Review

Immune responses to implants e A reviewof immunomodulatory biomaterials

Sandra Franz a,d,*, Stefan Rammelt b,d, Dieter ScharnwaDepartment of Dermatology, Venerology and Allergology, University Leipzig, 04103 LeipzigbDepartment of Trauma and Reconstructive Surgery, Dresden University Hospital Carl GuscMax Bergmann Center for Biomaterials, University of Technology Dresden, 01069 Dresden,dCollaborative Research Center (SFB-TR67), Matrixengineering, Department of Dermatology

a r t i c l e i n f o

Article history:Received 15 March 2011Accepted 26 May 2011

Immune response to biomaterialsImmunomodulation

a b s t r a c t

A key for long-term survival anresponse. As biomaterials can hto design biomaterials that are able to trigger desired immunological outcomes and thus support the

host tissue.This review outlines the current knowledge on immune

responses to biomaterials and debates on biomaterials designed toelicit appropriate host immune responses. Evidence will be pre-sented that biomaterials mimicking the physiological extracellularmatrix (ECM) may have the potential to modulate the immunesystem by enhancing or suppressing normal immune cellfunctions.

morphonuclear leukocytes; PRR, pattern recognition receptors; RGD,arginineeglycineeaspartic acid; ROS, reactive oxygen species; TF, tissue factor;TGF-b, transforming growth factor beta; TH1, T helper 1 cells; TIMP, tissue inhibitorof metalloproteinases; TLR, toll-like receptors; TNFa, tumor-nekrose-faktor alpha;TReg, regulatory T lymphocytes; VEGF, vascular endothelial growth factor.* Corresponding author. Department of Dermatology, Venerology and Allergol-

ogy, Max Brger Research Centre, University Leipzig, Johannisallee 30, 04103Leipzig, Germany. Tel.: 49 341 9725880; fax: 49 341 9725878.

Contents lists availab

Biomat

journal homepage: www.elsev

Biomaterials 32 (2011) 6692e6709E-mail address: sandra.franz@medizin.uni-leipzig.de (S. Franz).designed to minimize potentially deleterious host responses, Wil-liams recently dened a biomaterial as .a substance that has beenengineered to take a form which, alone or as part of a complex system,is used to direct, by control of interactions with components of livingsystems, the course of any therapeutic or diagnostic procedure,. [1].All materials when implanted into living tissue initiate a hostresponse that reects the rst steps of tissue repair. Modernimplant design is directed on making use of this immune responseto improve implant integration while avoiding its perpetuationleading to chronic inammation and foreign body reactions, andthus loss of the intended function [2]. Directing these processesrequires an in-depth understanding of the immunologicalprocesses that take place at the interface between biomaterial and

Abbreviations: aECM, articial ECM; BMP, bone morphogenetic protein; CCL, CCchemokine ligand; COX-2, cyclooxygenase-2; CS, chondroitin sulfate; CXCL, CXCchemokine ligand; DC, dendritic cells; DC-STAMP, dendritic cell-specic trans-membrane protein; ECM, extracellular matrix; EGF, epidermal growth factor; ENA-78, epithelial cell-derived neutrophil attractant-78; ERK, extracellular signal-regulated kinases; FBGC, foreign body giant cells; FBR, foreign body response;FGF, broblast growth factor; GAGs, glycosaminoglycans; GM-CSF, granulocytemacrophage colony stimulating factor; HA, hyaluronic acid; HMW-HA, highmolecular weight HA; IL, interleukin; IL-1ra, IL-1 receptor antagonist; IFNg, inter-feron gamma; LPS, lipopolysaccharide; MAPK, mitogen-activated protein kinase;MCP, monocyte chemotactic protein; MDC, macrophage-derived chemokine; MFI,mean uorescence index; MHC-II, major histocompatibility complex class IImolecules; MIP, macrophage inammatory protein; MMPs, matrix metal-loproteinases; NO, nitric oxide; NOS2, nitric oxide synthase 2; PAMPs, pathogen-associated molecular patterns; PDGF, platelet-derived growth factor; PEG, poly-ethylene glycol; PEO, poly(ethylene oxide); PGs, proteoglycans; PMN, poly-Extracellular matrixBiomaterial integration0142-9612/$ e see front matter 2011 Elsevier Ltd.doi:10.1016/j.biomaterials.2011.05.078Understanding the complex interactions of host response and material implants reveals the need for andalso the potential of immunomodulating biomaterials. Based on this knowledge, we discuss strategiesof triggering appropriate immune responses by functional biomaterials and highlight recent approachesof biomaterials that mimic the physiological extracellular matrix and modify cellular immune responses.

2011 Elsevier Ltd. All rights reserved.

1. Introduction

The paradigm on the nature of biomaterials has been intenselyrened within the last decades. Rather than being an inert materialKeywords:

inammatory and wound healing response to implanted materials.One focus of this review is to outline the up-to-date knowledge on immune responses to biomaterials.Available online 28 June 2011 healing process. However, engineering such biomaterials requires an in-depth understanding of the hostAll rights reserved.of the implications for the design

eber c,d, Jan C. Simon a,d

, Germanytav Carus, 01307 Dresden, GermanyGermany, Venereology and Allergology, Philipp-Rosenthal-Str. 23, 04103 Leipzig, Germany

d function of biomaterials is that they do not elicit a detrimental immuneave profound impacts on the host immune response the concept emerged

le at ScienceDirect

erials

ier .com/locate/biomateria ls

rials2. Host responses toward biomaterials

Biomaterial implantation is always accompanied by injurythrough the surgical procedure. Injury to the tissue or organ initi-ates an inammatory response to the biomaterial starting with theformation of a provisional matrix. Implantation of engineeredcellematerial hybrids elicits an adaptive immune reaction towardthe cellular component that inuences the host response to thematerial component. When degradable devices are applied theimmune response is additionally affected by degradation productsand surface changes of the biomaterial that occur due to thedegradation process.

2.1. Onset of the inammatory response: blood proteins andalarmins induce granulocyte/monocyte activation

Nanoseconds after the rst contact with tissue, proteins fromblood and interstitial uids adsorb to the biomaterial surface. Thislayer of proteins determines the activation of coagulation cascade,complement system, platelets and immune cells and guides theirinterplay which results in the formation of a transient provisionalmatrix and the onset of the inammatory response [3,4] (Fig. 1A).The following section only delineates the complex interactions ofcoagulation, complement and platelets that occur on biomaterialsurfaces. For a more detailed depiction we refer to recent reviewson biomaterial-associated protein adsorption, coagulation andcomplement activation [3,4].

2.1.1. Coagulation, complement and adhesion proteins e signals forintegrins

Factor XII (FXII) and tissue factor (TF) are the initiators of theintrinsic and extrinsic system of the coagulation cascade, respec-tively. The intrinsic system is induced by contact activation of FXIIon negatively charged substrates followed by a downstreamcascade of protein reactions resulting in the release of thrombin[4,5]. Indeed, auto-activation of FXII has been shown to be cata-lyzed by surface contact with biomaterials [6]. It was suggested thatcontact activation of FXII is specically promoted by biomaterialswith anionic surfaces by imposing specic orientation on thesurface adsorbed FXII facilitating its activation [7]. However, newndings demonstrate differences in the displacement of competingproteins on hydrophilic and hydrophobic surfaces to be the causefor enhanced FXII activation at negatively charged biomaterials[8,9]. Although auto-activated FXII on biomaterials initiates thegeneration of thrombin, the amount produced is not sufcient toinduce clot formation [10]. Blood coagulation on biomaterials hasbeen recently demonstrated to require the combination of bothcontact activation and platelet adhesion and activation [10,11].Thrombin is one of the main activators of platelets. Low concen-trations of thrombin released by FXII on biomaterials are suggestedto activate platelets which in turn release mediators of the coagu-lation process and expose negatively charged phospholipids, thusproviding the supposed catalytic surface for the coagulationcascade [10,12]. Platelet-derived prothrombinases (factor Va andfactor Xa) of the extrinsic coagulation pathway assemble on theactivated platelet surface and become activated. Subsequentthrombin formation occurs [12] that further activates platelets andcoagulation factors of the intrinsic and extrinsic system amplifyingthe coagulation cascade on biomaterial surfaces [13]. Furthermore,thrombin initiates the cleavage of brinogen to brin forming theprimary brous mesh around the biomaterial.

Fibrinogen is also known to spontaneously adsorb to biomate-rial surfaces [14,15]. It was shown that an adhesion relatedconformational change of brinogen resulted in the exposure of

S. Franz et al. / Biomatetwo integrin binding domains capable to activate phagocytes[14,15]. It is suggested that phagocytes sense brinogen adherent tobiomaterials as brin thus launching an inammatory responseoccurring physiologically following clot formation [14]. Adsorbedbrinogen also serves as adhesion substrate for platelets [16,17].Besides thrombin and the adsorbed brinogen, TF expressed ondamaged cells or on activated leukocytes at the implantation site aswell as activated complement can induce activation of biomaterialattached platelets [18].

It is well established that complement is activated upon contactwith biomaterial [19e21]. In inammation complement activationoccurs via three distinct pathways, the alternative, the classical andthe lectin pathway which all converge on the level of C3 convertaseactivation that mediates formation and release of the anaphylatox-ins C3a andC5a (reviewed in Ref. [22]). Activation of complement onbiomaterial surfaces has been shown to be predominantly triggeredby the classical and alternative pathway although there are alsoreports of lectin mediated complement activation [19e21].However, complement activation is always associated with thebiomaterial adsorbed protein layer. Attached IgG has been demon-strated to bind C1q resulting in the assembly of C1, the rst enzymeof the classical pathway that promotes the initiation of the classicalC3 convertase [23]. Additionally, C3 can spontaneously adsorb to thebiomaterial surface while adopting a new conformation mimickingthatof surface boundC3bof the alternative pathway [24]. TheboundC3 thenpromotes the assembly of the initiating C3 convertase of thealternative pathway. C3b arising from the proteolytic cleavage of C3bycoagulation factors including FXIIa and thrombin canalso directlyattach to biomaterial surfaces and form an initiating C3 convertase[4,25]. Action of the C3 convertase results in the generation of C3bthat binds to the biomaterial protein layer to form more C3 con-vertases thus launching the amplication loop of the alternativecomplement pathway [19,24]. Once the complement cascade isstarted high amounts of C3a and C5a are generated at the implan-tation site [26]. Both anaphylatoxins can contribute to the onset ofinammatory responses at implantation sites through their multi-tude of effector functions including triggering of mast cell degran-ulation, increasing vascular permeability, attracting and activatingof granulocytes and monocytes and inducing granulocyte ROS(reactive oxygen species) release [22]. The induced complementproteins have also been shown to support platelet adhesion andactivation on biomaterial surfaces [27] which in turn can propagatethe coagulation cascade [10,12] but also promote TF expression bymonocytes and granulocytes on biomaterial surfaces [28,29]. Itbecomes apparent that the coagulation cascade and complementsystem closely interact on the biomaterial surface and modulateeach others activity [18]. The cross-talk between both systemsoperates synergistically in inammatory cell activation.

Adhesion proteins of the ECM including bronectin and vitro-nectin have also been described to attach to biomaterial surfaces[30]. Whereas brinogen and complement mainly contribute to theactivation of inammatory cells, bronectin and vitronectin arecritical in regulating the inammatory response to biomaterials.Both proteins have been described to promote the fusion ofmacrophages to foreign body giant cells on biomaterial surfaces[31,32]. However, these proteins also support adhesion andspreading of osteoblastic cells to biomaterials which representsa crucial step for integration of orthopedic implants [33]. Theseseemingly opposed functions of bronectin and vitronectinexemplarily reect the high effector capacity of the adsorbedprotein layer. Depending on the environment at the implant site,proteins adhering to biomaterials may either initiate and fosterinammation or assist healing.

Cell adhesion and activation on biomaterials primarily occurs viainteraction of adhesion receptors with the adsorbed proteins. In this

32 (2011) 6692e6709 6693review we focus on the interaction of biomaterials with

Fig. 1. Immune response toward biomaterials. A) Adsorption of blood proteins and activation of the coagulation cascade, complement and platelets result in the priming andactivation of PMNs, monocytes and resident macrophages. B) Danger signals (alarmins) released from damaged tissue additionally prime the immune cells for enhanced functionvia PRR engagement. C) The acute inammatory response is dominated by the action of PMNs. PMNs secrete proteolytic enzymes and ROS, corroding the biomaterial surface. IL-8

S. Franz et al. / Biomaterials 32 (2011) 6692e67096694

immunocompetent cells. Integrins represent the major adhesionreceptors of leukocytes [34]. Protein ligands of integrins includebrinogen, factor X, iC3b, bronectin, vitronectin [35]e all of whichhave been shown to attach to biomaterial surfaces [36]. Indeed, celladhesion to protein-coated biomaterials and subsequent cell acti-

cells as well as injured cells (Fig. 1A,B). Mast cell degranulation andassociated histamine release have been shown to play a role indirecting PMNs and monocytes to implanted biomaterials in miceand humans [50,51]. Reaching the implantation site, PMNsencounter the protein-coated biomaterial surface and subsequent

hronhronndand

S. Franz et al. / Biomaterials 32 (2011) 6692e6709 6695vation have been described to be mediated by integrins [37e40].Initial adhesion and spreading of phagocytes are primarily achievedthrough b2 integrins [14,37] which in turn leads to a change in thereceptor prole including the up-regulation and enabling of furtherintegrins [34]. Adherent monocytes differentiating to macrophagesinitiate b1 integrins [41] which in conjunction with b2 integrinsmediate adhesion during biomaterial-associatedmacrophage fusion[32,40,42]. Clustering of integrins may also induce motility, phago-cytosis, degranulation, as well as the release of ROS and cytokines,events which play an important role in the inammatory responsetoward biomaterials.

2.1.2. Danger signals and pathogen recognition receptorBesides recognition of biomaterials through adhesion receptors,

there are other receptoreligand interactions activating immuno-competent cells. Danger signals (also referred to as alarmins) havelong been ignored as potential activators of leukocytes followingbiomaterial implantation (Fig. 1B). The role of alarmins in inam-mation has been reviewed in detail elsewhere [43]. Alarmins arethe endogenous equivalent of pathogen-associated molecularpatterns (PAMPs) and include heat shock proteins, HMGB1 (highmobility group box 1), ATP and uric acid. They are rapidly releasedfollowing injury by cells dying in a non-programmedway (necrosis)to signal associated tissue damage. Like PAMPs, alarmins arerecognized by cells of the innate immune system such as macro-phages and dendritic cells (DCs) via pattern recognition receptors(PRR) such as scavenger receptors, toll-like receptors (TLR) and C-type lectins which promote inammation and immunity.

There is evidence that induced danger signals are capable ofimmune cell activation at biomaterial surfaces [44,45]. Uponbiomaterial application alarmins may be released or induced bycells at the implant site that had been damaged due to the surgicalprocedure. Proteolytic enzymes leaking from the injured cells mayadditionally trigger the generation of extracellular danger signals[46]. These signals include brinogen or cleaved ECM componentssuch as collagen peptides, hyaluronic acid (HA), bronectin andlaminin that adsorb to biomaterials. Thus, a coated biomaterialsurface itself might act as a danger signal (Fig. 1A). Soluble orbiomaterial-associated alarmins can interact with PPRs, preferen-tially TLRs on leukocytes and propagate the inammatory response.Indeed, recent studies demonstrate a role of TLR4 in the hostresponse to biomaterials in vitro and in vivo [47,48].

2.1.3. PMN activationImmediately following injury and protein deposition, inam-

matory cells e predominantely polymorphonuclear leukocytes(PMNs, granulocytes)emigrate from the blood toward the implantsite. Circulating PMNs are rapidly recruited to infection sites byhost- and pathogen-derived mediators, acting as a rst line ofdefense against invading pathogens. The role of granulocytes in theinnate immune response is extensively reviewed elsewhere [49].PMN recruitment to implantation sites is triggered by host derivedchemoattractants released from activated platelets and endothelial

released from PMNs enhances PMN inux and priming. In the transition from acute to cand activation of monocytes and macrophages. D) Macrophages are the driving force of c1 results in permanent activation of macrophages. Fusion-inducing stimuli like IL-4 aenvironment on the biomaterial surface. Furthermore, FBGC promote ECM remodeling

Macrophage-derived cytokines and PRR engagement activate DCs on the biomaterial surfatolerogenic subtypes, amplifying or suppressing the inammatory response.engagement of integrins and PRRs on the PMN surface triggersa phagocytic response and degranulation [52,53] (Fig. 1C). PMNsusually secrete proteolytic enzymes and ROS to promote and fosterpathogen killing [49]. At the implant site, the destructive agentsmay corrode material surfaces as described for polyurethane [54].Furthermore, the cytotoxic components damage surroundingtissue, prolonging the inammatory response. Another adverseeffect of biomaterial-induced PMN activation is the metabolicexhaustion and depletion of the granulocytes oxidative resources.Due to the continuous release of ROS the microbial killing capacityof PMNs is dramatically reduced, which has been related to severebiomaterial-centered infections [55].

PMNs also represent a signicant source of immunoregulatorysignals which they synthesize upon activation [56], interleukin-8(IL-8) being among the most prominent chemokines. The primarytargets of IL-8 are PMNs themselves. Various studies have reportedgranulocyte migration and prolonged presence of granulocyteswithin chitosan materials [57,58] due to persistent autocrine PMNattraction by IL-8 [59]. Activated PMNs also secreteMCP-1 andMIP-1b [49]. Both chemokines are known as potent chemoattractantsand activation factors for monocytes, macrophages, immature DCsand lymphocytes [60]. Increased release of these chemokines byPMNs suppresses further PMN inltration in favor of mononuclearcell inux [61]. Due to a lack of further activation signals PMNundergo apoptosis after having done their job as phagocytes andare engulfed by macrophages [61]. Within the rst two days afterbiomaterial implantation PMN typically disappear from implanta-tion sites [62].

2.2. Chronic inammation: dual role of macrophages asinammatory mediators and wound healing regulators in theforeign body reaction

Chronic inammation develops as inammatory stimuli persistat the implant sitewithmacrophages representing the driving forcein perpetuating immune responses [63]. Monocytes arriving at theimplantation site undergo a phenotypic change differentiating tomacrophages. Their activation leads to further dissemination ofchemoattractants. Macrophages attached to the biomaterial canfoster invasion of additional inammatory cells by secreting che-mokines like IL-8, MCP-1, MIP-1b [64] (Fig. 1D).

Macrophages also play a critical role in wound healing andtissue regeneration. Phagocytosis of wound debris, release ofenzymes important for tissue reorganization and of cytokines andgrowth factors inducing migration and proliferation of broblastsare mediated by macrophages and constitute the initial stepstoward effective tissue regeneration [65]. These different functionsare typically promoted by different macrophage subsets, originallyreferred to as M1 (classically activated) and M2 (alternativelyactivated) macrophages [66]. Based on fundamental macrophagefunctions involved in maintaining homeostasis these subsets havebeen reclassied into classically activated, regulatory, and wound-healing macrophages [67] (Fig. 2). The latter two arise from

ic inammation, PMNs stop secreting IL-8 in favor of cytokines promoting immigrationic inammation. Constant release of inammatory mediators like TNFa, IL-6, and MCP-IL-13 promote the fusion of macrophages to FBGC, which form a highly degradativebroblast activation resulting in excessive brosis and biomaterial encapsulation. E)

ce. Depending on the nature of the stimulus, DCs mature to either immunogenic or

a subdivision of the alternatively activated macrophage subset.Activation and function of these macrophage phenotypes isexcellently reviewed by Mosser and Edwards [67]. The differentmacrophage populations are generated in response to eitherendogenous stimuli released by damaged cells or innate immunecells following injury or infection or to adaptive immune signalsproduced by antigen-specic immune cells [68,69]. Classicallyactivated macrophages are typically triggered by interferon-g(IFNg) released by T helper 1 (TH1) cells during adaptive immu-nity or by natural killer (NK) cells during innate immunity and byTNFa produced by antigen presenting cells [67,68]. Classicalstimulation prime macrophages to secrete inammatory cytokinesand to perform microbicidal activity mediated by increasedsynthesis of ROS and nitrogen radicals making them to a crucialpart of host defense [67,68]. Wound-healing macrophages aregenerated in response to IL-4 produced by basophils, mast cellsand granulocytes in early innate immune responses or by TH2cells during adaptive immune responses [67,70]. Interleukin-4programs macrophages to down-regulate pro-inammatorymediators and to promote wound healing processes by contrib-uting to the production of ECM and by activation of broblasts[67,68,70e72]. Although wound-healing macrophages exert anti-inammatory activities they are not capable of down-regulatingimmune responses. Regulatory macrophages also arise duringinnate and adaptive immune responses. They are triggeredin response to a variety of signals including apoptotic cells, pros-taglandins, IL-10, immune complexes, glucocorticoids [67,73].

emigrate from the site of inammation to give rise for regulatoryand wound-healing macrophages [75] or whether the macro-phages alter their functional phenotype in response to progressivechanges of signals during the course of inammation [76]. There arereports showing that macrophages retain their functional adap-tivity and adjust their phenotype to changing environmentalstimuli [76,77]. The remarkable plasticity of macrophages is makingthem to an interesting target in the context of immunomodulation.

2.2.1. Foreign body giant cell formationMacrophages that attach and recognize a foreign material show

typically a classically activated phenotype secreting inammatorycytokines, ROS, and degradative enzymes and displaying highphagocytic capacity. Single macrophages are able to phagocytoseparticles up to a size of 5 mm [78]. If the particle size is larger,macrophages attempt to coalesce to FBGCs. The cytokines IL-4 andIL-13 have been identied to induce macrophage fusion onbiomaterial surfaces in vivo and in vitro [79e81]. ActivatedT lymphocytes (CD4 cells) at the implant site are assumed to bethe source of both IL-4 and IL-13 and have been shown to enhancemacrophage fusion on biomaterials [82]. However, a recent studyinvestigating synthetic biomaterials in nude mice revealed CD4cells not to be essential for induction of a foreign body response(FBR) [83]. Although IL-4 was not present in the T cell-decientsetting, macrophage fusion was not impaired due to unaffectedlevels of IL-13 at the implant site. It was suggested that mast cellsmost likely serve as a source of IL-13 at the onset of inammation

S. Franz et al. / Biomaterials 32 (2011) 6692e67096696However, to become fully activated the macrophages needa second signal such as a PRR-ligand [67,73,74]. The main task ofregulatory macrophages is to limit inammation and to dampenimmune responses which they achieve by release of high levels ofIL-10, a very potent immunosuppressive cytokine [67,73,74].

An immune response involves the action of all types of macro-phages, classical activated macrophages in the early phase andregulatory and wound-healing macrophages in the resolutionstage. It is still debated whether inammatory macrophagesFig. 2. Classication of macrophages. Macrophages represent a heterogeneous population owound healing, and immune regulation. This plasticity enables macrophages to adapt theirand also sustain IL-13 production during the chronic inammatoryresponse to the biomaterial. Chemoattractant CCL2 was alsoreported to be involved in FBGC formation [84] though not byrecruiting cells to the implant site but rather by guiding macro-phage chemotaxis toward each other [85].

Moreover, the properties of the biomaterial surface are impor-tant for FBGC formation. Since biomaterials are immediatelycovered, it is the adsorbed protein layer that renders the surfacefusogenic. A variety of proteins including collagen, bronectin,f cells with different phenotypic proles performing distinct functions in host defense,functions to environmental cues. Adapted from Refs. [67,68].

rialslaminin, brinogen, and vitronectin have been tested on theircapability to promote FBGC formation [32]. Although all proteinsmediate initial monocyte adhesion, only vitronectin supportsmacrophage adhesion and fusion [32]. It has been shown that b1and b2 integrins play a role in macrophage adhesion duringbiomaterial-associated fusion [42]. Nevertheless, both integrinsseem not to mediate the crucial adhesion step required formacrophage fusion [86]. It is proposed that fusion-inducing stimuliand initial adhesion to the biomaterial induces the expression ofmultiple fusogenic molecules including mannose receptor (CD206)[87], CD44 [88], CD47 [89], DC-STAMP [90], and E-cadherin [91]rendering macrophages capable for fusion [85].

2.2.2. Release of ROS, degradative enzymes and MMPsIf the FBGCs do not succeed in phagocytosing the foreign

material, they remain at the biomaterialetissue interface and shapepodosomal structures forming a closed compartment betweentheir surface and the underlying substrate [92]. Various studieshave shown that following fusion to FBGC, macrophages displaya reduced phagocytic activity in coincidence with enhanceddegradative capacity [93], a phagocyte-specic phenomenonreferred to as frustrated phagocytosis. In an attempt to resorb thenon-phagocytosable biomaterial, FBGCs secrete protons, enzymes,and ROS into the compartment between them [94,95]. This willlead to resorption of materials that are susceptible to degradation[96]. If the biomaterial is completely resorbed, associated inam-mation may resolve as the causative agent is no longer present.

Matrix metalloproteinases (MMPs) are macrophage-derivedproteolytic enzymes involved in the foreign body reaction tobiomaterials [97e100]. The collagenases MMP-8 and MMP-13 andthe gelatinases MMP-2 and MMP-9 could be detected in macro-phages adhering to collagen disks post explantation [97]. Combinedaction of both gelatinases and collagenases has been suggested topromote degradation of collagen implants with MMP-9 as keyenzyme. MMP-9 was also found to be induced in macrophages andbroblasts during tissue remodeling in response to naturalhydroxyapatite implanted in rats [100]. In this study MMP-9 alonewas unable to breakdown the xenograft implant, but was involvedin angiogenesis and ECM remodeling of the peri-implant connec-tive tissue. Macrophages and FBGC cultured in vitro on variousbiomaterials have been shown to express MMP-9 [98]. The studyfurther demonstrated reduced macrophage fusion by pharmaco-logical inhibition of MMP-1, -8, -13, and -18 implicating a role ofthese MMPs in the fusion process. Although the MMP blockingassays of this study did not show an involvement of MMP-9 in theFBGC formation [98] comprehensive investigations analyzing theforeign body reaction in MMP-9 null mice clearly demonstrated aninvolvement of MMP-9 in macrophage fusion [99]. The study alsorevealed that MMP-9 may play a pivotal role in biomaterialencapsulation and angiogenesis [99].

The action of MMPs at the implant site is assumed to dramati-cally change the cellular environment around the biomaterial andto modify migration, differentiation and active state of macro-phages and other immune cells [101]. Increased levels of MMP-9have been suggested to be indicative of inammation and associ-ated with poor wound healing [102]. An environment rich in MMP-9 at the biomaterial site may thus perpetuate inammation and actcounter-regulatory on the wound healing process.

2.2.3. Fusion-induced macrophage phenotype switchFusion to FBGCs is typically associated with a phenotype

switch of the macrophages from a classical to a more alter-native activation state. The fusion-inducing cytokines IL-4 and IL-13 are known to promote wound-healing macrophages during

S. Franz et al. / Biomateinammation [67,71,72]. This transition is reected by alterationsof the cytokine prole at the biomaterial site [64,103e105]. Earlyupon biomaterial recognition macrophages release IL-6, IL-1b,TNFa, IL-8, MCP-1, RANTES, ENA-78 similar to classically acti-vated macrophages [64]. Over time the majority of theseinammatory cytokines is down-regulated in favor of IL-10, TGF-b, MDC, Eotaxin-2 as well as IL1 receptor antagonist (IL1ra) thatresembles the anti-inammatory cytokine/chemokine prole ofalternatively activated macrophages. However, the activationstate of fusing macrophages is not completely identical to thealternative phenotype since they still produce pro-inammatoryRANTES and MCP-1 [64], ROS and degradative enzymes. More-over, biomaterial-adherent macrophages sustain or even increasetheir IL-6 and TNFa production on materials that do not promotemacrophage fusion without mandatorily undergoing a pheno-type switch [64,105]. These ndings support the dogma thatimplant surface properties dictate macrophage responses. Moreimportantly, they show that macrophage behavior is predomi-nantly governed by the fusion event rather than adhesion.

2.2.4. Impaired wound healing and excessive brosisLittle knowledge exists on the involvement of macrophages/

FBGCs in the healing response to biomaterials. Successful tissuerepair requires resolution of the inammation through the releaseof anti-inammatory mediators, cleavage of chemokines, down-regulation of inammatory mediators and receptors, andapoptosis of immune cells [106]. At implantation sites, FBGCsproduce anti-inammatory cytokines (IL-10, IL-1ra) [64,103e105].However, their immunosuppressive activity may be counter-regulated by the proteolytic and pro-oxidant microenvironmentdue to continuous release of ROS and degradative enzymes aroundthe biomaterial [106,107]. FBGCs that have formed in response toIL-4 may release pro-brotic factors such as transforming growthfactor beta (TGF-b) and platelet-derived growth factor (PDGF) thattrigger the action of broblasts and endothelial cells as shown forIL-4 induced alternatively activated macrophages in vitro [71,72].Activated broblasts start to synthesize and to deposit collagen thatoften results in material encapsulation [108,109]. However, therelease of pro-brotic factors by FBGCs has never been clearlydescribed. Although the mechanism is not well understood, it isassumed that continuous action of FBGCs result in prolongedbroblast activation and excessive biomaterial-associated matrixdeposition [110,111].

2.3. Dendritic cell responses to biomaterials

Synthetic biomaterials including ceramics, polymers or metallicmaterials are typically not immunogenic and are thought not toinitiate an adaptive immune response, except for cases of metalhypersensitivity. However, lymphocytes have been found at sites ofsynthetic implants [44,112] suggesting their involvement inimmune responses tobiomaterials. DCs are keyplayers of innate andadaptive immunity. They elicit adaptive immune responses by theirability of antigen presentation and T cell priming. Additionally, DCspossess immunoregulatory capacities as they play a role in theinduction of antigen-specic T cell tolerance, T cell anergy and theactivation and expansion of regulatory T cells (TReg) (reviewed inRefs. [113,114]). Whether DCs induce an immunogenic or a tolero-genic T cell response depend on many factors including the state ofDCmaturation and the cytokineenvironment [114]. According to thecurrent concept of DC functions immature and semi-mature DCs arepromoters of tolerance whereas fully mature DCs induce immunity[113]. Antigen presentation in the absence of co-stimulation byimmature DCs typically promotes T cell anergy [115,116]. Semi-mature DCs expressing MHC and co-stimulatory molecules but

32 (2011) 6692e6709 6697unable to produce pro-inammatory cytokines such as IL-12, TNFa,

rialsIL-6 and IL-1b have been described to convert nave T cells intoCD4/CD25 TReg or IL-10 secreting CD4 T cells (Tr1) [117e119].The tolerogenic activity of semi-mature DCs can be additionallyenhanced by their release of IL-10 [113,114]. Cytokines have beenshown to play the major part in the induction of tolerogenicity.Besides IL-10 and TGF-b, two important immunosuppressive regu-lators, various cytokines and growth factors including IL-6 and TNFaat low concentrations as well as IL-16, granulocyte colony stimu-lating factor and hepatocyte growth factor have been reported togenerate DCs with tolerogenic activity [114,120e125]. With respectto biomaterial application induction of tolerogenic DCs at theimplant site would provide a powerful means to limit the immuneresponse and to promote wound healing and biomaterialintegration.

The role of DCs in biomaterial application has mostly beenaddressed in the presence of an immunogenic biological compo-nent. Biomaterials were found to exert adjuvant effects since theypotentiated immune responses toward co-delivered antigens [126].However, acting as an adjuvant requires the capability to prime DCsthat has been shown to necessitate direct cell-material contact[127]. It is assumed that biomaterials activate DCs by triggeringreceptors and signaling cascades of the pathogen recognitionsystem [44] (Fig. 1B and E). As described above, DCs may sensematerials using PRRs including toll-like receptors and C-type lec-tins. The receptor-activating ligands are danger signals constitutedby proteins and carbohydrate moieties in the adsorbed proteinlayer. A recent study suggests a substrate-dependent DC activation.Albumin or whole serum stimulated DCs to the release of IL-10 or toprime IL-4 producing T cells as typical for Th2 type responses [128].In contrast, DCs cultured on collagen and vitronectin generatedhigh levels of IL-12p40 that correlated with the release of IFNg by Tcells indicating a Th1 type response [128]. Because collagen andvitronectin function both as danger signals and adhesion substratesfor integrins, integrin signaling cannot be ruled out as an alterna-tive mechanism of DC activation to PRR engagement. Indeed, DCintegrin binding to ECM proteins and subsequent DC maturationhas been shown, and for bronectin a dependency of DC matura-tion on b1 integrin binding was demonstrated [129,130].

Various biomaterial polymers (alginate, agarose, chitosan, HA,poly(lactic-co-glycolic acid)) have been shown to exert differentialeffects on DC maturation and activation [44,131,132]. DCs activatedupon biomaterial contact develop an immunogenic phenotypesimilar to LPS-activated DCs which is characterized by increasedexpression of co-stimulatory molecules (CD80/CD86), major histo-compatibility complex class II (MHC-II) molecules and the DCmaturation marker CD83 [44,127]. Biomaterial-matured DCs arecapable to promote T cell proliferation and secrete inammatorycytokines (TNFaand IL-6)knownto furtheramplifyDCmaturationbyautocrine stimulation [44,127]. Nevertheless, some of the syntheticmaterials tested (e.g. alginate or HA) restrained DCmaturation [131].These cells develop a tolerogenic phenotype [125] that, uponencountering and presenting antigens, induce T cell tolerance.

Depending on which PRR is engaged, DC maturation can bepromoted or inhibited leading to immunity or tolerance, respec-tively [133]. Whereas immunogenic DCs may prolong the immuneresponse to biomaterials and delay wound healing, tolerogenic DCsare capable to down-regulate the immune cells and resolveinammation. Thus, induction of tolerogenic DC by designing thesurface chemistry appears to be a promising strategy of modulatingimmune responses to biomaterials to improve biocompatibility.

2.4. T lymphocyte responses to biomaterials

T lymphocytes have been shown to adhere to synthetic

S. Franz et al. / Biomate6698biomaterials in vitro [62]. In co-culture with macrophages, theywere found attached predominantly to macrophages and not to thebiomaterial surface [82]. T lymphocytes have been demonstrated topromote macrophage adhesion and fusion via paracrine effects [82](Fig. 1D). However, close association of lymphocytes and macro-phages also suggests direct signaling which has been shown todominate at later time points of their interaction [134]. During theinitial response to biomaterials, lymphocytes and macrophagespredominantly release inammatory mediators [134e136]. Theseinclude the cytokines IL-1b, IL-6, TNFa and the chemokines IL-8,MCP, MIP-1b, ENA-78 all of which attract and activate inamma-tory effector cells such as neutrophils, monocytes, T lymphocytesand natural killer cells. Interestingly, release of IL-1b and TNFadeclined over time in favor of IL-10 and MMP-9, tissue inhibitor ofMMPs (TIMP)-1 and TIMP-2 e important mediators for ECMremodeling in wound healing [135]. These data nicely demonstratethe capability of T lymphocyteemacrophage interactions in guidingthe inammatory phases of the foreign body reaction. The absenceof T cell activation in in vitro studies, however, questions the effectof lymphocytes in these processes. Neither IL-2 nor IFNg weredetected as a response to lymphocytes alone or in co-culture withmacrophages to different synthetic biomaterials including siliconerubber, elasthane 80A or polyethylene terephthalate [135e137]. Onthe other hand, activated T cells were identied during inamma-tory biomaterial responses in vivo, suggesting the presence of thecomplete inammatory environment as requirement for T cellactivation in response to synthetic materials [137,138]. Neverthe-less, the question remains how T cell activation is mediated duringforeign body reaction. Synthetic biomaterials do not serve as anantigen. Given that the biomaterial is not degradable and that nobacteria transiently attach to its surface, T cell activation via antigenpresentation does not occur. It has been suggested, that syntheticbiomaterials may present functional groups on their surfaces actingas mitogens [137]. Mitogens are lectins that can trigger lympho-cytes by cross-linking of glycoproteins on the lymphocyte surface.To date, mitogenic capabilities of biomaterials have not beendemonstrated.

3. Extracellular matrix e a native modulator of cell activity inimmune responses and tissue repair

The direct interactions of biomaterials and components of thehosts immune system described so far occur in vivo in the presenceof the ECM. Several ECMproteins are potent regulators ofmonocyte/macrophage adhesion and activation and are involved in DCmigration and maturation [139,140]. Bidirectional interactions ofECM, cells andgrowth factors/cytokines determine cellular behaviorduring all phases of biomaterial integration and wound healingincluding inammation, cell proliferation and tissue remodeling.

Moreover the ECM provides mechanical support and a three-dimensional scaffold for cellular organization. In addition it regu-lates cell behavior, storage and mobilization of signaling moleculesas well as proteolytic degradation [141] (Fig. 3A). Collagen, elastinand brin form a brous structure that provides tensile strength orelasticity [142]. Non-brous proteins (predominantly bronectinand laminin) linked to this scaffold supply domains for cellematrixinteraction [143]. The protein scaffold is embedded in a gelatinous,negatively charged matrix composed of glycosaminoglycans(GAGs). GAGs are long, unbranched carbohydrate chains consistingof repeating disaccharide units [144]. The sugar chains can bemodied by sulfate groups as in chondroitin sulfate (CS), heparansulfate, dermatan sulfate and keratin sulfate. HA is the only non-sulfated GAG. It is also not attached to a protein core whereas thesulfated GAGs are linked to serine rich proteins to form proteo-glycans (PGs) [144]. The glycan matrix of the ECM serves as lubri-

32 (2011) 6692e6709cant and provides a reservoir for signaling molecules. Additionally,

rialsS. Franz et al. / Biomateit participates in a variety of biological processes includingcellematrix interactions and activation of enzymes and mediatorslike growth factors and cytokines [145].

Cells including those involved in inammation attach to the ECMvia various surface receptors including integrins, selectins, synde-cans and CD44 (Fig. 3B). They either recognize specic adhesivedomains in the protein scaffold or bind to components of the glycanmatrix e.g. HA, heparan sulfate and CS. This usually triggers intra-cellular signals that direct adhesion, migration, proliferation,differentiation, protein synthesis, and secretion [139,141,146]. Thesecellular functions are additionally regulated by signaling molecules(growth factors, cytokines and chemokines) that are trapped in theECM. However, the ECM not only serves as a reservoir for signalingmolecules, it rather regulates their distribution and mode of action.Components of the ECM (mostly PGs) retain the soluble mediatorsfor example via electrostatic interactions between the negativelycharged sulfate groups of the PGs and the positively charged surfaceof the signaling molecule [147]. This interaction has different

Fig. 3. Structure and function of the ECM. A) ECM composition. B) Cellematrix interaction: enintracellular signaling events. C) Cytokineematrix interaction: proteoglycans bind growth fenzymatic cleavage, enriches their local concentration and enhances or reduces their actio32 (2011) 6692e6709 6699biological consequences since it affects the local concentration,biological activity, and stabilization of growth factors [148,149](Fig. 3C). Secreted growth factors usually have a very short half-lifedue to their high susceptibility to proteolytic degradation. Linkageto the ECM protects them from enzymatic cleavage. Moreover,binding to the ECM prevents growth factor diffusion within thecompartment, providing a local store of functional molecules thatpersists long after their release has stopped [148]. Thismay result ina local concentration of growth factors needed for effective receptorsignaling. Additionally, growth factor activity may be enhanced bylocalization within the ECM, allowing interaction with its specicligands [149]. On the other hand, some growth factors may becomeinactive when bound to the ECM and can only act on their targetwhen released by matrix proteolysis [141]. This requires action ofECM-degrading enzymes expressed by cells regulated by growthfactors and ECM adhesion domains. Thus, growth factors and ECMproteins collaborate in creating a distinct cellular environment orniche that regulates tissue regeneration [150].

gagement of cell surface receptors with proteins and proteoglycans of the ECM triggersactors and cytokines. Interaction with the ECM protects the signaling molecules fromn on cells.

growth factors and cytokines and thus inuencing cellularbehavior [152]. For example, repressed T cell activity due to

rialsa down-regulation of the IL-2 receptor is mediated by MMP-9 thatcleaves the cytokine receptor from the T cell surface [152,153].Membranetype-MMP-1-mediated shedding of syndecan-1,a transmembrane protein involved in cellecell and cellematrixadhesion has been shown to stimulate cell migration [154].Vascular endothelial growth factor (VEGF) restrained in the ECMbecomes mobilized by MMP-induced proteolysis to exert itsangiogenic activity and to serve as chemotactic signal for osteo-clasts [155,156]. Several MMPs (MMPs-2, -9, -13 and -14) areinvolved in the release of the ECM-bound latent TGF-b complex andits processing into an active ligand that controls proliferation,differentiation and other functions [151,157,158]. MMP-mediatedproteolysis of bronectin generates fragments inducing cellmigration but blocking cell proliferation [151]. Angiogenic frag-ments are released from collagen by MMP-2 and -9 [151,159].Laminin-5 when cleaved by MMP-2, -3, -12, -13 and -14 exposesneo-epitopes promoting cell migration whereas MMP-8-derivedlaminin-5 fragments do not [151,160].

The action of MMPs is controlled at two levels, during tran-scription by cytokines and integrin clustering and after secretion byendogenous inhibitors including a2-macroglobulin and TIMPs[161e163]. Whereas cytokine and integrin signaling up- or down-regulate MMP expression, the protease inhibitors and here specif-ically TIMPs bind to the MMPs thereby determining their activity[163]. Precise control of MMP activity is crucial to ensure ECMremodeling under normal physiological conditions as dysregula-tion has been implicated in many diseases such as brosis, cancer,arthritis and vascular disease [162,164e166].

Taken together, the ECM not only provides support and a three-dimensional scaffold for cells, it also plays a highly functionalizedrole in directing cell behavior by spatially and temporarilyconcerted interactionwith other cells, ECM components, degradingenzymes and signaling molecules all of which being relevant totissue repair, host responses to biomaterial and ultimately bioma-terial integration [167].

4. Modulating immune responses to biomaterials

For a long time biomaterial engineering focused on inertbiomaterials. The concept of biologically inert materials is tominimize the host response by avoiding cellematerial interactions.Inert biomaterials are recognized as foreign by the host but remainessentially unchanged and tolerated due to their encapsulation inbrous tissue [168]. However, it has been realized that permittingspecic cell responses may in fact be benecial for biomaterialintegration and improve implant performance [169]. For exampleosseointegration of titanium implants, classically inert biomate-rials, could be improved by modication of the material surfaceallowing migration and adhesion of bone-forming cells from theProteolytic degradation of the ECM is also an essential feature oftissue repair and remodeling, and enables cell migration andinvasion [148]. Numerous cell-secreted and cell-activated enzymeswith ECM-degrading capacities have been identied includingMMPs, elastase, and plasmin. MMPs play a predominant role inECM degradation since they process and degrade virtually allstructural ECM proteins. There are over 25 known MMPs groupedinto collagenases, gelatinases, stromelysins, matrilysins, mem-branetype (MT)-MMPs and others performing multiple, sometimesoverlapping, functions in ECM proteolysis (reviewed in Ref. [151]).However, MMPs are not only involved in ECM breakdown they alsotarget non-ECM proteins, cell surface molecules and ECM-bound

S. Franz et al. / Biomate6700surrounding tissues onto the implant [170].Controlled tissue responses at the implant site are assumed toencourage wound healing. Increasing comprehension of the healingprocess point out that immune responses associated with inamma-tion and macrophage activation are crucial for tissue repair [171,172].With growing knowledge on the processes ofwoundhealing andhostresponses to biomaterials the eld of biomaterial engineering hasbegun to address the development of materials with immunomo-dulating capacities. Ideally, the materials should affect normalimmune cell function such that they promote healing and implantintegration while sustaining specic implant function [2]. There areseveral publications reporting of modulated immune responsesinduced by silicons and blends of polydioxanon (PDO) and elastin orcollagen [173e175]. The very comprehensive studies addressingmodulationof functions of innate and adaptive immune cells revealedsuppressed NK cell activity in response to all biomaterials whereasmacrophage functions remained unaffected [173e175]. Blends of PDOand elastin or collagenwere also shown to exert immunosuppressiveeffects on T and B cell-mediated immunity [174,175].

On the one hand the studies clearly point out the need for testingof biomaterials on their effects on both acquired and innate immuneresponses as a component of biocompatibility assessment [174,175].On the other hand the studies nicely demonstrate the potential ofbiomaterials to modulate immune cell function encouraging thedesign of biomaterials capable of eliciting appropriate immuneresponses at implantation sites. Current strategies in the design ofsuch biomaterials include alteration of material surface propertieseither passively via physicochemical features or actively with mole-cules or matrices designed to systematically target cell behavior.

4.1. Immunomodulation by surface modications of biomaterials

Passive modulation of biomaterial surface properties aims atlimiting macrophage adhesion, activation and fusion to FBGCs(Fig. 4A). Type, level and conformationof serumproteins that adsorbto biomaterial surfaces depend on the terminal chemistry of thebiomaterial [36,38,39,176]. The adsorbed protein layer usuallyprovides binding sites for protein-specic receptors (integrins,PRRs) on PMNs, monocytes and macrophages. Surface chemistry-dependent modulation of the protein layer may thus enabledifferent receptor binding and signaling in the immune cells leadingto altered cellular responses. Indeed, comprehensive proteomicsstudies revealed macrophages to change their protein expressionproles and cytokine/chemokine responses when cultured onsurface-modied polymers displaying hydrophobic, hydrophilic,and/or ionic chemistries [64,177]. Macrophages attaching tohydrophilic and anionic biomaterial surfaces providing low integrinbinding sites were shown to experience low integrin-mediated cellspreading, leading to macrophage apoptosis [178]. These ndingsprovide a clue for surface modication of biomaterials to elicitdesired PMN and macrophage activities.

Another strategy for guiding macrophage responses to bioma-terials is the variation of roughness and surface topography[179,180]. In their natural environment cells respond to ECMcomponents in the nanometer scale in terms of adhesion, prolif-eration, migration, and gene expression. Imprinting of patterns atmicron and nanometer scales on material surfaces may mimic thenatural topography of the ECM [179]. Topographic patterns areknown to affect function of broblasts [181], epithelial cells [182],and endothelial cells [183]. A differing response of macrophages tomicron-structured biomaterials has been demonstrated recently inthe context of the foreign body reaction [184]. Parallel gratingsimprinted on polymeric surfaces with line width ranging from250 nm to 2 mmwere shown to affect macrophage morphology andcytokine secretion in vitro, and macrophage adhesion in vivo

32 (2011) 6692e6709independent of the biomaterial surface chemistry [184].

rialsS. Franz et al. / Biomate4.2. Immunomodulation by incorporation of bioactive molecules

Current methods of biomaterial functionalization includespecic surface coatings and the incorporation of bioactive mole-cules such as adhesion sites, growth factors, anti-inammatorymediators or drugs either alone or combined [185].

4.2.1. Providing of integrin adhesion sitesFunctionalization of biomaterial surfaces with specic integrin

binding sites represents a powerful strategy in directing responses

Fig. 4. Current strategies in the design of immunomodulating biomaterials. A) Alterationtionalization of biomaterials by incorporation of bioactive molecules like integrin adhesbiomaterials with articial ECM e mimicking the biofunctions of the natural ECM as a tooladhesion receptors (e.g. integrins). Proteoglycans are capable to interact with endogenous32 (2011) 6692e6709 6701of inammatory cells (Fig. 4A). Attachment of short oligopeptidesequences that make up receptor binding domains within adhesiveproteins have been shown to promote cell-specic adhesion andfunction on biomaterials with arginineeglycineeaspartic acid(RGD) being the most prominent domain (reviewed in Refs.[186,187]). The RGD and PHSRN (proline-histidine-serine-arginine-asparagine) domains of bronectin were identied to be crucial inregulating macrophage function via a5b1 and anb3 integrinsignaling in vitro and in vivo [188,189]. Both domains, whenimprinted at a specic orientation on the biomaterial, were found

s of physicochemical features of biomaterials like chemistry or topography. B) Func-ion sites as well as growth factors and anti-inammatory mediators. C) Coating offor modulating immune cell behavior. Collagen provides natural binding sites for cellcytokines and growth factors allowing for their specic presentation to target cells.

rialsto initiate distinct intracellular outside-in signal pathwaysmediating macrophage adhesion and function upon ligation withintegrins [190]. As one consequence, RGD and PHSRN domainsmediated the formation of FBGC on biomaterial surfaces [189,190].

The incorporation of integrin ligands to biomaterials is oftencombined with polyethylene glycol (PEG) coatings (PEGylation),which renders biomaterial surfaces non-interactive (also referredto as non-fouling) [191,192]. Surfaces modied with PEG resistprotein adsorption and are thus protected against passive cellattachment and subsequent cell activation [193]. The advantage ofthese combined approaches relies on the prevention of nonspeciccellematerial interaction in favor of specic cell activation elicitedby recognition of integrin adhesion sites on the biomaterial surface[194,195]. Further non-fouling coatings include dextran-based gels,poly(ethylene oxide) (PEO), and alginate [196e198].

4.2.2. Coupling of anti-inammatory drugs to biomaterialsAnother method of rendering biomaterials immunomodulatory

is the incorporation of anti-inammatory factors (Fig. 4A). Gluco-corticoids are potent suppressors of immune responses (reviewedin Ref. [199]). They inhibit inammatory cell activation by abro-gating the synthesis of inammatory mediators including severalcytokines and chemokines, prostaglandins, leukotrienes, proteo-lytic enzymes, free oxygen radicals and nitric oxide (NO). Simul-taneously, they promote resolution of inammation and of adaptiveimmune response by enhancing anti-inammatory cytokinerelease and suppressing cellular [T helper (Th)1-directed] immu-nity in favor of humoral (Th2-directed) immunity and tolerance.Indeed, delivery at the implantation site of dexamethasone viacoupling to biomaterials results in reduced implant-associatedinammation as shown by decreased numbers of PMNs in theinitial inammatory phase and the absence of macrophages,lymphocytes, and brous capsule formation in the later phase[200,201]. However, an unwanted side effect of dexamethasonetreatment is the reduction of VEGF in the surrounding tissueimpeding angiogenesis and delaying wound healing [200,202]. Bycombined administration of dexamethasone and VEGF this anti-angiogenic effect could be overcome [202]. Though, dexametha-sone needs to be supplied by the biomaterial throughout itslifetime. As soon as delivery is subsided, the anti-inammatoryeffects fade and PMNs and macrophages start an inammatoryresponse [203].

Loading of biomaterial surfaces with NO-releasing coatings hasevolved as an attractive strategy for durable control of immuneresponses [204]. Continuous and slowly liberation of NO results inreduced inammatory cell recruitment and performance at theimplant surface that sustains even after exhaustion of the NOreservoir [205]. Down-regulation of inammatory cytokines suchas IL-6 and MCP-1 as well as induction of nitrosated proteins arediscussed to cause NO-mediated immunosuppression [206].Furthermore, NO may induce macrophages to produce NO them-selves explaining its long-lasting anti-inammatory activity [207].

4.2.3. Delivery of growth factorsA complex signaling network of growth factors, including

epidermal growth factor (EGF), broblast growth factor (FGF),granulocyte macrophage colony stimulating factor (GM-CSF), TGF-b, VEGF, and PDGF control adhesion, migration, proliferation, anddifferentiation of broblasts, keratinocytes, and endothelial cells inwound healing (reviewed in Ref. [208]). Although tissue cells arethe primary targets of the growth factors, biomaterials decoratedwith these bioactive molecules can still be considered as immu-nomodulatory (Fig. 4A). Wound healing in adult tissue is alwaysassociated with an inammatory response [209] and there is a tight

S. Franz et al. / Biomate6702cross-talk between immune cells and tissue cells regulating thehealing process [210e215]. Fibroblasts, for example, have beenshown to suppress MIP-1a release by activated macrophages [211]and to differently modulate IL-10 and IL-12 production by mono-cytes [212]. Monocytes develop broblast modulating activity asseen by enhancedMMP-2 expression of broblasts only in responseto monocyte-derived GM-CSF. Vice versa, the activated broblastsamplify GM-CSF production in monocytes suggesting synergisticinteractions during matrix remodeling [210]. Thus, modulatingbroblast, endothelial cell and keratinocyte function by loadingbiomaterials with growth factors also feeds back on the activity ofmonocytes and macrophages [202,216e219]. Moreover, growthfactors also act directly on the immune cells, as shown for TGF-b1and PDGF onmacrophage chemotaxis and activation during woundrepair [220].

4.3. Designing immunomodulating biomaterials based onarticial ECM e concepts and recent ndings

The majority of functionalized biomaterials provide a singlebioactive signal. Presentation of several bioactive factors is closer tothe natural environment of the cells and may help tuning thedesired responses. Interaction with the ECM regulates cellularbehavior including migration, differentiation and proliferationmaking the ECM an attractive tool for endowing biomaterials withphysiologic biofunctions (Fig. 4B).

4.3.1. HydrogelsConsiderable effort has been directed at mimicking the physi-

ologic ECM microenvironment in the design of tissue engineeringmatrices. Natural and synthetic hydrogels are attractive matricesdue to their high water content and three-dimensional structureresembling soft tissue [221]. Synthetic hydrogels are composed ofpolymers like PEG that are bio-inert. In different approaches thesynthetic scaffolds have been rendered ECM-mimetic by graftingthem with typical biofunctions including the presentation ofreceptor binding ligands for specic cell adhesion, the suscepti-bility to proteolytic degradation and remodeling by cell-derivedproteases and the capability of binding and presentation ofgrowth factors (reviewed in Ref. [222]).

Hydrogels that are made of materials from natural sourcesincluding ECM proteins (collagen, brin, gelatin) and poly-saccharides (GAGs, dextran, alginate, chitosan) are alreadyprovided with ECM-derived biofunctions [223]. Although theyusually have a low toxicity and rarely induce chronic inammatoryresponses, the potential of pathogen transmission and immuno-genicity of the materials restrain their use. A comparative studyevaluating the morphologic host tissue response to ve commer-cially available ECM-derived biologic scaffolds revealed that thescaffolds elicited distinct host responses depending on species oforigin, tissue of origin, processing methods, and/or method ofterminal sterilization [224]. However, natural ECM scaffolds arewidely used for tissue engineering in regenerative medicine due totheir simple design and economic fabrication in contrast tosynthetic hydrogels.

4.3.2. Articial ECM coatings for synthetic implantsThe concept of modulating cellular responses through ECM-

mimetic biomaterials has also been exploited to improve the bio-logical acceptance and integration of synthetic implants. The rstcoatings that were developed made either use of adhesive domainsof ECMproteins or ECM-derivedproteins as awhole. Various studiesreported of collagen coatings that supported cell attachment andactivity on implants in vitro and signicantly improved bonematuration and mineralization at implantetissue interfaces in vivo

32 (2011) 6692e6709[225,226]. Early appearance of mononuclear phagocytozing cells

and higher expression of bone-specic matrix proteins associatedwith increased early bone remodeling around the modied bioma-terials were observed [227] indicating a modulatory capacity of thecollagen matrices on both immune and tissue cells. Collagen coat-ings seem to promote specic cell implant interactions via integrinb1aspresented for rat calvarial osteoblasts [228]. Although, collagenwas shown to mediate adhesion and spreading of osteoblasts to theimplants, it had no effect on later processes such as proliferation,differentiation, and mineralization of the osteoblasts [229].

Two conclusions can be drawn from these ndings. First, cellularfunctions are not controllable only by promoting adhesion anddirect interaction with integrins and other cell surface receptors. Italso requires indirect modulation through inducing specic growthfactor responses. Second, to effectively modulate cell activitybiomaterial coatings should comprise the action of signalingmolecules. Biodegradable coatings locally releasing incorporatedgrowth factors like bone morphogenetic protein (BMP), insulin-likegrowth factor (IGF), and TGF-b have been successfully tested tostimulate fracture healing and to improve biomaterial performance[230]. However, to yield proper biological effects, the growthfactors have to be supplied in non physiologically high amounts atenormous costs. A more sophisticated approach therefore is toutilize the growth factor regulating property of the ECM. ArticialECM (aECM) coatings were developed by modifying collagenmatrices with GAGs and PGs [231] in order to create a biomaterialenvironment mimicking the situation in vivo (Fig. 5A).

Proteoglycans are suggested to be important mediators of osteo-

additional sulfate groups to GAGs should improve the biologicalproperties of aECM coatings. For HA it was demonstrated thatmodication with sulfate groups increased the binding afnity forhuman BMP-4 [234]. Implant coatings containing sulfated GAGsmaythus allow for enhanced osteoinductive activity by specically inter-acting with bone cell stimulating factors.

In vitro experiments revealed aECM containing sulfated GAGs likeCS promote cell adhesion as well as cell proliferation [235,236].Several in vivo studies reported that addition of CS to collagen coat-ings improved the osteoconductive properties of both titaniumimplants [231,237] and hydroxyapatite bone cements [238,239].When compared to uncoated implants and collagen matrices, boneremodeling and new bone formation around the implants wereincreasedwhen CSwas incorporated into the coatings (Fig. 5BeD). Itwas suggested that CS mediates attachment of growth factors to theECM or cell surface thus stimulating both bone resorbing and bone-forming cells [231,237e239]. However, whether a specic growthfactor response was modulated by the aECM coatings remainsunclear. An interesting hint in this respect was provided by a studyevaluatingosseointegrationof implants coatedwith aECMcomposedof collagen, GAG and BMP-4 [240]. Interestingly, collageneGAGcoatings aloneproved tobe as effective as thosepreloadedwithBMP-4. The similar effects of both coatings might be due to the fact thatonly a very low amount of growth factor was used. Another expla-nation is the interaction of the collageneGAG matrix with endoge-nous growth factors. The aECM coating might have stored growthfactors at the implant site and presented them to cell receptors

ngs:n arplanfacecoatew bo

S. Franz et al. / Biomaterials 32 (2011) 6692e6709 6703blast attachment and adhesion. Osteoblasts cultured on titaniumwere shown to synthesize sulfated GAGs that permeate thecellebiomaterial interface and promote adhesion [232]. Additionally,PGs should act as mediator betweenmatrix and endogenous growthfactors, thus improving cell activity around implants and inuencingbiomaterial integration [233]. Recent investigations focus on modi-cations of GAGs to inuence their interaction with specic growthfactors.Within the natural ECM the sulfate groups in GAGs representone important growth factor binding site. Thus, incorporation of

Fig. 5. Use of aECM coatings for synthetic implants. A) General assembly of aECM coatiare incorporated. B) Signicantly greater bone contact and increased new bone formatioas compared with uncoated titanium rods (Ti) in a rat tibia model [231] 7 days after imnumber of TRAP-positive osteoclasts (red) around the newly formed bone (B) at the sura sign of increased bone remodeling 7 days after implantation [231]. Around the un(enzymehistochemical stain against TRAP, original magnication 100). D) Increased ne

(Ti Coll CS) as compared to uncoated screws (Ti) in the medullary canal of shemagnication 16).(For interpretation of the references to colour in this gure legend, thebeneting bone formation. Besides GAGs, other ECM components,eitherwholeproteins (osteocalcin [241]), peptides (phosphoserineasanactive part of osteopontin [242]), or functionalities (sodiumcitrate[242]) had positive effects on bone remodeling around hydrox-yapatiteecollagen composite bone cements.

4.3.3. Immunomodulating effects of aECM coatingsECM proteins are potent regulators of immune cell activity

[139,140]. Thus, aECM coatings may have the capability to control

collagen brils serve as protein substrate in which distinct glycosaminoglycans (GAGs)ound titanium implants coated with type I collagen and chondroitin sulfate (Ti Coll CS)tation (Goldner Trichrome stain, original magnication 25). C) Signicantly increasedof titanium implants coated with type I collagen and chondroitin sulfate (Ti Coll CS) asd pins there is less bone contact and a brous interface (F) at the same time pointne formation around Schanz screws coated with type I collagen and chondroitin sulfate

ep tibiae [237] 6 weeks after implantation (Goldner Trichrome stain, originalreader is referred to the web version of this article).

the inammatory response which in turn has an important inu-ence on bone regeneration and wound healing (Fig. 4B). Naturaloccurring GAGs have been identied to bind and modify inam-matory factors such as interleukins and chemokines [243e245]. Animportant immunoregulatory cytokine is IL-10, exhibiting bothsuppressive and stimulatory effects on the immune system. IL-10acts anti-inammatory on macrophages and DCs by reducing theproduction of pro-inammatory cytokines and presenting antigensto T cells. Sulfated GAGs have been reported to bind human IL-10and to modulate its activity [244]. Interestingly, GAG-promotedlinkage of IL-10 to the surface of monocytes/macrophages resul-ted in enhanced activity of IL-10 on the immune cells [244]. Thebinding capability of the sulfated GAG was determined by grade ofsulfation, position of the sulfate groups (C4 or C6) and linkage of thesulfate group to the sugar chain (N- or O-sulfation) indicating thatsubtle differences in the GAG structure and/or sequence might besensed by signaling molecules and thus guides their interactionwith the ECM. Differential binding to subsets of various sulfatedGAGs was also observed for chemokines such as IL-8 [245]. Giventhat GAG structures vary among tissues and that chemokineinteraction with GAG mediates their presentation to leukocytes, itis conceivable that GAGs participate in determining the specicityof leukocyte recruitment in vivo. This might be an important aspectfor the design of aECM coatings to modulate the invasion ofimmune cells to the implantation site.

However, GAGs not only modulate the activity of inammatorymediators they are also capable to directly interact with immunecells. In this respect, CS and HA are of special interest because theyare known to act immunosuppressive on various cells. CS, forexample, is used as a therapeutic agent in osteoarthritis since itexerts strong anti-inammatory effects in the joint [246]. It was

found that the immunosuppressiveactivityof CS isbasedon reducednuclear translocation of NF-KB, a key transcription factor of variouspro-inammatorymediators, due to an inhibition of the ERK1/2 andp38MAPK pathway by CS [247]. Subsequent down-regulation ofexpression of MMPs, IL-1b, TNFa, COX-2 and NOS2 resulted ina reduced inammatory reaction. Of special note is that the anti-inammatory activity of CS was not only found at the level ofchondrocytes and synovial cells but also on circulating peripheralbloodmononuclear cells [248] suggesting anubiquitousactionof CS.Thus, CS may have the capacity to act as an immunosuppressiveagent in a variety of inammatory conditions associated with NF-KBactivation including the foreign body reaction. Furthermore, both CSand HA were identied as antioxidants capable of reducing freeradicals, protecting both cells andmaterials fromROSdamage [249].CS may therefore provide biomaterials with both immunosuppres-sive and antioxidant properties likely tomodulate the inammatoryresponse toward resolution and healing.

The use of HA for aECM biomaterial coatings, however, is morecomplex. HA features different immunomodulatory activitiesdepending on its molecular size [250]. Large HA polymers, as seenunder physiologic conditions, have anti-inammatory propertiesand promote tissue integrity and quiescence. However, duringtissue injury and inammation, HA becomes cleaved into smallerfragments that function as pro-inammatory and immunostimu-latory mediators potentially serving as endogenous danger signaltriggering PRR of immunocompetent cells. Several studies havereported of low molecular weight HA activating inammatorycytokine and chemokine production in PMNs and monocytes/macrophages as well as maturation of DCs via CD44 and/or TLR4signaling, respectively [251e254]. Implant coating with HA mighttherefore seem counterproductive, since cleavage may cause

f theCD

10 nDC v

S. Franz et al. / Biomaterials 32 (2011) 6692e67096704Fig. 6. Immunomodulating effects of aECM coatings on dendritic cells. A) Composition ocollagen and LPS-stimulation: immature DCs (iDCs) were generated by culture of humanfor 24 h on either aECM or collagen alone with and without co-stimulation with LPS atactivation of T cell immunity were assessed by ow cytometry and ELISA, respectively.

after 24 h culture on aECM was the same as in controls (DC cultured on plastic) and in all eperformed. ManneWhitney U test was used for statistical analysis (*p< 0.05; **p< 0.005)tested aECM coatings. B, C) Articial ECM coatings attenuate DC maturation induced by14 monocytes for 4 days in the presence of GM-CSF and IL-4. The iDCs were culturedg/ml. Expression of DC maturation markers and release of IL-12, the main cytokine foriability was assessed by staining with annexin V and propidium iodide. Viability of DC

xperiments over 85% (data not shown). At least three independent experiments were. Error bars represent standard deviation.

rialsfurther tissue injury. However, high molecular weight (HMW)-HAmight continue acting anti-inammatory even at sites of inam-mation and promote and healing. In a model of non-infectious lunginjury, up-regulation of endogenous HMW-HA production pro-tected against acute inammation due to reduced epithelial cellapoptosis and accelerated tissue repair [252]. HMW-HA has alsobeen shown to promote the function of CD4CD25 TReg [255]. Inuninjured or healing tissue, HMW-HA provides an important signalfor TReg populations to persist, to resolve inammatory processesand maintain homeostasis. There are reports of HA cross-linking tocables at sites of inammation as part of a protective mechanism ininammatory processes (reviewed in Ref. [256]). These HA cablesare pro-adhesive to leukocytes, but suppress their activation byimpeding the interaction with inammation-promoting receptorson the underlying tissues. Moreover, the cross-linked HA structuresare suggested to sequester pro-inammatory mediators and to bemore resistant to degradation in injured tissue, both being impor-tant events for limiting inammation.

Although all these ndings clearly emphasize the potential ofHMW-HA and CS to gear inammatory into regulatory processessuggesting their potential use for immunomodulating biomaterialcoatings, data on controlling function of immunocompetent cellsby aECM composed of collagen and GAGs are lacking. Our grouprecently addressed the immunomodulatory effects of aECM thatwas generated utilizing the natural self-assembly potential of type Icollagen in combination with either HA or CS. HA was additionallymodied by attaching of sulfate groups at low (S1) or high levels(S3) providing binding sites for endogenous growth factors andinammatory mediators (Fig. 6A). Induction of tolerogenic DCfunction by aECM would provide a powerful tool to promoteresolution of inammation and to modulate T cell activity at theimplantation site. In accordance with other studies [128] we foundthat collagen alone provokes DC maturation (Fig. 6B). Culture ofimmature DCs on collagen induces up-regulation of the co-stimulatory molecules CD80 and CD86 (Fig. 6B) as well as releaseof IL-12p40 (Fig. 6B), signals through which DCs direct differenti-ation of T cells toward an immune response. Of note, incorporationof the GAG derivates into the collagen matrix attenuated thecollagen driven DC maturation. Release of IL-12p40 and expressionof maturation marker (MHC, CD86, CD80) were down-regulatedfollowing DC interaction with aECM (Fig. 6C). Moreover, DCmaturation induced by LPS, a potent activator of DCs, is alsodiminished in the presence of aECM as seen by reduced expressionlevel of MHC and CD86 as well as decreased secretion of IL-12p40(Fig. 6C). Dendritic cells that have been prevented to mature areprone to develop a tolerogenic phenotype [113,114]. T cell stimu-lation in an environment lacking of co-stimulation (as typical fortolerogenic DC) can lead to T cell clonal anergy and adaptivetolerance [257]. As one consequence the inammatory activity of Tcells at the implantation site could become compromised either byT cell growth arrest or by down-regulating their effector functions.Additionally T cells with a regulatory phenotype could be inducedcapable to suppress nave T cell responses but also the activity ofinnate immune cells like DC andmacrophages through their releaseof immunosuppressive IL-10 and TGF-b. We are currently investi-gating T cell responses triggered by aECM-activated DC. Since ourdata suggest immunomodulatory capacities of aECM for DC wewillfurther expand our investigations on innate immune cells includingPMNs and macrophages.

5. Conclusions

For a long time, biomaterial science focused on inert materialswhich are recognized as foreign by the host but remain essentially

S. Franz et al. / Biomateunchanged and tolerated due to their encapsulation in broustissue. However, it has been realized that permitting specic cellresponses may be benecial for biomaterial integration andimprove implant performance. Cellular responses at the implan-tation site will also include immune responses to biomaterialsresulting in their rejection or degradation. With expandingknowledge on these processes, biomaterial engineers have begunto develop materials capable of avoiding such detrimental immuneresponses. Current strategies for the design of these biomaterialsfocus on alteration of material surfaces either by modication oftheir physicochemical features or by rendering them biologicallyactive to systematically target cell behavior. Among the latter,functionalization of biomaterials with articial ECM coatingsappears to be particularly promising.

Acknowledgements

This work was funded by the German Research Council (DFGSFB-TR67, projects A3, B3, B5). We thank Markus Karsten for hisoutstanding help in designing the graphics.

References

[1] Williams DF. On the nature of biomaterials. Biomaterials 2009;30:5897e909.[2] Williams DF. On the mechanisms of biocompatibility. Biomaterials 2008;29:

2941e53.[3] Wilson CJ, Clegg RE, Leavesley DI, Pearcy MJ. Mediation of biomaterialecell

interactions by adsorbed proteins: a review. Tissue Eng 2005;11:1e18.[4] Gorbet MB, Sefton MV. Biomaterial-associated thrombosis: roles of coagu-

lation factors, complement, platelets and leukocytes. Biomaterials 2004;25:5681e703.

[5] Schmaier AH. Contact activation: a revision. Thromb Haemost 1997;78:101e7.

[6] Zhuo R, Siedlecki CA, Vogler EA. Autoactivation of blood factor XII athydrophilic and hydrophobic surfaces. Biomaterials 2006;27:4325e32.

[7] Chen X, Wang J, Paszti Z, Wang F, Schrauben JN, Tarabara VV, et al. Orderedadsorption of coagulation factor XII on negatively charged polymer surfacesprobed by sum frequency generation vibrational spectroscopy. Anal BioanalChem 2007;388:65e72.

[8] Vogler EA, Siedlecki CA. Contact activation of blood-plasma coagulation.Biomaterials 2009;30:1857e69.

[9] Zhuo R, Siedlecki CA, Vogler EA. Competitive-protein adsorption in contactactivation of blood factor XII. Biomaterials 2007;28:4355e69.

[10] Sperling C, Fischer M, Maitz MF, Werner C. Blood coagulation on biomaterialsrequires the combination of distinct activation processes. Biomaterials 2009;30:4447e56.

[11] Fischer M, Sperling C, Werner C. Synergistic effect of hydrophobic andanionic surface groups triggers blood coagulation in vitro. J Mater Sci MaterMed 2010;21:931e7.

[12] Heemskerk JW, Bevers EM, Lindhout T. Platelet activation and blood coag-ulation. Thromb Haemost 2002;88:186e93.

[13] Johne J, Blume C, Benz PM, Pozgajova M, Ullrich M, Schuh K, et al. Plateletspromote coagulation factor XII-mediated proteolytic cascade systems inplasma. Biol Chem 2006;387:173e8.

[14] Hu WJ, Eaton JW, Ugarova TP, Tang L. Molecular basis of biomaterial-mediated foreign body reactions. Blood 2001;98:1231e8.

[15] Tang L. Mechanisms of brinogen domains: biomaterial interactions.J Biomater Sci Polym Ed 1998;9:1257e66.

[16] Savage B, Bottini E, Ruggeri ZM. Interaction of integrin alpha IIb beta 3 withmultiple brinogen domains during platelet adhesion. J Biol Chem 1995;270:28812e7.

[17] Rodrigues SN, Goncalves IC, Martins MC, Barbosa MA, Ratner BD. Fibrinogenadsorption, platelet adhesion and activation on mixed hydroxyl-/methyl-terminated self-assembled monolayers. Biomaterials 2006;27:5357e67.

[18] Fischer M, Sperling C, Tengvall P, Werner C. The ability of surface charac-teristics of materials to trigger leukocyte tissue factor expression. Biomate-rials 2010;31:2498e507.

[19] Nilsson B, Ekdahl KN, Mollnes TE, Lambris JD. The role of complement inbiomaterial-induced inammation. Mol Immunol 2007;44:82e94.

[20] Lhotta K, Wurzner R, Kronenberg F, Oppermann M, Konig P. Rapid activationof the complement system by cuprophane depends on complementcomponent C4. Kidney Int 1998;53:1044e51.

[21] Hed J, Johansson M, Lindroth M. Complement activation according to thealternate pathway by glass and plastic surfaces and its role in neutrophiladhesion. Immunol Lett 1984;8:295e9.

[22] Sarma JV, Ward PA. The complement system. Cell Tissue Res 2011;343:227e35.

32 (2011) 6692e6709 6705[23] Tengvall P, Askendal A, Lundstrom II. Ellipsometric in vitro studies on theactivation of complement by human immunoglobulins M and G after

rialsadsorption to methylated silicon. Colloids Surf B Biointerfaces 2001;20:51e62.

[24] Andersson J, Ekdahl KN, Larsson R, Nilsson UR, Nilsson B. C3 adsorbed toa polymer surface can form an initiating alternative pathway convertase.J Immunol 2002;168:5786e91.

[25] Johnson RJ. Complement activation during extracorporeal therapy:biochemistry, cell biology and clinical relevance. Nephrol Dial Transplant1994;9(Suppl. 2):36e45.

[26] Andersson J, Ekdahl KN, Lambris JD, Nilsson B. Binding of C3 fragments ontop of adsorbed plasma proteins during complement activation on a modelbiomaterial surface. Biomaterials 2005;26:1477e85.

[27] KarpmanD,ManeaM,Vaziri-Sani F, StahlAL,KristofferssonAC.Platelet activationin hemolytic uremic syndrome. Semin Thromb Hemost 2006;32:128e45.

[28] Gorbet MB, Sefton MV. Material-induced tissue factor expression but notCD11b upregulation depends on the presence of platelets. J Biomed MaterRes A 2003;67:792e800.

[29] Morley DJ, Feuerstein IA. Adhesion of polymorphonuclear leukocytes toprotein-coated and platelet adherent surfaces. Thromb Haemost 1989;62:1023e8.

[30] McFarland CD, Thomas CH, DeFilippis C, Steele JG, Healy KE. Proteinadsorption and cell attachment to patterned surfaces. J Biomed Mater Res2000;49:200e10.

[31] Keselowsky BG, Bridges AW, Burns KL, Tate CC, Babensee JE, LaPlaca MC,et al. Role of plasma bronectin in the foreign body response to biomaterials.Biomaterials 2007;28:3626e31.

[32] McNally AK, Jones JA, Macewan SR, Colton E, Anderson JM. Vitronectin isa critical protein adhesion substrate for IL-4-induced foreign body giant cellformation. J Biomed Mater Res A 2008;86:535e43.

[33] Kilpadi KL, Chang PL, Bellis SL. Hydroxylapatite binds more serum proteins,puried integrins, and osteoblast precursor cells than titanium or steel.J Biomed Mater Res 2001;57:258e67.

[34] Hynes RO. Integrins: bidirectional, allosteric signaling machines. Cell 2002;110:673e87.

[35] Lowell CA, Berton G. Integrin signal transduction in myeloid leukocytes.J Leukoc Biol 1999;65:313e20.

[36] Jenney CR, Anderson JM. Adsorbed serum proteins responsible for surfacedependent human macrophage behavior. J Biomed Mater Res 2000;49:435e47.

[37] McNally AK, Anderson JM. Complement C3 participation in monocyteadhesion to different surfaces. Proc Natl Acad Sci U S A 1994;91:10119e23.

[38] Keselowsky BG, Collard DM, Garcia AJ. Integrin binding specicity regulatesbiomaterial surface chemistry effects on cell differentiation. Proc Natl AcadSci U S A 2005;102:5953e7.

[39] Keselowsky BG, Collard DM, Garcia AJ. Surface chemistry modulates focaladhesion composition and signaling through changes in integrin binding.Biomaterials 2004;25:5947e54.

[40] McNally AK, Macewan SR, Anderson JM. Alpha subunit partners to beta1 andbeta2 integrins during IL-4-induced foreign body giant cell formation.J Biomed Mater Res A 2007;82:568e74.

[41] Werr J, Eriksson EE, Hedqvist P, Lindbom L. Engagement of beta2 integrinsinduces surface expression of beta1 integrin receptors in human neutrophils.J Leukoc Biol 2000;68:553e60.

[42] McNally AK, Anderson JM. Beta1 and beta2 integrins mediate adhesionduring macrophage fusion and multinucleated foreign body giant cellformation. Am J Pathol 2002;160:621e30.

[43] Bianchi ME. DAMPs, PAMPs and alarmins: all we need to know about danger.J Leukoc Biol 2007;81:1e5.

[44] Babensee JE. Interaction of dendritic cells with biomaterials. Semin Immunol2008;20:101e8.

[45] Bennewitz NL, Babensee JE. The effect of the physical form of poly(lactic-co-glycolic acid) carriers on the humoral immune response to co-deliveredantigen. Biomaterials 2005;26:2991e9.

[46] Kono H, Rock KL. How dying cells alert the immune system to danger. NatRev Immunol 2008;8:279e89.

[47] Grandjean-Laquerriere A, Tabary O, Jacquot J, Richard D, Frayssinet P,Guenounou M, et al. Involvement of toll-like receptor 4 in the inammatoryreaction induced by hydroxyapatite particles. Biomaterials 2007;28:400e4.

[48] Rogers TH, Babensee JE. Altered adherent leukocyte prole on biomaterialsin toll-like receptor 4 decient mice. Biomaterials 2010;31:594e601.

[49] Kobayashi SD, Voyich JM, Burlak C, DeLeo FR. Neutrophils in the innateimmune response. Arch Immunol Ther Exp (Warsz) 2005;53:505e17.

[50] Tang L, Jennings TA, Eaton JW. Mast cells mediate acute inammatoryresponses to implanted biomaterials. Proc Natl Acad Sci U S A 1998;95:8841e6.

[51] Zdolsek J, Eaton JW, Tang L. Histamine release and brinogen adsorptionmediate acute inammatory responses to biomaterial implants in humans.J Transl Med 2007;5:31.

[52] Nimeri G, Ohman L, Elwing H, Wettero J, Bengtsson T. The inuence ofplasma proteins and platelets on oxygen radical production and F-actindistribution in neutrophils adhering to polymer surfaces. Biomaterials 2002;23:1785e95.

[53] Nimeri G, Majeed M, Elwing H, Ohman L, Wettero J, Bengtsson T. Oxygenradical production in neutrophils interacting with platelets and surface-

S. Franz et al. / Biomate6706immobilized plasma proteins: role of tyrosine phosphorylation. J BiomedMater Res A 2003;67:439e47.[54] Labow RS, Meek E, Santerre JP. Neutrophil-mediated biodegradation ofmedical implant materials. J Cell Physiol 2001;186:95e103.

[55] Patel JD, Krupka T, Anderson JM. iNOS-mediated generation of reactiveoxygen and nitrogen species by biomaterial-adherent neutrophils. J BiomedMater Res A 2007;80:381e90.

[56] Scapini P, Lapinet-Vera JA, Gasperini S, Calzetti F, Bazzoni F, Cassatella MA. Theneutrophil asa cellular sourceof chemokines. ImmunolRev2000;177:195e203.

[57] Hidaka Y, Ito M, Mori K, Yagasaki H, Kafrawy AH. Histopathological andimmunohistochemical studies of membranes of deacetylated chitin deriva-tives implanted over rat calvaria. J Biomed Mater Res 1999;46:418e23.

[58] VandeVord PJ, Matthew HW, DeSilva SP, Mayton L, Wu B, Wooley PH.Evaluation of the biocompatibility of a chitosan scaffold in mice. J BiomedMater Res 2002;59:585e90.

[59] Park CJ, Gabrielson NP, Pack DW, Jamison RD, Wagoner Johnson AJ. The effectof chitosan on the migration of neutrophil-like HL60 cells, mediated by IL-8.Biomaterials 2009;30:436e44.

[60] Yamashiro S, Kamohara H, Wang JM, Yang D, Gong WH, Yoshimura T.Phenotypic and functional change of cytokine-activated neutrophils:inammatory neutrophils are heterogeneous and enhance adaptive immuneresponses. J Leukoc Biol 2001;69:698e704.

[61] Gilroy DW. The endogenous control of acute inammation e from onset toresolution. Drug Discov Today Ther Strat 2004;1:313e9.

[62] Anderson JM, Rodriguez A, Chang DT. Foreign body reaction to biomaterials.Semin Immunol 2008;20:86e100.

[63] Hamilton JA. Nondisposable materials, chronic inammation, and adjuvantaction. J Leukoc Biol 2003;73:702e12.

[64] Jones JA, Chang DT, Meyerson H, Colton E, Kwon IK, Matsuda T, et al. Pro-teomic analysis and quantication of cytokines and chemokines frombiomaterial surface-adherent macrophages and foreign body giant cells.J Biomed Mater Res A 2007;83:585e96.

[65] Broughton G, Janis JE, Attinger CE. The basic science of wound healing. PlastReconstr Surg 2006;117:12Se34S.

[66] Mantovani A, Sica A, Sozzani S, Allavena P, Vecchi A, Locati M. The chemo-kine system in diverse forms of macrophage activation and polarization.Trends Immunol 2004;25:677e86.

[67] Mosser DM, Edwards JP. Exploring the full spectrum of macrophage activa-tion. Nat Rev Immunol 2008;8:958e69.

[68] Martinez FO, Sica A, Mantovani A, Locati M. Macrophage activation andpolarization. Front Biosci 2008;13:453e61.

[69] Zhang X, Mosser DM. Macrophage activation by endogenous danger signals.J Pathol 2008;214:161e78.

[70] Stein M, Keshav S, Harris N, Gordon S. Interleukin 4 potently enhancesmurine macrophage mannose receptor activity: a marker of alternativeimmunologic macrophage activation. J Exp Med 1992;176:287e9