Anti‐VEGFs hinder bone healing and implant ... · Ranibizumab on bone healing and implant...

Anti-VEGFs hinder bone healingand implant osseointegration inrat tibiaeAl Subaie A, Eimar H, Abdallah M-N, Durand R, Feine J, Tamimi F, Emami E.Anti-VEGFs hinder bone healing and implant osseointegration in rat tibiae. J ClinPeriodontol 2015; 42: 688–696. doi: 10.1111/jcpe.12424

AbstractAim: To assess the effect of anti-vascular endothelial growth factors (VEGF) onbone healing (defect volume) and implant osseointegration (bone-implant contactper cent) in rat tibia.Materials and Methods: In Sprague–Dawley rats (n = 36), a unicortical defectwas created in the right tibia and a titanium implant was placed in the left tibiaof each rat. Rats were assigned into three groups and received either anti-vascularendothelial growth factor neutralizing antibody, Ranibizumab or saline (control).Two weeks following surgery, rats were euthanized and bone samples wereretrieved. Bone healing and osseointegration were assessed using micro-CT andhistomorphometry. One-way ANOVA followed by the Tukey’s test was used fordata analyses.Results: The volume of the bone defects in the anti-VEGF group(2.48 � 0.33 mm3) was larger (p = 0.026) than in the controls (2.11 � 0.36 mm3)as measured by l-CT. Bone-implant contact percent in the anti-VEGF(19.9 � 9.4%) and Ranibizumab (21.7 � 9.2%) groups were lower (p < 0.00)than in the control group (41.8 � 12.4%).Conclusions: The results of this study suggest that drugs that inhibit the activityof vascular endothelial growth factor (i.e. anti-VEGF) may hinder bone healingand implant osseointegration in rat tibiae.

Ahmed Ebraheem Al Subaie1,2,Hazem Eimar1, Mohamed-Nur

Abdallah1, Robert Durand2,Jocelyne Feine1, Faleh Tamimi1,†

and Elham Emami1,3,†

1Faculty of Dentistry, McGill University,

Montreal, QC, Canada; 2College of Dentistry,

University of Dammam, Dammam, Saudi

Arabia; 3Faculty of Dentistry, University of

Montreal, Montreal, QC, Canada

†Indicates equal contributions.

Key words: angiogenesis; anti-VEGF; bone

healing; osseointegration; Ranibizumab;

VEGF

Accepted for publication 8 April 2015

Endosseous implants are widely usedto rehabilitate bone impairments (i.e.fractures) and to restore anatomicalstructures (i.e. lost teeth; Blanchaert

1998, Scully et al. 2007). Successof implant therapies relies on func-tional bone formation in directcontact with the implant, this

process is defined as osseointegration(Albrektsson et al. 1981).

Upon surgical placement,implants create a bone wound, and

Conflict of interest and financial disclosureThe authors revealed no conflict of interest.This study was supported by grants from the International Team for Implantology (ITI) foundation (# 871-2012) and the RoyalEmbassy of Saudi Arabia in Ottawa. Dr Emami was supported by a CIHR clinician-scientist salary award. Ahmed Al Subaiewas sponsored by a scholarship from the Saudi Arabian Cultural Bureau in Canada and the University of Dammam. HazemEimer was supported by scholarships from Fonds de recherch�e Sant�e (FRQS) and Reseau de Recherch�e en Sant�e Bucco-dentaire et Osseuse (RSBO). Mohamed-Nur Abdallah was supported by scholarships from RSBO and Natural Sciences andEngineering Research Council of Canada (NSERC). Dr. Tamimi is supported by, CIHR, NSERC and the Canadian Founda-tion for Innovation (CFI).

© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd688

J Clin Periodontol 2015; 42: 688–696 doi: 10.1111/jcpe.12424

info:doi/10.1111/jcpe.12424

the mechanisms underlying theirosseointegration are very similar tothose occurring during bone repairand fracture healing (Davies 2003,Eriksson et al. 2004). Following afracture or installment of Ti implantsin bone, a blood clot is formed andplatelets release cytokines and growthfactors including vascular endothelialgrowth factors (VEGF). VEGFs area family of growth factors that playsa crucial role in the restoration ofvascular supply during the bone heal-ing process (Ribatti 2005, Beameret al. 2010) and attracts osteoblastsand osteoblastic progenitor cells,stimulating their differentiation(Street et al. 2002).

Despite its importance in osseoin-tegration and bone healing, upregu-lation of VEGFs is associated withpathologic angiogenesis such as thatseen in cancer- and age-relatedmacular degeneration. Therefore,anti-VEGF therapies are very valu-able in clinical oncology as a first-line of treatment for various types ofcancer such as breast cancer, colo-rectal cancer and malignant glioma(Ferrara et al. 2004, Miletic et al.2009). Furthermore, they are used tocontrol aberrant angiogenesis in neo-vascular eye diseases such as age-related macular degeneration (AMD;Kvanta et al. 1996, Avery et al.2006). Nowadays the standard ofcare for this condition is Rani-bizumab, a fragment of a recombi-nant humanized monoclonalantibody Fab (48 kDa) that inhibitsthe activity of human VEGF-A (Ro-senfeld et al. 2006, Lu & Adelman2009, Mitchell 2011, Schmuckeret al. 2012). For the reasons men-tioned above, millions of peopleworldwide are currently using drugsthat target VEGF, and this numberis expected to reach over 500 mil-lions in the next decades (Carmeliet2005, Roskoski 2007).

Inhibiting angiogenesis may havea negative impact on bone healingand implant osseointegration (Davies2003, Eriksson et al. 2004). Indeedanti-angiogenic agents such as TNP-470 have a negative impact on peri-implant bone formation. For thisreason, testing the effect of anti-VEGF on osseointegration has beenstrongly suggested (Mair et al. 2007).Despite its importance, knowledgeon the potential effect of inhibitingVEGF on implant osseointegration

is scarce and this may expose suscep-tible patients to higher risk of fail-ures in implant rehabilitations.

To address this issue, we investi-gated the effect of anti-VEGF andRanibizumab on bone healing andimplant osseointegration in rats. Sim-ilarity between rats and humans interm of bone biology as well as thedifficulty to extrapolate in vitro datainto humans, make rats a very accu-rate model for assessing bone healingand implant osseointegration (Baronet al. 1984, Turner 2001, Pearce et al.2007, Stadlinger et al. 2012). Ourhypothesis was that anti-angiogenicdrugs such as anti-VEGF and Rani-bizumab will negatively affect bonehealing and implant osseointegration.

Materials and Methods

The process of osseointegration andbone healing involves many cells (i.e.osteoblast, osteoclasts), growth fac-tors (i.e. VEGF) and systems (i.e.immune system) that cannot be sim-ulated in vitro and can only beassessed in vivo. Moreover, the Euro-pean Medicines Agency and theFood and Drug Administrationrequire preclinical investigationsbefore human trials. For these rea-sons, we conducted this study toassess the effect of anti-VEGF andRanibizumab on bone healing andimplant osseointegration using ratstibiae model.

Ethical approval for this studywas obtained from McGill AnimalEthics Board (#2012-7269). The studywas conducted on a group of 36female, 10 to 12-week-old Sprague–Dawley rats (Charles River Labora-tories, Montreal, QC, Canada)weighing 200–250 g. The rats werehoused in the Genome Animal Facil-ity of McGill University in a standardenvironmental 22°C with 12-hourlight/dark cycles and a relativehumidity of 30–70%. Every two ratswere caged in a standard sterile cagebedded with wood chips that main-tained them dry and clean. Rodentbreeding diet and water were pro-vided ad libitum. Animals wereallowed to acclimatize to thisenvironment for 2 weeks prior to sur-gical intervention. Animals’ conditionand welfare were assessed daily forany sign of pain, infection, dehis-cence, loss of appetite, weight lossor restricted movement. Analgesics

(Carprofen 5–10 mg/kg) were admin-istered when animals showed signs ofpain. Prior to surgical intervention,rats were randomly divided into threegroups: (i) control (n = 12); (ii)Ranibizumab (n = 12); and (iii) anti-VEGF (n = 12).

Surgical intervention

The surgical part of this study wascarried out on November andDecember, 2013. We performed thisstudy in rats tibia because it is con-sidered a good experimental locationto assess the systemic effects of drugson implant osseointegration (Stadlin-ger et al. 2012). During the surgicalintervention the surgeon was blindedto group allocation and all surgicalinterventions were performed by onesurgeon using standardized instru-ments (i.e. burr and Ti implants) toassure consistent bone defects andimplant placement among all ani-mals. The animals were anaesthe-tized with isoflurane (3–5% atinduction, 2–2.5% at maintenanceperiod). After showing signs of beingfully anesthetized, the legs wereshaved and disinfected with chlorh-exidine scrub and the animals werecovered with a sterile drape. A longi-tudinal full thickness skin incisionwas created over the tibial metaphy-sis. The proximal tibialis muscle andthe periosteum were dissected andconserved (Fig. 1a,c). A unicorticaldefect was created on the medialaspect of the left tibial metaphysis,7–12 mm distal to the knee joint,using a cylindrical burr (1.5 mmø)adapted to a handpiece drill (Stry-ker, Hamilton, ON, Canada) underconstant saline irrigation. A custom-made (1.5 mmØ 9 2.0 mm long)titanium implant was placed in theleft defect (Fig. 1d). The incision wasclosed using 5-0 monocryl sutures.The procedure was repeated on theright tibia creating a 2.5 mmØ defectthat was left empty (Fig. 1b). Per-forming two surgeries in each animalallowed us to reduce the number ofanimals by half. Animals interven-tion and husbandry were refined tominimize pain and distress. Carpro-fen (Pfizer Animal Health, Montr�eal,QC, Canada) was injected (5–10 mg/kg) subcutaneously 30 min. prior tosurgery and 24 h after surgery for2 days in order to provide analgesiato the rats.

© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

Ranibizumab hinders osseointegration 689

Postoperative drug administration

Following surgical intervention, thefirst group of rats received salineinjections (control); the second groupreceived Ranibizumab (Lucentis;Genentech Inc., San Francisco, CA,USA); and the third group receivedanti-VEGF (R&D Systems Inc., Min-neapolis, MN, USA). Ranibizumabwas administered intra-peritoneallyin a single injection of 15 lg dilutedin 1.5 mL of saline (equivalent to therecommended human dose 0.5 mg;Lu & Adelman 2009). Ranibizumabis a humanized antibody that mightnot be effective on rats, therefore,rat-specific anti-VEGF, which hassimilar action to Ranibizumab inrats, was included as a third group inthe study. This group received intra-peritoneal injections of anti-VEGF,4 lg diluted in 1.5 mL of saline, threetimes per week (a dose that shown toaffect blood vessels in rats). The anti-VEGF that we used is a 150 kDaanti-VEGF antibody that neutralizesthe effect of VEGF164 with somereactivity with VEGF120, VEGF165

and <2% cross reactivity with recom-binant human VEGF-B, C, and D(Io et al. 2004, Rocher et al. 2011).

After the intervention, the ratswere left to heal for 2 weeksbefore being euthanized using CO2

overdose and had the tibiaeretrieved. Right tibiae were preservedin paraformaldehyde solution 4% inPBS (Santa Cruz Biotech, Dallas,TX, USA) while the left tibiae werepreserved in 10% neutral-bufferedformalin (Richard Allan Scientific,Kalamazoo, MI, USA). Theretrieved samples were code labelledand samples analyses were per-formed by one person that wasblinded to group allocation.

Assessment of bone healing

A micro-CT analysis was conductedto assess bone healing in the defects.All tibiae with empty defects werescanned using a micro-CT (Sky-Scan1172; Bruker, Kontich, Bel-gium) set at a 12.7 lm resolution, a50 kV voltage, a 0.5 rotation step, a10 random movement and a 0.5 mmaluminium filter. The area of thebone defect was determined andincluded in the region of interest(ROI). The ROI included the entireoriginal defect (2.5 mm Ø, unicorti-cal). The ROI was reconstructed andanalysed by Skyscan CT-analyser(SkyScan1172; Bruker). Threedimensional bone parameters includ-ing tissue volume, bone volume,trabecular thickness, trabecular num-ber and trabecular separation were

obtained from the 3-D reconstructedimages. The volume of the defectwas determined by subtracting thebone volume from the ROI.

Assessment of Blood vessels density and

osteoclast number

All right tibiae samples (tibiae withbone defects) were decalcified using10% EDTA for 3 weeks. The decal-cified bone samples were dehydratedin ascending concentrations of etha-nol (70–95%) using the automatedparaffin tissue processor (ASP300-Leica, Wetzlar, Germany) andcleared with Xylene. The dehydratedsamples were pre-infiltrated withParaplast X-Tra wax at 58°C. Thepre-infiltrated samples were embed-ded in paraffin block wax using anembedding center (EG1160-Leica)and cut into 5 lm thick sectionsusing a microtome (RM2265-Leica).At least five coronal sections wereobtained from each experimentaldefect. Two sections per defect werestained with von Willebrand Factor(vWF) (Millipore’s Blood VesselsStaining Kit, Millipore, Billerica,MA, USA) to assess the blood ves-sels density in the bone defects. ThevWF-stained histological sectionswere recorded using a digital opticalmicroscope (Carl Zeiss Microscopy,Jena, Germany). The total numberof blood vessels stained by vWFwere quantified manually, the aver-age blood vessels density per defectwas obtained, and the data werepresented as blood vessels densityper mm2 mean value � standarddeviation. Another two sections perdefect were stained with TartrateResistance Acid Phosphatase(TRAP) to assess the number of os-teoclasts. The rational for thisassessment is that VEGFs stimulatebone osteoclastic activity (Aldridgeet al. 2005), accordingly anti-VEGFcould have an negative effect bonehealing and implant osseointegrationthrough osteoclast. The histologicalsections were recorded and the totalnumber of osteoclasts was quantifiedusing an imaging software (ZEN2012 SP2, Zeiss, Jena, Germany).The average number of osteoclastsper defect were obtained and thedata were presented as osteoclastnumbers per mm2 of mineralized tis-sue mean value � standard devia-tion.

(a) (b)

(c) (d)

Fig. 1. Surgical Intervention. Rat tibia metaphysis exposed: (a, c) left and right tibiaebefore creating the bone defect; (b) 2.5 mm ø unicortical bone defect on right tibia,(d) 1.5 mm ø x 2.0 mm Ti implant on left tibia.


690 Al Subaie et al.

Assessment of osseointegration

Histology and histomorphometrywere conducted to assess osseointe-gration. All left tibiae (with implants)were dehydrated in ascending con-centrations of ethanol (70–100%)before embedding them in poly-methyl methacrylate histologicalresin (Technovit 9100, Heraeus Kul-zer, Wehrheim, Germany). Afterpolymerization, the osseointegratedimplants were sectioned into histo-logical slides (30 lm thickness) usinga diamond saw (SP1600, LeicaMicrosystems GmbH, Wetzlar, Ger-many) and stained using basic fuch-sine and methylene blue. Due to thesmall size of the implants, only onehistological section through the mid-dle of each implant was obtained andscanned (Carl Zeiss Microscopy).Histomorphometrical measurementswere performed using the ImageJsoftware (Wayne Rasband; NationalInstitute of Health, Bethesda, MD,USA). Implant osseointegration wasdefined as bone-implant contact area(BIC) and was calculated by divid-ing the bone-covered implant perime-ter (BIP) by the total implantperimeter (TIP) as showed in thisequation: BIC = BIP/TIP%. Twoanalyses of BIC were performed;first) BICtotal included the corticaland trabecular peri-implant area and;second) BICtrabecular included the tra-becular peri-implant area only. Allhistomorphometric measurementswere calculated as percentage val-ues � standard deviation.

Statistical analysis

Sample size was calculated toachieve a power of 80% at signifi-cance level of 5% in order to be ableto reject the null hypothesis thatthere is no difference between Rani-bizumab, anti-VEGF and control interms of bone healing and implantosseointegration. A 10% differencebetween groups was considered to beclinically relevant, and a 12% poten-tial standard deviation was assumedbased on a pervious study (Du et al.2009). Accordingly, a total of 11 ratsper group were determined to be suf-ficient. However, one rat was addedto each group to compensate for10% potential dropouts (animal dieout). Descriptive statistics were con-ducted and data distribution was

Fig. 2. Micro-CT Scans. Coronal, sagittal, trans-axial sections and 3-D reconstructionsof the micro-CT scans of bone defects shows larger defects in Ranibizumab and anti-VEGF treated rats compared to the control group. Scale bar = 200 lm.

(a) (b)

(c) (d)

Fig. 3. Micro-CT Analyses. Micro-CT data analysis of bone defects in Ranibizumaband Anti-VEGF treated rats compared to the control group: (a) Defect volume, (b)Trabecular thickness, (c) Trabecular number (d) Trabecular separation. Statisticalanalysis by one-way ANOVA and Welch F test, n control = 11 n Ranibizumab = 11, nAnti-VEGF = 12.



tested for normality using the Kol-mogorov–Smirnov test; the testshowed that all outcomes were nor-mally distributed. The Levene statis-tic test was used to assesshomogeneity of variance. One-wayANOVA followed by the Tukey’s test,were used when the variance washomogenous, which was the case forbone defect volume, blood vesselsdensity, osteoclast number andosseointegration, and data were pre-sented as mean � standard devia-tion. When homogeneity of variancewas rejected, the Welch F test fol-lowed by Games Howell test wereused and data were presented asmean � standard deviation, medianand range values. This was the casefor trabecular number, thickness andseparation. Data analyses were car-ried out using SPSS 17 (SPSS Inc.,Chicago, IL, USA). Statistical signif-icance was set at p < 0.05.

Results

Surgeries proceeded without compli-cations, however, one rat from theRanibizumab group and one from thecontrol group were excluded becauseof post-operative bone fracture. Noother animal drop-out was registered.

Micro-CT: bone healing

The size of the bone defect was sig-nificantly different among groups asindicated by one-way ANOVA

(p = 0.026). Moreover, Tukey’s testindicated that the bone defect vol-ume was significantly larger in theanti-VEGF group (2.48 � 0.33 mm3,p = 0.022) than in the control group(2.11 � 0.36 mm3). However, no sig-nificant differences were observedbetween Ranibizumab (2.35 � 0.23mm3) and control (p = 0.152) as wellas between Ranibizumab and anti-VEGF groups (p = 0.66; Figs 2 and3a).

Trabecular thickness was differ-ent among groups as determined byWelch F test (p = 0.024). GamesHowell test showed that trabeculaewere thinner in the anti-VEGFgroup (0.20 � 0.04 mm, median =0.20 mm, range: 0.19–0.21 mm,p = 0.041) than in the controls(0.288 � 0.096 mm, median = 0.30mm, range: 0.18–0.36 mm). No sig-nificant differences were detectedbetween Ranibizumab (0.23 � 0.05mm, median = 0.24 mm, range:0.20–0.27 mm) and control groups(p = 0.253) as well as between Rani-bizumab and anti-VEGF groups(p = 0.184; Fig. 3b).

Trabecular number was differentamong groups (Welch F test;p < 0.001). The number of trabecu-lae was higher (Games Howell test)in the anti-VEGF group (2.60 � 0.394/mm, median = 2.64/mm, range: 2.34–2.91/mm, p = 0.007) and Rani-bizumab group (2.23 � 0.477/mm,median = 2.06/mm, range: 1.81–2.52/mm, p = 0.035) than in the controlgroup (1.26 � 1.21/mm, median =0.77/mm, range: 0.31–2.29/mm).However, no significant differencewas observed between Rani-bizumab and anti-VEGF (p = 0.073;Fig. 3c).

Trabecular separation was alsodifferent among groups as deter-mined by Welch F test (p < 0.001).Games Howell test indicated that themean trabecular separation waslower in the anti-VEGF group(0.148 � 0.039 mm, median = 0.15mm, range: 0.12–0.18 mm, p =0.013) and the Ranibizumab (0.215� 0.074 mm, median = 0.20 mm,range: 0.14–0.28 mm, p = 0.027)than in the controls (0.6 mm �0.420 mm). No significant differencewas observed between the Rani-bizumab and anti-VEGF groups(0.097; Fig. 3d).

Angiogenesis and osteoclastogenesis

Ranibizumab and anti-VEGF inhib-ited the formation of new blood ves-sels as determined by one-way ANOVA

(p < 0.001). Tukey’s test showed thatblood vessel density was lower in theanti-VEGF group (7.8 � 4.0/mm2,p < 0.001) and Ranibizumab group(11.7 � 4.3/mm2, p < 0.001) than inthe control group (28.0 � 6.5/mm2;Fig. 4). No significant difference wasobserved between anti-VEGF andRanibizumab groups (p = 0.90). Os-teoclasts numbers in the anti-VEGFgroup (28.5 � 11.4 osteoclast/mm2)and Ranibizumab group (21.5 � 11.9osteoclast/mm2) were not signifi-cantly different from the controlgroup (19.2 � 5.2 osteoclast/mm2;one-way ANOVA) p = 0.929 (Fig. 5).

Osseointegration

Bone implant contact area was dif-ferent among groups, the averageBICtotal and BICtrabecular were signifi-cantly different among groups asdetermined by one-way ANOVA

(a) (b)

(d)

(c)

Fig. 4. Blood Vessels Density. Coronal histological sections of tibial bone defects of:(a) control rat, (b) Ranibizumab-treated rat, (c) anti-VEGF-treated rat, red dotted cir-cles indicate blood vessels. Scale bar = 100 lm. (d) Histomorphometric analysis of theblood vessels density. Statistical analysis was done by one-way ANOVA, n control = 11,n Ranibizumab = 11, n Anti- VEGF = 12.



(p < 0.001). Tukey’s test indicatedthat the BICtotal was significantlylower (p < 0.001) in the anti-VEGF(19.9 � 9.4%, p < 0.001) and Rani-bizumab groups (21.7 � 9.2%,p < 0.001) than in the control group(41.8 � 12.4%). Tukey’s test alsoindicated that BICtrabecular in theanti-VEGF (18.6 � 6.1%, p < 0.001)and Ranibizumab groups(19.8 � 4.4%, p < 0.001) were signif-icantly lower than in the controls(35.3 � 10.7%; Fig. 6).

Discussion

This study provides the first evi-dence indicating that anti-VEGFsimpair bone healing and implantosseointegration in rats tibiae. Manyrisk factors could influence osseoin-tegration (Lang & Jepsen 2009,Moy et al. 2004, Misch 2008, Espos-ito et al. 1998, Bornstein et al. 2009,

Heitz-Mayfield & Huynh-Ba 2009,Salvi & Bragger 2009). These riskfactors can be categorized into (i)very high, and; (ii) significant riskfactors. Very high risk factorsinclude serious systemic diseases(rheumatoid arthritis, osteomalacia,osteogenesis imperfect, HIV) andimmunosuppressive medications(corticosteroids, chemotherapy orimmunosuppressive drugs) as well asdrug and alcohol abuse. Significantrisk factors include irradiated bone,severe and uncontrolled diabetes,bleeding disorders, drug-inducedanticoagulation and heavy smoking(Buser et al. 2000). However, arecent systematic review on thistopic concluded that the level of evi-dence indicating absolute and rela-tive contraindications for oralimplant therapy due to systemicconditions and medications is lowand that no data exist for all medi-

cal conditions (Bornstein et al.2009).

Vascular endothelial growth fac-tors can improve bone healing byenhancing angiogenesis (Zhang et al.2014), and by stimulating bone turn-over through osteoclasts chemotaxisand activity (Niida et al., 1999, Nak-agawa et al., 2000, Deckers et al.2000, 2002, Street et al. 2002a,b,Ramazanoglu et al. 2013). Accord-ingly, anti-VEGF could affect bonehealing by suppressing angiogenesisor osteoclasts. However, our datashowed that the compromised bonehealing and osseointegration amongthe anti-VEGF group, as well as thecompromised osseointegration amongthe Ranibizumab group, are likely tobe caused only by the down-regula-tion of angiogenesis since neitherdrugs affected osteoclastic number.

Our findings are in agreementwith previous work that found thatanti-angiogenic drug (TNF-470) hada negative effect on peri-implantbone formation although did not sig-nificantly affect the direct BIC prob-ably because the dose they used. Infact, TNF-470 can only inhibit theblood vessel formation at doses threetimes higher than the one used(30 mg/kg, every other day) whichwas at least three times more thanthe dose (10 mg/kg, three times aweek) Mair’s study (Castronovo &Belotti 1996, Mair et al., 2007). Andhere, we saw significant impairmentof BIC probably because the numberof newly formed blood vessels wasinhibited significantly by Rani-bizumab and anti-VEGF.

In this study, anti-VEGF andRanibizumab did not affect osteo-clast numbers, probably because theanti-bodies that we used have spe-cific affinity to only some VEGFsisoforms (i.e. 164, 165 and 121) butnot to all of them. Moreover, Rani-bizumab which is a fragment of arecombinant humanized monoclonalantibody Fab (48 kDa) is even morespecific and it only inhibits the bio-logic activity of human VEGF-A.Osteoclasts express two types ofVEGF receptors: VEGFR-1 (Fit1-1)and VEGFR-2 (FLk-1). These recep-tors are sensitive to many VEGFisoforms besides those inhibitedby the antibodies that we used inthis study (i.e. 198, 206, beside164, 165, 121 and VEGF-A; Streetet al. 2002a,b). Therefore, specific

(a) (c)(b)

(d)

(g)

(f)(e)

Fig. 5. Osteoclastogenic Activity. Coronal histological sections of tibial bone defectsin: (a and d) control rat; (b and e) Ranibizumab-treated rat; and (c and f) anti-VEGF-treated rat. Black arrows indicate osteoclasts. Scale bar = 100 lm. (g) histomorpho-metric analysis of osteoclast number. Statistical analysis by one-way ANOVA, n con-trol = 11 n Ranibizumab = 11, n Anti-VEGF = 12.



inhibition of VEGF 164, 165, 121 andVEGF-A by anti-bodies we testedmight not be enough to affect osteo-clasts since its receptors will still beactivated by the other spared VEGFisoforms (Ferrara et al. 2003, Mitch-ell et al., 2011, Rosenfeld et al. 2006,Kanczler et al., 2008, Lu et al.,2009, Schmucker et al. 2012).

Anti-VEGF and Ranibizumabalso affected bone quality in the heal-ing defects by reducing trabecularthickness and volume while increas-ing the trabecular number. Theseobservations indicate that inhibitionof angiogenesis seems to hinder tra-becular development and thickening,and this is probably compensated byincreasing the trabecular number.This observation could be explainedby the fact that among the cytokinesand growth factors involved in bonehealing and growth, VEGFs are thekey regulators of angiogenesis (Konet al. 2001, Schliephake 2002, Dimi-triou et al. 2005). Therefore, VEGFinhibition during bone formationwould suppress angiogenesis andconsequently reduces trabecular bone

volume, probably due to deprivationof oxygen and nutrients (Maes et al.2002), but it would have a lessereffect on osteoblast chemotaxis anddifferentiation. Indeed, inhibition ofVEGFs suppresses capillaries inva-sion, reduces chondroclasts recruit-ment and increases the hypertrophiccartilaginous zone in growth plates,without affecting bone mineralizationaround the cartilaginous zone (Ger-ber et al. 1999). These key observa-tions along with our findings suggestthat bone formation in bone defectsof animals treated with anti-VEGFsgets started but it is then interrupteddue to the deprivation of nutrientsand oxygen caused by a deficientblood supply (Gerber et al. 1999,Kanczler & Oreffo 2008).

The average BIC on themachined Ti implants used in thisstudy was 42% among control rats.This BIC value is consistent withprevious studies using similar materi-als, and it is suitable for functionalloading (Wong et al. 1995).However, anti-VEGFs treatmentdecreased BIC almost by half,

reaching levels that are not compati-ble with successful mechanical func-tionalization of implants (Wonget al. 1995). This magnitude of BICreduction is comparable to thatobserved in animal exposed to radia-tion therapy which is a well knowncause of angiogenesis deteriorationand failure of dental implants (Moyet al., 2004, Weinlaender et al. 2006,Kaigler et al. 2006). Accordingly, itcould be speculated that anti-VEGFtherapies could be problematic intreatments with osseointegratedimplants, however, future studies areneeded to confirm this.

The anti-VEGF had strongereffects on bone healing than Rani-bizumab. This observation wasexpected because the anti-VEGFused in this study was rat specific;whereas, Ranibizumab is human spe-cific (Rosenfeld et al. 2006, Loweet al. 2007, Lu & Adelman 2009,Mitchell 2011, Schmucker et al.2012). Nevertheless, Ranibizumabstill had a negative effect on bonehealing and osseointegration in ourstudy, probably due to the similarityin structure between rodent andhuman VEGF (Carmeliet et al.1999). Indeed, Ranibizumab hasbeen found to inhibit angiogenesis inrats (Ostendorf et al. 1999, Arevalo2013).

In order to extrapolate ourresults to different age groups, wechose young growing rats. Youngerrats have faster bone healing (Bak &Andreassen 1989) and should be lesssusceptible to drugs that delay bonehealing (Aguirre et al. 2010).Accordingly, since anti-VEGF andRanibizumab did have a negativeeffect on young animals, we wouldexpect an even stronger effect onolder animals that already havelower serum level of VEGF (Linet al. 2008), although future studieswill have to be performed to confirmthis hypothesis.

As with any animal model, thereare some limitations to be acknowl-edged. These include inherent varia-tion between animals, however, thisissue was compensated by sufficientsample size. Also extrapolation ofour results to clinical practice couldbe limited by potential differencesbetween humans and rats (e.g.bone pattern and bone quality).Moreover, the effects of anti-VEGFson bone healing and implant

(a) (b) (c)

(d) (e)

Fig. 6. Osseointegration %. Sagittal histological sections of Ti implants of tibiae: (a)control rat; (b) Ranibizumab-treated rat; and (c) anti-VEGF-treated rat. Black arrowsindicate newly formed bone. White arrows indicate soft tissue. The anti-VEGF-treatedrats show the least newly formed bone around the implants, followed by Ranibizumaband controls, respectively. Scale bar = 200 lm. (d and e) Histomorphometric analysisof osseointegration. Statistical analysis by one-way ANOVA, n control = 11, n Rani-bizumab = 11, n Anti-VEGF = 12.



osseointegration were assessed onlyat one time point (two weeks follow-ing surgery), thus it did not allow toevaluate the long-term effects ofthese antibodies on bone healing andimplant osseointegration, these issueswill have to be addressed in futurestudies. The 2-week time point wasselected because healing defect neo-angiogenesis and VEGFs expressionare high in this period (Dimitriouet al. 2005). Although this studyshowed that anti-VEGF drugsaffected bone healing and osseointe-gration by compromising blood ves-sels formation, a future knockoutmice model may be needed toconfirm this hypothesis. One morelimitation in this study is that theinjections were given intraperitone-ally instead of intravitreally, how-ever, intravitreal-administered drugs(bevacizumab) do migrate from thevitreous cavity (Bakri et al. 2007)and cause systemic complication(Shima et al. 2008).

Conclusions

Anti-VEGFs inhibit osseointegrationof Ti implants and delay bone heal-ing.

Acknowledgments

The authors acknowledge the finan-cial support from InternationalTeam for Implantology (ITI), theCanadian Institute of HealthResearch (CIHR), the Saudi ArabianCultural Bureau in Canada, the Uni-versity of Dammam, Fonds de rech-erch�e Sant�e (FRQS), the NaturalSciences and Engineering ResearchCouncil of Canada (NSERC) andthe Canada Foundation of Innova-tion (CFI). Moreover, the authorsare grateful to Mr Pierre Rompr�efor assistance in statistical analyses.

References

Aguirre, J. I., Altman, M. K., Vanegas, S. M.,Franz, S. E., Bassit, A. C. F. & Wronski, T. J.(2010) Effects of alendronate on bone healingafter tooth extraction in rats. Oral Diseases 16,674–685.

Albrektsson, T., Br�anemark, P.-I., Hansson, H.-A. & Lindstr€om, J. (1981) Osseointegrated tita-nium implants: requirements for ensuring along-lasting, direct bone-to-implant anchoragein man. Acta Orthopaedica 52, 155–170.

Aldridge, S., Lennard, T., Williams, J. & Birch,M. (2005) Vascular endothelial growth factorreceptors in osteoclast differentiation and

function. Biochemical and Biophysical ResearchCommunications 335, 793–798.

Arevalo, J. F. (2013) Diabetic macular edema:current management 2013. World Journal ofDiabetes 4, 231–233.

Avery, R. L., Pieramici, D. J., Rabena, M. D.,Castellarin, A. A., Nasir, M. A. A. & Giust,M. J. (2006) Intravitreal bevacizumab (Avastin)for neovascular age-related macular degenera-tion. Ophthalmology 113, 363–372.e5.

Bak, B. & Andreassen, T. T. (1989) The effect ofaging on fracture healing in the rat. CalcifiedTissue International 45, 292–297.

Bakri, S. J., Snyder, M. R., Reid, J. M., Pulido,J. S. & Singh, R. J. (2007) Pharmacokinetics ofintravitreal bevacizumab (Avastin). Ophthal-mology 114, 855–859.

Baron, R., Tross, R. & Vignery, A. (1984) Evi-dence of sequential remodeling in rat trabecularbone: morphology, dynamic histomorphometry,and changes during skeletal maturation. TheAnatomical Record 208, 137–145.

Beamer, B., Hettrich, C. & Lane, J. (2010) Vascu-lar endothelial growth factor: an essential com-ponent of angiogenesis and fracture healing.HSS Journal 6, 85–94.

Blanchaert, R. H. (1998) Implants in the medi-cally challenged patient. Dental Clinics of NorthAmerica 42, 35–45.

Bornstein, M. M., Cionca, N. & Mombelli, A.(2009) Systemic conditions and treatments asrisks for implant therapy. International Journalof Oral and Maxillofacial Implants 24 (Suppl.),12–27.

Buser, D., Arx, T., Bruggenkate, C. & Weingart,D. (2000) Basic surgical principles with ITIimplants Note. Clinical Oral Implants Research11, 59–68.

Carmeliet, P. (2005) Angiogenesis in life, diseaseand medicine. Nature 438, 932–936.

Carmeliet, P., Ng, Y.-S., Nuyens, D., Theilmeier,G., Brusselmans, K., Cornelissen, I., Ehler, E.,Kakkar, V. V., Stalmans, I. & Mattot, V.(1999) Impaired myocardial angiogenesis andischemic cardiomyopathy in mice lacking thevascular endothelial growth factor isoformsVEGF164 and VEGF188. Nature Medicine 5,495.

Castronovo, V. & Belotti, D. (1996) TNP-470(AGM-1470): mechanisms of action and earlyclinical development. European Journal of Can-cer 32, 2520–2527.

Davies, J. E. (2003) Understanding peri-implantendosseous healing. Journal of Dental Educa-tion 67, 932–949.

Deckers, M. M., Karperien, M., Van Der Bent,C., Yamashita, T., Papapoulos, S. E. & L€owik,C. W. (2000) Expression of vascular endothelialgrowth factors and their receptors duringosteoblast differentiation. Endocrinology 141,1667–1674.

Deckers, M. M., Van Bezooijen, R. L., Van DerHorst, G., Hoogendam, J., Van Der Bent, C.,Papapoulos, S. E. & L€owik, C. W. (2002) Bonemorphogenetic proteins stimulate angiogenesisthrough osteoblast-derived vascular endothelialgrowth factor A. Endocrinology 143, 1545–1553.

Dimitriou, R., Tsiridis, E. & Giannoudis, P. V.(2005) Current concepts of molecular aspectsof bone healing. Injury 36, 1392–1404.

Du, Z., Chen, J., Yan, F. & Xiao, Y. (2009)Effects of Simvastatin on bone healing aroundtitanium implants in osteoporotic rats. ClinicalOral Implants Research 20, 145–150.

Eriksson, C., Nygren, H. & Ohlson, K. (2004)Implantation of hydrophilic and hydrophobic

titanium discs in rat tibia: cellular reactions onthe surfaces during the first 3 weeks in bone.Biomaterials 25, 4759–4766.

Esposito, M., Hirsch, J. M., Lekholm, U. &Thomsen, P. (1998) Biological factors contrib-uting to failures of osseointegrated oralimplants. (I). Success criteria and epidemiology.European Journal of Oral Sciences 106, 527–551.

Ferrara, N., Gerber, H. & Lecouter, J. (2003) Thebiology of VEGF and its receptors. NatureMedicine 9, 669–676.

Ferrara, N., Hillan, K. J., Gerber, H.-P. & Nov-otny, W. (2004) Discovery and development ofbevacizumab, an anti-VEGF antibody fortreating cancer. Nature Reviews Drug Discovery3, 391–400.

Gerber, H.-P., Vu, T. H., Ryan, A. M., Kowalski,J., Werb, Z. & Ferrara, N. (1999) VEGF cou-ples hypertrophic cartilage remodeling, ossifica-tion and angiogenesis during endochondralbone formation. Nature Medicine 5, 623–628.

Heitz-Mayfield, L. J. & Huynh-Ba, G. (2009) His-tory of treated periodontitis and smoking asrisks for implant therapy. International Journalof Oral and Maxillofacial Implants 24 (Suppl),39–68.

Io, H., Hamada, C., Ro, Y., Ito, Y., Hirahara, I.& Tomino, Y. (2004) Morphologic changes ofperitoneum and expression of VEGF in encap-sulated peritoneal sclerosis rat models. KidneyInternational 65, 1927–1936.

Kaigler, D., Wang, Z., Horger, K., Mooney, D.J. & Krebsbach, P. H. (2006) VEGF scaffoldsenhance angiogenesis and bone regeneration inirradiated osseous defects. Journal of Bone andMineral Research 21, 735–744.

Kanczler, J. & Oreffo, R. (2008) Osteogenesis andangiogenesis: the potential for engineeringbone. Eur Cell Mater 15, 100–114.

Kon, T., Cho, T. J., Aizawa, T., Yamazaki, M.,Nooh, N., Graves, D., Gerstenfeld, L. C. &Einhorn, T. A. (2001) Expression of osteopro-tegerin, receptor activator of NF-jB ligand(Osteoprotegerin Ligand) and related proin-flammatory cytokines during fracture healing.Journal of Bone and Mineral Research 16,1004–1014.

Kvanta, A., Algvere, P., Berglin, L. & Seregard,S. (1996) Subfoveal fibrovascular membranes inage-related macular degeneration express vas-cular endothelial growth factor. InvestigativeOphthalmology & Visual Science 37, 1929–1934.

Lang, N. P. & Jepsen, S. (2009) Implant surfacesand design (Working Group 4). Clinical OralImplants Research 20 (Suppl. 4), 228–231.

Lin, S., Roguin, A., Metzger, Z. & Levin, L.(2008) Vascular endothelial growth factor(VEGF) response to dental trauma: a prelimin-ary study in rats. Dental Traumatology 24,435–438.

Lowe, J., Araujo, J., Yang, J., Reich, M., Olden-dorp, A., Shiu, V., Quarmby, V., Lowman, H.,Lien, S., Gaudreault, J. & Maia, M. (2007) Ra-nibizumab inhibits multiple forms of biologi-cally active vascular endothelial growth factorin vitro and in vivo. Experimental Eye Research85, 425–430.

Lu, F. & Adelman, R. A. (2009) Are intravitrealbevacizumab and ranibizumab effective in a ratmodel of choroidal neovascularization? GraefesArchive for Clinical and Experimental Ophthal-mology 247, 171–177.

Maes, C., Carmeliet, P., Moermans, K., Stock-mans, I., Smets, N., Collen, D., Bouillon, R. &Carmeliet, G. (2002) Impaired angiogenesis andendochondral bone formation in mice lacking



the vascular endothelial growth factor isoformsVEGF164 and VEGF188. Mechanisms of Devel-opment 111, 61–73.

Mair, B., Fuerst, G., Kubitzky, P., Tangl, S.,Bergmeister, H., Losert, U., Watzek, G. &Gruber, R. (2007) The anti-angiogenic sub-stance TNP-470 impairs peri-implant bone for-mation: a pilot study in the rabbit metaphysismodel. Clinical Oral Implants Research 18,370–375.

Miletic, H., Niclou, S. P., Johansson, M. &Bjerkvig, R. (2009) Anti-VEGF therapies formalignant glioma: treatment effects andescape mechanisms. Informa Healthcare 13,455–468.

Misch, C. E. (2008) Contemporary implant den-tistry. St.Louis, Missouri: Mosby Elsevier.

Mitchell, P. (2011) A systematic review of the effi-cacy and safety outcomes of anti-VEGF agentsused for treating neovascular age-related macu-lar degeneration: comparison of ranibizumaband bevacizumab. Current Medical Researchand Opinion 27, 1465–1475.

Moy, P. K., Medina, D., Shetty, V. & Aghaloo,T. (2004) Dental implant failure rates and asso-ciated risk factors. International Journal of Oraland Maxillofacial Implants 20, 569–577.

Nakagawa, M., Toshio, K., Toshiya, A., Shuichi,M., Takuya, S., Takeo, Y., Koji, H.,Masayoshi, K. & Yoshiyuki, H. (2000) Vascu-lar endothelial growth factor (VEGF) directlyenhances osteoclastic bone resorption and sur-vival of mature osteoclasts. FEBS letters 473,161–164.

Niida, S., Masato, K., Hitoshi, A., Hisahiro, Y.,Hiroshi, K., Satomi, N., Kazuo, T., Norihiko,M., Shin-Ichi, N. & Hiroaki, K. (1999) Vascu-lar endothelial growth factor can substitute formacrophage colony-stimulating factor in thesupport of osteoclastic bone resorption. TheJournal of experimental medicine 190, 293–298.

Ostendorf, T., Kunter, U., Eitner, F., Loos, A.,Regele, H., Kerjaschki, D., Dwight, H., Neb-ojsa, J. & Floege, J. (1999) VEGF165 mediatesglomerular endothelial repair. The Journal ofClinical Investigation 104, 913–923.

Pearce, A., Richards, R., Milz, S., Schneider, E.& Pearce, S. (2007) Animal models for implantbiomaterial research in bone: a review. Eur CellMater 13, 1–10.

Ramazanoglu, M., Lutz, R., Rusche, P., Trabzon,L., Kose, G. T., Prechtl, C. & Schlegel, K. A.(2013) Bone response to biomimetic implantsdelivering BMP-2 and VEGF: an immunohisto-chemical study. Journal of Cranio-Maxillo-Facial Surgery 41, 826–835.

Ribatti, D. (2005) The crucial role of vascularpermeability factor/vascular endothelialgrowth factor in angiogenesis: a historicalreview. British Journal of Haematology 128,303–309.

Rocher, N., Behar-Cohen, F., Pournaras, J.,Naud, M., Jeanny, J., Jonet, L. & Bourges, J.(2011) Effects of rat anti-VEGF antibody in arat model of corneal graft rejection by topicaland subconjunctival routes. Molecular Vision17, 104–112.

Rosenfeld, P. J., Brown, D. M., Heier, J. S., Bo-yer, D. S., Kaiser, P. K., Chung, C. Y. & Kim,R. Y. (2006) Ranibizumab for neovascular age-related macular degeneration. New EnglandJournal of Medicine 355, 1419–1431.

Roskoski, J. R. (2007) Sunitinib: a VEGF andPDGF receptor protein kinase and angiogene-sis inhibitor. Biochemical and BiophysicalResearch Communications 356, 323–328.

Salvi, G. E. & Bragger, U. (2009) Mechanical andtechnical risks in implant therapy. InternationalJournal of Oral and Maxillofacial Implants 24(Suppl), 69–85.

Schliephake, H. (2002) Bone growth factors inmaxillofacial skeletal reconstruction. Interna-tional Journal of Oral and Maxillofacial Surgery31, 469–484.

Schmucker, C., Ehlken, C., Agostini, H. T.,Antes, G., Ruecker, G., Lelgemann, M. &Loke, Y. K. (2012) A safety review and meta-analyses of bevacizumab and ranibizumab: off-label versus goldstandard. PLoS ONE 7,e42701.

Scully, C., Hobkirk, J. & Dios, P. D. (2007) Den-tal endosseous implants in the medically com-promised patient. Journal of Oral Rehabilitation34, 590–599.

Shima, C., Sakaguchi, H., Gomi, F., Kamei, M.,Ikuno, Y., Oshima, Y., Sawa, M., Tsujikawa,M., Kusaka, S. & Tano, Y. (2008) Complica-tions in patients after intravitreal injection ofbevacizumab. Acta Ophthalmologica 86, 372–376.

Stadlinger, B., Pourmand, P., Locher, M. &Schulz, M. (2012) Systematic review of animalmodels for the study of implant integration,assessing the influence of material, surface anddesign. Journal of Clinical Periodontology 39(s12), 28–36.

Street, J., Bao, M., Deguzman, L., Bunting, S.,Peale, F. V. Jr, Ferrara, N., Steinmetz, H.,Hoeffel, J., Cleland, J. L., Daugherty, A., VanBruggen, N., Redmond, H. P., Carano, R. A.& Filvaroff, E. H. (2002) Vascular endothelialgrowth factor stimulates bone repair by pro-

moting angiogenesis and bone turnover. Pro-ceedings of the National Academy of SciencesUSA 99, 9656–9661.

Turner, A. S. (2001) Animal models of osteoporo-sis—necessity and limitations. Eur Cell Mater1, 13.

Weinlaender, M., Beumer, J. 3rd, Kenney, E. B.,Lekovic, V., Holmes, R., Moy, P. K. & Plenk,H. Jr (2006) Histomorphometric and fluores-cence microscopic evaluation of interfacialbone healing around 3 different dental implantsbefore and after radiation therapy. Interna-tional Journal of Oral and MaxillofacialImplants 21, 212–224.

Wong, M., Eulenberger, J., Schenk, R. & Hunzi-ker, E. (1995) Effect of surface topology on theosseointegration of implant materials in trabec-ular bone. Journal of Biomedical MaterialsResearch 29, 1567–1575.

Zhang, L., Zhang, L., Lan, X., Xu, M., Mao, Z.,Lv, H., Yao, Q. & Tang, P. (2014) Improve-ment in angiogenesis and osteogenesis withmodified cannulated screws combined withVEGF/PLGA/fibrin glue in femoral neck frac-tures. Journal of Materials Science. Materials inMedicine 25, 1165–1172.

Address:Elham EmamiFaculty of DentistrySchool of Public HeathDepartment of Social and PreventiveMedicineUniversit�e de Montr�ealMontr�eal, CanadaPavillon Roger-Gaudrylocal C-2192900, �Edouard-MontpetitMontr�eal, QC H3T 1J4CanadaE-mail: [email protected] TamimiFaculty of DentistryRoom M60C, Strathcona Anatomy & Dent3640 University StreetMontrealQuebecH3A 0C7CanadaE-mail: [email protected]

Clinical Relevance

Scientific rationale for the study:Many ophthalmology and oncol-ogy patients are treated with anti-body therapies against vascularendothelial growth factors. Thesegrowth factors are essential forbone healing and osseointegration,accordingly drugs that hinder thieractivity such as Ranibizumab andanti-VEGF, might have a negative

effect on bone healing and implantosseointegration.Principal findings: Drugs that hinderthe activity of vascular endothelialgrowth factors, such as anti-VEGFsdelayed bone healing and compro-mised implant osseointegration inrats’ tibiae.Practical Implications: Anti-bodiesthat used to treat age-related macu-lar degeneration of the eye and can-

cer might expose patients to higherrisk of implant and bone surgerycomplications. Therefore, thesemedications should be consideredas potential risk factors for implantosseointegration and bone healingin dental and orthopedic surgicalinterventions.



Anti‐VEGFs hinder bone healing and implant ... · Ranibizumab on bone healing and implant...

Documents

Transcript of Anti‐VEGFs hinder bone healing and implant ... · Ranibizumab on bone healing and implant...