Research Collection

Doctoral Thesis

Towards receptor-specific targeting of antigen presenting cellswith functionalized stealth microparticles

Author(s): Wattendorf, Uta

Publication Date: 2008

Permanent Link: https://doi.org/10.3929/ethz-a-005582085

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.

ETH Library

DISS. ETH NO. 17540

TOWARDS RECEPTOR-SPECIFIC TARGETING OF ANTIGEN PRESENTING CELLS WITH

FUNCTIONALIZED STEALTH MICROPARTICLES

A dissertation submitted to

ETH ZURICH

for the degree of

Doctor of Sciences

presented by

UTA WATTENDORF

Diplom - Chemikerin, University of Heidelberg

born September 1st, 1977

citizen of Germany

accepted on the recommendation of

Prof. Dr. Hans P. Merkle, examiner Prof. Dr. Marcus Textor, co-examiner Prof. Dr. Janos Vörös, co-examiner

2008

"The phagocytes won't eat the microbes

unless the microbes are nicely buttered

for them. (...) That butter I call opsonin."

George Bernard Shaw, "The Doctor's Dilemma" (1906)

TABLE OF CONTENTS I

ABBREVIATIONS AND NOTATIONS III

BACKGROUND AND PURPOSE 1

SUMMARY 7

ZUSAMMENFASSUNG 11

CHAPTER I 17PEGylation as a Tool for the Biomedical Engineering of Surface Modified Microparticles

CHAPTER II 53Phagocytosis of PLL-g-PEG Coated Microspheres by Antigen Presenting Cells: Impact of Grafting Ratio and PEG Chain Length on Cellular Recognition

CHAPTER III 89Stable Stealth Function for Hollow Polyelectrolyte Microcapsules through a Polyethylene Glycol Grafted Polyelectrolyte Adlayer

CHAPTER IV 123Impact of Mannose-Based Molecular Patterns on Human Antigen Presenting Cells: A Study with Stealth Microspheres

GENERAL DISCUSSION AND OUTLOOK 169

APPENDIX 177

CURRICULUM VITAE 197

ACKNOWLEDGEMENTS 201

II

ABBREVIATIONS AND NOTATIONS III

ABBREVIATIONS AND NOTATIONS AF488 Alexa Fluor® 488

AF633 Alexa Fluor® 633

APC antigen presenting cells

a PLL:b PLL-PEG coatings with a% PLL and b% PLL-PEG

a PLL:b PLL-PEG/Mannan a secondary coating with mannan onto a PLL:b PLL-PEG

BSA bovine serum albumin

CLR C-type lectin receptor

ConA Concanavalin-A

DC dendritic cells

DC-SIGN dendritic cell-specific ICAM-3 grabbing non-integrin

EDC 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide

grafting ratio, g total number of lysine monomers divided by the number of PEG side chains

HEPES buffer 10 mM 4-(2-hydroxyethyl)piperazine-1-ethane-sulfonic acid, pH 7.4

iDC immature DC

IgG immunoglobulin G

mDC mature DC

MF melamine formaldehyde

MΦ macrophages

MHC (I or II) major histocompatibility complex (I or II)

mPEG monomethoxy-PEG (= the monomethyl-ether of PEG)

MPS mononuclear phagocytic system

IV ABBREVIATIONS AND NOTATIONS

MR mannose receptor

NHS N-hydroxysulfosuccinimide

PAH polyallylamine hydrochloride

PBS buffer 10 mM phosphate buffered saline, pH 7.4

PEG poly(ethylene glycol)

PEO poly(ethylene oxide)

PGA poly-L-glutamic acid

PGA(x)-g[y]-PEG(z) copolymer of PEG (MW = z kDa) grafted to PGA (MW = x kDa) at a grafting ratio y

PI propidium iodide

PLL poly-L-lysine

PLL(x)-g[y]-PEG(z) copolymer of PEG (MW = z kDa) grafted to PLL (MW = x kDa) at a grafting ratio y

PLL-PEG PLL(20)-g-PEG(2), a copolymer of PEG (MW = 2 kDa) grafted to PLL (MW = 20 kDa) at a grafting ratio g = 3.5 - 4.0

PLL-PEG-ligand(x%) PLL-g-PEG(2)/PEG(3.4)-ligand(x%), a PLL-PEG copolymer with x% of the side chains exchanged by ligand-coupled PEG chains of 3.4 kDa

PEG-RDG PLL-g-PEG(2)/PEG(3.4)-RDG(9%)

PLL-PEG-RGD PLL-g-PEG(2)/PEG(3.4)-RGD(8%)

PLL-PEG-mannoside collective abbreviation for PLL-g-PEG(2)/ /PEG(3.4)-mannoside(x%), with mannoside as general term for different ligands (mono- and tri-mannose) and linkers (propyl and ethylene glycol)

ABBREVIATIONS AND NOTATIONS V

PLL-PEG-λMan(x%) and PLL-PEG-λTri-Man(x%)

PLL-PEG-ligand(x%) with mono- or tri-mannose as a ligand and λ indicating the respective linker (propyl or ethylene glycol)

Pluronic BASF trade name for poloxamers

Poloxamer PEOa-PPOb-PEOa

PS polystyrene

PSS polystyrene sulfonate

RPMI medium RPMI 1640 (Roswell Park Memorial Institute) with L-glutamine

SSC side scattering (measured by flow cytometry)

TLR Toll-like receptor

TRITC tetramethylrhodamine isothiocyanante

VI

BACKGROUND AND PURPOSE 1

BACKGROUND AND PURPOSE Microparticles, such as microspheres or microcapsules, are of

considerable interest in drug delivery, vaccination and as diagnostic agents. Among the advantages of microparticulate formulations in drug delivery are the stabilization of the drug and its controlled and sustained release, as well as the potential for localized delivery of the drug to a particular body compartment [1, 2]. In general, it is possible to distinguish two different approaches for the targeting of particulate carriers: (i) physical targeting and (ii) targeting through decoration of the microparticles with suitable ligands [3]. An example for the first approach would be the accumulation of magnetic microparticles at specific sites in the body by means of an external magnetic field. The targeting to specific cells, or cellular receptors, respectively, is feasible through surface modification with suitable ligands.

The latter concept has gained much attention for vaccine delivery and immunomodulation [4]. It is generally recognized that an engagement of surface modified particulate carriers with specific receptors on professional antigen presenting cells (APCs) could potentially affect the immune response that will be elicited. Microparticulates of about 1 to 10 µm in diameter are logical targeting systems for vaccines and immunomodulators, as they undergo rapid phagocytosis by APCs, an exclusive feature of phagocytes such as dendritic cells (DC) and macrophages (MΦ). Among the many APC receptors, such as Toll-like receptors (TLR), C-type lectins (CLR), scavenger receptors (SR) or immunoglobulin receptors (FcR), especially TLRs and CLRs have been extensively studied for their capacity to recognize pathogen associated molecular patterns (PAMPs) [5, 6]. The targeting of CLRs on APCs was identified as a promising strategy for immunomodulation, as documented by several studies showing that immune responses were enhanced or modified by coupling mannose type ligands, such as mannosides or mannan, to proteins or particulates [7-12].

Upon entry into the organism, however, microparticulate carriers are instantly recognized as foreign and are efficiently cleared via phagocytosis by the mononuclear phagocytic system (MPS), which comprises (in addition to other specialized organs and tissues) migrating phagocytes such as dendritic

2 BACKGROUND AND PURPOSE

cells and macrophages [13]. At first glance, for the targeting of APCs this seems quite advantageous. However, for the targeting of specific cell surface receptors, it is important to prevent recognition by competing receptors. Otherwise, it would be difficult to specify whether an observed effect on phagocytosis or immunomodulation was exclusively due to a specific ligand-receptor interaction, or resulted from a combination of more than a single recognition signal. Since it is particularly the adsorption of opsonic proteins onto the particles´ surface which promotes recognition by various phagocytic receptors [14-16], receptor specific targeting requires surface modifications that are repellent to the adsorption of proteins. It has been hypothesized that this could be accomplished by microparticulate carriers that bear suitable ligands coupled to an MPS resistant corona, such as poly(ethylene glycol) (PEG) [3].

This PhD thesis aims at the development of such a microparticulate platform that allows the targeting of CLRs on APCs. We chose the PEG grafted polymer poly(L-lysine)-graft-PEG (PLL-g-PEG)a to immobilize a protein repellent PEG corona onto microparticles. Due to its positive charge at physiological pH, the PLL backbone adsorbs strongly onto negatively charged substrates through electrostatic interaction, forming a less than 10 nm thick adlayer of PLL-g-PEG stretching its PEG chains out perpendicularly to the surface [17, 18]. Appropriate choice of copolymer architecture, i.e. molecular weight and degree of PEG substitution, renders such coated surfaces protein-repellent even in the presence of full human serum [17, 18]. In addition, it was shown that phagocytosis of microspheres could be abolished with PLL-g-PEG coatings [19].

For the aim of targeting cell surface receptors, the PEG side chains of PLL-g-PEG can be terminally functionalized with bioligands, while retaining the repellent character of its adlayer to unspecific adsorption of proteins [20, 21]. The approach chosen in this dissertation involves (i) the immobilization of mono-and tri-mannoside functionalized PLL-g-PEG onto negatively charged microparticulate carriers, and (ii) the deposition of the polysaccharide mannan

a For simplicity, in all of Chapter IV PLL-g-PEG was denoted by PLL-PEG, and PLL-g-PEG/PEG-ligand as PLL-PEG-ligand.

BACKGROUND AND PURPOSE 3

onto mixed PLL and PLL-g-PEG coatings. We hypothesized that the presentation of mannoside ligands linked to the surface of microspheres at high densities via flexible PEG chains could be a strategy to mimic high molecular weight mannose based PAMPs with synthetic small molecular ligands. Mannan coatings onto mixed PLL and PLL-g-PEG adlayers were to be tested as controls.

Following a review on the properties of PEG, its immobilization on the surface of microparticles, and the biomedical applications of surface modified microparticles (Chapter I), the impact of PLL-g-PEG architecture on the protein repellence and resistance to phagocytosis of surface modified polystyrene (PS) model microspheres is investigated using a small library of such copolymers (Chapter II). As a further step towards possible applications as drug delivery system, it is shown that PS microspheres, here used as a model, can be replaced by other microparticulates such as polyelectrolyte microcapsules (Chapter III). Introduced in the late 1990s through layer-by-layer assembly of oppositely charged polyelectrolytes onto a solid microparticulate template that is subsequently dissolved, hollow polyelectrolyte microcapsules are currently investigated for their prospects towards biomedical application [22]. So far, a broad range of chemically diverse compounds has been successfully incorporated into hollow microcapsules, among them enzymes, nucleic acids, and magnetic nanoparticles [23-26]. Along with transferring the PEGylation technology to other microparticulates, this chapter also confirms the stability of the coatings for at least three weeks. Chapter IV finally deals with receptor specific targeting of APC. Different mannoside ligands (chemically coupled mono- and tri-mannose; or adsorbed mannan from S. cerevisiae) were immobilized onto stealth model microspheres via adsorption of PLL-g-PEG. Aspects that are covered in this chapter include both the formulation of such microspheres and their specific recognition by APC, as well as potential effects of mannoside substitutions on DC maturation.

4 BACKGROUND AND PURPOSE

REFERENCES 1. Langer, R., New methods of drug delivery. Science, 1990. 249(4976): p.

1527-1533.

2. Putney, S.D., and Burke, P.A., Improving protein therapeutics with sustained-release formulations. Nat. Biotechnol., 1998. 16(2): p. 153-157.

3. Arshady, R., and Monshipouri, M., Targeted delivery of microparticulate carriers. Microspheres, Microcapsules & Liposomes, 1999. 2(Medical & Biotechnology Applications): p. 403-432.

4. Gogolák, P., Réthi, B., Hajas, G., and Rajnavölgyi, É. Targeting dendritic cells for priming cellular immune responses. J. Mol. Recognit., 2003. 16(5): p. 299-317.

5. Proudfoot, O., Apostolopoulos, V., and Pietersz, G.A., Receptor-mediated delivery of antigens to dendritic cells: anticancer applications. Mol Pharm, 2007. 4(1): p. 58-72.

6. Cambi, A., and Figdor, C.G., Dual function of C-type lectin-like receptors in the immune system. Curr. Opin. Cell Biol., 2003. 15(5): p. 539-546.

7. Toda, S., Ishii, N., Okada, E., Kusakabe, K.I., Arai, H., Hamajima, K., Gorai, I., Nishioka, K., and Okuda, K., HIV-1-specific cell-mediated immune responses induced by DNA vaccination were enhanced by mannan-coated liposomes and inhibited by anti-interferon-gamma antibody. Immunology, 1997. 92(1): p. 111-117.

8. Cui, Z., Han, S.-J., and Huang, L., Coating of Mannan on LPD Particles Containing HPV E7 Peptide Significantly Enhances Immunity Against HPV-Positive Tumor. Pharm. Res., 2004. 21(6): p. 1018-1025.

9. Jain, S., and Vyas, S.P., Mannosylated niosomes as adjuvant-carrier system for oral mucosal immunization. J Liposome Res, 2006. 16(4): p. 331-345.

10. Apostolopoulos, V., Pietersz, G.A., Tsibanis, A., Tsikkinis, A., Drakaki, H., Loveland, B.E., Piddlesden, S.J., Plebanski, M., Pouniotis, D.S., Alexis, M.N., McKenzie, I.F., and Vassilaros, S., Pilot phase III immunotherapy study in early-stage breast cancer patients using oxidized

BACKGROUND AND PURPOSE 5

mannan-MUC1 [ISRCTN71711835]. Breast Cancer Res, 2006. 8(3): p. R27.

11. White, K., Rades, T., Kearns, P., Toth, I., and Hook, S., Immunogenicity of liposomes containing lipid core peptides and the adjuvant Quil A. Pharm. Res., 2006. 23(7): p. 1473-1481.

12. Hattori, Y., Kawakami, S., Lu, Y., Nakamura, K., Yamashita, F., and Hashida, M., Enhanced DNA vaccine potency by mannosylated lipoplex after intraperitoneal administration. J Gene Med, 2006. 8(7): p. 824-834.

13. Juliano, R.L., Factors affecting the clearance kinetics and tissue distribution of liposomes, microspheres and emulsions. Adv. Drug Deliv. Rev., 1988. 2(1): p. 31-54.

14. Thiele, L., Diederichs, J.E., Reszka, R., Merkle, H.P., and Walter, E., Competitive adsorption of serum proteins at microparticles affects phagocytosis by dendritic cells. Biomaterials, 2003. 24(8): p. 1409-1418.

15. Absolom, D.R., Opsonins and dysopsonins: an overview. Methods Enzymol., 1986. 132(Immunochem. Tech., Pt. J): p. 281-318.

16. Patel, H.M., Serum opsonins and liposomes: their interaction and opsonophagocytosis. Crit. Rev. Ther. Drug Carrier Syst., 1992. 9(1): p. 39-90.

17. Huang, N.-P., Michel, R., Vörös, J., Textor, M., Hofer, R., Rossi, A., Elbert, D.L., Hubbell, J.A., and Spencer, N.D., Poly(L-lysine)-g-poly(ethylene glycol) Layers on Metal Oxide Surfaces: Surface-Analytical Characterization and Resistance to Serum and Fibrinogen Adsorption. Langmuir, 2001. 17(2): p. 489-498.

18. Pasche, S., De Paul, S.M., Vörös, J., Spencer, N.D., and Textor, M., Poly(L-lysine)-graft-poly(ethylene glycol) assembled monolayers on Niobium Oxide surfaces: a quantitative study of the influence of polymer interfacial architecture on resistance to protein adsorption by ToF-SIMS and in situ OWLSu OWLS. Langmuir, 2003. 19(22): p. 9216-9225.

19. Faraasen, S., Vörös, J., Csúcs, G., Textor, M., Merkle, H.P., and Walter, E., Ligand-Specific Targeting of Microspheres to Phagocytes by Surface Modification with Poly(L-Lysine)-Grafted Poly(Ethylene Glycol) Conjugate. Pharm. Res., 2003. 20(2): p. 237-246.

6 BACKGROUND AND PURPOSE

20. VandeVondele, S., Vörös, J., and Hubbell, J.A., RGD-grafted poly-L-lysine-graft-(polyethylene glycol) copolymers block non-specific protein adsorption while promoting cell adhesion. Biotechnol. Bioeng., 2003. 82(7): p. 784-790.

21. Müller, M., Vörös, J., Csúcs, G., Walter, E., Danuser, G., Merkle, H.P., Spencer, N.D., and Textor, M., Surface modification of PLGA microspheres. J. Biomed. Mater. Res., Part A, 2003. 66A(1): p. 55-61.

22. Sukhorukov, G.B., Rogach, A.L., Zebli, B., Liedl, T., Skirtach, A.G., Köhler, K., Antipov, A.A., Gaponik, N., Susha, A.S., Winterhalter, M., and Parak, W.J., Nanoengineered polymer capsules: tools for detection, controlled delivery, and site-specific manipulation. Small, 2005. 1(2): p. 194-200.

23. Tiourina, O.P., Antipov, A.A., Sukhorukov, G.B., Larionova, N.L., Lvov, Y., and Möhwald, H., Entrapment of alpha-chymotrypsin into hollow polyelectrolyte microcapsules. Macromolecular Bioscience, 2001. 1(5): p. 209-214.

24. Kreft, O., Georgieva, R., Bäumler, H., Steup, M., Müller-Röber, B., Sukhorukov, G.B., and Möhwald, H., Red blood cell templated polyelectrolyte capsules: A novel vehicle for the stable encapsulation of DNA and proteins. Macromol Rapid Comm, 2006. 27(6): p. 435-440.

25. Shchukin, D.G., Patel, A.A., Sukhorukov, G.B., and Lvov, Y.M., Nanoassembly of Biodegradable Microcapsules for DNA Encasing. J. Am. Chem. Soc., 2004. 126(11): p. 3374-3375.

26. Zebli, B., Susha, A.S., Sukhorukov, G.B., Rogach, A.L., and Parak, W.J., Magnetic targeting and cellular uptake of polymer microcapsules simultaneously functionalized with magnetic and luminescent nanocrystals. Langmuir, 2005. 21(10): p. 4262-4265.

SUMMARY 7

SUMMARY Microparticles, such as microspheres or microcapsules, are of

considerable interest in drug delivery, vaccination and as diagnostic agents. A special focus in the field of vaccination are targeted microparticles. Specifically, the targeting of C-type lectins (CLRs) on antigen presenting cells (APCs) was identified as a promising strategy for immunomodulation. Therefore, the objective of this dissertation was the development of a microparticulate platform that allowed the receptor specific targeting of CLRs on APCs. To achieve this aim, two requirements had to be met: (i) the microparticles had to be shielded from opsonization, phagocytic clearance, and engagement of cell surface receptors other than those to be targeted; and (ii) specific ligands had to be immobilized on the surface of the microparticles, leading to specific recognition by the targeted CLRs, and without compromising the particles´ stealth character.

It has been hypothesized that the specific targeting of cell receptors could be accomplished by microparticulate carriers that bear suitable ligands coupled to the free ends of a of poly(ethylene glycol) (PEG) corona. As an introduction to this PhD thesis, Chapter I reviews the unique properties of PEG, which make it the polymer of choice to render surfaces resistant to both adsorption of proteins and cellular recognition. Furthermore, different strategies for the preparation of PEGylated microparticles are presented. Both the incorporation of PEG during microparticle formation and the immobilization of PEG onto preformed microparticles are discussed. Finally, miscellaneous applications of PEGylated microparticles are reviewed, focusing on applications in the field of drug delivery. In particular, a section of this chapter is dedicated to the targeting of PEGylated stealth microparticles, such as through the surface immobilization of suitable ligands.

For the immobilization of a PEG corona on microparticles we chose PEG grafted poly(L-lysine) (PLL-g-PEG) copolymers. Due to the positive charge of the PLL backbone at physiological pH, it spontaneously adsorbs from aqueous solution onto negatively charged substrates through electrostatic interactions. The impact of PLL-g-PEG architecture on the recognition of coated microspheres by APC was explored in Chapter II. A small library of PLL-g-PEG copolymers was coated onto model microspheres (5 μm carboxylated

8 SUMMARY

polystyrene) to investigate resistance to phagocytosis depending on both grafting ratio (g = 2 to 20) and molecular weight (1 to 5 kDa) of the PEG chains. Phagocytosis by human monocyte derived dendritic cells (DC) and macrophages (MΦ) was quantified by phase contrast microscopy and by analysis of the cells’ side scattering in a flow cytometer. Supporting experiments included the determination of the particles´ ζ-potentials, their repellence to protein adsorption, and cell adhesion experiments on PLL-g-PEG coated glass slides. In addition, the efficient internalization of control microspheres, in contrast to adsorption to the outer cell membrane, was confirmed by confocal microscopy. Generally, increasing grafting ratios impaired the protein repellence of coated microspheres, leading to higher phagocytosis rates. For DC, long PEG chains of 5 kDa decreased the phagocytosis of coated microspheres even in the case of considerable IgG adsorption. In addition, preferential adsorption of dysopsonins is discussed as another potential factor for decreased phagocytosis rates. Remarkably, DC and MΦ were found to adhere to relatively protein-repellent PLL-g-PEG adlayers, whereas phagocytosis of microspheres coated with the same copolymers was inefficient. Overall, the copolymer with 2 kDa PEG chains, a 20 kDa PLL backbone and a grafting ratio of 3.5, PLL(20)-[3.5]-PEG(2), was identified as the copolymer of optimal architecture to ensure resistance to both phagocytosis and cell adhesion. Therefore, this polymer was subsequently used throughout this PhD thesis. Finally, this chapter also shows the feasibility of preparing mixed coatings from different PLL-g-PEG type copolymers. This concept is exploited in the last chapter to control ligand densities by co-adsorption of ligand-modified and ligand-free PLL-g-PEG.

As a further step towards possible applications as drug delivery systems, Chapter III shows that the model microspheres can be replaced by other particulates such as polyelectrolyte microcapsules. Layer-by-layer assembled hollow microcapsules that were composed of alternating layers of polystyrene sulfonate (PSS) and polyallylamine hydrochloride (PAH) were coated with adlayers of PLL-g-PEG. In addition, a PEG-grafted copolymer with an anionic backbone, poly(L-glutamic acid)-graft-PEG (PGA-g-PEG), was investigated. PGA-g-PEG coatings had no significant effect on phagocytosis, which may be explained by insufficient PEG density of the adlayer. Contrary, PLL-g-PEG

SUMMARY 9

effectively blocked phagocytosis of coated microcapsules. Moreover, PAH/PSS microcapsules did not impair phagocyte viability. The results demonstrate that layer-by-layer assembled polyelectrolyte microcapsules coated with PLL-g-PEG represent a promising platform for a drug delivery system that escapes fast clearance by the MPS. In addition, this chapter also confirms the stability of PLL-g-PEG coatings upon storage for at least three weeks, even in the presence of human serum.

Chapter IV finally deals with receptor specific targeting of CLRs on APCs. Different mannoside ligands (adsorbed mannan from S. cerevisiae; or chemically coupled mono- and tri-mannose) were immobilized at controlled concentrations onto stealth microspheres via adsorption of PLL-g-PEG. We hypothesized that the presentation of mannoside ligands linked to the surface of microspheres at high densities via flexible PEG chains could be a strategy to mimic high molecular weight mannose based pathogen associated molecular patterns (PAMPs) with synthetic small molecular ligands. Aspects that are covered in this chapter include both the formulation of such microspheres and their specific recognition by APC, as well as potential effects of mannoside substitutions on DC maturation. In addition, combinations of mannoside ligands with an integrin targeting RGD peptide ligand were investigated. Due to their PEG and mannan coronas, the coatings were repellent to unspecific protein adsorption, thus inhibiting opsonization and allowing specific ligand-receptor interactions. Mannose density was a major factor for the phagocytosis of mannosylated microspheres, though with limited efficiency. This strengthened the recent hypothesis by other authors that the mannose receptor (MR) can only act as a phagocytic receptor in collaboration with yet unidentified partner receptors. Finally, analysis of DC maturation revealed that surface assembled mannan on the microspheres could not stimulate DC maturation. Thus, phagocytosis upon recognition by CLRs alone cannot trigger DC activation towards a T helper response.

Overall, the microparticulate platform established in this work represents a promising tool for a systematic investigation of specific ligand-receptor interactions on phagocytes, including the screening for potential ligands and ligand combinations, and the analysis of immunomodulatory effects thereof. We

10 SUMMARY

hope that this platform will be useful to deepen our knowledge about the impact of the first encounter of APCs with micron-sized pathogens or microparticles on the downstream immune cascade, which may ultimately lead to improved vaccine formulations.

ZUSAMMENFASSUNG 11

ZUSAMMENFASSUNG Mikropartikel, seien es solide Partikel oder Mikrokapseln, finden in der

pharmazeutischen Forschung vielseitige Anwendungen. Ein besonderer Fokus ist das sogenannte Targeting von Mikropartikeln, worunter generell ein gezieltes Ansteuern bestimmter Zellen oder Gewebe verstanden wird, im engeren Sinne aber auch das Ansprechen spezieller Zell-Rezeptoren. Der letztgenannte Ansatz wird seit einiger Zeit auch zur Verbesserung der Immunantwort bei Impfungen verfolgt. Insbesondere das Targeting einer speziellen Art von Lektin-Rezeptoren auf der Oberfläche von antigenpräsentierenden Zellen (APC), den Lektin-Rezeptoren vom C-Typ (CLR), wurde als vielversprechende Strategie für die Modulation der Immunantwort erkannt. Ziel der vorliegenden Dissertation war das spezifische Targeting solcher CLR. Um dies zu erreichen, waren zwei Bedingungen zu erfüllen: (i) die Mikropartikel mussten gegen die Adsorption von Proteinen, die sog. Opsonisierung, passiviert werden, um Wechselwirkungen der adsorbierten Proteine mit anderen Zell-Rezeptoren auszuschliessen; und (ii) es mussten spezifische Liganden für die CLR auf der Oberfläche der Mikropartikel immobilisiert werden, ohne dabei deren Passivierung gegen Proteinadsorption zu beeinträchtigen.

Einer Hypothese aus der Literatur zufolge kann spezifisches Targeting von Zell-Rezeptoren erreicht werden, indem Polyethylenglykol-(PEG)-Ketten an die Oberfläche von Mikropartikeln gebunden werden, an deren Ende Liganden für die Rezeptoren chemisch gebunden wurden. Eine Einführung in die Thematik der PEGylierung von Mikropartikeln ist Gegenstand von Kapitel I. Zunächst wird auf die einzigartigen Eigenschaften von PEG eingegangen, aufgrund derer es wie kaum ein anderes Polymer zur Passivierung von Oberflächen gegen Proteinadsorption und zelluläre Wechselwirkungen geeignet ist. Im Folgenden werden verschiedene Strategien zur Herstellung von Mikropartikel mit PEGylierten Oberflächen vorgestellt, wobei PEG entweder bereits während der Herstellung der Mikropartikel eingesetzt wird, oder erst in einem weiteren Schritt auf der Oberfläche bereits bestehender Partikel immobilisiert wird. Ein Schwerpunkt des Kapitels liegt schliesslich auch auf den verschiedenen Anwendungsmöglichkeiten PEGylierter Mikropartikel im bio-medizinischen Bereich, wobei insbesondere auf das Targeting PEGylierter

12 ZUSAMMENFASSUNG

Mikropartikel mit Hilfe geeigneter, auf der Oberfläche immobilisierter Liganden eingegangen wird.

Zur Adsorption einer PEG-Schicht auf den Mikropartikeln wird in dieser Dissertation grundsätzlich ein Pfropfpolymer (engl.: graft copolymer) vom Typ Poly(L-Lysin)-graft-PEG (PLL-g-PEG) verwendet. Da das Rückgrat dieses Polymers (PLL) bei physiologischem pH positiv geladen ist, adsorbiert PLL-g-PEG aus wässrigen Lösungen spontan auf negativ geladene Substrate unter Ausbildung elektrostatischer Wechselwirkungen. Kapitel II untersucht den Einfluss der molekularen Struktur des Pfropfpolymers auf die Phagozytose der Mikropartikel durch APC, wobei das Molekulargewicht der PEG-Ketten von 1 bis 5 kDa und das Pfropfverhältnis (engl.: grafting ratio) von 2 bis 20 variiert wurde. Letzteres ist definiert als die Anzahl PLL-Monomere geteilt durch die Anzahl gepfropfter PEG-Ketten. Das Propfverhältnis ist damit eine Kennzahl für die PEG-Dichte des jeweiligen Pfropfpolymers. Die Phagozytose von derart überzogenen Mikropartikeln wurde mit Hilfe von dendritischen Zellen und Makrophagen untersucht, welche aus humanen Monozyten gewonnen wurden. Die Quantifizierung erfolgte mittels Phasenkontrastmikroskopie und Durchflusszytometrie, wobei als Modell Mikropartikel aus carboxyliertem Polystyrol mit einem Durchmesser von 5 μm Verwendung fanden. Unterstützende Experimente umfassten die Bestimmung der ζ-Potentiale beschichteter Mikropartikel und des Ausmasses der Proteinadsorption, sowie Zelladhäsionsexperimente auf beschichtetem Glas. Zusätzlich wurde auch die Phagozytose der (Kontroll)-Mikropartikel, im Gegensatz zur Adhäsion an die äusseren Zellmembran, mittels konfokaler Mikroskopie nachgewiesen. Grundsätzlich führten grössere Pfropfverhältnisse, also geringere PEG-Dichten, zu einer verringerten Resistenz gegen die Adsorption des Modell-Proteins Immunoglobulin G (IgG), was mit einer erhöhten Phagozytose einherging. Im Fall von dendritischen Zellen und langen PEG-Ketten (5 kDa) nahm die Phagozytose beschichteter Mikropartikel darüber hinaus selbst dann ab, wenn beträchtliche Mengen IgG adsorbiert wurden. In diesem Zusammenhang wird auch die möglicherweise bevorzugte Adsorption von Dysopsoninen diskutiert, welche ein weiterer wichtiger Faktor für die verminderte Phagozytose durch dendritische Zellen oder Makrophagen sein könnte. Interessanterweise konnten

ZUSAMMENFASSUNG 13

beide Zellarten auch auf einigermassen protein-resistenten PLL-g-PEG Schichten adhärieren, während keine effiziente Phagozytose von Mikropartikeln möglich war, wenn diese mit den gleichen Pfropfpolymeren modifiziert wurden. Insgesamt legt dieses Kapitel das Fundament für alle weiteren Arbeiten, indem das Pfropfpolymer mit einem Pfropfverhältniss von 3.5 bei einem PEG-Molekulargewicht von 2 kDa als optimal identifiziert wurde, um sowohl Passivierung gegen Proteinadsorption als auch verminderte Zelladhäsion und Phagozytose zu verleihen. Darüber hinaus wird gezeigt, dass auch Mischungen verschiedener PLL-g-PEG-Polymere auf Mikropartikel aufgebracht werden können. Dieser Aspekt wird im letzten Kapitel Anwendung finden, um durch Adsorption einer Mischung von Ligand-modifiziertem und unmodifizierem PLL-g-PEG eine kontrollierte Variation der Ligandenkonzentration zu erreichen.

Als weiterer Schritt in Richtung möglicher Anwendungen wird in Kapitel III gezeigt, dass die als Modelle verwendeten Mikropartikel auch durch andere Partikel, wie z.B. durch Polyelektrolyt-Mikrokapseln, ersetzt werden können. Diese Mikrokapseln wurden durch schichtweise abwechselnde Adsorption der Polymere Polystyrolsulfonat (PSS) und Polyallylamin Hydrochlorid (PAH) auf mikropartikuläre Template und nachfolgende Auflösung der Template hergestellt, und in einem zweiten Schritt mit PLL-g-PEG beschichtet. Alternativ wurde ein ähnliches Pfropfpolymer mit negativ geladenem Rückgrat, Poly(L-Glutaminsäure)-graft-PEG (PGA-g-PEG), aufgebracht. PGA-g-PEG hatte keinen Einfluss auf die Phagozytose, was wahrscheinlich auf eine zu geringe PEG-Dichte zurückzuführen war. Auf der anderen Seite konnte die Phagozytose der Mikrokapseln durch das Aufbringen einer PLL-g-PEG-Schicht erfolgreich gehemmt werden, ohne die Viabilität der Phagozyten zu beeinträchtigen. Darüber hinaus konnte gezeigt werden, dass Beschichtungen mit PLL-g-PEG über einen Zeitraum von mindestens drei Wochen stabil waren, selbst in Gegenwart von humanem Serum.

Kapitel IV beschäftigt sich schliesslich mit dem Rezeptor-spezifischen Targeting von CLRs auf APC. Dafür wurden verschiedene Mannosid-Liganden (Mannan von S. cerevisiae, sowie mono- und tri-Mannose) in kontrollierten Konzentrationen mit Hilfe von PLL-g-PEG auf Mikropartikeln immobilisiert.

14 ZUSAMMENFASSUNG

Die Erkennung hochmolekularer Mannose-Strukturen, sogenannter Pathogen-assoziierter molekularer Muster (engl.: pathogen associated molecular patterns, PAMPs), spielt eine wichtige Rolle in der Immunabwehr von Pathogenen. Unserer Hypothese zufolge sollte die Präsentation der Mannoside am äusseren Ende flexibler PEG-Ketten die Nachahmung dieser hochmolekularen PAMPs durch kleine synthetische Liganden ermöglichen. Die in diesem Kapitel behandelten Aspekte umfassen sowohl die Formulierung derartiger Mikropartikel als auch ihre spezifische Erkennung durch APC, sowie mögliche Effekte der Mannosid-Substitution auf die Maturierung dendritischer Zellen. Darüber hinaus wurden Kombinationen der Mannosid-Liganden mit einem Integrin-bindenden Liganden, einer RGD-Peptidsequenz, aufgebracht. Dank der Beschichtung mit PEG und/oder Mannan waren die modifizierten Mikropartikel gegen Proteinadsorption passiviert, so dass spezifische Wechselwirkungen zwischen Liganden und Rezeptoren untersucht werden konnten. Die Phagozytose-Raten der modifizierten Mikropartikel wurden in erster Linie von der Mannosid-Dichte bestimmt, allerdings wurde insgesamt nur eine geringe Effizienz beobachtet. Dies unterstützt eine vor kurzem veröffentlichte Hypothese, derzufolge der Mannose-Rezeptor (MR) nur in Gemeinschaft mit noch nicht identifizierten Co-Rezeptoren als phagozytierender Rezeptor wirkt. Ein weiteres Ergebnis der Studie war, dass die Aufnahme mit Mannan beschichteter Mikropartikel keine Maturierung dendritischer Zellen auslöste. Dies bedeutet, dass die Aktivierung von CLRs allein nicht in der Lage ist, dendritische Zellen in Richtung einer T-Zell-Aktivierung zu stimulieren, sondern dass dafür offensichtlich eine Stimulierung weiterer Rezeptoren und/oder die gleichzeitige Verarbeitung eines Antigens erforderlich sind.

Der in dieser Dissertation etablierten mikropartikulären Plattform kommt eine vielversprechende Rolle im Hinblick auf die Erforschung spezifischer Ligand-Rezeptor-Wechselwirkungen auf phagozytierenden Zellen zu, einschliesslich der Identifizierung potentieller Liganden und Ligandenkombinationen, sowie der Untersuchung der durch diese Wechselwirkungen ausgelösten immunmodulierenden Effekte. Dies wird das Wissen um die ersten Schritte der Kontakte zwischen APC und Mikropartikeln oder Pathogenen und die daraus resultierende Immunkaskade um wertvolle

ZUSAMMENFASSUNG 15

Details vertiefen, von denen wir annehmen, dass sie zu verbesserten Delivery-Systemen für Impfstoffe führen können.

16 ZUSAMMENFASSUNG

CHAPTER I 17

CHAPTER I

PEGylation as a Tool for the Biomedical Engineering of Surface Modified Microparticles

Uta Wattendorf and Hans P. Merkle

Institut for Pharmaceutical Sciences, ETH Zurich, 8093 Zurich, Switzerland

Journal of Pharmaceutical Sciences, 2008. DOI 10.1002/jps.21350.

18 CHAPTER I

ABSTRACT

Microparticles are of considerable interest for drug delivery, vaccination and diagnostic imaging. In order to obtain microparticles with long circulation times, or to provide the prerequisite for tissue specific targeting through decoration with suitable ligands, their surfaces need to be modified such that they become repellent to the adsorption of opsonic proteins and resistant to unspecific phagocytosis. The currently most considered strategy relies on the immobilization of a poly(ethylene glycol) (PEG) corona onto the microparticles’ surface. In the first chapter of this review, we discuss the unique physicochemical properties of PEG, which make it the polymer of choice to render the surfaces of microparticles repellent to the adsorption of proteins and resistant to cellular recognition. Furthermore, we present various technologies for the preparation of microparticles with PEGylated surfaces. Another aspect is the decoration of the PEGylated surfaces with suitable ligands for cell specific recognition and targeting. Finally, we review miscellaneous applications of PEGylated microparticles, mainly focusing on the fields of drug delivery, targeting and vaccination. Although still in its infancy, the PEGylation of microparticles holds promise towards future biomedical applications.

CHAPTER I 19

INTRODUCTION

Microparticles, along with other particulate carriers such as nanoparticles and liposomes, are of considerable interest for drug delivery, vaccination and diagnostic imaging [1-3]. Among the advantages of microparticulate formulations in drug delivery are the stabilization of the drug and its controlled and sustained release, as well as the potential for localized delivery of the drug to a particular body compartment [4, 5]. Microparticles of about 1 to 10 µm in diameter are also logical delivery and targeting systems for vaccines, as they undergo rapid phagocytosis by antigen presenting cells (APC), an exclusive feature of phagocytes such as dendritic cells and macrophages. Moreover, upon ingestion by antigen presenting cells, microparticles have the ability to release antigen in a controlled and sustained manner, such that the number of booster injections required to provide full immunological protection may be reduced [6]. Microparticles have also been associated with immunomodulatory properties. In fact, microparticulate antigen delivery systems, when phagocytosed by antigen presenting cells, have been shown to cause cross presentation on both MHC I and II molecules [7] and thus elicit cytotoxic T lymphocyte (CTL) immunity in addition to T helper cell and antibody responses [8-11].

Upon entry into the organism, microparticulates are instantly recognized as foreign and are efficiently removed and digested via unspecific phagocytosis by the mononuclear phagocytic system (MPS), which (in addition to other specialized organs and tissues) also comprises the phagocytes of the innate immune system [12]. Extent and rate of phagocytic clearance depends on both particle size and surface properties of the carrier. For instance, phagocytosis is enhanced by charged or hydrophobic surfaces [13-15]. It is particularly the adsorption of opsonic proteins onto the particles´ surface that promotes phagocytosis [16-18]. Therefore, in order to obtain microparticles with long circulation times, or to ensure tissue specific targeting through decoration with suitable ligands, their surfaces need to be modified such that they become resistant to unspecific phagocytosis after opsonization.

While there are various approaches available to prevent unwanted opsonization and cellular recognition [19-21], most strategies rely on the

20 CHAPTER I

immobilization of a poly(ethylene glycol) (PEG) corona onto the microparticles’ surface. Due to the uncharged and highly hydrophilic character of the PEG corona, in combination with the low toxicity and immunogenicity of PEG, such particles adopt stealth character which makes them attractive tools for biomedical applications [15, 22]. In addition to PEG, stealth character may be also conferred by other synthetic polymers, such as poly(acryl amide) or poly(vinyl pyrrolidone), and by polysaccharides, such as dextranes [23, 24].

In the first chapter of this review, we discuss the unique physicochemical properties of PEG, which make it the polymer of choice to render surfaces resistant to both the adsorption of proteins and cellular recognition. Furthermore, we present various technologies for the preparation of microparticles with PEGylated surfaces. Finally, in a third section, we review miscellaneous applications of PEGylated microparticles, mainly focusing on the fields of drug delivery, targeting and vaccination.

PROPERTIES OF PEG

There are various approaches to prevent unwanted adsorption of serum proteins and cellular recognition [19-21, 24], with most of them relying on the immobilization of poly(ethylene glycol) (PEG) at the surface. Alternative polymers include poly(acryl amide), poly(vinyl pyrrolidone), or polysaccharides [23, 24]. Polymers with the capacity to confer protein repellence to surfaces share a number of properties such as hydrophilicity, presence of hydrogen-bond acceptors but absence of hydrogen-bond donors, and electrical neutrality [25]. PEG, also denoted as poly(ethylene oxide) (PEO), is a polyether composed of the simple repeating unit HO-(CH2-CH2-O)n-H with a monomolecular mass of 44 g/mole. Thus, a PEG molecule of 1 kDa consists of approximately n = 23 repeating units. PEG is an uncharged, hydrophilic polymer that is soluble in water as well as in many organic solvents. Due to its low toxicity and immunogenicity, PEG is highly suitable for biomedical applications [22, 26, 27]. The solubility of PEG in water is unusual, considering the fact that both its two neighbours in the homologous series, poly(methylene glycol) and poly(propylene glycol), are insoluble in water. This is believed to result from the

CHAPTER I 21

Figure 1. The trans and gauche conformations of PEG, and their implications in waterbonding.

preferred (polar) gauche conformation of PEG in water [28], which offers two hydrogen bond acceptors in ideal distance for hydrogen bonding with water (Figure 1). This conformation leads to extensive hydration in aqueous environments which, along with good conformational flexibility and high chain mobility, causes a steric exclusion effect that prevents the adsorption of proteins [22, 29, 30].

As demonstrated theoretically [30-33] and experimentally [34-39] protein repellence of PEG coatings depends on both chain length and chain density which jointly determine the thickness of the adlayer. The conformation of surface grafted PEG chains is dictated by the relation between the distance of two adjacent chains (D) and their radius of gyration (Rg), the latter being related to the chain length. The polymer forms a so called “mushroom” conformation for D > 2 Rg, while a “brush” like conformation appears at D << Rg when - for steric reasons - the polymer chains stretch out perpendicularly from the surface. However, these two conformations do not represent sharply separated regimes, but undergo smooth transitions through mushroom/brush intermediates for 2 Rg > D > Rg (Figure 2) [29, 40].

Protein repellence becomes increasingly efficient the more the PEG chains overlap [37]. However, longer chains were found to have the ability to fill the gaps between less densely grafted PEG chains [38]. Theoretical studies have revealed two modes of protein adsorption [33]: (i) primary adsorption at the PEG-substrate interface and (ii) secondary adsorption at the top of the PEG brush. Small proteins may penetrate the PEG brush and adsorb onto the

22 CHAPTER I

Figure 2. Schematic representation of different conformations of surface-grafted polymers resulting from increasing chain densities.

underlying surface. Increasing PEG chain density is expected to reduce the degree of primary adsorption. Repellence of larger proteins, which may undergo secondary adsorption, requires a sufficiently thick PEG layer to screen long range forces such as, e.g., electrostatic protein-substrate interactions.

PEGylation of the surfaces of microparticles not only has the capacity to reduce the adsorption of proteins, but also to cut off the particles´ cellular recognition. In vivo, most injected microparticulate drug carriers are instantly recognized as foreign material and efficiently removed via phagocytic clearance by the mononuclear phagocytic system (MPS), comprising phagocytes residing, e.g., in the liver, spleen, lungs, and bone marrow, but also patrolling phagocytes such as dendritic cells and macrophages [12, 40]. In contrast to macromolecules and nanoparticles, which may be endocytosed through alternative mechanisms such as pinocytosis and macropinocytosis, microparticles are exclusively internalized by phagocytes [41], making them a preferred platform for vaccine delivery. Extent and rate of phagocytic clearance is dependent on size and surface properties. For instance, phagocytosis is enhanced by charged or hydrophobic surfaces [13-15]. Most importantly, adsorption of opsonic proteins onto the surface of hydrophobic particles promotes phagocytosis [16-18].

To date, a broad variety of PEGylated delivery systems has been reported in literature. PEGylated systems based on liposomes, nanoparticles and polymeric micelles have been extensively reviewed elsewhere [21, 40, 42-46]. Applications of PEGylated microparticles, among others, comprise long-circulation stealth particles or particles for ligand- and receptor-specific recognition or targeting, and will be discussed in detail in another section.

CHAPTER I 23

ENGINEERING PEGYLATED MICROPARTICLES

In general, there are two contrasting strategies to engineer microparticles carrying a PEG based hydrophilic corona: (i) formation of microparticles in the presence of copolymers bearing PEG moieties, or (ii) surface modification by chemical coupling of PEG moieties or physisorption of PEG copolymers onto the surface of earlier fabricated microparticles (Figure 3).

croparticles in the presence of copolymers bearing PEG moieties, or (ii) surface modification by chemical coupling of PEG moieties or physisorption of PEG copolymers onto the surface of earlier fabricated microparticles (Figure 3).

Formation of microparticles in the presence of PEG copolymers Formation of microparticles in the presence of PEG copolymers

When microparticles are fabricated in the presence of PEG copolymers, PEG moieties may be expected both in the bulk and on the surface of the microparticles. In this case the surface may not necessarily be densely packed with PEG chains. For biodegradable microparticles the presence of the PEG

When microparticles are fabricated in the presence of PEG copolymers, PEG moieties may be expected both in the bulk and on the surface of the microparticles. In this case the surface may not necessarily be densely packed with PEG chains. For biodegradable microparticles the presence of the PEG

Figure 3. Strategies for the formation of PEGylated surfaces on microparticles.

24 CHAPTER I

copolymer throughout the bulk of the microparticle has been proposed to ensure a sustained availability of surface exposed PEG moieties during the entire biodegradation and -erosion process of the particle [47]. As the fabrication of microparticles typically involves emulsification, precipitation or dispersion polymerization in aqueous media, PEG moities are likely to orient themselves preferrably towards the water phase, while the more hydrophobic portion of the copolymer remains either chemically bound or physically entangled with the polymer matrix, thus leading to microparticles with PEG enriched surfaces and PEG poor cores [47]. This requires sufficient surface activity of the copolymer. In special cases, such as the self-assembly of block copolymers, core-shell structures may be formed due to different solubilities of the building blocks in the dispersion medium [44, 45, 48].

In recent literature there are manifold examples for the fabrication of microparticles with PEGylated surfaces through, e.g., addition of PEG and PEG-derivatives during spray-drying or emulsion based solvent evaporation techniques [49-54], copolymerization of PEG with other matrix forming polymers [55, 56], and formation of microparticles from PEG block or random copolymers [57-61]. Rarer are microparticles entirely composed of PEG as matrix material, which are mainly prepared with the help of supercritical CO2 [62-64]. However, the reports on the latter type of microparticles are mostly of methodological nature. So far, information regarding their protein adsorption or their recognition by phagocytes is missing.

Surface modification with PEG on preformed microparticles

The second option, i.e., the surface modification of preformed microparticles with PEG leads to the formation of a more or less dense PEG corona. It can be accomplished by different strategies, which may be classified into covalent or non-covalent approaches.

Covalent coupling of PEG onto microparticles involves chemical protocols that are similar to those for the PEGylation of proteins, flat surfaces and nanoparticles, and have been reviewed extensively [26, 27, 46, 65, 66]. The predominant reaction scheme followed for covalent PEGylation of

CHAPTER I 25

Figure 4. Coupling via succinimidyl active esters.

microparticles is the coupling via succinimidyl esters (Figure 4) [67-71]. In this coupling reaction, an intermediate active ester is formed by reaction of a carboxyl group with a carbodiimide such as N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide (EDC). The intermediate active ester is then converted into the more stable succinimidyl ester with N-hydroxysuccinimide (NHS). Finally, the succinimide is replaced by an amino group to form the final adduct. This reaction is not necessarily restricted to the presence of carboxyl groups because these may be generated by reaction of, e.g., succinic anhydride with hydroxyl groups [66]. In contrast to other coupling reactions, the carbodiimid method can be performed under mild conditions. However, it usually requires reaction times in the range of hours [68, 69, 71]. This may be unfavorable when the particles were previously loaded with therapeutics which may then be eluted or inactivated.

The reaction can be performed in two ways: Either amino-PEGs are coupled to surface bound activated esters, or activated PEG-esters are reacted with surface-bound amino groups. The first strategy is advantageous because many types of microparticles with carboxylated surfaces are commercially available, that can be easily converted into NHS esters. Nevertheless, the second

26 CHAPTER I

strategy is interesting for microparticles that bear hydroxyl groups on their surface, which can be conveniently aminated with (3-aminopropyl)triethoxysilane (APTES, Figure 5).

PEG has two equivalent hydroxyl groups, which could lead to cross linking upon covalent coupling. Therefore, the monomethyl-ether of PEG, or monomethoxy-PEG (mPEG), is generally used such that coupling can only occur at one end of the PEG molecule [26, 46]. A general problem of the covalent grafting of PEG onto surfaces is the fact that owing to steric repulsion between the chains high PEG surface densities are difficult to achieve, unless special experimental conditions are used, such as carrying out the grafting reaction under cloud point conditions [38, 72].

An alternative approach to produce surfaces with high density PEG brushes is the physisorption of PEG grafted polymers. Non-covalent immobilization of PEG on a microparticle involves the adsorption of copolymers with either block or grafted architecture, such that the PEG part sticks out into the aqueous dispersion medium while the other part interacts with the microparticle´s surface. Such interactions can be of either hydrophobic or electrostatic nature, the latter requiring fixed complementary charges on the surface and the polymer.

A typical group of copolymers that can be immobilized via hydrophobic forces are the Poloxamers, also known as Pluronics, which are triblock copolymers consisting of two equal poly(ethylene oxide) chains linked to a

Figure 5. Surface amination with APTES.

CHAPTER I 27

middle segment of poly(propylene oxide) of the general formula PEOa-PPOb-PEOa ([73], applying the commonly used alternative terminology PEO for PEG). The most widely used Poloxamer for the surface PEGylation of microparticles in recent literature is Poloxamer 407 (= Pluronic F127; with a=101 and b=56). It carries PEO chains of about 4.4 kDa and has an overall molecular weight of approximately 12 kDa [74, 75]. Due to the high percentage of PEO, Poloxamers are well soluble in water or aqueous buffer solutions. When microparticles are incubated in such a solution, the more hydrophobic PPO segment adsorbs onto the surface, while the two hydrophilic PEO segments protrude into the aqueous medium. As a general trend, the PEO chain density of surface immobilized Poloxamers was found to be independent of the size of the PEO segments, and thus appeared to be governed primarily by the length of the PPO anchor block [73]. On the other hand, increasing the molecular weight of the PEO segments increases the thickness of the adlayer, and was generally found to improve repellence against protein adsorption [73].

A copolymer that can be immobilized via electrostatic interactions is, e.g., PLL-graft-PEG, which consists of PEG chains grafted to a fraction of the side chain amino groups of a poly(L-lysine) (PLL) backbone. Because of the positive charge of the remaining free amino groups at physiological pH, the polycationic copolymer spontaneously adsorbs efficiently onto negatively charged surfaces through electrostatic interactions, leading to a monolayer with the PEG chains stretched out perpendicularly to the surface [76]. Simulations using a self-consistent-field approach revealed the PEG end-segment distribution to be farther displaced from the PLL backbone for decreasing grafting ratios (g, defined as the ratio of the total number of L-lysine monomers devided by the number of PEG side chains) [77]. This is concluded to be a consequence of an increased stretch-out of the PEG chains due to increased steric repulsion between the PEG chains [77]. Appropriate choice of polymer architecture renders such coated surfaces protein repellent even in the presence of full human serum [39, 76]. The optimum architecture, i.e. 20 kDa backbone, 2 kDa PEG side chains, and a grafting ratio of 3.5, was found to efficiently minimize opsonization and recognition by phagocytes [78]. In fact, even after several weeks of pre-incubation of PLL(20)-g[3.5]-PEG(2) coated carboxylated

28 CHAPTER I

polystyrene microparticles in serum containing cell culture medium, only minor phagocytosis by human derived dendritic cells was observed [79]. In addition, the terminal ends of the PEG side chains of PLL-g-PEG could be functionalized with bioligands, such as, e.g., adhesion peptides or biotin, while retaining efficient repellence to unspecific protein adsorption [80-83].

The assembly of PEG copolymer adlayers by physisorption is a fast and convenient technique allowing very mild conditions, such as incubation during 15 to 60 minutes in neutral buffered solutions followed by short aqueous washings. This could be of general advantage over chemical surface modifications which usually require more extended reaction times and sometimes more rigorous reaction conditions [66], both of which may unfavorably affect the chemical and physical stability of embedded peptide and protein therapeutics. Nevertheless, a drawback of the physisorption of PEG copolymers was that under certain conditions the adsorbed copolymers may fully or partially desorb, leaving unprotected surface defects which were prone to opsonization [47].

FIELDS OF APPLICATIONS FOR PEGYLATED MICRO-PARTICLES

In the vast majority of cases in literature, the use of PEG in the fabrication of microparticles is not for surface modification. Instead, PEG is employed as an additive, e.g., to enhance drug stability during both the formation of the particles and their biodegradation after injection, or to modify drug release [46, 50, 53, 54, 59]. In addition, PEGylation of proteins has been considered as a tool to enhance drug solubility and stability upon microencapsulation [26, 66]. However, as elaborated in more detail in this section, here we focus on the use of PEG moieties for surface modification with the aim to reduce (unwanted) protein adsorption, platelet activation and clearance of microparticles by phagocytes.

CHAPTER I 29

Biophysical, biotechnological and biomedical applications

Apart from drug delivery applications, PEGylated microparticles were also found to be useful for biophysical, biotechnological and biomedical applications. For instance, PEGylation may improve assays that traditionally apply BSA as a blocking reagent to reduce unspecific interaction with proteins. Very recently, Nagasaki et al. [67] developed an oligoamine terminated PEG, pentaethylenehexamine-PEG (N6-PEG), which was immobilized onto the surface of carboxylated magnetic microbeads (2.2 μm) for application in an immunomagnetic ELISA. This strategy took advantage of both covalent bonding to NHS ester activated carboxyl groups and blocking of remaining sites through electrostatic interaction. Remarkably, a 9:1 mixture of two PEGs of 6 and 2.5 kDa, respectively, was shown to yield optimum results in blocking nonspecific binding of proteins. To develop an immunomagnetic ELISA, antibodies were covalently coupled to the microbeads which were subsequently surface modified with the N6-PEG mix. The PEGylated microbeads were found to have a 20-fold increased signal to noise (S/N) ratio as compared to conventional BSA blocking. This improvement was achieved both by reduction of the background due to improved blocking of unspecific binding, and by an increase in the signal due to improved orientation and accessibility of the immobilized antibodies within the PEG brush.

PEGylated microspheres were also used in microrheology, in order to probe local viscosity, elasticity and microstructure of biopolymer networks such as, e.g., entangled or cross-linked actin networks [69]. As an alternative to the standard blocking of binding sites by pre-incubation with BSA, carboxylated poly(styrene) microspheres (1 μm) were grafted with PEG via the carbodiimid method. It was found that PEGylation reduced binding of a model protein to approximately 2% of the uncoated control. Furthermore, PEGylated microspheres were able to freely move in the biopolymer networks, while BSA coated microspheres adhered to the filaments of the network, whereas untreated microspheres aggregated.

Recently, Wang et al. introduced PEGylated poly(glycidyl methacrylate-divinylbenzene) [P(GMA-DVB)] microspheres of 30-50 μm as a more rigid

30 CHAPTER I

alternative to conventional agarose bead matrices in protein purification [84]. The authors proposed PEGylation to prevent irreversible binding and denaturation of proteins, which may otherwise occur on hydrophobic polymer particles. PEGylation was achieved by reaction between the epoxy groups of the GMA moieties and the hydroxyl groups of the PEG in the presence of boron trifluoride as catalyst. In this study, hydrophilicity and PEG density of the microspheres was investigated as a function of both crosslinker concentration and PEG molecular weight, and the effect on the microspheres’ capacity to reversibly adsorb proteins was determined. The optimized formulation, containing 20% DVB and 4 kDa PEG, was able to adsorb approximately 50 mg protein per gram microsphere, and 97-99% of the adsorbed protein could subsequently be recovered. Thus, such surface modified microspheres show promise for the purification of proteins by hydrophobic interaction chromatography.

In the field of biomedical applications, chitosan microspheres containing the surface active triblock copolymer Pluronics (PEO-PPO-PEO) were suggested as potential cationic adsorbents for the removal of toxic compounds from plasma during extracorporeal circulation [51]. Microspheres of 15-30 μm were prepared by crosslinking chitosan with sodium triphosphate, with Pluronic added to the chitosan solution prior to crosslinking. Incorporation of Pluronics reduced fibrinogen adsorption to 15%, significantly reduced adhesion and activation of platelets, and even slightly enhanced the capacity for bilirubin adsorption.

PEGylation of larger microparticles (60-70 μm) has been studied for use in therapeutic embolization, a tool in the treatment of various vascular lesions [68]. As the efficiency of this method is limited by revascularization, which in turn is associated with an inflammatory reaction, PEGylation was studied as a means to reduce complement activation. Indeed, a grafted layer of PEG on the surface of the microparticles reduced complement activation by one third, and by combination of PEGylation and loading the microparticles with an anti-inflammatory drug, complement activation was almost fully suppressed.

CHAPTER I 31

Long circulating microparticles and drug delivery

Since the first studies by Illum et al. [85] in the late 1980s, interest has grown in the PEGylation of microparticles for drug delivery applications. Similar to nanoparticles, microparticles can be injected into the bloodstream if their largest diameter is smaller than 5 μm. PEGylation of such drug delivery systems reduces opsonization and strongly prolongs circulation times [42].

For the aim of long-circulating particles, Lacasse et al. added PEG-distearat to polyanhydride, PLGA or PLA solutions prior to spray-drying of microspheres. The resulting microsphere formulations (0.9-3.2 μm) were shown to shield the negative surface charge that was inherent to microspheres prepared without PEG-distearat. Moreover, PEGylation reduced phagocytosis by mouse macrophages.

A similar result was found with microbubbles, which were prepared by sonication of fluorocarbon gas in an aqueous solution containing lipidic stabilization agents [86]. Due to their surface charge negatively charged microbubbles own the problem of size-independent microvascular retention. It was shown that the mean number of retained microbubbles could be decreased by 70% through addition of lipid-bound PEG [86]. Interestingly, the negative ζ-potential of the microbubbles was not shielded by addition of PEGylated lipid, and complement (C3b) attachment was only reduced by about 30%. As only a fraction of the stabilizing lipid was PEGylated, it is probable that only a limited PEG density was achieved, which was insufficient to confer full stealth properties to the microbubbles.

Gbadamosi et al. correlated the ability of microspheres for prolonged circulation with the density of the PEG corona [87]. Microspheres were PEGylated in a two step process. First, a monolayer of BSA was adsorbed onto 1 μm poly(styrene) microspheres, followed by covalent binding of 5 kDa monomethoxy-PEG with cyanuric chloride. Hydrophobic interaction chromatography revealed that this process yielded different subpopulations of microspheres, which exhibited different degrees of resistance against phagocytosis in a murine macrophage cell line. The authors found a linear relationship between the surface charge of the subpopulations and the extent of

32 CHAPTER I

phagocytosis, with more neutral subpopulations being more resistant to phagocytosis. It was concluded that populations bearing PEG predominantly in a high-density mushroom-brush intermediate and/or brush configuration were most resistant to phagocytosis, while populations with predominant mushroom regimes of PEG were potent activators of the complement system and prone to phagocytosis.

In addition to the importance of high PEG density and sufficient molecular weight, Makino et al. [60] stressed the importance of softness. In their study, hydrophilic gel microcapsules (15-30 μm) were polymerized using two different compositions of three monomers (i.e. N,N-dimethylacrylamide and 4-(aminomethyl)-styrene copolymerized with N,N-dimethylaminopropyl-acrylamide or 2-[(methacryl-oyloxy)ethyl] trimethylammonium chloride, respectively), with or without addition of PEG to each of the two sets. PEGylation was found to decrease the charge of the outer layer and to increase the softness of the microcapsules, as well as to reduce the haemolytic interaction with red blood cells. The correlation of surface charge, hydrophilicity and softness of the four types of microcapsules with their haemolytic effects revealed that the interaction with red blood cells was independent of the outer surface charge but strongly reduced with increasing softness. Therefore, the authors postulated that above all synthetic polymer surfaces should be soft and hydrophilic to control their interaction with cell surfaces.

Coupling of amino-PEGs to carboxylated poly(styrene) microspheres via carbodiimide chemistry was chosen by Meng et al. [70, 71] Amino-PEGs of different molecular weights (1.5, 3.4 or 5 kDa) and with different endgroups (amino-, hydroxy- or methoxy-) were used. The authors found that protein adsorption from human plasma decreased with increasing PEG chain density and with increasing PEG molecular weight. However, no effect of the endgroups was observed. Protein adsorption was inhibited by 90-95% at most. Moreover, PEGylation of microspheres not only led to decreased total amounts of adsorbed protein but also changed the composition of adsorbed proteins towards typical dysopsonins, such as albumin and apolipoproteins, thus confirming results from an earlier study by Lück et al. [61] on microspheres formed by spray-drying of PEG-containing triblock copolymers. Finally, interaction with human umbilical

CHAPTER I 33

vein endothelial cell (HUVEC) layers was investigated to simulate interaction of injected particles with vascular endothelium [71]. In line with the prediction from the protein adsorption studies, PEGylated microspheres were generally found to strongly reduce adhesion to HUVEC layers, with longer chains being more effective than short chains. Again, no effect of the endgroups was observed.

Other authors relied on the self-assembly of PEO-PPO-PEO triblock copolymers (Poloxamers or Pluronics) to prepare stealth microparticles. Jones et al. [88] coated small (2-3 μm) PLGA microparticles with Poloxamer 338 aiming at an increase in bioavailability of orally administered drugs through an inhibition of the phagocytosis of respirable microparticles. Poloxamer 338, also known as Pluronic F108, features PEG end segments of approximately 6 kDa attached to the middle PPO segment of approximately 2.5 kDa, and adsorbs onto microparticles via hydrophobic interactions. In a study with three different macrophage cell models, some reduction in the number of phagocytosed coated microparticles or the percentage of phagocytosing cells was found. However, there was no general trend for all three studied cell models, and the reduction was less as compared to older studies [73, 89] by other authors. The authors explained their results by assuming defective domains on the microparticles’ surface that remained uncovered by the Poloxamer. Such defects in surface modification would render the microparticles accessible for ingestion by the cells. Regrettably, no experiments were reported to characterize the polymeric adlayer or its protein repellent character. Thus, it is difficult to judge whether the failure in resistance to phagocytosis should be attributed to the specific cell models, to the preparation of the coatings, or to the chosen type of Poloxamer. One could also speculate that the stealth character was incomplete because the Poloxamer 338 adlayer was not dense enough due to the increased steric repulsion of the relatively long PEG chains used in this study.

Noticeably, another Poloxamer (Poloxamer 407 or Pluronic® F127), bearing a similarly sized PPO middle segment (approx. 3 kDa) but smaller PEO end segments (approx. 4 kDa), was successfully immobilized on microparticles made of various matrix materials as reported by Jackson et al. [74] The

34 CHAPTER I

formulations included microparticles of poly(lactic acid) (PLA), poly(ε-caprolactone) (PCL), poly(methyl methacrylate) (PMMA) or a 50:50 blend of PLA and poly(ethylene-co-vinyl acetate) (PLA:EVA) with comparable size distributions and average diameters of 40 μm. With the final aim to reduce the inflammatory potential of microparticulate drug delivery systems, the authors studied both the adsorption of IgG and plasma proteins onto bare and surface modified microparticles, and the activation of neutrophils when incubated with the microparticle formulations. Uncoated microparticles were found to adsorb considerable amounts of protein, which was accompanied by a significant activation of neutrophils in vivo. On the other hand, Poloxamer coated microparticles were found to repel protein adsorption, thus inhibiting the opsonization effect of IgG and plasma on neutrophil activation. The authors concluded that pretreatments of microparticles with Pluronic® F127 reduced the inflammatory potential of microparticles in vivo, since neutrophil activation is part of the inflammation process.

A system that has been studied in detail during the past 5 years consists of PLL-g-PEG coated microparticles. This polymer is synthesized by grafting PEG chains to the amino side groups of the polycationic PLL backbone, and assembles readily onto negatively charged microparticles, such as carboxylated poly(styrene) or PLGA microparticles, via electrostatic interactions [39, 80]. First studies on the protein repellence of PLL-g-PEG coated microparticles for drug delivery applications were performed by Müller et al. [80] The PLL-g-PEG used was a copolymer with PEG chains of 2 kDa grafted to a PLL backbone of 20 kDa at a grafting ratio of 3.5, which was known for its protein repellence from a previous study [90] on flat metal oxide surfaces. Spray-dried PLGA microspheres (1-10 μm) were coated with PLL-g-PEG and subsequently incubated with the model proteins human serum albumin, fibrinogen, immunoglobulin G, and fibronectin. A similar study was carried out by Heuberger et al. [91] with PLL-g-PEG coated hollow polyelectrolyte microcapsules and the model protein streptavidin. In both studies, quantification of the fluorescently labelled model proteins by confocal microscopy revealed a reduction in protein adsorption by two to three orders of magnitude.

CHAPTER I 35

From the aforementioned studies on protein repellence it was hypothesized that PLL-g-PEG coated microparticles could be shielded from phagocyte recognition and uptake. This was expected due to the protection of the microparticles from opsonization, i.e. the adsorption of plasma proteins, which represents the first step that finally leads to recognition and uptake of the microparticles through phagocytes. The validity of this hypothesis was demonstrated by Faraasen et al. with human monocyte derived macrophages and dendritic cells representing professional antigen presenting cells [92]. Both adhesion of the cells to PLL-g-PEG coated glass slides and phagocytosis of PLL-g-PEG coated polystyrene or PLGA microspheres (5 μm) were shown to be strongly reduced as compared to uncoated controls.

We recently continued this work by screening a small library of PLL-g-PEG copolymers with PLL backbones of 20 kDa, with grafting ratios ranging from 2 to 20, and PEG chain lengths varying from 1 to 5 kDa [78]. After coating, poly(styrene) microspheres were characterized by their ζ-potential, IgG adsorption and phagocytosis by dendritic cells and macrophages. PLL-g-PEG with a grafting ratio of 3.5 and 2 kDa PEG chains was identified as the best suited copolymer of this library to inhibit opsonization and phagocytosis of coated microspheres. Nearly complete resistance to phagocytic uptake for PLL-g-PEG coated microspheres versus striking uptake of non-coated microparticles is demonstrated in Figure 6.

Recently, we studied the physical persistence of such PLL-g-PEG coatings. Even after 3-4 weeks of incubation in buffer or serum supplemented cell culture medium, PLL-g-PEG coated microspheres and microcapsules showed no noticeable change in surface charge or resistance to phagocytosis [79].

Another potential field of application for PLL-g-PEG coated microparticles, which is currently studied in our laboratory, is the delivery of immunomodulatory signals to dendritic cells. Dendritic cells are equipped with a variety of conserved receptors that recognize pathogen associated molecular patterns (PAMPs) typical for parasite, bacterial or viral pathogens. Polyriboinosinic acid-polyribocytidylic acid, denoted by poly(I:C), resembles

36 CHAPTER I

Figure 6. Nearly complete resistance to phagocytosis for PLL-g-PEG coated microspheres (unlabelled) versus efficient phagocytosis of non-coated microspheres (fluorescently labelled). Reprinted from Ref. [78].

viral dsRNA, and is a specific ligand for the Toll-like receptor 3 (TLR3) on dendritic cells. Remarkably, poly(I:C) assembled onto previously PLL-g-PEG coated microparticles led to markedly better expression of maturation markers on dendritic cells after phagoctosis as compared to the same dose of soluble poly(I:C). Best results were achieved with relatively “open” PLL-g-PEG structures, i.e. with polymers of high grafting ratios g ≥ 10, that allowed the PLL-g-PEG adlayer on the microparticles to accommodate poly(I:C) in the voids between the PEG chains of the corona. This is assumed to protect poly(I:C) from premature degradation by ubiquitous RNase, a fundamental drawback of unprotected poly(I:C) (A. Hafner, unpublished data).

Using a similar but more complex approach, our group also aims at the co-delivery of immunomodulatory agents such as poly(I:C) (mimicking viral RNA) or CpG (mimicking bacterial DNA), with plasmid DNA or mRNA encoding protein antigens. For this purpose we prepared layer-by-layer coated microparticles carrying alternating nanoscale layers of a cationic polymer and a

CHAPTER I 37

negatively charged nucleic acid. Coating this construct with a third layer consisting of PLL-g-PEG would allow to additionally accommodate the immunomodulator as a fourth layer (N. Csaba, unpublished data). With such delivery systems it is hoped to further enhance the immunomodulatory potential of microparticulate vaccines.

PEGylated microparticles for drug targeting

An important aspect in drug delivery is the targeting of a drug bearing delivery systems to a specific tissue [93]. In this context, the PEG corona on a microparticle is not only expected to protect from unspecific phagocytosis as seen above, but also to provide a linker molecule for the decoration of the microparticle with a targeting ligand [22]. The role of the linker may not be limited to the coupling of the ligand. Another role could be that of a spacer to enhance the access of the ligand to its target receptor. The complexity of this concept is stressed by the assumption that the observed steric repulsion through the PEGylated surface of microparticles may not only repel opsonising proteins but also restrict ligand-receptor recognition [73, 94]. In a study with PEG-stabilized microbubbles bearing the model ligands biotin or FITC, respectively, Kim et al. [94] found that binding to corresponding surface-bound receptors, avidin or anti-FITC-antibody, could not take place unless the ligands were coupled via a sufficiently long spacer. In fact, the chain length of the PEG spacer (3.4 kDa) for the ligand exceeded that of the PEG corona by 1.6 kDa. The authors concluded that in order to prevent a ligand from being hidden in a PEG corona, the spacer should be sufficiently long.

The double function of PEG as stealth corona and spacer molecule was also studied with microspheres coated with a PLL-g-PEG copolymer that was terminally decorated with biotin. While PEG chains of 2 kDa were sufficient to ensure stealth function, PEG chains of 3.4 kDa were used for the coupling of the ligand to ensure its accessibility. The feasibility of this approach was shown by Müller et al. [80]. Accessibility of the biotin ligand was demonstrated with fluorescently labelled streptavidin. Increasing ligand densities led to increasing streptavidin binding.

38 CHAPTER I

The concept of biofunctionalization was further stressed by Heuberger et al., who studied biotin-substituted PLL-g-PEG coatings on hollow polyelectrolyte microcapsules, with about 30% of the PEG chains terminally substituted with biotin [91]. Streptavidin binding to coated and uncoated microcapsules was analysed by confocal microscopy. Specific binding of streptavidin to microcapsules coated with biotin-substituted PLL-g-PEG was observed. This study was the first to widen PLL-g-PEG surface coating technology from solid microparticles to hollow microcapsules.

The specific recognition of a biotin ligand on PLL-g-PEG coated microparticles was further investigated by Raichur et al. [95], measuring the apparent adhesion areas of hollow polyelectrolyte microcapsules on poly(ethylene imine) or streptavidin coated surfaces by reflective interference contrast microscopy (RICM). In contrast to the strong and fast adhesion of untreated microcapsules on such surfaces, PLL-g-PEG coated microcapsules did not engage in significant adhesion, whereas capsules functionalized with biotinylated PLL-g-PEG showed a significantly larger adhesion area.

Finally, the first report on a specific recognition of microspheres that were coated with PLL-g-PEG substituted with an adhesion peptide, RGD, was presented by Faraasen et al. [92] Because of the abundant occurance of RGD recognizing integrin receptors on many cell types, however, this peptide is unsuitable for studies aiming at cell specific targeting. Therefore, our current studies focus on mannoside-substituted PLL-g-PEG to investigate ligand- and receptor-specific phagocytosis by antigen presenting cells carrying mannose specific lectin receptors. Due to the PEG corona, microspheres coated with mono- or trimannose-substituted PLL-g-PEG were found to be repellent to unspecific protein adsorption, thus shielding opsonization and providing specific ligand-receptor interactions. In studies with human monocyte derived dendritic cells and macrophages, mannose density was identified as the major factor for phagocytosis of mannosylated microspheres, though with limited efficiency (U. Wattendorf, unpublished results). This strengthens the recent hypothesis by other authors [96] that the mannose receptor (MR) can only act as a phagocytic receptor in collaboration with yet unidentified partner receptors. The microparticulate platform established in this work represents a promising tool

CHAPTER I 39

for a systematic investigation of specific ligand-receptor interactions on phagocytes, including the screening for potential ligands and ligand combinations.

Other studies have focused on the targeting of inflamed endothelial cells with ligands for endothelial cell adhesion molecules (ECAMs), which are often enriched at sites of pathological inflammation. For instance, ligands for P-selectin were coupled to lipid-PEG stabilized microbubbles of 1-10 μm as an approach towards a targeted ultrasound contrast agent [97, 98]. Two ligands, namely an anti-P-selectin antibody and polymeric sialyl Lewisx, yielded selective adhesion of microbubbles to inflamed endothelial tissue in vitro and in vivo. In these studies, the ligands were immobilized via a biotin-streptavidin bridge. This makes it an experimentally highly flexible concept which allows the attachment of a wide variety of biotinylated ligands without the need for major changes in conjugation chemistry. On the other hand, because of the presumably high antigenicity of the protein component, biotin-streptavidin bridged constructs are a priori unsuitable for clinical development.

Using the same concept, Sakhalkar et al. [99, 100] coupled biotinylated anti-ECAM antibodies to microparticles (1-2.5 μm average diameter) earlier prepared by an emulsion method from a biotinylated block copolymer consisting of poly(lactic acid) and PEG (PLA-PEG-biotin). Due to phase separation during particle formation, the microparticles´ surface was enriched with biotinylated PEG. Similarly to the aforementioned studies, binding of anti-ECAM antibodies (via neutravidin) enabled specific targeting of the microparticles to inflamed endothelium both in vitro and in a mouse model for inflammatory pathologies.

Finally, targeting of microparticles to tissues is even possible without coupling suitable ligands, e.g. by means of a magnetic field. In a recent study, PEG was added to the aqueous phase in the w/o emulsion/solidification based preparation of magnetic BSA microparticles that were loaded with the anti-cancer drug 5-fluorouracil [49]. Owing to the co-encapsulated magnetic Fe3O4

nanocrystals, injected microparticles could be enriched in the desired tissue. The authors could show that PEGylation of the microparticles led to an increased

40 CHAPTER I

delivery of the drug to the target tissue as compared to the PEG-free formulation.

OUTLOOK

PEGylation of microparticles for biomedical application is still in its infancy. As compared to PEGylated nanoparticles, currently receiving prime attention as particulate delivery systems, research on PEGylated microparticles is still of limited scope.

Particularly promising for PEGylated microparticles seems to be the field of vaccination. Their size range is similar to those of bacterial or parasite pathogens, and they are ideally and exclusively suited to undergo phagocytosis by antigen presenting cells. Therefore, they can more closely than other vaccine formulations mimic the interactions of such pathogens with the cells of the innate immune system. PEGylated microparticles may open pathways to avoid unspecific opsonization and complement activation, and – through decoration with suitable ligands - pave the way towards specific recognition by receptors on antigen presenting cells that direct type and intensity of the subsequent immune response. In view of substantial advancement in cell and tissue targeting, the combination of stealth characteristics and targeting ligands is also an intriguing concept not only to be exploited for the delivery of drugs and antigens to their targets, but also to address selected cellular signaling pathways by activation of specific cell surface receptors. Last but not least, PEGylated microparticles bear potential for biotechnological applications, where they may improve existing protocols and techniques. This inherent combination of medical, pharmaceutical and technological applications, involving scientists from many different areas, holds promise towards future applications of PEGylated microparticulate systems.

CHAPTER I 41

REFERENCES 1. Ha, C.-S., and Gardella Joseph, A., Jr., Surface chemistry of

biodegradable polymers for drug delivery systems. Chem. Rev., 2005. 105(11): p. 4205-4232.

2. Thiele, L., Merkle, H.P., and Walter, E., Phagocytosis of synthetic particulate vaccine delivery systems to program dendritic cells. Expert Review of Vaccines, 2002. 1(2): p. 215-226.

3. Svenson, S., Carrier-based drug delivery. ACS Symp. Ser., 2004. 879(Carrier-Based Drug Delivery): p. 2-23.

4. Langer, R., New methods of drug delivery. Science, 1990. 249(4976): p. 1527-1533.

5. Putney, S.D., and Burke, P.A., Improving protein therapeutics with sustained-release formulations. Nat. Biotechnol., 1998. 16(2): p. 153-157.

6. Gander, B., Trends in particulate antigen and DNA delivery systems for vaccines. Adv. Drug Deliv. Rev., 2005. 57(3): p. 321-323.

7. Stuart, L.M., and Ezekowitz, R.A.B., Phagocytosis: Elegant complexity. Immunity, 2005. 22(5): p. 539-550.

8. Men, Y., Tamber, H., Audran, R., Gander, B., and Corradin, G., Induction of a cytotoxic T lymphocyte response by immunization with a malaria specific CTL peptide entrapped in biodegradable polymer microspheres. Vaccine, 1997. 15(12-13): p. 1405-1412.

9. Peter, K., Men, Y., Pantaleo, G., Gander, B., and Corradin, G., Induction of a cytotoxic T-cell response to HIV-1 proteins with short synthetic peptides and human compatible adjuvants. Vaccine, 2001. 19(30): p. 4121-4129.

10. Partidos, C.D., Vohra, P., Jones, D.H., Farrar, G.H., and Steward, M.W., Mucosal immunization with a measles virus CTL epitope encapsulated in biodegradable PLG microparticles. J. Immunol. Methods, 1996. 195(1-2): p. 135-138.

11. Kovacsovics-Bankowski, M., Clark, K., Benacerraf, B., and Rock, K.L., Efficient major histocompatibility complex class I presentation of

42 CHAPTER I

exogenous antigen upon phagocytosis by macrophages. Proc. Natl. Acad. Sci. U. S. A., 1993. 90(11): p. 4942-4946.

12. Juliano, R.L., Factors affecting the clearance kinetics and tissue distribution of liposomes, microspheres and emulsions. Adv. Drug Deliv. Rev., 1988. 2(1): p. 31-54.

13. Ahsan, F., Rivas, I.P., Khan, M.A., and Torres Suárez, A.I., Targeting to macrophages: role of physicochemical properties of particulate carriers-liposomes and microspheres-on the phagocytosis by macrophages. J. Controlled Release, 2002. 79(1-3): p. 29-40.

14. Foged, C., Brodin, B., Frokjaer, S., and Sundblad, A., Particle size and surface charge affect particle uptake by human dendritic cells in an in vitro model. International Journal of Pharmaceutics, 2005. 298(2): p. 315-322.

15. Thiele, L., Rothen-Rutishauser, B., Jilek, S., Wunderli-Allenspach, H., Merkle, H.P., and Walter, E., Evaluation of particle uptake in human blood monocyte-derived cells in vitro. Does phagocytosis activity of dendritic cells measure up with macrophages? J. Controlled Release, 2001. 76(1-2): p. 59-71.

16. Thiele, L., Diederichs, J.E., Reszka, R., Merkle, H.P., and Walter, E., Competitive adsorption of serum proteins at microparticles affects phagocytosis by dendritic cells. Biomaterials, 2003. 24(8): p. 1409-1418.

17. Absolom, D.R., Opsonins and dysopsonins: an overview. Methods Enzymol., 1986. 132(Immunochem. Tech., Pt. J): p. 281-318.

18. Patel, H.M., Serum opsonins and liposomes: their interaction and opsonophagocytosis. Crit. Rev. Ther. Drug Carrier Syst., 1992. 9(1): p. 39-90.

19. Kingshott, P., and Griesser, H.J., Surfaces that resist bioadhesion. Curr. Opin. Solid State Mater. Sci, 1999. 4(4): p. 403-412.

20. Moghimi, S.M., Hunter, A.C., and Murray, J.C., Long-circulating and target-specific nanoparticles: theory to practice. Pharmacol. Rev., 2001. 53(2): p. 283-318.

CHAPTER I 43

21. Gref, R., and Couvreur, P., Stealth and biomimetic core-corona nanoparticles. Encyclopedia of Nanoscience and Nanotechnology, 2004. 10: p. 83-94.

22. Harris, J.M., Zalipsky, S., and Editors, Polyethylene glycol: Chemistry and Biological Applications. ACS Symp. Ser., 1997. 680: p. 489 pp.

23. Torchilin, V.P., and Trubetskoy, V.S., Which Polymers Can Make Nanoparticulate Drug Carriers Long-Circulating. Adv. Drug Deliv. Rev., 1995. 16(2-3): p. 141-155.

24. Lemarchand, C., Gref, R., and Couvreur, P., Polysaccharide-decorated nanoparticles. Eur. J. Pharm. Biopharm., 2004. 58(2): p. 327-341.

25. Ostuni, E., Chapman, R.G., Holmlin, R.E., Takayama, S., and Whitesides, G.M., A survey of structure-property relationships of surfaces that resist the adsorption of protein. Langmuir, 2001. 17(18): p. 5605-5620.

26. Bentley, M.D., Bossard, M.J., Burton, K.W., and Viegas, T.X., Poly(ethylene) glycol conjugates of biopharmaceuticals in drug delivery. Modern Biopharmaceuticals, 2005. 4: p. 1393-1418.

27. Harris, J.M., and Chess, R.B., Effect of PEGylation on pharmaceuticals. Nat. Rev. Drug Discovery, 2003. 2(3): p. 214-221.

28. Bjorling, M., Karlstrom, G., and Linse, P., Conformational Adaption of Poly(Ethylene Oxide) - a C-13 Nmr-Study. J. Phys. Chem., 1991. 95(17): p. 6706-6709.

29. Vermette, P., and Meagher, L., Interactions of phospholipid- and poly(ethylene glycol)-modified surfaces with biological systems: relation to physico-chemical properties and mechanisms. Colloids Surf. B, 2003. 28(2-3): p. 153-198.

30. Morra, M., On the molecular basis of fouling resistance. J. Biomater. Sci. Polym. Ed., 2000. 11(6): p. 547-569.

31. Jeon, S.I., Lee, J.H., Andrade, J.D., and De Gennes, P.G., Protein-surface interactions in the presence of polyethylene oxide. I. Simplified theory. J. Colloid Interface Sci., 1991. 142(1): p. 149-158.

44 CHAPTER I

32. Jeon, S.I., and Andrade, J.D., Protein-surface interactions in the presence of polyethylene oxide. II. Effect of protein size. J. Colloid Interface Sci., 1991. 142(1): p. 159-166.

33. Halperin, A., Polymer Brushes that Resist Adsorption of Model Proteins: Design Parameters. Langmuir, 1999. 15(7): p. 2525-2533.

34. Zhu, B., Eurell, T., Gunawan, R., and Leckband, D., Chain-length dependence of the protein and cell resistance of oligo(ethylene glycol)-terminated self-assembled monolayers on gold. J. Biomed. Mater. Res., 2001. 56(3): p. 406-416.

35. Prime, K.L., and Whitesides, G.M., Adsorption of proteins onto surfaces containing end-attached oligo(ethylene oxide): a model system using self-assembled monolayers. J. Am. Chem. Soc., 1993. 115(23): p. 10714-10721.

36. McPherson, T., Kidane, A., Szleifer, I., and Park, K., Prevention of Protein Adsorption by Tethered Poly(ethylene oxide) Layers: Experiments and Single-Chain Mean-Field Analysis. Langmuir, 1998. 14(1): p. 176-186.

37. Malmsten, M., Emoto, K., and Van Alstine, J.M., Effect of chain density on inhibition of protein adsorption by poly(ethylene glycol) based coatings. J. Colloid Interface Sci., 1998. 202(2): p. 507-517.

38. Kingshott, P., Thissen, H., and Griesser, H.J., Effects of cloud-point grafting, chain length, and density of PEG layers on competitive adsorption of ocular proteins. Biomaterials, 2002. 23(9): p. 2043-2056.

39. Pasche, S., De Paul, S.M., Vörös, J., Spencer, N.D., and Textor, M., Poly(L-lysine)-graft-poly(ethylene glycol) assembled monolayers on Niobium Oxide surfaces: a quantitative study of the influence of polymer interfacial architecture on resistance to protein adsorption by ToF-SIMS and in situ OWLSu OWLS. Langmuir, 2003. 19(22): p. 9216-9225.

40. Vonarbourg, A., Passirani, C., Saulnier, P., and Benoit, J.P., Parameters influencing the stealthiness of colloidal drug delivery systems. Biomaterials, 2006. 27(24): p. 4356-4373.

CHAPTER I 45

41. Conner, S.D., and Schmid, S.L., Regulated portals of entry into the cell. Nature (London, United Kingdom), 2003. 422(6927): p. 37-44.

42. Sinha, V.R., Aggarwal, A., and Trehan, A., Biodegradable PEGylated microspheres and nanospheres. American Journal of Drug Delivery, 2004. 2(3): p. 157-171.

43. Moghimi, S.M., and Szebeni, J., Stealth liposomes and long circulating nanoparticles: critical issues in pharmacokinetics, opsonization and protein-binding properties. Prog. Lipid Res., 2003. 42(6): p. 463-478.

44. Kataoka, K., Harada, A., and Nagasaki, Y., Block copolymer micelles for drug delivery: design, characterization and biological significance. Adv. Drug Deliv. Rev., 2001. 47(1): p. 113-131.

45. Torchilin, V.P., Targeted polymeric micelles for delivery of poorly soluble drugs. Cell. Mol. Life Sci., 2004. 61(19-20): p. 2549-2559.

46. Bhadra, D., Bhadra, S., Jain, P., and Jain, N.K., Pegnology: a review of PEG-ylated systems. Pharmazie, 2002. 57(1): p. 5-29.

47. Owens, D.E., 3rd, and Peppas, N.A., Opsonization, biodistribution, and pharmacokinetics of polymeric nanoparticles. Int. J. Pharm., 2006. 307(1): p. 93-102.

48. Zhang, X., Chen, K., Xie, K., and Long, Y., Self-assembly of monodisperse polymer microspheres from PPQ-b-PEG rod-coil block copolymers in selective solvents. Science in China, Series B: Chemistry, 2005. 48(1): p. 42-48.

49. Jia, W.J., Liu, J.G., Zhang, Y.D., Wang, J.W., Wang, J., Sun, C.Y., Pan, Y.F., Yu, C., Li, G.Z., Zhang, H., Gou, M.L., Tu, M.J., and Qian, Z.Y., Preparation, characterization, and optimization of pancreas-targeted 5-Fu loaded magnetic bovine serum albumin microspheres. J. Drug Target., 2007. 15(2): p. 140-145.

50. Castellanos, I.J., Flores, G., and Griebenow, K., Effect of the molecular weight of poly(ethylene glycol) used as emulsifier on a-chymotrypsin stability upon encapsulation in PLGA microspheres. J. Pharm. Pharmacol., 2005. 57(10): p. 1261-1269.

46 CHAPTER I

51. McQueen, C., Silvia, A., Lai, P.K., and Amiji, M., Surface and blood interaction properties of poly(ethylene oxide)-modified chitosan microspheres. S.T.P. Pharma Sciences, 2000. 10(1): p. 95-100.

52. Lacasse, F.X., Filion, M.C., Phillips, N.C., Escher, E., McMullen, J.N., and Hildgen, P., Influence of surface properties at biodegradable microsphere surfaces: effects on plasma protein adsorption and phagocytosis. Pharm. Res., 1998. 15(2): p. 312-317.

53. Piao, M.G., Kim, J.-H., Kim, J.O., Lyoo, W.S., Lee, M.H., Yong, C.S., and Choi, H.-G., Enhanced Oral Bioavailability of Piroxicam in Rats by Hyaluronate Microspheres. Drug Dev. Ind. Pharm., 2007. 33(4): p. 485-491.

54. Rokhade, A.P., Patil, S.A., Belhekar, A.A., Halligudi, S.B., and Aminabhavi, T.M., Preparation and evaluation of cellulose acetate butyrate and poly(ethylene oxide) blend microspheres for gastroretentive floating delivery of repaglinide. J. Appl. Polym. Sci., 2007. 105(5): p. 2764-2771.

55. Chen, F.-m., Wu, Z.-f., Sun, H.-h., Wu, H., Xin, S.-n., Wang, Q.-t., Dong, G.-y., Ma, Z.-w., Huang, S., Zhang, Y.-j., and Jin, Y., Release of bioactive BMP from dextran-derived microspheres: A novel delivery concept. International Journal of Pharmaceutics, 2006. 307(1): p. 23-32.

56. Ahmad, H., and Tauer, K., Preparation of monodisperse micron-sized PEG-coated polymer particles utilizing different methods. Prog. Colloid Polym. Sci., 2004. 124: p. 107-111.

57. Luo, X., Qiu, D., He, B., Wang, L., and Luo, J., Biodegradable heparin-loaded microspheres: carrier molecular composition and microsphere structure. Macromolecular Bioscience, 2006. 6(5): p. 373-381.

58. Zhou, S., Liao, X., Li, X., Deng, X., and Li, H., Poly-d,l-lactide-co-poly(ethylene glycol) microspheres as potential vaccine delivery systems. J. Controlled Release, 2003. 86(2-3): p. 195-205.

59. Kim, J.-H., and Bae, Y.H., Albumin loaded microsphere of amphiphilic poly(ethylene glycol)/ poly(a-ester) multiblock copolymer. Eur. J. Pharm. Sci., 2004. 23(3): p. 245-251.

CHAPTER I 47

60. Makino, K., Umetsu, M., Goto, Y., Nakayama, A., Suhara, T., Tsujii, J., Kikuchi, A., Ohshima, H., Sakurai, Y., and Okano, T., Interaction between charged soft microcapsules and red blood cells: effects of PEGylation of microcapsule membranes upon their surface properties. Colloids Surf. B, 1999. 13(6): p. 287-297.

61. Lück, M., Pistel, K.-F., Li, Y.-X., Blunk, T., Müller, R.H., and Kissel, T., Plasma protein adsorption on biodegradable microspheres consisting of poly(d,l-lactide-co-glycolide), poly(l-lactide) or ABA triblock copolymers containing poly(oxyethylene) Influence of production method and polymer composition. J. Controlled Release, 1998. 55(2,3): p. 107-120.

62. Hao, J., Whitaker, M.J., Serhatkulu, G., Shakesheff, K.M., and Howdle, S.M., Supercritical fluid assisted melting of poly(ethylene glycol): a new solvent-free route to microparticles. J. Mater. Chem., 2005. 15(11): p. 1148-1153.

63. He, W., Suo, Q., Hong, H., Shan, A., Li, C., Huang, Y., Li, Y., and Zhu, M., Production of natural carotene-dispersed polymer microparticles by SEDS-PA co-precipitation. J. Mater. Sci., 2007. 42(10): p. 3495-3501.

64. Nalawade, S.P., Picchioni, F., and Janssen, L.P.B.M., Batch production of micron size particles from poly(ethylene glycol) using supercritical CO2 as a processing solvent. Chem. Eng. Sci., 2007. 62(6): p. 1712-1720.

65. Hamidi, M., Azadi, A., and Rafiei, P., Pharmacokinetic consequences of pegylation. Drug Deliv, 2006. 13(6): p. 399-409.

66. Delgado, C., Francis, G.E., and Fisher, D., The uses and properties of PEG-linked proteins. Crit. Rev. Ther. Drug Carrier Syst., 1992. 9(3-4): p. 249-304.

67. Nagasaki, Y., Kobayashi, H., Katsuyama, Y., Jomura, T., and Sakura, T., Enhanced immunoresponse of antibody/mixed-PEG co-immobilized surface construction of high-performance immunomagnetic ELISA system. J. Colloid Interface Sci., 2007. 309(2): p. 524-530.

68. Madani, F., Bessodes, M., Lakrouf, A., Vauthier, C., Scherman, D., and Chaumeil, J.-C., PEGylation of microspheres for therapeutic

48 CHAPTER I

embolization: preparation, characterization and biological performance evaluation. Biomaterials, 2007. 28(6): p. 1198-1208.

69. Valentine, M.T., Perlman, Z.E., Gardel, M.L., Shin, J.H., Matsudaira, P., Mitchison, T.J., and Weitz, D.A., Colloid surface chemistry critically affects multiple particle tracking measurements of biomaterials. Biophys. J., 2004. 86(6): p. 4004-4014.

70. Meng, F., Engbers, G.H.M., and Feijen, J., Polyethylene glycol-grafted polystyrene particles. J. Biomed. Mater. Res., Part A, 2004. 70A(1): p. 49-58.

71. Meng, F., Engbers, G.H.M., Gessner, A., Mueller, R.H., and Feijen, J., Pegylated polystyrene particles as a model system for artificial cells. J. Biomed. Mater. Res., Part A, 2004. 70A(1): p. 97-106.

72. Dalsin, J.L., Lin, L., Tosatti, S., Voros, J., Textor, M., and Messersmith, P.B., Protein resistance of titanium oxide surfaces modified by biologically inspired mPEG-DOPA. Langmuir, 2005. 21(2): p. 640-646.

73. Caldwell, K.D., Surface modifications with adsorbed poly(ethylene oxide)-based block copolymers. Physical characteristics and biological use. ACS Symp. Ser., 1997. 680(Poly(ethylene glycol)): p. 400-419.

74. Jackson, J.K., Springate, C.M.K., Hunter, W.L., and Burt, H.M., Neutrophil activation by plasma opsonized polymeric microspheres: inhibitory effect of Pluronic F127. Biomaterials, 2000. 21(14): p. 1483-1491.

75. Rokhade, A.P., Shelke, N.B., Patil, S.A., and Aminabhavi, T.M., Novel hydrogel microspheres of chitosan and pluronic F-127 for controlled release of 5-fluorouracil. J. Microencapsul., 2007. 24(3): p. 274-288.

76. Huang, N.-P., Michel, R., Vörös, J., Textor, M., Hofer, R., Rossi, A., Elbert, D.L., Hubbell, J.A., and Spencer, N.D., Poly(L-lysine)-g-poly(ethylene glycol) Layers on Metal Oxide Surfaces: Surface-Analytical Characterization and Resistance to Serum and Fibrinogen Adsorption. Langmuir, 2001. 17(2): p. 489-498.

77. Feuz, L., Leermakers, F.A.M., Textor, M., and Borisov, O., Bending Rigidity and Induced Persistence Length of Molecular Bottle Brushes: A

CHAPTER I 49

Self-Consistent-Field Theory. Macromolecules, 2005. 38(21): p. 8891-8901.

78. Wattendorf, U., Koch, M.C., Walter, E., Vörös, J., Textor, M., and Merkle, H.P., Phagocytosis of poly(L-lysine)-graft-poly(ethylene glycol) coated microspheres by antigen presenting cells: Impact of grafting ratio and poly(ethylene glycol) chain length on cellular recognition. Biointerphases, 2006. 1(4): p. 123-133.

79. Wattendorf, U., Kreft, O., Textor, M., Sukhorukov, G.B., and Merkle, H.P., Stable Stealth Function for Hollow Polyelectrolyte Microcapsules through a Polyethylene Glycol Grafted Polyelectrolyte Adlayer. Biomacromolecules, accepted.

80. Müller, M., Vörös, J., Csúcs, G., Walter, E., Danuser, G., Merkle, H.P., Spencer, N.D., and Textor, M., Surface modification of PLGA microspheres. J. Biomed. Mater. Res., Part A, 2003. 66A(1): p. 55-61.

81. VandeVondele, S., Vörös, J., and Hubbell, J.A., RGD-grafted poly-L-lysine-graft-(polyethylene glycol) copolymers block non-specific protein adsorption while promoting cell adhesion. Biotechnol. Bioeng., 2003. 82(7): p. 784-790.

82. Huang, N.-P., Vörös, J., De Paul, S.M., Textor, M., and Spencer, N.D., Biotin-derivatized poly(L-lysine)-g-poly(ethylene glycol): A novel polymeric interface for bioaffinity sensing. Langmuir, 2002. 18(1): p. 220-230.

83. Tosatti, S., De Paul, S.M., Askendal, A., VandeVondele, S., Hubbell, J.A., Tengvall, P., and Textor, M., Peptide functionalized poly(L-lysine)-g-poly(ethylene glycol) on titanium: resistance to protein adsorption in full heparinized human blood plasma. Biomaterials, 2003. 24(27): p. 4949-4958.

84. Wang, R., Zhang, Y., Ma, G., and Su, Z., Modification of poly(glycidyl methacrylate-divinylbenzene) porous microspheres with polyethylene glycol and their adsorption property of protein. Colloids Surf. B, 2006. 51(1): p. 93-99.

50 CHAPTER I

85. Illum, L., Jacobsen, L.O., Muller, R.H., Mak, E., and Davis, S.S., Surface characteristics and the interaction of colloidal particles with mouse peritoneal macrophages. Biomaterials, 1987. 8(2): p. 113-117.

86. Fisher, N.G., Christiansen, J.P., Klibanov, A., Taylor, R.P., Kaul, S., and Lindner, J.R., Influence of microbubble surface charge on capillary transit and myocardial contrast enhancement. J. Am. Coll. Cardiol., 2002. 40(4): p. 811-819.

87. Gbadamosi, J.K., Hunter, A.C., and Moghimi, S.M., PEGylation of microspheres generates a heterogeneous population of particles with differential surface characteristics and biological performance. FEBS Lett., 2002. 532(3): p. 338-344.

88. Jones, B.G., Dickinson, P.A., Gumbleton, M., and Kellaway, I.W., The inhibition of phagocytosis of respirable microspheres by alveolar and peritoneal macrophages. International Journal of Pharmaceutics, 2002. 236(1-2): p. 65-79.

89. Illum, L., Hunneyball, I.M., and Davis, S.S., The Effect of Hydrophilic Coatings on the Uptake of Colloidal Particles by the Liver and by Peritoneal-Macrophages. International Journal of Pharmaceutics, 1986. 29(1): p. 53-65.

90. Kenausis, G.L., Vörös, J., Elbert, D.L., Huang, N., Hofer, R., Ruiz-Taylor, L., Textor, M., Hubbell, J.A., and Spencer, N.D., Poly(L-lysine)-g-Poly(ethylene glycol) Layers on Metal Oxide Surfaces: Attachment Mechanism and Effects of Polymer Architecture on Resistance to Protein Adsorption. J. Phys. Chem. B, 2000. 104(14): p. 3298-3309.

91. Heuberger, R., Sukhorukov, G., Vörös, J., Textor, M., and Möhwald, H., Biofunctional polyelectrolyte multilayers and microcapsules: control of non-specific and bio-specific protein adsorption. Adv. Funct. Mater., 2005. 15(3): p. 357-366.

92. Faraasen, S., Vörös, J., Csúcs, G., Textor, M., Merkle, H.P., and Walter, E., Ligand-Specific Targeting of Microspheres to Phagocytes by Surface Modification with Poly(L-Lysine)-Grafted Poly(Ethylene Glycol) Conjugate. Pharm. Res., 2003. 20(2): p. 237-246.

CHAPTER I 51

93. Vyas, S.P., Singh, A., and Sihorkar, V., Ligand-receptor-mediated drug delivery: an emerging paradigm in cellular drug targeting. Crit. Rev. Ther. Drug Carrier Syst., 2001. 18(1): p. 1-76.

94. Kim, D.H., Klibanov, A.L., and Needham, D., The influence of tiered layers of surface-grafted poly(ethylene glycol) on receptor-ligand-mediated adhesion between phospholipid monolayer-stabilized microbubbles and coated class beads. Langmuir, 2000. 16(6): p. 2808-2817.

95. Raichur, A.M., Voeroes, J., Textor, M., and Fery, A., Adhesion of Polyelectrolyte Microcapsules through Biotin-Streptavidin Specific Interaction. Biomacromolecules, 2006. 7(8): p. 2331-2336.

96. Le Cabec, V., Emorine, L.J., Toesca, I., Cougoule, C., and Maridonneau-Parini, I., The human macrophage mannose receptor is not a professional phagocytic receptor. J. Leukoc. Biol., 2005. 77(6): p. 934-943.

97. Takalkar, A.M., Klibanov, A.L., Rychak, J.J., Lindner, J.R., and Ley, K., Binding and detachment dynamics of microbubbles targeted to P-selectin under controlled shear flow. J. Controlled Release, 2004. 96(3): p. 473-482.

98. Klibanov, A.L., Rychak, J.J., Yang, W.C., Alikhani, S., Li, B., Acton, S., Lindner, J.R., Ley, K., and Kaul, S., Targeted ultrasound contrast agent for molecular imaging of inflammation in high-shear flow. Contrast Media Mol I, 2006. 1(6): p. 259-266.

99. Sakhalkar, H.S., Hanes, J., Fu, J., Benavides, U., Malgor, R., Borruso, C.L., Kohn, L.D., Kurjiaka, D.T., and Goetz, D.J., Enhanced adhesion of ligand-conjugated biodegradable particles to colitic venules. FASEB J., 2005. 19(7): p. 792-794.

100. Sakhalkar, H.S., Dalal, M.K., Salem, A.K., Ansari, R., Fu, J., Kiani, M.F., Kurjiaka, D.T., Hanes, J., Shakesheff, K.M., and Goetz, D.J., Leukocyte-inspired biodegradable particles that selectively and avidly adhere to inflamed endothelium in vitro and in vivo. Proc. Natl. Acad. Sci. U. S. A., 2003. 100(26): p. 15895-15900.

52 CHAPTER I

CHAPTER II 53

CHAPTER II

Phagocytosis of PLL-g-PEG Coated Microspheres by Antigen Presenting Cells: Impact of Grafting Ratio and

PEG Chain Length on Cellular Recognition

Uta Wattendorf,a Mirabai C. Koch,a Elke Walter,a Janos Vörös,b,1 Marcus Textor,b and Hans P. Merkle

a

a Institut for Pharmaceutical Sciences, ETH Zurich, 8093 Zurich, Switzerland b Laboratory for Surface Science and Technology, ETH Zurich, 8093 Zurich,

Switzerland 1 Present address: Laboratory for Biosensors and Bioelectronics, Swiss Federal

Institute of Technology Zurich (ETH Zurich), 8092 Zurich, Switzerland. Biointerphases, 2006. 1(4): pp. 123-133.

54 CHAPTER II

ABSTRACT

Microparticulate carrier systems have significant potential for antigen delivery. We studied how microspheres coated with the polycationic copolymer poly(L-lysine)-graft-poly(ethylene glycol) (PLL-g-PEG) can be protected against unspecific phagocytosis by antigen presenting cells, a prerequisite for selective targeting of phagocytic receptors. For this aim we explored the influence of PLL-g-PEG architecture on recognition of coated microspheres by antigen presenting cells with regard to both grafting ratio and molecular weight of the grafted PEG chains. Carboxylated polystyrene microspheres (5 μm) were coated with a small library of PLL-g-PEG polymers with PLL-backbones of 20 kDa, grafting ratios from 2 to 20, and PEG side chains of 1 to 5 kDa. The coated microspheres were characterized by their ζ-potential and resistance to IgG adsorption. Phagocytosis of these microspheres by human monocyte derived dendritic cells (DC) and macrophages (MΦ) was quantified by phase contrast microscopy and by analysis of the cells’ side scattering in a flow cytometer. Generally, increasing grafting ratios impaired the protein resistance of coated microspheres, leading to higher phagocytosis rates. For DC, long PEG chains of 5 kDa decreased the phagocytosis of coated microspheres even in the case of considerable IgG adsorption. In addition, preferential adsorption of dysopsonins is discussed as another factor for decreased phagocytosis rates. For comparison, we studied the cellular adhesion of DC and MΦ to PLL-g-PEG coated microscopy slides. Remarkably, DC and MΦ were found to adhere to relatively protein-resistant PLL-g-PEG adlayers, whereas phagocytosis of microspheres coated with the same copolymers was inefficient. Overall, PLL(20)-[3.5]-PEG(2) was identified as the optimal copolymer to ensure resistance to both phagocytosis and cell adhesion. Finally, we studied coatings made from binary mixtures of PLL-g-PEG type copolymers that led to microspheres with combined properties. This enables future studies on cell targeting with ligand modified copolymers.

CHAPTER II 55

INTRODUCTION

Microparticulate carrier systems such as liposomes, microspheres or microcapsules are of considerable interest in drug delivery, vaccination and as diagnostic agents [1-3]. Notably, microparticulate antigen delivery systems have been shown to elicit cytotoxic T lymphocyte (CTL) immunity as well as antibody and helper T cell responses [4-7] through phagocytosis by antigen presenting cells (APC) and subsequent presentation of the antigen on MHC I (i.e. crosspresentation) or MHC II [8].

It is generally recognized that surface modifications of microparticulate antigen delivery systems with suitable peptides, proteins, oligosaccharides or nucleic acids are of potential interest to control their first encounter with APC and could thus affect the type of immune response that will be elicited [9-12]. In order to specifically target APC, ligand mediated interactions with suitable receptors would be necessary. In vivo, however, injected particulate drug carriers are instantly recognized as foreign material and efficiently removed via ligand-unspecific phagocytosis by the mononuclear phagocytic system (MPS), which comprises phagocytic cells such as dendritic cells (DC) and macrophages (MΦ), among others [13]. The extent and rate of phagocytic clearance is dependant on size and surface properties. For instance, phagocytosis is enhanced by charged or hydrophobic surfaces [14-16]. The adsorption of opsonic proteins onto the surface of hydrophobic particles promotes phagocytosis [17-19]. Therefore, receptor specific targeting requires a microparticulate carrier system that is able to escape the MPS due to a surface modification that renders them resistant to opsonization. Only under this condition can the decoration of the particles´ surface with specific ligands lead to a specific targeting of APC via surface receptors.

There are various approaches to prevent unwanted serum protein adsorption and cellular recognition [20-22]. Most strategies rely on the immobilization of poly(ethylene glycol) (PEG) coatings, either through covalent or non-covalent PEGylation of the surface. Due to its low toxicity and immunogenicity, PEG is highly suitable for biomedical applications [23, 24]. PEG is an uncharged, hydrophilic polymer that is soluble in water as well as in

56 CHAPTER II

many organic solvents. In an aqueous environment, its extensive hydration, good conformational flexibility and high chain mobility causes a steric exclusion effect [23]. Besides steric stabilization, hydration due to PEG-coupled, structured water is believed to contribute to the protein resistance of PEGylated surfaces [25, 26]. As demonstrated theoretically [26-29] and experimentally [30-35] protein resistance of PEG-based coatings depends on both chain length and density, the product of which determines the thickness of the PEG adlayer. Protein resistance becomes increasingly efficient the more the PEG chains overlap [33], and longer chains have been found to have the ability to fill in the gaps between less densely grafted PEG chains [34]. Theoretical studies have revealed two modes of protein adsorption [29]: (i) primary adsorption at the PEG-substrate interface and (ii) secondary adsorption at the top of the PEG brush. Small proteins may penetrate the PEG brush and adsorb onto the underlying surface. Increasing PEG chain density is expected to reduce the degree of primary adsorption. Resistance to larger proteins, which may undergo secondary adsorption, requires a sufficiently thick PEG layer to screen long range, e.g. electrostatic protein-substrate interactions.

A convenient method for the non-covalent immobilization of PEG on negatively charged surfaces is the adsorption of PLL-g-PEG copolymers, consisting of a polycationic poly(L-lysine) backbone whose side chain amino groups are partly grafted with PEG chains. Such polymers were previously found to spontaneously adsorb from aqueous solution onto negatively charged metal oxide surfaces [36, 37]. Due to its positive charge at physiological pH, the PLL backbone adsorbs strongly onto negatively charged substrates through electrostatic interaction, leading to a monolayer with PEG chains stretched out perpendicularly to the surface [36]. Simulations using a self-consistent-field approach revealed the PEG end-segment distribution to be farther displaced from the the PLL backbone for decreasing grafting ratios (defined as the ratio of the number of L-lysine monomers to the number of PEG side chains) as a consequence of increased stretch-out of the PEG chains due to increased PEG-PEG steric repulsion [38]. In addition, the PEG side chains can be functionalized with bioligands while retaining resistance to unspecific protein adsorption [39-42]. Appropriate choice of polymer architecture and assembly conditions

CHAPTER II 57

renders such coated surfaces protein-resistant even in the presence of full human serum [35, 36]. Optimum protein resistance of PLL-g-PEG adlayers on metal oxide surfaces was observed for architectures with a 20 kDa PLL backbone, 2 kDa PEG side chains, and a grafting ratio of approximately 3.5 [35, 37].

PEGylation of nano- and microparticles does not only reduce plasma protein adsorption but also inhibits the nonspecific internalization by phagocytes [22, 43-45]. Recent publications report the modification of microparticles with PLL-g-PEG and ligand modified PLL-g-PEG polymers that showed excellent protein resistance [39, 46]. In addition, it was shown for DC and MΦ cell cultures that phagocytosis of such particles could be abolished with PLL-g-PEG coatings [47]. However, these studies were performed only with the aforementioned “optimum” architecture, while no evaluation of the influence of polymer structure was reported so far.

In this work, we explored the influence of PLL-g-PEG architecture on phagocyte recognition of microspheres with regard to both grafting ratio and molecular weight of the grafted PEG chains. PLL-g-PEG coated polystyrene microspheres were first characterized by their ζ-potential and degree of IgG adsorption. In a second step, phagocytosis of PLL-g-PEG coated microspheres was examined in vitro in DC and MΦ by a microscopy-based counting protocol and by analysis of the cells’ light scattering characteristics. In a second part of the study we investigated the effect of PLL-g-PEG coatings on the spreading of DC and MΦ on flat surfaces and correlated the effect of surface chemistry on phagocytosis with cell adhesion.

MATERIALS AND METHODS Materials

As microparticulates we used carboxylated polystyrene (PS) microspheres (5 μm diameter) from Micromod Partikeltechnologie GmbH, Rostock, Germany. PLL-g-PEG copolymers were synthesized as outlined by Pasche et al.[35] The notation used for PLL(x)-g[y]-PEG(z) copolymers indicates the

58 CHAPTER II

average molecular weights of PLL (x) and PEG (z) and the grafting ratio g (y) which specifies the total number of lysine monomers divided by the number of PEG side chains as determined by 1H-NMR. The PLL-g-PEG copolymers consisted of a PLL-backbone of 20 kDa with grafted PEG chains of 1, 2 or 5 kDa and a grafting ratio of 6.5 (for 1 kDa), 2.2, 3.5, 5.7, 10.1 (for 2 kDa) and 18.7 (for 5 kDa). All chemicals used in this study were of analytical grade (from Fluka, Buchs, Switzerland) unless otherwise specified. Ultrapure water (NANOpure DiamondTM, Skan AG, Allschwil, Switzerland) was used for buffer preparation.

DC and MΦ Cell Culture

DC and MΦ were obtained from human peripheral blood monocytes according to Sallusto et al. [48] Briefly, peripheral blood monocytes were isolated from buffy coats (Bloodbank Zurich, Zurich, Switzerland) by density gradient centrifugation of Ficoll-PaqueTM PLUS (Amersham Biosciences, Uppsala, Sweden). Isolated peripheral blood monocytes were washed in PBS (Sigma, Buchs, Switzerland) and resuspended in RPMI 1640 + L-glutamine (Invitrogen, Basel, Switzerland) supplemented with 10% heat-inactivated (pooled) human serum (Bloodbank Zurich, Zurich, Switzerland), 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen, Basel, Switzerland) and then allowed to adhere for 2 h in culture flasks (25 cm2). Nonadherent cells were removed whereas adherent cells were further cultured in RPMI medium supplemented with 5% heat-inactivated (pooled) human serum, 100 U/ml penicillin and 100 μg/ml streptomycin in the presence of 1000 IU/ml interleukin-4 (Sigma, Buchs, Switzerland) and 50 ng/ml human granulocyte-macrophage colony stimulating factor (R+D Systems Europe Ltd., Abingdon, UK) to obtain DC. MΦ were obtained without additional supplements [14]. These media will further be referred to as supplemented RPMI media. Cultures of DC and MΦ were kept at 37°C in 5% CO2-humidified atmosphere. For phagocytosis studies, cells were mechanically removed from the flasks after 24 h and reseeded into 24-well plates (300,000-350,000 cells per well).

CHAPTER II 59

Surface antigen expression of DC and MΦ obtained by this protocol were previously validated by flow cytometry (FACScan, Becton Dickinson, Franklin Lakes, NJ) [47]. Cells were incubated for 45 min at 4°C with the following antibodies: CD11b (Mac-1, Clone ICRF44, Pharmingen, San Diego, CA, USA), CD14 (Clone UCM-1, Sigma), CD83 (Clone HB15e, Pharmingen), CD86 (Clone IT2.2, Pharmingen), mouse isotype control antibody IgG2a (UPC-10, Sigma) and subsequently (after washing) with the secondary antibody (anti-mouse IgG R-Phycoerythrin conjugate, Sigma). For further identification, the cells were also challenged with lipopolysaccharide (1 μg/ml, Escherichia coli 055:B5, Sigma) 48 h before the antibody labeling, which causes maturation of DC only. More than 90 % of the cells were identified as DC, and in MΦ cultures more than 90% of the cells were DC14+ and CD83-, even in the presence of lipopolysaccharide [47].

Adhesion Studies

Adhesion of DC and MΦ on modified surfaces was tested on glass coverslips coated with the respective PLL-g-PEG polymers. To rule out variances originating from donor-to-donor variability, experiments on each cell batch were always performed in parallel with all six polymers simultaneously. Before coating, 2-propanol cleaned glass coverslips (12 mm diameter) were oxygen-plasma treated in a Plasma Cleaner/Sterilizer PDC-32G instrument (Harrick Scientific Corporation, Ossining, NY) for 2 min. Coating with PLL-g-PEG was performed in 24-well plates under aseptic conditions with 0.1 mg/ml of the respective polymers dissolved in 10 mM HEPES buffer, pH 7.4, for 30 min on a shaker. All solutions were sterile filtered through a 0.2 μm membrane filter before use. Control coverslips were incubated in HEPES buffer only. After coating, the coverslips were rinsed twice for 5 min with phosphate buffered saline (PBS, Sigma, Buchs, Switzerland) and once with supplemented RPMI medium. Either DC or MΦ incubated for 1 week in cell culture flasks were randomly added to each well (200,000-300,000 cells in 600 μl of cell medium) containing either coated or control coverslips. After 4 hours cells were washed twice (DC) or three times (MΦ) with supplemented RPMI medium, and

60 CHAPTER II

cellular adhesion was checked by microscopy (Axiovert 35, Zeiss, Germany). 5 - 6 randomly acquired images were taken at 10-fold magnification using a CCD camera (CF 8/1 DXC, Kappa, Gleichen, Germany).

Coating of Microspheres

PS microspheres (5 mg/ml) were dispersed under aseptic conditions in sterile filtered HEPES buffer (10 mM, pH 7.4). The relatively low ionic strength of the buffer was chosen to avoid loop formation of the PLL backbone upon adsorption of the polymer to the microsphere surface [49]. For microsphere coating the dispersions were mixed at equal volumes with the respective PLL-g-PEG or mixed PLL-g-PEG solutions (1mg/ml in sterile filtered HEPES buffer) and incubated under gentle mixing for 15 min at room temperature. Surface-coated microspheres were centrifuged (5 min, 2655 x g, Eppendorf, Centrifuge 5417R, Eppendorf-Netheler-Hinz GmbH, Hamburg, Germany) and redispersed in half of the previous volume in sterile filtered HEPES buffer.

ζ-Potential of PLL-g-PEG Coated Microspheres

PLL-g-PEG coated and uncoated PS microspheres (65 μg/ml) were dispersed in filtered HEPES buffer (10 mM, pH 7.4), HEPES buffer supplemented with 150 mM NaCl, or RPMI medium. Measurements were performed on a Zetasizer 3000 HS (Malvern Instruments, Worcestershire, UK; five measurements per sample).

Protein Resistance of PLL-g-PEG Coated Microspheres

To determine the protein resistance of surface coated PS microspheres we used a fluorescence labelled immunoglobulin, TxRed-goat anti-rabbit IgG(H+L) (Molecular Probes Inc., Eugene, OR, USA). The protein solution was diluted in filtered HEPES buffer (10 mM, pH 7.4) to yield a final concentration of 150 μg/ml, and added to an equal volume of microsphere dispersion (1 mg/ml in filtered HEPES buffer). After 45 minutes of incubation in the dark, samples were centrifuged (10 min, 2655 g, Eppendorf, Centrifuge 5417R, Eppendorf-

CHAPTER II 61

Netheler-Hinz GmbH, Hamburg, Germany), washed and redispersed in filtered HEPES buffer at 250 μg/ml. Fluorescence of the microspheres as a consequence of antibody binding was determined in a flow cytometer (FACSCalibur from Becton Dickinson, Franklin Lakes, NJ). For data analysis, median fluorescence values were calculated using the Cytomation Summit software (v3.1, Cytomation Inc., Fort Collins, CO, USA). Relevant quenching of IgG-TxRed fluorescence on the particles was excluded by an incubation experiment with concentrations up to 160 μg/ml, resulting in an approximately linear relationship of fluorescence intensity and IgG-TxRed concentration.

Phagocytosis Studies

Studies were performed with both DC and MΦ. To rule out variances originating from donor-to-donor variability, experiments on each cell batch were always performed in parallel with all six polymers simultaneously. MΦ cultured in 24-well plates were washed once with RPMI medium and then covered with 600 μl of supplemented RPMI medium. Because DC adhered less strongly to the cell culture plates, they were not washed but directly exposed to the microspheres. To exclude compromised viability, we stained the DC with annexin-V-APC (BD Biosciences, Basel, Switzerland) and propidium iodide (BD Biosciences, Basel, Switzerland) according to the manufacturer´s protocol. Analysis by flow cytometry revealed that more than 80 % of the cells turned out to be viable upon double staining, and less than 15 % were positive to propidium iodide.

For the assessment of phagocytosis, microspheres dispersed in filtered HEPES buffer were added to either DC or MΦ cultures and incubated at 37°C in 5% CO2-humidified atmosphere for 4 h. The medium was carefully exchanged to remove non-adherent cells. Cells were immediately inspected for phagocytosis by light microscopy (phase contrast; Axiovert 35, Zeiss, Germany). Seven to eleven randomly acquired images were taken from three samples per experiment at 20-fold magnification using a CCD camera (CF 8/1 DXC, Kappa, Gleichen, Germany). The number of phagocytosed microspheres per cell was counted manually. To account for the variance of phagocytotic

62 CHAPTER II

activity of cells from different donors, numbers of microspheres per cell were normalized to PS positive controls.

Following microscopic analysis, the cells were mechanically recovered in a test tube and additionally evaluated by flow cytometry (FACScan, Becton Dickinson, Franklin Lakes, NJ) [50]. For data analysis, mean SSC values were calculated with Cytomation Summit software (v3.1, Cytomation Inc., Fort Collins, CO, USA) [51].

Significance Testing

For evaluation of statistical significance, samples were evaluated by ANOVA followed by a post-hoc assessment using the Tukey HSD method. Differences were considered significant when equal or less than p=0.05 using SPSS software (SPSS Inc., Chicago, IL).

RESULTS ζ-Potential of Coated Microspheres

We measured the ζ-potential to monitor the coatings of PLL-g-PEG copolymers on carboxylated polystyrene microsphere surfaces. Uncoated microspheres showed a strongly negative ζ-potential (-45 ± 5 mV in 10 mM HEPES-buffer). All coatings gave rise to a slightly positive ζ-potential of the microspheres when measured in 10 mM HEPES buffer (pH = 7.4), thus providing evidence for effective coating of the microspheres (Table 1). While the ζ-potentials of most of the polymers were around 5 mV, PLL(20)-[2.2]-PEG(2) and PLL(20)-[6.5]-PEG(1) yielded slightly different values of 1 ± 2 mV and 14 ± 6 mV, respectively. Only the latter value, however, was statistically different from the others. When the ζ-potential of coated microspheres was determined at the higher ionic strength used in the cell assays, the data were no longer statistically significant: The ζ-potentials of coated microspheres were all 5 ± 2 mV when measured in 10 mM HEPES supplemented with 150 mM

CHAPTER II 63

sodium chloride (pH = 7.4) or in cell culture medium (RPMI medium supplemented with 5 % human serum).

Resistance of Coated Microspheres to IgG adsorption

The protein resistance of the PLL-g-PEG coated microspheres was assessed by incubation with a goat-anti rabbit IgG(H+L) serving as a model protein (Table 1 and Fig. 1). For polymers with PEG chains of 2 kDa we found the adsorption of IgG to increase linearly with the grafting ratio. Very low amounts of IgG adsorbed on coatings of polymers with the lowest grafting ratios (< 4 % of the amount detected on uncoated polystyrene microspheres), while a significant amount of protein adsorbed at the highest grafting ratio (18.2 ± 6.5 %). Interestingly, the intermediate grafting ratio (g = 5.7) that had been qualified as protein resistant in the literature [35] showed considerable protein adsorption.

At constant ethylene-glycol-monomer surface densities no significant difference in IgG adsorption was found between a PEG chain length of 1 kDa (16.9 ± 2.5 %) and 2 kDa (18.2 ± 6.5 %). Interestingly, much lower adsorption of IgG was found with 5 kDa PEG chains (6.8 ± 0.4 %).

Table 1. Microsphere Characterization. ζ-potential of and IgG adsorption on PLL-g-PEG coated microspheres in 10 mM HEPES buffer. ζ-potential in mV (± SD); IgG adsorption in % of uncoated control microspheres as determined by flow cytometry (median of fluorescence ± SD).

64 CHAPTER II

Figure 1. Protein Resistance of PLL-g-PEG Coated PS Microspheres. Fluorescence of PLL-g-PEG coated microspheres after incubation with TxRed-goat anti-rabbit IgG(H+L) in 10 mM HEPES buffer and subsequent washing. Data represent median fluorescence values from flow cytometry normalized to uncoated control microspheres.

Figure 2. Phagocytosis of PLL-g-PEG Coated PS Microspheres. Phagocytosis of PLL-g-PEG coated microspheres by DC (black) and MΦ (grey). Numbers of internalized microspheres were counted after 4 hours of incubation and normalized to uncoated control microspheres.

CHAPTER II 65

Phagocytosis of PLL-g-PEG Coated PS Microspheres by DC and MΦ

To investigate whether PLL-g-PEG coatings would render microspheres resistant to phagocytosis, we coated PS microspheres (5 μm) and incubated them in DC or MΦ cultures, respectively. The extent of phagocytosis was first analyzed by microscopy (Table 2 and Fig. 2) [52]. As a reference we used uncoated microspheres which were efficiently internalized within the timeframe of the experiment (4 h) by about 70% of the cell population. The effect of the grafting ratio of PLL-g-PEG type copolymers was examined using four copolymers of the general architecture PLL(20)-g-PEG(2) with grafting ratios of 2.2, 3.5, 5.7 and 10.1, respectively. The first three have been reported to resist protein adsorption but not the latter [35]. As expected, the numbers of PLL(20)-[10.1]-PEG(2) coated microspheres that were subject to phagocytosis were close or equal to those for uncoated reference microspheres, while polymers classified by Pasche et al. [35] as protein resistant reduced the phagocytic uptake to about 10 to 20 %. Interestingly, the lowest numbers of phagocytosed microspheres (about 10%) were observed for g = 5.7 and increased for smaller grafting ratios, though without statistical significance. No difference in phagocytosis was found between DC and MΦ.

In addition to the effect of the grafting ratio, the influence of the PEG chain length was investigated as well. For this purpose, we chose polymers with PEG chains of 1, 2 and 5 kDa, which result in roughly the same ethylene-glycol-monomer surface density when assembled on a surface (as determined by Pasche et al., on Nb2O5 surfaces [35]), i.e. PLL(20)-[6.5]-PEG(1), PLL(20)-[10.1]-PEG(2) and PLL(20)-[18.7]-PEG(5). Equal to the g = 10.1 polymer, relatively high amounts of serum proteins were observed to adsorb to flat surfaces coated with the other two polymers [35]. Consistently, microspheres coated with any of the three polymers were internalized by MΦ about as efficiently as reference microspheres. However, DC and MΦ behaved differently: While microspheres coated with the 2 kDa PEG chain polymer were phagocytosed to a similar extent as uncoated reference microspheres, both longer (5 kDa) and shorter (1 kDa) PEG chains significantly (p < 0.01) impaired the microspheres´ internalization to about 50 and 20 %, respectively.

66 CHAPTER II

Ta

ble

2. A

dhes

ion

and

Phag

ocyt

osis

. Cel

l adh

esio

n (n

umbe

r of

adh

eren

t cel

ls)

on P

LL-g

-PEG

coa

ted

glas

s sl

ides

in %

of

unco

ated

gla

ss b

y m

icro

scop

y (m

ean

coun

ts ±

SD

); p

hago

cyto

sis

of P

LL-g

-PEG

coa

ted

mic

rosp

here

s de

term

ined

by

flow

cy

tom

etry

(m

ean

of S

SC ±

SD

) an

d by

mic

rosc

opy

(mea

n co

unts

± S

D)

in %

of

unco

ated

mic

rosp

here

s. Ev

alua

tion

afte

r 4

hour

s of i

ncub

atio

n.

CHAPTER II 67

Figure 3. Correlation of Flow Cytometry and Microscopy. Correlation of DC phagocytosis data obtained from flow cytometry (mean of SSC ± SD) vs. data obtained from phase contrast microscopy (mean counts ± SD). PS microspheres coated with PLL-g-PEG copolymers of different architectures were incubated with DC for 4 h. Samples were successively analysed by both techniques.

Quantification of Phagocytosis by Side Scatter Analysis

For the quantification of phagocytic efficiency we focused not only on phase contrast microscopy but also sought for a less time consuming technique. Therefore we successfully established a flow cytometric protocol that was based on the shift of the side scattering signal as a result of phagocytosis [51, 53, 54]. Because MΦ adhere very strongly to cell culture dishes and are difficult to remove without damaging them, only DC were used in these assays. Our results (Table 2 and Fig. 3) demonstrate an excellent correlation between the mean of the side scatter and the values obtained by microscopic analysis. Thus side scatter analysis by flow cytometry can be used as a convenient method to quantify phagocytosis, at least within the experimental framework of our study and for DC.

Influence of Polymer Mixtures on Phagocytosis of Coated Microspheres

In addition to varying the architecture of the PLL-g-PEG it is also possible to assemble the polymers from mixtures of two polymers, which may

68 CHAPTER II

result in microspheres with surface characteristics that lie either between those of the individual polymers or are dominated by one of them. As an example we investigated the phagocytosis of microspheres coated with varying ratios of the two polymers PLL(20)-[5.7]-PEG(2) and PLL(20)-[10.1]-PEG(2), the former being the most phagocytosis-resistant one among the polymers used in this study, while microspheres coated with the latter were internalized about as efficiently as uncoated reference microspheres. Up to a fraction of 40 % of the non-resistant polymer we observed a direct dependence of phagocytic efficiency and polymer proportion. Above that proportion phagocytosis was close to 100% of control (Fig. 4).

Adhesion of DC and MΦ on PLL-g-PEG Coated Glass Slides

To investigate the effect of polymer coatings on the adhesion of APC on flat surfaces, glass coverslips coated with the respective polymers were used as substrates and subsequently incubated with DC or MΦ for 4 hours prior evaluation by microscopy (Table 2 and Fig. 5). As a reference uncoated coverslips were used, to which both cell types showed a high affinity as indicated by good spreading and strong adhesion.

As expected for the group of 2 kDa PEG chains, the numbers of adherent cells on the polymer with the relatively high grafting ratio of 10.1 was close or equal to the number on the reference glass slides. Interestingly, although both g = 3.5 and 5.7 yielded significantly (p < 0.01) reduced numbers of adherent cells, only the polymer with a grafting ratio of 3.5 was found to completely abolish cellular adhesion for both cell types. Both the higher and the lower grafting ratios of 2.2 and 5.7 were found to be less resistant to cellular adhesion with a high variance between different samples and positions in case of g = 5.7.

Coatings of the non-protein resistant group of polymers of different PEG chain lengths were not expected to impede cellular adhesion to a great extent. Consistently, all polymers showed about the same numbers of cellular adhesion which was close to control values. Only the polymer with the short PEG chains (1 kDa) and the relatively low grafting ratio of g = 6.5 featured significantly (p < 0.05) less DC adhesion.

CHAPTER II 69

Figure 4. Phagocytosis of PS Microspheres Coated with Binary PLL-g-PEG Mixtures. PS microspheres coated with binary mixtures of PLL(20)-[5.7]-PEG(2) and PLL(20)-[10.1]-PEG(2) were incubated with DC. Phagocytosis was quantified by side scatter analysis (mean of SSC ± SD) after 4 hours of incubation and normalized to uncoated control microspheres.

Figure 5. Adhesion of PLL-g-PEG Coated Glass Slides. Adhesion of DC (black) and MΦ (grey) on PLL-g-PEG coated glass slides. Numbers of adherent cells were counted after four hours of incubation and normalized to uncoated control slides. Both cell types did not adhere on glass slides coated with PLL(20)-[3.5]-PEG(2).

70 CHAPTER II

DISCUSSION

Recognition and subsequent phagocytosis by APC represent important steps in the first encounter of microparticulate antigen delivery systems with the innate immune system. Unspecific recognition of microparticulates by APC is supported by prior adsorption of opsonizing serum proteins on their surface [17, 18]. With the aim of establishing design rules for PLL-g-PEG surface coatings in terms of their ability to shield microspheres from recognition by APC, we studied the impact of PLL-g-PEG copolymer architecture on protein resistance and phagocytosis of coated PS microspheres. In a first set of experiments, we examined four polymers of the general architecture PLL(20)-g-PEG(2) having grafting ratios g of 2.2, 3.5, 5.7 and 10.1. The first three polymers have previously been reported to provide protein resistant flat surfaces while the latter was not [35]. However, earlier work suggested that PLL-g-PEG copolymers with a grafting ratio of 5 to 6 were less protein resistant than those with the grafting ratio of 3.5 [35, 37]. In addition to the effect of the grafting ratio we were also interested in a potential effect of the PEG chain length. To this end, we chose polymers with PEG chains of 1, 2 and 5 kDa. As protein resistance on flat surfaces has previously been related to the amount of ethylene-glycol-monomer density per surface area rather than to PEG chain length or grafting ratio alone [35], we chose three polymers having about the same ethylene-glycol-monomer density per surface area, i.e. PLL(20)-[6.5]-PEG(1), PLL(20)-[10.1]-PEG(2) and PLL(20)-[18.7]-PEG(5). As mentioned above for the g = 10.1 copolymer, the protein resistance of flat surfaces coated with these polymers was relatively low [35].

To monitor the effects of PLL-g-PEG coatings on phagocytosis, we used carboxylated PS microspheres as core material. Microspheres of 5 μm in diameter were chosen which is within the optimum size range for phagocytosis by APC [7, 55]. Due to their negative surface charge, carboxylated PS microspheres adsorb PLL-g-PEG copolymers by electrostatic interaction resulting in a PEG corona [36, 47].

For the quantification of phagocytic efficiency we first focused on phase contrast microscopy. Counting numbers of phagocytosed microspheres per cell

CHAPTER II 71

from several statistically acquired images has the advantage of yielding absolute numbers without the need of further sample processing. However, manual acquisition of statistically significant cell counts by a counting protocol turned out to be extremely time consuming. Flow cytometry has been routinely used for fast and reliable quantification of fluorescently labelled microspheres [56]. A less common flow cytometric approach to monitor particle internalization takes advantage of the high sensitivity of the side scattering signal to changes in cell morphology as a consequence of phagocytosis [51, 53, 54], thus bypassing the need for fluorescent dyes. The direct use of fluorescently labelled microspheres was discarded because of an interference of the PLL-g-PEG coatings with the label, potentially due to inhomogeneities in the negative surface charge distribution on such spheres. Because MΦ adhere very strongly to cell culture dishes, we exclusively used DC to examine the feasibility to quantify particle internalization by evaluation of the side scattering signal. Our results demonstrate an excellent correlation between the mean of the side scatter and the values obtained by microscopic analysis. It is a general drawback of flow cytometry that arbitrary units are obtained instead of the absolute numbers of phagocytosed microspheres per cell. However, this is not necessarily a disadvantage as numbers can be normalized to unmodified control microspheres. In addition to being less time-consuming, the developed protocol also yields statistically more reliable data than microscopy, based on the several thousands of cells per sample that can be analyzed in the former case.

Characterization of Coated Microspheres

The coating process could be conveniently monitored by measuring the microspheres’ ζ-potential. Upon PLL-g-PEG adsorption the strongly negative surface charge of uncoated carboxylated polystyrene microspheres is counterbalanced by the positive charges of the PLL-backbone. Furthermore, remaining interfacial charges are shielded by ions within the PEG corona. Thus, coating of the particles is expected to yield a ζ-potential close to zero [46]. Experimentally, slightly positive values (5 ± 2 mV) were obtained for all coated microspheres when measured in solutions of physiological ionic strength. At

72 CHAPTER II

lower ionic strength (10 mM HEPES), a statistically significant difference was only observed for PLL(20)-[6.5]-PEG(1). The slightly higher ζ-potential obtained for this polymer (14 ± 6 mV) probably results from its shorter 1 kDa PEG chains which may be assumed to be less efficient in shielding the remaining positive interfacial charge than the 2 kDa or 5 kDa PEG chains of the other coatings. However, the measured increase in surface charge is small compared to formulations in other studies that observed an enhanced phagocytosis for positively charged particle formulations [14-16]. Hence, a major influence of the surface charge on the results of our cell adhesion or phagocytosis experiments is unlikely. Overall, our observations are consistent with the successful formation of a polymeric adlayer on PS microspheres whose surface is efficiently shielded by a PEG corona.

In principle, polymeric coatings can be characterized by the size increase of the coated microspheres. PLL-g-PEG, however, has been shown to form a monolayer when adsorbed to a negatively charged surface [36] with a thickness of less than 10 nm [57]. Therefore, the size increase during the coating of the microspheres was too small to be detected by light scattering. Moreover, the negligible size increase discards size as a factor for phagocytic efficiencies in our study.

In its physiological context, it is well established that phagocytosis of microspheres requires prior opsonization by immunoglobulins and small proteins of the complement system [8, 18]. Amount and nature of adsorbed proteins is believed to affect extent and rate of phagocytosis [13, 19, 43, 58]. To mimic opsonization in vitro we chose immunoglobulin G, goat-anti rabbit IgG(H+L), as model protein. Despite the variable charge and size of proteins present in full serum, IgG is an established model to check for resistance to serum protein adsorption [39, 45]. In particular, class G immunoglobulins are the principal “heat-stable” opsonic proteins [17-19]. In addition, opsonization with complement was found to be insufficient for phagocytosis unless the studied particles also contained IgG [59].

Hypothetically, a direct relationship between the previously published adsorption of whole serum on flat, polymer-coated surfaces [35] and our data of

CHAPTER II 73

IgG adsorption on microspheres coated with the same type of polymers could be expected. However, the number of available data pairs does not allow a statistically significant confirmation (Fig. 6). Deviations from such a direct relationship may be due to the different serum proteins used in both essays, i.e. IgG or whole serum. For example, in the case of surfaces with dense PEG chains (low [g] value), the very small gaps between the PEG chains would be expected to be accessible for the primary adsorption of very small proteins only. As well, a steric effect of long PEG chains (5 kDa) could be assumed. Indeed, this seems to be the case for the 5 kDa PEG copolymer used in this study. As its grafting ratio [18.7] is higher than that of any of the 2 kDa PEG copolymers [10.1], and IgG adsorption was found to increase with increasing grafting ratio, even higher amounts of IgG would be expected to adsorb on microspheres coated with the 5 kDa PEG copolymer. However, we found a reduced IgG adsorption, which thus can only be attributed to the longer PEG chains. This would be in accordance with a recent publication claiming that longer PEG chains (≥ 5 kDa) have the ability to fill in effectively the space between less densely spaced PEG chains [34]. In contrast to IgG with its rather large molecular size, smaller serum proteins may still be able to reach the surface, thus leading to an enhanced protein adsorption as documented in a previous study in serum [28, 35]. It is interesting to note that we found a linear increase in IgG adsorption with increasing grafting ratio, whereas previous studies [35, 37] led to ambiguous results for g = 5.7.

Cellular Recognition of Coated Microspheres: Phagocytosis

The major aim of our study was to investigate the capacity of PLL-g-PEG coatings to resist phagocytosis of the PS microspheres. We hypothesized that this should correlate with the resistance of the polymeric adlayers to unspecific protein adsorption. The latter has previously been shown to depend primarily on the ethylene-glycol-monomer surface density, which is a function of both grafting ratio and PEG chain length of the polymer [35]. To verify this hypothesis, both DC and MΦ were incubated with surface-coated microspheres. Number of internalized microspheres per cell were determined by phase contrast microscopy and normalized to uncoated control microspheres.

74 CHAPTER II

Figure 6. Correlation of IgG Adsorption on Coated PS Microspheres with Serum Adsorption on Coated Flat Nb2O5. Median fluorescence values (± SD) from flow cytometry after incubation of uncoated and PLL-g-PEG coated PS microspheres with TxRed-goat anti-rabbit IgG(H+L) were normalized to uncoated control microspheres and correlated with published data on protein mass adsorbed from whole serum on flat Nb2O5 surfaces coated with PLL-g-PEG copolymers of the same architectures (from Pasche et al. [35]).

Concerning the effect of the grafting ratio, while keeping the general architecture of PLL(20)-g-PEG(2) polymers constant, we found relatively low phagocytosis rates (about 10 to 20 % of control levels) with grafting ratios of 2.2, 3.5 and 5.7 whereas the phagocytosis efficiency of PLL(20)-[10.1]-PEG(2) coated microspheres was close or equal to the uncoated control. This is in general agreement with the previous work of Pasche et al. on flat surfaces, who qualified the first three as protein-resistant (i.e., ≤ 10 ng/cm2 adsorbed serum), but not the latter [35]. Within the [g] range of 2.2 to 5.7, however, IgG protein adsorption was found to increase with the grafting ratio g, which in turn should have led to increasing average numbers of phagocytosed microspheres even for these small values. Experimentally, with respect to phagocytosis the data showed an opposite trend to the one seen in the IgG adsorption experiments. (Although statistically insignificant, the same tendency was found in most experiments). A possible explanation for this apparent discrepancy would be the assumption of a preferential adsorption of dysopsonins between the grafted PEG

CHAPTER II 75

chains for polymers with g = 5.7 when exposed to the cell culture medium. Although difficult to identify, dysopsonins have been shown to impair the phagocytosis of various particulates such as latex microspheres or liposomes [17, 18, 60-62]. In a recent study, a PEG corona was demonstrated to not only reduce protein adsorption but also to shift the composition of the adsorbed proteins towards dysopsonins [43]. In particular, albumin is abundant in body fluids and has been shown to have dysopsonizing properties [19, 43]. Further studies should therefore concentrate on the identification of the adsorbed proteins. Another important issue to consider is the polymer surface coverage. Pasche et al. [35] showed for Nb2O5-coated surfaces that the surface coverage of the polymer with the very low grafting ratio, g = 2.2, was significantly reduced while the polymers with g ≥ 3.5 were approximately identical. This was ascribed to the smaller positive charge of the PLL backbone and especially to the onset of strong steric repulsion between the densely packed PEG chains. Concomitantly, a slight increase in serum adsorption relative to the polymer with the next higher grafting ratio, g = 3.5, was observed. Thus, internalization of microspheres coated with this polymer would be expected to increase relative to those with a slightly higher grafting ratio.

In a second set of experiments we varied the PEG chain length from 1 to 5 kDa, while the grafting ratio was adjusted such that the ethylene-glycol-monomer surface density was kept approximately constant. The rationale of this procedure was to examine a potential effect of the PEG chain length with polymers of comparable ethylene-glycol-monomer density. By choosing a set of non-protein-resistant polymers we expected to find more drastic effects, if any, than for polymers that resist the adsorption of proteins. Interestingly, in this set of experiments we encountered differences between the two cell types. In the case of MΦ, coatings of the copolymers PLL(20)-[6.5]-PEG(1), PLL(20)-[10.1]-PEG(2) and PLL(20)-[18.7]-PEG(5) yielded no significant deviation from control microspheres. This was not surprising, given the fact that relatively high amounts of protein adsorbed on all three coatings. For DC, however, the results were unexpected in the sense that both the 1 kDa and the 5 kDa PEG grafted polymers significantly reduced microsphere internalization. In fact, the 1 kDa PEG copolymer reduced phagocytosis to nearly the same extent as the protein-

76 CHAPTER II

resistant ones. These findings clearly indicate that although the ethylene-glycol-monomer density per surface area seems to be the decisive factor for protein resistance [35], it is not a sufficient criterion to predict phagocytic efficiency of APC. As discussed above, this effect cannot be ascribed to differences in the microspheres´ surface charge because these were found to be all quite low for the coated particles. In addition, residual positive charges would have led to the opposite result since cationic surfaces are believed to enhance phagocytosis [14-16]. In the case of the 5 kDa PEG copolymer, its long PEG chains may constitute a steric barrier around the microspheres reducing the degree of interaction with the opsonins [63]. This was corroborated by our IgG adsorption study in which unexpectedly low levels of IgG were seen to adsorb onto the microspheres coated with this long PEG chain copolymer. On the other hand, microspheres coated with the 1 kDa PEG copolymer adsorbed high amounts of protein. Thus, the explanation for the highly reduced phagocytosis of these microspheres must probably be rather found in type and composition of the adsorbed proteins. Indeed, the grafting ratio of this polymer (g=6.5) is close to the grafting ratio of PLL(20)-[5.7]-PEG(2) for which we speculated that it would favor preferential adsorption of dysopsonins. The fact that these observations only apply to DC but not to MΦ is likely to reflect the generally superior phagocytic efficiency of MΦ [14, 64].

In summary, two main conclusions can be drawn from our phagocytosis study: Firstly, microsphere resistance to recognition by phagocytes is primarily correlated with resistance of its surface to protein adsorption. While high PEG chain densities are crucial for low cellular recognition, lower PEG chain densities may to some extent be balanced if the PEG chains are sufficiently long (5 kDa). Secondly, our data suggest that for polymer-coated surfaces with low but not negligible protein adsorption, the nature of the adsorbed proteins could be crucial. Although further studies with a variety of proteins would be needed to corroborate this assumption, in special cases adsorption of dysopsonins could be even more effective in reducing phacocytic activity than minimal protein adsorption.

CHAPTER II 77

Phagocytosis of Microspheres versus Cellular Adhesion

As both phagocytosis of microspheres (that do not expose specific ligands) and cellular adhesion rely primarily on protein binding to the respective surfaces prior to recognition by APCs, a correlation between both processes could be expected depending on the physico-chemical properties of the respective surfaces, i.e. the architecture of the PLL-g-PEG polymer in our case. Preliminary results from Faraasen et al. [47] supported this hypothesis. With the more extensive data set of this work, we found, however, that the correlation between phagocytic activity and cell adhesion is nonlinear. As seen in Fig. 7, both DC and MΦ were able to adhere to some extent to more or less protein-resistant polymer adlayers, whereas phagocytosis of microspheres coated with the same copolymers was yet inefficient. Two factors may account for this finding: Firstly, cells that are seeded onto a surface undergo sedimentation and secrete proteins to remodel the surface for improved adhesion. The resulting high interfacial concentration of proteins will increase the probability of protein adhesion to structural defects in the coating, resulting in an increase of cell adhesion to the remodelled surfaces. Moreover, cellular adhesion on PEG coated surfaces has been attributed to reversible adsorption of proteins, which was reported to support weak, reversible cellular adhesion [30]. Secondly, cells require a minimum surface density of “adhesive spots” for efficient spreading [65], while phagocytosis relies on a minimum number of activated receptors to recognize the microspheres as a foreign material [8]. Obviously, these numbers need not be the same, and it is possible that higher ligand densities are needed for phagocytosis than for spreading. In addition, receptors that mediate adhesion are not necessarily active in phagocytosis, and dysopsonins may interfere with microsphere recognition.

Thus, adhesion studies yield supplementary information rather than merely confirming phagocytosis data. In the present study, only PLL(20)-[3.5]-PEG(2) showed no cellular adhesion, which is another hint that this is the only polymer, within the tested series, with an architecture that ensures optimum protein resistance. This finding corresponds with results from other authors on comparable flat or microparticulate systems [35, 37, 47].

78 CHAPTER II

Figure 7. Correlation of Phagocytosis of PLL-g-PEG coated PS Microspheres with APC Adhesion on PLL-g-PEG Coated Glass Slides. Correlation of data obtained from phagocytosis (mean counts ± SD) vs. data obtained from adhesion experiments (mean counts ± SD). DC (black) and MΦ (grey) are plotted in the same graph. The outlier data point refers to PLL-(20)-[2.1]-PEG(2). Dashed curve indicates trend of data.

Binary Polymer Mixtures

In the previous section we could show that the bioresponsiveness of surface coatings of microspheres could be tailored by using PLL-g-PEG with different architecture. In addition, we were able to demonstrate that the degree of resistance to phagocytosis can be quantitatively tailored by assembling mixed PLL-g-PEG coatings based on two different polymers rather than just one type. In the present study, binary mixtures consisting of a protein-resistant (PLL(20)-[5.7]-PEG(2)) and a non-resistant copolymer (PLL(20)-[10.1]-PEG(2)) were tested. We observed microspheres to become increasingly resistant to phagocytosis with increasing proportion of the protein resistant PLL-g-PEG copolymer in the mixed adlayer. The correlation is, however, not linear (Fig. 4), with the phagocytotic activity increasing most steeply at the lowest fraction of the (non-protein-resistant) PLL(20)-[10.1]-PEG(2), and levelling off once the proportion of this polymer in the coating has reached 60%. Since protein adsorption is expected to correlate approximately linearly with the fraction of

CHAPTER II 79

the non-protein-resistant polymer, this implies that (much) less than a monolayer of opsonins at the microsphere surface is required for efficient recognition and uptake by APC.

CONCLUSION

In this work, we studied the impact of PLL-g-PEG copolymer architecture on the resistance of coated microspheres to protein adsorption and phagocytosis by APC. As expected, protein adsorption was found to increase with increasing grafting ratio. Lower adsorption was found for longer (5 kDa) PEG chains, which suggests a steric effect of the long PEG chains on the adsorption of the relatively big model protein, IgG.

Two conclusions could be drawn from phagocytosis and cellular adhesion studies: Firstly, microsphere resistance to recognition by phagocytes was found to be mostly dependent on the protein resistance of the surface coatings. Although a too low PEG density could to some extent be balanced by longer PEG chains, high PEG chain densities were crucial for low cellular recognition. From the copolymer architectures used in this study, PLL(20)-[3.5]-PEG(2) was identified as optimal to ensure resistance to phagocytosis. Secondly, our data suggested that for polymers with low but not negligible protein adsorption, the nature of these proteins would be of essential importance. Although rather speculative at the moment, in special cases adsorption of dysopsonins seemed to be even more effective to impair phagocytosis than minimal protein adsorption. In addition, PLL-g-PEG surface coverage may become important for very high PEG chain densities as in the case of PLL(20)-[2.2]-PEG(2), where the PEG interchain repulsion is believed to reduce the surface coverage of the copolymer thereby decreasing the protection from cellular recognition.

In this work, we successfully established quantification of DC phagocytosis by side scatter analysis using a flow cytometer. Application of this technique to microspheres coated with binary PLL-g-PEG mixtures revealed the feasibility to modify microsphere surfaces with a coating made up from a

80 CHAPTER II

mixture of different PLL-g-PEG type copolymers. This is of particular interest for the future use of these microspheres in cell targeting. For instance, it would be possible to decorate the microspheres’ surface with a mixture of cell targeting ligands, which would be individually linked to PEG side chains of PLL-g-PEG copolymers. Different surface compositions could then be obtained by variation of the proportions of differently functionalized and unfunctionalized copolymers in the coating mixture. Thus our microsphere system would allow for a convenient adjustment of ligand density and composition. At the same time, a stealth background could be ensured by choice of the appropriate copolymer architecture.

ACKNOWLEDGEMENTS

The authors would like to thank Stéphanie Pasche for synthesis of PLL-g-PEG copolymers.

ABBREVIATIONS

The abbreviations used are: MPS, mononuclear phagocytic system; APC, antigen presenting cells; DC, dendritic cells; MΦ, macrophages; PEG, poly(ethylene glycol); PLL, poly-L-lysine; PLL(x)-g[y]-PEG(z), copolymer of PEG (MW=z kDa) grafted to PLL (MW=x kDa) at a grafting ratio y; grafting ratio, g, number of lysine monomers divided by the number of PEG side chains; IgG, immunoglobulin G; PS, polystyrene; HEPES buffer, 10 mM 4-(2-hydroxyethyl)piperazine-1-ethane-sulfonic acid, pH 7.4; PBS buffer, 10 mM phosphate buffered saline, pH 7.4; RPMI medium, RPMI 1640 (Roswell Park Memorial Institute) with L-glutamine; SSC, side scattering (measured by flow cytometry).

CHAPTER II 81

REFERENCES 1. Ha, C.-S., and Gardella Joseph, A., Jr., Surface chemistry of

biodegradable polymers for drug delivery systems. Chem. Rev., 2005. 105(11): p. 4205-4232.

2. Thiele, L., Merkle, H.P., and Walter, E., Phagocytosis of synthetic particulate vaccine delivery systems to program dendritic cells. Expert Review of Vaccines, 2002. 1(2): p. 215-226.

3. Svenson, S., Carrier-based drug delivery. ACS Symp. Ser., 2004. 879(Carrier-Based Drug Delivery): p. 2-23.

4. Men, Y., Tamber, H., Audran, R., Gander, B., and Corradin, G., Induction of a cytotoxic T lymphocyte response by immunization with a malaria specific CTL peptide entrapped in biodegradable polymer microspheres. Vaccine, 1997. 15(12-13): p. 1405-1412.

5. Peter, K., Men, Y., Pantaleo, G., Gander, B., and Corradin, G., Induction of a cytotoxic T-cell response to HIV-1 proteins with short synthetic peptides and human compatible adjuvants. Vaccine, 2001. 19(30): p. 4121-4129.

6. Partidos, C.D., Vohra, P., Jones, D.H., Farrar, G.H., and Steward, M.W., Mucosal immunization with a measles virus CTL epitope encapsulated in biodegradable PLG microparticles. J. Immunol. Methods, 1996. 195(1-2): p. 135-138.

7. Kovacsovics-Bankowski, M., Clark, K., Benacerraf, B., and Rock, K.L., Efficient major histocompatibility complex class I presentation of exogenous antigen upon phagocytosis by macrophages. Proc. Natl. Acad. Sci. U. S. A., 1993. 90(11): p. 4942-4946.

8. Stuart, L.M., and Ezekowitz, R.A.B., Phagocytosis: Elegant complexity. Immunity, 2005. 22(5): p. 539-550.

9. Walter, F., Scholl, I., Untersmayr, E., Ellinger, A., Boltz-Nitulescu, G., Scheiner, O., Gabor, F., and Jensen-Jarolim, E., Functionalisation of allergen-loaded microspheres with wheat germ agglutinin for targeting enterocytes. Biochem. Biophys. Res. Commun., 2004. 315(2): p. 281-287.

82 CHAPTER II

10. Kempf, M., Mandal, B., Jilek, S., Thiele, L., Vörös, J., Textor, M., Merkle, H.P., and Walter, E., Improved Stimulation of Human Dendritic Cells by Receptor Engagement with Surface-modified Microparticles. J. Drug Target., 2003. 11(1): p. 11-18.

11. Copland, M.J., Baird, M.A., Rades, T., McKenzie, J.L., Becker, B., Reck, F., Tyler, P.C., and Davies, N.M., Liposomal delivery of antigen to human dendritic cells. Vaccine, 2003. 21(9-10): p. 883-890.

12. Keegan, M.E., Whittum-Hudson, J.A., and Saltzman, W.M., Biomimetic design in microparticulate vaccines. Biomaterials, 2003. 24(24): p. 4435-4443.

13. Juliano, R.L., Factors affecting the clearance kinetics and tissue distribution of liposomes, microspheres and emulsions. Adv. Drug Deliv. Rev., 1988. 2(1): p. 31-54.

14. Thiele, L., Rothen-Rutishauser, B., Jilek, S., Wunderli-Allenspach, H., Merkle, H.P., and Walter, E., Evaluation of particle uptake in human blood monocyte-derived cells in vitro. Does phagocytosis activity of dendritic cells measure up with macrophages? J. Controlled Release, 2001. 76(1-2): p. 59-71.

15. Foged, C., Brodin, B., Frokjaer, S., and Sundblad, A., Particle size and surface charge affect particle uptake by human dendritic cells in an in vitro model. International Journal of Pharmaceutics, 2005. 298(2): p. 315-322.

16. Ahsan, F., Rivas, I.P., Khan, M.A., and Torres Suárez, A.I., Targeting to macrophages: role of physicochemical properties of particulate carriers-liposomes and microspheres-on the phagocytosis by macrophages. J. Controlled Release, 2002. 79(1-3): p. 29-40.

17. Absolom, D.R., Opsonins and dysopsonins: an overview. Methods Enzymol., 1986. 132(Immunochem. Tech., Pt. J): p. 281-318.

18. Patel, H.M., Serum opsonins and liposomes: their interaction and opsonophagocytosis. Crit. Rev. Ther. Drug Carrier Syst., 1992. 9(1): p. 39-90.

CHAPTER II 83

19. Thiele, L., Diederichs, J.E., Reszka, R., Merkle, H.P., and Walter, E., Competitive adsorption of serum proteins at microparticles affects phagocytosis by dendritic cells. Biomaterials, 2003. 24(8): p. 1409-1418.

20. Kingshott, P., and Griesser, H.J., Surfaces that resist bioadhesion. Curr. Opin. Solid State Mater. Sci, 1999. 4(4): p. 403-412.

21. Moghimi, S.M., Hunter, A.C., and Murray, J.C., Long-circulating and target-specific nanoparticles: theory to practice. Pharmacol. Rev., 2001. 53(2): p. 283-318.

22. Gref, R., and Couvreur, P., Stealth and biomimetic core-corona nanoparticles. Encyclopedia of Nanoscience and Nanotechnology, 2004. 10: p. 83-94.

23. Harris, J.M., Zalipsky, S., and Editors, Polyethylene glycol: Chemistry and Biological Applications. ACS Symp. Ser., 1997. 680: p. 489 pp.

24. Harris, J.M., and Chess, R.B., Effect of PEGylation on pharmaceuticals. Nat. Rev. Drug Discovery, 2003. 2(3): p. 214-221.

25. Vermette, P., and Meagher, L., Interactions of phospholipid- and poly(ethylene glycol)-modified surfaces with biological systems: relation to physico-chemical properties and mechanisms. Colloids Surf. B, 2003. 28(2-3): p. 153-198.

26. Morra, M., On the molecular basis of fouling resistance. J. Biomater. Sci. Polym. Ed., 2000. 11(6): p. 547-569.

27. Jeon, S.I., Lee, J.H., Andrade, J.D., and De Gennes, P.G., Protein-surface interactions in the presence of polyethylene oxide. I. Simplified theory. J. Colloid Interface Sci., 1991. 142(1): p. 149-158.

28. Jeon, S.I., and Andrade, J.D., Protein-surface interactions in the presence of polyethylene oxide. II. Effect of protein size. J. Colloid Interface Sci., 1991. 142(1): p. 159-166.

29. Halperin, A., Polymer Brushes that Resist Adsorption of Model Proteins: Design Parameters. Langmuir, 1999. 15(7): p. 2525-2533.

30. Zhu, B., Eurell, T., Gunawan, R., and Leckband, D., Chain-length dependence of the protein and cell resistance of oligo(ethylene glycol)-

84 CHAPTER II

terminated self-assembled monolayers on gold. J. Biomed. Mater. Res., 2001. 56(3): p. 406-416.

31. Prime, K.L., and Whitesides, G.M., Adsorption of proteins onto surfaces containing end-attached oligo(ethylene oxide): a model system using self-assembled monolayers. J. Am. Chem. Soc., 1993. 115(23): p. 10714-10721.

32. McPherson, T., Kidane, A., Szleifer, I., and Park, K., Prevention of Protein Adsorption by Tethered Poly(ethylene oxide) Layers: Experiments and Single-Chain Mean-Field Analysis. Langmuir, 1998. 14(1): p. 176-186.

33. Malmsten, M., Emoto, K., and Van Alstine, J.M., Effect of chain density on inhibition of protein adsorption by poly(ethylene glycol) based coatings. J. Colloid Interface Sci., 1998. 202(2): p. 507-517.

34. Kingshott, P., Thissen, H., and Griesser, H.J., Effects of cloud-point grafting, chain length, and density of PEG layers on competitive adsorption of ocular proteins. Biomaterials, 2002. 23(9): p. 2043-2056.

35. Pasche, S., De Paul, S.M., Vörös, J., Spencer, N.D., and Textor, M., Poly(L-lysine)-graft-poly(ethylene glycol) assembled monolayers on Niobium Oxide surfaces: a quantitative study of the influence of polymer interfacial architecture on resistance to protein adsorption by ToF-SIMS and in situ OWLSu OWLS. Langmuir, 2003. 19(22): p. 9216-9225.

36. Huang, N.-P., Michel, R., Vörös, J., Textor, M., Hofer, R., Rossi, A., Elbert, D.L., Hubbell, J.A., and Spencer, N.D., Poly(L-lysine)-g-poly(ethylene glycol) Layers on Metal Oxide Surfaces: Surface-Analytical Characterization and Resistance to Serum and Fibrinogen Adsorption. Langmuir, 2001. 17(2): p. 489-498.

37. Kenausis, G.L., Vörös, J., Elbert, D.L., Huang, N., Hofer, R., Ruiz-Taylor, L., Textor, M., Hubbell, J.A., and Spencer, N.D., Poly(L-lysine)-g-Poly(ethylene glycol) Layers on Metal Oxide Surfaces: Attachment Mechanism and Effects of Polymer Architecture on Resistance to Protein Adsorption. J. Phys. Chem. B, 2000. 104(14): p. 3298-3309.

CHAPTER II 85

38. Feuz, L., Leermakers, F.A.M., Textor, M., and Borisov, O., Bending Rigidity and Induced Persistence Length of Molecular Bottle Brushes: A Self-Consistent-Field Theory. Macromolecules, 2005. 38(21): p. 8891-8901.

39. Müller, M., Vörös, J., Csúcs, G., Walter, E., Danuser, G., Merkle, H.P., Spencer, N.D., and Textor, M., Surface modification of PLGA microspheres. J. Biomed. Mater. Res., Part A, 2003. 66A(1): p. 55-61.

40. VandeVondele, S., Vörös, J., and Hubbell, J.A., RGD-grafted poly-L-lysine-graft-(polyethylene glycol) copolymers block non-specific protein adsorption while promoting cell adhesion. Biotechnol. Bioeng., 2003. 82(7): p. 784-790.

41. Huang, N.-P., Vörös, J., De Paul, S.M., Textor, M., and Spencer, N.D., Biotin-derivatized poly(L-lysine)-g-poly(ethylene glycol): A novel polymeric interface for bioaffinity sensing. Langmuir, 2002. 18(1): p. 220-230.

42. Tosatti, S., De Paul, S.M., Askendal, A., VandeVondele, S., Hubbell, J.A., Tengvall, P., and Textor, M., Peptide functionalized poly(L-lysine)-g-poly(ethylene glycol) on titanium: resistance to protein adsorption in full heparinized human blood plasma. Biomaterials, 2003. 24(27): p. 4949-4958.

43. Meng, F., Engbers, G.H.M., Gessner, A., Mueller, R.H., and Feijen, J., Pegylated polystyrene particles as a model system for artificial cells. J. Biomed. Mater. Res., Part A, 2004. 70A(1): p. 97-106.

44. Gref, R., Luck, M., Quellec, P., Marchand, M., Dellacherie, E., Harnisch, S., Blunk, T., and Muller, R.H., 'Stealth' corona-core nanoparticles surface modified by polyethylene glycol (PEG): influences of the corona (PEG chain length and surface density) and of the core composition on phagocytic uptake and plasma protein adsorption. Colloids Surf. B, 2000. 18(3,4): p. 301-313.

45. Lacasse, F.X., Filion, M.C., Phillips, N.C., Escher, E., McMullen, J.N., and Hildgen, P., Influence of surface properties at biodegradable

86 CHAPTER II

microsphere surfaces: effects on plasma protein adsorption and phagocytosis. Pharm. Res., 1998. 15(2): p. 312-317.

46. Heuberger, R., Sukhorukov, G., Vörös, J., Textor, M., and Möhwald, H., Biofunctional polyelectrolyte multilayers and microcapsules: control of non-specific and bio-specific protein adsorption. Adv. Funct. Mater., 2005. 15(3): p. 357-366.

47. Faraasen, S., Vörös, J., Csúcs, G., Textor, M., Merkle, H.P., and Walter, E., Ligand-Specific Targeting of Microspheres to Phagocytes by Surface Modification with Poly(L-Lysine)-Grafted Poly(Ethylene Glycol) Conjugate. Pharm. Res., 2003. 20(2): p. 237-246.

48. Sallusto, F., Cella, M., Danieli, C., and Lanzavecchia, A., Dendritic cells use macropinocytosis and the mannose receptor to concentrate macromolecules in the major histocompatibility complex class II compartment: downregulation by cytokines and bacterial products. J. Exp. Med., 1995. 182(2): p. 389-400.

49. Pasche, S., Mechanisms of protein resistance of adsorbed PEG-graft copolymers, Doctoral Thesis, Swiss Federal Institute of Technology Zurich, 2004,

http://e-collection.ethbib.ethz.ch/show?type=diss&nr=15712.

50. See Appendix I or EPAPS Document No. E-BJIOBN-003604 for details on the gating strategy. This document can be reached through a direct link in the online articles's HTML reference section or via the EPAPS homepage (http://www.aip.org/pubservs/epaps.html).

51. Stringer, B., Imrich, A., and Kobzik, L., Flow cytometric assay of lung macrophage uptake of environmental particulates. Cytometry, 1995. 20(1): p. 23-32.

52. See Appendix I or EPAPS Document No. E-BJIOBN-003604 for phase contrast and confocal microscopy images on the intracellular localization of the microspheres. This document can be reached through a direct link in the online articles's HTML reference section or via the EPAPS homepage (http://www.aip.org/pubservs/epaps.html).

CHAPTER II 87

53. Parks, D.R., and Herzenberg, L.A., Fluorescence-activated cell sorting: theory, experimental optimization, and applications in lymphoid cell biology. Methods Enzymol., 1984. 108(Immunochem. Tech., Pt. G): p. 197-241.

54. Olivier, V., Riviere, C., Hindie, M., Duval, J.L., Bomila-Koradjim, G., and Nagel, M.D., Uptake of polystyrene beads bearing functional groups by macrophages and fibroblasts. Colloids Surf. B, 2004. 33(1): p. 23-31.

55. Rejman, J., Oberle, V., Zuhorn, I.S., and Hoekstra, D., Size-dependent internalization of particles via the pathways of clathrin- and caveolae-mediated endocytosis. Biochem. J., 2004. 377(1): p. 159-169.

56. Stewart, C.C., Lehnert, B.E., and Steinkamp, J.A., In vitro and in vivo measurement of phagocytosis by flow cytometry. Methods Enzymol., 1986. 132: p. 183-192.

57. Pasche, S., Textor, M., Meagher, L., Spencer, N.D., and Griesser, H.J., Relationship between interfacial forces measured by colloid-probe atomic force microscopy and protein resistance of poly(ethylene glycol)-grafted poly(L-lysine) adlayers on niobia surfaces. Langmuir, 2005. 21(14): p. 6508-6520.

58. Gbadamosi, J.K., Hunter, A.C., and Moghimi, S.M., PEGylation of microspheres generates a heterogeneous population of particles with differential surface characteristics and biological performance. FEBS Lett., 2002. 532(3): p. 338-344.

59. Ishida, T., Harashima, H., and Kiwada, H., Liposome clearance. Biosci. Rep., 2002. 22(2): p. 197-224.

60. Johnstone, S.A., Masin, D., Mayer, L., and Bally, M.B., Surface-associated serum proteins inhibit the uptake of phosphatidylserine and poly(ethylene glycol) liposomes by mouse macrophages. BBA: Biomembranes, 2001. 1513(1): p. 25-37.

61. Moghimi, S.M., Muir, I.S., Illum, L., Davis, S.S., and Kolb-Bachofen, V., Coating particles with a block copolymer (poloxamine-908) suppresses opsonization but permits the activity of dysopsonins in the serum. Biochim. Biophys. Acta, 1993. 1179(2): p. 157-165.

88 CHAPTER II

62. Moghimi, S.M., and Szebeni, J., Stealth liposomes and long circulating nanoparticles: critical issues in pharmacokinetics, opsonization and protein-binding properties. Prog. Lipid Res., 2003. 42(6): p. 463-478.

63. Mori, A., Klibanov, A.L., Torchilin, V.P., and Huang, L., Influence of the steric barrier activity of amphipathic polyethylene glycol and ganglioside GM1 on the circulation time of liposomes and on the target binding of immunoliposomes in vivo. FEBS Lett., 1991. 284(2): p. 263-266.

64. Kiama, S.G., Cochand, L., Karlsson, L., Nicod, L.P., and Gehr, P., Evaluation of phagocytic activity in human monocyte-derived dendritic cells. Journal of aerosol medicine: official journal of the International Society for Aerosols in Medicine, 2001. 14(3): p. 289-299.

65. Massia, S.P., and Hubbell, J.A., An RGD spacing of 440 nm is sufficient for integrin alpha V beta 3-mediated fibroblast spreading and 140 nm for focal contact and stress fiber formation. J. Cell Biol., 1991. 114(5): p. 1089-1100.

CHAPTER III 89

CHAPTER III

Stable Stealth Function for Hollow Polyelectrolyte Microcapsules through a Polyethylene Glycol Grafted

Polyelectrolyte Adlayer

Uta Wattendorf,a Oliver Kreft,b Marcus Textor,c Gleb B. Sukhorukov,d and Hans P. Merklea

a Institut for Pharmaceutical Sciences, ETH Zurich, 8093 Zurich, Switzerland b Max-Planck Institute of Colloids and Interfaces, Research Campus Golm,

14424 Potsdam, Germany

c Laboratory for Surface Science and Technology, ETH Zurich, 8093 Zurich, Switzerland

d Department of Materials, Queen Mary University of London, Mile End Road, E1 4NS, London, UK.

Biomacromolecules, 2008. 9(1): pp. 100-108.

90 CHAPTER III

ABSTRACT

Prospective biomedical applications of hollow polyelectrolyte microcapsules, e.g., as drug delivery systems, require surface modifications that help to escape clearance by the mononuclear phagocytic system (MPS). Layer-by-layer assembled microcapsules that were alternatingly composed of polystyrene sulfonate (PSS) and polyallylamine hydrochloride (PAH) were coated with adlayers of PEG-grafted poly-L-lysine (PLL-g-PEG) and poly-L-glutamic acid (PGA-g-PEG). Their effects on MPS recognition was studied in primary cell cultures of human monocyte derived dendritic cells and macrophages. PGA-g-PEG coatings had no significant effect on cellular recognition, which may be explained by insufficient PEG density of the adlayer. Contrary, PLL-g-PEG effectively blocked phagocytosis of coated microcapsules. In addition, PLL-g-PEG coatings showed efficient adlayer stability for at least three weeks, and PAH/PSS microcapsules did not impair phagocyte viability. Our results demonstrate that layer-by-layer assembled polyelectrolyte microcapsules coated with a PEG-grafted polyelectrolyte, PLL-g-PEG, represent a promising platform for a drug delivery system that escapes fast clearance by the MPS.

CHAPTER III 91

INTRODUCTION

Hollow microcapsules made by layer-by-layer assembly of oppositely charged polyelectrolytes on particulate templates [1, 2] have been previously considered for their prospects towards biomedical application [3]. Among other particulate drug delivery systems such as micelles [4, 5] and polymeric micro- and nanoparticles [6, 7] - just to mention a few - polyelectrolyte microcapsules offer unique opportunities for both targeting and delivery of encapsulated materials. So far, a broad range of chemically diverse compounds has been successfully incorporated into hollow microcapsules, among them drugs and dyes [8, 9], as well as biopolymers such as proteins [10, 11] and nucleic acids [12, 13]. Recently, the development of a microcapsule-based sensor-system was demonstrated and utilized to measure pH-values inside human breast cancer cells and fibroblasts [14]. These cells internalize microcapsules in a non-specific process and store them in endosomal/lysosomal compartments [15-17].

Weakly cross-linked, monodisperse melamine formaldehyde (MF) particles have been extensively studied as templates for the fabrication of hollow polyelectrolyte microcapsules. After shell formation is complete, MF particles can be rapidly dissolved in organic solvents such as DMSO or via acid catalyzed decomposition of the polymer network [18, 19]. The resulting microcapsules are of spherical shape and narrow size distribution, and commercially available in size ranges between 0.5 and 12 µm.

A typical material used to functionalize hollow polyelectrolyte microcapsules are, e.g., nanoparticles. The accumulation of superparamagnetic Fe3O4 nanoclusters in the target tissue was achieved by means of a magnetic field [20]. Rupture of phagocytosed microcapsules functionalized with nanoparticulate silver or gold could be triggered by near infrared laser irradiation, thereby discharging the encapsulated cargo into the cells [21]. In addition, it was shown that once inside the cells partial defoliation of the microcapsules allowed for transfection with a GFP encoding plasmid that was immobilized within the polyelectrolyte multilayer of microparticles [17]. While the standard protocol to prepare such microcapsules involves bioincompatible and nondegradable components, namely polyallylamine hydrochloride (PAH)

92 CHAPTER III

and polystyrene sulfonate (PSS), it has been demonstrated that the underlying concept can be extended to biocompatible and biodegradable polymers as well [16, 22].

Typically, upon entry into the organism, particulate carriers are instantly recognized as foreign and efficiently removed via unspecific phagocytosis by the mononuclear phagocytic system (MPS), which comprises (in addition to other specialized organs and tissues) migrating phagocytes such as dendritic cells (DC) and macrophages (MΦ) [23]. Extent and rate of phagocytic clearance depends on both particle size and surface properties of the carrier. For instance, phagocytosis is enhanced by charged or hydrophobic surfaces [24-26]. It is particularly the adsorption of opsonic proteins onto the particles´ surface which largely promotes phagocytosis [27-29]. An important strategy to prevent serum protein adsorption and cellular recognition of particles is the installation of protein resistant coatings, e.g., through immobilization of a polyethylene glycol (PEG) corona. Due to the low toxicity and immunogenicity of PEG, such stealth particles are attractive tools for biomedical applications [26, 30].

A convenient method to immobilize PEG on polyelectrolyte microcapsules is the use of copolymers with PEG chains grafted to a polyelectrolyte backbone. Heuberger et al. [31] used PLL-g-PEG, consisting of a polycationic poly-L-lysine (PLL) backbone whose side chain amino groups were partially grafted with PEG chains, and assembled it on PSS-terminated polyelectrolyte microcapsules. Due to its positive charge at physiological pH, the PLL backbone adsorbs strongly onto negatively charged substrates through electrostatic interaction, leading to a monolayer with PEG chains stretched out perpendicularly to the surface [32]. In addition, the PEG side chains can be functionalized with bioligands while retaining resistance to unspecific protein adsorption [33-36].

Recently we studied the impact of PLL-g-PEG architecture on the resistance of surface modified polystyrene (PS) microspheres against protein adsorption and internalization by phagocytic cells, namely DC and MΦ. The stealth function achieved with a small library of such polymers was most prominent with PLL(20)-[3.5]-PEG(2), a polymer with a 20 kDa PLL backbone,

CHAPTER III 93

2 kDa PEG side chains, and an average of 3.5 lysine monomers between two PEG chains [37]. The grafting ratio g, defined as the total number of lysine monomers divided by the number of PEG side chains, was found to have a major influence on the stealth effect, with an optimum around g = 3.5. A structurally analogous approach is the use of a negatively charged poly-L-glutamic acid (PGA) backbone instead of PLL, leading to PGA-g-PEG. This polymer has been recently used to improve the anti-microbial protection of polyelectrolyte films [38].

In this work, we investigated PAH/PSS microcapsules that were surface modified with two PEG-grafted copolymers, PLL-g-PEG and PGA-g-PEG. The stealth effects of respective adlayers against cellular recognition and phagocytosis were studied in primary cell cultures of human monocyte derived DC and MΦ. PGA-g-PEG had no significant effect on cellular recognition, which may be explained by its insufficient PEG density. On the other hand, PLL-g-PEG effectively blocked phagocytosis of coated microcapsules, and the PLL-g-PEG coatings were stable within the investigated time frame of three weeks. Moreover, phagocytosed PAH/PSS microcapsules did not impair the viability of our cell model. Our results contribute to the design of layer-by-layer assembled microcapsules as promising drug delivery system that can be surface modified with a PEG grafted polymer, PLL-g-PEG, to escape fast clearance by the MPS.

MATERIALS AND METHODS Materials

Sodium polystyrene sulfonate (PSS, MW = 70 000), polyallylamine hydrochloride (PAH, MW = 70 000), poly-L-glutamic acid (PGA) sodium salt (MW = 50,000-100,000), poly-L-lysine (PLL) hydrobromide salt (MW = 30,000-70,000), 1-ethyl-3-(3-dimethylamino-propyl) carbodiimide (EDC) and tetramethylrhodamine isothiocyanate (TRITC) were purchased from Sigma-Aldrich (Munich, Germany). Methoxypoly(ethylene glycol) amine (MePEGNH2, MW = 5,000) and N-hydroxysulfosuccinimide sodium salt

94 CHAPTER III

(NHS) were from Fluka (Seelze, Germany). Melamine formaldehyde (MF) microparticles (1.2 and 4.8 μm) were obtained from Microparticles GmbH (Berlin, Germany). Alexa Fluor® 488 carboxylic acid (AF488), succinimidyl ester (mixed isomers) was purchased from Invitrogen (Karlsruhe, Germany). TRITC-PAH and AF488-PAH were synthesised according to the literature [39, 40]. Polystyrene (PS) microspheres (5 μm) were obtained from Micromod Partikeltechnologie GmbH (Rostock, Germany). All chemicals used in this study were of analytical grade (Fluka, Buchs, Switzerland) and used as received unless otherwise specified. Ultrapure water (NANOpure DiamondTM, Skan AG, Allschwil, Switzerland) was used for all experiments.

Synthesis of PEGylated copolymers

PLL-g-PEG [= PLL(20)-g[3.5]-PEG(2)] was received from SurfaceSolutionS GmbH (Zurich, Switzerland) and consisted of a PLL backbone of 20 kDa with grafted PEG chains of 2 kDa and a grafting ratio (specifying the total number of lysine monomers divided by the number of PEG side chains) of g = 3.5. Henceforth this polymer is abbreviated as PLL-g-PEG.

PGA-g-PEG [= PGA(50/100)-g[5.5]-PEG(5)] was synthesized according to the strategy described by Boulmedais et al. [38] Briefly, 100 mg of PGA (=0.67 mmol of acid functions), 5 mg of NHS and 600 mg of MePEGNH2 (0.12 mmol of amine functions), respectively, were dissolved in 5 ml of sodium tetraborate buffer (50 mM, pH 8.5). Then, 36.5 mg of EDC were added to the mixture with stirring. After 6 h at room temperature the product was dialyzed twice, first against phosphate buffer (2 l, 0.1 M, pH = 7.4, Na2HPO4/NaH2PO4) for 24 h and then against deionised water (2 l) for 24 h (cut off at MW

= 20,000). The product was freeze-dried and stored at -20 °C. A coupling rate of approximately 18% (= 18 PEG side chains per 100 monomers of glutamic acid) was determined by 1H-NMR and corresponded to a grafting ratio of g = 5.5. This polymer is referred to as PGA-g-PEG.

CHAPTER III 95

Microcapsule preparation

Microcapsules having a diameter of 1.2 µm or 4.8 µm, respectively, were prepared according to the literature [18, 19]. Briefly, by adsorption from aqueous solutions (2 mg/ml in 0.5 M NaCl), MF microparticles were consecutively coated with eight alternating layers of either PSS or PAH. Polyelectrolytes were adsorbed for at least 15 min followed by 3 washings with water. For fluorescent labelling the middle layers were made with PAH-TRITC or PAH-AF488 instead of PAH. After adsorption of a total of eight layers, the resulting core-shell particles were treated with 0.1 M HCl to decompose the MF cores. To facilitate an efficient and homogeneous adsorption of the two investigated types of PEGylated polymers, having either positively or negatively charged backbones, microcapsule preparations were split in two parts which were coated with either PAH or PSS, respectively, resulting in (PSS/PAH)4 and (PSS/PAH)4-PSS microcapsules. Particle numbers per volume were determined by either CLSM in combination with the software package "Igor Pro" (Wave Metrics, Lake Oswego, OR), or by single particle light scattering (see below). Further information regarding the characterization of microcapsules can be found in Appendix II.

Coating of microcapsules

PAH-terminated microcapsules were surface coated with PGA or PGA-g-PEG, and PSS-terminated microcapsules with PLL or PLL-g-PEG, respectively. In brief, 4.8 μm capsules were dispersed under aseptic conditions in sterile filtered polymer solution (1.33 mg/ml in 10 mM HEPES buffer supplemented with 150 mM sodium chloride, pH 7.4) at a concentration of 1.25 x 107 microcapsules per ml. 1.2 μm capsules were dispersed in 1.67 mg/ml polymer solutions at a concentration of 5 x 108 microcapsules per ml. After 1 hour of incubation under gentle mixing at room temperature, surface coated microcapsules were centrifuged (4.8 μm: 30 min at 350 x g; 1.2 μm: 60 min at 1450 x g; Centrifuge 5417R, Eppendorf-Netheler-Hinz GmbH, Hamburg, Germany) and washed with sterile filtered 10 mM HEPES buffer.

96 CHAPTER III

Coating of microspheres

PS microspheres (5 mg/ml) used for adlayer stability studies were dispersed under aseptic conditions in sterile filtered PLL-g-PEG solutions (0.5 mg/ml in 10 mM HEPES buffer, pH 7.4) and incubated under gentle mixing for 15 min at room temperature. Surface coated microspheres were centrifuged (5 min, 2655 x g, Centrifuge 5417R, Eppendorf-Netheler-Hinz GmbH, Hamburg, Germany) and washed with sterile filtered 10 mM HEPES buffer.

Single particle light scattering (SPLS)

SPLS measurements were conducted on a custom-made photometer equipped with an argon laser (Innova 305, Coherent, Dieburg, Germany) with power track, as described earlier [41]. For analysis the dispersion was focused through a thin capillary with an orifice of 0.1 mm diameter at the end. Hydrodynamic focusing was applied. Capsule concentration was adjusted to minimize coincidences. The scattering volume was 1.7 x 10-9 cm-3. Forward scattered light pulses recorded from capsules flowing through the scattered volume were detected between 5° and 10°. Intensity distributions were obtained with a resolution of about 0.5%.

Electrophoretic mobility

Electrophoretic mobilities of microcapsules and microparticles were measured using a Malvern Zetasizer 3000HS (Malvern Instruments, Worcestershire, UK). The mobility u was converted into a ζ-potential using the Helmholtz-Smoluchowski relation (ζ = ν / η ε E; ν = particle mobility, η, ε = viscosity and permittivity of the solution, E = electric field strength). All measurements were performed in filtered 10 mM HEPES buffer at pH 7.4.

DC and MΦ cell culture

Primary cultures of DC and MΦ were obtained from human peripheral blood monocytes according to Sallusto et al. [42] Both cell types efficiently internalize foreign particulate material such as bacteria or microparticles, even if

CHAPTER III 97

the phagocytic efficiency of MΦ is generally superior [26, 43]. Briefly, peripheral blood monocytes were isolated from buffy coats (Bloodbank Zurich, Zurich, Switzerland) by density gradient centrifugation of Ficoll-PaqueTM PLUS (Amersham Biosciences, Uppsala, Sweden). Isolated peripheral blood monocytes were washed in PBS (Sigma, Buchs, Switzerland) and resuspended in RPMI 1640 + L-glutamine (Invitrogen, Basel, Switzerland) supplemented with 10% heat-inactivated (pooled) human serum (Bloodbank Zurich, Zurich, Switzerland), 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen, Basel, Switzerland) and then allowed to adhere for 2 h in culture flasks (25 cm2). Nonadherent cells were removed whereas adherent cells were further cultured in RPMI medium supplemented with 5% heat-inactivated (pooled) human serum, 100 U/ml penicillin and 100 μg/ml streptomycin in the presence of 1000 IU/ml interleukin-4 (Sigma, Buchs, Switzerland) and 50 ng/ml human granulocyte-macrophage colony stimulating factor (R+D Systems Europe Ltd., Abingdon, UK) to obtain DC. MΦ were obtained without additional supplements [26]. These culture media will further be referred to as supplemented RPMI media. Cultures of DC and MΦ were kept at 37 °C in 5% CO2-humidified atmosphere. For phagocytosis studies, cells were mechanically removed from the flasks after 24 h and reseeded into 24-well plates (3 x 105 cells per well).

Phagocytosis studies

Phagocytosis was studied with both DC and MΦ. To rule out variances originating from donor-to-donor variability, each cell batch was tested with all formulations in parallel. For the assessment of phagocytosis, 3 x 107 (1.2 μm) or 3 x 106 (4.8 μm) capsules dispersed in filtered HEPES buffer were added to either DC or MΦ cultured in 24-well plates (3 x 105 cells per well) and incubated at 37 °C in 5% CO2-humidified atmosphere for 4 h. Control samples were incubated at 4 °C in parallel to identify microcapsules adsorbed to the outer cell membrane. Cells were immediately inspected for phagocytosis by phase contrast and fluorescence microscopy (Axiovert 35, Zeiss, Germany). Following microscopic analysis, the cells were mechanically recovered in a test tube and additionally evaluated by flow cytometry (FACSCanto, BD

98 CHAPTER III

Biosciences, Basel, Switzerland). For data analysis, mean fluorescence values were calculated with FlowJo software (v7.1, Tree Star Inc., Ashland, OR, USA) and corrected for background values at 4 °C. For flow cytometric analysis of non-fluorescent PS microspheres, mean SSC values were calculated as established previously [37].

Internalization of microcapsules analyzed by confocal microscopy

To confirm phagocytosis of microcapsules, in contrast to adhesion to the outer cell membrane, cells were cultured in Lab-Tek 8 well chamber glass slides (Nalge Nunc Int., IL, USA) at a concentration of 200,000 cells per well and analyzed by confocal laser scanning microscopy (CLSM). Phagocytosis experiments were performed as described above. After incubation with the microspheres, cells were washed three times with 10 mM PBS (Sigma, Buchs, Switzerland). To label the cell membranes, the cells were incubated with 100 μl of 50 μg/ml ConcanavalinA-AF633 (Molecular Probes Inc., Eugene, OR, USA) in 10 mM PBS for 2 minutes. Cells were washed three times with PBS and fixed with 3% paraformaldehyde (Sigma, Buchs, Switzerland) in PBS for 15 minutes. To label the actin cytoskeleton, cells were washed with PBS, permeabilized with Triton-X (0.1% in PBS supplemented with 1% BSA) for 20 min and stained with Phalloidin-OregonGreen488 (6 nM in PBS with 1% BSA; Molecular Probes Inc., Eugene, OR, USA) for 20 minutes. After three washings with PBS, the cells’ nuclei were stained with 1 μg/ml Hoechst-33342 (Molecular Probes Inc., Eugene, OR, USA) in PBS for 5 minutes. The samples were washed three times and mounted for CLSM using the ProLong® Gold antifade reagent from Molecular Probes (Eugene, OR, USA). Samples were examined with a Zeiss LSM 510 META (Carl Zeiss AG, Feldbach, Switzerland). Image processing was performed with Imaris (Bitplane, Zurich, Switzerland), a 3-D multi-channel image processing software for CLSM images.

Viability studies

Cell viability after 3 days of microsphere incubation was assessed by staining with propidium iodide (PI; BD Biosciences, Basel, Switzerland). PI

CHAPTER III 99

stains cells whose plasma membranes were compromised due to necrosis or apoptosis. Phagocytosis experiments were performed as described above. PI staining was performed according to the manufacturer’s protocol. In brief, cells were washed with cold PBS (Sigma, Buchs, Switzerland), and incubated with 5 μl of PI per 100,000 cells for 10 minutes on ice. 200 μl of cold PBS were added and samples were analyzed within 1 hour by flow cytometry (FACSCanto, BD Biosciences, Basel, Switzerland). For positive controls, cells were fixed with 10% neutral buffered formalin solution (Sigma, Buchs, Switzerland) before PI staining. For data analysis, cells were gated for the DC or MΦ population, respectively, and percentages of positively stained cells were calculated with the instrument’s BD FACSDivaTM software, using untreated cells as background control.

Stability studies

To assess the stability of the PLL-g-PEG coatings, microcapsules (4.8 μm) and PS microspheres (5 μm) were coated as described above and stored in 10 mM HEPES buffer at 4 °C for up to 27 days. In addition, samples of PEGylated PS microspheres were stored in 5% serum supplemented RPMI medium at 37 °C. Coated samples and uncoated controls were washed with sterile filtered 10 mM HEPES (pH 7.4) before ζ-potential measurements and incubation with MΦ. Phagocytosis experiments and ζ-potential measurements were performed as described above.

Significance testing

For evaluation of statistical significance, samples were evaluated by ANOVA followed by a post-hoc assessment using the Tukey-HSD method. Differences were considered significant when equal or less than p=0.05 using SPSS software (SPSS Inc., Chicago, IL).

100 CHAPTER III

RESULTS

ζ-potentials of microcapsules

As expected, ζ-potentials (measured in 10 mM HEPES, pH 7.4) of both 1.2 μm and 4.8 μm capsules corresponded to the charge of the outermost polymer layer (Table 1). Microcapsules with a positive polyelectrolyte as the top layer, PAH or PLL, had positive ζ-potentials of about 20 to 30 mV, while ζ-potentials of microcapsules terminated by PSS or PGA, were negative (-30 to -50 mV, depending on the different formulations). Assembly of the PEG-grafted copolymers on the oppositely charged microcapsules reduced the ζ-potential to values close to zero, with PLL-g-PEG on PSS-terminated microcapsules yielding slightly positive potentials (3 or 4 ± 1 mV) and PGA-g-PEG on PAH-terminated microcapsules approximately neutral or slightly negative potentials.

Table 1. ζ-potential of microcapsules in mV (± SD), measured in 10 mM HEPES buffer (pH 7.4).

Phagocytosis studies

To investigate whether coatings with the PEG-grafted polymers PGA-g-PEG and PLL-g-PEG would render 1.2 and 4.8 μm microcapsules resistant to phagocytosis, we incubated them with DC and MΦ cultures. As references we

CHAPTER III 101

Figure 1. Phagocytosis of coated microcapsules. DC (left) and MΦ (right) were incubated with 4.8 μm (left) or 1.2 μm (right) microcapsules for 4 hours. Images were obtained by an overlay of phase contrast and fluorescence channels. A, B: PSS coated microcapsules; C, D:PLL coated microcapsules; E, F: PLL-g-PEG coated microcapsules. Scale bars = 50 μm.

102 CHAPTER III

used the corresponding uncoated microcapsules, i.e. PAH terminated microcapsules for PGA-g-PEG and PSS terminated microcapsules for PLL-g-PEG, as well as microcapsules only coated with the backbone polymers PGA and PLL. First, cell culture samples were qualitatively evaluated by phase-contrast and fluorescence microscopy. Figure 1 conclusively shows the efficient phagocytosis of PSS- and PLL-terminated microcapsules (upper and middle row, respectively) in contrast to the PLL-g-PEG coated microcapsules that largely resisted internalization (bottom row). Cellular internalization or association of PLL-g-PEG coated microcapsules was minor, while the vast majority of the microcapsules remained freely dispersed throughout the well.

Further steps were carried out to investigate phagocytosis in more detail. Flow cytometric analyses (Figure 2) revealed no significant effect on uptake efficiency for PGA-g-PEG. There was only one exception in the case of 1.2 μm capsules incubated with DC, which was significant (p<0.01) compared with the PGA terminated control, though insignificant compared with the PAH terminated control. However, there was a marked drop in phagocytosis for PLL-g-PEG coated microcapsules (p<0.01). The effect was particularly pronounced for 4.8 μm capsules (Figure 2a). Here, internalization was reduced to about 10-15% of the control microcapsules carrying no PEG corona. For 1.2 μm capsules, the reduction was less distinct. We observed uptake rates that were decreased by about two thirds and half for MΦ and DC, respectively (Figure 2c). However, also in the case of 4.8 μm capsules, we observed individual microcapsule batches with a weaker impact of the PEG corona on uptake rates (Figure 2b). Here, the moderate stealth function was similar to that with 1.2 μm capsules. We attribute this observation to yet elusive batch-to-batch variations during microcapsule preparation, as will be discussed in detail in the Discussion section.

As flow cytometry per se cannot distinguish between internalized microcapsules and microcapsules that adhere to the outer cell membrane, all experiments were accompanied by control experiments carried out at 4 °C. Incubation at low temperature inhibits phagocytosis. Accordingly, signals collected at 4 °C corresponded to adhesion only, and were subtracted from the

CHAPTER III 103

values obtained at 37 °C. Overall, the contribution of microcapsule adhesion to the overall signal turned out to be negligible (data not shown) indicating efficient phagocytosis once cell contacts were established.

Figure 2. Phagocytosis of coated microcapsules by flow cytometry. DC (grey) and MΦ (black) were incubated with 4.8 μm capsules (A, B) or 1.2 μm capsules (C) for 4 hours. Phagocytosis was quantified by flow cytometry (mean fluorescence intensity ± SD). Significances are reported as ∗∗, p≤0.01; ∗, p≤0.05; n.s., not significant. A and C are the means of three independent experiments on cells from three different donors. B shows the representative results of a different microcapsule fabrication batch; the degree of reduction of phagocytotic uptake is less than for batches A (see Discussion section).

104 CHAPTER III

Figure 3. Phagocytosis of coated microcapsules by confocal microscopy. DC (left) and MΦ (right) were incubated with 4.8 (A-D) or 1.2 (E-H) μm capsules for 4 hours and stained for confocal microscopy. Images were obtained by an overlay of the red (capsules), green (membrane or cytoskeleton) and blue (nucleus) channels. Scale bars = 10 μm.

CHAPTER III 105

Internalization of microcapsules analyzed by confocal microscopy

In addition to the quantitative assessment of phagocytosis by flow cytometry above, we studied in more detail by CLSM whether the microparticles were internalized or only adhered to the outer cell membrane (Figure 3). Cells were incubated with TRITC-labelled microcapsules (red) and stained with Phalloidin-OregonGreen488 or ConA-AF488 (green) to label the cells´ cytoskeleton or membrane, respectively, so that the microcapsules appeared as red dots within the green outline of the cells. For further clarification, the nuclei were stained with Hoechst-33342 (blue). Both the x-y optical sections and the z-projections indicated massive internalization of both 1.2 or 4.8 μm microcapsules by both DC and MΦ. Microcapsules were found throughout the cells, except for the nucleus. Without performing further stainings, no information as to the subcellular localization of the microcapsules was available. Nevertheless, in a recent study microcapsules were found to internalize into acidic compartments, suggesting their transfer into phago-lysosomes [14]. Efficient phagocytosis of PS microspheres by DC and MΦ was previously demonstrated [37, 44].

Viability study

An important aspect of drug delivery systems is the potential cytotoxicity of the carrier. Therefore, we analyzed the viability of both DC and MΦ over up to three days after addition of 1.2 or 4.8 μm capsules. Even after three days, deviations of cell viability were statistically not significant compared to untreated control cells (Figure 4).

Stability of PLL-g-PEG coatings

In view of future biomedical applications, we were interested in the long-term stability of PLL-g-PEG coatings. Previously, PS microspheres were successfully implemented as model system for microparticles [37, 44]. They are monodisperse, stable and easy to isolate by centrifugation. Therefore, we first investigated the stability of PLL-g-PEG coatings on 5 μm PS spheres under two

106 CHAPTER III

Figure 4. Viability of MΦ (grey) and DC (black) was determined by PI staining after three days of incubation with PS microspheres (5 μm), PS/PLL-g-PEG, and PAH- or PSS-terminated microcapsules (1.2 or 4.8 μm). Control cells were stained without addition of microspheres or microcapsules. Data are presented as percentage of viable cells in the whole cell population (± SD). Significant differences from control cells are indicated as ∗∗, p≤0.01; ∗, p≤0.05; n.s., not significant.

different conditions: (i) when kept in 10 mM HEPES buffer (pH 7.4) at 4 °C or (ii) in RPMI medium supplemented with 5% human serum at 37 °C. Throughout the whole timeframe of the experiment, ζ-potentials did not show relevant changes, and resistance to phagocytosis remained unimpaired (Figure 5). There was only a slight decrease noted in the ζ-potentials of microspheres kept in cell culture medium as compared to those kept in buffer. This minor effect is most probably due to the fact that even repeated washings cannot completely remove salts and other components of the cell culture medium. No evidence was found for either polymer desorption or increasing cellular recognition as a consequence of the long-term exposure to buffer and serum. Thus, the PLL-g-PEG coating was concluded to be stable for at least four weeks, even at 37 °C and in the presence of serum proteins or physiological ionic strength. In agreement with

CHAPTER III 107

the results obtained from solid PS microspheres, we found both ζ-potentials and resistance to phagocytosis unchanged throughout a period of 3 weeks (Figure 6) for the PLL-g-PEG coated 4.8 μm capsules.

Figure 5. Stability of PLL-g-PEG coatings on PS microspheres. PS microspheres (5 μm) were coated with PLL-g-PEG and stored in 10 mM HEPES buffer at 4°C (light grey) or in cell culture medium (5% serum) at 37°C (dark grey). Controls were freshly coated (empty) and uncoated (black) microspheres. A: After storage for the indicated time, PLL-g-PEG coated microspheres and uncoated controls were incubated with MΦ for 4 hours. Phagocytosis was quantified by flow cytometry (mean SSC ± SD). B: Corresponding ζ-potentials in mV (± SD; measured in 10 mM HEPES).

108 CHAPTER III

Figure 6. Stability of PLL-g-PEG coatings on microcapsules. 4.8 μm PSS-terminated capsules were coated with PLL-g-PEG and stored in 10 mM HEPES buffer at 4°C. A: After storage for the indicated time, PLL-g-PEG coated capsules (grey) and PSS terminated controls (black) were incubated with MΦ for 4 hours. Phagocytosis was quantified by flow cytometry (mean fluorescence intensity ± SD). B: Corresponding ζ-potentials in mV (± SD; measured in 10 mM HEPES).

CHAPTER III 109

DISCUSSION

Prospective applications of layer-by-layer assembled, hollow polyelectrolyte microcapsules as drug delivery system require surface modifications that help to escape from unspecific clearance by the phagocytes of the MPS (such as DC or MΦ) and, thus, prolong residence time at the site of delivery. As the unspecific removal by professional phagocytes depends mainly on the adsorption of opsonic proteins, we expected the formation of a PEG corona immobilized via deposition of PEG grafted polyelectrolytes to protect the polyelectrolyte microcapsules from premature phagocytosis and clearance by the MPS. This aspect is of fundamental interest for any further development of such drug containers, e.g., if functionalized with bioligands for targeted delivery. It was previously shown that deposition of PLL-g-PEG onto PSS terminated microcapsules improved protein resistance by three orders of magnitude [31]. In the present study, we studied the effect of PEG grafted polyelectrolyte coatings on the recognition of microcapsules by professional phagocytic cells, i.e. DC and MΦ.

Both types of PEG grafted polymers, PLL-g-PEG and PGA-g-PEG, could be successfully adsorbed onto oppositely charged polyelectrolyte microcapsules, as evidenced by the microcapsules’ ζ-potential. Major differences between the adlayer coatings were observed in phagocytosis experiments. Here, PLL-g-PEG coatings on 4.8 μm capsules caused an efficient shielding of cellular recognition, as demonstrated by a drop in phagocytosis by up to 85-90%. These results are in good agreement with previous studies regarding effects of PLL-g-PEG-coatings on protein adsorption and phagocytosis [37, 44]. In control experiments using flow cytometry and CLSM, unmodified microcapsules were efficiently internalized. However, the degree of shielding from phagocytosis after PLL-g-PEG deposition was found to be subject to batch-to-batch variations in microcapsule fabrication. In general, two main factors could contribute to such variations: (i) mechanical instability of the microcapsules and (ii) their surface charge distribution. Both could be affected by the adsorption of traces of MF oligomers released during the dissolution of the core material. Measurements of ζ-potentials were not sensitive enough to detect meaningful differences in

110 CHAPTER III

nanoscale distributions of surface charges among those batches. To minimize the potential adsorption of MF oligomers, the core was always dissolved after an equal number of adsorbed layers, such that both MF oligomers and the microcapsule wall were positively charged [18]. The additional layer of PSS prior to PLL-g-PEG adsorption should further cover any adsorbed MF oligomers. However, even small defects in the final PEG-layer may lead to recognition by phagocytes. This conclusion was suggested by the observation that on some types of microspheres, which did not resist phagocytosis in spite of PLL-g-PEG coatings, fluorescently labelled protein adsorbed in a patchlike manner (U. Wattendorf, unpublished results).

A further point of concern may be the degree of MF crosslinking that typically increases over time during particle storage [18, 19, 45]. Accordingly, size and charge-density of the MF oligomers generated during the dissolution-process may vary among batches of different age and, as a consequence, show varying adsorption properties towards charged surfaces, such as the polyelectrolyte network of a microcapsule shell. In the present study, different batches of MF cores were utilized for the production of different microcapsule batches. Therefore, we assume that the observed inconsistencies could be due to chemical and mechanical changes in the microcapsule shell induced by an inevitable electrostatic adsorption of MF oligomers. In addition, non-dissolved MF core material remaining in the inner void of the microcapsule could cause local differences in thickness of the microcapsule’s shell. This would have major impact on the microcapsule’s integrity during the subsequent coating process, as it might cause unbalanced mechanical stress to the shell during centrifugation. Based on microscopic analysis, less than 5% of the microcapsules seemed to be deformed. However, as pointed out above, even submicron sized defects in the final PEG-layer may lead to recognition by phagocytes. Therefore, future experiments should involve other template cores that can be dissolved more efficiently than MF. Especially silica and carbonate template microparticles were recently shown to yield microcapsules of reproducible quality [14, 46, 47]. Both types of template cores could be completely dissolved, without affecting the integrity of the capsule wall or its mechanical properties.

CHAPTER III 111

With regard to particle-size, we found the stealth function of the PLL-g-PEG corona on the 1.2 μm capsules to be considerably less efficient as compared to 4.8 μm capsules. A possible explanation could be the batch-to-batch variation discussed above. However, it is also possible that our results may have been biased by the internalization mechanism. In general, microparticulates are internalized predominantly via phagocytosis, requiring interactions between the particulate and receptors on the cell membrane [48]. On the other hand, macropinocytosis does not require such interactions. While particulates in the size range of 5 μm are too big for other mechanisms than phagocytosis, 1 μm capsules could be additionally internalized via macropinocytosis, being characterized by the sampling of large volumes of extracellular milieu through fusion of membrane protrusions with the plasma membrane [48, 49]. Thus, if macropinocytosis was involved in the internalization of the 1 μm microcapsules, the adsorbed PEG corona would only interfere with the phagocytic mechanism, whereas a fraction of microcapsules undergoing macropinocytosis would be internalized independent of surface properties.

Interestingly, PGA-g-PEG coatings on PAH-terminated capsules of both sizes had no significant impact on the phagocytosis of coated microcapsules, although ζ-potential measurements clearly indicated successful coating and effective charge neutralization. Although one formulation (1.2 μm capsules in DC) was significantly different from one of the controls, it failed to be significantly different from the other control. In principle, one would expect similar effects as seen with PLL-g-PEG, given the fact that both copolymers adsorb via their charged backbones, with the PEG chains stretching out perpendicularly to the surface [32]. Nevertheless, considering the somewhat higher grafting ratio of PGA-g-PEG (g = 5.5) versus the optimum value [37] used with PLL-g-PEG (g = 3.5), some degree of opsonization could have occurred into hypothetic voids in-between the somehow less dense PEG corona of PGA-g-PEG. This contrasts with the observation of Pasche [50] that for 5 kDa PEG chains, PLL-g-PEG coatings with a grafting ratio of 5.5 could fully repel serum protein adsorption on flat surfaces. Another point to consider is the molecular weight of the PEG chains (5 kDa). While longer PEG chains are generally expected to improve protein resistance as compared to shorter chains

112 CHAPTER III

[51], they may as well lead to a drop in the adsorption of the copolymer due to steric repulsion [50]. Indeed, preliminary experiments on the adsorption of PLL-g-PEG and PGA-g-PEG onto flat niobium oxide surfaces using a quartz crystal microbalance (QCM-D) indicated that the adsorbed mass of PGA-g-PEG was only about 70-85% of the adsorbed mass of PLL-g-PEG (O. Kreft, unpublished results). Consequently, PEG density per surface area could be too low, which may explain the observed failure of protein resistance and subsequent cellular recognition and uptake. Finally, the higher molecular weight of the PGA backbone (50-100 kDa), as compared to the PLL backbone (20 kDa), may have contributed to the failure of the PGA-g-PEG copolymer. In the case of PLL-g-PEG copolymers with long (300 kDa) PLL backbones, it was proposed that the adsorption onto surfaces resulted in partial loop formation of the backbone, while copolymers with short (20 kDa) PLL backbones formed a smooth adlayer [52]. Although the molecular weight of the PGA backbone was still much lower, it is still possible that some loop formation occurred, possibly hindering the formation of a smooth and dense PEG corona. Overall, at this point a mechanistic interpretation of the observed contrasts between PLL-g-PEG and PGA-g-PEG is elusive. Whereas our data corroborates the previously found optimum architecture of the PLL-g[3.5]-PEG(2) copolymer [37], it also suggests that a PGA-g-PEG of similar architecture could lead to comparable shielding.

An important aspect of drug carrier systems is their effect on cell viability. Recent studies in different kidney cell models (such as HEK cells or NRK cells) reported negligible toxicity for polyelectrolyte microcapsules [17, 53]. Moreover, due to its low toxicity and immunogenicity PEG is generally considered as highly suitable for biomedical applications [26, 30]. Consistently, neither DC nor MΦ showed significantly impaired cell viability upon phagocytosis of the microcapsules, at least within the framework of this study.

Finally, we were interested in the stability of the coatings, a crucial aspect for many biomedical applications. Under cell culture conditions, PLL-g-PEG adlayers were previously found to be stable for at least two weeks on flat substrates made of tissue culture polystyrene or niobium oxide, while a much less stable adlayer formed on silicon oxide or titanium oxide surfaces [54]. The authors found cellular processes to be comparatively unimportant for the

CHAPTER III 113

stability of the adlayer, and believed that the observed decomposition of the coatings was due to desorption and exchange processes by unidentified serum components. Therefore, we assumed the stability of the PLL-g-PEG adlayer to be solely governed by the strength of its interaction with the substrate. To assess adlayer stability in our system, we first studied carboxylated PS microspheres (5 μm) under different conditions, i.e. in buffer at 4 °C and in serum supplemented cell culture medium at 37 °C. PS microspheres represent a successfully implemented model system [37, 44] because it is monodisperse, stable and easy to isolate by centrifugation. In fact, we found PLL-g-PEG adlayers on PS microspheres to be stable for nearly one month even in the presence of serum at physiological ionic strength and at elevated temperature. Similar stability was found for the PLL-g-PEG coated microcapsules.

Obviously, providing microcapsules with a stealth coating is only the first step towards receptor-mediated drug delivery. For future studies we suggest to functionalize the PEG corona with suitable ligands to ensure specific microcapsule recognition by target cells [3]. Recently, a biotin-functionalized PLL-g-PEG was immobilized onto polyelectrolyte microcapsules, and the availability of the biotin ligand was demonstrated through its recognition by streptavidin [31]. Moreover, the functionalization of PLL-g-PEG with the integrin targeting peptide sequence RGD, and its recognition by various cell types (fibroblasts, epithelial cells, or osteoblasts) was demonstrated through cell adhesion on coated substrates [35, 55]. What remains to be demonstrated is that adlayers of suitably functionalized PLL-g-PEG copolymers mediate specific recognition and uptake by target cells.

CONCLUSION

In this work, we could show that layer-by-layer assembled polyelectrolyte microcapsules can be effectively shielded from recognition of professional phagocytes by adsorption of a PEG-grafted polymer, PLL-g-PEG, to form a PEG corona. The coated microcapsules could be stored over several weeks without apparent loss of stealth function. Importantly, internalized

114 CHAPTER III

microcapsules did not measurably impair phagocyte viability. Considering their potential for drug loadings [56], PLL-g-PEG coated microcapsules appeared as a highly interesting drug delivery system for further investigation, e.g., as containers for localized drug delivery.

ACKNOWLEDGEMENTS

The financial support by the Volkswagen Foundation (I/80 051-054) is kindly acknowledged. We are grateful to Annegret Praast for assistance in microcapsule preparation and to Heidi Zastrow for SPLS measurements.

ABBREVIATIONS

The abbreviations used are PSS: polystyrene sulfonate; PAH: polyallylamine hydrochloride; MPS: mononuclear phagocytic system; DC: dendritic cells; MΦ: macrophages; PEG: polyethylene glycol; PLL: poly-L-lysine; PGA: poly-L-glutamic acid; PLL(x)-g[y]-PEG(z): copolymer of PEG (MW = z kDa) grafted to PLL (MW = x kDa) at a grafting ratio y; PGA(x)-g[y]-PEG(z): copolymer of PEG (MW = z kDa) grafted to PGA (MW = x kDa) at a grafting ratio y; grafting ratio, g: total number of lysine monomers divided by the number of PEG side chains; EDC: 1-ethyl-3-(3-dimethylamino-propyl) carbodiimide; NHS: N-hydroxysulfo succinimide sodium salt; TRITC: tetramethylrhodamine isothiocyanante; AF488: Alexa Fluor® 488; AF633: Alexa Fluor® 633; PS: polystyrene; MF: Melamine formaldehyde; HEPES buffer: 10 mM 4-(2-hydroxyethyl)piperazine-1-ethane-sulfonic acid, pH 7.4; PBS buffer: 10 mM phosphate buffered saline, pH 7.4; RPMI medium: RPMI 1640 (Roswell Park Memorial Institute) with L-glutamine; PI: propidium iodide; SSC: side scattering (measured by flow cytometry).

CHAPTER III 115

REFERENCES 1. Donath, E., Sukhorukov, G.B., Caruso, F., Davis, S.A., and Möhwald, H.,

Novel hollow polymer shells by colloid-templated assembly of polyelectrolytes. Angew. Chem., Int. Ed. Engl., 1998. 37(16): p. 2202-2205.

2. Peyratout, C.S., and Dähne, L., Tailor-made polyelectrolyte microcapsules: from multilayers to smart containers. Angew. Chem. Int. Ed. Engl., 2004. 43(29): p. 3762-3783.

3. Sukhorukov, G.B., Rogach, A.L., Zebli, B., Liedl, T., Skirtach, A.G., Köhler, K., Antipov, A.A., Gaponik, N., Susha, A.S., Winterhalter, M., and Parak, W.J., Nanoengineered polymer capsules: tools for detection, controlled delivery, and site-specific manipulation. Small, 2005. 1(2): p. 194-200.

4. Kataoka, K., Harada, A., and Nagasaki, Y., Block copolymer micelles for drug delivery: design, characterization and biological significance. Adv. Drug Deliv. Rev., 2001. 47(1): p. 113-131.

5. Torchilin, V.P., Targeted polymeric micelles for delivery of poorly soluble drugs. Cell. Mol. Life Sci., 2004. 61(19-20): p. 2549-2559.

6. Kikuchi, A., and Okano, T., Pulsatile drug release control using hydrogels. Adv. Drug Deliv. Rev., 2002. 54(1): p. 53-77.

7. Tonnesen, H.H., and Karlsen, J., Alginate in drug delivery systems. Drug Dev. Ind. Pharm., 2002. 28(6): p. 621-630.

8. Ai, H., Jones, S.A., de Villiers, M.M., and Lvov, Y.M., Nano-encapsulation of furosemide microcrystals for controlled drug release. J Control Release, 2003. 86(1): p. 59-68.

9. Antipov, A.A., Sukhorukov, G.B., Donath, E., and Möhwald, H., Sustained release properties of polyelectrolyte multilayer capsules. J. Phys. Chem. B, 2001. 105(12): p. 2281-2284.

10. Brown, J.Q., Srivastava, R., and McShane, M.J., Encapsulation of glucose oxidase and an oxygen-quenched fluorophore in polyelectrolyte-coated

116 CHAPTER III

calcium alginate microspheres as optical glucose sensor systems. Biosens. Bioelectron., 2005. 21(1): p. 212-216.

11. Tiourina, O.P., Antipov, A.A., Sukhorukov, G.B., Larionova, N.L., Lvov, Y., and Möhwald, H., Entrapment of alpha-chymotrypsin into hollow polyelectrolyte microcapsules. Macromolecular Bioscience, 2001. 1(5): p. 209-214.

12. Kreft, O., Georgieva, R., Bäumler, H., Steup, M., Müller-Röber, B., Sukhorukov, G.B., and Möhwald, H., Red blood cell templated polyelectrolyte capsules: A novel vehicle for the stable encapsulation of DNA and proteins. Macromol Rapid Comm, 2006. 27(6): p. 435-440.

13. Shchukin, D.G., Patel, A.A., Sukhorukov, G.B., and Lvov, Y.M., Nanoassembly of Biodegradable Microcapsules for DNA Encasing. J. Am. Chem. Soc., 2004. 126(11): p. 3374-3375.

14. Kreft, O., Javier, A.M., Sukhorukov, G.B., and Parak, W.J., Polymer microcapsules as mobile local pH-sensors. J. Mater. Chem., 2007: p. DOI: 10.1039/b705419j.

15. Munoz Javier, A., Kreft, O., Piera Alberola, A., Kirchner, C., Zebli, B., Susha, A.S., Horn, E., Kempter, S., Skirtach, A.G., Rogach, A.L., Rädler, J., Sukhorukov, G.B., Benoit, M., and Parak, W.J., Combined atomic force microscopy and optical microscopy measurements as a method to investigate particle uptake by cells. Small, 2006. 2(3): p. 394-400.

16. De Geest, B.G., Vandenbroucke, R.E., Guenther, A.M., Sukhorukov, G.B., Hennink, W.E., Sanders, N.N., Demeester, J., and De Smedt, S.C., Intracellularly degradable polyelectrolyte microcapsules. Advanced Materials (Weinheim, Germany), 2006. 18(8): p. 1005-1009.

17. Reibetanz, U., Claus, C., Typlt, E., Hofmann, J., and Donath, E., Defoliation and plasmid delivery with layer-by-layer coated colloids. Macromol Biosci, 2006. 6(2): p. 153-160.

18. Gao, C., Moya, S., Donath, E., and Möhwald, H., Melamine formaldehyde core decomposition as the key step controlling capsule integrity: Optimizing the polyelectrolyte capsule fabrication. Macromolecular Chemistry and Physics, 2002. 203(7): p. 953-960.

CHAPTER III 117

19. Gao, C., Moya, S., Lichtenfeld, H., Casoli, A., Fiedler, H., Donath, E., and Möhwald, H., The decomposition process of melamine formaldehyde cores: the key step in the fabrication of ultrathin polyelectrolyte multilayer capsules. Macromolecular Materials and Engineering, 2001. 286(6): p. 355-361.

20. Zebli, B., Susha, A.S., Sukhorukov, G.B., Rogach, A.L., and Parak, W.J., Magnetic targeting and cellular uptake of polymer microcapsules simultaneously functionalized with magnetic and luminescent nanocrystals. Langmuir, 2005. 21(10): p. 4262-4265.

21. Skirtach, A.G., Munoz Javier, A., Kreft, O., Köhler, K., Piera Alberola, A., Möhwald, H., Parak, W.J., and Sukhorukov, G.B., Laser-induced release of encapsulated materials inside living cells. Angew. Chem. Int. Ed. Engl., 2006. 45(28): p. 4612-4617.

22. Shenoy, D.B., Antipov, A.A., Sukhorukov, G.B., and Möhwald, H., Layer-by-Layer Engineering of Biocompatible, Decomposable Core-Shell Structures. Biomacromolecules, 2003. 4(2): p. 265-272.

23. Juliano, R.L., Factors affecting the clearance kinetics and tissue distribution of liposomes, microspheres and emulsions. Adv. Drug Deliv. Rev., 1988. 2(1): p. 31-54.

24. Ahsan, F., Rivas, I.P., Khan, M.A., and Torres Suárez, A.I., Targeting to macrophages: role of physicochemical properties of particulate carriers-liposomes and microspheres-on the phagocytosis by macrophages. J. Controlled Release, 2002. 79(1-3): p. 29-40.

25. Foged, C., Brodin, B., Frokjaer, S., and Sundblad, A., Particle size and surface charge affect particle uptake by human dendritic cells in an in vitro model. International Journal of Pharmaceutics, 2005. 298(2): p. 315-322.

26. Thiele, L., Rothen-Rutishauser, B., Jilek, S., Wunderli-Allenspach, H., Merkle, H.P., and Walter, E., Evaluation of particle uptake in human blood monocyte-derived cells in vitro. Does phagocytosis activity of dendritic cells measure up with macrophages? J. Controlled Release, 2001. 76(1-2): p. 59-71.

118 CHAPTER III

27. Thiele, L., Diederichs, J.E., Reszka, R., Merkle, H.P., and Walter, E., Competitive adsorption of serum proteins at microparticles affects phagocytosis by dendritic cells. Biomaterials, 2003. 24(8): p. 1409-1418.

28. Absolom, D.R., Opsonins and dysopsonins: an overview. Methods Enzymol., 1986. 132(Immunochem. Tech., Pt. J): p. 281-318.

29. Patel, H.M., Serum opsonins and liposomes: their interaction and opsonophagocytosis. Crit. Rev. Ther. Drug Carrier Syst., 1992. 9(1): p. 39-90.

30. Harris, J.M., Zalipsky, S., and Editors, Polyethylene glycol: Chemistry and Biological Applications. ACS Symp. Ser., 1997. 680: p. 489 pp.

31. Heuberger, R., Sukhorukov, G., Vörös, J., Textor, M., and Möhwald, H., Biofunctional polyelectrolyte multilayers and microcapsules: control of non-specific and bio-specific protein adsorption. Adv. Funct. Mater., 2005. 15(3): p. 357-366.

32. Huang, N.-P., Michel, R., Vörös, J., Textor, M., Hofer, R., Rossi, A., Elbert, D.L., Hubbell, J.A., and Spencer, N.D., Poly(L-lysine)-g-poly(ethylene glycol) Layers on Metal Oxide Surfaces: Surface-Analytical Characterization and Resistance to Serum and Fibrinogen Adsorption. Langmuir, 2001. 17(2): p. 489-498.

33. Tosatti, S., De Paul, S.M., Askendal, A., VandeVondele, S., Hubbell, J.A., Tengvall, P., and Textor, M., Peptide functionalized poly(L-lysine)-g-poly(ethylene glycol) on titanium: resistance to protein adsorption in full heparinized human blood plasma. Biomaterials, 2003. 24(27): p. 4949-4958.

34. Huang, N.-P., Vörös, J., De Paul, S.M., Textor, M., and Spencer, N.D., Biotin-derivatized poly(L-lysine)-g-poly(ethylene glycol): A novel polymeric interface for bioaffinity sensing. Langmuir, 2002. 18(1): p. 220-230.

35. VandeVondele, S., Vörös, J., and Hubbell, J.A., RGD-grafted poly-L-lysine-graft-(polyethylene glycol) copolymers block non-specific protein adsorption while promoting cell adhesion. Biotechnol. Bioeng., 2003. 82(7): p. 784-790.

CHAPTER III 119

36. Müller, M., Vörös, J., Csúcs, G., Walter, E., Danuser, G., Merkle, H.P., Spencer, N.D., and Textor, M., Surface modification of PLGA microspheres. J. Biomed. Mater. Res., Part A, 2003. 66A(1): p. 55-61.

37. Wattendorf, U., Koch, M.C., Walter, E., Vörös, J., Textor, M., and Merkle, H.P., Phagocytosis of poly(L-lysine)-graft-poly(ethylene glycol) coated microspheres by antigen presenting cells: Impact of grafting ratio and poly(ethylene glycol) chain length on cellular recognition. Biointerphases, 2006. 1(4): p. 123-133.

38. Boulmedais, F., Frisch, B., Etienne, O., Lavalle, P., Picart, C., Ogier, J., Voegel, J.C., Schaaf, P., and Egles, C., Polyelectrolyte multilayer films with pegylated polypeptides as a new type of anti-microbial protection for biomaterials. Biomaterials, 2004. 25(11): p. 2003-2011.

39. Klitzing, R., and Möhwald, H., Proton Concentration Profile in Ultrathin Polyelectrolyte Films. Langmuir, 1995. 11(9): p. 3554-3559.

40. Haugland, R.P., Handbook of Fluorescent Probes and Research Biochemicals. 2004, Eugene, OR, USA: Molecular Probes, Inc.

41. Lichtenfeld, H., Knapschinsky, L., Sonntag, H., and Shilov, V., Fast coagulation of nearly spherical ferric oxide (haematite) particles. Part 1. Formation and decomposition of aggregates: experimental estimation of velocity constants. Colloids Surf. A, 1995. 104(2/3): p. 313-320.

42. Sallusto, F., Cella, M., Danieli, C., and Lanzavecchia, A., Dendritic cells use macropinocytosis and the mannose receptor to concentrate macromolecules in the major histocompatibility complex class II compartment: downregulation by cytokines and bacterial products. J. Exp. Med., 1995. 182(2): p. 389-400.

43. Kiama, S.G., Cochand, L., Karlsson, L., Nicod, L.P., and Gehr, P., Evaluation of phagocytic activity in human monocyte-derived dendritic cells. Journal of aerosol medicine: official journal of the International Society for Aerosols in Medicine, 2001. 14(3): p. 289-299.

44. Faraasen, S., Vörös, J., Csúcs, G., Textor, M., Merkle, H.P., and Walter, E., Ligand-Specific Targeting of Microspheres to Phagocytes by Surface

120 CHAPTER III

Modification with Poly(L-Lysine)-Grafted Poly(Ethylene Glycol) Conjugate. Pharm. Res., 2003. 20(2): p. 237-246.

45. Gao, C., Donath, E., Möhwald, H., and Shen, J., Spontaneous deposition of water-soluble substances into microcapsules: phenomenon, mechanism, and application. Angew. Chem. Int. Ed. Engl., 2002. 41(20): p. 3789-3793.

46. Köhler, K., Shchukin, D.G., Möhwald, H., and Sukhorukov, G.B., Thermal behavior of polyelectrolyte multilayer microcapsules. 1. The effect of odd and even layer number. J Phys Chem B, 2005. 109(39): p. 18250-18259.

47. Antipov, A.A., Shchukin, D., Fedutik, Y., Petrov, A.I., Sukhorukov, G.B., and Möhwald, H., Carbonate microparticles for hollow polyelectrolyte capsules fabrication. Colloid Surface A, 2003. 224(1-3): p. 175-183.

48. Conner, S.D., and Schmid, S.L., Regulated portals of entry into the cell. Nature (London, United Kingdom), 2003. 422(6927): p. 37-44.

49. Kovacsovics-Bankowski, M., Clark, K., Benacerraf, B., and Rock, K.L., Efficient major histocompatibility complex class I presentation of exogenous antigen upon phagocytosis by macrophages. Proc. Natl. Acad. Sci. U. S. A., 1993. 90(11): p. 4942-4946.

50. Pasche, S., De Paul, S.M., Vörös, J., Spencer, N.D., and Textor, M., Poly(L-lysine)-graft-poly(ethylene glycol) assembled monolayers on Niobium Oxide surfaces: a quantitative study of the influence of polymer interfacial architecture on resistance to protein adsorption by ToF-SIMS and in situ OWLSu OWLS. Langmuir, 2003. 19(22): p. 9216-9225.

51. Kingshott, P., Thissen, H., and Griesser, H.J., Effects of cloud-point grafting, chain length, and density of PEG layers on competitive adsorption of ocular proteins. Biomaterials, 2002. 23(9): p. 2043-2056.

52. Wagner, M.S., Pasche, S., Castner, D.G., and Textor, M., Characterization of poly(L-lysine)-graft-poly(ethylene glycol) assembled monolayers on niobium pentoxide substrates using time-of-flight secondary ion mass spectrometry and multivariate analysis. Anal. Chem., 2004. 76(5): p. 1483-1492.

CHAPTER III 121

53. Kirchner, C., Javier, A.M., Susha, A.S., Rogach, A.L., Kreft, O., Sukhorukov, G.B., and Parak, W.J., Cytotoxicity of nanoparticle-loaded polymer capsules. Talanta, 2005. 67(3): p. 486-491.

54. Lussi, J.W., Falconnet, D., Hubbell, J.A., Textor, M., and Csucs, G., Pattern stability under cell culture conditions - A comparative study of patterning methods based on PLL-g-PEG background passivation. Biomaterials, 2006. 27(12): p. 2534-2541.

55. Schuler, M., Owen, G.R., Hamilton, D.W., de Wild, M., Textor, M., Brunette, D.M., and Tosatti, S.G., Biomimetic modification of titanium dental implant model surfaces using the RGDSP-peptide sequence: a cell morphology study. Biomaterials, 2006. 27(21): p. 4003-4015.

56. Sukhorukov, G.B., From polyelectrolyte capsules to drug carriers. ACS Symp. Ser., 2004. 879(Carrier-Based Drug Delivery): p. 249-266.

122 CHAPTER III

CHAPTER IV 123

CHAPTER IV

Impact of Mannose-Based Molecular Patterns on Human Antigen Presenting Cells: A Study with Stealth Microspheres

Uta Wattendorf,a Géraldine Coullerez,b Janos Vörös,c

Marcus Textor,b and Hans P. Merklea

a Institute for Pharmaceutical Sciences, ETH Zurich, 8093 Zurich, Switzerland b Laboratory for Surface Science and Technology, ETH Zurich, 8093 Zurich,

Switzerland c Laboratory for Biosensors and Bioelectronics, ETH Zurich, 8093 Zurich,

Switzerland.

124 CHAPTER IV

ABSTRACT

The targeting of antigen presenting cells has recently gained strong attention for both targeted vaccine delivery and immunomodulation. Here, we present a versatile technology to prepare surface modified microspheres that display various mannose-based ligands and graded ligand densities to target C-type lectin receptors (CLRs) on human dendritic cells (DC) and macrophages. Surface modification was by electrostatic assembly of poly(L-lysine)-graft-poly(ethylene glycol) (PLL-PEG) copolymers onto negatively charged microspheres forming a PEG corona with stealth character. Decoration with carbohydrate ligands was achieved (i) through PLL-PEG copolymers carrying small substructure mannoside ligands, such as mono- and tri-mannose, as terminal substitution of the PEG chains, or (ii) by secondary coating with mannan onto previously formed adlayers of PLL or a mix of PLL and PLL-PEG. Microspheres carrying mannoside ligands were also studied in combination with an integrin targeting RGD peptide ligand. Due to the presence of a PEG (or mannan) corona such microspheres were protected against protein adsorption and opsonization thus allowing the formation of specific ligand-receptor interactions. Mannose density was the major factor for the phagocytosis of mannoside decorated microspheres, though with limited efficiency. This strengthens the recent hypothesis by other authors that the mannose receptor (MR) only acts as a phagocytic receptor when in conjunction with yet unidentified partner receptor(s). Finally, analysis of DC surface markers for maturation revealed that neither surface assembled mannan nor mannoside modified surfaces on the microspheres could stimulate DC maturation. Thus, phagocytosis upon recognition by CLRs alone does not trigger DC activation towards a T helper response. The microparticulate platform established in this work represents a promising tool for systematic investigations of specific ligand-receptor interactions on phagocytes, including the screening for potential ligands and ligand combinations.

CHAPTER IV 125

INTRODUCTION

The targeting of receptors on antigen presenting cells (APCs) has recently gained much attention for both vaccine delivery and immunomodulation [1]. Among the many APC receptors, such as Toll-like receptors (TLR), C-type lectins (CLR), scavenger receptors (SR) or immunoglobulin receptors (FcR), especially TLRs and CLRs have been extensively studied for their capacity to recognize pathogen associated molecular patterns (PAMPs) [2, 3]. Hence, such APC receptors have been referred to as pathogen-recognition receptors (PRRs). PAMPs are invariant structures that are unique to pathogens (and are not produced by the host) and include diverse molecular structures such as lipopolysaccharides (LPS), mannans, aldehyde-derivatized proteins, and bacterial DNA, among others [4, 5]. One may hypothesize that the cross-talk between a vaccine delivery system decorated with PAMP like motives and APC receptors could be a strategy to affect type and extent of APC maturation. Here we challenge this hypothesis using stealth microparticles with mannose enriched surface coatings.

The CLR family comprises a number of PRRs that bind to carbohydrate ligands in the presence of Ca2+ ions. The calcium dependent primary binding site is selective for a single monosaccharide residue, although variations especially in C-2 position of the monosaccharide are often accepted. However, secondary binding sites exist adjacent to the primary site that strongly enhance the binding affinity and fine-tune the specificity for certain types of linkages and substitutions [6].

On APCs, two CLRs have been characterized as mannose binding, namely the mannose receptor (MR) and DC-SIGN (dendritic cell-specific ICAM-3 grabbing non-integrin) [7]. From these two receptors, MR mainly recognizes terminal mono-mannose or di-mannose [8, 9]. On the contrary, DC-SIGN has high affinity for internal (rather than terminal) sugars within more complex mannose structures, including mannan, as well as several fucose containing oligosaccharides (i.e. blood group substances and Lewis antigens) [7, 10, 11]. Moreover, α(1,3)-α(1,6)-mannotriose was shown to bind to DC-SIGN, however with relatively low affinity [12]. It was recently demonstrated that high

126 CHAPTER IV

affinity binding to DC-SIGN requires additional interactions of the lectin with terminal α(1,2)-mannose moieties [13]. The ligand specificity of Dectin-2, another member of the CLR family, is still controversial. Recently, however, McGreal et al. presented evidence for specific recognition of high-mannose structures (Man9GlcNAc2 > Man8GlcNAc2 > Man7GlcNAc2) [14]. Targeting CLRs on APCs was identified as a promising strategy for immunomodulation, as documented by several studies showing that immune responses were enhanced or modified by coupling mannose type ligands, such as mannose moieties or mannan, to proteins or particulates [15-20].

However, most of these approaches used either soluble protein with covalently coupled ligands, or ligand-decorated liposomal carriers. Thus, there were multiple opportunities for ligand-receptor interactions, e.g. through proteinic epitopes or opsonization of the carrier. Hence, it is often difficult to specify in detail whether an observed effect on phagocytosis or immunomodulation is exclusively due to a specific ligand-receptor interaction, or results from a combination of more than a single recognition signal. Therefore, the aim of this work was to develop a versatile microparticulate platform that allowed to study recognition signals of APCs in more detail, displaying various ligands on the surface of a stealth microsphere. Microparticulates of about 1 to 10 µm in diameter are logical targeting systems for vaccines and immunomodulators, as they undergo rapid phagocytosis by APCs, an exclusive feature of phagocytes such as dendritic cells (DC) and macrophages (MΦ).

Poly(ethylene glycol) (PEG) coatings of appropriate architecture are commonly used to prepare microspheres with stealth surfaces that are repellent to serum protein adsorption and cellular recognition. It is the uncharged and highly hydrophilic character of such a PEG corona and its low toxicity and immunogenicity that makes it most suitable for biomedical applications [21]. Negatively charged surfaces can be conveniently coated with a PEG corona by electrostatic adsorption of poly(L-lysine)-graft-poly(ethylene glycol) copolymers, here further denoted as PLL-PEG, consisting of a polycationic PLL backbone whose side chain amino groups are partly grafted with PEG chains

CHAPTER IV 127

[22, 23]. The degree of PEG substitution of such polymers is indicated by the grafting ratio g, which is defined as the number of L-lysine monomers divided by the number of PEG side chains. Due to its positive charge at physiological pH, the PLL backbone adsorbs strongly onto negatively charged substrates through electrostatic interaction, forming a less than 10 nm thick adlayer of PLL-PEG stretching its PEG chains out perpendicularly to the surface [23, 24]. In fact, we could recently show that polystyrene (PS) microspheres which were surface coated with PLL(20)-[3.5]-PEG(2) - i.e. a copolymer with a 20 kDa PLL backbone, 2 kDa PEG side chains, and a grafting ratio of approximately 3.5 - were efficiently shielded from both serum protein adsorption and recognition by phagocytes [25].

It has been further shown that the PEG side chains of PLL-PEG can be terminally functionalized with bioligands, such as adhesion peptides or biotin, while retaining the repellent character of its adlayer to unspecific protein adsorption [26-28]. Most recently, terminal subsitution of the PEG side chains has been also demonstrated for selected carbohydrates [29].

Here we investigate two prototype microsphere formulations for their capacity to address APC receptors, the first coated with stealth adlayers of mannoside substituted PLL-PEG, the second coated with mixtures of PLL and PLL-PEG and then functionalized with oligomannose by adsorption of mannan from S. cerevisiae (Figure 1). Mannan is a polysaccharide PAMP consisting of an α(1,6) linked mannose backbone grafted with α(1,2) and α(1,3) linked mannose side chains, some of which contain phosphate groups [30]. Mannan is assumed to be the major antigen of yeast cell walls [31] and is a well-established multivalent ligand for the receptors TLR4 and DC-SIGN [7, 32]. We hypothesized that the presentation of mannoside ligands linked to the surface of microspheres at high densities via flexible PEG chains could be a strategy to mimic high molecular weight mannose based PAMPs with synthetic small molecular ligands.

Aspects that will be covered in this work are: (i) the formulation of the two microparticulate prototypes, (ii) the availability of the surface exposed mannose-based ligands to lectin binding, (iii) the capacity of the formed

128 CHAPTER IV

Figure 1. Schematic representation of carbohydrate modified microspheres. Carboxylated PS microspheres were surface modified through adsorption of a PLL-PEG-mannoside copolymer (A), or sequential adsorption of PLL and mannan (B). In both cases, thecarbohydrate density on the microspheres’ surface could be controlled through theadsorption of polymer mixtures with (ligand-free) PLL-PEG (C, D). The chemical structures of PLL-PEG-mannoside and mannan are indicated in E and F, respectively.

CHAPTER IV 129

adlayers to repel protein adsorption, (iv) the ζ-potentials of such particles, and, in particular, (v) the specific recognition and phagocytosis of the microparticulates by both DC and MΦ as APCs, as well as (vi) potential effects of mannoside substitutions on APC maturation. Although specific recognition of the two prototype microspheres could be demonstrated, we will reject the paradigm that by themselves mannose-based molecular patterns on microparticulates are recognized as PAMP like motives.

MATERIALS AND METHODS Materials

PLL(20)-g[3.5]-PEG(2), PLL(20)-g[4.0]-PEG(2)/PEG(3.4)-RGD(8%) and PLL(20)-g[3.7]-PEG(2)/PEG(3.4)-RDG(9%) were obtained from SurfaceSolutionS GmbH (Zurich, Switzerland) and consisted of a PLL backbone of 20 kDa with grafted PEG chains of 2 kDa and a grafting ratio (which specifies the total number of lysine monomers divided by the number of PEG side chains) of g = 3.5 – 4.0. The notation used for ligand modified copolymers PLL-g-PEG(2)/PEG(3.4)-ligand(x%) indicates the percentage of ligand covalently coupled to 3.4 kDa PEG side chains relative to the total number of PEG chains as determined by NMR. The full sequences of the coupled peptides were N-acetyl-GCRGYGRGDSPG-amide (biologically active peptide) and N-acetyl-GCRGYGRDGSPG-amide (inactive control). For simplicity and readability of this paper the polymers are henceforth abbreviated as PLL-PEG, PLL-PEG-RGD and PLL-PEG-RDG, respectively. As microparticulates we used negatively charged, carboxylated polystyrene (PS) microspheres (5 μm diameter) from Micromod Partikeltechnologie GmbH, Rostock, Germany. All chemicals used in this study were of analytical grade (from Fluka, Buchs, Switzerland) unless otherwise specified. Ultrapure water (NANOpure DiamondTM, Skan AG, Allschwil, Switzerland) was used for buffer preparation.

130 CHAPTER IV

Synthesis of mannoside substituted PLL-PEG copolymers

Sulfhydryl-tri(ethylene glycol) derivatized mannose and branched α(1,3)-α(1,6)-trimannose were prepared by methods analogous to those described in the literature by Ratner et al. [33] and kindly provided by Riccardo Castelli of Prof. Seeberger’s lab at ETH Zurich. Mannose elongated with a sulfhydryl-propyl linker was synthesized by allylation reaction as described by Houseman et al. [34]

The mannoside ligand (1.2 equiv.) dissolved in 50 mM sodium tetraborate buffer at pH 8.5 (∼1.1 wt%) was stirred for 1 h with tris(2-carboxyethyl)phosphine hydrochloride (TCEP) (1 equiv.) to reduce disulfide bonds. An α-,ω-functionalized PEG (3.4 kDa) with maleimide and N-hydroxysuccinimide functional groups (MAL-PEG-NHS, 1 equiv.) dissolved in anhydrous diglyme (∼9.5 wt%) was then added followed by a solution of poly-

L-lysine hydrobromide (PLL-HBr, 20 kDa, ∼ 2.5 wt%) in sodium tetraborate buffer. After reacting for ca. 4 h at room temperature, the appropriate amount of methoxy-PEG succinimidyl propionate (mPEG-SPA, 2 kDa) powder was added to the solution to adjust the overall lysine to PEG grafting ratio. After stirring for an additional 12 h, the solution was dialyzed to remove unreacted PEGs (molecular weight cutoff 6-8 kDa, Spectrum Laboratories, Inc.) and lyophilized against Millipore water to get a white powder. The polymers were synthesized with various molar ratios of PEG-mannoside to methoxy-PEG, targeting nominal percentages of mannoside-functionalized PEG of 20, 50, or 100%. A polymer architecture with a grafting ratio of approximately 3.5 was chosen as it is known to provide a highly protein resistant adsorbed layer [24, 25]. The degrees of substitution were adjusted by the variation of the molar ratios of the PEGs to PLL-HBr and of MAL-PEG-NHS to mPEG-SPA. The products were recovered with yields of 63%, 67%, and 73% for propyl-mannose functionalized polymers with a nominal functionalization of 20%, 50%, and 100%, respectively; and with yields of 56% and 77% for (ethylene glycol)-derivatized mono- and tri-mannose with a nominal functionalization of 50%. The polymers were stored at -20°C until further use.

CHAPTER IV 131

For simplicity and readability of this paper, the PLL-g-PEG(2)/PEG(3.4)-λmannoside(x%) polymers are henceforth abbreviated as PLL-PEG-λMan(x%) and PLL-PEG-EGTri-Man(x%), with x% indicating the experimentally determined ratio between mannoside coupled PEG chains (3.4 kDa) and unmodified PEG chains (2 kDa), and λ indicating the respective linker, propyl (Pr) or ethylene glycol (EG). The term PLL-PEG-mannoside will be used to refer to all glycopolymers collectively.

Characterization of mannoside substituted PLL-PEG copolymers Characterization by 1H-NMR.

The lyophilized glycopolymers were dissolved in D2O and the spectra were recorded on 300 MHz and 500 MHz spectrometers (Bruker). Recordings were at room temperature (300 K) and 323 K. The conformational change of the polymers with the increase in temperature induced a shift of the resonance of the anomeric proton of the mannose to 5.2 ppm apart from the D2O resonance (4.8 ppm), which allowed for quantification of the carbohydrate moieties. Main PLL-PEG peaks were assigned according to previously published spectra of PLL-g-[PEG/PEG-biotin] [26]. The degrees of mannoside substitution of the synthesized glycopolymers are given in Table 1.

In situ analysis of polymer and protein adsorption on flat surfaces by optical methods.

Dual Polarization Interferometry (DPI) and Optical Waveguide Lightmode Spectrometry (OWLS) have been used to monitor in situ the adsorbed mass of glycopolymers, of Concanavalin A, and of human serum proteins. For all PLL-PEG-mannoside polymers, adsorption of serum protein was below 10 ng/cm2. Compared to untreated niobium oxide control surfaces (590 ± 57 ng/cm2) [24], protein adsorption was reduced by more than 98%. Thus, all polymers displayed a high degree of protein repellence. Further information on methodology and results is given as supporting information in Appendix III.

132 CHAPTER IV

Table 1. Degree of substitution of the synthesized PLL-PEG glycopolymers determined by 1H-NMR.

Coating of microspheres

For surface coatings with ligand linked PLL-PEGs, PS microspheres (final conc. 2.5 mg/ml) were dispersed under aseptic conditions in sterile filtered PLL-PEG-ligand solutions (final conc. 0.5 mg/ml in 10 mM HEPES buffer, pH 7.4; sterile filtered) and incubated under gentle mixing for 20 min at room temperature. Surface-coated microspheres were centrifuged (5 min, 2655 x g; Centrifuge 5417R, Eppendorf-Netheler-Hinz GmbH, Hamburg, Germany) and redispersed in sterile filtered HEPES buffer.

For coatings with mannan from Saccharomyces cerevisiae (MW 34-62.5 kDa; Sigma, Buchs, Switzerland), PS microspheres were first coated with defined mixtures of PLL and PLL-PEG following the same protocol as with ligand linked PLL-PEGs (see above). After washing, microspheres (final conc. 2.5 mg/ml) were subjected to a secondary coating with mannan (final conc. 0.5 mg/ml in 10 mM HEPES buffer, pH 7.4; sterile filtered) for 20 min at room temperature, centrifuged and redispersed in sterile filtered HEPES buffer.

CHAPTER IV 133

Negative controls were prepared by replacement of mannan with D-(+)-mannose (Sigma, Buchs, Switzerland) instead of the secondary coating with mannan. The exact compositions of the investigated formulations will be encoded as: a PLL:b PLL-PEG, a PLL:b PLL-PEG/Mannan, and a PLL:b PLL-PEG/Mannose, with a indicating the percentage (w/w) of PLL, and b the percentage (w/w) of PLL-PEG, further with /Mannan denoting the secondary coating with mannan, and with /Mannose representing a secondary coating with mannose as indicated above.

Availability of mannose on coated microspheres

To assess the availability of mannose residues on surface coated PS microspheres we used fluorescence labelled Concanavalin-A (ConA-AlexaFluor488, Molecular Probes Inc., Eugene, OR, USA) [35]. Microspheres (final conc. 1 mg/ml) were incubated with ConA-AF488 (final conc. 0.3 mg/ml in 10 mM HEPES buffer supplemented with 150 mM sodium chloride, 1 mM calcium chloride and 1 mM magnesium chloride) for 45 minutes in the dark, centrifuged (5 min, 10,000 g, Centrifuge 5417R, Eppendorf-Netheler-Hinz GmbH, Hamburg, Germany), washed and redispersed in filtered HEPES buffer. Fluorescence of the microspheres as a consequence of ConA adsorption was determined in a flow cytometer (FACSCanto, BD Biosciences, Basel, Switzerland). For data analysis, mean fluorescence values were calculated with FlowJo software (v7.1, Tree Star Inc., Ashland, OR, USA).

Protein repellence of coated microspheres

To determine the protein repellence of surface coated PS microspheres we used fluorescently labelled bovine serum albumin (BSA; BSA-TxRed, Molecular Probes Inc., Eugene, OR, USA). Microspheres (final conc. 1 mg/ml) were incubated with BSA-TxRed (final conc. 1 mg/ml in 10 mM HEPES buffer) for 45 minutes in the dark, centrifuged (5 min, 10,000 g, Centrifuge 5417R, Eppendorf-Netheler-Hinz GmbH, Hamburg, Germany), washed and redispersed in filtered HEPES buffer. Additional experiments were performed in HEPES buffer supplemented with 150 mM NaCl to simulate physiological ionic

134 CHAPTER IV

strength. Fluorescence of the microspheres as a consequence of BSA adsorption was determined in a flow cytometer (FACSCanto, BD Biosciences, Basel, Switzerland). For data analysis, mean fluorescence values were calculated with FlowJo software (v7.1, Tree Star Inc., Ashland, OR, USA).

ζ-potential of coated microspheres

Coated and uncoated microspheres (65 μg/ml) were dispersed in filtered HEPES buffer (10 mM, pH 7.4). Measurements were performed on a Zetasizer 3000 HS (Malvern Instruments, Worcestershire, UK; five measurements per sample).

DC and MΦ cell culture

DC and MΦ were obtained from human peripheral blood monocytes according to Sallusto et al. [36] Briefly, peripheral blood monocytes were isolated from buffy coats (Bloodbank Zurich, Zurich, Switzerland) by density gradient centrifugation of Ficoll-PaqueTM PLUS (Amersham Biosciences, Uppsala, Sweden). Isolated peripheral blood monocytes were washed in PBS (Sigma, Buchs, Switzerland) and resuspended in RPMI 1640 + L-glutamine (Invitrogen, Basel, Switzerland) supplemented with 10% heat-inactivated (pooled) human serum (Bloodbank Zurich, Zurich, Switzerland), 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen, Basel, Switzerland) and then allowed to adhere for 2 h in culture flasks (25 cm2). Nonadherent cells were removed whereas adherent cells were further cultured in RPMI medium supplemented with 5% heat-inactivated (pooled) human serum, 100 U/ml penicillin and 100 μg/ml streptomycin in the presence of 1000 IU/ml interleukin-4 (Sigma, Buchs, Switzerland) and 50 ng/ml human granulocyte-macrophage colony stimulating factor (R+D Systems Europe Ltd., Abingdon, UK) to obtain DC. MΦ were obtained without additional supplements [37]. These media will be further denoted as supplemented RPMI media. For cell culture, DC and MΦ were kept at 37°C in 5% CO2-humidified atmosphere. For phagocytosis studies, cells were mechanically removed from the flasks after 24 h and reseeded into 24-well plates (300,000 cells per well).

CHAPTER IV 135

Phagocytosis studies

Rates of phagocytosis were studied with both DC and MΦ. 50 μg microspheres dispersed in filtered HEPES buffer (10mM, pH 7.4) were added to either DC or MΦ cultured in 24-well plates (3 x 105 cells per well) and incubated at 37°C in 5% CO2-humidified atmosphere for 4 h. Control samples were incubated at 4°C in parallel to identify microspheres adsorbed to the outer cell membrane. Cells were immediately inspected for phagocytosis by phase contrast microscopy (Axiovert 35, Zeiss, Germany). Following microscopic analysis, the cells were mechanically recovered in a test tube and additionally evaluated by flow cytometry (FACSCanto, BD Biosciences, Basel, Switzerland). For data analysis, mean SSC values were calculated with FlowJo software (v7.1, Tree Star Inc., Ashland, OR, USA) as a measure for the number of microspheres per cell, as established and validated previously [25]. Values obtained at 4°C were subtracted from the corresponding values at 37°C to exclude microspheres that were only adsorbed to the outer cell membrane but not internalized. All experiments were repeated with cultures from three different donors.

Intracellular localization of phagocytosed microspheres by confocal microscopy

To confirm the intracellular localization of phagocytosed microspheres by confocal laser scanning microscopy (CLSM), cells were cultured in Lab-Tek 8 well chamber glass slides (Nalge Nunc Int., IL, USA) at a concentration of 200,000 cells per well. Phagocytosis experiments were performed as described above. However, instead of unlabelled microspheres we used rhodamine labelled PS microspheres (surface labelled; Micromod Partikeltechnologie GmbH, Rostock, Germany). After incubation with the microspheres, cells were washed three times with 10 mM PBS (Sigma, Buchs, Switzerland) and fixed with paraformaldehyde (3% in PBS; Sigma, Buchs, Switzerland) for 15 minutes. To label the actin cytoskeleton, the cells were washed with PBS, permeabilized with Triton-X (0.1% in PBS supplemented with 1% BSA) for 20 minutes and stained with Phalloidin-OregonGreen488 (6 nM in PBS with 1% BSA;

136 CHAPTER IV

Molecular Probes Inc., Eugene, OR, USA) for 20 minutes. After three washings with PBS, the cells’ nuclei were stained with 1 μg/ml Hoechst-33342 (Molecular Probes Inc., Eugene, OR, USA) in PBS for 5 minutes. The samples were washed three times and mounted for CLSM using the ProLong Gold antifade reagent from Molecular Probes (Eugene, OR, USA). Samples were examined using confocal microscopy (CLSM; Zeiss LSM 510 META (Carl Zeiss AG, Feldbach, Switzerland). Image processing was performed with Imaris (Bitplane, Zurich, Switzerland), a 3-D multi-channel image processing software for CLSM images.

Characterization of DC and assessment of maturation

Differentiation into DC was confirmed by analysis of surface marker expression. Cells were incubated for 45 min on ice in 1% serum supplemented RPMI medium with the following antibodies: CD11c-PE (IgG1k), CD14-FITC (IgG2a), CD83-FITC (IgG1k), CD86-FITC (IgG1k), MHC II-FITC (IgG2a) and the corresponding isotope matched control antibodies (all mouse anti-human, BD Biosciences, Basel, Switzerland). Samples were washed, fixed with 2% neutral buffered formalin solution (Sigma, Buchs, Switzerland) and analyzed by flow cytometry (FACSCanto, BD Biosciences, Basel, Switzerland). Data analysis was performed with FlowJo software (v7.1, Tree Star Inc., Ashland, OR, USA). For further identification, cells were also challenged with lipopolysaccharide (2.5 μg/ml, Escherichia coli 055:B5, Sigma, Buchs, Switzerland) 3 days before the antibody labelling, which causes maturation of DC. More than 95% of the cells were identified as CD11c+ CD14- DC. To assess the ability of coated microspheres to stimulate DC maturation, DC were incubated with the respective microsphere formulation for 3 days and stained for surface antigen expression following the above described protocol. Controls were performed with control microspheres missing the outer coating layer, and with soluble polymers. All experiments were repeated in cells from three different donors.

CHAPTER IV 137

Cell Viability

Cell viability after 3 days of microsphere incubation was assessed by staining with propidium iodide (PI, BD Biosciences, Basel, Switzerland). PI detects dead cells that were subject to necrosis or apoptosis. Phagocytosis experiments were performed as described above. PI staining was performed according to the manufacturer’s protocol. In brief, cells were washed with cold PBS (Sigma, Buchs, Switzerland), and incubated with 5 μl of PI per 100,000 cells for 10 minutes on ice. 200 μl of cold binding buffer was added and samples were analyzed within 1 hour by flow cytometry (FACSCanto, BD Biosciences, Basel, Switzerland). For positive controls, cells were fixed with 10% neutral buffered formalin solution (Sigma, Buchs, Switzerland) before PI staining. For data analysis, cells were gated for the DC population, and percentages of positively stained cells were calculated with the instrument’s BD FACSDivaTM software, using untreated cells as background control.

Significance Testing

For evaluation of statistical significance, samples were evaluated by ANOVA followed by a post-hoc assessment using the Tukey HSD and LSD methods. Differences were considered significant when equal or less than p=0.05 using SPSS software (SPSS Inc., Chicago, IL).

138 CHAPTER IV

RESULTS

This work aimed at the formulation and testing of a microsphere based delivery system with the capacity to specifically target selected phagocytic receptors on antigen presenting cells. Two prototypes of microparticulates were used, both based on physical adsorption of mannoside functionalized polyelectrolytes onto 5 μm carboxylated and thus negatively charged PS microspheres (Figure 1). The first prototype was prepared through adsorption of a PEGylated PLL graft copolymer, PLL-PEG, that was further functionalized with mono- or tri-mannose groups coupled to the free termini of the PEG-chains. The second prototype consisted of PS microspheres that were first coated with PLL and PLL-PEG adlayers which were subsequently functionalized by incubation with mannan (from S. cerevisiae), a polysaccharide consisting of an α(1,6) linked mannose backbone grafted with α(1,2) and α(1,3) linked mannose side chains, some of which carrying phosphate groups [30]. Owing to the negative charge of the phosphate groups, mannan was readily adsorbed onto positively charged (PLL coated) microspheres but not onto uncoated (carboxylated) PS microspheres (data not shown).

For optimization, surface densities of the ligand(s) were systematically varied for both prototypes. For microspheres coated with mannoside substituted PLL-PEG, ligand density was either varied by the degree of mannoside substitution on PLL-PEG copolymers, or by mixing mannoside substituted PLL-PEG at defined ratios with ligand-free copolymers. On the other hand, for the mannan coated microspheres, the degree of mannan assembly was controlled by the primary coatings with defined mixtures of PLL and (ligand-free) PLL-PEG. Because PEGylated surface domains protect from adsorption of mannan, the proportions of PLL and PLL-PEG used for the primary coatings determined the degree of mannan adsorption during the secondary coating.

Availability of mannose type carbohydrates on coated microspheres

To assess the exposure of the carbohydrate moieties on surface-coated microspheres, we incubated them with fluorescently labelled Concanavalin A (ConA), which specifically recognizes mannose residues [35]. For mannan

CHAPTER IV 139

coated microspheres (encoded by a PLL:b PLL-PEG/Mannan), we observed an approximately linear increase of ConA binding with increasing fractions of PLL, until a plateau was reached at a ratio of about 45% PLL, i.e. for 45 PLL:55 PLL-PEG/Mannan (Figure 2 A). Even 15% PLL in the primary coating solution was sufficient to demonstrate the presence of mannan on the microspheres´ surface. Control experiments using microspheres without secondary mannan coating confirmed that ConA adsorption was indeed due to the presence of mannan and not due to unspecific adsorption to PLL. Similarly, upon mannose instead of mannan treatments (a PLL:b PLL-PEG/Mannose) no increase in ConA adsorption relative to control microspheres was found, indicating very low, if any, adsorption of mannose (data not shown).

In the case of mannoside-substituted PLL-PEG, we found a general trend for increasing ConA adsorption on coated microspheres with increasing degrees of mannoside substitution of the polymers (Figure 2 B). Increasing densities of mannoside substitution from 0% to 28%, 34% and 45% resulted in a stepwise increase in ConA binding. Similarly, for microspheres coated with mixtures of any of the copolymers with unmodified PLL-PEG we found a linear dependence of ConA binding on the mixing ratio. After coating with PLL-PEG-EGTri-Man(45%) the amount of ConA bound to the microspheres was more than doubled, as compared to the highest values reached with the mono-mannose substituted copolymers, reflecting a higher degree of multivalent binding of ConA [38]. In addition, the efficiency of ConA binding also depended on the linker: EG-linked mannose was more weakly recognized in comparison with Pr-linked mannose.

The results of the ConA binding assay were further corroborated by measurements of ConA binding onto coated flat niobium oxide waveguides. A strong correlation was found between the results of the two complementary assays (Figure 3).

In summary, the carbohydrate moieties of both the mannan and the mannoside-substituted PLL-PEG adlayers could be successfully immobilized on the microspheres´ surface at quantitatively controlled, graded surface densities and were readily available for specific binding by ConA.

140 CHAPTER IV

Figure 2. Availability of mannose type carbohydrates on coated microspheres. Surface modified microspheres presenting mannosides or mannan were incubated with ConA to prove the binding availability of the ligands. Specific binding of the fluorescently labeled lectin was measured by flow cytometry and presented as arbitrary units [± SD]. A: Mannan coated microspheres; ConA adsorption is plotted vs. the ratio of PLL in the mixture with PLL-PEG during the primary adsorption step, which was followed by the adsorption of mannan. B: PLL-PEG-mannose and PLL-PEG-trimannose coated micropsheres; ConA adsorption is plotted vs. the ratio of the respective PLL-PEG-mannose in the mixture with ligand free PLL-PEG.

Figure 3. Correlation of ConA adsorption on microspheres vs flat waveguides. ConA adsorption on PLL-PEG-mannose coated PS microspheres [a.u. ± SD] was plotted versus ConA adsorption on coated flat niobium oxide waveguides [ng/cm2; single measurements]. A strong correlation between the two complementary assays was obtained.

CHAPTER IV 141

Repellence of coated microspheres to BSA adsorption

To assess the repellence of the microspheres against serum protein adsorption we incubated surface coated microspheres, in analogy to an earlier approach [25], with fluorescently labelled BSA as a model protein at low ionic strength (10 mM HEPES buffer). Similar to uncoated (carboxylated PS) microspheres, high amounts of BSA adsorbed onto PLL coated microspheres (Figure 4 A). In contrast, microspheres with an additional coating of mannan led

Figure 4. Adsorption of BSA on coated microspheres. Surface modified microspheres were incubated with BSA-TxRed to assess the repellence to protein adsorption. Adsorption of thefluorescently labeled BSA was measured by flow cytometry and presented as arbitrary units(± SD). A: Mannan coated microspheres (10 mM HEPES, pH 7.4); B: PLL-PEG-mannose and PLL-PEG-RGD coated microspheres (10 mM HEPES, pH 7.4); C: PLL-PEG-mannose coated microspheres (10 mM HEPES + 150 mM NaCl, pH 7.4).

142 CHAPTER IV

to much lower BSA adsorption, with no significant difference to microspheres coated with pure PLL-PEG, which is considered to form a stealth corona that is repellent to protein adsorption [24, 25]. As expected from the ConA study, no stealth effect was observed for microspheres incubated with mannose instead of mannan (a PLL:b PLL-PEG/Mannose; data not shown).

The BSA repellence of coatings with PLL-PEG copolymers that were terminally coupled with peptide ligands, i.e. PLL-PEG-RGD and PLL-PEG-RDG, was similar to that of ligand-free PLL-PEG (Figure 4 B). Coatings with PLL-PEG-mannoside also lowered the binding of BSA, however to a lesser extent. Although there was no linear trend, higher ligand densities seemed to impair the repellence to BSA adsorption.

All BSA adsorption assays as described above were performed at low ionic strength (10 mM HEPES buffer), which does not correspond to physiological conditions. Therefore, additional experiments were performed in buffer supplemented with 150 mM sodium chloride (Figure 4 C). Under this condition, the repellence to BSA adsorption upon PLL-PEG-mannoside coatings was greatly improved and indistinguishable from unsubstituted PLL-PEG controls. Thus, at least under simulated physiological conditions, PLL-PEG-mannoside coated microspheres were largely repellent to BSA.

ζ- potential of coated microspheres

Most mannan-coated microspheres had a slightly negative ζ-potential (approx. -5 ± 5 mV, measured in 10 mM HEPES buffer, pH 4.7; Table 2). Only formulations coated with very low fractions of PLL, and thus very small amounts of mannan, had a slightly positive ζ-potential, with a transition from positive to negative values at 15% PLL. For comparison, uncoated carboxylated PS microspheres had a highly negative ζ-potential of – 36 ± 8 mV, and microspheres coated with PLL and (ligand-free) PLL-PEG had ζ-potentials of -1 ± 3 mV and 5 ± 1 mV, respectively. The ζ-potential of PLL-PEG coated microspheres remained unchanged for copolymers terminally substituted with the oligopeptides RGD or RDG. Similarly, all microspheres coated with carbohydrate substituted PLL-PEG copolymers had slightly positive ζ-potentials

CHAPTER IV 143

(between zero and + 5 mM). Thus, ζ-potential measurements further corroborated the presence of the polymeric adlayers.

Phagocytosis by MΦ and DC

All mannan coated microspheres were readily phagocytosed by both MΦ and DC. Little mannan seemed to be needed for recognition, because even microspheres with only 15% PLL in the primary coating step were equally phagocytosed to the same extent as (100%) PLL coated control microspheres (Figures 5 A and 6 A). In fact, this formulation - 15 PLL:85 PLL-PEG/Mannan - was that with the lowest PLL content for which the presence of mannan in the coating could be demonstrated with the ConA assay.

Table 2: ζ-potential of coated microspheres. Measured in 10 mM HEPES (pH 7.4). Values in mV ± SD.

144 CHAPTER IV

Figure 5. Phagocytosis of coated microspheres by MΦ. MΦ were incubated with surface modified microspheres for 4 hours and phagocytosis was assessed by flow cytometry. SSC data are presented as arbitrary units [± SD]. Significant differences relative to PLL-PEG are reported as ∗∗ (p≤0.01, HSD method) and (∗) only significant with LSD method (p≤0.05) but n.s. (p>0.05) with HSD method. A: Mannan coated microspheres, with controls; B: PLL-PEG-mannoside coated microspheres, with PLL-PEG coated and uncoated PS controls; C: Microspheres coated with mixtures of PLL-PEG-mannose and ligand-free PLL-PEG; D: Microspheres coated with PLL-PEG-RGD and 1:1 mixtures of PLL-PEG-mannose and PLL-PEG-RGD.

CHAPTER IV 145

Figure 6. Phagocytosis of coated microspheres by DC. DC were incubated with surface modified microspheres for 4 hours (50 μg; A-C) or 3 days (100 μg; D) and phagocytosis was assessed by flow cytometry. SSC data are presented as arbitrary units (± SD). Significant differences relative to PLL-PEG are reported as ∗∗ (p≤0.01, HSD method). A: Mannan coated microspheres, with controls; B: PLL-PEG-mannoside coated microspheres, with PLL-PEG coated and uncoated PS controls; C: Microspheres coated with PLL-PEG-RGD and 1:1 mixtures of PLL-PEG-mannose and PLL-PEG-RGD; D: Microsphere formulations selected for maturation experiments, after incubation for 3 days.

146 CHAPTER IV

In the case of microspheres coated with carbohydrate-substituted PLL-PEG, MΦ behaved differently from DC. Enhanced phagocytosis was observed for MΦ with the two polymers having the highest densities of mannose substitution, PLL-PEG-PrMan(45%) and PLL-PEG-PrMan(34%), reaching about 50% of the phagocytosis values obtained for PLL control microspheres (Figure 5 B). However, statistically, the values were only marginally different from the PLL-PEG controls, with p ≤ 0.05 for the weaker LSD post-hoc method, but without a significant difference when using the HSD method. While all other formulations where not significantly different from the negative controls (ligand-free PLL-PEG), we still observed tendencies that were worth to consider further. In fact, the human APCs used in this study were subject to high inter-individual differences between donors, which could account for the relatively high standard deviations observed. Thus, further repetitions of the experiments, in excess of the three different donors used in this study, might actually shift some of these tendencies towards statistical significance. As demonstrated for PLL-PEG-PrMan(45%) (Figure 5 C), we observed an interesting trend between increasing phagocytosis and increasing carbohydrate content when mixing carbohydrate substituted with ligand-free PLL-PEG. While it is impossible to determine the minimum density of the ligand required for phagocytosis, it is obvious that increasing rates of receptor-ligand binding events increased the rate of phagocytosis.

It is equally interesting to note that PLL-PEG-EGTri-Man(35%) did not enhance phagocytosis rates as compared to other PLL-PEG-mannoside polymers, although this formulation had the highest available carbohydrate content of all formulations used in this study. Because PLL-PEG-EGMan(45%) coated microspheres were also far less efficiently internalized than PLL-PEG-PrMan(45%) coated microspheres, we conclude that this linker was not optimal for receptor recognition. A similar linker effect was already observed in the mannose recognition assay with ConA.

Finally, we were interested in a potential synergy in phagocytosis when simultaneously presenting both the integrin-targeting RGD ligand and the lectin-targeting carbohydrate ligands [39]. PLL-PEG-RGD alone induced, on average,

CHAPTER IV 147

a slight increase in phagocytosis relative to ligand-free controls, without any effect observed for the scrambled control sequence (PLL-PEG-RDG) (Figure 5 D). Thus, phagocytosis seemed to be affected by the integrin targeting ligand, however far less than in the study by Faraasen et al observing full restoration of phagocytosis up to the level of uncoated microspheres [40]. When we incubated MΦ with microspheres coated with equal amounts of PLL-PEG-RGD and PLL-PEG-mannoside, the degree of phagocytosis was approximately the same as with microspheres coated with PLL-PEG-mannoside alone. Keeping in mind that phagocytosis dropped approximately by half when PLL-PEG-mannoside was mixed with ligand-free PLL-PEG, an additive effect of both ligands may be concluded. However, there was no indication of mutual cooperativity between both ligands, which would have enhanced phagocytosis above the levels of the single components.

Phagocytosis rates with DC were found to be generally lower as compared to MΦ, which is typical for these cells [37, 41]. This may explain why none of the carbohydrate substituted PLL-PEG formulations was phagocytosed by DC (within four hours) to an extent significantly different from ligand-free PLL-PEG controls (Figure 6 B and C). While the data may suggest similar trends as seen above with MΦ, clear-cut conclusions cannot be drawn. Therefore, it was instructive to analyze the phagocytosis rates obtained during the maturation studies (see below), where twice the amount of microspheres was incubated with DC for a relatively long timeframe (three days); under such conditions the relative internalization of PLL-PEG-mannoside as compared to control microspheres was drastically enhanced (Figure 6 D). In fact, microspheres coated with PLL-PEG-PrMan(45%) or mixtures of PLL-PEG-PrMan(45%) and PLL-PEG-RGD were internalized as efficiently as uncoated control microspheres. Nevertheless, PLL-PEG-EGTri-Man(45%) was recognized to a much lower extent, confirming the findings with MΦ.

148 CHAPTER IV

Figure 7. Intracellular localization of microspheres by confocal microscopy. MΦ (A, C, E, F) and DC (B, D) were incubated with surface modified microspheres (red) for 4 hours, fixed and stained for the actin cytoskeleton (green) and the nucleus (blue). The images show both x-y optical sections and x-z or y-z projections. A, B: 45 PLL: 55 PLL-PEG/Mannan; C, D: PLL coated microspheres; E: PLL-PEG-PrMan(45%); F: PLL-PEG-PrMan(34%).

CHAPTER IV 149

Intracellular localization of phagocytosed microspheres

Neither can flow cytometry distinguish microspheres that are phagocytosed from those that are only attached to the outer cell membrane, nor is it possible to unambiguously ascertain their intracellular localization by phase-contrast or fluorescence microscopy. Therefore, control cells were incubated at 4°C with the microspheres to determine the contribution of attached microspheres to the overall signal. In general, the contribution of attached microspheres was found to be negligible. Nevertheless, the data presented above were corrected to account for phagocytosed microspheres only.

In addition, selected samples were analyzed by confocal microscopy (Figure 7). We used rhodamine labelled microspheres and stained the actin cytoskeleton with OregonGreen-488 labelled phalloidin, staining the cells green, while (surface labelled) microspheres appeared as red circles. In addition to x-y optical sections, we acquired z-stacks of whole cells, which were projected to x-z and y-z sections, respectively. Thereby, intracellular localization of the microspheres was unambiguously confirmed by their embedment into the cells´ actin cytoskeleton.

No stimulation of DC maturation upon receptor specific phagocytosis

To test our microparticulate platforms in an immunologically relevant setting, we investigated the effect of various surface coatings on DC stimulation. We selected PLL-PEG-PrMan(45%), i.e. mannose substituted PLL-PEG with the highest proportion of attached ligand, and PLL-PEG-EGTri-Man(45%), to compare the effect of oligo- vs. mono-mannose ligands. In addition, PLL-PEG-PrMan(45%) was mixed with PLL-PEG-RGD to study possible cooperative effects of two different phagocytic receptors. Also PLL-PEG-RGD alone was examined. For the mannan coated prototype, we chose a formulation with little mannan on the surface, 15 PLL:85 PLL-PEG/Mannan, as well as 45 PLL:55 PLL-PEG/Mannan, being the formulation just sufficient to achieve surface saturation with mannan (as indicated by the plateau in the ConA assay; see Figure 2 A), and finally the fully mannan coated formulation, 100 PLL:0 PLL-PEG/Mannan. In addition, the effect of several control

150 CHAPTER IV

formulations was assessed, including uncoated microspheres, coated microspheres lacking the respective ligand (PLL, PLL-PEG), and soluble polymers used for the different coating steps.

To exclude potential effects of microsphere toxicity on DC maturation, we first examined cell viability. Toxicity was generally found to be statistically insignificant, with the exception of two control formulations (uncoated and PLL coated carboxylated PS microspheres) which caused a moderate drop in viability of about 16% or 11%, respectively. Thus, significant toxicity of the formulations could be ruled out.

As shown in Figure 8, the expression of the characteristic surface markers for DC maturation, namely CD83, CD86, and MHC-II, was found to be identical to those of immature DC (iDC). This was most obvious for CD83, which is hardly expressed by iDC but becomes highly expressed on mature DC (mDC). On the contrary, CD86 and MHC-II are present on both iDC and mDC but are

Figure 8. Carbohydrate modified microspheres do not stimulate DC maturation. Cells were stained with antibodies against CD83, CD86, and MHC-II or matched isotype control antibodies after incubation with selected microsphere formulations for three days. Control cells wereincubated with LPS to generate mDC (positive control) or left untreated (iDC, negative control).There were no significant differences between iDC and any of the cells incubated withmicrosphere samples. Dotted lines: unstained cells; dashed lines: isotype controls; light greytinted: iDC; dark grey tinted: mDC; black lines: microspheres coated with PLL-PEG-EGTri-Man(45%), PLL-PEG-PrMan(45%), a 1:1 mixture of PLL-PEG-PrMan(45%) and PLL-PEG-RGD, or different amounts of mannan.

CHAPTER IV 151

overexpressed upon DC stimulation. Nevertheless, also for these two surface markers, the profiles of microsphere treated DC clearly corresponded to those of iDC. Full DC functionality was confirmed by an LPS challenge which induced an mDC profile showing an associated upregulation of all three surface markers. Our results implicate that DC did not mature upon any of the tested microsphere formulations. There was no indication that any of the studied mannose-based molecular patterns on the microspheres led to any level of maturation. Consistently, soluble polymers did not change the maturation state either, when studied at concentrations corresponding to those used for the various coating steps.

DISCUSSION

Specific targeting of APC receptors is increasingly recognized for its attractive potential in immunomodulation [1]. Among the various APC receptors, TLRs and CLRs were extensively studied due to their capacity to recognize pathogen associated molecular patterns (PAMPs) [2, 3]. Most approaches to target an APC receptor either use a soluble protein covalently coupled to a ligand, or ligand modified liposomal carriers [2, 42]. Commonly, there are multiple possibilities for ligand-receptor interactions, e.g., through proteinic epitopes or opsonization of the carrier. Nonetheless, especially for the immunomodulatory aspect of APC receptor targeting, it is often difficult to determine whether an observed effect is exclusively triggered by an interaction of the ligand with its receptor, or results from a combination of different signalling events involving, among others, both ligand and antigen recognition. Therefore, in this work we explored a versatile microparticulate platform that displays specific ligands on the surface of a stealth microsphere. We used PS microspheres of 5 μm to mimic the phagocytosis of microbial pathogens. The microparticulates developed in this work were tested for their abilities to target APCs, induce phagocytosis and stimulate maturation of iDC. To this aim, the microspheres were surface modified in such a way that various carbohydrate ligands were displayed on the outer interface of a protein repellent PLL-PEG

152 CHAPTER IV

adlayer. The mechanism for protein repellence (or stealth function) of micro- and nanoparticulates through a PEG corona is well understood [43, 44]. Its underlying concept is to prevent unspecific opsonization of the particulates and subsequent interactions between opsonins and APC receptors. Similar to PEG, the polysaccharide mannan has also been associated with protein repellence [45], obviously owing to its hydrophilic and highly hydrated nature. Moreover, mannan is well known for its immunomodulatory potential [17, 18].

Preparation of surface modified microspheres

In the present study, we concentrated on mannose-based ligands to target lectin receptors on DC and MΦ. To this aim, we developed microspheres whose surface was either functionalized with mono- and tri-mannose ligands, expected to target MR, or with the poly-mannose mannan, a well-established multivalent ligand for the receptors TLR4 and DC-SIGN [7, 32]. In both cases, a PLL-PEG adlayer was adsorbed onto the microspheres to provide a protein repellent and phagocytosis resistant background [25].

Mannan (from S. cerevisiae) is an oligosaccharide consisting of an α(1,6) linked mannose backbone grafted with α(1,2) and α(1,3) linked mannose side chains, some of which contain free phosphate groups [30]. Therefore, mannan readily adsorbs onto positively charged PLL coated microspheres. The presence of mannan, and its availability as a ligand to lectin receptors, could be directly proven by the specific adsorption of ConA, a plant lectin (Figure 2 A). Moreover, incubation of the microspheres with a model protein, BSA, gave evidence for the protein repellent character of mannan coatings (Figure 4 A). The latter effect closely corresponds with experiments by Vukovic et al. on interactions between mannan and other cell wall components, namely glucan, chitin, and proteins [45]. Their results suggested that mannan represents an inert cell wall component that does not interact with proteins and carbohydrates. However, to our knowledge, mannan has not as yet been used to passivate microspheres against opsonization.

Next to PLL coated microspheres fully covered with mannan, we were interested in microspheres that exhibited graded mannan surface concentrations.

CHAPTER IV 153

Because mannan was readily adsorbed onto PLL coated microspheres, we speculated that partial coatings with mixtures of PLL and PLL-PEG should enable to control mannan surface concentrations, because no adsorption of mannan should occur onto PEG shielded domains on the surface. The results from Figure 2 A clearly demonstrate increasing amounts of mannan adsorption with increasing fractions of PLL whereas no detectable adsorption of mannan was found on exclusively PLL-PEG coated microspheres. A plateau was reached at a fraction of approximately 55% PLL. This outcome may be explained by two mechanisms: firstly and most importantly, the indicated mixing ratios refer to concentrations in the primary coating solutions, and are supposed to differ from those of the adsorbed adlayers. Adsorption of linear PLL is presumably favoured over the adsorption of “bottle brush” type PLL-PEG. Therefore, it is likely that for the maximum PLL surface concentration a fraction of 55% PLL in the coating solution is sufficient. Secondly, because of its large molecular weight adsorbed mannan might be able to bridge between adjacent PEGylated spots. Importantly, mixed PEG and mannan coatings were as protein repellent as after individual coatings (Figure 4 A).

In addition to the surface modification through a mannose-rich natural polysaccharide, mannan, we also aimed at a synthetic approach to generate mannose-rich surface modifications on PS microspheres. The underlying principle was the terminal conjugation of mono- or oligo-mannose to the PEG chains of a PLL-PEG copolymer that was then adsorbed onto the microspheres´ surface to form mannoside decorated adlayers. Protocols for the terminal conjugation of bioactive molecules to PLL-PEG were previously published by some of us, such as with biotin [27, 46] or RGD-peptides as ligands. [26, 28] More recently, we reported on the coupling of mannoside ligands to PLL-PEG [29]. For the present work, we used mono- and tri-mannose ligands at different degrees of substitution, denoted as PLL-PEG-λMan(x%) and PLL-PEG-λTri-Man(x%), with x representing the experimental percentages of substitution (0, 28, 34 and 45% conjugation). Two different linkers between the PEG moieties and the ligands were tested, a propyl linker (denoted as PrMan) and an ethylene glycol linker (EGMan). In addition to the coatings with single copolymers ligated with different percentages of ligand, we also prepared mixed

154 CHAPTER IV

coatings composed of ligand-modified and ligand-free PLL-PEG. Both strategies yielded microspheres with graded mannose surface concentrations (Figure 2 B). However, we expected the local distribution of the ligand after coating to be different, with a more statistical distribution for PLL-PEG-mannoside(x%) as compared to a more patchlike distribution for mixed adlayers of ligand-modified and ligand-free PLL-PEG.

Unlike mannan coated microspheres, PLL-PEG-mannoside coated microspheres did not fully repel model protein (BSA) adsorption at low ionic strength (10 mM HEPES, Figure 4 B). However, this ionic strength was far below physiological conditions. Moreover, the ability of PLL-PEG to repel protein adsorption was previously shown to depend on the ionic strength [47]. Accordingly, we repeated the experiment in HEPES buffer supplemented with 150 mM NaCl. Under these conditions, the adsorption of BSA on PLL-PEG-mannoside was as low as on PLL-PEG (Figure 4 C). Therefore, we conclude that under the conditions of the cell assays, opsonization on PLL-PEG-mannoside should be negligible.

We were finally interested to study not only single ligands, but also combinations of two or more ligands. PLL-PEG-RGD coated microspheres have been previously used to target integrins on APCs [40]. Here we coated PS microspheres with PLL-PEG-RGD, the negative control sequence PLL-PEG-RDG, and combinations of PLL-PEG-RGD with PLL-PEG-mannoside. The presence of the polymer was demonstrated by ζ-potential and repellence to protein adsorption.

Targeting of APC surface receptors: Phagocytosis in DC and MΦ

Next to the design of a versatile microparticulate platform to target APCs, we analyzed the effects of ligands, ligand densities and ligand combinations on phagocytosis. Mannan is a polysaccharide ubiquitously found in the cell walls of yeasts, and a well-known ligand for at least two APC receptors, TLR4 and DC-SIGN [7, 32]. Because of the reproducible adsorption of graded amounts of mannan onto PLL and PLL-PEG precoated microspheres, we were interested to study the effect of mannan surface concentrations on recognition by both types

CHAPTER IV 155

of APCs. Interestingly, our results followed a “yes or no” like pattern. All mannan surface concentrations that were above the detection limit of the ConA assay were internalized as efficiently as unmodified or PLL coated control microspheres (Figures 5 A and 6 A). This suggests that very little mannan is needed to trigger internalization by APCs.

At this point no information regarding the spatial arrangement of surface immobilized mannan on our formulations is available. Two contrasting scenarios may be considered: On the one hand adsorbed mannan may form a smooth adlayer, but on the other hand it is equally possible that long chains or loops may protrude from the surface. In the latter case, far more binding sites would have been available for receptor recognition. Moreover, mannan is a natural polysaccharide with concomitantly variable molecular size and structure. In addition, more than one APC receptor is known to recognize mannan. In fact, members of two different receptor families, namely TLR and CLR, are capable of recognition. Thus, although mannan is a natural ligand for APC receptors [7, 32], due to flexible ligand structure and receptor promiscuity there may be several options when it comes to the interpretation of single receptor-ligand interactions.

A synthetic approach to generate mannose-rich surface modifications on PS microspheres with PLL-PEG-mannoside could overcome some of these drawbacks. Both experimental [23] and theoretical [48] studies have previously shown that the PEG side chains of adsorbed PLL-PEG graft copolymers extend perpendicularly from the surface, exposing a coupled ligand to the surrounding environment. Small carbohydrate substructures representing a minimum recognition unit should be ideally suited to target specialized types of receptors. For instance, among other carbohydrate recognizing receptors, there are two receptors found on APCs that have been characterized as mannose binding, namely MR and DC-SIGN [7]. From these two receptors, MR mainly recognizes terminal mono-mannose or di-mannose ligands [8, 9]. In contrast, DC-SIGN has high affinity for internal (rather than terminal) sugars within more complex mannose structures, including mannan, as well as fucose containing oligosaccharides (blood group substances and Lewis antigens) [7, 10, 11]. α(1,3)-α(1,6)-mannotriose was also shown to bind to DC-SIGN, however with

156 CHAPTER IV

relatively low affinity [12], and seems to be essential for carbohydrate recognition by DC-SIGN [10, 13]. It was recently demonstrated that high affinity binding to DC-SIGN requires additional interactions of the lectin with terminal α(1,2)-mannose moieties [13]. The ligand specificity of Dectin-2, another member of the CLR family, is still controversial. However, recently McGreal et al. [14] presented evidence for specific recognition of high-mannose structures (Man9GlcNAc2 > Man8GlcNAc2 > Man7GlcNAc2). Thus, in contrast to mannan, the mono- and tri-mannoside substituted PLL-PEG coated formulations used in our study were expected to target primarily to the MR.

Our data indicated that the efficiency of phagocytosis increased with increasing amounts of mannose exposed on the microspheres´ surface (Figure 5 B and C). This is in agreement with recent findings of other groups with other delivery platforms, such as mannosylated emulsions [49] or liposomes [50]. In general, however, we found mannose mediated phagocytosis to be relatively poor. While the MR is usually considered as professional phagocytic receptor, first doubts about this classification have been previously raised by Le Cabec et al. [51] In their work, expression of human MR in three different cell types with a principle capacity for phagocytosis, yielded endocytosis of soluble ligands rather than phagocytosis of ligand modified particulates. Therefore, the authors proposed that the role of MR is that of a binding receptor that requires additional partner(s) for phagocytosis. Our present results may be interpreted as further evidence for this hypothesis. Even though mannose mediated phagocytosis was observed to some degree, its overall efficiency was minor. At this point, information about partner receptor(s) is still missing.

Generally, there are two complementary ways to identify such partner receptors. A first strategy is to restrict the types of available receptors. This can be accomplished by receptor overexpression in appropriate cells, as in the aforementioned work of Le Cabec et al. [51]. A second option is to restrict the available types of ligands. The latter approach would be represented by particulates that are surface modified with the respective ligands. So far, however, most studies used mannosylated particulates that were not passivated against opsonization, thus enabling uncontrolled interactions with other cell

CHAPTER IV 157

surface receptors. Therefore, high internalization efficiencies observed in many studies do not necessarily imply that phagocytosis is mediated by the MR, but seem to be the combined effect of MR binding and interactions with other cell surface receptors. The novelty of our approach to this issue is the PEG based stealth function having the capacity to block interactions other than with the coupled ligand. Thus, by mixing potential ligands, receptors may be identified that are either constitutively active in phagocytosis or have additive or cooperative effects on phagocytosis.

As a first example for this concept, mixed adlayers presenting both mannose and RGD ligands were prepared (Figure 5 B and D). When MΦ were incubated with microspheres coated with equal amounts of PLL-PEG-RGD and PLL-PEG-mannoside, levels of phagocytosis equalled those of microspheres fully coated with PLL-PEG-mannoside alone. On the other hand, the internalization of microspheres coated with either PLL-PEG-RGD or a mixture of equal parts of PLL-PEG-mannoside and ligand-free PLL-PEG was comparatively small. Therefore, we conclude that there was no cooperativity between the two types of ligands. Phagocytosis levels of the mixtures did not exceed those of microspheres with pure coatings.

It is interesting to note that our results could not fully reproduce a previous study by Faraasen et al. [40] made in our group. In this work a RGD-functionalized PEG-corona was found to restore the phagocytosis of coated microspheres close to control levels. Due to the limited methodology used by Faraasen et al., however, it is difficult to quantitatively compare our data with that study. In particular, the presence and quality of the respective coatings of the former study were not further characterized (e.g. by ζ-potential or protein adsorption assays), and cell studies were analyzed only by direct light and fluorescence microscopy which is inherently more susceptible to biases in image acquisition than flow cytometry. Moreover, donor-to-donor differences inherent to the cell culture model may have also contributed to the observed differences. In the present study, we found only little uptake mediated by the RGD-ligand when terminally coupled to the copolymer.

158 CHAPTER IV

No stimulation of DC upon receptor specific targeting

Interactions between surface modified microspheres and APCs were also investigated for the potential to affect their stage of maturation. For this purpose we selected representative surface modified formulations for an evaluation of the maturation patterns of DC after phagocytosis, which would be indicative of the type of DC stimulation. However, none of the formulations stimulated any marker of maturation in DC (Figure 8).

The negative outcome of this experiment was quite unexpected, given the fact that mannan is an established immunopotentiator [15-18]. Even soluble (unconjugated) mannan was recently shown to activate DC, with different modifications of mannan yielding different immune biases [52]. In fact, oxidized as well as unmodified mannan stimulated Th1, while reduced mannan stimulated Th2. Nonetheless, the mannan concentrations used in this study [52] were extremely high, and expression of maturation markers even at concentrations as high as 800 μg/ml soluble mannan was much lower than with LPS control (1 μg/ml). Importantly, using similarly high concentrations of soluble mannan in our experiments comparable degrees of maturation could be achieved (data not shown).

According to the literature, most studies combining mannan with particulates were based on liposomes with mannan being covalently linked to certain membrane lipids, such as cholesterol [15, 16], or palmitoyl-mannan [17]. For another approach, mannan was covalently coupled to (soluble) proteins. For instance, a conjugate of the cancer epitope MUC1 with (oxidized) mannan was reported to be in phase III clinical trials [18]. The usual immunological reaction to such mannan conjugates was towards a Th1 response. However, in the case of conjugated mannan, both the proteinic and the lipidic moieties in the conjugates may be processed by the DC for T-cell presentation via MHC or CD1, respectively. Thus, the final polarization of the DC was not necessarily the result of an activation of phagocytic receptors alone, but was probably also affected by signals stemming from the processing of the antigenic part of the conjugate. On the other hand, by themselves, carbohydrates are poor immunogens that may activate only B cells through a thymus-independent (TI-2) mechanism, but may

CHAPTER IV 159

not be made available to T-cells [53]. To our knowledge, we report here for the first time on mannan being non-covalently immobilized on an inert microparticulate carrier. Our findings lead us to the hypothesis that mannan coated microspheres, somehow mimicking an “artificial yeast”, cannot provide a major danger signal to DC. This implies that the manifold observed Th1/Th2 biases that were thought to be mediated through mannan conjugation per se may not only represent the result of its recognition by either TLR4 or CLRs alone, but require concomitant antigen processing. To challenge this hypothesis, further steps should implement the co-encapsulation of a model antigen into mannan coated microspheres in addition to the empty carrier.

CONCLUSION

Here we introduced a versatile technique to prepare microspheres with surface assembled adlayers presenting mannose-based ligands and graded ligand densities for cellular recognition. Microspheres decorated with either mannan or smaller substructure ligands, such as mono- and tri-mannose, could be readily and rapidly fabricated through fast and convenient self-assembly protocols. While the small substructure mannosides in this study were expected to target to the MR only, the targeting of other CLRs, such as DC-SIGN, would be equally feasible. One possibility would be substitution with higher oligo-mannosides. Alternatively, the increased affinity of carbohydrate ligands to DC-SIGN through additional interaction with terminal α(1,2)-mannose moieties has potential to be mimicked through mixed coatings of PLL-PEG substituted with both tri-mannose and α(1,2)-mannose. Combinations of different types of ligands are equally feasible, as demonstrated with a combination of a mannoside ligand and an adhesion peptide ligand, such as RGD.

Most importantly, the coatings turned out to repel unspecific protein adsorption, shielding unspecific opsonization and providing an opportunity for specific cellular interactions between the microspheres´ surface and receptors on APCs. Therefore, this microparticulate construct represents a promising tool for

160 CHAPTER IV

further investigations of specific ligand-receptor interactions on phagocytes, including the screening of potential ligands and ligand combinations.

Mannose density was the major factor for the phagocytic internalization of mono- and tri-mannose decorated microspheres, though with limited efficiency. This strengthens recent doubts raised by other authors [51] about MR acting as a constitutive phagocytic receptor, i.e. without further interactions with partner receptors. Further studies including combinations of various ligands to address potential co-receptors would certainly help to identify such receptors.

Finally, an analysis of DC maturation revealed that surface assembled mannan on the microspheres could not stimulate DC maturation. Therefore, we conclude that recognition by CLRs alone followed by phagocytosis of the microspheres cannot trigger DC activation towards a T helper response by itself. To this end the co-processing of a suitably encapsulated antigen could to be mandatory. Therefore, to obtain further insight into the opportunities of immunomodulation by surface modified particulates co-delivery of antigens may be helpful.

ACKNOWLEDGEMENTS

Financial support by ETH Zurich (Grant THG-38/04-2) is gratefully acknowledged. The authors would like to thank Katrin Barth for assistance in glycopolymer synthesis and Riccardo Castelli for providing derivatized mannose ligands.

CHAPTER IV 161

ABBREVIATIONS

The abbreviations used are APC: antigen presenting cell; TLR: Toll-like receptor; CLR: C-type lectin receptor; MR: mannose receptor; DC-SIGN: dendritic cell-specific ICAM-3 grabbing non-integrin; MΦ: macrophages; DC: dendritic cells; iDC: immature DC; mDC: mature DC; PS: polystyrene; PEG: poly(ethylene glycol); PLL: poly(L-lysine); PLL-PEG: PLL(20)-g-PEG(2), a copolymer of PEG (MW=2 kDa) grafted to PLL (MW=20 kDa) at a grafting ratio g = 3.5 - 4.0; grafting ratio (g): number of lysine monomers divided by the number of PEG side chains; PLL-PEG-ligand(x%): PLL-g-PEG(2)/PEG(3.4)-ligand(x%), a PLL-PEG copolymer with x% of the side chains exchanged by ligand-coupled PEG chains of 3.4 kDa; PLL-PEG-RGD: PLL-g-PEG(2)/ /PEG(3.4)-RGD(8%); PLL-PEG-RDG: PLL-g-PEG(2)/PEG(3.4)-RDG(9%); PLL-PEG-mannoside: collective abbreviation for PLL-g-PEG(2)/PEG(3.4)-mannoside(x%), with mannoside as general term for different ligands (mono- and tri-mannose) and linkers (propyl and EG); PLL-PEG-λMan(x%) and PLL-PEG-λTri-Man(x%): PLL-PEG-ligand(x%) with mono- or tri-mannose as a ligand and λ indicating the respective linker (Pr or EG); a PLL:b PLL-PEG: coatings with a% PLL and b% PLL-PEG; a PLL:b PLL-PEG/Mannan: a secondary coating with mannan onto a PLL:b PLL-PEG; ConA: Concanavalin-A; BSA: bovine serum albumin; HEPES buffer: 10 mM 4-(2-hydroxyethyl)piperazine-1-ethane-sulfonic acid, pH 7.4; PBS buffer: 10 mM phosphate buffered saline, pH 7.4; SSC: side scattering (measured by flow cytometry); PI: propidium iodide.

162 CHAPTER IV

REFERENCES 1. Gogolák, P., Réthi, B., Hajas, G., and Rajnavölgyi, É. Targeting dendritic

cells for priming cellular immune responses. J. Mol. Recognit., 2003. 16(5): p. 299-317.

2. Proudfoot, O., Apostolopoulos, V., and Pietersz, G.A., Receptor-mediated delivery of antigens to dendritic cells: anticancer applications. Mol Pharm, 2007. 4(1): p. 58-72.

3. Cambi, A., and Figdor, C.G., Dual function of C-type lectin-like receptors in the immune system. Curr. Opin. Cell Biol., 2003. 15(5): p. 539-546.

4. Medzhitov, R., and Janeway, C.A., Jr., Decoding the patterns of self and nonself by the innate immune system. Science, 2002. 296(5566): p. 298-300.

5. Clark, R., and Kupper, T., Old meets new: the interaction between innate and adaptive immunity. J. Invest. Dermatol., 2005. 125(4): p. 629-637.

6. Van Vliet, S.J., Gruen, C.H., and Van Kooyk, Y., C-type lectin receptors that regulate pathogen recognition through the recognition of carbohydrates. Protein-Carbohydrate Interactions in Infectious Diseases, 2006: p. 106-124.

7. Cambi, A., Koopman, M., and Figdor, C.G., How C-type lectins detect pathogens. Cell Microbiol, 2005. 7(4): p. 481-488.

8. Geijtenbeek, T.B., van Vliet, S.J., Engering, A., t Hart, B.A., and van Kooyk, Y., Self- and nonself-recognition by C-type lectins on dendritic cells. Annu. Rev. Immunol., 2004. 22: p. 33-54.

9. Frison, N., Taylor, M.E., Soilleux, E., Bousser, M.-T., Mayer, R., Monsigny, M., Drickamer, K., and Roche, A.-C., Oligolysine-based oligosaccharide clusters. Selective recognition and endocytosis by the mannose receptor and dendritic cell-specific intercellular adhesion molecule 3 (CiCAM-3)-grabbing nonintegrin. Vol. 278 (2003) 23922-23929. J. Biol. Chem., 2003. 278(41): p. 40415.

10. Feinberg, H., Mitchell, D.A., Drickamer, K., and Weis, W.I., Structural basis for selective recognition of oligosaccharides by DC-SIGN and DC-

CHAPTER IV 163

SIGNR. Science (Washington, DC, United States), 2001. 294(5549): p. 2163-2166.

11. Guo, Y., Feinberg, H., Conroy, E., Mitchell, D.A., Alvarez, R., Blixt, O., Taylor, M.E., Weis, W.I., and Drickamer, K., Structural basis for distinct ligand-binding and targeting properties of the receptors DC-SIGN and DC-SIGNR. Nat Struct Mol Biol, 2004. 11(7): p. 591-598.

12. Appelmelk, B.J., van Die, I., van Vliet, S.J., Vandenbroucke-Grauls, C.M.J.E., Geijtenbeek, T.B.H., and van Kooyk, Y., Cutting Edge: Carbohydrate profiling identifies new pathogens that interact with dendritic cell-specific ICAM-3-grabbing nonintegrin on dendritic cells. J. Immunol., 2003. 170(4): p. 1635-1639.

13. Feinberg, H., Castelli, R., Drickamer, K., Seeberger, P.H., and Weis, W.I., Multiple modes of binding enhance the affinity of DC-SIGN for high mannose N-linked glycans found on viral glycoproteins. J. Biol. Chem., 2007. 282(6): p. 4202-4209.

14. McGreal, E.P., Rosas, M., Brown, G.D., Zamze, S., Wong, S.Y., Gordon, S., Martinez-Pomares, L., and Taylor, P.R., The carbohydrate-recognition domain of Dectin-2 is a C-type lectin with specificity for high mannose. Glycobiology, 2006. 16(5): p. 422-430.

15. Toda, S., Ishii, N., Okada, E., Kusakabe, K.I., Arai, H., Hamajima, K., Gorai, I., Nishioka, K., and Okuda, K., HIV-1-specific cell-mediated immune responses induced by DNA vaccination were enhanced by mannan-coated liposomes and inhibited by anti-interferon-gamma antibody. Immunology, 1997. 92(1): p. 111-117.

16. Cui, Z., Han, S.-J., and Huang, L., Coating of Mannan on LPD Particles Containing HPV E7 Peptide Significantly Enhances Immunity Against HPV-Positive Tumor. Pharm. Res., 2004. 21(6): p. 1018-1025.

17. Jain, S., and Vyas, S.P., Mannosylated niosomes as adjuvant-carrier system for oral mucosal immunization. J Liposome Res, 2006. 16(4): p. 331-345.

18. Apostolopoulos, V., Pietersz, G.A., Tsibanis, A., Tsikkinis, A., Drakaki, H., Loveland, B.E., Piddlesden, S.J., Plebanski, M., Pouniotis, D.S.,

164 CHAPTER IV

Alexis, M.N., McKenzie, I.F., and Vassilaros, S., Pilot phase III immunotherapy study in early-stage breast cancer patients using oxidized mannan-MUC1 [ISRCTN71711835]. Breast Cancer Res, 2006. 8(3): p. R27.

19. White, K., Rades, T., Kearns, P., Toth, I., and Hook, S., Immunogenicity of liposomes containing lipid core peptides and the adjuvant Quil A. Pharm. Res., 2006. 23(7): p. 1473-1481.

20. Hattori, Y., Kawakami, S., Lu, Y., Nakamura, K., Yamashita, F., and Hashida, M., Enhanced DNA vaccine potency by mannosylated lipoplex after intraperitoneal administration. J Gene Med, 2006. 8(7): p. 824-834.

21. Harris, J.M., and Chess, R.B., Effect of PEGylation on pharmaceuticals. Nat. Rev. Drug Discovery, 2003. 2(3): p. 214-221.

22. Kenausis, G.L., Vörös, J., Elbert, D.L., Huang, N., Hofer, R., Ruiz-Taylor, L., Textor, M., Hubbell, J.A., and Spencer, N.D., Poly(L-lysine)-g-Poly(ethylene glycol) Layers on Metal Oxide Surfaces: Attachment Mechanism and Effects of Polymer Architecture on Resistance to Protein Adsorption. J. Phys. Chem. B, 2000. 104(14): p. 3298-3309.

23. Huang, N.-P., Michel, R., Vörös, J., Textor, M., Hofer, R., Rossi, A., Elbert, D.L., Hubbell, J.A., and Spencer, N.D., Poly(L-lysine)-g-poly(ethylene glycol) Layers on Metal Oxide Surfaces: Surface-Analytical Characterization and Resistance to Serum and Fibrinogen Adsorption. Langmuir, 2001. 17(2): p. 489-498.

24. Pasche, S., De Paul, S.M., Vörös, J., Spencer, N.D., and Textor, M., Poly(L-lysine)-graft-poly(ethylene glycol) assembled monolayers on Niobium Oxide surfaces: a quantitative study of the influence of polymer interfacial architecture on resistance to protein adsorption by ToF-SIMS and in situ OWLSu OWLS. Langmuir, 2003. 19(22): p. 9216-9225.

25. Wattendorf, U., Koch, M.C., Walter, E., Vörös, J., Textor, M., and Merkle, H.P., Phagocytosis of poly(L-lysine)-graft-poly(ethylene glycol) coated microspheres by antigen presenting cells: Impact of grafting ratio and poly(ethylene glycol) chain length on cellular recognition. Biointerphases, 2006. 1(4): p. 123-133.

CHAPTER IV 165

26. Huang, N.-P., Vörös, J., De Paul, S.M., Textor, M., and Spencer, N.D., Biotin-derivatized poly(L-lysine)-g-poly(ethylene glycol): A novel polymeric interface for bioaffinity sensing. Langmuir, 2002. 18(1): p. 220-230.

27. VandeVondele, S., Vörös, J., and Hubbell, J.A., RGD-grafted poly-L-lysine-graft-(polyethylene glycol) copolymers block non-specific protein adsorption while promoting cell adhesion. Biotechnol. Bioeng., 2003. 82(7): p. 784-790.

28. Müller, M., Vörös, J., Csúcs, G., Walter, E., Danuser, G., Merkle, H.P., Spencer, N.D., and Textor, M., Surface modification of PLGA microspheres. J. Biomed. Mater. Res., Part A, 2003. 66A(1): p. 55-61.

29. Coullerez, G., Seeberger, P.H., and Textor, M., Merging organic and polymer chemistries to create glycomaterials for glycomics applications. Macromol Biosci, 2006. 6(8): p. 634-647.

30. Vinogradov, E., Petersen, B., and Bock, K., Structural analysis of the intact polysaccharide mannan from Saccharomyces cerevisiae yeast using 1H and 13C NMR spectroscopy at 750 MHz. Carbohydr. Res., 1998. 307(1-2): p. 177-183.

31. Domer, J.E., and Garner, R.E., Fungal wall components and immunostimulation. NATO ASI Ser., Ser. H, 1991. 53(Fungal Cell Wall Immune Response): p. 415-430.

32. Roeder, A., Kirschning, C., Rupec, R., Schaller, M., Weindl, G., and Korting, H.C., Toll-like receptors as key mediators in innate antifungal immunity. Med. Mycol., 2004. 42(6): p. 485-498.

33. Ratner, D.M., Plante, O.J., and Seeberger, P.H., A linear synthesis of branched high-mannose oligosaccharides from the HIV-1 viral surface envelope glycoprotein gp120. Eur J Org Chem, 2002(5): p. 826-833.

34. Houseman, B.T., Gawalt, E.S., and Mrksich, M., Maleimide-functionalized self-assembled monolayers for the preparation of peptide and carbohydrate biochips. Langmuir, 2003. 19(5): p. 1522-1531.

35. Kalb, A.J., and Helliwell, J.R., Concanavalin A. Handbook of Metalloproteins, 2001. 2: p. 963-972.

166 CHAPTER IV

36. Sallusto, F., Cella, M., Danieli, C., and Lanzavecchia, A., Dendritic cells use macropinocytosis and the mannose receptor to concentrate macromolecules in the major histocompatibility complex class II compartment: downregulation by cytokines and bacterial products. J. Exp. Med., 1995. 182(2): p. 389-400.

37. Thiele, L., Rothen-Rutishauser, B., Jilek, S., Wunderli-Allenspach, H., Merkle, H.P., and Walter, E., Evaluation of particle uptake in human blood monocyte-derived cells in vitro. Does phagocytosis activity of dendritic cells measure up with macrophages? J. Controlled Release, 2001. 76(1-2): p. 59-71.

38. Gestwicki, J.E., Cairo, C.W., Strong, L.E., Oetjen, K.A., and Kiessling, L.L., Influencing receptor-ligand binding mechanisms with multivalent ligand architecture. J. Am. Chem. Soc., 2002. 124(50): p. 14922-14933.

39. Stuart, L.M., and Ezekowitz, R.A.B., Phagocytosis: Elegant complexity. Immunity, 2005. 22(5): p. 539-550.

40. Faraasen, S., Vörös, J., Csúcs, G., Textor, M., Merkle, H.P., and Walter, E., Ligand-Specific Targeting of Microspheres to Phagocytes by Surface Modification with Poly(L-Lysine)-Grafted Poly(Ethylene Glycol) Conjugate. Pharm. Res., 2003. 20(2): p. 237-246.

41. Kiama, S.G., Cochand, L., Karlsson, L., Nicod, L.P., and Gehr, P., Evaluation of phagocytic activity in human monocyte-derived dendritic cells. Journal of aerosol medicine: official journal of the International Society for Aerosols in Medicine, 2001. 14(3): p. 289-299.

42. Vyas, S.P., Singh, A., and Sihorkar, V., Ligand-receptor-mediated drug delivery: an emerging paradigm in cellular drug targeting. Crit. Rev. Ther. Drug Carrier Syst., 2001. 18(1): p. 1-76.

43. Vonarbourg, A., Passirani, C., Saulnier, P., and Benoit, J.P., Parameters influencing the stealthiness of colloidal drug delivery systems. Biomaterials, 2006. 27(24): p. 4356-4373.

44. Meng, F., Engbers, G.H.M., Gessner, A., Mueller, R.H., and Feijen, J., Pegylated polystyrene particles as a model system for artificial cells. J. Biomed. Mater. Res., Part A, 2004. 70A(1): p. 97-106.

CHAPTER IV 167

45. Vukovic, R., and Mrsa, V., Structure of the Saccharomyces cerevisiae cell wall. Croat. Chem. Acta, 1995. 68(3): p. 597-605.

46. Tosatti, S., De Paul, S.M., Askendal, A., VandeVondele, S., Hubbell, J.A., Tengvall, P., and Textor, M., Peptide functionalized poly(L-lysine)-g-poly(ethylene glycol) on titanium: resistance to protein adsorption in full heparinized human blood plasma. Biomaterials, 2003. 24(27): p. 4949-4958.

47. Pasche, S., Voeroes, J., Griesser, H.J., Spencer, N.D., and Textor, M., Effects of Ionic Strength and Surface Charge on Protein Adsorption at PEGylated Surfaces. J. Phys. Chem. B, 2005. 109(37): p. 17545-17552.

48. Feuz, L., Leermakers, F.A.M., Textor, M., and Borisov, O., Bending Rigidity and Induced Persistence Length of Molecular Bottle Brushes: A Self-Consistent-Field Theory. Macromolecules, 2005. 38(21): p. 8891-8901.

49. Yeeprae, W., Kawakami, S., Yamashita, F., and Hashida, M., Effect of mannose density on mannose receptor-mediated cellular uptake of mannosylated O/W emulsions by macrophages. J Control Release, 2006. 114(2): p. 193-201.

50. White, K.L., Rades, T., Furneaux, R.H., Tyler, P.C., and Hook, S., Mannosylated liposomes as antigen delivery vehicles for targeting to dendritic cells. J. Pharm. Pharmacol., 2006. 58(6): p. 729-737.

51. Le Cabec, V., Emorine, L.J., Toesca, I., Cougoule, C., and Maridonneau-Parini, I., The human macrophage mannose receptor is not a professional phagocytic receptor. J. Leukoc. Biol., 2005. 77(6): p. 934-943.

52. Sheng, K.-C., Pouniotis, D.S., Wright, M.D., Tang, C.K., Lazoura, E., Pietersz, G.A., and Apostolopoulos, V., Mannan derivatives induce phenotypic and functional maturation of mouse dendritic cells. Immunology, 2006. 118(3): p. 372-383.

53. Lucas, A.H., Apicella, M.A., and Taylor, C.E., Carbohydrate moieties as vaccine candidates. Clin. Infect. Dis., 2005. 41(5): p. 705-712.

168 CHAPTER IV

GENERAL DISCUSSION AND OUTLOOK 169

GENERAL DISCUSSION AND OUTLOOK The objective of this dissertation was the development of a

microparticulate platform that allowed the receptor specific targeting of C-type lectins (CLRs) on antigen presenting cells (APC). To achieve this aim, two requirements had to be met:

(i) The microparticles should have stealth character, i.e. they had to be shielded from opsonization, phagocytic clearance, and engagement of cell surface receptors other than those to be targeted.

(ii) Specific ligands had to be immobilized on the surface of the microparticles, leading to specific recognition by the targeted CLRs without compromising their stealth character.

Targeted microparticles are of special interest in the field of vaccination. Of a similar size range as bacteria and parasites, they may better mimic interactions of such pathogens with immune cells than other vaccine formulations. Moreover, it is generally recognized that engagement of APC receptors by surface modified particulate carriers could potentially affect the type of immune response that will be elicited [1]. Specifically, the targeting of CLRs was identified as a promising strategy for immunomodulation, as documented in several studies with enhanced or modified immune responses through the coupling of mannose type ligands to proteins or particles [2-7]. However, as it is particularly the adsorption of opsonic proteins onto the particles’ surface which promotes recognition by various phagocytic receptors, receptor specific targeting requires surface modifications that are repellent to the adsorption of proteins. It has been hypothesized that this could be accomplished by microparticulate carriers that bear suitable ligands coupled to an MPS resistant polymer, such as poly(ethylene glycol) (PEG) [8].

We chose PLL-g-PEG1 to immobilize a protein repellent PEG corona onto the microparticles. This was for two reasons. First, PLL-g-PEG self-assembles onto negatively charged surfaces from aqueous, neutral buffered

1 For simplicity, in all of Chapter IV PLL-g-PEG was denoted by PLL-PEG, and PLL-g-PEG/PEG-ligand as PLL-PEG-ligand.

170 GENERAL DISCUSSION AND OUTLOOK

solutions yielding dense PEG brushes. Thus, the PEGylation process is fast and uses very mild conditions, which is an important aspect for future incorporation of less stable entities, e.g. antigens. Using appropriate copolymer architecture, the PEG brush was shown to be dense enough to repel the adsorption of proteins from human serum [9]. Secondly, it was already shown that bioligands could be covalently linked to the PEG chains prior to the adsorption step [10, 11].

PLL-g-PEG coated stealth microparticles

As hypothesized, we found that the resistance of surface modified microspheres to phagocytosis was mostly dependent on their repellence to the adsorption of proteins. Screening a small library of PLL-g-PEG copolymers, we found protein adsorption to decrease with decreasing grafting ratio and increasing PEG chain length, which was in agreement with the findings for different PEGylation strategies by other authors. We could also confirm that PLL(20)-g[3.5]-PEG(2), i.e. 2 kDa PEG chains grafted to a 20 kDa PLL backbone at a grafting ratio of g=3.5, has an ideally suited architecture to achieve protein repellence and resist cell adhesion and phagocytosis of coated microspheres.

Importantly, we could prove that mixed adlayers composed of different PLL-g-PEG copolymers could be assembled onto the surface of microspheres. This is of particular importance for the use of such microspheres to target cells, as it enables the decoration of the microspheres’ surface with a mixture of cell targeting ligands, which can be linked to PEG side chains of individual PLL-g-PEG copolymers. Controlled surface compositions can thus be obtained through variation of the proportions of differently functionalized and unfunctionalized copolymers in the coating mixture. Thus, our microsphere system allows for a convenient adjustment of ligand density and composition.

We further studied the stability of the PLL-g-PEG adlayer to exclude its disintegration upon storage, or during the experiments. The adlayers turned out to be stable even under cell culture conditions (i.e. stored at 37°C in cell culture medium supplemented with 5% human serum) for at least one month, without

GENERAL DISCUSSION AND OUTLOOK 171

apparent loss of stealth function. This aspect was not only important in view of the experimental setup, but was indispensable for future applications.

As a further step towards possible applications as drug delivery system, we were interested to replace the model PS microspheres by other microparticles. Considering that the adsorption mechanism was based on electrostatic interaction, it should be possible to apply PLL-g-PEG coatings to any type of microparticles, provided that they bear a sufficient and homogeneously distributed negative surface charge. Indeed, we demonstrated that hollow microcapsules prepared from layer-by-layer assembly of oppositely charged polyelectrolytes could be efficiently passivated against phagocytosis through adsorption of a PLL-g-PEG adlayer. Two particle sizes (1 and 5 μm) were successfully tested.

It should be noted, however, that the transferability of the PEGylation process to any kind of microparticles cannot be taken for granted. A first indication for a limitation to microparticles of certain surface characteristics is discussed in Chapter III, where we observed different degrees of stealth for different microcapsule batches. Moreover, in a previous study [12] it was documented that the internalization of PLL-g-PEG coated PLGA microparticles by phagocytes was strongly reduced as compared to uncoated PLGA microparticles, with a similar degree of reduction as those reported in this dissertation. However, in spite of manifold attempts, this work could never be fully reproduced in our laboratory. Even the adsorption of three double layers of PAH/PSS prior to PEGylation, intended to stabilize the microparticles’ surface in analogy to the polyelectrolyte microcapsules, yielded highly irreproducible results; i.e. samples were protected from phagocytosis in some experiments but easily phagocytosed in others (B. Hedinger and U. Wattendorf, unpublished results). Furthermore, while PLL-g-PEG coatings on PS microspheres that were obtained from one company (Micromod Partikeltechnologie GmbH, Germany) were perfectly reproducible, microspheres obtained from another company (Polyciences Europe GmbH, Germany) could not be efficiently coated (U. Wattendorf, unpublished results). Therefore, it seems that the quality of the microspheres’ surface is of vital importance for the outcome of the PEGylation. Although we could not provide an ultimate proof for our hypothesis, we

172 GENERAL DISCUSSION AND OUTLOOK

postulated that a homogeneous surface charge distribution is pivotal for the formation of a smooth and defect-free adlayer of PLL-g-PEG, or indeed of any graft copolymer. This appears reasonable given the fact that small surface defects in the PEG brush may lead to partial opsonization, which in turn enables internalization by phagocytes.

Targeting of APC receptors

Microspheres exposing either mannan or small substructure mannoside ligands, such as mono- and tri-mannose, could be readily and rapidly fabricated through fast and convenient self-assembly protocols. While substructure ligands were covalently coupled to the PEG chains prior to PLL-g-PEG/PEG-ligand adsorption, mannan was electrostatically adsorbed onto PEG-free, PLL coated sites of the microspheres. In both cases, ligand densities were conveniently controlled by the amount of (unmodified) PLL-g-PEG in the adsorption mixture; i.e. either PLL-g-PEG/PEG-ligand or PLL (the latter for subsequent mannan adsorption) were mixed at defined ratios with PLL-g-PEG. Combinations of different types of ligands were equally feasible, as demonstrated for a combination of a mannoside ligand with an integrin homing RGD peptide. Most importantly, all of the coatings were repellent against unspecific adsorption of proteins, thus providing an opportunity for specific cellular interactions between the ligands exposed on the microspheres’ surface and receptors on APCs.

Although mannose density was identified as the major factor for the phagocytosis of mono- and tri-mannose decorated microspheres, the efficiency of internalization was limited. While the mannose receptor (MR) was traditionally considered a constitutive phagocytic receptor, i.e. not requiring interactions with partner receptors, this view was recently questioned by other authors [13]. Such doubt was further strengthened by our results. Overall, our microparticulate platform will enable further studies including combinations of various ligands to address potential co-receptors, which would certainly help to identify such receptors.

Finally, a striking outcome of our study was the failure of surface assembled mannan to stimulate DC maturation which contrasts with its well

GENERAL DISCUSSION AND OUTLOOK 173

documented capacity for immunomodulation. Therefore, we concluded that recognition by CLRs alone and subsequent phagocytosis of the microspheres could not trigger DC activation towards a T helper response by itself. Instead, the co-processing of a suitably encapsulated antigen and/or the engagement of further cell surface receptors seems to be mandatory.

In summary, the investigated microparticulate constructs represent a promising tool for the investigation of specific ligand-receptor interactions on phagocytes, including the screening of potential ligands and ligand combinations, and the analysis of immunomodulatory effects thereof. Also the effect of co-delivery of antigens will be highly interesting, and may be investigated through the use of hollow microcapsules or biodegradable microparticles. We assume our microparticulate platform to prove useful in order to deepen the general knowledge on the impact of the first encounter of APC with micron-sized microparticles on the downstream immune cascade, which may ultimately lead to improved vaccine formulations.

174 GENERAL DISCUSSION AND OUTLOOK

REFERENCES 1. Gogolák, P., Réthi, B., Hajas, G., and Rajnavölgyi, É. Targeting dendritic

cells for priming cellular immune responses. J. Mol. Recognit., 2003. 16(5): p. 299-317.

2. Toda, S., Ishii, N., Okada, E., Kusakabe, K.I., Arai, H., Hamajima, K., Gorai, I., Nishioka, K., and Okuda, K., HIV-1-specific cell-mediated immune responses induced by DNA vaccination were enhanced by mannan-coated liposomes and inhibited by anti-interferon-gamma antibody. Immunology, 1997. 92(1): p. 111-117.

3. Cui, Z., Han, S.-J., and Huang, L., Coating of Mannan on LPD Particles Containing HPV E7 Peptide Significantly Enhances Immunity Against HPV-Positive Tumor. Pharm. Res., 2004. 21(6): p. 1018-1025.

4. Jain, S., and Vyas, S.P., Mannosylated niosomes as adjuvant-carrier system for oral mucosal immunization. J Liposome Res, 2006. 16(4): p. 331-345.

5. Apostolopoulos, V., Pietersz, G.A., Tsibanis, A., Tsikkinis, A., Drakaki, H., Loveland, B.E., Piddlesden, S.J., Plebanski, M., Pouniotis, D.S., Alexis, M.N., McKenzie, I.F., and Vassilaros, S., Pilot phase III immunotherapy study in early-stage breast cancer patients using oxidized mannan-MUC1 [ISRCTN71711835]. Breast Cancer Res, 2006. 8(3): p. R27.

6. White, K., Rades, T., Kearns, P., Toth, I., and Hook, S., Immunogenicity of liposomes containing lipid core peptides and the adjuvant Quil A. Pharm. Res., 2006. 23(7): p. 1473-1481.

7. Hattori, Y., Kawakami, S., Lu, Y., Nakamura, K., Yamashita, F., and Hashida, M., Enhanced DNA vaccine potency by mannosylated lipoplex after intraperitoneal administration. J Gene Med, 2006. 8(7): p. 824-834.

8. Arshady, R., and Monshipouri, M., Targeted delivery of microparticulate carriers. Microspheres, Microcapsules & Liposomes, 1999. 2(Medical & Biotechnology Applications): p. 403-432.

9. Pasche, S., De Paul, S.M., Vörös, J., Spencer, N.D., and Textor, M., Poly(L-lysine)-graft-poly(ethylene glycol) assembled monolayers on

GENERAL DISCUSSION AND OUTLOOK 175

Niobium Oxide surfaces: a quantitative study of the influence of polymer interfacial architecture on resistance to protein adsorption by ToF-SIMS and in situ OWLSu OWLS. Langmuir, 2003. 19(22): p. 9216-9225.

10. VandeVondele, S., Vörös, J., and Hubbell, J.A., RGD-grafted poly-L-lysine-graft-(polyethylene glycol) copolymers block non-specific protein adsorption while promoting cell adhesion. Biotechnol. Bioeng., 2003. 82(7): p. 784-790.

11. Müller, M., Vörös, J., Csúcs, G., Walter, E., Danuser, G., Merkle, H.P., Spencer, N.D., and Textor, M., Surface modification of PLGA microspheres. J. Biomed. Mater. Res., Part A, 2003. 66A(1): p. 55-61.

12. Faraasen, S., Vörös, J., Csúcs, G., Textor, M., Merkle, H.P., and Walter, E., Ligand-Specific Targeting of Microspheres to Phagocytes by Surface Modification with Poly(L-Lysine)-Grafted Poly(Ethylene Glycol) Conjugate. Pharm. Res., 2003. 20(2): p. 237-246.

13. Le Cabec, V., Emorine, L.J., Toesca, I., Cougoule, C., and Maridonneau-Parini, I., The human macrophage mannose receptor is not a professional phagocytic receptor. J. Leukoc. Biol., 2005. 77(6): p. 934-943.

176 GENERAL DISCUSSION AND OUTLOOK

APPENDIX 177

APPENDIX I: Supplementary Figures for Chapter II Figure A.I-1: Phagocytosis of Microspheres by DC and MΦ. (A, B) DC were incubated with PLL-g-PEG coated and uncoated (carboxylated) PS microspheres for 4 h. Samples were not washed to show microspheres that were not phagocytosed. The microphotographs demonstrate the close packing of uncoated control microspheres inside the DC (A), while PLL-g-PEG coated microspheres (B) were not phagocytosed within the timeframe of the experiment. (C) Coincubation of both uncoated and PLL-g-PEG coated PS microspheres confirmed the capacity of MΦ for efficient distinction between uncoated (greeen fluorescence) and PLL-g-PEG coated PS microspheres (no fluorescence), so that only uncoated PS microspheres were phagocytosed by MΦ. Experimental details: (A) DC with uncoated 5 μm PS microspheres. (B) DC with 5 μm PS microspheres coated with PLL(20)-[3.5]-PEG(2). (C) MΦ with 5 μm PS microspheres coated with PLL(20(-[3.5]-PEG(2), and uncoated 4.5 μm FITC-labelled PS microspheres added 1 hour after the coated microspheres.

Figure A.I-2: Intracellular Localization of Phagocytosed Microspheres. Intracellular localization of phagocytosed microspheres is demonstrated by combination of CLSM and DIC. (A) x-y overlay of DIC and fluorescence channels showing an accumulation of phagocytosed microspheres tightly enclosed by the MΦ´s cell membrane; (B) actin cytoskeleton staining in combination with nucleus staining; and (C) cell membrane staining in combination with nucleus staining. (B and C) In all CLSM projections the non-fluorescent microspheres appear as dark areas enclosed by cytoskeleton and cell membrane. Experimental details were as follows: (A) DIC image overlayed with x-y optical sections of Phalloidin-OregonGreen488,

ConcanavalinA-AlexaFluor633 and Hoechst-33342 channels. (B) x-y optical section and x-z and y-z projections of Phalloidin-OregonGreen488 and

Hoechst-33342 channels. (C) x-y optical section and x-z and y-z projections of ConcanavalinA-AlexaFluor633 and

Hoechst-33342 channels.

MΦ were cultured in Lab-Tek 8 well chamber glass slides (Nalge Nunc Int., IL, USA) and incubated with uncoated PS microspheres or PLL(20)-[10.1]-PEG(2) coated microspheres as described in the Materials section. Cells were stained with ConcanavalinA-AlexaFluor633 (cell membrane), Phalloidin-OregonGreen488 (actin cytoskeleton) and Hoechst-33342 (nucleus; all Molecular Probes Inc., Eugene, OR, USA) and examined using a Zeiss LSM 510 META (Carl Zeiss AG, Feldbach, Switzerland; lasers: He-Diodes 405 nm, Ar 488/514 nm, HeNe 543 nm, He/Ne 633 nm). Image processing was performed with Imaris (Bitplane, Zurich, Switzerland). DC yielded similar results.

178 APPENDIX 178 APPENDIX

Figure A.I-1

APPENDIX 179

Figure A.I-2

180 APPENDIX

Figure A.I-3

Figure A.I-3: Side Scatter Analysis for the Assessment of Phagocytosis by DC. To quantify phagocytosis by side scatter analysis, the DC population was identified in the SSC/FSC dotplot (region R1 in A, region R2 in B). SSC-histograms were gated on this region and the mean SSC was calculated by the evaluation software (Cytomation Summit, Cytomation Inc., Fort Collins, Co, USA). (A) DC in the absence of PS microspheres; (B) DC after 4 h incubation and phagocytosis of carboxylated PS microspheres; (C) overlay of gated SSC from A (green) and B (red).

APPENDIX 181

APPENDIX II: Characterization of Microcapsules Figure A.II-1: Confirmation of polyelectrolyte deposition by ζ-potentials.

To monitor the deposition of the polyelectrolyte layers, ζ-potentials (in water) were determined after washing. Electrophoretic mobilities were measured using a Malvern Zetasizer 3000HS and converted into ζ-potentials using the Helmholtz-Smoluchowski relation (ζ = ν / η ε E, where ν = particle mobility, η, ε = viscosity and permittivity of the solution, E = electric field strength).

The figure shows a representative preparation of PSS terminated 1.2 μm capsules. Four double layers of PAH / PSS were deposited onto the MF template, indicated by alternating ζ-potentials. No significant change was observed during the dissolution of the template. Finally, a last layer of PSS was applied, with a corresponding drop in the ζ-potential.

Figure A.II-2: Integrity of microcapsules by CLSM.

To confirm the microcapsules’ integrity and the homogeneity of the fluorescent labelling, samples were visualized with a confocal laser-scanning system TCLS attached to an inverse microscope from Leica (Wetzlar, Germany) that was equipped with a 100x oil immersion objective having a numerical aperture of 1.4. A: 4.8 μm capsules; B: 1 μm capsules (left: transmitted light; right: fluorescence image.)

182 APPENDIX

Figure A.II-3: Average Layer Thickness by Scanning Force Microscopy (SFM).

SFM measurements were performed in air at room temperature using a Nanoscope III Multimode SFM (Digital Instruments Inc.) operating in tapping mode. Samples were prepared by placing a drop of the sample solution onto a freshly cleaved mica substrate and drying it at room temperature. The thickness of the (collapsed) capsule walls was measured from the flat regions of the profile, and the average layer thickness was calculated from division by the number of layers.

The figure shows representative images for 4.8 μm (A-C) and 1.2 μm (D-F) capsules. The increased noise observed for samples of 1.2 μm capsules is due to the smaller capsule size. Average layer thicknesses of about 2.5 and 2.3 nm (±0.2 nm) for microcapsules of 4.8 and 1.2 μm diameter, respectively, were obtained. A, D: 3D-images; B, E: 2D-images indicating a cross-section for height determination; C, F: cross-sections indicating heights of mica substrates and collapsed microcapsules.

APPENDIX 183

APPENDIX III: Supporting Information for Chapter IV In situ analysis of polymer and protein adsorption on flat surfaces by optical methods

Dual Polarization Interferometry (DPI) and Optical Waveguide Lightmode Spectrometry (OWLS) were used to monitor in situ the adsorbed mass of polymers and proteins. Both are optical methods using the evanescent field of a laser light to sense biomolecules up to ∼ 100 nm from the surface. To calculate the adsorbed amounts the DPI measures the changes of an interference pattern while the biomolecules are deposited onto the surface. The OWLS measures the changes of the angle of incidence at which the laser light is incoupled into a waveguide. In both methods the measurements give information about the refractive index and thickness of the adsorbed layer, which can be used to calculate the adsorbed amounts. More detailed descriptions about the instrumentation and theories behind the OWLS method [1-3] or the DPI method [4, 5] can be found elsewhere. The de Feijter formula (Equ.1) [6] was used to quantify the adsorbed mass (ng/cm2) from the film thickness da (nm), refractive index of the film (na) and refractive index of the buffer (nc). A refractive index increment (dn/dc) of 0.182 cm3/g was used for the protein solutions [6] and 0.145 cm3/g for the PLL-g-PEG/PEG-mannoside polymers as determined by refractometry.

Γ = dana − nc

dn/dc (Equ. 1)

OWLS instrumentation

The measurements were carried out with the OWLS 110 system with OW2400 sensor chips (Micro Vacuum Ltd., Budapest, Hungary). The adsorbing substrates are OWLS sensor chips consisting of a 200 nm thick Si0.75Ti0.25O2 waveguiding surface layer deposited onto a 1 mm thick AF45 glass substrate purchased from Micro Vaccum, Budapest. The glycopolymer-functionalized interfaces were prepared in situ using a flow-through chamber (16 μl) followed by incubation of the surfaces with the protein solutions.

184 APPENDIX

DPI instrumentation

The measurements were carried out with the AnaLight Bio200® DPI instrument from Farfield Sensors Ltd. (Crewe, UK) with triple channel chips mounted in a thermostated flow-cell. Two channels were used for adsorption experiments, while the third channel was used only as a reference containing no liquid flow. The flow of the buffer solution in the sensing region was varied according to the experimental needs as followed: Polymers and proteins were injected at a flow rate of 15 μl/min followed by incubation at a flow rate of 5 μl/min and rinsing with buffer at a flow rate of 50 μl/min.

Substrates

A thin layer of a high refractive index niobium oxide (Nb2O5, thickness ≈ 6 nm) was sputter coated on the top of the waveguides using reactive magnetron sputtering (Leybold Z600, PSI Villigen, Switzerland). Prior experiments substrates were cleaned for 10 min in 0.1 N HCl, sonicated for 10 min in 2-propanol, rinsed with Millipore water, dried under a nitrogen stream and exposed to oxygen plasma (plasma cleaner/sterilizer PDC-32G instrument, Ossining, NY) for 2 min. The cleaned OWLS waveguides were equilibrated overnight in HEPES buffer solution prior to the measurements. Before starting the experiment the DPI chips were calibrated using a degassed solution of 80/20-ethanol/water solution to rate the sensitivity of the waveguide to the bulk refractive index allowing for a higher accuracy in the quantification of the adlayer refractive index and thickness.

Adsorption experiments

The glycopolymers (0.1 mg/ml in HEPES buffer, 150 mM NaCl) were adsorbed on the waveguides (OWLS, DPI) during ca. 45 min and rinsed with HEPES buffer (150 mM NaCl) for 15 min. The surfaces were then tested against nonspecific protein adsorption from a solution of human serum (Precinorm U, Roche Diagnostics GmbH, Mannheim, Germany) for 15 min. For bioaffinity assays, freshly prepared surfaces were incubated with a solution of the lectin concanavalin A (ConA, Sigma-Aldrich, molecular weight of the tetramer

APPENDIX 185

∼ 104 kDa) for 20 min (0.1 mg/ml in HEPES buffer, 150 mM NaCl, 1 mM CaCl2, 1 mM MnCl2). After incubation with biomolecules, the surfaces were flushed with HEPES buffer for ca. 15 min for removal of unbound polymers or proteins. The protein adsorption studies to the PLL-PEG-PrMan were performed with the DPI and OWLS. The adsorption experiments with the PLL-PEG-EGMan (50%) were performed with the OWLS only. The adsorption experiments with the PLL-PEG-EGTri-Man(50%) were performed with the DPI only.

The availability of the mannose ligands for specific molecular recognition was demonstrated by specific interaction with the lectin ConA. A typical experiment is shown in Figure A.III-1. The data are summarized in Table A.III-1 and Figure A.III-2.

The adsorption of human serum proteins to PLL-PEG-mannoside coated waveguides is summarized in Table A.III-1. For all glycopolymers, adsorption of serum protein was reduced by more than 98% compared to untreated niobium oxide control surfaces (590 ± 57 ng/cm2) [7]. A typical experiment is shown in Figure A.III-3.

Degree of substitution of the synthesized PLL-PEG glycopolymers determined by 1H-NMR

Details to the methodology and a summary of the results can be found in the manuscript in the Methods section and in Table 1, respectively. Further information on the methodology and corresponding NMR spectra are presented in Table A.III-2 and Figures A.III-4 to 9 in this supporting information.

186 APPENDIX

REFERENCES 1. Höök, F., Vörös, J., Rodahl, M., Kurrat, R., Böni, P., Ramsden, J.J.,

Textor, M., Spencer, N.D., Tengvall, P., Gold, J., and Kasemo, B., A comparative study of protein adsorption on titanium oxide surfaces using in situ ellipsometry, optical waveguide lightmode spectroscopy, and quartz crystal microbalance/dissipation. Colloid Surface B, 2002. 24(2): p. 155-170.

2. Kurrat, R., Textor, M., Ramsden, J.J., Böni, P., and Spencer, N.D., Instrumental improvements in optical waveguide light mode spectroscopy for the study of biomolecule adsorption. Rev. Sci. Instrum., 1997. 68(5): p. 2172-2176.

3. Vörös, J., Ramsden, J.J., Csucs, G., Szendro, I., De Paul, S.M., Textor, M., and Spencer, N.D., Optical grating coupler biosensors. Biomaterials, 2002. 23(17): p. 3699-3710.

4. Freeman, N.J., Peel, L.L., Swann, M.J., Cross, G.H., Reeves, A., Brand, S., and Lu, J.R., Real time, high resolution studies of protein adsorption and structure at the solid-liquid interface using dual polarization interferometry. J Phys-Condens Mat, 2004. 16(26): p. S2493-S2496.

5. Cross, G.H., Reeves, A.A., Brand, S., Popplewell, J.F., Peel, L.L., Swann, M.J., and Freeman, N.J., A new quantitative optical biosensor for protein characterisation. Biosensors & Bioelectronics, 2003. 19(4): p. 383-390.

6. Defeijter, J.A., Benjamins, J., and Veer, F.A., Ellipsometry as a Tool to Study Adsorption Behavior of Synthetic and Biopolymers at Air-Water-Interface. Biopolymers, 1978. 17(7): p. 1759-1772.

7. Pasche, S., De Paul, S.M., Vörös, J., Spencer, N.D., and Textor, M., Poly(L-lysine)-graft-poly(ethylene glycol) assembled monolayers on Niobium Oxide surfaces: a quantitative study of the influence of polymer interfacial architecture on resistance to protein adsorption by ToF-SIMS and in situ OWLSu OWLS. Langmuir, 2003. 19(22): p. 9216-9225.

APPENDIX 187

Table A.III-1: Summary of the adsorption data determined by OWLS and DPI for the PLL-g-PEG/PEG-mannoside on Nb2O5 as substrate surface.

Table A.III-2: Summary of the degrees of substitution of the PLL-PEG-PrMan glycopolymers determined by 1H-NMR data and from the calibration curve of ConA adsorption on PLL-PEG-PrMan coated OWLS waveguides.

188 APPENDIX

Figure A.III-1: Typical in situ measurement of the adsorption of PLL-PEG-PrMan(50%) from solution onto a Nb2O5-coated waveguide chip and subsequent exposure to the lectin ConA, monitored by OWLS.

Figure A.III-2: Comparative adsorption of the protein ConA to the glycopolymers PLL-g-PEG/PEG-mannoside immobilized onto Nb2O5 as substrate surface, monitored by OWLS.

APPENDIX 189

Figure A.III-3: Typical in situ measurement of the adsorption of PLL-PEG-PrMan(50%) from solution on Nb2O5-coated waveguide chip and subsequent exposure to human serum, monitored by DPI.

190 APPENDIX

Figure A.III-4: Calibration curve for the adsorption of ConA to PLL-PEG-PrMan glycopolymers monitored by OWLS a) versus the mannoside density of functionalized PLL-PEG-PrMan(x%) as determined by 1H-NMR, and b) versus the molar fraction of PEG-PrMan per glycopolymer as determined by 1H-NMR.

APPENDIX 191

Figure A.III-5: 1H-NMR spectrum of PLL-PEG-PrMan(20%) (500 MHz, 300 K) and assignments of the peaks according to the chemical formula.

192 APPENDIX

Figure A.III-6: 1H-NMR spectrum of PLL-PEG-PrMan(50%) (D2O, 300 K, 500 MHz).

APPENDIX 193

Figure A.III-7: 1H-NMR spectrum of PLL-PEG-PrMan(100%) (D2O, 300 K, 500 MHz).

194 APPENDIX

Figure A.III-8: 1H-NMR spectrum of PLL-PEG-EGMan(50%) (D2O, 300 K, 500 MHz).

APPENDIX 195

Figure A.III-9: 1H-NMR spectrum of PLL-PEG-EGTri-Man(50%) (D2O, 300 K, 500 MHz).

196 APPENDIX

CURRICULUM VITAE 197

CURRICULUM VITAE

Uta Wattendorf born in Münster, Germany

on September 1st, 1977

Education and Postitions held 03/2003 – present Postgraduate Studies and Teaching Assistant Swiss Federal Institute of Technology Zurich (ETH Zurich), Switzerland Institute of Pharmaceutical Sciences Drug Formulation and Delivery (Prof. H. P. Merkle) 03/2002 – 09/2002 Diploma Thesis and Student Research Assistant

„Einschränkung unspezifischer Wechselwirkungen von Lipidmembranen auf nanostrukturierten Substraten“

Ruprecht-Karls-Universität University of Heidelberg, Germany Institute of Physical Chemistry Biophysical Chemistry (Prof. J. P. Spatz) 04/2001 – 11/2001 Student Research Assistant Ruprecht-Karls-Universität University of Heidelberg, Germany Institute of Physical Chemistry Biophysical Chemistry (Prof. J. P. Spatz)

198 CURRICULUM VITAE

10/1999 – 08/2000 Undergraduate Studies in Chemistry Universidad Complutense, University of Madrid, Spain

10/1996 – 09/2002 Undergraduate Studies in Chemistry

Ruprecht-Karls-Universität, University of Heidelberg, Germany

06/1996 – 08/1996 Internship at Wessling Laboratories GmbH, Altenberge, Germany 09/1988 – 06/1996 High School Gymnasium Arnoldinum, Steinfurt, Germany

Scientific Contributions Journal Publications U. Wattendorf, M. C. Koch, E. Walter, J. Vörös, M. Textor and H. P. Merkle Phagocytosis of PLL-g-PEG coated microspheres by antigen presenting cells: Impact of grafting ratio and PEG chain length on cellular recognition. Biointerphases, 2006. 1(4): pp. 123-133. U. Wattendorf, O. Kreft, M. Textor, G. B. Sukhorukov, and H. P. Merkle Stable Stealth Function for Hollow Polyelectrolyte Microcapsules through a Polyethylene Glycol Grafted Polyelectrolyte Adlayer. Biomacromolecules, 2008. 9(1): pp. 100-108. U. Wattendorf and H. P. Merkle PEGylation as a Tool for the Biomedical Engineering of Surface Modified Microparticles. Journal of Pharmaceutical Sciences, 2008. DOI 10.1002/jps.21350.

CURRICULUM VITAE 199

U. Wattendorf, G. Coullerez, J. Vörös, M. Textor, and H. P. Merkle Impact of Mannose-Based Molecular Patterns on Human Antigen Presenting Cells: A Study with Stealth Microspheres. In preparation. Oral Presentations U. Wattendorf, et al., Targeting of antigen presenting cells with carbohydrate functionalized microspheres. Zurich-Geneva Joint Seminar Series on Drug Formulations & Drug Delivery, Geneva, Switzerland, February 15-16, 2007 U. Wattendorf, et al., Targeting of dendritic cell receptors with surface modified microparticles, Zurich-Mainz Joint Seminar Series on Polymers, Mainz, Germany, September 21-11, 2006 U. Wattendorf, et al., Surface modified microspheres for the targeting of antigen-presenting cells, 5th World Meeting on Pharmaceutics, Biopharmaceutics and Pharmaceutical Technology, Geneva, Switzerland, March 27-30, 2006 U. Wattendorf, et al., Surface Modified Microspheres for the Targeting of Antigen-Presenting Cells, Doktorandentag ETH Zurich, Zurich, Switzerland, October 26, 2005 U. Wattendorf, et al., Specific Targeting of DC Receptors with Surface Modified Microparticles, Workshop: Internalisation of colloidal particles and capsules with biological cells, Munich, Germany, December 02, 2004 U. Wattendorf, et al., Specific Targeting of DC Receptors with Surface Modified Microparticles, Zurich-Geneva Joint Seminar Series on Drug Formulations & Drug Delivery, Zurich, Switzerland, October 07-08, 2004 U. Wattendorf, et al., Specific Targeting of DC Receptors with Microparticulate Carriers, International Conference & Workshop on Physical Chemistry of Bio-Interfaces, Barossa Valley, Australia, May 23-26, 2004

200 CURRICULUM VITAE

Poster Presentations U. Wattendorf, et al., Surface Targeting of DC Receptors with Sugar Modified Microparticles, 1st MRC (Materials Research Center) Graduate Symposium, ETH Zurich, Zurich, Switzerland, June 29, 2006 U. Wattendorf, et al., Specific Targeting of DC Receptors with Sugar Modified Microparticles, Polymers in Life Sciences (PILS), Basel, Switzerland, March 22-23, 2005 U. Wattendorf, et al., Specific Targeting of DC Receptors with Microparticulate Carriers, Joining Forces: Intersections between Biology, Chemistry, Engineering, Informatics and Physics, ETH Zurich, Zurich, Switzerland, October 14, 2004 U. Wattendorf, et al., Surface Modified PLGA Microspheres as a DNA Vaccine Delivery System, Pharma-Day, Pharmacenter Basel, Basel, Switzerland, July 11, 2003 U. Wattendorf, et al., Surface Modified PLGA Microspheres as a DNA Vaccine Delivery System, Biological Surfaces and Interfaces (EURESCO), Castelvecchio Pascoli, Italy, June 21-26, 2003 U. Wattendorf, et al., Biochemical Surface Functionalization of Carriers for Targeted Drug Delivery Applications, Joining Forces: Intersections between Biology, Chemistry, Engineering, Informatics and Physics, ETH Zurich, Zurich, Switzerland, March 2003

ACKNOWLEDGEMENTS 201

ACKNOWLEDGEMENTS

Above all, I would like to thank my supervisor Prof. Hans P. Merkle, Head of the Drug Formulation & Delivery Group at ETH Zurich, for the opportunity to perform my PhD studies in this highly interesting field on the intersection of immunology and material science. I am grateful for his continuous support and valuable advice.

In addition, I am deeply grateful to my co-supervisor Prof. Marcus Textor, Head of the BioInterface Group at ETH Zurich, for “adopting” me as a member of his own group, and for his support and profound advice on all materials science’s aspects of this PhD thesis. Furthermore, I am thankful for his encouragement during difficult phases of my project.

My sincerest thanks also go to my co-examiner Prof. Janos Vörös, Head of the Laboratory of Biosensors and Bioelectronics at ETH Zurich. I especially appreciated his numerous valuable ideas, and his motivating enthusiasm.

I’d also like to thank PD Dr. Elke Walter (ETH Zurich) for invaluable advice particularly with regard to the initial phase of the project.

Special thanks go to Dr. Oliver Kreft (MPI for Colloids and Interfaces, Germany) for the excellent collaboration in the field of polyelectrolyte microcapsules. I especially enjoyed the many discussions on scientific (and non-scientific) topics – both at the MPI, at ETH, and in countless phone calls – and his readiness to help wherever needed.

I’d also like to express my gratitude to Dr. Géraldine Coullerez (BioInterface Group at ETH Zurich) for synthesis and characterization of the mannoside functionalized PLL-g-PEG copolymers and for her input in the corresponding chapter of this thesis. I’d also like to thank Katrin Barth for assisting in the synthesis, and especially for some very interesting discussions. And I’m equally grateful to yet another (former) member of the BioInterface Group, Dr. Stéfanie Pasche, who introduced me to the “magic” of PLL-g-PEG copolymer architectures, and who synthesized the library of PLL-g-PEG’s that was used in the second chapter of this thesis.

202 ACKNOWLEDGEMENTS

I gratefully acknowledge the contributions from my students: Barbara Hediger (diploma thesis), Emanuela Fusi (semester & diploma thesis) and Mirabai Koch (Hilfsassistentin). It was a pleasure to accompany them during their projects, and although not all projects led to scientific publications, also the “negative results” nonetheless contributed to the methodology and the final outcome of this thesis.

Many thanks go to all members of the Drug Formulation & Delivery Group. First of all, I’d like to thank the “surface-modification sub-group” people: Dr. Noémi Csaba, Stefan Fischer, Prof. Bruno Gander, Annina Hafner, and (most recently) Dr. Silvina Bravo. It was especially nice when Annina and Noémi joined to share the experiences on the dendritic cells. Next to them, I’d like to mention the people with whom I had the pleasure to share the office/lab: Christina Förg, Michael Herbig, and Katrin Weller in the “original” J-398 lab; Noémi, Annina and Silvina later in J-398; and Annette Koch, Rachel Tréhin, and (once more) Christina in the early days at K-28. Likewise, I am very happy for the time spent – for work, lunch, coffee breaks or evenings – with Ruth Alder, Sergio Freitas, Marcos García Fuentes, Henri Hagenmüller, Sandra Hofmann, Chunmei Li, Vera Luginbühl, Srinivas Madduri, Lukas Pfister, Lorenz Uebersax, Anne Wandrey, and Esther Wenk. Next to our “Pharma Group”, a big “thank you” to all people from the LSST – unfortunately far too many to name you all, but I definitely enjoyed the great atmosphere whenever I could make it to join.

Last but not least, I’d like to thank all my friends, my family, and especially Nils. Without all of you, I would not be who I am, nor would I have succeeded.