Synthesis and characterization of self-curing hydrophilic bone cements for protein delivery ndez1 E. Franco-Marquès,1 J. Parra,2 M. A. Pèlach,1 J. A. Me 1 LEPAMAP Group, Universitat de Girona, Maria Aurèlia Capmany 61, 17071, Girona, Spain   n Clınica y Biopatologıa Experimental, Complejo Asistencial de Avila Unidad Asociada de I1D al CSIC de Investigacio  s del Gran Poder 42, 05003 Avila, Spain (SACYL), Jesu 2 Received 30 April 2014; revised 25 July 2014; accepted 19 August 2014 Published online 00 Month 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/jbm.b.33283 Abstract: New formulations of acrylic bone cements for bone defect reparation, based on self-hardening methyl methacrylate (MMA)/methacrylic acid (MAA), with a high capacity for protein delivery, have been developed. The self-curing formulations were prepared by partial substitution of solid phase PMMA microparticles by newly obtained PMAA microspheres. The PMAA microspheres were prepared by inverse suspension polymerization of their monomer and were crosslinked with N,N’-methylene-bis-acrylamide (MBA) (10–15 wt %) to produce stable systems in contact with aqueous media. PMAA microspheres were loaded with hydrolyzed collagen (HC) as a model protein to simulate bone morphogenetic protein delivery useful for hard tissue reconstruction. Solid phase PMMA microparticles in the formulation were partially substituted by new PMAA-HC microspheres and were characterized to determine viability as an acrylic bone cement in minimally invasive surgery. The incorporation of PMAA-HC microspheres decreased peak temperature by 20 C, which minimized thermal necrotic risk after implantation. Mechanical compression tests revealed a behavior, under dry conditions, close to ISO 5833 standard requirements. However, a drastic drop in mechanical strength, 64%, was obtained after 15 days of immersion in simulated physiological conditions (37 C and pH 7.4) and was attributed to water absorption and a subsequent plasticizing effect. The increase in water uptake and retention enhanced the capability for controlled protein delivery. Finally, the biocompatibility of the cements was determined; some toxicity of the material during the first hours of culture incubation was observed. Later, toxicity was observed to decrease due to nonreacted monomer leaching, which ensured the low toxicity of the already C 2014 Wiley Periodicals, Inc. J Biomed Mater polymerized phase. V Res Part B: Appl Biomater 00B:000–000, 2014. Key Words: self-curing bone cements, minimally invasive surgery, hydrophilic cements, protein delivery, dorsalgia ndez JA. 2014. Synthesis and characterization of self-curing How to cite this article: Franco-Marquès E, Parra J, Pèlach MA, Me hydrophilic bone cements for protein delivery. J Biomed Mater Res Part B 2014:00B:000–000. INTRODUCTION Current acrylic bone cements, which are used for bone reparation and reconstruction, are based on poly(methyl methacrylate) (PMMA). These types of materials are obtained by radical polymerization, which is initiated at low temperature with a redox initiating system, of a liquid phase based on methyl methacrylate monomer (MMA) in the presence of an already polymerized solid phase based on PMMA microparticles. From the beginning, these materials have been used in hip and knee total arthroplasty,1,2 dental surgery,3,4 bone defect reparation and reconstruction5, and even in the treatment of tumors.6 The use of plain PMMA systems is focused on the prosthesis fixation or bone defect stabilization without any other expectation of interaction with living tissue. The interaction of biological activity with damaged or newly born tissue is a broad field of research that is constantly growing. Mendez et al. developed an acrylic bone cement that was modified with a phosphate derivative of salicylic acid; antiinflammatory activity on the implant’s surrounding tissue diminished the inflammatory process because of the presence of scanty multinucleated giant cells.7 Included antibiotics8,9 have effectively diminished the risk of infection after surgery by controlled delivery during the first period of implantation. Vitamins such as a-tocopherol have also been included both by mixing with biodegradable polymers10 and by copolymerizing with MMA11 to decrease oxidation phenomena during in situ polymerization, which improves biocompatibility. One of the handicaps of this biomedical technology is related to the low capacity of interaction of PMMA with aqueous fluids, such as biological ones, to promote the interaction between living tissue and PMMA implants. PMMA is a very hydrophobic polymer with a water uptake ndez; e-mail: jalberto.mendez@udg.edu Correspondence to: J. A. Me Contract grant sponsor: MEC of Spain; contract grant number: MAT2010-18155 Contract grant sponsor: University of Girona; contract grant number: BR-07/05 C 2014 WILEY PERIODICALS, INC. V 1 capacity lower than 2.0–2.5 wt %, which avoids the diffusion of water through the material and limits its use as a carrier for hydrophilic medicaments.9,12 In this sense, many of the medicaments, with useful activity for osteoinduction and osteoconduction, are highly hydrophilic and once incorporated to the cement, have serious limitations in migrating from the material to the surrounding tissue. The common behavior of the delivery process from highly hydrophobic materials is characterized by a fast release of the drug close to the implant surface,8,9 which is limited to the outer layers of the material. The release profile presents a burst effect, leaching up to 90% of the maximum medicament content during the first hours of incubation/implantation, and retaining a high percentage inside the material after reach ing the equilibrium release d½drug dt 50 . Release behavior depends on the drug-polymer interaction. Caracciolo et al. synthesized polyurethane networks (PUNs) based on oligomers of caprolactone (CL) triol or tetronic 701 as a polyol and poly(ethyleneglycol) (PEG) as an extender cross-linked using hexamethylene diisocyanate (HDI) to determine its use as a drug delivery system of paracetamol.13 The high hydrophilicity of the polymer matrix and its low interaction with the medicament gave rise to release profiles with 20% of the total drug delivered in the first 10 h of immersion in phosphate buffered solution. Thus, the equilibrium between the hydrophobic and hydrophilic ratio must be controlled to reach the desired profiles. The application of the self-curing acrylic material prepared in this work is designed to repair vertebra bone defects derived from degenerative pathologies such as osteoporosis. Osteoporosis appears with irreversible damage of bone microarchitecture and a reduced bone mass, which increases the risk of fractures. The disease is mainly due to a dysfunction of bone remodeling and allows the formation of bone cavities produced by the loss of bone. Most patients suffer strong back pain (dorsalgia) related with the vertebrae crushing affected by osteoporosis.14 One surgical technique used to diminish dorsalgia is vertebroplasty, where an acrylic bone cement is injected directly into the vertebra to reconstruct its structure and support the body’s weight. In this work, a new self-curing acrylic material has been obtained to allow protein delivery, which minimizes burst effect and increases release time. Solid phase PMMA microspheres of the acrylic formulation have been substituted by poly(methacrylic acid) (PMAA) cross-linked microspheres to increase water uptake and maintain the methacrylic nature of the polymer matrix to improve the chemical compatibility. PMAA-based microspheres have been prepared by means of inverse suspension polymerization using methacrylic acid (MAA) as a monomer cross-linked with N,N’methylene-bis-acrylamide (MBA) to avoid solubilization in aqueous media. The PMAA microspheres were characterized by different microscopic and spectroscopic techniques and loaded with hydrolyzed collagen (HC) as a model protein to simulate the delivery of bone morphogenetic proteins (BMPs) with a high capacity for osteoinduction and osteoconduction.15–17 Finally, the biocompatibility of the formula- 2  ET AL. FRANCO-MARQUES tions was characterized by means of studying the in vitro toxicity of samples in contact with human osteoblasts. MATERIALS AND METHODS Materials Materials for synthesis and physical-chemical characterization. Poly(methyl methacrylate) (PMMA) microspheres (DegacrylTM MW 332) were provided by Degussa, with an average diameter of 60 lm. Methacrylic acid (MAA) (SigmaAldrich) was purified by vacuum distillation and neutralized with a concentrated solution of NaOH to avoid monomer distribution between the aqueous and organic phases during inverse suspension polymerization. Methyl methacrylate (MMA) (Acros Organics) and N,N0 -methylene-bis-acrylamide (MBA) (Sigma-Aldrich), used as received, were used as acrylic monomer and the crosslinking agent, respectively. Benzoyl peroxide (BPO), purchased from Scharlau, purified by fractional crystallization from methanol (mp 104 C), and 4,40 -(dimethyl amino)diphenyl carbinol (BZN) (Fluka), used as received, were used as a low temperature redox initiator system. Hydrolyzed collagen (HC) has been kindly given by Protein (Celra, Spain) and is commercialized under the commercial name of Colnatur. Potassium peroxodisulphate (K2S2O8) and Span 80 (sorbitan monooleate) (SigmaAldrich) were used as received as the polymerization initiator and nonionic surfactant. The other reagents for synthesis were of analytical grade and used without any further purification. Materials for toxicity characterization. Nunc ThermanoxTM coverslips were provided by Thermo Scientific. Foetal bovine serum (FBS) and Triton X-100 were supplied by Gibco and Merck, respectively. Dimethylsulfoxide (DMSO) was acquired from Scharlau. All other reagents were supplied by Sigma-Aldrich. Methods Preparation of PMAA microspheres by inverse suspension polymerization. A typical procedure for the inverse suspension polymerization. Aqueous phase. MAA (17.0 mL, 17.2 g, 200 mmol) were drop-wise neutralized with an alkaline aqueous solution (NaOH, 8 g, 200 mmols), and dissolved in 20 mL of distilled water in an ice cooling bath. MBA was added to the totally neutralized MAA solution in a 100-mL beaker. After mild stirring for few minutes, K2S2O8 (1% w/w regarding monomer content) was added to the monomer solution. To induce complete dissolution of the polymerization initiator, the minimum volume of distilled water was added to the solution at room temperature until obtaining a clear and saturated homogeneous solution. Oil phase. 8.7 mL of Span 80 (5.5% v/v regarding oil phase) were added to 150 mL of toluene in a 250-mL three-neck jacketed reactor mechanically stirred (600 rpm) and heated at 70 C. The mixture was purged with nitrogen for 10 min prior to the reaction. The aqueous phase was placed in a dropping funnel and flushed with nitrogen for 15 min. Later, the aqueous phase SELF-CURING HYDROPHILIC BONE CEMENTS FOR PROTEIN DELIVERY ORIGINAL RESEARCH REPORT was drop-wise added to the oil phase, also previously flushed with nitrogen. The reactor was stirred (600 rpm) and heated for 4 h. Once this time had elapsed, the temperature was increased to 80 C for two more hours. Isolation of the PMAA microspheres. The oil phase (upper phase) was removed by a Pasteur pipette and the obtained PMAA microspheres, together with the aqueous phase, were added to 400 mL of acetone in a 1L-beaker and stirred for 20 min. Later, the microspheres were filtrated under vacuum. This dispersion/filtration process was repeated three times. The microspheres were washed with distilled water and lyophilized to constant weight (Virtis lyophilizer, New York). Preparation of PMAA-HC microspheres. Depending on the intended HC content, 2 or 5 g of HC were dissolved in 200 mL of distilled water at room temperature. The solution was added to 20 g of PMAA-microspheres under strong stirring. Once the solution was completely absorbed by the microspheres, the newly formed PMAA-HC microspheres were lyophilized (Virtis lyophilizer, New York) to eliminate water. Afterwards, PMAA-HC microspheres were stored until use in an anhydrous atmosphere. Characterization of PMAA and PMAA-HC microspheres. Particle size distribution. The particle size distribution was evaluated by laser scattering using a Coulter LS320 (Beckman). About 10 mg of each powdered material were thoroughly dispersed in 2 mL of acetone by magnetic stirring for 10 min and finally, an average of three measurements was recorded for each sample. Surface characterization. The prepared microparticles and the derived acrylic self-curing formulation were characterized by scanning electron microscopy (SEM). Each sample was previously conditioned by freezing with liquid nitrogen and sputter-coating with gold (K550 from Emitech, Ashfort, UK). Finally, the samples were observed under a Zeiss DMS 960 model electronic microscope. Chemical composition. Qualitative chemical composition of the microparticles was determined by means of FTIR spectrophotometry. KBr-supported pellets were prepared with 2–3 wt % of each sample and the spectra were recorded, from 400 to 4000 cm21 wavenumber, using a Mattson-Satellite FTIR spectrophotometer. Thermal characterization. Differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA) were performed. Samples of 10 mg were submitted to a heating program from 25 to 200 C in a Mettler Toledo 822e calorimeter (Switzerland). Additionally, samples of 8–10 mg were heated from 50 to 650 C using a Mettler Toledo TGA/ DTA 851e thermobalance (Switzerland). Both tests were carried out in independent experiments and were performed under inert atmosphere (nitrogen 40 mL min21). Preparation of hydrophilic self-curing formulations. The solid phase of the self-curing formulation was based on three main components: commercial PMMA beads, and/or the prepared PMAA/PMAA-HC microspheres (50 : 50), and BPO as the polymerization initiator (1.5 wt % regarding solid phase). The liquid phase was composed of the monomer (MMA) and BZN (1.0 wt %). To obtain low viscous and easy handling formulations, the solid:liquid ratio of compositions (S:L) was adjusted to 1 : 1. Polymerization started with the addition of the liquid to the solid phase. Initially, the reacting mass was hand-stirred with a spatula at a very low mixing rate (20–30 r.p.m.) to avoid trapping air bubbles. Once the reacting mass was not stuck to the surgical glove, it was molded inside PTFE-moulds equipped with pressing plates. Finally, the mould was placed in an oven at 37 C for 1 h, to simulate physiological conditions. Characterization of hydrophilic self-curing formulations. Curing parameters. The energetic behavior of the polymerizing process was determined by recording temperature-curing time values of the curing process, performed within a cylindrical PTFE-made mould at 25 C, according to ISO 5833 standard specifications. A thermocouple was placed in the centre of the polymerizing mass, at a height of 3 mm in the internal cavity of the mould, to detect the variation of temperature with curing time. The temperature-time data were recorded with a highly sensitive thermotester with automatic data acquisition (Testo 177, Barcelona, Spain). In vitro behavior. Water sorption (%WS) of the hydrophilic acrylic formulations was evaluated at 37 C and pH 7.4 (PBS). This parameter was studied after the immersion of discs of each formulation (15 mm in diameter and 1 mm in thickness) during different time periods. An average of a minimum of three independent experiments was obtained. The determination of %WS was according to Eq. (1):   ðWt 2W0 Þ %WS5  100 W0 (1) where W0 is the weight of the sample before immersion and Wt is the weight at immersion time t. Mechanical characterization. A minimum of five cylindrical-shaped specimens of each formulation, 6 mm in diameter and 12 mm in height, were tested at room temperature under compression stresses in an IDM DTC-10 dynamometer according to ISO 5833 standard specifications. Evaluation of protein delivery. Hydrolyzed collagen, HC, was used as a protein model to determine the capacity of the system to be used as a protein carrier. Polymerized disc-shaped specimens (15 mm in diameter and 1 mm in thickness) of formulations loaded with HC were immersed in 15 mL of PBS (37 C and pH 7.4). PBS was totally replaced at specific periods of time by a fresh buffer solution and the HC content was monitored by UV-VIS spectrophotometry (205 nm) (Schimadzu UV-160). Toxicity evaluation. The biological response to the experimental formulations was tested using cultures of human foetal osteoblasts (HOb; Health Protection Agency Culture Collections). The culture medium was Dulbecco’s Modified Eagle’s Medium/Nutrient Mixture F-12 Ham (DME/F12 1 : JOURNAL OF BIOMEDICAL MATERIALS RESEARCH B: APPLIED BIOMATERIALS | MONTH 2014 VOL 00B, ISSUE 00 3 1 mixture) modified with HEPES (4-(2-hydroxyethyl)21piperazine ethanesulfonic acid) and supplemented with 10% fetal bovine serum (FBS), 200 mM L-glutamine (Sigma), 100 units/mL penicillin and 100 lg/mL streptomycin. HOb cultures were maintained at 37 C in humidified air with 5% CO2, and the culture medium was changed every 2 days with care to cause little disturbance to culture conditions. R coverslips with a diameter of 13 mm (TMX) ThermanoxV and Triton X-100 (0.5% aqueous solution) were used as negative and positive controls, respectively. Square-shaped samples of the experimental formulations, with a side length of 12 mm and a thickness of 2 mm, were used to quantify the cytotoxicity. All specimens were sterilized with ethylene oxide. The in vitro biological behavior of the analyzed systems was assessed with a MTT (3-(4,5-dimethylthiazol-2yl)22,5-diphenyltetrazolium bromide) assay, using a HOb culture (subculture 4) and the extracts eluted from both the experimental formulations and TMX samples. For determination of the extracts’ toxicity, the samples were set up in 5 mL of FBS free medium, and they were placed on a roller mixer at 37 C. The medium was removed and cryopreserved at different periods of time (1, 3, and 7 days) and replaced with the other 5 mL of FBS free fresh medium at 1 and 3 days. All of the extracts were obtained under sterile conditions. Osteoblasts suspended in complete culture medium were seeded into sterile 96-well culture plates at a density of 1 3 105 cellsmL21 (100 mL/well) and were incubated to confluence during 24 h. Next, the medium was replaced with the corresponding extract (n 5 6), and the cultures were incubated at 37 C in humidified air with 5% CO2 for 24 h. A solution of MTT was prepared in warm PBS (0.5 mg/mL), which, after removal of the sample extracts, was added to the cultures (100 mL/well). Then, the plates were incubated at 37 C for 4 h. Excess medium and MTT were removed, and dimethyl sulphoxide was added to all wells to dissolve formazan resulting from the reduction of MTT by the cells. After mixing for 10 min, the absorbance was measured with a Biotek ELX808IU detector using a test wavenumber of 570 nm and a reference wavenumber of 630 nm. The cell viability was calculated using Eq. (2): . RELATIVE VIABILITY ð% TMXÞ5ðODS 2ODB Þ 3100 (2) ðODC 2ODB Þ where ODS, ODB, and ODC are the optical density of formazan production for the sample, blank (medium without cells) and TMX negative control, respectively. Analysis of variance (ANOVA) of the results was performed with respect to TMX at p < 0.05, p < 0.01, and p < 0.001 significance levels. RESULTS Preparation of PMAA and PMAA-HC-microspheres In this work, methacrylic acid (MAA) has been used as a hydrophilic monomer to prepare PMAA microspheres and was subsequently loaded with hydrolyzed collagen as a model molecule for controlled delivery of bone morphogenetic proteins. A difunctionalized monomer (MBA) was also 4  ET AL. FRANCO-MARQUES added to MAA to induce crosslinking and avoid dissolution of PMAA in contact with simulated physiological fluids. Two different systems of microspheres were designed depending on MBA content: M10 and M15, where MBA was added by 10 and 15 wt % regarding the monomer weight, respectively. The addition of lower percentages of the crosslinking agent led to water soluble materials, which were not useful for our purposes. In both cases, microsphere particles were obtained, as can be observed in Figure 1(A,B); microsphere particles were insoluble in water but had a high capacity for water retention. Characterization of the microspheres Once the PMAA microspheres were dried, both formulations M10 and M15 were swollen with a solution of hydrolyzed collagen (HC) in distilled water. Four formulations of PMAA/HC-loaded microspheres were obtained, depending on the microsphere composition, which are summarized in Table I. The subsequent lyophilization process produced a material loaded with HC spherical-shaped microparticles as can be observed in Figure 1(C,D). Laser scattering revealed a different particle diameter average compared with SEM observation. Via laser scattering, no substantial dependence was observed between the particle size distribution and the content of MBA because the maximum population of microparticles of both sizes’ dispersion curves were centered very closely. M10 and M15 showed a particle diameter centered at approximately 80– 90 lm [Figure 2(A)], although M15 had a broader dispersion. However, M10 exhibited a particle size in the range of 30–40 lm, as measured by SEM (n 5 100 particles). These differences in particle size and statistical distribution are related with the establishment of particle aggregates as observed in Figure 2(B). PMAA and PMAA-HC microspheres were submitted to a physicochemical characterization to determine their properties and compositions. An FTIR spectrum of PMAA microspheres revealed the presence of methacrylate salts instead of methacrylic acid functionality, corroborating the previous neutralization of the monomer (Figure 3). Signals at 1540 and 1392 cm21 correspond to symmetric and antisymmetric stretching of carboxylate functionality (COO2). The presence of the signal 1197 cm21 (C-COO2 stretching) also confirms this hypothesis.18 The results of weight loss versus temperature, obtained by TGA-analysis, of the formulations of composites reinforced with PMAA and with PMAA-HC microspheres are shown in Figure 4. There is a progressive degradation, in all cases, of the material over 100 C to 400 C, far above the use temperature for these materials (37 C). The degradation is more intensive in the case of materials loaded with HC because of its lower thermal resistance. The thermal decomposition under inert conditions left a high content in char residue at temperatures lower than 500 C, which was in the range of 40–50 wt % in all cases. DSC profiles of the materials prepared in this work are shown in Figure 5. Taking into account the percentage of crosslinker used for their formulation, no glass transition SELF-CURING HYDROPHILIC BONE CEMENTS FOR PROTEIN DELIVERY ORIGINAL RESEARCH REPORT FIGURE 1. SEM microphotographs of PMAA-microspheres: (A) M10, (B) M15, (C) M10-10, and (D) M10–25. was observed in the range of temperatures assayed in this work. Intermolecular crosslinking reduces the mobility of the polymer segments, giving rise to an increase in the glass transition that, in all cases, was higher than the degradation temperature. Preparation and physicochemical characterization of the self-curing formulations Four different acrylic cements were prepared using each formulation of PMAA-based microspheres loaded with TABLE I. Composition of PMAA Microsphere (Mxx) and PMAA-HC Microcapsules (Mxx-yy) Formulation M10 M15 M10-10 M10–25 M15-10 M15–25 a b PMAA (wt %)a MBA (wt %)a HC (wt %)b 90 85 90 90 85 85 10 15 10 10 15 15 0 0 10 25 10 25 Regarding weight of PMAA microspheres. Regarding weight of PMAA-HC microspheres. hydrolyzed collagen (HC). Moreover, four formulations named FC-10, FC-25, FM-10, and FM-25 were also prepared for comparison purposes. The compositions of all of the cement formulations are summarized in Table II. To acquire easily handled reacting masses and to allow an optimal molding of the material, a typical 1 : 1 solid phase:liquid phase composition was used. All of the formulations were cured at 37 C for 1 h and stiff materials were obtained in all cases. The study of the curing process allowed for the determination of the peak temperature (Tpeak) and the setting time (tsetting), derived from the exothermic polymerization process, which are summarized in Table III. The peak temperature of the cements modified with PMAA-microspheres, and specifically, with PMAA-HC microspheres, was lower than that of the reference (PMMA) and lower than the maximum value accepted by ISO 5833 (90 C). The tsetting was not influenced by the composition of the microspheres; in all cases, tsetting was in the range of 13–14 min. The fracture surface of composites was evaluated by SEM microscopy and the micrographs are shown in Figure 6. It is possible to observe the interphase between the continuous PMMA matrix, synthesized during the curing JOURNAL OF BIOMEDICAL MATERIALS RESEARCH B: APPLIED BIOMATERIALS | MONTH 2014 VOL 00B, ISSUE 00 5 FIGURE 2. (A) Particle size distribution obtained by laser scattering; (B) SEM microphotograph of M10 aggregates. process, and the PMAA-microspheres, characterized by the hollows between both components, where the transmission of loads is not optimized. The micrograph also shows the dispersion and individualization of the PMAA-based microparticles inside the material, helping the maintenance of the material’s mechanical properties. The materials were tested under compression stresses, in a dry state and after 15 days of immersion in PBS, and the results are summarized in Table IV. A decrease in the compression yield stress (rc) has been observed when the material is modified with PMAA or PMAA-HC microspheres because of the incorporation of water inside the structure of the composite. This phenomenon is more evident in formulations modified with the obtained PMAA microspheres than in the reference (PMMA), due to their higher water uptake capacity. In the case of the formulations modified with hydrophilic microspheres, the value of rc was close to ISO 5833 standard requirements (70 MPa), which was an acceptable result considering the high level of chemical structure modification of the cement (50 wt % regarding the solid phase). The immersion of the samples in PBS (37 C) led to an important decrease in rc. The incorporation of water in the internal structure of the samples acts like a plasticizer, and decreases the mechanical capacity of the material. Increasing hydrophilicity of the acrylic cements has been corroborated by means of monitoring the in vitro behavior of the material, after immersion in PBS for different time periods at 37 C. PMMA, FC-10 and FC-25 showed a typical hydrophobic behavior of common acrylic bone cements as shown in Figure 7. All formulations modified with PMAA or PMAA-HC microspheres reached higher values of water uptake without modification of the samples’ dimensions. These water uptake results (WS%), in the range of 50–55% for the formulations modified with PMAA microspheres, are slightly higher, in the range of 35–50%, than those of materials modified with PMAA-HC microspheres. To keep the dimensions of the samples stable, the absorbed water is expected to be located in the holes/pores created in the PMMA-PMAA interfaces and is expected to interact with polar groups of the microparticles. FIGURE 3. FT-IR spectra of hydrolysed collagen (HC), M15 PMAA microspheres and M15–25 PMAA-HC microspheres. FIGURE 4. TGA thermograms of PMAA microspheres (M10 and M15) and PMAA-HC microspheres (not distinguished). 6  ET AL. FRANCO-MARQUES SELF-CURING HYDROPHILIC BONE CEMENTS FOR PROTEIN DELIVERY ORIGINAL RESEARCH REPORT ered a toxic species due to the high viability of cells in formulations FC-10 and FC-25. DISCUSSION FIGURE 5. DSC thermograms of PMAA and PMAA-HC microspheres. The profiles of HC delivery with immersion time in PBS are shown in Figure 8, and follow a similar water sorption profile trend. The drug delivery capacity is strongly dependent on the presence of PMAA in the chemical structure of the cement. The nonmodified formulation, made of plain PMMA in the solid phase (PMMA, FC-10, and FC-25), showed the lowest capacity of protein delivery with a maximum dosage of 0.33 mg/mL delivered during the first 24– 48 h of immersion. However, the formulations modified with PMAA-HC microspheres deliver higher protein content, which is higher for formulations with higher HC content in the composition. Characterization of the toxicity of the self-curing formulations A partial toxicity of formulations modified with PMAAmicrospheres has been detected and is shown in Figure 9. A decrease of the cell viability during the first hours of incubation for formulations obtained with PMAA microspheres can be observed, but after 3 or 7 days, a decrease in toxicity was found, which suggests an initial toxicity attributed to material components that are washed at the beginning of the test. Additionally, hydrolyzed collagen was not considTABLE II. Composition of the Solid Phase of Self-Curing Formulations Loaded with HC as Protein Model Solid phase a Formulation PMMA FC-10 FC-25 FM10 FM15 FM10-10 FM10–25 FM15-10 FM15–25 a b PMMA (wt %) 98.6 93.5 86.1 49.3 49.3 49.3 49.3 49.3 49.3 PMAAa microspheres (wt %) HCa 5 12.5 49.3 49.3 44.2 36.7 44.2 36.7 (M10) (M15) (M10) (M10) (M15) (M15) Regarding weight of the solid phase. Regarding full weight of the cement. 5 12.5 10 12.5 HCb (wt %) 0 2.5 6.25 0 0 2.5 6.25 2.5 6.25 The self-curing formulations, also named acrylic bone cements, have been designed for several decades as fixation systems in orthopaedic surgery. Most of these systems also offer the capacity to deliver pharmaceutical species for different therapies.5,8,9,19 However, in many cases, such species show delivery difficulties because of the low chemical interaction between PMMA and the physiological media. To increase the hydrophilic character and the delivery capability of the medicament, chemical modification of the polymer matrix could be an interesting alternative. Because of the singular particles’ shape of PMMA in this application (microspheres), this chemical modification should be compatible with the preparation of microspheres for subsequent loading with the desired pharmacological molecule. In this sense, a polymerization reaction based on a suspension polymerization system was intended. The polar characteristic of methacrylic acid led to development of the suspension polymerization reaction in an inverse way, where the dispersant is the organic phase. An inverse suspension polymerization reaction is one of the most useful methodologies to obtain highly hydrophilic microsphere-shaped polymers.20,21 This hydrophilic character, together with the crosslinking structure inducted by MBA, allowed us to obtain microspheres nonsoluble in water with a high capacity for water swelling. Moreover, spherical shape maintenance, after water addition and lyophilization, is additional evidence of the crosslinking process induced by MBA during polymerization. Depending on the amount of HC added to each formulation, two different surface morphologies were obtained. The formulations loaded with higher HC content (M10–25 and M15–25) showed an irregular surface morphology attributed to a saturation phenomenon of the HC content inside the microparticle. The addition of 25 wt % of HC led to materials in which HC could not be totally absorbed by the microsphere and the excess remained precipitated on the particle’s surface after lyophilization. However, in formulations loaded with lower content of HC (M10-10 and M15-10), HC was almost completely absorbed inside the particle as shown in the drug delivery results. One of the most significant observations of the already polymerized PMAA microspheres was the high tendency for particle-particle interactions giving rise to aggregates. This adhesion capacity has been previously reported in adhesive materials for dentine defect reparation22 and promotes aggregation character. The establishment of these aggregates does not represent a disadvantage to protein delivery, as the main objective of the work was focused on the design of a polar system to increase water uptake and act as a carrier of hydrophilic species such as proteins. The chemical composition of the microspheres and microparticles was characterized by FTIR. The spectrum of M15–25 reveals the presence of HC as well as PMAA in the composition of the PMAA-HC-based microspheres. NAH and JOURNAL OF BIOMEDICAL MATERIALS RESEARCH B: APPLIED BIOMATERIALS | MONTH 2014 VOL 00B, ISSUE 00 7 TABLE III. Curing Parameter of Control and Self-Curing Formulations Loaded with HC Formulation PMMA FC-10 FC-25 FM10 FM15 FM10-10 FM10–25 FM15-10 FM15–25 Tpeak ( C) [s.d] 92.3 [1.1] 85.8[0.6] 75.4 [1.6] 62.0 [0.7] 65.8 [2.4] 64.8 [1.0] 63.4 [0.6] 61.1 [0.7] 67.3 [1.3] Tsetting (min) [s.d.] 14.4 13.1 13.1 15.1 13.5 14.1 14.8 14.5 13.4 [0.9] [0.9] [0.3] [0.1] [0.3] [0.6] [0.1] [2.4] [0.3] [s.d.]: standard deviation. C@O symmetric stretching signals at 3300 and 1643 cm21 (Ref. [23), respectively, together with the above-mentioned signals of PMAA, informs qualitatively about the presence of both components. TGA-analysis shows a different degradation process for PMAA and PMAA-HC microspheres, as expected. Two major behaviors were observed: (1) High content of char after a heating rate of 10 C/min and (2) a lower thermal resistance of the material when modified with collagen. The first observation corroborates the presence of poly(sodium methacrylate) instead of poly(methacrylic acid), and suggests the salt nature of the polymer as previously reported.24 The lower thermal resistance of these formulations is directly related with the lower thermal resistance of collagen compared with that of the cross-linked matrix of PMAA. DSC was performed to demonstrate the cross-linked nature of the microsphere structure. The presence of crosslinking linkage in the structure shifts the glass transition (Tg) to higher temperatures, even beyond degradation temperature, as an answer to a decrease in the polymer segment mobility. Brock et al. found that in the case of poly(metha- FIGURE 6. SEM microphotograph of the surface of the self-curing formulation FM10. 8  ET AL. FRANCO-MARQUES crylic acid), the addition of 0.43% of a crosslinking agent shifts Tg by 7 C (from 178.4 to 185.2 C).25 If the amount of crosslinker rises to 3%, Tg shifts to 222.2 C. In our case, the crosslinker (MBA) is added in a range of 10–15 wt %, because lower quantities led to partially soluble materials. This result, together with the maintenance of the spherical shape of the PMAA microsphere after hydration/lyophilization, confirms the cross-linked structure of the microspheres. ISO 5833 suggests a maximum peak temperature lower than 90 C to minimize possible necrotic effects in the surrounding tissue close to the implant. All formulations designed in this work showed a maximum curing temperature of the material lower than standard requirements. In fact, the formulations modified with PMAA microspheres showed a peak temperature close to 20 C lower than control formulations (PMMA, FC-10, and FC-25). The presence of interfaces PMMAPMAA decreases the peak temperature acting as a barrier to the diffusion of the exothermic polymerization heat from inner cement to external medium, diminishing the necrotic thermal risk.12 The formation of such interfaces is demonstrated by SEM microscopy as shown in Figure 6, where hollows around PMAA microspheres are formed after curing. The incorporation of the hydrophilic microparticles did not produce a significant increase of tsetting, and was, in all cases, close to 13–14 min. This behavior also helps the surgeon to properly handle the material, prior to vitrification, before its introduction in the bone cavity or defect. A compression test was carried out to determine the mechanical strength of these materials. Other mechanical tests, such as tensile or bending tests, have also been considered by researchers to determine the mechanical behavior of the material. In the case of the systems prepared in this work, only the compression test has been considered as compression is the most important stress that the material supports after vertebroplasty and kiphoplasty surgery. When the materials were tested in dry conditions, the formulations modified with PMAA-microspheres, PMAA-HC-microspheres or even HC powder showed a decrease in compression yield stress, compared with that of the PMMA control. The formation of the polymer– polymer interfaces led to a decrease in the mechanical strength of the material as mentioned above. Additionally, the presence of HC in the formulation also decreases this mechanical property. This decrease is also attributed to the important difference of molecular weight between HC and PMMA microparticles: 5103 and 200103 Da, respectively. HC could be considered a plasticizer of PMMA giving rise to a material with diminished mechanical properties. In the case of the formulations modified with the PMAAHC microspheres, the value of rc was close to ISO 5833 standard requirements (70 MPa) and was an acceptable result considering the high level of chemical structure modification of the cement. Previously, Franco-Marquès et al. reported an important decrease of mechanical properties of acrylic bone cements modified with polymer particles not compatible with PMMA, which is related with the existence of interphases between PMMA and the modifying polymer.12 In the formulations of this work, the addition of PMAAmicrospheres with a chemical structure close to that of SELF-CURING HYDROPHILIC BONE CEMENTS FOR PROTEIN DELIVERY ORIGINAL RESEARCH REPORT TABLE IV. Compressive Strength of Self-Curing Formulations Loaded with HC rc (MPa) Formulation PMMA FC-10 FC-25 FM10 FM15 FM10-10 FM10–25 FM15-10 FM15–25 a b a Dry 82.1 75.2 77.6 74.3 67.4 68.7 66.4 67.0 64.9 [1.5] [0.8] [1.9] [2.7] [0.8] [2.6] [0.9] [3.0] [0.6] Wetb 78.8 71.4 60.5 27.3 29.7 28.7 28.8 29.0 28.8 [1.8] [1.9] [0.9] [3.5] [1.3] [1.8] [1.4] [1.7] [1.2] Before immersion. After 15 days of immersion in PBS. PMMA led to a lower loss of material mechanical strength under dry conditions. The materials were also tested under wet conditions after 15 days of immersion in PBS at 37 C. In all cases, a drastic decrease in strength was observed and was attributed to the hydrophilic character of PMAA. This material is expected to interact more intensively with external aqueous media than PMMA, and give rise to a plasticizing behavior by the incorporation of water. Moreover, the penetration of the external medium is favored by the polymer–polymer interfaces, creating a network of micropores to increase water diffusion inside the material. The average loss of rc was in the range of 63.3–65.3% without significant differences between the formulations. The hydrophobic chemical structure of PMMA does not allow the absorption of water beyond 2–2.5 wt %. In the cases of FC-10 and FC-25, water uptake was slightly higher than PMAA cement because of the high solubility of HC in water at experimental conditions (pH 7.4). The dissolution of HC led to a porous material where water can diffuse, but cannot interact by any chemically compatible functionality. However, the addition of PMAA or PMAA-HC microspheres induces the absorption and retention of higher amounts of water due to the higher hydrophilic character and the cross- FIGURE 7. Water absorption (in vitro behavior) of formulations of acrylic bone cements modified with PMAA microspheres and PMAAHC microspheres. FIGURE 8. Controlled delivery profiles of hydrolyzed collagen from self curing formulations modified with PMAA-HC microspheres. linked nature of the microspheres. The water interaction is higher for the formulations FM-10 and FM-25 due to – COONa hydrophilic groups of PMAA that do not have to compete with the hydrophilic functionality of HC for water. Thus, the average of water absorption of FM-xx formulations is in the range of 48–54 wt % and that of FMxx-yy, loaded with HC, is in the range of 38–50 wt%. This higher capacity for water retention has been used to improve controlled delivery of HC from the formulations. In terms of HC delivery, the formulations based on plain PMMA showed a very low capacity for delivery during a very short period of time. This behavior suggests a delivery of the protein closer to the surface of the samples.7,8 A material with higher hydrophilicity is required to allow water diffusion inside the specimens, which is the case for materials where part of the PMMA was substituted by PMAA microspheres. This hydrophilic polymer drives the external aqueous medium towards the inner part of the samples, allowing the dissolution of the protein and its delivery to the surrounding tissue outside the sample. This behavior involves values of HC delivery close to 1.6 and 2.7 mg/mL, depending on the initial composition of HC in the material. Thus, a higher composition of HC, previously absorbed by the PMAA-microspheres, leads to higher profiles of release and longer delivery periods. FM15–25 is able to deliver 2.7 mg/mL of protein for more than 20 days. No significant influence on the composition of MBA was observed in the delivery of HC and was only considered as a structural agent to avoid the dissolution of the PMAA-microspheres in the aqueous media. The toxicity of the cements formulated with PMAA or PMAA-HC microspheres is not directly related with the already polymerized phase, but with the used monomers MMA, MAA, and MBA. Cell viability shows a minimum value after 1 day of incubation but a significantly higher value after the seventh day. In the case of the four formulations modified with PMAA-HC microspheres, cell viability rises from the first to seventh incubation day; cell viability in contact with FM10–25 rises significantly (F1,10 5 7.28; p < 0.05) from 64.6 6 4.3% to 74.6 6 8.5%. This result suggests a partial toxicity of the residual monomers during the first hours of incubation that subsequently is diminished JOURNAL OF BIOMEDICAL MATERIALS RESEARCH B: APPLIED BIOMATERIALS | MONTH 2014 VOL 00B, ISSUE 00 9 FIGURE 9. MTT results for determination of cell viability in presence of the extracts of PMMA and curing formulations modified with PMAA and PMAA-HC microspheres eluted in culture medium. The diagram includes the mean (n 5 6), the 0.95 confidence interval for the mean, and the ANOVA results with respect to TMX extracts samples (*p < 0.01, **p < 0.001). due to the experimental assay where the culture medium is substituted with fresh medium after proper periods of time. The same result would be obtained under in vivo conditions due to the dynamic behavior of fluids in the body. This characterization procedure simulates body behavior where the body fluid washes the implanted material and removes the secreted residues. After seven days of incubation, these toxicity levels were within the limits considered by ISO 10993 to determine the cytotoxicity of medical devices (70%). CONCLUSIONS Inverse suspension polymerization is a useful polymerization system for microsphere preparation of hydrophilic polymers for biomedical purposes. PMAA-based microspheres have been prepared and characterized as an interesting carrier for protein delivery. PMAA-HC-based microspheres improve protein delivery from two different points of view: increase of protein content delivery to the surrounding tissue and maintenance of its release during a longer period of time compared with common formulations of cements. Because of this behavior, a decrease in compression stress has been found to increase the macromolecule mobility derived from the plastifying effect induced by absorbed water. Finally, a decrease of cell viability of human osteoblasts has been obtained as a result of the reacting monomers’ toxicity. This negative reaction is diminished with the incubation time because of the washing effect related with the simulated testing technique; the toxicity after the seventh incubation day is within the limits assumed by ISO 10993. ACKNOWLEDGMENTS The authors thank the financial support to MEC-Spain (MAT2010-18155) and Protein for HC supply. Franco-Marquès thanks UdG for doctorate grant (BR-07/05). REFERENCES 1. Charnley J. Anchorage of the femoral head prosthesis to the staft of the femur. J Bone Jt Surg B 1964;42:28–30. 2. Charnley J. Fracture of femoral prosthesis in total hip replacement. Clin Orthop Rel Res 1975;111:105–120. 10  ET AL. FRANCO-MARQUES 3. Sweeny WT. Acrylic resins in prosthetic dentistry. Dent Clin North Am 1958;2:593–601. zquez B, Deb S, San Roman J. Eugenol derivatives 4. Rojo L, Va immobilized in auto-polymerizing formulations as an approach to avoid inhibition interferences and improve biofunctionality in dental and orthopedic cements, Acta Biomaterialia 2009;5:1616–1625. 5. Lopes P, Corbellini M, Ferreira BL, Almeida N, Fredel M, Fernandes MH, Correia R. New PMMA-co-EHA glass-filled composites for biomedical applications: Mechanical properties and bioactivity. Acta Biomaterialia 2009;5:356–362. 6. Nelson DA, Barker ME, Hamlin, BH. Thermal effects of acrylic bone cements at bone sites. Int J Hyperthermia 1997;13:287–306. 7. Mendez JA, Fernandez M, Gonzalez-Corchon A, Salvado M, Collia F, de Pedro JA, Levenfeld BL, Lopez-Bravo A, Vazquez B,San Roman J. Injectable self-curing bioactive acrylic-glass composites charged with specific anti-inflammatory/analgesic agent. Biomaterials 2004;25:2381–2392. 8. Mendez JA, Abraham GA, Fernandez MD, Vazquez B, Roman JS. Self-curing acrylic formulations containing PMMA/PCL composites: Properties and antibiotic release behavior. J Biomed Mater Res 2002;61:66–74.  E. Anti9. Miola M, Bistolfi A, Valsania MC, Bianco C, Fucale G, Verne biotic-loaded acrylic bone cements: An in vitro study on the release mechanism and its efficacy. Mater Sci Eng C 2013;33:3025–3032.  F, Aina V, Gatti S, Cannas M. Effect of vitamin E addition to 10. Reno poly(D,L)-lactic acid on surface properties and osteoblast behaviour. Biomaterials 2005;26:5594–5599. 11. Mendez JA, Aguilar MR, Abraham GA, Vazquez B, Dalby M, Di Silvio L, San Roman J. New acrylic bone cements conjugated to vitamin E: Curing parameters, properties, and biocompatibility. J Biomed Mater Res 2002;2:299–307. 12. Franco-Marques E, Mendez JA, Girones J, Ginebra MP, Pelach MA. Evaluation of the influence of the addition of biodegradable polymer matrices in the formulation of self-curing polymer systems for biomedical purposes. Acta Biomaterialia 2009;5:2953–2962. 13. Caracciolo PC, Sanz Pita C, Abraham GA, Mendez JA, Molera JG. Synthesis, characterization and applications of amphiphilic elastomeric polyurethane networks in drug delivery. Polym J 2013;45:331–338. ~o 14. Palomino B, Jimenez L, Ferrero A. El dolor lumbar en el an  n 2010;44:69–81. 2009. Rehabilitacio 15. Forriol F. Proteinas morfogenicas oseas y su aplicacion clinica. ~ ola de Cirugia Ortopedica y Traumatologia54, SupRevista Espan plement 2010;1:2–10. 16. Chen DI, Zhao M, Mundy GR. Bone Morphogenetic Proteins. Growth Factors 2004;22:233–241. 17. Abdalla MA, Suhaimi F, Das S, Shuid AN, Mohamed Z, Soelaiman IN. Expression of TGF-B1 in the blood during fracture repair in an estrogen-deficient rat model. Clinics 2011;66:2113–2119. 18. Kim B, Peppas NA. Analysis of molecular interactions in P(MAAg-EG) hydrogels. Polymer 2003;44:3701–3707. 19. Luginbuehl V, Meinel L, Merkle HP, Gander B. Localized delivery of growth factors for bone repair. Eur J Pharma Biopharma 2004; 58:197–208. 20. Renard B, McKenna TF. Kinetics of polymerization of partially neutralized acrylic acid in inverse suspensions. Macromol Symp 2000;150:251–257. 21. Dowding PJ, Vincent B. Suspension polymerization to form polymer beads. Colloids Surf A Physicochem Eng Asp 2000;161:259–269. 22. Solhi L, Atai M, Nodehi A, Imani M, Ghaemi A, Khosravi K. Poly(acrylic acid) grafted montmorillonite as novel fillers for dental adhesives: Synthesis, characterization and properties of the adhesive. Dental Mater 2012;28:369–377. 23. Muyonga JH, Cole CGB, Duodu KG. Fourier transform infrared (FTIR) spectroscopic study of acid soluble collagen and gelatin from skins and bones of young and adult Nile perch (Lates niloticus). Food Chem 2004;86:325–332. 24. Xue T, Wilkie CA. Thermal degradation of Poly(styrene-g-acrylonitrile). Polym Degrad Stabil 1997;56:109–113. 25. Brock J, Creecy C, McGinity JW, Peppas NA. Synthesis and properties of lightly crosslinked poly((meth)acrylic acid) microparticles prepared by free radical precipitation polymerization. Polym Bull 2006;17:11–20. SELF-CURING HYDROPHILIC BONE CEMENTS FOR PROTEIN DELIVERY