o-

Dc,

d A

, Br

ity

Received 23 November 2007; received in revised form 8 May 2008; accepted 2 July 2008Available online 24 July 2008

Keywords: Keratoprosthesis; In vitro resorption; Corals; Ceramics

diseases not amenable to conventional PK, e.g. limbaltransplants, stem cell or ex vivo cultured epithelial trans-plants using autologous limbal, conjunctival or oral muco-sal epithelium. These new emerging procedures havelimited long-term follow up and diering success rates.For example in the dry eye limbal stem cell transplantation

* Corresponding author. Tel.: +44 121 359 3611; fax: +44 121 359 2792.E-mail address: b.j.tighe@aston.ac.uk (B. Tighe).

1 Present address: Bone and Joint Group, Developmental Origins ofHealth and Disease, Southampton General Hospital, Southampton SO166YD, UK.2 Present address: Turku Biomaterials Centre, University of Turku,

FI-20520 Turku, Finland.

Available online at www.sciencedirect.com

Acta Biomaterialia 5 (2009) 4384521. Introduction

The general treatment for serious corneal disease is cor-neal graft by penetrating keratoplasty (PK). Such trans-plants are common and have a success rate in excess of90% in ordinary patients. However, PK failure is virtuallycertain when the ocular surface severely compromised. This

includes patients suering from corneal alkali burn (a seri-ous industrial hazard), StevensJohnson syndrome orrecurrent graft failure, and is a distinct possibility withdry eye, abnormal intraocular pressure (i.e. glaucoma) orongoing ocular inammation. Keratoprosthesis representsthe only viable option for restoring sight in these patients.However, various forms of ocular surface epithelial trans-plantation procedures are now available for ocular surfaceAbstract

Osteo-odonto-keratoprostheses (OOKP) is a unique form of keratoprosthesis involving surgical removal of a tooth root and sur-rounding bone from the patient which are then used to construct an osteo-odonto lamina into which an optical cylinder is cemented.The OOKP procedure is successful and capable of withstanding the very hostile ocular environments found in severe StevensJohnsonsyndrome, pemphigoid, chemical burns, trachoma and multiple corneal graft failure. The existing procedure is complex and time con-suming in terms of operative time, and additionally involves sacrice of the oral structures. This paper discusses the rational search for asynthetic analogue of the dental lamina, capable of mimicking those features of the natural system that are responsible for the successof OOKP. In this study the degradation of selected commercial and natural bioceramics was tested in vitro using a purpose-designedresorption assay. Degradation rate was compared with tooth and bone, which are currently used in OOKP lamina. At normal physio-logical pH the degradation of bioceramics was equivalent to tooth and bone; however, at pH 6.55.0, associated with infectious andinamed tissues, the bioceramics degrade more rapidly. At lower pH the degradation rate decreased in the following order: calciumcarbonate corals > biphasic calcium phosphates > hydroxyapatite. Porosity did not signicantly inuence these degradation rates. Suchdegradation is likely to compromise the stability and viability of the synthetic OOKP. Consequently more chemically stable materials arerequired that are optimized for the surrounding ocular environment.Crown Copyright 2008 Published by Elsevier Ltd. on behalf of Acta Materialia Inc. All rights reserved.Towards a synthetic oste

Reeta Viitala a,2, Valerie Franklin a,Andrew Lloyd

aAston Biomaterials Research Unit, School of Engineering anbSussex Eye Hospital

cBiomedical Research Group, Univers1742-7061/$ - see front matter Crown Copyright 2008 Published by Elseviedoi:10.1016/j.actbio.2008.07.008odonto-keratoprosthesis

avid Green a,1, Christopher Liu b,Brian Tighe a,*

pplied Science, Aston University, Birmingham B4 7ET, UK

ighton BN2 5BF, UK

of Brighton, Brighton BN2 4GJ, UK

www.elsevier.com/locate/actabiomatr Ltd. on behalf of Acta Materialia Inc. All rights reserved.

matand other forms of ocular surface transplantation may notperform optimally [13].

Keratoprostheses are penetrating total replacements ofthe cornea and therefore require the maintenance of a clearvisual window whilst allowing sucient cellular invasion tox the implant rmly in place. This represents a veryconsiderable biomaterials challenge. Extrusion of the ker-atoprosthesis as a result of internal globe pressure is acommon problem, resulting from a failure to promotewound healing at the implanttissue interface. Althoughthe concept of a keratoprosthesis can be traced back to1789 and is attributed to the French ophthalmologistGuillaume Pellier de Quengsy, the rst functional kerato-prosthesis was not fabricated until nearly two centurieslater. This device, produced by Cardona and co-workers[4], used a PMMA nut and bolt design to allow thedevice to be secured into the ocular surface by means ofa trepanned bolt hole.

It is apparent that since Guillaume Pellier de Quengsysspeculative proposal, the articial cornea has had a verychequered developmental history. Some 300 designs of arti-cial cornea have been suggested, but success in implemen-tation has been extremely low. Additionally, success ratesvary widely; data published in 1988 indicated that between8% and 53% of inserted articial corneas fail completelywithin 3 months to 5 years [5]. This contrasts with a 95%success rate for donor corneas in the rst 5 years afterimplantation. In the succeeding decade and beyond, bettersurgical techniques and materials with enhanced propertieshave improved the success of more recent keratoprostheses,such as those developed by Legeais et al. [6], Pintucci et al.[7] and Chirila et al. [8].

The level of interest and considerable level of dicultyin creating a universal replacement for severely damagedcorneas continues. There are still many long-standing com-plications to be completely addressed following the implan-tation of the articial cornea. They include bacterialinfection, enzyme degradation of surrounding tissue, pro-liferation of membranes, raised intraocular pressure andpoor stabilization.

A unique approach to the articial cornea problem, theosteo-odonto-keratoprosthesis (OOKP), was developed inItaly by Strampelli in 1963 [9]. In examining the use of var-ious autologous biological tissues to act as a frame for apolymethyl methacrylate (PMMA) optical cylinder,Strampelli found the tooth root most eective. Strampellisdevice used a lamina prepared from a single-root toothwith surrounding bone sawn from the patients jawbone,into which the PMMA optic was cemented. The devicewas implanted in a subcutaneous pocket below the eye inthe patients cheek for 3 months after which the devicewas removed and surgically inserted into the cornea belowa covering of buccal mucosa. Despite recent advances inthe keratoprosthesis eld, OOKP surgery as pioneered byStrampelli not only provides eective control of most of

R. Viitala et al. / Acta Biothe biological complications, but remains the most success-ful approach for the keratinized eye.Following the pioneering work of Strampelli, Falcinelli[1013] made stepwise improvements to the original tech-nique, ensuring good long-term visual and retention: 75%of his patients achieved 6/12 vision or better. Long-termfollow up has shown good retention results (85% in 18years). As a consequence, this is currently the only proce-dure used in the UK and, because of its cost and complex-ity, is carried out in only one centre, which has a dedicatedophthalmologist working in conjunction with an experi-enced dental surgeon [14].

Modern OOKP surgery is usually performed in twostages, spaced 24 months apart. The rst stage involvesocular surface reconstruction and fashioning of an osteo-odonto lamina and its optical cylinder. The ocular surfaceis reconstructed by suturing a piece of explanted buccalmucosa to the patients sclera. A tooth root and surround-ing bone is surgically removed from the patient and workedinto a lamina with dentine on one side and bone on theother. A hole is drilled through the dentine to accommo-date a PMMA optical cylinder, which is cemented in place.The resultant osteo-odonto lamina is placed into a sub-muscular pocket under the lower lid of the fellow eye, inorder to acquire a soft tissue covering and to allow the lam-ina to recover from any thermal damage caused by the dril-ling. Any infections introduced from the oral cavity can betreated whilst the lamina is sub-muscular prior to implan-tation in the eye. Longer sub-muscular implantationperiods have the potential to lead to signicant resorptionof the lamina. The second stage of the OOKP surgeryessentially involves the retrieval of the osteo-odonto laminafrom its sub-muscular pocket and implantation under thebuccal mucosa. The optic is tted through a trephined holein the patients cornea and the lamina is sutured onto thecornea and sclera. The procedure is completed by cuttinga hole in the overlying buccal mucosa to allow the protru-sion of the anterior part the optical cylinder.

The physiological environment where the OOKP isimplanted is quite complex and depends partly on the med-ical history of the patient. The lateral and anterior sides ofthe OOKP support frame are in contact with the buccalmucous membrane graft, allowing integration due togrowth of the soft tissue into the pores of the bone enablingimproved xation of the implant. The lower part of thesupport frame is in contact with the cornea and possiblewith aqueous humour, which could percolate round theposterior part of the optical cylinder. In seeking to buildon and improve the performance of current OOKPs, it isimportant to examine the advantages, unique featuresand shortcomings of the procedure. The particular struc-tural features of the OOKP that might explain its successare the porosity of the bone and the suspensory periodontalligament linking it to the tooth, thereby forming a semi-rigid block into which is anchored the optical cylinder. Thiscomposite structure of tooth and bone is called the osteo-odonto or dental lamina. The peripheral interconnected

erialia 5 (2009) 438452 439pore spaces of the bone provide xation of supportingtissue and is the key issue, whilst the periodontal ligament

Whilst recognizing that the selection of an appropriateosteo-odonto lamina analogue is a key element in takingOOKP forward, it must be recognized that other factors,such as the use of buccal mucosa to provide an eectivebiological seal around the penetrating implant, contributeto success. The mucosal seal is critical in acting as abarrier to infection and conjunctival downward growth,and protects the articial cornea from dislodgement bymechanical forces. Thus, stable, long-term retention ofthe OOKP is believed to be based upon the specic mate-rial, structural and biochemical properties of these threeintegrated tissues that solve many of the common reasonsfor keratoprosthesis failure, and which are summarized inTable 1. A number of synthetic analogues, such asaluminium oxide, hydroxyapatite ceramic and glass cera-mic, have been fabricated with varying degrees of struc-tural and functional similarity to the OOKP archetype[2226]. None has produced the clinical success of theOOKP paradigm, however. On the other hand, positive

materialia 5 (2009) 438452and cementum appear to provide a barrier, preventingmembrane encapsulation of the device and the downwardsmigration of the mucosal epithelium, which may lead toseparation of the optic from the bone/dentine lamina.

Despite the success of the OOKP procedure and the factthat it is capable of withstanding the very hostile ocularenvironments found in severe StevensJohnson syndrome,pemphigoid, chemical burns, trachoma and multiplecorneal graft failure, it is not without problems. One signif-icant issue is raised intraocular pressure (IOP), which is aconsistent problem, and another, which has been morerecently highlighted, is the problem of long-term bone bior-esorption [1518]. Lamina degradation can cause looseningof the optical part and can eventually lead to the failure ofthe prosthesis. Chronic inammation is reported to play acrucial role in lamina degradation. It has been stated in theliterature that the pH of acutely inamed tissue becomesacidic. Based on animal tests, pHs as low as 5.8 have beenreported in studies relating to pH changes during inam-mation [19]. Inammation can promote the appearanceof an acidic microenvironment in tissues, presumably inresponse to high metabolic activities of inammatory cellsinltrate, and as a consequence of production of acidicproducts of bacterial metabolism [20]. Studies using a ratmodel have shown that Porphyromonas gingivalis, Trepo-nema denticola and Tannerella forsythia not only exist asa consortium that is associated with chronic periodontitisbut also exhibit synergistic virulence resulting in theimmunoinammatory bone resorption of periodontitis[21]. It is also likely that in the eye, in the case of OOKPrelated infections, the pH of the surrounding tissues maybe reduced, which may lead to changes in the degradationbehaviour of the lamina.

Furthermore, the existing procedure is complex andtime consuming in terms of operative time, and the oralstructures (buccal mucous membrane, tooth and bone)are sacriced. In addition, HLA-matched allografting isrequired for edentulous patients, which relies on the good-will of a relative and involves long-term use of immunosup-pression of the recipient with the increased likelihood ofrejection and resorption of the OOKP lamina. The dimen-sional limitations of the lamina restrict the size and designof the optic with consequent restriction of visual eld.

The development of an implantable device that could beused in place of the OOKP lamina would revolutionizeOOKP surgery and optic design, oering the opportunityto widen the use of the procedure to other patient groups.The removal of the requirement for orthodontic surgery aspart of the procedure would signicantly reduce the cost tothe primary healthcare and oer an opportunity for thewider use of this complex technique worldwide. Thus, asynthetic analogue of the dental lamina would makethe OOKP less complex to use and more widely available,whilst allowing more freedom in design of both optic andsupport frame; there are therefore clear surgical, economic

440 R. Viitala et al. / Acta Bioand therapeutic advantages in searching for alternativestructures.in vivo results with porous bioceramic support frameshave been obtained. Leon, for example, used coral skele-ton (thermally transformed into carbonated apatite)shaped to the contours of a healthy cornea and gluedthe internal edge onto a PMMA optical cylinder. Goodtissue integration was achieved when such a supportframe was implanted into the corneal stroma [27].

The aim of this paper is to provide an overview of therequirements for synthetic OOKP lamina materials andto present some potential candidate materials for use aslaminae. The structure of an entirely synthetic keratopros-thesis is also discussed. We have selected natural and com-mercial calcium carbonate, and materials based onhydroxyapatite (HA) and tricalcium phosphate (TCP),which mimic the chemical and/or pore structure of boneand tested the in vitro degradation of these materials insimulated aqueous humour. A simulated aqueous humour,

Table 1Suggested reasons for OOKP success in common areas of keratoprosthesisfailure

Articial corneas: commonreasons for failure

OOKP: reasons for success

Lack of vascularization Porous bone encouragesvascularization

Degradation of tissue by enzymes Antiproteases brought in byblood

Inltration of leucocytes Hindered by protrusion ofoptical cylinder and tightmucosal seal at edges

Proliferation of conjunctiva

Unsupported cornea Cornea held in place by mucosain tension and in contact withbase of the articial cornea

Proliferation of epithelium Inhibited by tight epithelium sealand contact with periodontalligamentum and cementum

Shear forces on supporting frame Periodontal ligament absorbscompressive forcesLack of exibility Mucosa cushions againstcompressive shear forces

which mimics the inorganic component of human aqueoushumour, was developed during this study. Degradation ofmaterials was measured at physiological pH 7.4 and at pH6.5, 6.0 and 5.0. The lower pHs were used to demonstratethe probable situation in the case of infected tissue. In vitrodegradation of these materials was compared with cur-rently used OOKP support frame materials: tooth andbone. This provided a unique possibility to compare thein vitro degradation rate of currently used autogenoustooth and bone with natural and synthetic materials, andtherefore to assess their likely long-term viability. This

R. Viitala et al. / Acta Biomatinformation will enable the future selection of a synthetickeratoprosthesis lamina material for an OOKP that signif-icantly lessens the complexity of surgery, reduces costs,lessens pain and saves the patient from having to sacricea healthy tooth and a part of their jawbone.

1.1. Requirements for a dental lamina analogue

It is clear that the OOKP represents an archetype fromwhich biological lessons of integration can be understood.The many instances of failed attempts to produce syntheticanalogues of the OOKP indicate that a more rationalassessment is necessary to identify the key features requiredby structures that would be able to successfully mimic thephysical behaviour and key integrative features of theOOKP support frame. Two obvious areas of importanceare the porosity and the mineral constitution of the naturalsystem. Table 2 summarizes the chemical and structuralfeatures of the OOKP lamina, which mimics the currenttooth and bone laminae.

Extensive investigations into the design of keratopros-theses have demonstrated the importance of the morphol-ogy of the support frame, particularly the role of specicpore sizes, their distribution and the degree of interconnec-tions between them. In recent times, two important kerato-prosthetic designs, those of Chirila [28] and Legeais [29],have focused on obtaining an optimal pore size to maxi-mize tissue integration and minimize the number of likelycomplications. Chirila found that pore size has a directbearing upon the speed and amount of tissue interpenetra-tion. This may explain the moderate clinical successes ofporous DacronTM and polytetrauoroethylene felts [7,29].Legeais found that pore size inuenced the amount of

Table 2Structural requirements of an OOKP support frame mimic

OOKP lamina property Indicated requirements

Porosity (tissue integration) Graded interconnected porosity inthree dimensionsMicroporosity: 1540 lmMacroporosity: 50150 lm

Mineral constitution (biologicaltolerance and stability)

High crystallinityModerate ionicityCalcium phosphate/carbonate

Ca/P ratio of 1.51.67Non-stoichiometric HAcollagen synthesized and the number of cells per unit vol-ume. Pore structure has been an important structural ele-ment in the design of a successful support frame, and thisnding is supported by studies demonstrating a direct cau-sal link between local environment and tissue responses.Often porosity is given as one of the key merits of a tem-porary tissue replacement. Growth into the porosityreduces encapsulation. The parameters of porosity directlyinuence the form that the tissue takes and the rate ofinward growth. Pore sizes must range from 15 lm to allowbrovascular inward growth and 150 lm to facilitate oste-oid development [30].

An implant will function for longer if these propertiesare maximized, i.e. there is rapid and extensive integrationand strong attachment. Both porous solids and brousmeshes are structures that provide coherent frameworksfor strong and stable tissue attachment and residence. Thisis because the structure of the immediate surroundings hasa direct inuence on tissue function. Both the arrangementof spaces in a structure and surface texture convey informa-tion that directly inuences cell phenotype as a result ofstresses placed on the cell membrane and the connectingintracellular framework [31]. Structure therefore, has thecapacity to reorganize damaged tissue that has lost manyof the cues for growth orientation and direction necessaryfor repair.

Material type also has a strong inuence on how sur-rounding tissue behaves. Stromal tissue behaves in a regu-lar manner in and around uoropolymer prostheses,although the rates of extracellular matrix deposition arenot rapid enough to ensure complete enclosure before theattachment of bacteria or brous encapsulating mem-branes. Polymers previously used to construct articialcornea support frame structures were not designed with asurface chemistry that prevents tissue melting or alterna-tively encourages a coordinated reconstruction of extracel-lular matrix. It is important to pursue the development ofmore sophisticated polymer-based support frame for kera-toprostheses, and a second obvious and pragmaticapproach is to seek alternative inorganic analogues of thedental lamina. The OOKP is interesting in that it providesa pore system that allows the rapid and extensive inwardgrowth of tissue (cells and matrix) and a surface biologicalcompatibility.

The logical requirement for a natural artefact is a por-ous structure that matches dental alveolar bone in the sizeof pores and their arrangement. On this basis, suitable can-didates must possess the following ideal pore attributes,if it is to be successful:

a uniform pore diameter that induces custom cellularresponses in space and time;

a diameter of pore interconnections equal to the porediameter this gives an exceptionally high permeability,maximizes transport of nutrients and waste, and maxi-

erialia 5 (2009) 438452 441mizes the number of possible cell contacts for a mini-mum expenditure;

a solid-to-void ratio of one, which means: (i) it optimizesspace lling for support and (ii) maximizes availablespace for cell distribution.

Potential candidates for dental lamina analogues can befound from the natural world and from synthetic materials.In the natural world there is a vast array of structures tochoose from since the animal and plant kingdoms arereplete with foam structures and materials for a varietyof disparate functions. In general, sedentary (non-mobile)

closer examination to nd a species with a skeletal frame-work with the correct average porosity, mineral composi-tion and amenability to processing for a keratoprosthesissupport frame prototype. Marine-derived skeletal frame-works have already been successfully used in bone replace-ment therapies. Table 3 presents a list of the species, fromtwo groups of marine invertebrates with skeletal frame-works, that have been used to replace damaged anddiseased tissue in the human body [38]. Coral skeletonstend to be more widely available and can be cut into a

ee

ens

ens

442 R. Viitala et al. / Acta Biomaterialia 5 (2009) 438452invertebrates dominate, because they possess more elabo-rate porous structures than higher organisms, which donot rely upon them for support and defence [32].

Porous structures with pore dimensions matching thoseof alveolar bone (a type of bone which possesses a cancel-lous texture) are widespread amongst lower marine inverte-brates. As an added bonus, many of these creatures have asimilar mineral composition to bone, with high fractions ofcalcium and phosphate salts. Amongst invertebrates, withthe exception of insects, substantial calcied tissues arefound in most species. Coral skeletons, sea urchin spines,bamboo culms and certain types of marine sponge possesspore geometries (the specic arrangement of cells) very sim-ilar to common types of bone, such as primary lamellar andhaversian bone, with either open or closed cells [3234].Little information exists regarding resorption of marineskeleton matrices in vitro or in vivo. The only data weare aware of relates to nacre, which is remodelled andbecomes welded with native bone. Degradation occursslowly due to the tight ultrastructure and mineral composi-tion. The degradation rate depends on the size and shape ofthe implanted nacre and the cellular environment in theliving system. It is shown that in bone tissue in sheep theraw nacre implants persist even after 9 months of implan-tation [3537]. Echinoderm structures are much moresoluble. Strength and resistance to resorption are increasedby hydrothermal conversion of skeletal matrices to CaP.Nothing is known of marine sponge resorption except forpreliminary data of collagenous marine sponges whichshow no signs of degradation in vivo (unpublished data).

A particularly diverse group of sponges possesses astructure of interwoven strands that build to form feltsand webs, similar to those employed in previous kerato-prosthesis designs, such as that of Legeais et al. [6]. Thesevarious and disparate groups of marine organisms require

Table 3Naturally derived frameworks applied in human reconstructive surgery

Type Structure

Echinoderm spine: Heterocentrotus trigonarius Periodic minimal surfacEchinoderm skeletal plates and spines:

Acanthaster plancii and NardoaPeriodic minimal surfac

Madrepore coral: Porites porites Vronoi forma: three dimhighly interconnected

Madrepore coral: Goniopora lonata Vronoi forma: three dim

highly interconnected

a Vronoi honeycombs are formed by competitive growth of active cells.greater number of possible shapes than tubular sea urchinspines. In contrast to coral skeletons, sea urchin (echino-derm) spines are rarely employed as tissue replacements,perhaps because of the greater eort in nding andcollecting suitable material [39].

1.2. Coral skeletons

Coral is a name for a variety of marine animals belong-ing to the phlyum Coelenterata. Coelenterates are a prim-itive group of invertebrates with a very simple body planconsisting of two tissue layers (which are mesoderm andectoderm) and perfect symmetry [40]. From a pragmaticas well as analytical point of view, the phylum Coelenteratahas attractive attributes in the search for a dental laminasubstitute in OOKP surgery. In the aragonite skeletons ofreef-building corals there are very many dierent types ofintricate structure. In contrast to sponges, corals are morerigid and more highly structured (pore and channel wallson average three times as thick) due to the larger quantitiesof ceramic deposits. In the genus Porites, for example, thethickness of interconnecting struts and pore sizes closelymatches the open foam of long bone and jawbone. Poritesdeposits calcium carbonate in the form of small coralliteseither in a radial manner or perpendicular to the axis ofgrowth. Both types essentially resemble trabecular bone(lamellar and Haversian types) but possess contrastingspace geometries. Species of Seriatopora closely matchthe porous structure of ner bone types (thin strutsand many interconnections), such as alveolar bone. Beforeany coral can be safely implanted into humans, all organicmaterial needs to be removed and the material has to besterilized. The clinical literature describing the results ofcoral skeletons implanted into the human body is verypositive [4144]. Both hard tissues and soft connective

Biomedical applications

Screw and plugs for bone replacement (experimental only)Screws and plugs for bone replacement (experimentalonly)

ional and Replacement and augmentation of maxillofacial bones,vertebrae and orbital implants

ional and Replacement and augmentation of maxillofacial bones

tissue penetrate the pores and channels rapidly, and remainhealthy therein. Inammation is uncommon because of thehealthy state of tissue and this also inhibits bacterial con-tamination. Coral skeleton has been previously used aloneas a keratoprosthesis support frame but subsequentimplantation trials have shown it to break in situ. Thesemechanical shortcomings can be signicantly improvedby a number of routes. Judicious selection of coral typeswith dierent crystal structures to the Porites coral usedin the previous example can provide structures withimproved resistance to breakage. This all has to do withcrystallite size and organization. Secondly, complete inte-gration with the dentine analogue (we propose a ceramerconstituent to overcome an abrupt boundary between poly-mer and ceramic) along the lower surface of the implantwill stien the overlying coral zone. Thirdly, there exist

attributes of cellular frameworks are reasonably uniform.In relation to fabrication, coral skeletons are easy to cutand render into moderately intricate shapes. This isbecause the mechanical properties of coral are unexpect-edly good for the given amount of porosity (5070% ormore) and mineral constituency.

Given that the primary feature of a coral skeleton thatmakes it highly suitable as a tissue scaold in providing

Table 4Pore sizes (diameter in lm) of common trabecular bone-like corals

Name of coral Macropore diameter Micropore diameter

Seriatopora sp. ca. 400 ca. 105Porites porites ca. 175 ca. 30Goniopora lobata ca. 135 ca. 90Stylophora sp. ca. 245 ca. 140

R. Viitala et al. / Acta Biomaterialia 5 (2009) 438452 443technologies using HA solgel that can be used to coatcomplex pre-formed skeletons without compromising thepore architecture. Such coatings can improve the mechan-ical strength of corals by 120% [45].

All coral skeletons possess open porosity in which coral-lites are connected by pores. In general, interconnectivitydid not vary greatly enough for it to feature prominentlyin selection. A signicant advantage for selection of anappropriate coral skeleton is that they are relatively simpleto identify. Much is now known of the most abundantspecies of coral, their taxonomy and distribution; thereare limits on the exploitation of the less common coral spe-cies, however, because it could irreversibly reduce theirpopulations. This restricts the number of choices for thesupport frame. The habitat conditions in which the coralgrows determines the degree of uniformity in structure.In general, corals can only thrive in uniform conditions,but slight changes do occur and these aect form and struc-ture to a considerable degree. Corals closer to the equatorand distant from land masses are less exposed to alterationsin sea temperature and nutrient loading. Corals expressgreat morphological plasticity in response to slight envi-ronmental dierences. Despite this fact, the physicalFig. 1. SEM images comparing two open cell natural foams of Goniopora lobatstruts and a highly ordered and interconnected pore system.support and space for cell migration and proliferation isits ordered uniform porosity, the logical criteria for selec-tion of suitable coral candidates are (i) porosity, (ii) degreeof structural uniformity, (iii) naturally abundance and (iv)ease of identication and collection. Table 4 indicates theporosity characteristics of a number of corals with anacceptable growth form, abundancy and structural unifor-mity. The majority of coral skeletons possess pores up to500 lm, easily identiable under low magnication. Fromthe group in Table 4 it is possible to reduce the numberof candidates by considering the pore structures in moredetail. Although Seriatopora, Porites, Goniopora and Stylo-phora all possess pore dimensions within acceptable limitsfor tissue inward growth and attachment, by mechanicalinterlock, Seriatopora and Stylophora do not possess anopen enough interconnected porosity, due to a greaterdegree of pore compaction and mineral aggregationaround adjoining struts.

On the basis of this information, Porites and Gonioporacorals appear to be the candidates with the best potentialfor a keratoprosthesis support frame, by virtue of theirextensive and substantial growth form (massive reef build-ers), their structure (highly porous) and their availability ina (left) and Porites porites (right). Both structures consist of relatively thin

the required amounts (they are both widespread and abun-dant). Both Porites and Goniopora occur over a wide rangeof habitats across the globe, from the Indo-Pacic to theCaribbean, and possess pore sizes well within the limitsof pore sizes for alveolar bone. Direct comparison of thetwo structures by eye show the contrasting concentrationof macropores; in Porites these are less obvious. A directcomparison of the two corals is presented in Fig. 1 andTable 5.

Compared with other natural artefacts, corals are usedextensively in bone replacement therapies, spinal surgeryand maxillofacial repairs, and have been for some 20 years.Because of this, corals are easier to obtain in signicant

444 R. Viitala et al. / Acta Biomatquantity from a number of commercial manufacturers.Biocoral is a thermally treated derivative of a species ofPorites coral skeleton, with the zoological name Poritesporites. It is a highly biocompatible biological materialand functions as a resorbable bone graft substitute owingto the many chemical, mineral (crystallographic regularity)and morphological similarities with natural bone. It has thefollowing physical properties: pore diameter of 100200 lm, porosity of 2050%, density of 1.42.8 g cm3

and Youngs modulus of 8 GPa. In vivo, Biocoral, likeany type of thermally treated coral skeleton, initiallyinduces a transient inammatory reaction, perhaps causedby friction between the new developing tissue and theimplant. Importantly, there is no acute chronic inamma-tory or infectious reaction present, particularly with neu-trophils, nor is there brous encapsulation [43]. In thisstudy we have chosen the following corals for degradationexperiments: echinoderm spines, Goniopora, Stylophoraand Biocoral.

1.3. HA-based materials

The mineral part of bone contains calcium phosphate inthe form of HA, and the development of calcium phos-phate ceramics and other related biomaterials for bonegrafting involves better control of biomaterial resorptionand bone substitution. HA is extremely well tolerated bythe body and HA-based materials are a logical alternativefor a synthetic OOKP lamina, because their crystal struc-

Table 5Physical characterization of Goniopora and Porites

Property Goniopora Porites

Contact Open OpenZe/Zf 06/02 06/02Cell shape Ovoid SD below 1 Oblong SD above 1Symmetry Radial RadialFraction of material at edges 0.7 0.5L1/L2 3.7 3.3S.D Ratio 0.7 1.3Density 0.5 0.2 g cm3 0.9 0.3 g cm3

Goniopora and Porites are seen to be physically similar, despite the clear

dierences in the openness of pore volume. The cellular framework hasmany features in common.ture is close to bone, which is currently used in OOKPlamina. HA-basedmaterials can also be derived from corals,like Pro Osteon, which is derived from reef-dwelling coralskeletons. It is chemically transformed from a calcium car-bonate crystalline form into HA by hydrothermal process-ing. Pro Osteon has the following physical properties:pore diameter of 190600 lm, porosity of 5075%, densityof 0.751.1 g cm3 and Youngs modulus of 10 GPa.

Both coral-derived materials Pro Osteon and Biocor-al are sterilized with gamma radiation to reduce infection,and organic components are removed to prevent rejection.Neither Pro Osteon nor Biocoral has been found tocause adverse biological reactions when implanted. Bothmaterials contain interconnected network of pores, whichfall into an optimal range of sizes that allow for the pene-tration and rapid inward growth of structural proteins suchas elastin and collagen matrices (the foundation for theconstruction of a collagen framework) and a well-formedvasculature, essential to wound healing. Both coral deriva-tives have been clinically validated in other applications. Ithas been shown that the rst-generation Pro Osteon hasan extremely slow resorption rate, sometimes taking1520 years to obtain complete resorption when used inthe spinal surgery [46]. It is appropriate to assess in moredetail the suitability of the two materials as OOKP ana-logues. The primary dierence between Pro Osteon andBiocoral is density. To reduce the load exerted on the cor-nea the density should be as low as possible, so ProOsteon might be a slightly better candidate in that sense.The crystals of HA are more robust and are resorbed at aconsiderably slower rate than calcium carbonate. On thisbasis alone, Pro Osteon is therefore the preferred materialfor an articial cornea, which must remain intact through-out its lifetime (over 10 years if the OOKP model can bereproduced accurately). Other HA biomaterials are alsoavailable, like Endobon and PermaBoneTM. PermaBoneTM

is a synthetic and porous HA, whereas Endobon isderived from the spongy bone of cows and possesses acompletely dierent pore architecture to Pro Osteon andBiocoral. Endobon has the following physical proper-ties: pore diameter of 1001500 lm, porosity of 3080%,density of 0.41.3 g cm3 and Youngs modulus of 7 GPa.

The biphasic calcium phosphate ceramic concept isdetermined by an optimal balance of the more stable HAand the more soluble TCP. The material is soluble in thebone environment and gradually dissolves in the body,seeding new bone formation as it releases calcium andphosphate ions into the biological medium. The mainattractive feature of biphasic calcium phosphate ceramicsis their ability to form a direct bond with host bone, result-ing in a strong interface. The formation of this dynamicinterface is the result of a sequence of events involvinginteractions with cells and the formation of carbonatehydroxyapatite (similar to bone mineral) by dissolutionand precipitation processes [47]. There is less information

erialia 5 (2009) 438452on the tissue replacement mechanism in the eye and theuse of bone replacement materials in the eye, parts of which

remain unknown. For example, there is no calcium in theaqueous humour, so in the eye environment it is less likelythat partly dissolved bioceramic would be replaced with car-bonate hydroxyapatite-like structures. On the other hand,because tooth and bone are currently used for OOKP, a log-ical option is to try to replace them with bone substitutematerials in order to nd a suitable synthetic option forautogenous bone. ReproboneTM and Biceram are typicalbiphasic ceramics, and both contain 60% HA and 40%TCP. In this study, the following materials from the syn-thetic HA-based materials group were chosen for the degra-dation experiments: Endobon, Biceram, PermaBoneTM,ReproBoneTM and dense research-purpose HA.

2. Materials and methods

Corals (calcium carbonate), HA- and TCP-based mate-rials were chosen for the in vitro degradation studies, asshown in Table 6. Goniopora and Stylophora are naturalcorals, echinoderm spines were taken from a tropical seaurchin, and Biocoral (Inoteb) is a commercially availablecalcium carbonate, which is derived from natural coral.Three HA ceramics were used: a dense HA previouslydeveloped for research [48], porous Endobon (Merck),which is derived from animal bone, and synthetic Perma-

wisdom tooth was a healthy adult tooth, which wasremoved by the dentist during a routine dentist control.The human tooth was cut into smaller pieces in the follow-ing way: rst the enamel part was removed and the dentinfrom the root was used. The root dentin was cut into smal-ler pieces and used in the short-term degradation experi-ments. Immature pig teeth were found inside the jawboneof a young pig and cut into smaller pieces. The pig teethwere also used in the short-term degradation experiments.The jawbone of the same pig was used as a source of cor-tical and trabecular bone, which were separated by cuttingand studied separately. This jawbone was obtained from abutcher.

A simulated aqueous humour, which mimics the inor-ganic composition of human aqueous humour, was devel-oped and used as the in vitro model uid. The followingchemicals were used to prepare the simulated aqueoushumour: sodium chloride 99.5% (Sigma), potassiumchloride ACS reagent (Sigma), sodium bicarbonate 99.5%(Sigma), potassium phosphate dibasic trihydrate99% (Sigma), tris(hydroxymethyl)aminomethane 99.8%(Aldrich) and citric acid 99% (Acros Organics). Table 7 liststhe inorganic composition of the human aqueous humourand the in vitro model uid. Zinc is present in the humanaqueous humour; however, it could not be added to the sim-

PO4perch

PO4sed f

a10a3(setic

PO4

s

etic

a10a3(

R. Viitala et al. / Acta Biomaterialia 5 (2009) 438452 445BoneTM (Ceramisys). Two biphasic calcium phosphateceramics comprising 60% HA and 40% TCP were used:Biceram (SEM science and medicine) and ReproBoneTM

(Ceramisys). Pig jawbone, immature pig teeth and a humanwisdom tooth were used as reference materials. The human

Table 6Materials used in degradation studies and their chemical composition

Corals HA-based materials

Echinodermspines

CaCO3 Hydroxyapatite Ca10(Natural Dense

Resea

Goniopora CaCO3 Endobon Ca10(

Natural PorouDerivbone

Stylophora CaCO3 Biceram 60% C

Natural 40% CPorouSynth

Biocoral CaCO3 PermaBoneTM Ca10(

Derived from naturalcorals

Porou

Synth

ReproBoneTM 60% C40% C

PorousSyntheticulated aqueous humour because it caused precipitation ofthe uid. Two dierent buers, TRIS and citric acid, wereused, depending on the pH required. The TRIS was usedto buer the solution to pH 7.4, whilst citric acid was usedto buer the solution to pH 6.5, 6.0 and 5.0.

Bone and tooth

)6(OH)2 Human tooth 70% Ca10(PO4)5CO3(OH)2llets 20% collagenpurpose 10% water

Dentin

)6(OH)2 Pig tooth 70% Ca10(PO4)5CO3(OH)220% collagen

rom animal 10% water

Immature dentin, inside pigjawbone

(PO4)6(OH)2 Cortical bone 70% Ca10(PO4)6(OH)2PO4)2 20% collagen

10% waterPig jawbone

)6(OH)2 Trabecularbone

70% Ca10(PO4)6(OH)2

Pig 20% collagen

10% waterPig jawbone

(PO4)6(OH)2PO4)2

tion time of corals is presented in Fig. 2. The degradationrate of corals increased as the pH decreased. At all studiedpHs the total degradation time was shortest for the echino-derm spines, which had a total degradation time of lessthan 200 h at pH 6.5. Biocoral, Goniopora and Stylophorahave quite similar total degradation times of about 400 h atpH 6.5. At pH 6.0 the total degradation time of Stylophoraand Goniopora is about 350 h and for Biocoral it is about230 h. The total degradation time of corals at pH 5.0increases in the following order: echinoderm spines