lancet Infect Dis 5: 568-580.

38. Okeke IN, Laxminarayan R, Bhutta ZA, Duse AG, Jenkins P, O'Brien TF, Pablos-Mendez A, Klugman KP. (2005). Antimicrobial resistance in developing countries. PartI: recent trends and current status. Lancet Infect Dis 5: 481-493.

39. World Health Organization. (2002). WHO Global Strategy for Containment ofAntimicrobial Resistance. WHO Report.

40. Zetola N, Francis JS, Nuermberger EL, Bishai WR. (2005). Community-acquiredmeticillin-resistant Staphylococcus aureus: an emerging threat. Lancet Infect Dis 5:275-286.

41. Hiramatsu K. (2001). Vancomycin-resistant Staphylococcus aureus: a new model ofantibiotic resistance. Lancet Infect Dis 1: 147-155.

42. Livermore DM. (2005). Minimising antibiotic resistance. Lancet Infect Dis 5: 450-459.

Introduction

21

43. Garvin KL, Hinrichs SH, Urban JA. (1999). Emerging antibiotic-resistant bacteria.Their treatment in total joint arthroplasty. Clin Orthop 110-123.

44. Harris LG, Richards RG. (2006). Staphylococci and implant surfaces: a review. Injury37: S3-S14.

45. Richards RG, Harris LG, Schneider E, Haas N. (2006). Antiseptics and antibiotics onimplants. Injury 37: S113-S116.

46. von Eiff C, Peters G, Becker K. (2006). The small colony variant (SCV) concept-the roleof staphylococcal SCVs in persistent infections. Injury 37: S26-S33.

47. van de Belt H, Neut D, Schenk W, van Horn JR, Der Mei HC, Busscher HJ. (2001).Staphylococcus aureus biofilm formation on different gentamicin-loadedpolymethylmethacrylate bone cements. Biomaterials 22: 1607-1611.

48. Hancock RE. (1997). Peptide antibiotics. Lancet 349: 418-422.49. Ganz T. (2001). Fatal attraction evaded. How pathogenic bacteria resist cationic

polypeptides. J Exp Med 193: F31-F34.50. Yeaman MR, Yount NY. (2003). Mechanisms of antimicrobial peptide action and

resistance. Pharmacol Rev 55: 27-55.51. Helmerhorst EJ, Breeuwer P, Van't Hof W, Walgreen-Weterings E, Oomen LC,

Veerman EC, Amerongen AV, Abee T. (1999). The cellular target of histatin 5 onCandida albicans is the energized mitochondrion. J Biol Chem 274: 7286-7291.

52. Helmerhorst EJ, Troxler RF, Oppenheim FG. (2001). The human salivary peptidehistatin 5 exerts its antifungal activity through the formation of reactive oxygen species.Proc Natl Acad Sci U S A 98: 14637-14642.

53. Helmerhorst EJ, Van't Hof W, Breeuwer P, Veerman EC, Abee T, Troxler RF,Amerongen AV, Oppenheim FG. (2001). Characterization of histatin 5 with respect toamphipathicity, hydrophobicity, and effects on cell and mitochondrial membraneintegrity excludes a candidacidal mechanism of pore formation. J Biol Chem 276: 5643-5649.

54. Helmerhorst EJ, Murphy MP, Troxler RF, Oppenheim FG. (2002). Characterization ofthe mitochondrial respiratory pathways in Candida albicans. Biochim Biophys Acta1556: 73-80.

55. Faber C, Stallmann HP, Lyaruu DM, de Blieck JM, Bervoets TJ, Nieuw Amerongen AV,Wuisman PIJM. (2003). Release of antimicrobial peptide Dhvar-5 frompolymethylmethacrylate beads. J Antimicrob Chemother 51: 1359-1364.

56. Stallmann HP, Faber C, Slotema ET, Lyaruu DM, Bronckers AL, Nieuw AmerongenAV, Wuisman PIJM. (2003). Continuous-release or burst-release of the antimicrobialpeptide human lactoferrin 1-11 (hLF1-11) from calcium phosphate bone substitutes. JAntimicrob Chemother 52: 853-855.

57. Faber C, Hoogendoorn RJ, Lyaruu DM, Stallmann HP, van Marle J, Nieuw AmerongenAV, Smit TH, Wuisman PIJM. (2005). The effect of the antimicrobial peptide, Dhvar-5,on gentamicin release from a polymethyl methacrylate bone cement. Biomaterials 26:5717-5726.

58. Lupetti A, Nibbering PH, Welling MM, Pauwels EK. (2003). Radiopharmaceuticals:new antimicrobial agents. Trends Biotechnol 21: 70-73.

59. Lupetti A, Welling MM, Pauwels EK, Nibbering PH. (2003). Radiolabelledantimicrobial peptides for infection detection. Lancet Infect Dis 3: 223-229.

60. Ding L, Yang L, Weiss TM, Waring AJ, Lehrer RI, Huang HW. (2003). Interaction ofantimicrobial peptides with lipopolysaccharides. Biochemistry 42: 12251-12259.

61. Cole AM, Hong T, Boo LM, Nguyen T, Zhao C, Bristol G and others. (2002).Retrocyclin: a primate peptide that protects cells from infection by T- and M-tropicstrains of HIV-1. Proc Natl Acad Sci U S A 99: 1813-1818.

62. Cole AM. (2003). Minidefensins and other antimicrobial peptides:candidate anti-HIVmicrobicides. Expert Opin Ther Targets 7: 329-341.

63. Cole AM, Lehrer RI. (2003). Minidefensins: antimicrobial peptides with activity againstHIV-1. Curr Pharm Des 9: 1463-1473.

64. Lupetti A, Paulusma-Annema A, Welling MM, Dogterom-Ballering H, Brouwer CP,Senesi S, Van Dissel JT, Nibbering PH. (2003). Synergistic activity of the N-terminalpeptide of human lactoferrin and fluconazole against Candida species. AntimicrobAgents Chemother 47: 262-267.

Chapter 1

22

65. Nekhotiaeva N, Elmquist A, Rajarao GK, Hallbrink M, Langel U, Good L. (2004). Cellentry and antimicrobial properties of eukaryotic cell-penetrating peptides. FASEB J 18:394-396.

66. Koczulla R, von Degenfeld G, Kupatt C, Krotz F, Zahler S, Gloe T and others. (2003).An angiogenic role for the human peptide antibiotic LL-37/hCAP-18. J Clin Invest 111:1665-1672.

67. Lorget F, Clough J, Oliveira M, Daury MC, Sabokbar A, Offord E. (2002). Lactoferrinreduces in vitro osteoclast differentiation and resorbing activity. Biochem Biophys ResCommun 296: 261-266.

68. Giles FJ, Miller CB, Hurd DD, Wingard JR, Fleming TR, Sonis ST and others. (2003). Aphase III, randomized, double-blind, placebo-controlled, multinational trial of isegananfor the prevention of oral mucositis in patients receiving stomatotoxic chemotherapy(PROMPT-CT trial). Leukemia and Lymphoma 44: 1165-1172.

69. Isaacson RE. (2003). MBI-226. Micrologix/Fujisawa. Current Opinion in InvestigationalDrugs 4: 999-1003.

70. Levin M, Quint PA, Goldstein B, Barton P, Bradley JS, Shemie SD and others. (2000).Recombinant bactericidal/permeability-increasing protein (rBPI21) as adjunctivetreatment for children with severe meningococcal sepsis: a randomised trial. rBPI21Meningococcal Sepsis Study Group. Lancet 356: 961-967.

23

Continuous-release or burst-release ofthe anti-microbial peptide hLF1-11 fromcalcium phosphate bone substitutes

Hein P. Stallmann, Chris Faber, Eveline T. Slotema, D.M. Lyaruu,Antonius L.J.J. Bronckers, Arie V. Nieuw Amerongen,Paul I.J.M. Wuisman

J Antimicrob Chemother (2003), 52: 853-855

Background In order to identify possible drug delivery systems againstresistant bone infection, we determined the release of the antimicrobialpeptide (AMP) human lactoferrin 1-11 (hLF1-11) from commerciallyavailable bone substitutes.Methods We combined six calcium phosphate cements and six granule-types with 5 mg/g hLF1-11 and measured the availability and release invitro from cements (seven days) and granules (three days). The integrityand antimicrobial activity of the hLF1-11 that was released during thefirst 24 hours were measured, using mass spectrometry, and a killingassay on methicillin resistant Staphylococcus aureus (MRSA).Results Most of the cements showed burst release followed by low-levelcontinuous release, whereas the coated granules showed high burstrelease for 24 hours. After release the peptide was active (in 9 of 12materials) and intact.Conclusions Different release profiles may be obtained by choosing theappropriate carrier, which supports the feasibility of biodegradablecarriers releasing AMP’s against resistant infections.

2

Chapter 2

24

ntimicrobial resistance will probably complicate futuretreatment of bone infection. Treatment requires local andsystemic antibiotics and often several surgical interventions.The non-degradable polymethyl-methacrylate beads that areused currently as antibiotic carriers demand operative

removal and may induce resistant bacteria.1

We aimed to address this growing problem by combining an AMP ofhuman origin (hLF1-11) with biodegradable carriers - which obviates theneed for operative removal - and analysing its availability and release.These carriers, consist of calcium-phosphate ceramics such astricalcium-phosphate, Ca3(PO4)2. They are slowly replaced by ingrowingbone after implantation.2

AMP’s form a novel class ofantimicrobial agents of naturalorigin and have been identified invirtually all forms of life as part ofthe antimicrobial defence system. These positively charged peptides killby forming pores in the negatively-charged bacterial cell-membrane andtargeting intracellular organelles. They rarely cause development ofresistance.3 Moreover, they seem to have an immuno-modulating effect,which kills microorganisms at lower concentrations in vivo(nanograms/mL) than in vitro (micrograms/mL).4,5

Materials and MethodsPeptide

HLF1-11-peptide (GRRRRSVQWCA, 1375 Da) was manufactured bysolid-phase peptide synthesis using Fmoc (9-fluorenylmethoxycarbonyl)chemistry as described previously.6 Reanalysis of peak fractions byReversed phase HPLC resulted in one major peak revealing at least 90%purity. The authenticity was confirmed by electrospray-ionisationquadrupole-time-of-flight mass-spectrometry (Q-TOF MS, MicromassInc.). Thermal stability in solution, adhesion to polystyrene and solubilitywere tested as described previously.7

Release ExperimentAfter mixing cement powder and liquid containing 5 mg hLF1-11 per

gram of powder (liquid-to-powder ratio as specified by themanufacturer) cylindrical specimens hardened overnight at 37 °C in6x5mm moulds. The cements were: Biobon (Biomet Merck), Calcibon(Biomet Merck), Biofil, (experimental, DePuy CMW), Bonesource(Stryker-Leibinger), Chronos Inject, (Mathys) and Norian SRS (Mathys).

Granules were immersed in 1.0 mL of dH2O containing 5 mg of hLF1-11 per gram of material and lyophilised. After removal of the granules,the residual hLF1-11 in the vessel was measured. The granules were:Bonesave (Stryker-Leibinger), Biosorb (Science for Biomaterials,Lourdes, France), Allogran-R (Orthos, Bristol, UK), Vitoss (OrthovitaMalvern, PA, USA), Cerasorb (Curasan, Kleinostheim, Germany), andBicalphos (Medtronic).

peptides kill by forming pores in

the negatively-charged bacterial

cell-membrane

hLF1-11 release in vitro

25

Specimens were immersed in 500 µL dH2O and kept in sealedpolystyrene 48-well plates (Costar) at room temperature on a shakingdevice (180 rpm). The water was replaced at regular intervals: 30, 90 and180 min. on day 1 and then 24 hourly for 7 days (cements) or 3 days(granules) and stored at –20 °C.

After production and after release three specimens per group werefinely ground and suspended in 5 mLdH2O containing 1M NaCl and theinitial and residual hLF1-11 weredetermined. The hLF1-11 remaining inthe vessel after lyophilisation of the granules was also determined.

The hLF1-11 concentration was measured using a bicinchoninic acidprotein assay (Pierce) and read at 540 nm (Biorad) (hLF1-11 serialdilutions were used as a reference, the detection limit was 2.5 µg). Theaccuracy of the assay was calculated as 6.5% (mean error of true value),the precision as 3.5% (coefficient of variance). Control samples withouthLF1-11 were used for background correction. The authenticity of thehLF1-11 of the first day release samples was analysed by Q-TOF MS andsequencing of the 1375 Da peak in one of the samples.Antimicrobial Activity

Samples of 500 µL release-medium (taken after 24 h) werelyophilised (without carrier material) and 106 cfu of an MRSA clinicalisolate (ATCC BAA-811) in PBS containing 0.01% Brain Heart Infusion

was added. After 60 min. at 37 °C, these were plated on blood agar

(n=6), and colonies were counted after 18 h at 37 °C. The percentage ofkilling was calculated: [1-(cfu in sample/cfu in control)] × 100%. Over90% killing was considered active. Samples without hLF1-11 were usedas controls, samples from the first day were used because these allcontained adequate amounts of peptide.

Results

Tests of hLF1-11 confirmed stability at 37 °C in water for 26 weeks,solubility up to 150 mg/mL and no adhesion to polystyrene. Sample-weight showed only a small variation (<0.01g in samples of 0.150g). Thetable shows the amount of hLF1-11 that wasinitially available for dissociation, the quantitythat was actually released from the carrier,and the part that was left in the carrier. From the cements, less hLF1-11was extractable (2.65 - 3.43 mg/g) than had been incorporated (5 mg/g).For the granules the amount of available peptide varied (1.5 - 4.6 mg/g),the rest was recovered from the vessel after lyophilisation.

All cements displayed a sustained release profile for several days, withBiobon, Biofil and Chronos releasing significantly more hLF1-11 thanthe three other cements (Figure). All granules had burst release profiles inthe first day only; Bicalphos, Bonesave and Vitoss released significantlymore peptide than the other three granule-types (not shown).

after release three specimens

per group were finely ground

from the cements, less

hLF1-11 was extractable

Chapter 2

26

After release, a 1375 Da peak on MS-QTOF, of the correct amino-acidsequence, confirmed the integrity of hLF1-11 (control samples had no1375 Da peak). The antimicrobial activity was >90% in nine of twelvecarriers (Table).

hLF1-11 (mg/g)

Carrierreleased

biologicallyavailable

extractedafter release

Killing(%)

Cements

Biobon 0.57 ± 0.10 3.25 ± 0.18 2.47 ± 0.07 98

Biofil 0.57 ± 0.19 2.70 ± 0.13 1.96 ± 1.08 72

Bonesource 0.10 ± 0.05 2.69 ± 0.06 2.65 ± 0.45 97

Calcibon 0.22 ± 0.02 2.89 ± 0.27 2.25 ± 0.45 88

Chronos 1.69 ± 0.30 3.43 ± 0.07 2.19 ± 0.24 92

Norian 0.12 ± 0.10 2.65 ± 0.12 2.66 ± 0.30 83

Granules

Allogran 2.52 ± 0.38 2.64 ± 0.57 0.00 99

Bicalphos 4.25 ± 0.30 4.31 ± 0.60 0.07 99

Biosorb 2.53 ± 0.68 2.63 ± 0.84 0.00 99

Bonesave 1.58 ± 0.21 1.83 ± 0.40 0.00 99

Cerasorb 1.50 ± 0.20 1.52 ± 0.18 0.11 99

Vitoss 4.43 ± 1.46 4.61 ± 1.00 0.00 99

Table The hLF1-11 that was released by the carriers and theantimicrobial activity against MRSA are shown. The amount of hLF1-11available immediately after production (biologically available) and afterthe release-experiment (extracted after release) was determined by finelygrinding and addition of 1M NaCl. The values represent the mean ± SDfrom three experiments, totalling at least 12 samples per carrier-material. Sterile cultures were conservatively calculated as 99% killing,due to the lower detection limit of the assay

DiscussionLow systemic toxicity and high local drug concentrations are the main

advantages of local antibiotic delivery systems. These can be used bothfor the treatment of chronic bone infection, requiring prolonged highantibiotic concentrations and for the prevention, in which a short-duration local antibiotic peak concentration may suffice.8


27

0.5

1.0

1.5

2.0

2.5

0 0.5 1.5 3.0 24 48 72 96 120 144 168time (h)

hLF1-11(mg/g)

Chronos inject

NorianBiofil

BonesourceBiobon

Calcibon

Figure The cumulative release of hLF1-11 from cements is expressed inmg hLF1-11 per g carrier material. The increments between thedifferent time-points represent the release for that specific interval, the x-axis is interrupted between 3.0 and 24 hours. Total release from highrelease cements was: 33.8% for Chronos Inject (�), 11.5% for Biobon (�)and 11.4% for Biofil (�). This was significantly higher (p<0.05,Student’s T-test) than from low release cements: 2.7% for Calcibon (�)and 2.5% for Norian (�) 1.9% for Bonesource (�).

Using methods described by Kühn, adapted for biodegradablespecimens, we determined the available hLF1-11 in samples before andafter release.9 Not all hLF1-11 added to cement was available forextraction, even after finely grinding and adding high salt concentrationsto decrease charge-dependent binding. This suggests that part of thehLF1-11 strongly binds to the hardening cement leading to sequestrationand unavailability for release, but it might still become available afterosteoclastic resorption.

A variable amount of hLF1-11 attached to the granules duringlyophilisation, the rest was detected in the coating vessel. High-porosity

Chapter 2

28

granules (Vitoss) bound most hLF1-11 indicating that the surface area ofthe carrier material might determine the loading capacity.

Thus, two different release profiles were observed in this study, whichsupport a two-phase explanation of the release process. The initial burst-release is predominantly determined by the release of the drug from thesurface of the carrier; the second phase shows the more gradual diffusionof the drug from deeper layers, determined by the porosity of thecarrier.9 The hLF1-11 released in the first day was active in nine of thetwelve materials; of these Chronos was the highest releasing cement-typeand Vitoss the highest releasing granule-type (Table). The amount ofhLF1-11 released may be relevant in vivo, as Nibbering and co-workersdemonstrated that hLF1-11 concentrations as low as 0.1 ng/mL killMRSA in vivo.

Limitations of this study can be found in the extrapolation of the invitro results to in vivo concentrations in bone and surrounding tissues.Instead of using simulated-body-fluid as the release medium, theexperiment was performed using small volumes of water. This allowedthe use of a simple protein assay for the detection of hLF1-11.Disadvantages of this method are: 1) pH and ionic content were differentfrom the in vivo situation, 2) proteases that could degrade the hLF1-11were not present, 3) any positive interaction ofAMP with the adaptive immune system couldnot be included.5 The strength of in vitrorelease studies of this kind lies in a qualitative comparison of differentcarrier materials, allowing the identification of the most suitablematerials for in vivo testing. In spite of the in vitro shortcomings apositive correlation between in vivo and in vitro results has beenreported. 10

To conclude, in nine of the twelve combinations basic requirementsfor designing a drug delivery system (controlled release of an activesubstance) have been met.8 From these results, investigations usinganimal models to study the in vivo efficacy of AMP-releasing drug may bedesigned.

References

1. van de Belt H, Neut D, Uges DR, Schenk W, van Horn JR, van der Mei HC, BusscherHJ. (2000). Surface roughness, porosity and wettability of gentamicin-loaded bonecements and their antibiotic release. Biomaterials 21: 1981-1987.

2. Blokhuis TJ, Termaat MF, den Boer FC, Patka P, Bakker FC, Haarman HJ. (2000).Properties of calcium phosphate ceramics in relation to their in vivo behavior. JTrauma 48: 179-186.


4. Nibbering PH, Ravensbergen E, Welling MM, van Berkel LA, van Berkel PH, PauwelsEK, NuijensJH.(2001). Human lactoferrin and peptides derived from its N terminus arehighly effective against infections with antibiotic-resistant bacteria. Infect Immun 69:1469-1476.

5. van 't Hof W, Veerman ECI, Helmerhorst EJ, van Nieuw Amerongen A. (2001).Antimicrobial peptides: properties and applicability. Biol Chem 382: 597-619.

identification of the

most suitable materials


29

6. van 't Hof W, Driedijk PC, van den BM, Beck-Sickinger AG, Jung G, Aalberse RC.(1991). Epitope mapping of the Dermatophagoides pteronyssinus house dust mitemajor allergen Der p II using overlapping synthetic peptides. Mol Immunol 28: 1225-1232.

7. Faber C, Stallmann HP, Lyaruu DM, de Blieck JM, Bervoets TJ, van Nieuw AA,Wuisman PI. (2003). Release of antimicrobial peptide Dhvar-5 frompolymethylmethacrylate beads. J Antimicrob Chemother 51: 1359-1364.


9. Kühn, K.-D. (2000) Bone cements : up-to-date comparison of physical and chemicalproperties of commercial materials. Springer, Berlin, London

10. Baro M, Sanchez E, Delgado A, Perera A, Evora C. (2002). In vitro-in vivocharacterization of gentamicin bone implants. Journal of Controlled Release 83: 353-364.

Chapter 2

30

31

In vitro gentamicin release fromcommercially available calcium phosphatebone substitutes

Influence of carrier type on durationof the release profile

Hein P. Stallmann, Chris Faber, Antonius L.J.J. Bronckers,Arie V. Nieuw Amerongen, Paul I.J.M. Wuisman

BMC Musculoskelet Disord (2006), 7: 18

Background Polymethyl-methacrylate (PMMA) beads releasingantibiotics are used extensively to treat osteomyelitis, but require surgicalremoval afterwards because they do not degrade.Methods As an alternative option, this report compares the in vitrogentamicin release profile from clinically used, biodegradable carrier-materials: six injectable cements and six granule-types. Cement cylindersand coated granules containing 3% gentamicin were submerged in dH2Oand placed in a 48-sample parallel drug-release system. At regularintervals (30, 90, 180 min. and then every 24 h, for 21 days), the releasefluid was exchanged and the gentamicin concentration was measured.The activity of released gentamicin was tested on Staphylococcus aureus.Results All combinations showed initial burst-release of activegentamicin, two cements had continuous-release (17 days). The relativerelease of all cements (36-85%) and granules (30-62%) was higher thanpreviously reported for injectable PMMA-cements (up to 17%) andcomparable to other biodegradable carriers. From the cements residualgentamicin could be extracted, whereas the granules released allgentamicin that had adhered to the surface.Conclusions The high release achieved shows great promise for clinicalapplication of these biodegradable drug-carriers. Using the appropriatecombination, the required release profile (burst or sustained) may beachieved.

3

Chapter 3

32

steomyelitis is a refractory condition, potentially leading toamputation or even death. Treatment often requires multiplesurgical interventions and local or systemic antibiotictherapy.1 It predominantly affects both extremes of age;acute haematogenous infection occurs mainly in children

and chronic osteomyelitis in the elderly.2 Chronic osteomyelitis oftenrequires surgical debridement and local antibiotic treatment.Disadvantages of currently used non-biodegradable polymethyl-methacrylate (PMMA) carriers include low antibiotic release by cementsand the requirement of surgical removal in the case of PMMA-beads.3

Moreover, resistant bacteria may appear on the carrier-surface duringlater stages of low-level antibiotic release.4 In contrast, biodegradableantibiotic carriers may result in high release and obviate the need forremoval; they are gradually replaced by ingrowing tissue.5,6 Furthermore,secondary release of the antibiotic may occur during the degradationphase of the carrier, this could increase the antimicrobial efficacycompared to non-biodegradable carriers.7

To our knowledge, this is the first study to describe in vitro antibiotic-release from twelve clinically available biodegradable carriers. Therelease profile of the drug-carrier combination determines the clinicalefficacy for an important part. For prevention of infection, a high initialburst-release has been suggested, whereas treatment of chronic infectionmay require a sustained antibiotic-concentration.3,8 The materials used,six injectable cements and six pre-shaped granule types, are all calcium-phosphate based bone-defect fillers, mainly for non-weight-bearingapplications. The chemical composition of these materials is highlysimilar to natural bone; all consist of calcium-phosphate ceramics andare slowly replaced by ingrown bone.6

MethodsSample productionGentamicin-sulphate (GS) loaded cements were produced by mixing

cement powder and liquid containing 30 mg GS (Biomet MerckBioMaterials, Darmstadt, Germany) per gram powder (liquid-powder-ratio according to the manufacturer). Cylinders hardened overnight (37°C) in 6x5mm moulds. The cements were: Biobon (Biomet MerckBioMaterials), Calcibon (Biomet Merck BioMaterials), Biofil, (DePuyCMW, Blackpool, UK), Bonesource (Stryker-Leibinger, Freiburg,Germany), Chronos Inject, (Synthes, Bettlach, Switzerland) and NorianSRS (Synthes).One gram of granules was immersed in 2.0 mL of dH2O containing 30

mg GS per gram carrier-material and freeze-dried. After removal of thegranules, the residual GS in the vessel was measured. The granules were:Allogran-R (Orthos, Bristol, UK), Bicalphos (Medtronic, Memphis, TN,USA), Biosorb (Science for Biomaterials, Lourdes, France), Bonesave(Stryker-Leibinger), Cerasorb (Curasan, Kleinostheim, Germany) andVitoss (Orthovita Malvern, PA, USA).

In vitro release

33

Release experimentAs in previous experiments, specimens were immersed in 500 µL

dH2O, in sealed polystyrene 48-well plates (Costar, High Wycombe, UK)at room temperature on a shaking device (180 rpm).9,10 The water wasreplaced regularly: 30, 90 and 180 minutes on day 1, and then 24 hourlyfor 21 days (cements) or five days (granules) and stored at –20 °C.Additional time points were used in the first 24 hours in order to detect apossible burst-release pattern. Immediately after the sample-production(control) and after the release experiment (residual) three specimens pergroup were analysed. They were finely ground and suspended in 4 mLdH2O containing 1M NaCl to determine the extractable amounts of GS.The cumulative release on day 21 (cements) or five (granules) wascompared with Student’s t-test (two-tailed, p<0.05).Gentamicin determinationGentamicin concentrations were measured using an AxSym System

(Abbott Laboratories, Irving, TX, USA), which allows accurate

measurement of 0.30 to 30.00 µg/mL. Aliquots were therefore diluted inphosphate buffered saline (PBS); the values reported were corrected forthe dilution factor. All measurements were calculated as GS quantitiesfor transparent comparison to the amount of powder added duringproduction. The accuracy was calculated as 2.9 % (mean error of truevalue) and the precision as 1.6% (coefficient of variance), it was sensitiveto 0.27 µg/mL with 95% confidence.7 The values represent the mean±SD from three experiments, totalling at least nine samples per carrier-material. Cements were divided in high release and low release; thethreshold was arbitrarily set at 20 mg/g or 67% release. Since allgranules released <20 mg/g this division was not made for granules.Antimicrobial activityA killing assay was used to ascertain that gentamicin activity was

undiminished by possible interaction with the calcium-phosphate. Asdescribed earlier, samples of 500 µL release-medium (taken after 24hours) were freeze-dried and 106 cfu Staphylococcus aureus ATCC 10834were added in 100 µL of 1 mM potassium-phosphate-buffer (pH 8.0)containing 1.0% Brain Heart Infusion.10 After 60 minutes incubation at

37 °C, these were diluted in PBS and quantitatively cultured. Thepercentage of killing was calculated: [1-(cfu in sample/cfu in control)] ×100%, (>90% was considered active). Sterile cultures wereconservatively calculated as 99.8% killing, which was the detection limitof the assay. Samples without GS were used as controls; samples fromthe first day were used as representative samples for the drug-carriercombination since these all contained sufficient amounts of gentamicin.

Chapter 3

34

1

Figure 1 Production mould and CaP cylinders 6x5mm (A); the cylindersare slightly conical to facilitate extrusion. The scale bars are 5 mm.Figures 2-7 Samples of the granule types illustrate the differences inshape and macro-porosity: 2 Allogran, 3 Bicalphos, 4 Bonesave, 5Cerasorb, 6 Biosorb, 7 Vitoss

Results

The cumulative release varied from 36% to 85% (cements) and 30% to62% (granules) of the GS added. High killing percentages (99-100%)indicated intact antimicrobial activity (Table 1). Figures 1-7 show thecement-cylinders, the mould and the different irregularly porous granule-types. All cements showed burst-release in the first day, two cements(Biobon and Bonesource) showed a continuous-release for 17 days. Fromall cements residual gentamicin could be extracted, which was highest inCalcibon and Norian (Table 1). The high release cements (Chronos,Bonesource and Biofil) released more GS than low release cements(Norian and Biobon) and than low release granules Bonesave and Biosorb(p<0.05, Table 1). The release from Chronos was also higher than fromlow release cement Calcibon (p<0.05), the differences betweenBonesource, Biofil and Calcibon were not significant.In Figure 8 the release of GS from cements is expressed as a fraction

of the total cumulative release. In Figure 9 release results from thegranules are plotted. The initial release, both from cements and fromgranules, followed square-root-of-time kinetics. The granules all hadburst-release during the first day, which was related to the amount of GSthat had associated with the carrier-material during production (Table 1,control column). The release from the granules showed a larger variationthan from cements, no significant differences between the differentgranule types were observed.

The GS that did not associate with the granules during production was

2

3

4

5

6 7

In vitro release

35

recovered from the vessels in which the samples had been freeze-dried;no residual GS was extracted from the granules.

Gentamicin (mg/g)Carrier cumulative

release(%) residual control

Killing(%)

Cements

Biobon 14.0 ± 3.6 46.7 2.1 ± 2.1 17.1 ± 0.0 99.8

Biofil 23.2 ± 4.0 77.3a 0.9 ± 0.3 21.3 ± 3.9 99.7

Bonesource 25.3 ± 2.5 84.3a 1.8 ± 2.0 23.4 ± 4.5 99.8

Calcibon 16.1 ± 5.9 53.7 5.8 ± 3.8 21.2 ± 0.5 99.8

Chronos 25.6 ± 3.3 85.3a,b 0.5 ± 0.3 22.0 ± 5.5 99.1

Norian 10.8 ± 3.8 36.0 6.4 ± 0.0 16.9 ± 8.4 99.8

Granules

Allogran 15.2 ± 9.6 50.7 0.0 16.5 ± 10.0 99.8

Bicalphos 13.4 ± 8.2 44.7 0.0 15.3 ± 9.9 99.8

Biosorb 9.6 ± 2.7 32.0 0.0 11.5 ± 6.0 99.8

Bonesave 9.1 ± 5.5 30.3 0.0 10.0 ± 4.6 99.7

Cerasorb 10.8 ± 7.1 36.0 0.0 11.1 ± 7.1 99.8

Vitoss 18.6 ± 8.3 62.0 0.0 18.6 ± 10.1 99.4

Table 1 All carriers combined high cumulative release with intactantimicrobial activity, only the cements had residual GS after release.The total release is also given as percentage of the 30 mg/g that wasadded during production. (a)Total release from three high release cements(Biofil, Bonesource and Chronos) was significantly higher than from lowrelease cements Biobon and Norian and than low release granulesBonesave and Biosorb. (b)The release from Chronos was also higher thanfrom low release cement Calcibon (all p<0.05). Only two cements showedcontinuous release (Biobon and Bonesource for 17 days). Forcomparison, available gentamicin extracted from samples groundimmediately after production is shown in the control column. Values aremeans ± SD of at least three experiments in triplicate.

Chapter 3

36

Figure 8 Gentamicin released from biodegradable cements. Thecumulative released GS fraction as part of the total release is plotted as afunction of the square root of time. Initial release was linear (R2 > 0.9).Data represent cumulative values from a complete fluid exchangeexperiment.

DiscussionA safe release profile would feature a short duration of antibiotic

release after which the release should stop completely to prevent sub-inhibitory gentamicin concentrations, which could induce resistantbacteria.11 A discrete burst after implantation eradicating thecontaminating bacteria in the surgical site could be sufficient forprophylactic therapy. This would fit the burst-release seen in all carrier-materials of the current study. The continuous-release profile for 17 daysseen in Biobon and Bonesource could conform well with treatment ofestablished osteomyelitis in which a concentration ≥ 8 µg/mL for severalweeks has been propagated.12,13 Remarkably, all twelve carriers showed ahigher release of active GS than reported from injectable PMMAcements. The release in the first seven days was 36-78% for cements and30-62% for granules. Based on similar experiments both Kühn and vande Belt reported up to 17% release after seven days.13-15 Using the sameconditions as the current study, Faber and co-workers reported 18%release from small sized PMMA-cement samples.16

The GS content is a major determinant of cumulative release. Thisstudy used 3.0%, a review of current antibiotic containing PMMAcements reports a range of 2.3-3.8%.14 Another important factor is thesurface to volume ratio; our study used 6×5 mm cylinders (diameter ×height). Compared to our study, both Kühn and van de Belt used larger

Released gentamicin fraction (%)

25

50

75

100

0 5 10 15 20

(time) 1/2 [hour 1/2]

Chronos Calcibon

Biofil Bonesource

Norian Biobon

In vitro release

37

sized samples, which may have influenced the total release.13,14 Theimportance of carrier size was demonstrated in vivo by Walenkamp;small PMMA-beads established a markedly higher gentamicin release(over 90%) than larger ones.17 In contrast to biodegradable materials,PMMA-beads allow staged treatment and require surgical removal aftereach treatment episode.

Released gentam icin fraction (% )

25

50

75

100

0 1 2(tim e)

1/2 [hour

1/2]

Cerasorb B iosorb

Vitoss Allogran

Bonesave B icalphos

Figure 9 Gentamicin released from biodegradable granules. A burstrelease pattern was observed in all sample groups.

As anticipated, the antimicrobial activity was retained; gentamicin is aheat-stable aminoglycoside and has been shown to release intact fromseveral carrier-materials.5,18,19 From the cements, not all GS could beextracted, suggesting incorporation of gentamicin into the cement. Themechanism could be by interaction of the positively charged gentamicin-base with the calcium-phosphate crystal-structure during the settingreaction.10,12 This interaction could have been of influence on thecontinuous release seen in Biobon and Bonesource. Some of the (non-released) residual GS was not strongly bound to the cement matrix andcould be extracted by grinding the cement samples and high saltconcentrations to decrease charge dependant binding. This fraction mayhave been sequestered in the deeper layers, inaccessible to the releasefluid. In vivo, it could cause prolonged release from the carrier duringbiodegradation, when ingrown cells and crack formation increase thecarrier-surface exposed to extra cellular fluid. Carrier-surfacecharacteristics and drug-sequestration appear to significantly influence

Chapter 3

38

the release profile; which could be used to tailor the therapeutic effect ofdrugs.13 Favourable release properties should result in consistent safetherapeutic drug concentrations. Although all carriers used exist ofcalcium-phosphate salts, the chemical composition of the individualproducts differs. This could influence release properties by differences inmaterial characteristics as surface roughness, wettability and chemicalreactivity of the surface.13

The variation in total release was smaller in cements than in granules,possibly indicating that the cement mixing procedure gives a morehomogeneous drug-distribution than drug-adsorption onto the irregularporous granules-surface. Both the initial burst from the cements and thecomplete burst-release from the granules follows square-root-of-timekinetics (Figures 8 and 9). Such kinetics would indicate that initialrelease from these carriers is predominantly by diffusion.12,20 Thissuggests attachment of the drug both to the carrier-surface and to theinternal surface of the pores. During preparation, not all GS associatedwith the granules, indicating that the granules’ porosity and surfaceproperties may limit the amount of GS that will attach.This in vitro study is limited by the extrapolation of the results to

clinical concentrations in bone and surrounding tissues. Many differentmodels to analyse drug-release have been described, all optimised forcertain factors that influence release including: release medium, fluidexchange volume, flow rate, and interfacial gap.21-23 In this model testingconditions were rigidly kept constant for all groups, but not allphysiological parameters were mimicked: clearance of the drug by bloodflow, buffered pH and ionic content, and the high concentration in thecarrier-bone interface were not included in this model.21 This experimentused complete exchange of small volumes of water as the releasemedium, allowing comparison with our earlier experiments that reporteddose-dependent release of antimicrobial agents from PMMA andbiodegradable carriers.9,10 In previous experiments, we successfully usedcalcium-phosphate cement loaded with gentamicin to prevent infectionin a well characterised rabbit model of osteomyelitis. Bone cultures, thegold standard in osteomyelitis detection, showed a significant reductionin pathogenic bacteria.24 This stressesthe potential therapeutic options ofsuch composites.25 As a technicaladvantage, the current set-up will testseveral combinations in one experimental run, since it accommodates alarge number of isolated samples.The strength of in vitro release studies of this kind lies in a qualitative

comparison of different carrier materials, allowing the identification ofthe most suitable drug-carrier composites. Although recognising thelimitations of in vitro models, a positive correlation between in vivo andin vitro results has been reported for gentamicin-containingbiodegradable implants.26,27 This study presents several high releasecarriers. Possible advantages of cements include high and continuous

a qualitative comparison of

different carrier materials

In vitro release

39

release and easy injection into the relevant anatomic site. We havesuccessfully used antibiotic containing cement in small animal models.25

On the other hand, the granules completely released all gentamicinpresent, markedly decreasing the risk of inducing antimicrobialresistance.Prolonged in-hospital systemic antibiotic treatment of osteomyelitis

implies a heavy burden for both patient and healthcare organisation. Thelarge number of currently available non-biodegradable antibiotic carriersunderscores the demand for sound local antibiotic treatment options. Incontrast, only few approved biodegradable antibiotic delivery devices arecurrently available. This emphasizes the clinical potential ofbiodegradable antibiotic carriers and the need for comparative studies.By using materials approved for implantation in bone, the gap betweenpre-clinical and clinical studies may perhaps be narrowed, thus allowingearly adaptation and implementation of these devices.

ConclusionsFor patients requiring surgery for bone defects, the burst-release

pattern observed in all carriers, may offer clinical applications for theprevention of osteomyelitis. The prolonged release-profile from two ofthe cements might be an effective option in the concept of treatingosteomyelitis.8,28 Since all carrier-materials are commercially available,these results may readily be extended to in vivo studies to determine theoptimal combination for different clinical situations. The presentedresults support the ongoing clinical exploration of antibiotic containingbiodegradable drug-delivery systems.18,25,29

References

1. Cierny G, III, Mader JT, Penninck JJ. (2003). A clinical staging system for adultosteomyelitis. Clin Orthop Relat Res 7-24.

2. Le Saux N, Howard A, Barrowman NJ, Gaboury I, Sampson M, Moher D. (2002).Shorter courses of parenteral antibiotic therapy do not appear to influence responserates for children with acute hematogenous osteomyelitis: a systematic review. BMCInfect Dis 2: 16.



5. Wichelhaus TA, Dingeldein E, Rauschmann M, Kluge S, Dieterich R, Schafer V, BradeV. (2001). Elution characteristics of vancomycin, teicoplanin, gentamicin andclindamycin from calcium sulphate beads. J Antimicrob Chemother 48: 117-119.

6. Blokhuis TJ, Termaat MF, den Boer FC, Patka P, Bakker FC, Haarman HJ. (2000).Properties of calcium phosphate ceramics in relation to their in vivo behavior. JTrauma 48: 179-186.

7. Humphrey JS, Mehta S, Seaber AV, Vail TP. (1998). Pharmacokinetics of a degradabledrug delivery system in bone. Clin Orthop Relat Res 218-224.

8. Nijhof MW, Stallmann HP, Vogely HC, Fleer A, Schouls LM, Dhert WJ, Verbout AJ.(2000). Prevention of infection with tobramycin-containing bone cement or systemiccefazolin in an animal model. J Biomed Mater Res 52: 709-715.

Chapter 3

40

9. Faber C, Stallmann HP, Lyaruu DM, de Blieck JM, Bervoets TJ, Nieuw Amerongen AV,Wuisman PI. (2003). Release of antimicrobial peptide Dhvar-5 frompolymethylmethacrylate beads. J Antimicrob Chemother 51: 1359-1364.

10. Stallmann HP, Faber C, Slotema ET, Lyaruu DM, Bronckers AL, Nieuw AmerongenAV, Wuisman PI. (2003). Continuous-release or burst-release of the antimicrobialpeptide human lactoferrin 1-11 (hLF1-11) from calcium phosphate bone substitutes. JAntimicrob Chemother 52: 853-855.

11. van de Belt H, Neut D, van Horn JR, van der Mei HC, Schenk W, Busscher HJ. (1999).Antibiotic Resistance - to treat or not to treat? Nat Med 5: 358-359.

12. Bohner M, Lemaitre J, Van Landuyt P, Zambelli PY, Merkle HP, Gander B. (1997).Gentamicin-loaded hydraulic calcium phosphate bone cement as antibiotic deliverysystem. J Pharm Sci 86: 565-572.


14. Kühn, K. D. (2000) Bone cements : up-to-date comparison of physical and chemicalproperties of commercial materials. Springer, London, UK

15. van de Belt H, Neut D, Schenk W, van Horn JR, van der Mei HC, Busscher HJ. (2000).Gentamicin release from polymethylmethacrylate bone cements and Staphylococcusaureus biofilm formation. Acta Orthop Scand 71: 625-629.

16. Faber C, Hoogendoorn RJ, Lyaruu DM, Stallmann HP, van Marle J, Nieuw AmerongenAV, Smit TH, Wuisman PI. (2005). The effect of the antimicrobial peptide, Dhvar-5, ongentamicin release from a polymethyl methacrylate bone cement. Biomaterials 26:5717-5726.

17. Walenkamp G. (1989). Small PMMA beads improve gentamicin release. Acta OrthopScand 60: 668-669.

18. Ipsen T, Jorgensen PS, Damholt V, Torholm C. (1991). Gentamicin-collagen sponge forlocal applications. 10 cases of chronic osteomyelitis followed for 1 year. Acta OrthopScand 62: 592-594.

19. Soriano I, Evora C. (2000). Formulation of calcium phosphates/poly (d,l-lactide) blendscontaining gentamicin for bone implantation. J Control Release 68: 121-134.

20. Higuchi T. (1963). Mechanism of sustained-action medication. Theoretical analysis ofrate of release of solid drugs dispersed in solid matrices. J Pharm Sci 52: 1145-1149.

21. Hendriks JG, Neut D, van Horn JR, van der Mei HC, Busscher HJ. (2003). The releaseof gentamicin from acrylic bone cements in a simulated prosthesis-related interfacialgap. J Biomed Mater Res 64B: 1-5.

22. McLaren AC, McLaren SG, Nelson CL, Wassell DL, Olsen KM. (2002). The effect ofsampling method on the elution of tobramycin from calcium sulfate. Clin Orthop 54-57.

23. Perry AC, Rouse MS, Khaliq Y, Piper KE, Hanssen AD, Osmon DR, Steckelberg JM,Patel R. (2002). Antimicrobial release kinetics from polymethylmethacrylate in a novelcontinuous flow chamber. Clin Orthop 49-53.

24. Zuluaga AF, Galvis W, Jaimes F, Vesga O. (2002). Lack of microbiological concordancebetween bone and non-bone specimens in chronic osteomyelitis: an observational study.BMC Infect Dis 2: 8.

25. Stallmann HP, Faber C, Bronckers AL, Nieuw Amerongen AV, Wuisman PI. (2004).Osteomyelitis prevention in rabbits using antimicrobial peptide hLF1-11- orgentamicin-containing calcium phosphate cement. J Antimicrob Chemother 54: 472-476.

26. Baro M, Sanchez E, Delgado A, Perera A, Evora C. (2002). In vitro-in vivocharacterization of gentamicin bone implants. J Control Release 83: 353.

27. Torrado S, Frutos P, Frutos G. (2001). Gentamicin bone cements: characterisation andrelease (in vitro and in vivo assays). Int J Pharm 217: 57-69.

28. Nelson CL, McLaren SG, Skinner RA, Smeltzer MS, Thomas JR, Olsen KM. (2002).The treatment of experimental osteomyelitis by surgical debridement and theimplantation of calcium sulfate tobramycin pellets. J Orthop Res 20: 643-647.

29. Walenkamp GH. (2001). Gentamicin PMMA beads and other local antibiotic carriers intwo-stage revision of total knee infection: a review. Journal of Chemotherapy 13 SpecNo 1: 66-72.

41

Histatin and lactoferrin derivedpeptides: antimicrobial properties andeffects on mammalian cells

Hein P. Stallmann, Chris Faber, Antonius L.J.J. Bronckers,Jolanda M.A. de Blieck-Hogervorst, Carlo P. J. M. Brouwer,Arie V. Nieuw Amerongen, Paul I.J.M. Wuisman

Peptides (2005), 26(12): 2355-2359

Background In order to analyse the clinical potential of twoantimicrobial peptides, human lactoferrin 1-11 (hLF1-11) and synthetichistatin analogue DHVAR-5, we measured the killing effect on bacteria,and the potential toxicity on erythrocytes and bone cells.Methods The antibacterial activity was determined in a killing assay onsix strains, including methicillin resistant Staphylococcus Aureus. Theeffect on human erythrocytes and mouse bone cells was measured with ahaemolytic assay and a viability assay respectively.Results Both hLF1-11 and DHVAR-5 dose-dependently killed allbacterial strains, starting at concentrations of 6 µg/mL. hLF1-11 had noeffect on mammalian cells at concentrations up to 400 µg/mL, butDHVAR-5 induced significant haemolysis (37% at 200 µg/mL) and bonecell death (70% at 400 µg/mL).Conclusion This indicates that both peptides are able to kill variousresistant and non-resistant bacteria, but DHVAR-5 may exert a cytotoxiceffect on host cells at higher concentrations.

4

Chapter 4

42

ntimicrobial peptides (AMPs), short positively chargedpeptides, have been found in virtually all forms of life.1 Theyform a non-immune first-line of defence against invadingpathogens. Their fast killing of a broad range ofmicroorganisms, including resistant species, offers a great

opportunity for the development of newantibiotic agents.2 Potent antimicrobialshave been produced by using naturalpeptides as a template, and by isolatingactive domains of larger proteins.3 The mechanism of action of AMPs isstill controversial. Membrane permeation and intracellular killing eventshave initially been described.4 Recently, reports shift focus to immuno-modulation and stimulation of the host response to invading micro-organisms.5,6

For a possible clinical application, a high therapeutic index -indicating high antimicrobial but low cytotoxic activity - is essential. Thisindex often compares the lytic activity on human red blood cells to theantimicrobial activity.7 Peptides with a high cytotoxic activity, such asthe bee venom mellitin, can thus be distinguished from the peptides witha low cytotoxic and high antimicrobial activity.

In this study we determined the antimicrobial and cytotoxic effect oftwo well-characterised AMPs: human lactoferrin N-terminal amino acids1-11 (hLF1-11) and a synthetic histatin analogue (DHVAR-5).5,7,8 As anindication for cytotoxicity in possible intravenous therapy, the lyticactivity on human erythrocytes was measured. For the application ofAMPs in the local treatment of osteomyelitis, the effect on mouseosteoblast-like cells was determined. Osteomyelitis, a refractorycondition in which drug resistant bacteria are increasingly encountered,is a focus of our AMP-related research.8-10

Materials and methodsPeptides

Freeze-dried hLF1-11-peptide (GRRRRSVQWCA, 1375 Da) andDHVAR-5-peptide (LLLFLLKKRKKRKY, 1847 Da) were obtained fromUCB-Bioproducts (Braine-l’Alleud, Belgium). As in previous studies,Protegrin-1 (RGGRLCYCRRRFCVCVVGR, 2161 Da), a defensin-likemolecule isolated from pig leukocytes, was used as a positive control forthe antimicrobial activity assay.5,11 Cystatin-S (SSSKEENRIIPGGIYDA,1938 Da), manufactured by solid-phase peptide-synthesis using Fmoc (9-fluorenyl-methoxycarbonyl) chemistry, was used as a negative control.Reanalysis of peak fractions by reversed phase HPLC resulted in onemajor peak revealing at least 90% purity.12 The authenticity wasconfirmed by electrospray-ionisation quadrupole-time-of-flight mass-spectrometry (Q-TOF MS, Micromass Inc., Manchester, UK). Peptideswere freeze-dried in a rotational vacuum concentrator (Christ RVC 2-25,Osterode am Harz, Germany) and stored at -20°C until use.

the mechanism of action of

AMPs is still controversial.

Toxicity and activity

43

BacteriaThe following clinical isolates were obtained (Department of

Infectious Diseases Leiden University Medical Center, the Netherlands):Acinetobacter baumannii (6034) (E-test MIC-values: piperacillin >256mg/L, ciprofloxacin >32 mg/L, cotrimoxazol >32 mg/L), Escherichia coli(M1CK), methicillin resistant Staphylococcus Aureus (MRSA) ATCC BAA-811, Pseudomonas aeruginosa (7545), S. aureus ATCC 10834 andStaphylococcus epidermidis (TY 2163) (E-test MIC-values: vancomycin 6mg/L, oxacillin >256 mg/L, penicillin >32 mg/L). The bacteria werecultured overnight in brain heart infusion broth (BHI) at 37°C, diluted inBHI and cultured to mid-logarithmic growth phase for 2.5 h on a shakerat 37°C.Antimicrobial activity

The bacteria were washed three times with PBS and diluted to~5×106 cfu/mL in 1mM potassium phosphate buffer containing 1.0%BHI, as described previously.5 Equal volumes of this bacterial suspensionwere added to the peptides, (final concentration 6.25-100 µg/mL). After

60 min. at 37 °C, a serial dilution was made, plated on blood agar, and

colonies were counted after 18 h at 37 °C. The percentage of killing wascalculated: [1-(cfu sample/cfu control)] × 100%. Cystatin (100 µg/mL) orno peptide served as negative controls; Protegrin (20 µg/mL) as a positivecontrol.Haemolytic activity

Freshly collected heparin-supplemented blood from 12 healthyindividuals (after informed consent and approval of the institutionalethical committee) was centrifuged for 15 min. at 100 g to remove thebuffy coat.3,7 The erythrocytes were washed three times with PBS,centrifuged for 10 min. at 1000 g, and suspended in PBS to 1% (vol./vol.).Peptides were serially diluted in PBS and 100 µL was added in triplicateto 100 µL of the erythrocyte suspension (final concentration 6.25-200µg/mL), incubated for 1 h at 37°C and then centrifuged for 5 min. at1000 g. PBS and 1% Tween-20 were used to establish 0 and 100%haemolysis. Of the supernatants 150 µL was transferred to a flat-bottom96-well plate, and the haemoglobin release was determined by lightabsorbance at 450 nm. The percentage of intact erythrocytes wascalculated: [1 - (A450 peptide – A450 PBS) / (A450 Tween – A450 PBS)] ×100%.Cytotoxicity on bone cells

The MC3T3-E1 cells were cultured to near-confluence in α-ModifiedEagle’s Medium (Gibco, Paisley, UK) containing 10% fetal bovine serum(Gibco), 50 mg/L ascorbic acid (Merck, Darmstadt, Germany), 10 mMβ-glycerophosphate-disodium salt hydrate (Sigma, St Louis, MO, USA),300 mg/L L-glutamine (Merck), 50 mg/L gentamicin (Gibco), and1.25 mg/L fungizone (Gibco). The cells were then harvested and seededat 2.5 ×104 cells/well in a flat bottom 96-well plate (Greiner Bio-One,Solingen, Germany) in 90 µL complete medium. After 24 h (37°C, 5%

Chapter 4

44

CO2), peptides were serially diluted in PBS and 10 µL of each solutionwas added in triplicate to the wells (final concentration 6.25-400 µg/mL).Since local antibiotic treatment results in high concentrations inadjacent bone, higher peptide concentrations were added to bone cellsthan to erythrocytes. As controls PBS (no killing) and 0.1 % H2O2 (100%killing) were used.13

Peptide (µg/mL) A. baumannii E.coli P. aeruginosa

hLF1-11

100 33±6.7×102 a 14±2.1×103 a 33±3.9×102 a

50 32±3.9×103 a 52±4.1×103 a 16±3.7×103 a

25 24±3.9×104 a 32±5.4×104 a 13±2.0×104 a

12.5 51±3.2×104 a 55±4.4×104 a 38±3.2×104 a

6.25 49±1.0×105 91±3.9×105 15±0.8×105

DHVAR-5

100 6.7×102 a 6.7×102 a 6.7±3.9×102 a

50 13±6.7×102 a 2.7±2.8×103 a 4.0±1.0×103 a

25 3.0±1.4×104 a 15±9.7×104 a 5.3±2.6×104 a

12.5 4.7±1×105 56±9.1×104 a 25±9.1×104 a

6.25 27±8.8×105 57±8.5×105 4.9±3.1×105

Control 98±7.9×105 110±8.7×105 83±1.8×105

Table 1A Both hLF1-11 and DHVAR-5 caused a dose-dependent decreasein cfu of all strains tested (a: P<0.05, two-tailed Student’s t-test). In short,106 cfu/mL of various strains were incubated 60 min. at 37 ºC with hLF1-11 or DHVAR-5. As controls inactive peptide (Cystatin-S) or no peptidewere used. Results are means ± SD of at least three independentexperiments.

The effect on the morphology of cells was assessed by light-microscopy at regular intervals (30 min., 3, 6 and 24 h). After 24 h, theviability of the cells was measured by adding 10 µL cell-proliferationreagent (WST-1, Roche, Almere,Netherlands). WST-1 is cleaved toformazan by mitochondrial enzymespresent in metabolically active cells. After4 h, light absorption was measured at 450 nm; background absorption at570 nm was subtracted. The percentage of viable cells was calculated:[(A450 - A570) peptide-treated cells - (A450 - A570) medium] / [(A450 - A570)untreated cells - (A450 - A570) medium] × 100%. To preclude a synergisticeffect of medium additives, control experiments were performed in pureculture medium, yielding similar results.

WST-1 is cleaved to formazan

by mitochondrial enzymes


45

ResultsAntimicrobial activity

Both hLF1-11 and DHVAR-5 caused a dose-dependent decrease in allstrains; over 90% was killed at 12.5 µg/mL. At 6.25 µg/mL, hLF1-11showed a lower percentage of killing for S. aureus (67%) and P.aeruginosa (74%), DHVAR-5 still eradicated over 90% in all strains(Table 1). Negative controls (no peptide or Cystatin-S peptide) did notreduce bacterial count, whereas the positive control (Protegrin) caused a>90% reduction for all strains (typically 97% killing).

Peptide (µg/mL) MRSA S. aureus S. epidermidis

hLF1-11

100 2.0±2.7×103 a 8.0±1.4×103 a 6.7±6.7×102 a

50 11±2.5×103 a 27±6.6×103 a 8.7±2.0×103 a

25 5.5±2.0×104 a 21±1.3×104 a 46±1.7×103 a

12.5 37±5.0×104 a 53±5.8×104 a 15±1.3×104 a

6.25 17±4.4×105 5.3±1.0×106 36±8.4×104

DHVAR-5

100 6.7±3.9×102 a 6.7×102 a 6.7±3.9×102 a

50 3.3±1.0×103 a 1.3±2.4×103 a 2.0±1.0×103 a

25 5.3±1.0×104 a 11±2.9×103 a 19±3.7×103 a

12.5 25±1.4×104 a 10±4.4×104 a 24±1.4×104 a

6.25 26±1.7×105 16±3.2×105 11±4.1×105

Control 61±2.8×105 64±5.8×105 3.1±9.0×105

Table 1B Both hLF1-11 and DHVAR-5 caused a dose-dependentdecrease in cfu of all strains tested (a: P<0.05, two-tailed Student’s t-test).

Haemolytic effecthLF1-11 and Cystatin-S did not cause haemolysis at concentrations

up to 200 µg/mL (Figure 1A). DHVAR-5 induced significant haemolysis(37%), only at the highest concentration tested, 200 µg/mL. This was alsomacroscopically visible by orange-discoloration of the samples.Interaction with bone cells

No effect on the morphology and viability of bone cells was observedfor either hLF1-11 or Cystatin-S up to 400 µg/mL (Figure 1B). DHVAR-5did not induce changes up to 200 µg/mL; however, at 400 µg/mLDHVAR-5 induced morphologic changes consistent with cell-death. Afterthree hours the majority of the cells exposed to DHVAR-5 started toappear increasingly rounded with chromatin condensation and vacuoleformation (Figure 2). A significant decrease in viable cells was measuredafter 24 hours with the WST-1 assay (Figure 1B).

Chapter 4

46

intact erythrocytes (%)

0

20

40

60

80

100

6 50 100 200

peptide concentration (µg/mL)

*

hLF1-11 neg. control DHVAR-5

A

viable bonecells(%)

0

20

40

60

80

100

6 50 100 200 400peptide concentration (µg/mL)

*

hLF1-11 neg. control DHVAR-5

BFigure 1 The effect of hLF1-11 and DHVAR-5 on mammalian cells isshown as percentage of intact human erythrocytes (A) and viable mousebone cells (B). Values represent absorption of light by haemoglobin fromlysed erythrocytes (A) and by formazan produced by metabolically activecells after WST-1 addition (B). hLF1-11 or Cystatin-S (negative control)did not significantly alter the number of cells, but DHVAR-5 significantlyreduced intact erythrocytes and viable bone cells at the highestconcentrations tested (*: P<0.01, two-tailed Student’s t-test). Resultsrepresent at least three determinations in triplicate; average standarddeviations were 4% (intact erythrocytes) and 11% (viable bone cells).


47

DiscussionThe unique properties of AMPs may offer several options for clinical

application: broad spectrum antimicrobial activity on bacteria and fungimay result in antibiotics for treatment of drug-resistant infection.2,5

Furthermore, induction of pore-formation may facilitate the entry ofconventional drugs. Immuno-modulation and LPS-binding capacitiesmight improve the treatment of systemic infections.1,3,14

A B

C D

E

Figure 2 Bone cells exposed to 400 µg/mL hLF1-11 (A), cystatin (B) or toPBS (C) did not show degenerative changes at light microscopy(magnification 50 times). At high concentrations (400 µg/mL), DHVAR-5(D) resulted in vacuole formation and rounding of the MC3T3 bone cells,similar to addition of 0.1% H2O2 (E).

Chapter 4

48

A swift increase in drug-resistant bacterial strains, leading toincreased morbidity and costs, urgently requires new antibiotics.1,14 Inthis study, hLF1-11 and DHVAR-5 killed different strains, includingmulti drug-resistant A. baumannii. Of these two, only DHVAR-5displayed toxicity on erythrocytes and bone cells, at 16-32× theconcentration that was required for 90% bacterial killing (Figure 1A and1B).

The considerable window observed between therapeutic and toxicconcentrations could actually be larger, as AMPs may kill bacteria atlower concentrations in vivo (ng/mL) than in vitro (µg/mL).5 Importantaspects of the methods used in thisstudy include: (i) antimicrobial activitywas tested at low ionic strength toprevent possible interference ofcalcium and magnesium ions (ii) haemolytic testing was performed onblood from twelve individuals, all blood types were included (iii) a mousebone cell line was used for cytotoxicity testing, which constitutes areproducible, well described cell system but the reaction pattern of thesecells might differ from primary human bone cells. Although these factorslimit the extrapolation to in vivo effects, the consistent outcomes of theseassays, using six different bacterial strains and two different cell types,acknowledge their suitability in analysing different AMP-drugcandidates. Furthermore previous experiments using the same low ionicstrength conditions report a correlation between in vivo and in vitroactivity of AMPs.5

Without histological signs of toxicity, hLF1-11 administered in bonecement was able to prevent S. aureus bone infection in rabbits.10 Thelimited number of bacterial strains, human blood donors and differentmammalian cells characterises the transparent set-up of this study. Inorder to expand on the current results and optimise the conditions forantimicrobial therapy, a set-up involving larger numbers of bacterialstrains and cell cultures will be appropriate. On a different perspective, anumber of antimicrobial peptides have shown antitumouric activity insimilar cell-culture experiments. Therefore including tumour cell linesinto future experiment could add significantly to the understanding ofthe role of AMPs in vivo.3,15

To conclude, both hLF1-11 and DHVAR-5 killed a range of bacteria atconcentrations at which no toxicity was observed on bone cells orerythrocytes. These results underline the potential of hLF1-11 andDHVAR-5 for clinical use, either systemic or local. However, thecytotoxicity of DHVAR-5 at higher concentrations, might limit its clinicalapplication.

AMPs may kill bacteria at lower

concentrations in vivo


49

References




4. Helmerhorst EJ, Breeuwer P, Van't Hof W, Walgreen-Weterings E, Oomen LC,Veerman EC, Amerongen AV, Abee T. (1999). The cellular target of histatin 5 onCandida albicans is the energized mitochondrion. J Biol Chem 274: 7286-7291.


6. Powers JP, Hancock RE. (2003). The relationship between peptide structure andantibacterial activity. Peptides 24: 1681-1691.

7. Helmerhorst EJ, Reijnders IM, van 't HW, Veerman EC, Nieuw Amerongen AV. (1999).A critical comparison of the hemolytic and fungicidal activities of cationic antimicrobialpeptides. FEBS Lett 449: 105-110.

8. Stallmann HP, Faber C, Slotema ET, Lyaruu DM, Bronckers AL, Amerongen AV,Wuisman PI. (2003). Continuous-release or burst-release of the antimicrobial peptidehuman lactoferrin 1-11 (hLF1-11) from calcium phosphate bone substitutes. JAntimicrob Chemother 52: 853-855.

9. Faber C, Stallmann HP, Lyaruu DM, de Blieck JM, Bervoets TJ, van Nieuw AA,Wuisman PI. (2003). Release of antimicrobial peptide Dhvar-5 frompolymethylmethacrylate beads. J Antimicrob Chemother 51: 1359-1364.

10. Stallmann HP, Faber C, Bronckers AL, Nieuw Amerongen AV, Wuisman PI. (2004).Osteomyelitis prevention in rabbits using antimicrobial peptide hLF1-11- orgentamicin-containing calcium phosphate cement. J Antimicrob Chemother 54: 472-476.

11. Kokryakov VN, Harwig SS, Panyutich EA, Shevchenko AA, Aleshina GM, Shamova OV,Korneva HA, Lehrer RI. (1993). Protegrins: leukocyte antimicrobial peptides thatcombine features of corticostatic defensins and tachyplesins. FEBS Lett 327: 231-236.

12. van 't Hof W, Driedijk PC, van den BM, Beck-Sickinger AG, Jung G, Aalberse RC.(1991). Epitope mapping of the Dermatophagoides pteronyssinus house dust mitemajor allergen Der p II using overlapping synthetic peptides. Mol Immunol 28: 1225-1232.


14. Yeaman MR, Yount NY. (2003). Mechanisms of antimicrobial peptide action andresistance. Pharmacol Rev 55: 27-55.

15. Kim S, Kim SS, Bang YJ, Kim SJ, Lee BJ. (2003). In vitro activities of native anddesigned peptide antibiotics against drug sensitive and resistant tumor cell lines.Peptides 24: 945-953.

Chapter 4

50

51

In vivo release of theanti-microbial peptide hLF1-11from calcium phosphate cement

Hein P. Stallmann, Ronald de Roo, Chris Faber,Arie V. Nieuw Amerongen, Paul I.J.M. Wuisman

Submitted

Objectives We studied the release of human lactoferrin 1–11 (hLF1-11), apotent antimicrobial peptide, in a rabbit model.Methods Calcium phosphate cement with 50 mg/g hLF1-11 was injectedinto the femoral canal of twelve rabbits. One, three and seven days laterfour animals were terminated and the femora removed. Sections of boneand cement were removed for histological analysis and semi-quantitativehLF1-11 determination by Liquid Chromatography-MassSpectrometry/Mass Spectrometry (LC-MS/MS). Blood samples weredrawn for leukocyte count and differentiation to identify a potentialimmunomodulating effect of hLF1-11.Results After an initial burst release, the hLF1-11 concentration incement and bone decreased steadily over seven days. This in vivo releaseprofile is consistent with earlier in vitro studies. We observed signs ofearly bone ingrowth into the cement without signs of inflammation.Leukocytosis or a shift in differentiation did not occur.Conclusions The carrier released over 99% of the hLF1-11, resulting inpeak concentrations at the cement-bone interface. This indicates thathLF1-11 could become a valuable prophylactic agent in osteomyelitistreatment.

5

Chapter 5

52

ntimicrobial peptides (AMPs), also named host defencepeptides, play a crucial role in the innate immune system.1,2

These small peptides kill various bacteria and fungi, includingresistant strains.3-5 We describe the release kinetics of a potentAMP: hLF1-11. It consists of N-terminal amino acids one to

eleven of human lactoferrin (LF). LF is an iron-binding protein that issecreted as an effector of innate immunity in e.g. tears, saliva and milk.Besides antibacterial activity, other functions of LF have been described:biofilm permeation, complement inhibition and bone formation.6-8

In spite of increasing diagnostic and treatment options, bone infectionremains one of the most serious complications in orthopaedic surgery.Its multi-factorial incidence relates to patient (immune status), pathogen(resistance) and prosthesis (surface colonisation).9 Infection rates of 1-3% have been reported after arthroplasty. A recent study of the NorwegianArthroplasty register reports favourable results of antibiotic containingcement compared to plain cement.10 After cemented prosthesisimplantation an increased infection incidence was observed, comparedto uncemented implants. The infection rate was equal after antibioticcontaining cement and uncemented prostheses. The paper advises acombination of local and systemic antibiotics.

Staphylococcus aureus and coagulase-negative staphylococci (CoNS)have been the main cause of bone infection; during the last two decadesa shift toward more CoNS and resistant strains has been observed.11,12

Current treatment of osteomyelitis usually includes a thoroughdebridement and placement of polymethyl-methacrylate (PMMA) beadsreleasing an antibiotic e.g. gentamicin. However, the occurrence ofgentamicin resistant bacteria in patients treated with genta-beads hasrecently been reported.13 Therefore new treatment options, includingnovel antibiotics which are slowly released from carrier-matrices, arebeing explored.14-20

To our knowledge, this is the first description of in vivo hLF1-11concentrations in a controlled release study. We used an injectablecalcium phosphate (CaP) cement as the drug-carrier matrix. Thisbiodegradable cement has beenapproved for clinical treatment ofbone defects.21 Although theefficacy of this combination hasbeen demonstrated in animal studies of osteomyelitis, the mechanism ofhLF1-11 release remained unclear.22,23 By analysing the in vivo releaseprofile and duration, we aim to optimise local treatment with AMPs.

AMPs are positively charged peptides of 10 to 50 amino-acids thatform a first line-of-defence against invading microorganisms. They havebeen identified in virtually all plants and animals.24 Most AMPs have abroad spectrum of antibacterial activity, which includes resistantpathogens. Some also show antifungal or antiviral activity.4,25

Surprisingly, anti-retroviral activity has also been described: inhibition ofHIV proviral DNA formation and protection of primary human CD4-

they have been identified in

virtually all plants and animals

In vivo release

53

positive lymphocytes.26,27 AMPs also seem to have a central role inregulatory processes, including immunomodulation and angiogenesis.3,28

Bacterial resistance to AMPs is quite rare, which makes them excitingdrug candidates.29

Recent advances in peptide-chemistry will allow the manufacture oflarge quantities of highly purified AMPs from individual amino acids.30

This obviates the need for animal-based peptide-production with itsinherent risks. AMPs could be instrumental in countering the growingproblem of bacterial resistance to current antibiotics. The remainingobstacles for the practical use of peptides for clinical application mayinclude impaired local bio-availability after systemic administration andpossible degradation by proteolytic enzymes. The bio-availability of drugsmay be increased by using local release systems.22,31,32

In the current study, we evaluated the release of hLF1-11 into bone toassess in vivo kinetics in an established rabbit model. We previouslydemonstrated intact activity after in vitro release of AMPs from severalcarriers.14,32 We did not anticipate problems caused by immunologicalreactions in the rabbits to the human peptide. Firstly, previous studiesdemonstrated that transgenic production of the human LF protein byrabbits caused no inflammatory complications.33 Secondly, the hLF1-11

peptide is only a small fraction(eleven amino acids) of thecomplete LF protein (711 aminoacids), which might prevent

inflammation. The exact amino acid sequence of rabbit LF has not beendescribed yet, therefore the differences with human LF are not known.The transferrin protein, however, shows large analogies between humanand rabbit amino acid sequences.34

Resistant strains are increasingly encountered in orthopaedicsurgery.35 The spectrum of hLF1-11 includes most offendingmicroorganisms, including methicillin resistant Staphylococcus aureus(MRSA) strains.36 The already difficult treatment of osteomyelitis(debridement and antibiotic administration) is undermined by a risingnumber of resistant microorganisms.37 As a reaction to the presence ofprostheses, slow growing antibiotic tolerant strains occur.38 Thepresented drug-release system has the potential for use in the preventionof osteomyelitis caused by resistant microorganisms, a growing clinicalproblem and a focus of our AMP-related research.22,39

Materials, Animals and MethodsPeptide

HLF1-11-peptide (GRRRRSVQWCA, 1375 Da) was produced undercGMP conditions by PeptiSyntha Inc. (Los Angeles, CA, USA). Thepeptide was provided as an amorphous white powder. Stability,antimicrobial activity and solubility of the peptide have been describedpreviously.32 The powder was produced and handled under aseptic

the spectrum of hLF1-11 includes

most offending microorganisms

Chapter 5

54

conditions from single sterile batches. Handling of the hLF1-11 wasperformed in a laminar flow cabinet to prevent contamination.Animals

Fourteen healthy (specific pathogen free) adult female New ZealandWhite rabbits (Harlan, Horst, the Netherlands) weighing between 3.0-3.5kg were kept in group housing. They were assigned to the followinggroups: D1 (24 h follow-up, n = 4), D3 (72 h follow-up, n = 4), D7 (168 hfollow-up, n = 4) and control (168 h follow up, n = 2). The rabbits in thecontrol group (sham operation) received cement without hLF1-11, theothers received cement containing hLF1-11. The guidelines according tothe Dutch Act on Animal Experiments (1985) have been observed.Cement preparation

Syringes filled with 1.00 g Bonesource cement (Stryker-Leibinger,Freiburg, Germany) and 50.0 mg hLF1-11 were prepared aseptically in alaminar flow cabinet. The syringes and the cement mixing fluid werestored at 4°C during the operation, to improve the injectability of thecement.Operative procedure

The temperature and weight were measured pre-operatively, and 1.0mL blood was drawn from the central ear artery. Additional bloodsamples were taken at 3 h, 24 h, and at termination. On all samples, anautomated leukocyte count (WBC) and manual leukocyte differentiationwere determined.

The operation was performed on the animals as described by Nijhofand coworkers.40 The trochanter tertius of the right femur was surgicallyexposed and a hole to the medullary canal drilled. The marrow cavitywas reamed, flushed with saline and suctioned dry immediately prior tocement injection. The cement was made by adding 0.40 mL of mixingsolution to the previously prepared syringes containing 1.00 g of cementpowder. The cement paste was manually mixed for 2 min. in the syringe,the homogeneous paste was then gently injected. The syringe wasweighed before and after injection, to determine the amount of cementinjected. To confirm cement position, radiographs of the right femurwere made.Autopsy and sample acquisition

The rabbits were kept in individual cages during follow-up. Aftertermination with a pentobarbital overdose, both femora were asepticallyremoved. Transverse sections measuring ± 0.5 cm were sawn off theproximal femora.

Both the right femoral sections containing cement, and the leftfemora (controls) were analysed. After careful cement removal thesamples were stored at – 20°C until analysis of hLF1-11 content. Thecement in the samples for histology was not removed.Histology

We performed histological evaluation of the cement-bone interface forpossible signs of tissue reaction or infection. Sections for histology weredecalcified, embedded in paraffin, H&E stained and mounted on glass

In vivo release

55

slides. Signs of tissue reaction or infection were noted using Smeltzer’ssystem, in which a maximum score of 16 signifies severe osteomyelitis.41

Two independent observers (HPS and CF) scored all samples.hLF1-11 determination by LC-MS/MS

The samples were weighed and finely ground in 50.0 mL saline atmaximum speed in an Omnimix high-speed grinder (Sorvall, DupontInstruments, Newton, CT). We used three 20-second intervals followedby one 60-second interval. The supernatant wasremoved and stored at – 20 ºC until further analysis.

Before the measurements, the samples werediluted five times in a 0.01% acidic acid solution. Astandard dilution series was made using known concentrations of hLF1-11 and controls. LC-MS/MS determinations were performed on a triplequadrupole mass spectrometer (API3000, Applied Biosystems) by TNOQuality of Life (Business Unit Analytical Sciences, Zeist, TheNetherlands).32,42 Control bone samples without hLF1-11 were used forbackground correction.Statistics

The hLF1-11 concentrations were compared between groups with theWilcoxon signed ranks test (p<0.05 was considered significant). Forclarity, the controls (sham operated controls, and unoperated leftfemora) are presented as a pooled control after separate statisticalanalysis indicated no differences in hLF1-11 concentration. The accuracyof the assay was calculated as the mean error of the true value, and theprecision was calculated as the coefficient of variance.

WBC (109/l)Group

-0.5 h 3 h 24 h 72 h 168 h

D1 6.3 ± 1.9 6.3 ± 2.2 6.5 ± 1.5

D3 4.7 ± 1.6 4.7 ± 2.7 7.8 ± 3.0 6.4 ± 2.6

D7 7.9 ± 3.2 6.6 ± 2.4 9.1 ± 1.6 6.4 ± 1.3

Total 6.4 ± 2.5 6.5 ± 2.2 7.7 ± 2.2

Table 1 The WBC values are presented for the three rabbit groups attheir respective follow up intervals. A reaction to hLF1-11: leukocytosisor shift in differentiation was not found. The data represent mean ±SDfrom four rabbits per time-point.

ResultsThe operation was well tolerated by all animals. Clinical signs of

inflammation or impairment were not observed. Post-operatively allanimals started weight baring within 24 h. Radiographs confirmedcorrect injection of cement into the femora in all cases.

Leukocytosis or a shift in white cell differentiation did not occur, andthe average leukocyte count was 6.7 ± 0.6 × 109/mL. Table 1 shows theaverage leukocyte counts per rabbit group at each interval. Histological

a triple quadrupole

mass spectrometer

Chapter 5

56

analysis of the bone sections with hLF1-11 containing cement yielded nosigns of inflammation compared to the controls. The tissue reaction tothe drug release system was scored as minimal (<1 point out of 16),equivalent to the control score (<1 point). This low inflammatory scoreafter recent surgery confirms the minimally invasive techniques used.

At light microscopy, we observed ample cement-filling in the femoralcanal (Figure 1). The bone-cement-interface showed close contactbetween the cement and the bone (Figure 2). Histological signs of blood-vessel ingrowth and signs of cement remodelling were observed in someof the bone sections at 7-day follow-up (Figure 3). This illustrates thebiodegradability and porosity of the CaP material, which allows earlybone ingrowth.21

hLF1-11 concentration (µg/g)Group

Bone Cement Control

D1 40.9 ± 59.9 1.50 × 103 ± 1.11 × 103 a 2.3 ± 4.3

D3 0.0 ± 2.11 128 ± 86.9 -1.6 ± 1.3

D7 1.36 ± 3.46 15.8 ± 5.13 -1.7 ± 1.5

Table 2 The hLF1-11 content of samples from the right femur decreasedover time. Samples were determined by LC-MS/MS after 24, 72 and 168h. The data represent the average and the standard deviation from fourrabbits per time-point. The cement was injected into the proximal rightfemur. Sham operated rabbits and left femora were included as controls.Subtraction of a reference value from a standard dilution series wasperformed to correct for background. This resulted in some negativevalues for the controls. a: The hLF1-11 concentration in cement,measured after 24 h, was significantly higher than in control samples.

After 24 h, hLF1-11 was present in all bone and cement samples, butthe concentration decreased in later samples (Table 2). The accuracy ofthe LC-MS/MS determinations was calculated as 10.7 µg hLF1-11 pergram bone, the precision as 10.8 %.

LC-MS/MS determination was performed on a triple quadrupole massspectrometer (API3000 , Applied Biosystems) by TNO Quality of Life(Business Unit Analytical Sciences, Zeist, The Netherlands).32,42

StatisticsThe hLF1-11 concentrations were compared between groups with the

Wilcoxon signed ranks test, p<0.05 was considered significant. Theaccuracy of the assay was calculated as the mean error of the true value,the precision as the coefficient of variance. Control samples withouthLF1-11 were used for background correction.

In vivo release

57

DiscussionThe rapid and complete release of the antibiotic agent is pertinent for

adequate prophylactic treatment without risk of inducing antimicrobialresistance. Most of the available hLF1-11 peptide was released withinseven days, suggesting burst-pattern kinetics in a specified interval. TheLC-MS/MS identification of hLF1-11 indicates the release of structurallyintact peptide. The drug content of the cement is a major determinant ofthe release profile. This study used 5.0%, which is in line with a range of2.3– 3.8% reported in a review of current antibiotic containing PMMAcements.43

Table 2 shows the initial high release followed by low level release.The concentration measured in cement also showed a steady decrease inhLF1-11. Possible mechanisms for late release include crack formationand degradation of the carrier. These mechanisms allow dissolution ofthe hLF1-11 from deeper layers of the injected cement. The data wouldsuggest a slightly increased release between 72 and 168 h, but thedifferences were not significant. As the concentrations were at thedetection limit, the accuracy did not allow statistical differentiationbetween release data from 72 and 168 h. The variability of the resultsmay be improved in future studies by increasing the number of animals,using highly sensitive detection methods, and increasing the number ofintervals.

Figure 1 Biodegradable cement (arrow) adequately fills the rabbit femur

This validates earlier in vitro studies that demonstrated the release ofantimicrobially active hLF1-11from CaP and PMMA carriersusing the same dose.22,32 In theseearlier studies, the duration of therelease was mainly determined bythe choice of carrier type and material.32 The initial release processinvolved diffusion of the active peptide from the surface of the carrier.32,44

biodegradable cement combines

osteoconductivity with convenient

injectability

Chapter 5

58

The biodegradable cement combines osteoconductivity withconvenient injectability. The porous structure that results fromhardening of the cement accommodates early ingrowth of tissue andsubsequent remodelling.21

The absence of tissue inflammation or necrosis on histologicalexamination is in line with previous in vitro studies that reported nohLF1-11 cytotoxicity on blood or bone cells.5 Furthermore, the relativesmall size of the peptide (eleven amino acids) may not incite antibodyproduction as readily. This was also our experience when trying todevelop suitable antibodies for an ELISA test.

The tentative signs of early bone remodelling in the cement illustratethe versatility of such biodegradable substances as drug carriers.Naturally, to quantitatively identify a stimulating effect on bonemineralisation, which has been reported for LF in rabbit bone cellculture, extensive histomorphometry and immuno-histology would berequired.45

In other animal experiments, the same hLF1-11 dose successfullyprevented osteomyelitis by Staphylococcus aueus.22 Furthermore, in amodel of chronic infection, hLF1-11 demonstrated its therapeutic abilityto cure MRSA osteomyelitis.23 Another AMP: the synthetic histatinanalogue DHVAR-5, was also quite effective against chronic boneinfection caused by MRSA.46 These findings strengthen the concept oflocal use of AMPs to counter resistant bacteria. Further investigationsare still in order as the AMPs succeeded in a reduction of bacterial load,but complete eradication of all pathogens present was not achieved.

Radio-active labelling of AMPs with Tc99m offers exciting imagingopportunities. The real-timebiodistribution of AMPs can thus bemonitored in vivo. After intravenousinjection, hLF1-11 showed rapid renal clearance and preferentialaccumulation in liver and gut.47 In the presence of living bacteria, thelabelled AMPs accumulated at the site of infection.47 Diagnosticapplications could include evaluation of aseptic loosening of orthopaedicimplants and rejection of transplanted organs. Because of their ability todiscriminate between sterile inflammation and active infection, labelledAMPs have been proposed as a new standard for infection imaging.28

Limitations of this study are inherent to the experimental approachthat was chosen. To allow comparison to previous rabbit experiments,the rabbit femur-infection model was chosen for this study. Other studiesdescribe favourable results of AMPs in animal models of pneumonia,infected burns, muscle infection and wound infection.36,48-50

radio-active labelling of AMPs

In vivo release

59

Figure 2 A tight interface, the cement ‘hugs’ the bone; the arrowsindicate the close approximation of the cement to the bone cells..

AMPs have several mechanisms of direct antimicrobial action:because of their negative charge they may interact with positivelycharged membranes resulting inmembrane perturbation and poreformation. After entry into the cellkilling may occur by electrolytedisregulation and mitochondrial disruption.51-53 Cholesterol-rich andnegatively charged mammalian cell membranes are not affected byAMPs.3 Although AMPs rarely induce resistance and readily kill a rangeof multi-drug-resistant pathogens, some authors have reportedmechanisms of primary or acquired resistance. These include adaptationof membrane potential and modification of intracellular targets.54,55

Indirect action includes immunomodulation: in a mouse model ofthigh muscle infection hLF1-11 killed microorganisms at lowerconcentrations in vivo (ng/mL) than in vitro (µg/mL). This effect requiredthe presence of living pathogens.36 Since in our study, the animals werenot infected, immunomodulation may not have been present. Otherfactors than leukocyte count or differentiation may also be relevant. Inexperimental Klebsiella Pneumoniae infection, Bartlett and co-workersreported increased levels of interleukin 6 after administration ofnovispirin G10 peptide.56 Furthermore, the LPS binding capacity ofAMPs has been linked to treatment of sepsis. Angiogenesis, boneformation, regulation of iron absorption and prevention of biofilmformation: all appear sensitive to AMPs.6,8,57,58 These intriguing propertieshave expanded the area of AMP-research into the domain of signallingand regulating effects.

killing may occur by electrolyte

disregulation and mitochondrial

disruption

Chapter 5

60

Figure 3 Blood vessel ingrowth (arrow) into the cement demonstratesosteoconductive properties of the CaP cement.

Since AMPs have been successful in different animal models - mouse,pig, rat and rabbit - extrapolation to clinical results appears feasible.Although clinical studies on AMP efficacy are still scarce, a number oftrials underlines the potential of AMP-based drugs: treatment ofsuperficial infections, oral mucositis and paediatric sepsis.59-61

ConclusionsThis indicates that the intervention described constitutes a safe and

reliable model for the study of in vivo release of hLF1-11. In seven days,the hLF1-11 was released from the injected cement, constituting a burst-release system. These results validate thetechnique of in vivo measurement ofhLF1-11-tissue-concentrations by LC-MS/MS to analyse drug kinetics. Afterseven days, only minimal residual hLF1-11 was measured in the bonetissue. This reduces the risk of clinical induction of resistant strains. Weobserved no signs of toxicity due to local hLF1-11 release. On thecontrary, vascular growth into the cement indicated biocompability andthe potential of complete remodelling into bone.

The upsurge in resistant strains in surgical infections decreases thenumber of therapeutic options for the management of musculoskeletalinfection. This study adds to the proof of principle for treatment ofinfection with natural or designer AMPs. Additional research efforts tooptimise release kinetics from biodegradable carriers will further theevolution of drug-release systems. A balanced drug release system wouldfeature the complete release of a powerful antibiotic agent in a specifiedtime-frame, complemented by rapid remodelling of the carrier by theingrowing host cells.

treatment of infection with

natural or designer AMPs

In vivo release

61

References


2. Jenssen H, Hamill P, Hancock RE. (2006). Peptide antimicrobial agents. ClinMicrobiol Rev 19: 491-511.

3. Zasloff M. (2002). Antimicrobial peptides of multicellular organisms. Nature 415:389-395.

4. Lupetti A, Paulusma-Annema A, Welling MM, Senesi S, Van Dissel JT, NibberingPH. (2000). Candidacidal activities of human lactoferrin peptides derived fromthe N terminus. Antimicrob Agents Chemother 44: 3257-3263.

5. Stallmann HP, Faber C, Bronckers AL, Blieck-Hogervorst JM, Brouwer CP,Nieuw Amerongen AV, Wuisman PIJM. (2005). Histatin and lactoferrin derivedpeptides: Antimicrobial properties and effects on mammalian cells. Peptides.

6. Cornish J. (2004). Lactoferrin promotes bone growth. Biometals 17: 331-335.7. Samuelsen O, Haukland HH, Ulvatne H, Vorland LH. (2004). Anti-complement

effects of lactoferrin-derived peptides. FEMS Immunol Med Microbiol 41: 141-148.

8. Singh PK, Parsek MR, Greenberg EP, Welsh MJ. (2002). A component of innateimmunity prevents bacterial biofilm development. Nature 417: 552-555.

9. Lew DP, Waldvogel FA. (2004). Osteomyelitis. Lancet 364: 369-379.10. Engesaeter LB, Espehaug B, Lie SA, Furnes O, Havelin LI. (2006). Does cement

increase the risk of infection in primary total hip arthroplasty? Revision rates in56,275 cemented and uncemented primary THAs followed for 0-16 years in theNorwegian Arthroplasty Register. Acta Orthop 77: 351-358.

11. Ostendorf, M., Malchau, H., and Dhert, W. (2001) Revisions for infection from theSwedish National Hip Registry. Abstract 749

12. van de Belt H., Neut D, van Horn JR, van der Mei HC, Schenk W, Busscher HJ.(1999). Antibiotic Resistance - to treat or not to treat? Nat Med 5: 358-359.

13. van de Belt H., Neut D, van Horn JR, van der Mei HC, Schenk W, Busscher HJ.(1999). Antibiotic Resistance - to treat or not to treat? Nat Med 5: 358-359.

14. Faber C, Stallmann HP, Lyaruu DM, de Blieck JM, Bervoets TJ, NieuwAmerongen AV, Wuisman PIJM. (2003). Release of antimicrobial peptide Dhvar-5 from polymethylmethacrylate beads. Journal of Antimicrobial Chemotherapy 51:1359-1364.

15. Stallmann HP, Faber C, Slotema ET, Lyaruu DM, Bronckers AL, Amerongen AV,Wuisman PI. (2003). Continuous-release or burst-release of the antimicrobialpeptide human lactoferrin 1-11 (hLF1-11) from calcium phosphate bonesubstitutes. J Antimicrob Chemother 52: 853-855.

16. Castro C, Sanchez E, Delgado A, Soriano I, Nunez P, Baro M, Perera A, Evora C.(2003). Ciprofloxacin implants for bone infection. In vitro-in vivocharacterization. J Control Release 93: 341-354.

17. Nijhof MW, Stallmann HP, Vogely HC, Fleer A, Schouls LM, Dhert WJ, VerboutAJ. (2000). Prevention of infection with tobramycin-containing bone cement orsystemic cefazolin in an animal model. J Biomed Mater Res 52: 709-715.

18. Nijhof MW, Dhert WJ, Fleer A, Vogely HC, Verbout AJ. (2000). Prophylaxis ofimplant-related staphylococcal infections using tobramycin-containing bonecement. J Biomed Mater Res 52: 754-761.

19. Nijhof MW, Fleer A, Hardus K, Vogely HC, Schouls LM, Verbout AJ, Dhert WJ.(2001). Tobramycin-containing bone cement and systemic cefazolin in a one-stagerevision. Treatment of infection in a rabbit model. J Biomed Mater Res 58: 747-753.

20. Vogely HC, Oosterbos CJ, Puts EW, Nijhof MW, Nikkels PG, Fleer A, Tonino AJ,Dhert WJ, Verbout AJ. (2000). Effects of hydroxyapatite coating on Ti-6A1-4Vimplant-site infection in a rabbit tibial model. J Orthop Res 18: 485-493.

Chapter 5

62

21. Bloemers FW, Blokhuis TJ, Patka P, Bakker FC, Wippermann BW, Haarman HJ.(2003). Autologous bone versus calcium-phosphate ceramics in treatment ofexperimental bone defects. J Biomed Mater Res B Appl Biomater 66: 526-531.

22. Stallmann HP, Faber C, Bronckers AL, Nieuw Amerongen AV, Wuisman PIJM.(2004). Osteomyelitis prevention in rabbits using antimicrobial peptide hLF1-11-or gentamicin-containing calcium phosphate cement. J Antimicrob Chemother 54:472-476.

23. Faber C, Stallmann HP, Lyaruu DM, Joosten U, von Eiff C, Nieuw AmerongenAV, Wuisman PIJM. (2005). Comparable efficacies of the antimicrobial peptidehuman lactoferrin 1-11 and gentamicin in a chronic methicillin-resistantStaphylococcus aureus osteomyelitis model. Antimicrob Agents Chemother 49:2438-2444.

24. van 't Hof W, Veerman ECI, Helmerhorst EJ, van Nieuw Amerongen A. (2001).Antimicrobial peptides: properties and applicability. Biological Chemistry 382:597-619.

25. Cole AM, Lehrer RI. (2003). Minidefensins: antimicrobial peptides with activityagainst HIV-1. Curr Pharm Des 9: 1463-1473.

26. Cole AM, Hong T, Boo LM, Nguyen T, Zhao C, Bristol G and others. (2002).Retrocyclin: a primate peptide that protects cells from infection by T- and M-tropic strains of HIV-1. Proc Natl Acad Sci U S A 99: 1813-1818.

27. Cole AM. (2003). Minidefensins and other antimicrobial peptides:candidate anti-HIV microbicides. Expert Opin Ther Targets 7: 329-341.


29. Hancock RE. (1997). Peptide antibiotics. Lancet 349: 418-422.30. van 't Hof W, Driedijk PC, van den BM, Beck-Sickinger AG, Jung G, Aalberse RC.

(1991). Epitope mapping of the Dermatophagoides pteronyssinus house dust mitemajor allergen Der p II using overlapping synthetic peptides. Molecularimmunology 28: 1225-1232.

31. Humphrey JS, Mehta S, Seaber AV, Vail TP. (1998). Pharmacokinetics of adegradable drug delivery system in bone. Clinical orthopaedics and relatedresearch 218-224.

32. Stallmann HP, Faber C, Slotema ET, Lyaruu DM, Bronckers AL, NieuwAmerongen AV, Wuisman PIJM. (2003). Continuous-release or burst-release ofthe antimicrobial peptide human lactoferrin 1-11 (hLF1-11) from calciumphosphate bone substitutes. J Antimicrob Chemother 52: 853-855.

33. Li L, Shen W, Min L, Dong H, Sun Y, Pan Q. (2006). Human lactoferrintransgenic rabbits produced efficiently using dimethylsulfoxide-sperm-mediatedgene transfer. Reprod Fertil Dev 18: 689-695.

34. Ghareeb BA, Thepot D, Delville-Giraud C, Houdebine LM. (1998). Cloning andfunctional expression of the rabbit transferrin gene promoter. Gene 211: 301-310.

35. Berger-Bachi B, McCallum N. (2006). State of the knowledge of bacterialresistance. Injury 37 Suppl 2: S20-S25.

36. Nibbering PH, Ravensbergen E, Welling MM, van Berkel LA, van Berkel PH,Pauwels EK, Nuijens JH. (2001). Human lactoferrin and peptides derived fromits N terminus are highly effective against infections with antibiotic-resistantbacteria. Infect Immun 69: 1469-1476.

37. Garvin KL, Hinrichs SH, Urban JA. (1999). Emerging antibiotic-resistantbacteria. Their treatment in total joint arthroplasty. Clin Orthop 110-123.

38. von Eiff C., Peters G, Becker K. (2006). The small colony variant (SCV) concept --the role of staphylococcal SCVs in persistent infections. Injury 37 Suppl 2: S26-S33.

39. Coast J, Smith RD, Millar MR. (1996). Superbugs: should antimicrobialresistance be included as a cost in economic evaluation? Health Economics 5:217-226.

40. Nijhof MW, Stallmann HP, Vogely HC, Fleer A, Schouls LM, Dhert WJ, VerboutAJ. (2000). Prevention of infection with tobramycin-containing bone cement or

In vivo release

63

systemic cefazolin in an animal model. Journal of Biomedical Materials Research52: 709-715.

41. Smeltzer MS, Thomas JR, Hickmon SG, Skinner RA, Nelson CL, Griffith D, ParrTR, Jr., Evans RP. (1997). Characterization of a rabbit model of staphylococcalosteomyelitis. J Orthop Res 15: 414-421.

42. Nagle GT, Jong-Brink M, Painter SD, Li KW. (2001). Structure, localization andpotential role of a novel molluscan trypsin inhibitor in Lymnaea. Eur J Biochem268: 1213-1221.


44. Stallmann, H. P., Faber, C, Bronckers, A. L., Nieuw Amerongen, A. V., andWuisman, P. I. J. M. (2005) In vitro gentamicin release from commerciallyavailable calcium-phosphate bone substitutes Influence of carrier type onduration of the release profile Biomed Central Musculoskeletal Disorders,

45. Lorget F, Clough J, Oliveira M, Daury MC, Sabokbar A, Offord E. (2002).Lactoferrin reduces in vitro osteoclast differentiation and resorbing activity.Biochem Biophys Res Commun 296: 261-266.

46. Faber C, Hoogendoorn RJ, Stallmann HP, Lyaruu DM, Nieuw Amerongen AV,Wuisman PIJM. (2004). In vivo comparison of Dhvar-5 and gentamicin in anMRSA osteomyelitis prevention model. J Antimicrob Chemother 54: 1078-1084.

47. Welling MM, Lupetti A, Balter HS, Lanzzeri S, Souto B, Rey AM and others.(2001). 99mTc-labeled antimicrobial peptides for detection of bacterial andCandida albicans infections. J Nucl Med 42: 788-794.

48. Ceccarelli AV, Cole AM, Park AK, Tahk S, Yoshioka D, Ganz T. (2001).Therapeutic effect of a pig-derived peptide antibiotic on porcine woundinfections. Comp Med 51: 75-79.

49. Song Z, Wu H, Mygind P, Raventos D, Sonksen C, Kristensen HH, Hoiby N.(2005). Effects of intratracheal administration of novispirin G10 on a rat model ofmucoid Pseudomonas aeruginosa lung infection. Antimicrob Agents Chemother49: 3868-3874.

50. Steinstraesser L, Klein RD, Aminlari A, Fan MH, Khilanani V, Remick DG, SuGL, Wang SC. (2001). Protegrin-1 enhances bacterial killing in thermally injuredskin. Crit Care Med 29: 1431-1437.

51. Helmerhorst EJ, Breeuwer P, Van't Hof W, Walgreen-Weterings E, Oomen LC,Veerman EC, Amerongen AV, Abee T. (1999). The cellular target of histatin 5 onCandida albicans is the energized mitochondrion. J Biol Chem 274: 7286-7291.

52. Helmerhorst EJ, Troxler RF, Oppenheim FG. (2001). The human salivary peptidehistatin 5 exerts its antifungal activity through the formation of reactive oxygenspecies. Proc Natl Acad Sci U S A 98: 14637-14642.

53. Helmerhorst EJ, Murphy MP, Troxler RF, Oppenheim FG. (2002).Characterization of the mitochondrial respiratory pathways in Candida albicans.Biochim Biophys Acta 1556: 73-80.

54. Ganz T. (2001). Fatal attraction evaded. How pathogenic bacteria resist cationicpolypeptides. J Exp Med 193: F31-F34.


56. Bartlett KH, McCray PB, Jr., Thorne PS. (2003). Novispirin G10-induced lungtoxicity in a Klebsiella pneumoniae infection model. Antimicrob Agents Chemother47: 3901-3906.

57. Norrby K. (2004). Human apo-lactoferrin enhances angiogenesis mediated byvascular endothelial growth factor A in vivo. J Vasc Res 41: 293-304.

58. Iyer S, Lonnerdal B. (1993). Lactoferrin, lactoferrin receptors and ironmetabolism. Eur J Clin Nutr 47: 232-241.

59. Giles FJ, Miller CB, Hurd DD, Wingard JR, Fleming TR, Sonis ST and others.(2003). A phase III, randomized, double-blind, placebo-controlled, multinationaltrial of iseganan for the prevention of oral mucositis in patients receivingstomatotoxic chemotherapy (PROMPT-CT trial). Leuk Lymphoma 44: 1165-1172.

Chapter 5

64

60. Isaacson RE. (2003). MBI-226. Micrologix/Fujisawa. Curr Opin Investig Drugs 4:999-1003.

61. Levin M, Quint PA, Goldstein B, Barton P, Bradley JS, Shemie SD and others.(2000). Recombinant bactericidal/permeability-increasing protein (rBPI21) asadjunctive treatment for children with severe meningococcal sepsis: arandomised trial. rBPI21 Meningococcal Sepsis Study Group. Lancet 356: 961-967.

65

Osteomyelitis prevention in rabbitsusing antimicrobial peptide hLF1-11or gentamicin containingcalcium phosphate cement

Hein P. Stallmann, Chris Faber, Antonius L.J.J. Bronckers,Arie V. Nieuw Amerongen, Paul I.J.M. Wuisman

J Antimicrob Chemother 2004, 54: 472-476

Background The efficacy of prophylactic treatment with humanlactoferrin 1-11 (hLF1-11), a broad-spectrum antimicrobial peptide, wasstudied in a rabbit model of femur infection.Methods Calcium-phosphate cement with 50 mg/g hLF1-11 orgentamicin was injected into the femoral canal, after inoculation withStaphylococcus aureus. Three weeks later, slices of the proximal femorawere sawn for quantitative bacterial culture and histology.Results Treatment with hLF1-11 (p<0.038) or gentamicin (p<0.008)caused a reduction of cfu compared to the untreated control rabbits. Thenumber of sterile cultures was higher in hLF1-11 (3/7) and gentamicin(5/6) treated animals than in controls (1/7). Radiological and histologicalanalysis showed early bone ingrowth into the cement cracks. Rabbitswith positive cultures had only moderate pathological changes.Conclusions Local prophylaxis with hLF1-11 effectively reduceddevelopment of osteomyelitis in a rabbit model but gentamicin resultedin a larger number of sterile femora.

6

Chapter 6

66

one infection remains one of the most serious complicationsin orthopaedic surgery, in spite of increasing diagnostic andtreatment options. Its high impact is suffered equally interms of increased morbidity, and comprehensive use ofexpensive healthcare resources.1 Staphylococcus aureus and

coagulase negative staphylococci (CoNS) have been the main causativeorganisms; during the last two decades a shift toward more CoNS andresistant strains has been observed.2,3 Current treatment of osteomyelitisusually includes a thorough debridement and placement of polymethyl-methacrylate (PMMA) beads releasing an antibiotic e.g. gentamicin.However, the induction of gentamicinresistant bacteria in patients treatedwith genta-beads has recently beenreported.3 Therefore new treatmentoptions, including novel antibioticswhich are slowly released from carrier-matrices, are being explored.4-10

Antimicrobial peptides (AMPs), short positively-charged peptides,have been found in virtually all forms of life as a first line-of-defenceagainst invading microorganisms.11,12 They combine pore-formation innegatively-charged bacterial cell-membranes, with intracellular killingevents and a minimal propensity to induce resistance.13-15 Their fastkilling-effect on a broad range of microorganisms offers a greatopportunity for the development of new antibiotic agents.14 Thesepeptides have been proposed as a novel class of natural antibiotic agents,and might be used clinically to treat infections caused by resistantbacterial strains.11 Potent antimicrobials have been produced by usingnatural peptides as a template, and by taking the active domains of largerproteins.16-18

The antimicrobial peptide hLF1-11 consists of N-terminal amino acidsone to eleven of human lactoferrin, and has a broad antimicrobialspectrum.5,19,20 It has shown in vivo bactericidal activity after systemicadministration against a variety of microorganisms, includingmethicillin-resistant S. aureus (MRSA).20 Biodegradable calcium-phosphate (CaP) cements have been studied as carriers for the controlledrelease of many different antibiotics.6 We previously showed that hLF1-11 was still able to kill MRSA after being released from CaP carriers invitro.5 The presented study describes the efficacy of hLF1-11-loaded andgentamicin-loaded CaP-cement for the prevention of osteomyelitis in awell-characterised rabbit model.7-9

Materials and MethodsBacterial strainS. aureus (Wood 46, ATCC10832) was used to inoculate the rabbit

femora (gentamicin MIC 8 µg/mL by broth dilution method; hLF1-11LC50 6.25 µg/mL by cfu assay adapted for salt sensitive AMPs).

5,20 Afterculture in Brain Heart Infusion (BHI), a stock of aliquots for single usewas frozen. Preoperatively, samples containing approximately

treatment of osteomyelitis

usually includes a thorough

debridement

Osteomyelitis prevention

67

107 cfu/mL PBS were prepared; a volume of 0.1 mL (106 cfu) wasinjected into the femoral canal. Previous studies have shown this dosageto result in a high rate of infection in the presence of PMMA cement.9

Post-operatively the inoculum was serially diluted and plated on bloodagar; colony count after 18 hours at 37ºC confirmed accurate bacterialcontent.Antimicrobial agentsHLF1-11-peptide (GRRRRSVQWCA, 1375 Da) was manufactured by

solid-phase peptide synthesis using Fmoc (9-fluorenyl-methoxycarbonyl)chemistry as described previously.21 Reanalysis of peak fractions byreversed phase HPLC resulted in one major peak revealing at least 90%purity; the authenticity of the peptide was confirmed by electrospray-ionisation quadrupole-time-of-flight mass-spectrometry (Q-TOF MS,Micromass Inc., Manchester, UK).Gentamicin-sulphate powder (GS) was a gift from Biomet Merck

BioMaterials (Darmstadt, Germany).AnimalsTwenty-six healthy (specific pathogen free) adult female New Zealand

White rabbits (Harlan, Horst, the Netherlands) of 3.23 ± 0.24 kg werekept in group-housing. They were assigned to the following treatmentgroups: hLF1-11 (n=7), gentamicin (n=7), control (infection, noantibiotic agent, n=9) and sham (operated but not infected, n=3).The guidelines according to the Dutch ‘Act on Animal Experiments’

(1985) have been observed.Operative procedureThe operation was performed as described by Nijhof and associates.7

After sedation, blood samples were drawn to determine the leukocytecount (WBC) and the erythrocyte sedimentation rate (ESR). Thetrochanter tertius of the right femur was surgically exposed and a hole tothe medullary canal was drilled. Subsequently, 0.1 mL of bacterialsuspension (hLF1-11, gentamicin and control groups) or PBS (shamgroup) was injected, followed byinjection of the cement-paste. Thecement was prepared immediatelybefore inoculation of one gramBonesource cement-powder (Stryker-Leibinger, Freiburg, Germany), 50 mg of hLF1-11 (hLF1-11 group) orGS (gentamicin group) and 0.40 mL mixing-solution (Stryker-Leibinger).The control and sham groups received plain cement without antibioticagent. The cement-syringe was weighed before and after injection, todetermine the amount of injected cement.As a base-line image, radiographs of the right femur were made post-

operatively. Additional blood samples were drawn on days 7 and 21.Autopsy and sample acquisitionAfter 21 days the animals were killed with a pentobarbital overdose,

both femora were aseptically excised and radiographed.7 Signs of

the trochanter tertius of

the right femur was

surgically exposed

Chapter 6

68

osteomyelitis were scored using Norden’s system in which a maximumscore of seven corresponds with maximal pathogenic changes.22

Of the proximal femoral diaphysis, two transverse slices of ± 0.5 cmwere sawn; the proximal sample was used for histological and the distalone for bacteriological determination. Proximal bone sections weredecalcified, embedded in paraffin, H&E stained, and mounted on slides.Signs of infection were scored using Smeltzer’s system, in which amaximum score of sixteen signifies severe osteomyelitis.23 After carefulremoval of the cement from the distal sample, it was weighed andground in 50 mL PBS. This suspension was serially diluted and plated onblood agar; colony count after 18 hours at 37ºC was calculated torepresent cfu/g bone. Sterile cultures were conservatively calculated as4000 cfu/g (the average detection limit). A bacterial load greater than 105

cfu/g (arbitrarily set at 10% of the inoculum dose) was considered a signof full-blown infection.

cfu/g

101

102

103

104

105

106107

0 hLF1-11 control gentamicin

* *

Figure 1 Bone-cultures show a significant reduction in bacteria for boththe hLF1-11 (p<0.038) and the gentamicin (p<0.008) groups comparedto the control group (indicated by *). The average bacterial counts pergroup are shown as a horizontal line

StatisticsThe samples from all animals that completed the 21-day follow-up

period were randomised in a blinded manner and evaluated by twoobservers. The differences between groups were analysed using a two-tailed Mann-Whitney U test, p<0.05 was considered significant (valuesare means ± SD).


69

ResultsThree animals were excluded due to post-operative complications.

One rabbit in the control group died immediately post-operatively,autopsy revealed massive pulmonary embolisms; two animals sustained afemoral fracture during the follow-up period, one in the gentamicingroup and one in the control group. The other 23 animals recovered welland appeared to slightly increase in weight (0.20 ± 0.27 kg) withoutshowing signs of systemic infection.

GroupCulture

(10log cfu/g)Histology(0-16)

Radiology(0-7)

Weight change(kg)

hLF1-11 4.2 ± 1.1a 7.5 ± 2.8 2.9 ± 2.9 +0.21 ± 0.29

gentamicin 3.7 ± 0.3b 3.9 ±1.1c 1.6 ± 1.7 +0.12 ± 0.27

control 6.2 ± 1.3 6.8 ± 2.8 2.8 ± 1.5 +0.04 ± 0.20

Table 1 Characteristics of S. aureus inoculated rabbit femora afterprophylactic treatment with cement containing 50 mg/g hLF1-11 orgentamicin compared to controls. ap<0.038, bp<0.008, cp<0.045.

Quantitative culture results indicated that addition of gentamicin orhLF1-11 resulted in a significant decrease in viable bacteria (Figure 1).Table 1 summarises the results for all treatment groups: cultures, amountof injected cement, body-weight change, histological score andradiological score. Bacteria were cultured from 11 of the 23 operatedfemora: four of seven in the hLF1-11 group, one of six in the gentamicingroup and six of eight in the control group. The non-inoculated shamgroup and the un-operated left femora had sterile cultures. The amountof cement injected during the operation was 0.83 ± 0.34 g; there was nocorrelation between the amount of injected cement and the cultureresult. The pathological changes of infected rabbit femora included

intramedullar abscess formation, bone necrosis and periosteal boneformation (Figure 2). However, not every femur with positive bacterialcultures had all histological signs of infection. The biodegradability of thecement resulted in early remodelling by ingrowing bone in some of thesamples (Figure 2). The group of femora with positive bone cultures hada significantly higher histological score (7.5 ± 2.5, p=0.028) than thosewith sterile cultures (4.7 ± 2.5). The gentamicin treated rabbits had asignificantly lower score than controls; hLF1-11 treatment did not resultin a statistical difference with controls (Table 1).Overall, the femora with positive culture results had a higher averageradiographic score (3.4 ± 1.2, p=0.006) compared to sterile femora (2.1± 1.7). However, this did not result in significant radiological differencesbetween the treatment groups (Table 1). Figure 3 shows radiographs of a

Chapter 6

70

femur with pathological changes and a control femur. The ESR andWBC values showed no significant post-operative changes withinanimals or between groups (Figure 4).

A B

CFigure 2 Histological sections illustrate periosteal new bone formation(A, arrowheads, 5× magnification), empty lacunae, a sign of bonenecrosis (B, arrowheads, 10×) and an intramedullary abscess (B, arrow).The allocated disease severity score of the presented sections was:periosteal inflammation 3 points (A), chronic inflammation 3 points (B)and destruction of bone 3 points (B). Section C displays bone ingrowthinto the cement cracks (C, 10×), indicative of early remodelling 21 dayspost-operatively.

DiscussionTo our knowledge this is the first

study that describes the efficacy of alocally released AMP, hLF1-11, forprevention of osteomyelitis. Injectionof calcium-phosphate cementcontaining hLF1-11 or gentamicin into a previously inoculated femoralcanal significantly reduced the bacterial load compared to plain cement.

the number of published

animal studies is still

relatively small


71

Although gentamicin resulted in more sterile bone cultures thanhLF1-11, this difference was not statistically significant.The in vivo killing effect of natural and synthetic AMPs has been

demonstrated on a variety of microorganisms, but the number ofpublished animal studies is still relatively small.20,24-26 Intravenousinjection of low concentrations (ng/mL) of hLF1-11 effectively reducedinfection by resistant S. aureus strains in a mouse model of thigh-muscleinfection.20 Synthetic and bovine analogues of hLF1-11 havedemonstrated antimicrobial activity in mouse models by decreasingEscherichia coli bladder infection, and Toxoplasma gondii induced braincysts.24,25 Furthermore, intradermal injection of 1-6 mg Novispirin G10effectively reduced infection in rats with Pseudomonas aeruginosainfected burns.26 Finally, a number of clinical trials underlines thepotential of future AMP-based drugs for treatment of superficialinfections, oral mucositis and paediatric sepsis.27-29

A

B Figure 3 White lines indicate the saw cuts over the proximal diaphysis ofa rabbit femur without infection (A). In an X-ray of an infected rabbitfemur (B) with positive bone cultures, arrowheads indicate aninvolucrum (bone formed beneath elevated periosteum) and periostealnew bone formation. The arrow indicates the cement, which is clearlyvisible in the proximal femoral canal.

Besides direct antimicrobial killing activity through pore-formationand intracellular-targeting, AMPs are claimed to have animmunomodulating effect.15,20 This results in higher in vivo than in vitroantimicrobial activity by specific activation of signalling-cascades in thehost immune system.11,20 The in vivo antimicrobial activity afterintravenous administration of hLF1-11 at minimal (ng/mL)concentrations supports the role of an immuno-modulating effect of this

Chapter 6

72

human-derived peptide.20 Although some AMPs, like the bee venommellitin, are toxic to mammalian cells, exposure of human erythrocytesto high concentrations of hLF1-11 up to 200 µg/mL did not indicate invitro cytotoxicity of the peptide (unpublished data). Furthermore, humanallergic reactions would not be likely since hLF1-11 consists of a smallfragment of the human lactoferrin protein.

WBC (109/l)

0

2

4

6

8

10

pre-op 7 days 21 days

hLF1-11 control gentamicin

A

ESR (mm/h)

0

1

2

3

4

5

pre-op 7 days 21 days

hLF1-11 control gentamicin

B

Figure 4 Seven days post-operatively, the WBC (A) was elevated in allgroups, the ESR (B) was elevated in hLF1-11 and gentamicin treatedanimals. However, due to a large variation, no significant changes inESR and WBC during the follow-up period were found.

Rabbits are among the most frequently used animals for experimentalbone infection studies.30 Most studies report complications similar tothose described in the current study, either related to infection, tosurgery or to cement-injection.9,10 Animal models of osteomyelitis


73

prevention require a high rate of infection in the (untreated) controlanimals in order to detect differences with the treated animals. In thepresented study, infection was established in 85% of the (untreated)control animals; similar rates of infection have been reported in otherrabbit infection studies.9,31 The infection rate in this study compares wellwith other studies, modification of several factors could further improveresults and decrease the number of complications.23,30,32 Virulence of thebacterial strain, inoculum dose, presence of a biomaterial to whichbacteria preferentially adhere, surgical technique, and local tissuetrauma all influence the infection rate in untreated animals.8,31,33

The presence of tissue necrosis or a foreign material is essential in thedevelopment of osteomyelitis in animal models.8,32 The competitionbetween host-cells and bacteria to establish predominance on theimplant-surface (the race for the surface) determines whether or not theimplant becomes infected.34 The result of this conflict depends highly onthe material properties of the implant-surface. Although speculative, injectable CaP cement may influence bacterial

growth by its setting reaction and subsequent osteoconductive properties,which might explain that no infection was detected in one of sevencontrol animals (infected but untreated).In the present study, clinical results of the rabbits showed mild to

moderate signs of infection; the small increase in body-weight, the minorvariation in ESR and WBC and the absence of signs of systemic illnessall support that the induced infection was contained within the femur.Furthermore, the relatively small number of pathological changesrecorded on histological and radiological observation could be consistentwith an infection of intermediate severity. The cement in the proximalfemur could have obscured some of the pathological signs onradiographs, nonetheless the low histological scores appear to confirmthe contained and limited nature of the femur infection. Nevertheless, theshort follow-up period (three weeks) may not have been sufficient forboth the disappearance of tissue reaction due to the surgical trauma andthe full development of signs of infection in all rabbits.We used an established technique for histology and culture, but a

sample error may still haveoccurred, which could haveresulted in missed cases ofinfection.7 The difference ininfection reduction between hLF1-11 and gentamicin was not significant;adaptation of the drug delivery system may increase the cure rate. Futureimprovements could include alterations to the AMP dose, the peptidestructure and the carrier material to optimise release kinetics in vivo. Wepreviously demonstrated that part of the added hLF1-11 may bind to thecement during the setting reaction, which could limit the release. Morecomplete release occurred when higher amounts were added to thecement.4,5 These experiments support using a high dose of hLF1-11,

the added hLF1-11 may bind to the

cement during the setting reaction

Chapter 6

74

however, there are practical limitations to the scale of experimentalpeptide production and purification methods.21

In conclusion, both hLF1-11 and gentamicin reduced the number ofinfected animals and the bacterial count in infected femora. The recentobservation of resistant bacterial strains on explanted gentamicin-containing PMMA-beads (in 18 of 20 patients) highlights the clinicalimpact of antibiotic-resistance in orthopaedic infections.35 Our results,using a non-resistant strain, support the concept of using hLF1-11 in theprevention of osteomyelitis and thus warrant pre-clinical studies of boneinfection caused by resistant strains, comparing the efficacy of AMPs tocurrent antibiotic agents.

References

1. Sculco TP. (1993). The economic impact of infected total joint arthroplasty. InstrCourse Lect 42: 349-351.

2. Ostendorf, M., Malchau, H., and Dhert, W. (2001) Revisions for infection from theSwedish National Hip Registry. Sixty-eighth Annual Meeting of the American Academyof Orthopaedic Surgeons, Rosemont,IL,USA, Abstract 749

3. van de Belt H., Neut D, van Horn JR, van der Mei HC, Schenk W, Busscher HJ. (1999).Antibiotic Resistance - to treat or not to treat? Nat Med 5: 358-359.

4. Faber C, Stallmann HP, Lyaruu DM, de Blieck JM, Bervoets TJ, Nieuw Amerongen AV,Wuisman PIJM. (2003). Release of antimicrobial peptide Dhvar-5 frompolymethylmethacrylate beads. Journal of Antimicrobial Chemotherapy 51: 1359-1364.

5. Stallmann HP, Faber C, Slotema ET, Lyaruu DM, Bronckers AL, Amerongen AV,Wuisman PI. (2003). Continuous-release or burst-release of the antimicrobial peptidehuman lactoferrin 1-11 (hLF1-11) from calcium phosphate bone substitutes. JAntimicrob Chemother 52: 853-855.

6. Castro C, Sanchez E, Delgado A, Soriano I, Nunez P, Baro M, Perera A, Evora C.(2003). Ciprofloxacin implants for bone infection. In vitro-in vivo characterization. JControl Release 93: 341-354.

7. Nijhof MW, Stallmann HP, Vogely HC, Fleer A, Schouls LM, Dhert WJ, Verbout AJ.(2000). Prevention of infection with tobramycin-containing bone cement or systemiccefazolin in an animal model. J Biomed Mater Res 52: 709-715.

8. Nijhof MW, Dhert WJ, Fleer A, Vogely HC, Verbout AJ. (2000). Prophylaxis of implant-related staphylococcal infections using tobramycin-containing bone cement. J BiomedMater Res 52: 754-761.

9. Nijhof MW, Fleer A, Hardus K, Vogely HC, Schouls LM, Verbout AJ, Dhert WJ. (2001).Tobramycin-containing bone cement and systemic cefazolin in a one-stage revision.Treatment of infection in a rabbit model. J Biomed Mater Res 58: 747-753.

10. Vogely HC, Oosterbos CJ, Puts EW, Nijhof MW, Nikkels PG, Fleer A, Tonino AJ, DhertWJ, Verbout AJ. (2000). Effects of hydroxyapatite coating on Ti-6A1-4V implant-siteinfection in a rabbit tibial model. J Orthop Res 18: 485-493.


12. van 't Hof W, Veerman ECI, Helmerhorst EJ, van Nieuw Amerongen A. (2001).Antimicrobial peptides: properties and applicability. Biological Chemistry 382: 597-619.

13. Groisman EA. (1994). How bacteria resist killing by host-defense peptides. TrendsMicrobiol 2: 444-449.

14. Hancock RE. (1997). Peptide antibiotics. Lancet 349: 418-422.15. Helmerhorst EJ, Breeuwer P, Van't Hof W, Walgreen-Weterings E, Oomen LC,

Veerman EC, Amerongen AV, Abee T. (1999). The cellular target of histatin 5 onCandida albicans is the energized mitochondrion. J Biol Chem 274: 7286-7291.


75

16. Helmerhorst EJ, Van't Hof W, Veerman EC, Simoons-Smit I, Nieuw Amerongen AV.(1997). Synthetic histatin analogues with broad-spectrum antimicrobial activity.Biochem J 326 (Pt 1): 39-45.


18. van der Kraan MI, Groenink J, Nazmi K, Veerman EC, Bolscher JG, Nieuw AmerongenAV. (2004). Lactoferrampin: a novel antimicrobial peptide in the N1-domain of bovinelactoferrin. Peptides 25: 177-183.



21. van 't Hof W, Driedijk PC, van den BM, Beck-Sickinger AG, Jung G, Aalberse RC.(1991). Epitope mapping of the Dermatophagoides pteronyssinus house dust mitemajor allergen Der p II using overlapping synthetic peptides. Molecular immunology28: 1225-1232.

22. Norden CW, Myerowitz RL, Keleti E. (1980). Experimental osteomyelitis due toStaphylococcus aureus or Pseudomonas aeruginosa: a radiographic-pathologicalcorrelative analysis. Br J Exp Pathol 61: 451-460.

23. Smeltzer MS, Thomas JR, Hickmon SG, Skinner RA, Nelson CL, Griffith D, Parr TR,Jr., Evans RP. (1997). Characterization of a rabbit model of staphylococcalosteomyelitis. J Orthop Res 15: 414-421.

24. Haversen LA, Engberg I, Baltzer L, Dolphin G, Hanson LA, Mattsby-Baltzer I. (2000).Human lactoferrin and peptides derived from a surface-exposed helical region reduceexperimental Escherichia coli urinary tract infection in mice. Infect Immun 68: 5816-5823.

25. Isamida T, Tanaka T, Omata Y, Yamauchi K, Shimazaki K, Saito A. (1998). Protectiveeffect of lactoferricin against Toxoplasma gondii infection in mice. J Vet Med Sci 60:241-244.


27. Giles FJ, Miller CB, Hurd DD, Wingard JR, Fleming TR, Sonis ST and others. (2003). Aphase III, randomized, double-blind, placebo-controlled, multinational trial of isegananfor the prevention of oral mucositis in patients receiving stomatotoxic chemotherapy(PROMPT-CT trial). Leuk Lymphoma 44: 1165-1172.

28. Isaacson RE. (2003). MBI-226. Micrologix/Fujisawa. Curr Opin Investig Drugs 4: 999-1003.


30. Cremieux AC, Carbon C. (1997). Experimental models of bone and prosthetic jointinfections. Clin Infect Dis 25: 1295-1302.

31. Mayberry-Carson KJ, Tober-Meyer B, Lambe DW, Jr., Costerton JW. (1992).Osteomyelitis experimentally induced with Bacteroides thetaiotaomicron andStaphylococcus epidermidis. Influence of a foreign-body implant. Clin Orthop Relat Res289-299.

32. Norden CW, Niederriter K. (1988). Treatment of experimental chronic osteomyelitisdue to Staphylococcus aureus with LY146032. Infection 16: 27.

33. van de Belt H, Neut D, Schenk W, van Horn JR, Der Mei HC, Busscher HJ. (2001).Staphylococcus aureus biofilm formation on different gentamicin-loadedpolymethylmethacrylate bone cements. Biomaterials 22: 1607-1611.

Chapter 6

76

34. Gristina AG, Oga M, Webb LX, Hobgood CD. (1985). Adherent bacterial colonization inthe pathogenesis of osteomyelitis. Science 228: 990-993.

35. Neut D, van de BH, Stokroos I, van Horn JR, van der Mei HC, Busscher HJ. (2001).Biomaterial-associated infection of gentamicin-loaded PMMA beads in orthopaedicrevision surgery. J Antimicrob Chemother 47: 885-891.

77

Summary and Discussion 7

Chapter 7

78

esults of laboratory and animal experiments on the clinicalapplication of antimicrobial peptides (AMPs) are presentedin this thesis. The release from drug carriers, the toxicity andantimicrobial activity were examined. The peptide hLF1-11was tested as an antibiotic agent for the prevention of bone

infection.

The first Chapter familiarises the reader with several concepts used inthis thesis: bone infection, antimicrobial resistance and AMPs. Thechapter was published as a review article in a supplement to Injury. Inthis supplement the Arbeitsgemeinschaft für Osteosynthesefragen (AO)presented a selection of papers on pre-clinical and clinical aspects ofbone and implant infection.

Current treatment of bone infection involves surgical debridementand local and systemic antibiotic therapy. In the case of infectedimplants, removal may be required in spite of prolonged treatment withantibiotics. Several bacterial survival strategies frustrate effectiveantimicrobial therapy: Firstly, genetic mutations may thicken the cellularmembrane and make it impermeable to antibiotics. Secondly, bacteriamay enter a hibernation-like state of slow metabolism, which reducestheir susceptibility. Finally, bacteria cover implant surfaces withprotective slime (biofilm), which is inaccessible to most antibiotics. Theseand other mechanisms have increased the burden of infectious diseaseand have further weakened last-resort antibiotics such as vancomycinand teicoplanin. To oppose this accumulation of universal antibioticresistance, structurally novel classes of antibiotics will be required.

AMPs have several distinguishing properties that make theminteresting drug candidates:ubiquitous presence in nature,broad spectrum of activity, andonly minimal potential to induceresistance. They form a naturalfirst line-of-defence in all living organisms and show promisingtherapeutic results in several animal studies and a small number ofclinical studies. Additional interesting properties of AMPs includeimmuno-modulation, promotion of drug entry into cells and angiogenicproperties.

The focus of our experiments was local antibiotic therapy: insertion ofa drug carrier that released hLF1-11. Local therapy has severaladvantages: it results in high antibiotic concentrations without risk ofsystemic toxicity. The disadvantages are the increased morbidity and costassociated with surgery. The hLF1-11 peptide is an N-terminal fragmentof the first eleven amino acids of human lactoferrin, which is found inmilk, saliva and tears. As part of the innate immune system it protectsvulnerable mucosal surfaces from bacterial invasion. The spectrum ofantimicrobial activity includes several resistant pathogens capable ofcausing bone infection. The peptide was produced by organic chemical

immuno-modulation, promotion

of drug entry into cells and

angiogenic properties

Summary & Discussion

79

synthesis from individual amino acids. In our experiments gentamicin,an aminoglycoside antibiotic commonly used in bone cements, was usedas a control drug.

Chapters two and three compare the release of hLF1-11 andgentamicin from twelve different carrier materials. We usedbiodegradable carriers made of calcium phosphate salts. These materialsare used clinically as bone defect fillers; after implantation they areslowly resorbed by ingrowing bone cells. Two shapes of carrier wereused: cements and pre-fabricated granules. The antibiotic was mixedthrough the cement samples which hardened into cylinders. The granuleswere superficially coated with antibiotic by freeze drying. The carrierswere put into small volumes of water and the amount of drug release wasmeasured at set intervals.

All carriers showed an initial burst release pattern. Several cementsexhibited sustained release of hLF1-11 or gentamicin. Significantdifferences appeared between duration and quantity of release. Theantimicrobial activity of both gentamicin and hLF1-11 remained intactafter the release process. Since all carrier materials used have alreadybeen approved for clinical use, progression to clinical application asdrug release systems appears to be straightforward. The release profileand intact activity of the peptide were important aspects for selection ofa carrier for subsequent animal experiments. Handling properties alsoinfluenced this choice such as behaviour during mixing and injection.Bonesource cement showed good general characteristics and allowedoptimal cement injection after lowering the temperature and increasingthe liquid-to-powder ratio.

The fourth Chapter addresses the biological safety and toxicity ofAMPs. We analysed the haemolytic, cytotoxic and antimicrobialproperties of hLF1-11 and several comparable peptides. Humanerythrocytes from twelve volunteers were exposed to peptides to studypotential systemic effects after intravenous injection. After sixty minutes,haemolysis was measured by light absorbance. As an indicator for localpeptide toxicity after implantation, cultured mouse bone cells were used.Metabolic activity and morphology of the bone cells was measured up totwenty-four hours after exposure.

The results revealed that even at high peptide concentrations,hLF1-11 showed no negative effects on bone cells or erythrocytes.Contrariwise, DHVAR-5, which is a potent AMP derived from salivaryhistatin, showed some toxicity on both cell types at eight times thetherapeutic concentration. The antimicrobial activity was determined oncommon problematic hospital pathogens. All strains were decimatedafter 60 minutes exposure to hLF1-11 or DHVAR-5. This validates thebroad range of activity and rapid destruction mechanism of thesepeptides. Furthermore, the combined results from the previousexperiments encouraged us to set up animal experiments.

Chapter 7

80

Chapters five and six describe the release and antimicrobial action ofhLF1-11 in rabbit experiments. The rabbit model combines severalqualities: The size accomodates more extensive surgery than in rodentsand housing and peri-operative care are less elaborate than required forsheep or goats. Furthermore, the surgical intervention was minimallyinvasive and, as such, allowed swift recovery of the rabbits. Finally, theavailability of many other published rabbit experiments strengthens theinterpretation of our results and allows for comparative analysis.

After surgical exposure of the proximal right femur, the medullarycanal was injected with cement containing hLF1-11. The first animalstudy describes the hLF1-11 concentrations in bone and cement duringone week after implantation. At one, three and seven days after cementinjection, the samples were removed, ground and analysed for peptidecontent. A swift release of the peptide into bone was established, yielding

early peak concentrations. Thisconcurs with the release profilefound in vitro. The bones werealso histologically analysed, no

signs of inflammatory reaction were found. As a possible sign of earlyremodelling, bone and vascular ingrowth into the cement was present insome of the samples at seven days.

The second animal study evaluates the prevention of bone infectionusing hLF1-11. By the same operative approach, S. aureus, a majorcause of bone infection, was injected into the medullary canal.Immediately afterwards, we injected cement containing 5% hLF1-11, 5%gentamicin or no active ingredient. After three weeks the animals werekilled and the femora were excised. Histology and radiology showedsome signs of infection or inflammation in all groups. For quantitativebacterial culture, the cement was removed and the bone sections wereground and cultured. Both hLF1-11 and gentamicin significantlyreduced infection compared to untreated controls, and gentamicinresulted in a larger number of sterile cultures than hLF1-11.

DiscussionThese results confirm the principal conditions of an effective drug

release system: controlled release resulting in therapeutic concentrationsof a locally acting drug.1 Unlike PMMA, biodegradable carriers allowcomplete drug release in vivo due to degradation of the material byingrowing cells. We chose Bonesource cement for our animalexperiments after in vitro results had shown intact release, and notoxicity of hLF1-11.2,3 Since we were using a higher drug concentrationin vivo, an additional in vitro experiment was performed that showedadequate release of hLF1-11 at 5% drug per carrier. Current antibioticcontaining bone cements use similar drug concentrations (2-4%).4 Theeffective release of hLF1-11 was measured both by tissue concentrationand by a reduction of infection. The pharmacokinetics were dependenton the carrier type (cement or granules) and on the drug concentration

antimicrobial action of hLF1-11

in rabbit experiments


81

added to the carrier. Consequently, by altering the carrier configurationdifferent release profiles could be designed to fulfil specific clinicalrequirements.2,5 Further experiments may focus on optimisation of thedosage and carrier characteristics for specific situations.

The comparison of biodegradable carriers was performed under strictreproducible conditions in a multi-sample set-up. This resulted inconsistent results between in vitro and animal experiments. The initialburst release of hLF1-11 was followed by sustained release from severalcements. This type of release is described in several other drug-releaseexperiments. Initial burst release is considered to be a phenomenonlargely related to superficial dissolution of the drug, whereas sustainedrelease involves diffusion from deeper layers of the carrier.4,6

Prevention of infection usually involves short duration high antibioticconcentrations whereas management of existing bone infection demandsprolonged antibiotic therapy after radical surgical debridement.7 In theseexperiments, prevention of infection with hLF1-11 was explored. Inaddition, Faber and coworkers described the treatment of chronicosteomyelitis by injection of cement containing hLF1-11.8 This resultedina significant reduction of bone infection by MRSA, in a comparablerabbit model. Furthermore, DHVAR-5, another AMP, also proved to be avaluable option. Following its addition to PMMA cement the mechanicalproperties remained stable and adequaterelease was achieved both in vitro and inanimal experiments.9,10 Finally, the additionof DHVAR-5 to gentamicin containingPMMA-beads significantly increased the concomitant release ofgentamicin.10 This could prove to be beneficial in penetrating bacterialbiofilms, which may even grow on gentamicin containing PMMAcement.11,12 In models of oral biofilms, DHVAR-5 and other AMPs haveshown remarkable activity on a broad range of microorganisms.13,14

The detection of hLF1-11 in biological samples posed some analyticalchallenges. The small size of the peptide - eleven amino acids –hampered the development of selective antibodies and ELISA testing.Secondly, the presence of rabbit tissue proteins caused a relatively highbackground signal, which interfered with the measurements. Thesamples were, therefore, analysed using a combined technique of massspectrometry and liquid chromatography. This allowed for a semi-quantitative analysis of bone and cement samples from rabbits that wereearlier injected with hLF1-11 containing cement. These experiments, inwhich no infection was induced, showed tentative signs of bone ingrowthinto the cement after seven days.

In our evaluation of the animal studies we need to take into accountseveral inherent differences with the clinical situation. The small size ofthe injected cement sample favours fast release of hLF1-11. In patients,larger bones would require injection of larger cement quantities, thismight increase the duration of hLF1-11 release. The described‘immunomodulating effect’ of hLF1-11 might differ between rabbits and

beneficial in penetrating

bacterial biofilms

Chapter 7

82

humans. Nevertheless, the bactericidal effect caused byimmunomodulation of hLF1-11 has also been reported in mouseexperiments, suggesting a common theme in innate immunity.15-17

The safety was assessed in vitro on bone cells and blood cells.Extensive toxicity analysis requires several animal models from differentspecies, before allowing human toxicity studies. Currently, hLF1-11 hasbeen analysed in mice, rats and rabbits. Until now no adverse eventshave been observed at therapeutic conditions. The human origin of thedrug (lactoferrin) suggests safe use, but this should be confirmed byformal toxicity studies. Continuation of the project is therefore required.

Musculoskeletal applicationsAs a vanguard or first-line defence of the immune system, AMPs

protect susceptible body surfaces such as skin and mucosa from bacterialinvasion. Similarly, surface protection is the key to prevention andtreatment of bone and implant infection. Winning the ‘race for thesurface’ precludes colonisation of bone and implant by pathogens. Thedemonstrated efficacy of AMPs inanimal models of bone infectionsupports this mechanism: bothprevention and treatment ofexperimental infection with hLF1-11 adequately reduced the number ofoffending bacteria.8,9,18 In addition, Gabriel and coworkers demonstratedthat AMPs can be bound covalently to implant material, while retainingtheir biological activity.19

AMPs could be used clinically for detection, prevention and treatmentof infection. Radiolabelled peptides have been proposed as the goldstandard for scintigraphic imaging of infection. A consistent labellingmethod with radioactive technetium allows discrimination betweeninfection and sterile inflammation. This may shed light on the suspicionthat many cases of aseptic loosening actually involve overlookedprosthesis infection.20 Conversion from animal experiments to clinicalapplication of this technique could, therefore, be valuable for earlytreatment of occult infection. The application of this configuration in thecase of suspected implant infection might prove a valuable clinical optionin prosthesis retaining strategies.

Prevention of bone infection by local application of either hLF1-11 orDHVAR-5 was a successful approach in animal experiments. Preventionof infection by either peptide resulted in a significant reduction ofbacteria. Even in established osteomyelitis, a local application ofhLF1-11 effectively decimated the MRSA bacteria.8 The favourable effectof AMPs is not necessarily limited to bone infection. Nibbering andcoworkers report that treatment with hLF1-11 eradicated severalresistant bacterial or fungal strains in experimental thigh muscleinfections.15-17 Steinstraesser and associates report effective treatment ofskin burns by Novispirin G10 or Protegrin-1 after contamination withresistant Pseudomonas strains.21-23 Most studies show a significant

a vanguard or first-line defence

of the immune system


83

decrease in bacteria, however, complete eradication of all offendingbacteria will require further development of the methods of treatment byAMPs. A small number of clinical trials stresses the potential of suchdrugs, in for example, superficial infections, oral mucositis andpaediatric sepsis.24-26

The non-infective properties of AMPs are not yet completelyunderstood. These include stimulation of bone and blood vessel growth,suppression of tumour cells and formation of pores to promote drugentry.27-29 Moreover, as a part of lactoferrin, hLF1-11 may also influencethe iron metabolism. Further study of these mechanisms will improveour knowledge of immunity, infection and tissue repair.

DevelopmentsBone infection is a problem that has challenged clinicians over the

years. Controversy, however, still exists on aspects of prevention andtreatment: the value of prophylactic use of antibiotic containing cementwith its potential risk of inducing resistant bacteria, the duration ofintravenous and oral antibiotic treatment, the indication for one-stage ortwo-stage revision of infected implants.Long-term studies of Scandinavianimplant registries supply us with someanswers that may serve as a guide inclinical practice. The multi-factorial nature of bone and implant infectionneeds to be our focus in treating the individual patient. As mentionedbefore: the patient, the prosthesis and the pathogen all determine theoutcome. Efforts to optimise each factor (nutritional status, surfacecoating, antibiotic agents) will benefit the clinical outcome.

Since the first description of AMPs in the 1980s, several hundredshave been described. A comprehensive directory of AMPs is presented athttp://www.bbcm.univ.trieste.it/~tossi/antimic.html. As a first-linedefence, they seem to have arrived early in the evolution of species.Consequently, they have been found in all lifeforms studied for theirpresence. The initial focus on antibacterial activity has expanded sinceangiogenic, anti-tumouric, anti-viral and immuno-modulating propertiesemerged. The growing number of biological functions started a searchfor the correct name for AMPs. Alternatives such as cationic peptide,immune defence peptide, and peptide antibiotic were proposed. Thesewould more accurately address certain properties or functions of AMPs.As one of the first AMP scientists Robert Lehrer stated at the 2003Gordon Conference: “The device I have in my hand has many functions.It opens cans, and cuts paper, it has a toothpick and also a smallmagnifying glass. We still call it a Swiss army pocket-knife, andeverybody understands what we are talking about. Why complicatethings by changing its name every time a new function is added? Let’sstick with AMP.”

The process of AMP research slowly matures from exploration tocharacterisation and ultimately application. Initial studies screened

Scandinavian implant registries

supply us with some answers

Chapter 7

84

thousands of biological samples for active compounds. For example,sources as diverse as amphibianskin secretions, pig leukocytes,and human tears, all containAMPs. These peptides can be synthesised from individual amino acidsand altered to improve activity. Further characterisation of AMPsinvolved the interaction with cellular membranes, intracellular processesand intercellular signalling.30 Mechanisms of action include poreformation, disruption of mitochondria, chemotaxis and cytokine release.Finally, a number of toxicity and animal studies has focused on potentialtherapeutic applications. Diverse animal models include: skin burninfection, pneumonia, bone and soft tissue infection.8,17,31,32 These modelsdescribe both local and systemic therapy: topical application, localinjection, tracheal instillation and controlled release of AMPs.

Future considerationsIn our view, the robust performance of AMPs in animal experiments

dictates further exploration in clinical studies. However, the anticipationof new antibiotics should not postpone execution of urgent measures tocontain antimicrobial resistance: restricted antibiotic use in healthcare, abar on inappropriate use in animals, and improving healthcare facilitiesin developing countries.33

Multi-resistant ‘superbugs’ such as MRSA and VRSA have created asense of urgency in drug development.34,35 Nevertheless, introduction ofAMP-based drugs should be preceded by established drug-efficacy,toxicity and safety studies. The central role of AMPs in innate immunityrequires special examination before clinical introduction. The completerange of biological signalling effects of AMPs still remains a considerablechallenge to peptide researchers. It is, therefore, worthwhile to considerwhether clinical use of antibiotics based on AMPs could affect our innateimmune system. Furthermore, antimicrobial resistance to AMPs appearsto be difficult but is not impossible.36 One could speculate on thedifficulties our body’s defence system could have in dealing with AMP-resistant ‘superbugs’. Consequently, pathogens that have becomeresistant to AMPs might present a serious medical situation.

As often in medicine, solutions present themselves in emulatingnature’s methods. Using the shotgunapproach by developing a diversepalette of AMP-derived antibioticsfrustrates bacterial development ofresistance. Understanding the mechanism of immunomodulation byAMPs will improve their specific and selective therapeutic application.Finally, similar to the body combining several defence systems, synergyof old and new antibiotics may effectively kill pathogens. New antibioticsmay perforate the bacteria to allow entry of established drugs. Thesedevelopments require further analysis of drug kinetics and dose-response

call it a ‘Swiss army pocketknife’

perforate the bacteria to allow

entry of established drugs


85

optimisation. Depending on the clinical application, several methods ofAMP administration may be used.9,17,21,31

ConclusionsPrevention of bone infection is a major concern in orthopaedic

surgery. Prophylactic use of antibiotic containing implants may causeresistant strains. AMPs could be a valuable treatment option because oftheir low tendency to induce resistance.

Calcium phosphate cements and granules may be used as antibioticcarriers with predictable pharmacokinetics. They combinebiodegradability and osteoconductivity with proven clinical safety.Because they are replaced by bone, surgical removal is not required aftertreatment.

hLF1-11 may kill a range of pathogens, including resistant strains andhas shown effectivity in animal models of osteomyelitis and other types ofinfection. Further development of this concept may decrease the highmorbidity and costs of orthopaedic infection. Its reported diagnosticapplication to distinguish between infection and sterile inflammation isalso a major topic of further research.

References





5. Faber C, Stallmann HP, Lyaruu DM, de Blieck JM, Bervoets TJ, Nieuw Amerongen AV,Wuisman PIJM. (2003). Release of antimicrobial peptide Dhvar-5 frompolymethylmethacrylate beads. Journal of Antimicrobial Chemotherapy 51: 1359-1364.


7. Engesaeter LB, Lie SA, Espehaug B, Furnes O, Vollset SE, Havelin LI. (2003).Antibiotic prophylaxis in total hip arthroplasty: effects of antibiotic prophylaxissystemically and in bone cement on the revision rate of 22,170 primary hipreplacements followed 0-14 years in the Norwegian Arthroplasty Register. Acta OrthopScand 74: 644-651.



10. Faber C, Hoogendoorn RJ, Lyaruu DM, Stallmann HP, van Marle J, Nieuw AmerongenAV, Smit TH, Wuisman PIJM. (2005). The effect of the antimicrobial peptide, Dhvar-5,

Chapter 7

86

on gentamicin release from a polymethyl methacrylate bone cement. Biomaterials 26:5717-5726.




14. Rothstein DM, Helmerhorst EJ, Spacciapoli P, Oppenheim FG, Friden P. (2002).Histatin-derived peptides: potential agents to treat localised infections. Expert OpinEmerg Drugs 7: 47-59.





19. Gabriel M, Nazmi K, Veerman EC, Nieuw Amerongen AV, Zentner A. (2006).Preparation of LL-37-grafted titanium surfaces with bactericidal activity. BioconjugChem 17: 548-550.

20. Nelson CL, McLaren AC, McLaren SG, Johnson JW, Smeltzer MS. (2005). Is asepticloosening truly aseptic? Clin Orthop Relat Res 25-30.





25. Giles FJ, Miller CB, Hurd DD, Wingard JR, Fleming TR, Sonis ST and others. (2003). Aphase III, randomized, double-blind, placebo-controlled, multinational trial of isegananfor the prevention of oral mucositis in patients receiving stomatotoxic chemotherapy(PROMPT-CT trial). Leuk Lymphoma 44: 1165-1172.

26. Isaacson RE. (2003). MBI-226. Micrologix/Fujisawa. Curr Opin Investig Drugs 4: 999-1003.




87


30. Mangoni ML. (2006). Temporins, anti-infective peptides with expanding properties.Cell Mol Life Sci 63: 1060-1069.




34. Coast J, Smith RD, Millar MR. (1996). Superbugs: should antimicrobial resistance beincluded as a cost in economic evaluation? Health Economics 5: 217-226.



Chapter 7

88

89

Samenvatting en Discussie 8

Hoofdstuk 8

90

esultaten van laboratorium en dierexperimenteel onderzoeknaar de klinische toepassing van antimicrobiële peptiden(AMPs) worden in dit proefschrift beschreven. Wijbestudeerden de afgifte uit een dragerstof, de toxiciteit en deantibacteriële werking. Hiermee werd de toepassing van het

peptide hLF1-11 onderzocht voor de preventie van botinfectie.

Het eerste hoofdstuk brengt een aantal kenmerkende begrippen naarvoren: botinfectie, resistente bacteriën en AMPs. Het hoofdstuk is als eenoverzichtsartikel opgenomen in een supplement van het blad Injury. Indit themanummer presenteert de Arbeitsgemeinschaft fürOsteosynthesefragen (AO) een set artikelen over wetenschappelijke enklinische kanten van bot- en implantaatinfectie.

De huidige behandeling van botinfectie combineert chirurgischebehandeling van de infectie met lokale en systemische antibiotischetherapie. Ondanks langdurige behandeling met antibiotica is het vaaknoodzakelijk om geïnfecteerde prothesen te verwijderen. Effectievebehandeling wordt op meerdere manieren door de bacteriëngedwarsboomd: Allereerst kunnen genetische mutaties een dikkerecelwand opleveren, ondoordringbaar voor antibiotica. Verder kunnenbacteriën in ‘winterslaap’ gaan, doorhun vertraagde metabolisme wordenzij dan ongevoelig voor antibiotica.Op implantaten, ten slotte, bedekkenzij zich met slijm (biofilm), waarin demeeste antibiotica niet kunnen doordringen. Naast deze drie, bestaannog meer mechanismen, die infecties bevorderen en zelfs de krachtigsteantibiotica zoals vancomycine en teicoplanine onwerkzaam maken. Omprogressie naar universele antibioticaresistentie tegen te gaan, zijncompleet nieuwe antibioticumsoorten nodig.

Een aantal nuttige eigenschappen maken AMPs interessant voor deontwikkeling van geneesmiddelen: het zijn natuurlijke afweerstoffen,actief tegen veel verschillende bacteriën, en veroorzaken nauwelijksantibioticumresistentie. Zij vormen een primaire afweer in alle levendeorganismen en hebben in een aantal dierstudies en klinische studiesveelbelovend gepresteerd. Daarnaast hebben AMPs nog een aantalbelangwekkende eigenschappen: zij hebben reguleren de afweer,bevorderen de groei van bloedvaten en helpen andere geneesmiddeleneen cel binnen.

De nadruk ligt bij onze experimenten op lokale therapie met antibiotica:het operatief inbrengen van een drager die hLF1-11 afgeeft. Devoordelen van lokale afgifte zijn: hoge concentratie lokaal zonderbijwerkingen elders in het lichaam. De nadelen zijn: de vereiste operatiemet bijbehorende risico’s en kosten. Het hLF1-11 peptide is een fragment

Effectieve behandeling wordt

op meerdere manieren door

de bacteriën gedwarsboomd

Samenvatting & Discussie

91

van het eiwit lactoferrine. Lactoferrine is in de mens o.a. in melk,speeksel en tranen aanwezig. De eerste elfaminozuren (vanaf de N-zijde) vormenhLF1-11. Het is een onderdeel van de afweertegen ziekten en beschermt de kwetsbareslijmvliezen tegen infectie. Het werkingsgebied omvat meerdereresistente soorten die botinfecties zouden kunnen veroorzaken. Hetpeptide is gemaakt door chemische synthese van losse aminozuren. In deexperimenten gebruikten wij gentamicine, een standaard antibioticum inbotcement, ter vergelijking.

Hoofdstukken twee en drie vergelijken de afgifte van hLF1-11 engentamicine uit twaalf verschillende dragers. Het waren biologischafbreekbare dragers van calciumfosfaat. Deze worden gebruikt ombotdefecten te vullen en worden na de operatie langzaam vervangen doorlevend bot. Wij gebruikten twee soorten: cement en korrels. Hetantibioticum werd door het cement gemengd, wat daarna uithardde totcylinders. De korrels, daarentegen, werden bedekt met een dun laagjeantibioticum met een vries-droog techniek. Op vaste tijden werdvervolgens de afgifte gemeten in kleine volumes water.

Alle geteste combinaties lieten meteen hoge antibioticum afgifte zien.Een aantal cementen vertoonde langduriger afgifte van hLF1-11 ofgentamicine. Er bleek sprake van significante verschillen in duur enhoeveelheid van afgifte tussen de verschillende materialen. Deantibiotische activiteit van hLF1-11 en van gentamicine bleef intact.Aangezien de gebruikte dragers al zijn goedgekeurd voor medischetoepassing, lijkt de stap naar gebruik als systeem voor gecontroleerdeafgifte eenvoudig. Voor de selectie van een drager voor dedierexperimenten waren afgifte en intacte activiteit uiteraard van belang.Verder speelden praktische overwegingen een rol; gemak bij mengen,hanteren en injecteren. Bonesource cement had goede algemeneresultaten en gaf bij lage cement temperatuur een hoge vloeibaarheid enoptimale injectie resultaten.

Het vierde hoofdstuk behandelt de veiligheid,toxiciteit en activiteit van AMPs. Wij bestudeerdende effecten van hLF1-11 en andere peptiden op rodebloedcellen, botcellen en bacteriën. Twaalf vrijwilligers stonden bloed afwat blootgesteld werd aan peptiden, om mogelijke effecten naintraveneuze injectie te analyseren. Na 60 minuten werd vervolgens hetaantal intacte rode bloedcellen gemeten. Voor het meten van een lokaaleffect op bot na implantatie gebruikten wij muizecellen in weefselkweek.Eén dag na blootstelling aan de peptiden werd de stofwisseling van decellen gemeten en werd onder de microscoop gezocht naar tekenen vanceldood.

lactoferrine is een onderdeel

van de afweer tegen ziekten

twaalf vrijwilligers

stonden bloed af

Hoofdstuk 8

92

Zelfs bij de hoogste concentraties veroorzaakte hLF1-11 geen schade aande bot- of bloedcellen. DHVAR-5, een krachtige AMP afgeleid van hetspeekselpeptide histatine, toonde echter schade aan beide celtypes bij eenconcentratie van achtmaal de therapeutische dosis. De antimicrobiëleactiviteit was bepaald op veroorzakers van moeilijk behandelbare‘ziekenhuisinfecties’. Zestig minuten blootstelling van deze bacteriën aanhLF1-11 of DHVAR-5 gaf een sterke afname in aantal. Dit bevestigt hetbrede spectrum en de snelle werking van deze peptiden. Uitgaande vande gecombineerde resultaten van de eerder genoemde studies, hebbenwij vervolgens de dierproeven opgezet.

Hoofdstukken vijf en zes beschrijven de afgifte en antibiotische werkingvan hLF1-11 in konijnen. Hetkonijn heeft enkele voordelen: deafmeting laat grotere operatiestoe dan bij ratten of muizen, maarverzorging en huisvesting zijn minder uitgebreid dan bij schapen ofgeiten. Gezien de beperkte operatiebelasting was snel herstel van dekonijnen te verwachten. Ten slotte tonen vele andere konijnstudies debruikbaarheid van dit model aan en versterken daarmee de analyse vanonze resultaten.

Na chirurgische benadering van het bovenste stuk van het rechterdijbeen, werd cement met hLF1-11 in de mergholte van het bot gespoten.De eerste dierstudie beschrijft metingen van hLF1-11 concentraties inbot en cement tijdens de eerste week na de operatie. Na één, drie enzeven dagen werden stukjes bot en cement verwijderd, gemalen en dehLF1-11 concentraties geanalyseerd. Het peptide werd snel afgegeven inhet bot, dit gaf vroege piek-concentraties en daarna weer snelle afname.Dit past bij het profiel dat in de eerdere laboratorium-experimenten werdgevonden. De botten werden ook onder de microscoop beoordeeld: erwaren geen tekenen van ontstekingsreactie. Als mogelijke tekenen vanvroege weefsel-ingroei in het cement was er in enkele gevallen sprakevan bot- en vaatingroei.

De tweede dierstudie evalueert de preventie van botinfectie met hLF1-11.Met dezelfde operatie werd éérst Staphylococcus aureus, een belangrijkeveroorzaker van botinfectie in het mergkanaal gespoten. Meteen daarnawerd cement geïnjecteerd met 5% hLF1-11, 5% gentamicine of zonderantibioticum. Na drie weken werden de dieren gedood en de geopereerdebotten verwijderd. In alle groepen werden tekenen van ontsteking ofinfectie gezien op röntgenfoto’s en microscopiebeelden. De botten en hetverwijderde cement werden gemalen en gekweekt voor bacterietelling.Zowel hLF1-11 als gentamicine gaven een significante afname vaninfectie vergeleken met de onbehandelde controlegroep, gentamicine gafeen groter aantal steriele kweken dan hLF1-11.

het konijn heeft enkele voordelen: de

afmeting laat grotere operaties toe


93

Discussie

Deze resultaten tonen de essentiële voorwaarden van een effectiefafgiftesysteem voor geneesmiddelen: gecontroleerde afgifte met alsgevolg therapeutische concentraties van een lokaal werkzaam middel.1

De biologische afbreekbaarheid van het hier gebruikte calciumfosfaatmateriaal impliceert volledige afgifte van het geneesmiddel na ingroeivan botcellen, in tegenstelling tot niet afbreekbaar PMMA cement. Wijhebben Bonesource cement met hLF1-11 gekozen voor dedierexperimenten op basis van de eerdereresultaten. De activiteit van hLF1-11 bleefintact na afgifte uit cement, en er was geentoxiciteit van hLF1-11 aantoonbaar. De afgiftevan hLF1-11 was adequaat in de dosering van 5% die werd gebruikt bijde dierexperimenten.2,3 De huidige botcementen met antibioticumbevatten vergelijkbare concentraties (2-4%).4

De effectieve hLF1-11 afgifte werd op twee manieren bepaald. Allereerstwerd de concentratie in bot en in cement gemeten, daarnaast werd eenreductie van het aantal bacteriën gezien. In deze studies werd hetafgifteprofiel beïnvloed door de keuze van drager (cement of korrels) endoor de dosis antibioticum die werd toegevoegd. Hieruit volgt dat deconfiguratie van de drager de mogelijkheid biedt om verschillendeafgifteprofielen te creëren, afhankelijk van klinische eisen.2,5 Verdereexperimenten zouden de optimalisatie van de dosis en de dragereigenschappen voor specifieke situaties betreffen.

De analyse van afbreekbare dragers werd onder duidelijkreproduceerbare omstandigheden verricht met gelijktijdige meting vanverschillende materialen. Dit gaf eenduidige resultaten, dieovereenkwamen tussen laboratorium- en dierproeven. De aanvankelijkesnelle afgifte van hLF1-11 werd bij enkelecementen gevolgd door langdurigeafgifte. De eerste afgiftepiek wordt metname gerelateerd aan het oplossen vande hoeveelheid geneesmiddel gelegen aan de oppervlakte van de drager.Langdurige afgifte wordt verklaard als diffusie van het middel vanuitdieper gelegen lagen van de drager.4,6

De preventie van infectie met antibiotica vereist meestal kortdurendehoge concentraties. Behandeling van bestaande infectie, daarentegen,vereist operatieve sanering en langdurige antibiotische behandeling.7 Inde voorgaande experimenten werd de preventie van infectie met hLF1-11uitgezocht. Daarnaast beschrijven Faber c.s. de behandeling vanlangdurig bestaande botinfectie door injectie van cement met hLF1-11. 8

Dit gaf een duidelijke afname van MRSA infectie in een vergelijkbaarexperiment met konijnen. Bovendien bleek DHVAR-5, een ander AMP,

de afgifte van hLF1-11

was adequaat

diffusie van het middel vanuit

dieper gelegen lagen

Hoofdstuk 8

94

een nuttige behandeloptie. Toegevoegd aan PMMA-cement werdadequate afgifte in laboratorium- en dierproeven gezien, zonder nadelenvoor de mechanische sterkte van het cement.9,10 Tenslotte bleek degentamicine afgifte te worden verhoogd na toevoegen van DHVAR-5 aanPMMA-cement met gentamicine.10 Dit kan gunstig zijn om in bacteriëlebiofilms door te dringen; deze kunnen zelfs op gentamicine houdendPMMA cement groeien.11,12 In modellen van biofilms lieten DHVAR-5 enandere AMPs opmerkelijke effectiviteit tegen een groot aantal bacteriënzien.13,14

Het detecteren van hLF1-11 in konijnebot vormde een analytischeuitdaging. De ontwikkeling van antilichamen voor een ELISA-test bleekte stuiten op beperkingen door de kleine afmeting van het peptide,(slechts elf aminozuren lang). Verder ontstond er veel storend signaaldoor de aanwezige eiwitten van het konijn. De analyse werd derhalvegedaan met een gecombineerde techniek van massaspectrometrie envloeistofchromatografie. Dit gaf een semi-quantitatieve analyse uit bot encement van de konijnen die met hLF1-11 in cement behandeld waren.Deze experimenten – waarin geen infectie was opgewekt – toonden nazeven dagen vroege tekenen van botingroei.

Bij de evaluatie van de dierexperimele resultaten houden wijnoodzakelijkerwijs rekening met inherente verschillen met de klinischepraktijk. De kleine afmetingen van het geïnjecteerde cement bevorderensnelle afgifte van hLF1-11. In patiënten is meer cement nodig, dit zou desnelheid van afgifte kunnen vertragen en de duur van afgifte verhogen.Verder zou het positieve effect van hLF1-11 op de werking van de afweer(immuno-modulatie) kunnen verschillen tussen konijnen en mensen.Anderzijds is dit effect ook beschreven inexperimenten met muizen, wat eengemeenschappelijk basisthema suggereert inde werking van de aangeboren afweer.15-17

De veiligheid werd bepaald met experimenten op bloed- en botcellen.Uitgebreide analyse van de toxiciteit vereist meerdere diermodellen enmeerdere diersoorten voordat menselijke studies kunnen worden gedaan.Er zijn resultaten van hLF1-11 in muizen, ratten en konijnen. Hierbij zijnnog geen bijwerkingen gerapporteerd. De menselijke oorsprong van hetpeptide (humaan lactoferrine) suggereert een bepaalde veiligheid, ditmoet uiteraard bevestigd worden door middel van standaard onderzoeknaar toxiciteit. De voortzetting van het project is derhalve van significantbelang.

Orthopedische toepassingenAls frontsoldaten van het afweersysteem beschermen AMPs kwetsbarelichaamsoppervlakken als huid en slijmvlies tegen bacterie-invasie.Bescherming van het oppervlak is evenzeer de kern van preventie en

experimenten op bloed-

en botcellen


95

behandeling van bot- en prothese-infectie. Winst van de ‘race voor hetoppervlak’ voorkomt kolonisatie door bacteriën. De bewezen effectiviteitvan AMPs in diermodellen van botinfectie bevestigt dit mechanisme:zowel preventie als behandeling met hLF1-11 verminderde deinfectie.8,9,18 Daarnaast kunnen AMPs op prothesen bevestigd worden metbehoud van hun activiteit, zoals is beschreven door Gabriel c.s.19

AMPs zouden klinisch gebruikt kunnen worden voor diagnostiek,preventie en behandeling van infectie. Radioactief gelabelde AMPsworden gepresenteerd als de nieuwe gouden standaard voor hetafbeelden van infecties.18,20,21 Deze betrouwbare methode van labelen metradioactief technetium maakt in dierstudies duidelijk onderscheid tusseninfectie en steriele ontsteking. Dit kan nieuwe inzichten opleveren naaraanleiding van de stelling dat veel gevallen van steriele loslating van deprothese toch berusten op niet ontdekte infecties.22 De klinischeuitwerking van deze dierexperimentele resultaten is waardevol voor devroege behandeling van infecties. De toepassing van deze techniek bijverdenking van prothese-infectie lijkt een waardevolle diagnostischeoptie bij het voorkomen van de noodzaak tot verwijderen van deprothese.

Preventie van infectie door lokale behandeling met hLF1-11 of DHVAR-5was een succesvolle benadering in de dierexperimenten. Beide peptidenzorgden preventief voor een significante afname van de bacterie-aantallen. Zelfs in het geval van bestaande infectie werden de resistenteMRSA-bacteriën gedecimeerd door lokale behandeling met hLF1-11.8

Het nut van AMPs beperkt zich niet tot botinfecties. Nibbering c.s.beschrijven dat behandeling met hLF1-11 een aantal resistente bacterie-en gist-stammen uitroeide in experimentele bilspierinfecties.15-17 Deeffectieve behandeling met verschillende AMPs van brandwonden nainfectie met Pseudomonas-bacteriën wordt gerapporteerd doorSteinstraesser c.s.23,24

De andere eigenschappen van AMPs,niet gerelateerd aan infectie, zijn nogniet volledig doorgrond. Hierondervallen groeistimulatie van bot en bloedvaten, remming van tumoren envorming van poriën in celwanden waardoor geneesmiddelen kunnenbinnenkomen.25-27 Daarnaast zou hLF1-11 invloed hebben op hetijzermetabolisme. Aanvullende studies naar deze mechanismen kunnenleiden tot betere kennis van de menselijke afweer, infectie enweefselherstel.

OntwikkelingenBotinfectie is een oud probleem, toch is er nog discussie over de optimalepreventie en behandeling: Het belang van antibioticum in cement metinherent risico op resistente bacteriën, de duur van intraveneuze en orale

groeistimulatie van bot en

bloedvaten, remming van tumoren

Hoofdstuk 8

96

antibioticum behandeling, de indicatie voor directe of uitgestelde revisiebij geïnfecteerde prothesen. Lange termijn studies van de Scandinavischeimplantaten registers biedenrichtlijnen die de dagelijkse praktijkkunnen sturen. De behandeling vande individuele patiënt met bot- of prothese-infectie dient zich te richtenop verschillende factoren. Zoals eerder vermeld: patiënt, prothese enpathogeen (bacterie) bepalen de uitkomst. Optimalisatie van deze driefactoren (voedingstoestand, speciaal oppervlak, antibiotica) zal hetbeloop maximaal beïnvloeden.

Sinds de eerste vermelding van AMPs in de jaren 80 zijn er al honderdenbeschreven. Een uitgebreide lijst is te vinden op het internet:http://www.bbcm.univ.trieste.it/~tossi/antimic.html. Als een eersteverdedigingslinie, blijken AMPs relatief vroeg in de evolutie te zijnontstaan. Inmiddels zijn AMPs gevonden in elk organisme waarin isgezocht (mens, dier, plant, zelfs in bacteriën). Aanvankelijk was hetonderzoek vooral gericht op de antibiotische werking. Nieuweeigenschappen – stimulatie van vaatgroei, remming van tumoren,virusremming en afweerregulatie – krijgen steeds meer aandacht. Hetstijgend aantal biologische funkties van AMPs doet discussie ontbrandenover de correcte naam. Daarom werden er steeds alternatieven bedachtdie de andere eigenschappen beter uitdrukten. In 2003 beëindigde AMP-guru Robert Lehrer de discussie:”Dit voorwerp in mijn hand kan alles:blikopenen, papier knippen, kurketrekken en vishaakjes verwijderen.Toch noem ik het een Zwitsers zakmes en begrijpt iedereen waar ik hetover heb. Wat doen we nou moeilijk? We houden het op AMP en nemenniet steeds een nieuwe naam wanneer er een funktie bijkomt.”

Het onderzoek naar AMPs ontwikkelt zich langzaam van ontdekkingnaar beschrijving en uiteindelijk toepassing.De eerste studies onderzochten duizendenbiologische monsters naar actieve moleculenBijvoorbeeld, bronnen zo verschillend alskikkerslijm, varkensbloed en mensentranen, allen bevatten AMPs. Dezeeiwiten kunnen synthetisch gemaakt worden van losse aminozuren enaangepast worden om de funktie te verbeteren. Verdere analyse vanAMPs betrof de interactie met celmembranen, celprocessen en signalen.28

De verschillende AMPs werken door middel van: porievorming in decelmembraan, sabotage van het energiesysteem (mitochondriën), ensignalen naar het menselijk immuunsysteem (chemotaxis en cytokineafgifte). Tenslotte beschrijft een groeiend aantal dierstudies de mogelijketoepassingen en toxiciteit van AMPs. De verschillende diermodellenbeschrijven: brandwondinfectie, longontsteking, bot- enspierinfectie.8,17,29 De modellen gebruiken zowel lokale als systemischetherapie met AMPs: huidcrème, injectie, longspoeling, en afgifte uit eendrager.

patiënt, prothese en pathogeen

kikkerslijm, varkensbloed

en mensentranen


97

ToekomstmuziekWij menen dat de solide prestaties van AMPs in dierexperimenten eenduidelijk vervolg in klinische studies vereisen. Anderzijds, moetenurgente maatregelen om resistentie te beperken niet worden uitgesteld:antibioticumgebruik sterk beperken, een verbod op risicovol gebruik inde veeteelt, en het verbeteren van de gezondheidszorg inontwikkelingslanden.30

Meervoudig resistente bacteriestammen zoals MRSA en VRSA hebbenongeduld veroorzaakt in het geneesmiddelenonderzoek.31,32 Toch moet deintroductie van AMP-gebaseerde middelen met de gebruikelijkewerkzaamheids-, toxiciteits- en veiligheidsstudies gepaard gaan. Decentrale rol van AMPs in de aangeboren afweer vereist bijzondereaandacht voor klinische introductie van middelen. Het completespectrum van biologische signalen beïnvloed door AMPs is nog eenaanzienlijke uitdaging voor peptidewetenschappers. Daarom is het nodigom te overwegen of gebruik van AMP gebaseerde antibiotica onze afweerzou kunnen beïnvloeden. Sterker nog, bacteriële resistentie tegen AMPsblijkt weliswaar moeilijk maar is niet onmogelijk.33 Derhalve zou onzeafweer in de problemen kunnen komen wanneer het AMP-resistente‘superbacteriën’ moet opruimen. Deze bacteriën zouden een aanzienlijkmedisch probleem kunnen vormen.

Zoals vaak in de geneeskunde liggen er oplossingen in het nabootsen vande natuur. Met een ‘schot hagel’ aanpak – de ontwikkeling van eengevarieerd pakket van AMP-antibiotica – wordt ontwikkeling vanresistentie tegengegaan. Meer inzicht in het mechanisme van regulatievan de afweer door AMPs zal hunspecifieke en selektieve toepassingkunnen verbeteren. En tenslotte, netzoals in het lichaam meerdereafweersystemen gebruikt worden, kan synergie tussen klassieke ennieuwe antibiotica effectief zijn bij resistente infecties. Nieuweantibiotica kunnen de bacteriën lek maken zodat de klassieke middelennaar binnen kunnen en zelfs in lage dosering hun werk weer kunnendoen. Deze ontwikkelingen vereisen nadere bestudering van interactiesen perfectionering van dosering. Afhankelijk van de situatie, zijnalternatieve toedieningswegen mogelijk (op de huid, longspoeling,injectie).9,17,23,29

ConclusiesInfectiepreventie is essentieel in de orthopedie. Preventieve antibioticakunnen resistentie veroorzaken. De waarde van AMPs ligt in hun lageneiging to inductie van resistente bacteriën.

nieuwe antibiotica kunnen de

bacteriën lek maken

Hoofdstuk 8

98

Calciumfosfaat cementen en korrels kunnen als antibioticum dragers meteen betrouwbare, beïnvloedbare afgifte worden ingezet. Dit combineertbiologische afbreekbaarheid met botgeleiding en bekende klinischeveiligheid. Aangezien calciumfosfaat wordt vervangen door bot isoperatieve verwijdering niet nodig.

hLF1-11 doodt een ruime hoeveelheid bacteriestammen, inclusiefresistente soorten en heeft aangetoonde effectiviteit in experimenten metbotinfectie en andere infecties. Verdere ontwikkeling van dit concept zoude ziektelast en de kosten van infecties kunnen helpen verminderen. Dediagnostische toepassing voor infectie detectie is een andere belangrijkeonderzoeksrichting.

Referenties





5. Faber C, Stallmann HP, Lyaruu DM, de Blieck JM, Bervoets TJ, Nieuw Amerongen AV,Wuisman PIJM. (2003). Release of antimicrobial peptide Dhvar-5 frompolymethylmethacrylate beads. J Antimicrob Chemother 51: 1359-1364.


7. Engesaeter LB, Lie SA, Espehaug B, Furnes O, Vollset SE, Havelin LI. (2003).Antibiotic prophylaxis in total hip arthroplasty: effects of antibiotic prophylaxissystemically and in bone cement on the revision rate of 22,170 primary hipreplacements followed 0-14 years in the Norwegian Arthroplasty Register. Acta OrthopScand 74: 644-651.



10. Faber C, Hoogendoorn RJ, Lyaruu DM, Stallmann HP, van Marle J, Nieuw AmerongenAV, Smit TH, Wuisman PIJM. (2005). The effect of the antimicrobial peptide, Dhvar-5,on gentamicin release from a polymethyl methacrylate bone cement. Biomaterials 26:5717-5726.




99


14. Rothstein DM, Helmerhorst EJ, Spacciapoli P, Oppenheim FG, Friden P. (2002).Histatin-derived peptides: potential agents to treat localised infections. Expert OpinEmerg Drugs 7: 47-59.





19. Gabriel M, Nazmi K, Veerman EC, Nieuw Amerongen AV, Zentner A. (2006).Preparation of LL-37-grafted titanium surfaces with bactericidal activity. BioconjugChem 17: 548-550.


21. Lupetti A, Nibbering PH, Welling MM, Pauwels EK. (2003). Radiopharmaceuticals:new antimicrobial agents. Trends Biotechnol 21: 70-73.

22. Nelson CL, McLaren AC, McLaren SG, Johnson JW, Smeltzer MS. (2005). Is asepticloosening truly aseptic? Clin Orthop Relat Res 25-30.






28. Mangoni ML. (2006). Temporins, anti-infective peptides with expanding properties.Cell Mol Life Sci 63: 1060-1069.



31. Coast J, Smith RD, Millar MR. (1996). Superbugs: should antimicrobial resistance beincluded as a cost in economic evaluation? Health Economics 5: 217-226.



Hoofdstuk 8

100

101

Acknowledgements

102

romoveren is teamsport, mijn dank gaat uitnaar de basis-spelers, invallers en supporters.

Allereerst bedank ik mijn promotoren professorWuisman en professor van Nieuw-Amerongen,

en mijn copromotor professor Veerman voor de goedgestructureerde, altijd stimulerende begeleiding van hetproject. Paul, het enthousiasme was absoluut besmettelijk.Arie, je positieve en snelle begeleiding heb ik gewaardeerd.Enno, de wondere wereld der peptiden-chemie werd metjouw hulp inzichtelijker.

Natuurlijk bedank ik iedereen van de afdeling OraleCelbiologie: Cor (voor zijn betrokkenheid en gezelligheid),Dirk-Jan (voor zijn onvermoeibare bepalingen opkonijnenbotjes), Don (voor een kritische blik), Els (de opzetvan het TS projekt), Jenneke (de vele flitsende discussies),Jolanda (voor alle ELISA’s en het onthouden van deverjaardagen), Marion (als brenger van orde in de TS-registratie), Ton (voor de rustige, goede begeleiding bij hetschrijven), Vincent (voor kritisch advies) en Wan (voorprachtige histologie).

Van de afdeling waar de AMP’s worden klaargestoomd,Orale Biochemie, wil ik graag bedanken: Jan (ondersteuningvan mijn basale peptiden-kennis), en Kamran (heerser vande HPLC, voor zijn geduldige biochemisch advies en depeptide-productie).

Van de afdeling orthopaedie wil ik nog graag noemen: Elise,Margreet (maakten altijd tijd vrij voor besprekingen met deprofessor), Marco en Theo (voor hun kritische kijk op hetproject en ondersteuning binnen STEGA). Ruud, bedanktvoor zorgvuldige beheer van de TS-fondsen.

De grote inzet van Arie, Esther, Ger, Klaas, en de anderen bijde dierexperimenten in het KDL heb ik zeer gewaardeerd.De altijd parate expertise, inzet en gezelligheid maken hetKDL een parel van de VU campus.Ook van de dieren ethische commissie; Jan, Klaas en Paul(advies en overleg over de opzet van dierexperimenten) hebik veel geleerd. Verder wil ik nog bedanken: Ferry(microbiologisch advies), Paul (submitten ATTC stam) Koraen Hein (voor hulp en advies bij de scans in het AMC), Micken Peter (voor goede samenwerking met de peptidenonderzoekers in het LUMC).

103

De onderzoeks-studenten, van wie ik minstens zo veelgeleerd heb als zij van mij: Anje (voor het zorgvuldig scorenvan de histologie), Carlijn (inzet bij de opzet van het MRSAproject), Crispijn (assistentie bij de eerste konijn-operaties),Diana (de opzet van de haemolyse assay), Eveline (de eersterelease experimenten), Thijs (voor het scannen van dekonijnen en de handgreep van Plokker ter lediging van deblaas).

Verder gaat mijn dank uit naar de Herensteeg; Guido(verzorging van de opmaak van dit boekje), Rutger enRonald (chirurgische assistentie op OK).

AM-Pharma: Carlo, Ewald, Martin, Paul en Robert (voor hetstimulerende effect dat uitgaat van samenwerken met eenjong biotech bedrijf).

Meerdere experts hebben het manuscript op fouten enleesbaarheid beoordeeld. Dank hiervoor aan mijn ouders,Sophie en Sue Weijhenke.

Uiteraard was het niet ’All work and no play’. Het dagje uitmet de afdeling, de borrels, het VU-beach café, decongressen, squashen, een bowling of skate middag metSTEGA. Misschien nog belangrijker dan onderzoek, eengezellige, stimulerende tijd op OCB: Agnes, Astrid, Aviral,Behrouz, Clara, Djien, Emiel, Ilara, Gerald, James, Jack,Marco, Margriet, Marlene, Matthijs, Mel, Roel, Saskia.

Als laatste bedank ik mijn kamergenoot Chris, samen succesen congres, Londen en Lunteren meegemaakt; het was eenwereldtijd!

Antimicrobial peptides Experimental prevention of osteomyelitis · Osteomyelitis Osteomyelitis is a...

Documents

Transcript of Antimicrobial peptides Experimental prevention of osteomyelitis · Osteomyelitis Osteomyelitis is a...