lable at ScienceDirect

ARTICLE IN PRESS

Biomaterials xxx (2010) 1–9

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

Biocompatibility and safety of a hybrid core–shell nanoparticulate OP-1 deliverysystem intramuscularly administered in rats

Ziyad S. Haidar a,b,c,d,e,1, Reggie C. Hamdy c,d,e,2, Maryam Tabrizian a,b,c,d,*

a Faculty of Dentistry, McGill University, Montreal (QC), H3A 2B4 Canadab Department of Biomedical Engineering, Faculty of Medicine, McGill University, Montreal (QC), Canadac Center for Biorecognition and Biosensors, McGill University, Montreal (QC), Canadad Centre for Bone and Periodontal Diseases Research, McGill University, Montreal (QC), Canadae Shriners Hospital for Children and Division of Orthopaedic Surgery, Montreal Children’s Hospital, McGill University, Montreal (QC), Canada

a r t i c l e i n f o

Article history:Received 30 September 2009Accepted 13 December 2009Available online xxx

Keywords:BiocompatibilityBMPDrug deliveryForeign body responseHaemocompatibilityNanoparticle

* Corresponding author. Faculty of Dentistry, McGiStreet, Montreal (QC), H3A 2B4 Canada. Tel.: þ1 514 39

E-mail address: maryam.tabrizian@mcgill.ca (M. T1 740 Rue Dr. Penfield, Lab # 4300, Montreal (QC) H

398 3469; fax: þ1 514 398 7469. ziyad.haidar@mcgill2 1529 Cedar Avenue, Montreal (QC) H3A 1A6 Canadþ1 514 842 7553. rhamdy@shriners.mcgill.ca.

0142-9612/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.biomaterials.2009.12.034

Please cite this article in press as: HaidarBiomaterials (2010), doi:10.1016/j.biomateri

a b s t r a c t

A hybrid, localized and release-controlled delivery system for bone growth factors consisting of a lipo-somal core incorporated into a shell of alternating layer-by-layer self-assembled natural polyelectrolyteshas been formulated. Hydrophilic, monodisperse, spherical and stable cationic nanoparticles (�350 nm)with an extended shelf-life resulted. Cytocompatibility was previously assayed with MC3T3-E1.4 mousepreosteoblasts showing no adverse effects on cell viability. In this study, the in vivo biocompatibility ofunloaded and loaded nanoparticles with osteogenic protein-1 or OP-1 was investigated. Young maleWistar rats were injected intramuscularly and monitored over a period of 10 weeks for signs ofinflammation and/or adverse reactions. Blood samples (600 mL/collection) were withdrawn followed byhematological and biochemical analysis. Body weight changes over the treatment period were noted.Major organs were harvested, weighed and examined histologically for any pathological changes. Finally,the injection site was identified and examined immunohistochemically. Overall, all animals showed noobvious toxic health effects, immune responses and/or change in organ functions. This hybrid core–shellnanoparticulate delivery system localizes the effect of the released bioactive load within the site ofinjection in muscle with no significant tissue distress. Hence, a safe and promising carrier for therapeuticgrowth factors and possibly other biomolecules is presented.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Bone morphogenetic proteins (BMPs) are cytokines that are ableto induce new bone formation in vitro and in vivo [1]. BMP-7 (alsoknown as osteogenic protein-1 or OP-1) has been shown to accel-erate the formation of new bone in numerous pre-clinical andclinical studies [2]. Nonetheless, the clinical efficacy of OP-1 stilldepends on the carrier or delivery system used to ensure a sus-tained, multi-step, and prolonged delivery of adequate proteinconcentrations to the desired site of tissue repair or restoration[3,4]. The foremost limitations include the rapid diffusion of OP-1away from the site of application and loss of its bioactivity, resulting

ll University, 3775 University8 8129; fax: þ1 514 398 7461.abrizian).3A 1A4 Canada. Tel.: þ1 514

.ca.a. Tel.: þ1 514 842 4464; fax:

All rights reserved.

ZS, et al., Biocompatibilityals.2009.12.034

in orthotopic/heterotopic bone formation or suboptimal localinduction and hence failure of bone regeneration. Consequently,supra-physiological, unsafe and expensive dosages of OP-1 in themilligram range for satisfactory bone healing continue to berequired [5]. Liposomes are the commonly investigated vehicles fordelivery of therapeutic compounds, such as enzymes and proteins[6,7] because of their biocompatibility and appealing ability tocarry hydrophobic and hydrophilic drugs. Nonetheless, stability invivo remains a setback, mainly due to their high tendency todegrade or aggregate leading to leakage of the entrapped drugduring storage or after administration with considerable toxiceffects. Additionally, they are rapidly cleared from circulation viauptake by the reticulo-endothelial system [8]. To overcome suchproblems, modifying the surface by means of coating the liposomewith a single layer of hydrophilic polymers has been investigated[8–10]. In an initial in vitro study [10], we formulated monodispersenanoparticles (NPs) constituting a core of cationic liposomes(L) and a shell constructed through the layer-by-layer (l-b-l) self-assembly of alternating layers of naturally-available anionicalginate (AL) and cationic chitosan (CH). The NPs had a cumulative

and safety of a hybrid core–shell nanoparticulate OP-1 delivery...,

Z.S. Haidar et al. / Biomaterials xxx (2010) 1–92

ARTICLE IN PRESS

size of 383 � 11.5 nm and a Zeta (z) potential surface charge of44.61 � 3.31 mV, suitable for complex formation with anionicproteins such as OP-1 [11]. The choice of reducing particle size ofbiomaterials from the microscale to the nanoscale is mainly toimprove the bioavailability of the encapsulated drug once admin-istered in situ and hence allowing for the effective use of muchlower and safer dosages [12]. In a subsequent work [11], the cyto-biocompatibility (effect on MC3T3-E1.4 preosteoblast cell viability)and capability of the hybrid core–shell NPs to encapsulate a rangeof concentrations of bioactive OP-1 for their potential administra-tion via a parental injection was investigated. The system exhibitedhigh physical stability in simulated physiological media allowingfor immediate protein loading prior to administration, thuspreventing degradation or loss of the entrapped growth factor. Asustained tri-phasic linear release of the water-soluble and readilydiffusible positively-charged OP-1 was evident for an extendedperiod of 45 days with the bioactivity of the protein maintained viapromoting preosteoblast differentiation with no evident cytotoxiceffects (96% cell viability). In vivo, the localized and release-controlled potential of the delivery system were then demonstratedin rabbits where enhanced new bone regeneration resultedfollowing a single injection of the NPs loaded with doses as low as1.0 mg OP-1 [13]. This established the efficiency as well as thepotential biocompatibility and safety of the delivery system whereno adverse effects or behavioral changes were noted in all animalsover a period of w21 days. Besides the known advantages includingthe size property, longer shelf-life, favorable preparation methodsand subsequent ability to entrap more bioactive drugs [14], nano-sized delivery systems have the advantage of residing longer incirculation when compared to microparticles [15], raising safetyconcerns where acidic by-products, foreign body responses andheterotopic bone formation in undesirable tissues have beenreported; especially with injectable nanoparticulate deliverysystems [16,17].

In this study, we further evaluate the biocompatibility and safetyof the injectable nanoparticulate OP-1 delivery system in healthyrats to investigate any potential hazards and/or adverse effects. Thecommonly-measured blood markers, clinical signs and majororgans (and site of injection in quadriceps muscle) were examinedfor any pathological abnormalities.

2. Materials and methods

2.1. Formulation of hybrid nanoparticles and evaluation of OP-1 encapsulation

The formulation and characterization of the hybrid core–shell nanoparticulateprotein delivery system have been previously described [10,11]. Briefly, a lipid phasewas prepared via thin-film hydration by dissolving 1, 2-Dipalmitoyl-sn-glycero-3-phosphocholine (Genzyme Pharmaceuticals, Switzerland), cholesterol and a cationicsurfactant; dimethyldioctadecyl-ammonium bromide (DDAB) obtained fromSigma–Aldrich Chemical in a chloroform–methanol (Fisher Scientific) mixture (4:1,v/v). DDAB was used in a 4% molar concentration to tailor the surface charge of thephospholipids. The solvent mixture was then removed from the lipid phase byrotary evaporation under vacuum resulting in the deposition of a homogenous drylipid film. The film was hydrated with 18.2 MU cm�1 highly-pure water (HPW),vortexed to obtain a suspension of positively-charged multi-lamellar vesicles andtransferred into a mini extruder (Avanti� Polar Lipids, Inc.) for filtration throughdouble 200 nm pore size 19 mm polycarbonate filters (GE Osmonics). For the l-b-lshell coating, fresh AL (sodium salt; 2% viscosity) and CH (85% deacetylated withmolecular weight of 91.11 kDa) obtained from Sigma–Aldrich Chemical wereprepared in HPW (1 mg/mL). The final pH of the CH solution was then adjusted with1 M NaOH to 5.5. The cationic liposomes (L) were coated with alternating layers ofanionic AL and cationic CH until the desired number of polyelectrolyte layers wasachieved; 6 layers: L(AL-CH)3, herein and after denoted as NPs. With the depositionof each polymeric layer, the solution was incubated at room temperature for 60 minand centrifuged at 1600g for 15 min for washing. Prior to protein loading, aliquots ofnanoparticle suspensions were freeze-dried in the presence of 10% w/w sucroseserving as a cryoprotectant at�54 �C for 48 h (Modulyo D-115, Thermo Savant, MA).The recombinant human (rh) osteogenic protein-1/rhOP-1 (15.7 kDa molecularweight, lyophilized) was purchased from bio-WORLD, OH and stored at �20 �C until

Please cite this article in press as: Haidar ZS, et al., BiocompatibilityBiomaterials (2010), doi:10.1016/j.biomaterials.2009.12.034

use, according to the manufacturer’s instructions. Lyophilized NPs were rehydratedback to the original volume with different concentrations of OP-1 solution(0.0–1.0 mg/mL in 0.3 mL HPW). Average size and surface charge changes uponloading were then determined using low-angle laser light-scattering (DLS-HPPS,Malvern Instruments, UK) and laser Doppler anemometry (Zeta-Plus, BrookhavenInstruments, NY), respectively.

2.2. Animals

A total of 22 young and healthy (10–15 weeks old, 350–400 g at the start of theacclimatization period) male normal Wistar rats (Harlan Sprague Dawley�) wereincluded in this study. The animals were housed individually in type III Macroloncages under conventional hygienic conditions, at 20–24 �C and 30–70% relativehumidity and with natural day/night light rhythm. Rats were fed a pelleted diet,allowed access to tap water ad libitum and inspected daily by fit animal health carepersonnel. The housing, care and experimental protocol were approved by theMcGill University Animal Care and Ethics Committee.

2.3. Experimental study design and analyses

2.3.1. Study protocolAll animals were randomized to each receive a single injection (0.3 mL total

volume) in the right quadriceps muscle using latex-free micro-fine� IM syringes(25G½ 0.36 mm� 13 mm, Becton Dickinson and Co., NJ) as follows: A. Control group1: saline (n ¼ 2); B. Control group 2: 0.5 mg OP-1 (n ¼ 2); C. Control group 3: 1.0 mgOP-1 (n ¼ 2); D. Experimental group 1: blank/unloaded NPs in highly-pure water(n ¼ 4); E. Experimental group 2: NPs loaded with 0.5 mg OP-1 (n ¼ 6); and F.Experimental group 3: NPs loaded with 1.0 mg OP-1 (n ¼ 6). No anesthesia wasrequired. Rats were examined daily for signs of infection, inflammation and adverseeffects by visual observation. Body weight changes were measured and recordedover a period of 10 weeks. All animals were euthanized by carbon dioxide either ond28 (n ¼ 11; 50%) or d70 (n ¼ 11; 50%).

2.3.2. Serum biochemical and hematological analysisBlood samples were collected (6 mL/kg/3 weeks: 600 mL per collection) on day

0 (baseline: pre-injections) and post-injections on days, 1, 7, 14, 28, 56 and 70. Usinga biochemical autoanalyzer (VITALAB, Merck, Netherlands), serum biochemicalanalysis was carried out to determine the serum level of total protein, albumin, totalbilirubin, aspartate transaminase (AST), alanine transaminase (ALT), alkaline phos-phatase (ALP), glucose, cholesterol, triglyceride, blood urea nitrogen (BUN), lacticacid dehydrogenase, creatinine, sodium, potassium, chloride, calcium, phosphorusand magnesium. Hematological parameters consisting of erythrocytes, leukocytes,neutrophils, lymphocytes, monocytes, eosinophils, basophils, hemoglobin, hemat-ocrit and platelets were determined using a hematological analyzer (Coulter T540hematology system, Fullerton, CA).

2.3.3. Histopathalogical analysisThe brain, liver, lungs, kidneys, heart, spleen and right quadriceps muscle (site of

injection) were removed from each animal at time of euthanasia (50% on d28 andthe other 50% on d70), weighed, sectioned and then immersed-fixed in a buffered(0.4 M phosphate buffer, pH 7.61) 4% paraformaldehyde solution for at least 24–48 h.Tissue sections (3 mm) were prepared after dehydration and embedded in paraffin.Of these, random samples were stained with hematoxylin and eosin (H&E) andprocessed for comparative histopathological examination under a light microscopyby a qualified and blinded veterinary pathologist to eliminate any bias.

2.3.4. ImmunohistochemistryThe specimens of the right quadriceps muscle assigned for immunohisto-

chemistry were sectioned, fixed in 4% paraformaldehyde overnight, decalcified in20% ethylenediamine tetra-acetic acid for 3 weeks, and embedded in methylmethacrylate or MMA. Seven micrometer sections of a random selection of blocksfrom every experimental group were then cut. After de-paraffinization and hydra-tion, endogenous peroxidase was blocked with 3% hydrogen peroxide for 10 min.Non-specific binding was blocked by incubation in phosphate-buffered salinecontaining 10% normal horse serum and Triton X-100 for 20 min. For immunos-taining, commercially available polyclonal goat antibodies (Santa Cruz Biotech-nology Inc., Santa Cruz, CA) were used for the qualitative verification of the presenceof genes involved in the BMP signaling pathway during bone formation. Those werecategorized according to ligands (BMP-2, -3 and -7), receptors (BMPR-I, BMPR-IIA,BMPR-IIB), transcription factors (Smads 1–5 and Sox-9), differentiation marker(Collagen-II) and antagonists (BMP-3 or osteogenin and Noggin). Sections wereincubated with these primary antibodies at a dilution of 1:100 in phosphate-buff-ered saline with 1% normal horse serum. Overnight incubation at 4 �C in a humidi-fied chamber followed. As a secondary antibody, a biotinylated horse anti-goatantibody (Vector Labs, Burlingame, CA) at a dilution of 1:400 was used. Sectionswere stained using the avidin–biotin complex method along with 3,30-dia-minobenzidine tetrachloride for 30 min, followed by DAB-peroxidase revelation.Finally, the sections were counterstained with Goldner Trichrome, mounted withPermount, imaged and semi-quantified as described below under optical

and safety of a hybrid core–shell nanoparticulate OP-1 delivery...,

Z.S. Haidar et al. / Biomaterials xxx (2010) 1–9 3

ARTICLE IN PRESS

microscopy. Negative controls were prepared similarly however excluding theprimary antibody.

2.3.5. Quantitative and quantitative grading of the immunostained sectionsA semi-quantitative analytical method to describe immunohistochemistry

images was previously developed [13,18–20]. The number of cells expressing theproteins was assessed by stained cell counting and graded by a blinded observer asfollows: � no staining; þ/� staining in less than 25% of cells; þ staining in 25% ofcells; þþ staining in 25–50% of cells; þþþ staining in 50–75% of cells; þþþþstaining in more than 75% of cells. Chondrocytes (C) and fibroblastic (F) cells wereidentified morphologically when detected. It is noteworthy that this set of immu-nohistochemical preparation/analysis on samples of the harvested muscle was notprimarily planned as this study does not aim to evaluate the potential of the NPs ininducing ectopic bone formation; especially given the difficulty in identifying thesite of injections in muscle.

2.4. Statistical analysis

Findings are expressed as means � standard error (SE) of the mean. Forcomparisons among treatment groups, a 2-tail Student’s t-test was used withdifferences deemed significant at p-values less than 0.05. Immunohistochemistrydata are represented as an average score from blind observations.

3. Results and discussion

Combining the advantages of phospholipid vesicles with thoseof l-b-l nano-assembled systems, we introduced previously inBiomaterials a novel hydrophilic protein delivery system consistingof a suspension of core–shell NPs [10]. Fig. 1 displays a cryo-TEM(cryo-Transmission Electron Microscopy) image of the NPsdemonstrating their spherical morphology. We later reported thecytocompatibility (96% cellular viability), encapsulation efficiency(80% of loaded protein) and drug release kinetics (linear and sus-tained release over 45 days period) demonstrating the favorablepotential of the NPs in delivering bioactive growth factors such asosteogenic protein-1 or OP-1 [10,11,13]. To further investigate thebiocompatibility of the NPs (unloaded and OP-1 loaded) ina smaller (than rabbits) and more frequently employed animalspecies in such studies, timely blood and organ function analysis

Fig. 1. Cryo-Transmission Electron Microscopy (Cryo-TEM) of coated liposomes form-ing the nanoparticles. Bar ¼ 100 nm. The polymeric shell coating the liposomal coremaintains the spherical morphology and supports the stability of the nanoparticlesfollowing the aggressive preparatory steps prior to imaging.

Please cite this article in press as: Haidar ZS, et al., BiocompatibilityBiomaterials (2010), doi:10.1016/j.biomaterials.2009.12.034

over a period of 70 days were performed. Table 1 summarizes thechange in average size and Zeta (z) potential surface charge beforeand after freeze-drying and rehydration/loading of the NPs withbioactive OP-1; prior to their administration into the rats.

3.1. Clinical signs and body weight changes

Throughout the study period, animals showed no apparent signsof dehydration, muscle loss or anorexia; symptoms associated withanimal toxicity [21]. Body weight loss, particularly when in excessof 10–20% is also indicative of toxicity. All animals irrespective ofthe injection received did not show any such adverse effects ontheir growth as evident in their normal body weight gain (Fig. 2)observed over the experimental time span of 10 weeks with nostatistically significant differences (p > 0.05) detected amonggroups including control rats. Further, no abnormal physical signsand/or behaviors in any groups were reported by the animal healthcare technicians supervising the animals.

3.2. Serum biochemical and hematological findings

Timely complete blood count and total panel serum biochemicalanalysis for various cytokines and enzymes (a total of 50 parame-ters) were performed for the collected blood samples in order toassess any potential inflammatory responses and/or organ injuryfollowing the intramuscular administration of saline (for thecumulative physiological and psychological/behavioral effect ofneedle insertion), free OP-1 solution prepared in HPW (0.5 and1.0 mg), unloaded and loaded NPs with 0.5 and 1.0 mg OP-1. None ofthe measured parameters showed definitive changes by theexperimental time points signifying the absence of inflammatoryresponse(s) or organ toxicity and dysfunction post-injections. Inaddition, hematological analysis demonstrated consistency in theresults where no alarming variations in the normal values of any ofthe evaluated key assessors of biocompatibility such as ALP levelsand counts of red blood cells, white blood cells and platelets aroseas shown in Fig. 3. ALT, mostly in red blood cells provides infor-mation about the status of the heart and liver. A level increase ordecrease may indicate hepatitis, ischemic cirrhosis or the presenceof a liver tumor, for instance. On the other hand, ALP enzymes arenormally present regularly in the bile ducts, liver and bone;therefore alterations will often suggest acute liver damage or bonefracture. Moreover, an elevation in BUN levels indicates kidneyand cardiac damage [22,23]. Other blood parameters that areknown to indicate an immune response or reflect changes inorgans function were also unchanged in the animals receiving thecore–shell NPs whether unloaded or loaded with OP-1 (p > 0.05).Similarly, no trends or indications of change in enzymatic activityreflective of inflammatory response or organ toxicity and failurewere noted over the extended period of 70 days. This can beattributed to the localized characteristics of the nanoparticlesconfining their encapsulant effect to the site of injection, hereby inmuscle. Growth factors such as BMPs are known to be dosage-dependent [1,4,24] and large supra-physiological concentrations

Table 1Summary of the characterization parametric data measured before and after rehy-dration and loading the nanoparticles with OP-1. Values are reported as mean (*)and standard error (SE) of the mean for the average hydrodynamic diameter (nm)and Zeta potential surface charge (mV).

Pre-loading with OP-1 Post-loading with OP-1

Size (nm)* SE Zeta (mV)* SE Size (nm)* SE Zeta (mV)* SE

345 10.9 þ37.88 1.3 386 9.4 þ38.77 1.5

and safety of a hybrid core–shell nanoparticulate OP-1 delivery...,

Fig. 2. Average animal body weight changes (in g) pre- and post-IM injections for all animals up to day 70 (where 50% of animals per group were sacrificed on day 28). The datashow a consistent body growth with no significant differences among groups (p > 0.05) including control rats. All animals continued in good health with normal food/waterconsumption and energy intake until time of sacrifice and organ harvest.

Z.S. Haidar et al. / Biomaterials xxx (2010) 1–94

ARTICLE IN PRESS

often in the milligram range continue to be required to obtain aneffect [5,13], as mentioned earlier. For example, when ‘low’ dosages(2 mg) of BMP-2 were locally delivered to the dorsum of rats frombiphasic calcium phosphate granules, significant molecularchanges were not induced while ‘higher’ (50 mg) dosages did [25];mainly due to the bolus and initial burst release, rapid protein

Fig. 3. Selective representation of the most important biochemical and hematological paramcore–shell nanoparticulate system. No significant differences detected. No alarming variationsuch as alanine transaminase (ALT), alkaline phosphatase (ALP), blood urea nitrogen (BUN) anto include the largest number of animals. No significant differences were noted on days 56

Please cite this article in press as: Haidar ZS, et al., BiocompatibilityBiomaterials (2010), doi:10.1016/j.biomaterials.2009.12.034

dispersal away from intended site and/or loss of the bioactivity(given the short half-lives) especially when administered into morefluid environments requiring slower release rates [1–5,13]. Thus,this can also elucidate why no biochemical or hematologicalchanges were detected in our animals receiving pure OP-1, takinginto consideration that the dosages used herein were very low in

eters analyzed from the collected blood from the rats demonstrating the safety of thes in normal values of key biocompatibility assessors indicative of major organ functiond counts of white blood cells (WBCs) were detected. Data is shown up to day 28 so thatand 70 for the rats that continued to the end of the study.

and safety of a hybrid core–shell nanoparticulate OP-1 delivery...,

Table 2Summary of the measured organ weights on animal sacrifice and harvest (in g) on experimental day 70 (n ¼ 11). The range is based on information from the animal supplierand others from the literature. No significant differences were detected among any of the groups indicating no pathological findings.

Organ Weight range (g) Weight mean � SD (g)

Saline 0.5 mg OP-1 1.0 mg OP-1 NPs þ 0.0 mg OP-1 NPs þ 0.5 mg OP-1 NPs þ 1.0 mg OP-1

n ¼ 1 n ¼ 1 n ¼ 1 n ¼ 2 n ¼ 3 n ¼ 3

Brain 1.75–1.83 1.79 1.76 1.77 1.77 � 0.024 1.78 � 0.026 1.80 � 0.022Heart 0.48–0.54 0.50 0.52 0.54 0.52 � 0.039 0.5 � 0.049 0.49 � 0.010Liver 9.35–11.04 9.71 10.21 10.97 9.80 � 0.452 10.79 � 0.402 10.56 � 0.748Lung 0.96–1.99 1.21 1.34 0.99 1.40 � 0.424 1.58 � 0.434 1.36 � 0.397Spleen 0.50–0.734 0.58 0.66 0.72 0.74 � 0.212 0.79 � 0.124 0.92 � 0.061Kidney 1.851–2.267 1.88 1.98 2.13 2.05 � 0.071 2.12 � 0.141 1.99 � 0.275

Z.S. Haidar et al. / Biomaterials xxx (2010) 1–9 5

ARTICLE IN PRESS

comparison; 0.1 mg and 0.5 mg only. Besides, although host-response to foreign materials was reported to be more intensive inperitoneum-supported organs than in subcutaneous sites [17] forinstance, it appears that our core–shell NPs, if reached, were rapidlycleared from the un-targeted tissues and organs such as the liverand kidney. This might be either due to their small size or morepossibly so because of their naturally-biocompatible compositionwhere the rapid degradation of such polymers (and any ensuing by-products are neutralized within the body) was shown to bea contributing factor in numerous in vivo studies [17,26,27].

3.3. Organ weights

Organs were harvested on d28 and d70. Organ weights wereconsistent among all rats at both endpoints. Table 2 summarizesthese values (in g) for the organs harvested at day 70. These resultswere compared to other works on toxicity of different nanoparticleformulations in similar animal strains [12,16,26] where no trends orindications of any pathological changes were noted.

3.4. Histopathological findings

Post-mortem examinations at all time points revealed no visiblesigns of internal inflammation. The systemic distribution and fate ofthe different formulations were evaluated via studying the majororgans (brain, heart, liver, kidneys, lungs, spleen as well as theinjection site in muscle) histologically and histomorphometricallyunder a light optical microscope. Data from such studies wouldprovide pharmacokinetic and biokinetic descriptions of our

Fig. 4. Selective histology (H&E staining) images of sections of some of the harvested organsthe lungs of groups 2 and 3 than the other groups. Rare inflammation in the liver was notedrats given exposure to environmental changes. Site of injection in the right quadriceps muscells shown are endogenous muscular regenerative cells following needle insertion.

Please cite this article in press as: Haidar ZS, et al., BiocompatibilityBiomaterials (2010), doi:10.1016/j.biomaterials.2009.12.034

nanoparticulate protein delivery system essential for the possibleextrapolation of data from animals to humans for potential appli-cations in bone tissue engineering. The performed comparativepathology in this study focused on identifying any abnormal tissuechanges in the organs harvested such as bone metaplasia and isdisplayed in Fig. 4. Osteoid formation was detected in the lungs ofgroups 2 through 3 and ranged from rare (5–10%) to moderate(25%–50%). However, in groups 4 through 6, no such observationswere evident except in one animal (1/6) from group 5 sacrificed atthe 70 days time point and with a diminutive mass. Multifocalinflammation and lymphoid hyperplasia was also observed in thelungs of most animals including controls and therefore are non-specific to one group or another. Similarly, alveolar histiocytosiswas observed however seems to be a common finding in this strainof rats, according to Sells et al. [28]. In the heart, cardiomyocytedegeneration with mononuclear inflammatory infiltrate was notedas rare (5–10%) to mild (5–20%) as was tubular degeneration andregeneration in the kidneys (no calcifications detected), multifocalmixed inflammatory infiltrate in the liver and lymphoid hyper-plasia (white pulp) in the spleen. These observations were commonin the control rats and were also reported to be a classic anddistinctive-pathological finding in Wistar rats, perhaps environ-mentally-related [29]. In the harvested brains of the animals, nochanges or lesions in the studied cerebrums, cerebellums or brainstems were detected whatsoever. Finally, the H&E stained histo-pathological analysis performed for the initial collection of randomsamples of the harvested hind leg muscles of the rats revealed thepresence of mixed lymphohistiocytic inflammatory infiltratesassociated with skeletal muscle degeneration and regeneration

(magnification ranges from 20� to 400�) on d70. Osteoid formation is more evident inhowever is consistent in all animals and has been reported to be normal in this strain ofcle can be identified in the displayed images with no evident pathologies. The pools of

and safety of a hybrid core–shell nanoparticulate OP-1 delivery...,

Z.S. Haidar et al. / Biomaterials xxx (2010) 1–96

ARTICLE IN PRESS

most likely indicating the site of the injections and the associatedinjury/tear upon needle insertion. This is an expected inflammatoryresponse normally occurring within 2–3 weeks triggered by injuredvascularized connective tissue [17,30].

3.5. Molecular changes in muscle for esthetic purposes

3.5.1. Histology for site of injectionSince the pioneering work of Marshall R. Urist in 1965 intro-

ducing the concept of bone formation by auto- or osteoinduction[3,31], several pre-clinical and clinical studies have then continuedto report caution when using local growth factor application in thetreatment of bone-related conditions explicitly as a result of ectopic

Fig. 5. Selective histology (Goldner Trichrome staining) and immunohistochemistry imageswith 0.5 mg OP-1 representative of findings in animals of groups 5 and 6 and demonstrating tThe top 2 rows (histology) show the injection site and pools of osteoblastic cells (top) and tdetected in situ; suggested by the presence of healthy nerve and blood vessels in the injecbottom 2 rows (immunohistochemistry) shows expression of different ligands, their receptorbone formation.

Please cite this article in press as: Haidar ZS, et al., BiocompatibilityBiomaterials (2010), doi:10.1016/j.biomaterials.2009.12.034

bone formation in extra skeletal sites and soft tissues such asmuscle [24,32–39]. This is highly undesired because it often leadsto dysfunction of the normal tissue and therefore should beexcluded in safety studies prior to clinical use. For the proof ofectopic bone formation with BMPs, rat and mice models are themost commonly used [32,33,39,40]. Previous studies in rats havedemonstrated that ectopic bone formation, induced by BMP-2,occurs through an endochondral series of events [36]. In addition,bone formation has been reported after 21 days in different extraskeletal sites in rats [34,41]. Therefore, we thought of investigatingpotential ectopic bone formation especially that the animalssurvived for a period of 70 days. In our preliminary samples of theleg muscle, no bone metaplasia was observed histologically and the

for the site of injection in the right quadriceps muscle of a rat injected with NPs loadedhe biocompatibility, safety, localization and osteogenic potential of the delivery system.ransverse sections of nerves and blood vessels (bottom). No considerable damage wasted area upon examining the histology slides under higher magnification (200�). Thes and transcription factors (brown staining for chondrocytes) indicating the initiation of

and safety of a hybrid core–shell nanoparticulate OP-1 delivery...,

Z.S. Haidar et al. / Biomaterials xxx (2010) 1–9 7

ARTICLE IN PRESS

injection site could not be always clearly identified; thereforefurther recuts of blocs at different depths were done and preparedas described earlier for histopathological and immunohistochem-ical analysis using Goldner Trichrome staining. As shown in the toptwo rows of Fig. 5, histologically, signs of bone formation wereevident in groups 5 and 6 represented by the presence of pools ofpotential bone-forming cells which were absent in any of the othergroups and controls. It is known that undifferentiated mesen-chymal cells have the potency to differentiate into chondrocytes,fibrocytes and osteocytes [24]. Furthermore, it has been suggestedthat because the effect of growth factors is dosage-dependent,BMPs should be used in dosages high enough to evoke cellularresponses, but not ectopic bone formation. Several studies havereported that for osteoinduction to occur, a minimum dosage of1.0 mg BMP-2 (more for other BMPs) would be required [42–44].Yet, studies continue to require much larger and expensive dosagesin part due to the physico-chemical and release characteristics ofthe carriers used. For example, in one study [37] performed in non-human primates, BMP-2 (50, 250 and 1250 mg) in combination witha type I collagen carrier were implanted into pouches prepared inthe calf muscle of macaque monkeys. The authors reportedosteoinduction in only one third of the animals treated with thehighest protein concentration radiographically, while no boneformation was evident in any of the other two animal groupsbesides slight enlargement of the nuclei of muscle cells that mightbe indicative of a cellular response [37]. In a more recent study [39]using OP-1, synthetic alginate discs were prepared and implantedin the back muscle of mice. Four concentrations were investigatedranging from 3.0 to 30 mg OP-1. Results once more demonstratedectopic bone formation only with the use of the highest dosage ofOP-1 [39]. This basically further confirms the role of our core–shellNPs in localizing, confining and restricting the effect of the release-controlled OP-1 within the site of administration. Moreover, thebiocompatibility and safety of the delivery system can be recog-nized where no considerable or irreversible damage was detectedin situ; suggested by the clear presence of healthy nerve and bloodvessels in the injected area upon examining the histology slidesunder higher magnification. BMPs have been reported to promoteangiogenesis both in vitro and in vivo [45,46]. Angiogenesisprecedes bone healing and formation as vasculature supplies thestem cells, nutrients, mineral elements and cytokines vital forosteogenesis. Deckers et al. [47] demonstrated that BMPs stimulateangiogenesis through enhancing the production of vascular

Table 3Summary of immunolocalization in the harvested right quadriceps muscle of the injected rpositive, (þþ) ¼ – ½ of cells stained positive, (þþþ) ½ – G of cells stained positive, (þþþþgrouped together.

Saline 0.5/1.0 mg OP-1 NPs þ

C F C F C

LigandsBMP-2 � � þ þ þ/�BMP-3 � � þ/� þ þ/�BMP-7/0P-1 � � þ/� þ þ/�

ReceptorsBMPR-I � � þ/� þ/� þ/�BMPR-IIA � � þ þ/� þ/�BMPR-IIB � � þ/� þ/� þ

Transcription factorsSmads 1–5 � � þ/� þ/� þ/�Sox-9 � � � þþ þþ

OthersCollagen-II � � þ þ þ/�Noggin � � þþ þþþ þ

Please cite this article in press as: Haidar ZS, et al., BiocompatibilityBiomaterials (2010), doi:10.1016/j.biomaterials.2009.12.034

endothelial growth factor A (VEGF-A) by osteoblasts and havesuggested that low doses of BMPs may be sufficient for initiatingangiogenesis but not osteogenesis. The capability of the localizedand release-controlled hybrid NPs in promoting the bioavailabilityof the encapsulated growth factor and in very low, safe and cost-effective dosages (0.5 mg OP-1) is thus further demonstrated in ratsin the present study as was established formerly in our rabbit tibialdistraction osteogenesis model [13].

3.5.2. Immunohistochemistry for site of injectionThe expression of 15 genes, representing the BMP signaling

pathway involved in de novo bone formation [48] was evaluated atthe protein level using immunohistochemistry, as displayed in thebottom two rows of Fig. 5. Table 3 summarizes our findings. Theobtained data showed that the investigated genes were expressedat various levels in the six animal groups. Cellular localization wasmore in chondrocytes than in fibroblast-like cells for groups 2through 6 with no response whatsoever in the control animals ofgroup 1. And so, because the amount of mineralized bone wasinconsistent and insufficient, perhaps depending on the site anddepth of the injections, it was difficult to obtain comparative datafrom all samples and prepared slides. Generally, most proteins wereup-regulated post-injections especially in groups 5 and 6 suggest-ing that the localized and release-controlled application of exoge-nous OP-1 most likely did stimulate bone formation processes viaincreasing the rate of expression of several factors involved in theBMP signaling pathways including receptors such as BMPR-I andBMPR-II, antagonists such as Noggin and BMP-3 and transcriptionfactors such as Smads 1–5. This was not as evident in the animalsreceiving pure OP-1 injections. However, data could not confirmthe presence of formed bone in the muscles of these rats wherewhile group 6 appears to have much more induced activity in theexpression of BMP-2, OP-1 and their receptors than the othergroups, the expression of the antagonist Noggin well-known toblock osteogenesis was also slightly more up-regulated, thoughwith no significant differences detected (p� 0.05). Nonetheless, theherein observed under-expression of R-Smads1/5/8 has beenreported earlier to accompany an over-expression of Noggin [49].Although many in vitro studies have been conducted, it is not fullyclear yet in the literature how these BMP-receptor antagonistsregulate in vivo bone formation and therefore the specificity ofBMP-receptor signaling is an area that remains to be explored [48].In addition, such analysis has been mainly carried out and limited

ats: (�) no positive staining, (þ/�)<¼ of cells stained positive, (þ) ¼ of cells stained) more than G of cells stained positive. Groups 2 and 3 yielded the same data and are

0.0 mg OP-1 NPs þ 0.5 mg OP-1 NPs þ 1.0 mg OP-1

F C F C F

þ þ þþ þþ þþ/� þ/� þ þ þþ/� þ þþ þþ þþ

þ þ þþ þþ þþ/� þ/� þ þþ þþþ/� þ þ þ þþ

þ þ þþ þþ þþ/� þþþ þ þþþ þ

þ þþþ þþþ þþþþ þþþþ þ þ þþ þþþ

and safety of a hybrid core–shell nanoparticulate OP-1 delivery...,

Z.S. Haidar et al. / Biomaterials xxx (2010) 1–98

ARTICLE IN PRESS

to fracture healing and de novo bone regeneration models; furtherrestricting a comparison of findings with other delivery systemsand carriers. Finally, the effect of BMPs varies with species, genderand age to name a few. For example, when the effects of BMP-2were evaluated in rats and primates, inconsistent ectopic boneformation resulted in the primate muscle independent on proteindosage whereas this was not the case in the rat muscle showingdose-dependency [24,50]. This discrepancy might be also attrib-uted to differences in the amount of specific BMP-receptors presentin the muscle sites of animal species [43,44] which as well might beconsequential to variations in impact of the trauma to tissues andtheir surroundings from the different modes of application;surgical implantation versus a single injection as is in this study, forinstance. Finally, the degree of vascularization will entail theintensity or moderation of cellular responses [45–47,50].

4. Conclusions

Injectable delivery systems for growth factors are in highdemand due to ease of application, localization at defect site andimproved patient comfort. Besides the previously reported in vitrocytocompatibility, the present study is to the best of our knowledge,the first investigation of the in vivo biocompatibility and safety ofan injectable hybrid nanoparticulate rhOP-1 delivery system con-sisting of a liposomal core and a shell constructed by the layer-by-layer self-assembly of two natural polymers; alginate and chitosan,based on electrostatic interactions. The analytical findings fromevaluating animal body weight and behavioral changes, a total of 50blood markers and hematological levels along with the healthstatus of the six major organs suggested that the nanoparticles, thereleased bioactive load as well as the resulting effects wererestricted to the site of administration with no considerablecomplications or reactions from any degradation by-products. Wehave previously demonstrated the effect of the physico-chemical,localized and sustained release-controlled characteristics of thenanoparticles in enhancing de novo bone regeneration andconsolidation in a long bone distraction osteogenesis model withvery low doses of rhOP-1. The results from this study furtherestablished the possibilities of (a) maintaining the bioactivity ofproteins with short half-lives over the required therapeutic timeand consequently (b) alleviating the often need for using precariousand expensive large supra-physiological dosages and (c) overcomethe need and issues from the invasive surgical implantation ofcurrent scaffolds and carriers for BMP-induced bone regeneration.

Acknowledgements

The authors would like to thank Mrs. Dominique Lauzier forher assistance in immunohistochemistry, Dr. Fereshteh Azari forher assistance in cryo-TEM imaging and Dr. Marilene Paquetfor her comparative pathology analysis. This work was supported byan operating grant from the Shriners of North America, Fonds de laRecherche en Sante du Quebec, the National Science and Engi-neering Research Council (NSERC), the Canadian Institutes of HealthResearch (CIHR) – Regenerative Medicine/Nanomedicine and theCenter for Biorecognition and Biosensors (CBB), McGill University,Montreal, Quebec, Canada. Dr. Haidar acknowledges scholarshipsfrom the Center for Bone and Periodontal Diseases Research and theShriners Hospital, Montreal, Quebec, Canada.

Appendix

Figures with essential color discrimination. Figs. 3–5 in thisarticle are difficult to interpret in black and white. The full color

Please cite this article in press as: Haidar ZS, et al., BiocompatibilityBiomaterials (2010), doi:10.1016/j.biomaterials.2009.12.034

images can be found in the on-line version, at doi:10.1016/j.biomaterials.2009.12.034.

References

[1] Bessa PC, Casal M, Reis RL. Bone morphogenetic proteins in tissue engineering:the road from the laboratory to the clinic, part I (basic concepts). J Tissue EngRegen Med 2008;2:1–13.

[2] Mont MA, Ragland PS, Biggins B, Friedlaender G, Patel T, Cook S, et al. Use ofbone morphogenetic proteins for musculoskeletal applications: an overview. JBone Joint Surg 2004;86:41–55.

[3] Termaat MF, Den Boer FC, Bakker FC, Patka P, Haarman HJ. Bone morphoge-netic proteins: development and clinical efficacy in the treatment of fracturesand bone defects. J Bone Joint Surg Am 2005;87:1367–78.

[4] Bessa PC, Casal M, Reis RL. Bone morphogenetic proteins in tissue engineering:the road from laboratory to clinic, part II (BMP delivery). J Tissue Eng RegenMed 2008;2:81–96.

[5] Luginbuehl V, Meinel L, Merkle HP, Gander B. Localized delivery of growthfactors for bone repair. Eur J Pharm Biopharm 2004;58:197–208.

[6] Illum L, Davis SS. Drug delivery. Curr Opin Biotechnol 1991;2:254–9.[7] Chaize B, Colletier JP, Winterhalter M, Fournier D. Encapsulation of enzymes in

liposomes: high encapsulation efficiency and control of substrate perme-ability. Artif Cells Blood Substit Immobil Biotechnol 2004;32:67–75.

[8] Takeuchi H, Kojima H, Yamamoto H, Kawashima Y. Polymer coating of lipo-somes with a modified polyvinyl alcohol and their systemic circulation andRES uptake in rats. J Control Release 2000;68:195–205.

[9] Charrois GJ, Allen TM. Rate of biodistribution of STEALTH� liposomes to tumorand skin: influence of liposome diameter and implications for toxicity andtherapeutic activity. Biochim Biophys Acta 2003;1609:102–8.

[10] Haidar ZS, Hamdy RC, Tabrizian M. Protein release kinetics for core-shellhybrid nanoparticles based on the layer-by-layer assembly of alginate andchitosan on liposomes. Biomaterials 2008;29:1207–15.

[11] Haidar ZS, Azari F, Hamdy RC, Tabrizian M. Modulated release of OP-1 andenhanced preosteoblast differentiation using a core-shell nanoparticulatesystem. J Biomed Mater Res A 2009;31:1817–24.

[12] Kim JS, Yoon TJ, Yu KN, Kim BG, Park SJ, Kim HW, et al. Toxicity and tissuedistribution of magnetic nanoparticles in mice. Toxicol Sci 2006;89:338–47.

[13] Haidar ZS, Tabrizian M, Hamdy RC. A hybrid OP-1 delivery system enhancesnew bone regeneration and consolidation in a rabbit model of distractionosteogenesis. Growth Factors, in press [Epub ahead of print].

[14] Gref R, Minamitake Y, Perracchia MT, Trubeskoy V, Torchilin V, Langer R.Biodegradable long-circulating polymeric nanospheres. Science1994;263:1600–3.

[15] Desai MP, Labhasetwar V, Amidon GL, Levy RJ. Gastrointestinal uptake of biode-gradable microparticles: effect of particle size. Pharm Res 1996;13:1838–45.

[16] Fabian E, Landsiedel R, Ma-Hock L, Wiench K, Wohlleben W, vanRavenzwaay B. Tissue distribution and toxicity of intravenously administeredtitanium dioxide nanoparticles in rats. Arch Toxicol 2008;82:151–7.

[17] De Souza R, Zahedi P, Allen CJ, Piquette-Miller M. Biocompatibility of injectablechitosan–phospholipid implant systems. Biomaterials 2009;30:3818–24.

[18] Hamdy RC, Amako M, Beckman L, Kawaguchi M, Rauch F, Lauzier D, et al.Effects of osteogenic protein-1 on distraction osteogenesis in rabbits. Bone2003;33:248–55.

[19] Haque T, Mandu-Hrit M, Rauch F, Lauzier D, Tabrizian M, Hamdy R. Immu-nohistochemical localization of BMP signaling Smads during long-bonedistraction osteogenesis. J Histochem Cytochem 2005;54:407–15.

[20] Mandu-Hrit M, Haque T, Lauzier D, Kotsiopriftis M, Rauch F, Tabrizian M, et al.Early injection of OP-1 during distraction osteogenesis accelerates new boneformation in rabbits. Growth Factors 2006;24:172–83.

[21] Stokes WS. Humane endpoints for laboratory animals used in toxicity testing.ILAR J 2002;43:31–8.

[22] Aubert AE, Vandeput S, Beckers F, Liu J, Verheyden B, Van Huffel S. Complexityof cardiovascular regulation in small animals. Philos Transact A Math Phys EngSci 2009;367:1239–50.

[23] Wang HP, Liang YJ, Long DX, Chen JX, Hou WY, Wu YJ. Metabolic profiles ofserum from rats after subchronic exposure to chlorpyrifos and carbaryl. ChemRes Toxicol 2009;22:1026–33.

[24] Wildemann B, Kandziora F, Krummrey G, Palasdies N, Haas NP, Raschke M,et al. Local and controlled release of growth factors (combination of IGF-I andTGF-beta I, and BMP-2 alone) from a polylactide coating of titanium implantsdoes not lead to ectopic bone formation in sheep muscle. J Control Release2004;5:249–56.

[25] Oda S, Kinoshita A, Higuchi T, Shizuya T, Ishikawa I. Ectopic bone formation bybiphasic calcium phosphate (BCP) combined with recombinant human bonemorphogenetic protein-2 (rhBMP-2). J Med Dent Sci 1997;44:53–62.

[26] Takeuchi H, Kojima H, Yamamoto H, Kawashima Y. Evaluation of circulationprofiles of liposomes coated with hydrophilic polymers having differentmolecular weights in rats. J Control Release 2001;75:83–91.

[27] Yang KC, Wu CC, Lin FH, Qi Z, Kuo TF, Cheng YH, et al. Chitosan/gelatinhydrogel as immunoisolative matrix for injectable bioartificial pancreas.Xenotransplantation 2008;15:407–16.

[28] Sells DM, Brix AE, Nyska A, Jokinen MP, Orzech DP, Walker NJ. Respiratorytract lesions in noninhalation studies. Toxicol Pathol 2007;35:170–7.

and safety of a hybrid core–shell nanoparticulate OP-1 delivery...,

Z.S. Haidar et al. / Biomaterials xxx (2010) 1–9 9

ARTICLE IN PRESS

[29] Mohr U, Dungworth DL, Capen CC. In: Mohr U, Dungworth DL, editors.Pathobiology of the aging rat. Washington, DC: Intl Life Sciences Inst (ILSI)Press; 1992. p. 485–8.

[30] Anderson JM. Biological responses to materials. Annu Rev Mater Res 2001;31:81–110. <http://arjournals.annualreviews.org doi:10.1146/annurev.matsci.31.1.81>.

[31] Urist MR. Bone: formation by autoinduction. Science 1965;150:893–9.[32] Yamamoto M, Tabata Y, Ikada Y. Ectopic bone formation induced by

biodegradable hydrogels incorporating bone morphogenetic protein. JBiomater Sci Polym Ed 1998;9:439–58.

[33] Nakagawa T, Tagawa T. Ultrastructural study of direct bone formation induced byBMPs-collagen complex implanted into an ectopic site. Oral Dis 2000;6:172–9.

[34] Okubo Y, Bessho K, Fujimura K, Konishi Y, Kusumoto K, Ogawa Y, et al.Osteoinduction by recombinant human bone morphogenetic protein-2 atintramuscular, intermuscular, subcutaneous and intrafatty sites. Int J OralMaxillofac Surg 2000;29:62–6.

[35] Kurokawa I, Kusumoto K, Bessho K, Okubo Y, Senzaki H, Tsubura A. Immu-nohistochemical expression of bone morphogenetic protein-2 in piloma-tricoma. Br J Dermatol 2000;143:754–8.

[36] Nakagawa T, Sugiyama T, Kamei T, Murata T, Tagawa T. An immuno-light- andelectron-microscopic study of the expression of bone morphogenetic protein-2 during the process of ectopic bone formation in the rat. Arch Oral Biol2001;46:403–11.

[37] Kusumoto K, Bessho K, Fujimura K, Akioka J, Okubo Y, Wang Y, et al.Osteoinduction by recombinant human bone morphogenetic protein-2 inmuscles of non-human primates. J Int Med Res 2002;30:251–9.

[38] Wysocki RW, Cohen MS. Ectopic ossification of the triceps muscle afterapplication of bone morphogenetic protein-7 to the distal humerus forrecalcitrant nonunion: a case report. J Hand Surg Am 2007;32:647–50.

[39] Nanno K, Sugiyasu K, Daimon T, Yoshikawa H, Myoui A. Synthetic alginate isa carrier of OP-1 for bone induction. Clin Orthop Relat Res 2009;467:3149–55.

[40] Stoeger T, Proetzel GE, Welzel H, Papadimitriou A, Dony C, Balling R, et al. Insitu gene expression analysis during BMP2-induced ectopic bone formation in

Please cite this article in press as: Haidar ZS, et al., BiocompatibilityBiomaterials (2010), doi:10.1016/j.biomaterials.2009.12.034

mice shows simultaneous endochondral and intramembranous ossification.Growth Factors 2002;20:197–210.

[41] Yoshida K, Bessho K, Fujimura K, Kusumoto K, Ogawa Y, Tani Y, et al.Osteoinduction capability of recombinant human bone morphogeneticprotein-2 in intramuscular and subcutaneous sites: an experimental study. JCraniomaxillofac Surg 1998;26:112–5.

[42] Uludag H, D’Augusta D, Golden J, Li J, Timony G, Riedel R, et al. Implantation ofrecombinant human bone morphogenetic proteins with biomaterial carriers:a correlation between protein pharmacokinetics and osteoinduction in the ratectopic model. J Biomed Mater Res 2000;50:227–38.

[43] Bessho K, Carnes DL, Cavin R, Ong JL. Experimental studies on bone inductionusing low-molecular-weight poly(DL-lactide-co-glycolide) as a carrier forrecombinant human bone morphogenetic protein-2. J Biomed Mater Res2002;61:62–5.

[44] Yamamoto M, Takahashi Y, Tabata Y. Controlled release by biodegradablehydrogels enhances the ectopic bone formation of bone morphogeneticprotein. Biomaterials 2003;24:4375–88.

[45] Carano RAD, Filvaroff EH. Angiogenesis and bone repair. Drug Discov Today2003;8:980–9.

[46] David L, Feige JJ, Bailly S. Emerging role of bone morphogenetic proteins inangiogenesis. Cytokine Growth Factor Rev 2009;20:203–12.

[47] Deckers MML, van Bezooijen RL, van der Horst G, Hoogendam J, van derBent C, Papapoulos SE, et al. Bone morphogenetic proteins stimulate angio-genesis through osteoblast-derived vascular endothelial growth factor A.Endocrinology 2002;143:1545–53.

[48] Rosen V. BMP and BMP inhibitors in bone. Ann N Y Acad Sci 2006;1068:19–25.[49] Tsialogiannis E, Polyzois I, Oak Tang Q, Pavlou G, Tsiridis E, Heliotis M, et al.

Targeting bone morphogenetic protein antagonists: in vitro and in vivo evidenceof their role in bone metabolism. Expert Opin Ther Targets 2009;13:123–37.

[50] Aspenberg P, Turek T. BMP-2 for intramuscular bone induction: effect insquirrel monkeys is dependent on implantation site. Acta Orthop Scand 1996;67:3–6.

and safety of a hybrid core–shell nanoparticulate OP-1 delivery...,