improving oral implant osseointegration in a murine model via wnt signal amplification

Improving oral implantosseointegration in a murinemodel via Wnt signalamplificationMouraret S, Hunter DJ, Bardet C, Popelut A, Brunski JB, Chaussain C, BouchardP, Helms JA. Improving oral implant osseointegration in a murine model via Wntsignal amplification. J Clin Periodontol 2014; 41: 172–180. doi: 10.1111/jcpe.12187.

AbstractAim: To determine the key biological events occurring during implant failure andthen we use this knowledge to develop new biology-based strategies that improveosseointegration.Materials and Methods: Wild-type and Axin2LacZ/LacZ adult male mice underwentoral implant placement, with and without primary stability. Peri-implant tissueswere evaluated using histology, alkaline phosphatase (ALP) activity, tartrate resis-tant acid phosphatase (TRAP) activity and TUNEL staining. In addition, miner-alization sites, collagenous matrix organization and the expression of bonemarkers in the peri-implant tissues were assessed.Results: Maxillary implants lacking primary stability show histological evidenceof persistent fibrous encapsulation and mobility, which recapitulates the clinicalproblems of implant failure. Despite histological and molecular evidence offibrous encapsulation, osteoblasts in the gap interface exhibit robust ALP activity.This mineralization activity is counteracted by osteoclast activity that resorbs anynew bony matrix and consequently, the fibrous encapsulation remains. Using agenetic mouse model, we show that implants lacking primary stability undergoosseointegration, provided that Wnt signalling is amplified.Conclusions: In a mouse model of oral implant failure caused by a lack ofprimary stability, we find evidence of active mineralization. This mineralization,however, is outpaced by robust bone resorption, which culminates in persistentfibrous encapsulation of the implant. Fibrous encapsulation can be prevented andosseointegration assured if Wnt signalling is elevated at the time of implantplacement.

Sylvain Mouraret1,2, Daniel J.Hunter1, Claire Bardet1,3, Antoine

Popelut2, John B. Brunski1,Catherine Chaussain3, Philippe

Bouchard2 and Jill A. Helms1

1Division of Plastic and Reconstructive

Surgery, Department of Surgery, Stanford

School of Medicine, Stanford, CA, USA;2Department of Periodontology, Service of

Odontology, Rothschild Hospital, AP-HP,

Paris 7 – Denis, Diderot University, U.F.R. of

Odontology, Paris, France; 3Dental School

University Paris Descartes PRES Sorbonne

Paris Cit�e, Montrouge, France

Key words: Axin2; bone; dental;

fibrointegration; fibrous encapsulation;

histology; in vivo; maxilla; mice; model; oral

cavity; Wnt

Accepted for publication 29 September 2013

Osseointegration is deemed arequirement for the clinical successof an implant (Branemark et al.1977), but accurately assessing theextent of osseointegration can often-times be problematic. Osseointegra-tion is defined as a direct anchorageof the implant to surrounding bone

under functional loading, and simplemethods to assess this status includeradiographs and torque wrenches(Albrektsson et al. 1986). More com-plex methods include devices thatuse vibrational frequency to measureimplant stability (Meredith et al.1996), but these have inconsistent

Conflict of interest and source of

funding statement

The authors declare that they have noconflict of interest. This study wassupported by a grant to J.A.H. fromthe California Institute of Regenera-tive Medicine (CIRM) TR1-01249.

© 2013 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd172

J Clin Periodontol 2014; 41: 172–180 doi: 10.1111/jcpe.12187

results (Huwiler et al. 2007). Thus,clinicians are often left with obscurecriteria by which to determinewhether osseointegration has beenachieved.

Although not explicitly stated,there is a long-standing assumptionthat implants require primary stabil-ity in order to osseointegrate (Mere-dith 1998). The clinical literature,however, provides examples ofimplants that exhibit primary stabil-ity at the time of placement but thentransiently lose this stability only toregain it at some later time-point(reviewed in Raghavendra et al.2005). The biology underlying thisseemingly contradictory sequence ofevents is unknown.

In previous studies, we directlytested the assumption that primarystability is required for osseointegra-tion by creating a gap interfacebetween an implant and the sur-rounding cortex bone (Leucht et al.2007, Wazen et al. 2013). A gap-typeinterface necessarily creates an envi-ronment in which the implant lacksprimary stability and thus providedus with an experimental model inwhich to assess implant failure. Wefound, however, that when a gap-type interface was created in thetibia, osseointegration still occurred,without any apparent intermediatestep of fibrous encapsulation, pro-vided the implant was stabilized(Leucht et al. 2007, 2012, Wazenet al. 2013). There are, however,notable differences between longbones and craniofacial bones includ-ing embryonic origins (Leucht et al.2008a), their osteogenic plasticity(Leucht et al. 2008a, Richardson2009), and in the presence (orabsence) of an osteogenic marrowcavity. We began this project withthe objective of understanding bio-logical criteria of successful oralimplant osseointegration, using amurine maxillary model rather thana tibial model. We then explored thepossibility of preventing implant fail-ure by creating an environment ofelevated Wnt signalling around theimplant. We specifically employedWnt reporter (Axin2LacZ/LacZ) micefor this purpose because of they rep-resent a robust model of enhancedWnt signalling (Minear et al. 2010,Liu et al. 2013). In doing so, wegained new insights into the biologyof fibrous encapsulation and the

tantilizing possibility that this end-stage failure state may be reversible.

Materials and Methods

Animal husbandry

All procedures were approved bythe Stanford Committee onAnimal Research. Wild-type andAxin2LacZ/LacZ skeletally maturemale mice weighing an average of28 g were obtained from the JacksonLaboratory (Bar Harbor, ME). Ani-mals were given a soft diet food (BioServ product #S3472) and waterad libitum. No antibiotics were givento the operated animals and therewas no evidence of infection or pro-longed inflammation at the surgicalsite.

Implant surgery

Forty adult mice wild type and tenAxin2LacZ/LacZ mice (males, 3–5months old) were anaesthetized withan intraperitoneal injection of Keta-mine (80 mg/kg) and Xylazine(16 mg/kg). The mouth was rinsedusing a povidone-iodine solution for1 min followed by a sulcular incisionthat extended from the maxillaryfirst molar to the mid-point on thealveolar crest. A full-thickness flapwas elevated; a pilot hole was madeto prepare the implant bed on thecrest 1.5 mm in front of the firstmaxillary molar, using a Ø 0.3 mmpilot drill bit (Drill Bit City, Chi-cago, IL, USA). Then animals weredivided into two groups, those inwhich the implant had primary sta-bility (24) and those in which pri-mary stability was lacking (17), seeTable 1. For the former cases, thepilot drill hole was followed with adrill bit of Ø 0.45 mm, using a low-speed dental engine (800 rpm). Thesurgical site was carefully rinsed andthe titanium implant (0.6 mm diame-ter titanium-6 Aluminium-4 Vana-dium alloy “Retopins”; NTI Kahla

GmbH, Kahla, Germany) was cut atlength of 2 mm and was inserted inthe bed preparation.

In cases where the implant lackedprimary stability, the pilot drill holewas followed with a drill bit up to afinal size of Ø 0.65 mm into which a0.6 mm implant was placed. The lar-ger hole ensured that the implantlacked primary stability at the timeof placement.

The flap was closed using non-absorbable single interrupted sutures(Ethilon monofilament 9-0, Johnson& Johnson Medical, Piscataway, NJ,USA). Following surgery, micereceived subcutaneous injections ofbuprenorphine (0.05–0.1 mg/kg) foranalgesia once a day for 3 days.Mice were killed at 7, 14, 21 and28 days post-surgery.

Sample preparation, processing, histology

Maxillae were harvested, the skinand outer layers of muscle wereremoved and the tissues fixed andprocessed as described (Mouraretet al. 2013). Histological stains, Mo-vat’s pentachrome, Aniline blue andPicrosirius red were preformed.

Cellular assays

Alkaline phosphatase (ALP) activityand tartrate-resistant acid phospha-tase (TRAP) activity were performedusing standard procedures (Mouraretet al. 2013). After developing, theslides were dehydrated in a series ofethanol and xylene and subsequentlycover-slipped with Permount mount-ing media.

Immunohistochemistry

Immunostaining was performedusing standard procedures (Mouraretet al. 2013). Antibodies used includeproliferating cell nuclear antigen(PCNA, Invitrogen kit, Camarillo,CA, USA, 1/2000), Osteocalcin (Ab-cam ab93876, Cambridge, MA,

Table 1. Design features of implants

Implant type Genotype Number of implants post-surgery

Day 7 Day 14 Day 21 Day 28

Direct contact CD1 wild type 6 6 6 6Gap interface CD1 wild type 3 7 4 3Gap interface Axin2LacZ/LacZ 2 3 4 1

© 2013 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

Molecular biology of implant failure 173

USA, 1/2000), Osteopontin (NIHLF 175, Bethesday, MD, USA, 1/4000), Decorin (NIH LF 113, 1/1000) and Fibromodulin (LF 149, 1/2000). Negative controls were per-formed during each experiment.Only when there was no evidence ofstaining were the results consideredinformative and included in themanuscript.

Histomorphometric analyses

Maxillae were collected on post-surgical day 21 to quantify the bone-implant contact for the direct contact,gap interface, and Axin2LacZ/LacZ

mice. The 0.6 mm implant was repre-sented across 20 tissue sections; atleast four sections were used to quan-tify the bone-implant contact. Allsections were stained with Anilineblue, which labels osteoid matrix. Thesections were photographed using aLeica digital imaging system at thesame magnification. The resultingdigital images were analysed withAdobe Photoshop CS5. The ruler toolwas used to make all measurements.For each section, the length alongthe implant surface that was directlycontacting the bone was determinedfor both the surfaces. Then, the actualsurface length along each of the twosides was measured. Implants werefabricated from titanium and thus

had to be removed prior to tissuesectioning. Therefore, the void left bythe implant was designated as theimplant surface. We defined bone-implant contact (BIC) as the lack ofintervening soft tissue between histo-logical bone and this void where theimplant had resided. The percentageof BIC was then calculated for bothsides of the implant on each of thesections obtained from each implant,using the BIC length as the numera-tor and the measured implant lengthof the implant as the denominator.

Statistical analyses

Results are presented as themean � SEM. Student’s t-test wasused to quantify differencesdescribed in this article. The p ≤ 0.05was considered significant.

Results

Osseointegration occurs by post-surgery

day 14 in a mouse oral implant model

Our objective at the outset was torecapitulate a condition of oralimplant osseointegration (i.e. directbone-implant contact) and in doingso, capture the limitations and issuesthat confront clinicians. An implantwas placed immediately anterior tothe first maxillary molar, along the

alveolar crest in the edentulous space(Fig. S1).

Histological analyses showed thaton post-surgery day 7 evidence ofbone formation was detectable in theperi-implant space (Fig. 1a). The newbone appeared to extend from theperiosteal surfaces of the native max-illary bone (Fig. 1a′,a″). On day 14,more new bone apparently arisingfrom the periosteum was in contactwith the implant surface (Fig. 1b–b″),indicating that the implant was osseo-integrated at this point. The maxi-mum amount of bone in contact withthe implant surface was achieved byday 21 (Fig. 1c–c″). At day 28, thebone continued to mature and devel-oped a more lamellar organization(Fig. 1d–d″). Of 24 implants placedwith this direct bone-implant contact,23 exhibited primary stability (96%),on par with what is reported for oralimplant osseointegration in humans(Pjetursson et al. 2012). By histologi-cal assessment, 17 achieved osseointe-gration (74%).

Murine peri-implant tissue responses are

similar to large animal responses

In a separate paper, we describe theperi-implant tissue reaction to themaxillary surgery, and the molecularand cellular mechanisms underlyingoral implant osseointegration in this

(a)

(a´) (a´´) (b´) (c´) (d´)(b´´) (c´´) (d´´)

(b) (c) (d)

Fig. 1. Chronology of implant osseointegration in the oral cavity. (a) Representative sagittal tissue section through a maxillaimplant site on post-surgery day 7, stained with Pentachrome. The yellow/blue colour denotes collagen-rich matrix in the defect site.(a′ and a″) High magnification images of the bone in contact with the implant; black dotted lines outline original bone. (b) Maxillaimplant site on post-surgery day 14. (b′ and b″) High magnification image of the bone in contact with the implant. (c) Representa-tive sagittal tissue section through a maxillary implant. (c′ and c″) High magnification images on post-surgery day 21. (d) Represen-tative sagittal tissue section through a maxillary implant site on post-surgery day 28, stained with Pentachrome. (d′ and d″) Highmagnification image of the bone in contact with the implant. Scale bars: (a–d), 200 lm; (a′–d′) 50 lm.


174 Mouraret et al.

model (Mouraret et al. 2013). Here,we focused on differences betweenimplants that had a direct contact withthe maxillary bone (i.e. those that hadprimary stability; Fig. 2a) and thoseimplants that had a gap-type interfaceand consequently lacked primary sta-bility (Fig. 2b). Even after 28 days,implants placed in gap-type interfacesremained surrounded by fibrous con-nective tissue with no evidence ofosseointegration (Fig. 2c).

Bone formation was not curtailedaround implants that lacked primarystability. In addition to the nativemaxillary cortical bone (dotted whiteline, Fig. 2d), aniline blue stainingidentified the newly mineralizedmatrix induced by implant bed prepa-ration (Mouraret et al. 2013) (arrow,Fig. 2d). Despite this robust boneformation, a fibrous, non-mineralizedtissue still occupied the peri-implantspace (arrow, Fig. 2d). Using Picrosi-rus red staining and polarized light(Dayan et al. 1989), the collagenousmatrix in the new bone was evident(arrow, Fig. 2e) as was the persistent

lack of an organized collagenousmatrix in the gap interface (Fig. 2e).Most of the fibrous tissue encapsulat-ing the implant was immunopositivefor Decorin (Weis et al. 2005), an pro-teoglycan involved in connective tis-sue matrix assembly (Fig. 2f).

Molecular differences between

osseointegration and fibrous

encapsulation

Around implants lacking primarystability we detected no appreciablechange in the organization of theperi-implant tissues between days 7and 28, suggesting that fibrousencapsulation represented some end-stage process of cell differentiation(Figs 1 and 2). To determine if thisindeed was the case we evaluated theperi-implant tissues for evidence ofbone turnover. Our controls werethe primary stability cases, where adirect contact between bone and theimplant existed. In these direct-con-tact cases, alkaline phosphatase(ALP) activity (Stucki et al. 2001)

marked sites of active bone minerali-zation around the implant (arrowsFig. 3a). ALP activity was specifi-cally located to sites of new bonedeposition, outlining the lacunae andperiosteal surfaces where activemineralization was taking place(Fig. 3a).

Around implants lacking primarystability (e.g. gap-interface cases),ALP activity was evident around thenew bone (red arrow, Fig. 3b) butwas conspicuously absent in the 125micron-wide zone of fibrous tissueimmediately adjacent to the implant(blue arrow, Fig. 3b). A thin zone ofALP activity was consistentlyobserved in a 50 micron wide zoneadjacent to the native maxillary bone(yellow arrow, Fig. 3b). Thus, boneformation and ALP activity wereconsistent features in the tissuesaround implants that lacked primarystability, but their distribution wasabsent from the peri-implant tissues.

What repressed ALP activity andconsequently, new bone formation inthe zone immediately adjacent to the

(a) (b) (c)

(d) (e) (f)

Fig. 2. Comparative histology of osseointegrated and failed oral implants. (a) Representative tissue section through a direct contactmodel maxilla implant site on post-surgery day 14, stained with Pentachrome; black-dotted line indicates native bone. The yellow/blue colour denotes collagen-rich matrix. (b) Gap interface implant model on post-surgery day 14 and (c) day 28 stained with Pen-tachrome. (d) Aniline blue staining of maxillary gap interface implant model on post-surgery day 28, the red arrow denotes newbone formation. (e) Picrosirius Red staining of maxillary gap interface implant model on post-surgery day 28, observed with polar-ized light; the white arrow denotes the orientation of the new collagen. (f) Representative tissue section through a gap interfaceimplant model on post-surgery day 28, immunostained for Decorin. Scale bars: (a–d), 100 lm.



failed implants? We evaluated TRAPstaining as a measure of osteoclastactivity (Ashton et al. 1993). Indirect contact cases, TRAP activitywas tightly restricted to the new bonesurfaces adjacent to the implant(Fig. 3c). In implant cases lackingprimary stability, TRAP activity wasmuch broader and nearly encom-passed the entire peri-implant spaceincluding the gap (Fig. 3d).

Cell proliferation continues around failed

implants

We found evidence of cell turnoverin cases with a gap-interface. UsingProliferating Cell Nuclear Antigen(PCNA) immunostaining to identifycells undergoing active division, wefirst characterized the control, direct-contact cases, where proliferationwas largely constrained to the cellsalong the edges of the new bone(Fig. 3e). In implants with a gap-interface, cell proliferation wasfound in the same general region(Fig. 3f), indicating that cell turn-over and new bone formation is aongoing process, even around failedimplants. Thus, tissues around animplant lacking primary stability arefar from non-viable. Instead, cell pro-liferation as well as osteogenic differ-entiation potential and active boneturnover are all active cellular pro-cesses found in implant cases charac-terized by fibrous encapsulation.

A genetic approach to enhancing Wnt

signalling prevents implant failure

Given the active state of cell and tis-sue turnover around implants lack-ing primary stability, we wondered ifthe fibroblasts surrounding thesefailed implants could be convertedinto matrix-secreting osteoblasts. Weapproached this question by focusingon the contribution of the pro-osteo-genic Wnt signalling pathway tobone formation (Babij et al. 2003,Day et al. 2005, Gaur et al. 2005).To test if Wnt signalling plays a rolein osseointegration, we used a strainof Wnt reporter mice, Axin2LacZ/+

(Lustig et al. 2002). In the homozy-gous state (i.e. Axin2LacZ/LacZ), micelack both copies of the negativefeedback regulator, Axin2 and con-sequently, endogenous Wnt signal-ling is amplified (Minear et al. 2010).

(a) (b)

(c) (d)

(e) (f)

(g) (h)

Fig. 3. Molecular differences between osseointegration and fibrous encapsulation. (a)Direct contact implant model on post-surgery day 14, stained with Alkaline phospha-tase (ALP) staining, the activity is detectable in the newly mineralizing bone matrix;dotted line indicates native bone, red arrows indicate areas high ALP expression andnew bone formation. (b) Gap interface implant model on post-surgery day 28, stainedwith Alkaline phosphatase staining; yellow double arrow indicates zone of ALP activ-ity, blue double headed arrow indicates area of fibrous tissue, red single arrow indi-cates new bone formation and dotted blue line indicates implant interface. (c)Representative sagittal tissue section through a direct contact implant on post-surgeryday 14, stained with Tartrate resistance acid phosphatase (TRAP) staining. The pinkcolour denotes the osteoclast activity and dotted line indicates native bone. (d) TRAPstaining of gap interface implant model on post-surgery day 28. (e) Direct contactimplant model on post-surgery day 14, immunostained for Proliferating Cell NuclearAntigen (PCNA). (f) PCNA staining of gap interface implant model on post-surgeryday 28. (g) Fibromodulin (FIBRO) staining of direct contact implant model on post-surgery day 14. (h) FIBRO staining of gap interface implant model on post-surgeryday 28. Scale bars: (a–d; g–h), 100 lm; (e–f) 50 lm.


176 Mouraret et al.

We placed implants in control(wild-type and Axin2LacZ/+) andAxin2LacZ/LacZ mice. All implantswere placed into a gap-type interfaceand thus we anticipated that theywould all undergo fibrous encapsula-tion as we had previously docu-mented (Figs 1–3). In wild-type andAxin2LacZ/+ mice, this hypothesiswas correct : all implants placed intogap-type interfaces showed evidenceof fibrous encapsulation of theimplant (17/17; see Table 1 and

Figs 2 and 3). In sharp contrast,when implants with the samegap-interface were placed intoAxin2LacZ/LacZ mice, most underwentosseointegration (9/10; Fig. 4a).

In Axin2LacZ/LacZ peri-implanttissues, Pentachrome and Anilinestaining clearly outlined the lamellararchitecture of the mature bone (dot-ted line, Fig. 4a,b) that was distin-guished from the woven osteoidmatrix occupying the peri-implantspace (double headed arrows,

Fig. 4a,b). Polarized light and Picros-irius red staining demarcated thenative bone (dotted line) from the newosteoid matrix around the implant(arrows, Fig. 4c).

In Axin2LacZ/LacZ mice, ALPactivity was evident around the sur-faces of the newly forming bone nextto the implant and around the bloodvessels in the bone (Fig. 4d). In thesesame samples, TRAP activity wasevident on the remodelling surfacesof the new osteoid matrix and

(a) (b) (c)

(g) (h) (i)

(d) (e)(f)

Fig. 4. A genetic approach to enhancing Wnt signalling prevents implant failure. (a) Representative sagittal tissue section through agap interface model implant on post-surgery day 21 on a Axin2LacZ/LacZ mouse, stained with Pentachrome; dotted line indicatesnative bone and double headed arrow indicates area of new bone around implant. (b) Aniline blue stain, the dark blue colourdenotes collagen in the osteoid tissue, red double headed arrow indicates area of new bone. (c) Picrosirius Red stain, observed withpolarized light; the red colour denotes the orientation of the collagen, white arrows indicate organization of new bone. (d) Alkalinephosphatase (ALP) activity is detectable in the newly mineralizing bone matrix. (e) Tartrate resistance acid phosphatase (TRAP)staining, is evident around newly formed bone. (f) Immunostaining for Osteopontin, osteogenic marker. (g) Osteocalcin immuno-staining marks osteogenic cells. (h) Decorin immunostaining identifies connective tissue matrix assemblies. (i) The bone-implantcontact for three groups were quantified using histomorphometric measurements: group 1 consisted of implants that, when placed,had direct bone contact; group 2 were implants that had a gap-type interface; and group 3, which also had a gap-type interface butwere placed in Axin2LacZ/LacZ mice. Scale bars: (a–e), 100 lm; (f–h) 25 lm. Results are presented as the mean � SEM. * p ≤ 0.05.



around the lacunae associated withthe blood vessels (Fig. 4e). Immuno-staining for the pro-osteogenicmarkers Osteocalcin (Hoffmannet al. 1996) and Osteopontin(Rodan 1995) confirmed that inAxin2LacZ/LacZ mice cells in the gap-type interface were differentiating intoosteoblasts (Fig. 4f,g). TheAxin2LacZ/LacZ peri-implant tissues was devoid ofthe fibrous connective tissue marker,Decorin expression (Fig. 4h).

Finally, we used histomorphome-try to quantify the amount of peri-implant bone that formed in all ofthe test cases. Implants that had pri-mary stability at the time of place-ment had the highest bone-implantcontact (BIC) on day 21, in compari-son to implants that lacked primarystability at the time of implant place-ment (Fig. 4i).

We then compared the BICbetween Axin2LacZ/LacZ mice withwild-type mice. Recall that in bothstrains of mice, the implants lackedprimary stability at the time ofplacement. Nonetheless, on day 21Axin2LacZ/LacZ mice had 70% BICversus 25% BIC in wild-type mice(Fig. 4i). Thus, in cases where pri-mary stability is lacking, elevatingthe level of endogenous Wnt signal-ling was sufficient to induce implantosseointegration.

Discussion

The act of cutting into bone, evenwhen using a low-speed drill runningat 800 rpm, results in osteocyte celldeath. Our data demonstrate emptylacunae evident within 24 h of anosteotomy (Fig. S3, and Mouraretet al. 2013). TUNEL staining of theosteocytes near the cut edge of thealveolar bone indicate that additionalosteocytes are undergoing pro-grammed cell death (Mouraret et al.2013). This cell death triggers a rapidresponse: osteoclasts immediatelybegin to resorb the dead bone (Fig. 3)and via the well-described RANKLfeedback loop (O’Brien et al. 2013),osteogenesis ensues. Consequently,even in cases where there is directcontact, the alveolar bone resorbsand new bone is deposited. Histologi-cally, the native bone can be distin-guished from the new bone by itslamellar architecture (Fig. 3). In caseswhere direct contact does not exist,the initial cellular responses (i.e. an

osteotomy creates a zone of celldeath, osteoclasts begin to resorb themineralized matrix around the emptylacunae, and new bone deposition isinitiated) remain the same. The differ-ence, however, is in the extent towhich this mineralization occurs:implants lacking primary stabilityshow an abrupt halt in mineraliza-tion, falling short of the implant sur-face (Fig. 3a,b). What factorscontribute to this halt in mineraliza-tion remain to be determined.

Fibrous encapsulation can arisebecause of over-loading (Isidor 1996,1997), a lack of initial primary stabil-ity (Fig. 2), or because of an infec-tion around an implant (i.e. peri-implantitis; Jung et al. 2012, Bordinet al. 2009). The principal featuresthat distinguish among these threepossible causes of fibrous encapsula-tion are evidence of initial primarystability and evidence of inflamma-tion and/or infection. In cases offibrous encapsulation secondary toover-loading, the implant would ini-tially exhibit primary stability, andno evidence of bacterial infection(e.g. suppuration, swelling, redness).In cases of fibrous encapsulationsecondary to peri-implantitis, theimplant would initially exhibit pri-mary stability, there may also beevidence of new bone formation incontact with the implant surface butthis would gradually be lost. In allcases of peri-implantitis, infection isevident early on. Here, we presenta model of fibrous encapsulationsecondary to the lack of primarystability. At no time-point during thehealing phase did we observe bone incontact with the implant surface, ordid we observe evidence of infectionor inflammation (Fig. S2). By virtueof the model, where the implant beddiameter was larger than the implantitself, implants lack primary stability.

When such irreversible failures ofimplants occur, clinical standardsdictate that the unstable implantmust be removed and, if sufficientbone stock remains, replaced. Thereis little guarantee that the nextattempt will be successful largelybecause the underlying aetiology ofthe initial failure is oftentimes diffi-cult to ascertain. This is understand-able because factors such as thepatient’s health status and behaviours(Tonetti 1998), conditions of inap-propriate loading (Brunski 1999),

excessive surgical trauma andinfection can all contribute to implantfailure (reviewed in Esposito et al.1998). What if, instead of removal,fibrous encapsulated implants couldbe induced to osseointegrate?

Defining fibrous encapsulation of an

implant

In this study, we relied on histologi-cal, cellular and molecular tools,rather than radiographs or physicalexaminations, to assess the extent ofimplant osseointegration. We alsoemployed cellular markers of osteo-genic differentiation and bone turn-over, as well as histological stains, toevaluate the quality and quantity ofbone surrounding the implant. Thus,the term “osseointegration” could berigorously applied to implants placedin direct contact with the maxillarybone (Figs 1, 2a and 3a,c,e).

In wild-type mice, an initial lackof primary stability resulted in fibrousencapsulation as early as day 7, thehistological appearance of which didnot change over the course of theexperiment (Fig. 2). Thus, the pres-ence of fibrous encapsulation aroundan implant at day 28 represents a per-sistent state. Furthermore, inflamma-tory and immunocompetent cellswere absent from the fibrous tissue,which is characteristic of late-stagefailed implants (Esposito et al. 1997).

Can fibrous encapsulation be reversed?

Despite persist fibrous encapsulationon day 28 we still found evidence ofactive bone mineralization aroundfailing implants (Fig. 3). The body’sattempt at mineralization neverculminated in new bone formation,however, apparently because ofsimultaneous robust osteoclast activ-ity (Fig. 3). Cell turnover, as mea-sured by proliferation markers,indicated the dynamic state of cellsin the fibrous sheath but this still didnot culminate in bone formationaround the implant. These data sug-gest that cells in a gap-type interfaceretain the potential to differentiateinto osteoblasts, if given the properphysical and/or biological stimuli.

Wnt signalling as an osteogenic stimulus

The Wnt signalling pathway plays acentral role in bone development


178 Mouraret et al.

and homeostasis. In most cases, Wntligands promote bone growth (Kimet al. 2007, Leucht et al. 2008b,2013, Minear et al. 2010), andincrease the speed of osseointegra-tion in the tibia (Popelut et al.2010). We gained insights into themechanism by which Wnt signallingregulates adult bone repair throughthe use of the Axin2LacZ/LacZ mousestrain, in which the cellular responseto Wnt signalling is increasedbecause of the removal of a negativeWnt regulator (Minear et al. 2010).Unlike wild-type littermates,implants with a gap-interfaceundergo osseointegration in Axin2-LacZ/LacZ mice (Fig. 4).

We did not observe fibrousencapsulation of implants placed inAxin2LacZ/LacZ mice. Even thoughthere was no initial primary stability(a state which dictates that the max-illary implant will fail in wild-typemice), bone formed in the gap-typeinterface created in Axin2LacZ/LacZ

mice (Fig. 4). From these data, weconclude that primary stability is nota prerequisite for implant osseointe-gration. At least in this model sys-tem, elevated Wnt signalling canovercome the disadvantages resultingfrom a lack of primary stability.Elevated Wnt signalling, as seen inAxin2LacZ/LacZ mice, preventedimplant failure. This finding suggeststhat delivery of a pro-osteogenicstimulus around an implant mayimprove the chances of osseointegra-tion especially in those cases whereanatomy or previous disease hascompromised the implant bed.

References

Albrektsson, T., Zarb, G., Worthington, P. & Eri-ksson, A. R. (1986) The long-term efficacy ofcurrently used dental implants: a review andproposed criteria of success. International Jour-nal of Oral and Maxillofacial Implants 1, 11–25.

Ashton, B. A., Ashton, I. K., Marshall, M. J. &Butler, R. C. (1993) Localisation of vitronectinreceptor immunoreactivity and tartrate resis-tant acid phosphatase activity in synoviumfrom patients with inflammatory or degenera-tive arthritis. Annals of the Rheumatic Diseases52, 133–137.

Babij, P., Zhao, W., Small, C., Kharode, Y., Ya-worsky, P. J., Bouxsein, M. L., Reddy, P. S.,Bodine, P. V., Robinson, J. A., Bhat, B., Mar-zolf, J., Moran, R. A. & Bex, F. (2003) Highbone mass in mice expressing a mutant LRP5gene. Journal of Bone and Mineral Research 18,960–974.

Bordin, S., Flemmig, T. F. & Verardi, S. (2009) Roleof fibroblast populations in peri-implantitis.

International Journal of Oral and MaxillofacialImplants 24, 197–204.

Branemark, P. I., Hansson, B. O., Adell, R., Bre-ine, U., Lindstrom, J., Hallen, O. & Ohman,A. (1977) Osseointegrated implants in the treat-ment of the edentulous jaw: experience from a10-year period. Scandinavian Journal of Plasticand Reconstructive Surgery. Supplementum 16,1–132.

Brunski, J. B. (1999) In vivo bone response to bio-mechanical loading at the bone/dental-implantinterface. Advances in Dental Research 13, 99–119.

Day, T. F., Guo, X., Garrett-Beal, L. & Yang, Y.(2005) Wnt/beta-catenin signaling in mesenchy-mal progenitors controls osteoblast and chon-drocyte differentiation during vertebrateskeletogenesis. Developmental Cell 8, 739–750.

Dayan, D., Hiss, Y., Hirshberg, A., Bubis, J. J. &Wolman, M. (1989) Are the polarization colorsof picrosirius red-stained collagen determinedonly by the diameter of the fibers? Histochemis-try 93, 27–29.

Esposito, M., Hirsch, J. M., Lekholm, U. &Thomsen, P. (1998) Biological factors contrib-uting to failures of osseointegrated oralimplants. (II) Etiopathogenesis. European Jour-nal of Oral Sciences 106, 721–764.

Esposito, M., Thomsen, P., Molne, J., Gretzer,C., Ericson, L. E. & Lekholm, U. (1997)Immunohistochemistry of soft tissues surround-ing late failures of Branemark implants. Clini-cal Oral Implants Research 8, 352–366.

Gaur, T., Lengner, C. J., Hovhannisyan, H.,Bhat, R. A., Bodine, P. V., Komm, B. S.,Javed, A., van Wijnen, A. J., Stein, J. L., Stein,G. S. & Lian, J. B. (2005) Canonical WNT sig-naling promotes osteogenesis by directly stimu-lating Runx2 gene expression. Journal ofBiological Chemistry 280, 33132–33140.

Hoffmann, H. M., Beumer, T. L., Rahman, S.,McCabe, L. R., Banerjee, C., Aslam, F., Tiro,J. A., van Wijnen, A. J., Stein, J. L., Stein, G.S. & Lian, J. B. (1996) Bone tissue-specifictranscription of the osteocalcin gene: role of anactivator osteoblast-specific complex and sup-pressor hox proteins that bind the OC box.Journal of Cellular Biochemistry 61, 310–324.

Huwiler, M. A., Pjetursson, B. E., Bosshardt, D.D., Salvi, G. E. & Lang, N. P. (2007) Reso-nance frequency analysis in relation to jawbonecharacteristics and during early healing ofimplant installation. Clinical Oral ImplantsResearch 18, 275–280.

Isidor, F. (1996) Loss of osseointegration causedby occlusal load of oral implants. A clinicaland radiographic study in monkeys. ClinicalOral Implants Research 7, 143–152.

Isidor, F. (1997) Histological evaluation of peri-implant bone at implants subjected to occlusaloverload or plaque accumulation. Clinical OralImplants Research 8, 1–9.

Jung, S. R., Bashutski, J. D., Jandali, R., Prasad,H., Rohrer, M. & Wang, H. L. (2012) Histo-logical analysis of soft and hard tissues in aperiimplantitis lesion: a human case report.Implant Dentistry 21, 186–189.

Kim, J. B., Leucht, P., Lam, K., Luppen, C., TenBerge, D., Nusse, R. & Helms, J. A. (2007)Bone regeneration is regulated by wnt signal-ing. Journal of Bone and Mineral Research 22,1913–1923.

Leucht, P., Jiang, J., Cheng, D., Liu, B., Dhamd-here, G., Fang, M. Y., Monica, S. D., Urena,J. J., Cole, W., Smith, L. R., Castillo, A. B.,Longaker, M. T. & Helms, J. A. (2013) Wnt3areestablishes osteogenic capacity to bone grafts

from aged animals. Journal of Bone and JointSurgery. American Volume 95, 1278–1288.

Leucht, P., Kim, J. B., Amasha, R., James, A.W., Girod, S. & Helms, J. A. (2008a) Embry-onic origin and Hox status determine progeni-tor cell fate during adult bone regeneration.Development (Cambridge, England) 135, 2845–2854.

Leucht, P., Kim, J. B. & Helms, J. A. (2008b)Beta-catenin-dependent Wnt signaling in man-dibular bone regeneration. Journal of Bone andJoint Surgery. American Volume 90 (Suppl 1),3–8.

Leucht, P., Kim, J. B., Wazen, R., Currey, J. A.,Nanci, A., Brunski, J. B. & Helms, J. A. (2007)Effect of mechanical stimuli on skeletal regen-eration around implants. Bone 40, 919–930.

Leucht, P., Monica, S. D., Temiyasathit, S., Len-ton, K., Manu, A., Longaker, M. T., Jacobs,C. R., Spilker, R. L., Guo, H., Brunski, J. B.& Helms, J. A. (2012) Primary cilia act asmechanosensors during bone healing aroundan implant. Medical Engineering & Physics 35,392–402.

Liu, B., Hunter, D. J., Rooker, S., Chan, A., Pau-lus, Y. M., Leucht, P., Nusse, Y., Nomoto, H.& Helms, J. A. (2013) Wnt signaling promotesMuller cell proliferation and survival afterinjury. Investigative Ophthalmology & VisualScience 54, 444–453.

Lustig, B., Jerchow, B., Sachs, M., Weiler, S., Pie-tsch, T., Karsten, U., van de Wetering, M.,Clevers, H., Schlag, P. M., Birchmeier, W. &Behrens, J. (2002) Negative feedback loop ofWnt signaling through upregulation of conduc-tin/Axin2 in colorectal and liver tumors.Molecular and Cellular Biology 22, 1184–1193.

Meredith, N. (1998) Assessment of implant stabil-ity as a prognostic determinant. The Interna-tional Journal of Prosthodontics 11, 491–501.

Meredith, N., Alleyne, D. & Cawley, P. (1996)Quantitative determination of the stability ofthe implant-tissue interface using resonance fre-quency analysis. Clinical Oral ImplantsResearch 7, 261–267.

Minear, S., Leucht, P., Jiang, J., Liu, B., Zeng,A., Fuerer, C., Nusse, R. & Helms, J. A.(2010) Wnt proteins promote bone regenera-tion. Science Translational Medicine 2, 29ra30.

Mouraret, S., Hunter, D. J., Bardet, C., Brunski,J. B., Bouchard, P. & Helms, J. A. (2013) Apre-clinical murine model of oral implantosseointegration. Bone, doi:10.1016/j.bone.2013.07.021.

O’Brien, C. A., Nakashima, T. & Takayanagi, H.(2013) Osteocyte control of osteoclastogenesis.Bone 54, 258–263.

Pjetursson, B. E., Thoma, D., Jung, R., Zwahlen,M. & Zembic, A. (2012) A systematic review ofthe survival and complication rates of implant-supported fixed dental prostheses (FDPs) aftera mean observation period of at least 5 years.Clinical Oral Implants Research 23 (Suppl 6),22–38.

Popelut, A., Rooker, S. M., Leucht, P., Medio,M., Brunski, J. B. & Helms, J. A. (2010) Theacceleration of implant osseointegration byliposomal Wnt3a. Biomaterials 31, 9173–9181.

Raghavendra, S., Wood, M. C. & Taylor, T. D.(2005) Early wound healing around endosseousimplants: a review of the literature. Interna-tional Journal of Oral and MaxillofacialImplants 20, 425–431.

Richardson, M. (2009) A developmental biologisthighlights potential pitfalls of using stem cellsthat can “remember” their origins [WWW doc-ument].



Rodan, G. A. (1995) Osteopontin overview.Annals of the New York Academy of Sciences760, 1–5.

Stucki, U., Schmid, J., Hammerle, C. F. & Lang,N. P. (2001) Temporal and local appearance ofalkaline phosphatase activity in early stages ofguided bone regeneration. A descriptive histo-chemical study in humans. Clinical OralImplants Research 12, 121–127.

Tonetti, M. S. (1998) Risk factors for osseodisin-tegration. Periodontology 2000 17, 55–62.

Wazen, R. M., Currey, J. A., Guo, H., Brunski,J. B., Helms, J. A. & Nanci, A. (2013) Micro-motion-induced strain fields influence earlystages of repair at bone-implant interfaces.Acta Biomaterialia 24, 313–324.

Weis, S. M., Zimmerman, S. D., Shah, M., Cov-ell, J. W., Omens, J. H., Ross, J. Jr, Dalton,N., Jones, Y., Reed, C. C., Iozzo, R. V. &McCulloch, A. D. (2005) A role for decorin inthe remodeling of myocardial infarction.Matrix Biology 24, 313–324.

Supporting Information

Additional Supporting Informationmay be found in the online versionof this article:

Figure S1. An failure implant modelin mice. (a) Pre-operative photo-graph of the alveolar crest, anteriorto the maxillary first molar; black-

dotted line indicates incision place-ment. (b) The intrasulcular incisionextends from the lingual surface ofthe maxillary first molar anteriorly,to the crest of the edentulous space.(c) A full-thickness flap is elevatedto expose the alveolar bone. (d) A0.65 mm hole is prepared on thecrest, 1.5 mm anterior of the firstmaxillary molar. (e) The 0.6 mmdiameter titanium alloy implant. (f)The implant is placed manually, fol-lowed by careful rinsing. (g) Woundsare closed with non-absorbable sin-gle interrupted sutures. (h) Skeletalpreparation showing location of themaxillary implant relative to thedentition and bones of the skull. (i)Soft tissue covered the healingimplant days post-surgery. M1, max-illa first molar; ab, alveolar bone.Scale bars: (a–d, f–i) 900 lm; (e)2500 lm; (h) 1500 lm.Figure S2. Analysis of the Osteoto-my. (a) Representative sagittal tissuesection through a maxilla cut bonesite using a low-speed drill runningat 800 rpm, post-surgery day 1, cutsite indicated by white rectangle. (b)Stained with DAPI; blue indicates

cell nuclei and empty lacunae weredetected within 24 h of the osteoto-my. (c) TUNEL staining, observedwith fluorescent light, the greencolour denotes cells undergoingapoptosis near the cut site. (d)TRAP staining is evident around cutbone indicate osteocytes near thecut. Scale bars: (a–e), 100 lm.Figure S3. Inflammation of theImplant. (a) Direct contact implantmodel on post-surgery day 7, stainedwith GR1 antibody, indicates granu-locytes and peripheral neutrophils.(b) Representative sagittal tissue sec-tion, stained with GR1, through agap interface model implant on post-surgery day 7. Scale bars: (a–b),100 lm.

Address:Jill A. Helms257 Campus DrivePSRL BuildingStanfordCA 94305USAE-mail: [email protected]

Clinical Relevance

Scientific rationale for the study:Implants can fail, but discriminat-ing between the ones that are des-tined to fail and those that willultimately succeed can be difficult.This represents a real and ongoingchallenge to clinicians, who areoftentimes forced to place implantsinto patients with pre-disposingfactors, or in oral sites where suffi-cient bony support is unavailable.Principal findings: We created amouse model of implant failure by

ensuring that the implant lacked pri-mary stability. The characteristics ofthis model closely mirror what isobserved in large animal models andin humans, but has a distinct advan-tage: a wide array of molecular, cel-lular and genetic tools can be usedto characterize the process of fibrousencapsulation. Our study revealedthat even in a fibrous capsule sur-rounding a loose implant, there isstill evidence of cell proliferationand active bone turnover. Weshowed that fibrous encapsulation

could be prevented, even in caseswhere there is no primary stability,by modulating a stem cell growthfactor that is responsible for boneinduction.Practical implications: This mousemodel can be used to test the influ-ence of implant surface modifica-tions, time to loading, andbiological stimuli on the ability toreverse or prevent implant instabil-ity, which may then inform theclinical practice of implantology inhumans.


180 Mouraret et al.

improving oral implant osseointegration in a murine model via wnt signal amplification

Documents