Pre-clinical
in vivo models for the
screening of bone biomaterials
for oral/craniofacial indications:
focus on small-animal models
A
N D R E A SS
T A V R O P O U L O S, A
N T O NS
C U L E A N, D
I E T E RD. B
O S S H A R D T, D
A N I E LB
U S E R& B
J €OR NK
L I N G EThe need to regenerate bone lost because of dis-ease, trauma and/or surgery has, during recent years, become common in periodontology and oral surgery, as patients are increasingly requesting tooth longevity or rehabilitation with dental implant treat-ment for improved function and esthetics. Autoge-nous bone is generally considered as the ‘gold standard’ in bone-reconstructive surgery, primarily because of its inherent osteogenetic, osteoconduc-tive and osteoinducosteoconduc-tive capacities and its dynamic capability for remodeling (3). Limited amounts of autogenous bone can be harvested at the primary surgical site; therefore there is often the need for a second surgical intraoral site or, in cases of insuffi-cient amounts of intraoral bone, even for an extra-oral site (e.g. the iliac crest). As inclusion of a second surgical site necessitates extra time, incurs increased costs and is associated with morbidity, it is accepted with reluctance by patients; use of a second surgical site also requires increased operator skills and/or training, which limits the generalizabil-ity of the procedure.
With the aim of avoiding the use of autogenous bone or reducing the amounts needed, various bone-substitute materials, alone or in combination with bioactive substances, are increasingly being consid-ered as stand-alone approaches or as adjuncts to autogenous bone. After positive in vitro testing for hypothesis generation, evaluation of proof-of-princi-ple concepts (i.e. evaluation of potential efficacy), safety and possible unwanted reactions of candidate materials, experimental pre-clinical in vivo studies are performed before proceeding to clinical testing.
Pre-clinical
in vivo modeling:
rationale and relevance
The use of pre-clinical in vivo models as an analog to the human is considered necessary, not only because of ethical concerns regarding safety but often also as a result of practical, including economic, issues, as well as regulatory concerns. Pre-clinical in vivo exper-iments are less cumbersome and are usually less expensive compared with large-scale clinical studies. Cell and/or tissue and organ cultures alone cannot predict the biological activity or possible toxic/delete-rious effects of a biomaterial in humans; and such models cannot accurately recreate the complex in vivo situation in which a multitude of molecular factors, signals and cells interact in a three-dimen-sional environment. Furthermore, to evaluate healing responses or possible side effects it is often necessary to harvest biopsies for further analysis; this may not be possible in humans. In fact, the European Medi-cines Agency (www.ema.europa.eu) and the US Food and Drug Administration (www.fda.org) often require a series of targeted pre-clinical in vivo evaluations before human trials may be initiated. Also, the Decla-ration of Helsinki from 2008 states that medical research involving human subjects has to be based, as appropriate, on animal experimentation, while the welfare of animals used for research must be respected (http://www.wma.net/en/30publications/ 10policies/b3/).
No single pre-clinical in vivo model is ideal for eval-uation of bone biomaterials for oral indications in the sense that it completely recreates the anatomical,
physiological, biomechanical and functional environ-ment of the human mouth and jaws. Variations among species include, among other parameters, the anatomy and the dimensions of the jaws, the alveolar processes and the teeth; occlusion; the amount and character of the gingiva and the mucosa; animal behavior; and the healing rate. It is therefore impor-tant to recognize that differences (small or large) between a pre-clinical in vivo model and the human situation will always be present. A pre-clinical in vivo model is based on similarities and analogies between the process and system under study in the animal and the human; in general, the more similarities in the process/system under study– in health and disease, in the animal and in the human– the more suitable the model (i.e. the information/data obtained can be extrapolated to the human situation more reliably). It is broadly perceived that a close phylogenetic rela-tionship and/or anatomic similarity indicate an iden-tical molecular background and biochemical mechanisms, and comparable physiology; although this may usually be true, it is not always the case. A striking example of this is aspirin, which may be tera-togenic in several animals, including rats, dogs, guinea pigs and monkeys, but is evidently not in humans, despite frequent consumption by pregnant women (3, 44). Nevertheless, the more basic the feature, function or response under study, the more valid the compari-son between animals and humans; basic biological mechanisms are usually common among species, including species far away from humans in terms of taxonomy. For example, significant knowledge about the molecular mechanisms and functions of bone morphogenetic proteins has been obtained in studies performed in Drosophila melanogaster (fruit-fly) (36).
Data extrapolation does not imply that numerical values obtained in pre-clinical in vivo experiments can be directly translated to the human situation; it rather means that a response similar in essence with that observed in the animal would also be expected in the human. Negligence or misconception regarding the meaning of data extrapolation from pre-clinical in vivo models may often result in methodology typolatry at the expense of efficiency and reasoning. Substantial resources are frequently spent to identify the exact degree of biological response of a biomate-rial implanted in an animal; in most cases such precise information cannot relate to the human situa-tion by an order of magnitude. For example, a given amount (in mm3) of bone regeneration within the rabbit sinus augmented with a biomaterial would not translate to the same amount, or just ‘X-times’ the amount, if the same procedure was performed in a
patient. Therefore, performing a complicated histomor-phometric analysis, involving registration of several parameters and evaluation of a large number of serial sections to obtain the exact volume of bone regener-ated within the rabbit sinus, would not provide any additional information of relevance (i.e. is not translatable) compared with that obtained from the evaluation of fewer systematically uniform random sampled sections according to stereological principles (10). The latter approach would instead provide a precise estimate of bone volume, allowing valid comparisons among treatment modalities in a much faster and easier way.
Size issues relevant to pre-clinical
in vivo models
Discovery and characterization of biological mecha-nisms often involves the use of in vivo models, including animals remote from humans in taxonomy (e.g. fruit-fly). However, evaluation of bioactivity, bio-compatibility, toxicity, potential adverse reactions, or viability and efficacy of a biomaterial intended to enhance bone formation, requires models involving animals closer to humans. The use of small animals has several advantages over the use of skeletally large animals. The fact that they are small in size circum-vents the issues regarding the major housing require-ments, in terms of space, facilities and husbandry, which are associated with skeletally larger animals. This allows inclusion of large numbers of animals, thereby facilitating collection of sufficient data to per-form proper analyses and calculations, on a more cost-effective basis. Small animals are easier to han-dle compared with large animals; for example, prepa-ration for, and recovery from, surgery usually takes less time. Small animals can also endure more com-plex surgical procedures with apparently much less discomfort and with fewer complications compared with large animals. Small animals are more resistant to disease and infections, thus reducing the risk of losing experimental units (animals) during the course of the experiment, and (in contrast to monkeys) they do not pose any risk for zoonotic disease transmission (42). They have well-defined and controlled genetic backgrounds and there is less variation among the individual animals in terms of biological response. This is considered to be an important advantage because it means that a smaller number of experi-mental units is likely to be required to achieve statisti-cally valid data. However, in perspective, it must be appreciated that pre-clinical in vivo models attempt
to give answers that should translate to the highly variable biological background of humans. This is one of the reasons why regulatory agents almost always require additional proof of effect, at least regarding medicinal products, in skeletally larger animals before proceeding to clinical studies in humans (50). Finally, the use of small animals for biomedical research apparently has fewer ethical concerns for society and does not meet the same resistance as the use of com-panion animals or nonhuman primates; notably, this attitude may be difficult to rationalize.
As defined above, numerical data from animal experiments should not be directly extrapolated to the human situation; this does not imply that size never matters! Testing for clinical efficacy and rele-vance of biomaterial properties necessitates the use of skeletally large animals, offering the possibility of having sites of clinically relevant dimensions. Use of a ‘critical size defect’ [i.e. an osseous defect that does not heal with bone completely during the lifetime of the animal unless subjected to an intervention (41)], whilst an important aspect regarding the ability of a model to disclose the osteopromotive potential of a biomaterial, may not provide translatable results unless it is of relevant dimensions. For instance, when testing the possible effect of a bone biomaterial on buccal bone resorption in a postextraction socket, the use of a rat extraction socket model (11) would proba-bly not provide much insight into the effect-size of a specific biomaterial when used in the human clinical situation; more clinically relevant information would be obtained if addressing this question in a skeletally larger animal, for example by using the well-charac-terized extraction socket model in dogs (1). Defect size limitations in small-animal models is also prob-lematic regarding the evaluation of cell-therapy approaches; diffusion distances or void volumes, important aspects for characterization of mass trans-port limitations (e.g. diffusion of oxygen) relevant to the survival of transplanted cells, need to have certain minimal dimensions in order to be meaningful (31).
Small size usually does not allow the use (testing) of clinical-type devices (e.g. oral implants and related instruments) and hence requires custom-manufac-ture of such equipment, thus increasing costs. Only a limited number of experimental sites/defects per unit (i.e. per animal) are feasible in small animals, which, in turn, implies that more animals are needed to compare several groups within the same experi-ment; however, this may not be very important as a result of the realistic potential for using larger num-bers of animals. Predictability of procedures, in the sense that procedures can be reproduced without
much deviation or with few complications in the vast majority of animals included in an experiment, may be difficult in small animals. Extraction of molar teeth in rats, for instance, may often lead to tooth and socket fracture (especially in the mandible), apparently because of the narrow alveolar ridge in this animal (33). Small size may also preclude repeated biopsy taking or large-volume blood sam-pling for longitudinal evaluation. Furthermore, small animals present dissimilar biomechanical challenges compared with those of humans. If the question was, for instance, related to loading of oral implants inserted in biomaterial-augmented sites, meaningful evaluation would require the use of skeletally large animals [e.g. monkey (27) or mini pigs (7)], with anatomy and dimensions similar to those of humans, allowing installation of clinical-size oral implants and providing an analogous functional environment (i.e. allowing loading) compared with small animals. Finally, the histotechnical procedures indicated are also not always straightforward in small animals, especially regarding oral/dental structures.
Except for the above-mentioned differences between small laboratory animals and larger animals and/or humans, a further major difference that should also be taken into consideration is the different metabolic rates among species (12). As a rule of thumb, the smaller the animal, the higher the metabolic rate compared with that of a human. Thus, in an experiment involving the evaluation of a drug for its potential to enhance bone regeneration, a relevant dose adjust-ment should consider not only body weight but also differences in metabolic rate, determining distribu-tion of the drug in the tissues and excredistribu-tion rates. The metabolic rate of the animal also determines the rate and amount of a given response (e.g. bone healing), which, in turn, dictates the relevant time-points for data collection during the course of the experiment. Shorter observation periods (i.e. faster data sampling) are usually required when small animals are used compared with larger animals, simply because small animals heal faster. The faster metabolic rate, on the other hand, implies that small animals age faster than larger ones, which may be considered as an advan-tage, allowing evaluations in an environment with aging-related changes in a relatively short time span.
Heterotopic and orthotopic testing
sites
There are basically two distinct categories of test sites for bone biomaterials in small animals: heterotopic
(or ectopic) sites (i.e. within the soft tissues); and orthotopic sites (i.e. involving resident bone). Whether a heterotopic site or an orthotopic site is used depends primarily on the mechanism of action of the biomaterial under test. By definition, osteoin-duction implies that factors/signals interact with local primitive, undifferentiated and pluripotent cells that are somehow stimulated to develop toward a bone-forming cell lineage (51). Obviously, proof of osteoin-ductive potential is more clearly illustrated in a het-erotopic site (e.g. by implantation of the biomaterial in a intramuscular pouch) where no bone cells from the neighborhood can contribute to any potentially observed‘de novo’ bone formation. Intramuscular or epimuscular, subcutaneous and intrafatty implanta-tion sites have also been used, but intramuscular implantation seems to elicit the clearest response (32). Heterotopic sites are also used for evaluation of pharmacological/pharmacokinetic properties or the possible toxicity of factors or devices (43). The mus-cles of the thorax, back and thighs of the rat and the rabbit are commonly used for such tests.
Orthotopic sites are primarily used when the mechanism of action of the biomaterial under test is osteoconduction (i.e. the biomaterial enhances bone formation by providing a template for local factors and cells, including cells from resident bone, to attach onto) (3). Of course, this does not pre-clude the use of biomaterials with osteoinductive properties in orthotopic sites as well. Orthotopic sites may be intraosseous or periosseous, or extra-skeletal. Intraosseous/periosseous sites include sur-gically created defects in a variety of locations [e.g. calvarium (15), mandible (37) and long bones (14)] or naturally existing cavities (i.e. sinus) (52), and the latter involve the use of space-providing devices [e.g. capsules/domes on the mandible (45) or cal-varium (19)] extending beyond the original skeletal envelope. Selection of a particular orthotopic model depends on a variety of factors, such as the mecha-nism of action (i.e. osteoconduction or osteoinduc-tion), the macro-configuration and/or physical properties of the bone biomaterial (e.g. particulated or block form; solid or within a liquid carrier) and the intended clinical indication (i.e. inlay or onlay; weight-bearing or nonweight-bearing). As an exam-ple of test-site selection being dependent on the mechanism of action of a bone biomaterial, it seems better to test osteoinductive materials in the bone-marrow-abundant diaphysial sites of long bones, whereas it seems better to test osteoconduc-tive substances in the trabecular bone of metaphyses or epiphyses.
The majority of bone biomaterials for oral/cranio-facial indications are in granular form. Bearing in mind the above strengths and limitations, well-char-acterized pre-clinical in vivo screening models in rodents, relevant for testing such bone biomaterials for oral/craniofacial indications, are presented here-after; less well-characterized models [e.g. the rabbit maxillary defect model (22) and the rabbit extraction socket model (30)], and models exclusively for testing block grafts/substitutes or involving long-bones, will not be considered in this context.
Intraosseous and periosseous site
models
Critical size calvarial defect
The critical size calvarial defect in rodents (mice and rats) and rabbits, originally developed to simulate fracture nonunions in long bones, is perhaps the most widely used pre-clinical in vivo model for screening bone biomaterials (41). The model consists of a cylindrical full-thickness defect involving the parietal and/or occipital bone; occasionally, in rab-bits, only a partial-thickness defect involving the outer diploe is used. Various dimensions of calvarial defects (depending upon species, strain and age) have been proposed as being of ‘critical size’: 3– 5 mm in mice, 5–8 mm in rats and 15 mm in rabbits are usually mentioned (35). Modifications, usually rel-evant only in rats and rabbits, involve the creation of two or more defects (Fig. 1). Defects receiving experi-mental treatment can thus be compared with defects receiving control treatment or with nontreated defects within the same animal (Fig. 2). Access to the calvarium is made by means of a straight (midline) or L-shaped (lateral) skin incision extending from the nasofrontal area to the occipital protuberance. After skin elevation, the subcutaneous fascia is incised and the calvarial bone surface is exposed by a blunt dis-section through the periosteum. The structure of the calvarium and anatomical location allows the estab-lishment of a uniform, standardized and easily repro-ducible defect, commonly created using a trephine bur. During drilling, care should be taken not to dam-age the dura mater or the underlying blood vessels and cranial sinus; the resulting calvarial disk should be carefully removed to avoid tearing of the subjacent cranial structures. Nevertheless, avoiding trauma to the dura mater in mice is not straightforward; (larger) trauma to the dura mater, beyond any detrimental effect on its bone-regenerative potential (9), results in
protrusion of the brain into the defect during healing, thereby minimizing/eliminating the experimental space. At the end of the procedures, the periosteum, fascia and skin are sutured in layers. The dura and the overlying periosteum and skin provide a relatively stable environment for implanted materials, although membranes or other space-providing devices (e.g. titanium mesh) are sometimes used for better con-tainment of the implanted materials. A possible limi-tation of the model is that it does not normally allow assessment of the biological response of an implanted material to physiological biomechanical loading.
The symmetrical morphology of the defect and its localization not only allow easy access and ease of handling during the surgical procedures, but also
facilitate a straightforward analysis, including the possibility for radiographic and histological evalua-tions. In cases with a good bone-regenerative response and the absence of– or largely resorbed – biomaterials, it may occasionally be difficult to clearly recognize the defect margins by means of incandes-cent light, and polarized light might be needed. The model is well characterized, in the sense that there is information on expected healing outcomes of empty untreated sites for several breeds, defect dimensions and healing times. Usually, a 4- to 8-week observation period is used as it has been demonstrated that bone formation in empty control critical-size calvarial defects reaches a plateau during this period, with only minimal bone formation at longer observation peri-ods (4, 8) (Fig. 3). Possible outcome variables in the critical-size calvarial defect model include: extent of defect closure (length/area); thickness of the osseous bridge; and relative and absolute volumes of tissue components (e.g. bone, marrow, connective tissue and biomaterial).
Rabbit sinus
Maxillary sinusfloor augmentation, nowadays a rou-tine procedure, aims at creating bone of adequate quantity and quality to facilitate the installation, osse-integration and functional loading of implants. As already mentioned, functional loading of implants can be properly simulated only with skeletally large animals. However, evaluation of the potential of a biomaterial to enhance bone formation within the sinus cavity can be relatively realistically simulated in rabbits (2, 18, 52, 53). Rabbit maxillary sinuses have a
Fig. 1. The cranium of an adult rat is several times smaller than that of the rabbit. (Yellow bar= 1 cm.)
A B C
Fig. 2. Using a trephine bur, a 5-mm-diameter defect can be created in each parietal bone (A). The two defects are full thickness (B) and allow comparison of a test (i.e. grafted) site with an empty control site (C) within the same animal.
well-defined ostium opening to their nasal cavities, and air-pressure measurements are similar to those in humans, both for absolute pressures and for syn-chronicity with the respiratory cycle (38). The rabbit sinus cavity is surrounded by a thin layer of cortical bone, which is covered by a sinus mucosa containing numerous serous glands and lined with a pseudostr-atified ciliated epithelium. Also, the sinus cavity in rabbits has a size (e.g. for Japanese white rabbits: 2 cm9 1 cm 9 1.5 cm [anterior–posterior direction 9 width 9 height]) (2) that permits a fair amount of biomaterial to be implanted. The sinus cavity is very easily accessible in the rabbit as it is located just beneath the nasal bone (Fig. 4). A straight incision is made on the skin, and the soft tissues, including the periosteum, are elevated to expose the nasal bone (i.e. the upper border of the sinus); two cylindrical (or ovoid) windows (4–6 mm in diameter) can then be created bilaterally using a round diamond bur or a trephine. The sinus membrane is elevated from the lateral walls and a bone biomaterial can be easily inserted into the resulting compartment; the access
opening must be covered with a membrane to con-tain the biomaterial and to avoid collapse of soft tis-sue into the sinus cavity. The rabbit sinus model has also been used to study implant osseointegration after osteotome sinusfloor elevation (34).
The model is well characterized with information on the sequence of events at several time-points after surgery (from 1 week and up to 6 months) in nonmented sinuses (i.e. negative control) and those aug-mented with autografts (i.e. positive control) sinuses (2, 18, 49, 53). Moreover, the model has a good dis-criminating capacity because bone formation in non-augmented (filled with blood coagulum only) sinuses is minimal even after long periods of time – c. 10% after 6 months. The possibility to create two defects per animal allows for intra-individual comparisons (e.g. test group vs. control group), thus reducing the number of animals needed. Finally, harvesting the samples after the animal has been killed is trouble-free, while the outline of the sinus cavity on histologi-cal sections is, in general, readily identifiable because of the cortical bone margins and it is thus easy to dis-tinguish between resident and newly formed bone. Possible outcome variables include the amount of augmentation, the height of augmentation, relative and absolute volumes of tissue components (e.g. mineralized bone, marrow, connective tissue and bio-material) and bone-to-implant contact.
Extraskeletal site models
Space-providing devices on the
calvarium
The realization that isolation of bone defects using a mechanical barrier (i.e. a membrane) predictably enhances osteogenesis (i.e. the guided bone regen-eration principle) (5, 39) led to attempts to augment
A
B
C
Fig. 3. Histomicrographs of an ungrafted critical size calv-arial defect in the rat after (A) 1, (B) 2 and (C) 4 months of healing. Arrowheads indicate the margins of the original defect. Minimal bone formation originating from the mar-gins of resident bone can be observed even after 4 months of healing. (Toluidine blue stain.)
Fig. 4. The sinus cavity (arrowheads) in the rabbit is located just beneath the nasal bone and is thus easily accessible.
resident bone volume by generating new bone in areas where it is anatomically not present (i.e. guided bone augmentation). The first attempts involved expanded polytetrafluoroethylene hemi-spherical domes of various internal dimensions (5–8 mm) and/or porosities, and with a brim of 1–2 mm, that were placed on the denuded and wounded (scratched) left parietal bone in rats (19, 55). The domes were held in position by the origin of the temporalis muscle and the periosteum and skin, which were repositioned over the domes and sutured. Such blood-filled domes presented with various amounts of new-bone formation after 6– 12 weeks of healing (c. 30–60% of the dome space, respectively), and extending the healing period to 24 weeks did not dramatically alter the results. These studies demonstrated the potential to create bone beyond the genetically determined skeletal envelope, and the model was indeed used in several other experiments, including testing of bone-pro-moting agents (56). Nevertheless, it was also realized in those studies that the model has the major draw-back of lack of stabilization potential of the domes; several domes were found to be displaced in the above-mentioned studies and a large portion of the created space was filled with invading soft-connec-tive tissue. In order to avoid the displacement of the device, screw-retained silicone rings were employed (21). Nevertheless, a common drawback of the model, irrespective of the type of device, was the almost unavoidable involvement of the mid-sag-ittal skull suture, simply because of the relative dimensions of the device and the rat skull.
With an analogous rationale to studies in the rat, space-providing devices (domes or cylinders) have been placed on the rabbit calvarium (23, 40). Rabbit calvarium is much larger and thicker, and it features a clear outer and inner cortical plate and a diploe, in contrast to the rat calvarium, where basically the two cortical plates are fused. These characteristics, and the fact that rabbits are larger than rats, make the sur-gical procedures (e.g. defect creation and groove preparation) much easier and permit the use of larger devices with additional features (such asflanges, slits, threads and screws), facilitating the tight adaptation of the device onto the bone. The model permits the creation of more than one experimental site (usually up to four) per animal. Such space-providing devices placed on the uninjured rabbit skull result in progres-sive bone formation in a rather consistent manner; for example, 5 mm9 3 mm (diameter 9 height) titanium domes are consistently filled with new trabecular bone within 1–3 months (23, 26), while
penetration of the external cortical plate accelerates bone formation (24, 28).
The new bone on the rabbit calvarium shows, in general, rather thin trabeculae and large marrow spaces. Nevertheless, the model also allows simula-tion of osteoporotic condisimula-tions in a meaningful vol-ume: following proper preparation of the animal (i.e. bilateral ovariectomy associated with a low-calcium diet) (29), new-bone formation under the devices is characterized by trabecular thinning – reduced tra-becular connectivity coupled with tratra-becular micro-fractures (26)– compared with healthy animals. Use of this model for testing substances and/or devices, including bone biomaterials, in relation to osteoporo-sis within the maxillofacial skeleton, seems more rele-vant compared with the use of rats or of long-bones (e.g. tibia and femur). This is not only because skele-tal maturity in rats is never achieved and normal in-tracortical remodeling may not be inhibited by ovariectomy (54), but is also a result of the structural differences between the long bones in rodents, consisting basically of a thick cortex and no trabecu-lar bone, and the target tissues in humans (for review of models on medically compromised animals (see Donos et al. (6) in this volume of Periodontology 2000).
The localization allows easy access and ease of han-dling during surgical procedures, while the symmetri-cal morphology of the devices employed, and the fact that these are included in the histological sections, facilitate a straightforward analysis. Possible outcome variables in the model include: the amount of aug-mentation; the height of augaug-mentation; and relative and absolute volumes of tissue components (e.g. mineralized bone, marrow, connective tissue and bio-material).
The
‘capsule model’ in the rat mandible
In order to avoid the drawbacks associated with the placement of devices on the rat calvarium, a model was developed that involves the placement of polytet-rafluoroethylene domes (capsules) on the lateral sur-face of the mandibular ramus on either side of the jaw in rats (17, 45). In the beginning, the capsules were created by hollowed polytetrafluoroethylene spheres (17, 20), but this approach was problematic with respect to standardization of the internal diame-ter of the capsule and also with its stabilization onto the jaw; poor stabilization frequently resulted in soft-tissue invasion. Thereafter, the model was improved by changing the hollowed spheres to custom-made, standardized, rigid, hemisphericalpolytetrafluoroeth-ylene capsules (45). The capsules, with afixed internal diameter (6 mm), possess a 1-mm peripheralflexible collar that facilitates their close adaptation to, and stabilization on, the bone surface. The capsule is placed with its open part facing the bone on the lat-eral aspect of the ramus – exposed by an extraoral approach – and is filled with the biomaterial under test or left empty to serve as a control, and then stabi-lized with silk sutures passing through the peripheral collar and through holes drilled through the ramus (Fig. 5). A secluded space is thus created adjacent to an essentially uninjured bone surface, and at the same time the surrounding soft connective tissues are excluded from participating in the healing process. The space created by the capsule is voluminous con-sidering the local anatomy; the bone at the lateral surface of the ramus in the rat is paper-thin (<0.5 mm). However, compared with models involv-ing the rat calvarium, the surgical part of this model is demanding.
The possibility of having test and control capsules in the same animal allows intra-individual compari-sons, thus reducing the number of animals needed. Use of rigid, noncollapsible capsules creates an exper-imental region that is clearly defined and unchange-able throughout the course of the study period, and the fact that the polytetrafluoroethylene capsules are usually incorporated in the histological sections facili-tates an unbiased analysis in regard to the outline of the experimental region. The model is well character-ized with information on the sequence of events at several time-points after surgery (2 weeks, and 1, 2, 4 and 12 months) in empty control capsules (25, 45, 46). Such nonaugmented capsules present gradually increasing new-bone formation emanating from resi-dent bone at the base of the capsule (i.e. the lateral aspect of the ramus) and occupying about 30% of the capsule space at 4 months (Fig. 6). After 12 months the capsules are consistently completely filled with bone (Fig. 7). This extraskeletally generated bone has
been shown to be stable on a long-term basis (47) and to respond to mechanical stimuli/trauma [e.g. implant placement (48)] in the same way as pristine bone, suggesting that the results obtained with the capsule model are probably also valid for other parts of the skeleton.
Possible outcome variables in the capsule model include: the amount of augmentation; the height of augmentation; relative and absolute volumes of tissue components (bone, marrow, connective tissue and biomaterial); and bone-to-implant contact.
Critical-size defect models:
relevance and pitfalls
A defect is defined as of ‘critical size’ when it has ade-quate dimensions so that it does not heal completely with bone during the lifetime of the animal, unless subjected to intervention (41). More strict definitions describe defects that show bone regeneration of less than 10% during the lifetime of the animal, or for practical reasons within a year from intervention (13), as‘critical defects’. The ‘lifetime’ of the animal in this context is determined by the study design (i.e. it cor-responds to the duration of the experiment and in the vast majority of cases is much shorter than the bio-logically determined lifetime of the animal). Even a very small size of defect would be of critical size, pro-vided that the experiment is of short enough dura-tion. Meaningful results regarding the ability of a biomaterial to enhance bone formation can be pro-duced only if the defects have relevant dimensions. Critical evaluation of defect size relative to healing time is thus of utmost importance when interpreting the results of an experiment employing a‘critical size defect’. A relevant point of criticism regarding the critical size calvarial defect model is that the absence of bone formation in empty control defects is largely caused by collapse of the overlying soft tissues. This,
A B C
Fig. 5. The capsules are placed with their opening toward the lateral aspect of the mandibular ramus and can be left empty to serve as controls (A), orfilled with the biomaterial under test (B) and then stabilized with silk sutures (C).
in turn, does not necessarily reflect the innate bone-regenerative potential of the site. Enhanced bone formation in grafted defects may solely be a function of space provision of the biomaterial and not a result of the bone-promoting capacity of the biomaterial per se. Use of space-providing measures on the extracranial aspect may thus be relevant when using the critical size calvarial defect.
Rigid space-creating devices on a bone surface (e.g. the‘capsule model’ in rats) do not qualify as critical size defect models, according to the original defini-tion (41), because ultimately these become com-pletelyfilled out with bone. Use of ‘long’ observation periods in such models, where potentially both‘test’ and‘control’ sites show advanced/complete healing, may fail to disclose the potential of the biomaterial. Nevertheless, when choosing an appropriate observa-tion period (e.g. shorter than 6–12 months for the capsule model), the potential of a biomaterial to pro-mote bone-tissue formation can readily be discrimi-nated: the amount of bone formed in augmented (test) capsules placed on one side of the jaw can be compared with the amount of bone formed under ‘optimal’ conditions (i.e. in empty [control] capsules placed on the contralateral side of the jaw), which represent the innate bone-forming potential of the model for the specific time period. Capsules
contain-ing materials that favor bone regeneration become filled earlier and/or present larger amounts of newly formed bone compared with originally empty cap-sules. By contrast, capsules filled with biomaterials not enhancing bone formation will becomefilled later and/or present lower amounts of newly formed bone compared with empty control capsules at a given time-point. In addition, the fact that 12-month non-augmented capsules are consistently completelyfilled out with bone offers the possibility – in contrast to the critical size calvarial defect model– of identifying biomaterials/substances that may inhibit bone for-mation. Lack of completefilling in a biomaterial-aug-mented critical size calvarial defect model, even after a long observation period, may be regarded as the habitual healing outcome of the model. An osteoin-hibitory biomaterial may thus be overlooked with the critical size calvarial defect model, while this can be captured with the capsule model provided that a suf-ficiently long observation period is used.
Concluding remarks
Pre-clinical in vivo models involving small animals are necessary platforms for screening biomaterials for their potential to enhance bone formation. No single
A B C
Fig. 6. Originally empty capsules present gradually increasing new-bone formation at 1 (A), 2 (B) and 4 months (C), emanating from resident bone at the base of
the capsule (i.e. the lateral aspect of the ramus); at 4 months the newly formed bone occupies about 30% of the capsule space. (Toluidine blue stain.)
A B C
Fig. 7. At re-entry at 12 months (A), the space under the capsule is consistently completely filled out with bone (B) that has a trabecular appearance (C). (van Gieson’s picro fuchsin stain.)
model can completely recreate the anatomical, physiological, biomechanical and functional environ-ment of the human mouth and jaws. Model selection, including defect size and observation times, should be performed carefully, taking into account the mech-anism of action and the macroconfiguration and/or physical properties of the bone biomaterial, as well as the potential clinical indications. Small-animal mod-els, however, are generally not suitable for evaluating the clinical efficacy of bone biomaterials. Model selec-tion should be based not on the‘expertise’ or capaci-ties of the team, but on a scientifically solid rationale, and the animal model selected should reflect the question for which an answer is sought. It is acknowl-edged that all animal experiments should be designed with care and include sample-size and study power calculations, thus allowing the generation of meaningful data, and that they are subject to ethical application to the relevant authority and are subsequently approved. The post-operative handling and care, including post-operative analgesics, is always taken into consider-ation and should follow best practice (16).
References
1. Araujo MG, Lindhe J. Dimensional ridge alterations follow-ing tooth extraction. An experimental study in the dog. J Clin Periodontol 2005: 32: 212–218.
2. Asai S, Shimizu Y, Ooya K. Maxillary sinus augmentation model in rabbits: effect of occluded nasal ostium on new bone formation. Clin Oral Implants Res 2002: 13: 405–409. 3. Burchardt H. The biology of bone graft repair. Clin Orthop
Relat Res 1983: 174: 28–42.
4. Cooper GM, Mooney MP, Gosain AK, Campbell PG, Losee JE, Huard J. Testing the critical size in calvarial bone defects: revisiting the concept of a critical-size defect. Plast Reconstr Surg 2010: 125: 1685–1692.
5. Dahlin C, Linde A, Gottlow J, Nyman S. Healing of bone defects by guided tissue regeneration. Plast Reconstr Surg 1988: 81: 672–676.
6. Donos N, Dereka X, Mardas N. Experimental models for guided bone regeneration in healthy and medically com-promised conditions. Periodontol 2000 2015: 68: 99–121. 7. Fenner M, Vairaktaris E, Fischer K, Schlegel KA,
Neukam FW, Nkenke E. Influence of residual alveolar bone height on osseointegration of implants in the maxilla: a pilot study. Clin Oral Implants Res 2009: 20: 555–559.
8. Gosain AK, Song L, Yu P, Mehrara BJ, Maeda CY, Gold LI, Longaker MT. Osteogenesis in cranial defects: reassessment of the concept of critical size and the expression of TGF-beta isoforms. Plast Reconstr Surg 2000: 106: 360–371; dis-cussion 372.
9. Greenwald JA, Mehrara BJ, Spector JA, Chin GS, Steinbrech DS, Saadeh PB, Luchs JS, Paccione MF, Gittes GK, Longaker MT. Biomolecular mechanisms of calvarial bone induction:
immature versus mature dura mater. Plast Reconstr Surg 2000: 105: 1382–1392.
10. Gundersen HJ, Jensen EB. The efficiency of systematic sam-pling in stereology and its prediction. J Microsc 1987: 147: 229–263.
11. Hahn E, Sonis S, Gallagher G, Atwood D. Preservation of the alveolar ridge with hydroxyapatite-collagen implants in rats. J Prosthet Dent 1988:60: 729–734.
12. Hau J, Poulsen O. Doses for laboratory animals based on metabolic rate. Scand J Lab Anim Sci 1988: 15: 81–85. 13. Hollinger JO, Kleinschmidt JC. The critical size defect as an
experimental model to test bone repair materials. J Cranio-fac Surg 1990: 1: 60–68.
14. Hollinger JO, Schmitt JM, Buck DC, Shannon R, Joh SP, Ze-gzula HD, Wozney J. Recombinant human bone morphoge-netic protein-2 and collagen for bone regeneration. J Biomed Mater Res 1998: 43: 356–364.
15. Hollinger JO, Schmitz JP, Mark DE, Seyfer AE. Osseous wound healing with xenogeneic bone implants with a bio-degradable carrier. Surgery 1990: 107: 50–54.
16. Klinge B, Jonsson J. Animal models in oral health science. In: Hau J, Schapiro SJ, editors. Handbook of laboratory ani-mal science. Boca Raton: CRC Press, 2012: 387–418. 17. Kostopoulos L, Karring T, Uraguchi R. Formation of
jawbone tuberosities by guided tissue regeneration. An experimental study in the rat. Clin Oral Implants Res 1994: 5: 245–253.
18. Lambert F, Leonard A, Drion P, Sourice S, Layrolle P, Rom-pen E. Influence of space-filling materials in subantral bone augmentation: blood clot vs. autogenous bone chips vs. bovine hydroxyapatite. Clin Oral Implants Res 2011: 22: 538–545.
19. Linde A, Thoren C, Dahlin C, Sandberg E. Creation of new bone by an osteopromotive membrane technique: an experimental study in rats. J Oral Maxillofac Surg 1993: 51: 892–897.
20. Lioubavina N, Kostopoulos L, Wenzel A, Karring T. Long-term stability of jaw bone tuberosities formed by “guided tissue regeneration”. Clin Oral Implants Res 1999: 10: 477–486.
21. Lundgren A, Lundgren D, Taylor A. Influence of barrier oc-clusiveness on guided bone augmentation. An experimental study in the rat. Clin Oral Implants Res 1998: 9: 251–260. 22. Lundgren AK, Sennerby L, Lundgren D. An experimental
rabbit model for jaw-bone healing. Int J Oral Maxillofac Surg 1997: 26: 461–464.
23. Lundgren D, Lundgren AK, Sennerby L, Nyman S. Augmen-tation of intramembraneous bone beyond the skeletal envelope using an occlusive titanium barrier. An experi-mental study in the rabbit. Clin Oral Implants Res 1995: 6: 67–72.
24. Majzoub Z, Berengo M, Giardino R, Aldini NN, Cordioli G. Role of intramarrow penetration in osseous repair: a pilot study in the rabbit calvaria. J Periodontol 1999: 70: 1501– 1510.
25. Mardas N, Kostopoulos L, Stavropoulos A, Karring T. Osteo-genesis by guided tissue regeneration and demineralized bone matrix. J Clin Periodontol 2003: 30: 176–183.
26. Mardas N, Schwarz F, Petrie A, Hakimi A-R, Donos N. The effect of SLActive surface in guided bone formation in oste-oporotic-like conditions. Clin Oral Implants Res 2011: 22: 406–415.
27. Marukawa E, Asahina I, Oda M, Seto I, Alam M, Enomoto S. Functional reconstruction of the non-human primate man-dible using recombinant human bone morphogenetic pro-tein-2. Int J Oral Maxillofac Surg 2002: 31: 287–295. 28. Min S, Sato S, Murai M, Okuno K, Fujisaki Y, Yamada Y, Ito
K. Effects of marrow penetration on bone augmentation within a titanium cap in rabbit calvarium. J Periodontol 2007: 78: 1978–1984.
29. Mori H, Manabe M, Kurachi Y, Nagumo M. Osseointegra-tion of dental implants in rabbit bone with low mineral density. J Oral Maxillofac Surg 1997: 55: 351–361; discussion 362.
30. Munhoz EA, Bodanezi A, Cestari TM, Taga R, Ferreira Junior O, de Carvalho PSP. Biomechanical and microscopic response of bone to titanium implants in the presence of inorganic grafts. J Oral Implantol 2011: 37: 19–25.
31. Muschler GF, Raut VP, Patterson TE, Wenke JC, Hollinger JO. The design and use of animal models for translational research in bone tissue engineering and regenerative medi-cine. Tissue Eng Part B Rev 2010: 16: 123–145.
32. Okubo Y, Bessho K, Fujimura K, Konishi Y, Kusumoto K, Ogawa Y, Iizuka T. Osteoinduction by recombinant human bone morphogenetic protein-2 at intramuscular, intermus-cular, subcutaneous and intrafatty sites. Int J Oral Maxillo-fac Surg 2000: 29: 62–66.
33. Pellegrini G, Seol YJ, Gruber R, Giannobile WV. Pre-clinical models for oral and periodontal reconstructive therapies. J Dent Res 2009: 88: 1065–1076.
34. Rahmani M, Shimada E, Rokni S, Deporter DA, Adegbembo AO, Valiquette N, Pilliar RM. Osteotome sinus elevation and simultaneous placement of porous-surfaced dental implants: a morphometric study in rabbits. Clin Oral Implants Res 2005: 16: 692–699.
35. Raut VP, Patterson TE, Wenke JC, Hollinger JO, Muschler GF. Assessment of biomaterials: standardized in vivo test-ing. In: Hollinger JO, editor. An Introduction to Biomateri-als. Boca Raton, FL, USA: CRC Press, 2012: 157–172. 36. Reddi AH. Bone morphogenetic proteins: from basic
sci-ence to clinical applications. J Bone Joint Surg Am 2001: 83-A (Suppl. 1): S1–S6.
37. Sandberg E, Dahlin C, Linde A. Bone regeneration by the osteopromotion technique using bioabsorbable mem-branes: an experimental study in rats. J Oral Maxillofac Surg 1993: 51: 1106–1114.
38. Scharf KE, Lawson W, Shapiro JM, Gannon PJ. Pressure measurements in the normal and occluded rabbit maxillary sinus. Laryngoscope 1995: 105: 570–574.
39. Schenk RK, Buser D, Hardwick WR, Dahlin C. Healing pat-tern of bone regeneration in membrane-protected defects: a histologic study in the canine mandible. Int J Oral Max-illofac Implants 1994: 9: 13–29.
40. Schmid J, H€ammerle CH, Olah AJ, Lang NP. Membrane per-meability is unnecessary for guided generation of new bone. An experimental study in the rabbit. Clin Oral Implants Res 1994: 5: 125–130.
41. Schmitz JP, Hollinger JO. The critical size defect as an experimental model for craniomandibulofacial nonunions. Clin Orthop Relat Res 1986: 205: 299–308.
42. Schou S, Holmstrup P, Kornman KS. Non-human primates used in studies of periodontal disease pathogenesis: a review of the literature. J Periodontol 1993: 64: 497–508. 43. Schuh JCL. Medical device regulations and testing for
toxi-cologic pathologists. Toxicol Pathol 2008: 36: 63–69. 44. Shepard TH. Teratogenicity of therapeutic agents. Curr
Probl Pediatr 1979: 10: 1–42.
45. Stavropoulos A, Kostopoulos L, Mardas N, Nyengaard JR, Karring T. Deproteinized bovine bone used as an adjunct to guided bone augmentation: an experimental study in the rat. Clin Implant Dent Relat Res 2001: 3: 156–165.
46. Stavropoulos A, Kostopoulos L, Nyengaard JR, Karring T. Deproteinized bovine bone (Bio-Oss) and bioactive glass (Biogran) arrest bone formation when used as an adjunct to guided tissue regeneration (GTR): an experimental study in the rat. J Clin Periodontol 2003: 30: 636–643.
47. Stavropoulos A, Kostopoulos L, Nyengaard JR, Karring T. Fate of bone formed by guided tissue regeneration with or without grafting of Bio-Oss or Biogran. An experimental study in the rat. J Clin Periodontol 2004: 31: 30–39.
48. Stavropoulos A, Nyengaard JR, Kostopoulos L, Karring T. Implant placement in bone formed beyond the skeletal envelope by means of guided tissue regeneration: an exper-imental study in the rat. J Clin Periodontol 2005: 32: 1108– 1115.
49. Sun X-J, Zhang Z-Y, Wang S-Y, Gittens SA, Jiang X-Q, Chou LL. Maxillary sinusfloor elevation using a tissue-engineered bone complex with OsteoBone and bMSCs in rabbits. Clin Oral Implants Res 2008: 19: 804–813.
50. Thompson DD, Simmons HA, Pirie CM, Ke HZ. FDA Guide-lines and animal models for osteoporosis. Bone 1995: 17: 125S–133S.
51. Urist MR. Bone: formation by autoinduction. Science 1965: 150: 893–899.
52. Watanabe K, Niimi A, Ueda M. Autogenous bone grafts in the rabbit maxillary sinus. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1999: 88: 26–32.
53. Xu H, Shimizu Y, Asai S, Ooya K. Grafting of deproteinized bone particles inhibits bone resorption after maxillary sinus floor elevation. Clin Oral Implants Res 2004: 15: 126–133. 54. Yamazaki I, Yamaguchi H. Characteristics of an
ovariecto-mized osteopenic rat model. J Bone Miner Res 1989: 4: 13– 22.
55. Zellin G, Linde A. Effects of different osteopromotive mem-brane porosities on experimental bone neogenesis in rats. Biomaterials 1996: 17: 695–702.
56. Zellin G, Linde A. Bone neogenesis in domes made of expanded polytetrafluoroethylene: efficacy of rhBMP-2 to enhance the amount of achievable bone in rats. Plast Rec-onstr Surg 1999: 103: 1229–1237.
